Volume 1 photovoltaic solar energy 1 08 – environmental impacts of photovoltaic life cycles

Volume 1 photovoltaic solar energy 1 08 – environmental impacts of photovoltaic life cycles Volume 1 photovoltaic solar energy 1 08 – environmental impacts of photovoltaic life cycles Volume 1 photovoltaic solar energy 1 08 – environmental impacts of photovoltaic life cycles Volume 1 photovoltaic solar energy 1 08 – environmental impacts of photovoltaic life cycles Volume 1 photovoltaic solar energy 1 08 – environmental impacts of photovoltaic life cycles Volume 1 photovoltaic solar energy 1 08 – environmental impacts of photovoltaic life cycles Volume 1 photovoltaic solar energy 1 08 – environmental impacts of photovoltaic life cycles

1.08 Environmental Impacts of Photovoltaic Life Cycles VM Fthenakis, Columbia University, New York, NY, USA; Brookhaven National Laboratory, Upton, NY, USA HC Kim, Brookhaven National Laboratory, Upton, NY, USA © 2012 Elsevier Ltd All rights reserved 1.08.1 Introduction 1.08.2 Background 1.08.3 Life Cycle of Photovoltaics 1.08.4 Life-Cycle Inventory 1.08.4.1 Modules 1.08.4.2 Balance of System 1.08.5 Energy Payback Times and Greenhouse Gas Emissions 1.08.5.1 Energy Payback Time 1.08.5.2 Greenhouse Gas Emissions 1.08.6 Criteria Pollutant and Heavy Metal Emissions 1.08.6.1 Criteria Pollutant Emissions 1.08.6.2 Heavy Metal Emissions 1.08.6.2.1 Direct emissions 1.08.6.2.2 Indirect emissions 1.08.7 Life-Cycle Risk Analysis 1.08.7.1 Risk Classification 1.08.7.2 Risks of Accidents in the Photovoltaic Life Cycle 1.08.7.3 Comparison with Other Energy Technologies 1.08.7.4 Limitation of the Study 1.08.8 Conclusion Acknowledgment References Nomenclature a-Si amorphous silicon AC alternate current BOS balance of system c-Si crystalline silicon CdS cadmium sulfide CdTe cadmium telluride CIGS copper indium gallium selenide CIS copper indium diselenide DC direct current EPBT energy payback time ESP electrostatic precipitator FBR fluidized bed reactor GaAs gallium arsenide GHG greenhouse gas GWP global warming potential HCl hydrogen chloride HF hydrogen fluoride LCA life-cycle analysis (or assessment) LCI life-cycle inventory LPG liquefied petroleum gas Comprehensive Renewable Energy, Volume 144 144 144 146 146 147 147 147 148 150 150 151 151 151 153 153 153 155 157 157 158 158 multi-Si multicrystalline silicon mono-Si monocrystalline silicon NAICS North American Industry Classification System NG natural gas NOx nitrogen oxide PM particulate matter PR performance ratio PSA probabilistic safety assessment PSI Paul Scherrer Institute PV photovoltaic RMP risk management program SiH4 silane SiHCl3 trichlorosilane SOx sulfur oxide TeO2 tellurium dioxide TPE thermoplastic elastomer UCTE Union for the Co-ordination of Transmission of Electricity VTD vapor transport deposition doi:10.1016/B978-0-08-087872-0.00107-4 143 144 Economics and Environment 1.08.1 Introduction Currently, the main commercial photovoltaic (PV) materials are multicrystalline silicon (multi-Si), monocrystalline silicon (mono-Si), amorphous silicon (a-Si), and cadmium telluride (CdTe) A typical PV system consists of the PV module and the balance of system (BOS) structures for mounting the PV modules and power-conditioning equipment that converts the generated electricity to alternate current (AC) electricity of proper magnitude for usage in the power grid Life-cycle analysis (LCA) is a framework for considering the environmental inputs and outputs of a product or process from cradle to grave It is employed to evaluate the environmental impacts of energy technologies, and the results are increasingly used in decisions about R&D funding and in formulating energy policies In this chapter, we summarize the results of PV LCAs based on current data for two silicon and one thin-film technologies, emphasizing basic metrics including energy payback times (EPBTs), greenhouse gas (GHG) emissions, criteria pollutant emissions, toxic metal emissions, human injuries, and fatalities 1.08.2 Background Early life-cycle studies report a wide range of primary energy consumption for Si PV modules [1] On the basis of normalized energy consumption per square meter, the researchers reported 2400–7600 MJ of primary energy consumption for multi-Si and 5300–16 500 MJ for mono-Si modules Besides uncertainties in the data, these differences are due to different assumptions and allocation rules that each author adopted for modeling the purification and crystallization stages of silicon [1, 2] Selecting only those process steps needed to produce solar-grade silicon, Alsema’s [1] own estimates were 4200 and 5700 MJ m−2 for multi- and mono-Si modules, respectively These values correspond to an EPBT of 2.5 and 3.1 years and life-cycle GHG emissions of 46 and 63 g CO2-eq kWh−1 for rooftop-mounted multi-Si PV with 13.2% efficiency and mono-Si with 14% efficiency, respectively, under Southern European (Mediterranean) conditions: insolation of 1700 kWh m−2 yr−1 and a performance ratio (PR; the ratio between the ideal and the actual electricity output) of 0.75 The BOS components, such as a mounting support, a frame, and electrical components, account for additional ∼0.7 years of EPBT and ∼15 g CO2-eq kWh−1 of GHG emissions Meijer et al [3] more recently assessed a slightly higher energy expenditure of 4900 MJ m−2 to produce a multi-Si module They assumed that the 270 µm thick PV cells with 14.5% cell efficiency were fabricated from electronic-grade high-purity silicon, which entails greater energy consumption Their corresponding EPBT estimate for the module was 3.5 years excluding BOS components, that is, higher than Alsema’s earlier determination of 2.5 years The increase stems mainly from the low level of insolation in the Netherlands (1000 kWh m−2 yr−1) compared with the average for Southern Europe (1700 kWh m−2 yr−1), and, to a lesser degree, from the higher energy estimation for silicon [1, 3] Jungbluth [4] reported the life-cycle metrics of various PV systems under environmental conditions in Switzerland in 2000 He considered the environmental impacts for 300 µm thick multi- and mono-Si PV modules with 13.2% and 14.8% conversion efficiency, respectively Depending on which of the two materials he evaluated, and their applications (i.e., faỗade, slanted roof, and flat roof), he arrived at figures of 39–110 g CO2-eq kWh−1 of GHG emissions and 3–6 years of EPBT for the average insolation of 1100 kWh m−2 yr−1 in that country He assumed that the source of silicon materials was 50% from off-grade silicon and 50% from electronic-grade silicon, which is distant from the composition of the current (2010s) PV supply [2, 4] There are fewer life-cycle studies of thin-film PV technologies; evaluations of the life-cycle primary energy consumption of a-Si ranged between 710 and 1980 MJ m−2 [1] The differences are largely attributed to the choice of substrate and encapsulation materials The lowest estimate, made by Palz and Zibetta, considered a single glass structure, while the highest one by Hagedorn and Hellriegel was based on a double-glass configuration to protect the active layer [1, 5, 6] For CdTe PV, Hynes et al [7] based their energy analysis on two alternative technologies employed at that time The first employed nonvacuum electrodeposition of a 1.5 µm absorber layer (CdTe), in conjunction with chemical-bath deposition of the 0.2 µm window layer (cadmium sulfide (CdS)); the second method deposited both these layers by thermal evaporation yielding an ∼5 µm thick absorber layer and an ∼1.7 µm thick window layer Their primary energy estimate for the first technology was 993 MJ m−2 and that for the second was 1188 MJ m−2 Kato et al.’s [8] energy estimates were pertinent to the scale of annual production; they suggested that energy consumption will decline as the scale of production rises; they cited values of 1523, 1234, and 992 MJ m−2 for frameless modules with annual capacities of 10, 30, and 100 MWp (peak power), respectively However, these earlier estimates fall far short of describing present-day commercial-scale CdTe PV production, which, unlike previously, now encompasses many large-scale production plants 1.08.3 Life Cycle of Photovoltaics The life cycle of PVs starts from the extraction of materials from the ground (cradle) and ends in the disposal or recycling of the endof-life products (grave) The main stages are (1) the production of