Greenhouse gas emissions embodied in the U.S. solar photovoltaic supply chain

Solar photovoltaic (PV) electricity is considered to be an important source of electricity generation in the quest for net-zero carbon emissions. However, the growth of solar electricity is creating both increased material demands and increased greenhouse gas (GHG) emissions from silicon and PV manufacturing (also referred to as embodied GHG emissions of solar electricity). Here we analyze the silicon and solar PV supply chain for the United States (U.S.) market and find that the embodied GHG emissions of solar PV panel materials (such as silicon), manufacture, logistics, and installation in the U.S. given the current supply chain are 36 g CO2e kWh−1 of solar electricity generated. Eighty-five percent of the embodied GHG emissions are from PV panel production processes in China and other Asia–Pacific countries. Moving the silicon and PV manufacturing to the U.S. would reduce the embodied GHG emissions of solar electricity by 16% from its current level, primarily because of the lower GHG emission intensity of the U.S. electrical grid and the lower GHG emissions for aluminum electrolysis in North America. Future scenario analysis shows that by 2030, with the U.S. PV domestic supply chain and its decarbonized grid electricity and aluminum production, as well as improving PV conversion efficiency, the embodied GHG emissions of solar electricity in the U.S. will be reduced to 21 g CO2e kWh−1.


Introduction
Decarbonizing the electricity grid is an important means of reducing economy-wide greenhouse gas (GHG) emissions (Bistline 2021, Fankhauser et al 2022) while replacing fossil fuels with renewable energy, such as solar photovoltaic (PV) energy is the key to electric sector decarbonization (Margolis 2021, USDOE 2021).To achieve the U.S. government's goal of reducing the carbon footprint of the U.S. electricity grid by 95% from the 2005 level by 2035 and reaching net-zero carbon emissions by 2050, the installed capacity of solar PV power would need to increase from the 2020 level of 74 GW by 8.5-11.5 times in 2035 and by 12-19 times in 2050 to meet 40 and 45% of the U.S. electricity demand, respectively (U.S.Department of Energy (USDOE) Solar Energy Technologies Office 2021, Margolis 2021).That is, solar PV energy will become the major source of U.S. electricity generation.
Although electricity generation from a solar PV system has almost zero emissions during its lifespan of 25-30 years, the manufacturing and installation of solar PV systems require energy/material inputs and produce emissions (U.S.Department of Energy (USDOE) Solar Energy Technologies Office 2021, Margolis 2021).The so-called embodied emissions are being debated for inclusion in the holistic evaluation of solar electricity and its downstream use such as hydrogen production via electrolysis (Palmer et al 2021).As the U.S. is now incentivizing the establishment of the domestic solar PV supply chain for its ambitious goal of solar electricity growth, it is important to examine embodied GHG emissions of the U.S. solar PV supply chain in the future when economywide decarbonization efforts will be underway.
The Greenhouse Gases, Regulated Emissions, and Energy Use in Technologies (GREET) model developed at Argonne National Laboratory includes the life cycle analysis (LCA) of solar electric generation, among other electric generation technologies.GREET LCA includes supply chains of energy resources (coal, natural gas, etc) for power generation (called fuel cycle analysis) and infrastructures establishment of energy systems (power plant construction, wind turbine, and solar PV manufacture, etc) (called facility cycle analysis).In this study, we expanded the GREET module for the solar PV facility cycle by analyzing the global PV supply chain for the U.S. market.We collected data on material/energy use of solar PV manufacturing from literature including statistical reports, technical papers, and research articles for the GREET expansion to generate up-todate embodied GHG emissions of solar electricity in the U.S. We explored potential GHG emissions reduction by tracking emissions sources in each production stage.We further conduct scenario analyses to quantify GHG mitigation effects for measures of electricity grid decarbonization, increased aluminum recycling, and improved PV conversion efficiency.Results of the GREET solar PV module can be used in identifying hot spots of solar PV manufacturing for GHG reduction opportunities, holistically evaluating GHG emissions of low-GHG technologies such as hydrogen and synthetic fuels from CO 2 resources and solar electricity, and estimating the trend of GHG emission intensity of electricity grid over time.

