Life Cycle Assessment Comparison of Acrylonitrile Production Methods

Journal of Cleaner Production 69 2014 1725 pd Contents lists available at ScienceDirect A Journal of Cleaner Production Ti ata ELSEVIER journal homepage wwwelseviercomlocatejclepro ae Life Cycle Assessment comparison of two ways for acrylonitrile CroseMark production the SOHIO process and an alternative route using propane Daniele Cespi Fabrizio Passarini Esmeralda Neri Ivano Vassura Luca Ciacci Fabrizio Cavani Department of Industrial Chemistry Toso Montanari ALMA MATER Studiorum University of Bologna Viale del Risorgimento 4 40136 Bologna Italy ARTICLE INFO ABSTRACT Article history The aim of this study is to apply the LCA methodology to the industrial chemical sector in order to Received 31 July 2013 compare the traditional process for acrylonitrile production from propylene with alternative routes Received in revised form starting from propane while assessing which one is the cleaner production in terms of sustainability 13 January 2014 from a life cycle perspective The model created refers to the production of 1 kg of acrylonitrile System Accepted u January 2014 boundaries of each scenario include all mass flows into and out of the reactor all mass and energy flows Available online 28 January 2014 into and out of the heat exchanger of the fluid bed the amount of raw material for the production of each ye catalyst the avoided impacts resulting from energy and mass recovery and all transportations Also Keywords sya Acrylonitrile average infrastructure processes that refer to land use building and disposal of the chemical plant were Ammoxidation of propylene not included The life cycle evaluation was performed using the ReCiPe 2008 method v107 showing Ammoxidation of propane results in terms of midpoint categories such as Climate Change Fossil Fuel Depletion and Metal Catalyst Depletion The results from the inventory show that alternative synthetic routes starting from propane Chemical process have higher impacts than the traditional processes in terms of fossil fuel depletion and climate change Environmental sustainability categories due to higher consumption of reactants caused by the lower efficiency of catalytic systems Conversely impacts associated with the metal depletion category have an irregular trend due mainly to the extraction of different percentages of resources for the catalyst production The results were vali dated by a sensitivity analysis using the Monte Carlo method This study suggests that the LCA meth odology may be used as a scientific approach to identify the environmental issues associated with the chemical production of a product In particular it is useful in comparing alternative ways of synthesis and evaluating which process is more sustainable and which production stage should be improved in order to ensure greater environmental sustainability 2014 Elsevier Ltd All rights reserved 1 Introduction to the environment and aiming for prevention as the inspiring feature in every human activity JiménezGonzalez and Constable The chemical industry is one of the sectors that contribute most 2011 In this sense ecoefficiency may be intended as synony to the economy in terms of revenues trade and employment while mous with industrial efficiency reducing the consumption of energy chemistry is considered to be at the forefront of the transition to a and resources and orienting processes toward a selective production more sustainable development as it takes part in all economies do put the ecodesign principles into practice with the prefix eco through the furnishing of products Sus Chem 2013 applying in both ecological and economic terms The need to move away from the old anthropic production led to The 20th century has been defined among other ways as the the definition of Green Chemistry and the publishing of the Twelve plastic century due to the primary role played by polymers in Principles of Green Chemistry which claim to support sustainability influencing the human culture and way of living While considering through for instance avoiding or limiting the use and production of the wide use of plastics and rubbers in our society as critical ele hazardous substances enhancing heat recovery and energy effi ments in the depletion of fossil fuels it is not surprising that among ciency improving the closure of waste flows and reducing emissions the most promising fields of improvement for a greener chemistry is the study of alternative building blocks and innovative processes to replace traditional feedstock routes from oil Cavani 2010a Corresponding author Tel 39 0512093863 Acrylonitrile ACN is an example of a chemical whose use Email address fabriziopassariniuniboit F Passarini increased dramatically after its first application in plastic and 09596526 see front matter 2014 Elsevier Ltd All rights reserved httpdxdoiorg101016jjclepro201401057 18 D Cespi et al Journal of Cleaner Production 69 2014 1725 rubber manufacturing in the 1930s The production of acrylic fibers in replacing materials with other renewablesbased compounds La acrylonitrilebutadienestyrene ABS adiponitrile nitrile Rosa et al 2013 and it was also used in combination with costs butadiene NB copolymers acrylamide for watertreatment poly evaluation eg life cycle costing in chemical industries Kawauchi mers and carbon fibers are among the main applications that have and Rausand 2002 LCA and marginal prevention cost applied to determined the commercial significance of ACN Brazdil 2012 In acrylonitrile plant MoralesMora et al 2012 and risk assessment 2010 world ACN production amounted to 6 Mt and a 37 annual tools eg combining LCA methodology with Risk Phrase Askham growth over the years 20082018 mainly driven by ABS and et al 2013 the use of LCA and risk assessment tool to identify styreneacrylonitrile SAN resin manufacturing has been esti chemicals and workplace that will require more investigations mated Brazdil 2012 Garmston 2009 About 90 of the total ACN Demou et al 2011 to suggest improvements in the green chem production follows the Standard Oil of Ohio SOHIO process which icals using LCA and risk assessment tool Herrchen and Klein 2000 is based on the ammoxidation of propylene The reaction is highly Furthermore in the coming years the trend may be expected to selective fast and leads to efficient yields of ACN requiring no increase in particular regarding the life cycle of processes sector further recycling steps Langvardt 2011 In any case the cost of which is growing with respect to a single product assessment ACN production has recently increased due to the market price of Jacquemin et al 2012 reagents in particular propylene which accounts for more than Specifically in this study five scenarios describing different ACN 70 of the total cost Brazdil 2006 Thus there is an increasing production routes