Perovskite/silicon tandem solar cells: from first demonstrations to efficient, scalable and durable (?) devices

Christophe Ballif

EPFL Photovoltaics and Thin-Film Electronics Laboratory (PV-Lab), Switzerland

CSEM Sustainable Energy Center, Switzerland

Abstract

In barely a decade, monolithic perovskite/silicon tandem solar cells have progressed from a 13.7% proof of concept (2015) to certified efficiencies of 35.5% (LONGi, 2026), well beyond the practical limit of single-junction silicon. This keynote will first review that worldwide journey: the early two terminal demonstrations on planar silicon heterojunction cells, the move to fully textured bottom cells, the advent of self-assembled monolayer hole contacts and advanced interface passivation, the successive records by academic and industrial teams in Europe, the Middle East and China, and the first steps towards large-area cells, modules and pilot production. It will also trace the progress of perovskite/perovskite/silicon triple junctions, first demonstrated in 2018. We will report on the key steps undertaken at EPFL and CSEM in Neuchâtel to break several efficiency barriers on small-scale devices [1–4] with heterojunction and Topcon bottom, and discuss new results for fully screen-printed solar cells of 2 × 2 cm² and 60 cm². We will also report on the various deposition techniques that allow for high-efficiency devices [5], and on the impact of different front silicon surface textures on the optical properties of the devices [6]. In addition, we will comment on the role of photon recycling in measuring the quantum efficiency of tandem cells [7].

While efficiency records open the prospect of extending the efficiency curve of c-Si modules over the next decade, the decisive challenges now lie elsewhere: operational stability under heat, light, reverse bias and day–night cycling, where mobile ions, halide segregation in wide-bandgap absorbers and weak interfaces dominate [8]. We will discuss a few degradation modes, in particular lateral ion migration and its impact on cell (in)stability measurements under light soaking [9] and how smart automation can lead to faster progress towards device with enhanced stability [10] We will then discuss the importance of crystalline defects [11] and introduce new results for FAPbI₃ devices [12]. The usual stabilisation of α-FAPbI₃ relies on alloying with Cs⁺ or MA⁺, which unavoidably creates nanoscale A site heterogeneity and stacking faults. We will show how, by careful tuning of the layer stacks and of the crystallisation process, it is possible to achieve phase-pure α-FAPbI₃ with a uniform lattice parameter throughout the film, a (001) orientation and a tenfold increase in photoluminescence quantum yield. Inverted cells reach 25.8% efficiency. Unencapsulated devices retain their full performance for more than 2,400 h at 85 °C under 1-sun maximum-power-point tracking in ambient air without a UV filter. Minimodules lose less than 4% after 1,500 h at 75 °C, and outdoor tests in Berlin maintain a performance ratio of 0.9–1 over eight months.

Finally, we will conclude on what these advances mean for bankable, industrial perovskite/silicon tandem technology.

References：

1. F. Sahli et al., Nat. Mater. 17, 820 (2018) - fully textured monolithic tandems.

2. X. Y. Chin et al., Science 381, 59 (2023) - 31.25% via interface passivation.

3. K. Artuk et al., Nat. Commun. 16, 8672 (2025) - J Hurni,et al EES Solar 1 (2025) .

4. K. Artuk et al., Nature (2026), - triple-junction cells at 30.02% certified.

5. A. G. Kuba et al., ACS Energy Lett. 11, 3164 (2026) — two-step evaporation/CSS absorbers.

6. D. Turkay et al. ACS Energy Lett 10(2025) Beyond Flat. 

7. S. Zeder et al. Solar RRL 9 (2025) – impact of PR on IQE.

8. D. A. Jacobs et al., Energy Environ. Sci. 15, 5324 (2022) — lateral ion migration.

9. C Ongaro,et al. EES solar 2 (2026), wide bandgap degradation modes.

10. HA Can, et al Digital Discovery 4 (10), 2804-2815.

11. M. Othman et al., Energy Environ. Sci. 17 (2024); Adv. Mater. 37, 2502142 (2025)。 

12. M. Othman et al., manuscript submitted (2026).