DT1C 7;LE COPY

LO

0 OFFICE OF NAVAL RESEARCH

in CONTRACT N00014-82-K-0317

ITECHNICAL REPORT NO. 56

Chemistry and Superconductivity in Thallium-Based Cuprates

I By

M. Greenblatt, S. Li, L.E.H. McMills and K. V. Ramanujachary

Prepared for Publication

in

"Studies of High Temperature Superconductors", Ed. A.V. Narlikar, Nova Science

Rutgers, The State University of New Jersey

Department of Chemistry

New Brunswick, NJ 08903

June 1, 1990

Reproduction in whole, or in part, is permitted forany purpose of the United States Government.

This document has been approved for public releaseand sale; its distribution is unlimited.

Accossion For

NTIS GRA&I

DTIC TABOnaiiou:.c ed D T ICJusti ifI ,ctio, ELECTE

• :.: ,AUGO0 7 1990

Distril"itton/

Avail2bility Codcs

Irist I IdIor I

TMGl11~v(

$1ECURITV CI.ASS4.'ATION OF T,,$I PAGE

REPORT DOCUMENTATION PAGEI& REPORT SECURITY CLASSIFICATION Ib RESTRICTIVE MARKINGS

Uriclass ified

2. SECURITY CLASSIFICATION AUTWORITY 3 DISTRIBUTION / AVAILABILITY OF REPORT

Zb casfe ECSICCEApproved for Public Distribution-unlimited2b O0[CLASSiFicAT*Ni OWNGRADINtG SCHEDULE

4 PERFORMING ORGANIZATION REPORT NuMBER(S) 5 MONITORING ORGANIZATION REPORT NUMRER(S)

Go NAME OF PERFORMING ORGANIZATION |60 OFFICE SYMBOL 7a NAME OF MONITORING ORGANIZATIONRutgers, The State University j (If 800icowe) Electrochemical Sciences

of New Jersey 4B883 Office of Naval Research

kc. ADDRESS (City, State, and ZIPCode) ?b ADDRESS (CIty, State. and ZIP Code)P.O. Box 939 800 North Quincy St.Piscataway, NJ 08854 Arlington, VA 22217-5000

Be NAME OF PUNDING/SPONSORING f8b OFFPCE SYMBOL 9 PROCUREMENT INSTRUMENT IDENTIFICATION NUMBERORGANIZATION I (If applhcable)

Office of Naval Research I 1 IIFS_.. ..k. ADDRESS (City,. Stare. and ZIP Code) 10 SOURCE OF FUNDING NUMBERS

800 North Quincy St. PROGRAM ELPROJECT O TASK UNT

Arlington, VA 22217-5000 ELEMENT NO NO I NO ACCESSION NoN00014 1413d008

11 TITLE (Include Securrty ClaSSibcateon)

Chemistry and Superconductivity in Thallium-Based Cuprates

12 PERSONAL AUTHOR(S)M. Greenblatt, S. Li, L.E.H. McMills and K.V. Ramanujachary

13a TYPE OF REPORT 13b TIME COvERED 14 DATE OF REPOP (Year, Month, Day) S PAGE COuNTTechnical IROM 6/1/89 TO 5/30/90 June 1, 1990

16 SOPP,,EMENTARY NOTATION

7 COSAT, CODES '8 58,ECT TERMS Contirue on reverse if necessary and identify by block number);'E.D j GROUP SuB-GROUP

'9 ABSRACT (Continue on reverse ,f necessarV am identify by block number)

A review of the synthesis, structure, substitutions and properties of Tl-base copper oxide

superconductors up to October 1, 1990 with 113 references.

5 S'1 8'0% vvO..Ar..- 0 AOSS A C I' ; -85;;C sEC_ _ ' ; C,C %C- .-C..ASS ;O' '" D [ SAME AS C _s-'s.'.'AE 0: -IESD0%S B.E %: ,A.: EDO%E lsnci~oe Area C):oc: ~~ ;

Martha Greenblatt (201) 932-3277

00 FORM 1473, B: VAR 3 A00 C " , -a, e oC" ea.sea SEC- v'v C ASSF CA' 0% 0; '- -A :*e' to ! o-s are 0osoete

CHEMISTRY AND SUPERCONDUCTIVITY OF THALLIUM-BASED

CUPRATES

M. Greenblatt, S. Li, L. E. H. McMills, and K. V. Ramanujachary

Department of Chemistry

Rutgers, the State University of New Jersey, Piscataway, NJ 08855-0939

1. Introduction

Following the discoveries of ligh temperature superconductivity in the rare-earth copper oxide systems at r40K by Bednorz and MUller in 1986111 and at.490K by other researchersf2,3) in 1987, Sheng and Hermann, in 1988, discoveredsuperconductivity in the thallium-alkaline-earth copper oxide systems withcritical temperatures as high as -120K[4,51. All of the Tl-based compounds canbe described by the general formula, TlmA2Can-lCunO2n+m+2, where m=1 or 2;n=1~5; A=Ba, Sr. For convenience, the names of these compounds areabbreviated as 2223 for T12 Ba 2Ca2Cu 3OO, where each number denotes thenumber of T1, Ba(Sr), Ca and Cu ions per formula, respectively. Thecompounds with m=1 and m=2 are usually referred to as single and doubleTI-O layered compounds, respectively. The highest superconducting transitiontemperature known so far was found in Tl2Ba2Ca2Cu 3 010 at ,i25K[6,7 -

The thallium-based compounds are structurally related to the rare-earth-based cuprates and present more interesting chemical and physical features. Todate, about twenty-five thallium-copper oxide systems have been investigated.A summary of those studies is given in Table 1. Current investigations of theseTl-based high Tc superconducting oxides have mostly focused on the synthesis,structural and electronic properties and the correlations between the chemistryand their superconducting properties.

Unlike the rare-earth-based compounds, the TI-based copper oxides arethermally unstable phases, and hence, are difficult to prepare as pure phases.Moreover, the thallium compounds are severely nonstoichiometric and

2

Table 1. Superconducting Thalli um-copper-oxi des and Related Systems

Compound Tc (K) Related Systems [Tc (K)1 Ref.

712 Ba2CuO6 0-90 8-14

TI2 Ba2CaCu2O8 110 10,11,15,16

TI2Ba2-xSrxCaCu2O8 [40-1101 17-19

T12-xKx~a2CCu2O8 1110-1301 20

fl 2 -x/3TIIxBal.sxLnCu2O8 [01 21

TI2Ba2Ca2Cu3OI0 115-125 6,7,10,11,16,22,23

T12Ba2-xSrxCa2Cu3O)10 [1141 19

T12Ba2Ca3Cu4O1O 104-110 24,25

TlBa2CuO5 metallic 10,11,26

T1Ba2-xLnxCuO5[V401 27

T1Ba2CaCu2O7 50-90 10,11,26,28,29

TlBa2LflCu2O7 [0] 30,31

TISr2,(Ln,Ca)Cu207/ [60-901 32-35

TIBa2Ca2Cu3O)9 110-117 36,37

TIBa2Ca3Cu4O1I -120 38-41

TIBa2Ca4Cu5O13 -110 42-44

MT,Bi)Sr2CuO5 metallic 45

M,Bi)Sr2-xCaxCuO5 [-501 46

M,Bi)Sr2CaCu207 90 47,48

(TI,Bi)Sr2Ca2Cu3O9 -120 49

M,Pb)Sr2Cu05 0-20 45,50

(TlPb)-Sr-LTI-Cu-O 120-701 51-54

M,Pb)Sr2CaCu207 80-855,6

MT,Pb)Sr2Ca2Cu309 115 55

T1Sr2CaCu2O7 100 57

TI-Sr-Ca-Cu-O [20-1001 58-60

TI-Sr-Ln-Cu-O [0-401 61

phases with n;-4 110-160? 62,63

3

contain a considerable concentration of structural defects, which significantlyaffects the physical properties. The copper and thallium chemistry areparticularly interesting in terms of the Cu and T1 valences and the correlationsbetween their valences and superconductivity of Tl-based cuprate materials.

This article is intended to review the present status of the chemistry of thethallium based high Tc superconducting oxides, including their syntheses,structures, nonstoichiometries, T1 and Cu valences and their correlations withthe superconducting properties. In view of the extensive studies carried out onthese phases, it is not possible to include all the results in this paper.

2). Synthetic Chemistry of TlmA2Can-lCUnO2n+m+2

Thallium and its compounds are among the most toxic inorganic materials.Safety procedures are extremely important in the preparation of Tl-containingcompounds. T120 3 , which is one of the starting oxides used to prepare the highTc superconducting thallocuprates, is highly volatile at temperatures above itsmelting point of -710°C. In preparing these high Tc superconducting oxides,prevention of thallium loss is the most important factor, not only for safetypurposes, but also for maintaining the stoichiometry of the products to ensurethe formation of the desired phase.

i. Reactions in Open Systems

The thallium based high temperature superconducting compounds werefirst synthesized by Sheng and Hermann[4-5], by heating a Ba-Ca-Cu-Oprecursor oxide with T120 3 powder at -880-910'C for -3-5 minutes in flowingoxygen. This technique is extremely simple, convenient and unusually quickfor a solid state reaction; and it is now generally used with modifications of thereaction temperature, time and the starting chemical composition, to preparemost of the well characterized thallium based compounds[18,19,22-24,64,65].However, this method carries some difficulties inherent in the svnthesis ofpure phase samples due to the volatility of the thallium compinent and severedecomposition of products at the high reaction temperatures as described by:

high temp.

