Grooved electrodes for high-power-density fuel cells

The study addresses the challenge of improving power density, durability, efficiency, and cost of PEMFCs for transportation—especially heavy-duty vehicles, where fast refueling, low weight, and high robustness are required. Conventional flat electrodes, formed by mixing Pt/C and ionomer into a porous layer, suffer from tortuous ionomer and pore networks that limit O2 and H+ transport and reduce catalyst utilization. Increasing ionomer content improves proton conduction but penalizes O2 transport, forcing a trade-off. The research hypothesizes that a grooved electrode architecture—high-ionomer-content ridges for H+ transport separated by void grooves for rapid O2 diffusion—can decouple and optimize transport pathways, thereby enhancing performance across a wide range of humidity and operating conditions and improving durability relevant to heavy-duty applications.

Conventional PEMFC electrodes (Wilson and Gottesfeld architecture) remain largely unchanged and are limited by mass transport and catalyst utilization. Prior efforts have explored alternative architectures via micro/nanofabrication, inkjet printing, template-based patterning, and thin-film deposition to improve mass activity, transport, and durability. However, many improvements are confined to narrow operating windows and have not been validated over the broader conditions encountered in transportation (notably ≥0.7 V and low RH). Previous macropore designs often enhanced transport of a single reactant; in contrast, the present work proposes a multifunctional, ordered ridge-and-groove structure to simultaneously enhance H+ and O2 transport, aiming to overcome the traditional trade-off between ionomer content and O2 transport resistance.

- Concept and morphology: Quantified reduction in effective O2 diffusion length by computing Euclidean distances from electrode interior to surface for flat vs grooved morphologies; average distance reduced from 6.0 μm (flat) to 2.5 μm (12 μm spacing) and 0.73 μm (3 μm spacing). Fabricated grooved electrodes by depositing Pt/C and ionomer onto patterned Si templates and transferring to Nafion membranes. Verified ridge/groove morphology via nanoscale X-ray tomography and HAADF-STEM/EDS. - Groove geometries: Tested groove sets 2 μm/6 μm, 1.5 μm/4 μm, 1 μm/3 μm (groove width/period), with grooves perpendicular to flow channels. Ionomer-to-carbon (I/C) ratios studied: 0.9, 1.2, 1.5. Cathode Pt loading ~0.3 mg cm−2. - Performance testing: Polarization curves at 150 kPa, 80 °C, H2/air under 100%, 75%, and 40% RH; current normalized by 5 cm2 geometric area. Evaluated impact of I/C and groove period on performance. - Transport diagnostics: Measured sheet resistance (Rsheet) via EIS in H2/N2 to assess proton transport; O2 transport resistance (Ro2,total) via limiting current in H2/0.5% O2 at multiple absolute pressures and 60% RH; mass-transport resistance (RMT) via EIS in air and helox during operation to separate pressure-dependent and -independent components. - Durability testing: Conducted carbon corrosion AST (500 cycles, 1.0–1.5 V, H2/N2, 100% RH, 80 °C). Monitored CO2 evolution, mass activity, ECSA, HFR, Ro, and Rsheet pre- and post-AST. Verified groove integrity post-degradation via SEM/HAADF-STEM. - Fabrication details: Silicon templates produced by photolithography and deep reactive ion etching (SF6–C4F8 mixed process) to 10–13 μm depth. Catalyst inks (TEC10E40E cathode; TEC10V20E anode) prepared with Chemours ionomer dispersions; coatings via wire-wound rod (cathode) and ultrasonic spraying (anode). MEAs formed by hot pressing and template removal in NaOH; reprotonation in H2SO4. - Modelling: 2D finite element multiphysics simulations in COMSOL integrating electrochemistry, gas/liquid/ionomer water transport, heat, and charge transport across GDL/MPL/electrodes/membrane. Model calibrated on flat electrodes then applied to grooved geometry with identical parameters; groove transport treated as porous domain with pore fraction 1. Predicted iORR and O2 concentration distributions. - Machine learning optimization: Adaptive learning with an ANN (3 hidden layers, 64 neurons, tanh, ridge regularization 0.05) trained on multiphysics outputs for groove design parameters (period w, depth h, opening d1, base d2 parameterized as l1=d1/w, l2=1−d2/d1). Bootstrapped uncertainty quantification (1,000 models). Global optimization via expected improvement to propose new simulations. Dataset grew from 50 to 350 designs; final model performance: RMSE 13.6±3.6 mA cm−2, MAPE 0.6±0.1%, R2 0.993±0.003.

