Related Experiment Video
Updated: Jul 16, 2026
11:38
Influence of Hybrid Perovskite Fabrication Methods on Film Formation, Electronic Structure, and Solar Cell Performance
Published on: February 27, 2017
Data-Driven Design of Self-Assembled Monolayers for High-Efficiency Perovskite Solar Cells.
Mingyu Song1, Lei Liu1, Peidong Chen1
1MOE Key Laboratory of Low-grade Energy Utilization Technologies and Systems, School of Energy and Power Engineering, Chongqing University, Chongqing, China.
Advanced Materials (Deerfield Beach, Fla.)
|July 15, 2026
Summary
A data-driven approach using machine learning optimized self-assembled monolayers (SAMs) for perovskite solar cells (PSCs). This strategy identified key molecular fragments, leading to a new SAM (S1) that achieved a 26.21% power conversion efficiency.
Area of Science:
- Materials Science
- Renewable Energy
- Artificial Intelligence
Background:
- Self-assembled monolayers (SAMs) are crucial for enhancing perovskite solar cell (PSC) performance.
- Rational design of SAMs is challenging due to the complex relationship between molecular structure and device efficiency.
Purpose of the Study:
- To develop a data-driven strategy for designing high-performance SAMs for PSCs.
- To identify dominant molecular fragments influencing SAM performance in PSCs.
Main Methods:
- Utilized a curated dataset of SAMs and their PSC efficiencies, encoding molecular structures into anchor-linker-head group fragments.
- Employed ensemble learning and SHapley Additive exPlanations (SHAP) for interpretability within a cross-validated framework.
- Recombined high-value fragments to create an expanded molecular library for virtual screening.
Main Results:
- Identified the head group as the primary driver of SAM performance.
- Virtually screened an expanded library, identifying a promising SAM molecule, S1.
- Experimental validation showed S1 forms an ordered monolayer on NiOx, optimizing electronic coupling and enabling defect passivation.
- Achieved a champion power conversion efficiency of 26.21% with S1.
Conclusions:
- Established a machine learning paradigm integrating fragment-based encoding and explainable AI for SAM design.
- Demonstrated a data-driven approach for optimizing interfaces in high-performance PSCs.
- S1 represents a significant advancement in SAM design for efficient perovskite solar cells.
