Walk into a typical physics lab and you’ll confront a mad profusion of electrical cables, lasers, computers, vacuum pumps, and who knows what else. Enter Robert Cava’s lab, though, and the first thing you’ll see is a six-inch-long dart with green plastic fins stuck into a chart tacked onto the side of a locker. A closer look reveals the chart to be that familiar chemistry-class prop, the periodic table of the elements. The dart has pierced a little square slightly to the right of center, the one belonging to element 29--copper.
Throwing darts at the chart, Cava explains, is how he chooses the raw ingredients for his experiments.
He is joking, but he’s not straying all that far from the truth. Cava, a materials scientist at AT&T; Bell Laboratories’ sprawling research complex in Murray Hill, New Jersey, studies superconductors. Unlike all other substances known to us, superconductors have the seemingly magical ability to conduct electricity without resistance. Physicists have hooked up batteries to superconductors, then removed the batteries from the circuit and watched the current continue to flow as if the batteries were still in place--in one case for as long as four years, at which point the physicists simply tired of the experiment. Prolonging it would not have been advisable. Calculations suggest that superconducting currents, like some Energizer Bunny from hell, can persist for millions of times longer than the universe has thus far existed.
Such extraordinary materials, naturally, could make for an extraordinary material life. Our electricity could flow through superconducting power lines without any loss of energy. Levitating high- speed trains could float on powerful superconductor-generated magnetic fields. Computers, medical scanners, motors, and generators with superconducting components could be made shockingly smaller and more powerful. The potentially immense practical payoff of these resistance-free materials is one reason so many physicists, like Cava, can’t resist studying them these days.
Just a few years ago the field was populated by a mere handful of researchers. Certainly others were aware of superconductivity, but the phenomenon had been observed only at temperatures near absolute zero, the point at which atoms themselves nearly freeze in place. In 1987, however, physicists discovered materials that became superconductors at dramatically higher temperatures. Suddenly it seemed possible that this arcane laboratory phenomenon might break out into the real world. Superconductors were front-page news.
There haven’t been many headlines since. It’s not that the new superconductors were a bust exactly, but they have turned out to be more problematic than anyone had anticipated. Even today, as some important practical applications are at last beginning to emerge from the lab, champions of the technology are more cautious than they were eight years ago. As it turns out, the new superconductors are odd, brittle compounds, and shaping them into useful products--like a wire--has proved an enormous challenge. But even more fundamental than the practical problem is the theoretical one: no one really understands how superconductors work. Physicists cannot agree on a theory that explains the mysterious phenomenon of superconductivity, let alone specify how an experimentalist might search for new superconducting compounds. And this takes us back to Cava’s little joke.
Our job is to find new superconducting materials, says Cava. How do we figure out what to do? This is our answer-- he points to the dart. There is no real theoretical basis for understanding what’s going to happen when we mix different things together. There are no rules at all that direct you. Only your own experience. So in the end, 99 percent of what we try doesn’t do what we want.
Cava’s predicament typifies the murky state of superconductor research today. It’s a contentious field, with theorists falling for the most part into two surprisingly hostile camps, both convinced, like religious zealots, that the truth is theirs. Although to an outsider their debates sometimes recall the hairsplitting disputes of medieval scholastics, many physicists consider superconductivity the problem of their field.
It is one of the deepest and most fundamental problems in physics, says Philip Anderson, a Nobel laureate physicist at Princeton and one of the leading theorists on superconductivity. Perhaps it doesn’t have the glamour or general interest that cosmology has, or quantum gravity, but in terms of fundamentally understanding how the world works, this is very deep and very difficult.
It wasn’t always like this. At one time physicists thought they had superconductors all figured out.
For most of this century the study of superconductivity was an unhurried affair, carried out in the quiet backwaters of physics. The very first superconductor was discovered in the Netherlands, in 1911, by Heike Kamerlingh Onnes, a physicist interested in the properties of materials at extremely low temperatures. At the time he was the only physicist in the world who could liquefy helium, and the only one able to closely approach absolute zero. He used his exclusive supply of liquid helium to chill mercury down to about -452 degrees Fahrenheit, or 7 degrees above absolute zero. One of the properties he then measured was the metal’s electrical resistance.
Kamerlingh Onnes found that as he lowered the temperature, mercury’s resistance decreased steadily until, at about 7 degrees above absolute zero, it suddenly plunged and vanished entirely. He was at a loss to explain what he had seen, but not for a name: supergeleiding, or superconductivity.
