Flux growth of Sr1−xCaxCuO2 single crystals

Flux growth of Sr1−xCaxCuO2 single crystals

NH CRYSTAL GROWTH Journal ofCrystal Growth 140 (1994) 72—78 ELSEVIER Flux growth of Sr1 _~Ca~CuO2 single crystals C.T. Lin ~, W. Zhou, A.P. Mackenzi...

1MB Sizes 4 Downloads 160 Views

NH CRYSTAL GROWTH Journal ofCrystal Growth 140 (1994) 72—78

ELSEVIER

Flux growth of Sr1 _~Ca~CuO2 single crystals C.T. Lin ~, W. Zhou, A.P. Mackenzie, F. Gauthier, W.Y. Liang IRC in Superconductiiily, UnOersity of Cambridge, Cambridge CB3 OHE, UK (Received 4 November 1993; manuscript received in final form 10 January 1994)

Abstract A series of single crystals of Sr1 _~Ca~CuO2, x = 0.19—0.83, has been grown using a low cooling rate in a flux containing excess CuO. Values of the temperature gradient, amount of flux, soaking temperature, soaking time, crucible material and cooling rate have been optimized. The chemical compositions for the as-grown crystals were determined by means of energy dispersive X-ray analysis (EDS) in a transmission electron microscope and confirmed by electron probe microanalysis (EPMA), while the SrCuO2-type structure having orthorhombic symmetry is confirmed by high resolution transmission electron microscopy (HREM). Both the Ca/Sr ratios in the crystals and the unit cell constants vary with a linear relationship with the Ca/Sr ratios in the initial melt.

1. Introduction

reports

Since the first reports of the “infinite-layer” (Sr,Ca)Cu02 compound which consists of alternate layers of Cu02 and (Sr,Ca) by Greaves [1]

and Siegrist et al. [21, many researchers have attempted to synthesize such a defect-type perovskite in the system Sr1 ~Ca~CuO2 and to make the compound superconducting. In 1991, the first “infinite-layer” electron-doped superconductor, Sr1 _1Nd~CuO2[3], having 1~at 40 K was synthesized using a high pressure technique. This was followed by the discovery of the first p-type “infinite-layered” superconductor with T~ near 110 K and a nominal composition (Sr0 7Ca~3)()9CuO2 [41,as well as other compositions in the series Sr1 ~Ca~CuO2 (x 0—0.95) [5,61, also synthesized by a similar high pressure method. Recent =

*

Corresponding author.

[5,7]suggest

that the superconductivity

may be related to intergrowth layers of unknown structure which fulfil the role of charge reservoirs. On the other hand, specimens prepared at low pressure have a different structure, known as the “SrCuO2” structure, in which square coordinated Cu02 units share edges with each other instead of the corner sharing configuration observed in the superconducting phase prepared at high pressure. Other methods, such as thin film growth [8—10]and conventional synthesis [11] have also been applied to form such compounds. However, nearly all the studies so far have been made on sintered specimens, or poorly characterized polycrystals. The reported compositions of the specimens were inferred from the starting composition of the powders, despite the inhomogeneous state of the specimens. This leaves considerable uncertainties when interpreting the electrical and superconducting properties of the samples.

0022-0248/94/$07.00 © 1994 Elsevier Science B.V. All rights reserved SSDI 0022-0248(94)00087-3

CT. Lin et al. /Journal of Crystal Growth 140 (1994) 72—78

In order to clarify the chemical and physical properties of the compound studied, it is desirable to work with single crystals. In this paper we report the development of a growth method for single crystals of Sr 1 ~Ca~CuO2, employing self-

73

tion was ground in an agate mortar. The compositions of the starting mixtures are given in Table 1. High density alumina crucibles (type 001, 5 cm diameter) were mainly used. In order to minimize

contamination, these were preheated at 1400°C for one day before they were used for crystal growth. ZrO2 and SnO, crucibles were also tried, but they were found to be less satisfactory.

flux to decrease the possibility of contamination as well as a low cooling rate to improve crystal quality and dimensions.

