Arrhenius temperature dependence of the crystallization time of deeply supercooled liquids

arXiv:2303.07817 · cond-mat.soft, cond-mat.dis-nn · Submitted 2023-03-14 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Condensed Matter. The day in condensed matter: materials, phases and the measurements under them.

Kai: We're looking at the paper: "Arrhenius temperature dependence of the crystallization time of deeply supercooled liquids".

Mira: Here's the gist of the paper. Arrhenius temperature dependence of the crystallization time of deeply supercooled liquids uses molecular dynamics simulations to show that while equilibrium relaxation time increases in a super-Arrhenius manner with decreasing temperature, the crystallization time shows an Arrhenius temperature dependence at very low temperatures This finding contradicts conventional formulas based on Classical Nucleation Theory (CNT) Furthermore, the estimated energy barrier for crystallization is surprisingly small compared to that for equilibrium dynamics.

Kai: Critique of Classical Nucleation Theory Classical nucleation theory (CNT) predicts a nonmonotonic temperature dependence for the nucleation time τN, which has a minimum value at a certain temperature.

Mira: The paper notes that the CNT predicts a similar nonmonotonic temperature dependence for the crystallization time τcry, which is the time required for crystalline regions to occupy a major part of the sample.

Kai: However, molecular dynamics simulations on soft spheres show that although equilibrium relaxation time depends on temperature in a super-Arrhenius manner, the crystallization time shows an Arrhenius-like temperature dependence at very low temperatures.

Paper summary: Kai: So to summarize what this paper on "Arrhenius temperature dependence of the crystallization time of deeply supercooled liquids" claims, they use molecular dynamics simulations on soft spheres right after an instantaneous quench from a high temperature to a low one.

Mira: The core thesis is that although equilibrium relaxation time increases in a super-Arrhenius manner as the temperature drops, the crystallization time shows an Arrhenius-like dependence at very low temperatures.

Lev: They found that this finding clearly contradicts equation two from classical nucleation theory, which predicts a different behavior for both relaxation and crystallization times.

Kai: On top of that, they showed that the energy barrier for this crystallization process is surprisingly small compared to the energy barrier associated with equilibrium dynamics of those liquids.

Mira: This points to a mechanism where crystallization happens much faster than the overall equilibration process in this specific low-temperature regime, which is what they are focusing on.

Lev: The paper also details how the crystallization process at low temperatures consists of an early stage where the system falls into its nearest inherent structure and a late stage involving transitions between these structures.

Kai: They argue that the late process of crystallization is what dominates the total time required for crystalline regions to take up a major part of the sample.

Mira: And they highlight that this coupling between aging and crystallization is essential for understanding rapid crystallization in deeply supercooled liquids or glasses.

Lev: This coupling leads them to conclude that the relaxation time in these systems can be quantitatively equivalent to the actual crystallization time when aging and crystallization are coupled.

Conclusion: Kai: So, looking at "Arrhenius temperature dependence of the crystallization time of deeply supercooled liquids," it really comes down to this paper showing how the dynamics don't follow classical nucleation theory predictions in this regime.

Mira: The main implication is that we need to look at the late process of crystallization and its coupling with aging when studying rapid crystallization in deeply supercooled liquids or glasses.

Lev: They establish that the exponential decay time of crystallization, tau exp, is the unique time scale for these intermittent dynamics in the low-temperature regime they studied.

Kai: This suggests that for fragile glass formers, crystallization can proceed very rapidly with an Arrhenius dependence even though they are fragile glass formers.

Mira: The work really focuses on how temperature acts as a key control parameter when studying atomic and molecular glasses where temperature is the key control parameter.

Lev: The paper also notes that its method only studies the instantaneous quench of a monodisperse system, which is an important limitation to keep in mind for applying these findings generally.

Yuki Takaha, *Hideyuki Mizuno, Atsushi Ikeda

Graduate School of Arts and Sciences, The University of Tokyo · Research Center for Complex Systems Biology, Universal Biology Institute, The University of Tokyo

cond-mat.soft, cond-mat.dis-nn

Submitted: 2023-03-14

Updated: 2023-03-14

Comments: 8 pages, 4 figures

DOI: 10.1103/PhysRevResearch.6.013040

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 85/100

The gist: Arrhenius temperature dependence of the crystallization time of deeply supercooled liquids uses molecular dynamics simulations to show that while equilibrium relaxation time increases in a…

Key concepts

Equilibrium Relaxation Time
This describes how long it takes for a deeply supercooled liquid to reach its stable, equilibrium state at a given temperature. It typically increases in a super-Arrhenius manner as the temperature drops, meaning relaxation slows down dramatically at lower temperatures.
Crystallization Time ($ au_{cry}$)
This is the time required for crystalline regions to occupy a significant portion of the liquid sample. The paper finds that at very low temperatures, this time exhibits an Arrhenius dependence, which differs from predictions made by Classical Nucleation Theory (CNT).
Coupling of Aging and Crystallization
This refers to the crucial link where the process of aging (structural relaxation) is accelerated by crystallization. This coupling means that the resulting relaxation time is quantitatively equivalent to the measured crystallization time, suggesting crystallization drives rapid structural change.
Energy Barrier ($ riangle E$)
This represents the energy required for a process to occur, calculated using an Arrhenius relationship ($ au ext{ ~ exp}( riangle E/T)$). The paper finds that the energy barrier for crystallization (0.16) is surprisingly small compared to the barriers for equilibrium relaxation (1.0 or 6.5), indicating rapid crystallization is energetically accessible.

