Diagarama de Fase de Los Cupratos

Prog. Crystal Growth and Charact. Vol. 36, Nos. 1-2. pp.
1-97, 1998
Pergamon © 1998 Published by Elsevier Science Ltd. All rights reserved
Printed in Great Britain
096D-8974/98 $19.00 + 0.00
PII: S 0 9 6 0 - 8 9 7 4 ( 9 8 ) 0 0 0 0 8 - 4
PHASE DIAGRAM AND ITS APPLICATION

TO THE CRYSTAL GROWTH OF
HIGH T c OXIDE SUPERCONDUCTORS
Chen Changkang
Clarendon Laboratory, Oxford University. E-mail: c.chenl @physics.ox.ac.uk

Contents
I. High Tc oxide superconducting phase and crystal growth 2

II. Phase diagram, phase rule and crystal growth 8
Ill. Experimental determination of phase diagram 17
1. Sample preparation 17
2. Thermal analysis and its improvements 21
3. Determination of crystallization temperature by decanting experiments 25
4. Other methods 25
5. Experimental conditions 27
( 1) Crucible 27
(2) Atmosphere 30
(3) Equilibration 30
IV. Application of phase diagram in the crystal growth of oxide superconductors 31
1. Phase diagram and crystal growth in multiple phases of Bi-Sr-Ca-Cu-O system 31
2. Peritectic reaction and seed for melt-texturing growth 37
3. Liquidus line and crystal growth of (La,Sr)2CuO 4 43
4. Selection of appropriate system for the crystal growth of YBCO 48
5. Phase diagram and growth technique 61
(l) Crystal growth of PSYCCO in an immiscible flux system 61
(2) Top-seeding technique 65
(3) Crystal growth of Lil+xTi2_xO4 and electric crystallization 68
(4) Eutectic temperature and crystal separation 70
6. Phase diagram and superconductivity 74
(1) Structure transition and superconductivity in YBCO crystal 74
(2) Crystal chemistry and superconductivity in Bi-2201crystal 78
(3) Substitution and superconductivity in BPBO crystal- 8O
(4) Cross substitution anti superconductivity in CLBCO crystal 83
V. Conclusions 86
Acknowledgement 88
References 89
Chen Changkang
PHASE DIAGRAM AND ITS APPLICATION TO THE CRYSTAL GROWTH

OF HIGH Te OXIDE SUPERCONDUCTORS
I. HIGH Te OXIDE SUPERCONDUCTING PHASE AND CRYSTAL GROWTH
Since the great discovery of the physical phenomenon of superconductivity in mercury by

Kamerlingh Onnes [1 ] in 1911, the development of superconducting materials has passed
from the simple metal elements (Hg, Sn, Pb, Tc, Nb..,), through the binary element
compounds (NbC, NbN, Nb3Sn, V3Si, Nb3Ge..,) to the complex oxide or organic
compounds. It was a breakthrough for the superconducting oxide by the report of rare
earth cuprate system, (La,Ba)2CuO 4 [2] , in 1986 and (La,Sr)2CuO 4 [3, 4] in 1987 with
Tc=25-40 K. Based on the cuprate system, the superconducting transition temperature
was raised rapidly by the discovery of a series of high Tc compounds:
YBa2Cu307_y(YBCO) [5] and other rare earth 123 phases with Tc above the boiling
temperature of nitrogen, non-rare earth cuprate system of Bi-Sr-Ca-Cu-O [6],
T1-Ba-Ca-Cu-O [7] and Hg-Ba-Ca-Cu-O [8] with highest Tc above 100 K. The progress of
superconducting materials is shown in Fig. 1.
,,o] • l t g - 1223
Sheng and Herman
Tl~Ba;Ca~Cu30~ o (Feb 88)
120 B,~S{~Ca~Cu30, o Maeda (Jan .58)
110
T12Ba~CaCu.Q e
10£
Y'Ba.gCu~O ~, Chu and Wu (Ma' 87)
9£
8i~SrgCaCu;© e .
8O . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Iio t,u(! ~,'
7£ layelec CuD,ale ~
~6C
LaSrCu© C~[J
5O cinder p,essure
1 ~, I:a ( J,f
rtt,1 C , ~ t S e o ' 91)
Be~Jnorz a,'~ ~ '.A Ilef iSep .b
I aBaCu
3O ~-~ BaKB,C,, iDa'd ~t~'
NDAIG¢~ ND C:
Nb:lS~ V]S' ~ . . . . .
2O
............... L,'K": ...... "~ ' ...... °"~'
10
14 I
191C 1930 t950 lq7p 1QQO ;' ~f!c "loar
Fig. I. The progress of Tc for superconducdn.~, matcrialx

High T c Oxide Superconductors 3
However, as the complexity of the composition for oxide superconductors increased, they
exhibited less and less stability. Nearly all of the nigh Tc oxide compounds decompose
before melting (incongruent melting). Fig. 2 shows the dependence of decomposition
temperature(Td) or melting temperature(Tm) for the cuprate superconductors with layer
structure compared with superconducting metals or intermetallic A I5 alloys with three
dimensional chain structure. Thermal and chemical stabilities of oxide superconductors are
much poorer than A15 alloys. For the same family, the decomposition temperature or
stable range of the superconducting phase reduces as Tc increases, which brings more
difficult5, to the material preparation and fabrication. Now we are facing two challenging
strategies -- to improve present compounds and to find new ones. The knowledge of
phase diagram is fundamental data to understand, to control and to improve the material
stabilitv. Phase diagram also pro~4des basic information to starting the preparation of new
type of materials.
%~o1(° C):
2400
2200 beSn
200O
Nb.Ge
1800
1600
1400 LSCO
NCCO "-"
1200
C L B C O "~- -~.
"\ BI>B0 * Y B C O "~ ---.
1000 \ I 1 2201r0~ - ] 1-2212 ~:ttg !_2:5
~i 22~J~\..... r'.~vcco~,r172~3
800 ~ Bi 2212 Bi 2')2'~
--- I~KB'] ~ '
0 2(1 /10 1) 1 8() I(l() 1'21} ]:l(I

"I, ( ~< )
Fig. 2. Superconducting transition temperature (Tc) and deconlposition (Td) or melting

temperature (Tin) of conventional alld layered cuprate superconductors
Phase diagram shows the relation t)ctwecn phases, components under wtri~ms conditions
(temperature, pressure and electric field elc.) at a state of thermodynamical equilibration.
AI therm()dvnamical e.quilibr,ition a system does not change, which includes: (a} The
lemperature does not ('hanfe .... thermal equilibration: (b) The pressure ~H ,~ther exlra
4 Chen Changkang
force does not change -- mechanical equilibration and (c) The chemical potential of any
component in all the phases is the same -- chemical equilibration. A phase is a
homogeneous body wi',h same composition, structure and properties. Different adjacent
phase are distinguished by a boundary where the chemical and physical properties change
discontinuously. A system may contain one or more phases, consisting of various element
or compound compositions. The independently changeable composition is cailed as
component. The main element compositions for the compounds are located at several
groups in the periodic table, which decides the characters of the high Tc oxide
superconductors and its relative phase diagram. As an oxide, the nonmetallic element is
oxygen. But sometimes it can be partially replaced by its neighbour element fluorine to
form metal oxyfluoride such as Sr2CuO2F2+y (Tc~46 K). And the partial substitution of
oxygen in Cu-O layer by carbonate CO3 -2 or nitrate NO3-2 groups may introduce new
superconducting compounds. The main metal components consist of four groups of
elements:
1. Copper metal: Cu (Ni)
2. Alkali and alkali earth metals: Li, Na, K, Rb and Ca, Sr, Ba
3. Y and rare earth metals
4. Heavy metals: Bi, T1, Hg, Pb, W, Mo, Re
The chemical properties of the elements in each group are similar, which decides some
common natures of the oxide superconducting phase and their phase relation as well.
From the chemical point of view, the oxide superconductors may be divided into the
following categories:
(a) Oxide compounds without copper, like LiTi20 4 [9] with spinel structure, Ba(Pb,Bi)O 3
(BPBO) [10] and (Ba,K)BiO 3 ( B K B O ) I l l ] with cubic structure. They are three
dimensional superconductors.
(b) Oxide compounds with copper-oxygen layer structure, in which the superconductivity
always exhibits a two dimensional nature. They can be further divided into two groups:
cuprates with or without rare earth elements. The late one holds the highest Tc record
above 100 K in the superconductor world.
The high Tc superconductors discovered in recent years have the following distinct
characters:
(1) Nonstoichiometry because of the present of component elements with changeable
valence. As shown in the Periodic Table, valences of alkali or alkali earth elements and Y
or some rare earth elements are normally fixed. But copper, h e a w metal elements and
other rare earth elements, even if the oxygen, have variable valences. The changing
situation is more complex for the superconductors than that indicated in the Periodic
Table. For instance, cuprate superconductors might have Cu +1 , Cu +2, Cu+3; Pb +l , Pb +2,
Pb +3, Pb +4 and Bi +3, Bi +4, Bi +5 in the compounds.
(2) Multiple phases in the same family, the same compound and even in the same
structure of the same compound. As well known, the 2201, 2212, 2223 phases in the"
High Tc Oxide Superconductors 5
Bi-Sr-Ca-Cu-O family easily coexist. And Bi-2201 has two phases, only one is
superconducting. The orthorhombic YBCO at least has two type of structures, orth-I and
orth-II, which correspond to two steps in superconducting transitions.
(3) Doped compounds with unequal valence substitution, such as (La,Sr)2CuO 4 (LSCO),
(Nd,Ce)2CuO 4 (NCCO), (Ca,La,Ba)Cu307_y (CLBCO), Pb2Sr2(Y,Ca)Cu308+y
(PSYCCO), Ba(Pb,Bi)O 3 (BPBO) and (Ba,K)BiO3 (BKBO). Superconducting behaviour in
these compounds occurs within a range of dopant concentration bounded on one side by
semiconducting or insulating phases and on the other side by metallic phase. Fig. 3
illustrates some typical phase diagrams of doped oxide superconductors.
//
/"
/
/
/
/
lnsulatin~ ~ Mcl;dllc Mclalti~ /
tV
13, /
/
( u oxidation stale . . . . . . BaPbO3., X -- ~'~ t3a't~iO
(c)
Metallic
Semiconducling
X ....
Fig. 3. Schematic phase diagrams of oxide superconductors
(4) Phase transition in solid state like
YBa2Cu306(Tet. ) + 1/402 = YBa2Cu306.5(orth-II ) + I / 4 0 2 = YBa2Cu307(orth-I ) (1)
The tetragonal phase is semiconducting and the Orth-I and orth-II phases are
6 Chen Changkang
superconducting with Tc~90 K and 60 K respectively.
(5) The ambient atmosphere can bring about obvious effects on the phase diagram and on
crystal growth and superconducting properties as well.
(6) Incongruent melting makes single crystal growth much more difficult than preparation
by a ceramic sinter. A molten salt or flux has to be used to lower the growth temperature
zone below its decomposition temperature. Inhomogeneity of dopant concentration always
occurs in the crystal.
The history of solid state physics and chemistry shows that a detailed understanding and
improving of the functional materials usually depends upon the availability of single
crystals. The crystal growth of high Tc oxide superconductors is essential not only for
fundamental investigations, but for potential applications as well in order to:
(1) Explore the relation between the composition, the structure and the superconducting
mechanism and properties.
(2) Determine the anisotropy of physical properties.
(3) Understand the role of impurity, defect and grain or twin boundaries on
superconductivity for manufacture or treatment of ceramic wire, tape or bulk materials.
(4) Develop single crystal, crystal film and textured materials for electrical devices.
(5) Clarify the growth mechanism of crystals with layer structure.
Because of the incongruent melting, crystal growth has to be carried out at lower
temperature than the decomposition temperature and Czochralski method from a
stoichiometric melt is not possible. The methods for crystal growth include:
(1) Solid-state growth: almost immediately after the discovery of the superconducting
cuprate, only tiny single crystals by solid-state growth were available for structure
identification by X-ray diffraction. They were too small for other investigation. Then small
amounts of alkali halide [ 12] or alkali carbonate [ 13] were used to encourage bulk growth.
But these crystals show large misorientation [14], cracking and inclusions.
(2) Vapour growth: Electron beam and thermal evaporation or molecular beam epitaxy
(MBE) of YBCO on some substrates produces epitaxial thin films for electronic devices.
But no large crystal has been grown by this method up to date. Single crystals of
Hg-Ba-Ca-Cu-O with suhmillimeter size were grown either by vapor-solid reaction or by
solid - solid reaction in a sealed quartz tube, where high vapour pressure created by
mercury decomposed from HgO at 500 C [15,16]
(3) Hydrothermal growth was successfully used for the crystal growth of superconducting
BPBO from aqueous solvents [ 17]. But due to its high reactivity with water:
2YBa2Cu307 + 3 H 2 0 ===3Ba(OH) 2 + 5CuO + Y2BaCuO5 + 0 . 5 0 2 (2)
no YBCO phase was obtained in the hydrothermal experiments [18].