TlmA2Can1CunO2n+2+m h T1203 + ACaCuOy (1)

In order to prevent severe thallium loss durin- the reaction, samples havebeen wrapped in noble metal foils, such as Au, Ag, Pt. Ni foils have also beenused in the synthesis of thallium-based superconductors, however, in ourexperience, Ni foils become highly reactive at the temperatures needed tosynthesize those compounds. We have found that very thin (<0.025mm thick)Au foil works better than other foils, in terms of preventing Tl loss, because it isvery inert and melts slightly at the high reaction temperatures, forming a tightseal which prevents severe thallium loss.

4

In open reaction systems, phase transformation from a TI-rich to a relativelyTi-poor phase was observed as a result of TI loss. For example, starting with a2223 stoichiometry, continuous transformations of 2201--2212----*2223[65] and2223---+1223 - 1234[40,44,65] phases as a function of time have been observedby powder X-ray diffraction diffraction. Under controlled reaction conditions,single phase samples can be obtained.

ii. Reactions in Closed Systems

Thallium-copper oxide samples can also be prepared using a completelysealed reaction system, which was developed by researchers at DuPont[7,15,361.In this method, starting materials are sealed into a gold tube and heated at-900°C for a period of 0.5-3 hours; both pure powder samples and single crystalscan be prepared, since a much longer reaction time may be used under closedreaction conditions. This technique is, in our experience, the best way toreproducibly prepare pure phase samples with nearly the desired stoichiometry.Silver tubes have also been used successfully as a less expensive substitute forgold tubes.

Sealed quartz tubes [6,10,11,26,66-68] or autoclave-like inert metalcontainers[691 have also been employed for the preparation of Tl-basedcompounds. However, the quartz t abe can be seve rely attacked by the thalliumoxide vapor, resulting in deviation from the desired stoichiometry. In addition,there is always a danger of explosion at high temperatures due to the highvapor pressure of thallium oxides in the quartz tube. Special precautionsshould be taken when using this route. Preparing materials in an autoclave, inprinciple, should work as well as the sealed-gold tube technique, however, inconventional autoclaves, the sealing does not hold at the reaction temperaturesused(840-930"C), resulting in deviations from the stoichiometry. Theadvantage of using quartz tubes is obvious when vacuum conditions arerequired.

Using the above techniques, T12Ba2CuO618-14], T12Ba2CaCu 2O8110,11,15,16]and T12Ba2Ca2Cu3OlO[7,10,11,16](2201, 2212 and 2223, respectively),(TI,Pb)Sr2CaCu2O7155,56], and (TI,Pb)Sr 2Ca 2Cu309[551 have been successfullyprepared in both bulk pure powder and single crystal forms.T1Ba 2Ca2Cu3O9(1223) single crystals have also been grown by the sealed-goldtube technique[36]. Although single crystal growth of TI2 Ba2CaCu 2O 8 andT12Ba2Ca2Cu3OO were reported using an open reaction system[701 with excessCaO and CuO, it is very difficult to reproducibly obtain good quality singlecrystals this way.

iii. Synthetic Chemistry

All of the members of TlmBa2Can-lCunO2n+m+2 are thermally unstable. At-820"C, they begin to form at a relatively rapid rate, but decomposition occurs

5

almost simultaneously (decomposition can take place slowly even at 750"C). Atapproximately 880-900"C, the compounds melt incongruently, according toequation (1). If an open reaction system is used, thallium loss in thistemperature region is severe and irreversible. The ideal temperature region, inour experience, for the preparation of single phase samples is -840-870"C for 1-6hours, regardless of the nature of starting materials and the composition ofproducts.

The single Tl-O layered compounds are much less stable than the doublelayered counterparts, and thercfore, the preparation of pure single TI-O layeredcompounds is extremely difficult. Although, the existence ofTIBa2Can-lCunO2n+3 with n up to 5 has been observed[10,11,26,62,63,71-75] bymeans of high resolution electron microscopy, none of the compounds hasbeen prepared in pure powder form. The exception is TlSr 2CaCu20757], whichwas prepared by a hot press technique from a nonstoichiometric startingcomposition. However, single phase analogues of these unstable compoundscan be prepared by partial substitutions of TI by Bi or Pb, or Ca by rare-earthelements, forming (Tl,Bi,Pb)Sr2Can.CunO 4n+1(n=1, 2 and 3)[45-561 andTIBa2(Caj-xLnx)n-1CunO4n+1 (0.0<x<1.0; n=1 and 2)[30-351. The reason for theinstability of the single TI-O layered compounds is probably due to the highformal valence (much higher than 2+) of copper in these compounds. Theexistence of strontium analogues in the T12Ba2Can-lCunO2n+4 family has notbeen reported; although there is evidence of Tl2Sr 2CaCu 20 8 in a mixed phasesample[17].

Another way to prevent severe loss of TI is by the use of Tl-containingprecursor oxides[14,45,471. Unlike the Ba-Ca-Cu-O precursors, which have to beprepared at high temperatures prior to the reaction with T1203, the TI-containing precursors can be prepared by preheating all of the starting oxides at-580-600°C for 1-2 hours. CuO does not react with the other components at

such low temperatures. The Ti-containing precursor oxides observed includeBa 2TI2Os, Sr 4 TI207 and an unidentified Ca-TI-O phase, all of which arerelatively less volatile than T120 3 at the high reaction temperatures.

Most preparations of Tl-based cuprates are carried out in pure oxygenatmosphere. However, post-annealing of samples in controlled atmosphere hasbeen reported to be an important factor for the preparation of pure phasesamples and the enhancement of the superconducting properties for some ofthe Tl-based cuprates[14,68,69,761. For example, the preparation of thesuperconducting TI2 Ba2CuO 6 phase (Tc-80-90K), which has a tetragonal crystalsymmetry, requires an oxygen-poor atmosphere or fast quenching in air fromthe reaction temperature (-820-850"C) to room temperature14]. If the sample of2201 is prepared in pure oxygen atmosphere or in a closed reaction system, theproduct has an orthorhombic symmetry and either lower Tc(-20K) or metallicbehavior[10-14] The nonsuperconducting (or low Tc) phase can be convertedinto a superconducting phase (or with enhanced Tc), by post-annealing thesample in an inert atmosphere at -300°C[14]. This transformation is

6

accompanied by an orthorhombic to tetragonal phase transition, which isbelieved to be related to the oxygen content and/or ordering in Tl2Ba 2CuO6 andis critical in determining the superconducting properties. The effects ofcontrolled atmosphere on the superconductivity of Ti-based cuprates were alsofound in other systems, such as in the 2223 and will be discussed later in detail.

(a) 12CI (b) 1212 (:) 1223

(c) 123

. Cu

eTI(f) 22.2 (g) 2223 0 Ca••0 Ba(Sr)

(e) 2201

Figure 1. Schematic presentation of crystal structures of 1201, 1212, 1223, 123, 2201, 2212and 2223.

'7

3. The Structural Properties of TImA2A'n-lCunO2n+m+2

i. Basic Structural Properties

Characterization of the new thallium copper oxides were carried out bysingle crystal X-ray diffraction, Rietveld analysis of powder neutron/X-raydiffraction, electron diffraction and high resolution electron microscopy. Thestructures of T12Ba2Can.lCunO 2n+4 phases (n=l, 2, and 3)[7-9,15,161,(Tl,Pb)Sr2Can-lCunO4n+l (n=2 and 3)[55,56] and T1Ba2Ca 2Cu3O9[36] have beenwell established by single crystal X-ray and/or powder neutron diffractionanalyses. Figure 1 illustrates schematically the crystal structures of thesecompounds. The structure of YBa2Cu 3O 7 (denoted as 123) is shown forcomparison. Figure 2 shows the powder X-ray diffraction patterns of pure 1201,1212, 2201, 2212 and 2223 phases prepared in our laboratory[14,45,47,77. Thestructures of the other members of this family were characterized only by highresolution electron microscopy, powder X-ray analysis and electron diffractiontechniques, especially for the phases with m=1 or n>4162,63,71-751 (also see Table2). High resolution electron microscopy studies have shown that the phaseswith n>4 consist of intergrowths of the phases with n-<3, and exist only inmicroscopic regions of the samples[73].

