An abandoned metal may help to make cleaner fuel for airplanes

Airplanes and cargo ships are among the most difficult sectors to decarbonise because batteries are often too heavy. Researchers are therefore searching for ways to make energy-dense liquid fuels from CO₂ and renewable electricity instead of oil.

For aviation fuel, that means building long hydrocarbon chains that can store enough energy for long-distance flight. But most electrochemical systems can only stitch together very short carbon chains before the reactions stall.

“We know that we are stuck at C3,” says Brian Seger, Professor of Physics at the Technical University of Denmark. “We can never get a fourth carbon on.”

So Seger and colleagues turned to a metal that most electrochemists had largely abandoned: cobalt.

Now, Seger and international collaborators show that cobalt at elevated temperatures can produce molecules extending into the C6+ range – still not aviation fuel, but a step closer to the longer chains aviation fuel requires.

“It was definitely a gamble,” Seger recalls. “If you are not gambling in research, well, that is kind of the point of what we are doing as researchers.”

Why everybody kept using copper

For more than a decade, electrochemical CO₂ conversion has largely revolved around copper because it can build multi-carbon molecules.

“That was really why we turned to Fischer-Tropsch chemistry,” Seger says. “Those catalysts already know how to build long carbon chains.”

In Fischer-Tropsch chemistry, catalysts such as cobalt and iron convert carbon monoxide (CO) and hydrogen into liquid hydrocarbons at high temperature and pressure. Yet despite powering industrial fuel production for decades, the chemistry still holds a central mystery: how the carbon chains actually grow.

“What is fascinating is that Fischer-Tropsch chemistry is more than 100 years old, and we still argue about the mechanism,” Brian Seger explains. “Electrochemistry may give us an entirely new way to study these reactions.”

Cobalt, meanwhile, had long been considered a poor candidate for electrochemical CO₂ reduction because it tends to produce hydrogen gas instead of carbon-based molecules.

“People assumed that cobalt would just make hydrogen,” Seger says. “That is why very few people really explored it.”

Hints had started appearing that cobalt and nickel could sometimes produce longer hydrocarbons after all.

“It is really hard to make anybody do anything creative at all,” Seger adds. “If everybody keeps doing variations of the same copper chemistry, you mostly end up making marginal gains.”

That raised the possibility that an entirely different catalytic regime existed outside normal copper chemistry.

Could heat change the chemistry?

Another largely neglected variable was temperature.

“There is surprisingly little work at intermediate temperatures in electrochemistry,” Seger says. “Most people either stay near room temperature or jump all the way to solid-oxide systems at 600 to 800 degrees.”

Unlike thermal catalysis, electrochemistry is usually studied close to room temperature.

“Copper does not hold onto carbon strongly enough to build very long chains,” Seger explains. “But cobalt may actually hold onto the products too strongly.”

Because cobalt holds onto CO much more strongly than copper does, higher temperatures might help the surface build longer chains without trapping the products completely.

“If that is true, then this may not only be about cobalt,” Seger notes. “There may be many other catalytic reactions in which adding a little electrochemistry completely changes the mechanism.”

“So many catalytic reactions are just thermal reactions,” Seger says. “Why not just put a little potential on these same systems and see what happens?”

Moving beyond room-temperature electrochemistry

To test this idea, Seger and colleagues moved away from the conditions normally used in CO₂ electrolysis.

“Everybody just keeps doing copper at room temperature,” Brian Seger notes. “We wanted to move into a completely different parameter space.”

Instead of feeding CO₂ directly into the reactor, they started with CO – an intermediate already formed during CO₂ electrocatalysis. That let them isolate the chain-growth chemistry while avoiding carbonate formation in the electrolyte.

“We wanted to isolate the chemistry itself,” Seger says. “Using CO gives a cleaner window into those mechanisms.”

“If you are dealing with CO₂ in water, you start forming carbonates and bicarbonates everywhere,” Seger explains. “The system basically starts clogging itself chemically, and the reactor becomes much harder to operate.”

Pushing the reactor towards Fischer-Tropsch conditions

The experiments used a thin cobalt layer deposited onto a gas-diffusion electrode inside a membrane electrode assembly reactor operating at commercially relevant current densities.

The reactor was heated to 30–80°C – unusually high for electrochemical CO₂ conversion.