raw materials, (2) their processing and purification, (3) the manufacture of modules and BOS components, (4) the installation and use of the systems, and (5) their decommissioning and disposal or recycling (Figure 1) Environmental Impacts of Photovoltaic Life Cycles M, Q M, Q Raw material acquisition M, Q Material processing M, Q Manufacturing Use M, Q M, Q Decommissioning Treatment /disposal 145 M, Q E E E E E E Recycling M, Q: material and energy inputs E: effluents (air, water, solids) E Figure Flow of the life-cycle stages, energy, materials, and effluents for PV systems (a) (b) Quartz mining Metallurgical-grade Si Zn ores Cu ores Cd Te Solargrade Si Mono-Si crystal Multi-Si ingot Ribbon Mono-Si wafer Multi-Si wafer CdTe powder CdS powder Thin-film CdS/CdT Cell Encapsulation Frame Module Module BOS BOS PV system PV system Figure Detailed flow diagram from raw material acquisition to the manufacturing stage of PVs [9] (a) Silicon PVs and (b) CdTe PVs (frameless) BOS, balance of system; mono-Si, monocrystalline silicon; multi-Si, multicrystalline silicon Production starts with mining of the raw materials (i.e., quartz sand for silicon PV; Zn and Cu ores for CdTe PV), and continues with their processing and purification (Figure 2) [9] The silica in the quartz sand is reduced in an arc furnace to metallurgical-grade silicon, which must be purified further into ‘electronic-grade’ or ‘solar–grade’ silicon, typically through a ‘Siemens’ process Crystalline silicon (c-Si) modules typically are framed for additional strength and easy mounting The recent LCAs of c-Si are based on life-cycle inventory (LCI) data provided, collectively, by 11 European and US PV companies participating in the European Commission’s CrystalClear project The data sets were published in separate papers by Alsema and de Wild-Scholten [2] and by Fthenakis and Alsema [10] The LCIs of the minor metals used in thin-film PVs such as Cd, In, Mo, and Se are closely related to the production cycle of base metals (Zn, Cu) The allocations of emissions and energy use between the former (Cd, In, Mo, and Se) and the latter (Zn, Cu) during mining, smelting, and refining stages are described elsewhere [11] Fthenakis [12] described the material flows of cadmium (Cd) and emissions from the entire life-cycle stages of CdTe PV The life cycle starts with the production of Cd and Te, which are by-products, respectively, of smelting of Zn and Cu ores (Figure 2) Cadmium is obtained from the Zn waste streams, such as particulates collected in air pollution control equipment and slimes collected from Zn electrolyte 146 Economics and Environment purification stages Cadmium is further processed and purified to meet the four or five 9s purity required for synthesizing CdTe Tellurium is recovered and extracted after treating the slimes produced during electrolytic copper refining with dilute sulfuric acid; these slimes also contain Cu and other metals After cementation with copper, CuTe is leached with caustic soda to produce a sodium telluride solution, which is used as the feed for Te and TeO2 (tellurium dioxide) Additional leaching and vacuum distillation gives Cd and Te powders of semiconductor grade (i.e., 99.999%) The LCI data on thin-film CdTe PV were provided by First Solar, the largest manufacturer of CdTe PV modules, using vapor transport deposition (VTD) to deposit the CdTe layer 1.08.4 Life-Cycle Inventory Life-cycle assessments require data on material and energy use as well as emissions during the various stages of the life cycle of PVs These data, called LCIs, are typically available from different databases for the modules and the BOS 1.08.4.1 Modules The material and energy inputs and outputs during the life cycles of Si PVs, namely, multi- and mono-Si, and also thin-film CdTe PV were investigated in detail based on actual measurements from PV production plants Alsema and de Wild-Scholten recently updated the LCI for the technology for producing c-Si modules in Western Europe under the framework of the CrystalClear project, a large European integrated project focusing on c-Si technology, cofunded by the European Commission and the participating countries [2, 13] Fthenakis and Kim [14] reported the LCI data for CdTe thin-film technology taken from the production data from First Solar’s plant in Perrysburg, OH, USA Table presents the simplified LCIs for 2006, compiled from the data from 11 European and US plants along with values in the literature [2, 14] The typical thickness of multi- and mono-Si PVs is 200 and 180 µm, respectively; 60 individual cells of 243 cm2 (156 mm  156 mm) comprise a module of 1.6 m2 for all Si PV types The conversion efficiency of multi- and mono-Si modules is taken as 13.2%, and 14.0%, respectively On the other hand, as of 2009, the frameless, double-glass, CdTe modules of 1.2 m  0.6 m, manufactured by First Solar, are rated at 10.9% photon-to-electricity conversion efficiency with ∼3 µm thick active layer The data for Si PVs extend from the production stage of solar-grade Si to the manufacturing stage of the module, while those for CdTe PV correspond to the deposition of the CdTe film and the manufacturing stage of the module The metallurgical-grade silicon that is extracted from quartz is purified into solar-grade polysilicon by either a silane (SiH4)- or a trichlorosilane (SiHCl3)-based process The energy requirement for this purification step is significant for c-Si PV modules, accounting for ∼30% of the primary energy used for fabricating multi-Si modules [15] Two technologies are currently employed for producing polysilicon from silicon gases: the Siemens reactor method and the fluidized bed reactor (FBR) method In the former, which accounts for the majority (∼90% in 2004) of solar-grade silicon production in the United States, silane or trichlorosilane gas is introduced into a thermal decomposition furnace (reactor) with high-temperature (∼1100–1200 °C) polysilicon rods [16–18] The silicon rods grow as silicon atoms in the gas deposit onto them, up to 150 mm in diameter and up to 150 cm in length [16] The data on Si PVs in Table are based on averages over standard and modified Siemens reactors The former primarily produces the electronic-grade silicon with a purity of over nine 9s, while the latter produces the solar-grade silicon with a purity of six to eight 9s, consuming less energy than the former [19] The scenario involving the scrap silicon from electronic-grade silicon production is not considered as the market share of this material accounts for only 5% in 2005 [20] Table Materials and energy inputs for PV systems to produce m2 of module including process loss, compiled in 2006 (excluding the frame for Si modules) [9] Category Inputs Multi-Si Mono-Si CdTe Components (kg) Cell materials Glass Ethylene vinyl acetate Others Gases Liquid Others Electricity (kWh) Oil (l) Natural gas (MJ) 1.6 9.1 1.0 1.8 2.2 6.8 4.3 248 0.05 308 1.5 9.1 1.0 1.8 7.8 6.6 4.3 282 0.05 361 0.065 19.2 0.6 2.0 0.001 0.67 0.4 59 0.05 – Consumables (kg) Energy Environmental Impacts of Photovoltaic Life Cycles 1.08.4.2 147 Balance of System Little attention has been paid to the LCA studies of the BOS, and so inventory data are scarce Depending on the application, solar cells are either rooftop- or ground-mounted, both operating with a proper BOS Silicon modules need an aluminum frame of 2.6 kg m−2 for structural robustness and easy installation, while a glass backing performs the same functions for the CdTe PV produced in the United States [2, 14] For a rooftop PV application, the BOS typically includes inverters, mounting structures, cable, and connectors Large-scale ground-mounted PV installations require additional equipment and facilities, such as grid connections, office facilities, and concrete A recent analysis of a 3.5 MWp mc-Si installation at the Springerville Generating Station in Arizona affords a detailed material and energy balance for a ground-mounted BOS (Table 2) [21] For this study, Tucson Electric Power (TEP) prepared the BOS bill of material and energy consumption data for its mc-Si PV installations The life expectancy of the PV metal support structures is assumed to be 60 years Inverters and transformers are considered to last for 30 years, but parts must be replaced every 10 years, amounting to 10% of their total mass, according to well-established data from the power industry on