Session R02: Molecular Glasses and Liquids

Live

Computer simulations give unique insights into the microscopic behavior of amorphous materials. It became recently possible to generate ultrastable glass configurations using a simple Monte Carlo algorithm for a broad variety of model glass-formers. In this talk, we show that this discovery has allowed a deeper understanding of the rheological, thermodynamic and thermal properties of amorphous solids.   

Live

Stable glasses (SGs) are highly dense, low energy glasses that are made by physical vapor deposition (PVD). Because of their desirable properties and potential for wide range of applications, they are extensively studied glasses. However, estimating their dynamics can be difficult since they are produced directly into the SG state during PVD. We introduce a novel and indirect method of measuring the dynamics of SGs using solvent vapor annealing (SVA), combined with in-situ ellipsometry. Using this technique, we have shown that the mechanism by which solvent penetrates a molecular glass film, depends on the solvent vapor pressure. Above a certain threshold of the solvent vapor pressure, the swelling front moves linearly into the film in a Case II diffusion process. By monitoring the front’s motion across the film, we can compare the dynamics of SG films of various stability produced at different deposition temperatures. The results correlate well with thermal transformation growth fronts but in this case, can be used to characterize glasses with a wider range of stability and with more accuracy. The SVA process is also able to provide more details of the interface dynamics and their differences compared to the bulk glass.   

Live

Glass films created using physical vapor deposition (PVD) have been recently shown to exhibit enhanced stability properties as compared to ordinary glasses formed via standard liquid-quenching. However, the mechanism behind the formation of PVD stable glasses has yet to be fully uncovered. In this work, we use molecular dynamics simulations to study a coarse-grained model glass-former based on the molecule 9-(3,5-di(naphthalen-1-yl)phenyl)anthracene (α,α-A), within which we vary the strength of the dihedral potential on the respective side-groups to systematically study the effects of intramolecular degrees of freedom on molecular packing. By studying the PVD process as it is occurring, we've observed a denser, lower-energy free surface state that gradually expands into its bulk-like state as the film is formed. By looking at PVD glass formation with this new perspective, we can gain insight into the role of the free surface, the substrate, and the effect each has on molecular packing and mobility, both in and out of plane.   

Live

Understanding the structure of vapor deposited (PVD) glasses at buried interfaces is important for organic electronic applications. Here, we combine coarse-grained molecular dynamics (MD) simulation and experiments to study the molecular structures at the buried interfaces of PVD glass of an organic semiconductor, DSA-Ph (1,4-di-[4-(N,N-diphenyl)amino]styrylbenzene), on different substrates. For simulations, custom interaction potentials are used to control anchoring and orientation on different substrates. Simulation results show that the length scale over which the substrate perturbs the structure of a vapor-deposited glass is approximately 5σ (3 nm). In experiments, the structure of PVD glasses of DSA-Ph are examined as a function of film thickness with grazing incidence X-ray scattering. Experiments show that interfacial molecular packing is more disordered compared to the bulk, and they suggest that the film thickness at the buried interface that exhibits a modified molecular packing is ∼8 nm, consistent with the simulation results. In both simulations and experiments, the deposition temperature controls glass structure beyond this interfacial layer of a few nanometers.   

Live

Glasses age towards equilibrium in a manner characterized by signatures identified by Kovacs over 50 years ago. Here we probe the range of validity of these signatures via simulations of aging in polymer glasses, driven much farther from equilibrium and at higher temperatures than typically possible in experimental glasses. Simulated glasses exhibit Kovacs signatures of glassy aging only when both the temperature at which they depart equilibrium and their aging temperature are below the onset temperature T_A of non-Arrhenius equilibrium dynamics. For higher temperatures, Kovacs signatures are lost, signaling that the substance is not genuinely ‘glassy’ despite being removed from equilibrium. Below T_A, the we find that the timescale for aging in overdense glasses can be predicted with zero parameters from equilibrium behavior via an Arrhenius interpolation between T_A and the temperature of departure from equilibrium, again identifying T_A as key to understanding the nature of glasses. We compare these findings to expectations of aging behavior from the Tool-Narayanaswamy-Moynihan model.   

Live

In this communication we report on a promising candidate for a structural feature that correlates with the emergence of super-Arrhenian behavior in glass forming materials. MD simulation were carried out for a binary mixture of 80% AB and 20% CD dumbbells for a wide temperature range encompassing liquid and super-cooled liquid state, where the interaction potentials and particle diameters of the dumbbells are chosen to suppress crystallization. The relaxation time is evaluated via rotational and translational diffusion, where the relaxation time exhibits Arrhenian temperature dependence above T_A and super-Arrhenian dependence below T_A. The full orientational correlator F₀ for the long axes of the AB dumbbells was calculated, where the correlator has a value of unity if orientations of the molecules are perfectly random. At high temperatures above T_A F₀=1; however, at T_A the correlator F₀ departs from unity.   

Live

The dielectric loss spectrum of many glass-forming liquids exhibits an excess contribution in the high-frequency part of the α-relaxation peak, referred to as an ``excess wing’’. The physical origin of the excess wing has been long debated, but a microscopic understanding remains elusive.

By combining the swap Monte Carlo algorithm to long molecular dynamics simulations, we provide a clear microscopic picture for the equilibrium relaxation dynamics of a model atomic supercooled liquid, down to the experimental glass transition at T_g. Near T_g, structural relaxation is initiated by dilute clusters of mobile particles. The distribution of appearance time of these clusters displays a power law at times much shorter than the structural relaxation time, which is responsible for the emergence of an excess wing in the relaxation spectrum. We show that large dynamically correlated regions at τ_α originate from the small initial clusters forming the excess wing, via dynamic facilitation.

We devise a simple model based on these observations, and successfully reproduce the qualitative shape of dielectric loss spectra, including the excess wing.   

