Related Experiment Video
Updated: Aug 31, 2025
11:38
Influence of Hybrid Perovskite Fabrication Methods on Film Formation, Electronic Structure, and Solar Cell Performance
Published on: February 27, 2017
18.6K
Data-Driven Analysis of Hole-Transporting Materials for Perovskite Solar Cells Performance
Marcos Del Cueto1, Charles Rawski-Furman1, Juan Aragó2
1Department of Chemistry, University of Liverpool, Liverpool L69 3BX, U.K.
Summary
A machine learning model predicts perovskite solar cell performance using a dataset of 269 cells and hole-transporting material features. This model accurately identifies high and low-performing materials, aiding in the discovery of efficient solar cell components.
Area of Science:
- Materials Science
- Renewable Energy
- Computational Chemistry
Background:
- Perovskite solar cells (PSCs) are a promising photovoltaic technology.
- Hole-transporting materials (HTMs) are critical for PSC efficiency and stability.
- Developing new HTMs with optimized properties is essential for advancing PSC technology.
Purpose of the Study:
- To create a comprehensive dataset of PSCs and HTM features.
- To develop and validate a machine learning model for predicting PSC performance based on HTM characteristics.
- To identify key HTM features and chemical fragments that correlate with power conversion efficiency (PCE).
Main Methods:
- Compilation of a dataset of 269 perovskite solar cells, including perovskite family, cell architecture, and HTM features (fingerprints, additives, structural, and electronic properties).
- Development of a predictive machine learning model trained on the compiled dataset.
- Analysis of data biases and their impact on model accuracy.
- Investigation of specific chemical fragments (e.g., arylamine, aryloxy, thiophene) and their correlation with PCE.
Main Results:
- The machine learning model achieved reasonable accuracy in predicting PSC performance.
- The model successfully identified most top-performing and lowest-performing HTMs within the dataset.
- Specific chemical fragments, such as arylamine and aryloxy groups, showed a positive correlation with cell efficiency.
- Thiophene groups exhibited a negative correlation with power conversion efficiency (PCE).
Conclusions:
- A data-driven machine learning approach can effectively screen and predict HTMs for PSCs.
- Understanding the structure-property relationships of HTMs is crucial for optimizing PSC performance.
- The developed model and identified chemical fragment correlations provide valuable insights for designing next-generation perovskite solar cells.
