ALUMINUM RECYCLING AND PROCESSING FOR ENERGY CONSERVATION AND SUSTAINABILITY

ALUMINUM RECYCLING ANDPROCESSING FOR ENERGY

CONSERVATION AND SUSTAINABILITY

JOHN A.S. GREENEDITOR

ASM International®

Materials Park, Ohio 44073-0002www.asminternational.org

Copyright © 2007by

ASM International®

All rights reserved

No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by anymeans, electronic, mechanical, photocopying, recording, or otherwise, without the written permission of thecopyright owner.

First printing, December 2007

Great care is taken in the compilation and production of this book, but it should be made clear that NO WAR-RANTIES, EXPRESS OR IMPLIED, INCLUDING, WITHOUT LIMITATION, WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, ARE GIVEN IN CONNECTIONWITH THIS PUBLICATION. Although this information is believed to be accurate by ASM, ASM cannot guaran-tee that favorable results will be obtained from the use of this publication alone. This publication is intended for useby persons having technical skill, at their sole discretion and risk. Since the conditions of product or material useare outside of ASM’s control, ASM assumes no liability or obligation in connection with any use of this informa-tion. No claim of any kind, whether as to products or information in this publication, and whether or not based onnegligence, shall be greater in amount than the purchase price of this product or publication in respect of whichdamages are claimed. THE REMEDY HEREBY PROVIDED SHALL BE THE EXCLUSIVE AND SOLE REM-EDY OF BUYER, AND IN NO EVENT SHALL EITHER PARTY BE LIABLE FOR SPECIAL, INDIRECT ORCONSEQUENTIAL DAMAGES WHETHER OR NOT CAUSED BY OR RESULTING FROM THE NEGLI-GENCE OF SUCH PARTY. As with any material, evaluation of the material under end-use conditions prior tospecification is essential. Therefore, specific testing under actual conditions is recommended.

Nothing contained in this book shall be construed as a grant of any right of manufacture, sale, use, or repro-duction, in connection with any method, process, apparatus, product, composition, or system, whether or notcovered by letters patent, copyright, or trademark, and nothing contained in this book shall be construed as adefense against any alleged infringement of letters patent, copyright, or trademark, or as a defense againstliability for such infringement.

Comments, criticisms, and suggestions are invited, and should be forwarded to ASM International.

Prepared under the direction of the Aluminum Advisory Group (2006–2007), John Green, Chair.

ASM International staff who worked on this project include Scott Henry, Senior Manager of Product and ServiceDevelopment; Steven R. Lampman , Editor; Ann Briton, Editorial Assistant; Bonnie Sanders, Manager of Pro-duction; Madrid Tramble, Senior Production Coordinator; Patti Conti, Production Coordinator; and KathrynMuldoon, Production Assistant.

Library of Congress Control Number: 2007932444ISBN-13: 978-0-87170-859-5

ISBN-10: 0-87170-859-0SAN: 204-7586

ASM International®

Materials Park, OH 44073-0002www.asminternational.org

Printed in the United States of America

Preface....................................................................................................................vii

Chapter 1 Life-Cycle Engineering and Design ......................................................1Life-Cycle Analysis Process Steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2Application of Life-Cycle Analysis Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5Case History: LCA of an Automobile Fender . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Chapter 2 Sustainability—The Materials Role ....................................................15Some History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17The Materials Role in Industrial Ecology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19The U.S. Government Role—Organizational . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23The U.S. Government Role—Technical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24The Role of Professional Societies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28Summary and Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

Chapter 3 Life-Cycle Inventory Analysis of the North AmericanAluminum Industry ................................................................................................33

Life-Cycle Inventory Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35Inventory Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40Primary Aluminum Unit Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46Secondary Aluminum Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51Manufacturing Unit Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53Results by Product System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56Interpretation of LCI Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

Chapter 4 Life-Cycle Assessment of Aluminum: Inventory Data for the Worldwide Primary Aluminum Industry ............................................67

Data Coverage, Reporting, and Interpretation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67Data Quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68Unit Processes and Results by Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73Aluminum Life-Cycle Assessment with Regard to Recycling Issues . . . . . . . . . . . . . . . . . 83

Chapter 5 Sustainable Development for the Aluminum Industry ........................91Recycling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92Perfluorocarbon Emissions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94Fluoride Emissions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

Contents

Energy Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97Aluminum in Transportation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99Natural Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

Chapter 6 Material Flow Modeling of Aluminum for Sustainability ..................103Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103Key Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

Chapter 7 Recycling of Aluminum ....................................................................109Industry and Recycling Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110Recyclability of Aluminum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114The Recycling Loop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115Technological Aspects of Aluminum Recycling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116Process Developments for Remelting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118Developing Scrap Streams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119Can Recycling Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122Automobile Scrap Recycling Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125Building and Construction Recycling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128Aluminum Foil Recycling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128Impurity Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129Molten Metal Handling and Safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130

Chapter 8 Identification and Sorting of Wrought Aluminum Alloys ..................135Sources of Aluminum Raw Material for Alloy Sorting . . . . . . . . . . . . . . . . . . . . . . . . . . 136Improving Recovery for Wrought and Cast Fractions . . . . . . . . . . . . . . . . . . . . . . . . . . 136Pilot Processes for Improved Wrought Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

Chapter 9 Emerging Trends in Aluminum Recycling ..........................................147Objectives and Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148The Nature of Recycled Metal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148Recycling Aluminum Aerospace Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150Alloys Designed for Recycling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152Developing Recycling-Friendly Compositions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153Conclusions and Looking Ahead . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154

Chapter 10 U.S. Energy Requirements for Aluminum Production: Historical Perspective, Theoretical Limits and New Opportunities ......................................157

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158Aluminum Production and Energy Consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160Methodology, Metrics, and Benchmarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164Aluminum Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168Primary Aluminum Raw Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171Primary Aluminum Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178Advanced Hall-Heroult Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191Alternative Primary Aluminum Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197Secondary Aluminum (Recycling) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204Aluminum Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208

iv

Appendixes ..........................................................................................................223Appendix A Energy Intensity of Materials Produced in the United States . . . . . . . . . . . . 223Appendix B Energy Values for Energy Sources and Materials . . . . . . . . . . . . . . . . . . . . . 225Appendix C Hydroelectric Distribution and Electrical Energy Values . . . . . . . . . . . . . . . 229Appendix D Emission Data and Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231Appendix E U.S. Energy Use by Aluminum Processing Area . . . . . . . . . . . . . . . . . . . . . 237Appendix F Theoretical Energy Data and Calculations . . . . . . . . . . . . . . . . . . . . . . . . . 245Appendix G Aluminum Heat Capacity and Heat of Fusion Data . . . . . . . . . . . . . . . . . . 251Appendix H Impact of Using Different Technologies on Energy

Requirements for Producing Aluminum . . . . . . . . . . . . . . . . . . . . . . . . . . . 253Appendix I Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257

Index ....................................................................................................................261

v

ALUMINUM AND ENERGY—energy and aluminum—the two have been intimately linked sincethe industry started in 1886. That was when both Charles Martin Hall, in the United States, and PaulHeroult, in France, working independently, almost simultaneously discovered an economical processto produce aluminum from a fused salt using electrolysis. From this relatively recent discovery, theuse of aluminum has grown rapidly and overtaken other older metals, such as copper, tin, and lead. Itis now the second most widely used metal after steel.

Although the actual chemistry of the winning of aluminum from its oxide, alumina, has notchanged greatly since 1886, the growth of the industry has brought about huge changes in productionscale and sophistication. Also, there have been considerable reductions in the amount of energy usedper unit of production. However, the production of aluminum is still energy-intensive, and the smelt-ing process requires approximately 15 MWh per metric ton of aluminum production. For the UnitedStates, aluminum production consumes approximately 2% of the total industrial energy used.

For most of the 120 years of aluminum production, the growth of aluminum and energy productionfrom hydroelectric sources were essentially symbiotic in nature. Aluminum production requires largequantities of stable and low-cost power, while hydroelectric projects need steady baseline users toensure the viability of a hydroelectric project. Nowhere was this linkage between the aluminum indus-try and hydroelectric power producers better demonstrated than in the Pacific Northwest of the UnitedStates. There is now a concentration of both smelters and hydroelectric dams in the Columbia Riverbasin. Although this mutually beneficial relationship was tested on occasion by market recession,drought, or lack of sufficient snowpack, the linkage persisted for several decades. It was not until 2000and 2001 that the severe economic recession, coupled with the extreme energy crisis in California,caused the linkage between the industry and hydroelectric power producers to finally rupture. At thistime, several aluminum smelters “mothballed” their operations, and power producers discontinued theirsupply arrangements with the aluminum smelters. About this same time, the importance of recycledsecondary aluminum grew, and, in fact, in 2002, the percentage of recycled metal exceeded the primarysmelted metal in the total U.S. metal supply for the first time.

Recycling of aluminum is vitally important to the sustainability of the aluminum industry. Whenthe metal has been separated from its oxide in the smelting process, it can be remelted and recycledinto new products numerous times, with only minimal metal losses each time. In fact, as the life-cycle and sustainability studies discussed in Chapters 3, 4, and 5 indicate, the recycling of aluminumsaves ~95% of the energy used as compared to making the metal from the original bauxite ore. Thisenormous energy savings has accounted for the continuing growth of the secondary industry and hasled to the concept that aluminum products can be considered as a sort of “energy bank.” The energyembedded in aluminum at the time of smelting remains in an aluminum product at the end of its use-ful life and effectively can be recovered through the recycling process. Probably the best example ofthis is the ubiquitous aluminum beverage can that, on average, is recovered, recycled, and fabricatedinto new cans that are put back on the supermarket shelves in approximately 60 days!

With an increasing awareness of environmental and climate-change issues in the public arena, it isconsidered that the publication of this sourcebook will be most timely. The purpose of this book is toprovide a comprehensive source for all aspects of the sustainability of the aluminum industry. It is

vii

Preface

anticipated that issues of sustainability will become increasingly important in the next couple ofdecades as individuals, companies, various agencies, governments, and societies in general strive toseek a responsible balance between using materials to maintain and improve living standards whilenot despoiling the planet of its limited mineral and material resources for future generations.

This publication is a collection of basic factual information on the modeling of material flow in thealuminum industry, the life-cycle materials and energy inputs, and the products, emissions, andwastes. The energy savings involved with recycling, various scrap-sorting technologies, and futureenergy-saving opportunities in aluminum processing are outlined. Finally, the positive impact of thegrowing use of lightweight aluminum in several segments of the transportation infrastructure and itsbenefit on greenhouse gas production is also highlighted. This book should provide much-neededbasic information and data to reduce speculation and enable fundamental analysis of complex sus-tainability issues associated with the aluminum industry.

Regarding the specific contents of this book, Chapter 1 is a brief introduction to the concept oflife-cycle analysis by Hans Portisch and coworkers. Portisch has pioneered in the field of life-cyclestudies and has helped to establish many of the life-cycle protocols developed by the EuropeanUnion and International Standardization Organization (ISO) for working groups. Chapter 1, entitled“Life-Cycle Engineering and Design,” is an opportunity for the reader to become familiar with theconcept of life-cycle analysis and its terminology that will be important in appreciating several ofthe subsequent chapters.

Chapter 2, entitled “Sustainability—The Materials Role,” by Lyle Schwartz, is probably the realintroduction to the complex subject of sustainability. This chapter was first presented by the author asthe Distinguished Lecture in Materials and Society in 1998. The chapter sets out the case for sustain-ability and life-cycle analysis and is introductory in nature. The huge worldwide growth of the auto-mobile is used to illustrate the enormity of the materials and sustainability issues facing the technicalcommunity and society in general. The chapter traces some recent history and proposes several pathsfor future direction, such as:

• Cleaner processing• The development of alternative materials• Dematerialization, or the use of less material per capita to accomplish the necessary material

requirements• Reuse and recycling

The latter, recycling, is of course one of the key attributes of aluminum. This chapter also contrastsother materials, such as magnesium, advanced steels, and polymer composites, with aluminum in thecontext of reducing the weight of automobiles to enhance fuel efficiency. The chapter ends with a callto action by the professional societies and the individual materials scientists. It is indeed a rallying callfor materials responsibility!

The Life-Cycle Inventory for the North American Aluminum Industry, discussed in Chapter 3, repre-sents the original (year of 1995) study of the industry and is probably still the most comprehensive. Ithas since become the basis for future studies by the International Aluminum Institute (IAI). The studywas conducted in response to a request from Chrysler, Ford Motor Company, and General Motorsunder the United States Automotive Materials Partnership. This automotive materials partnership wasenabled by the PNGV program established by the U.S. Government. Recently, the PNGV activitieshave transitioned to FreedomCAR and its emphasis has been expanded to include other light materi-als, e.g. Mg, Ti and composites, as well as aluminum. The purpose of this study was to provide theparticipating companies with detailed life-cycle inventories of the various processes within thealuminum product life cycle. This information provides a benchmark for improvements in the man-agement of energy, raw material use, waste elimination, and the reduction of air and water emissions.Although this is titled a North American study, it was in fact global in reach due to the internationaloperations of the 13 companies taking part. The study incorporated data from 15 separate unitprocesses located in 213 plants throughout North and South America, Africa, Australia, Europe, andthe Caribbean. The results were tabulated by an independent contractor (Roy F. Weston, Inc.) andwere peer reviewed by a distinguished panel of experts prior to publication in accord with ISOmethodologies. One excellent feature of this chapter is the graphical presentation of the results. Forexample, for any particular process, such as aluminum extrusion or cold rolling, it is possible to see at

viii

a glance what the materials and energy inputs are and what are the products, air emissions, and wastesgenerated to that stage in the fabrication process.

Following the publication of the comprehensive life-cycle study discussed in Chapter 3, the leadersof the aluminum industry vested in the IAI, based in London, the responsibility of maintaining andextending the database to include significant areas of aluminum production that were not included inthe initial study, namely Russia and China. Also, the IAI was requested to develop several globalperformance indicators and to track these indicators toward key sustainability goals agreed upon bythe international industry. The global performance indicators chosen include such items as primaryproduction; electrical energy used for production; emissions of greenhouse gases during electrolysis;specific emissions of perfluorocarbon gases, which are potent global-warming gases; consumption offluoride materials; as well as injury rates and loss time severity rates. The considerable progress thatthe industry has made toward achieving many of these voluntary objectives is described in Chapter 5.The addendum report, updated to the end of 2005, illustrates quantitatively the industry’s progresstoward the 12 voluntary objectives. Significant progress has been achieved and documented.

The sixth chapter, entitled “Material Flow Modeling of Aluminum for Sustainability,” by KennethMartchek of Alcoa, describes the development of a global materials flow model. Annual statisticaldata since 1950 from all the significant market segments have been combined with the most recentlife-cycle information from the IAI to develop this global model. The model has demonstrated goodagreement between estimated and reported worldwide primary production over the past threedecades. Probably one of the most interesting features of the chapter is the table citing the worldwidecollection rates and recycle rates for each market segment. The model also demonstrates that approxi-mately 73% of all aluminum that has ever been produced is contained in products that are currentlyin service—surely a good testament to the recyclability and versatility of the metal!

Chapter 7 is devoted to a detailed discussion of the recycling of aluminum. As noted previously,recycling is a critical component of the sustainability of aluminum because of the considerableenergy savings and the equivalent reduction in emissions from both energy and metal production.The chapter starts with a discussion of the recycling process and reviews the steps to remelt, purifythe molten metal, and fabricate new products. The chapter also contains a discussion of the life-cycle trends in each major market area and how these factors impact recyclability. For example, onesignificant development that is discussed is the growing importance of automotive scrap. It is nowestimated from modeling approaches that automotive scrap became more dominant than the tradi-tional recycling of beverage containers at some stage during the 2005 to 2006 time period. Thistransition has occurred because of the marked increases of aluminum being used in automotives toenhance fuel efficiency, the fact that auto shredders are now commonly used, and shredder scrapcan be economically sorted on an industrial scale. The transition has also occurred because the ratesfor the collection of can scrap have recently declined from the peak values of 1997, when ~67% ofall cans were bought back by the industry, to the time of writing, when the recycling rates arehovering around 50%.

One dominant issue in the recycling of automotive scrap is the control of impurities, especiallyiron and silicon, that inevitably build up during the recycling process. Cast aluminum alloys, withtheir higher silicon content, are better able to tolerate this increase of impurity content than wroughtalloys. Future trends, potential solutions, and research directions to resolve this issue of impuritycontrol are outlined in this chapter. Also, the chapter mentions the potential impact of governmentregulations in European Union countries that now mandate that vehicles be 95% recyclable by theyear 2015. Chapter 7 concludes with a brief discussion of the safety issues related to melting andcasting aluminum.

Aluminum products are formed from an extremely wide array of alloys. These range from the softalloys used in foil and packaging material, to the intermediate alloys used in the construction of boatsand trains, to the hard alloys used in aircraft and aerospace applications. It is inevitable that someamount of all these alloys will end up in the products from the industrial shredder. Accordingly, toachieve the optimum recycling, it is most economical to identify and separate scrap and to reuse thespecific alloying elements in the most advantageous manner. This is why the recent advances in scrapsorting by Adam Gesing and his coworkers at Huron Valley Steel Corporation are so significant. Thesedevelopments are detailed in Chapter 8. This chapter is a comprehensive discussion of the complexi-ties of automotive alloys and recycling issues. The chapter provides a state-of-the-art description of

ix

sorting technologies and demonstrates alloy sorting by color, x-ray absorption, and laser-inducedbreakdown spectroscopy technology. Color sorting of cast material from wrought alloys is now fullyestablished on a commercial scale, and LIBS sorting has become commercially viable in the pastcouple of years.

The next chapter, Chapter 9, explores some of the emerging trends in municipal recycling from theperspective of the operation of a municipal recycling facility. More importantly, the chapter discussesat length the issue of impurities and alloy content and how best to assimilate the recycled materialstream into the existing suite of aluminum alloys. At present, sorted material can contain a widerange of elemental content, and this can modify and impact the physical, chemical, and mechanicalproperties of recycled alloys. This chapter, contributed by Secat and the University of Kentucky withpartial support of the Sloan Foundation, suggests several routes to optimize the economical and prop-erty benefits achieved through recycling. It also explores the development of aluminum alloys thatare more tolerant of recycling content, that is, recycling friendly alloys.

The final chapter of the sourcebook, Chapter 10, was originally prepared by BCS, Inc. for theU.S. Department of Energy in Washington, D.C., in February 2003 but has since been updatedwith the latest available data as of early 2007. This chapter looks at the whole production systemfor aluminum, from the original bauxite ore, through refining of alumina and smelting of alu-minum, to various rolling, extrusion, and casting technologies. From an historical perspective, thechapter explores the energy requirements for aluminum production. The theoretical energy limitsfor each process step are compared to the actual current industry practice, and new opportunitiesfor saving energy are highlighted. The original report was commissioned as the baseline study ofthe industry by the Department of Energy and contains extensive discussions of potential advancesin aluminum processing and fabrication. For example, the potential of wettable cathodes, inertanode technology, carbothermic reduction, and various melting and fabrication technologies arediscussed at length. Finally, the chapter is most valuable because it is supported by numerous ap-pendixes with almost 50 years of industry data and statistics. Much of the energy data used for theenergy calculations evaluating competing technologies is drawn from the industry life-cycle out-lined in Chapter 3.

It is hoped that this compilation of published material can be a contribution to the sustainabilitydebate and, specifically, can help to increase the understanding about the sustainability and recycla-bility of aluminum. The availability of credible information can only help sustain rational debateand the development of optimal actions and policies for the future.

Much progress has been made in recent years, although a lot still remains to be achieved. At thetime of writing, the Baltimore Sun newspaper (dated January 24, 2007), in an article entitled “PlaneTrash,” refers to a report by the National Resources Defense Council that says that the aviation industry is pitching enough aluminum cans each year to build 58 Boeing 747s! This is blamed on alack of understanding and on a mishmash of conflicting regulations and procedures at various air-ports around the country. While many airports are in fact recycling much of their trash and therebyreducing operating costs and landfill fees, many airlines and airports are not doing so. Under thepresent conditions and with the potential gains of energy and environmental emissions that areavailable through recycling of beverage cans, this situation seems remarkably shortsighted, espe-cially when all cans are collected before the termination of a flight! On the other hand, enormousprogress has been made in recycling and sustainability. Especially, it is noteworthy that computermodels now indicate that the aluminum industry will become “greenhouse gas neutral” by the year2020. This is indicated by the fact that the potential savings in emissions of greenhouse gases fromthe transportation use of aluminum for lightweighting of vehicles and increased fuel efficiency isgrowing at a faster rate than the emissions from the production of the aluminum itself. For all of uswith children and grandchildren, this is indeed a hopeful sign.

John Green, Ellicott City, MDJanuary 2007

x

ENVIRONMENTAL CONSIDERATIONSplay an increasingly important role in design anddevelopment efforts of many industries. “Cradle-to-grave” assessments are being used not onlyby product designers and manufacturers but alsoby product users (and environmentalists) to con-sider the relative merits of various availableproducts and to improve the environmental acceptability of products.

Life-cycle engineering is a part-, system-, orprocess-related tool for the investigation ofenvironmental parameters based on technicaland economic measures. This chapter focuseson life-cycle engineering as a method for evalu-ating impacts, but it should be noted that othertechniques also can be used to analyze the life-cycle costs of products (e.g., see the article“Techno-Economic Issues in Materials Selec-tion” in Materials Selection and Design, Volume20, ASM Handbook, 1997.

Products and services cause different environ-mental problems during the different stages oftheir life cycle. Improving the environmentalperformance of products may require thatindustry implement engineering, process, andmaterial changes. However, a positive changein one environmental aspect of a product (suchas recyclability) can influence other aspectsnegatively (such as energy usage). Therefore, amethodology is required to assess trade-offsincurred in making changes. This method iscalled life-cycle analysis or assessment (LCA).

Life-cycle analysis aims at identifying im-provement possibilities of the environmentalbehavior of systems under consideration bydesigners and manufacturers. The whole lifecycle of a system has to be considered. There-fore, it is necessary to systematically collect andinterpret material and energy flows for all rele-vant main and auxiliary processes (Fig. 1.1).

Life-cycle analysis methods have been devel-oped by governmental, industrial, academic,and environmental professionals in both NorthAmerica and Europe. Technical documents onconducting LCA have been published by theSociety of Environmental Toxicology andChemistry (SETAC), the U.S. EnvironmentalProtection Agency (EPA), the Canadian Stan-dards Association (CSA), the Society for thePromotion of LCA Development (SPOLD), andvarious practitioners.

For meaningful comparisons of the life-cycleperformance of competing and/or evolvingproduct systems, it is important that associatedLCAs be conducted consistently, using thesame standards. Although the common metho-dologies developed by SETAC, EPA, CSA, andSPOLD are a step in that direction, a broad-based international standard is needed. Such aneffort is being undertaken by ISO 14000 series(TC207).

Life-cycle thinking and techniques can beapplied to products, processes, or systems invarious ways: it can help assess life-cycle eco-nomic costs (LCAecon), social costs (LCAsoc), orenvironmental costs (LCAenv).

A primary objective of LCA is to provide atotal life-cycle “big-picture” view of the interac-tions of a human activity (manufacturing of aproduct) with the environment. Other majorgoals are to provide greater insight into theoverall environmental consequences of industrial

CHAPTER 1

Life-Cycle Engineering and Design*

*Adapted from an article by Hans H. Portisch, Krupp VDMAustria GmbH (Committee Chair), with contributions fromSteven B. Young, Trent University; John L. Sullivan, FordMotor Company; Matthias Harsch, Manfred Schuckert, andPeter Eyerer, IKP, University of Stuttgart; and Konrad Saur,PE Product Engineering, which was published in MaterialsSelection and Design, Volume 20, ASM Handbook, ASMInternational, 1997, p 96–104.

Aluminum Recycling and Processing for Energy Conservation and Sustainability John A.S. Green, editor, p 1-14 DOI: 10.1361/arpe2007p001

Copyright © 2007 ASM International® All rights reserved. www.asminternational.org

activities and to provide decision makers with aquantitative assessment of the environmentalconsequences of an activity. Such an assessmentpermits the identification of opportunities forenvironmental improvement.

Life-Cycle Analysis Process Steps

Life-cycle analysis is a four-step process;each of these steps is described in detail as fol-lows. The process starts with a definition of thegoal and scope of the project; because LCAsusually require extensive resources and time,this first step limits the study to a manageableand practical scope. In the following steps of thestudy, the environmental burdens (includingboth consumed energy and resources, as well asgenerated wastes) associated with a particularproduct or process are quantitatively invento-ried, the environmental impacts of those bur-dens are assessed, and opportunities to reducethe impacts are identified.

All aspects of the life cycle of the product areconsidered, including raw-material extractionfrom the earth, product manufacture, use, recy-cling, and disposal. In practice, the four steps ofan LCA are usually iterative (Fig. 1.2).

Step 1: Goal Definition and Scoping. Inthe goal definition and scoping stage, the pur-poses of a study are clearly defined. Subse-quently, the scope of the study is developed,

which defines the system and its boundaries,the assumptions, and the data requirementsneeded to satisfy the study purpose. For rea-sons of economy and brevity, the depth andbreadth of the study is adjusted, as required, toaddress issues regarding the study purpose.Goal definition and project scope may need tobe adjusted periodically throughout the courseof a study, particularly as the model is refinedand data are collected.

Also during this stage, the functional unit isdefined. This is an important concept because itdefines the performance of a product in meas-ured practical units and acts as a basis forproduct system analysis and comparison tocompeting products. For example, the carrying

2 / Aluminum Recycling and Processing for Energy Conservation and Sustainability

Goal definitionand scoping

Inventory(data collection)

Improvementassessment(companyresponse)

Impactassessment

(environmentalevaluation)

Specification • Technical • Economic • Ecological

Balances • Materials • Waste • Energy • Emissions • Sewage

Processing

Disposal

Impact assessment and valuation

Improvement

Exploitation

Synthesis

Recycling

Utilization

Fig. 1.1 Factors considered in the life-cycle engineering approach. Source: Ref 1.1

Fig. 1.2 The life-cycle assessment triangle. Source: Ref 1.2

Chapter 1: Life-Cycle Engineering and Design / 3

capacity of a grocery bag may be a sensiblefunctional unit.

Finally, the quality of the life-cycle data mustbe assessed in order to establish their accuracyand reliability. Typically, factors such as dataage, content, accuracy, and variation need to bedetermined. Clearly, data quality affects thelevel of confidence in decisions that are basedon study results.

Step 2: Inventory Analysis. The secondstage of LCA is a life-cycle inventory (LCI). Itis in this stage that the various inputs and out-puts (energy, wastes, resources) are quantifiedfor each phase of the life cycle. As depicted inFig. 1.3, systems boundaries are defined in sucha way that the various stages of the life cycle ofa product can be identified. The separation ofburdens (inputs and outputs) for each stagefacilitates improvement analysis.

For the purposes of LCI, a “product” should bemore correctly designated as a “product system.”First, the system is represented by a flowchartthat includes all required processes: extractingraw materials, forming them into the product,using the resulting product, and disposing ofand/or recycling it. The flowchart is particularlyhelpful in identifying primary and ancillarymaterials (such as pallets and glues) that arerequired for the system. Also identified are thesources of energy, such as coal, oil, gas, or elec-tricity. Feedstock energies, which are defined ascarbonaceous materials not used as fuel, arealso reported.

After system definition and materials and en-ergy identification, data are collected and model

calculations performed. The output of an LCI istypically presented in the form of an inventorytable (an example is shown in Table 1.1), accom-panied by statements regarding the effects ofdata variability, uncertainty, and gaps. Alloca-tion procedures pertaining to co-product gener-ation, re-cycling, and waste treatment processesare clearly explained.

Step 3: Impact Assessment and Interpreta-tion. Impact assessment is a process by whichthe environmental burdens identified in theinventory stage of an LCA are quantitatively orqualitatively characterized as to their effectson local and global environments. Morespecifically, the magnitude of the effects onecological and human health and on resourcereserves is determined.

Life-cycle impact assessment is, at this time,still in an early phase of development. Al-though some impact assessment methods havebeen advanced as either complete or partialapproaches, none has been agreed upon. Never-theless, an approach to impact analysis, knownas “less is better,” is typically practiced. Withthis approach, process and product changes aresought that reduce most, if not all, generatedwastes and emissions and consumed resources.However, situations in which such reductionsare realized are not yet typical. Usually, achange in product systems is accompanied bytrade-offs between burdens, such as moregreenhouse gases for fewer toxins. A fullydeveloped impact analysis methodology wouldhelp in the environmental impact assessment ofsuch cases.

Inputs Outputs

Materials production

Usable products

Water effluents

Air emissions

Solid wastes

Other impacts

Product manufacturing

Energy

Raw materials

Product use

Product disposal

System boundary

Fig. 1.3 Generalized system boundaries for a life-cycle inventory of a generic product. Source: Ref 1.2


As advanced by SETAC, impact analysiscomprises three stages:

• Classification: In this stage, LCI burdens areplaced into the categories of ecologicalhealth, human health, and resource deple-tion. Within each of these categories, theburdens are further partitioned into subcate-gories, for example, greenhouse gases, acidrain precursors, and toxins of various kinds.Some burdens may fall into several cate-gories, such as sulfur dioxide, whichcontributes to acid rain, eutrophication, andrespiratory-system effects. Environmentalburdens are sometimes called stressors,which are defined as any biological, chemi-cal, or physical entity that causes an impact.

• Characterization: In the characterizationstep of impact assessment, the potential im-pacts within each subcategory are estimated.Approaches to assessing impacts include re-lating loadings to environmental standards,modeling exposures and effects of the bur-dens on a site-specific basis, and developingequivalency factors for burdens within animpact subcategory. For example, all gaseswithin the global-warming category can beequated to carbon dioxide, so that a total ag-gregate “global-warming potential” can becomputed.

• Valuation: In the valuation step of impactassessment, impacts are weighted and com-pared to one another. It should be noted thatvaluation is a highly subjective process withno scientific basis. Further, attaching weight-ing factors to various potential impacts forcomparison purposes is intrinsically difficult.For example, what is more important: therisk of cancer or the depletion of oil reserves?Who would decide this? Because a consen-sus on the relative importance of differentimpacts is anticipated to be contentious, awidely accepted valuation methodology isnot expected to be adopted in the foreseeablefuture, if ever.

It is important to recognize that an LCAimpact assessment does not measure actualimpacts. Rather, an impact in LCA is generallyconsidered to be “a reasonable anticipation of aneffect,” or an impact potential. The reason forusing impact potentials is that it is typically dif-ficult to measure directly an effect resulting fromthe burdens of a particular product. For example,are the carbon dioxide emissions of any individ-ual’s vehicle specifically causing the world toget warmer? It is unlikely that this could ever beshown, although it is reasonable to assume thatany individual vehicle contributes its share to thepossible effect of global warming caused byhuman-generated carbon dioxide in proportionto the amount of emissions.

Inventory Interpretation. It is argued bysome that, due to the difficulties cited previ-ously, the notion of impact assessment shouldbe dropped and replaced by inventory interpre-tation. Classification and characterization couldstill be used, but all suggestion that environ-mental effects are assessed is avoided. In com-parative assessments, “less is better” is theprinciple in identifying the environmentallypreferable alternative.

Table 1.1 Example of a life-cycle inventory foran unspecified productSubstance Amount Substance Amount

Inputs OutputsEnergy from Air emissions, mgfuels, MJ Dust 2000

Coal 2.75 CarbonOil 3.07 monoxide 800Gas 11.53 Carbon dioxide 11�105

Hydro 0.46 Sulfur oxides 7000Nuclear 1.53 Nitrogen oxides 11,000Other 0.14 Hydrogen Total 19.48 chloride 60

Hydrogen Energy from fluoride 1feedstocks, MJ Hydrocarbons 21,000

Coal �0.01 Aldehydes 5Oil 32.75 Other organics 5Gas 33.59 Metals 1Other �0.01 Hydrogen 1Total feedstock 66.35 Solid wastes, mgTotal energy Mineral waste 3100

input, MJ 85.83 Industrial waste 22,000Slags and ash 7000

Raw materials, mg Toxic chemicals 70Iron ore 200 Nontoxic Limestone 150 chemicals 2000Water 18�106 Water effluents, mgBauxite 300 COD 1000Sodium chloride 7000 BOD 150Clay 20 Acid, as H+ 75Ferromanganese �1 Nitrates 5

Metals 300Ammonium ions 5Chloride ions 120Dissolved organics 20Suspended solids 400Oil 100Hydrocarbons 100Phenol 1Dissolved solids 400Phosphate 5Other nitrogen 10Sulfate ions 10

COD, chemical oxygen demand; BOD, bacteriological oxygen demand. Source:Ref 1.2


Step 4: Improvement Analysis. This stepinvolves identifying chances for environmentalimprovement and preparing recommendations.Life-cycle assessment improvement analysis isan activity of product-focused pollutionprevention and resource conservation. Oppor-tunities for improvement arise throughout anLCA study. Improvement analysis is oftenassociated with design for the environment ortotal quality management. With both of thesemethodologies, improvement proposals arecombined with environmental cost and otherperformance factors in an appropriate decisionframework.

Application of Life-Cycle Analysis Results

The results of an LCA can be used by a com-pany internally, to identify improvements inthe environmental performance of a productsystem; and externally, to communicate withregulators, legislators, and the public regardingthe environmental performance of a product.For external communications, a rigorous peer-review process is usually required. Virtually allof the peer-reviewed studies conducted to daterepresent analyses of simple product systems.However, studies for systems as complicated asautomobiles are being conducted.

Whether used qualitatively or quantitatively,LCAs often lead to products with improvedenvironmental performance. In fact, an often-over-looked, important qualitative aspect ofLCA is that it engenders a sense of environ-mental responsibility. Beyond this develop-ment within manufacturers, LCA has thepotential to become a tool to regulate products,or perhaps even for “eco-labeling.” However,such uses are contentious and are expected toremain so.

The bulk of LCA efforts to date have beenfocused on preparing LCIs, with the impactassessment stage currently seen as the weakestlink in the process. Indeed, some companieshave even decided to skip this phase of theprocess altogether, opting to carry out a brieflife-cycle review before moving straight on tothe improvement stage.

Large or small companies and other users willfind LCA of value at a number of different levels.Indeed, groups such as SETAC and SPOLD nowsee LCA playing a key role in three main areas:

• Conceptually: As a framework for thinkingabout the options for the design, operation,and improvement of products and systems

• Methodologically: As a set of standards andprocedures for the assembly of quantitativeinventories of environmental releases orburdens—and for assessing their impacts

• Managerially: With inventories and—whereavailable—impact assessments serving as aplatform on which priorities for improve-ment can be set

Not surprisingly, perhaps, the bulk of currentLCA efforts is devoted to the second of theseareas, particularly initiatives such as the 1993Code of Practice by SETAC (Ref 1.3). How-ever, the scope of LCA is rapidly spreading toembrace the other two application areas. The“supplier challenges” developed by companiessuch as Scott Paper, which has incorporatedenvironmental performance standards in itssupplier selection process, underscore the veryreal implications of the managerial phase forsuppliers with poor environmental perform-ances. Also, the “integrated substance chainmanagement” approach developed by McKin-sey & Company Inc. (Denmark) for VNCI(Association of the Dutch Chemical Industry),covering three chlorine-base products, showsthat LCA can produce some fairly pragmatictools for decision making.

Longer term, the prospects for LCA are excit-ing. Within a few years, product designers world-wide may be working with “laptop LCAs”—small, powerful systems networked with largerdatabases and able to steer users rapidly aroundthe issues related to particular materials, prod-ucts, or systems. This process would be greatlyaided by a widely accepted, commonly under-stood environmental accounting language.

In the meantime, however, LCA is still quitefar from being simple or user-friendly, as isillustrated in the following example.

Example: Life-Cycle Analysis of a Pencil.Anyone who has had even a brief encounter withan LCA project will have seen flow charts rathersimilar to the one in Fig. 1.4, which shows thekey life-cycle stages for one of the simplest in-dustrial products, a pencil. Most such diagramsare much more complicated, but, as is evident inthe figure, even the humble pencil throws an ex-traordinarily complex environmental shadow.

For example, imagine the flow chart inFig.1.4 is on the pencil maker’s PC screen as thecomputer menu for an electronic information


system. When the pencil maker clicks on “Tim-ber,” a wealth of data begins to emerge thatmakes one realize things are not as simple asmay have been imagined. Not only is there apotential problem with tropical timber becauseof the rain forest issue, but the pencil makernow notes that suppliers in the U.S. PacificNorthwest have a problem with the conflictbetween logging operations and the habitat ofthe Northern Spotted Owl.

At this point, a pencil maker recognizes theneed to examine the LCAs produced by thecompanies supplying timber, paints, andgraphite. Working down the flowchart, the pencilmaker sees a total of ten points at which otherLCA data should be accessed. This is wherecomplex business life gets seriously compli-

cated. At the same time, however, LCA projectscan also be fascinating, fun, and a potential goldmine of new business ideas.

Different Approaches to LCA. As Fig. 1.5indicates, the LCA practitioner can look at thelife cycle of a product through a number oflenses, focusing down of life-cycle costs orfocusing out to the broader sociocultural effects.One example is the Eco-Labeling Scheme (Fig.1.6 administered by the European CommissionDirectorate General XI (Environment, NuclearSafety, and Civil Protection). This scheme iscommitted to assessing environmental impactsfrom cradle to grave.

The sheer variety of data needs, and of datasources, makes it very important for LCAproducers and users to keep up to date with the

Other primaryproduction via LCAs

Primaryproduction

LCAsLCAForest

Timber

Energy

Plant

Building

Othermaterials

Manufacture Packaging

Retail

Consumer

Disposal

Impacts in use

Biosphere

Pencil manufacture

A PENCIL

Transport

Intermediate production

GraphitePaints

Intermediateproduction

LCAsLCA

Other LCALCA

LCA

LCA

LCA

LCA

LCA

LCA

LCA

Biosphere

Various impactsBiosphere

Disposal

Emissions

Discharges

Co-products

Biosphere

Biosphere

Biosphere

Fig. 1.4 Simplified life-cycle analysis (LCA) process for a pencil. Source: Ref 1.4


debate and build contacts with other practition-ers. Among the biggest problems facing theLCA community today are those associatedwith the availability of up-to-date data and thetransparency of the processes used to generatesuch data.

Most LCA applications, however, focus—andwill continue to focus—on single products andon the continuous improvement of theirenvironmental performance. Often, too, signifi-cant improvements will be made after arelatively simple cradle-to-grave, or perhapscradle-to-gate, analysis.

A detergent company, for example, may findthat most of the energy consumption associatedwith a detergent relates to its use, not its

manufacture. So, instead of just investing in asearch for ingredients that require less energy tomake, the company may decide to develop adetergent product that gives the same perform-ance at lower wash temperatures.

In short, LCA is not simply a method forcalculation but, potentially, a completely newframework for business thinking.

Case History: LCA of anAutomobile Fender

A detailed LCA for an automotive fenderas performed by IKP (University of Stuttgart,Germany) and PE Product Engineering

1. Cost

2. Value/ performance

4. Environment: beyond compliance

5. Socio- economic

6. Socio- cultural

3. Environment: compliance

Rawmaterial

acquisition

Bulkprocessing

Engineeredmaterials

processing

Assemblyand

manufacture

Useand

service

Retirement Treatmentand

disposal

Fig. 1.5 Matrix showing some possible different approaches to LCA. Source: Ref 1.4

ENVIRONMENTAL FIELDS Preproduction Production Distribution Utilization Disposal

Waste relevance

Water contamination

Air contamination

Noise

Consumption of energy

Consumption of naturalresources

Effects on ecosystems

Soil pollution and degradation

Product life cycle

Fig. 1.6 The European Community eco-labeling scheme “indicative assessment matrix.” Source: Ref. 1.4


(Dettingen/Teck, Germany) is included to illus-trate the present status and limitations of thismethodology (Ref 1.1).

Goal and Scope. The specific goal of thisinvestigation was to compare four differentfender designs for an average compact classautomobile in Germany. The comparisonshould result in the identification of the bestmaterial in terms of resource use, impact onglobal climate, and recyclability.

The four options were steel sheet; primaryaluminum sheet; an injection-molded polymerblend of polyphenylene oxide and nylon(PPO/PA); and sheet molding compound(SMC), a glass-fiber-reinforced unsaturatedpolyester resin. The mechanical requirementsfor the four fenders were identical; this en-sures that the functional unit is well definedand that they are equivalent. Table 1.2 showsthe materials and weights of the four differentfender designs.

Data Origin and Collection. Data in thiscontext means all pieces of information thatmay be relevant for the calculation of processesand materials. Such information includes mate-rial and energy flows of processes, processdescriptions, materials and tools, suppliers,local energy supply, local energy production,production and use of secondary energy carriers(e.g., pressurized air, steam), and location ofplants. Which processes are the most relevantand must be considered in more detail dependson the goal and scope of the study. Within thisstudy, the following information (supplier spe-cific, if possible) had to be identified, collected,and examined:

• Production processes, with all links in theprocess chain

• Primary data concerning energy and materialflow with respect to use of energy carriers(renewable and nonrenewable), use ofmineral resources (renewable and nonrenew-able), emissions into the air, waterborne

emissions, and waste and productionresidues

• Coupled and by-products as well as entriesfrom other process steps (internal loops)

• Transportation needs with respect to dis-tance, mode, and average utilization rate

• Primary energy carriers and their means ofproduction and distribution

• Secondary energy carriers and their meansof production and distribution

• Air and water treatment measures and dis-posal of residues

Data collection is not a linear process. Gooddata collection and evaluation requires iterationsteps for identifying relevant flows or addi-tional information, and experience is needed tointerpret the collected data. Calculation ofmodules should be carried out with specialregard to the method of data collection (e.g.,measured, calculated, or estimated) and thecomplexity of the system.

Materials production is an important factor.The consideration of aluminum shows that notonly the main production chain has to be consid-ered but also the process steps for alumina pro-duction (Fig. 1.7). The steps in electrolysis mustbe calculated, and the energy use connected withcaustic soda and the anode coke has to be

Alumina production

Emissions

Anode coke

Energy

Others

Red mud

CaCO3

NaOH

Emissions

Emissions

Press shop

Aluminum production(electrolysis)

Processor

Sheet

Al-loop withsalt slag recycling

Bauxite mining

Bauxite

Al2O3

Al

Fig. 1.7 Main material flow for the production of aluminumsheet parts. Source: Ref 1.1

Table 1.2 Material and weight of the differentfender designs

Thickness WeightMaterial mm in. kg lb

Steel 0.7 0.0275 5.60 12.35Sheet molding

compound 2.5 0.10 4.97 11.00Polyphenylene

oxide/polyamide 3.2 0.125 3.35 7.40Aluminum 1.1 0.043 2.80 6.20

Source: Ref 1.1


examined. The four steps shown in Fig. 1.7,which must be considered along with a long listof others, demonstrate the difficulty of balancingcosts and environmental impacts.

In electrolysis, the source of electric power isimportant because of the differences in carbondioxide emissions between plants that arewater-power driven and those that burn fossilfuels. Another significant factor is how electrol-ysis is controlled. Modern plants use technolo-gies that prevent most of the anode effectsresponsible for the production of fluorocarbongases, but many older plants emit four or fivetimes as much. This shows the importance ofcalculating on a site-specific or at least on acountry-specific basis.

Because aluminum is globally merchandised,the user frequently does not know the exactsource of the metal. The solution to this prob-lem is to calculate the average aluminum importmix. However, this calculation requires detailedinformation about the different ways aluminumis produced all over the world.

Material weight must also be considered. Inselection of automotive parts, the usage phase isof great interest. The main environmental factorduring this phase is weight difference. Each partcontributes to the energy demand for operatingan automobile. The share a fender contributesdepends only on its mass. However, no data areavailable for the same car carrying differentfenders. Therefore, this study calculated the fuelconsumption assuming a steel fender, becauseaverage fuel consumption is known for the com-plete car with the traditional fender. In the sameway, possible weight savings are known. Mea-surements and judgments from all automobileproducers show that the assumptions for fuel re-duction from weight savings vary within a rangeof 2.5 to 6% fuel reduction per 10% weight sav-ings. For this study, 4.5% was assumed to be anaverage value for the kind of cars considered.

Recycling of the SMC fender shows anotherweight-related issue. After the useful life of theproduct, a decision has to be made about whetherthe part should be dismantled for recycling orotherwise disposed of. Within this study, the re-cycling solution was considered because theSMC part can be dismantled easily and groundinto granules. Furthermore, SMC can replacevirgin material as reinforcement, and granulescan be used as filler up to 30%. In addition to thepossibility of using recycled material in newparts, the SMC recycling process offers anotheradvantage because the reformulated material has

a lower density than the primary material. Thismeans that the use of recycled SMC leads to fur-ther weight savings of approximately 8%, whilefulfilling the same technical requirements. Thisexample shows that recycling is not only usefulfor the purpose of resource conservation butmany provide other benefits as well. However,successful recycling requires more than techni-cal feasibility—it is highly dependent on viableeconomics.

Inventory Results. The discussion of thewhole inventory process is not possible here,because it includes up to 30 resource parame-ters, approximately 80 different emissions intothe air, more than 60 water effluents, and manydifferent types of waste. Therefore, this exampleconcentrates on energy demand, selected air-borne emissions, and resource use (recyclability).

Energy use is one of the main parameters toconsider when selecting automotive parts. It is areliable basis for judgment because energy usegenerates waste and emissions, and it requiresdepletion of resources. Figure 1.8 shows theenergy demand for the different fender materi-als over two complete usage phases, includingproduction out of raw material and recycling forthe second application.

The values at the zero kilometer line repre-sent the energy needed for both material andpart production. It is easy to see that aluminumhas the highest energy demand of all four mate-rials. This comes mainly from the electrolysisprocess and the alumina production process.

SMC has the lowest energy demand, needingapproximately one-third of the energy requiredfor the aluminum fender. This is due to the factthat SMC is a highly filled material in whichthe extender is a heavy, relatively inexpensivematerial. Second best is steel, which requiresonly a little more energy than SMC. Some-where in the middle is the PPO/PA blend; thereason for the relatively high energy demand isthe feedstock energy of the materials used inpolymer production.

The ascending gradients represent thedifferences arising from the weights of the fend-ers. The larger the gradient, the higher theweight. It is easy to see that steel, as the heaviestmaterial, loses a lot of its advantage from theproduction phase. This points out the impor-tance of lightweight designs. The energydemand for the usage phase is approximatelyfour times higher than that required for partproduction. As a result, the most significantimprovements can be made in the usage phase.


Nevertheless, SMC still has the lowest energydemand after the first usage phase, and alu-minum is still the worst.

After the first life cycle of the fender, it isrecycled into a new part. The energy neededfor recycling of SMC, steel, and aluminum isrelatively low; PPO/PA requires much moreenergy for recycling. The disadvantage ofPPO/PA is that although recycling is possibleand very energy efficient, the production of the70% virgin material required in the part is veryenergy intensive.

The second utilization phase shows the sameresults as the first. In the final analysis, steel

turns out to be the most energy-intensive mate-rial, followed by the PPO/PA blend. While steelhas the disadvantage of its weight, the polymerblend has disadvantages concerning recyclabilityfor external body parts. The situation would betotally different if more material could be recy-cled, or if the polymer blend could be usedmore extensively in heavier cars with a longerusage phase. The weight advantage is especiallyhigh for aluminum. However, SMC turns out tobe the most energy-efficient material over-all.

Emissions of carbon dioxide, nitrogen oxides,sulfur dioxide, and fluorocarbons were esti-mated for each material because of their effects

2800

1400

1037

682

357299

00 100

SMCSteel

PPO/PAAluminum

Distance traveled by automobile, km × 103

Ene

rgy,

MJ

240200

180 300

26392478

21692286

26392478

21692286

Fig. 1.8 Energy consumption for the production, use, recycling, and reuse of different fender materials considering the distancetraveled by the automobile. PPO/PA, polyphenylene oxide and nylon; SMC, sheet molding compound. Source: Ref 1.1

150

200

100

Em

issi

ons

per

fend

er, k

g

Em

issi

ons

per

fend

er, g

50

0

250

300

200

150

100

50

CO2 Dust

179.8

143.0131.7

134.1

83.3

287.7

11.413.5

291.1

138.2

12.417.6

108.1 109.0

155.1

75.3

129.3

218.1

91.6

120.3

40.2

54.555.0

111.9

CO NOx SO2 NMVOC

0Steel

Al PPO/PA

SMC

Steel

Al PPO/PA

SMC

Steel

Al PPO/PA

SMC

Steel

Al PPO/PA

SMC

Steel

Al PPO/PA

SMC

Steel

Al PPO/PA

SMC

ReuseRecyclingFirst useProduction

Fig. 1.9 Selected airborne emissions for the production, use, recycling, and reuse of different fender materials. NMVOC, non-methane volatile organic compound; PPO/PA, polyphenylene oxide and nylon; SMC, sheet molding compound.Source: Ref 1.1


on ozone depletion and global warming (Fig.1.9). These pollutants were also chosen becausethey are generated by nearly every manufacturingprocess, all over the world.

As mentioned before, a high percentage ofatmospheric emissions is caused by energy gen-eration. In the case of polymers, emissions arelower than expected because so much energy isstored as material feedstock. Aluminum is thematerial with the highest energy demand, butemissions are comparatively low becausewater power is used for a high percentage ofaluminum electrolysis. The highest levels ofcarbon dioxide are emitted during steel produc-tion, mainly from the ore reduction process.Carbon dioxide emissions for the production ofboth polymers are dominated by hydrocarbonprocessing and refining.

For aluminum, most emissions come fromearlier process steps. Alumina is producedmainly in bauxite mining countries, where theleast expensive locally available energy is typi-cally generated by burning heavy fuel andcoal. Carbon dioxide emissions from alu-minum production are dominated by thissource, plus the electric power demand ofthose electrolysis processes that are not basedon water power.

Carbon dioxide emissions during usage aredirectly related to fuel consumption: heavierfenders result in the generation of more carbondioxide. This is also true for all other emissionsconsidered here. One important approach for apossible improvement is certainly to reduce thismain impact on global warming.

Impact assessment is a special step withinthe framework of LCA. Based on the results ofthe inventory, conclusions can be drawn, andjudgments and valuations are possible. The im-pact assessment supplies additional informationthat enables the practitioner to interpret theresults from the inventory.

Impact assessment also should allow thepractitioner to draw the right conclusions con-cerning improvement approaches. However, itshould be noted that consideration of environ-mental effects as a consequence of environmentalreleases is additional information that is notcovered by the inventory step. This case historyprovides only a brief overview.

Impact assessment involves three steps. Firstis the definition of “environmental problems”or “themes.” The problems to be addressed aredefined in the scope of the project. Second,emissions are grouped to show their specific

contribution to the environmental themes.Third, their shares are calculated. A standardlist covers the following themes, which aremore or less identical with most of the ap-proaches taken in LCA literature:

• Global criteria: Resource use (energy carri-ers and mineral resources, both renewableand nonrenewable, and water and land use),global warming, ozone depletion, and releaseof persistent toxic substances

• Regional criteria: Acidification and landfilldemand

• Local criteria: Spread of toxic substances,eutrophication, and formation of photo-chemicals

• Others: Noise, odor, vibration, and so on

In most of the studies conducted by IKP andPE Engineering, resource use and the global cli-mate problems are considered. The methodologyfor their consideration is broadly accepted.Sometimes, acidification or eutrophication isconsidered as well. All others are more difficultto handle, and appropriate methods are stillunder discussion.

For the fender example, the contribution tothe global-warming problem is calculated bytaking into account production, use, recycling,and second use of each material (Fig. 1.10). Theresults are mainly influenced by carbon dioxideemissions and energy use and show that light-weight materials have advantages during uti-lization. However, aluminum is far worse thanthe others during production because electroly-sis is accompanied by fluorocarbon emissions(CF4 and C2F6), which have a very high global-warming potential.

Valuation. The second step in the judgmentof the environmental impacts is the valuationstep. This step may be divided into the normal-ization process and the final weighing.

Normalization involves scaling absolute con-tributions to single environmental themes on thesame level, because absolute numbers may varywithin six to ten decades. The effect scores arenormalized with the amount of the annualglobal effect score or the contribution of oneprocess to the theme per year, and so on.

Final weighing involves a personal judgmentabout the importance of each environmentaltheme, and the effect of each score on overallimpact. This final step is part of the decision-making process. Scientists create tools for thisprocess and help decision makers use and


understand them, but the final decisionsdepend on company policies, not scientific orconsultancy work.

Improvement Options. From this study, thefollowing conclusions for improvement can bedrawn:

• The usage phase is dominated by fuel con-sumption and the resulting carbon dioxideemissions. For other emissions, the produc-tion phase and recycling is also of great importance.

• Reducing part weight may improve energyuse and reduce the contribution to globalwarming. However, reducing part weightmay require higher environmental invest-ments during production or recycling. Insome cases, these investments are veryuseful.

• Recycling is more important for expensiveand energy-intensive materials.

Experience gained from the evaluation of fendermaterials shows that the following general con-clusions can be made:

• The fuel production has great impact and isnot well known today.

• The best basis for decision making is asupplier-specific LCA.

• Close cooperation between producers andsuppliers is necessary to find processes thatwill reduce environmental impacts.

Conclusions

Life-cycle engineering—in particular, LCA—is gaining importance for design and materialsengineers because environmental considerationsare increasingly important factors in design andmaterials selection. The creation and develop-ment of environmental management systems,including extended producer responsibility andproduct stewardship responsibility, pollution pre-vention strategies, “green” procurement guide-lines, and eco-labeling programs, are evidence ofthe growing importance of life-cycle concerns.

To make a proper assessment, the total lifecycle of a material, all forms of energy use,waste production, reuse, and recycling have tobe considered. Many of these factors are sitespecific, which complicates calculations andcomparisons. While LCAs for simple productshave been performed, more complicated sys-tems are only now being tackled.

Many industry trade organizations have de-veloped or are in the process of developingLCI databases for their products. The Associa-tion of Plastics Manufacturers in Europe(APME), the European Aluminum Association(EAA), Finnboard, and the International Ironand Steel Institute are just a few examples.This publication addresses efforts with respectto aluminum and energy conservation.

A wide variety of reports and software pack-ages containing inventory data are available.

100

200

50 35

26

14

79

0100

183

0 200Distance traveled by automobile, km × 103

GW

P p

er fe

nder

300

163

138

PPO/PASMCAluminumSteel

147150

Fig. 1.10 Calculated contribution to global warming for the production, use, recycling, and reuse of different fender materials considering the distance traveled by the automobile. GWP, greenhouse warming potential (CO2 equivalents); PPO/PA,

polyphenylene oxide and nylon; SMC, sheet molding compound. Source: Ref 1.1


Examples are given in Table 1.3 (Ref 1.5). Inaddition, a large number of national and interna-tional database projects exist. A comprehensivelisting can be found in Ref 1.6. Also, Ref 1.4,while concentrating on Europe, gives an excel-lent overview and many useful examples andaddresses.

Steps of a complete LCA are being standard-ized with respect to methods and data. Simpli-fication and standardization will lead to morereliable, timely, and cost-effective LCIs. Anoperational guide for ISO standards is given inRef 1.7. When consensus about an acceptableimpact assessment methodology is reached,LCA for simple and then more complex unitsand systems will be possible.

ACKNOWLEDGMENTS

Portions of this article were adapted from Ref1.1 and 1.2. The authors wish to thank Sustain-ability Ltd. (United Kingdom) and the Secretariatof SPOLD (Belgium) for allowing the use ofsome of their information.

REFERENCES

1.1. M. Harsch et al., Life-Cycle Assessment,Adv. Mater. Proc., June 1996, p 43–46

1.2. J.L. Sullivan and S.B. Young, Life CycleAnalysis/Assessment, Adv. Mater. Proc.,Feb 1995, p 37–40

1.3. “Guidelines for Life Cycle Assessment: ACode of Practice,” Society of Environmen-tal Toxicology and Chemistry (SETAC),Europe (Brussels), 1993

1.4. The LCA Sourcebook, Sustainability Ltd.,London, 1993

1.5. “Life Cycle Assessment: Principles andPractice,” EPA/600/R-06/060, ScientificApplications International Corporation(SAIC), May 2006

1.6. “Directory of Life Cycle Inventory DataSources,” Society for the Promotion ofLCA Development (SPOLD), Brussels,Nov 1995

1.7. J.B. Guinée, Handbook on Life CycleAssessment: Operational Guide to the ISOStandards, Kluwer Academic Publishers,2002

Table 1.3 Life-cycle assessment and life-cycle inventory software toolsTool Vendor URL

BEES 3.0 NIST Building and Fire Research Laboratory http://www.bfrl.nist.gov/oae/software/bees.htmlBoustead Model 5.0 Boustead Consulting http://www.boustead-consulting.co.uk/products.htmCMLCA 4.2 Centre of Environmental Science http://www.leidenuniv.nl/cml/ssp/software/

cmlca/index.htmlDubo-Calc Netherlands Ministry of Transport, Public Works http://www.rws.nl/rws/bwd/home/www/cgi-bin/

and Water Management index.cgi?site=1&doc=1785Ecoinvent 1.2 Swiss Centre for Life Cycle Inventories http://www.ecoinvent.chEco-Quantum IVAM http://www.ivam.uva.nl/uk/producten/product7.htmEDIP PC-Tool Danish LCA Center http://www.lca-center.dkeiolca.net Carnegie Mellon University http://www.eiolca.netEnvironmental ATHENA Sustainable Materials Institute http://www.athenaSMI.ca

Impact IndicatorEPS 2000 Design Assess Ecostrategy Scandinavia AB http://www.assess.se/

SystemGaBi 4 PE Europe GmbH and IKP University of Stuttgart http://www.gabi-software.com/software.htmlGEMIS Öko-Institut http://www.oeko.de/service/gemis/en/index.htmGREET 1.7 DoE’s Office of Transportation http://www.transportation.anl.gov/software/

GREET/index.htmlIDEMAT 2005 Delft University of Technology http://www.io.tudelft.nl/research/dfs/idemat/index.htmKCL-ECO 4.0 KCL http://wwwl.kcl.fi/eco/softw.htmlLCAIT 4.1 CIT Ekologik http://www.lcait.com/01_1.htmlLCAPIX vl.1 KM Limited http://www.kmlmtd.com/pas/index.htmlMIET 3.0 Centre of Environmental Science http://www.leidenuniv.nl/cml/ssp/software/miet/index.htmlREGIS Sinum AG http://www.sinum.com/htdocs/e_software_regis.shtmlSimaPro 6.0 PRé Consultants http://www.pre.nl/simapro.htmlSPINE@CPM Chalmers http://www.globalspine.comSPOLD The Society for Promotion of Life- http://lca-net.com/spold/

Cycle AssessmentTEAM 4.0 Ecobalance http://www.ecobalance.com/uk_lcatool.phpUmberto ifu Hamburg GmbH http://www.ifu.com/en/products/umbertoUS LCI Data National Renewable Energy Lab http://www.nrel.gov/lci

Source: Ref 1.5

SELECTED REFERENCES

• “Life Cycle Assessment: Inventory Guide-lines and Principles,” U.S. EnvironmentalProtection Agency (EPA), Office of Re-search and Development, Cincinnati, OH,1993

• “Life Cycle Design Manual: EnvironmentalRequirements and the Product System,”University of Michigan, 1993

• “A Technical Framework for Life CycleAssessment,” SETAC USA, Washington,D.C., 1991


HARVEY BROOKS (Ref 2.1) focused on thelinks between materials, energy, and the envi-ronment. At that time, the issue of sustainabilityhad emerged but appeared first as an explorationof the possibility of materials depletion in theface of predicted population growth. Thistheme of sustainability is made more relevantby the current debate about greenhouse gasesand the global climate, but one which is centralto the materials industries and the users of theirproducts. One cannot do justice to such a broadtopic in this brief overview, but several themesdo establish the basis for several recommenda-tions. There are roles to be played by industry,by academia, by government, and by each of usindividually and collectively through our soci-eties. With some passing remarks about the otherarenas, the aim of the recommendations is forgovernment and professional societies, basedon current knowledge.

Our economy and that of the other nations inthe world depends on materials to an extent thatmost nontechnical persons do not realize.Materials are, after all, “the stuff that things aremade of.” At issue today, is can the current levelof use of this “stuff” maintain? If, as expected,both economic activity and population willgrow significantly worldwide in the next half-century, the current per capita use of materialswill certainly be unsustainable. Indeed, theOrganization for Economic Cooperation andDevelopment (OECD) recently adopted a long-range goal that industrial countries shoulddecrease their materials intensities by a factorof 10 over the next four decades. That would be

equivalent to using only 66 lb of materials per$100 gross domestic product (GDP) comparedto the present value of approximately 660 lbper $100 GDP. If achievable at all, much ofthese savings would come from the construc-tion and mining industries, the largest users ofmaterials in tonnage, but opportunities for moreefficient use of materials would need to befound in every aspect of our industrial andservice economy. Are such efficiencies achiev-able, and, if so, can they be obtained throughthe application of existing technologies, ormust new technology be developed? There isno comprehensive answer to such questions,yet there is hope that governments would turnto members of the materials community foranswers as they make public policy on issuesrelating to sustainable development.

Today, there is a sensitivity to the desire toachieve a sustainable world economic and so-cial system, which both satisfies human needsand does not despoil the earth. Brown air in thecapitals of many of the world’s countries, holesin the ozone layer, and increases in greenhousegases, with their attendant climactic conse-quences, have made some of the negative im-pacts of current technology apparent to all. Thecentrality of materials usage to this subjectshould make the achievement of sustainabilityour issue, but environmental issues have not al-ways been visible on our lists as we identifiedour priorities for future attention. Indeed, it ismore common today to hear the mechanical andelectrical engineering communities discussingdematerialization and alternate materials tech-nologies with ecologists and economists. Whathas been our role in this increasingly importantarena of human concern, and what should it be?

Think back to the times in which this series oflectures originated. Those were the days of

CHAPTER 2

Sustainability—The Materials Role*

* Adapted from the 1998 Distinguished Lecture in Materi-als and Society of ASM International, by Lyle H. Schwartz,Metallurgical and Materials Transactions A, Vol 30, April1999, p 895.




awakening consciousness: the first Earth Day onApril 22, 1970; the development of “green”political parties and action organizations; the be-ginning of a period of rapidly escalating regula-tion; and the inception of a heightened awarenessof the links between materials, energy, the envi-ronment, and rising population. Two landmarkstudies captured the thinking of that era about therole of materials. In 1973, the report of theCongressionally mandated National Commissionon Materials Policy appeared. Entitled “Materi-als Needs and the Environment Today andTomorrow” (Ref 2.2), this document contained adetailed discussion of the materials cycle andmany recommendations for government action.Shortly thereafter, a major Nuclear RegulatoryCommission (NRC) study was published. Thismonumental effort was a product of the ad hoccommittee on the study of materials (COSMAT),chaired by Morris Cohen. Titled “Materials andMan’s Needs,” this text (Ref 2.3) set the stage forserious consideration about materials science andengineering for the next 20 years, clarified con-cerns with structure-property relationships, andgraphically focused attention on the whole mate-rials cycle. Indeed, it was from the COSMATstudy that the representation of the materialscycle (Fig. 2.1) was derived, which still definesthe scope of our field. These two documents, the

first emphasizing policy and the second explor-ing technical issues, education, and research anddevelopment (R&D), represented high-watermarks for focus of attention by the materialscommunity on the links between what they doand the consequences to the environment.

While many of the specific recommendationsin these volumes are a bit dated, the generalprinciples still apply and are worth repeatinghere. The three summary directives for policymakers were as follows (Ref 2.4):

“Strike a balance between the ‘need to pro-duce goods’ and the ‘need to protect the envi-ronment’ by modifying the materials systemso that all resources, including environmental,are all paid for by users.”

“Strive for an equilibrium between the sup-ply of materials and the demand for their useby increasing primary production and byconserving materials through acceleratedwaste recycling and greater efficiency-of-useof materials.”

“Manage materials policy more effectivelyby recognizing the complex interrelation-ships of the materials-energy-environmentsystem so that laws, executive orders, andadministrative practices reinforce policy andnot counteract it.”

Fig 2.1 The total materials cycle

THE TOTAL MATERIALS CYCLE

ARENA OFMINERAL AND

AGRICULTURALSCIENCES ANDENGINEERING

ARENA OFMATERIALS SCIENCES AND ENGINEERING

RAWMATERIALS

ORE COAL

BULK MATERIALSENGINEERING

MATERIALS

WASTEJUNK

PERFORMANCESERVICE

USE

DESIGNMANUFACTURING

ASSEMBLY

CRYSTALS

EXTRACTREFINE

PROCESS

METALS

CHEMICALS PAPER

CEMENT FIBERS

PROCESS

RECYCLE

ALLOYS CERAMICSPLASTICS CONCRETE

TEXTILES

DISPOSE

SAND WOOD OILPLANTSROCK

THE EARTH

OIL WOODORE

MINEDRILL

HARVEST

PRODUCTS

MACHINES

STRUCTURES

DEVICES

Chapter 2: Sustainability—The Materials Role / 17

While significant progress since the early1970s may be found on all three fronts, theseprinciples could well be taken today as guid-ance for our continued efforts in both public andprivate arenas.

As the decade of the 1970s passed, other lec-turers in this series returned to environmentalissues. James Boyd devoted his 1973 lecture(Ref 2.5) to resource limitation in the context ofwhat he termed the resource trichotomy: materi-als, energy, and the environment. Michael Ten-nebaum (Ref 2.6) included references to the en-vironment in his 1975 lecture, describing whathe saw as lack of balance in regulation efforts,with too much focus on the near term. HerbertKellogg (Ref 2.7) in 1978 returned to the subjectof conservation and anticipated the current termdematerialization to describe the more efficientuse of materials. However, with the exception ofseveral passing allusions to the need to be envi-ronmentally sensitive in everything we do, laterspeakers in this series have not devoted any seri-ous consideration to the subject.

The visible evidence of our apparent disin-terest in the subject was most clearly presentedto the world in our comprehensive self-studyorganized by the National Academy of Sciences-National Academy of Engineering-NuclearRegulatory Commission in the late 1980s. It isdifficult to even find the word environment inthat volume entitled Materials Science andEngineering for the 1990’s; Maintaining Com-petitiveness in the Age of Materials (Ref 2.8).What were the reasons for the disappearanceof environmental concern from the center ofour focus?

Some History

There have certainly been many issues ofconcern in the past quarter-century in terms ofeconomic challenges, dramatic restructuring ofindustrial sectors, shifts of employment onmassive scales, and globalization of manufac-turing and R&D. Governmental focus alsoshifted from energy and the environment in the1970s to strategic defense concentration in the1980s, and onto industrial competitiveness andtechnology transfer in the 1990s. Most dramat-ically, this era brought the end of the Cold Warand a world increasingly focused on economiccompetition and the desires of all people toachieve the standards of living exemplified bythat small fraction living in the so-called

developed lands. It is no wonder that in themidst of this major restructuring in the UnitedStates and the apparent invincibility ofalternate technology policy and strategies inseveral Asian nations, the last major study onmaterials by the NRC focused its attention onmaintaining industrial competitiveness.

The past quarter of a century has been one toreckon with, and our colleagues have exploredthe ramifications of many of these societalchanges in the series of lectures that precededthis one. Many of these issues have drawn ourattention away from the necessity of organizingour activity to achieve a sustainable economicsystem.

Certainly, the issue has not gone away. We’vemade progress and expanded our knowledgeand environmental sensitivities, and our atten-tion has moved beyond cleaning up the waterand the air and controlling pollution. We nowexplore a systems approach to the developmentof a world in which human society and com-merce can be sustained throughout the cominggenerations. Emphasis on pollution and envi-ronmental impact caused us to focus our atten-tion on effluents and process modification. Bycontrast, emphasis on sustainability will directour attention increasingly toward materialsusage and the flow of materials through whathas been called the materials cycle.

This transition in the way environmental goalsare viewed was made concrete in the 1992United Nations Conference on Environmentand Development (UNCED) held in Rio deJaneiro, Brazil. The strategies emphasized toachieve a sustainable state of developmentwere as follows: improving efficiency and pro-ductivity through frugal use of energy andmaterials, substituting environmentally detri-mental materials with ones that are less so, aswell as recycling and reusing products at theend of their lives. These are remarkably similarto the agenda espoused by COSMAT almosttwo decades earlier and remain the technicalframework today. It was particularly strikingthat UNCED defined waste as “material out ofplace.” In that simple phrase, the challenge forthe materials community is clearly etched.While we may be increasingly in agreementaround the world about desirable outcomes, weare not always in such close accord about thepath to achieve those goals, and the record ofactions is more variable yet.

A comprehensive picture of where we standon the road toward achieving sustainability


could fill a book, and indeed many articles andbooks have been written on the subject. Whenwe move from generalities to specifics, there isno consensus on even the definition of sustain-ability. We cannot precisely define it, but we willknow it when we see it. In lieu of such a precisedefinition, the following discussion extracts afew facts from one of those many articles: thatwritten by Ausubel et al. (Ref 2.9), who, in 1995,explored “The Environment Since 1970.”

During this period, population grew from 3.7to 5.7 billion and is expected to continue togrow to a steady state of double or triple currentlevels. Global per capita commercial energyconsumption has stayed level, but because ofpopulation growth, 8 billion tons of oil wereused in 1995 compared to 5 billion in 1970. Theoil “crisis” of the 1970s has not returned, asproven oil reserves have increased from 600 to1000 billion barrels during the period, whileusing more than 500 billion barrels, which werepumped from the ground. Meanwhile, technol-ogy has been shifting our source of energytoward the less polluting natural gas, a form ofdecarbonification, while proven reserves of natu-ral gas have tripled. In short, while we may seesome improvement in reduction of greenhousegases per unit of energy extracted, we cannotcount on imminent shortages of fuels to drive usin the direction of seeking alternate, higher-priced sources of energy, even if they mayrepresent more “sustainable” options.

Automobiles are, of course, a major aspect forstructural and functional materials of all sorts.Their manufacture, repair, and recycle occupy alarge fraction of the working populace. Theyuse substantial amounts of energy, contributinggreatly to greenhouse gas production. Since1970, as the world became more affluent and thepopulation grew, the number of motor vehiclesmore than doubled to the staggering figure ofapproximately 600 million, more than offsettingany gains in fuel efficiency that may have beenachieved. As the developing nations strive toemulate our affluent style of living, we may ex-pect to see the number of vehicles double again,and once again see the fruits of more efficientfuel utilization per unit negated by the sheernumbers of units.

Ausubel et al. (Ref 2.9) summarize the situa-tion by noting that “production, consumption,and population have grown tremendously since1970. . . . Globally, and on average, economicand human development appears to have out-paced population growth.”

There can be no doubt of the continuing presence of major global consequences ofhumankind’s expanding numbers and currentlifestyle. Among these, we may number increas-ing emission of greenhouse gases with projectedclimate change, depletion of stratospheric ozonelayer by chlorofluorocarbons, tropical forestdepletion, and, with it, decreases in biologicaldiversity, air quality in densely populated cities,and waste disposal in increasingly limited land-fill space.

Public concern for the environment has takenmany forms. In the United States, the number offederal laws for environmental protection hasmore than doubled since 1970, and governmentinvolvement in industrial operations has beencorrespondingly increased. As one indicator ofthat impact, spending on pollution abatementhas also doubled and exceeded $90 billion annu-ally by 1995. Nongovernmental environmentalorganizations in the United States have roughlytripled since 1970, and, with their presence, moreinformation is available in the popular press, notall of sound scientific basis.

Again quoting Ausubel et al. (Ref 2.9), “Peo-ple are demanding higher environmental quality.The lengthening list of issues and policyresponses reflects not only changing conditionsand the discovery of new problems, but alsochanges in what human societies define as prob-lems and needs” (Ref 2.9). This observation isnowhere more appropriate than when applied tothe automobile. In the 25-year period we arelooking at, cars were lightened by 30%; catalyticconverters dramatically cleaned up the noxiouseffluents of combustion; gas mileage increasedby a factor of 2; corrosion protection and dam-age-tolerant materials lengthened the use of ve-hicles, contributing to dematerialization; and, atthe same time, the vehicle became safer andmore attractive to owners. Materials substitutionhas reduced the weight of the average familysedan from approximately 4000 to approxi-mately 3000 lb and dramatically changed themix of materials. At the same time, however, wehave seen the development of changing con-sumer buying practice with rapid growth in themore materials-intensive small trucks, sportutility vehicles, and vans. In fact, since 1973, allof the increase in U.S. highway fuel consump-tion has been due to these other vehicles.

To offset these effects and to make a dramaticchange in materials usage and energy use willrequire a total redesign of the vehicles and theirpower trains. In the United States, the Big Three


auto manufacturers have joined together inR&D partnerships to explore many of therequisite new technologies under the banner ofthe United States Council for Automotive Re-search (USCAR) and are intensively engagedwith the federal government in a Partnership forNew-Generation Vehicles (PNGV) to achievesuch radically new vehicles. The situation intransportation has become even more complexas we confirm the consequences of increases ofgreenhouse gases on the global environment. Inthis context, the basis for power in most vehi-cles, the gasoline engine itself, is viewed as aculprit, and we must look to other powersources or at least to much greater efficiency viaburning carbon in power plants and then usingelectricity to power the vehicle.

In summary, there has certainly been muchaccomplished in the last 25 years, but instead ofthe issues of materials and the environmentdisappearing, they have become more demand-ing, raising questions in many arenas aboutwhether our advanced technologically-basedprosperity is in fact sustainable.

So, is it that others have taken this issue andmade it their agenda while we have remained inthe background? Many other communities haveembraced the issue of sustainability with openarms. Although the general initial reaction of in-dustry has been to resist regulations, many havenow recognized that environmentally friendlymanufacturing can be a plus to the bottom line.This practical observation and the desirablestrategic responses have now been codified in theterm industrial ecology.

Selected with obvious reference to the com-plex interactive natural world around us, thisterm has as many definitions as there are com-mentators, like that advanced by Frosch andUenohara (Ref 2.10):

“Industrial ecology provides an integratedsystems approach to managing the environ-mental effects of using energy, materials, andcapital in industrial ecosystems. To optimizeresource use (and to minimize waste flowsback to the environment), managers need abetter understanding of the metabolism (useand transformation) of materials and energyin industrial ecosystems, better informationabout potential waste sources and uses, andimproved mechanisms (markets, incentives,and regulatory structures) that encouragesystems optimization of materials and energyuse.”

The systems approach leads to thinking aboutboth the productive output and the waste fromone industrial arena as the input for others; itleads to a demand for extensive databases onmaterials flow throughout the total materialscycle, and it leads to the need for sophisticateddecision-making tools to enable enlightenedtechnical and business decisions. The materialscommunity has been engaged in such activitiesto a greater or lesser extent throughout the lastdecades, but we have always given too littlepriority to such activities in favor of more glam-orous, and usually more “scientific,” efforts innew materials development and characteriza-tion. Increasingly, as industrial rather thangovernment priorities dictate, we will have tochange our focus as well.

The Materials Role in Industrial Ecology

The U.S. government has been a major factorin driving environmental fixes for some timenow. The quarter-century since the first EarthDay has frequently been characterized as one ofregulatory command and control. Certainly, thisphilosophy has led to the development of quitean array of new technology, but much of thishas been aimed at “end-of-pipe” and cleanup.These old regulatory strategies, enacted as aquick dose of strong medicine, may have playedthemselves out. Many now question whether thecost/benefit of further regulatory reform is sup-portable. Increasingly over the last 10 years orso, many in the government have recognized theopportunities for new technology developmentas the next phase in the achievement of sustain-able development. This new philosophy wasexpressed in the first 2 years of the Clinton-GoreAdministration as an Environmental TechnologyInitiative (ETI) and described in glossybrochures as the National Environmental Tech-nology Strategy (Ref 2.11, 2.12). Products ofthe National Science and Technology Council,these referenced documents take a broad-brushview of the societal requirements to achievesustainability and touch lightly on the technicalspecifics. For example, the following general-izations about materials are made (Ref 2.11):

“Many materials will need to be discontinuedand new materials employed in order toachieve environmental technologies that con-form with the principles of industrial ecol-ogy. Advances in the materials used in the


manufacturing process and the developmentof lighter materials for transportation have thepotential to reduce the production of wastes,minimize the extraction and use of virginorganic resources, mitigate pollution, and im-prove energy efficiency. These new materialswould have a predictable lifespan and, whenthey are retired from their original use wouldbe designed to be used for other purposes.”

To translate such lofty goals into reality wouldtake planning and execution over decades. And itwould take dedicated resources over a long timehorizon. And, of course, it would take public-private cooperation on a scale not yet achievednor perhaps even imagined. This environmentalstrategy envisioned a broad multiagency effort,which was to include those agencies in whichmuch of the materials research is funded; how-ever, most of the “new” money appeared in theEnvironmental Protection Agency (EPA), whichwas designated lead agency. Not surprisingly,given the regulatory mission and historical strat-egy of the EPA, these funds were targeted at ex-pansions of pre-existing projects to implementbest existing practices, and little found its wayinto the development of new materials technolo-gies. The ETI was launched with an initialappropriation of $36 million in fiscal year (FY)1994, grew to $68 million in FY 1995, but wasthen slashed to $10 million for FY 1996 by the104th Congress, which directed that the remain-ing appropriation be directed toward environ-mental verification. This start-stop record isconsistent with the general history of technol-ogy investment by the federal government, butit is further complicated by the lack of claritywithin the Congress regarding their regulatoryrole. In no statute has Congress explicitly giventhe regulatory agencies any mission to encour-age technological change as a means towardenvironmental improvement (Ref 2.13).

We may properly ask here about the desiredtechnical agenda. If new funds were to be madeavailable, toward what ends may they be ap-plied? The technical agenda has not reallychanged much since it was outlined in the COS-MAT report in the early 1970s. In that overview,the subject was divided into effluent abatement,materials substitution, functional substitution (orredesign), waste disposal, and increased use ofrecycling. This list, expanded to the next level ofdetail, is included as Table 2.1. The specific is-sues vary in importance for different industries,and the degree of progress made since 1973 is

similarly quite varied; however, when the issuewas addressed in detail in a workshop held in1995 (Ref 2.14), the same general list of topicsemerged. Using a slightly more up-to-dateterminology, we may say that the opportunitiesmay be found in cleaner (greener) processing,alternative materials, dematerialization (lighterand less to do “same” function), and reuse orrecycle. Since the 1970s, substantial progresshas been made in each of these areas, but muchmore will clearly have to be done if we are toachieve a sustainable level of materials usage.

Some general comments about each areafollow.

Cleaner Processing. Industry has clearly beenthe leader here; driven largely by regulation, dra-matic improvements have been made in reduc-ing effluents, cleaning up scrap, and minimizingenergy uses. Linked strongly to the recyclingissue, new technologies have been introduced,which have radically transformed the materials-producing industries. Recall that before the1970s, the economic prowess of a nation wasschematized by a picture of tall smokestacks,belching smoke and other noxious fumes intothe atmosphere. How differently we regard thatimage today, and how much more common it isto see photos of grass-covered campuses onwhich green factories ply their work. Manufac-turing moves ever closer to a scrap-free environ-ment, as we introduce one after another of thedesirable net-shape technologies. ContinuedR&D in this arena is certain to produce furtherbenefits, but a delicate balance must be struckhere. In many of the industries where processimprovement is most likely to be of greatestimpact, such as casting, coatings, and specialtyalloys, the disaggregated nature of the businessand the small size of most companies makeresearch difficult to do and unlikely to pay off inthe near term. Regulations intended to force

Table 2.1 COSMAT list of materials tasks forenvironmental issuesEffluent abatement• Process restructuring• Containment• Recycling

Materials substitution• Through alteration of existing devices• Through substitute devices

Functional substitutionWaste disposal• Increased degradability• Reduction in noxiousness

Increased recyclability• Through design• Through suitable materials choice


cleaner processing in these industries may havethe undesirable alternative of forcing the compa-nies to locate off-shore, with a consequent lossof jobs to the United States and no net improve-ment in the world’s environmental position. Forthis, if for no other reason, close cooperationmust be sought within government between theregulators and the technology developers.

Alternative Materials. This topic appearssimilar to dematerialization, and some definitionsare appropriate. It is useful to distinguish be-tween those technologies intended to eliminate“bad” things from the waste stream (referred tohere as alternative materials) and those demateri-alization technologies intended to reduce mate-rial usage, either through reduction in weight inthe product or in scrap. When considering alter-native materials, we must focus our attention onthe materials producers and be sensitive to thesource of R&D funding. In the case of poly-mers, it has clearly been industry that has footedthe bill, with some rare exceptions. It is quite in-teresting that these companies have stronglylinked their business goals to their perceptionsof an environmentally conscious public. Refer,for example, to the Dupont ads that appear inevery issue of Scientific American and describethe newest biodegradeable polymers as alterna-tives for manufacturing. Dupont clearly believesthat developing new materials from renewablesources is in their best corporate interests aswell as those of the planet.

On the other hand, in the case of metals, wesee a mixed picture, due in part to the source orlack of R&D funding. At one time, extensivealloy development could be expected from theprivate sector. This went in two directions. Onthe commercial side, autos and beer cans, withtheir significant environmental implications,have produced fierce competitions for marketshare, which have driven both aluminum andsteel. On the other hand, in the aerospace arena,it is government funding that was dominant asthe driver for materials development. We haveseen the dramatic improvement in materialsproperties over the last 25 years, largely drivenby performance needs and often financed by thefederal government. Higher-temperature opera-tion to achieve higher performance has been themotivation behind programs focused on enginematerials and largely funded by the Departmentof Energy (DoE) and Department of Defense(DoD). While higher performance may translateinto higher fuel efficiency with positive environ-mental implications, this was clearly not the

driver. As DoD procurement needs decrease, weare finding them less ready to invest in suchcostly and time-consuming efforts as newmaterials development, and few of our industrialconcerns can capture enough economic benefitto justify private sector investment. This subjectwas addressed in some detail in the 1997 Distin-guished Lecture by Jim Williams (Ref 2.15). AsDoD disappears from the picture, will DoE fillthe void, and can we expect the motivation forenergy reduction, coupled with the desire forsustainability, to achieve a substantial new in-vestment by that agency?

In any discussion of alternative materials, wemust deal with the materials selection in thecontext of a system, and that the system must beexplored over its full life cycle. Life-cycle analy-sis (LCA) is the generic terminology to describesuch thought processes and accounting exercises,but current versions of this technology have notyet reached the levels of user-friendliness andeconomy, which would make LCA commonpractice for all designers. The weaknesses notedby many authors include: full analysis is toocostly to be justified, and, instead, limited envi-ronmental impacts usually suffice; results aretoo sensitive to input data and critical boundaryconditions, but sensitivity analysis is not readilyaccomplished; and no standards exist, limitingintercomparisons of alternate methodologies.These latter failings often lead to conflicting con-clusions, which decrease technical and publicacceptance of the results. If we cannot resolvethe paper cup versus plastic cup choice, howcan we be expected to use this methodology tochoose between aluminum and steel in autobodies?

We have a long way to go to resolve theseissues, but progress continues on the technicalissues of data and sensitivity analysis. Datalimitations include proprietary ownership andlack of accurate, materials-specific information,but the most significant uncertainties in LCAarise from options in the attribution of environ-mental burden between the original materialsproducers and the ultimate users. This attribution,in turn, depends on assumptions regarding antici-pated recycling prospects. In one article (Ref2.16), Clark et al. describe a technique they dub“product stream life-cycle inventory” for quanti-fying this allocation. Most importantly, thisapproach allows the analyst to consider the sen-sitivity of results to assumptions. Extendingthis sensitivity analysis to the full LCA wasdone by Newell in his Ph.D. thesis (Ref 2.17),


creating the new technique he calls Explicit LCA(XLCA).

It remains to be seen whether XLCA will beaccepted by those engaged in LCA, but strategieslike this and other steps, most notably the intro-duction of standards, will be required to trans-form this technology from a subjective one to atruly objective one. Only then can we expect tosee LCA take its place as a requisite tool on thepalette of every design engineer, and only thenwill we see materials selection done in a mannermost consistent with sustainability.

Dematerialization. Some writers suggestthat advanced technologies have already led todematerialization, a reduction in the per capitaconsumption of materials needed to sustain oursocietal economic hunger. Others are more cau-tious, because the data are hard to acquire andeven more difficult to interpret. For example, ina study of the change over time of per capitalead usage in the United States, carried out bythe United States Geological Survey (USGS)(Ref 2.18), an erroneous observation of demate-rialization could be obtained if consumptionwere estimated by tracking the mining produc-tion to manufacturer materials flow, rather thandoing the much more difficult, bottoms-upanalysis of actual consumption. What has actu-ally been happening is complicated by themovement of lead acid battery manufacture off-shore. In fact, U.S. per capita consumption hasactually increased during the period. Similarly,while we have seen more efficient use of mate-rials in home construction in the United States,lifestyle demands for leisure and comfort haveled to expanded space per person and consequentincreases in per capita materials usage. The obvi-ous dematerialization in the weight of some auto-mobiles is easily demonstrated in the reduction inbody weight of a standard passenger vehicle byone-quarter during this period. However, duringthe same period, the number of vehicles percapita has increased, and the product mix on theroad has dramatically changed. Clearly, we can-not begin to evaluate real progress toward dema-terialization without significant improvement inour worldwide materials flow database.

Reuse/Recycle. One of the great successes ofthe last quarter-century is the growth of recyclingas a natural way of life for consumers. More than80% of the states in the United States have com-prehensive recycling laws, and curbside recy-cling programs have grown from a few hundredto more than 4000. There can be no doubt thatyounger generations will be ever more desirous

of eliminating the path from use to landfill. Thereare television programs and children’s cartoonsdedicated to the subject of recycling. Thesetelevision programs, and many others like it, aresponsored in part by funds from the NationalScience Foundation (NSF) and the DoE, andrepresents an aspect of the federal role in promot-ing recycling.

Increasingly in other nations, the governmentalrole in recycling is moving from promotion tomandating, and with such a transition comesopportunity and challenge. Germany has gonefarther than most with its “take-back” legislation.The Closed Substance Cycle and Waste Man-agement Act of 1996, requires manufacturersto recover, recycle, or dispose of assembledproducts such as automobiles, electronics, andhousehold appliances when consumers retirethem. Japan has introduced similar require-ments in its Law Promoting the Utilization ofRecycled Resources. In the global economy,which now governs manufacturing, productsmanufactured in the United States will not besold in such lands if they do not conform tolocal standards, so we are already feeling theimpact of such take-back philosophy withoutany legislative mandate in this country. Sharedownership, manufacturing in the lands of sale,and now international mergers such as that ofChrysler and Daimler-Benz will accelerate thistrend. The technical, business, and political is-sues underlying this trend will demand thecontinued intense involvement of the materialsscience and engineering community.

Ultimately, the success of any recyclingprogram depends on creating markets for therecycled material, and therein we find much ofthe accomplishments of materials engineeringin the last quarter-century. Recycle/reuse of theautomobile is one of the most visible and mostcomplex of these stories. A discarded car, sentto a junkyard, is first denuded of reusable com-ponents, then crushed, shredded, and separated.Approximately 75% by weight is recycled. Theremainder, known as “fluff,” is treated as wasteand buried in landfills. This process hasspawned and depends for its success on anindustrial infrastructure composed of disassem-blers, distributors, and reusers; an infrastruc-ture which is driven by profit and sensitive tochange. This infrastructure is continually atrisk as the product changes; the cost of labor,landfill, and transportation varies; and the na-ture and market for the recovered componentsis modified.


Evolution of the automobile in response todesires to improve fuel economy and reducepollution has not only lightened the vehicle buthas also changed the materials mix. More com-plex alloy steels, while “reusable” in the sensethat they appear once again as useful products,are not truly “recycled” in the sense of beingused again and again for the same product.While this is certainly better than no reuse atall, when viewed in the context of a sustainableeconomy, it must be viewed as falling short ofthe desired goal. The current mix of materialswill soon be disrupted significantly as a newgeneration of highly fuel-efficient vehiclesemerges. Whatever the specific design andpower source of a given car, such vehiclestaken as a fleet must be lighter yet than today’s(2007) vehicles, ensuring the use of a stilllarger fraction of high-strength steels or substi-tution on a massive scale by aluminum alloysand organic-matrix composites. One of themajor bogeys to be met in any such redesignshould be an improvement in the reuse of com-ponents and true recycling of a larger fractionof materials.

Research on materials substitution in automo-biles has been a continuing subject for originalequipment manufacturers (OEMs) and theirsuppliers for many years, but in recent years,prompted in large part by the generic, precom-petitive nature of such research from theperspective of the OEMs, cooperative jointresearch ventures have become more common.As part of their more general cooperation underthe banner of USCAR, Chrysler, Ford, and Gen-eral Motors have formed a vehicle recyclingpartnership for R&D to recover and recycle ma-terials from scrap autos and to develop tools toevaluate the recyclability of new designs. It iscertain that this area of R&D will grow andspread to other manufactured products. Thechallenge is to maintain profitability and utilityin the product while sustaining the recycle/reuseinfrastructure so that the consequences of gov-ernment mandates for socially desirable goalsare not merely passed on to the consumer ashigher prices.

The U.S. Government Role—Organizational

The U.S. federal government has a role insupport of the technology base for sustainability.The organizational structure includes several

agencies, each with their own missions and eachwith their complex arrays of governing Con-gressional committees and various private sec-tor constituencies and customers. It represents asystem that may certainly better be character-ized as a collection rather than an organization.And yet, as we approach such issues as complexas sustainability, defining the proper role for thefederal government demands an organizedapproach. In the late 1990s, such a cross-agencyinvolvement was displayed in the committeeefforts that led to the policy positions on enviro-mental issues cited earlier and to more detailedefforts focused on research and on informationabout materials flows.

For some time, the efforts to achieve organi-zational approach within government have beenapplied to R&D through the Frederal Coordi-nating Council on Science and Technology, es-tablished in the mid-1970s, and the NationalScience and Technology Council (NSTC),which replaced it in 1993. Under the NSTC, thematerials R&D was coordinated by the intera-gency committee called Materials Technology(MatTec). MatTec organized around the areas offocus of the civilian technologies identified bythe NSTC and developed working groups toconsider materials needs in automotive, buildingand construction, electronics, and aeronautics.While environmental issues came up in each ofthese areas, it was felt that a more comprehensiveview of sustainability demanded a working groupdevoted to defining the issues for “materials andthe environmental.” Formed in 1996, this groupbegan to work with others to identify the cross-agency and government/private sector issues thatmust be addressed as the government role in sus-tainability evolves. During 1997 to 1998, thiscommittee collaborated with the Federation ofMaterials Societies in two workshops intendedto intensify interest in the subject, identify activ-ities that societies may fruitfully pursue individ-ually and collectively, and search for closersociety/public sector interaction.

One of the roles of interagency groups such asMatTech and Environmental Management andTechnology (EMAT) is to examine the portfolioof federal R&D programs in the relevant areas ofscience and technology; identify opportunitiesfor synergism through cooperative ventures; and,where appropriate, reveal gaps in the portfolio.No such inventory has yet been made by EMAT,so we may only make rough statements about themagnitude of federal funding in this area. Twouseful sources were used for this “analysis.”


Teich (Ref 2.19) examined the data for fiscal1995 and estimated a total of approximately$5 billion on “environmental research” as socharacterized in nondefense agency budget justi-fications. This substantial sum is concentratedprimarily in the National Aeronautics and SpaceAdministration (NASA), DoE, NSF, Departmentof Interior (DoI) and Department of Agriculture(DoA). However no matter how hard we look,we are not going to find much materials researchincluded in this total, because this surveyfocused on programs relating to pollution controland abatement, conservation, and managementof natural resources. To avoid double counting insuch inventories, agencies went out of their wayto put all materials research in the “advancedmaterials and processing” category, so what weare searching for must be found among theapproximately $2 billion included in the surveypublished by MatTech (Ref 2.20).

When this materials survey was put together,no funding breakdown identifying environmentalissues was made, so no quantitative informationcan be derived. The report does call out manyexamples of such activity, especially in the DoEand DoA, with some other examples in otheragencies. What is most apparent, however, isthat few of these examples are being justifiedprimarily because of their environmental impact,and words such as sustainability have not evenpenetrated into the vocabulary used to describethis work. It is generally believed by members ofEMAT that most of the current materials R&Dfunding, which is primarily justified as environ-mental in nature, is focused on issues associatedwith cleanup, and little is aimed at prevention.As the issues of global change become moresignificant in a political sense, we may expect tosee a relabeling of projects, currently justified bytheir energy savings, as leading to sustainableuse of materials. It would be very interesting tofollow this relabeling over the next several yearsto see what is really new in the government’sresearch portfolio. In any case, it would be desir-able to assess the current portfolio to clarify whatthe federal government is now doing to supportnew technology development for sustainable useof materials.

The U.S. Government Role—Technical

Federal Support of Civilian R&D. As weexplore the role of the federal government insupporting research that may influence sustain-

ability, we must confront head-on the continuingdebate regarding the place of federal investmentsin the private sector product arena. We all knowtoo well the use of the term corporate welfare todenigrate any direct taxpayer funding of researchthat may have an impact on corporate profit. Cer-tainly, there are enough examples of situations inwhich we already do this, so that one must recog-nize that the federal role is specific to the situa-tion, not easily generalized. Arguments favoringfederal investment in technology developmentare usually based on “market failures,” idiosyn-cracies of the technology/regulation/capital en-vironment that impede the development ofeconomically and/or socially desirable technol-ogy. New materials and new materials processingmay be an example of market failure, triggeredparticularly by the mismatch between the 10- to20-year development times of these technologiesvis-à-vis the increasingly short developmenttimes for the products that may take advantage ofsuch new materials technologies. New materialstechnologies suitable for improving our goal ofsustainable development would then appear to beideal candidates for federal investment in tech-nology development, if there is to be any at all.

By way of example, the federal role in pro-viding national security justifies not only R&Dbut also test and evaluation of actual products,which are then sold to the government. The col-lateral benefit to the aircraft industry and itscommercial return to its stockholders is a pleas-ant side benefit that seems to bother none of thefree-marketeer critics of corporate welfare. Simi-larly, in the field of agriculture, a federal role haslong been recognized, supported by the extensiveresearch establishment of the Agriculture Depart-ment and augmented by the many states throughthe land-grant college system. What then shouldbe the federal role in supporting the developmentof environmental technology?

Certainly, basic research and even applied re-search with a focus on more environmentallyfriendly technologies should be encouraged.When we get to product, the answer lies inassessing whether the desired social good—inthis case, a more environmentally friendly prod-uct—can be expected without the involvementof the government. If the answer to that ques-tion is no, a government mission exists and maybe addressed by one of three alternatives: taxes(financial incentive to consumer or disincentiveto producer to follow desirable paths), regula-tions (product guidelines mandated to achievethe desired end with technology choice left to


the producer), or assistance in technologydevelopment (recognizing that the free marketwill not invest if no near-term profit is believedto be forthcoming).

On the automotive scene, we see the mix ofthese three policy choices vying for ascendencyin our contentious political arena. Regulationhas been the predominant strategy in the UnitedStates for the last quarter-century, with signifi-cant visible results in both cleaner emission andlower fuel consumption. This path is by nomeans exhausted because new clean air regula-tions continue to be debated in the Congress,and many individual states place ever greaterrestrictions on the allowable emissions from ve-hicles. In most of the rest of the world, taxationhas also been a significant factor in governmen-tal attempts to minimize the ecological impactand energy usage for transportation. With higherresultant energy costs to the consumer as a keydriver, smaller, fuel-efficient vehicles are morereadily marketable. The political obstacles toachieving even a 5-cent increase on gasolinemake such an approach unlikely to play a signif-icant role in the United States.

Finally, in recent years, the U.S. federal gov-ernment has expanded its research agenda toinclude the commercial automotive arena. First,through several unconnected efforts in the DoE,National Institute of Standards and Technology(NIST), and elsewhere, and then through thehighly visible PNGV, the government and theautomotive Big Three joined in a historic partner-ship to develop technology that would satisfy theindividual customer’s desire for transportationand the collective societal desire for a “greener”vehicle. The success of this partnership in both atechnical and a political sense will have profoundimplications for the continued participation insuch research in other industrial sectors. It isworth looking in a bit more detail at the technicalopportunities and challenges in the PNGV.

The challenge of the PNGV can be summa-rized in three goal statements: advancement ofmanufacturing practices, implementation of in-novations on current vehicles, and developmentof a vehicle with up to 3 times current fuel effi-ciency. It is this third goal and its restrictionsthat make up the grand challenge of the PNGV.It is possible today (2007) to build a passengervehicle with 3 times the fuel efficiency of today’spassenger sedan, but the formidable challenge setby the PNGV is to do this while maintaining thepassenger and storage capacities; satisfying thedriver’s desires for speed, acceleration, and

driving range; meeting all existing emission reg-ulations; achieving recyclability of 80%; main-taining manufacturability; and perhaps, mostdaunting, maintaining affordability. To achievethis result will require dramatic mass reductionof the vehicle and simultaneous significant in-creases in power source efficiency. The designspace allows for many solutions to this problem,and we expect that each of the Big Three willfind its own unique combination of parameters.Nevertheless, it seems apparent that vehicularweight reductions of the order of 40% will berequired. One scenario for distribution of theweight reductions among vehicle components isdisplayed in Table 2.2. Such dramatic reductionsin weight can only be achieved through majormaterials substitution and with significant designand manufacturing changes. Some of the re-search topics under study are summarized inTable 2.3. This list leaves no doubt that ouragenda remains formidable. One of the mostcomplex tasks facing the materials and manufac-turing communities is to achieve the requirementof recyclability while further increasing the mixof materials. The competition now underwayamong the principal contestants for such alterna-tive technologies has added new vitality to themetals and polymer composites industries andwill certainly lead to significant technologicalchange in automobile manufacture for years tocome.

Energy and the Environment. Our attentionwas diverted from environmental issues in themid-1970s by the development of the so-called“energy crisis,” a supply/demand trauma attrib-utable to the price escalation by the Organizationof the Petroleum Exporting Countries (OPEC), anear-monopolistic trade organization. As weexplored alternative energy sources, new andimproved materials came to the forefront. Wefound that no new energy technology was goingto become a serious price competitor to then-dominant fuels without dramatic reductions in

Table 2.2 Vehicle mass reduction targets forPartnership for New-Generation Vehicles(PNGV) goal 3

Current PNGVvehicle vehicle target

MassSystem kg lb kg lb reduction,%

Body 514 1134 257 566 50Chassis 499 1101 249 550 50Powertrain 394 868 354 781 10Fuel/other 62 137 29 63 55Curb weight 1470 3240 889 1960 40


materials costs or improvements in conversionefficiencies. As the “crisis” passed, and with thearrival of the Reagan administration, budgets fellat the DoE, and the focus on alternative energysources was reduced. Has anything changed inthe subsequent 15 years? There is certainly nodoubt that petroleum prices will someday rise assupplies decrease—only the timeframe is de-bated. Published studies by Campbell and La-herrère (Ref 2.21) and MacKenzie (Ref 2.22)conclude that the peak of global production ofconventional oil is only one or at most twodecades away. With the subsequent decline inproduction and increase in price, the pressure foralternative sources of energy will accelerate.Among those alternatives, we continue to listnatural gas and coal-derived fuels, but are thesestill realistic alternatives?

In the last several years, we have come to rec-ognize that the accumulation of greenhousegases in the atmosphere will very likely causesubstantial climactic change. While much moreneeds to be done to clarify the consequences ofsuch change, public policy is already movingtoward control of greenhouse gas emissions.With this new reality in mind, our attentionshould increasingly focus on alternatives to fossilfuels as the only long-term sustainable strategyfor expanded energy needs of a growing popula-tion with growing per capita economic progress.Consequently, in an even more intense way thanearlier, the issues of energy and environmenthave become intertwined. Not surprisingly, theDoE is the largest source of funds for environ-mental programs in the federal government. Noris it surprising that the strong dependency of en-ergy technology on materials technology has

made the DoE by far the largest federal funder ofmaterials science and engineering. The range oftechnical programs in the DoE portfolio is far toobroad to be covered in such a brief overview asthis, because it includes some level of effort in allfossil and nonfossil alternatives to petroleum asan energy source. The DoE is determined to takea lead role in the development of new technologyfor sustainability and has brought the efforts ofseveral of its national laboratories to bear on thesubject of R&D strategies. The volume entitled“Scenarios of U.S. Carbon Reductions: Poten-tial Impacts of Energy Technologies by 2010and Beyond” (Ref 2.23) is a gold mine of ideasand R&D needs, with many focused on materi-als issues. A second excellent summary of R&Dopportunities was produced by a joint effort ofthe DoE and NSF. This report on “Basic Re-search Needs for Environmentally ResponsiveTechnologies of the Future” (Ref 2.14) linksthe needs to the various industrial sectors andgives significant attention to the materialsresearch agenda. Rather than attempt to sum-marize these technical options, the focus here isinstead on the public-private interaction neededto bring such technology to bear on the issue ofsustainability.

During the 1980s, the general antipathy to-ward “demonstration” projects left over fromthe years of the energy crisis was followed by abipartisan recognition that new methods neededto be developed to garner the fruits of federallyfunded research. A series of experiments arenow underway in an effort to link the variousgovernment programs more closely with indus-try. Many of these programs remain controver-sial, and the demise of the DoD’s Technology

Table 2.3 Partnership for New-Generation Vehicles widely applicable materials—ChallengesPolymer composite Polymer composite

Priority of challenges Aluminum Magnesium components body Steel

High Feedstock cost Feedstock cost Low-cost carbon Low-cost carbonfiber fiber

High-volume Weight-reductionmanufacturing concepts

High-volume Joiningmanufacturing

High-temperature alloy Analytical design Analytical designIn-service inspection In-service inspection

RecyclingMedium Casting Recycling Lightweight

Forming Improved design and technology manufacturing (incremental)

Joining Properties in service Properties in serviceRecycling Machining

RecyclingLow Extrusion


Reinvestment Program and reduced funding forthe DoE’s Cooperative Research and Develop-ment Agreements and the Department of Com-merce’s Advanced Technology Program areindications that these programs are still viewedas experimental by many in Congress. Whatevermay come of the funding of these programs inthe future, they have had significant impacts onmany of the materials producers and users andwill likely continue to do so in the future, if notthrough direct funding, then through changes in business practice. This is particularly true for many of the materials-producing and -processing industries with their disaggregatedorganizational structures.

One particular activity of note in the DoE wasin the Office of Industrial Technology (OIT),which had an exciting new program called theIndustries of the Future Program. This effortinvolved federal-private partnerships with sevenindustrial communities in the development ofvisions for their future development and tech-nology roadmaps required to get there. We aremost familiar with the roadmap concept in theelectronics industry, where, guided by the rate-determining predictions of Moore’s law, thatindustry can look 10 years ahead, define wherethe industry’s technology will be, and thendevelop a map of development work necessaryto get to that desired endpoint.

The DoE-OIT selected for its industrial part-ners the seven most energy-intensive industrialsectors, collectively using more than 60% of theenergy consumed in the United States. Theseindustries are also among the most intensive con-tributors to other waste streams besides green-house gases. Partnerships have been developedwith the aluminum, chemical, forest products,glass, metal casting, steel, and agricultureindustries, with supplementary agreements withpetroleum refining, heat treating, and forging.The heart of the materials-producing commu-nity is displayed in this list. Thus, as theseindustries develop their roadmaps, identifyingtheir technology needs for the future, they arealso laying out a rough outline of the materialsresearch needed to achieve sustainability in thenext decades.

