Techno-Economic Optimization of Different Solid Oxide Electrolysis Cell Architectures

Solid oxide electrolysis cells are manufactured in three different architectures: Electrolyte supported (ES), Anode supported (AS) and Metal supported (MS). The three architectures differ in which cell layer functions as the mechanical support and which material is used. Over the years, quite some work has been done on optimizing the microstructure of the different architectures [1], [2], [3], but these studies focus on single parameter optimization for only one objective. This objective is always connected to performance, e.g. current density, while recently the environmental impact becomes increasingly more important. Conversely, most life cycle analyses [4], [5], [6] deal with a single membrane electrode assembly (MEA) and are limited by the number of structures and materials they can use for their models. To that end, we developed a parallelized multiphysics MEA model to simulate a very broad range of microstructures and materials for the three SOEC architectures. The aim of this model is to identify an optimal microstructure in terms of price, CO 2 emissions, and critical raw material usage for a 1MW system. The developed model is used to do a parametric sweep over the microstructural parameters: thickness, porosity, particle size and phase fraction. A LSCF/CGO/8YSZ/Ni-8YSZ is considered as a base case for the MEA. In the ES MEA, the 8YSZ electrolyte is additionally replaced by a 3YSZ and 1Ce10SSZ as alternative electrolytes. For the AS and MS, CGO is also used as a substitute of 8YSZ inside the functional fuel electrode. The microstructural sweep is performed for each of the three different architectures with different materials at thermoneutral voltage and different temperatures. Each MEA is evaluated based on the costs, CO 2 emissions from its manufacture and the critical raw material usage relative to the architecture itself. The critical raw material usage is evaluated based on the GeoPolRisk MidPoint, which is a measure of critical raw material supply disruption. The best MEA scores the lowest average across all categories. The parametric sweeping shows that MEAs, where both electrodes are smaller than 5 µm, have a very low current density regardless of their architecture. To get a good current density at 700 ̊C, the air electrode should be around 10-15 µm thick and have a porosity of ~10%. The active fuel electrode should be around 4 µm thick and have a porosity of ~15%. At 600 ̊ C, the ideal thickness for the air electrode increases to 40-50 µm and the fuel electrode to 10 µm. At the single cell level, the main contributors to the costs, CO 2 emissions, and material usage per m 2 for ES and AS architectures are YSZ and Ni as they are the majority of the supporting layers, as shown in Figure 1 (left) and Figure 2 (left). The MS MEAs are very effective in reducing YSZ and Ni usage by replacing the support layer with a cheap and available material such as stainless steel. Consequently, a reduction of 3-5 times is achieved in costs and material usage per m 2 . For the full 1MW system, around 90% of the costs and CO 2 emissions come from the Crofer interconnects. Interestingly, the Ni-CGO electrodes, while being more expensive per m 2 , have a lower cost at the 1MW level than the Ni-YSZ due to their better current density. The AS and MS MEAs are a lot better than the ES MEAs, as expected. However, when degradation is factored in, the case for the ES architecture improves significantly and the three architectures are a lot closer together in price, see Figure 1 (right). While the MS MEAs show great promise, their relatively faster degradation does make them less competitive in terms of economics. The MS architecture is still the best option for reducing the critical raw material usage (see Figure 2). In the end, an ES MEA using a good ionic conductive electrolyte like 1Ce10SSZ seems to be the most optimal choice when considering a 1MW SOE system operating for, at least, 4 years (see Figures 1 and 2 (right)). [1] F. Ramadhani, M. A. Hussain, H. Mokhlis, and S. Hajimolana, “Optimization strategies for Solid Oxide Fuel Cell (SOFC) application: A literature survey,” 2017, Elsevier Ltd . doi: 10.1016/j.rser.2017.03.052. [2] B. Hu, G. Lau, D. Song, Y. Fukuyama, and M. C. Tucker, “Optimization of metal-supported solid oxide fuel cells with a focus on mass transport,” J Power Sources , vol. 555, Jan. 2023, doi: 10.1016/j.jpowsour.2022.232402. [3] V. A. C. Haanappel et al. , “Optimisation of processing and microstructural parameters of LSM cathodes to improve the electrochemical performance of anode-supported SOFCs,” J Power Sources , vol. 141, no. 2, pp. 216–226, Mar. 2005, doi: 10.1016/j.jpowsour.2004.09.016. [4] E. Schropp, G. Naumann, and M. Gaderer, “Hydrogen production via solid oxide electrolysis: balancing environmental issues and material criticality,” Advances in Applied Energy , p. 100194, Oct. 2024, doi: 10.1016/j.adapen.2024.100194. [5] L. Zhao and J. Brouwer, “Life Cycle Analysis of Ceramic Anode-Supported SOFC System Manufacturing Processes,” ECS Trans , vol. 42, no. 1, pp. 247–263, Apr. 2012, doi: 10.1149/1.4705501. [6] L. Smith, T. Ibn-Mohammed, F. Yang, I. M. Reaney, D. C. Sinclair, and S. C. L. Koh, “Comparative environmental profile assessments of commercial and novel material structures for solid oxide fuel cells,” Appl Energy , vol. 235, pp. 1300–1313, Feb. 2019, doi: 10.1016/j.apenergy.2018.11.028. This project receives a Dutch National Growth Fund contribution from the programme NXTGEN HIGHTECH Figure 1