Understanding Thermal Diffusion in Liquid Mixtures and Macromolecular Solutions via Molecular Dynamics Simulations

该项目的目的是通过分子动力学模拟进一步深入了解流体混合物中热扩散(Soret效应)的分子起源。基于在Lennard-Jones液体混合物上获得的现有经验，将进行两个方面的研究。首先，将模拟扩展到真实的液体混合物的原子模型。这样做的目的是将结果与最近已有或正在进行的准确测量进行比较，从而从定量上验证分子动力学方法。这将使我们能够检查从Lennard-Jones模型混合物的早期工作中得出的概念，例如质量和内聚能量密度的差异导致的Soret效应的可加性，是否仍然适用于现实系统。详细的原子模拟的另一个目的是理解Köhler等人最近的一项实验发现，即Soret效应也有来自组件转动惯量差异的贡献。其次，用链状低聚物溶液(单原子溶剂中的珠弹性型大分子)的简单模型进行模拟。这组模拟的目的是研究大分子溶液中Soret效应的特性，特别是链长度的热扩散系数的独立性和Soret系数的符号反转随溶剂质量的变化。这些效应是非常普遍的，并且已经在许多大分子溶液中被发现。因此，最初使用简化的齐聚物模型是明智的，它允许模拟大型系统。大系统是必要的，因为这些溶液是非常稀薄的，所以必须包括许多显式溶剂。这两种模拟方法都将与PD西蒙·维甘德博士(FZ Jülich)的实验小组密切合作执行。

The aim of the proposed project is to gain further insight into the molecular origin of thermal diffusion (Soret effect) in fluid mixtures by means of molecular dynamics simulations. Building on existing experiences gained on mixtures of Lennard- Jones liquids, two lines of research are to be pursued. Firstly, the simulations are to be extended to realistic atomistic models of liquid mixtures. The purpose of this is to compare results with accurate measurements, which recently have become available or are being conducted, and to thereby validate the molecular dynamics approach quantitatively. This will allow us to check if concepts derived from the earlier work on Lennard-Jones model mixtures, such as the additivity of Soret effects from the differences in mass and in cohesive energy density, still hold in realistic systems. Another purpose of the detailed atomistic simulations is to understand a recent experimental finding by Köhler et al., namely that the Soret effect also has a contribution arising from the difference of moments of inertia of the components. Secondly, simulations are to be conducted with simple models of chain-like oligomer solutions (bead-spring type macromolecule in a monoatomic solvent). This set of simulations is aimed at studying the peculiarities of the Soret effect in macromolecular solutions, in particular the independence of the thermal diffusion coefficient of the chain length and the sign reversal of the Soret coefficient with changing solvent quality. These effects are quite generic and have been found for many macromolecular solutions. It is therefore sensible to initially use a simplified oligomer model, which allows the simulation of large systems. Large sys- tems are necessary, as these solutions are very dilute, so much explicit solvent has to be included. Both simulation approaches will be executed in close collaboration with the experimental group of PD Dr. Simone Wiegand (FZ Jülich).