In general, the Tl-based superconducting oxides can be considered to formoxygen deficient perovskite or Aurivillius related structures. The unit cell ismade up of alternating arrays of one to n Cu-O corner-sharing square-planarand/or square-pyramidal layer(s) and one single or double edge-sharing T1-Ooctahedral layer. The Cu-O and TI-O polyhedra are interconnected to formcovalent framework slabs; the A cations are located in the interstices of TI-Oand Cu-O slabs and the Ca 2+ ions are between the Cu-O layers (Figure 1). Theoxygen coordination number of TI is 6, of Cu is 4 or 5, of Ca is 8, and of A is 9 inthe double and 12 in the single TI-O layered compounds. In all of thesecompounds, the Tl-O bond length is -2.0k along the 001 and -2.5-3.0k alongthe 110 direction; the Cu-O bond length is -1.90-1.92k in the ab plane; thedistance of adjacent Cu-O layers is -3.1-3.2A. The apical Cu-O distance is -2.7Ain TlmBa2Can-lCunO 2n+4 and -2.3k in (Tl,M)Sr2Can-lCunO 2 n+3 (M=Pb or Bi). TheCu-O-Cu bond angle is close to 180' in all the structurally characterized phases,indicating that the Cu-O layer is almost ideally planar.

Table 2 summarizes the lattice parameters of the known thallium copperoxides. Tetragonal symmetry was found to be common in all of thesecompounds, as determined by X-ray and neutron diffraction. However, thereare some uncertainties about the oxygen or thallium positions in the structure.In the double TI-O compounds, the oxygen atom in the basal plane of the Tl-Olayer can be refined only at the 16(n) position of space group 14/mmm with 1/4occupancy, instead of the ideal, fully occupied 4(e) site[7-9,15]. In the single TI-Olayer compounds, the oxygen refines best at 4(n), not the ideal 2(g) position; andTI must be placed at a position slightly deviating from the ideal 4(l) site[55]. Thedisplacements of the oxygen and thallium atoms from their ideal locations, in

(a) TJ.2Ba 2Ca2 Cu3Ol 0

(b) T' 2 Ba 2 CaCu 2 39

(c) T22 Ba 2 _Z,;L

(e) t

Figure 2. Powder X-ray diffraction patterns of a. TI2Ba2Ca2Cu 3OlO,b. TI2Ba2CaCu2O8, c. TI2Ba2CUO6(tetragonal), d. (TI,9i)Sr2CaCu2O7, ande. (TI,Bi)Sr2CuO 6.

9

Table 2. Lattice parameters of thallium copper oxides

a(A) b(A) c(A) 9G Refs.TI2Ba2CuO6(ortho) 5.4967(3) 5.4651(3) 23.2461(1) Abmaa 9T12Ba2CuO6(tetra) 3.866(1) 23.239(6) 14/mmmt 8T12Ba2CaCu2O8 3.8550(6) 29.318(4) 14/mmmt 15

3.8559(1) 29.4199(10) 14/mmm' 16T12Ba2Ca2Cu3OlO 3.8503(6) 35.88(3) 14/mmm t 7

3.8487(1) 35.6620(15) 14/rnmm 16TI2Ba2Ca3Cu4O2 3.854() 42.070(11) 14/mmmt 24T1Ba2CuO5 3.896(2) 9.694(9) P4/mmm4 10,26TlBa2CaCu2O7 3.8505(7) 12.728(2) P4/mmm 10,11TIBa2Ca2Cu3O9 3.8429(6) 15.871(3) P4/mmmt 10,37TlBaiCa 3Cu 4Oll 3.85 18.73 P4/ mmnu 41TISr2CaCu 207 3.805(6) 12.14(6) P4/mmm 57(TI,Bi)Sr2CuOS 3.759(1) 9.01(1) P4/mnn- 45(TI,Bi)Sr 2CaCu 2O7 3.796(1) 12.113(1) P4/mmm 47,48(T1,Pb)Sr2CuO5 3.731(1) 9.01(1) P4/mmnv 4 45(TI,Pb)Sr2CaCu2O7 3.80 12.05 P4/mmmt 55(Tl,Pb)Sr 2Ca 2Cu 3O9 3.808 15.232 P4/mmmt 55phases with n >4 3.80-3.85 *? 38,T9

t: Single crystal X-ray diffraction datat: Powder X-ray diffraction dataa: Neutron diffraction data*: The c lattice parameters can be estimated by formulae c = 23.2 + 6.4(n-1) for the m=2 phases;c = 9.6 + 3.2(n-1) for the m=1, A=Ba phases; and c = 9.0 + 3.1(n-1) for the m=1 A=Sr phases. Thestructural symmetry is suggested to be tetragonal.

Figure 3. Structure of 1212(a), considering oxygen ordering in the T1-O basal plane. ATI-O chain is shown and compared with the Cu-C chain in the 123 structure(b).

I0

all of the Tl-based cuprates, result in the formation of four different TI-O bondsin the basal plane and severely distorted Tl-O octahedra. It is assumed that the16(n) oxygens in the double layered compounds and the T1 atoms in the singlelayered compounds are randomly distributed around the ideal sites of thetetragonal lattice. Ordering of these atoms will lead to the formation of a localT1-O chain, as shown in Figure 3, and a concomitant change in the structuralsymmetry and/or formation of a superlattice. However, both X-ray and neutrondiffraction studies lack conclusive evidence for oxygen ordering inT12Ba 2CaCu2O8 and T12Ba2Ca2Cu 3Olo[16].

ii. Local Structural Distortions and Cation Deficiencies

High resolution electron microscopy and electron diffraction techniqueshave been used to investigate the local structural properties of Ti-basedsuperconductors. Reports of structural modulations, similar to those found inthe Bi-based high Tc superconducting oxides, have not been unambiguouslyconfirmed[78-84]. The origin of such modulations and their correlations withthe physical properties are not well understood. It has been suggested that themodulated structure is due to the local ordering of cation defects or oxygenatoms in the T1-O basal plane[84,86]. However, these effects are notreproducible, and the presence of the modulations is dependent upon thesample exposure to the electron beam[81-84].

Consistent with X-ray and neutron diffraction results, high resolutionelectron microscopy and microprobe chemical analysis show significant Ti andCa deficiencies in most of the thallium copper oxides [67,85,861. The amounts ofTi/Ca deficiencies depend on the preparation conditions and the type ofstructure. In general, the deficiencies tend to increase with more Cu-O layers inthe structure[87]. Hibble et al. reported that in T12Ba2CaCu2O8 andTi2Ba2Ca2Cu3010, the deficiencies are 12%TI/30%Ca and 8.5%TI/28%Ca,respectively. No Ti/Ca deficiencies were observed in TlBa 2Ca 2Cu 309[85].However, in these studies, barium and copper deficiencies were not observed.The Ti/Ca deficiencies cause deviations from the ideal stoichiometry and arelikely to have significant effects on the superconducting properties of thesecompounds. Partial occupancies of TI and Ca on each other's sites have alsobeen observed by the X-ray and neutron dffraction analysb f7,15,16,55].Nominally, 50 percent of TI could be substituted by Ca in 2223[881.

iii. Structural Phase Transition

All of the Ti-based cuprates crystallize with tetragonal symmetry, exceptT12Ba2CuO6, which exists in both tetragonal and orthorhombic form,depending on the preparation conditioa and/or the post annealing treatment.Ramanujachary et al.[141 reported a reversible structural phase transition undercontrolled atmosphere at -300°C by high temperature X-ray diffractiontechniques. A small but measurable oxygen loss was observed, by both DTA and

II

TGA analysis (Figure 4), accompanying the phase transition from orthorhombicto tetragonal symmetry. Figure 5 illustrates the X-ray powder diffraction peakprofiles of the 200/020 reflection of the tetragonal phase as a function oftemperature when the tetragonal sample is heated in oxygen. As thetemperature exceeds 300*C, the reflection splits into a doublet, which clearlyindicates the tetragonal-orthorhombic transition. The oxygen-rich phase isorthorhombic with lower Tc(or only metallic if the samples are prepared inthe closed reaction system), while the oxygen-poor phase is tetragonal withhigh Tc(55-85K). Although the difference in the oxygen content of these twophases is small, estimated at <0.1 oxygen atom per formula unit, it appears to bea very important factor in determining the superconducting and structuralproperties of 2201. Oxygen ordering may also be important in effecting thephysical properties of this compound. The correlation betweensuperconductivity and preparation conditions of the 2201 phase is wellillustrated in Figure 6.

99.60 -, -0.30

99.48 - -0.35

t 99.36 - -0.40

99.24 -0.45 <

99.12 -0.50

99.00 I I-0.55

260 285 310 335 360

Temperature (K)

Figure 4. TGA(left) and DTA(right) profiles of a 2201 sample heated in Aratmosphere.

12

S00C

32.0 32.5 33.0

20

Figure 5. Peak profiles of 200/020 reflection of a tetragonal 2201 sample when heatedin oxygen atmosphere at various temperatures.

