Machine-learned dynamic disorder of electron transfer coupling

Area of Science:

• Physical Chemistry
• Computational Chemistry
• Biochemistry

Background:

• Electron transfer (ET) is crucial in chemical and biological processes.
• Electronic coupling dictates ET rates but is sensitive to nuclear dynamics, especially intermolecular movements.
• Dynamic disorder in ET coupling is poorly understood, limiting insights into charge transport.

Purpose of the Study:

• To investigate dynamic disorder in hole transfer coupling between ethylene and naphthalene dimers.
• To elucidate the role of intermolecular movements in electronic coupling dynamics.
• To characterize the spectral density of the coupling and its temperature dependence.

Main Methods:

• Utilized molecular dynamic (MD) simulations to model system dynamics.
• Employed machine-learning models to analyze electronic coupling.
• Calculated spectral density and identified dominant low-frequency modes.

Main Results:

• Low-frequency modes, driven by intermolecular rotation and translation, dominate coupling dynamics.
• Translational motion's contribution increases with temperature.
• The coupling exhibits sub-Ohmic spectral density with a cut-off frequency around 10^2 cm^-1.

Conclusions:

• Machine learning and MD simulations provide a powerful approach to study dynamic disorder in electronic coupling.
• Understanding these dynamics is key to advancing charge transport in complex systems.
• The findings offer new perspectives on the factors influencing electron transfer rates.