were modeled and compared in terms of interest in finding alternative more economic ways to produce contribution to midpoint impact categories Climate Change ACN Specifically propane ammoxidation seems to be the most damage on human health CH and on ecosystem CE Fossil Fuel promising alternative process Langvardt 2011 In 2012 PTT Asahi Depletion FD and Metal Depletion MD One kilogram of ACN was Kasei Chemical Co Ltd a joint venture of PTT Asahi Kasei Chemical assumed as the reference flow while the impact analysis was car and Marubeni started up the first acrylonitrile production facility ried out using SimaPro 733 software Pré Consultant 2010 and in Thailand The plant which achieves an acrylonitrile productivity the ReCiPe 2008 method v107 Goedkoop et al 2012 of about 200000 ty also includes in project the production of methyl methacrylate and ammonium sulfate about 70000 ty and 11 Ammoxidation reactions to acrylonitrile 160000 ty respectively Chemsystems 2013 Also the use of propane would imply a significant cost advan Ammokxidation reaction consists of a catalytic oxidation of hy tage over propylene the difference between propylene and pro drocarbons in the presence of ammonia to produce organic nitriles pane was estimated in the range of 9001000 per Mt in the year and water where mixed metal oxides with or without support are 2012 Dow 2011 Although it might be argued that such a differ used as catalysts Chauvel and Lefebvre 1989 Typical reactants are ence would not be stable enough because it is influenced by the alkenes The reaction involves three main steps hydrocarbons fluctuations in market prices anyhow it is certain that the change oxidation to form intermediates on the active sites nitrogen from propylene to propane ammoxidation produces a reduction in insertion and oxidative dehydrogenation of the Nbonded species the number of steps required as the latter is commonly produced Cavani et al 2009 A description of both the conventional SOHIO by a direct fractional distillation of petroleum whereas with pro process and the most innovative routes for the production of ACN is pylene cracking or catalytic dehydrogenation operations are reported hereafter necessary Moreover it is reported that the direct transformation of the alkane into acrylonitrile affords a more rational use of energy 111 The SOHIO process propylenebased scenario since the strongly endothermic process for propylene production is The INEOS Technologies ACN technology also known as the avoided Cavani and Teles 2009 Moreover the development of SOHIO process involves the catalytic oxidation of propylene in the new catalysts will lead to an improved exploitation of light alkanes presence of ammonia and air as schematically depicted in the by means of processes conducted in mild condition Cavani 2010b Equation A1 The classes of catalyst used in this process are Multi Aiming to evaluate whether such a difference between pro Metal Molybdate MMM made up of Mo Bi Fe Ni Co and addi cesses might eventually entail a reduction of loads to the envi tives eg Cr Mg Rb K Cs P B Ce Sb and Mn dispersed in silica ronment or in other words whether the alternative 50 ww and Antimonate with rutile structure made up of four ammoxidation process is greener than the conventional one the metal antimonate cations and a redox couple of Fe Ce U and Cr Life Cycle Assessment LCA methodology was used to evaluate the Cavani et al 2009 These classes of catalyst led to an increase in cleaner production in terms of sustainability from a life cycle the production of ACN starting from 1960 Grasselli 2002 for this perspective LCA is an internationally standardized tool ISO 14040 reason they were deeply investigated in literature in order to 140442006 that helps to identify and quantify environmental identify their physical and chemical structures and the proposed pressures benefits balances and potentials for achieving im mechanisms for propylene and propane ammoxidation reaction provements along the whole life cycle of goods and products The Grasselli 2005 application of the Life Cycle Assessment in the chemical sector is The reaction is conducted in a fluidizedbed reactor system not new In fact it is well known that LCA methodology is useful for operating at 30200 kPa and with a temperature range of 400 comparing different processes in order to better understand the 500 C Brazdil 2012 environmental impacts at all stages of the production process of a chemical substance Anastas and Lankey 2000 LCA has already CH3CHCH2 3202 NH3 CH2CHCN 3H20 A1 been used to investigate several aspects of the chemical sector such as for example the production process of different products and The reaction is highly exothermic AHp 515 kJmol Chauvel their best recovery option polymer materials AlSalem et al and Lefebvre 1989 thus the fluidizedbed configuration is neces 2012 dimethylcarbonate Aresta and Galatola 1999 maleic an sary to remove the heat in excess and achieve a propylene con hydride Doménech et al 2002 cellulose mnanowhiskers version higher than 95 as well as a selective yield in ACN close to Figueirédo et al 2013 carbon nanotubes Griffiths et al 2013 80 Also numerous steam coils are contained inside the reactor to acrylic acid Holman et al 2009 sodium chromate Kowalski et al keep the temperature as constant as possible in order to prevent 2007 astaxanthin PérezLopez et al 2013 in oil refining and the runaway phenomena Cavani et al 2009 Usually the heat pharmaceutical sector Portha et al 2010 ab Jodicke et al 1999 recovered is used to produce highpressure steam Langvardt D Cespi et al Journal of Cleaner Production 69 2014 1725 19 2011 Reactants are fed separately in order to avoid homoge Also the propane ammoxidation reaction A2 is extremely neous reactions with high purity eg 90 for propylene and exothermic AH at reaction temperature is around 813 kJmol it is 995 for ammonia Cavani et al 2009 and with a molar ratio close conducted in a fluidizedbed reactor under the same process condi to the stoichiometric one A1 Higher ammoniapropylene molar tions temperature and pressure as the SOHIO process as well as with ratios are not common because of the occurrence of side reactions the use of procedures that neutralize unreacted ammonia and recover The hot effluent outputs of the fluidizedbed reactor are acrylonitrile from byproducts Fig 2 shows a simplified scheme of quenched in a water absorber here unreacted ammonia is rapidly propanebased process to ACN Botha higher performance in terms of neutralized by sulfuric acid to produce ammonium sulfate Chauvel selectivity and conversion from molybdate catalysts and the lower and Lefebvre 1989 while unreacted propylene is vented The ACN price of propane make the configuration without recirculation pref produced is recovered from the organic phase and then purified of erable Fig 2 