Still in Suspense

Home , Heike Kamerlingh Onnes

ven bouncers in New York City night clubs were aware of our notoriety,” says Paul Grant, thinking back to the “E 1987 March meeting of the American Physical Society (APS). The hype had been building for months, as newspapers, magazines and morning tele vision talk shows heralded jaw-dropping announcements from physics labs. A techno STILL IN logical revolution seemed at hand, promising an era of levitating trains, coin-sized comput ers and power lines that could span continents without losing energy. When the meeting finally convened, says Grant, a physicist at the energy consulting firm W2AGZ Technolo CONSORTIUM/SPL BIRMINGHAM HIGH TC D. PARKER/IMI/UNIV. gies in San Jose, California, anyone with an APS badge who arrived at a trendy club aptly SUSPENSE named ‘The Limelight’ was ushered straight to the front of the queue. A quarter of a century after the discovery of Yet the public’s excitement was nothing com pared with the eager frenzy of the physicists. On high-temperature superconductivity, there the evening of Wednesday 18 March, more than 1,800 APS attendees squeezed into a ballroom at is still heated debate about how it works. the New York City Hilton (while another 2,000 milled outside) to watch a marathon set of pres entations that lasted more than 7 hours. At the BY ADAM MANN sometimes-raucous symposium — dubbed the ‘Woodstock of physics’ — researchers devoured the latest findings on what was easily the most astonishing discovery their field had seen in a generation: materials that became supercon ductors at high temperatures. ‘High-temperature’ was a relative term: even the best of the materials would not transition to become superconducting — having no resist ance to an electric current — until it was chilled below 93 K (roughly 200 °C below room tem perature). But that was nearly four times higher than the transition temperature of any previ ously observed superconducting material, and shattered what had once seemed to be a solid theoretical upper limit of 30 K. Everyone in the

A sample of a high-temperature superconductor hovers in a magnetic field. 1911 1957 1986

A century of superconductivity Georg Bednorz (left) and Alex Müller find Heike Kamerlingh Onnes (seated centre front) a copper oxide material that becomes and his colleagues discover superconductivity. superconducting at 35 K. He receives the Nobel prize in 1913.

John Bardeen, Leon Cooper and Robert Schrieffer

AIP EMILIO SEGRE VISUAL ARCHIVES VISUAL AIP EMILIO SEGRE (left to right) publish a theory of superconductivity

that predicts a maximum transition temperature of ARCHIVES VISUAL IBM, AIP EMILIO SEGRE GIFT OF B. PIPPARD;

30 K. They are awarded the Nobel prize in 1972. COLLECTION, PHYSICS TODAY ARCHIVES, VISUAL AIP EMILIO SEGRE

ballroom knew that, whatever was going on, it bound inner core of electrons. But their loosely happens at very low temperatures, their quan was something profoundly new. attached outer electrons become unbound, tum-mechanical wavefunctions align, draw Better still, they knew that 93 K could be collecting into a mobile ‘electron sea’. Under ing the pairs into a collective state known as a achieved easily with cheap, plentiful liquid the influence of an electric field, this ocean of condensate. Once there, they keep one another nitrogen as a coolant, instead of the expensive, free electrons will drift throughout the lattice, in check because breaking up one pair would tricky-to-handle liquid helium required by forming the basis of conductivity. raise the energies of all the others. The net the earlier superconductors. Suddenly, appli In a normal metal, this motion isn’t always result is that they all flow together without cations of superconductivity such as lossless predictable: no matter how cold it gets, random interruption, creating superconductivity. power lines seemed economically feasible. And thermal fluctuations scatter the electrons, inter The theory was very successful, making the room was alive with an even more electrify rupting their forward motion and dissipating many predictions that were quickly confirmed ing idea: could there be materials that super energy — thereby producing electrical resist by experiment. But it also implied that the conduct without any refrigeration at all? ance. But as some metals are cooled to tem forces binding the Cooper pairs were very fee But 25 years after the publication of the first peratures close to absolute zero, the electrons ble, so they would be ripped apart by thermal paper on high-temperature superconductivity1, vibrations at anything other than extremely such materials remain a dream. So do most of low temperatures. “Armies of researchers in the the miraculous-sounding applications. And “It’s not that there’s 1950s and ’60s worked on improving the tem so does a deep understanding of what is going perature range,” says Jan Zaanen, a theoretical on. Despite increasingly refined experimen no theory; there are physicist at Leiden University. “But they soon tal techniques and nearly 200,000 published lots of theories — realized that they could not give rise to super papers, physicists still do not have a complete conductivity above 25 K or 30 K” — tempera theoretical explanation for high-temperature just none that most tures that generally require elaborate cooling superconductivity. “It’s not that there’s no systems for liquid helium, which boils at 4.2 K. theory; there are lots of theories — just none people agree on.” This did not stop the use of superconduct that most people agree on,” says John Tran ing wires and films in certain high-value appli quada, a physicist at the Brookhaven National cations such as medical magnetic resonance Laboratory in Upton, New York. suddenly shift into a highly ordered state and imaging (MRI) machines and particle collid travel collectively without deviating from their ers. But the expense seemed to rule out any SLOW PROGRESS path. Below a critical temperature that is unique wider application. Still, history offers some reassurance. Physi to each of these metals, the electrical resistance Then, in June 1986, physicists Georg Bed cists took 50 years to understand conventional falls to zero and any current flows practically norz and Alex Müller at the IBM Laboratory superconductivity — which was discovered forever. They become superconductors. in Zurich, Switzerland, reported1 that they had 100 years ago in the laboratory of Heike But why does this ordered state form? In created a material that became superconduct Kamerlingh Onnes, at Leiden University in February 1957, three physicists — John Bar ing at 35 K. The finding was dramatically con the Netherlands (see ‘A century of supercon deen, Leon Cooper and Robert Schrieffer, all firmed in January 1987, when physicists in the ductivity’). On 8 April 1911, after testing for then at the University of Illinois in Urbana- United States found a material in the same class electrical resistance in a sample of mercury at Champaign — published the first complete that became superconducting at 93 K (ref. 3). 3 K, Onnes wrote down “Kwik nagenoeg nul answer2. The Woodstock of physics followed barely two (Mercury practically zero)”, marking the first According to their proposal, now known months later. observation of a superconductor. as BCS theory, an electron moving through One of the many astonishing aspects of A step towards an explanation of supercon a positively charged lattice of atomic nuclei Bednorz and Müller’s work was that they ductivity came in the 1920s with the develop leaves behind a small wake, like the deforma were looking not at metals, but at insulating ment of quantum mechanics, which provided tion caused by a bowling ball rolling across materials called copper oxides, which physi an underlying model for the structure of a mattress. The distortion pulls in another cists would soon dub cuprates. In particular, ordinary metals. Metal atoms form a regu electron, and the two become what is known they were investigating what happens when lar crystalline lattice and hang on to a tightly as a Cooper pair. If many such pairs form, as a cuprate is ‘doped’, or has foreign elements 1987 1991 2008

