Physica C 349 (2001) 103±112 www.elsevier.nl/locate/physc

Kinetics study of the Bi-2223 grain growth thickness V. Garnier *, I. Monot-La€ez, G. Desgardin Laboratoire CRISMAT-ISMRA, CNRS UMR 6508, 6 Bd. Mar echal Juin, 14050 Caen Cedex, France Received 16 April 2000; received in revised form 4 June 2000; accepted 4 July 2000

Abstract A kinetic study of the Bi-2223 phase formation mechanism has been done at temperatures ranging from 830°C to 865°C and for long sintering times (25±100 h), since the formation of Bi-2223 phase slows down. Above a sintering temperature of 850°C, the Bi-2223 decomposition begins to occur. The reaction order, n, depends on the sintering temperature; from 830°C to 850°C, n is roughly equal to 0.45 and the related mechanism corresponds to the thickening of plates after their edges have impinged. For temperatures higher than 850°C, n decreases to reach 0.13 at 865°C. Using the Johnson±Mehl±Avrami equation, an activation energy of 1437 kJ/mol has been found for temperatures between 835°C and 850°C where the partial melting becomes more signi®cant. This activation energy has been shown to correspond to the thickening of the Bi-2223 grain. Ó 2001 Elsevier Science B.V. All rights reserved.

1. Introduction With ®ve elements involved, the Bi±Pb±Sr±Ca± Cu±O system is complex, and its phase diagram is not well understood, even though a number of studies have been carried out in order to successfully determine the phase relations [1]. In this system, the high Tc phase (Bi-2223) possesses the best superconducting properties and is thus the most promising for applications. However, any mechanism proposed [2] for this phase formation is accepted unanimously: · A disproportionation reaction of the Bi(Pb)2212 phase to form the Bi-2223 and Bi-2201 phases [3,4].

* Corresponding author. Tel.: +33-2-31-45-29-15; fax: +33-231-95-1600. E-mail address: [email protected] (V. Garnier).

· An intercalation process in which the pre-existing Bi(Pb)-2212 crystals transform directly in Bi(Pb)-2223 platelets via the insertion of the CuO2 /Ca bilayers into the CuO2 /Ca/CuO2 blocks of 2212 [5,6]. · A Bi-2223 precipitation from a partially molten phase [7]. · Formation of Bi/Pb rich mobile liquid droplets which migrate over growing platelets [8]. · A dissolution±precipitation process where the Bi-2223 phase is formed from both the 2212 phase, CaO, and a liquid, the last two components arising from the decomposition of Ca2 PbO4 [9]. · A two-dimensional growth with decreasing nucleation rates. A liquid phase generated in the powder reacts with the 2212 matrix via the diffusion of Ca, Cu and Pb ions, resulting in the nucleation of the Bi-2223, which di€uses into the 2212 matrix, maintaining the outline of the original crystal shape [10±14].

0921-4534/01/$ - see front matter Ó 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 1 - 4 5 3 4 ( 0 0 ) 0 1 5 3 4 - 3

104

V. Garnier et al. / Physica C 349 (2001) 103±112

Whatever be the mechanism involved, the Bi2223 synthesis is time consuming and does not allow one to obtain a pure phase. In this study, we will discuss the di€erent possible mechanisms that may occur during sintering. Kinetics studies of the Bi-2223 formation have given dissimilar results. Indeed, di€erent reaction orders from 0.2 [15] to 2 [16] have been found. The large di€erence between these values leads us to be careful concerning their comparisons, and makes us wonder about their meanings. Di€erent powder precursor preparation methods (de®ned precursors [17], solid state [18], spray pyrolysis [19], liquidphase process [20,21]) result in di€erent phase equilibria with di€erent intermediates phases formation [22,23]. The same behavior di€erence is also observed for di€erent initial stoichiometries, even close [24,25]. The partial lead substitution [26], the presence of silver [27] and the oxygen partial pressure [28,29] also induce a change in the phases equilibrium. However, taking into account all these parameters, one can try to explain the meaning of what seems to be contradictory results for the reaction-order value. It appears fundamental to differentiate the reaction-order value obtained with di€erent temperatures and with di€erent periods of sintering time. It is clear that the kinetics study of the Bi-2223 formation cannot be strictly done if the reaction order is calculated without separating short and long sintering times. Indeed, the kinetics changes a lot as one considers the sintering beginning where the Bi-2223 formation is fast, and the sintering end where the Bi-2223 formation slows down. These last considerations may imply an evolutionary mechanism of the Bi-2223 formation during sintering and/or evolutionary conditions during sintering. Most of the previous studies use the whole sintering time to calculate the reaction order. Among the results obtained, Sung and Hellstrom [12] and Matsubara [10], using the powder-in-tube (PIT) process where the Bi-2223 phase forms inside Ag sheath of tapes, obtained, respectively, n ˆ 1 and n ˆ 1:3. The di€erence may be due to the process using Ag sheath and the reduced oxygen partial pressure sintering atmosphere which increases the partial melting and decreases the reaction order. The liquid-phase process preparation is considered to lead to a lower n value (0.4

