The Miracle That Lasted Five Weeks

In March 1989, two chemists stood at a folding table in Utah and announced they'd solved the energy crisis with a jar of heavy water and a strip of metal. The world believed them for exactly as long as it took anyone else to check.

On the afternoon of March 23, 1989, two chemists sat down at a table in front of television cameras in Salt Lake City and told the world they had built a nuclear fusion reactor. Not a machine the size of a stadium, ringed with superconducting magnets and costing billions of dollars, like the ones physicists had been chasing for decades. Theirs fit in a laboratory jar. It ran at room temperature. And if it worked the way they said it did, it meant the end of the energy crisis — clean, virtually limitless power, extracted from ordinary water, using equipment a well-funded high school could have assembled.

It was one of the most sensational scientific claims of the twentieth century. It took the rest of the scientific world about five weeks to conclude it almost certainly wasn’t true.

Fusion, the hard way and the easy way

To understand why the announcement caused the reaction it did, it helps to know what fusion normally requires. It’s the reaction that powers the sun: light atomic nuclei — usually isotopes of hydrogen — smash together with enough force to fuse into a heavier nucleus, releasing tremendous energy in the process. Doing that on purpose, on Earth, is brutally hard. The nuclei involved are positively charged and repel each other, so forcing them to fuse means overcoming that repulsion with extreme heat and pressure: tens of millions of degrees, in stars, or in the massive, expensive machines — tokamaks, laser-ignition facilities — that fusion researchers have spent seventy years and tens of billions of dollars trying to perfect, still without a working power plant to show for it.

Martin Fleischmann and Stanley Pons were not fusion physicists. Fleischmann, a Fellow of the Royal Society based at the University of Southampton, and Pons, chair of the chemistry department at the University of Utah, were electrochemists — specialists in reactions that happen at the surface of electrodes. Their idea, which they had been quietly testing since the mid-1980s using their own money, was that a palladium electrode might be able to pack deuterium atoms (a heavy isotope of hydrogen, the “D” in “heavy water”) so densely into its crystal lattice, through ordinary electrolysis, that the deuterium nuclei would be squeezed close enough together to fuse — at room temperature, with no need for a star’s worth of heat.

For years they ran the experiment in near-secrecy: electrolysis cells with palladium cathodes sitting in heavy water, current running for weeks at a stretch, careful calorimetry to measure exactly how much heat came out compared to how much electrical energy went in. Most of the time, the numbers balanced the way ordinary chemistry predicted. But some of the time, they said, the temperature of the cell would suddenly jump — heat appearing that the electrical input couldn’t account for. To them, “excess heat” of that kind, sustained over days, could only be explained by a nuclear process. They believed they had found a way to fuse atoms on a lab bench.

A race nobody was supposed to know about

Fleischmann and Pons weren’t the only people in Utah chasing low-temperature fusion. Forty miles south, at Brigham Young University, physicist Steven Jones was investigating a genuinely established (if impractical) phenomenon called muon-catalyzed fusion, which does produce occasional fusion events at low temperatures but far too few, and far too faint, to ever generate usable power. When Jones and the Utah pair discovered they were both submitting grant proposals that touched on “cold fusion,” university officials arranged a meeting to compare notes — and the two teams realized their claims, while built on different physics, overlapped enough to be a problem for anyone hoping to claim priority.

They agreed, amicably, to submit their papers to the journal Nature on the same day, meeting at a Utah airport on March 24 to send them off together by courier.

Then the University of Utah’s administration got involved. If Pons and Fleischmann were right, the discovery wouldn’t just be a physics landmark — it could be worth an almost unimaginable amount of money in patents, and the university wanted its name attached first and unambiguously. Rather than wait for the peer-reviewed paper, or even for the planned joint submission, the university pushed Fleischmann and Pons to hold a press conference on March 23 — a day before the agreed airport meeting with Jones’s team, and before their own paper (hastily submitted instead to the Journal of Electroanalytical Chemistry on March 11) had been reviewed by anyone outside the university. Fleischmann later said in a television interview that the public announcement had been the university’s idea, and that he regretted going along with it.

So the discovery that would normally have crept into public awareness through months of quiet peer review — other labs checking the data, a paper appearing in print, cautious commentary from experts — instead arrived fully formed at a press conference, ahead of the science that was supposed to support it.

The world takes the bait

The reaction was immediate and enormous. Newspapers around the globe ran the story on their front pages. Pons and Fleischmann were treated, briefly, as the two men who had solved the world’s energy problems from a chemistry lab. The University of Utah moved fast to try to capture the moment: state lawmakers voted to give the university millions of dollars to fund a new National Cold Fusion Institute, and the university’s lawyers filed for patents.

