Oxygen excess and superconductivity at 45 K in La2CaCu2O6+y

Physica C 170 (1990) 153-160 North-Holland

Oxygen excess and superconductivity at 45 K in

La2CaCu206+y

A. Fuertes, X. O b r a d o r s , J.M. N a v a r r o , P. G o m e z - R o m e r o , N. Ca,sail-Pastor, F. P~rez, J. F o n t c u b e r t a , C. Miravitlles, J. R o d r i g u e z - C a r v a j a l a a n d B. M a r t t n e z Institut de Cibncia de Materials, CSIC, Marti i Franqu~s s/n, 08028 Barcelona, Spain i lnstitut Laue Langevin, BP 156X, 38042 Grenoble, France

Received 22 June 1990 Revised manuscript received 13 July 1990

Preparation of single-phase stoiehiomctric La2CaCu206+yfrom oxide precursors is reported along with high-resolution neutron powder diffraction studies. To date this is the only route that allows the stoichiometric phase to be obtained. An air-heated sample having y=0.0378 (8) displays a transition onset at 45 K to diamagnetic susceptibility. Nevertheless, the maximum amount of superconducting phase inferred from these flux exclusion experiments is only 1% in volume. It is also found that the diamagnetic signal is not substantially modified by changing the annealing atmosphere at normal pressures. Neutron diffraction data show a high atomic ordering of La and Ca ions with a strong preference (75%) of Ca ions for sites eight-fold coordinated located between the Cu--O2 planes. The other 25% is occupied by La ions, around which the excess oxygen is located with partial occupancies, yielding a higher coordination number for some of these La ions. Comparison of this structure with that of the nonsuperconducting oxide LaL9CamCu206+, suggests that the observed small superconductivity islands are related to clustered oxygen excess intercalated between the two Cu-O2 planes, along with the La ions. The small overall concentration of defects observed here, and thus the small number of holes, is responsible for the absence of bulk superconductivity in La2CaCu206+r

1. Introduction As has already been pointed out by several authors [ 1 - 4 ] , the nonobservance o f any superconducting behavior in two-dimensional LaE_xMl+xCu206+y (M = Sr, Ca) oxides appears very puzzling according to the crystallochemical correlations systematically found for high-temperature superconductors. These phases are the n = 2 members o f the so-called Ruddlesden-Popper series ( A C u O s ) n ( A ' O ) [5 ] with A = La, Sr or Ca, corresponding to an intergrowth o f single rock-salt-type layers with double perovskite layers. As in the high-To superconducting cuprate oxides, copper displays square pyramidal coordination in CuO2 planes which are separated by La, Sr or Ca ions. Moreover, the same structural relationship exists a m o n g these c o m p o u n d s and the superconducting La2_~SrxCuO4+y that exists between the onecopper-layer and the two-copper-layer superconducting oxides in the series (Bi, Tl)m(Sr, Ba)2Can_lCu~O2n+4. In the case o f Sr oxides, La2_xSrl+xCu206+~,, an

oxygen excess o f up to y = 0 . 2 was obtained by high oxygen pressure annealing [ 1 ]. Even though metallic behavior was induced with this treatment, no superconducting behavior was detected. This is not, however, a p r o o f o f the nonexistence o f superconductivity in the LaEMCu206+ysystem, since strong modifications in interatomic distances are observed when comparing the Sr and the Ca oxide, and the latter has not yet been doped adequately. Several authors have reported [ 6 - 8 ] the impossibility o f preparing the stoichiometric Ca oxide, La2CaCu206, and only an oxide with excess Ca, Lal.9Cal.iCu206, has been investigated to date. This phase could not accept additional oxygen even under oxygen pressure (40 atm) and displays semimetallic behavior [ 6 ]. We show here a method o f synthesis that allows the preparation o f the stoichiometric phase, a relevant phase since it proves that superconductivity may also exist for the n = 2 m e m b e r o f the (ACuO3)nAO series.