Over the next couple of decades Kamerlingh Onnes and other physicists observed the phenomenon in many metals besides mercury. In all cases, it occurred only at temperatures barely above absolute zero. But because liquid helium, the required coolant, was expensive and difficult to handle, superconductors remained a laboratory curiosity--strange and fascinating, but of no practical importance. On the theoretical side, also, superconductivity remained tantalizing but elusive. More than 40 years would pass after its discovery before a trio of physicists would come up with what seemed to be a definitive explanation of how it worked.
The three theorists--John Bardeen, Leon Cooper, and Robert Schrieffer--won a Nobel Prize in 1972 for what is now known as the BCS model of superconductivity. (Bardeen also shared an earlier Nobel for his role in inventing the transistor.) Their theory, published in March 1957, when all three were at the University of Illinois, essentially described the behavior of electrons in metals at very low temperatures.
The atoms in all metals are arranged in stacks of orderly, Levittown-like arrays. Unlike the electrons around an isolated atom in free space, the electrons in a crystal lattice of a metal, like copper, no longer belong to one atom. Instead the electrons become communal property, forming a sort of sea that engulfs the metal lattice. At normal temperatures heat from the environment keeps this atomic lattice constantly vibrating. When a battery, for example, starts a current flowing in the electron sea of a conventional conductor, the electrons in the current encounter resistance as they are slowed or deflected by the electric fields of the atoms in the vibrating lattice. Physicists knew that very low temperatures would quench these thermal vibrations, but they couldn’t see exactly how that would lead to superconductivity.
What Bardeen, Cooper, and Schrieffer proposed was that a two-step process underlies superconductivity. First, at low enough temperatures the disordered electron sea starts to break apart as the electrons team up in loosely bound pairs. This pairing at first seems to defy reason. Electrons, after all, are negatively charged and should repel one another. But a more subtle process comes into play when things get very cold.
Picture a single electron traveling through a metal lattice at low temperatures. When it passes between two neighboring atoms in the lattice, the BCS model says, the negatively charged electron distorts the lattice slightly; the positively charged atoms of the lattice get pulled, just a bit, toward the passing electron. This distortion creates a region of positive charge in the wake of the passing electron. A second electron, attracted to this little zone of positive charge, follows in the wake of the first much as one race car might be pulled along by the slipstream of another.
The chain stops there, though. Electron triplets never form; the pairing process is delicate--the distortion of the lattice that bonds the pairs is fleeting--and nature doesn’t seem to be able to forge stable links between more than two electrons at a time. BCS theory predicts that pairing can happen only at very low temperatures. Heat the lattice to more than about 70 degrees above absolute zero (to -389 degrees Fahrenheit) and thermal energy breaks up the pairs as effectively as a gunshot scatters a flock of birds; the extra heat makes the electrons so energetic that they simply part company.
Once the electrons have all paired up, the stage is set for act 2 of the BCS drama, which features one of quantum mechanics’ bizarre descriptions of particle behavior. Physicists lump all the particles in the universe into two broad families: fermions and bosons. Electrons, protons, and neutrons--that is, just about all the familiar components of matter-- are fermions. Photons--particles of light--are examples of bosons. All these particles have a property that physicists call spin, which describes the tiny amount of momentum generated by a spinning particle. Fermions have exactly half the amount of spin of bosons.
Quantum mechanics holds that fermions and bosons have very different properties. Bosons are gregarious, like inebriated conventioneers. They can all join together in a unified state of energy, as photons do in a laser beam. Fermions, and hence electrons, are loners. No two can occupy exactly the same energy state. Like dancers in a raucous nightclub, the electrons in a metal at normal temperatures are a jumble of distinct energies. Disorder prevails.
But at low temperatures, a radical change takes place in the electrons’ behavior. The wild dance becomes an ordered waltz. After electrons pair off, they act like a different type of particle altogether. Since the spin of two fermions equals the spin of one boson, paired electrons essentially become bosons. They can now all occupy a single energy level, and they tumble into this ground state like rush-hour commuters filling a train. Physicists describe the transition as a phase change, one equivalent to water suddenly turning to ice. The entire superconductor at this point very closely resembles a giant atom. All the electrons move in unison; individual electrons no longer get deflected randomly in the lattice. Just as electrons whirl eternally around a nucleus, the electron pairs course ceaselessly throughout a superconductor.
