Deep-learning-based inverse design of three-dimensional architected cellular materials with the target porosity and stiffness using voxelized Voronoi lattices.

Architected cellular materials are a class of artificial materials with cellular architecture-dependent properties. Typically, designing cellular architectures paves the way to generate architected cellular materials with specific properties. However, most previous studies have primarily focused on a forward design strategy, wherein a geometry is generated using computer-aided design modeling, and its properties are investigated experimentally or via simulations. In this study, we developed an inverse design framework for a disordered architected cellular material (Voronoi lattices) using deep learning. This inverse design framework is a three-dimensional conditional generative adversarial network (3D-CGAN) trained based on supervised learning using a dataset consisting of voxelized Voronoi lattices and their corresponding relative densities and Young's moduli. A well-trained 3D-CGAN adopts variational sampling to generate multiple distinct Voronoi lattices with the target relative density and Young's modulus. Consequently, the mechanical properties of the 3D-CGAN generated Voronoi lattices are validated through uniaxial compression tests and finite element simulations. The inverse design framework demonstrates potential for use in bone implants, where scaffold implants can be automatically generated with the target relative density and Young's modulus.

Light-Induced Topological Patterning toward 3D Shape-Reconfigurable Origami.

Shape-reconfigurable materials are crucial in many engineering applications. However, because of their isotropic deformability, they often require complex molding equipment for shaping. A polymeric origami structure that follows predetermined deformed and non-deformed patterns at specific temperatures without molding is demonstrated. It is constructed with a heterogeneous (dynamic and static) network topology via light-induced programming. The corresponding spatio-selective thermal plasticity creates varied deformability within a single polymer. The kinematics of site-specific deformation allows guided origami deployment in response to external forces. Moreover, the self-locking origami can fix its geometry in specific states without pressurization. These features enable the development of shape-reconfigurable structures that undergo on-demand geometry changes without requiring bulky or heavy equipment. The concept enriches polymer origamis, and could be applied with other polymers having similar chemistries. Overall, it is a versatile material for artificial muscles, origami robotics, and non-volatile mechanical memory devices.

Topological alternation from structurally adaptable to mechanically stable crosslinked polymer.

Stimuli-responsive polymers with complicated but controllable shape-morphing behaviors are critically desirable in several engineering fields. Among the various shape-morphing materials, cross-linked polymers with exchangeable bonds in dynamic network topology can undergo on-demand geometric change via solid-state plasticity while maintaining the advantageous properties of cross-linked polymers. However, these dynamic polymers are susceptible to creep deformation that results in their dimensional instability, a highly undesirable drawback that limits their service longevity and applications. Inspired by the natural ice strategy, which realizes creep reduction using crystal structure transformation, we evaluate a dynamic cross-linked polymer with tunable creep behavior through topological alternation. This alternation mechanism uses the thermally triggered disulfide-ene reaction to convert the network topology - from dynamic to static - in a polymerized bulk material. Thus, such a dynamic polymer can exhibit topological rearrangement for thermal plasticity at 130°C to resemble typical dynamic cross-linked polymers. Following the topological alternation at 180°C, the formation of a static topology reduces creep deformation by more than 85% in the same polymer. Owing to temperature-dependent selectivity, our cross-linked polymer exhibits a shape-morphing ability while enhancing its creep resistance for dimensional stability and service longevity after sequentially topological alternation. Our design enriches the design of dynamic covalent polymers, which potentially expands their utility in fabricating geometrically sophisticated multifunctional devices.

Deep Learning in Mechanical Metamaterials: From Prediction and Generation to Inverse Design.