transformers and electronic components The inverters are utility-scale Xantrec PV-150 models with a wide-open frame, allowing failed parts to be easily replaced The LCI includes the office facility’s material and energy use for administrative, maintenance, and security staff, as well as the operation of maintenance vehicles Aluminum frames are shown separately, since they are part of the module, not of the BOS inventory; there are both framed and frameless modules on the market De Wild-Scholten et al [22] studied two classes of rooftop mounting systems based on an mc-Si PV system called SolarWorld SW220 with dimensions of 1001 mm  1675 mm and 220 Wp: they are used for on-roof mounting where the system builds on existing roofing material and in-roof mounting where the modules replace the roof tiles The latter case is credited in terms of energy and material use because roof tile materials then are not required Table details the LCI of several rooftop mounting systems, cabling, and inverters Two types (500 and 2500 W) of small inverters adequate for rooftop PV design were inventoried A transformer is included as an electronic component for both models The amount of control electronics will become less significant for inverters with higher capacity (>10 kW), resulting in less material use per PV capacity [22] 1.08.5 Energy Payback Times and Greenhouse Gas Emissions The most frequently measured life-cycle metrics for the environmental analysis of PV systems are the EPBT and the GHG emissions 1.08.5.1 Energy Payback Time EPBT is defined as the period required for a renewable energy system to generate the same amount of energy (either primary or kWh equivalent) that was used to produce the system itself EPBT ẳ Emat ỵEmanuf ỵEtrans ỵEinst þEEOL Eagen −Eaoper where Emat is the primary energy demand to produce materials that constitute PV system, Emanuf the primary energy demand to manufacture PV system, Etrans the primary energy demand to transport materials used during the life cycle, Einst the primary energy Table Mass balance of major components for the 3.5 MW Tucson Electric Power Generating plant in Springerville, AZ, based on 30 years of operation [21] BOS BOS PV support structure Module interconnections Junction boxes Conduits and fittings Wire and grounding devices Inverters and transformers Grid connections Office facilities Concrete Miscellaneous Total Framea a Based on 12.2% rated efficiency for mc-Si module BOS, balance of system Mass (kg MWp−1) Percentage of total 16 821 453 385 561 648 10.3 0.3 0.8 4.0 3.4 28 320 726 20 697 76 417 806 163 834 18 141 17.3 1.1 12.6 46.6 3.5 100.0 148 Economics and Environment Table Life-cycle inventory of balance of system [21,22] (a) Mounting system (kg m−2) On-roof Low-alloy steel Stainless steel Aluminum Concrete Frame In-roof Phönix, TectoSun Schletter Eco05+EcoG Schletter, Plandach Schweizer, Solrif 0.49 0.54 3.04 0.72 0.97 0 0.28 1.21 0 0.08 1.71 0 (b) Cabling (g m−2) Copper Thermoplastic elastomer (TPE) PVC Helukabel, Solarflex 101, mm2, DC Helukabel, NYM-J, mm2, AC 83.0 64.0 0.0 19.9 0.0 16.9 Philips PSI 500 (500 W) Mastervolt SunMaster 2500 (2500 W) 78 682 68 148 9800 1400 (c) Inverters (g) Steel Aluminum Copper Polycarbonate Acrylonitrile butadiene styrene (ABS) Other plastics Printed circuit board Connector Transformers, wire-wound Coils Transistor diode Capacitor, film Capacitor, electrolytic Other electric components 5.4 100 50 310 74 10 72 54 20 1800a 5500 a Including electric components AC, alternate current; DC, direct current demand to install the system, EEOL the primary energy demand for end-of-life management, Eagen the annual electricity generation in primary energy term, and Eaoper the annual energy demand for operation and maintenance in primary energy term Calculating the primary energy equivalent requires knowledge of the country-specific, energy conversion parameters for fuels and technologies used to generate energy and feedstock The annual electricity generation (Eagen) is represented as primary energy based on the efficiency of electricity conversion at the demand side The electricity is converted to the primary energy term by the average conversion efficiency of 0.29 for the United States and 0.31 for Western Europe [23, 24] 1.08.5.2 Greenhouse Gas Emissions The GHG emissions during the life-cycle stages of a PV system are estimated as an equivalent of CO2 using an integrated time horizon of 100 years; the major emissions included as GHG emissions are CO2 (GWP = 1; GWP stands for global warming potential and is an indicator of the relative radiative effect of a substance compared to CO2, integrated over a chosen time horizon), CH4 (GWP = 25), N2O (GWP = 298), and chlorofluorocarbons (GWP = 4750–14 400) [25] Electricity and fuel use during the PV material and module production are the main sources of the GHG emissions for PV cycles Upstream electricity generation methods also play an important role in determining the total GHG emissions For instance, the GHG emission factor of the average US electricity grid is 40% higher than that of the average Western European (UCTE (Union for the Co-ordination of Transmission of Electricity)) grid although emission factors of fossil fuel combustion are similar, resulting in higher GHG estimates for the US-produced modules [23, 24] Environmental Impacts of Photovoltaic Life Cycles 149 With material inventory data from industry, Alsema and de Wild-Scholten [2] demonstrated that the life-cycle primary energy and GHG emission of complete rooftop Si PV systems are much lower than those reported in earlier studies Primary energy consumption is 3700 and 4200 MJ m−2, respectively, for multi- and mono-Si modules Fthenakis and Alsema also report that the GHG emissions of multi- and mono-Si modules corresponding to 2004–05 production are within 37 and 45 g CO2-eq kWh−1, with an EPBT of 2.2 and 2.7 years for a rooftop application under Southern European insolation of 1700 kWh m−2 yr−1 and a PR of 0.75 (Figures and 4, respectively) [2, 10] We note that in these estimates, the BOS for rooftop application accounts for 4.5–5 g CO2-eq kWh−1 of GHG emissions and 0.3 years of EPBT De Wild-Scholten recently updated these estimates based on thinner modules and more efficient processes, reporting an EPBT of ∼1.8 years and GHG emissions of ∼30 g CO2-eq kWh−1 for both multi- and mono-Si PVs Note that these figures include the effect of ‘take back and recycling’ of PV modules but not take into account the frame that is typically required for structural integrity in single glass modules After accommodating these factors, the GHG emissions correspond to 28 and 29 g CO2-eq kWh−1, respectively, for multi- and mono-Si PVs, whereas the EPBT is ∼1.7 years for both Si PV systems (Figures and 4) These calculations were based on the electricity mixture for the current production of Si, within the context of the CrystalClear project For CdTe PV, we previously estimated the energy consumption as 1200 MJ m−2, based on the actual production data of the year 2005 from the First Solar’s 25 MWp plant in Ohio, USA, close to the early studies reviewed [7, 8] The GHG emissions and EPBT of ground-mounted CdTe PV modules under the average US insolation condition, 1800 kWh m−2 yr−1, were determined to be 24 g CO2-eq kWh−1 and 1.1 years, correspondingly These estimates include g CO2-eq kWh−1 of GHG and 0.3 years of EPBT contribution from the ground-mounted BOS [14] On the other hand, Raugei et al [26] estimated a lower primary energy GHG (g CO2-eq kWh–1) 35 30 25 20 BOS 15 Frame 10 Module Multi-Si 13.2% Mono-Si 14.0% CdTe, 10.9% CdTe, 10.9% European US production production* Figure Life-cycle greenhouse gas (GHG) emissions from silicon and CdTe PV modules, wherein BOS is the balance of system, that is, the module supports, cabling, and power conditioning [2, 10, 13, 14, 26] Unless otherwise noted, the estimates are based on rooftop-mounted installation, Southern European insolation of 1700 kWh m−2 yr−1, a performance ratio of 0.75, and a lifetime of 30 years *Based on ground-mounted installation, average US insolation of 1800 kWh m−2 yr−1, and a performance ratio of 0.8 Mono-Si, monocrystalline silicon; multi-Si, multicrystalline silicon 1.8 EPBT (years) 1.6 1.4 1.2 BOS 0.8 Frame 0.6 Module 0.4 0.2 Multi-Si 13.2% Mono-Si 14.0% CdTe, 10.9% CdTe, 10.9% European US production production* Figure Energy payback time (EPBT) for