Bifa PV Eco-EfficBifa PV Eco-efficiency 2013

8/21/2019 Bifa PV Eco-EfficBifa PV Eco-efficiency 2013

1/76

Eco-efficiency Analysis ofPhotovoltaic Modules

Matthias SeitzDr. Magorzata KrobanThorsten PitschkeDr. Siegfried Kreibe

November 2013 bifa-Text No. 62 ISSN 2198-8056


2/76

Imprint

All rights (including the rights of reproduction, distribution and translation) are reserved.This work is protected by copyright. No part of bifa-Text may be reproduced or electronically stored,processed, duplicated or be disseminated in any form without permission of the Publisher.

Publisherbifa Umweltinstitut GmbH (bifa environmental institute)Am Mittleren Moos 4686167 Augsburg (Germany)

AuthorMatthias SeitzDr. Magorzata KrobanThorsten PitschkeDr. Siegfried Kreibe

Funded by

Bavarian State Ministry of the Environment and Consumer ProtectionMunich (Germany)

1st edition 2013; english version 2014

bifa Umweltinstitut GmbH


3/76

Eco-efficiency analysis of photovoltaic modules I

CONTENTS

1 Summary .............................................................................................................. 12 Project background and task .......................................................................... 33 Methodological principles ............................................................................... 44 Notes about the scope ...................................................................................... 65 Lifecycle modeling of PV systems .................................................................. 66 End of Life Disposal and Recycling .......................................................... 147 Results of the eco-efficiency analysis ........................................................ 168 Conclusions ....................................................................................................... 399 Literature ........................................................................................................... 44A Appendix: Methodological principles ......................................................... 45B Appendix: Impact category results .............................................................. 53


4/76


5/76

Eco-efficiency analysis of photovoltaic modules 1

1 Summary

This examination was conducted on behalf of the Bavarian State Ministry of the Environment andConsumer Protection. It analyzes a future-oriented picture of the impact of PV systems on the envi-ronment based on the status quo and considers the complete life cycle as well as the costs for operat-ing PV systems. The manufacture of general PV systems (module and system components), their opera-tion in various cases and their end of life treatment are assessed. Current production processes areportrayed for cell-based silicon technologies (monocrystalline and multicrystalline) as well as for thin-film technologies (CdTe1and CIS2). Based on that, the impact of further developments is examined,e.g., the use of various recycling methods, improvements in basic material manufacturing and the pro-cessing techniques. Unforeseen incidents such as fires or unauthorized disposal of defective or old

modules are not a subject to review. The examination was supported by a technical advisory board,whose representatives belong to companies active in the photovoltaics value chain as well as to repre-sentatives of the ordering customer.

The ecological assessment covers the analysis of seven effects on the environment, which are combinedto show their overall impact on the environment: the ecology index. Due to the very dynamic marketdevelopment and the cost information only available from tendered prices, the electricity generationprices are approximated technology-independent using the feed-in tariffs in the German RenewableEnergy Act (EEG). Accordingly, all technologies are classified equally based on the specified classifica-tion of the examined application cases pursuant to EEG (rooftop up to 30 kWp, rooftop starting from100 kW and ground-mounted others).

Within the framework of the examination, the PV technologies are depicted based on production facil-ities for module manufacturing with very different production capacity and technological maturity.The respective overall impact on the environment is influenced decisively by economies of scale inproduction. There are only slight differences with respect to the impact on the environment for theexamined silicon-based PV systems (monocrystalline and multicrystalline). This analysis is based on dataconcerning module manufacturing with high production capacities (larger than 500 MWp). The analy-sis of CdTe-based systems is based on data from a manufacturer with stable production (capacity>500 MWp), which already ensures high-quality recycling of its modules today and a mid-sized manu-facturer, which is setting up a new production line (>50 MWp). The difference in the ecology index ofapprox. 23% between the two examined CdTe systems is due to the different state of technologicaldevelopment in laminate manufacturing. In addition, the differences are also a result of one of themanufacturers having already established high-quality PV module recycling. Within the context of theexamination, the CIS-based PV systems are analyzed based on two different module/laminate typesthat are produced in facilities with comparably small capacities (>100 and


6/76

2 Eco-efficiency analysis of photovoltaic modules

Eco-efficiency analysis of technologies based on the status quo

The energy payback timesof the examined PV systems, which range from 0.55 to 1.3 years, take into

account the total amount of energy used for manufacturing the systems and vary according to appli-cation types and technology. Accordingly, during their operating life all module types generate a mul-tiple of the energy required in their production. The values apply to the selected installation site, Nu-remberg, a central European site with good solar energy influx values. Better values would result fromsouthern European sites.

The overall environmental impact of generating electricity from photovoltaics is 10 to 20 times smallerthan the impact of fossil fuel electricity generation. However, it must be noted that "electricity fromphotovoltaics" and "electricity from non-renewable energy sources" differ in other aspects besidestheir environmental impact, e.g. the power variability of current photovoltaics. In this context, theenvironmentally-related differences between the individualPV technologies can be interpreted as

small. However, the composition of the ecology index still varies due to the fundamentally differentproduction processes of wafer-based and substrate-coated PV modules/laminates. Semiconductor andmodule manufacturing are decisive factors for the former, and the expenses in glass manufacturingare the focus for the latter.

When comparing the differentapplication types of PV systems, the large industrial rooftop applica-tion has the best ecology index assessment with comparably low feed-in tariffs. Within this applica-tion, CdTe modules have the least environmental impact, closely followed by monocrystalline siliconmodules or the examined CIS modules of smaller manufacturers. Ground-mounted systems representthe least expensive use for renewable electricity generation with photovoltaics (lowest EEG incentives).However, the overall environmental impact of ground-mounted systems is greater than that of large

industrial rooftop applications. The environmentally-related difference between monocrystalline siliconand CdTe modules decreases with ground-mounted systems. The BOS expenses related to ground-mounted applications3 is lower for silicon technologies. For small residential rooftops, as with largeindustrial rooftops, there are similarly higher feed-in tariffs and smaller differences in the overall envi-ronmental impact.

Environmentally-related adjustment options and potentials

From an environmental perspective, the BOSplays a significant role depending on the technology andthe application type in particular. The fact that the module technologies became increasingly moreefficient in the past has resulted in the environmentally-related contribution of the BOS becomingmore significant in the overall result. Correspondingly, the use of existing infrastructure is evaluated as

positive, e.g. attachment on existing slanted rooftops or the integration of PV systems into existingelectronic systems (use of available transformer capacities).

The potential of high-quality recyclingof PV modules is proving to be significant for the environment.In contrast to status quo recycling which is primarily aimed at the fulfillment of mass-related recyclingquotas, optimal (high-value) recycling of silicon modules for example can result in twice as many envi-ronmental benefits. Overall, the environmental impact can be reduced by approx. 10% through theestablishment of high-quality PV recycling.

3 BOS "Balance of System": all system peripherals required for system operation


7/76


An increase in efficiencywas found for all PV technologies from 2011 until 2012. This is primarilydue to the increase in efficiency, both in terms of module efficiency and economies of scale with theincreased throughput rate of production lines. Savings of approx. 9 to 17% on the ecology index areachieved in this way.

With regards to mid-term improvements,targeted efficiency increases in production (along with op-timization of thickness and quality of glass and reduction of plastic) will result in further improve-ments of the same magnitude as during 2011/2012. In terms of the forecasted development of effi-ciencies, an improvement potential of 23 to 33% is possible with the same amount of production ex-penses.

The specific annual yieldper application type and PV technology has a decisive influence on the envi-ronmental impact assessment. In addition to the different PV technologies, this assessment is also de-termined by distribution of production and specific features in the respective designof the overall

PV systems, e.g. the selection of the size and number of inverters or string distribution. The environ-mental impact would be probably reduced, especially for newer thin-film technologies of smaller man-ufacturers thanks to a more specific design of the total system related to PV technologies.

2 Project background and task

Photovoltaics represents an important component of sustainable energy supply and is contributingdecisively to Germanys energy transition. However, critics of solar energy technology still questionelectricity generation from PV modules from an environmental viewpoint.

With the study "Eco-efficiency Analysis of Photovoltaic Modules", bifa provides a comprehensive pic-ture of the environmental and economic impact that is related to the manufacture, use and disposal ofcurrently available PV technologies. The objectives of the study are:

1. Depiction of current production processes and the operating scenarios for selected PV moduletechnologies

2. Determination of the environmental and economic impact using the methodology of an eco-efficiency analysis

3. Identification of adjustment options for improving eco-efficiency in the comparison of differ-ent PV modules

4. Identification of future possibilities for further improvements in eco-efficiency, e.g. throughimproved processes for obtaining of raw materials, production or recycling

The studys results support the strategic further development of the PV industry. The great variety ofavailable PV module technologies cannot be depicted completely within the context of this


8/76


examination. Therefore, the following PV module technologies are the object of this eco-efficiencyanalysis:

a) Silicon-based technologies: monocrystalline and multicrystalline

b) Thin-film technologies: CdTe and CIS

The project was supported by a technical advisory board, whose representatives belong to companiesactive in the photovoltaics value chain as well as representatives of the Institute. Support was in thecontext of workshops for discussing the future course, assumptions and results as well as providingindustry expertise and current, basic data.

