New insights into the electrochemical hydrogen oxidation and evolution reaction mechanism

Let me start with the riddle that has bothered electrocatalysis for decades. Take the simplest reaction we ask a catalyst to perform with hydrogen: split H2 into protons and electrons, or run it in reverse to make H2. Same metal, same basic steps. Yet in acid, it's blazingly fast, and in base, it slows down by roughly two orders of magnitude. That's not a subtle tweak. That's a brick wall. So what's changing when you swing the pH? The metal's intrinsic chemistry, or the microscopic environment just outside the surface? Durst, Siebel, Simon, Hasché, Herranz, and Gasteiger decided to force that question to ground. Their wager was simple: if you can strip away mass transport and local pH gradients, and if you measure on the three best monometallic workhorses—platinum, iridium, and palladium on carbon—you can see whether alkaline conditions really demand a different mechanism or just reshape the same one. They put one more conceptual stake in the ground: the hydrogen you sense in voltammetry at low potentials, the so-called underpotential-deposited hydrogen, H-UPD, is the relevant surface intermediate. You don't need to invoke an extra "overpotential" hydrogen, H-OPD, to explain the rates. To get clean kinetics, they built a three-tool setup that's as much about what you don't measure as what you do. First, a hydrogen pump polymer electrolyte membrane fuel cell runs in a mode where hydrogen is supplied on both sides. In that configuration, charge transfer is the only game in town; diffusion cannot bottleneck the current. Second, a rotating disk electrode in alkaline electrolyte containing zero point one molar sodium hydroxide, where they can push convection hard and correct for any remaining mass transport losses. And third, impedance spectroscopy, because the shape of the frequency response tells you whether you're watching electrons hop across the interface or mass sloshing around in solution. There's a reason for all this belt-and-suspenders caution. As Auinger and colleagues showed earlier, if you don't control transport, local pH changes near the electrode can masquerade as kinetic pH effects. You'll think the elementary step got slower in base, when in fact the environment just got weird. When they put platinum, iridium, and palladium through this gauntlet at three hundred thirteen kelvin, a clear picture snapped into focus. In both electrolytes, platinum is the fastest, then iridium, then palladium. In acid, the platinum on carbon shows an exchange current density—think of it as the baseline turnover rate at zero overpotential—of about two hundred milliamps per square centimeter of platinum surface. Palladium and iridium are way lower in acid, roughly zero point two to zero point eight milliamps per square centimeter for palladium and about zero point two milliamps per square centimeter for iridium. Now comes the kicker: swing to base and everything drops by around a hundred-fold. On platinum in zero point one molar sodium hydroxide, the exchange current lands near one milliamps per square centimeter. The ranking doesn't flip. It just shifts down, hard. That's the phenomenology. But what does it say about mechanism? Impedance provides a clue by connecting what you see in hydrogen adsorption—H-UPD—to what you see in hydrogen oxidation and evolution. On platinum surfaces in acid, the charge-transfer resistance you infer from the H-UPD feature is tiny: about zero point zero three to zero point zero five ohm square centimeters for platinum one hundred eleven and polycrystalline platinum. Translate that into the equivalent exchange current for the elementary charge-transfer step—the Volmer step, where a proton plus an electron adsorb as a surface H—and you get ballpark values of roughly eight hundred fifty and five hundred milliamps per square centimeter. That's the same order of magnitude as platinum's overall hydrogen oxidation and evolution exchange current in acid. This suggests you're watching the same thing in two mirrors: H-UPD kinetics and the rate-determining Volmer chemistry are one and the same. Carry that logic into base and it stays consistent. The platinum exchange current collapses by roughly two orders of magnitude, and the voltammetric estimates of exchange current on stepped platinum in alkaline—on the order of zero point two to zero point five milliamps per square centimeter—sit close to what they measure for platinum on carbon in base, about one milliamps per square centimeter. The rate-limiting step, the Volmer proton-electron transfer to or from an adsorbed hydrogen, looks like the bottleneck. Not a separate recombination of two surface H atoms—the classic Tafel step. Not some exotic, hydroxide-assisted detour. Even the kinetic fingerprints agree: the observed behavior lines up with a Volmer or Heyrovsky control, not a Tafel-limited regime. So if the steps don't change, what does? The binding energy of surface hydrogen shifts with pH. That's the through-line in this work. You can actually watch it in the voltammetry. On platinum single crystals with steps—platinum five hundred fifty-three and platinum five hundred thirty-three—and on polycrystalline platinum, the H-UPD peak creeps positive by about ten to eleven millivolts per pH unit. That's a small move per unit, but carry it from pH zero to pH thirteen and it adds up to a change in hydrogen binding energy of roughly twelve point five to thirteen point five kilojoules per mole for a one-electron process. If you then invoke a standard kinetic link—the Brønsted-Evans-Polanyi relationship, which says that activation energy scales with binding energy—you predict a rate change on the order of one hundred twenty to two hundred fold. Durst and colleagues measure an actual drop of about two hundred ten fold in the exchange current over that same pH swing. That agreement isn't an accident. It's a quantitative sign that hydrogen binding is the right descriptor, and pH shifts the energy landscape rather than the machinery of the