These roadmaps, unguided by any Moore’slaw for the metals and chemical industries, areless well defined than that of the electronicsindustry, and a great deal of work lies aheadbefore they will yield the required detail ofresearch agenda. These roadmaps are now publicdocuments, available through the internet and

from the industries and the DoE. Among othercommon features, they all share a strong com-mitment to sustainable development. As oneexample, consider statements made by the Steelroadmap in the chapter entitled “Environment”(Ref 2.24).

Over the past 25 years, the investments of $6billion on capital investments on environmentalprojects have led to reduced discharges of airand water pollutants by 90% and a reduction ofsolid waste production by more than 80%. Nev-ertheless, “further improvements to pollutionprevention technologies are needed to reducecosts, improve profitability, and facilitate com-pliance with changing Federal regulations. Thesteel industry’s goal is ‘to achieve further reduc-tions in air and water emissions and generationof hazardous wastes,’ and the development ofprocesses ‘designed to avoid pollution ratherthan control and treat it.’” The report then goeson to list desired technical developments incokemaking, ironmaking, steelmaking, refiningand casting, forming, and finishing in this 34-pagechapter on environmental technologies.

The Industries of the Future TechnologyRoadmaps represent a fertile area for planning,not only for the industry itself but also for theuniversity and government organizations thatwould interact with these industries. Materialsscientists and engineers must “mine” these plansfor the concepts and then fill in the details to de-velop a materials research agenda. Professionalsocieties also are deeply involved in developingroadmaps. The TMS and ASM Internationalhave both been playing a role with the metalsindustries, and part of our societal agenda shouldbe to assure that we will hear more about theseefforts in the meeting sessions and hallways inyears to come. We must make our meetings thetechnical home for the results of R&D in theIndustries of the Futures Program.

Materials Flow Data. The collective actionof the federal government has been of benefit toindustry and the citizenry in yet another technicalarena of relevance to this discussion: the area ofinformation. Many organizations within govern-ment assemble, organize, and disperse informa-tion of a technical nature. Most visible to us asscientists and engineers has been evaluatedtechnical data such as thermodynamic, crystal-lographic, or materials properties, which allowsthe scientific enterprise to proceed with com-mon basis for quantification. One concern inthis presentation is with another broad arena ofdata: data on materials flow. When we think of


the materials cycle, it is usually in a qualitativesense, but a small, growing number of our col-leagues are beginning to look at the full lifecycle of particular materials in a quantitativesense. Understanding how materials are de-rived, used, and disposed of may often focus ourattention on alternative strategies and technolo-gies to achieve an environmentally friendly andsustainable manufacturing enterprise. There aresome economists who believe that in the future,materials flows will develop as much impor-tance in economic analysis as have energy flowsin the last several decades.

The focus on materials flow within the gov-ernment is centered now in the EPA and in theUSGS, conducted there by a few folks who areamong the remnants of the gone-but-not-forgot-ten Bureau of Mines. However, there is interestin this subject in many other agencies. An inter-agency working group on industrial ecology,material, and energy flows has constructed areport on materials flow (Ref 2.25). This docu-ment significantly influences the visibility andsignificance this topic will receive in future gov-ernment effort.

We need to be “mining” these data in twosenses: on the one hand, by looking for opportu-nities to identify research needs and, on theother, by literally mining in the sense of identi-fying materials flows that will be sources of“raw” materials for processing. Consider the ex-ample of silver in water, sediment, and tissue offish and marine mammals in the San FranciscoBay (Ref 2.25). The source, located after adetailed materials flow study by the Universityof California at Los Angeles, was traceable tophotographic and radiographic materials usedby the service sector, including dentists, x-raylabs, hospitals, photo shops, and so on, not toheavy industry. The solution was found, in part,through changes in the regulatory system and, inpart, through the development of cost-effectiveprocessing and recovery systems. One suchexample is the establishment by Kaiser Perma-nente of a centralized silver recovery and fixerreprocessing plant in Northern California. Thiswas a profit-making system that reduced silverloadings to waste treatment plants and water-ways, and paid for itself in less than 1 year.

This is only one example of what could be amyriad of profit-making efforts to recover valu-able minerals from what are now considered tobe waste streams. Allen and Behmanesh (Ref2.26) encourage the view that these wastestreams are raw materials that are often signifi-

cantly underused. They emphasize that one ofthe research challenges of industrial ecologywill be to identify productive uses for suchmaterials currently considered wastes. In theiranalysis, they focus on materials flow datagenerated by the EPA from the National Haz-ardous Waste Survey but emphasize that exist-ing data represent only 5 to 10% of the totalflow of industrial wastes. Detailed informationabout concentrations in the waste streams com-pared with prices for raw materials forces theconclusion that the concentrations of metalresources in many waste streams currentlyundergoing disposal are higher than for typicalvirgin resources. Among these materials arelead, antimony, mercury, and selenium. Withappropriate R&D, we may expect to find moresymbiotic developments in which small recoveryplants are sited at the sources of these wastestreams. We may also expect that, as in the caseof silver in the San Francisco Bay, R&D willlikely have to be supplemented by changes in theregulatory policies to enable efficient handling ofthese materials as potential resources, not wastes.

The Role of Professional Societies

Each of the individual professional societieswith an interest in materials carries out activi-ties focused on their particular constituency toa degree consistent with their resources andtheir perception of their member needs. In thearena of materials and sustainability, as in somany other areas of broad common interest, theindividual efforts are significant but seem to befar less than what may be accomplished if wecould find ways to pool our resources and tal-ents. The Federation of Materials Societies(FMS), formed for just such collective action,has been just as strong and effective as itsmember societies are willing to make it. Thishas meant that it is not nearly as strong and ef-fective as it needs to be. This area of materialsand sustainability may be one in which ourcommon interest will override our competitivenatures to that degree required for some seriouscollective action.

The FMS has begun to set the stage for suchaction in its strategic planning and through threemeetings held in the late 1990s. In the first twoof these meetings, workshops were organizedjointly with the interagency government groupEMAT. At the December 1996 meeting, societyand government agency representatives laid out


a map of their current efforts on environmentalissues in the materials arena. It was a sort of“get-to-know you” meeting, one which was use-ful in identifying the “best practices” that couldbe found in each of the professional societies.The report of that meeting, available to all par-ticipants and to other societies that request it,can act as a template against which to examineone’s own efforts and a challenge to achieve thelevel of best practice in each society.

The second FMS-EMAT workshop, held inDecember 1997, targeted the technical arenaand identified action items for government andsocieties. The third meeting was held in May1998. At this, the 15th Biennial on MaterialsPolicy, the general theme was “Maximizing Re-turn on Investment in R&D: Case Studies inMaterials,” but materials and the environmentwere significant in the discussion. Here too,calls for action resulted, but no plan for actionemerged.

This has too often been our history, as weusually tend to put too many items on theagenda for action and are then unable to beginany one of them. Furthermore, while it is allwell and good to suggest things for others to do,it is time for us to take action ourselves. Wemust search for one good area where cooperationamong the societies can be carried through andjust begin together and do it. A specific sugges-tion for action is included in the summary, butthis section lists some areas of current societyactivity and challenges the reader to determinewhether ASM International and TMS may bedoing more than they currently do.

Professional societies all engage in promotingthe R&D agenda, in education and in publicity,but not all use the same mechanisms and not allhave environmental activity in each area. Thestrongest efforts include some central commit-tee(s) and staff with well-defined responsibilitiesfor environmental activity. Promoting the R&Dagenda includes the following: helping to de-fine that agenda, for example, by participatingin roadmapping exercises; creating the forumfor discussing results, for example, meetings,workshops, and trade shows; making informa-tion widely available, for example, throughpublications of technical proceedings and or-ganized data; and, in several instances, directlyfacilitating needed R&D, for example, by actingas manager for federally- or industry-fundedprojects. Some societies have workshops or sym-posia focused on environmental issues at everyannual meeting; others do so rarely, if at all.

Educational activities include kindergartento high school and beyond, including specificvocational training, and include formal curricularinvolvement as well as preparation of materialthat may supplement the core curriculum. Somesocieties now incorporate environmental activityin such educational efforts, while many do littlemore than pay lip service to the link betweenmaterials issues and the environment.

By publicity, that means telling our story to theworld. Besides being an element of education,there is a special need for communication withindividuals who know little about the scienceand technology that have played such a signifi-cant part in providing the standard of living towhich we have now become accustomed. Suchindividuals, be they voters or heads of powerfulCongressional appropriations committees, mustunderstand the positive role materials has playedand can continue to play, or we will fail in anyattempt to achieve sustainable development. Sci-entists and engineers have not had a very goodtrack record in this publicity arena, but manysocieties have turned their attention to this effortof public education. Specifically, the prepara-tion of “success stories” and the direct contactwith Congressional representatives and staff canhelp tell them what we have done for them latelyand what we may yet do if given the opportu-nity. Many of the materials societies are stillbehind the curve in this activity, and in only afew cases have the successes of our efforts beenfocused on the positives to the environment.There is much yet to be done here.

Summary and Recommendations

In summary of efforts over the last quarter ofthe twentieth century, how are we doing? Ingeneral, we have made considerable progress.We have increased lifetimes for many materials,especially when we include recycling as anextension of total “life.” We have cleaned up theprocessing lines and reduced scrap considerablyin many instances, and we have lightened trans-portation vehicles through some extensive ma-terials substitution. However, it should be clearfrom earlier remarks that we have mostly beengathering low-hanging fruit. If we are to reallyachieve a worldwide sustainable manufacturingparadigm, the hardest work is ahead. Some ofthat work will be technical, but much will beorganizational. We have a pretty good trackrecord on the technical agendas that do not need


any further discussion here. The organizationalagenda is a major concern and should be ad-dressed in two arenas—government and profes-sional societies.

On the government side, it is time to revitalizethe environmental technology initiative focusedon addressing the technical opportunities andneeds associated with sustainability. Heaton andBanks (Ref 2.13) argue persuasively for such aneffort and suggest three reasons why we cannotdepend solely on the private sector to producesuch technologies. First, the current regulatorysystem acts as an impediment both to the inno-vation and the long-term predictability thatR&D investments demand; second, much of theenvironmental R&D that does occur is concen-trated in the large firms, beyond the reach ofmany small companies and disaggregated in-dustries most in need of improvement; andthird, the kinds of research that may help mostover the long term often produce results of suchgeneric applicability that no single companycan justify the investment. The federal invest-ment seems justified in such instances as oneelement of a strategy to achieve the desiredsocial goal; however, this time, it should not belodged within the EPA, but rather in the DoE orNIST Advanced Technology Program. Thiswould eliminate the conflict between regulationand the best technical solutions, which alwaysplagues any agency. In anticipation of such aninitiative and to clarify the present baseline, theMatTec/EMAT committee should carry out aninventory of materials R&D in support of envi-ronmental sustainability. Even with all of the un-certainties inherent in such an analysis of exist-ing programs, most of them currently justified onsome other, mission-focused basis, a beginningmust be made to clarify the likely impacts of thegovernment’s investment.

What should we demand of our professionalsocieties? We have seen in the FMS-sponsoreddiscussions over the last several years that thereare an enormous number of environmental-related activities underway when one examinesall of the societies collectively, but few are doingall things, and many things are less than ade-quately covered even by the separate, disaggre-gated efforts of the individual societies. We needmore technical activity and we need to do moreto tell the story of what we have already accom-plished. On the technical side, we need to give ahome to the R&D activity on sustainability,through special symposia and in our regularmeetings. In many instances, this may require

collaboration among several societies to broadenthe coverage of science through engineering orrange of materials, and, in all cases, it willrequire more visibility and overt statement ofthe agenda of sustainability. The technicalagenda also extends to the arena of helping inthe process of identifying the needed researchagenda. Here, the most significant needed stepswill require closer cooperation between the pro-fessional societies and trade associations. Wehave seen significant steps here in the AmericanCeramic Society, the Heat Treat and ThermalSpray Societies within ASM International, theAmerican Welding Society, and the AmericanFoundry Society. Whether it be through such for-mal arrangements or through informal coopera-tion, as happened with the chemical industryassociations and the American Chemical Societyand the American Institute of Chemical Engi-neers in the development of the chemical indus-try roadmap, such future vision exercises demandan industrial perspective to give broad outlines,followed by detailed technical analysis to trans-late those guidelines into useful bases for actionby researchers.

Finally, and to close with a specific call foraction, we must tell the world about our suc-cesses. For several years, stimulated largely bythreats against continuing funding of science,several large societies, most notably the Ameri-can Chemical Society and American PhysicalSociety have been working to develop carefullycrafted, well-documented, graphically attractiveexamples of how basic research in their fields hasled to technological advances that the generalpublic would recognize as positively affectingtheir lives. In general, while those efforts haveincluded some materials examples, the materi-als story has yet to be told. Several materials so-cieties have made starts on such an effort, butthe high cost and too diffuse an agenda haveboth contributed to a weak showing. On theother hand, a focused effort on success storiesabout materials and sustainability carried outjointly by several of our societies and coordi-nated by the FMS can succeed and would be ofsignificant value on several fronts. It wouldincrease the armamenta that we need to makethe case on the importance of materials researchto all levels of government; it would help usclarify to the public at large and to young peo-ple in particular that we wear the “white hats” inthe environmental arena; it would provide aresource for educational use in early schoolyears; and it would provide a solid, well-defined


project for the diverse members of the FMS toengage in to test, once and for all, if we arecapable of working together for our commongood. Such a project would require resourcesthat no single society is likely to advance butthat would be feasible through collective action,particularly if the trade associations join theaction. This idea has already been brought up at the FMS Biennial meeting with generallypositive reactions from those assembled, butothers, representing the FMS and its membersocieties, will have to come together to giveflesh to this skeleton of an idea.

The targets are becoming clear; our role iscentral; it is time for the materials community toaccept the burden and step forward as leaders inthe quest for sustainable development.

REFERENCES

2.1. H. Brooks, Metall. Trans., Vol 3, 1972,p 759–768

2.2. “Final Report of The National Commis-sion on Materials Policy,” Library ofCongress Card No. 73-600202, June 1973

2.3. “Report of the Committee on the Surveyof Materials Science and Engineering,”National Academy of Sciences, Washing-ton, D.C., 1974, followed by four appen-dixes published in 1975

2.4. “Final Report of The National Commis-sion on Materials Policy,” June 1973, p1–4

2.5. J. Boyd, Metall. Trans., Vol 5, 1974, p5–10

2.6. M. Tenenbaum, Metall. Trans. A, Vol 7,1976, p 339–350

2.7. H.H. Kellogg, Metall. Trans. A, Vol 9,1978, p 1695–1704

2.8. “Report of the Committee on MaterialsScience and Engineering,” NationalAcademy Press, Washington, D.C., 1989

2.9. J.H. Ausubel, D.G. Victor, and I.K. Wer-nick, Consequen. Nature Implic. Environ.Change, Vol 1 (No. 3), 1995, p 2–15

2.10. R.A. Frosch and M. Uenohara, IndustrialEcology: US/Japan Perspectives, Na-tional Academy of Engineering, Wash-ington, D.C., 1994, p 2

2.11. “Technology for a Sustainable Future,”Report of the National Science and Tech-nology Council, Office of Science andTechnology Policy, U.S. Govt. PrintingOffice, July 1994

2.12. “Bridge to a Sustainable Future,” Reportof the National Science and TechnologyCouncil, Office of Science and Technol-ogy Policy, U.S. Govt. Printing Office,April 1995

2.13 G.R. Heaton, Jr. and R.D. Banks, in In-vesting in Innovation, L.M. Branscomband J.H. Keller, Ed., MIT Press, Cam-bridge, MA, 1998, p 276, 292

2.14 P.M. Eisenberger, Ed., Basic ResearchNeeds for Environmentally ResponsiveTechnologies of the Future, Materials In-stitute, Princeton, NJ, 1996

2.15. J. Williams, Metall. Mater. Trans. A, Vol29, 1998, p 0000–00

2.16. J. Clark, S. Newell, and F. Field, LifeCycle Engineering of Passenger Cars,VDI Verlag GmbH, Duesseldorf, 1996, p1–19

2.17. S.A. Newell, Ph.D. thesis, MIT, June1998

2.18. M. Sullivan and L. Wagner, “MaterialsFlow Studies, Total Materials Consump-tion—An Estimation Methodology andExample Using Lead. U.S. GeologicalSurvey,” presented at the Eighth Int.Symp. on Raw Materials, Federal Insti-tute for Geosciences and Natural Re-sources, Hannover, Germany, 1997

2.19. A.H. Teich, Linking Science and Technol-ogy to Society’s Environmental Goals,National Academy Press, Washington,D.C., 1966, p 345

2.20. “The Federal R&D Program in MaterialsScience and Engineering, Data for Fiscal1995,” www.msel.nist.gov

2.21. C.J. Campbell and J.H. Laherrère, Sci.Am., March 1998, p 78–83

2.22. J.J. MacKenzie, Iss. Sci. Technol., Sum-mer 1996, p 48–54

2.23. Scenarios of U.S. Carbon Reductions,Office of Energy Efficiency and Renew-able Energy, DoE, Washington, D.C.,1998

2.24. “The Steel Industry TechnologyRoadmap,” The American Iron and SteelInstitute

2.25. “Materials,” Report of the InteragencyWorkgroup on Industrial Ecology, Mate-rials and Energy Flows, Fall 1998. Forcopies, contact 1-800-363-3732

2.26. D.T. Allen and N. Behmanesh, TheGreening of Industrial Ecosystems,National Academy Press, Washington,D.C., 1994, p 69–89

LIFE-CYCLE ASSESSMENT (LCA) is amethodology that uses a systems approach tounderstand the environmental consequences ofa product, process, or activity from initial extrac-tion of raw materials from the Earth until thepoint at which all residuals are returned to theEarth. The goal of LCA is to quantify, evaluate,and then identify opportunities to reduce theoverall environmental effects of the systemunder study.

The LCA methodology, as defined by theInternational Standards Organization (ISO), istypically divided into four separate and interre-lated components:

• Life-cycle scoping and goal definition:Includes the clear statement of the purpose ofthe study, the system to be studied, the in-tended use of the results, limitations on itsuse for other purposes, data quality goals, re-porting requirements, and the relevant typeof review process. The scope also defines adescription of the geographical and temporalboundaries, system boundaries, data require-ments, decision rules, and other assumptions.

• Life-cycle inventory (LCI) analysis: Thephase of LCA involving the compilation andquantification of inputs and outputs throughthe life cycle of a product or service, includ-ing the stages of resource extraction, manu-facturing, distribution, use, recycling, andultimate disposal.

• Life-cycle impact assessment: The phase ofthe LCA aimed at understanding and evalu-ating the magnitude and significance of the

potential environmental impacts of a productsystem.

• Life-cycle interpretation: The phase of theLCA technique in which the findings of theinventory analysis and impact assessmentare combined in line with the defined goalsand scope. The findings may take the formof conclusions and recommendations to de-cision makers, consistent with the goal andscope of the study.

This LCI study of the North American alu-minum industry has the objective of generatinga resource consumption and environmental pro-file of the aluminum-producing industry. Thepurposes are to:

• Provide the participating companies withdetailed LCIs of the various processes andactivities within the aluminum product lifecycle. This information provides a baselinefor improvements in the management ofenergy, raw material use, waste elimination,and the reduction of air and water emissions.

• Develop a comprehensive and up-to-dateLCI of primary and secondary aluminumproducts produced in North America for theUnited States Automotive Materials Partner-ship (USAMP) initiative.

The domain of this study on behalf of theNorth American automotive industry aluminumsystem includes operations in the chain of plantsthat supply metal to the Big Three automobilemanufacturers. This study contains data col-lected from 213 reporting locations representing15 unit operations. The specific unit operationsfor the primary industry are shown in Fig. 3.1and for the secondary industry in Fig. 3.2 Unit

CHAPTER 3

Life-Cycle Inventory Analysis of theNorth American Aluminum Industry*

*Adapted with permission from The Aluminum AssociationReport AT-2: “Life Cycle Inventory Report for The NorthAmerican Aluminum Industry,” November 1998




operations ranged from bauxite mining in Aus-tralia, Africa, Brazil, and Jamaica to primary andsecondary processes and manufacturing forNorth America. For each reporting location, in-formation on energy and materials used, waterconsumption, and releases to air, water, and landwas gathered. The study follows the USAMPLCI methodology (Ref 3.1) developed by Roy F.Weston, Inc. and ISO 14000 series (Ref 3.2)LCA guidelines.

A key element of this study is the scope of thedata-gathering efforts. The results of this LCIreport are based on input and output data of en-vironmental relevance collected at productionfacilities. This information was gathered by theAluminum Association, Inc. and Association del’Industrie de l’Aluminium of Canada from 13primary and secondary aluminum-producingcompanies in North America. The data were as-sessed in detail by both Weston Inc. and the

Manufacturescrap

transport

Shreddedaluminumtransport

Consumerscrap

transport

Shreddingand

decoating

Secondaryingot

casting

Toaluminumextruding

Toaluminum

rolling

Toshapecasting

Frompart

manufacture

Fromcomponentdismantling

Fromhulk

shredder

Fig. 3.2 Simplified process flow diagram for secondary aluminum products

Electricityproduction

Bauxitemining

Anodeproduction

Aluminarefining

Aluminumsmelting

Ingotcasting

Hotrolling

Coldrolling

Shapecasting

Aluminumextruding

Primaryextrudedaluminum

Primarycast

aluminum

Primaryrolled

aluminum

Fig. 3.1 Simplified process flow diagram for primary aluminum products

Chapter 3: Life-Cycle Inventory Analysis of the North American Aluminum Industry / 35

Aluminum Association LCA Task Force. Apanel of external recognized life-cycle expertsselected by the USCAR steering committee alsoreviewed the report.

The results of the LCI have confirmed andreinforced the progress that the aluminum in-dustry has been making, especially in regard toenergy efficiency, recycling, and the reductionof CO2 emissions. Improvements in smeltercell operation and practice have continued toimprove overall energy efficiency and have sig-nificantly lowered the releases of gases withglobal-warming potential. The enormous sig-nificance of recycling, which now has grown torepresent approximately 37% (1995 figure) ofthe metal supply, has again been emphasized,because the LCI demonstrates the great energybenefits associated with recycling aluminum ascompared to producing aluminum from thebauxite ore. This comparison is illustrated inFig. 3.3 for the integrated performance of theNorth American industry with regard to energy-use components for both primary and secondaryaluminum ingots.

The results of this LCI study are intended to:

• Establish a database describing the resourceconsumption and environmental releases ofmaterials and processes

• Improve understanding of the environmentalimplications of aluminum product manufac-ture

• Facilitate assessment of the life-cycle envi-ronmental inventory of alternative productdesign options (for example, alternativefuels, process design, etc.), compare corre-sponding data sets, and guide the evaluationof modifications for improvement

• Provide environmental information for usein strategic planning

In terms of limitations, there is potential formisuse of the database and results. Specifically,the inventory database and results shall not be:

• Used as the sole criterion in raw materialsselection decisions

• Quoted or published without prior consentof The Aluminum Association, Inc., exceptin the case where a participant* or designatewished to quote or publish their own sub-mission of data

• Used as a basis for impact assessment

Life-Cycle Inventory Methodology

Life-cycle inventory methodology is reason-ably well established. The techniques that havebeen used to model other systems are directly ap-plicable to almost all aspects of the product sys-tem. However, LCI is assumption based, andmany decisions are involved in the developmentof an inventory. Therefore, for the credibility ofthe results and conclusions drawn from the study,it is essential to have clearly defined assump-tions, decision rules, and data quality assessment.

Aluminum product systems include:

• Primary ingot• Primary rolled aluminum• Primary extruded aluminum

2%

71%

6%

21%

57%

2%13%

27%

Primary ingotTotal energy per 1000 kg=186,262 MJ

FeedstockProcess: nonelectric Process: electric Transportation

Secondary IngotTotal energy per 1000 kg=11,690 MJ

Fig. 3.3 Energy consumption for primary and secondary ingots

*The Aluminum Association member companies thatparticipated in this study include: Alcan Aluminum, Ltd.;Alumax, Inc.; Aluminum Company of America (ALCOA);Commonwealth Aluminum; Kaiser Aluminum & ChemicalCorporation; Century Aluminum Corporation; ReynoldsMetals Company; Southwire Company; Wabash Alloys;and Werner Company. The members of Association del’Industrie de l’Aluminium of Canada that participatedinclude: Aluminerie Alouette Inc.; Aluminerie de Bécan-cour, Inc.; Aluminerie Lauralco Inc.; and CanadianReynolds Metals, Inc.


• Primary cast aluminum• Secondary ingot• Secondary rolled aluminum• Secondary extruded aluminum• Secondary cast aluminum

The simplified process flow diagrams foreach of these processes are shown in Fig. 3.1and 3.2. The results of the rolled-up data foreach of these products (modeled for 1000 kg ofthe product) are presented in the section “Re-sults by Product System” in this chapter.

In a comprehensive cradle-to-gate analysis,the system’s “envelope” includes the extractionof raw materials from the Earth (bauxite min-ing), materials processing (anode production,alumina refining, aluminum smelting, primaryingot casting), manufacture (hot and cold rolling,extruding, and aluminum casting), and recov-ery/recycling (consumer aluminum transport,manufactured aluminum transport, shreddedaluminum transport, shredding and decoating,and secondary ingot casting).

All of the primary data provided, with oneexception, were from the calendar year 1995.One location in the bauxite mining unit processwas from 1992 to 1993 from a previous alu-minum industry study that met the data qualityrequirements for this study. Data for ancillarymaterials flows that were included in the studywere taken from the Weston Inc. database repre-senting data that were, for the most part, from1995 or 1996. All ancillary data satisfied therequirements defined by the methodology.

Geographic Coverage

The goal of the USAMP study was to develop acomprehensive LCI of the North Americangeneric vehicle. As such, the initial step in thealuminum analysis of the domain for this studyincluded an evaluation of the sources of supplyto the automotive industry for semifabricatedand fabricated products (i.e., rolled, extruded,and cast aluminum). With respect to automotiverolled aluminum, the aluminum industry taskforce evaluated the North American rollingplants to determine those that were capable ofsupplying the auto industry. Only these plantswere designated as part of the domain. When itcame to the extruded and cast aluminum supply,there were numerous plants that could qualifyfor inclusion in the domain. For reasons of prac-ticality, and recognizing that there is only minorvariation in the extruding and casting processes,

only a sample of these operations was selectedfor inclusion in the domain.

The next step in the analysis was to definethe source of aluminum ingot supply to thefabrication processes; these included both pri-mary and secondary sources. This process ofdefining the supply chain within the NorthAmerican domain was extended to the bauxitemines that ultimately feed the automotive sub-systems. The domain for the subsystems andthe geographic coverage is summarized inTable 3.1.

Technology Coverage

The domain identified for the study representedthe full range of technological differences in thevarious unit processes. These differences weredefined in an analysis of the available technolo-gies, and the review of actual reporting loca-tions did not indicate any biases or distortionsfor one type of technology versus another. Asummary of the unit process descriptions maybe found in Table 3.2 and the following text.

Table 3.1 Domain and geographical coverageGeographical Total by unit

Source Unit process regions process

Primary Bauxite mining Australia, Africa, 10Brazil, Jamaica

Electricity North America 8generation

Alumina Australia, Brazil, 13refining Germany, Jamaica,

United StatesAnode North America 27

productionAluminum North America 29

smeltingPrimary ingot North America 23

castingSubtotal 110

Manufacturing Hot rolling North America 14Aluminum North America 21

extrudingCold rolling North America 17Aluminum North America 6

castingSubtotal 58

Secondary Manufactured North America 9scrap transport

Consumer scrap North America 12transport

Shredded North America 12aluminum

transportShredding and North America 13

decoatingSecondary ingot North America 39

casting and melting

Subtotal 85

Total 253


Bauxite mining activities mainly take place intropical and subtropical areas of the Earth. Mostall bauxite is mined in an open-pit mine. Theknown reserves of aluminous ore will sustain thepresent rate of mining for 300 to 400 years.

Commercial bauxite can be separated intobauxite composed of mostly alumina trihydrates

and those composed of alumina monohydrates.The trihydrate aluminas contain approximately50% alumina by weight, while monohydrates areapproximately 30%. Monohydrates are normallyfound close to the surface in grassy plains (e.g.,Australia), while trihydrates tend to be at deeperlevels and in heavily forested areas (e.g., Brazil).

Table 3.2 Unit process description in aluminum life-cycle inventory analysisUnit process Operations of the unit process Output

Bauxite mining • Extraction of bauxite-rich minerals for the site Bauxite that is transported to an• Beneficiation activities, such as washing, screening, or drying alumina refinery• Treatment of mining site residues and waste• Site restoration activities, such as grading, dressing, and planting• Excludes material, energy, and releases associated with personnel infrastructure

Electricity production • Management of reservoirs Electrical energy for uses at• Generation of electrical energy production facility• Transformation and transmission of high-voltage energy• Distribution, transformation, and rectification of low-voltage energy• Maintenance and repair of plant and equipment• Treatment of process air, liquids, and solids

Alumina refining • Bauxite grinding, digestion, and processing of liquors Smelter-grade alumina • Alumina precipitation and calcination transported to primary • Maintenance and repair of plant and equipment aluminum smelter• Treatment of process air, liquids, and solids

Anode production • Recovery of spent anode materials Rodded anodes or briquettes • Anode mix preparation; block or briquette forming and baking for transport to aluminum • Rodding of baked anodes smelter• Maintenance and repair of plant and equipment• Treatment of process air, liquids, and solids

Aluminum smelting • Recovery, preparation, and handling of process materials Hot metal transported to ingot• Manufacture of major process equipment (e.g., cathodes) casting facility• Process control activities (metal, bath, heat)• Maintenance and repair of plant and equipment• Treatment of process air, liquids, and solids

Ingot casting • Pretreatment of hot metal (cleaning and auxiliary heating) Packaged aluminum ingots • Recovery and handling of internal process scrap or hot metal transported to an• Batching, metal treatment, and casting operations aluminum fabricating facility• Homogenizing, sawing, and packaging activities• Maintenance and repair of plant and equipment• Treatment of process air, liquids, and solids

Aluminum extruding • Preheating and cutting or shearing of billet lengths Extruded profiles shipped to• Extruding of shapes, cooling, stretching, and cutting component fabricators• Heat treating, aging, anodizing, or painting• Finishing and packaging activities• Maintenance and repair of plant and equipment• Treatment of process air, liquids, and solids• Excludes material, energy, or releases associated with secondary ingot casting

Rolling hot mill • Preheating and scalping of ingot Aluminum coils or plate shipped• Hot mill rolling of coils to component fabricators• Maintenance and repair of plant and equipment • Treatment of process air, liquids, and solids• Excludes material, energy, or releases associated with secondary ingot casting

Rolling cold mill • Cold mill rolling of coils Aluminum coils or plate shipped• Heat treating, leveling, anodizing, or painting to component fabricators• Finishing and packaging activities• Maintenance and repair of plant and equipment• Treatment of process air, liquids, and solids• Excludes material, energy, or releases associated with secondary ingot casting

Shape casting • Pretreatment of hot metal (cleaning and auxiliary heating) Semifabricated aluminum • Recovery and handling of internal process scrap components• Preparation and forming of cores and molds• Batching, metal treatment, and casting operations• Finishing and packaging activities• Maintenance and repair of plant and equipment• Treatment of process air, liquids, and solids

(continued)


The only significant processing difference inbauxite mining is the need for beneficiation.Beneficiation occurs with ores from forestedareas, while the grassland type typically doesnot require washing. The wastewater fromwashing is normally retained in a settling pondand recycled for continual reuse.

Electricity Production. The aluminumsmelting process is energy-intensive, and assuch, primary aluminum smelters are locatedclose to large sources of reliable energy. TheCanadian aluminum smelting industry is lo-cated in Quebec and British Columbia due tothe abundant source of hydroelectric energy. In-deed, 100% of primary smelting capacity issupplied with hydroelectricity. Prior to enteringthe smelting process, the alternating current isconverted into direct current through a series ofrectifiers.

Alumina Refining. In alumina refining,bauxite is converted to aluminum oxide usingthe Bayer process. Most refineries blend bauxiteto provide feedstock with consistent properties.Bauxite is ground and blended into a recycledliquor. The liquor contains sodium carbonateand sodium hydroxide plus supernatant liquorrecycled from red mud holding ponds. Theslurry is heated and pumped to digesters, whichare heated pressure tanks. In digestion, iron andsilicon impurities form insoluble oxides calledred mud. The red mud settles out, and a richconcentration of sodium aluminate is filteredand seeded to form hydrate alumina crystals inprecipitators. These crystals are then heated in acalcining process. The heat in the calcinersdrives off combined water, leaving alumina.

The major differences in processing are atthe calcinations stage. Two types of kilns areused: rotary and fluid bed. The fluid bed or sta-tionary kiln is newer and significantly moreenergy-efficient.

Significant reductions of energy per unit ofalumina have been achieved over the last 15years with the introduction of higher-pressure di-gesters and fluid flash calciners. Energy require-ments have almost been halved in that period.

Anode Production. There are two types ofaluminum smelting technologies that are distin-guished by the type of anode that is used in thereduction process: Soderberg and prebake.

Soderberg design has a single anode that cov-ers most of the top surface of a reduction cell(pot). Anode paste (briquettes) is fed to the topof the anode, and as the anode is consumed inthe process, the paste feeds downward by grav-ity. Heat from the pot bakes the paste into amonolithic mass before it reaches the elec-trolytic bath interface.

The prebake design has prefired blocks ofsolid carbon suspended from axial busbars.The busbars both hold the anodes in place andcarry the current for electrolysis.

The process for making the aggregate for bri-quettes or prebake blocks is identical. Coke iscalcined, ground, and blended with pitch toform a paste that is subsequently extruded intoblocks or briquettes and allowed to cool. Whilethe briquettes are sent directly to the pots forconsumption, the blocks are sent to a separatebaking furnace.

Baking furnace technology has evolved fromsimple pits that discharged volatiles to the

Table 3.2 (continued)

Unit process Operations of the unit process Output

Manufacturing • Bailed, bundled, and segregated aluminum scrap Scrap transported to aluminumscrap transport on loading dock of auto part dismantler recycling facility

• Transportation by various modes

Consumer scrap • Containerized, bundled, and segregated aluminum Scrap transported to aluminumtransport scrap on loading dock of fabricator recycling facility


Shredded aluminum • Shredded and segregated aluminum scrap on Scrap transported to aluminumtransport loading dock of auto hulk shredder recycling facility


Shredding and • Recovery, preparation, and handling of process materials Shredded aluminum to second-decoating • Shredding and decoating activities ary ingot casting

• Maintenance and repair of plant and equipment • Treatment of process air, liquids, and solids

Secondary ingot • Recovery and handling of internal process scrap Packaged ingots or hot metalscasting • Batching, metal treatment, and casting operations transported to fabrication

• Homogenizing, sawing, and packaging activities facilities• Maintenance and repair of plant and equipment• Treatment of process air, liquids, and solids


atmosphere during the baking cycle to closed-loop-type designs that convert the caloric heatof the volatile into a process fuel that reducesenergy consumption for the process.

Aluminum Smelting. Molten aluminum isproduced from alumina by the Hall-Heroultelectrolytic process that dissolves the aluminain a molten cryolite bath and passes currentthrough this solution, thereby decomposing thealumina into aluminum and oxygen. Aluminumis tapped out of the reduction cell (pot) at dailyintervals, and the oxygen combines with thecarbon to form carbon dioxide and carbonmonoxide.

The pot consists of a steel shell lined with re-fractory insulation and with a hearth of carbon.This is known as the cathode. The cathode isfilled with a cryolite bath and alumina, and ananode is suspended in the bath to complete thecircuit for the pot. When started, a pot will runcontinuously for the life of the cathode, whichmay last in excess of 10 years. The current in apot varies from 60,000 to over 300,000 A at avoltage drop of 4.2 to 5.0 V. Pots produce ap-proximately 7.3 +/– 0.3 kg/day (16.2 +/– 0.6lb/day) of aluminum for each kiloampere at anoperating efficiency of 91 +/– 4%.

All North American aluminum smelters usesome type of air pollution control system to re-duce emissions. The primary system is typicallya scrubber. Some plants use dry scrubbers withalumina as the absorbent, which is subsequentlyfed to the pots and allows for the recovery ofscrubbed materials. Other plants use wet scrub-bers, which recirculate an alkaline solution toabsorb emissions. Unlike dry scrubbers, wetscrubbers absorb carbon dioxide, nitrogenoxide, and sulfur dioxide that are entrained inthe wastewater liquor.

Ingot Casting. Molten metal siphoned fromthe pots is sent to a resident casting complexfound in each smelter. In some cases, due toproximity, molten metal is transported directlyto a shape casting foundry. Molten metal istransferred to a holding furnace, and the compo-sition is adjusted to the specific alloy requestedby a customer. In some instances, depending onthe application and on the bath composition inthe pots, some initial hot metal treatment to re-move impurities may be done.

When the alloying is complete, the melt isfluxed to remove impurities and reduce gas con-tent. The fluxing consists of slowly bubbling acombination of nitrogen and chlorine or carbonmonoxide, argon, and chlorine through the

metal. Fluxing may also be accomplished withan in-line degassing technology, which per-forms the same function in a specialized de-gassing unit.

Fluxing removes entrained gases and inor-ganic particulates by flotation to the metal sur-face. These impurities (typically called dross)are skimmed off. The skimming process alsotakes some aluminum, and as such, drosses arenormally further processed to recover the alu-minum content and to make products used in theabrasives and insulation industries.

Depending on application, metal is thenprocessed through an in-line filter to removeany oxides that may have formed. Metal is thencast into ingots in a variety of methods: openmolds (typically for remelt ingot), direct chillmolds for various fabrication shapes, electro-magnetic molds for some sheet ingots, and con-tinuous casters for aluminum coils.

Aluminum Extruding. The extrusion processtakes cast extrusion billet (round bar stock pro-duced from direct chill molds) and producesextruded shapes. The process begins with an in-line preheat that takes the temperature of thebillet to a predetermined level, depending on thealloy. The billet is then sheared, if not already cutto length, and deposited into a hydraulic press.The press squeezes the semiplastic billet througha heated steel die that forms the shape. The shapeis extruded into lengths defined by the take-offtables and is either water quenched or air cooled.The shape is then clamped and stretched to forma solid, straightened length.

The straightened lengths are cut to finallength multiples and placed in an aging furnaceto achieve a desired temper. Lengths are thenfinished (drilled and shaped) and placed into acoating process. The types of coatings includeanodized, painted, and lacquered finishes.

There are over 250 extrusion plants in NorthAmerica, some with multiple presses. The tech-nology is relatively mature, and variation inprocess efficiency is minor.

Depending on the shape and desired perform-ance characteristics of the extrusion, some pro-files are put through an impact extrudingprocess using considerably higher pressures toform the final parts.

Aluminum hot rolling processes may begin with direct chill cast ingot or with contin-uous cast coils. Direct chill ingots are typically45 to 66 cm (18 to 26 in.) thick and weigh from13 to 27 metric tons (15 to 30 tons) each. Theingots are preheated to approximately 540 °C


(1000º F) and may be scalped on their rollingsurfaces or directly fed to a reversing hot mill.In the reversing mill, the ingot passes back andforth between the rollers, and the thickness isreduced to between 25 to 50 mm (1 to 2 in.),with a corresponding increase in length. Fol-lowing the reversing mill, the coils are fed to acontinuous hot mill where the thickness is fur-ther reduced to less than 6.3 mm (0.25 in.).Continuous cast coils would also be included atthis stage.

Aluminum Cold Rolling. Prior to the coldmill, the coils may be annealed to give the metalworkability; some plants have moved towardself-annealing, which requires no additionalenergy input. Coils are then passed through acoil mill to further reduce the thickness to thecustomer’s requirements and may also gothrough leveling, heat treating, slitting, cuttingto length, and coating processes to meet cus-tomer requirements.

Shape Casting. Aluminum castings are pro-duced in aluminum foundries that are designedto process a range of old and new scrap qualitiesthat contain alloying elements.

Aluminum castings represent over 20% oftotal semifabricated production and are growingat an annualized rate of 4.8%. North Americanproduction is over 1 million tonnes. More than80% of aluminum castings are used in the auto-motive industry.

The majority of castings are produced by sec-ondary smelters, of which there are more than70 production plants in North America. Thetechnology employed is dependent on theshapes produced and uses both permanent andnonpermanent molds. Newly constructed plantsinclude purpose-built facilities that supply theautomotive market.

Manufacturing Scrap Transport. Manufac-turing scrap is generated when semifabricatedaluminum products are manufactured into finalcomponents or parts. The amount varies withapplication and will be defined by the compo-nent and final product manufacturers. This scrapis normally returned to the semifabricated alu-minum producers.

Consumer Scrap. This material is typicallygenerated by automotive dismantlers prior to theautomotive hulks being sent to a hulk shredder.Given the high residual value of aluminum scrapand the premiums associated with segregatedclean scraps, dismantlers invest the additionaltime and effort to remove easily distinguishablealuminum parts, such as body sheet, bumpers,

wheels, brakes, and heat exchangers. Dependingon market conditions, specific efforts to removeall aluminum engines, transmissions, and addi-tional aluminum engine and suspension partswould also be included. In this study, the infor-mation on travel mode, load sizes and distanceswas defined by the secondary smelters.

Shredded Aluminum Scrap. All remainingaluminum automotive scraps are recoveredfrom shredder residues using various media-separation technologies. These include flotationand eddy current methods. As noted, informa-tion on travel mode, load sizes, and distanceswas defined by the secondary smelters.

Shredding and Decoating. In some in-stances, a secondary aluminum smelter willprocess manufacturing and postconsumer scrapthrough a shredding and decoating process.Such processes generate smaller pieces, assistwith the removal of impurities such as ferrousmetals and other unwanted materials, and, incases where the metal has been coated or fin-ished, the organic cover may be removedthrough different types of decoating equipment(i.e., belt decoaters and fluidized bed decoaters).The main purpose for such processes is to im-prove the recovery and efficiency of recyclingand to reduce environmental impacts of remelt-ing coated materials.

Secondary Ingot Casting. The major dis-tinction between secondary casters and primarycasters is in the melting technology employed.Secondary smelters have a variety of purpose-built melting furnaces: top-loaded closedmelters, rotary melters, and sidewell-feedingmelting furnaces. These melting furnaces havedifferent efficiencies and recoveries. Some ofthe melters also use a saltcake to cover the meltand reduce oxidation.

Inventory Analysis

Each unit process in the system is character-ized and documented by a list of input materi-als, energy, and emissions, as shown in Fig. 3.4.

As noted, the primary source of data was ob-tained from questionnaires prepared and issuedto the 13 participating companies. The evalua-tion of the data focused on the informationreceived from the completed responses to thesequestionnaries.

The data categories (material inputs, energyinputs, environmental releases, material out-puts) were predefined for each unit process.


These lists were identified by Weston, Inc. andthe Aluminum Association and were completedbased on the results of the survey.

The data categories used to present the resultsof this study are based on those specified by theUSAMP methodology (Ref 3.1) document(Table 3.3). Additional data categories, at unitprocess level, were collected for Aluminum As-sociation benchmarking purposes.

Calculation Procedures

In addition to the numerous assumptions thatwere made to simplify the data collectionprocess, there were several special calculation

procedures that were used to refine and integratethe information in the inventory. For example,all material and water consumption as well asenvironmental emissions were converted to met-ric units: mass in kilograms (kg); volume inliters (L); gaseous volume in cubic meters (m3);surface area in square meters (m2); and energy inmega joules (MJ). Other calculation methods ad-dress allocation procedures, deliberate omis-sions, and anomalies.

Allocation procedures are used to partitioninputs and/or outputs within a specific productsystem. An allocation procedure is requiredwhen a unit process within a system shares acommon pollution treatment infrastructure orwhere multiple products or co-products are pro-duced in a common unit process.

For allocation of utilities and services com-mon to several processes, allocations weremade to reflect relative use of the service. Forexample, in instances where different unitprocesses (smelter and alumina refinery) werehoused in the same effluent treatment facilities,allocation was determined by the percent oftreatment load generated by each process.

Similar techniques based on mass of the unitprocess outputs were used to allocate some com-mon solid waste streams, such as general refuse.When different unit processes shared a commonenergy source, for example, an electric grid, themetered energy to a specific unit process wasused to allocate the common use items, such asplant and office lighting or space heating.

Deliberate Omissions. Capital assets (plantand equipment), transportation infrastructure(roads, street lights, service stations), andhuman activity (food and transport) associatedwith the production of materials are omittedfrom this study because the relative contributionto the product system for the functional unitbeing considered would be minor.

Unitprocess

Intermediate materialsor final products

Ancillarymaterials

Energy

Water

Air emissions

Water effluent

Solid waste

Raw or intermediate materials

Fig. 3.4 Life-cycle inventory unit process template

Table 3.3 Data categories specified by theUnited States Automotive Materials Partnershipmethodology document

1. Energy(a): 4. Water emissions:Fossil Dissolved solidsNonfossil Suspended solidsProcess (electrical and Heavy metals(c)

nonelectrical) Oils and greasesTransportation Other organicsFeedstock Phosphates,Total ammonia

2. Water consumption: 5. Solid wastes:Ground Total solid wasteSurface Sanitary/municipal

3. Air emissions: 6. Raw materials Dust and particulates consumed:

(includes metals) All significant inputs asCarbon dioxide defined by the decisionCarbon monoxide rulesSulfur oxidesNitrogen oxidesNMHC (includes halogenated

organic compounds)(b)MethaneAcid gases: HCl and HFLead

(a) All energies will include precombustion contributions.(b) All organics other than methane are reported in the aggregate

here.(c) This includes arsenic, cadmium, chromium, cobalt, copper, lead,

mercury, nickel, and zinc.


The manufacturing and postconsumer scrap isassumed to be ready for transport at the ship-ping location. Estimates for selective handlingand processing (e.g., component dismantling)have not been included. These aspects are con-sidered to be minor.

Treatment of Anomalies/Missing Data.Anomalies are extreme data values within a dataset. Anomalies/missing data values are a resultof misinterpreted requests for data input, misre-ported values, improper analysis, or simply notavailable from a reporting location. After an ini-tial review of the data, additional informationwas requested by the facilities to verify certainmisreported/missing values. The missing datafeedback received from the individual plantswas incorporated into the original data set.Where an anomaly was traced to process irregu-larities or accidental release, it was included inthe data set. If an explanation could not befound, the anomaly was removed from the dataset. When all attempts to secure actual inputswere exhausted, a calculated value was usedbased on the average of reported values from aunit process with similar technology.

Data Integration and Presentation

Life-cycle inventory modeling software calcu-lates the LCI profile for each product system.Each of the unit processes and subsystems of

the product system is aggregated to provide asummary cradle-to-gate LCI profile. Figure 3.5shows the integration of unit process data withprecombustion, transportation, fuel use, elec-tricity, and ancillary LCI data to provide theproduct system LCI profile. Primary data forseparate facilities concerning the same productor process were averaged and presented in afashion that ensures confidentiality of individ-ual company data. These aggregated results(weighted-average numbers normalized foreach unit process) are given in the sections“Primary Aluminum Unit Processes” and“Secondary Aluminum Processes” in this arti-cle, along with unit process description, as-sumptions, and data quality assessment results.

Emissions from Fuel. The assumed energyvalues used are characteristically the same asfuels extracted in North America. The environ-mental releases (Table 3.4) from the combus-tion of these fuels include the precombustionenergy for these fuels. Their emissions areadded to the process emissions in the productsystem rollup.

Precombustion energy is defined as the en-ergy associated with the acquisition, processing,and delivery of fossil fuels. Precombustionenergy values were included for all fuels ad-dressed within the scope of this LCI study.

Electrical Energy. The generation and deliv-ery of electricity from electric utilities on a

•Fuel and emission database•Electric database•Ancillary database•Transport database

Normalizeddata

Rollup

Solidwaste

Summary

Fuelanalysis

Transportationanalysis

Electricanalysis

Ancillaryanalysis

Fig. 3.5 Life-cycle inventory engine structure


regional basis involves a mix of energy sources,electricity generation efficiency, transmissionlosses, and environmental releases. Data weredetermined for the ten census divisions in theU.S. and Canadian national grid based oninformation published by the Energy Informa-tion Administration (Ref 3.3) and the NorthAmerican Electric Reliability Council (Ref3.4). A hydropower profile was used to calcu-late LCI releases from facilities using hy-dropower. (A hydroelectric LCI profile wasdeveloped by Roy F. Weston, Inc. consistentwith decision rules and data processing require-ments with those of this study. The profile isbased on confidential 1995 primary data fromfive hydropower plants in North America.) Forsmelter-based activities, a North Americansmelter electrical power mix, reported by theInternational Primary Aluminum Institute(IPAI) (Ref 3.5) was used. Where electricalgrid information from offshore locations wasnot available, the U.S. average profile wasused. Precombustion energy factors (for utili-ties) were included in this study.

Transportation Energy. The four modes oftransportation identified in this study are truck,

rail, barge, and ocean vessel. The environmentalreleases and the energy requirements (includingprecombustion factors) to deliver one metrictonne-kilometer (tonne-km) for each of the trans-port modes are presented (Table 3.5). Estimatesof distance travel and mode of travel are providedby the reporting locations. Transportation has notbeen modeled independently in the life-cyclestages (primary or secondary stages) but ratherhas been incorporated into the product systemrollup.

Feedstock Energy. The energy content ofmaterial resources, is calculated as the grosscalorific value (high heat) of the energy resourcesremoved from the Earth’s energy reserves. It isthe calorific value of the inputs to the system asopposed to the calorific value of the output.

Ancillary Material Analysis

The following is the decision rule process forcompleting this analysis:

1. Identify all potential ancillary material flowsfor a unit process that are greater than 1% bymass of the output of the unit process. Whenthe ancillaries have been identified, a mass

Table 3.4 Typical industrial facility factors with fuel processing burden (per unit input)Heating fuel oil, Diesel, Gasoline, Natural gas, Propane, Coal,

kg L L m3 L kg

Mode Industrial boiler Industrial engine Industrial engine Industrial boiler Industrial boiler Industrial boiler

Water use L 1.20�10–1 1.02�10–1 1.04�10–1 2.32 � 10–7 . . . 3.54 � 10–4

Air emissionsParticulate matter kg/MJ 3.30�10–3 1.58�10–3 2.38�10–3 4.92�10–5 2.41�10–4 1.30�10–3

CO kg/MJ 9.00�10–4 1.27�10–2 9.41�10–1 6.64�10–4 1.33�10–3 6.87�10–3

CO2 kg/MJ 3.86 3.05 2.85 1.97 1.62 2.65SOX kg/MJ 4.78�10–2 3.71�10–3 5.58�10–3 1.86�10–2 2.25�10–3 1.53�10–2

NOX kg/MJ 8.45�10–3 4.75�10–2 2.57�10–2 9.01�10–3 9.16�10–3 7.74�10–3

Nonmethane kg/MJ 1.03�10–3 2.24�10–3 4.72�10–2 2.38�10–5 2.77�10–3 8.08�10–4

hydrocarbonsMethane kg/MJ 0 0 0 0 . . . 0Lead kg/MJ 0 0 0 0 . . . 0HCl kg/MJ 0 0 0 0 . . . 0HF kg/MJ 0 0 0 0 . . . 0

Water effluentsBOD(a) kg/MJ 4.07�10–4 3.44�10–4 7.85�10–4 9.76�10–10 . . . 1.17�10–6

COD(b) kg/MJ 0 0 0 0 . . . 0Heavy metals kg/MJ 0 0 0 0 . . . 0Other organics kg/MJ . . . . . . . . . . . . . . . . . .Suspended solids kg/MJ 1.85�10–3 1.56�10–3 3.56�10–3 4.43�10–9 . . . 5.30�10–6

Dissolved solids kg/MJ 0 0 0 0 . . . 0Oil and grease kg/MJ 2.34�10–4 1.98�10–4 3.60�10–4 5.13�10–10 . . . 6.57�10–7

Phosphates kg/MJ . . . . . . . . . . . . . . . . . .NH4+ kg/MJ 0 0 0 0 . . . 0Solid waste kg/MJ 1.71 1.45 1.28 3.48�10–6 . . . 2.35�10–1

(a) BOD, bacteriological oxygen demand. (b) COD, chemical oxygen demand

References1. Coal Industry Annual 1995, DOE/EIA-0584(95), Energy Information Administration, 19962. Petroleum Supply Annual 1995, Vol 1, DOE/EIA-0340(95)/1, Energy Information Administration, May 19963. Natural Gas Annual 1995, DOE/EIA-0131(95), Energy Information Administration, Nov 19964. Stationary Point and Area Sources, U.S. EPA Compilation of Air Pollutant Emission Factors, AP-42, 5th ed., Vol 1


balance for the subsystems being analyzed isperformed and normalized to the outputfrom the subsystem.

2. Ancillary materials are then classified for in-clusion in the scope of the study based on ananalysis of their contribution to the mass ofthe subsystem, energy of the subsystem, andtheir environmental relevance.

3. All ancillary materials of a ranked ancillarylist that have a cumulative mass contributionof up to 95% of the subsystem are includedin the scope of the study.

4. A further decision rule is used to classifyenergy contribution. All ancillary materialsthat have a cumulative contribution of 99%of the total subsystem energy are included inthe scope of the study, regardless of theirmass ranking.

5. The remaining ancillary materials that arenot selected as a result of the mass andenergy criteria are analyzed for their envi-ronmental relevance. Specifically, an analy-sis is performed to determine if the ancillary

material flow contributes more than 15% toan environmental release data category.

Data Quality

A comprehensive LCI involves the collectionand integration of thousands of pieces of dataregarding the product, process, or activityunder study. To assess the data quality in thisstudy, a number of data quality indicators(DQIs) were used.

Quantitative DQIs include mean and stan-dard deviation values for measures of the vari-ability of the data set values within each unitprocess. These precision measures provide theindustry with an understanding of the variabilityof unit process performance and enable theindustry to define opportunities for improve-ment from a benchmarking perspective.

Another DQI quantitative measure is datacompleteness, defined as a measure of the pri-mary data values used in the analysis divided bythe number of possible data points associatedwith a data category (within the domain). The

Table 3.5 Transportation emission factors (includes precombustion burdens) per tonne-kmHeavy truck Rail Barge Ocean Vessel

Mode Unit (18 tonne) (70 tonne) (12,000 tonne) (35,000 tonne)

Fuel ... Diesel Diesel Diesel Fuel oilFuel consumption MJ/tonne-km 1.04 3.35�10–1 3.42�10–1 9.72�10–2

Water use L/MJ 3.40�10–3 3.40�10–3 3.53�10–3 3.53�10–3

Air emissionsParticulate matter kg/MJ 8.25�10–5 6.25�10–5 6.50�10–5 9.62�10–5

CO kg/MJ 6.04�10–4 3.42�10–4 3.55�10–4 2.06�10–5

CO2 kg/MJ 1.01�10–1 1.01�10–1 1.05�10–1 1.04�10–1

SOX kg/MJ 1.28�10–4 1.28�10–4 1.33�10–4 1.74�10–3

NOX kg/MJ 6.24�10–4 1.71�10–3 1.78�10–3 1.47�10–4

Non-methane hydrocarbons kg/MJ 1.58�10–4 1.08�10–4 1.12�10–4 4.40�10–5

Methane kg/MJ 5.53�10–5 5.27�10–5 5.03�10–5 5.03�10–5

Lead kg/MJ 1.60�10–9 1.60�10–9 1.67�10–9 1.67�10–9

HCl kg/MJ 5.97�10–7 5.97�10–7 6.21�10–7 6.21�10–7

HF kg/MJ 7.47�10–8 7.47�10–8 7.76�10–8 7.76�10–8

Water effluentsBOD(a) kg/MJ 1.15�10–5 1.15�10–5 1.20�10–5 1.20�10–5

COD(b) kg/MJ 9.73�10–5 9.73�10–5 1.01�10–4 1.01�10–4

Heavy metals kg/MJ 7.38�10–7 7.38�10–7 7.67�10–7 7.67�10–7

Other organics kg/MJ . . . . . . . . . . . .Suspended solids kg/MJ 5.22�10–5 5.22�10–5 5.44�10–5 5.44�10–5

Dissolved solids kg/MJ 1.38�10–3 1.38�10–3 1.44�10–3 1.44�10–3

Oil and grease kg/MJ 6.63�10–6 6.63�10–6 6.89�10–6 6.89�10–6

Phosphates kg/MJ . . . . . . . . . . . .NH4+ kg/MJ 1.68�10–6 1.68�10–6 1.75�10–6 1.75�10–6

Solid waste kg/MJ 4.85�10–2 4.85�10–2 5.04�10–2 5.04�10–2

(a) BOD, bacteriological oxygen demand. (b) COD, Chemical oxygen demand

References1. Emissions data derived from EPA Compilation of Air Pollutant Emission Factors—Volume II: Mobile Sources. Specific data used obtained from Table II-3-2 forsteamship burning residual oil and distillate oil. Mode of operation assumed was cruising. Sulfur content assumed was 1%.2. Emissions data derived from EPA Compilation of Air Pollutant Emission Factors—Volume II: Mobile Sources. Specific data used obtained from Table II-3-3 for dieselvessel with an internal combustion engine. Size of engine was assumed to be 3600 hp.3. Data from Transportation Energy Data Book, 11th ed. Table 2.15 for Class I railroads and waterborne commerce4. Emissions data derived from EPA Compilation of Air Pollutant Emission Factors—Volume II: Mobile Sources. Specific data used obtained from Table II-2-15. Average miles per gallon consumed by a heavy-duty truck obtained from “National Transportation Statistics, Annual Report,” Sept 1993. Data for combination trucksin 19916. Emission factors developed from emission estimates in “EPA National Air Pollutant Emission Estimates, 1900–1991” and from energy consumption data for heavy-duty trucks in the Transportation Energy Data Book, 11th ed.


completeness of reporting from the locationsdefined in the domain is highlighted in Table 3.6.

Qualitative DQIs include consistency andrepresentativeness of the data. Consistency is aqualitative understanding of how uniformly thestudy methodology is applied to the variouscomponents of the study. This quality measureis one of the most important to manage in theinventory process. There are a number of stepsthat must be taken to ensure consistency. Themost significant of these is communication. Ina study that involves a number of differentcompanies, which in turn collected data fromover 200 different sites in different countriesand continents, there must be a clear under-standing of what data are being requested, howthey are measured, how they are reported, andhow they are used. For this study, a trainingsession was conducted with company represen-tatives to familiarize them with the key aspectof LCI analysis. The session focused on key re-quirements of the methodology and the datacollection efforts. In addition, detailed instruc-tion sheets were appended to each data inputsheet, describing what was to be includedunder each data category. The reporting loca-tions were encouraged to describe fully whatthey included in their data values and how itwas measured, where possible, highlighting thesampling and analysis protocols that wereemployed. These descriptions were used to ver-ify the consistency in data collection betweenthe companies.

A further mechanism to enhance the consis-tency was the establishment of a hotline forreporting locations to obtain answers for spe-cific questions or to address any difficulties theymay be facing. Verbal and written communica-tions flowed between the consultant team andthe reporting locations, sometimes on severaloccasions to a single location, to verify the datainputs. This verification process also involved acritical review of each data category for eachunit process.

Representativeness is a measure of the degreeto which the data values used in the analysispresent a true and accurate measurement of theaverage processes that the study is examining.The degree of representativeness is normallyjudged by the comparison of values determinedin the study with existing reported values in otheranalyses or published data sources dealing withthe subject matter. Any major variances identifiedshould be examined and explained.

The results of the study and each unit processwithin the subsystems were compared with thatof previous studies produced by the aluminumindustry within both North America and Eu-rope. Indeed, one of the purposes of the alu-minum industry inventory was to quantify theimprovements that have occurred since thoseanalyses were completed. The comparison didnot yield any major variances other than thoserelated to the regional energy and energy-related emissions data. The other factor thatleads to confidence in the representativeness of

Table 3.6 Data completeness by unit processUnit Total Total by unit Percent

Source process reported process complete(a)

Primary Bauxite mining 8 10Electricity generation 7 8Alumina refining 11 13Anode production 25 27Aluminum smelting 29 29Primary ingot casting 22 23Subtotal 102 110 93%

Manufacturing Hot rolling 10 14Aluminum extruding 16 21Cold rolling 12 17Aluminum casting 3 6Subtotal 41 58 71%

Secondary Manufactured scrap transport 9 9Consumer scrap transport 12 12Shredded aluminum 12 12

transportShredding and decoating 9 13Secondary ingot casting and melting 28 39

Subtotal 70 85 82%Total 213 253 84%

(a) The percentage complete is calculated by dividing the number of plants submitting data with the total number of plants for that unit operation within the domain.


the data is the high level of completeness, asmentioned previously.

Primary Aluminum Unit Processes

The life-cycle stage of primary aluminummaterial processing includes the unit processesof bauxite mining, alumina refining, anode pro-duction, smelting, and primary ingot castingFig. 3.1. As previously noted, transportation hasnot been modeled independently but rather hasbeen incorporated into the product systemrollup. For each life-cycle stage, the unitprocess is described, with assumptions made inthe data aggregation. The aggregated data arepresented as “mean,” which is the weighted av-erage of the normalized data for 1000 kg of theoutput from the unit process. Data quality meas-ures, such as precision, completeness, consis-tency, and the qualitative DQIs (average DQIand DQI range), are also shown for the unitprocesses of primary ingot casting.

The energy consumption by unit process in theprimary production chain is shown in Fig. 3.6 per1000 kg of process output. The figure shows theminimum weighted average and maximum valueof energy consumption for each process andclearly illustrates that smelting is the dominantenergy-consuming operation. Alumina refining

uses almost an order of magnitude less energythan smelting per unit of product output.

Figure 3.7 illustrates the CO2 emissions per1000 kg of output for the primary unit processes.With a close relationship between energy use andCO2 releases, it is not surprising that smelting isthe unit operation with the highest CO2 emission.The baking of anodes is also a process with sig-nificant CO2 emission. In this figure, the largevariability in the data is due to major differencesin production technology.

Bauxite Mining

Bauxite is composed largely of hydrated alu-minum oxides, normally including significantimpurities. The major impurities in commerciallymined bauxite include silicon, titanium, ironcompounds, and other materials. The most valu-able bauxite sources tend to be found near theequator. Some of the more important sources arefound in Australia, Guinea, Brazil, and theCaribbean.

Bauxite is typically found at a depth of 0 to 180 m (600 ft) beneath the surface, with an aver-age of 4 m (80 ft). The overburden is removed andstockpiled for use in restoring the site after thebauxite has been removed. The bauxite depositmay be loosened with explosives, depending onits hardness and other local conditions. In some

1.4×105

1.2×105

1×105

8×104

6×104

4×104

2×104

0

MJ

MinWt. avg.Max 1.0×105

1.12×103

2.36×103

1.56×104

Bauxitemining

Aluminarefining

Anodeproduction

Smelting Primary ingotcasting

Fig. 3.6 Primary unit processes energy consumption per 1000 kg of output. Variation due to difference in technology is notaccounted for.


2.00×103

1.80×103

1.60×103

1.40×103

1.20×103

1.00×103

8.00×102

6.00×102

4.00×102

2.00×102

0.00

kg

Bauxitemining

Aluminarefining

Anodeproduction


MinWt. Avg.Max

N/A N/A

9.3×10−1

1.52×103

3.885×102

Fig. 3.7 Primary unit processes CO2 emission per 1000 kg of output. Variation due to difference in technology is not accountedfor. Note: CO2 values presented here represent process-related emissions only.

cases, the bauxite is crushed with dust-controlequipment and/or treated with water to removeimpurities before it is shipped. The washingprocess is called beneficiation. Beneficiatedbauxite typically is dried prior to shipment to therefinery. Ten bauxite mines supplied the majorsources of ore for the thirteen alumina refineriesproviding data for this study. The operationsassociated with this unit process include:

• Extraction of bauxite-rich minerals from thesite

• Beneficiation activities, such as washing,screening, or drying

• Treatment of mining site residues and waste• Site restoration activities, such as grading,

dressing, and planting

The output of this unit process is the bauxitethat is transported to an alumina refinery. The re-source consumption and environmental releasesfor bauxite mining are shown in Fig. 3.8.

Alumina Refining

In alumina refining, bauxite is converted to alu-minum oxide (alumina) using the Bayer process.Most refineries use a mixture of blended bauxiteto provide feedstock with consistent properties.

The mixture is ground and blended with recycledplant liquor. This liquor contains dissolvedsodium carbonate and sodium hydroxide recov-ered from previous extraction cycles plus super-natant liquor recycled from the red mud holdingponds. The aluminum trihydrate is recovered byprecipitation and is filtered, dried, and calcined toform alumina.

Following refining, the alumina is trans-ported to the smelter. There were a total of 13refineries that participated in this study. Four ofthese refineries are in North America and ninewere offshore. This introduces a large variationin both the bauxite and alumina transportationdistance reported by the operation locations.The boundaries for this study begin with theunloading of process materials to their storageareas on site. The operations associated withthis unit process include:

• Bauxite grinding, digestion, and processingof liquors

• Alumina precipitation and calcination• Maintenance and repair of plant and equip-

ment• Treatment of process air, liquids, and solids

The output of this unit process is smelter-grade alumina transported to primary aluminum


smelters in North America. The resource con-sumption and environmental releases for aluminarefining are shown in Fig. 3.9.

Anode Production

There are two types of aluminum smelting tech-nologies that are distinguished by the type ofanode that is used in the reduction process:Soderberg and prebake. Soderberg design has asingle anode that covers most of the top surfaceof a reduction cell (pot). Anode paste (bri-quettes) is fed to the top of the anode, and as theanode is consumed in the process, the pastefeeds downward by gravity. Heat from the potbakes the paste into a monolithic mass before itgets to the electrolytic bath interface. The

prebake design has prefired blocks of solidcarbon suspended from axial busbars. The bus-bars both hold the anodes in place and carry thecurrent for electrolysis.

The process for making the aggregate for bri-quettes or prebake blocks is identical. Coke iscalcined, ground, and blended with pitch toform a paste that is subsequently extruded intoblocks or briquettes and allowed to cool. Whilethe briquettes are sent directly to the pots forconsumption, the blocks are sent to a separatebaking furnace. Baking furnace technology hasevolved from simple pits that dischargedvolatiles to the atmosphere during the bakingcycle to closed-loop-type designs that convertthe caloric heat of the volatile into a process fuelthat reduces energy consumption for the process.

Aluminarefining

Materials

Energy

Air emissions

Products

Wastes

0.6 kg PM

1000 kg alumina

944 kg bauxite residue170 kg other wastes

109 kWh electricity 225 m3 natural gas 93 kg fuel oil

2640 kg bauxite 74 kg caustic soda 46 kg lime1480 L water

Fig. 3.9 Alumina refining—resource consumption and environmental releases (per 1000 kg of unit process output). Note: Difference in total mass of inputs and outputs is attributable to free and bound water in bauxite and residues. PM,

particulate matter

Bauxitemining

Materials

Energy

Air emissions

Products

Wastes

2.4 kg PM

1000 kg bauxite

136 kg residue

0.4 kWh electricity 1 kg fuel oil 4 L diesel

1154 kg bauxite-rich soil 0.6 m2 land use 18 kg lube oil 205 L water

Fig. 3.8 Bauxite mining—resource consumption and environmental releases (per 1000 kg of unit process output). PM, particulate matter


Anodeproduction

Materials

Energy

Air emissions

Products

Wastes

388 kg CO2 3 kg SOx 3 kg PM

1000 kg anode

23 kg residue


820 kg calcined coke 231 kg pitch 85 kg green coke667 L water

Fig. 3.10 Anode production—resource consumption and data quality (per 1000 kg of unit process output). PM, particulate matter

Twenty-seven anode facilities have partici-pated in this study. This unit process (as appliedto this study) begins with the unloading ofprocess materials to their storage areas on site.The operations associated with this unit processinclude:

• Recovery of spent anode materials• Anode mix preparation; block or briquette

forming and baking• Rodding of baked anodes• Maintenance and repair of plant and equip-


The output of this unit process is rodded an-odes or briquettes transported to a primary alu-minum smelter. The resource consumption andenvironmental releases for anode production areshown in Fig. 3.10.

Aluminum Smelting

Molten aluminum is produced from alumina bythe Hall-Heroult electrolytic process (Ref 3.6).This involves two steps: dissolving the aluminain a molten cryolitic bath, and passing electriccurrent through this solution, thereby decom-posing the alumina into aluminum and carbondioxide.

There are two generic types of reductioncells: prebake and Soderberg. The prebaketechnology has two variants referring to howalumina is added, namely, center-worked pre-bake and side-worked prebake. Also, the Soder-berg technology has two variants based on how

the electricity is introduced to the cell, namely,vertical stud Soderberg and horizontal studSoderberg. All plants built since approximately1970 are of the prebake type. The two technolo-gies are differentiated by the type of anodesthey consume.

The prebake design has prefired blocks ofsolid carbon suspended from axial busbars. Thebusbars both hold the anodes in place and carrythe current required for electrolysis. The anodeblocks are consumed until they can no longer besafely held in place without attack on the stubsand are then replaced. The butts of the anodesare returned to the anode plant for recycling intonew anodes by way of adding calcined petro-leum coke and pitch.

The Soderberg design has a single anode thatcovers most of the top surface of the pot. Theanode is baked in place in a large rectangularsteel shell.

All aluminum smelters in North America usesome type of air pollution control system to re-duce emissions. The primary system is typicallya scrubber. Most plants use dry scrubbers, withalumina as the absorbent. The alumina is thenreturned to the pots as feedstock. This allows re-covery of the scrubbed materials. Other plantsuse wet scrubbers, which recirculate an alkalinesolution to absorb emissions. Unlike dry scrub-bers, wet scrubbers absorb the emitted carbondioxide, nitrogen oxide, and sulfur dioxide.Plants with wet scrubbers have a wastewaterstream from the wet scrubber blowdown.

All 29 North American smelters (within thedomain of this study) have participated in this


study. The unit process (as defined for this study)begins with the unloading of process materialsto their storage areas on site. The operationsassociated with this unit process include:

• Recovery, preparation, and handling ofprocess materials

• Manufacture of major process equipment(e.g., cathode shells)

• Process of control activities (metal, bath,heat)

• Maintenance and repair of plant and equip-ment

• Treatment of process air, liquids, and solids

The output of this unit process is hot metaltransported to an ingot casting facility. This unitprocess excludes material, energy, and environ-mental releases associated with anode productionand casting operations. The resource consump-tion and environmental releases for aluminumsmelting are shown in Fig. 3.11. In addition,environmental releases of perfluoro carbon andcarbonyl sulfide generated in aluminum smeltingare summarized in the following sections.

Perfluoro Carbon (PFC) Generation inAluminum Smelting. The PFC values are cal-culated from the 1990 and 1996 values reportedin the IPAI-PFC report (Ref 3.7). Table 3.7summarizes the average PFC emissions fromaluminum smelting based on IPAI estimatesand actual 1995 production rates for primaryaluminum taken from the domain analysis. ThePFC calculation is based on the assumption that92% CF4 (perfluoromethane) and 8% C2F6

(perfluoroethane) are emitted. CO2 equivalentsare calculated based on 6500 (kg) global-warming potential for CF4 and 9200 (kg)global-warming potential for C2F6. Globalwarming potential is defined as the total impactover time of adding a unit of a greenhouse gasto the atmosphere. For the production of every1000 kg of primary aluminum, 0.38 kg of PFCis emitted to the atmosphere.

Carbonyl Sulfide (COS) Generation inAluminum Smelting. The COS values are cal-culated from the reported 1995 values (Ref 3.8).For the production of every 1000 kg of primaryaluminum, 1.12 kg of COS is emitted to theatmosphere.

Primary Ingot Casting

This unit process begins with molten primaryaluminum and ends with sheet ingot suitable forrolling, extruding, or shape casting. There areother types of ingot, called T-ingot, as well aspigs or sows. Aluminum is commonly shippedin these solid forms if it is necessary to transportit a long way from a smelter or reclaimer to aningot casting facility. This process descriptiondoes not include information on any type ofingot besides sheet ingot from primary metaland additives. Ingots that contain manufacturingor recycling aluminum are evaluated in the re-cycling life-cycle stage.

Twenty-two North American primary ingotcasting facilities participated in the study. Theunit process (as defined for this study) begins

Aluminumsmelting

Materials

Energy

Air emissions

Products

Wastes

1520 kg CO2 67 kg CO 17 kg SOx 0.4 kg CF4 0.6 kg HF 0.4 kg PFC 1.2 kg COS

1000 kg aluminum metal

46 kg spent potliner13 kg other wastes

15,400 kWh electricity

1930 kg alumina 19 kg AlF3 446 kg anode carbon 9 kg cathode carbon 827 L water

Fig. 3.11 Aluminum smelting—resource consumption and environmental releases (per 1000 kg of unit process output). PFC, perfluoro carbon; COS, carbonyl sulfide


with the unloading of process materials to theirstorage areas on site. The operations associatedwith this unit process include:

• Pretreatment of hot metal (cleaning and aux-iliary heating)

• Recovery and handling of internal processscrap

• Batching, metal treatment, and casting oper-ations

• Homogenizing, sawing, and packaging andcasting operations



The output of this unit process is packagedaluminum ingots transported to an aluminumfabricating facility. The resource consumptionand environmental releases for primary ingotcasting are shown in Fig. 3.12. Data precisionand completeness for resource consumption andenvironmental releases data for primary ingotcasting are shown in Tables 3.8 and 3.9.

Secondary Aluminum Processing

The life-cycle stage of secondary aluminumprocessing (Fig. 3.2) includes the unit processesof manufacturing scrap transport, consumerscrap transport, shredding and decoating, andsecondary ingot casting. Like the LCI assess-ment of primary processing, transportation hasnot been modeled independently but rather hasbeen incorporated into the product systemrollup. The aggregated data are presented as“mean,” which is the weighted average of thenormalized data for 1000 kg of the output fromthe unit process. Data quality measures, such asprecision, completeness, consistency, and thequalitative DQIs (average DQI and DQI range),are also shown for secondary ingot casting.

Secondary Aluminum Transport

Secondary aluminum transport includes manu-factured scrap, consumer scrap, and shredded

Table 3.7 Perfluoro carbon (PFC) summary for 1996

PFC generation rates (kg of PFC per tonne of aluminum)

CWPB 0.15SWPB 1.00VSS 1.50HSS 0.45

Average North American 0.38

PFC generation (kg of PFC)

CWPB 691,350SWPB 146,000VSS 1,254,000HSS 310,014Total 2,401,364

PFC generation (tonnes of CO2 equivalents)

16,127,561

Aluminum production

Smelter type

CWPB 4609SWPB 146VSS 836HSS 689

North America 6280

CWPB, center-worked prebake; SWPB, side-worked prebake; VSS, verticalstud Soderberg; HSS, horizontal stud Soderberg

Primaryingot

casting

Materials

Energy

Air emissions

Products

Wastes

0.9 kg CO20.02 kg Cl20.05 kg HCl

1000 kg ingot

20 kg skim and dross


1020 kg metal and alloys 3609 L water

Fig. 3.12 Primary ingot casting—resource consumption and environmental releases (per 1000 kg of unit process output)


aluminum. Because data for scrap transportoperations were provided by less than threecompanies, data for these unit operations arenot presented here, but normalized data fromthese operations were given to the individual(participating) companies.

Manufactured Scrap Transport. The unitprocess (as defined for this study) begins withthe containerized, bailed or bundled, and segre-gated aluminum manufacturing scrap on theloading dock of a component or final productmanufacturer. The unit process includes thetransport of manufacturing scrap by varioustransport modes. The output of the unit processis manufacturing scrap transported to an alu-minum recycling facility.

Consumer Scrap Transport. This unit processbegins with disassembled aluminum componentscontainerized, bailed or bundled, and segregatedon the loading dock of an automobile part dis-mantler. The unit process includes the transportof consumer scrap by various transport modes.The output of the unit process is consumer scraptransported to an aluminum recycling facility.

Shredded Aluminum Transport. This unitprocess begins with the containerized, bailed orbundled, and segregated aluminum shreddedscrap on the loading dock of a component orfinal product manufacturer. The unit process in-cludes the transport of shredded scrap by vari-ous transport modes. The output of the unitprocess is shredded scrap transported to an alu-minum recycling facility.

Shredding and Decoating

This unit process begins with the unloading ofsegregated aluminum scrap to designated storage

areas on site. The operations associated with thisunit process include:


• Shredding and decoating activities• Maintenance and repair of plant and equip-


The output of this unit process is shreddedaluminum transported to a secondary ingotcasting facility. As for scrap transportation,data for the shredding and decoating operationwere provided by less than three companiesand are not presented here. Normalized datafrom these operations were provided to partici-pating companies.

Secondary Ingot Casting

The major distinction between secondary cast-ers and primary casters is in the melting tech-nology employed. Secondary smelters have avariety of melting furnaces: top-loaded closedmelters, rotary melters, and sidewell-feedingmelting furnaces. These melting furnaces havedifferent efficiencies and recoveries. Some ofthe melters also use a saltcake to cover the meltand reduce oxidation.

Thirty-nine secondary ingot casting facilitiesparticipated in this study. This unit process be-gins with the unloading of process materials totheir storage areas on site. The operations asso-ciated with this unit process include:


• Batching, metal treatment, and casting oper-ations

Table 3.8 Primary ingot casting—resource consumption and data quality (per 1000 kg of unit processoutput)Inputs Units Mean Standard deviation Completeness,% DQI average(a) DQI range(a)

MaterialLube oil kg 5.35�10–3 2.45�10–2 13 3.8 1–5Filter media kg 1.37 3.46 33 3.6 1–5Primary metal kg 1.00�103 1.19�10 100 4.8 3–5Alloying additives kg 1.74�10 1.25�10 100 4.0 1–5Grain refiners kg 2.27 4.18 63 3.7 1–5Surface water L 2.86�103 4.59�103 67 2.4 1–5Ground water L 7.49�102 2.40�103 17 3.9 1–5

EnergyLight fuel oil kg 1.74�10 3.04�10 100 4.3 3–5Diesel L 1.84�10–1 2.96�10–1 100 3.0 1–5Gasoline L 7.46�10–2 1.19�10–1 100 3.2 1–4Natural gas m3 5.18�10 4.82�10 100 4.3 1–5Propane L 7.98�10–1 1.24 100 4.0 2–5

Purchased/self-generated kWh 2.11�102 1.87�102 100 3.1 1–5electricity

(a) Ranking and definitions of data quality indicators (DQIs).


• Homogenizing, sawing, and packaging andcasting operations



The output of this unit process is packagedaluminum ingots transported to an aluminumfabricating facility. The resource consumptionand environmental releases for secondary ingotcasting are shown in Fig. 3.13. Data precisionand completeness for resource consumption andenvironmental releases data for secondary ingotcasting are shown in Tables 3.10 and 3.11.

Manufacturing Unit Processes

The life-cycle stage of manufacturing includesthe unit processes of shape casting, aluminum

extruding, hot rolling, and cold rolling. Ingotsare provided to this life-cycle stage from bothprimary and secondary aluminum processing.As previously noted for primary and secondaryprocessing, transportation has not been modeledindependently but rather has been incorporatedinto the product system rollup. The aggregateddata are presented as “mean,” which is theweighted average of the normalized data for1000 kg of the output from the unit process.

Shape Casting. Aluminum castings are pro-duced in aluminum foundries that are designedto process a range of old and new scrap qualitiesthat contain alloying elements. They representover 20% of total semifabricated production andare growing at an annualized rate of 4.8%.North American cast aluminum production isover a million metric tons. More than 80% ofaluminum castings are used in the automotive

Table 3.9 Primary ingot casting—environmental releases and data quality (per 1000 kg of unit process output)

DQI DQIReleases Units Mean Standard deviation Completeness, % average(a) range(a)

Air emissionsParticulate matter kg 5.77�10–2 5.41�10–2 71 3.6 1–5CO kg 6.01�10–3 1.31�10–2 21 3.4 1–5CO2 kg 9.34�10–1 4.52 13 3.0 3–3SOX kg 1.10�10–2 2.85�10–2 17 3.5 1–5NOX kg 2.62�10–2 5.53�10–2 25 3.2 1–5Cl2 kg 1.76�10–2 4.96�10–2 58 3.4 1–5HCl kg 5.27�10–2 6.37�10–2 58 3.6 1–5F kg 1.92�10–2 8.56�10–2 8 5.0 5–5HF kg 2.30�10–3 9.57�10–3 13 4.7 4–5Nonmethane hydrocarbons kg 3.82�10–3 3.00�10–2 17 3.0 1–5Organics kg 1.05�10–2 NA 4 3.0 3–3Lead kg 9.37�10–6 NA 4 5.0 5–5Metals kg 1.63�10–3 4.26�10–3 17 3.3 1–5

Water effluentsCOD(b) kg 2.11�10–1 2.93�10–1 58 2.9 1–5BOD(c) kg 4.14�10–2 5.41�10–2 29 3.0 1–5Lead kg 3.18�10–6 2.88�10–5 17 2.7 1–5Iron kg 8.59�10–4 1.19�10–2 17 2.5 2–3NO3 kg 4.71�10–4 4.46�10–3 17 3.0 2–5Mercury kg 1.36�10–8 NA 4 3.0 5–5Metals kg 2.55�10–3 9.57�10–3 29 3.4 1–5NH4+ kg 3.63�10–4 3.65�10–3 17 3.0 2–5Cl– kg 7.80�10–3 NA 4 3.0 3–3CN– kg 1.73�10–6 1.49�10–5 13 3.3 3–4F– kg 2.66�10–3 6.09�10–3 33 3.8 3–5S kg 5.70�10–4 NA 4 3.0 3–3Dissolved organics kg 1.26�10–2 5.76�10–2 17 3.0 1–5Suspended solids kg 6.66�10–2 8.36�10–2 71 3.1 1–5Detergents kg 1.47�10–5 1.91�10–4 8 2.5 2–3Dissolved chlorine kg 8.78�10–5 1.14�10–3 13 2.7 2–4Phenol kg 1.20�10–6 1.47�10–5 13 2.5 1–4Dissolved solids kg 1.80�10–1 4.50�10–1 42 2.0 1–4Oil and grease kg 2.33�10–2 7.42�10–2 75 3.1 1–5

Solid wasteProcess Related kg 1.89�10 1.22�10 100 3.9 1–5Environmental abatement kg 7.16�10–1 2.62 25 3.5 1–5Municipal kg 1.61 2.01 88 2.9 1–5

(a) See preceding table for ranking and definitions of data quality indicators (DQIs). (b) COD, chemical oxygen demand. (c) BOD, bacteriological oxygen demand


industry. The majority of castings are producedby secondary smelters, of which there are morethan 70 production plants in North America.The technology employed is dependent on theshapes produced and employs both permanentand nonpermanent molds.

This unit process (as defined for this study)begins with the unloading of process materialsto their storage areas on site. The operations as-sociated with this unit process include:

• Pretreatment of hot metal (cleaning and aux-iliary heating)


• Preparation and forming of cores and molds• Batching, metal treatment, and casting oper-

ations

• Finishing and packaging activities• Maintenance and repair of plant and equip-


The output of this unit process is semifabri-cated aluminum components that are transportedto a component or final product manufacturer.The resource consumption and environmentalreleases for aluminum shape casting are shownin Fig. 3.14.

Extruding. The extrusion process takes castextrusion billet (round bar stock produced fromdirect chill molds) and produces extrudedshapes. The process begins with an in-line pre-heat that takes the temperature of the billet to apredetermined level, depending on the alloy.The billet is then sheared, if not already cut to

Secondaryingot

casting

Materials

Energy

Air emissions

Products

Wastes

66 kg CO223 kg CO

1000 kg ingot

80 kg residue

115 kWh electricity126 m3 natural gas

1043 kg metal and alloys 8 kg water treatment chemicals 9 kg treatment salts 960 L water

Fig. 3.13 Secondary ingot casting—resource consumption and environmental releases (per 1000 kg of unit process output)

Table 3.10 Secondary ingot casting—resource consumption and data quality indicator (DQI)(per 1000 kg of unit process output)Inputs Units Mean Standard deviation Completeness, % DQI average DQI range

MaterialLube oil kg 5.35 � 10–3 2.45�10–2 13 3.8 1–5Filter media kg 1.37 3.46 33 3.6 1–5Primary metal kg 1.00�103 1.19�10 100 4.8 3–5Alloying additives kg 1.74�10 1.25�10 100 4.0 1–5Grain refiners kg 2.27 4.18 63 3.7 1–5Surface water L 2.86�103 4.59�103 67 2.4 1–5Ground water L 7.49�102 2.40�103 17 3.9 1–5

EnergyLight fuel oil kg 1.74�10 3.04�10 100 4.3 3–5Diesel L 1.84�10–1 2.96�10–1 100 3.0 1–5Gasoline L 7.46�10–2 1.19�10–1 100 3.2 1–4Natural gas m3 5.18�10 4.82�10 100 4.3 1–5Propane L 7.98�10–1 1.24 100 4.0 2–5

Purchased/self-generated electricity kWh 2.11�102 1.87�102 100 3.1 1–5


length, and deposited into a hydraulic press.The press squeezes the semiplastic billetthrough a heated steel die that forms the shape.The shape is extruded into lengths defined bythe take-off tables and is either water quenchedor air cooled. The shape is then clamped andstretched to form a solid, straightened length.The straightened lengths are cut to final lengthmultiples and placed in an aging furnace toachieve a desired temper. Lengths are then fin-ished (drilled and shaped) and placed into acoating process. The types of coatings includeanodized, painted, and lacquered finishes.

There are over 250 extrusion plants in NorthAmerica, some with multiple presses. The tech-nology is relatively mature, and variation inprocess efficiency is minor. Depending on theshape and desired performance characteristicsof the extrusion, some profiles are put throughan impact extruding process that uses consider-ably higher pressures to form the final parts.

The operations associated with this unitprocess include:

• Preheating and cutting or shearing of billetlengths

• Extruding of shapes, cooling, stretching, andcutting

• Heat treating, aging, anodizing, or painting• Finishing and packaging activities• Maintenance and repair of plant and equip-

ment and treatment of process air, liquids,and solids

The output of this unit process is semifabri-cated or finished extruded profiles transportedto a component or final product manufacturer.The unit process excludes materials, energy, andenvironmental releases associated with second-ary ingot casting operations, if any. The re-source consumption and environmental releasesfor aluminum extruding are shown in Fig. 3.15.

Table 3.11 Secondary ingot casting—environmental release and data quality indicator (DQI)(per 1000 kg of unit process output)Releases Units Mean Standard deviation Completeness, % DQI average DQI range

Air emissionsParticulate matter kg 2.38�10–1 3.43�10–1 81 2.9 1–5CO kg 2.35�10 4.91�10 70 3.0 1–5CO2 kg 6.59�10 6.08�102 15 2.0 1–4SOx kg 1.02�10–2 3.65�10–2 56 3.0 1–5NOx kg 2.02�10–1 7.58�10–1 67 2.9 1–5Cl2 kg 6.34�10–2 8.55�10–2 41 3.5 1–5HCl kg 1.71�10–1 1.51�10–1 44 3.3 1–5HF kg 1.07�10–2 1.44�10–2 15 4.3 4–5Nonmethane hydrocarbons kg 8.91�10–2 1.80�10–1 41 2.8 1–5Organics kg 1.71�10–2 5.21�10–2 19 3.8 2–5Lead kg 1.99�10–4 2.43�10–4 33 3.4 3–5Metals kg 1.74�10–4 4.14�10–4 11 3.5 1–5

Water effluentsCOD(a) kg 2.92�10–2 5.80�10–2 100 3.7 2–5BOD(b) kg 3.76�10–2 6.74�10–2 100 3.6 2–5Lead kg 8.06�10–7 1.07�10–5 22 3.3 1–5Iron kg 9.31�10–6 2.58�10–4 15 3.5 2–5NO3 kg 2.12�10–4 3.19�10–3 11 2.7 2–3Metals kg 2.12�10–4 3.37�10–3 33 2.8 1–5NH4+ kg 1.42�10–5 4.95�10–4 15 3.2 2–5F– kg 1.06�10–4 1.48�10–3 22 2.8 1–5Dissolved organics kg 3.63�10–4 5.06�10–3 15 3.3 1–5Suspended solids kg 2.25�10–2 1.14�10–1 100 3.5 1–5Detergents kg 3.65�10–6 7.23�10–5 11 2.7 2–3Hydrocarbons kg 4.65�10–4 1.08�10–3 11 1.3 1–2Dissolved solids kg 1.29�10–3 NA 4 4.5 4–4Phosphates kg 9.61�10–6 NA 4 2.0 3–3Other nitrogen kg 1.86�10–4 2.38�10–3 19 3.6 2–5Oil and grease kg 2.56�10–2 1.93�10–2 52 3.6 1–5

Solid wasteProcess related kg 7.87�10 7.54�10 85 2.9 1–5Environmental abatement kg 6.54�10–1 1.82 19 3.7 1–5Municipal kg 4.81 2.23�10 67 2.6 1–5

(a) COD, chemical oxygen demand. (b) BOD, bacteriological oxygen demand


Hot and Cold Mill Rolling. The ingots aretypically 45 to 66 cm (18 to 26 in.) thick andweigh 13 to 27 � 103 kg (15 to 30 tons). The in-gots are preheated to approximately 540 °C(1000 °F) and fed through a hot reversing mill.In the reversing mill, the coil passes back andforth between rollers, and the thickness is re-duced from the initial thickness to 25 to 50 mm(1 to 2 in.), with a corresponding increase inlength. Following the reversing mills, the slabsare fed to a continuous hot mill, where thethickness is further reduced to less than 6 mm(1/4 in.) in thickness. The metal, called reroll orhot coil, is rolled into a coil and is ready to betransferred to the cold mill.

Prior to the cold mill, the coils may be an-nealed to give the metal the workability for

downstream working. The coils are then passedthrough multiple sets of continuous rollers to re-duce the gage. The coils are cut to the width andlength as required by the customer. The coils arepackaged to prevent damage to the metal inshipping.

The resource consumption and environmentalreleases for aluminum rolling are shown in Fig.3.16 (hot rolling) and 3.17 (cold rolling).

Results by Product System

Figures 3.18 and 3.19 summarize the cradle-to-gate profiles of primary and secondary prod-uct systems. The product systems include the rawmaterial acquisition, processing, and product

Aluminumextruding

Materials

Energy

Products

Wastes

1000 kg extruded aluminum 440 kg manufactured scrap

5 kg residue

590 kWh electricity104 m3 natural gas 11 kg anthracite

1445 kg metal 1180 L water

Fig. 3.15 Aluminum extruding—resource consumption and environmental releases (per 1000 kg of unit process output)

Aluminumshape casting

Materials

Energy

Air emissions

Products

Wastes

8 kg PM3 kg NOx

1000 kg cast aluminum 990 kg manufactured scrap

212 kg residue

4 kWh electricity240 m3 natural gas

2202 kg metal 8 kg treatment gases 335 kg silica5450 L water

Fig. 3.14 Shape casting—resource consumption and environmental releases (per 1000 kg of unit process output). PM, particulatematter


Hotrolling

Materials

Energy

Air emissions

Products

Wastes

16 kg NOx23 kg CO2

1000 kg hot roll 208 kg manufactured scrap

13 kg residue

265 kWh electricity 33 m3 natural gas

1221 kg metal 3 kg rolling oil1352 L water

Fig. 3.16 Hot rolling—resource consumption and environmental releases (per 1000 kg of unit process output)

Coldrolling

Materials

Energy

Air emissions

Products

Wastes

0.4 kg CO2

1000 kg rolled aluminum 188 kg manufactured scrap

8 kg residue

350 kWh electricity 25 m3 natural gas

1196 kg hot roll 4 kg rolling oil 3 kg pallets 135 L water

Fig. 3.17 Cold rolling—resource consumption and environmental releases (per 1000 kg of unit process output)

manufacture life-cycle stages. Material con-sumption includes the consumption of miner-als, water, and petrochemical feedstocks. En-ergy consumption includes the use of energy atthe plant (purchased/self-generated); energyused in transportation of raw materials, prod-ucts and solid waste; and precombustion en-ergy (energy expended to extract, refine, anddeliver a fuel).

The outputs of the product systems are prod-ucts, air emissions, water effluents, and solidwaste. The air emissions quantified include dustand particulates (includes metals), CO2, CO,SOX, NOX, nonmethane hydrocarbons (includeshalogenated organic compounds), CH4, acid

gases (HCl and HF), and lead. The water efflu-ents quantified include dissolved solids, sus-pended solids, heavy metals, oils and greases,other organics, and phosphates and ammonia.The solid waste produced by the product systemwas aggregated based on the waste managementmethod used. Waste management options in-clude landfilling, recycling, incineration, andbeneficial use.

Automotive Aluminum Recycling. Figure3.20 portrays the mass flow of extruded, rolled,and shape cast aluminum products through thevarious steps of vehicle assembly, use, disman-tling, and recycling. Secondary aluminum re-covered during recycling is returned for the


Primaryingot

Ancillarymaterials

Rawmaterials

Energy

Air emissions

Products

Wastes

11,564 kg CO2 74 kg CO 85 kg SOx 40 kg NOx 43 kg PM

1000 kg ingot

4590 kg residues

5090 kg bauxite 3 m2 land16,577 L water

143 kg caustic soda366 kg calcined coke103 kg pitch 93 kg lube oil 88 kg lime

186,262 MJ total energy 69,406 MJ nonfossil116,856 MJ fossil

(a)

Primaryrolled aluminum

Ancillarymaterials

Rawmaterials

Energy

Air emissions

Products

Wastes


1000 kg rolled aluminum

4810 kg residues




(b)

Fig. 3.18 Resources and environmental releases (per 1000 kg output) for primary aluminum products. (a) Primary ingot. (b) Primary rolled product. (c) Primary extrusions. (d) Primary shape castings. PM particulate matter

manufacture of new components. This chartwas developed through the combined efforts ofrepresentatives of the Auto and Light TruckCommittee, the Recycling Committee, and theLife-Cycle Task Group of the Aluminum Asso-ciation. Figure 3.20(a) shows the situation forthe 1995 calendar year, and Fig. 3.20(b) por-trays the future steady-state situation, when afleet of more aluminum-intensive vehicles isrecycled. Currently, because the amount of alu-minum, especially wrought aluminum, used in

each vehicle is growing, there is more alu-minum being consumed in each new vehiclethan is now being recovered from used vehi-cles. However, as the more aluminum-intensivevehicles reach the end of their useful life of ap-proximately 10 to 12 years, larger quantities ofaluminum will be recycled, and a highersteady-state situation should be reached forwrought products. This assertion is based onthe economical and technical performance ofthe recycling infrastructure. This steady-state


Primaryextruded aluminum

Ancillarymaterials

Rawmaterials

Energy

Air emissions

Products

Wastes


1000 kg extruded aluminum

4780 kg residues




(c)

Primaryshape cast aluminum

Ancillarymaterials

Rawmaterials

Energy

Air emissions

Products

Wastes


1000 kg shape cast aluminum

5114 kg residues




(d)

Fig. 3.18 (continued) (c) Primary extrusions. (d) Primary shape castings. PM, particulate matter

situation will probably be attained during thesecond decade of the next century. Some esti-mates for the use of aluminum in vehicles atthat time range from 180 to 315 kg (400 to 700lb) per vehicle.

The results released to USAMP will con-tribute in achieving the objectives of the genericautomobile LCI effort. The energy and environ-mental profiles modeled and submitted to theautomakers are based on the present industrypractice, where only 11% recycled aluminum is

used in wrought products and 85% is used incast aluminum products. Currently, the distribu-tion of aluminum use by product form in auto-motive applications is made up of 73.8% castaluminum, 22.8% extruded aluminum, and3.4% rolled aluminum. Accordingly, using thesepercentages, a composite energy and environ-mental profile for the generic USAMP vehiclecan be derived for automotive aluminum. Theresults for the three product systems are pre-sented in Fig. 3.21.


Interpretation of LCI Results

The aluminum LCI results described hereare based on 1995* actual average data col-lected from 213 facilities (13 companies and15 primary and secondary unit operations). In-formation on energy used for the primary ingot(Fig. 3.3) product system revealed that 84% of

the total energy along the life cycle was fromprocess energy (nonelectric and electric). Elec-tric process energy accounted for 57% of thetotal energy. Production of secondary ingotrequires 92% of its energy from process (non-electric and electric), with electric processenergy accounting for 21% of the processenergy.

With regard to recycling, when the energy as-sociated with transportation is discounted, therelative energy consumption between secondary

*There was one location where actual 1992 to 1993 datawere used in this analysis.

Secondaryingot

Ancillarymaterials

Rawmaterials

Energy

Air emissions

Products

Wastes

617 kg CO2 1 kg CO 5 kg SOx 3 kg NOx 1 kg PM

1000 kg ingot

388 kg residues

1133 kg metal (scrap) 963 L water

21 kg alloys 9 kg treatment salts 8 kg water treatment chemicals

11,690 MJ total energy 108 MJ nonfossil11,583 MJ fossil

(a)

Secondaryrolled aluminum

Ancillarymaterials

Rawmaterials

Energy

Air emissions

Products

Wastes



729 kg residues


30 kg alloys13 kg treatment salts12 kg water treatment chemicals 7 kg rolling oil


(b)

Fig. 3.19 Resource consumption and environmental releases (per 1000 kg output) for products from secondary aluminum. (a) Ingot. (b) Rolled aluminum. (c) Extruded aluminum. (d) Shape castings. PM, particulate matter


Secondaryextruded aluminum

Ancillarymaterials

Rawmaterials

Energy

Air emissions

Products

Wastes



686 kg residues


30 kg alloys 3 kg cardboard12 kg water treatment chemicals


(c)

Secondaryshape cast aluminum

Ancillarymaterials

Rawmaterials

Energy

Air emissions

Products

Wastes



1190 kg residues


46 kg alloys 19 kg treatment salts 18 kg water treatment chemicals 334 kg silica


(d)

Fig. 3.19 (continued) (c) Extruded aluminum. (d) Shape castings. PM, particulate matter

(recycled) and primary ingot indicates that only6.3% of the energy is required to produce sec-ondary ingot as compared to primary ingot (asshown in Fig. 3.22). Accordingly, the energy ef-ficiency and environmental performance of theindustry as a whole will continue to improve asthe extent of recycling increases.

Areas for Process Improvement. TheUSAMP partners in the LCI study consider thedatabase suitable for benchmarking and processimprovement, but it should not be used for the

purposes of materials selection. For materialsselection in actual application, other factors,such as cost, materials availability, and perform-ance, must also be considered.

This LCI study has provided opportunities forbenchmarking and subsequent process improve-ment. For example, each company contributingdata to the study has received its own data, to-gether with the industry average data, in thesame format. Thus, each company can directlycompare its performance in a given unit process


to the benchmark performer and to the industryaverage. It is believed that this study will helpidentify several areas where reductions in rawmaterials and energy consumption, as well asenvironmental loadings, can be achieved in alu-minum production processes.

To facilitate the benchmarking process, theuse of an improvement potential index has beenexplored. The improvement potential has beendefined as:

Improvement potential = Standard deviation� Mean � Contribution

For each data category, the standard deviationof the input or output was divided by the meanof the input or output and then multiplied by thecontribution that the input or output value hadon the data category. Conceptually, the im-provement potential could be used as a meansof prioritizing or selecting the process improve-ment efforts or investments.

It is critical to recognize, however, that thisLCI database is enormous in its scope, and ineach unit process, it arbitrarily compares datafrom a wide range of operations in differentlocations, using raw materials of varying quality,

that is, bauxite type and alumina quality. Mostimportantly, the report often compares oldertechnologies with newer variations of the sametechnology or even with radically different tech-nologies that employ different operating princi-ples, that is, fluid bed calcination versus rotarycalcination of alumina trihydrate.

Given that the input data were provided underconditions of anonymity, no attempt is made tosuggest specific areas of process improvementin view of the enormous number of raw mate-rial, process, and technology variables. It is con-sidered that specific process improvements willemerge when knowledgeable scientists and en-gineers sift through the detailed database and,using it as a background, can apply their ownunique knowledge of their companies’ specificprocesses to define the best opportunities forprocess improvement.

However, with the aforementioned caveats inmind, it was considered instructive to evaluatethe improvement potential for the category ofenergy use. Figure 3.23 shows the improvementpotential, expressed as a percentage, for energyuse as a function of the unit process. The rela-tively high value for alumina refining (39.1%) isprobably due to a combination of factors. For

(a)

Primary aluminum

Extruded Al

Rolled Al

Shape cast

25

25

23 15%

primary

primary

288

89%27

Componentmfg.

Vehicleassembly

56

158

Vehicleoperation

Componentdismantling

Hulkshredder

7

20

166153

21

Secondary aluminum

Extruded Al

Rolled Al

Cast Al

13

(b)

Primary aluminum

Extruded Al

Rolled Al

Shape cast

14

15

30 19%

primary

primary

281

52%48

Componentmfg.

Vehicleassembly

56

158

Vehicleoperation

Componentdismantling

Hulkshredder

29

20

144153

21

Secondary aluminum

Extruded Al

Rolled Al

Cast Al

Fig. 3.20 Automotive aluminum recycling. (a) Current practice. (b) Steady-state future practice


Automotiverolled aluminum

Ancillarymaterials

Rawmaterials

Energy

Air emissions

Products

Wastes



4480 kg residues


329 kg calcined coke128 kg caustic soda 93 kg pitch 83 kg lube oil 79 kg lime


(a)

Automotiveextruded aluminum

Ancillarymaterials

Rawmaterials

Energy

Air emissions

Products

Wastes



4441 kg residues


328 kg calcined coke128 kg caustic soda 93 kg pitch 83 kg lube oil 79 kg lime


(b)

Fig. 3.21 Resource consumption and environmental releases (per 1000 kg output) for aluminum automotive products. (a) Rolled.(b) Extruded. (c) Shape cast. (d) Composite. PM, particulate matter

example, in alumina refining, the quality of thebauxite ore dictates whether low- or high-tem-perature digestion conditions must be used todissolve the alumina. Also, the calcination ofthe resulting aluminum trihydrate intermediateproduct can be conducted using either the tradi-tional rotary kiln or the more recent and energy-efficient fluid bed process.

The relatively low value for smelting (2.1%)is believed to reflect the fact that smelting hasalready been extensively optimized through re-search and development and plant practice, be-

cause this process step has long been recog-nized as energy-intensive (Fig. 3.6, 3.7). Insmelting, energy improvements to some newlevel of efficiency will require some fundamen-tal development in advanced electrode technol-ogy, for example, wettable cathode and/or inertanode technology and related cell design im-provements. This is consistent with the fact thatthe highest long-term research and developmentpriorities in the technology roadmap of the in-dustry relate to the development of advancedelectrode technology.