The recycling option may become interesting when a hydrogen cyanide and heavy impurities The main byproducts are high propane conversion rate is not achieved Cavani et al 2009 acetonitrile and hydrogen cyanide 002011 kg and 015020 kg The next section offers a detailed description of scenarios per kg of ACN respectively which could be recovered and used for modeled and compared in the LCA methodology other applications eg solvent raw material Brazdil 2012 A schematic view of the process is shown in Fig 1 2 Materials and methods 112 Alternative synthetic process propanebased scenario 21 Introduction to the LCA methodology The general stoichiometry of the reaction based on propane feedstock is shown in Equation A2 Life Cycle Assessment tool is an objective and standardized methodology able to investigate the environmental behavior of CH3CH2CH3 20 NH3 CHCHCN 4H20 A2 product process or system during the entire life cycle The general LCA framework defined by the ISO 14040 and Many companies are involved in the industrial conversion of 14044 consists of four conceptual phases namely Goal and Scope propane to ACN and although the synthetic reaction is the same definition Life Cycle Inventory LCI Life Cycle Impact Assessment each of them has undertaken different production routes with LCIA and Interpretation The simulation is carried out using LCA distinct process conditions The first patent developed by Standard software accredited database and analysis method able to identify Oil now BP plc uses propanerich conditions while both the environmental pressures and benefits associated to the ecosystem Mitsubishi Chemical Corporation MCC and Asahi Kasei Chemicals and the human health In accordance with the ISO series principles Corporation AKCC use propanelean compositions Cavani et al and recommendations in the next paragraphs the main assump 2009 Major differences between these two approaches lie in tions and data collected for the scenario modeling are discussed both the propane conversion which is lower in the former case and the choice of catalyst MCC and AKCC use a molybdatebased 22 System boundaries and functional unit catalyst so that the system may achieve higher selectivity and yield while BP has preferred an antimonate catalyst However ina The goal and scope definition represents the first stage of an later patent BP also suggested the use of propanelean conditions LCA in which researchers define the aim of the study by identifying which in any case require higher temperatures around 50 C the system boundaries and the reference unit for all flows In this higher Both catalytic systems used are defined as belonging to the study LCA methodology was applied as a scientific tool aimed at the multifunctional system category due to their structure with identification of the cleaner technology in the production of ACN different kinds of active sites that have the ability to produce ACN For this reason the system boundaries cover single unit processes via a propylene intermediate Nevertheless the catalysts for pro involved in each scenario for the ACN production Specifically they pane ammoxidation differ from the systems used in the ammox include energy and mass flow into and out of the reactor reactants idation of propylene due to the presence of an oxidative enhancer auxiliary chemicals electricity and heat consumed for utilities a halogen promoter or a strong oxidant such as Vanadium emissions into the air and water energy dissipation all mass and Fluidbed Absorber Acrylonitrile Acetonitrile Lights Product reactor recovery recovery column column column column Crude Product acrylonitrile Crude acrylonitrile Of eas Crude a acetonitrile H0 HP steam BFW Air Ammonia H0 Propene Reavy impurities Fig 1 SOHIO process adapted from Brazdil 2010 energy flows into and out of the heat exchanger of the fluidized bed the amount of raw material for the production of catalysts transportation processes and benefits resulting from energy and mass recovery expressed as avoided impacts Average infra structure processes were not considered because of their low representativeness in quality for database data Also the typical lifetime of a chemical plant is commonly very long so any envi ronmental loads connected with the functional unit chosen would in this case be negligible System boundaries are plotted in Fig 3 One kilogram of ACN produced was assumed as the reference flow and thus used as a base to measure inputs and outputs of the systems studied As reported previously the ammoxidation pro cesses produce acetonitrile and HCN as main byproducts Both species are recovered and used in downstream applications either as a solvent acetonitrile or as a reactant for the synthesis of other chemicals such as acetone cyanhydrine which is the intermediate for methylmetacrylate synthesis However although the integration with other processes may greatly contribute to the success of a chemical production the further extension of the system boundaries including also the downstream use of byproducts would greatly increase both the system complexity and the uncertainty of the final result due to the lack of detailed data for the additional inputs This extension would require a specific study which is intended to be tackled as a widening of the present work Therefore on the basis of each selectivity rate for ACN it has been decided to apply mass allocation criteria to ACN production only thus including any up stream transformation but without taking into account acetonitrile and HCN management and related burdens Mass shares for each scenario investigated were estimated in accordance with both company reports and process specifications reported in patents 23 Inventory analysis The inventory analysis represents the more timeconsuming phase of the entire methodology It consists in the data collection to create a model which should be able to depict as objectively as possible the system studied In this study LCI includes data collection for five scenarios by modeling different ACN production routes the first scenario describes the conventional SOHIO process from propylene while the others focused on alternative processes having propane as the raw material Each scenario is characterized by different catalyst systems and process specifications eg selectivity conversion and yield that influence mass and energy balances The modeling phase was carried out using SimaPro 733 software Pré Consultant 2010 A further description of the sce narios is reported below 231 Scenario 1 the SOHIO process Two propylene production methods were primarily investi gated one in which the olefin is produced by naphtha steam cracking known also as thermal cracking and the other in which propylene is synthesized by catalytic cracking Fluid Catalytic Cracking FCC The former entails the use of heat to obtain the desired product