0.001a

0.000

2 b-0.001

:S -0.002C

4W

S -0.003

-0.0040 20 40 60 80 100 120

Temperature (K)

Figure 6. Superconductivity of 2201 samples prepared at different conditions. a.prepared in closed reaction system, which is orthorhombic and metallic, b. furnace cooledin air, which is orthorhombic with Tc -20K, and c. quenched in air to room temperature,which is tetragonal and Tc is -85K.

13

The structural difference between the orthorhombic and tetragonal phases isattributed to differences in the local symmetry around the thallium atoms.According to crystallographic studies[8,9], in the orthorhombic form of 2201, thedeviation from the tetragonal symmetry is essentially restricted in the T1-Olayers. The oxygen coordination of TI in the orthorhombic form of 2201 isnearly a square-pyramid[9]. Bordet et al. noted differences in the superlatticestructure of these two phases by means of electron diffraction: the superlatticecorrelation length in the orthorhombic phase is much larger than that in thetetragonal phase[80]. It was suggested that the tetragonal-orthorhombic phasetransition is related not only to the oxygen content, but also to the oxygenordering in the Tl-O basal plane[141.

iv. Nonstoichiometry of Oxygen

The importance of oxygen nonstoichiometry in Tl-based cuprates has beenreported in several studies[14,68,69,761. However, there is no general consensuson how the oxygen nonstoichiometry affects the superconductivity. Hervieu etal.[681 and Ramanujachary et al.[14] reported that argon annealing of 1212 and2201 samples enhances the Tc, while other investigators found that oxygenannealing can improve the superconducting properties of 2223 and that heattreatment in an inert atmosphere is detrimental to superconductivity[69,76].These results imply that the effects of oxygen nonstoichiometry onsuperconductivity might be different in different compounds, depending uponthe structure, oxygen content/ordering and other nonstoichiometric factors.

The presence of excess oxygen in all of the Ti-based compounds seemsquestionable, because wherever excess oxygen is presumed to be in the lattice,unacceptably short metal-oxygen bonds appear to form. Neutron diffractionstudies[16] showed that oxygen deficiencies are possible in the 2223 phase andare most likely to be located in the Tl-O basal plane, which is disordered in thetetragonal lattice.

The effects of oxygen nonstoichiometry were observed in only a few phases,most profoundly in the 2201 and 2223 systems. Figure 7 shows the effects of thecooling and the post-annealing conditions on the superconductivity of 2201samples. These effects are related to the oxygen content in the structure asmentioned earlier. In contrast, when a 2223 sample is annealed in Aratmosphere at low temperatures, the Tc is depressed; and an oxygen post-annealing will restore the Tc back to -120K[76]. In 2212, oxygen/Ar annealing attemperatures between 250-450°C has no significant effect on Tc.

Due to the complexity of the thallium-copper systems, the oxygen content ofthese compounds has not been directly determined by chemical methods.Schilling et al. reported[69] a qualitative evaluation on the correlation betweenthe oxygen content and the superconducting transition temperature, using TG

14

measurements, but the exact oxygen content in the samples could not bedetermined from the study.

0.001

furnace cooled sample0.000 - €

-0.001

* -0.002

cn -0.003 . Ar post-annealed sample

-0.004 r0 20 40 60 80 100 120

Temperature (K)

Figure 7. Enhancement of Tc by Ar post-annealing of a 2201 orthorhombic low Tc sample.

4. Form A1 Oxidation States of Copper and Thallium

The formal oxidation states of copper and thallium are by far the mostinteresting features of the thallium-copper oxide superconductors. In the highTc superconducting cuprate materials(except in the newly found electron-dopedcuprate superconductors[89,901), the existence of Cu 2 5 (5 is equivalent to thehole concentration in the compound) has been considered to be a critical factorfor superconductivity. One of the central issues concerning the Tl-based cupratesuperconductors is how Cu 2 + 8 is generated in these compounds and how it iscorrelated with superconductivity.

In TlBa2Can-lCunO2n+3, (assuming T13 + , Ba 2+, Ca 2+ and 02-), the valence ofthe copper cation is (2+1 /n)+, as calculated from the ideal stoichiometry, andvaries with the number of Cu-O layers in the structure. However, for the idealstoichiometry of T12Ba2Can-lCunO2n+4, the copper valence is 2+ for all themembers of this family. The origin of Cu 2+5 in T12Ba2Can-lCunO2n+4 has beenattributed to the following: a. the Ti/Ca deficiencies in the structure; b. partial

15

occupancy of Ca on the T1 sites; c. the internal redox reaction T13+ + Cu 2 + 15

T13-8 + Cu 2 +8 (details will be discussed in later sections); and d. possibly,although not proven, excess oxygen in the structure (like in the Bi-based highTc materials).

The crystallographic data[7,8,15,36,55,611 show that in the Tl-O octahedra,there are two very short T1-O bonds(-2.0 A) along the c crystallographicdirection; two nearly regular (-2.5A) and two extremely long Tl-O bonds(3.OA)in the basal plane. Compared with known 13+-O and Tl+-O bonddistances(-2.4A and -3.0, respectively), the formal valence of TI in thesematerials appears to be between 3+ and 1+ and/or corresponds to a veryanisotropic electronic charge distribution on the TI cation. XPS resultsindicate[91-93] that the effective charge on the Ti cations could be less than 3+ inseveral Ti-based cuprates. These results suggest mixed valency of TI, which isconsistent with the band structure calculations and the crystallographic data.

Table 3. Calculated Copper and Thallium Valences inSelected Compounds

Compounds

Valence & Tc 1201" 1223t 12 23t 2201 2212 2223

VTI.0 1 0.67 0.83 0.97 1.02 1.05 1.19VTI 2.09 2.35 3.10 2.65 2.60 2.86

Vcu 2.76 2.07 2.29 2.07 2.07 2.08Tc (K) 0-40 110 120 80-90 110 125

Reference 61 55 36 8 15 7

t: (TI.Pb)Sr 2Ca2Cu 3O9

t: T1Ba 2 Ca2Cu 3O9*: TISr2-xPrxCuO5

Others are TI-Ba-Ca-Cu-O phases.Calculations are based on the equation V = Y(d/do)-N. do and N values for both Tl3 and Cu 2+ arefrom ref. 94. Bond distances, d, are from the corresponding references.

The valences of copper and thallium ions can be estimated using therelationship between bond-order and bond-length in oxide materials[94].Selected values calculated from the crystallographic data are tabulated in Table3, together with the superconducting transition temperature of eachcompound. It is seen that the valence of TI is less than 3 (less than 3.5 at the Ti

16

site for the (T10.5Pb0. 5)Sr 2Ca 2Cu 3O9 ) and that the copper valence is greater than2 in all of the selected compounds. In addition, the copper valences are muchhigher in the single T1-O layered compounds than those in the double T1-Olayered compounds. It is interesting to note that the superconducting transitiontemperatures of these compounds are related to the TI-O 1 (the TI-O bond alongthe c crystallographic direction) bond order. In both single and double T1-Olayered compounds, Tc increases with increasing TI-O1 bond order. Since theT1-O1 bonds are connected to the Cu-O layer(s) where the holes are presumablylocated, the correlation between the TI-0 1 bond order and the Tc may implythat the internal electronic communication between the Tl-O and Cu-O layers isan important factor in effecting superconductivity in the Tl-based cuprates.

5. Substitutional Studies

Substitutional studies have been carried out in all of the known Tl-basedcuprate phases, in part, to explore new superconducting oxides in the TI-relatedsystem, and in part, to understand the relationship between the chemistry ofthese oxides and superconductivity. The major focus of substitutional studieshas been to manipulate the formal oxidation states of copper and thallium,alternatively, to vary the concentration of holes by chemical substitution.Because of the complexity and thermal instability of Tl-based compounds,cationic substitution often leads to mixed phase products. In the single T1-Olayered compounds, substitutions at the T1 site by Bi or Pb appears to stabilizethe phases via reducing the formal copper valence, and hence, have beenstudied more extensively than the double TI-O layered compounds.

i. Substitutions in the TIA2Can-lCunO2n+3 Family(A=Ba, S)

As previously discussed, in the ideally stoichiometric T1Ba2Can-lCunO2n+3(n=l, 2 and 3), the formal copper valence is always higher than 2+, while inTl2Ba2Can-lCunO2n+4, it is always 2+. The copper valence in TlBa2Can-lCunO2n+3is nominally 3.0+ for n=1, 2.50+ for n=2 and 2.33+ for the n=3 phaserespectively. The high copper valence may be one of the reasons for thestructural instability observed in the single T1-O layered compounds. With theexception of single crystal TIBa 2Ca2Cu 309136], pure phase powder materialshave not been obtained for TIA 2Can-lCunO2n+3 family without substitution.The T1A2Can-lCunO2n+3 system is an ideal system for studying the correlationsbetween the copper valence/hole concentration and superconductivity,provided that pure phase samples can be obtained. Replacement of T13+ byPb 4+/Pb2+, or Bi5 /Bi 3 , or Ca 2 by Ln3+(Ln=rare-earth elements) would beexpected to reduce the copper valence with concomitant stabilization of thestructure.