(4) Flux growth is the major method for oxide superconductors. Tal)le I summarizes
fluxes for crystal growth.
High Tc Oxide Superconductors
Table 1. Flux growth of oxide superconductors
Crystal Flux Note Ref.
Rare-earth cuorate:
La2CuO 4, :Sr, :Ba CuO S-F 19-21
L i 4 B 2 0 5,Na4B20 5 O-F 22
Na2CuO 4, :Ce CuO S-F 23
YBCO, RBCO CuO S-F 24
CuO-BaO S-F 25, 26
In20 3 O-F 27
Y-247, Y-124 CuO-BaO High pressure S-F 28
CLBCO CuO-BaO S-F 99
PSYCCO PbO-PbF 2 P.S-F 29,30
PbO-NaC1 Immiscible sys. P.S-F 31,32
PbO-CuO S-F 33
PbO-B20 3 P.S-F 34
PSRCCO PbO-NaCI P.S-F 35
None-rare-earth cuorate:
Bi-Sr-Ca-Cu-O KF-KCI O-F 36,37
NaCI-KCI O-F 38
PbO-CuO P.S-F 39
CuO-CaO S-F 40
TI-Ca-Ba-Cu-O CuO-CaO S-F 41
Non-copper oxide:
LiTi20 4 NaBO2-LiBO2-NaF E-C + P.S-F 42
Ti Sealed S-F 43
BaPbO3, :Bi, :Sb PbO S-F 44
PbO 2 S-F 45
PbO-PbO2-Bi20 3 S-F 46
KCI O-F 47
(Ba,K)BiO 3 KOH S-F 48
E-C + S-F 49
NaxWO 3 WO 3 E-C + S-F 50
S-F: self-flux, P.S-F: Partially self-flux

O-F: Other flux E-C: Electrocrystallization
The main flux as shown in Table 1 for the crystal growth is the self-flux system, because
8 Chen Changkang
oxide superconductors with layer structure normally have low stability and high ability to
react with other compounds resulting in other crystalline phases. By using a flux of the
component species, a "self-flux", we can avoid additional contamination from the fitLY,
which also provides a good opportunity to use high purity starting materials for crystal
growth. The superconducting properties are very sensitive to the impurity, most of them
decrease or even kill the superconductivity. In some case the other flux, like alkali halide
or borade, may be used. The ionic radius of alkali or boron are largely different from the
component elements of oxide superconductors. In the case of non-cuprate oxide or organic
superconductors, the electric field might facilitate crystal growth as a driving force for
crystallization and variation in the charge of the element species as well.
I1. THE PttASE RULE, PHASE DIAGRAM AND CRYSTAL G R O W T H
Phase diagram characterises the equilibrium state which is non<ime dependent. This state
is difficult to reach during practical experiments. Stnctty speaking, most phase diagrams
are inferred from conditions of non-equilibrium although close to the equilibrium state. In
many cases we prefer the growth procedure close to equilibrium in order to produce high
quality single crystals. Theories of crystal growth was developed based on thermodynamic
equilibrium or stable state. A phase diagram or even some initial knowledge of phase
relation are essential for selecting the crystal growth method and for establishing
optimum of growth conditions, such as the starting composition and the temperature
program.
Consider a system with multiple phases in a state of equilibrium state, where the number
of phases is P and each phase consists of C components, x i a represents the fractional
content of the i molecule in a phase, so that
c
xia = I
i=l
The total number of the parameters, which describes the whole system, is P(C-I). Under
the condition of phase equilibrium, the chemical potential Nia in any phase is equal:
Nla= Nlb= NlC . . . . . . . . NlP "1

N2a= N2b= N2C= - = N2P c(P ~) (3)
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Nca= Ncb= NcC . . . . . . . . . . NcP
Therefl)re the number of indcpcndenl parameters is P(C-1)-C(P-1) = C-P, With the

number of the two external conditions (temperature and pressure), the degree of freedom
F = C-P+2 (4)
This is the phase rule governing the phase relation in the system. Due to the character of
oxide superconductors mentioned before, the external conditions have great effect on the
superconducting phases. The main external conditions are:
(1) Temperature: The temperature always is the dominant parameter in the phase
diagrams, especially for the growth system. Flux growth is normally carried out by slow
cooling. For multiple phases in the same system, certain phase exists only in a narrow
temperature range. In the period after crystal growth, an additional temperature
programme or anneal procedure may be carried out in order to encourage or avoid a
specific structure transition. A twin pattern may result from the structure transition. A
specific technique of de-twinning may then be applied if single domain is required.
(2) Pressure: The stability of oxide superconductors can be varied with pressure. At
atmospheric pressure and above oxygen will stabilize the Y-123 phase and can increase
of the decomposition temperature:
2YBa2Cu307_y ===Y2BaCuO5 + 3BaCuO 2 + 2Cud + (0.5-x)O 2 (5)
However in the high oxygen pressure Y-123 becomes unstable and decomposes into
Y-247 or Y- 124 phases:
3YBa2Cu307_y ===Y2BaCuO5 +YBa2Cu408+3BaCuO2+CuO+(O.5-1.5x)O 2 (6)
4YBa2Cu307-y===Y2BaCuO5+Y2Ba4Cu7014+x+3BaCuO2 +Cud+( 1-2y-x/2)O 2 (7)
(3) Atmosphere: Due to non-stoichiometry of oxygen and elements with various valences
in the oxide superconductors, changes in the atmosphere surely will affect the phase
diagram of the growth system by changing the peritectic or eutectic points and the slope
of the liquidus line. Thus only the oxidized state of LSCO and YBCO (P-type
superconductors) is superconducting. On the other hand, the reduced state of NCCO
(N-type superconductors) and PSYCCO (P-type) is superconducting.
(4) Electric field: This can be provided by an electric current or voltage between the
cathode and the anode in the crucible. The main functions are to generate an
electrochemical reaction or to bring ahout electrocrystallization, which is currently not
well understood for the crystal growth of superconductors. The new' eqoilibrium state
created by the electric field may change the phase relationships. Table 2 shows the
NaxWO 3 with various structures was produced by electrolysis of the flux system of
Na2WO4_WO 3 [50], which is completely different from Fig. 4, the phase diagram
without the electric field [51]. Under electric field non-stoichiometric crystals with various
amounts of Na were grown, instead of stoichiometric compounds of Na2W()4, N a 2 W 2 0 7
or Na2W4OI 3 in Fig. 4.
10 Chen Changkang
~4(:-:,c, " "°

. . . . . . • T T ] .... r0
20C -
;" r
1:.,, ~. ,, 72C ~ bl
t,~ L,
Fig. 4. The system of Na2WO 4 - WO 3
Table 2. Phases of NaxWO 4 grown from the flux system of N a 2 W O 4 - W O 3 by

electrolysis as functions of melt composition and temperature [50]
W O 3 in melt (mol.%) <5(:) 50-62 62-70 >70
Temp. range ( C ) 700-900 740-790 790-890 >890

Phase Cubic Tetra.-I Tetra.-II Orth.
Composition x 0.93-0.49 0.49-0.25 0.25-0.07 < 0.07
The oxide superconductors and their relatives contain two or more oxide components,
LiTi204, La2CuO 4 or (La,Sr)2CuO 4, Nd2CuO 4 or (Nd,Ce)2CuO 4, (Sr,Nd)BaO 2,
BaPbO3, BaBiO3, KBiO2. 5 or Ba(Pb,Bi)O 3, (Ba,K)BiO3, which can be regarded as binary
oxide systems. Y-Ba-Cu-O, Bi2Sr2CuO 6 and TI2Ba2CuO 6 are ternary systems. And
Ca-La-Ba-Cu-O, Pb-Sr-(Y,Ca)-Cu-O, Bi-Sr-Ca-Cu-O, T1-Ba-Ca-Cu-O and Hg-Ba-Ca-Cu-O
can be considered as quaternary oxide systems. Here the dopant with the oxide which it
partly replaces is regarded as one component. According to the phase rule a multiple
component system (C = 2 to 4) with various external conditions is difficult to represent in
the form of a phase diagram without simplification. Pseudo-diagrams are widely used fl~r
oxide superconducting materials. In these cases the phase rule is not strictly followed.
HighTcOxide Superconductors 11
I. Single system -- P-T diagram:
In a single system with external conditions of tenl~:"rature and pressure, the degree of
freedom F = 3-P. The maximum degree is F = 2. The phase diagram is two dimensional
with Pressure and temperature as co-ordinates. The growth system~ of oxide
superconductors are not formally single component systems, but the P-T diagram can lye
used to express the phase relationship in multiple system as the external conditions (P, T)
are changed. Fig. 5 shows a pseudo-phase diagram for the flux growth system of
YBa3Cu7.50 x [52] At low pressure of oxygen, there is a temperature range for the
crystal growth of Y-123. As the pressure increases, the range becomes smaller and
smaller. At about 100 bars, the stable range for Y-247 appears. But this range keeps very
narrow as the pressure increases which indicates the difficulty in obtaining single Y-247
phase material without intergrowth with Y-123 or Y-124. And after the crystal growth,
fast cooling is required in order to avoid the phase transition from Y-247 into Y-124 as
the temperature passes the stable range of Y-124. At a pressure of 400-600 bars, both
phases Y-247 and Y-124 can coexist. The single phase of Y-124 exists at higher
pressure and the single crystal of Y-124 has been grown by a flux method under high
pressure [53].
O O o
LO 0 LO
o T/oc/
c .S 4 ,;i
~'2 ?2'4 c. i 2 3 ;
•/.o
~, 0 2 3 1
1000 + _21 i !
......... i
Po2
/bar/
: L+ S
L
100
L Jt I
1111' !
10
,v ," 7 ? .:. ,'-; 7 4 7.5 7.6 7 7
IO000/T{K}
Fig. 5. P-T diagram for the flux growth system of YBa3Cu3.50 x [52]
12 Chen Changkang
2. Binary system -- X-T diagram:
This kind of phase diagram is mos: commonly used for crystal growth, because the flux
growth is normally performed by slow cooling under constant atmospheric pressure. In
this case, the degree of freedom still is F = 3-P and the diagram is expressed by the
composition and the temperature as co-ordinates. Binary systems of La203-CuO,
Nd203-CuO and BaO-PbO are related to the crystal growth of La2CuO 4, Nd2CuO 4 and
BaPbO 3 respectively, which will be described later.
Many oxide superconductors or flux have two or more oxide components. A simplification
is occasionally used for the growth system. Fig. 6 shows a typical pseudo-binary diagram
of superconductors (S C) - flux (F). Due to the similarities of the composition and the
structure of cuprate superconductors, their phase relations are typically a combination of
eutectic and peritectic systems. For crystal growth, the starting composition t~) could be
selected along de, the liquidus line, in the eutectic system. We also can allocate the
starting composition (~) in the peritectic range, a-d, where the first crystalline phase is D
until the peritectic reaction occurs at the peritectic temperature (Tp):
D + L ......... > S C (8)
(2)
I
D+L
Tp
sc+L d~
n+SC
n SC
Cd Ce Flux
Fig. 6. Typical phase diagram for the growth systems of oxide superconductors
Here the composition of the crystalline phase "jumps" from D into (S C). As the reaction
takes place at the interface between the solid (D) and the liquid (L), resulting in a coating
of D by (S C), whirl" r)revents further reaction due to the difficulty of diffusion in the solid
layer. As the temperature continually decreases, crystal growth enters into the eutectic
area and terminates above the eutectic temperature (Te). However, the real situation is
more complicated because (a) the real system always contains multiple phases with more
components, (b) the equilibration state is difficult to reach, (c) the corrosion of crucible
materials can not be neglected and (d) the influence of the local environment should be
considered.These factors are probably the reason for the considerable differences found in
the experiment results from various laboratories.
3. Ternary system:
Under constant pressure, the degrees of freedom are given by F = 4-P. Fma x = 3. The
phase diagram is shown in three-dimensional space as a triangular prism with the
concentration triangle as the base and the vertical axis as the temperature. Fig. 7 shows a
T-X triangular prism for the ternary system of Y2BaCuO5(211) - Ba2Cu305(023 ) -
CuO(O01) at Po2 = 1 bar, [52] which indicates stable ranges of Y-123, 247 and 124 in
the vertical section A( 1 2 11/4) - (001) line.
T = 1000°C
900°C
800°C
211
700"C
)1
023
Fig. 7. Triangle prism for ternary Y2BaCuO5(211) - Ba2Cu305(023) - CuO(O01)

system at Po 2 = I I)ar [52]
14 Chen Changkang
The three phases are located at:
(1 2 3 ) = 0.8( 1 2 1 1 / 4 ) - 0.2(0 0 1) (9)
0.5(24 7)=0.571(1 2 11/4)-0.429(00 1) (1o)
(1 2 4) = 0.444( 1 2 11/4) - 0.556(0 0 1) (11)
Under Po2 = 1 bar, Y-123 decomposes at 1000 °C by the reaction (5). The instability
may occur on the low temperature side (< 700 °C) with an extremely slow decomposition
rate [57]. Y-247 exists in the range between 910 °C and 800 °C with reactions:
Y2Ba4Cu7014+y .... > 2YBa2Cu307_ x + CuO + 0.5(2x+y-1)O 2 (12)
Y2Ba4Cu7014+y ----> YBa2Cu307_ x + YBa2Cu408 + 0.5(x+y-1)O 2 (13)
Y-124 decomposes into Y-247 + CuO at 890 °C and is stable down to room
temperature. Fig. 8(1) shows the ternary system of YO 1.5 - BaCuO2 - CuO surrounding
the Y-123 phase. The temperature vs composition diagrams along (211)-(123),
( 123)-A(3 BaCuO2+CuO ) and ( 123)-CuO lines are shown in Fig. 8 (2-4). No coexis ten ce of
the (123) phase + liquid occurs in the phase diagram of (211) - (123). But phase
diagrams of (123)-A and (123) -CuO did show such coexist ranges. The phase diagram of
YBCO -BaCuO 2 [60] revealed a liquidus line of (123) + L at 1030 - 910 °C. Therefore it
is clear that the starting compositions for the crystal growth of Y-123 by self-flux are
within triangle area of YBCO - BaCuO 2 - CuO and close to the line of BaCuO 2 - CuO.
6oCuO z
&;2
yO U 2," 4-[/ h 2 [,~,p8 5 ~0 LuO
F,? .8(1)
211--A Sectxon
1400
000
(211 ).( I 2~ )Te " (125)Te*(0 ~I*

0~2)
6OO
(12~} ~.(Ll'i 1 *"
20(}
2~
(2)
"-- i
1]OCI
f
i Y2BoCuOs + L
/
I
I
1100
X
~
YzBoCuO~
\\
+
~ unWnown
*
x L
, x
' i
1020 ° : 10
i
~: i',YBa~Cu307. ,8~CuO~ ~,~ gcCu02
:CCC ~FIn( ,0, , I
'~ 'O00i;,, , .CuC.I
: :.:;~ I t0 ..... '* I '~ s:o..,lo 1

I
r' I
',F:, ,7,: , • ~sCuC, + [ I
i : ,, Y~%Cus07 T*CuO*L I
:, j
c~b~ L
,~r,o/ _ . .Y:~_LP__~_ . . . . . . . . . . . .
! . . . . C, , ', r', - G,O
] vEc ;¢u~3~ ~* BOCUC'~* CuO
!
IF] :" 4 C} IS:: ~C ( :0 40 ~)~ BO 100
~,~, ~,~ A (~,,: ) I, L ,~1,( n ' ~ tool % Cu()
(3) (4)
Fig. g (l).Ternary system of Y O I . 5 - BaCuO 2 - CuO at 9 5 0 - 1 0 0 0 °C.

