A deep neural network is constructed to yield in principle exact exchange–correlation potential. It requires merely the electron densities of small molecules and ions and yet can determine the exchange–correlation potentials of large molecules. We train and validate the neural network based on the data for H2 and HeH+ and subsequently determine the ground-state electron density of stretched HeH+, linear H3 +, and H–He–He–H2+. Moreover, the deep neural network is proven to model the van der Waals interaction by being trained and validated on a data set containing He2. Comparisons to B3LYP are given to illustrate the accuracy and transferability of the neural network.

There is growing experimental and theoretical evidence that vibronic couplings, couplings between electronic and nuclear degrees of freedom, play a fundamental role in ultrafast excited-state dynamics in organic donor–acceptor hybrids. Whereas vibronic coupling has been shown to support charge separation at donor–acceptor interfaces, so far, little is known about its role in the real-space transport of charges in such systems. Here we theoretically study charge transport in thiophene:fullerene stacks using time-dependent density functional tight-binding theory combined with Ehrenfest molecular dynamics for open systems. Our results reveal coherent oscillations of the charge density between neighboring donor sites, persisting for ∼200 fs and promoting charge transport within the polymer stacks. At the donor–acceptor interface, vibronic wave packets are launched, propagating coherently over distances of more than 3 nm into the acceptor region. This supports previous experimental observations of long-range ballistic charge-carrier motion in organic photovoltaic systems and highlights the importance of vibronic coupling engineering as a concept for tailoring the functionality of hybrid organic devices.

Using an innovative quantum mechanical method for an open quantum system, we observe in real time and space the generation, migration, and dissociation of electron-hole pairs, transport of electrons and holes, and current emergence in an organic photovoltaic cell. Ehrenfest dynamics is used to study photoexcitation of thiophene:fullerene stacks coupled with a time-dependent density functional tight-binding method. Our results display the generation of an electron-hole pair in the donor and its subsequent migration to the donor-acceptor interface. At the interface, electrons transfer from the lowest unoccupied molecular orbitals (LUMOs) of thiophenes to the second LUMOs of fullerene. Further migration of electrons and holes leads to the emergence of current. These findings support previous experimental evidence of coherent couplings between electronic and vibrational degrees of freedom and are expected to stimulate further work toward exploring the interplay between electron-hole pair (exciton) binding and vibronic coupling for charge separation and transport.

When it comes to predicting experimental values of molecular properties with deep learning, the key problem is the lack of sufficient experimental data for training. We propose a method that consists of pretraining a graph neural network that aims to reproduce first-principles quantum mechanical results, followed by fine-tuning of a fully connected neural network against experimental results. The combined pretraining and fine-tuning model is expected to yield molecular properties close to experimental accuracy. This is made possible because first-principles quantum mechanical methods are often qualitatively correct or semiquantitatively accurate; thus, a calibration of the calculation results against high-precision but limited experiment data can improve accuracy greatly. Moreover, the method is highly efficient, as first-principles quantum mechanical calculation is bypassed. To demonstrate this, we apply the combined model to determine the experimental heats of formation of organic molecules made of H, C, O, N, or F atoms (up to 30 atoms), where mere 405 experimental data are used. The overall mean absolute error is 1.8 kcal/mol for these molecules.

The paper mainly simulates the effects of layer thickness and deposition velocity on thermal stress in fused deposition modelling (FDM). Different values of these variables are considered in this process. The simulation results show the trends of thermal stress with the two variables and the simulation results provide a guidance for the practical fabrication.