T13, Pb4+ and Bi5+ have identical electronic configuration and similarchemistry. Substitutions of TI by Bi and Pb in the TI-Sr-Ca-Cu-O systems were

17

carried out very successfully. (TI,Bi)Sr2Can-lCunO2n+3(n=l and 2)[45,47,48) and(Tl,Pb)Sr2CuO5145I were prepared as pure polycrystalline phases, and(Tl,Pb)Sr2Can-lCunO2n+4 (n=2 and 3)[55] in both powder and single crystal forms.In the substituted (T1,M)Sr2Can-lCunO2n+3 (M=Pb or Bi) compounds, theTl/Pb(Bi) ratio is close to 1.0. No solid solution region was reported in the(Tl,M)Sr2Can-1CunO2n+3 phases, except in nonsuperconductingTllxBixSr2CuO5[45], where 0.0<x<0.50.

The Pb and Bi valences are of interest in the (Tl,Pb(Bi))Sr2Can-lCunO2n+3series. There is no direct method available to determine the formal valence ofBi and Pb in these compounds. The existence of Bi3+6 is apparent inTll-xBixSr2CuO5 by a careful analysis of lattice parameters as a function of x. Asshown in Figure 8, the lattice parameters of Tll.xBixSr 2CuO5 decrease in theregion 0.0<x<0.20, and then increase with x in the region 0.20<x<0.50. Theinitial decrease of the lattice parameter a is attributed to the substitution of T13+

by Bi 3 + 8, while the increase is due to the substitution of T13- by Bi 3+ since bothBi5+ and T13+ are smaller than Bi3 (rBi3+=1.17A, rBi5+=0.90A and rTl3+=1.03A forCN=6). Therefore, in the region 0.20<_x_<0.50, Bi3+ 5 and Bi3 + are likely to coexist.

3.78- - 130.0- a

V-129.0

3.76-

128.0° 3.74 0< *

127.0

3.72 "126.0

3.70" 125.0

0.00 0.10 0.20 0.30 0.40 0.50 0.60

Value of x

Figure 8. Lattice parameter of Tll-xBixSr2CuO5 as a function of Bi content x.

Substitution of trivalent rare-earth cations for Ca 2+ would also be expected toreduce the copper valence and stabilize the single T1-O layered phases. Acomplete substitution of Ca by rare-earth cations in TlA2CaCu 207 (A=Ba or Sr)stabilizes the phase but destroys the superconductivity, resulting insemiconducting compounds[30,311. Partial substitutions of Ca by rare-earth ions

18

yield superconductivity with a lower Tc than that of the parent compounds[32-351. However, Yb substitution for Ca was reported to be not deleterious tosuperconductivity[33]. Although TlBa2CuO5 is only metallic[10],superconductivity was reported in TIBa2.xLnxCuO5 (Ln=La, Nd) with Tc-40K[27]. Reports of superconductivity in the T1-Ln-Ba(Sr)-Ca-Cu-O mixedphase systems[35,51-54] may be due to the presence of either T1A2-xLnxCuO5 orTlA 2CaCu207 or both[33,52].

ii. Substitutions in the Tl2Ba2Can-lCunO2n+4 Family

Following the discovery of superconductivity in the TI-Ba-Ca-Cu-O systems,attempts to prepare Sr analogues have been undertaken. Superconductivity wasreported[57-601 in some Tl-Sr-Ca-Cu-O phases, however, further confirmationof this is needed. Partial substitutions of Ba by Sr in the 2212 and 2223 systemwere studied by Hayri et al117l and Soeta et al.[191, respectively. It has been foundthat in both systems, increasing the Sr content will decrease the Tc and that thelattice parameters of 2212 and 2223 decrease linearly with increasing Sr content.The reasons for the depression of Tc by Sr substitution in both 2212 and 2223oxides are not clear at present. It is possible that the substitution of Ba by Sralters the TI-O-Cu bonding characteristics, resulting in an unfavorableelectronic communication between the Tl-O and Cu-O layers, and subsequentdepression in the superconducting properties.

Substitutions of TI by Bi or Pb in the double TI-O layer compounds werewithout success, due to the formation of BaBiO3 or BaPbO 3[82,951. Attempts tosubstitute K for Ti, in the TI-O layers were carried out by Sequeira et al[20].These authors supposed that the valence of Ti in TI-O layers is a strongadmixture of T13+ and Tl1+ and hence assumed that K can be used to replacemono-valent thallium. However, their work does not provide conclusiveevidence that such substitutions do take place, since there are no correspondingchanges in the structural parameters as inferred from their neutron diffractionstudy[201. Potassium sublimes at the high reaction temperatures employed intheir synthesis, and an off-stoichiometric composition can still yield 2212 or2223-type oxides, which can account for the superconducting properties of K"substituted" phases.

Bourgault et al. reported partial substitution of Ba2 by T1 in theTI 2 x, 3TlI,.xBal+xLnCu 208 (Ln=Pr and Nd) system[21]. A nonsuperconductingpure phase sample was isolated corresponding to x-0.25. Based on powder X-raydiffraction data, the authors concluded that TI and Ba2 ions are randomlydistributed at th- Ba sites in the 2212 related structure. This observation impliesthat partial occupancies of Tl11 at the Ba sites provide another source ofnonstoichiometry in addition to: a) apparent T13+ and Ca 2+ deficiencies, and b)partial occupancy of Ca 2 on the T13+ site in the Tl-based cuprates.

iii. Others

19

Most substitutional studies focused on various replacements in the TI, Baand Ca sites. A few investigations were also carried out on substitution for theCu site, primarily by Fe ions for M6ssbauer spectroscopic studies[96l. One of thereasons for the lack of other transition metal substitutions at the Cu-sites maybe related to the difficulty in preparing samples at the relatively low reactiontemperatures and short reaction times that are required to prevent the loss ofT1. However, if the synthetic difficulties can be overcome, it would beworthwhile to study substitutions at the Cu site, which would provide clues,vital for the understanding of electronic correlations between the T1-O and Cu-0 layers.

6. Physical Properties and T, Correlations

Due to the difficulties encountered in preparing high quality samples andgrowing large single crystals of thallium-copper oxide superconductors, thephysical properties have not been well established. There are somediscrepancies surrounding tLe superconductivity and other related physicalproperties of these compounds, probably due to subtle differences (such asinhomogeneity and nonstoichiometry) in the samples and measurementconditions. For instance, in the 2201 system, regardless of the crystal symmetry,the superconducting transition temperatures have been reported to beanywhere between 0-90K. In addition, there is no general agreement on thesuperconducting transition temperatures of polycrystalline 2212 and 2223specimens. Measurements on impure/mixed phase samples can often lead tomisleading conclusions and compound the confusion, which is already presentin thallo-cuprates. Although the correlations between the chemistry andsuperconductivity in these phases are far from being established, most of theknown physical properties of the thallium-copper oxides are at leastqualitatively comparable with those of the rare-earth based copper oxides.

i. Transport and Magnetic Properties

The thallium copper oxides are metallic in their normal state, including thenonsuperconducting members of TlA 2CuO5(A=Ba and Sr). TlBa2LnCu2Oy isthe only known exception, which is semiconducting between 4-300K. Thenormal state resistivity of thallo-cuprates is very similar to that of the 123-type

materials (of the order of -10 -3 fj"cm at room temperature). Thermoelectricpower measurements[97,981 on polycrystalline 2223 samples show a positiveSeebeck coefficient, which indicates that the dominant charge carriers in thesecompounds are holes. The thermoelectric power is small, about 2-4 4V/K, andincreases linearly with decreasing temperature upto Tc. A small value ofthermal power is in agreement with the observed metallic property, however,the temperature dependence of the Seebeck coefficient can not be explained bysimple band models.

20

Hall effect measurements[99,100], provide further evidence that holes arethe major charge carriers in the superconducting thallium cuprates. The chargecarrier concentration in a 2223 sample, calculated from the Hall coefficient, wasestimated to be -2 x 1021 /cm 3, somewhat smaller than that of rare-earth-basedhigh Tc compounds (n - 6 x 1021 /cm 3). The complex RH I/T temperaturedependence, however, has not been explained.

Although the thermoelectricity and Hall effect measurements indicate thatthe major charge carriers are holes in 2212 and 2223 phases, Vijayaraghavan etal., recently, reported possible electron superconductivity inT1Cal.xLnxSr2Cu2O7-8 when x=0.25[33]. Superconductivity was found in thissystem and the Tc and the type of the charge carriers were found to bedependent on the composition and the nature of the rare-earth ions.

The measured transport properties are somewhat qualitative, because thepolycrystalline powder samples used for the measurements usually containimpurities, which may affect the quantities of physical properties and theirtemperature dependencies. Measurements on single crystal samples arepreferred.