the Ecoinvent process Propylene at plantRER is set up as a default process for propylene production in the SimaPro software Ecoinvent Centre 2009 The FCC process is characterized by lower process temperatures while the cracking reaction is car ried out using zeolite as the catalyst As the process was not present in the database a specific inventory was created by using mass and energy balances Plastics Europe 2005 and the Ecoinvent process Zeolite powder at plantRER was used to model the catalyst pro duction Ecoinvent Centre 2009 An average amount of catalyst was estimated on the basis of literature reports 097 kgton of propylene produced Raseev 2003 No catalyst regeneration stages were included The screening comparison between the two propylene pro duction procedures showed for the FCC process a 10 reduction of total impacts thanks to a lower operating temperature thus cata lytic cracking was used in Scenario 1 to simulate the production process for propylene supply Data inventory for the ammoxidation of propylene to ACN was collected from literature which refers to an existing plant At the industrial scale a small excess of ammonia with respect to the stoichiometric procedure is necessary ammoniato propylene molar ratio varies between 105 and 12 while the range 19e21 applies to the airpropylene share usually oxygen enriched air is used Cavani et al 2009 The catalyst was modeled considering the typical empirical structure of molyb dates KCs01NiMgMn75FeCr23Bi05Mo12Ox The excess of molybdenum is important for the catalyst performance because it functions as a molecular bridge from the molybdates and provides a reserve of the metal that is partly depleted during the Fig 3 System Boundaries of the study Fig 2 Propane ammoxidation recycleno recycle configuration adapted from Brazdil 2006 D Cespi et al Journal of Cleaner Production 69 2014 17e25 20 D Cespi et al Journal of Cleaner Production 69 2014 1725 21 redox cycle Cavani et al 2009 This type of active phase is or tc AH supported over SiOz typically in the range of 50 of the entire mmc3 moles of propene or prapane reacted A4 catalyst weight Considering data on the plant productivity and AH heat of reaction catalyst makeup 07 kgt of ACN IPPC 2003 an amount of 1 g of catalyst per kg of acrylonitrile produced was estimated The u e model designed for the catalyst assumed raw material inputs Qs MH0CPH0 Tho Tho necessary for assembling the system while the amount of each My0 moles of water in coils element was calculated on the basis of the catalyst stoichiom Cpy0 heat capacity of water in coils etry As the model the composition of the catalyst C49MC Tyo temperature of water input e and output u the coils developed by Standard Oil in 1991 was chosen Friedrich et al A5 1991 Ko15 Cso05 Ni4o Coo5 M825 Fe20 Bios Wo5 Moi2 Ox and Loa SiOz 50 ww The process includes the extraction phase only The model created assumes that the heat exchanged in coils is No information was available on production steps for which ecovered with an efficiency of 50 652E 03 kj and is used for impacts and flows were assumed to be negligible from the plant utilities half of it to produce heat 326E 03 kJ and the rest perspective of the entire life cycle Again the phase of catalyst converted into electricity 281E 01 kWh with a 31 conversion regeneration was not taken into account due to its negligible ficiency Domenech et al 2002 impact on the total value All details about the SOHIO process inventory are grouped in ACN selectivity and propylene conversion yield were assumed Table 1 to be equal to 83 and 98 respectively Brazdil 2012 deter mining an 81 molar yield for ACN Molar proportions and pro 232 Scenario2 propane ammoxidation to acrylonitrile cess specifications were used to calculate the reactant amounts alternative synthetic process put into the reactor ie 086 kg of propylene and 039 kg of As previously reported shifting from propylene to propane in ammonia per kg of product Furthermore emissions were estl volves a reduction in the production steps which could produce an mated by assuming that the unreacted olefin 1S burnt with 999 environmental benefit In particular a screening comparison shows combustion efficiency about 002 kg while the remaining a cut in the total impact of about 16 if propane production is fraction escapes into the air as fugitive leakage 172E05 kg In compared with the catalytic cracking operation to manufacture order to evaluate the cleaner production we performed the propylene this difference is much more appreciable as concerns comparison on the basis of the best available techniques on the climate change for which the environmental load is reduced of market and we assumed that the heat from combustion 1S about 42 Therefore to assess whether this reduction is main recovered and used for plant utilities ie steam and electricity tained even during ammoxidation reaction different scenarios the model counts environmental benefits from enersy Tecovely were created aiming to model alternative routes for ACN produc such as avoided impacts from natural gas 0022 m extraction tion from propane ammoxidation As just mentioned only the and use Energy produced from the propylene combustion and AKCC started to develop a commercial process for the production of LHV Low Heating Value values were used to estimate the ACN from propane ammoxidation by modifying an existing plant amount of natural gas avoided i Cavani et al 2009 However many companies are looking into the Model assumed that 80 of unreacted ammonia is neutralized replacement of propylene with alkanes mainly by exploring 004 kg Mass balances were used to calculate the amount of different combinations of catalytic systems and reaction conditions sulfuric acid consumed during the neutralization phase and the Thus four scenarios were modeled according to the most advanced resulting ammonium sulfate produced 012 kg and 016 kg results and called AKCC MCC BP poor lowpropane concentra respectively Considering that ammonium sulfate is recovered in a tion and BP rich propanerich concentration Each of them refers solution to about one third by weight models include the energy to a specific process and to the catalysts developed by the com consumed for the production of salt as a byproduct deriving from panies The modeling phase was performed using information and the main operations water evaporation centrifugation and data reported on patents which remain constant even on an in dehydration Nemecek and Kagi 2007 It was assumed that dustrial scale reaction conditions catalyst composition process ammonium sulfate is sold as fertilizer so the avoided impact specifications yield and selectivity and feed molar ratio for each deriving from the production of a nitrogen fertilizer was taken into system As reported in literature a double catalyst makeup with account in our scenarios