2.2. Growth procedure 2. Procedure for crystal growth and characterisation

A good environment for crystal growth is the presence of a sharp thermal gradient. Its effects are well known for the Bridgman and vapour

2.1. Sample preparation

transport methods. We have therefore carried out the growth in a 10 cm diameter tube furnace, working in the region with the strongest temperature gradient, between the central uniform ternperature zone and the ends of the tube. In the

The crystals were grown from a mixture of CaCO3, CuO and SrCO3 (3N or better). A total mass of 50 g having the desired starting composi-

Table I Starting compositions and compositions of the majority crystalline phase; numbers in brackets are standard deviations affecting the last digit of the concentration No

Starting composition of cation (at% fraction)

Composition of majority crystalline phase

Minority phases identified

SrO

CaO

CuO

Sr

Ca

Cu

30.0

3.3

66.7

0.80 (2)

0.19 (2)

1.02 (3)

(Sr,Ca)Cu02 CaSrO2

(Sr,Ca)Cu203 CuO. SrO 2

26.7

6.6

66.7

0.69 (4)

0.32 (2)

1.07(3)

(Sr,Ca)CuO,; CaSrO2 (Sr,Ca)Cu203 CuO, CaO

3

23.3

10.0

66.7

0.63 (3)

0.37 (3)

1.01 (2)

4

20.0

13.3

66.7

0.60 (4)

0.41 (2)

0.99 (2)

SrCurOy (x 0.06—0.2); Sr,Cu03 CuO SrCu,03 Sr,Cu,O~(Sr.Ca)~Cu7O~ Ca0 2Sr~3CuO;CaSrO3 CuO, (Sr,Ca)3Cu 205

5

16.7

16.6

66.7

0.50 (2)

0.50 (4)

1.00 (2)

(Sr,Ca)CuO,; CaCu,03 CaCuO,; CaSrO,;

(Sr,Ca)Cu,03 CuO 6

13.3

20.0

66.7

0.36 (2)

0.62 (3)

1.02 (2)

(Sr,Ca)CuO,; (Sr,Ca)Cu,03 CaCu203 (Sr,Ca)5CuO4: (Sr,Ca)5Cu,0,; CuO

7

10.0

23.3

66.7

0.32 (1)

0.70 (1)

0.98 (1)

(Sr,Ca)Cu02 (Sr,Ca)2Cu505 (Sr,Ca)5Cu04 (Sr,Ca),Cu20~

8

7.0

26.3

66.7

0.21 (4)

0.79 (5)

0.95 (2)

Ca,Sr03 (Sr,Ca)CuO,; (Sr,Ca)2Cu 305, (Sr,Ca)3Cu04 (Sr,Ca)3Cu,05 (Sr,Ca)4Cu3O7

74

C. T Lin et al. /Journal of Crystal Growth 140 (1994) 72—78

(EPMA). The results are in good agreement with the measurements by EDS.

12007 0.3 C/h 900

-

-

2.4. Structure and microstnicture study

-

600-

The structures and the lattice parameters of the as-grown crystals were determined by a focus-

30&

ing Guinier camera after grinding the crystal to a

//

I

50

370

Time (hrs)

Fig. 1. Temperature programme for growing single crystals: (1) cool from 1030 to 980°Cat 1—2°C/h;(2) cool from 980 to 900°C at a rate of 0.3°C/h for growth; (3) further cool to room temperature at 50°C/h.

powder and mixing with silicon as a calibration standard. High resolution electron microscopy (HREM) was performed on a JEOL EM-200CX operating at 200 keV to study the detailed structure.

3. Results and discussion 3.1. Flux growth

horizontal direction, it was found that the temperature was highest at the centre and decreased

The difficulty in crystal growth is to obtain the

towards both ends symmetrically. A second thermal gradient was also found along the radius of the tube with the furnace walls being approximately 8°C hotter than the centre. This furnace therefore provides us with strong thermal gradients along both the horizontal and vertical direc-

optimum growth parameters of temperature gradient, amount of flux, soaking time, soaking temperature, crucible material, cooling rate and crystal separation, because of the complexity of the flux growth mechanism in multi-element oxides. Our aim is to investigate these growth parameters

tions. In the horizontal direction the gradient was 20°C/cm in the two 10 cm long regions either

and to obtain the optimum growth conditions.

side of the central zone and remained unchanged in the temperature range between 900 and 1000°C used for the crystal growth. The crucibles con-

3.1.1. CuO flux A traditional growth method using CuO as the flux was employed. This self-flux method lowers

taming the charge were placed in these high thermal gradient regions and heated in air to

the melting point and minimizes contamination since no other elements are introduced. We in-

1030°C at a rate of 200°C/h. They were then allowed to soak for 24 h, followed by a growth

vestigated the possibility of influencing the cation ratio of the resulting crystals by adding excess CuO. It was found that increasing the amount of CuO flux resulted in higher levels of CuO inclu-

procedure in three steps, as shown in Fig. 1.

