Life cycle analysis within pharmaceutical process ... · development. Nevertheless, it is highly promising to rethink, redesign, and optimize process strategies as early as possible

Life cycle analysis within pharmaceutical process optimizationand intensification: Case study of an API productionCitation for published version (APA):Ott-Reinhardt, D., Kralisch, D., Dencic, I., Hessel, V., Laribi, Y., Perrichon, P., Berguerand, C., Kiwi-Minsker, L.,& Loeb, P. (2014). Life cycle analysis within pharmaceutical process optimization and intensification: Case studyof an API production. ChemSusChem, 7(12), 3521-3533. https://doi.org/10.1002/cssc.201402313

DOI:10.1002/cssc.201402313

Document status and date:Published: 01/01/2014

Document Version:Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can beimportant differences between the submitted version and the official published version of record. Peopleinterested in the research are advised to contact the author for the final version of the publication, or visit theDOI to the publisher's website.• The final author version and the galley proof are versions of the publication after peer review.• The final published version features the final layout of the paper including the volume, issue and pagenumbers.Link to publication

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, pleasefollow below link for the End User Agreement:www.tue.nl/taverne

Take down policyIf you believe that this document breaches copyright please contact us at:[email protected] details and we will investigate your claim.

Download date: 08. Apr. 2021

https://doi.org/10.1002/cssc.201402313https://doi.org/10.1002/cssc.201402313https://research.tue.nl/en/publications/life-cycle-analysis-within-pharmaceutical-process-optimization-and-intensification-case-study-of-an-api-production(20c0defd-a62e-4297-89fb-a4ad22838531).html

DOI: 10.1002/cssc.201402313

Life Cycle Analysis within Pharmaceutical ProcessOptimization and Intensification: Case Study of ActivePharmaceutical Ingredient ProductionDenise Ott,*[a] Dana Kralisch,[a] Ivana Denčić,[b] Volker Hessel,[b] Yosra Laribi,[c]

Philippe D. Perrichon,[c] Charline Berguerand,[d] Lioubov Kiwi-Minsker,[d] and Patrick Loeb[e]

Introduction

To quote Lynn R. Goldman, a well-known American publichealth physician, epidemiologist, and pediatrician, as well asformer Assistant Administrator for toxic substances at the U.S.Environmental Protection Agency: “Green Chemistry is preven-tative medicine for the environment.”[1] But, how green is me-dicinal chemistry?

Although pharmaceuticals are usually produced batchwise,yielding only a few tons annually, environmental impacts areindeed significant. If comparing the estimated total waste re-lated to annual production, pharmaceutical processes aremuch less environmentally benign. As estimated by the Ameri-

can Chemical Society Green Chemistry Institute PharmaceuticalRoundtable (ACS GCIPR), the E factor[2, 3] is comparable to thefine chemical sector, but many times higher than that for oilrefining and bulk chemical industry.[4] This is due to the long-chain production pathways and high efforts for active pharma-ceutical ingredient (API) isolation and purification.

The ACS GCIPR was launched by the ACS Green ChemistryInstitute and global pharmaceutical corporations in 2005, withthe aim of encouraging innovation by simultaneously imple-menting green chemistry[5] and green engineering[6] ideals inthe pharmaceutical industry.[7, 8] Thereby, process intensification(PI), coupled with microreaction technology and flow process-ing, is described as an important “revolutionary” tool in thepharmaceutical and fine chemical sectors, offering, for exam-ple, shorter and more efficient process development or scaleupactivities, safer and more efficient processes, as well as lowertotal costs.[9–11] In the meantime, striking improvements, for ex-ample, of the syntheses of ibuprofen, sertraline, pregabalin,and others, have been published;[12–18] this highlights the valueof process optimization and redesign (“rethinking”) of pharma-ceutical processes with respect to environmental impacts.

If the implementation of process optimization and intensifi-cation concepts is accompanied by environmental and eco-nomic assessment tools from the start, the possibility of devel-oping more environmentally benign and economically compet-itive processes can be enhanced significantly.

This approach was followed within the project POLYCAT.POLYCAT is a large-scale integrating collaborative FP7-EU proj-ect; its acronym stands for “modern polymer-based catalystsand microflow conditions as key elements of innovations infine chemical synthesis”.[19] By combining novel heterogeneouscatalytic systems and improved process conditions, as well as

As the demand for new drugs is rising, the pharmaceutical in-dustry faces the quest of shortening development time, andthus, reducing the time to market. Environmental aspects typi-cally still play a minor role within the early phase of processdevelopment. Nevertheless, it is highly promising to rethink,redesign, and optimize process strategies as early as possiblein active pharmaceutical ingredient (API) process development,rather than later at the stage of already established processes.The study presented herein deals with a holistic life-cycle-

based process optimization and intensification of a pharma-ceutical production process targeting a low-volume, high-valueAPI. Striving for process intensification by transfer from batchto continuous processing, as well as an alternative catalyticsystem, different process options are evaluated with regard totheir environmental impact to identify bottlenecks and im-provement potentials for further process development activi-ties.

[a] Dr. D. Ott, Dr. D. KralischDepartment of Pharmaceutical TechnologyFriedrich Schiller University JenaOtto-Schott-Strasse 41, 07745 Jena (Germany)E-mail : [email protected]

[b] I. Denčić, Prof. Dr. V. HesselLaboratory of Chemical Reactor EngineeringMicro Flow Chemistry and Process TechnologyDepartment of Chemical Engineering and ChemistryEindhoven University of TechnologyP.O. Box 513, 5600 MB Eindhoven (The Netherlands)

[c] Y. Laribi, P. D. PerrichonSanofi Recherche and DeveloppementChemistry and Biotechnology DevelopmentChemistry Vitry, Centre de Recherche de Vitry-sur-Seine13, quai Jules Guesde, BP 14, 94403 Vitry-sur-Seine (France)

[d] C. Berguerand, Prof. Dr. L. Kiwi-MinskerGroup of Catalytic Reaction EngineeringEcole Polytechnique F�d�rale de Lausanne (GGRC-ISIC-EPFL)1015 Lausanne (Switzerland)

[e] Dr. P. LoebFraunhofer ICT-IMM, Carl-Zeiss-Strasse 18–2055129 Mainz (Germany)

� 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemSusChem 2014, 7, 3521 – 3533 3521

CHEMSUSCHEMFULL PAPERS

utilizing enabling processing techniques based on microreac-tion technology, more sustainable production pathways wereaimed for, which will be finally implemented in the Evonik Evo-trainer.[20] This modular, compact container plant is seen as oneof enabling concepts for the “factory of the future”.

To provide decision support during these technical develop-ments targeted at process optimization and intensification,and to ensure the competitiveness, environmental compatibili-ty, and thus, sustainability of POLYCAT-based processes, lifecycle assessment (LCA) and cost analyses were performed.[19]

Herein, the results of process optimization obtained are pre-sented by taking an established production of an anticancerdrug at the Sanofi plant site in Vitry-sur-Seine, France, as start-ing and reference points. The study was performed in parallelto experimental investigations to find out the most relevantweak points of the current process and to show promisingpathways to improve its overall environmental balance fordirect implementation in development. The eco-efficiency ofthe process alternatives were taken into account as well toensure the competiveness of the newly developed, environ-mentally benign alternative.

Integration of evaluation approaches within pharmaceuticalprocess design, optimization, and intensification

Moving toward the application of green chemistry principlesand following sustainable practices in pharmaceutical chemis-try has been intensively pursued over the past decade, see, forexample, refs. [21–26] , and references therein.

To implement life cycle thinking into practical decision guid-ance, the methodology of LCA can be applied.[27, 28] This holisticapproach includes the whole life cycle of a product or process,starting from extraction of resources, production, and use upto product recycling, reuse, or disposal. Although LCA method-ologies are described thoroughly in literature, the applicationis still not practiced widespread within pharmaceutical manu-facturing.[26, 29] Life cycle invento-ry (LCI) data are often rarelyavailable for pharmaceutical pro-cesses, mainly due to the use offine chemicals with complex mo-lecular structures stemming fromlong production chains. There-fore, the ACS GCI PharmaceuticalRoundtable has defined processmass intensity (PMI; defined asthe total mass of materials perunit mass of product) as the keymass-based green metric forbenchmarking the greenness ofpharmaceutical processes.[26, 30]

Unfortunately, PMI does not pro-vide a holistic LCA view, but theuse of such a mass-based metriccan function as a supportingtool within initial process designdecision making. However, in

recent years, the intention to apply more sophisticated and ho-listic LCA methodology has increased significantly in manyareas of research, development, and manufacturing, see, forexample, refs. [29, 31–34].

