From ionic surfactants to Nafion through convolutional neural networks

TL;DR

This paper uses deep convolutional neural networks to analyze 3D soft matter data, enabling automatic detection of hydration levels and structural similarities between ionic surfactants and Nafion, offering a data-driven approach to disordered materials.

Contribution

The study introduces a transfer learning approach with CNNs to classify hydration levels and compare structures of ionic surfactants and Nafion without relying on prior models.

Findings

CNN accurately detects water content in surfactant structures

Transfer learning reveals structural similarities between surfactants and Nafion

Structure factor of Nafion can be expressed as a superposition of surfactant structures

Abstract

We have applied recent machine learning advances, deep convolutional neural network, to three-dimensional (voxels) soft matter data, generated by Molecular Dynamics computer simulation. We have focused on the structural and phase properties of a coarse-grained model of hydrated ionic surfactants. We have trained a classifier able to automatically detect the water quantity absorbed in the system, therefore associating to each hydration level the corresponding most representative nano-structure. Based on the notion of transfer learning, we have next applied the same network to the related polymeric ionomer Nafion, and have extracted a measure of the similarity of these configurations with those above. We demonstrate that on this basis it is possible to express the static structure factor of the polymer at fixed hydration level as a superposition of those of the surfactants at multiple water contents. We suggest that such a procedure can provide a useful, agnostic, data-driven, precise description of the multi-scale structure of disordered materials, without resorting to any a-priori model picture.