Lastly the remaining ammonia around respect to the scenario from propylene was assumed Pavone and 20 of the unreacted amount about 001 kg was assumed to be Schwaar 1989 entailing a consumption of 17 g of catalyst per kg oxidized to produce N2 and H20 of ACN produced Further details on these scenarios are given in Both the heat exchanged and the water amount used in reactor Table 2 coils were calculated by using energy balance respectively 764 kg The amount of propane necessary for the reaction was esti and 130E 04 kJ The equations shown below apply to each mated from the ACN yield reported In the LCA model alkane scenario production as a fraction of the distillation of petroleum and wot yn Cp T TOR ym Cpl T TSR Q Qs 0 m moles of each substance input e and output u of the system Cp heat capacity of each substance input e and output u of the system A3 T temperature input e and output u of the system TOR temperature of reference state 0C 22 D Cespi et al Journal of Cleaner Production 69 2014 1725 Table 1 3 Results and discussions Inventory analysis for propene ammoxidation scenario Processes Unit Scenario Impact analysis was carried out using the ReCiPe 2008 method SOHIO22OOCOCS v107 which was followed for the assessment of environmental DOAT we burdens for the midpoint categories CH CE FD and MD Goedkoop Feed molar ratio CaNHaairinertH20 101120 et al 2012 The decision to neglect other midpoint impact cate Catalyst in silica SOwt Ko15 Cso05 Niao Coo5 Mg25 Fe 9 Bigs Wos M012 Oy gories is related to aim of the study In fact chemical processes Catalyst amount g 10 involve very wide geographical areas considering those in which ACN selectivity 830 raw materials are extracted and sold transportation and the final ACN yield 810 use for synthetic purposes So in our opinion fossil fuel depletion Propylene conversion 980 Propylene input kg 086 metal depletion and climate change are the main important cate Propylene burned 999 efficiency kg 002 gories for this study since they represents damages on a global Propylene in air 001 kg 172E05 scale Ammonia input kg 039 These midpoint scores may further be grouped into three end Neutralized ammonia 80 of unreacted kg 004 points based on Damages to Human Health units of measurement Oxidized ammonia 20 of unreacted kg 001 Sulfuric acid input ke 012 disability adjusted life years DALYs Ecosystem Quality Ammonium sulfate kg 016 measured in potentially disappeared fractions of species spe Heat recovered KJ 326E 03 ciesyr and Resource Consumption in terms of increased costs of se an ae extraction 8 Sativa 88 avonee The results of the characterization analysis at the midpoint level are reported in Table 3 Fig 4 shows the results of the character naphtha was assumed considering it to be a cheaper raw material ization analysis in a radar chart while Fig 5 offers a comparison for synthesizing ACN instead of using it to produce propylene by among the five ammoxidation scenarios in terms of ReCiPe single dehydrogenation Cavani et al 2009 score The Climate Change category includes process contributions Molar ratios were used to calculate the mass balance of input to both human health and the ecosystem damage categories Each and output flows Also in this case 999 unreacted propane is sent impact category is described in detail in the following paragraphs to combustion and avoided impacts from heat recovery are taken into account similar to previous scenario avoided extraction and 31 Climate change and fossil fuel depletion combustion of natural gas Table 2 As previously assumed for the process from propylene byproducts acetonitrile and hydrogen The impact on climate change and fossil fuel depletion cate cyanide were not considered in these models Mass allocation for gories is closely linked the use of fossil fuels to generate energy each scenario was obtained using selectivity values as reported in produces relevant carbon dioxide emissions and entails green Table 2 Energy consumption for the production of ammonium house effects In fact both categories show a growing trend moving sulfate and the avoided impacts derived from N fertilizer produc from the SOHIO process to propane ammoxidation scenarios tion were included in each model The remaining unreacted Table 2 The increase in impacts is mainly due to the amounts of ammonia about 20 is assumed to be oxidized to produce N2 and input and output substances of the models There are two major H20 causes first moving from the SOHIO process to the systems Equation 3 was used to estimate the energy exchanged in involving propane ammoxidation a decrease in yield occurs 81 reactor coils by assuming that 50 of the total is recovered and for propylene ammoxidation and around 60 for Asahi and Mit reused for the plant utilities eg steam and electricity subishi processes down to the lowest values for the BP processes Table 2 Inventory analysis for propane ammoxidation scenarios Processes Unit Scenarios 1 AKCC 2 MCC 3 BP Poor 4 BP Rich Feed molar ratio C3NHs3airinertH20 101230148 11515 1215573 51281 Catalyst in silica 50 wt M0310 Vo33 Nbo11 Teo22 Ox M010 Vo3 Nbo12 Teo23 Ox V Sbs Wos Teo5 SNos Ox V Sbi4 SNo2 Tio2 Ox Catalyst amount g 17 17 17 17 ACN selectivity 655 655 567 619 ACN yield 590 596 390 89 Propane conversion 900 910 688 145 Propane input kg 141 139 213 926 Propane burned 999 efficiency kg 014 012 066 791 Propane in air 001 kg 141E 04 125E04 665E04 792E03 Ammonia input kg 065 081 165 071 Neutralized ammonia 80 of unreacted kg 013 025 086 016 Oxidized ammonia 20 of unreacted kg 003 006 022 004 Sulfuric acid input kg 038 073 249 045 Ammonium sulfate kg 050 099 335 061 Heat recovered kj 579E 03 571E 03 6 OOE 03 3 18E 03 Electricity recovered kWh 050 049 052 027 Natural gas avoided m 018 016 086 1022 1 Asahi Kasei Chemical Corporation US PAT 6143916 2 Mitsubishi Chemical Corporation EU PAT 529853 3 BP Poor US PAT 4788317 4 BP Rich US PAT 5094989 D Cespi et al Journal of Cleaner Production 69 2014 1725 23 Table 3 06 Impact assessment results for each category considered Fossil fuel depletion Impact Categories Unit SOHIO AKCC MCC BP poor BP rich 05 m Metal depletion Climate change DALYs 305E06 369E06 412E06 672E06 767E06 m Climate change on human 04 health Climate change speciesyr 173E08 209E08 233E08 380E08 435E08 on ecosystems 03 Metal depletion 893E03 537E03 537E03 374E03 478E03 Fossil fuel 240E01 315E01 337E01 498E01 511E01 02 depletion 01 from 9 to 40 Consequently the amount of organics introduced into the reactor changes for instance in the case of the BPrich soulo axec mec BP poor BP rich scenario the amount of propane used is more than ten times higher than the propylene in the SOHIO process The large amount Fig 5 Five ammoxidation scenarios compared in terms of ReCiPe single score of reactants means that a higher quantity of organics has to be extracted to satisfy supply requirements