silicon and CdTe PV modules, wherein BOS is the balance of system, that is, the module supports, cabling, and power conditioning [2, 10, 13, 14, 26, 27] Unless otherwise noted, the estimates are based on rooftop-mounted installation, Southern European insolation of 1700 kWh m−2 yr−1, a performance ratio of 0.75, and a lifetime of 30 years *Based on ground-mounted installation, average US insolation of 1800 kWh m−2 yr−1, and a performance ratio of 0.8 Mono-Si, monocrystalline silicon; multi-Si, multicrystalline silicon 150 Economics and Environment 1400 1210 Materials 1200 GHG (g CO2-eq kWh–1) Operation Transportation 1000 Fuel production 880 760 800 600 400 200 24 18a 28b Nuclear, baseline [28] PV, CdTe PV, mc-Si [50] [50] Coal [49] Natural gas [49] Petroleum [49] Figure Comparison of emissions from PV with those from conventional power plants Based on the average US insolation of 1800 kWh m−2 yr−1 and a performance ratio of 0.8; bbased on the Southern European insolation of 1700 kWh m−2 yr−1 and a performance ratio of 0.75 GHG, greenhouse gas; mc-Si, multicrystalline silicon a consumption, ∼1100 MJ m−2, and thereby less GHG emissions and lower EPBT than ours, based on the data of the year 2002 from the Antec Solar’s 10 MWp plant in Germany However, the latter estimates are obsolete as their plant ceased producing CdTe PV With continued growth in efficiency and reduction of electricity use in the new production lines, Fthenakis et al [50] updated CdTe PV’s environmental indicators using new data from the plant in Perrysburg, OH, and two studies based on data from the plant in Frankfurt-Oder, Germany (Figures and 4) The latest EPBT and GHG emissions based on the standard system boundary are 0.8 years and 18 g CO2‐eq kWh−1 for CdTe PV for typical rooftop installation in Southern Europe, that is, with irradiation of 1700 kWh m−2 yr−1 and a PR of 0.75 Note that the US estimates of CdTe in Figures and include R&D and administrative energy consumptions which are not included in other PV LCAs These updated EPBT and GHG figures are 30–35% lower than the previous estimates by Fthenakis and Kim [14] and Fthenakis et al [9] We note that this picture is not a static one and expect that improvements in material and energy utilization and recycling will continue to improve the environmental profiles A recent, major improvement is a recycling process for the sawing slurry, the cutting fluid that is used in the wafer cutting [27] This recycling process recovers 80–90% of the silicon carbide and polyethylene glycol, which used to be wasted On the other hand, any increases in the electrical conversion efficiencies of the modules will entail a proportional improvement of the EPBT Figure compares these emissions with those of conventional fuel-burning power plants, revealing the considerable environmental advantage of PV technologies The majority of GHG emissions are from the operation stage for the coal, natural gas, and oil fuel cycles, while the material and device production accounts for nearly all the emissions for the PV cycles The GHG emissions from the nuclear fuel cycle are mainly related to the fuel production, that is, mining, milling, fabrication, conversion, and enrichment of uranium fuel The details of the US nuclear fuel cycle are described elsewhere [28] 1.08.6 Criteria Pollutant and Heavy Metal Emissions 1.08.6.1 Criteria Pollutant Emissions The emissions of criteria pollutants during the life cycle of a PV system are largely proportional to the amount of fossil fuel burned during its various phases, in particular, PV material processing and manufacturing; therefore, the emission profiles are close to those of the GHG emissions (Figure 6) Toxic gases and heavy metals can be emitted directly from material processing and PV manufacturing, and indirectly from generating the energy used at both stages Accounting for each of them is necessary to create a complete picture of the environmental impact of a technology An interesting example of accounting for the total emissions is that of cadmium flows in CdTe and other PV technologies, as discussed next Environmental Impacts of Photovoltaic Life Cycles 151 (a) 70 Emissions (mg kWh–1) 60 50 40 BOS 30 Frame Module 20 10 Multi-Si, 13.2% Mono-Si, 14.0% CdTe, 10.9% (b) 140 Emissions (mg kWh–1) 120 100 80 BOS 60 Frame Module 40 20 Multi-Si, 13.2% Mono-Si, 14.0% CdTe, 10.9% Figure Life-cycle emissions of (a) NOx and (b) SOx from silicon and CdTe PV modules, wherein BOS is the balance of system (BOS), that is, module supports, cabling, and power conditioning The estimates are based on rooftop-mounted installation, Southern European insolation of 1700 kWh m−2 yr−1, a performance ratio of 0.75, and a lifetime of 30 years It is assumed that the electricity supply for all the PV system is from the UCTE grid Mono-Si, monocrystalline silicon; multi-Si, multicrystalline silicon 1.08.6.2 1.08.6.2.1 Heavy Metal Emissions Direct emissions Cadmium is a by-product of zinc and lead, and is collected from emissions and waste streams during the production of these major metals The largest fraction of cadmium, with ∼99.5% purity, is in the form of a sponge from the electrolytic recovery of zinc This sponge is transferred to a cadmium recovery facility and is further processed through oxidation and leaching to generate a new electrolytic solution After selectively precipitating the major impurities, cadmium of 99.99% purity is recovered by electrowinning It is further purified by vacuum distillation to the five 9s purity required for CdTe PV manufacturing The emissions during each of these steps are detailed elsewhere [12] They total up to 0.02 g GWh−1 of PV-produced energy under Southern European condition (Table 4) On the other hand, cadmium emissions during the life span of a finished CdTe module are negligible; the only conceivable pathway of release is if a fire broke out Experiments at Brookhaven National Laboratory that simulated real fire conditions revealed that CdTe is effectively contained within the glass-to-glass encapsulation during the fire, and only minute amounts (0.4–0.6%) of Cd are released The dissolution of Cd into the molten glass was confirmed by high-energy synchrotron X-ray microscopy [29] 1.08.6.2.2 Indirect emissions The indirect emissions here are those emissions associated with the production of energy used in mining and industrial processes in the PV life cycle Reporting indirect emissions separately from direct ones not only improves transparency in analyses but also allows calculating emissions for a certain mix of energy options as demonstrated in a recent study by Reich et al [30] Coal- and oil-fired power plants routinely generate Cd during their operation, as it is a trace element in both fuels According to the data from US Electric Power Research Institute (EPRI), under the best/optimized operational and maintenance conditions, burning coal for electricity releases into the air between and g Cd GWh−1 [31] In addition, 140 g Cd GWh−1 inevitably collects as fine dust in boilers, baghouses, and electrostatic precipitators (ESPs) Furthermore, a typical US coal-powered plant emits per GWh about 1000 tonnes of CO2, tonnes of SO2, tonnes of NOx (nitrogen oxide), and 0.4 tonnes of particulates The emissions of Cd from heavy oil-burning power plants are 12–14 times higher than those from coal plants, even though heavy oil contains much less Cd than coal (∼0.1 ppm), 152 Economics and Environment Table Direct, atmospheric Cd emissions during the life cycle of the CdTe PV module (allocation of emissions to coproduction of Zn, Cd, Ge, and In) Mining of Zn ores Zn smelting/refining Cd purification CdTe production PV manufacturing Operation Disposal/recycling Total Air emissions (g Cd tonne−1 Cd a) Allocation (%) Air emissions (g Cd tonne−1 Cd a) mg Cd GWh−1 b 2.7 40 6 0.3 0.58 0.58 100 100 100 100 100 0.016 0.23 6 0.3 15.55 0.02 0.3 9.1 9.1 4.5 0.3 23.3 a Tonne of Cd produced Energy produced assuming average Southern European insolation (i.e., 1700 kWh m−2 yr−1), 9% electrical conversion efficiency, and a 30-year life for the modules b because these plants not have particulate control equipment Cadmium emissions are also associated with the life cycle of natural gas and nuclear fuel because of the energy used in the processing and material production of the associated fuel [23] We accounted for Cd emissions in generating the