- Geometry and transport path length: Grooves reduced average Euclidean distance to surface from 6.0 μm (flat) to 2.5 μm (12 μm spacing) and 0.73 μm (3 μm spacing), enabling shorter O2 diffusion paths. - Performance vs I/C and RH: Flat electrodes showed optimal I/C≈0.9; higher I/C degraded performance due to O2 transport penalties. Grooved electrodes achieved peak performance at I/C=1.2, surpassing optimized flat electrodes across RH conditions. At 0.7 V, when RH decreased from 100% to 75%: flat I/C=0.9 dropped 20% (1.19→0.95 A cm−2; 0.83→0.66 W cm−2), while grooved I/C=1.2 dropped only 11% (1.28→1.14 A cm−2; 0.90→0.80 W cm−2). Performance improved with decreasing groove period. - Proton transport: Rsheet decreased with increasing I/C for both flat and grooved electrodes; grooved I/C=1.2 achieved ~60% lower H+ transport resistance at 100% RH vs the flat baseline (as summarized in the conclusion). - O2 transport resistance (limiting current): In flat electrodes, pressure-independent Ro2 increased by +96% (I/C 1.2) and +530% (I/C 1.5) vs I/C 0.9. Introducing grooves reduced pressure-independent Ro2 by >50% at I/C 1.2 and 1.5. Grooved I/C=1.2 and flat I/C=0.9 had nearly identical Ro2, indicating grooves restore facile O2 transport at higher I/C. - Mass transport resistance (EIS): Increasing I/C from 0.9 to 1.2 in flat electrodes raised RMT by +430% (air) and +350% (helox). Adding grooves (I/C=1.2) decreased RMT by 66% (air) and 68% (helox). At optimized I/C (flat 0.9 vs grooved 1.2), RMT in helox was nearly identical, confirming effective pressure-independent O2 diffusion in grooved electrodes. - Local reaction uniformity (modelling): Grooved electrodes at I/C=1.2 and 1.5 exhibited more uniform iORR and O2 concentration distributions and lower H+ ohmic losses relative to flat designs; iORR peaked at I/C=1.2 grooved. - Durability (AST 500 cycles): At 0.7 V EOT, grooved (1 μm/3 μm, I/C=1.2) outperformed flat (I/C=0.9 and 1.2) by +44% and +170% current density, respectively. Despite similar carbon loss, ECSA, mass activity, and HFR among EOT electrodes, grooved EOT had lower O2 transport resistances: pressure-dependent and -independent Ro were 12% and 38% lower than flat I/C=0.9 EOT. Grooved Rsheet at 100% RH was 57% lower vs flat I/C=0.9 EOT. Grooves remained structurally present after AST. - Machine learning optimization: Predicted potential performance gains up to +60% vs baseline flat (I/C=0.9) and up to +36% vs current grooved (I/C=1.2) by employing wider and deeper grooves, within fabrication and mechanical constraints.

The grooved architecture directly addresses the longstanding trade-off in PEMFC electrodes between high ionomer content (for H+ transport) and O2 transport resistance. By spatially partitioning the transport pathways—ionomer-rich ridges for protons and void grooves for oxygen—the design preserves rapid proton conduction at higher I/C while mitigating O2 diffusion limitations. Experimental diagnostics (limiting current, EIS) and modelling jointly corroborate that grooves restore pressure-independent O2 transport at high I/C and yield more uniform local ORR rates across the catalyst layer, enhancing catalyst utilization. The benefits are amplified under low RH, a critical regime for automotive-grade PEMFCs, where both O2 and H+ transport typically degrade. Durability testing shows the grooved architecture maintains superior transport characteristics post carbon corrosion, counteracting compaction-induced porosity loss that otherwise impairs O2 access in flat electrodes. Thus, the findings demonstrate that deliberate mesoscale structuring can decouple and optimize multiphase transport, enabling higher power density, better robustness over humidity ranges, and improved end-of-life performance—key advances for heavy-duty transportation applications.

The work introduces a grooved PEMFC electrode architecture comprising high-ionomer-content catalyst ridges separated by empty grooves that provide dedicated pathways for H+ and O2 transport, respectively. Relative to an optimized flat baseline, grooved electrodes reduce H+ transport resistance by about 60% at 100% RH without increasing O2 transport resistance, deliver higher performance across humidity conditions with particularly strong gains at low RH, and exhibit improved durability with lower O2 transport resistance and sheet resistance after carbon corrosion AST. Multiphysics simulations confirm more uniform ORR rates and reduced transport losses in grooved structures. An adaptive machine learning framework identifies substantial headroom—up to 60% performance improvement over baseline and 36% over current grooved designs—by optimizing groove depth and width. Future work should focus on fabricating higher-aspect-ratio, mechanically robust grooved structures and refining manufacturing approaches to realize ML-predicted designs, advancing high power density PEMFCs for transportation.

- Fabrication constraints: Current template-based methods are near their limits for 1 μm/3 μm geometries; wider and deeper grooves predicted by ML may require new fabrication processes. Templates were dissolved during release and not reusable. - Mechanical stability: Increasing ridge aspect ratio may compromise structural integrity under cell compression; systematic studies of compression effects are needed. - Residual transport limits at very high I/C: Even with grooves, I/C=1.5 exhibited lower performance due to local ionomer/agglomerate-induced O2 transport barriers. - Modelling approximations: Simulations used 2D unit-cell representations and treated groove domains as porous media with unity porosity; while validated against experiments, full 3D effects and detailed microstructure may introduce deviations.