January: High-temperature Philippe Monthoux, Alexander superconductivity is Balatsky and David Pines confirmed in cuprates, publish the spin fluctuation this time at a temperature theory of high-temperature of 93 K. superconductivity.

March: The American Physical Society hosts the AIP EMILIO SEGRE VISUAL ARCHIVES VISUAL AIP EMILIO SEGRE 1993 'Woodstock of physics' (pictured). And Phillip Researchers discover a J. BOBROFF, LPS/ORSAY, FRANCE/WIKIMEDIA LPS/ORSAY, J. BOBROFF, Anderson posits the material that becomes resonating-valence-bond superconducting at 135 K, Hideo Hosono and his theory as the mechanism setting a world record for co-workers discover a new for high-temperature the highest transition class of superconductors, superconductivity. temperature. iron pnictides (pictured).

21 JULY 2011 | VOL 475 | NATURE | 281 © 2011 Macmillan Publishers Limited. All rights reserved NEWS FEATURE such as lanthanum or barium introduced into that of its neighbour: one electron will have Moreover, finding the pnictides has reas the parallel planes of copper and oxygen that its spin up, the next down, the next up, and sured researchers that they might be able to comprise its structure. What they found was so on. The magnetic fields produced by the find other high-temperature superconductors, that the foreign atoms freed up the outer elec spins lock the electrons in place. But in doped providing more information or perhaps even tron of some of the copper atoms, which then cuprates, the foreign atoms break up this rigid a path to the elusive room-temperature super flowed through the lattice. If the cuprate was chequerboard pattern, giving the spins room conductor. “Once there’s two, there’s a high then cooled enough — to a temperature that to wobble. A passing electron can then set up probability of there being more,” says Andrew depended on how it was doped — the elec a pulsating pattern of spins analogous to the Millis, a physicist at Columbia University in trons would flow freely, and the material would lattice distortions of conventional supercon New York. become superconducting. ductivity. This disturbance then draws moving Researchers have made progress in practical This strange state of affairs — superconduc electrons together, allowing them to associate applications. In the past five years, for example, tivity in an insulator — quickly led physicists to into Cooper pairs and achieve a superconduct they have managed to string cuprate materials re-examine their basic ideas about condensed ing state. into superconducting tape that can be used in matter. But because some of the experiments In the early days, says Tranquada, advocates power transmission cables or MRI machines were unknowingly done on impure samples, of these two mechanisms were at loggerheads cooled with liquid nitrogen. people were having trouble reproducing the as much as anyone else in this field. But after results. “The first years of the field were very a while, he says, “it becomes easier to relax a THE ROOT OF THE MATTER confusing,” says Patrick Lee, a physicist at the little bit and try to start discussing where the No one is predicting a full understanding of Massachusetts Institute of Technology in Cam points of agreement are and where the points high-temperature superconductivity any time bridge. Hypotheses invoking bizarre and exotic of disagreement are. We can get beyond opin soon — not least because such an account physics cropped up, often without much evi ions and try to make some progress by agree would have to make sense of the huge num dence to back them up. ber of papers. “A rich enough theory should The field soon broke up into competing explain everything and not just cherry pick,” camps, each advocating a different theory. “Ladies and says David Pines, a physicist from the Univer Researchers would often ignore data that did sity of Illinois at Urbana-Champaign. not jibe with their pet theory, clinging almost gentlemen, that But it’s not always clear exactly what needs to religiously to their ideas, and attacking those man is a liar — don’t be explained. Roughly 15 years ago, for exam who believed otherwise. ple, researchers discovered that some high- Kathryn Moler, a physicist at Stanford Uni listen to a word he’s temperature superconductors allow electron versity in California, recalls a colloquium in pairs to form above the transition tempera which a scientist in the audience stood up, saying!” ture. In