Manufacturing-to-Validation Workflows for Repeatable PEM Fuel Cells: A Review and Technology Perspective

1. INTRODUCTION

Hydrogen is often presented as an energy carrier that can bridge primary energy resources and end-use services through an integrated value chain: production, storage and transport, and conversion at the point of use as shown in Figure 1.^[1] While interest in hydrogen is rising across policy and industry, the practical impact of hydrogen depends on how these stages fit together in real systems, particularly when cost, infrastructure readiness, and delivered hydrogen quality are considered.

Close

Figure 1: A simplified hydrogen value chain from production to end-use conversion.

Export to PPT

In this context, proton exchange membrane fuel cells (PEMFC) are among the most mature fuel-cell technologies for transportation and distributed power. Still, moving from “good lab results” to reliable, scalable products is not primarily a chemistry problem—it is a manufacturing and workflow problem. In real production settings, two cells built from the same bill of materials can perform differently because small variations in coating quality, porosity, interfaces, compression, hydration history, and even test conditioning accumulate across the membrane electrode assembly (MEA) and stack. For PEMFCs, therefore, performance and durability are tightly coupled to how components are fabricated, assembled, activated, and validated.

This review takes a practical, workflow-centric view of PEMFC development. Instead of discussing each component in isolation, it connects (i) fabrication routes and their critical process parameters, (ii) typical defects and failure modes, (iii) how those defects appear in characterisation and diagnostics, and (iv) how the full workflow shapes repeatability, yield, and long-term stability. The focus is on manufacturing-relevant steps: GDL, CCM/CL, and flow fields and bipolar plates—followed by quality control, assembly, activation (break-in), and performance evaluation. Where relevant, we highlight what changes when methods scale from laboratory to industrial practice (e.g., roll-to-roll coating, in-line inspection, and basic statistical process control). The workflow-centric process map is shown in Figure 2.

Close

Figure 2: Workflow-centric process map for proton exchange membrane fuel cells (PEMFC) development and validation.

Export to PPT

A narrative literature search is performed using major scientific databases and publisher platforms, supported by citation tracking from key papers and reviews. The selection prioritised studies with clearly reported processing and operating conditions and explicit links between manufacturing choices, diagnostic signatures, and PEMFC performance or durability.

The paper is organised as follows: Section 2 introduces PEMFC architecture and reviews fabrication and process evolution of key components (GDL, CCM/catalyst layers, and flow fields/bipolar plates) and discusses quality control and characterisation. Section 3 covers assembly and repeatability drivers. Section 4 summarises activation protocols, and Section 5 consolidates testing and performance evaluation practices. Section 6 concludes the review and outlines a technology development perspective with near-term priorities.

2. PEMFC ARCHITECTURE AND MANUFACTURING-RELEVANT COMPONENTS

In this section, a PEMFC is introduced in its simplest architectural form as a layered device built around the membrane–electrode assembly (MEA) and the adjacent hardware. Its structure is shown in Figure 3.^[2] The polymer electrolyte membrane acts as the ion-conducting separator; the catalyst layers provide the reaction sites; and the porous transport layers (typically the GDL) enable gas access and support water/heat transport while maintaining electronic contact. On the hardware side, flow fields distribute reactants and help remove water and heat, while bipolar plates provide current collection and mechanical support at the stack level. Building on this architecture, the following review focuses on the manufacturing-relevant layers and interfaces that most strongly control repeatability and performance, with emphasis on the fabrication and evolution of the GDL, the CCM and catalyst layers, and the flow fields and bipolar plates.

Close

Figure 3: Proton exchange membrane fuel cells (PEMFC) layers in detail. GDL: Gas diffusion layer, CCM: Catalyst-coated membranes, MPL: Microporous layer.

Export to PPT

To maintain a consistent framework across the reviewed PEMFC components, each layer/component is discussed according to the sequence outlined in Figure 4.

Close

Figure 4: Schematic framework showing the structured sequence used to review the main PEMFC layers/components, BPP: Bipolar plate, HFR: High-frequency resistance, ECSA: Electrochemically active surface area, ICR: Interfacial contact resistance ASR: Area-specific resistance.

