Identifying methods to reduce emission intensity of centralised Photovoltaic deployment for net zero by 2050: Life cycle assessment case study of a 30 MW PV plant

Photovoltaics (PV) is one of the most effective and necessary energy sources to mitigate climate change. The broad electrification scenario projects the PV market to grow from 1 TW in 2022 to over 63 TW in 2050. While increasing PV production will significantly reduce the emission intensity of electricity generation, it is still important to minimise the overall environmental impact of such a large industry. In this study, we investigated the intensity of greenhouse gas (GHG) emissions of a 30 MW PV plant using a life cycle assessment (LCA). Based on the LCA, we propose a roadmap to reduce emissions from PV manufacturing and deployment. Decarbonising significant factors like aluminium and concrete production or the electricity demand to produce PV modules can greatly reduce the carbon budget related to PV production. Our study shows that the global warming potential (GWP) per kWh can be reduced from 11.2 to 1.7 g CO2‐eq/kWh over the lifetime of the PV system (85% reduction). Using the aspects to decarbonise PV production, the roadmap is demonstrated. The cumulative GWP to reach 63 TW is initially estimated to be 44 Gt CO2‐eq. Our decarbonising roadmap demonstrated that such significance can be reduced by over 37 Gt CO2‐eq, equivalent to a whole year emission in year 2022.


| INTRODUCTION
The world is facing one of the most unprecedented events.Recently, more frequent and severe weather events are occurring due to climate change.The increasing intensity of hot and cold climate incidents has led to global consequences, including food shortages, drought, bushfires, and tornadoes.The world is also expected to reach a global temperature 1.5 C above that of the pre-industrial period of 1850 within the coming decade, despite whether the pledged scenario by all developed nations is reached by 2030 [1].Therefore, we need a more urgent, effective, and immediate approach to mitigate climate change.
Increased greenhouse gas (GHG) emissions, which are primarily from human activities, are the cause of climate change.Cumulative Moonyong Kim and Storm Drury contributed equally to this study.
GHG emissions in the atmosphere from 1751 to 2019 are estimated to be 1615 Gt, where 51.5% was emitted in the past 40 years, correlating with the beginning of noticeable changes in global temperature [2].The remaining carbon budget for the 1.5 C global warming scenario is estimated to be 500 billion metric tons CO 2 -eq [3].Total annual GHG emissions in 2021 were 36.3Gt CO 2 -eq [4], representing the highest ever annual increase after a slight reduction during the start of the pandemic in 2020.During the first year of the pandemic in 2020, GHG emissions were only reduced by 6% compared with the previous year despite the many travel restrictions around the world, implying that we need a more fundamental rectification of emission sources used to reduce global GHG emissions.One of the primary GHG sources is electricity generation.The emissions intensity from the worldwide power system is approximately 0.436 kg CO 2 -eq/kWh based on last year's prediction [5], where most electricity generation is based on emission-intensive sources like fossil fuels.As most power generation is fossil fuel-based, the emissions can be between 0.49 and 0.82 kg CO 2 -eq/kWh based on gas or coal power plants, respectively [6].Considering the world electricity demand was as high as 26.4 PWh last year in 2021 alone [7], less carbon-intensive power generation methods are needed to reduce global GHG emissions and mitigate climate change.
One of the most established approaches to mitigate climate change is using renewable energy like hydropower, photovoltaic (PV), wind, and geothermal.PV is one of the most available and abundant renewable power sources.It also has emissions as low as 0.018 kg CO 2 -eq/kWh over its lifetime [6], effectively a factor of 22 and 55 lower than gas and coal power plants, respectively.PV contributed about 3-5% of total electricity, where about 250 GW of PV was installed in 2022 [8].PV installations have been growing rapidly at 20-30% annually, and this growth is expected to continue, giving hope to a scenario of slowing or reducing global annual GHG emissions [9].While many scenarios had projected different PV market scales from 2 to over 60 TW by 2050, the most ambitious scenario must be considered for rapid transition.As the world is already expected to reach well over 1.5 C above the pre-industrial period, a conservative scenario is no longer adequate to combat climate change.One of the most aggressive scenarios is the broad electrification scenario from the International Technology Roadmap for Photovoltaic (ITRPV) [10], originally by Ram et al. [11], which estimated cumulative PV installed capacity in 2030 and 2050 to be 10 and 63 TW, respectively.Such a scenario projected a 100% global electrification scenario, one of the most promising scenarios to solve the climate change problem.The potential carbon budget savings from replacing the electricity source with PV can be calculated (see Supporting

