Materials Science · Climate Tech
Thin air went in.
Graphite came out. Scientists finally watched it happen.
For decades, chemists knew molten salt could turn atmospheric CO2 into solid graphite — but nobody had actually seen the reaction happen at the molecular level. A Berkeley-led team just did, using a technique that let them watch bonds rearrange in real time. What they found could turn today's expensive trial-and-error carbon production into something engineers can actually design on purpose.
Inside a 500°C bath of molten salt, carbon dioxide from the air is stripped apart and rebuilt into solid carbon.
Scientists Just Watched Air Turn Into Graphite
Imagine pulling a raw material for batteries directly out of the air around you.
Not mined. Not shipped from a distant refinery. Just pulled from the atmosphere itself.
It sounds like a stretch. It isn't.
Chemists have known for years that molten salt electrolysis can do exactly this, converting carbon dioxide into solid carbon.
What nobody had actually done was watch it happen. Not at the molecular level, in real time.
A team led by chemical physicist Sander Ratso at the University of California, Berkeley, just changed that.
Published in Nature Communications, their work captured the entire transformation as it unfolded, bond by bond.
"Very simply put, when before we were cooking with a closed oven, we put a window in it," Ratso told ScienceAlert.
That's a deceptively simple way to describe something researchers had been trying to see clearly for decades.
And now that the window is finally open, the view is more interesting than anyone expected.
What Molten Salt Electrolysis Actually Does
Start with the basics. Molten salt electrolysis means exactly what it sounds like.
Salts are heated until they liquefy, then poked with electrodes carrying an electric current.
We're not talking about a gentle simmer, either. The reaction runs at roughly 500 degrees Celsius — nearly 932 degrees Fahrenheit.
Inside that scalding liquid, something remarkable happens to carbon dioxide.
The electric current rearranges the bonds holding carbon and oxygen atoms together.
Piece by piece, the gas gets pulled apart and rebuilt — locking the carbon into a solid state.
Sometimes that solid comes out as graphite. Sometimes it takes other carbon forms entirely.
Either way, a gas that started out floating freely in the air ends up as something you could hold in your hand.
This isn't a brand-new idea. A few companies already run versions of this process at industrial scale, producing kilograms to tons of carbon material.
Ratso himself co-founded one of them.
What was missing wasn't the process. It was understanding exactly what happens inside it while it's running.
A Mystery That Sat Unsolved for Decades
Knowing that a reaction works is one thing. Knowing exactly how it works is another.
For this particular reaction, that second part stayed stubbornly out of reach.
Scientists have wondered for decades what's actually going on at the molecular level, deep inside that 500-degree bath.
The problem wasn't a lack of curiosity. It was a lack of visibility.
Molten salt at 500 degrees Celsius isn't exactly a friendly environment for delicate lab instruments.
Most methods of studying the resulting carbon had to happen afterward — pulling material out, letting it cool, then analyzing it separately.
That's a bit like trying to understand how bread rises by only ever examining the finished loaf.
You can guess at the process. You just can't actually confirm it.
One idea, in particular, had been floating around since 1999: that a peroxide compound might form as a fleeting middle step.
For over two decades, that stayed a hypothesis. Nobody had the tools to catch it happening.
Putting a Window in the Oven
The breakthrough came down to a technique called operando .
In plain terms, it lets researchers shine light on a reaction while it's actively happening, and read the vibrations bouncing back.
Confocal Raman microscopy lets researchers read the vibrational "fingerprints" of molecules as the reaction unfolded, live, inside the molten salt.
Every molecule vibrates in its own distinctive way. Those vibrations act like a fingerprint, revealing exactly what's present.
Point that technique directly at a live reaction, and suddenly you're not guessing from the finished product anymore.
You're watching intermediate molecules appear, evolve, and disappear, in real time, exactly as the reaction happens.
That's the "window" Ratso was describing — a genuine, live view into a process that used to be a sealed black box.
"Understanding how this reaction proceeds and how the structure of the carbon evolves over time is the first step in controlling" it, Ratso explained.
Mike Whittaker, an Earth geoscientist at Lawrence Berkeley National Laboratory, called it a major win for the broader effort to synthesize critical battery materials from molten salts.
And the detail this method provided wasn't a small improvement. It was unprecedented — enough to reveal a full two-step reaction pathway that had never been directly confirmed before.
The Hidden Middle Step Nobody Had Confirmed
Remember that 1999 hypothesis about a peroxide intermediate? This is where it finally gets its proof.
As the researchers watched carbon gradually deposit during electrolysis, something specific showed up in the Raman signal.
A distinct band of light, matching carbon-adsorbed peroxide, appeared reliably, every single time.
And crucially, it didn't show up randomly. It grew stronger in exact step with carbon deposition itself.
That timing is the smoking gun. It's direct evidence that peroxide is the mechanistic middle step driving CO2 into solid carbon.
Not a side reaction. Not a coincidence. The actual bridge between gas and graphite.
A hypothesis that sat unconfirmed for more than 25 years finally had direct, real-time evidence behind it.
That might sound like a small technical detail. It isn't.
Once you know the actual pathway a reaction takes, you can start asking how to speed it up, slow it down, or steer it toward a specific outcome.
Before this, engineers were essentially optimizing a process without knowing its actual internal logic.
Same Reaction, Four Very Different Carbons
Here's where the study gets genuinely surprising. The exact same reaction didn't always produce the same carbon.
Swap the cathode, and the output completely changes shape.
On a pure nickel cathode, the team found carbon forming as nanotubes, alongside flat, platelet-like structures.
Switch to an electrode made of Inconel 600, a nickel-based alloy, and the output shifted almost entirely to platelets.
