Journal of Crystal Growth HO (1991) 994—998 North-Holland
Growth by self-flux method and post-annealing effect of Bi2Sr2Ca1Cu2O~single crystals Tatsuhiko Fujii, Yoshihisa Nagano and Junji Shirafuji Department of Electrical Engineering, Faculty of Engineering. Osaka University, Suita, Osaka 565, Japan
Received 1 August 1990; manuscript received in final form 29 November 1990
Post-annealing effects on superconducting characteristics have been studied in Bi2Sr2Ca1Cu2O~single crystals grown by a conventional flux method. Also, favorable growth conditions and the effect of the pre-sintering process on the starting materials for flux growth have been examined. The best superconducting behavior is obtained in post-annealed crystals grown from pre-sintered 5A/cm2 (HIIc) and.l~ 5X104A/cm2 at 4.2 Khysteresis at zero field. powder materials. The critical current density estimated from(Hic) magnetization in annealed crystals grown with pre-sintered materials is roughly 8X10
Bi-based superconducting copper oxides have, as is well known, two superconducting phases with 7~’shigher than the boiling point of liquid nitrogen, the highest 7~phase (7~ 105 K) and a lower T~phase (T~= 85 K). Materials with T~ 105 K have, so far, been prepared only by solid-state reaction processes, while single crystals solidified from the liquid phase have been obtained only in the lower 7~, 85 K, phase. However, it is worthwhile to note that 85 K crystals sustain a high critical current density, J~,up to magnetic fields as high as 20 T, for crystals kept near liquid helium temperature, in contrast with conventional hard superconductors such as Nb3Sn which shows a steep decrease in J~at magnetic fields higher than 15 T. The critical current density J~is determined by pinning effects in the superconductor; thus we need high quality single crystals in order to clarify the origin of the pinning centers and to study how to control the pinning effect for obtaining high J~. Moreover, single crystals are useful for investigating the highly anisotropic nature of high 7~superconductors. Bulk single crystals of Bi--Sr---Ca—Cu—O system have been grown by self-flux [1—7] and alkali chloride flux methods [8—10],and by the floating0022-0248/91/$03.50 © 1991
zone technique [11,12]. All of the single crystals grown from the liquid phase by these methods were predominately the 85 K phase, Bi 2Sr2Ca1 Cu2O~,although some cases showed a fractional component of the high 7~phase, Bi2Sr2Ca2Cu3O~, in the resistivity—temperature characteristic [5,10]. Most of the as-grown crystals display low 7~superconductivity in contrast with solution-grown YBa2Cu3O7_,~single crystals which require postannealing. However, the transition width at the superconducting transition was much less sharp than has been observed in conventional metal superconductors; the transition width reported was 5 K at best or more [1,3,61. The worst crystals showed a significant tail to the low temperatures in the resistivity—temperature characteristic [2,5], indicating inhomogeneity in the crystals. This inhomogeneity may arise partly from the use of carbonates of Ca and Sr as starting materials, because these carbonates are difficult to fully decompose to ensure a highly homogeneous melt [1,2,4,5]. When a mixture of starting materials consisting typically of Bi203, SrCO3, CaCO3 and CuO is directly melted in the crucible, it is very difficult to obtain a fully homogenized melt [1,2,5]. Thus, the thermal annealing of grown crystals should be effective in improving their supercon-
Elsevier Science Publishers B.V. (North-Holland)
T Fujii et al.
SrCO3 CaCO3,CuO Weighing,Mixing~0rying
Growth by self-flux and post-annealing effect of Bi
Starting Materials (Process 1)
3 Steps Sintering 760 C 24 h Grinding,Mixing 800C, 24 ~. Grinding .Mixing 820 °C,24 h Grindirig.Mixing
Pre-Sintering (Process 2]
Growth (Process 3]
R T Annealing
Post -Annealing (Process 4)
800 ~C,100 h
Fig. 1. Diagram for showing the procedures used in the present experiment.
