Structure-based prediction of gas diffusion property of catalytic layer of proton exchange membrane fuel cells via manifold learning and X-ray ptychographic nano-computed tomography
Source abstract
Abstract
Proton-exchange membrane fuel cells (PEMFCs) have attracted significant attention as a promising technology for clean and efficient power conversion, contributing to the realization of a sustainable, carbon-neutral future. To advance this technology, a data-driven framework is proposed to construct structure–property relationships between gas diffusion and the three-dimensional structure of the catalyst layer (CL) within the catalyst-coated membrane of a PEMFC. Our approach uses manifold learning to extract the intrinsic structural features of the CL implicitly, thereby eliminating the need for manual feature engineering. To significantly enhance prediction performance, manifold learning is performed in a high-dimensional space defined by the logarithmic power spectra of the porous structures, which effectively capture features relevant to gas diffusion. The framework accurately predicts gas-diffusion coefficients with relative errors below 10%, even in low-resolution regimes where direct simulations fail. Validated using hard X-ray ptychographic nano-CT data acquired at the NanoTerasu facility, these findings establish a robust foundation for a new paradigm in the data-driven design of materials and devices with complex hierarchical architectures and in the development of digital twins for complex functional materials.
Source: openalex_inverted_index · Quality: baseline
Research context
OpenAlex topics: Fuel Cells and Related Materials · Electrocatalysts for Energy Conversion · Machine Learning in Materials Science
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
References (69)
Show source-provided references
1. Structure, property, and performance of catalyst layers in proton exchange membrane fuel cells10.1007/s41918-022-00175-1 ↗
2. Enhanced water management in the cathode of an air-breathing PEMFC using a dual catalyst layer and optimized gas diffusion and microporous layers10.1016/j.ijhydene.2015.01.091 ↗
3. Influence of structural modification of micro-porous layer and catalyst layer on performance and water management of PEM fuel cells: hydrophobicity and porosity10.1002/fuce.201900203 ↗
4. Effect of the porosity distribution on liquid water transport in the gas diffusion layer of PEMFC10.1016/j.electacta.2021.137814 ↗
5. Effect of through-plane pore gradient GDL on water distribution of PEMFC10.1016/j.ijhydene.2017.12.007 ↗
6. Pt content and pore structure gradient distributed catalyst layer to improve PEMFC performance10.1016/j.ijhydene.2016.11.002 ↗
7. Effect of Pt loading and catalyst type on pore structure of porous electrodes in PEM fuel cells10.1016/j.energy.2018.02.134 ↗
8. Analysis of catalyst layer microstructures: from imaging to performance10.1002/fuce.201600008 ↗
9. Using MRT lattice Boltzmann method to simulate gas flow in simplified catalyst layer for different inlet–outlet pressure ratios10.1016/j.ijheatmasstransfer.2015.04.031 ↗
10. X-ray ptychography on low-dimensional hard-condensed matter materials10.1063/1.5045131 ↗
11. An improved ptychographical phase retrieval algorithm for diffractive imaging10.1016/j.ultramic.2009.05.012 ↗
12. Reconstructing state mixtures from diffraction measurements10.1038/nature11806 ↗
13. Phase retrieval based on total-variation-regularized poisson model for X-ray ptychographic imaging of low-contrast objects10.1107/S1600576722005234 ↗
14. Subgradient-projection-based stable phase retrieval algorithm for X-ray ptychography10.1107/S1600576724004709 ↗
15. Ptychographic phase retrieval via deep-learning-assisted iterative algorithm10.1107/S1600576724006897 ↗
16. Chemistry of cements revealed by synchrotron ptychographic nano-imaging10.1107/S2052252519003774 ↗
17. Capturing meso- and nanoscale structural evolution in clay minerals during dehydration10.1038/s41467-022-32373-2 ↗
18. Multiscale operando analysis of Li-ion battery cathodes by X-ray ptychography and computational modeling10.1021/acs.jpclett.1c01445 ↗
19. Visualization of crystallization pathways in nanoscale detail by in situ X-ray ptychography10.1039/D0CC06101H ↗
20. Cellular distribution and chemical composition of silica in sugarcane by X-ray ptychography10.1038/s41598-020-63093-6 ↗
21. Nanoscale dynamics of living cells monitored by X-ray ptychography
22. Three-dimensional imaging of biological tissue by cryo X-ray ptychography10.1038/s41598-017-05587-4 ↗
23. Mapping organelle architecture in cells with ptychographic coherent diffractive imaging10.1038/srep06796 ↗
24. Correlative 3D imaging of subcellular structures using cryo-ET and X-ray ptychography
