Accurate Electronic and Optical Properties of Organic Doublet Radicals Using Machine Learned Range-Separated...

Area of Science:

• Computational Chemistry
• Materials Science
• Organic Electronics

Background:

• Organic semiconducting doublet-spin radicals exhibit unique optical properties due to minimal spin-flipping intersystem crossing (ISC).
• Their multiconfigurational nature poses challenges for traditional single-reference density functional theory (DFT) calculations.

Purpose of the Study:

• To extend the ML-ωPBE functional, developed using stacked ensemble machine learning (SEML), to accurately calculate properties of doublet-spin organic radicals.
• To assess the performance of ML-ωPBE for predicting optical properties like absorption and fluorescence energies.

Main Methods:

• Extended the ML-ωPBE range-separated hybrid (RSH) exchange-correlation (XC) functional to a dataset of 64 doublet-spin radicals.
• Utilized a new training set including 3926 closed-shell molecules and the 64 radicals.
• Employed linear-response time-dependent DFT (TDDFT) to evaluate absorption (Eabs) and fluorescence (Efl) energies, comparing ML-ωPBE with nine other XC functionals.

Main Results:

• ML-ωPBE accurately predicted the molecule-dependent range-separation parameter (ω) with a small mean absolute error (MAE) of 0.0197 a0−1, comparable to OT-ωPBE but with significantly reduced computational cost (2.46 orders of magnitude less).
• Demonstrated excellent domain adaptation of ML-ωPBE for diverse organic semiconducting species.
• ML-ωPBE reproduced experimental absorption and fluorescence energies with small MAEs of 0.299 eV and 0.254 eV, respectively, closely matching OT-ωPBE's performance.

Conclusions:

• Successfully extended the SEML framework and the ML-ωPBE functional from closed-shell molecules to doublet-spin organic radicals.
• ML-ωPBE provides a computationally efficient and accurate method for calculating optical properties of organic semiconductors using single-reference TDDFT.
• This work opens new avenues for the computational study of emergent optical materials like organic spin radicals.