Critical current densities for 2212 and 2223 phases were measured in thinfilm and polycrystalline powder samples. In a 2212 thin film[101], the transportcritical current densities J, are 103 A/cm2 at 95K, 104-5 A/cm 2 at 77K and -106A/cm 2 at 4K; while in a powder sample[102], J, is only -300 A/cm 2. The lowercritical field Hcl is about -3 x 10-2 T, and the upper critical field Hc2 is about 50 Tat 77K for a 2223 sample. The critical fields Hcl and Hc2 are estimated to be -5 x10-2 T and 100-130 T, respectively at OK. The upper critical field of 2223 is muchlower than 205 T, estimated as the Pauli paramagnetic limiting field. From themagnetization studies on 2223[103,104], the coherence length 4 was estimated tobe -26A at 77K and 16-18A at OK. The penetration depth, X is atout -2200A at77K and 1400A at OK. Magnetic relaxation[105] and flux pinning[1021 have alsobeen investigated.

Surprisingly, in most of the thallium-copper oxides, evidence ofantiferromagnetic ordering was not found, except in TlBa2YCu2O7. Lowtemperature neutron diffraction study on semiconducting TIBa2YCu 207 showsevidence of antiferromagnetic ordering[106]. A three dimensional magneticordering, similar to that seen in YBa2Cu3O 6 and Bi2Sr2YCu2Ox, was establishedin this compound with the Nel temperature >350K. The absence of magneticordering in most of the thallium cuprates raises concern about the proposedrelationship between superconductivity and antiferromagnetic ordering.

ii. Electronic Structures

The electronic band structures of T12Ba2Can-lCunO2n+4 (n = 1, 2 and 3) serieswere calculated by several groups[107-109]. In general, the electronic band

21

structures are similar to those of the Bi-based compounds[110,111]. The focus ofthese studies has been on the special features of electronic structure introducedby the Tl-O layers . Although, in the various calculations, differentapproximation schemes were used, the calculated results are comparable anduniform in some features. Figure 9 shows a schematic diagram of theT12Ba2CuO6 band structure.

Bands I and II are the Tl-O antibonding bands, consisting of T16s and 02porbitals. III and IV are the Cu-0 bands, made up of Cu3d and 02p orbitals. Asthe number of Cu-O layers increases in the structure, all the correspondingbands remain at essentially the same energies, but the degeneracy of Cu-O bands(III and IV) increases. The presence of two Cu-O bands is a distinctive feature inall of the thallium-copper oxides. The broad band (band III, in Fig. 9) isessentially the Cu-O band arising from the Cu-O equatorial plane whichdisperses across the Fermi level, indicating that this band is the mainconduction band. The narrow Cu-O band is of interest. It is very close to orpossibly slightly overlaps with the Fermi level. If this narrow band doesoverlap with the Fermi level, it may provide electrons to the Fermi level. Themost important feature of the Tl-O bands is that they may cross Ef, and provideholes to the Cu-O bands. If this is the case, the TI-O layers are expected

Ef

r X P G3 Z r GI

Figure 9. Schematic band diagram of T12Ba2CuO6 (after ref. 109).

to be metallic. The broad Cu-O conduction bands are relatively insensitive tostructural changes, such as deviation from the ideal stoichiometry, local

22

distortion or substitutions. However, the Tl-O bands and the narrow Cu-Obands are very sensitive to small structural changes. An isoelectronicsubstitution of Sr for Ba will dramatically change the band structure near theFermi level, as revealed by the test calculations[1091.

The calculated electronic band structures provide some insight for theunderstanding of the physical properties of thallium-copper oxidesuperconductors. First, they present a mechanism by which Cu 2 8 might beproduced, in addition to the nonstoichiometry or structural defects. Second, thecoexistence of broad Cu-O and narrow TI-O bands near the Ef reveal thatelectronic conduction in these compounds may occur by plasmon-like(orexciton)[109] or charge fluctuation(transfer)-like mechanisms[112]. Third, theband structures show that the sizes of either the hole reservoir(the TI-O bands)or the electron reservoir(the narrow Cu-O bands) are not monotonic functionsof the number of Cu-O layers[109], which indicates that the basic band structuredoes not change with the number of Cu-O layers. Finally, these calculationsdemonstrate great similarities among the band structures of cupratesuperconductors; this implies that they all may have a similar conductionmechanism.

Herman et al.[1131 reported band structure calculations on the single Tl-Olayer compounds. Their results show features of the Cu-O band which arecomparable to those seen in the double Tl-O layer compounds. It was suggestedthat the T1-O narrow band can be moved away from Ef without significantlyaltering the Fermi surface, implying that the TI-O layers are not necessarily thehole reservoirs.

The band structures, however, can provide only very qualitativeinformation about the relationship between the chemistry andsuperconductivity in these compounds, given the fact that these calculationsoversimplify the complex structural features, such as nonstoichiometry andlocal structural distortions. Therefore, oversimplification of the actual structurein order to perform the calculation may lead to ambiguities. For example, mostof the calculations do not take into consideration the electronic spin functions,so that a calculated metallic state could actually be an antiferromagneticinsulating state.

Electron photoemission experiments[91-931 often provide some confirmationof the calculated band structures. These experiments indicate that the electronicstates near the Fermi level are composed of mainly Cu3d, 02p and the T16sorbitals.

iii. Tc Correlations

The discovery of high Tc superconductivity in the cuprate materials haschallenged the conventional BCS theory of superconductivity, because BCS

23

theory fails to predict such high critical temperatures. It has been established byflux quantization measurements that in these high Tc materials the thesuperconducting current carriers are paired holes/electrons. To date, there hasnot been a general consensus on the superconducting mechanisms andcorrelations between the chemistry of materials and the superconductivity.

However, some common features have been fairly well established in thesuperconducting cuprates: a) all the cuprates are nonstoichiometriccompounds, containing oxygen or cation defects; b) mixed Cu valence in thecompounds provides the charge carriers; c) all of the compounds haveorthogonal (tetragonal or orthorhombic) symmetry, providing the mosteffective metal3d-oxygen2p overlapping for conduction; d) the superconductingcuprates are poor metals (p-10 - 3 0-cm at room temperature) in their normalstate, which indicates that they are near a metal-semiconductor boundary; ande) in some cases, the semiconducting form of the superconducting phase ordersantiferromagnetically at ambient temperatures, which implies that thesuperconducting mechanism may be related to the magnetic properties.

In the thallium-copper oxides, all of the above features are present, inaddition to the effects due to the Tl-O layers. It _ppears that thesuperconductivity is correlated with elect-onic interactions between the Tl-Oand Cu-O layers. Many structural properties could also be considered to berelated to the transition temperature, including the in-plane and apical Cu-Odistances, the polarizability of the lattice, the distance between Cu-O layers,number of Cu-O layers, the location and concentration of charge carriers.Although, there have been insufficient data to support any of thesecorrelations, such empirical data could provide further understanding of thesematerials.

7. Future Work

The new superconducting oxide family of TlmA2A'n-lCunO2n+2+n presentsmany structural and electronic features in common with othersuperconducting cuprates and provides new opportunities for understandingthe chemical, structural and electronic property relationships required for hightemperature superconductivity. However, more work is needed to fullyunderstand details of the structural defects including oxygen content/orderingand distribution of cation deficiencies in these compounds. Novel substitutionstudies would be helpful in clarifying some of these issues.

Synthetic techniques, which can reproducibly yield pure phase samples withcontrolled stoichiometry, need to be developed. Investigations of new ternarylayered thallium transition metal oxides would be useful for modeling the roleof T1-O layers in the Tl-based cuprate materials.

24

Acknowledgments

We would like to thank Mr. M. H. Pan and Miss. Z. Zhang for providingtheir results for this paper. This work was supported by the Office of NavalResearch and National Science Foundation-Solid State Chemistry Grant DMR-87- 14072.

Ref erences

[I] J.G. Bednorz and K.A. Muller, Z. Phys. B -64, (1986) 189.[21 M.K. Wu, J.R. Ashburn, C.J. Torng, P.R. Hor, R.L. Meng, L. Gao, Z.J. i.uang, Y.Q. Wang and

C.W. Chu, Phys. Rev. B 58, (1987) 908.(31 R.J. Cava, B. Batlogg, R.B. van Dover, D.W. Murphy, S. Sunshine, T. Siegrist, J.P.

Remeika, E.A. Rietman, S. Zahurak and G.P. Espinosa, Phys. Rev. B 58, (1987) 1676.[4) Z.Z Sheng and A.M. Hermann, Nature 332, (1988) 55.(51 Z.Z. Sheng and A.M. Hermann, Nature 332, (1988) 138.[61 S.S.P. Parkin, V.Y. Lee, E.M. Engler, A.I. Nazzal, T.C. Huang, G. Gorman, R. Savoy and R.

Beyers, Phys. Rev. Lett. 60, (1988) 2539.[71 C.C. Torardi, M.A. Subramanian, I.C. Calabrese, J. Gopalakrishnan, K.J. Morrissey, T.R.