[15]) than the solid-state method (0.93 [18]). Additional deduction would be hazardous. Nevertheless, reaction-order variations have been observed during sintering. Kanai et al. [30], Grivel and Fl ukiger [26] and Huang et al. [16] have determined a slope variation at 10, 12 and 8 h respectively, showing that one must be careful concerning the previous results. For their ®rst sintering part, they found n ˆ 1:5 [26±30] and n ˆ 2 [16], showing a two-dimensional nucleation and growth mechanism. For the second sintering part, the reaction order signi®cantly decreases, n ˆ 0:85 for 12 < t < 35 h [26]. A prolonged sintering time should further decrease this value. In this study, we are interested in the third part of the sintering process (25±100 h), when the slowest Bi2223 formation rate is reached. It is precisely with long sintering times that the Bi-2223 synthesis is achieved. This kinetics study will allow one to determine the formation mechanism and the activation energy corresponding to the permanent rate of the slowing down of the formation of Bi-2223. 2. Experimental The powder precursor was prepared by a liquidphase process, the polymer matrix method, detailed elsewhere [21]. The nominal composition Bi1:85 Pb0:35 Sr2 Ca2 Cu3:1 O10‡d (BPSCCO) was suggested by Maeda et al. [31] and is adopted by many other groups [32±34]. After ®ring on a hot plate, the precursor powder is crushed and hand milled in an agate mortar to be calcined at 820°C/24 h under air, milled again and pelletized (200 MPa, 16 mm diameter, 3 g) to be sintered under air for 25± 100 h at temperatures ranging from 830°C to 865°C. The microstructure of the sintered pellet was observed on the fractured part of the sample using a scanning electron microscope (SEM Philips XL30). The phases present in the sintered pellets were analyzed by X-ray di€raction powder (Philips PW 3710, kCu[Ka1]). The relative percentage evolution of the 2201, 2212 and 2223 phases during sintering was estimated with the peak area integration, but the secondary phases were not taken into consideration. This estimation has been possible thanks to the Schmahl's program [35] which enables us to take into account

V. Garnier et al. / Physica C 349 (2001) 103±112

105

several peak area ((0 0 10)2223 ; (1 1 5)2223 ; (0 0 6)2212 ; (1 1 5)2212 ; (0 0 6)2201 ; (1 1 5)2201 ) with their relative intensity and the preferential orientation phenomena of these di€erent phases. The X-ray diffraction method is widely employed for the 2223 phase estimation. However, some authors [14,30] prefer to determine the fractional conversion of the 2223 phase from susceptibility measurements. From our viewpoint, this method is not suciently accurate since the low-temperature susceptibility is strongly dependent on the grain size and on the grain morphology (demagnetizing factor). On the other hand, the X-ray di€raction method is generally used to estimate the weight fraction conversion of 2212 in 2223 simply by the ratio of the strong (0 0 l) or (1 1 5) di€raction intensities [16,27,29]. This simple procedure ignores the problem of preferred orientation of the plate-like BPSCCO powder, and does not assume the structure factor di€erence between the di€erent phase peaks [35]. Another problem commonly encountered concerns the (0 0 8) peak of the 2212 phase which is often used and which is superimposed by (1 1 1)2223 , (0 1 7)2223 and (1 1 1)2212 peaks, making it unsuitable as a reference peak. Moreover, the 2201 phase must also be taken into consideration, since this phase is all the more present as much as the temperature increases. Most of the works analyze the 2223 isothermal formation kinetics using the Avrami (Eq. (1)) [26,27] or the Johnson±Mehl±Avrami (Eq. (2)) [13,14] equations:

The experimental values of n correspond to particular Bi-2223 formation mechanism. Theoretical models have been established to determine the speci®c mechanism in conformity with typical n value. The Hulbert model [38] is very often used [14±16,18] and corresponds to a nucleation-growth model. We would not use this model considering that our study is conducted between 25 and 100 h sintering time, when the nucleation step is over, since this nucleation has been completed at the beginning of the sintering [10]. The Rao and Rao model [39], already used by Grivel and Fl ukiger [26], will be preferably employed to explain the Bi2223 phase formation, after the nucleation step, because this model correspond to a di€usion-controlled transformation. According to this model, the reaction order can take the following values:

C…t† ˆ 1 ÿ exp…ÿKtn †;

The results plotted in Fig. 1 show the Bi-2223 phase fraction rate evolution at di€erent temperatures ranging from 830°C to 865°C and at different quenching time ranging from 25 to 100 h. The Bi-2223 calculated ratio shows that sintering at 850°C leads to the largest amount of Bi-2223, 75% in 100 h without intermediate milling. A lower sintering temperature is less ecient to obtain rapidly the 2223 formation, and, on the other hand, a higher sintering temperature induces the 2223 decomposition [9] and element losses [9,40,41] specially Pb, resulting in the nominal composition changes. To study the kinetics of the 110 K phase formation, Eqs. (3) and (4) could be used according to Eqs. (1) and (2), respectively, as follows:

…1†

(Ref. [36]); n

C…t† ˆ 1 ÿ exp…ÿ…Kt† †;

…2†

(Ref. [37]) where C is the 2223 phase fraction after quench time t, n is the reaction order and K ˆ K0 exp…ÿ…Ea =RT †† is the Arrhenius rate constant. For both Eqs. (1) and (2) the plots of ln …ÿ ln …1 ÿ C…t††† against ln…t† give n as the slope value, but for the activation energy, Eq. (2) shows an intercept value of (n ln K) compared to (ln K) for Eq. (1), leading to appreciable di€erent Ea results for both graphs.

1. n ˆ 2:5 ± initial growth of particles nucleated at constant rate, 2. n ˆ 1:5 ± initial growth of particles nucleated only at the start of transformation, 3. n ˆ 1 ± growth of isolated plates or needles of ®nite size, 4. n ˆ 0:5 ± thickening of plates after their edges have impinged.

3. Results and discussion 3.1. Reaction order and associated mechanism

106

V. Garnier et al. / Physica C 349 (2001) 103±112

Fig. 1. Volume fraction of the Bi-2223 phase (%) as a function of the sintering time depending on the temperature.

ln ‰ÿln…1 ÿ C…t††Š ˆ n ln t ‡ lnK;

…3†

ln ‰ÿln…1 ÿ C…t††Š ˆ n ln t ‡ n lnK:

…4†

Fig. 2 shows the ln‰ÿ ln…1 ÿ C…t††Š versus ln t. The value of the reaction order n corresponds to the slopes, and of course no di€erence between Eqs. (3) and (4) exist for the n values. The straight lines obtained show a single rating for the considered time range (25±100 h) at a given temperature. If such straight lines are observed for T < 850°C, the lines become less well ®tted for temperatures above 850°C, meaning that the amount of 2223 phase reaches saturation earlier than in samples heat treated at lower temperatures. As shown in Fig. 3, the reaction order depends on the sintering temperature. From 830°C to 850°C, n is roughly equal to 0.45 on average, and decreases for higher sintering temperature to reach 0.13 for 865°C. This decrease in the n value is related to the Bi-2223 phase decomposition, or more precisely to the phase equilibrium change which

leads to a more and more signi®cant amount of the Bi-2201 phase at the expense of the Bi-2223 phase, starting at 850°C. The reaction order 0.45 is thought to correspond to ``the thickening of plates after their edges have impinged'', referring to the Rao model introduced in the Section 2. In fact, as shown in Fig. 4 with the SEM micrographs, the platelets are more and more embedded and entangled since the sintering time increases from 25 h (Fig. 4A and B) to 100 h (Fig. 4C and D), and so the platelets impingement on one another is more and more e€ective, leading to the slowing down (and even the stopping) of the (ab) plane growth. In this way, the grain growth would occur according to the c-axis direction. The structural similarity of the 2201, 2212 and 2223 phases suggests that the phase intergrowth of 2212 from 2201 and 2223 from 2212 through an intercalation di€usion of Ca and Cu atoms occur. Some evidence of this formation mechanism for the 2212 phase have already been proposed [42] considering that the thickness of the 2212 platelets increases with the heating time of about 25% on

V. Garnier et al. / Physica C 349 (2001) 103±112

Fig. 2. ln‰ÿ ln…1 ÿ C…t††Š versus ln…t† plots for temperatures ranging from 830°C to 865°C.

