Machine learning assisted insights into the mechanical strength of nanocrystalline graphene oxide

Abstract

The mechanical properties of graphene oxides (GOs) are of great importance for their practical applications. Herein, extensive first-principles-based ReaxFF molecular dynamics (MD) simulations predict the wrinkling morphology and mechanical properties of nanocrystalline graphene oxides (NCGOs), with intricate effects of grain size, oxidation, hydroxylation, epoxidation, grain boundary (GB) hydroxylation, GB epoxidation, GB oxidation being considered. NCGOs show brittle failures initiating at GBs, obeying the weakest link principle. By training the MD data, four machine learning (ML) models are developed with capability in estimating the tensile strength of NCGOs, with sorting as Extreme Gradient Boosting (XGboost) > Multilayer Perceptron (MLP) > Gradient Boosting Decision Tree (GBDT) > Random Forest (RF). In the XGboot model, it is revealed that the strength of NCGOs is greatly dictated by oxidation and grain size, and the hydroxyl group plays more critical role in the strength of NCGOs than the epoxy group. These results uncover the pivotal roles of structural signatures in the mechanical strength of NCGOs, and provide critical guidance for mechanical designs of chemically-functionalized nanostructures.

Data-Driven Machine Learning to Predict Mechanical Properties of Monolayer Transition-Metal Dichalcogenides for Applications in Flexible Electronics

“…One of the major goals of ML models is to accomplish high-throughput measurement of critical characteristics of materials under different conditions . Previously, ML algorithms such as support vector machine (SVM) and feed forward neural network (FFNN) have been used to predict the mechanical properties of 2D materials. ,,− Wang et al used SVM to predict the mechanical properties of MoSe 2 . Xu et al investigated four ML models for predicting the mechanical strength of nanocrystalline graphene oxide .…”

“…15 Xu et al investigated four ML models for predicting the mechanical strength of nanocrystalline graphene oxide. 55 Among these four ML models, they found that the eXtreme Gradient Boosting algorithm is more capable of accurate predictions. 55 Mortazavi et al incorporated machine learning interatomic potentials (MLIP) in multi-scale modeling to efficiently bridge ab initio modeling to the continuum scale.…”

“…55 Among these four ML models, they found that the eXtreme Gradient Boosting algorithm is more capable of accurate predictions. 55 Mortazavi et al incorporated machine learning interatomic potentials (MLIP) in multi-scale modeling to efficiently bridge ab initio modeling to the continuum scale. 58,59 Kastuar et al proposed an ML-based toolkit (ElasTool toolkit) for studying the temperaturedependent mechanical properties of 2D MoS 2 -WS 2 heterostructures.…”

“…Previously, ML algorithms such as support vector machine (SVM) and feed forward neural network (FFNN) have been used to predict the mechanical properties of 2D materials. ,,− Wang et al used SVM to predict the mechanical properties of MoSe 2 . Xu et al investigated four ML models for predicting the mechanical strength of nanocrystalline graphene oxide . Among these four ML models, they found that the eXtreme Gradient Boosting algorithm is more capable of accurate predictions .…”

Hydrogenated Graphene with Tunable Poisson’s Ratio Using Machine Learning: Implication for Wearable Devices and Strain Sensors

“…In recent years, numerous new materials and metamaterials have been successfully discovered. To determine their properties, machine learning (ML) is considered a promising tool. − The ML method was used to search for materials and structures with desirable properties such as composite materials with optimal thermal conductivity, toughness, and strength, conductor materials for batteries, , graphene kirigami with high stretchability, , inflatable soft membranes, porous graphene structure with optimal thermal conductivity, thermoelectricity, and thermal transport properties of nanostructures, and other applications. − Several studies have focused on searching for metamaterials with auxeticity using the ML method. , Based on the results of finite element analysis, Wilt et al used the ML method to optimize an auxetic porous metamaterial with a re-entrant honeycomb unit cell at the macro scale.…”