Information S1).
Despite the potential significant GHG emissions reduction achievable through using PV, such rapid expansion can introduce new challenges, including material sustainability and emission intensity from the PV manufacturing process.As a transition from fossil fuel-based power sources to PV occurs, the type and the amount of materials required also shift.One study suggested that over 20 Mt of copper (Cu) is required to meet 8.5 TW of PV only [12], which would correspond to over 148 Mt of Cu for a broad electrification scenario.While the main issue with using more materials is a scarcity of rare materials, such as silver (Ag), indium (In), and bismuth (Bi), other abundant materials are still critical to consider and reduce.Aluminium (Al) and Cu, as examples, mainly rely on a secondary production source for use in PV, which has a significantly lower global warming potential than the primary source by a factor of 22 [13] and 8 [14], respectively.Therefore, the surge in material demand from the PV industry may unintentionally amplify the environmental impact of these materials.Many stakeholders are already taking action to reduce the embodied carbon in PV systems [15,16].This paper aims to determine an emissions reduction roadmap for deploying PV utilities as we head towards TW annual production through a range of strategies to reduce both material consumption (kt/GW) and the associated carbon dioxide equivalent (CO 2 -eq) emission intensity (Gt CO 2 -eq/TW of PV produced).First, we establish approximate ranges for the consumption of key materials required for deploying centralised PV systems based on publicly available environmental production declaration (EPD) of a 30 MW centralised PV plant [17].The environmental impact was quantified in global warming potential (GWP).We then use LCA to determine the required material, electricity demands, and GWP.This is followed by approximate estimates for the emissions generated by key materials required for PV based on known emissions intensities.
Moreover, other carbon-and energy-intensive processes like transportation, refinery, and manufacturing processes are considered, which may contribute to significant emissions from the production.
Finally, we propose a roadmap to reduce the emission intensity of PV for the TW scale.

| Life cycle assessment
Life cycle assessment (LCA) is a methodology used to evaluate the potential environmental impacts of products or services throughout their entire life cycle, with a "cradle-to-grave" approach based on International Organization for Standardization (ISO) 14040 and 14044 [18,19].The LCA can be used to assess the environmental impacts of the products and the processes.This method is advantageous for evaluating energy production systems because the procedure measures all direct and indirect energy consumed.The indirect component comes from the electricity, which depends on the emission intensity of electricity (kg CO 2 -eq/kWh).The LCA method consists of four main phases: the goal and scope definition, the life cycle inventory analysis, the life cycle impact assessment, and the interpretation steps [9,11].The values can then be used to inform the public sector, stakeholders, and manufacturers.Our study focusses on the global warming potential of PV power plants using a 30 MW PV plant case study.
The LCA result is then used to identify the scope to reduce the emissions.
2.2 | The goal, scope, and functional unit

| Goal
This LCA seeks to determine the cumulative environmental impacts of the production of solar power plant systems, including silicon wafers, modules, inverters, and mounting systems.An LCA for a 30 MW PV plant as a case study is used to identify strategies to reduce the emission intensity of centralised PV deployment and highlight the benefits of a sustainable and circular PV economy as we approach TW-scale production.

| Scope
The environmental impact the study focusses on is the global warming potential (GWP) in carbon dioxide equivalent (CO 2 -eq).The stages considered are the manufacturing and mounting/installation processes.Life cycle assessments (LCAs) [17,20] have identified current methods for the following materials for each PV system component: • Solar-grade (SOG) silicon wafers for silicon wafers and cells; • Silver for solar cell metallisation contacts; • Aluminium for framing; • Stainless steel and concrete for reinforced mounting; • High-purity glass for module; and • Copper for interconnection wires and inverters.
Some key assumptions for the system prior to implementing strategies to reduce emissions include the following: • The emissions are based on material and manufacturing phases.
Any additional emissions from transportation, operation, and end of life are not considered.
• The use of 100% primary Al emphasises the importance of accounting for larger contributions of primary Al as we approach TW-scale production [13].While a large fraction of Al mainly comes from secondary sources, it may not be sufficient to meet the rapidly increased material demand.
• Electricity consumption for the key processes of Al, steel, and silicon wafer production follows China's average energy profile, mainly consisting of fossil fuels at 0.531 kg CO 2 -eq/kWh [5].Such values are higher than the global emission intensity of 0.436 kg CO 2 -eq/kWh [5], but the value can still provide a larger estimation of GHG emissions.Moreover, material production predominantly occurs in China, so China's energy profile is the most representative value.
• The emissions and the electricity demand for silicon cells are based on the production of all production, including silica sand, quartz, metallurgical silicon (MG-Si), poly-silicon (poly-Si), Czochralski mono-crystalline silicon (Cz-Si) ingot, and cell production that is involved in making the cells.The value does not include emissions or electricity demand for silver production.

