NEWSFOCUS on July 21, 2011

After 2 decades of monumental effort, still cannot explain 100,000 papers on the materials. Several theo- rists claim they have deciphered them— high-temperature . But they may have identified the although their explanations clash. Still, high- puzzles they have yet to solve temperature superconductivity has refused to submit to some of the world’s best minds. “The theoretical problem is so hard that there isn’t an obvious criterion for right,” says High T : The Mystery Steven Kivelson, a theorist at Stanford Univer- www.sciencemag.org c sity in Palo Alto, California. Experimenters are producing a flood of highly detailed data, but physicists struggle to piece the results together, That Defies Solution says Joseph Orenstein, an experimenter at the University of California, Berkeley, and TWENTY YEARS AGO, A FIRESTORM OF giddy enthusiasm. “We had prominent people Lawrence Berkeley National Laboratory. “It

discovery swept through the world of . saying it would all be explained quickly and must be close to unique to have so much infor- Downloaded from German experimenter J. Georg Bednorz and that we would have superconducting power mation and so little consensus on what the his Swiss colleague Karl Alexander Müller lines and levitating trains,” Ashcroft says. questions should be,” Orenstein says. kindled the flames in September 1986 when Ashcroft himself had doubts, however, as The problem is more than a sliver under they reported that an odd called he told a class of graduate students a few the nail. High-temperature superconductiv- lanthanum barium copper oxide carried elec- months later. (I was a member of the class.) ity has shown that physicists’ conceptual tricity without any resistance at a temperature The materials comprised four and five ele- tools can’t handle materials in which elec- of 35 kelvin—12 degrees above the previous ments and possessed elaborate layer-cake trons shove one another so intensely that it’s record for a superconductor. The blaze ran wild structures. They broke the rules about what impossible to disentangle the motion of one a few months later when Paul Chu of the Uni- should make a good superconductor. In short, from that of the others. Such “strongly cor- versity of Houston, Texas, and colleagues syn- Ashcroft predicted, high-temperature super- related” electrons pop up in nanodevices thesized yttrium barium copper oxide, a com- conductivity would remain the outstanding and novel magnets, organic conductors and pound that lost resistance at an unthinkable problem in for the other exotic superconductors. “High- 93 K—conveniently warmer than liquefied air. next 25 years. temperature superconductivity is the A frenzy of slapdash experimenting and That prognostication is coming true. Two stumbling block of the whole discipline of sensational claims ensued, says Neil Ashcroft, decades after high-temperature super- condensed matter physics,” says Peter a theorist at Cornell University. He organized a conductors were discovered, physicists still do Abbamonte, an experimenter at the Univer- session on the new high-temperature super- not agree on how electrons within them pair sity of Illinois, Urbana-Champaign. conductors at the meeting of the American to glide through the materials effortlessly at In spite of the difficulty of the puzzle, many Physical Society in New York City the follow- temperatures as high as 138 K. Researchers physicists say they are closing in on a solution. ing March. The “Woodstock of physics” haven’t failed for lack of trying. According to Most now agree on certain key properties of

stretched until 4 a.m. and bubbled over with some estimates, they have published more than the materials. Precision experiments are AND M. O. JONES/UNIVERSITY OF OXFORD EDWARDS CREDITS (TOP TO BOTTOM): P.

1072 17 NOVEMBER 2006 VOL 314 SCIENCE www.sciencemag.org Published by AAAS HIGH-TEMPERATURE SUPERCONDUCTIVITY TURNS 20  NEWSFOCUS

Hot stuff. The structure of mercury barium calcium conductors shove one another so mightily High Tc Superconductivity copper oxide, a superconductor at 138 K. that they tend to jam up with one electron on each copper ion, like gridlocked commuters. revealing surprising details of the compounds. That impasse can be broken only by tweaking And computer simulations—and perhaps even the material’s chemical composition to siphon mockups fashioned of ultracold atoms and away some of the electrons to create positively laser light—could soon show physicists charged “holes,” a process called doping. whether their basic model of the problem is cor- The challenge then is to explain how elec- rect. “If I had to make a prediction,” Kivelson trons that fiercely repel each other manage to says, “I would say that in 10 years time the pair anyway. Some researchers argue that problem will be solved.” waves of magnetism play a similar role to the Conventional Superconductivity one play in conventional supercon- The ultimate chess game ductors. Others focus solely on how the elec- Even “conventional” superconductivity, which trons shuffle past one another in a quantum- was discovered in 1911, is mind-bending. Elec- mechanical game of chess. Still others say that trons in a metal move in quantum waves of dis- patterns of charge or current, or even phonons, tinct energies. Quantum mechanics prohibits play a crucial role. Pairing might even require two electrons from occupying the same wave or all of these things in combination, which “state,” so they stack into the states from the would be many physicists’nightmare scenario. lowest energy on up. But when metals such as lead and niobium are cooled to near absolute Familiar solutions zero, the electrons in them can lower their total Some of the theories being refined today Copper Oxygen Electron Niobium energy by pairing like ballroom dancers. That emerged soon after Bednorz and Müller’s

