Comparing Environmental Impacts of Single-Junction Silicon and Silicon/Perovskite Tandem Photovoltaics-A Prospective Life Cycle Assessment

Abstract

Tandem photovoltaics applying perovskite on silicon are considered to be a possible route to sustaining continuous efficiency improvements and price reductions. A meaningful market share for such tandems is, however, at least a decade away. Herein, a comprehensive prospective life cycle assessment was conducted, comparing the full life cycle of monofacial and bifacial silicon/perovskite tandem panels with single-junction silicon panels produced up to 2050. The end-of-life included the recovery of silicon and silver. Climate change impacts per kilowatt hour were projected to decrease by two-thirds over time. Tandem panels are expected to reach impacts of 8-10 g CO₂-eq/kWh in 2050, while single-junction panels may reach 11-13 g CO₂-eq/kWh in 2050. Other midpoint impact categories with substantial contributions to damaging human health and ecosystem quality were toxicity, particulate matter formation, and acidification, with tandems having lower impacts in each category. Reductions in impacts over time are mainly the result of grid mix decarbonization and panel efficiency improvements. Balance-of-system and recycling were found to contribute substantially to these impact categories. To ensure that tandem panels provide environmental benefits, annual degradation rates should not exceed 1% for monofacial or 3% for bifacial tandems, and refurbishment of panels with advanced degradation is crucial.

Figure 1

Configuration of the four product systems and schematic of the light absorption of the bifacial silicon/perovskite tandem panel. EVA, ethylene vinyl acetate; ITO, indium tin oxide; PCBM, [6,6]-phenyl C61 butyric acid methyl ester; PERC, passivated emitter and rear contact; PET, polyethylene terephthalate; POE, polyolefin elastomer; PVD, physical vapor deposition; PVT, polyvinyl toluene; sALD, spatial atomic layer deposition; SDP+TC, slot die printing and thermal cure; and SnO₂, tin(IV) oxide.

Figure 2

Flowchart for the product systems of this study, where either one of four panels is installed on either a slanted or flat roof, as indicated with dashed arrows. System boundaries are indicated by dotted lines. Materials recovered in waste treatment are cut off at the indicated point. (CN), process occurs in China; (EU), process occurs in Europe; T, transport.

Figure 3

Midpoint to endpoint contribution analysis for the endpoints of (a) human health in DALY/kWh and (b) ecosystem quality in species·year/kWh using the hierarchist (H) perspective of the ReCiPe 2016 LCIA method. Percentages represent the decrease in impact between 2023 and 2050 for each respective background scenario (i.e., shared socioeconomic pathway 2 (SSP2) – baseline, SSP2 – representative concentration pathway (RCP) 2.6, and SSP2 – RCP 1.9, corresponding to 3–4, 1.6–1.8, and 1.2–1.4 °C global mean surface temperature increases by 2100, respectively). Other assumptions: panel design, monofacial; insolation, 1391 kWh/m²/year; geographic scope, China (production) and Europe (use and end-of-life); annual degradation rate for perovskite, 0.4–0.5% (relative).

Figure 4

Process contribution analysis for the impact category of climate change in kg of CO₂-eq/kWh using the EF v3.1 method. The x axis represents the year of production for the panel. Impacts from end-of-life were modeled to occur after the economic lifespan of the panel. For example, end-of-life for the panel produced in 2023 was modeled to occur in 2053. Assumptions: panel design: (a) monofacial single-junction, (b) monofacial tandem, (c) bifacial single-junction, or (d) bifacial tandem; insolation: 1391 kWh/m²/year; geographic scope: China (production) and Europe (use and end-of-life); annual degradation rate for perovskite: 0.5% (relative); background scenario: shared socio-economic pathway 2 - representative concentration pathway 2.6 (corresponding to a 1.6–1.8 °C global mean surface temperature increase by 2100).

Figure 5

Variance decomposition analysis for the results of all midpoint impact categories for (a) the single-junction panel or (b) the tandem panel using the hierarchist (H) perspective of the ReCiPe 2016 LCIA method.