| Electricity generation over the lifetime
The electricity generation of the 30 MW PV plant is calculated using the following equations: where deg is the annual linear degradation rate of modules, n is the nth year of the PV system operation, E ideal is ideal annual production from the PV system without degradation, P system is the PV system power, P module is the PV module power in kW, N mode is a number of modules, LT is the lifetime of the PV system, and Irr is the average daily irradiation.Table 1 shows the input values and the references used in this study.The total module areas are 1.45-1.56Â 10 5 m 2 .
Based on the assumptions and the input values, the total electricity generation is estimated to be 1.51 Â 10 9 kWh or 1.51 TWh over a lifetime of 30 years.

| Life cycle inventory data
Life cycle inventory (LCI) data were collected from various sources.
The foundational database was sourced from EcoInvent [25] and the International Energy Agency [26].Module data and system quantities were provided by EPDs from both Trina and Jinko Solar [21,22] [17, 20], which were in consumption per area (CPA).Therefore, to calculate consumption per unit of power, CPP, the following equation is used: where i is module number (1 or 2).
The values of calculated CPA can be found in Table S1.The table also contains the values of other materials, like ethylene vinyl acetate (EVA) and glass, required to produce PV modules.EVA is the most used material for encapsulants [8].Other encapsulants that contain polyolefin are also commonly used and projected to dominate the market [8] due to their advantage of material endurability and stability [27].The polyolefin-based materials, however, are known to have higher energy and GHG emissions [28,29], which can be minimised by creating a circular economy.However, this study does not consider the scope of alternative material for encapsulants.The electricity to produce 1 kg of each material is summarised in Table 2.The consumption per power for mounting systems, cables, and inverters is calculated from the literature [13,21,22,25,26].For silicon wafers, the values of electricity demand are based on the literature [23], which combines the electricity demand for all processes from quartz silica to silicon wafers.Only the electricity demand values are shown from cell production and module fabrication.Concrete has a wide range from 0 to 263 kt/GW.The usage varies significantly with the ground condition and the type of mounting systems.The midvalue of 131.5 kt/GW is used for this study.Note that the usage of concrete depends heavily on the ground condition and the mounting system types.The electricity consumption for silicon wafers is the sum of all processes required to make silicon from quartz.The processes include metallurgical silicon production, poly-silicon purification, ingot growing, and wafer cutting.The silver consumption is based on the usage of different cell technologies such as passivated emitter and rear contact (PERC), tunning oxide passivating contact (TOPCon), and silicon heterojunction (SHJ) in ITRPV [8].This was done to highlight the different ranges of silver usage where the n-type technology has more sustainable problems with greater silver usage [32].
Electricity generation can be sourced from different carbonintensive sources.The carbon emissions in GWP per electricity are summarised in Table 3.The emission intensity of electricity from PV is around 0.018 kg CO 2 -eq/kWh, which is expected to be dependent on the electricity mix [35].

| Emissions and energy intensity per weight
Other than the emissions from the electricity generation to meet electricity demand, there are emissions directly from producing the materials.Table 4 shows the GWP values used in this study, where the emission intensity of silver was the highest.Other materials like copper and glass had relatively lower GWP per kg of materials.The carbon intensity of aluminium is based on primary aluminium, which is significantly higher than that of secondary [13].The emission intensity depends significantly on the location of the production with different T A B L E 1 Key system assumptions and input values for the 30 MW PV plant [21,22].Average in China 0.531 [5]-calculated based on annual emissions and the electricity electricity mixes, whereas decarbonising electricity is expected to lower the intensity significantly more [42].
Concrete and steel can be reduced using a greener approach of geopolymer concrete (up to 90% reduction) [38] and "green" steel involves using waste CO 2 as a carbon source to make steel with iron or using green hydrogen which would correspond to 0.08 kg CO 2 -eq/ kg [40,41].However, the value for steel depends heavily on the emissions intensity of the electricity as the green steel process requires about five times more electricity demand per kg of steel (see Supporting Information S1) [31].

| Material consumption and the sustainability of PV production
Material consumption to build a 30 MW PV power plant is calculated.
Using the consumption per power from Table 2, Figure 1 shows the total material amount from each material in the log scale.This was done to compare the different material demands in several orders of difference, like concrete to silver.The consumption per power, which was normalised with 30 MW of the system, is also shown in the second y-axis.The concrete amount showed the highest amount 63%, while the percentage of silver was below 0.01%.Other materials like glass and steel are also heavily required, known to be produced using high energy-intensive processes.
More importantly, the values need to be compared with the available material supply.As the PV market expands, the surge in material demand can lead to a material shortage.As PV production is estimated to be over 1 TW annually, the current technology may not be sustainable without significant silver reduction.Companies have successfully reduced silver consumption over the years, but it is still not sufficient compared with the rapid rate of PV production.