Mechanical metamaterials are meticulously designed structures with exceptional mechanical properties determined by their microstructures and constituent materials. Tailoring their material and geometric distribution unlocks the potential to achieve unprecedented bulk properties and functions. However, current mechanical metamaterial design considerably relies on experienced designers' inspiration through trial and error, while investigating their mechanical properties and responses entails time-consuming mechanical testing or computationally expensive simulations. Nevertheless, recent advancements in deep learning have revolutionized the design process of mechanical metamaterials, enabling property prediction and geometry generation without prior knowledge. Furthermore, deep generative models can transform conventional forward design into inverse design. Many recent studies on the implementation of deep learning in mechanical metamaterials are highly specialized, and their pros and cons may not be immediately evident. This critical review provides a comprehensive overview of the capabilities of deep learning in property prediction, geometry generation, and inverse design of mechanical metamaterials. Additionally, this review highlights the potential of leveraging deep learning to create universally applicable datasets, intelligently designed metamaterials, and material intelligence. This article is expected to be valuable not only to researchers working on mechanical metamaterials but also those in the field of materials informatics.

Simulation of Abnormal Grain Growth Using the Cellular Automaton Method.

The abnormal grain growth of steel, which is occurs during carburization, adversely affects properties such as heat treatment deformation and fatigue strength. This study aimed to control abnormal grain growth by controlling the materials and processes. Thus, it was necessary to investigate the effects of microstructure, precipitation, and heat treatment conditions on abnormal grain growth. We simulated abnormal grain growth using the cellular automaton (CA) method. The simulations focused on the grain boundary anisotropy and dispersion of precipitates. We considered the effect of grain boundary misorientation on boundary energy and mobility. The dispersion state of the precipitates and its pinning effect were considered, and grain growth simulations were performed. The results showed that the CA simulation reproduced abnormal grain growth by emphasizing the grain boundary mobility and the influence of the dispersion state of the precipitate on the occurrence of abnormal grain growth. The study findings show that the CA method is a potential technique for the prediction of abnormal grain growth.

Two- and Three-Dimensional Modeling and Simulations of Grain Growth Behavior in Dual-Phase Steel Using Monte Carlo and Machine Learning.

Dual-phase (DP) steel has been widely used in automotive steel plates with a balance of excellent strength and ductility. Grain refinement in DP steel is important to improve the properties further; however, the factors affecting grain growth need to be well understood. The remaining problem is that acquiring data through experiments is still time-consuming and difficult to evaluate quantitatively. With the development of materials informatics in recent years, material development time and costs are expected to be significantly reduced through experimentation, simulation, and machine learning. In this study, grain growth behavior in DP steel was studied using two-dimensional (2D) and three-dimensional (3D) Monte Carlo modeling and simulation to estimate the effect of some key parameters. Grain growth can be suppressed when the grain boundary energy is greater than the phase boundary energy. When the volume fractions of the matrix and the second phase were equal, the suppression of grain growth became obvious. The long-distance diffuse frequency can promote grain growth significantly. The simulation results allow us to better understand the factors affecting grain growth behavior in DP steel. Machine learning was performed to conduct a sensitivity analysis of the affecting parameters and estimate the magnitude of each parameter's effects on grain growth in the model. Combining MC simulation and machine learning will provide one promising research strategy to gain deeper insights into grain growth behaviors in metallic materials and accelerate the research process.

Evaluation of Austenite-Ferrite Phase Transformation in Carbon Steel Using Bayesian Optimized Cellular Automaton Simulation.

Austenite-ferrite phase transformation is a crucial metallurgical tool to tailor the properties of steels required for particular applications. Extensive simulation and modeling studies have been conducted to evaluate the phase transformation behaviors; however, some fundamental physical parameters still need to be optimized for better understanding. In this study, the austenite-ferrite phase transformation was evaluated in carbon steels with three carbon concentrations during isothermal annealing at various temperatures using a developed cellular automaton simulation model combined with Bayesian optimization. The simulation results show that the incubation period for nucleation is an essential factor that needs to be considered during austenite-ferrite phase transformation simulation. The incubation period constant is mainly affected by carbon concentration and the optimized values have been obtained as 10-24, 10-19, and 10-21 corresponding to carbon concentrations of 0.2 wt%, 0.35 wt%, and 0.5 wt%, respectively. The average ferrite grain size after phase transformation completion could decrease with the decreasing initial austenite grain size. Some other parameters were also analyzed in detail. The developed cellular automaton simulation model combined with Bayesian optimization in this study could conduct an in-depth exploration of critical and optimal parameters and provide deeper insights into understanding the fundamental physical characteristics during austenite-ferrite phase transformation.