在传统晶体中，扩散通常是动力学均匀的，使用单值或者两个特征激活能去表征。但在原子尺度上，这个认识并不适用于最近热门的高熵合金。高熵合金是一种由至少五个以上元素等比例或接近等比例组成的复杂浓缩合金，其多主元随机占据晶格位置的奇特结构特性可以引起显著的局部化学不均匀性。对高熵合金扩散的研究有助于我们去了解其相稳定性以及变形行为，从而扩展对高熵合金的应用。本文中我们通过结合分子动力学，分子静力学和激活-弛豫方法发现并解耦了典型高熵Cantor合金(CoCrFeMnNi)中可能存在的适应晶格扩散的不均匀性。

首先，为了探索高熵合金中激活能的上限和下限，我们分别建立了单空位和空位饱和的模型。这组模型有望为高熵合金中的晶格扩散定义一个与实验数据相近的激活能范围。然而我们发现，两个模型得到的激活能与实验值相比偏小，需要修正。

然后，我们使用分子静力学去求高熵合金中各元素原子的空位形成能；同时，又使用激活-弛豫方法去求空位饱和模型中空位近邻原子的迁移能。由于高熵合金多主元随机占据晶格位置的特性，每个原子的局部化学环境都不一样，我们最后得到各元素原子空位形成能和迁移能的分布。在得到了空位形成能后，我们对激活能进行了修正，得到了更为合理的值。

最后，我们将模拟得到的空位形成能和迁移能的分布，文献中扩散激活能的实验值与模拟得到的两个模型的扩散激活能进行比较。我们认为Arrhenius方程中得到的扩散激活能等于空位形成能加迁移能这一传统假设在高熵合金中难以适用。迁移能的广泛分布范围可能是高熵合金乃至多主元合金的扩散不均匀性的重要来源。指前因子也可能是扩散不均匀性的重要来源之一。与传统固溶体中均匀扩散的经典理论相比，基于原子尺度的对扩散不均匀性的理解，突出了扩散路径的复杂性以及一般的复杂浓缩合金中化学、拓扑和动力学不均匀性之间的密切关系。

Diffusion in the traditional crystalline solids is dynamically homogeneous characterized by a single-value activation energy. However, such a scenario breaks down at atomic scale in the recently advanced high-entropy alloy. High-entropy alloy is a kind of complex concentrated alloy composed of at least five elements in equal proportions or close to equal proportions. Its unique structural features with multi-principal elements randomly occupying on lattice sites that may induce strikingly local chemical heterogeneity. The research on the high-entropy alloys can help us to understand their phase stability as well as deformation behavior, and extend the application of them. In the paper we uncover and decouple the dynamic heterogeneity accommodating the lattice diffusion in an archetypical high-entropy Cantor alloy (CoCrFeMnNi) via combined molecular dynamics, molecular statics and activation-relaxation technique.

Initially, we set up a single-vacancy and a vacancy-saturated model to explore the upper bound and lower bound of activation energies. The models are expected to define a possible range of activation energies for the lattice diffusion in high-entropy alloys, which are comparable to experimental data. However, we found the activation energies derived from two models are smaller than the experimental value hence modifications are necessary.

Afterwards, we used molecular statics to get the vacancy formation energy of each type atom in high-entropy alloy while use the activation-relaxation technique to figure out the migration energy of the nearest-neighbor atoms of the vacancies in vacancy-saturated model. Due to the feature of high-entropy alloy that multi-principal elements randomly occupying on lattice sites, each atom in the model has a different local chemical environment, we got the distribution of vacancy formation energy for each element as well as migration energy finally. We modified the activation energies and got a more reasonable value after obtaining the vacancy formation energy.

Finally, we compared the distributions of migration energy and vacancy formation energy obtained from the simulation, the experimental value of diffusion activation energy from the literature and activation energy of two models from the simulations. We argue that the conventional hypothesis of diffusion activation energy estimated from Arrhenius equation as a sum of vacancy formation energy and migration energy becomes intractable in high-entropy alloy. The wide distribution of the migration energy may be an important source of the diffusion heterogeneity of high-entropy alloys and even multi-principal element alloys. The pre-exponential factors may be another source of the diffusion heterogeneity. The atomic-scale insights to diffusion heterogeneity, in contrast to the classical theory of homogeneous diffusion in conventional solid solutions, highlight the complexity of diffusion pathway and the intimate correlation between chemical, topological and dynamic heterogeneity in the generic complex concentrated alloys.