Track synthesis, materials and catalysis as it is published.
Search papers, labs, and topics across Lattice.
This paper investigates the time reparameterization hypothesis for Swap Monte Carlo in a polydisperse glass former, finding that while collective overlaps exhibit approximate time reparameterization, the associated clock is dependent on probe length and relaxation stage. This dependence implies that swap dynamics do not define a unique, observable-independent material time, and importantly, alter the length-scale dependence of relaxation within the accessible $\alpha$-relaxation regime. Consequently, the collective reparameterization is only approximate, challenging the notion of exact dynamical equivalence between swap and reference relaxation.
Swap Monte Carlo alters the length-scale dependence of liquid relaxation, meaning its accelerated dynamics don't map to a single, universal material time.
Swap Monte Carlo dramatically accelerates the equilibration of deeply supercooled liquids while preserving the equilibrium Boltzmann distribution. Recent work has argued that swap and other accelerated algorithms act by reparameterizing time: different dynamics follow the same slow relaxation process, but with different clocks. Here we test this idea in a continuously polydisperse glass former. We confirm that mean collective overlaps satisfy an approximate time reparameterization, in agreement with earlier results. However, the associated clock depends systematically on the collective probe length and on the relaxation stage. Thus the observed collective collapse does not define a unique, observable-independent material time. Over the numerically accessible $α$-relaxation regime, swap changes not only the rate of relaxation, but also its length-scale dependence. The resulting collective reparameterization is therefore only approximate and does not establish an exact dynamical equivalence between swap and reference relaxation.