Proton exchange membrane water electrolysis (PEMWE) offers a promising route to low-carbon hydrogen because it can achieve high current densities, uses compact equipment, and adapts quickly to fluctuations in renewable electricity. Nevertheless, PEMWE durability depends not only on the continued presence of its costly platinum-group-metal (PGM) catalysts; at the anode, a nanometer-thick oxide film can break electrical contact between the catalyst and the current collector, so the valuable material remains physically present but becomes electrochemically inaccessible.
This contrasts with a common way of interpreting electrolyzer aging. An increasing operating voltage or falling current does not necessarily indicate that the catalyst has been used up; in some cases, oxidation of the interfacial titanium can hide remaining catalytic activity. Therefore, to improve stack lifetime, reduce hydrogen production costs, and recover scarce iridium, it is essential to distinguish between electrical isolation and true catalyst degradation.
The anode in the PEMWE system features an acidic environment and highly oxidizing potentials, conditions suitable for iridium-based oxygen evolution catalysts and for corrosion-resistant titanium porous transport layers (PTLs). The PTL supplies water, removes the oxygen produced, and carries electrons away from the catalyst layer. Titanium survives mainly because of a protective oxide layer; the native film on its surface, generally a few nanometers thick, prevents further corrosion. The Pilling–Bedworth ratio of titanium dioxide is about 1.7, which helps explain its tendency to form a protective covering layer. However, this ratio alone does not guarantee the film's long-term integrity.
This same protection creates an electrical liability. Stoichiometric TiO₂ is a wide-bandgap semiconductor and has relatively low conductivity, while oxygen-deficient titanium oxides can be considerably more conductive. When anodic polarization is applied, variations in oxide thickness, composition, and defect concentration can increase resistance at the interface between the catalyst and the PTL. As a result, titanium can remain structurally intact while becoming a less effective current collector. The electrode shape worsens this issue. The porous catalyst layer contacts only a few discrete points on a network of titanium fibers or particles. Electrons must pass through the catalyst network to reach these limited contact points, so losing a small number of important connections can render the entire larger catalytic area inoperative.
Electrons can pass through when the oxide layer is thin enough via tunneling; in a simple barrier model, the tunneling probability decreases exponentially with thickness. As a result, even small oxide growth can greatly reduce electrical conduction. Band alignment is also important because interfaces between defective TiO₂ and conductive catalyst materials can create depletion regions and Schottky-like barriers. The behavior of these interfaces depends not only on the work functions of the ideal bulk materials but also on oxide stoichiometry, surface states, interfacial chemistry, and local geometry. As oxidation proceeds, transport may include a combination of tunneling, defect-assisted conduction, and thermally activated processes.
In practice, this increases interfacial contact resistance. Current shifts to better-connected areas, increasing local electrical and transport stresses. Ultimately, some catalyst domains might remain with iridium and appear structurally intact but may lack a conductive pathway sufficient to allow oxygen evolution. This situation is an example of electrical masking, not catalyst exhaustion. Dissolution, agglomeration, detachment, and chemical restructuring are still real degradation mechanisms and can occur simultaneously. The key point is that a performance decline alone is not enough to determine which mechanism dominates.
A credible diagnosis requires complementary measurements, not a single indicator. Electrochemical impedance spectroscopy can reveal increasing resistive losses. A rise in high-frequency resistance is consistent with deteriorating electrical contact, but it is not unique proof: membrane hydration, compression, temperature, and other interfaces also contribute. Comparisons between coated and uncoated PTLs, supported by independent contact-resistance measurements, strengthen the interpretation.
Transmission-line models help identify ionic transport, electronic conduction, and charge-transfer limitations in porous electrodes. Yet the fitted parameters must be checked against physical reality, since different equivalent circuits can yield similar spectra. Online ICP-MS detects iridium dissolved from the system. The low amount of dissolved iridium, compared with the considerable loss in performance, suggests investigating mechanisms other than catalyst removal. Still, outlet measurements do not provide a complete mass balance, since iridium may redeposit or move into the membrane and other components. X-ray spectroscopy and electron microscopy offer complementary information regarding the oxidation state, coordination, particle morphology, and growth of the interfacial oxide. The fact that the catalyst structure is retained supports the idea that catalytic activity has been retained, but it does not prove it on its own. Furthermore, specialized operando measurements do not always match the full-stack conditions. The most compelling evidence for masking is the combination of a retained PGM inventory, increasing interfacial resistance, direct oxide characterization, and performance recovery after reestablishing electrical contact.
Thin platinum or iridium coatings can maintain conductive contact and reduce titanium passivation; however, the advantages of such coatings must be weighed against issues of coating continuity, adhesion, material cost, and durability under realistic operating cycles. Platinum may dissolve during certain potential transients and then redistribute itself. Once dissolved, platinum can precipitate within the membrane when it contacts crossover hydrogen; these deposits may influence gas recombination and peroxide chemistry, and in the presence of certain metal impurities, peroxide can produce radicals that attack the polymer.
A secondary degradation pathway may occur, so the outcome is not certain based on platinum coating; platinum may also cause harmless gas recombination or peroxide decomposition. The overall effect varies with particle distribution, operating conditions, and contamination. PGM-free methods include surface treatments using titanium hydride and conductive titanium suboxides. Some studies show that hydride-forming treatments can reduce resistance growth, but the long-term effects of oxidation and mechanical action still need evaluation. Conductive Magnéli phases such as Ti₄O₇ offer another possibility, but they are not immune to oxidation into less-conductive TiO₂. Similarly, niobium and titanium nitride coatings must show they maintain conductivity and chemical stability over time, not just good initial performance.
Electrical masking wastes both energy and catalyst. When Faradaic efficiency remains approximately constant, hydrogen production is proportional to current, while extra voltage primarily increases electricity consumption. An additional 0.10 V per cell is equivalent to about 2.7 extra kWh per kilogram of hydrogen, or around $0.08 per kilogram at $0.03 per kWh, before accounting for downtime or replacement costs. Although early stack replacement adds further expenses, its impact on the levelized cost of hydrogen varies with utilization, electricity prices, the scope of replacement, and financing arrangements. It is not possible to conclude that passivation alone creates large, universal cost penalties.
A correct diagnosis also helps recover resources. When a considerable amount of iridium stays in a failed assembly, refurbishment or recycling could recover a significant value. Material recovery yield should be determined through material accounting, not assumed based on the failure mechanism. Ultimately, PEMWE durability depends on maintaining both catalytic activity and electrical access to it. A catalyst that is intact but lies behind a resistive titanium oxide interface represents a resource that is not fully used and is not necessarily exhausted. Focusing on the catalyst–PTL interface for both design and diagnosis offers a practical way to achieve longer-lasting stacks and use scarce PGMs more effectively.
