PhysicaC213 (1993) 17-25 North-Holland Phase diagram of the T12Ba2Cu06 compounds in the zyxwvutsrqponmlkjihgfedcb T, p( 0,) plane C. Opagiste a*b,G. T&one b, M. Couach a, T.K. Jondo ‘, J.-L. Jorda ‘, A. Junod b, A.F. Khoder a and J. Muller b ’ CEA/DRFM C/SPSM S/LCP, BP 85 X, F- 38041 Grenoble Cedex, France b Dkpartement de Phy sique de la M at&e Condenske, 24 Quai E.-Ansermet. UniversitP de Gen&ve, CH- 1211 GenPve 4, Switzerland ’ Laboratoire de Phy sico- chimie M in&ale II, VniversitC Claude Bernard Ly on I, F- 69622 Villeurbanne Cedex, France Received 27 April 1993 We present a detailed investigation of the structural and superconducting properties of TlrBa$ZuO, compounds as a function of oxygen pressure and annealing temperature prior to quenching. The crystal structure depends both on the oxygen concentration and on thallium deficiency. At the stoichiometric cation composition zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPO Tl : Ba: Cu = 2 : 2 : 1, the parameters p( 0,) and Tqumch uniquely define the structure and the superconducting T,. With adequate heat treatment sequences, reversible transitions from the tetragonal to the orthorhombic phase have been observed. 1. Introduction In the Tl,BaJZa,,_ ,CU,O~+~,, series of thallium superconducting phases, the n = 1 Tl,Ba,CuO, (Tl2201) material is particularly interesting. The crystallographic structure of this compound is either orthorhombic (Ccc2) [ 1 ] ) or tetragonal (14/mmm [ 2,3 ] ). Both phases are superconducting [ 4,5 ] with Tc’s ranging from 0 to 92 K. For thallium deficient samples, only the tetragonal phase can be synthesized. The difference between the two structures is subtle. Both contain T1202 double layers and a CuOZ plane separated by a layer of barium atoms. The small difference resides in the T1202 double slabs which are poorly defined; the positions of the atoms belonging to this layer have been refined with very anisotropic temperature factors and/or with a static displacement from their ideal sites. Superconductivity may depend on the occupancy of the interstitial oxygen O(4) in the thallium layers [ 21. Synthesizing thallium compounds at ambient pressure is a difficult task because of the volatility of this metal. Often in the literature, the structure of the Tl-220 1 superconducting samples prepared at ambient pressure is found to be tetragonal. Since the orthorhombic phase is stoichiometric in the thallium 0921-4534/93/$06.00 composition, this result is probably related to a loss of thallium during the synthesis. The thallium content being then lower than 2, it is not possible, without adding T1203, to obtain the stoichiometric orthorhombic phase. In such a preparation process, the samples produced are not in an equilibrium state and increasing the annealing time will result in completely decomposed samples. To avoid thallium losses a high pressure process has been developed, allowing one to work at much higher reaction temperatures [ 3,6]. The advantage is obvious particularly in view of an improved microstructure. 2. Sample preparation 2.1. Sy nthesis of the T12Ba2Cu06 compounds To prepare the 2201 samples, we used as starting components T12Baz05, a congruent melting oxide [ 7,8 1, and CuO powders. Two different batches were prepared from a stoichiometric amount of powders. For batch A the powders were first kept for three days at 750” C under 1 bar of oxygen, whereas batch B was not pre-treated. After this process, batch A contained the TlzBazCu,06 (Tl-220 1) phase and the two 0 1993 Elsevier Science Publishers B.V. All rights reserved. 18 C. Opagiste et al. / T12Ba2Cu0, phase diagram starting powders ( T1zBa205 and CuO). Plasma emission spectroscopy indicates no variation of the overall cation stoichiometry. Under 1 bar of oxygen, the thallium losses appear only at 800-850°C. The advantage of using the T12Ba205 precursor rather than directly T1203, BaO and CuO is to avoid the formation of BaCuO, known for its magnetic CurieWeiss behavior. The powders were then pressed into pellets under 5 kbar and reacted in a high pressure, high frequency furnace. Finally the samples were quenched at an initial rate of about SOO”C/min. Firing at 850°C under 100 bar of argon or helium during 12 min yields tetragonal superconducting samples (denoted TS in the following) with a critical temperature near 90 K. Using 100 bar of oxygen and a temperature of 930°C during 30 min, the samples are orthorhombic and non-superconducting (denoted ONS). The weight loss amounts to 1.5-2% and 0.8% for the TS and ONS samples, respectively, and is attributed to the metallic elements and to the change in oxygen content [ 3,6 1. For both high pressure treatments resulting in TS or ONS samples, we find that melting occurs at higher temperatures; i.e. > 850°C under argon or helium and > 930°C under oxygen. We also recall the striking fact that the TS samples thus prepared have an average oxygen concentration below 6 per formula unit [3]. 