QSPR modeling for the prediction of the triplet yield of singlet fission materials

Singlet excitons fission (SF) into triplet excitons can enhance the efficiency of solar cells while reducing the heat loss in light absorption systems, but requires photostable materials with high triplet yield. We use Support Vector Regression (SVR) to establish a Quantitative Structure-Property Relationship (QSPR) model for perdition the triplet yield (Φ_SF) of singlet fission materials (tetracene, pentacene, and perylenedimide derivatives, etc.). Minimum redundancy maximum relevance (mRMR), Floating front search (FFS) and Grid-search were used to select property molecular descriptions and parameter optimization, then the SVR model was established to predict the Φ_SF of these compounds. The correlation coefficient R values of leave one out cross-validation (LOOCV) test and external validation test set were 0.88 and 0.93, respectively. The results show that the SVR-QSPR model has strong generalization ability, can predict the Φ_SF value of singlet fission materials effectively and guide significance for developing highly efficient solar cells.