Figure 3.24 shows the improvement pon-tential data for CO2 emissions by unit process.Neither bauxite mining nor alumina refininghas any process-related CO2 emissions; here,the CO2 emissions are derived from the fuelsused, for example, during calcination. How-ever, during anode production and smelting,there appears to be significant potential for im-provement. Again, raw material variations andtechnology differences may account for a large

portion of the variation. For example, duringanode baking, the newest furnace technologycaptures the hydrocarbon offgases from thebaking process and recycles them as a fuel sup-plement. In the case of smelting, technologydifferences probably account for a significantportion of the 10.8% improvement potentialvalue. Note that the smelter data incorporatedata from both vertical stud and horizontal studSoderberg processes as well as side-worked

Automotiveshape cast aluminum

Ancillarymaterials

Rawmaterials

Energy

Air emissions

Products

Wastes



1398 kg residues

764 kg bauxite 0.4 m2 land 9912 L water

334 kg silica 56 kg calcined coke 45 kg alloys 25 kg caustic soda

63,603 MJ total energy12,543 MJ nonfossil51,059 MJ fossil

(c)

Automotivealuminum composite

Ancillarymaterials

Rawmaterials

Energy

Air emissions

Products

Wastes


1000 kg automotive aluminum

2196 kg residues


127 kg calcined coke 52 kg caustic soda 36 kg pitch 32 kg lube oil 31 kg lime

96,374 MJ total energy69,457 MJ nonfossil26,917 MJ fossil

(d)

Fig. 3.21 (continued) Resource consumption and environmental releases (per 1000 kg output) for aluminum automotive prod-ucts. (a) Rolled. (b) Extruded. (c) Shape cast. (d) Composite. PM, particulate matter


200,000

180,000

160,000

140,000

120,000

100,000

80,000

60,000

40,000

20,000

−

Ene

rgy,

MJ

Primary ingot Secondary ingot

182,439

94% reductionin energy use

10,988

Fig. 3.22 Energy consumption comparison of recycled (secondary) ingot to primary ingot. Excludes transportation energy

50.00

45.00

40.00

35.00

30.00

25.00

20.00

15.00

10.00

5.00

0.00

Ene

rgy,

%

Bauxitemining

Aluminarefining

Anodeproduction


8.59%

39.11%

3.37%2.10%

6.03%

Fig. 3.23 Improvement potential for energy consumption in primary aluminum processing

and center-worked prebake cells. Emissionsof CO2 can also be influenced by materialsvariability, for example, coke quality, and bycell operating practices.

In summary, it is anticipated that energy andenvironmental improvements will continue to

be made in aluminum production by dedicatedplant and process engineers who exploit thisLCI database as a tool, together with their ownspecific insights of different productionprocesses, to achieve overall performance im-provements.


REFERENCES

3.1. “USCAR/USAMP Generic AutomobileLife Cycle Inventory Methodology Docu-ment”, Roy F. Weston, Inc., 1997

3.2. “Environmental Management—Life CycleAssessment—Goal and Scope Definitionand Inventory Analysis”, Draft Interna-tional Standard ISO/DIS 14041, Interna-tional Standards Organization, 1997

3.3. Electric Power Annual 1995, Vol 1,DOE/EIA-0348(95)/1, Energy InformationAdministration, July 1996

3.4. “Electricity Supply and Demand 1995–2004, Summary of Electric Utility Supplyand Demand Projections,” North AmericanElectric Reliability Council, June 1995

3.5. “Electric Power Used in Primary Alu-minum Production,” International PrimaryAluminum Institute (IPAI) Form ES 002

3.6. K. Grjotheim, et al., Aluminum Electroly-sis, Fundamentals of the Hall-HeroultProcess, Chap. 4, 5, Aluminum-Verlag,Dusseldorf, 1982

3.7. “PFC Generation in 1996,” InternationalPrimary Aluminum Institute

3.8. COS Generation in 1995, Summary ofMajor Air Emissions of U.S. AluminumSmelters (1995 reporting year), Environ-mental Defense Fund; http:// www.score-card.org (accessed June 2007)

12.0

10.0

8.0

6.0

4.0

2.0

0.0

CO

2, %

N/A N/A

Bauxitemining

Aluminarefining

Anodeproduction


9.36%

10.83%

0.27%

Fig. 3.24 Improvement potential for CO2 emissions in primary aluminum processing

THE INTERNATIONAL ALUMINUM IN-STITUTE (IAI) has been involved in the world-wide collection of aluminum data to be used inlife-cycle assessments in order to “develop ascomplete an understanding as possible of thepositive contributions that aluminum makes tothe environmental and economic well-being ofthe world’s population; of any negative economicor environmental impacts that its production maycause; and of the balance between these posi-tives and negatives during the entire life cycle ofthe material” (Ref 4.1). This chapter containsresults from a March 2003 report (Ref 4.2) withthe purpose of collecting all significant life-cycle inventory data for primary aluminum (i.e.,raw materials and energy use, air and wateremissions, solid waste generated), with world-wide coverage (except Russia and China). Cu-mulative inputs and outputs of environmentalsignificance (air emissions, waste generation,resource consumption) are defined for producingprimary aluminum ingot from bauxite ore. Itdoes not include the inputs and outputs associ-ated with the further processes related to theproduction of final products from ingots and therecycling of the end-of-life product to obtainrecycled aluminum ingots.

The intended purpose of this life-cycle inven-tory is to accurately characterize resource con-sumption and significant environmental aspectsassociated with the worldwide production of pri-mary aluminum as a globally traded commodity.The primary aluminum production covered by

this life-cycle inventory includes the followingunit processes:

• Bauxite mining• Alumina production • Anode production: production of prebaked

anodes, production of Soderberg paste • Electrolysis• Ingot casting

The collected data will serve as a crediblebasis for subsequent life-cycle assessments ofaluminum products. In addition, the last sectionprovides guidance on key features about how totreat aluminum in life-cycle assessments whererecycled aluminum is used for aluminum prod-ucts. No specific additional unit processes, inparticular about energy production, transport,petroleum coke and pitch production, causticsoda production, and so on, have been added tothe process in order to avoid nonelementaryflows. Practitioners of life-cycle assessmentwho will use the data of this report may includesuch additional unit processes from their owndatabases (see also the section “Data Interpreta-tion Items” in this chapter).

Data Coverage, Reporting, and Interpretation

Data from worldwide aluminum producerswere requested for the calendar year 2000. Datafor the life-cycle survey were obtained from:

• Eighty-two worldwide aluminum electrol-ysis plants producing 14.7 million metric tons

CHAPTER 4

Life-Cycle Assessment of Aluminum:Inventory Data for the Worldwide Primary Aluminum Industry*

*Adapted with permission of the International AluminumInstitute from Ref 4.2




(tonnes, t) of primary aluminum, representingapproximately 60% of the worldwide alu-minum smelting operations. Base: primaryaluminum of 24,464,400 tonnes (t) from theWorld Bureau of Metal Statistics (WBMS).

• Twenty-three worldwide alumina facilitiesproducing 30.8 million metric tons of alu-mina, representing approximately 59% ofthe worldwide alumina operations. Base:world alumina production of 52,419,000 t,as 48,119,000 t from IAI plus 4,300,000 testimated for China.

• Seventy-two worldwide aluminum casthouses producing 14.0 million tonnes (metrictons) of primary aluminum ingot, represent-ing approximately 57% of the worldwidealuminum ingot casting operations. Totalworld aluminum production in the year 2000from WBMS was 32,623,800 t, of which24,464,400 t primary aluminum was takenas the cast house production base (plus8,159,400 t secondary aluminum not rele-vant here).

Overall, primary aluminum energy returnsrepresented approximately 69% of the worldprimary aluminum production; alumina energyreturns represented approximately 70% of theworld alumina production; and perfluorocarbonreturns represented approximately 66% of theworld primary aluminum production. The datawere reasonably evenly distributed on a world-wide basis, with the nonavailability of data fromChina and Russia being mainly responsible forthe comparatively poor coverage of Asia andEurope. For example, the survey’s coverage ofelectrolysis plants in terms of reported primaryaluminum production as a percentage of total pri-mary aluminum production was approximately82% in Africa, 75% in North America, 58% inLatin America, 34% in Asia (data for China notavailable), 52% in Europe (data for Russia andother Commonwealth of Independent Statescountries not available), and 100% in Oceania.

Collected data were combined together intoinventory summaries of input and output(Tables 4.1 and 4.2) resources to produce 1000kg of aluminum ingot (see also the section“Unit Processes and Results by Process” in thischapter). Selection of variables (Table 4.3) forthis inventory was based on their environmentalrelevance, either specific to the primary alu-minum production or as generally acknowledgedenvironmental issues. It should also be notedthat only direct energy-consumption figures

were documented for this inventory and thatCO2 emission data were not included. Compre-hensive energy data and CO2 emission data,including those associated with the generation ofelectricity, the production of fuel (precombus-tion), and the combustion of fuel, are covered inRef 4.3 and summarized in Table 4.4.

Data Quality

Data reporting includes quantitative andqualitative data quality indicators calculated foreach data item. All values presented representproduction-weighted mean values. Complete-ness is a quality indicator that covers the possi-bility for when not all respondents provided ananswer. It is a ratio of the number of responsesfor the data item versus the total number ofresponses to the survey. Data statistics assumea normal distribution of results.

Data Consistency. During data review, a sig-nificant scatter in individual data items wasnoticed. Beyond obviously unrealistic outliers,high- and low-value outlier responses appearedlikely to wrongly influence the final weightedaverage values. To address this issue, all individ-ual answers beyond two standard deviationsfrom the average value have been considered asoutliers. Every individual outlier respondent hasbeen queried accordingly with a request to checkthe response item for correction or confirmation.All outliers were adjusted according to answersreceived. If no answer had been received to theoutlier query (approximately 60% of outliers),the individual outlier item was removed from thesurvey. The effect of this correction from out-liers is summarized in Table 4.5. The tableshows the percent difference in inventory resultsobtained by reintroducing the outliers (individ-ual answers beyond two standard deviationsfrom the average value, identified in the initialsurvey results and not commented on query).Reasons for the occurrence of outliers rangefrom reporting mistakes (e.g., one nonrealisticlow alumina consumption report, bringing inthe same percent difference in all alumina-production-related data) to allocation issuesbetween operations within plants (e.g., waterinput and output, steel input) and to measurementissues (e.g., fluoride, polycyclic aromatic hydro-carbons, mercury air and water emissions, benzo-a-pyrene air emissions, and other by-productsfor external recycling), the latter being generallyresponsible for the largest differences.

Chapter 4: Life-Cycle Assessment of Aluminum / 69

(continued)

Table 4.1 Resource inputs from inventory data to produce 1000 kg of primary aluminumProcess

Inputs Bauxite mining Alumina production Anode production Electrolysis Ingot casting Total

Raw materialsBauxite 5168 . . . . . . . . . . . . 5168 kgCaustic sode . . . 159 . . . . . . . . . 159 kgCalcined lime . . . 86 . . . . . . . . . 86 kgAlumina 1925 kg

Petroleum coke . . . . . . 349 . . . . . . 349 kgPitch . . . . . . 92 . . . . . . 92 kgAnode 441 kg

Aluminum fluoride . . . . . . . . . 17.4 . . . 17.4 kgCathode carbon . . . . . . . . . 6.1 . . . 6.1 kgAluminum (liquid metal) 1000 kg

Alloy additives . . . . . . . . . . . . 20 20 kgChlorine . . . . . . . . . . . . 0.068 0.068 kgCast ingot 1000 kg

Other raw material inputsFresh water . . . 6.4 0.5 2.95 3.15 13.0 m3

Sea water . . . 6.5 0.001 20.8 0.2 27.5 m3

Refractory materials . . . . . . 5.5 6 . . . 11.5 kgSteel (for anodes) . . . . . . 1.4 . . . . . . 1.4 kgSteel (for cathodes) . . . . . . . . . 5.5 . . . 5.5 kg

Fuels and electricityCoal . . . 185 0.9 . . . . . . 186 kgDiesel oil 10.3 1.2 1.4 . . . 0.1 13.0 kgHeavy oil . . . 221.4 6.2 . . . 10 238 kgNatural gas . . . 233 23 . . . 52 308 m3

Electricity . . . 203 62 15,365 81 15,711 kWh

Table 4.2 Resource outputs from inventory data when producing 1000 kg of primary aluminumProcess

Outputs Bauxite mining Alumina production Anode production Electrolysis Ingot casting Total

Air emissionsFluoride gaseous (as F) . . . . . . 0.02 0.55 . . . 0.57 kgFluoride particulate (as F) . . . . . . 0.004 0.5 . . . 0.50 kgParticulates 12.2 1.2 0.1 3.3 0.08 16.9 kgNOx (as NO2) . . . 2.24 0.13 0.35 0.12 2.8 kgSO2 . . . 10.2 0.7 13.6 0.2 24.7 kgTotal polycyclic aromatic . . . . . . 0.02 0.13 . . . 0.15 kg

hydrocarbons (PAH)Benzo-a-pyrene (BaP) . . . . . . 0.1 5.0 . . . 5.1 gCF4 . . . . . . . . . 0.22 . . . 0.22 kgC2F6 . . . . . . . . . 0.021 . . . 0.021 kgHydrogen chloride (HCl) . . . . . . . . . . . . 0.067 0.067 kgMercury . . . 0.00020 . . . . . . . . . 0.00020 kg

Water emissionsFresh water . . . 6.4 . . . 3.2 3.8 13.4 m3

Sea water . . . 6.6 . . . 20.9 . . . 27.5 m3

Fluoride (as F) . . . . . . . . . 0.2 . . . 0.20 kgOil/grease . . . 0.13 . . . 0.008 0.009 0.15 kgPAH (six Borneff components) . . . . . . . . . 3.77 . . . 3.77 gSuspended solids . . . 1.43 . . . 0.21 0.02 1.66 kgMercury . . . 0.0018 . . . . . . . . . 0.0018 kg

By-products for external recyclingBauxite residue . . . 2.3 . . . . . . . . . 2.3 kgDross . . . . . . . . . . . . 13.0 13.0 kgFilter dust . . . . . . . . . . . . 0.57 0.57 kgOther by-products . . . 3.5 2.8 5.1 . . . 11.4 kgRefractory material . . . . . . 3.1 0.5 0.5 4.1 kgScrap sold . . . . . . . . . . . . 2.2 2.2 kg


Table 4.2 (continued)

Process

Outputs Bauxite mining Alumina production Anode production Electrolysis Ingot casting Total

Spent potliner (SPL) . . . . . . . . . 9.9 . . . 9.9 kgcarbon fuel—reuse

SPL refractory bricks—reuse . . . . . . . . . 5.5 . . . 5.5 kgSteel . . . . . . 1.7 6.9 . . . 8.6 kg

Solid wasteBauxite residue (red mud) . . . 1905 . . . . . . . . . 1905 kgCarbon waste . . . . . . 2.4 4.6 . . . 7.0 kgDross—landfill . . . . . . . . . . . . 7.7 7.7 kgFilter dust—landfill . . . . . . . . . . . . 0.4 0.40 kgOther landfill wastes 703 47.5 2.7 7.3 1.3 762 kgRefractory waste—landfill . . . . . . 2.5 1.2 0.7 4.4 kgScrubber sludges . . . . . . 0.8 13.7 . . . 14.5 kgSPL—landfill . . . . . . . . . 17.3 . . . 17.3 kgWaste alumina . . . . . . . . . 4.7 . . . 4.7 kg

Table 4.3 Units of input and output data for inventory of primary aluminum unit processInputs Unit Outputs Unit

Raw materials Air emissionsBauxite kg Fluoride gaseous (as F) kgCaustic soda (for alumina production) kg Fluoride particulate (as F) kgCalcined lime (for alumina production) kg Particulates kg

NOx (as NO2) kgPetroleum coke (for anode production) kg SO2 kgPitch (for anode production) kg Total Polycyclic aromatic

hydrocarbons (PAH) kgBenzo-a-pyrene (BaP) kg

Aluminum fluoride (for electrolysis) kg CF4 kgCathode carbon (for electrolysis) kg C2F6 kg

Hydrogen chloride (HCl) kgAlloy additives (for ingot casting) kg Mercury kgChlorine (for ingot casting) kg

Other raw material inputs Water emissionsFresh water m3 Fresh water m3

Sea water m3 Sea water m3

Refractory materials kg Fluoride (as F) kgSteel (for anodes) kg Oil/grease kgSteel (for cathodes) kg PAH (six Borneff components) g

Suspended solids kgMercury kg

Fuels and electricity By-products for external recyclingCoal kg Bauxite residue kgDiesel oil kg Dross kgHeavy oil kg Filter dust kgNatural gas m3 Other by-products kgElectricity kWh Refractory material kg

Scrap sold kgSpent potliner (SPL) carbon fuel—reuse kg

SPL refractory bricks—reuse kgSteel kg

Solid wasteBauxite residue (red mud) kgCarbon waste kgDross—landfill kgFilter dust—landfill kgOther landfill wastes kgRefractory waste—landfill kgScrubber sludges kgSPL—landfill kgWaste alumina kg


Table 4.4 Average greenhouse gas (GHG) emissions for primary aluminum unit process included inthis study

Bauxite Refining Anode Smelting Casting

GHG emissions [kg of CO2 equivalents per 1000 kg of process output]

Process 0 0 388 1626 0Electricity 0 58 63 5801 77Fossil fuel 16 789 135 133 155Transport 32 61 8 4 136Ancillary 0 84 255 0 0Perfluorocarbons 0 0 0 2226 0Total 48 991 849 9789 368

Source: Ref 4.3

Table 4.5 Difference on the inventory results from outlier reintroductionInputs Difference(a),% Outputs Difference(a),%

Raw material inputs Air emission outputsBauxite – 0.7 Fluoride gaseous (as F) 16Caustic soda – 0.7 Fluoride particulate (as F) 17Calcined lime – 0.7 Particulates 20Alumina – 0.7 NOx (as NO2) 14

SO2 2Petroleum coke (for anode production) 0 Total polycyclic aromatic hydrocarbons (PAH) 23Pitch (for anode production) 0 Benzo-a-pyrene (BaP) 6Anode 0 CF4 0

C2 F6 0Aluminum fluoride 3 Hydrogen chloride (HCl) 8Cathode carbon 6 Mercury 139Aluminum (liquid metal)

Alloy additives (for ingot casting) 0 Water emission outputsChlorine (for ingot casting) 6 Fresh water 28Cast ingot Sea water –0.2

Fluoride (as F) 39Oil–grease 62

Other raw material inputs PAH (six Borneff components) 102Fresh water 46 Suspended solids 73Sea water –0.2 Mercury 879Refractory materials 38Steel (for anodes) 12 By-products for external recyclingSteel (for cathodes) 0.7 Bauxite residue –0.7

Dross 32Fuel and electricity inputs Filter dust –2Coal –0.7 Other by-products 123Diesel oil –0.06 Refractory material 33Heavy oil –0.6 Scrap sold 20Natural gas –0.5 Spent potliner (SPL) carbon fuel—reuse 0Electricity –0.01 SPL refractory bricks—reuse 0

Steel 10

Solid waste outputsBauxite residue (red mud) 7Carbon waste 17Dross—landfill –2Filter dust—landfill 7Other landfill wastes 14Refractory waste—landfill 26Scrubber sludges 45SPL—landfill 3Waste alumina 53

(a) This table shows the percent difference in inventory results obtained by reintroducing the outliers (individual answers beyond two standarddeviations from the average value, identified in the initial survey results and not commented on query). Reasons for the occurrence of outliersrange from reporting mistakes (e.g., one nonrealistic low alumina consumption report, bringing in the same percent difference in all alumina-production-related data) to allocation issues between operations within plants (e.g., water input and output, steel input) and to measurementissues (e.g., fluoride, polycyclic aromatic hydrocarbons, mercury air and water emissions, other by-products for external recycling), the latterbeing generally responsible for the largest differences.


Missing Process Data Supplemented.During the course of the survey, it was realizedthat the life-cycle survey forms distributed tocompanies had omitted two sets of neededprocess data:

• Bauxite mining life-cycle data • Ingot casting energy-consumption data

Missing data have been worked out fromcurrently available information, as described inTables 4.6 and 4.7. Life-cycle data thusselected were considered quite representative.

Data Interpretation Items

As previously noted, the life-cycle inventorydoes not include data on nonelementary unitprocesses such as energy production, transport,petroleum coke and pitch production, causticsoda production, and so on. Other databases orreports can be considered for data on additionalunit processes. Examples of other data sourcesinclude:

• The global breakdown by source of electricityused at primary aluminum smelters in the IAIenergy survey for 2000 was as follows: hydro,52.5%; coal, 31.6%; oil, 0.8%; natural gas,9.0%; and nuclear, 6.1%.

• Environmental aspects from transport can beillustrated with a case from Ref 4.4, whichyields the following air emission levels fromtransport in proportion of those generatedfrom primary aluminum production: particu-lates, 1.1%; hydrocarbons, 29%; NOx,18.5%; and SO2, 6.7%.

Other important items on data interpretationare summarized as follows. When constructing alife-cycle assessment related to aluminum (whichincludes aluminum production, fabrication, prod-uct usage, and product end-of-life issues), amethodology in accordance with internationallyaccepted practice (ISO 14040 series standards)should be used.

Missing data for bauxite mining and ingotcasting energy have been worked out, as dis-cussed previously in the section “MissingProcess Data Supplemented” in this chapter.

Production-weighted mean values, whichcan be derived from the effective responses tothe survey, are normally reported. In severalsituations, however, this led to inconsistentresults. Data reporting then took place asindustry-weighted use, that is, the ratio of totalconsumption or emission reported from the

survey (total of all responses) to the total cor-responding industry production. This was usedfor the following:

• Petroleum coke and pitch consumption for an-odes (see Table 4.11 (a) is industry weighted,because mean values from anode productiontables are altered due to anode butt recycling.

• For fresh water and sea water use (inputsand outputs), this correction avoided theapparent input-output imbalance visiblewhen using weighted mean values.

• For fuel and electricity consumption foralumina and anode production, the rationalewas that the mix of fuel use is typically com-pany-specific; that is, using weighted meanvalues led to overestimated results.

Some data on air emissions (particulates,SO2 and NOx emissions) from fuel combustion(an energy unit process not documented in thepresent work) are included along with processemissions in the results reported from plants.This arises from the fact that the two emissiontypes (from fuel combustion and from process)do occur together. An improved reliability onactual emission levels can correspond to fuelcombustion as compared to general fuel combus-tion emission data. In particular, the actual sulfurcontent of fuel oil used, for example, in aluminaproduction, is thus most accurately accountedfor in SO2 emissions. Accordingly, it is recom-mended that life-cycle assessment practitionersusing data sets about emissions from fuel com-bustion remove collected data on particulatesand SO2 and NOx emissions in order to avoiddouble counting. (Note: This applies only tofuel combustion and not to precombustion datasets.)

Other Exceptions. A review of the resultscould raise remarks for apparent inconsistency orweakness of evidence. Noted examples are thesteel input-output imbalance (likely to be attrib-uted to a used steel output from general mainte-nance work), the ingot casting mass balance(probably linked to the cold metal contribution),and a low response rate for dioxin emissions iningot casting. These cases can probably be con-sidered as relatively nonsignificant, given theoverall purpose of the present inventory. It is alsoclear that the present survey produced the bestcurrently available knowledge for the worldwideprimary aluminum industry.

Dioxin emissions from ingot casting have notbeen included in the final inventory calculation,because they are related to aluminum scrap


remelting, which was outside the scope of thesurvey. Chlorine present in aluminum scrap(from painting or coating residues) is the originelement for dioxin emissions during aluminumremelting, and there is no chlorine in primaryaluminum. The relation with aluminum scrapremelting has been confirmed from the surveyresults, with higher dioxin emissions results rea-sonably correlated with higher scrap use, despitethe limited number of answers received (6 over72 cast houses) and a high scatter in values.

Unit Processes and Results by Process

Results from the IAI Aluminum Life-CycleSurvey 2000 are presented for the followingunit processes:

• Alumina production• Anode production (prebake) • Paste production (Soderberg) • Reduction (electrolysis) • Ingot casting

The life-cycle data are reported either asproduction-weighted mean values, which is thebasic situation, or as industry-weighted use, as

discussed in the section “Data InterpretationItems” in this chapter. The applicable definitionsare as follows:

• Production-weighted mean (wt. mean): Totalconsumption or emission reported, dividedby total corresponding industry productionof those plants that have reported data

• Industry-weighted use (industry wt. use):Total consumption or emission reported,divided by total corresponding industryproduction

For each unit process, the reference flow is 1metric tonne. The interrelationships of these unitprocesses are shown in Fig. 4.1, which providesan overview of material flows in the primaryaluminum production. As previously noted, nospecific additional unit processes, in particularabout energy production, transport, petroleumcoke and pitch production, caustic soda produc-tion, ansd so on, have been added to the processin order to avoid nonelementary flows.

Inventory data for the worldwide primary alu-minum are consolidated in Tables 4.1 and 4.2.The flow diagram of Fig. 4.2 combinesprocesses with an input of 5168 kg of bauxitefor a 1000 kg aluminum output. The maincomponents of this mass distribution of 1000 kg

Table 4.6 Supplemental bauxite mining energy dataNorth American

life-cycleTwo Australian Nine mines inventory study

Selected mines (Aachen) 1998*

Inputs

Diesel 2 kg/t(a) 0.90 kg/t 0.67–1.8 kg/t 4.37 kg/tOther oil . . . (1.16 kg/t medium fuel oil, 0.27 kg/t gasoline) 1.43 kg/tElectricity . . . 0.002 kWh/t . . . 0.4 kWh/tFresh water . . . 0.031 m3/t . . . . . .

Outputs

Particulates 2.35 kg/t(a) . . . 0.002–0.005 kg/t 2.35 kg/tSolid waste 136 kg/t(a) . . . . . . 136 kg/t

(a) Per tonne (103 kg) of bauxite output*Note: See chapter 3 for more information on the North American life-cycle inventory study 1998.

Table 4.7 Supplemental ingot casting energy dataEuropean Aluminum Assoc.

(25 cast houses. 3.3 Mio t, North American life-cycle Input Selected typical primary) Alcoa US. (K. Martchek) Two Australian (J. Pullen) inventory 1998*

Fuel oil 10.0 kg/t(a) 10.9 kg/t approx. 10 kg/t . . . 17.4 kg/tDiesel 0.1 kg/t(a) 0.1 kg/t . . . 0.9 L/t < 0.2 kg/tGas 52 m3/t(a) 17.6 m3/t 52 m3/t 84 m3/t 52 m3/tElectricity 81 kWh/t(a) 16 kWh/t 111 kWh/t 81 kWh/t 211 kWh/t

(a) Per tonne of cast metal output*Note: See chapter 3 for more information on the North American life-cycle inventory study 1998.


aluminum output from 5168 kg of bauxite input(Fig. 4.2) need some explanation, because thisis not an exact calculation. For example, there isalways a significant water component in thebauxite (5168 kg), typically approximately 20%(1034 kg). The non-aluminum-containing partof the bauxite is disposed of as bauxite residue(red mud, 1905 kg). The mass balance out of thealumina production process would be approxi-mately 2200 kg, after deduction of watercomponent and bauxite residue.

Aluminum oxide (alumina) is converted inthe electrolysis process (primary aluminumsmelting) by the following reaction:

2Al2O3 + 3C = 4Al + 3CO2

with a stoichiometric minimum requirement of1890 kg Al2O3 for 1000 kg of primary aluminum.The actual production process could be describedas an alumina breakdown by electrolysis, pro-ducing 1000 kg of aluminum and releasing

Fig 4.1 Interrelationships and materials flow of unit processes in aluminum primary production. Unit material processes not documented in the survey data are also shown. The production flow is as follows: aluminum is extracted from bauxite as

aluminum oxide (alumina); this oxide is then broken down through an electrolysis process into oxygen, emitted as CO2 by reactionwith a carbon anode, and aluminum as liquid metal; next, aluminum is cast into an ingot, the usual form suitable for further fabrica-tion of semifinished aluminum products.

Limestone mining

Petroleum cokeproduction

Pitch production Transport

Transport

Transport

Transport

Transport

Transport

Transport

TransportTransportCaustic sodaproduction*

Limestonecalcination

Bauxite mining

Alumina-production

Electrolysis

Anode production(including recycling

of anode buttsfrom electrolysis)

Cathode carbonproduction

Aluminum fluorideproduction

Alloy additivesproduction

Chlorineproduction

Remelt ingot*

Outside scrap*

Ingot casting scrap(Òrun-aroundÓ scrap)

Ingot casting

Primary aluminum to fabricating plant

*Input from remelt or recycled aluminum (cold metal) isexcluded as not representative for primary aluminum.

5168 kgbauxite

1925 kgalumina

1000 kg

1000 kg

liquid metal

slab, billet, etc.

0.068 kg

20 kg

17.4 kg

5.9 kg

441 kg anode

86 kg calcined lime

349 kg

92 kg

159 kg


oxygen on the carbon anode as CO2 (where 441kg coke and pitch input, considered as carbonweight, yield 1176 kg oxygen by difference, with1617 kg CO2 corresponding output).

Bauxite Mining

This unit process begins with the removal ofoverburden from a bauxite-rich mining site.Reusable topsoil is normally stored for latermine site restoration. The operations associatedwith this unit process include:

• Extraction of bauxite-rich minerals fromthe site

• Beneficiation activities, such as washing,screening, or drying

• Treatment of mining site residues and waste • Site restoration activities, such as grading,

dressing, and replanting

The output of this unit process is the bauxitethat is transported to an alumina refinery.

Bauxite mining activities mainly take place intropical and subtropical areas of the Earth.

Almost all bauxite is mined in an open-pit mine.The known reserves of alumina-containing orewill sustain the present rate of mining for 300 to400 years. Commercial bauxite can be sepa-rated into bauxite composed of mostly aluminatrihydrates and those composed of aluminamonohydrates. The trihydrate aluminas containapproximately 50% alumina by weight, whilemonohydrates are approximately 30%. Mono-hydrates are normally found close to the surface(e.g., Australia), while trihydrates tend to be atdeeper levels (e.g., Brazil).

The only significant processing difference inbauxite mining is the need for beneficiation.Beneficiation occurs with ores from forestedareas, while the grassland type typically doesnot require washing. The wastewater fromwashing is normally retained in a settling pondand recycled for continual reuse.

Alumina Production

This unit process begins with the unloading ofprocess materials to their storage areas on site.The operations associated with this unit processinclude:

• Bauxite grinding; digestion and processingof liquors

• Alumina precipitation and calcination• Maintenance and repair of plant and

equipment • Treatment of process air, liquids, and solids

The output of this unit process is smelter-gradealumina transported to an electrolysis plant (pri-mary aluminum smelter).

In alumina production, also commonly calledalumina refining, bauxite is converted to alu-minum oxide using the Bayer process, whichuses caustic soda and calcined lime (limestone)as input reactants. Bauxite is ground andblended into a liquor containing sodium car-bonate and sodium hydroxide. The slurry isheated and pumped to digesters, which areheated pressure tanks. In digestion, iron andsilicon impurities form insoluble oxides calledbauxite residue. The bauxite residue settles out,and a rich concentration of sodium aluminate isfiltered and seeded to form hydrate aluminacrystals in precipitators. These crystals are thenheated in a calcining process. The heat in thecalciners drives off combined water, leavingalumina. Fresh water (input taken conserva-tively whether the water used is from fresh,

Fig 4.2 The flow diagram of processes from 5168 kg of bauxiteinput to produce 1000 kg primary aluminum ingot

5168 kg

Alumina production

1925 kgAnode production

441 kg

Electrolysis

1000 kg

1000 kg

Ingot casting

Bauxite mining


underground, mine wastewater, or other sources)or sea water is used as the cooling agent.

The major differences in processing are at thecalcination stage. Two types of kilns are used:rotary and fluid bed. The fluid bed or stationarykiln is newer and significantly more energyefficient. Energy requirements (coal, diesel oil,heavy oil, natural gas, electricity) have almostbeen halved over the last 15 years with the in-troduction of higher-pressure digesters and fluidflash calciners.

Air emissions primarily arise from the calci-nation stage (particulates; NOX, as NO2 and SO2from fuel combustion; mercury found in bauxiteores), while water emissions come from coolinguse (fresh water, sea water, oil/grease) or arelinked with the digestion stage (suspendedsolids, mercury found in bauxite ores).

Most of the bauxite residue currently turnsout as solid waste, while a small but growingfraction is reused. Other by-products for exter-nal recycling are reaction chemicals. Otherlandfill wastes are typically inert componentsfrom bauxite, such as sand, or waste chemicals.

Survey results of inputs and outputs for alu-mina production are summarized, respectively,in Tables 4.8(a) and (b) based on productionnumbers of:

• 30,786,116 tonnes (metric tons) with life-cycle survey data from 23 refineries (whichis 59% of the estimated survey coverage ofthe total world production)

• 36,911,495 tonnes with the IAI energy sur-vey from 31 refineries (which is 70% of theestimated survey coverage of the total worldproduction)

Anode Production

Survey results for anode production are summa-rized for prebake (Tables 4.9a–c) and for paste

(Soderberg) anode production (Tables 10a–c).This unit process begins with the unloading ofprocess materials to their storage areas on site.The operations associated with this unit processinclude:

• Recovery of spent anode materials • Anode mix preparation; anode block or bri-

quette forming and baking• Rodding of baked anodes• Maintenance and repair of plant and


The output of this unit process is rodded an-odes or briquettes transported to an electrolysisplant (primary aluminum smelter).

There are two types of aluminum smeltingtechnologies that are distinguished by the typeof anode that is used in the reduction process:Soderberg and prebake. Soderberg design uses asingle anode that covers most of the top surfaceof a reduction cell (pot). Anode paste (bri-quettes) is fed to the top of the anode, and as theanode is consumed in the process, the pastefeeds downward by gravity. Heat from the potbakes the paste into a monolithic mass before itreaches to the electrolytic bath interface. Theprebake design uses prefired blocks of solid car-bon suspended from steel axial busbars. Thebusbars both hold the anodes in place and carrythe current for electrolysis.

The process for making the aggregate for bri-quettes or prebake blocks is identical. Petroleumcoke is calcined, ground, and blended with pitchto form a paste that is subsequently formed intoblocks or briquettes and allowed to cool. Whilethe briquettes are sent directly to the pots forconsumption, the blocks are sent to a separatebaking furnace.

Baking furnace technology has evolved fromsimple pits that discharged volatiles to the

Table 4.8a Alumina production inputs to produce 1000 kg of aluminaIndustry- Standard DQI DQI

Inputs weighted use Wt. mean deviation Unit(a) Response rate, % Min Max average(b) range(b)

Raw MaterialsBauxite . . . 2685 513 kg/t 100 2102 3737 1.1 1–2Caustic soda . . . 82 30 kg/t 100 29 181 . . . . . .Calcined lime . . . 45 44 kg/t 100 5 221 . . . . . .Fresh water 3.3 3.5 4.4 m3/t 91 0.003 14 1.4 1–3Sea water 3.4 11.4 17 m3/t 22 1.2 42 1.2 1–2

Fuels and electricityHeavy oil 115 154 137 kg/t 81 0.5 437 . . . . . .Diesel oil 0.6 1.7 2.4 kg/t 35 0.1 8 . . . . . .Gas 121 222 118 m3/t 35 2 384 . . . . . .Coal 96 374 266 kg/t 23 127 843 . . . . . .Electricity 106 170 188 kWh/t 65 8 749 . . . . . .

(a) Per tonne of alumina. (b) DQI, data quality indicator


Table 4.8b Alumina production outputs to produce 1000 kg of aluminaIndustry- Standard DQI DQI

Outputs weighted use Wt. mean deviation Unit(a) Response rate, % Min Max average(b) range(b)

Air emissionsParticulates . . . 0.63 1.9 kg/t 91 0.12 7.2 1.6 1–3SO2 . . . 5.3 8.9 kg/t 87 0.000003 24 1.6 1–2NOx (as NO2) . . . 1.17 0.68 kg/t 74 0.39 2.3 1.8 1–3Mercury . . . 0.10 0.06 g/t 26 0.004 0.16 2.1 1–3

Water emissionsFresh water 3.3 3.7 7.7 m3/t 87 0.58 33 1.9 1–3Sea water 3.4 11.4 17 m3t 22 1.2 42 1.3 1–2Suspended solids . . . 0.74 1.1 kg/t 52 0.0002 3.7 1.6 1–3Oil and grease/

total hydrocarbons . . . 0.069 0.10 kg/t 48 0.000001 0.27 1.9 1–3Mercury . . . 0.00094 0.0035 g/t 30 0.00003 0.0095 1.9 1–3

By-products (for external recycling)Bauxite residue . . . 1.2 0.5 kg/t 13 0.33 1.3 1.5 1–3Other . . . 1.8 3.2 kg/t 43 0.07 10.6 1.8 1–3

Solid wasteBauxite residues (red mud) . . . 990 407 kg/t 96 204 1916 1.7 1–3Other landfill wastes . . . 24.7 37 kg/t 87 0.18 161 1.9 1–3

(a) Per tonne of alumina. (b) DQI, data quality indicator

Table 4.9a Inputs to produce 1000 kg of baked anodeIndustry- Standard DQI DQI

Inputs weighted use Wt. mean deviation Unit(a) Response rate, % Min Max average(b) range(b)

Raw materialsPetroleum coke . . . 683 57 kg/t 100 599 888 . . . . . .pitch . . . 160 15 kg/t 100 135 200 . . . . . .Total . . . 843 . . . kg/t . . . . . . . . . . . . . . .

Fuels and electricityCoal 2.0 66 65 kg/t 4 4.1 96 . . . . . .Heavy oil 17.0 75 32 kg/t 23 14 160 . . . . . .Diesel oil 3.9 24 30 kg/t 13 0.02 72 . . . . . .Gas 63 90 33 m3/t 70 13 217 . . . . . .Electricity 158 185 96 kWh/t 86 0.19 450 . . . . . .

Other inputsFresh water 1.2 2.2 3.4 m3/t 56 0.004 12 1.8 1–3Sea water 0.0022 0.22 NA m3/t 2 0.22 0.22 1 1Refractory material . . . 12.5 18 kg/t 72 0.16 98 1.9 1–3Steel . . . 3.1 3.0 kg/t 44 0.04 10 1.9 1–3

From life cycle survey data with total production (baked) of 6,443,997 tonnes from 54 anode plants; and IAI energy survey data with totalproduction (baked) of 6,694,481 tonnes from 56 anode plants. (a) Per tonne anode. Note: Recycled anode butts account for the raw materialmass balance. (b) DQI, data quality indicator

Table 4.9b Outputs when producing 1000 kg of baked anodeStandard DQI DQI

Outputs Average deviation Unit(a) Response rate, % Min Max average(b) range(b)

By-products for externalrecycling

Refractory 6.9 6.2 kg/t 39 0.5 22.4 1.4 1–3Steel 3.9 2.4 kg/t 37 0.3 9.3 1.8 1–3Other 6.5 9.4 kg/t 31 0.0006 33.2 1.6 1–3

Solid waste not recycledWaste carbon or mix 6.0 4.9 kg/t 57 0.1 16.6 1.5 1–3Scrubber sludges 1.9 2.4 kg/t 15 0.04 6.5 1.4 1–3Refractory

(excluding spent potliner) 5.7 9.5 kg/t 46 0.1 33.6 1.9 1–3Other landfilled waste 5.3 6.4 kg/t 43 0.4 23.5 1.6 1–3

From life-cycle survey data with total production of 6,443,997 tonnes from 54 anode plants. (a) per tonne anode. (b)DQI, date quality indicator


Table 4.9c Air emissions when producing 1000 kg of baked anodeStandard DQI DQI

Outputs Wt. mean deviation Unit(a) Response rate, % Min Max average(b) range(b)

Air emissionsParticulates 0.33 0.39 kg/t 89 0.02 1.8 1.4 1–3SO2 1.8 1.6 kg/t 72 0.001 6.2 1.6 1–3NOx (as NO2) 0.31 0.23 kg/t 69 0.02 1.3 1.7 1–3Particulate fluoride (as F) 0.010 0.02 kg/t 57 0.000002 0.06 1.4 1–3Gaseous fluoride (as F) 0.046 0.16 kg/t 76 0.00001 0.9 1.2 1–3Total polycyclic

aromatic hydrocarbons 0.060 0.11 kg/t 76 0.00003 0.5 1.5 1–3Benzo-a-pyrene 0.27 0.68 g/t 54 0.00003 3.4 1.7 1–3

From, life-cycle survey data with total production of 6,443,997 tonnes from 54 anode plants. (a) Per tonne anode. (b) DQI, data quality indicator

Table 4.10a Inputs to produce 1000 kg of anode pasteIndustry- Standard

Inputs weighted use Wt. mean deviation Unit(a) Response rate, % Min Max

Raw materialsPetroleum coke 713 26 kg/t 100 669 760Pitch 284 25 kg/t 100 240 331Total 997 . . . kg/t . . . . . . . . .

Fuels and electricityCoal 2.6 65 NA kg/t 4 65 65Heavy oil 0.40 6.4 83 kg/t 11 0.6 145Diesel oil . . . . . . . . . kg/t . . . . . . . . .Gas 6.6 22 12 m3/t 19 5 37Electricity 65 106 69 kWh/t 63 33 264

Other inputsFresh water 1.5 2.0 2.6 m3/t 53 0.0004 8.2Sea water . . . . . . . . . m3/t . . . . . . . . .Refractory material . . . . . . . . . kg/t . . . . . . . . .

From life-cycle survey data with total production of 948,457 tonnes from 17 paste plants; and IAI energy survey data with total production of1,466,841 tonnes from 27 paste plants. (a) Per tonne anode

Table 4.10b Outputs when producing 1000 kg of anode pasteStandard DQI DQI

Outputs Average deviation Unit(a) Response rate,% Min Max average(b) range(b)

By-products for externalrecycling

Refractory . . . . . . kg/t . . . . . . . . . . . . . . .Steel . . . . . . kg/t . . . . . . . . . . . . . . .Other 0.63 NA kg/t 6 0.63 0.63 . . . . . .

Solid waste not recycledWaste carbon or mix 0.11 0.01 kg/t 24 0.13 0.17 1.7 1–3Scrubber sludges . . . . . . kg/t . . . . . . . . . . . . . . .Refractory

(excluding spent potliner) . . . . . . kg/t . . . . . . . . . . . . . . .Other landfilled waste 14.7 30 kg/t 18 0.4 53 1.7 1–3

From life-cycle survey data with total production of 948,457 tonnes from 17 paste plants. (a) Per tonne anode. (b) DQI, data quality indicatorBy-products for external recycling

Table 4.10c Air emissions when producing 1000 kg of anode pasteStandard DQI DQI

Air emissions Wt. mean deviation Unit(a) Response rate, % Min Max average(b) range(b)

Particulates 0.11 0.20 kg/t 71 0.01 0.72 1.3 1–3SO2 1.0 1.1 kg/t 41 0.002 2.57 1.6 1–2NOx (as NO2) 0.11 0.09 kg/t 29 0.05 0.26 2.4 2–3Particulate fluoride (as F) . . . . . . kg/t . . . . . . . . . . . . . . .Gaseous fluoride (as F) . . . . . . kg/t . . . . . . . . . . . . . . .Total polycyclic

aromatic hydrocarbons 0.0092 0.010 kg/t 53 0.001 0.03 1.1 1–2Benzo-a-pyrene 0.079 0.084 g/t 41 0.0002 0.23 1.6 1–3

From life-cycle survey data with total production of 948,457 tonnes from 17 paste plants. (a) Per tonne anode. (b) DQI, data quality indicator


atmosphere during the baking cycle to closed-loop-type designs that convert the caloric heatof the volatile into a process fuel that reducesenergy consumption for the process. Bakingfurnaces use refractory materials for linings,and fresh water (input taken conservativelywhether the water used is from fresh, under-ground, mine wastewater, or other sources) (orpossibly sea water) as the cooling agent. Bakingfurnaces account for most of the energy con-sumption (coal, diesel oil, heavy oil, naturalgas, electricity).

Air emissions—fluoride gaseous (as F), orfluoride particulate (as F)—arise from recoveredspent anode materials (unused anode ends, oranode butts) from electrolysis (see the followingsection) that are recycled within prebake anodeproduction. Particulates—NOX (as NO2) on SO2— typically come from fuel combustion. Totalpolycyclic aromatic hydrocarbons (PAH), whichincludes benzo-a-pyrene (BaP), are air emis-sions generated from the basic anode productionprocess. A common practice for their preventionand monitoring is water scrubbing, a processusing fresh water (input taken conservativelywhether the water used is from fresh, under-ground, mine wastewater, or other sources) orsea water as input and resulting in correspondingfresh water or sea water discharges (the latteraccounted for in the inventory together with theelectrolysis water discharges from scrubbing;see the following section).

By-products for external recycling includethe recovery of used steel from anode bars or

used refractory material from baking furnaces.Various other by-products are also recovered,for example, carbon recovered for reuse.

Solid waste that is not recycled (landfill)includes waste carbon or a mixed residue fromanode production, scrubber sludges arising fromwater scrubbing used for control of air emissionsmentioned previously, refractory waste from bak-ing furnaces, and other landfill wastes that ariseas various residues, for example, carbon fines.

Electrolysis

Survey results for electrolysis (smelting) are datathat came from all existing major technologytypes (Tables 4.11a–c). Approximately 15% ofthe total capacity surveyed was from Soderbergfacilities, and the remaining 85% was producedin prebake facilities. The supplied aluminaproduction data came from facilities currentlyin operation.

The unit process of electrolysis begins with theunloading of process materials to their storageareas on site. The operations associated with thisunit process include:


• Manufacture of major process equipment(e.g., cathodes)

• Process control activities (metal, bath, heat) • Maintenance and repair of plant and


Table 4.11a Inputs to produce 1000 kg of liquid aluminum from reduction (electrolysis)Industry- Wt. Standard Response DQI DQI

Inputs weighted use mean deviation Unit(a) rate, % Min Max average(b) range(b)

Raw materialsAlumina (dry) . . . 1925 27 kg/t 96 1871 2033 1.3 1–3Anode prebake (net) . . . 426 25 kg/t 65 388 546 . . . . . .Soderberg paste . . .` 510 32 kg/t 29 440 584 . . . . . .Anodes (net)/Soderberg paste 441 . . . . . . kg/t 95 . . . . . . . . . . . .Petroleum coke 349 . . . . . . kg/t . . . . . . . . . . . . . . .Pitch 92 . . . . . . kg/t . . . . . . . . . . . . . . .

Electricity consumptionElectricity 15,365 1179 kWh/t 100 13,405 19,446 . . . . . .

Other inputsFresh water 2.9 3.9 7.8 m3/t 73 0.002 32 1.6 1–3Sea water 20.7 163 203 m3/t 16 0.5 594 1.3 1–3Cathode carbon . . . 6.1 3.4 kg/t 93 1.1 16 1.6 1–3Refractory material . . . 6.0 4.2 kg/t 88 0.2 18 1.8 1–3Steel . . . 5.5 5.0 kg/t 85 0.1 35 1.8 1–3AlF3 . . . 17.4 5.4 kg/t 94 6.9 32 1.1 1–2

From life-cycle survey data with total production of 14,692,748 tonnes from 82 smelters representing approximately 60% total world produc-tion, of which Soderberg was used for 2,001,454 tonnes (29 smelters), with 68 smelters using baked anodes; and from IAI energy survey datawith total (production) of 16,822,420 tonnes from 112 smelters representing approximately 69% coverage of total world production. (a) Pertonne liquid aluminum. (b) DQI, data quality indicator


The output of this unit process is hot metaltransported to an ingot casting facility.

Molten aluminum is produced from alumina(aluminum oxide) by the Hall-Heroult elec-trolytic process that dissolves the alumina in amolten cryolite bath (requiring aluminum fluo-ride input) and passes current through this solu-tion, thereby decomposing the alumina into alu-minum and oxygen. Aluminum is tapped out ofthe reduction cell (pot) at daily intervals, and theoxygen combines with the carbon of the anode toform carbon dioxide.

The pot consists of a steel (for cathodes) shelllined with refractory materials insulation and

with a hearth of carbon (cathode carbon forelectrolysis). This is known as the cathode. Thecathode is filled with a cryolite bath and alu-mina, and an anode is suspended in the bath tocomplete the circuit for the pot. When started, apot will run continuously for the life of the cath-ode, which may last in excess of 10 years. Atthe end of its life, each pot is completely refur-bished. Steel from used cathodes is recoveredfor recycling. Refractory materials are eitherrecycled as by-products or landfilled (refractorywaste—landfill). Spent potlinings (SPLs),which include a carbon-base (SPL carbon) anda refractory-based part (SPL refractory bricks),

Table 4.11b Outputs when producing 1000 kg of liquid aluminum with electrolysisIndustry- Wt. Standard Response DQI DQI

Outputs Weighted use mean deviation Unit(a) rate, % Min Max average(b) range(b)

Water dischargesFresh water 3.1 4.7 8.0 m3/t 71 0.02 31.8 1.7 1–3Sea water 20.9 192 206 m3/t 15 0.1 594 1.4 1–3Suspended solids . . . 0.21 0.56 kg/t 61 0.0002 2.7 1.4 1–3Oil and grease/

total hydrocarbons . . . 0.0078 0.014 kg/t 41 0.00002 0.05 1.6 1–3Fluorides (as F) . . . 0.20 0.7 kg/t 70 0.00001 3.9 1.4 1–3Polycycylic aromatic

hydrocarbons (six Borneff components) . . . 3.8 9.3 g/t 32 0.000002 32 1.5 1–3

By-products for external recyclingSpent potlinen (SPL) carbon . . . 9.9 11 kg/t 35 0.25 41 1.2 1–3SPL refractory . . . 5.5 6.1 kg/t 26 0.60 19 1.2 1–3Refractory (other) . . . 0.53 0.8 kg/t 12 0.11 2.6 1.3 1–3Steel . . . 6.9 4.8 kg/t 74 0.13 20 1.6 1–3Other . . . 5.1 7.7 kg/t 49 0.13 26 1.5 1–3

Solid waste not recycledSPL . . . 17.3 12 kg/t 79 0.09 53 1.4 1–3Waste alumina . . . 4.7 7.3 kg/t 43 0.06 30 1.5 1–3Waste carbon or mix . . . 4.6 5.4 kg/t 40 0.01 20 1.4 1–3Scrubber sludges . . . 13.7 20 kg/t 16 0.04 50 1.3 1–3Refractory (excluding SPL) . . . 1.2 1.6 kg/t 40 0.05 6 1.6 1–3Other landfilled waste . . . 7.3 9.1 kg/t 71 0.06 33 1.6 1–3

From life-cycle survey data with total production of 14,692,748 tonnes from 82 smelters representing approximately 60% of total worldproduction, of which Soderberg had 2,001,454 tonnes (with 29 smelters) and 68 smelters with prebaked anodes. (a) Per tonne liquidaluminum. (b) DQI, data quality indicator

Table 4.11c Air emission when producing 1000 kg of liquid aluminum from electrolysisWt. Standard Response DQI DQI

Air emissions mean deviation Unit(a) rate, % Min Max average(b) range(b)

Particulates 3.3 5.0 kg/t 89 0.04 26 1.3 1–3SO2 13.4 6.6 kg/t 89 0.5 25 1.7 1–3NOx (as NO2) 0.35 0.8 kg/t 52 0.000004 3.9 1.8 1–3Particulate fluoride (as F) 0.50 0.8 kg/t 88 0.005 3.7 1.3 1–3Gaseous fluoride (as F) 0.55 0.9 kg/t 91 0.02 4.7 1.2 1–3Total polycyclic aromatic hydrocarbons 0.13 0.4 kg/t 44 0.0001 1.3 1.8 1–3Benzo-a-pyrene 5.0 16 g/t 35 0.0001 59.4 1.8 1–3CF4 0.22 0.40 kg/t 100 0.007 1.8 . . . . . .C2F6 0.021 0.040 kg/t 100 0.001 0.18 . . . . . .

From life-cycle survey data with total production of 14,692,748 tonnes representing 60% of total world production with 82 smelters, of whichsoderberg smelters (with 2,001,454 tonnes production) and 68 smelters with prebaked anodes; and from IAI per fluorocarbon survey data withtotal production of 16,079,065 tonnes representing 66% of total world production from 115 smelteres. (a) Per tonne liquid aluminum. (b) DQI,data quality indicator


are either recycled as by-products (SPL carbonfuel—reuse, SPL refractory bricks—reuse) orlandfilled (SPL—landfill).

The current in a pot varies from 60,000 to over300,000 A at a voltage drop of 4.2 to 5.0 V. Potsproduce approximately 16.2 +/– 0.6 lb per dayof aluminum for each kiloampere, at an operatingefficiency of 91 +/– 4%. Electricity consumptionis the major energy aspect of electrolysis.

Aluminum smelters typically use an air pollu-tion control system to reduce emissions. Theprimary system is usually a scrubber. Someplants use dry scrubbers, with alumina as theabsorbent that is subsequently fed to the potsand allows for the recovery of scrubbed materi-als. Other plants use wet scrubbers, whichrecirculate an alkaline solution to absorb emis-sions; the wet scrubbing process uses fresh water(input taken conservatively whether the waterused is from fresh, underground, mine waste-water, or other sources) or sea water as inputand results in corresponding fresh water or seawater discharges. Unlike dry scrubbers, wetscrubbers absorb carbon dioxide, nitrogen oxide,and sulfur dioxide that are entrained in thewastewater liquor (which is subsequently treatedprior to final discharge). Scrubber sludges arelandfilled.

Air Emissions. Specific aluminum electrolysisprocess emissions are fluoride gaseous (as F)and fluoride particulate (as F), which arise fromthe molten bath; total PAH, including BaP,which arises from anode consumption; and CF4and C2F6, commonly reported as perfluorocar-bons, are gases generated with an uncontrolledanode overvoltage situation called anode effect.Particulates, NOX (as NO2), and SO2, typicallycome from fuel combustion.

Water Emissions. Fluoride (as F) and PAH(Six Borneff components that are monitoredbecause of their particular environmentaleffect) arise from the same origin as theiraforementioned air emission equivalents. Sus-pended solids and oil and grease (or totalhydrocarbons) are monitored in water dis-charges from wet scrubbing.

Solid Waste. Other landfill wastes typicallyconsist of approximately 60% environmentalabatement waste (such as dry scrubber filterbags) and 40% municipal waste (Ref 4.5).

Ingot Casting

Survey results for ingot casting are summarizedin Tables 4.12(a) to (c). With the focus on pri-mary production, secondary ingots are not in-cluded, and contributions of remelt or recycledaluminum are excluded. Survey results for theprocess ingot casting yielded a higher weightoutput (1000 kg) than the correspondingelectrolysis metal input (874 kg), due to a “coldmetal” input contribution from remelt aluminum(133 kg remelt ingot) and recycled aluminum(101 kg outside scrap). Because the scope ofthis inventory report is primary aluminum andnot remelt or recycled aluminum, data for theunit process ingot casting were calculatedexcluding the contribution from “cold metal,”that is, all inputs and outputs from the surveyaverage were adjusted by a factor of 0.79 (inputratio [electrolysis metal + alloy additives = 892kg] / [total metal input = 1126 kg]—see Table4.12a). According to the ISO standards on life-cycle assessment, this can be described as asituation of joint process where a mass alloca-tion approach is applied.

Table 4.12a Inputs for casting 1000 kg of aluminum ingotIndustry- Wt. Standard Response DQI DQI

Inputs weighted use mean deviation Unit(a) rate, % Min Max average(b) range(b)

Electrolysis metal(c) . . . 874(d) 165 kg/t 96 373 1229 1.03 1–2Remelt ingot . . . 133 101 kg/t 67 0.09 451 1.2 1–3Outside scrap . . . 101 107 kg/t 47 0.5 388 1.1 1–3Alloy additives(c) . . . 18(e) 14 kg/t 90 0.004 67 1.06 1–3Fresh water 3.9 4.6 10.7 m3/t 89 0.001 48 1.6 1–3Sea water 0.23 10.5 12.7 m3/t 3 0.8 19 1.2 1–3Chlorine . . . 0.086 0.11 kg/t 61 0.001 0.42 1.3 1–3

From life-cycle survey data with total production of 14,016,405 tonnes representing 57% of total world production from 72 cast houses.(a) Per tonne of aluminum.(b) DQI, data quality indicator.(c) Metal input adjusted to exclude contribution from cold metal for 1000kg ingot output.(d) Electrolysis metal adjusted to 1000 kg/t for share of ingot casting inputs and outputs.(e) Alloy additives adjusted to 20 kg/t for primary aluminum life-cycle calculation


This unit process begins with the unloadingof process materials to their storage areas onsite. The operations associated with this unitprocess include:

• Pretreatment of hot metal (cleaning andauxiliary heating)


• Batching, metal treatment, and castingoperations

• Homogenizing, sawing, and packagingactivities

• Maintenance and repair of plant andequipment


The output of this unit process is packagedaluminum ingots or alloyed hot metal trans-ported to an aluminum fabricating facility.

Molten metal siphoned from the pots (elec-trolysis metal) is sent to a resident casting com-plex found in each smelter. In some cases, due toproximity, molten metal is transported directlyto a shape casting foundry. Remelt ingot andoutside scrap may also be used as metal input.

Molten metal is transferred to a holding furnace,and the composition is adjusted, by use of alloyadditives, to the specific alloy requested by acustomer. In some instances, depending on theapplication and on the bath composition in thepots, some initial hot metal treatment to removeimpurities may be done.

When the alloying is complete, the melt isfluxed to remove impurities and reduce gascontent. The fluxing consists of slowly bub-bling a combination of nitrogen and chlorine orcarbon monoxide, argon, and chlorine throughthe metal (chlorine use results in hydrogenchloride, or HCl, air emissions). Fluxing mayalso be accomplished with an in-line degassingtechnology, which performs the same functionin a specialized degassing unit.

Fluxing removes entrained gases and inorganicparticulates by flotation to the metal surface.These impurities (typically called dross) areskimmed off. The skimming process also takessome aluminum, and as such, drosses are nor-mally further processed to recover the aluminumcontent and to make products used in the abra-sives and insulation industries.

Table 4.12c Air emissions per 1000 kg of cast aluminum ingotWt. Standard Response DQI DQI

Air emissions mean deviation Unit(a) rate, % Min Max average(b) range(b)

Particulates 0.10 0.13 kg/t 71 0.001 0.53 1.5 1–3SO2 0.29 0.89 kg/t 51 0.00005 3.2 1.9 1–3NOx (as NO2) 0.16 0.14 kg/t 78 0.001 0.72 1.9 1–3HCl 0.085 0.085 kg/t 39 0.0003 0.33 1.7 1–3Dioxin/furans 0.0061 0.014 mg/t 8 0.0000003 0.035 2.5 1–3

From life-cycle survey data with total production of 14,016,405 tonnes representing 57% of total world production with 72 cast houses. (a) Pertonne aluminum. (b) DQI, data quality indicator

Table 4.12b Outputs per 1000 kg of cast aluminum ingotIndustry- Wt. Standard Response DQI DQI

Outputs weighted use mean deviation Unit(a) rate, % Min Max average(b) range(b)

Water dischargesFresh water 4.8 7.6 13 m3/t 72 0.001 43 1.7 1–3Suspended solids . . . 0.027 0.047 kg/t 47 0.0002 0.20 1.8 1–3Oil and grease/

total hydrocarbons . . . 0.011 0.026 kg/t 46 0.0000004 0.10 1.9 1–3

By-products for external recyclingDross . . . 16 7.4 kg/t 92 3.2 36 1.04 1–3Filter dust . . . 0.72 0.5 kg/t 10 0.2 1.4 1.2 1–3Scrap sold . . . 2.8 3.3 kg/t 29 0.08 10 1.1 1–3Refractory . . . 0.61 0.41 kg/t 11 0.014 1.2 1.4 1–3

Solid waste not recycledDross . . . 9.7 10 kg/t 18 2.0 30 1.3 1–3Filter dust . . . 0.50 0.5 kg/t 22 0.001 1.6 1.6 1–3Refractory . . . 0.81 0.7 kg/t 49 0.04 4.3 1.7 1–3Other landfilled waste . . . 1.6 2 kg/t 49 0.01 7.8 1.7 1–3

From life-cycle survey data with total production of 14,016,405 tonnes representing 57% of total world production with 72 cast houses. (a) Pertonne aluminum. (b) DQI, data quality indicator


Depending on the application, metal is thenprocessed through an in-line filter to removeany oxides that may have formed. Metal is thencast into ingots in a variety of methods: openmolds (typically for remelt ingot), direct chillmolds for various fabrication shapes, electro-magnetic molds for some sheet ingots, andcontinuous casters for aluminum coils. Freshwater (input taken conservatively whether thewater used is from fresh, underground, minewastewater, or other sources), seldom sea water,is used for cooling (often with recirculationthrough a cooling tower and water treatmentplant) and is subsequently discharged, wheresuspended solids and oil/grease (or total hydro-carbons) are monitored.

Energy used for ingot casting is electricity,natural gas, or heavy oil. Diesel oil is normallyused for internal plant transport.

While recovery and handling of internalprocess scrap is usually included in the ingotcasting operation, as mentioned previously, someprefer to sell it out (scrap sold as by-productfor external recycling). Dross, filter dust frommelting furnace air filtration, and refractorymaterial from furnace internal linings areeither recovered as by-products for external recy-cling or landfilled (dross—landfill, filter dust—landfill, refractory waste—landfill).

Solid Waste. Other landfill wastes typicallyconsist of approximately 80% of environmen-tal abatement waste (such as metal filter boxand baghouse) and 20% municipal waste(Ref 4.5).

Aluminum Life-Cycle Assessment withRegard to Recycling Issues

This section illustrates the ISO rules andfocuses on the major crucial aspects of aluminumlife-cycle assessment (LCA) in terms of energyaspects, the high recycling rates, and the highvalue of recycled aluminum. In the past, LCAstudies varied to a high degree because ofdifferent methodological approaches, whichcaused market risks when different materialswere compared. The ISO 14040 series of stan-dards now have set common methodologicalrules, which should be applied to all LCAstudies, including those dealing with aluminum.

The recommended method in assessing life-cycle factors with recycling is the substitutionmethod, which is explained and illustrated byexamples. Because all metals are characterized

by their ability to maintain their inherentproperties after recycling, contrary to wood,paper, concrete, or plastics, this method can beapplied for all metals, including aluminum. Inaddition to this substitution method, the so-called value-corrected substitution method isalso presented. This method tries to take intoaccount the loss of substitution ability of therecycled material in special cases where thevalue of the material is not maintained by recy-cling. More detailed guidance and scientificbackground, particularly on recycling issues, isgiven by Werner (Ref 4.6).

General

Any LCA study, especially for comparativepurposes, should be based on methodologieswithin the framework of the following inter-national standards:

• ISO 14040 Life-Cycle Assessment—Principles and Framework

• ISO 14041 Life-Cycle Assessment—Goaland Scope Definition and Inventory Analysis

• ISO 14042 Life-Cycle Assessment—Life-Cycle Impact Assessment

• ISO 14043 Life-Cycle Assessment—Life-Cycle Interpretation

In these standards, LCA is considered as atechnique to assess the environmental aspectsand potential impacts associated with a productor a service on a life-cycle basis. The LCAstudy includes four different phases:

• Goal and scope definition • Life-cycle inventory analysis • Life-cycle impact assessment • Interpretation

The LCA deals with product systems thatcomprise the full life cycle of a product, in-cluding raw material acquisition, fabrication,transportation, use, recycling/disposal, and theoperations of energy supply, ancillary materialsupply, and transports. Ideally, such a productsystem should only have input and output,which are elementary flows, that is, material orenergy that is drawn from the environment or isdiscarded to the environment without subsequenthuman transformation.

In LCA studies where aluminum is comparedto other materials, this comparison should bebased on the same functional unit; for example,1 kg of aluminum in a car may, in a specific


case, correspond to 1.8 kg of conventional steel,in order to fulfill the same function.

Quantitative aggregation differs over impactcategories; for example, the calculation of asingle score is not permitted by ISO 14042 forstudies to be used for comparative assertionsthat are made available to the public, where theoverall environmental superiority or equivalenceof one product versus a competing product thatperforms the same function is claimed.

A thorough interpretation of the results of anLCA study is required. This may include theneed of additional information about the dataof this report, that is, data quality or further in-formation about the data, for example, a tem-poral or spatial differentiation of the potentialenvironmental impact.

Energy Flows

Inventory data should not be added up if theyrepresent different types of potential environ-mental impact. It is not appropriate to add up allemissions, for example, on per kilogram basisor all energy flows in megajoules. A cumulativeenergy, which is understood as the sum ofrenewable (hydro) energy and nonrenewablefossil energies, has no ecological sense.

In the impact-assessment phase, data repre-senting elementary flows are aggregated to so-called indicator results if they belong to thesame impact category, possibly after havingbeen multiplied with a characterization factor.The elementary flow data related to energy con-sumption can be determined as the mass or theenergy content of the relevant energy resource.

Electricity supply data cannot be consideredas elementary flow data, because electricity isnot directly extracted from nature. The elemen-tary flows of the relevant power plant, includingextraction of fossil energy resources, must beconsidered instead.

An example of an indicator result for the im-pact category “extraction of energy resources” isthe low calorific value of the different fossil fuelsextracted for the supply of a certain quantity ofenergy (see ISO/TR 14047). If, for example, 100MJ of energy from natural gas are consumed in aplant, this may indicate an extraction of gasresources of 110 MJ, because 10 MJ may be usedfor the gas supply system or lost by leakage.

It is not appropriate to assign solar radiation ordam water, the elementary flow input of powerplants based on renewable energy, to the impactcategory “extraction of energy resources.”

Renewable energy flows have a different associ-ated type of impact (if considered) than fossil ornuclear energy flows. Generally, it is difficult tojustify an impact category that comprises ele-mentary flows from both renewable and nonre-newable energy in accordance with the criteriaas formulated in ISO 14042. On the other hand,other environmental impacts of power plantsbased on renewable energy, for example, theland use of hydroelectric power plants, must beconsidered.

Recycling

For most aluminum products, aluminum is notcompletely consumed but rather used. There-fore, a life cycle of an aluminum product is notcradle-to grave but rather “cradle-to-cradle.This means that the life cycle of an aluminumproduct usually ends when the recycled alu-minum is rendered in a form usable for a newaluminum product, for example, as an ingot.

According to ISO 14041, allocation princi-ples and procedures where input and output ofspecific processes must be shared by more thanone product system also apply to recycling situ-ations. In such cases, the environmental bur-dens related to extraction and processing of rawmaterials and final disposal of products mayhave to be shared with subsequent product lifecycles. This can also be addressed by using thesubstitution method, based on parameters suchas recycling rates and related metal yields. If achange in inherent properties is considered, thevalue of recycled material may play a role.

System Expansion and Substitution. Systemexpansion, an ISO 14041 recommended proce-dure, means expanding the system under studyto include the end-of-life recycling, resultingin substitution of recycled aluminum to primaryaluminum. The system boundaries must bedefined in the goal and the scope definitionphase of an LCA study. For the design for envi-ronment of aluminum products, it is appropri-ate to define the system boundaries in a way thatthey include all processes that can be influencedby the design. For aluminum products, mate-rial losses from recovery operations and metalyield by remelting depend on the form of theproduct, for example, wall thickness and the wayaluminum is joined with other materials thatcan be significantly influenced by the design.This means that the end-of-life recyclingprocesses of the product under study should beincluded, and the recycled aluminum should


leave the system in the form of the recycledingot.

A closed-loop allocation procedure is not onlyapplicable for really closed-loop product sys-tems, but also applies to open-loop product sys-tems, for example, when aluminum, after use, isrecycled into a raw material that has the same in-herent properties as primary aluminum. In thiscase, the system expansion and substitutionmethod can be applied. Unlike materials such aswood, paper, plastics, or concrete, aluminum hasthe ability to readily maintain its inherent prop-erties through recycling. The inherent propertiesof pure aluminum are not changed by remelting.

In the case of the open-loop recycling ap-proach, the substitution method can only beapplied if the recycled raw material is similarenough to the primary raw material that itsubstitutes. In many instances, however, therewill be a difference between a recycled materialand the primary material it may substitute.Strict interpretation of this rule thus limits theapplication of the substitution method to onlyspecial cases.

In practice, aluminum products followgenerally an open-loop recycling scheme. Recyc-ling operations that include collecting, sorting,remelting, and refining produce recycledaluminum that fulfills the requirements forprimary aluminum. Within these markets, recy-cled aluminum perfectly substitutes primaryaluminum. As a result, in most cases, the substi-tution method is fully applicable to the LCA ofaluminum products. However, in some particu-lar cases, the recycling operations can lead to asignificant change of the inherent properties ofthe recycled material compared with primaryaluminum, for example, by the presence ofmetallic impurities that are entrapped duringthe remelting operation of poorly sorted orcontaminated scrap. In this case, an allocationprocedure with a value-corrected approach isappropriate (see the sections “Value-CorrectedSubstitution” and “Substitution Method withRecycled Aluminium as Input” in this chapter.

For example, if 100 kg of primary aluminum isrequired for a product system, then 80 kg of recy-cled aluminum ingots (with the same inherentproperties as primary aluminum) are obtainedafter recovery of the end-of-life product andscrap remelting. The 20 kg of aluminum differ-ence is due to lost metal, for example, littered orlandfilled. In this case, only the environmentalburdens of the production of the lost aluminum(20 kg of primary metal) must be charged to the

product system under study, together with theburdens of the recovery operations. The environ-mental burdens of the production of 80 kg ofprimary aluminum must be charged to the nextuser(s) of the 80 kg of recycled aluminum.

Value-Corrected Substitution. The so-calledvalue-corrected substitution method considersthat the recycled metal is not able to fully substi-tute primary metal. This method assumes that thesubstitution ability is reflected by the ratiobetween the market prices of the recycled andprimary material. As a result, if the price of therecycled material is 90% of the price of the pri-mary material, 1 kg of recycled material will sub-stitute only 900 g of primary material. In the caseof aluminum, this method assumes a proportionof the environmental loads caused by primaryaluminum production and final disposal of alu-minum and the value change of the recycledmetal. This procedure is in line with ISO 14041,whereby allocation must be considered when oneor more unit processes are shared by differentproduct systems, including the product systemunder study. In this case, it is required to identifythese multifunctional processes and to selectthe appropriate allocation factors. The option touse economic parameters, for example, marketprices, for the calculation of allocation factors,is permitted by the standard, provided that:

• No option to avoid allocation can be justified.• No physical factor can be justified as the

allocation factor.

In the case of recycling, the ISO 14041standard offers the option to share, betweensubsequent product systems, inputs and outputsrelated only to the processes of raw materialacquisition and final disposal. This means thatonly specific unit processes within the productsystem are considered to be shared with otherfuture product systems, which can be unknownat the LCA study stage. This allocation principleavoids double-counting of inputs and outputsand ensures that the product system is chargedaccording to its actual material consumption.Moreover, guidance is given in this standard touse economic value proportions (e.g., scrapvalue versus primary value). This option makessense as long as reasonably free market pricesof the primary and recycled material exist.Nevertheless, other methodologies, such as howto treat open-loop recycling, may be justified asalternatives. In this case, a comparison througha sensitivity analysis is required. If aluminum iscompared with other materials, then it must be


clarified that the value-corrected substitutionmethod can be applied to the other materials aswell.

For example if from 100 kg of primary alu-minum in a product system, 80 kg of recycledaluminum ingots have 90% of the value ofprimary aluminum, then:

• 20 kg of aluminum are lost, for example, lit-tered or landfilled.

• An additional loss by value correction is10% of 80 kg (= 8 kg).

The total value-corrected loss is thus 28 kg.After value correction, the recycled aluminumingots (which have the value level of 72 kg ofprimary ingots) substitute 72 kg of primaryaluminum ingots. The environmental burdensof the production of only the lost aluminum,that is, 28 kg of primary metal, must be chargedto the product system under study, togetherwith the burdens of the recovery operations.The environmental burdens of the productionof 72 kg of primary aluminum must be chargedto the next user(s) of the 80 kg of recycledaluminum.

In many cases, scrap from different productsand different alloys is melted together in onefurnace batch, and alloying elements may beadded. If, for example, pure alloy scrap ismelted together with AlMg3 (EN-AW-5754)scrap, then this melt may be cast to AlMg1.5(EN-AW 5050) rolling ingots. In this case, onecould argue that the input material and theoutput material have different inherent proper-ties. However, not considering the value of thealloying elements, it can be shown that theAlMg1.5 rolling ingots have the same marketvalue as the unalloyed ingots and the AlMg3ingot that were the origin of the scrap.

If the market value analysis shows that themarket value of the aluminum ingots obtainedfrom recycling of the end-of-life product is thesame as the market value of primary aluminum,then a value correction is not necessary. In thiscase, the substitution can take place as in thecase of identical inherent properties, effectivelytreating the product system as a closed-loop one.

Substitution Method with Recycled Alu-minum as Input. ISO 14041 requires that allo-cation procedures must be uniformly applied tosimilar inputs and outputs of the system underconsideration. The rules on how to treat incom-ing recycled aluminum must correspond withthe methods for treating recycled metal that

leaves the system. If the substitution method isapplied, there is no need to consider the incom-ing portion of recycled aluminum, because onlythe metal loss during the complete life cycle ofthe product is considered. This means that if100 kg of aluminum ingots enter the system,and 100 kg of recycled aluminum with the sameinherent properties leave the system, then theenvironmental loads associated with the inputmetal and those associated with the outputmetal should be considered to be the same, evenif the products from which the input metal isrecycled are not known.

For a product, if 100 kg of a secondary rawmaterial, usually recycled from a mixture ofproduction scrap and postconsumer scrap, withthe same value of primary aluminum is usedand no recycling of this product happens, thenthe environmental loads of the production of100 kg of primary aluminum must be charged tothe product. For example, with 100 kg of alu-minum used for a product system, it may con-sist of 50 kg of primary aluminum and 50 kg ofrecycled aluminum with the same inherentproperties as the primary aluminum. If 80 kg ofrecycled aluminum ingots (with the same inher-ent properties as the primary aluminum) resultfrom recycling, including remelting (20 kg ofaluminum are lost), the environmental burdensof only the production of the lost aluminum,that is, 20 kg of primary metal, must be chargedto the product system under study, together withthe burdens of the recovery operations. Theseenvironmental burdens are valid no matter therecycled content of the product system, as longas no value correction is necessary. The envi-ronmental burdens of the production of 80 kg ofprimary aluminum must be charged to the nextuser(s) of the 80 kg of recycled aluminum.

Value-Corrected Substitution Methodwith Recycled Aluminum as Input. If thevalue-corrected substitution method is appliedfor recycled metal at the output side, then thevalue of the incoming recycled metal must beconsidered as well. For a product, if 100 kg of asecondary raw material, usually recycled from amixture of production scrap and old scrap, witha value of 90% of the value of primary alu-minum is used and no recycling of this producthappens, then the environmental loads of theproduction of 90 kg of primary aluminum mustbe charged to the product.

If end-of-life recycling can be considered,then a (possibly value-corrected) substitution


can occur, as described previously. In this case,the mass of the lost aluminum on the value levelof primary aluminum, that is, the value-correctedmass of the input aluminum minus the value-corrected mass of the output aluminum, mustbe calculated, and the environmental burdensof the production of this quantity of primaryaluminum must be charged to the product systemunder study.

For example, 100 kg of aluminum required fora product system consist of 40 kg of primary alu-minum and 60 kg of recycled aluminum ingotswith 90% of the value of primary aluminum. Ofthe primary aluminum (100 kg primary) used asthe source for the recycled aluminum, 20 kg arelost (for example, littered or landfilled), while80 kg of recycled aluminum ingots have 90%of the value of primary aluminum as a resultfrom recycling, including remelting. The netaluminum loss, based on the value level of pri-mary aluminum, is calculated as:

• Value-corrected mass of input aluminumminus value-corrected mass of outputaluminum, both on the value level ofprimary aluminum

• Value-corrected mass of input aluminum is40 kg + (60 kg � 0.9) = 94 kg aluminumingots of the value level of primary ingots

• Value-corrected mass of output aluminum is80 kg � 0.9 = 72 kg aluminum ingots of thevalue level of primary ingots

The net aluminum loss, based on the valuelevel of primary aluminum, is 94 kg – 72 kg =22 kg.

The environmental burdens of the productionof the lost aluminum, that is, 22 kg of primarymetal, must be charged to the product systemunder study, together with the burdens of the re-covery operations. The environmental loads ofthe production of 6 kg of primary aluminum hadalready been charged to the producer of the60 kg of recycled aluminum. The environmentalburdens of the production of 72 kg of primaryaluminum must be charged to the next user(s) ofthe 80 kg of recycled aluminum.

Long Lifetime Products. The substitutionmethod for recycling may apply for any lifetimeof a product, not only for aluminum cans butalso for aluminum in buildings. Aluminumproducts often have extended lifetimes becauseof their high corrosion resistance, often furtherenhanced by specific measures of corrosion pro-tection. Such products should not be mistreated

in LCA studies by omitting recycling credits, asdescribed earlier in the section “System Expan-sion and Substitution” in this chapter.

Any uncertainty about recycling rates andrecycling techniques for long-life aluminumproducts is not sufficient to refuse recyclingcredits. Rather, it must be dealt with by apply-ing different recycling scenarios in the form ofsensitivity analyses, which should include thestate-of-the art recycling technique for theproduct under study.

Moreover, it may be necessary in the inter-pretation phase of comparative studies to con-sider the temporal aspects of the environmentalimpacts of the different materials, for example,current greenhouse gas (GHG) emissions com-pared with GHG emissions in 50 years.

The Recycled Material Content Approach.There have been cases of LCA studies in whichthe recycling of a product is disregarded and norecycling credits are given on the output side,even in the case of closed-loop recycling, thatis, when the recycled aluminum is used for thesame product from which it has been recovered.Recycling credits are only given on the inputside if the aluminum product contains a certainamount of recycled aluminum (recycled mate-rial content approach).

There have also been cases of LCA studies inwhich a different form of the recycled materialcontent approach is applied. Here, recyclingcredits according to the substitution principleare given only in the case where the closed-looprecycling approach can be applied. In the caseof open-loop recycling, no or only limited recy-cling credits are given. In addition, recyclingcredits are given on the input side if the alu-minum product contains a certain amount ofrecycled aluminum.

According to ISO 14041, allocation princi-ples and procedures also apply to recycling situ-ations. The recycled material content approachis not mentioned in this standard as a method toavoid allocation. This approach can only beused in specific cases where the omission ofspecific outputs of recycled material can be jus-tified according to well-defined cut-off criteria,for example, if recycled concrete is given free ofcharge for road construction.

The high value of aluminum recycled fromend-of-life products can be demonstrated by themarket price of recycled aluminum ingots,which is identical or close to the price of pri-mary aluminum. In this context, methods such


as the recycled material content approach thatdo not consider the end-of-life recycling or eventhe value of recycled material should be avoided.If, under certain conditions, the recycled con-tent method can be justified, the LCA studyshould at least include a sensitivity analysiswith other methods that better address the end-of-life recycling of aluminum products.

For example, an LCA study of different win-dow frame materials (aluminum, aluminum/wood, steel, copper, polyvinyl chloride) wasperformed by K. Richter et al. (Ref 4.7) in 1996.The study used the recycled material contentapproach and was based on a window measuring1650 by 1300 mm (5.4 by 4.3 ft) of two wingsand a use time of 30 years. For the aluminumframe that had a mass of 40 kg, two differentrecycling scenarios were assumed:

• A realistic scenario, assuming a recycledaluminum content of 35% (alternative 1)

• A future scenario, assuming a recycled alu-minum content of 85% (alternative 2)

The study did not show general disadvan-tages of the two aluminum alternativescompared to other window frame materials.Nevertheless, the greenhouse potential ofalternative 1 was significantly higher (385 kgCO2 equivalents); only alternative 2 showedresults equivalent to the other materials (172 kgCO2 equivalents).

From this study, the conclusion could bedrawn that the aluminum industry should in-crease the recycled aluminum content, for ex-ample, by buying recycled aluminum on themarket. Nevertheless, because recycled alu-minum is an expensive commodity available in arelatively constant quantity, any increase of therecycled aluminum content in a window framewould lead to a reduction of the recycled alu-minum content in other products, which, accord-ingly, need more primary metal. Finally, newscrap and old scrap often have the same inherentproperties and therefore are mixed together,which makes the determination of the recycledmetal content impossible.

After the publication of the comparative win-dow frame study, ISO 14041 was published.Therefore, an additional study on aluminumwindow frames was performed by K. Richterand F. Werner, where the ISO rules were prop-erly applied according to the guidance given inthis paper (Ref 4.8). Because the recyclingprocesses of window frames had been cut off in

the first study, a main part of the second studywas to identify the unit processes for the end-of-life recycling of window frames and to definethe value of recycled aluminum when leavingthe system boundaries. This work included visitsto various recycling plants and the acquisitionof data from such locations.

The recycling studies showed that twodifferent types of aluminum window framesexisted:

• Type 1 window frames: Contain componentsof zinc and brass or small steel parts thatcould not be separated from the aluminumfraction

• Type 2 window frames: Contain only alu-minum, steel, and plastic/rubber parts that canbe easily separated by shredding, magneticsorting, and eddy-current sorting

The greenhouse potential of these windowframes has been calculated as:

• 235 kg CO2 equivalents for type 1 windows • 179 kg CO2 equivalents for type 2 windows

One recommendation of the new study was toapply design for recycling in such a way that type2 window frames should be built in a mannerwhere aluminum can be easily separated fromother materials and contamination of the metalwith impurities can be avoided.

REFERENCES

4.1. Board Resolution, International AluminumInstitute, 1998

4.2. “Life Cycle Assessment of Aluminum: In-ventory Data for the Worldwide PrimaryAluminum Industry,” International Alu-minum Institute, March 2003

4.3. “Aluminum Applications and Society: LifeCycle Inventory of the Worldwide Alu-minum Industry with Regard to EnergyConsumption and Emissions of GreenhouseGases, Paper 1—Automotive,” InternationalAluminum Institute, May 2000

4.4 “Environmental Profile Report for theEuropean Aluminum Industry,” April 2000

4.5 “North American Aluminum AssociationLCI Report,” 1998

4.6 F. Werner, “Treatment of Aluminum Re-cycling in LCA; Development and Evalua-tion of the Value-Corrected Substitution


Procedure Applied to Window Frames,”Research and Work Report 115/47, SwissFederal Laboratories for Materials Testingand Research (EMPA), Duebendorf, 2002

4.7 K. Richter, T. Künninger, and K. Brunner,“Ökologische Bewertung von Fensterkon-

struktionen,” Swiss Central Office forWindows on Facade Construction, 1996

4.8 F. Werner and K. Richter, Economic Alloca-tion in LCA: A Case Study about AluminumWindow Frames, Int. J. LCA, Vol 5 (No. 2),2000, p 79

ALUMINUM HAS ADVANTAGES to helptake on the challenge of sustainable developmentand environmental impact assessments, whichcan only be judged in terms of a full cradle-to-grave life-cycle analysis. It is a material that canbe profitably recycled without loss of quality foruse by future generations, and the use of alu-minum has the potential to help control CO2emissions from weight reduction of transporta-tion vehicles. For industry, the concern for theenvironment involves reducing the consumptionof resources. This includes minimizing the use ofenergy, materials, water, and land while enhanc-ing recycling and product durability. It alsomeans reducing the impact on nature by mini-mizing air emissions, water discharges, wastedisposal, and the dispersion of toxic substances.

To achieve sustainable development goals,the aluminum industry needs to focus on thefollowing:

• Continuing the reduction program for green-house gas emissions

• Increasing the recycling rate of used alu-minum

• Minimizing the use of hazardous or toxicmaterials and finding alternatives to landfilldisposal for spent potlining and bauxiteresidues

• Achieving further energy-efficiency improve-ments

• Establishing consistently understood andapplied sustainability metrics that can beused as leading indicators of progress

• Developing performance indicators on thesocial aspects of sustainability

• Encouraging an industry-wide approach tosustainability through awareness training,benchmarking, monitoring, and reporting

With these objectives, aluminum producerscontinue to develop models of sustainabilitythat are being integrated into daily companyoperational management practice.

The aluminum industry has a number of ele-ments in its current environment, health, andsafety program that reflect a pioneering role inthe field of global industrial reporting on sustain-able development. For example, the InternationalAluminum Institute (IAI) has been monitoringand reporting on a number of widely used sus-tainable development indicators, such as energyuse, greenhouse gas emissions, and safety per-formance. Since 1997, the IAI has collectedcomprehensive benchmarking data on safetyperformance in the global aluminum industry,with shared information on accidents, nearmisses, and their causes. Such benchmarkingdata are a driving force for continual improve-ment. The IAI is also currently working on:

• Generating global scrap and recycling statis-tics to help identify the scope for increasedrecycling

• Publishing annual global surveys of the alu-minum industry’s energy consumption

• Publishing an annual global survey of theindustry’s perfluorocarbon (PFC) emissionsand PFC reduction performance, backed upby benchmarking graphs and training semi-nars, to encourage plants to match the bestperformers

• Conducting regular surveys of land use forbauxite mining and rehabilitation, and re-porting the results every 5 years

CHAPTER 5

Sustainable Development for theAluminum Industry*

*Adapted from Ref 5.1 and 5.2




• Collecting annual accident statistics globallyand issuing global benchmarking reports toencourage participating companies to matchthe best performance of their colleagues

• Collecting data on the industry’s globaleconomic contribution

This chapter summarizes some IAI reportson sustainability progress from surveysthrough the year 2000 (Ref 5.1) and a 2005 up-date (Ref 5.2). Survey data on sustainabilityperformance are collected from IAI membercompanies, collectively responsible for over70% of global primary aluminum productionand approximately 20% of recycled metal pro-duction. The 2005 survey data were collectedfrom:

• Seventy-five smelters representing almost 18million tonnes of primary aluminum produc-tion, a 21% increase on 2004 coverage andequivalent to 55% of total global aluminumproduction

• Twenty-four refineries representing over 36million tonnes of smelter-grade alumina, a41% increase on 2004 coverage and equiva-lent to approximately 60% of total globalalumina production

• Eleven mines representing over 81 milliontonnes of bauxite, a 13% increase on 2004coverage and equivalent to approximately50% of total global bauxite production

Recycling

Aluminum is a metal that can be recycledand reused with as little as 5% of the energythat would be required to produce it from rawmaterials. In effect, the world’s increasingstock of aluminum acts like an energy resourcebank, over time delivering more and morepractical use and value from the energy em-bodied in the primary production for futuregenerations by conserving energy and othernatural resources. For example, recycling ofpostconsumer aluminum now saves an esti-mated 84 million tonnes of greenhouse gasemissions per year, equivalent to the annualemissions from 15 million cars. Since its in-ception, the recycling of postconsumer alu-minum scrap has already avoided over 1 bil-lion metric tonnes of CO2 emissions.

The contribution of scrap metal recovery hassteadily increased in terms of global output ofaluminum metal (Fig. 5.1). The percentage ofglobal output of aluminum metal from scrap re-covery has also increased from 17% in 1960 to33% today (2007) and is projected to rise to al-most 40% by 2020 (Fig. 5.2). Of an estimatedtotal of over 700 million tonnes of aluminumproduced in the world since commercial manu-facture began, approximately three-quarters (be-tween 400 and 500 million tonnes) is still in pro-ductive use. As an “energy bank,” the aluminumin use accounts for almost 50,000 petajoules of

35

30

25

20

15

10

5

0

Mill

ion

tonn

es

World aluminum usage 1950-1999

1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 1999

Primary production

Scrap recovery

Total usage

Fig. 5.1 Scrap and primary components of aluminum usage. Source: Ref 5.1

Chapter 5: Sustainable Development for the Aluminum Industry / 93

energy. This is higher than the current combinedannual energy demand of Africa and LatinAmerica and is equivalent to the annual totalelectrical energy generated globally from coal. Ifthis metal is recycled, the banked energy is aresource not just once but repeatedly, and overfuture generations. If landfilled, this valuableresource is potentially lost forever.

Global aluminum recycling rates are high,approximately 90% for transport and approxi-mately 60% for beverage cans. In Europe, alu-minum has high recycling rates, ranging from41% in beverage cans to 85% in constructionand 95% in transportation. In Japan, the recy-cling rate for cans was 79%, and Brazilachieved 78%. A kilo of used cans is worthmore to collectors than 15 kilos of plastic or 10kilos of paper. Americans recycled 62.6 billionaluminum cans in 2000, for a beverage canrecycling rate of 62%. For many groups, usedaluminum cans turn into new-found money,thanks to aluminum can industry initiatives.Initiatives include the American AluminumAssociation’s partnership with Habitat forHumanity, called “Aluminum Cans BuildHabitat for Humanity Homes,” designed toboost public interest in aluminum can recy-cling while helping volunteers and familiesbuild homes.

The recycling rate for end-of-life aluminumdepends on the product sector, the lifetime ofeach product, and society’s commitment tocollect aluminum. Each application requires itsown recycling solutions. Just over 15 milliontonnes of recycled aluminum were producedworldwide in 2004, which met 33% of theglobal demand for aluminum. Of the almost 7million tonnes of aluminum recycled fromend-of-life products, 28% came from packaging,44% from transport, 7% from building, and21% from other products.

Aluminum enjoys a high recycling rate of85% in the building industry. The global indus-try is keen to increase collection rates and isworking with producers of building applicationsto enable even more efficient collection of scrapfrom demolished buildings. In 2004, Delft Uni-versity of Technology conducted a study into thealuminum content of, and collection rates from,demolished buildings in six European countriesand found that the average collection rate foraluminum was close to 96%. The transport sec-tor also has high rates of recycling, currentlyapproximately 90% globally, because disman-tlers and recyclers recognize the high intrinsicvalue of end-of-life aluminum products.

The aluminum industry is working withmanufacturers to enable easier dismantling of

80

70

60

50

40

30

20

10

0

Tota

l alu

min

um p

rodu

ctio

n, m

illio

n to

nnes

/yea

r

1950 1960 1970 1980 1990 2000 2010 2020

90

80

70

60

50

40

30

20

10

0

Rat

io o

f prim

ary

and

recy

cled

alu

min

um p

rodu

ctio

n, %

Primary production

Total metal production

Recycling production

Fig. 5.2 Percentage of recycling component in aluminum production. Source: Ref 5.2


aluminum automotive components to improvethe sorting and recovery of scrap aluminum.The Japanese Aluminum Association, forexample, is currently undertaking a study ofadvanced separation techniques to identifyways in which aluminum components fromShinkansen (“Bullet Train”) carriages can beefficiently separated from other materials, thusincreasing the recycling rate and the quality ofscrap collected. Electromagnetic sorting is apowerful tool, currently used in many regionsof the world, that separates even the smallestaluminum shards from waste material. Apply-ing heat to end-of-life vehicle parts to separatethe lacquer from the aluminum also facilitatesrecycling.

Can collection is approximately 60% globally,and in some countries, the collection rate isalready above 80%. Sweden and Switzerlandcollect 86 and 88% of their aluminum beveragecans, respectively. Sweden’s success lies in adeposit/refund system, whereas in Switzerland, avoluntary prepaid recycling charge covers thecosts of collection. In Japan, a collection rate of92% for used beverage cans is achieved with avoluntary system. In Brazil, recycling is notmandatory, but every region has a recycling mar-ket that facilitates the collection and transporta-tion of end-of-life products. This has encouragedcommunities, supermarkets, condominiums, shop-ping centers, and clubs to collect. As a result, in2005, 96% of cans were recycled in Brazil, andthis is considered the world’s highest recyclingrate for used beverage cans.

As a tool, the IAI has designed a mass flowmodel (critically reviewed by Yale and DelftTechnical Universities) to track aluminumthroughout its life cycle, from mining to productuse to recycling. The main objective for creatingthe model is to identify present and future recy-cling flows and the scope for further recycling.The model traces the flow of aluminum from1888 to the present (2007) along the completevalue chain of aluminum processing: bauxitemining, alumina refining, aluminum and alu-minum ingot production, fabrication (rolling,extrusion, and casting), manufacturing (produc-tion and assembly of finished products), use, andrecycling. New scrap is generated immediatelyduring the production and processing stages, nothaving yet reached the use phase. Old scrap isgenerated when an aluminum-containing prod-uct reaches its end of life and is collected forrecycling. To calculate the amount of aluminumstill in productive use and leaving the use stage,

a product residence time model is applied. Here,the average lifetime of each of the main productsin which aluminum is used and the historicaltonnage of aluminum in those products are con-sidered. The results for 2004 are shown in amass flow diagram (Fig. 5.3).

At present, approximately one-third of worlddemand is met from recycled metal. Much betterglobal scrap statistics are required to improvethe use of available scrap material and ensurethat valuable metal is not going to the landfill.The IAI and associations from the recycling in-dustry have established a Global AluminumRecycling Committee to monitor scrap flowsand recycling rates. A study is also beingconducted to evaluate the contents in landfills ofvarious regions.

Perfluorocarbon Emissions

Two PFC compounds (CF4 and C2F6) are pro-duced during the electrolysis process from briefupset conditions known as anode effects. ThesePFCs have long atmospheric lifetimes and arepotent greenhouse gases, with a greenhouse gaswarming potential of 6500 and 9200 times thatof CO2 for CF4 and C2F6, respectively. From an-nual IAI surveys of PFC emissions, there hasbeen a steady trend in the reduction of PFCemissions (Fig. 5.4a), and a voluntary industryobjective is to obtain an 80% reduction in PFCgreenhouse gas emissions per tonne of alu-minum produced for the industry as a whole by2010 versus 1990 levels. The trend and 2010target of CO2 equivalents for greenhouse gasemissions per tonne of aluminum is plotted inFig. 5.4(b).

Data on PFC emissions from the annual IAIsurveys allow benchmarking and comparisons,so individual plants can monitor their perform-ance with other nonidentified plants using thesame technology. Those companies respondingto requests for anode effect data increased from61% of the world’s primary aluminum produc-tion in 1990 to 66% in 2000. In 1997, the in-dustrial processes of the primary aluminum in-dustry emitted 110 million tonnes of CO2equivalents, of which approximately 50 milliontonnes (45%) originated from the two PFCcompounds. Preliminary results of the IAI sur-veys for the years 1998, 1999, and 2000 for IAImember companies reporting anode effect dataindicate a continuing declining trend with PFCemissions. The data indicate a reduction of


Fig. 5.3 Mass flow model of aluminum throughout its life cycle, from mining to product use and recycling. Values in millions ofmetric tonnes. Values may not add up, due to rounding. Production stocks not shown. 1, aluminum in skimmings; 2, scrap

generated by foundries, rolling mills, and extruders. Most is internal scrap and not taken into account in statistics: 3, such as powder,paste, and deoxidation aluminum (metal property is lost); 4, area of current research to identify final aluminum destination (reuse, re-cycling, or landfilling); 5, calculated. Includes, depending on the ore, between 30 and 50% alumina; 6, calculated. Includes, on aglobal average, 52% Al; 7, scrap generated during the production of finished products from semis; 8, landfilled, dissipated into otherrecycling streams, incinerated, incinerated with energy recovery. Source: Ref 5.2

Bauxite3 153.7

Alumina3 57.3

Material flow Metal flow

Primaryaluminum used

30.2Remelted

aluminum 31.6

Bauxite residues 56.7and water 39.7 Metal losses 1.3

Tradednew

scrap1 1.3

Fabricatorscrap2

16.5 Traded newscrap3 7.7

Oldscrap7.4

Otherapplications3

1.1

Ingots 61.8

Fabricated andfinishedproducts (input)60.5

Finishedproducts(output)36.3

Buildings31%

Transport 28%

Net Addition 2004: 21.1

Packaging 1%

Other 11% Engineering

and cable 29%

Total productsstored in usesince 1888538.5

Not recycled in 20043 3.4 Under investigation4 3.3

Recycledaluminum

15.1

Automotive16%

specific CF4 emission rates by 60% between1990 and 2000, even though total primary pro-duction increased from 20 to over 30 milliontonnes per annum in the same period. Over thesame period, the specific emission rate for C2F6was reduced by 62%. Between 1990 and 2005,PFC emissions were reduced by 76% per tonneof aluminum produced, and the industry objec-tive is to continue with this beneficial trend.

Worldwide estimates of PFC emissions havebeen based on an extrapolation of the IAI surveydata, using knowledge of the reduction technolo-gies at those facilities that did not report anodeeffect data. These results show that, while world-wide aluminum production has increased by ap-proximately 24% since 1990, there has still beenan overall reduction in the total annual emissionsof PFCs. These reductions in PFC emissions tothe atmosphere are estimated to amount to over34 million tonnes as carbon dioxide equivalents,a reduction of approximately 39% from the 1990baseline for worldwide PFC emissions. This isone of the few examples where the global emis-sions of a greenhouse gas from an industrysector are actually in decline.

The surveys also show that smelters in thedeveloping world, which often use state-of-the-art technology, are performing as well as, if not

better than, some plants in Europe or NorthAmerica. Voluntary agreements between gov-ernment and industry have played a significantrole in encouraging this reduction in PFC emis-sions in many countries, such as Australia,Bahrain, Brazil, Canada, France, Germany,New Zealand, Norway, and the United King-dom. Together, they represent approximately50% of world production.

Fluoride Emissions

For decades, fluoride emissions (as gases andparticulates) were considered to be the singlemost important pollutant from aluminumsmelters. Depending on the local conditions,fluorides could have a serious environmentalimpact on the local flora and fauna. Fluoridesaccumulate in vegetation and can cause damageto coniferous trees. They also accumulate in theteeth and bones of ruminants eating fluoride-contaminated forage.

Figure 5.5(a) shows the development in fluo-ride reductions through three “generations” ofaluminum smelters. Plants with modern controlsystems to remove and recycle the fluorides donot generate local concerns. Optimal fume


collection from the electrolytic cells, coupledwith specific workplace-related training of theemployees, should also lead to further improve-ment in the future. Figure 5.6 illustrates thereduction of atmospheric fluoride emissions inNorwegian smelters from 1960 to 2000, despitethe increase in production from 185,000 tonnesper year in 1960 to 1,030,000 tonnes in 2000.

In Russia over the last decade, specific energyconsumption decreased by 11%, and hydrogenfluoride and sulfurous anhydride emissions werereduced by 45 and 59%, respectively. In the caseof the majority of plants around the world that

use modern control equipment with efficient gascollection and wet/dry scrubbing systems, thedamage from fluoride emissions to surroundingflora and fauna has largely been eliminated. Flu-oride control systems are operated as closed-loop systems with no residual wastes.

Data collected from facilities representing83% of IAI member company production indi-cate a reduction in total fluoride emissions(gaseous and particulate) of over 50% per tonneof aluminum produced between 1990 and 2005.A voluntary objective is to obtain a minimum ofa 33% reduction in fluoride emissions by IAI

Fig. 5.4 Global emission of greenhouse gases from electrolysis reduction. (a) Percentage (from 1990 baseline) of specific perfluoro-carbon (PFC) emissions. Source: Ref 5.1 (b) CO2 equivalents including PFC components of CO2 equivalents. Source: Ref 5.2

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

Global specific PFC emissions

(a)

Spe

cific

CF

4 em

issi

ons,

per

cent

of 1

990

base

line

100

90

80

70

60

50

40

30

20

10

0

1990

Tonn

es C

O2

equi

vale

nt p

er to

nne

alum

inum

pro

duce

d 5

4

3

2

1

0

4.4

1.0

0.9

1990

1992

1994

1996

1998

2000

2002

2004

2006

2008

2010

(b)


member companies per tonne of aluminum pro-duced by 2010 versus 1990.

Energy Efficiency

Energy represents a large part (~25%) of thecosts associated with primary aluminum pro-duction. The two main components of energyconsumption for primary aluminum occur in theproduction of metallurgical alumina and elec-trolysis. Trends in global energy consumptionrates for these two components of primary pro-duction are given in Fig 5.7 and 5.8.

The aluminum industry has a long tradition ofimproving energy efficiency. Figure 5.9 shows asteady reduction in electrical power used in pri-mary aluminum production from 1899 to 1999.The average energy consumption and subse-quent emissions per tonne of production fell by70% over those one hundred years, due to re-search and continuing process developments. Inthe 1950s, it took, on average, approximately 21kWh (kilowatt-hours) to make a kilogram ofaluminum from alumina.

The three main energy sources in aluminarefining are fuel combustion, electricity produc-tion, and energy use in lime production. These

kg F

per

tonn

e A

l

16

14

12

10

8

6

4

2

0

1940 1950 1960 1970 1980 1990 2000

1940 1950 1960 1970 1980 1990 2000

Development of fluoride emissions

Consumption of fluoride

(a)

(b)

kg F

per

tonn

e A

l

100

90

80

70

60

50

40

30

20

10

0

Fig. 5.5 Trends in fluoride emissions and consumption. (a) Fluoride emissions from first, second, and third generations of alu-minum electrolysis (smelting) technology. (b) Fluoride consumption due to losses to air, water, spent potlinings, and waste

at a representative smelter. Source: Ref 5.1


three sources are monitored together as the totalenergy used in alumina production. The averageenergy used to produce 1 tonne of metallurgicalalumina decreased by 5% between 1990 and2004 (Fig. 5.10). The IAI is developing a quan-titative voluntary objective for alumina refiningenergy efficiency. The objective of IAI membercompanies is to seek reduction in greenhousegas emissions from the production of aluminaper tonne of alumina produced.

Smelters in the 1990s used one-third less elec-tricity per tonne than equivalent plants in the1950s. In 1999, it took one of the newestsmelters just 13 kWh. The average electrical en-ergy required to smelt 1 tonne of aluminum fromalumina was cut by 5% between 1990 and 2005,mainly through investment in modern, more effi-cient technologies. New smelters generally usethe best available technologies, no matter wherein the world. The trend of improving energy

kg fl

uorid

e pe

r to

nne

of a

lum

inum

5.4

4.4

2.2

0.75

0.46

1960 1970 1980 1990 2000

1960 1970 1980 1990 2000

1000

tonn

e pe

r ye

ar

1

2.2

1.4

0.6

0.4

(a)

(b)

Fig. 5.6 Reduction of atmospheric fluoride emissions in Norwegian smelters from 1960 to 2000. (a) Specific emissions. (b) Totalemissions. Source: Ref 5.1


40

38

36

34

32

30

Mill

ions

of t

onne

s

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

(a)

(b)

Global metallurgical alumina production

14

13

12

11

Meg

ajou

les

per

kg a

lum

ina Global energy for metallurgical

alumina production

Fig. 5.7 Metallurgical alumina production. (a) Global tonnage. (b) Global energy consumption rates

efficiency continues today (2007), and the indus-try objective is a 10% reduction in averagesmelting energy use by IAI member companiesper tonne of aluminum produced by 2010 versus1990. For benchmarking to encourage improvedperformance, the IAI carries out an annualenergy consumption survey covering 70% of theworld’s primary production facilities.

Aluminum in Transportation

The global use of aluminum in the automo-tive sector increased from 2.5 million tonnes in1991 to nearly 4.5 million tonnes in 1999. Forthe first decade of this millennium, the use of

aluminum in cars is predicted to double becauseof more cars worldwide and more aluminum incars. The U.S. automobile fleet contained 116 kgof aluminum per vehicle in 2000 and is ex-pected to contain 159 kg by 2010. The cumula-tive impact of using aluminum in this way willyield greenhouse gas savings of 180 milliontonnes per year by 2010.