The BCS model seemed to explain superconductivity satisfactorily. It even predicted a firm upper temperature limit for the phenomenon, a barrier that seemed to relegate superconductors to the status of a permanent physics freak-show attraction. Only liquid helium could chill materials to temperatures as low as -389 degrees, and besides costing five dollars a quart, liquid helium requires bulky refrigeration equipment.
In the decades following the formulation of the BCS theory, a few physicists toiled away at searching for mixtures that might break the theoretical barrier. But no one came close. The best anyone could come up with, in 1974, was a mixture of niobium and germanium that became superconducting at about 41 degrees above absolute zero, or -418 degrees.
Then suddenly in 1987, in a series of stunning, almost accidental discoveries, a number of labs around the world smashed through the old BCS barrier. New materials appeared that became superconducting at -235 degrees, or 224 degrees above absolute zero. That is, of course, still cold by any normal standard, but in the world of superconductors it was a milestone: it meant that instead of liquid helium, physicists could use liquid nitrogen to chill their exotic compounds. Liquid nitrogen is readily available, easy to handle, and at 30 cents a quart, as physicists like to say, cheaper than beer.
The discoveries elated physicists. By 1988 some were predicting that in a few years they might develop room-temperature superconductors, which wouldn’t require any coolants.
Obviously, that didn’t happen. For one thing, these new superconductors were strange--they were ceramics, which to materials scientists means not just things like clay but also nearly any type of solid that is neither purely a metal nor a carbon compound. Normally ceramics are insulators. They don’t carry electric currents at all. (Your toilet bowl doesn’t conduct electricity, Cava points out. Nor do your dinner plates.) Another problem is that ceramics are brittle, and some of the most important applications for superconductors--electromagnets, generators, and motors--require flexible wires.
Besides these practical problems, theorists now found themselves nearly back at square one. Although physicists still believe that the BCS model explains the old superconductors, it can’t explain the newer ones: the chief problem is that superconductivity in these compounds persists at temperatures well beyond where theory predicts it should break down completely.
Theorists, though, perhaps even more than nature, abhor a vacuum, so there has been no shortage of ideas to explain what’s happening. Of all of them, two have emerged as the major challengers to the old BCS theory. Supporters in each camp cite numerous experiments to support their ideas, claiming the other side should wake up and recognize that the battle has been fought and won.
There are many people who are quite sure that they are not confused, but they are in fact totally confused, says Philip Anderson of Princeton. Anderson thinks some of the people working on the problem have let their egos get in the way of good science. Some of the elder statesmen have been the worst, he says. There are people who think that they own the field of superconductivity. After reflecting for a moment he adds, I suppose a lot of people think that I play that role.
Anderson does seem to take a proprietary view of the field. Perhaps that’s because he feels the problem at hand has broad implications for physics beyond understanding superconductivity. He thinks that if physicists can figure out how superconductors work, they will gain new insight into how electrons behave in solids.
Anderson has hammered out a theory for the new superconductors with Sudip Chakravarty, a physicist at UCLA. Like most other theorists, they are focusing their attention on ceramic compounds of copper and oxygen--in particular, a powdery black superconductor called yttrium- barium-copper-oxide.
It is no accident that the dart stuck in the table of elements in Cava’s lab punctures copper. Most of the new high-temperature superconductors consist of sheets of copper and oxygen atoms layered between various other elements, usually yttrium and barium, a formula hit upon by experimenters eight years ago after much trial and error. The copper and oxygen atoms seem to be the crucial ingredients; no one really understands what role the yttrium and barium have. Cognoscenti usually just refer to these materials as the cuprates.
To explain how superconductivity operates in the cuprates, Anderson and Chakravarty propose a radically different way of looking at electrons. In the conventional model of solids, electrons sail through a lattice past other electrons, suffering the odd collision here, the repulsion of another electron there. None of these interactions seriously affects the integrity of the electron as a discrete particle. It still roams proud and free through the solid.
But in the copper-oxide sheets in the new superconductors, and perhaps in other materials as well, Anderson and Chakravarty believe that this description of electron behavior is fundamentally inaccurate. They maintain that it is unrealistic to assume that the electron is not strongly affected by other electrons and atoms in the solid. Such interactions, they calculate, could be quite powerful--powerful enough that in some cases the electron separates into two more-fundamental particles, which they call the holon and the spinon. The holon keeps the electron’s charge; the spinon keeps the electron’s spin.
Holons and spinons require energy to remain separate, and thus exist only when the material is in its normal state. When the material cools and becomes superconducting, the holons and spinons recombine into electrons again; then, just as in BCS theory, the electrons form pairs. In effect, the electrons are transformed into bosons and can again all settle into one coherent energy state. But now there is a crucial departure from the BCS picture. In Anderson and Chakravarty’s model, the electron pairs perform another quantum mechanical trick.