Background and literature review
Solar PV systems have a global supply chain, with China dominating due to low production costs for silicon and PV products and relevant raw materials (Woodhouse et al 2019, Smith et al 2021).Chinese production shares in global production of crystalline silicon, silicon wafers, solar cells, and solar panels increased from 42%, 77%, 60%, and 66% in 2010 to 76%, 96%, 79%, and 77% in 2021, accompanied by a reduced market share and production capacity for suppliers in Europe and the U.S. (Smith et al 2021, Markit 2022).However, such a shift in the global supply chain might lead to increased GHG emissions and environmental impacts (Woodhouse et al 2019, Smith et al 2021) due to the relatively higher embedded GHG emissions for electricity and other materials produced in China (Smith and Margolis 2019).With more than 60% of China's electricity generation from coal in 2020 (IEA 2022a), its electricityintensive processes of polysilicon and solar PV production might generate a higher carbon footprint than the same production in Europe and the U.S.
Thus, analysis of the embodied emissions associated with solar PV electricity in the U.S. must consider global supply chains and regional variations in energy and material production.
Prior studies have analyzed the embodied GHG emissions of solar PV systems (Fthenakis et al 2008, Fthenakis and Kim 2009, 2011, Peng et al 2013, Lunardi et al 2018, Antonanzas et al 2019, Liu and Van den Bergh 2020, Anctil 2021, Müller et al 2021, Heidari et al 2022, Sinha and Hammann 2022).However, most of them did not explore the effects of globalized supply chains, regional variations in solar PV manufacturing, and/or upstream material production; instead, they usually focused on the same region for the entire material production and solar PV manufacturing process (Liu and Van den Bergh 2020), or comparing the current dominating technology with an advanced PV cell technology (Lunardi et al 2018).Some simply assumed the whole solar PV manufacturing process in China (Antonanzas et al 2019).More recent studies (Anctil 2021, Müller et al 2021, Heidari and Anctil 2022) recognized and analyzed the variations in production technologies, energy intensities, and carbon footprints of polysilicon and solar PV manufacturing across countries.However, these studies either focused on one specific production stage, for instance, the metallurgical grade silicon production stage (Heidari and Anctil 2022), the module design (Müller et al 2021), or only compared production in China versus the U.S., instead of analyzing the completed production processes and the whole global supply chain.The International Energy Agency's (IEA) LCA of solar PV manufacturing analyzed global supply chains and regional variations for each production stage, but only evaluated differences in electricity generation mixes but not other upstream processes, such as aluminum and glass production manufacturing (Krebs et al 2020, Rolf Frischknecht 2020).In addition, the IEA's analysis was intended for the entire global PV market.But due to unique U.S. policies, the supply chain for the U.S. market could be quite different from that for the global market.Thus, there may be different embodied emissions for solar PV electricity generation in the U.S., which are addressed specifically in this study.

