Ring polymers that are knotted form melts with tunable glass-forming properties. The tuning parameter is the knot: trefoils, figure-eights, and more complex topologies each produce a different glass transition temperature, a different fragility, and a different cooperativity length scale — all from the same monomer chemistry at the same molecular weight.
The paper demonstrates through molecular dynamics that multiple theoretical models of glass formation — the string model (cooperative rearrangements as string-like excitations), the localization model (caging by neighbors), and the shoving model (elastic cost of local expansion) — all successfully describe the segmental relaxation across the full range of knot complexities. Quantitative fits to all three models work simultaneously, with each model's characteristic parameters varying systematically with knot complexity.
The surprise is not that one model works but that all three do. These models emphasize different physical mechanisms — cooperative motion, localization, elastic response — yet they track each other across the knot series. More complex knots produce tighter localization, longer cooperative strings, and larger shoving barriers, and the three measures remain mutually consistent.
The implication: the three models may not be competitors describing different physics. They may be different projections of a single underlying phenomenon. Knot topology provides the control parameter that varies all three simultaneously, revealing their common dependence on a quantity — the topological constraint density — that none of the models explicitly contains. The glass transition has one mechanism viewed from three angles.