Beyond Silicon: Thin-Film Tandem for Photovoltaics (2025) PDF

Solar Energy Materials & Solar Cells 279 (2025) 113212 Contents lists available at ScienceDirect Solar Energy Materials and Solar Cells...

Solar Energy Materials & Solar Cells 279 (2025) 113212 Contents lists available at ScienceDirect Solar Energy Materials and Solar Cells journal homepage: www.elsevier.com/locate/solmat Beyond silicon: Thin-film tandem as an opportunity for photovoltaics supply chain diversification and faster power system decarbonization out to 2050 Alessandro Martulli a,*, Fabrizio Gota b, Neethi Rajagopalan c,d,e, Toby Meyer g , Cesar Omar Ramirez Quiroz h,i, Daniele Costa d,e , Ulrich W. Paetzold b,f, Robert Malina a,j, Bart Vermang e,k,l, Sebastien Lizin a a Centre for Environmental Sciences, Hasselt University, Martelarenlaan 42, 3500, Hasselt, Belgium b Light Technology Institute, Karlsruhe Institute of Technology, Engesserstrasse 13, 76131, Karlsruhe, Germany c Life Cycle Assessment Center of Expertise, Dow Silicones Belgium SRL, Parc Industriel, Rue Jules Bordet Zone C, 7180, Seneffe, Belgium d Smart Energy and Built Environment, Flemish Institute for Technical Research (VITO), Boeretang 200, 2400, Mol, Belgium e Energyville, Thor Park 831, 3600, Genk, Belgium f Institute of Microstructure Technology, Karlsruhe Institute of Technology, Hermann-von-Helmholtz-Platz 1, 76344, Eggenstein - Leopoldshafen, Germany g Solaronix SA, Rue de l’Ouriette 129, 1170, Aubonne, Switzerland h FOM Technologies, Artillerivej 86, 1., 2300, Copenhagen S, Denmark i NICE Solar Energy GmbH, Alfred-Leikam-Strasse 25, 74523, Schwaebisch Hall, Germany j Laboratory for Aviation and the Environment, Department of Aeronautics and Astronautics, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA, 02139, USA k Imec Division IMOMEC (partner in Solliance), Wetenschapspark 1, 3590, Diepenbeek, Belgium l Institute for Material Research (IMO, partnerin Solliance), Hasselt University, Wetenschapspark 1, 3590, Diepenbeek, Belgium A B S T R A C T In the last decade, the manufacturing capacity of silicon, the dominant PV technology, has increasingly been concentrated in China. This coincided with PV cost reduction, while, at the same time, posing risks to PV supply chain security. Recent advancements of novel perovskite tandem PV technologies as an alternative to traditional silicon-based PV provide opportunities for diversification of the PV manufacturing capacity and for increasing the GHG emission benefit of solar PV. Against this background, we estimate the current and future cost-competitiveness and GHG emissions of a set of already commercialized as well as emerging PV technologies for different production locations (China, USA, EU), both at residential and utility-scale. We find EU and USA-manufactured thin-film tandems to have 2–4 % and 0.5–2 % higher costs per kWh and 37–40 % and 32–35 % less GHG emissions per kWh at residential and utility-scale, respectively. Our projections indicate that they will also retain competitive costs (up to 2 % higher) and a 20 % GHG emissions advantage per kWh in 2050. 1. Introduction over 80 % per kWh of electricity generated in the last decade , which came at the expense of relocating PV manufacturing capacity away from Over the last decade, photovoltaics (PV) deployment has grown other locations, especially Europe and the United States (USA). In 2021, significantly to reach 1 TW of global cumulative PV capacity installed 89 % of European solar PV modules were imported from China.. By 2050, capacity is expected to surpass 40 TW [2–4]. Approximately 50 % of modules in the USA are imported from China, Single-junction (SJ) silicon solar modules are the dominant PV tech­ Singapore, Taiwan, and Vietnam. nology, accounting for 95 % of the PV market. Between 2014 and The COVID-19 pandemic and the Ukraine war have recently shown 2021, China invested more than USD 50 billion in scaling up the do­ that global supply chains are vulnerable to shocks and that import de­ mestic manufacturing capacity of silicon PVs. Currently, China holds pendency can threaten security of supply, thereby limiting availability over 75 % of the production capacity in all stages of manufacturing PV and increasing prices [11,12]. For instance, after experiencing declines modules, and almost half of its modules are sold to Europe. This, for many years, PV module prices increased by approximately 25 % along with technological learning, has contributed to cost reductions of between 2020 and 2022 due to the increase in material input prices and * Corresponding author. E-mail address: [email protected] (A. Martulli). https://doi.org/10.1016/j.solmat.2024.113212 Received 11 August 2024; Received in revised form 7 October 2024; Accepted 8 October 2024 Available online 23 October 2024 0927-0248/© 2024 Elsevier B.V. All rights are reserved, including those for text and data mining, AI training, and similar technologies. A. Martulli et al. Solar Energy Materials and Solar Cells 279 (2025) 113212 supply chain disruptions [13,14]. 2. Methods Supply chain and market concentration-related price increases are an impediment to the rapid decarbonization of the energy system required 2.1. Device architecture and steps in the production process to reach climate goals. This is widely realized by policymakers. For example, in 2022, the International Energy Agency called for a diver­ The PV device architectures under consideration for manufacturing sification of the PV supply chain. In the USA, the Inflation Reduction are shown in Fig. 1. Due to the wide range of possible choices in cell Act contains provisions to stimulate domestic PV manufacturing ca­ architectures and fabrication methods employed, we focus on PV con­ pacity ramp-up through investment credits. Similar plans have figurations that are representative of industrial-scale production to been announced in the European Union (EU). reflect the purpose of this study. The selected SJ configurations comprise As traditional SJ PV technologies approach the theoretical efficiency two alternative silicon-based technologies, passivated emitter and rear limit of 30 % , one pathway to long-term cost competitiveness for a cell (PERC) and silicon heterojunction (HJT). We specifically focus on domestic PV module manufacturing sector may lie in the earlier devel­ these two silicon SJ technologies due to their strong presence in the opment and deployment of novel, advantageous technologies , such market and the best efficiency performance. PERC is the market- as PV tandem configurations. Perovskites have been growingly applied dominant c-Si PV technology that currently accounts for about 85 % as top-cell in such tandem configurations due to their advantageous of the world market share, whereas HJT is the most-efficient silicon PV bandgap tunability and rapid increase in power conversion effi­ technology. HJTs currently account for only 2 % of the market and ciency (PCE) in the last decade. Perovskite tandem configurations are expected to grow in the following years. The PERC technology combining bottom-cell based on silicon technologies have shown PCEs substituted the mainstream Al-BSF solar cell in the last decade by surpassing 30 %. Yet, thin-film tandems coupling perovskite with overcoming the recombination losses at the rear side, leading to higher CI(G)S bottom-cell have also been demonstrated with promising effi­ PCEs; the theoretical limit of PERC’s PCE is around 25 % due to bandgap ciencies above 25 % [21–23]. Moreover, thin-film tandems employing narrowing, Auger, and contact recombination losses. The HJT perovskite as top and bottom cell, with efficiencies above 25 % have also technology reduces contact recombination losses by using passivated emerged recently [24,25]. The latter results, combined with established contact layers (e.g., a-Si: H) between the c-Si wafer and doped silicon thin-film R&D centers and equipment suppliers