Synthesis and structural characterization of a superconducting GdSr2RuCu2O8−δ single crystal grown by partial melting

We successfully synthesized a superconducting GdSr2RuCu2O8−δ (RuGd-1212) single crystal via the partial melting technique. Structural analysis of this crystal indicated the substitution of Cu and Gd at the Ru and Sr sites, respectively. The disappearance of superconductivity in non-superconducting samples could be attributed to the substitution of Ru in the CuO2 planes and/or Gd at the Sr sites. The distance between the CuO2 planes in the superconducting RuGd-1212 single crystal was shorter, indicating that the valence of Cu in the CuO2 plane had decreased. A superconducting transition with Tc-onset of 23.0 K and Tc-zero of 6.2 K was observed.

T he discovery of oxypnictide superconductors has triggered the development of new superconducting materials based on FeAs and BiS 2 as the superconducting layers. [1][2][3] Since iron is a magnetic element, much attention has been paid to the magnetism and superconductivity of iron-based superconductors. Furthermore, the coexistence of the antiferromagnetic order of Eu 2+ and superconductivity has been reported in the EuFe 2 (As,P) 2 system. 4,5) In addition, simultaneous superconductivity and spontaneous magnetization were recently observed in Ce(O,F)BiS 2 and EuAFe 4 As 4 (A = Rb and Cs). 6,7) Therefore, the coexistence of magnetism and superconductivity has attracted renewed attention.
Rutheno-cuprate materials such as RESr 2 RuCu 2 O 8−δ (RuRE-1212) and RE 2−x Ce x Sr 2 RuCu 2 O 10−δ (RuRE-1222), where RE denotes a rare-earth element, have also been shown to behave as superconducting ferromagnets. RuRE-1212 and RuRE-1222 were first synthesized and determined to be superconducting by Braun et al. in 1995. 8) Subsequently, Felner et al. demonstrated the coexistence of magnetism and superconductivity in RuRE-1222. 9) The crystal structure of RuRE-1212 is similar to that of YBa 2 Cu 3 O 7 (Y-123), 8) in which the CuO chain, the Y ions sandwiched between the CuO 2 planes, and the Ba ions of the blocking layer are replaced by RuO 2 octahedra, rare-earth ions, and Sr ions, respectively. The magnetic ordering temperature, T M (130-150 K), of RuRE-1212 is higher than its superconducting transition temperature, T c (30-50 K). The T M value of RuRE-1212 is much higher than those of the aforementioned EuFe 2 (As,P) 2 and Ce(O,F)BiS 2 . Thus, various investigations have been performed to understand the origin of the coexistence of magnetism and superconductivity in the ruthenocuprates. [10][11][12][13][14][15][16] These studies revealed that the observed ferromagnetism and superconductivity originate from the Ru and the CuO 2 planes, respectively. Specifically, the magnetism and superconductivity of RuRE-1212 were understood from the perspective of the stacking structure of SrRuO 3 (an itinerant ferromagnet with a T M of approximately 160 K) and the CuO 2 planes (responsible for the superconductivity of high-T c cuprates). Furthermore, rutheno-cuprates are considered to be potential candidates for π junctions, and it has been experimentally demonstrated that π-junction qubits exhibit coherence times comparable with traditional superconducting phase qubits. 17) In particular, RuGd-1212 has the potential to facilitate the realization of robust qubits in the future. However, numerous questions concerning rutheno-cuprates remain unanswered, such as their broad superconducting transition width and the absence of a bulk Meissner effect. Subtle differences in the sample preparation conditions can greatly affect the superconducting and magnetic properties of rutheno-cuprates, which complicates the investigation of these materials. [18][19][20][21] Furthermore, it was reported that the superconductivity of the RuEu-1222 system may be attributable to the intergrowth of the satellite EuSr 2 Cu 3−x (CO 3 ) x O 7+z phase. 22) It was also suggested that the superconductivity of RuGd-1212 may originate from its decomposition at high temperature. 23) Therefore, it is still unclear whether the observed superconductivity is truly a bulk intrinsic property or not. The investigation of high-quality single crystals of RuRE-1212 compounds would be indispensable with regard to answering these questions; however, the difficulty of crystal growth and the variability of the growth conditions have hindered progress in the study of these systems.