2.3. Compositional analysis We carried out compositional analysis by en-

ergy dispersive X-ray spectrometry (EDS) on a JEOL 2010 CX electron microscope. Single crys-

sions as well as increasing the amount of secondary phases. We have, therefore, kept to only a small excess of CuO, as shown in Table 1,

in order to obtain good quality crystals of Sr1 _~Ca~CuO2.

tals and other specimens taken from the residual

melt were ground and 20 fine particles were chosen randomly for examination. Al contamina-

3.1.2. Soaking temperature and soaking time The effect of the composition of the starting

tion was not observed to the detection limits of 2

material on the Ca/Sr ratio in the crystals was

at%. The compositions of the single crystals were also examined by electron probe microanalysis

investigated in a series of samples, as listed in Table 1. During these experiments it was found

( - f. liii

it

al. ‘ .lournal o/ Crvital Growth 140 (1904) 2

8

75

1cm

1mm

F 1g. 2. 1 a) T~pica I is—en n~ii Single rvsta Is ol Sr

0 ,, ,Ca II

~_

( ii 0

,

and ( b I a cleas ed single crvsta I nt Sr1

Ca,, ( u()

76

C. T Line! al. /Journal of Crystal Growth 140 (1994) 72-78

that a high Ca/Sr ratio in the initial melt led to a higher melting point, and thus a higher soaking ratio is less than 1 (runs 1—4) melting can be completed by soaking at 1030°Cfor one day and temperature would be necessary. When the Ca/Sr the end product consists of a flat solidified melt in the crucible. If the ratio was greater than or equal to 1 (runs 5—8), complete melting could not be achieved at this temperature and the solidified melt consisted of a condensed mixture of solid particles. By raising the soaking temperature to 1060°C,complete melting was then achieved and these runs were therefore soaked at this higher temperature for the same length of time. 3.1.3. Crucible Because of the high soaking temperature and low cooling rate, crucibles of Al203, Sn02 and Zr02 were found to be corroded by the very aggressive melt. Our experiments confirmed that for soaking and growing time of over 100 h, particles such as (Sr,Ca)Sn03, (Sr,Ca)Zr03 and (Sr,Ca)Al03, would be formed in the melt, causing a high viscosity and finally stopping the growth of the single crystals of Sr1 _~Ca5CuO2.Alumina is the most corrosion-resistant of the three and was chosen for the crystal growth project. 3.1.4. Cooling rate In an effort to minimize the Al contamination, we worked with a relatively short soaking time of 24 h, and used a high cooling rate of 1—2°C/h immediately after soaking to 980°C.It was found that a cooling rate of 0.3°C/h between 980 and 900°C was favourable for growing large crystals. Increasing the cooling rate in this temperature region to 0.5 and 1°C/hresulted in a lower yield of crystals a cooling rate the > 1°C/hwould give no crystals and at all. Therefore low cooling rate of 0.3°C/his one of the optimum parameters for growing Sr1 _~Ca~CuO2 crystals. 3.1.5. Majority and minority phases

!

~ ~ ~

-

o

compositions of the initial materials and the final

major crystalline products as well as the other secondary phases identified are listed. The single

3

2

4

Ca/Sr ratio in initial melt

Fig. 3. Ca/Sr ratio in Sr 1 _rCa~Cu02 single crystals as a tunction of the Ca/Sr ratio in the initial melt.

crystals formed as a majority phase, mainly at the bottom of the crucible. They can be seen to grow along the direction of the large thermal gradient. Thus crystal growth proceeds upward from the bottom as cooling continues. Some secondary phases are found on the surface of the melt, but the main product is (Sr,Ca)Cu02. As more calcium oxide was added (runs 5—8), the amount of secondary phases increased and the phases themselves became more complex. This is probably the result of the higher soaking temperature employed. Traces of CuO, CaO and SrO were also found due to incomplete reaction during melting. 3.2. Crystal morphology and growth behatiour The as-grown crystals can be separated mechanically from the solidified flux and they show

4.4

230

4.2

~

221)

2 0

30

20))

3.6

iso

All eight experimental runs reported in Table

1 have successfully produced single crystals. The

i



0.!

~

0.2

0.3

0.4

0.5

(1.6

Ca ronteni in crssiais

1)7

(.1)

((.9

Fig. 4. Cell parameters a, 6, c and V versus the content (If Ca in the Sr1 ,Ca 1Cu0~single crystals showing the retention of the SrCu0~-typestructure.

CT. Lin et at /Journal of Crystal Growth 140 (1994) 72—78

a layered-type morphology with smooth and shiny surfaces. In most cases, crystals were found to form together in a large block up to several cm3 similar to the “rock-salt” shape and the (001) face can be easily cleaved using a scalpel, as

77

0.3 mm. The dimension in the b direction can be as long as 1 cm. Fig. 2b shows a typical crystal cleaved from a large block. It is found that Ca has a high solubility forming Sr 1...~Ca~CuO2single crystals with the wide range of 0.19
shown in Fig. 2a. The crystals grown with a slow cooling rate of 0.3°C/hhave thicknesses up to 0.5 mm (c direction), while higher cooling rates of 0.5 to 1°C/h produced thinner crystals of 0.1 to

• *

020



• 001



020•

S



• $

• •

.