2.2. Variation of the oxy gen concentration In order to change and homogenize the oxygen concentration, we annealed tetragonal samples (T,= 90 K) at temperatures between 350 and 700°C and under an oxygen pressure ranging from 10p6 to 1 bar. The equilibrium oxygen content is then frozen by quenching the samples into liquid gallium. One problem with this technique is that traces of gallium around the samples render mass measurements after the quench rather worthless. At high pressure with partial oxygen pressure 100 bar), it was unfortu(P tota’=p(02) +p(He)= nately not possible to obtain sharp superconducting transitions. This is probably due to insufficient final cooling rates (about SOO”C/min (initial) ) responsible for macroscopic oxygen inhomogeneities in the samples. The sample codes and all relevant data are listed in table 1. Figure 1 is a representation diagram in the T, p (0,) plane. of the phase 3. X-ray diffraction, micrographic and microprobe analysis and superconducting transitions Microstructural investigations were carried out after polishing the samples using diamond paste. The surface was examined in a light microscope and in a scanning electron microscope (SEM, Cambridge Instruments Stereoscan 360). The crystal structure was studied at ambient temperature by X-ray diffraction in a Guinier camera with Cu Ku radiation. Silicon was added to the powdered samples as an internal standard (a= 5.4308 A). The lattice parameters of the T12Ba2CUG6+6 orthorhombic (Ccc2 or A2aa [ 1 ] ) and tetragonal (14/mmm [ 2,3] ) phases were determined from least-square tits using more than 20 lines ( 10” < 48~ 160” ). All X-ray diffraction patterns presented in this paper contain the peaks of the standard and two camera lines at 48~84.84 and 98.83”. They are also sensitive to the preferential grain orientation resulting from the preparation of the X-ray sample. The Meissner effect (field cooling) was measured by means of a RF SQUID magnetometer with an external magnetic field of 20 Oe. The Meissner flux expulsion ratio f= - 4rcx,,was evaluated with an effective sample volume given by m/p, where m is the sample mass and p the X-ray density ( z 8.0 g/cm3). A geometric demagnetization factor D, depending on the sample shape, was taken into account. We use the cgs system where B=H+4nM and xU=M IH=p& The corrected susceptibility reads x, =x,“/ ( l4rrDpxF ), where x,” is the measured susceptibility. The external magnetic field was calibrated with a high purity Pb sphere in the superconducting state. 3.1. Sample characterization after sy nthesis 3.1.1. Samples from batch A (preannealed powders) ONS and TS samples are X-ray pure. EDAX in- vestigations show that the average and local stoichiometries are the same and exclude an excess of thallium at the grain boundaries. Some traces of CuO and BazCu,O, impurities are detected; both are estimated to be < 0.1% by volume. Optical micro- Sample T-2-550 T-4-550 T-55-550 T-3T-5-600 T-2450 T-1-700 T-3-700 TOE= A A TlI24SOc~ B ITS-HeI T-5diOObl B e, 600 1 TS-He I 930 I loo 930 1 100 300 10m2 1 2days 02 1 30’ 02 I 30’ 07 7 days quench 1 +0.31 -0.27 quench 174.9 130.810.21114/mmm( lol- 0 - - 14/mmm - Ckc2 b8 (a.c. mceptiity)] a=b=3.8658(6) 123.231(4)1 0 f= a=b=3.8616(5) 23.130(3) I 0 ;; t 0.81 % A 15.4471(12)15.4915(12)~ 23.137(4) 1 Ccc2 5.4518(6) 5.4899(7) 23.194(3) 0.70 Ccc2 5.4502(7) 5.4870(5) 23.192(2) 0.67 Cd? 5.4563(14)5.4959(10) 23.163(4) 0.72 Ccc2 5.45700 5.4783(10) 23.198(2) 0.39 Ccc2 5.4474(10)5.4954(11) 23.135(4) 0.88 Gx2 5.4597(9) 5.4823(9) 23.219(3) 0.41 14/mmm1 a=b=3.8652(6) 123.185(4) 1 0 21 C. Opagiste et al. /zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJI TlzBa2Cu06 phasediagram 800 700 600 500 400 