In seeking to reduce the environmentalimpacts caused by transport, aluminum cancontribute to improved performance fromweight reduction. Current estimates show thatglobally there will be a 35% increase in CO2emissions from all vehicles by the year 2020.An increased use of aluminum could reducethis figure to 28% and thus help toward making


30

25

20

15

10

0

Mill

ions

of t

onne

s

17

16

15

14

kWh

per

kg A

l

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

(a)

(b)

Global primary production

Global electric energy for electrolysis

Fig. 5.8 Primary aluminum production. (a) Global tonnage. (b) Global electric energy rates for electrolysis. Source: Ref 5.1

the transportation sector more sustainable. Thisuse of aluminum for automobiles in 1999 alonehad the potential, over the lifespan of the vehi-cles, to reduce overall greenhouse gas (GHG)emissions by 90 million tonnes, assuming thatall this aluminum was used to replace densermaterials.

Annually, aluminum shipments for use intransport are being monitored by industry inorder to track aluminum use for reducing GHGemissions from road, rail, and sea transport. Dif-ferent life-cycle assessments have shown that1 kg of aluminum in a car body, replacing 2 kgof steel in a conventional car body, saves, duringthe lifetime of the car, approximately 20 kg ofGHG emissions (in CO2 equivalents). In addi-tion, if, for example, 1000 kg of GHG is saved bythe use of less gasoline in a lightweight vehicle,

this also means a reduction of other potentialenvironmental impacts, including:

• 15,800 MJ of crude oil resources; 933 kg ofwater

• 1.8 kg ethylene equivalents of ozone-forminghydrocarbons

• A savings of 2.1 kg SO2 equivalents of acidi-fication potential

Greater opportunities for reduction of GHGemissions also are possible for vehicles such asbuses and long-haul trucks (which operate overmuch longer distances during their lifetimes,often five times longer than that of a typicalpassenger car). For example, the use of 1 kg ofaluminum replacing 1.5 kg of steel in a typicalbus or truck reduces the GHG emission by ap-proximately 40 kg over its lifetime (i.e., twice


kWh

per

tonn

e of

alu

min

um

60

50

40

30

20

10

0

1899 1909 1919 1929 1939 1949 1959 1969 1979 1989 1999

Energy consumption for primary aluminum production

Fig. 5.9 Energy consumption of primary aluminum from 1899 to 1999. Source: Ref 5.1

13.0

12.8

12.6

12.4

12.2

12.0

11.8

11.6

11.4

11.2

11.0

GJ

per

tonn

e al

umin

um p

rodu

ced

1991

1990

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

Fig. 5.10 Energy use in alumina production. Source: Ref 5.2

as much as for a car). Finally, a railroad train,during its lifetime, runs a distance of between30 and 50 times the distance of a typical car.The use of 1 kg of aluminum replacing 1.6 kgof steel in a railway car reduces the GHG emis-sions by more than 200 kg (i.e., ten times asmuch as for a car). Reduction in the “deadweight” of ships through the use of lighter-weight aluminum can also result in energysavings and GHG reductions.

Natural Resources

Land Use. The use of mineral resources forthe production of aluminum is quite modest ascompared with many other materials. Between 4and 5 tonnes of bauxite are required to produce1 tonne of alumina, and 2 tonnes of alumina arerequired to produce 1 tonne of aluminum metal.Globally, bauxite mining disturbs approximately25 km2 a year, an area equivalent in size to


one-third of Manhattan Island, NY. Every year,approximately 20 km2 is rehabilitated. Thearea of land rehabilitated as a percentage ofland mined since operations began, in cur-rently operating mines, is 70%.

There are numerous bauxite deposits, mainlyin tropical and subtropical regions, that are gen-erally extracted by open cast mining from strata,typically some 4 to 6 m (13 to 20 ft) thick undera shallow covering of topsoil and vegetation. Inmost cases, the topsoil is removed and stored.Of the four main climate groups, the distribu-tion of bauxite mining locations in 1998 was:temperate, 0.5%; Mediterranean, 39%; tropical48%; and subtropical, 13%. The vegetationtypes disturbed each year are mainly 76%forests, 19% agriculture and pasture, and 2%shrubland. Postmining land use shows 70%being returned to native forest, 3% to commer-cial forest, 17% to pasture and agriculture, and7% used for urban and industrial development,housing, and recreational purposes.

The IAI’s first Bauxite Mine RehabilitationSurvey was prepared in 1991, covering 65% oftotal world bauxite production. It described theoverall impact of bauxite mining on the environ-ment and the rehabilitation programs in placeat 18 mining locations. A second survey wascarried out in 1998 and covered 27 mining loca-tions covering 72% of the total world bauxiteproduction. The results of both of these surveysindicate that the bauxite mining industry isbroadly in harmony with the approach set out inthe United Nations Environment Programme(UNEP) “Guidelines for the Environmental Man-agement of Alumina Production.”

The area disturbed for mine developmenthas increased by some 14% over the pastdecade. In 1991, it was approximately 1400hectares, and by 1998, it had risen to 1591hectares. The output of bauxite ore per hectarehad increased from 52 thousand tonnes perhectare to 56.5 thousand tonnes per hectare in1998. Of the total area mined per year in 1998,80% was identified as wildlife habitat, and 175hectares were tropical rainforest. The majority

of mines have introduced measures such aswildlife reserves and corridors, wildlife re-search, and monitoring programs. In Jamaica,the industry has created orchid sanctuaries tohelp preserve the local orchid species. The 577hectares of mined area were identified as pos-sessing important fauna species; of these, 527hectares, or 91%, will be restored to wildlifehabitat by the planting of suitable vegetation.Companies representing 60% of global bauxiteproduction had their own plant nurseries forthe purposes of reforestation and revegetation.

Rehabilitation plans are in place at mostmines (25 out of 27, or 90%, as compared to10 out of 18, or 55%, in 1991). A very highproportion of mine managements have long-term plans for the mine areas that will leave aself-sustaining ecosystem in place when allmining operations have been completed. Theindustry continues to seek increases in the pro-portion of bauxite mining land rehabilitatedannually.

Water Use. Besides energy productionfrom hydroelectric facilities, water in the alu-minum industry is used for cooling in certainprocesses and sanitary purposes. The local pro-grams to reduce water consumption vary fromlocation to location, depending on the ease ofaccess to water and on water consumptioncosts. A voluntary objective of IAI membercompanies is to reduce their fresh water con-sumption per tonne of aluminum and aluminaproduced. The IAI member companies willconcentrate efforts to minimize fresh waterconsumption where there are limited availablefresh water resources.

REFERENCES

5.1. “The Aluminum Industry’s SustainabilityDevelopment Report,” International Alu-minum Institute, 2003

5.2. “Aluminum for Future Generations—Sus-tainability Update 2006,” InternationalAluminum Institute, 2006

STATISTICS ON THE RESOURCE FLOWSrequired to produce primary and recycled alu-minum are incomplete on a global basis. Inaddition, there has not been a consensus or quan-titative estimate of future resource flows relatedto aluminum production or the potential avail-ability of less resource-intense end-of-life alu-minum (recycled) metal to meet ever-increasingconsumer and developing nation needs.

A model was developed by Alcoa Inc. (Ref6.1) to provide a quantitative understanding ofhistoric and today’s (year 1950 through year2003) worldwide aluminum mass flows andsystems losses. In addition, current and futureresource requirements and life-cycle inventoryflows were estimated by coupling these globalaluminum mass flows with global, average life-cycle inventory intensity data developed from amajority of producers in a report (Ref 6.2) bythe International Aluminum Institute (IAI) (seealso Chapter 4, “Life-Cycle Assessment of Alu-minum: Inventory Data for the WorldwidePrimary Aluminum Industry” in this book). Themodel was also developed to provide quantita-tive scenario-development capability to deter-mine the positive impact of enhanced recycling,lower resource-intense production, and productusage scenarios. The model and key resultsinformation were developed to be shared with

global aluminum industry technical experts,executives, and external stakeholders to betterunderstand potential paths to more globally sus-tainable aluminum.

Modeling

Modeling of historic and current flows wasbuilt around aluminum product net shipmentsstatistical data provided by governments, suchas the U.S. Geological Survey (Ref 6.3–6.5), orregional aluminum associations, such as theEuropean Aluminum Association, AustralianAluminum Council, the Japan Aluminum Asso-ciation, or the North American Aluminum Asso-ciation (Ref 6.6). The data were gathered,starting in year 1950, into a comprehensivespreadsheet model by year, by region (EuropeanUnion, South America, China, etc.), and inaccordance with the following customer (mar-ket) segmentation:

• Building and construction • Transportation—auto and light truck, aero-

space, and other (heavy trucks, trains, etc.) • Packaging—aluminum containers and other

packaging (foil, etc.) • Machinery and equipment • Electrical—cable and other electrical • Consumer durables • Other (such as aluminum used for propellant

or steel deoxidation)

From the product net shipments, the modelestimates both internal (runaround) aluminum

CHAPTER 6

Material Flow Modeling of Aluminum for Sustainability*Kenneth J. Martchek, Alcoa Inc., Alcoa Corporate Center

*Adapted with permission of International Journal of LifeCycle Assessment from the following article by Kenneth J.Martchek: “Modelling More Sustainable Aluminium: CaseStudy,” International Journal of Life Cycle Assessment,Volume 11 (No. 4), 2006.




facility recycle flows and new (prompt, fabrica-tor) customer recycle flow amounts, based onestimation of average use, yield, and melt lossrates identified in the literature and reviewedand agreed on by a subteam of global aluminumtechnical experts.

This subteam of experts was commissionedby the IAI Global Aluminum Recycling Com-mittee (GARC). Postconsumer (end of productlife) aluminum flows are estimated from prod-uct net shipments in previous years, estimatedproduct lifetimes (worldwide by market byyear), scrap recollection rates (by region bymarket by year), and recovery factors againbased on industry statistics, published litera-ture, and review and agreement by the IAIGARC committee. An illustration of some ofthese data on scrap recovery rates and meltingrecovery efficiency is provided in Table 6.1.

Modeling of future aluminum and resourceflows is based on literature (Ref 6.7, 6.8) andexpert projections of life-cycle inventory inten-sity rates (Ref 6.2, 6.9) and aluminum productshipments by market (currently with a weighted-average compounded annual growth rate of2.5% per year). The availability of recycle flowsto meet these market demands is based onprojected use, yield, melt loss, recovery rates,postconsumer recycling rates, and anticipatedfuture product lifetimes. Primary aluminum pro-duction is then calculated to determine the

market demand for additional primary capacityand resulting resource requirements.

Validity Checks. There are two validitychecks in the model:

• Comparison of estimated postconsumer andnew scrap by year with published values byyear. (This check is informational becausepublished values for global recycled metalare considered to be incomplete.)

• Comparison of the estimated market demandfor primary aluminum by year with pub-lished primary production by year

Figure. 6.1 shows the aluminum productionthat is estimated by the model to have beenrequired for the years 1970 to 2003, based onproduct net shipments less postconsumer and

Met

ric to

ns×1

03

45,000

40,000

35,000

30,000

25,000

20,000

15,000

10,000

5,000

01970 1980 1990 2000 2010 2020

Estimated/predicted

Reported

Fig. 6.1 Estimated versus reported worldwide primary production

Table 6.1 Example of global average worldwidecollection (recycle) rates and melting recoveriesby market

Collection, % Collection, %Market 1990 2000 Melting recovery, %

Buildings 69 70 96Autos and 75 75 96light trucks

Aerospace 76 75 96Other transport 76 75 96Containers 61 59 85 (net of 4 cycles/yr)Packaging––foil 13 16 30Machinery 40 44 96Electrical cable 45 51 96Electrical other 30 33 96Consumer durables 20 21 96

Chapter 6: Material Flow Modeling of Aluminum for Sustainability / 105

new recycle flows and system losses. This isshown to be in fairly good agreement withreported worldwide primary aluminum produc-tion for the years 1970 to 2003. Required pri-mary aluminum production is then projected tothe year 2020, with assumptions about eachmarket segment growth rates, postconsumerscrap collection rates, and anticipated recover-ies based on latest trends.

Key Results

The model assessment of global aluminummass flows is shown schematically in Fig. 6.2.

The size (area) of the circles illustrates relativevolume of flows. In year 2003, recovered post-consumer and new customer recycled metalsupplied 33% of the global aluminum industry’sproduct net shipment supply.

An estimated 516 million metric tons ofaluminum, approximately 73% of all of thealuminum ever produced, is contained in cur-rent transportation, cable, and building productinventory (in service), as illustrated in Fig. 6.3.The model also projects future product inven-tory volumes by market segment to year 2020.

Additional key results also included systemlosses, such as aluminum packaging lost in

Values in millions of metric tons

Primaryaluminum27.4

Recycled

Metal losses1.4

Al inskimmings1.2

Ingots56.2

Internalscrap15.0

Newscrap6.9

Oldscrap7.0

Fabricated andmanufactured products55.0

Finishedproducts33.1

Oxidized inapplication0.8

Landfill

516.1Net addition: 200319.0

Underinvestigation

Total productsin use 2003

Fig. 6.2 Global aluminum mass flows for the year 2003

Met

ric to

ns×1

03

1,000,000

900,000

800,000

700,000

600,000

500,000

400,000

300,000

200,000

100,000

0

Other (e.g., destructive uses)Consumer durablesElectrical–otherElectrical–cableMachinery and equipmentPackaging–other (foil)Packaging–cansTransportation–otherTransportation–aerospaceTransportation–auto and light truckBuilding and construction

1990 1995 2000 2005 2010 2015 2020

Fig. 6.3 Worldwide aluminum product inventory by market


landfills or metal oxidized to aluminum oxidewhen used as a propellant or for deoxidizingsteel melts. The model also provides an assess-ment of past, current, and projected energy andemissions intensity of aluminum semifabricatedshipments, as illustrated in Fig. 6.4, based onthe latest assessment of average greenhouse gas(GHG) emissions intensity for aluminum proc-esses (Ref 6.9).

On average, worldwide aluminum productsare becoming less GHG intense on a per-ton-shipped basis due to two reasons:

• Increase in the percent recycled metal rela-tive to primary metal. Only 5% of the energyand GHG emissions is required to producealuminum ingot compared to primary (baux-ite/Al2O3/electrolysis) aluminum.

• Lower emissions from primary aluminumfacilities due to reductions in energy inten-sity and significant reductions in perfluoro-carbon emissions

The results of the model were initially sharedwith the IAI Board of Directors in May 2004.Model development, the GARC expert reviewand contribution, and results of global aluminumflow were described, including a quantitativeassessment of GHG emissions from globalindustry aluminum facilities today (2003) andprojected into the future as a guide to the indus-try’s contribution to climate change effects. Atthat time, the Board requested a “what-if” casescenario that later indicated that industry factoryand indirect emissions from purchased electricitycould be stabilized, despite significant industrygrowth by 2020, based on currently projectedrecycled metal flows and moving all of the globalindustry toward today’s (2003) global benchmarktechnologies and operating best practices.Furthermore, the model indicated that fuel effi-ciency and emissions savings due to additional

aluminum transportation products had thepotential to surpass the global industry’s pro-duction emissions by 2020.

In May 2005, additional model results wereshared with the IAI Board, and they added thefollowing voluntary objective to their list ofsustainable development quantitative goals inrecognition of the ecological and economic valueof enhanced recycling to reduce natural resourceconsumption and life-cycle inventory effects:

“The IAI has developed its SustainabilityMaterial Flow Model to identify futurerecycling flows. The model projects thatglobal recycled metal supply (back to theindustry) will double by 2020 from today’s(2005) level of 6.4 million metric tons. Thealuminum industry will report annually onits global recycling performance.”

The IAI also continues to develop and improvethe model, collect supporting life-cycle inventoryintensity data, and use the scenario capability toquantitatively assess current and future produc-tion paths and sustainable strategies.

REFERENCES

6.1. P.R. Bruggink and K.J. Martchek, World-wide Recycled Aluminum Supply and En-vironmental Impact Model, Light Metals,Proceedings of the 2004 Annual Meeting,March 14–18, 2004 (Charlotte, NC), A.T.Tabereaux, Ed., TMS (The Minerals, Met-als and Materials Society), p 907–911

6.2. “Life Cycle Assessment of Aluminum: In-ventory Data for the Worldwide PrimaryAluminium Industry,” International Alu-minum Institute, London, U.K., March2003; http://www.world-aluminum.org/iai/publications/documents/lca.pdf (accessedJuly 2007)

6.3. P.A. Plunkert, Aluminum: Minerals Year-book, U.S. Geological Survey, U.S. De-partment of the Interior, Reston, Virginia,2001, p 6.1–6.19; http://minerals.er.usgs.gov/minerals/pubs/commodity/aluminum/050401.pdf (accessed July 2007)

6.4. P.A. Plunkert, Aluminum: Minerals Yearbook, U.S. Geological Survey, U.S.Department of the Interior, Reston, Virginia,2002, p 5.1–5.18; http://minerals.er.usgs.gov/minerals/pubs/commodity/aluminum/alumimyb02r.pdf (accessed July 2007)

12

10

8

6

4

2

0

Mto

nne

CO

2-e/M

tonn

e A

l

1950 1990 2000 2010 2020

11.3

9.68.9

7.76.8

Fig. 6.4 Greenhouse gas emissions intensity of aluminumshipments

Chapter 6: Material Flow Modeling of Aluminum for Sustainability / 107

6.5. P.A. Plunkert, Aluminum: Minerals Year-book, U.S. Geological Survey, U.S. Depart-ment of the Interior, Reston, Virginia,2003, p 5.1–5.16; http://minerals.er.usgs.gov/minerals/pubs/commodity/aluminum/alumimyb03.pdf (accessed July 2007)

6.6. “Aluminum Statistical Review for 2002,”The Aluminum Association Inc., Washing-ton, D.C., 2003

6.7. G. Rombach, Future Availability of Alu-minum Scrap, Light Metals, Proceedings ofthe 2002 Annual Meeting and Exhibition,Feb 17–21, 2002 (Washington, D.C.),

W. Schneide, Ed., TMS (The Minerals,Metals and Materials Society), p 1011–1018

6.8. Economics of Aluminium, 8th ed., RoskillInformation Services, London, U.K.,2003

6.9. “Life Cycle Inventory of the WorldwideAluminium Industry with Regard to EnergyConsumption and Emissions of Green-house Gases,” International AluminumInstitute, London, U.K., March 2000; http://www.world-aluminum.org/iai/publications/documents/expanded_summary.pdf(accessed July 2007)

AMONG THE MORE IMPORTANT CHAR-ACTERISTICS of aluminum is the instanta-neous formation of a self-limiting tenaciousoxide film in liquid and solid states. The protec-tive oxide makes the metal impervious to degra-dation or weight loss, except in cases of galvaniccorrosion or specific chemical attack. For thisreason, most of the aluminum produced from thebeginning remains recoverable as a vital resourcein satisfying current and future industry metalsupply requirements.

Aluminum recycling became a distinct manu-facturing activity less than 20 years after thecommercialization of the Hall-Heroult process in1888, driven by high value and growing demand.In the early days of the developing aluminumindustry, primary producers attempted to maxi-mize new metal sales, reduce the unit price, anddevelop the alloys and processes that were essen-tial to become more competitive with existingmaterials. They were not interested in scrap recy-cling, leaving that activity to others, who in timedeveloped an independent secondary industry.

As both industries grew, their objectiveschanged. Primary producers began acquiring andconsuming scrap in their operations, and second-ary producers began producing more sophisti-cated end products, thus reducing the originallydistinctive differences between the two indus-tries. Similar developments occurred in the tech-nology arena.

Secondary producers were originally low-capital salvage operators. Some still operate in

that mode, whereas others have become moresophisticated enterprises using advanced tech-nology to maximize recoveries, reduce costs, andproduce competitive-quality remelt ingot and dif-ferentiated products. The latter group patternedits development after that of the primary industry,which has taken a more capital-intensive ap-proach to recycling challenges.

With the modern demands of energy conser-vation and environment, the aluminum industryis a leading proponent of global sustainabilityand strongly advocates the use of recycled metal.Aluminum can be recycled with large energyand emission savings and essentially withoutany loss in potential material properties. Further,it can be recycled repeatedly without significantdegradation, although there are oxidation lossesand some contamination by impurity elements.This recycling advantage is so dominant foraluminum that it has become a key factor in theindustry’s sustainability. In addition to signifi-cant energy savings, recycling requires relativelylow-capital-cost facilities when compared to thecapital required for primary production.

Recycling advantages explain the aluminumindustry’s proactivity in promoting and exploit-ing the recovery of aluminum scrap in all forms.Aluminum cans and other aluminum productsare the most valuable components of the munici-pal waste stream, often providing the only eco-nomic justification for separation processes.Legislation mandating minimum standards forend-of-life recycling promotes the increased useof highly recyclable aluminum in transportationapplications. Of course, recycling is not the onlyrelevant factor. The use of aluminum in trans-portation is a major factor in light-weighting ofvehicles, and lighter vehicles are more fuel effi-cient. It has been demonstrated that a 10%reduction in vehicle weight results in a 6 to 8%

CHAPTER 7

Recycling of Aluminum*

*Adapted from “Recycling of Aluminum” by Elwin L. Rooyand J.H.L. Van Linden in Properties and Selection: Non-ferrous Alloys and Special-Purpose Materials, Volume 2,ASM Handbook, 1990, and “Recyclable Aluminum RolledProducts” by John Green and Michael Skillingberg in LightMetal Age, Aug 2006, p 33. Adapted with permission ofLight Metal Age.




reduction in fuel use and thus in CO2 emissions.The growth in the use of aluminum in carsand light trucks, 81 lb/vehicle in 1973 to 319lb/vehicle in 2006, reflects recognition of theseessential advantages. Studies suggest that thecontribution of aluminum in improving fuel effi-ciency results in rates of reduction in emissionsof greenhouse gases (GHG) from the transporta-tion sector that exceed emissions from aluminumproduction. This has led to the claim that thealuminum industry will become GHG neutralby the year 2020 (Ref 7.1).

Recycled aluminum is a vital supply compo-nent. It minimizes the need for imported metalin the U.S. domestic market, saves ~95% of theenergy, reduces ~95% of the emissions associ-ated with the production of metal from ore, andconserves raw materials. These benefits affectvirtually all aluminum products but are espe-cially pertinent to rolled products that includesheet, plate, and foil, which constitute 45.7% ofall U.S. and 39.9% of North American aluminumindustry shipments (Ref 7.2). However, a domi-nant issue in the use of recycled metal in theproduction of rolled aluminum products is thecontrol of impurities, especially iron and siliconthat increase modestly with each remelting cycle.Alloys used in the production of engineeredcastings typically contain much higher siliconconcentrations, so that mixing scrap can sub-stantially and negatively affect the chemistry ofthe wrought alloy stream. Iron impurities gener-ally increase as results of extraneous contamina-tion and refractory reactions.

In spite of the issues with impurity control,there is absolutely no doubt that the advantagesinherent in recycling aluminum are enormous.The “energy bank” concept and the opportunityto regain almost all the embedded energy in theproduct again and again when the metal is recy-cled will drive further growth in the secondaryindustry. In effect, the world’s increasing stock ofaluminum acts like an “energy resource bank,”over time delivering more and more practicaluse and value from the energy embodied in theprimary production for future generations byconserving energy and other natural resources.For example, recycling of postconsumer alu-minum now saves an estimated 84 million tonnesof GHG emissions per year, equivalent to theannual emissions from 15 million cars. Since itsinception, the recycling of postconsumeraluminum scrap has already avoided over 1billion metric tonnes of CO2 emissions. Forthese reasons, it is clear that recyclable rolled

products will continue to be building blockstoward sustained manufacturing viability.

Industry and Recyling Trends

Several generalizations from the early 1990sconcerning the U.S. and world aluminum indus-tries continue to be relevant today:

• The aluminum industry today is truly inter-national: Primary aluminum is produced invirtually every global region, and the metalproduced competes in global markets. Alu-minum alloy scrap is now traded internation-ally as well.

• Domestic U.S. primary aluminum produc-tion will not expand beyond the capacity ofexisting smelting facilities: Energy costs,labor rates, and environmental standards inthe United States suggest that there will besubstantial decreases in primary output in theabsence of any forseeable new, more eco-nomical smelting technologies.

• New risks and uncertainties: The Europeanand, to a lesser but significant extent, theU.S. aluminum industries face new risks anduncertainties caused by the growing geo-graphic separation of primary productionfrom major fabricating facilities and markets.

• Aluminum is a U.S.-dollar-based commodity:Exchange rate fluctuations represent a majorcomplication in stabilizing prices and regu-lating international competition and metalsupply.

• The world aluminum production capacity willexpand: This will occur in countries with lowenergy costs, such as Canada, Venezuela,Brazil, Australia, and parts of the MiddleEast.

• Recycling continues to have increasing impor-tance: For the United States, and ultimatelyfor the rest of the aluminum-consumingworld, recycling and resource recovery willplay an increasingly important strategic rolein ensuring a reliable and economical metalsupply.

• The United States will import aluminum: Onthe basis of the best assumption, the UnitedStates will become an importing nation asaluminum requirements exceed domesticsmelting capacity. Most imports will consistof unalloyed smelter ingot for remelting, cast-ing, and fabrication in North American facili-ties. Some products will be imported but only

Chapter 7: Recycling of Aluminum / 111

for specialty applications with unique and/orcost advantages over domestic products.

• World competition for scrap units will inten-sify: This competition will be based on therelative costs and availability of primary alu-minum and scrap.

The U.S. sources of aluminum have undergoneconsiderable change. Primary production fromthe reduction of alumina dominated the U.S.domestic metal supply for 112 of 115 years ofcommercial production since the first year ofrecorded statistics in 1893 (when some 91,000metric tons of metal were produced). In recentyears, the domestic metal supply reflects increas-ing imports and higher levels of secondaryrecovery. Recycled aluminum and imports haveactually exceeded primary production in theUnited States since 2002. These changes are

shown in Table 7.1 (Ref 7.2) and graphically inFig. 7.1 (Ref 7.2).

Overall, the United States reached the highestprimary production at 4653 thousand metric tonsin 1980; the largest U.S. metal supply on recordoccurred in 1999 with a combined value of11,154 thousand metric tons. Year 2000 showsthe marked reduction in primary metal supplyresulting from the effects of energy costs andavailability and competitively priced metal in theinternational market. These conditions nowappear to be permanent and have resulted in theshutdown of much primary capacity, particularlyin the Pacific Northwest. Restarting this capacitywill require access to low-cost power, unantici-pated changes in the supply of unalloyed metalfrom foreign sources, or government actioninvolving national security issues.

Overall aluminum production and the usecycle are strongly linked for the United Statesand Canada, and the combined U.S. and Cana-dian production of primary and secondaryaluminum is shown in Table 7.2 (Ref 7.3). Pri-mary production is still nearly 50% of the totalNorth American metal supply, and the portionsupplied from recycled metal drops to a littleover 30%. Still, the long-term trend has been anincrease in the importance of recycled and im-ported material. Accordingly, from an energy-efficiency standpoint, it is logical to increase the

Table 7.1 U.S. metal supplyMetric Primary Secondary

Year tons x103 production,% Imports,% recovered,%

1960 2406 78.0 7.5 16.51970 4950 72.9 8.8 18.31980 6833 68.1 8.9 23.01990 7863 51.5 18.1 30.41999 11,154 33.9 33.0 33.12000 10,699 34.2 33.5 32.22002 9579 28.5 40.8 30.72004 10,112 24.9 45.1 30.0

Met

ric to

ns×1

03

12,000

10,000

8,000

6,000

4,000

2,000

1944

1942

1946

1948

1950

1952

1954

1956

1958

1960

1962

1964

1966

1968

1970

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

2000

2002

2004

1996

1998

Imports Secondary recovery Primary

Fig. 7.1 U.S. aluminum supply. Source: Ref 7.2


amount of recycled aluminum in the metal sup-ply reservoir to the maximum extent possible.

Recycling Trends in the 1980s (Adaptedfrom Ref 7.4). The objective of all collectionactivities is the conversion of scrap forms toproducts having the highest commercial value.One trend in the 1980s was the consumption ofscrap in secondary fabricating facilities. A num-ber of extruders, foundries, and minimill opera-tions produce billet, castings, and common alloysheet products directly from scrap. The die cast-ing industry relies heavily on secondary compo-sitions for the production of automotive andother parts, and the larger die casters have, inmany cases, expanded metal supply throughdirect scrap purchases.

A number of industry segments are in compe-tition for the available aluminum scrap, althoughnot necessarily the same types of scrap. Asshown in Fig. 7.2, primary producers experi-enced the largest increase in scrap consumptionduring the decades of 1970 to 1990. The sporadicdata available for the period prior to 1940 sug-gest that recycling of aluminum grew steadily toapproximately 15% of total shipments. WorldWar II disrupted the pattern drastically, but thestockpile reduction in the years following thewar reduced the recycling rate to the prewarlevel (Fig. 7.3).

Current Trends (Adapted from Ref 7.5).The contribution of scrap metal recovery hassteadily increased in terms of global output ofaluminum metal (Fig. 7.4). The percentage ofglobal output of aluminum metal from scraprecovery has also increased from 17% in 1960to 33% today and is projected to rise to almost40% by 2020 (Fig. 7.5) (Ref 7.6). Of an estim-ated total of over 700 million tonnes of alu-minum produced in the world since commercialmanufacture began, approximately three-quarters(between 400 and 500 million tones) is still inproductive use. As an “energy bank,” the alu-minum in use accounts for almost 50,000 peta-joules of energy. This is higher than the current

Table 7.2 North American metal supplyMetric Primary Secondary

Year tons production,% Imports,% recovered,%

1960 2978 84.6 1.9 13.5(a)1970 5650 81.1 2.4 16.6(a)1980 7511 76.2 2.0 21.91990 8562 65.6 5.5 28.91999 11,765 52.4 15.0 32.52000 11,438 52.8 15.7 31.52002 10,320 52.5 17.4 30.22004 10,639 48.0 21.5 30.5

(a) U.S. + estimate of Canadian secondary. Source: Ref 7.3

U.S

. con

sum

ptio

n, to

nnes

×103

U.S

. con

sum

ptio

n, to

ns×1

03

2180

1910

1640

1360

1090

820

545

270

0

2400

2100

1800

1500

1200

900

600

300

0

Chemical producers

Foundries

Fabricators

Primary producers

Secondary smelters

1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986

Year

Fig. 7.2 U.S. aluminum scrap consumption by type of company for the years 1972 to 1986. Source: U.S. Bureau of Mines


Am

ount

of a

lum

inum

rec

ycle

d, %

of s

hipm

ents 40

30

20

10

0

1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Year

Fig. 7.3 Aluminum recycling trends in the United States. The percentage of shipments recycled is only now approaching the peakexperienced during World War II

Mill

ion

tonn

es

35

30

25

20

15

10

5

01950 1955 1960 1965 1970 1975 1980 1985 1990 1995 1999

Primary production Scrap recovery total usage

World aluminum usage 1950-1999

Fig. 7.4 Scrap and primary components of aluminum world usage. Source: Ref 7.5

Tota

l alu

min

um p

rodu

ctio

n,m

illio

n to

nnes

/yea

r

80

70

60

50

40

30

20

10

0

Rat

io o

f prim

ary

and

recy

cled

alum

inum

pro

duct

ion,

%

90

80

70

60

50

40

30

20

10

01950 1960 1970 1980 1990 2000 2010 2020

% from primary production

Total metal production

% recycling

Fig. 7.5 Percentage of recycling component in aluminumproduction. Source: Ref 7.6

combined annual energy demand of Africa andLatin America and is equivalent to the annualtotal electrical energy generated globally fromcoal. If this metal is recycled, the banked energyis a resource not just once but repeatedly andover future generations. If landfilled, this valu-able resource is potentially lost forever.

Global aluminum recycling rates are high,approximately 90% for transport and approxi-mately 60% for beverage cans. In Europe, alu-minum has high recycling rates, ranging from41% in beverage cans to 85% in construction and95% in transportation. In Japan, the recyclingrate for cans was 79%, and Brazil achieved 78%.


A kilo of used cans is worth more to collectorsthan 15 kilos of plastic or 10 kilos of paper.Americans recycled 62.6 billion aluminum cansin 2000, for a beverage can recycling rate of62%. For many groups, used aluminum cansturn into new-found money, thanks to aluminumcan industry initiatives. Initiatives include theAmerican Aluminum Association’s partnershipwith Habitat for Humanity, called “AluminumCans Build Habitat for Humanity Homes,”designed to boost public interest in aluminumcan recycling while helping volunteers and fami-lies build homes.

The recycling rate for end-of-life aluminumdepends on the product sector, the lifetime ofeach product, and society’s commitment tocollect aluminum. Each application requires itsown recycling solutions. Just over 15 milliontonnes of recycled aluminum were produced in2004 worldwide, which met 33% of the globaldemand for aluminum. Of the almost 7 milliontonnes of aluminum recycled from end-of-lifeproducts, 28% came from packaging, 44% fromtransport, 7% from building, and 21% fromother products.

Aluminum enjoys a high recycling rate of 85%in the building industry. The global industry iskeen to increase collection rates and is workingwith producers of building applications to enableeven more efficient collection of scrap from de-molished buildings. In 2004, Delft University ofTechnology conducted a study into the aluminumcontent of, and collection rates from, demolishedbuildings in six European countries, which foundthat the average collection rate for aluminumwas close to 96%. The transport sector has highrates of recycling, currently approximately 90%globally, because dismantlers and recyclers rec-ognize the high intrinsic value of end-of-lifealuminum products.

Can collection is approximately 60% globally,and in some countries, the collection rate isalready above 80%. Sweden and Switzerlandcollect 86 and 88% of their aluminum beveragecans, respectively. Sweden’s success lies in adeposit/refund system, whereas in Switzerland,a voluntary prepaid recycling charge covers thecosts of collection. In Japan, a collection ratefor used beverage cans of 92% is achieved witha voluntary system. In Brazil, recycling is notmandatory, but every region has a recyclingmarket that facilitates the collection and trans-portation of end-of-life products. This hasencouraged communities, supermarkets, condo-miniums, shopping centers, and clubs to collect.

As a result, in 2005, 96% of cans were recycledin Brazil, and this is considered the world’shighest recycling rate for used beverage cans.

Recyclability of Aluminum

Once produced, aluminum can be considereda permanent resource for recycling, preferablyinto similar products. It is highly recyclable, dueto the following characteristics:

• It is resistant to corrosion under most environ-ment conditions and thus retains a high levelof metal value after use, exposure, or storage.

• The energy required to remelt aluminum isonly 5% of the energy required for its primaryproduction.

• The versatility of aluminum alloying hasresulted in a large number of commercialcompositions, many of which were designedto accommodate impurity contamination.

The objective of recycling is to produce asalable commercial aluminum alloy product.Currently, more than 300 compositions coveringwrought and cast alloys are registered with theAluminum Association. Many of these alloys aredesigned to tolerate the variations in composi-tion and ranges in impurity contents that may beexperienced in the recovery of scrap.

In the early decades of the 20th century, theoutput of the secondary aluminum industry waslargely tied to the consumption of castings bythe automotive industry. As shown in Fig. 7.6,the recycling of aluminum has increasedsteadily from 1950 through the 1980s, despiterecessions and energy crises. This increase is

Qua

ntity

of a

lum

inum

, met

ric to

ns×1

06

Qua

ntity

of a

lum

inum

, lb×

109

7.2

6.3

5.4

4.5

3.6

2.7

1.8

0.9

0

16

14

12

10

8

6

4

2

01940 1950 1960 1970 1980 1990

Total

Domesticprime

Recycling

Import

Year

Fig. 7.6 U.S. aluminum supply distribution for the years1940 to 1990. Source: Aluminum Association


the result of growth in the automotive market aswell as the development of new and significantrecycling applications, such as the consumptionof used beverage cans (UBCs) in the manufac-ture of sheet for new cans.

In recent years, environmental concerns havecontributed to an increased awareness of theimportance of scrap recycling. Today (2007),the recyclability of aluminum is a major advan-tage to the aluminum industry in materials com-petitions for major product markets. The currentdriving forces for aluminum recycling are:

• Regulatory actions taken by governmentagencies to encourage resource conservation,energy conservation, and waste reductionthrough mandatory segregation and depositprograms

• Consumer sensitivity to environmental is-sues and the solid-waste crisis

• Competitive pressures from other materials • Economic advantages based on the relative

value and availability of aluminum scrap

Energy Savings. The recognized energy savi-ngs are well known for aluminum recyclingwhen compared with the energy consumed inprimary aluminum production. For example, theactual ratio of primary total energy to recycledtotal energy for used beverage cans is 28.5:1(Ref 7.7). A definitive study of energy savingsby recycling aluminum was obtained in the late1990s in a comprehensive project mandated bythe U.S. Council for Automotive Research underthe Partnership for a New Generation of Vehiclesprogram. At the request of the automotive com-panies, the aluminum industry was challenged toproduce a rigorous life-cycle assessment (seeChapter 3, “Life-Cycle Inventory Analysis of theNorth American Aluminum Industry”). This as-sessment represents a comprehensive evaluationof the aluminum industry operations involving15 separate unit processes (refining, smelting,hot rolling, shredding and decoating, andsecondary ingot casting are examples of unitprocesses) located in 213 plants throughoutNorth and South America, Africa, Australia,Europe, and the Caribbean.

Specifically, the life-cycle assessment demon-strated that the production of primary aluminum,when all the electrical generation, transmissionlosses, and transportation are accounted for,requires ~45 kWh and emits ~12 kg of CO2 perkilogram, whereas recycled aluminum onlyrequires ~2.8 kWh of energy (~5%) and onlyemits ~0.6 kg (~5%) of CO2 for each kilogram

of metal. In this sense, the intrinsic energy of themetal is recovered when the material is recycled,hence the “energy bank” concept.

The Recycling Loop

Some amount of aluminum recycling hasalways occurred. Producers remelted internalscrap, and foundries, the first importantcustomers of the emerging aluminum industry,remelted gates and risers and scrap castings. Intime, the recovery of new and old scrap becamea function of secondary smelters, whose prin-cipal customers were the aluminum castingsindustry. The emergence of the aluminum bev-erage can in 1965 and the initiation of “buy-back” programs by primary producers, whowere the source of rigid-container sheet (RCS),created the first significant public awareness ofthe importance and value of product recycling.The importance of aluminum recyclability wasnot lost on the producers of RCS nor on canmakers, who realized the economic and envi-ronmental advantages over competing containermaterials. With the success and magnitude ofthe aluminum can market, recycling efforts bymajor producers quickly spawned a new phasein recycling based on closed-loop processingand the development of radically new tech-nologies specifically designed to maximizerecoveries from light-gage coated scrap. Thelessons of can recycling were quickly adaptedto other scrap forms. The primary industry thathad been single-mindedly devoted for decadesto the promotion and sale of primary metal andproducts realized the comprehensive advantagesof maximizing recycling as a means of reducingcosts, serving the environment, and conservingraw materials and energy.

The reclamation of aluminum scrap is a com-plex interactive process involving collectioncenters, primary producers, secondary smelters,metal processors, and consumers. Figure 7.7depicts the flow of metal originating in primarysmelting operations through various recyclingactivities. The initial reprocessing of scrap takesplace in the facilities of primary producers. In-process scrap, generated both in casting andfabricating, is reprocessed by melting and re-casting. Increasingly, primary producers arepurchasing scrap to supplement primary metalsupply; an example of such activity is the pur-chase or toll conversion of UBCs by primary


producers engaged in the production of rigidcontainer stock.

Scrap incurred in the processing or fabricationof semifabricated aluminum products representsan additional source of recyclable aluminum.Traditionally, this form of new scrap has beenreturned to the supplier for recycling, or it hasbeen disposed of through sale on the basis ofcompetitive bidding by metal traders, primaryproducers, and secondary smelters.

Finished aluminum products, which includesuch items as consumer durable and nondurablegoods; automotive, aerospace, and military prod-ucts; machinery; miscellaneous transportationparts; and building and construction materials,have finite lives (see Chapter 6, “Material FlowModeling of Aluminum for Sustainability,” inthis book for additional life-cycle discussion).In time, discarded aluminum becomes availablefor collection and recovery. So-called old scrap,metal product that has been discarded afteruse, can be segregated into classifications thatfacilitate recycling and recovery.

Scrap specifications have been developedthat allow the convenient definition of scraptypes for resale and subsequent reprocessing.

Those developed by the Institute of ScrapRecycling Industries are in broad use (Ref 7.8).These specifications, however, are, for themost part, physical descriptions of scrapcategories useful to dealers. Chemistry specifi-cations and limits on harmful impurities arenot defined, and treatment of extraneous con-taminants is inconsistent. More comprehensivescrap specifications are being developed by theindustry.

Technological Aspects of Aluminum Recycling

The widely accepted range of aluminum recy-cling includes scrap collection and preparation,remelting, refining, and the upgrading of moltenmetal to a ready-to-cast condition. Multiplealloys and product forms are consumed in thegeneration of new product.

Molten metal treatment processes that reducethe concentration of reactive elements, such asmagnesium, sodium, and calcium, to complywith specification limits, remove dissolvedhydrogen, and filter entrained oxides and other

Imports Imports Exports Imports Imports

Imports

Exports

Imports ExportsImports Exports

Bauxite Aluminaproduction

Primaryaluminum(smelter)

Primary aluminum

(molten or ingot)

UBCprocessing

facility

Foundriesfabricators, andchemical plants

Secondaryaluminum

smelter

Dross generated

Dross generated

Secondaryaluminum

(molten or ingot)

Supply of scrap (old and new)

Scrap recyclingindustry

Totalaluminum

supply

Productionof wroughtproducts

New scrapgenerated

Productionof cast

products

New scrapgenerated

Old scraprecovered

End use

Containers andpackaging, 28%

Building andconstruction, 19%

Transportation, 19%

Electrical, 9%

Consumerdurables, 8%

Machinery andequipment, 5%

Other, 4%

Exports, 8%

Fig. 7.7 Flow diagram for aluminum in the United States, showing the role of recycling in the industry. Scrap recycling (lower left)includes scrap collectors, processors, dealers and brokers, sweat furnace operators, and dross reclaimers. UBC, used

beverage can. Source: U.S. Bureau of Mines


nonmetallics play an extremely important role inenabling the reuse of often heavily contaminatedscrap. Environmental restrictions have resultedin the development of process alternatives toreactive gas fluxing for these purposes. Variousmethods, including magnetic separation, areused to reduce the introduction of extraneouscontamination in the solid state.

After the invention of the resealable taphole,the coal-, natural gas-, or oil-fired reverberatory,or open hearth furnace, evolved as the predomi-nant remelting method. There are two reasonsfor the long-term success of the direct-firingmethod in which the combustion chamber andthe solid charge are combined into one openhearth. First, aluminum oxidation progressesslowly due to the protective, flexible thin oxidefilm that forms instantaneously when a freshaluminum surface is exposed to any oxygen-containing atmosphere. Second, the surface-to-volume ratio of scrap particles was generallysmall, because the scrap usually consisted oflarge sections or parts and heavy-gage sheet.Therefore, the penalty for direct exposure to themelting environment was generally acceptable.

The open hearth furnace has been the mainstayof the scrap-remelting business for over 50 years,but as the application of aluminum becamemore widespread and diverse, the need foreffective and efficient recycling technologygrew rapidly. The corresponding growth anddiversification of the scrap-recycling industryhave been driven by several factors, includingincreased production requirements, alloy devel-opment, scarcity of energy and resources, andincreased availability of mixed scrap.

Increased Production Requirements. Large,modern casting stations may use up to 250,000lb per production cycle. This requires not only ahigh melt rate in large melting furnaces but alsoa minimal charge time. The present rectangularopen hearth furnace construction is practicallylimited in size to approximately a 30 m2 (320ft2) bath area. Cylindrical furnace constructionallows for hearth areas of up to 75 m2 (800 ft2)and features a removable lid. This configurationallows for very fast, well-distributed overheadcharging with preloaded dump buckets.

Alloys. Magnesium, in many important com-positions that include castings, extrusions,sheet, and plate, results in higher oxidation ratesand melt losses. Among more important appli-cations for alloys containing significant magne-sium concentrations are light-gage sheet forautomotive and container products, aerospace

and other high-strength applications, andmarine applications. The prevalence of thistype of scrap has stimulated the developmentand use of methods that minimize the exposureof scrap to the furnace atmosphere; these meth-ods include continuous melting, rapid ingestionsystems, induction melting, and, more recently,isothermal melting.

Energy and Resources. The energy crisis of1973 highlighted the enormous energy advan-tage of using scrap rather than primary metal andexposed the corresponding need to minimizemelt losses. Primary producers began process-ing more-difficult-to-melt scrap using newlydeveloped technologies, forcing secondary pro-ducers to seek new scrap sources, including less-desirable scrap types, which are often processedin rotary salt furnaces previously employed inthe reclamation of skim and dross.

Concerted efforts have been made to incre-mentally improve the energy efficiency of rever-beratory furnaces through adaptation of heatrecuperation, preheating combustion air andsolid charge components, mechanically or elec-tromagnetically stirring the melt, modifyingburner designs, and incorporating sensors andautomated controls to optimize furnace opera-tion. New melting processes that target energyefficiency and reduced melt loss have beendeveloped that include flotation and isothermalmelting.

Mixed Scrap. Since the early 1960s, thevolume of scrap from discarded life-cycle prod-ucts such as automobiles, home constructionmaterials, trailers, and household goods, ofteninseparably mixed or joined with other metals,has steadily increased. Substantial amounts ofotherwise lost aluminum are recovered fromsuch scrap by sweat melting. This is a selectiveprocess that involves melting the scrap in asloped hearth that exploits the melting tempera-ture differences of dissimilar metals and which,during operation, melts, drains, and collectsmetals based on melting points.

Alloy Integrity. In addition to preservation andmelt loss reduction, a key factor in successful re-cycling is the maintenance of alloy integrity. Theideal way is closed-loop recycling in which scrapis recycled into identical or similar products. Ex-amples are the conversion of UBCs into new-cansheet and the conversion of auto body sheet intonew auto body sheet. Mixed scrap that requiresextensive treatment to chemically react and re-move magnesium or that requires the addition andadjustment of alloying elements is usual but far


less economically attractive. Accordingly, greatefforts are made at all times to keep in-processand purchased scrap identified and segregated.

The greatest challenge for the recycling com-munity is finding the most economical way toseparate and prepare scrap for melting so that itcan be used in the least-degraded form with theleast number of postmelt treatments for alloyor quality adjustments. Because UBC, automo-tive, and municipal scrap loops are importantand represent different challenges, technologicaldevelopments in preparation and melting forthese streams is further described.

Process Developments for Remelting

Although melt loss had become the majorcost factor in ingot production, it was, in fact,the soaring cost of energy during the 1973energy crisis that triggered the search for moreefficient remelt processes. This effort alsosought to develop less labor-intensive andhigher-productivity solutions to scrap melting.It was recognized that the open hearth furnacewas designed for remelting bulky scrap, whichnot only requires extensive charging time butalso creates large amounts of skim. Skim,comprising oxidation products and metal, actsas an insulating blanket between radiation fromsuperheated refractories, which is the dominantheat-transfer mechanism, and the melt, severelyreducing thermal efficiency. It became clear thatthe commitment to recycling necessitated afundamental review of processing methods.

Reverberatory furnaces, into which high-capacity natural gas burners introduce thermalenergy, are the mainstay of primary and second-ary industries and large engineered castingfoundries for melting, alloying, processing, andholding. These furnaces range in capacity from10,000 to more than 250,000 lb. Heat transferoccurs predominantly by radiation. The rate ofheat transfer is a function of melt surface areaand is maximized by design to enable the furnaceto operate at high melt rates, typically 60 lb/h-ft2.As a consequence, the depth of metal in the fur-nace hearth may approximate only 76 cm (30in.). These design considerations impose signifi-cant process disadvantages, some of which arenoted as follows:

• Aluminum in the molten state is highly reac-tive and readily oxidizes when exposed to theatmosphere and products of combustion. Thethickness of the oxide layer depends on alloy,

temperatures, and time but represents at alltimes an insulating barrier to the absorptionof radiant energy by the melt. Charging andmelting of scrap increases oxide formation asa function of surface-to-volume ratio.

• Heat concentrates at the melt surface, accel-erating oxidation, increasing oxide barrierthickness, and forming more harmful oxidespecies, such as magnesium aluminate spinelin magnesium-containing alloys that grow atgeometric rates with increasing temperature.

• Stirring of the melt to accelerate alloying,assure thermal and chemical homogeneity,and remove the oxide-metal skim layer isusually accomplished by industrial truck- orrail-mounted tools. Their use requires thatthe furnace design includes large accessdoors that, when opened, result in substan-tial heat losses and, even when closed, mayresult in pressure and heat losses that reduceenergy efficiency. Too-frequent stirringexposes nascent surfaces to oxidation, result-ing in greater metal losses and increaseddissolved hydrogen levels.

Charges may include solid and moltenaluminum, internal and purchased scrap, metal-lurgical metals, and master alloys. These areintroduced to the hearth, after which the furnacedoors or lids are closed, and firing for meltingbegins. Energy from combustion is absorbed byexposed refractories, from which radiant heat isreflected to the charge.

After melting, additional metal treatmentsteps are typically required. These metal treat-ments become more problematic with highproportions of scrap use and increase as unsegra-gated or mixed scrap are included in the furnacecharge. The melt must be stirred to assure com-plete solution and uniform distribution of alloy-ing elements. Adjustments in metal chemistrymay be required to meet specification require-ments mandating additional alloying, stirring,and sampling. In many cases, the melt is treatedwith solid or gaseous fluxes. Finally, the on-composition melt may be held for varying peri-ods at controlled temperatures before transferringto other furnaces or casting stations. Additionalstirring and/or skimming may be required beforetransfer. With each iteration, more energy isrequired, additional oxide losses are experienced,and dissolved hydrogen levels increase.

Because of the dramatic negative effects ofshredded can and light scrap melted through con-ventional methods in open hearth furnaces, it wasreasoned that low-bulk-density scrap must not be


exposed directly to the furnace atmosphere andthat scrap must be submerged quickly to opti-mize efficiencies and reduce melt losses.

Continuous Melting. Accordingly, continu-ous melting processes, distinct from typical batchoperations, were developed specifically to avoidthe conditions that contribute to increased meltlosses. Reverberatory furnaces used extensivelyfor scrap melting are often modified to incorpo-rate an exposed charging or side bay into whichscrap can be introduced without opening the fur-nace and without directly exposing the charge tothe combustion atmosphere. The rapid immer-sion of scrap is aided by tamping devices calledwell walkers, or by pumping/recirculation/inges-tion methods, and oxidation of the melt is re-duced by introducing a nitrogen gas stream.While these simple side-bay configurationsapproach the objectives, further process develop-ments provide more advanced configurations foroptimizing melting efficiencies and reducingmelt losses. The advantage of the latter processesis that skim can also be efficiently captured orseparated and removed outside the hearth.Steady-state conditions involving heat input thatbalances melting energy requirements and heatlosses can be established and maintained.

Three different methods (Ref 7.9–7.11) basedon pump design variations have been devel-oped. These designs create the scrap-ingestingvortex either in the pump bay itself or in an optimized adjacent charge bay (Fig. 7.8). Although these methods differ in such areas asproduction capacity, scrap size tolerance, andhardware simplicity, the skim generation for aspecific scrap type is equally low for all threemethods. Several variations of this principlehave been reported by other researchers (Ref7.12–7.14), suggesting that the method is ap-pealing to others in the industry who are alsostruggling to reduce melt losses.

Reverberatory furnace energy efficienciesvary with the extent to which process modifica-tions have been adopted. Incremental improve-ments result from recuperation of exhaust gasheat, cogeneration, preheating of charge com-ponents and combustion air, the use of induc-tion or mechanical circulation/stirring, and byautomated systems that monitor and control keyoperating parameters. All are commerciallyavailable technologies. Numerous burner varia-tions are in use, and other fuel options, such asoxy-fuel and oxygen-enriched burner operation,have been explored. An approximation of thebest energy efficiency obtained by reverberatory

furnaces in good condition with the mostadvanced features and operated competently is45%. More typically, energy efficiencies acrossthe industry are in the range of 20 to 30%.

Alternatives to Reverberatory Furnaces.The challenges of efficient low-oxidation-lossmelting have also resulted in the use of differentfurnace types.

Coreless and channel induction furnaces areroutinely used to melt fine scrap charged con-tinuously by automated conveying systems.Machining chips, delaquered UBCs, and shred-ded light-gage scrap are immediately drawn intothe melt by convective forces. Unfortunately,this means that oxides that may otherwise con-tribute to and remain in the skim layer are insteadentrained, requiring effective downstream meltprocessing for their removal.

New process developments now underindustrial-scale evaluation promise quantumadvances in both efficiency and melt loss control.They include flotation melting, based on theprinciples of cupola operation, and isothermalmelting, employing high-watt-density/high-efficiency electric resistance immersion heaters.Both developments have been supported by theU.S. Department of Energy.

Developing Scrap Streams

The traditional flow of scrap through the pri-mary and secondary aluminum industries isdominated by three major scrap streams: UBCs,automotive scrap, and municipal scrap. Therecycling of cans has continued to grow and hasestablished precedents and industrial infrastruc-ture. Now, the more extensive recovery ofaluminum from the transportation sector and theeffective separation and recovery of aluminumfrom the municipal scrap stream are of imminentimportance.

In the last two decades, some 10,000 recyclingcenters have been established in the UnitedStates for aluminum cans and other aluminumproducts. More recently, there has been consoli-dation in the recycling centers and also dramaticgrowth in municipal curbside collection of recy-clables that emphasize the economic value ofaluminum cans. Communities are now supply-ing aluminum cans to more than 300 municipalrecycling facilities covering half the U.S. popu-lation. This has developed a subsidized dealermarket for aluminum cans and other aluminumproducts that can effectively offset the costs for


bins, collection, hauling, separation, aggregation,and remarketing.

The time and rate at which material becomesavailable for recycling is a function of the prod-uct and can be modeled (Ref 7.3). For instance,it is generally accepted that rolled aluminumcomponents in buildings, automobiles, and bev-erage cans have useful lives that are respectively~50, ~15, and ~0.2 years. The quantity of metalthat is sold into these markets is known and canbe forecast. The useful lifetime can be esti-mated and used to determine when the materialbecomes available as old, postconsumer scrap.In this way, a metal flow model has been con-structed to estimate the availability of futurescrap. This model used existing data wheneverpossible and projects future metal flows basedon recent data trends.

Specifically, the scrap model identifies usedscrap based on aluminum product markets. Thepercentage of total metal market and the percentof annual growth over 1990 to 2000, respectively,for these markets were assumed as follows:transportation (37%, 9.8%), containers and

packaging (23%, 0.3%), building and construc-tion (15%, 2.6%), consumer durables (8%,4.8%), electrical (8%, 3.0%), machinery andequipment (7%, 4.8%), and others (2%, 1.7%).

The shear number of cars, sport utility vehi-cles, trucks and trailers produced, approximately15 to 18 million per year, all containing an appre-ciable and growing amount of aluminum, causesthe transportation segment to now dominate thescrap supply. Accordingly, in developing anymodel for recycling aluminum, it is critical tounderstand the changes in aluminum content invehicles over time and to estimate when thismetal may return to the recycle stream. Theamount of aluminum in vehicles is shown bythe annual Ducker Research Company surveysto have grown steadily. Specifically, it hasgrown from a value of 81 lb per vehicle in 1973to 274 lb in 2001, and the latest number for2006 is 319 lb per vehicle. Growth in the 1990swas higher than in the previous decades due toincreased emphasis on fuel efficiency, light-weighting, safety, and recyclability. At the sametime, however, vehicles are lasting longer (the

Pump/chargebay

Metal out

UBCs in

Skimbay

Skim outFurnace Furnace FurnacePump/charge

bay

Skimbay

Chargebay

Pumpbay

Metal

Furnace Furnace Furnace

Pump Scrap

Dryskim

(a)

(b)

Fig. 7.8 Melting processes for used beverage can (UBC) scrap. (a) Early can scrap melter. (b) More advanced swirl scrap chargemelter, which uses a continuous melting process


previous lifespan of ~12 years is now trendingtoward 15 to 18 years). Data show the signifi-cant lengthening of the lifespan of vehicles from1980 to 1990 in part is due to the application ofbetter electrocoat paint systems and bodydesign improvements, including the increaseduse of aluminum, that significantly reduce bodyrusting.

Figure 7.9 shows the model results for trans-portation metal flows and scrap material from1960 through the year 2030. When the results ofthe transportation metal flows are evaluated(Ref 7.3), it indicates that at some stage during2005, the total volume of scrap from thetransportation sector vehicles exceeded thatfrom the beverage cans and all other scrapsources. Scrap from the transportation sector isexpected be the dominant form of aluminumscrap for the foreseeable future.

Beverage Cans. The recycling of UBCs is aremarkable success story (Fig. 7.10). It is not anexaggeration to suggest that recycling hasplayed a major role in the market growth of alu-minum beverage cans and their penetration intoa market previously dominated by competingmaterials. In 1976, the aluminum can accountedfor 21% of the beverage container market; 46.4billion containers were produced, of which 4.9billion were recycled. In 1986, 72.9 billion canswere produced, and 33.3 billion cans were recy-cled. In 1988, 94% of all beverage can bodieswere produced from aluminum, and virtually100% of all cans featured the aluminum easy-open end.

Aluminum producers, can makers, and thepublic have invested in the return system.Large-scale consumer advertising campaignshave emphasized energy savings and resource

conservation. Under deposit legislation, canshave generally proven most convenient to handleby consumers, retailers, bottlers, and whole-salers. Collection activities include reversevending machines, mobile return centers, andpublic information and educational campaigns.Recycling centers are active in the developmentof thematic programs to promote the concept ofrecycling. These programs often associate recy-cling with civic causes and medical programs,and encourage volunteer, service, and commu-nity groups and individuals to maximize thereturn of UBCs for charitable benefits.

Recycling of UBCs is based on the inher-ently high scrap value of aluminum and itsconvertibility into new-can stock. For the mostpart, UBCs are consumed by primary producersof rigid-container sheet and are employed inthe regeneration of can stock. The growth inmarkets such as Western Europe, coupled withhigh energy rates in some countries (Japan, forexample), has also created an active marketinvolving the export of UBCs and can scrap.The alloy compatibility of the components ofthe can makes the UBC uniquely suitable forthe closed-loop recycling concept, and it isresponsible for the consistent high value ofUBCs as well as the ever-increasing volume ofUBCs in new-can sheet.

Automotive Scrap. Increased activity in therecycling of spent automobiles is based largelyon new steel technologies that make scrap con-version economically attractive. However, theexceptional value of the nonferrous metals thatare being separated through the efforts ofmetal recyclers and the secondary industryalso contributes significantly to this recyclingactivity.

6,000

5,000

4,000

3,000

2,000

1,000

0

Met

ric to

ns×1

03

1960 1970 1980 1990 2000 2010 2020 2030

Total transportation metalVehiclemetal

Total consumerscrap

Available vehicle scrap

Captured vehicle scrap

Fig. 7.9 Transportation metal flows


Because the use of aluminum in automobilesis increasing (Fig. 7.11), the aluminum valuerecovered from shredded automobiles is of spe-cial importance to the metal castings industry.The potential for aluminum use in automobilesis limited only by how aggressively materialadvantages are pursued by the manufacturers.Growth to the present level of aluminum use inthe automotive industry has been largely asso-ciated with construction concepts emphasizingcastings. Aluminum alloy forgings, sheet, andextrusions are presently being developed byinternational automobile manufacturers incooperation with primary aluminum producersto enhance fuel efficiency.

In 1990, 10 million retired automobiles yieldedas much as 500,000 metric tons (550,000 tons) ofaluminum; this figure is based on a recovery of55 kg (120 lb) per vehicle, approximately the1980 average. In 1995, recoverable aluminum increased to 820,000 metric tons (900,000 tons).Trucks and buses, in which aluminum is usedmore intensively, have not been included.

Municipal Scrap. The separation of metalsfrom municipal refuse has not grown as origi-nally expected. The best available informationindicates that more than 12 million metric tons(13.5 million tons) of metals are lost annuallythrough refuse disposal in the United States.

Metals make up only 9% of total refuse. Ofthe total metals available in metal municipalrefuse, less than 1% is recovered. Because the

cost of separating all components of the refusestream must be based on the value of reusablematerials and energy content, a logical conclu-sion is that municipal refuse processing has notyet become economically viable. Future expan-sion of such processing will depend either ongovernment subsidies, changes in energy costs,or the availability of raw materials.

The solid-waste crisis has exposed the need toincrease all forms of recycling activity. In 1988,the United States generated 145 million metrictons (160 million tons) of waste. In 1979, thenation had 18,500 landfills in operation. Thisnumber was reduced to 6500 by 1988, and theU.S. Environmental Protection Agency (EPA)estimates that 80% of existing landfills willclose in the next 20 years. The EPA and otheragencies have argued for responsible reductionsin waste, increased emphasis on recycling, andthe use of incineration to generate energy andgreatly reduce the reliance on landfill capacity.At present, only 10% of waste in the UnitedStates is recycled. The value of aluminum scrapin the municipal waste stream can be expectedto significantly affect the success of plans toincrease the recycling rate.

Clearly, the recyclability of any materialenhances its attractiveness in commercial appli-cations relative to materials that are not recy-clable or that can be recycled only at excessivecost. The inherent recyclability of aluminumand its value after recycling supports and isfully consistent with national environmentaland waste-reduction goals.

Can Recycling Technology

Beverage can recycling has succeeded be-cause of economic and environmental incen-tives. It can be conducted by individuals, bycharitable fundraising groups, by municipalities,and by state mandates. Individuals can earnapproximately 1 cent per can for themselves, orthey can choose to donate or aggregate via anarray of groups, from Boy Scouts and otherorganizations to the largest of corporations andcharities. All channels can make money for indi-viduals, municipalities, businesses, and charitieswhile at the same time saving energy, preservingthe environment, and recycling the metal.

Large amounts of UBCs are either toll con-verted for primary producers or remelted bysecondary operators for resale as remelt ingot orfor use in casting alloys. However, the majority

Num

ber

of c

ans

colle

cted

×109

50

40

30

20

10

0‘78 ‘79 ‘80 ‘81 ‘82 ‘83 ‘84 ‘85 ‘86 ‘87 ‘88

Year

Fig. 7.10 Increase in recycling of aluminum used beveragecans from 1978 to 1988. The calculation for the

number of cans collected is based on a can weight surveyconducted by the Aluminum Association


(approximately 80%) of UBCs are returneddirectly to the primary industry, and UBC recy-cling is thus a prime example of closed-loopproduct recycling. The flow diagram in Fig. 7.12(Ref 7.15) shows the captive nature of the canmanufacturing and recycling loop. The scrappreparation and melt technologies described inthis section are considered the most advancedand should not be viewed as industry standards.However, many UBC converters are using simi-lar, rather sophisticated technologies.

In the recent past, the most successful bever-age can recycling year was 1997, when 66.5%of all cans produced were bought back by theindustry and some 2.05 billion 1b of metal wererecovered. The rate of can recycling declineduntil 2004, when the trend reversed and slightlymore than 50% of all cans were collected.

To encourage higher percentages of can collec-tion, the industry has further strengthened the recycling infrastructure and embarked on exten-sive marketing and publicity efforts. For example,November 15 each year has been declared“America Recycles Day,” and localities and com-munities have been challenged to recycle more.Some nations, notably Brazil, Japan, Sweden, andSwitzerland, have already achieved can collectionrates at the level of +90% due to a combination ofeconomic, social, and regulatory forces.

In conjunction with the Can ManufacturersInstitute, the Aluminum Association has formedthe Aluminum Can Council (ACC) to develop

educational programs and encourage collectionthrough school, community, and municipal recy-cling promotions. The ACC formed the CurbsideValue Partnership to enhance curbside collection,public awareness, and increased household binuse. The Aluminum Association has also estab-lished “Cans for Habitat” in conjunction withHabitat for Humanity, and since 1997, some $4million dollars has been collected and 92 homeshave been built for needy families (Ref 7.16).

These activities, along with a strong municipalcurbside recycling network established by thepaper and fiber industry, have started to drive upthe aluminum recycling rates and have resultedin several marketing slogans to encourage theindividual collector. Some examples are that“recycling each beverage can is equivalent tosaving the energy content of 6 oz of gasoline,”or that “recycling a can saves enough electricityto run an average television set for 3 hours,” or“recycling buys a nail for a Habitat for Humanityhouse.” All of this has encouraged and simplifiedthe rationale for recycling.

When collected by industry, the cans areshredded, remelted, cast, and rolled again toform new rigid-container sheet from which newcans are made. These new cans are back on thesupermarket shelves in as little as 60 days.Over 95% of recycled UBCs go back into cansheet, forming a closed-loop recycled materialproduct. Cans in the United States currentlycontain an average of 49% recycled material

Am

ount

of a

lum

inum

per

veh

icle

, kg

75

60

45

30

15

0

165

132

99

66

33

0

Am

ount

of a

lum

inum

per

veh

icle

, lb

1945 1950 1955 1960 1965 1970 1975 1980 1985 1990

Automotive model year

United States domestic

Imported

Fig. 7.11 Average use of aluminum in automobiles from 1946 to 1988. U.S. automakers have used more aluminum than worldproducers since the mid-1950s


from this closed-loop process. Other recycledmaterial products, such as paper coffee cupsand polyethylene terephthalate soda bottles,boast approximately 10% recycled material. Alternatively, the recycled material can berolled into other alloys to form additional build-ing blocks of material for use in transportation,packaging, building and construction, or otherapplications.

Collection. The UBCs are received fromcollection centers as bales weighing 400 kg (880lb) or as briquettes with a maximum density of500 kg/m3 (31 lb/ft3). These bales and briquettesare broken apart and the cans shredded to ensurethat no trapped liquid or extraneous material willreach the melters and cause serious damage orinjuries. From the shredder, the material passesvia a magnetic separator that removes ferrouscontaminants and through an air knife that sepa-rates more dense nonferrous and nonmagneticferrous materials, such as lead, zinc, stainlesssteel, and high-nickel iron. The shredded alu-minum cans are then typically delacquered.

Delacquering. There are two basic approachesto thermal delacquering. One is based on a rela-tively long exposure time at a safe temperature,and the other is based on staged temperatureincreases to just below melting for as short an

exposure time as possible. The first approachuses a pan conveyor on which a bed of pre-crushed and shredded UBCs approximately 200mm (8 in.) deep moves through a chamber heldat approximately 520 °C (970 °F). The chambercontains products of combustion (POC) that arediluted with air to provide the proper atmosphereand temperature for the pyrolysis and combus-tion. The second approach employs a rotary kilnwith a sophisticated recirculating system forPOC gases at various entry points. The tempera-ture in the last stage is near 615 °C (1140 °F),which is very close to the temperature at whichincipient melting occurs in the aluminum-magnesium (5xxx-series) alloys typically usedin can lids and tabs.

Precise practices and controls are vital forthese delacquering units, which treat approxi-mately 18 metric tons (20 tons) of scrap (~1.25million UBCs) per hour.

Alloy Separation. Hot, delacquered UBCsthen move into the thermomechanical separationchamber, which is held at a specific temperatureand contains a nonoxidizing atmosphere. In thischamber, a gentle mechanical action breaks upthe alloy 5182 lids into small fragments alonggrain boundaries, which have been weakened bythe onset of incipient melting. An integrated

Fig. 7.12 Flow diagram of closed-loop used beverage can (UBC) recycling and manufacturing. HDC, horizontal direct chill; EMC,electromagnetic casting. Source: Ref 7.15

UBCs andcan-manufacturing

scrap Scrappreparation

Shreddinghammermill

high-triporque

Heavy-metalremovalmagnetsair knife

Delacqueringrotary Kiln

bed conveyor

Scrap

Mill scrap Meltingcontinuous

melters

Holdingfluxing

In-linetreatment

filters

CastingHDCEMC

Ingot Rollingbody stock

ThermomechanicalseparationBody-3004Lid-5182

3004

5182

Skim

Skim

Recovered metalSkim

treatmentcooling/preparation

Rotary furnace

Melt loss

Pot room metal

Mill scrap

MeltingContinuousoperation

Holdingfluxing

In-linetreatment

filters

CastingHDCEMC

Rollinglid stock

Ingot

Scrap

Body sheet

Lid sheetCan

manufacturerCans

Scrap

Bottlerbeer

soft drinksRetailer Comsumer


screening action removes the fragments as soonas they can pass the screen, to avoid overfrag-mentation. This process requires a very narrowoperating range to avoid melting and assure maxi-mum segregation. The screened-out alloy 5182particles are transported to the lid stock melters,and the larger alloy 3004 particles continuedirectly into the body stock melters.

Melting, Preparation, and Casting. At pres-ent, most melting facilities for UBCs throughoutthe industry are dedicated units designed to han-dle the enormous volumes and to minimize themelt losses inherent in melting thin-walled mate-rial. Larger companies have developed their ownoften-proprietary processes. Significant amountsof skim—the mixture of metal, oxides, othercontaminants, and trapped gas that floats on topof the melt—are removed and treated for metalrecovery. The metal from these dedicated meltersis often transferred to on-line melting furnaces,where additional bulky scrap is remelted and pri-mary unalloyed metal is charged to create the de-sired volume of the proper alloy composition.From these furnaces, the metal is transferred tothe holding furnaces, where minor compositionadjustments are made and metal quality treat-ments are performed (for example, gas fluxing toremove hydrogen). Some metal treatment, forexample, inclusion removal, can be done in so-called in-line treatment units; again, most majorcompanies have developed their own preferredmethods and technology. The clean and on-composition metal is cast into essentially rectan-gular cross-sectional ingots that are typicallyscalped to remove nonuniform surface and subsurface structure and thermomechanicallyprocessed to finished sheet dimensions. Processscrap in the form of scalpings, edge and end trim,and discrepant product is in-house, or runaround,scrap, which is returned to the melters.

Sheet is shipped to can manufacturers, whowill generate approximately 20% skeleton andother forms of scrap that are most typicallyreturned to the aluminum supplier. On a globalbasis, 55% of a melt consists of new (produc-tion-related) scrap. If all cans were returned asUBCs and total melt losses were 7% of themelt, this 7% would be the only makeup metalrequired from primary smelters to close the loop(provided the market remained constant).Figure 7.13 illustrates this interrelationship.

As the recycling rate continues to increase,composition control and the correspondingcontamination avoidance become technical chal-lenges as important as melt loss reduction. These

developments were predicted in a mathematicalmodel of the recycling system (Ref 7.17). Itappears that prudent use of salt fluxes may holdthe key to improvements in these areas.

Automobile Scrap Recycling Technology

In contrast with the modern aspects of closed-loop can recycling, the recycling of aluminumfrom automobiles has existed for more than acentury. It has traditionally been a multifacetedscavenging activity carried out by independentscrap yards. Materials selection by car manufac-turers has been based on cost and performance,with little consideration given to the economicor environmental value of recyclability.

Transportation is currently the largest (33.9%in 2004) and fastest-growing market of the alu-minum industry (Ref 7.2) and achieves greaterthan 90% recovery of recycled material back tothe industry. Market growth rates for aluminumin transportation were 9.8% over the decade1990 to 2000 and 6.1% from 2001 to 2004. Thisgrowth is driven by the need for light-weighting,increased fuel efficiency, and safety. It has beenestimated that the replacement of heavier com-ponents in vehicles by aluminum saves ~16 kWhannually for each kilogram of aluminum used.Incidentally, this number implies that, on a life-cycle basis, all the energy required to produceprimary aluminum is recovered in vehicle fuelsavings within three years, and in less than threemonths when recycled metal is used. This largelight-weighting advantage is expected to con-tinue to drive the high growth of aluminum inthe transportation sector. In essence, this growingquantity of metal is a very large and growingmobile “urban mine” of metal available forrecycling (Ref 7.3).

The time and rate at which the urban minematerial becomes available for recycling is afunction of the product and can be modeled, aspreviously noted in the section “DevelopingScrap Streams” in this chapter. The useful life-time can be estimated and used to determinewhen the material becomes available as old orpostconsumer scrap. In this way, an urbanmine metal flow model has been constructed toestimate the availability of future scrap. Themodel considers the number of vehicles pro-duced, the amount of aluminum per vehicle,vehicle life, growth rate of aluminum use, im-pact of dismantlers, the role of automotiveshredders, and so on.


As mentioned before, the amount of alu-minum in cars has steadily increased since 1946(Fig. 7.11). Substantial amounts of aluminumhad been applied in early car manufacturing, butbetween 1925 and 1946, its use was minimal.Until 1975, most growth was in castings, but asa consequence of the energy crisis, more wroughtalloy hang-on parts began to be used for weightreduction, and sheet metal panels and spaceframe constructions are now under development.An average 1980-model car contains approxi-mately 70 kg (155 lb) of aluminum, whichaccounts for only 5% of the total weight of theaverage car. Yet, it is economically an importantfraction.

Figure 7.14 is a schematic showing one ofseveral possible paths for the recovery of themajority of the materials presently used in cars.The aluminum fraction is usually sold to anautomotive cast shop, which uses open hearth aswell as induction furnaces and occasionallyrotary salt kilns. The fraction of inseparable mul-timetallic particles must be treated in a sweat fur-nace. In some schemes, without heavy-mediumor eddy current separation, the entire nonferrousfraction goes to the sweat melter.

The can recycling loop is totally dedicated toa single product of two compatible aluminumalloys. Automotive recyclers, on the otherhand, must deal with a number of fractionswith different destinations and relatively lowvalues. For automobile recycling to become aseffective as can recycling, a cooperative effortis required by the scrap collectors, handlers,

and manipulators to introduce advanced scrapseparation and upgrading technology. A num-ber of alternative scrap recovery methods havebeen developed for other mixed-aluminumscrap sources, and these methods could beadapted for use in car recycling. They includeimproved preprocessing methods to concen-trate aluminum fractions and mechanical,physical, and chemical separation processesfor upgrading scrap mixtures and recoveringaluminum from low-grade sources.

One major factor that has strongly impactedvehicle recycling is the advent of automotiveshredders for the industrialized collection ofautomotive scrap. Nowadays, most vehicles arefirst processed by dismantlers and scrap proces-sors, and then the residual hulk is shredded. Aftershredding, followed by air separation to removefluff, magnetic separation of ferrous material,sink-float density separation, and eddy currentseparation, it is possible to obtain a nonferrousmetal stream containing some 90 to 95% of theavailable metal. In other words, shredders haveindustrialized and automated the scrap collectionprocess and made it more efficient. Two addi-tional developments by Huron Valley Steel Corporation (Ref 7.18, 7.19) have further ad-vanced this technology. By use of color sorting, itis now possible to sort cast aluminum alloys fromwrought material, and by using laser-inducedbreakdown spectroscopy (LIBS), it is possible toanalyze the emitted plasma light from a sampleand determine its chemical composition. Thisanalysis of individual pieces of scrap has now

Can-manufacturing scrap, 13% (new scrap)

Mill scrap, 42%(runaround scrap)

Melt, 100%

Virgin metal Cans,41.5%

Cans not returned

Recycled cans (old scrap)

Process losses

Fig. 7.13 Process by which recycled cans replace virgin metal in the beverage container market


been demonstrated at a production scale and isshown schematically in Fig. 7.15. Although notyet economically feasible, it is technically possi-ble to separate and sort aluminum on an alloy-by-alloy basis by virtue of these developments.

High-Temperature Process. Besides thedelacquering processes for can scrap describedearlier, other high-temperature processes havebeen considered for separating aluminumscrap. However, these high-temperature meth-ods were not effective or cost too much energy.Methods included a fluid-bed rotary furnace toremove paint, plastic, and other combustiblesfrom aluminum (Ref 7.20). Another was the so-called Granges box used in Sweden, which isan oven consisting of two chambers, one forcontaining scrap, the other for combustion ofgases and fumes released from the scrap. Thesehigh-temperature batch processes are no longerused.

Other Separation Technologies. Some otherseparation technologies that have been used inEurope include low-temperature separation andgravity separation methods.

Cryogenic separation has been performed com-mercially in Belgium on shredded automotivescrap. The method is based on the difference inductility of nonferrous and ferrous metals at ex-tremely low temperatures. Below –65 °C (–85°F), ferrous metals become very brittle and can befragmentized easily, whereas nonferrous metalsremain ductile. Simple screening achieves separa-tion. This method is expensive and should be usedonly for separation of mixed shredder fragments(of steel attached to aluminum, for example).

Gravity seperation methods include thecommon heavy-media/sink-float process men-tioned earlier and a process developed in theNetherlands (Ref 7.21). The latter uses thesame heavy medium (ferrite/water suspension),

New materials

Engine blocks

SteelNew cars

Electrical anddie cast parts

PartsRepair

Retired carscollection

Dismantler

HulkShredder

Air classifier

Magnet

Heavy medium

Eddy current

Sweat melterResidue(discard)

Rubber, glass(to landfill)

Copper, zinc

(to melters)

Fluff (to landfill)

Pig aluminum

CasterAluminum

Ferrous

Steelmaker

Baler

Fig. 7.14 Recycling loop for aluminum automotive components. Castings make up the bulk of aluminum automotive scrap


but it is not a passive sink-float method.Instead, the scrap is charged in a cyclonethrough which the heavy medium is pumped.In this manner, the sensitivity of the process isincreased, and inaccuracies due to shapedifferences of the particles are decreased.

Building and Construction Recycling

The recycling of rolled products from thebuilding and construction area has long beensomething of an unknown quantity of recycledmaterial, mostly because buildings have a longlife cycle (~50 years) and the data are sparse.In an attempt to provide some more specific information for this activity, the European Alu-minum Association recently commissioned astudy by the University of Delft to provide infor-mation on the content and collection rate in sev-eral European buildings (Ref 7.22). Buildings inFrance, Germany, Italy, The Netherlands, Spain,and England were closely monitored, and com-prehensive data were gathered. It was found thatthe collection rates of aluminum in this marketsector varied between 92 and 98%, therebydemonstrating enormous potential for recyclingand highlighting the role of aluminum in strivingfor full sustainability. With the enforcedreduction of landfill areas and the introductionof increased landfill fees, it is anticipated that therecycling rates of building materials will tend toincrease.

A survey of U.S. aluminum producers in late2003 indicated that the total recycled content ofdomestically produced, flat rolled products forthe building and construction market wasapproximately 80 to 85%. A subsequent surveyof the producers indicated that, on average,approximately half of the recycled content (or40 to 42% of the total content) is from postcon-sumer sources (Ref 7.23).

Aluminum Foil Recycling

While aluminum foil is fully recyclable, theamount of foil that has actually been recycledhas been somewhat limited. With its thin crosssection and high surface area, foil tended to berapidly oxidized into dross, and product recoverywas poor. The recycling of the combined plastic-foil aseptic packaging especially had been anissue. However, in May 2005, a partnershipinvolving Alcoa Alumino, Tetra Pak, Klabin, andTSL Ambiental announced a solution to the recycling of aseptic packaging. Using a high-temperature plasma jet to heat the plastic andaluminum mixture under an inert atmosphere,the plastic is converted to paraffin, and the alu-minum is recovered in the form of high-purityingot, which can then be recycled to manufac-ture new foil. The emission of pollutants duringthe recycling of materials is reported to beminimal, and the energy efficiency rate is closeto 90%.

Fig. 7.15 Schematic of scrap-sorting laser-induced breakdown spectroscopy (LIBS)

Al concentratecar shredder

LIBS

Camera

Laser

Spectrometer

Al alloy

Al

λ (nm)


Impurity Control

As noted earlier, a dominant issue in the recy-cling of rolled aluminum products is the controlof impurities, especially iron and silicon. Bothof these elements tend to increase modestly themore frequently the metal has been recycled.Casting alloys, with a generally much highersilicon content, can markedly increase impuritylevels of a wrought alloy recycle stream. Ironimpurities generally build up due to wear andtear of process equipment.

These impurities can adversely impact physi-cal, chemical, and mechanical properties, and,as a result, definite alloy compositional limits areset for aluminum alloys. This is especially truein the aerospace market, where there are criticalrequirements for fatigue, toughness, creep, andstress corrosion, and the impurity levels mustbe maintained at relatively low levels. High-performance rolled, automotive alloys generallyrestrict both silicon and iron maximum limitsbut at levels considerably above those of theaerospace alloys (Ref 7.24). Better control ofimpurities or improved sorting of recycledmaterials is necessary for the widespread appli-cation of recycled alloys, and the industry is exploring several options. At present, the recy-cled material is “sweetened” with more pure,primary ingot to attain the required impuritylimits for specific alloys.

The better the impurities can be controlled,the more economical the whole recyclingprocess will become. Of course, most recycledwrought alloys tend to fit well within the compo-sitional limits of casting alloys, which typicallyare significantly higher, especially in regards tosilicon. However, this approach does not satisfythe need to recycle wrought alloys back intowrought material. Several approaches for copingwith the wrought alloy impurity issues come tomind.

It may be possible to modify the compositionsof certain alloys with recycling in mind. Con-ceivably, these alloys may be used as paintedstock where corrosion issues are less critical(soffits, downspouts, etc.) and where mechanicalproperty needs are less demanding. Already,some alloys such as 3105 and 6061 have highercompositional limits for iron (0.7% maximumin each case) (Ref 7.25) and are good alloys forpotential scrap additions. Perhaps the composi-tion limits of such alloys can be modified with-out detriment to their physical or mechanicalproperties.

Another approach is to encourage the designand assembly community to design with eventualrecycling in mind. For example, it is importantto minimize the use of dissimilar metals injoining and assembly. Some early automotivedesigns employed steel rivets or clinching toattach an aluminum roof to the aluminum body-in-white, a design that can be detrimental to thequality of the recycled feedstock after the even-tual shredding of the vehicle.

With the successful development of the LIBStechnology, it is quite feasible technically,although not yet economical, to conduct alloy-by-alloy sorting. Before this stage is reached, itwould be possible to set the computer limits tosort and discard pieces of scrap with excessivelyhigh iron content, thereby producing a muchpurer melt composition subsequently. The high-iron material could then be targeted for thepreparation of “de-ox” alloys and sold to the steelindustry, thereby eliminating the material fromthe aluminum process stream.

Recently, the corrosion performance of mod-ern zirconium-refined magnesium alloys hasbeen markedly improved and is now close to thatof aluminum alloys. This has been achieved bycontrolling the level of iron impurities to be < 0.005% in the alloy, and this in turn has im-proved the corrosion performance of the oxide.Zirconium effectively controls iron levels byforming heavy iron-zirconium intermetallicparticles that sink to the bottom of the crucibleduring processing (Ref 7.26). Can such a similar“density separation” concept be made to workfor the removal of iron from aluminum melts?

Another option may be to “learn to live withthe impurities” by modifying production prac-tices. For example, with partial support fromthe Department of Energy, a multipartner teaminvolving the University of California atIrvine/Davis, Colorado School of Mines, IdahoNational Laboratory, Alcoa Technical Center,Pechiney Rolled Products (now part of Alcan),Inductotherm Corp., and Metals Technology,Inc., is currently developing a spray rollingaluminum strip process (Ref 7.27). This sprayrolling process is an innovative strip-castingtechnology that combines the benefits of twinroll casting with conventional spray forming(Fig 7.16). The project description says:

“Compared to conventional ingot processing,spray rolling significantly reduces energy con-sumption and cost by eliminating direct chillcasting, homogenization, and hot rolling unit


operations. Compared to twin-roll casting,spray rolling significantly increases productionrate and is applicable to a broader range ofalloys. Tensile properties of strip processed byspray rolling have met or exceeded those ofingot-processed material.”

Currently, the process has been demonstrated ata strip width of 20 cm (8 in.) and is to be scaledup to 51 cm (20 in.) in the next stage. Moresignificantly, the project description goes on tosay:

“The resultant improvements in fuel effi-ciency and anticipated improvements inrecyclability and tramp element tolerance ofspray-rolled sheet would result in evengreater energy savings.”

In general, rapid-solidified materials such asthose generated by spray forming have beenshown to tolerate higher-impurity contentsbecause the tramp elements are frozen in thestructure and do not have as much opportunity tosegregate to the grain-boundary regions, wherethey can form intermetallic precipitates andcause localized corrosion or embrittlement sites.It will be important to demonstrate whether thisspeculation about spray rolling having greatertolerance for impurities is in fact true.

Molten Metal Handling and Safety*

The major safety concerns specific to the alu-minum industry involve molten metal and pot

bath and finely divided aluminum particles,which, under certain circumstances, can give riseto explosions, with serious consequences. Thehazards of producing and handling these materi-als are well recognized, and the industry hasundertaken a number of programs and datacollection efforts to better understand and con-trol the risks of burns, explosions, injuries, andparticular hazards, depending on the processesand/or products involved.

Molten Metal Explosions

The single greatest safety concern in the alu-minum industry is a molten metal explosion.While millions of pounds of molten aluminumare produced and handled without incidentevery day, explosions do infrequently occur.Most are the result of the molten metal trappingand reacting with water, but other contaminantscan also bring about catastrophic incidents.

In general, whenever two liquids at widelydifferent temperatures are brought into intimatecontact, an explosion can result, particularly ifthe colder liquid has a relatively low boilingpoint. Examples of these potential explosionscan be found in many different industries. Thepaper industry has long had concerns aboutsmelt-water reactions. The nuclear power indus-try has had serious concerns and sponsored aconsiderable amount of research about steamexplosions that could arise from meltdown ofcore and/or cladding reacting with coolingwater in a nuclear reactor. And, of course, themetals industry is concerned about reactions be-tween molten metal and water.

In most instances of explosions involvingmolten metals and water, the water is trapped

Liquid metal

Consolidationrolls

Spray-rolledstrip

Atomized spray

Atomizer nozzle

Inlet gas

Fig. 7.16 Schematic of spray rolling process

*Adapted from the paper “World-Wide Efforts to Prevent Explosions in the Aluminum Industry” by Seymour G.Epstein, Technical Consultant, The Aluminum Association


between the molten metal and a solid surface. Itquickly turns to steam and, in so doing, expandsmore than a thousand times. This rapid expan-sion, accompanied by a rise in temperature, is anexplosion that can throw metal a large distance,injure employees, and damage equipment.

With aluminum, there is an additional factor tobe recognized. Aluminum is a very reactivechemical element with a great affinity for oxy-gen, with which it is almost always combined innature. Just as it requires a large amount ofenergy to break the aluminum-oxygen bonds andproduce metallic aluminum in a reduction cell,that energy will be released if the aluminum isable to recombine with the oxygen from eitherwater or air. From thermodynamic calculations, ithas been estimated that the energy released from1 lb of aluminum fully reacting with oxygen isequivalent to detonating 3 lb of trinitrotoluene(TNT).

Research on Causes and Prevention ofExplosions. For more than 50 years, the industryhas studied molten aluminum-water explosionsrelevant to what may happen if a bleed-out ofmolten metal were to occur during the directchill (DC) casting process. These studies havebeen conducted at company laboratories (Ref7.28–7.30), sponsored by the industry atindependent research laboratories (Ref 7.31),and conducted at government laboratories withindustry cooperation (Ref 7.32) to gain an un-derstanding of these phenomena, how they areinitiated, and how they may be prevented. Theresults of these studies have been published inthe open literature and presented at workshopsand international conferences (Ref 7.33).

These early studies revealed that certainorganic coatings applied to the exposed surfacesin a DC casting pit will prevent explosionsresulting from bleed-outs. Recently, a series ofstudies were undertaken to gain a better under-standing of how the coatings afforded that pro-tection and to identify new coatings to replacethe long-used Tarset Standard, whose productionhad been discontinued (Ref 7.34–7.36). Thestudies were conducted at the Alcoa TechnicalCenter, with a supporting effort at the OakRidge National Laboratory, sponsored by aninternational consortium arranged through theAluminum Association and including aluminumproducers, coating suppliers and an equipmentmanufacturer.

Based on the results of these efforts, safetyprocedures were developed and implemented,guidelines and training aids were prepared and

distributed, and a number of workshops andconferences were held in this country andabroad to provide pertinent information to thosein need of it. The efforts are continuing, and,while explosions have not been eliminated, theindustry has made considerable progress in bet-ter protecting its employees.

Molten Metal Incident Reporting

For more than 20 years, the Aluminum Asso-ciation has conducted an industry-wide moltenmetal incident reporting program to glean asmuch information as possible from events.Currently, the program has 230 participantsreporting for approximately 300 plants locatedin 20 countries (Ref 7.37).

Scrap-Charging Safety. One important find-ing emerged from the incident reporting pro-gram. Many of the most severe explosionsresulted from charging scrap into a remeltfurnace, no doubt a reflection of the greatlyincreased worldwide recycling of aluminumsince the program was initiated. In particular,contamination of the charge was suspected inmany of the events. As a result, the Associationbegan working closely with the Institute of ScrapRecycling Industries (ISRI) in the United Statesto increase awareness of the potential hazards ofscrap contamination (Ref 7.38). Several work-shops were conducted for the ISRI membership,guidelines for scrap receiving and inspectionwere prepared and distributed along with othertraining aids (Ref 7.39), a scrap rejection notifi-cation program was implemented (Ref 7.40), andinformation on scrap melting safety continues tobe disseminated (Ref 7.41).

Sow Casting and Charging. The reportingprogram also identified sow charging as the sec-ond leading cause of furnace explosions. A sowis an ingot weighing from 700 to more than2000 lb that has been cast in a steel mold. Often,solidification of the ingot leaves a shrinkagecavity below the surface. If water enters the cav-ity during storage and is not completely drivenoff by preheating and/or sectioning, an explo-sion can occur when the sow is charged into thefurnace heel of molten metal. Reports were alsoreceived of sows exploding on a dry furnacehearth or during preheating, because the trappedmoisture could not conveniently escape. Need-less to say, the violence is greatest when the sowexplodes under molten metal.

The Association sponsored efforts at the KaiserCenter for Technology and the Alcoa Technical


Center to develop and validate a computer modelfor metal solidification in a sow mold to hopefullylead to improved sow design (Ref 7.42, 7.43).Guidelines for sow casting and charging, basedon these efforts and company practices, were pub-lished by the Association in 1998 (Ref 7.44).

Guidelines and Training Aids

Based on the research findings and on com-pany experiences, comprehensive guidelines forhandling molten aluminum were first publishedand distributed by the Aluminum Association in1980. They were updated, and an expandedsecond edition was published by the Associa-tion in 1990. Through the efforts of an editorialboard with considerable industry experienceand expertise, the third edition of the guidelineswas published in July 2002 by the Association(Ref 7.45).

Awareness of potential hazards is a key objec-tive. The Association publishes and distributesinformation, guidelines, and audiovisual trainingaids. Presentations and short courses on safe han-dling of molten aluminum are arranged with pro-fessional societies, such as TMS, and workshopsare conducted at members’ plants. Casthousesafety workshops are held at regular intervals inthe United States and have also been conductedin Europe and Asia as a joint activity with the Eu-ropean Aluminum Association and the Interna-tional Aluminum Institute. In the United States,these are held regionally for the convenience ofthe plants located in those areas. In addition, pot-room personal protective equipment workshops,sponsored by the Association’s Primary Division,were held at members’ plants.

Finely Divided Aluminum

Because of its chemical activity, affinity foroxygen, and the large amount of surface areaavailable for reaction, finely divided aluminumrepresents a hazard that must be recognized. Aswith other reactive materials, such as flour,starch, and coal dust, fine particles of aluminumdispersed in the air as dust clouds can ignite orexplode when triggered by a spark or otherenergy source (Ref 7.46).

Finely divided aluminum is encountered pri-marily in one of several forms: powder atomizedfrom molten metal; flake produced from ball-milling operations and made into a paste forpaints; and fines generated by various fabricatingor finishing operations such as grinding, sawing,

cutting, sanding, brushing, and so on. Generally,aluminum powders with a particle size of 420 µm(40 mesh) or smaller present a fire and explosionhazard; the finer the particles, the greater theexplosion hazard.

Approximately 25 years ago, the Safety andProperty Protection Committee of the Alu-minum Association’s Pigments and PowderDivision developed Recommendations forStorage and Handling of Aluminum Powdersand Paste (Ref 7.47). Now in its fourth edition,the publication is continually updated andmade available to producers and users. TheCommittee was also involved in the preparationof Guidelines for Handling Aluminum FinesGenerated during Various Aluminum FinishingOperations (Ref 7.48) and has representationon appropriate panels of the National FireProtection Association that publishes standardsfor handling these products. In addition, work-shops are held to help exchange information onprograms and experiences.

REFERENCES

7.1. R. Overbey, “Sustainability—What MoreShould Companies Do?” presentation toAlcoa Conference Board, June 2005

7.2. Aluminum Statistical Review, 2004, TheAluminum Association, Arlington, VA,2005, p 5

7.3. W.T. Choate and J.A.S. Green, “Modelingthe Impact of Secondary Recovery (Recy-cling) on the U.S. Aluminum Supply andNominal Energy Requirements,” paper pre-sented at TMS 2004 (Charlotte, NC), 2004

7.4. E.L. Rooy and J.H.L. Van Linden,Recycling of Aluminum, Properties andSelection: Non Ferrous Alloys andSpecial-Purpose Materials, Vol 2, ASMHandbook, ASM International, 1990

7.5. “The Aluminium Industry’s Sustainabil-ity Development Report,” InternationalAluminum Insitute, 2003

7.6. “Aluminum for Future Operations, Sus-tainablity Update 2006,” International Alu-minum Institute, 2006.

7.7. P.R. Atkins, Recycling Can Cut EnergyDemand Drastically, Eng. Min. J., May 1973

7.8. “Scrap Specifications Circular,” Instituteof Scrap Recycling Industries, Inc., 1988

7.9. J.H.L. Van Linden, J.R. Herrick, and M.J.Kinosz, Metal Scrap Melting System,U.S. Patent 3,997,366, 1976


7.10. J.H.L. Van Linden, R.J. Claxton, J.R. Her-rick, and R.J. Ormesher, Aluminum ScrapReclamation, U.S. Patent 4,128,415, 1978

7.11. J.H.L. Van Linden and J.B. Gross, VortexMelting System, U.S. Patent 4,286,985,1981

7.12. A.G. Szekely, Vortex Reactor and Methodfor Adding Solids to Molten Metal There-with, U.S. Patent 4,298,377, 1981

7.13. D.V. Neff, in Proceedings of TMS FallMeeting, AIME, 1985, p 57–72

7.14. R.J. Claxton, Method for Submerging,Entrainment, Melting and CirculatingMetal Charge in Molten Media, U.S.Patent 4,322,245, 1982

7.15. J.H.L. Van Linden, Aluminum Recycling—Everybody’s Business, Light Metals 1990,TMS, 1990

7.16. “2004 Sustainability Report,” Alcoa, p 26,www.alcoa.com/global/en/about_alcoa/sustainability_report_2004/images/sr_online_final.pdf (accessed July 2007)

7.17. J.H.L. Van Linden and R.E. Hannula, inLight Metals 1981, AIME, p 813–825

7.18. A. Gesing et al., “Separation of WroughtFraction of Aluminum Recovered fromAutomobile Shredder Scrap,” paper pre-sented at TMS meeting (Pittsburgh, PA),Nov 2000

7.19. A. Gesing, L. Berry, R. Dalton, andR. Wolanski, “Assuring Continued Recy-clability of Automotive Aluminum Alloys,”presentation at TMS Meeting (SeattleWA), Feb 2002

7.20. J. Butson, A Market Study for the EnergyEfficiency Office, Alum. Recycl., 1986

7.21. “Observations at Stamicarbon Pilot Plantat Dalmeyer’s Salvage Yard,” Nieuwerkerka/d Yssel, The Netherlands

7.22. “Collection of Aluminum from Buildingsin Europe,” Delft University of Technol-ogy, European Aluminum Association,Brussels; www.eaa.net (accessed July2007)

7.23. “LEED Fact Sheet—Aluminum Sheetand Plate for the Building and Construc-tion Market,” The Aluminum Association,2004

7.24. G. Kaufman, private communication toJ. Green, Dec 2005

7.25. “Aluminum Standards and Data 2003,”Section 6–5, Table 6.2, The AluminumAssociation

7.26. N. Jeal, High–Performance Magnesium,Adv. Mater. Process., Sept 2005, p 67

7.27. K. McHugh and E. Lavernia, “SprayRolling Aluminum Strip,” presentation toDept. of Energy, Industries of the FutureReview, Oak Ridge National Laboratory,Oct 2005

7.28. G. Long, Explosions of Molten Aluminumand Water—Cause and Prevention, Met.Prog., Vol 71, 1957

7.29. P.D. Hess and K.J. Brondyke, Causes ofMolten Aluminum—Water Explosionsand Their Prevention, Met. Prog., Vol 95,1969

7.30. P.D. Hess et al., Molten Aluminum/WaterExplosions, Light Metals 1980, TMS,1980

7.31. A.W. Lemmon, Explosions of MoltenAluminum and Water, Light Metals 1980,TMS, 1980

7.32. L.S. Nelson, M.J. Eatough, and K.P. Guay,Why Does Molten Aluminum Explode atUnderwater or Wet Surfaces?, Light Met-als 1989, TMS, 1989

7.33. S.G. Epstein, Causes and Prevention ofMolten Aluminum-Water Explosions,Light Metals 1991, TMS, 1991

7.34. R.T. Richter, D.D. Leon, and T.L. Leven-dusky, Investigation of Coatings WhichPrevent Molten Aluminum/Water Explo-sions—Progress Report, Light Metals1997, TMS, 1997

7.35. R.P. Taleyarkhan, V. Georgevich, andL. Nelson, Fundamental Experimentationand Theoretical Modeling for Preventionof Molten Metal Explosions in CastingPits, Light Met. Age, June 1997

7.36. D.D. Leon, R.T. Richter, and T.L. Leven-dusky, Investigation of Coatings WhichPrevent Molten Aluminum/Water Explo-sions, Light Metals 2001, TMS, 2001

7.37. S.G. Epstein, A Summary of Findingsfrom Twenty Years of Molten Metal Inci-dent Reporting, Light Metals 2005, TMS,2005

7.38. S.G. Epstein, Recycling Aluminum Safely,Scrap, Vol 57, 2000

7.39. Guidelines for Aluminum Scrap Receivingand Inspection Based on Safety and HealthConsiderations, 2nd ed., Aluminum Asso-ciation Publication GSR, 2002

7.40. D.C. Pierce and C.H. Kenney, PromotingAwareness of Potential Safety, Hazards inAluminum Scrap, Light Metals 1997,TMS, 1997

7.41. M.D. Bertram, F.R. Hubbard, and D.C.Pierce, Scrap Melting Safety—Improving,


but Not Enough, Light Metals 2005, TMS,2005

7.42. J.R. Payne, “Sow Mold Thermal Analy-sis,” final report to The Aluminum Asso-ciation, 1988

7.43. A. Giron and J.E. Jacoby, “ExperimentalVerification of the Aluminum Association’sThermal Analysis Model,” final report toThe Aluminum Association, 1992

7.44. “Guidelines for Aluminum Sow Castingand Charging,” Aluminum AssociationReport GSC, 1998

7.45. Guidelines for Handling Molten Alu-minum, 3rd ed., Aluminum AssociationPublication 69, July 2002

7.46. D.D. Leon, Casthouse Safety—A Focuson Dust, Light Metals 2005, TMS, 2005

7.47. Recommendations for Storage and Han-dling of Aluminum Powders and Paste, 4thed., Aluminum Association PublicationTR-2, 2006

7.48. Guidelines for Handling Aluminum FinesGenerated during Various AluminumFabricating Operations, AluminumAssociation Publication F-1, 2000

SELECTED REFERENCE

• M.E. Schlesinger, Aluminum Recycling,Taylor & Francis, 2006

THE IMPORTANCE OF ALUMINUM as avehicle construction material has been increas-ing quickly over the last decade. The trend ofincreasing aluminum content in automobiles iswell established and will have a dramatic impacton the recycling industry. On average, aluminumalready constitutes ~10% of a vehicle weightand over 50% of the value of the material avai-lable for recovery from 0the end-of-life vehicle(Ref 8.1).

Although automotive aluminum is currentlycompletely recycled, recycling technologies needto be improved to assure continued, high-valuerecyclability of all present and future automotivealloys—particularly wrought alloys, because theuse of these alloys increases in the aluminum-intensive vehicle. For this reason, the recyclingindustry is working diligently to provide techni-cal solutions of new scrap-sorting technologiesto deliver high-value recyclability of aluminumscrap from end-of-life vehicles. While disman-tling is economically justifiable for part rebuild-ing and reuse, material recovery is enabled byparticle liberation via shredding and bulk sepa-ration and sorting techniques. The recyclingprocess for recovery of aluminum from autoshred and its separation into wrought and castproducts (Ref 8.2–8.4) has already broadened

the market for aluminum recycled from shred-der metal concentrate, in particular, options toimprove fractions of recovered wrought alu-minum alloys, because sheet and extrusion alloysare increasingly used as structural and bodycomponents.

This chapter briefly describes a developmenteffort by Huron Valley Steel Corporation for asystem to culminate in the sorting of the wroughtaluminum alloy concentrate from aluminum-intensive vehicle hulks supplied by auto manu-facturers. This development effort is intended toinclude decoating and precleaning followed byseparation of 5182/5754/6111/6061/6082 froman aluminum alloy mix. Three identification/sorting techniques are discussed for further sort-ing of the mixed wrought aluminum alloys toenable recycling of these scrap alloys intowrought alloy products:

• Color sorting of particles tinted by etching inan NaOH solution

• X-ray absorption imaging • Chemical-composition-based alloy sorting

These techniques are suitable for use inwrought alloys that are typically batched fromprime metal. Solid-particle chemical-composi-tion-based sorting, although technically chal-lenging, is the method of choice for furthersorting of the mixed wrought aluminum alloys.This development work has been done as part ofHuron Valley Steel Corporation’s participationin an Automotive Aluminum Alliance programdesigned to demonstrate the ability of newscrap-sorting technologies to deliver high-valuerecyclability of aluminum scrap from end-of-lifevehicles.

CHAPTER 8

Identification and Sorting of WroughtAluminum Alloys*

*Adapted with permission of the Minerals, Metals andMaterials Society (TMS) from the article by A. Gesing,L. Berry, R. Dalton, and R. Wolanski, “Assuring ContinuedRecyclablity of Automotive Aluminum Alloys: Grouping ofWrought Alloys by Color, X-Ray Absorption and ChemicalComposition-Based Sorting,” TMS 2002 Annual Meeting:Automotive Alloys and Aluminum Sheet and Plate Rollingand Finishing Technology Symposia, Feb 18–21, 2002(Seattle, WA), Minerals, Metals and Materials Society(TMS), 2002, p 3–17.




Sources of Aluminum Raw Material forAlloy Sorting

Wrought aluminum scrap can be separatedfrom several sources of scrap material, such as:

• Shredder product fractions• Old dealer scrap• Manufacturing prompt scrap

Shredder fractions include nonmagnetic shred-der fraction (NMSF) and aluminum shredderfraction (Al-SF). North American NMSF sup-ply contains approximately 50,000 metrictons/month (~100 million lb/month) of metalat a concentration of 20 to 95%, of which 60to 70% is aluminum, with 30 to 50% of thealuminum-containing portion comprised ofwrought aluminum alloys (9000 to 11,000 met-ric tons/month, or 20 to 25 million lb/month).North American Al-SF supply (approximately2000 metric, or 5 million lb, per month) is grow-ing quickly because shredders are installing eddycurrent separators to separate a portion of alu-minum directly from shred. The Al-SF producttypically contains >95% metal, >95% Al in thismetal, and 40 to 60% wrought alloys in the aluminum (900 to 1400 metric tons, or 2 to 3million lb, per month).

