Rapid Parameter Prediction in Perovskite Solar Cells via AI-Assisted Performance Analysis
Source abstract
Abstract
High Resolution Image Download MS PowerPoint Slide Metal halide perovskites are highly promising for photovoltaics, yet their complex physics makes comprehensive device characterization challenging. Here, we introduce a machine learning approach that enables holistic, high-precision characterization of perovskite solar cells from an impedance (EIS) measurement. By training a Random Forests Median Leaf Value model on over 50,000 drift-diffusion simulations, our model simultaneously predicts 12 critical device parameters, including ion dynamics and densities, carrier mobilities, and recombination properties, with high accuracy ( R 2 > 0.99). This approach advances prior work by extracting a more complete parameter set while alleviating the nonuniqueness problem through information-rich EIS input features. We rigorously validate the model against experimental values and closed-loop verification. Applying this technique to triple-cation perovskites with varying compositions reveals intriguing underlying trends, demonstrating its power to uncover complex structure–property relationships. This work establishes ML-driven frameworks for rapid, reliable device analysis to accelerate photovoltaic optimization.
Source: openalex_inverted_index · Quality: baseline
Research context
OpenAlex topics: Perovskite Materials and Applications · Machine Learning in Materials Science · Advanced Memory and Neural Computing
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Institute of Physics and Astronomy, University of Potsdam, Karl-Liebknecht-Str. 24-25, D-14476 Potsdam-Golm, Germany
École Polytechnique Fédérale de Lausanne (EPFL), Institute of Electrical and Microengineering (IEM), Photovoltaics and Thin Film Electronics Laboratory (PV-Lab), Rue de la Maladière 71b, 2000 Neuchâtel, Switzerland
Electronic Engineering Department, The Chinese University of Hong Kong, Chung Chi Rd, Ma Liu Shui, 00000 Hong Kong SAR, China
Heterostructure Semiconductor Physics, Paul Drude Institute for Solid State Electronics, Hausvogteiplatz 5-7, 10117 Berlin, Germany
Young Investigator Group Perovskite Tandem Solar Cells, Helmholtz-Zentrum Berlin für Materialien und Energie GmbH, Hahn-Meitner-Platz 1, D-14109 Berlin, Germany
Funding
Research Grants Council, University Grants Committee · Innovation and Technology Commission · National Natural Science Foundation of China · Chinese University of Hong Kong · Volkswagen Foundation
References (59)
Show source-provided references
1. Best Research-Cell Efficiency Chart. NREL. https://www.nlr.gov/pv/cell-efficiency (accessed 2018-07-20).
2. 10.1063/1.513275410.1063/1.5132754 ↗
3. 10.1039/C6EE00409A10.1039/C6EE00409A ↗
4. 10.1126/science.abd801410.1126/science.abd8014 ↗
5. 10.1038/nature1413310.1038/nature14133 ↗
6. 10.1021/acsami.9b0529310.1021/acsami.9b05293 ↗
7. 10.1021/acsami.4c1419410.1021/acsami.4c14194 ↗