300 200 1 10” 1 10-3 ~(0,) 10’ (atm.) Fig. 1. Phase diagram in the T, ~(0~) plane. The full and open symbols represent the superconducting and non-superconducting samples, respectively. The numbers are the critical temperature defined as the onset of the field cooling superconducting transition. The tetragonal and orthorhombic samples are distinguished by squares and circles, respectively. The light lines are iso-T. curves. graphs show only Tl-2201 grains in both phases as can be seen in figs. 2 (a) and (b). The stoichiometry determined by various methods of analysis including plasma emission spectroscopy and X-ray diffraction refinement of a tetragonal superconducting single crystal is Tl: Ba: Cu= 2 : 2: 1for both structures [ 3,6]. Figures 3(a) and (b) show the X-ray diffraction patterns for the highest and lowest oxygen concentrations, i.e. after the synthesis under high pressure in pure oxygen (ONS, sample TKO-930) and in pure helium (TS, sample TA2-850). The splitting of e.g. the tetragonal line (1 1 0) at 48=65.49” into (0 2 0) and (2 0 0) or of the tetragonal line ( 1 1 10) at 4~9=103.24” into (02 10) and (20 10) are clearly resolved. The maximum difference in the oxygen content between the ONS and TS samples has been estimated to amount to 0.4 per formula unit [ 3 1. 3.1.2. Samples from batch B (powders without pretreatment) Optical micrographs show in polarized light some yellow inclusions ( < 5% by volume) of the unreacted T12Baz05 precursor and small CuO grains. For these samples, T12Ba205 impurities are sometimes also detected in the X-ray diffraction pattern. In contrast to the orthorhombic structure, previous Fig. 2. Optical micrographs of as-synthesised non-superconducting orthorhombic (TKO-930), and superconducting tetragonal (TA2-850) samples. The third sample T-4-350 was annealed at 350°C (section 3.2). 22 C. Opagiste et al. / TlzBa,CuO, phase diagram perconductivity. In the latter, thallium losses of the order of lOoh were measured by EDAX. Superconccc2 ductivity could be restored by a second low ~(0,) non-supemnductmg zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA annealing, with slightly lower T, and lower Meissner fraction 1: Figure 4 illustrates the degraded microstructure with CuO, BaO, and BaCuO, around the eroded Tl-220 1 grains. TKO-930 50 60 70 80 90 100 110 120 40 Fig. 3. X-ray diffraction patterns of the samples corresponding to fig. 2. investigations [ 3,6,9] indicated the possibility of some homogeneity domain (Tl, +Cu,) in the tetragonal structure. 3.2.2. Equilibrium state at T< 600°C A number of experiments were devised to ascertain the equilibrium nature of compounds with cation stoichiometry Tl: Ba : Cu = 2 : 2 : 1 heat treated at low enough temperature. We refer again to the data of table 1. First, two starting samples with different symmetry, i.e. ONS and TS, were annealed for a long time at 350°C in the 10e5 bar range of oxygen pressure. Figure 5 demonstrates that both specimens are 3.2. Effect of post-annealing on structure and superconductivity As a general rule, irrespective of the initial symmetry of a given specimen, annealing at temperatures below 600°C does not alter the microstructural aspect of the 220 1 phases and leaves the cation stoichiometry unchanged. The only exception was found in samples of batch B containing some TlzBaz05. The latter inclusions, at ~(0~) below 10m2 bar, change color under polarized light, indicating a change of the Tl oxidation state. 3.2.1. Annealing at temperatures T2 600” C In this temperature range and at low oxygen pressure ( < low2 bar), the thallium losses are no longer negligible. Very long annealing times even lead to decomposition to 2201 into TlzBa20,, CuO, BaO, BaCuO, and T&OS?. In order to document the irreversible evolution due to heat treatment under low p(OZ), we performed several experiments, notably those listed in table 1 under codes TH3-850a through d. In short, part of a TS sample as synthesized under 100 bar He was subjected to annealing at 600°C under low4 bar of OZ. Both parts were then briefly exposed to 100 bar O2 at 930°C. The unannealed part changed its structure into orthorhombic, as expected, whereas the annealed part remained tetragonal but also lost its su- Fig. 4. Optical micrographs of partly (above) and totally (below) decomposed 2201 compounds, see