Lin et al. reported on the growth of RuGd-1212 single crystals with dimensions up to approximately 200 × 200 × 50 µm 3 using the self-flux method. 24) Furthermore, bulk superconductivity and intrinsic Josephson junction properties were successfully demonstrated. 15,25) It is surprising that the SrRuO 3 layer serves as a blocking layer given that SrRuO 3 is a metallic itinerant ferromagnet. It is plausible that RuGd-1212 exists in an underdoped state in analogy with the Y-123 system. [26][27][28] This possibility appears to be consistent with the dispersion of T c in the RuRE-1212 systems. However, the quality of the previously obtained RuGd-1212 single crystals 15,24,25) was not high, as demonstrated by the fluctuation of the T c value and the broad transition width. The observed T c value exceeds 50 K, which suggests the partial substitution of Cu at the Ru sites. 20,29) It should be noted that for GdSr 2 -Ru 1−x Cu 2+x O 8−δ maximum T c of 72 K was observed at x = 0.4, whereas T c of stoichiometric RuGd-1212 is 30-50 K. Such variations in T c are ideally understood in relation to the crystal structure. However, a detailed structural analysis of a RuRE-1212 single crystal has not been reported to date.
Recently, we successfully synthesized cubic-shaped single crystals of Gd 1+y Sr 2−y Ru 1−x Cu 2+x O 8−δ with typical lengths of 100-150 µm using the partial melting technique. 30,31) However, these single crystals did not exhibit superconductivity. 31) Single-crystal X-ray diffraction (XRD) measurements and structural refinement indicated that the absence of superconductivity may have originated from the increase in the lattice parameter a. 31) Since the superconductivity of the RuRE-1212 materials is highly sensitive to the sample preparation conditions, we considered that it would be valuable to reexamine the synthetic conditions used to prepare RuGd-1212 by the partial melting technique, and to characterize the acquired single crystals in more detail.
Polycrystalline precursors of GdSr 2 Ru 0.7 Cu 2+0.3n O 6.9+0.3n with n = 0, 1, 2, and 3 were prepared by conventional solidstate reactions using RuO 2 (Ru 76.45 wt %), SrCO 3 , Gd 2 O 3 , and CuO (3N) powders. Subsequently, RuGd-1212 single crystals were synthesized using the partial melting technique. In this synthesis process, n represents the ratio of additional Sr : Gd : Cu flux. In other words, n = 0 corresponds to a Sr : Gd : Cu flux ratio of 2 : 1 : 2, whereas n = 3 corresponds to a Sr : Gd : Cu flux ratio of 2 : 1 : 5. The partially melted flux promotes the growth of single crystals. In this study, the crystal structures and conductivities of the RuGd-1212 single crystals were evaluated with a focus on superconductivity. The detailed synthesis procedure is described elsewhere. 30,31) Figure 1(a) shows the XRD patterns of the GdSr 2 Ru 0.7 -Cu 2+0.3n O 6.9+0.3n samples with n = 0, 1, 2, and 3 prepared by the partial melting technique, along with that of polycrystalline RuGd-1212 as a reference. A two-step sintering procedure was used for all samples with T 1 = 1130°C followed by T 2 = 1070°C for 48 h. This two-step sintering procedure promotes the melting of the precursors during sintering at T 1 (1130°C) and stabilizes the RuRE-1212 phase during sintering at T 2 (1070°C). These temperatures were selected based on the optimum conditions for obtaining the largest single crystals of RuGd-1212 as reported in previous studies. 30,31) The main peaks in the diffraction patterns for the samples with n = 0, 1, 2, and 3 coincided with those for the RuGd-1212 polycrystal. In addition, for the samples with n = 0 and 2, additional impurity phases were clearly observed indicating the presence of the SrRuO 3 (+) and Sr 2 GdRuO 6 (2116) (+) phases. 32) Figure 1(b) shows scanning electron microscopy (SEM) images of the samples with n = 1 and 3. These images demonstrate that the single crystals possess a cubic shape with a length of approximately 100 µm in both cases. However, in the case of n = 3, crystals without a cubic shape were also observed. An increase in the amount of CuO decreased the melting point of the precursor, which inhibited the growth of larger cubic RuRE-1212 single crystals. Table I is a summary of the maximum size and composition of the samples, as determined by SEM=energydispersive X-ray spectroscopy measurements. It can be seen that the largest single crystals were obtained for n = 1. Although we succeeded in synthesizing single crystals with a length exceeding 100 µm, these samples did not exhibit superconductivity down to 4.2 K. Since prolonged sintering in oxygen was previously reported to lead to superconductivity in the RuRE-1212 system, 18,19,33) we also investigated the prolonged sintering of the n = 1 sample using the partial melting technique.
