News & Views | Published:


Putting the squeeze on lithium

At low temperature, some elements are superconductors under normal pressure. Others become superconducting if the pressure is raised. Lithium is the latest low-temperature, high-pressure superconductor.

The report by Shimizu et al.1 (page 597 of this issue) of superconductivity in lithium, a light metal in group I of the periodic table, is noteworthy for being at odds with an old historical imperative in superconductivity. This remarkable phenomenon was discovered in 1911, and over the following 40 years a considerable database of superconductors was amassed that included both elements and compounds. Its scope was sufficient to invite a search for signs of systematics, and in 1955 Bernd Matthias2 summarized the situation with a plot of the superconductors' transition — or superconducting — temperature as a function of the average number of valence electrons. One of its more dispiriting features was the apparently resolute exclusion of the putatively 'simple' metals of valence 1.

The half-century since then has seen a veritable avalanche of new superconductors identified, including the extraordinary but still somewhat mystifying class of high-temperature cuprate superconductors, and it became clear that attempts to find systematic effects would require far more searching assessments than that of Matthias. Nevertheless, suspicions that valence-1 elements would not be part of the unfolding story of superconductivity seem to have lingered on.

This past half-century has produced another significant development, namely the relentless drive to subject elements and compounds to ever-increasing static compression, in some cases approaching an order of magnitude above normal densities. Among other achievements, these pressures have succeeded in turning non-superconductors (and even non-conductors) into superconductors. The success of Shimizu et al.1 now brings the tally of pressure-induced superconducting elements to 23 (Fig. 1). This is a very considerable achievement considering that at ordinary pressures the number is just 29.

Figure 1: Superconductors under pressure.

The colour code of this periodic table (adapted from ref. 13) shows elements that superconduct under normal, atmospheric pressure conditions (purple) and those that superconduct when subjected to high pressure (orange). Shimizu et al.1 confirm the superconductivity of lithium at high pressure, bringing the number of such elements to 23. Under normal pressure conditions, 29 elements are superconductors.

The changes wrought by pressure are often striking. To make the point, if conditions of one million atmospheres or so were normal on Earth, some characteristic features of the periodic table would look very different from the present one-atmosphere construct. In particular, sulphur, in group VI, would be a metal, and it would also be an exceptional elemental superconductor, with a transition temperature reaching 17 K (ref. 3). Once again this is a case of historic note, for valence-6 elements were also excluded in Matthias's early one-atmosphere systematization.

So now we have the important diamond-cell experiment on compressed lithium, a group-I element, which Shimizu and colleagues1 find to be superconducting. At pressures as high as 48 gigapascals (roughly 480,000 atmospheres), the transition temperature rises to 20 K. Notable though it is, the discovery is not entirely unheralded, because in 1986, with a quite different apparatus, Lin and Dunn4 reported resistance changes in lithium at high pressures and low temperatures, and in interpreting their results they did not rule out the onset of a superconducting state.

Lin and Dunn apparently had no easy way of applying a magnetic field in their experiment. But, for many workers in this area, a sine qua non for establishing a new superconductor is the observation of the Meissner–Ochsenfeld (MO) effect — when an element or compound enters the superconducting state, magnetic flux that previously permeated the sample is suddenly excluded. Also, in certain superconductors, applying a magnetic field of sufficient magnitude can suppress superconductivity, and this is exactly the effect seen by Shimizu and colleagues. Although they also hold observation of the MO effect to be “indispensable” and were unable to achieve this, their demonstration of this suppression is nevertheless an important pointer.

Another indication of the onset of superconductivity is a drop in electrical resistance, which Shimizu et al.1 observed. This, however, is a slightly indeterminate procedure, as can be seen from the authors' data, and future experiments may well lead to some minor revisions of the transition temperature they have measured. Nevertheless, the evidence seems compelling that a superconducting state in lithium has been attained. As the authors note, the reported transition temperature makes lithium (for now) the element with the highest superconducting transition temperature — with, so to speak, not a d-electron in sight.