partnership produces superconductivity, as discovery, and the dividing lines that run Shall we dance? Instead of the motion of ions, the on July 21, 2011 explained in 1957 by theorists , through the field were drawn in those heady subtle waltz of electrons along atomic planes may , and . days. For example, as early as 1987 some cause pairing in high-temperature materials. The pairing alters the spacing of the rungs theorists argued that high-temperature super- on the energy ladder, creating a gap near the top conductivity arose not from phonons but tions, creating an up-down-up-down pattern of the stack. To break from its partner, an elec- from the interaction of the electrons alone. known as . The electrons tron must jump the gap to an empty state. There But even those who agree on that principle can tilt and flip, and waves of wobble coursing isn’t enough energy around to allow that, so the often disagree on the details. through this arrangement can provide the glue pairs glide along unperturbed. Something must The idea that waves of magnetism drive for pairing, says David Pines, a theorist at Los

glue the pairs together, and according to the the superconductivity is based on the fact that Alamos National Laboratory in New Mexico www.sciencemag.org Bardeen-Cooper-Schrieffer (BCS) theory, the electrons act like little magnets. Those on and the University of California, Davis. adhesive is quantized vibrations of the crys- adjacent copper ions point in opposite direc- But Philip Anderson, a theorist at Princeton talline material, or “phonons.” A University, says that no glue is necessary. Just passing electron attracts the slower- “The theoretical problem is so months after the discovery of high-temperature moving ions in the crystal lattice, hard that there isn’t an obvious superconductors, he proposed a scheme known which squeeze together to produce a as the resonating valence bond (RVB) theory, knot of positive charge that attracts criterion for right.” which focuses on subtle quantum connections Downloaded from another electron (see diagram). —Steven Kivelson, Stanford University between electrons on neighboring copper ions. High-temperature materials lit- In the theory, no waves of any kind pass erally take superconductivity to a between electrons, Anderson says. new plane. The compounds con- Thanks to the weird rules of quantum tain planes of copper and oxygen mechanics, each electron can point both up ions that resemble chess boards, and down simultaneously. Moreover, neigh- with a copper ion at every corner of boring electrons can join in an odd quantum a square and an oxygen ion along state called a singlet in which both electrons each side. Electrons hop from cop- point both up and down at once, but the two per ion to copper ion. Between the electrons always point in opposite direc- planes lie elements such as lan- tions—either down-up or up-down. When ; LINDA A. CICERO/STANFORD NEWS SERVICE NEWS ; LINDA A. CICERO/STANFORD thanum, strontium, yttrium, bis- enough holes open in the plane, singlets form

SCIENCE muth, and thalium. But it is along and begin to slide freely past one another, the copper-and-oxygen planes that eventually producing superconductivity. the electrons pair and glide. Others contend that both the magnetic Just how that happens is any- fluctuation and RVB theories leave out some thing but clear. The electrons in an essential piece of physics. Stanford’s Kivelson ordinary metal hardly notice one believes stripes of electric charge on the planes, another and interact mainly with which have been seen in some materials, may phonons. In contrast, the electrons be necessary to trigger the pairing. Chandra in high-temperature super- Varma, a theorist at the University of California, CREDITS (TOP TO BOTTOM): P. HUEY/ CREDITS (TOP TO BOTTOM): P.

www.sciencemag.org SCIENCE VOL 314 17 NOVEMBER 2006 1073 Published by AAAS NEWSFOCUS

Riverside, proposes that loops of leagues present data online in Science this week current flowing inside each copper- (www.sciencemag.org/cgi/content/abstract/ and-oxygen square are key. 1134742) consistent with this interpretation. Still others argue that high- Preformed pairs are too much to swallow temperature superconductivity for other researchers, who say the pseudogap is may not have one root cause. a sign of something else that clashes with “There is no silver bullet that is Strange metal superconductivity. For example, Zhi-Xun e r

going to explain everything,” says u Shen of Stanford University and colleagues t a Thomas Devereaux, a theorist at r argue online in Science this week (www. e the University of Waterloo in p sciencemag.org/cgi/content/abstract/1133411) m