The categories of the dealer old scrap that con-tain mixed wrought alloys and thus are suitablefor sorting include old sheet and old extrusion.Manufacturing prompt scrap is often commin-gled for practical reasons and thus becomes asuitable feed for an alloy sorter. Commerciallytraded categories of wrought aluminum mixed-alloy manufacturing scrap that are suitable as analuminum alloy sorter feed include mixed stamp-ing scrap, mixed clip, and mixed low-copper clip.

Upgrading automotive scrap to recover andsort aluminum for material recovery involvesshredding, followed by magnetic separation ofiron and metal-nonmetal separation based onelectrical conductivity. The aluminum is thensink-float separated from denser metals. In 2001,a proprietary sorting technology for separationof wrought aluminum shred from the cast alu-minum shred was implemented industrially ascommercial state-of-the-art automated alu-minum sorting (Ref 8.3, 8.4). Typical hand-sortclassification of wrought aluminum productsis given in Table 8.1. Figure 8.1 shows theseproducts, and the elemental composition of thealuminum melt recovered from these productsis given in Table 8.2. These products are now

finding application in use as steel deoxidants,3105 secondary sheet, and in an expanded rangeof foundry alloys. Because these products arenow typically made from segregated manufac-turing scrap, the value added to the mixed-alloyaluminum scrap product by wrought-cast sort-ing is modest. A significantly higher value wouldbe available when the scrap product compositionwill allow it to replace prime metal in the newalloy batch. In order to achieve that, furtherupgrading is necessary.

Improving Recovery for Wrought andCast Fractions

Various alternatives or additional technolo-gies can be considered for further upgradingaluminum recovered from postconsumer scrap.Focusing in particular on the quickly increasingvolume of aluminum in auto shred, potentialmethods to further upgrade wrought and castaluminum fractions fall into two categories:melt processing and solid-particle sorting. Tech-nically feasible approaches that have been triedare reviewed. This is followed by descriptionsof three processes being piloted for commercialdevelopment of improvements in recoveringfrom wrought scrap.

Melt Processing

Melt processing routes include fractionalcrystallization, electrorefining, fluxing, and alu-minum chloride chemistry. Fractional crystal-lization is a commercial process for precipitationof 4N-purity aluminum by controlled cooling ofmolten prime aluminum. When applied toscrap-based melt, controlled cooling results inprecipitation and segregation of complex inter-metallic precipitates containing iron, manganese,

Table 8.1 Hand-sort classification of the products of separation of wrought and cast aluminumSort of aluminum mix from heavy-media float (HMF)

Sort of wrought aluminum concentrate

Al mix from Cast Wrought Wrought Wrought HMF, % Al, % Al conc., % Al-1, % Al-2, %

Al wrought 40.11 10.16 69.74 91.02 95.20Al cast 54.12 85.67 26.03 3.47 3.31Al other 2.82 1.50 2.43 0.47 0.82Other metal 2.34 1.90 1.50 4.72 0.48Nonmetal 0.61 0.78 0.31 0.33 0.19Al + Fe 4.86 7.74 2.56 0.54 0.96Al painted 4.50 3.28 8.44 11.08 14.24

Chapter 8: Identification and Sorting of Wrought Aluminum Alloys / 137

chromium, and titanium in addition to alu-minum. The multicomponent phase diagramplaces limits on achievable upgraded productcomposition. A considerable fraction of alu-minum ends up with the intermetallics, a down-graded residue. Tight control of thermal gradi-ents and convection currents in the melt iscritical to obtaining a product composition ap-proaching that predicted by theory, and recoveryis invariably significantly lower than expecteddue to trapping of aluminum melt among theintermetallic crystals.

Electrorefining is a commercial process forproduction of 5N-purity aluminum from primealuminum in three-layer electrolytic cells. Both

anode and cathode are molten pools of metal, adense anode of AlCu melt and a low-densitycathode of superpurity aluminum product.Because both electrodes are molten, they needto be separated by a layer of electrolyte that issignificantly thicker than that in the Hall-Heroultprimary aluminum cells. Electrolysis transfersaluminum from the anode to the cathode, and,because the entire product is electrolyzed, energyconsumption in electrorefining is higher thanthat required for prime production. Although itis technically feasible to refine scrap in thethree-layer refining cells, scrap electrorefiningcan never compete economically with primeproduction.

Melt fluxing refers to the removal of sodium,calcium, lithium, and magnesium by directchlorination or by exchange reaction with Al3+

containing salt flux. Magnesium and lithiumalloying elements are converted to chlorides andend up with the dross. Other contaminants andalloying elements are less reactive to chlorinethan aluminum and thus cannot be removed fromthe melt by fluxing. This is the standard commer-cial way of controlling the levels of alkali andalkaline earth contaminants in the secondaryalloy melts. However, the use of the toxic Cl2gas flux can be dangerous to the workers and harmful to the environment. Although

Table 8.2 Optical emission spectroscopic elemental composition of melts of recoveredaluminum products

Al mix from Wrought Wrought Element HMF(a),% Cast Al,% Al conc.,% Al-2,%

Al 88 83.5 93.1 97.1Cu 3 4.4 1.2 0.1Fe 1 1.1 0.7 0.6Mg 0.6 0.4 0.7 0.8Mn 0.3 0.3 0.3 0.2Si 5 8.0 2.6 0.5Zn 2 1.9 1.2 0.4Other … 0.4 0.2 0.2

(a) HMF, heavy-media float

Fig. 8.1 Shredded aluminum products recovered from nonmagnetic shredder fraction. (a) Wrought and cast aluminum mix. (b) Cast aluminum product. (c) Wrought Al-2 product


chlorination is a relatively low-cost commercialprocedure, the environmental considerations willfavor routes that reduce the need for fluxing inpoorly controlled conditions of a melting orholding furnace.

Aluminum chloride chemistry was alsoexplored as a route to making prime aluminummetal by Alcoa, Alcan, and other companies. Itis technically feasible to refine impure scrapaluminum melt using the technology developed.One would chlorinate the aluminum out ofimpure scrap melt to yield either AlCl3 at lowtemperatures or AlCl at higher temperatures.Pure aluminum metal is recovered by eitherAlCl decomposition to Al � AlCl3 or by AlCl3electrolysis in a multipolar inert electrode cell.Although both processes have been piloted andabandoned for primary aluminum production,first by Alcan and second by Alcoa, they shouldnot be summarily dismissed. Toth AluminumCorporation has recently developed a method ofpurifying impure AlCl3 produced by low-costchlorination of clay (Ref 8.5). Toth estimatesthat, combined with the Alcoa multipolar inertelectrode cell, this technology could reduce thecost of production of prime by ~50%. Regardlessof whether or not this technology is applied toscrap aluminum melts, the significant productioncost reduction for prime metal production wouldhave a large economic impact on the aluminumrecycling system.

Solid-Particle Sorting

Solid-particle sorting can be based on anumber of particle attributes. Shape and sizecorrelations are widely used in hand-sortingduring dismantling and can be used within lim-its to differentiate between wrought and castcomponents. These correlations, however, arecontext-specific and are not generally useful todifferentiate between shredded pieces of sheetalloys.

Physical property correlations with density,electrical conductivity, magnetic susceptibility,color, or x-ray translucence form the basis of themechanized separation of aluminum from othermetals and nonmetals. However, because nounique property correlation exists with chemicalcomposition for the wrought aluminum alloys,the usefulness of physical-property-basedapproaches for further grouping of wroughtalloys is quite limited. One of the most promisingphysical-property-based approaches is the cor-relation with the chemical reactivity of the alloy.

Tinting by selective etching and/or heat treatmentallows the commercial color sorter to group thetinted pieces into additional metal-purity cate-gories. Another approach is based on thecorrelation between the x-ray mass absorptioncoefficient and the average atomic number of thematerial. These approaches will be discussed inmore detail as follows.

Chemical-composition-based sorting relieson rapid, noncontact quantitative chemical com-position analysis. Three different noncontactspectroscopic methods have been applied to theelemental analysis of aluminum alloys: x-rayactivation with x-ray fluorescence spectroscopy,neutron activation with gamma-ray fluorescencespectroscopy, and laser activation with opticalemission spectroscopy (laser-induced breakdownspectroscopy).

X-ray fluorescence is hampered by the lowenergy of the light element fluorescence x-raysand their mutual interference. The characteristicMg-, Al-, and Si-K x-ray peaks overlap and areall too easily absorbed by any material, includingair. This limits x-ray fluorescence to the determi-nation of the concentration of heavy alloyingelements (iron, copper, manganese, and zinc) inthe aluminum particles. Pilot scrap-sorting unitsbased on x-ray fluorescence analysis were builtin the 1980s in Germany and the USSR; how-ever, none were ever commercialized.

Neutron activation analysis depends on theirradiation of the particles with a highly pene-trating neutron flux and on detecting the fluores-cence of the gamma rays characteristic to theatoms in the particles. Here, the light elementcharacteristic gamma rays are sufficiently dis-tinct and sufficiently energetic to yield quantita-tive concentration results. The highly penetratingnature of both neutrons and gamma rays permitsthe analysis of the total volume of the materialstream, and commercial units recently have beenbuilt that accurately determine the average com-position of a metal-particle stream carried on aconveyor through a source-detector unit. How-ever, the neutron flux provided by a radioactiveisotope of californium is limited in intensity,leading to long exposure times not suitable forparticle-by-particle sorting, which demands highspeeds for industrial application.

Laser-induced breakdown spectroscopy(LIBS) is the illumination of a spot on the particlewith a focused laser beam to generate a plasmaspark, combined with the optical emission spec-troscopy of the fluorescence photons. Thismethod can yield quantitative analytical results


for all alloying elements of interest in aluminumand magnesium, and at a rate that is suitable forindustrial particle-by-particle sorting. There havebeen a few publications on LIBS analysis ofaluminum alloys showing calibration curves withsufficient sensitivity throughout the concentra-tion range of interest for all the major alloyingelements (Ref 8.6–8.9). Most of these studieswere done by signal averaging a large numberof laser shots on a flat, clean, stationary samplesurface. The challenge to the application of thistechnique to industrial particle-by-particle sort-ing is to:

1. Locate a suitable target spot on a randomlyshaped and randomly placed rapidly mov-ing aluminum particle

2. Deliver a laser shot containing both spotcleaning and analysis pulses to this selectedsingle spot

3. Collect the light from the plasma emissionfrom this random location

4. Simultaneously determine the quantitativeconcentration of all the alloying elementsof interest

5. Make the choice of the target output bin6. Mechanically divert each particle individu-

ally into the target bin

This sequence needs to be repeated for each 20to 50 g particle at a rate that will permit an indus-trial sorting throughput of 3 to 6 tonnes/h (1 to 2kg/s). This works out to 40 to 50 particles persecond. The following reports both pilot-plantand industrial-prototype results that show howclose this challenge is to being overcome.

Pilot Processes for Improved Wrought Recovery

As noted, three processes have been exam-ined for commercial development of improve-ments in recovery from wrought scrap:

• Wrought aluminum tinting and color grouping• X-ray absorption imaging• Aluminum alloy chemical-composition-

based sorting

Pilot projects demonstrate that solid-particlechemical-composition-based sorting, althoughtechnically challenging, is the method of choicefor further sorting of the mixed wroughtaluminum alloys to enable recycling of thesescrap alloys into wrought alloy products.

Wrought Aluminum Color Grouping

One promising way to obtain further separa-tion of mixed wrought alloys is to tint the par-ticles by etching and then use a commercialcolor sorter (Fig. 8.2). This Alcoa-patentedetching process involves a degreasing wash,hot water rinse, hot caustic solution etch, and asecond water rinse followed by hot air drying(Ref 8.10). This process has been further opti-mized by etching samples of known wroughtalloys and choosing the conditions that maxi-mized the color contrast between the variousalloys while minimizing the etching time andthe weight loss.

The selected etch was a 90s etch at 80 °C (175 °F) in a 10 wt% NaOH-water solution. Thedried pieces were sorted into several groups byan industrial color sorter of Huron Valley SteelCorporation. Four color groups were distinguish-able by the color sorter. The color-sorting resultson the etched known alloys are summarized inTable 8.3, which lists recoveries of a givenknown alloy in a particular output stream. Re-covery is defined here as the weight fraction ofthe given alloy found in the particular outputstream. Figure 8.3 shows the tint-etched parti-cles of the known alloys.

The etching method was then applied to an~400 kg pilot-scale sample of wrought alu-minum concentrate with an unknown mix ofalloy feed. See Tables 8.1 and 8.2 for the typical

WroughtAl concentrate

NaOH Al(OH)3

Brightwrought

Al

Graywrought

Al

Darkwrought

Al

Tint etch

Color sort

Fig. 8.2 Block diagram of the tint-etching color-sortingprocess


composition of this feed stream. This time, threecolor groups were distinguished by the colorsorter, as shown in Fig. 8.4. Table 8.4 lists thecomposition of these three products in terms ofhand-sort categories, while Fig. 8.5 shows theelemental composition of ~60 kg crucible meltsrecovered from the products.

In summary, caustic-etched aluminum alloyparticles are tinted gray by silicon, iron, andmanganese, and dark by copper and zinc. Thereis a slight weight loss due to etching and gener-ation of aluminum trihydrate by-product. Colorsorting of tinted wrought shred mix resulted ingrouping into three or four additional productsthat have different aluminum purity but not

directly correlated to alloy families. Color-sorting results are affected by paint, anodizing,and plating.

X-ray Absorption Imaging

X-ray absorption correlates directly to theaverage atomic number of the material illumi-nated by the x-ray flux. Figure 8.6 shows afalse color image in which the hue is set by theaverage atomic number, and the luminescenceis inversely correlated with the particle thick-ness. It demonstrates that aluminum particleswith a higher content of copper, iron, man-ganese and zinc (dark gray) can be separatedfrom those containing only light elements—aluminum, magnesium, and silicon (light gray).This type of imaging can give similar alloygroupings as tinting and color sorting withoutmetal loss and sensitivity to surface contamina-tion. Figure 8.7, however, implies that x-ray ab-sorption imaging cannot differentiate betweenalloys or material mixtures that have the sameaverage atomic number for a different combi-nation of elements. This requires recourse toparticle sorting based on quantitative elementalcomposition analysis.

Fig. 8.3 Tint-etched shredded particles of known wrought aluminum alloys

Table 8.3 Product stream recoveries for colorsorting of known wrought aluminum alloys

Recovery, number percent

Product 1, Product 2, Product 3, Product 4,Component bright light gray gray dark

5182 100 … … …4147 … 100 … …6061 … 70 30 …3003 … 28 72 …6111 10 20 70 …7129 10 20 70 …7016 … 30 70 …2036 … … 20 807003 … … … 100


Composition-Based Particle Sorting

Composition-based particle-sorting technol-ogy has been in development since 1993, build-ing both on earlier work at Los Alamos NationalLaboratory (Ref 8.11–8.13), Metallgesselshaft(Ref 8.14–8.16), Alcan (Ref 8.17–8.19), and onindustrial color-sorting technology. The devel-opment included high-speed, noncontact,quantitative LIBS as an analytical method andthe integration of LIBS analysis with robustcolor-sorter technology developed by Huron

Fig. 8.4 Color-sorted products of tint-etched wrought aluminum concentrate. Labels identify the elemental compositionrecovered melts

Table 8.4 Hand-sort category classification of the color-sort products from the etch-tintedwrought aluminum concentrate feedCategory Bright, % Gray, % Dark, %

Al wrought 91.69 98.02 87.65Al cast 2.08 0.25 10.18Al other 5.64 1.74 1.54Al irony 3.10 3.27 1.04Al painted 19.13 21.37 12.94Cu 0.51 0 0.14Stainless steel 0.08 0 0Wire 0 0 0.05Nonmetal 0 0 0.45

Dark

Gray

Bright

0 1 2 3 4 5 6

Alloying element concentration, wt%

MgMnFeSiCuZnOther

Fig. 8.5 Alloying element concentration in the melt productsrecovered by color sorting of etch-tinted wrought

aluminum concentrate

Fig. 8.6 X-ray absorption image showing separation of aluminum alloys based on the average atomic num-

ber of each particle. Courtesy of the Technical University ofDelft, The Netherlands (see Acknowledgments)


Valley Steel Corporation. In 1998, a full-scaleLIBS alloy-sorter pilot plant (see Fig. 8.8 for ablock diagram) was commissioned to:

• Calibrate magnesium, copper, and siliconspectral lines on the pilot sorter

• Optimize signal-to-noise ratio • Simultaneously acquire spectral lines and

determine elemental particle composition in< 10 ms

• Achieve a resolution of ~0.1% for ~1% con-centration at that rate

• Demonstrate 6111-5182 alloy sort

• Study bulk pre-cleaning and laser spotcleaning

The pilot facility continues to be used for LIBSdevelopment, and topics under currentinvestigation include a study of apparent spectralline shifts and a survey of LIBS spectral emis-sion lines for all major aluminum alloying ele-ments. In 2001, an industrial prototype LIBSsorter was commissioned to demonstrate multi-stream diversion on a full industrial scale andspeed and for accurate targeting of scrap parti-cles with a laser at production speed. One

WroughtAl conc

Alloy 1 Alloy 2 Alloy 3Al alloy

mixAlloy 6Alloy 5Alloy 4

Al alloymix

LIBS alloysorter

LIBS alloysorter

Fig. 8.8 Block diagram of the Huron Valley Steel Corp. full-scale industrial prototype of the laser-induced breakdownspectroscopy (LIBS) alloy sorter

2

1.9

1.8

1.7

1.6

1.5

1.4

1.3

1.2

1.1

1

Mas

s ab

sorp

tion

coef

ficie

nt

Mg

pure

Al p

ure

Al 1

100

Al 6

063

Al 5

754

Al 5

052

Al A

356

Al 6

016

Al 5

182

Al 3

56

Al 3

105

A 3

003

Al 3

004

Al 2

036

Al 3

19

Al 2

024

Al 3

90

Al 7

016

Al 7

129

Al 3

80

Al 7

075

Alloy type

Max

Avg

Min

Fig. 8.7 Theoretical alloy grouping based on the correlation of the average atomic number with the x-ray absorption coefficient.Courtesy of the Technical University of Delft, The Netherlands (see Acknowledgments)


demonstration was for a successful industrialprototype of a 6111-5182 alloy sort.

Bulk-Particle Cleaning. To evaluate bulk-particle-cleaning requirements, LIBS intensityratios were compared for magnesium-aluminumand silicon-aluminum pieces with bare surfacescleaned by various methods with the LIBSintensity ratios for a freshly ground spot. Asillustrated in Fig. 8.9, a stable intensity ratio isobtained from the first laser shot on a groundsurface. With all cleaning methods tried, rangingfrom water washing to caustic etch, more thanone multipulse laser shot is still necessary toachieve an intensity ratio representative of bulkalloy composition that is obtained on a groundsurface. A stable value was obtained after a sec-ond laser shot on an unpainted 6111 alloy surfaceprecleaned by a proprietary method, leading thedevelopment to focus on optimizing laser spot-cleaning conditions rather than searching formore aggressive bulk-cleaning methods. Thenext challenge for the bulk cleaning is the evalua-tion of decoated painted surfaces, including theeffect of pigment residues and chromate andphosphate pretreatment layers.

6111-5182 Alloy Sort. Alloys 6111 and 5182are among the most important sheet alloys in thebody sheet and body structure of the new alu-minum-intensive vehicles. In the sorting of thesetwo alloys, a quantitative analysis that permits thecalibration of the optical emission spectral lines isbeing investigated. In a pilot, a simultaneousanalysis based on magnesium, copper, and siliconcalibration allowed reproducible sorting of 6111(< 1% Mg) from 5182 (~5% Mg). This sort was

accomplished under the following conditions.Sheet samples of known 6111 ~1% Mg and 5182~5% Mg samples were placed randomly on a beltmoving at 1.5 m/s (5 ft/s) on the pilot plant sorteror 2 m/s (6.5 ft/s) on the industrial prototypesorter. During the sort, aluminum, magnesium,copper, and silicon spectral line intensities weremonitored simultaneously, and the aluminum linewas used as the internal standard to normalize theremaining spectral lines. Because the composi-tion of these two alloys differed primarily in themagnesium concentration, the sort strategy wasbased mainly on this element.

Table 8.5 lists the spectral line intensity ratiodata and their precision for the measurementsdone on the pilot plant sorter. Figure 8.10 illus-trates that, even with this limited precision, it isquite easy to separate these two alloys based on

0.9

0.8

0.7

0.6

0.5

0.4

0.3

Mg/

Al i

nten

sity

rat

io

1 2 3 4 5

Laser shot

Fig. 8.9 Spectral line intensity ratio as a function of the number of laser shots for � washed and � ground preparation of a 6111alloy sample surface

Table 8.5 Pilot plant laser-induced breakdownspectroscopy sorter spectral line intensity ratio(IX/IAl) data for the 5182-6111 alloy sort

Alloy

Element Intensity ratio data 5182 6111

Magnesium, IMg/IAl

Concentration, wt% 5 1Average 0.934 0.575Standard deviation 0.042 0.043Coefficient of 5 8

variance (standard deviation/average), %

Silicon, ISi/IAl

Concentration, wt% 0.1 1Average 0.243 0.532Standard deviation 0.065 0.143Coefficient of 27 27

variance, %


LIBS measurement of magnesium concentra-tion. Table 8.3 and Fig. 8.7 show that such cleanseparation is not possible by tint-etching or x-rayabsorption methods. Both pilot plant and indus-trial prototype sorters were system integrated tocombine image acquisition, laser targeting,LIBS analysis, sort decision, and particle diver-sion. During the sort on both machines, the5182 alloy particles were consistently divertedto bin 1 and 6111 alloy particles to bin 2.

Full-Scale Industrial Prototype LIBS Sorter.A full-scale and operational prototype sorter(built in 2001 by Huron Valley SteelCorporation) has been used to demonstrateaccurate, reproducible particle image acquisitionand laser targeting at a belt speed of 2 m/s (6.5ft/s), with 50 laser shots being delivered to up to70 particles passing the target area in a second.Further, high-accuracy particle diversion intofour output streams was also demonstrated atthese production rates. Three diverters wereused, with the first one giving 98% diversion ofthe targeted particles into the first productstream, the second one 95%, and the third one~90%. The particles not diverted end up in thefourth output stream, which is an alloy mix or arecycle stream, and thus have only a minor effecton the composition of the three sorted products.Detailed analytical calibration of the prototype

for all alloying elements is in progress. It isalready possible to reproduce the 6111 and 5182sort on the prototype.

Continued Alloy Sorting Process Devel-opment. To satisfy the objectives of the AutoAluminum Alliance project on AutomotiveAluminum Scrap Sorting (Ref 8.19–8.23),alloy sorting process developments are aimedat separation of known bare alloys. Separationobjectives are between the following alloyfamilies and also separation within the fami-lies. Alloy families for separation are:

• 1100 • 2036• 3003 • 4147• 6111• 5182• 7129

Separation objectives within these families are:

• Manganese: 3003/3004/3105• Magnesium: 5052/5754/5182• MgSi: 6063/6061/6082

When these sorting objectives are met, the nextstep is sorting of bare mixed-alloy shred of un-known composition. Grouping by major alloyingelements (e.g., low zinc, copper, silicon) would

1.2

1.0

0.8

0.6

0.4

0.2

0

Mg/

Al i

nten

sity

rat

io

0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

Mg concentration in sample, wt%

Fig. 8.10 Magnesium calibration of pilot plant laser-induced breakdown spectroscopy sorter based on known 6111-1% Mg and5182-5% Mg alloys. The width of the calibration line represents confidence limits, and the wide vertical bars show

the precision of concentration determination for a given value of the intensity ratio


be the first step, then moving to sorting out parti-cles compatible with a particular alloy, and finally attempting real-time particle-by-particlebatching of particular alloy composition. Soft-ware for real-time optimized batching is cur-rently under development. The accuracy of thesorting will be verified by elemental analysis of60 kg crucible test melts of the sorted products.This development effort is intended to culminatein the sorting of the wrought aluminum alloyconcentrate from aluminum-intensive vehiclehulks supplied by auto manufacturers. It will in-clude decoating and precleaning followed byseparation of 5182/5754/6111/6061/6082 fromthe remaining aluminum alloy mix.

ACKNOWLEDGMENTS

The editors thank the Huron Valley Steel(HVS) Corp. and the Minerals, Metals and Ma-terials Society (TMS) for permission to publishthis paper. Information presented here is a com-pilation of the work of several teams generatedover an extended development timeframe. Theseteams included:

• Wrought-cast aluminum sorting: GeorgeHopson, Tim Good, Richard Grieve, ScottCostello, Sam Lambert (HVS, Belleville, MI)

• Wrought aluminum color grouping: GeorgeHopson, Sam Lambert, Scott Costello (HVS,Belleville, MI) and Paul Schultz (AlcoaTechnical Center, Pittsburgh, PA)

• X-ray absorption imaging: Tako de Jong,Wijnand Kuilman, and Wijnand Dalmijn(Technical University of Delft, The Nether-lands), and Bahram Farahbakhsh (Columbus,OH)

• Composition-based particle sorting: ChrisStewart, Ben AuBuchon, Sam Lambert,Arnold Petz, and Kerry Lang (HVS,Belleville, MI); Philip MacNutt (Austin,Tx); David Hoadley (Farmington Hills, MI);and Paul Torek (Ann Arbor, MI)

REFERENCES

8.1. “Passenger and Light Truck AluminumContent Report,” Ducker Research Com-pany, 1999

8.2. A.J. Gesing et al., Scrap Preparation forAluminum Alloy Sorting, 2000 TMS Fall

Extractive Meeting (Recycling of Metalsand Engineered Materials), Oct 22–25,2000 (Pittsburgh, PA), p 1233–1249

8.3. A.J. Gesing and R. Wolanski, New Devel-opments in Recycling of Light Metalsfrom End-of-Life Vehicles, J. Met., Nov2001, p 21–23

8.4. A.J. Gesing et al., Separation of WroughtFraction of Aluminum Recovered fromAutomobile Shredder Scrap, Aluminum2001 (Proceedings of the TMS 2001 An-nual Meeting, Aluminum Automotive andJoining Symposia), p 31–42

8.5. C. Toth et al., “TAC’s Bipolar Clay-to-Aluminum Technology and Project:The New Way to Make Aluminum,”seminar edition, Toth Aluminum Corpora-tion, Feb 2001

8.6. M. Sabsabi and P. Ceilo, QuantitativeAnalysis of Aluminum Alloys by Laser-Induced Breakdown Spectroscopy andPlasma Characterization, Appl. Spectrosc.,Vol 49 (No. 4), 1995

8.7. M. Sabsabi and P. Ceilo, QuantitativeAnalysis of Copper Alloys by Laser-Produced Plasma Spectrometry, J. Anal.At. Spectrom., Vol 10, Sept 1995

8.8. A.K. Rai et al., Parametric Study of aFiber-Optic Laser-Induced BreakdownSpectroscopy Probe for Analysis ofAluminum Alloys, Spectrochim. Acta B,At. Spectrosc., Vol 56, 2001, p2371–2383

8.9. B. Német and L. Kozma, Time-ResolvedOptical Emission Spectrometry of Q-Switched Nd:YAG Laser-Induced Plasmasfrom Copper Targets in Air at AtmosphericPressure, Spectrochim. Acta B, At. Spec-trosc., Vol 50, 1995, p 1869–1888

8.10. R.K. Wyss and P.B. Schultz, Color Sort-ing Aluminum Alloy Scrap for Recycling,Light Metals 1999, The CD-ROM Collec-tion, The Minerals, Metals and MaterialsSociety, 1999

8.11. “Laser-Induced Breakdown Spectroscopy:A New Spectrochemical Technique,”Technical Report LA-UR-82-465, LosAlamos National Laboratory, 1982

8.12. L.J. Radziemski and D.A. Cremers,Laser-Induced Plasmas and Applications,M. Dekker, New York, 1989

8.13. D.A. Cremers, The Analysis of Metals at aDistance Using Laser-Induced BreakdownSpectroscopy, Appl. Spectrosc., Vol 41,1987, p 572–579


8.14. M. Potzschke et al., Scrap Detector, U.S.Patent 5,042,947, 1991

8.15. H.-P. Sattler, Automatic Sorting of Non-Ferrous Metals from AutomobileShredders, Second International Sympo-sium on Recycling of Metals and Engineered Materials, The Minerals,Metals and Materials Society, 1990, p333–341

8.16. H.-P. Sattler and T. Yoshida, New SortingSystem for Recycling of Magnesium andIts Alloy After Use, First InternationalConference on Processing Materials forProperties, The Minerals, Metals andMaterials Society, 1993, p 861–864

8.17. A.J. Gesing and T. Shaw, Method ofSorting Pieces of Material, U.S. Patent5,813,543, 1998

8.18. A.J. Gesing and A. Rosenfeld, Composi-tion Based Sorting of Aluminum Scrapfrom Aluminum Intensive Vehicles,Developments in Aluminum Use forVehicle Design, SP-1164, Society ofAutomotive Engineers, Inc., Feb 1996, p29–38

8.19. “Automotive Aluminum Scrap Sorting,”Project Fact Sheet, Office of IndustrialTechnologies, Energy Efficiency andRenewable Energy, U.S. Department ofEnergy, July 2001

8.20. J.L. Broge, Managing and RemakingMetals, Automot. Eng. Int., Vol 109 (No. 8),2001, p20–31

8.21. L.E. D’Astolfo, Jr. and P.R. Bruggink,Preserving the Value Chain in AutomotiveAluminum Recycling, Light Metals 1994,The Minerals, Metals and Materials Soci-ety, 1994, p 1121–1127

8.22. L.E. D’Astolfo, Jr. et al., Recycling Auto-motive Aluminum Scrap into New Prod-ucts, First International Conference onProcessing Materials for Properties, TheMinerals, Metals and Materials Society,1993, p 839–842

8.32. Aluminum Industry Roadmap for the Au-tomotive Market: Enabling Technologiesand Challenges for Body Structures andClosures, The Aluminum Association,Inc., Energetics, Incorporated, May1999, p 51

THE ALUMINUM that goes into the produc-tion of aluminum alloy products for manyapplications comes increasingly from recycledproducts. In a recent modeling study of theindustry, Choate and Green (Ref 9.1) illustratedthat most of the increase comes from recycledautomotive components, which, in 2005, wasexpected to exceed for the first time the recy-cled metal coming from used beverage cans.The Aluminum Industry Roadmap (Ref 9.2)also illustrates the importance of these trendsand of the efforts to address the technologyfrom primary production to finished products.Fielding (Ref 9.3) illustrates how one segmentof the industry, the extrusion business, isapproaching the challenge.

As Choate and Green (Ref 9.1) demon-strated, the increase in available recycled metalis a positive trend, because secondary metalproduced from recycled metal requires onlyapproximately 2.8 kWh/kg of metal produced,while primary aluminum production requiresapproximately 45 kWh/kg of metal produced.It is to the industry and national advantage tomaximize the amount of recycled metal, bothfrom the standpoint of energy savings andreduction of dependence on overseas sources(now approximately 40% of U.S. consump-tion) and also from the ecological standpoint,because recycling emits only approximately4% as much CO2 as primary production.

A joint study of a representative Americancommunity by Secat, Inc., the Center for Alu-minum Technology, and the Sloan IndustryCenter for a Sustainable Aluminum Industryfound that for each 1% increase in the amountof aluminum cans recycled, the economic sav-ings to the U.S. economy is $12 million/year.This will approach $600 million/yr if all theavailable aluminum is recycled. The additionalrecycling also contributes to energy savings of1 trillion Btu/year. For the United States as awhole, recycling also has the potential tosignificantly decrease reliance on overseassources of primary aluminum metal.

Aluminum recycling in North America is oneof the well-developed and mature metals recy-cling economies in the world. While today’s(2007) recycling metals markets also includeferrous metals such as iron and steel and nonfer-rous metals such as copper and brass, aluminumrecycling is the engine of recycling economics.However, it is clear that more should be done tomaximize the advantages of a recycling-friendlyworld. The growth in aluminum usage in trans-portation applications, the decline in aluminumbeverage can recycling, and the increasingreliance of the domestic fabrication industry onsecondary aluminum have combined to createnew needs in both the materials design and pro-cessing realm.

Today (2007), in order to meet the perform-ance requirements of many alloy and productspecifications, much of the recycled metal mustbe “sweetened” with the more costly andenergy-intense primary metal before it can be

CHAPTER 9

Emerging Trends in Aluminum Recycling*Subodh K. Das, Secat, Inc.

*Adapted with permission of the Minerals, Metals andMaterials Society (TMS) from the article by Subodh K.Das, “Emerging Trends in Aluminum Recycling: Reasonsand Responses,” Light Metals 2006, TMS, 2006.




re-employed in many applications. The specialtyalloys required for a number of applicationsrequire such strict controls on impurities thatrecycled metal cannot be used without modifica-tion. The result is that, in many cases (exceptbeverage cans), recycled metal tends to be usedprimarily for lower-grade casting alloys andproducts. While a certain amount of this isacceptable, the recycling-friendly world will onlybe truly optimized when the recycle loop is closerto a closed loop within a number of product lines.

This chapter details the history and futureprojections for aluminum recycling, emphasiz-ing the increasing importance of mixed-scrapstreams in the makeup of secondary aluminum.For the most economical use of these scrapstreams, new approaches are needed to developacceptable materials processed to control prop-erties suitable for an expanded range of applica-tions. This chapter discusses how the aluminumenterprise, including industry, academia, andgovernment, can work together to meet theseimportant but aggressive targets and transformrecycling from strictly an environmental imper-ative to an economic development opportunity.

Objectives and Challenges

A robust recycling regime for aluminumwould include:

• Recycled products to approach the totalrequired for new consumption, with lessdependence on primary production

• Use of automatic sorting, shredding, and sep-aration technology to facilitate its reuse innew products

• Having a variety of existing and new alu-minum alloys with compositions suitablefor direct reuse of most recycled metal, sothe opportunities for direct use of the recy-cled, shredded, and sorted metal would beoptimized

• Developing a number of high-value applica-tions, such as beverage cans, into which therecycled metal would flow. In such situations,product made directly from the recycledmetal would readily meet both specificationcomposition and mechanical property limitsof the intended applications.

Among the key challenges to be met in creatingthis ideal recycling-friendly world are:

• Maximize recovery of used aluminum prod-ucts and components for recycling

• Identify more useful by-products to handleresidual elements; for example, aluminumalloys containing relatively high iron contentidentified for deoxidizing steel (de-ox)

• Automate and optimize the presorting,shredding, and separation technologies

• Broaden the number of available aluminumalloys whose specifications will readilyaccept recycled metal and will perform wellin high-quality, value-added products

Progress has been and is still being made inaddressing the first three challenges. Onesignificant new development in terms ofautomated sorting is the application of laser-induced breakdown spectroscopy (LIBS) forsorting of aluminum and aluminum alloys (Ref9.4). An overview of this technology, devel-oped and applied by Huron Valley Steel Corp.,is described in Chapter 8, “Identification andSorting of Wrought Aluminum Alloys” in thisbook.

The fourth challenge, the identification ofnew alloys that more readily accept recycledaluminum, has received little attention. How-ever, the potential economic and environ-mental benefits are sufficiently great for moreattention to this challenge. There are a numberof detailed challenges facing any effort toincrease the number of aluminum alloys andapplications suitable for direct production fromrecycled metal.

The Nature of Recycled Metal

As a starting point in considering thechallenges faced in directly using recycledaluminum scrap, it is appropriate to look atrepresentative compositions observed in suchmetal. Representative compositions of recycledaluminum are shown in Table 9.1 from presort-ing wrought and cast alloy scrap (Ref 9.4), sotwo samples were of wrought separations, twoof cast separations, and one mixed. These repre-sentative compositions illustrate several of thefundamental complications in directly reusingscrap aluminum:

• Even segregated wrought scrap can haverelatively widely varying compositions.Wrought lots 3 and 4 in Table 9.1, for exam-ple, have higher copper (from more 2xxxalloys) and higher zinc (from more 7xxxalloys) in the mix than did wrought lots 1 and2. It appears that auto bumper alloys such as

Chapter 9: Emerging Trends in Aluminum Recycling / 149

7029 and auto body sheet alloys such as 2036were more highly represented in wrought lots3 and 4.

• Some lots of wrought recycled metal (lots 1and 2) match existing wrought alloysreasonably well, for example, 3005, 3104,3105, and 6061, and can be readily reused;others, such as lots 3 and 4, will be moredifficult to use directly.

• Cast alloy scrap can vary greatly in composi-tion. Cast alloy scrap also differs signifi-cantly from wrought alloy scrap, notablywith higher total alloy content, higher siliconcontent, and, depending on which cast alloysare involved, higher copper (from 380.0 and390.0) and zinc (from 7xx.0 cast alloys).

• Compositions resulting from mixed wroughtand cast scrap will be more difficult to usedirectly because of their combinations ofhigher silicon, copper, and zinc.

With the exception of recycled beverage cans,most recycled aluminum involves a mixture ofalloys from a fairly wide variety of applications,including a selection of castings containingrather high percentages of silicon. While there isgenerally no problem recycling most of thismetal as castings, there is a significant challengein shredding, sorting, and, in some cases, furtherrefinement of the metal to achieve acceptable im-purity levels for products other than castings,including sheet, plate, forgings, and extrusions.

This is particularly true for any of the special-ized alloys produced today (2007), for example,those used in the aerospace industry whererequirements for exceptionally high ductility andtoughness are common. Such performancerequirements call for very tight composition con-trols on both iron and silicon. Impurity levels

above 0.15% Fe or 0.25% Si are unacceptable inpremium aerospace alloys such as 7050, 7055,and 7475. Similarly, some high-performanceautomotive alloys (e.g., 5457 and 6111) restrictboth silicon and iron to 0.40% maximum. Bothof these elements (iron and silicon) are difficultto control in recycled metal and tend to increasemodestly the more often the metal has beenrecycled.

Iron, in particular, can be a significantchallenge because of its tendency to increasegradually in metal recycled over and overagain, primarily from pickup from scrap-handling system equipment. As a result, iron isan ideal candidate for application to alternativeproducts, an excellent example being the useof high-iron-bearing aluminum as a deoxidiz-ing agent for steel production. It should benoted that elements other than iron may also beexpected to increase with repeated recyclingand to require special attention, for example,magnesium, nickel, and vanadium.

Cast Alloy Scrap. As noted earlier, recycledmetal from castings can often be used directly innew cast products, usually of the 3xx.0 and 4xx.0series, because the impurity limits tend to berelatively high, and almost all contain relativelyhigh silicon, which improves their flow charac-teristics in the dies. Examples of scrap-tolerantcasting alloys are shown in Table 9.2. Even theserelatively tolerant limits pose a challenge fordirect recycling reuse. For all except 336.0, forwhich no “Others” limit exists, the “Others” con-tents noted in scrap samples are higher thandesired. In the 4xx.x series, the tight magnesiumcontents will be a challenge. Nevertheless, cast-ing alloys as a whole have higher impurity limitsthan wrought alloys and will be more tolerant fordirect recycling.

Table 9.2 Sample of cast alloy scrap compositions Alloy Cu Fe Mg Mn Si Zn Others

B319.0 3.0–4.0 1.2 max 0.10– 0.50 0.8 max 5.5–6.5 1.0 max 0.50 max336.0 0.50–1.5 1.2 max 0.70–1.3 0.35 max 11.0–13.0 0.35 max . . .

C443.0 0.6 max 2.0 max 0.10 max 0.35 max 4.5–6.0 0.50 max 0.25 max

Table 9.1 Representative compositions of presorted wrought and cast scrap Lot Al Cu Fe Mg Mn Si Zn Others

Wrought 1 97.1 0.11 0.59 0.82 0.21 0.51 0.45 0.19 Wrought 2 96.7 0.30 0.60 0.60 0.20 0.90 0.50 0.10 Wrought 3 93.1 0.95 1.01 0.89 0.12 2.41 1.25 0.27 Wrought 4 93.1 1.20 0.70 0.70 0.30 2.60 1.20 0.20 Cast 1 83.5 4.40 1.10 0.40 0.30 8.0 1.90 0.40 Cast 2 86.0 3.90 1.00 0.10 0.20 6.30 2.30 0.30 Cast 3 88.4 2.50 0.75 0.58 0.26 5.18 1.27 1.09 Mixed wrought and cast 90.1 2.30 0.80 0.50 0.20 4.50 1.20 0.30


Wrought Alloy Scrap. The extent of the cur-rent opportunity as well as the challenge indirectly reusing recycled wrought alloy scrapwithout “sweetening” with primary metal ismuch greater. This can be illustrated by compar-ing the compositions in Table 9.2 with those ofseveral wrought commercial alloys already rec-ognized as good consumers of scrap (Table 9.3).

As noted earlier, material from wrought lots 1and 2 could reasonably be used in alloys such as3105 and 6061. Even for these alloys, the maxi-mum limit on “Others” may be a challenge,because that value for wrought 1 is slightly inexcess of the limit. Scrap from wrought lots 3and 4 could not be directly reused without flexi-ble impurity limits (see the section “DevelopingRecycling-Friendly Compositions” in this chap-ter). Material from the cast lots could not beused directly for any of these wrought alloys orany others, for that matter.

An additional elemental impurity problemneeds to be recognized, and that is the generaltrend for iron content in scrap reused over andover again to increase gradually, primarilythrough pickup from scrap-handling systemequipment. With only a few exceptions, iron isan impurity in wrought alloys today (2007) andis an ideal candidate for application to alternateproducts (e.g., a deoxidizing agent in steel pro-duction). Maximization of this capability willbenefit both the aluminum and steel industriesand add to the life-cycle benefits of aluminumoperations. Another possible approach to theincreased iron content is to make use of theaffinity of zirconium for iron, resulting in aheavy particle that sinks to the bottom ofcrucibles during processing. Combining thisiron-zirconium product with magnesium andperhaps other undesirable impurities such asnickel and vanadium may also improve theirimpact on the resulting recycling content.

Recycling Aluminum Aerospace Alloys (Ref 9.5)

As an example, consider the principal chal-lenges for effective recycling of aircraft. For

decades, thousands of obsolete aircraft havebeen sitting in “graveyards,” while the demandfor recycled aluminum continues to increase.The discarded aircraft provide a large source ofvaluable metal. However, cost-effective recy-cling of aircraft alloys is complex, becauseaircraft alloys are typically relatively high inalloying elements and contain very low levels of impurities to optimize toughness and otherperformance characteristics.

Thus, recycling of aluminum aerospace alloysrepresents a major challenge to both the alu-minum and aerospace industries. While therecycling of high percentages of aluminum frompackaging and automotive applications has beencommercialized and has become economicallyattractive (Ref 9.1–9.3), the unique compositionsand performance requirements of aerospacealloys have resulted in the delay of directlyaddressing techniques for cost-effectively recy-cling ignore those alloys. Aluminum remains themost economically attractive material fromwhich to make aircraft and space vehicles, andnew construction proceeds at a prodigious rate.However, the development of newer aircraftstructures has proceeded at such a pace that thou-sands of obsolete civil and military aircraft standidle in “graveyards” around the United States.Yet, it has been impractical to reuse the metal inthese planes because of the combination of thedifferences in compositions of older, obsolete air-craft and those of new aircraft, which often havespecial performance requirements for specializedalloy compositions.

Cost-effective means is the subject of a study(Ref 9.5) for recycling aluminum alloys used inthe production of private, civil, and militaryaircraft. The principal challenges that must bedealt with in creating this ideal aircraft-recyclingscenario include:

• Identifying decision options for dismantlingaircraft to simplify recycling

• Identifying and optimizing technologies forautomated shredding, sorting, and remeltingof those 2xxx and 7xxx alloys with relativelyhigh levels of alloying elements (sometimesin excess of 10%)

Table 9.3 Sample of wrought alloy scrap compositionsAlloy Cu Fe Mg Mn Si Zn Others

3005 0.30 max 0.7 max 0.20–0.8 1.0–1.5 0.6 max 0.25 max 0.15 max 3104 0.8 max 0.6 max 0.8–1.3 0.8–1.4 0.6 max 0.25 max 0.15 max 3105 0.30 max 0.7 max 0.20–0.8 0.30–0.8 0.6 max 0.40 max 0.15 max 6061 0.15–0.40 0.7 max 0.8–1.2 0.15 max 0.40–0.8 0.25 max 0.15 max


• Identifying the range of representative com-positions likely to be obtained from recyclingaircraft components, dependent on theamount of presorting that proves practical

• Identifying the combination of performancerequirements and compositions that wouldmake useful aircraft components from recy-cled metal, even though they may not achievethe highest attainable levels of toughness

• Identifying useful by-products to handle ele-mental residue unable to be used in recycledmetal, for example, iron

These challenges include many of the generalissues of aluminum recycling. As noted,progress is being made in the application ofLIBS for shredding and sorting of somealuminum alloys. In general, however, little ornothing has yet been done to apply recyclingtechnology to shredding, sorting, and reuse ofrecycled metal from obsolete aircraft and spacevehicle components. Implementation issuesaddressed in Ref 9.5 include:

• Dismantling and presorting strategies• Automated shredding, sorting, and remelting • Identifying the resulting compositions of

recycled aircraft components• Options for reuse of the metal from recycled

aircraft components in new aircraft • Options for reuse of the metal from recy-

cled aircraft components in nonaircraftapplications

• Options for reuse of the metal from recycledaircraft components in aluminum castings

Of these issues, the latter three options aredescribed as follows.

Options for Reuse of the Recycled AircraftComponents in Aircraft. Table 9.4 illustratescomposition examples from presorting of 2xxxand 7xxx alloys. Assuming that these estimatedcompositions are reasonably correct, it wouldappear that the resultant alloys could be used fora number of noncritical aircraft components,such as stiffeners, flaps, and other relatively low-to-moderately stressed components made ofsheet, plate, or extrusions. These may be usedin private, civil, and many military aircraft.Typically, these would be components that arenot designed based on fracture mechanics con-cepts employing fatigue crack growth rates andfracture toughness parameters. Alloys for frac-ture-critical areas may still have to be fabri-cated using primary metal. Further study isneeded to estimate the percentage of aircraftcomponents that do not require fracture-criticaldesign and whether it is broad enough to justifythe reuse of compositions likely to result fromrecycling.

Options for Reuse of the Recycled AircraftComponents in Nonaircraft Applications. Ifthe use of these compositions for non-fracture-critical components in new aircraft is too tightlylimited, that is, if the number of non-fracture-critical components in civil and military aircraftis not large enough to justify reuse of the recy-cle compositions, it is useful to look at the otheropportunities that may exist for use of thecompositions (Tables 9.5, 9.6).

To aid in addressing that question, the compo-sitions of several wrought 2xxx and 7xxx alloysused in other applications (including 2014, alsoan aircraft alloy) are presented in Table 9.7.Comparing compositions in Tables 9.5 and 9.7

Table 9.5 Potential compositions of somerecycled aircraft alloys, assuming presortingAlloy Al Cu Fe Mg Mn Si Zn Others

R2xxx ~93 4.4 0.5 1.0 0.7 0.5 0.1 0.2R7xxx ~90 2.0 0.4 2.5 0.2 0.2 6.0 0.2

Table 9.6 Potential composition of some recycled aircraft alloys, assuming no presortingAlloy Al Cu Fe Mg Mn Si Zn

R2 + 7xxx ~92 3.0 0.4 1.8 0.4 0.4 3.0

Table 9.4 Nominal compositions of some 2xxx and 7xxx alloysAlloy Al Cu Fe Mg Mn Si Zn

2014 ~93 4.4 0.7 max 0.50 0.8 0.8 0.15 max2214 ~93 4.4 0.3 max 0.50 0.8 0.8 0.15 max2024 ~93 4.4 0.5 max 1.5 0.6 0.5 max 0.25 max2324 ~94 4.1 0.12 max 1.5 0.6 0.1 max 0.15 max7050 ~89 2.3 0.15 max 2.2 0.1 max 0.12 max 6.27075 ~90 1.6 0.5 max 2.5 0.3 max 0.4 max 5.67475 ~90 1.6 0.12 max 2.2 0.06 max 0.1 max 5.77178 ~89 2.0 0.5 max 2.8 0.3 max 0.4 max 6.8


illustrates that it may be possible to reuserecycled aircraft metal in certain other products.Further study of this option is justified.

Options for Reuse of the Recycled AircraftComponents in Castings. Other opportunitiesfor the use of recycled metal from aircraft mayinclude aluminum alloy castings, especiallythose of the 2xx.0 and 7xx.0 series, aluminum-copper and aluminum-zinc, respectively. Exam-ples of such alloys are included in Table 9.8.Even these relatively tolerant limits posechallenges for direct reuse of recycled metal.Nevertheless, opportunity for study of new alloyoptions remains, and the properties of the alloysin Table 9.5, when produced as casting, shouldbe studied.

Alloys Designed for Recycling

As noted, an ideal component of resourcemaximization in recycling would be the avail-ability of a larger number of aluminum alloyssuitable for a wide variety of applications thatdo not require any extra purification step beforethey are recycled into high-performance prod-ucts. That is largely the case with beveragecans, if the recycle scrap is not mixed with anyother material.

Such an approach calls for “tailored” alloys.The goal of identifying new recycling-friendlyaluminum alloy compositions is to increase theopportunities to directly, or with only minormodification, reuse recycled scrap aluminumproducts. Such an approach requires composi-tions with relatively broad specification limitson major alloying elements, such as copper andmagnesium, plus more tolerant (i.e., higher)limits on iron, silicon, and other impurities,without significant restriction on performancecharacteristics for many applications.

Full development of this approach requiresseveral important steps. Application of LIBS,developed by Huron Valley Steel Corp., toscreen scrap with certain combinations of thedesired elemental additions may permit relax-ation of the broadest interpretation of the afore-mentioned guidelines, but it is beneficial to lookat the most useful long-term trends when suchtechnology may not be available. Adopting theapproach of alloy optimization for recyclingrequires several steps, which are potentiallyphases in a development program. The phasesmay include:

• Identify with increasing precision the rangeof expected current and future recycledmetal content, using feedback from organi-zations that are already capitalizing on theeconomics of recycling. Perform a massbalance to the extent practical, indicating therelative volumes of various scrap composi-tions to be expected

• Identify approximately five to seven basiccandidate compositions of alloys that wouldaccept recycled metal directly and haveacceptable/desirable performance character-istics for a wide variety of applications,including structural components for bridgesand buildings, high-temperature applica-tions, and architectural usage

• Evaluate the performance of these candidatealloys in representative production lots,including, in particular, the following, alongwith the usual tensile and design properties,in order to assess their abilities to meet therequirements of representative applicationsas compared to existing alloys (Ref9.6–9.11):

a. Atmospheric corrosion resistance b. Stress-corrosion crack growth

Table 9.7 Nominal compositions of some 2xxx and 7xxx alloys used in nonaerospace applicationsAlloy Application Cu Fe Mg Mn Si Zn

2014 Railroad; truck bodies 4.4 0.7 max 0.50 0.8 0.8 0.15 max2017 Rivets 4.0 0.7 max 0.60 0.7 0.5 0.25 max7129 Auto bumpers 0.7 0.3 max 1.6 0.1 max 0.15 max 4.7

Table 9.8 Some aluminum casting alloy compositions Alloy Al Cu Fe Mg Mn Si Zn Others

201.0 ~95 4.6 0.15 max 0.35 0.35 0.10 0.25 0.05 max242.0 ~94 4.1 0.6 max 1.4 0.10 max 0.6 max 0.10 0.05 max295.0 ~94 4.5 1.0 max 0.03 0.35 max 1.1 0.25 0.05 max710.0 ~93 0.5 0.5 max 0.7 0.05 max 0.15 max 6.5 0.05 max713.0 ~91 0.7 1.1 max 0.35 0.6 max 0.25 7.5 0.05 max


c. Toughness, with tear tests and/or fracturetoughness tests (for thick sections)

d. Formability tests, with bulge, minimumbend, and hemming tests

It is recognized that there will be some nega-tive effects; the question is the degree to whichsuch alloys are still useful for high-volume ap-plications.

Developing Recycling-Friendly Compositions

Based on what is known about recycledaluminum metal, preliminary candidates forrecycling-friendly alloys can be considered.Attention is given here primarily to wroughtalloy compositions, because, as noted earlier,casting alloys already can be produced in ratherlarge quantities from recycled cast products.The greatest challenge is direct use of recycledwrought products. The rationale is to definealloy specifications that can be readily metusing recycled aluminum with no addition ofprimary metal. The target applications for thenew recycling-friendly aluminum alloys includemany of the same as for their existing counter-parts, with tighter limits. Examples may include:

• 3xxx: heat-exchanger tubing, chemical piping • 4xxx: forged or cast engine parts • 5xxx: tankage plate; housing components • 6xxx: extruded structural components

While suitable performance requirements withthe higher level of impurities may not beentirely successful, steps in that direction willenable the aluminum industry to maximize itsrecycling opportunities. Therefore, addressingthe challenge of a new approach to alloy designis warranted.

Candidate Compositions for Recycling-Friendly Aluminum Alloys. In developingcandidate compositions, the following ratherbasic guidelines have been used:

• For major alloying elements in a particularseries (e.g., copper in the 2xxx series, siliconand magnesium in the 6xxx series, etc.), pro-pose relatively broad specification limits.Select alloying elements that are commonlyand successfully used in alloys of the vari-ous series, for example, 2024 or 2219 in the2xxx series, or 7005 in the 7xxx series.

• For limits on elements not usually added in-tentionally or on undesirable impurities (e.g.,iron, nickel, and vanadium), propose moretolerant (i.e., higher) limits to the degreepotentially practical, to the levels of thoseelements typically found in recycled metal.

Table 9.9 shows six preliminary candidatewrought alloy compositions that may reasonablybe made from recycled and shred-sortedwrought products with, at minimum, the addi-tion of some alloying elements. In this initial list,one composition has been selected from eachmajor alloy series. Other candidates may well bedevised by adjustments in the major alloyingelements and/or the addition of other minoralloying elements.

These are only preliminary candidates, and itis recognized that the completion of phase 1(described in the preceding section as identify-ing to a higher precision the compositions ofincoming scrap, current and future) may resultin significant changes in these candidate alloys.It may also lead to a focus on several differentcandidates from specific series that show maxi-mum fit with the incoming metal (e.g., the 3xxx,5xxx, and 6xxx series representing the highestvolumes of recycled metal).

The application targets for these candidatecompositions are much the same as their exist-ing counterparts with tighter limits, recognizingthat these are not likely to be suitable for themore fracture-critical items. However, thepossibility exists that they may perform quitesatisfactorily in such applications as chemicalplant piping (A-2xxx), heat-exchanger tubing(B-3xxx), forged or cast engine parts (C-4xxx),rolled and extruded structural components

Table 9.9 Candidate alloys for recycling Alloy Si Fe Cu Mn Mg Zn Others

A (2xxx) 0.7 0.6 5.5–7.0 0.2–0.4 0.7 0.5 0.3 B (3xxx) 0.7 0.6 0.4 1.0–1.5 0.8–1.5 0.5 0.3 C (4xxx) 10.0–14.0 1.0 0.5–1.5 0.3 0.8–1.5 0.5 0.3 D (5xxx) 0.7 0.6 0.3 0.05–0.35 2.0–3.0 0.5 0.3 E (6xxx) 0.3–1.0 0.6 0.3 0.3 0.4–1.0 0.5 0.3 F (7xxx) 0.5 0.6 0.5–1.2 0.3 2.0–2.8 4.0–6.0 0.3


(D-5xxx and E-6xxx), and even for noncriticalaircraft components (F-7xxx).

The application of LIBS technology to screenscrap with certain combinations of the desiredelemental additions may permit relaxation ofthe broadest interpretation of the aforemen-tioned guidelines, but it is beneficial to look atthe most useful long-term trends when suchtechnology may not be widely available.

Unialloy. Another approach that may beconsidered, and one that has been studied toconsiderable degree in the past, is the develop-ment of one or two “unialloys,” that is, alloysthat meet all the requirements for a largeapplication, such as aluminum beverage pack-aging, automotive components, or architecturalcomponents. The aforementioned approach ofselecting “recycle-optimized” alloys from eachseries may lead to a master list of severalrecycling-friendly alloys.

The unialloy concept was first developed byGolden Aluminum (Ref 9.12) in the late 1980sand focused on alloy AA5017 with 2% Mg and0.7% Mn. The concept was derived from theidea that unialloy had an average weighted com-position of can body alloy AA3004 and can endalloy AA5182. This recycling idea achievedlimited application due to economic and com-mercial factors. In the current environment ofrising prices of primary aluminum and itsalloying elements, such as magnesium, man-ganese, and copper, coupled with the societaldesire for enhancing recycling rates of products,it is time to rethink applicability and commer-cialization of the unialloy concept.

This has proven difficult to achieve becauseof the diverging performance requirements ofdifferent applications. Even within autobodypanels, for example, the differing requirementsfor dent resistance in outer panels and opti-mized formability for inner panels continue tolead to two different types of alloys being used (for example, 6111 heat treated for high outerpanel dent resistance, and 5754 annealed formaximum formability for inner panels).

Some Caveats. The large bank of alloydesign experience in the aluminum industrymay well result in skepticism about the proba-bilities of success and the seeming backwardmovement in the aforementioned approach ofpermitting higher levels of impurities in newalloy candidates. There is indeed some reasonfor the skepticism, because it is well known thatoptimizing fracture toughness, for example,requires tight impurity controls, especially on

iron and silicon. So, it is probably appropriate toacknowledge at the outset that it may be impos-sible to reach the stage where recycled metalwill be used untreated for all aluminum alloyproducts, such as fracture-critical-componentaerospace wings and wing spars; the fracturetoughness requirements on these are simplytoo stringent to be met without tight impuritycontrols.

However, aerospace applications account foronly approximately 8% of the total approxi-mated 14 to 18 � 106 metric tons (30 to 40 bil-lion lb) of aluminum used annually. Therefore,the adoption of a recycling-friendly alloysystem applicable to most other applicationswill still have a very great economic andecological benefit in world consumption.

The performance requirements of many high-volume aluminum products, such as buildingand highway structures and chemical industrycomponents, may well be satisfied by alloyswith higher levels of impurities than presentlymandated. In fact, most composition limits wereset when the greater volume of production wasprimary metal and there was no need for higherimpurity levels. The limits were not set byperformance requirements but by anticipatedincoming metal compositions.

Finally, it must be acknowledged that attemptsat obtaining suitable performance requirementswith higher levels of impurities may not besuccessful. However, any step in that directionwill better enable the aluminum industry tomaximize its recycling opportunities; therefore,the challenge of a new approach to alloy designis warranted.

Conclusions and Looking Ahead

There are significant economic and ecologicaladvantages in maximizing the recycling rates ofaluminum and the ready reuse of recycledmetal. Several important conclusions for thealuminum industry, both in the United Statesand throughout the world, include:

• Avenues for recovery of aluminum scrapfrom as many products as possible shouldcontinue to be exploited; there are massiveeconomic, energy, and ecological advan-tages to the communities and to the alu-minum industry.

• Development and application of enhancedshredding and sorting technologies shouldcontinue.


• A focus on the most cost-effective remeltingprocesses is justified, including the possibil-ity of combining iron with zirconium andother impurities such as nickel and vanadiumin high-density particles that could be easilyseparated from the melt.

• The production of alternative products suchas aluminum-iron deoxidizing agents shouldbe pursued to use that part of recycled alu-minum products that cannot cost-effectivelybe used in the production of new aluminumalloys, to the benefit of both the steel and thealuminum industries.

• The most overlooked aspect of maximizingrecycled metal appears to be the develop-ment of new alloys tailored to meet compo-sition and performance criteria whenproduced directly from recycled metal. Astudy of the type suggested in the section“Alloys Designed for Recycling” in thischapter should be carried out to potentiallyadd to the number of alloys available fordirect recycling. Some candidate composi-tions have been suggested for such alloys,with the intent to broaden this part of the dis-cussion and perhaps lead to interesting andeconomically attractive components to max-imize recycling efficiency and effectiveness.

Looking ahead, the challenge is to identifythe most successful means of implementing theadditional studies and development programsneeded to maximize the benefits of recycling.Accordingly, one step may be the formation ofan aluminum recycling consortium to consideroptions and opportunities to increase the overalleffectiveness and efficiency of aluminum recy-cling and its benefits. The consortium couldinclude representatives of:

• The municipalities and their representativesat the first line of collecting recycled metal

• The recyclers themselves (e.g., members ofthe Aluminum Association Recycling Divi-sion)

• Fabricators and distributors of recycledproducts (e.g., auto and beverage can pro-ducers)

• End-users of such recycled products

In this regard, Secat, Inc. is in the process offorming an aluminum recycling consortium tofurther the goals outlined in this chapter. Thefirst two charges of the consortium should be to:

• Identify all means of maximizing theamount of aluminum products entering the

recycling chain. Carry out the study of alloydesign optimization for recycling asdescribed in the section “Alloys Designedfor Recycling,” identifying the most likelyhigh-volume recycling compositions, carry-ing out the resultant mass balance, and fine-tuning several candidate compositions thatwould take advantage of the anticipated recycling content

• Explore opportunities to improve thequality of the melt by removal of high-density particles formed by the combinationof iron with zirconium and possibly otherhigh-density elements such as nickel andvanadium.

The next logical step would be a completeevaluation of the physical, mechanical, corro-sion, and fabricating characteristics of whateveroptimized recycling candidate alloys are gener-ated. Successful application of these variousapproaches to maximizing the cost-effective-ness and efficiency of recycling processesshould lead to increased opportunity to extendthe life-cycle advantages of aluminum alloys,increase the usefulness of directly recycledalloys, and therefore increase the amount ofmetal that is directly reused without the additionof primary metal.

ACKNOWLEDGMENTS

The author gratefully acknowledges valuableinput and advice from Gil Kaufman, Vice Presi-dent of Technology, The Aluminum Association(Retired), Dr. John A.S. Green, Vice Presidentof Technology, The Aluminum Association(Retired); Dr. Warren Hunt, President, TMS;and Dr. Wayne Hayden, Oak Ridge NationalLaboratory (Retired).

REFERENCES

9.1. W.T. Choate and J.A.S. Green, Modelingthe Impact of Secondary Recovery (Recy-cling) on the U.S. Aluminum Supply andNominal Energy Requirements, TMS, 2004

9.2. Aluminum Industry Technology Roadmap,The Aluminum Association, Washington,D.C., 2003

9.3. R.A.P. Fielding, Recycling Aluminum, Es-pecially Processing Extrusion Scrap, LightMet. Age, Aug 2005, p. 20–35


9.4. A. Gesing, L. Berry, R. Dalton, and R.Wolanski, Assuring Continued Recycla-bility of Automotive Aluminum Alloys:Grouping of Wrought Alloys by Color,X-Ray Absorption and Chemical Com-position-Based Sorting, TMS 2002 AnnualMeeting: Automotive Alloys andAluminum Sheet and Plate Rolling andFinishing Technology Symposia, Feb18–21, 2002 (Seattle, WA), Minerals,Metals and Materials Society (TMS),2002, p. 3–17

9.5, S.K. Das and J.G. Kaufman, Reycling Alu-minum Aerospace Alloys, TMS, 2007

9.6. Properties and Selection: NonferrousAlloys and Special–Purpose Materials, Vol2, ASM Handbook, ASM International,1990

9.7. J.E. Hatch, Ed., Properties and PhysicalMetallurgy, American Society for Metals,1984

9.8. J.R. Davis, Ed., Aluminum and AluminumAlloys, ASM International, 1993

9.9. J.G. Kaufman, Ed., Properties ofAluminum Alloys—Tensile, Creep, andFatigue Data at High and Low Tempera-tures, ASM International, 1999

9.10. J.G. Kaufsman, Fracture Resistance ofAluminum Alloys—Notch Toughness,Tear Resistance, and Fracture Toughness,ASM International, 2001

9.11. J.G. Kaufman and E.L. Rooy, AluminumAlloy Castings—Properties, Processes,and Applications, ASM International, 2004

9.12. D. McAuliffe and I.M. Marsh, LightMetals 1989, p739–741

THIS CHAPTER, based on a 2005 revisedversion of a 2003 report (Ref 10.1) for the U.S.Department of Energy, provides reliable andcomprehensive statistical data over the period1960 to 2005 for the evaluation of energy trendsand issues in the U.S. aluminum industry. Itshould be noted, however, that these trends needcareful interpretation because they incorporateunusual circumstances of a single year, that is,2001. During the summer of 2001, the extensiveheat wave in the western United States producedan increased demand for electricity. Simultane-ously, the ability to generate hydroelectric powerwas reduced due to historically low snow packsin the Columbia River basin and new regulationsmandating the spill of water to aid migratingsalmon. The combination of high electricitydemand and limited water supply contributed toa significant increase in the market price of elec-tricity during this time. This price increase in thePacific Northwest made it more economical foraluminum smelters to stop metal production andsell back power from their low-cost, fixed-priceelectric contracts to aid in minimizing the short-fall in energy supply. As a result, the majority ofaluminum smelting capacity in the Pacific

Northwest, representing approximately 43% ofall U.S. primary aluminum capacity, shut down.

The dramatic and relatively quick shutdownof a substantial amount of the U.S. primaryaluminum capacity can make certain trendnumbers appear misleading. Throughout thischapter, percentages are given for 10-yeartrends in various sectors throughout the alu-minum industry. For example, U.S. primaryaluminum production has a 10-year annualgrowth rate of –4.2%, decreasing from 3375thousand metric tons in 1995 to 2480 thousandmetric tons in 2005. These numbers wouldseem to imply a steady decline throughout thespecified time period, but this is not the case.In fact, the total drop in primary productionover this time period (895 thousand metrictons) is less than the 1031 thousand metric tondrop from 2000 (3668 thousand metric tons) to2001 (2637 metric tons) (Ref 10.2). This one-year drop coincides with the majority of thePacific Northwest shutdown. Other primaryproduction numbers in this chapter have simi-lar trends, and this drastic shutdown should betaken into account when considering thesenumbers.

It is currently too early to accurately assess thelong-term impact of these sudden shutdowns andchanging conditions on the aluminum industry. Itremains to be seen whether the shutdowns will

CHAPTER 10

U.S. Energy Requirements for Aluminum Production: Historical Perspective, Theoretical Limits, andNew Opportunities*William Choate, BCS Inc.

*Adapted with permission from Ref 10.1. The original refer-ence 10.1 was prepared for Industrial Technologies Program,Energy Efficiency and Renewal Energy, and funded by theU.S. Department of Energy.




lead to a permanent decline of primary metalproduction in the Pacific Northwest, or whetherthe industry will emerge based on higher alu-minum prices, new power contracts, or additionalself-generated power capacity and energy effi-ciency improvements. Between October 8, 2004,and January 2006, �110,000 metric tons/yr ofthe Pacific Northwest capacity has come back online. It is difficult to judge whether or not thiswill be the common trend. Whatever the future ofthe industry, it is clear that the local and globalpressures to increase overall energy efficiencywill determine its vitality. The energy-efficiencyopportunities discussed in this chapter are perti-nent to the future of the aluminum industry.

Another recent trend to make note of isChina’s rapid growth in both primary and sec-ondary aluminum production. While the UnitedStates has slipped from being the largest pro-ducer of primary metal in 2000 to being thefourth largest in 2005, China has taken theworldwide lead in primary aluminum produc-tion, producing 7,800,000 metric tons of pri-mary aluminum in 2005. Additionally, Chinahas been aggressively buying up much of thealuminum scrap supply. In 2003, the UnitedStates exported 568,721 metric tons of scrapand dross, with 43% (244,374 metric tons)going to China (Ref 10.2). These trends aremanifested in the slight decline of secondaryaluminum production within the United States.

Summary

The United States aluminum industry,processed 10.7 million metric tons of metal andproduced over $40 billion in products andexports in 2005. It operates more than 400plants in 41 states and employs more than145,000 people. Aluminum impacts every com-munity and person in the country, through eitherits use and recycling or the economic benefits ofmanufacturing facilities.

Energy reduction in the U.S. aluminum in-dustry is the result of technical progress and thegrowth of recycling. These two factors havecontributed 22 and 42%, respectively, to thetotal 64% energy reduction over the past 45years. By many measures, aluminum remainsone of the most energy-intensive materials toproduce. Only paper, gasoline, steel, and ethyl-ene manufacturing consume more total energyin the United States than aluminum. Aluminumproduction is the largest consumer of energy on a

per-weight basis and is the largest electric energyconsumer of all manufactured products. The U.S.aluminum industry directly consumes 42.3 � 109

kWh (0.16 quad, or 0.16 � 1015 Btu) of electric-ity annually, or 1.1% of all the electricity con-sumed by the residential, commercial, andindustrial sectors of the U.S. economy. This isequivalent to the electricity consumed by4,826,000 U.S. households annually, or theenergy equivalent of nearly 50 million barrelsof oil.

The aluminum industry has large opportunitiesto further reduce its energy intensity. The annualsum of all the energy required in the productionof aluminum metal and products in the UnitedStates is equivalent to 173 � 109 kWh (0.59quad). The difference between the gross annualenergy required and the theoretical minimumrequirement amounts to over 141 � 109 kWh(0.48 quad). This difference is a measure of thetheoretical potential opportunity for reducingenergy consumption in the industry, althoughachievable cost-effective savings are smaller.

This chapter provides energy performancebenchmarks for evaluating new process develop-ments, tracking progress toward performance tar-gets, and facilitating comparisons of energy use.It provides a basic description of the processesand equipment involved, their interrelationship,and their effects on the energy consumed andenvironmental impact of manufacturing alu-minum and aluminum products. This knowledgecan help identify and understand process areaswhere significant energy reductions and environ-mental impact improvements can be made.

This chapter examines and carefully distin-guishes between the actual on-site energy con-sumption values and gross or tacit energy values.The tacit or gross energy value accounts for thegeneration and transmission energy losses asso-ciated with electricity production, the feedstockenergy of fuels used as materials, and the processenergy used to produce fuels. On-site energyimprovements provide concomitant gross energysavings.

Primary aluminum is produced globally bymining bauxite ore, refining the ore to alumina,and combining the alumina and carbon in anelectrolytic cell to produce aluminum metal.Secondary aluminum is produced globallyfrom recycled aluminum scrap. Primary andsecondary aluminum metal are cast into largeingots, billets, T-bar, slab, or strip and thenrolled, extruded, shape cast, or otherwise formedinto the components and products used daily.

Chapter 10: U.S. Energy Requirements for Aluminum Production / 159

Fig. 10.1 Aluminum industry flow diagram

Figure 10.1 shows the major processing opera-tions required to produce aluminum and alu-minum products. This chapter examines theseprocesses and the energy they require.

Note on Tacit Energy Consumption. Acomplete accounting of the energy consumed inthe production of any product should includethe energy required to produce the fuels and

electricity that are used plus the feedstockenergy associated with any fuels that are used asmaterials (e.g., carbon used in anode production).This chapter, in many places, reports energy as aset of two numbers, for example, 1.0 (2.0tf) kWh.The first value represents the energy consumedwithin a facility (on-site energy consumption,or primary energy), while the value with “tf”


superscript (tacit-feedstock energy) is a measureof energy that includes the energy used to pro-duce and transmit the energy consumed within afacility and the raw material feedstock energy.This adjustment is very significant in processesthat consume electricity or use fuels as materials.

The U.S. average grid connection requiresapproximately 3 kWh (9890 Btu) of fuel energyto deliver 1 kWh (3412 Btu) of electrical energy(Table 10.1). The average U.S. grid is suppliedwith approximately 7% hydroelectric genera-tion. The aluminum industry is located nearlow-cost power sources, and many of these arehydroelectric. The average U.S. primary alu-minum plant connection is 39% hydroelectricgeneration. If it is assumed that the heat rate forhydroelectric power is 3412 Btu/kWh, then theaverage primary aluminum plant grid connectionrequires 2.2 kWh (7570 Btu) of fuel energy todeliver 1 kWh (3412 Btu) of electrical energy. Itshould be noted that values reported in this chap-ter use the U.S. average grid connection values.The use of U.S. average grid values results in ahigher reported tacit energy consumption valuethan the actual tacit value required for primaryaluminum production. The advantage of report-ing primary production energy consumptionbased on the U.S. average grid is that it allowseasier and same-basis comparisons to other U.S.manufacturing industries.

Aluminum Production and Energy Consumption

Aluminum is an essential material for modernmanufacturing. It is a lightweight, high-strength,corrosion-resistant metal with high electrical andthermal conductivity, and it is easy to recycle.The U.S. aluminum industry is the largest in theworld in terms of consumption. The U.S. alu-minum industry used 9,266,000 metric tons ofmetal in 2005 to produce an enormous variety

of products. The U.S. per capita consumptionwas 66 lb. The industry operated more than 400plants in 41 states and employed over 145,000people to make aluminum products (Ref 10.2).These products are shipped to thousands ofbusinesses in the United States, from which theyare distributed or are incorporated into otherproducts. Aluminum, per unit mass, is the mostenergy-intensive material produced in largequantities in the United States. Only paper, gaso-line, steel, and ethylene manufacturing consumemore total energy for manufacturing in theUnited States than aluminum (Appendix A).

Research and development (R&D) effortsto reduce energy consumption are important,because energy consumption correlates to manu-facturing economics, environmental impact, andU.S. dependence on imported energy sources.Identifying process areas where opportunitiesfor energy-use reduction exist and applyingresources to capture these opportunities willbenefit the industry and the nation.

The science and technologies associated withthe production of aluminum and aluminumproducts are complex. This report attempts toprovide the reader with a basic understanding ofthe science, technology, and energy usage of thealuminum industry. More detailed books (Ref10.3, 10.4) are available for the reader whorequires further in-depth study of the subject.

Energy and Environmental Overview. Theenergy consumption and environmental effectsassociated with product manufacturing and useare important measures of the impact of theproduct on society. Energy consumption andenvironmental-impact measures are becomingkey decision tools for consumers and corpora-tions when choosing a product. In the nearfuture, manufactured products will compete notonly on price and performance but also on theirimpact on society.

Aluminum manufacturing is energy-intensive,and approximately one-third of the cost to

Table 10.1 Electric tacit energy and emission data for fuels used for aluminum productionU. S. primary aluminum capacity Heat rate (2005), Carbon emission coefficient,

Electrical energy source Metric tons % Btu/kWh Mt/Qbtu

Hydro 1,633,696 39.4 3412 0Coal 2,415,826 58.2 10,303 21.25Oil 8,245 0.2 10,090 19.08Natural gas 31,120 0.8 7937 12.50Nuclear 60,112 1.4 10.420 0Total 4,149,000 100.0

Weighted averages based on aluminum capacity 7570 12.51Average U.S. grid 9890 13.56

See Appendix C for additional details


produce aluminum from ore is associated withthe use of energy and environmental compliance.The aluminum industry, in the past 40 years,reduced its overall energy intensity by nearly64% (Table 10.2). However, even with the largereduction in energy intensity, the industry con-sumes nearly three times the theoretical mini-mum energy required. Significant opportunitiesfor further energy improvements still remain.

Technologies, practices, and product use deter-mine the energy consumption and environmental

impact of aluminum. Many of the current tech-nologies and practices used to produce aluminummetal and aluminum products are mature. Newtechnologies and practices are being proposedand studied to improve aluminum manufacturingfrom an energy and environmental standpoint.The history and explanation of current state-of-the-art technologies and practices are presentedso the reader can appreciate the values and ben-efits that new technologies or practices maybring to the aluminum industry. Current U.S.

Table 10.2 Impact of secondary metal production on energy requirements for U.S. aluminum production

U.S. secondaryU.S. primary aluminum Market Total energy Effective

aluminum production, Smelting production, percent of ore-to-metal to smelting energy to produce thousand metric energy, thousand metric secondary produce combined combined metals,

Year tonnes kWh/kg tonnes metal, % metals, kWh/kg kWh/kg

1960 1828 23.1 401 18 67.6 19.21961 1727 22.9 445 . . . 65.1 18.51962 1921 22.7 533 . . . 63.6 18.11963 2098 22.6 601 . . . 62.7 17.91964 2316 22.4 648 . . . 62.6 17.81965 2499 22.3 774 . . . 60.7 17.31966 2693 22.1 833 . . . 60.3 17.21967 2966 21.9 821 . . . 61.4 17.51968 2953 21.8 935 . . . 59.1 16.81969 3441 21.6 1067 . . . 58.9 16.81970 3607 21.4 937 21 60.7 17.31971 3561 21.0 1004 . . . 58.6 16.71972 3740 20.6 1022 . . . 57.9 16.51973 4109 20.2 1127 . . . 56.8 16.11974 4448 19.9 1163 . . . 56.2 16.01975 3519 19.5 1121 . . . 52.8 15.01976 3857 19.1 1334 . . . 50.7 14.51977 4117 18.7 1456 . . . 49.4 14.11978 4358 18.3 1518 . . . 48.5 13.81979 4557 17.9 1612 . . . 47.3 13.51980 4653 17.5 1577 25 46.8 13.31981 4489 17.4 1790 . . . 44.4 12.71982 3274 17.2 1666 . . . 40.9 11.81983 3353 17.1 1773 . . . 40.1 11.51984 4099 16.9 1760 . . . 42.4 12.11985 3500 16.8 1762 . . . 40.0 11.51986 3039 16.6 1773 . . . 37.8 10.91987 3347 16.5 1986 . . . 37.2 10.71988 3945 16.3 2122 . . . 38.2 11.01989 4030 16.2 2054 . . . 38.5 11.11990 4048 16.1 2393 37 36.3 10.51991 4121 16.0 2286 . . . 36.9 10.61992 4042 15.9 2756 . . . 34.0 9.81993 3695 15.8 2944 . . . 31.7 9.21994 3299 15.7 3086 . . . 29.3 8.61995 3375 15.6 3188 . . . 29.0 8.51996 3577 15.5 3307 . . . 29.1 8.51997 3603 15.4 3547 . . . 28.1 8.21998 3713 15.3 3442 . . . 28.7 8.41999 3779 15.2 3695 . . . 27.8 8.12000 3668 15.1 3450 48 28.1 8.22001 2637 15.0 2970 . . . 25.6 7.52002 2705 14.9 2927 . . . 25.9 7.62003 2704 14.8 2820 . . . 26.2 7.72004 2517 14.6 3025 . . . 24.1 7.12005 2480 14.4 2990 55 23.7 7.0

The nominal U.S. energy required to produce aluminum metal has been rapidly declining as secondary aluminum production has grown Secondary aluminum requiresonly 6% of the energy necessary to manufacture primary aluminum (Appendix E, Table E.6). The total U.S. ore-to-metal primary energy values include: refining of halfthe alumina supply, anode manufacture, electrolysis, and ingot casting. Combining the energy requirements for U.S. production of primary and secondary metals lowersthe average energy associated with U.S. aluminum metal from over 40 kWh/kg for primary alone to approximately 21 kWh/kg for the combined metals.


production levels, historical production levels,and projected growth rates of aluminum arepresented. These production values are needed tomeasure the magnitude of the impact of a changein technology or practice. The energy and envi-ronmental impacts from the use of aluminumproducts are generally low and, in some applica-tions such as transpontation, may be significantlybetter than the impacts of alternative materials.As significant as these impacts are on the use ofaluminum products, they are beyond the scopeof this chapter.

The greatest impact on the future energy inten-sity of aluminum has been the structural changein the industry itself. More than 55% of the alu-minum produced by U.S. industry in 2005 camefrom recycled material. In 1960, recycled mate-rial was used to generate less than 18% of U.S.-produced aluminum (Table 10.3). Recoveringaluminum from wastes and scraps requires lessthan 6% of the energy of aluminum productionfrom bauxite mining (Table 10.4). This reportexamines how “urban mining” (recycling) willcontinue to change the structure of the aluminumindustry and continue to lower the overall energyassociated with aluminum production. Recyclingis the largest contributor to the reduction of theenergy intensity of U.S-produced aluminum.

Aluminum is an “energy bank” in that nearlyall of the original energy stored in the metal canbe recovered again and again every time theproduct is recycled. Small fractions of the recy-cled metal are lost to oxidation (melt loss) andentrapment in purifying fluxes (dross) duringthe recycling process. Aluminum can be recy-cled indefinitely, allowing this saved energy tobe collected again and again.

Greenhouse gas (GHG) emission reduction isa key environmental and sustainability issue forthe 21st century. Energy-intensive manufac-tured materials (such as aluminum) could besignificantly affected, both in terms of price anduse, by GHG emission-reduction policies. How-ever, contrary to common belief, aluminumproduction could be positively affected by GHGemission-reduction policies. A combination ofemission mitigation in production and signifi-cant GHG emission reduction further down theproduct chain enhance the attractiveness of alu-minum for end-use applications, particularly inthe fastest growing sector of transportation(5.4% annual growth from 2001 to 2005).

Additional energy and environmental savingscan be achieved in the aluminum product chainthrough the introduction of new alloys and

Table 10.3 U.S. supply of aluminum from 1960to 2005 (in thousand metric dry tons)

U.S. U.S. U.S. U.S. primary secondary aluminum total

Year aluminum aluminum imports aluminum

1960 1828 401 178 24071961 1727 445 230 24021962 1921 533 341 27951963 2098 601 419 31181964 2316 648 409 33731965 2499 774 543 38161966 2693 833 591 41171967 2966 821 466 42531968 2953 935 684 45721969 3441 1067 484 49921970 3607 937 406 49501971 3561 1004 582 51471972 3740 1022 684 54461973 4109 1127 523 57591974 4448 1163 511 61221975 3519 1121 453 50931976 3857 1334 608 57991977 4117 1456 683 62561978 4358 1518 907 67831979 4557 1612 711 68801980 4653 1577 603 68331981 4489 1790 782 70611982 3274 1666 823 57631983 3353 1773 1023 61491984 4099 1760 1376 72351985 3500 1762 1332 65941986 3039 1773 1843 66551987 3347 1986 1702 70351988 3945 2122 1467 75341989 4030 2054 1353 74371990 4048 2393 1421 78621991 4121 2286 1398 78051992 4042 2756 1573 83711993 3695 2944 2327 89661994 3299 3086 3136 95211995 3375 3188 2701 92641996 3577 3307 2572 94561997 3603 3547 2804 99541998 3713 3442 3264 10,4191999 3779 3695 3680 11,1542000 3668 3450 3580 10,6982001 2637 2970 3487 90942002 2705 2927 3947 95792003 2704 2820 4068 95922004 2517 3025 4565 10,1072005 2480 2990 5221 10,691

Average growth rate 1995 through 2005, %–4.2 –1.7 6.9 0.4

Percent of total supply (2005), %28 29 42 100

Source: Aluminum Statistical Review for 2003, The Aluminum Association,2004, p 7

Table 10.4 Energy saved with recycling (tacit energy values)60.5 kWh/kg to produce primary metal ingot2.8 kWh/kg to produce secondary metal ingot5.0% secondary-to-primary energy2,990,000 kg of secondary metal produced in 20031.72 � 1011 kWh/yr energy saved 20030.59 quads saved per year 200319,700 MW saved 2003


improved (lightweight) product design. Theseoptions are not considered in this chapter, buttheir potential is at least of the same order ofmagnitude as changes to production practicesand processes.

Identifying Energy-Reduction Opportuni-ties. Energy performance benchmarks, currentpractice, and theoretical minimums provide thebasis for evaluating energy-reduction opportu-nities. These benchmarks and gross energy con-sumed during aluminum production in theUnited States are summarized in Table 10.5.

Industrial processes that consume energy atsignificantly higher rates than their theoreticalrequirements are, on the surface, obvious targetsfor potential improvement. However, energyperformance is only one factor in identifyingthe best opportunities for improving energy

efficiency. Other factors, particularly marketdynamics, process economics, and forecasting offuture demand are very significant in identify-ing real opportunities. This report examines theenergy performance of the operations involvedin manufacturing aluminum products.

The amounts of energy used onsite in themajor processing operations of the U.S. alu-minum industry are shown in Fig. 10.2. Thebottom band on each bar shows the theoreticalenergy requirement, while the top band of eachbar shows the energy used above the theoreticalminimum. The size of the top band is an indica-tion of how large the opportunity is for energyreduction in that process step.

Smelting requires 43% of the total primaryenergy consumed in U.S. manufacturing of alu-minum (66% when electric generation and

Table 10.5 U.S. energy requirements and potential savingsTheoretical U.S. process Potential process Total U.S. Potential gross

minimum energy energy U.S. energy gross energytf U.S. energytf

U.S. annual requirement, required, savings, required(a), savings(a),production 2005, kWh (109)/yr kWh (109)/yr kWh (109)/yr kWh (109)/yr kWh (109)/yr

metric tons (quad) (quad) (quad) (quad) (quad)

Bauxite mining 4,419,000 0.62 (0.002) 16.64 (0.057) 16.02 (0.055) 18.03 (0.062) 17.41 (0.059)alumina refining

Anode production 1,128,000 11.12 (0.038) 14.44 (0.049) 3.33 (0.011) 15.20 (0.052) 4.08 (0.014)Al smelting 2,530,000 15.15 (0.052) 39.42 (0.135) 24.27 (0.083) 113.41 (0.387) 98.26 (0.335)Primary casting 2,480,400 0.82 (0.003) 2.50 (0.009) 1.68 (0.006) 3.55 (0.012) 2.73 (0.009)Secondary casting 2,990,000 0.99 (0.003) 7.47 (0.026) 6.48 (0.022) 8.37 (0.029) 7.37 (0.025)Rolling 4,842,600 1.55 (0.005) 3.04 (0.010) 1.49 (0.005) 5.91 (0.020) 4.36 (0.015)Extrusion 1,826,000 0.80 (0.003) 2.37 (0.008) 1.57 (0.005) 2.75 (0.009) 1.96 (0.007)Shape casting 2,289,000 0.76 (0.003) 5.85 (0.020) 5.09 (0.017) 6.04 (0.021) 5.28 (0.018)Total 31.81 (0.109) 91.74 (0.313) 59.93 (0.204) 173.3 (0.591) 141.4 (0.483)

(a) tf, tacit feedstock

Fig. 10.2 Process energy used in U.S manufacturing of aluminum products


transmission losses are considered). This processis the largest consumer of energy and the mosttechnically complex operation. Smelting requiresmore than twice its theoretical minimum energyand has the potential for the greatest energyreduction of all operations. Electricity is requiredfor smelting and accounts for over 98% of theenergy used in the process. Current R&D effortsto advance existing technology and to developalternatives to the existing smelting process havethe potential to lower smelting energy consump-tion by more than 30%.

Process heating accounts for 27% of the totalprimary energy consumed in U.S. manufactur-ing of aluminum. Process heating is required forholding, melting, purifying, alloying, and heattreating. It is used in nearly all aluminum pro-duction operations. Heating is the second largestenergy-consuming operation.

Recycled aluminum now accounts for overhalf of all U.S-produced aluminum. It requiresless than 5% of the energy to produce aluminumfrom mined bauxite and provides significantenvironmental benefits. The R&D efforts thatimprove the ability to recycle aluminum offersome of the greatest opportunities for energyreduction in the industry, because recycling dis-places aluminum produced by smelting.

The magnitude of the top bands of the energybars in Fig. 10.2 shows that large opportunitiesexist for lowering energy consumption in theindustry. The Aluminum Industry Vision: Sus-tainable Solutions for a Dynamic World (Ref10.5), published by the Aluminum Associationin 2001, recognizes these opportunities and setsindustry goals for achieving further energyreduction. In Hall-Heroult smelting technology,the most energy-intensive process, the industryhas set a target for reducing electrical energyusage to 11 kWh/kg of aluminum producedby the year 2020, a 27% reduction from 2000practices.

Evaluation of the many opportunities that existfor reducing energy consumption in the industrycan only be made by comparing processes usingconsistent system boundaries and measures. Thischapter provides data and information necessaryfor the reader to understand opportunities forenergy savings in the aluminum industry.

Methodology, Metrics, and Benchmarks

This chapter focuses on the most energy-intensive manufacturing operations for alu-minum: electrolysis (smelting) and process heat-ing operations. These two operations accountfor over 70% of the primary energy used by theindustry (Table 10.6). There is a large differencebetween the theoretical minimum energy require-ments and current practice energy values inelectrolysis and melting. The magnitude of theenergy consumed and the difference betweencurrent practice and theoretical energy levelsmeans improvement in electrolysis and processheating will have the largest impact on the per-formance of the industry. This chapter docu-ments existing operations and explores potentialnew technology opportunities.

The purpose of this chapter is to:

• Provide an understanding of the processesinvolved, the energy consumed, and the envi-ronmental impact of manufacturing alumi-num and aluminum products

• Provide a common set of terms, benchmarks,and values for comparing processes andissues related to the aluminum industry

• Identify process areas in which significantenergy reductions and environmental impactimprovements could be made

• Strengthen public and workforce awareness,education, and training (as an identifiedindustry goal in Ref 10.5)

Table 10.6 Smelting and heating fractions of total U.S. aluminum industry energy consumedSmelting fraction of total U.S. aluminum industry energy consumed Percent

Electrolysis Total industry of industry

Onsite kWh/yr 3.94 � 1010 9.17 � 1010 43Tacit kWh/yr 1.13 � 1011 1.73 � 1011 66

Aluminum metal process heating/melting fraction of total U.S. aluminum industry energy consumedRolling Total process Percent

Casting Hot Cold Extrusion Shape casting heating Total industry of industry

Btu/kg Al 9244 1212 905 4096 8704 2.42 � 104

kWh/kg Al 2.71 0.36 0.27 1.20 2.55 7.08Onsite kWh/yr 1.48 � 1010 8.60 � 108 6.42 � 108 2.19 � 109 5.84 � 109 2.44 � 1010 9.17 � 1010 27Tacit 1.73 � 1011 14

Note: The tacit increase for process heating is negligible.


There are a variety of metrics, measurements,benchmarks, boundaries, systems, and units thatare used differently by various analytical groups.These variations can cause confusion whencomparing values stated in one report to thosein another. Two commonly confused values arethe relationship between on-site and tacit energyvalues, and between U.S. energy requirementsand worldwide energy requirements. On-siteenergy values are based on physical measure-ments. Tacit energy values involve assump-tions, which can create large differences in thereported values. The on-site and tacit values usedhere are explained in the section “Tacit, Process,Feedstock, and Secondary Energies” in thischapter. The United States does not mine ore foraluminum production but does refine ore domes-tically. Energy consumption within the UnitedStates is the focus here. The total energy asso-ciated with production of metal from ore is animportant value and is reported as the “world-wide” energy requirement in this chapter.

Theoretical, Practical Minimum, andCurrent Practice Benchmarks. When examin-ing industrial processes, two metric values forenergy requirement are obtainable with littledebate: the process theoretical minimum valueand the current practice value. The theoreticalminimum energy requirement for chemicallytransforming a material is based on the netchemical reaction used to manufacture theproduct. In the case of aluminum made fromalumina (2Al2O3 4Al � 3O2), the theoreti-cal minimum energy is 9.03 kWh/kg ofaluminum produced (see Appendix F). Thisminimum value is simplistic and represents thethermodynamically ideal energy consumption.It requires any reaction to proceed infinitelyslowly. The theoretical minimum energy totransform a material from one shape to anothershape is based on the mechanical properties ofthe material. It is also an idealized value. Nei-ther chemical nor mechanical theoretical mini-mums can be realized in practice; however,these provide the benchmarks that no processcan surpass. (Analogy: The theoretical mini-mum score for a full round of golf is 18.)

The current practice value is the average of theactual measurements of existing processes andpractices. (Analogy: The current practice valuefor golf is the average score of every player,which is well above par.) The boundaries drawnaround the process or practice, the number ofsamples, sampling techniques, and so on deter-mine the precision and accuracy of this value.

The difference between the theoretical minimumand current practice metric is a valuable measureof the opportunities for energy efficiencyimprovement in that process or practice.

Practical minimum energy is a term in com-mon use. However, its definition varies widely.In some instances, it is used to describe theprocess energy value that represents the combi-nation of integrated unit operations using bestavailable technology and best energy manage-ment practices. In other instances, practical mini-mum energy is defined as the optimal designvalue projected with the adoption of new,advanced technology. Practical minimum energyvalues are, in reality, a moving target because itis not possible to predict the new technologies,practices, and materials that will impact anindustrial process. What is known about thepractical minimum energy value is that it liessomewhere between the current best availablevalue and the theoretical minimum value.(Analogy: The practical minimum score for golfis some value below par and over 18.)

The Aluminum Industry Vision (Ref 10.5) hasselected a goal of 11 kWh/kg of aluminum as itssmelting current practice value for the year2020. This represents a 27% reduction over the1995 value of 15.4 kWh/kg of aluminum. Theindustry envisions this as an obtainable andpractical minimum smelting energy goal for2020.

Tacit, Process, Feedstock, and ‘Secondary’Energies. Current practice process measure-ments are actual measurements taken within afacility on existing operations. These on-siteprocess measurements are valuable because theyare the benchmarks that industry uses to compareperformance between facilities and companies.More importantly, these on-site process meas-urements are used to assess the value of newprocesses and practices. These are the criticalvalues used in the decision-making process toadopt new technologies and practices. On-siteprocess measurements, however, do not accountfor the complete energy and environmentalimpact of manufacturing a product. A fullaccounting of the impact of manufacturing mustalso include the energy used to produce theelectricity, as well as the fuels and the rawmaterials used within a manufacturing facility.These secondary energy requirements for elec-tric power generation and transmission, for theenergy needed to produce fuels, and for theenergy values of feedstock materials are veryimportant from a regional, national, and global

�


energy perspective, but they are seldom analyzedor accounted for within an individual plant site.

The process energy or secondary energy asso-ciated with the fuels used in aluminum process-ing is presented in Table B.1 of Appendix B.The process energy adds approximately 3% tothe energy values of the fuels used (Appendix B,Table B.2). Feedstock energy represents the energy inherent in fuels that are taken into amanufacturing process but used as materialsrather than fuels. Aluminum production usescoke as a raw material in the production of car-bon anodes. The feedstock energy of coke issignificant and is equivalent to a 30% increasein the on-site energy consumption of the Hall-Heroult process (Appendix E, Table E.1). Theenergy contribution of feedstocks is expressed interms of calorific or fuel value plus the second-ary energy used to produce the feedstock. (Note:Fuel and feedstock tacit energy values used hereare the calorific fuel value plus the fuel process-ing energy, [Appendix B, Table B.1]).

Tacit energy is a term frequently used todescribe the combined total of on-site energyand the secondary energy requirements. Tacitelectrical energy and environmental impactmeasurements account for the fact that substan-tial electrical generation inefficiencies and trans-mission losses occur outside the facility. It cantake as much as four units of hydrocarbon orcoal calorific energy to produce one unit of elec-tric energy. Saving 1 kWh of on-site electricity isequivalent to saving nearly 2.9 kWh of theenergy contained in the petroleum or coal-basedfuels used to generate electrical power.

Tacit electric conversion factors are variablebecause they are dependent on the sources ofthe energy used to produce electricity. Eachmanufacturing facility has a different tacit con-version factor depending on its location. TypicalU.S. grid electricity requires approximately9890 Btu of energy to deliver 1 kWh of on-siteelectricity (3412 Btu) for use. Electricity pro-duction from coal requires 10,303 Btu to deliver1 kWh of on-site electricity (3412 Btu). Waterhas no fuel value, and hydroelectric facilitiescan be assumed to have a tacit energy require-ment of 3412 Btu to deliver 1 kWh of on-siteelectricity (3412 Btu) and near zero GHG emis-sions (Appendix C). The on-site and tacit electricenergy requirements for a facility operating onhydroelectric power are equal.

Comparing energy values for the varioussteps used in the production of aluminum prod-ucts is simpler when a common unit is used for

all processing steps. Because electricity is thesingle largest source of energy consumed in themanufacture of aluminum, the common units ofa kilowatt-hour (kWh) are used in this chapter.Process energy values for production steps thatconsume fuels are converted to kWh using theconversion factor of 3412 Btu/kWh.

The large variations in tacit electric energyconversion values, 10,303 Btu per on-site kWhfor coal compared to 3412 Btu per on-site kWhfor hydroelectric, have a dramatic influence onthe reported tacit energy profile of an industry.Aluminum smelting energy is 98% electricenergy. A modern smelter operating from ahydroelectric utility requires on-site energy of14.4 kWh/kg of aluminum produced and tacitenergy of 14.4 kWh/kg of aluminum, whereasan identical smelter operating from a coal-firedutility requires on-site energy of 14.4 kWh/kgof aluminum and tacit energy of 43.5 kWh/kg ofaluminum (Appendix B, Table B.3). The U.S.primary aluminum industry has approximately39% of its capacity connected to hydroelectricfacilities.

All values reported in this document use theU.S. average grid connection values (i.e., 9890Btu/kWh). The use of U.S. average grid valuesresults in a higher energy consumption valuethan the actual tacit value required for primaryaluminum production. The advantage of report-ing primary production energy consumptionbased on the U.S. average grid is that it allowseasier and same-basis comparisons to other U.S.manufacturing industries. For clarity, on-siteoperating energy values are distinguished fromthe secondary energy values (which includetacit/feedstock contributions with the use of asuperscript “tf”).

Life-cycle assessment (LCA) is recognized asthe most complete analysis model of the productimpact on energy, environmental, economic, andsocial values. The LCA of an industrial productextends from “cradle-to-grave,” that is, frommaterial acquisition and production; throughmanufacturing, product use, and maintenance;and finally, through the end of the product life indisposal or recycling. The LCA recognizes theimportance of considering energy, economic, andenvironmental factors not only during the produc-tion of a product but also over the complete lifecycle of the product, including use and disposal.The LCA is particularly useful in ensuring thatthe benefits derived in one area do not shift theimpact burden to other places within a productlife cycle.


The LCA “use and maintenance” factor foraluminum varies by end-product and, in manyapplications, is more significant in terms ofenergy and environmental impact than produc-tion. Aluminum, in some cases, provides LCA“use and maintenance” energy savings that aresignificantly greater than the energy used in itsproduction. For example, production of an equal-strength but lighter aluminum product in thetransportation sector saves significant amountsof transportation fuel and provides substantialreductions in GHGs during the “use” phase ofthe product when compared to traditional mate-rials. In 2001, an estimated 2.2 billion gallons(18.3 billion liters) of gasoline use and 20 mil-lion metric tons of CO2 emissions were reduceddue to the use of lightweight aluminum castingsin automobiles (Ref 10.6).

Complete LCA for aluminum products mustaccount for the significant portion of aluminumthat, in the acquisition phase, comes from “urbanmining” (recycling). The ability of aluminum tobe easily recycled is reflected in the fact thatover half of the U.S.-produced aluminum now

originates from recycled material. Recycling isthe best option for disposal of nearly everyproduct made from aluminum. This makes alu-minum a “cradle-to-cradle” LCA product.

Energy Value Chain Analysis. The energyvalues studied and presented are based on anenergy “value chain” analysis. The value chainanalysis or “cradle-to-shipping dock” analysisprovided is an integral part of an LCA. It pro-vides valuable information and data values fororganizations performing LCA on aluminumproducts. Value chain analyses are similar toLCA; however, they cover only a portion of atotal LCA. Figure 10.3 shows the global bound-aries of an LCA study and the boundaries for thisstudy’s value chain.

Value chain analysis allows for the capture ofthe direct energy and feedstock inputs of eachprocessing step (link) and builds a cumulativevalue of each product along the chain. This chap-ter looks at a portion of the LCA: the energyvalue chain from cradle-to-shipping dock. Itdoes not account for the LCA use and mainte-nance phase energy or for tertiary energy inputs

Fig. 10.3 Boundaries for life cycle and value chain


(i.e., the energy used to make the equipment orbuildings that house the process steps). Thecradle-to-shipping dock approach is valuablefor providing decision-making analyses withinthe manufacturing sphere.

Transportation Energy. The transportationenergy associated with acquiring raw materialsand distribution of intermediate products isimportant for a full LCA. Transportation energycan account for a significant portion of the totalenergy associated with manufacturing a finalproduct. The energy required to transport minedbauxite to refining operations, alumina to smelt-ing operations, ingots to metal processors, andscrap from collection to melting is not accountedfor in the process energy requirements that aredeveloped here. This chapter focuses on theenergy associated with the processing of rawmaterials and the processes employed in alu-minum production. The transportation energyassociated with these raw materials andprocesses is small in relation to the total energyconsumed in production.

Transportation energy calculations for rawmaterials that are mined globally are highlyvariable. They are a function of the locationand multiple modes of transportation, forexample, conveyors, trucks, trains, and oceanfreight. Transportation energy requirements areevaluated in Chapter 3, “Life-Cycle InventoryAnalysis of the North American AluminumIndustry” in this book. Transportation of rawmaterials accounted for 2% of the total energyassociated with primary aluminum productionin the United States.

Evaluation of the transportation energyrequirements associated with secondary alu-minum production is complicated. Consumerscrap can require considerable transportationenergy resulting from individual consumer drop-off, curbside collection, transfer station collec-tion, and the actual transportation to a secondaryprocessor. Transportation energy, associated withindustrial manufacturing scrap and scrap origi-nating at large automotive and white-good scrapprocessing centers, is more easily estimatedbecause its boundaries are easier to define. Chap-ter 3 estimates transportation energy from thesesources to account for 6 to 8% of the total energyassociated with the production of secondary alu-minum products.

Emissions. Energy use and GHG emissionsare closely related. This chapter provides overallcarbon emission data associated with fuels usedfor aluminum operations. Other fuel-related

emissions (e.g., nitrous oxides, sulfur dioxide,volatile organic compounds) are not consideredbecause their quantities are typically small ascompared to the carbon-base emissions. Emis-sions that are aluminum process-related (e.g.,perfluorocarbons, from cryolite) are reported.The Energy and Environmental Profile of theU.S. Aluminum Industry (Ref 10.8) providesdetailed emission data for aluminum operations.

Emission calculations for this chapter areshown in Appendix D. Greenhouse gases con-tribute to climate change by increasing the abilityof the atmosphere to trap heat. Gases differ intheir ability to trap heat. To express the green-house effect of different gases in a comparableway, atmospheric scientists use a weightingfactor, global-warming potential. The heat-trapping ability of 1 metric ton of CO2 is thestandard, and emissions are expressed in terms ofa million metric tons of CO2 equivalent, or 106

tonnes CO2e. This report uses carbon dioxideequivalents (CO2e). Emissions are also com-monly expressed in terms of a million metrictons of carbon equivalent (106 tonne Ce). Car-bon comprises of the mass of CO2; to convertfrom CO2 equivalent to carbon equivalent, mul-tiply the CO2 equivalent by 0.273.

Aluminum Production

Aluminum metal is classified as primary alu-minum if it is produced from ore and as second-ary aluminum if it is produced predominantlyfrom recycled scrap material. Primary alu-minum metal production consists of bauxitemining, refining bauxite to produce alumina,and finally, smelting alumina to produce alu-minum. Secondary aluminum is produced bysorting, melting, and treating scrap aluminum.Primary and secondary aluminum metal are fur-ther processed using traditional metalworkingtechnologies—rolling, extrusion, forging, shap-ing, and casting into thousands of products.