(a)

:(b):

200’

*

• •

-

S

Fig. 5. Typical electron diffraction patterns of Sn)) 0Ca050Cu0 single crystal viewed down the (a) [00] and (b) [0011directions. (c) HREM image corresponding to (a).

78

CT Lin

Ct

al. /Journal of Crystal Growth 140 (1994) 72—78

However, it should be noted that when the Ca content is low, there is considerable deviation from this linear relationship, as shown in Table 1. The calcium content increases continuously in the crystal up to x = 0.83 which is greater than the limit of solubility in ceramic specimens, x 0.7, according to a previous report [10]. This high solubility may in part be due to calcium having an ionic radius (Ca~20.99 A) which is only slightly smaller than that of Sr~2(1.12 A). However, other factors must also be important, as the solubility rapidly diminishes at the high soaking ternperatures of 1030 and 1060°C. =

3.3. Structural characterisation X-ray diffraction studies of single crystals revealed that all the compounds have orthorhombic symmetry, with a 3.5428—3.4117 A,0 b 16.2746—16.0584 A and c 3.8993—3.8583 A. The unit cell dimensions are strongly affected by the Ca/Sr ratio, with all the cell parameters, a, b, c and V, decreasing continuously as Ca content increases, as shown in Fig. 4, in accord with the smaller size of Ca substituting for Sr. The SrCuO2-type structure of the as-grown single crystals was also confirmed by HREM. Fig. 5 shows two typical selected area electron diffraction patterns, viewed down the [100] and [001] zone axes and a HREM image on the [100] projection from Sr0 5~Ca050CuO2 single crystal. The diffraction patterns and HREM images from all other compositions are almost identical, although the intensities of the diffraction spots vary as would be expected. Our HREM studies did not reveal any defect structure. =

=

=

3.4. Electrical characteristics Although this paper is concerned with single crystal growth, some preliminary transport measurements have also been made. These measurements indicated that all the crystals were semiconducting, in agreement with previous findings for this type of structure. Several attempts have been made to increase the oxidation state of the crystals, such as annealing in flowing oxygen at 600°Cfor 5 days or quenching into liquid nitro-

gen, having been annealed in flowing oxygen for I day at temperatures of 600, 700, 800 and 850°C. None of these attempts resulted in inducing metallic behaviour. It remains to be seen whether a structural phase transition can be induced by annealing the samples at high pressures. It might also be of interest to investigate introducing defects in the crystals by means of radiation damage, for example. 4 C

1

usion Crystals of Sr ..XCa(CuO2 with sizes up to 3 can 1be grown using CuO as of self-flux with a cooling rate of 0.3°C/h at 900—980°C in several cm high density alumina crucibles. The solubility of Ca in Sr 1 5Ca~CuO2single crystals is high, with x up to 0.83. Comparing the compositions in the crystals with those in the starting mixtures, the slope of Ca/Sr ratios is as large as 0.95. The unit cell parameters a, b, c, and thus volume, V, decrease linearly with increasing Ca concentration in the crystals. OflC

Acknowledgements We thank A. Carrington for carrying out the resistance measurements.

References [1] C. Greaves, Nature 334 (1988) 193. [2] T. Siegnist, SM. Zahurak, D.W. Murphy and R.S. Roth, Nature 334 (1988) 231. [3] MG. Smith, A. Manthiram, J. Zhou, J.B. Goodenough and J.T. Markert, Nature 351 (1991) 549. [41M. Azuma, Z. Hiroi, M. Takano. Y. Bando and Y. Takcda, Nature 356 (1992) 775. [5] S. Adachi, H. Yamauchi, S. Tanaka and N. Mon. Physica C 208 (1993) 226. [6] B. Okai, Jap. J. AppI. Phys. 28 (1989) L2251. [7] Z. Hinoi, M. Azuma, M. Takano and Y. Takeda, Physica C 208 (1993) 286 [8] K. Kobayashi, Y; Ishihara, S. Matsushima and G. Okada. Jap. J. AppI. Phys. 30 (1991) L1931. [9] X. Li, M. Kanai, T. Kawai and S. Kawai. Jap. J. App!. Phys. 31(1992) L217. [10] X. Li T. Kawai and S. Kawai.. Jap. J. AppI. Phys. 31 (1992) L934. [11] MT. Gambardella, B. Domenges and B. Raveau, Mater. Res. Bull. 27 (1992) 629.