in the EU and the USA films; this approach has demonstrated high efficiencies of around 26 % , may provide an opportunity to build new thin-film tandem PV. HJTs are the ideal bottom cell for high-efficiency tandem devices; module manufacturing capacity to provide the additional supply needed a tandem configuration that combines a perovskite top cell with an HJT due to expected growth of demand for solar PV. Consequently, this bottom cell currently holds the PCE record. Nevertheless, as would contribute to diversifying the PV supply chain and the PV prod­ highlighted by the first cost studies, tandems with HJT bottom cells may ucts available in the market. not be the most suitable from a cost perspective [27,32] due to the Previous analyses determined the conditions under which specific higher associated costs with HJT production. tandem PVs can compete with traditional SJ PVs [27–34]. However, As an alternative to PV technology entailing silicon, we consider SJ these studies did not provide an analysis that considers both a range of thin-films, being CIGS and perovskites, and thin-film tandem devices, SJ and tandem technologies and jointly covers cost and environmental being the 2T and 4T architectures for a perovskite/CIS tandem. As our impact aspects. Wikoff et al. addressed the embodied carbon of focus is on thin-film tandems employing the perovskite and CIGS tech­ single-junction c-Si and cadmium-telluride (CdTe) PV technologies in nologies, we exclude another relevant thin-film PV technology, the EU, USA, China and India. However, regional (dis)advantages and a cadmium-telluride (CdTe), currently present in the market. Moreover, perspective out to 2050 for tandem technologies were not addressed. among thin-film tandem technologies, we limit the analysis to perov­ Against this background, the goal of this paper is to define the con­ skite/CIS devices, thus excluding the all-perovskite tandem PV tech­ ditions under which thin-film tandem PVs, containing a perovskite top nology. The development of a 2T perovskite/CIS tandem has only cell and a CIS bottom cell, can be economically and environmentally recently been described. Its appeal derives from using a CIS bottom cell competitive with the market-dominant silicon PVs manufactured in that perfectly couples the p-i-n perovskite top cell, which allows for China. To obtain a harmonized study, we quantify the costs and the optimal bandgap tunability and therefore increases the efficiency. The lifecycle GHG emissions of different single-junction and tandem PV obtained PCE is close to 25 %, and values close to or above 30 % are technologies when manufactured either in the EU, China, or the USA as expected to be reached. Additionally, this device demonstrated a more if production of the modules were to take place now and out to 2050. For stable PCE. A detailed description of the manufacturing process of SJs, we consider PERC, HJT, CIGS, and perovskite PV modules. For the each of the devices shown in Fig. 1 can be found in the Supporting In­ tandems, beside perovskite/CIS we also focus on devices in a 2-terminal formation. All devices are fabricated on glass substrates encapsulated (2T) and 4-terminal (4T) architecture of perovskite//silicon (PERC or with EVA and glass. For the module assembly components, such as the HJT) modules. We start by quantifying present-day cost metrics for both substrates and encapsulation, the same deposition methods and mate­ the production of SJ and tandem modules as well as their deployment in rials used for silicon-based devices are assumed to be used for all PV systems across the three geographical areas considered (EU, China, thin-film devices. Below, for brevity, we describe the manufacturing and the USA) for the same climatic conditions and production scale. process for the thin-film perovskite/CIS tandem architectures, as these Since location also affects environmental impact, we proceed by quan­ are the most novel devices. tifying the lifecycle greenhouse gas (GHG) emissions associated with In its 2T configuration, the perovskite/CIS tandem consists of a p-i-n module production and the accompanying PV systems. Our results top cell deposited on the CIS bottom cell having ZnO and CdS, respec­ benefit from the output of an energy yield model that quantifies PV tively as a window layer and buffer layer, with the CIS absorber modules’ yield under realistic climate conditions. In a final step, we deposited on the molybdenum rear contact. The CIS and perovskite cells project the selected cost and GHG emissions indicators to 2050. We are connected through a recombination junction layer composed of account for growing worldwide solar capacity, cost reductions, power AZO. The top cell uses NiOx and a thin layer of 2PACz as the hole conversion efficiency (PCE) improvements, and changes in countries’ transport layer, which is deposited on ITO. The electron transport layer energy mixes. is composed of SnO2 and C60. The 4-terminal device employs the same materials as the 2-terminal, although it does not include a recombina­ tion junction and comprises an EVA spacer layer between the top and bottom cells. 2 A. Martulli et al. Solar Energy Materials and Solar Cells 279 (2025) 113212 Fig. 1. Configurations of PV technology assessed in this study. 2.1.1. Estimating energy yield given in Table 1 for the three selected locations. The PCE and EY figures The energy yield (EY) analysis calculates the annual energy output presented in Table 1 include the losses attributed to module in­ for each device under realistic irradiation conditions modeled using the terconnections within the active area. These are obtained by considering state-of-the-art energy yield platform open-source software “EYcalc” 5 % active area losses for each technology on the initial outcomes of the developed by KIT [40,41]. Four modules are integrated to create the EY simulations for the optical and energy yield simulations [44,45]. While platform: (i) an irradiance module, (ii) an optics module, (iii) an elec­ the differences in EY for different locations are due to different solar trics module, and (iv) an energy yield core module. The irradiance spectra, temperature differences are not considered across the various module computes the irradiance at selected locations with a time reso­ locations. We use TMY3 data to calculate a solar spectrum for each hour lution of 1 h. The irradiance is angularly and spectrally resolved, of the year. The TMY3 dataset contains meteorological data about, considering the meteorological conditions and the cloud coverage at the among others, humidity, cloud coverage, dry-bulb temperature, pres­ selected location. Meteorological data from the National Renewable sure, precipitable water and aerosol optical depth, which vary for each Energy Laboratory (NREL) are used. The optics module calculates the location. On top of that, the different latitude and longitude of the lo­ angularly and spectrally resolved absorptance for each layer of the solar cations lead to different irradiation conditions. Lastly, the different light cell stack. To this end, a combination of the transfer-matrix method absorption properties of each technology explain why certain (TMM) for thin, coherent layers and series expansion of the Lambert-Beer law for thick, optically incoherent layers is used. The Table 1 irradiance obtained from module (i) and the absorptance obtained from PCE and EY values under three climatic conditions for PV devices under module (ii) are then given as an input to the energy yield core module to consideration. compute the photogenerated current density in the absorber materials PCE Average Energy Energy Yield (kWh/m2) with a time resolution of 1 h. Using a one-diode analytical model, the (%) Yield (kWh/m2) electrics module then uses the time-resolved photogenerated currents to Desert Tropical Oceanic compute the maximum power point (MPP) for each year’s hour. In