Ask three hydrogen engineers which electrolyser technology will win and you will get three answers, each defensible on efficiency, footprint or cost per kilowatt. At gigawatt scale, though, a different question may decide things: can we get the materials, and can we keep using them? Here is how the three low-temperature technologies compare when you start from materials.
Alkaline: proven and abundant, but big and slow
Alkaline water electrolysis has been in industrial use for about a century. It runs on inexpensive, abundant electrode materials in concentrated potassium hydroxide, with a porous diaphragm separating the gases. Its strengths are maturity, long stack lifetimes and low material cost, with no precious metals required.
The trade-offs are physical. Current densities are typically well below 1 A/cm², so an alkaline plant needs a lot of active area, equipment and floor space per megawatt. The porous diaphragm lets gases mix more easily at low load, which limits turndown and makes it harder to follow fast swings in wind and solar output. Concentrated lye (around 30 wt %) also brings handling and corrosion demands for the balance of plant.
PEM: compact and dynamic, but tied to iridium and PFAS
Proton exchange membrane electrolysis solved alkaline's size and dynamics problems. A thin solid membrane allows 2 A/cm² and more, fast response and differential-pressure operation in a compact stack, which makes PEM a natural partner for variable renewables.
The cost is in the materials. The acidic environment demands iridium at the oxygen electrode, conventionally around 2 mg/cm². Iridium is one of the scarcest elements mined, recovered only as a by-product of other precious-metal mining, with world output in the single-digit tonnes per year. At conventional loadings, each gigawatt of PEM capacity needs on the order of half a tonne of it.
The other exposure is PFAS. Standard PEM membranes and ionomers are fluorinated polymers, and the EU is weighing a broad PFAS restriction. Even with exemptions, that creates regulatory and supply uncertainty over a stack's lifetime.
The most direct fix for iridium is to use far less of it. Our ultra-low iridium OXYGN-M anodes run at 0.2 mg Ir/cm², a tenth of the conventional loading, and have shown 1.86 V at 2 A/cm² for more than 1,000 hours without degradation (details here).
AEM: the attempt to have both
Anion exchange membrane electrolysis uses a solid polymer membrane like PEM, but the membrane conducts hydroxide ions, so the chemistry is alkaline. That combination is the whole appeal:
- No iridium or other PGMs needed. Alkaline conditions allow abundant catalysts. Our OXYGN™ and H2GN™ electrodes are PGM-free and free of critical raw materials.
- PFAS-free by design. AEM membranes and ionomers can be made from fluorine-free polymers, as our AionFLX™ materials are.
- Compact and dynamic like PEM. A dense membrane gives lower gas crossover than a diaphragm and higher current density. Public HYScale data show 2 A/cm² at about 2 V in 0.1 M KOH in cells with no PGMs.
- Dilute electrolyte. AEM systems run on pure water to dilute KOH, far gentler than alkaline lye.
Progress is fast: standard AionFLX™ has logged about 5,000 hours at around 20 µV/h in 0.1 M KOH, and the composite AionFLX™ Plus has passed 8,000 hours in 1 M KOH (results here). At system level, HYScale's 100 kW AEM system is built and in commissioning (the story so far).
Side by side
| Alkaline | PEM | AEM | |
|---|---|---|---|
| Maturity | Very high | High | Emerging (100 kW systems) |
| Typical current density | < 1 A/cm² | 2+ A/cm² | 1–2 A/cm² demonstrated |
| Precious metals | None | Iridium (anode), PGM (cathode) | None needed |
| Fluorinated polymers | Usually none | Yes (membrane, ionomer) | None needed |
| Electrolyte | ~30 % KOH | Pure water | Pure water to dilute KOH |
| Dynamic operation | Limited | Excellent | Good |
So which will lead by 2030?
Probably all three, in different niches. Alkaline will keep winning large, steady, cost-driven projects. PEM will stay strong where dynamics and footprint matter, provided iridium loadings fall fast enough, which is exactly what ultra-low iridium catalysts are for. AEM's case rests on removing the material constraints of the other two. If durability keeps improving at the current pace, it could be the one that scales most freely.
At CENmat, we work on both membrane routes, AEM and PEM, from materials to stacks and systems. Our iridium calculator estimates the metal needs of your own project, and our FAQ answers the most common questions about AEM and PEM materials.