(2-4). The pseudo-binary phase diagrams along (2) (211) - (I 23) [58]; (3) (123) -
A ( B a 3 C u 5 0 8 ) and (4) (123) - CuO lines [59] in fig. 8 (1).
16 Chen Changkang
4. Quaternary system:
Under conditions of constant temperature and pressure, the degrees of freedom are given
by F=4-P. Fma x = 3. The phase diagram shows in a tristatrahedron with four components
in four top points respectively. The composition of any point in the tristatrahedron is
expressed by the distances between the point and four triangle planes. Fig. 9 [10] (a)
represents the Bi-Sr-Ca-Cu-O quaternary system. The superconducting family of
Bi2Sr2Ca(n_l)CunO2(n+l) is located in Bi/Sr=l plane. A bold line in the plane connects
two compositions between Bi2Sr2CuO6(2201) and CaCuO2(0.5Ca,0.5Cu ). Fig. 9(b)
shows the preliminary pseudo-binary phase diagram along this line, which indicates the
suitable liquidus ranges for the crystal growth of 2201 and 2212 phases. But no liquidus
range with a 2223 phase was found. Even with the coexistence of 2212 and 2223, the
whole system would only partially melt. So, it is very difficult to grow a single crystal ()f
the 2223 phase without intergrowth. Compared to the Bi2Sr2CuO 6 - CaCuO 2 system, the
decomposition temperatures of *I'1-22()1, 2212, 2223 and 2234 phases in Tt2Ba2CuO 6 -
CaCuO 2 system [41] (shown in Fig. 10) are also arranged from high to low temperature.
But there is a narrow liquidus range for the crystal growth of the 2223 phase by the flux
method. Crystals of TI-2201 and 2212 or 2212 and 2223 have been grown successfully
by the flux method [135].
I c No melting
• a-" Partial melting
900
ca • Important melting
~2201 • m
~x.72212
.2,2,2. [ .. 1 + CaCu + CaPb + k
2223 -~ 8~q[) ~ 1+1~" \"IN " ¢- ~'- i,-
/
/ \
\ ~4(} l J .I J d) 4 (-~( u-(~AP~_ ~*Ca('u

(P ",)-;~
] 2 3 4 5 6 7 10 15
n value
Bi2Sr 2('uO~ -" (b) - (;aCu(),
(a)
Fig. 9(a). Bi 1.5-SrO-CaO-CuO quaternary system

(b). Binary system of Bi2SrCuOx-CaCuO 2 along the bold line in (a) [40]
(0 \
..... x 941
940
"%
x
X
\
" A
c) \\ I
SI+I. \i
o 920 \\ B
\ {
(2021) \ i
k
O
\ 907
...... -- ,, C D
E \ ]
O i
b- 9(~) {~ \ : i
. 1"3"9"
~_ ) \ 896 1
I ½ (2>2a).
S: ~ S~ S; + L (2223~
88,{}
__ S1 _a_~ 1 ....~. . . .S3+Ss
... ~5°(
: i ........
2() 41} S 2 ~ 60 80 (al. q )
CaO + CuO
Fig. 10. Pseudo-binary system of TI2Ba2CuO6-(CaO+CuO )

SI: 2201, $2: 2212, $3:2223 and $ 4 : 2 2 3 4 [41]
HI. EXPERIMENTAL DETERMINATION OF PHASE DIAGRAM
The experimental determination of phase diagrams generally includes two stages: the
preparation of samples and the measurement of phase relation parameters under
conditions as close as possible to the equilibrium state. The pseudo-binary diagram of a
superconductor - flux system is very useful for considering crystal growth. This phase
diagram shows the phase, especially the desirable superconducting phase, relation in the
temperature - composition coordinates, which is to be discussed in detail helow.
1. Sample Preparation
Normally the superconducting compeund and the flux are synthesized separately. Then
the mixtures of them in various proportions are treated in order to form equilibrium
phases, which are the samples for measurement. The superconductors can be prepared
either by solid, liquid or sol-gel methods. The liquid method involves a low temperature
solution or a high temperature melt. The precipitate from a low temperature solution, in
18 Chen Changkang
which the stoichiometric components are dissolved, has high homogeneity at molecular
level. Then the precipitate is separated out and sintered at high temperature to obtain the
superconducting phase through a solid state reaction. The high temperature melt method
involves a stoichiometric system, which is completely melted at very high temperature.
Due to the incongruent melting of oxide superconductors, quenching the melt to room
temperature is necessary to reduce the formation of decomposition phase and the
segregation of solute. Repeat grinding, mixing, pressing and heat treatment enhance the
homogeneity. However due to the high reactivity of cuprate superconductors in the liquid
state, no crucible has been found as an ideal container for melt synthesis. Contamination
from crucible materials inevitably occurs. The technique of melting without crucible is
recommended. The sol-gel method has characteristics between solid and liquid methods.
After drying further reaction may take place by heat treatment in the solid state. This
method is widely used for the preparation of bulk or thin film superconducting materials.
Solid state reaction is the traditional method for preparing samples for phase diagram
determination. Reaction occurs through diffusion in particles at solid state, and long
periods for reaction at various temperatures are required, involving several cycles of
grind-mix-press (into pellet) - sinter. In some circumstance, a precursor is used for the
synthesis. For the high Tc superconductors of Hg-Ba-Ca-Cu-O, the precursor of
Ba-Ca-Cu-O was used to react with HgO [61 ]. HgO decomposes at a tow temperature of
500 °C in ambient atmosphere. But the starting materials of BaCO 3 and CaCO 3 for the
precursor lose CO 2 at much higher temperatures by solid state reaction. Table 3 shows
the thermogravimetric(TG) results for the mixed powder of
2.28BaCO2+2CaCO3+3.72CuO. BaCO 3 and CaCO 3 have completely decomposed at a
temperature of 940 °C and 1230 °C respectively under experimental condition. Table 4
shows the TG result for the pressed pellet of the same composition using a step heating
program. Compared to Table 3, the decomposition temperature of BaCO 3 and CaCO 3 are
much lower, which indicates that the diffusion rate in the solid is speeded up by the
pressing. And the reaction rate also depends on the amount of sample, the degree of
mixing, the temperature and the time. However Hg-Ba-Ca-Cu-O superconductors can be
prepared in a single step without precursor [62,63]. Mixtures of HgO, BaO, CaO and CuO
were ground and pressed into a pellet, then they were sealed in an evacuated container.
As the temperature increased, the decomposition of HgO giving mercury vapour created a
high pressure which assisted the solid-state reaction and Hg-1223 was obtained.
Table 3. Thermogravimetr~c analysis of the starting mixture of 2.28 BaCO 3 + 2 CaCO 3 +

3.74 CuO at heating rate of 30°Ch -1 in air
TG step I II III IV V
T (°C) 740 940 960 1160 1230

CO 2 loss from BaCO 3 CaCO 3
Calculated(wt%) 10.6 9.3
measured (wt%) 10.6 9.4
Table 4. Thermogravimetric analysis of the pellet of 2.28 BaCO 3 + 2 CaCO 3 + 3.74 CuO
at step heating in air
Temp. step 1 2 3 4 5
T ( °C ) 660 730 780 830 880

t (hs) 6 6 6 6 60
CO 2 loss from BaCO 3 CaCO 3
Calculated(wt%) 10.6 9.3
measured (wt%) 10.5 9.2
BaCO 3 and CuO are two of the most common starting materials for the preparation of
oxide superconductors. Care has to be taken for the complete decomposition of BaCO 3 by
a solid state reaction. Sometimes precursors are used to avoid CO3-2 contamination in
the structure. The existence of a liquid phase greatly speeds up the decomposition. Table
5 shows the results of crystal growth from starting materials with the same composition
(1) with both precursors YBCO and 7BaO+18CuO; (2) without any precursor and (3)
with the precursor YBCO only. All three experiments give similar results using a proper
temperature program. The equilibration state is more easily approached in high
temperature solution. In the Ba-Pb-Bi-O system, PbO and Bi203 are excellent molten
solvents with low nielting points. The decomposition of BaCO 3 takes place at low
temperature. Our TGA experiment on the starting materials BaCO3-PbO-PbO2-Bi203
used for crystal growth shows that at 290 °C PbO 2 starts to decompose with a
corresponding weight loss:
PbO 2 . . . . > PbO + 0.5 0 2 (14)

20 Chen Changkang
BaCO 3 loses weight from 550 °C, followed by the formation of BaPbO 3 at 560 °C, which
introduces two further steps in the TG curve. The overall reaction is written as
BaCO 3 + PbO + 0 . 5 0 2 .... > BaPbO 3 + CO 2 (15)
which is completed at around 700 °C with a loss of CO 2 comparable to the calculated

value, indicating the complete loss of CO 2.
Table 5. Crystal growth of YBCO with and without precursors
No. Starting materials YBCO crystals

Solute Flux BaF 2 Max.size Yield
(13/6) (0.87/25) (wt%) (ram 3) (g)
1 YBCO 7BaO+ 18CuO 1 6x3×0.4 0.72

2 0.5Y203+2BaCO3+3CuO 7BaCO3+lSCuO 1 7x4×0.3 1.0
3 YBCO 7BaCO3+ 18CuO 1 6x4xO.3 1.1
1100 108(I 1(h'fft 1U--kt) 11)20
43.2
e~
~ .XI4
\ '\,
'\x
4) t~i \
"\
\\,
q gJ . . . .
1,"I d(!(l:K I)
Fig. 1 1. Decomposition temperature of CuO vs oxygen pressure

Care should also be taken to avoid the decomposition of CuO:
2 CuO .... > C u 2 0 + 0.5 0 2 (16)
which produces another component Cu20 to the system [64]. The composition of CuO -
C u 2 0 [65] is strongly dependent on the atmosphere, the oxygen pressure, the holding
temperature and time. Fig. 11 shows the linear dependence of log PO2 on l/Td. The
decomposition temperature Td = 1096 °C for lbar oxygen pressure is 70 °C high than for
Td = 1(/26 °C in air. The sintering temperature should be kept below the decomposition
temperature.
2. Thermal Analysis and its Improvements
The construction of T-X diagrams normally involves two stages. One is the determination
of phase boundaries and the other is identification of phases. Identification is mostly
carried out on quenched samples in various phase zones at different temperatures by
physico-chemical analysis, such as X-ray diffraction, microscopy or spectroscopy. The
composition of oxide superconductors can include multiple phases and with slow
equilibration, low stability and high reactivity. Many possibilities for phase identification
may need to be checked in connection with crystal growth. The determination of phase
boundaries also can be done by the identification of quenched samples at various
temperatures over a small interval [66]. However, the most extensive method is thermal
analysis.
Thermal analysis is a physico-chemistry method for measuring properties of a system as a
function of temperature during heating or cooling. The properties cover optical, electrical.
thermal, mechanical, acoustic and magnetic properties and weight, composition and the
dimensions of the substance. Among them, differential thermal analysis (DTA) [67,68]
and thermogravimetric analysis (TGA) [69] are the most common and convenient in use
for superconductor systems.
(1) DTA and its improvement:
In DTA, the temperature of a sample and the temperature difference between the sample
and the reference (an inert substance in the experimental temperature range, such as
NaC1, Ai20 3 or even an empty crucible) are recorded simultaneously. In the typical phase
diagram of growth system for oxide superconductors (Fig. 6), thermal effects at the
eutectic or peritectic lines are rather large allowing detection by DTA. But the effect along
the liquidus line is occasionally too small to be detected with use of a small amount of
sample (10 - 101 mg). Direct observation and analysis of quenched samples are helpful,
22 Chen Changkang
but it is time consuming. Improved DTA may provide an efficient way to determine the
dissolution temperature.
Improved DTA (IDTA) is an in situ analytical method, in which a DTA detector head is
placed directly into the growth system. Fig. 12 shows the schematic assembly of IDTA,
where one DTA thermocouple junction is inserted into the growth system and the other
junction sits outside the crucible. Sample temperature is measured by another
thermocouple attached on the crucible bottom. The temperature gradient between two
junctions of the DTA thermocouples is compensated by a minivoltmeter and then the
signal is inputted into an amplifier. As a large amount of material is used for crystal
growth, a large signal may result under slow heating or cooling rates. Fig. 13 shows the
IDTA endothermal peaks of starting materials of Tm203-BaCO3-CuO for the crystal
growth of TmBCO (luring heating at 1.7 °C/min. in air. The small peak at 801 °C
corresponds t{} the structure transition of BaCO 3 [7{)]. Above 944 °C the materials start
to react with partial dissolution. The shield tube provides some sites for nucleation and
crystal growth. Crystallisation can be detected in a stable system.
MV
Pt or AI203 tube . . . . .
DTA Ihermocouple . . . . . .
Crucible I~Q i
Flux ..... -II,~. DTA

s."
.__!
Pedestal . . . . . . . ~
Thermocoup]e f~
'i £e(.orde'l
Alp O~ tube . . . . . .
i
I
Fig. 12.Improved DTA assembly

:?
\\ ',,\
ll) I/',
Fig. 13. Improved DTA for TmBCO growth system
(2) Thermogravimetric analysis (TGA) and its improvement:
Normally DTA is used for qualitative analysis. Differential scanning calorimetry (DSC) is
suitable for quantitative analysis , but the operation temperature is usually below 800 °C.
Therefore TGA has become the most convenient method for quaJ~titative analysis over
wide temperature range. This technique has been used in connection with the control of
the crystal diameter when pulling a seeded crystal. For research on superconductors, TGA
has been used to understand and control the processes of preparation, sl()ichiometry and
annealing in different atmospheres.
The improved TGA is an in situ method also, which was developed for the determination
of solubility for flux growth [71, 72]. Fig. 14 shows the schematic arrangement of ITGA
[73]. A platinum wire hanging from an electrical balance, is immersed into a high
temperature ~olution to act as a nt?:-leation centre and the apparent weight is recorded.
The immersion depth can bc changed mechanically to study the nucleathm at different
depths of the solution. With a volatile flux nucleathm always appears on the surface,
where the Pt wire can be located. The wire, as a probe, may be located near the bott(;nl of
the crucible if local cooling is used to induce the nucleation [74]. As soon as the wire
contacts the base, the apparenl weight suddenly decreases, so that the depth of the
solution can be measured. After the determination of crystallization temperature in the
24 Chen Changkang
growth system, the crystal growth may be initiated on the wire at a low cooling rate.
Upon the termination of crystal growth, the crystals on the wire are separated from the
flux by lifting it from the solution.
3<, <,,,,,~.,i,,~ ,_kT-- _Ark: ,..,< ._J__
q ;: ~ i P rophyllitc plug l
IL '~l J t
Alurni . . . . . . bc .~7.~.---77~i. I i I V ~ :~-: ,., /
" rtJ ll}{C CICII]C[I "- t I #7 un~l a
] l I ,,!,: jl ~. :L~, }'ta,i ........ p .... ]- J
. . .,., . . . . . . . . . . . ! I ! ' /I t i ~ :
i ~ I "7U ! [ 17, ~ ! "
i ,.'~ u m m a it iw : ! " --~ ! 1 cmpcrapc-e
Electronic "it ............. plc I ~]--~"-4 i Racl" and pint°in

