The proton exchange membrane is an American invention with a strange history. It was invented twice, by two different American companies, for two different reasons in the early 1960s, and then almost nothing happened to it for thirty years.

The first version came out of General Electric in 1962. Two chemists named Leonard Niedrach and Thomas Grubb built the first practical PEM fuel cell using a polystyrene sulfonic acid membrane. It worked. NASA flew it on the Gemini program as the power source for the spacecraft. The membrane was the heart of the cell — it was what made the device work at all — but it was viewed as a means to an end rather than as an interesting object in its own right. NASA wanted power for its capsules. The membrane was the way to get the power.

The second version came out of DuPont in 1966. A chemist named Walther Grot, working on chlor-alkali separators, synthesized a perfluorosulfonic acid polymer that conducted protons better than anything that had come before. He named it Nafion. The original target market was the chlor-alkali industry, which uses electrolysis to make chlorine and sodium hydroxide from salt water .Nafion was a better separator than what the industry had been using, and it became the standard. Nafion is still the standard material today for chlor-alkali plants.

These two inventions — the PEM fuel cell and Nafion — converged in the 1980s when researchers realized that Nafion could be used as the membrane in the PEM fuel cell architecture. The combination became the foundation of the modern PEM industry. Nafion is the membrane in almost every PEM fuel cell, electrolyzer, and flow battery in commercial use today.

The thing I find interesting about this history is what didn't happen next. Once Nafion was established, the chemistry essentially stopped moving. There were optimizations — DuPont and 3M independently developed shorter-side-chain variants in the early 1990s that handled higher temperatures slightly better, and W.L. Gore added a porous mesh reinforcement in 1994 that made the membrane more mechanically robust — but the basic molecular architecture has not been redesigned in fifty years. The polymer in a commercial 2026 electrolyzer is, structurally, the polymer Walther Grot made in 1966 with minor process improvements.

Why? The answer, I think, has to do with the structure of incentives around a successful substrate. Once Nafion took the market, the customers —chlor-alkali, then fuel cells, then electrolyzers — built their entire engineering practice around its properties. Catalyst formulations were tuned to Nafion's ion exchange capacity. Stack architectures were designed around its water management quirks. Manufacturing equipment was bought to handle its mechanical behavior. The downstream industry got very good at working with Nafion specifically, which meant that any new membrane chemistry had to overcome a substantial switching cost just to be considered.

At the same time, DuPont's incentive to dramatically improve the substrate was muted. They had the dominant position. Incremental improvements were enough to keep that position. A radically better membrane would have been expensive to develop, would have cannibalized their existing product, and would have had to fight uphill against an installed base of equipment and engineering practice optimized for the old chemistry.

So the substrate sat. For thirty years. Through the dot-com boom, the rise ofChina, the entire global energy debate, three U.S. administrations, the Kyoto Protocol, the Paris Agreement, the IRA, and an enormous amount of academic research on every part of the electrochemistry stack except the part that gated all of it. Catalysts got better. Bipolar plates got better. System integration got better. The membrane stayed roughly where it had been since the Carter administration.

This is, I think, the standard fate of platform technologies in industries that aren't competitive enough at the substrate layer. Silicon has a similar story. Nobody was competing on the membrane. So the membrane didn't move.

What's changed in the last five or six years is that several downstream industries have started simultaneously hitting the wall that the 1966 substrate creates. PEM electrolyzers can't easily run at higher current densities because the membrane is thick and resistive. Fuel cell durability targets can't be met because the membrane degrades. Flow battery economics don't work because the membrane is expensive and ions cross over. CO₂ reduction reactors don't work at industrial scale because the membrane isn't selective enough for products over reactants.

A field full of researchers and companies, each independently trying to do something different, has all run into the same physical limit at roughly the same time. When that happens, somebody usually does the work to push the substrate forward.

That's what we're doing. The membranes we make aren't a tweak to Nafion. It's a different architecture, designed for what the industries above it actually need to do over the next fifty years rather than what the chlor-alkali industry needed in 1966. The history is interesting partly because it explains why this problem went unsolved for so long, and partly because it explains why the gap to the next generation is so large now that someone is finally trying to close it.

‍