Applications Of Machine Learning In Hydrogen-Bond Dynamics In Bulk Water And Polymer Chains' Structure Factor

The reorientation and diffusion of water molecules in liquid water are essential to a wide range of processes.Based on the Density Functional Theory-based Molecular Dynamics(DFTMD)simulation of bulk water,we designed a Recurrent Neural Network(RNN)based model to classify water molecule pairs' configuration changes as the hydrogen bond network evolves.Although we found that there are many types of configuration changes from the simulation trajectory,one has apparent characteristics: donor-acceptor(DA)exchange the configuration changes found in experiments or simulations related to water dimers.Besides,through our model,we determined the relative proportions of DA exchange and diffusion.Although the absolute number of the two processes will fluctuate with the increase of temperature,the relative proportion is basically invariable.This feature implies the universality of DA exchange processes of water molecules in the hydrogen bond network.As a substantial physical quantity to understand polymer chains' internal structure,the structure factor is studied both in theory and experiment.In this work,by training a deep neural network(NN),we obtained an efficient model to calculate the structure factor of polymer chains without considering different wavenumber and chain rigidity regions.Furthermore,based on the trained neural network model,we predicted the contour and Kuhn length of some polymer chains using scattering experimental data.We found our model can get pretty reasonable predictions.This work provides a method to obtain structure factor for polymer chains,which is as good as previous and more computationally efficient.Also,it provides a potential way for the experimental researchers to measure the contour and Kuhn length of polymer chains.