There has also been rapidly growing interest at Sanofi in in-vestigating more productive, competitive, and simultaneouslyenvironmentally benign process routes by applying greenchemistry, green engineering, and PI principles. In 2011, Sanofijoined the ACS GCI Pharmaceutical Roundtable. Within theiractive ingredient manufacturing, Sanofi is committed to im-proving its drug manufacturing processes, while minimizingenvironmental impacts. Therefore, Sanofi also tracks the green-ness of their processes by implementing solvent-selectionguidelines,[35] as well as metrics such as PMI, reaction mass effi-ciency (RME), or solvent consumption.[36] In addition, indicesdescribing waste production, carbon content, or potential CO2emissions related to the targeted product are used for compar-ing processes and quantifying progress during the process de-velopment phase.[24]

However, the study presented herein is the first holistic life-cycle-based process optimization and intensification of an es-tablished pharmaceutical production process at Sanofi.

API production process at Sanofi: Process analysis

The API production process investigated herein is based onstereo- and chemoselective reduction of aromatic nitro groupsin the presence of alkene functionalities to the respective Z-and E-amino compounds. The Z-isomeric compound is theeconomically valuable one. The production scale is 100 kg peryear, which indicates a low-volume, high-value API. For confi-dentiality reasons, no details about the chemical compoundscan be given. The final API is targeted for the treatment of ad-vanced-stage tumors. A generic process is shown in Figure 1.Conventionally, a powdery Pd@C catalyst is used for this heter-ogeneous hydrogenation reaction. First, charging and mixing

Figure 1. Simplified flow sheet of the API batch manufacturing process investigated in this study; Z-isomeric com-pound as economically valuable key output material (KOM).


CHEMSUSCHEMFULL PAPERS www.chemsuschem.org

www.chemsuschem.org

of the components take place followed by a batch reaction.Thereby, the functional unit of E/Z-isomeric starting material,which is referred to in the following as the key input material(KIM), amounts to 8.1 kg KIM per batch.

Hydrogenation is achieved by using ammonium formate asan in situ hydrogen source. After the hydrogenation reaction,the catalyst is filtered and washed. To separate the targetedKOM, that is, the Z-amino isomeric compound, the E-isomericcompound (which is the main byproduct) has to be isolatedby 1) solvent switching; 2) a second reaction (in the followingreferred to as reaction II), in which the E-amino isomer crystalli-zes; 3) E-amino isomer filtration and washing; and 4) secondsolvent switching, in which the Z-amino isomer crystallizes andis dried in vacuo. The overall yield of targeted KOM amountsto 47 %.

The main bottlenecks of the process are 1) low reaction effi-ciency, 2) high solvent consumption, 3) fresh catalyst is re-quired for each batch, 4) high losses of KOM along the processchain (mainly due to purification and crystallization steps), and,5) as a result, a high waste production rate.

The overall PMI of the investigated process amounts to115 kg of raw material (RM) input per kg of API in the currentstate. Solvents give the largest contribution to the mass inten-sity (about 96 %), which emphasizes the need for reducingand/or replacing hazardous solvents (see Figure 2, top).

Because PMI does not provide a holistic LCA view on the en-vironmental burdens of the pharmaceutical process, previousinvestigations were extended by using the cumulative energydemand (CED).[37, 38] CED represents the energy demand duringthe entire life cycle of a product or process, and is accepted asa suitable screening indicator predicting environmental bur-dens of production[39] and reflecting many energy-related life

cycle impact categories of an LCA, for example, global warm-ing potential (GWP) or abiotic resource depletion poten-tial.[40–43] Primarily, the CED for the KIM supply was predictedby using the Finechem software tool developed by Wernetet al.[44, 45] (and later on substituted by values resulting from LCImodelling). As a result, parallels between CED and PMI resultscould be identified, especially with regard to solvent consump-tion. In contrast to PMI, the CED values for the catalyst supply,mainly due to palladium, as well as for KIM supply, indicatehigh environmental burdens as well, although only in smallamounts compared with the overall mass streams (seeFigure 2, bottom).

Resulting from initial process analysis, different strategies ofoptimization became clear immediately: 1) development ofstrategies to reduce solvents and/or replace hazardous sol-vents; 2) fixation of the heterogeneous catalyst in the reactorinstead of using a catalyst suspension (thus avoiding additionalsolvent for catalyst separation); 3) optimizing the crystallizationstep for KOM isolation (lower solvent demand, higher crystalli-zation yield) ; 4) improvement of overall purification yields; and5) recovery of KOM losses.

To consider the latest PI issues, the following items wereconsidered to be conceivable for implementation as well :1) development and utilization of efficient novel nanoparticu-late catalysts based on a structured support ; and 2) continuousoperation mode by using intensified micro-/millireactors, inwhich the catalyst is implemented through deposition on wallsor filling of reactor channels (fixed bed), respectively.

Both ideas were followed within POLYCAT. With regard tofurther optimization potentials gained from full continuousprocessing, continuous distillation of solvents as well as micro-wave drying, to enhance the efficiency of solvent recycling andthe drying procedure, were considered as theoretical PI op-tions as well. In case of a transfer to continuous processing,molecular hydrogen instead of ammonium formate will beused at Sanofi. The catalytic three-phase hydrogenation of aro-matic nitro compounds has great importance in the manufac-turing of aromatic amines.[46–48]

Cost analysis studies recently done on selected process op-tions have shown that an intensified millireactor-based tech-nology for synthesis and downstream processing results inhigh economic benefits.[49] This is valid even if lower initial con-centrations of KIM are applied to avoid crystallization in pe-ripheral elements (i.e. , 30 g L�1).

In the next step, by means of a holistic, detailed LCA analy-sis, the following questions were addressed: Which processsteps and materials are the most influencing factors from a ho-listic environmental point of view? How can the current APIproduction process be improved? Which process parameterchanges will result in significant ecological improvements?Does the implementation of continuous operation mode im-prove overall ecological sustainability, although the concentra-tion of KIM in the solvent needs to be reduced to avoid clog-ging phenomena? Will other process optimization and intensi-fication options within novel process design outweigh nega-tive impacts of increased solvent consumption? What initialconcentrations of KIM should be targeted to outweigh the cur-

Figure 2. Shares of components in metrics PMI (top) and CED (bottom) forcurrent API production process.



www.chemsuschem.org

rent Sanofi batch process? How does the preparationof catalysts affect the overall environmental balance?What lifetimes and stability of heterogeneous cata-lysts utilized are required for flow conditions to beecologically competitive to the current slurry-batchAPI manufacturing process?

By performing detailed LCA analyses, bottlenecksor process improvement potentials can be investigat-ed and recommendations for further process devel-opment can be given from a holistic point of view,by taking into account a variety of partly contradicto-ry environmental effects.

Methodological approach

To support process redesign and optimization efforts,a comparative environmental assessment and holisticdecision-making procedure was followed accordingto previous studies, see, for example, refs. [50–52].Thus, focus was given on life cycle approaches thatcovers all life cycle stages of a process or productsystem, and which were combined with an iterativescreening and optimization procedure. The LCAmethodology was chosen to quantify and assess the potentialenvironmental impacts associated with the processes under in-vestigation. During the evaluation steps, the information wassummarized by using several criteria for process improvementto assess whether the requirements specified during the objec-tive definition step were met. Finally, the screening step endedwith a ranking of alternatives according to their overall level ofattractiveness by applying multicriteria decision-making tools.The key criteria for green process improvement determinedwere then taken into account within the next iterative step ofprocess design. In the end, the environmental impacts of theredesigned process alternatives were quantified and comparedwith a reference or benchmark process.