The second reason is the higher amount of the ammonia necessary for running the process 32 Metal depletion which is determined on the basis of different molar ratios for each process According to literature EFMA 2000 the designed model As known the catalyst efficiency influences the product yield assumed that 85 of ammonia synthesis gases Nz and H2 are and so the feasibility of processes on industrial scale aspect produced by the reforming of natural gas while the rest is obtained considered in the study through the evaluation of reactants amount from the partial oxidation of heavy fuel oil at high temperature and Consumed per kg of ACN produced However also the impacts pressure conditions about 500 C and 300 atm respectively Appl associated with each catalyst system were evaluated As reported in 2011 All process stages are energyintensive and entail a high the inventory analysis no information about the production steps consumption of fossil fuels were available eg energy consumption on industrial scale during Nevertheless a significant contribution to the results is due to the manufacture due to the confidential information linked with the avoided impact thanks to material and energy recovery The the catalyst making corporate knowhow electrical and thermal energy recovered from the exothermal re Nonetheless to have a simplified evaluation of impacts of the actions results in avoiding the extraction and consumption of fossil catalyst manufacture we assessed the environmental load only in fuels which contributes to reducing the overall impact score terms of resources extraction required for each system Results significantly In particular a low percentage of propane conversion expressed in form of metal depletion impact category show the means that a great deal of unconverted reagent a light alkane is Worst performance for the SOHIO process scenario Reasons must burned this translates into a direct contribution to methane be sought in the use of two catalyst systems for this scenario the savings first one zeolite is needed for the cracking while the second is Furthermore the avoided impacts deriving from the produc used in the ammoxidation step tion of ammonium sulfate contribute to reducing the global Catalysts developed by Mitsubishi Chemicals Corporation and environmental load for the two midpoint categories As expected Asahi Kasei Chemical Corporation seem to have similar empirical the higher amount of ammonium sulfate is produced by the BP composition identified by means of a combinatorial methodology scenario with propane lean conditions and the consequent excess Table 2 Molybdate systems are prepared preferably by hydro of ammonia in feeding Table 2 As already mentioned unreacted thermal synthesis which results in nucleation and the growth of ammonia is neutralized by sulfuric acid to produce salt the two phases M1 which is able to independently convert propane greater the amount of unreacted base the greater the quantity of into acrylonitrile and the cocatalyst M2 which is necessary for sulfate produced and consequently the greater the avoided imcreasing selectivity the promotion of intermediately formed impact propylene ammoxidation into acrylonitrile Cavani et al 2009 These M1 and M2 compositions are reported in literature Mo7g V12 Nb Teop94 O2g9 and Mo47 V133 Te182 01982 respectively Cavani et al 2009 Climate change 1 oo As already mentioned Fig 4 shows the results of the charac Metal depletion terization analysis in the form of a radar chart at each vertex of the 8 pentagon a scenario is reported and colored lines represent the Fossil fuel depletion percentage ratio for every impact category The closer the lines are Za to vertices 100 the higher the impact of the considered scenario BP rich Sw AKCC The red line represents the climate change category which includes 0 process contributions to both the human health and the ecosystem P damage categories the black line indicates the fossil fuel depletion TA and lastly the grey line shows metal consumption scores The results obtained through the comparison of the five ammoxidation scenarios were also expressed in terms of a ReCiPe single score Fig 5 Histograms show the overall results for the five a scenarios which were obtained from the cumulative sum of each BP poor McC impact category after conversion to a single point The comparison shows how the alternative synthetic routes starting from propane Fig 4 Radar chart showing results in terms of single points by percentage have total impacts higher than those of the conventional SOHIO process The cumulative results give overall measures of the envi ronmental performance for scenarios and permit weighting the relevance of each impact category in the total load on the envi ronment As shown the fossil fuel depletion and climate change categories make the highest ReCiPe Pt contribution to the total impact while the lowest contribution comes from metal depletion Although this might make it seem environmental implications from that midpoint category are negligible a further discussion is necessary Indeed the importance of taking into account the use of metals in catalytic systems is confirmed by literature Buchert et al 2009 as the situation for most metals eg platinum palladium and rhodium is critical because of supply risk environmental implications and vulnerability to supply restrictions Graedel et al 2012 and therefore their effects on the environment might be not negligible in the future 33 Sensitivity analysis LCA studies are commonly influenced by data quality used in the inventory analysis As reported SOHIO process scenario was created using data reported in literature instead no information for alternative route starting from propane are available except in the patents Also some information about several aspects are not available for all scenarios as for example detailed input and output com positions catalyst makeup and regeneration energy consumed during the several stages etc These information often represent corporate knowhow so they are confidential and it is not possible obtain them from literature So a sensitivity analysis was performed to evaluate the robust ness of the model created by focusing on the two scenarios fully operating at an industrial scale the SOHIO process and the AKCC process Uncertainty ranges for data used in the inventory analysis were determined by using the data pedigree matrix developed by Weidema and Wesnaes 1996 as reported in previously study Cespi et al 2013 The values obtained were used to perform a Monte Carlo analysis which is a statistical method for evaluating the models sensitivity Lognormal statistical distribution with a 95 confidence interval and an iterative calculation number of 1000 simulations was applied The results of this comparison are reported in Fig 6 The three impact categories