Henritzi Patrick, Bormuth André, Klameth Felix, Vogel Michael
Institut für Festkörperphysik, Technische Universität Darmstadt, 64289 Darmstadt, Germany.
J Chem Phys. 2015 Oct 28;143(16):164502. doi: 10.1063/1.4933208.
We perform molecular dynamics simulations for viscous liquids to study the relations between dynamical heterogeneity, structural (α) relaxation, and self-diffusion. For atomistic models of supercooled water, polymer melts, and an ionic liquid, we characterize the space-time characteristics of dynamical heterogeneity by the degree of deviations from Gaussian displacement statistics (α2), the size of clusters comprising highly mobile particles (S(w)), and the length of strings consisting of cooperatively moving particles (L(w)). Comparison of our findings with previous simulation results for a large variety of viscous liquids, ranging from monoatomic liquids to silica melt, reveals a nearly universal decoupling between the time scales of maximum non-Gaussian parameter (τ(α2)) and the time constant of the α relaxation (τ(α)) upon cooling, explicitly, τ(α2) ∝τ(α)(3/4). Such uniform relation was not observed between the peak times of S(w) or L(w) and τ(α). On the other hand, the temperature-dependent time scale of maximum string length (τ(L)) follows the inverse of the self-diffusion coefficient (D) for various systems at sufficiently low temperatures, i.e., τ(L) ∝ D(-1). These observations are discussed in view of a breakdown of the Stokes-Einstein relation for the studied systems. It is found that the degree of deviation from this relation is correlated with the stretching of the α relaxation.
我们对粘性液体进行分子动力学模拟，以研究动力学非均匀性、结构（α）弛豫和自扩散之间的关系。对于过冷水、聚合物熔体和离子液体的原子模型，我们通过偏离高斯位移统计量的程度（α2）、由高度移动粒子组成的簇的大小（S(w)）以及由协同移动粒子组成的链的长度（L(w)）来表征动力学非均匀性的时空特征。将我们的研究结果与先前对从单原子液体到二氧化硅熔体等多种粘性液体的模拟结果进行比较，发现在冷却时最大非高斯参数的时间尺度（τ(α2)）和α弛豫的时间常数（τ(α)）之间存在近乎普遍的解耦，具体而言，τ(α2) ∝τ(α)(3/4)。在S(w)或L(w)的峰值时间与τ(α)之间未观察到这种统一关系。另一方面，在足够低的温度下，各种系统中最大链长度的温度依赖时间尺度（τ(L)）遵循自扩散系数（D）的倒数，即τ(L) ∝ D(-1)。鉴于所研究系统的斯托克斯 - 爱因斯坦关系的失效，对这些观察结果进行了讨论。发现偏离该关系的程度与α弛豫的拉伸相关。