0921-4534/90/$03.50 © 1990 - Elsevier Science Publishers B.V. (North-Holland)

154

A. Fuertes et al. / Oxygen excess and superconductiviO, in Za2CaCu206+ r

2. Experimental

La2CaCu206+x was prepared using an oxide precursor route. In a first step, pure L a 2 C u O 4 w a s obtained by heating in air at 1100°C a 1 : 1 mixture of La203 and CuO (99.99% Aldrich) with an intermediate regrinding. Then, a 1 : 1 : 1 molar ratio of La2CuO4, CaCO3 (Baker 99.995%) and CuO were subjected to a treatment involving heating in air at 900°C for 12 h and further reheating at 1070°C with several intermediate regrindings. Finally the sample was furnace-cooled to room temperature. Portions of the sample were subjected to annealing in O2 ( P = 1 arm) for 48 h at 900°C and in Ar ( P = 1 atm) for 6 h at 900 ° C. [ C u - O ] global oxidation states were determined iodometrically according to the modification described by Nazzal [9]. The absolute error for those analyses is estimated to be _+0.03. Neutron powder diffraction measurements were performed at 1.5 K and 295 K with the high-resolution Diffractometer D2B at the Institut Laue-Langevin in Grenoble. Neutrons of wavelength 1.5945 /~ were selected for the experiments. The sample container was a vanadium can of about 1 cm diameter and 5 cm length. The data were collected in the angular range 0-160 o (20) and preselected monitor counts for a fixed position of the detector bank was 150 000. The analysis of the data was performed using the programs F U L L P R O F [ l0 ] and SHELX76 [ l l ] . AC magnetic susceptibility measurements were performed on powdered samples using a Lake Shore instrument. The experiments were carded out at a fixed frequency, v = 111. l Hz, and the AC ripple field was ho= l Oe. The observed magnetic susceptibilities were found to be independent of the AC excitation field within the range 0 . 1 0 e < ho < 10 Oe. Four-probe resisti~/ity measurements were performed by the standard procedure on bars with dimensions 1 × 1 × 13 m m 3 cut from the corresponding pellets, in the temperature range 300-10 K.

3. Results

Single-phase cation-stoichiometric La2CaCu206+y has been succesfully obtained by an oxide precursor

route involving L a 2 C u O 4 as the host lattice in which CaO and CuO are intercalated. It is remarkable that, up to date, only this procedure has allowed the preparation of the stoichiometric phase. Neutron diffraction refinements show an oxygen excess y = 0 . 0 3 7 ( 8 ) . On the other hand, the results of iodometric analyses for the samples treated in argon or oxygen at P = 1 atm, show small variation in oxygen content with respect to the air-heated phase, within the experimental error range, yielding values around y=0.02. The analyses of 1.5 K neutron diffraction data were performed following two separate refinement procedures as follows: ( 1 ) Program FULLPROF was used in a first step to extract integrated intensities from the powder diffractogram. (2) These intensities were used by SHELX for least-squares fitting of a stoichiometric model using the positions given by Nguyen et al. [ 7 ] as initial data, followed by a difference Fourier synthesis. This showed only a chemically reasonable m a x i m u m at (0, 0, ½) which was incorporated into the model as a partially occupied oxygen position. In addition a disorder model was established by refinement of a partial occupation by Ca of the La position La (Ca) 2, and partial occupation by La of the Ca position C a ( L a ) ( 1 ). This was done with the corresponding constraints (SF=stoichiometric factor, S F ( L a ( 2 ) ) + SF(Ca(2) ) =2.0, SF(Ca( 1 ) ) + SF(La( 1 ) ) = 1.0) and setting the thermal parameters at varying fixed values to avoid correlations. Later on, the found stoichiometric coefficients were fixed and a final refinement was performed with anisotropic temperature factors (unit weights, R=0.0276 for 108 observations and 21 parameters). The stoichiometry found atthispoint was L a ( 2 ) 1.71 (4) C a ( 2 ) 0.29 (4) Ca( 1 ) 0.75(3) L a ( 1 ) 0.25(3) Cu 2 0 6.02(2). (3) The final stage of the process consisted again in profile fitting analyses using program FULLPROF. We started with the structural model found in step 2, and performed refinements as described in the above step. The occupation factor of O ( 3 ) refined to a more precise value of 0.037 (8). This value was found to be not very sensitive to the corresponding isotropic temperature factor of O (3), which was fixed at the average value of B=0.55. All other atoms were refined anisotropicaUy during the final re-

