Modelling correlated electron-ion diffusion in nano-scale TiO2: beyond periodic model and density functional theory

当前研究的重大挑战之一是减少社会对化石燃料使用的依赖，从而降低二氧化碳排放水平。这一挑战的一个主要战略组成部分是用太阳能电池、氢燃料电池和电池等替代品取代液体燃料（汽油、柴油和煤油），这些燃料是汽车和航空航天应用的主要燃料来源。特别是，几款使用锂电池的汽车已经量产。然而，由于电池充电时间长、能量容量小、老化等原因，电动汽车占据的市场份额仍然很小。例如，在内伦敦居民中很受欢迎的G-Wiz，充电时间为8小时，最大续航里程约为50英里，而一款新的（2008年生产的）全电动跑车特斯拉跑车（Tesla Roadster）需要长达17小时的充电时间才能行驶240英里。锂电池特性的非增量改进需要大幅提高电动汽车的比例，从而大大降低噪音、污染和二氧化碳排放水平，这反过来又将对经济、环境和健康产生有益的影响。纳米结构电极材料，特别是纳米结构TiO2基化合物的使用，成为提高锂电池性能特征的一种有前途的策略。本研究项目旨在通过原子尺度机制的量子力学建模和Li离子迁移与电子跳变耦合的动力学，促进新电极材料的发展。我们将通过比较它们单独和成对迁移的动力学来研究Li+ -电子相互作用的影响。为了研究纳米结构的影响，我们将比较大块和晶界附近Li+和e-物种的迁移动力学。了解Li+/e-耦合和界面结构的影响将为我们提供控制和优化能量容量和充电速度的各种协议。其中一些方案可能包括选择TiO2纳米结构的特征尺寸和控制电子注入。氧化物中相关离子-电子迁移是一个普遍的问题，但在量子力学水平上还没有得到适当的解决。这是因为广泛使用的基于密度泛函理论（DFT）的从头算方法严重低估了绝缘体和半导体的带隙，从而导致对捕获电子性质的定性描述不正确。此外，这些方法采用周期边界条件模型，不适用于复杂非周期结构的建模。在本提案中，我们将突破这些限制，并应用专门为不规则表面和界面中的缺陷建模而设计的嵌入式聚类方法。作为该提案的一部分，我们将开发新形式的一致的远程静电和短程嵌入电位。这将使我们能够应用精确的量子化学方法进行电子结构计算，而不会受到传统DFT技术的缺点的影响。这些嵌入势可以用于与TiO2体积、表面、界面和纳米结构相关的许多电子现象的其他计算研究。为了更广泛地利用它们，我们将与量子化学计算计算机软件包的开发人员合作。

A principal focus of one of the current research grand challenges is on reducing society's dependence on the use of fossil fuels and thus decreasing the CO2 emission levels. A major strategic component of this challenge is replacing liquid fuels (petrol, diesel, and kerosene), as main fuel sources for automotive and aerospace applications, with alternatives such as solar cells, hydrogen fuel cells, and electric batteries. In particular, several models of automobiles operating on Li batteries are already mass-produced. However, the market niche taken up by the electric cars remains small due to long battery charging times, small energy capacity, and aging. For example, the G-Wiz, which is popular among inner London residents, has a charging time of 8 hours and maximum range of ~50 miles, while a new (2008 production) all-electric sports car the Tesla Roadster needs up to 17 hours of charging time for a 240 mile journey. Non-incremental improvements of Li battery characteristics are needed to dramatically increase the fraction of electric automobiles and, thus, considerably reduce noise, pollution, and CO2 emission levels, which, in turn, will have beneficial economical, environmental, and health implications. The use of nano-structured electrode materials, in particular, nano-structured TiO2 based compounds, emerged as a promising strategy for increasing performance characteristics of Li batteries. This research project aims to facilitate the development of new electrode materials through quantum-mechanical modelling of the atomic-scale mechanisms and kinetics of Li ion migration coupled with electron hopping. We will investigate the effects of the Li+ - electron interaction by comparing the kinetics of their migration individually and as a pair. To investigate the effects of nano-structuring, we will compare the migration kinetics of the Li+ and e- species in the bulk and in the vicinity of the grain boundary. Understanding the effect of the Li+/e- coupling and that of the interface structure would provide us with a variety of protocols for controlling and optimising the energy capacity and charging speed. Some of these protocols may include selecting characteristic size of TiO2 nano-structures and controlling electron injection. Despite being a generic problem, correlated ion-electron migration in oxides has not been properly addressed on the quantum mechanical level. This is because widely used ab initio methods based on the density functional theory (DFT) severely underestimate the band gap of insulators and semiconductors, which results in a qualitatively incorrect description of the properties of trapped electrons. In addition, these methods employ a periodic boundary conditions model, which is not applicable to modelling complex non-periodic structures. In this proposal, we will break through these limitations and apply an embedded cluster method specifically designed for modelling defects in irregular surfaces and interfaces. As a part of this proposal, we will develop new forms of consistent long-range electrostatic and short-range embedding potentials. This will allow us to apply accurate quantum-chemical methods for electronic structure calculations, which do not suffer from drawbacks of conventional DFT techniques. These embedding potentials can be used in other computational studies of numerous electronic phenomena associated with TiO2 bulk, surfaces, interfaces, and nano-structures. To facilitate their wider availability, we will collaborate with developers of computer packages for quantum-chemical calculations.