A Numerical Study on the Impact of Low Electronic Conductivity on PEMWE Electrolyser Performance

Due to the high scarcity and cost of the catalysts used in proton exchange membrane water electrolysis (PEMWE), i.e. platinum and iridium, it is of paramount importance to maximise their utilisation and lifespan, particularly for the anode catalyst layer (ACL) where iridium is commonly used to catalyse the oxygen evolution reaction (OER). Maximising utilisation requires understanding how the reaction is distributed within the catalyst layer (CL), which is affected by the layer electronic and protonic conductivity, in addition to the activity of the catalyst [1]. Recently, it has been shown that a CL composed of a commonly used IrOx catalyst from Tanaka Kikinzoku Kogyo (TKK) has an electronic conductivity that is three orders of magnitude lower than the protonic [2]. Such a low conductivity may result in the reaction being extremely concentrated in the ACL and therefore allow for a reduction of the catalyst loading. Such a reduction has been demonstrated in the literature [3], where ACLs with loadings of the order of 0.1 mg/cm2 still provide excellent performance when compared to ACLs with the more commonly used loadings of 1-5 mg/cm2 [4]. The improved performance was attributed to the improved distribution of the catalyst due to the use of an optimised deposition method. The impact of the low electronic conductivity was not studied, as the measured ohmic resistance was dominated by the NRE-117 membrane, and the through plane reaction distribution cannot be determined experimentally. As such, this work uses numerical modelling to investigate the impact of the low electronic conductivity on the ohmic resistance of the cell and on the reaction distribution in the ACL. A two-dimensional, macro-homogeneous PEMWE model is implemented in OpenFCST [5]. Charge transport is accounted for using Ohm’s Law, and multi-step reaction kinetic models are used for the hydrogen evolution reaction [6] and the OER [7]. The conductivities of the protonic and electronic phases are taken from recently published ex-situ measurements [2]. The ohmic heating method [8] is used to compute the voltage losses incurred from charge transport. The numerical model is compared to in-house experimentally obtained polarisation curves using a 5 cm2 cell, using a TKK IrOx catalyser in the ACL and an NRE 211 membrane. The results show a close agreement between the experimentally and numerically obtained polarisation curves, with the electronic transport in the ACL incurring the highest voltage loss in the cell. The reaction distribution shows that it is strongly concentrated at the ACL/porous transport layer interface, due to the low electronic conductivity of the IrOx. The model shows that the catalyst loading of the layer to be reduced from 1 mg/cm2 to 0.025 mg/cm2, without significantly reducing the kinetic performance. The overall resistance of the layer was reduced, though further reductions in loading causes kinetic losses to dominate. These trends are in agreement with the data shown by Taie et al. [3]. However, the concentrated reaction distribution causes large gradients in electronic potential within the ACL. As such, part of the CL experiences potential differences between the phases as large as 1.6 V at 1.8 A/cm2, creating a strongly oxidising environment for the catalyst. Tan et al. [9] showed the TKK catalyst degrades significantly faster at 1.6 V compared to 1.53 V, so high current density operation with this catalyst may cause shorter lifespans. The maximum potential difference experienced by the ACL can be reduced if the conductivities of the phases are of a similar order of magnitude. For example, if the ACL has an electronic conductivity ten times smaller than the protonic, instead of one thousand times [2], but still has the same performance at 1.8 A/cm2, the maximum potential difference is reduced to 1.51 V, which could result in a significantly lower degradation rate [9]. This suggests that the conductivity of the catalyst may be crucial to achieving lower degradation rates. References: [1] K. Neyerlin et al., J. Electrochem. Soc., 2007, 154 B631. [2] M. Mandal et al., ACS Appl. Mater. Interfaces, 2020, 12, 44, 49549–49562 [3] Z. Taie et al. ACS Appl. Mater. Interfaces, 2020, 12, 47, 52701–52712 [4] M. Carmo et al., Int. J. Hy. Ener., 2013, 38(12), 4901–4934 [5] M. Secanell, et al., ECS Trans, 2014, 64 (3) , 655 [6] K. Elbert et al., ACS Catalysis, 2015, 5(11), 6764–6772 [7] Z. Ma et al., J. of Electroanalytical Chem., 2018819, 296–305 [8] A. Kosakian et al., Electro. Acta, 2020, 350, 136204 [9] X. Tan et al., Journal of Catalysis, 2019 371, 57–70 Figure 1

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.