Fig. 3. Reaction order n depending on the temperature range.

107

108

V. Garnier et al. / Physica C 349 (2001) 103±112

Fig. 4. SEM micrographs of the fracture part of the pellets sintered for (A, B) 25 h at 850°C and (C, D) for 100 h at 850°C.

average from 25 to 100 h [43]. For the Bi-2223 phase, Bian et al. [5] showed that, at the early sintering stage, an intergrowth mechanism occurs from the 2212 phase. But during this ®rst step, the predominant mechanism is the two-dimensionalnucleation growth. The Hulbert model [38] is a more accurate model to describe what happens when Bi-2223 starts to form, because it is from the nucleation viewpoint. Studies [14,16] using this model, and others [10,12,26,44,45] have determined a reaction order between 1 and 2, corresponding to a two-dimensional-nucleation growth with decreasing nucleation rates. Kanai et al. [30] has also shown that addition of Bi-2223 nuclei decreases the reaction-order value, and so changes the nucleation growth mechanism. Therefore, since

the sintering time increases, the reaction order decreases [16,26,30]. In fact, at the beginning of sintering, small Bi-2212 platelets are present, resulting from the previous milling. These platelets have a bidimensional growth which leads to the decrease of the density [30], and at the same time, rapid 2223 nucleation growth occurs on the 2212 grains being formed. Thereafter, during sintering, the former 2212 grains, which become larger, are in each otherÕs way, blocking the (ab) plane growth. The 2223 phase closely follows the same grain impingement. This impediment to the 2223 expansion changes the 110 K phase formation mechanism, and results in a decrease of the reaction-order value. The slow intergrowth mechanism [18], which was occulted by the rapid two-dimen-

V. Garnier et al. / Physica C 349 (2001) 103±112

109

mechanism proposed by Morgan et al. [8], which could also explaine the thickness of the 2223 platelets via the sequential deposition of new material on the top of pre-existing crystals by the intermediate of a liquid phase. 3.2. Activation energy

Fig. 5. SEM microgaph showing the Bi-2223 grain thickness growth.

sional growth at the early sintering stage, becomes the single 2223 growth possibility, and the increase in the Bi-2223 grain thickness can occur for prolonged sintering times, as illustrated in Fig. 5. This photograph could also support the view of a

The activation energy of the Bi-2223 formation for 25±100 h sintering is obtained using either the modi®ed Avrami equation (3) or the modi®ed Johnson±Mehl±Avrami equation (4), both combined with the Arrhenius law explained in Section 2. Figs. 6 and 7 correspond to the ln K versus 103 /T plots using the Avrami and the Johnson±Mehl± Avrami equations, respectively, the activation energy, Ea , values being deduced from the slopes. In Fig. 6, it is shown that the ®tted line has a bad correlation factor ( 835°C and 1900 kJ/mol for T < 835°C). This di€erence results in a large sintering time range (0±200 h) used in Ref. [15] while di€erent mechanisms occur for early and long sintering. Therefore, these values must be an average of di€erent activation energy contributions. Con®rmation of a lower activation energy in the early sintering stage has been observed by Grivel et al. [26] at 17 h sintering time with 500 kJ/mol in air for T > 835°C, and by Luo et al. [11], at 17 h sintering time also, with 460 kJ/mol in 7.5% O2 for 819°C < T < 830°C. The temperature range difference between both these results is due to the reduced oxygen partial pressure sintering atmosphere which increases the partial melting and consequently decreases the required sintering temperature. The activation energy transition temperature under 7.5% O2 is then 819°C according to Luo et al. [11] (Ea ˆ 2100 kJ/mol for T < 819°C) and 825°C according to Sung and Hellstrom [12] (Ea ˆ 2900 kJ/mol for T < 825°C and Ea ˆ 890 kJ/ mol for T > 825°C). In the same way, the higher Ea value (890 kJ/mol) obtained by Sung et al. is due to the large sintering time range (0±96 h) used for the calculation, which included di€erent activation energy contributions. Generally speaking, one can say that, without observing the activation energy transition temperature and considering short and long sintering times together, the Ea values averaged 1513 kJ/mol [13], 1900 kJ/mol [14] and 1904 kJ/mol [18]. Our study was conducted at long sintering time, and so higher Ea values were expected. Therefore, a value of 4700 kJ/mol was obtained below 835°C, where a sluggish Bi-2223 phase formation occurs since only little partial melting exists. On the other hand, above 835°C, and still by considering long sintering times, the Ea value decreases to 1437 kJ/mol due to the melt formation. We have shown in Section 3.1 that the mechanism of occurrence was the thickening of plates after their edges had impinged. Consequently, 1437 kJ/mol may correspond to the thickness growth of the Bi-2223 grain. Danusantoso and Chaki [43] has shown that the Bi-2212 platelet thickness follows an activation energy of 80 kJ/mol. However, the 2212 grain growth exhibits an activation energy of 22.1 kJ/mol, imply-