“The membranes used in most electrochemical systems simply cannot tolerate high temperatures,” Seger says.

The system was pushed as far as the membrane stability limits allowed.

“We wanted to move toward Fischer-Tropsch conditions,” Seger adds. “80°C is actually not that high – but already there we started to see chemistry that basically was not happening at room temperature.”

The reactor started making unexpected molecules

The reactor also operated at around 250 milliamps per square centimetre – conditions more relevant for practical fuel synthesis than idealised proof-of-concept experiments.

As the experiments progressed, the product mixture became unexpectedly complex.

The analysis detected 28 compounds, including hydrocarbons, alcohols, ketones and carboxylic acids.

“Normally in CO₂ electrolysis you focus on a relatively small number of products,” Seger says. “Here we suddenly had a whole spectrum of molecules appearing.”

One major challenge was proving that the hydrocarbons genuinely originated from electrochemical reactions on cobalt rather than contamination or ordinary thermal catalysis inside the reactor.

“These are unusual products for electrochemistry,” Seger explains. “So we had to rule out every obvious alternative explanation.”

Ruling out contamination and ordinary catalysis

To rule out contamination or ordinary thermal Fischer-Tropsch chemistry, the feed gases were analysed for impurities, CO was replaced with argon gas, blank electrodes without cobalt were tested and heat alone was examined as a possible explanation.

The controls consistently pointed to the same conclusion: the hydrocarbons were being generated electrochemically on the cobalt surface.

One important question concerned the electrolyte surrounding the catalyst. In copper systems, simply changing the positively charged ions in the liquid can dramatically alter which molecules form.

But cobalt behaved very differently.

“With copper, the choice of cation is absolutely critical,” Seger says. “Here it basically did not matter.”

To understand why, the experiments were combined with theoretical calculations and machine-learning-assisted simulations.

Initially, cobalt was expected to behave more like classical Fischer-Tropsch catalysis.

When the calculations stopped making sense

But the calculations pointed somewhere else entirely.

“The theoreticians basically told us: ‘this mechanism should not work under our conditions,’” Seger notes. “And the copper mechanism also did not explain the data.”

“At that point we really had no idea what was going on,” he says. “Without the theoreticians, we would probably still have no idea.”

Instead, the calculations suggested that water molecules may help to remove oxygen step by step during chain growth rather than CO first having to split apart completely on the catalyst.

“That was really the breakthrough moment,” Seger explains. “We suddenly realised we were not looking at pure copper electrochemistry and not looking at classical Fischer-Tropsch chemistry either.”

“So now this is fun.”

The simulations also revealed that unsaturated hydrocarbons bind extremely strongly to cobalt surfaces.

“Our modelling suggests that the rate-limiting step may actually just be getting the products off the surface,” Seger says. “The catalyst may simply hold onto them too strongly.”

A product distribution electrochemists rarely see

The first products already included hydrocarbons extending into the C6+ range.

The reactor generated 28 products, including hydrocarbons, alcohols, alkenes, ketones and carboxylic acids – including butane, pentane and hexane-range compounds rarely seen under comparable electrochemical conditions.

“Usually in electrochemistry you try to simplify the product distribution,” Seger notes. “Here the complexity was actually telling us something important.”

“We were suddenly looking at chemistry that resembled Fischer-Tropsch far more than classical CO₂ electrolysis,” he says.

As the reactor temperature increased from 30°C to 80°C, hydrocarbon production rose almost linearly.

“On copper, higher temperatures usually have only a modest effect on what products are formed,” Seger adds. “Here we observed something much more dramatic.”

Cobalt was not behaving like copper

An analysis showed an Anderson-Schulz-Flory product distribution – common in Fischer-Tropsch catalysis – thus revealing a characteristic stepwise chain-growth pattern.

“The product distribution itself contains mechanistic information,” Seger says.

The electrolyte offered another surprise. In copper systems, changing the cation strongly alters hydrocarbon formation.

“With copper, the cation effect is enormous,” Seger explains. “Cesium works really well. Lithium is horrible.”

But for cobalt, the effect was unexpectedly small.

“We could use basically any of them,” Seger adds.

The simulations suggested that direct splitting of CO on cobalt remained energetically inaccessible.