order ​ ​ ​ PERC 21.8 370 461 372 278 to estimate the temperature of the cells, we use the Nominal Operating HJT 22.8 377 451 375 295 Cell Temperature (NOCT) model , assuming a NOCT of 48 ◦ C (valid CIGS 20.2 343 426.5 345 256.7 for an open rack configuration), while the insolation on the cell and the Perovskite 20.4 362 455.8 365.3 264.9 ambient air temperature is extracted from TMY3 data. Ultimately 2T Perovskite/ 29.3 489 618 495 355 PERC temperature coefficients for the open circuit voltage (VOC) and short 4T Perovskite/ 29.2 498 630 499 365 circuit current density (JSC) are used to update the electrical simula­ PERC tions’ current density – voltage (J-V) characteristics as a function of the 2T Perovskite/ 29.6 503 636 504 368 cell temperature, which, as previously mentioned, was computed via the HJT 4T Perovskite/ 29.5 494 624 500 359 NOCT model. HJT EY simulations in realistic irradiation conditions are performed for 2T Perovskite/ 28.8 483 610 487 353 the ten devices under consideration. Three locations representing very CIS different climatic conditions are selected: Phoenix (desert), Miami 4T Perovskite/ 28.9 492 624 492 361 (tropical), and Seattle (temperate oceanic). The obtained EY results are CIS 3 A. Martulli et al. Solar Energy Materials and Solar Cells 279 (2025) 113212 technologies perform better than others for specific solar spectrum modules. conditions. The assumptions of this study follow the Methodology Guidelines on LCA of PV. Additionally, we use the most recent global warming 2.2. Estimating cost competitiveness and greenhouse gas emissions potential (GWP100) factors published by the IPCC. Depending on where the PV system is deployed, we use balance of system (BOS) cost To quantify and compare the cost competitiveness and GHG emis­ data for the EU, China, and the USA. Besides the differences in sions for manufacturing the PV devices shown in Fig. 1 across the three system area, the utility-scale and residential-scale differ in terms of locations, we apply the principles of an environmental-techno economic system component costs (Table 43 in Supporting Information). With assessment (ETEA) [46,47]. This technology assessment method in­ regards to emissions of the PV system components, the mounting tegrates a life-cycle assessment (LCA) with a techno-economic assess­ structure GHG associated emissions are modeled for the three regions ment (TEA) using the exact system boundaries. Here, we assess the PV (USA, EU and China); for the other components, a default global value devices from manufacturing to deployment in the PV systems without from the environmental impact database employed (ecoinvent 3.9) is considering transportation between stages, as its impact is negligible considered. Degradation rates are assumed equal across PV systems, due. The assessment comprises indicators computed at the PV module to the unavailability of such data for other than silicon PVs, with a rate and system level. At the module-level, the cost competitiveness and GHG of 0.88 %/year, 0.78 %/year, and 0.48 %/year respectively, for a desert, intensity are respectively determined by computing the minimum sus­ tropical and oceanic climate. Given that such a degradation rate has tainable prices (MSP) in USD$ per watt) and the global warming po­ not been demonstrated for perovskites, we also explore the LCOE and tentials (GWP) in kgCO2-eq per watt) for each PV device being GEF as a function of the perovskite cell’s degradation rate. For 4T tan­ manufactured in each of the selected locations today. These two in­ dem applications, we assume that the perovskite cell contributes to 65 % dicators do not include the evaluation of cost and GHG emissions of the overall PCE and thus of the EY performance, based on recent derived from the use of the PV devices. The module-level analysis has findings that place the top-cell influence between 60 % and 70 %. thus a cradle-to-gate approach. In contrast, for 2T tandems, we reckon that the perovskite’s degradation For each technology and region, we model manufacturing with an rate determines the lifetime of the entire device, as for this device the annual production capacity of 100 MW. We opt for this scale because, two sub-cells are in series. Since our paper is about diversifying the PV for the novel technologies, we assumed that additional capacity needs to supply chain, we compare the LCOE and GEF indicators for PV modules be added, and the industry often uses this scale before ramping up its manufactured in the EU, USA, or China and installed in a PV system in capacity. This additionally allows for an analysis that excludes the cost the EU or USA. advantage coming from the scale. However, for the more established We account for two types of uncertainty while quantifying the in­ technologies, being PERC and HJT SJ, we also consider larger dicators mentioned above. Technical uncertainty is caused by diver­ manufacturing capacities for a single plant to capture effects derived gence in layer thicknesses for the considered device stack architectures, from manufacturing silicon PV cells and modules at larger scales. In the equipment type employed, and the consequent material and energy particular, the indicators are also computed for 1 GW and 500 MW use. The input parameters’ ranges are larger for the novel technologies, production of PERC and HJT in China. We refrained from performing such as tandems or perovskite SJ devices, compared to market-available such an analysis to expand CIGS SJ manufacturing capacity as this technologies, such as silicon-based and CIGS SJ, due to more standard­ would be a break in an ongoing trend of shutdowns. Furthermore, ized production routes. Technical uncertainty affects both the cost and we assume that the cost of capital is equally costly across the technol­ GHG emission indicators. Price-related uncertainty is mainly related to ogies and the regions considered due to the implementation of the materials and capital equipment cost variation. We, therefore, estimate Inflation Reduction Act in the USA and the Green Deal Industrial Plan in these indicators’ distribution by performing a Monte Carlo analysis with the EU. Additionally, as the cost of capital for PV projects decreases over 50 000 iterations. In every iteration, for every uncertain input param­ time, this would likely be similar across regions. eter, a value is randomly drawn from a triangular distribution created Cost differentials across regions are assumed to be related to labor, with the most likely, maximum, and minimum values found describing energy, and scale. Data regarding energy and labor cost differences are that specific parameter. readily available in contrast to regional price differences for materials This method allows for the identification of ranges for indicators like and equipment. For the latter, we consider that manufacturers based in ours that are dependent on multiple, uncertain input parameters. Be­ China have a cost advantage compared to those based in the EU or the sides, the approach to executing the MC we describe above allows for the USA due to the larger scale of raw material and equipment identification of cost and environmental impact uncertainties using the manufacturing and associated concentration of supply chain activities. minimal resources available (e.g. technical and price data) for devel­ Specifically, we assume a 20 % cost advantage for acquiring equipment oping the cost and environmental impact analysis. of silicon-based PVs manufactured in China whereas, no equipment In a final step, we extend our analysis by projecting the LCOE and cost differences are accounted for thin-film PVs due to the available GEF results to 2050. The selected time frame is based on the expectation equipment manufacturers in the USA and the EU. We also assume 10 % that tandem technologies may begin to gain a portion of the photovol­ price advantage for sourcing materials for