Methods
We used GREET to develop facility cycle pathways for silicon solar PV manufacturing, maintenance, and end-of-life treatment, and to estimate the embodied GHG emissions of solar electricity generation in the U.S. (Wang 1999, 2022, Wang et al 2022).Since cSi PV accounts for ∼95% of the current global solar PV module installations (Fraunhofer ISE 2022), we focus on cSi PV technology here.Figure 1 shows the LCA system boundary.The system boundary of LCA of the cSi solar PV system (transportation activities, not shown in the figure, are also included in the system boundary).The LCA analysis includes the production of metallurgical grade silicon, solar grade polysilicon, single-and multi-cSi ingots and wafers, single-and multi-cSi cells and panels, and other relevant materials and energy, solar PV system installation, and solar PV system's end-of-life treatment.
Recognizing the huge variations in the GHG intensity levels of electricity and other materials (e.g.aluminum and glass) among regions, we considered the global supply chain of silicon and PV manufacturing (as shown in figure 2) (IEA 2022b) and the effects of regionalized electricity, aluminum, and glass production on the emissions from each stage of silicon and PV manufacturing.Figure 2 shows the supply chain of each stage in silicon and PV manufacturing for solar PV panels used in the U.S. (IEA 2022b).
As shown in figure 2, Chinese production met only 17% of U.S. cSi solar PV cell demand (including direct exports to the U.S. and exports to other countries for solar PV panel production used in the U.S.), and 5% of U.S. cSi solar PV panel demand.This differs from the global market share, where China accounted for around 80% of cSi solar PV cell and panel supply (Smith et al 2021, IHS Markit 2022, IEA 2022b).U.S. policy has extended tariffs on Chinese products since 2011, reducing China's direct exports of solar PV cells and panels to the U.S. APAC countries like Malaysia, Vietnam, Thailand, and South Korea met 76% and 64% of U.S. solar PV cell and panel demand from 2017 to 2021, respectively.Despite this, China still supplies 63% and 82% of the polysilicon and silicon wafer needs, respectively, for U.S. solar PV systems.As the main producer of silicon wafers globally (96% in 2021), China remains critical to provide silicon wafers for solar PV cell production in the U.S. and APAC countries.
For production processes in different regions, the corresponding regional electricity mix was applied to estimate upstream emissions associated with electricity input (IEA 2022a).Since China and APAC regions account for a major part of the supply chain and there are significant variations in electricity sources among different provinces in China and different countries in the APAC region, we applied the production-weighted average electricity mix for these two regions.The production volume of polysilicon, silicon wafers, PV cells, and PV panels of different Chinese provinces and APAC countries was obtained from the literature (Smith et al 2021, Basore andFeldman 2022).The consumption-based electricity mixes for different provinces of China were estimated based on earlier studies (CEC 2018, CNBS 2018, Gan et al 2021).The electricity mix for APAC countries was sourced from the electricity information by (IEA 2022a).Emissions associated with aluminum and glass input for silicon and PV manufacturing in China were derived from a China-specific GREET model (Gan et al 2020(Gan et al , 2021)).For aluminum production in other regions besides China, we applied the region-specific electricity mix for aluminum electrolysis reported by the International Aluminum Institute (IAI 2022) in the GREET model to estimate the varied regional upstream emissions of aluminum production.
Energy efficiency and material input data for silicon and PV manufacturing processes were mainly obtained from the IEA's PV systems life cycle assessment report (Rolf Frischknecht et al 2020), except for production processes in China where energy efficiency data were sourced from Chinese studies (Itten and Frischknecht 2014, Yao et al 2014, Fu et al 2015, Yang et al 2015).Material inventory for solar PV panels is shown in table S1 in the supplemental information (SI).We accounted for emissions from material transportation by both on-road and ocean shipping.The energy and material flow data for each process are presented in GREET 2022 (Wang et al 2022).
GHG emissions in this study, presented in CO 2 equivalent (CO 2 e) emissions, include CO 2 , CH 4 , and N 2 O emissions combined with their global warming potentials from Intergovernmental Panel on Climate Change 6th Assessment Report.
Embodied emissions per kWh of electricity output are calculated as the life cycle emissions of the solar PV system divided by the lifetime solar electricity generation.The lifetime solar electricity generation is calculated by the area of the solar panel, the solar radiation intensity, the system performance ratio and its degradation over time, and the system lifespan.Section 6 in the SI presents the detailed methods, data, and assumptions for calculating the lifetime electricity generation of solar PV systems.