A. Fuertes et aL / Oxygen excess and superconductivity in LazCaCu206+~

finement, and yielded temperature factors (table I) very similar to those obtained in step 2. The stoichiometry found was La(2) 1.77(4) Ca(2) 0.23(4)

Ca(l) 0.77(2) La(l) 0.23(2) Cu 2 0 6.037(8). The analysis of the data measured at 295 K was performed starting form the low-temperature model and fixing the occupation factors, as above. Figure 1 shows the observed and calculated diffractograms for 1.5 k as well as their difference. Figure 2 shows an ORTEP diagram of the unit cell (ellipsoids of 7 5% probability). Table I summarizes crystallographic parameters and table II selected interatomic bond

155

distances and angles. AC magnetic susceptibility measurements were carried out in the air-heated sample studied by neutron diffraction and in samples treated under oxygen or argon. As it may be observed in fig. 3, a diamagnetic onset transition is clearly observed at 45 k for the air-heated sample, a transition not observed for La2CuO4 treated in a similar way. The corresponding oxygen and argon annealed samples display a slightly reduced onset temperature and also some reduction of the maximum diamagnetic signal, and yield remarkably similar curves. Electrical resistivity

Table I Crystallographic parameters.

I = 1.5 K

•=295 K

Space group

14/mmm

14/mmm

3.82820(5) 19.4498(4)

3.83350(5) 19.5169(4)

Cell parameters ( /~ )

a c Atom Coordinates

Ca( 1 ) La( 1 ) La (2) Ca(2) Cu O ( 1) 0(2) 0(3)

x 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

z 0.0000 0.0000 0.1756(1) 0.1756(1) 0.5853( 1 ) 0.0821(1) 0.7040( l ) 0.5000

y 0.0000 0.0000 0.0000 0.0000 0.0000 0.5000 0.0000 0.0000

Number o f Reflections

OCC. 0.77(2) 0.23(2) 1.77(4) 0.23(4) 2.00 4.00 2.00 0.037(8)

Ca( 1 ) La( 1 ) La(2) Ca(2) Cu O( 1 ) 0(2) 0(3)

x 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

y 0.0000 0.0000 0.0000 0.0000 0.0000 0.5000 0.0000 0.0000

117

z 0.0000 0.0000 0.1759(1) 0.1759(1) 0.5853 ( 1 ) 0.0821 ( 1 ) 0.7039(2) 0.5000

Occ. 0.77 0.23 1.77 0.23 2.00 4.00 2.00 0.037

117

U V W

0.092(3) -0.106(8) 0.171(4)

U V W

0.072(3) -0.102(7) 0.175(3)

Reliability Factors: Rp R~o Re RB,-,ss

6.90 8.93 4.82 3.32

Rp Rwp Rc RB~

6.51 9.32 3.11 3.32

Thermal parameters ( × 104) according to the expression: exp [ - (h 2#11 +k2#22+12#33+2hk#12 +2hl#13 +2kl#23) ]. In all cases #12 =

#13 =#23 =0Ca(La)(l) La(Ca)(2) Cu O(1) 0(2) B(O(3))

#11

#22

fl33

120(13) 54(7) 41(5) 58(8) 359(11) 0.55

120(13) 54(7) 41(5) 57(8) 359(11)

6.8(5) 3.2(2) 3.2(2) 8.0(2) 5.4(3)

Ca(La)(i) La(Ca) (2) Cu O(1) 0(2) B(O(3))

#.

#22

#33

151(14) 117(8) 61(5) 65(8) 462(13) 0.9

151(14) 117(8) 61(5) 121(8) 462(13)

7.5(6) 5.0(2) 6.2(3) 10.6(3) 5.9(4)

156

A. Fuertes et al. I O x y g e n excess a n d superconductivity in L a 2CaCu 2O n +r

~ I*0I 8

0.5

0.0 I

I

I1

I I

II

I

I

III

In

I Ullll

I

I

I III

I

I I1|11

III

I| |1

a

I IIII

I II

I

II

II

I Illlllll

IIIIl

In

Itl

IIII

0.0 ,,.,,,.d,,t,,,,,,l,,,,,,,,,l.,,,,,,.h,,,,,,,,h,t,,,,,,h,,,,.,.h,,,,,,,,l ......... l,,,,,,,,,h,,,,t,,,h,,,,,,,,h,,,,,,,,l,~u 0 20 30 dO 50 60 70 80 90 100 110 120 130 ldO TWO THETA

