高温铁基费托合成是我国当前最主要的煤炭间接液化技术。从分子水平深入理解原位条件下催化剂的结构、活性位点与反应机理是开发下一代更高性能催化剂的基础。限于现有实验技术与常规理论模拟方法，无法获得反应条件下催化剂表面活性层的动态结构，因而难于实现基于理性设计的高效催化剂开发。本项目拟采用基于原子热力学的智能杂化遗传算法，预测原位条件下的催化剂表面活性层的组成和结构，揭示反应条件下催化剂表面活性层结构的调控机制。在此基础上，采用第一性原理增强采样分子动力学方法研究原位条件下的重要费托合成基元反应，获取原位条件下的反应机理。该项目的实施不仅对费托合成新一代催化剂的开发具有重要意义，而且对其它多相催化过程有普适性，在理论催化研究中具有一定的引领作用。

High temperature iron-based Fischer-Tropsch Synthesis (FTS) is the major coal-to-liquid technology in China. Further development of next-generation FTS catalysts requires atomic-level understanding of the active sites and the reaction mechanism in operando conditions. However, limited by the experimental techniques and the defects of the current theoretical modelling methods, in-depth understanding of the dynamic structure of the active surface layers of the catalyst at atomic level in the working condition is lacking, which greatly hindered the rational design of FTS catalysts. In this project, we propose to use an atomistic thermodynamics based hybrid genetic algorithm to predict the composition and structure of the active surface layers of the catalyst in-situ, and to understand how it is modulated by the reaction conditions. Furthermore, we propose to use first-principle enhanced sampling methods to study the reaction mechanism of key FTS reactions on the predicted active surface layers in operando conditions. The implement of this project is not only of fundamental importance to the development of next generation FTS catalyst, but the strategy used in this project is also applicable in other heterogeneous catalysis problems.