powders of starting materials prepared in process 1 were subjected to pre-sintering in three steps under the conditions indicated as process 2 in fig. 1 to assure homogeneous solid-state reaction with asintering. special care not tognnding melt theand matenals Thorough mixing dunng procedures were employed after each pre-sintenng step. Appropriate amounts of raw materials prepared only by process 1 or pre-sintered materials provided through process 2 was placed in a highgrade alumina crucible with an open end diameter of about 4 cm and a height of about 5 cm. The crucible filled with powdered materials was heated in air to 9500 C for 2 h, then slowly cooled at a rate of 2°C/h down to 8000 C where the melt was entirely solidified, followed by quick cooling to room temperature. This growth process is mdicated as process 3 in fig. 1. When the atomic ratio of the starting matenals for the growth was Bi : Sr : Ca: Cu 2.2: 1.8: 1.0:2.0, the most satisfactory result was obtained. Most of the solidified bowl was occupied by a pile of thick plate-like crystals of the 85 K phase Bi2 Sr2Ca~Cu20X. Typical dimensions of the crystals removed mechani3, which cally from the bowlalong were the 10 xa—b 5 x crystallographic 0.5 mm grew dominantly =
ducting properties. However, little has been attempted except for a brief description by Takagi et al. , where they observed no annealing effect on In thethis superconducting properties. communication, we report growth procedures of bulk single crystals of Bi 2Sr2Ca3Cu 20X 3 by self-flux method as 10 xof5 annealing x 0.5 mm on the superconductandlarge the as effect ing characteristic of the crystals. Moreover, the effect of three steps of pre-sintering of starting materials prior to crystal growth is also studied in order to improve the homogeneous distribution of constituents in the solution. It has been well established that in ceramic superconductors, alternate sintering--grinding procedures to enhance homogeneous solid-state reaction result in good bulk materials. High purity materials of Bi 203 (99.9%), 5rCO3 (99.5%), CaCO3 (99.0%) and CuO (98.0%) were weighed in proper atomic compositions, mixed with a mortal and pestle in pure anhydrous ethanol and dried in air (process 1 in fig. 1). A variety of atomic ratios of Bi, Sr, Ca, and Cu were examined to grow single crystals of Bi-based superconductors. The atomic ratio, Bi: Sr: Ca : Cu 2.2: 1.8: 1.0 : 2.0, gave the best results in terms of the crystal size attained. In some growth runs, mixed =
plane bottom the crucible the top of the from bowl.the The broadofsurface of the toplate-like crystals was perpendicular to the c-axis with a lattice constant of 30.4 A. The size, especially in the length along the long edge, of the grown crystals depended dominantly on the volume of the charged material in the crucible; the larger the volume, the longer the crystals. Although the thickness of the removed single crystals was at most about 1 mm at present, because of a very low growth rate along the c-axis, an intentional ternperature gradient in the molten solution and/or the control of nucleation site may grow more bulky crystals. In the present growth experiment, the most favorable atomic ratio of the starting materials was found to be Bi : 5r: Ca: Cu 2.2: 1.8: 1.0: 2.0. In addition, it was confirmed that moderately large plate-like single crystals can be grown from the melt with atomic ratios near the best one, for example Bi: Sr: Ca: Cu 2.3—2.4: 1.6—1.5 : 1.0: 1.9—1.8. These observations are consistent with =
T. Fujii eta!.
/ Growth by self-flux and post-annealing
the results reported by Ren et al. ,in which the atomic ratio of the starting materials has been chosen at 2 : 2: 1: 2. Furthermore, the starting atomic ratio of 2.2 : 1.8 : 1.0: 2.0 in the present experiment is the same as those employed as the solvent zone in the traveling solvent floating zone method . It was also observed in the present experiment that a Bi 203-nch solution was not adequate for single crystal growth, and a CuO-rich solution gave only small platelet crystals embedded in the solidified matrix in accordance with previous reports [1—4,7]. An ICP analysis revealed the actual atomic composition of the grown crystals to be Bi : Sr: Ca: Cu 2.12: 1.77 : 0.98 : 2.0, being very close to the composition of the starting materials (2.2: 1.8: 1.0: 2.0). This is consistent with the fact that the pile of crystals occupies nearly the whole volume of the solidified bowl except for the thin crust touching the alumina wall and the surface region which solidifies last. The atomic compositionSr: deviated the been stoichiometric one, Bi: Ca: Cuslightly 2 : 2:from 1 : 2 has also observed by other researchers [3,6,12]. Conventional transmission electron microscopy (TEM) was carried out to study macroscopic imperfections in the =
• • •
50 TEMPERATURE (K )
crystals. TEM samples were prepared by ion-beam thinning of thin cleaved crystals. Twins were not found in the samples investigated so far, in contrast to the case of Y-based superconductors in which twins are predominantly observed . Instead, peculiar honeycomb-like dislocation networks as reported by Li et al.  were encountered, together with the frequent existence of defect free regions. Some of the crystals were further annealed in air at 800°C for 100 h (process 4 in fig. 1). Through these procedures as mentioned above, we prepared 4 classes of single crystalline samples from A to D, as illustrated in fig. 1. The magnetic moment and resistivity as a function of temperature were measured on these four classes of samples. A SQUID magnetometer was used to obtain the magnetization—temperature characteristic under zero-field-cooled (shielding) conditions. The applied field was 40 G in the configuration of H 1 C. The typical size the of the specimens was 0.5 X 3. For resistivity—temperature 0.5 X 0.2 mm measurement, Ag was evaporated characteristic onto bar-shape samples. Resistivity measurement was carried out at zero magnetic field. Fig. 2 shows the comparison of the zero-field-
effect of Bi
I • ••
Fig. 2. Zero-field cooled magnetization characteristics for four samples prepared by different procedures. Samples A—D correspond to those shown in fig. 1.
