Temperature-Dependent Pre-nucleation Dynamics in Polymer Melts Using Event-Aligned Simulation Analysis

Polymer nucleation remains difficult to characterize because precursor states are highly localized, transient, and structurally heterogeneous, and these features collectively govern the emergence of the crystalline nucleus. Previous molecular dynamics studies of polyethylene have shown that the organization of pre-nucleation involves local chain alignment, conformational straightening, and densification and that the structural motifs at the growth front vary with the prevailing kinetic regime. Based on these observations, we establish an event-aligned, lead-time-resolved analytical workflow that connects mechanistic descriptors to temperature-dependent precursor dynamics. We first quantify nucleation kinetics across a series of supercooled temperatures to determine how thermal driving forces modulate the onset of ordering. We then evaluate conventional structural parameters by aligning the trajectories to the nucleation event and comparing their precursor signatures. To probe spatially resolved configurational organization, we employ directional entropy band (DEB) features that encode local anisotropic entropy gradients across concentric radial shells. Low-dimensional embeddings of these entropy band features reveal a progressive separation between atoms belonging to the incipient nucleus and those in the surrounding melt well before nucleation occurs. This structural separation systematically intensifies as the nucleation event approaches. Lead-time analysis further shows that the emergence of strong precursor differentiation shifts to earlier times with increasing undercooling, indicating that deeper undercooling enhances the temporal window over which nucleation-relevant ordering can be detected. Finite-time chain-mobility analysis indicates that accelerated melt dynamics contributes to—but does not fully account for—the observed temperature dependence of precursor differentiation, consistent with an additional role for cooperative structural organization in shaping early-stage nucleation pathways. Complementary supervised classification analyses further demonstrate that DEB features retain predictive discriminability at controlled lead times prior to nucleation onset, with the predictive horizon contracting systematically as the temperature increases. Direct comparison with the combined conventional descriptors {${\bar{S}}_i$⁠, p₂, q₆} reveals the same temperature-dependent reduction of the precursor horizon. However, conditional-association tests and nested-model evaluations show that outer-shell DEB features preserve complementary predictive information beyond these scalar measures, highlighting their ability to capture spatially distributed configurational cues not encoded in traditional order parameters. In general, the event-aligned structural framework introduced here provides a reproducible methodology to quantify precursor dynamics and establishes a foundation for relational and spatiotemporal modeling of polymer nucleation pathways.