Although it may be regarded as a simple element in atomic terms, in structural terms lithium in its condensed state at high pressures is complex. It seems to be the case5 (and indeed it was predicted6) that, under pressure, lithium takes on structures quite unlike the familiar body-centred-cubic phase, including structures that have an even number of atoms in the unit cell. How this structural complexity connects with the theory of superconductivity is an intriguing question, for perhaps one of the most difficult theoretical problems in the field of condensed-matter physics is accurately predicting superconducting transition temperatures. Nevertheless, general arguments (based on the density of states, the strength of the electron–phonon interaction, the scale of the Coulomb pseudopotential, and so on) can bolster the expectation that a state of significant superconductivity will occur, and did in fact make a strong case for lithium6.

But calculations7 of lithium's transition temperature seem to come up with a value that is a factor of four or so above Shimizu and colleagues' measurement. In a way, this discrepancy could be seen as a propitious opportunity — the pressure experiments on lithium and on other light-element systems may at last offer the prospect for systematic studies of the behaviour of the dynamic electron–electron interaction and in particular of the Coulomb pseudopotential, the manifestation of direct electron–electron repulsion that generally works against electron pairing (unlike the electron–phonon interaction, which in lithium probably promotes it).

It may be that the early Matthias rule should simply be rephrased to read that single-band, rather than single-valence, elements are in general not favoured for superconductivity, for there is accumulating evidence that structures resulting in multiple bands are predominant among superconductors. If structural anisotropy in layered structures is a route to higher-temperature superconductors, as Ginzburg and Kirzhnits have suggested8 and as seems to be the case with the superconducting latecomer MgB2 (whose transition temperature is 38 K), then we could consider going further, for example to near-linear structures such as LiBx (ref. 9) at high pressure. Under pressure, boron is also an excellent superconductor in its own right10.

In a curious historical throwback, a lithium compound was among the class of systems that spawned the idea that electron pairs might be associated with superconductivity. This was Ogg's proposal11, in 1946, based on his studies of metal amines at low temperatures. So perhaps the lithium amine Li(NH3)4 should be squeezed as well. And the possibility of dropping down two further notches in the periodic table, to hydrogen, seems to be back in contention in the quest for higher-temperature superconductors12, as Shimizu et al. cogently observe.


  1. 1

    Shimizu, K., Ishikawa, H., Takao, D., Yagi, T. & Amaya, K. Nature 419, 597–599 (2002).

  2. 2

    Matthias, B. T. Phys. Rev. 97, 74–76 (1955).

  3. 3

    Struzhkin, V. V., Hemley, R. J., Mao, H.-K & Timofeev, Y. A. Nature 390, 382–384 (1997).

  4. 4

    Lin, T. H. & Dunn, K. J. Phys. Rev. B 33, 807–811 (1986).

  5. 5

    Hanfland, M., Syassen, K., Christensen, N. E. & Novikov, D. L. Nature 408, 174–178 (2000).

  6. 6

    Neaton, J. B. & Ashcroft, N. W. Nature 400, 141–144 (1999).

  7. 7

    Christensen, N. E. & Novikov, D. L. Phys. Rev. Lett. 86, 1861–1864 (2001).

  8. 8

    Ginzburg, V. L. & Kirzhnits, D. A. High-Temperature Superconductivity (Consultants Bureau, New York, 1982).

  9. 9

    Wörle, M. & Nesper, R. Angew. Chem. 112, 2439–2443 (2000).

  10. 10

    Eremets, M. I. et al. Science 293, 272–274 (2001).

  11. 11

    Ogg, R. A. Phys. Rev. 69, 243–244 (1946).

  12. 12

    Ashcroft, N. W. Phys. Rev. Lett. 21, 1748–1749 (1968).

  13. 13

    Hemley, R. J. & Mao, H.-K in Proceedings of the International School of Physics 'Enrico Fermi': High Pressure Phenomena (eds Hemley, R. J., Bernasconi, M., Ulivi, L. & Chiarotti, G.) (Italian Phys. Soc., Bologna, in the press).

Download references

Author information

Correspondence to N. W. Ashcroft.

Rights and permissions

Reprints and Permissions

About this article

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.