all PV technologies manu­ taic market after 2030. Additionally, substantial efforts towards factured in China , including 100 MW and large-scale silicon PV decarbonizing the countries’ energy mixes target the year 2050. In manufacturing (1 GW PERC and 500 MW HJT). We consider equivalent doing so, we consider forecasts for cumulative PV capacity that estimate efficiency performance for PV technologies originating in the EU, USA worldwide capacity to be 42 TW in 2050 [2,59]; we then estimate and China as connections between equipment manufacturers and R&D modules and BOS costs by employing a learning rate approach. We institutes facilitate technology diffusion on a global scale [51,52]. assume that future cost reductions are related to the industry learning Analogously, at system-level, we quantify the levelized cost of elec­ rate, and therefore, we disregard potential cost reductions due to an tricity (LCOE) in USD$cents per kWh and GHG emission factor (GEF) in increase in manufacturing scale for each technology. Besides the kgCO2-eq per kWh for residential scale (30 m2) and utility-scale (0.5 learning rate, module cost reductions are affected by PV technologies km2). With the system-level analysis, we consider the effect of the en­ market shares as shown in the Supporting Information (Table 59). When ergy yield performance of each PV technology considered in this study. a specific PV technology is present in the market, the same learning rate Both indicators are computed over a period given by the years until the is applied; this implies that novel configurations such as perovskite/CIS PV system provides 80 % of the actual energy output for a maximum of are assumed to have, once entering the market, the same learning rate as 25 years. The inverter’s lifespan is taken at 15 years. All the other traditional crystalline silicon technologies (e.g., PERC) even though in components are considered to have the same life expectancy as the PV reality this could be different. However, we decided to adopt this 4 A. Martulli et al. Solar Energy Materials and Solar Cells 279 (2025) 113212 approach as there is no information about learning rates for tandem and material related). Large-scale manufacturing (1 GW) of PERC PVs in thin-film PV technologies yet. China results in the lowest MSPs being, on average, at 0.27 USD$/W. Table 59 in the Supporting Information provides the assumed market Although tandems generally show higher MSPs, the thin-film-based, shares for each technology. For perovskite/silicon tandem technologies, perovskite/CIS are likely to reach similar MSP values as HJT devices, the module cost reductions are not only driven by the perovskite (tan­ ranging between 0.36 and 0.44 USD$/W. In contrast, perovskite/silicon dem) share in the market but also by the PERC and HJT shares tandems, if manufactured in Europe (EU) or the USA, have higher prices (depending on whether the tandem considered is perovskite/PERC or than thin-film tandems, with perovskite/HJT MSPs over 0.49 USD$/W perovskite/HJT). Thus, the module cost fraction of perovskite/silicon and perovskite/PERC MSP above 0.45 USD$/W. Although the distri­ tandems dependent on the PERC (or HJT) technology is assumed to bution of novel thin-film tandems MSPs, such as perovskite/CIS tan­ decrease based on the PERC (or HJT) market share. The module cost dems, is noticeably wider than those of SJ, the probability of perovskite/ fraction dependent on the perovskite technology is assumed to decrease CIS manufactured in the EU to having a lower MSP than the mean MSP based on the perovskite tandem market share. Perovskite/CIS module of perovskite/PERC (0.39 USD$/W) and perovskite/HJT (0.43 USD cost reductions, are only driven by the perovskite (tandem) market $/W), manufactured in China, is 26 % and 48 %, respectively. USA- share. based manufacturing of thin-film-based tandems may result in lower Moreover, we account for expected PCE improvements by fitting a costs compared to the perovskite/HJT tandem and may result in com­ logistic function to the cell efficiencies data provided [20,61]. The parable costs with perovskite/PERC tandems with manufacturing based percentage increase in PCE is then taken as a proxy for the expected in China; in this case, the probability of perovskite/CIS manufactured in relative increase in the EY values in each climatic location. Additionally, the USA to have a lower MSP than the mean MSP of perovskite/PERC we consider changes in all three countries’ electricity mixes de­ manufactured in China is 51 %. As seen on the right-hand side of Fig. 2, velopments according to scenarios of energy system transition pathways when considering the GHG emissions, the differences are more clear-cut developed by the Lawrence Berkeley National Laboratory, NREL, and across the manufacturing locations with a clear advantage to European the EU Commission [62–64]. Mathematical formulae for computing the production. In this location, only the perovskite SJ device obtains lower indicators, the model specification for the learning rate and logistic GHG emissions per watt than the perovskite/CIS tandems (EU), which in function, and the data used as input for their calculation can be found in turn presents GWP reduction of around 30 %, 32 %, 33 %, and 29 % Supporting Information. compared to perovskite/HJT, perovskite/PERC, PERC SJ and HJT SJ. Furthermore, the mean GWP of perovskite/CIS tandems manufactured 3. Results and discussion in the EU is approximately 58 % and 56 % lower than PERC and HJT SJ modules produced in China. The mean GWP of perovskite/CIS modules 3.1. Module-level indicators produced in the USA is 49 % and 48 % less than PERC and HJT modules made in China. In sum, contrasting results are thus found for the MSP The resulting present-day MSPs and GWPs are shown in Fig. 2. As can and GWP across regions. Whereas Chinese PV module production is be seen on the left-hand side of Fig. 2, SJ photovoltaics are found to have associated with the lowest price per watt, European and US-based the lowest MSP with the perovskite device resulting in approximately manufacturing cause the lowest GHG emissions. The former can be 0.25–0.34 USD$/W, depending on the manufacturing location. Despite explained by the lower material and capital costs assumed in this study the lower efficiency than silicon SJ technologies, perovskite SJ can for Chinese producers and the markedly lower labor costs. As for energy achieve low MSP due to the associated low manufacturing costs (energy costs, EU-based manufacturers are greatly influenced by the high energy Fig. 2. Minimum Sustainable Price and GHG emissions associated with the production of each PV technology under consideration in the EU, China, and the US. Results are based on 100 MW manufacturing capacity; for comparison, MSPs corresponding to manufacturing capacities of 1 GW and 500 MW (in China), respectively, for PERC and HJT SJ, are provided. 