I thermomeler
Fig. 14. The schematic diagram of ITGA
oo (a)
g
1270
I{K] (b)
1260
1250
o
c~
~ o
o 123C'
22O
0 C
Cooling rate { C ° / i n t o )
Fig. 15 (a) ITGA curve for the mixture of YOl.5+2BaCO3+3CuO

(b) Dependence of precipitation temperature on cooling rates
Fig. 15(a) shows the chart record showing crystallisation and dissolution for an ITGA
curve for a n,~'lt of YO 1.5+2BaO+3CuO in a Pt crucible. As the solid phase precipitates
on the piatinum wire during cooling, the apparent weight increases rapidly because of the
surface tension at the solid-liquid interface and the density difference between the solid
and liquid phases. In the reverse situation, during heating the weight decreases because of
dissolution. The faster the cooling rate, the lower the precipitation temperature. Fig. 15(b)
shows the dependence of the precipitation temperature on the cooling rate. The liquidus
temperature, 1265 °C, of the stoichiometric melt Y-123 is given by extrapolation, which
agrees with the phase diagram [76] data fairly well.
3. Determination of Crystallisation Temperature by Decanting Experiments
During cooling crystal growth can be terminated by decanting the crucible at various
temperatures. The solvent is poured out and the crystals are left in the crucible. Since
some crystals might be poured out with the flux. A better means of separation of the flux
can be achieved by using a porous ceramic base in the furnace. Table 6 shows the result
of decanting experiments at various temperature for the system of O.13(I/6)YBCO -
0.87(0.28BaO+0.72CuO) + lwt.%BaF2 in alumina crucible, which indicates that the
crystallisation temperature is in between 935 °C and 930 °C. Using such simple
experiments, the crystallisation temperatures of (Y,Pr)BCO, NBCO and CLBCO have also
been determined in the author's laboratory.
Table 6. Determination of crystallisation temperature of YBCO by decanting experiments
No. Decanting temperature YBCO crystal

(°C) Max.size (mm 3) Picked weight (g)
1 976 none none

2 953 none none
3 935 none none
4 930 3xlx0.2 0.1
5 920 3.lx2.5x0.5 0.55
4. Other Methods:
In order to have more information, combined DTA-TGA are occasionally used for the
superconductor system. TGA, evolved gas analysis (EGA), thermomechanical analysis
(TMA) and X-ray powder diffraction (XRD) were used to characterize the oxygen
26 Chen Changkang
stoichiometry and structural transition [78] of YBCO as well as the calculation of oxygen
content based on the reduction reaction in a hydrogen atmosphere at high temperature:
2YBa2Cu30 x + (2x-7)H 2 .... > 4BaO + Y203 + 6Cu + (2x-7)H20 (17)
Direct observation of melting and solidification in Bi-Pb-Sr-Ca-Cu-O [77] and Y-Ba-Cu-O

[78] systems was viewed using a high temperature optical microscope. The observed
results correlated well with the phase diagram of YBCO system. For the
Bi-Pb-Sr-Ca-Cu-O system, a series of decomposition were observed under the microscope
as:
high viscosity liquid 965°C Ca CuO

2223910°C 2212965°C 2 2 0 1 1 0 0 0 ° C J 2 3 (18)
low viscosity liquid 860°C 2212
No single crystals of 2223 was grown in this system. The in-situ high temperature X-ray
diffraction combined with TGA indicated that the orthorhombic to tetragonal transition of
RBCO was not the first order transition [69].
The solid-liquid equilibrium could be estimated from the results of a soaking experiment
[79, 80]. The solid residue was separated out by arranging for the liquid to soak into a
substrate at a given temperature after reaching equilibrium. Then its phase and
composition were determined by X-ray analysis and chemical analysis. The composition
of the liquid phase can be calculated using the lever rule:
(Xc° - XcS ) N s = (Xcl - Xc° ) N 1 (19)
Where Xc°, XcS and Xcl are the molar fractions of starting composition and the
compositions in the solid and liquid phases respectively. N s and N 1 are mole numbers of
solid and liquid phases. The weight fraction of solid phase
W s = NsMs/[(Ns+N1)M o] (20)
M o, M s are the "average molecular weights" of the starting composition and the solid
phase. Combining (I 9) and (20), we have
Xcl = (Xc°M s - XcSMoWs) /(M S - MoWs) {2])
A small drop of liquid could be separated during the dipping experiment [137]. The
compositions of the liquid were analysed by inductively-coupled plasma atomic emission
spectrometry to dermine the liquidus lines of Y-Ba-Cu-O and Sm-Ba-Cu-O systems.
5. Experimental Conditions
(1) Crucible
The ideal crucible material for the study of the phase diagram and crystal growth should
have no chemical reaction or dissolution with the sample at the working temperature.
Unfortunately, no such crucible exists for the high Tc oxide systems at present. The
reactivity of the molten or partial molten state of the oxides is very high, resulting in
contamination of the system or the formation of new phases. In practice, it is better to
determine the phase diagram using the same crucible as is used for crystal growth. Table
7 lists crucible materials used for the crystal growth of high Tc superconductors. There
are basically five types of crucible material:
Table 7. Crucible materials for the crystal growth of oxide superconductors
No. Crucible materials Crystal
a Element
Au YBCO, PSYCCO, TIBCCO, BPBO
Pt YBCO, PSYCCO, CLBCO, LSCO, NCCO,
BSCCO, TIBCCO, BPBO, BKBO
Mo, Fe LTO
C (graphite) LTO, BKBO
b Oxide
BeO:Mg YBCO
MgO YBCO, BSCCO
ZrO2:Y, Mg YBCO, NCCO
SnO2:Zn, Ti YBCO
A1203 YBCO, CLBCO, NCCO, BSCCO, TIBCCO. LTO
ThO 2 YBCO
BaZrO 3 RBCO (R = Y, Pr)
c. Self-composition
Y203:Mg,Ti,Ba YBCO
RBCO RBCO
CuO YBCO
YBCO YBCO
d. None YBCO
e. Polymer Teflon BKBO
28 Chen Changkang
a, Metal or nonmetal elements: The noble metal Pt has been used for the crystal growth
of YBCO from the self-flux of BaO-CuO. The high concentration of BaO, that was used,
resulted in platinum being transported to the melt surface and forming a layer attached to
the wall. The long period at high temperature caused a platinum reaction with the flux
which produced Pt-compounds of Y2Ba3Cu2PtO10 and Ba4(Pt,Cu)O 6 [81]; Mo, Fe or
graphite crucibles were used with a LiTi204 flux growth in a reduced atmosphere in
order to avoid oxidation. Graphite crucibles can be used at low temperature (~300 °C) for
the BKBO-KOH system in an ambient atmosphere.
b. Oxides with high melting points: Alumina crucibles have been extensively used with
less expense. The dissolution of A1 increases the solubility of YBCO in the flux and
decreases the solution viscosity resulting large single crystal. But the AI contamination of
the crystal depresses the Tc. The reaction products of yellow-greenish YBa3A1207. 5 and
transparent BaAI204 continuously changes the composition of the system. The
contamination of crystals with Zr +4 or Sn +4 through the use of Z r O 2 or SnO 2 crucibles
is rather small. But, the system can solidify at high temperature quite quickly by the
following reactions:
BaO + Z r O 2 .... > BaZrO 3 (22)
BaO + CuO + SnO 2 . . . . > Ba(Sn,Cu)O 3 (23)
BaZrO 3 does not contaminate the crystal. Crystal growth of YBCO and PrBCO in
BaZrO 3 crucible showed that the crucible was unreactive with a self-flux of BaO-CuO
[119].
Fig. 16 shows the pseudo-binary phase diagram of YBCO-BaCuO 2 by differential
scanning calorimetry (DSC) under an oxygen atmosphere in (a) A1203 crucible in which 5
wt% A1 content was found in the melts; (b) Y-stabilized ZrO 2 crucible [60]. The liquidus
line for YBCO as the primary phase using a ZrO2:Y crucible is much narrower than that
found using an A1203 crucible and moves towards the low concentration side. When the
same A1 content (~5wt.%) was added into an ZrO2:Y crucible, both phase relations
became identical, which indicates the dissolved AI203 from the crucible plays a role as a
component of the flux.
Wetting and creeping are the another problems to use the most "of crucibles. Wetting is
caused by a chemical reaction between the crucible and the flux. "Creeping up" is
stimulated by the high surface tension of the melt. The surface tension may be increased
by preferential adsorption of a selected component in the melt [86]. The temperature
gradient enhances the creep of flux along the direction from high to low temperature. A
homogeneous temperature field is preferable for limiting creep or hold up of the solution.
For the same crucible material, a high density and a good quality crucible with a smooth
surface normally reduces the wetting or creep. Various binders used in the sintering
procedure for the preparation of ceramic crucible may introduce different results: a
MgO:Si crucible had little or no wetting, while a MgO:Y crucible showed considerable
creep [87].
-[ - ~ [ - -
t ]. . . . .
/
\ P \l
i0~0 711 , t \ t 1050 711 + L
/ \
• it~ I
• \
:-5 h • °;°-2
~- ] U,5(.
]
\
c]_ 377 + l \f +
~+,+
12] , k . . . . . . . \
E
FSO J *
i
• • tNI i
9O0 90O
Y;EN;'CU30~ t * guEuO ~ -~
BoCuO2. Y;Bo~Lu~O~ ~
y~BQ~Cu30) e 60 70 80 90 Y~Ba2Cu]07.6 60 70 80 90 goCuO

BaCu02
rnot % BaCu02 ,~ot % Bc~CuO:
(a) (b)
Fig. 16. The phase diagrams of YBCO-BaCuO 2 under an 0 2 atmosphere in

(a) an AI203 crucible
(b) an ZrO2:Y crucible
c nnd d" ~ e l f - c n m p n ~ i t i n n ernelhle and nn erneihle.. A d d i t i o n a l contamination can be

avoided by using self-composition or no crucible. A YBCO crucible was fabricated by an
e x p l o s i o n t e c h n i q u e w i t h s u b s e q u e n t sintering. C u O is f o r m e d b y o x i d a t i o n of a copper
crucible at high temperature. Both crucibles have relatively low melting temperatures and
are rarely used. By using a yttria crucible, a single crystal with a size up to 1 cm and high
Tc has been obtained [131]. The dissolution of Y 2 0 3 from the crucible changes the
composition of the system. The crucible binders of Mg, Ti cause a little pollution to the
crystals. High purity Y 2 0 3 crucibles without a binder or with Ba as a binder are
recommended for crystal growth. Significant improvements in the quality of crucibles are
needed for use in experiments carried out over a long period of time.
e Polymer: A teflon crucible with high chemical stability is used for the crystal growth of
BKBO by the electrocrystallization method below 400 °C [83,84].
3O Chen Changkang
(2) Atmosphere
Quite a number of elements with variable valences are used and can show
non-stoichiometry in the preparation of high Tc superconductors. CuO is very often
selected as a main flux composition for crystal growth. The decomposition reaction 2CuO
---> C u 2 0 + 0 . 5 0 2 is strongly dependent on the partial pressure of oxygen. The
atmosphere obviously has a controlling influence on the parameters of phase diagrams.
Table 8 shows the peritectic (Tp) and eutectic (Te) temperatures of Nd2CuO4-CuO,
NCCO-CuO and LSCO-CuO [85] in different atmospheres [54, 55]. Both Tp and Te are
higher when determined in an oxygen than in air environment. General speaking, most of
oxide superconductors are more stable in oxygen than in a reduced atmosphere.
Table 8. Peritectic (Tp) and eutectic (Te) temperatures of Na2CuO4-CuO,

Nd1.85Ce0.15CUO4 -CuO and La 1.85Sr0.15CUO4 -CuO in different atmospheres
System Na2CuO 4-CuO NCCO-CuO LSCO-CuO

Atmosphere 02 Air 0 2 Air N2 02 Air
Tp ( °C ) 1281 1240 1324 1290 1261 1350 1343
Te ( °C ) 1068 1040 1072 1040 1050 1040
(3) Equilibration
A phase diagram is a graphical representation of a thermodynamic state of equilibrium,

which is unique and constant in time. The variety of none equilibrium states are unlimited.
Complex and sluggish reactions lead to very slow equilibration for most of oxide
superconductor systems. It takes a long time to completely lose CO 2 from BaCO 3 using a
solid pellet of BaCO3+CuO flux. Although the reaction is greatly speeded up in high
temperature solution, the holding experiments showed that ten hours at 930 °C was still
needed for the complete dissolution of a 2 g eutectic mixture of BaO+CuO. Slow
equilibration may account for large differences of results between various authors. Table 9
shows the reported eutectic points of the BaO+CuO system in an A120 3 crucible in air.
Both eutectic compositions and temperatures are recorded differently from various
laboratories. In our ITGA experimentg [92], a mixture of more than 50 g BaCO3+CuO
was used. The eutectic composition and temperature were determined to be
0.28BaO+0.72CuO and <880 "C. Three major problems occur in the system of BaO-CuO:
Slow equilibration, flux wetting or creeping and crucible corrosion. With a completely
dissolved system, ITGA can speed up equilibration. The effects of wetting and creeping
are obviously reduced by using large amounts of mixture. The solubility of oxide
superconductors is increased in the eutectic flux and the temperature for crystal growth is
relatively low, which may relieve the crucible corrosion to some degree.
Table 9. The eutectic point of BaO-CuO system in an A I 2 0 3 crucible in air
No. 1 2 3 4 5 6
Eutec. comp.(BaO tool.%) 22.2 29+1.5 37 28 28
Eutec. Temp. ( °C ) 910 910+10 925 900 920 <880
Method DTA Obser. DTA Obser. DSC ITGA
Ref. 89 88 90 58 91 92
IV. APPLICATION OF PHASE DIAGRAM IN THE CRYSTAL GROWTH OF OXIDE

SUPERCONDUCTORS
Since it is one of the fundamental parameters for crystal growth, phase diagrams always
appear in review articles about the crystal growth of high Tc superconductors
[18,81,86,100-102]. See for example the publications of our review papers about
physico-chemical investigations on crystal growth of oxide superconductors [103] and of
superconducting YBCO [ 104]. The following is a summary of the application of phase
diagrams to crystal growth of oxide superconductors, including our research in Clarendon
laboratory, University of Oxford.
1. Phase Diagrams and Crystal Growth in Multiple Phases of Oxide Superconductor

