Available online at www.sciencedirect.com
Physics Procedia 45 (2013) 37 – 40
ISS2012
Superconducting gap structure in out-of-plane-disordered Bi2Sr2CaCu2O8+ as studied by Raman spectroscopy N. Muraia, T. Masuia*, M. Ishikadob, c, S. Ishidab, H. Eisakid, S. Uchidab, S. Tajimaa a Department of Physics, Osaka University, Toyonaka, Osaka 560-0043, Japan b Department of Physics, University of Tokyo, Tokyo 113-0033, Japan. c Japan Atomic Energy Agency, Tokai, Naka, Ibaraki 319-1195, Japan d National Institute of Advanced Industrial Science and Technology (AIST), Ibaraki, 305-8568, Japan
Abstract We measured electronic Raman spectra of nonstoichiometric Bi2+xSr2 xCaCu2O8+ with x=0.1, 0.3 and stoichiometric Bi2Sr2Y0.08Ca0.92Cu2O8+ and investigated the effect of out-of-plane disorder on the superconducting B1g Raman response. We observed that the B1g peak energy is enhanced in Y-substituted Bi2212, compared with that in nonstoichiometric Bi2212. By contrast, for nonstoichiometric Bi2212, the B1g peak energy remains constant, even if Tc is significantly suppressed by disorder. This anomalous behavior is discussed in terms of the interplay between the pseudogap and superconducting gap.
© 2013 The Authors. Published by Elsevier B.V. © 2013 The Authors. Published by Elsevier B.V. Selection and/or peer-review under responsibility of ISS Program Committee. Selection and/or peer-review under responsibilty of ISS Program Committee. Keywords: High-Tc superconductivity; Electronic Raman scattering; Disorder effect
1. Introduction In spite of tremendous amounts of studies, there remain a number of questions which are yet to be resolved in high-Tc superconducting cuprates. One of them is that Tc of the cuprates is significantly material dependent, ranging from 30 K from 150 K. Recently, Eisaki et al. have demonstrated that disorder outside the CuO2 plane substantially reduces Tc and suggested that disorder outside the CuO2 plane might be responsible for significant variation of Tc within the cuprates family [1]. Since carriers are doped either by varying oxygen content in blocking layer or chemical substitution, almost all the cuprates are inherently disordered. Therefore, it is important to understand the effect of outof-plane disorder on the superconducting property. In order to investigate the effect of out-of-plane disorder on the superconducting gap structure, we used Raman spectroscopy. Electronic Raman scattering (ERS) is one of the powerful methods to probe the electronic excitation in the different region of the Fermi surface. Recently, precise doping dependence of Raman spectra has been reported for HgBa2CuO4- [2]and Bi2Sr2CaCu2O8+ [3], which reveal that the superconducting gap extracted from B 1g symmetry (dominated by the antinodal region) does not scale with Tc, but rather follows the doping dependence of the pseudogap temperature (T*). Nevertheless, the B1g peak is observed only below Tc. In order to study the origin of the B 1g gap, it is necessary to compare the results of samples with different Tc. One of the ways to tune Tc is to introduce disorder outside the CuO2 plan. In this study, we focus on Bi2Sr2CaCu2O8+ (Bi-2212) and evaluate the effect of out-of-plane disorder on superconducting property using ERS. We prepared Bi2+xSr2 xCaCu2O8+ with x=0.1, 0.3 and Bi2Sr2Y0.08Ca0.92Cu2O8+ . In the former materials, Bi/Sr nonstoichiometery is a dominant source of disorder outside the CuO2 plane, while the latter is almost stoichiometric in Bi/Sr ratio. Although Ca site is disordered in Y-substituted Bi2212, it is well known that the disorder at Ca site has negligible effect on Tc [4]. Here, the effect of out-of-plane disorder on the *Corresponding author. Tel.:+81-6-6850-5584; fax: +81-6-6850-5585 E-mail address:
[email protected]