Export to PPT

3. ASSEMBLY: PROCESS EVOLUTION AND KEY REPEATABILITY DRIVERS

MEA assembly has evolved from empirical hot-pressed gas diffusion electrode (GDE), catalyst-coated substrate (CCS) structures toward CCM, decal-transfer, printed, roll-to-roll, and automated five-/seven-layer assemblies, where ink preparation, coating, drying, cutting, framing, alignment, and pressing are treated as an integrated manufacturing chain.^[20,30,55] The literature shows that compression is the dominant repeatability driver: insufficient load increases leakage and interfacial/contact resistance,^[56,57] whereas excessive pressure over-compresses the GDL, reduces porosity, increases mass-transport resistance, and distorts oxygen, water, temperature, and current distributions.^[58,59] Gasket/sub-gasket design, bolt-load distribution, layer alignment, swelling, and dimensional tolerances further control sealing integrity, edge protection, pressure uniformity, and local mechanical stress.^[31,60] In parallel, MEA development has shifted from simple layer bonding to deliberate interface engineering through CCM architectures, Nafion/ionomer interlayers, hot pressing, patterned membranes, direct membrane deposition, and improved PEM/CL contact.^[55,61] These improvements reduce ohmic and transport losses, enhance catalyst utilisation, and improve durability through better adhesion and lower delamination risk.^[20,55] Accordingly, QC now relies on pressure mapping, leakage/crossover, EIS/HFR, CV/ECSA, polarisation, thickness/loading uniformity, adhesion, IR/XRF inspection, rheology, and in-situ water or stress diagnostics.^[30,31,62] For manufacturing translation, these assembly descriptors should be treated as integration release criteria that connect alignment tolerance, gasket/sub-gasket dimensions, clamping pressure, compression distribution, bonding conditions, and stack-up variation to allowable limits in leakage/crossover, interfacial resistance, pressure uniformity, adhesion, and early electrochemical response. For porous media and interfaces, such screening should explicitly include thickness reduction under load, through-plane/contact resistance, and compression-dependent diagnostic response, since these measurements are more predictive of assembled-cell behaviour than free-standing properties alone. This distinction is important because assembly-induced deviations can otherwise be misattributed to catalyst, membrane, or GDL degradation, although their origin is mechanical, interfacial, or tolerance-related. A summarised comparison is shown in Supplementary Table 2.

4. ACTIVATION AND BREAK-IN PROTOCOLS

The activation process of PEMFCs has evolved from conventional, time-consuming procedures to faster and more efficient protocols. Early studies focused on voltage-control and current-control methods, which were effective but costly and lengthy.^[63] These approaches mainly aimed to establish the triple-phase boundary and improve proton conductivity within the membrane electrode assembly (MEA).^[64] With increasing demand for faster conditioning, hybrid methods were introduced to combine activation strategies while reducing time. For instance, hydrogen pumping combined with current control shortened activation to nearly 30 minutes and improved electrochemical behaviour by reducing ohmic resistance and Tafel slope.^[65] Later, high-current pulse activation and gas-starvation methods further decreased activation time and hydrogen consumption, making them more attractive for industrial use,^[66] while gas starvation also improved cell voltage.^[67] More recent studies highlight combined protocols involving temperature control, hydrogen pumping, and air starvation as cost-effective, environmentally favourable, and suitable for large-scale PEMFC production.^[68] Overall, activation research has shifted toward rapid combined methods that improve performance and economic efficiency.^[63] From a manufacturing-to-validation perspective, the value of these rapid protocols is not only their shorter duration, but also their potential to reduce operator- and laboratory-dependent conditioning variability when they are applied within clearly reported limits. Therefore, hydrogen pumping combined with current control, high-current pulse activation, and controlled gas-starvation or air-starvation approaches should be considered as practical activation options where compatible with the MEA design and test objective. To avoid procedural drift, studies should report the activation sequence, duration, gas composition, humidity, temperature, current/voltage limits, and stabilisation criteria before subsequent performance comparison.

5. TESTING AND PERFORMANCE EVALUATION

Testing practices for PEMFCs have evolved from laboratory-specific performance measurements toward harmonised, application-relevant protocols as the technology moved closer to deployment.^[69] Early testing approaches relied on polarisation curves and steady-state measurements to rank MEA performance, but large scatter between laboratories limited cross-study comparability and obscured manufacturing-driven effects.^[70] This motivated the development of standardised single-cell test frameworks, particularly for automotive-relevant conditions, where operating windows, controlled deviations, and durability metrics are explicitly defined. Harmonised protocols formalised testing around reference operating conditions, controlled deviations, and defined performance checkpoints representing kinetic, ohmic, and mass-transport regimes. Testing therefore shifted from reporting absolute performance to assessing sensitivity to operating conditions, reversible versus irreversible degradation, and repeatability under dynamic load cycles. More recent studies further extended testing beyond protocol definition toward diagnostic-rich testing strategies and improved reporting discipline under varying operating conditions,^[2] reframing testing from a validation step into a process-aware diagnostic tool for benchmarking and technology readiness assessment.