Aluminum is the most abundant metallic ele-ment in the Earth’s crust. However, it is neverfound in natural deposits as a free metal, such ascopper and gold. Aluminum is typically foundas one of several aluminum oxides or silicatesmixed with other minerals and must be processedto be recovered in its pure form. All commercialprimary aluminum is produced from one rawmaterial, bauxite, and by one process, elec-trolytic reduction. For economic and strategicreasons, the aluminum industry continues to

1244


perform research and development on alternativeraw materials (e.g., kaolin clay) and processes(e.g., chemical reduction). Although these alter-natives hold promise for reducing costs, energyconsumption, and environmental impacts, noneare near commercialization.

The markets for the aluminum industry’s rawmaterials and products are global. Global pri-mary aluminum production has been growing ata rate of 4.5% annually from 1995 to 2005 (Ref10.2). The U.S. aluminum total supply grew atan annual rate of 0.5% over the period of 1995to 2005 (Table 10.3). Aluminum is still in thegrowth phase of the product cycle. Demand foraluminum is increasing, mainly due to aluminumsubstitution for other materials in the transporta-tion sector and other lightweight applications. Itslight weight, corrosion resistance, and processingpossibilities, coupled with its ease and value forrecycling, strengthen its position as the materialof choice in many applications. Measured ineither mass produced or economic value, theuse of aluminum exceeds that of any other metalexcept iron. It is important in virtually all seg-ments of worldwide manufacturing.

The global estimate for economically recov-erable bauxite reserves is 22,000,000,000 met-ric tons. This quantity can address the demandsfor the next century. Two countries have nearlyhalf of the world’s identified bauxite resources(Guinea has 25%, and Australia has 20%).Bauxite is no longer mined in the United Statesas a commercial feedstock for aluminum pro-duction. Domestic ore, which accounts for lessthan 1% of the U.S. requirement for bauxite, isused in the production of nonmetallurgicalproducts, such as abrasives, chemicals, flameretardants, and refractories.

Alumina is produced by refining bauxite in awet caustic chemical leaching process (Bayerprocess). Imported bauxite is refined in theUnited States, the largest importer of bauxite andthe second largest bauxite refiner after Australia.Alumina production is continuing to rise in Aus-tralia, Brazil, Jamaica, Surinam, Venezuela, andIndia, all countries with large indigenous bauxite

reserves. The trend in alumina production istoward placing refining capacity near the mineralresources, thereby reducing transportation energyand costs and adding more value to exports.

Primary aluminum (aluminum from ore) isglobally produced by the Hall-Hèroult process,a method that involves electrolysis or smelting ofalumina. Companies choose their smelting loca-tions where production conditions are favorable,such as the availability of skilled labor, proximityto a consumer market, and provision for ahighly-developed infrastructure, and, especially,for low-cost and reliable energy. Hydroelectricpower accounts for 50% of the energy usedworldwide for electrolysis of aluminum (Table10.7). The bulk of energy use in aluminum pro-duction is related to the electricity required forprimary electrolysis. Because energy costs areapproximately one-third of the total cost ofsmelting primary aluminum, smelter productionhas been moving from sites close to consumermarkets to sites with low electricity costs. Mostof the primary aluminum industry restructuringbegan in the late 1970s and continues to thisday. China and Russia have emerged as majormetal producers; other countries entering theworld market include Canada, Australia, Brazil,Norway, and countries in the Persian Gulf area,all areas with low energy costs.

Secondary aluminum is produced from scrapor recycled aluminum. The world’s averageshare of secondary aluminum production isroughly one-quarter of total aluminum produc-tion. The United States produces over half of itsaluminum from recycled aluminum scrap(Table 10.3). Aluminum recycling is concen-trated in the countries where the scrap is gener-ated, with the exception of Asia, which importssignificant amounts of aluminum scrap (drivenby the demand for cast aluminum in the Asiancar industry).

Primary and secondary aluminum are used tomanufacture numerous products, ranging fromaircraft components to household and packagingfoils. Each product requires processing, suchas heating, melting, alloying, and mechanical

Table 10.7 Energy sources of electrical power in 2002 (electrical power used in gigawatt hours)Africa, North America Latin America, Asia, Europe, Oceania, Total, Grand total,

Electric energy source GWh GWh % GWh GWh GWh GWh GWh %

Hydro 6444 50,312 64 32,207 3030 29,969 7380 129,342 50Coal 13,443 27,248 35 0 10,080 15,773 24,120 90,664 35Oil 0 93 0 0 109 1279 0 1481 1Natural gas 38 351 0 1343 16,336 6199 0 24,267 9Nuclear 189 678 1 0 0 12,672 0 13,539 5Total 20,114 78,682 100 33,550 29,555 65,892 31,500 259,293 100


working. Figure 10.4 shows the tacit energyconsumption of the major processes in the alu-minum production chain. Production of primaryaluminum accounts for 86.7tf% of the energyconsumed by the U.S. industry; production ofsecondary aluminum for 4.8tf%; rolling for3.4tf%; extrusion for 1.6tf%; and shape casting forthe remaining 3.5tf% (Appendix E, Table E.4).

Two operations, electrolysis and the heat-ing/melting of aluminum, account for over 70(80tf)% of energy consumed in aluminum pro-cessing (Appendix E, Table E.8). Heating andmelting technologies are used for holding,alloying, and treating metal as well as for recy-cling. Programs that improve thermal efficiencyof heating and melting while minimizing theformation of aluminum oxide and/or dross pro-vide a much larger impact on decreasing indus-try energy usage than their energy consumptionindicates.

The U.S. aluminum supply of 10,692,000metric tons in 2005 originated from three basicsources: primary aluminum (domestically pro-duced), secondary aluminum (recycled domesticmaterial), and aluminum imports. This consistedof 2,480,000 metric tons of primary aluminum,2,990,000 metric tons of secondary aluminum,and 5,221,000 metric tons of importedaluminum (Ref 10.9). From 1995 to 2005, theannual U.S. growths of these supplies were –4.2,–1.7, and 6.9%, respectively. Since 1995, thetotal U.S. supply has risen at an annual rate of

approximately 0.5%. Figure 10.5 shows the dis-tribution of these supplies over the past 45 years.

The United States is the fourth leadingproducer of primary aluminum metal in theworld. However, its dominance in the globalindustry has declined. The U.S. share of worldproduction in 1960 accounted for slightly morethan 40% of the primary aluminum produced.By 2005, the U.S. share of world productionhad decreased to 7.8%. The U.S. primary pro-duction peaked in 1980 and over the past 25 years has been gradually declining. Signifi-cant year-to-year variations occur as a result ofU.S. electrical costs and global market changes.

Secondary (recycled) aluminum is of growingimportance to the U.S. supply. In 1960, only401,000 metric tons of aluminum were recov-ered. In 2005, 2,990,000 metric tons of alu-minum were recovered. For the years 1995through 2005, the secondary production of alu-minum has dropped at an annual rate of –1.7%(Table 10.3). Recently, the U.S. secondary alu-minum growth rate has been slowing due to acombination of factors. Scrap collection pro-grams are beginning to reach their maturationstage, and the market growth of scrap sourceshas slowed. Additionally, the U.S. has nowbecome a net exporter of scrap and dross. In1993, the U.S. had net imports of scrap anddross of 98,540 metric tons, while in 2003, theU.S. had net scrap and dross exports of 152,000metric tons. The main cause for this change is

Fig. 10.4 Energytf consumption of U.S. aluminum operations


again China’s radical increase in demand forscrap. In 2003, 244,000 metric tons of the567,700 metric tons of scrap and dross exportedby the U.S. (43%) went to China (Ref 10.9).These trends of decreasing U.S. secondary alu-minum production are, however, expected tochange. Use of aluminum in the automotiveindustry grew at over 5.4% annually between1995 and 2005. This large and growing supplyis now beginning to enter the scrap markets andwill spur new growth in secondary aluminum.

Imported aluminum is the fastest growingsource of U.S. supply, with an annual rate of6.9% over the 1995 to 2005 time frame (Table10.3). New primary aluminum facilities are beinglocated outside the United States, near newsources of low-cost electricity.

Primary Aluminum Raw Materials

The total energy associated with producingthe raw materials required for aluminum pro-duction from bauxite ore was approximately 8.20(14.22tf) kWh/kg of aluminum. This accounts for28% of the total energy required to produce pri-mary aluminum metal and consists of:

• 0.32 (0.34tf) kWh/kg aluminum for bauxitemining

• 7.27 (7.87tf) kWh/kg aluminum for bauxiterefining to alumina

• 0.61 (6.01tf) kWh/kg aluminum for carbonanode production

A complete account of the energy require-ments and environmental impacts to produceany product must include the energy require-ments and environmental impact associated

with the production of the raw materials used.The raw material energy requirements andenvironmental impacts associated with primaryaluminum production can be divided into themajor operations required to produce it. Theseare bauxite mining, bauxite refining, and carbonanode manufacturing. Approximately 5900 kgof earth are mined to produce 5100 kg of baux-ite, which is refined into 1930 kg of alumina.The 1930 kg of alumina are electrolyticallyprocessed with 446 kg of carbon to produce 1 metric ton (1000 kg) of aluminum (AppendixE, Table E.1).

Cryolite and other fluoride salts are used asthe electrolytic bath for aluminum production.These materials are theoretically not consumedin the process or combined as part of the finalproduct. However, approximately 19 kg of bathmaterial is lost for every metric ton of aluminumproduced (Appendix E, Table E.1). These lossesare a result of process upsets and bath drag-outwhen molten aluminum is removed from thesmelting operation. These salts represent only asmall portion of the energy requirement for pro-ducing the raw materials required for aluminumproduction and are not addressed here.

Bauxite

Aluminum, never found as a free metal, occursnaturally in the form of hydrated aluminumoxides or silicates. Because the silicates aremixed with other metals such as sodium, potas-sium, iron, calcium, and magnesium, and it ischemically difficult and expensive to extractaluminum from them, the silicates are not apractical source of aluminum. The oxides are,therefore, used for producing aluminum. The

Fig. 10.5 U.S. aluminum supply from 1960 to 2005


aluminum oxides commonly found as naturallyoccurring minerals include:

• Corundum (alumina, Al2O3) • Böehmite (α-Al2O3 � H2O, a monohydrate

containing 85 wt% alumina) • Diaspore (β-Al2O3. � H2O, chemically the

same as böehmite but with a different crystalstructure)

• Gibbsite (γ-Al2O3 � 3H2O, a trihydrate con-taining 65.4 wt% alumina)

• Alumina, used for the production of alu-minum, is obtained from bauxite deposits.

Bauxite is not a true mineral but a rock thatcontains mostly böehmite and gibbsite alongwith diaspore, corundum, and numerous impu-rities (mostly compounds of iron, silicon, andtitanium). Bauxite commonly appears as a col-lection of small, reddish-brown nodules in alight-brown, earthy matrix. The alumina avail-able in commercial bauxite ranges from 30 to60 wt%. Bauxite is typically classified accordingto its intended commercial application: abrasive,cement, chemical, metallurgical, refractory, andother end uses. The bulk of world bauxite pro-duction (approximately 85%) is metallurgicaland used as feedstock for the manufacture ofaluminum.

The United States mines less than 1% of thebauxite it uses annually, virtually all of which isused in the production of nonmetallurgical prod-ucts, such as abrasives, chemicals, and refracto-ries (Ref 10.10). Nearly all bauxite consumed inthe United States is imported. Figure 10.6 tracksthe domestic and imported components of U.S.bauxite supply from 1960 to 2005 (Table 10.8).

In 2005, the United States imported a total of10,400,000 metric tons of bauxite. Approxi-mately 95% of the imported bauxite is refined toproduce alumina, and approximately 91% of the

refined alumina is used to produce primaryaluminum.

Bauxite Energy Requirements (Onsite andTheoretical). Approximately 0.32 (0.34tf) kWhof process energy were required to produce the5.1 kg of bauxite needed to produce 1.0 kg ofaluminum. Approximately 16.7 kg of carbondioxide equivalent (CO2e) were released foreach metric ton of bauxite mined.

The energy demand associated with theextraction of bauxite is typical of most miningoperations. Bauxite ore is generally strip-minedby removing the overburden (the soil on top ofthe deposit) and excavating it with mechanicalequipment. The overburden is saved for recla-mation operations, which are extensively prac-ticed to ecologically restore mined areas. Thesoft, earthy nature of many bauxite depositsgenerally does not require drilling or blastingoperations. After mining, the bauxite is crushed,sometimes washed and dried, and transported torefining plants via ship, barge, rail, truck, orconveyor belt.

Approximately 5.1 kg of bauxite are requiredto produce a kilogram of aluminum. The energyrequirement per kilogram of mined bauxite is0.06 kWh for typical extraction (Ref 10.11).Because electricity accounts for less than 1% ofthe energy used in bauxite production, the tacitaddition is negligible (Appendix E, Table E.1).

Calculation of a theoretical minimum energyrequirement for mining bauxite is dependent onthe system boundaries applied and the processesused. The laws of thermodynamics state thatseparating the constituents of a mixture, such asbauxite from bauxite-rich soil, requires a certainminimum expenditure of energy. Bauxite is themajor constituent of bauxite-rich soils, and thereis no change in the chemical nature of bauxite inthe mining process. So, the theoretical minimum

1960 1965 1970 1975 1980 1985 1990 1995 2000 2005

20

15

10

5

0

Mill

ion

met

ric to

ns

Imports of bauxiteU.S. mined bauxite

Fig. 10.6 U.S. bauxite supply from 1960 to 2005


energy for preparing bauxite is negligible. Inaddition, because it is theoretically possible tofind bauxite on the surface, the theoretical mini-mum energy requirement to produce bauxite isvery close to zero. In the interest of simplicity,the theoretical minimum energy requirement formining bauxite is assumed to be zero.

Emissions from fuels used in the extraction ofbauxite are listed in Appendix D, Table D.2.These emissions are typically from surface-mining operations and result from a variety of

fuels used in the production of bauxite. Nearly0.0167 kg CO2e are emitted for each kilogramof bauxite mined.

Alumina (Al2O3)

Theoretically, from the stoichiometric equation(2Al2O3 � 3C 4Al � 3CO2), 1.89 kg ofalumina is required to produce 1 kg of alu-minum. In practice, a very small portion of thealumina supply is lost, and the industry requires

�

Table 10.8 U.S. supply of bauxite and alumina from 1960 to 2005 (in thousand metric dry tons)A B C D E F G H I

Estimated alumina Estimated net Imports U.S. mined Exports Net bauxite production from Imports of Exports of alumina for

Year of bauxite bauxite of bauxite supply bauxite(a) alumina alumina electrolysis(b)

1960 8879 2030 29 10,880 4651 80 ... 42661961 9354 1248 153 10,449 4467 171 ... 41911962 10,745 1391 263 11,873 5076 158 ... 47261963 9408 1549 206 10,751 4596 173 ... 43091964 10,518 1627 283 11,862 5071 191 ... 47551965 11,601 1681 149 13,133 5614 206 290 49691966 11,928 1825 63 13,690 5852 443 290 54201967 12,010 1681 2 13,689 5852 865 499 56331968 11,359 1692 7 13,044 5576 1190 780 54291969 12,355 1873 5 14,223 6080 1730 885 63171970 13,039 2115 3 15,151 6477 2840 998 71711971 12,837 2020 35 14,822 6336 2190 980 69131972 12,803 1841 29 14,615 6248 2590 797 74161973 13,618 1909 12 15,515 6633 3090 694 83651974 15,216 1980 16 17,180 7344 3290 927 89731975 11,714 1801 20 13,495 5769 3180 933 74391976 12,749 1989 15 14,723 6294 3290 1050 79051977 12,989 2013 26 14,976 6402 3760 857 86651978 13,847 1669 13 15,503 6628 3970 878 90571979 13,780 1821 15 15,586 6663 3840 849 89881980 14,087 1559 21 15,625 6680 4360 1140 92321981 12,802 1510 20 14,292 6110 3980 730 87491982 10,122 732 49 10,805 4619 3180 590 67471983 7601 679 74 8206 3508 4030 602 65851984 9435 856 82 10,209 4364 4290 648 75701985 7158 674 56 7776 3324 3830 316 65061986 6456 510 69 6897 2948 3600 487 57671987 9156 576 201 9531 4075 4070 1130 66071988 9944 588 63 10,469 4475 4630 1040 76181989 10,893 525 44 10,849 4638 4310 1330 71541990 12,142 490 53 12,089 5168 4070 1260 74611991 12,300 460 58 12,242 5233 4590 1350 79501992 11,400 400 68 11,332 4844 4700 1140 79201993 11,900 378 92 11,808 5048 3940 1240 72431994 11,200 350 137 11,063 4729 3120 1040 63361995 10,800 300 120 10,680 4566 4000 1040 70691996 10,700 270 154 10,546 4508 4320 918 74601997 11,300 230 97 11,203 4789 3830 1270 68701998 11,600 200 108 11,492 4913 4050 1280 71921999 10,400 180 168 10,232 4374 3810 1230 65172000 9030 150 147 8883 3797 3820 1090 61862001 8670 125 88 8582 3669 3100 1250 51892002 7710 125 52 7658 3274 3010 1270 47192003 8860 W(c) 89 8771 3750 2310 1090 46322004 10,500 W(c) 75 10,425 4457 1650 1230 44762005 10,400 W(c) 62 10,338 4419 1860 1200 4682

Note: 4,419,000 metric tons or 94% of the 4,682,000 metric tons of alumina needed to produce aluminum in the United States was refined in the United States in 2005. (a) Calculated based on United States Geological Survey (USGS) data. The average alumina content for bauxite is 45%, and 95% of bauxite is converted to alumina. Thecalculation is column F = 0.95 � 0.45 � column E. The numbers may be slightly overestimating because slightly less than 100% of the alumina content of bauxite is ac-tually recovered. (b) Net is calculated based on USGS data. 90% of alumina produced is metallurgical alumina used in aluminum production. (c) W, data withheld.Source: United States Geological Survey, Minerals Information, Statistical Compendium


approximately 1.93 kg of alumina for productionof each kilogram of aluminum. In 2005, theUnited States produced 4,419,000 metric tonsof alumina from bauxite and imported an addi-tional 1,860,000 metric tons of alumina to makealuminum.

Figure 10.7 shows the alumina supply sourcesfrom 1960 to 2005. The United States had fourBayer refineries in operation in 2005. Theserefiners processed approximately 10,400,000metric tons of bauxite into 4,419,000 metrictons of alumina. Approximately 9% of the alu-mina produced is used to manufacture abrasive,refractory, and other products. Approximately4,883,000 metric tons of U.S.-refined aluminawere transferred to the primary aluminum indus-try. This quantity of alumina was not sufficient tosupply the U.S. demand for alumina; therefore,an additional 1,200,000 (net) metric tons of alu-mina were imported.

All commercial alumina is refined from baux-ite using the Bayer refining process. Theprocess, developed by Karl Bayer in 1888, con-sists of four major steps: digestion, clarification,precipitation, and calcination Bauxite composi-tion varies, and refining plant designs areslightly different to account for the site-specificquality of the bauxite.

In digestion, crushed, ground, and sized baux-ite is dissolved under pressure with a hot (180 to250 °C, or 360 to 480 °F) sodium hydroxide andsodium carbonate solution in a series of steam-heated digesters. The concentrations, tempera-tures, and pressures employed vary depending onthe properties of the bauxite. Gibbsite is solublein caustic soda above 100 °C (212 °F), whileböehmite and diaspore are soluble in caustic sodaabove 200 °C (390 °F). Because the treatments ofböehmite and diaspore require higher tempera-tures and longer digestion times, they are moreexpensive than the treatment of gibbsite. The

aluminum oxides in the bauxite react to formsoluble sodium aluminate or “green liquor”:

Al2O3 � H2O � 2NaOH 2NaAlO2 � 2H2O and

Al2O3 � 3H2O � 2NaOH 2NaAlO2 � 4H2O

Silicas in the bauxite are detrimental to thedigestion efficiency. They react to form sodiumaluminum silicate, which precipitates. This pre-cipitate chemically binds the aluminum fromthe bauxite and the sodium from the sodiumhydroxide into a solid from which the aluminacannot be economically recovered. This de-creases the yield of alumina and increases thecosts associated with sodium hydroxide. Chem-ical additions and the adjustment of refiningpractices can effectively provide desilicationand decalcification of specific alumina streams.

In clarification, the green liquor produced bydigestion is clarified to remove sand, undis-solved iron oxides, titanium oxides, silica, andother impurities. The insoluble materials, calledbauxite residue or red mud, are thickened,washed, and dewatered to recover sodiumhydroxide. Bauxite residue is a large-quantitywaste product that is generally stored adjacentto refinery sites in landfills or lagoons. Afterweathering, the landfills can sustain vegetation.

In precipitation, the clarified liquid that results from clarification is cooled and “seeded”with crystals of gibbsite to aid precipitation ofalumina trihydrate (Al2O3 � 3H2O). This is thereverse of the digesting step (2NaAlO2 � 4H2O

(Al2O3 � 3H2O � 2NaOH). However, bycarefully controlling the seeding, temperature,and cooling rate, specific physical properties canbe given to the precipitating alumina trihydrate.

In calcination, alumina trihydrate is typicallycalcined in a fluid bed or rotary kiln at approxi-mately 980 to 1300 °C (1800 to 2370 °F) to

�

�

�

Fig. 10.7 U.S. alumina supply from 1960 to 2005


remove the water of crystallization and producethe dry white powder, alumina (Al2O3 � 3H2O

Al2O3 � 3H2O). Calcining rates and tem-peratures are carefully controlled and varydepending on the final physical properties spec-ified for the alumina.

Alumina used for electrolysis not only has achemical purity specification but also a physicalspecification on particle size, surface area, bulkdensity, and attrition behavior. These propertiesaffect the free-flowing properties of alumina(how it flows in feeders), the rate at which it dis-solves in cryolite, dust levels, the strength of thealumina crust, its insulating properties, and otherproperties important in the aluminum electroly-sis cell operation. The bulk density of aluminais 880 to 1100 kg/m3, while its specific gravityis 3.9 g/cm3.

Bauxite residue (red mud) is a by-product ofthe Bayer process and contains the insolubleimpurities of bauxite. The amount of residuegenerated per kilogram of alumina producedvaries greatly depending on the type of bauxiteused, from 0.3 kg for high-grade bauxite to 2.5 kg for low-grade bauxite. Its chemical andphysical properties depend primarily on thebauxite used and, to a lesser extent, the mannerin which it is processed.

Although a great deal of effort has beenexpended over several decades to find anddevelop uses for bauxite residue, a cost-effective,large-scale bulk application has yet to be found(Ref 10.12). Numerous attempts have beenmade to recover additional metals from theresidue, such as iron, titanium, and gallium.Other possible uses for the residue have includedproduction of ceramic bricks or tiles, use asroadbed material or as filler material for plas-tics, or production of cement. Accordingly, thecurrent industry efforts focus on minimizingthe amount of residue generated and improvingits storage conditions.

Probably the most promising and recentapplication for the residue has occurred inWestern Australia, where it is being evaluatedas a soil amendment or conditioner. The soilsin Western Australia are sandy and drainfreely, allowing fertilizers to leach into water-ways where they boost nutrient levels and canlead to problems such as algal blooms. Theapplication of bauxite residue to these sandysoils aids the retention of phosphates andmoisture and reduces the need to apply limefor soil pH adjustments. As part of the Bauxite

Roadmap project on bauxite residue, amodel farm has been planned to run for 5 to 7years with the application of red mud, and allits aspects and results will be studied in greatdetail.

Alumina Energy Requirements. Approxi-mately 7.27 (7.87tf) kWh of energy were requiredand 1.62 kg CO2e were released to refine the1.93 kg of alumina from bauxite needed to pro-duce 1 kg of aluminum in 2003.

The energy required to produce aluminafrom bauxite in 1985 was estimated to rangefrom 2 to 9 kWh/kg of alumina (Ref 10.3). Thisbroad range of energy intensity reflects bothbauxite quality (alumina content) and refinerydesign. It was estimated that in 1991, U.S.refiners averaged 3.66 kWh/kg of alumina pro-duced (Ref 10.13). The most recent availablerefining data list 3.76 kWh/kg of alumina for1995 (Ref 10.11). Calcination is the most energy-intensive operation of the Bayer process. Onaverage, 1.93 kg of alumina are consumed tomake 1 kg of aluminum (Ref 10.11). The alu-mina energy requirement for 1 kg of aluminumcan be estimated as 3.76 kWh/kg of aluminatimes 1.93 kg alumina, or 7.27 kWh/kg ofaluminum, a tacit value of 7.87tf kWh/kg ofaluminum.

The Alumina Technology Roadmap (Ref10.14) provides insight into the high-priorityresearch and development needs of the globalalumina industry. The Roadmap recognizes theneed and the opportunity for a 25% energyreduction by 2020 and improved, more sustain-able handling of bauxite residues. Better chemicalprocess knowledge, waste heat utilization, andcogeneration are opportunities for energy reduc-tion in the refining process.

Emissions from fuels used in the refiningprocess are listed in Appendix D, Table D.2.These emissions are predominantly related tonatural gas and coal consumption for digestionand calcination. Nearly 1.62 kg CO2e are emit-ted for each kilogram of refined alumina.

Alumina Theoretical Minimum EnergyRequirements. The theoretical minimum energyrequired to produce alumina is 0.13 kWh/kg ofaluminum produced.

The theoretical minimum energy requirementsto produce metallurgical-grade alumina frombauxite can be calculated from the reactionsrequired in the process. The minimum energyrequirements for the digestion, clarification,and precipitation steps are related to the two

�


chemical reactions that take place during theseprocessing steps:

Al2O3 � 3H2O � 2NaOH 2NaAlO2 � 4H2O

Al2O3 � 3H2O � 2NaOH

However, because there is no net chemical ortemperature change for the combined reactions,the theoretical minimum energy requirement forthese steps is near zero. The final reaction, calci-nation, requires energy to remove the hydratedwater from alumina (Al2O3

. 3H2O Al2O3 �3H2O). The theoretical minimum energyrequirement to calcine (dehydrate) the aluminais 0.13 kWh/kg of aluminum produced (Appen-dix F, Table F.10). This value assumes that theprecipitated alumina trihydrate is completelybone dry before entering the calcining process.At 8% moisture, this would add approximately0.01 kWh/kg of aluminum to the minimumrequirement.

Carbon Anode

The United States consumed 1,128,000 metrictons of carbon anode in 2005. Approximately0.45 kg of carbon anode were needed to pro-duce 1 kg of aluminum (Appendix E, TableE.1). All commercial production of aluminumuses carbon as the anode material for the elec-trical reduction of alumina to aluminum. Thecarbon anode net reduction reaction (2Al2O3 �3C 4Al � 3CO2) requires three carbonatoms for the reaction to free four aluminumatoms. The theoretical minimum anode con-sumption is 0.33 kg of carbon per kilogramaluminum [(3 � 12.01 carbon molecular weight)(4 � 26.98 aluminum molecular weight)]. Anodematerial quality is important because all theimpurities dissolve into the bath and ultimatelycontaminate the molten aluminum. The physi-cal quality of the anode also affects both theenergy efficiency and productivity of smeltingcells.

Anode consumption rates in practice, typicallyapproximately 0.45 kg carbon per kilogram ofaluminum, are 35% higher than the theoreticalrequirement. Excess carbon usage results fromthe need to protect the iron electrical connectionwithin the carbon anode, and from air burningand dusting. The surface of the carbon is hotenough at cell operating temperatures to oxidizeat a slow rate. This is minimized by coating theanode surface and covering the anode with alu-mina, which insulates it from air exposure.

Dusting, which is breaking off small particlesof carbon into the air or bath, can account forhalf of the excess carbon used in the smeltingprocess. Dusting is a direct function of theanode material uniformity. It is caused by selec-tive electrolytic oxidation and air burning of thebinder pitch, which releases aggregate carbonparticles into the bath. Anode carbon dust isunavailable for aluminum production.

Two different anode technologies are used bythe U.S. industry. Prebaked carbon anodesaccount for more than 86% of the U.S. capacity.Older, in situ baked Soderberg anodes accountfor the remainder of the capacity. New prebakedanode reduction cells have surpassed Soderberganodes in terms of current efficiency and emis-sions control. Only three operational smelters inthe United States are currently using Soderberganodes. No new Soderberg cells are being built,and those that exist are progressively being re-placed, converted, or shut down. This chapterfocuses mainly on prebaked anodes for smeltingtechnology.

Carbon prebaked anodes are made by mixingground used carbon anodes, calcined petroleumcoke, and coal tar or petroleum pitch. Pitch actsas a binder to hold the anode mass in a “green”formed shape. Compacting the anode by usingvacuum and vibrating the mixture when formingproduces a denser, more conductive and lower-dusting anode. Baking carbonizes the pitch andcreates a solid bond between the particles ofcalcined coke and used anode material. Castiron is poured into preformed sockets in thebaked anode to form an electrical connection.Prebaked anodes can weigh as much as 1250 kgand have a working face size of approximately0.70 by 1.25 m (2.3 by 4.1 ft) and a 0.5 m (1.6 ft)height. Prebaked anodes are removed beforethey are completely consumed. Used anodes arerecycled into the anode production system torecover the carbon and the iron rods used forelectrical connections. Used anodes can accountfor 15 to 30% of the mass used in green anodemakeup.

Calcined petroleum coke is a by-product of thecrude oil refining industry. Green or raw cokecontains 8 to 10% moisture and 5 to 15% volatileorganic materials. Raw coke must be calcined atapproximately 1200 to 1350 °C (2190 to 2460°F) in gas-fired kilns or rotary hearths to removethe moisture, drive off volatile matter, and to increase the density, strength, and conductivity ofthe product (Ref 10.15). Calcined coke has a bulkdensity of approximately 800 kg/m3. Worldwide,

�

�

��


approximately 25% of all raw coke is calcined,and approximately 70% of all calcined coke goesto aluminum production. Modern calcininghearth and kiln designs capture and use thevolatile organic matter in raw coke as their majorfuel source.

Carbon Anode Energy Requirements.Approximately 0.61(6.01tf) kWh of energy wererequired and 0.12 kg CO2e were released in themanufacturing of the 0.45 kg of carbon anodeneeded to produce 1 kg of aluminum in 2003.

Anode blocks are typically baked in a naturalgas-fired furnace for several weeks. Quality an-odes depend on careful baking controls to gradu-ally raise the temperature to approximately1250 °C (2280 °F). Volatile hydrocarbons fromthe pitch are gradually released during the bakingprocess. Theoretically, these volatile compoundscould provide sufficient heat for anode bakingand no additional energy would be required,However, in practice, volatile organic com-pounds account for only 46% of the energy inputto the prebake ovens. The remaining 54% of theenergy needed comes from fuel. Only approxi-mately 30% of the input energy goes into makingthe anode; 24% is lost from oven surfaces, 29%goes up the stack, and 17% is lost in other ways.New prebake furnace ovens with computer con-trols are more efficient, with both regenerativeand recuperative elements (Ref 10.3).

Emissions associated with prebaked anodesresult mainly from the combustion of natural gasand the volatile organic compounds contained inthe pitch. These amount to 0.27 kg CO2e per

kilogram of anode or 0.12 kg CO2e per kilogramof aluminum (Appendix D, Table D.2).

The process energy used to produce a carbonanode is 1.36 (1.66tf) kWh/kg of anode, from themost recent available U.S. data for 1995 (Ref10.15). The total energy (Table 10.9) to produce acarbon anode is shown in Table 10.10. The energyper kilogram of aluminum produced is obtainedby multiplying 0.45 times the total energy required to produce a kilogram of carbon anode.

Soderberg anodes use green coke and arebaked in situ. These anodes can be baked onlyto the maximum cell-operating temperature,which results in an anode with 30% higher elec-trical resistivity and a greater dusting propensitythan a prebaked anode (Ref 10.3). The cellemission control system is also more complexthan for a prebake cell, because emission sys-tems must be designed to handle the 5 to 15%volatile organic material content of the greencoke (Ref 10.15). Some plants employ both wetand dry gas scrubbing systems to meet the envi-ronmental regulations.

Carbon Anode Theoretical Energy Values.The minimum theoretical energy requirementto manufacture a carbon anode is the energy

Table 10.9 Energy associated with aluminum industry carbon anode manufacturingMass of material input(a), Material input energy(b),

kg/1000 kg anode Btu/kg kWh/kg anode kWh/kg aluminum

PitchMass 231 . . . . . . . . .Feedstock energy . . . 8813 2.58 1.15Process energy . . . 9 0.003 0.001Calcined cokeMass 820 . . . . . . . . .Feedstock energy . . . 27,569 8.08 3.60Process energy . . . 395 0.12 0.05Green cokeMass 85 . . . . . . . . .Feedstock energy . . . 2664 0.78 0.35Process energy . . . 43 0.01 0.01Carbon anode bakingProcess energy(a) . . . . . . 1.90 0.85Total raw materials and energyMass (kg/1000 kg anode) 1136 . . . . . . . . .Feedstock energy (kWh) . . . 39,046 11.44 5.10Process energy (kWh) . . . 446 2.03 0.90

Total tacit energy input 13.47 6.01

(a) Values from Appendix E, Table E.1. (b) Values from Appendix B, Table B.1

Table 10.10 Energy associated with carbonanode manufacturing

kWh/kg of anode kWh/kg of aluminumA 0.45 � A

Pitch 0.003 (2.58tf) 0.001 (1.15tf)Coke 0.13 (8.99tf) 0.06 (4.01tf)Anode 1.36 (1.90tf) 0.61 (0.85tf)Total 1.49 (13.47tf) 0.66 (6.01tf)


necessary to convert the coal tar pitch by destruc-tive distillation to a coke-based binder. Approxi-mately one-third of the pitch binder mass is lostin the baking process (Ref 10.16). A portion ofthis loss consists of volatile components, whilethe remainder is the carbonization of the pitch.Pitch contains approximately 85% C. Approxi-mately 79% of the pitch is actually carbonized,while 21% is volatilized. The fuel energy valueof the pitch that is carbonized, 0.75 kWh/kg ofanode, is a measure of the theoretical energyrequired for anode manufacturing. In addition,the tacit or inherent energy of 11.55tf kWh/kgof anode must be accounted for as part of thetotal theoretical requirement. Therefore, thetotal theoretical minimum energy associatedwith the production of prebaked carbon anodesis 12.30tf kWh/kg of carbon anode. Further,because 0.33 kg of carbon anode are required toproduce a kilogram of aluminum in theory, theminimum energy requirement to produce an-odes is 4.1 kWh/kg of aluminum.

Primary Aluminum Production

The total energy associated with primary alu-minum production from bauxite ore was approx-imately 23.78 (45.21tf) kWh/kg of aluminum.This consisted of:

• 8.20 (14.22tf) kWh/kg aluminum for rawmaterials

• 15.58 (44.82tf) kWh/kg aluminum for elec-trolytic reduction

Alumina is insoluble in all ordinary chemicalreagents at room temperature and has a highmelting point (above 2000 °C or 3630 °F).

These properties make conventional chemicalprocesses used for reducing oxides difficult andimpractical for conversion of alumina intoaluminum.

Commercial primary aluminum is producedby the electrochemical reduction of alumina.Charles Martin Hall in the United States andPaul Lewis Toussaint Heroult in France inde-pendently developed and patented a commer-cially successful process for alumina reductionin 1886. This process, commonly referred to asthe Hall-Heroult process, is still the primarymethod in use for aluminum production.Although the engineering has improved vastly,the process fundamentals are basicallyunchanged today. The Hall-Heroult processtakes place in an electrolytic cell or pot. The cellconsists of two electrodes (an anode and a cath-ode) and contains a molten bath of fluoridecompounds (cryolite), which serves as anelectrolyte and solvent for alumina. An electriccurrent is passed through the bath, whichreduces the alumina to form liquid aluminumand oxygen gas. The oxygen gas reacts with thecarbon anode to form carbon dioxide. Moltenaluminum collects at the cathode in the bottomof the cell and is removed by siphon.

Production, Capacity, and Growth

In 2005, 9 companies operated 19 primary aluminum production facilities in the UnitedStates. These facilities, having a production capacity of approximately 3,708,000 metrictons, operated at approximately 67% capacity toproduce 2,480,000 metric tons of aluminum in2005 (Ref 10.9).

Figure 10.8 shows variations in the produc-tion of primary aluminum in the United States

Fig. 10.8 U.S. production of primary aluminum from 1960 to 2005


from 1960 to 2003. These production varia-tions are more representative of the costs toproduce aluminum than of the domestic demand.Primary aluminum is traded on a global mar-ket, and global supply has been growing steadilyat a rate of 4.5% annually for the past tenyears. The United States accounted for 7.8% ofthe world’s primary aluminum production in2005.

Historical Hall-Heroult Energy Utilization

“The first commercial aluminum cells atNeuhausen, Switzerland (Heroult), and Pitts-burgh, Pennsylvania (Hall), required more than40 kWh/kg of aluminum produced and had cur-rent efficiencies ranging from 75 to 78%” (Ref10.17). The Hall-Heroult process is still electricenergy-intensive. Because electricity costs arean important component (approximately one-third) of the total production costs, energyefficiency continues to be a major area of focusfor the aluminum industry.

The electrical energy consumed in a primaryaluminum cell is measured by the number ofwatts consumed over a period of time. Wattageis determined by multiplying the cell voltage bycell amperage. Figure 10.9 shows the significantelectrical energy improvements made between1900 and 2000. Total electricity usage (excludingtacit generation and transmission losses) variesfrom less than 13 kWh/kg of aluminum for thestate-of-the-art plants up to more than 20 kWh/kg for older Soderberg facilities (U.S.plants in 1995 averaged approximately 15.4kWh/kg of aluminum). The theoretical mini-

mum energy requirement for carbon anode alu-minum electrolysis is approximately 5.99kWh/kg of aluminum. Compared to theoreticalvalues, U.S. facilities are operating at approxi-mately 38% energy efficiency.

Significant engineering changes in cell designand operation have occurred over the past 50 years. Table 10.11 shows the changes in operating parameters of a typical cell from 1948to 1999 (Ref 10.18). Each new or updated pri-mary facility tries to increase productivity andincorporate energy-reducing technologies tolower production costs. This results in gradualchanges in the industry. The most significantchange is the new equipment and techniques al-lowing for smaller and more frequent aluminaadditions (point feeders). This, combined withhigher amperage, lower current density, andlarger cells, has dramatically improved currentefficiencies and productivity.

There is a minimum cell amperage (electricalcurrent) required to produce aluminum. Produc-tion in the United States now operates atapproximately 95% current efficiency, a signifi-cant improvement over the past several years, asshown in Fig. 10.10. The high current efficiencyof existing technologies leaves little opportunityfor process or technology improvements to further reduce amperage and save additional energy. Because current efficiency is high, low-ering the voltage requirements of cells presentsthe largest challenge and the best opportunity forimproving Hall-Heroult efficiencies. The voltagerequirements of a cell are described in the sec-tion “Voltage Requirements” of the “Hall-Her-oult Reduction Process” in this chapter.

50

40

30

IAI 2001

Average energy Consumption

Industry Range

Industry Range(Ref 10.3, 10.17)

U.S.IndustryRange (Ref 10.7)

Advanced anodeCathode systems

Alternative technologies

20

10

01900 1920 1940 1960 1980 2000

kWhkg

Fig. 10.9 Primary aluminum electric energy consumption from 1900 to 2000


Table 10.11 Ore-to-metal comparison of near-and midterm technology improvementsWetted Inert anode Kaolinite

Modern prebaked cathode wetted cathode Carbothermic (AlCl3) Energy input kWh/kg Al Hall-Heroult ACD=2.0 ACD=2.0 reduction reduction

On-site energy demandsRaw materials

Bauxite-alumina 7.59 7.59 7.59 7.59 . . .Kaolinite . . . . . . . . . . . . 8.14Anode materials 0.61 0.61 0.76 . . . 0.76Reaction carbon . . . . . . . . . 0 0Total 8.20 8.20 8.35 7.59 8.91

Reaction energyReaction thermal . . . . . . . . . 7.71 –1.90Furnace losses . . . . . . . . . 1.36 0.40Reaction 3.76 3.76 6.90 . . . 6.48Cell ohmic 10.67 7.62 6.20 . . . 2.93Total reaction 14.43 11.38 13.11 9.07 7.91Total onsite 22.63 19.58 21.46 16.66 16.82Percent energy Reactions 21% 9% 37% 45%savings Reactions and anode 20% 8% 40% 42%

Reactions, anodes, and ore 13% 5% 26% 26%

Tacit energy demandsRaw materials

Bauxite-alumina 8.21 8.21 8.21 8.21 . . .Kaolinite . . . . . . . . . . . . 8.81Anode materials 6.01 6.01 0.76 . . . 0.76Reaction carbon . . . . . . . . . 8.41 11.34Total 14.22 14.22 8.98 16.63 20.92

Reaction energyReaction thermal . . . . . . . . . 17.10 –1.90Furnace losses . . . . . . . . . 3.02 0.60Reaction 10.91 10.91 20.00 . . . 18.79Cell ohmic 30.91 22.07 17.98 . . . 8.48Total 41.83 32.99 37.99 20.12 25.98Total tacit 56.05 47.21 46.96 36.74 46.90Percent energy Reactions 21% 9% 52% 38%savings Reactions and anode 18% 19% 58% 44%

Reactions, anodes, and ore 16% 16% 34% 16%

ACD, anode-cathode distance

Fig. 10.10 Primary aluminum current efficiency from 1900 to 2000


Theoretical Minimum Energy Requirementfor Reduction

All current primary production facilities andmost alternative processes for aluminum produc-tion use alumina as their raw material. Anyprocess that starts with alumina to producealuminum has the same theoretical energy-requirement. Different processes do not offer anytheoretical energy advantage. However, they dooffer significant tradeoffs between efficiencies,emissions, footprints, and sources of input energy(electricity, carbon, and fuels). The theoreticallimits required to manufacture aluminum providea valuable insight into Hall-Heroult cell opera-tion and potential future reduction processes.

The product of primary reduction process ismolten aluminum. This report calculates thetheoretical minimum energy by assuming thereactants enter and the by-products leave thesystem at room temperature and that moltenaluminum leaves the system at 960 °C (1760 °F).The molten metal temperature, 960 °C (1233 °K),is an approximation of an average commercialoperating cell. Figure 10.11 illustrates the theo-retical boundaries for a system that reduces alu-mina to form aluminum and oxygen. Changesin the operating temperature of a cell have aminor effect on the theoretical energy require-ments. Operating changes of 100 °C (180 °F) ina Hall-Heroult cell, operating in the range of700 to 1100 °C (1290 to 2010 °F), result in lessthan a 1% change in the theoretical minimumenergy requirements (Appendix F, Table F.7).

Some studies assume that the gases evolvedduring reduction leave the system at the moltenmetal temperature. In these studies, the theoreti-cal minimum is 2.5 to 3% higher than the valuespresented in this chapter (Appendix F). Theoret-ically, it is possible to capture all the energyassociated with these gaseous emissions. Inpractice, however, the gas stream is collectedfrom hundreds of hooded pots and treatedbefore release to the atmosphere. Only a verysmall portion of the heat is actually absorbedand returned to the system.

Three energy factors must be examined in theproduction of aluminum: the energy required todrive the reduction reaction forward, the energyrequired to maintain the system at constant pres-sure and temperature, and the energy required tochange the temperature of the reactant and/orproduct. The thermodynamics and chemicalequilibrium of the reactions are described by thefollowing equation:

ΔG � ΔH � TΔS

The energy required to drive the reaction for-ward is the energy for the electrolytic reductionof alumina (2Al2O3 4Al � 3O2) and is givenby the change in the Gibbs free energy value(ΔG). The energy required to maintain systemequilibrium is the difference between the heatof reaction (ΔH) and the Gibbs free energyvalue (ΔG), which equals the entropy term(TΔS). Because the Gibbs free energyrequirement is less than the heat of reaction for

�

Fig. 10.11 Alumina-to-aluminum theoretical minimum energy


alumina reduction, additional energy must beadded to the system to maintain the systemtemperature. Otherwise, the system will cool asthe reaction proceeds. Hence, for the aluminareduction reaction, the ΔH term provides theminimum theoretical energy requirement (reac-tion and equilibrium). Reduction cells operate atatmospheric conditions, and no pressure changeresults from the reduction. Numeric values forthese thermodynamic measures for the elementsand compounds common to aluminum process-ing and the calculations used to determine thetheoretical minimum energy requirement aregiven in Appendix F. The energy required tochange the temperature of reactants and productsis calculated from their heat capacities (Cp),which are provided in Appendix F.

Faraday’s law provides the minimum amper-age requirement for electrolytic reduction. Thislaw states that 96,485 coulombs of electricityare passed through a cell to produce a 1 g equiv-alent of an element or compound. Aluminumhas an atomic weight of 26.98, a charge of 3�,and therefore has an 8.99 g equivalent weight.Faraday’s law is converted to more commonmeasurements:

The value 2980 Ah/kg of aluminum is the the-oretical minimum amperage (current) requiredfor production. This value assumes perfect con-ditions, where there are no reverse or parasiticreactions that consume amperage and no limita-tion to the ionic species availability to react atthe electrodes (no concentration gradients orgas bubbles). The Gibbs free energy (ΔG)divided by the Faraday amperage provides theminimum voltage required to drive the reactionforward. Cell voltage and current efficiency arevariables that are controllable by design, andthey determine the electrical power required forreducing alumina. In practice, electrolytic cellshave significant inefficiencies and operateabove the minimum voltage requirement. Thisexcess voltage provides the thermal energyrequired to maintain system equilibrium (ΔH –ΔG) and to produce molten material (Cp).

In the case of aluminum made directly fromalumina (2Al2O3 4Al � 3O2), shown in Fig.10.11, the energy required to drive the reaction

forward (ΔG) is 8.16 kWh/kg, the thermal energy(ΔH – ΔG) required to maintain thermal equilib-rium is 0.48 kWh/kg, and the thermal energy (Cp)associated with producing the molten aluminumis 0.39 kWh/kg of aluminum. The theoreticalminimum energy requirement is 9.03 kWh/kg ofaluminum. (Note: If the gas emission at 960 °C,or 1760 °F, is included, the total theoreticalminimum energy requirement is 9.30 kWh/kgof aluminum, (Appendix F, Table F.2).

Theoretical Energy for Hall-Heroult Car-bon Anode Reduction. The theoretical mini-mum energy requirement for producing moltenaluminum at 960 °C (1760 °F) in a Hall-Heroultcell with a carbon anode is 5.99 kWh/kg.

All commercial aluminum production uses acarbon anode in a Hall-Heroult cell. The carbonis consumed during the electrolytic process andsupplies part of the energy necessary for thereduction of alumina. This gives the Hall-Heroultcarbon anode process a lower energy require-ment than the direct reduction of alumina toaluminum. The theoretical energy required forreduction is the same for prebaked or Soderbergcarbon anodes.

The net reaction for the carbon anode Hall-Heroult process is 2Al2O3 � 3C 4Al � 3CO2.Figure 10.12 shows an idealized Hall-Heroultcell for the production of aluminum. In this cell,it is assumed that the reactants (alumina and car-bon) enter the cell at 25 °C (75 °F), the carbondioxide by-product leaves the cell at 25 °C (75 °F), and the aluminum product leaves asmolten metal at the cell operating temperature of960 °C (1760 °F). The reaction is assumed tooccur under perfect conditions, where there areno reverse reactions, no parasitic reactions con-suming additional anode carbon, no limitationsto the ionic species reacting at the electrodes,and no heat or energy losses external to thesystem.

Appendix F, (Table F.1) details the calcula-tion of theoretical minimum energy for thisreaction. The results show that the energyrequired to drive the reaction forward (ΔG) is5.11 kWh/kg, the thermal energy required tomaintain equilibrium is 0.49 kWh/kg, and thethermal energy associated with the molten alu-minum is 0.39 kWh/kg of aluminum. Therefore,the theoretical minimum energy requirement forthe reduction of alumina in a carbon anode celladds up to 5.99 kWh/kg of aluminum. (Note: Ifthe CO2 gas emission at 960 °C, or 1760 °F, isincluded, the total theoretical minimum energyrequirement is 6.16 kWh/kg of aluminum).

�

�

96 485, coulombs

g equivalent

Ampere/s

C

⎡

⎣⎢

⎤

⎦⎥

ooulomb

Hour

s

g equivalen

⎡

⎣⎢⎤

⎦⎥⎡

⎣⎢⎤

⎦⎥

×

3600

tt

8.99 g

g

kg2980 Ah/kg

⎡

⎣⎢

⎤

⎦⎥

⎡

⎣⎢

⎤

⎦⎥ =

1000


In actual carbon anode cell operations, cur-rent efficiencies of less than 100% result fromreverse oxidation reactions between part ofthe aluminum metal that is dissolved in thecryolite and carbon dioxide gas produced(2Al � 3CO2 Al2O3 � 3CO) and, to a lesserextent, between the carbon dioxide gas and thecarbon anode (CO2 � C2 CO). Current effi-ciency losses can also result from direct short-ing of the anode to the aluminum pad.

Today’s (2007) state-of-the-art reduction cellsare achieving current efficiency levels of morethan 96% and energy consumption levels of lessthan 13.0 kWh/kg of aluminum (Ref 10.19).The theoretical minimum energy requirement at100% current efficiency is 5.99 kWh/kg of alu-minum. The energy efficiency levels of presentstate-of-the-art carbon anode reduction cells areapproximately 46%.

Hall-Heroult Reduction Process

The engineering, materials, and process knowl-edge of existing components and processes formthe foundation for developing new components,processes, and techniques for producing alu-minum. The Hall-Heroult cell works as a systemwhere all the components perform together.Therefore, improving one component may notnecessarily result in an improved cell or a moreenergy-efficient operation. To understand the im-pact of component and system changes on thecell performance, it is helpful to know about theHall-Heroult process in terms of its components,operations, and interrelationships. This section

explains the process by describing a typical prebake anode operation.

Typical Hall-Heroult Cell Operation.A typical modern aluminum electrolysis Hall-Heroult reduction cell (pot) is a rectangular steelshell 9 to 12 m long, 3 to 4 m wide, and 1 to 1.5m deep (30 to 40 ft long, 10 to 13 ft wide, and 3to 5 ft deep) (Fig.10.13). It has an inner lining ofcarbon, which is surrounded by refractory ther-mal insulation, that keeps it isolated thermallyand electrically from the steel shell. Commercialcells range in capacity from 60,000 A to morethan 500,000 A and can produce more than 450to 4000 kg of aluminum per day, respectively.

A cell typically operates at 950 to 980 °C(1740 to 1800 °F) and yields molten aluminumand carbon dioxide. The molten aluminum has ahigher density than the electrolyte (cryolitebath) and settles to the bottom of the cell, on topof the carbon lining. Molten aluminum at ap-proximately 99.7% purity is periodically tappedby a vacuum siphon from the cell bottom. Thetapped metal is transferred to holding furnaceswhere the metal is alloyed, and entrained gasesand impurities are removed prior to casting. Thecarbon dioxide and other gases generated in thecell during the reduction process are collectedand treated to meet environmental regulations.

Electric current enters the cell through the car-bon anode and flows through 3 to 6 cm (1.2 to2.4 in.) of electrolyte (bath) to the aluminum padand carbon lining cathode. The aluminum pad isin intimate contact with the carbon lining andserves as the charged surface of the cathode.Steel collector bars are set near the bottom of the

�

�

Fig. 10.12 Alumina-and-carbon-to-aluminum theoretical minimum energy


carbon lining to conduct the current to the anodeof the next cell.

The 950 to 980 °C (1740 to 1800 °F) moltencryolite and aluminum used in a typical reduc-tion cell are corrosive. Molten cryolite has lowviscosity and interfacial tension that allows it toeasily penetrate any porosity in the cell lining.To protect the carbon lining, the thermal insula-tion is adjusted to provide sufficient heat loss tofreeze a protective coating of the electrolyte,known as a ledge, on the inner walls. The moltenaluminum pad protects the carbon bottom of thecell. The cell is never tapped completely dry ofmolten aluminum. It is essential that no aluminaor frozen ledge form under the metal pad. Thecarbon cathode must remain bare for good elec-trical contact with the aluminum pad and to min-imize wear of the carbon surface.

The reduction reaction is continuous, and alu-mina must be supplied to the bath at a controlledrate to maintain constant conditions. This is ac-complished with automatic feeders that breakthe surface crust and deposit alumina into themolten bath, where it is dissolved for reaction.Alumina is also used to cover the carbon anodesand the frozen bath surface. The alumina cover-ing serves as thermal insulation and as a protec-tive cover to reduce air burning of the anode.

The electrolytic reaction in a Hall-Heroultcell consumes the carbon anode. Approximately0.45 kg of the carbon anode are consumed foreach kilogram of aluminum produced. The car-bon anodes provide a necessary part of the energyrequired to operate a cell. The distance betweenthe carbon anode and the metal pad is kept con-stant by adjusting the anode as it is consumed

and as the aluminum is tapped. The consumablecarbon anodes must be replaced periodically,typically about every four weeks in a modernplant. The frequency of anode changing dependson the anode design and the cell operation.Anode changing represents the most frequentcell and productivity disruption. The removedportion of an anode (known as a butt) is recycledor sold as a fuel. The pot cover, which is part ofthe gas collection system, must be removed, theused anode must be pulled from the frozen sur-face crust, and the new anode must be insertedinto the space of the consumed anode. This mustbe accomplished without significant pot crustbreakage or alumina falling into the bath. Anodechanging is the single largest thermal, current,and magnetic disturbance in cell operation.

Cells are arranged in long rows called potlines.They are placed as close as possible to each otherwhile maintaining sufficient room for anodechanging, alumina feeding, and reasonably lowelectromagnetic interference. The cells are con-nected electrically in series. Rectifiers, whichconvert alternating current (ac) to direct current(dc), are chosen to minimize capital investmentand typically provide approximately 700 V (dc).Typically, a reduction cell design requiresapproximately 4.6 V (dc), so that a potline ofroughly 150 to 180 cells would be used.

Voltage Requirements. The energy con-sumed in an electrolytic reaction is a functionof the voltage used and the current efficiency ofthe operating cell (the minimum current is fixedby Faraday’s law; see the section “TheoreticalMinimum Energy Requirement for Reduction”in this chapter). Modern Hall-Heroult cellsoperate at high (�96%) current efficiencies.The approximate voltage components of a con-ventional cell are shown in Fig. 10.14.

The electric current flows through the cell,and the cell voltage components can be describedas a set of resistors in series:

E = Cell reaction � Overvoltage � Bath � Cathode

� Anode � Connectors

The cell reaction voltage is a function of tem-perature and, at 960 °C (1760 °F), is fixed at 1.2V (dc) (Ref 10.20). This is the theoretical mini-mum voltage required for the reduction reactionto take place, and no cell can operate at 960 °C(1760 °F) below this voltage. The total cell oper-ating voltage includes the addition of voltagesrequired to overcome the ohmic resistance of the

Fig. 10.13 Typical Hall-Heroult cell––half-section view


other cell components. These are described inthe following section.

Cell subsystems and variables include:

• Busbars and pot connectors• Electrolyte• Anode-cathode distance• Aluminum pad• Cathode• Current density• Cell polarization/overvoltage• Anode effects• Alumina feed• Cell operating temperature• Heat balance

Busbars and pot connectors electrically con-nect in series all the cells of a single potline,which typically contains more than 150 pots orcells. They are fabricated from highly conduc-tive aluminum alloy and are sized for minimumoverall system cost. Any voltage drop in the bus-bar and connector system results in energy loss.

The electrolyte or bath used in Hall-Heroult iscryolite (Na3AlF6). This is modified with the-addition of aluminum fluoride (AlF3), calciumfluoride (CaF2), and other additives to control theoperating temperature, solubilities, activities ofionic species, conductivity, viscosity, interfacialtension, bath density, vapor pressure, hardness

of the crust, and other factors. Bath chemistry,physical properties, and thermodynamics arevery complex. The bulk electrical conductivityof the bath is influenced not only by its compo-sition and temperature but also by the presenceof anode gas bubbles and carbon dust.

Aluminum fluoride (AlF3) is the most commonbath additive. It lowers the operating tempera-ture, the solubility of the reduced aluminum,surface tension, viscosity, and density. However,it has the undesirable effect of decreasing alu-mina solubility and electrical conductivity. Theweight ratio of NaF to AlF3 is referred to as thebath ratio. Controlling this ratio is important forefficient cell operation. (Note: Outside the UnitedStates, many countries use the cryolite or molarratio, which is twice the bath ratio.)

The fluid bath circulates within the cell. Asgas molecules are formed at the anode, theyaccumulate and coalesce into fine bubbles thataggregate into larger bubbles. These bubbles col-lect and move across the anode surface to escapearound the edges of the anode. The buoyancy ofthe gas creates movement, which contributes tothe motion of the bath and pad. The bath move-ment results from, in decreasing order of magni-tude, gas bubble drag and electrolyte densitydifference caused by the bubbles generated at theanode, electromagnetic forces on the moltenmetal pad, and temperature gradients. Thismotion influences the concentration gradients ofdissolved alumina and affects current efficiency.The motion also influences the heat transfer fromthe bath to the protective frozen ledge.

The anode-cathode distance (ACD) is the dis-tance between electrode surfaces. In the Hall-Heroult cell, it is the distance from the lower faceof the carbon anode to the top surface of the aluminum pad. This distance is typicallyapproximately 4 to 5 cm (1.6 to 2.0 in.). Theelectrolytic bath occupies the space between thecarbon anode and the aluminum pad. The voltagerequired for current to pass through the bath is related to the bath conductivity and the distancebetween the anode and the cathode. Decreasingthe ACD lowers the voltage and energy require-ments of the cell. The operating ACD is a com-promise between keeping a low value of bath resistance, while at the same time enabling elec-trolyte rich in alumina ionic species to reach thecharged surfaces and allowing reactant gas bub-bles to escape. The ACD also must be largeenough to ensure that the liquid metal does notcontact the anode and short-circuit the cell.

Fig. 10.14 Voltage distribution in a Hall-Heroult cell


The heat required to keep the bath molten is inpart supplied by the electrical resistance of thebath as current passes through it. The amount ofheat developed depends on the current path or theACD. Changing the ACD is one method of con-trolling the desired bath operating temperature.

The molten aluminum pad that forms at thebottom of the cell is the cathodically-chargedsurface for the reduction reaction. The largeamperage flowing through the cell creates electromagnetic forces and torques that causethe metallic pad to rotate. These magnetohydro-dynamic (MHD) forces create this motion,which deforms the molten aluminum/cryolite interface. The MHD forces cause local metal velocities of approximately 5 cm/s (2 in./s) tooccur. Control of MHD effects is one of the keyfactors for successful cell operation with highcurrent efficiency and low energy consumption.Cells are designed to minimize these forces, andtoday (2007), velocities are typically one-thirdof what was common 50 years ago. Movementof the aluminum pad is mainly caused by elec-tromagnetic forces in the cell and, to a smallerextent, by the interfacial drag of the bath fluid.Joint discontinuities in the carbon blocks createadditional flow disturbances in the moving pad.The combination of all these forces causes thepad to undulate or roll, which could result inwaves forming on the surface of the pad. Thesewaves can approach the anode and result in anelectrical short-circuit. The current that flowsduring this shorting produces no aluminum andresults in a major loss of power and productivity.The motion of the aluminum pad also produceserosion of the carbon lining and shortens celllife. Because the distance between the anode andcathode is constantly changing as a result of theundulating pad, the ACD is kept large enough toavoid contact between the anode and the pad.This requires that the anode be backed awayfrom its optimal position. Designing systems tominimize movements of the metal pad is a keyfactor in the efficient operation of a cell. A stablepad surface will allow the ACD to be decreased.

The newer concept of using a drained cathodecell is an approach to circumvent the difficultiesassociated with keeping the metal pad stable.Essentially, the bulk of the metal is drained to asump, and the cathode is left wetted only by athin metal sheet (see the section “AdvancedHall-Heroult Cells” in this chapter).

The bottom lining of the cell also serves as acathode and carries current from the moltenmetal pad. Because 10% of the total cell

voltage drop is in the carbon cathode blocks, itis important that they have the highest densityand electrical conductivity possible.

In an attempt to reduce the resistance of thecathode, some have experimented with cathodeblocks containing a higher content of graphiticcarbon. However, while less resistive, graphiteis also less wear resistant, and this compromisesthe life of the cathode. The cathode life generallydetermines cell life, because cathode replace-ment requires the complete dismantling of a cell.The advent of hard titanium diboride (TiB2)coatings may offer an opportunity to increase thegraphitic content of cathode blocks, and lowercell resistance, without reducing the cell life.Under optimal conditions, the life of a cathodeor cell is in the range of 7 to 10 years.

Current density is a measure of the produc-tivity of a cell. It is calculated by dividing theamperage supplied to an anode by the geomet-ric face area of the anode. It is generallyexpressed in amperes per square centimeter(A/cm2). Most potlines operate in the range of0.8 to 1.0 A/cm2. The quantity of aluminumproduced per cell increases with increasingcurrent density. The tradeoff is that as currentdensity and productivity increase, current effi-ciency decreases, which results in a higherenergy consumption per unit of metal produced.Lower current densities are more energy effi-cient but increase capital and labor costs perunit of output.

The reactions occurring at the anode and thecathode create localized conditions that are dif-ferent from the bulk of the bath. The reactionsdeplete the supply of reactants and increase thequantity of products. This creates concentrationgradients, which in turn cause concentrationpolarization. Additionally, the gas generated atthe anode forms bubbles, which lower the effec-tive bath conductivity. Localized conditionsat the anode and cathode are unavoidable andrequire a voltage higher than the minimum reac-tion voltage to be applied to the cell.

No free aluminum (Al3+) or oxygen (O2–)ions are present in the bath. Alumina dissolvesand dissociates into salt complexes in the bath.The dynamics at the anode are complicated bythe release of gas bubbles. Oxygen-containingionic species are transported through the bathand discharged on the carbon anode. Theseanode reactions are:

Al O F + 2F + C CO + 2AlF + 4 and

Al

2 2 62– –

2 4– –→ e

22 2 42– –

2 4– –O F + 4F + C CO + 2AlF + 4→ e


At least three phenomena have been identi-fied as contributing to the anode overvoltage:

• An increase in the local current density dueto the presence of gas bubbles adjacent tothe anode surface, which displace the elec-trolyte bath

• An ohmic component from an increase inthe resistance of the electrolyte due to thepresence of the bubbles

• The concentration polarization overvoltage(Ref 10.21)

Polarization effects at the cathode contributemuch less to overvoltage than at the anode. Thecathode reactions are and . The aluminumion complexes, and have higherionic mobility than their anodic counterparts,which lowers the concentration polarizationeffect. In addition, no gas bubbles, which influ-ence both resistance and concentration polariza-tion, are produced at the cathode.

Control of the quantity of alumina dissolved inthe bath is important for proper operation of acell. Alumina saturation is reached at approxi-mately 7% alumina dissolved in a typical bath.The normal operating level is approximately 3%alumina. If the level goes above 4%, some of theadded alumina may not dissolve rapidly and cansettle to form a sludge on the cell bottom, therebyreducing the cell conductivity. If it falls below1% alumina, the cell is starved of the reactant,and an anode effect ensues. When this occurs, theproduction of metal is interrupted, and fluorine(F2), hydrogen fluoride (HF), and two perfluoro-carbon gases—tetrafluoromethane (CF4) andhexafluoroethane (C2F6)—are discharged insteadof carbon dioxide. These perfluorocarbon gaseshave 6000 to 9000 times the global-warmingpotential of carbon dioxide and require gasscrubbing to meet environmental regulations.

Significantly, alumina is a good absorbent offluorine and hydrogen fluoride but not perfluo-rocarbon gases. This allows primary productionfacilities to use their alumina raw material as anabsorbent in dry gas scrubbing systems. The flu-orides absorbed on the alumina in the scrubbersare recycled into the potfeeding systems, so thatboth the alumina and fluorides can be reused inthe process.

Alumina is ideally added to the cell at a ratethat exactly replaces the alumina that has beenreduced. If alumina is fed too fast or in largeincrements, it may not dissolve and can formsludge. Sludge affects fluid flows within the cell

and contributes to erosion of the cathode blocksurface. Underfeeding the cell results in ananode effect. There is no current technology forin situ, real-time bath analysis to provide pre-cise control of alumina concentration in the bathand alumina feed rate.

Alumina is fed by automatic handling and con-veyor systems. The specifications for aluminaare complex, with numerous tradeoffs. It mustdissolve rapidly; it should contain few impuritiesand have a high surface area, yet be a relativelylarge particle (this apparent inconsistency isovercome by the particle having significantporosity). The particle must also be robust, createlittle dust, and resist breakage during handling.The introduction of dry scrubbing systems forthe cell offgases, wherein the alumina is used asan adsorbent for fluoride emissions, further com-plicates the alumina specifications.

Alumina is side-worked in older prebake celldesigns and Soderberg cells. Side-worked cellsintroduce alumina into the bath using automatedcrust-breaking and feeding machines that movealong the length of the cell. Side-working istime-consuming and can take 1 to 4 h before themachine feeds the same section of the cellagain. Newer prebaked cell designs use thespaces between the anodes to feed alumina intothe cells. Point feeders pierce the crust and dis-pense small quantities of alumina at numerouspoints, typically in the center of the cell. One-minute intervals between point feeding arecommon. This frequent addition of small quan-tities of alumina takes advantage of the motionof the bath and provides significantly bettercontrol of local alumina concentration. Thisprovides for better current efficiency, feweranode effects, and less erosion caused by solidalumina. Point feeders have proven to be signifi-cantly better and have replaced most feedingsystems in both prebaked and Soderberg cells.

Bath chemistry controls the operating temper-ature of the Hall-Heroult process. Most com-mercial cells operate near 960 °C (1233 K).Reducing the cell operating temperature is anobvious approach to saving energy and reducingcapital costs by lowering insulation require-ments. However, controlling the operatingvariables of a cell becomes more critical astemperature is lowered. Dissolved alumina mustbe available for reduction in the bath to have highcell productivity and minimal energy-consumingconcentration polarization and anode effects.Lowering the cell temperature lowers the solu-bility range for alumina in a cryolite bath. The

AlF4−AlF6

3−( )– –AlF Al 4F4

–+ → +3e( )– –AlF Al 6F6

–3 3+ → +e


solubility or phase diagram for a cryolite alu-mina system is V-shaped with temperature on they-axis and alumina concentration on the x-axis,as shown in Fig. 10.15. Cells must operate withinthe “V”; operations to the left will precipitatecryolite, and those to the right will precipitatealumina. Decreasing the temperature narrowsthe cell-operating range with respect to aluminasolubility. In practice, bath chemistry is modifiedwith the addition of aluminum fluoride, calciumfluoride, and other salts. These modificationslower the liquid-phase temperature. However,the basic V-shape of the diagram is retained.

Controlling the thermal balance of the cell isof prime importance for efficient operation andlong cell life. There is no commercially availablematerial that can retain its insulating value andresist penetration and chemical attack by cryo-lite at cell-operating temperatures. Accordingly,the method used to protect the side walls is toallow some heat loss so that the temperature ofthe exposed surface of the cell lining is belowthe freezing point of the bath. This creates afrozen layer or ledge of bath that protects thelinings. Side-wall heat losses can account for35 to 45% of the total heat loss in a cell (Ref10.3). The frozen ledge, due to phase relation-ships, differs in composition from the bath. If thetemperature rises above steady state, the ledgebegins to dissolve, and the bath ratio (sodiumfluoride to aluminum fluoride) changes. If thetemperature is allowed to fall too much, ledgeformation is excessive, anodes are changed withgreat difficulty, alumina does not dissolve as

readily, and the bath ratio is affected. The frozenelectrolyte ledge also provides the electricalinsulation of the side walls. The thermal con-ductivity of frozen cryolite is an order of magni-tude lower than that of the molten cryolite.

Large cell designs require less energy tomaintain operating temperatures because of thelower ratio of cell surface area to volume. Thisis one factor in the trend to use larger cells innewer smelting plants.

Environmental Considerations

The Energy and Environmental Profile of theU.S. Aluminum Industry (Ref 10.8), compiledby the U.S. Department of Energy, Office ofIndustrial Technologies Program, and LifeCycle Inventory Report for the North AmericanAluminum Industry (Ref 10.7), published byThe Aluminum Association, provide detailedenvironmental information on the overall alu-minum production process. The by-productsgenerated in the Hall-Heroult process that areof environmental concern can be grouped intothree areas: electrolysis, anode production, andcell waste products.

Electrolysis. Electrolysis green house gas(GHG) emissions in the Hall-Heroult processcan be split into three groups: reduction reactionemissions, carbon dioxide (CO2) and carbonmonoxide (CO); process upset perfluorocarbonsemissions; and hydrogen fluoride (HF) formedfrom the inclusion of moisture (H2O) in the rawmaterials. Hydrogen fluoride gas is almostcompletely captured and returned to the cellsby the alumina dry scrubbing system used inmodern facilities.

The carbon-base emissions associated withthe reduction reaction come from three sources:

• Reaction products: The reaction producesoxygen that reacts with the carbon anode toproduce CO2 and small quantities of CO;this reaction produces 1.22 kg of carbondioxide equivalents for each kilogram of alu-minum produced (Appendix D, Table D.4).

• Air burning: The carbon anode loses mass tooxidation with the atmosphere. This produces0.30 kg of CO2 for each kilogram of alu-minum produced (Appendix D, Table D.4).

• Electricity generation and transmission: Theemissions related to the fuels used in elec-tricity generation for U.S. primary facilitiesare 5.38 kg of carbon dioxide for each kilo-gram of aluminum produced (Appendix D,Table D.2). Fig. 10.15 Cryolite-alumina phase diagram


A total of 6.90 kg of CO2 gas is generatedfrom the reduction process for each kilogram ofaluminum produced in an average U.S. primaryfacility. It should be noted that electricity-related emissions for specific potlines varywidely. Potlines operating on electricity obtainedfrom coal-fired power plants produce 16.0 kg ofCO2 gas for each kilogram of aluminum pro-duced, while potlines using electricity fromhydropower plants produce close to zero CO2gas emissions.

The perfluorocarbon emissions are related tothe anode effect. If the concentration of aluminain the bath becomes too low, other reactions be-tween the carbon anode and the bath occur, andtetrafluoromethane (CF4) and hexafluoroethane(C2F6) are generated. These gases have a highglobal-warming potential (GWP). The GWP of aGHG is a ratio developed to compare the abilityof each GHG to trap heat in the atmosphere rela-tive to CO2 gas. The GWP of CF4 and C2F6 is6500 and 9200, respectively. In other words, 1 kgof CF4 released to the atmosphere is equivalentin its warming potential to 6500 kg of CO2.

Aluminum smelting is the principal quantifi-able source of perfluorocarbon in the UnitedStates. The U.S. Environmental ProtectionAgency (EPA) estimates U.S. emissions fromaluminum production at 500 metric tons of CF4gas and 50 metric tons of C2F6 gas in 2003 (Ref10.22). In 1995, the aluminum industry enteredinto a Voluntary Aluminum Industry Partner-ship (VAIP) with the EPA to reduce perfluoro-carbon emissions by 46% during the nextdecade. Reductions in primary aluminum pro-duction and efficiency improvements to reduceanode effects have reduced emissions of CF4and C2F6 gases since 1990 by 71% each. TheU.S. aluminum industry and EPA are continu-ing the VAIP to seek further GHG reductionsbeyond the original achievements. The total2002 U.S. aluminum perfluorocarbon emis-sions are 2.0 CO2 equivalent metric tons per tonof aluminum.

The emissions of perfluorocarbon for olderside-fed cells are 1 order of magnitude higherthan the emissions for cells with point feeders.The industry continues to improve point feedsystems. Ultimately, with the entire industryon point feeders and advanced cell control sys-tems, it should be possible to virtually elimi-nate anode effects and hence perfluorocarbongeneration.

Anode Production. The CO2 emissions forthe carbon anode manufacturing amount to

approximately 0.13 kg of CO2 equivalents perkilogram of aluminum. In this report, the feedstock portion of the CO2 emissions for car-bon anode manufacturing is included in theelectrolysis. More specific information on thistopic can be found in the comprehensive life-cycle information published by The AluminumAssociation (Ref 10.7).

Cell Waste Products. Aluminum electrolysiscarbon slime and spent potlining (SPL) areunavoidable by-products of the aluminumsmelting process and are listed in the UnitedStates as hazardous wastes. Development workis underway to mitigate problems associatedwith SPLs. Most development efforts attemptto combust the carbon linings to destroy anyremaining toxic chemicals, to recover the valu-able fluoride as AlF3, and to render the remainingmaterial inert through vitrification. In some areasof the world, SPL is destroyed by combustion inthe production of cement.

Technological Change in the Next Decade

The Hall-Heroult electrolysis process, using acarbon anode and cryolite bath, is a maturetechnology. However, gradual improvements inboth productivity and environmental perform-ance are still possible. The typical Hall-Heroultcell life ranges from 7 to 10 years. Adoptionrates of new technology and systems are gov-erned to some degree by the cell life. There is aslow, autonomous efficiency improvement inthe Hall-Heroult process because of the contin-ued adoption of advanced cell designs, feedingsystems, bath composition, control systems, andother practices. This trend has resulted in agradual decline in energy consumption in therange of 0.2 to 0.5% per year. Energy savingsare actively pursued by aluminum producersbecause electricity costs constitute a high per-centage (20 to 30%) of total production costs.Because current efficiencies are already over95%, the goal is generally to reduce the over-voltage in the aluminum cells to increase theoverall electric efficiency.

A number of technological and engineeringimprovement options exist and are being adoptedby the aluminum industry. These include:

• Point Feeders: Point feeders enable moreprecise, incremental alumina feeding forbetter cell operation. Point feeders are gen-erally located in the center of the cell andthereby cut down on the diffusion requiredto move dissolved alumina to the anodic


reaction sites. The controlled addition ofdiscrete amounts of alumina enhances the dissolution process, which aids in improv-ing cell stability and control, minimizinganode effects, and decreasing the formationof undissolved sludge on the cathode. Inthe jargon of modern commerce, pointfeeders enable “just-in-time alumina sup-ply” to permit optimal cell operation. Pointfeeder improvements continue to be madeas more accurate cell controllers becomeavailable.

• Improved process controls: Advanced processcontrollers reduce the frequency of anodeeffects and control operational variables,particularly bath chemistry and aluminasaturation, so that cells remain at theiroptimal conditions.

• Slotted anodes: The use of transversally slot-ted anodes helps to facilitate the removal ofanode gases from underneath the anode. Thisconsequently lowers the overall anode polari-zation, allowing for incremental improve-ments in productivity and cell amperage.

Changes to the Hall-Heroult Cell and Alternative Technologies. Two innovativetechnological changes to the Hall-Heroultprocess, the wetted drained cathode and theinert anode, are on the near-term horizon forimproving energy efficiency. These technolo-gies can, with cell modifications, be retrofit intoexisting potlines and supporting infrastructure.Wetted cathodes are anticipated to lower energyconsumption of a Hall-Heroult cell by 15tf%when compared to a modern Hall-Heroult cell.This report defines a modern cell as one thatoperates at 4.6 V (dc) and 95% current efficiencywith the voltage distribution shown in Fig.10.14. The combination of an inert anode with awetted cathode could provide a 18tf% reductionin energy consumption and the elimination ofcell CO2 emissions. These technologies aredescribed in the section “Advanced Hall-Her-oult Cells” in this chapter, and the energy im-pacts are calculated in Appendix H. Multipolarcells using Hall-Heroult chemistry require theuse of inert anode and wetted cathodetechnologies. The multipolar design allows fora more compact, more productive cell withsignificant thermal energy savings. The section“Advanced Hall-Heroult Cells” also describesmultipolar electrolytic cells.

Two alternative technologies to the Hall-Heroult process, carbothermic reduction and

kaolinite reduction, have been studied by severalgroups for many years. These alternative tech-nologies could displace existing Hall-Heroultcells in the future. They are described in the sec-tion “Alternative Primary Aluminum Processes”in this chapter. Both these processes could po-tentially change where and how the aluminumindustry operates, while lowering energy con-sumption. These alternatives consume more car-bon and have higher on-site carbon emissionsthan the Hall-Heroult process. However, theirelectrical demands are lower, which results inlower overall (utility-to-metal) CO2 emissions.The carbothermic process is anticipated to save21tf% in energy and be economical at a muchsmaller scale ( ) than the Hall-Heroult facili-ties. The kaolinite reduction process is antici-pated to save approximately 15tf% of the energyrequired for a modern Hall-Heroult system.This value is impacted by the need to prepareadditional ore mass and carbon for the process.However, the kaolinite reduction process iscommercially interesting because of its loweron-site energy demands, domestically availableore, and lower-cost raw materials.

Table 10.11 summarizes the estimated energyperformance of these near- and midterm tech-nologies (Appendix H, Table H.4). The on-siteand tacit energy values in the table allow theprocesses to be compared on a reaction, rawmaterial, or a complete ore-to-metal basis. Thetable provides the energy associated with anodeproduction and feedstock energies. The energyperformance of the near-term technologies,wetted cathode and inert anode, are based onvoltage changes in the electrolysis cell (Appen-dix H, Tables H.1 and H.2). These voltagechanges are supported by theory and reportedexperiments and provide a good estimate ofenergy use. The energy values reported formidterm technologies, carbothermic reductionand kaolinite reduction, are approximationsbased on the theoretical energy requirementsand assumed reactor inefficiencies. Bothmidterm technologies involve multiple reactionand separation zones. To date, no fully inte-grated reactor systems have been built. Thesemidterm energy approximations assume thatthere is significant heat integration (recovery)within a facility.

Many technical hurdles remain to be solvedin these new processes before they becomecommercially viable. Wetted cathodes and inertanodes will be adopted as their performancesare proven and existing cells need rebuilding.

110


Industry will require significant demonstrationtime before adopting any alternative to Hall-Heroult technology.

Advanced Hall-Heroult Cells

The Hall-Heroult cell system efficiency canbe improved with the adoption of new cell tech-nologies, described subsequently. Wetted drainedcathode technology offers the most significantimprovement in energy efficiency, but its adop-tion will be gradual. Typical cell life is seven toten years. The industry will optimize the capitalinvested in existing cells before retrofitting withnew technology. Inert anodes offer significantenvironmental benefits, lower maintenancecosts, and can be retrofit into the existing cells.Inert anodes could be adopted more quickly byindustry because the carbon anodes are replacedapproximately every four weeks. However, theirsuperior performance must still be proven inindustrial trials.

Wetted Drained Cathode

Wetted drained cathodes allow the anode-cathode distance to be reduced and are expectedto result in a 20 (18tf)% reduction in the elec-trolysis energy required to produce aluminum.

Molten aluminum does not wet the carbonlining of a Hall-Heroult cell. The aluminum padrests on an extremely thin sheet of cryolite bath.This creates an electrical junction similar to theair gap between two metal bars and causes asmall voltage drop. If the bars are clampedtighter, providing better contact, the junctionvoltage drop decreases. The thicker (heavier)the aluminum pad is, the thinner the cryolitesheet becomes and the lower the junction volt-age drop. Modifying the cathode surface tomake it more wettable would allow the sameelectrical contact with a decreased thickness inthe pad. A thinner pad would be more hydrody-namically stable, have lower wave height, andallow a decrease in the anode-cathode distance(ACD). A decrease in the ACD results in energysavings. A cell lining that is completely wettedand inert to cryolite would be even more effi-cient. This combination of properties wouldallow the aluminum pad to be drained out of theanode-cathode spacing. Removing the unstablealuminum pad would allow the ACD to beconsiderably reduced and provide significantenergy savings.

Titanium diboride (TiB2) has been found tobe a wettable cathodic material, and several ap-proaches to incorporate TiB2 into a Hall-Heroultcell are being studied (Ref 10.23). Cathodesmade wettable with TiB2 appear to increase celllife by making the cathode less susceptible topenetration by bath material. This is a consider-able benefit, because cell rebuilding costs are amajor contribution to primary aluminum opera-tional costs. Longer-life cells also generate lesscell lining waste material (spent potliner) perton of aluminum produced. These benefits haveto be balanced with the higher costs associatedwith the TiB2 material. Recent evidence alsosuggests that the wetted cathodes reduce theformation of sludge (undissolved alumina) onthe cell bottom and improve cell operations (Ref10.24). Chalco is testing and has made wettedcathode blocks available to other primary alu-minum producers.

Several concepts for wetted cathode anddraining cells have been proposed. Figure 10.16shows:

• A conventional cell• A low metal pad with a wetted cathode in a

conventional cell• A hybrid cell with a metal sump• A fully drained slanted cell

The decrease in ACD is shown for each con-figuration. These configurations offer attractivealternatives because they can be incorporatedas a retrofit to the existing facilities. Some ofthese configurations are currently being evalu-ated in commercial cells and could soon beavailable (Ref 10.24). However, as ACD isdecreased, less bath is available for circulationand mixing. This requires retrofit cell designsto account for the change in bath dynamics.Designs must ensure that dissolved alumina isavailable across the anode surface to maintainhigh current efficiency and productivity and toavoid anode effects. Designs also need to compensate for the heat energy lost due to thelower cell voltage operation.

Conceptually, the fully drained inclined cath-ode design offers the greatest potential forenergy savings. By inclining the anode and cath-ode a few degrees, the molten metal pad (withall its complexities) is removed completely to asump. Without the pad, the ACD is dimension-ally stable and can be narrowed, which signifi-cantly reduces the total electrical resistance ofthe bath. The alumina feeding of the bath is notcompromised, because the buoyancy of the gas


bubbles generated during reduction causes bathcirculation, and fresh alumina is drawn into theACD gap and across the electrode face.

Energy Savings for Wetted Cathode Tech-nologies. Decreasing the ACD results in a pro-portional decrease in the voltage drop associatedwith the electrolytic bath. Energy is saved whenthe ACD reduction is matched with the ability tomaintain current efficiency and heat balance.The energy savings for the different wetteddesigns shown in Fig. 10.16 are estimated inAppendix H, Table H.1.

The modern Hall-Heroult cell, shown in Fig.10.16(a) has an ACD of 4.5 cm (1.8 in.) and avoltage distribution similar to that shown in Fig.10.14. The traditional cell voltage distributionhas 1.75 V (dc) associated with the ACD, whichaccounts for approximately 38% of the total 4.60V (dc) drop across the cell. As discussed in theprevious section, a wetted cathode provides bet-ter electrical contact between the metal pad andthe cathode, allowing the ACD to be decreased. Ifthe ACD were lowered from 4.5 to 3.5 cm (1.8 to1.4 in.), as shown in Fig. 10.16(b), the voltageassociated with the bath would be proportionallylowered to 1.36 V (dc). The total voltage woulddecrease to 4.21 V (dc), which would provide an8% reduction in the electrical energy usage.

Draining the metal pad into a sump wouldeliminate the unevenness of the pad and permitan even smaller ACD with greater energy sav-ings. If the ACD were lowered from 4.5 to 2.5cm (1.8 to 1.0 in.), as shown in Fig. 10.16 (c),the voltage associated with the bath would beproportionally lowered to 0.97 V (dc). The totalvoltage would decrease to 3.82 V (dc), providinga 16% reduction in the electrical energy usage.

Reducing the ACD is limited by the ability totransport the reactant (dissolved alumina) to theelectrode interface and to remove products (alu-minum and carbon dioxide) from the electrodeinterface. Sloping the electrode interfaceslightly, as shown in Fig. 10.16(d), removesproducts and supplies reactants more effectivelyby using the buoyancy of the gas to induce bathcirculation. It is estimated that under the bestconditions, the ACD for a sloped configurationcould be reduced to as little as 2.0 cm (0.8 in.)(Ref 10.25, 10.26), in which case, the voltageassociated with the resistance of the bath wouldbe 0.78 V (dc). The total voltage required wouldthen decrease to 3.63 V (dc), providing nearly a20% reduction in the electrical energy usage.

Environmental Impacts for Wetted Cath-ode Technologies. The by-products generatedin the Hall-Heroult process that are of environ-mental concern are grouped into three areas:electrolysis, anode production, and cell wasteproducts. Of these, the wetted cathode technol-ogy impacts electrolysis and cell waste products.

Electrolysis GHG emissions can be split intothree groupings: reduction reaction emissions,process upset perfluorocarbons emissions, andhydrogen fluoride emissions from the bath. Wet-ted cathode technology does not change theemissions related to process upsets or bath emis-sions. The reduction reaction emissions comefrom three sources: the reaction products, anodeair burning, and electricity generation and trans-mission. Wetted cathodes do not change thereaction products or anode air burning. Both wet-ted and traditional cathodes will produce 1.22 kgof carbon dioxide from the reduction reaction(2Al2O3 � 3C 4Al � 3CO2), 0.30 kg of�

Fig. 10.16 Conceptual wetted cathode cells. (a) Traditional Hall-Heroult cell. (b) Wetted cathode with reduced metal pad. (c) Wetted drained cathode. (d) Wetted sloped drained cathode. ACD, anode-cathode distance


carbon dioxide for each kilogram of aluminumproduced from air burning of the carbon anode,and 0.12 kg of carbon dioxide for each kilogramof aluminum for the fuels associated with man-ufacturing the anodes.

The environmental benefit of wetted cathodetechnology is related to the emissions associatedwith the electricity production. A wetted cathodelowers the electrical energy requirement, whichin turn reduces the emissions related to the fuelsused in electricity generation and transmission.The electricity production (14.4 kWh/kg ofaluminum) for a modern Hall-Heroult cell emits5.04 kg of carbon dioxide equivalents (CO2e) foreach kilogram of aluminum produced. A wetted-sloped cathode cell with a 2.0 cm (0.8 in.) ACDwill lower CO2e associated with electricity gener-ation and transmission by nearly 21% to 3.98 kgCO2e/kg of aluminum produced. This lowers thetotal CO2e associated with a wetted-sloped cath-ode cell from 10.62 to 7.90 kg CO2e/kg of alu-minum produced (Appendix D, Table D.4).

One of the benefits of using wetted cathodesis longer cell life. Rebuilding cells less fre-quently will lower the quantity of spent potlinerwaste per unit of aluminum produced.

Hall-Heroult Inert Anode

Carbon anodes are consumed in the Hall-Heroultprocess, making the continuous manufacture ofnew anodes and constant changing of the con-sumed anodes necessary. Anode changing upsetsthe stability, production, and energy efficiencyof the cell. For more than 125 years, Charles Halland other primary metal producers haveattempted to find an inert anode that would elim-inate the manufacturing and handling of consum-able anodes (Ref 10.20). The material demandsof an inert anode require that it be highly conduc-tive, and be thermally and mechanically stable at800 to 1000 °C (1470 to 1830 °F). Further, itshould not react or dissolve to any significantextent in the cryolite, or react or corrode in the800 to 1000 °C (1470 to 1830 °F) oxygen-containing atmosphere. Any undesirable reactionwith the bath must occur at a very slow rate,because this will result in the anode materialcontaminating the aluminum product.

Few materials are truly inert under the extremeconditions of a cell. Research has focused onthree classes of inert anode materials: ceramics,cermets, and metals. The challenges to findingthe most efficient material are substantial.Ceramics typically have poor thermal-shock

resistance, are not mechanically robust, displaypoor electrical conductivity, and are difficult to connect electrically. Metals have good thermal,mechanical, and electric properties but areattacked by the hot oxidizing atmosphere. Metaloxides are somewhat soluble in cryolite and canresist the hot oxygen atmosphere, but theyexhibit lower electrical conductivity than metals.When in solution, they electrochemically reduceand contaminate the aluminum metal. Metaloxide solubility can be reduced dramatically incryolite by operating with a bath saturated in alu-minum oxide and alumina. However, this pres-ents significant bath feeding challenges. Cermetscombine the advantages and disadvantages ofceramics and metals. In addition to overcomingtechnical hurdles, the likely higher cost of manu-facturing inert anodes of commercial size mustbe compensated with longer life, lower energyconsumption, and higher productivity.

An inert anode would enable greater controlof the critical ACD, which represents thelargest voltage drop in the cell (Fig. 10.14).When used in conjunction with a drained cath-ode, it is estimated that an inert anode may saveup to 21tf% of the energy required for reducingaluminum. Inert anodes offer a major environ-mental advantage and the potential of produc-ing a valuable coproduct. Replacing the carbonanode with an inert material results in oxygenbeing discharged rather than carbon dioxide. Ifa significant market for oxygen is near thereduction facility, the oxygen produced couldbe collected and sold as a coproduct. Carbon credits are an unknown but potentially largeeconomic force that could hasten the develop-ment of inert anodes.

Inert anode technology could potentially beretrofit into the existing cells with limitedchanges and use the existing alumina- andaluminum-handling infrastructures. Some elec-trical infrastructure changes would be requiredbecause the inert anode will operate at a highervoltage than the carbon anodes. Because frequentaccess to the cells is not required for changinganodes, cells can be sealed more effectively toprovide better gas collection and treatments.

Notable progress in the production and test-ing of potential inert anode materials has beenmade in recent years. (Ref 10.27). Some com-panies are now conducting trials with relativelystable materials that offer the promise of inertanode performance. Using an inert anode andwetted cathode could also lead to the design ofmultipolar, vertical electrode cells, which would


increase productivity and further reduce energyusage.

Theoretical Energy for Hall-Heroult InertAnode. The theoretical minimum energyrequirement for an ideal Hall-Heroult cell usingan inert anode can be calculated from the ther-modynamics of the reaction, 2Al2O3 4Al �3O2 (Appendix F, Table F.2). The production ofaluminum shown in Fig. 10.17 assumes thatthe reactant (alumina) enters the cell at 25 °C(75 °F), the oxygen by-product leaves the cell at25 °C (75 °F), and the aluminum product leavesthe cell as molten metal at 960 °C (1760 °F).The theoretical minimum energy requirement iscalculated under perfect conditions: an environ-ment with no reverse reactions, parasitic reac-tions, heat/energy losses external to the system,or limitations to the ionic species reacting at theelectrodes. The results show that the energyrequired to drive the reaction forward (ΔG) is8.16 kWh/kg, the thermal energy (ΔH – ΔG)required to maintain equilibrium is 0.48 kWh/kg,and the thermal energy (Cp) associated with themolten aluminum is 0.39 kWh/kg of aluminum.The total theoretical minimum energy require-ment for an inert anode adds up to 9.03 kWh/kgof aluminum. (Note: If the O2 gas emission at960 °C, or 1760 °F, is included, the total theo-retical minimum energy requirement is 9.30kWh/kg of aluminum.)

In actual cell operations, current efficiencies ofless than 100% can result from reverse oxidationreactions between part of the metallic aluminumand the oxygen gas produced by the reaction.Inert anode cell designs must ensure that oxygen

bubbles are discharged from the system withoutcontacting the molten aluminum being produced.

Comparative Energy Requirements forCarbon and Inert Anode Hall-Heroult Cells.The theoretical minimum energy requirementfor an inert anode reaction is 51% higher thanthe carbon anode requirement. Both the Hall-Heroult carbon and inert anode processes usethe same starting raw material, alumina, andrequire the same total theoretical energy, 9.03kWh/kg of aluminum. The Hall-Heroult carbonanode is consumed as part of the reduction reac-tion and provides part of the energy to the sys-tem. The energy provided by oxidizing thecarbon (3.04 kWh/kg of aluminum) lowers thecarbon anode system theoretical energy require-ment to 5.99 kWh/kg of aluminum. The inertanode must supply the full minimum energy re-quirement for alumina reduction. The inertanode system requires 2.2 V (dc) compared tothe Hall-Heroult reaction requirement of 1.2 V(dc) (Ref 10.28). This 83% increase in reactionvoltage supplies the additional 3.04 kWh/kg ofaluminum minimum energy required by theinert anode system to reduce alumina.

Although the inert anode reaction voltage is1 V (dc) higher than a carbon anode Hall-Heroultcell, three factors are expected to provide theinert anode with an overall improved operationalenergy performance than the carbon anodes:

• Elimination of carbon anode manufacturing• Reduction of anode polarization overvoltage• Reduction in ACD that results from a stable

anode surface

�

Fig. 10.17 Theoretical energy schematic for an inert anode


The impacts of these three factors are shownin Table 10.12, and calculations are presented inAppendix H, Table H.2.

The energy inherent in the carbon and requiredfor carbon anode manufacturing exceeds theadditional theoretical energy need for the inertanode system. A complete account of the energyassociated with the carbon anode productionshows that 6.02tf kWh/kg of aluminum producedis associated with the anode manufacture andfuel value. This value is higher than the 3.04kWh/kg of aluminum that the carbon anode sup-plies in the reaction for theoretical reduction.Because the actual values for manufacturing inertanodes are unknown at this time, a conservativeestimate is made in Appendix H, Table H.3.This estimate, approximately 0.77 kWh/kg ofaluminum, provides a basis for comparing thetwo processes. The total energy associated withan inert anode operating cell is 0.77tf kWh/kgfor anode manufacturing, 3.14 kWh/kg for theon-site additional voltage requirement, and 3.86tf

kWh/kg for the generation and transmissionlosses associated with the additional anode volt-age. The inert anode requires 7.77tf kWh/kg ofaluminum produced, 29tf% more than the tacitenergy associated with the Hall-Heroult carbonanode requirement.

The total energy consumption associated withthe cell operation of a simple inert anode is 18.3(51.7tf) kWh/kg of aluminum. This includes a5.25tf kWh/kg energy savings for anode manu-facturing, as well as added energy consumptionof 3.14 kWh/kg for the additional on-site volt-age requirement, and 3.86tf kWh/kg for the gen-eration and transmission losses associated withthe additional anode voltage requirement. Over-all, the inert anode cell consumes an additional3.14 (9.10tf) kWh/kg of aluminum producedover the traditional Hall-Hèroult carbon anoderequirement, representing an increase in energyconsumption of 22 (8tf)%.

As gas molecules evolve on the anode surface,they coalesce and form bubbles, which creates alayer of bubbles adjacent to the horizontal elec-trode. These bubbles jostle and move along as aresult of their density difference with the bathand by the motion imparted from the aluminumpad. Experimental evidence shows that the oxy-gen gas evolved in the bath has a differentfroth/foam dynamic than carbon dioxide. Themolar volumes of gases that result from thereduction reaction are equal. However, becauseoxygen is 27% lighter than carbon dioxide, lessmass moves through the electrode area. Thephysical properties, in practice, contribute to a

Table 10.12 Energy impact of inert anode technologyInert anode

Inert anode 1) Elimination of Inert anode 1) Elimination carbon anode

Modern prebaked 1) Elimination of of carbon anode 2) Low polarization Hall-Heroult carbon anode 2) Low polarization 3) Wetted cathode

Energy input kWh/kg Al ACD = 4.5 ACD = 4.5 ACD = 4.5 ACD = 2.0


Bauxite-alumina 7.59 7.59 7.59 7.59Anode materials 0.61 0.76 0.76 0.76Total 8.20 8.20 8.20 8.20

Reaction energyReaction electrolysis 3.76 6.90 6.90 6.90Cell ohmic 10.67 10.67 9.25 6.20Total reaction 14.43 17.57 16.15 13.11

Total on-site 22.63 25.76 24.35 21.30Percent energy Reactions �22% �12% 9%savings Reactions and anode �22% �13% 8%

Reactions, anodes, and ore �14% �8% 6%Tacit energy demands

Raw materialsBauxite-alumina 8.21 8.21 8.21 8.21Anode materials 6.01 0.76 0.76 0.76Total 14.22 8.98 8.98 8.98

Reaction energyReaction electrolysis 10.91 20.00 20.00 20.00Cell ohmic 30.91 30.91 26.82 17.98Total reaction 41.83 50.92 46.83 37.99

Total tacit 56.05 59.89 55.80 46.96Percent energy Reactions �22% –12% 9%savings Reactions and anode �8% 1% 19%

Reactions, anodes, and ore �7% 0% 16%



lower anode overvoltage (Ref 10.29). The inertanode can also be shaped to allow for betterrelease of the oxygen generated (Fig. 10.18). Thelower overvoltage of the inert anode, comparedto a carbon anode, reduces the energy consumed.The overvoltage reduction, combined with theelimination of carbon anode manufacturing, isestimated to provide a 5% energy improvement.

A more stable inert anode surface will allowthe ACD to be reduced and will be simpler tocontrol. Carbon anodes must be lowered as theyare consumed to maintain an optimal ACD.Combined with a wetted cathode, the inert anodeprobably offers the greatest potential for bathvoltage reduction. Inert anodes will also elimi-nate cell disruptions that occur with carbonanode changing.

Overall, inert anodes, when combined with awetted cathode and compared to the traditionalHall-Heroult cells, are expected to provide:

• A 10% reduction in operating costs (elimi-nation of carbon anode plant and labor costsassociated with replacing anodes) (Ref 10)

• A 5% increase in cell productivity (Ref 10.30)• A 32% reduction in GHG emissions

(Appendix D, Table D.4)

These practical scenarios provide the inertanode cell with an overall lower energy require-ment than the state-of-the-art Hall-Heroult cell.However, the challenging engineering designsfor the inert anode cell systems must incorporateeffective approaches for minimizing thermallosses from the reduction cells, the current-carrying bus systems, and the connectors externalto the cell.

Multipolar Cells

Present Hall-Heroult industrial cells consist of asingle cathode surface and essentially a singleanode surface immersed horizontally, one overthe other, in a bath. This single-pole arrange-ment makes aluminum reduction a capital-in-tensive process. Use of multipolar electrolyticcells would greatly increase productivity-per-unit-reactor volume by placing multiple elec-trodes in a single reactor. This arrangementwould also provide better control of heat losses.For practical engineering reasons, a multipolarcell requires that the ACD be stable, which inturn requires an inert/stable anode. To date, thelack of suitable materials of construction haveruled out multipolar cells using molten fluorideelectrolytes.

Two variations of electrochemical multipolarcells are possible, as shown conceptually in Fig.10.19. One system uses a cell with an anodicsurface on one side, multiple bipolar electrodesin the center portion, and a cathodic surface onthe opposite side, as in Fig. 10.19(a). One surfaceof a bipolar electrode plate acts anodically,while the opposite surface acts cathodically.Bipolar electrodes must be electrically isolatedand sealed to the cell walls to avoid bypasscurrent efficiency losses. The second systemuses independent pairs of electrodes immersedin the same cell, as shown in Fig. 10.19(b).Multipolar cell designs are complicated by theneed to remove liquid metal and collect thegaseous reaction products.

The bipolar-multipolar concept was success-fully demonstrated for electrolytic aluminumchloride reduction (Ref 10.31). In 1976, Alcoa

Fig. 10.18 Inert anode schematic (right) compared with conventional carbon anode (left)


Inc. began operating a multipolar, prototypeplant with a capacity of over 18,000 kg of alu-minum per day. Alcoa obtained 90% current effi-ciency at 9.26 kWh/kg of aluminum. The Alcoamultipolar cell was electrically 40% more effi-cient than a present-day Hall-Heroult cell. How-ever, the prototype plant was eventually shutdown because of the combination of the costs toproduce anhydrous aluminum chloride feed, thefailure to reach full design capacity, the need toremove and destroy trace amounts of chlorinatedbiphenal by-products, the reactor capital costs,and general maintenance costs (Ref 10.32).

The Alcoa cell had distinct technical advan-tages over the classic Hall-Heroult cell: it workedat substantially lower temperatures, it had rela-tively high current densities, its carbon anodeswere not consumed, it had no fluoride emis-sions, and the plant had a smaller footprint. Theimprovement in electrical energy efficiency wasa result of the higher electrical conductivity ofthe electrolyte and the small interpolar distance.The cell feed consisted of 3 to 10% of purifiedaluminum chloride (AlCl3), obtained by car-bochlorination. The electrolyte consisted mainlyof sodium chloride and potassium chloride, orlithium chloride. Electrolysis was performed in asealed cell consisting of 12 to 30 bipolar carbonelectrodes stacked vertically at an interpolar dis-tance of 1 cm (0.4 in.). The electrodes remainedimmersed in the electrolyte at an operatingtemperature of 700 � 30 °C (1290 � 55 °F). Acurrent density of 0.8 to 2.3 A/cm2 and a single-cell voltage of 2.7 V (dc) were typical operatingconditions. An operating life of nearly threeyears was claimed for the electrodes.

A useful feature of the Alcoa cell was that thebuoyant chlorine created a flow in the narrowgap between the bipolar electrodes. This aidedin sweeping the aluminum from the cathodesand enhancing the formation of aluminum

droplets. The gas also helped maintain a contin-uous flow of electrolyte across the cell. Chlorinewas collected at the top of the cell, and moltenaluminum was collected in the bottom. Thechlorine gas generated by the reduction reactionwas completely recycled to produce new alu-minum chloride feed for the cell.

A Hall-Heroult multipolar design, involvingmultiple inert anodes and wettable cathodes ar-rayed vertically in a fluoride electrolyte cell, hasbeen explored by Northwest Aluminum (Ref10.33). In this concept, the operating tempera-ture was 750 °C (1380 °F), much lower thanHall-Heroult technology. Alumina saturation atthe lower temperature of the electrolyte bathwas maintained by controlling a suspension offine alumina particles in the bath. This technol-ogy offered all the benefits of reduced energyconsumption (low ACD), elimination of carbonanodes and associated emissions, as well as asignificant increase in productivity per cell, as aresult of the multipolar design. This work cameto an unfortunate halt with the closure of North-west Aluminum.

Another multipolar cell design using a modi-fied bath composition has been under investiga-tion at Argonne National Laboratory (Ref 10.34).The modified bath allows for the cell-operatingtemperature to be lowered to 700 °C (1290 °F).This temperature enables the use of commer-cially available metal anodes (e.g., aluminumbronzes) with wetted cathodes in a vertical (non-horizontal) arrangement to produce high-purityaluminum metal product and oxygen gas.

Alternative Primary AluminumProcesses

Carbothermic reduction of alumina and chlo-ride reduction of kaolinite clay have been under

Fig. 10.19 Multipolar cells


study for more than 45 years. These technolo-gies have the potential to be alternative processesto the Hall-Heroult process. The nonelectrolytic,carbothermic reduction of alumina is estimatedto produce aluminum with significantly lessenergy consumption, reduced emissions, andlower investment costs. Multipolar-reductiontechnology to produce aluminum using alu-minum chloride obtained from low-grade domes-tically available clays is also estimated to reduceenergy consumption, capital costs, and emissions.The aluminum industry has yet to adopt theseprocesses for producing aluminum, mainlybecause of the uncertain capital and operatingeconomics associated with large-scale plants.

Carbothermic Technology

Carbothermic reduction of alumina is the onlynonelectrochemical process that has shown po-tential for aluminum production. This technologyhas been an industry objective and the subject ofextensive research for more than 45 years (Ref10.35). Carbothermic technology is projected toproduce aluminum from ore at 36.7tf kWh/kg, or34% below a modern Hall-Heroult carbon anodetechnology (Table 10.13).

Carbothermic reduction produces aluminumusing a chemical reaction that takes placewithin the volume of a reactor. This volumetricreaction requires much less physical space thanthe area reaction of Hall-Heroult. Specifically,with this distinction of volumetric versus areaprocessing, the Hall-Heroult reduction elec-trolytic reaction occurs on the surface of anelectrode in a reactor (cell). Capacity is doubledsimply by doubling the electrode surface area.

Nonelectrolytic reduction chemistry occurswithin a volume of fluid. A spherical volumetricreactor can be doubled by increasing its radius1.4 times. This difference between volume andsurface area reactions provides significant sav-ings in capital cost for a reactor.

