RT @_Drayerv2: (the answer is no). read here: feel free to DM if you cannot access!

These findings are inconsistent with chain end dilution models of the molecular weight Tg dependence. They suggest a need to reopen the problem of the Tg molecular weight dependence, with attention given to more modern theories of Tg and its dependence on molecular weight.

At least one model, a substantial Tg dependence on molecular weight is observed even when the ends are dynamically identical to the interior. Even in models where a chain end effect is observed, its importance weakens on cooling such that it has little effect on Tg.

We used MD simulations to ask a foundational question - are chain ends actually more mobile than chain interiors, and is this effect big enough to account for the molecular weight dependence of Tg? It turns out that, across multiple models, the answer is 'no'.

Chain ends are thus expected to act like plasticizers - the more there are, the faster the mean dynamics of the system due to averaging effects. We all learned these models in our introductory polymer classes - they are a core component of our education in polymer science.

For decades, the Tg(M) dependence has been understood based on chain end dilution models. Their basic idea is that chain ends are less dynamically constrained, and thus more mobile, than chain interiors.

Very excited about this new paper from our group (former student @_Drayerv2 ), on the origins of the molecular weight dependence of Tg. We think these findings will ultimate require the textbook understanding of this trend to be rewritten. 🧵.

RT @danteshepherd: The Boston Globe printed my letter today:

RT @tarak_patra: Our new paper in @Nature with @sanatkk and @TomislavRovis, and Eugene Chen reports a new method that compatibilizes immisc….

RT @pierrekawak: Pre-print out! So proud of the culmination of my PhD work in this paper.

The shorter-range gradient is of mixed nature, combining elastic effects with a gradient in segmental localization length scale. Our results thus point to a dual-barrier mechanism of glass-formation, with contributions from local caging and from long-ranged collective elasticity.

This finding provides compelling new support for the physics underlying the ECNLE model of glass formation. In essence, this long-range 1/z gradient is a direct signature of collective elastic physics within the ECNLE theory.

and predicted by the ECNLE theory at a remarkably detailed level and without adjustment.

tail on the log(relaxation time) gradient at large distances from the interface. Here our simulations reveal this tail for the first time. New more detailed analytic and numerical ECNLE predictions from Anh Phan and Ken Schweizer demonstrate that this is all consistent with. .

In this study, our group developed a strategy to simulate these gradients much further from the interface. This was inspired by a prediction several years ago by Ken Schweizer and coworkers, based on their ECNLE theory of glass formation, that there should be a 1/z power law. .

The goal has been to find a testable signature of the physics underlying glass formation in this gradient shape. It has been observed that the gradient in the segmental relaxation time near interfaces is double-exponential . This has considerable theoretical down-selection power.

More context: there has long been hope that dynamical correlations believed to underlie the glass transition might reveal themselves near interfaces. We and many others have been studying the shape of gradients of altered dynamics near interfaces in part for this reason.

Our new paper in @NaturePhysics, with K. Schweizer and A. Phan, reveals a signature of a combined caging and elasticity origin of the #glasstransition within dynamics at the surface of glass-forming liquids. Surface gradients have long-range tails! 🧵.