To evaluate the entire 10 step synthesis holistically, LCI dataare necessary. In Figure 3, the system boundary and overall ret-rosynthetic steps for KIM supply down to very basic com-pounds with existing LCI data (green points, Figure 3) areshown schematically. In some synthetic steps, valuable byprod-ucts (orange points, Figure 3) are formed to which environ-mental burdens can be allocated. Data for the KIM productionstep were from Sanofi in-house processes. If there were nodirect production data from manufacturers available for LCIupstream chain modelling, information platforms such as Sci-Finder,[53] Web of Science,[54] or Espacenet[55] were used. Fur-thermore, information within textbooks, for example, Ullmann’sEncyclopedia of Industrial Chemistry,[56] Handbook of Petro-chemicals and Processes,[57] and the references therein, wereused. These data sources allowed a mass balance of all mass-based input and output flows for each synthetic step to begenerated. In the case of low data availability, the LCI analysisapproach by Hischier et al.[58] was followed; this allowed forthe integration of average values with regard to transport,[59]

energy and water consumption,[60] and plant infrastructure. Forthe investigated fine chemical upstream processes, an average

organic solvent recovery rate of 71 % was assumed.[61] For KIMpreparation, which is the last process step before the API pro-duction process under study, no solvent recovery is actuallyimplemented at Sanofi ; this was also considered within theLCA investigations. In a later section, the impact of solvent re-covery within Sanofi processes (KIM and API manufacturing) isalso considered. Due to a lack of information and to ensureconsistency, LCI data for hydrogenation catalyst preparationwere also modeled according to Ref. [58] . RM input for catalystpreparation was modeled directly based on extracted metal(instead of, e.g. , inorganic metal salt precursors).

In general, process waste (liquid, solid) was considered to bedisposed of by hazardous waste incineration. Distilled solventwaste was considered to be disposed of by conventionalwaste incineration. All processes were modeled by consideringinfrastructure (Ecoinvent v.2.2[62] module “infrastructure, chemi-cal plant, organics”).

The differences in the impact of the chosen reaction equip-ment (stirred tank reactor compared with tubular reactor (e.g. ,H-Cube[63] system, ThalesNano, Hungary) etc.) was not evaluat-ed separately due to the estimated negligibly low life cycleimpact of the overall plant infrastructure and transport (in totalless than 1 % of the environmental impact in all impact catego-ries considered) compared with the production performedwithin the reactor (see Figures 4 and 7, below). This is in linewith earlier results of a comprehensive LCA study, in which thesupply chain of the alternative reactors (batch and continuous)were considered.[64]

LCI analysis and impact assessment were supported by thesoftware tools Umberto 5.6 and the LCI database Ecoinventv.2.2, respectively.[62, 65] Life cycle impact assessment (LCIA) wasconducted by applying the LCIA methodology ReCiPe 2008 de-veloped by Goedkoop et al.[66] Thus, 10 ReCiPe impact poten-tials, namely, GWP (in kg of CO2 equivalents per functional unit

Figure 3. Upstream chain analysis of KIM for LCI data collection.



www.chemsuschem.org

(FU)), ozone depletion potential (ODP; in kg of chlorofluorocar-bon-11 (CFC-11) per FU), terrestrial acidification potential (TAP;in kg of SO2 equivalents per FU), freshwater eutrophication po-tential (FEP; in kg of P equivalents per FU), human toxicity po-tential (HTP; in kg of 1,4-dichlorobenzene (1,4-DCB) per FU),photochemical oxidant formation potential (POFP; in kg ofnon-methane volatile organic compounds (NMVOC) per FU),terrestrial ecotoxicity potential (TETP; in kg of 1,4-DCB per FU),natural land transformation potential (NLTP; in m2 per FU),metal depletion potential (MDP; in kg of Fe equivalents perFU), and fossil fuel depletion potential (FDP; in kg of oil equiv-alents per FU), were addressed at the midpoint level and hier-archist perspective. This perspective is in accordance with themost common policy principles with regard to time frame andother issues.[66] Most frequently, the 100 year timeframe is usedtherein. All LCIA potentials were weighted equally during thedecision making process. In some cases, LCIA and interpreta-tion were simplified by applying only selected impact poten-tials, namely, GWP, HTP, MDP, and TAP. For example, globalwarming due to anthropogenic emissions is one of the mosturgent environmental problems and is directly related to cost-intensive energy consumptions; thus particularly awakeningstakeholder’s interests. In addition, HTP addresses toxicologicalissues related to potential emissions that occur within the con-sidered system boundary. Ingeneral, environmental impactswere estimated in a cradle-to-gate boundary for European pro-duction. The LCIA results pre-sented in the following arebased on FU = 1 kg API.

Results and Discussion

Screening for optimization andintensification: Process alterna-tives in focus

After establishing LCI data forupstream chains, the holisticevaluation of the current pro-duction at Sanofi, as well as pos-sible process alternatives, was

performed. Thereby, different process alternatives were consid-ered to investigate the environmental effects of single andcoupled process parameter variations, regarding 1) changingthe reduction agent, 2) changing the crystallization approach(direct crystallization of KOM in acetonitrile (ACN)/benzyl alco-hol), 3) isolating the Z isomer prior to the hydrogenation reac-tion (due to melting point differences of E/Z isomers) and re-ducing KOM losses (i.e. , higher purification/crystallizationyield), 4) changing the drying procedure (microwave dryingversus vacuum drying), 5) changing the operation mode (frombatch to continuous reaction and processing), accompanied by6) changing the catalyst. Tables 1 and 2 provide a summary ofthe batch and continuous process alternatives consideredherein. For simplification, the reference process at Sanofi iscalled the actual process (AP), whereas the rest are batch (BP)and continuous process alternatives (CP).

Inventory analysis was based on process flow sheet informa-tion from Sanofi (related to existing batch procedures and firsttrial processes with a continuous H-Cube[63] mini system, Tha-lesNano, Hungary, as well as Aspen process simulation re-sults).[49] In case of microwave and vacuum drying steps, theenergy demand was calculated thermodynamically.

According to Kemp,[67] to achieve drying, at least the latentheat of evaporation must be supplied to turn moisture into

Table 1. Summary of the main process characteristics of considered batch scenarios.[a]

Process OperatingT [8C]

Solvent Initial KIMc [g L�1]

Reducingagent

Catalyst Crystallizationmethod

Z-separation +reduction of KOM losses

KOM drying Overallyield [%]

AP 30 ACN 87 NH4HCO2 Pd@C powder, 5 wt % Pd indirect – vacuum 47BP 1 30 ACN 87 NH4HCO2 Pd@C powder, 5 wt % Pd direct – vacuum 50BP 2 50 MeOH/H2O 71 Na2S2O4 – indirect – vacuum 42BP 3 50 MeOH/H2O 71 Na2S2O4 – direct – vacuum 44BP 4[b] 30 ACN 87 NH4HCO2 Pd@C powder, 5 wt % Pd indirect X vacuum 54BP 5[b] 30 ACN 87 NH4HCO2 Pd@C powder, 5 wt % Pd indirect – microwave 47BP 6[b] 30 ACN 30 NH4HCO2 Pd@C powder, 5 wt % Pd indirect – vacuum 47

[a] Common process conditions: p = 1 bar; reaction time�24 h; reactor volume, V = 90 L. [b] Hypothetical scenarios for demonstration of effects of param-eter variations, based on Aspen simulation and/or estimations.

Table 2. Summary of main process characteristics of continuous process alternatives.[a]

Process Operationmode

Catalyst Crystallizationmethod

Z-separation +reduction of KOM losses

Overall yield[%]

CP 1 continuous reaction Pt/V@C indirect – 52CP 2 continuous reaction Pt/V@C direct – 55CP 3[d] continuous reaction Pt@ZnO indirect – 52CP 4[e] continuous reaction Pt/V@C indirect X 60CP 5[e] continuous reaction Pt/V@C direct X 64CP 6[f] continuous process Pt/V@C indirect – 52

[a] Common process conditions: T = 55 8C; p = 35 bar; liquid flow rate, Fl = 41 mL min�1; residence time 9 s; esti-

mated reactor volume, V = 10 mL[b] for assumed bed porosity of 60 % (Pt/V@C);[c] reaction solvent: ACN; KIM in-itial concentration c0 = 30 g L

�1; reducing agent: H2 ; catalyst: fixed bed; Pt/V@C: 1 wt % Pt/2 wt % V @C,Pt@ZnO: 2 wt % Pt @ZnO; KOM drying by means of a vacuum dryer. Liquid flow rate, Fl, calculated accordingto P = c0Fl ; P : targeted productivity [kg h

�1] , c0 : KIM initial concentration. [b] For H-Cube or similar. [c] Bed po-rosity of Pt@ZnO still to be determined. [d] In accordance to CP 1, hypothetical scenario considering newly de-veloped hydrogenation catalyst. [e] Hypothetical scenarios, forecasted to demonstrate the effect of optimizedcontinuous processing. [f] Hypothetical scenario, RM demand adapted to known process output, based onAspen simulation.



www.chemsuschem.org

vapor. Thus, the absolute minimum amount of energy fora drying process was determined by the solvent’s mass, heatcapacity, and its enthalpy of vaporization. An overall drying ef-ficiency of 40 % was assumed.[68] Regarding microwave drying,power consumption was calculated according to previouswork,[49] considering an overall efficiency of 25 %.[69, 70] The re-maining solvent amount to be dried was assumed to be2.5 kg kg�1 KOM in each case.

