Single crystals of LnFeAsO_1−xF_x (Ln = La, Pr, Nd, Sm, Gd) and Ba_1−xRb_xFe₂As₂: Growth, structure and superconducting properties

Abstract

A review of our investigations on single crystals of LnFeAsO_1−xF_x (Ln = La, Pr, Nd, Sm, Gd) and Ba_1−xRb_xFe₂As₂ is presented. A high-pressure technique has been applied for the growth of LnFeAsO_1−xF_x crystals, while Ba_1−xRb_xFe₂As₂ crystals were grown using a quartz ampoule method. Single crystals were used for electrical transport, structure, magnetic torque and spectroscopic studies. Investigations of the crystal structure confirmed high structural perfection and show incomplete occupation of the (O, F) position in superconducting LnFeAsO_1−xF_x crystals. Resistivity measurements on LnFeAsO_1−xF_x crystals show a significant broadening of the transition in high magnetic fields, whereas the resistive transition in Ba_1−xRb_xFe₂As₂ simply shifts to lower temperature. The critical current density for both compounds is relatively high and exceeds 2 × 10⁹ A/m² at 15 K in 7 T. The anisotropy of magnetic penetration depth, measured on LnFeAsO_1−xF_x crystals by torque magnetometry is temperature dependent and apparently larger than the anisotropy of the upper critical field. Ba_1−xRb_xFe₂As₂ crystals are electronically significantly less anisotropic. Point-Contact Andreev-Reflection spectroscopy indicates the existence of two energy gaps in LnFeAsO_1−xF_x. Scanning Tunneling Spectroscopy reveals in addition to a superconducting gap, also some feature at high energy (∼20 meV).

Introduction

Since the first report on superconductivity at 26 K in F-doped LaFeAsO at the end of February 2008, the superconducting transition temperature has been quickly raised to about 55 K and several new superconductors of a general formula LnFeAsO_1−xF_x (Ln = La, Ce, Pr, Nd, Sm, Gd, Tb, Dy), abbreviated as Ln1111, have been synthesized [1], [2], [3], [4], [5], [6], [7], [8]. These compounds crystallize with the tetragonal layered ZrCuSiAs structure, in the space group P4/nmm. The structure consists of alternating LnO and FeAs layers, which are electrically charged represented as (LnO)^+δ(FeAs)^−δ (Fig. 1). Covalent bonding is dominant in the layers, while ionic bonding dominates between layers. Electron carriers can be introduced by substituting F for O or by oxygen deficiency [1], [2], [3], [4], [5], [6], [7], [8]. By substituting Sr²⁺ for La³⁺ in La1111, holes are introduced [9].

More recently, superconductivity in AFe₂As₂ (A = Ca, Sr, Ba) (called A122) with ThCr₂Si₂-type structure and maximum T_c = 38 K has been reported [10]. These compounds have a more simple crystal structure in which (Fe₂As₂)-layers, identical to those in Ln1111 are separated by single elemental A layers (Fig. 2). Up to date superconductivity has been found in hole-doped Sr_1−xK_xFe₂As₂ and Sr_1−xCs_xFe₂As₂ [11], Ca_1−xNa_xFe₂As₂ [12], Eu_1−xK_xFe₂As₂ [13], and Eu_1−xNa_xFe₂As₂ [14], as well as in electron-doped Co-substituted BaFe₂As₂ [15] and SrFe₂As₂ [16], and Ni-substituted BaFe₂As₂ [17]. Furthermore, pressure induced superconductivity has been also discovered in the parent compounds CaFe₂As₂ [18], [19], SrFe₂As₂ [20], [21], and BaFe₂As₂ [21].

Besides KFe₂As₂ and CsFe₂As₂, which are superconductors with T_c’s of 3.8 K and 2.6 K [11] respectively, RbFe₂As₂ is known to exist as well [22]. Therefore, it seemed natural to us to explore the BaFe₂As₂–RbFe₂As₂ system in order to search for superconductivity.

It is interesting to explore the important parameters which govern the superconducting properties of new superconductors. In the case of high-T_c cuprates it is the number of carriers doped into the CuO₂ layers. In analogy in the pnictides it is the number of carriers doped into the FeAs layers. There are some similarities between the new pnictide superconductors and the cuprate superconductors due to the layered structure and the fact that both Fe and Cu are 3d-elements. However, there are important differences. First, doping on the Fe site in Sr122 or Ba122 by substitution of Fe by Co leads also to the appearance of superconductivity [15], [16] in contrast to cuprates, where substitution for Cu suppresses superconductivity. Second, in cuprates, introducing of one oxygen atom is equivalent to introducing of two fluorine atoms [23]. In Ln1111 there is a significant difference in carrier doping between oxygen deficiency and fluorine substitution. One expects that one oxygen atom deficiency provides two electrons while substitution of F⁻ for O²⁻ provides one electron. However, according to [24], [25] oxygen deficiency is much less effective as a source of electrons than F-substitution. Structural parameters play also an important role for obtaining high T_c’s. There is dependence between T_c and the As–Fe–As bond angle of the FeAs₄ tetrahedron: maximum T_c is achieved when As–Fe–As bond angle is close to 109.47° corresponding to an ideal tetrahedron [25].

