Data-driven approach towards identifying dyesensitizer molecules for higher power conversion efficiency in solar cells

Abstract

The machine learning (ML) research based on the quantitative structure-property relationship (QSPR) have been applied to the development of high-efficient dye-sensitized solar cells (DSSCs). This study brings forward a robust...

Application of machine learning‐based read‐across structure‐property relationship (RASPR) as a new tool for predictive modelling: Prediction of power conversion efficiency (PCE) for selected classes of organic dyes in dye‐sensitized solar cells (DSSCs)

“…We have also calculated the different cross-validation metrics like leave-one-out cross-validated R 2 (Q 2 LOO ), leave-one-out mean absolute error (MAE LOO ), and leave-one-out mean squared error (MSE LOO ), using training sets to see whether the generated models are overfitted or not [37]. The mathematical formula for all these metrics is shown in Equations ( 1)- (11).…”

“…Therefore, by changing only the molecular structures of the dye and keeping other parts of the DSSCs constant, the performance of the DSSCs can be improved. Different in-silico methods including various machine learning (ML) tools are being used for the design and development of new dye molecules by appropriate structural modifications in the field of materials science [9][10][11][12][13][14][15][16][17][18][19].…”

Deep-learning-assisted photovoltaic performance prediction of sensitizers in dye-sensitized solar cells

“…Very recently, influential descriptors extracted from the molecular structure and excited properties have been chosen to construct ML models, resulting in high prediction accuracy for power conversion efficiency. 25,26,31 However, there remain several limitations when applying ML to the development of DSSCs. One issue is that partial microscopic features or descriptors for the molecules are obtained from high-accuracy quantum calculations on excited states, which are computationally expensive and unsuitable for fast virtual screening or large-scale prediction.…”

Machine learning enabled high-throughput screening of inorganic solid electrolytes for regulating dendritic growth in lithium metal anodes

“…ML is the science of getting computers to perform calculations and improve themselves continuously without being explicitly programmed. It has a wide variety of applications in areas such as speech recognition, 31 autonomous driving, 32 solar cells, 33 and medicine, 34 where conventional algorithms perform poorly or are not suitable. ML tools are extensively being deployed for generating, testing, and refining scientific models.…”