Representative chemistry and strain analysis of fuel cell catalyst nanoparticles via machine learning and DFT modelling

氢气的使用越来越多地被视为实现我们的净零碳能源目标的关键组成部分。氢燃料电池用于将氢转化为电能，但受到燃料电池阴极处缓慢的氧气还原反应的限制。以铂为基础的催化剂被用来提高反应速度，但铂是一种昂贵的金属，限制了氢基能源的经济可行性。用更便宜的贱金属合金化铂可以减少所需的铂的质量，令人惊讶的是，它可以提高活性，而不是铂。活性增强的原因尚不完全清楚，可能与物种的组成聚集、混合金属的化学效应或晶格应变的效应(如果组成不均匀)有关。挑战在于，在目前的实验能力范围内，测量纳米颗粒中的应变或成分都是恰到好处的，特别是在我们需要能够检查许多颗粒以了解整体性质的方法的情况下。该项目将利用最先进的电子显微镜成像和光谱技术来确定成分，包括氧化程度和产生的菌株。扫描透射式电子显微镜(STEM)(一项在英国已有大量投资的技术)将成为主要的实验工具。STEM将用于形成原子分辨率图像，并使用电子能量损失光谱(EELS)和能量色散X射线(EDX)光谱同时测量成分。机器学习将被用来分析更多的粒子。以前的工作表明，存在高水平的剪切应变，但这对催化剂电子结构的影响尚未被研究。密度泛函理论模型将被用来理解结构和活性之间的联系，基于使用电子显微镜研究测量的结构、应变和组成。充分了解结构和活性之间的联系是开发新催化剂系统的重要一步。该项目与EPSRC能源主题非常一致，该主题将重点放在“能源主题的总体目标是赞助研究和博士培训，以确保低碳的未来，通过创建可靠的、经济上可行的能源系统，同时保护自然环境、资源和生活质量。”这项工作的方法论方面还将支持更一般的物理科学主题内的其他活动。

The use of hydrogen is increasingly seen as a key component to meeting our goals on net zero carbon energy sources. Hydrogen fuel cells are used to convert hydrogen to electricity, but are limited by the sluggish oxygen reduction reaction at the fuel cell cathode. Catalysts based on platinum are used to increase this reaction rate, but Pt is an expensive metal limiting the economic viability of hydrogen-based energy. Alloying Pt with cheaper base metals reduces the mass of Pt required, and surprisingly can enhance activity beyond that of Pt. The origins of the enhanced activity are not fully understood and may be associated with compositional clustering of species, the chemical effects of mixing metals or the effect of lattice strain if the composition is inhomogeneous. The challenge is that measuring either strain or composition within a nanoparticle is right at the limits of current experimental capabilities, especially as we need methods that can examine many particles to understand the ensemble properties. This project will make use of state-of-the-art electron microscope technologies for imaging and spectroscopy to determine composition including degree of oxidation and the resulting strain. Scanning Transmission Electron Microscopy (STEM) (a technique in which there has been substantial investment in the UK) will be the primary experimental tool. STEM will be used to form atomic-resolution images and to simultaneous measure composition using electron energy-loss spectroscopy (EELS) and energy-dispersive X-ray (EDX) spectroscopy. Machine learning will be used to allow a larger numbers of particles to be analysed. Previous work has shown that high levels of shear strain are present, but the effect of this on the electronic structure of the catalysts has not been studied. Density functional theory modelling will be used to understand the link between structure and activity, based on the structure, strain and composition that is measured using the electron microscope studies. Developing a full understanding of the link between structure and activity is an important step in the development of new catalyst systems.This project aligns very closely with the EPSRC Energy theme, which states as a priority "The overarching goal of the Energy theme is to sponsor research and PhD training to secure a low-carbon future, through the creation of reliable, economically viable energy systems while protecting the natural environment, resources and quality of life." The methodological aspects of the work will also support other activities within the more general Physical Sciences theme.