Bi2Sr2Can_ 1CunO2(n+2) System
There are multiple superconducting phases in this system and a series of decomposition
reactions as the temperature increases [40]:
830 °C 860 °C 870 °C 880 °C

2212 - - 2223 - - 2212 - - 2201 - - 2201+"23.9 A" phase (24)
According to the pseudo-binary system of Bi2SrCuOx-CaCuO 2 shown in Fig. 9, Table 10

shows temperature ranges of Bi-2201, 2212 and 2223. These ranges are narrow and
close each other, which occasionally introduces the intergrowth of multiple phases. The
decomposition temperatures of 2201 and 2212 are near to their liquidus temperatures
("an important melting condition"). Crystals of 2201 and 2212 can be obtained from
either stoichiometric or non-stoichiometric compositions or other flux systems, such as
alkali halide [105,106]. The incongruent melting of 2212 phase is a decomposition
reaction:
Bi2Sr2CaCu20 8 .... > Bi2Sr2CuO 6 + CaCuO 2 + L (25)

32 Chen Changkang
2201 is the first crystalline phase and 2212 is stable up to 875 °C. Our experimental
results [ 115] for crystal growth from a self-flux basically agree with the phase diagram.
Table 10. Temperature range of Bi-2201,2212 and 2223 in Bi2Sr2CuOx-CaCuO 2
No. 1 2 3 4
Temp. range ( °C ) 875-890 865-875 850-865 830-850

Phase 2201+L 2201 +2212+L 2212+2223+L 2223+S
Flux growth Yes Yes Yes No
Note Single phase intergrowth intergrowth Solid reaction
l 1 t I i i I 1 I I
i200
Cu20 ss-
~ ss + Liq.
~CuO
i000
CuO + Liq. ~ Liquid
xx
~ C u O ss 820_.20 ° ~J 2 ss + - : q . /~
750 ° '
eoo J--
CuO ss + I: 2 o
¢n
400 °1 / Bi20 ~
CPl t'
I L [ I ~, I I I J
20 40 60 80
CuO Blur33
i~401%
Fig. 17. System of Bi203-CuO [107]

i I
tiquld
o by ITGA 200°Ch-1 zI
-°k
900
• by DTA 600°Ct~1 'x
x11770 e
II
T (°CI CuO+L °N~ 02X)

8Bb03+ ~ c ~
850
----o--o --OL ~
700 5:¢JZ, + CuC
I
800 CuO*CuBI2Q ! a{570., + ~'2':a:',
i
i i
I
I
750
C~,O ~'o ~'o 7'o '
60 CuBi2Ot,
Cu0 ( tool °/,)
(a) (b)
Fig. 18 (a) The phase diagram of CuO-CuBi204 [ 108]

(b) The phase diagram of C u B i 2 0 4 - B i 2 0 3 [109]
Because of their low melting points, many authors employ Bi203 and CuO as a self-flux
to grow Bi-Sr-Ca-Cu-O single crystals. Fig. 17 shows an early report of the Bi203-CuO
system [ 107], where only the existence of one compound of CuBi407 was reported. Based
on this diagram, we firstly chose the starting composition of 0.57Bi203+0.43CuO to
grow CuO crystals. As a result no CuO crystal was obtained. Instead we got CuBi204
crystals, which were not shown in the phase diagram. By ITGA and DTA, we established
the phase diagram of CuO - CuBi204 [108] shown in Fig. 18 (a) together with another
published diagram of CuBi204 - Bi203 in Fig. 18(b) [109].The CuO - CuBi204 system
has a eutectic temperature of 845 °C and the eutectic composition of CuO : Bi203 = 0.53
: 0.47 (tool.). A wide liquidus range exists for the crystal growth of CuO and the solubility
of CuO follows the Arrhenius dependence:
S = Soexp(-H/RT) (26)
Where H=52.1 Kcal/mol is the heat of crystallization of CuO in the system.
Fig. 19 and Fig. 20 show binary phase diagrams of Bi203 - SrO and B i 2 0 3 - CaO
respectively. As the compositions move toward the rich-Bi203 direction, the dissolution
temperatures decrease. An excess of Bi203 plays a role of a flux for the crystal growth of
34 Chen Changkang
Bi-Sr-Ca-Cu-O. The addition of Bi203 causes a marked decrease of liquidus temperature.

With respect to solvent properties, Bi203 exhibits some similarity to that of PbO [112].
FiE 21 shows the partial phase diagram of B i 2 0 3 PbO [95]. Bi and Pb are
neighbouring elements in the Periodic Table. They can form a solid solution in a certain
wide range. Many experiments have indicated that Pb could partially substitute the Bi site
in the structure of Bi-Sr-Ca-Cu-O superconductors, and stabilize the Bi-2223 phase. Fig.
22 shows the system of PbO - CaO [113]. At low temperature (~680 °C), the following
reaction may occur:
PbO + 2CaO .... > Ca2PbO 3 (27)
Referring to the phase diagram of CuO -CaO [114] shown in Fig. 23, the eutectlc
temperature in PbO - Ca2PbO 3 is 200 °C lower than in the CuO - Ca2CuO 3 system. The
addition PbO of decreases the melting temperature of Bi-Sr-Ca-Cu-O system which
benefits the formation of the Bi-2223 phase by the diffusion of Ca +2 and Cu +2.
i l i I E ] i I f
,ooo; + , , , , ,'- , i ! 950

\'COO + Liqz'
3 7 + CoO
3'SS
900 // ~
r,'/I I +++0°
SrO
/
j_ _ _ _~v_47ss I
+
3:7
850 /
(i's$
/ Y # a'.+ 5
Ce7BieOle
+
CoO
' ~ss
750 --
;--,~"~ 1 I
700 ~ o
\ V p2ss I -,I I
I %1 ~.-'1 "a c5
aBip~ I rl ~1 '
650 p~ss , ,q, g, ~ S
_ I ul ~ _ 1 2 . i . , ......J_ /1 ir ~L__±__.d _ _ __[ I

" 2C 4C 6:? 5> 1I}O 2(' 49 [0 88 i%
KIo
ai.~os S,, Bi20~ CoO
Fig. 19. The system of Bi203-SrO [ 110] Fig. 20. The system of Bi203-CaO [ 111 ]
2 C ss+ Lic
~. CC ~!cc f
730* I
]
i
[ _ qu c
+css 11 , OOC
CoO * t ,q I
I i
600
90C~-
866~
PC2 • L,'4
t I
•. 822 ~ PbO 2 C 7 2 i
&12C~ ( 575~, ; 61b" I -J
I ___J. . . . . . . . .
5: 6 : ~ 24 c',aD •c-O ~)
Me1%
Bi2O~ f'30i % PbO
Fig. 21. The system Bi203-PbO [95] Fig. 22. The system PbO-CaO [112]
I l I I I ' I\\ /~,,&.u£ +

2:1 + liquid 1:2 .t Liquid
CoO + Liquid 1o30' ~ ~020~
".~ 'v,-,o, a.
1026 a
1000 Co2CuO 3 + 1:2 ~_. 1:2 + CuO /__-~

Liqui~ + CuO
2:1 4 CuO
800
@
t
CoO + 2:1
600 --. 2:1 4 1:1 tl 1:1 + CoO
400 - - - - - I ....J. . . . . k _ ~ . . . . . . . . . . L. . . . 1 I __ I . . . . . .
0 2O 4 g. 60 80 lOO
C,JO Mol % CuO
Fig. 23. The system CuO-CaO [114]

36 Chen Changkang
For alkali halide flux crystal growth is carried out by a combination of temperature
cooling and solvent evaporation [! 16]. KCI, NaC1 or the mixture of KCI+NaC1 and
KCI+KF were used as flux. Fig. 24(a) and (b) show the system of KC1 - NaCI [117] and
KCI - KF [ 118]. Interestingly, the 2212 phase was grown from a KC1 flux (melting point
of 776 °C) [81,115]. The binary flux system of KC1 + NaC1 (The lowest co-melting
temperature of 680 °(2) produced two phases, 2212 and 2201 [81]. And the binary fluxes
of KCI+KF with the eutectic temperature of 600 °C resulted in 2201 single phase [106].
Fig. 25 (a) and (b) show rectangular crystals of Bi-2201 and 2212 grown from the flux.
800 900 I I '" t 1
600 SS 800
700
400
600
2-Cryslolli ne Phoses
i
200
500
0 I I , ] L _ .... I I I 1 I
0 20 40 60 80 I00
KCt Mol. % NoCI
0 20 40 60 80 I00
KCI Mol. °/o KF
Fig. 24. (a) The system KCI-NaCI [117] (b) The system KCI-KF [118]
I ]
Fig. 25. (a) Crystals Bi2Sr2CuO 6 (b) Crystals Bi2Sr2CaCu208

(scale of grid: 2xl mm 2) (2x2 mm 2 in size)
2. Peritectie Reactions and the Seed Crystal for Melt-Texturing Orowth
One of the major problems for high Tc oxide superconductors is the need to avoid
material having low critical current density, a consequence of the weak link between grain
boundaries. Seeded melt-texturing growth(SMTG) is a promising technique to obtain large
grain YBCO having a high critical current density in a high magnetic field, which has
practical applications in high-field engineering technology, such as magnetic bearings and
levitation.
SMTG involves two opposing directional peritectic reactions:
Tp(l)
2YBa2Cu306. 5 (123) Y2BaCuO5 (211) + L (3BaCuO2+2CuO) (28)
Tp(2)
(1) Above the peritectic temperature(Tp), YBCO decomposes into the "green" phase of
211 and a Ba-Cu-O liquid phase. Our ITGA experiments for the stoichiometric YBCO
system indicates that even at higher temperatures than Tp, there still exists a balance
between reactions (1) and (2) [75]. (2) As the temperature falls through Tp, YBCO grains
form by peritectic reaction and texturing growth develops on the seed crystal in the
semi-solid mixture under a suitable temperature gradient and temperature program. The
dispersion of unreacted or additional 211 particles [120] in the 123 matrix could play a
role as a magnetic flux pinning centre which enhances the critical current density. Small
additions of PtO 2 or CeO 2 have been used to modify the size and morphology of the 211
particles [ 121 ].
Single crystals of NdBCO [ 122] and SmBa2Cu307_y(SmBCO ) [123] are the most suitable
seeds for SMTG as they have the same lattice structure as YBCO and higher
decomposition temperatures than YBCO. Table 11 lists the decomposition
temperature(Td) for rare-earth 123 compounds with the rare-earth ionic radius. Except for
La and Pr, the decomposition temperature increases with ionic radius. Among them,
NdBCO and SmBCO have the highest Td with high solubilities and wide liquidus ranges
for crystal growth. The exceptions of the Td correlation for LuBCO and PrBCO might be
caused by different decomposition reactions to that of reaction (28):
PrBa2Cu307_y .... > PrBaO 3 + L(BaCuO 2 + 2CuO) (29) [ 1241
2LaBa2Cu307_y .... > La2BaCu206 + L(3BaCuO 2 + CuO) (30) [125]
The Td of single crystal PrBCO is about 70 °C higher than that of sintered powder,
probably due to the existence of impurity phases in the sinter, Unlike the other rare-earth
"123" phases, it is difficult to prepare single phase PrBCO by solid state reaction [127].
Fig. 26 shows the dependence of Td for (Yl-xPrx) BCO on the dopant concentration x in
38 Chen Changkang
the starting materials. Td increases with increasing Pr concentration from pure YBCO up
to x~0.4 (Td=1064 °C) and then falls to a constant value of 1052 °C from x=0.5 to pure
PrBCO. Td for single crystal (Y0.6Pr0.4)BCO reaches as high as 1063 °C [128].
Therefore, it is also possible to use single crystals of (Y,Pr)BCO and PrBCO as seeds for
SMTG. Single crystals of (Y,Pr)BCO over the whole doping range have been obtained
from one end of YBCO to the other end of PrBCO [128]. Fig. 27 shows that the Pr
concentration in crystals depends on the concentration in the starting composition. The
crystal composition was analysed by electron probe microanalysis (EPMA) or X-ray
microanalysis (XMA). The distribution coefficient of Pr is ideally equal to 1.
Table 11. Decomposition temperature (Td) of RBa2Cu307_y [ 103,129,143]
R La Pr Nd Sm EuGd Dy Y Er
Ionic radius (A) 1.18 1.14 1.12 1.09 ].07 1.04 1.03 1.02 1.00
Td(°C)inO 2 1054 1090 1078 1073 1056 1020
in Air 1050 1052" 1070 1060 1060 1010 1005 1000 995
984** 1080
* for single crystal grown by flux

** for sintered sample by solid reaction
1060 /
.~" °\\\ e
1040
/
oc
10201
/
~000
i
98O0 t
0-2
I
0 z~
i
O-E
I
08
i
1-0
Pr in start#ng composition
Y*Pr
Fig. 26. Decomposition temperature of (Y,Pr)BCO crystals as a function of Pr dopant

concentration in the starting materials
]-0
/ •
0"8 /
0 /
(.P
rn /
/ •
0,6 ¸
/
>c /
/
0 /
cJ /
e/
/•
(._3
c-- 0"2 j
r~~-
,¢.', •
0 I j
0 0-2 O.L 0.6 0-8 l,O

Pr in solute (Y, Pr )BCO of s t a r t i n g
materiats.
Fig. 27. Concentration of Pr in single crystal (Y,Pr)BCO as a function of its

concentration in the starting materials. • • EPMA; A • X-ray microanalysis.
In order to produce large dimension textured samples, seed crystals having a large (001)
face are required. Fig. 28 and Pig. 8(3) show T-X phase diagrams of NdBCO [130] and
YBCO [59] for the same flux system of Ba3Cu508. The liquidus range for the crystal
growth of NdBCO is wider than that of YBCO. A series of pseudo-binary phase diagrams
of (YI_xPrx)BCO ( x = 0.01, 0.3, 1 ) have been established by DTA using the eutectic flux
of 7BaO+18CuO and are shown in Fig. 29(a) [131], (b) [128] and (c) [126] respectively.
The phase diagram of (Y0.99Pr0.01)BCO - (7BaO+l 8CuO) is quite similar to the system
YBCO - (7BaO+18CuO). Compared to Fig. 8(3), the liquidus range for YBCO is
obviously widened by using the self-flux system having the eutectic composition. A similar
situation also appears in the crystal growth of NdBCO. Table 12 show,: the crystal growth
from fluxes with various ratio of BaO:CuO. As the composition approaches the eutectic
composition from both sides, higher solubility of the solute and lower viscosity of the
solution results in larger yield and bigger crystals. At the end of crystal growth, decanting
the crucible in order to separate crystals from flux is also easier for the flux system with
the eutectic composition. Compared to YBCO in the same flux system, (Yo.7Pr0.3)BCO
and PrBCO have higher solubilities in the flux. From a knowledge of the phase diagrams.
40 Chen Changkang
suitable starting compositions and temperature programs were selected to grow single
crystals of PrBCO, (Y,Pr)BCO, NdBCO and SmBCO. Table 13 summarizes experimental
conditions and results.
xX
Na2BaCuOs+ L \
\k
\
k\
1100 1090 ° -+ lO x
\
i
C NdBa2CusO v 4 1
\
o
loo0
970 ° -+ l O \
NdBa2CuaOT_y+ BaCuO 2 + L
900 c ± 10
9O0 i//
NdBa2Cu3OT.y÷ BaCuO 2 + CuO

