Abstract
Diblock copolymers have been shown to undergo microphase separation due to an interplay of repulsive interactions between dissimilar monomers, which leads to the stretching of chains and entropic loss due to the stretching. In thin films, additional effects due to confinement and monomer-surface interactions make microphase separation much more complicated than in that in bulks (i.e., without substrates). Previously, physics-based models have been used to interpret and extract various interaction parameters from the specular neutron reflectivities of annealed thin films containing diblock copolymers (J. P. Mahalik, J. W. Dugger, S. W. Sides, B. G. Sumpter, V. Lauter and R. Kumar, Interpreting neutron reflectivity profiles of diblock copolymer nanocomposite thin films using hybrid particle-field simulations, Macromolecules, 2018, 51(8), 3116; J. P. Mahalik, W. Li, A. T. Savici, S. Hahn, H. Lauter, H. Ambaye, B. G. Sumpter, V. Lauter and R. Kumar, Dispersity-driven stabilization of coexisting morphologies in asymmetric diblock copolymer thin films, Macromolecules, 2021, 54(1), 450). However, extracting Flory-Huggins χ parameters characterizing monomer-monomer, monomer-substrate, and monomer-air interactions has been labor-intensive and prone to errors, requiring the use of alternative methods for practical purposes. In this work, we have developed such an alternative method by employing a multi-layer perceptron, an autoencoder, and a variational autoencoder. These neural networks are used to extract interaction parameters not only from neutron scattering length density profiles constructed using self-consistent field theory-based simulations, but also from a noisy ad hoc model constructed previously. In particular, the variational autoencoder is shown to be the most promising tool when it comes to the reconstruction and extraction of parameters from an ad hoc neutron scattering length density profile of a thin film containing a symmetric di-block copolymer (poly(deuterated styrene-b-n-butyl methacrylate)). This work paves the way for automated analysis of specular neutron reflectivities from thin films of copolymers using machine learning tools.
| Original language | English |
|---|---|
| Pages (from-to) | 7280-7291 |
| Number of pages | 12 |
| Journal | Nanoscale |
| Volume | 15 |
| Issue number | 16 |
| DOIs | |
| State | Published - Mar 10 2023 |
Funding
Analysis of NR using machine learning tools was supported by the Center for Nanophase Materials Sciences, (CNMS), which is a US Department of Energy, Office of Science User Facility at Oak Ridge National Laboratory. DE and MJ acknowledge the sponsorship of this research by the US Department of Energy, Office of Science through the Science Undergraduate Laboratory Internship (SULI) and the management of this program through the ORISE. RK acknowledges discussions about NR with Dr James Browning, Dr Hanyu Wang, Dr Brad Lokitz, and Dr John F. Ankner. This research used the resources of the Compute and Data Environment for Science (CADES) at the Oak Ridge National Laboratory, which are supported by the Office of Science of the U.S. Department of Energy under contract no. DE-AC05-00OR22725. This research used resources at the Spallation Neutron Source, a Department of Energy Office of Science User Facility operated by the Oak Ridge National Laboratory. This manuscript has been authored by UT-Battelle, LLC, under contract DE-AC05-00OR22725 with the US Department of Energy (DOE). The US government retains and the publisher, by accepting the article for publication, acknowledges that the US government retains a nonexclusive, paid-up, irrevocable, worldwide license to publish, reproduce the published form of this manuscript, or allow others to do so, for US government purposes. DOE will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan ( https://energy.gov/downloads/doe-public-access-plan ).