Bridging atomistic simulations and experiments: A potential-energy-landscape-based analytical model for the strain-rate-dependent mechanical behaviour of epoxy resins

Nanoparticle-reinforced composites can improve stiffness, strength and thermal stability, but nanoscale defects such as particle agglomeration can reduce performance. This behaviour is hard to study because simulations and experiments work at very different strain rates. Conventional molecular dynamics (MD) is limited to very high strain rates and nanosecond times, while experiments use much lower rates. To reduce this gap, this work develops a coarse-grained potential-energy-landscape (PEL) framework for epoxy deformation. In this study, a PEL–Eyring model is proposed. In this model, autonomous basin climbing samples a spectrum of metastable stress states, and an Eyring-type activated-process model turns this spectrum into the rate- and temperature-dependent response. All model inputs are physical constants, independently reported literature values or PEL-derived quantities, and no parameter is fitted to the validation curves. The coarse-grained model is first checked against all-atom simulations at high strain rates and is then validated against independent low-rate tensile experiments. It reproduces both rate hardening and thermal softening, with yield-stress errors of 2.7–3.6%, normalized area errors below 0.10 and R 2 ≥ 0.938 across two strain rates and three temperatures. It also matches an independent tensile strength scale without adjustment. The work gives a parameter-free and physically clear link from molecular-scale PEL information to experimentally accessible mechanical behaviour. It shows that the nonlinear, rate- and temperature-dependent tensile response of epoxy can be predicted beyond the time limits of conventional MD.