5 A. Martulli et al. Solar Energy Materials and Solar Cells 279 (2025) 113212 prices, which started rising in the second half of 2021. This aspect residential scale. This can be explained by the larger influence of BOS not only results in higher costs for European PV modules but also en­ costs for lower-capacity plants, as for these applications, tandems have a larges this difference for technologies for which the energy costs have clear advantage due to their higher PCEs and hence the possibility to more relevance in the total manufacturing costs, such as thin-film PVs offset the higher residential system cost. The tandem residential cost (tandem and SJ). The GWP differences are primarily due to the high advantage is more evident in areas characterized by higher system costs, carbon intensity of the Chinese electricity generation mix, which con­ such as the USA. At the utility scale, the tandem cost advantage is less tributes to higher GWP values compared to the European counterparts. marked, and except for the USA, where higher BOS costs are present, the perovskite SJ technology has the lowest LCOE. For the EU and USA, Perovskite/CIS PVs not only have the lowest LCOE at the residential 3.2. System-level indicators scale but also present the lowest LCOEs among tandem PVs at the utility- scale. Furthermore, they obtain comparable values with the widely Building on the obtained module-level estimates, we present the commercialized PERC SJ PVs. Additionally, we observe that perovskite/ estimated regional present-day LCOEs and GEFs for PV systems in Fig. 3. CIS are often found closer to the quadrant’s bottom left corner, indi­ For each region, results pertaining to the utility-scale are presented on cating the lowest LCOE and GEF. Perovskite/CIS tandems, manufactured the left-hand side, whereas results applying to the residential scale are in the EU, have about 23 % and 26 % lower GEF than PERC SJ manu­ plotted on the right-hand side. The values shown for each technology factured in the EU, respectively, at the residential and utility-scale; The result from an average of LCOE and GEF computed for the three climatic LCOE is 4 % and 1 % lower than PERC SJ respectively at the residential conditions (desert, tropical and oceanic). This cancels out the effect of and utility-scale. The cost competitiveness of the thin film tandems is where the PV system is installed on the presented metrics. even more relevant for the systems produced in the USA; here, the LCOE Generally, tandem PVs show the best economic performance at the Fig. 3. LCOE and GEF for all technologies under study. Results show the utility and residential scale PV systems in the EU, the USA, and China. Each technology is assumed to be produced in the same region where it is deployed. Each LCOE and GEF point is representative of the average of three climatic conditions (desert, tropical and oceanic). 6 A. Martulli et al. Solar Energy Materials and Solar Cells 279 (2025) 113212 is 9 % and 6 % lower than PERC SJ (manufactured in the USA) at the This suggests thin-film tandems to be a promising alternative to silicon residential and utility scale. SJ PVs and an investment opportunity for EU-based PV manufacturers to This confirms the potential of thin-film tandems to be cost- develop and commercialize the technology, diversifying the supply competitive with mature PV technologies, such as PERC, when dis­ chain. Nevertheless, emerging technologies such as perovskite SJ and regarding the option to import modules. Furthermore, the potential perovskite/PERC manufactured in China, if installed in EU PV systems, contribution to reducing the GHG emissions of electricity generation is would have a cost advantage compared to thin-film tandems manufac­ considerably higher for thin-film tandem PVs. tured in the EU. At the utility and residential scale, perovskite/PERC As for perovskite-based PV, despite recent advancements , sta­ (China) PVs have an LCOE of 4 % and 2 % lower than perovskite/CIS bility represents a significant barrier to commercialization. For this (EU). Moreover, at the utility scale, PERC SJ (CN), with manufacturing reason, the LCOE and GEF are recalculated (Fig. 4) by varying degra­ capacities of 1 GW, presents the lowest LCOE among all PVs, approxi­ dation rates of perovskite-based PVs (SJ and tandems), between 0.50 mately 9 % lower than perovskite/CIS made in the EU. In contrast, EU %/year and 1.20 %/year. Degradation rates for PERC SJ PVs are the PV systems deploying perovskite/CIS manufactured in the EU would same as the baseline values assumed in this study, namely 0.88 %/year, have a 40 % and 37 % lower GEF at the utility and residential scale 0.78 %/year, and 0.48 %/year respectively, for a desert, tropical and compared to perovskite/PERC produced in China. Similar findings can oceanic climate. Generally, the findings suggest that all be drawn from Fig. 6 for USA PV systems. In this case, at the utility-scale, perovskite-based PVs should have an average yearly degradation rate USA-manufactured perovskite/CIS tandems exhibit more competitive below 0.80 %/year to have both a competitive LCOE and GEF with PERC LCOEs compared to PERC (1 GW) and perovskite/PERC tandems man­ SJ PVs. However, the maximum degradation value for competitive ufactured in China. Here, the perovskite/CIS LCOE is only 2 % higher LCOE/GEF varies among the perovskite PV technologies. 4T Per­ than PERC (1 GW) and has approximately the same LCOE compared to ovskite/CIS tandems have the highest limit for degradation rates, as for perovskite/PERC tandems made in China. values approaching 1.00 %/year, the LCOE is only 3 % higher than PERC SJ. On the contrary, the GEF is always lower than that for PERC SJ. 3.3. Forecasted system-level indicators Among silicon-based tandems, the degradation analysis suggests that the use of PERC as bottom-cell is favored over that of HJT, given their The forecasted LCOEs and GEFs out to 2050 at residential and utility cost and performance estimates. scales are shown in Figs. 7 and 8, respectively, for PV systems installed in Fig. 5 presents the results when considering imports. It depicts an the EU and the USA using domestic production or imports from China. LCOE and GEF comparison between EU-manufactured and Chinese- As this paper’s goal is to verify whether thin-film tandems manufactured manufactured PV modules installed in an EU PV system, entailing EU in the EU and the USA provide room for supply chain diversification and BOS cost for both utility (left-hand side) and residential scale (right- faster decarbonization, we here focus on perovskite/CIS made in the EU hand side). This allows verifying whether installing thin-film PV mod­ and USA vs. silicon alternatives made in China. These objectives are ules made in the EU could be competitive with installing imported motivated by the anticipated further increase of silicon PV production Chinese PV modules that currently dominate the European market. At capacity in China to further benefit from economies of scale. Thus, the residential scale, EU-manufactured thin-film tandem PV systems perovskite/CIS made in China is excluded even though it may represent may be cost-competitive with PV systems deploying SJ PVs produced in a valid low-cost alternative, as can be seen in Figs. 5 and 6. China. We can see that a PV system deployed in Europe utilizing EU- Despite the significant transitions towards low-carbon energy sour­ manufactured thin-film tandems’ LCOE distribution overlaps with its ces expected in the following decades for the Chinese energy mix, the counterpart utilizing PERC manufactured at large scale (1 GW) in China. mean GEF of perovskite/CIS tandem manufactured in the EU or USA will Fig. 4. Relative LCOE and GEF differences as a function of the perovskite cell degradation rate LCOE and GEF increase/decrease (utility-scale, average at three climatic locations) at various degradation rates, compared to PERC SJ. 7 A. Martulli et al. Solar Energy Materials and Solar Cells 279 (2025) 113212 Fig. 5. LCOE and GEF for EU-PV systems, with modules manufactured in the EU and China. Values refer to 100 MW manufacturing capacity; for comparison, results for PERC 1 GW and HJT 500 MW (both in China) manufacturing capacity are provided. Fig. 6. LCOE and GEF for USA-PV systems, with modules manufactured in the USA and China. Values refer to 100 MW manufacturing capacity; results for PERC 1 GW and HJT 500 MW (both in China) manufacturing capacity are provided for comparison. remain significantly lower than silicon SJ and tandem PV technologies compete with cheaper PV technologies manufactured in China. produced in China. In 2050, for an EU PV system, the mean GEF of thin- film tandems would still be 34 % and 39 % lower than PERC SJ (man­ 4. Conclusions ufactured in China) at the residential and utility-scale, respectively. The mean GEF of perovskite/CIS made in the EU is also lower, approximately In the following