Stochastic Microstructure Reconstruction of a Binder/Carbon Fiber/Expanded Graphite Carbon Fiber Paper for PEMFCs Applications: Mass Transport and Conductivity Properties

Abstract

Mass transport and conductivity properties of the carbon fiber papers (CFPs) used in gas diffusion layers are among the critical parameters that affect a proton exchange membrane fuel cell (PEMFC) performance. Here, instead of carbonization/graphitization in the commercial CFPs manufacturing steps, the expanded graphite (EG) is numerically added to the CFP substrate to compensate and enhance the conductivity. The microstructures of different polymer binder/carbon fiber (CF) composites with and without EG are reconstructed using a digital 3D reconstruction technique. The binder phase is discriminated as an individual phase using image processing, and the effects of different binder percentages are demonstrated. In comparing the properties, the reconstructed CFP exhibited satisfactory properties against the values of the commercial CFPs in the literature and our conducted experimental tests. For the binder/CF/EG CFPs, 1.9 W/mK of effective thermal conductivity in the through-plane direction and also 30.2 m .cm of the in-plane electrical resistivity are reached, compared to 0.48-1.8 W/mK (for Sigracet and Toray papers) and 12.9 m .cm (for untreated Toray TGP-H-60), in the conventional CFPs. Overall, adding EG to the CFPs substrate provides valuable properties and can be considered as a proper replacement in future design and manufacturing of CFPs.

Tailored Porous Transport Layers for Optimal Oxygen Transport in Water Electrolyzers: Combined Stochastic Reconstruction and Lattice Boltzmann Method

“…Based on available data in the literature and online by the manufacturer (Bekaert), the typical porosity of Ti felt PTLs is in the range of 0.40–0.90, the fiber diameter 10 μm–50$${{\rm { \hskip0.17em\hskip0.17em{\rm \mu} m}}}$$ , and the thickness 100 μm–2 $${{\rm { mm}}}$$ [9,10] . To generate the virtual PTLs with customized structures, a stochastic reconstruction algorithm implemented in MATLAB R2021b was applied, based on well‐described methods from literature [53–55] . Figure 2b shows the flow chart of the generation of a titanium felt‐based PTL within this work.…”

“…[9,10] To generate the virtual PTLs with customized structures, a stochastic reconstruction algorithm implemented in MATLAB R2021b was applied, based on well-described methods from literature. [53][54][55] Figure 2b shows the flow chart of the generation of a titanium felt-based PTL within this work. In order to customize the structural characteristics of stochastically generated PTLs, the PTL structural parameters need to be input in the code, including matrix domain size, porosity, fiber radius, fiber length, and anisotropy parameter.…”

Gas Diffusion Layer for Proton Exchange Membrane Fuel Cells: A Review

“…Carbon fiber paper is the most widely used material for MPS due to its ease in shaping a thin layer with a uniform porous structure, its high electrical/thermal conductivity and its excellent chemical/thermal stability [34][35][36]. The carbon fibers used in MPS are generally 5-15 µm in diameter, and the thickness of the carbon fiber paper thereby prepared typically has a thickness within 200-400 µm and an average pore size of tens of micrometers [37][38][39].…”

Mechanical and Thermal Behavior of Fibrous Carbon Materials

“…As the temperature of the carbon fiber production process increases, the concentration of structural defects is reduced and crystal order is improved [ 21 ], which results in increased thermal conductivity. Recent studies on commercially available polyacrylonitrile (PAN)-based carbon fibers have shown that, at room temperature, the thermal conductivity of the fibers along their axis ranges from about 5 up to 156 W·m −1 ·K −1 [ 22 , 23 ]. The thermal conductivity values of carbon fibers based on regenerated cellulose (Rayon) along the axis have been found to be in the range of 5 to 15 W·m −1 ·K −1 [ 24 ].…”