electricity used in producing a CdTe PV system [32] The assessment of electricity demand for PV modules and BOS was based on the LCI of each module and the electricity input data for producing BOS materials Then, Cd emissions from the electricity demand for each module were assigned, assuming that the life-cycle electricity for the silicon and CdTe PV modules was supplied by the UCTE (European) grid The indirect Cd emissions from electricity usage during the life cycle of CdTe PV modules (i.e., 0.2 g GWh−1) are an order of magnitude greater than the direct ones (routine and accidental) (i.e., 0.016 g GWh−1) The complete life-cycle atmospheric Cd emissions, estimated by adding those from the electricity and fuel demand associated with manufacturing and material production for various PV modules and BOS, are compared with the emissions from other electricity-generating technologies (Figure 7) [9] Undoubtedly, displacing the others with Cd PV markedly lowers the amount of Cd released into the air Thus, every GWh of electricity generated by CdTe PV modules can prevent around g of Cd air emissions if they are used instead of, or as a supplement to, the UCTE electricity grid Also, the direct emissions of Cd during the life cycle of CdTe PV are 10 times lower than the indirect ones due to electricity and fuel use in the same life cycle, and about 30 times less than those indirect emissions from crystalline PVs [9] Furthermore, we examined the indirect heavy metal emissions in the life cycle of the three silicon technologies discussed earlier, finding that, among PV technologies, CdTe PV with the lowest EPBT has the lowest heavy metal emissions (Figure 8) [9] 50 43 45 Cd Emissions (g GWh–1) 40 35 30 25 20 15 10 6.2 4.1 3.1 0.7 Si lti- Mu 0.6 i o-S n Mo 0.2 Te Cd 0.2 al rd Ha co e nit Lig N as lg atu 0.5 Oil ar cle Nu 0.03 dro avg TE UC Hy Figure Life-cycle atmospheric Cd emissions for PV systems from electricity and fuel consumption, normalized for a Southern Europe average insolation of 1700 kWh m−2 yr−1, a performance ratio of 0.8, and a lifetime of 30 years A ground-mounted balance of system is assumed for all PV systems [9] Mono-Si, monocrystalline silicon; multi-Si, multicrystalline silicon Environmental Impacts of Photovoltaic Life Cycles 153 25 Multi-Si,13.2% Emissions (g GWh–1) 20 Mono-Si,14% CdTe,10.9% 15 10 Arsenic Chromium Lead Mercury Nickel Figure Emissions of heavy metals due to electricity use, based on European UTCE averages (Ecoinvent database) Emissions are normalized for Southern European average insolation of 1700 kWh m−2 yr−1, a performance ratio of 0.8, and a lifetime of 30 years Each PV system is assumed to include a ground-mounted balance of system as described by Mason et al [21] Mono-Si, monocrystalline silicon; multi-Si, multicrystalline silicon 1.08.7 Life-Cycle Risk Analysis 1.08.7.1 Risk Classification Perhaps the greatest potential risks of the PV fuel cycle are linked with chemical usage during material production and module processing [33, 34] Although rare, accidents like leakage, explosion, and fire of toxic and flammable substances are the major concerns to be addressed to ensure public acceptance of PV technology But little is known about their frequency and scale in terms of human fatalities, injuries, and economic losses during the PV fuel cycle In its burgeoning stage, the PV industry has not experienced major accidents of the same scale as other energy sectors Furthermore, the stages of the PV fuel cycle are closely connected to the semiconductor processes characterizing PV risks, that is, equitably allocating risks to the PV and semiconductor industries poses many difficulties The risks associated with energy technologies can be classified into four types based on their scale, frequency, and the severity of harmful events [35] Risk during normal operation – risks the consequences of which are typically accepted, for example, GHG emissions, toxic chemical emissions, routine radioactive emissions, chemical/radioactive waste, and resource (fuel, water) depletion Risk of routine accidents – risk of accidents with high frequency and low consequences, for example, small-scale leakage of chemicals, small-scale explosion/fires, transportation accidents, and small-scale radioactivity release Risk of severe accidents – risk of accidents with low frequency and high consequences, for example, core meltdown, collapse of dam, and large-scale fire/explosion Difficult-to-evaluate risks – subject to, and sometimes reinforced by, perception These include terrorist attacks on reactor/used fuel storage, geopolitical instability, military conflicts, energy security/national independence, and nuclear proliferation The first type of risk is characterized as routine under normal conditions as opposed to accidents, and its impact is usually limited by safety measures often established by regulation This type of risk is usually determined by analytical tools, such as LCA and impact pathway analysis The second and third types cover anomalous events or accidents The general public is more concerned about the third type of risk, low-probability catastrophic events, than the second type, high-probability less severe accidents The fourth type of risk is often associated with the public’s perception, and the probability and consequences are difficult to evaluate [36] 1.08.7.2 Risks of Accidents in the Photovoltaic Life Cycle The greatest potential risks of the PV life cycle involve the toxic and flammable chemicals used for producing PV materials and for manufacturing modules [34] Fthenakis and Kim undertook a life-cycle risk analysis based on accident records from a national database, the risk management plan (RMP) The study focused on the stages of handling and using the chemicals while making solar cell materials and the modules [35] The risks of potential accidents during installation, operation, and disposing/recycling of PV modules may be negligible, as then chemical usages are little or unnecessary 154 Economics and Environment According to the Clean Air Act Amendments of 1990, the US Environmental Protection Agency (EPA) is to publish regulations and guidance for facilities that use extremely hazardous substances to prevent chemical accidents The US EPA developed the RMP rule to implement section 112 (r) of the amendments This rule requires the facilities that produce, use, or handle toxic or flammable chemicals of concern in quantities over specific thresholds to submit an RMP Regulated substances and their threshold quantities are listed under section 112 (r) of the Clean Air Act in 40 CFR Part 68 The RMP should include a 5-year accident history detailing the cause of accidents and the damages to the employees, environment, and local public Other RMP components include a hazard assessment for the most likely and the worst-case accident scenarios, a prevention program covering safety precautions and maintenance, monitoring, and employee training measures, and an emergency response program that describes emergency health care, employee training measures, and procedures for informing the public and response agencies [37] In 1999, around 15 000 relevant facilities in the United States first submitted their accident records during the previous 5-year period, along with other documents The current (2005) updated RMPs mostly cover accidents that occurred since mid-1999 for 17 000 country-wide facilities The program must be resubmitted every years and revised whenever there is a significant change in usage Fthenakis et al [35] estimated the risks of the chemicals used in the PV life cycle, that is, accidents, fatalities, and injuries during the production, storage, and delivery of the flammable and toxic chemicals listed in 40 CFR Part 68, based on the current and past histories in the RMPs reported Table shows the number of accidents/death/injuries between 1994 and 2004 for substances involved in PV material and module production Since there are very few data on the production/consumption of trichlorosilane (SiHCl3), estimates are given based on polysilicon production and capacity data About 2.5 kg of polysilicon is needed to produce a multi-Si module of  10 cells of 156 mm  156 mm and an input of 11.3 kg of trichlorosilane is required to generate kg of polysilicon [2, 19] No accidents were reported in the RMP database for phosphine and diborane in these years We note that the number of occurrences shown in Table represents the total for all the US facilities that produce, process, handle, and store these chemicals in quantities over their specified thresholds These statistics may include data irrelevant to the PV industry For example, the majority (61%) of accidents/death/injuries attributed to hydrogen fluoride (HF) occurred in petroleum refineries (NAICS 32411 (North American Industry Classification System)), although refinery use accounts for only 6% of the US consumption [38] The petrochemical industry uses 100% HF under high pressure in large, multicomponent units (e.g., alkylation units with many pipes, fittings, valves, compressors, and pumps) from where two-phase releases may occur, whereas the PV industry uses only aqueous solutions of HF (typically 49 wt.