111

ing that the Ea value is 3.62 times larger for the thickness growth compared to the (ab) plane natural growth. Considering the Ea value obtained by Luo et al. [11] (460 kJ/mol for the Bi-2223 grain growth at early sintering time, corresponding to the two-dimensional-nucleation and growth mechanism) and our value of 1437 kJ/mol, the corresponding multiplying factor is about 3.12. The correspondence between the 2212 and the 2223 phase formation concerning the activation energy ratio for the thickness growth and the (ab) plane growth is rather correct, assuming all the more that the two-dimensional growth could still occur and, can consequently decrease our Ea value.

4. Conclusion Using the Johnson±Mehl±Avrami equation, a kinetics study of the Bi-2223 phase formation mechanism was conducted. The calcined powder was milled, pelletized and sintered in air for 25± 100 h at temperatures ranging from 830°C to 865°C. Our study was conducted when the slowing down of the Bi-2223 phase formation occurs. The reaction order (n) has been shown to depend on the sintering temperature, from 830°C to 850°C, n is roughly equal to 0.45 and decreases for higher sintering temperatures to reach 0.13 for 865°C, this decrease being related to the Bi-2223 phase decomposition. The value 0.45 corresponds to the thickening of the plate grains after their edges have impinged. This impediment to the Bi-2223 grain (ab) plane expansion for prolonged sintering times, results in the increase of the Bi-2223 grain thickness. The activation energy of this growth mechanism corresponds to 1437 kJ/mol for temperatures ranging from 835°C to 850°C as the partial melting increases.

Acknowledgements The authors would like to acknowledge J. Lecourt for his help in sample preparation and Dr. M. Korzenski for the fruitful discussions.

112

V. Garnier et al. / Physica C 349 (2001) 103±112

References [1] K. Majewski, Supercond. Sci. Technol. 10 (1997) 453. [2] S. Nhien, Ph.D. Thesis, University of Caen, France, 1996. [3] H. Nobumasa, K. Shimizu, Y. Kitano, T. Kawai, Jpn. J. Appl. Phys. 27 (1988) L846. [4] N. Kijima, H. Endo, J. Tsuchiya, A. Sumiyama, M. Mizumo, Y. Oguri, Jpn. J. Appl. Phys. 27 (1988) L1852. [5] W. Bian, Z. Yimei, Y.L. Wang, M. Suenaga, Physica C 248 (1995) 119. [6] Z.X. Cai, Y. Zhu, D.O. Welch, Phys. Rev. B 52/17 (1995) 13035. [7] T. Hatano, K. Aota, S. Ikeda, K. Nakamura, K. Ogawa, Jpn. J. Appl. Phys. 27 (1988) L2055. [8] P.E.D. Morgan, R.M. Housley, J.R. Porter, J.J. Ratto, Physica C 176 (1991) 279. [9] Y.L. Chen, R. Stevens, J. Am. Ceram. Soc. 75 (1992) 1142. [10] Y.C. Guo, H.K. Liu, S.X. Dou, Physica C 235±240 (1994) 1231. [11] J.S. Luo, W. Merchant, V.A. Maroni, D.M. Gruen, B.S. Tani, W.L. Carter, G.N. Riley Jr., Appl. Supercond. 1 (1993) 101. [12] Y.S. Sung, E.E. Hellstrom, J. Am. Ceram. Soc. 78 (1995) 2003. [13] W. Zhu, P.S. Nicholson, J. Mater. Res. 7 (1992) 38. [14] I. Matsubara, R. Funahashi, T. Ogura, H. Yamashita, Y. Uzawa, K. Tanizoe, T. Kawai, Physica C 218 (1993) 181. [15] V. Rouessac, S. Nhien, J. Wang, G. Desgardin, Physica C 282±287 (Part 2) (1997) 511. [16] Y.T. Huang, D.S. Shy, L.J. Chen, Physica C 294 (1998) 140. [17] A. Sotelo, P. Majewsky, H.-S. Park, F. Aldinger, Physica C 272 (1996) 115. [18] S. Nhien, G. Desgardin, Physica C 272 (1996) 309. [19] I. Van Driessche, R. Mouton, S. Hoste, Mater. Res. Bull. 31 (1996) 979. [20] C.Y. Shei, R.S. Liu, C.T. Chang, P.T. Wu, Mater. Lett. 9 (1990) 105. [21] V. Garnier, I. Monot, G. Desgardin, Supercond. Sci. Technol. 13 (2000) 602. [22] S. Tirumala, D.F. Lee, D.M. Kroeger, K. Salama, Supercond. Sci. Technol. 10 (1997) 686. [23] M.G. Smith, D.S. Phillips, D.E. Peterson, J.O. Willis, Physica C 224 (1994) 168.