One of the strangest elements of quantum mechanics, demonstrated by many experiments, is that electrons and all other particles sometimes don’t seem to be solid particles at all but instead behave like waves. The well-defined particle becomes a smear. In Anderson and Chakravarty’s model, this waviness enables electron pairs to extend from one copper-oxygen layer to another, tunneling through an intermediate layer of atoms, which in many of the new superconductors consists of yttrium and barium. As the pairs cross these layers, they lose a little kinetic energy and become more stable. The pairs are harder to break up when heated, so superconductivity persists at higher temperatures. But at normal temperatures, most of the electrons in the lattice exist in the form of holons and spinons. Electrons at these higher temperatures can’t form the pairs that lead to superconductivity, because they are bound up--or even split up--by their strong interactions in the copper-oxygen lattice.
Anderson and Chakravarty claim that a number of experiments support their theory. Two seem especially striking. First, says Chakravarty, several labs have reported that as you increase the number of copper-oxygen layers in a superconductor, the compound will remain superconducting at higher temperatures. This jibes with their theory, which says that the layers make the electron pairs more stable.
In a second group of more subtle experiments, researchers have shone infrared light on cuprate materials and measured the amount of reflection. These experiments found that the very same compound will reflect more light while it is superconducting than it will at higher temperatures. The results seem puzzling, but Anderson and Chakravarty say that their theory explains them naturally--they are a consequence of the state of the electrons in the material at different temperatures.
When light hits something, it jiggles electrons. The jiggled electrons immediately emit radiation, which we see as reflected light. But in Anderson and Chakravarty’s model, at normal temperatures the electrons have split into holons and spinons. So the light jiggles fewer intact electrons, and less light gets reflected. When the temperature drops and the material becomes a superconductor, holons and spinons merge, the light jostles many electron pairs, so more light bounces back.
That experiment to my mind is conclusive, says Anderson. It isn’t to everybody else, of course.
Anderson knows his audience.
I don’t think there’s a chance in hell that Anderson’s interlayer pairing has a thing to do with this, says Douglas Scalapino, a physicist at the University of California at Santa Barbara.
While Anderson and Chakravarty believe that a fundamental shift in our understanding of the electron is needed to explain the new superconductors, Scalapino and David Pines, a theorist from the University of Illinois, believe that less drastic measures will suffice.
Much simplified, their model seems quite similar to BCS theory. But they maintain that it’s not the negative charge of the electron that distorts the lattice of the superconductor but the electron’s tiny magnetic field. The magnetic field of the passing electron tugs on nearby atoms in the lattice, distorting the local magnetic field of the copper atoms. The slight magnetic distortion in the electron’s wake attracts another electron. Again, once electrons pair up, they all fall into the same energy state, just as in the BCS picture. Pines’s calculations show that this attraction holds the electrons together at higher temperatures than are possible in the older theory.
If this seems like a minor adjustment to the BCS theory, Anderson and Chakravarty agree. You just take over the BCS equations from a textbook, says Chakravarty. All you are doing is changing one thing. You’re changing the attractive interaction. My commonsense point of view tells me that there is something very special about these materials, that you cannot just tweak textbook equations.
Nevertheless, Pines says that numerous experiments support their theory. You’ve got a cast of thousands, and they all get the same answer, he says.
Many of those experiments test one of Pines and Scalapino’s key predictions. In their theory, as in Anderson and Chakravarty’s, the wavelike nature of the electron becomes very important; pairs of electrons combine to produce a single traveling wave. It turns out that the subtle calculations of quantum mechanics show that the crests of a pair wave traveling in one direction through the copper-oxygen grid will always coincide with the troughs of a pair wave traveling perpendicular to the first pair. The two waves will cancel, and the superconducting current will vanish.
Dale Van Harlingen, an experimental physicist at the University of Illinois, put this odd prediction to the test. He attached a tiny C- shaped piece of lead to a pepper-grain-size square of an yttrium-barium- copper-oxide superconductor. The arms of the lead C touched two adjacent sides of the cuprate square. Then he lowered the temperature and started a current in the little piece of lead, which is a conventional BCS superconductor. The idea was that the current from the lead would enter the cuprate superconductor at right angles. If Pines and Scalapino are right, the perpendicular currents should cancel out in the cuprate because the electron pair waves will annihilate each other. That is exactly what Van Harlingen observed. As a control, he checked to see whether a current would flow through the cuprate if the current came through one side of the square only. It did.