This distinction gives carbothermic technol-ogy a smaller footprint ( the size of Hall-Heroult) and makes it much less dependenton the economies of scale (large and long pot-lines) required for an economically efficientHall-Heroult facility. If successful, this tech-nology could transform the structure of the alu-minum industry, allowing industry the freedomto relocate from regions of inexpensive power tocenters of manufacturing. Aluminum produc-tion “mini-mills” could be placed adjacent to orwithin aluminum casting facilities. These mini-mill installations could provide molten aluminumand/or specifically alloyed metal directly to thecasting operations. This would provide additionalenergy, economic, and environmental benefits tothe industry (Ref 10.36, 10.37).

Carbothermic reduction of alumina to alu-minum is a multistep chemical reaction process.The thermodynamic optimization for the reac-tions requires a multizone furnace operating atvery high temperatures (Fig. 10.20). In the first-stage net reaction, alumina and carbon form analumina/aluminum carbide slag at ~1900 °C(3450 °F) (2Al2O3 � 9C Al4C3 � 6CO). Inthe next-stage net reaction, aluminum carbide isreduced by alumina to form aluminum metal at~2000 °C (3630 °F) (Al4C3 � Al2O3 6Al �3CO). The thermodynamics (temperatures andchemical equilibria) of these reactions are verycomplex. A significant portion of aluminumevolves as gas-phase components (aluminumand Al2O) at these operating temperatures.Careful process control is necessary to mini-mize the generation of volatiles. Recovery ofthese components in the form of Al4C3 in avapor recovery system is required for theprocess to be economically viable. If the alu-minum and Al2O back-react with CO to formAl2O3, then the productivity of the process isdecreased, and the Al4C3 required to satisfy theprocess stoichiometry is deficient.

The complex thermodynamic controls, sophis-ticated equipment, and construction materialsrequired to successfully develop an economicalcommercial system have eluded the industry sofar. Current R&D efforts are reevaluating car-bothermic technology in hopes of capitalizingon new, advanced, high-intensity electric-arc

�

�

110

Table 10.13 Comparison of Hall-Heroult andcarbothermic reduction

Modern prebaked Carbothermic Energy input kWh/kg Al Hall-Heroult reduction


Bauxite-alumina 7.59 7.59Kaolinite . . . . . .Anode materials 0.61 . . .Reaction carbon . . . 0Total 8.20 7.59

Reaction energyReaction thermal . . . 7.71Furnace losses . . . 1.36Reaction 3.76 . . .Cell ohmic 10.67 . . .Total reaction 14.43 9.07

Total on-site 22.63 16.66Percent energy Reactions 37%savings Reactions and anode 40%

Reactions, anodes, and ore 26%


furnace technology, advanced thermodynamicand system modeling techniques, and animproved understanding of the process dynamics(Ref 10.38).

At a minimum, the carbothermic process willhave at least twice the inorganic metallic impuri-ties (and possibly more, because petroleum cokeis a waste product of overall refining operations),which will end up in the metal. In addition, anyerosion or corrosion of furnace liners or planthardware, including wear of an auger that may beused to force aluminum carbide from the vaporrecovery stage back into the stage 1 and 2 reac-tors, will likewise end up contaminating themetal. Probably, it will have to be experimentallydetermined if the dilution of the metal with scrapor virgin metal additions during the decarbona-tion stage is sufficient to meet the required metalpurity standards without an additional impurityremoval or filtration step, which would adverselyimpact the process economics.

Theoretical Energy for CarbothermicReduction of Alumina. The theoretical mini-mum energy requirement for producing alu-minum by the carbothermic reduction of aluminais 7.32 kWh/kg.

Carbon is a reactant in the carbothermicreduction reaction process and supplies part ofthe energy necessary to drive the reaction for-ward. This gives the carbothermic reductionprocess a lower theoretical energy requirementthan the direct electrolytic reduction of aluminato aluminum.

The net reaction for the carbothermic reduc-tion is Al2O3 � 3C 2Al � 3CO. It is assumed

that the reactants (alumina and carbon) enter thereactor system boundaries at 25 °C (75 °F), thecarbon monoxide byproduct leaves at 25 °C(75 °F), and the aluminum product leaves thereactor system as molten metal at 960 °C(1760 °F) (Fig. 10.21). The assumed moltenmetal temperature is lower than the actual reactormetal discharge temperature. Theoretically andpractically, the energy in the higher-temperaturereactor discharge can be recovered efficiently(e.g., by mixing high-temperature reactor alu-minum with solid metal scrap to recover the heatenergy and lower the temperature). To comparethe theoretical limits of the various aluminumproduction processes, the same molten producttemperature of 960 °C (1760 °F) is used through-out this chapter. The theoretical reaction occursunder perfect conditions when there are noreverse reactions, parasitic reactions, or heat/energy losses external to the system.

The calculation of theoretical minimum energyrequirement for the carbothermic reaction isdetailed in Table F.3 of Appendix F. The resultsshow that the energy required to drive the reac-tion forward (ΔG) is 6.03 kWh/kg, and thisenergy is supplied as thermal energy versus theelectrical energy used in a Hall-Heroult cell.The thermal energy (ΔH � ΔG) required tomaintain equilibrium is 0.90 kWh/kg, and thethermal energy (Cp) associated with the moltenaluminum is 0.39 kWh/kg of aluminum. Thetheoretical minimum energy requirement underthese conditions adds up to 7.32 kWh/kg of alu-minum. (Note: If the CO gas emission at 960 °C,or 1760 °F, is included, the total theoretical�

Fig. 10.20 Carbothermic reactor


minimum energy requirement is 7.51 kWh/kgof aluminum.)

The carbon monoxide by-product has a fuelvalue and would likely be captured and used tosupply thermal energy to the carbothermicfacility.

Comparative Benefits for CarbothermicReactors and Hall-Heroult Cells. Electric fur-nace technology provides 90% thermal efficiencyin many applications. Heat losses are limited toconduction and radiation losses from the furnaceshell. Flue gas heat losses are eliminated, as areany undesired reactions between flue gases andmolten aluminum metal. It is reasonable toassume that the carbothermic furnace/reactor canbe designed with more than 85% thermal effi-ciency. If the thermodynamics of the reaction andoffgas recovery can be controlled within 95% ofthe theoretical requirements, the electrical energyrequired for an operating carbothermic reactorwould be (7.32 kWh/kg � [0.85 � 0.95]) or 9.07kWh/kg of aluminum produced. This representsa 37% reduction in energy use when compared tothe 14.4 kWh/kg of aluminum from a modernHall-Heroult cell. Table 10.13 shows the on-siteand tacit energy savings potential of carbother-mic technology.

In addition to electrical energy savings, car-bothermic technology is expected to provideother benefits, such as:

• A reduction in capital costs by 50% or moreas a result of volumetric processing throughhigh-intensity smelting

• A reduction in production costs by 25%through lower electrical demand and elimi-nation of carbon anode manufacturing andhandling

• The production of a high-purity CO/CO2stream for coproduct sale or energy integration

• The potential to use small blocks of electri-cal power due to a high turndown ratio

• The potential of widely locating mini-millswith integral, captive smelters deliveringmolten metal

Environmental Impacts of CarbothermicTechnology. The carbothermic process, whencompared to Hall-Heroult technology, results insignificantly reduced electrical consumptionand the elimination of perfluorocarbon emis-sions that result from carbon anode effects,hazardous spent potliners, and hydrocarbonemissions associated with the baking of con-sumable carbon anodes.

The total carbon dioxide emissions fromcarbothermic reduction, as with Hall-Heroult,depend on the source of electricity. Hydroelec-tric power generation emits almost no carbondioxide, whereas the carbon dioxide emissionsassociated with the average U.S. grid electricityare 0.49 kg CO2e/kWh (Appendix D, TableD.1). Because 39% of the U.S. primary industryoperates on hydroelectric power, the aluminumindustry’s average electrical generation emissionrate is 0.35 kg CO2e/kWh. The carbothermicreaction results in the generation of carbon-baseGHGs, mainly carbon monoxide (CO), at twicethe rate of the Hall-Heroult reaction. However,the carbothermic process only requires electricityfor heating and not for the reduction reaction.Assuming carbothermic technology is used inmini-mills operating off the average U.S. electricgrid, the total GHG emissions from “utility-to-metal” for the carbothermic process are reducedrelative to the average U.S. Hall-Heroult sys-tem, because of carbothermic lower electricalintensity. Table 10.14 (Appendix D, Table D.4)shows the lower electrical demand of carbother-mic technology resulting in lower total carbon

Fig. 10.21 Carbothermic theoretical minimum energy


dioxide equivalent emissions (CO2e) than for aHall-Heroult system. Other studies (Ref 10.39)have also shown a reduction in CO2e.

Kaolinite Reduction Technology

Production of pure aluminum by reduction ofaluminum chloride was discovered before theHall-Heroult process, in 1825. Alumina conver-sion to aluminum chloride and reduction to alu-minum using bipolar technology was demon-strated in the late 1970s, but it was notcommercialized because of problems with theproduct purity and projected high capital andoperating costs. New construction materials,improved thermodynamic understanding, andthe potential to use low-cost alumina containingclays have maintained the interest in chloridereduction technology for producing aluminum.

Compared to the current Bayer refining andHall-Heroult fluoride-base smelting of alumina,the chlorination of alumina-containing clayspromises many potential advantages:

• Raw materials are widely available, inexpen-sive, and indigenous to the United States.

• Thermodynamics provide high-speed, high-conversion reactions with lower electricaldemand.

• No bauxite residue is produced, althoughthere is waste from the clay chlorination.

• Conventional materials-of-construction (i.e.,mild steel) can be used.

Industry has spent more than 35 years develop-ing process technologies for the chlorinationof widely available, low-grade kaolin clays.These clays contain kaolinite (hydrated aluminasilicate, i.e., Al2O3

.2SiO2.2H2O), significant

amounts of titanium dioxide (1 to 5%), and othermaterials. Titanium tetrachloride and other metalchloride by-products are also produced whenprocessing kaolin clays.

The basic steps of the clay-to-aluminumprocess are shown in Fig. 10.22 (Ref 10.40).

First, the kaolin clay and process coke aredried. All feed materials to the clay chlorina-tion units must be dried to minimize chlorineabsorption in any water and to reduce the cor-rosive effects of moisture in the chlorinationoffgas stream. The drying process involvescontrolled, catalyzed heating of the finelyground clay at ~800 °C (1470 °F) in a fluidizedreactor, using coke and air to provide theheating energy. The hot reactor produces adehydrated calcined clay, which is sent tocarbochlorination.

The carbochlorination step is a high-speed,catalyzed exothermic reaction of calcined claywith chlorine and coke:

3(Al2O3.2SiO2) � 14C � 21Cl2 6AlCl3 � 6SiCl4

� 7CO � 7CO2

Chlorine is injected into the bottom of thefluidized reactor with the clay oxides and coke.The metallic oxides, principally those ofaluminum and silicon, are converted to their

�

Table 10.14 Carbon dioxide equivalent compar-ison of Hall-Heroult and carbothermic reduction

Modern Hall-Heroult, Carbothermic,Emission sources kg CO2e/kg Al kg CO2e/kg Al

Carbon anode 1.68 2.45Reaction energy requirements:

kWh/kg Al 14.43 9.07kg Co2e/kWh 0.492 0.492

7.10 4.46Process 2.20 0.0Total 10.98 6.91

Fig. 10.22 Clay-to-aluminum process schematic


chlorides. The chlorides exist as vapors at thereaction temperatures. The next process step isdesigned to suppress the formation of silicontetrachloride. The offgases of the chlorinationreactor are treated with a small amount of cata-lyst vapor and reacted in a second fluid bedreactor with additional calcined clay. The alu-mina portion of the clay is converted to alu-minum chloride, and silicon tetrachloride isconverted to solid silica and discharged (3SiCl4� 2Al2O3 4AlCl3 � 3SiO2). The net effectof this step is the preferential chlorination ofalumina relative to silica. The hot vapors fromthe second reactor are cooled, and crudealuminum chloride is recovered. Several by-products are produced as an extension of thechlorination step. Titanium tetrachloride, silicontetrachloride, and boron trichloride are not con-densed in the cooler and can be recovered insubsequent processing. Silicon tetrachloride canbe reacted with oxygen to recover chlorine,which is recycled back to the chlorination step,as follows:

4SiCl4 � 5O2 2SiO2 � 3O2 � 8Cl2

The crude aluminum chloride must be puri-fied for ease of operation of the electrolyticreduction cells and for the final aluminumquality. The main impurity is iron, which ispresent as ferric chloride at levels as high as~50,000 ppm. Impurities are removed bychemical treatment.

The electrolysis of aluminum chloride to alu-minum takes place in an aluminum chloridesmelting cell, which comprises a stack of hori-zontal bipolar graphite electrodes, betweenwhich the aluminum chloride is converted intohigh-grade aluminum and chlorine gas (2AlCl3

2Al � 3Cl2). The electrodes are immersedin a chloride bath, which is contained in fullyenclosed, thermally lined vessels. The bipolarelectrodes are supported and separated by inertspacers resting on the electrode below it, andthe entire stack, consisting of bottom cathode,bipolar electrodes, and top anode, is supportedby the walls of the cell.

Theoretical Energy for Kaolinite Reductionof Alumina. The theoretical minimum energyrequirement for producing aluminum fromkaolinite is 5.76 kWh/kg of aluminum produced.The theoretical minimum for the chloride reduc-tion step of aluminum chloride is 7.66 kWh/kgof aluminum.

The theoretical minimum energy requirementcan be calculated from the net chemical reactionin the kaolinite-to-aluminum process:

7(Al2O3.2SiO2) � 14C 14Al � 14SiO2 � 7CO

� 7CO2

It is assumed that the reactants (kaolinite andcarbon) enter the system boundaries at 25 °C(75 °F), the carbon monoxide and carbon diox-ide by-products leave at 25 °C (75 °F), and thealuminum product leaves the system as moltenmetal at 960 °C (1760 °F). The theoretical netreaction occurs under perfect conditions whenthere are no reverse reactions, parasitic reactions,or heat/energy losses external to the system.These assumptions yield a theoretical minimumenergy requirement of aluminum productionfrom kaolinite as 5.76 kWh/kg of aluminum(Appendix F, Table F.5). This assumes a purekaolinite feedstock.

The kaolinite process is accomplished in twosteps: carbochlorination and aluminum chloridereduction. The electrolytic aluminum chloridereduction process used in the kaolinite processrequires the exact same minimum amperage(2980 Ah/kg of aluminum) as any electrolyticprocess for reducing aluminum. The theoreticalchloride reduction reaction occurs under perfectconditions (where there are no reverse reactions,parasitic reactions, or heat/energy losses externalto the system) and requires 7.66 kWh/kg of alu-minum produced. The theoretical minimumenergy requirement used in this chapter is calcu-lated at 960 °C (1760 °F) to provide comparisonwith the other processes. The proposed multipo-lar system operates at approximately 700 °C(1290 °F), which would lower the theoreticalenergy demand by approximately 0.09 kWh/kgof aluminum. The exothermic nature of thecarbochlorination reaction (�1.90 kWh/kg ofaluminum) results in the overall kaolinite-to-aluminum theoretical energy requirement beinglesser than the minimum energy requirement forthe aluminum chloride reduction.

Comparative Benefits for Kaolinite Reduc-tion and Hall-Heroult Cells. The kaolinitereduction process offers potential advantageswhen compared to the Hall-Heroult process.Table 10.15 shows the on-site and tacit energycomparison for the Hall-Heroult and kaoliniteprocesses.

The combined clay carbochlorination andbipolar aluminum chloride smelting process is

�

�

�

�


estimated to be 16tf% more tacit energy efficientthan the Hall-Heroult, has a smaller plant foot-print, can be more flexible in regard to use of off-peak power and power fluctuations, and producesfewer emissions and process wastes. On-siteenergy use is approximately 42% lower than amodern Hall-Heroult cell. The large tacit energyimprovement results from a decrease in electricaluse. Compared to the current Hall-Heroult smelt-ing technology, the bipolar aluminum chloridesmelting process consumes less electricity, yield-ing 60% more metal for the same electrical input.Bipolar cell designs provide significantly lowerreactor volume per unit of product output. Thislower volume allows the cell to idle and holdtemperature much more efficiently than a singleelectrode cell. These properties allow multipolarcells to take better advantage of off-peak electri-cal costs. Chloride cells also operate at lowertemperatures, providing additional savings.

Raw materials used in the kaolinite processrequire approximately 47% more energy com-

pared to those in a modern Hall-Heroult process(Table 10.15). Kaolinite requires more energyfor extraction on an aluminum basis, and thecarbochlorination step requires 2.5 times thecarbon of a Hall-Heroult carbon anode system.The raw material energy expenses may be com-pensated for by the lower-cost domestic supplyof kaolinite and the ability to use lower-cost car-bon than Hall-Heroult systems.

It is important to emphasize that while signif-icant elements of the kaolinite process havebeen studied and developed, no integrated pro-duction of aluminum from kaolinite clays hasyet been attempted. Significant work in devel-oping an integrated production process has beenreported.

Environmental Impacts of Kaolinite Tech-nology. Table D.4 in Appendix D tabulatesCO2e for the kaolinite-to-aluminum process.Operating off the average U.S. electrical grid,this process would produce 29% fewer CO2ethan a typical Hall-Heroult facility.

Table 10.15 Comparison of Hall-Heroult and kaolinite reductionModern prebaked Kaolinite (AlCl3)

Energy input kWh/kg Al Hall-Heroult reduction


Bauxite-alumina 7.59 ...Kaolinite ... 8.14Anode materials 0.61 0.76Reaction carbon ... 0Total 8.20 8.91

Reaction energyReaction thermal ... –1.90Furnace losses ... 0.40Reaction 3.76 6.48Cell ohmic 10.67 2.93Total reaction 14.43 7.91

Total onsite 22.63 16.82Percent energy Reactions 45%savings: Reactions and anode 42%

Reactions, anodes, and ore 26%Tacit energy demands

Raw materialsBauxite-alumina 8.21 ...Kaolinite ... 8.81Anode materials 6.01 0.76Reaction carbon ... 11.34Total 14.22 20.92

Reaction energyReaction thermal ... –1.90Furnace losses ... 0.60Reaction 10.91 18.79Cell ohmic 30.91 8.48Total 41.83 25.98

Total tacit 56.05 46.90Percent energysavings:

Reactions 38%Reactions and anode 44%Reactions, anodes, and ore 16%


Secondary Aluminum (Recycling)

The production of primary aluminum ingotsfrom bauxite ore requires approximately 23.8(59.1tf) kWh/kg of aluminum. Recoveringaluminum from scrap to produce secondary alu-minum ingot consumes approximately 5% ofthe energy required to produce primaryaluminum (Ref 10.11). This significant energydifference drives the emphasis placed onaluminum recycling in today’s (2007) societyand in the aluminum industry.

Recycling in the United States saved morethan 172 � 109 kWh (0.59 quad) of energy in2005, the equivalent of 19,700 Mw. Each kilo-gram of aluminum that is recovered by recyclingsaves 57.3tf kWh of the 60.5tf kWh of energyconsumed in producing a primary aluminumingot from bauxite ore (Appendix E, Table E.6).Any process that improves the recovery of scrapaluminum is effectively making an order ofmagnitude change in the energy associated withaluminum production.

The growth of aluminum recycling repre-sents the greatest change in the structure of theindustry and in the energy associated with alu-minum manufacturing. A common practice sincethe early 1900s, recycling was a low-profileactivity until 1968, when aluminum beveragecan recycling vaulted the industry into publicconsciousness. In 1960, recycled aluminumaccounted for 18% of the nation’s total alu-minum supply (401,000 metric tons). Over thenext 45 years, production of recycled aluminumrose by almost 746% to 2,990,000 metric tons.During those same 45 years, the total U.S.aluminum metal supply increased 300%(Table 10.3). In 2005, 75 plants in the UnitedStates produced secondary ingots. Over half—54.7%—of the aluminum metal produced in theUnited States in 2005 was from recycled mate-rial (Ref 10.9).

The growth of the market for recycled alu-minum is due in large measure to economics. Itis cheaper, faster, and more energy-efficient torecycle aluminum than to manufacture it fromore. Recovered aluminum is easily melted atrelatively low temperatures (aluminum alloystypically melt at temperatures below 660 °C, or1220 °F). Producing a recycled aluminum ingotconsumes only approximately 5% of the energyrequired to produce a primary aluminum ingotfrom bauxite ore. In addition, to achieve a givenoutput of ingot, recycled aluminum requires onlyapproximately 10% of the capital equipment

costs compared with those required for the pro-duction of primary aluminum.

Aluminum products are corrosion resistant,which allows them to be easily and repeatedlyrecycled into new products. The corrosionresistance is due to the metal properties. Whenthe surface of aluminum is exposed to air, it rap-idly forms a tenacious, self-limiting, protectiveoxide layer. Other surface treatments can beapplied to further enhance the corrosion resist-ance of aluminum.

Figure 10.23 shows the annual growth ratesof the three sources of metal supply between the1995 and 2005 period, together with the growthin the major aluminum product markets. Alu-minum imports are growing at a rate of 6.9%per year, while both U.S. primary and second-ary productions are in decline (Table 10.3).

Aluminum scrap is categorized as new or old.New scrap is generated when aluminum prod-ucts are manufactured. It includes defectiveproducts; scalping chips; edge and end trimfrom rolling processes; skeleton scrap fromstamping and blanking operations; flash, gates,and risers; extrusion butts and ends; and turn-ings and borings. Old scrap (post-consumer orobsolete products) comes from discarded, used,worn-out, or out-of-date products that includeautomotive parts, white-good parts, containerssuch as used beverage cans and closures, wires,cables, and building materials. Runaround scrapis new scrap that is recycled by the same com-pany that generated it. Because runaround isusually not sold or marketed, it is not reportedin the U.S. recycling statistics.

Secondary Aluminum Production. Sec-ondary aluminum producers represent a sepa-rate and vital segment of the aluminum industrywhose principal activities are converting pur-chased scrap and metal recovered from skimand dross generated in molten metal operationsinto usable aluminum alloy products. Recyclingin primary aluminum operations is typicallyconfined to in-house or runaround scrap andmanufacturing scrap returned directly fromtheir customers. Secondary aluminum produc-ers specialize in melting and processing a widerange of new and old, segregated and mixed,high- and low-quality scrap.

The most desirable form of recycling isclosed-loop, in which scrap from specific prod-uct applications are returned for remanufactur-ing of the same products. Rigid container stockused in beverage cans is an example of closed-loop recycling, because used beverage cans are


processed exclusively to remanufacture cansheet. Secondary aluminum producers areinvolved in closed-loop recycling as contractorsto primary producers for this and other products.

Scrap segregated by alloy has greater valuethan mixed scrap because it can be used pre-dictably and most efficiently in the productionof the highest value-added compatible composi-tions. Mixed scrap presents the greatest chal-lenge, generally requiring added steps in melting,composition identification, and often casting intoingot form before consumption.

Casting alloys are the largest and most impor-tant but not the only market for secondary alu-minum. Scrap is used, with primary metal, in theproduction of extrusion billet and fabricatingingot. Reclaimed smelter ingot (RSI) is producedfrom scrap and dross, often on a toll conversionbasis.

Alloys used by aluminum foundries in theproduction of shape or engineered castingsinclude compositions designed to facilitate pro-duction from scrap. These alloys, which havebeen historically prominent, typically specifybroader element ranges and higher impuritylimits than the alloys developed for more spe-cialized purposes and whose compositionsrequire production from primary metal sources.The increased use of aluminum in structuralapplications in the ground transportation sectortypically requires primary compositions, but the

expanded use of aluminum in engine blocks,cylinder heads, and other power train partsrelies on castings in secondary alloys. Theemphasis on light-weighting for improved fuelefficiency and reduced environmental degrada-tion is encouraging the adoption of variouswrought aluminum, products, such as auto bodysheet, fuel tanks, seat backs, extruded drive shaftsand stringers, and forged connecting rods. Asthe proportion of aluminum increases in groundtransportation designs, segregating and blend-ing scrap into saleable alloy compositions willbecome common practice.

New scrap-sorting technology is developing,including those related to the recovery of alu-minum from the large transportation andwhite-goods markets (Ref 10.41). New tech-nologies with computer screening are usingcolor segregation and laser-induced breakdownspectroscopy to sort wrought from cast and, insome instances, one alloy from another (seeChapter 8). Aluminum alloy separations willallow scrap to be segregated into more specificalloy groupings with higher economic value.

Production, Capacity and Growth. Alu-minum scrap is widely recycled and supports alarge secondary aluminum industry. In 2005,the United States produced 2,990,000 metrictons of secondary aluminum, amounting tonearly one-third of its total supply of aluminum.This market had an annual growth rate of

Fig. 10.23 U.S. aluminum markets and growth over the period 1995–2005


–1.7% over the last ten years (Fig. 10.24). Thesupply of scrap metal has become more precious,and U.S. secondary producers must compete ina global market to obtain scrap.

Market demand for recycled aluminum willremain strong due to its inherent low energy costrelative to primary metal. Aluminum use in con-sumer products has become widespread in atrend described as the “urban mine.” Challengesin scrap recovery, alloy sorting, and impurityremoval can be addressed with the currenttechnologies. The limitation to secondary metalmarket growth is the economic supply of scrap.

The urban mine has been accumulating scrapproducts for the past 100 years. There is a lagtime from when aluminum leaves the shippingdock to when it becomes available for recycling.This lag time, or use-phase life, varies signifi-cantly among aluminum products. For packagingproducts such as beverage cans, this lag is onlyapproximately 65 days. Long-life products, suchas building frames, can have a lag time of 40 to50 years. Automobile lag time is 12 to 15 years.Aluminum in automobiles accounted for 5 to10% of scrapped automobile weight in 2000 butrepresents 35 to 50% of its scrap value. Thissource will grow rapidly as new automobiledesigns continue to use a greater proportion ofaluminum. Studies of the future supply of scrapin the United States show the potential for signif-icant shortfalls (Ref 10.42).

Considering the large amounts of aluminumthat are stored in long-life products (accumu-lated within the urban mine) and the continuedgrowth in aluminum demand, recycling willcontinue to increase and be a significant contrib-utor to the U.S. metal supply. In fact, recyclinghas overtaken primary production as the main

source of domestically produced aluminum inthe United States.

Recycling Processes. The objectives of therecycling process are maximum metal recovery,minimum contamination, and lowest conversioncost. Safety is of prime importance because com-ponents mixed with scrap can present the risk ofexplosive reactions and other concerns. Becausemoisture is a safety concern, remelt ingot andRSI are routinely preheated before charging tothe furnace hearth. Preheating methods varyfrom heating on the charging doorsill to usingdedicated preheating ovens. Preheating standardsfrom times of exposure and temperature alsovary. In either case, energy is consumed for thepromotion of operational safety.

Scrap for recycling is available in many forms.Light scrap is typically baled or briquetted toreduce transportation costs. Bales and briquettesare typically split for inspection of integrity andcontamination and/or are crushed, shredded, or“shripped” (sheared and ripped) to controlledflowable particle sizes. Conveyor systems segre-gate particle fines for separate processing, pro-vide magnetic separation, and allow for visual orautomated inspection.

Large volumes of aluminum scrap containpaint, enamel, lacquer, or porcelain coatings.These coatings, which contain oxidizing com-pounds, would significantly reduce metal recov-ery if not removed before melting. Conveyorfurnaces or rotary kilns operating at tempera-tures near the melting point are required fortheir removal. Turnings, borings, and other finescrap may contain oil from cutting fluids thatmust be removed for satisfactory metal recoveryby swarf (metal fines and chips) drying, inwhich the oil contributes to the energy required,

Fig. 10.24 U.S. production of secondary aluminum from 1960 to 2005


or by elevated-temperature treatment in rotatingkilns. Fine scrap is conveyed to external charg-ing wells for submergence by mechanical meth-ods, including the use of recirculating moltenmetal pumps. Light scrap is charged directly tothe furnace hearth and is covered by additionalheavier charge components.

Melting is typically accomplished in gas-firedfurnaces ranging in size from 75,000 to morethan 250,000 lb holding capacity. Corelessinduction furnaces are in common use for rap-idly melting fine scrap. Molten metal from thesefurnaces is transported to the main furnaces forfurther processing. After iterative alloying steps,molten metal is held and processed before cast-ing. Furnaces have thermal efficiencies rangingfrom approximately 20 to 45%. Rotary kilnsand conveyor furnaces operate in essentially thesame efficiency range. Induction melting ismore efficient, at approximately 90% but fur-nace capacities are limited, and additional stepsare required to address oxide concentrationscreated by electromagnetic stirring. In-furnaceand in-line molten metal treatments areemployed to remove dissolved hydrogen andentrained oxides and other nonmetallics beforecasting. In-line systems are usually internallyheated using gas-air or electric resistanceimmersion elements. Troughing and filter basinsare preheated using air-gas torches or resistanceelements. Open molds for casting remelt ingotand sow are routinely preheated before use.

The aluminum, recoverable from skim anddross obtained from foundries and primary andsecondary molten metal processing units, is typi-cally extracted in secondary producers. Whileother dross treatment processes have been devel-oped, rotary furnaces in which dross and salts aremixed and heated remain the most commonlyused. The products of dross recovery treatmentby this process are aluminum and black dross, amixture of unrecovered aluminum, metallicoxides, and salt. Tertiary processes have beendeveloped for separating the components ofblack dross into saleable salt fluxes, metal, andvalue-added nonmetallic derivatives, such as cal-cium aluminate and additives for low-densitycement. The economics of tertiary processing arestrongly and adversely influenced by the currentlow-cost alternative of landfill disposal.

In summary, the energy-intense operationscommon in recycling are:

• Swarf drying for removal of combustiblecontaminants

• Kiln or conveyor delacquering of coatings • Preheating charge components • Melting• Holding and melt processing• Rotary furnace operation

Theoretical Energy for Secondary Alu-minum. The theoretical minimum energyrequired to produce secondary aluminum at960 °C (1760 °F) is 0.39 kWh/kg of aluminum.On a theoretical and a practical basis, the energyrequired to produce secondary aluminum is lessthan 6.5% of the energy required to produceprimary metal.

If the system boundaries are drawn around asecondary aluminum facility, the material enter-ing and leaving is aluminum metal. Because nochemical change has occurred, the theoreticalminimum energy requirement for scrap conver-sion is only the energy required to melt andraise the metal temperature to that required forcasting. Molten metal furnace temperaturesvary depending on the furnace, alloys, and otherproprietary factors. A molten metal temperatureof 960 °C (1760 °F), the same production tem-perature as for primary aluminum, is used inthis chapter for secondary aluminum to easilycompare aluminum process technologies. Inreality, normal pouring temperatures are muchlower, in the range of 650 to 750 °C (1200 to1380 °F).

The theoretical minimum energy requirementto bring room-temperature (25 °C, or 75 °F)aluminum to its molten form at 960 °C(1760 °F) is 0.39 kWh/kg, which is less than6.5% of the theoretical energy requirement forthe primary production of aluminum. The theo-retical energy required to heat aluminum fromroom temperature to its melting point, melt it,and raise the molten aluminum to a highertemperature is calculated and explained inAppendix G. Pure aluminum melts at 660 °C(1220 °F) and requires 0.30 kWh/kg to melt,23% lower than the value used in this chapter.

The energy efficiency of the entire aluminumindustry can be further increased by capturing agreater percentage of material for recycling andby improving technology for scrap handling andmelting. Nontechnological and nonmarket fac-tors are also important for the continued growthof recycling. Two of these factors, consumerawareness and incentives, can contribute sig-nificantly to the recycling volume. Consumerawareness requires continuing public educationabout the energy and environmental benefits of


recycling. Offering incentives will aid in thereturn of aluminum scrap to the manufacturingbase.

Recycling energy efficiency will be enhancedby developing technologies that minimize oxi-dation and improve thermal inefficiencies inscrap processing and melting. Improved collec-tion systems and separation devices (e.g., eddycurrent, color sorting, laser sorting) can increasealuminum scrap recovery by 20 to 30%.Improved technology can be applied to increasescrap recovery rates, especially with regard toaluminum in municipal solid waste. Incrementalimprovements in existing furnaces can furtherreduce the recycling energy requirements. Theycan be achieved by recuperating stack gasenergy for preheating combustion air and metalfeedstock, by modifying burner and furnacedesigns, and by controlling furnace practice andoperating conditions (Ref 10.43, 10.44). Newtechnologies must also be developed to ensuremore significant progress.

Aluminum Processing

The aluminum industry can be divided intometal- and product-producing sectors. In 2005,the metal-producing sector manufacturedapproximately 2,480,000 metric tons of primarymetal and 2,990,000 metric tons of secondarymetal. The product-producing sector processesthese domestically produced and imported met-als into approximately 5,497,000 metric tons ofrolled products, 1,719,000 metric tons of extru-sions, and 2,513,000 metric tons of shape cast-ings (Appendix E, Table E.4) (Ref 10.43). Wirerod and bar, forgings, impacts, and powder prod-ucts comprise less than 8% of the aluminumproduct market and have not been studied as apart of this chapter.

Recoveries (yields) in each product-processingsector are less than 100%. Handled, melted,and worked tonnage exceeds that implied bythe product statistics. The manufacturing of cir-cles and blanks from aluminum sheet, forexample, may recover as little as 35% of theoriginal ingot weight as shipped product. Thegross-to-net weight ratio in gravity castings isoften 2:1; in other words, 50% of the originalcast weight is automatically designated forremelting. Furthermore, many products have agenealogy of repeated melting and castingsteps. Ingot cast at a primary smelter isremelted at a secondary producer for casting

remelt ingot, which is again remelted at analuminum foundry for casting production. It isestimated that more than 11,000,000 metrictons of aluminum alloys are melted each yearto support aluminum industry shipments. Somecasting operations are located sufficiently closeto primary smelters and secondary aluminumproducers to allow the molten metal to beshipped in insulated and protected crucibles,saving the energy of remelting.

Minimizing planned and unplanned scraprepresents a large opportunity for energy- andcost-savings. Each kilogram of metal that doesnot go into a final product must be remelted,recast, and reworked. Remelting also leads tothe loss of a percentage of metal to oxidation,which must be replaced with energy-intensiveprimary metal. Melting and melt-processingoperations are the most energy-intense of allpostsmelting processes.

Melting, Alloying, and Melt Treatment

In an aluminum primary smelting facility,molten metal is transferred from the smeltingcells to furnaces for alloying and melt treatmentprior to casting. In secondary and other castingplants, ingot, metallurgical metals, and masteralloys must be melted and alloyed. The meltingarrangement in most larger plants provideshigh-heat-input, high-melt-rate furnaces formelting, and separate holding furnaces to whichmolten metal is transferred for final alloyingand preparation for casting. Some operationshave combination melting/holding furnaces.

Furnaces and Melt Treatment. There are awide variety of furnace types and designs formelting aluminum. Furnace choice depends onthe required melt volume, melt rate, availability,cost of fuel and electrical energy, and emissionstandards. The most common in primary andsecondary operations are natural gas-firedreverberatory furnaces reaching capacities ofmore than 120,000 kg. Crucible furnaces withcapacities ranging from 160 to 4500 kg aremore common in small- and medium-sizedfoundries. Other furnace types include corelessand channel induction, electrical resistance,and radiant tube furnaces.

Reverberatory furnaces are box-shaped andconsist of an insulated steel shell with a refrac-tory lining. Fuel-fired reverberatory furnaces areused when the melt rate and/or capacity arelarge. The fuel-fired reverberatory furnace firesnatural gas, propane, or oil directly into the


furnace from either the roof or, more typically,the sidewall. The heat is transferred to the sur-face of the molten aluminum predominantly byrefractory radiation and some convection. Thereare a large number of reverberatory furnacedesign variations: charging and access doors,refractory specifications, sidewells for chargingand/or recirculation, hearth or sidewall inductionstirring, split hearths, dry hearths, divided zonesfor melting and holding, and various burnercapacities and types. Recuperation conceptsinclude charge preheating, preheating combus-tion air, and cogeneration.

The growth in recycling has resulted in anumber of specialized processes, furnaces, andsystems to improve metal recovery from scrap.Molten metal pumps have been incorporatedinto these designs to provide rapid ingestion offine scrap and more rapid melting of largerscrap forms. Salt flux additions maintain systemcleanliness and aid in the separation of oxides.Pump-induced flow external to the furnace mayinclude provisions for melt treatment and theseparation of oxides as well as for melting.

Natural gas or oil-fired reverberatory furnacesuse approximately 0.87 to 1.96 kWh/kg of alu-minum (Ref 10.44). In addition, a gas furnaceincreases metal losses due to oxidation. Gas fur-naces have 5 to 8% metal loss compared to 0.5to 3% loss in electric furnaces. Recent designinnovations in fossil-fuel reverberatory furnaceshelp capture the waste heat in the stack gas topreheat incoming materials. This increasesenergy efficiency and reduces the time requiredto melt the metal. Recuperated waste heat canalso be used to preheat combustion air. Thesetechnologies can reduce fuel usage to less than0.57 kWh/kg of aluminum. Large commercialgas-fired furnaces are reported with efficienciesas high as 56% (0.30 kWh/kg) (Ref 10.45).

Crucible furnaces are more versatile withregard to alloy changes and melt quantities.Combustion occurs between an insulated steelshell and a crucible of silicon carbide, graphite,clay graphite, or other refractory materialresistant to molten aluminum attack. Electricalresistance elements can be substituted for gasburners in crucible furnace designs.

In recent years, much research has been doneon using electric immersion heaters as a way toremelt aluminum (Ref 10.46). Immersion heatershave very low rates of heat loss (~97% thermalefficiency) and have energy usage levels of lessthan 0.50 kWh/kg of aluminum. This high ther-mal efficiency is obtained on site but is offset by

the lower efficiency (30 to 40%) for electricitygeneration and transmission.

Oxide naturally and rapidly forms on the sur-face of molten aluminum, resulting initially in athin protective film. With increase in time andtemperature, the thickness of the oxide layerincreases. Turbulence and agitation accelerateoxide formation and result in the intermixing ofmetal and oxides. Oxidation rates are influencedby alloy content and increase with temperature,especially when magnesium is present in thealloy. The oxide layer also effectively insulatesthe bath from radiation heat transfer and mustbe periodically removed to maintain thermalefficiency in reverberatory furnaces.

If fluxes are employed to treat the skim beforeit is removed from the furnace, the oxides aretypically dewet, and a large portion of themolten aluminum entrained in the skim layerseparates to the melt. In either case, untreated(skim) or flux-treated (dross) contains entrainedfree metal as a result of the skimming action.Efforts are usually made to recover entrainedfree aluminum after skimming. While still hot,metal can be drained from the skim gravimetri-cally and with vibration. Alternatively, skimmay be rapidly cooled by inert-gas quenchingor in rotating water-cooled steel drums, afterwhich free aluminum may be physically sepa-rated. The residue comprising unrecovered alu-minum and oxides is normally further processedfor its metal content.

Gross-melt losses typically range from 1 to ashigh as 8% in small, poorly designed furnaces.The magnitude of the loss is dependent on thetype of furnace and burners, the surface-to-volume ratio, the practices that are used, and thematerial being melted (Ref 10.9). Melt loss hassignificant economic impact, because oxidizedmetal must be replaced in the supply chain withnew primary aluminum metal.

Specific elements or combinations of elementsare added to molten aluminum to produce alu-minum alloys. Alloying provides the basis for aremarkable range of physical and mechanicalproperty capabilities not displayed by unalloyedaluminum or by any other metal system. Amongthe elements and combinations of elementsadded to molten aluminum are modifiers andrefiners, which provide finer grain structures andcontrolled microstructural features, includingmetallurgical phases that influence recrystalliza-tion behavior.

Molten aluminum contains dissolved hydro-gen, entrained oxides, and other nonmetallic


inclusions that, if not removed, would adverselyaffect metal acceptability and performance.Treatment with salt fluxes or active fluxing gaseschanges the interfacial relationship of includedparticles with the melt so that gravitationalseparation is facilitated. Fluxing with argon,nitrogen, and/or other gases results in flotation ofentrained matter, while dissolved hydrogen isreduced by partial pressure diffusion. Metal treat-ment takes place in the melting furnace, holdingfurnace, or in line between the furnace and thecasting unit. Rotary degassers have been devel-oped to provide the finest dispersion and inter-mixing of metal and fluxes for these purposes.

Molten metal filtration for the removal ofparticulate contamination was introduced in the1950s and has grown in importance and appli-cation since that time. The first and still amongthe most effective filtration processes are deepbed filters using tabular alumina as the filtrationmedium. Crushed carbon beds are capable offine inclusion removal and a reduction insodium content that is important for manyproducts. Porous foamed ceramics are widelyused for commercial-grade and higher-qualityrequirements. Fused ceramic and refractory fil-tration elements are also available.

Energy Requirements for Melting Alu-minum. Melting is an energy-intensive process;it requires nearly the same amount of energy toraise 1 kg of aluminum to a molten 700 °C(1290 °F) state as it does to raise 1 kg of iron to1500 °C (2730 °F). However, nearly three timesthe volume of aluminum is produced comparedto iron, because of density differences.

The energy requirements for melting alu-minum are presented in the section, “TheoreticalEnergy for Secondary Aluminum” in this chap-ter. When the system boundaries are drawnaround an aluminum melting facility, the mate-rial entering and leaving is aluminum metal.Because no chemical change has occurred, thetheoretical minimum energy requirement is onlythe energy required to melt the metal. Moltenmetal furnace temperatures vary depending onthe furnace, alloys, and other proprietary factors.The theoretical minimum energy requirement tobring room-temperature (25 °C, or 75 °F) alu-minum to a molten 960 °C (1760 °F) metal is0.39 kWh/kg. The theoretical energy requirementfor melting pure aluminum to molten metal atvarious temperatures is presented in Appendix G.

Basic natural gas or oil-fired reverberatoryfurnaces range in efficiencies from approxi-

mately 20 to 45%. The more efficient furnacesemploy recuperation of stack gas heat forreduced melting energy requirements throughcharge preheating or for more efficient burneroperation through preheating combustion air.Furnace condition and operating practices havelarge effects on energy performance. Becauseheat transfer in reverberatory furnaces takesplace principally through radiation, melt surfacetemperatures are considerably hotter, leading tomore rapid oxidation and higher melt losses.

Electric furnaces, typically used in small pro-cessing operations, do not require a flue, andtheir heating chambers can be made nearly air-tight. A sidewell is provided for charging metaland alloying materials. The sidewell removesthe need to open the furnace door and prevents amajor convective heat loss. Energy losses (ex-cluding electrical generation and transmission)in electrical furnace heating are principally dueto conduction and radiation losses from the ex-posed furnace shell. Losses are typically 0.49 to0.81 kWh/kg of aluminum. Induction furnacesare typically more than 90% energy efficient,while gas-fired crucibles are 15 to 28%, andelectrically heated crucibles 83% efficient interms of on-site energy use.

Technological Change in the Next Decade.The industry is constantly evaluating, adopting,and improving furnace technologies and prac-tices. This provides not only energy and environ-mental benefits but also cost savings. New burnertechnologies and oxygen-enhanced combustionsystems are being developed and evaluated tofurther improve efficiency and reduce the costsof melting without increasing emissions. Allindustry segments have transitioned to greaterreliance on scrap and recycling for their metalneeds. Technologies for sorting, handling, andremelting scrap in all forms with optimal metalrecoveries and lowest costs are continuouslybeing developed and refined. New technologiesfor melting thin-gage material to minimize oxi-dation losses are being developed and imple-mented. Also, industry is continually seekingbetter methods to recover the aluminum that istrapped in dross. Additional efforts are directed atthe closed-loop recycling of dross-related wastes,including saltcake.

New laser technologies will speed the insitu chemical analysis of molten metal andminimize the processing time required foralloy compliance. The development of better,longer-lasting ceramic materials for furnace


linings is ongoing and will reduce the timerequired for furnace maintenance. The aluminumindustry has recently published a technologyroadmap in conjunction with the advancedceramics industry, to encourage the develop-ment of superior furnace construction materials(Ref 10.47).

Finally, new melting technologies now underdevelopment offer the prospects for revolution-ary improvements in melting efficiencies thatmay be applicable to the scale and operationaldemands of much of the industry. One excitingdevelopment is that of immersion heating withhigh-watt-density elements for melting as wellas temperature maintenance. The developmentof commercial immersion heaters for aluminumremelting is very likely to occur within the nextfew years.

Ingot Casting

Ingot casting is the solidification of moltenalloys into shapes that are suitable for subse-quent thermomechanical processing or forremelting. Ingots are made by controlled solidi-fication in molds designed to produce thedesired geometrical configuration and metallur-gical characteristics.

Ingot casting is, by itself, not energy-intensive; however, casting is typically a batchprocess, and large quantities of molten metal areheld in furnaces in which alloying, fluxing, anddegassing are performed. Accordingly, conduc-tive and radiant heat losses occur from thesefurnace operations.

Ingot for wrought product applications isalmost universally cast by the semicontinuousdirect-chill (DC) casting process. The processincludes different means of introducing andcontrolling the flow of molten metal into themold, lubrication methods, the use of insula-tion in mold construction, and the injection ofair or imposition of an electromagnetic fieldfor reducing or eliminating contact betweenmolten metal and the mold. The process pro-duces rectangular cross-sectional ingots forrolling; round loglike billets for extrusion;squares for wire, rod, and bar products; andvarious shapes as fabricating ingot in forging.

The DC casting process begins when alu-minum flows from the furnace through troughsto the casting station. At the casting station, thealuminum flows into one or multiple water-cooled stationary molds that rest on the casting

station table. The DC ingot molds are only afew inches deep and form the cross section ofthe ingot or the billet. The ingot is initiallyformed in the water-cooled mold. Whenperimeter solidification has begun, the castingtable is gradually lowered into the casting pit,while additional molten aluminum is suppliedto the top of the mold. The water-cooled moldremains at the top of the pit and continues toshape the casting. Water sprays impinging onthe solid shell continue the solidificationprocess of the molten ingot core. The castingtable is lowered into a casting pit until thedesired length is achieved. After casting, ingotintended for wrought fabrication may be stressrelieved, scalped, cut to length, and homoge-nized. Cutting, shearing, forming, and othermechanical operations, as well as melting,heating, casting, heat treating, and other thermaloperations are used by the product-producingsector.

Casting operations attempt to control thecrystal/grain structure and composition gradi-ent of cast products. Grain size and boundariesare important factors affecting the physical andmechanical properties of the material in castand wrought form. However, because there is asignificant temperature profile across the ingotcross section during solidification, grain struc-ture and composition can vary from surface tocenter. For subsequent fabricating operations,it is usually necessary to remove the skin layerby scalping so that the final product has con-sistent physical properties. The amount ofsurface to be removed is dependent on shape,surface quality, and the depth of undesirablegrain structure and segregation. Scalpings areremelted and reprocessed, which results inadditional energy usage and metal losses dueto oxidation.

A percentage of DC ingots are rejected forquality reasons. Cracking may occur during orafter solidification. Surface defects may form,which affect the acceptability of the ingot forwrought processing. Other specialized stan-dards concern grain structure, segregation, andmicrostructure. At times, ingots are found toexceed alloy specification limits. The process-ing energy used to produce the ingot is then lost,and additional energy is required for remeltingand reprocessing.

Energy Requirements. More than 2,480,000metric tons of primary aluminum were cast intoingots in 2005 (Table 10.3). The tacit energy


consumed in aluminum processing operationscan be divided into three categories:

• Fossil-thermal energy use, which includesfurnaces, heating, and heat treatment opera-tions

• Electrical energy required for heating, saw-ing, and scalping and for motor, pump, andcompressor operation

• Other fuel-consuming operations, includingtransportation

Primary ingot casting has typical metal yieldsfrom 88 to 98% and requires approximately1.01 (1.46tf) kWh/kg of cast ingot product (Appendix E, Table E.2).

The theoretical minimum energy requirementsfor primary and secondary castings are the sameat 0.33 kWh/kg of aluminum (Appendix E, TableE.3). The difference in their actual energy usageresults from their respective initial materials. Inprimary casting, the initial material is molten alu-minum in a holding furnace, while the initialmaterial in secondary casting is metal scrap. Thescrap must first be melted before it enters aholding furnace, giving secondary casting ahigher actual energy use than primary casting.Secondary aluminum was cast into 2,990,000metric tons of ingots in 2005. Secondary castinghas typical yields of 96% and requires approxi-mately 2.50 (2.81tf) kWh/kg of product.

Technological Change in the Next Decade.Ingot casters have strived to improve processyields and to refine practices to provide moreconsistent internal and surface quality inwrought ingot manufacture. These efforts haveresulted in significant progress in surface, subsurface, and metallurgical improvements. Grainrefining by heterogeneous nucleation agents hasbenefited from decades of constant research and

development by primary producers in coopera-tion with master alloy suppliers.

Molten metal fluxing and filtration processescontinue to undergo changes, leading to greaterefficiencies, higher product quality, reduced en-vironmental impact, and reduced costs.

Numerous research programs are directed atthe modeling and prediction of the solidificationprocess for reduced cracking incidence andimproved structural uniformity. The evolutionof mold designs capable of improved surfaces,reduced scalping, and higher production ratescontinues in all wrought ingot production.

All of these developments reflect advances insensors and instrumentation that permit moreaccurate monitoring and comprehensive controlof the casting process.

Rolling

Rolling is the process of reducing ingot thick-ness by passing it between counterrotating steelrolls. Aluminum-rolled products include plate(typically >0.6 cm, or 0.2 in., thick), sheet (typ-ically 0.02 to 0.6 cm , or 0.008 to 0.24 in.,thick), and foil (typically <0.02 cm, or 0.008in., thick). Figure 10.25 shows the unit opera-tions of a typical rolling mill.

Hot, cold, and continuous rolling operationsare used industrially to shape material into thebroad category of flat rolled products. Hot rollingis generally performed at various temperaturesexceeding the recrystallization temperature fordiffering rolling alloys. Cold rolling takes placeat room temperature, but the heat generated canresult in metal temperatures as high as 150 °C(300 °F). Continuous rolling pours moltenmetal directly into strip used for sheet and foiland is significantly more energy efficient thancold or hot rolling. The United States has more

Fig. 10.25 Typical rolling mill processing operations. DC, direct chill


than 50 aluminum rolling mills, 11 of which arecontinuous casters.

Large, high-strength alloy ingot may requirethermal stress relief after casting. Most ingotsare homogenized to reduce intergranular segre-gation and to modify intermetallic particle form.Homogenization may be integral to preheatingfor hot rolling, or the preheating may take placeseparately. Although some ingot may be rolledwith the cast surface intact, rolling surface facesare normally removed by scalping, and the headand butt of the ingot are cropped to assure fin-ished product uniformity.

Hot rolling begins with repetitive reversingmill reductions on stock preheated to tempera-tures in the range of 400 to 500 °C (750 to 930°F). The slab is passed repeatedly back andforth through the rolls until the desired reduc-tion in thickness is achieved. While heat is gen-erated during deformation, the number of passesrequired at the reversing mills may—because ofreducing section thicknesses, time, and coolantapplication—cause losses in temperature thatrequire reheating before continuing the hotrolling process. Slabs are lifted from the hot lineand charged in reheating furnaces until workingtemperatures are restored. End crops are takenduring breakdown rolling as required to main-tain squareness.

Edge-trimming knives remove stock from theedges of the sheet before coiling. The amount ofedge trim required is determined by the depth ofedge cracks and the ragged conditions associ-ated with the unrestrained deformation thatoccurs during hot rolling. The amount of edgetrim required with cropping losses represents asignificant percentage of planned scrap thatmust be returned to the cast shop for remelting.Edge trim losses can be minimized by improvedingot quality and by edge rolling.

The hot rolling process completely changesthe microstructure formed during casting andelongates the grain structure in the direction ofrolling. Even though deformation temperaturesare typically greater than the recrystallizationtemperature, there is inevitably some degree ofequivalent cold work, so that annealing at rerollgage will result in recovery or recrystallizationbefore cold rolling.

Energy Requirements for Rolling Alu-minum. Approximately half of U.S. rolledaluminum products are cold rolled. In 2003,2,421,300 metric tons of cold-rolled productswere produced (Appendix E, Table E.4). Coldrolling has typical yields of approximately

84% and requires approximately 0.64 (1.35tf)kWh/kg of rolled product. Hot rolling has typi-cal yields of approximately 82% and requiresapproximately 0.62 (1.16tf) kWh/kg of rolledproduct.

Theoretically, it is possible to roll productswithout the need for heat treatments and with noloss of material due to trimming or slitting. Inthis case, the minimum theoretical energy to rolla product is composed of only two components:

• The energy required to heat starting stock tothe rolling temperature

• The energy required to deform the shape

The hotter the material, the lower the deforma-tion energy required. The heat capacity equationsrequired to calculate the energy requirement forheating pure aluminum are listed in Table 10.9.The energy required for deforming is given bythe equation E � εσc, where ε is the strain ordeformation defined as ε = ln (ti/tf), where tirepresents the initial and tf the final dimension, σdenotes the yield stress, and c denotes a constantdescribing the shape of the stress-strain curve.The yield stress value for aluminum can vary byas much as a factor of 10 over the hundreds ofalloys that are used by the industry.

The very large variations in alloy properties,particularly the shape and the magnitude of thestress-strain curves, make it possible to calculatetheoretical minimum energy requirements forrolling aluminum only for a specific rollingprocess with a specific alloy. There are a largenumber of heavy-equipment requirements inrolling mill operations that are not confined torolling. Roller and tension levelers, slitters,stretchers, roll formers, and paint or coating linesare examples of energy-consuming operationsthat rely on pumps, motors, and compressors aswell as on mechanical design for efficiency. Arough approximation of the rolling sector mini-mum energy can be made by assuming overallprocess heating efficiencies and electric/-hydraulic system efficiencies and by looking atthe entire rolling sector yield and energy con-sumption. If the overall sector heating efficiencyis 50% and the electric/hydraulic system effi-ciency is 75%, the estimate of the minimumenergy requirement is 0.31 kWh/kg of productfor hot rolling and 0.33 kWh/kg of product forcold rolling. These assumptions imply that coldrolling is operating at approximately 52% overallenergy efficiency and hot rolling at approxi-mately 50% overall efficiency (Appendix E,Table E.3).


Advanced Rolling Technology. Hot rollingtypically requires numerous passes through therolling mills and is energy-intensive. Oneapproach to improving productivity and reducingheating energy is to continuously cast moltenmetal into slab or strip (Ref 10.48). Goingdirectly to thin strip, continuous strip and slabcasting saves the energy required for homoge-nization, scalping, preheating, end and sidetrim, and multiple passes through rolling mills.

Continuous strip casting is in wide currentuse for some sheet and foil specifications andhas demonstrated energy savings of more than25% relative to conventional ingot rolling.These casters convert molten metal directly intoreroll gage sheet at 1 to 12 mm (0.04 to 0.5 in.)thickness. Continuous strip casters employ twincounterrotating water-cooled rolls or belts toaccomplish solidification. Strip casting is nowrestricted to certain low-alloy-content composi-tions. Technology is being developed for morecomplex or more highly alloyed compositions.Very high solidification rates and the extrusioncomponent of casting between cylindrical rollsresult in high degrees of segregation of solute tothe centerline and in cracking tendencies.

While continuous strip casting is recognized asalloy-constrained, an alternative continuous pro-cess for coiled reroll, slab casting is highly alloy-tolerant and results in metallurgical structuresthat closely correspond to those of the ingot andwrought products produced by the DC process.In slab casting, solidification takes place betweenwater-cooled belts or blocks. Slab thicknessvaries but typically corresponds to continuoushot mill entry gages of 75 to 150 mm (3 to 6in.). An individual slab-casting line has a pro-duction capacity ten times that of the largeststrip caster. The cast slab is directly fed into oneor more in-line, low-speed, high-torque hotreduction mills that reduces thickness to coilablegage. Slab-casting processes have successfullyproduced high-strength aluminum alloy sheetas well as challenging products such as beveragecan sheet, but subtle differences in productperformance—especially in formability and

anisotropy—continue to favor the use of conven-tionally hot-rolled sheet in these applications.

A further advancement in rolling technologyis spray rolling. In this process, molten metaldroplets are sprayed directly into the nip of twinrolls, and the material is solidified and consoli-dated directly into sheet material in one step. Inconcept, this offers the most energy-efficientprocess, and recent projects have sought toadvance this technology.

Extrusion

Extrusion is the process of forcing an alu-minum ingot or billet through a steel die toform an elongated shape of consistent crosssection. Extruded products include rods, bars,tubes, and specialized products interchange-ably called shapes, sections, or profiles. Figure10.26 shows the unit operations of a typicalextrusion plant. After rolling, extrusion is thesecond most common processing technique foraluminum. Aluminum extrusion is remarkablebecause the process combines high productivitywith an essentially infinite variety of extremelycomplex shapes, cross sections, or profiles thatcannot be economically duplicated in any otherprocess. Furthermore, aluminum can be readilyextruded; this process is either extremely difficultor impractical for many other metals. There are190 extrusion plants in the United States. Manyof these plants run multiple extrusion presses.

It is possible to produce almost any cross-sectional shape, within wide limits. Throughthe use of hollow stock and floating or fixedmandrels, hollow shapes or cross sections withcomplex enclosed configurations can be pro-duced. The extrusion process is capable of pro-ducing a cross section with a weight of a fewgrams to more than 300 kg/m, a thickness ofless than 1 to over 250 mm (0.04 to over 98in.), circumscribing diameters of 5 to 1000 mm(0.2 to 39 in.), and lengths in excess of 30 m(98 ft). The appropriate choice of alloy andextrusion conditions can result in an optimalcombination of properties for a particular

Fig. 10.26 Typical aluminum extrusion processing operations. DC, direct chill


application. Such properties may include ten-sile strength, toughness, formability, corrosionresistance, and machinability. Countless extru-sion products are made in the United States.These include frame and supporting shapes forwindows and doors, carpet strips, householdbath enclosures, screens, bridge structures,automotive parts, aerospace components, andmany other consumer products.

Operations at individual plants vary widelydepending on the cross sections and alloys pro-duced. Extrusion presses range in capacity up to15,000 tons, but the most common are perhaps2500 tons. Presses may extrude vertically orhorizontally, but virtually all modern extrusionpresses are horizontal. Typically, billets indiameters up to 275 mm (11 in.) are preheatedto temperatures ranging from 450 to 550 °C(840 to 1020 °F) depending on the alloy, prod-uct design, and the desired mechanical charac-teristics. The preheated billet is charged to theextrusion press container and forced by hydraulicpressure through the extrusion die.

There are essentially two processes for extru-sion production. In the direct extrusion process,the billet is hydraulically pressed through thedie, while in the indirect extrusion process, thedie is forced over the billet. In direct extrusion,the billet surface is retained in the extrusioncontainer and contributes to butt loss, whichmay total 8% of the starting billet weight.Because the billet surface is not extruded tobecome part of the product, scalping is notrequired. For indirect extrusion, the billet sur-face becomes an integral part of the product, andso, scalping before extrusion is essential.

The single largest area for energy improve-ment in extrusion technology is the reduction ofprocess scrap, including butt losses. Extrusionproduct specifications include significant sur-face-quality and chemical-finishing criteria.Defects include torn surface; die pickup, whichis often related to the billet microstructure; andnonfill. Variable response to chemical finishingresults in appearance and color mismatches thataffect product acceptability.

Energy Requirements for Extruding Alu-minum. The United States produced over1,826,000 metric tons of extruded aluminumproducts in 2003 (Appendix E, Table E.4).Extrusion processes have typical yields of 69%and require approximately 1.30 (1.52tf) kWh/kgof extruded product.

Theoretically, it is possible to extrude productswithout the need of additional heat treatments

and without loss of material. The minimum theo-retical energy to extrude a product, in such acase, is composed of only two components:

• The energy required to preheat the billet toextrusion temperature

• The energy required to deform the materialthrough a die

The hotter the material, the lower the defor-mation energy required. The simpler the die, thelower the extrusion-energy requirement. Theheat capacity equations needed to calculate theenergy requirements for heating pure aluminumare listed in Appendix G. The energy required todeform the material through the die is highlydependent on the size and shape of the productand the die design. Calculation of the minimumextrusion force is very complex and can only beestimated with theoretical and empirical models.Typical formulae have the following simplifiedform:

F = Ao(σm/η) ε

where ε, the strain, corresponds to the reductionarea, ε = ln(Ao/Af), σm is the mean stress for thestrain, Ao is the original cross-sectional area, andη is an efficiency factor.

The very large variations in alloy properties,particularly the infinite numbers of possibleshapes, make it impossible to calculate a theoret-ical minimum energy requirement. This valuecan only be determined by analyzing a specificprocess, a specific extruded profile, and a specificalloy. The perimeter of the profile and the radiusof intersecting edges have a large influence onthe force required for extrusion. A rough approx-imation of a minimum energy value can be madeby examining the entire aluminum extrusionindustry yield and energy values and assumingan overall process heating efficiency and electric/hydraulic-system efficiency. This approach pro-vides an estimate of the minimum energy, 0.44kWh/kg of aluminum, when efficiencies areassumed to be 50% for heating and 75% for theelectric/hydraulic system. These assumptionsimply that overall extrusion facilities operate atapproximately 34% energy efficiency.

Shape Casting

Shape casting, or the casting of engineereddesigns, enables the production of simple andcomplex parts that meet a wide variety of needs.


The process produces parts weighing ounces toparts weighing several tons. Over 1000 compa-nies cast aluminum products. Some of these arevery small garage-shop operations that supplyniche market casters, for example, specialty toy-makers. Over 15 plants have capacities to pro-duce >25,000 tons of product per year. Figure10.27 shows the unit operations of a typical alu-minum shape-casting foundry. These operationsvary significantly depending on the size of theoperation, the processes employed, the com-plexity of the parts, the alloy compositions, andthe type of castings being produced.

The basic casting process consists of meltingand alloying aluminum and pouring or injectingmolten metal into molds containing single ormultiple cavities of the desired shape. Theimportant casting processes for engineered alu-minum castings are pressure die, permanentmold, green and dry sand, plaster, and invest-ment casting.

In pressure die or die casting, metal is injectedat pressures up to 10,000 psi into water-cooledsteel dies. Productivity rates are high, and theprocess can be highly automated. While surfacequality and dimensional accuracy are excellent,the typical die casting contains a degree of inter-nal unsoundness associated with nondirectionalsolidification, entrapped gases and inclusions re-sulting from turbulent metal flow, and the pres-ence of air and lubricants in the die cavity.

Permanent mold or gravity die castings areproduced by introducing molten metal by grav-ity or countergravity means into iron or steelmolds. Productivity, surface quality and dimen-sional accuracy are lower than in pressure diecasting. Internal soundness, depending on theextent to which sound molten metal treatmentfor hydrogen elimination and oxide removal arepracticed and the principles of directionalsolidification are employed, can meet the mostchallenging quality standards. Low-pressurecasting is usually considered a variation of thepermanent mold process, even though dry sandand plaster cast parts have been produced. Inthis process, molten metal is forced by theapplication of pressure to rise through a tubeinto a mold mounted over the furnace. It has the

advantages of a significantly reduced gross-to-net weight ratio and correspondingly lowertrimming costs.

Green and dry sand casting can also yieldhigh-integrity parts. In green sand casting, sand,binders such as clays, and moisture are blendedto provide the molding medium. Patterns mayinclude loose pieces, wood models, cast match-plates, and molded Styrofoam (Dow ChemicalCompany). For dry sand molding, air or thermo-setting chemicals coat sand particles so that thefinished mold after curing offers superior sur-faces, dimensional accuracy, and shelf life.

Investment molds are produced by repetitiveimmersion of plastic, wax, or other low-temper-ature-melting and volatile pattern material intoceramic slurries. After drying, the hardenedcasing containing the pattern is heated to atemperature at which the pattern material iseliminated. Typically, the mold is preheatedbefore pouring and may be filled under vac-uum. Investment castings offer extremely finefinishes, thin walls, and excellent dimensionalaccuracy. Plaster molding offers the sameadvantages and may be used for the productionof thicker sections, large parts for whichinvestment is less suited.

Molten aluminum is required as the basis forall foundry production. High-volume castingoperations may acquire part or all of their metalrequirements as molten metal delivered over-the-road in insulated crucibles. Other foundriesmelt and process prealloyed ingot, RSI, andinternal scrap, including gates and risers. Allmelting and melt-processing technologies andconsiderations described in the section “Melt-ing, Alloying, and Melt Treatment” are applica-ble to foundry operations.

Energy Requirements for Shape CastingAluminum. Nearly 2,413,000 metric tons ofshape-cast products were produced in 2003.Shape casting has typical yields of only 45%and requires approximately 2.56 (2.64tf)kWh/kg of cast product. The energy consumedis almost exclusively related to furnace andheating operations (Appendix E, Table E.4).

The theoretical minimum energy requirementfor shape casting can be calculated from the

Fig. 10.27 Typical aluminum product shape-casting operations


energy required to go from room temperature toliquid metal plus some superheat value (seethe section “Theoretical Energy for SecondaryAluminum” in this chapter). Pure aluminummelts at 660 °C (1220 °F). The minimumenergy required to produce liquid aluminum at660 °C (1220 °F) is approximately 0.3 kWh/kg.

Alloy composition; superheat requirements;mold sprue; gates, runners, and riser systems;and postcasting heat treatments vary by molddesign and casting practices. A rough approxi-mation of the aluminum shape cast sector min-imum energy requirement can be made bylooking at the entire sector yield and energyvalues (Appendix E, Table E.2) and assumingan overall process heating efficiency and elec-tric/hydraulic-system efficiency. The estimateof the minimum energy requirement using thisapproach is 0.60 kWh/kg, when a 50% overallheating efficiency and a 75% electric/hydraulicsystem efficiency are assumed. These assump-tions imply that shape casting facilities operateat approximately 23% overall energy efficiency.

Technological Change in the Next Decade.Castings are among the most cost-effective andversatile solutions to part design and perform-ance challenges. The range of available alloysand properties provides combinations of manu-facturability and product characteristics forone-of-a-kind, prototype, limited, or high-volumeapplications. Castings are near-net shape, withthe potential for precise integral internal pas-sages and complex shapes. Aluminum is cast inmore alloys with a wider range of physical andmechanical properties by more processes thanany competing metal system.

The most important trend affecting aluminumcasting production is its continuous growth inautomotive applications. The advantages of alu-minum for many powertrain components, includ-ing transmission cases, oil pans, pistons, intakemanifolds, cylinder heads, and engine blocks,have been reflected in its wide adoption.

Squeeze casting and semisolid forming haveemerged as candidates for a new generation ofprocess capabilities producing heat treatable,high-integrity pressure die castings that uselower-impurity compositions, vacuum, and drylubrication.

Thermal Treatments

A significant component of energy use in thealuminum industry is the heat treatment of metaland products. The physical and mechanical

properties of aluminum alloys in any productform can be controllably altered by thermaltreatment. Thermal treatments are used to softenthe material and to recrystallize the grain struc-ture. Other aluminum alloys, principally thosecontaining copper, magnesium, silicon, and zinc,can be thermally treated to significantly improvestrength through the dissolution and reprecipita-tion of soluble phases. Aluminum alloys are cat-egorized as heat treatable if thermal treatmentshave significant hardening benefits or non-heattreatable if the alloy is unresponsive to thermaltreatments for hardening purposes. Among thelatter are alloys dependent on work hardeningfor strengthening and those whose properties areessentially fixed after solidification.

All product types, including sheet, plate, foil,wire, rod, bar, extrusions, forgings, and cast-ings, are produced in heat treatable alloys. Themajority of extrusions, forgings, and a largepercentage of plate and castings are heat treated.Heat treatment facilities are integral to largeroperations. Commercial firms also provide con-tracted heat treatment services.

Annealing is performed at temperatures from300 to 500 °C (570 to 930 °F) to reduce strength,improve formability and ductility, lower residualstress levels, and improve dimensional stabilityin cast and wrought products. Because electricalconductivity is adversely affected by elementsretained in solution, annealing is also used inelectrical and electronic applications. Deep-drawn sheet is normally annealed. Intermediateannealing is the usual practice in rolled productmanufacture to permit subsequent cold reduc-tions. The final temper of non-heat treatablerolled products may include annealing, partialannealing, or stabilization treatments. Mostannealing is a batch operation.

Heat treatable aluminum alloys containintermetallic metallurgical phases that can bedissolved at elevated temperature (up to550 °C, or 1020 °F) and retained in solid solu-tion by rapid quenching. Solution heat treat-ment is batch or continuous. Sheet and foil canbe heat treated continuously through accumu-lator towers, and plate, castings, and forgingsby conveyer furnaces. Coiled sheet, plate, andextrusions are more typically batch treated. Ineither case, the product must be held at solu-tion temperature long enough for completesolution to occur and for desirable changes inthe shape or form of the insoluble intermetallicsthat are present in the microstructure. Thequench medium is generally water but may be


glycol where greater control is required atroom temperature. The retained metastablesolid solution permits precipitation hardeningat intermediate temperatures (170 to 300 °C, or340 to 570 °F) for fully hardened, partiallyhardened, or overaged conditions. Each offerscombinations of strengths, ductilities, toughness,stability, resistance to stress corrosion, andhardness not achievable through solidificationor work hardening. Precipitation-hardeningfurnaces are also batch or continuous.

The high rate of solidification in many castingprocesses results in a degree of solution reten-tion in heat treatable compositions, permittingage or precipitation hardening to be employedwithout solution heat treatment. This method isextensively used in extrusions that can be pressquenched by forced air or water mist to improvesolution retention.

Recommended solution heat treatment prac-tices have been standardized to reflect worst-caseconditions. The cycle defines the minimum timeat the required temperature for successfully treat-ing the part requiring the longest exposure. Asafety factor is usually also applied to assure thatreheat treatment will not be necessary. Practicesare typically generic, applying to specific alloysand tempers without regard for section thicknessor degree of metallurgical refinement. Thinnerwall castings, forgings, and extrusions generallyrespond to heat treatment more rapidly. Finergrain and dendrite cell sizes are reflected insmaller, more dispersed solute particles that canbe more rapidly dissolved. Solution heat treat-ment time can therefore be patterned to specificproducts and manufacturing processes and thecombination of finer metallurgical structures.Decreased variability can result in substantiallyreduced cycle times and energy costs.

New heating technologies are being studied toreduce energy requirements through more rapidheating to treatment temperature. Fluidized bedsand infrared heating can be used to shorten heat-up times but do little to accelerate either the rateof solution or microstructural change when solu-tion temperatures are reached.

Another approach being investigated is theuse of sensible heat to reduce energy require-ments. While there are metallurgical concerns,extrusions, castings, and forgings can be placedin heat treatment furnaces directly from themold or die, thereby preserving the latent heatof the casting or final forming operation.

Energy Requirements for Thermal Treat-ment. The theoretical energy required for all

thermal treatments can be calculated from thespecific heat or heat capacity of the various alu-minum alloys. For example, it requires 0.06kWh/kg (95 Btu/lb) to heat A356 to its solutionheat treatment temperature, 0.05 kWh/kg (80Btu/lb) to reach annealing temperature for 1100alloy, and 0.02 kWh/kg (32 Btu/lb) to reach pre-cipitation-hardening temperature for alloy 2024.

Time at temperature for each procedure variesdepending on alloy and product form but canexceed 12 h. While most energy is consumed inraising the product and oven to temperature,additional energy is required to maintain the tem-perature for the duration of the cycle. Sustainingenergy requirements is exclusively a function offurnace design and condition. Standard quenchtemperatures include room temperature: 65, 80,and 100 °C (150, 180, and 212 °F). Large vol-umes of water must be heated and maintained attemperature by steam or other means for thelatter practices.

REFERENCES

10.1. “Report for U.S. Department of Energy,”by W. Choate and J. Green, BCS Inc.;http://www.eere.doe.gov (accessed July2007)

10.2. Aluminum Statistical Review for 2005,The Aluminum Association, Inc., Wash-ington, D.C., 2006, p7 and Table 16.The Aluminum Association’s statisticalreviews are issued annually and assem-ble, in one document, important detaileddata from primary production, to mar-kets for finished goods, to the recoveryof scrap. Both U.S. and world informa-tion are included.

10.3. A.R. Burkin, Ed., Production of Alu-minium and Alumina, John Wiley &Sons, 1987, p 38, 49, 57, 63. This book,published on behalf of the Society ofChemical Industry, is one of the onlytexts to provide a comprehensiveoverview of the state-of-the-art of all sci-ence and technologies associated withprimary aluminum production in a singlevolume. It is a detailed, academic volumethat covers chemistry, thermochemistry,fluid dynamics, process dynamics, and soon from both a theoretical and practicalperspective.

10.4. D. G. Altenpohl, Aluminum: Technology,Applications, and Environment; A Profile


of a Modern Metal, 6th ed., The Alu-minum Association, Inc. and the Miner-als, Metals and Materials Society, 1998.This book is described in its foreword ashaving “global recognition as the defini-tive educational text and reference bookfor aluminum industry participants; abroad range of aluminum fabricators andusers; students; and the scientific, engi-neering, and academic community.”

10.5. Aluminum Industry Vision: SustainableSolutions for a Dynamic World, The Alu-minum Association, Inc., Washington,D.C., Nov 2001, p 3, 22. The Vision doc-ument identifies key needs of the industryand outlines a comprehensive aluminumresearch and development agenda.

10.6. R. D. Naranjo, E.-P.H. Fu, and M. Gwyn,Castings Drive Fuel Efficiency, Mod.Cast., Vol 94 (No. 9), Sep 2004, p. 20

10.7. Life Cycle Inventory Report for theNorth American Aluminum Industry,The Aluminum Association, Inc., Wash-ington, D.C., 1998. This report providesinformation on a life-cycle inventorystudy of the North American aluminumindustry in 1995. The report is the resultof extensive surveying and contains thebest and most complete industry per-formance information of any recentstudy. Chapter 3 is an adapted versionand Appendix E is a summary of thereport.

10.8. Energy and Environmental Profile ofthe U.S. Aluminum Industry, U.S.Department of Energy, Office of Indus-trial Technologies Program, Washington,D.C., July 1998. This report reviewsthe energy and environmental charac-teristics of the key technologies used inthe major processes of the aluminumindustry.

10.9. Aluminum Statistical Review for 2003,The Aluminum Association, Inc.,Washington, D.C., p 7, 10–12, 19–23,38, 42, 92

10.10. “Bauxite and Alumina Statistics andInformation,” U.S. Geological Survey,Minerals Information; http://minerals.usgs.gov/minerals/pubs/commodity/bauxite/ (accessed July 2007)

10.11. Life Cycle Inventory Report for theNorth American Aluminum Industry,The Aluminum Association Inc., Wash-ington, D.C., p E-3, E-6, E-8, E-12

10.12. Technology Roadmap for BauxiteResidue Treatment and Utilization, TheAluminum Association, Washington,D.C., Feb 2000. This booklet contains acomprehensive discussion of bauxiteresidue.

10.13. Energy and Environmental Profile of theU.S. Aluminum Industry, U.S. Depart-ment of Energy, Office of IndustrialTechnologies Program Washington, D.C.,July 1998, p 29

10.14. Alumina Technology Roadmap, TheAluminum Association, Inc., Washing-ton, D.C. Nov 2001. This documentidentifies performance goals and de-scribes 25 research and developmentareas that are a priority for the globalalumina industry.

10.15. P.J. Ellis, “Tutorial: Petroleum CokeCalcining and Uses of Calcined Petro-leum Coke,” Speech given at AIChE2000 Spring National Meeting, ThirdInternational Conference on RefiningProcesses, March 2000, p 13

10.16. The Manufacture of Carbon andGraphite, Industrial Graphite Engineer-ing Handbook, UCAR Carbon Company,Inc., Parma, OH, 2001, p 1–6

10.17. W. Haupin, History of Electrical Con-sumption by Hall-Heroult Cells, Hall-Heroult Centennial, W. Peterson and R.Miller, Ed., The Metallurgical Societyof AIME, New York, 1986, p 106. Thisarticle provides a detailed historical-review of energy usage in the Hall-Heroult process, from the first cells to1986.

10.18. R.P. Pawlek, 75 Years of Developmentof Aluminum Electrolysis Cells, Alu-minium, Vol 75 (No. 9) 1999, p 734–743

10.19. B.J. Welch, Aluminum Production Pathsin the New Millennium, J. Met., Vol51(No. 5), Feb 1999, p 24–28

10.20. Appendix A-9: Thermodynamics ofElectrochemical Reduction of Alu-minum, Report of the American Societyof Mechanical Engineers: TechnicalWorking Group on Inert Anode Tech-nologies, U.S. Department of Energy,Office of Industrial Technologies, Wash-ington, D.C., July 1999, p 3, 28. Thisreport reviews the literature and patentsconcerning inert anode technologies. Itpresents a theoretical discussion of alu-minum reduction and has input from


over eleven experts in the field. Carbonanode voltage equation E � 1.898 �0.0005728 � T, where E is in V (directcurrent), and T is in degrees Kelvin.

10.21. N. Jarrett, W.B. Frank, and R. Keller,Advances in Aluminum Smelting, Met-all. Treatises AIME, Vol 1 (No. 93), 1981,p 137

10.22. U.S. Environmental Protection Agency,Office of Air and Radiation; www.epa.gov/globalwarming/ (accessed July 2007)

10.23. G.D. Brown, M.P. Taylor, G.J. Hardie,and R.W. Shaw, “TiB2 Coated AluminumReduction Cells: Status and FutureDirection of Coated Cells in Comalco,”paper presented at the Queenstown Alu-minum Smelting Conference, Nov 26,1998

10.24. M.P. Taylor, G.J. Hardie, F.J. StevensMcFadden, and W. Uru, “Use ofRefractory Hard Cathodes to ReduceEnergy Consumption in AluminumSmelting,” paper presented at the SecondInternational Conference on ProcessingMaterials for Properties, TMS 2000Technical Program

10.25. L.Boxall, A.V. Cooke, and W. Hayden,TiB2 Cathode Material: Application inConventional VSS Cells, J. Light Met.,Vol 36 (No. 11), Nov 1984, p 35–39

10.26. Low Energy Aluminum Reduction Cellwith Induced Bath Flow, U.S. Patent4,602,990

10.27. T.M. Leeuwen, An Aluminum Revolu-tion, Desk Notes, Credit Suisse FirstBoston Corporation, June 22, 2000

10.28. Appendix A-9: Thermodynamics of Elec-trochemical Reduction of Aluminum,Report of the American Society ofMechanical Engineers: Technical Work-ing Group on Inert Anode Technologies,U.S. Department of Energy, Office ofIndustrial Technologies, July 1999. Inertanode voltage equation, E � 2.922 �0.0005712 � T, where E is in V (directcurrent), and T is in degrees Kelvin

10.29. Appendix A-9: Thermodynamics ofElectrochemical Reduction of Alu-minum, Report of the American Societyof Mechanical Engineers: TechnicalWorking Group on Inert Anode Tech-nologies, U.S. Department of Energy,Office of Industrial Technologies, July1999

10.30. Inert Anode Roadmap: A Framework forTechnology Development, The AluminumAssociation, Inc. and the U.S Departmentof Energy, Office of Industrial Technolo-gies Program, Feb 1998, p 1. This docu-ment describes the potential benefits, per-formance targets, technical barriers, andtechnology development steps associatedwith the research and development andproduction of inert anodes

10.31. D.B. Rao, U.V. Choudry, T.E. Erstfeld,R.J. Williams, and Y.A. Chang, Extrac-tion Processes for the Production of Alu-minum, Titanium, Iron, Magnesium, andOxygen from Nonterrestrial Sources,Space Resources and Space Settlements,NASA Ames Research Center SummerStudy, Moffett Field, CA, 1977, p 15

10.32. W. Haupin, Reflections on the Hall-Her-oult Process, Molten Salts Bulletin, 59thed., Ecole Polytechnique, Universitarirede Marseille

10.33. C. Brown, Next Generation VerticalElectrode Cell, J. Met. Vol 53 (No. 5),May 2001, p 39–42

10.34. J.N. Hryn, “Inert Metal Anode,” projectsummary form from the 2004 AnnualAluminum Portfolio Review, Argonne,IL, Oct 19–20, 2004

10.35. V. Garcia-Osorio, B.E. Ydstie, andT. Lindstad, “Dynamic Model for VaporRecovery in Carbothermic AluminumProcess,” Carnegie-Mellon Universityand SINTEF Materials Technology

10.36. W.T. Choate, A.M. Aziz, and R. Fried-man, New Technologies Will Sustain theU.S. Primary Aluminium Industry, LightMetals 2005, H. Kvande, Ed. TMS, p 495

10.37. W. Choate and J. Green, Technoeco-nomic Assessment of the CarbothermicReduction Process for AluminumProduction, Light Metals 2006, Vol. 2:Aluminum Reduction Technology, T.J.Galloway, Ed., TMS, p 94

10.38. K. Johansen, J.A. Aune, M. Bruno, andA. Schei, “Carbothermic Aluminum—Alcoa and Elkem’s New ApproachBased on Reactor Technology to MeetProcess Requirements,” paper presentedat the Sixth International Congress onMolten Slags, Fluxes, and Salts, June2000 (Stockholm)

10.39. H. Myklebust and P. Runde, Green-house Gas Emissions from Aluminum


Carbothermic Technology Compared toHall-Heroult Technology, Light Metals2005, H. Kvande, Ed., TMS, p 519

10.40. “Toth Aluminum Corporation’s BipolarClay-Aluminum Technology and Pro-ject,” seminar ed., paper presented atTMS annual meeting, Feb 11, 2001 (NewOrleans, LA)

10.41. New Technology is a Breakthrough forAutomotive Aluminum Recycling,Alum. Now, Nov/Dec 2000, p 28

10.42. W.T. Choate and J.A.S. Green, Model-ing the Impact of Secondary Recovery(Recycling) on U.S. Aluminum Supplyand Nominal Energy Requirements,Light Metals 2004, TMS, p 913–918

10.43. J.H.L. Van Linden, Melt Loss Reductionin Recycling Processes, Katholieke Uni-versiteit Leuven, May 1988

10.44. E.L. Rooy, Aluminum Dross—Liabilityinto Opportunity, Light Met. Age, June1995

10.45. Hertwich Engineering, “AdvancedRemelt Technology at Hydro AluminumPortalex,” Aluminum 2004 (Chicago, IL)

10.46. New Aluminum Melting Process Pro-vides Energy Savings, JOM, June 2006,p 6

10.47. Metal Melting Practices and Procedurefor Efficiency and Effectiveness, Califor-nia Cast Metals Association, sponsoredby the California Energy Commission,Summer 2001, p 20

10.48. Metal Melting Practices and Procedurefor Efficiency and Effectiveness, Califor-nia Cast Metals Association, sponsoredby The California Energy Commission,Summer 2001. This and its companionbooklet, Foundry Energy Use Study andConservation Manual, provide detailedfurnace design information and opera-tional survey data on melting and fur-nace practices for nonferrous metals inthe State of California

ALUMINUM, by several measures, is one ofthe most energy-intensive (Btu/lb) materialsproduced, ranking at the top among the majorproducts of the United States. Table A.1 shows acomparison of the on-site energy requirementsfor several major products manufactured in theUnited States. These values do not include theenergy content of fuels used as materials or thegeneration and transmission losses associatedwith electricity production.

The data in Table A.2 are the on-site processenergy requirements to produce the correspon-ding products plus the energy content of fuelsused as materials and the generation and trans-mission losses associated with electricity pro-duction. Examples of the energy content of fuelsused as materials (feedstock energy) are: 22,681Btu/lb is the feedstock energy for ethylene;wood products are commonly assumed to haveno feedstock value because they are renewable

APPENDIX A

Energy Intensity of Materials Produced in the United States

Table A.1 Energy requirements to produce materials in the United States (2002)Material Btu/yr Btu/lb lb/yr Data sources

Paper and paper board 2.75 � 1015 15,590(a) 1.76 � 1011 Ref A.1Gasoline 2.41 � 1015(b) 2659 9.07 � 1011 Ref A.2, A.3Iron and steel 1.79 � 1015(b) 8700 2.06 � 1011 Ref A.4, A.5Ethylene 4.22 � 1014(b) 8107 5.21 � 1010 Ref A.6, A.7Aluminum (primary ingot) 2.66 � 1014(b) 44,711 5.95 � 109 Appendix E and Table 10.3 in Chap 10Distillate 3.63 � 1014(b) 990 3.67 � 1011 Ref A.2, A.3Ammonia 3.53 � 1014(b) 12,150 2.90 � 1010 Ref A.6, A.7Propylene 4.30 � 1013(b) 1351 3.18 � 1010 Ref A.6, A.7Jet fuel 1.46 � 1014(b) 990 1.47 � 1011 Ref A.2, A.3Coal 1.29 � 1014(b) 60 2.14 � 1012 Ref A.8Benzene 2.03 � 1013(b) 1255 1.62 � 1010 Ref A.6, A.7

(a) Calculated from Btu/yr value and lb/yr value. (b) Calculated from Btu/lb value and the value for lb/yr

Table A.2 Gross energy requirements to produce materials in the United States (2002)Material Btu/yr Btu/lb lb/yr Data sources

Paper and paper board 2.75 � 1015 15,590(a) 1.76 � 1011 Ref A.1Gasoline 2.22 � 1015(b) 2659 8.34 � 1011 Ref A.2, A.3Iron and steel 1.79 � 1015(b) 8700 2.06 � 1011 Ref A.4, A.5Ethylene 1.74 � 1015 33,470(a) 5.21 � 1010 Ref A.6, A.7Propylene 8.90 � 1014 27,978(a) 3.18 � 1010 Ref A.6, A.7Ammonia 6.95 � 1014 23,917(a) 2.90 � 1010 Ref A.6, A.7Benzene 5.43 � 1014 91,317(a) 5.95 � 109 Ref A.6, A.7Aluminum (primary ingot) 5.38 � 1014(b) 33,305 1.62 � 1010 Appendix E and Table 10.3 in Chap 10Distillate 3.63 � 1014(b) 990 3.67 � 1011 Ref A.2, A.3Coal 1.29 � 1014(b) 60 2.14 � 1012 Ref A.8Jet fuel 1.46 � 1014(b) 990 1.47 � 1011 Ref A.2, A.3

(a) Calculated from Btu/yr and lb/yr data. (b) Calculated from Btu/lb and lb/yr data




resources; and petroleum-calcined coke used asa raw material in aluminum production has afeedstock energy of 15,250 Btu/lb.

REFERENCES

A.1. “Manufacturing Energy Consumption Sur-vey,” Energy Information Agency, DOE

A.2. H. Brown, Energy Analysis of 108Processes, Drexel University, 1996

A.3. “Petroleum Annual 2003,” Energy Infor-mation Agency, DOE

A.4. J. Stubbles, Energy Use in U.S. Steel Indus-try: A Historical Perspective and FutureOpportunities, Sept 2000, p 23

A.5. Steel Industry of the Future, FY 2004 An-nual Report, DOE, p 1

A.6. Energy and Environmental Profile of theU.S. Chemical Industry, DOE-OIT, May2000, p 28, 30, 32

A.7. Guide to the Business of Chemistry, Ameri-can Chemical Council, 2003, p 31

A.8. “Coal Industry Annual 2003,” Energy In-formation Agency, DOE

CALORIFIC ENERGY VALUES are the en-ergy content inherent to the material. Except forpure materials, for example, propane, these val-ues vary depending on the raw materials usedand the final product formulations.

Process energy is a measure of the energy re-quired to manufacture the material. Process en-ergy values are also variable and depend onequipment efficiency estimates and systemboundaries. These values for crude-oil-derived

products are variable and depend on the specificcrude processed, refinery configuration, localproduct specifications, and refinery efficiency.Process energy values from the California En-ergy Commission include the energy for pro-cessing as well as factors for raw material, forexample, delivery of crude to refiners and trans-porting fuels to their point of use.

Tacit or gross energy is the sum of thecalorific and process energy values.

APPENDIX B

Energy Values forEnergy Sources and Materials




Tabl

e B

.1En

ergy

val

ues

for

ener

gy s

ourc

es a

nd m

ater

ials

Cal

orifi

c en

ergy

val

ues

(pri

mar

y en

ergy

)P

roce

ss e

nerg

y va

lues

(se

cond

ary

ener

gy)

Taci

t en

ergy

Btu

per

In

put

Btu

per

Btu

per

Dat

aB

tu p

er c

omm

on u

nit

Btu

per

Dat

aA

ppen

dix

E

Inpu

tE

nerg

y so

urce

unit

com

mon

uni

tin

put

unit

sour

cere

quir

ed t

o pr

oduc

ein

put

unit

sour

cein

put

unit

unit

Fue

ls a

nd f

uels

use

d as

mat

eria

ls

Fuel

oil,

med

ium

kg13

9,00

0 B

tu/g

al43

,380

Ref

B.1

5000

Btu

/gal

1321

Ref

B.2

44,7

00kg

Fuel

oil,

light

kg15

0,00

0 B

tu/g

al46

,813

Ref

B.1

5000

Btu

/gal

1321

Ref

B.2

48,1

30kg

Die

sel

L5,

670,

000

Btu

/bbl

35,6

67R

ef B

.352

00 B

tu/g

al13

74R

ef B

.437

,040

LK

eros

ene

L5,

670,

000

Btu

/bbl

35,6

67R

ef B

.552

00 B

tu/g

al13

74R

ef B

.437

,040

LG

asol

ine

L5,

198,

000

Btu

/bbl

32,6

98R

ef B

.512

,000

Btu

/gal

3170

Ref

B.4

35,8

70L

Nat

ural

gas

m3

1027

Btu

/ft3

36,2

68R

ef B

.530

Btu

/ft3

1059

...37

,330

m3

Bitu

min

ous/

subb

itum

inou

skg

11,1

10 B

tu/lb

24,4

93R

ef B

.660

Btu

/lb13

2R

ef B

.724

,630

kgC

alci

ned

coke

kg15

,250

Btu

/lb33

,620

Ref

B.8

179

Btu

/lb39

5R

ef B

.734

,010

kgPi

tch

kg6,

065,

000

Btu

/bbl

38,1

52R

ef B

.718

Btu

/lb40

Ref

B.7

38,1

90kg

Gre

en c

oke

kg14

,200

Btu

/lb31

,305

Ref

B.8

Btu

/lb50

0R

ef B

.231

,810

kgPr

opan

eL

3,82

4,00

0 B

tu/b

bl24

,055

Ref

B.9

Btu

/lb..

...

.24

,050

LC

oal

kg10

,240

Btu

/lb22

,574

Ref

B.5

60 B

tu/lb

132

...

22,7

10kg

Ele

ctri

city

Ele

ctri

ckW

h1

kWh

3412

Ref

10

Hyd

roel

ectr

ic u

tility

0R

ef B

.10

3412

kWh

Avg

. U.S

. ele

ctri

c(a)

kWh

1 kW

h34

12R

ef 1

0A

vera

ge U

.S. g

rid

6478

Tac

it B

tu98

90kW

hPr

imar

y A

l ele

ctri

c(b)

kWh

1 kW

h34

12R

ef 1

0Pr

imar

y A

l ele

ctri

c(b)

4158

Tac

it B

tu75

70kW

hC

oal-

fired

ele

ctri

ckW

h1

kWh

3412

Ref

10

Coa

l-fir

ed u

tility

6891

Tac

it B

tu10

,300

kWh

(a)

Taci

t Btu

acc

ount

s fo

r ge

nera

tion

and

tran

smis

sion

ene

rgy

loss

es f

or a

vera

ge U

.S. g

ener

atio

n. (

b) T

acit

Btu

for

ele

ctro

lysi

s an

d an

ode

man

ufac

ture

is lo

wer

then

gen

eral

ele

ctri

city

bec

ause

of

the

larg

e (5

0.2%

) hy

dro

com

pone

nt.

Appendix B: Energy Values for Energy Sources and Materials / 227

The process energy (secondary energy) val-ues contribute approximately 3% additional en-ergy to the total carbon-base fuel energy of theU.S. aluminum industry. This includes the en-ergy expended worldwide for the production ofaluminum in the United States. The UnitedStates does not mine bauxite but refines it tosupply 47% of the alumina needed and importsthe remaining 53% of alumina required for alu-minum production.

Tacit electric energy conversion factors varysignificantly depending on the fuel source usedto generate electricity. The large variation inelectric tacit conversion factors creates the needfor careful analysis when comparing differentstudies. A completely coal-fired electric-basedsmelting operation requires 2.5 times greatertacit energy than a completely hydroelectricsmelting operation. This report uses an averagetacit based on the electric fuel sources of alu-minum smelting operations. (Appendix C,Table C.1)

REFERENCES

B.1. Cogeneration Technologies, Houston, TX.http://www.cogeneration.net (accessed July2007)

B.2. Data derived from Drexel University study.B.3. Energy Information Agency, DOEB.4. California Energy CommissionB.5. Annual Energy Outlook 2004, Energy In-

formation Agency, Jan 2004, p 262B.6. Coal Industry Annual 2000, Energy Infor-

mation Administration, Jan 2002, p 284.This report uses an average value for bitu-minous and subbituminous coals

B.7. “Energy and Environmental Profiles, Alu-minum and Mining,” DOE-ITP, p 52

B.8. Data from Mid-Continent Coal & Coke Co.B.9. Emissions of Greenhouse Gases in the

United States 1987–1992, Energy Infor-mation Agency, Oct 1994, Appendix A

B.10. International Organization for Standardi-zation

Table B.2 Fuels consumed worldwide (kWh/yr) for U.S. aluminum processing, excluding electricityand coke feedstock energy

Anode with Primary Secondary Hot Cold ShapeMining Refining feedstock Electrolysis casting casting rolling rolling Extrusion casting Total

No process (secondary energy) inputs for fuelskWh/yr 8.13�108 1.79�1010 1.23�109 4.59�108 1.98�109 7.13�109 8.62�108 6.93�108 2.20�109 5.84�109 3.91�1010

With process (secondary energy) inputs for fuelskWh/yr 8.45�108 1.84�1010 1.27�109 4.72�108 2.04�109 7.37�109 8.87�108 7.17�108 2.26�109 6.01�109 4.02�1010

Percent process (secondary energy) fuel contribution4 3 3 3 3 3 3 3 3 3 3

Table B.3 Impact of electric tacit conversion factors on kWh/yr consumed worldwide for U.S. aluminum production

Anode with Primary Secondary Hot Cold ShapeUnit Mining Refining feedstock Electrolysis casting casting rolling rolling Extrusion casting Total

Average U.S. aluminum processing facilities7570 Btu of energy required to produce 1 kWh of on-site electrical power for an average U.S. aluminum smelting, anode, and primary

casting facility9890 Btu of energy required to produce 1 kWh of on-site electrical power for all other U.S. aluminum facilities

kWh/yr 8.60�108 1.99�1010 1.50�1010 1.13�1011 3.55�109 8.37�109 2.75�109 3.17�109 2.75�109 6.04�109 1.76�1011

Percent increase relative to facilities with 100% hydroelectric energy resource

% 4 8 4 188 42 12 83 106 16 3 86

U.S. aluminum processing facilities with 100% electrical power generated by hydroelectric utilities3412 Btu of energy required to produce 1 kWh of on-site electrical power from hydroelectric resources for all U.S. aluminum facilities

kWh/yr 8.26�108 1.84�1010 1.44�1010 3.94�1010 2.50�109 7.47�109 1.50�109 1.54 �109 2.37�109 5.85�109 9.43�1010

U.S. aluminum processing facilities with 100% electrical power generated by coal-fired utilities10,300 Btu of energy required to produce 1 kWh of on-site electrical power from coal resources for all U.S. aluminum facilities

kWh/yr 8.62�108 2.01�1010 1.52�1010 1.29�1011 3.77�109 8.51�109 3.01�109 3.51�109 2.82�109 6.04�109 1.93�1011

Percent increase relative to facilities with 100% hydroelectric energy resource

% 4 10 5 228 51 14 100 128 19 3 105

SIGNIFICANT ENERGY is consumed in thegeneration and transmission of electricity. Tacitelectric energy conversion factors include theenergy associated with production, processing,and distribution of the primary energy sourcesused in the production of electricity. Tacit values

vary significantly depending on the source of energy. In the United States, 39.4% of aluminumsmelting capacity uses hydroelectric power.

The electrical power used in primary alu-minum production worldwide is shown in TablesC.2 and C.3.

Table C.1 Electric tacit energy and emission data for fuels used for aluminum productionElectrical U.S. primary aluminum capacity(a) Heat rate Carbon emission energy (2005)(b), coefficient(c),source Metric tons % Btu/kWh Mt/Qbtu

Hydro 1,633,696 39.4 3412 0Coal 2,415,826 58.2 10,303 21.25Oil 8245 0.2 10,090 19.08Natural gas 31,120 0.8 7937 12.50Nuclear 60,112 1.4 10,420 0Total 4,149,000 100.0

Weighted averages based on aluminum capacity 7570 12.51Average U.S. grid(d) 9890 13.56

(a) The distribution of electrical sources for U.S. primary aluminum capacity is obtained by subtracting the Canadian capacity of 2.827 million metric tons (2002),which is 100% hydroelectric, from the North American totals presented in Table C.2.

(b) Heat rate values are derived using the primary energy values in Appendix B, Table B.1, and the net generation and consumption values listed in the Annual EnergyReview 2005, Energy Information Administration (see Table C.4).

(c) Table C.4 of this appendix.(d) 9890 tacit Btu/kWh is used here for all electricity consumed in U.S. aluminum processing operations. This value overstates the actual aluminum-related tacit values

(which is 7570 Btu/kWh), but it is more useful for comparing the aluminum industry to other U.S. manufacturing operations.

APPENDIX C

Hydroelectric Distribution and Electrical Energy Values

Table C.2 Energy sources of electrical power in 2002 (electrical power used in gigawatt hours)North America Latin Grand

Africa, America, Asia, Europe, Oceania, Total, total,Electric energy source GWh GWh % GWh GWh GWh GWh GWh %

Hydro 6444 50,312 64 32,207 3030 29,969 7380 129,342 50Coal 13,443 27,248 35 0 10,080 15,773 24,120 90,664 35Oil 0 93 0 0 109 1279 0 1481 1Natural gas 38 351 0 1343 16,336 6199 0 24,267 9Nuclear 189 678 1 0 0 12,672 0 13,539 5Total 20,114 78,682 100 33,550 29,555 65,892 31,500 259,293 100Source: International Aluminum Institute




Significant energy is consumed in the genera-tion and transmission of electricity. Tacit electricenergy conversion factors (Btu/kWh) include theenergy associated with production, processing,and distribution of the primary energy sourcesused in the production of electricity. Tacit valuesvary significantly depending on the source of energy to produce electric power.

REFERENCES

C.1. Annual Energy Review 2005, Table 8.2aElectricity net generation, p 228. Note:

Net generation includes transmissionlosses.

C.2. Annual Energy Review 2005, Table 8.5aConsumption of combustible fuels forelectricity generation. p 242

C.3. Annual Energy Outlook 2004, EnergyInformation Agency, Jan 2004, p 262

C.4. Annual Energy Review 2005, Table 2.1fElectric power sector energy consump-tion, p 43

C.5. Data from Mid-Continent Coal & CokeCo. for green petroleum coke

C.6. Data from Nuclear Energy Institute

Table C.3 Sources of supply of electrical power in 2002 (electrical power used in gigawatt hours)North America Latin Grand

Electric source Africa, America, Asia, Europe, Oceania, Total, total,of supply GWh GWh % GWh GWh GWh GWh GWh %

Self-generated 0 27,702 35 4386 28,228 9193 1244 70,753 27

Purchased—grid 20,114 39,015 50 28,863 1327 56,698 23,237 169,254 65

Purchased—other 0 11,965 15 301 0 1 7019 19,286 7

Total 20,114 78,682 100 33,550 29,555 65,892 31,500 259,293 100

Self-generated 0 351 . . . 253 2662 1 0 3267 . . .other purposes

Source: International Aluminum Institute

Table C.4 Average U.S. grid connection tacit energyU.S. electricity netgeneration in 2005 Energy Btu/kWh

(Ref C.1) Units consumed consumed, net Source Billion kWh % (Ref C.2) Heat content 109 Btu generation(a)

Coal/ 2014.2 50 1051.2 � 106 short tons 20,620,000 Btu/ton (Ref C.3) 20,752,000 (Ref C.4) 10,303Petroleum . . . . . . 21.9 � 106 bbl distillate 5,825,000 Btu/gal (Ref C.3) 127,626(b) . . .

. . . . . . 146.8 � 106 bbl residual 6,287,000 Btu/bbl (Ref C.3) 923,126(b) . . .

. . . . . . 3.7 � 106 bbl other liquids 5,670,000 Btu/bbl (Ref C.3)(c) 20,786(b) . . .

. . . . . . 8.5 � 106 short tons coke 14,200 Btu/lb (Ref C.5) 241,684(b) . . .Subtotal 121.9 3 1,230,000 (Ref C.4) 10,090Natural gas 751.5 19 6466 � 109 ft3 1019 Btu/ft3 (Ref C.3) 5,965,000 (Ref C.4) 7937Nuclear 780.5 19 10,000 Btu/kWh (Ref C.6) 8,133,000 (Ref C.4) 10,420Hydroelectric 265.1 7 10,117 Btu/kWh(d) 2,682,000 (Ref C.4) 10,117Renewable 92.1 2 957 � 1012 Btu 1,004,000 (Ref C.4) 10,901Other 3.7 0 27 � 1012 Btu 22,000 (Ref C.4) 5946Totals 4029.0 100 39,851,000 (Ref C.4) 9891

Note: Transmission and distribution losses = 1.31 of 13.78 quad Btu generated. Therefore 13.78/12.58, or 1.1051 times delivered-to equals generation (Annual EnergyReview 2005, p 223). (a) 109 Btu ÷ Billion kWh. (b) Heat content � units consumed. (c) Assumes diesel fuel is the majority of other fuels. (d) Assumes that a substitutefor hydropower is equivalent to the average of the other sources of electricity

APPENDIX D

Emission Data and Calculations

Table D.1 presents the carbon dioxide equiva-lent emission values for the fuels and materialsassociated with the production of aluminummetal and aluminum products.

Table D.2 uses the units of energy input fromAppendix E, Table E.1, and calculates the CO2efor each energy input based on the CO2e valuesfor fuels presented in Table D.1.

Table D.3 uses the units of energy input fromAppendix E, Table E.2, and calculates the car-

bon dioxide equivalent (CDE) (CO2e) emissionsfor each energy input base on the CDE valuesfor fuels presented in Table D.1.

Table D.4 provides comparison of the ore-to-metal carbon dioxide equivalent emis-sions for a modern Hall-Heroult cell to thosefor improved Hall-Heroult and alternativetechnologies.




Tabl

e D

.1C

arbo

n di

oxid

e eq

uiva

lent

em

issi

on c

oeff

icie

nts

for

fuel

s as

soci

ated

wit

h al

umin

um p

rodu

ctio

nK

ilogr

am c

arbo

nA

ppen

dix

EB

tu p

erC

arbo

n em

issi

onC

arbo

n em

issi

onP

erce

ntca

rbon

AP

I gr

avit

yD

ensi

ty,(

Ref

D.1

)’M

illio

n B

tu/b

bldi

oxid

e eq

uiva

lent

Sour

cein

put

unit

inpu

t un

itco

effi

cien

t(a)

,Mt/

Qbt

uco

effi

cien

t so

urce

(Ref

D.1

)(R

ef D

.1)

lb/g

al(R

ef D

.1)

per

inpu

t un

it

Fue

lFu

el o

il,he

avy

(#6)

kg..