reaction itself. This has a second, sharper edge because it directly undercuts a popular storyline: that more oxophilic metals—those that hold onto oxygenated species or hydroxide more strongly—should be better in base because hydroxide helps the reaction along. If that were true, iridium should shine in alkaline media. It doesn't. Across both acid and base, the order is platinum, then iridium, then palladium. The iridium surface may be more hospitable to oxygenated adsorbates. That does not translate into faster hydrogen oxidation or evolution in base. Durst's team goes one step further and asks why the oxophilicity idea got traction in the first place. One suspect is surface history. If you start collecting polarization curves before the surface has stabilized, you can be fooled. They point to early-cycle data reported by Strmcnik and colleagues on bulk platinum-ruthenium alloys, where the initial impression was that more ruthenium—more oxophilicity—helped. But when you look at the corrected surface compositions, the numbers tell a different story: the so-called platinum zero point five ruthenium zero point five crystal had about eighty-five atomic percent platinum at the surface, and the platinum zero point one ruthenium zero point nine had about ninety percent platinum. The most active surface was actually the least ruthenium-rich. In other words, the least oxophilic of the bunch. Let's pause here. What we have is a single, pH tunable knob—the hydrogen binding energy—that seems to unify behavior across metals and electrolytes. The microscopic steps, the Volmer and Heyrovsky moves that put hydrogen onto or take it off the surface, keep their character as you swing pH. And when you measure carefully, in a regime where mass transport can't fake a kinetic effect, the numbers line up with that picture. A quick word about how those numbers are pulled out, because it matters. The exchange current comes from fitting the current-potential curve to the Butler-Volmer relationship, which is just the mathematical way of saying that the forward and backward rates of electron transfer grow exponentially with the driving force. Near zero overpotential—on the order of ten millivolts versus the reversible hydrogen electrode—you can linearize that curve and read off the slope. That slope is proportional to the exchange current density. The team did those fits at three hundred thirteen kelvin, and they cross-checked the fuel-cell data against the rotating disk in alkaline solution after correcting for any ohmic losses. The impedance spectra close the loop by showing a clean charge transfer signature, not a diffusion-limited arc. And behind the scenes, they made sure they were comparing like with like—nanoparticle catalysts with well-characterized surface areas and stable voltammetric fingerprints—so they weren't conflating morphology with mechanism. If you're wondering whether any of this reopens the door to hydroxide helping out under special conditions, the authors are pretty clear about what their data do and don't say. They do not see evidence that you need to invoke an extra, oxide-associated hydrogen species—H-OPD—to explain hydrogen rates in base. H-UPD, the hydrogen you count in voltammetry, maps directly onto the Volmer step, which sets the pace in alkaline media. They do see that the interfacial environment, including how water molecules arrange themselves and how the local field shifts with pH, can change the apparent barrier. That's not a different mechanism. That's the same mechanism running over a slightly reshaped energy surface. And they add one more cautionary flag for anyone measuring in unbuffered solutions: local pH gradients are sneaky. You can measure something that looks like a kinetic pH dependence and be staring at a near-electrode pH that's drifted away from the bulk. Auinger's work is the exemplar here, and Durst's team designs around it by leaning on a mass-transport-free hydrogen pump as the primary kinetic lens and using the rotating disk only as a corroborating view in alkaline electrolyte. So what's the payoff? If hydrogen binding is the lever, and the Volmer step is the bottleneck in base, then catalyst design should focus on tuning that binding—not on making the surface more oxophilic in the hope that hydroxide will pull you through. That's a different mindset for alkaline hydrogen electrodes. It also reframes how we model. If your density functional theory calculation in vacuum says platinum binds hydrogen just so, and then your experiment in zero point one molar sodium hydroxide says the rate plummets, you shouldn't rush to add a whole new reaction path. You should ask how pH perturbs the binding energy you feed into your rate law. The team even gives you a quantitative target: capturing a binding-energy shift on the order of twelve point five to thirteen point five kilojoules per mole from pH zero to thirteen reproduces a roughly two hundred-fold swing in rate. There's room to push this further. We still need better experimental handles on hydrogen binding energy under reaction conditions, not just in a clean voltammogram, and better computational models that put pH and interfacial water on the same footing as the bare metal surface. But those are next steps, not rescue lines for a failing idea. The central claim holds: in both acid and base, on platinum, iridium, and palladium, hydrogen oxidation and evolution run through the same microscopic steps. The huge pH dependence doesn't force a new mechanism. It reflects how the interface tilts the same landscape. If you take nothing else away, take this. In their two thousand fourteen Energy and Environmental Science paper, Durst and colleagues give us a unifying picture: platinum beats iridium, which beats palladium; swing from pH zero to pH thirteen and the exchange current falls by about two orders of magnitude; the Volmer step governs in base; and the one descriptor that stitches it all together is the hydrogen-binding energy. It's a clean story. And for once in catalysis, the clean story survives the numbers.