Supply chain emissions contribution of silicon PV panels consumed in the U.S
Using the supply chain material flow analysis in figure 2, we estimated the life cycle GHG emissions embodied in single-cSi and multi-cSi solar PV panels for the U.S. market, as shown in figures 3 and S1 in the SI, respectively.Producing a 1 m 2 single-cSi solar panel for the U.S. results in 127.3 kg CO 2 e of GHG emissions,with 33.1,20.3,13.6,and 58.3 kg of emissions coming from polysilicon, wafer, solar PV cell, and PV panel production stages, respectively.The embodied emissions of a single-cSi solar PV panel are further broken down by emission source and production regions (figure 3(A)) and the corresponding raw material and energy input (figure 3(B)).
Emissions from production in China, APAC countries, the U.S., and Europe respectively account for 38%, 47%, 12%, and 1.3% of life cycle emissions for single-cSi solar PV panels sold in the U.S. market.Emissions from shipping, including ocean and on-road transportation, account for 1.6% of total embodied emissions.
China dominates polysilicon and silicon wafers production.China provides 63% and 82%, respectively, of the polysilicon and silicon wafer for PV panels used in the U.S., yet it is responsible for 74% and 85%, of the emissions at the polysilicon and silicon wafer production stage, respectively.Emissions from solar PV cell and PV panel manufacturing stages for the U.S. market are mainly from production in APAC countries.APAC countries account for 76% and 64% of the direct and indirect solar PV cell and PV panel consumption in the U.S. respectively, yet accounting for 73% and 71% of the solar PV cell and PV panel emissions, respectively.For the PV panel stage, domestic production in the U.S. represents 30% of market value, yet 22% of the emissions.U.S. production accounts for a minor share of emissions at other stages.
The U.S. tariff policy led to a shift of solar PV component supply from China to APAC countries, resulting in reduced embodied emissions.This study found that single-cSi PV panels produced in APAC countries have 12% lower embodied emissions than those produced in China.For PV panel production in APAC countries, emissions related to electricity consumption, aluminum, and glass inputs are 30%, 14%, and 6% lower than those in China.
In figure 3(B), emissions related to different energy and material inputs are shown.Emissions related to electricity used in solar PV manufacturing account for 45% of the total embodied GHG emissions.Aluminum and glass production contribute 25% and 12% of emissions, respectively.Other inputs contribute 16%, and transportation contributes the remaining emissions.Note that the electricity-related emissions here account only for emissions related to direct electricity consumption at each stage of silicon and solar PV manufacturing and do not include emissions related to upstream electricity use for raw material production.For instance, emissions related to electricity used for aluminum smelting are included in aluminum-related emissions rather than electricity-related emissions in figure 3(B).Silicon and solar PV manufacturing are electricity-intensive processes.Electricity-related emissions represent 78%, 84%, 81%, and 14% of emissions in the polysilicon, wafer, cell, and panel production stages, respectively.In the panel production stage, aluminum and glass contribute 49% and 26% of emissions, respectively.
Figure 3 tracks emissions sources along the supply chain of solar PV panels, and identifies the key regions and materials associated with major emission releases in the processes, thus providing utility for sustainable supply chain management to reduce lifecycle emissions.China is currently the major producer of polysilicon and solar PV products for the global market.Although the U.S. tariff policy has greatly impeded final product imports from China to the U.S. market, the upstream supply chain still depends on Chinese production.Compared to Europe and the U.S., China has relatively higher GHG emissions for electricity generation because of the high share of GHG-intensive coal usage.Moreover, China relies primarily on coal power for aluminum electrolysis, leading to higher embedded emissions for aluminum refining in China.The high share of Chinese production in the silicon solar PV supply chain increases the embodied GHG emissions of solar PV power.

