Reducing energy requirements of gas separations by computational material design of metal-organic frameworks

气体分离是世界上能源最密集的化学过程之一。通过用基于吸附的工艺[1]取代传统的低温蒸馏，这些分离的环境足迹将大大减少。这种情况尚未发生的主要原因是缺乏具有理想特性的吸附剂材料，使每种气体分离工艺可行。该项目旨在利用实验和计算模型之间的协同努力，探索新的纳米多孔金属有机框架（MOF）材料，用于具有挑战性的气体分离。这些是结晶有机金属框架，允许孔结构和表面化学的可调设计，以适应特定的应用。近年来，在使用含有协调不饱和位点（CUS）[2]的mof方面取得了令人兴奋的进展。这些位点是在合成过程中去除附着在金属位点上的溶剂分子后在MOF材料上产生的。因此，材料激活产生了“空”位点，这些位点可以通过配位型键强烈而特异性地结合给电子分子，极大地提高了材料的选择性。存在的各种不同的MOF结构（~70000）或可能被合成的MOF结构（>500000）使得使用昂贵且耗时的实验来筛选如此大量的材料是不可能的。像分子模拟这样的计算工具可以快速廉价地筛选特定应用的mof。然而，阻碍高通量筛选工作转化为实际计算材料设计的主要障碍是缺乏适当的分子模型，无法准确描述气体分子与CUS之间的相互作用。事实上，已有研究表明，传统模型无法处理这种相互作用，需要将量子化学（QM）与经典蒙特卡罗（MC）模拟相结合的多尺度方法来解决这种复杂的问题。本项目将建立在Jorge集团[5]中开发的混合QM/MC模型的基础上，并将其应用于预测在工业相关分离条件[1]下，CUS在大量mof中的吸附。案例研究将涉及存在水蒸气的分离（例如碳捕获），这是设计中最具挑战性的过程之一。这些模拟的结果将用于训练机器学习算法，与科罗拉多州[3]的Gómez-Gualdrón小组合作，探索现有和假设的mof的全部集合。一旦确定了一小部分最有前途的mof，它们将在弗莱彻实验室进行实验合成、表征和吸附分离测试，从而结束材料设计周期的循环。[1] shall, and Lively；自然科学学报，2016,32 (5):435-437Bachman等人；j。化学。系统工程学报，2017,39(3):15363-15370.链接本文Anderson等人；化学。材料学报，2018,30,6325-6337.链接本文Fischer等人；摩尔。同时。[j] .中国农业科学，2014,40,537-556豪尔赫·m·;印第安纳州,Eng。化学。科学通报，2014,53(3):15475-15487。

Gas separations are some of the most energy-intensive chemical processes worldwide. The environmental footprint of these separations would be immensely reduced by replacing the conventional cryogenic distillation with an adsorption-based process [1]. The main reason why this has not yet happened is the lack of adsorbent materials with the ideal characteristics to make each gas separation process viable. This project aims to deploy a synergistic effort between experiments and computational modelling to explore new nanoporous Metal-Organic Framework (MOF) materials for challenging gas separations. These are crystalline organometalic frameworks that allow for tunable design of both pore structure and surface chemistry to suit particular applications. Exciting developments have been recently reported on using MOFs that contain Coordinatively Unsaturated Sites (CUS) [2]. These sites arise on MOF materials after the removal of solvent molecules attached to the metal sites during synthesis. Material activation thus yields "vacant" sites that can bind strongly and specifically to electron-donating molecules through coordination-type bonds, dramatically increasing the material's selectivity. The wide variety of different MOF structures that exist (~70000) or can potentially be synthesised (>500000) makes it impossible to screen over such a large number of materials using expensive and time-consuming experiments. Computational tools like molecular simulation allow fast and cheap screening of MOFs for a particular application [3]. However, the main hurdle preventing high-throughput screening efforts to be turned into de facto computational material design is the lack of appropriate molecular models that can accurately describe the interactions between gas molecules and the CUS. Indeed, it has been shown that conventional models are unable to handle this kind of interaction, and that multi-scale approaches that combine quantum chemistry (QM) with classical Monte Carlo (MC) simulations are needed to address such complex problems [4]. The present project will build upon a hybrid QM/MC model developed in the Jorge group [5] and apply it to predict adsorption in a large number of MOFs with CUS under industrially relevant separation conditions [1]. Case studies will involve separations in the presence of water vapour (e.g. carbon capture), which are among the most challenging processes to design. The results of these simulations will be used to train a machine-learning algorithm, in collaboration with the group of Gómez-Gualdrón at Colorado [3], to explore the full set of existing and hypothetical MOFs. Once a small subset of MOFs with the most promising performance is identified, they will be experimentally synthesised, characterised and tested for adsorption separations in the Fletcher lab, thus closing the loop on the material design cycle. [1] Sholl, and Lively; Nature 2016, 532, 435-437.[2] Bachman et al.; J. Am. Chem. Soc. 2017, 139, 15363-15370.[3] Anderson et al.; Chem. Mater. 2018, 30, 6325-6337.[4] Fischer et al.; Mol. Simul., 2014, 40, 537-556.[5] Jorge, M.; Ind. Eng. Chem. Res., 2014, 53, 15475-15487.