Sintering of this sample at T 2 for 150 h led to T c-onset of 23 K. However, zero resistance was not observed down to 4.2 K. This sample contained approximately two to five single crystals, as observed using an optical microscope. XRD measurements revealed that the RuGd-1212 phase was retained even after prolonged sintering at T 2 for 150 h. Therefore, prolonged sintering is considered to be useful for achieving superconductivity in RuGd-1212. To investigate the origin of superconductivity in RuGd-1212, the crystal structure was refined based on single-crystal XRD measurement data for a single crystal selected from a region of the superconducting sample. The intensity data were collected using a chargecoupled device (CCD)-type single-crystal diffractometer at 295 K. In the analysis, the initial model was determined using the SHELXT software 34) and the final values of R and R w were 2.81 and 6.58%, respectively. The Sr and Ru sites were    Table II. The observed lattice parameters [a = 3.8420(2) Å and c = 11.4918(6) Å] were slightly different from those of previously reported for polycrystalline RuGd-1212 samples. 8,12,19,21,29) These differences were further increased by prolonged sintering at T 2 . 31) The smaller lattice parameter c is consistent with polycrystalline RuGd-1212 in which the Ru sites are substituted by Cu. 20,29,31) Furthermore, the superconducting GdSr 2 Ru 1−x Cu 2+x O 8−δ sample was determined to have a lattice parameter a that was larger than that of the nonsuperconducting sample in our previous report, 31) and also similar to that reported previously for nonsuperconducting RuGd-1212 samples doped with La at the Sr sites. 35) Elongation of the lattice parameter a by doping with La at Sr sites is consistent with the results of our single-crystal analysis in which Gd substitutes at Sr sites. Since some of the samples with an elongated lattice parameter a exhibits superconductivity, it is necessary to reconsider the origin of superconductivity in RuGd-1212. Table III presents a structural comparison of the superconducting single crystal of RuGd-1212 synthesized in this work, two previously reported superconducting polycrystals, 12,29) and our previously reported non-superconducting single crystal. 31) Polycrystal I has a stoichiometric composition, 12) whereas polycrystal II was synthesized under high pressure and contains a partial substitution of Cu at the Ru sites. 29) The superconducting and nonsuperconducting single crystals have a larger lattice parameter a than polycrystals I and II owing to the substitution of Gd at the Sr sites. It is considered that the partial melting technique leads to a different site occupancy compared with the typical solid-state reaction because in the former method, the sample is sintered at the decomposition temperature of the RuRE-1212 phase for a short time. Furthermore, since no substitution of Ru on the CuO 2 plane was detected for the superconducting samples, the occupancy of Ru on the CuO 2 plane appears to be correlated with the emergence of superconductivity in RuGd-1212 single crystals. There is also the possibility that the substitution of Gd at the Sr sites inhibits superconductivity, because the T c-onset value of the superconducting single crystal synthesized in this work is lower than that reported for polycrystal II. This is because the Sr in the blocking layer is adjacent to the CuO 2 plane and hence, the large magnetic moment of Gd at the Sr sites may prevent superconductivity. In general, high-T c cuprates are relatively robust against impurities such as Co, Ni, and Zn, [36][37][38][39] but the results of this study indicate that the presence of Ru in the CuO 2 layer or Gd at the Sr sites may destroy the superconductivity of the CuO 2 plane. This loss of superconductiv-  (5) Note 1: Final R values: R = 2.81%, R w = 6.58%, a = 3.8420(2) Å, and c = 11.4918(6) Å (295 K). Note 2: The constraints used for the cation occupancies were g(Sr) + g(GdA) = 1 and g(Ru) + g(CuA) = 1. Note 3: No Ru substitution at Cu sites forming the CuO 2 plane was observed on refinement [Cu : Ru ¼ 1:000ð13Þ : 0:000ð13Þ]. ity due to Ru and Gd substitution is reasonable given the broad transition width of rutheno-cuprate magnetic superconductors; that is, the heterogeneity with respect to site substitution may exert a strong influence on T c . Therefore, since the Ru and Gd substitution in the single crystal is heterogeneous, the transition width becomes broad. Furthermore, the Cu-Cu distance between the CuO 2 planes is shorter for the superconducting single crystal than that of the nonsuperconducting single crystal. This suggests that the valence of Cu in the CuO 2 plane decreased. A bond valence sum analysis indicates a high hole concentration of ∼0.4 in the CuO 2 plane. 