Terminology

Summary

Arrhenius temperature dependence of the crystallization time of deeply supercooled liquids uses molecular dynamics simulations to show that while equilibrium relaxation time increases in a super-Arrhenius manner with decreasing temperature, the crystallization time shows an Arrhenius temperature dependence at very low temperatures This finding contradicts conventional formulas based on Classical Nucleation Theory (CNT) Furthermore, the estimated energy barrier for crystallization is surprisingly small compared to that for equilibrium dynamics. p. 0

Critique of Classical Nucleation Theory

Classical nucleation theory (CNT) predicts a nonmonotonic temperature dependence for the nucleation time τN, which has a minimum value at a certain temperature. The paper notes that the CNT predicts a similar nonmonotonic temperature dependence for the crystallization time τcry, which is the time required for crystalline regions to occupy a major part of the sample. p. 0 However, molecular dynamics simulations on soft spheres show that although equilibrium relaxation time depends on temperature in a super-Arrhenius manner, the crystallization time shows an Arrhenius-like temperature dependence at very low temperatures. p. 0 This is contrary to Eq. p. 0 (2).

Dynamics of Crystallization Processes

The crystallization process at low temperatures is shown to consist of an early process, in which the system falls into the nearest inherent structure, and a late process, in which transitions between inherent structures occur. The crystallization time is dominated by the late process. p. 1 The early process corresponds to dynamics in which the system falls into the nearest IS, and this is consistent with observations that the dynamics in the early process hardly depend on temperature. p. 4

Coupling of Aging and Crystallization

The paper demonstrates that a coupling between aging and crystallization is key for understanding rapid crystallization. The inequality τa τcry implies that the deeply supercooled liquid crystallizes before equilibration. p. 4 In the case of aging with crystallization, the relaxation is accelerated by crystallization, and the resulting relaxation time is quantitatively equivalent to the crystallization time. p. 6

Energy Barrier Analysis

The authors estimate characteristic energy barriers for these processes by fitting τ ∼ exp(∆E/T). The energy barrier ∆E for the crystallization is 0. p. 016 while ∆E for the equilibrium relaxation of the high-temperature and low-temperature liquids are 1. p. 10 and 6.5, respectively. p. 1 This reveals that the energy barrier for crystallization is surprisingly small compared to those for equilibrium dynamics. p. 0

Universal Nature of Dynamics

The intermittent crystallization observed in soft spheres at low temperature suggests that these peculiar crystallization dynamics are universal in deeply supercooled liquids or glasses. The work reveals several new aspects of this crystallization, including the role of temperature and the Arrhenius temperature dependence with a very small energy barrier. p. 6 The findings are relevant to understanding the crystallization of atomic and molecular glasses where temperature is the key control parameter. p. 6 The results suggest that crystallization in such systems proceeds very rapidly with an Arrhenius temperature dependence even if they are fragile glass formers. p. 6 The paper focuses solely on the instantaneous quench of the monodisperse system. p. 6 p. 6

Where the gains come from

The key finding is that a coupling between aging and crystallization is essential for understanding rapid crystallization in deeply supercooled liquids or glasses. This coupling leads to the relaxation time being quantitatively equivalent to the crystallization time. p. 6 The paper suggests that this coupling is the key for understanding rapid crystallization of glasses. p. 5 The results establish that intermittent crystallization takes place not only in hard spheres but also in soft spheres. p. 2

Pruning redundant reasoning tokens

The early process is not relevant for the crystallization time and the late process dominates the crystallization time. The relaxation times in the aging dynamics do not give a direct explanation of the crystallization time, as they are far from it. p. 5 The relaxation time τaging,mono has a similar temperature dependence as the crystallization time, and fitting it yields ∆E = 0. p. 516 which coincides with that for τcry. p. 5 This suggests that the crystallization itself accelerates the relaxation dynamics, and the resulting relaxation time gives the crystallization time. p. 5 The coupling between aging and crystallization is the key for understanding rapid crystallization of glasses. p. 0 p. 0 p. 0

