Deploying solar photovoltaic energy first in carbon-intensive regions brings gigatons more carbon mitigations to 2060

The study addresses how the spatial and temporal deployment of the solar PV industrial chain—from manufacturing to installation—affects life-cycle greenhouse gas (GHG) mitigation. Despite rapid PV cost declines and capacity growth, prior analyses have under-examined how evolving geographical specialization, varying grid emission factors, solar resources, and technology improvements jointly shape both emissions from PV manufacturing and avoided emissions from PV generation. The research aims to quantify historical (2009–2019) and prospective (2020–2060) net GHG mitigation at country level using a spatial-temporal LCA, and to assess how alternative manufacturing and installation strategies can maximize global net mitigation while supporting carbon neutrality targets.

Prior work has used LCA and emission inventories to estimate PV-related GHG emissions or mitigation in specific countries (notably China) and globally. However, existing studies often lack: (1) explicit spatiotemporal characterization of both mitigation and emission intensities across the full industrial chain; (2) integration of evolving factors such as manufacturing energy intensity, conversion efficiency, grid decarbonization, and solar resource variation; and (3) scenario-based projections that reflect uncertainties in industrial deployment patterns. The paper positions itself to fill these gaps with a unified, spatialized, dynamic LCA framework comparable across countries and times.

The study employs a cradle-to-cradle, spatialized, dynamic life-cycle assessment (LCA) of crystalline silicon PV (mono- and multi-/polysilicon) covering manufacturing (metallurgical silicon to modules and balance-of-system), assembly, operation, transportation (international and domestic), and end-of-life (disposal and recycling). Silicon ore mining is excluded due to negligible contributions. Emissions are quantified using 100-year GWP (CML 2001), with a 1 kW system as the functional unit under standard test conditions. Base life-cycle inventories come from IEA PVPS Task 12 and literature, implemented in GaBi (v10.6.1.35; DB 2023.1). The model dynamically characterizes technical parameters (electricity/steam use, silicon use efficiency, solar-to-power conversion efficiency) and spatializes emissions via national power grid emission factors. Historical analysis (2009–2019): Annual emissions are computed for each supply-chain step and country via activity data (IEA PV trends) multiplied by country- and step-specific emission intensities. Mitigation is estimated as avoided grid emissions based on installed capacity, country-specific capacity factors, and grid emission factors (WIND database for capacity; capacity factors per Lu et al.). Net mitigation equals mitigation minus emissions. Future projections (2020–2060): Nine scenarios combine three installation patterns and three manufacturing patterns: Installation scenarios—C1 (lock-in current national distribution), C2 (equitable shares across nations), C3 (targeted to highest grid emission factors). Manufacturing scenarios—M1 (concentration in China: 90% by 2030), M2 (relocation to Asia-Pacific outside China: 90% split among India, Vietnam, Malaysia, South Korea, Philippines), M3 (relocation to Europe and North America: 90% across USA, Germany, Italy, France, Spain). New manufacturing distributions are established by 2030 and held constant thereafter; PV generation capped at 70% of national electricity demand and within national technical PV potential. Power system emissions: The grid emission factor combines non-PV sources and PV (PV set to zero for operational EF to align with IEA accounting). Three non-PV decarbonization trajectories (optimistic, moderate, conservative) reflect national neutrality targets (non-PV EF reaches net-zero 10, 5, or 0 years earlier than national neutrality date; otherwise annual EF declines of 1%, 0.5%, 0.1%). Power demand is downscaled by nation from IEA regional projections using GDP and population (SSP1). PV generation equals installed capacity times capacity factor times 8760 h. C3 allocation uses an iterative algorithm prioritizing countries with highest marginal mitigation intensity (driven by current grid EF), respecting the 70% PV share and technical potential constraints, and updating grid EFs after each allocation step. End-of-life assumes recycling rates increasing from 10% (2020) to 100% (2045). Transportation emissions are included. Sensitivity/uncertainty ranges reflect grid EF scenarios and PV utilization hours (1000–2000 h for intensity tabulations).