Fig. 1. Observed and calculated neutron diffractograms of La2CaCu206.037 at 1.5 K. Experimental points are indicated by crosses. The calculated diffractogram is drawn as a continuous line. The curve at the bottom is the difference pattern Y(obs) - Y( calc) and small bars indicate the angular positions of the allowed Bragg reflections.

measured at i = 1 mA does not show any evidence of superconductivity at temperatures close to 50 K. Instead, a semiconductor-like behavior is observed over the whole temperature range studied. The room temperature resistivity is about 90 m ~ cm. An activated dependence of the resistivity can be nicely fitted over the 80-300 K temperature interval, giving E , ~ 10 meV. Both values can be compared to those reported for La2_xSrxCuO4 [ 12 ] oxides, where p_~ 50 mfl cm (for x = 0 ) , and p ~ 3 mfl cm (for x - 0 . 0 6 ) and E,~,-- 10 meV. At lower temperatures, the activation energy decreases, thus pointing to a new channel for electrical conduction. A 3D Variable Range Hopping model can be use6 to fit the data in the narrow temperature range T < 20 K.

4. Discussion

The main structural features of La2CaCu206+y remain similar to those of Lal.9Cal.lCU206 and La2SrCu206 [ 13,14 ]. As in nonstoichiometric Lal.9CalACU206, partial occupation is observed among La and Ca ions in the respective nine- and eight-coordinated sites. However, refinements show a 25% occupancy of La ions in the usually eight-coordinated sites between Cu-O2 planes (site 1 ), versus the 12% found for the nonstoichiometric phase.

In addition, a disordered oxygen is found, close to those positions. Very possibly this oxygen, 0 (3), increases the coordination number of some of the La ions, since this ion clearly prefers high coordination numbers. Occupancies in the nine-coordinated positions (sites 2), close to the apical oxygens, remain basically equal in both structures: 88.5% for La and 11.5% for Ca ions. Thus, the main structural difference between Ca stoichiometric and nonstoichiometric phases is the higher occupancy of La in sites 1 and the presence of excess oxygen between the CuP2 planes. On the other hand, at room temperature, La2CaCu206+y shows cell parameters ( a = 3.834 A, c = 19.517 A) only slightly larger than those observed for Lal.9Cal.iCu206 (a=3.825 A, c = 19.420 A). The corresponding Cu-Cu distances between layers (3.330 A), O(1 )-O(1 ~ interlayer distance (3.205 A) and the ( L a , C a ) ( 2 ) - O ( 1 ) distance (2.651 A), are also larger (Cu-Cu: 3.306 A, (La, C a ) ( 2 ) - O ( l ) : 2.638 A, O ( 1 ) - O ( 1 ) : 3.191 A for Lal.9Cal.~Cu206). Apparently, the presence of a larger proportion of tripositive ions, with larger ionic radius, between Cu-O2 planes plus the presence of a slight excess of oxygen are able to expand the structure by pushing away the Cu ions. The excess oxygen atom O (3) is found trans to O (2) at a distance of 1.66 A from Cu. This distance is, in principle, too

A. Fuertes et al. / Oxygen excess and superconductivity in La2CaCuzO6+y

157

Table II Selected bond lengths (/~) and angles (degrees). T= 1.5 K

La(2)-O(l) La(2)-O(2) La(2)-O(2)

O(1)-La(2)-O(1) O(1)-La(2)-O(1) O(1)-La(2)-O(2) O(1)-La(2)-O(2) O(1)-La(2)-O(2) O(2)-La(2)-O(2) O(2)-La(2)-O(2) O(2)-La(2)-O(2)

2.640(2) 2.7627(9) 2.342(4) 92.93(6) 61.68(4) 68.62(9) 129.8(1) 133.5(3) 87.71(3) 156.9(1) 78.5(1)

T=295 K 2,651(2) 2.7652(9) 2.346(4) 92.63(6) 61.51(4) 68.61(9) 129.6(1) 133.7(3) 87.76(2) 157.2(1) 78.6(1)