I I I i : 1
0 20 40 60 80 100
l Ndga,~Cu~O7 y
-~- tool % / Ba~Cu50~
Fig. 28. Binary phase diagram of NBCO-Ba3Cu50 8 [ 130]
Table 12. Crystal growth of NBCO from the system

(0.13/6)NBCO-0.87[ 1/(x+y)](xBaO+yCuO) [ 132]
No. 1 2 3 4 5
BaO:CuO (mole) 0.375:0.625 0.30:0.70 0.27:0.73 0.25:0.75 0.17:0.83
Growth range (°C) 1040-927 1056-943 1050-9.35 1050-940 1050-940
Decant Not well Well Well Well Can't
NBCO crystals Small Lots Lots Lots Few
Max.size (mm 3) <0.5 4x3x0.25 5x3x0.3 6x6x0.2 <1
T(C°) - -
~020
o o L
1000 °"%,~o
98O (Y,Pr )BCO*L 'c~
o
£60 ',>
£L0 x
(Y,P,)BCO+ CuO*L
920
900 (Y, 8 )BCO*Ba Cu Q'~*CuO*L
880
860
8LO
820
800 t I; 2~0 30 L0 50 60 70 8 90
g ~Y099P''~01',BCO 2~{TBoO÷18CuO)
(a)
\
\
•.m T(C°)
1040
t Singte crystet melts inconyruent|y ot
1052 °C
I000
1020 \
(y, pr)BCO ÷ L \ 980, ~-- L
1000
\ 960 Pr BCO÷L "x\
o~ 980 .^.\ o
._ ..... ._.,..,,~ 940 ....... o ' o-- o ~ ° o~.~1
~- 960 i
Pr BCO÷CuO*L I
940 (Y, Pr )BCO-"CuO ÷ L 920 o o--O~ °-°-o°-°cb
920 900 P, BCO+BQCuO2*CuO*L
900 ( Y, Pr )BCO*B°CuO2÷CuO" L
880
860
840 !
820
800 L I I I I I I I [
{-(Y, Pr)BC0 20 L0 6~0 8J0 2~(7BaO'lSCuO) 10 20 30 ~,0 50 60 70 80 90
gl P, BCO 2~(7BaO*18CuO)
(b) (c)
Fig. 29. (a) Binary phase diagram of (Yo.99Pro.o I)BCO-(7BaO+ 18CuO) [ 131 ]
(b) Binary phase diagram of (Y0.7Pr0.3)BCO-(7BaO+I 8CuO) [128]
(c) Binary phase diagram of PrBCO-(?BaO+ ! 8CuO) [ 126]
42 Chen Changkang
Table 13. Crystal growth of PrBCO, (Y,Pr)BCO, NBCO and SmBCO
No. Crystal Starting composition Grow.range Cool.rate Crystal Ref.

(1/6)RBCO 0.28BaO+0.72CuO Max.size
( mol.% ) ( °C ) ( °Ch-1 ) (ram 3)
1 PrBCO 0.15 0.85 1040-920 0.3 7x3.5xO.6 126
2 (Y,Pr)BCO 0.15 0.85 1040-958 0.5 8x7.5xO.5 128
3 NBCO 0.35 0.65 1100-970 1 5x5x3 132
4 SmBCO 0.15 0.85 1050-980 1

980-899 0.5 6.2x4x2.5 133
Fig. 30 is a picture of single crystal of (Y,Pr)BCO grown from the eutectic flux system.
The doping concentration in the starting material was Pr/(Y+Pr) = 40% (at.). These
crystals reached to the highest Td = 1064 °C shown in Fig. 26.
i .4
0 i tp) rT~
Fig. 30. Single crystal of (Y,Pr)BCO

3. Liquidus Line and Crystal Growth of (La,Sr)2CuO 4
(La,A)2CuO 4 (A = Ba or Sr) is regard as a binary ox~ae system, in which La atoms are

partially substituted by Ba or Sr. Fig. 31 (a), (b) and (c) show the binary systems of L a 2 0 3
- BaO [ 1 3 8 ] L a 2 0 3 _ CuO [94] and SrO - CuO [114, 189]. Amongst the four oxides
L a 2 0 3, BaO, SrO and CuO, copper oxide is the only compound with a relatively low
melting temperature, which has been used as a self-flux for crystal growth.
oof
. /,"
/
//
1573
I,K
L~20~, L 1585 I<
\\
\
x
q
\
,'L'L'~I l/l/ ]
/
//
/
] \
J I
/ ]
/ I
\\ / / 1373
,bOC 1325 K i / r
177~,
[
1600
H75 i 4
So'S' 20 40 60 80 !2
L%03 f-O 70 80 90 103
,.~g rno[ Cu3 0~0
(a)
(b)
-- -q . . . . . ] ~-
14(,'1() -
15©0
1225±5o ~ \
U 12QO ! - ~ .... ~-- - ~ ' "
o\
/!
1026~
~ 1000
q()O t
I I
6()Or[ ~ l I. _ l . . . . I ~ .J _ _ L _ . _ _ ~ , ~ .....
HIO
",r() M,I Sr ¢ t.O
(c)
Fig. 31, Binary systems of (a) L a 2 0 3 - BaO. (b) L a 2 0 3 - CuO. (c) SrO - CuO
44 Chen Changkang
In the system of L a 2 0 3 - CuO there is a narrow liquidus range for the crystallization of
La2Cu20 5, which is formed by the peritectic reaction at 1050 °C [179]:
La2CuO 4 ( S ) + CuO ( L ) .... > La2Cu205 ( S ) (3])
In the crystal growth of LSCO by slow cooling or temperature gradient transport, LSCO
crystals were found to be partly dissolved by the reaction, resulting in rectangular plate
crystals of (La,Sr)CuO3_y with a perovskite structure. This phase was not observed in the
phase diagram established by DTA, because a much higher heating or cooling rate was
used. In the system of SrO - CuO there exists the interesting compounds, the infinite-layer
compounds Sr2CuO 3 and SrCuO 2, a~s well as the ladder compound SrI4CU24041. After
high pressure treatment at high temperature or fluorination at lower temperature, Sr-Cu-O
system posses high Tc superconductivity.
From the information of phase diagrams, it is possible to grow (La2_xSrx)CuO 4 crystals
with various Sr-doping concentrations ranging from La2CuO 4 on one side to Sr2CuO 3 on
the other. Superconductivity of LSCO under normal condition only appears over a short
St-doping range, with the highest Tc ~ 35-40 K for x ~ 0.15. In contrast to sintered
material, it is very difficult to grow the single crystal having a high Tc. Flux grown
crystals always exhibit a low Tc and a broad transition width or not even exhibiting
superconductivity. This problem has not yet been solved although some explanations
[139] based on points of structure, composition or contamination from crucibles have
been made. It seems that the doping structure is strongly dependent upon the
experimental conditions. Studies on the crystal growth of La2CuO 4 - Sr2CuO 3 is a route
to understand the relationship between crystal growth, structure and superconductivity.
By DTA Two similar phase diagrams of La 1.85Sr0.15CUO4 - CuO and La2CuO 4 - CuO
in air and in 0 2 respectively are shown in Fig. 32(a-d) [85, 183]. They are typical
eutectic systems with incongruent melting of LSCO:
LSCO(S) .... > L a 2 0 3 + L (32)
The peritectic and eutectic temperatures in 0 2 are a little higher than that in air. Below
the eutectic temperature there is a solid line which probably corresponds to the
decomposition reaction 2CuO ---> Cu20 + Y 2 0 3 , v~here Td in 0 2 is higher than in air.
The phase diagram indicates that CuO is an excellent flux for LSCO crystal growth. The
starting concentration could be set up along the liquidus line from ~10% to 40% of CuO.
But due to the decomposition of CuO during soaking, the real system is a ternary one
containing LSCO, CuO and Cu20. The heating or cooling rates in DTA experiments are
much higher than in crystal growth. And flux wetting or creeping appear very seriously
when using small DTA crucibles with only 100 - 101 mg samples. The liquidus line in the
phase diagram includes many uncontrollable and indefinable factors.
.-ka) f ° ~'" °- (b) - ~
1300
1200 i t., C~O [ L + LSCO 12oo i. ~i^ ~ L÷ LSCO

E
1100
o
-o.Z-Jn ~ u o ~ - - o-- o -,~-d
100(2 r v ' c ~ - c ~ e n v E 0
n
,oo~
!
900]- CuO+ LSCO CuO LSCO +

900 I
0
L .
20
I
40
J
60 80
I '...
100 0 20
J
40
J
60 80
~ [
] ~g)
CuO tool % LSCO LSCO CuO mol % LSCO LSCO
(a) (b)
T( ° C)! - - - - T( ° C)
1400 1400
0
/ fi ± : •
I
1300 I
1300 1
.I
I /
i/
1200 L ÷ LCO 1200 M~
I
/
!
,' L +LCO
I I
t100 r-'u o "
1100
g ii- • it m L m •_,¢ . . . . .
1000 1000
900 900
o 2'6- 4 G'o 50 loo o 20 4'0 do 80 lOO

Cu0 Mot.% I.~CuO, Cu0 Mol.% La~CuO,
(c) (d)
Fig. 32. The system of LSCO-CuO (a) in air; (b) in 0 2 and La2CuO4-CuO (c) in air; (d)
in 0 2
46 Chen Changkang
In order to determine the liquidus line for crystal growth, an situ ITGA was carried out
using more than 50 g starting materials. The high sensitivity and stability of the electric
balance enabled the cooling rate to be greatly reduced. The crystallization temperatures
have been determined using cooling rate as low as 0.2 °Ch-1. Table 14 shows the
crystallization and dissolution data for LSCO from the same starting composition by
ITGA. As the cooling rate decreases from 136 °Ch-I to 0.5 °Ch -1, the degree of
supercooling decreases and the crystallization temperature slowly increases from 1210 to
1219 °C. In the mean time, the apparent TG slope, which corresponds to the
crystallization rate, decreases rapidly from 100 to 0.01 mg/min. The larger the heating
rate, the higher the dissolution temperature. From a comparison between the
crystallization and dissolution temperatures, we conclude that the saturation temperature
for the system is located at 1219 - 1220 °C.
Table 14. Crystallization (Tk) and Dissolution(Td) temperatures of LSCO crystals by

ITGA
Starting compositions(mol.) Cooling rate Tk Heating rate Td

LSCO Cud Sr/La (oCh- 1) (oc) (oCh- 1) (oc)
0.23 0.77 0.24 136 1210 365 1245

32.3 1212 136 1242
13.6 1215 13.6 1223
0.5 1219 0.5 1220
Using in-situ thermogravimetric analysis, the liquidus line was built up and is shown in
Fig. 33(a). Fig. 33(b) shows that the solubility S of LSCO at various temperatures follows
the Arrhenius equation (26), from which the heat of crystallization of LSCO is determined
to be 20 kcal./mol, by the least square method. Large crystals have been grown either by
spontaneous nucleation or top-seeding along the liquidus line. Crystals were separated
from the solution either by decanting the crucible or by lifting the crystals above the
solution surface at a temperature higher than eutectic one. Fig. 34(a) and (b) show
La2CuO 4 and LSCO crystals grown by the flux method. Recently a Tc of 37 K with
transition width of 3 K for an as-grown crystal has been obtained by the author. After
oxygenation the Tc increased to 38.5 K and the transition width reduced to 1.5 K. Fig. 35
shows the AC susceptibility measurement for the LSCO crystal [181]. The
superconductivity is better than that for crystals grown by floating-zone [182] which has
been regarded as the only method for growing high Tc LSCO crystals up to date. The
insulating crystal of La2CuO 4 becomes superconducting after oxygenation under high
oxygen pressure [ 183].
3-6
\
o
1300
A euteclic point
3-2
\o
oa c
~- 1200 2.8
I
c
o
1100 %
N
%
/ 2.LI-
I \o
2.
c3 1ooo l 1 4 2.0[ k
o 10 20 30 4O 62 6.6 7.0 7./. 7-8
tool °Io L S C O I/T (IOt'K-11
[o)
Fig. 33(a) Solubility of LSCO in LSCO-CuO system

(b) Arrhenius plot of the solubility
(La,Sr)2CuO4.y
La2CuO 4
Fig. 34 (a) La2CuO 4 (b) LSCO crystals grown by flux method

48 Chen Changkang
,¢..~
~d
4-
# o
. ,...~
cj)
25 30 ,~,~
"": 40 ~:rl5 "]( H }
Fig. 35. AC susceptibility measurement of as-grown crystal LSCO
4. Selection of an Appropriate System for the Crystal Growth of YBCO
( l ) Self-flux
YBCO is one of the most important high Tc superconductors. The crystal growth has
been extensively studied by large numbers of investigators. Because of the low stability
and the high reactivity of YBCO and the corrosion of crucibles by flux, it is difficult to
find an ideal flux for crystal growth. Almost all types of practical molten solvents that are
used for the growth of oxide crystals [72], including heavy metallic oxides, borate,
vanadate, molybdate, halide and carbonate [140-142], are found to be unsuitable for the
crystal growth of YBCO, Up to date the most useful fluxes are the self-fluxes CuO and
BaO-CuO. Fig. 36 shows the ternary phase diagram (at 950 - 1000 °C) [143] with three
edge binary systems of Y O I . 5 - BaO, BaO - C'uO and CuO - YO1. 5 [144]. Due to the
high melting temperatures of Y 2 0 3 (~2400 °C) and BaO (1918 °C), Y 2 0 3 - BaO could
not be used as a flux. The melting point of CuO is much lower (~1200 °C) and eutectic
temperatures of CuBaO 2 - CuO and Y 2 C u 2 0 5 - CuO in rich-CuO side are below the
decomposition temperature of YBCO, which offers the possibility of their use as fluxes.
The eutectic system (e l) of CuO - BaCuO 2 has a lower melting temperature and wider
liquidus range than those in the eutectic system (e2) of CuO - Y2Cu205 [68], so that it's
more useful for crystal growth. The dashed line in the ternary system indicates a region of
partial melting, where the three lines YBCO-el, YBCO-CuO and YBCO-e 2 pass. A
completely dissolved system was obtained by using the eutectic flux of CuO - BaCuO 2
w i t h a l o w concentration of solute. The ternary eutectic c o m p o s i t i o n of Y B C O - BaCuO 2