Despite all these differences, the similarities due to the layered crystal structure are important as well. So far, all high-T_c superconductors have a layered crystal structure leading to pronounced anisotropic physical properties. All cuprate superconductors have been characterized by a well-defined effective mass anisotropy parameter γ [26]

$$\gamma =\sqrt{m_c^{\ast }/m_{\mathit{ab}}^{\ast }}={\lambda }_c/{\lambda }_{\mathit{ab}}={\xi }_{\mathit{ab}}/{\xi }_c=H_{c2}^{\Vert \mathit{ab}}/H_{c2}^{\Vert c}\text{,}$$

where $m_i^{\ast }$ denote the effective mass, λ_i the magnetic penetration depth, ξ_i the coherence length and $H_{c2}^{\mathit{IIi}}$ the upper critical field in the magnetic field direction i. Nevertheless, the understanding of high-temperature superconductivity was challenged by the observation of two distinctly different and temperature dependent anisotropies in MgB₂ single crystals [27], [28], [29]:

$${\gamma }_{\lambda }={\lambda }_c/{\lambda }_{\mathit{ab}}\text{,}$$

$${\gamma }_H=H_{c2}^{\Vert \mathit{ab}}/H_{c2}^{\Vert c}\text{.}$$

A straight forward interpretation based on a two-band model was quickly developed, which also lead to a further understanding of the temperature and field dependence of the anisotropy parameters in MgB₂, mirroring the complex inter- and intraband mechanism of the two superconducting gaps [30]. For having an overall comparison with the other high-temperature superconductors (e.g. cuprates and MgB₂) a detailed knowledge of γ in the oxypnictides is required. Various attempts were made to estimate anisotropy in the oxypnictide superconductors [31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41], [42], leading to a wide range of results. Nevertheless, from both experimental and theoretical sides there is clear evidence that superconductivity in the pnictides involves more than one-band [31], [32], [33], [43], [44], [45], [46], [47], [48] and therefore it is expected that anisotropy is temperature dependent.

Not only the anisotropic properties are, obviously, best investigated on single crystals. Single crystals are also required for spectroscopic techniques such as Scanning Tunneling Spectroscopy (STS), Angle-Resolved Photoemission Spectroscopy (ARPES), Point-Contact Andreev-Reflection (PCAR) spectroscopy, optical spectroscopy, etc. A number of recent investigations of oxypnictides have focused on the multi-band superconductivity [31], [32], [33], [43], [44], [45], [46], [47], [48]. Answering the question of whether the different Fermi surface sheets are associated with different gaps is of crucial importance in order to identify the mechanism of superconductivity in these compounds. In this regard, experimental techniques such as ARPES, PCAR, and STS are very powerful methods since they allow a direct determination of the energy gap(s). STS is a suitable technique to study this issue since it probes the quasiparticle excitation spectrum in the superconducting state, a direct measure of the local density of states and therefore of the fundamental properties of the superconducting order parameter. This technique was successfully applied to MgB₂, where multi-band superconductivity was unambiguously demonstrated through directional STS measurements [49].

We succeeded in growing of the first free standing FeAs-oxypnictide crystals (SmFeAsO_1−xF_x) using a high pressure technique and NaCl/KCl flux [50]. The NaCl/KCl flux has very low solubility at temperatures below 1000 °C used for processes in quartz ampoules, therefore crystal growth at this temperature is extremely slow [51]. In order to increase the solubility in NaCl/KCl flux for more efficient crystal growth higher temperature should be used, but Ln1111 becomes unstable. This trend can be counteracted by applying high pressure, which then tends to stabilize the structure of Ln1111 at high temperature.

Single crystals of AFe₂As₂ can be grown from Sn flux, similar to many other intermetallic compounds [52], [53]. Tin is practically the only metal that dissolves iron reasonably well and does not form stable unwanted compounds. Due to high solubility in Sn flux at temperatures compatible with quartz ampoules, large, millimeter-sized crystals of A122 have been grown, which allowed extensive measurements of their physical properties. The disadvantage of the Sn flux technique is that crystals usually contain ∼1 at.% Sn. Another method of growing AFe₂As₂ crystals is the high-temperature growth from FeAs flux [54].

Here, we report on the crystal growth using both the high-pressure, high-temperature method with NaCl/KCl flux for Ln1111 and the quartz ampoule method with Sn flux for A122 [55]. The results of structure investigations on series of Ln1111 crystals (Ln = Sm, Nd, Pr, La, Gd) and Ba_1−xRb_xFe₂As₂ are presented. Electrical resistivity measurements, investigations of the anisotropy parameter and spectroscopic studies are also summarized.