decades, additional PV manufacturing capacity will 19 %, than silicon-based tandems manufactured in China. Similar GEF need to be installed in the EU and the USA to meet the growing demand results are found for PV systems in the USA, with domestically manu­ for solar PV. In this paper, we therefore compared the production of factured perovskite/CIS tandems having 22 % and 19 % fewer emissions perovskite/CIS, perovskite/PERC, perovskite/HJT tandems and single- than perovskite/PERC made in China, respectively, at the utility-scale junction PVs such as PERC, HJT, perovskite, and CIGS based on eco­ and residential scale. nomic and environmental indicators that were calculated for PV mod­ For EU utility-scale PV systems, perovskite/HJT (CN) is expected to ules as if they were manufactured and installed today in the EU, the USA, be the most cost-competitive technology in 2050 by gradually shrinking and China under the same climatic conditions. It was presupposed that the gap in the following decades as market share and PCE increase. the performance of perovskite/CIS tandems obtained for lab-scale de­ Silicon tandems (CN) are the most cost-competitive for EU residential- vices can be achieved with module-area devices while being manufac­ scale PV systems. Yet, with perovskite market share growth coming tured using processing techniques that are amenable to upscaling. In into effect after 2035, LCOE reductions are envisioned for EU-produced addition, the results took as a given that perovskite-based devices ach­ thin-film tandems out to 2050, causing them to catch up. For PV systems ieve degradation rates lower than 1 %/year. These are requirements that based in the USA, the cost disadvantage of locally manufactured thin- have not yet been demonstrated and that may prove challenging to film tandems is less evident, making this technology more attractive to meet. We then sought to predict these indicators out to 2050. Our PV investors. findings show that the development of production capacity for emerging These results confirm the contrast between cost and GHG emission thin-film tandems, in particular perovskite/CIS, could provide a cost- performance of PV modules. Technologies such as perovskite/CIS tan­ competitive way to enable PV supply chain diversification and faster dems could provide considerable GHG emissions benefits to the PV in­ way to achieve power system decarbonization for the EU and the USA. dustry. Nevertheless, from a cost perspective, these would need to The main findings that support this statement are the following. 8 A. Martulli et al. Solar Energy Materials and Solar Cells 279 (2025) 113212 Fig. 7. Projected LCOE and GEF for EU PV systems (utility and residential scale) until 2050. As in Fig. 3, values are representative of averages at three climatic conditions. Perovskite/CIS tandems (with 100 MW manufacturing capacity) manufactured in the EU are compared to PERC SJ (with 1 GW manufacturing capacity), HJT SJ (with 500 MW manufacturing capacity), perovskite/PERC (with 100 MW manufacturing capacity), and perovskite/HJT PVs (with 100 MW manufacturing capacity) manufactured in China and imported to the EU. First, perovskite/CIS modules showed one of the best GHG emission price reductions are based on forecasts of each PV technology market performances as the associated GHG emissions were quantified as low as share until 2032, which are then extrapolated to 2050. Second, for 0.21 kgCO2eq./W if manufactured in the EU, approximately 56–58 % projections out to 2050, the PCE (which impacts both the economic and lower than PERC-HJT SJ PVs manufactured in China. Second, although the environmental metrics used) was extrapolated by fitting a logistic having higher values than SJ PVs produced at GW scale in China, their function to historical PV technology efficiency data. Higher future effi­ module price at 100 MW scale, ranging between 0.36 and 0.44 USD$/W ciency increases in specific tandem technologies (e.g., perovskite/HJT) for 2T and 4T tandems, was found to be competitive with perovskite/ would make these more attractive compared to others from both a cost HJT tandems manufactured in China for a same-sized plant. Third, for and GHG impact perspective, which is not covered in our study. degradation rates lower than 1.0 %/year, perovskite/cis tandems man­ Furthermore, the learning rate is assumed to be equal among the various ufactured in Europe entail the lowest GHG emissions per kWh across all tandem technologies; differences in technologies’ learning rates would PV systems. Fourth, their LCOE can be competitive with EU or USA- affect the economic metrics. Finally, this work is limited to costs of manufactured perovskite/PERC and PERC SJ. Fifth, although having a production and GHG emissions and the trade-offs between the two. higher LCOE than PERC SJ made in China (9 %) when compared to Future work could broaden the scope of the environmental impacts and perovskite/PERC tandems made in China, for the same manufacturing use aggregation methods for the environmental impacts (e.g., multi- scale, the perovskite/CIS tandems made in the EU showed slightly criteria analysis) to compare costs and environmental impacts more higher LCOEs (2–4 %) while at the same time having GHG emission holistically. reductions of 37–40 %. When looking out to 2050, thin-film PV pro­ duction in the EU and the USA continues to show lower GHG emissions CRediT authorship contribution statement than silicon PVs manufactured in China, even if the energy transition plans of China come to fruition. The LCOE results also exhibited that this Alessandro Martulli: Writing – review & editing, Writing – original carbon emission reduction brought by deploying diversified PV products draft, Visualization, Software, Project administration, Methodology, could come at similar costs compared to perovskite/silicon tandems, Formal analysis, Data curation, Conceptualization. Fabrizio Gota: provided strongly increasing market shares of thin-film tandem Writing – review & editing, Writing – original draft, Resources, Formal technologies. analysis, Data curation. Neethi Rajagopalan: Formal analysis, Data Due to the current large-scale silicon PV deployment, silicon tandems curation, Conceptualization. Toby Meyer: Validation, Resources, Data are the usual suspect for bringing tandem PVs to the market. Nonethe­ curation, Conceptualization. Cesar Omar Ramirez Quiroz: Validation, less, our results indicate that an equivalent investment in expanding Resources, Data curation, Conceptualization. Daniele Costa: Writing – thin-film tandem manufacturing capacity could be cost-competitive, review & editing, Validation. Ulrich W. Paetzold: Writing – review & assuming a comparable lifespan, while also ensuring significantly editing, Writing – original draft, Validation, Resources, Data curation, lower GHG emissions. Conceptualization. Robert Malina: Writing – review & editing, Writing We close by noting the limitations of this study. First, the module – original draft, Validation, Supervision, Conceptualization. Bart 9 A. Martulli et al. Solar Energy Materials and Solar Cells 279 (2025) 113212 Fig. 8. Projected LCOE and GEF for USA PV systems (utility and residential scale) until 2050. As in Fig. 3, values are representative of averages at three climatic conditions. Perovskite/CIS tandems (with 100 MW manufacturing capacity) manufactured in the EU are compared to PERC SJ (with 1 GW manufacturing capacity), HJT SJ (with 500 MW manufacturing capacity), perovskite/PERC (with 100 MW manufacturing capacity), and perovskite/HJT (with 100 MW manufacturing ca­ pacity) PVs manufactured in China and imported to the USA. Vermang: Writing – review & editing, Validation, Supervision, Re­ References sources, Project administration, Funding acquisition, Conceptualization. Sebastien Lizin: Writing – review & editing, Writing – original draft, J.F. Weaver, World has installed 1TW of solar