It would be unfair to call it a philosopher’s stone, yet there is something beguiling about the search for a room-temperature superconductor. This material would be able to transmit electricity perfectly, without any resistance. It could pick up renewable energy where it is abundant and deliver it efficiently to faraway cities, going a long way towards solving the climate crisis.
No wonder, then, that when not one, but two such materials were supposedly discovered last year, the physics world went into a frenzy. In March 2023, researchers reported a material known as “red matter” that could purportedly do the business at 21°C (70°F), albeit only at incredible pressures. A matter of weeks later, news broke of another substance called LK-99 that apparently worked at both room temperature and ambient pressure. Alas, all that glitters is not gold – both claims have now been widely dismissed.
But the fuss over those studies obscures a more subtle and interesting truth: broader research in pursuit of a practical superconductor is racing forwards and there is a sense that, finally, the search is turning a corner. In the past few years, there have been more experimental breakthroughs than you can shake a stick at, while theorists are honing a wealth of methods to predict the composition of new superconducting materials from scratch. “Folks my age can remember when it was absolutely certain: there will never be a room-temperature superconductor,” says , a physicist at the University of Oxford. “Only now we’re realising how wrong we were.”
Advertisement
It was back in 1911 that physicist Heike Kamerlingh Onnes discovered that at -270°C – just 3°C above absolute zero, the coldest possible temperature – the electrical resistance of mercury suddenly vanishes. No one expected this behaviour. All the top physicists of the day, including Albert Einstein, had a go at explaining it. But it wasn’t until nearly 50 years later that a trio of physicists – John Bardeen, Leon Cooper and John Robert Schrieffer – cracked the puzzle. BCS theory, which bears their initials, is now regarded as a pinnacle of 20th-century science: beautiful in its simplicity, formidable in its predictive power.
Cooper pairs
To get your head around the idea, imagine zooming into the innards of a superconducting material, where negatively charged electrons create an electric current by moving through a lattice of positively charged atomic nuclei. As an electron moves, it attracts those nearby nuclei, setting them off in ripples of positive charge in its wake. This attracts another electron, dragging it behind the first, as if it were on a leash. These paired electrons – known as a Cooper pair – are then immune to the vibrations in the lattice of atoms that make up the material, which usually cause electrical resistance.
Unless, that is, the lattice is vibrating so strongly that it has enough energy to snap the Cooper pair’s leash. In physics, more heat equals more vibrations, which explains why, according to BCS theory, superconductivity generally occurs at very low temperatures. Like any good hypothesis, BCS also made a prediction. It suggested that materials in which the lattice atoms were relatively light would superconduct at higher temperatures. Being lighter, these atoms would ripple more easily, creating a more robust leash between the paired electrons.
The theory explained all the superconductors then known, metals such as lead, niobium and tin, as well as the original mercury. For all of these, the point at which superconduction began – the so-called critical temperature – was within a few degrees of absolute zero.
Then came 1986, and the surprise discovery by and K. Alex Müller at the IBM Zurich Research Laboratory in Switzerland of superconductivity in a copper oxide-based material, or cuprate, at a relatively balmy -238°C. Within a few years, other groups found similar materials that worked at even higher temperatures, up to -180°C. This was quite a shock. Not only were these materials regarded as insulators that don’t usually conduct electricity at all, but they were utterly failed by BCS theory and its insistence that only materials made of very light atoms could superconduct at higher temperatures. Copper and oxygen didn’t fit the bill. These cuprates became known as “unconventional” superconductors because they defied the BCS orthodoxy.
The materials kept getting better, though only up to a point. Today, the best unconventional superconductors operate at about -140°C. Still, that is good enough for some applications. They are used to make extremely strong magnets, for example, such as those found in MRI machines. We cool them to frigid temperatures using liquid nitrogen.
But what everyone really longed for was a superconductor that could be used anywhere, a technological panacea that would, among other things, revolutionise electricity networks. To get towards that, we needed a way to go beyond mere theory and test what the electrons inside unconventional superconductors are doing.
Davis and his colleagues managed exactly this in work published in 2022 that used a scanning tunnelling microscope, an instrument in which a metal needle scans the surface of a sample, with electrons hopping from it to the material being examined. They analysed one particular cuprate superconductor using two slightly different needles, one of which was itself superconducting. This allowed the researchers to obtain . “It only took me just slightly less than 30 years,” jokes Davis. “Nobody had ever visualised Cooper pairs before, by any technique whatsoever.”
The team saw that the Cooper pairs were most numerous where hopping between the material and the needles was easiest. This, according to Davis, is strong evidence in support of a particular hypothesis of unconventional superconductivity put forward by the late Nobel laureate Philip Anderson. He said that electrons pair up in cuprates not by moving continuously, as in BCS theory, but in such a way that they correlate a quantum mechanical property called spin – one electron spins down, the next up, and so on.
The experiment will have to be repeated on a wide array of unconventional superconductors before Anderson’s hypothesis is fully accepted. If it is, researchers can then confidently use the idea to predict the structures of new, better superconductors. That won’t be easy though – Anderson’s idea involves much more complicated interactions between the electrons than BCS theory and the simulations are incredibly onerous for even the best supercomputers.
AI and superconductors
Then again, maybe we don’t need to bother with predictions from first principles at all. This is the hope offered by artificial intelligence, which is increasingly seen as a way to find better superconducting substances based on trends in existing experimental data.
Last year, a team led by at Johns Hopkins Applied Physics Laboratory in Maryland trained an AI algorithm using SuperCon, a database of the compositions of more than 16,000 known superconductors and the temperatures at which they start to work. In its initial computation, the AI predicted dozens of possible superconducting materials. The researchers already knew from the literature that some of these weren’t really superconductors, so they put those predictions aside and made a handful of the other compounds suggested.