Askew, R.B. Flippen, U. Chowdhry and A.W. Sleight, Sience 24 0,0l9 88 ) 631.[81 C.C. Torardi, M.A. Subramanian, J.C. Calabrese, J. Gopalakrishnan, E.M. McCarron, K.J.

Morrissey, T.R. Askew, R.B. Flippen, U. Chowdhry and A.W. Sleight, Phys. Rev. B 38,(1988) 225.

[91 I.B. Parise, 1. Gopalakrishnan, M.A. Subramanian and A.W. Sleight, 1. Solid State Chem.76,0(988) 432.

[101 R. Beyers, S.S.P. Parkin, V.Y. Lee, A.I. Nazzal, R. Savoy, G. Gorman, T.C. Huang and S.LaPlaca, Appi. Phys. Lett. 53, (1988) 432.

[111 S.S.P.Parkin, V.Y. Lee, A.!. Nazzal, R. Savoy, T.C. Huang, G. Gorman and R. Beyers,Phys. Rev. B 38, (1988) 6531.

[121 A.W. Hewat, P. Bordet, J.1. Capponi, C. Chaillout, J. Chenavas, M. Godinho, E.A. Hewat,J.L. Hodeau and M. Marezio, Physica C 156, (1988) 369.

[131 T.C. Huang, V.Y. Lee, R. Karimi, R. Beyers and S.S.P. Parkin, Mat. Res. Bull. 23,(1988) 1307.

[14] K.V.Ramanujachary, S. Li and M. Greenblatt, Phvsica C to be published.(151 M.A. Subramanian, J.C. Calabrese, C.C. Torardi, J. Gopalakrishnan, T.R. Askew, R.B.

Flippen, K.J. Morrissey, U. Chowdhry and A.W. Sleight, Nature 332, (1988) 420.[161 D.E. Cox, C.C. Torardi, M.A. Subramanian,J. Gopalakrishnan and A.W. Sleight, Ph'ys.

Rev. B 38, (1988) 6624.[171 E.A. Hayri and M. Greenblatt, Physica C 156, (1988) 775.-[181 M. Kuroda and M. Araki, 112n. 1. Ap212. Phys. Lett. 28,0(989) L1154.[191 A. Soeta, T. Suzuki, S. Takeuchi, T. Kamo, K. Usami and S.P. Matsuda, 112n. 1. Ap21I. Phvs.

Lett. 8, (1989) L1 186.[201 A. Sequeira, H. Rajagopal, I.K. Gopalakrishnan, P.V.P.S.S. Sastry, G.M. Phatak, J.V.

Yakhmi and R. M. lyer, Physica C 156, (1988) 599.[211 D. Bourgault, C. Martin, C. Michel, M. Hervieu and B. Raveau, Physica C 158, (1989) 511.[221 M. Kikuchi, T. Kajitani, T. Suzuki, S. Nakajima, K. Hiraga, N. Kobayashi, H. Iwasaki, Y.

Yono and Y. Mu to, Ipn. T. Ap212. Phys. 28, (1989) L282.[231 Y. Shimakawa, Y. Kubo, T. Manako, Y. Nakabayashi and H. Igarashi, Physica C 156,

(1988) 97.

25

[241 Y. Tang, B. Lin, D. Zhou, W. Zhu, F. Chen, N. Li, K. Chen and C. Lu, Modemn Phys. Lett. B3,0(989) 853.

[251 M. Hervieu, A. Maignan, C. Martin, C. Michel, J. Provost and B. Raveau, Modern Phys.Lett. B2, (1988) 1103.

[261 S.S.P. Parkin, V.Y. Lee, A.I. Nazzal, R. Savoy and R. Beycrs and S. J. LaPlaca, Phys. Rev.Left 61, (1988) 750.

[271 T. Manako, Y. Shirnakawa, Y. Kubo, T. Satoh and H. Igarashi, Physica C 158, (1989) 143.[281 B. Domeng~s, M. Hervieu and B. Raveau, Solid State Commun. 69, (1989) 1085.[291 A. K. Ganguli, G. N. Subbanna and C. N. R. Rao, Physica C 156, (1988) 116.[301 T. Manako, Y. Shimakawa, Y. Kubo, T. Satoh and H. Igarashi, Phvsica C 156, (1988) 315.[311 C. Martin, D. Bourgault, C. Michel, MI. H-erveau and B. Raveau, Mod. Phys. Lett. B 3, (1989)

93.[321 J. Liang, Y.L. Zhang, G.H. Rao, X.R. Cheng, S.S. Xie, and Z.X Zhao, Solid State Commun.

70, (1989) 661.[331 R. Vijayaraghavan, A. K. Ganguli, N. Y. Vasanthacharya, M. K. Rajumon, G. U. Kujlkarni,

G. Sankar, D. D. Sarma, A. K. Sood, N. Chandrabhas and C. N. R. Rao, Supercond. Sci.Technol. 2, (1989) 195.

[34] J. K. Liang, Y. L. Zhang, G. H. Rao, X. R. Cheng, S. S. Xie and Z. X. Zhao, Z. Phys. B 76,(1989) 277.

[35] Z.Z. Sheng, L. Sheng, X. Fei and A.M. Hermann, Phys. Rev. B 39, (1989) 2918.1361 M.A. Subramanian, J.B. Parise, J.C. Calabrese, C.C. Torardi, J. Gopalakrishnan and A.W.

Sleight, V. Solid State Chem. 77, (1988) 192.[37] B. Morosin, D.S. Ginlev, J.E. Schirber and E. L. Venturini, Physica C 156, (1988) 587.[381 P. Haldlar, K. Chen, B. Maheswaran, A. Roig-Janicki, N. K. Jaggi, R.S. Markiewicz and

B.C. Giessen, Science 241, (1988) 1198.[391 H. Ihara, R. Sugise, M. Hirabavashi, N. Terada, M. Jo, K. Hayashi, A. Negishi, M.

Tokumoto, Y. Kimura and T. Shimomura, Nature 334,0(988) 510.[40] R. Sugise, M. Hirabayashi, N. Terada, M. Jo, T. Shimomura and H. Ihara, lpn. T. App12.

Phys._27, (1988) L1709.[411 J. Q. Huang, J. K. Liang, Y. L. Zhang, S. S. Xie, X. R. Chen and Z. X. Zhao, Modemn Phys.

Lett. B 3, (1989) 41.[421 R. Sugise, M. Hirabayashi, N. Terada, M. Jo, T. Shimomura and J. Ihara, Phvsica C 157,

(1989) 131.[431 S. Nakajima, NI. Kikuchi, Y. Syono, T. Oko, D. Shindo, K. Hiraga, N. Kobayashi, H.

Iwasaki and Y. Muto, Physica C 158, (1989) 471.[441 R. Sugise and H. Ihara, [12n. 1. A;2p2. Phvs. 28, (1989) 334.[451 S. Li, M. Pan and M. Greenblatt, to be published.[461 P. Haldlar, A. Roig-Janicki, S. Sridhar and B.C. Giessen, Mater. Lett. 7, (1988) 1.1471 S. Li and M. Greenblatt, Physica C 157, (1989) 365.[481 P. Haldar, S. Sridhar, A. Roig-janicki, W. Kennedy, D.H. Wu, C. Zahopoulos and B.C.

Giessen, J. Superconductivity 1, (1988) 211.[491 0. Inoue, S. Adachi and S. Kawashima, 112n. 1. AppI. Phvs. 28 , (1989) L1 167.[501 J. C. Barry, Z. Iqbal, B. L. Ramakrishna, R. Sharmna, H. Eckhardt and F. Reidinger, 11 111

Phys. 65, (1989) 5207.[511 T. Itoh and H. Uchikawa, Phys. Rev. B 39, (1989) 4690.[521 T. Itoh and H. Uchikawva, Jp2n. T. A12p2. Phvs. 28, (1989) L200.[531 S. Adlachi, 0. Inoue, H. Hirano, Y. Takahashi and S. Kawashima, 112n. V. A1212. Ehy. . 28,

(1989) L775.[541 C.N.R. Rao, A.K. Ganguli and R. Vijavaraghavan, Phys. Rey. B 40, (1989) 2565.[551 M.A. Subramanian, C.C. Torardi, J. Gopalaknishnan, P.L. Gai, J.C. Calabrese, T.R. Askew,

R.B. Flippen and A.W. Sleight, Science 242, (1988) 249.[561 A.K. Ganguli, K.S. Nan jundaswamv and C.N.R. Rao, Physica C 156, (1988) 788.[571 W.L. Lechter, M.S. Osofskv, R.J. Soulen Jr., V. M. LeTourneau, E. F. Skelton, S. B. Qadri,

W. T. Elam, H.A. Hoff, R.A. Hein, L. Humphreys, C. Skowronek, A.K. Singh, J.V. Gilfrich,L.E. Toth and S.A. Wolf, Solid State Commun. 68, (1988) 519.

26

[581 S. Matsuda, S. Takeuchi, A. Soeta, T. Suzuki, K. Aihara and T. Kamo, 112n. 1. A12p2. Phs..27, (1988) 2062.