[24] A. Sotelo, L.A. Angurel, M.T. Ruiz, A. Larrea, F. Lera, G.F. de la Fuente, Solid State Ionics 63±65 (1993) 883. [25] H. Liu, L. Liu, H. Yu, Y. Zhang, Z. Jin, J. Mater. Sci. 34 (1999) 4329. [26] J.-C. Grivel, R. Fl ukiger, J. Alloys Compds. 235 (1996) 53. [27] Y.D. Chiu, C.H. Kao, T.S. Lei, M.K. Wu, Physica C 235± 240 (1994) 485. [28] K. Aota, H. Hattori, T. Hatano, K. Nakamura, K. Ogawa, Jpn. J. Appl. Phys. 28 (1989) L2196. [29] Y.E. High, Y. Feng, Y.S. Sung, E.E. Hellstrom, D.C. Larbalestier, Physica C 220 (1994) 81. [30] T. Kanai, T. Kamo, S.P. Matsuda, Jpn. J. Appl. Phys. 28 (1989) L2188. [31] A. Maeda, K. Noda, K. Uchinokura, S. Tanaka, Jpn. J. Appl. Phys. 28 (1989) L576. [32] B.-S. Hong, T.O. Mason, C.K. Chiang, S.W. Freiman, N.M. Huang, Appl. Supercond. 1 (1993) 109. [33] H. Sasakura, S. Minamigawa, K. Nakahigashi, M. Kogachi, S. Nakanishi, N. Fukuoka, M. Yoshikawa, S. Noguchi, K. Okuda, A. Yanase, Jpn. J. Appl. Phys. 28 (1989) L1163. [34] J. Wang, M. Wakata, T. Kaneko, S. Takano, H. Yamauchi, Physica C 208 (1993) 323. [35] W.W. Schmahl, M. Lehmann, S. Rath, M. Gerards, R. Riddle, Supercond. Sci. Technol. 11 (1998) 1269. [36] M. Avrami, J. Chem. Phys. 7 (1939) 1103. [37] W.A. Johnson, R.F. Mehl, Trans. AIME 135 (1939) 416. [38] S.F. Hulbert, J. Br. Ceram. Soc. 6 (1969) 11. [39] C.N.R. Rao, K.J. Rao, Phase transitions in Solids, McGraw-Hill, 1978, p. 93. [40] L.C. Pathak, S.K. Mishra, D. Bhattacharya, K.L. Chopra, J. Mater. Sci. 34 (1999) 1619. [41] T. Sata, K. Sakai, S. Tashiro, J. Am. Ceram. Soc. 75 (1992) 805. [42] A. Sotelo, H. Szillat, P. Majewski, F. Aldinger, Supercond. Sci. Technol. 10 (1997) 717. [43] J. Danusantoso, T.K. Chaki, Supercond. Sci. Technol. 4 (1991) 509. [44] C.H. Kao, Y.D. Chiu, S.L. Dung, L.C. Lee, T.S. Lei, M.K. Wu, Mater. Lett. 20 (1994) 83. [45] Q.Y. Hu, H.K. Liu, S.X. Dou, Physica C 250 (1995) 7. [46] V. Rouessac, Ph.D. Thesis, University of Caen, France, 1997. [47] L. Yalu, L. Chengren, Z. Lian, Physica C 235±240 (1994) 487.