A range of techniques have been used to show that the behaviour of the superconductivity in LiFeAs is quite different from that in the other iron arsenide superconductors, and the system shows some similarities with the FeSe system.
LiFeAs1–3 is a member of a large family of iron-based superconductors which has received a large amount of attention since the discovery of high temperature superconductivity in LaFeAsO1–xFx. A wide range of compounds containing antifluorite-type FeAs or FeSe layers exhibit superconductivity at temperatures up to 50 K. The high critical temperatures lie beyond the limits of the Bardeen-Cooper-Schrieffer (BCS) theory of superconductivity and the precise origin of the superconductivity in these materials is the subject of much debate. The compounds have a formal iron oxidation state close to +2 and in stoichiometric phases such as LaFeAsO, BaFe2As2 and NaFeAs, semimetallic properties are exhibited at ambient temperatures and there is a transition to an itinerant antiferromagnetically ordered state at low temperatures which drives a tetragonal-to-orthorhombic structural distortion. In order to realise superconducting properties it is normally required that chemical substitution be carried out or that the stoichiometric compounds be subjected to elevated pressures. Derivatives of BaFe2As2 have received much attention because of the relative ease with which single crystals can be obtained and there are two chemical substitution strategies which lead to suppression of magnetic order and the emergence of superconductivity. Formal “hole” doping in the series Ba1–xKxFe2As2 yields optimal superconducting properties with Tc ~ 38 K with x ~ 0.4. This type of substitution is analogous to that used to drive layered cuprates from the antiferromagnetic insulating state to the superconducting metallic state. In contrast to the cuprates however, direct substitution of Fe by a neighbouring transition metal, especially Co, Ni or Cu, is also found to result in superconductivity in for example BaFe2–xCoxAs2. Presumably in this case the magnetic order is suppressed by introducing disorder and the superconducting ground state becomes relatively more stable. In this series there is mounting evidence that the transition from magnetic order to superconductivity as a function of Co content proceeds via an unusual region in which each Fe atom is involved in both magnetic order and in superconductivity. In the entire class of iron-based superconductors there is also evidence that compounds with regular FeAs4 tetrahedra generally tend to show optimal superconducting properties.
LiFeAs with the anti-PbFCl structure type was first synthesised in 1968.1 Superficially it resembles the other members of the iron-based superconducting series, but there are some differences in detail. The first is that because of the small size of the Li+ ion, the FeAs4 tetrahedra are very far from regular, being highly compressed in the basal direction. The second is that LiFeAs is a superconductor (Tc ~ 17 K)2–3 when stoichiometric with Fe in the formal oxidation state of +2. There is no evidence for long range magnetic order in LiFeAs. The structural and superconducting properties of LiFeAs resemble quite closely those of FeSe4.
Figure 1: Rietveld refinement of the structure of Li0.98Fe1.02As against Ill data (top) and the correlation between Tc derived from magnetic susceptibility measurements and unit cell volume for Li1-xFe1+xAs.
We have probed the behaviour of LiFeAs and its derivatives using neutron and synchrotron X-ray powder diffraction and muon-spin rotation spectroscopy (mSR) at Diamond Light Source, the ISIS facility and the Paul Scherrer institute as well as using SQUID magnetometry. Experiments performed on I11 have shown that there is an extremely sharp dependence of the superconducting properties on composition. It was reported in the original literature1 that a series of compounds could be made which were simultaneously slightly poor in lithium and slightly rich in iron. We used the high quality data available on I11 (Fig. 1) to determine the trends in lattice parameters over a very narrow range of compositions for the series Li1–xFe1+xAs with x = 0.04. These data were used together with high resolution neutron powder diffraction data from HRPD at ISIS to calibrate the unit cell volume against the refined composition. Measurement of the magnetic superconducting properties of these compounds using SQUID magnetometry showed that the superconductivity is very rapidly suppressed as the amount of Fe located on the Li sites between the superconducting FeAs layers increases. The critical concentration of interlayer Fe, beyond which superconductivity is completely suppressed by these paramagnetic impurities is x ~ 0.02. The parallels between LiFeAs and FeSe thus persist in their sensitivity to these interlayer dopants – the superconductivity in FeSe is similarly susceptible to suppression by the presence of interlayer Fe ions4.
Cobalt and Nickel doping on the iron site in LiFeAs also serve to suppress superconductivity (Fig. 2), and so far no chemical substitutions in LiFeAs have been found to increase Tc. We have probed the properties of the superconducting state in LiFe1–xCoxAs and LiFe1–xNixAs compounds using µSR. This technique offers a direct measure of the London penetration depth in a superconductor which is related to the superfluid stiffness. Our measurements reveal that in LiFeAs derivatives the superfluid is stiffer than expected for a given Tc (or Tc is lower than expected for a given superfluid stiffness) based on comparison with the other members of the iron arsenide family. The range of behaviour exhibited by the iron arsenides in a plot of Tc against superfluid stiffness (Uemura plot - Fig. 3) is similar to that exhibited by electron and hole-doped cuprate superconductors.
Figure 2: Evolution of structural parameters (left) and magnetic susceptibility with Co and Ni doping in LiFeAs (right).
In summary our measurements show that the superconductivity in stoichiometric LiFeAs is very rapidly suppressed either by the incorporation of small amounts of transition metal ions between the FeAs layers, or by the substitution of Fe by a later transition metal. Since the superconductivity is also suppressed by the application of hydrostatic pressure5, LiFeAs seems to exhibit its optimal properties when it is stoichiometric and at ambient pressure. Furthermore the relationship between penetration depth and Tc differs from that exhibited by the other iron arsenides. The reason for the difference between the behaviour of LiFeAs and its derivatives and the other members of the wider series of iron arsenide compounds is likely a consequence of the more highly compressed FeAs4 tetrahedra and the associated short Fe–Fe distance in LiFeAs. But a great deal of further work is required before the relationship between magnetism and superconductivity and the interplay of electron count, structural parameters and disorder on the competition between the magnetic ordering and superconductivity in this class of compounds becomes clear.
Figure 3: Uemura plot of superfluid stiffness (?s) against Tc for LiFe1-xMxAs, where M = Co, Ni. In LiFeAs derivatives Tc is lower than expected for a given superfluid stiffness than the other iron arsenide superconductors. The diversity in the iron arsenides is similar to that exhibited by the layered cuprates.
Pitcher, M. J., Lancaster, T., Wright, J. D., Franke, I., Steele, A. J., Baker, P. J., Pratt, F. L., Trevelyan Thomas, W., Parker, D. R., Blundell, S. J.& Clarke, S. J. Compositional control of the superconducting properties of LiFeAs. J. Am. Chem. Soc. 132, 10467–10476 (2010)
References
- Juza, R. & Langer, K. Z. Anorg. Allg. Chem. 361, 5 (1968).
- Pitcher, M. J. et al. Chem. Commun. 5918 (2008).
- Tapp, J. H. et al. . Phys. Rev. B. 78, 060505 (2008).
- McQueen, T. M. et al. Phys. Rev. B. 79, 014522 (2009).
- Mito, M. et al. . J. Am. Chem. Soc. 131, 2986 (2009).
Funding Acknowledgement
We acknowledge the UK Engineering and Physical Sciences Research Council for financial support, and Diamond Light Source, ISIS facility, ESRF and Paul Scherrer Institute for the award of beamtime.