A century on from the discovery of superconductivity, we still do not know how to design superconductors that can be really useful in the everyday world. Despite this seemingly downbeat statement, I remain enthusiastic about the search for new superconducting materials. Although my own research in this area has had its share of null results and knock-backs, in that I am in good company with the true leaders in the field. Optimism abounds, and the past couple of years have seen a renewed passion, with researchers worldwide wanting to work together to find a way to design new materials that we know in advance will function as superconductors.

That would be very different from most of the discoveries in superconductivity, which have often been serendipitous. Indeed, the main quest of Heike Kamerlingh Onnes was to liquefy gases, and only after managing to liquefy helium in 1908 did he set his Leiden lab to work on a study of the properties of metals at low temperature. The choice of sample was fortunate – mercury was used because it is a liquid at ambient temperature and so could easily be purified. The discovery of its dramatic drop in resistance when cooled to 4 K, which we now know to be the critical temperature, T_c, was an unexpected and fortuitous surprise.

In subsequent years, increasing the critical temperature was achieved by systematic experimental tests of elements, alloys and compounds, predominantly led by Bernd Matthias from about 1950, who in doing so became the first researcher to discover a new class of superconductors. To begin with, the only known superconductors were elements, but Matthias found superconductivity in various combinations of elements that on their own are non-superconducting. The earliest of these was the ferromagnetic element cobalt combined with the semiconductor silicon to form CoSi₂. What changed the game was the discovery by John Hulm and his graduate student George Hardy at the University of Chicago in 1952 of the vanadium–silicon compound V₃Si, the first of the then-called high-T_c superconductors. This was a completely new class of superconductors – known as the A15s (a particular crystal structure of the chemical formula A₃B, where A is a transition metal) – and it enabled Matthias to discover more than 30 compounds of this type, with values of T_c that ranged up to 18 K in the case of Nb₃Ge.

Increasing the critical superconducting temperature is certainly what most interests the media, but it is not the only property with which to rank new superconductors. The A15s were the first family of superconductors that maintained a high critical current density, J_c, in the presence of strong magnetic fields, which is crucial for all current-carrying applications. In 1963 Hulm, then with co-workers at the Westinghouse Research Laboratories, developed the first commercial superconducting wires, based on random alloys of niobium–titanium, a material discovered at the Rutherford Appleton Laboratory in the UK. Although niobium–titanium alloys exhibit a lower T_c and J_c than the A15s, they were chosen for wires because they are malleable, reliable and can be used in nearly all practical applications, including the medical technique of magnetic resonance imaging (MRI). Despite the importance of a high J_c, achievements in this area receive little recognition compared with progress in increasing T_c. But as my former boss, John Rowell, stated at the retirement party of Jack Wernick, who is noted for the discovery of several A15s, “High-T_c wins Nobel prizes; high J_c saves lives.” So although the search for new families of higher-T_c superconductors is what makes the headlines, what really matters when it comes to applications is a high value of J_c and mechanical properties that are good for making wires.

Figure 1

In 1979 Frank Steglish and colleagues discovered superconductivity in materials containing rare earths (elements with a 4f electron orbital) or actinides (those with a 5f electron orbital). These compounds are called the “heavy fermions”, because with antiferromagnetic ground states, and at low temperatures, the itinerant electrons behave as if they have masses up to 1000 times larger than the free-electron mass. This discovery was significant because the heavy fermions were the first truly tunable superconductors, through a competition between superconductivity and magnetic order. But what was even more important was that heavy-fermion superconductors did not follow the rule book: for the first time, the brilliant Bardeen–Cooper–Schrieffer (BCS) theory of superconductivity was shown to break down. BCS theory explains what is happening at the microscopic level – it involves paired electrons known as “Cooper pairs” travelling around the crystal lattice – and this part of the theory remains robust in all the known superconductors. But the microscopic mechanism for superconductivity in all previously found superconductors was attributed in BCS to electron–phonon coupling, which was not sufficient to cause the electron pairing in the new heavy-fermion superconductors (figure 1).

Before the heavy fermions were discovered, it was accepted that any kind of magnetism would harm the superconducting state. But in this new class of superconductor the magnetism appeared integral to the strength of the superconductivity. Another exciting aspect of this class is that higher-T_c heavy-fermion superconductors – in particular the “115” series beginning with the discovery of CeCoIn₅ – were not discovered purely by serendipity, but driven by guidelines learned from many preceding substitution and pressure studies.

New classes

Enter the high-T_c oxides. First was the sensational revolution of the copper oxides, or “cuprates”: Georg Bednorz and Alex Müller discovered LaBaCuO in 1986 with a T_c of 40 K, and subsequently Maw-Kuen Wu and Ching-Wu (Paul) Chu discovered YBa₂Cu₃O_7–d, or “YBCO”, with a T_c of more than 90 K. (For more about the high-T_c revolution, see “Resistance is futile” on page 33, print edition.) These transformative discoveries again relied on guidelines put together by thoughtful and talented physicists, but serendipity certainly played a factor. Indeed, I believe the only discovery of a high-T_c system that was driven predominantly by theory is Ba_1–dK_dBiO₃, or BKBO (to date at least): Len Mattheiss and Don Hamann at Bell Labs used electronic-structure calculations of an earlier low-T_c system, Ba(Pb,Bi)O₃, to predict and then make BKBO, for which their colleague Bob Cava drove the T_c to a respectable 30 K.

