What a catalyst does, and why the electrolyte decides which one you can use

Every electrolyzer datasheet names its catalyst materials. Some say iridium, some say nickel and iron, and almost none say why. The answer is rarely a preference. It is set by the electrolyte the cell runs in, and once the technology is chosen the catalyst options are largely chosen with it.

It is also worth being precise about what a catalyst does, because the common shorthand is wrong. A catalyst does not split water and it does not make hydrogen. The current does that. What the catalyst changes is the voltage you have to apply before the reaction proceeds at a useful rate, and voltage is electricity you pay for.

A better catalyst does not produce more hydrogen. It produces the same hydrogen at a lower voltage, which is the same thing as less electricity.

Where the voltage actually goes

Splitting water has a thermodynamic floor. At room temperature the reversible cell voltage is 1.23 V, and no catalyst moves that number, because it is a property of the reaction rather than of the electrode. Run a cell adiabatically and the relevant figure is the thermoneutral voltage, 1.48 V, at which the reaction supplies its own heat.

Real cells run above both. The gap is what the catalyst, the membrane and the cell design compete over.

That third band is where the engineering lives, and it splits into three parts: the overpotential of the oxygen evolution reaction on the anode, the overpotential of the hydrogen evolution reaction on the cathode, and the ohmic losses of the membrane, the contacts and the electrolyte.

Those three parts are not equal. Oxygen evolution is the slow half of water splitting, and it carries most of the kinetic loss in the cell. That is why most catalyst research in the field is on the oxygen side, and why an anode is the component a materials team spends its time on. Hydrogen evolution is fast enough that it is usually not what limits a cell.

The electrolyte decides the catalyst

This is the part that is worth understanding before comparing quotes, because it explains why two suppliers name completely different materials without either of them being wrong.

PEM does not use iridium because iridium is the best catalyst in the abstract. It uses iridium because acidic oxygen evolution at industrial current densities destroys almost everything else. The material is a consequence of the environment.

AEM keeps the solid polymer membrane that makes PEM attractive and moves the chemistry back into alkaline conditions, which is exactly what reopens the catalyst question. That is the single largest cost difference between the two, and it is a difference in what is available rather than in what is preferred.

Two directions, and what they have in common

The reason catalysts are an active field rather than a settled one is supply. On the acidic side the constraint is a single scarce metal, and it is not a constraint engineering can design around.

Catalyst work today moves along two tracks in response, and it is worth knowing which one a supplier is on.

• Use less of the scarce metal. Thrifting iridium down by an order of magnitude or dispersing it to near-atomic loadings is an active field with real published results. The recurring difficulty is that lowering loading and keeping durability tend to pull against each other.
• Use none of it. Non-noble oxygen catalysts work in alkaline conditions, and this is established rather than speculative. The open problem is long-term stability, because transition metal catalysts can restructure, dissolve or lose active surface over thousands of hours.

What both tracks share is where they get tested. Serious attempts to leave the noble metal market are run in alkaline cells, because that is the only environment in which non-noble oxygen catalysts survive at all. The direction of travel in catalyst research and the case for AEM are the same argument seen from two sides.

How to read a catalyst number

Three figures are usually quoted, and they trade against each other, which is why any one of them alone is close to meaningless.

• Overpotential. How much voltage above the thermodynamic floor the reaction needs. Lower is less electricity, and the condition it was measured at travels with it.
• Current density. How much current the electrode carries per unit area at a given voltage. Higher means more hydrogen from the same size of machine, and it is also the condition under which durability gets harder.
• Durability. How much of the performance survives. This is the figure most often quoted without the conditions that make it comparable.

Ask for the electrolyte, the temperature, the current density, the membrane, the hours and the voltage the run ended at. A number without them cannot be compared with anyone else's.

There is also a difference between a catalyst result and a cell result. A half-cell measurement tells you about the material. It does not tell you what happens once that material is bonded into an electrode, compressed against a porous transport layer and run against a membrane. Numbers rarely survive that transition unchanged.

Our own long runs are a useful illustration of why the cell matters more than the component. A HydroXpand grown nickel-iron anode ran 6,358 hours in a 16 cm² two-cell short stack at 0.625 A cm⁻², degrading at 14.7 µV/h per cell on a full-run linear fit. In the same campaign, changing only the membrane gave 1,234 hours. Five times the life, same anode.

What the published durability data suggests

One pattern in the literature is worth stating, with its limitation attached. Across published AEM durability reports, the runs using iridium-oxide anodes cluster below 2,000 hours, while the longest runs, 4,968 and 8,875 hours, use nickel-iron and copper-cobalt oxide anodes with no platinum-group metal at all.

The limitation: every one of those points is a different laboratory, membrane, electrolyte and temperature, so this is a distribution and not a controlled comparison. And hours alone mislead. That 8,875-hour run ended at 2.89 V at 0.6 A cm⁻², which is a cell that had largely stopped being useful.

Read carefully, the conclusion is narrow but real: in alkaline conditions, iridium is not what makes a cell last. Nobody should conclude from this that non-noble catalysts have no durability problem. They have a different one.

What we use, and what we still buy

Our anode is nickel-iron layered double hydroxide grown directly on nickel foam. There is no binder and no platinum-group metal in it, the substrate porosity is 95 to 98 percent, and the measured surface area is about 25 times that of the bare foam. It was chosen for durability and cost rather than for the first measurement, and the trade is real: on a freshly assembled cell, a well-made iridium-oxide anode will give a higher current density at the same voltage.

Our cathode is not noble-metal-free, and we would rather say so plainly than let a page about iridium imply otherwise. It is Pt/C or PtRu/C on carbon paper at 0.2 mg of platinum-group metal per cm². Hydrogen evolution is the fast half of the reaction, the loading required is small, and platinum is neither as scarce nor as geographically concentrated as iridium. Removing it entirely is a research direction rather than a claim we make about a product we ship.

Both electrodes are sold as standard 200 by 300 mm sheets with datasheets on this site, and both are the electrodes inside our own stacks. We do not keep a better version for ourselves.