Diffusion in Multi Component Mixtures

温度梯度在人类和工业活动中普遍存在，并可引起广泛的热传输现象、多组分混合物中的质量分离(热扩散)或溶液中的溶质运动(热电泳法)。这些现象对于确定油藏的组成、短DNA片段的复制和积累，或设计微米和纳米级的、用于处理从纳米到微米的小体积液体或溶质的热电泳器具有实际意义。在地下油气藏中，温度在垂直方向上变化，有时在水平方向上变化(即可能存在不可忽略的水平梯度)。随着时间的推移，由于热扩散，一些烃类组分向冷侧分离，另一些组分向储集层热侧分离。热分离还会引发由浓度梯度和自然对流驱动的分子扩散。因此，确定油藏中流体成分的变化是至关重要的。目前在油藏模拟器中使用的模型是基于菲克定律的，该定律不能解释这些组成相互作用和物种的非理想行为。悬浮液或流体混合物对温度梯度的响应以一种微妙且基本上未知的方式取决于分子相互作用。了解这些分子相互作用是如何决定热传输的，对于推动具有技术和工业意义的设备和工艺的发展至关重要。为了能够预测这种机制，需要一个准确的分子多组分扩散模型。我们建议在实验上测量这些扩散系数，并首次引入结合不可逆热力学理论的分子动力学模型。将实验结果与数值结果进行比较，将有助于我们建立一个准确的分子扩散模型。这种实验测量已经在国际空间站上进行，但到目前为止在地面实验室取得的成功有限。这是由于对流的存在，产生了混合，因此破坏了分离过程。针对空间实验成本高、获取空间受限的问题，提出了一种创新性的实验设计，在实验室内测量不同平均混合温度下的扩散系数。这种实验方法需要精确的样品设计来克服重力效应(即导致混合的对流)。实验室测量扩散系数的成功将导致准确的分子多组分扩散模型，这在油藏模拟中是非常需要的。

Thermal gradients are ubiquitous in human and industrial activities, and can induce a wide range of thermal transport phenomena, mass separation in multicomponent mixtures (thermodiffusion), or solute motion in solutions (thermophoresis). These phenomena are of practical interest in defining the composition of oil reservoirs, the replication and accumulation of short DNA fragments, or the design of microscale and nanoscale thermophoretic devices for the manipulation of small volumes of liquids or solutes in scales ranging from nano to micrometer. In subsurface hydrocarbon reservoirs, the temperature varies vertically and sometimes horizontally (i.e. there may be non-negligible horizontal gradients). Over time, some hydrocarbon components segregate towards the cold side and others towards the hot side of the reservoir due to thermodiffusion. Thermal segregation will also initiate molecular diffusion driven by concentration gradients and natural convection. As a result, it is crucial to determine the fluid compositional variation in the oil reservoir. Current models that are used in oil reservoir simulators are based on Fick's law, which cannot account for these compositional interactions and the non-ideal behaviour of the species. The response of a suspension or a fluid mixture to a temperature gradient depends in a subtle and essentially unknown way on the molecular interactions. Understanding how these molecular interactions determine thermal transport is essential to advance the development of devices and processes of technological and industrial interest. An accurate molecular multicomponent diffusion model is needed to be able to predict this mechanism. We propose to measure experimentally these diffusion coefficients and introduce for the first time a molecular dynamics model combined with irreversible thermodynamic theory. Comparison between experimental and numerical results will help us develop an accurate molecular diffusion model. Such experimental measurements have been performed on board the International Space Station but sofar had a limited success in ground laboratory. This is due to the presence of convection, which creates mixing and therefore destroys the separation process. With a high cost and limited access to space experiment, an innovative experimental design to measure these diffusion coefficients in our laboratory for different average mixture temperature is proposed. This experimental approach requires a precise sample design to overcome the gravity effect (i.e. convection which leads to mixing). Succeeding in measuring the diffusion coefficients in laboratory will lead to an accurate molecular multicomponent diffusion model, which is in great needs in oil reservoir simulation.