The catalyst became too good at holding onto products

Instead, the bottleneck seemed to emerge later in the process.

Long-chain unsaturated hydrocarbons bound extremely strongly to cobalt surfaces.

“Copper does not hold onto carbon strongly enough to build very long chains,” Seger explains. “But cobalt may actually hold onto the products too strongly.”

The products then began coating the catalyst surface itself, gradually blocking the sites needed to keep the reaction running.

“The catalyst almost becomes too good at holding onto the products,” Seger notes.

The experiments supported this interpretation directly: hydrocarbon production declined steadily during operation, with roughly 80% loss in selectivity after three hours.

But short oxidative pulses during electrolysis temporarily restored catalytic activity.

“It was almost like cleaning the catalyst surface while the reaction was running,” Seger says.

The bottleneck may be letting go

The pulses likely stripped away carbon-rich deposits and strongly bound hydrocarbons, temporarily exposing fresh catalytic sites.

The Arrhenius analysis added another clue.

The apparent activation energy extracted from the temperature-dependent experiments was only around 0.26 electronvolts – far lower than values typically associated with classical thermal Fischer-Tropsch chemistry.

This suggested that the slowest step was probably not building the carbon chains themselves but getting the finished molecules back off the catalyst surface.

“For years we focused on how to make longer hydrocarbons,” Seger recalls. “We may discover that the real challenge is preventing the catalyst from loving them too much.”

“The mechanism does not look like normal copper electrochemistry – but it does not fully behave like classical Fischer-Tropsch chemistry either,” he says. “We seem to be sitting somewhere in between those two worlds.”

The chemistry still falls far short of aviation fuel

The study does not yet offer a practical route to achieving sustainable aviation fuel.

Hydrogen production still dominates the reaction, hydrocarbon efficiency remains relatively low and the catalyst gradually deactivates as products accumulate on the surface.

But Seger believes that the importance of the work lies less in current performance than in the chemistry itself.

“We may have become very focused on copper because it was the first system that worked well,” Seger notes. “But once you move into a different parameter space, entirely different mechanisms may start to emerge.”

The cobalt experiments suggest that electrochemists may have overlooked entirely different kinds of reaction chemistry by focusing so heavily on copper.

“If we could push this toward C8 to C12 molecules – the range used in aviation fuel – then suddenly you have something very different from most electrochemical CO₂ systems,” Seger says.

The work also blurs the boundary between electrochemistry and thermal catalysis.

“I think that is the thing that really excites me,” Seger adds. “There are so many catalytic reactions that are just thermal reactions. Why not just put a little potential on those same systems and see what happens?”

Now the researchers want to push the temperature far higher

Seger believes that the new approach could potentially extend far beyond aviation fuel synthesis.

“There are lots of industrial catalytic reactions – methanol synthesis, hydroformylation and carboxylation,” Seger says. “They will not all work electrochemically, and that is okay. We only need one or two to really influence the world.”

The next step is pushing toward much higher temperature and pressure.

“80°C is actually not that high,” Seger explains. “We saw basically nothing at 25°C, and then at 80°C we suddenly started seeing chain growth.”

This next phase will require solving major engineering problems.

“The limitation now is not really catalysis,” Seger says. “It is the membrane. Most electrochemical membranes simply cannot survive those conditions.”

Seger and his colleagues are developing pressurised reactor systems that can operate above 100°C while still keeping water liquid inside the reactor.

“Does this 1% jump to 50%, or are we still down at 5%?” Seger asks.

A completely different electrochemical world may exist beyond copper

Another question is whether chain growth itself can be actively controlled.

“With electrochemistry, I can change the potential in a microsecond,” Seger says. “So maybe we can pulse the catalyst and actually control chain length dynamically.”

The work also points toward a broader class of strongly CO-binding metals – including nickel, iron and ruthenium – that may act very differently from copper under electrochemical conditions. Current efficiency remains far too low for industrial aviation fuel production, and Seger is careful not to overstate how close the technology is to practical deployment.

But Seger argues that discovering a fundamentally different way for the chemistry to work may ultimately matter more than immediate optimisation.

“In science, sometimes the most important thing is not optimisation,” Seger concludes. “Sometimes it is discovering that the chemistry behaves differently than we thought.”