considered in this study are shown on the yaxis while the xaxis shows the percentage values ach ieved by the scenarios at the end of iterative simulations Green bars show the number of times the AKCC process has higher impact than the SOHIO process conversely the blue bars represent the opposite situation As shown the Monte Carlo method confirms the reliability of the results obtained in the characterization analysis the AKCC process proves to have higher impacts with regard to fossil fuel consumption and climate change whereas the SOHIO process has a poorer performance in the metal depletion category A marked preference between the two scenarios at a 95 con fidence level cannot be clearly identified the scores show that the poorer results achieved by the AKCC process are frequently confirmed in the range of 60e66 climate change and fossil fuel depletion respectively Conversely in the case of the metal depletion category the SOHIO process seems to have a negligibly higher impact about 48 for the Asahi process and 52 for the SOHIO process These scores are due mainly to the close results in the characterization analysis 4 Conclusions The study presents a scientific approach through which inves tigate the environmental footprint of the chemical production sector In particular the industrial production of acrylonitrile was studied considering the main stage of the manufacturing process mass and energy inputoutput from reactor heat exchanger mass and energy recovery and the catalyst making The LCA methodology was applied as scientific tool to compare the traditional synthesis of acrylonitrile by propylene ammox idation SOHIO process and the less expensive alternative routes that use propane as the precursor also in terms of production steps in fact propane production is performed with a onestep process the distillation of petroleum whereas propylene production in volves two steps distillation and cracking steam or catalytic This savings in production steps may suggest that the alternative processes would be allegedly greener than traditional methods thus leading to a reduction in the environmental load However in order to define the most sustainable and cleaner route it is necessary to have a complete view of the entire process while evaluating the main stages and flows involved LCA answers well the need for quantitatively assessing the environmental sustain ability of an industrial process in a life cycle perspective According to Curran 2013 LCA is a fundamental supporting tool for the chemical industry as it makes it possible to extend the assessment of environmental implications from the production process to the entire life cycle thus avoiding the deduction of partial or limited evaluations The model created was analyzed using ReCiPe 2008 midpoint oriented method able to assess environmental loads of each sce nario in terms of climate change fossil fuel depletion and metal depletion categories As shown by characterization analysis alternative processes starting from propane generally seem to have higher impacts especially in terms of fossil fuel depletion and climate change categories Although not reported in the study the comparison between the five scenarios was also done using Ecoindicator 99 as analysis method Scores confirmed results obtained by ReCiPe 2008 method underlining higher impacts of the alternative routes starting from propane in particular in terms of fossil fuel depletion This outcome is mainly due to the lower activity of the commer cially developed catalyst systems entailing both larger amounts of reactants and a heavier load on the ecosystem thus resulting in the lower sustainability of alternative processes Therefore a very crucial role is played mainly by the commercially developed cata lyst system which can modify the yield of the entire process thus influencing selectivity and conversion As we said in the inventory analysis a model for different catalyst systems was created e including only the extraction phase of metals e starting from in formation about makeup and composition reported in patents Nevertheless to make this model complete more details would be needed regarding catalyst production steps for example energy consumption and emissions and the regeneration of catalysts but Fig 6 Monte Carlo Analysis in terms of midpoint impact categories D Cespi et al Journal of Cleaner Production 69 2014 17e25 24 D Cespi et al Journal of Cleaner Production 69 2014 1725 25 this kind of information is often confidential as it is part of corpo Friedrich MS Seely MJ Suresh DD European Patent 0 437 056 A2 1991 Oli Assigned to the Standard Oil Company OH USA rate know how to which there Is limited access Garmston S 2009 Acrylonitrile and Derivatives World SupplyDemand Report As we Said LCA studies are influenced by data quality In this 2009 PCI Acrylonitrile Ltd wwwpciacrylocom accessed 090114 study was no possible to obtain primary data for each scenario so Goedkoop M Heijungs Huijbregts M De Schryver A Struijs J van Zelm R to evaluate the robustness of model created a sensitivity analysis 2012 ReCiPe 2008 a Life Cycle Impact Assessment Method Which Comprises 1 hod fe d Harmonised Category Indicators at the Midpoint and the Endpoint Level first using Monte Carlo method was per orme ed revised Ministry of Housing Spatial Planning and the Environment In order to enhance the potentiality of LCA applied to the VROM Netherlands chemical industry it may be appropriate to discuss how to establish cre i i Cee Ne has sok pe yantstoffersen we se riedlander E Henly C Jun C Nassar NT Schechner D Warren S Yang M closer relationships between companies and research institutions Zhu C 2012 Methodology of metal criticality determination Environ Sci such as for instance the development of an accredited database as Technol 46 10631070 well as a standardized approach to ensure a reliable inventory Grassell BiG 2002 Fundamental principles of selective heterogeneous oxidation catalysis Top Catal 21 7988 analysis and unambiguous assessment of environmental impacts Grasselli RK 2005 Selectivity issues in ammoxidation catalysis Catal Today 99 2331 References Griffiths OG OByrne JP TorrenteMurciano L Jones MD Mattia D McManus MC 2013 Identifying the largest environmental life cycle impacts Al Salem SM Mechleri E Papageorgiou LG Lettieri P 2012 Life cycle assess during carbon nanotube synthesis via chemical vapour deposition J Clean ment and optimization on the production of petrochemicals and energy from Prod 42 180189 polymers for the Greater London Area Comput Aided Chem Eng 30 101106 Herrchen M Klein W 2000 Use of lifecycle assessment LCA toolbox for an envi Anastas PT Lankey RL 2000 Life cycle assessment and green chemistry the yin ronmental evaluation of production processes Pure Appl Chem 72 12471252 and yang of industrial ecology Royal Soc Chem 2 289295 