Zeolite is a class of microporous crystalline materials widely used in heterogeneous catalysis. Over the past decades, theoretical simulations, particularly those based on first principles calculations, have advanced significantly the understandings on zeolite, from structure to adsorption kinetics and to catalytic reactivity. The machine learning (ML) methods developed in recent years further boost the ability of theory for unraveling the interplay between the synthetic conditions and the zeolite structure and functionality. This short review overviews the theoretical insights into the role of zeolite framework in zeolite stability and catalysis revealed from atomic simulation in recent years. We will mainly focus on two key aspects: (i) the theory on zeolite stability, including the templating effect of structure directing agents and the zeolite bonding pattern analysis; (ii) the confinement effect of zeolite pores that affects the catalytic conversion of molecules in zeolite. The future directions of theoretical simulation are also discussed.

沸石是一类广泛用于多相催化的微孔结晶材料。在过去的几十年里，理论模拟，特别是那些基于第一性原理计算的理论模拟，极大地促进了对沸石的理解，从结构到吸附动力学，再到催化反应。近年来开发的机器学习 (ML) 方法进一步提高了解开合成条件与沸石结构和功能之间相互作用的理论能力。这篇简短的综述概述了近年来原子模拟揭示的沸石骨架在沸石稳定性和催化中的作用的理论见解。我们将主要关注两个关键方面：(i) 沸石稳定性理论，包括结构导向剂的模板效应和沸石键合模式分析；(ii) 沸石孔隙的限制作用，影响分子在沸石中的催化转化。还讨论了理论模拟的未来方向。