Canada. The fact that materials Te that there may be two different gaps. Either with very similar structures have Pseudogap way, the strange state might hold the key to

very different critical tempera- o m agne t Metal explaining high-temperature superconductivity, tures shows that the copper-and- says Michael Norman, a theorist at Argonne oxygen planes are not the whole ti fe r National Laboratory in Illinois. “The thing that story, he says. Phonons may still An Superconductor explains the pseudogap may explain the super- play an essential role, such as conductivity as well,” he says. driving up the critical tempera- Doping (holes per copper ion) ture, Devereaux says. Terra incognita. The mysterious and controversial pseudogap Computers and cold atoms As in the beginning, the field phase may hold the key to explaining superconductivity. Ultimately, the mystery of high-temperature is contentious. Recent experi- superconductivity may be solved not in the lab ments have hinted at current loops. “If these undoped material, and it’s an antiferro- or at the theorist’s chalkboard but in the heart are accepted, the theoretical game is over,” magnetic insulator. Dope it to draw between of a computer. Some theorists have turned to 6% and 22% of the electrons out of the planes, Varma says. “That’s why no one wants to numerical simulations of the electrons hop- on July 21, 2011 accept it.” Anderson is equally convinced that and it’s a superconductor. Dope it more, and it ping around the copper planes. If everything his RVB theory is correct—and underappreci- becomes an ordinary metal. springs from the interactions between the ated. “Eighty percent of the field is against These properties can be plotted on a electrons alone, then all the different phases anything—especially anything that might “phase diagram” that, like some medieval and perhaps even the pairing mechanism solve the problem,” he says. map, charts the mysteries physicists face (see should emerge from such simulations, much figure, above). “To solve the whole problem, as the double helix, genes, and the mechanism Mapping out the mysteries you’re going to need to understand the whole of transcription arise from chemical inter- In spite of the discord, researchers have made phase diagram,” says Séamus Davis, an exper- actions between the building blocks of DNA.

progress—especially the experimenters. For imenter at Cornell University. “It could be that The mathematics can vary, but theorists www.sciencemag.org example, in 1994, John Kirtley and Chang focusing on the mechanism is the reason that generally study a scheme known as the Tsuei of IBM’s T. J. Watson Research Center the mechanism hasn’t been found.” Hubbard model, in which the only adjustable in Yorktown Heights, New York, probed the Most intriguingly, at low doping a gap parameters are the ease with which the elec- shape of the cloudlike quantum wave that opens in the ladder of electron energy levels trons hop and the strength with which they describes the paired electrons. In a conven- even at temperatures far above the super- repel each other. Tracking electrons on a grid tional superconductor, electrons can pair in conducting transition. That “pseudogap” sug- might sound easy, but the complexity of the gests that electrons pair at such toasty tempera-

any direction and can sit on top of each other, quantum-mechanical calculations limits Downloaded from so the wave is a sphere. In high-temperature tures, and that superconductivity arises when researchers to grids of a few dozen lattice superconductors, Kirtley and Tsuei found, the “preformed” pairs gather into a single quan- sites. And still they must use approximation the cloud is shaped like a four-leaf clover. That tum wave, some researchers say. “Everything schemes to keep the calculation manageable. “d-wave” shape means that paired electrons we have seen goes in that direction,” says Øys- Such simulations have begun to reproduce sit on adjacent copper ions and tein Fischer, an experi- pairing, stripes, and features of the pseudogap, never on the same ion. menter at the University says André-Marie Tremblay, a theorist at the D-wave pairing would be of Geneva, Switzerland. University of Sherbrooke in Canada. Unfortu- hard to explain with phonons, And Tonica Valla of nately, different approximation methods can but it had been predicted by Brookhaven National lead to different results for the same parame- Anderson and others who Laboratory in Upton, ters, says Douglas Scalapino of the University favored purely electronic theo- New York, and col- of California, Santa Barbara. But that’s not ries. As a result, even most of necessarily a bad thing, he says, those who say phonons play a “Eighty percent of because that very sensitivity sug- role do not believe that they gests that the Hubbard model can alone cause pairing. the field is against produce a variety of effects with By dint of a variety of anything—especially only a little tweaking, just as high- experiments, researchers have anything that might temperature superconductors do. also agreed upon the proper- “I interpret that to mean we have ties common to all the mate- solve the problem.” the right model,” he says. rials, which change with the —Philip W. Anderson, Meanwhile, a wild new kind of

amount of doping. Cook up an Princeton University simulation could be gearing up to CREDIT (BOTTOM): BOB MATTHEWS