capacity, PV Magazine (2022). htt ps://www.pv-magazine.com/2022/03/15/humans-have-installed-1-terawatt-of- Validation, Supervision, Project administration, Methodology, solar-capacity/. Conceptualization. D. Bogdanov, M. Ram, A. Aghahosseini, A. Gulagi, A.S. Oyewo, M. Child, U. Caldera, K. Sadovskaia, J. Farfan, L.D.N.S. Barbosa, M. Fasihi, S. Khalili, T. Traber, C. Breyer, Low-cost renewable electricity as the key driver of the global Declaration of competing interest energy transition towards sustainability, Energy 227 (2021), https://doi.org/ 10.1016/j.energy.2021.120467. The authors declare that they have no known competing financial F. Creutzig, P. Agoston, J.C. Goldschmidt, G. Luderer, G. Nemet, R.C. Pietzcker, The underestimated potential of solar energy to mitigate climate change, Nat. interests or personal relationships that could have appeared to influence Energy 2 (9) (2017), https://doi.org/10.1038/nenergy.2017.140. the work reported in this paper. N.M. Haegel, H. Atwater, T. Barnes, C. Breyer, A. Burrell, Y.M. Chiang, S. De Wolf, B. Dimmler, D. Feldman, S. Glunz, J.C. Goldschmidt, D. Hochschild, R. Inzunza, I. Kaizuka, B. Kroposki, S. Kurtz, S. Leu, R. Margolis, K. Matsubara, A.W. Bett, Acknowledgements Terawatt-scale photovoltaics: transform global energy, Science 364 (6443) (2019) 836, https://doi.org/10.1126/science.aaw1845. This project has received funding from the European Union’s Hori­ FraunhoferISE, Photovoltaics report. https://www.ise.fraunhofer.de/conte zon 2020 research and innovation programme under grant agreement nt/dam/ise/de/documents/publications/studies/Photovoltaics-Report.pdf, 2022. IEA, Special report on solar PV global supply chains. https://www.iea.org/reports/ No 850937. solar-pv-global-supply-chains, 2022. R. Chen, China Module Export Analysis, InfoLink Consulting, 2022. https://www. Appendix A. Supplementary data infolink-group.com/energy-article/solar-topic-chinas-module-exports-drop-fur ther-in-november. IRENA, Renewable power generation costs in 2021. https://irena.org/publicat Supplementary data to this article can be found online at https://doi. ions/2022/Jul/Renewable-Power-Generation-Costs-in-2021, 2022. org/10.1016/j.solmat.2024.113212. Eurostat, International trade in products related to green energy, Eurostat (2022). https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Internation al_trade_in_products_related_to_green_energy&oldid=575162#Overall.2C_the_EU_i Data availability mports_more_green_energy_products_than_it_exports. EIA, 2021 annual solar photovoltaic module shipments report. https://www.eia. gov/renewable/annual/solar_photo/pdf/pv_full_2021.pdf, 2022. Data is provided in Supporting Information file A. Goldthau, L. Hughes, Protect global supply chains for low-carbon technologies, Nature 585 (7823) (2020) 28–30, https://doi.org/10.1038/d41586-020-02499-8. OECD, OECD economic outlook, interim report march 2022. https://doi.org/10.1 787/4181d61b-en, 2022. IEA, Energy Technology Perspectives 2023, IEA, 2023. https://iea.blob.core.windo ws.net/assets/d5a18261-96c5-4f3c-b052-f7405a93cf10/EnergyTechnologyPerspe ctives2023.pdf. 10 A. Martulli et al. Solar Energy Materials and Solar Cells 279 (2025) 113212 pvXchange, Market Analysis December 2022 - a new year in PV – Part 1: what M. Hermle, F. Feldmann, M. Bivour, J.C. Goldschmidt, S.W. Glunz, Passivating retailers and manufacturers expect from 2023, pvXchange (2022). https://www. contacts and tandem concepts: approaches for the highest silicon-based solar cell pvxchange.com/Market-Analysis-December-2022-A-new-year-in-PV-Part-1-Wh efficiencies, Appl. Phys. Rev. 7 (2) (2020), https://doi.org/10.1063/1.5139202. at-retailers-and-manufacturers-expect-from-2023. A.W.Y. Ho-Baillie, J.H. Zheng, M.A. Mahmud, F.J. Ma, D.R. McKenzie, M.A. Green, R.F. Kennedy, Anne, The role of solar in the inflation reduction Act. PV Magazine, Recent progress and future prospects of perovskite tandem solar cells, Appl. Phys. 2022. https://pv-magazine-usa.com/2022/11/14/the-role-of-solar-in-the-inflation Rev. 8 (4) (2021), https://doi.org/10.1063/5.0061483. -reduction-act/. R. Schmager, M. Langenhorst, J. Lehr, U. Lemmer, B.S. Richards, U.W. Paetzold, European Commission, 16/03/2023), Net-Zero Industry Act: Making the EU the Methodology of energy yield modelling of perovskite-based multi-junction home of clean technologies manufacturing and green jobs (2023). https://ec.euro photovoltaics, Opt Express 27 (8) (2019) A507–A523, https://doi.org/10.1364/ pa.eu/commission/presscorner/detail/en/ip_23_1665. Oe.27.00a507. A. Richter, M. Hermle, S.W. Glunz, Reassessment of the limiting efficiency for R. Schmager, U.W. Paetzold, M. Langenhorst, F. Gota, EYcalc - energy yield crystalline silicon solar cells, IEEE J. Photovoltaics 3 (4) (2013) 1184–1191, calculator for multi-junction solar modules with realistic irradiance data and https://doi.org/10.1109/JPHOTOV.2013.2270351. textured interfaces. https://dx.doi.org/10.5281/zenodo.4696257, 2021. A. Bettoli, Tomas Nauclér, Thomas Nyheim, Andreas Schlosser, Christian Staudt, M.C.A. Garcia, J.L. Balenzategui, Estimation of photovoltaic module yearly Building a Competitive Solar-PV Supply Chain in Europe, M. Company, 2022. https temperature and performance based on Nominal Operation Cell Temperature ://www.mckinsey.com/industries/electric-power-and-natural-gas/our-insights/ calculations, Renew. Energy 29 (12) (2004) 1997–2010. Go to ISI>://WOS: building-a-competitive-solar-pv-supply-chain-in-europe. 000222344700005. H. Li, W. Zhang, Perovskite tandem solar cells: from fundamentals to commercial S.W.a.W. Marion, Innovation for Our Energy Future Users Manual for TMY3 Data deployment, Chem. Rev. 120 (18) (2020) 9835–9950, https://doi.org/10.1021/ Sets, 1994. acs.chemrev.9b00780. B. Abdollahi Nejand, D.B. Ritzer, H. Hu, F. Schackmar, S. Moghadamzadeh, NREL, Best research-cell efficiencies. https://www.nrel.gov/pv/assets/pdfs/best-re T. Feeney, R. Singh, F. Laufer, R. Schmager, R. Azmi, M. Kaiser, T. Abzieher, search-cell-efficiencies-rev220630.pdf, 2022. S. Gharibzadeh, E. Ahlswede, U. Lemmer, B.S. Richards, U.W. Paetzold, Scalable T. Feeney, I.M. Hossain, S. Gharibzadeh, F. Gota, R. Singh, P. Fassl, A. Mertens, two-terminal all-perovskite tandem solar modules with a 19.1% efficiency, Nat. A. Farag, J.-P. Becker, S. Paetel, E. Ahlswede, U.W. Paetzold, Four-terminal Energy 7 (7) (2022) 620–630, https://doi.org/10.1038/s41560-022-01059-w. perovskite/copper indium gallium selenide tandem solar cells: unveiling the path M.F. Stuckings, A.W. Blakers, A study of shading and resistive loss from the fingers to >27% in power conversion efficiency, Sol. RRL 6 (12) (2022) 2200662, https:// of encapsulated solar cells, Sol. Energy Mater. Sol. Cell. 59 (3) (1999) 233–242, doi.org/10.1002/solr.202200662. https://doi.org/10.1016/s0927-0248(99)00024-0. M. Jošt, E. Köhnen, A. Al-Ashouri, T. Bertram, Š. Tomšič, A. Magomedov, A. Martulli, N. Rajagopalan, F. Gota, T. Meyer, U.W. Paetzold, S. Claes, A. Salone, E. Kasparavicius, T. Kodalle, B. Lipovšek, V. Getautis, R. Schlatmann, C. J. Verboven, R. Malina, B. Vermang, S. Lizin, Towards market commercialization: A. Kaufmann, S. Albrecht, M. Topič, Perovskite/CIGS tandem solar cells: from lifecycle economic and environmental evaluation of scalable perovskite solar cells. certified 24.2% toward 30% and beyond, ACS Energy Lett. 7 (4) (2022) Progress in Photovoltaics, 2022, https://doi.org/10.1002/pip.3623. 