3 Methodological principles

The decisive components of the examination are an ecological assessment of the environmental im-pact, cost considerations and their combination within the context of an eco-efficiency analysis. Figure1 summarizes the steps for determining and considering the joint environmental and economic impact.

Eco-EfficiencyAnalysis

Cost considerationsEcological assessment

Life cycle inventory analysis

(elementaryflows)Life cycle impact assessment

(impactcategories)Combined environmental

impact (normalization,grouping,weighting

Costs balance sheet

Goal and scope definition

(systemboundaries,functionalunit,allocations)Data collection for all relevant processes

(quantityflows,emissiondata,operationaldata,costs)

Costs [/t]

Figure 1: Steps for considering the joint environmental and economic impact within the context of the eco-efficiency analysis


9/76


The ecological assessment is based on life cycle assessment standards and as a result provides individu-al values which characterize the environmental impact of the processes. The analysis of the environ-mental impact is depicted in the impact categories shown in Table 1.

Table 1: Selection of impact categories, allocation of environmental lifecycle inventory analysis to theindividual impact categories and unit of the impact indicator results

Impact category Impact Lifecycle inventory analysisparameters

Global warming impact[kg CO2equivalents]

Warming of the Earth's atmosphere CO2, CH4, N2O

Acidification[kg SO2equivalents]

Emission of acidifying substances NOXas NO2, SO2, NH3, (HCl, HF)

Terrestrial eutrophication

[kg PO43-equivalents]

Excessive nutrient feed in soil NOXas NO2, NH3

Photochemicalozone creation[kg C2H4equivalents]

Creation of ground-level ozone(summer smog)

CH4, NMVOC, VOC unspec. (benzene, for-maldehyde)

Human toxicity

[cases/kgemitted]

Toxic damage of people andorganisms

Air emissions: As, Pb, Cd, Cu, Ni, Hg, Zn

Water emissions: Cd, Cr(VI)

Eco-toxicity[PAFm3day/kg]

Toxic damage of organisms andeco-systems

Air emissions: As, Pb, Cd, Cu, Ni, Hg, Zn

Water emissions: Cd, Cr(VI)

Resource use[kg iron equivalents]

Consumption of primary fossil fuelsand metal from deposits

Petroleum, natural gas, coal and metals

PAF: potentially affected fraction of species

The environmental impacts were aggregated using a method developed by bifa. The principles of thismethod are that the procedure should be based on the life cycle assessment requirements of the Ger-man Federal Environment Agency (UBA) to the greatest extent possible. It should be transparent andcomprehensible and provide an overall value as result.4

The electricity generation prices are approximated for the examined PV systems using the feed-in tar-iffs in the German Renewable Energy Act (EEG).

The eco-efficiency analysis then provides the result of the life cycle assessment with specific costs for

electricity generation from PV systems (here: EEG feed-in tariffs).

4 Refer to Appendix A for detailed explanations about the methodological principles.


10/76


4 Notes about the scope

The study considers a total lifecycle composed of production, regular operation and subsequent dispos-al. Unforeseen (including illegal) incidents such as fires or unauthorized disposal etc. are not part ofthis study. The differently designed PV systems generate different amounts of current during theiroperating lives. These are standardized to the amount of current (per GWh) to improve comparison ofthe results. The data polling refers to the status quo (2011/2012) as well as future developments basedon that. The focus is placed on the manufacture and use in Germany. The examined application casesdiffer with respect to facility size, inclination angle of the PV modules to the sun as well as to othersystem components used such as inverters, flat rooftop mounting system and wiring of the PV facility.

5 Lifecycle modeling of PV systems

5.1 Lifecycle of a PV system

The analysis of the environmental impact covers the complete lifecycle of PV systems in various areasof use with the following lifecycle stages:

a) Production phase

Obtaining and refining raw materials producing starting materials

PV module manufacturing depending on the technology BOS5Manufacturing inverters, attachment components and wiring

PV system production installation of modules and BOS in the respective system

b) Utilization phase

The amount of current generated during the operating life time of a PV system

Solar energy influx

Degradation6of module efficiency

c) Disposal and recycling phase

The system boundary for the balancing of PV systems covers the obtaining of raw materials andproviding them for technical processes as well as the emission of elementary flows to the environmen-tal media of water, air and soil. Additional use from residues as well as from recycling of waste are

5 BOS "Balance of System [component]" (for an explanation of the term, cf. Chapter 5.2.3)

6

PV modules are subject to performance reduction over time due to aging of materials. The degradation must be consideredin calculating yield [PHOR 2013a].


11/76


calculated as equivalent functional utilization7. Provision and maintenance of infrastructure (construc-tion, service and repair of buildings, machine, industrial facilities, transport means and traffic routes)are not considered.

The production phase of a PV system has been examined for the following technologies:

a) Silicon-based PV modules: monocrystalline and multicrystalline

b) Thin-film modules: CdTe and CIS

The data for obtaining and refining raw materials were mainly taken from the Eco-Invent database[ECOI 2010].

5.2 System components

5.2.1 Silicon-based semiconductors

The base material for producing silicon-based solar cells in the form of quartz sand is available in suffi-cient quantity. According to different purification methods, so-called polysilicon is available in varyingquality for different areas of use (metallurgy, solar and semiconductor industries).

Monocrystalline and multicrystalline modules with respective power capacity of an average 245 Wpwere selected as reference products. The data for multicrystalline modules are based on informationfrom a manufacturer with global production capacity for monocrystalline and multicrystalline moduleswith each >500 MWp. The production of monocrystalline modules only differs with respect to the Sistarting material. A production process diagram is shown in Figure 2.

Figure 2: Overview of the production process for silicon-based PV modules.

7 The conventional manufacturing or production process of additional utilization is called an equivalent process or equivalent

system. For each quantifiable additional utilization, a specific equivalent system is modeled, which generates the same orequivalent functional utilization.


12/76


5.2.2 Thin-film technologies

Thin-film PV modules are based on semiconductors, which are placed directly on a substrate (mostly on

a glass pane).

5.2.2.1 Cadmium telluride technology

This technology uses cadmium telluride (CdTe) as semiconductor material, a crystalline compound ofthe metals cadmium and tellurium. Cadmium is a waste product of zinc, copper and lead productionand can only be used further in its metallic form to a very limited extent due to its toxic properties.CdTe is created via chemical refining with the semi-metal tellurium, which is extracted during copperand lead production from anode sludge and can be mined directly in a few mines worldwide. [CALY2010] Figure 3 provides an overview of the production process for CdTe PV modules.

Figure 3: Overview of the production process for CdTe PV modules

Two CdTe products are referenced for the lifecycle modeling: The data of CdTe module 1 are based oninformation from a producer with production capacities in Germany >500 MWp and that of CdTemodule 2 on information from a mid-sized company (capacity >50 MWp).

5.2.2.2 CIS technology

Copper-based thin-film modules are among the newest thin-film technologies marketed commercially.

"CIS" is an abbreviation for the group of compound semiconductors. The semiconductor layer is com-posed of copper (Cu), indium (In) and selenium (Se) as well as gallium (Ga) and sulfur (S). The basicproduction steps are shown in Figure 4.