Anderson and Chakravarty’s theory, says Pines, may have seemed plausible a year or two ago, but he believes the tide has turned. Matters have changed dramatically now, and there are some 40 to 50 experiments that support us.
Anderson and Chakravarty say that their own theory can be adjusted to comply with these experiments as well, but that the other camp’s model can’t explain the importance of the layers in the superconductors or their light-reflecting properties.
Who’s right? And will it matter to physicists like Cava?
These are all really good people, says Robert Dynes, an experimental physicist at the University of California at San Diego. They’re not doing foolish things. Dynes, although he has done experiments that seem to support Anderson and Chakravarty’s experiment, believes neither side fully understands superconductors yet.
We don’t know what’s causing superconductivity, he says. At least I don’t. Some of these people think they do. No theory has comfortably described all these experiments. I’ll probably offend a lot of my friends with that statement. But all the cards haven’t fallen into place.
Dynes also says that more is at stake here than figuring out how superconductors work. The thing that theorists like Phil and David Pines and--the list just goes on as long as your arm--the thing that’s got them excited, and that causes all the strife, is the underlying belief that we’re entering a new era in our understanding of electrons. It’s a much, much broader problem. I think people really smell a new era.
The confusion over superconductivity, says Dynes, betrays a serious gap in physicists’ understanding of the most basic properties of matter. Physicists like to think they understand why gases change to liquids, and liquids to solids--and why some materials conduct electricity and others don’t. We understand copper, he says. We understand conduction in aluminum. Why don’t we understand it in copper-oxides?
The theorists may be stymied, but that hasn’t stopped progress on the practical front. A handful of research teams have finally managed to wring something practical from the brittle new superconductors. In April, for example, a team at Los Alamos National Laboratory announced that it had made wires from yttrium-barium-copper-oxide. The researchers claim that their wire--it actually looks more like a thin piece of tape--can carry nearly 100 times the current of any other superconducting wire, and more than 1,000 times as much current as a typical household copper wire. Whether the enormous technological potential of superconductors--the levitating trains, the superconducting power lines, and more--will ever be realized depends largely on whether superconducting wires can be manufactured cheaply and reliably.
The Los Alamos researchers feel they’ve made a promising start. The question is, says Paul Arendt, one of the physicists who developed the tape, is there a market? We think there is. Will people use these? We’re certain they will.
One of the problems the Los Alamos team overcame, one that has slowed progress in the field for years, was to make sure that the superconducting crystals that make up their tape were properly aligned. If the crystal grains don’t line up, the current is drastically reduced. To avoid that problem, the physicists first coated a thin, flexible strip of nickel alloy with a layer of perfectly aligned crystals of a mineral called zirconia. The zirconia acts like a buffer and prevents the nickel from contaminating the superconductor, which gets deposited next. It also acts like a template for it. When a laser vaporizes the cuprate superconductor, the vaporized grains line up naturally as they nestle into the aligned zirconia crystals. Finally the researchers wrap the coated nickel strip around a narrow tube filled with liquid nitrogen to chill the superconductor.
Despite this auspicious beginning, the Los Alamos team still has some years of work ahead of it. So far the tape it’s made is only two inches long and a half-inch wide. Arendt says his team doesn’t see any problems in making the tape longer.
We’re scaling up now, he says. We’re doing this on a shoestring, and we’re hoping to get industrial investors.
Cava meanwhile presses on in his search for new materials. While a successful theory might help guide him and other experimentalists, Cava is more interested in working outside the boundaries set by the theorists.
We are very driven by intuition and empiricism, he says of his lab staff. It’s a very successful method for finding new things. If you theorize too much, you can find only the things that you understand. If you do things beyond what you can theorize about, you find something that’s a surprise. That’s what we want. Surprises.
Cava recently found an unusual boron-carbon superconductor. Although it becomes a superconductor at temperatures well below that of the cuprates, it serves as a reminder that the periodic table may yet hold some surprises for him and that the theorists shouldn’t get too cocky. One of our big goals in here is to find the thing that’s going to open everybody’s eyes, he says.
He walks over to a drawer and pulls it open. The drawer is filled with nearly a hundred small jars that contain almost every element in the universe in one form or another--a sort of messy cosmic medicine cabinet. The small jars tumble about and clack together as the drawer slides out. Cava picks one up and muses: There’s a lifetime--ten lifetimes--of things to try that nobody has tried before.