Break-junction tunneling on MgB2

https://doi.org/10.1016/S0921-4534(02)02317-1Get rights and content

Abstract

Tunneling data on magnesium diboride, MgB2, are reviewed with a particular focus on superconductor–insulator–superconductor (SIS) junctions formed by a break-junction method. The collective tunneling literature reveals two distinct energy scales, a large gap, ΔL∼7.2 meV, close to the expected BCS value, and a small gap, ΔS∼2.4 meV. The SIS break junctions show clearly that the small gap closes near the bulk critical temperature, Tc=39 K. The SIS spectra allow proximity effects to be ruled out as the cause for the small gap and therefore make a strong case that MgB2 is a coupled, two-band superconductor. While the break junctions sometimes reveal parallel contributions to the conductance from both bands, it is more often found that ΔS dominates the spectra. In these cases, a subtle feature is observed near ΔS+ΔL that is reminiscent of strong-coupling effects. This feature is consistent with quasiparticle scattering contributions to the interband coupling which provides an important insight into the nature of two-band superconductivity in MgB2.

Introduction

Approximately one and a half years after the discovery [1] of superconductivity in MgB2 a wealth of information about its properties has been collected. While its high critical temperature and simple crystal structure generated an initial flurry of activity, there is now an increasing interest in the nature of superconductivity in MgB2, as it appears to be one of the rare examples of a two-band superconductor.

Theoretically, such two-band behavior was proposed for MgB2 because the electronic system consists of two qualitatively different types of charge carriers, derived from boron π- and σ-bands, respectively [2]. The π-bands are three dimensional (3D), while the σ-bands are effectively restricted to two dimensions (2D). For convenience, these actual four bands are henceforward treated as two effective bands. Superconductivity is proposed to arise from electron–phonon coupling of the 2D band with a specific boron bond-stretching mode. That implies, that superconductivity originates from the σ-band and that superconductivity in the π-band is driven by that primary interaction. As a result, the properties of superconductivity are proposed to be different in both bands, provided the material is in the clean limit and the electronic systems do not strongly mix. The 2D band shows a large gap, ΔL∼7.2 meV, whereas the 3D band shows a small gap, ΔS∼2.4 meV, both closing at a joint critical temperature, Tc. In the dirty limit, only one gap of intermediate magnitude is expected to be observed, closing at a reduced Tc.

There now is sufficient experimental evidence to strongly support such a two-band scenario, and there is a growing consensus in the community to this end. However, a large fraction of these experiments involved tunneling spectroscopy which is susceptible to surface imperfections. Specifically, surface proximity effects can mimic the effects of two-band superconductivity. In this paper, we will present an overview of tunneling results, and will use this together with our own experimental data to argue in favor of the two-band nature of superconductivity in MgB2.

Section snippets

Techniques of tunneling spectroscopy

Tunneling spectroscopy traditionally presents one of the most direct probes of the superconducting energy gap [3]. In superconductor–insulator–normal metal (SIN) tunnel junctions, the conductance at low temperature is equivalent to the density of states (DOS) near the Fermi energy, EF. At finite temperature, the conductance may be calculated from:dIdV=−σN∫N(E)∂f(E+eV)∂eVdE,where N is the superconducting DOS and σN the conductance of the junction in the normal state. To account for

Experimental

A tremendous amount of tunneling work has been performed on MgB2, and even though such a narrow view certainly does not mirror the full depth of information obtained from these studies, we for now want to focus only on the inferred energy gap values. Table 1 gives a compilation of these values, and it immediately becomes evident that most studies agree with the presence of two gaps around ΔS=2.0–2.8 meV and ΔL=7.0–7.5 meV. This observation of two gaps, as well as their values, are in nice

Temperature dependence of ΔS

The first unexpected result from tunneling spectroscopy on MgB2 was the observation of a small energy gap, ΔS=2.0 meV [8], much smaller than the weak coupling limit would allow for Tc=39 K. The simplest scenario to result in a reduced gap value is a surface layer of reduced superconductivity with a gap anywhere in between the bulk value and zero. This scenario can positively be ruled out by studying the temperature dependence of the small gap, since such a reduced gap should still scale with

Spectroscopy of ΔL

Such current contributions from the 2D band are expected to show up at eV=ΔL in SIN junctions and at ΔS+ΔL and 2ΔL in SIS junctions. However, it is necessary to very carefully analyze whatever extra features are observed. Both the spectra in Fig. 6 and Fig. 7a, respectively, show additional features at bias close to ΔS+ΔL, yet the shape of these features is entirely different. Whereas in the first case a distinct peak is observed, the second spectra displays a dip at about the same position.