13/76


Figure 4: Overview of the production process for CIS PV modules

Two CIS products are referenced for the lifecycle modeling: copper-indium-disulphide with a capacityof 125 Wp and copper-indium-gallium-diselenide with a capacity of 90 Wp per module. Both modulesare produced in a mid-sized company quantity (production capacities >100 and


14/76


Table 2: Overview of the system parameters for the three application cases of residential rooftop, largeindustrial rooftop and ground-mounted systems

Application Type Parameter Value

General Site Nuremberg, climate data record (1981-2000)

Directional alignment Directional alignment South

Module capacity Si: 245 Wp (mono) / 245 Wp (multi)

CdTe: 85 Wp / 80 Wp

CIS: 125 Wp / 90 Wp

Residentialrooftop (RTR)

System size 50 m

Installation angle 30

Attachment Configuration 30 rooftop (tiles)

Inverters Approx. 10 kWCable DC: from 30 to 70 m, AC: 5 m

Medium voltage power cable present

Transformer None, because part of the existing infrastructure

Large industrialrooftop (RTC)

System size 1,500 m


Attachment Configuration 15 rooftop (trapezoidal corrugatedsheets)

Inverters Approx. 250 kW

Cable DC: from 850 to 4,000 m, AC: 25 m

Medium voltage power cable present

Transformer None, because part of the existing infrastructure

Ground-mountedsystem (GM)

System size 5 MWp


Attachment Flat surface configuration advisory board

Inverters Approx. 10 kW

Cable DC: from 70,000 to 170,000 m,

AC medium voltage power cable: 5,000 mTransformer Yes, combined with inverter

Table 3 summarizes the detailed design for PV system modeling. The upper part shows the selectednumber of modules and the inverter design for the application case. The installed system capacity andthe return specifics are also shown. The two pieces of information were determined for the base year2011. The current revenues of the scenarios 2012 and mid-term improvements are derived via a pro-


15/76


jection of the module capacity9 correspondingly. The lower part of the figure summarizes the pieceweights of the attachment components, anchoring, enclosure and cable lengths.

Table 3: Overview of the (above) number of PV modules and inverter design for the accounted systemconfigurations with additional information about PV capacity and specific annual return from

PV*Sol; (below) piece weights of the attachment components, anchoring, enclosure and cable

lengths; all values are standardized to the reference unit in a second step (Standardization col-

umn)

Szenariom m pcs. pcs. pcs. pcs. pcs. pcs. kW MW MW kWp kWh/kWp

RTR_poly 50 50.3 30 8 7.35 1,001RTR_mono 50 50.3 30 8 7.35 996RTR_CdTe_1 50 46.8 65 6 5.53 1,050RTR_CdTe_2 50 46.8 65 6 5.20 1,016RTR_CIS_1 50 48.23 45 6 5.63 935RTR_CIS_2 50 46.35 56 5 5.04 916

m m pcs. pcs. pcs. pcs. pcs. pcs. kW MW MW kWp kWh/kWpRTC_poly 1,500 1,475 880 240 216 990

RTC_mono 1,500 1,475 880 240 216 964RTC_CdTe_1 1,500 1,426 1,980 165 168 1,017RTC_CdTe_2 1,500 1,426 1,980 150 158 982RTC_CIS_1 1,500 1,447 1,350 181 169 921RTC_CIS_2 1,500 1,449 1,750 168 158 880

MWp MWp pcs. pcs. pcs. pcs. pcs. pcs. kW MW MW kWp kWh/kWpGM_poly 5 5.304 21,648 4.80 4.80 5,304 1,086GM_mono 5 5.304 21,648 4.80 4.80 5,304 1,064GM_CdTe_1 5 5.310 62,496 4.56 4.56 5,312 1,047GM_CdTe_2 5 5.310 62,496 4.56 4.56 5,000 1,065

Inverter,1

0kW,scalable,for

PV(approx.40kg)

Inverter,1

MW,scalable,for

PV(approx.6t)

solarmodule,CIS(1)

InstalledP

Vcapacity

Specificannualyield

Transform

er,mediumvoltage,

1MW

(approx.14.5t)

solarmodule,Simono

solarmodule,Simulti

Normalization

Reference

1

solarlaminate,CdTe(2)

solarlaminateCdTe(1)

solarlaminate,CIS(2)

9Cf. parameter P in Table 5 (page 17)


16/76


Szenariom m kg kg kg kg kg m m m m m m

RTR_poly 50 51.92 48 59 4.6 0.1 7.4 50.3 70 5 0RTR_mono 50 51.92 48 59 4.6 0.1 7.4 50.3 70 5 0RTR_CdTe_1 50 49.17 67 78 11 0.1 8.5 46.8 110 5 0RTR_CdTe_2 50 49.17 67 78 11 0.1 8.5 46.8 110 5 0

RTR_CIS_1 50 51.39 48 61 3.2 0.1 7.8 48.2 130 5 0RTR_CIS_2 50 48.30 59 73 0.0 0.1 8.1 46.4 150 5 0m m kg kg kg kg kg m m m m m m

RTC_poly 1,500 1,523 0 1,012 131 1.8 33 1,475 840 25 0RTC_mono 1,500 1,523 0 1,012 131 1.8 33 1,475 840 25 0RTC_CdTe_1 1,500 1,498 7.4 2,226 335 4.0 0.2 1,426 3,960 25 0RTC_CdTe_2 1,500 1,498 7.4 2,226 335 4.0 0.2 1,426 3,960 25 0RTC_CIS_1 1,500 1,542 0 1,525 201 2.7 60 1,447 2,796 25 0RTC_CIS_2 1,500 1,510 6.3 1,975 530 3.6 0.2 1,449 3,180 25 0

MWp MWp t Fe kg Al kg PVC t concrete t Fe MWp km km kmGM_poly 5 1 38.3 340 706 76.3 4.9 5 4.6 66.3 30.0GM_mono 5 1 38.3 340 706 76.3 4.9 5 4.6 66.3 30.0GM_CdTe_1 5 1 91.8 1,136 529 43.4 4.8 5 14.8 103.0 30.7GM_CdTe_2 5 1 91.8 1,136 529 43.4 4.8 5 14.8 103.0 30.0

BOSaccessories,click-in

block,plastic

Screws,stainlessste

el,milled

Fence,forground-mo

unted

systems

Foundations/anchorin

gs,for

ground-mountedsystems

Cable,mediumvoltag

e,AC

Cable,directcurrent(DC)

Cable,alternatingcurrent(AC)

Reference2

Reference3

Roofhook,stainlesssteel,

welded

Supportsection,alum

inium,

extruded

Moduleclamp,alumin

ium,

drilled

Normalization

For example, if you compare the building component weights for attachment (Table 3, below, greenarea) in the application case of large industrial rooftop (RTC), you can see that approx. twice as manyaluminum support sections and more than twice as many module clamps and snap-on components areused for the ground-mounted, smaller thin-film modules. With respect to the limited area (1,500 m),however, facility capacity of only 165 kWp (Table 3, above, PV capacity) can be achieved with thin-filmmodules. The value is 216 kWp is for silicon modules, which is due to the higher efficiency of the mod-ules.

Another comparison illustrates the extra expense for installation of a ground-mounted system (GM)compared to installation on a large rooftop (RTC). Almost 9 kg of iron metal aluminum and plastic areused per installed kWp in the silicon-based ground-mounted system. The values for a large industrialrooftop (RTC) are only approx. 5.5 kg/kWp (60%). Almost 20 kg/kWp are used for the CdTe ground-mounted systems. The values for a large industrial rooftop (RTC) are only approx. 15 kg (79%) per in-stalled kWp. Compared to a rooftop construction, ground-mounted (GM) systems also have expensesfor anchoring the table with foundations. Finally, a statically sufficient attachment is required. Anexisting roof already provides this function in part.

Yield calculation

A simulation in the program PV*Solwas conducted for the yield calculation of the selected systemconfigurations. A centrally located site in Nuremberg, Bavaria was selected with intermediate but nev-

ertheless good solar energy influx values.


17/76


The specific annual yields calculated by PV*Solhave a degradation of 0.7% of the initial yield annuallydue to aging of the modules. This is considered in calculating the total current yield over the set termof 30 years. All calculated, specific annual yields are already summarized for the respective operatingscenarios and PV technologies in Table 3.

5.4 Calculation of costs

This study was conducted in the environment of very dynamic market developments. The actual systemcosts of electricity generation can only be approximated based on prices. The levelized costs of elec-tricity (LCOE) consequently resulted at the time of the valid EEG feed-in tariffs for the respective ap-plication types. There are essentially identical LCOE for all technologies, which is explained by the ori-entation of the consumer prices to the feed-in tariff.

With respect to consideration of economic aspects within the context of the eco-efficiency assess-

ment, the comparative discussion of electricity generation costs on a price-based LCOE basis is notconsidered meaningful for the various PV technologies. Consequently, the technologies were assessedeconomically according to EEG, relative to the feed-in tariff valid at the respective time. Table 4 dis-plays the feed-in tariffs for application types and time periods to be assessed [BNA 2013a, BNA 2013b].