.21

.49

Ref

D.1

85.7

17.0

...

6.28

7..

.Fu

el o

il,m

ediu

mkg

44,7

0020

.72

(b)

...

...

...

...

3,40

Fuel

oil,

light

(#2

)kg

48,1

3019

.95

(c)

86.3

33.9

7.06

45.

825

3.52

Die

sel

L37

,040

19.9

5R

ef D

.186

.335

.57.

064

5.82

52.

71K

eros

ene

L37

,040

19.7

2R

ef D

.286

.141

.4..

.5.

670

2.68

Gas

olin

eL

35,8

7019

.34

Ref

D.2

86.6

58.6

...

5.25

32.

54N

atur

al g

asm

337

,330

14.4

7R

ef D

.2..

...

...

...

.1.

98B

itum

inou

s/su

bbitu

min

ous

kg24

,630

25.8

1R

ef D

.1..

.N

/A..

...

.2.

33C

alci

ned

coke

kg34

,010

27.8

5(d

)..

...

...

...

.3.

47Pi

tch

kg38

,190

20.6

2(e

)85

.825

.6..

...

.2.

89G

reen

cok

ekg

31,8

1027

.85

Ref

D.2

92.3

N/A

9.54

36.

024

3.25

Prop

ane

L24

,050

17.2

0R

ef D

.1..

...

...

...

.1.

52C

oal

kg22

,710

25.7

6R

ef D

.2..

.N

/A..

...

.2.

15

Ele

ctri

city

Hyd

roel

ectr

ickW

h34

120.

0R

ef D

.2..

...

...

...

.0.

00A

vera

ge U

.S. e

lect

ric

kWh

9891

13.5

6R

ef D

.2..

...

...

...

.4.

92 �

10�

1

Prim

ary

Al e

lect

ric

kWh

7620

12.5

1(f

)..

...

...

...

.3.

49 �

10�

1

Coa

l-fir

ed e

lect

ric

kWh

10,3

0321

.25

Ref

D.2

...

...

...

...

8.03

� 1

0�1

(a)

Mt/Q

btu,

mill

ion

met

ric

tons

per

qua

drill

ion

Btu

(10

6m

etri

c to

ns/1

015B

tu).

(b)

Med

ium

fue

l oil

valu

es a

re th

e av

erag

e of

the

light

and

hea

vy f

uel o

il va

lues

.(c

) D

iese

l and

ligh

t fue

l oil

is a

ssum

ed to

hav

e th

e sa

me

carb

on c

oeff

icie

nt in

Ref

D.1

.(d

) G

reen

cok

e an

d ca

lcin

ed c

oke

are

assu

med

to h

ave

the

sam

e ca

rbon

coe

ffic

ient

.(e

) Pi

tch

is a

ssum

ed to

hav

e th

e sa

me

carb

on c

oeff

icie

nt a

s as

phal

t,re

port

ed in

Ref

D.1

.(f

) Pr

imar

y al

umin

um c

arbo

n di

oxid

e eq

uiva

lent

s ar

e ba

sed

on th

e in

dust

ry’s

mix

of

fuel

s (A

ppen

dix

C,T

able

C.1

).

Appendix D: Emission Data and Calculations / 233

Tabl

e D

.2C

arbo

n di

oxid

e eq

uiva

lent

(C

O2e

) em

issi

ons

asso

ciat

ed w

ith

prim

ary

alum

inum

pro

duct

ion

Inpu

t/ou

tput

Min

ing

Refi

ning

Ano

deE

lect

roly

sis

Pri

mar

y al

umin

um

Ric

h so

ilkg

1150

...

...

...

Cum

ulat

ive

CO

2em

issi

ons

Tota

l CO

2em

issi

ons

Bau

xite

kg10

0026

40..

...

.fr

om m

anuf

actu

ring

the

from

man

ufac

turi

ng th

eA

lum

ina

kg..

.10

00..

.19

30ra

w m

ater

ials

req

uire

d to

raw

mat

eria

ls a

nd

Cal

cine

d co

kekg

...

...

820

...

prod

uce

alum

inum

elec

trol

ytic

red

uctio

n re

quir

ed to

pro

duce

al

umin

umPi

tch

kg..

...

.23

1..

.G

reen

cok

ekg

...

...

85..

.A

node

kg..

...

.10

0044

6Fl

uori

des

kg..

...

...

.18

.6A

lum

inum

kg..

...

...

.10

00R

atio

to a

lum

inum

Bau

xite

5.1

0A

limin

a 1.

93

Ano

de 0

.45

Alu

min

um 1

.00

Ene

rgy

inpu

ts p

erkg

equ

ival

ent

kg e

quiv

alen

tkg

equ

ival

ent

kg e

quiv

alen

t10

00 k

gU

nits

of C

O2

Uni

tsof

CO

2U

nits

of C

O2

Uni

ts/k

g A

llof

CO

2U

nits

/kg

Al

kg C

O2/

kg A

lU

nits

/kg

Al

kg C

O2/

kg A

l

Fuel

oil,

med

ium

kg1.

160

3.94

93.2

003.

17�

102

3.79

01.

29�

10..

.18

7.48

6.37

�10

218

7.47

76.

37�

102

Fuel

oil,

light

(#2

)kg

...

...

0.81

42.

874.

310

1.52

� 1

00.

361.

284.

673

1.65

�10

Die

sel

L4.

370

1.18

� 1

01.

670

4.52

0.11

02.

98 �

10-1

1.84

04.

9925

.54

6.92

� 1

027

.378

7.42

� 1

0K

eros

ene

L..

...

...

...

...

.0.

00..

.0.

00G

asol

ine

L0.

274

6.97

� 1

0�1

0.02

46.

05 �

10�

20.

046

1.17

� 1

0�1

0.28

57.

25 �

10�

11.

463.

721.

748

4.45

Nat

ural

gas

m3

...

...

225

4.46

� 1

0297

.11.

92 �

102

7.63

1.51

� 1

047

7.56

9.46

� 1

0248

5.18

79.

61 �

102

Bitu

min

ous/

kg..

.8.

592.

00 �

10

...

...

...

...

16.5

83.

86 �

10

16.5

793.

86 �

10

subb

itum

inou

s C

alci

ned

coke

kg

...

...

0.00

134

4.65

� 1

0�3

820

2.85

� 1

03..

...

.36

5.72

1.27

� 1

0336

5.72

31.

27 �

103

Pitc

h kg

...

...

...

...

231

6.67

� 1

02..

...

.10

3.03

2.97

� 1

0210

3.02

62.

97 �

102

Gre

en c

oke

kg..

...

...

...

.85

2.76

� 1

02..

...

.37

.91

1.23

� 1

0237

.910

1.23

� 1

02

Prop

ane

L..

...

...

...

.0.

136

2.06

� 1

0-12.

724.

130.

069.

20 �

10-2

2.78

14.

22C

oal

kg..

...

...

...

...

...

...

...

...

...

...

...

.E

lect

ric

kWh

0.4

1.97

� 1

0�1

109

5.36

� 1

026

61.

31 �

102

15,4

007.

57 �

103

1.57

� 1

041.

63 �

102

31,0

66.0

07.

74 �

103

Tota

l per

100

0 kg

Tota

l kg

CO

2e1,

67 �

10

8.40

� 1

024.

13 �

103

7.61

� 1

033.

55 �

103

1.56

� 1

04

Man

ufac

turi

ng

1.67

� 1

08.

40 �

102

3.39

� 1

027.

61 �

103

1.86

� 1

038.

33 �

103

port

ion

kg C

O2e

Feed

stoc

k po

rtio

n C

O2e

...

...

3.79

� 1

03..

.1.

69 �

103

7.32

�10

3

Vol

atile

orga

nic

com

poun

d4.

99 �

10�

1

kg C

O2e

/kg

Al

0.08

51.

622

1.84

7.61

3.55

15.6

5To

tal m

ater

ial p

rodu

ctio

n in

the

U.S

. for

alu

min

um m

anuf

actu

ring

met

ric

tons

04,

419,

000

1,12

8,00

02,

530,

000

...

...

Tota

l U.S

. pro

duct

ion

carb

on d

ioxi

de e

mis

sion

sm

etri

c to

ns C

O2e

03,

713,

552

4,65

9,14

619

,263

,471

9,24

5,13

428

,508

,605

Tota

l mat

eria

l pro

duct

ion

wor

ldw

ide

for

use

in t

he U

.S. f

or a

lum

inum

man

ufac

turi

ngm

etri

c to

ns12

,891

,000

4,88

3,00

01,

128,

000

2,53

0,00

0..

...

...

.To

tal w

orld

wid

e pr

oduc

tion

em

issi

ons

Met

ric

tons

CO

2e21

4,93

84,

103,

479

4,65

9,14

619

,263

,471

11,0

92,8

4430

,356

,315

kg C

O2e

/ kg

AL

...

...

...

...

4.38

12.0

0


Tabl

e D

.3C

arbo

n di

oxid

e eq

uiva

lent

(C

O2e

) em

issi

ons

asso

ciat

ed w

ith

alum

inum

pro

duct

ion

Pri

mar

y in

got

Seco

ndar

y in

got

Inpu

t/ou

tput

cast

ing

cast

ing

Hot

rol

ling

Col

d ro

lling

Ext

rusi

onSh

ape

cast

ing

Ele

ctro

lysi

s m

etal

kg10

20...

...

...

...

...

Allo

y ad

ditiv

eskg

...

21..

...

...

...

.Sc

rap

kg..

.67

6..

...

...

...

.Pr

imar

y in

got

kg..

.34

649

847

455

9..

.Se

cond

ary

ingo

tkg

...

Yie

ld..

.Y

ield

723

Yie

ld72

2Y

ield

885

Yie

ld22

00Y

ield

Prod

uct

kg10

0098

%10

0096

%10

0082

%10

0084

%10

0069

%10

0045

%

Ene

rgy

inpu

ts

kg e

quiv

alen

tkg

equ

ival

ent

kg e

quiv

alen

tkg

equ

ival

ent

kg e

quiv

alen

tkg

equ

ival

ent

per

1000

kg

Uni

tsof

CO

2U

nits

of C

O2

Uni

tsof

CO

2U

nits

of C

O2

Uni

tsof

CO

2U

nits

of C

O2

Fuel

oil,

med

ium

kg..

...

...

...

...

...

...

.Fu

el o

il,lig

ht (

#2)

kg17

.461

42.3

149

...

...

0.00

20

...

...

Die

sel

L0.

184

...

31.4

6485

0.04

00.

109

0.03

80

0.41

71

...

...

Ker

osen

eL

...

...

0.00

70

...

0.02

80

...

...

Gas

olin

eL

0.07

50

12.7

5732

0.00

70

2.15

05

0.04

40

...

...

Nat

ural

gas

m3

51.8

103

126.

025

033

.466

24.8

4910

3.0

204

240.

047

5B

itum

inou

s/kg

...

...

...

...

...

...

...

...

...

...

...

...

subb

itum

inou

sC

alci

ned

coke

kg..

...

...

...

...

...

...

...

...

...

...

...

.Pi

tch

kg..

...

...

...

...

...

...

...

...

...

...

...

.G

reen

cok

ekg

...

...

...

...

...

...

...

...

...

...

...

...

Prop

ane

L0.

798

11.

941

30.

034

00.

238

04.

460

7..

...

.C

oal

kg..

...

...

...

....

...

.....

.11

.200

24..

...

.E

lect

ric

kWh

211

103.

7711

5.00

056

.56

265

130.

3334

917

1.64

9345

.74

41.

98To

tal p

er 1

000

kgC

O2e

2.69

� 1

025.

76 �

102

1.97

� 1

022.

27 �

102

2.82

� 1

024.

77 �

102

Tota

l mat

eria

l pro

duct

ion

in t

he U

.S. f

or a

lum

inum

man

ufac

turi

ngM

etri

c to

ns2,

480,

400

2,99

0,00

02,

421,

300

2,42

1,30

01,

826,

000

2,28

9,00

0To

tal U

.S. p

rodu

ctio

n ca

rbon

dio

xide

em

issi

ons

Met

ric

tons

CO

2e66

7,29

71,

721,

552

476,

218

548,

896

514,

666

1,09

2,60

2K

ilogr

ams

of c

arbo

n di

oxid

e eq

uiva

lent

per

kilo

gram

of

prod

uct

kg/k

g A

l12

.27

0.58

0.20

0.23

0.28

0.48

Appendix D: Emission Data and Calculations / 235

Tabl

e D

.4

Car

bon

diox

ide

equi

vale

nt (

CO

2e)

emis

sion

s as

soci

ated

wit

h ne

w a

lum

inum

pro

duct

ion

tech

nolo

gies

Chl

orid

e re

duct

ion

of k

aolin

ite

Typ

ical

mod

ern

Hal

l-H

erou

lt

Car

both

erm

ic r

educ

tion

cl

ays,

elec

trol

ysis

cur

rent

ce

ll op

erat

ing

at 9

5% c

urre

nt

Iner

t an

ode

oper

atin

g at

wit

h a

reac

tion

ef

fici

ency

95%

,rea

ctio

n ef

fici

ency

T

ypic

al m

oder

n H

all-

Her

oult

effi

cien

cy,

retr

ofitt

ed w

ith

a sl

oped

95%

cur

rent

eff

icie

ncy

wit

h ef

fici

ency

of

95%

and

a

95%

,and

hea

ting

eff

icie

ncie

s 85

%.

cell

oper

atin

g at

95%

and

wet

ted

cath

ode

surf

ace,

oxyg

en p

olar

izat

ion

diff

eren

ces

and

furn

ace

effi

cien

cy o

f 85

%.

On-

site

ele

ctri

c fu

rnac

e an

d cu

rren

t ef

fici

ency

and

wit

hal

umin

um s

ump,

and

a re

duce

d A

CD

= 2

.0 c

m (

0.8

in.)

usi

ngE

lect

ric

furn

ace

oper

atin

gel

ectr

olys

is c

ell o

pera

ting

on

the

an A

CD

= 4

.5 c

m (

1.8

in.)

AC

D (

AC

D =

2.0

cm

or

0.8

in.)

wet

ted

cath

ode

tech

nolo

gyon

the

ave

rage

U.S

. gri

dav

erag

e U

.S. g

rid

Min

eral

mat

eria

lkW

h/kg

Al

kgC

O2/

kg A

lkW

h/kg

Al

kgC

O2/

kg A

lkW

h/kg

Al

kgC

O2/

kg A

lkW

h/kg

Al

kgC

O2/

kg A

lkg

CO

2/kg

Al

kWh/

kg A

l

Taci

t ene

rgy

requ

ired

8.21

1.71

8.21

1.71

8.21

1.71

8.21

1.71

8.81

1.83

On-

site

rea

ctio

n en

ergy

req

uire

men

tsR

eact

ion

ther

mal

...

...

...

...

...

...

7.71

3.79

–1.9

0–0

.93

Furn

ace

loss

es..

...

...

...

...

...

.1.

360.

670.

400.

20R

eact

ion

elec

trol

ysis

3.76

1.85

3.76

1.85

6.90

3.39

...

...

6.48

3.19

Cel

l ohm

ic10

.67

5.25

7.62

3.75

6.20

3.05

...

...

2.93

1.44

Tota

l rea

ctio

n en

ergy

14.4

37.

1011

.38

5.60

13.1

16.

459.

074.

467.

913.

89

Ano

de-r

elat

ed e

mis

sion

sA

node

man

ufac

turi

ng0.

610.

150.

610.

150.

760.

27..

...

...

...

.A

node

use

Ano

de r

eact

ion

0.33

kg

1.22

...

1.22

...

...

...

...

...

...

Ano

de e

xces

s (a

ir b

urni

ng)

0.11

kg

0.30

...

0.30

...

...

...

...

...

...

Tota

l ano

de..

.1.

68..

.1.

68..

...

...

...

...

...

.

Pro

cess

-rel

ated

em

issi

ons

Car

bon

reac

tant

...

...

...

...

...

...

0.66

kg2.

450.

89 k

g3.

27Pe

rfluo

roca

rbon

...

2.20

...

0.55

...

0.55

...

0.00

...

0.00

Tota

l CO

2em

issi

ons

Min

eral

mat

eria

l..

.1.

71..

.1.

71..

.1.

71..

.1.

71..

.1.

83R

eact

ions

...

7.10

...

5.60

...

6.45

...

4.46

...

3.89

Car

bon

...

1.68

...

1.68

...

0.00

...

2.45

...

3.27

Proc

ess

...

2.20

...

0.55

...

0.55

...

0.00

...

0.00

Tota

l..

.12

.68

...

9.53

...

8.70

...

8.61

...

8.99

AC

D,a

node

-cat

hode

dis

tanc

e


REFERENCES

D.1. Emissions of Greenhouse Gases in theUnited States 1987–1992, Energy Infor-mation Agency, Oct 1994, Appendix A

D.2. Emission of Greenhouse Gases in theUnited States 2000, Energy InformationAgency, Nov 2001, p 140

THE ENERGY INPUT UNITS and raw mate-rial quantities used in Tables E.1 and E.2 arefrom the Life Cycle Inventory Report for theNorth American Aluminum Industry, publishedby the Aluminum Association in November1998. The report is the result of extensive sur-veying and contains the best and most com-plete industry performance information of anyrecent study. The report is consistent with ISOprocedures and was favorably peer-reviewedby groups outside the industry. The AluminumAssociation has provided permission to usethese data.

The conversion of input units to commonenergy units are based on the values presentedin Appendix B, Table B.1.

Table E.3 lists the theoretical minimum energyrequirements to produce raw materials andaluminum products. It also shows the processefficiency for each operation. Mining, refining,anode manufacturing, and electrolysis mini-mum energy values are based on the net chemi-cal changes that result from these processes.Primary casting, secondary casting, and shapecasting minimum energy values are based onthe energy required to produce molten purealuminum at 775 °C (1425 °F). The minimumenergy requirements for rolling and extrusionoperations are estimated from their yield andon-site energy consumption values in TableE.2, and from an overall assumed electric andhydraulic system efficiencies of 75% and ther-mal furnace and heating efficiencies of 50%.

Table E.4 lists U.S. production quantitiesand energy and tacit energy consumption asso-ciated with producing aluminum productswithin the United States. The United Statesdoes not consume energy to produce metallur-

gical bauxite and consumes only 56% of theenergy required for alumina. This report distin-guishes between the worldwide energy valuesand the U.S. values in order to measure the im-pact of market or process changes to the energydemands within the United States. Table E.9lists the worldwide values. The theoreticalmagnitude of the potential U.S. energy savingscan be measured by subtracting the theoreticalenergy requirements from the actual energyconsumption numbers.

Table E.5 shows the total energy values con-sumed by the U.S. aluminum industry and theenergy savings possible if:

• It was possible to reduce energy consump-tion to the theoretical limit

• Energy were reduced by 30%• Electrolysis electrical energy were reduced

from 15 to 11 kWh/kg of aluminum

Table E.6 shows 94% reduction in energyconsumption associated with secondary metalproduction. Recycling aluminum essentially re-captures all the energy associated with mining,refining, and smelting.

Table E.7 shows the electric energy consump-tion of the U.S. aluminum industry and com-pares it to the total electric energy produced inthe United States.

Electrolysis and process heating are the twomost energy-intensive process areas of the alu-minum industry. Electrolysis accounts for 43%of all on-site and 66% of all tacit energy use.Process heating accounts for 27% of all on-siteand 14% of all tacit energy use in the industry.

Table E.9 lists the worldwide productionquantities, energy, and tacit energy consumptionassociated with producing aluminum products

APPENDIX E

U.S. Energy Use by Aluminum Processing Area



238 / Aluminum Recycling and Processing for Energy Conservation and SustainabilityTa

ble

E.1

Mat

eria

ls a

nd e

nerg

y as

soci

ated

wit

h pr

imar

y m

etal

ele

ctro

lysi

sIn

put/

outp

utM

inin

gR

efini

ngA

node

Ele

ctro

lysi

sP

rim

ary

alum

inum

Ric

h so

ilkg

1150

.....

...

.C

umul

ativ

e en

ergy

Tota

l ene

rgy

requ

irem

ent

Bau

xite

kg10

0026

40..

...

.re

quir

emen

t for

fo

r m

anuf

actu

ring

the

raw

Alu

min

akg

...

1000

...

1930

man

ufac

turi

ng th

e ra

wm

ater

ials

and

ele

ctro

lytic

Cal

cine

d co

kekg

...

...

820

...

mat

eria

ls r

equi

red

tore

duct

ion

requ

ired

toPi

tch

kg..

...

.23

1..

.pr

oduc

e al

umin

umpr

oduc

e al

umin

umG

reen

cok

ekg

...

...

85..

.A

node

kg..

...

.10

0044

6Fl

uori

des

kg..

...

...

.18

.6A

lum

inum

kg..

...

...

.10

00R

atio

to a

lum

inum

Bau

xite

5.1

0A

lum

ina

1.93

Ano

de 0

.45

Alu

min

um 1

.00

Ene

rgy

inpu

ts p

er 1

000

kg o

f m

ater

ial p

rodu

ced

Uni

tsB

tuU

nits

Btu

Uni

tsB

tuU

nits

Btu

Uni

ts/k

g A

lB

tu/k

g A

lU

nits

/kg

Al

Btu

/kg

Al

Fuel

oil,

med

ium

kg1.

160

50,3

2193

.200

4,04

3,05

73.

790

164,

412

...

...

187.

488,

132,

824

187.

477

8,13

2,82

4Fu

el o

il,lig

htkg

...

...

...

...

0.81

438

,106

4.31

020

1,76

60.

3616

,995

4.67

321

8,76

1D

iese

lL

4.37

015

5,86

51.

670

59,5

640.

110

3923

1.84

065

,627

25.5

491

0,87

327

.378

976,

501

Ker

osen

eL

...

...

...

...

...

...

...

...

...

...

...

...

Gas

olin

eL

0.27

489

590.

024

778

0.04

615

040.

285

9319

1.46

47,8

221.

748

57,1

41N

atur

al g

asm

3..

...

.22

58,

160,

344

97.1

3,52

1,64

27.

6327

6,72

647

7.56

17,3

20,1

1748

5.18

717

,596

,843

Bitu

min

ous/

kg

...

...

8.59

210,

396

...

...

...

...

16.5

840

6,06

416

.579

406,

064

subb

itum

inou

sC

oke

kg..

...

.0.

0013

445

...

...

...

...

0.00

387

0.00

387

Prop

ane

L..

...

...

...

.0.

136

3271

2.72

65,4

290.

061,

459

2.78

166

,888

Coa

lkg

...

...

...

...

...

...

...

...

...

...

...

...

Ele

ctri

ckW

h0.

413

6510

937

1,90

826

690

7,59

215

,400

52,5

44,8

0033

1.04

1,12

9,52

215

,731

.04

53,6

74,3

22Ta

cit e

lect

ric

kWh

1.2

...

316

...

771

...

44,6

38..

.10

88..

.45

,726

...

Tota

l ene

rgy

inpu

ts (

ener

gy p

er k

g of

mat

eria

l out

put)

Btu

/kg

217

12,8

4646

4053

,164

Not

e:T

he o

n-si

te o

r pr

imar

y el

ectr

ical

ene

rgy

kWh/

kg0.

063.

761.

3615

.58

requ

irem

ent t

o pr

oduc

e al

umin

um f

rom

one

acc

ount

sTa

cit k

Wh/

kg0.

074.

081.

9044

.82

for

66%

of

the

tota

l ene

rgy

requ

irem

ent.

Rat

io a

djus

ted

ener

gy in

puts

(en

ergy

per

kg

of a

lum

inum

)

Btu

/kg

1103

24,7

9320

7053

,164

27,9

6681

,129

kWh/

kg0.

327.

270.

6115

.58

8.20

23.7

8

Taci

t kW

h/kg

0.34

7.87

0.85

44.8

29.

0653

.89

Fee

dsto

ck e

nerg

y pe

r kg

Al (

ener

gy in

here

nt in

fue

ls u

sed

as m

ater

ials

)F

eeds

tock

rat

io a

djus

ted

Btu

/kg

...

...

17,4

14..

...

.98

,544

kWh/

kg..

...

.5.

10..

...

.28

.88

Taci

t kW

h/kg

Tota

l tac

it (T

able

10.

9) (

proc

ess

0.90

+ f

eeds

tock

5.1

0)6.

01..

.14

.22

59.0

5

Appendix E: U.S. Energy Use by Aluminum Processing Area / 239

Tabl

e E.

2M

ater

ials

and

ene

rgy

asso

ciat

ed w

ith

alum

inum

man

ufac

turi

ng o

pera

tion

sP

rim

ary

ingo

t Se

cond

ary

ingo

t H

ot

Inpu

t/ou

tput

cast

ing

cast

ing

rolli

ngC

old

rolli

ngE

xtru

sion

Shap

e ca

stin

g

Ele

ctro

lysi

s m

etal

kg10

20..

...

...

...

...

.A

lloy

addi

tives

kg..

.21

...

...

...

...

Scra

pkg

...

676

...

...

...

...

Prim

ary

ingo

tkg

...

346

498

474

559

...

Seco

ndar

y in

got

kg..

.Y

ield

...

Yie

ld72

3Y

ield

722

Yie

ld88

5Y

ield

2200

Yie

ldPr

oduc

tkg

1000

98%

1000

96%

1000

82%

1000

84%

1000

69%

1000

45%

Ene

rgy

inpu

ts p

er 1

000

kgU

nits

Btu

Uni

tsB

tuU

nits

Btu

Uni

tsB

tuU

nits

Btu

Uni

tsB

tu

Fuel

oil,

med

ium

kg..

...

...

...

. ..

...

...

...

...

...

...

...

.Fu

el o

il,lig

htkg

17.4

814,

554

42.3

1,98

1,34

6..

...

...

...

.0.

002

109

...

...

Die

sel

L0.

184

6563

31.4

641,

122,

230

0.04

014

410.

038

1362

0.41

714

,873

...

...

Ker

osen

eL

...

...

...

...

0.00

724

9..

...

.0.

028

988

...

...

Gas

olin

eL

0.07

52,

439

12.7

5741

7,11

50.

007

213

2.15

070

,301

0.04

414

42..

...

.N

atur

al g

asm

351

.81,

878,

693

126.

04,

569,

793

33.4

1,21

1,35

824

.889

9,45

110

3.00

03,

735,

624

240.

000

8,70

4,36

7B

itum

inou

s/kg

...

...

...

...

...

...

...

...

...

...

...

...

subb

itum

inou

sC

oke

kg..

...

...

...

...

...

...

...

...

...

...

...

.Pr

opan

eL

0.79

819

,196

1.94

146

,692

0.03

481

80.

238

5725

4.46

010

7,28

5..

...

.A

nthr

acite

kg..

...

...

...

...

...

...

...

.11

.200

252,

829

...

...

Ele

ctri

ckW

h21

171

9,93

211

539

2,38

026

590

4,18

034

91,

190,

788

9331

7,31

64

13,7

50

Tota

l ene

rgy

inpu

ts (

ener

gy p

er k

g of

out

put)

Btu

/kg

3,44

185

3021

1821

6844

3087

18

kWh/

kg1.

012.

500.

620.

641.

302.

56

Taci

t en

ergy

inpu

ts (

ener

gy p

er k

g of

out

put)

Btu

/kg

4887

9548

3870

4462

5145

8999

kWh/

kg1.

432.

801.

131.

311.

512.

64


Tabl

e E.

3Th

eore

tica

l min

imum

ene

rgy

requ

irem

ents

to

prod

uce

raw

mat

eria

ls a

nd a

lum

inum

pr

oduc

tsA

node

wit

h P

rim

ary

Seco

ndar

yH

ot

Col

d Sh

ape

Min

ing

Refi

ning

feed

stoc

kE

lect

roly

sis

cast

ing

cast

ing

rolli

ngro

lling

Ext

rusi

onca

stin

g

kWh/

kg..

.0.

1412

.97

5.99

0.33

0.33

0.31

0.33

0.44

0.33

Rat

ioed

to A

l kW

h/kg

Al

...

0.27

4.40

5.99

0.33

0.33

0.31

0.33

0.44

0.33

Cur

rent

pra

ctic

e on

-site

..

.4

7738

3313

5052

3413

proc

ess

effi

cien

cy,%

Tabl

e E.

4U

.S. e

nerg

y us

e in

the

pro

duct

ion

of d

omes

tic

alum

inum

Ano

de w

ith

Pri

mar

y Se

cond

ary

Hot

C

old

Shap

e To

tal

Min

ing

Refi

ning

feed

stoc

kE

lect

roly

sis

cast

ing

cast

ing

rolli

ngro

lling

Ext

rusi

onca

stin

gen

ergy

Mat

eria

l pro

duct

ion

wit

hin

the

Uni

td S

tate

s

Met

ric

tons

04,

419,

000

1,12

8,00

02,

530,

000

2,48

0,40

02,

990,

000

2,42

1,30

0(a)

2,42

1,30

0(a)

1,82

6,00

0(a)

2,28

9,00

0. .

.

On-

site

ene

rgy

cons

umed

in p

rodu

ctio

n fo

r U

.S. a

lum

inum

indu

stry

kWh/

kg..

.3.

7612

.80

15.5

81.

012.

500.

620.

641.

302.

56R

atio

ed to

Al(

a)kW

h/kg

Al

. . .

7.27

5.71

15.5

81.

012.

500.

620.

641.

302.

56To

tal

kWh/

yr. .

.1.

66�

1010

1.44

�10

103.

94�

1010

2.50

�10

97.

47�

109

1.50

�10

91.

54�

109

2.37

�10

95.

85�

109

9.17

�10

10

Prop

ortio

n of

U.S

. ene

rgy

used

,%. .

.18

1643

38

22

36

100

80%

tota

l pri

mar

y U

.S. p

ropo

rtio

n

Tota

l tac

it e

nerg

y co

nsum

ed in

pro

duct

ion

for

U.S

. alu

min

um in

dust

ry

kWh/

kg. .

.4.

0813

.34

44.8

21.

432.

801.

131.

311.

512.

64. .

.R

atio

ed to

Al(

a)kW

h/kg

Al

. . .

7.87

6.01

44.8

21.

432.

801.

131.

311.

512.

64. .

.To

tal

kWh/

yr. .

.1.

80�

1010

1.50

�10

101.

13�

1011

3.55

�10

98.

37�

109

2.75

�10

93.

17�

109

2.75

�10

96.

04�

109

1.73

�10

11

Prop

ortio

n of

U.S

. ene

rgy

used

,%. .

.10

.48.

765

.52.

14.

81.

61.

81.

63.

510

086

.7%

tota

l pri

mar

y U

.S. p

ropo

rtio

n

The

oret

ical

mag

nitu

de o

f op

port

unit

ies

for

U.S

. ene

rgy

savi

ngs

kWh/

yr. .

.1.

60�

1010

–1.8

2�10

82.

43�

1010

1.68

�10

96.

48�

109

7.56

�10

87.

36�

108

1.57

�10

95.

09�

109

5.64

�10

10

Taci

t kW

h/yr

. . .

1.74

�10

104.

23�

108

9.83

�10

102.

73�

109

7.37

�10

92.

00�

109

2.36

�10

91.

96�

109

5.28

�10

91.

38�

1011

(a)

Dat

a ba

sed

on y

ear

2000

val

ues,

the

last

yea

r nu

mbe

rs s

peci

fic to

U.S

. pro

duct

ion

are

avai

labl

e. A

fter

200

0,re

port

ed n

umbe

rs a

re b

ased

on

Nor

th A

mer

ican

dat

a,no

t spe

cific

ally

U.S

. pro

duct

ion.

Aft

er 2

000,

U.S

. pro

duct

ion

decl

ined

with

the

gene

ral

econ

omy.

It i

s es

timat

ed th

at in

200

5,pr

oduc

tion

rose

to 2

000

leve

ls.


within the United States. These values providea full measure of the energy associated withproducing aluminum and products. The actualenergy consumed within the United States islower because no aluminum metallurgicalbauxite is mined in the United States, and ap-proximately 95% of the alumina required is re-fined in the United States.

Table E.10 divides energy uses in aluminumcasting and semifabrication manufacturing operations into three areas: electric, heating,and miscellaneous fuels. The following percent-ages of energy use are derived from the energyvalues listed in Table E.2. Fuel oils, natural gas,and propane are assumed to be associated with

heating (thermal operations). Miscellaneousfuels include all other fuels, and it is assumedthat these are fuels used in nonheating opera-tions, such as transportation and collection ofmaterials. (Note: Rounding causes some totalsto differ from 100%.)

The total U.S. primary aluminum capacityoperates on over 39% hydroelectric power com-pared to the average U.S. electric grid use of 7%hydroelectric power. It should be noted that, inreality, the industry has lower tacit energy values than reported in the other tablesof this appendix. Using the average U.S. gridconnection is the best way to compare variousindustries.

Table E.5 Total U.S. aluminum industry energy consumption and potential savingsU.S. aluminum industry kWh/yr Quads/yr MW Households bbl crude per year

Total energy use On-site 9.17�1010 0.31 10,500 10,500,000 54,000,000Tacit 1.73�1011 0.59 19,800 19,800,000 101,900,000

Theoretical requirement On-site 3.53�1010 0.12 4100 4,100,000 20,800,000Tacit 3.53�1010 0.12 4100 4,100,000 20,800,000

Energy efficiency On-site 39% . . . . . . . . . . . .Tacit 20% . . . . . . . . . . . .

Theoretical opportunity On-site 5.64�1010 0.19 6400 6,400,000 33,200,000Tacit 1.38�1011 0.47 15,700 15,700,000 81,100,000

Practical goal (vision) On-site 2.75�1010 0.09 3150 3200 16,200,000Tacit 5.19�1010 0.18 5940 6000 30,600,000

Electrolysis savings of 4 kWh/kg On-site 1.01�1010 0.03 1155 1200 6,000,000Tacit 2.25�1010 0.08 2563 2600 13,300,000

Table E.6 Energy impact of recyclingEnergy saved with recycling

Tacit energy values60.5 kWh/kg to produce primary metal ingot2.8 kWh/kg to produce secondary metal ingot5.0% secondary-to-primary energy2,990,000 kg of secondary metal produced in 20031.72�1011 kWh/yr energy saved 20030.59 quads saved per year 200319,700 MW saved 2003


Tabl

e E.

7U

.S. e

lect

ric

on-s

ite

ener

gy c

onsu

mpt

ion

in t

he a

lum

inum

indu

stry

Ano

de w

ith

Pri

mar

y Se

cond

ary

Hot

C

old

Shap

eTo

tal e

lect

ric

Refi

ning

feed

stoc

kE

lect

roly

sis

cast

ing

cast

ing

rolli

ngro

lling

Ext

rusi

onca

stin

gen

ergy

Met

ric

tons

200

04,

419,

000

1,12

8,00

02,

530,

000

2,48

0,40

02,

990,

000

2,42

1,30

02,

421,

300

1,82

6,00

02,

289,

000

kWh/

kg0.

109

0.26

615

.400

0.21

10.

115

0.26

50.

349

0.09

30.

004

kWh/

yr4.

82�

108

3.00

�10

83.

90�

1010

5.23

�10

83.

44�

108

6.42

�10

88.

45�

108

1.70

�10

89.

22�

106

4.23

�10

10

quad

0.14

Tota

l U.S

. Al i

ndus

try

elec

tric

al u

se4.

23�

1010

kWh/

yr

On-

site

4826

MW

4,82

6,00

0ho

useh

olds

Net

U.S

. gen

erat

ion:

3.85

8�10

12

Taci

t49

,900

,000

bbl c

rude

Alu

min

um %

of

U.S

. net

:1.1

%

Sour

ce:E

nerg

y In

form

atio

n A

genc

y,D

OE

,off

ice

of C

oal,

Nuc

lear

,Ele

ctri

c,an

d A

ltern

ate

Fuel

s; h

ttp://

ww

w.e

ia.d

oe.g

ov/c

neaf

/ele

ctri

city

(ac

cess

ed J

uly

2007

)

Tabl

e E.

8 Sm

elti

ng a

nd h

eati

ng fr

acti

ons

of t

otal

U.S

. alu

min

um in

dust

ry e

nerg

y co

nsum

edTo

tal

Per

cent

of

Smel

ting

fra

ctio

n of

tot

al U

.S. a

lum

inum

indu

stry

ene

rgy

cons

umed

indu

stry

indu

stry

Ele

ctro

lysi

sO

n-si

tekW

h/yr

3.94

�10

109.

17�

1010

43Ta

cit

kWh/

yr1.

13�

1011

1.73

�10

1166

Alu

min

um m

etal

pro

cess

hea

ting

/mel

ting

fra

ctio

n of

tot

al U

.S. a

lum

inum

indu

stry

ene

rgy

cons

umed

Rol

ling

Tota

lSh

ape

proc

ess

Tota

l P

erce

nt o

f C

asti

ngH

otC

old

Ext

rusi

onca

stin

ghe

atin

gin

dust

ry

indu

stry

Btu

/kg

Al

9244

1212

905

4096

8704

2.42

�10

4

kWh/

kg A

l2.

710.

360.

271.

202.

557.

08O

n-si

tekW

h/yr

1.48

�10

108.

60�

108

6.42

�10

82.

19�

109

5.84

�10

92.

44�

1010

9.17

�10

1027

Taci

t1.

73�

1011

14

Not

e:T

he ta

cit i

ncre

ase

for

proc

ess

heat

ing

is n

eglig

ible

.


Tabl

e E.

9To

tal w

orld

wid

e pr

oduc

tion

and

ene

rgy

cons

umpt

ion

asso

ciat

ed w

ith

prod

ucin

g al

umin

um in

the

Uni

ted

Stat

esA

node

wit

hP

rim

ary

Seco

ndar

y H

ot

Col

d Sh

ape

Tota

lM

inin

g(a)

Refi

ning

(a)

feed

stoc

k(a)

Ele

ctro

lysi

s(a)

cast

ing

cast

ing

rolli

ngro

lling

Ext

rusi

onca

stin

gen

ergy

Tota

l wor

ldw

ide

prod

ucti

on r

equi

red

to p

rodu

ce a

lum

inum

in t

he U

nite

d St

ates

Met

ric

tons

200

512

,891

,000

4,88

3,00

01,

128,

000

2,53

0,00

02,

480,

400

2,99

0,00

02,

421,

300(

b)2,

421,

300(

b)1,

826,

000(

b)2,

289,

000

...To

tal o

n-si

te e

nerg

y us

e co

nsum

ed in

wor

ldw

ide

prod

ucti

on fo

r U

.S. a

lum

inum

indu

stry

kWh/

kg0.

063.

7612

.80

15.6

1.01

2.50

0.62

0.64

1.30

2.56

...R

atio

ed to

Al(

b) k

Wh/

kg A

l0.

327.

275.

7115

.61.

012.

500.

620.

641.

302.

56...

Tota

lkW

h/yr

8.18

�10

81.

84�

1010

1.44

�10

103.

94�

1010

2.50

�10

97.

47�

109

1.50

�10

91.

54�

109

2.37

�10

95.

85�

109

9.43

�10

10

Perc

ent r

elat

ive

to e

lect

roly

sis(

b)2.

146

.636

.610

0.0

6.3

19.0

3.8

3.9

6.0

14.8

...Pr

opor

tion

of e

nerg

y us

ed,%

0.9

19.5

15.3

41.8

2.7

7.9

1.6

1.6

2.5

6.2

100.

0

80%

tota

l pri

mar

y pr

opor

tion

Tota

l tac

it e

nerg

y co

nsum

ed in

wor

ldw

ide

prod

ucti

on fo

r U

.S. a

lum

inum

indu

stry

kWh/

kg0.

074.

0813

.34

44.8

1.4

2.8

1.1

1.3

1.5

2.6

...R

atio

ed to

Al(

b) k

Wh/

kg A

l0.

347.

876.

0144

.81.

42.

81.

11.

31.

52.

6...

Tota

lkW

h/yr

8.60

�10

81.

99�

1010

1.50

�10

101.

13�

1011

3.55

�10

98.

37�

109

2.75

�10

93.

17�

109

2.75

�10

96.

04�

109

1.76

�10

11

Perc

ent r

elat

ive

to e

lect

roly

sis

0.8

17.6

13.3

100.

03.

17.

42.

42.

82.

45.

3...

Prop

ortio

n of

ene

rgy

used

,%0.

511

.38.

664

.52.

04.

81.

61.

81.

63.

410

0.0

87%

tota

l pri

mar

y pr

opor

tion

(a)

Rat

ios

to p

rim

ary

alum

inum

:Mix

ing,

5.10

; refi

ning

,1.9

3; a

node

with

fee

dsto

ck,0

.45;

ele

ctro

lysi

s,1.

00. (

b) D

ata

base

d on

yea

r 20

00 v

alue

s th

e la

st y

ear

num

bers

spe

cific

to U

.S. p

rodu

ctio

n ar

e av

aila

ble.

Aft

er 2

000,

repo

rted

num

bers

are

bas

ed o

nN

orth

Am

eric

an d

ata,

not s

peci

fical

ly U

.S. p

rodu

ctio

n. A

fter

200

0,U

.S. p

rodu

ctio

n de

clin

ed w

ith th

e ge

nera

l eco

nom

y. I

t is

estim

ated

that

in 2

005,

prod

uctio

n ro

se to

200

0 le

vels

.

Tabl

e E.

10Pe

rcen

t br

eaku

p of

diff

eren

t ty

pes

of e

nerg

y us

ed in

alu

min

um p

rodu

ctio

n op

erat

ions

Pri

mar

y Se

cond

ary

Hot

C

old

Shap

e E

nerg

yca

stin

gca

stin

gro

lling

rolli

ngE

xtru

sion

cast

ing

On-

site

ele

ctri

c,%

215

4355

70

Hea

ting,

%79

7757

4287

100

Mis

cella

neou

s fu

els,

%0

180

36

0Ta

cit e

lect

ric,

%36

962

7272

14H

eatin

g,%

6373

3826

2680

Mis

cella

neou

s fu

els,

%0

170

33

6


Tabl

e E.

11U

.S. e

nerg

y us

e in

the

pro

duct

ion

of d

omes

tic

alum

inum

—ta

cit

elec

tric

ity

alum

inum

pri

mar

y gr

id c

onne

ctio

nA

node

wit

hP

rim

ary

Seco

ndar

y H

ot

Col

d Sh

ape

Tota

lM

inin

gR

efini

ngfe

edst

ock

Ele

ctro

lysi

sca

stin

gca

stin

gro

lling

rolli

ngE

xtru

sion

cast

ing

ener

gy

Mat

eria

l pro

duct

ion

wit

hin

the

Uni

ted

Stat

es

Met

ric

tons

04,

419,

000

1,12

8,00

02,

530,

000

2,48

0,40

02,

990,

000

2,42

1,30

02,

421,

300

1,82

6,00

02,

289,

000

...

On-

site

ene

rgy

cons

umed

in p

rodu

ctio

n fo

r U

.S. a

lum

inum

indu

stry

kWh/

kg..

.3.

7612

.80

15.5

81.

012.

500.

620.

641.

302.

56..

.

Rat

ioed

to A

l kW

h/kg

Al

...

7.27

5.71

15.5

81.

012.

500.

620.

641.

302.

56..

.To

tal

kWh/

yr..

.1.

66 �

1010

1.44

� 1

0103.

94 �

1010

2.50

� 1

097.

47 �

109

1.50

� 1

091.

54�

109

2.37

� 1

095.

85 �

109

9.17

� 1

010

Prop

ortio

n of

U.S

. ene

rgy

used

,%..

.18

1643

38

22

36

100

80%

tota

l pri

mar

y U

.S. p

ropo

rtio

nTo

tal a

vera

ge U

.S. g

rid

taci

t en

ergy

con

sum

ed in

pro

duct

ion

for

U.S

. alu

min

um in

dust

ry

kWh/

kg..

.4.

0815

.19

34.3

51.

432.

801.

131.

311.

512.

64..

.

Rat

ioed

to A

l kW

h/kg

Al

...

7.87

6.77

34.3

51.

432.

801.

131.

311.

512.

64..

.To

tal

kWh/

yr..

.1.

80 �

1010

1.71

� 1

0108.

69 �

1010

3.55

� 1

098.

37 �

109

2.75

� 1

093.

17 �

109

2.75

� 1

096.

04 �

109

1.49

� 1

011

Prop

ortio

n of

U.S

. ene

rgy

used

,%..

.12

.111

.558

.52.

45.

61.

82.

11.

94.

110

0

84.5

% to

tal p

rim

ary

U.S

. pro

port

ion

The

oret

ical

mag

nitu

de o

f op

port

unit

ies

for

U.S

. ene

rgy

savi

ngs

kWh/

yr..

.1.

60 �

1010

–1.8

2 �

108

2.43

� 1

0101.

68 �

109

6.48

� 1

097.

56 �

108

7.36

� 1

081.

57 �

109

5.09

� 1

095.

64 �

1010

Taci

t kW

h/yr

...

1.74

� 1

0104.

82 �

108

7.53

� 1

0102.

73 �

109

7.37

� 1

092.

00 �

109

2.36

� 1

091.

96 �

109

5.28

� 1

091.

15 �

1011

THE THEORETICAL MINIMUM energyrequirement for producing any chemical is deter-mined based on the net chemical reaction used toproduce the product. It is defined as the energyrequired to synthesize a substance in its standardstate from substances also in their standard states.It can be calculated by summing the reactionenergies of the products minus the energies of thereactants.

This report calculates the theoretical mini-mum energy by assuming the reactants enterand the by-products leave the system at roomtemperature and that molten aluminum leavesthe system at 960°C (1760 °F). This report haschosen 960°C (1233 K) as the molten metaltemperature. This value is an approximation ofthe average operating cell temperatures of in-dustrial cells.

Some studies assume that the gases evolvedduring reduction leave the system at the moltenmetal temperature. In these studies, the theoreticalminimum is 2.5 to 3% higher than the numberscalculated here. The additional energy for heatingof the emissions is shown in the individual tables.Theoretically, it is possible to capture all the en-ergy associated with these gaseous emissions.However, there is currently no available economicmeans of recovering this energy. In practice, theemission gas stream is diluted with air in the cellsto lower the temperature that the cell hoods andducts are exposed to. The emission gas is col-lected from hundreds of hooded pots and treatedbefore release to the atmosphere. Only a very

small portion of the heat is actually absorbed andreturned to the system.

The minimum theoretical energy requirementfor aluminum production requires the evalua-tion of three energy factors: the energy requiredto drive the reduction reaction forward, the en-ergy required to maintain the system at constantpressure and temperature, and the energy re-quired to change the temperature of the reactantand/or product. The thermodynamics and chem-ical equilibrium of reactions are described bythe equation ΔG = ΔH–TΔS, and the numericvalues are given in Table F.6. The energy re-quired to drive the reaction forward is the Gibbsfree energy value (ΔG). The energy required tomaintain system equilibrium is the differencebetween the heat of reaction (ΔH) and the Gibbsfree energy value (ΔG), which equals the en-tropy term (TΔS). Because the Gibbs free en-ergy requirement is less than the heat of reac-tion for alumina reduction, additional energymust be added to the system to maintain the sys-tem temperature. Otherwise, the system wouldcool as the reaction proceeds. (Reduction cellsoperate at atmospheric conditions, and no pres-sure change results during reduction.) Thenumeric values for G, H, and S are given inTable F.6.

A detailed discussion of the theoretical re-quirements is made in “Current and Energy Ef-ficiency of Hall-Héroult Cells—Past, Presentand Future,” by Warren Haupin and WilliamFrank, published in Light Metal Age, June 2002.

APPENDIX F

Theoretical Energy Data and Calculations




Table F.1 Theoretical minimum energy for Hall-Heroult carbon anode system(Products-

Reactants Products reactants)temp. temp. Reaction thermodynamics at 298 K cal/g mole Al

(2 Al2O3 + 3 C) to (4 Al + 3 CO2) = Net

2 Al2O3 4 Al ΔG –756,358 0 0 –282,779 118,39525 °C 960 °C ΔH –801,000 0 0 –282,155 129,711

Carbon ΔS 24.3 4.1 27.1 153.2 11,311

anode Process theoretical minimum energy requirements kWh/kg3C cell 3CO2 Electrolytic work requirement (ΔG) 5.1125 °C 25 °C Thermal energy for temperature maintenance (ΔH–ΔG) 0.49

Thermal energy for Al at 960 °C 0.39Theoretical minimum 5.99

Thermal energy for CO2 at 960 °C 0.17

Note: Thermodynamic values for G, H, and S are from Table F.6. Heat capacity data are from Table F.9 and Appendix G.

Table F.2 Theoretical minimum energy for Hall-Heroult inert anode system(products-


(Al2O3) to (2 Al + 1.5 O2) = NetΔG –378,179 0 0 189,089

2 Al ΔH –400,500 0 0 200,250960 °C ΔS 12 14 74 11,155

InertAl2O3 Process theoretical minimum energy requirements kWh/kg

25 °C anodecell Electrolytic work requirement (ΔG) 8.16

1.5O2 Thermal energy for temperature maintenance (ΔH –ΔG) 0.4825 °C Thermal energy for Al at 960 °C 0.39

Theoretical minimum 9.03Thermal energy for O2 at 960 °C 0.27


Table F.3 Theoretical minimum energy for carbothermic reduction(Products-


(Al2O3 + 3 C) to (2 Al + 3 CO) = NetAl2O3 2 Al ΔG –378,179 0 0 –98,424 139,878

25 °C 960 °C ΔH –400,500 0 0 –79,247 160,626

Carbo- ΔS 12.2 4.1 13.5 141.9 21,347

thermic

Process theoretical minimum energy requirements kWh/kgreactor Work requirement (ΔG) 6.03

3C 3CO Thermal energy for temperature maintenance (ΔH–ΔG) 0.9025 °C 25 °C Thermal energy for Al at 960°C 0.39

Theoretical minimum 7.32Thermal energy for CO at 960°C 0.19


Appendix F: Theoretical Energy Data and Calculations / 247

Table F.4 Theoretical minimum energy for reduction of aluminum chloride(products-


(2 AlCl3) to (2 Al + 3 Cl2)= NetΔG –300,574 0 0 150,287

2 Al ΔH –336,616 0 0 168,308Inert 960 °C ΔS 48.1 13.5 159.9 18,664

2 AlCl3 anode Process theoretical minimum energy requirements kWh/kg

25 °Ccell Electrolytic work requirement (ΔG) 6.48

3Cl2 Thermal energy for temperature maintenance (ΔH –ΔG) 0.78

25 °C Thermal energy for Al at 960 °C 0.39Theoretical minimum 7.65

Thermal energy for Cl2 at 960 °C 0.19


Table F.5 Theoretical minimum energy for kaolinite/aluminum chloride reduction(Products-

Reactants Products Net aluminum reaction reactants)temp. temp. Carbochlorination (960 °C) cal/g mole Al

(3 Al2O3.2SiO2 + 14 C + 21 Cl2) to (6 AlCl3 + 6 SiCl4 + 7 CO + 7 CO2) = Net

ΔG –2,374,053 0 0 –901,721 –884,799 –229,655 –659,819 –50,323ΔH –2,531,520 0 0 –1,009,847 –942,161 –184,910 –658,363 –43,960

7Al2O3.2SiO2 14 Al ΔS 19 1119 144 474 331 357

25 °C Kaolinite 960 °C

Chloride reduction (700 °C)to (2 AlCl3) to (2 Al + 3 Cl2) = Net

ΔG –300,574 0 0 150,287aluminium 14SiO2 ΔH –336,616 0 0 168,308

chloride 25 °C ΔS 48 14 160 18,664

Process theoretical minimum energy requirements kWh/kg

to 7CO2 Thermal work for AlCl3 production –1.90

25 °C Electrolytic work requirement (ΔG) 6.48reduction Thermal energy for temperature maintenance (ΔH–ΔG 0.78

Thermal energy for Al at 960 °C 0.39system Theoretical minimum 5.76

14 °C 7CO Thermal energy for CO2 at 960 °C 0.1125 °C 25 °C Thermal energy for CO at 960 °C 0.25


248 / Aluminum Recycling and Processing for Energy Conservation and SustainabilityTa

ble

F.6

Ther

moc

hem

istr

y da

ta fo

r el

emen

ts a

nd c

ompo

unds

ass

ocia

ted

wit

h al

umin

um p

rodu

ctio

nC

hem

ical

Abs

trac

ts

Serv

ice

Reg

istr

atio

n N

o.

Mol

ecul

arC

hem

ical

H

(s)

G(s

)S(

s)(C

AS

RN

)w

eigh

tfo

rmul

aJ/

mol

cal/m

olJ/

mol

cal/m

olJ/

mol

Kca

l/mol

KJ/

mol

Kca

l/mol

K

Alu

min

um74

29-9

0-5

26.9

8A

l0

00

028

.36.

764

24.3

55.

82A

lum

inum

chl

orid

e74

46-7

0-0

133.

34A

lCl 3

–704

,200

–168

,308

–628

,800

–150

,287

100.

724

.061

91.8

421

.95

Cor

undu

m13

34-2

8-1

101.

96A

l 2O3

–1,6

75,7

00–4

00,5

00–1

,582

,300

–378

,179

50.9

12.1

6579

18.8

8G

ibbs

ite(a

)..

.15

5.96

Al 2O

3. 3H

2O–1

,293

,100

–309

,058

–1,1

54,9

00–2

76,0

28..

...

...

...

.K

aolin

ite(a

)13

32-5

8-7

...

Al 2O

3. SiO

2. 2H

2O–4

,119

,000

–984

,465

–3,7

93,9

00–9

06,7

64..

...

...

...

.K

aolin

ite,m

eta

...

162.

04A

l 2O3. S

iO2

...

–843

,840

...

–791

,351

...

...

...

...

Gra

phite

7440

-44-

012

.01

C0

00

05.

71.

361

...

...

Chl

orin

e..

.70

.91

Cl 2

(g)

00

00

222.

953

.286

...

...

Car

bon

mon

oxid

e63

0-08

-028

.01

CO

(g)

–110

,523

–26,

416

–137

,268

–32,

808

197.

947

.301

...

...

Car

bon

diox

ide

124-

38-9

44.0

1C

O2

(g)

–393

,513

–94,

052

–394

,383

–94,

260

213.

651

.061

...

...

Oxy

gen

7782

-44-

732

.00

O2

(g)

00

00

205.

049

.003

...

...

Wat

er..

.18

.00

H2O

(g)

–241

,826

–57,

798

–228

,582

–54,

632

188.

845

.132

33.5

98..

.Si

lica

diox

ide

1480

8-60

-760

.08

SiO

2–9

10,7

00–2

17,6

63..

.–2

20,6

1541

.46

9.90

9..

.Si

licon

tetr

achl

orid

e10

026-

04-7

169.

90Si

Cl 4

(g)

–657

,000

–157

,027

–617

,000

–147

,467

330.

7079

.039

90.3

21.6

Sour

ce:D

.R. L

ide,

Ed.

,Han

dboo

k of

Che

mis

try

and

Phy

sics

,80t

h ed

.,C

RC

.(a

) So

urce

:S. K

. Sax

ena,

Ed.

,Adv

ance

s in

Phy

sica

l Geo

chem

istr

y,da

ta c

ourt

esy

of M

r. B

. Hem

ingw

ay a

t U.S

.G.S

.

Tabl

e F.

7C

hang

es in

hea

t of

form

atio

n va

lues

as

a fu

ncti

on o

f tem

pera

ture

Tem

pera

ture

Hea

ts o

f fo

rmat

ion

in c

alor

ies

per

gram

mol

e°C

KA

l 2O3(

a)C

(a)

Al(

a)C

O2(

a)C

O(a

)O

2

2529

8–4

00,3

000

0–9

4,05

0–2

6,40

00

727

1000

–404

,400

2310

18,7

10–9

4,40

0–2

6,75

092

4982

711

00–4

04,0

0033

2021

,710

–94,

250

–26,

900

10,5

1592

712

00–4

03,6

0038

5024

,740

–94,

300

–27,

000

11,8

4396

012

33–4

03,5

0040

3025

,750

–94,

300

–27,

100

12,2

9510

2713

00–4

03,2

0043

9027

,790

–94,

300

–27,

300

13,2

3311

2714

00–4

02,8

0049

3030

,850

–94,

300

–27,

350

14,6

85

(a)

Sour

ce:V

alue

s fr

om T

echn

ical

Wor

king

Gro

up o

n In

ert A

node

Tec

hnol

ogie

s,A

ppen

dix

A.9

,p 1

1

Cp

(s)

Appendix F: Theoretical Energy Data and Calculations / 249

Table F.8 Theoretical minimum energy for gibbsite dehydration(Products-

Reactants Products Gibbsite dehydration reaction reactants)temp. temp. cal/g mole Al2O3

(Al2O3.3H2O) to (Al2O3 + 3H2O) = Net

Al2O3 ΔH –618,117 –400,500 –173,393 44,223

25 °CAl2O3

.3H2O Kiln

25 °C Process theoretical minimum energy requirements

3 H2O kWh/kg alumina

25 °C 0.50


Table F.9 Heat of capacity equations for gasesassociated with aluminium productionStandard molar heat capacity, C=a + bT + cT2

cal/mol K a b�103 c�107

O2 6.148 3.102 –9.23CO 6.420 1.665 –1.96CO2 6.214 10.396 –35.45Cl2 7.576 2.424 –0.65

Source: W.F. L udar, A Different Approach to Thermodynamics, 1967T in degrees K

Table F.10 Theoretical minimum energy for kaolinite dehydrationReactants Products (products-temp. temp. Reaction reactants)

Al2O3.2SiO2 Al2Si2O5 (OH)4 to (Al2O3

.2SiO2 + 2 H2O) = Net cal/g mole Al2O3.2SiO2

25ΔG –906,764 –791,351

Al2Si2O5 (OH)4 Kiln ΔH –984,465 –843,840 –115,596 25,029

25 ΔSkWh/kg

2 H2O Theoretical minimum 0.18

25

APPENDIX G

Aluminum Heat Capacity andHeat of Fusion Data

Table G.1 Formulas for calculating heat capacity and heat of fusion

where t = K/1000 and A, B, C, D, E, F, G, and H are constants

Source: Standard Reference Data Program, National Institute of Standard, and Technology, http://webbook.nist.gov (accessed July 2007)

Table G.2 Formula constants for aluminumSolid (298 to 933.45 K, 1 atm):

A B C D E F G H6.71348 –1.29418 2.04599 0.819161 –0.066294 –2.18623 14.7968 0

Liquid (933.45 to 2790.812 K, 1 atm):

A B C D E F G H7.588681 9.40685�10–9 4.26987�10–9 6.43922�10–9 1.30976�10–9 –0.226024 17.5429 2.524381

Heat capacity (cal/mol K) A B C DE2⋅ = + ⋅ + ⋅ + ⋅ +t t t

t

2 3

Standard enthalpy (Kcal/mol) AB C

= ⋅ +⋅

+⋅

tt t( ) (

2

3))

3 4+

⋅− + −

(D4

) EF H

t

t

Standard entropy (cal/mol K) A ln BC

⋅ = ⋅ + ⋅ +⋅

( )(

t tt22

2

3

3 22

) ( )

( )+

⋅−

⋅+

D EG

t

t



252 / Aluminum Recycling and Processing for Energy Conservation and Sustainabilityca

l/mol

K

9.00

8.50

8.00

7.50

7.00

6.50

6.00

5.50

5.00

4.50

4.0015000 500 1000

Temperature, °C

Solid Liquid

Table G.3 Aluminum energy requirements for heating and meltingEnergy Cumulative Energy Cumulative

Heat capacity at for step energy to raise for step energy to raiseTemperature temperature, change, from 25 °C, change, from 25 °C,°C K cal/mol k kWh/kg kWh/kg Btu/lb Btu/lb

25 298 Solid 5.79 0.00 0.00 . . . . . .660 933 7.88 0.19 0.19 286 286660 . . . Fusion 94.5 (cal/g) 0.11 . . . 170 . . .660 933 Liquid 7.59 . . . 0.29 . . . 456775 1048 7.59 0.04 0.33 58 515960 1233 7.59 0.06 0.39 94 6082000 2273 7.59 0.34 0.73 526 1134

Smelting 25 to 960 °C Total 0.39 . . . 608Furnace melting 25 to 775 °C Total 0.33 . . . 515

Note: Heat capacity for solid aluminum varies significantly with temperature, whereas for molten aluminum, it is nearly constant.

Heat capacity as a function of temperature

WETTED CATHODES allow the anode-cathode distance to be reduced. This results in alowering of the voltage requirement for amperage

to pass through the cryolite bath. The followingtable shows the effect of lowering the anode-cathode distance on energy consumption.

APPENDIX H

Impact of Using Different Technologies on Energy Requirements for Producing Aluminum

Table H.1 Impact of wetted cathode technology on primary metal electrolysisTypical modern prebaked Typical modern prebaked

Typical modern prebaked Hall-Heroult cell operating Hall-Heroult cell operatingHall-Heroult cell operating at 95% current efficiency at 95% current efficiency

Typical modern at 95% current efficiency and retrofitted with a and retrofitted with a slopedprebaked Hall-Heroult and retrofitted with a wetted cathode surface, and wetted cathode surface,

Wetted cathode cell operating at 95% wetted cathode surface aluminum sump, and aluminum sump, and technology current efficiency and reduced ACD reduced ACD reduced ACD

ACD = 4.5 cm ACD = 3.5 cm ACD = 2.5 cm ACD = 2.0 cm

Energy requirements V (dc) kWh/kg Al V (dc) kWh/kg Al V (dc) kWh/kg Al V (dc) kWh/kg Al

Reaction 1.20 3.76 1.20 3.76 1.20 3.76 1.20 3.76Additional energy requirements:

External 0.15 0.47 0.15 0.47 0.15 0.47 0.15 0.47Anode 0.30 0.94 0.30 0.94 0.30 0.94 0.30 0.94Anode polarization 0.55 1.73 0.55 1.73 0.55 1.73 0.55 1.73Cathode polarization 0.05 0.16 0.05 0.16 0.05 0.16 0.05 0.16Cryolite bath 1.75 5.49 1.36 4.27 0.97 3.05 0.78 2.44Cathode 0.45 1.41 0.45 1.41 0.45 1.41 0.45 1.41Other 0.15 0.47 0.15 0.47 0.15 0.47 0.15 0.47

Onsite energy valuesCell total 4.60 14.43 4.21 13.21 3.82 11.99 3.63 11.38

Energy savings, % . . . 8 17 21Anode manufacturing 0.61 0.61 0.61 0.61Total on-site cell and anode 15.04 13.82 12.60 11.99

Energy savings, % . . . 8 16 20

Tacit energy valuesCell total 4.60 41.83 4.21 38.29 3.82 34.75 3.63 32.99

Energy savings, % . . . 8 17 21Anode manufacturing 6.01 6.01 6.01 6.01Total on-site cell and anode 47.83 44.30 40.76 38.99

Energy savings, % . . . 7 15 18





Table H.2 Impact of inert anode and wetted cathode technology on primary metal electrolysisA B C D

Inert anode operating Inert anode operatingTypical modern at 95% current efficiency at 95% current efficiency,

Hall-Heroult cell operating Inert anode direct with polarization differences with oxygen polarizationInert anode and wetted at 95% current efficiency substitution operating reported for oxygen differences and an ACD of

cathode technology and ACD of 4.5 cm at 95% current efficiency generation 2 cm using wetted cathode

Energy requirements V (dc) kWh/kg Al V (dc) kWh/kg Al V (dc) kWh/kg Al V (dc) kWh/kg Al

Reaction 1.20 3.76 2.20 6.90 2.20 6.90 2.20 6.90Additional energy requirements:

External 0.15 0.47 0.15 0.47 0.15 0.47 0.15 0.47Anode 0.30 0.94 0.30 0.94 0.30 0.94 0.30 0.94Anode polarization 0.55 1.73 0.55 1.73 0.10 0.31 0.10 0.31Cathode polarization 0.05 0.16 0.05 0.16 0.05 0.16 0.05 0.16Cryolite bath 1.75 5.49 1.75 5.49 1.75 5.49 0.78 2.44Cathode 0.45 1.41 0.45 1.41 0.45 1.41 0.45 1.41Other 0.15 0.47 0.15 0.47 0.15 0.47 0.15 0.47

On-site energy valuesCell total 4.60 14.43 5.60 17.57 5.15 16.15 4.18 13.11

Energy savings, % . . . –22 –12 9Anode manufacturing 0.61 0.76 0.76 0.76Total on-site cell and anode 15.04 18.33 16.92 13.87

Energy savings, % . . . –22 –13 8

Tacit energy valuesCell total 4.60 41.83 5.60 50.92 5.15 46.83 4.18 37.99

Energy savings, % . . . –22 –12 9Anode manufacturing 6.01 0.76 0.76 0.76Total on-site cell and anode 47.83 51.68 47.59 38.75

Energy savings, % . . . –8 1 19

ACD, anode-cathode distance. Inert anodes by themselves require additional reaction voltage, as can be seen by comparing columns A and B. However, this additionalenergy requirement is offset by the elimination of carbon anode manufacturing, along with the elimination of the feedstock energy associated with carbon. The esti-mate for the energy associated with the manufacture of inert anode is shown in Table H.3. Column C shows the impact that results from the lower anode polarizationand the ability to design to release gas more effectively. Wetted cathodes combined with inert anodes can provide additional savings, as shown in column D.

Table H.3 Estimate of energy requirement for manufacturing an inert anodeThe energy requirements for manufacturing an inert anode are significantly less than the total manufacturing energy of the many consumablecarbon anodes that it replaces. Below is an estimate of the energy required to manufacture an inert anode. Assuming an inert anode has a celllife of two years, the equivalent carbon energy requirements can be calculated. Using 1 cm2 as a basis, the following can be calculated:

0.85 A/cm2 is the typical current density for a modern Hall-Heroult cell current density.2980 Ah/kg Al is the theoretical amperage required to produce aluminum.95% is the typical efficiency of a modern Hall-Heroult carbon anode cell.0.446 kg of carbon are required to produce 1 kg of aluminum.

From the above data, the amount of carbon consumed and aluminum produced per cm2 over a two-year period is calculated to be:

14.82 kg of carbon are consumed per cm2 over a two-year operating period.33.23 kg of aluminum are produced per cm2 over a two-year operating period.

From Table 10.9:

6.01 kWh/kg of Al produced is the tacit energy associated with carbon anode. The total energy associated with the two-year operation ofthe 1 cm2 of carbon anode can be calculated.

200 kWh of tacit energy are consumed for anode manufacture and use.2.27 g/cm3 is the density of carbon anode. The height of the 1 cm2 of carbon can be calculated as below.6529 cm is the height of carbon anode with a 1 cm2 base or anode face for two years of operation.

The materials under consideration for inert anodes have no inherent fuel value, as does the carbon anode. The tacit energy requirement associ-ated with the manufacturing of an inert anode is related to the extraction of materials and the inert anode manufacturing process. Assuming thatthe inert anode is 5 cm thick per cm2 of anode face and that it requires five times the total tacit energy of a carbon anode (which includes itsfuel value) to manufacture, it can be calculated that:

0.76 tacit kWh/kg of aluminum will be required to produce an inert anode.

Appendix H: Impact of Using Different Technologies on Energy Requirements / 255

Tabl

e H

.4Es

tim

ate

of e

nerg

y re

quir

emen

t to

man

ufac

ture

alu

min

um u

sing

diff

eren

t te

chno

logi

esT

ypic

al m

oder

n H

all-

Her

oult

ce

ll op

erat

ing

at 9

5% c

urre

ntIn

ert

anod

e op

erat

ing

at 9

5%T

ypic

al m

oder

n H

all-

Her

oult

effi

cien

cy,r

etro

fitte

d w

ith

a sl

oped

curr

ent

effi

cien

cy w

ith

oxyg

enC

hlor

ide

redu

ctio

n of

kao

linit

e

preb

aked

ano

de c

ell o

pera

ting

and

wet

ted

cath

ode

surf

ace,

pola

riza

tion

dif

fere

nces

and

Car

both

erm

ic r

educ

tion

wit

hcl

ays:

elec

trol

ysis

cur

rent

A

lum

inum

pro

duct

ion

at 9

5% c

urre

nt e

ffic

ienc

y an

dal

umin

um s

ump,

and

a A

CD

of

2.0

cm u

sing

wet

ted

a re

acti

on e

ffic

ienc

y of

95%

effi

cien

cy 9

5%,r

eact

ion

effi

cien

cyte

chno

logi

esw

ith

AC

D o

f 4.

5 cm

redu

ced

AC

D o

f 2.

0 cm

cath

ode

tech

nolo

gyan

d a

furn

ace

effi

cien

cy o

f 80

%95

%,a

nd h

eati

ng e

ffic

ienc

y 85

%

Raw

mat

eria

lskW

h/kg

Al

kWh/

kg A

lkW

h/kg

Al

kWh/

kg A

lkW

h/kg

Al

(a)

On-

site

ene

rgy

requ

ired

Alu

min

a7.

59A

lum

ina

7.59

Alu

min

a7.

59A

lum

ina

7.59

Kao

linite

8.14

Taci

t ene

rgy

requ

ired

Alu

min

a8.

21A

lum

ina

8.21

Alu

min

a8.

21A

lum

ina

8.21

Kao

linite

8.81

On-

site

rea

ctio

n en

ergy

req

uire

men

ts

Rea

ctio

n th

erm

al..

...

...

.7.

71–1

.90

Furn

ace

loss

es..

...

...

.1.

36(b

)0.

40V

(dc)

V(d

c)V

(dc)

V(d

c)R

eact

ion

elec

trol

ysis

1.20

3.76

1.20

3.76

2.20

6.90

...

2.07

6.48

Cel

l ohm

ic3.

4010

.67

2.43

7.62

1.98

6.20

...

0.93

2.93

Tota

l rea

ctio

n en

ergy

4.60

14.4

33.

6311

.38

4.18

13.1

19.

073.

007.

91A

node

man

ufac

turi

ng0.

610.

610.

760

0.76

On-

site

ene

rgy

savi

ngs,

%..

.20

840

42

Tota

l on-

site

ene

rgy

requ

irem

ents

incl

udin

g ra

w m

ater

ials

Raw

ore

mat

eria

ls7.

597.

597.

597.

598.

14R

eact

ions

14.4

311

.38

13.1

19.

077.

91C

arbo

n/el

ectr

odes

0.61

0.61

0.76

(c)

0(d

)0.

76To

tal e

nerg

y sa

ving

s,%

...

135

2626

Tota

l tac

it e

nerg

y re

quir

emen

ts in

clud

ing

raw

mat

eria

ls

Raw

ore

mat

eria

ls8.

218.

218.

218.

218.

81R

eact

ions

41.8

332

.99

37.9

926

.28

24.3

6C

arbo

n/el

ectr

odes

6.01

6.01

0.76

(c)

8.41

(d)

12.1

0To

tal e

nerg

y sa

ving

s,%

...

1616

2319

AC

D,a

node

-cat

hode

dis

tanc

e.(a

)K

aolin

ite c

onta

ins

27.2

% A

l by

wei

ght.

Bau

xite

con

tain

s 45

% A

l by

wei

ght.

Bot

h m

ater

ials

con

tain

app

roxi

mat

ely

the

sam

e pe

rcen

tage

of

impu

ritie

s. F

or th

is e

stim

ate,

it is

ass

umed

that

the

proc

essi

ng a

nd c

alci

natio

n en

ergy

of

kaol

inite

is th

e sa

me

as b

auxi

te. H

owev

er,6

6% m

ore

kaol

inite

mus

t be

proc

esse

d to

pro

duce

a k

ilogr

am o

f al

umin

um th

an b

auxi

te. K

aolin

ic c

lays

con

tain

val

uabl

e tit

aniu

m in

add

ition

to th

eir

silic

on c

onte

nt. T

hese

mat

eria

ls w

ill li

kely

be

reco

vere

d in

the

proc

essi

ngpl

ant a

nd a

ccou

nt f

or a

ppro

xim

atel

y 35

% o

f ka

olin

ic c

lay

cont

ent.

Thi

s ta

ble

allo

cate

s 65

% o

f th

e to

tal r

aw m

ater

ial e

nerg

y re

quir

emen

t to

the

kaol

inite

mat

eria

l use

d fo

r al

umin

um m

anuf

actu

ring

.(b

)T

he c

arbo

chlo

rina

tion

reac

tion

is e

xoth

erm

ic (

–1.9

0 kW

h/kg

Al)

. How

ever

,Tot

h A

lum

inum

rep

orts

that

,bas

ed o

n pi

lot p

lant

exp

erie

nce

and

the

chal

leng

es o

f ca

ptur

ing

the

offg

as e

nerg

y,a

smal

l qua

ntity

of

ener

gy (

0.4

kWh/

kg A

l) is

req

uire

d to

mai

ntai

n th

e re

actio

n sy

stem

tem

pera

ture

.(c

)T

he c

arbo

ther

mic

rea

ctio

n re

quir

es tw

ice

the

carb

on a

s th

e H

all-

Her

oult

reac

tion.

The

Hal

l-H

erou

lt re

actio

n,on

a th

eore

tical

bas

is,r

equi

res

0.33

kg C

/kg

Al.

In th

is e

stim

ate,

it is

ass

umed

that

the

carb

othe

rmic

rea

ctio

n ut

ilize

d ca

rbon

at a

95%

eff

i-ci

ency

. Unl

ike

a ca

rbon

ano

de,t

he c

arbo

n us

ed d

oes

not r

equi

re th

e en

ergy

ass

ocia

ted

with

car

bon

anod

e m

anuf

actu

ring

. Hen

ce,o

nly

the

fuel

or

seco

ndar

y en

ergy

req

uire

men

t of

the

carb

on is

use

d.(d

)T

he c

arbo

chlo

rina

tion

reac

tion

requ

ires

0.8

9 kg

of

carb

on p

er k

ilogr

am o

f al

umin

um p

rodu

ced.

In

this

est

imat

e,it

is a

ssum

ed th

at th

e ca

rboc

hlor

inat

ion

reac

tion

utili

zed

carb

on a

t a 9

5% e

ffic

ienc

y. I

t is

also

ass

umed

that

the

ener

gy r

equi

red

and

fuel

valu

e of

the

carb

on is

the

sam

e on

a w

eigh

t bas

is a

s th

e ca

rbon

util

ized

for

car

both

erm

ic r

educ

tion

of a

lum

inum

(se

e fo

otno

te c

).

A

alumina. An oxide of aluminum (Al2O3), andthe compound from which aluminum metalis commercially obtained.

aluminum. A versatile, silvery-white metal.When exposed to the atmosphere, alu-minum rapidly forms an oxide film that pre-vents it from reacting with air and water.This gives it exceptional corrosion-resistantproperties. Aluminum is not found in natureas a free metal, like gold, but is chemicallybound to other elements. Aluminum is themost abundant metal in the Earth’s crust(8.1%). Atomic number, 13; atomic mass,26.982; melting point, 993.52 K; boilingpoint, 2698 K.

anode. A positively charged mass or surface thatattracts negatively charged ions (anions). Theanode used in the Hall-Heroult process iscomposed of carbon. The oxygen-containinganions react on the anode surface, releasingoxygen that consumes the carbon to form carbon dioxide.

anode-cathode distance (ACD). The geometriclinear distance between the anode and thecathode is a critical measurement in an elec-trolytic cell. This distance affects the voltageand energy requirement of a cell.

anode effect. An aluminum-industry idiom usedto describe a process upset where the anodereaction shifts from producing oxygen to flu-orine, and the cell voltage increases. Anodeeffects are primarily the result of having in-sufficient alumina dissolved in the bath andavailable at the anode for reduction.

B

bath. An aluminum-industry idiom referring tothe cryolite-based electrolyte pool in the re-duction cell.

bauxite. A prime source of alumina, found as acollection of small, reddish-brown nodulesin a light-brown, earthy matrix. Commercialbauxite ore contains 30 to 60 wt% alumina.

Bayer process. A process developed by KarlBayer in 1888 that refines bauxite ore intoalumina grains. It is the process currently inuse worldwide.

C

calcining. The process of heating a material toa sufficiently high temperature to drive offvolatile components or to oxidize the mate-rial without fusing it. The aluminum indus-try uses calcining in the Bayer process toproduce alumina and to prepare coke for anodes.

carbon dioxide equivalents (CDE). The pre-ferred unit of measure used to compare theimpact of different greenhouse gases. It is cal-culated by multiplying the quantity of a green-house gas emission by the global-warmingpotential of the gas. The results are commonlyexpressed in terms of a million metric tons ofcarbon dioxide equivalent (106 TCDE).

carbon equivalents (CE). A unit of measureused to compare the impact of different green-house gases. It is calculated by multiplyingthe carbon dioxide equivalent by 12/44, themass ratio of carbon to carbon dioxide.

APPENDIX I

Glossary




carbothermic reduction. An alternative processto electrolytic reduction. The carbothermicprocess reduces alumina in a high-tempera-ture furnace with carbon.

castings. Metal objects that are cast into ashape by pouring or injecting molten/liquidmetal into a mold. This book divides cast-ings into ingot and shape categories. Ingotcastings are produced in molds of very sim-ple cross section, and shape castings arecomplex structures.

cathode. A negatively charged surface that at-tracts positively charged ions (cations). Thecathode surface in the Hall-Héroult process isthe molten aluminum pad, which rests directlyon the cell carbon lining. The aluminum-containing cations react on the cathode sur-face, releasing the aluminum as free metal.

chloride reduction. An alternative process toalumina electrolytic reduction in which alu-minum chloride is used as the feed to thereduction cell.

coke. A carbon product of the crude oil refiningindustry. Green or raw coke contains 8 to10% moisture and 5 to 15% volatile organicmaterials. Coke is calcined in thermal kilnsto remove moisture and volatile organic materials.

cryolite. Na3AlF6, a mineral that, when molten,dissolves alumina to form aluminum andoxide ions. It is the main component used inthe electrolyte bath for aluminum production.

D

dross. The material that forms on the surface ofmolten aluminum as it is held in a furnace. Itis composed of impurities that have surfacedas a result of gas fluxing, oxidized aluminumthat is the result of molten aluminum exposureto the furnace atmosphere, and aluminum thatbecomes entrapped in the surface material.Dross is periodically skimmed off the surfaceof molten aluminum and processed to recoverits aluminum content.

dusting. An aluminum-industry idiom used todescribe fine carbon anode particles that arelost in the electrolyte bath or atmosphereduring electrolytic reduction. Dusting resultsin a loss of productivity.

E

electrolysis. An electrochemical process inwhich the charged species in an electrolyte

are attracted to electrodes, where they reactwith the electrons of the electrical current.Positively charged ions migrate to the cath-ode, and negatively charged ions migrate tothe anode.

electrolyte. A nonmetallic electrical conductorin which current is carried by the movementof ions.

extrusion. The process of forcing the metalingot (or billet) to flow through a die to createa new cross section.

F

feedstock energy. These values represent theenergy inherent in a fuel that is used as mate-rial. For example, aluminum production usescoke as the raw material in carbon anodes.The energy contribution of a feedstock is ex-pressed in terms of calorific or fuel valueplus the tacit/process energy used to producethe feedstock.

G

global-warming potential (GWP). Green-house gases differ in their abilities to trapheat. Global-warming potential is used to ex-press the greenhouse effect of different gasesin a comparable way. The heat-trapping abilityof one metric ton of CO2 is the commonstandard, and emissions are expressed interms of a million metric tons of CO2 equiva-lent, or 106 TCDE.

greenhouse gases (GHG). Atmospheric gasesthat contribute to climate change by increas-ing the ability of the atmosphere to trap heat.

H

Hall-Heroult process. An electrolytic processfor reduction of alumina, developed inde-pendently by Charles Martin Hall and PaulLewis Toussaint Héroult in 1886. Thisprocess is commonly referred to using bothnames, the Hall-Heroult process. It is theprocess used worldwide for commercialaluminum production.

I

ingot. Ingot, as used in this book, describes analuminum casting of simple shape. It in-cludes billets, pigs, sows, T-bar, and othersimple cast semifinished shapes.

Appendix I: Glossary / 259

K

kilowatt-hour (kWh). A unit of energy.

L

life-cycle assessment (LCA). An internationallyrecognized analysis model of the impact of aproduct on energy, environment, economic,and social values. The LCA extends from“cradle-to-grave”: from material acquisitionand production; through manufacturing,product use, and maintenance; and finally,through the end of the product life in disposalor recycling. The LCA is particularly usefulin ensuring that benefits derived in one areado not shift the impact burden to other placeswithin a product life cycle.

O

on-site energy. The energy used within a facil-ity. This is sometimes called primary energy.Electrical on-site energy is the kilowatthours used and does not include the secondaryenergy required for generation and transmis-sion of electricity. Fuel on-site energy use isbased on the calorific heating value of thefuel and does not include the secondaryenergy required to produce and transportthe fuel.

P

pad. An aluminum-industry idiom used to de-scribe the body of molten aluminum that accumulates within the Hall-Heroult elec-trolytic cell.

polarization. The nonuniform concentrationgradients that form near electrodes duringthe reduction process. The reactions occur-ring at the anode and the cathode createlocalized conditions that are different fromthe bulk of the bath. The reactions depletethe supply of reactants and increase thequantity of products. Additionally, in aluminum electrolysis, gas is generated atthe anode, which lowers the effective bathconductivity. An electric overpotential is required to overcome the effects of polar-ization.

pot. An aluminum industry idiom used to describe an electrolytic cell. The term wasderived from the shape of the first cells.

potline. An aluminum-industry idiom that de-scribes the arrangement of a long row of in-terconnected electrolytic cells (pots).

potlining. An aluminum-industry idiom that de-scribes the refractory and carbon materialsused to line the interior of the cell (pot).

primary aluminum. Refers to aluminum metalproduced directly from alumina feedstock bychemical reduction.

Q

quad. A common abbreviation for a quadrillionBtu (1 quad = 1015 Btu).

R

red mud. The residue of insoluble materialsthat results from extracting alumina frombauxite ore. It is also referred to as bauxiteresidue.

reduction cell. A container holding single ormultiple anodes, cathodes, and an electrolyticbath used for reducing a material.

reverberatory furnace. The most commonlyused furnace type in the aluminum industry.The furnace is box-shaped and consists of asteel shell with refractory lining. Fuel is fireddirectly into the box, either from the roof or,more typically, from the sidewall. Heat istransferred to the molten metal with convec-tion and radiation.

rolling. A process that results in the reductionof the cross-sectional area of a metal shapeas it is passed through rotating rolls.

S

secondary aluminum. Aluminum metal that isproduced from recycled aluminum productsand wastes.

T

tacit energy. A term used to describe an energyvalue that equals the combination of on-siteenergy (primary energy) consumption, theprocess energy required to produce and trans-mit/transport the energy source (secondaryenergy), and feedstock energy (energy inher-ent in fuels used as materials). This book usesthe superscript “tf” to denote any value thatincludes the tacit and feedstock energy contri-butions. The book does not include the energy


used to make the equipment or buildings thathouse the process steps (tertiary energy).

U

urban mining. A term that describes the largesource of aluminum available through urbanrecycling programs as compared to bauxitemining.

V

value chain analysis. A method that capturesthe energy and material inputs and outputs ofeach processing step (link) and builds the cu-mulative value for each product along thechain. A value chain analysis, or “cradle-to-shipping dock” analysis, is an integral part ofa life-cycle analysis.

AACC (Aluminum Can Council), 123 ACD (anode-cathode distance),185–186, 191 alternative materials, 21–22alumina

energy requirements, 175 production, 75–76(T)refining, 38, 47–48(F),173–175(F) theoretical minimum energy requirements, 175–176

aluminum, definition of, 257 aluminum, LCA assessment of

data coverage, reporting and interpretation, 67–68,69–70(T), 71(T)

data qualitydata consistency, 68, 71(T) data interpretation items, 72–73(T) introduction, 68missing process data supplemented, 72, 73(T)

introduction, 67recycling issues

energy flows, 84 general, 83–84 introduction, 83 recycling, 84–88

unit processes/results by process alumina production, 75–76(T) anode production, 76, 77–78(T), 79bauxite mining, 75 electrolysis, 79–81(T) ingot casting, 81–83(T) introduction, 73–75(F,T) solid waste, 83

aluminum, recycling ofaluminum foil recycling, 128automobile scrap recycling technology, 125–127(F),

128(F) building/construction recycling, 128 can recycling technology, 122–125(F) energy savings, 115 impurity control, 129–130(F) industry and recycling trends, 110–112(F,T)

in 1980s, 112(F), 113(F) current trends, 112–114(F)

introduction, 109–110

molten metal handling/safety, 130–132 recyclability, 114–115(F) recycling loop, 115–116(F)remelting, process developments for, 118–119, 120(F)scrap specifications, 116 scrap systems, developing, 119–122(F)technological aspects

alloy integrity, 117–118 alloys, 117energy and resources, 117 increased production requirements, 117 introduction, 116–117 mixed scrap, 117

U.S. aluminum supply, 111(F) U.S. metal supply, 111(T)

aluminum aerospace alloys, recycling introduction, 150–151recycled aircraft components in aircraft, 151(T) recycled aircraft components in castings, 151–152(T) recycled aircraft components in nonaircraft application,

151–152(T)Aluminum Can Council (ACC), 123 aluminum chloride chemistry, 138 aluminum extruding, 34(F), 37(T), 39,45(F),

55–56(F) aluminum foil recycling, 128 aluminum heat capacity and heat of fusion data,

251–252(T) aluminum industry, sustainable development for

energy efficiency, 97–99(F), 100(F), 101(F) fluoride emissions, 95–97(F), 98(F) introduction, 91–92 natural resources, 101–102 PCF emissions, 94–95, 96(F) recycling, 92–94(T), 95(F) in transportation, 99–101 water use, 102

aluminum processingextrusion, 214–215(F) ingot casting, 211–212 introduction, 208 melting, alloying, melt treatment

energy requirements for, 210furnaces and melt treatment, 208–210technological change in next decade, 210–211

Index

262 / Index

rolling, 212–214(F) shape casting, 215–217 thermal treatments, 217–218

aluminum processing area, energy use by, 237–244(T)aluminum production, U.S. energy requirements for

aluminum processing, see aluminum processing aluminum production, primary

Hall-Heroult carbon anode reduction, theoreticalenergy for, 182–183(F)

Hall-Heroult energy utilization, 179–180(F,T)Hall-Heroult process, 183–191(F)introduction, 178production, capacity, growth, 178–179(F)theoretical minimum energy requirement,181–182(F)

aluminum supply, U.S., 170–171(F) carbon anode, 176–178 Hall-Heroult cells, advanced, 191–197(F,T) impact of different technologies on, 253–255(T) introduction, 157–158 methodology, metrics, benchmarks

benchmarks (theoretical, practical minimum, currentpractice), 165

emissions, 168energy chain value analysis, 167–168(F)introduction, 164–165(T)LCA, 166–167tacit, process, feedstock, and ‘secondary’ energies,

165–166transportation energy, 168

overview, 168–170(F,T) primary aluminum processes, alternative

carbothermic technology, 198–201(F,T)introduction, 197–198kaolinite reduction technology, 201–204(F,T)

production/energy consumptionenergy/environmental overview, 160–163(T)energy-reduction opportunities, 163–164(F,T)introduction, 160

raw materials, primaryalumina refining, 173–175bauxite, 171–172(F), 173(T)introduction, 171

secondary aluminum (recycling), 204–208(F) summary, 158–159(F)tacit energy consumption, note on, 159–160(T)

aluminum recycling, emerging trends inalloys designed for recycling, 152–153aluminum aerospace alloys

introduction, 150–151recycled aircraft components in aircraft, 151(T)recycled aircraft components in castings, 151–152(T)recycled aircraft components in nonaircraft

application, 151–152(T)conclusions/looking ahead, 154–155introduction, 147–148objectives/challenges, 148recycled metal, nature of

cast alloy strips, 149(T)introduction, 148–149(T)wrought alloy scrap, 150(T)

recycling-friendly compositions, developingcandidate compositions for, 153–154(T)caveats, 154introduction, 153unialloy, 154

aluminum smeltingCOS generation, 50description of, 39introduction, 49–50(F)PCF generation, 50, 51(T)

aluminum supply, U.S., 111(F)America Recycles Day, 123anode, definition of, 257anode effect, 81, 94, 187, 189, 191, 257anode production, 38–39, 48–49(F), 76, 77–78(T), 79anode-cathode distance (ACD), 185–186, 191Association of Plastics Manufacturers in Europe,

(APME), 12automobile scrap recycling technology

high-temperature process, 127overview, 125–127(F), 128(F)separation technologies, other, 127–128

automotive aluminum recycling, 57–59, 62(F), 63–64(F)

Bbath, 257bath ratio, 185bauxite, 171–172(F,T), 257bauxite mining, 37–38, 46–47, 48(F), 75Bayer process, 75–76(T), 77(T), 174–175bulk-particle cleaning, 143(F)

Ccalcining process, 38, 177, 257can recycling technology

alloy separation, 124–125collection, 124delacquering, 124melting, preparation, and casting, 125, 126(F)overview, 122–124(F)

Canadian Standards Association (CSA), 1Cans for Habitat, 123carbon anode

energy requirements, 177(T)overview, 176–177theoretical energy values, 177–178

carbon equivalent (CE), 257carbonyl sulfide (COS) generation, 50carbotheric reduction, 258. see also carbothermic

technologycarbothermic technology

carbothermic reactors/Hall-Heroult cells, benefits for,198(T), 200

environmental impacts of, 200–201(T)overview, 198–199(F,T)theoretical energy for, 199–200(F)

case history: LCA of automobile fenderdata origin/collection, 8–9(F)goal and scope, 8(T)

aluminum processing (continued)

Index / 263

impact assessment, 11,12(F)improvement options, 12introduction, 7–8inventory results, 9–11(F)valuation, 11–12

castingsdefinition of, 258ingot casting

description of, 40, 211–212primary, 50–51(F), 52(T), 53(T)secondary, 52–53, 54(F,T), 55(T)survey results for, 81–83(T)

shape castingdescription of, 215–217overview, 40process, 53–54, 56(F), 59(F), 61(F)

cathode, definition of, 258center-worked prebake, 49chain value analysis or cradle-to-shipping dock

analysis, 167–168(F)chloride reduction, 197–198, 201, 202, 258Closed Substance Cycle and Waste Management

Act of 1996, 22closed-loop allocation procedure, 85coke, definition of, 258cold rolling, 40committee on the study of materials (COSMAT), 16composition-based particle sorting

6111-5182 alloy sort, 143–144(F,T)bulk-particle cleaning, 143(F)full-scale industrial prototype LIBS sorter, 144introduction, 141–143(F)sorting process development, continued, 144–145

continuous melting, 119, 120(F)corporate welfare, 24COS (carbonyl sulfide) generation, 50cradle-to-shippping dock analysis or chain value

analysis, 167–168(F)crucible furnaces, 209cryolite or molar ratio, 185, 258CSA (Canadian Standards Association), 1Curbside Value Partnership, 123current density, 186

Ddata collection, 8data quality indicators (DQIs), 44–45delacquering, 124dematerialization, 17, 21, 22Department of Defense (DoD), 21Department of Energy (DoE), 21direct extrusion process, 215direct-chill (DC) casting process, 211dross, definition of, 258dry sand molding, 216“dusting”, 176, 258

Eelectricity production, 38electrolysis

definition of, 258steps in, 8–9(F)unit processes/results by process, 79–81(T)

electrolyte, definition of, 258electrorefining, 137EMAT (Enviromental Management and

Technology), 23emission data and calculations, 231–236(T)energy bank concept, 110, 112–113energy intensity of materials produced in U.S.,

223–224(T)energy requirements for producing aluminum, impact of

different technologies on, 253–255(T)energy resource bank, 92, 110energy sources and materials, energy values for,

2l5–227(T)Environmental Management and Technology

(EMAT), 23Enviromental Protection Agency (EPA), 1, 20, 189Enviromental Technology Initiative (ETT), 19European Aluminum Association (EAA), 12Explicit LCA (XLCA), 22extruding, 56(f)extrusion, 214–215(F), 258

FFaraday’s law, 182Federation of Materials Societies (FMS), 28–29feedstock energy, 3, 43, 166, 223–224, 258finely divided aluminum, 132Finnboard, 12fluid bed or stationary kiln, 38, 63, 76, 257fluoride emissions, 95–97(F), 98(F)fluxing, 39, 82, 210, 212

GGHG emission reduction, 162Global Recycling Committee (GARC), 104global-warming potential (GWP), 189, 258Granges box, 127green sand casting, 216greenhouse gas (GHG) emission, 87, 258gross domestic product (GDP), 15

HHall-Heroult cells, advanced

carbon and inert anode, energy requirements for,194–196(F,T)

inert anode, 193–194, 194(F)introduction, 191multipolar cells, 196–197(F)wetted drained cathodes, 191–193(F)

Hall-Heroult inert anode, 193–194Hall-Heroult process

alternative technologies, 190, 191cell operation, typical, 183–184(F)cell subsystems and variables, 185–188(F)description of, 39, 178

264 / Index

energy utilization, 179–180(F,T)environmental considerations

anode production, 189cell waste products, 189GHG emissions, 188–189

technological change in next decade, 189–191technological changes to, 190, 191voltage requirements, 184–185(F)

horizontal stud Soderberg, 49hot rolling processes, 39–40hot/cold mill rolling, 56, 57(F)hydroelectric distribution and electrical energy values,

229–230(T)

Iindirect extrusion process, 215industrial ecology, 19Industries of the Future Program, 27industry-weighted use (industry wt. use), 73ingot, definition of 258ingot casting

description of, 40, 211–212primary, 50–51(F), 52(T), 53(T)secondary, 52–53, 54(F,T), 55(T)survey results for, 81–83(T)

International Aluminum Institute (IAI), 91International Iron and Steel Institute, 12International Primary Aluminum Institute (IPAI), 43International Standards Organization (ISO), 33investment molds, 216

Kkaolinite reduction technology

environmental impacts of, 203introduction, 201–202(F)kaolinite reduction/Hall-Heroult cells, benefits for,

202–203(T)theoretical energy for, 202

kilowatt-hour (kWh), definition of, 259

Llaser-induced breakdown spectroscopy (LIBS), 126,

138–139, 148, 152LCI (life-cycle inventory). see life-cycle inventory (LCI)LIBS sorter, full-scale industrial prototype, 144life-cycle analysis or assessment (LCA)

definition of, 1(F), 259description of, 166–167goals of, 1–2

life-cycle economic costs (LCAecon), 1life-cycle engineering/design

conclusions, 12–13introduction, 1–2LCA and LCI software tools, 13(T)LCA results, application of

case history: LCA of automobile fender. see case history: LCA of automobile fender

different approaches to, 6–7(F)example: LCA of a pencil, 5–6(F)introduction, 5

LCA triangle, 2(F)process steps

goal definition/scoping, 2–3impact assessment/interpretation, 3–4introduction, 2(F)inventory interpretation, 4life-cycle inventory (LCI), 3,4(T)

life-cycle environmental costs (LCAenv), 1

life-cycle inventory (LCI)ancillary material analysis, 43–44data integration/presentation

electrical energy, 42–43emissions from fuel, 42, 43(T)feedstock energy, 43introduction, 42(F)precombustion energy, 42transportation energy, 43, 44(T)

data qualityqualitative DQIs, 45quantitative DQIs, 44–45representativeness, 45–46

definition of, 3, 4(T)introduction, 33–35(F)inventory analysis

allocation procedures, 41anomalies/missing data, treatment of, 42calculation procedures, 41–42deliberate omissions, 41–42introduction, 41–42(F,T)

manufacturing unit processesextruding, 54–55, 56(F)hot/cold mill rolling, 56, 57(F)introduction, 53shape casting, 53–54, 56(F)

methodologyalumina refining, 38aluminum smelting, 39anode production, 38–39bauxite mining, 37–38cold rolling, 40consumer scrap, 40electricity production, 38extruding, 39geographic coverage, 36(T)Hall-Heroult process, 39hot rolling, 39–40ingot casting, 39introduction, 35–36manufacturing scrap transport, 40secondary ingot casting, 40shape casting, 40shredded aluminum scrap, 40shredding and decoating, 40technology coverage, 36, 37–38(T)

primary aluminum unit processes,46–51(F)

product system, results by

Hall-Heroult process (continued)

Index / 265

automotive aluminum recycling, 57–59, 62(F),63–64(F)

introduction, 56–57, 58–59(F), 60(F)results, interpretation of

introduction, 65(F)process improvement, areas for, 61–66(F)

secondary aluminum processing, 51–53(F,T)life-cycle social costs (LCAsoc), 1

Mmagnetohydrodynamic (MHD) forces, 186material flow modeling

modelingintroduction, 103–104(T)key results, 105–106(F)validity checks, 104–105(F)

overview, 103materials produced in U.S., energy intensity of,

223–224(T)MatTec (Materials Technology), 23, 24“mean”, 53mega joules (MJ), 41melt fluxing, 137–138melt processing, l37–138metal supply, U.S., 111(T)MHD (magnetohydrodynamic) forces, 186molar or cryolite ratio, 185molten metal explosions, 130–131

causes/prevention of, 131incident reporting

scrap-charge safety, 131sow casting/charging, 131–132

introduction, 130–131molten metal handling/safety

explosions, causes/prevention of, 131finely divided aluminum, 132guidelines/training aids, 132incident reporting, 131–132introduction, 130molten metal explosions, 130–131

multipolar cells, 196–197(F)

NNational Aeronautics and Space Administration

(NASA), 24National Institute of Standards and Technology

(NIST), 25National Science and Technology Council

(NSTC), 23Natural Science Foundation (NSF), 22neutron activation analysis, 138Nuclear Regulatory Commission (NRC), 16

OOffice of Industrial Technology (OIT), 27on-site energy values, 165, 259open-loop recycling approach, 85

Organization for Economic Cooperation andDevelopment (OECD), 15

Organization of the Petroleum Exporting Countries(OPEC), 25

original equipment manufacturers (OEMs), 23

Ppad, definition of, 259Partnership for New-Generation Vehicles (PNGV),

19, 25perfluorocarbon (PCF), 50, 51(T), 91perfluorocarbon (PCF) emissions, 94–95, 96(F)pigs or sows, 50point feeders, 189–190polarization, definition of, 259polyphenylene oxide and nylon (PPO/PA), 8pot, definition of, 259potline, definition of, 259potlining, definition of, 259practical minimum energy, 165prebake design, 38primary aluminum unit processes

alumina refining, 47–48(F)aluminum smelting, 49–50(F)anode production, 48–49(F)bauxite mining, 46–47, 48(F)introduction, 46, 47(F)primary ingot casting, 50–51, 52(T), 53(T)

product stream life-cycle inventory, 21product system, 3production-weighted mean (wt. mean), 73

Qquad, definition of, 259qualitative DQIs, 45quantitative DQIs, 44–45

Rreclaimed smelter ingot (RST), 205recycled metal, nature of

cast alloy strips, 148–149(T)introduction, 150(T)wrought alloy scrap, 149(T)

recyclingof aluminum. see aluminum, recycling ofintroduction, 84long lifetime products, 87recycled material content approach, 87–88substitution method, 86system expansion and substitution, 84–85value-corrected substitution, 85–86value-corrected substitution method, 86–87

red mud, 38, 174, 175, 259reduction cell, 39, 259remelting, process developments for

continuous melting, 119, 120(F)overview, 118–119

266 / Index

reverberatory furnaces, 118reverberatory furnaces, alternatives to, 119

research and development (R&D), 16reverberatory furnaces, 117, 118, 119, 208–209, 210, 259reverberatory furnaces, alternatives to, 119rigid-container sheet (RCS), 115rolling

advanced technology, 214definition of, 259energy requirements, 213overview, 212–213(F)

rotary kiln, 38, 76, 124, 207

Sscrap systems, developing

automotive scrap, l21–122, 123(F)municipal scrap, 122overview, 119–121(F)UBCs, 121, 122(F)

secondary aluminum procesingintroduction, 51secondary aluminum transport, 51–52secondary ingot casting, 52–53, 54(F,T), 55(T)shredding and decoating, 52

secondary aluminum (recycling)introduction, 204, 205(F)production, 204–205production, capacity, growth, 205–206(F)recycling processes, 206–207theoretical energy for, 207–208

secondary ingot casting, 40shape casting

description of, 215–217overview, 40process, 53–54, 56(F), 59(F), 61(F)

sheet molding compound (SMC), 8shredding and decoating, 40side-worked prebake, 49slotted anodes, 190Society for the Promotion of LCA Development

(SPOLD), 1, 5Society of Environmental Toxicology and Chemistry

(SETAC), 1Soderberg technology

anode, description of, 38horizontal stud Soderberg, 49vertical stud Soderberg, 49

solid-particle sortingchemical-composition based, 138introduction, 138laser-induced breakdown spectroscopy (LIBS), 138–139neutron activation analysis, 138physical property correlations, 138x-ray florescence, 138

spent potlinings (SPLs), 80–81, 189SPL carbon, 80–81SPL refractory bricks, 80–81SPLs (spent potlinings), 80–81sustainability—the materials role

alternative materials, 21–22cleaner processing, 20–21COSMAT list of materials tasks for environmental

issues, 20(T)dematerialization, 22history, 17–19in industrial ecology, 19–20introduction, 15–17professional societies, role of, 28–29summary/recommendations, 29–31total materials cycle, 16(F)U.S. government role—organizational, 23–24U.S. government role—technical

civilian R&D, federal support of, 24–25, 25(T), 26(T)energy and environment, 25–27materials flow data, 27–28

Ttacit energy, 166, 259–260tacit energy values, 165tacit feedback energy, 160tailored alloys, 152theoretical energy data/calculations, 245–249(T)T-ingot, 50trinitrotoluene (TNT), 131type 1 window frames, 88type 2 window frames, 88

Uunialloy, 154United Nations Conference on Environment and

Development (UNCED), 17United Nations Environment Programme (UNEP), 102United States Automotive Materials Partnership

(USAMP) initiative, 33United States Council for Automotive Research

(USCAR), 19United States Geological Survey (USGS), 22urban mine, 125, 206urban mining, 162, 167, 260U.S. aluminum supply, 111(F)U.S. metal supply, 111 (T)used beverage cans (UBCs), 115

Vvalue chain analysis, l67–168(F), 260vehicle recycle partnership, 23vertical stud Soderberg, 49Voluntary Aluminum Industry Partnership (VAIP), 189

Wwaste, defined (UNCED), 17wetted drained cathodes

energy savings for, 192(F)environmental impacts for, 192–193overview, 191–192

World Bureau of Metal Statistics (WBMST), 68

remelting, process developments for (continued)

Index / 267

wrought aluminum alloys, identification/sorting of improving recovery for

aluminum chloride chemistry, 138electrorefining, 137introduction, 137melt fluxing, 137–138melt processing, 137–138solid-particle sorting, 138–139

introduction, 135raw material, sources of, 136(T),

137(F,T)

wrought recovery, pilot processes forcolor grouping, 139–140(F), 141(F,T)composition-based particle sorting, 141–145(F)introduction, 139x-ray absorption imaging, 140, 141(F), 142(F)

XXLCA (Explicit LCA), 22x-ray absorption imaging, 140, 141(F), 142(F)x-ray florescence, 138

Documents

ALUMINUM RECYCLING AND PROCESSING FOR ENERGY CONSERVATION AND SUSTAINABILITY