Comparison of GHG emissions of solar PV manufacturing in the U.S. and China
Figure 4 compares the lifecycle GHG emissions for each production stage if the PV supply chain were to occur either entirely in the U.S. or entirely in China.The GHG emissions for polysilicon, single-cSi wafers, single-cSi PV cells, and single-cSi PV panels produced in the U.S. are 49%, 37%, 44%, and 40% lower than those in China, respectively.Overall, the life cycle GHG emissions of PV panels produced entirely in the U.S. are 42% lower than those in China, implying that completely shifting silicon and solar PV manufacturing from China to the U.S. can reduce the embodied GHG emissions in solar PV panels significantly.
As mentioned above, lifecycle GHG emissions of solar PV panels produced in China are higher than those in the U.S., mainly because of China's coaldominated electricity mix.Moreover, solar PV production in China is concentrated in a few provinces.For each production stage, production from the top five provinces accounts for more than 80% of total production in China and more than 60% of global production.Most provinces with large-scale silicon and solar PV manufacturing in China are located in the northwest-e.g.Xinjiang, Inner Mongolia, and Shanxi-and have an even higher share of coal power and GHG emissions intensity for electricity than the national average.This further increases the GHG emissions for silicon and solar PV manufacturing in China.Figure 5 shows the GHG intensity supply curve for polysilicon production in China.The top five provinces for polysilicon production in China are Xinjiang, Jiangsu, Inner Mongolia, Sichuan, and Shanxi.The share of coal power is up to ∼90% in Shanxi, Inner Mongolia, and Xinjiang; thus the production-weighted national average GHG emission intensity of polysilicon production there is 10% higher than the estimate if we simply assume the national average electricity generation mix is used for all polysilicon production in China.
It is worth noting that in the early years, China's solar PV production was mainly in the eastern provinces with developed economies and technologies and relatively less GHG-intensive electricity, such as Jiangsu, Zhejiang, etc, while the production capacity added in recent years is concentrated highly in the northwestern provinces with lower electricity and labor costs but a higher proportion of coal-fired power, such as Xinjiang, Inner Mongolia, Qinghai, Ningxia, etc (CNBS 2018, Gan et al 2021, IHS Markit 2022).This indicates that the migration of the silicon and PV supply chain within China recently has further increased the country's carbon footprint for silicon and solar PV manufacturing.
Other influencing factors aside, the geographic relocations of silicon and PV supply chain to regions with cheap coal-based electricity would reduce the cost of solar PV panels but increase the embodied GHG emissions of solar power, which runs counter to the original intent of developing solar PV technology for GHG emissions reduction.This trend reveals that achieving sustainable development in the silicon and PV industry might depend on more policy interventions to account for the cost of environmental and social externalities in the supply chain.