1875-3892 © 2013 The Authors. Published by Elsevier B.V. Selection and/or peer-review under responsibilty of ISS Program Committee. doi:10.1016/j.phpro.2013.04.046
38
N. Murai et al. / Physics Procedia 45 (2013) 37 – 40
superconducting gap in the antinodal region is presented and discussed. 2. Experimental High-quality single crystals of Bi2212 were grown by a traveling-solvent-floating-zone (TSFZ) method [4]. In this study, we prepared (1) Bi2+xSr2 xCaCu2O8+ with disorder at Sr sites (x=0.1, 0.3) and (2) Bi2Sr2Y0.08Ca0.92Cu2O8+ with disorder at Ca site. It is well established that substitution of Y for Ca helps to make Bi/Sr ratio close to stoichiometry and to raise Tc up to 96 K for y=0.08[1,4]. Since Bi/Sr and Y/Ca substitution change the doping level, we need to control oxygen content in order to fix a doping level for all the samples. Here, careful attention was paid to maximize Tc values by changing oxygen content. The Tc values are 91.5 K for x=0.1, 77 K for x=0.3, and 96 K for Ysubstituted B-i2212. These are the maximum Tc values for each sample, confirming that all of them are optimally doped. Raman scattering experiments were carried out in pseudo-back scattering configuration using an Ar laser line (514.5 nm) and T64000 Jobin-Ybon triple grating spectrometer equipped with a nitrogen-cooled charge coupled device (CCD) detector. The B1g Raman spectra have been obtained using cross polarization tilted by 45° from Cu-O bond direction. ERS for B1g symmetry is sensitive to the electronic excitations from the antinodal regions (along the X direction) of the Brillouin zone, where the amplitude of the superconducting d-wave gap reaches its maximum [5]. All the spectra have been corrected for Bose-Einstein factor. 3. Results and discussion In Fig. 1, we show the temperature evolution of B1g Raman spectra of three Bi-2212 crystals. All the sharp peaks are due to phonons. The responses due to superconductivity are seen in the electronic continua. Below Tc, a broad peak appears in the Raman spectra as a result of superconductivity-induced redistribution of the electronic continuum. This broad peak is associated with a breaking of Copper-pair and thus located at 2 for B1g symmetry, (a) B1g
80K 60K 30K 10K
2
500
Raman shift (cm )
Intensity (arb. units)
1
0 0
Intensity (arb. units)
2
Bi2.3Sr1.7CaCu 2O8+ Tc= 77 K
1
100K 80K 75K 60K 50K 10K 500
Raman shift (cm )
(d)
(b) B1g Bi2.1Sr1.9CaCu 2O8+ Tc= 91.5 K
100K 80K 70K 60K 45K 10K 500
Raman shift (cm )
1000
1000
Bi2.3Sr1.7CaCu 2O8+ Tc=77 K Bi2.1Sr1.9CaCu 2O8+ Tc=91.5 K
1
1
0 0
(c) B1g Bi2Sr2Y0.08Ca0.92Cu2O8+ Tc=96 K
0 0
1000
Peak intensity
Intensity (arb. units)
2
Bi2Sr2Y0.08Ca0.92Cu2O8+ Tc=96 K 0.5
0 0
0.5
1
1.5
Tc/T
Fig.1. Temperature dependence of the B1g Raman spectra of Bi-2212 with different degree of Bi/Sr nonstoichiometry. (a) Bi2.3Sr1.7CaCu2O8+ with Tc=77 K. (b) Bi2.1Sr1.9CaCu2O8+ with Tc=91.5 K. (c) Bi2Sr2Y0.08Ca0.92Cu2O8+ with Tc=96 K. (d) Temperature dependence of the B1g peak intensity.
39
N. Murai et al. / Physics Procedia 45 (2013) 37 – 40
where is superconducting gap amplitude [5]. The spectra show a gradual suppression of superconductivity-induced peak intensity with increasing temperature up to Tc and its disappearance at Tc. In Fig. 1(d), we plotted normalized B 1g peak intensity as a function of T/Tc. Here, the peak intensity is defined by a peak area deduced from the difference between the superconducting and normal state Raman response. Since peak intensity disappears at Tc, this peak can be interpreted as a pair-breaking peak, although it is known that doping dependence of the B 1g peak energy is different from that of Tc [2, 3]. As shown in Fig. 1(d), the B 1g peak of Bi2Sr2Y0.08Ca0.92Cu2O8+ evolves more rapidly than that of Bi2.1Sr1.9CaCu2O8+ . The B1g peak intensity of the former reaches nearly its maximum at T/Tc~0.5, while that of the latter at T/Tc~0.5 is just half of its maximum intensity.