Try gold instead, and the carbon came out looking nothing like either of those — deposited in loose, disordered clumps researchers simply called "amorphous."
Tungsten produced its own amorphous carbon too, but with a genuinely striking bonus: tiny structures nicknamed "carbon nano-onions."
Four electrodes. Four distinctly different carbon products. One underlying reaction.
That's not a minor footnote. It's arguably the most practically useful discovery in the entire study.
It means the electrode material itself functions like a dial, adjustable to steer what kind of carbon comes out the other end.
Batteries, smartphones, and industrial equipment don't all need the same kind of carbon. Now there's a mechanism for choosing.
Why the Electrode Changes Everything
It's worth pausing on why this cathode effect matters so much for anyone trying to actually build something with this technology.
Not all carbon is created equal. Nanotubes behave very differently from platelets, and both behave differently from amorphous clumps.
Battery-grade graphite has specific structural requirements. So does the carbon used in certain industrial coatings or specialty electronics.
Before this study, getting a specific carbon structure meant trial and error — synthesizing a batch, testing it, tweaking a variable, and trying again.
That loop is slow. It's also expensive, and it limits how many carbon variants anyone can realistically explore.
Ratso is candid about how this has held the whole field back.
Existing commercial approaches, he explained, still rely on synthesis conditions "inferred from ex situ trial-and-error measurements."
Now, with real-time visibility into exactly how the electrode shapes the outcome, that guesswork has an actual mechanism behind it.
Instead of hoping a given electrode produces the right structure, engineers can start reasoning from cause to effect.
That's the shift from craft to engineering — and it rarely happens without exactly this kind of direct observation first.
From Trial-and-Error to Designed Carbon
Put the peroxide discovery and the electrode discovery together, and something bigger comes into view.
This isn't just a curiosity about how one reaction happens to work. It's a toolkit for controlling it.
Know the mechanism, and you can ask sharper questions. Adjust the electrode, and you can steer the outcome.
Together, that starts to look less like discovery, and more like design.
Ratso frames the potential plainly: these insights "could make the process much more efficient."
Efficient in the literal sense — fewer wasted batches, faster iteration, less expensive experimentation.
But also efficient in a more ambitious sense: purpose-built carbon, made to order, rather than discovered by accident.
Whittaker's framing points toward exactly that future — one where the process runs "at low temperatures with really cheap salts," widely deployable, feeding real battery supply chains.
That's still a leap from today's lab conditions. But it's a leap researchers can now aim toward deliberately, instead of stumbling toward blindly.
A reaction understood is a reaction that can finally be engineered.
A Battery Material That Could Erase Its Own Footprint
There's a climate story tucked inside this chemistry story, and it's arguably the bigger headline.
Producing the carbon used in today's batteries and is not a clean process.
Battery-grade graphite normally carries a heavy carbon footprint of its own — sourcing it from CO2 instead could flip that equation.
Conventional methods carry enormous CO2-equivalent costs, measured in kilograms of emissions per kilogram of material produced.
Sourcing that same graphite from atmospheric CO2 instead flips that entire equation on its head.
Instead of adding emissions to make a battery material, you'd be removing carbon dioxide that's already floating in the air.
That's not a small distinction. It's the difference between a product that contributes to warming and one that actively works against it.
Ratso is careful not to oversell this as a climate cure-all on its own. It isn't, and researchers broadly agree alone won't solve the climate crisis.
But paired with clean electricity, the math genuinely improves.
"The molten salt electrolysis method can make these materials with negative CO2 equivalents when coupled with wind, solar, or nuclear power," Ratso said.
He goes further still, calling it "among the most important technologies that could have a real impact on the levels of CO2 in the environment."
That's a bold claim from someone who's spent years inside this exact reaction. It's also one this new molecular-level understanding makes easier to actually test.
What Still Has to Happen
None of this makes molten salt electrolysis a finished technology. It makes it a better-understood one.
The next step is translating this new mechanistic clarity into actual process improvements, not just lab-bench insight.
Can the reaction run efficiently at lower temperatures? Can cheaper, more widely available salts replace the current ones?
Those are engineering questions this study didn't answer, but did make far more approachable.
There's also the matter of scale. A handful of companies already run this commercially, producing kilograms to tons of material.
Feeding a global battery supply chain would mean scaling that up by orders of magnitude.
And crucially, that scaling needs to run on genuinely clean electricity to deliver the negative-emissions outcome Ratso describes.
Run the same reaction on a dirty power grid, and much of that climate benefit quietly disappears.
None of these hurdles are small. But they're now hurdles researchers can approach with an actual map of the reaction in hand.
That's a considerably better starting point than guessing in the dark, which is exactly where this field stood a year ago.
Manohar Insight
What stays with me here is Ratso's own phrase: putting a window in a closed oven.
It's such a simple image for something that took over two decades to actually achieve.
And it's a reminder that plenty of "unsolved" problems in science aren't unsolvable. They're just unobservable, with the tools available at the time.
A hypothesis from 1999 waited 27 years for someone to finally build the instrument capable of checking it.
That's a long wait. But it's also exactly how patient, cumulative science tends to work.
What excites me more than the chemistry, honestly, is the framing. A material we usually think of as pollution, reimagined as a raw ingredient.
Not stored away and forgotten, like most carbon capture proposals. Actually rebuilt into something batteries, phones, and industry need.
The real breakthrough isn't that we can pull carbon from the air. It's that we can finally watch it happen closely enough to control it.
Join the Signal
If graphite pulled straight from the air could realistically feed battery supply chains at scale, would that change how you think about the materials inside your next phone or EV — or does the climate math still feel too good to be true? Share your take in the comments.
Discussion