Giovambattista Nicolas, Buldyrev Sergey V, Stanley H Eugene, Starr Francis W
Center for Polymer Studies and Department of Physics, Boston University, Massachusetts 02215, USA.
Phys Rev E Stat Nonlin Soft Matter Phys. 2005 Jul;72(1 Pt 1):011202. doi: 10.1103/PhysRevE.72.011202. Epub 2005 Jul 13.
We study the spatially heterogeneous dynamics in water via molecular dynamics simulations using the extended simple point charge potential. We identify clusters formed by mobile molecules and study their properties. We find that these clusters grow in size and become more compact as temperature decreases. We analyze the probability density function of cluster size, and we study the cluster correlation length. We find that clusters appear to be characterized by a fractal dimension consistent with that of lattice animals. We relate the cluster size and correlation length to the configurational entropy, S(conf). We find that these quantities depend weakly on 1/ S(conf). In particular, the linearity found between the cluster mass n() and 1/ S(conf) suggests that n() may be interpreted as the mass of the cooperatively rearranging regions that form the basis of the Adam-Gibbs approach to the dynamics of supercooled liquids. We study the motion of molecules within a cluster, and find that each molecule preferentially follows a neighboring molecule in the same cluster. Based on this finding we hypothesize that stringlike cooperative motion may be a general mechanism for molecular rearrangement of complex, as well as simple liquids. By mapping each equilibrium configuration onto its corresponding local potential energy minimum or inherent structure (IS), we are able to compare the mobile molecule clusters in the equilibrium system with the molecules forming the clusters identified in the transitions between IS. We find that (i) mobile molecule clusters obtained by comparing different system configurations and (ii) clusters obtained by comparing the corresponding IS are completely different for short time scales, but are the same on the longer time scales of diffusive motion.
我们使用扩展的简单点电荷势，通过分子动力学模拟研究水中的空间非均匀动力学。我们识别由可移动分子形成的簇，并研究它们的性质。我们发现，随着温度降低，这些簇的尺寸会增大且变得更加紧凑。我们分析了簇尺寸的概率密度函数，并研究了簇相关长度。我们发现，簇似乎具有与晶格动物一致的分形维数特征。我们将簇尺寸和相关长度与构型熵S(conf)联系起来。我们发现这些量对1/S(conf)的依赖性较弱。特别是，在簇质量n()与1/S(conf)之间发现的线性关系表明，n()可以解释为形成过冷液体动力学的Adam-Gibbs方法基础的协同重排区域的质量。我们研究了簇内分子的运动，发现每个分子优先跟随同一簇中的相邻分子。基于这一发现，我们假设线状协同运动可能是复杂液体以及简单液体分子重排的一种普遍机制。通过将每个平衡构型映射到其相应的局部势能最小值或固有结构(IS)，我们能够将平衡系统中的可移动分子簇与在IS之间的转变中识别出的形成簇的分子进行比较。我们发现：(i) 通过比较不同系统构型获得的可移动分子簇和(ii) 通过比较相应的IS获得的簇在短时间尺度上完全不同，但在扩散运动的较长时间尺度上是相同的。
