What an AEM Electrolyser Asks of Its Membrane

Green hydrogen has a cost problem, and a good deal of that cost is metal. Proton exchange membrane electrolysers work well, respond quickly to a variable renewable supply, and run on iridium and platinum, which are among the least abundant elements anyone has built an industrial plan around. Traditional alkaline electrolysers avoid the precious metals entirely and give up responsiveness, footprint and current density in exchange.

Anion exchange membrane electrolysis is the attempt to have both: the nickel-based catalysts of alkaline chemistry inside the compact, fast-responding architecture of a membrane cell. The concept is sound and has been for years. What has held it back is not the catalysts and not the cell design. It is the membrane.

What the membrane is being asked to do

In an AEM cell, hydroxide ions have to travel from the cathode, where water is split and hydrogen evolves, through the membrane to the anode, where oxygen is released. The membrane sits between the two, and it has to do four things at once.

It has to conduct hydroxide quickly, which means a high density of fixed positive charge along the polymer backbone. It has to keep the two gas streams apart, because hydrogen crossing into oxygen is both a safety problem and an efficiency loss. It has to hold its mechanical integrity while swollen with alkaline water and while the pressure across it changes. And it has to do all of this for tens of thousands of hours in hot caustic.

Those requirements pull against each other, which is the whole difficulty. Raising the charge density raises conductivity and also raises water uptake. A more swollen membrane conducts better, tears more easily and passes more gas. Cross-linking the polymer to control the swelling costs conductivity and usually costs toughness too. Every AEM development programme is a walk along that trade, not a search for a single missing ingredient.

Alkaline stability is the binding constraint

The cationic groups that make the membrane conduct are the same groups the environment attacks.

Quaternary ammonium is the usual choice, and hydroxide degrades it by two well-understood routes: direct substitution at the carbon adjacent to the nitrogen, and elimination where a hydrogen sits in the wrong position on the neighbouring carbon. Both remove the fixed charge. Conductivity falls, and it does not come back.

This is why the honest question to ask of any AEM material is not what conductivity it achieved. It is what conductivity it retained, at what temperature, in what concentration of hydroxide, after how many hours. A membrane measured fresh, at room temperature, tells you almost nothing about a stack that has to run at sixty degrees Celsius for five years.

The structural responses are known. Remove the hydrogens that permit the elimination route. Put steric bulk around the nitrogen so the substitution route is hindered. Move the cation off the backbone on a flexible spacer so that when it is eventually lost, the backbone survives. Consider cations other than ammonium, at the cost of a harder synthesis. None of these is free, and each shifts the conductivity trade.

The properties worth measuring, and in what order

Characterisation of a membrane can absorb unlimited effort. The measurements that change a decision are these.

• Ion exchange capacity fixes the ceiling on conductivity and predicts the swelling. Measure it first, because it explains most of what follows.
• Water uptake and swelling ratio, measured hot rather than at ambient, tell you whether the mechanical properties you measured dry mean anything in service.
• Through-plane conductivity in hydroxide form, at the intended operating temperature. In-plane numbers are easier to obtain and are not the number the cell experiences.
• Retained conductivity after alkaline soak, at temperature, over hundreds of hours. This is the single most informative measurement in the whole set and the one most often left until last.
• Hydrogen permeability, which sets the lower bound on how thin the membrane can go before the safety case fails.
• Mechanical properties in the wet, swollen state, because that is the only state in which the membrane operates.

Running the alkaline soak early reorders a development programme, usually by eliminating candidates that looked excellent on day one.

Where the honest uncertainty sits

Two problems remain genuinely open, and it is worth being direct about them rather than implying the field is further along than it is.

The first is the ionomer in the catalyst layer, as distinct from the membrane sheet. It has the same stability problem and a harder job, since it must also bind the catalyst and let gas escape. Cell performance frequently degrades there first while the membrane itself is still sound.

The second is that carbon dioxide from air, dissolving into the electrolyte, converts hydroxide to carbonate and bicarbonate. These are larger, slower ions, and cell voltage rises accordingly. It is reversible, but it means an AEM system's feed water and gas management are part of the design rather than a utility detail.

Where this sits for us

Membranes for gas separation, high-performance polymer formulation, and nanocomposite materials are areas we work in, and AEM electrolyser development is a line we are actively building rather than one with a long delivery record behind it. We would rather describe it that way.

If you are developing an electrolyser and the stack is losing performance without an obvious cause, the useful experiment is usually the boring one: put the membrane and the ionomer through a hot alkaline soak and measure what is left. It is a fortnight of work and it tends to settle the argument.