8. 10.1002/aenm.20220110910.1002/aenm.202201109 ↗
9. 10.1002/adfm.20221427110.1002/adfm.202214271 ↗
10. 10.1038/s41524-021-00495-810.1038/s41524-021-00495-8 ↗
11. 10.1038/s42256-023-00736-z10.1038/s42256-023-00736-z ↗
12. 10.3390/photonics1003027110.3390/photonics10030271 ↗
13. 10.1016/j.xcrp.2021.10034610.1016/j.xcrp.2021.100346 ↗
14. Knapp, E.; Battaglia, M.; Stadelmann, T.; Jenatsch, S.; Ruhstaller, B. XGBoost Trained on Synthetic Data to Extract Material Parameters of Organic Semiconductors. 2021 8th Swiss Conference on Data Science (SDS) ; IEEE, 2021; pp 46–51. 10.1109/SDS51136.2021.00015.10.1109/SDS51136.2021.00015 ↗
15. 10.1063/5.013970710.1063/5.0139707 ↗
16. 10.1002/solr.20210077210.1002/solr.202100772 ↗
17. 10.1002/aenm.20170277210.1002/aenm.201702772 ↗
18. 10.1038/s41467-020-19769-810.1038/s41467-020-19769-8 ↗
19. 10.1002/pssa.20200029110.1002/pssa.202000291 ↗
20. 10.2307/229128210.2307/2291282 ↗
21. Owen A. B. · Statistica Sinica · 1992 · 2 · 439
22. 10.1063/1.173603410.1063/1.1736034 ↗
23. 10.1002/solr.20210021910.1002/solr.202100219 ↗
24. 10.1039/C9EE02020A10.1039/C9EE02020A ↗
25. 10.1080/10485252.2012.71516110.1080/10485252.2012.715161 ↗
26. 10.1016/j.mtener.2017.07.01410.1016/j.mtener.2017.07.014 ↗
27. 10.1038/s41563-023-01771-210.1038/s41563-023-01771-2 ↗
28. 10.1002/aenm.20240327910.1002/aenm.202403279 ↗
29. 10.1038/s41598-017-14920-w10.1038/s41598-017-14920-w ↗
30. 10.1039/D2EE02330J10.1039/D2EE02330J ↗
31. 10.1002/admi.20160069410.1002/admi.201600694 ↗
32. 10.1126/science.adg375510.1126/science.adg3755 ↗
33. 10.1038/s41586-023-05992-y10.1038/s41586-023-05992-y ↗
34. 10.1021/acs.jpclett.3c0053010.1021/acs.jpclett.3c00530 ↗
35. 10.1038/s41560-024-01487-w10.1038/s41560-024-01487-w ↗
36. 10.1039/D4EE02494J10.1039/D4EE02494J ↗
37. 10.1021/acsenergylett.5c0169010.1021/acsenergylett.5c01690 ↗
38. 10.1038/s41467-022-34203-x10.1038/s41467-022-34203-x ↗
39. 10.1021/ja903481810.1021/ja9034818 ↗
40. 10.1126/sciadv.adt117110.1126/sciadv.adt1171 ↗
41. 10.1039/C8EE03395A10.1039/C8EE03395A ↗
42. 10.1021/acsenergylett.7b0027610.1021/acsenergylett.7b00276 ↗
43. 10.1038/s41467-023-44418-110.1038/s41467-023-44418-1 ↗
44. 10.1002/aenm.20210277610.1002/aenm.202102776 ↗
45. 10.1039/C6EE03014A10.1039/C6EE03014A ↗
46. 10.1038/s41560-018-0219-810.1038/s41560-018-0219-8 ↗
47. 10.1039/C7SE00603A10.1039/C7SE00603A ↗
48. 10.1002/admi.20200036610.1002/admi.202000366 ↗
49. 10.1021/acsaem.2c0216110.1021/acsaem.2c02161 ↗
50. 10.1021/acsenergylett.0c0110410.1021/acsenergylett.0c01104 ↗
51. 10.1021/acsenergylett.0c0264210.1021/acsenergylett.0c02642 ↗
52. 10.1038/s41467-019-10468-710.1038/s41467-019-10468-7 ↗
53. 10.1021/acsami.6b0269210.1021/acsami.6b02692 ↗
54. 10.1063/1.497203210.1063/1.4972032 ↗
55. 10.1063/5.026801310.1063/5.0268013 ↗
56. 10.48550/arXiv.2511.0692910.48550/arXiv.2511.06929 ↗
57. 10.1021/acs.jpcc.4c0230410.1021/acs.jpcc.4c02304 ↗
58. 10.1002/advs.20200251010.1002/advs.202002510 ↗
59. 10.1021/acsami.3c0026910.1021/acsami.3c00269 ↗