For a few weeks, it looked like it might hold up. A handful of other laboratories — at Texas A&M and the Georgia Institute of Technology among them — announced they had seen hints of the same effect: extra heat, or faint signs of nuclear byproducts. Given how simple the experiment sounded, physicists and chemists across the world tried to replicate it themselves, sometimes within days of the announcement, using whatever palladium and heavy water they could get their hands on.

But almost as quickly as the positive reports arrived, they started falling apart. Georgia Tech’s team retracted its finding of fusion neutrons after realizing the signal had come from a miscalibrated instrument, not the experiment. Texas A&M’s excess-heat measurements ran into serious questions about the accuracy of their calorimetry. Lab after lab reported nothing unusual at all. Crucially, nobody could reliably produce the byproducts — neutrons, tritium, gamma rays — that a real fusion reaction, even a strange new kind, should have produced in quantities matching the reported heat. The heat and the physics didn’t agree with each other, and neither reliably showed up twice.

By the end of April, roughly five weeks after the press conference, the tide had turned decisively. Major newspapers, including The New York Times, were already describing cold fusion as dead. On May 1, at a tense session of the American Physical Society’s annual meeting in Baltimore, physicist after physicist stood up to dismantle the claim; of nine invited panelists, eight said the evidence didn’t hold up. Steven Koonin, a Caltech physicist, told the room the Utah results were best explained by “the incompetence and delusion of Pons and Fleischmann.” It was about as brutal a verdict as a scientific meeting delivers.

That November, a special panel convened by the U.S. Department of Energy released its own review: it found no convincing evidence that cold fusion, as described, was a real phenomenon, though it stopped short of recommending that all research into the general area be defunded.

A quieter, messier postscript

The story didn’t end there, and not everyone accepted the verdict gracefully. Two years later, in 1991, a science writer at MIT named Eugene Mallove — who had become a believer in cold fusion — resigned from his job at the university’s news office and publicly accused MIT researchers of altering their own calorimetry data in July 1989 to hide a positive result and help bury the Utah claim, a serious allegation the researchers involved denied and that never gained traction with the broader physics community. Mallove went on to found a magazine, Infinite Energy, dedicated to advocating for cold fusion research, and remained a vocal critic of what he saw as premature dismissal for the rest of his life.

For Pons and Fleischmann themselves, the fall from grace was swift and largely permanent. The University of Utah’s National Cold Fusion Institute, established with such fanfare, quietly closed in 1991 after its own scientists failed to reliably reproduce the original results. No patent was ever granted on the discovery, despite the university’s efforts. In 1992, Pons and Fleischmann left the United States for the south of France, where a subsidiary of Toyota funded a new laboratory for them to continue the research, reportedly investing more than ten million dollars over several years. It closed in 1998 without producing results that satisfied outside scientists. Pons took early retirement, settled in France, and eventually became a French citizen; he has rarely spoken publicly since. Fleischmann continued giving talks and publishing on the subject until his death in 2012.

“Cold fusion” became, for most of the scientific world, a byword for what happens when extraordinary claims bypass ordinary scrutiny — science announced by press release instead of by publication, reviewed by reporters instead of referees. In a 2004 review, the Department of Energy revisited the question and found the community still split roughly down the middle on whether something unusual was happening in these experiments, even as the panel agreed the evidence remained far short of proving nuclear fusion. Nobody, in over three decades, has produced a cold fusion cell that a skeptical outside laboratory could run and reliably get the same striking results from.

The twist nobody expected: Google went looking too

Which makes the last chapter of the story an unlikely one. Between 2015 and 2019, a team backed by Google — not a fringe believer, not a university desperate for a headline, but one of the best-funded technology companies on Earth — spent an estimated ten million dollars quietly running one of the most careful replication efforts cold fusion had ever received. They built new, highly precise calorimeters. They consulted with veteran cold-fusion researchers to make sure they weren’t missing some subtlety of the original technique. They tried, deliberately, to recreate the exact conditions Fleischmann and Pons had described, and then pushed further with related experiments of their own design.

Published in Nature in 2019, the verdict was unambiguous: no evidence of cold fusion. The team couldn’t even reliably reproduce the underlying conditions the 1989 claim depended on, let alone the fusion itself. Thirty years and a fortune in patient, rigorous funding later, the answer was the same one the American Physical Society had delivered in a hotel ballroom in 1989.

There was, however, a small and unexpected silver lining. In chasing a discovery that almost certainly wasn’t there, the Google-backed team built genuinely useful new tools — calorimeters that hold their precision under extreme conditions, and techniques for making and studying materials packed unusually full of hydrogen — that researchers in adjacent, very real fields of energy science have since put to use. It was, as one of the project’s own retrospectives in Nature put it, a program that failed to find cold fusion and still counted as a success.