this ‘pseudogap’ regime, the material pointed a finger at the speaker and shouted, spontaneously organizes itself into stripes: lin “Liar! Liar! Ladies and gentlemen, that man ear regions that act like rivers and carry elec is a liar — don’t listen to a word he’s saying!” ing on some experiments or calculations that tron pairs through the insulating landscape Igor Mazin, a physicist at the Naval Research may help.” Most researchers now broadly agree where electrons remain stuck in place. “It’s a Laboratory in Washington DC, remembers a on many aspects, such as the importance of precursor state to the superconducting state conference in 1989 when physicists promoting magnetic interaction. and is therefore fundamental to understand the different theories stood on stage “yelling Things have also calmed down a bit in ing this problem,” says Ali Yazdani, a physicist like schoolchildren”. the laboratory, as improved techniques have at Princeton University. Not so, says Pines, who Eventually, the cacophony sorted itself into helped researchers to weed out the more exotic thinks the pseudogap state “interferes with the two theories with which most physicists theories and refine those that remain. A good superconductivity but is not responsible for it”. now work. The first, resonating-valence-bond example is angle-resolved photoemission Much as physicists had to wait for highly theory4, is largely the creation of Philip Ander spectroscopy (ARPES), a method that uses developed quantum-mechanical tools to son, a condensed-matter physicist at Princeton high-energy photons to probe what electrons unlock the secret behind traditional supercon University in New Jersey. The theory states that are doing. “In 1993, the best we could do was ductivity, researchers today may require future the electron-pairing mechanism is imprinted four spectra in 12 hours,” says Zhi-Xun Shen, ideas to complete their task. in the cuprates’ structure. Neighbouring a physicist at Stanford University who works If nothing else, the field’s early quarrels copper atoms can become linked through with ARPES. “One of vastly superior quality have ensured that only the most determined chemical valence bonds, in which they share now takes 3 seconds.” researchers have stayed. Those remaining are electrons with opposite spins. Typically, the And in 2008, Hideo Hosono and his col perhaps humbled by their experiences. “I think bonding locks these spin pairs in place, pre leagues at the Tokyo Institute of Technology our biggest problem has been human fallibil venting any current from being carried. But in Japan discovered a second class of high- ity,” says Anderson. And perhaps these initial when the material is doped, the pairs become temperature superconducting material — this difficulties have helped to forge theories that mobile and the valence bonds become Cooper time based on iron and arsenic — called pnic can stand the test of time. “In the end, it’s your pairs that condense into a superconductor. tides6. These materials superconduct at lower competitor that makes you strong,” says Shen. ■ The second theory, called spin fluctua temperatures than most cuprates — often only tion5, has the most support in the community. below 40 K— but they have given theorists a Adam Mann is a freelance writer based in Devised by Philippe Monthoux of the Univer new arena for testing their ideas. Oakland, California. sity of Edinburgh, UK, Alexander Balatsky “It’s almost like a do-over,” says Thomas from Los Alamos National Laboratory in New Maier, a physicist at Oak Ridge National Lab 1. Bednorz, J. G. & Müller, K. A. Z. Phys. 64, 189–193 (1986). Mexico and David Pines from the University oratory in Tennessee. Pnictides have a more 2. Bardeen, J., Cooper, L. N. & Schrieffer, J. R. Phys. of Illinois–Urbana Champaign, it posits that complex structure than cuprates, but they Rev. 106, 162–164 (1957). without doping, cuprates are locked into an might help to uncover which phenomena are 3. Wu, M. K. et al. Phys. Rev. Lett. 58, 908–910 (1987). 4. Anderson, P. W. Science 235, 1196–1198 (1987). ordered state called an antiferromagnet. That central to high-temperature superconductiv 5. Monthoux, P., Balatsky, A. V. & Pines, D. Phys. Rev. means that the outer electron on each copper ity, and which are simply due to the copper Lett. 67, 3448–3451 (1991). atom lines up such that its spin is opposite to oxide structure. 6. Takahashi, H. et al. Nature 453, 376–378 (2008).