What question did this study set out to answer?

The research aims to link the molecular interfacial structure of elastomers to their macroscopic mechanical performance.

February 28, 2026Open Access

Bound Rubber as a Transferable Structural Descriptor: Connecting MD-Derived Interfacial Scaling to Continuum Reinforcement Models

Key Points

The research aims to link the molecular interfacial structure of elastomers to their macroscopic mechanical performance.
Utilized a hierarchical multiscale framework combining coarse-grained molecular dynamics (MD) and dynamic mechanical analysis (DMA).
Investigated a scaling relation for chain length in coarse-grained simulations to establish structural descriptors.
Applied statistical models to compare macro-rheological predictions using different datasets.
Identified a significant bound-layer scaling relation in MD simulations for N=50 chains.
Demonstrated that the single-phase fractional Maxwell model is statistically preferred for specific datasets.
Achieved a reduction in error for cross-scale prediction using an MD-derived effective-volume-fraction baseline, achieving 7.3% error in bridging models.

Abstract

Filled elastomers often exhibit a low-frequency power-law storage modulus (G-prime), yet quantitative links between molecular interfacial structure and macroscopic reinforcement remain unresolved. This gap is addressed using a hierarchical multiscale framework that integrates coarse-grained molecular dynamics (MD) and dynamic mechanical analysis (DMA). Overall, MD contributes transferable structural descriptors rather than direct macro-rheology prediction. MD simulations yield a bound-layer scaling relation for chain length N=50 in coarse-grained simulations serving as a structural probe. For EPDM master curves, the single-phase fractional Maxwell model is statistically preferred (Delta AICc > 147, n = 56), reflecting limited statistical power; larger datasets (e.g., PC/ABS, n = 952) favor the dual-phase formulation. For cross-scale prediction, an MD-derived effective-volume-fraction baseline (MAPE = 54.1%) provides a structural prior; the regime-partitioned bridge model absorbs relaxation physics not resolved at the MD scale, reducing error to 7.3% (blocked-CV MAPE = 9.5%, with a 2.3% fold-to-fold spread). Linear-viscoelastic constraints improve nonlinear PTT calibration, reducing die-swell error by 87%.

AIに質問

Bookmark

View Full Paper