%

Ca(1)-O(1) Ca(l)-O(3)

i i

T 1lc"°'"" ~

(::~alLa),(11

II

O(l)-Ca(1)-O(1) O(l)-Ca(1)-O(l) O(l)-Ca(l)-O(l) O(I)-Ca(1)-O(1) O(l)-Ca(l)-O(l) O(1)-Ca(l)-O(3) O(l)-Ca(1)-O(3) Cu-O(l) Cu-O(2) Cu-O(3) O(1)-Cu-O(l) O(l)-Cu-O(l) O(I)-Cu-O(2) O(1)-Cu-O(3) Cu-O(1)-Cu

2.493(1) 2.7069 100.33(5) 65.77(3) 79.67(9) 180.00(9) 114.2(1) 57.11(3) 122.89(6) 1.9151(1) 2.309(4) 1.659(2) 176.3(1) 89.939(4) 91.9(2) 88.14(8) 176.3(1)

2.498(1) 2.7107 100.21(5) 65.71(3) 79.79(9) 180.00(9) 114.3(1) 57.15(3) 122.86(6) 1.9178(1) 2.315(4) 1.665(2) 176.3(1) 89.939(4) 91.9(2) 88.13(8) 176.3(1)

Fig. 2. ORTEPdiagram of the unit cell for La2CaCu206.o37 ( 1.5 K), showing75e ellipsoids. Dashed circles indicate the partially occupied position of O (3).

short but can be understood if we consider a model o f local defects of stoichiometry La2CaCu206+ 6 ( J > y ) diluted in a bulk phase La2CaCu206 so that in the former the c parameter has expanded to accommodate O(3 ) at a more reasonable distance from Cu. In the case o f Lal.9CaI.ICu206 [13] the authors concluded that superconductivity is not observed in this phase simply because the concentration of holes is too small ( y = 0 . 0 5 per [ C u - O ] unit). When comparing with other superconducting oxides, such as La2_xSrxCuO4, it may indeed be agreed that this is just the concentration necessary to destroy corn-

pletely the long-range antiferromagnetic ordering so that no insulator-metal transition is observed. However, it must be emphasized that this argument cannot be accepted in a general way because it assumes a random distribution of the "defect" creating a hole, i.e., in this case the additional Ca ions. The behavior of La2CuO4+v [ 15 ], with excess oxygen, is an example o f this. In this case, a vanishingly small oxygen excess leads to the appearance of some superconducting behavior. After careful neutron diffraction studies [ 16,17 ] and I~SR experiments [ 18 ] it has been concluded that a phase separation occurs below about 300 K, involving an oxygen segregation. In this way some domains of a nearly stoichiometric, antiferromagnetic phase coexist with superconduct-

158

A. Fuertes et al. / Oxygen excess and superconductivity in La2CaCu206+ y 1E-005

oxygen •

~--1 E--O05

ann.

-"

]

~.

.-.

: ........

'io

io'4'o'5'o'sb'7'0'a'o'9'o ',oo

r(r)

Fig. 3. Magnetic susceptibility observed for the air-treated (squares) and oxygen-annealedsamples (triangles). The argonannealed sample showed a magnetic susceptibility practically equal to that of the oxygen-annealedsample. Inset: expanded onset region.

ing domains

having a composition

of about

L a 2 C u O 4 . o 4 $ [17]. Therefore, even if the mean ox-

ygen excess, and thus the hole concentration, is very low, superconducting behavior with reduced flux exclusion may be observed. A study of the relationship among superconducting phase volumes determined by diffraction and from the shielding volume, clearly showed that a few percent of perfect diamagnetism may correspond to about 30% of superconducting phase volume because of the strong flux penetration effects occurring in the small size domains [ 19]. In this work we show that a small superconducting portion with T¢___45 K may indeed be obtained, possibly because a small oxygen excess is naturally introduced in the so-called "cation-stoichiometric" La2CaCu206+r We observe that the eight-fold coordinated sites have 25% occupancy by La ion and this is probably the reason why some excess oxygen is introduced, since La ions have a strong tendency to have a higher coordination number. It is very likely that, given the strong diffusivity of oxygen ions observed in other layered copper oxides even at low temperatures, this oxygen suffers some clustering as in La2CuO4+r and, thus, leads to small domains having a mean hole concentration high enough to achieve the requirements for superconductivity in the Cu--O2

