Accelerated catalyst layer design and fabrication: Materials discovery through machine learning

交通和能源部门的脱碳对于减缓人为气候变化至关重要，因为仅交通运输一项就占全球与能源相关的二氧化碳排放量的23%。虽然目前的燃料电池和电解槽已经达到了用于商用乘用车和轻型汽车的最先进性能，但这些技术的广泛使用需要大幅减少催化剂负载量，以降低贵金属的高成本。此外，用于二氧化碳电解槽的催化剂材料将大气中的二氧化碳转化为合成气等有用形式的燃料，由于其复杂的离子-电子-反应物传输机制，在减少负载方面需要大量的工作。 催化层是包括聚合物电解质膜(PEM)燃料电池、水电解槽和CO2电解槽在内的所有电化学能量转换装置的核心部件。由于所需的高催化活性，贵族金属(PGM)，如铂或Ir，通常用作催化剂材料；然而，成本高昂，无法广泛采用，而超低负载量的催化剂层将大大降低这些清洁电化学能量转换技术的成本。我们将把机器学习应用于多孔材料的随机材料生成和多相流模拟，以确定优化传输行为的关键参数和材料结构。所要求的系统将用于研究新的催化剂层设计，以实现更低的催化剂负载量和增强的稳定性。 我们正在申请资金，以支持一个制造和测试系统，用于设计燃料电池和电解槽的催化剂层，以减少负载和提高稳定性。具体地说，这笔资金将用于基于超声波的催化剂涂层系统以及分段式电化学电池的表征工具，该工具将与Bazylak教授的原位X射线断层扫描系统相结合，建立催化剂材料的制造、测试和成像实验室。该实验室将为Bazylak教授和她的学员提供一个国际上独一无二的基于机器学习的材料发现系统，用于高通量、同时进行操作测试和新型催化剂层的表征。其影响将是为负担得起的燃料电池和电解槽发现超低负载催化层结构。这里提出的研究将导致燃料电池和电解槽的新催化剂设计和制造工艺，并将传播到更广泛的研究社区。这项研究的结果将加强加拿大在催化剂开发方面的领导地位，并吸引工业界和学术界合作伙伴推动清洁电化学能源转换。

Decarbonizing the transport and energy sectors is critical for mitigating anthropogenic climate change, as transportation alone accounts for 23% of global energy-related carbon dioxide emissions. While current fuel cells and electrolyzers are already reaching state of the art performances to be used in commercial passenger cars and light vehicles, the widespread use of these technologies demands a substantial reduction in catalyst loading to reduce the high costs of precious group metals. Moreover, the catalyst materials for CO2 electrolyzers, which convert atmospheric CO2 into useful forms of fuel such as syngas, require a substantial amount of work in terms of reducing the loading due to its complicated ion-electron-reactant transport mechanisms. A catalyst layer is the core component in all electrochemical energy conversion devices including polymer electrolyte membrane (PEM) fuel cells, water electrolyzers, and CO2 electrolyzers. Due to the high catalytic activity required, precious group metals (PGMs) such as Platinum or Iridium are typically used as catalyst materials; however, the costs are prohibitive for widespread adoption, and ultra-low loading catalyst layers would drastically reduce the costs for these clean electrochemical energy conversion technologies. We will apply machine learning to stochastic material generation and multiphase flow simulations of porous materials to identify the key parameters and material architectures for optimal transport behaviour. The requested system will be used to investigate new catalyst layer designs to achieve reduced catalyst loadings with enhanced stability. We are requesting funds to support a fabrication and testing system for designing catalyst layers with reduced loadings and high stability for fuel cells and electrolyzers. Specifically, the funds will be used for an ultrasonic based catalyst coating system as well as a characterization tool for a segmented electrochemical cell, which will be coupled with Prof. Bazylak's in situ X-ray tomography system to establish a fabrication, testing, and imaging laboratory for catalyst materials. This laboratory will provide Prof. Bazylak and her trainees with an internationally unique system for machine learning based materials discovery for high throughput, concurrent in operando testing and characterization of novel catalyst layers. The impact will be the discovery of ultra-low loading catalyst layer architectures for affordable fuel cells and electrolyzers. The research proposed here will lead to new catalyst designs and manufacturing processes for fuel cells and electrolyzers that will be disseminated to the wider research community. The outcomes of this research will strengthen Canada's leadership in catalyst development and attract industrial and academic partners for the advancement of clean electrochemical energy conversion.