The

Two-band fitting of SIS spectra

Most tunneling work that determined ΔL, or at least found evidence for its presence, did so by studying direct tunneling contributions from the 2D band in SIN junctions (double-peak structures). In the previous section we illustrated how such contributions do, even though rarely, influence SIS junctions. However, the more frequently seen extra feature in SIS junctions is entirely different in shape and origin. Fig. 7 gives an example for a well-developed dip near ΔS+ΔL, in contrast to the peak

Ruling out proximity effects

However, the applicability of this model a priori does not prove two-band superconductivity to be the origin of the dip-feature, as McMillan’s model originally was developed to describe the proximity effect, which therefore needs to be considered as a possible cause. In the proximity effect, a thin surface layer of reduced (or no) intrinsic superconductivity is influenced by an adjacent superconductor to show enhanced superconductivity (or some superconductivity at all). McMillan’s equations

Conclusion

We have reproducibly observed a small gap, ΔS=2.5 meV, in break-junction tunneling on MgB2 and traced it to high temperatures, where it closes near the bulk Tc. Only in rare cases, we also observe direct tunneling contributions showing a large gap, ΔL=7.6 meV. These findings give evidence for, and are interpreted in terms of a two-band model. A commonly observed dip feature at ΔS+ΔL is analyzed using a specific two-band model. We argued, that this feature gives evidence for quasiparticle

Acknowledgments

We are indebted to B. Jankó and C.P. Moca for valuable discussions and sharing unpublished results from their two-band calculations. We want to thank K. Scharnberg for allowing us to use Fig. 8a. We further want to thank M.A. Belogolovskii, G. Burnell, D.J. Kang, J.-I. Kye, D. Mijatovic, A. Saito, and K. Ueda for sending their manuscripts prior to publication. This research is supported by the US Department of Energy, Basic Energy Sciences––Materials Sciences, under contract no. W-31-109-ENG-38.

References (69)

  • H. Suderow

    Physica C

    (2002)
  • S. Lee

    Physica C

    (2002)
  • R.S. Gonnelli

    J. Phys. Chem. Solids

    (2002)
  • T. Takasaki

    Physica C

    (2002)
  • Z.Z. Li

    Physica C

    (2002)
  • L. Ozyuzer et al.

    Cryogenics

    (1998)
  • N. Schopohl et al.

    Solid State Commun.

    (1977)
  • J. Nagamatsu

    Nature

    (2001)
  • A.Y. Liu

    Phys. Rev. Lett.

    (2001)
  • I. Giaever

    Phys. Rev. Lett.

    (1960)
  • R.C. Dynes et al.

    Phys. Rev. Lett.

    (1978)
  • A.F. Andreev

    Zh. Éksp. Teor. Fiz.

    (1964)
  • G.E. Blonder et al.

    Phys. Rev. B

    (1982)
  • A. Plecenı́k

    Phys. Rev. B

    (1994)
  • G. Rubio-Bollinger

    Phys. Rev. Lett.

    (2001)
  • A. Sharoni

    J. Phys.: Cond. Mat.

    (2001)
  • A. Sharoni

    Phys. Rev. B

    (2001)
  • G. Karapetrov

    Phys. Rev. Lett.

    (2001)
  • G. Karapetrov et al., in: A.V. Narlikar (Ed.), Studies of High Temperature Superconductors, vol. 38, Nova Sci. Publ.,...
  • M. Iavarone

    Phys. Rev. Lett.

    (2002)
  • R.J. Olsson et al., unpublished...
  • F. Giubileo

    Phys. Rev. Lett.

    (2001)
  • F. Giubileo

    Europhys. Lett.

    (2002)
  • F. Giubileo

    Int. J. Mod. Phys. B

    (2002)
  • J. Karpinski et al., unpublished...
  • M.R. Eskildsen

    Phys. Rev. Lett.

    (2002)
  • P. Seneor

    Phys. Rev. B

    (2001)
  • A. Plecenik

    Physica C

    (2001)
  • F. Laube

    Europhys. Lett.

    (2001)
  • A.I. D’yachenko et al., unpublished...
  • Z.Z. Li

    Supercond. Sci. Technol.

    (2001)
  • Z.-Z. Li

    Phys. Rev. B

    (2002)
  • P. Szabó

    Phys. Rev. Lett.

    (2001)
  • Yu.G. Naidyuk

    Pis. Zh. Éksp. Teor. Fiz.

    (2002)

    • Cited by (0)

    View full text
    Copyright © 2002 Elsevier Science B.V. Published by Elsevier B.V. All rights reserved.