- C u O is very d o s e to the binary eutectic system. A r o u n d Y B C O three crystalline phases,
CuO, B a C u O 2 a n d Y 2 B a C u O 5, occasionally a p p e a r in the s y s t e m t o g e t h e r with Y B C O
crystals.
i
f
(.
Y )I 5 Y2E~I 4 Y2 8020"5 YZ Ra407 8(10

t Y4 e°3~9 . . . . . . . ~ 0~7
I : |
i 0~ t I
+ I OH, +d'l Id-
,
I
i 01,7, I m
I I
. . . . . . . . .
1t,'_.
1 009
oo], + 0 t'6 i I 010 *01.1,
t . . . . I. . . . c +0-b}
!. . . . . . . . -I
O0~t
i
O £ b + 0 l,O
O09t
0 01, + OC._lz
O00g
[3o]Z
Fig. 36. S y s t e m of Y O 1.5 - B a O - C u O

5O Chen Changkang
(2) Eutectic flux system
The published binary eutectic points of BaCuO 2 - CuO from various laboratories [88]
show large differences. The reason is the different experimental conditions that have been
used to determine the eutectic points, such as the measurement methods, crucible
materials, atmospheres and difficulties to reach the equilibration state and to control the
corrosion of crucibles. The normal methods for determining the phase diagram are DTA,
direct observation or quench experiments combined with phase identification. In the
author's laboratory, ITGA has been adopted for determining the liquidus line. Data were
obtained starting started from a 50 g mixture with the composition BaO:CuO = 0.33:0.67
(Mol.) and then certain amounts of CuO were added step by step to change the
composition gradually. Fig. 37 shows the binary system BaCuO 2 - CuO. Combining the
above ITGA data with that of other authors [145, 6 5 ] we confirm the eutectic
composition of 0.28BaO+0.72CuO or (1/25)(7BaO+18CuO) and that the eutectic
temperature is below 900 °C.
1100 c,
/
L +CueC~ L /
T(°C ,~ o/
/
~oo( ~9-2-e-°c--k /
/.
°1
\ ,!
/
!
L*CuO ~ iI L÷BaCu02
I
I
900 °t
/,
!
CuO ÷ BaCuO 2
800
CuO 90 ~0 70 60 BctCu02
~CuO(mol.°/o)
Fig. 37. BaCuO2-CuO phase diagram
There are two liquidus ranges of CuO+L and BaCuO2+L, which are suitable for crystal
growth of CuO and BaCuO 2 respectively. Fig. 38 shows two BaCuO 2 crystals with sizes
up to 5.5x5.5x3 mm 3 sitting on the surface of a single crystal of CuO with a size of
30x5x4 mm 3 grown from the starting composition of 0.80CuO - 0.20BaO. This

phenomenon has been called "eutectic jump" in the author's laboratory and can be
explained as follows: as the temperature decreases and approaches the eutectic point, the
viscosity of the solution increases and a local rich concentration of BaCuO 2 is created on
the growth front of the CuO crystal. Because the two liquidus lines are steep and close to
each other near the eutectic point, the crystalline phase can jump from the CuO side to
the BaCuO 2 side, resuldfig in the nucleation and growth of BaCuO 2 on the surface of the
CuO crystal.
Fig. 38. BaCuO 2 crystals setting on CuO single crystal
The eutectic temperature is the lowest melting temperature for the binary system. As the
composition departs from the eutectic point along both sides, the dissolution temperature
increases. Thus, the flux composition has a great effect on crystal growth. Table 15 shows
the dependence of the liquidus range for YBCO as the primary crystalline phase on the
flux composition. As the composition approaches the eutectic point from both sides, the
solubility of YBCO in the flux increases and the concentration range for crystal growth
gets wider. There are three major problems associated with crystal growth of YBCO.
They are slow equilibration, low solubility and crucible corrosion. By using the eutectic
flux, the solubility is enhanced and complete dissolution speeds up the equilibration
process, which makes the experimental results more reproducible. As the growth
52 Chen Changkang
temperature is relatively low compared to that using the non-eutectic flux at the same
solute concentration, the corrosion of the crucible might be relieved to some degree. The
author considers that the eutectic flux of 0.28BaO+O.72CuO is a good choice for crystal
growth.
Table 15. Liquidus range for YBCO as the primary phase in the system (1/6)YBCO -
1/(x+y) (xBaO+yCuO)
No. BaO:CuO C*(MoI.%) AC*(Mol.%) T (°C) AT* (°C) Ref.
1 0:1 0 0 55
2 0.20:0.80 7-4 3 (1020-970)+10 50+20 55
3 0.28:0.72 20-4 16 1010-950 60 40
4 0.30:0.70 15-3 12 (995-920)+ 10 75+20 55
5 0.375:0.625 12-10 2 (1020-970)+10 50+20 55
C* : (1/6)YBCO
AC* & AT* : Concentration or temperature ranges where YBCO is the primary phase
respectively.
The pseudo-binary phase diagram of (Y0.99Pr0.01)BCO - (0.28BaO+0.72CuO) shown in

Fig. 29 is quite similar to the eutectic growth system of YBCO. From the diagram we can
introduce some important parameters for crystal growth. Below the peritectic point of
~0.20(1/6)Y-BCO and 1010 °C, there is a liquidus range for the crystal growth of YBCO
as primary crystal phase. Below 950 °C the second and then the third crystalline phases
of CuO and BaCuO2 appear, but YBCO phase still continues to grow until eutectic
solidification. Crystals can be separated from the flux by decanting the molten solution
above the eutectic temperature.
(3) Additives
Although the solubility of YBCO is increased by using the eutectic flux, the absolute value
remains quite low. For instance, in the system of 0.10(1/6)YBCO -
0.90(0.28BaO+0.72CuO) the concentration ratio for the starting materials is
Y203:BaCO3:CuO = 1:34.1:83.8, which is very close to the ternary eutectic point [ 104].
In order to further improve the growth system, the author and his colleagues have carried
out extensive investigations on the effect of additives [ 104, 131, 146, 174].
As described above nearly all the conventional fluxes, except the self-flux, are unsuitable
for use in the crystal growth of YBCO. Instead they produce other crystal phases. But by
reducing them to a certain low level as additives in the seE-flux system, it is possible to
reach a situation where YBCO remains as the primary phase. Sometimes, a minor amount
of a special additiw, greatly improves parameters of phase diagram resulting in nucleation
and growth of the desirable phase.
(a) The effects of additives on crystal growth.
Firstly, crystal growth was studied with various single additives such as BaF 2, BaCI2,
Bi20 3 or B 2 0 3. The results show that a small amount of BaF 2 as an additive is clearly
beneficial for the crystal growth of YBCO. It increases the solubility of the solute and
reduces the solution viscosity resulting in a higher yield and larger single crystals, this
also makes the separation of crystals from the flux by decanting much easier. In contrast,
additives of B 2 0 3 or BaCI2 decreased the yield of crystals and resulted in the preferred
production of other crystalline phases, such as C u t crystals. Experiments with double
additives of BaF2+B20 3 or BaF2+Bi203 confirmed that the main improvement was
achieved by using BaF 2. Therefore more investigations have been focused on the effects
of BaF 2.
In fact by using BaF 2 as a flux, the main crystalline phase is not YBCO, but C u t , BaF2
et al [ 141 ]. However by using a small amount of BaF 2 as an additive in the eutectic flux
system, the primary crystalline phase of YBCO remains and its crystal growth can be
improved. Table 16 shows the dependence of crystal growth on the additive
concentrations under the same system of 0.13(1/6)YBCO - 0.87(0.28BaO+0.72CuO)
from an alumina crucible in the same furnace. The optimum concentration of BaF 2 for the
growth of larger crystals appears to be around 1 wt.%. As it increases over 1 wt.%, the
size of YBCO crystals is reduced because of its enhanced solubility. Ultimately the
primary crystalline phase gradually becomes C u t . With more then 2 wt.%, the
crystallization temperature moves down below 900 °C and the cooling range has to be
expanded to 880 °C. No YBCO crystal is obtained in the cooling range from 1000 to 880
°C as the BaF 2 concentration reaches 3 wt.% or more.
54 Chen Changkang
Table 16. Crystal growth of YBCO using various concentrations of BaF 2 as additive
No. BaF 2 Cooling YBCO crystals Other phase

Range Rate Max.size Max.wt
(wt.%) ( °C ) (°ch-l) (mm 3) (mg)
1 0.5 1000-900 0.5 4x3.5x0.3 22 CuO

2 1.0 ,, , 5x3xO.6 36 None
3 1.5 , ,, 4x3xO.4 31 CuO
4 2.0 ,, , <lxO.5 <1 CuO
5 2.2 1000-9000 2
900-880 0.2 3x2xO.6 17 Mainly CuO
6 2.45 1000-9000 2
900-880 0.2 3x1.6x0.4 11 Mainly CuO
7 3.0 1000-9000 2
900-880 0.1 None 0 CuO
8 3.5 1000-9000 2
900-880 0.1 None 0 CuO
(b) The effects of additives on the composition, structure and superconductivity of YBCO
crystals.
The addition of BaF 2 improves not only the crystal growth by modification of parameters
of the phase diagram, but also the quality of the crystal with improvements of purity,
stoichiometry, structural ordering and morphology, and hence of the superconductivity.
Table 17 shows the composition and Tc of YBCO crystals grown in an alumina crucible
from the eutectic flux with various additives. Compared to the crystal without any additive,
the addition of BaF 2 reduces the contamination of the crystal by A1 from the crucible
material and leads to a Ba content close to stoichiometry, which results in high Tc. The
contamination is caused by the partial substitution of Cu atoms in the structure by AI.
Bi20 3 may have the same positive effect, but it did not benefit the growth of the crystal
[131]. BaCI 2 improves the stoichiometry, but does not suppress the AI impurity. B 2 0 3
does not make any improvement either in the contamination or in the stoichiometric
composition, resulting in crystal having the lowest Tc. Only specific additives appear to
be useful for the growth of any one crystal.
Table 17. Composition and Tc of YBCO crystals grown with various additives
No. Additive Composition of crystals Tc after anneal*

Y Ba Cu A1
(wt.%) (at.) (K)
1 None 0.99 1.88 2.88 0.12 57

2 1 B20 3 0.99 1.86 2.86 0.14 53
3 ! BaF 2 0.98 1.94 2.94 0.06 90
4 1 Bi20 3 0.98 1.94 2.94 0.06 80
5 0.9 BaC12 0.98 1.97 2.89 0.11 77**
* • in oxygen at 500 °C for 24 hs

**: in oxygen at 600-400 °C for 48 hs
Further studies confirmed the effects of BaF 2 on AI contamination and the

superconductivity of the crystal. Fig. 39 (a) and (b) show the relationship between the
BaF 2 concentration in the starting materials and the AI content in the crystal and the Tc
and ATc of as-grown crystals. With the addition of more BaF2, the aluminium content in
the crystals decreases, whilst Tc increases and the width of the transition reduces.
L (K) AL (K)
O •
At (at. per formula)
\
90 ~-~o o
0.06 25
O
0.04
Q ~ o
O~O
0.02
0.5
I I
1.0 1.5
I
2.0
I
2.5
I _._ 60I~l
0 0.0-2
I
0.04
i
0.06 0.08
BaF2 (wt.%)in starting rnoleriel.~;
AI (at. per formula)
Fig. 39 (a) The dependence of A1 content in YBCO crystals on BaF 2 concentration

(b) The dependence of Tc and A Tc on the AI contamination
56 Chen Changkang
The nanostructural morphology and the atomically resolved image obtained by scanning
tunnelling microscope (STM) [147] have provided some information about the relation
between the structure and properties of YBCO crystals grown with various additives. It
was found that noticeable distortions in the Cu-O chains and Ba-O planes of the structure
occurred with B 2 0 3 as an additive. As boron has a much smaller ionic radius than any
elemental components in YBCO, boron may enter the lattice as an interstitial. This
introduces structural distortions in both layers and causes a stoichiometric deviation of the
Ba content. The distortion and the non-stoichiometry together with the AI contamination
cause a large Tc degradation of the sample. The X-ray study of the structure refinement
[148] revealed that AI only partially substituted copper atoms in the Cu-O chains. Due to
the difference of ionic size of A1 and Cu, the substitution increases the surface roughness
of the chain layer. The nanostructural roughness of the Cu-O chain has been measured by
STM for four kinds of crystals grown by using none, B203, BaF 2 and Bi203 as additives
in an alumina crucible respectively. The results are shown in Table 18. The crystal with
BaF 2 additive has the lowest degree of roughness, corresponding to its high
superconducting transition temperature.
Table 18. Nanostructural roughness of Cu-O chains in YBCO crystals with various
additives
No. Starting composition Roughness Tc

(1/6)YBCO 0.28BaO+0.72CuO Additive of Cu-O chain As-grown Anneal
(wt. %) (arb. units) (K )
1 O. 12 0.88 None 90 56 57
2 ,, ,, 1 B20 3 88 39 53
3 ,, ,, 1 BaF 2 43 81 90
4 . . . . 1 Bi20 3 56 65 80
The conductance spectra obtained using STM allows further interpretation of the
depression of AI contamination and the improvement in Tc by the use of BaF 2 additive.
The spectra of crystals with a low AI content grown using the additive BaF 2 indicates a
typical quasi-metallic nature in the Cu-O chain. In contrast, the distinct notch structure in
the conductance spectra for crystals grown with B 2 0 3 suggests an insulating character in
the chain layer. Table 19 shows the schematic arrangement of atomic layers in the YBCO
unit cell. As is well known, the Cu-O plane in the structure is directly responsible for the
superconductivity and the Cu-O chain acts as a charge reservoir. The nearest Cu-O planes
across the Y-plane are separated by a distance of 3.4 A. However, their separation across
the Cu-O chain is 8.2 A, which is much larger than the superconducting coher~,nce length
(5~6 A) in the c-direction. If the intervening chains are insulating, it is very difficult for
the superconducting wave function to overlap each other which results in a lower Tc.
Table 19. Schematic arrangement of atomic layers in the YBCO unit cell
Atomic layer Distance (A) Layer character Distance (A) Unit cell (A)
Cu-O plane Conducting layer