These were nothing special either, so the team fed all the negative data back into the algorithm and reran it. Still nothing.
The researchers repeated this process and by the fourth iteration, the AI spat out what they wanted: predictions for six superconductors, five of which were already known to be genuine (though not included in the original training data). Tentatively, they synthesised the other one, an untested alloy of the metals zirconium, indium and nickel. Lo and behold, when they cooled it below -264°C, . Not warm enough to rock anyone’s world, but proof of AI’s potential.
Stiles doesn’t know the mechanism by which the zirconium alloy works, and neither does the algorithm. It doesn’t have to: like all AIs, its prediction boils down to statistical analysis. And it can only get better. “With this approach, the more data you feed in, the more predictive it gets,” says Stiles. “Unlike me – the more data I feed on, the more I forget the earlier data.”
Or perhaps we go back to BCS superconductors. For decades, that didn’t seem promising because none operated much above absolute zero. But they did at least have a clear rule that pointed the way to higher-temperature operation: use lighter atoms. Indeed, the best performing superconductor of all may well be a fabled metallic form of the lightest element, hydrogen, which scientists have been trying and failing to synthesise for nigh-on 90 years. The next-best contenders could be the superconducting hydrides, alloys that contain as much hydrogen as possible – and it is these that have suddenly set the field alight.
Hydrides and nickelates
Their rise began in 2015 when, on the basis of a promising prediction a year earlier, a team at the Max Planck Institute for Chemistry in Mainz, Germany, experimented with hydrogen sulphide. Subjecting the sample to pressures of 90 gigapascals, roughly a million times atmospheric pressure, the researchers found that it began superconducting at -83°C, nearly 60°C warmer than the best cuprate. Other results swiftly piled in. Lanthanum superhydride, an alloy that contains 10 hydrogen atoms for every lanthanum one, has proved especially intriguing. By 2018, it was exhibiting signs of . A year later, it was made to work at (9°ąó).
With things inching so close to room temperature, superconductivity research is becoming more competitive than ever. Perhaps that goes some way to explaining the furore over red matter and LK-99. The red matter paper has now been retracted at the request of some of its co-authors, who, among other issues, cited inaccuracies in the data. Meanwhile, repeated attempts to replicate the LK-99 experiments have failed. żěè¶ĚĘÓƵs complain that they no longer know what to trust, and at present there is no consensus on what the highest temperature at which a superconductor works is, or in what material.
Arguably, that is irrelevant. All the superconducting hydrides investigated experimentally only work at high pressures, up to around 300 gigapascals, more or less the squeeze you would find at the centre of Earth. Achieving that requires a tiny sample and a diamond anvil press – like a thumbscrew, but with a pair of opposing diamonds to inflict the torturous force. Clearly, even if a high-pressure hydride does pass the room-temperature milestone, practical applications will be few.
Stalemate, then? Theorists widely agree that simple hydrides involving hydrogen and another element can’t superconduct much beyond room temperature without being subjected to very high pressures. Equally, the original non-BCS superconductors, cuprates, are at something of a dead end too: progress plateaued in the mid-1990s. “It’s tantalising,” says , a theorist at the University of Cambridge. “You see these nice curves of recorded critical temperatures go shooting up, but then they top out.”
However, there are other avenues to explore. In 2019, a team at Stanford University in California discovered that materials similar to cuprates, but with nickel atoms substituting for copper, . At present, the maximum critical temperatures for these “nickelates” is around -193°C at ambient pressure, placing them well behind the pack. Yet this finding opens up a whole new set of materials for physicists to play with and optimise.
On top of that, we discovered another family of superconductors based on iron back in 2006, though the best they can manage so far – again, at standard atmospheric pressure – is around -217°C. But the point is that this expands the space in which we can experiment. “We went from one family of high-temperature superconductors to four,” says at the University of Bristol, UK. “This demonstrates that multiple routes can lead to room-temperature superconductivity, and makes me very optimistic that room-temperature superconductivity is possible at ambient pressure.”
Much hope now centres on the hydrides – but more complex ones, involving two elements in addition to hydrogen, rather than just one. Calculations suggest that the extra elements can help stabilise the atomic structures, rendering high pressures unnecessary. Earlier this year, an international group conducted a theoretical sweep of more than a million such hydrides and found that a subset that contain magnesium ought to superconduct at up to . At the same time, another international group of researchers, including Pickard, corroborated the result and also identified a particular magnesium-iridium hydride that should . “If we’re right, we’ve beaten the cuprates at ambient pressure,” says Pickard.
As he admits, these studies can be taken both ways. The good news is that higher-temperature superconductivity is possible without high pressures. The bad news is that the extensive search still didn’t turn up any room-temperature candidates. “It’s good,” says Pickard, “but not good enough.”
Yet, as always in superconductivity research, surprises may await. At a meeting of the American Physical Society in March, Adam Denchfield at the University of Illinois Chicago presented the results of a new way of searching for promising hydrides. Rather than trying to work out how to reduce the pressures necessary to stabilise known superconducting hydrides, he and his colleagues do the opposite: start with those known to be stable at ambient pressures and see how they can be altered for high-temperature superconductivity. They found that a particular yttrium hydride with a scattering of lithium should, with some tweaking of the lithium content, .
Still not room temperature, you might think. But there are error bars on that number and “they go both ways”, Denchfield said at the meeting. He pointed out that the same theoretical prediction for lanthanum superhydride gave a critical temperature 60°C cooler than its measured value. Might -53°C be a similar underestimate for yttrium hydride? If so, room temperature suddenly isn’t so far off, after all.
Article amended on 9 May 2024
We corrected Christopher Stiles’ affiliation