planes, that is around 0.15 holes per [ C u - O ] group. It should be noted that his oxygen segregation process would not actually involve very long distances. For instance, in La2CuO4.032 (stoichiometry implying one interstitial oxygen every 15 unit cells), we need an oxygen defect every 10 unit cells in order to achieve the observed fraction of superconducting phase (La2CuO4.04s). In the present case, La2CaCu206.037, there is an average of one extra oxygen every 14 unit cells, while we need one every 5 unit cells if the same hole concentration is to be reached (0.10 holes per C u - O unit). The different behavior observed by us for La2CaCu206+ywith respect to that observed by other authors for Lat.9Cal.lCu206, can be explained on the basis of the reduced La occupancy in the eight-fold coordinated sites for the latter, and so the elimination of the clustering process. The above constitutes the simplest explanation for the observed behavior and, also, the most similar to the case of superconductivity in the La2CuO4+y system. However, more complex, alternative schemes may exist. For instance, one could be the presence of an inhomogeneous distribution of calcium and lanthanum ions leading to microdomains such as LaE_xCaxCal_xLaxCu2II06 (as the main phase), LaECa t_xLaxcu21106+x/2 and La2_xCaxCaCu2II'IIIO6 (as minority phases). The first two domains would contain only C u ( I I ) and would not superconduct, whereas the third one would exhibit mixed-valence and would be responsible for the observed superconductivity. We should note however, that our explanation would lead as well to a coexistence of antiferromagnetism and superconductivity, similarly to the case ofLaECuO4+ r Within the counting statistics accuracy of the powder neutron diffraction experiment we could not detect any sign of antiferromagnetic order at 1.5 K. This negative result, however, does not contradict our interpretation because the observation of antiferromagnetic order in the 2D, S = 1/2 Heisenberg systems appears very problematic, due to the small intensity of the Bragg lines, associated with a strong quantum reduction of the magnetic moments. The situation becomes even worse if the antiferromagnetic domains have a small size, as may be anticipated from the proposed oxygen segregated structure. Other experiments with higher sensitivity

A. Fuertes et al. / Oxygen excess and superconductivity in La:CaCuzO6+ y

to the perturbed magnetically ordered system such as B-spin rotation should be used. According to this explanation, the structure reported here for La2CaCu206+, then corresponds basically to that of the antiferromagnetic phase. Comparison among distances observed by us and by other authors for superconductors with a pyramidal Cu coordination, has to be done carefully. Thus, although Cu-O(2) distances fall in the range of CuO(apical oxygen) found for other superconductors [20], Cu-O(1) is clearly shorter than the corresponding values for superconductors. However, no discussion is possible at this point on that matter since no correlations exist for n = 2 members of this series of compounds [ 21 ]. With respect to the electrical properties observed for our phase, it is worth noting that activation energies as low as ca. 10 meV are not usually associated with interband transitions. Instead, they are very similar to the energies involved in spin fluctuations (with an order of magnitude of k T N ~- 10 meV). It is then tempting to associate this activation energy with the motion of charge carriers through a background of an antiferromagnetically coupled 2D lattice. Indeed, E~_ 10 meV is similar to the values observed in La2_~SrxCuO4 where the existence of 2D antiferromagnetic correlations within the Cu-O2 planes is well established. The nonobservance of superconductivity in our resistivity data is consistent with the small fraction of superconducting islands already discussed, and their dilution within the structure. As a summary, our work shows that, indeed, hightemperature superconductivity may exist in the La2CaCu206-type structure if an appropriate doping procedure is carried out. Our work predicts that the superconducting fraction may be increased if the oxygen excess is further increased. At the moment of submitting this article for publication, we have learned that a new superconducting phase (T~ of about 50 K) with the type of structure described here, and doped with Sr, has been obtained by Cava et al. [22], work that will be published in Nature.

N o t e a d d e d in p r o o f

159

Acknowledgements This work has been supported by Grants from the Spanish CICYT, from the MIDAS Program and by the European Economic Community.