Table 4: Feed-in tariffs according to EEG for selected time periods (change in %) [BNA 2013a, BNA 2013b]

Feed-in tariffs [ ct/kWh]according to

Dec2011

Jan2012

Jun2012

Dec2012

Rooftop, up to 30 kWp 28.74 -15% 24.43 -22.6% 18.92 -7.8% 17.45

Rooftop, starting from

100 kWp25.86 -15% 21.98 -27.2% 16.01 -7.7% 14.77

Ground-mounted, other 21.11 -15% 17.94 -27.0% 13.1 -7.8% 12.08


18/76


6 End of Life Disposal and Recycling

6.1 Current situation

Today, the disposal of PV modules is limited to defective modules or those damaged during transportor installation. This is due to the long operating life of PV modules. With a predicted operating life of30 years, larger quantities of PV modules will only have to be disposed of in future decades. On onehand, there is currently the overlapping take-back and recycling program PV-Cycle, to which approx.90% of all manufacturers on the European solar energy market belong [SONN 2012], and on the otherhand individual manufacturers voluntarily commit to taking-back and recycling modules themselves.The EU Directive 2012/19/EU10(WEEE Directive) also specifies the requirements placed on manufactur-

ers with respect to collection and recycling values.

The BOS components are mainly those from the construction and electronics industries, for which es-tablished take-back and recycling systems exist.

6.2 Overview of PV module disposal

When considering recycling end-of-life PV modules, it is important to recover the particularly highglass content (up to 90%) and metal contents in the product. The recycling of semiconductor materialsis only possible after complex processing. The study analyzes two scenarios of PV module disposal.

6.2.1 Status quo recycling scenario minimum recycling option

The current approach to the disposal of PV systems is the basis of the status quo recycling scenario. Forsilicon modules, this scenario already complies with the requirements foreseen in the WEEE Directive asof 2018. The processing techniques, which are already used today for PV silicon modules, enable almostcomplete recycling of glass and metal. After collection of silicon PV modules, the connection jacks andcables are separated from the module and recycled. The module is shredded, and the remaining frac-tions sorted and crushed. The main components glass, aluminum and copper are recycled. The glass isrecycled into glass fiber and glass insulation materials, and aluminum and copper are recycled as metal.There is no material recycling of silicon.

However, this type of photovoltaic module processing is not suitable for thin-film modules because theglass is contaminated with the semiconductor layer and cannot be recycled without removal of thesemiconductor. Because thin-film modules have only been installed over the past few years and thewaste quantity is very small, there are currently no commercial waste reclamation plants for this mod-ule category. The largest global supplier of CdTe modules is an exception in terms of the treatment ofmanufacturing/installation waste. It has operated its own recycling facility since 2005 and ensures freetake back and recycling of all of its modules.

10

Directive 2012/19/EU of the European Parliament and Council of July 4, 2012 concerning old electric and electronic devices(new version)


19/76


6.2.2 High-quality recycling scenario optimal recycling option

This scenario depicts the potential of high-quality recycling processes, which are already used today in

pilot plants and in industrial processes.The experiences of a leading module producer of Si-based modules are considered in the depiction ofthe disposal processes, which achieve high-quality recycling (of glass, aluminum, copper and brokensolar cells). The process starts with separating connection jacks and any remaining cables. The connec-tion jacks are collected as electronic scrap, and the cables are recycled. The module then undergoesthermal compound separation in the recycling furnace facility. The organic components in the moduleare completely converted in the gaseous phase. Following the thermal decomposition of the modulebond, the inorganic components are separated and sorted. A combination of several sorting units re-sults in the purity of the broken solar cell >99.99% [SOCY 2012].

The potential of the high-quality recycling processes for CdTe thin-film modules are considered using a

system already in operation on an industrial scale. The technology employed [FIRS 2009] enables high-quality recycling of glass, aluminum and copper as well as semiconductor materials, which are pro-cessed further into CdTe. According to information from the manufacturer, more than 95% of thesemiconductor material could be recycled in a metal-containing filter cake as well as approx. 90 to95% of the glass, already in high quality in part. Once the jacks and cables are separated, the modulesare fed into a dry process in which they are shredded and crushed (separation of the laminating filmfrom glass). A downstream wet process handles the removal of semiconductor layers as well as theseparation of solids (glass-laminate material separation) and liquids (precipitation and dehydration).Then the semiconductor materials from the filter cake are processed downstream.

The high-quality recycling options for CIS modules are examined based on data from a pilot plant of a

chemical company. The company uses a chemical combination for precipitation of the semiconductormaterials [LOSE 2013], in which the semiconductor bonds are detached in varying forms of purity (e.g.obtaining indium with 95% purity). [PVM 2011] Output materials of this process are high-quality flatglass, laminated film, metal and semiconductors, which can be processed again.


20/76


7 Results of the eco-efficiency analysis

The environmental impact for each technology examined in the study was determined using the meth-odology described in Chapter 3 (also refer to Appendix A). All required parameters and assumptions aredepicted in scenarios for creating the material flow model.

7.1 Overview of scenarios

The following tables summarize the aspects described in Chapters 5 and 6. Table 5 provides an over-view of the accounted application cases and the two recycling scenarios at the lifecycle end.

Table 5: Design application cases for the scenarios to be analyzed

Application according to market categories

Residential rooftop 50 m roof area, tile roof, directional alignment 30 south

Large industrial rooftop 1,500 m roof area, trapezoidal corrugated sheets,directional alignment 15 south,

Ground-mounted system 5 MWp installed facility capacity, flat surface with flat mounting,directional alignment 30 south

Recycling scenarios

Rec.sq Status quo recycling option = minimum recycling 1

Rec.opt Optimal recycling option = high-quality recycling

1 Exception: Recycling facility of leading CdTe manufacturer already in operation; consequently, high-quality recycling is therefore availa-ble for CdTe (1)

Table 6 shows an overview of the specific module features for the actual status of PV module produc-tion in 2011 (scenario 2011). The short-term improvements, which could already be achieved in 2012starting from the situation in 2011, are reflected in the 2012 scenario.

In two additional scenarios, the potential of two improvements expected in the mid-term are shown.On one hand, the possibility of reducing material use in module production through higher manufac-turing efficiency is considered. On the other hand, an outlook of the probable improvement of moduleefficiency is provided. Mid-term implementation within the next five years can be considered probablefor both scenarios. In the sense of a conservative consideration, both scenarios are considered sepa-rately and not combined.


21/76


Table 6: Scenario overview and important parameters

Scenarios

2011 2012

Mid-term

improvementsProduction efficiency

Mid-term

developmentsModule efficiency*

Module production

Silicon mono monocrystalline:P: 245 WpA: 1.68 m

P: 265 Wp material saving:- cover glass: 25%substitution:- lead: completely- silver: 50% pro rata

P: 350 Wp

Silicon multi multicrystalline:P: 245 WpA: 1.68 m

P: 255 Wp material savings:- cover glass: 25%substitution:

- lead: completely- silver: 50% pro rata

P: 340 Wp

CdTe (1) CdTe:P: 85 WpA: 0.72 m

P: 95 Wp material savings:- substrate glass: 25%

P: 125 Wp

CdTe (2) CdTe:P: 80 WpA: 0.72 m

P: 85 Wp material saving:- substrate glass: 25%

P: 120 Wp

CIS (1) CIGS:P: 125 WpA: 1.09 m

P: 125 Wp,new design

material saving:- substrate glass: 25%

P: 185 Wp

CIS (2) CIGSe:P: 90 WpA: 0.81 m

no information no information no information

Material and energyutilization(for all technologies)

module-specific no change material saving:- metal: 1%- plastic: 5%energy saving:- electricity: 5%

no change

Balance of System

Roof hooks, screws made ofstainless steel

no change made of individuallygalvanized steel

no change

PV cable(DC/AC) 100% copper no change 80% copper,20% aluminum no change

Medium voltagepower cable:

100% copper no change 20% copper,80% aluminum

no change

P: Rated current of the PV modules

A: PV module area (outer dimensions)

* Expected capacity due to efficiency increases


22/76


7.2 Explanations concerning the depiction of results

Depictions were employed for assessing the scenarios to be compared as explained in Figure 5.

-2

-1

0

1

2

3

4

5

6

Silicon

mono

Silicon

multi

CdTe(1)

CdTe(2)

CIS(1

)

CIS(2)

[2011]

NETTO

BOS recycling

module recycling

inverter

wiring

attachment

frame

module/laminate

semiconductor

glass

Figure 5: Explanation of the chart for assessing the contributions of the examined scenarios for the con-sidered impact categories and individual parameters

The life cycle assessment provides three sets of results. The respective bars on the left of Figure 5 showthe gross expense results (environmental impact bar upward) on one hand and on the other hand thecredits (environmental benefits bar downward). The sectoral display in color-coded sections makes it

possible to identify the sub-systems (sections) with relevant contributions to the overall result.1. The PV modules are thereby composed of the following sections:

a) Glass = production process from raw materials

b) Semiconductors = production process from raw materials

c) Module/laminate = solar module/laminate production with upstream chains for raw materials,auxiliary materials, consumables, energy consumption and process emissions

d) Frames = production process from raw materials

Consequently, the sum of the items listed above corresponds to the production expenses for the solar

modules or solar laminates. In other words, these represent the environmental impact from a manufac-turer's viewpoint, which are directly involved (analog to the approach used in various manufacturers'balance sheets). The expenses for the BOS represent an indirect environmental impact for a PV modulefrom a manufacturer's viewpoint.