Life cycle impact assessment

Ecological profile of reference API process

First, in Figure 4, an ecological fingerprint of the current APIproduction process AP at Sanofi is shown. The main result isthat the ecological impact is mostly affected by the supply ofKIM, the supply of ACN as a reaction and purification solvent,and the supply of the catalyst. Despite the fact that thedemand of the Pd@C catalyst amounts to only 0.1 kg kg�1 KIM,the environmental effect is indeed significant, not least be-cause the catalyst is disposed of after each batch. The high en-vironmental burden of catalyst supply is mainly due to thesupply of palladium (>99 %).

Thus, from a life cycle point of view, the main key drivers forenhancing sustainability of the existing AP production processare the improvement in reaction efficiency, for example,through the increase in reaction, purification, and/or crystalliza-tion yield, as well as reduction of solvent demand and/or theintegration of efficient solvent recycling strategies. Moreover,lower catalyst consumption, efficient catalyst recycling/reacti-vation strategies, or the use of “greener” catalysts can signifi-cantly contribute to more environmentally benign processing.

In the following section, theeffects of selected process pa-rameter variations on ecologicalimpacts are discussed in detail.

Process parameter variations forbatch optimization

First, focus was given to batchprocess optimization strategies.Figure 5 provides a summary ofthe results for scenarios AP andBP 1–BP 6 (see Table 1) withregard to selected LCIA poten-tials. For demonstration purpos-es and a clear arrangement, allscenarios are standardized to thereference case AP (blue deca-gon, Figure 5). The larger thedecagon areas in Figure 5, thehigher the environmental bur-dens associated with the respec-tive process alternative.

Thus, it can be seen thatchanging the reduction agent

from NH4HCO2 to Na2S2O4 (BP 2, BP 3) does not result in overallecological benefits ; this is mainly due to lower yields. In addi-tion to higher RM consumption, especially the use of dichloro-methane, which is required in these scenarios for post-treat-ment after hydrogenation, causes a significant peak in ODP.However, simultaneous reduction within HTP, POFP, MDP, TAP,and FEP also reveals environmental benefits of a catalyst-freepreparation. In BP 2 and BP 3 alternatives, no metal-based hy-drogenation catalyst was used; this is the key reason for fewerenvironmental burdens within these categories (cf. Figure 4).Instead, the application of process alternatives BP 1 and BP 4results in fewer environmental burdens within all impact cate-gories. Within these scenarios, the KOM yield is improved asa result of more effective downstream processing. Primarily,this is realized by a direct crystallization approach that enhan-

Figure 4. Ecological profile of conventional batch API manufacturing (pro-cess AP) at Sanofi. Scaled effects.

Figure 5. LCIA: a comparison of conventional API batch processes versus batch process alternatives. Bottom right:complete graphical display of ODP. Scaled effects.



www.chemsuschem.org

ces the crystallization yield (BP 1). The direct crystallizationroute further requires much less solvent than the referenceroute.

Second, this result can be gained by strengthening effortsfor KOM recovery from waste streams (BP 4). As shown inFigure 4, the share of KIM supply in noncatalyst dominatedLCIA items is >40 %. Thus, yield enhancement is another keydriver, favoring scenario BP 4.

Because energetic aspects do not seem to play a crucial rolewithin this production process, the switch to microwavedrying has a negligibly low impact on the overall environmen-tal effects (see Figure 5, BP 5, which has nearly the same envi-ronmental impact as AP).

In conclusion, the following strategies are to be focused onin further process development: 1) overall yield enhancement;2) reduced solvent demand; and 3) changing the catalyst, pref-erably switching to wall-coated or fixed-bed catalysis.

In accordance with the PI approach followed withinPOLYCAT, the next step evaluated was the transition frombatch to continuous processing by using hydrogen and a fixedheterogeneous, more effective catalyst. Thus, the solventdemand for catalyst separation and product stream purificationis reduced, and possibly improves the environmental balanceas well. For safety precautions, the use of hydrogen as a directreducing agent was not applicable in the case of batch proc-essing. In continuous processing, changing the reducingagent, that is, by using hydrogen instead of ammonium for-mate or sodium dithionite, will result in an environmentallymore benign procedure that features less emissions and down-stream efforts.[2, 3] Figure 6 presents the resulting environmentalimpacts for the supply of those alternative reducing agents uti-lized in, for example, AP, BP 2, and a continuous process alter-native with hydrogen, as discussed in the following (CP 1, seeTable 2).

However, the transfer from batch to continuous processingmay also result in a worse environmental balance. If assuminga KIM concentration in ACN of 30 g L�1, which is the currentstatus within continuous process development, the environ-mental impact would clearly be worsened if combined withbatch processing (see BP 6, Figure 5). Thus, the critical questionis, whether the environmental benefits gained by continuous,heterogeneously catalyzed processing can counterbalance oreven outweigh current constraints in process engineering.

Comparison of crude batch and continuous operation: Ecolog-ical profile

In Figure 7, the LCIA of both batch (AP, see Table 1) and cur-rent continuous operation processing with a benchmark cata-lyst (CP 1, see Table 2) is compared. The current continuousprocessing (CP 1) will not bring immediate environmental ben-efits with regard to GWP, TETP, NLTP, and FDP. However, thisholistic life cycle analysis reveals significant benefits withinother LCIA categories, such as HTP, POFP, MDP, TAP, and FEP.Thus, despite featuring higher solvent and energy consump-tion in continuous reaction, changing the catalyst and a yieldenhancement outweigh the negative environmental burdens

already in the current state of development. Beyond this, sev-eral additional actions can strengthen the greenness of the re-designed continuous API production process, as discussed inthe following.

Process parameter variations for continuous optimization:Prognosis

Figure 8 reflects a prognosis for GWP and HTP with regard todifferent parameter variations that can be realized in combina-tion with continuous processing. Starting from current batchprocess AP, alternative continuous mode processes (seeTable 2) were evaluated comparatively. As discussed above,transferring from batch to continuous reaction without any fur-ther process optimization results in 1 and 41 % reduction onGWP and HTP, respectively (compare AP and CP 1 in Figures 7and 8).

According to Sanofi, a process including continuous reactionfollowed by a downstream processing option (“direct crystalli-zation”) would be easy to implement (scenario CP 2, seeTable 2). Thus, an additional reduction in GWP and HTP by 12and 4 %, respectively, can be gained (compared to CP 1).Recent catalyst developments within POLYCAT also led to an-other promising and stable hydrogenation catalyst, namely,Pt@ZnO,[71, 72] which was favored, for example, over Pt/V@C, forthe following reasons: 1) The Pt/V@C catalyst indeed demon-strates high selectivity (up to 99 %) towards the desired prod-uct, but only with system modifiers. However, the presence ofadditives, even in small concentrations, could be a problem forproduct purity, see, for example, Ref. [73]. Usually, such modifi-ers are sulfur-containing organic compounds and are verytoxic. 2) Furthermore, the recycling of Pt/V@C is difficult, andthe presence of toxic vanadium is a big issue, especially in thecase of drug production. The developed Pt@ZnO catalyst does

Figure 6. Comparison of reducing agent supply (Na2S2O4, NH4HCO2, H2) perkg of KOM (according to scenarios AP, BP 2, and CP 1) from a life cycle per-spective. Scaled effects, logarithmic scale.



www.chemsuschem.org

not require any additives to attain close to 100 % selectivity atup to 99 % conversion, and ensures the purity of the finalproduct without special cleaning steps. 3) Moreover, the cata-lyst was confirmed to be stable in batch and continuous oper-ation. Consequently, confidence in the choice of the Pt@ZnOcatalyst for further process design was given.

