RESEARCH

I. Multiscale Multiphysics Constitutive Modeling

Mechanisms-based constitutive modeling serves as a critical link between solid mechanics and materials physics. Our group develops predictive models that not only describe complex material behaviors but also reveal the underlying physical mechanisms across multiple scales. This modeling framework provides reliable tools for interpreting nonlinear, time-dependent, and coupled responses in materials, which in turn supports material innovation in areas such as intelligent materials, soft robotics, and high-performance composites. 

We have constructed constitutive models for a wide spectrum of polymeric and soft materials, including carbon black-filled rubbers, semi-crystalline polyethylene, liquid crystal elastomers, double-network hydrogels, and polymers with transient networks. These models are derived from micromechanical principles at the chain and molecular level and are rigorously upscaled to the macroscopic level using statistical mechanics. They accurately capture diverse behaviors such as viscoelasticity, plasticity, anisotropy, self-healing, and thermo-electro-mechanical coupling, while also reflecting key multiscale phenomena—e.g., strain-induced crystallization, chain scission, dynamic bond reformation, mesogen rotation, and morphology evolution. Complementing these efforts, we develop multiscale and multiphysics numerical tools, including user-defined subroutines for commercial finite element platforms, enabling simulation-driven material design and the exploration of complex physical processes.

II. Fracture of Soft Materials

Soft materials—such as elastomers, hydrogels, and soft composites—play a central role in cutting-edge technologies including stretchable electronics, soft robotics, tissue engineering, and biomedical devices. However, their applications are often limited by their vulnerability to fracture during fabrication and operation. Understanding and predicting fracture in soft materials is inherently challenging due to their large deformations, nonlinear time-dependent behavior, and strong coupling between mechanical, thermal, and chemical processes. Fracture mechanisms often originate at molecular or network levels and manifest at macroscopic scales in experimental and practical scenarios.

Our research addresses the fracture initiation and propagation in a broad class of soft materials exhibiting viscoelasticity, anisotropic damage, phase transitions, and multiphysics interactions. These include conventional rubbers, liquid crystal elastomers (LCEs), carbon fiber-reinforced composites, and mechanochromic polymers. We have developed a fatigue fracture theory for viscoelastic solids, including a dynamic scaling law that explains why cracks grow significantly faster under cyclic loading than under static loading. This theory has been recognized and adopted by Michelin’s R&D division for improving crack resistance in next-generation tires. Additionally, we investigate fracture mechanisms in LCEs by uncovering the interaction between energy dissipation and mesogen rotation at crack tips, offering fundamental insights for designing tougher, more reliable soft robotic actuators and functional materials.

III. Fatigue Assessment and Life Prediction

Fatigue is one of the most critical failure modes in structural materials, yet conventional fatigue testing methods are time-consuming and resource-intensive—often requiring weeks or months of repetitive loading experiments. To meet the growing demand for faster and more efficient fatigue characterization in modern industry, our research explores alternative methods rooted in the physics of energy dissipation and self-heating phenomena during fatigue.

Our work establishes a new framework for rapid fatigue evaluation by linking microscale damage mechanisms with measurable thermal responses. We have developed a highly accurate estimation model for intrinsic dissipation in fatigue-prone materials, utilizing a dual-exponential regression of temperature rise while accounting for environmental influences. Building on this, we proposed a novel energy-based methodology that leverages intrinsic dissipation as a reliable fatigue damage indicator. Remarkably, this approach enables high-cycle fatigue assessment within 24 hours using only two specimens. Furthermore, we introduced a theoretical model elucidating the origin of fatigue-induced dissipation, grounded in dislocation and point defect dynamics. This model provides critical insights for advancing fast fatigue evaluations of metals via self-heating analysis.

IV. Reduced Order Modeling of Lithium-Ion Batteries

We focus on developing multiscale reduced-order models to simulate the mechanical behavior of lithium-ion batteries (LIBs) subjected to mechanical abuse, such as crushing, indentation, or penetration. This research addresses the anisotropic, rate-dependent, and temperature-sensitive responses of various LIB components, including active particles, binders, current collectors, and separators. To manage the strong heterogeneity of constituent behaviors, we have restructured the eigenstrain-based reduced-order homogenization method to accommodate distinct constitutive models across different phases. This approach enables rapid yet detailed simulations, achieving up to two orders of magnitude speedup while preserving accuracy in damage evolution and fracture prediction.

In parallel, we are developing a dedicated multiscale and multiphysics framework for modeling the LIB separator—a porous polymer membrane critical to battery integrity. This model captures key chemo-mechanical coupling effects, including electrolyte interaction, porosity evolution, and swelling-induced stresses. By incorporating these coupled phenomena, our framework offers a more realistic depiction of internal mechanical responses and failure risks within LIBs. These advancements provide powerful tools for understanding battery safety and guiding the design of more robust energy storage systems.