2. The expenses for the BOS are composed as follows:

a) Attachment = provision of raw materials and production expenses for roof hoods, support sec-tions, module clamps and screws; additionally for ground-mounted systems: for poles, founda-tion and enclosures

b) Wiring = provision of raw materials and production expenses for direct, alternating and medi-um voltage power cables


23/76


c) Inverters = provision of raw materials and production expenses for inverters; additionally forground-mounted systems: for transformers

3. A distinction is made according to module and the BOS components for recycling (cf. 1. and 2.)and both the

a) expenses of the preparation process and

b) the resulting credits

are displayed.

The net result for the considered scenario results from the offsetting of the environmental impact andbenefit, which is displayed respectively in the gray bar on the right in Figure 5. It shows whether theenvironment suffers (bar upward) or sustains benefits (bar downward) due to the contribution of thescenario.

In the following, all results are presented and discussed based on the display of the ecology indexpoints, standardized to the PV electricity produced in GWh. The ecology index is produced by combin-ing all seven, analyzed individual impact categories (Table 1). The results of this are shown in AppendixB.

7.3 Comparison of Si-based PV systems

7.3.1 Technology 2012 and recycling scenarios

The environmental impact from manufacture, utilization and recycling of Si-based PV systems (mono-

crystalline and multicrystalline) is displayed for the scenario 2012 below. The manufacturers have sys-tems with production capacity of more than 500 MWp. The two options status quo recycling (Rec.sq)and optimal recycling (Rec.opt) are compared for the recycling lifecycle phase described in Chapter 6.

Figure 6 displays the ecology index as an overall environmental score relating to the production ofGWh electricity on a residential roof for the Si-based PV systems.

The ecology index in the scenario with status quo recycling (Rec.sq) is 2.5 points for monocrystallinesilicon and 2.6 points for multicrystalline silicon. The ecology index in the scenario with high-qualityrecycling (Rec.opt) is 2.3 points for monocrystalline silicon and 2.4 points for multicrystalline silicon.


24/76


Ecology Index

[Ecology Index points / GWh]

2,5

2,3 2

,6

2,4

-2

-1

0

1

2

3

4

2012

Rec.sq

NET

2012

Rec.opt

NET

2012

Rec.sq

NET

2012

Rec.opt

NET

Silicon mono Silicon multi

NET

BOS recycling

module recycling

inverter

w iring

attachment

frame

module/laminate

semiconductor

glass

Figure 6: Contributions of the sections of the monocrystalline silicon and multicrystalline silicon scenariosto the ecology index (module type: silicon, application case: residential rooftop)

Figure 6 shows that there are only slight differences between the Si-based technologies with respectto the environmental impact. Optimal recycling provides the potential to reduce the overall environ-mental impact by approx. 8% compared to status quo recycling for both Si technologies.

7.3.2 Short- and mid-term improvements in efficiency

This chapter analyzes the environmental benefits Si-based PV systems achieved thanks to efficiencyincreases from 2011 to 2012. In addition, it considers which mid-term improvements over the next fiveyears are to be expected for Si-based PV systems (cf. Table 6). The production efficiency scenario refersto the conservation of raw materials in the production of modules and the BOS, for example. Themodule efficiency scenario refers to the expected power increase of the produced modules. The devel-opment of the considered technologies is displayed based on residential rooftop installations.


25/76


Ecology Index


2,9

2,5

2,2

1,9

2,9

2,6

2,3

2,0

-2

-1

0

1

2

3

4

5

2011NET

2012NET

Prod.Eff.NET

Mod.Eff.NET

2011NET

2012NET

Prod.Eff.NET

Mod.Eff.NET


NET

BOS recycling

module recycling

inverter

w iring

attachment

frame

module/laminate

semiconductor

glass

Figure 7: Contributions of the sections of the 2011, 2012, production and module efficiency scenarios tothe ecology index for monocrystalline and multicrystalline modules (module type: silicon, applica-tion case: residential rooftop)

Figure 7 shows that the efficiency increases already implemented in 2011/2012 and the mid-term im-provements in particular have substantial potential for reducing the environmental impact. For exam-

ple, the overall environmental impact could be reduced by an average of approx. 12% in 2011/2012.

Mid-term improvements provide the potential to reduce the environmental impact compared to 2011by more than one-third. The potential is mainly in module manufacturing (improved production tech-nology, increased efficiency, material savings and substitution).

7.3.3 Application types

Figure 8 shows the environmental impacts quantified in the ecology index of Si-based PV systems forthe three application cases of residential rooftop, large industrial rooftop and ground-mounted sys-tems

The ecology index in the residential rooftop scenario is 2.5 and 2.6 points for monocrystalline and mul-ticrystalline silicon respectively. The ecology index in the large industrial rooftop scenario is 2.0 pointsand for ground-mounted systems 2.8 and 2.9 points respectively. Consequently, the different applica-tion types have substantial differences in their environmental impact.


26/76


Ecology Index 2012


2,5

2,0

2,8

2,6

2,0

2,9

-3

-2

-1

0

1

2

3

4

56

RTR

NET

RTC

NET

GM

NET

RTR

NET

RTC

NET

GMNET


NET

BOS recycling

module recycling

inverter

w iring

attachment

frame

module/laminate

semiconductor

glass

Figure 8: Contributions of the sections of the scenarios to the ecology index; comparison of the applicationtypes for monocrystalline and multicrystalline modules (module type: silicon, application case:residential rooftop, large industrial rooftop and ground-mounted systems)

The comparison shows that the large industrial rooftop scenario has the lowest ecology index for bothSi-based technologies and as a result has the lowest environmental impact. The residential rooftop

scenario has a greater environmental impact, which results from the contributions of the attachmentand wiring sections.

The ground-mounted system scenario has double the environmental impact compared to the largeindustrial rooftop. The greater environmental impact results from the substantially larger materialquantities, which are required for the attachment components and wiring of ground-mounted systems(attachment and wiring sections). This is especially expressed in the impact categories of resource utili-zation and human toxicity. With respect to human toxicity, the substantial use of galvanized steel forthe BOS has considerable impact on ground-mounted systems.

7.4 Comparison of CdTe PV systems

7.4.1 Technology 2012 and recycling scenarios

The environmental impact from the manufacture, utilization and recycling of CdTe PV systems (mono-crystalline and multicrystalline) are displayed below for the 2012 scenario. The manufacturers havesystems with capacities of >500 MWp CdTe (1) and >50 MWp CdTe (2). The status quo recycling(Rec.sq) and high-quality recycling (Rec.opt) options are compared for the recycling lifecycle phasedescribed in Chapter 6.

Figure 9 displays the ecology index as an overall environmental score relating to the production ofGWh electricity on a residential roof for CdTe PV systems.


27/76


Ecology Index[Ecology Index points / GWh]

2,0

2,0

2,6

2,3

-2

-1

0

1

2

3

4

2012

Rec.sqNET

2012

Rec.optNET

2012

Rec.sqNET

2012

Rec.optNET

CdTe (1) CdTe (2)

NET

BOS recycling

module recycling

inverter

w iring

attachment

frame

module/laminate

semiconductor

glass

Figure 9: Contributions of the sections of the CdTe(1) and CdTe(2) scenarios to the ecology index (moduletype: CdTe, application type: residential rooftop)

For CdTe (1), both recycling scenarios result in an identical environmental impact because the CdTemodule manufacturer, whose information is the basis for the CdTe (1) system, has already implementeda high-quality recycling process, the Rec.sq scenario also considers high-quality recycling that goes

beyond the quota specifications of WEEE (also compare Chapter 6.2.1).

The ecology index for CdTe (1) is 2.0 and for CdTe (2) 2.6 points in a comparison of the Rec.sq scenari-os. The difference in the ecology index of approx. 23% between the two examined CdTe systems is dueto the different state of technological development in laminate manufacturing. CdTe (1) is based oninformation of one manufacturer with stable production, while CdTe (2) is based on the situation of amid-sized manufacturer, which is setting up a new production line. The differences in the environmen-tal impact also result from the establishment of module recycling for CdTe (1). With high-quality recy-cling for CdTe (2) (Rec.opt scenario), the overall environmental impact is reduced by 0.3 ecology indexpoints or 12% compared to the Rec.sq scenario.