12,40) Since the Ru-NMR study of pure RuGd-1212 revealed that 40 and 60% of Ru have a valence of 4 and 5, respectively, the valence of Cu in the CuO 2 plane results in 2.20. 41) However, the electrical properties of pure RuGd-1212 are similar to those of heavily underdoped high-T c materials. 35) In our results, the decreased valence of Cu was observed for the superconducting RuGd-1212 single crystal.
Since the results presented above indicate that prolonged sintering leads to superconductivity, the n = 1 sample was sintered at T 2 for 300 h. Figure 2(a) shows the R-T characteristic for this RuGd-1212 single-crystal sample, as measured using the four-probe method. The inset shows an optical microscopy image of the sample, with a silver paste attached to the edge of the crystal surface along the ab plane. A superconducting transition with T c-onset of 23.0 K and T c-zero of 6.2 K was successfully observed. This constitutes the first observation of zero resistance for a single crystal of RuGd-1212 prepared using the partial melting technique. The superconductivity of this sample may originate from the prevention of Gd substitution at the Sr sites by the prolonged sintering, when compared with the structural analysis of the superconducting single crystal obtained after sintering for 150 h. However, the observed T c value was lower than those of polycrystalline samples in which the Ru site is partially substituted by Cu. This seems to be strongly related to the substitution of Gd to the Sr sites. 35) The R-T characteristic of the RuGd-1212 single crystal revealed metallic behavior above 85 K and semiconducting behavior below 85 K. This temperature dependence is consistent with the results obtained for our previous non-superconducting single crystal. 31) However, there is also the possibility that the electrical component of the c-axis contributed to the resistance, since the absolute value of resistance was relatively high compared with our previous samples. Figure 2(b) shows the I-V characteristic at 4.2 K. The I c value was estimated to be approximately 3 mA. The temperature dependence of the critical current was found to be linear (data not shown), which may reflect an underdoped state and=or electrical inhomogeneity in this single crystal. The inset in Fig. 2(b) shows the magnetic field dependence of the zero-field-cooled magnetization (M-H ) for the same RuGd-1212 single crystal. Since no diamagnetism was observed, this single crystal did not exhibit bulk superconductivity. The results appear to be consistent with those reported for a superconducting RuGd-1212 single crystal prepared by the self-flux method in a platinum crucible. 15) Since the T c value of our sample is low, further investigations should be performed to elucidate the correlation between the magnetic order, superconductivity, and crystal structure in more detail.
In conclusion, we have successfully grown single crystals of RuGd-1212 using the partial melting technique. The obtained single crystals possessed a cubic shape with a typical length exceeding 100 µm. Prolonged sintering was found to be effective for realizing superconductivity. The structural analysis results for the single crystal revealed that the Ru and Sr sites were partially substituted by Cu and Gd ions, respectively. The loss of superconductivity in non-superconducting samples may be attributable to Ru and Gd substitution of the CuO 2 planes and Sr sites, respectively. Furthermore, the distance between the CuO 2 planes in the superconducting sample was shorter than that in a non-superconducting single crystal, which suggests that the valence of Cu in the CuO 2 plane decreases. Finally, a superconducting transition with T c-onset of 23.0 K and T c-zero of 6.2 K was observed for the first time for a RuGd-1212 single crystal prepared by partial melting.