Holman PA Shonnard DR Holles JH 2009 Using life cycle assessment to guide Appl M 2011 Ammonia 3 Production Plants In Ullmanns Encyclopedia of In catalysis research Ind Eng Chem Res 48 66686674 dustrial Chemistry Published Online on wwwonlinelibrarywileycomdoi10 INEOS Wa wwwineostechnologiescom78redirectionhtm accessed 090114 een el weastola ML 1098 LOA ped tothe coecee ent of the environmental IPPC Integrated Pollution Prevention and Control 2003 Reference Document on impact of alternative synthetic processes The dimethylcarbonate case part 1 Best Available Techniques in the large Volume Organic Chemical Industry J Clean Prod 7 181193 Published Online on wwwepaiedownloadsadvicebrefsname 14517enhtml Askham C Gade AL Hanssen OJ 2013 Linking chemical risk information with accessed 090114 life cycle assessment in product development J Clean Prod 51 196204 Jacquemin L Pontalier P Sablayrolles Cc 2012 Life cycle assessment LCA applied Brazdil JF 2006 Strategies for the selective catalytic oxidation of alkanes Top 10 the process industry a review Int J Life Cycle Assess 7 10281041 Catal 38 289294 JimenezGonzalez C Constable DJC 2011 Green Chemistry and Engineering a Brazdil JF 2010 Acrylonitrile In KirkOthmer Ecyclopedia of Chemical Technol Practical Design Approach John Wiley Sons Inc Hoboken New Jersey USA ogy vol 1John Wiley Sons Ltd pp 397414 Published Online on www Jodicke G Zenklusen O Weidenhaup tA Hungerbithler K 1999 Developing onlinelibrarywileycomdoi10100204712389610103182502180126a01pub3 environmentallysound processes in the chemical industry a case study on abstract accessed 090114 pharmaceutical intermediates J Clean Prod 7 159166 Brazdil JF 2012 Acrylonitrile In Ullmanns Encyclopedia of Industrial Chemistry Kawaucht W Re aoa nds Mi 2002 A new approach fo production regular 3 388 Published Online on wwwonlinelibrarywileycomdoi10100214356007a01 i u Reliab eng yst 177pub3abstract accessed 090114 Kowalski Z Kulczycka J Wzorek Z 2007 Life cycle assessment of different var Buchert M Schuler D Bleher D OkoInstitut eV Freiburg Germany 2009 L ae oe tum chrom ve production A roland J Clean Pro a G Sustainable Innovation and Technology Transfer Industrial Sector Studies a S013 Life ne 7 hal sie f oul nib oT lace h ithenn 0 Ica oe Critical Metals for Future Sustainable Technologies and their Recycling Potential Clea Da 6076 of a novel hybrid glasshempthermoset composite United Nations Environment Programme UNEP United Nations University J add a ae rn Langvardt PW 2011 Acrylonitrile In Ullmanns Ecyclopedia of Industrial Chem Cavani F 2010a Catalytic selective oxidation faces the sustainability of challenge istry vol John Wiley Sons Ltd 365372 Published Online on www turning points objectives reached old approaches revisited and solutions still ie Libr il y doi 101002 oe 356007 01 177 pub2abstract requiring further investigations J Chem Technol Biotechnol 85 11751183 onunell TOOL 0i10 a01177pub2abstrac Cavani F 2010b Catalytic selective oxidation the forefront in the challenge for a accesse wee more sustainable chemical industry Catal Today 157 815 MoralesMora MA RosaDominguez E SuppenReynaga N Martinez Cavani F Teles JH 2009 Sustainability in catalytic oxidation an alternative Delgadil lo SA 2012 Piel and ecocosts life cycle assessment of an approach or a structural evolution Chem Sus Chem 2 508534 acrylonitrile process by capacity enlargement in Mexico Process Saf Environ Cavani F Centi G Marion P 2009 Catalytic ammoxidation of hydrocarbons on 90 2737 oe mixed oxides In Jackson SD Hargreaves JSJ Eds Metal Oxide Catalysis Nemecek T Kagi T 2007 Life Cycle Inventor tes of Agric ultural Production Sys WileyVCH Verlag GmbH Co KGaA Weinheim Germany httpdxdoiorg10 tems Ecoinvent Database report No 15 Zitrich and Dilbendorf CH er Pavone A Schwaar RH 1989 PEP Review No 8824 Acrylonitrile from Propane 10029783527626113ch20 Published Online on wwwonlinelibrarywileycom doi1010029783527626113ch20summary accessed 090114 by New BPSOHIO Patented Process SRI INTERNATIONAL Menlo Park Califor Cespi D Passarini F Ciacci L Vassura I Castellani V Collina E Piazzalunga A nia Published also Online on wwwihscomproductschemicaltechnology Morselli L 2013 Heating systems LCA comparison of biomassbased appli 2 pepreviews acrylonitrilefrompropaneaspx accessed 090114 PérezLopez P GonzalezGarcia S Jeffryes C Agathos SN McHugh E Walsh D ances Int J Life Cycle Assess httpdxdoiorg101007s1136701306113 in i ress accessed 090114 Murray P Moane S Feijoo G Moreira MT 2014 Life cycle assessment of the Chavwel A Lefebvre G 1989 Petrochemical Processes In Major Oxygenated production of the red antioxidant carotenoid astaxanthin by microalgae from Chlorinated and Nitrated Derivatives vol 2 Institut francais du pétrole publi lab to pilot scale J Clean Prod 64 332334 cations Edition Technip France ISBN 9782710805618 Plastics Europe 2005 Ecoprofiles of the European Plastics Industry Propylene Chemsystems wwwchemsystemscomaboutcsnewsitemsPERP09102 Published Online on wwwplasticseuropeorgplasticssustainabilityeco Acrylonitrilecfm 2710805618 accessed 090114 mort ones browse Dy listaspx accessed eran so10a Life evel Curran MA 2013 Life Cycle Assessment a review of the methodology and its ortha JF Jaubert JN Louret S Pons MN 2010a Life cycle assessmen application to sustainability Curr Options Chem Eng 2 3 273277 applied to naphtha catalytic reforming Oil Gas Sci Technol Rev IFP Energies Demou E Hellweg S Hungerbiihler K 2011 An occupational chemical priority nouvelles 65 5 793805 or list for future life cycle assessments J Clean Prod 19 13391346 Portha JF Louret S Pons MN Jaubert JN 2010b Estimation of the envi Doménech X Ayllon JA Peral J Rieradevall J 2002 How green is a chemical ronmen tal impact of a petrochemical process using coupled LCA and exergy reaction Application of LCA to Green Chemistry Environ Sci Technol 36 analysis Resour Conserv Recyl 54 291298 55175520 Raseev SD 2003 Thermal and Catalytic Processes in Petroleum Refining Cap 7 In Ecoinvent Centre formerly Swiss Centre for Life Cycle Inventories 2009 Ecoinvent dustrial Catalytic Cracking Marcel Dekker Inc New York United States of America 22 Database Sustainable Chemistry Sus Chem 2013 wwwsuschemorg accessed 090114 European Fertilizer Manufacturers Association EFMA 2000 Production of The Dow Chemical Company Dow wwwdowcominvestorspdfspresentations Ammonia Best Available Tecniques for Pollution Prevention and Control in the USGC101211pdffaccessed 090114 European Fertilizer Industries Booklet N1 Brussels Belgium Weidema BP Wesnaes M 1996 Data quality management for life cycle in Figueiredo de MCB Freitas Rosa de M Ugaya CMLL Souza Filho de 4 Moreira ventories an example for using data quality indicators J Clean Prod 4 3A4 de M Silva Braid da ACC Melo de LFL 2013 Life cycle assessment of 167174 cellulose nanowhiskers J Clean Prod 35 130139