Geometric Encoding of Cyclic Mechanical Fatigue in Model Glasses: Graph Neural Networks as Structural Probes for Battery-Relevant Degradation

Solid-state battery glass electrolytes degrade under repeated charge and discharge cycling, yet the atomic-scale structural changes underlying this fatigue remain poorly characterised. Here we investigate whether cyclic mechanical fatigue—simulated as oscillatory volumetric strain applied to a Lennard-Jones model glass—encodes a learnable geometric signature detectable by graph neural networks (GNNs). Using a GATv2 architecture and a systematic four-way ablation study, we find that classification between pristine (cycle 0) and heavily fatigued (cycles 300–400) glasses reaches 91.67 ± 4.22% accuracy using 8-dimensional bond-length statistics as node features, rising to 98.33 ± 1.49% (AUC = 0.9975 ± 0.0022) when per-sample instance normalisation is applied. The improvement under normalisation—which explicitly removes all global magnitude information—establishes that the discriminative signal resides in the relational geometry of local bond networks, not in trivial density or volume changes. Ablating distribution-shape features (skewness, quartiles) reduces accuracy by 5.3 percentage points and collapses cycle-number regression to near-random performance (R² = 0.006), confirming that asymmetric bond-length damage is the dominant structural fingerprint of fatigue. Continuous degradation tracking via regression achieves R² = 0.311 ± 0.039 under instance normalisation, with all five cross-validation folds converging to R² ≈ 0.28–0.32 given extended training. Structural analysis reveals that fatigue accumulates as monotonically increasing bond-length disorder that saturates around cycle 300, consistent with a metastable fatigued equilibrium. Per-atom energy validation reveals a striking inversion: the dominant classification feature (r_max, ΔAUCdrop = 0.601) correlates negligibly with local potential energy (Spearman ρ = −0.093), while the strongest energy proxy (r_min, ρ = −0.812) ranks seventh in fatigue discrimination—establishing that the GATv2 probe detects irreversible structural history encoded in stretched-bond geometry rather than the instantaneous energy state. A Kovacs memory effect experiment further validates the 8D feature vector as a genuine structural probe: bond-length standard deviation and skewness track the NVT Kovacs relaxation signal with the same timescale as the macroscopic energy recovery, while mean non-affine displacement continues to grow linearly after bulk equilibration, confirming that the GNN features occupy the physically privileged space between thermodynamic observables and full microscopic configuration. These findings extend the geometric-encoding framework established for thermal history to mechanical cycling history, and provide a methodological foundation for applying GNN-based structural probes to experimentally characterised solid electrolyte fatigue.