Accueil > English > News

Why do some “bad metals” become “good superconductors” ?


Surprisingly, the best superconductors discovered to date are often not very good metals. For physicists, a “bad metal” is a complex object that is not well described by current theories. We have used Angle Resolved Photoemission to reveal some anomalies of one of this bad metallic state, belonging to the family of iron superconductors, namely the formation of a “pseudogap” in the electronic structure.

In the newly discovered iron superconductors, the origin of superconductivity still raises many questions. There is a close proximity between superconducting and magnetic phases. This reminds the situation in another family of high temperature superconductors, cuprates, and suggests that, instead of destroying superconductivity, as it usually does, magnetism may help in certain cases to reach superconductivity at quite high temperatures. Understanding this in more details may help to create new types of superconductors, possibly at even higher temperatures.

To go further, it is necessary to understand better the nature of magnetism itself and the role of magnetic correlations in the metallic state. In cuprates, magnetism sets in within an insulating background, with one electron localized at each copper site. They give rise to localized magnetic moments that acquire an antiferromagnetic order. The Coulomb repulsion is so strong in these systems that it would cost too much energy to have two electrons on the same site and the system chooses instead to localize one electron per site (this state is called the Mott insulator). In iron superconductors, the situation is quite different. Magnetism sets in within a metallic background and there are six electrons per site. However, some of these phases are “bad metals” meaning it becomes very difficult for electrons to move from site to site. Why would “bad metals” become “good superconductors” ? This apparent contradiction may be at the heart of the formation of high temperature superconductivity.

Figure : (a) Energy vs momentum plot of one electron band in Fe1.06Te for different temperatures measured with ARPES. (b) ARPES spectrum at the Fermi vector kF (the momentum value k where the peak is closest from EF) for different temperatures. The thin line is symmetrized with respect to the Fermi level. The maximum should be at EF if there were no pseudogap (as it is at 20 K). (c) Value of the pseudogap as a function of temperature.

 

In a recent study, we investigated how this bad metallic behavior manifests itself in the electronic structure of Fe1.06Te. We imaged the electronic structure with angle resolved photoemission spectroscopy at the CASSIOPEE beamline of the SOLEIL synchrotron, both in the metallic paramagnetic phase (i.e. magnetically disordered, which occurs for temperatures T above 76 K) and in the magnetically ordered phase (T<76 K). Fig. a shows the evolution of one electronic band in this compound. The Femi level (the occupied state having the highest energy) is shown by the white line and the red intensity corresponds to states occupied with electrons. Normally, one gets a metal if there are bands crossing the Fermi level, meaning there are partially filled bands where electrons are free to move. This is the case for low temperatures : at 20 K, in Fig. a4, we see electrons up to the Fermi level and the spectra taken at this point (Fig. b) shows a “Fermi step” characteristic of the metal. At high temperatures, the intensity is small at the Fermi level and the peak in Fig. b moves away from it.

This situation can be called “pseudogap” as it is intermediate between the one expected for a metal (no gap) and an insulator (full gap). The fact that a good metallic state is recovered in the magnetic state is a very direct indication of the role of magnetic disorder to create the bad metallic state. The idea is that there is a local tendency at each Fe site to align the spins of the electrons in different orbitals and form a local magnetic moment. In the magnetic phase, these moments order in a way optimizing electronic conduction in certain directions and magnetic fluctuations are frozen. In the paramagnetic phase, the disordered moments interact with conducting electrons and hinder their motion. There is then in the heart of the paramagnetic metallic phase, strong magnetic correlations between electrons that impact the nature of the metallic state. How it creates a pseudogap remains to be understood. These correlations appear quite different from those present in the cuprates, which are mostly based on Coulomb repulsion. On the other hand, the metallic phase of the cuprates is also known to harbor a “pseudogap”. Whether there is any connection between these two new states of matter is an interesting challenge.

 

Reference :

Nature of the Bad Metallic Behavior of Fe1.06Te Inferred from Its Evolution in the Magnetic State
P.-H. Lin, Y. Texier, A. Taleb-Ibrahimi, P. Le Fèvre, F. Bertran, E. Giannini, M. Grioni, and V. Brouet
Physical Review Letters 111, 217002 (2013).

Corresponding author : Véronique Brouet