H-Bond State Spectroscopy: A Bitstring Framework for Characterizing Noncovalent Interaction Networks in Polymers
Topic
When and Where
Session Chairs
Presenter(s)
Co-Author(s)
Abstract
Noncovalent interactions (NCIs) govern key macroscopic polymer properties — mechanical resilience, self-healing, stimulus response — yet no standardized computational descriptor exists that captures the full statistical complexity of intramolecular NCI networks across diverse polymer chemistries. We present H-Bond State Spectroscopy (HBSS), an AI-assisted simulation framework that converts NCI dynamics into discrete binary fingerprints amenable to information-theoretic analysis and kinetic modeling.
The pipeline operates in five automated stages. (1) Machine-learning force field pre-relaxation (MACE-OFF) generates chemically valid starting geometries. (2) Four parallel GFN2-xTB NVT MD replicas (100 ps each, 300 K) sample the conformational ensemble of single-chain oligomers (100–400 atoms). (3) HBSS encodes each trajectory frame as a binary bitstring over the top-N ranked H-bond contacts (donor–acceptor cutoff 3.5 Å, D–H–A angle ≥120°), producing per-frame NCI state vectors. (4) Classical Mutual Information (CMI) quantifies pairwise coupling between contact sites, revealing cooperative and competitive H-bond motifs. (5) Kinetic Monte Carlo (KMC) propagates the state distribution across macroscopic timescales using transition rates derived from MD-sampled populations.
Applied to a poly(ether-thiourea) hexamer, HBSS resolves 239 unique NCI states across 4,004 trajectory frames, yielding a Shannon entropy of 6.72 bits — a quantitative measure of conformational NCI diversity unavailable from single-structure or average-property analyses. Clustering via UMAP and HDBSCAN identifies 65 distinct H-bond cluster states with population-resolved bitstring signatures.
Screening of 215 polymer chemistries reveals that 74% lack classical H-bond donors, necessitating a tiered applicability framework before pipeline deployment. For the 26% that qualify, HBSS provides chain-length-resolved, chemistry-specific NCI fingerprints directly interpretable for rational design of self-healing, responsive, and high-performance polymer systems.