%) in etching baths Moreover, the usage of HF and hydrogen chloride (HCl) in the US PV industry accounts for less than 0.1% of the total in the United States Therefore, we excluded HF use in the petroleum industries and the corresponding accidents from our analysis of risks in the PV fuel cycle Table Incident records and estimated production for substances used in PV operation and regulated under the RMP rule Substance Source Total average (1994–2000) US production (1000 tonnes yr−1) Toxic Arsine (AsH3) Boron trichloride (BCl3) Boron trifluoride (BF3) Diborane (B2H6) Hydrochloric acid (HCl) Hydrogen fluoride (HF) Hydrogen selenide (H2Se) Hydrogen sulfide (H2S) Phosphine (PH3) GaAs chemical vapor deposition (CVD) p-type dopant for epitaxial silicon p-type dopant for epitaxial silicon a-Si dopant Cleaning agent for c-Si Etchant for c-Si CIGS selenization CIS sputtering a-Si dopant Flammable Dichlorosilane (SiH2Cl2) Hydrogen (H2) Silane (SiH4) Trichlorosilane (SiHCl3) a-Si and c-Si deposition a-Si deposition/GaAs a-Si deposition Precursor of c-Si a Number of incidents (employees and public) in the RMP database (1994–2004) Incidents Injuries Deaths 23 NA NA ∼50 3500 190 N/A >110 N/A 28 165 (57)a 40 1 12 209 (70)a 17 47 0 0 1 N/A 18 000 m3 110 57 14 65 14 Number excludes incidents in petroleum refineries (NAICS 32411) a-Si, amorphous silicon; c-Si, crystalline silicon; CIGS, copper indium gallium selenide; CIS, copper indium diselenide; RMP, risk management program Reproduced from Williams E (2000) Global Production Chains and Sustainability: The Case of High-Purity Silicon and Its Applications in IT and Renewable Energy Tokyo, Japan: Institute of Advanced Studies, The United Nations University [19]; EPA (1992) Hydrogen Fluoride Study Report to Congress Washington, DC: Office of Solid Waste and Emergency Response, US Environmental Protection Agency [38]; Census of Bureau (2002) Current Industrial Reports –? Inorganic Chemicals (MQ325A) [39]; Board on Environmental Studies and Toxicology (BEST) (2003) Acute Exposure Guideline Levels for Selected Ariborne Chemicals, vol The National Academic Press [40]; U.S Geological Survey (2003) Mineral Commodity Summary: Arsenic [41]; Agency for Toxic Substances and Disease Registry (2004) Toxicological Profile for Hydrogen Sulfide [42]; Pichel JW and Yang M (2006) 2005 Solar Year-End Review & 2006 Solar Industry Forecast Piper Jaffray & Co [43] Environmental Impacts of Photovoltaic Life Cycles 155 101 Number of injuries (GWyr–1) 100 10–1 10–2 10–3 10–4 10–5 10–6 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 99–04 Figure Estimated injury rates of crystalline silicon (c-Si) per GWyr electricity produced in the United States based on the year of accidents for HF, HCl, and SiHCl3 (insolation = 1800 kWh m−2 yr−1, performance ratio = 0.8) To normalize the rate of accident/death/injuries, the figures are divided by the amount produced for those chemicals in the United States The risks of each substance used in a 25 MWp yr−1 scale PV industry are determined based on the amount of materials in Table Then, the number of accidents/death/injuries per GWyr of electricity generated is determined as a risk indicator of chemical accidents in the PV fuel cycle based on the average US insolation of 1800 kWh m−2 yr−1 and a PR of 0.8 (i.e., 20% system loss) Figure depicts the rates of accidents/death/injuries for c-Si per GWyr electricity produced based on the average US insolation, 1800 kWh m−2 yr−1 The last column with the most recent incident data (i.e., data submitted from 1999 onward) may better represent the current evolution of the fast-growing PV industry In general, for most chemicals, the numbers of accidents reported for the second cycle of the RMP (second half of 1999–2004) are lower than those reported in the first cycle (1994 to the first half of 1999) Specifically, only one accident involving silane was reported during the second cycle of the RMP, whereas four accidents occurred during the first Likewise, the number of accidents during the second RMP cycle involving trichlorosilane fell from 13 to between 1994–99 and 1999–2004 This across-the-board reduction of incidents likely represents improved safety in the whole US PV industry For c-Si PV modules, SiHCl3, the feedstock of polysilicon, presents greater risks than other chemicals (Figure 10) On the other hand, limited statistics suggest that silane poses the greatest risk in a-Si module manufacturing The fatalities in Figure 10 are related to HF However, this risk is not based on the real events in PV facilities but is a virtual risk, derived from accident rates in other industries and the amount consumed in the PV industry The real risk of HF in the PV industry, however, is likely to be lower than this current estimate due to the different characteristics of processes across industry sectors 1.08.7.3 Comparison with Other Energy Technologies Figure 11 compares fatality and injury rates across conventional electricity technologies and PV technologies The figures for the conventional fuel cycles were extracted from the compilation of severe accident records of the GaBE project by the Paul Scherrer Institute (PSI), Energy-Related Severe Accident Database (ENSAD), from 1969 to 2000 [44] Although such a direct comparison may not be entirely appropriate, the data imply that the expected risks in the PV fuel cycle are lower than the risks in other technologies, as we explain below Thus, there are several caveats in directly comparing our estimate of PV risks and the risks estimated by the PSI’s investigators First, Hirschberg et al.’s assessments are based on severe accidents only, and small-scale accidents are ignored [44]; the former are defined as events with at least fatalities, 10 injuries, $5 million of property damage, or 200 evacuees On the other hand, only 20 out of the 318 incidents (Table 5) can be classified as severe accidents under the same definition Therefore, the PSI’s values for risk of conventional energy technologies may be underestimates Also, we expect the safety records of the evolving PV industry to improve with time, whereas those for the mature conventional energy technologies are less likely to change The risk estimate for the nuclear cycle in Figure 11 relies on the probabilistic safety assessment (PSA) tool for the OECD countries rather than historical evidence due to insufficient experience related to the nuclear accidents We note that the figures should be updated with the recent nuclear accident in Fukushima, Japan Although PSA is a logical approach to identify plausible accidents, their rates, and consequences, it is not proven to provide absolute risk estimates due to its limitations associated with data uncertainty; for example, risks of unexpected events or those triggered by human errors are impossible to fully account for as evidenced in the recent nuclear 156 Economics and Environment Number of incidents (GWyr–1) 10–1 Death Injuries 10–2 10–3 10–4 10–5 10–6 HCl HF SiHCl3 c-Si (total) a-Si Figure 10 Estimated incident rates by chemicals used between 1999 and 2004 (insolation = 1800 kWh m−2 yr−1, performance ratio = 0.8) The rate of a-Si covers 1997–2004 a-Si, amorphous silicon; c-Si, crystalline silicon 102 Number of incidents (GWyr –1) 101 Fatalities Injuries 100 10–1 10–2 10–3 10–4 10–5 10–6 D D D D CD CD PSI* NL** NL** D D D CD CD CD E , E B B OE n-OEil, OE n-OEC, OECn-OEC, OECn-OEC, OECn-OEC , l , r O n-O PVc-Si, a-Si, a o , Co al, no O il, no NGG, no LPGG, no Hydr ro, nouclea ar, no , V PV P O N ucle d N Co LP Hy N Figure 11 Comparison of fatality and