text. 23 C. Opagiste et al. / Tl~BazCu06 phase diagram u50 60 70 80 90 100 110 120 40 Fig. 5. X-ray diffraction patterns of an initially orthorhombic (A) and an initially tetragonal (B) compound heat treated under equilibrium conditions as described in section 3.2.2. orthorhombic and superconducting, the slight difference in T, being explained by the somewhat different ~(0,). We conclude that the final state does not depend on the initial symmetry. A series of heat treatments similar to that described in section 3.2.1 was performed at 550°C lo-’ bar O2 (orthorhombic after annealing, T,=52.4 K) with an intermediate excursion to 930°C 100 bar O2 (table 1: codes TH2-850a through d). Figure 6 illustrates the states observed successively and proves that after transformation to a highly oxidized configuration, the superconducting T, is perfectly recovered, even without a degradation of the Meissner fraction. Note that a similar treatment but applied at 500°C 2x 10m5 bar O2 this time results in a tetragonal specimen ( T,= 7 1.9 K), again showing the reversible nature of the transformations. 3.3. Symmetry, lattice parameters and T, Figure 7(a) depicts the X-ray diffraction pattern of a TS sample (TKZR, batch A, T,z 90 K) annealed under 1 bar of oxygen at 350°C. Without any 50 60 70 80 90 loo II0 120 40 Fig. 6. Evolution of the structure and the T, in a sequence of equilibrium states, see text. lOOO ,x MN) 'g 600 2 4oO 2 200 0 1000 ,x ~00 '3 600 3 400 5 200 n "50 m 70 80 90 100 II0 I20 40 addition of thallium, the symmetry has changed to orthorhomic and no superconductivity is observed. Fig. 7. X-ray diffraction patterns of two compounds with perfect cation stoichiometry, reversibly transformed from tetragonal to orthorhombic (A) and from orthorhombic to tetragonal (B), respectively. We specify that this sample does not contain Tl,Ca205 or TlzO, impurities before the second treatment (batch A) and Rietveld refinements exclude the possibility of a two-phase orthorhombictetragonal structure after annealing. Figure 7 (b) represents the pattern of an ONS sample (TOE-840 ) annealed under 100 bar of helium at 840 oC (TOE- 840b). The structure has now changed to tetragonal and the sample is superconducting with a critical temperature close to 90 K, as for samples directly synthesized under inert gas. All our experiments show that in the absence of a modification of the cation stoichiometry (as verified by EDAX), the tetrago- zyxwvut 24 C. Opagisle et al. / TlzBatCuOs phase diagram nal+-+orthorhombic transition is reversible. A physically important question is that of the correlation of T, with structural parameters. The data of table 1 indicate that no such correlation exists with the in-plane lattice constants, including the degree of orthorhombicity. However, although the variation of the c-axis is very small, the extreme values observed differing by only 0.4%, the latter does appear to scale with T,, as illustrated in fig. 8. An interesting point is that the correlation is sensitive neither to the structure (tetragonal and orthorhombic) nor to a small thallium deficiency. A comparison of the measured c-parameter data with the equilibrium position in the phase diagram (fig. 1) yields an increase of c with decreasing average oxygen concentration. The trend is the same as found for YBa2Cu307_-xbut the analogy is not really relevant because in the latter case, only the chain sites allow for a variable occupation. Another correlation found in the present work is the noticeable decrease of the Meissner fraction with decreasing T, (fig. 9). This observation seems to be quite general for high-& systems (e.g. Bi,Sr,CaCuzO, [ lo], Y2Ba4Cu,0x [ 111, etc.) although the explanation is not straightforward. We finally wish to comment on a few particular points concerning superconductivity in the Tl-2201 series. Let us first consider tetragonal compounds with low T,. At the correct cation stoichiometry, the stable phase at low temperature is orthorhombic. In order to obtain a tetragonal superconductor with T,, 100 ‘I’I~I~I~I~,‘,‘,.,~ 80 - 23.16 23.19 -0.1 -0.2 x’ c -0.3 ti -0.4 -0.5 -0.6 0 20 40 60 80 100 Temperature zyxwvutsrqponmlkjihgf (K) Fig. 9. Field cooling susceptibility (Meissner effect) at H= 20 Oe. Ideal diamagnetism corresponds to 4%x”=- 1in these units. Symbols as in fig. 8. say, below 