To date, screening in the H-Cube system resulted in at leastcomparably high performances to that of CP 1. By evaluatingthe performance of this hydrogenation catalyst, the resultingdifference to CP 1 can be estimated to be negligibly low (seeCP 3, Table 2) (Due to a lack of data, catalyst consumptioncomparable to that of CP 1 was assumed, which is 0.23 g of

Pt@ZnO per kg of KIM. Thisvalue was confirmed by POLY-CAT members. Furthermore,a lower catalyst amount (1/3 ofthe above assumed value) wasindicated to be feasible as wellas within experimental condi-tions. However, this needs to bechecked in further experiments.)In batch alternatives the catalystshare in GWP and HTP is about10 and 42 %, whereas in continu-ous alternatives it is

ther improvement by 23 (GWP) and 17 % (HTP) could be ach-ieved (see CP 5, Figure 8). Performing the process in a full con-tinuous operation mode (reaction step and downstream proc-essing) will result in benefits of 6 (GWP) and 43 % (HTP) withregard to the current industrial process AP (see CP 6, Figure 8).Within future developments, this environmental profit couldbe extended by multiparameter changes (e.g. , continuous, het-erogeneously catalyzed processing + direct crystallization + im-proved downstream processing), as demonstrated for CP 5.

Another challenging task, but highly profitable from an envi-ronmental point of view—and also foreseen to be implement-ed at Sanofi site—would be the recycling and reuse of organicsolvents. Up to now, used solvents will enter the waste stream,that is, they are not fed back. Due to residual contaminations,quality requirements or legal restrictions (e.g. , good manufac-turing practice (GMP) standards), this is often practiced in thepharmaceutical industry.

Thus, in the following, the possibility of process-integratedsolvent recovery was assumed, namely, for the current bestcase described above (CP 5, see Table 2), to forecast theimpact of solvent recycling.

According to Capello et al. ,[74–76] the profit from solvent re-covery depends on the type of solvent. In the case of ACN, dis-tillation will be the environmentally superior treatment optionto incineration, also because of the high environmental bur-dens within virgin solvent production. In the case of methanol,this treatment option will be preferable only in the case ofgood solvent recovery. According to Capello et al. , a “good”solvent recovery rate of 0.84 kg kg�1 waste solvent was used asbasis for this estimation.[61, 76] Because the KIM manufacturingprocess also takes place at Sanofi, the targeted recovery ratewas also considered for solvents within the KIM synthesis step.

The resulting hypothetical scenario is named CP 5a (84 % re-covery rate), see Figure 8, and indicates high environmentalbenefits when integrating solvent recovery within further pro-cess design activities. Thus, compared with the benchmarkprocess AP, GWP, and HTP could be reduced by 66 and 64 %,respectively. Regarding other impact potentials, the followingreductions could finally be achieved: TETP 71 %, NLTP 65 %,POFP 91 %, MDP 85 %, ODP 26 %, TAP 99 %, FEP 82 %, and FDP67 %.

Sensitivity analyses

LCI Data of KIM

Figure 8 also provides error indicators for each scenario. Theseare determined by assuming that the environmental burdensfor KIM manufacturing and all associated upstream processesvary by �25 %. The error indicators reveal the high share ofKIM supply in most LCIA categories. Nevertheless, the resultingimprovements are significant and the trends discussed abovewill remain.

LCIA criteria contribution to the global preference score

As indicated above, process option CP 5a is favored withregard to LCIA categories GWP and HTP. To determine the con-tribution of all LCIA criteria to a global preference score, a deci-sion-making approach was followed based on a total ranking(equal weighting). The score shown in Figure 9 served asa basis for the subsequent eco-efficiency ranking. The higherthe global score for environmental efficiency, the higher thepreference for the process alternative.

Figure 9. Criteria contribution to the overall LCA preference ranking scores supported by D-Sight;[77] D-Sight parameters: LCA criteria weights : equal, linearminimization of LCIA criteria.



www.chemsuschem.org

The chart in Figure 9 shows how the different LCIA criteriaand their weights contribute to the global score of each pro-cess alternative to be discerned at a glance. Because in thecase of CP 5a all criteria have the highest contribution to theglobal score, this process option is preferred for every settingof weighting parameters. In other cases, a multicriteria out-ranking (without predefined weighting of environmental im-pacts) results in answers that are more complex. The followingorder of Pareto preferences can be found: CP 5a � CP 5 �CP 2��CP 4 �� BP 1 �� BP 4 � CP 6 � CP 1 �� CP 3 ��BP 3 � AP �� BP 5 � BP 2 � BP 6 (x�y : x is preferred to y,x��y : x is non-comparable with y, x~y : x is indifferent to y).Thus, careful weighting, for example, based on stakeholderopinions, expert knowledge or environmental impact boun-dary values have to be further taken into account to decidebetween non-comparable alternatives.

However, the results in Figure 9 already demonstrate thatthe focus on one or few selected parameters can be mislead-ing. For instance, BP 2 would switch from total ranking posi-tion 9 to 13 in the case of an exclusive focus on the typical keyindicator GWP, although the scenario could lead, in practice, tosignificant environmental im-provements within other impactcategories, for example, MDP orTAP, due to catalyst-free process-ing.

Catalyst deactivation scenarios

Last, but not least, the effect ofcatalyst lifetime (operation time)and catalyst deactivation ratewas investigated based on thecurrent status of continuous pro-cess design, CP 3. Although thelife cycle impact of the catalystsupply is identified to be negligi-ble for all continuous, heteroge-neously catalyzed options, theaim is to theoretically investigatelimiting values that counteractenvironmental benefits. To date,experimental tests on the long-term activity and stability of thePt@ZnO catalyst have not beenfinalized, and the results ob-tained by LCA are intended togive a recommendation ofwhich values need to be target-ed.

For this analysis, focus is givento the mainly catalyst-drivenLCIA category “metal depletion”.Because SO2 emissions are by farthe most important in platinumprimary production,[78] andacidification caused by SO2 rep-

resents the dominating environmental impact during platinummining,[79, 80] “terrestrial acidification” is chosen as the secondLCIA category.

As a starting point, in scenario CP 3a a catalyst amount com-parable to that applied in industrial process AP (0.1 kg kg�1

KIM) as well as a one-time use was assumed. Although theconsumption of KIM and catalyst could be reduced due to thehigher yields obtained, without fixation in the H-Cube minisystem and long-term use, the share of Pt@ZnO supply wouldbe the dominating part within MDP of process alternativeCP 3a. In general, LCI data of a metal consumer mix were ap-plied; this is an assembly of the primary (transported from pro-ducing mines for consumption in Europe) and secondarymetals (recovered and consumed in Europe).[62]

Figure 10 shows the results concerning KOM output andMDP in relation to different ranges of deactivation rate,k(deact) (0, 0.01, 0.05 h�1), and catalyst operation time, tc (0–2500 h); this is valid by assuming a first-order reaction and de-activation. In the case of an infinite turnover number (or con-stant turnover frequency), an operation time of 2400 h roughlyindicates the yearly operation time necessary to produce the

Figure 10. KOM output and resulting MDP in relation to different ranges of deactivation rate, ki (0, 0.01, 0.05 h�1),

and catalyst operation time, tc (0–2500 h), for scenario CP 3a as described in text. Plot (top) in logarithmic scale.Scaled effects of MDP.



www.chemsuschem.org

targeted amount of KOM (100 kg). Because of their majorimpact, analysis focuses on the environmental burdens causedby KIM, ACN, and catalyst supply only.

As expected, the environmental burdens increase with de-creasing catalyst activity and decreasing catalyst lifetime. By as-suming a catalyst deactivation rate of k(deact) = 0 h�1 and anoperation time of tc = 2400 h, the impact on MDP will decreaseby 92 % compared with CP 3a. By assuming a deactivation rateof k = 0.01 h�1/0.05 h�1, the operation time, tc, should be50 kgof KOM production).

To forecast eco-efficiency,[52, 81] which was the key enablingforce for competitive products and processes, LCA activitieswere further coupled with cost analysis. As previously shown,the direction of development was also associated with signifi-cant economic benefits.[49] Figure 12 gives a comprehensiveoverview of ecological and economic results (assessment ofeconomic impact was based on Ref. [49] ; the sale price of

Figure 11. Sensitivity analysis regarding catalyst lifetime and activity. Scaled effects of MDP and TAP; logarithmic scale. ! Effect of transition from batch tocontinuous reaction for equal catalyst consumption; !! effect of increasing catalyst operation time, assuming no catalyst deactivation; ! effect of plati-num supply: switching from platinum consumer mix (primary + secondary platinum) to secondary platinum; and ! effect of the current status of continuousprocess development (improved catalyst consumption).



www.chemsuschem.org

Pt@ZnO was estimated by the research group of Lioubov Kiwi-Minsker in Lausanne) for all scenarios discussed above. Thehigher the value in cost and environmental efficiency, thegreater the preference for the appropriate scenario. The follow-ing Pareto ranking of process design options was found: CP 5a� CP 5 � CP 4 �� CP 6 � CP 2 � CP 1 ~ CP 3 � BP 3 ��BP 4 � BP 1 �� BP 2 � BP 5 � AP � BP 6. Thus, a clear orderof preferences could be concluded for most of the alternativeprocesses considered. For CP 4 and CP 6, BP 3 and BP 4, andBP 1 and BP 2, the order may change based on a weightingpriority towards environmental protection or cost savings.