injury rates across electricity generation technologies The average US insolation of 1800 kWh m−2 yr−1 and a performance ratio of 0.8 were assumed The incident rates for coal, oil, natural gas (NG), liquefied petroleum gas (LPG), hydro, nuclear, and PV technologies given by Paul Scherrer Institute (PSI) are from Hirschberg S, et al (2004) Sustainability of Electricity Supply Technologies under German Conditions: A Comparative Evaluation Switzerland: Paul Scherrer Institute [44] *Module type is not specified and injuries are not estimated; **estimates of Brookhaven National Laboratory (BNL) based on US data [35] a-Si, amorphous silicon; c-Si, crystalline silicon disaster in Japan [45, 46] For a similar reason, the PSI’s estimate for the PV cycle relied on data of the chemical industries handling similar substances used in the PV cycle (B Peter, personal communication, 2006) We also examined the maximum consequences of each energy technology People are rarely neutral about risk; decision makers, risk analysts, and the public are more interested in unforeseen catastrophes, such as bridges falling, dams bursting, earthquakes and tsunami’s occurring, and nuclear reactors exploding, than in adverse but routine events, such as transportation accidents Comparing low-probability/high-damage risks with high-probability/low-damage events within one expected value frame often distorts the relative importance of consequences across technology options Therefore, describing the maximum consequence potential makes sense [36] Figure 12 shows the maximum fatalities: figures for the fossil fuel and nuclear cycles are from a database Environmental Impacts of Photovoltaic Life Cycles 157 105 9000–33000 104 Fatalities 3000 103 600 434 100 102 125 100 101 100 al Co Oil l) l) L G SI BN LP oby oby ,P V, PV ern ern P h h C C ar ( xcept cle e ( Nu r a cle Nu NG Figure 12 Maximum fatalities from accidents across energy sectors The number for Chernobyl includes latent fatalities The incident rates for coal, oil, natural gas (NG), liquefied petroleum gas (LPG), hydro, nuclear, and PV technologies given by Paul Scherrer Institute (PSI) are from Hirschberg S, Spikerman G, and Dones R (1998) Severe Accidents in the Energy Sector, 1st edn Switzerland: Paul Scherrer Institute [33] BNL, Brookhaven National Laboratory [33] and for PV from literature [35, 44] From a scale of consequence perspective, PV technology is remarkably safer than other technologies We anticipate that the maximum consequence will remain at the same level shown in Figure 12 unless there is a significant change in PV production technologies The nature of risks varies with different energy technologies The PSI analysis focuses on fuel mining, fuel conversion, power plant operation, and transportation of fuels On the other hand, fuel-related or power plant-related risks are nonexistent in the PV fuel cycle Instead, our analysis of PV risks focuses on accidents associated with feedstock materials as well as process consumables, that is, upstream risks, which the PSI analysis does not include 1.08.7.4 Limitation of the Study As discussed, the hazards of trichlorosilane when making modules of c-Si and that of silane for a-Si modules have dominated concerns over other chemicals due to the large amount required and their flammability Their limited number of incident records in the RMP database prevents accurate measurement of the safety of the PV industry Since death records are very rare for these chemicals, the risk of fatalities is highly sensitive to a single incident On the other hand, injury rates are relatively stable against the incremental number of injuries This illustrates that the scale of the PV industry in terms of capacity and employees is still small so that such analyses are inadequate to directly compare with other technologies PV technology is still in the early stage of commercial application, and, with time, risk management programs complied through experience eventually should stabilize the number of abnormal incidents Other technologies, such as nuclear power, experienced a similar period during the early years of commercialization [33] Although PV technology is at an early stage compared with other energy systems, it is rapidly growing within the context of EU policy to increase the contribution of renewable energies to the total EU energy mix [47, 48] Accordingly, there are strong efforts to boost the PV sector, and therefore, it should be treated more and more as an energy sector, avoiding comparisons with the chemical or semiconductor industry Our analysis of risk in the PV life cycle is at a level lower than many other technologies Given this situation, the PV industry should help by sharing information on risk, which, at the same time, could surely give PV more credibility and transparency 1.08.8 Conclusion This review offers a snapshot of the rapidly evolving life-cycle performances of PV technologies and underlines the importance of timely updating and reporting the changes During the life cycle of PV, emissions to the environment mainly occur from using fossil fuel-based energy in generating the materials for solar cells, modules, and systems These emissions differ in different 158 Economics and Environment countries, depending on that country’s mixture in the electricity grid and the varying methods of material/fuel processing The lower the EPBT, that is, the time it takes for a PV system to generate energy equal to the amount used in its production, the lower these emissions will be Under average US and Southern Europe conditions (e.g., 1700 kWh m−2 yr−1) with rooftop application, the EPBT of multi-Si, mono-Si, and CdTe systems was estimated to be 1.7, 1.7, and 0.8 years, correspondingly The EPBT of CdTe PV is the lowest in the group, although electrical conversion efficiency was also the lowest; this was due to the low energy requirement in manufacturing CdTe PV modules The indirect emissions of Cd due to energy used in the life of CdTe PV systems are much greater than the direct emissions CdTe PV systems require less energy input in their production than other commercial PV systems, and this translates into lower emissions of heavy metals (including Cd) as well as SO2, NOx, particulate matter (PM), and CO2 in the CdTe cycle than in other commercial PV technologies However, regardless of the particular technology, these emissions are extremely small in comparison to the emissions from the fossil fuel-based plants that PV will replace The greatest potential risks of the PV fuel cycle are linked with the use of several hazardous substances, although in quantities much smaller than in the process industries In an effort to measure the potential risks associated with PV fuel cycle in comparison with other electricity generation technologies, we used the US EPA’s RMP accident records that cover the entire major US chemical storage and processing facilities Our analysis shows that, based on the most recent records, the PV fuel cycle is much safer than conventional sources of energy in terms of statistically expected risks and by far the safest in terms of maximum consequence Acknowledgement The authors gratefully acknowledge support from the US DOE Office of Energy Efficiency and Renewable Energy, under contract DE-AC02-76CH000016, to Brookhaven National Laboratory References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] Alsema EA (2000) Energy pay-back time and CO2 emissions of PV systems Progress in Photovoltaics: Research and Applications 8: 17–25 Alsema E and de Wild-Scholten M (2005) Environmental impact of crystalline silicon photovoltaic module production In: Material Research Society Fall Meeting, Symposium G: Life Cycle Analysis Tools for “Green” Materials and Process Selection, pp 73–82 Boston, MA Meijer A, Huigbregts MAJ, Schermer JJ, and Reijnders L (2003) Life-cycle assessment of photovoltaic modules: Comparison of mc-Si, InGaP and InGaP/mc-si solar modules Progress in Photovoltaics: Research and Applications 11(4): 275–287 Jungbluth N (2005) Life cycle assessment of crystalline photovoltaics in the Swiss ecoinvent database Progress in Photovoltaics: Research and Applications 13(8): 429–446 Hagedorn G and Hellriegel E (1992) Umwelrelevante masseneinträge bei der herstellung verschiedener solarzellentypen – endbericht – teil I: konventionelle verfahren München, Germany: Forschungstelle für