TY - JOUR
T1 - 3D generation and reconstruction of the fuel cell catalyst layer using 2D images based on deep learning
AU - Liu, Xuanchen
AU - Park, Kayoung
AU - So, Magnus
AU - Ishikawa, Shota
AU - Terao, Takeshi
AU - Shinohara, Kazuhiko
AU - Komori, Chiyuri
AU - Kimura, Naoki
AU - Inoue, Gen
AU - Tsuge, Yoshifumi
N1 - Funding Information:
Owing to their high energy transfer efficiency, safety, and the fact that they do not emit CO or NOx, polymer electrolyte fuel cells (PEFCs) have attracted much attention as electric storage devices [1]. Among the PEFC subunits, the catalyst layer (CL) plays an important role in electron transfer, gas diffusion, proton transport, and moisture transport. The catalyst consists of carbon black (CB) particles, which act as electron conductors and as support for platinum nanoparticles; a polymer membrane that conducts the protons; platinum nanoparticles as the site of electrochemical reactions; and pores that provide a pathway for liquid water to transfer into or out of the CL. The structure of the CL is determined by its micropore structure, as defined by International Union of Pure and Applied Chemistry (IUPAC) according to pore diameter as follows: micropores (<2 nm), mesopores (2–50 nm), and macropores (>50 nm). Owing to the diverse functions of pores with varying diameters, the proportion of pores with varying diameters will lead to distinct battery performance. Salari et al. [2] measured the ex-situ and in-situ gas diffusion of ten different CL designs and found that increasing the CL porosity and primary pore size linearly increased the ionomer-water diffusivity and reduced the total oxygen diffusion resistance. In addition to primary pores, the role of secondary pores has also been thoroughly studied. Park et al. [3] studied the effect of secondary pores on the performance of cathode CLs, and demonstrated that abundant pore formation facilitates the transport of oxygen through the CL. To optimize the structure of the CL, Hou et al. [4] proposed an ideal structure design that can simultaneously ensure a large active catalyst area and extremely low transport loss, significantly enhancing the performance. Therefore, understanding the 3D microstructure of the catalyst layer is crucial for developing PEFCs.This research was supported by the New Energy and Industrial Technology Development Organization (NEDO), Japan. FIB-SEM fabrication/observation was supported by NIMS Low Carbon Research Network Japan sponsored by the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.
Funding Information:
This research was supported by the New Energy and Industrial Technology Development Organization (NEDO), Japan . FIB-SEM fabrication/observation was supported by NIMS Low Carbon Research Network Japan sponsored by the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.
Publisher Copyright:
© 2022 The Authors
PY - 2022/3
Y1 - 2022/3
N2 - The catalyst layer (CL) being the site of electrochemical reactions, is the core subunit of the membrane electrode assembly (MEA) in polymer electrolyte fuel cells (PEFCs). Thus, the porous structure of the CL has a significant influence on oxygen transfer resistance and affects the charge/discharge performance. In this study, the three-dimensional (3D) porous structure of the catalyst layer is reconstructed based on the deep convolutional generative adversarial network (DCGAN) deep learning method, utilizing focused ion beam scanning electron microscopy (FIB-SEM) microstructure graphs as training data. Each set of spatial-continuous microstructure graphs, generated by DCGAN with interpolation in latent space, is applied to build a unique 3D microstructure of the CL without the use of real FIB-SEM data. Meanwhile, distinct interpolation conditions in the DCGAN are discussed to optimize the ultimate structure by approaching the structural information to real data, including that of porosity, particle size distribution, and tortuosity. Moreover, the comparison of real and generated structural data reveal that the data generated by DCGAN shows an adjacency relationship with real data, indicating its potential applicability in the field of electrochemical simulation with reduced situational costs.
AB - The catalyst layer (CL) being the site of electrochemical reactions, is the core subunit of the membrane electrode assembly (MEA) in polymer electrolyte fuel cells (PEFCs). Thus, the porous structure of the CL has a significant influence on oxygen transfer resistance and affects the charge/discharge performance. In this study, the three-dimensional (3D) porous structure of the catalyst layer is reconstructed based on the deep convolutional generative adversarial network (DCGAN) deep learning method, utilizing focused ion beam scanning electron microscopy (FIB-SEM) microstructure graphs as training data. Each set of spatial-continuous microstructure graphs, generated by DCGAN with interpolation in latent space, is applied to build a unique 3D microstructure of the CL without the use of real FIB-SEM data. Meanwhile, distinct interpolation conditions in the DCGAN are discussed to optimize the ultimate structure by approaching the structural information to real data, including that of porosity, particle size distribution, and tortuosity. Moreover, the comparison of real and generated structural data reveal that the data generated by DCGAN shows an adjacency relationship with real data, indicating its potential applicability in the field of electrochemical simulation with reduced situational costs.
UR - http://www.scopus.com/inward/record.url?scp=85125660527&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85125660527&partnerID=8YFLogxK
U2 - 10.1016/j.powera.2022.100084
DO - 10.1016/j.powera.2022.100084
M3 - Article
AN - SCOPUS:85125660527
VL - 14
JO - Journal of Power Sources Advances
JF - Journal of Power Sources Advances
SN - 2666-2485
M1 - 100084
ER -