Experimental test of a predicted dynamics–structure–thermodynamics connection in molecularly complex glass-forming liquids

Abstract

Understanding in a unified manner the generic and chemically specific aspects of activated dynamics in diverse glass-forming liquids over 14 or more decades in time is a grand challenge in condensed matter physics, physical chemistry, and materials science and engineering. Large families of conceptually distinct models have postulated a causal connection with qualitatively different “order parameters” including various measures of structure, free volume, thermodynamic properties, short or intermediate time dynamics, and mechanical properties. Construction of a predictive theory that covers both the noncooperative and cooperative activated relaxation regimes remains elusive. Here, we test using solely experimental data a recent microscopic dynamical theory prediction that although activated relaxation is a spatially coupled local–nonlocal event with barriers quantified by local pair structure, it can also be understood based on the dimensionless compressibility via an equilibrium statistical mechanics connection between thermodynamics and structure. This prediction is found to be consistent with observations on diverse fragile molecular liquids under isobaric and isochoric conditions and provides a different conceptual view of the global relaxation map. As a corollary, a theoretical basis is established for the structural relaxation time scale growing exponentially with inverse temperature to a high power, consistent with experiments in the deeply supercooled regime. A criterion for the irrelevance of collective elasticity effects is deduced and shown to be consistent with viscous flow in low-fragility inorganic network-forming melts. Finally, implications for relaxation in the equilibrated deep glass state are briefly considered.

Some open challenges in polymer physics*

“…We have also evaluated [ 64 ] two “nondiverging” models that do not use a dependence of the dynamics on the entropy or configurational entropy: the Dyre shoving model [ 65–67 ] and the elastically collective nonlinear Langevin equation theory from Schweizer. [ 68–72 ] We find in those cases that the model predictions for the dynamics deviate strongly from the VFT/WLF temperature dependence, but that the experimental data still seem to fall below the theoretical predictions, similar to what was seen with the best‐case situations for the entropy‐based models. It was suggested [ 64 ] that there may be a need to incorporate additional mechanisms in the models to account for the greater than predicted molecular mobility seen in the experimental data.…”

Experimental Tests of a Theoretically Predicted Noncausal Correlation between Dynamics and Thermodynamics in Glass-forming Polymer Melts

“…Despite the system-specific complexities of real molecules compared to hard spheres, the alpha relaxation time data were shown to accord well with the theoretical prediction. 48 Agreement was also found for orthoterphenyl (OTP) liquids at fixed high pressures, for salol under isochoric conditions, and for network-forming strong glass formers (silica and boron oxide). Consequences of the idea that S 0 is operationally the key thermodynamic variable were shown to have major implications in T-space.…”

“…This includes the apparent inverse power law (with a large exponent) scaling of the effective activation barrier with temperature in deeply supercooled liquids, which was confirmed for ∼40 molecular liquids. 48 Given that the predicted dynamics−thermodynamics connection within ECNLE theory is noncausal and deduced based on the simple hard-sphere model, this level of experiment−theory agreement for chemically complex glass formers is perhaps surprising. We view it as suggesting that (i) there is a significant degree of "self-averaging" of chemical complexity in supercooled liquid relaxation and (ii) a thermodynamics-based picture of activated dynamics can "work" even though it is not causal because of the exact connections between the local structural pair correlation function and a specific thermodynamic property.…”

Understanding Creep Suppression Mechanism in Polymer Nanocomposites through Machine Learning

“…27,28 Both barriers are higher near rough surfaces 36 and decrease with external stress, 43 while the long range elastic barrier is more sensitive to stress. 36,44,45 However, the barriers in our model have a distinct microscopic origin. In ECNLE, the local cage and the elastic barrier are causally related, since the elastic barrier originates from the local cage expansion.…”