T. Fujii et al.
Growth by self-flux and post-annealing effect of Bi
83.6 aK slightly while sample without post-annealing shows round Asuperconductivity transi-
lionp—T the withcurve a lower in the 1~(end) normal 76.0 state K. for The the post-anslope of =
nealed B is larger than for sample A with nosample post-annealing; this that suggests that the
~ 04 0.2 76.0 K 0
10~K 150 200 0 _________________________ 50 T
• 0.4 0.2
B 83.6 K
100 150 200 T (K) Fig. 3. Normalized resistivity as a function of temperature for samples A with pre-sintering treatment and B with both presintenng and post-annealing processes.
cooled magnetization characteristic between 4 classes of samples prepared through different processes. As far as magnetic properties are concerned, the post-annealing (process 4) is effective compared to temperature the pre-sintering (process 2). The onset of theprocedure superconductivity is about 82 K for post-annealed crystals (samples B and C) accompanying relatively sharp transition in the negative magnetization, while crystals without post-annealing treatment (samples A and D) have somewhat lower onset temperature of about 80 K with a broad transition region. This may reflect a nonuniform distribution of constituents and/or a deficiency of oxygen content in as-grown crystals. The effect of post-annealing (process 4) on the superconductivity became also clear in the resistivity—temperature (p—T) characteristic in samples prepared with pre-sintered materials (A and B). Fig. 3 shows the results, where the post-annealed sample B exhibits a sharp superconductivity transition with the zero-resistivity temperature 7~(end)
post-annealing process may help to achieve better quality of crystals with a more metallic property in the normal state and sharp superconductivity transition with higher 7~.T he smaller slope of the p—T curve in sample A indicates the existence of inhomogeneities which may diminish the effective conducting cross-sectional area . In sintered Bi 2 Sr2CaCu 2°x ceramics, a prolonged annealing for 20 h at 800°C under various oxygen partial pressures has been found to enhance the onset transition temperature up to 94 K . However, in the present annealing experiment, no indication of the formation of a high lTc 2223 phase was observed at all. Samples C and D prepared without pre-sintering process showed poor electrical properties at the present stage of the experiment, suggesting a remarkable effect of the pre-sintering on the electrical properties of the grown crystals. The magnetization hysteresis cycle was measured at 4.2 K up to 5 T for samples B (pre-sintered plus post-annealed) in the configurations of both HIIc and H I c. The magnetic critical current density J~was estimated on the basis of Bean’s critical state model . of J~ obtained 2 The for values HIIc and 4.6 X iO~ were 7.7for x iO~ A/cm2 H I A/cm c at 4.2 K at zero field. These values are compared to those observed in single crystals grown from a CuO-rich melt  (J~ 1.2 x 10~A/cm2 for HIIc and J~ 4.7 x iO’ A/cm2 for H I c) and to those reported in crystals grown from KC1 solution  (J~ 8.5 x iO~A/cm2 for HIIc). The anisotropy of J~ (c~/c~ 17 25) observed by us and by Biggs et al.  is rather larger, while Agostinelli et al.  have reported a small anisotropy (J~/J~ 1.6) in single crystals grown from CuO-riJi solution. In summary, we have studied pre-sintering and post-annealing effects on Bi 2Sr2CaCu2O5 single crystals grown by the self-flux method. When the nominal atomic composition is Bi: Sr: Ca: Cu 2.2: 1.8: 1.0: 2.0, most of the charged material is solidified into a bulky aggregate of long plate-like .