Energiewirtschaft Palz W and Zibetta H (1991) Energy pay-back time of photovoltaic modules International Journal of Solar Energy 10: 211–216 Hynes KM, Baumann AE, and Hill R (1994) An assessment of the environmental impacts of thin film cadmium telluride modules based on life cycle analysis In: IEEE 1st World Conference on Photovoltaic Conversion (WCPEC), pp 958–961 Hawaii Kato K, Hibino T, Komoto K, et al (2001) A life-cycle analysis on thin-film CdS/CdTe PV modules Solar Energy Materials & Solar Cells 67: 279–287 Fthenakis VM, Kim HC, and Alsema E (2008) Emissions from photovoltaic life cycles Environmental Science & Technology 42: 2168–2174 Fthenakis V and Alsema E (2006) Photovoltaics energy payback times, greenhouse gas emissions and external costs: 2004–early 2005 status Progress in Photovoltaics: Research and Applications 14: 275–280 Fthenakis V, Wang W, and Kim HC (2009) Life cycle inventory analysis of the production of metals used in photovoltaics Renewable and Sustainable Energy Reviews 13: 493–517 Fthenakis VM (2004) Life cycle impact analysis of cadmium in CdTe PV production Renewable and Sustainable Energy Reviews 8(4): 303–334 de Wild-Scholten M and Alsema E (2005) Environmental life cycle inventory of crystalline silicon photovoltaic module production In: Material Research Society Fall Meeting, Symposium G: Life Cycle Analysis Tools for “Green” Materials and Process Selection, pp 59–72 Boston, MA Fthenakis VM and Kim HC (2006) Energy use and greenhouse gas emissions in the life cycle of CdTe photovoltaics In: Material Research Society Fall Meeting, Symposium G: Life Cycle Analysis Tools for “Green” Materials and Process Selection, pp 83–88 Boston, MA Alsema EA and de Wild-Scholten MJ (2007) Reduction of the environmental impacts in crystalline silicon module manufacturing In: 22nd European Photovoltaic Solar Energy Conference, pp 829–836 Milano, Italy Aulich HA and Schulze F-W (2002) Crystalline silicon feedstock for solar cells Progress in Photovoltaics: Research and Applications 10: 141–147 Woditsch P and Koch W (2002) Solar grade silicon feedstock supply for PV industry Solar Energy Materials & Solar Cells 72: 11–26 Maycock PD (2005) National survey report of PV power applications in the United States 2004 International Energy Agency Williams E (2000) Global Production Chains and Sustainability: The Case of High-Purity Silicon and Its Applications in IT and Renewable Energy Tokyo, Japan: Institute of Advanced Studies, The United Nations University Rogol M (2005) Silicon and the solar sector: Mapping a new world In: Presentation at the 2nd Solar Silicon Conference Munich, Germany, 11 April Mason JE, Fthenakis VM, Hansen T, and Kim HC (2006) Energy payback and life-cycle CO2 emissions of the BOS in an optimized 3.5 MW PV installation Progress in Photovoltaics: Research and Applications 14: 179–190 de Wild-Scholten MJ, Alsema EA, ter Horst EW, et al (2006) A cost and environmental impact comparison of grid-connected rooftop and ground-based PV systems In: 21st European Photovoltaic Solar Energy Conference, pp 3167–3173 Dresden, Germany Dones R, Bauer C, Bolliger R, et al (2003) Sachbilanzen von Energiesystemen Final report ecoinvent 2000, vol Swiss Centre for LCI, PSL, Dubendorf, CH Franklin Associates (1998) USA LCI Database Documentation Prairie Village, KS: Franklin Associates Forster P, Ramaswamy V, Artaxo P, et al (2007) Changes in atmospheric constituents and in radiative forcing In: Climate Change 2007: The Physical Science Basis Contributions of Working Group I to the Fourth Annual Assessment Report of the Intergovernmental Panel on Climate Change Cambridge UK and New York: Cambridge University Press Raugei M, Bargigli S, and Ulgiati S (2007) Life cycle assessment and energy pay-back time of advanced photovoltaic modules CdTe and CIS compared to poly-Si Energy 32: 1310–1318 Environmental Impacts of Photovoltaic Life Cycles 159 [27] Alsema EA, de Wild-Scholten MJ, and Fthenakis VM (2006) Environmental impacts of PV electricity generation – A critical comparison of energy supply options In: 21st European Photovoltaic Solar Energy Conference, pp 3201–3207 Dresden, Germany [28] Fthenakis VM and Kim HC (2007) Greenhouse-gas emissions from solar electric- and nuclear power: A life-cycle study Energy Policy 35: 2549–2557 [29] Fthenakis VM, Fuhrmann M, Heiser J, et al (2005) Emissions and encapsulation of cadmium in CdTe PV modules during fires Progress in Photovoltaics: Research and Applications 13: 713–723 [30] Reich NH, Alsema EA, van Sark WGJHM, et al (2011) Greenhouse gas emissions associated with photovoltaic electricity from crystalline silicon modules under various energy supply options Progress in Photovoltaics: Research and Applications 19: 603–613 [31] Electric Power Research Institute (EPRI) (2002) PISCES data base for US power plants and US coal [32] Fthenakis VM and Kim HC (2007) CdTe photovoltaics: Life cycle environmental profile and comparisons Thin Solid Films 515: 5961–5963 [33] Hirschberg S, Spikerman G, and Dones R (1998) Severe Accidents in the Energy Sector, 1st edn Switzerland: Paul Scherrer Institute [34] Fthenakis V (2003) Overview of potential hazards In: Markvart T and Gastaner L (eds.) Practical Handbook of Photovoltaics: Fundamentals and Applications, pp 857–868 Kidlington, UK: Elsevier [35] Fthenakis VM, Kim HC, Colli A, and Kirchsteiger C (2006) Evaluation of risks in the life cycle of photovoltaics in a comparative context In: 21st European Photovoltaic Solar Energy Conference, pp 3194–3201 Dresden, Germany [36] Haimes YY (1998) Risk Modeling, Assessment, and Management Hoboken, NJ: John Wiley & Sons [37] Environmental Protection Agency Risk Management Plan (RMP) Rule http://www.epa.gov/osweroe1/content/rmp/ (accessed September 2011) [38] EPA (1992) Hydrogen Fluoride Study Report to Congress Washington, DC: Office of Solid Waste and Emergency Response, US Environmental Protection Agency [39] Census of Bureau (2002) Current Industrial Reports – Inorganic Chemicals (MQ325A) [40] Board on Environmental Studies and Toxicology (BEST) (2003) Acute Exposure Guideline Levels for Selected Ariborne Chemicals, vol Washington, DC: The National Academic Press [41] U.S Geological Survey (2003) Mineral Commodity Summary: Arsenic [42] Agency for Toxic Substances and Disease Registry (2004) Toxicological Profile for Hydrogen Sulfide http://www.atsdr.cdc.gov/ [43] Pichel JW and Yang M (2006) 2005 Solar Year-End Review & 2006 Solar Industry Forecast Minneapolis, MN: Piper Jaffray & Co [44] Hirschberg S, Dones R, Heck T, et al (2004) Sustainability of Electricity Supply Technologies under German Conditions: A Comparative Evaluation Switzerland: Paul Scherrer Institute [45] Linnerooth-Bayer J and Wahlström B (1991) Applications of probabilistic risk assessments: The selection of appropriate tools Risk Analysis 11(2): 239–248 [46] Nuclear Energy Agency (1992) Probabilistic safety assessment: An analytical tool for assessing nuclear safety In: NEA Issue Brief: An Analysis of Principal Nuclear Issues, No 8, January 1992 [47] Directive 2001/77/EC of the European Parliament and of the Council (2001) Promotion of Electricity Produced from Renewable Energy Sources in the Internal Electricity Market, 27 September [48] European Commission (1997) Communication from the commission “Energy for the Future: Renewable Sources of Energy”, White Paper for a Community Strategy and Action Plan [49] Kim S and Dale BE (2005) Life cycle inventory information of the United States Electricity System International Journal of Life Cycle Assessment 10: 294–304 [50] Fthenakis V, Kim HC, Held M, et al (2009) Update of PV energy payback times and life-cycle greenhouse gas emissions In: 24th European Photovoltaic Solar Energy Conference, pp 4412–4416 Hamburg, Germany  ... Co [43] Environmental Impacts of Photovoltaic Life Cycles 15 5 10 1 Number of injuries (GWyr 1) 10 0 10 1 1 0–2 10 –3 10 –4 10 –5 10 –6 19 95 19 96 19 97 19 98 19 99 2000 20 01 2002 2003 2004 9 9–0 4 Figure... 0.8) The rate of a-Si covers 19 9 7–2 004 a-Si, amorphous silicon; c-Si, crystalline silicon 10 2 Number of incidents (GWyr 1) 10 1 Fatalities Injuries 10 0 10 1 1 0–2 10 –3 10 –4 10 –5 10 –6 D D D D CD... (MJ) 1. 6 9 .1 1.0 1. 8 2.2 6.8 4.3 248 0.05 308 1. 5 9 .1 1.0 1. 8 7.8 6.6 4.3 282 0.05 3 61 0.065 19 .2 0.6 2.0 0.0 01 0.67 0.4 59 0.05 – Consumables (kg) Energy Environmental Impacts of Photovoltaic Life