Computational modelling and design of nanoporous silica materials

纳米多孔材料，如沸石或活性炭，有着广泛的应用，从石油化工行业的气体分离，到空气或水的净化，再到控制药物输送等医疗用途。事实上，纳米多孔材料的市场规模估计为15亿英镑，并将在2017年上升至18亿英镑。尽管它们具有巨大的潜力，但由于我们缺乏对其合成过程的基本理解和控制，进一步的发展受到限制，大多数发现都是通过穷尽搜索或启发式方法的应用产生的。显然，有必要改变这种模式，以实现这些材料的有针对性的设计，而计算模型非常适合这一目的。纳米多孔材料的计算设计将使我们能够通过减少材料发现过程中必要的实验数量来节省时间和金钱，更重要的是，将使我们能够为特定的目标应用调整新材料的特性（例如，最大化材料对工业废水中给定污染物的亲和力）。本研究的主要目的是开发一种多尺度建模策略，可以描述纳米多孔材料从前驱体溶液到最终多孔固体的整个合成过程。我们将使用周期性介孔硅（PMS）作为原型系统，因为它们已经在实验中得到了广泛的研究，它们是使用模板合成工艺制成的（固体的结构由二氧化硅/表面活性剂液晶决定），它们的最终结构特别容易通过改变合成条件来调整。我们将在PI小组先前开创性研究的基础上，建立一个从量子力学水平到经典原子水平再到中尺度水平的模型层次结构，以降低复杂性（从而提高计算效率）。低级模型将根据高级模型和实验数据进行验证，保持必要的准确性，同时扩大长度和时间尺度的可访问范围。我们的想法是，使用最终模型，我们将能够仅基于初始合成条件的知识，生成PMS材料的完整虚拟模型——本质上是模仿计算机上的实际实验。至关重要的是，这一目标依赖于开发一种能够在这些复杂环境中处理二氧化硅化学反应的模型，这本身将构成计算材料科学领域的一项重大创新。

Nanoporous materials, like zeolites or activated carbon, are used in a wide range of applications, from gas separations in the petrochemical industry, to air or water purification, to medical uses like controlled drug delivery. Indeed, the market for nanoporous materials is estimated at ~£1.5 billion, and set to rise to ~£1.8 billion in 2017. Despite their tremendous potential, further developments are limited by our lack of fundamental understanding and control over their synthesis processes, with most discoveries arising from the application of exhaustive searches or heuristic approaches. It is clearly necessary to change this paradigm to enable targeted design of these materials, and computational models are ideally suited for this purpose. Computational design of nanoporous materials would allow us to save time and money by reducing the number of necessary experiments in the path to material discovery, and, more importantly, would enable us to tune the properties of a new material for a specific target application (for example, maximising the affinity of the material towards a given pollutant present in an industrial effluent). The main aim of this research is to develop a multiscale modelling strategy that can describe the entire synthesis process of a nanoporous material, from the precursor solution to the final porous solid. We will use periodic mesoporous silicas (PMS) as a prototype system, because they have been widely studied experimentally, they are made using a templated synthesis process (the structure of the solid is determined by silica/surfactant liquid crystals), and their final structure is particularly amenable to tuning by changing the synthesis conditions.We will build upon previous groundbreaking research in the PI's group to establish a hierarchy of models of decreasing degree of complexity (and thus increasing computational efficiency), ranging from the quantum-mechanical level, to the classical atomistic level, to the mesoscale level. Lower-level models will be validated against higher-level models and experimental data, maintaining the necessary accuracy while expanding the accessible range of length and time scales. The idea is that using the final model we will be able to generate a complete virtual model of a PMS material based only on knowledge of the initial synthesis conditions - essentially mimicking an actual experiment on the computer. Crucially, this goal relies on developing a model that can cope with chemical reactions of silica in these complex environments, which in itself will constitute a major innovation in the field of computational material science.