[591 Z.Z Sheng, A.M. Hermnann, D.C. Vier, S. Schultz, S.B. Oseroff, D.J. George and R.M.Hazen, Phys. Rv. 38, (1988) 7074.

[601 T. Nagashima, K. Watanabe, H. Saito and U. Fukai, [12n. f. App12. Phys. 27, (1988) L1077-[611 D. Bourgault, C. Martin, C. Michel, M. Hervieu, J. Provost and B. Raveau, 1. Solid State

Chm 8 (1989) 326.[621 P.T. Wu, R.S. Liu, J.M. Liang, W.H. Lee and L. Chang, L.J. Chen and C.T. Chang, Physica C

156,0(988) 109.[631 R.S. Liu, P.T. Wu, J.M. Liang and L.J. Chen, Phys. Rev. B 39, (1989) 2792.(641 M. Itoh, R. Liang, K. Urabe and T. Nakamura, 112n. 1. App!2. Phys. 27, (1988) L1672.[651 E. Ruckenstein and C.T. Cheung, T. Mater. Res. 4,0(989) 1116.[661 M. Hervieu, C. Michel, A. Maignan, C. Martin and B. Raveau, I. Solid State Chem. 74,

(1988) 428.[671 M. Hervieu, C. Martin, J. Provost and B. Raveau, f. Solid State Chem. 76, (1988) 419.[681 M. Hervieu, A. Maignan, C. Martin, C. Michel, J. Provost and B. Raveau, I. Solid State

[691 A. Schilling, H.R. Ott and F. Hulliger, Physica C 157, (1989) 144.[701 H. Takei, T. Kotani, T. Kaneko and K. Tada, In "Adv. in Superconductivity- Proceedings of

the Ist International Symposium on Superconductivi~ty'. (Eds. K. Kitazawa and T.Ishiguro), Springer-Verlag, Tokyo, (1989) p.229.

[711 Y. Liu, Y.L. Zhang, J.K. Liang, K.K. Fung, 1. Phys. C 21, (1988) L1039.[721 K. Hiraga, D. Shindo, M. Hirabayashi, M. Kikuchi, N. Kobayashi, and Y. Syono, IRDn.L.-

App!2. Phvs. 27, (1988) L1848.[731 M. Verwerft, G. Van Tendeloo and S. Amelinckx, Physica C 156, (1988) 607.[741 H. Ihara, R. Sugise, K. Hayashi, N. Terada, M. Jo, M. Hirabayashi, A. Negishi, N. Atoda,

H. Qyanagi, T. Shimomura and S. Ohashi, Phvs. Rev. B 38, (1988) 11952.[751 S. lijima, T. Ichihashi, Y. Shimakawa, T. Manako and Y. Kubo, [12n. T. App!. Phys. 27,

(1988) L837.[761 I.K. Gopalakrishnan, P.V.S.S. Sastry, H. Rajagopa], A. Sequeira, J.V. Yakhmi and R.M.

Iyer, Physica C in press.[771 L.E.H. McMills, S. Li, Z. Zhang and M. Greenblatt, to be published.[781 J.D. Fitz Gerald, R.L. Withers, J.G. Thompson, L.R. Wallenberg, J.S. Anderson and B.G.

Hyde, Phys. Rev. Lett. 60, (1988) 2797.[791 S. Iijima, T. Ichihashi, Y. Shimakawa, T. Manako and Y. Kubo, 11pn. 1. App!. Phvs. 27,

(1988) L1061.[801 P. Bordet, J.J. Capponi, C. Chaillout, J. Chenavas, M. Godinho, A. W. Hewat, E.A. Hewat,

J.L. Hodeau, A.M. Spieser, J.L. Tholence and M. Marezio, 1. Less-Common Metals 150, (1989)109.

1811 S. Amelinckx, C. Van Tendeloo, H.W. Zandbergen and J. Van Landuyt, 1. Less-CommonMetals 150, (1989) 71.

[821 H. W. Zandbergen, W.A. Groen, F.C. Mijlhoff, G.Van Tendeloo and S. Amelinckx, PhvsicaC-.156, (1988) 325.

[831 M. Hervieu, C. Michel and B. Raveau, I. Less-Common Metals 150,0(989) 59.[84] B. Domeng~s, M. Hervieu and B. Raveau, Phvsica C 68, (1988) 303.[851 STJ Hibble, A.K. Cheetham, A.M. Chippindale, P. Day and J.A. Hriljac, Physica C 156,

(1988)604.[861 H. W. Zandbergen, C. Van. Tendelo, J. Van Landuyt and S. Amelinckx, App!2. Phvs. A 46,

(1988) 233.[871 T. Zetterer, H. H. Otto, C. Lugert and K. F. Renk, Z. Phys. B 73, (1988) 321.[881 R.M. Iyer, G.M. Phatak, K. Gangadharan, M.D. Sastry, R.M. Kadam, P.V.P.S.S.Sastry and

J.V. Yakhmi, Physica C 160, (1989) 155.[891 Y. Tokura, H. Takagi and S. Uchida, Nature 337, (1989) 345.1901 H. Sawa, S. Suzuki, M. Watanabe, J. Akirnitsu, H. Matsubara, H. Watabe, S. Uchida, K.

Kokusho, H. Asano, F. Izumi and E. Takayama-Muromachi, Natu re 337, (1989) 347.

27

[911 H.M. Meyer 111, T.J. Wagener, J.H-. weaver and D.S. Ginley, Phs e.B.9 (1989) 7343.1921 P. Marksteiner, J. Yu, S. Massidda, A.J. Freeman, J. Redinger and P. Weinberger, Ph2 Rv

B 39, 1989) 2894.[931 T. Suzuki, M. Nagoshi, Y. Fukuda, S. Nakajima, M. Kikuchi, Y. Syono and M. Tachiki,

1PhyXicaC~ in press.1941 I.D. Brown, I n "Structure and Bond in& in Crystals 11", (Eds. M. O'Keeffe and A.

Navrotsky), Academic Press, Inc. New York (1981) p. I.1951 Y. Torni, T. Kotani, H. Takei and K. Taja, 1I2n. I. Api. Phys. 28, (1989) 1190.[96] Y. Zeng, X. Yan, M. Lin, Y. Ren, X. Meng, Y. Lu, Q. Tu, X. Wu and L. Sang, Phys. Rev. B

40,01989) 727.[971 L. Alc~cer, M. Almeida, U. Braun, A.P. Conqalves, S.M. Green, E.B. Lopes, H.L. Luo and C.

Politis, Mod. Phys. Lett. B 2, (1988) 923.[981 N. Mitra, J. Trefny, B. Yarar , G. Pine, ZZ Sheng and A.M. Hermann, Phys.Rev. B 38,

(1988) 7064.[991 J. Clayhold, N.P. Ong, P.H. Hor and C.W. Chu, Phys. Rev. B 38, (1988) 7016.[1001 P. Mandal, A. Poddar, A.N. Das, B. Ghosh, and P. Choudhury, Phvs. Rev. B 40, (1989) 730.[1011 J.F. Kwak, E.L. Venturini, R.J. Baughman, B. Morosin and D. S. Ginley, Physica C 156,

(1988) 103.[1021 J.R. Thompson, J. Brynestad, D.M. Kroeger, Y.C. Kim, S.T. Sekula, D.K. Christen and E.D.

Specht, Phvs. Rev. B 39, (1989) 6652.[1031 H. Iwasaki, N. Kobayashi, Y. Koike, M. Kikuchi, Y. Syono, K. Noto and Y. Muto, 12n.L

A12p2. Phys. 27, (1988) 1631.[1041 H. Kumakura, K. Togano, K. Takahashi, H. Shimizu, M. Uehara, H. Maeda and M.

Nakao, 112n 1. Ap212. Phys. 27, (1988) L857.[1051 M. E. McHenry, M. P. Maley, E. L. Ventunini and D. L. Ginley, Phys. Rev. B 39, (1989) 4784.[106] J. Mizuki, Y. Kubo, T. Manako, Y. Shimakawa, H. Igarashi, J.M. Tranquada, Y. Fujii, L.

Rebeisky and G. Shirane, Physica C 156, (1988) 781.[1071 J. Yu, S. Massidda and A.J. Freeman, Physica C 152,0(988) 273.(1081 D.R. Hamann and L.F. Mattheiss, Phys. Rev. B 38,0(988) 5138.[1091 R.V. Kasowski, W.Y. Hsu and F. Herman, Phys. Rev. B 38, (1988) 6470.[1101 H. Krakauer and W.E. Pickett, Phvs. R&y. Lett. 60, (1988) 1665.[1111 J. Ren, D. Jung, M.-H. Whangbo, J.-M. Tarascon, Y. Le Page, W. R. McKinnon and C.C.

Torardi, Physica C 159, (1989) 151.[1121 M. Singh and K.P. Sinha, Solid State Commun. 70, (1989) 149.[1131 F. Herman and R. V. Kasowski, Physica C, in press.