The outcome of the eco-efficiency ranking revealed the highenvironmental and economic benefits obtained by transferringfrom batch to continuous processing combined with the appli-cation of a heterogeneous catalyst fixed in a tubular reactor.Furthermore, it drew attention to the benefits that would addi-tionally be gained by combining full continuous processing(CP 6) and multiparameter optimization (as in CP 5a).

In conclusion, current saving potentials, for example, 765 kgCO2 equivalents (GWP) and 65 kg Fe equivalents (MDP) per kgof API, could be expected upon transitioning from the conven-tional API manufacturing process AP at Sanofi to the best sce-nario, CP 5a, discussed herein. This is complemented by cost-saving potentials of 33 %. Due to smaller reaction hold up,better heat management, and high reaction rates, better safetyand lower risk to human health and the environment are alsoto be expected. This will be also confirmed by the absence oftoxic vanadium compared with the heterogeneous benchmarkcatalyst.

Although this study already picked up on various process-in-fluencing parameters, the scope of topics addressed in thePOLYCAT project was even wider. A more detailed insight intothe methodological approach followed within the project forthe assessment of pharmaceutical and fine chemical processeswill be published in the near future. Furthermore, for example,different catalyst fixation methods and flow reactor conceptshave been investigated. The latest development activities bythe research group of De Bellefon at the Centre National de laRecherche Scientifique (CNRS; Lyon, France) integrate a catalyt-

ic foam-in-tube reactor[82] that is applicable for the same typeof hydrogenation reaction.

The results gained so far have now been implemented inthe newly developed compact container plant of Evonik.

Acknowledgements

We gratefully acknowledge financial support provided by the Eu-ropean Community’s 7th Framework Programme for Researchand Technological Development under grant agreement no. CP-IP 246095-2 POLYCAT.

Keywords: active pharmaceutical ingredient · process design ·green pharmacy · heterogeneous catalysis · life cycleassessment

[1] Quotation from Lynn R. Goldman, former Assistant Administrator of Pre-vention, Pesticides, and Toxic Substances, Environmental ProtectionAgency, 1996.

[2] R. A. Sheldon, Chem. Ind. 1992, 903 – 906.[3] R. A. Sheldon, Green Chem. 2007, 9, 1273 – 1283.[4] R. K. Henderson, J. Kindervater, J. B. Manley, Lessons Learned through

Measuring Green Chemistry Performance—The Pharmaceutical Experi-ence, , ACS Green Chemistry Institute Pharmaceutical Roundtable, 2007http://chemistry.org/greenchemistryinstitute/pharma_roundtable.html.

[5] P. T. Anastas, J. C. Warner, Green Chemistry: Theory and Practice, OxfordUniversity Press, Oxford, 1998.

[6] P. T. Anastas, J. B. Zimmerman, Environ. Sci. Technol. 2003, 37, 94A –101A.

[7] ACS GCI Pharmaceutical Roundtable, acs.org/content/acs/en/green-chemistry/industry-business/pharmaceutical.html.

[8] C. Jimenez-Gonzalez, P. Dell’orco, Chim. Oggi 2013, 31, 62 – 63.[9] EFCE Working Party on Process Intensification, Report on the European

Roadmap for Process Intensification, Creative Energy – Energy Transitions,2007, http://www.efce.info/EUROPIN.html.

[10] D. M. Roberge, B. Zimmermann, F. Rainone, M. Gottsponer, M. Eyholzer,N. Kockmann, Org. Process Res. Dev. 2008, 12, 905 – 910.

[11] F. Linz, Int. Pharm. Ind. 2012, 4, 88 – 90.[12] A. R. Bogdan, S. L. Poe, D. C. Kubis, S. J. Broadwater, D. T. McQuade,

Angew. Chem. Int. Ed. 2009, 48, 8547 – 8550; Angew. Chem. 2009, 121,8699 – 8702.

[13] P. J. Dunn, A. S. Wells, M. T. Williams, Green Chemistry in the Pharmaceuti-cal Industry, Wiley-VCH, Weinheim, 2010.

[14] G. P. Taber, D. M. Pfisterer, J. C. Colberg, Org. Process Res. Dev. 2004, 8,385 – 388.

[15] B. N. Roy, G. P. Singh, P. S. Lathi, M. K. Agrawal, R. Mitra, V. S. Pise, IndianJ. Chem. B 2012, 51, 1470 – 1488.

[16] P. J. Dunn, K. Hettenbach, P. Kelleher, C. A. Martinez in Green Chemistryin the Pharmaceutical Industry (Eds. : P. J. Dunn, A. S. Wells, M. T. Wil-liams), Wiley-VCH, Weinheim, 2010, pp. 161 – 177.

[17] A. Bhattacharya, R. Bandichhor in Green Chemistry in the PharmaceuticalIndustry (Eds. : P. J. Dunn, A. S. Wells, M. T. Williams), Wiley-VCH, Wein-heim, 2010, pp. 289 – 309.

[18] R. Ciriminna, M. Pagliaro, Org. Process Res. Dev. 2013, 17, 1479 – 1484.[19] FP7-EU-project POLYCAT, polycat-fp7.eu.[20] J. Lang, F. Stenger, H. Richert, Elements – Evonik Industries Quarterly Sci-

ence Newsletter, No. 37, Evonik Industries AG, Essen, Germany 2011,pp. 12 – 17.

[21] M. C. Bryan, B. Dillon, L. G. Hamann, G. J. Hughes, M. E. Kopach, E. A. Pe-terson, M. Pourashraf, I. Raheem, P. Richardson, D. Richter, H. F. Sned-don, J. Med. Chem. 2013, 56, 6007 – 6021.

[22] W. Zhang, B. Cue, Green Techniques for Organic Synthesis and MedicinalChemistry, Wiley, Chichester, 2012.

[23] K. Alfonsi, J. Colberg, P. J. Dunn, T. Fevig, S. Jennings, T. A. Johnson, H. P.Kleine, C. Knight, M. A. Nagy, D. A. Perry, M. Stefaniak, Green Chem.2008, 10, 31 – 36.

Figure 12. Eco-efficiency ranking supported by D-Sight:[73] multicriteria out-ranking of KOM production alternatives, according to their environmentaland cost efficiency. D-Sight parameters: criteria weights: equal, linear mini-mization of LCIA criteria and costs.



http://dx.doi.org/10.1039/b713736mhttp://dx.doi.org/10.1039/b713736mhttp://dx.doi.org/10.1039/b713736mhttp://dx.doi.org/10.1021/es032373ghttp://dx.doi.org/10.1021/es032373ghttp://dx.doi.org/10.1021/es032373ghttp://dx.doi.org/10.1021/op8001273http://dx.doi.org/10.1021/op8001273http://dx.doi.org/10.1021/op8001273http://dx.doi.org/10.1002/anie.200903055http://dx.doi.org/10.1002/anie.200903055http://dx.doi.org/10.1002/anie.200903055http://dx.doi.org/10.1002/ange.200903055http://dx.doi.org/10.1002/ange.200903055http://dx.doi.org/10.1002/ange.200903055http://dx.doi.org/10.1002/ange.200903055http://dx.doi.org/10.1021/op0341465http://dx.doi.org/10.1021/op0341465http://dx.doi.org/10.1021/op0341465http://dx.doi.org/10.1021/op0341465http://dx.doi.org/10.1021/op400258ahttp://dx.doi.org/10.1021/op400258ahttp://dx.doi.org/10.1021/op400258ahttp://dx.doi.org/10.1021/jm400250phttp://dx.doi.org/10.1021/jm400250phttp://dx.doi.org/10.1021/jm400250phttp://dx.doi.org/10.1039/b711717ehttp://dx.doi.org/10.1039/b711717ehttp://dx.doi.org/10.1039/b711717ehttp://dx.doi.org/10.1039/b711717ewww.chemsuschem.org

[24] J. P. Adams, C. M. Alder, I. Andrews, A. M. Bullion, M. Campbell-Crawford,M. G. Darcy, J. D. Hayler, R. K. Henderson, C. A. Oare, I. Pendrak, A. M.Redman, L. E. Shuster, H. F. Sneddon, M. D. Walker, Green Chem. 2013,15, 1542 – 1549.