1074 17 NOVEMBER 2006 VOL 314 SCIENCE www.sciencemag.org Published by AAAS HIGH-TEMPERATURE SUPERCONDUCTIVITY TURNS 20  NEWSFOCUS

leave computer simulations in the “It could be that temperature superconductivity dust. Physicists have begun to con- has already paid off hand- struct artificial crystals by sus- focusing on the somely. “It has led to the dis- pending ultracold atoms in corru- mechanism is the covery of new materials, of new gated patterns of laser light. In such states of matter, of new con- an “optical lattice,” the spots of reason that the cepts,” says Aharon Kapitulnik, light play the role of the ions, and mechanism hasn’t an experimenter at Stanford the atoms play the role of the hop- been found.” University. Shen says that in ping electrons. The setup might be their quest to unravel the phe- used to create an incarnation of the —Séamus Davis, nomenon, experimenters have Hubbard model with hundreds of Cornell University honed their techniques to new lattice sites and parameters that levels of sensitivity, precision, researchers can tune just by adjust- But even if and speed. “High-temperature super- ing the spacing and the brightness such simulations conductivity has completely changed the of the spots. do produce super- landscape of experimental condensed matter Several groups are already rac- conductivity, they physics,” he says. ing to produce such systems. “In may not yield a At the same time, condensed matter very quick succession, we have conceptual under- researchers have come to see high-temperature jumped over the first few hurdles, and maybe standing of the pairing, some researchers superconductivity as the gateway to a new area the number of hurdles ahead of us is not that say. Others question the relevance of the of study: strongly correlated electrons. “This much bigger than the number behind us,” says simulations to high-temperature super- problem of strongly correlated electrons is the , an experimenter at the conductors. “We don’t know that the new frontier,” says Argonne’s Norman, “and Massachusetts Institute of Technology in Hubbard model is what’s going on in the high-temperature superconductors have [materials],” says Cornell’s Davis. “That’s Cambridge. Using optical lattices, experi- brought it to the fore.” Two decades after their on July 21, 2011 menters could map out the phase diagram of a hypothesis.” discovery, high-temperature superconductors the Hubbard model within a few years, says are viewed less as a singular mystery and Henk Stoof, a theorist at Utrecht University in A threshold more as a threshold to new realms of physics. the Netherlands. “They have all the things they Even without a theory to explain the materials, Physicists hope it won’t take another 20 years need to do it,” he says. physicists agree that the pursuit of high- to cross it. –ADRIAN CHO

APPLICATIONS of Energy (DOE). “The engineering is just www.sciencemag.org beginning. I’ve been seeing a lot more engi- The Next Big Hurdle: Economics neering than before of motors, coils, and so on. That’s a good sign.” Researchers have solved most of the technical challenges. Now companies are struggling to make HTS devices competitive Slowing to a crawl Part of what made the HTS revolution so exciting was that the novel superconductors

Six months after J. Georg Bednorz and Karl nothing like its founders envisioned. “In my Downloaded from Alexander Müller discovered that a family of opinion, we oversold high-temperature super- looked and acted so differently from conven- could conduct electricity without conductivity,” says Lucio Rossi, who heads tional low-temperature superconductors the electrical equivalent of friction, the scien- the magnets and superconductors group at (LTS). The earlier materials were ductile tific buzz swelled into full-scale hype. News CERN, the European particle physics labora- metals, such as the alloy niobium-tin, that accounts gushed at the prospect of magneti- tory near Geneva, Switzerland. could be forged into wires for power cables cally levitated trains, novel sensors, superfast Today’s outlook is decidedly less rosy. or wound into spools for use in magnets, a superconducting computers, and of course, “It’s very difficult to make money on HTS,” key component for motors and generators. lossless electricity transmission cables. For a says John Rowell, a at Arizona State HTS materials, by contrast, are brittle generation that grew up watching the techno- University in Tempe, who notes that no ven- ceramics. In the early discoveries of HTS logical utopia of the Jetsons, the future, it ture capital–funded HTS company in the materials, researchers placed electrodes on seemed, was just around the corner. United States has ever had a year of prof- opposite sides of a millimeter-sized ceramic The trouble is, it’s still there. Two itability. Still, hope springs eternal, and after fleck or perhaps a few-centimeters-long film decades into the revolution, the effort to 20 years of development, HTS equipment of the material. That setup worked to show commercialize high-temperature super- makers seem to be finding ground beneath the drop in resistance characteristic of the conductors (HTS) is not for the faint- their feet. “It’s a slow process,” says Al Zeller, onset of superconductivity. But nobody knew hearted. Successful applications exist, a expert at Michi- how to turn these hard, brittle flecks into kilo- although with names and roles that few peo- gan State University in East Lansing. “But meters of wire. ple would recognize, such as current leads the applications are taking off.” “The materi- Part of the problem is that electrons pass- and cellular base station filters. And although als science in HTS has been terrific,” says ing through HTS materials, unlike those in those and other niche applications are turn- Bruce Strauss, who helps run the super- conventional superconductors, prefer to travel in particular directions through the

CREDIT: CORNELL UNIVERSITY PHOTOGRAPHY CREDIT: ing a profit for their owners, the field is conductivity program at the U.S. Department

www.sciencemag.org SCIENCE VOL 314 17 NOVEMBER 2006 1075 Published by AAAS