T. Fuji! eta!.
by self-flux and post-annealing effect of Bi
crystals with composition of Bi: Sr: Ca: Cu 2.12: 1.77 : 0.98 : 2.0 from which crystals as 3 are single removed. From large as 10cooled x 5 x 0.5 mm zero-field magnetization measurements, post-annealing of the grown single crystals gives rise to a considerable improvement of the superconductivity, while the pre-sintering process produces little effect. However, the advantage of presintering becomes obvious in the resistivity—temperature characteristic. In single crystals grown from pre-sintered materials and subject to postannealing, a sharp superconducting transition with 7~(onset) 96 K and T~(end) 83.6 K is attained. The magnetic critical current density J~ ~ estimated in pre-sintered and post-annealed crystals from the magnetization hysteresis cycle on the basis of Bean’s critical field model. The magnetic remanence shows a remarkable anisotropy. at 4.2 K and at zero field is 7.7 x iO~A/cm2 for HIIc and 4.6 x iO~A/cm2 for HI c, respectively. =
The authors would like to express their sincere thanks to Superconductivity and Low Temperature Research Center, Kobe Steel, Ltd., for magnetization hysteresis measurements and ICP analysis.
References  H. Takagi, H. Eisaki, S. Uchida, A. Maeda, S. Tajima, K. Uchinokura and S. Tanaka, Nature 332 (1988) 236.  J.Z. Liu, G.W. Grabtree, L.E. Rehn, U. Geiser, D.A. Young, W.K. Kwok, P.M. Baldo, J.M. Williams and D.M. Lam, Phys. Letters A127 (1988) 4.44. [31Y. Hidaka, M. Oda, M. Suzuki, Y. Maeda, Y. Enomoto and T. Murakami, Japan. J. AppI. Phys. 27 (1988) L538.
 G. Balestrino, U. Gambardella, Y.L. Liu, M. Marinelli, A. Paolleti,674. P. Paroli and G. Patemo, J. Crystal Growth 92 (1988)  Z. Ren, X. Yan, Y. Jiang and W. Guan, J. Crystal Growth 92 (1988) 677.  S. Nomura, T. Yamashita, H. Yoshino and K. Ando, Japan. J. Appl. Phys. 27 (1988) L1251.  J.J. Lin, EL. Benitez, S.J. Poon, MA. Subramaman, J.
Gopalakrishnan and A.W. Sleight, Phys. Rev. B38 (1988) 5095.  L.F. Schneemeyer, RB. van Dover, S.H. Glarum, S.A. Sunshine, R.M. Fleming, B. Batlogg, T. Siegrist, J.H. Marshall, J.V. Waszczak and LW. Rupp, Nature 332 (1988) 422. 19] A. Katsui, Japan. J. AppI. Phys. 27 (1988) L844.  T. Shishido, D. Shindo, K. Ukei, T. Sasaki, N. Toyota and T. Fukuda, Japan. J. Appl. Phys. 28 (1989) L791.
Takekawa, H. Nozaki, A. Umezono, K. Kosuda and M. Kobayashi, J. Crystal Growth 92 (1988) 687. [121 Y. Kubo, K. Michishita, Y. Higashida, M. Mi.zuno, H. Yokoyama, N. Shimizu, E. Inukai, N. Kuroda and H. Yoshida, Japan, J. Appl. Phys. 28 (1989) L606.  K. Jagannadham and J. Narayan, Phil. Mag. A59 (1989) 917.  Z.Q. Li, H. Shen, Y. Qin, J.Y. Jiang and J.J. Du, Phil. Mag. Letters 60 (1989) 123.  PB. Allen, Z. Fisk and A. Migliori, Normal state transport and elastic properties of high 1~,materials and related compounds, in: Physical Properties of High Temperature Superconductors, Vol. I, Ed. D.M. Ginsberg (World Scientific, Singapore, 1989) p. 213.  Y. Motoi, Y. Ikeda, H. Uwe and T. Sakudo, Physica C162—164 (1989) 929.  C.P. Bean, Phys. Rev. Letters 8 (1962) 250.  B.D. Biggs, MN. Kunchur, J.J. Lin, S.J. Poon, T.R. Askew, R.B. Flippen, M.A. Subramanian, J. Gopalakrishnan and A.W. Sleight, Phys. Rev. B39 (1989) 7309.  RB. van Dover, L.F. Schreemeyer, EM. Gyorgy and J.V. Waszczak, AppI. Phys. Letters 52 (1988) 1910. [201 E. Agostinelli, G. Balestrino, D. Fiorani, R. Muzi, P. Paroli, J. Tejada and AM. Testa, Physica C162—164 (1989) 319.