Exploring structural basis of photovoltaic dye materials to tune power conversion efficiencies: a DFT and ML analysis of Violanthrone

This study employs a systematic approach to modify Violanthrone (V) structures and analyze their impact on photovoltaic (PV) properties. We use cheminformatics based Python library based RDKit tool to calculate their structural descriptors for to correlate them with their PV parameters. Our analysis reveals a positive correlation for their Open-Circuit Voltage (V_oc) and Fill Factor (FF) for indicating that their higher voltage output is associated for their efficient charge carrier mobilities. We also predict their Power Conversion Efficiency (PCE) by drawing their their Scharber diagram which achieves their promising efficiency of up to 15 %. To further enhance the reliability our work, we conduct an extensive literature survey of such organic materials to predict their PCEs by their Machine Learning (ML) after utilizing various ML models. Among five tested ML models, it identifies the Random Forecast (RF) model and Gradient Boosting (GB) models as as the optimal one (R-squared value: 0.82). Their feature importance reveals that their FF is the most significant feature to impact their PCEs (importance value: 10.9). Furthermore, we observe a negative correlation between orbital interaction strength (E⁽²⁾) values and orbital energy differences E^(j)-E⁽ⁱ⁾ which indicates that their stronger orbital interactions are associated with their smaller energy differences. Our study provides valuable insights for their structural basis to PV material designs for enabling their design for efficient materials in energy conversion.