[25] R. K. Henderson, C. Jimenez-Gonzalez, D. J. C. Constable, S. R. Alston,G. G. A. Inglis, G. Fisher, J. Sherwood, S. P. Binks, A. D. Curzons, GreenChem. 2011, 13, 854 – 862.

[26] C. Jim�nez-Gonz�lez, P. Poechlauer, Q. B. Broxterman, B.-S. Yang, D.am Ende, J. Baird, C. Bertsch, R. E. Hannah, P. Dell’Orco, H. Noorman, S.Yee, R. Reintjens, A. Wells, V. Massonneau, J. Manley, Org. Process Res.Dev. 2011, 15, 900 – 911.

[27] DIN Deutsches Institut f�r Normung e.V. , DIN EN ISO 14040: Environ-mental Management, Life Cycle Assessment, Principles and Framework,Beuth, Berlin, 2006.

[28] DIN Deutsches Institut f�r Normung e.V. , DIN EN ISO 14044: Environ-mental Management, Life Cycle Assessment, Requirements and Guidelines,Beuth, Berlin, 2006.

[29] G. Wernet, S. Conradt, H. Isenring, C. Jim�nez-Gonz�lez, K. Hungerbueh-ler, Int. J. Life Cycle Assess. 2010, 15, 294 – 303.

[30] C. Jimenez-Gonzalez, C. S. Ponder, Q. B. Broxterman, J. B. Manley, Org.Process Res. Dev. 2011, 15, 912 – 917.

[31] A. Banimostafa, T. T. H. Nguyen, Y. Kikuchi, S. Papadokonstantakis, H. Su-giyama, M. Hirao, K. Hungerbuehler, Green Processing and Synthesis2012, 1, 449 – 461.

[32] T. Schaubroeck, R. A. F. Alvarenga, K. Verheyen, B. Muys, J. Dewulf, Envi-ron. Sci. Technol. 2013, 47, 13578 – 13586.

[33] W. De Soete, J. Dewulf, P. Cappuyns, G. Van der Vorst, B. Heirman, W.Aelterman, K. Schoeters, H. Van Langenhove, Green Chem. 2013, 15,3039 – 3048.

[34] J. Provo, J. Fava, S. Baer, Curr. Opin. Chem. Eng. 2013, 2, 278 – 281.[35] D. Prat, O. Pardigon, H.-W. Flemming, S. Letestu, V. Ducandas, P. Isnard,

E. Guntrum, T. Senac, S. Ruisseau, P. Cruciani, P. Hosek, Org. Process Res.Dev. 2013, 17, 1517 – 1525.

[36] Green Chemistry – Sanofi, CSR Factsheet, sanofi.com, 2013.[37] R. Frischknecht, R. Heijungs, P. Hofstetter, Int. J. Life Cycle Assess. 1998,

3, 266 – 272.[38] VDI-Standard: VDI 4600 Cumulative Energy Demand (KEA)—Terms, Defi-

nitions, Methods of Calculation, The Association of German Engineers(VDI), D�sseldorf, 2012.

[39] M. A. J. Huijbregts, S. Hellweg, R. Frischknecht, H. W. M. Hendriks, K.Hungerbuehler, A. J. Hendriks, Environ. Sci. Technol. 2010, 44, 2189 –2196.

[40] Final Report of the SETAC-Europe Screening and Streamlining Working-Group, Simplifying LCA: Just a Cut?, Society of Environmental Chemistryand Toxicology SETAC, Brussels, 1997.

[41] G. Fleischer, W.-P. Schmidt, Int. J. Life Cycle Assess. 1997, 2, 20 – 24.[42] M. A. J. Huijbregts, L. J. A. Rombouts, S. Hellweg, R. Frischknecht, J. Hen-

driks, D. Van de Meent, A. M. J. Ragas, L. Reijnders, J. Struijs, Environ. Sci.Technol. 2006, 40, 641.

[43] H. Sugiyama, M. Hirao, R. Mendivil, U. Fischer, K. Hungerbuehler, ProcessSaf. Environ. Prot. 2006, 84, 63 – 74.

[44] G. Wernet, S. Hellweg, U. Fischer, S. Papadokonstantakis, K. Hunger-buehler, Environ. Sci. Technol. 2008, 42, 6717 – 6722.

[45] G. Wernet, S. Papadokonstantakis, S. Hellweg, K. Hungerbuehler, GreenChem. 2009, 11, 1826 – 1831.

[46] Ullmann’s Encyclopedia of Industrial Chemistry, 7th ed. , Wiley-VCH, Wein-heim, 2009.

[47] R. S. Downing, P. J. Kunkeler, H. Van Bekkum, Catal. Today 1997, 37,121 – 136.

[48] A. M. Str�tz in Catalysis of Organic Reactions (Eds. : J. R. Kosak), MarcelDekker, New York, 1984, pp. 335.

[49] I. Dencic, D. Ott, D. Kralisch, T. Noel, J. Meuldijk, M. De Croon, V. Hessel,Y. Laribi, P. Perrichon, Org. Proc. Res. Devel. 2014, DOI : 10.1021/op5000573.

[50] D. Kralisch, I. Streckmann, D. Ott, U. Krtschil, E. Santacesaria, M. Di Serio,V. Russo, L. De Carlo, W. Linhart, E. Christian, B. Cortese, M. H. J. M. de C-roon, V. Hessel, ChemSusChem 2012, 5, 300 – 311.

[51] D. Kralisch, C. Staffel, D. Ott, S. Bensaid, G. Saracco, P. Bellantoni, P.Loeb, Green Chem. 2013, 15, 463 – 477.

[52] D. Kralisch, D. Ott, S. Kressirer, C. Staffel, I. Sell, U. Krtschil, P. Loeb, GreenProcessing and Synthesis 2013, 2, 465 – 478.

[53] SciFinder, scifinder.cas.org.[54] Web of Science (ISI Web of Knowledge), thomsonreuters.com.[55] Espacenet Patent search, worldwide.espacenet.com.[56] Ullmann’s Encyclopedia of Industrial Chemistry, Wiley-VCH, Weinheim,

2009.[57] M. Wells, Handbook of Petrochemicals and Processes, 2nd ed. , Ashgate

Publishing, Aldershot, England; Brookfield, VT 1999.[58] R. Hischier, S. Hellweg, C. Capello, A. Primas, Int. J. Life Cycle Assess.

2005, 10, 59 – 67.[59] H.-J. Althaus, G. Doka, R. Dones, T. Heck, S. Hellweg, R. Hischier, T. Nem-

ecek, G. Rebitzer, M. Spielmann, G. Wernet, Overview and Methodology,Data v2.0, Ecoinvent Report No. 1 (Eds. : R. Frischknecht, N. Jungbluth),Ecoinvent Centre, D�bendorf, 2007.

[60] T. Rosenberger-S�ß, G. Mayer, A. Schl�ter, U. Siebert, Umwelterkl�rung2012, Industriepark Werk GENDORF, InfraServ, Burgkirchen, 2011.

[61] C. Capello, S. Hellweg, B. Badertscher, K. Hungerbuehler, Environ. Sci.Technol. 2005, 39, 5885 – 5892.

[62] Ecoinvent Centre, Ecoinvent Data v2.2. Ecoinvent Reports Nos. 1 – 25,Swiss Centre for Life Cycle Inventories, D�bendorf, 2010.

[63] E. S. Borovinskaya, V. P. Reshetilowskii, Russian J. Appl. Chem. 2011, 84,1094 – 1104.

[64] D. Kralisch, G. Kreisel, Chem. Eng. Sci. 2007, 62, 1094 – 1100.[65] Umberto v.5.6, ifu Institut fuer Umweltinformatik, Hamburg; ifeu Institut

f�r Energie- und Umweltforschung, Heidelberg, 2010.[66] M. Goedkoop, R. Heijungs, M. Huijbregts, A. D. Schryver, J. Struijs, R. v.

Zelm, ReCiPe 2008 – A Life Cycle Impact Assessment Method Which Com-prises Harmonised Category Indicators At The Midpoint And The EndpointLevel, 1st ed. , Report I : Characterisation, Ministerie van Volkshuisvesting,Ruimtelijke Ordering en Milieubeheer, The Hague, 2009.

[67] I. C. Kemp in Modern Drying Technology Vol. 4: Energy Savings (Eds. : E.Tsotsas, A. S. Mujumdar), Wiley-VCH, Weinheim, 2012, pp. 1 – 45.