6. CONCLUSION AND TECHNOLOGY PERSPECTIVE

The reviewed evidence supports a practical shift from component-by-component optimisation toward workflow-controlled PEMFC manufacturing and validation. The central implication is that performance scatter and durability variability are not solely materials-limited; they are often workflow-amplified through small deviations in coating/drying history, porous-media compression state, interfacial contact, sealing strategy, and test conditioning. Technology programs should therefore translate scientific insights into manufacturable control plans: defined process windows for CL/CCM coating and drying, compression-aware specifications for diffusion media and interfaces, and geometry/tolerance control for flow fields and bipolar plates. In practice, this means converting “methods” into release criteria by pairing each critical process variable with a measurable acceptance metric and an allowable range. This control-plan logic should be applied across the main manufacturing and integration stages, including GDL/MPL fabrication, CL/CCM ink preparation, coating and drying, flow-field/bipolar-plate production, and MEA assembly, so that variability can be screened before it is amplified during activation, testing, or stack integration.

Three gaps repeatedly limit transfer from R&D to scalable production. First, industrial screening still relies heavily on limited or weakly predictive metrics, while defect-to-diagnostic links remain inconsistently implemented, reducing early detection of non-uniformity, cracking, interfacial losses, and transport constraints. Second, conditioning and activation practices vary widely across laboratories and organisations; inconsistent break-in and reporting can mask manufacturing effects and undermine comparability. Third, stack-level repeatability is often treated as a downstream integration problem rather than co-specified at the design stage: flow distribution, sealing/compression strategy, and plate tolerances are not always coupled to MEA requirements despite their direct influence on local hot spots, water management, and degradation.

Near-term progress is most likely through diagnostic-linked quality control with explicit go/no-go thresholds, supported by statistical process control and process-capability targets. Practical pathways include (i) spatial mapping of CL/CCM loading and thickness, (ii) quantitative crack/roughness and coating-defect metrology, (iii) compression-dependent resistance and thickness-under-compression checks for porous media and interfaces, and (iv) rapid electrochemical screening (e.g., EIS-informed metrics) that is tied to known defect signatures. Design-for-manufacturing should explicitly couple gasket selection, target compression distribution, and plate/channel tolerances to repeatability specifications early in the design cycle. In practical terms, gasket material and thickness should be selected together with the target GDL compression range, allowable pressure-distribution variation across the active area, and plate/channel dimensional tolerances. These specifications should be verified through tolerance-stack analysis, pressure mapping, leakage/crossover testing, contact-resistance measurements, and representative single-cell builds before stack-scale integration, enabling faster qualification cycles and higher yield. Finally, reproducibility will benefit from adopting a minimum reporting standard across fabrication, assembly, activation, and testing. At minimum, this should document ink and coating conditions, drying and transfer parameters, component thickness/loading and defect metrics, assembly alignment and compression conditions, activation sequence and stabilisation criteria, and testing operating windows. Such reporting would allow observed performance differences to be attributed to controllable workflow factors rather than undocumented procedural drift.

These directions align PEMFC scale-up with advanced manufacturing, industrial quality engineering, metrology, and data-driven process monitoring. Progress requires coordinated contributions from electrochemistry (diagnostic interpretation), materials science (structure–property relations under compression and hydration), mechanical design (sealing and tolerance management), manufacturing engineering (process windows and in-line inspection), and data science (SPC, capability analysis, machine learning, and defect-to-performance models). In this context, linked process, inspection, and electrochemical datasets can be used to identify which defect combinations are most predictive of HFR increase, ECSA loss, mass-transport limitation, voltage instability, or early durability decay, while SPC can track process drift in coating uniformity, compression distribution, plate tolerances, and diagnostic acceptance limits. Together, this interdisciplinary integration can shorten learning cycles, improve transferability across labs, and accelerate the emergence of reliable PEMFC product platforms.