7.4.2 Short- and mid-term improvements in efficiency and degree of effectiveness

This chapter analyzes the environmental benefits CdTe PV systems achieved thanks to efficiency in-creases from 2011 to 2012. In addition, it considers which mid-term improvements are to be expectedfor CdTe PV systems over the next five years (cf. Table 6). The production efficiency scenario describesthe conservation of raw materials in the production of modules and the BOS, for example. The moduleefficiency scenario describes the expected power increase of the produced modules. The developmentof the considered technologies is displayed based on residential roof installations.

Figure 10 shows that the efficiency increases already implemented in 2011/2012 and the mid-termimprovements in particular have substantial potential for reducing the environmental impact. For ex-

ample, the overall environmental impact could be reduced by 9% for CdTe (1) and 10% for CdTe (2) in


28/76


2011/12. The mid-term improvements provide the potential to reduce the environmental impact com-pared to 2011 by 32% for CdTe (1) and 34% for CdTe (2). The potential is mainly in module manufac-turing (improved production technology, increased efficiency, material savings and substitution) andthe substitution of materials such as copper and stainless steel in the BOS.

Ecology Index[Ecology Index points / GWh]

2,2

2,0

1,5

1,5

2,9

2,6

2,0

1,9

-2

-1

0

1

2

3

4

5

2011

NET

2012

NET

Prod.Eff.

NET

Mod.Eff.

NET

2011

NET

2012

NET

Prod.Eff.

NET

Mod.Eff.

NET

CdTe (1) CdTe (2)

NET

BOS recycling

module recycling

inverter

w iring

attachment

framemodule/laminate

semiconductor

glass

Figure 10: Contributions of the sections of the 2011, 2012, module and power efficiency scenarios to theecology index for CdTe systems (module type: CdTe, application type: residential rooftop)


Figure 11 shows the environmental impacts quantified in the ecology index of CdTe PV systems for thethree application types of residential rooftop, large industrial rooftop and ground-mounted systems.The 2012 ecology index for CdTe (1) in the residential rooftop scenario is 2.0 points, 1.4 for large in-dustrial rooftop and 2.5 for ground-mounted systems. The 2012 ecology index for CdTe (2) is at thesame order of scenarios with a somewhat higher environmental impact.


29/76


Ecology Index 2012[Ecology Index points / GWh]

2,0

1,4

2,5 2

,6

1,9

2,9

-4

-3

-2

-1

0

1

2

3

4

56

RTR

NET

RTC

NET

GM

NET

RTR

NET

RTC

NET

GM

NET

CdTe (1) CdTe (2)

NET

BOS recycling

module recycling

inverter

w iring

attachment

frame

module/laminate

semiconductor

glass

Figure 11: Contributions of the sections of the scenarios to the ecology index; comparison of the applicationcases for CdTe (1) and CdTe (2) (module type: CdTe, application case: residential rooftop, large in-dustrial rooftop and ground-mounted systems)

The comparison shows that the large industrial rooftop scenario has the lowest ecology index for bothCdTe technologies and as a result has the lowest environmental impact. The residential rooftop scenar-

io has a greater environmental impact, due to higher contributions from the attachment wiring, in-verters and BOS recycling sections.

The ground-mounted system scenario shows an increase of the environmental impact between 34%and 44% compared to large industrial rooftop. The greater environmental impact results from the sub-stantially larger quantities of materials which are required for the attachment components and wiringof ground-mounted systems.

7.5 Comparison of CIS PV systems

7.5.1 Technology 2012 and recycling scenariosThe environmental impact from the manufacture, utilization and recycling of CIS PV systems are de-scribed below. The 2012 scenario is considered for CIS (1); only the data for the 2011 scenario is avail-able for CIS (2). The manufacturers have systems with capacities of >100 MWp CIS (1) and


30/76


rails used on the back while CIS (2) is a frameless laminate. The ecology index in the scenario withhigh-quality recycling (Rec.opt) is reduced for CIS (1) by 17% to 2.5 points and for CIS (2) by 18% to2.8 points

kologie-Index[kologie-Index Punkte / GWh]

3,0

2,5

3,4

2,8

-3

-2

-10

1

2

3

4

5

6

2012

Rec.sq

NET

2012

Rec.opt

NET

2011

Rec.sq

NET

2011

Rec.opt

NET

CIS (1) CIS (2)

NET

BOS recycling

module recycling

inverter

w iring

attachment

frame

module/laminate

semiconductor

glass

Figure 12: Contributions of the sections of the CIS (1) and CIS (2) scenarios to the ecology index (module

type: CIS, application case: residential rooftop)

Figure 12 shows that there are only slight differences between CIS (1) and CIS (2) in terms of environ-mental impacts when considering the same recycling scenarios. However, there are differences in thesectoral composition of the environmental impact. The attachment rails made of galvanized steel(frame section) in CIS (1) result in an additional environmental impact compared to CIS (2). On theother hand, the glass section in CIS (2) has a substantially higher environmental impact compared toCIS (1).

7.5.2 Short- and mid-term improvements in efficiency and degree of effectiveness

This chapter analyzes the environmental benefits CIS PV systems achieved through efficiency increasesfrom 2011 to 2012. In addition, it considers which mid-term improvements are to be expected for CISPV systems over the next five years (cf. Table 6). The production efficiency scenario describes the sav-ings of raw materials in the production of modules and the BOS, for example. The module efficiencyscenario describes the expected power increase of the produced modules. The development of the con-sidered technologies is displayed based on residential rooftop installations. Corresponding data foranalysis of the development are only available for the PV system CIS (1).


31/76


kologie-Index[kologie-Index Punkte / GWh]

3,6

3,0

2,5

2,0

3,4

-2

-1

0

1

2

3

4

56

2011NET

2012NET

EffizienzNET

WGNET

2011NET

2012NET

EffizienzNET

WGNET

CIS (1) CIS (2)

NET

BOS recycling

module recycling

inverter

w iring

attachment

frame

module/laminate

semiconductor

glass

Figure 13: Contributions of the sections of the 2011, 2012, production and module efficiency scenarios tothe ecology index for CIS (1) systems and the 2011 scenario for CIS (2) systems (module type: CIS,application type: residential rooftop)

The ecology index for CIS (1) in the 2011 scenario is 3.6 points and 3.0 points in the 2012 scenario. Itwas possible to substantially reduce the environmental impact of module manufacturing (mod-

ule/laminate section), semiconductor manufacturing and glass through new product design and in-creased energy efficiency (2012 scenario).

Figure 13 shows that the implementation of mid-term improvements provide the possibility to reducethe environmental impact of CIS (1) compared to 2012 by 30% (production efficiency scenario) and44% (module efficiency scenario). The decisive potential is in the expected increase of module power.


Figure 14 shows the environmental impacts quantified in the ecology index of CIS PV systems for theapplication types of residential rooftops and large industrial rooftops.

The ecology index 2012 for CIS (1) in the residential rooftop scenario is 3.0 points and 2.3 for the largeindustrial rooftop scenario. The 2011 ecology index for CIS (2) results in the same order of scenarios.


32/76


kolog ie-Index 2012/11[kologie-Index Punkte / GWh]

3,0

2,3

3,4

2,5

-2

-1

0

1

2

3

4

56

RTR

NET

RTC

NET

GM

NET

RTR

NET

RTC

NET

GM

NET

CIS (1) CIS (2) [2011]

NET

BOS recycling

module recycling

converter

w iring

attachment

frame

module/laminate

semiconductor

glass

Figure 14: Contributions of the sections of the scenarios to the ecology index; comparison of the applicationtypes for CIS (1) and CIS (2) (module type: CIS, application type: residential rooftop and large in-dustrial rooftop)

The comparison shows that the large industrial rooftop scenario has the lowest ecology index for bothCIS technologies and as a result has the lowest environmental impact. The residential rooftop scenario

has a greater environmental impact, due to higher contributions from the attachment wiring, invertersand BOS recycling sections.

7.6 Overview of PV technologies

The previously considered separate results for PV technologies are displayed in an overview below. Toillustrate the relevance of the displayed result differences of the individual PV technologies, Figure 15shows a comparison of the electricity generation of the PV systems using the residential rooftop ex-ample with that of fossil fuels (Figure 15).11

11Data basis: Eco-Invent result data records "Electricity, [], at power plant [DE]" [ECOI 2010]


33/76


Ecology-Index 2012[Ecology Index points / GWh]

0

10

20

30

40

50

60

70

brown

coal

hard

coal

gas o

il

Silicon

mono

CdTe(1)

CIS(1)

Non-renewable sources Residential rooftop

electricity mix UTCE

Figure 15: Ecology index for electricity generation from the three PV technologies Si-mono, CdTe (1) andCIS (1) on a residential rooftop compared to electricity generation from fossil fuels (brown coal,coal, gas and oil)

Considering all application types, the overall environmental impact with electricity generation from

photovoltaics is 10 to 20 times less than to the impacts from conventional fossil fuel electricity gener-ation. The differences shown in the following chapters are to be interpreted against this background.However, it must be noted that the compared products "electricity from photovoltaics" and "electricityfrom non-regenerative energy sources" differ in several ways other than lifecycle impacts. An examplewould be the power variability of current photovoltaic technologies.