Embodied GHG emissions of solar electricity and potential emissions reduction
With these estimates of emissions from solar PV panel production, we further calculated the emissions associated with solar PV panel installation, including emissions embodied in solar PV panel mounting and supporting systems (e.g.inverter, electric components, etc), and evaluated the lifetime electricity generation of a solar PV system in the average solar irradiance conditions of the U.S., and thus estimated the embodied GHG emissions per kWh of solar electricity generated (black bar in figure 6).
With the current silicon and PV manufacturing supply chain, the average embodied GHG emissions of utility-scale solar electricity in the U.S. are 36 g CO 2 e kWh −1 , which are 3.7 and 3 times those of onshore and offshore wind power, respectively.The uncertainty estimates range from 26 to 52 g CO 2 e kWh −1 according to the location with varied average solar irradiation levels (see SI section 6 for the uncertainty estimate).With all-U.S. domestic supply, the embodied emissions of solar PVs can be reduced to 30 g kWh −1 (blue bar in figure 6, uncertainty ranging from 22 to 43 g kWh −1 ), while if the entire supply chain is in China the emissions would increase to 39 g kWh −1 (red bar in figure 6, uncertainty ranging from 29 to 56 g kWh −1 ), implying that moving solar PV manufacturing to the U.S. would help reduce GHG emissions embedded in solar electricity generation.
In the 2030 scenarios, due to grid electricity decarbonization (GED) caused by an increased share of renewable power, the embodied emissions of solar power are expected to decline by 11% and 6% in scenarios of entirely China or U.S. supply, respectively.GED has a stronger emission reduction effect on production in China because China currently has a GHG-intensive electricity grid and has set more aggressive electricity decarbonization goals for the future.The GED electricity grid mix projection here is developed based on the Stated Policies Scenario by the International Energy Agency (IEA 2022c).
As noted above, emissions associated with aluminum input account for a significant portion of embodied emissions in solar PV panels, so reducing upstream emissions from aluminum production would decrease embodied emissions of solar PV manufacturing.Increasing recycled content and using less GHG-intensive electricity for primary aluminum production are two major routes to aluminum sector decarbonization (ALD).According to our analysis, when coupled with GED, ALD further reduces solar power embodied GHG emissions by 12% and 6% for China and U.S. supply, respectively.In the ALD scenario, we assume 80% of aluminum is recycled and the rest is primary aluminum.Producing primary aluminum consumes more electricity than recycling aluminum, and most of the electricity for aluminum electrolysis is from captive power plants rather than the electricity grid, making the electricity mix for aluminum electrolysis different from the grid mix.In 2020, more than 80% of the electricity used for aluminum electrolysis in China is from captive coal-fired power plants.Phasing out captive coalfired power plants and using the less GHG-intensive grid power for aluminum electrolysis is considered a pathway for primary aluminum decarbonization in China and is assumed in the ALD scenario.However, the real situation does not suggest a decarbonizing trend.According to statistics from the International Aluminum Institute (IAI 2022), the proportion of coal power in the electricity mix for aluminum electrolysis in China rose from 80% in 2020% to 85% in 2021, making the prospects for primary aluminum decarbonization in China unclear.In contrast, the proportion of hydropower in the electricity mix for aluminum electrolysis in North America kept increasing in the past decade and reached 95% in 2021 (IAI 2022).Therefore, although we estimate higher potential emission reductions for solar PV manufacturing in China than in the U.S. in the ALD scenario, this will only be the case if the aluminum electrolysis sectors do, in fact, decarbonize.
Finally, improving PV conversion efficiency can increase the electricity yield per m 2 of solar PV panels and thus reduce the embodied emissions of solar electricity.The PV conversion efficiency of commercial solar panels is assumed as ∼20% in our estimate of 2020 status (Fraunhofer ISE 2022).PV conversion efficiency has been increasing and is expected to continue to improve in the future.The lab record of PV conversion efficiency has reached 26.7% for single-cSi technology and 24.4% for multi-cSi technology (Fraunhofer ISE 2022).Here, we created three scenarios of increased PV conversion efficiency (ICE) in 2030.When added to GED and ALD, increasing PV conversion efficiency to 23%, 25%, and 27% is estimated to decrease the embodied GHG emissions by 13%, 20%, and 26%, respectively.
Besides the future scenario discussed above, many other sustainable designs and technologies that are under development might further reduce the embodied GHG emissions of solar PV systems, for instance, the application of recyclable and biodegradable polymeric materials (Fiandra et al 2019), reduced usage of rare earth metal materials (Helbig et al 2016, Pavel et al 2017), and the use of novel PV materials such as quantum dot photovoltaic (QDPV) (Deng et al 2019).According to the study by (Engül and Theis 2011), the QDPV technology can reduce the embodied emissions of solar power to 5 g CO 2 e kWh −1 , which is much lower than the emission level of current PV technology.More in-depth studies and LCA modeling of these advanced technologies are required to explore the future GHG emissions reduction potential of solar PV systems.
As highlighted in the Introduction section, this study's primary originality lies in the comprehensive analysis of the U.S. specific supply chain of solar PV systems, considering diverse upstream emissions and production conditions associated with energy/material inputs (i.e.electricity, aluminum, and glass) from different regions throughout the PV manufacturing life cycle.By conducting a thorough supply chain analysis and incorporating regional variations in emissions associated with different solar PV components, this LCA study aims to generate more representative results for the U.S. compared to other previous studies that overlooked the entire supply chain and regional influences.Furthermore, by analyzing the emissions contributions by different regions at each production stage of PV manufacturing, the results of the present study will help identify emission hotspots at each production stage of PV manufacturing, facilitating a deeper understanding of GHG emissions reduction possibilities.It also enables the projection of future emission trends under distinct regional policies and production scenarios, including quantifying the potential impacts of geographic relocation of silicon and PV manufacturing.