25. Oxygen-diffusion-driven oxidation behavior visualized by X-ray spectro-ptychography10.1038/s42004-019-0147-y ↗
26. What can machine learning help with microstructure-informed materials modeling and design?10.1557/s43577-024-00797-4 ↗
27. Deep materials informatics: applications of deep learning in materials science10.1557/mrc.2019.73 ↗
28. Accurately predicting transport properties of porous fibrous materials by machine learning methods
29. Predicting permeability via statistical learning on higher-order microstructural information10.1038/s41598-020-72085-5 ↗
30. Manifold learning: what, how, and why10.1146/annurev-statistics-040522-115238 ↗
31. Nonlinear dimensionality reduction in molecular simulation: the diffusion map approach10.1016/j.cplett.2011.04.066 ↗
32. Trajectories of the ribosome as a Brownian nanomachine
33. Growth of cuprous oxide particles in liquid-phase synthesis investigated by X-ray laser diffraction10.1021/acs.nanolett.8b02153 ↗
34. Trajectories of structural dynamics in biomolecules by manifold embedding10.1038/s41467-020-18403-x ↗
35. Extraction of molecular information from experimental data on liquids using manifold learning10.1016/j.molliq.2024.126251 ↗
36. Interface-packing analysis of F1-ATPase using integral equation theory and manifold learning10.1016/j.physa.2024.130201 ↗
37. 1982
38. PorousGen: an efficient algorithm for generating porous structures with accurate porosity and uniform density distribution10.1016/j.commatsci.2025.114478 ↗
39. scikit-image: image processing in python10.7717/peerj.453 ↗
40. SciPy 1.0: fundamental algorithms for scientific computing in python10.1038/s41592-019-0686-2 ↗
41. Dynamic nanoimaging of extended objects via hard X-ray multiple-shot coherent diffraction with projection illumination optics10.1038/s42005-021-00539-x ↗
42. Electronic cam control technology for maintaining output beam height from symmetrical layout type double-crystal monochromator10.1088/1742-6596/3010/1/012091 ↗
43. Single-frame coherent diffraction imaging of extended objects using triangular aperture10.1364/OE.414341 ↗
44. Phase retrieval framework for direct reconstruction of projected refractive index applied to ptychography and holography10.1364/OPTICA.447021 ↗
45. Ptychographic overlap constraint errors and limits of numerical recovery using conjugate gradient descent10.1364/OE.22.001452 ↗
46. Translation position determination in ptychographic coherent diffraction imaging10.1364/OE.21.013592 ↗
47. Quantitative X-ray phase nanotomography10.1103/PhysRevB.85.020104 ↗
48. Adaptive local thresholding for detection of nuclei in diversity stained cytology images
49. Fiji: an open-source platform for biological-image analysis10.1038/nmeth.2019 ↗
50. A global geometric framework for nonlinear dimensionality reduction10.1126/science.290.5500.2319 ↗
51. Scikit-learn: machine learning in python
52. Scikit-learn Isomap documentation, https://scikit-learn.org/1.6/modules/generated/sklearn.manifold.Isomap.html (accessed January. 27, 2026).
53. Joblib documentation, https://joblib.readthedocs.io/en/stable/(accessed January 27, 2026).
54. Jolliffe · 2002
55. Scikit-learn PCA documentation, https://scikit-learn.org/1.6/modules/generated/sklearn.decomposition.PCA.html (accessed January. 27, 2026).
56. UMAP: uniform manifold approximation and projection for dimension reduction10.21105/joss.00861 ↗
57. OpenPNM: a pore network modeling package10.21105/joss.01296 ↗
58. Geodict Software, Math2Market GmbH, https://www.geodict.com (accessed January 6, 2026).
59. Array programming with NumPy10.1038/s41586-020-2649-2 ↗
60. Swin transformer-based transfer learning model for predicting porous media permeability from 2D images10.1016/j.compgeo.2024.106177 ↗
61. Universal neural network structure for predicting heat, flow, and mass transport in 3D porous media10.1016/j.ijheatmasstransfer.2025.126688 ↗
62. Machine and deep learning for estimating permeability of complex carbonate rock from X-ray micro-CT10.1016/j.egyr.2021.02.065 ↗
63. 3D convolutional neural network predicts permeability of gas diffusion layer materials from image data
64. Prediction of effective diffusivity of porous media using deep learning based on sample structure information self-amplification10.1016/j.egyai.2020.100035 ↗
65. Improved CNN for predicting porous media permeability from rock thin sections
66. Convolutional neural network-based prediction of effective diffusivity from microscope images
67. Machine learning assisted prediction of porosity and related properties using digital rock images10.1021/acsomega.3c10131 ↗
68. Deep learning for diffusion in porous media10.1038/s41598-023-36466-w ↗
69. Large-scale statistical learning for mass transport prediction in porous materials using 90,000 artificially generated microstructures10.3389/fmats.2021.786502 ↗