But what of materials with even higher transition temperatures? Through a tremendous amount of hard work worldwide by many talented physicists, transition temperatures in the cuprates have been pushed up to 135 K at ambient pressure and above 150 K at high pressure in HgBa₂Ca₂Cu₃O_8+d (also known as Hg-1223), which was discovered in 1993. We were then left with the idea that perhaps there were no other families of high-T_c superconductors. Could it be that the cuprate were the only high-T_c class we would ever find? The fear was that systematic studies had already found the highest possible T_c.

Figure 2

But we had guidelines and ideas. Many of these were published in a 2006 report for the US Department of Energy, Basic Research Needs for Superconductivity. Particularly of note in that report, which outlined the prospects and potential of superconductivity, was our canonical phase diagram (figure 2), which hinted that we knew where to look: at the boundary between competing phases. This personifies the concept of “quantum criticality”, where a phase transition occurs not because of thermal fluctuations as in a typical thermodynamic phase transition, but because of quantum-mechanical fluctuations at zero temperature. The phase diagram shows an antiferromagnetic insulator on the left and a normal metal on the right. Where they meet at the centre is the quantum critical point, and as that point is approached, the quantum fluctuations of the competing phases get stronger and a strange “emergent” state of matter appears – in this case, high-temperature superconductivity. The general rule was: the stronger the competing phases, the stronger the emergent phase. Those ideas remain, but where were these new families of superconductors? Had we hit a dead end?

Finally, in 2008, a second class of high-T_c superconductor was discovered. Hideo Hosono at the Tokyo Institute of Technology had discovered iron-based superconductors two years earlier, and in January 2008 his first “high-T_c” paper on these materials was published, which precipitated a renewed excitement and a frenzy of activity. Within four months, Zhongxian Zhao’s group at the Institute of Physics in Beijing created related materials that hold the record with a T_c of 58 K. Many of us were awestruck – here finally was a new class of high-temperature superconductors that broke the 22-year tyranny of cuprates, and in materials that no-one had predicted and were contrary to our basic notions of how superconductivity works. How could iron – the strongest ferromagnetic element in the periodic table – be a basis for superconductivity at all, let alone high-temperature superconductivity? There now exist whole arrays of iron-based superconductors – pnictides and chalcogenides – all found by clever, hard work, but originally discovered by serendipity.

Laying down the gauntlet

All of these families of superconductors have a great deal in common, yet also have unique properties. The physics seems to be growing more complex with time, and we continue to build more guidelines and structure into our search for new superconducting materials. Although the discovery of iron-based superconductors gave us a lot of research fodder, they will not necessarily tell us all we need to know about how to find new classes of superconductors. But one thing is for sure: the cuprates are not unique and as there is a second class of high-T_c superconductors, I believe there must be a third.

The discovery of iron-based superconductors – the first new class of high-T_c superconductors after more than two decades of only incremental progress – injected a new-found positivity into the field, rivalled only by the discovery of superconductivity in the cuprates. The resulting surge of global research, however, has a very different feel from that in 1986. In the early days of high-temperature superconductivity the competition was fierce – there was a real race to obtain higher transition temperatures. But now that zealous sense of urgency has been replaced by a more paced and considered approach.

Many scientists have been working on understanding novel superconductors for decades, often in productive collaborations. Recently, our research funding and support have been revitalized on a worldwide scale, in part because of the need to address the global energy crisis by significantly increasing the efficiency of power transmission. After 25 years of intense and fruitful work, the cuprates remain promising, but for various reasons may still not be the materials of choice to impact our power grid. The newly discovered iron-based high-temperature superconductors exhibit many positive aspects, but are likewise not yet in a position to impact power transmission. Another class of superconductors is needed.

For any one of us, putting all of our efforts towards attacking this problem of discovering a new superconductor is highly risky. If we want to find such a thing but do not manage this after three to five years – the typical length of most research grants – we seriously risk losing our funding. As a result, we focus most of our efforts on understanding the existing novel superconductors. So, I and my colleague Rick Greene (no relation) of the University of Maryland, aided by the Institute for Complex Adaptive Matter, have made a call to arms to the international community, which we are spreading via working groups at conferences and workshops: “It is time for us to join our expertise and resources together, on a worldwide scale, to search for that new class of superconductors.”

The gauntlet is being taken up with enthusiasm. With communication now flowing between different groups, and across funding and geographical barriers, we hope to soon reveal at last a clarified vision of high-temperature and novel superconductivity that will set us in the best possible stead in the quest for a new class.

It is gratifying to see superconductivity, at 100, finally growing up.