(a) B1g Intensity (arb. units)
2
Bi2.3Sr1.7CaCu 2O8+
(b) B1g Tc=77 K
1
0 0
500
Raman shift (cm )
Bi2.1Sr1.9CaCu 2O8+
Tc=91.5 K
1000
2
Intensity (arb. units)
Bi2.1Sr1.9CaCu 2O8+
Tc=91.5 K
Bi2Sr2Y0.08Ca0.92Cu2O8+ Tc=96 K
1
0 0
500
Raman shift (cm )
1000
Fig.2. Superconducting B1g Raman spectra of Bi-2212 at 10 K. (a) Effect of increasing Bi/Sr nonstoichiometyry on the B1g Raman response. (b) Effect of reducing Bi/Sr nonstoichiometry on the B1g Raman response. Fig. 2(a) and (b) summarize the B1g Raman spectra of Bi2Sr2Y0.08Ca0.92Cu2O8+ and Bi2+xSr2 xCaCu2O8+ in the superconducting state. First, we focus on the effect of increasing disorder due to Bi/Sr nonstoichiometry. Fig.2 (a) shows the superconducting B1g Raman response in Bi2+xSr2 xCaCu2O8+ with different x. Surprisingly, the B1g peak energy does not change with increasing disorder, but remains constant. This behavior is different from what one would expect for conventional superconductors where the superconducting gap scales with Tc. If we regard the B1g peak energy as a pair-breaking energy, it should scale with Tc. However, it is well known that the B1g peak energy does not show the same doping dependence as Tc but rather follows the doping dependence of T*, namely, the B1g peak energy monotonically increases with decreasing doping level [2, 3]. One possible explanation for this is that the pseudogaplike behavior of the B1g peak results from the interplay between the pseudogap and superconducting gap when the pseudogap is larger than the superconducting gap and its effect is predominant [6]. In fact, recent ERS experiments have revealed that B1g peak observed only below Tc gives the same energy as the pseudogap energy observed above Tc [7]. In this picture, the B1g peak energy is disconnected from superconductivity and largely affected by the pseudogap. In the present case, the superconducting gap itself is suppressed by disorder, as revealed by the suppression of the B2g gap in our previous work [8], while the B1g peak energy remains constant, because the pseudogap is unaffected by disorder at a fixed doping level and it determines the B1g peak energy. We note that the robustness of the B1g peak energy under the presence of disorder has been reported also for Zn- or Ni-substituted cuprates [9]. Next, we discuss the effect of reducing Sr site disorder shown in Fig. 2(b). With minimizing Bi/Sr nonstoichiometry by Y substitution, the B 1g peak energy shifts to higher energy, which is in remarkable contrast with the result in Fig. 2(a). At first glance, this seems to indicate an enhancement of the pseudogap energy in Y-Bi2212. However, this is unlikely. If the pseudogap energy of Y-Bi2212 is enhanced, compared with that of Bi2+xSr2 xCaCu2O8+ , the B2g peak energy should decrease as a result of loss of the quasiparticle weight in the antinode [2, 3]. Preliminary results of ERS measurements show that the B 2g gap in Y-Bi2212 is larger than that in nonstoichiometric Bi2+xSr2 xCaCu2O8+ , suggesting that the observed enhancement of the B1g peak energy should rather be attributed to the enhancement of the superconducting gap (not the pseudogap). Then, the question is why the change of the B1g peak energy is observed only for Y-substituted Bi2212. One possible interpretation is that the superconducting gap becomes larger than the pseudogap with reducing disorder, leading to switching of the dominant
40
N. Murai et al. / Physics Procedia 45 (2013) 37 – 40
source which affects the B1g peak. In that case, the observed B1g peak is expected to reflect the superconducting gap, and therefore scales with Tc. 4. Summary We measured ERS of nonstoichiometric Bi2+xSr2 xCaCu2O8+ with x=0.1, 0.3 and stoichiometric Bi2Sr2Y0.08Ca0.92Cu2O8+ to investigate the effect of out-of-plane disorder on the superconducting B1g Raman response. We observed that the B1g peak energy is enhanced in Y-substituted Bi2212, compared with that in nonstoichiometric Bi2212, as a result of reducing Bi/Sr nonstoichiometry. By contrast, for nonstoichiometric Bi2212, the B1g peak energy remains constant, even if Tc is significantly suppressed by disorder. This anomalous behavior can be explained by the contribution of the pseudogap to the superconducting B 1g Raman response. These results suggest a complicated interplay between the pseudogap and superconductivity. References [1] H. Eisaki, N. Kaneko, D. L. Feng, A. Damascelli, P. K. Mang, K. M. Shen et al., Phys. Rev. B 69, 064512 (2004). [2] M. Le Tacon, A. Sacuto, A. Georges, G. Kotliar, Y. Gallais, D. Colson et al., Nat. Phys. 2, 537 (2006). [3] S. Blanc, Y. Gallais, M. Cazayous, M. A. Measson, A. Sacuto, A. Georges et al.,Phys. Rev. B 82, 144516 (2010). [4] H. Hobou, S. Ishida, K. Fujita, M. Ishikado, K. M. Kojima, H. Eisaki, et al., Phys. Rev. B 79, 064507 (2009). [5] T. P. Devereaux, D. Einzel, B. Stadlober, R. Hackl, D. H. Leach, J. J. Neumeier, Phys. Rev. Lett. 72, 396 (1994). [6] W. Guyard, A. Sacuto, M. Cazayous, Y. Gallais, M. Le Tacon, D. Colson et al., Phys. Rev. Lett. 101, 097003 (2008). [7] A. Sacuto, Y. Gallais, M. Cazayous, S. Blanc, J. S. Wen, Z. J. Xu et al., C. R. Acad. Sci. 12, 480 (2011). [8] N. Murai, T. Masui, M. Ishikado, S. Ishida, H. Eisaki, S. Uchida et al., Phys. Rev. B 85, 020507(R) (2012). [9] M. Le Tacon, A. Sacuto, Y. Gallais, D. Colson, A. Forget, Phys. Rev. B 76, 144505 (2007).