• Historical (2009–2019): Global net GHG mitigation from PV was 1.288–1.289 Gt CO2e, achieved by 1.9688 Gt mitigation minus 0.6803 Gt manufacturing and related emissions. Regional shares of cumulative net mitigation: Asia-Pacific 51.5%, Europe 29.8%, North America 13.4%; others 5.3%. Top five national contributors: China (266.3 Mt), Japan (198.2 Mt), USA (167.9 Mt), Germany (157.3 Mt), India (144.0 Mt), together 72.5% of global total. Some manufacturing-heavy, installation-light countries (e.g., South Korea, Malaysia, Singapore) had negative net mitigation totaling −42.4 Mt by 2019.
• Mitigation intensity (kg CO2e per kW per year): Highest in the Middle East (avg. 2009–2019: 1293.9 kg CO2e kW−1 a−1); Africa and Asia-Pacific also high. Global average mitigation intensity declined 11.1% from 2009 to 2019, with pronounced declines in Europe (−24.9% overall; e.g., Denmark −75%). China fell 22.3% (1261.3 to 980.4 kg CO2e kW−1 a−1).
• Emission intensity (kg CO2e per kW functional unit): Global average fell 74.7% from 2009 to 2019. China declined by 2859.5 kg CO2e kW−1 (−79.2%), driven by improved conversion efficiency and reduced consumption of silicon, electricity, and heat, and lower grid EF; process contributions to China’s decline: polysilicon 58.4%, wafer 27.8%, cell 3.0%, system 10.8%.
• Future mitigation (2020–2060): Under installation scenarios, cumulative mitigation reaches 116.1 Gt CO2e (C1; range 86.5–145.8), 129.0 Gt (C2; 100.7–156.6), and 209.3 Gt (C3; 171.7–244.7). Moving from C1 to C3 raises mitigation by 92.1 Gt (≈1.8–1.9× 2020 global GHG emissions). Regions gaining under C3 include Asia-Pacific (+55.2 Gt), Middle East (+21.9 Gt), and Africa (+10.0 Gt), with large contributions from India, Indonesia, Russia, Saudi Arabia, Iran, and UAE.
• Future emissions (2020–2060): Cumulative life-cycle emissions differ primarily by manufacturing location: M3 (Europe & North America) 4.6 Gt (4.4–4.8), M1 (China) 6.2 Gt (5.9–6.5), M2 (Asia-Pacific ex-China) 8.9 Gt (8.4–9.2). Transportation adds ~0.45 Gt on average across scenarios. Recycling increasing to 100% by 2045 lowers emissions by 5.3–13.8% vs. landfill-only.
• Life-cycle intensity in 2030: 3.9–25.1 g CO2e kWh−1 across manufacturing–installation scenarios (driven mainly by manufacturing location and PV utilization hours).
• Power system effects of targeted deployment: Compared to C1, C3 reduces the global average grid emission factor by 17.6% (2030), 59.3% (2045), and 70.4% (2060), and narrows spatial EF gaps (e.g., by 34.8% in 2030). This convergence lowers cross-country manufacturing intensity differences and potential trade frictions (e.g., under CBAM-like policies).
• Carbon Payback Time (CPT): China’s CPT declined from 2.89 years (2009) to 0.83 years (2019). In 2030, CPT is shortest under M3 (0.77 year) and longest under M2 (1.30 years).
• Storage implications (if paired from 2030): Cumulative GHG from storage by 2060 could be 5.2–7.1 Gt (Li-ion) or 5.9–8.2 Gt (VRFB). Other environmental burdens from storage range from 0.1–1.3× those of the PV industry, and storage may increase PV investment by 9.4–33.1% depending on costs and deployment.
• Land use: PV land occupation was up to 7,288 km2 by 2019; projected 85–202 thousand km2 by 2060 (C1 and C3 use less land by prioritizing lower-latitude, higher land-occupation-ratio sites).

The analysis demonstrates that maximizing global net GHG mitigation from solar PV hinges on two coordinated levers: (1) targeting installations to countries with carbon-intensive power systems (C3), which delivers substantially higher avoided emissions without increasing global PV capacity; and (2) concentrating manufacturing in regions with low life-cycle emission intensities (M3), which minimizes embodied emissions. Together, these choices can raise 2020–2060 net mitigation from 107.2 Gt (C1–M2 worst case) to 204.7 Gt (C3–M3 best case), a gain of 97.5 Gt—about 1.9 times 2020 global GHG emissions. Targeted deployment also accelerates convergence of grid emission factors, lowering global disparities, reducing GHG leakage and potential trade barriers, and enabling broader co-benefits including improved energy access and air quality in late-adopting regions. While manufacturing emissions are small relative to operational mitigation, their spatial differences meaningfully affect total outcomes and potential CBAM-related costs. The findings support internationally coordinated PV industrial strategies that align installation with mitigation intensity and manufacturing with low-emission supply chains.

This study provides the first globally consistent, spatial-temporal LCA of the PV industrial chain, quantifying historical (2009–2019) and prospective (2020–2060) GHG emissions, mitigation, and net benefits under nine combined manufacturing–installation scenarios. It shows that prioritizing PV deployment in countries with high grid emission factors and situating manufacturing in low-intensity regions can unlock tens of gigatons of additional net mitigation without increasing total installed PV. Policy implications include: target installations to carbon-intensive grids; facilitate low-emission manufacturing hubs; accelerate recycling; and coordinate international finance and technology transfer to high-mitigation-potential regions. Future research should refine substitution dynamics (time-of-day/season), integrate storage and grid flexibility options more explicitly, explore intra-country siting heterogeneity, and leverage prospective LCA databases linked to IAMs (e.g., premise) under explicit climate targets.

Key limitations include: (1) Using average grid emission factors to represent displaced generation simplifies real-world temporal dispatch and marginal effects; (2) National averages for solar resource and grid conditions omit intra-country siting variability; (3) The PV share cap (70%) and exclusion of international electricity trade simplify system operation; (4) IEA grid emission factors may slightly underestimate full life-cycle emissions; (5) Silicon ore mining is excluded due to small contributions; (6) Future scenarios entail uncertainties in technology, costs, policy, and decarbonization pathways; (7) Storage assessments are approximate and not fully optimized within broader power system configurations; and (8) Results depend on assumed recycling trajectories and transportation patterns.