• Corpus ID: 247292694

Charge Transfer Simulations using Hamiltonian Elements and Forces from Neural Networks

  title={Charge Transfer Simulations using Hamiltonian Elements and Forces from Neural Networks},
  author={Philipp M. Dohmen and M. Kramer and Patrick Reiser and Pascal Friederich and Marcus Elstner and Weiwei Xie},
The trajectory surface hopping method has been widely used in the simulation of charge transport in organic semiconductors. In the present study, we employ the machine learning (ML) based Hamiltonian to simulate the charge transport in anthracene 1 ar X iv :2 20 3. 03 08 3v 2 [ ph ys ic s. ch em -p h] 2 3 M ar 2 02 2 and pentacene. The neural network (NN) based models are able to predict not just site energies and couplings but also the gradients of the site energy as well as off-diagonal… 

Figures and Tables from this paper



Machine learning enables long time scale molecular photodynamics simu- lations

Instead of expensive quantum chemistry during molecular dynamics simulations, deep neural networks are used to learn the relationship between a molecular geometry and its high-dimensional electronic properties and demonstrate that machine learning algorithms can outperform standard excited-state molecular dynamics approaches in their computational efficiency while delivering the same accuracy.

Automatic discovery of photoisomerization mechanisms with nanosecond machine learning photodynamics simulations†

The neural network photodynamics simulations of trans-hexafluoro-2-butene agree with the quantum chemical calculations showing the formation of the cis-product and reactive carbene intermediate and revealed subsequent thermal reactions in 1 ns.

Charge and Exciton Transfer Simulations using Machine-Learned Hamiltonians.

It is shown that kernel ridge regression models can be trained to predict electronic and excitonic couplings from semi-empirical DFTB reference data with very good accuracy and decreased the cost of exciton transfer simulations by one order of magnitude.

Machine learning and excited-state molecular dynamics

Recent advances for excited-state dynamics based on machine learning in quantum chemistry are surveyed to highlight successes, pitfalls, challenges and future avenues for machine learning approaches for light-induced molecular processes.

DFTB+, a software package for efficient approximate density functional theory based atomistic simulations.

An overview of the recently developed capabilities of the DFTB+ code is given, demonstrating with a few use case examples, and the strengths and weaknesses of the various features are discussed, to discuss on-going developments and possible future perspectives.

Combining SchNet and SHARC: The SchNarc Machine Learning Approach for Excited-State Dynamics

This work shows how deep learning can be used to advance the research field of photochemistry by learning all important properties—multiple energies, forces, and different couplings—for photodynamics simulations by incorporating spin–orbit couplings.

Quantum localization and delocalization of charge carriers in organic semiconducting crystals

By solving the time-dependent electronic Schrödinger equation coupled to nuclear motion for eight organic molecular crystals, the excess charge carrier forms a polaron delocalized over up to 10–20 molecules in the most conductive crystals.

Machine Learning for Predicting Electron Transfer Coupling.

This work systematically investigated the generality of the ML models, the choice of features and target labels, and the distance and orientation dependence of electronic coupling, and developed a machine learning approach to evaluate electronic coupling.

Thin Film Structures in Energy Applications

Thin film science and technology plays an important role in the development of devices in the future ranging from energy-efficient display devices to energy-harvesting and storage devices such as