| Electricity
As PV module production processes primarily require a significant amount of electricity, one of the main factors for GHG emissions, reducing electricity consumption will be necessary.The electricity demand per kg of material for each component is shown in Figure 2a.
Silicon and silver are the two main materials that require the most electricity per kg.However, as different materials are required for every power generation of PV production, as shown in the previous section, the electricity demand should be considered based on the PV production scale.For example, the electricity demand for last year's PV production of 30 MW from each material is estimated in report [44].These recent values reported this year appear to be

| Global warming potential (GWP)
The material consumption and electricity demand for each material were discussed.Now, each material's GWP is investigated using LCA.The GWP from each material source is assessed using LCA with support from previous studies [21,22,25,34,41].Based on the database of OpenLCA, the GWP values per kg of material are calculated (see Table 4).Using the CPP from Table 2 and the PV production of 30 MW, Figure 3 shows the extrapolated GWP values over the estimated material demand and the GWP per material weight.The values include the emissions from electricity usage (see Figure 2b) with the assumed emission intensity of electricity from China.While silver has the highest GWP per material, the amount used is at least two orders of magnitude lower than other materials, such that the GWP from the silver was insignificant.The main contributor of GWP is concrete, which is only required for the utility scale.
The emissions also come from cell production and module fabrication processes.Based on the electricity demand for 30 MW and the emission intensity of electricity, the estimated emissions from cell production and module fabrication are 0.84 and 0.22 kt of CO 2 -eq, respectively (not shown in the figure).3.4 | Roadmap to reduce material consumption and environmental impact

| Aspects of reducing environmental impacts
There are major aspects to reducing environmental impacts from PV production.Using a sensitivity analysis, we have identified methods to reduce the emissions by • Decarbonising the electricity usage for material production; • Replacing conventional steel and concrete with "green" materials using waste from other sectors, including ground glass and waste CO 2 [45]; • Reducing material consumption; and • Improving the module efficiency.

| Roadmap to reduce global warming potential from PV production
Previous sections investigated material consumption, electricity demands, and global warming potentials from different materials.While the LCA was based on a small scale, such results can be used to extrapolate the emissions in a larger scale of PV production.Moreover, they can be assessed to propose a roadmap for PV towards environmental impact reduction.We use LCA to identify the environmental impact of manufacturing and deployment for PV utilities in detail with a case study of a 30 MW PV plant to identify key areas of carbon intensity from cradle to grave, with an estimated emission intensity of ⁓11.2 g CO 2 -eq/kWh [22,34,46] generated over the system life.Such value depends on module efficiency and the type of modules (bifacial or monofacial).The value also includes the emissions from concrete where the usage of concrete depends on the ground condition and the mounting systems.When the concrete usage is not considered, then about 9 g CO 2 -eq/kWh is estimated.These values are lower than other LCA studies.One of the main reasons could be the significant improvement in electricity consumption for cell and module fabrication over the years (see Section 3.2).Other factors could also depend on the location and the type of mounting system, where the yield can vary with annual irradiance and the annual yields, especially with single-or dual-axis tracking systems.Using the aspects to reduce the environmental impacts (see the previous section), a pathway to reduce emissions can be seen in Figure 4.The first approach is by decarbonising electricity.There is viability in benefitting by decarbonisation from the circularity of PV to fabricate energy-intensive components via decarbonised electricity and green steel and concrete, using waste materials from other industries, such as ground glass and waste CO 2 [45].Decarbonising the electricity with the emission intensity of 18 g CO 2 -eq/ kWh reduced the emissions down to 5.34 g CO 2 -eq/kWh.
Green approaches for steel and concrete are the next step to reducing the emission intensity.Green steel is typically produced using green hydrogen as the fuel source, which can reduce emissions down to 0.08 kg CO 2 -eq/kg.For the concrete, replacing traditional cement with recycled aggregate can reduce the emission intensity to 0.087 kg CO 2 -eq/kg of concrete [31], equivalent to only 0.23 g CO 2eq/kWh of a PV system.
Lastly, the most significant emission reduction can come from module improvement, expected to improve to 30% efficiency, which is in line with the "30  the tandem structure has recently achieved over 30% efficiency, such performance may be viable [47].The summary of the roadmap can be found in Figure 4.