References [ 1 ] J.B. Torrance, Y. Tokura, A. Na77al and S.S.P. Parkin, Phys. Rev. Lett. 60 (1988) 542. [ 2 ] B. Raveau, C. Michel, M. Hervieu, D. Groult and J. Provost, Mat. Res. Soe. Syrup. Proc. 156 (1989) 129. [3] R.J. Cava, Science 247 (1990) 656. [4]T. Tamegai and Y. Iye, Physica C 159 (1989) 181. [5] S.N. Ruddlesdepn and P. Popper, Acta Crystallogr. 10 (1957) 538 and 11 (1958) 54. [6] IC Doverspike, J.H. Lin, K. Dwight and A. Wold, J. Solid State Chem. 82 (1989) 30. [7] N. Nguyen, L. Er-Rakho, C. Michel, J. Choisnets and B. Raveau, Mat. Res. Bull. 15 (1980) 891. [ 8 ] N. Nguyen, C. Michel, F. Studerand and B. Raveau, Mat. Chem. 7(1982) 413. [9] A.I. Nazzal, V.Y. Lee, R.D. Jacowitz, Y. Tokura and J.B. Torrance, Physiea C 153-155 (1988) 1367. [10]J. Rodrtguez-Carvajal, Program FULLPROF; Rietveld/ Pattern Matching Analysis of Powder Patterns (ILL, Grenoble, 1990). [ 11 ] G.M. Sheldrick, SHELX-76; Program for Crystal Structure Determination (University of Cambridge, 1976). [ 12 ] H. Takagi, Y. Tokura and S. Uchita in: Mechanisms of HighTemperature Superconductivity, Springer Verlag Materials Sciences Vol. 11, eds. H. Kamimura and A. Oshiyama (Springer, Berlin, 1989). [ 13 ] F. Izumi, E. Takayama-Muromachi, Y. Nakai and H. Asano, Physiea C 157 (1989) 89. [ 14] V. Caignaert, N. Nguyen and B. Raveau, Mat. Res. Bull. 25 (1990) 199. [15] J. Beille, R. Cabanel, C. Chaillout, B. Chevalier, G. Demazeau, F. Deslandes, J. Etourneau, P. Lejay, C. Michel, J. Provost, B. Raveau, A. Stdpice, J.L. Tholence and R. Tournier, C.R. Acad. SCi. Ser. II 304 (1987) 1097; J. Beille, B. Chevalier, G. Demazeau, F. Deslandes, J. Etourneau, O. Laborde, C. Michel, P. Lejay, J. Provost, B. Raveau, A. Sulpice, J.L. Tholence and R. Tournier, Physiea B 146 (1987) 307. [16]J.D. Jorgensen, B. Dabrnwski, S. Pei, D.G. Hinks, L. Soderholm, B. Morosin, J.E. Schirber, E.L. Venturini and D.S. Ginley, Phys. Rev. B38 (1988) 11337. [17] C. Chaillout, J. Chenavas, S.W. Cheong, Z. Fisk, M. Marezzio, B. Morosin and J.E. Schirber, Physiea C 162-. 164 (1989) 57; Physica C 170 (1990) 153. [ 18] E.J. Ansaldo, J.H. Brewer, T.M. Riseman, J.E. Schirber, E.L. Venturi, B. Morosin, D.S. Ginley and Sternlieb, Phys. Rev. B40 (1989) 2555.

160

A. Fuertes et al. / Oxygen excess and superconductivity in LazCaCu206+y

[19] B. Dabrowski, D.G. Hinks, J.D. Jorgensen and D.R. Richards, in: High-Temperature Superconductors: Relationships Between Properties, Structure and Solid-State Chemistry, vol. 156, eds. J.D. Jorgensen, K. Kitazawa, J.M. Tarascon, M.S. Thompson and J.B. Torrance (Materials Research Society Syrup. Proc. ) p. 69.

[20] IC Yvon and M. Francois, Z. Phys.-Condensed Matter 76 (1989) 413. [21 ] M.H. Whangbo, D.B. Kang, C.C. Torardi, Physica C 158 (1989) 371. [22] R.J. Cava, European Materials Research Society Meeting, May 1990, oral presentation.