Machine Learning to Accelerate Screening for Marcus Reorganization Energies

Abstract

Understanding and predicting the charge transport properties of π-conjugated materials is an important challenge for designing new organic electronic applications, including solar cells, plastic transistors, light-emitting devices, and chemical sensors. A key component of the hopping mechanism of charge transfer in these materials is the Marcus reorganization energy, which serves as an activation barrier to hole or electron transfer. While modern density functional methods have proven to accurately predict trends in intramolecular reorganization energy, such calculations are computationally expensive. In this work, we outline active machine learning methods to predict computed intramolecular reorganization energies of a wide range of polythiophenes and their use towards screening new compounds with low internal reorganization energies. Our models have an overall root mean square error of ±0.113 eV, but a much smaller RMSE of only ±0.036 eV on the new screening set. Since the larger error derives from high-reorganization energy compounds, the new method is highly effective to screen for compounds with potentially efficient charge transport parameters.

Evaluating Fast Methods for Static Polarizabilities on Extended Conjugated Oligomers

“…We then used an empirical descriptor of the geometric size of the largest conjugated π-system. 42,43 While better corre-lated to GFN2 polarizability than HOMO-LUMO gap, this information was not enough to meaningfully correct large polarizabilities (Figure 4B).…”

Evaluating Fast Methods for Static Polarizabilities on Extended Conjugated Oligomers

“…We then used an empirical descriptor of the geometric size of the largest conjugated ⇡-system. 45,46 While better correlated to GFN2/D4 polarizability than HOMO-LUMO gap, this information was not enough to meaningfully correct large polarizabilities (Figure 4B). Plotting DFT polarizabilities against GFN2/D4 polarizabilities for the PubChemQC subset and the "wide range" set, we examine the e↵ects of using a polynomial fit.…”

Strategies for Computer-Aided Discovery of Novel Open-Shell Polymers

“…By forcing a quinoidal bonding structure we can have a straightforward comparison with the aromatic bonding structure. We constructed the hexamer of each system, as previous studies show a high correlation with the calculated electronic energies of longer oligomers, 30,31 and followed the same geometry optimizations and single-point calculations of both singlet and triplet species as described in the Computational Methods section below. Figure 4(b), demonstrates that lower singlet HOMO-LUMO gap correlates with a lower ∆E T −S and a more stable triplet ground state.…”