1298–1307, https://doi.org/10.1021/acsenergylett.2c00274. G. Thomassen, M. Van Dael, S. Van Passel, F.Q. You, How to assess the potential of M.A. Ruiz-Preciado, F. Gota, P. Fassl, I.M. Hossain, R. Singh, F. Laufer, emerging green technologies? Towards a prospective environmental and techno- F. Schackmar, T. Feeney, A. Farag, I. Allegro, H. Hu, S. Gharibzadeh, B.A. Nejand, economic assessment framework, Green Chem. 21 (18) (2019) 4868–4886, V.S. Gevaerts, M. Simor, P.J. Bolt, U.W. Paetzold, Monolithic two-terminal https://doi.org/10.1039/c9gc02223f. perovskite/CIS tandem solar cells with efficiency approaching 25, ACS Energy Lett. S. Vorrath, Another blow to thin film, as Solar Frontier quits manufacturing and (2022) 2273–2281, https://doi.org/10.1021/acsenergylett.2c00707. switches sides, Renew Economy (2021). https://reneweconomy.com.au/another-bl Q. Jiang, J.H. Tong, R.A. Scheidt, X.M. Wang, A.E. Louks, Y.M. Xian, R. Tirawat, A. ow-to-thin-film-as-solar-frontier-quits-manufacturing-and-switches-sides/. F. Palmstrom, M.P. Hautzinger, S.P. Harvey, S. Johnston, L.T. Schelhas, B. ETIP-PV, Strategic research and innovation agenda on photovoltaics. https://medi W. Larson, E.L. Warren, M.C. Beard, J.J. Berry, Y.F. Yan, K. Zhu, Compositional a.etip-pv.eu/filer_public/85/68/8568e2ee-ad42-4198-8211-27b703e15e1a/sriapv texture engineering for highly stable wide-bandgap perovskite solar cells, Science -fullreport_web.pdf, 2022. 378 (6626) (2022) 1295–1300, https://doi.org/10.1126/science.adf0194. F. Egli, N. Orgland, M. Taylor, T.S. Schmidt, B. Steffen, Estimating the cost of B.A. Nejand, D.B. Ritzer, H. Hu, F. Schackmar, S. Moghadamzadeh, T. Feeney, capital for solar PV projects using auction results, Energy Pol. 183 (2023) 113849, R. Singh, F. Laufer, R. Schmager, R. Azmi, M. Kaiser, T. Abzieher, S. Gharibzadeh, https://doi.org/10.1016/j.enpol.2023.113849. E. Ahlswede, U. Lemmer, B.S. Richards, U.W. Paetzold, Scalable two-terminal all- A.C. Goodrich, D.M. Powell, T.L. James, M. Woodhouse, T. Buonassisi, Assessing perovskite tandem solar modules with a 19.1% efficiency, Nat. Energy 7 (7) (2022) the drivers of regional trends in solar photovoltaic manufacturing, Energy Environ. 620–630, https://doi.org/10.1038/s41560-022-01059-w. Sci. 6 (10) (2013) 2811–2821, https://doi.org/10.1039/c3ee40701b. CIGS-PV.net, CIGS white paper 2019. https://cigs-pv.net/wortpresse/wp-content/ A. de la Tour, M. Glachant, Y. Meniere, Innovation and international technology uploads/2019/04/CIGS_White_Paper_2019_online.pdf, 2019. transfer: the case of the Chinese photovoltaic industry, Energy Pol. 39 (2) (2011) N.L. Chang, J.H. Zheng, Y.L. Wu, H.P. Shen, F. Qi, K. Catchpole, A. Ho-Baillie, R. 761–770, https://doi.org/10.1016/j.enpol.2010.10.050. J. Egan, A bottom-up cost analysis of silicon-perovskite tandem photovoltaics, R. Frischknecht, M. Raugei, H.C. Kim, E. Alsema, M. Held, M. de Wild-Scholten, Progress in Photovoltaics 29 (3) (2021) 401–413, https://doi.org/10.1002/ Methodology Guidelines on Life Cycle Assessment of Photovoltaic Electricity, 12, pip.3354. IEA PVPS Task, Issue, 2020. https://iea-pvps.org/key-topics/methodology-guide E. Leccisi, V. Fthenakis, Life cycle energy demand and carbon emissions of scalable lines-on-life-cycle-assessment-of-photovoltaic-2020/. single-junction and tandem perovskite PV, Progress in Photovoltaics 29 (10) Ipcc, Climate Change 2021: the Physical Science Basis. Contribution of Working (2021) 1078–1092, https://doi.org/10.1002/pip.3442. Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Z.Q. Li, Y.Z. Zhao, X. Wang, Y.C. Sun, Z.G. Zhao, Y.J. Li, H.P. Zhou, Q. Chen, Cost Change, Cambridge University Press, 2021, https://doi.org/10.1017/ analysis of perovskite tandem photovoltaics, Joule 2 (8) (2018) 1559–1572, 9781009157896. In Press). https://doi.org/10.1016/j.joule.2018.05.001. D.C. Jordan, K. Anderson, K. Perry, M. Muller, M. Deceglie, R. White, C. Deline, M.M. Lunardi, A.W.Y. Ho-Baillie, J.P. Alvarez-Gaitan, S. Moore, R. Corkish, A life Photovoltaic fleet degradation insights, Progress in Photovoltaics 30 (10) (2022) cycle assessment of perovskite/silicon tandem solar cells, Progress in Photovoltaics 1166–1175, https://doi.org/10.1002/pip.3566. 25 (8) (2017) 679–695, https://doi.org/10.1002/pip.2877. S. Gharibzadeh, I.M. Hossain, P. Fassl, B.A. Nejand, T. Abzieher, M. Schultes, M. Roffeis, S. Kirner, J.C. Goldschmidt, B. Stannowski, L.M. Perez, C. Case, E. Ahlswede, P. Jackson, M. Powalla, S. Schaefer, M. Rienaecker, T. Wietler, M. Finkbeiner, New insights into the environmental performance of perovskite-on- R. Peibst, U. Lemmer, B.S. Richards, U.W. Paetzold, 2D/3D heterostructure for silicon tandem solar cells - a life cycle assessment of industrially manufactured semitransparent perovskite solar cells with engineered bandgap enables modules, Sustain. Energy Fuels 6 (12) (2022) 2924–2940, https://doi.org/ efficiencies exceeding 25% in four-terminal tandems with silicon and CIGS, Adv. 10.1039/d2se00096b. Funct. Mater. 30 (19) (2020), https://doi.org/10.1002/adfm.201909919. S.E. Sofia, H. Wang, A. Bruno, J.L. Cruz-Campa, T. Buonassisi, I.M. Peters, N.L. Chang, A.W.Y. Ho-Baillie, P.A. Basore, T.L. Young, R. Evans, R.J. Egan, Roadmap for cost-effective, commercially-viable perovskite silicon tandems for the A manufacturing cost estimation method with uncertainty analysis and its current and future PV market, Sustain. Energy Fuels 4 (2) (2020) 852–862, https:// application to perovskite on glass photovoltaic modules, Progress in Photovoltaics doi.org/10.1039/c9se00948e. 25 (5) (2017) 390–405, https://doi.org/10.1002/pip.2871. X.Y. Tian, S.D. Stranks, F.Q. You, Life cycle energy use and environmental IEA, Net zero by 2050. A roadmap for the global energy sector. https://www.iea.or implications of high-performance perovskite tandem solar cells, Sci. Adv. 6 (31) g/reports/net-zero-by-2050, 2021. (2020), https://doi.org/10.1126/sciadv.abb0055. J.C. Goldschmidt, L. Wagner, R. Pietzcker, L. Friedrich, Technological learning for Z.S.J. Yu, J.V. Carpenter, Z.C. Holman, Techno-economic viability of silicon-based resource efficient terawatt scale photovoltaics, Energy Environ. Sci. 14 (10) (2021) tandem photovoltaic modules in the United States, Nat. Energy 3 (9) (2018) 5147–5160, https://doi.org/10.1039/d1ee02497c. 747–753, https://doi.org/10.1038/s41560-018-0201-5. C. Breyer, A. Gerlach, Global overview on grid-parity, Progress in Photovoltaics 21 H.M. Wikoff, S.B. Reese, M.O. Reese, Embodied energy and carbon from the (1) (2013) 121–136, https://doi.org/10.1002/pip.1254. manufacture of cadmium telluride and silicon photovoltaics, Joule 6 (7) (2022) R. Preu, E. Lohmuller, S. Lohmuller, P. Saint-Cast, J.M. Greulich, Passivated 1710–1725. emitter and rear cell-Devices, technology, and modeling, Appl. Phys. Rev. 7 (4) VDMA, International technology roadmap for photovoltaic (ITRPV) 2021. https: (2020), https://doi.org/10.1063/5.0005090. //www.vdma.org/viewer/-/v2article/render/50902381, 2022. European Commission, D.-G. f. C. Action, D.-G. f. Energy, D.-G. f. Mobility, M.A. Green, The passivated emitter and rear cell (PERC): from conception to mass A. Transport, De Vita, P. Capros, L. Paroussos, K. Fragkiadakis, P. Karkatsoulis, production, Sol. Energy Mater. Sol. Cell. 143 (2015) 190–197, https://doi.org/ L. Höglund-Isaksson, W. Winiwarter, P. Purohit, A. Gómez-Sanabria, P. Rafaj, 10.1016/j.solmat.2015.06.055. L. Warnecke, A. Deppermann, M. Gusti, S. Frank, T. Kalokyris, EU Reference 11 A. Martulli et al. Solar Energy Materials and Solar Cells 279 (2025) 113212 Scenario 2020 : Energy, Transport and GHG Emissions : Trends to 2050, Eurostat, Energy prices on the rise in the euro area in 2021, Retrieved 12/09/2022 Publications Office, 2021, https://doi.org/10.2833/35750. from, https://ec.europa.eu/eurostat/web/products-eurostat-news/-/edn-202202 Lawrence Berkeley National Laboratory, China energy outlook: understanding 10-2, 2022. China’s energy and emissions trends. https://eta-publications.lbl.gov/sites/defa NREL, NREL-led breakthrough pushes perovskite cell to greater stability, ult/files/china_energy_outlook_2020.pdf, 2020. Efficiency, https://www.nrel.gov/news/press/2022/nrel-led-breakthrough-pushe NREL, 2022 standard scenarios report: a U.S. Electricity sector outlook, N. R. E. s-perovskite-cell-to-greater-stability-efficiency.html, 2022. Laboratory, https://www.nrel.gov/docs/fy23osti/84327.pdf, 2022. 12