[68] Expected Dryer Efficiencies in: Energy Manager Training, Best PracticeManual : Dryers, energymanagertraining.com.

[69] S. Kressirer, D. Kralisch, A. Stark, U. Krtschil, V. Hessel, Environ. Sci. Tech-nol. 2013, 47, 5362 – 5371.

[70] R. Hoogenboom, T. F. A. Wilms, T. Erdmenger, U. S. Schubert, Aust. J.Chem. 2009, 62, 236 – 243.

[71] C. Berguerand, A. Yarulin, A. Olasolo Alonso, I. Yuranov, L. Kiwi-Minsker,unpublished results.

[72] M. Consonni, D. Jokic, D. Y. Murzin, R. Touroude, J. Catal. 1999, 188,165 – 175.

[73] H.-U. Blaser, Chimia 2010, 64, 65 – 68.[74] C. Capello, U. Fischer, K. Hungerbuehler, Green Chem. 2007, 9, 927 – 934.[75] C. Capello, S. Hellweg, B. Badertscher, H. Betschart, K. Hungerbuehler, J.

Ind. Ecol. 2007, 11, 26 – 38.[76] C. Capello, S. Hellweg, K. Hungerbuehler, J. Ind. Ecol. 2008, 12, 111 – 127.[77] D-Sight Web, d-sight.com, Brussels, 2014.[78] M. Classen, H.-J. Althaus, Ecoinvent Reports, Part V: Platinum Group

Metals (PGM)—Platinum, Palladium, Rhodium, Data v2.1, EMPA, D�ben-dorf, 2009.

[79] M. Pehnt, Ganzheitliche Bilanzierung von Brennstoffzellen in der Energie-und Verkehrstechnik, VDI-Verlag Fortschrittsberichte Reihe 6 Nr. 476,2002.

[80] �ko-Institut e. V. , Daimler AG, TU Clausthal, Umicore, Ressourceneffi-zienz und ressourcenpolitische Aspekte des Systems Elektromobilit�t—Arbeitspaket 7 des Forschungsvorhabens OPTUM: Optimierung derUmweltentlastungspotenziale von Elektrofahrzeugen, oeko.de, Darm-stadt, Germany, 2011.

[81] I. Sell, D. Ott, D. Kralisch, ChemBioEng Rev. 2014, 1, 50 – 56.[82] J.-N. Tourvieille, R. Philippe, C. De Bellefon, unpublished results (http://

tel.archives-ouvertes.fr/docs/01/01/50/51/PDF/TH2014_Tourvieille_Jean-Noel.pdf).

Received: April 15, 2014Revised: June 26, 2014Published online on September 22, 2014


http://dx.doi.org/10.1039/c3gc40225hhttp://dx.doi.org/10.1039/c3gc40225hhttp://dx.doi.org/10.1039/c3gc40225hhttp://dx.doi.org/10.1039/c3gc40225hhttp://dx.doi.org/10.1039/c0gc00918khttp://dx.doi.org/10.1039/c0gc00918khttp://dx.doi.org/10.1039/c0gc00918khttp://dx.doi.org/10.1039/c0gc00918khttp://dx.doi.org/10.1021/op100327dhttp://dx.doi.org/10.1021/op100327dhttp://dx.doi.org/10.1021/op100327dhttp://dx.doi.org/10.1021/op100327dhttp://dx.doi.org/10.1007/s11367-010-0151-zhttp://dx.doi.org/10.1007/s11367-010-0151-zhttp://dx.doi.org/10.1007/s11367-010-0151-zhttp://dx.doi.org/10.1021/op200097dhttp://dx.doi.org/10.1021/op200097dhttp://dx.doi.org/10.1021/op200097dhttp://dx.doi.org/10.1021/op200097dhttp://dx.doi.org/10.1021/es4046633http://dx.doi.org/10.1021/es4046633http://dx.doi.org/10.1021/es4046633http://dx.doi.org/10.1021/es4046633http://dx.doi.org/10.1039/c3gc41185khttp://dx.doi.org/10.1039/c3gc41185khttp://dx.doi.org/10.1039/c3gc41185khttp://dx.doi.org/10.1039/c3gc41185khttp://dx.doi.org/10.1016/j.coche.2013.05.001http://dx.doi.org/10.1016/j.coche.2013.05.001http://dx.doi.org/10.1016/j.coche.2013.05.001http://dx.doi.org/10.1021/op4002565http://dx.doi.org/10.1021/op4002565http://dx.doi.org/10.1021/op4002565http://dx.doi.org/10.1021/op4002565http://dx.doi.org/10.1007/BF02979833http://dx.doi.org/10.1007/BF02979833http://dx.doi.org/10.1007/BF02979833http://dx.doi.org/10.1007/BF02979833http://dx.doi.org/10.1021/es902870shttp://dx.doi.org/10.1021/es902870shttp://dx.doi.org/10.1021/es902870shttp://dx.doi.org/10.1007/BF02978711http://dx.doi.org/10.1007/BF02978711http://dx.doi.org/10.1007/BF02978711http://dx.doi.org/10.1021/es051689ghttp://dx.doi.org/10.1021/es051689ghttp://dx.doi.org/10.1205/psep.04142http://dx.doi.org/10.1205/psep.04142http://dx.doi.org/10.1205/psep.04142http://dx.doi.org/10.1205/psep.04142http://dx.doi.org/10.1021/es7022362http://dx.doi.org/10.1021/es7022362http://dx.doi.org/10.1021/es7022362http://dx.doi.org/10.1039/b905558dhttp://dx.doi.org/10.1039/b905558dhttp://dx.doi.org/10.1039/b905558dhttp://dx.doi.org/10.1039/b905558dhttp://dx.doi.org/10.1016/S0920-5861(97)00005-9http://dx.doi.org/10.1016/S0920-5861(97)00005-9http://dx.doi.org/10.1016/S0920-5861(97)00005-9http://dx.doi.org/10.1016/S0920-5861(97)00005-9http://dx.doi.org/10.1002/cssc.201100445http://dx.doi.org/10.1002/cssc.201100445http://dx.doi.org/10.1002/cssc.201100445http://dx.doi.org/10.1039/c2gc36410ghttp://dx.doi.org/10.1039/c2gc36410ghttp://dx.doi.org/10.1039/c2gc36410ghttp://dx.doi.org/10.1065/lca2004.10.181.7http://dx.doi.org/10.1065/lca2004.10.181.7http://dx.doi.org/10.1065/lca2004.10.181.7http://dx.doi.org/10.1065/lca2004.10.181.7http://dx.doi.org/10.1021/es048114ohttp://dx.doi.org/10.1021/es048114ohttp://dx.doi.org/10.1021/es048114ohttp://dx.doi.org/10.1021/es048114ohttp://dx.doi.org/10.1134/S107042721106036Xhttp://dx.doi.org/10.1134/S107042721106036Xhttp://dx.doi.org/10.1134/S107042721106036Xhttp://dx.doi.org/10.1134/S107042721106036Xhttp://dx.doi.org/10.1016/j.ces.2006.11.009http://dx.doi.org/10.1016/j.ces.2006.11.009http://dx.doi.org/10.1016/j.ces.2006.11.009http://dx.doi.org/10.1021/es400085yhttp://dx.doi.org/10.1021/es400085yhttp://dx.doi.org/10.1021/es400085yhttp://dx.doi.org/10.1021/es400085yhttp://dx.doi.org/10.1071/CH08503http://dx.doi.org/10.1071/CH08503http://dx.doi.org/10.1071/CH08503http://dx.doi.org/10.1071/CH08503http://dx.doi.org/10.1006/jcat.1999.2635http://dx.doi.org/10.1006/jcat.1999.2635http://dx.doi.org/10.1006/jcat.1999.2635http://dx.doi.org/10.1006/jcat.1999.2635http://dx.doi.org/10.1039/b617536hhttp://dx.doi.org/10.1039/b617536hhttp://dx.doi.org/10.1039/b617536hhttp://dx.doi.org/10.1111/j.1530-9290.2008.00009.xhttp://dx.doi.org/10.1111/j.1530-9290.2008.00009.xhttp://dx.doi.org/10.1111/j.1530-9290.2008.00009.xhttp://dx.doi.org/10.1002/cben.201300007http://dx.doi.org/10.1002/cben.201300007http://dx.doi.org/10.1002/cben.201300007www.chemsuschem.org

Documents

Life cycle analysis within pharmaceutical process ... · development. Nevertheless, it is highly promising to rethink, redesign, and optimize process strategies as early as possible