| Emission reduction from PV production based on the roadmap
Based on the roadmap for reducing emissions in Figure 4 and the broad electrification scenario from Ram et al. [11], the emissions emissions of 0.7 Gt CO 2 -eq/TW.Decarbonising the electricity reduced the emissions per kWh of the system by $6.2 g CO 2 -eq/ kWh (see Figure 4), equivalent to over 23 Gt CO 2 -eq cumulative.
Green options additionally reduced the emissions by 11.3 Gt CO 2eq.Lastly, combining the efficiency boost to 30% with other decarbonisation options can result in only 6.8 Gt CO 2 -eq cumulative, which can prevent over 37 Gt CO 2 -eq emissions.Considering that such module efficiency improvement can also directly reduce the overall levelised cost of energy (LCOE), the result demonstrates another benefit of using better module efficiency.The roadmap for more affordable modules through cheaper electricity costs will be investigated in future study.
The projected cumulative global warming potential values presented here I have a few implications.First, the GWP value required to reach broad electrification is comparably smaller than the cumulative global CO 2 emissions in the coming decades with 30-40 Gt CO 2eq/year.Without more immediate and effective decarbonisation, the CO 2 will far exceed the current carbon budget, which is already too high.Decarbonising electricity will also naturally reduce the carbon footprint of other industry sectors, especially with high electricity demand.Second, the carbon emissions from PV manufacturing are expected to be reduced through improving module efficiency, reducing material consumption, using green approaches, decarbonising processing, and the electricity required.Therefore, future studies should focus on sustaining the rapid PV production rate to maximise the electricity mix with PV to reduce the emission intensity of electricity (Figure 5).

| CONCLUSION
This study performed a life cycle assessment of a 30 MW PV plant to assess material and electricity demand and the global warming potential required.The LCA was then used to identify strategies for reducing the emission intensity of centralised PV deployment and outline the benefits of a circulatory economy of PV towards broad electrification.High energy consumption resulting in indirect emissions released from the manufacturing and deployment of centralised PV modules and BoS are a cause for concern as we progress towards TW-scale PV production by 2030.However, our results show cumulative emission intensity can be reduced by 85% (from 11.18 to 1.71 g CO 2 -eq/kWh) with all our proposed strategies of emission reduction implemented, which can reserve over $37 Gt of CO 2 -eq.While such reduction may or may not be significant compared with the emissions that can be annually saved by PV power, decarbonising PV production should still be considered to minimise greenhouse gas emissions to mitigate climate change as effectively as possible.

Figure 2b .
Figure 2b.While silver had the second highest electricity demand per kg (see Figure 2a), it only contributed 0.3% of the total.Aluminium and silicon, in comparison, had the highest electricity demands, where they contributed over 93% of the total electricity demand.The result indicates that the two are the most energy-intensive materials for PV production.Additionally, the electricity demands for cell (p-type PERC) and module fabrications (53 and 13.5 MWh/MW, respectively) are from the China Photovoltaic Industry Association (CPIA)

Note:
Silicon emissions are based on direction emissions and indirection emissions with the emissions intensity of China.F I G U R E 1 The material consumption for 30 MW of PV modules from different components using the OpenLCA values and LCA scenario.The percentage indicates the total percentage of the weight with respect to the total module weight.Note the y-axis has a break at consumption = 1.6 kt.significantly lower than the values reported from the IEA PVPS report in 2020 (60 and 70 MWh/MW, respectively) [23].Such difference demonstrates such significant change in recent years in the PV industry.

F I G U R E 2
Electricity demand (a) per weight of materials or area and (b) for 30 MW of PV plant.Cell and module in (b) indicate the electricity demands from cell production and module fabrication, respectively.

F I G U R E 3
Estimated global warming potential (GWP) for 30 MW PV power plant as a function of the total material consumption and GWP per material.
that are required to reach a broad electrification scenario is demonstrated.The calculation for the electricity generation can be found in Figure S1a.The electricity generation from the installed PV was estimated using the broad electrification scenario and the irradiation average of 1.5 kWh/Wp annually.A large range of irradiation was used to consider the installation worldwide.The lifetime of installed PV systems is expected to be 30 years with 0.7%/year linear degradation.The cumulative emissions from PV production to reach broad electrification add up to $44 Gt CO 2 -eq by 2050 for an additional 60 TW of PV production and installation, with equivalent F I G U R E 4 A roadmap to reduce global warming potential per electricity generation (in g CO 2 -eq/kWh) for a 30 MW PV plant.F I G U R E 5 Cumulative global warming potential from PV installation using the roadmap values.
The electricity required to produce 1 kg of material, assumptions, and reference(s).