Accelerated exploration of efficient ternary solar cells with PTB7:PC₇₁BM:SMPV1 using machine-learning methods

Area of Science:

• Materials Science
• Organic Electronics
• Computational Chemistry

Background:

• Organic solar cells (OSCs) offer a promising renewable energy source.
• Predicting photoelectric conversion efficiency (PCE) is crucial for OSC development.
• Ternary OSCs, incorporating a third component, can enhance performance.

Purpose of the Study:

• To employ machine learning (ML) for predicting PCE in ternary OSCs.
• To identify the optimal doping concentration of the third component for improved PCE.
• To experimentally validate ML predictions and elucidate performance enhancement mechanisms.

Main Methods:

• Machine learning models including Random Forest (RF), K-nearest neighbors, and Support Vector Machine were utilized.
• RF demonstrated superior accuracy in PCE prediction for ternary OSCs with PC71BM.
• Ternary OSCs with PTB7:PC71BM:SMPV1 were fabricated to verify ML predictions.

Main Results:

• The RF model predicted a champion PCE of approximately 8.01% for ternary OSCs with 6 wt% SMPV1.
• Experimental fabrication yielded a maximum PCE of 8.83% with 7.5 wt% SMPV1.
• ML feasibility was confirmed for PCE prediction and doping concentration optimization in ternary OSCs.

Conclusions:

• Machine learning, particularly RF, is a viable tool for predicting PCE in ternary OSCs.
• Experimental results validated ML predictions, demonstrating potential for optimizing ternary OSC performance.
• Performance enhancement in ternary OSCs is attributed to increased photon capture and improved charge transport.