40 K, one has to anneal thallium deficient samples. Typically, we mention a TS sample treated at 600” C under 10e5 bar O2 for two days to reduce the Tl content. The next step was annealing at 300’ C under lo-* bar O2 for 7 days. The result was a tetragonal specimen (table 1: T-5-600b) with T,= 28 K. Without the first annealing at high temperature under low pressure, a similar T, is obtained but in the orthorhombic structure. Second, to obtain high-T, orthorhombic samples, one has to use a high inert gas pressure with a low partial oxygen pressure at high annealing temperature, followed by rapid quenching. This procedure was not optimized in the present work. Annealing e.g. a TS specimen at low temperature under a pure helium flow apparently does not lead to an equilibrium state as shown by an incipient decomposition. At 350°C a very minor mass loss is detected, but T, is diminished from 90 to 30 K. We suspect an oxygen rearrangement in the structure that depends on the annealing temperature. A low temperature heat treatment also modifies the environment of part of the copper sites, as clearly shown by the normal state susceptibility (fig. 10). The CurieWeiss contribution measured after the treatment corresponds to 1.5Ohof the Cu2+ moments. 23.22 c (A> Fig. 8. Correlation between T, and the c-axis lattice parameter. Squares and circles identify tetragonal and orthorhombic samples. 4. Conclusions We have clarified the essential features of the phase C. Opagiste et al. / TlzBatCuO, phase diagram 25 Acknowledgements The authors are grateful to J.A. Femandez, F. Lin- GJ2 iger and A. Naula for their technical assistance. This 3 E work was supported partially by the program EUI RODOC and the Fonds National Suisse de la Re7 a, zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA 0 V x” cherche Scientifique. 0 -I zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA 0 100 200 300 400 References Temperature (K) Fig. 10. Normal-state susceptibility vs. temperature at H=20 kOe of a tetragonal sample (squares), finally driven orthorhombic (circles), under pure helium flow at 350°C. Inset: field cooling Meissner transitions at H= 20 Oe. diagram of the T12Ba2Cu0, compounds in the T, plane. At perfect cation stoichiometry, either the tetragonal or the orthorhombic structure exists, according to the oxygen content. With increasing thallium deficiency, (Tl,_,Cu,,),Ba,CuO, samples only exhibit the tetragonal structure. Without Tl deficiency and at low oxygen pressure ( < 10e3 bar), the phase transition occurs between 350°C and 430°C and is reversible. Quenching allows one to preserve the high temperature tetragonal structure, indicating that the phase transition is most likely of first order. Regardless of the structural symmetry, the superconducting critical temperature (0 < T, < 92 K) only depends on the average oxygen concentration. The latter mainly affects the c-axis and a clear correlation between T, and c is observed. ~(0~) [ 1 ] A.W. Hewat, P. Bordet, J.J. Capponi, C. Chaillout, J. Chenevas, M. Godinho, E.A. Hewat, J.L. Hodeau and M. Marezio, Physica C 156 ( 1988) 369. [2] Y. Shimakawa, Y. Kubo, T. Manako, H. Igarashi, F. Izumi and H. Asano, Phys. Rev. B 42 (1990) 10165. [ 31 C. Opagiste, M. Couach, A.F. Khoder, R. Abraham, T.K. Jondo, J.-L. Jorda, M.Th. Cohen-Adad, A. Junod, G. Triscone and J. Muller, Alloys Compounds, E-MRS Strasbourg 1992, to be published. ]4] Y. Shimakawa, Y. Kubo, T. Manako and H. Igarashi, PhysicaC 185-189 (1991) 639. [ 51 Y. Shimakawa, Physica C 204 ( 1993) 247. [6] C. Opagiste, M. Couach, A.F. Khoder, T. Graf, A. Junod, G. TRiscone, J. Muller, T.K. Jondo, J.-L. Jorda, R. Abraham, M.Th. Cohen-Adad, L.A. But-sill, 0. Leckel and M.G. Blanchin, Physica C 205 ( 1993) 247. [ 71 T.K. Jondo, R. Abraham, M.T. Cohen-Adad, J.-L. Jorda, J. Alloys Compounds 186 ( 1992) 347. [ 81 J.-L. Jorda, T.K. Jondo, R. Abraham, M.T. Cohen-Adad, C. Opagiste, M. Couach, A.F. Khoder and F. Sibieude, Physica C205 (1993) 177. I9lR.S. Liu, S.D. Hughes, R.J. Angel, T.P. Hackwell, A.P. Mackenzie and P.P. Edwards, Physica C 198 ( 1992) 203. [ 10 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLK ] G. Triscone, J.-Y. Genoud, T. Graf, A. Junod and J. Muller, Physica C 176 ( 1991) 247. [ I I] G. Triscone, J.-Y. Genoud, T. Graf, A. Junod and J. Muller, PhysicaC 201 (1992) 1.