Conclusions
Solar PV electricity is considered an important source to achieve carbon neutrality in the electric sector.While low, embodied GHG emissions of solar PV electricity are non-trivial, especially when a large amount of solar electricity is used for downstream clean technologies such as clean hydrogen and synthetic fuel production.To quantify and further explore emissions reduction potentials for solar electricity generation in the U.S., we analyzed the silicon and solar PV supply chain for the U.S. market to conduct LCA of solar electricity with the current supply chain for the U.S. market.We then compared that to a U.S. domestic supply to examine the decarbonization potentials of solar electricity.
Results show that with the current supply chain, the US solar PVs have embodied GHG emissions of 36 g kWh −1 of electricity generated, of which, 38% and 47% are from processes in China and APAC countries, respectively.From an input materials and energy point of view, 45% of the embodied emissions of solar PV panels are associated with the electricity consumed during silicon and PV manufacturing, while 25% and 12% of the emissions are from emissions of aluminum and glass production, respectively.In the past decade, the silicon and PV manufacturing supply chain has shifted to China and APAC countries, which has increased the embodied GHG emissions of solar electricity generation in the U.S.This is primarily due to the high share of coal-based energy for both the grid electricity mix and the electricity mix for aluminum electrolysis in China and the main APAC production countries.In contrast, the U.S. has a less GHG-intensive electricity grid, and hydropower is used in more than 95% of North American aluminum electrolysis facilities.If the U.S. establishes its own supply chain for solar PV, the embodied GHG emissions of solar electricity generation would be 16% lower than the current supply chain.
Our future analysis shows that in 2030, shifting the supply chain to the U.S., decarbonizing both grid electricity and aluminum production, and improving PV conversion efficiency will help reduce embodied GHG emissions of U.S. solar electricity to 21 g kWh −1 , which is 43% lower than the current level.
The present study focuses on generating thorough estimates of embodied GHG emissions for the U.S. solar electricity to explore emission reduction opportunities.However, embodied GHG emission reductions are one of the considerations in solar PV production and commercialization.The implementation of solar PV manufacturing and supply chains is influenced by other factors such as the establishment of a domestic supply chain, and national and regional economic benefits.For instance, lower labor wages and electricity costs in countries like China and other Asia-Pacific nations made them more appealing for polysilicon and PV manufacturing historically.Analysis of these factors is out of the scope of this study.A more comprehensive exploration of these factors enables a nuanced evaluation of the geographic shift of solar PV supply chains.

Figure 1 .
Figure1.The system boundary of LCA of the cSi solar PV system (transportation activities, not shown in the figure, are also included in the system boundary).The LCA analysis includes the production of metallurgical grade silicon, solar grade polysilicon, single-and multi-cSi ingots and wafers, single-and multi-cSi cells and panels, and other relevant materials and energy, solar PV system installation, and solar PV system's end-of-life treatment.

Figure 2 .
Figure 2. The supply chain of solar PV panels used in the U.S. 2017-2021.The countries of origin for production are clustered into five groups: China (CN), Europe (EU), Asia-Pacific countries (APAC, China excluded), the U.S., and the rest of the world (ROW).

Figure 3 .
Figure 3. Emission contribution of the supply chain to single-cSi PV panel in the U.S. market (A) by emission source region and (B) by material/energy input.Emissions contributions of different regional production include both direct and indirect contributions through the supply chain network of U.S. solar PV systems shown in figure 2.

Y
Figure 4. Embodied GHG emissions for single-cSi PV manufacturing in the U.S. and China in 2020.

Figure 5 .
Figure 5. Provincial GHG intensity supply curve for polysilicon production in China.See table S4 in SI for the full list of provinces.SI figures S2-S7 show the GHG emissions intensity supply curves for multi-and single-cSi wafers, PV cells, and PV panels production in China.The provincial GHG emissions intensity shown here varies by the electricity mix of the corresponding province.The data on electricity mix, polysilicon, and PV production shares by province in China are from literature (CEC 2018, CNBS 2018, Gan et al 2021, IHS Markit 2022), and presented in table S4 in SI.

Figure 6 .
Figure 6.Current and potential changes in embodied GHG emissions of solar electricity in the U.S. in different future scenarios.For comparison, the figure presents estimates of embodied GHG emissions for the other renewable power alternative-wind power (Michael UNECE 2021, Wang 2022, Wang et al 2022)-as well as the embodied emissions of solar power assuming solar panels are entirely produced in the U.S. or entirely imported from China.Different future scenarios: GED: grid electricity decarbonization, ALD: aluminum decarbonization, ICE: improved PV conversion efficiency.The PV conversion efficiency in ICE Scenarios 1, 2, and 3 is assumed to be 23%, 25%, and 27%, respectively.This figure only presents results for utility-scale open-ground-mounted PV systems.The error bars show the uncertainty estimates of the results.See SI section 7 for the method of uncertainty estimates.Different mounting types, for instance, rooftop or open ground mounting would affect the solar irradiation received by the PV panel and manufacturing materials usage and thus affect the embodied emissions of solar electricity.See the SI section Solar PV mounting and supporting systems for more details about analyses of different mounting systems.