7.6.1 Environmental impact of PV technologies 2012

This chapter describes the overall environmental impacts of all examined PV technologies for the 2012scenario with respect to manufacturing, utilization and recycling. For end-of-life recycling, both thestatus quo12(Rec.sq) and high-quality recycling options (Rec.opt) were selected. The considered tech-

nologies are displayed based on residential rooftop installations.

Figure 16 displays the ecology index for all three technologies. It should be noted that the overview oftechnologies is based on production facilities with very different capacities (500 MWp)and technological maturity. The different overall impacts on the environment are influenced decisivelyby economies of scale in production.

12

Exception CdTe (1): because a module recycling system is already in operation, high-value recycling can already be assumedhere.


34/76


For the status quo recycling option (upper part of the display), the ecology index is close with mono-crystalline and multicrystalline silicon at 2.5, CdTe (1) and (2) between 2.0 and 2.6 and for CIS technol-ogies between 3.0 and 3.4 points.

The environmental impact of monocrystalline and multicrystalline silicon is essentially due to the com-parably high energy use for semiconductor production and semiconductor processing in module manu-facturing in particular (share of the module section: 30%). In addition, the Si-based modules haveframe parts. Overall, approx. 49% of the environmental impact from Si-based PV systems is due to themodule manufacturing section (includes all steps of raw material, wafer, cell and module manufactur-ing including their preliminary products).

Both thin-film technologies CdTe and CIS scenarios have considerably less environmental impactdue to module manufacturing, but have higher glass content and greater BOS requirements. The shareof PV modules on the ecology index is approx. 23% for the CdTe (1) and (2) scenarios. The impact due

to the BOS has an approx. 60% share. The CIS (1) module is supplied with mounting rails (frame sec-tion), similar to the attachment frames for the silicon modules. PV module manufacturing has a shareof environment impact between 30 and 35% for both CIS technologies.

The status quo recycling option (Rec.sq) reduces the environmental impact of all PV technologies by0.8 to 1.5 ecology index points. Recycling reduces the environmental impact by approx. 25% formonocrystalline and multicrystalline silicon, more than 28% for CdTe (2) and up to 29 and 33% forCIS (2) and CIS (1). The share of CdTe recycling for CdTe (1) is 38% due to the already established,comprehensive high-value recycling.

The environmental benefits of high-quality recycling for all PV technologies are illustrated in the lowerpart of Figure 16. With the high-quality recycling option (Rec.opt), the environmental impact is re-

duced by 31% for monocrystalline and multicrystalline silicon, 38% for CdTe (1) and (2) and 43 to 45%for CIS (2) and CIS (1). The ecology index scores are correspondingly similar for the examined PV tech-nologies and have values from 2.0 to 2.8.


35/76


Ecology Index 2012 - Status quo Recycling (Rec.sq)


2,5 2

,6

2,0 2,6 3

,0 3,4

-3

-2

-1

0

1

2

3

4

56

Silicon

mono

Silicon

multi

CdTe(1)

CdTe(2)

CIS (1)

CIS(2)

[2011]

Ecology Index 2012 - High-quality Recycling (Rec.opt)


2,3 2

,4

2,0 2

,3 2,5 2

,8

-3

-2

-1

0

1

2

3

4

5

6

Silicon

mono

Silicon

multi

CdTe(1)

CdTe(2)

CIS (1)

C

IS(2)

[2011]

2,52,6

2,0

2,63,03,4

SilmSilmCd(1)Cd(2)CI(1)CI[2

NET

BOS recyclingmodule recycling

inverter

wiring

attachment

frame

module/laminate

semiconductor

glass

Figure 16: Contributions of the sections of the 2012 scenarios for monocrystalline and multicrystalline sili-con, CdTe (1) and CdTe (2) as well as CIS (1) and (2) to the ecology index (application type: resi-dential rooftop); above: recycling option status quo; below: optimal recycling option

> 500 MW > 500 MW > 500 MW > 50 MW > 100 MW < 25 MW Production capacity DE



36/76


In the comparison of PV technologies for residential rooftop installations, it can be seen that:

The differences between the overall environmental impacts of the examined PV technologies are

only slight.

However, the composition of the ecology index varies in this context. This is due to the essentially

different production processes for wafer-based (cf. semiconductor and module sections) and sub-

strate-coated PV modules/laminates (cf. the glass section).

The efficiency of PV modules/laminates (capacity per area) has considerable impact on the assess-

ment of the BOS components (substructure, wiring, inverter size per area) (cf. the attachment and

wiring sections).

In addition, the different state of technical development in comparison of established manufactur-

ers with solid production and mid-sized manufacturers with relatively new production lines has an

effect. The significant environmental benefits of high-quality recycling demonstrates the potential that

can be achieved through the successful implementation of integrated and technologically advanced

take-back and recycling systems.

7.6.2 Environmental impact of PV technologies mid-term improvements

Analog to the depiction in Chapter 7.6.2 (System installation in 2012), an overview of PV technologiesfor the scenarios with mid-term efficiency improvements (with the same production expenses) is de-scribed (Figure 18). Compared to the consideration of system installation for 2012, almost all PV tech-

nologies are approximately the same with an ecology index of approx. 2 (Rec.sq) and 1.7 points(Rec.opt) respectively. There is an ecology index of 1.5 points for both CdTe (1) recycling options, whichis due to the already established high-quality recycling process in the status quo consideration.


37/76


Ecology Index 'Module efficiency improvement' - Rec.sq


1,9 2

,0

1,5 1,9 2

,0

-3

-2

-1

0

1

2

3

4

Silicon

mono

Silicon

multi

CdTe(1)

CdTe(2)

CIS (1)

CIS (2)

Ecology Index 'Module efficiency improvement' - Rec.opt


1,8

1,8

1,5 1

,6 1,7

-3

-2

-1

0

1

2

3

4

S

ilicon

mono

S

ilicon

multi

CdTe(1)

CdTe(2)

CIS (1)

CIS (2)

2,52,6

2,0

2,63,03,4

-

3

-

2

-

1

0

1

2

3

4

5

6


NET


inverter

wiring

attachment

frame

module/laminate

semiconductor

glass

Figure 17: Contributions of the sections of the mid-term improvement scenarios for monocrystalline andmulticrystalline silicon, CdTe (1) and CdTe (2) as well as CIS (1) and (2) to the ecology index (appli-cation type: residential rooftop); above: recycling option status quo; below: optimal recycling op-tion

7.6.3 Eco-efficiency of technologies

The eco-efficiency analysis compares the result of the life cycle assessment for the 2012 scenario withthe feed-in tariffs according to EEG (as of December 2012) for the generation of electricity from pho-tovoltaic systems. The specific EEG feed-in tariffs are summarized in Table 4. The life cycle assessment




38/76


(ecology index points) of the PV technologies in the portfolio corresponds to the results from Chapters7.3 ff.

Processes with low eco-efficiency (high price and high ecology indexes, i.e., greater environmentalimpact) can be found on the lower left in the portfolio; processes with high eco-efficiency (low priceand low ecology indexes, i.e., less environmental impact) can be found on the upper right.

Figure 18 displays the findings of Figure 15 as a portfolio graphic, using the example of a residentialrooftop installation. The differences between the individual PV technologies seem marginal with re-spect to the higher ecology indexes of electricity generation from non-renewable sources. The ecologyindex points for electricity generation from fossil fuels such as gas, coal and oil were determined basedon the Eco-Invent result data records "Electricity, [], at power plant [DE]" [ECOI 2010]. The corre-sponding classification of EEX prices13 is based on [FfE 2010]. The restrictions cited in Chapter 7.5.3with respect to electricity-generating technologies also apply here. The arrangement of the price axis

only shows the situation at a single point in time, because the feed-in tariffs for PV systems have beenadjusted downward monthly since April 1, 2012. The tendered prices for electricity from non-renewable energy sources depend on the respectively valid primary energy prices as well as the pricesfor the required CO2certificates.

13The European Energy Exchange (EEX) with headquarters in Leipzig is a marketplace for energy and energy-related products.

