Abstract:In this paper, a thermomechanical constitutive model was developed for the time-dependent behaviors of the glass transition of amorphous networks. The model used multiple discrete relaxation processes to describe the distribution of relaxation times for stress relaxation, structural relaxation, and stress-activated viscous flow. A non-equilibrium thermodynamic framework based on the fictive temperature was introduced to demonstrate the thermodynamic consistency of the constitutive theory. Experimental and theoretical methods were developed to determine the parameters describing the distribution of stress and structural relaxation times and the dependence of the relaxation times on temperature, structure, and driving stress. The model was applied to study the effects of deformation temperatures and physical aging on the shape-memory behavior of amorphous networks. The model was able to reproduce important features of the partially constrained recovery response observed in experiments. Specifically, the model demonstrated a strain-recovery overshoot for cases programmed below Tg and subjected to a constant mechanical load. This phenomenon was not observed for materials programmed above Tg. Physical aging, in which the material was annealed for an extended period of time below Tg, shifted the activation of strain recovery to higher temperatures and increased significantly the initial recovery rate. For fixed-strain recovery, the model showed a larger overshoot in the stress response for cases programmed below Tg, which was consistent with previous experimental observations. Altogether, this work demonstrates how an understanding of the time-dependent behaviors of the glass transition can be used to tailor the temperature and deformation history of the shape-memory programming process to achieve more complex shape recovery pathways, faster recovery responses, and larger activation stresses.

Toward Interpreting the Thermally Activated β Dynamics in Metallic Glass Using the Structural Constraint Neural Network

Towards interpreting the thermally activated $β$ dynamics in metallic glass with the structural constraint neural network

Modeling the Glass Transition of Amorphous Networks for Shape-Memory Behavior

Machine learning atomic dynamics to unfold the origin of plasticity in metallic glasses: From thermo- to acousto-plastic flow

Enhancing the Prediction of Glass Dynamics by Incorporating the Direction of Deviation from Equilibrium Positions

Learning molecular dynamics: predicting the dynamics of glasses by a machine learning simulator

What Do Deep Neural Networks Find in Disordered Structures of Glasses?

Machine-learning integrated glassy defect from an intricate configurational-thermodynamic-dynamic space

Predicting dynamic heterogeneity in glass-forming liquids by physics-inspired machine learning

Dynamic heterogeneity at the experimental glass transition predicted by transferable machine learning

Prediction and Interpretability of Glass Transition Temperature of Homopolymers by Data-Augmented Graph Convolutional Neural Networks

Rotation-equivariant Graph Neural Networks for Learning Glassy Liquids Representations

Graph Neural Networks for an Accurate and Interpretable Prediction of the Properties of Polycrystalline Materials

Machine learning atomic-scale stiffness in metallic glass

Predicting orientation-dependent plastic susceptibility from static structure in amorphous solids via deep learning

Machine-learning improves understanding of glass formation in metallic systems

Prediction of Glass Forming Ability of Bulk Metallic Glasses Using Machine Learning

Inverse design of glass structure with deep graph neural networks

Precision Glass Thermoforming Assisted by Neural Networks

Extracting governing system for the plastic deformation of metallic glasses using machine learning