Live

Previous work has shown that machine learning can identify a local structural variable, called softness, which is predictive of particle-scale dynamics in many disordered systems. In simulations of supercooled liquids, this quantity has been associated with a local energy barrier to rearrangement, and has been found to be strongly descriptive of structural aging out of equilibrium, remaining predictive of particle rearrangements and the structural relaxation time throughout aging. Thus, a theory of how softness evolves in time makes predictions about the aging of a glass out of equilibrium. Here we develop a phenomenological model for how the softness of particles evolves in time in and out of equilibrium. We test the predictions of this model against the aging behaviour and temperature dependence of observables in our MD simulations of a Kob-Andersen Lennard-Jones glass.   

Live

There is a “crossover” in the nature of relaxation in the middle of the supercooled liquid range. We have re-analyzed millisecond-long simulations of o-terphenyl [J. Phys. Chem. B 117, 12898 (2013)] with new statistical methods that provide more molecular detail about this crossover. These simulations reach α-relaxation times τ_α that are two orders-of-magnitude slower than at the mode-coupling temperature T_MCT. The rotational Green’s function shows that as the temperature drops below T_MCT, the rotational-jump size lengthen, and an unexpected jump-precursor state emerges. The rate-exchange time τ_ex is measured using multidimensional correlation functions. It is below the α-relaxation time, τ_ex < τ_α , at T_MCT, but slows to τ_ex = τ_α at the lowest temperature simulated. The same methods applied to single-molecule data [Phys. Rev. E 98, 040603(R) (2018)] show that the slowing continues to τ_ex = 22 τ_α close to the glass transition. Overall, a broad crossover is seen, which involves multiple phenomena, including ones unique to the crossover region.   

Live

Determination of the mechanism for super-Arrhenian behavior of glass forming liquids is one of the unsolved problems in condensed matter physics. Recent developments using experimental data showed the connection between the relaxation time as a function of temperature and pressure, and the excess internal energy which is the difference between internal energy of a liquid and a crystal (Phy. Rev. Mat., 2, 055604, 2018). It would be valuable to explore this connection using simulation. Atomistic molecular models are too complex and thus too slow to cover a sufficiently wide time interval. Simplified models such as binary Lennard-Jones mixtures or Kremer-Grest polymer models do not have a crystalline phase. In this paper we report on the development of molecular models which are simple enough for fast simulations, but that also possess stable super-cooled liquid and crystalline phases. The strategy for obtaining such models is to coarse-grain (CG) atomistic models of known glass formers using the iterated Boltzmann inversion procedure. The resulting two- and three-bead CG models also exist in both the super-cooled liquid and crystalline states. The translational and rotational mobility as well as internal energy have been determined for a several of these new CG models.   

Live

It is held by many that relaxation in supercooled liquids and glasses occurs through some sort of cooperative mechanism. Indeed, cooperative relaxation events are ubiquitously identified in computer simulation, and seem to be associated with motion on a length scale of approximately 0.2σ, and a timescale of approximately 1 ps. We have unambiguously identified the signature of these collective relaxations in neutron and light scattering data, and quantified the fraction (Φ_0) of molecules instantaneously involved in these relaxation events [1-3]. We further show that these motions are indeed the most fundamental element of relaxation [3]. We will present evidence through the temperature dependence of Φ_0 that these fundamental relaxation events (upon which all other relaxation events are built) are bosonic excitations in the liquid or glass. If this proves to be correct, it suggests an approach for understanding relaxation of liquids and glass from a new perspective.

[1] Cicerone et al., Physical Review Letters, 113:11, 117801 (2014)
[2] Cicerone et al., The Journal of Chemical Physics 146:5, 054502 (2017)
[3] Bender et al., Soft Matter 120:34, 9103-9114 (2020)   

Live

Pressure-dependent structural relaxation times for glass-forming systems, τ(T,V), depend on independent contributions from both temperature (T) and volume (V). Our recent work shows that free volume, V_free = V − V_hc, where V_hc is the system volume at close-packing, dictates the independent volume contribution. Specifically, isotherms of lnτ vs. 1/V_free are linear, with slopes that increase with decreasing T. We explained this behavior mechanistically through the cooperative free volume rate model (CFV). One way to calculate V_free is by applying our LCL equation of state to analyze pressure, volume, temperature (PVT) data to obtain V_hc from its molecular parameters. However, there is a very physically transparent alternative route, which is presented in this talk. V_hc values can be obtained by simple linear extrapolation of PVT isobars to T = 0. This new route for calculating V_free is another way to show the strong connection that exists between dynamics and PVT measurements. We will discuss results for a dozen polymer and small molecule glass forming systems.   

Live

In water, the existence of at least two different forms of glass, i.e., the low-density amorphous (LDA) and the high-density amorphous (HDA) ices, and of one anomalous liquid, points to a hidden connection between these states, whose understanding has the potential to shed light on the complex nature of water’s behavior. Here, we develop a Neural Network scheme capable of discerning local structures beyond tetrahedrality. Applied over a wide region of the water’s phase diagram, we show that the local structures that characterize both LDA and HDA amorphous phases are indeed embedded in the supercooled liquid phase. Remarkably, the rapid increase in the LDA-like population with supercooling occurs in the same temperature and pressure region where thermodynamic fluctuations are maximized, linking these structures with water’s anomalies. At the same time, the population of HDA-like environments rapidly increases with pressure, becoming the majority component at high density. Our results show that both LDA and HDA are genuine glasses, and provide a microscopic connection between the non-equilibrium and equilibrium phase diagrams of water.

F.Martelli, F.Leoni, F.Sciortino, J.Russo "Connection between liquid and non-crystalline solid phases in water" J. Chem. Phys. 153, 104503 (2020)