How the training signal is built

The analysis involves measuring the overlap function FO(t) to monitor relaxation dynamics. The short-time relaxation is not due to vibrations but rearrangements of particles, characterized by a time scale τini,poly. p. 4 The long-time relaxation is characterized by a time scale τaging,poly. p. 5 The short-time relaxation in the monodisperse system coincides with that in the polydisperse system, suggesting crystallization does not affect the first rearrangements of particles. p. 5

The crystallization time τcry diverges not only at the freezing point Tfreezing = 0.59 but also at lower temperatures and thus has a minimum at a finite temperature. p. 4 This is qualitatively consistent with conventional crystallization theory. p. 4 However, at sufficiently low temperature, the crystallization time becomes much smaller than the equilibrium relaxation time τa. p. 4 The Arrhenius temperature dependence of the crystallization time is contrary to Eq. p. 0 (2). The energy barrier for crystallization is surprisingly small compared to those for equilibrium dynamics. p. 0

The coupling between aging and crystallization is the key for understanding rapid crystallization in deeply supercooled liquids or glasses. The results establish that the exponential decay time τexp is the unique time scale of crystallization in the low-temperature regime where intermittent crystallization takes place, and this time scale can be well captured by τcry. p. 3 The irrelevance of the lag time suggests that any deterministic process, such as crystal growth, is absent or at least is not a rate-controlling process in this temperature regime. p. 3 The crystallization itself accelerates the relaxation dynamics, and the resulting relaxation time gives the crystallization time. p. 5 p. 5 p. 5

The results establish that the exponential decay time τexp is the unique time scale of crystallization in the low-temperature regime where intermittent crystallization takes place. p. 3 The lag time τlag does not increase at the lowest temperatures, and as a result, τcry is again dominated by τexp. p. 3 p. 3 The crystallization itself accelerates the relaxation dynamics, and the resulting relaxation time gives the crystallization time. The coupling between aging and crystallization is the key for understanding rapid crystallization of glasses.

The simulation uses an inverse-power-law (IPL) potential to model soft spheres. The degree of crystallization X(t) is calculated as the proportion of solid-like particles at a given time t. p. 1 The survival probability P(t) is defined as the proportion of noncrystallized samples at time t. p. 2 The crystallization dynamics are studied after an instantaneous quench from a high temperature to low temperatures. p. 1

The paper's findings are relevant to understanding the crystallization of atomic and molecular glasses where temperature is the key control parameter. The work suggests that crystallization in such systems proceeds very rapidly with an Arrhenius temperature dependence even if they are fragile glass formers. p. 6 The paper focuses solely on the instantaneous quench of the monodisperse system. The work is partly consistent with reports of intermittent and avalanche-like crystallization dynamics in hard spheres. p. 6

The coupling between aging and crystallization is the key for understanding rapid crystallization in deeply supercooled liquids or glasses. The results establish that the exponential decay time τexp is the unique time scale of crystallization in the low-temperature regime where intermittent crystallization takes place. The lag time τlag does not increase at the lowest temperatures, and as a result, τcry is again dominated by τexp. The crystallization itself accelerates the relaxation dynamics, and the resulting relaxation time gives the crystallization time. The coupling between aging and crystallization is the key for understanding rapid crystallization of glasses.

The early process corresponds to the dynamics in which the system falls into the nearest IS, and this is consistent with observations that the dynamics in the early process hardly depend on temperature. The late process of crystallization consists of transitions between the ISs. p. 4 p. 4 The late process dominates the crystallization time in deeply supercooled liquids. p. 4 This suggests that the early process (falling into ISs) is not relevant for the crystallization time. p. 4 The late process of crystallization consists of transitions between the ISs.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improved crystallization prediction for deeply supercooled liquids by using an Arrhenius temperature dependence for crystallization time at very low temperatures instead of relying on Classical Nucleation Theory (CNT). This allows AI systems to predict rapid crystallization events in molecular glasses with higher accuracy than CNT-based models, especially when the energy barrier is small.

  2. Development of a coupled aging and crystallization model for AI systems to understand rapid structural changes in amorphous materials. This system can now distinguish between the early process (falling into inherent structures) and the late process (transitions between inherent structures) dominating crystallization time.

  3. Enhanced analysis of relaxation dynamics by incorporating two distinct time scales: short-time rearrangement and long-time tail relaxation, as described in Figure 4(a). This allows AI to identify when particle rearrangements are due to fast initial movements versus slower, collective motions, improving predictions for dynamic heterogeneity.

  4. Quantitative estimation of the energy barrier for crystallization using the Arrhenius fit of the crystallization time against temperature. This provides a measure of how much smaller the crystallization barrier is compared to equilibrium dynamics, leading to more accurate calculations for molecular systems like those produced by vapor deposition.

  5. Ability to predict that relaxation times in aging dynamics do not directly explain crystallization time, but rather that a coupling between the relaxation and crystallization dynamics is the key. This informs AI systems on which dynamical process—aging or crystallization—is rate-controlling under specific low-temperature conditions.

Related papers