The EEX is the leading energy market in continental Europe with more than 200 exchange participants from currently 22countries. [WIKI 2013]


39/76


brown coal

hard Coal

gas

oil

residentialrooftop

largeindustrialrooftop

photovoltaics

ground-mountedsystem

-10

0

10

20

30

40

50

600102030

Ecology-Indexpoints

EEG/EEX-prices [ ct/kWh]

Environ-

mental

burdens

highercosts

lowercosts

lowerEco-efficiency

higherEco-efficiencyEn

viron-

mental

benefits

Figure 18: Eco-efficiency portfolio for three PV technologies for residential rooftops, large industrial rooftopsand ground-mounted systems and basic classification of electricity generation from non-regenerative sources (a lower ecology index signifies less environmental impact; a high ecologyindex signifies greater environmental impact; cost index: Incentives for PV systems according to

EEG (12.2012) and EEX prices for non-renewable energy sources [FfE 2010]). Reference unit: Gen-eration of 1 GWh electricity.

Figure 19 illustrates the detailed results (see the red border in Figure 18) for the examined PV technol-ogies in all three application types.


40/76


a

b

c

d

e

f

g

h

-1

0

1

2

3

4

5

101214161820

Ecology-Indexpoints

EEG incentive [ ct/kWh]

Environ-

mental

burdens

highercosts

lowercosts

higherEco-efficiencyE

nviron-

mental

benefits

lowerEco-efficiency

(a) residential rooftop Si mono

(b) residential rooftop CdTe(1)

(c) residential rooftop CIS(1)

(d) industrial rooftop Si mono

(e) industrial rooftop CdTe(1)

(f) industrial rooftop CIS(1)

(g) ground-mounted system Si mono

(h) ground-mounted system CdTe(1)

a

b

c

d

e

f

g

h

-1

0

1

2

3

4

5

101214161820

Ecology-Indexpoints

EEG incentive [ ct/kWh]

Environ-

men

tal

burd

ens

highercosts

lowercosts

higherEco-efficiencyE

nviron-

mental

benefits

lowerEco-efficiency

Figure 19: Eco-efficiency portfolio for three PV technologies Si-mono, CdTe (1) and CIS (1) for residentialrooftops, large industrial rooftops and ground-mounted systems, above: with status quo recyclingoption; below: with optimal recycling option (a lower ecology index signifies less environmentalimpact; a high ecology index signifies greater environmental impact; cost index: Incentives for PVsystems according to EEG, as of: Dec. 2012). Reference unit: Generation of 1 GWh electricity.


41/76


Large industrial rooftop installations have the best ecology index with comparably low feed-in tariffs.Within this application, CdTe (1) modules have the least environmental impact, closely followed bymonocrystalline silicon module and CIS modules.

Although ground-mounted systems for renewable electricity generation with photovoltaic representsthe most inexpensive application (lowest EEG incentive), the overall environmental impact is higherthan that of installations on industrial rooftops. The environmentally-related difference in the use ofmonocrystalline silicon modules decreases compared to CdTe (1) modules for ground-mounted systems.The expenses for the BOS related to an area are fewer for the silicon technologies. Compared toground-mounted systems, almost all other PV application types result in smaller environmental impact.For the area of small residential roofs, there are similarly smaller differences with higher EEG incentivesin the overall environmental impact compared to large industrial rooftops.

Direct cost information cannot be derived for the PV technologies from the eco-efficiency portfolio, as

explained in Chapter 5.4. The economic classification is identical for all technologies and is made ac-cording to the specified classification of the application types in EEG: "rooftop, up to 30 kWp", "roof-top, starting from 100 kWp" and " ground-mounted, other ".

7.6.4 Sensitivity consideration for the specific annual yield

The specific annual yield in kWh/kWp assumed per application type and PV technology has a decisiveinfluence on the environmental impact assessment. In addition to the different PV technologies, thisassessment is also determined by distribution of production and specific features in the respectivedesign of the overall PV systems, e.g. selection of the size and number of inverters or string distribution(Table 3). The demonstrated environmental impact would be probably reduced, especially for newer

thin-film technologies of smaller manufacturers, thanks to a more specific design of the total systemrelated to PV technologies. To estimate this influence, a specific annual yield of 1,050 kWh/kWp (cf.Table 3 for RTR CdTe (1)) is assumed in the following sensitive consideration independent of PV tech-nology for residential rooftop installations. Figure 20 displays the results for this marginal considera-tion, which assumes that similarly high annual yields can also be achieved for new technologiesthrough improved system design.


42/76


Ecology Index 2012 - Status quo Recycling (Rec.sq)


2,4 2

,5

2,0 2,5 2

,7 3,0

-3

-2

-1

0

1

2

3

4

5

Silicon

mono

Silicon

multi

CdTe(1)

CdTe(2)

CIS (1)

CIS(2)

[2011]

Ecology Index 2012 - High-quality Recycling (Rec.opt)


2,2 2

,3

2,0 2

,22,2 2

,4

-3

-2

-1

0

1

2

3

4

5

Silicon

mono

Silicon

multi

CdTe(1)

CdTe(2)

CIS (1)

C

IS(2)

[2011]

2,42,5

2,02,52,73,0


NET


inverter

wiring

attachment

frame

module/laminate

semiconductor

glass

Figure 20: Sensitivity consideration for the specific annual yield. Contributions of the sections to the ecologyindex with 1,050 kWh/kWp yield for monocrystalline and multicrystalline silicon, CdTe (1) andCdTe (2) as well as CIS (1) and (2) (application case: residential rooftop); above: Recycling option:Status quo; optimal recycling option below




43/76


8 Conclusions

8.1 Analysis methodology

Life cycle analysis of electricity generation by photovoltaic systems based on seven environmental im-pact categories, including consideration of avoiding use of fossil fuels and metals, enables a compre-hensive and balanced assessment. An environmental impact assessment of photovoltaic systems, whichwould only refer to the climate protection often discussed in isolation from other matters, would notobtain the targeted objective. Although global warming is an important factor in the description ofenvironment impact, a comparison that solely focuses on climate change would result in faulty inter-pretations, e.g. with respect to the environmental impact of the BOS. In order to assess human- and

eco-toxicity, relevant emissions of heavy metals were calculated with the factors of the current USEtoxtoxicity model (UNEP-SETAC). It must be noted that the model is still under development. This alsoapplies to the characterizing factors for metals. Consequently, this examination can only provide anassessment based on the currently available methodology. A sufficiently precise and current depictionof the metallurgic processing in Eco-Invent would have to be assumed as a general prerequisite14.

The contribution to the conservation of resources is also part of the environmental impact analyzed bybifa. The indicator of saved resources, developed by bifa for this study, combines consumption andconservation of fossil fuels and metal resources via an analysis of scarcity, substitution and use ofthese raw materials. The cost analysis is based on the valid incentive pursuant to the German Renewa-ble Energy Act (EEG). As a result, the economic estimate was conducted independent of technology for

the assessed application types.

8.2 New approaches and comparability

The objective of this study was to assess the environmental impact of different photovoltaic technolo-gies. The analysis was not only conducted on the PV module level, but instead covers the complete PVsystem including components of the BOS with the goal of determining the contributions of all PV sys-tem components.

The evaluation of PV systems was not only based on a snapshot analysis, but also includes an assess-ment of the technology improvements that occurred within a year (2011 and 2012) and for the first

time assesses the potential of mid-term improvements on targeted production efficiency increases andthe development of power capacity.

This study undertook an unprecedented assessment of the PV systems using the USEtox methodology.There are a number of PV life cycle assessments which focus on greenhouse gas emissions, stratospher-ic ozone depletion, photochemical ozone creation, acidification and eutrophication. There was previ-ously no assessment of human toxicity or eco-toxicity (cf. [WILD 2012], [MASO 2006]).

14Due to the dominated influence of zinc emissions on the impact categories of human- and eco-toxicity, the process data of

zinc and electric steel production as well as hot-dip galvanizing in Eco-Invent were checked and updated in part. The datarecord for indium was also affected due to co- allocation of expenses during zinc production.


44/76


When comparing the various study results, it is important to correctly classify the system parametersevery time. For example, it is important to consider the following:

Which reference points were selected for the study (module, area, yield)? Which solar irradiation values do the results apply to (standard conditions, southern Europe, Bavar-

ia)?

How large are the assessed systems (small systems, power plants)?

How should the ambient temperature of the modules be assessed (site, ventilated, open air)?

The differences of the possible constellations often make a direct comparison difficult.

8.3 Comparison of conventional electricity generation with PV technologies

A general comparison to the environmental impact of conventional energy generation was made inorder to better categorize the results of the environmental impact of PV systems. This helped to putthe differences between the PV technologies into perspective. The overall environmental impact of PVsystems is between 2.0 and 3.0 ecology index points (status quo recycling option) for residential roof-top installations, for example. As a result, the environmental impact associated with electricity genera-tion from photovoltaics compared to the average European electricity grid mix15is 10 to 20 times lessthan with fossil

Documents

Bifa PV Eco-EfficBifa PV Eco-efficiency 2013