T
1].7 1
"/-plane 3.4
1.7
Cu-O plane '
~ Conducting layer 1
2.3 l
Ba-O plane ~ 4.1 11.6
Cu-O chain 118

' Charge reservoir 1
Ba-O plane
'
1.8
I
1
4.1
2.3 ]
Cu-O plane ~ Conducting layer
(c) Mechanism of crystal growth
A model has been proposed to understand the functions of the additive BaF 2, which can
be summarized thus: (ij to improve crystal growth by increasing both the solubility and
the yield of crystals and reducing the viscosity of the solution; (ii) to improve the
superconductivity by depressing the A1 contamination and moving the composition
towards stoichiometry. The optical microscope observations by many authors have
occasionally shown spiral growth steps on the (001) face. The morphology recorded on a
nanometer scale by STM [149,150] confirmed the spiral growth features and revealed a
growth step with a height equivalent to C parameter of the unit cell. Also the topmost
atomic layer consisted of Cu-O chains. The impurity A1 substituted for copper atoms in
the Cn-O chain. All evidence indicates the crucial nature of the Cu-O chain in crystal
growth and superconductivity. The model suggests a screw dislocation growth mechanism
and a cluster model of monomolecular layers with a height of the C parameter controlling
the crystal growth. Figure 40 shows schematically the cluster model for the crystal growth
of YBCO. (a) There is an equilibration between the polymolecular clusters with different
58 Chen Changkang
numbers (n, m,...) which specify the molecular layers. The topmost atomic layers on the
clusters are Cu-O chains, which have many oxygen atom vacancies due to the use of a
high temperature solution. (b) Fluorine ions can occupy some of the 0 -2 sites in the
chains. Bond-breaking of the polymolecular layers into monomolecular layers occurs more
effectively, because F- ions have lower charge than 0 -2 and are more loosely bonded in
the clusters. (c) The clusters of monomolecular layers, which are the basic growth units,
incorporate into kinks at the growth steps by a screw dislocation mechanism for the (001)
face or probably two-dimensional nucleation mechanism for the (100) or (010) faces
[151]. After the incorporation. F- ions may re-enter the solution and be available for
further catalysis.
n>_2 n > m _>2 n=1

(o)
(b)
F- i~),
F
(c) \
\
\
\
\
\~F- S
~lilii~liil7
/
J ; <4 </////;:
~/////,I/'//i/ ,'. ,/.../ .-//// " ...-
"I//////__ " ' / / . / / ".......,
.//'/..'/'//'..///./'"
/ /
// / / / /./ / / / / / / .
Fig. 40. Cluster model for the crystal growth of YBCO

With the aid of this model, improvements in crystal growth and superconductivity through
the use of a BaF 2 additive can be understood: (i) The bond-breaking effect of the F- ions
reduces the solution viscosity and enhances the concentration of monomolecular clusters,
which results in a high yield and a large size of the crystals. The solute solubihty
increases and the crystallization temperature decreases. (ii) The occupation of the oxygen
sites in the Cu-O chains by F- ions depresses the substitution of Cu atoms in the chain by
aluminium, reducing the A1 contamination and the roughness of the chain layer. The lower
structural distortion may also promote the packing of Ba site stoichiometrically near the
chain. These effects result in the improvement of the superconductivity. There have been
numerous investigations on the effect of substitutions on the superconductivity of YBCO
materials. One of the most interesting studies is the fluorination by (1) solid state reaction
[152]; (2) ionic implantation [153] and (3) solid-gas reaction [154] which are all
performed in the solid state under a relatively lower temperature. The crystal growth in
high temperature solution with fluorine ion as additive provides a new route to study
fluorination of high Tc superconductors and the growth mechanism of crystals with layer
structure [ 174].
(4) Crystal growth in yttria crucibles from the self-eutectic flux with BaF 2 as an additive
Aluminium contamination in YBCO crystals can be completely avoided by using a yttria

crucible, so that the role of F- ions ,(other than the function of depressing the A1
contamination), could be identified. The author's studies show that the disappearance of
AI in the crystal leads to the composition moving closer to stoichiometry with a
significantly enhances orthorhombicity of the crystal, resulting in a higher and more
reproducible Tc. Figure 41 shows the dependence of Tc and transition width ATc on the
concentration of BaF 2 in the starting materials. Tc slightly increases with BaF 2 content
up to 93.7 K as 2 wt.% BaF 2 is added, and then decreases. The transition width is more
strongly dependent on the additive: 0.8 K for no additive which reduces to 0.1 K for 2
wt.% BaF 2 and afterward widens again. In the meantime, the main crystalline phase
gradually changes to CuO and the yield or size of YBCO crystals become smaller and
smaller. This suggests following the cluster model that bond-breaking by extra F- ions
occurs in other structural layers, such as Cu-O and Ba-O planes, destroying whole
structural framework resulting in the formation of simple CuO crystals.
Combining the research achievements described above, crystal growth of YBCO has been
carried out by using: (a) pure Y 2 0 3 or Y203:Ba crucibles to avoid any additional
contamination; (b) eutectic flux system of 0.28BaO+0.72CuO to widen the liquidus range
for the crystal growth; (c) BaF 2 as an additive to improve the crystal growth and the
superconductivity. As a result single crystals with sizes up to 18xlSx0.3 or 5x3x2.2 mm 3
were obtained by spontaneous nucleation. Fig 42 shows the AC susceptibility of one
crystal with Tc = 93.7 K and the transition width of 0.1 K.
60 Chen Changkang
1.2
] .0 94.0
- 93.{i
I,
<• ().6 - 93.2 ,,'%
O. 4 92.8
(}2 1 d 92..!
0 1 I I 92.0
() '2 10
BaF., (w1.9{,)
Fig. 41. Dependence of Tc and ATc on the BaF 2 concentration in the starting materials
used for YBCO crystals grown in yttria crucibles
U3
~4
c~
I 1 I I I i I I I | I |111 II • • ill dl ,m ill roll
• 92 93 .~ 94 T(K) 95
) Ol el ol ~ t- h I I IS I I oo N I I .O I I
Fig. 42. AC susceptibility of YBCO

5. Phase diagram and growth technique
(1) Crystal growth of Pb2Sr2(Y,Ca)Cu308 (PSYCCO) using an immiscibic flux system
Compared with YBCO, PSYCCO has a lower Tc of 70-80 K, but it is more stable
chemically and the as-grown crystals are optically untwined. The compound melts
incongruently and a flux has to be used for crystal growth. A m o n g s t the five oxide
components, PbO and CuO have the lower melting points and can be used as flux for
crystal growth. Figure 43 (a) and (b) show binary systems of PbO - CuO [97] and PbO -
C u 2 0 [98]. Both are simple eutectic systems with eutectic composition around 80 Mol.%
PbO. So the fluxes for crystal growth are always PbO rich. PSYCCO crystals have been
grown from flux systems of PbO - PbF 2 [29,30] PbO - CuO [33] and PbO - B 2 0 3 [341.
Either a reducing or mildly oxidizing atmosphere or sealed container for crystal growth
with subsequent annealing had to be used to obtain superconducting crystals. The
precursors of Sr2Y0.5Ca0.5Cu306+y(SYCCO) and even PSYCCO had to be prepared as
the starting materials for crystal growth. PbO - NaC1 [31,32] is a special flux system,
from which superconducting crystals were grown in air without using any precursor. The
growth system of PSYCCO - (PbO+NaCI) is an immiscible system, which is illustrated in
Fig. 44. The fluxed melt divides into two immiscible liquid layers with a distinct interface
in between. The lower layer is the flux system for crystal growth while the upper layer
contains mainly NaCI doped with Pb +2 or Cu +2. In the solidified melt, transparent NaC1
crystals are easily visible in the upper layer. PSYCCO crystals form in the lower layer
where a few small NaC1 crystals are also visible.
(B)
I i \l I t t I . . . . . . I
Mol %
_Cu20 + Liquid "N,b 0 20 40 60 80 I00
I 0 0 0 -- IO'26 *-~ ~,. L lqu~d
i200
CuO * Liquid ~'-...... P b O * LIQuid ~
812o ""...,o / f ~ " ~000 i
800
80O
600
CuO * PbO
600
6 S,C°
-. = £r, ~ Cv20* PbO
\Y-i .)
540'
CuzQ*Cc;C PbO C%0 PbO*Pt<

CuO * P b 2 Q 400
400 [ I ! I 1 .... " . . . . l
CoO 20 4o 6o 8• Pt-,(;, C u 20 20 40 6C 80 =~0
Mol °A
Fig. 43. Binary systems of (a) PbO - CuO and (b) PbO - C u 2 0
62 Chen Changkang
ALUMINA TUBE
FLUXED M E L T
PLATINUM
CRUCIBLE
PLATINUM
SUPPORT
PEDESTAL
ALUMINA TUBE
THERMOCOUPLE
Fig. 44. The flux growth system with two liquid layers
A detailed study of the phase diagram by DTA as well as its crystallization by I T G A have
been carried out in order to understand the function of the immiscible system [32]. Fig.
45 shows phase diagrams of (a) a one layer system of PSYCCO - PbO and (b) a two
layer system of PSYCCO - (PbO+NaC1) in air. Both systems are of the peritectic-eutectic
type, but there are large differences in their parameters. For the one layer system, the
eutectic point locates at a temperature of 800 °C and a composition of 10%PSYCCO -
90%PbO. A peritectic reaction occurs at ~900 °C. The liquidus line is nearly vertical at
the eutectic composition, which indicates that it is impossible to use PbO as a flux alone
for crystal growth without controlling the atmosphere. For the two layer system, some
important parameters for crystal growth have been changed dramatically: (1) The eutectic
temperature decreases to 770 °C as a small amount of NaCI enters into the growth
system: (2) The decomposition temperature of PSYCCO increases to ~930 °C as air is
insulated from the system by the covering of the upper layer of NaC1; (3) The solubility of
PSYCCO is greatly increased as NaCI is not only an air-insulating layer, but also provides
an excellent fltr~ in the system. Combining the three effects together, the liqidus range for
crystal growth is thoroughly improved. Although the allocation of the liquidus line
remains undetermined because the thermal effects are to() small to be detected by DTA,
single crystals of PSYCCO have been successfully obtained from the system with a
PSYCCO concentration as high as 33Mo1.%.
1100
Liquid L..I.I--O / f
- (e) d
I000
T(°C) oo/oo o
Liquid .," n o~
/
90£ /
900 O/QO 0 U Ill 0 n
/
J
g /
, L*PSYCCO z L + PSYCCO
I //
800 ~ U 0 0 ~ IllO u
80£
L)
o oo ° o o o o o "q
_,PbO
L + PbO
7(K
70C PbO÷PSYCCO
PbO 4- P S Y C C O
(~lt(I}
(+ Irace NaC]I
600
i
I
PbO
I
10
i
20
310 1
40
I
50
I
60
~
0 9'0 PSYCCO ql }.i P _ ~2___~
]0
........
20 3(}
~__._2
40 5 )
.
t)(i
.
71!
. .
b;() 90
M°I °' PbO tool 9~ PSYC(O

(+NaCI)
(a) (b)
Fig. 45 (a) Phase diagram of PSYCCO - PbO: a one layer system

(b) Phase diagram of PSYCCO - (PbO+NaC1): a two layer system
In ITGA experiments, the platinum wire is immersed as a probe into two layers to their
determine crystallization temperatures separately. Figure 46 s h o w s ITGA curves in two
layers. Figure 46(a) shows the crystallization temperature in the upper layer for the
starting composition of PSYCCO : PbO : NaC1 = 1 : 8 : 12 (Mol.) at various cooling rates.
The crystallization temperature ~ 7 7 0 °C is nearly unchanged with different cooling rates,
which suggests a very low supercooling in the liquid layer. The temperature is a little
lower than the melting point of pure NaCI because of the presence of s o m e other ions
such as Pb +2 and Cu +2. Figure 46(b) shows the TG curve of the lower layer for the
starting composition of SrCO 3 : Y 2 0 3 : CaCO 3 : CuO : PbO = 2 : 0 . 2 5 : 0.5 : 3 : 4 : 2 0
(Mol.). The crystallization temperature of 869 °C for such a high concentration system is
still lower than the decomposition temperature of PSYCCO.
Crystal growth has been carried out by spontaneous nucleation from the immiscible flux
system by using studies of phase diagrams and crystallization data. Figure 47 shows a
single crystal of PSYCCO with a size of 4x4xl mm 3, which was grown from the starting
composition of SrCO 3 : Y 2 0 3 : CaCO 3 : CuO : PbO : NaC1 = 2 : (i.25 : 0.5 : 3 : 4 : 30
(Mol.) in air without a precursor. As-grown crystals were superconducting with T c = 2 5 ~ 8 0
K. Unlike Y B C O no contamination by Pt from the crucible was detected as the growth
temperature is lower than that of YBCO. After cooling to room temperature, the top layer
64 Chen Changkang
of NaCI was easily removed mechanically and undamaged PSYCCO crystals were
separated from the bottom flux layer by water in an ultrasonic bath. The crystal was
investigated for the first :~rne by STM [154]. The observations provided evidence of a
screw dislocation growth mechanism. Ultrafine twin boundaries, which may serve as a
candidates for flux pinning centres, were found. The dislocation density was estimated to
be an order of magnitude lower than that of Y-123, Bi-2212 crystals or Y-123 thin films.
The crystal growth of PSYCCO from PbO-KC1 [ 175] has similar characteristics.
~o
Fig. 46 (a) ITGA curve in the top-layer (b) ITGA curve in the bottom-layer
ImPm'" ""slim
Ira' |mmnm
Im immm
Ill |Ulm
Iron,,.._ ..... .... _,,mmmmm
•
mmmm
- - In I - - IIII
mmmmm
-- -
Fig. 47. Single crystal of PSYCCO grown from the immiscible flux system (ram scale)
(2) Top-seeding technique
The top-seeding technique is the major route for the growth of large single crystals,
especially for crystals with a thickness more than millimetres in the c-direction. Suitable
composition and temperature fields are vital for crystal growth. Stirring the solution by
rotation of the seed or crucible for example by the accelerated crucible rotation technique
(ACRT)] can speed up the mass transport. However this should be controlled in order to
achieve a stable state at the interface. As usual, the seeding crystal grows along the
liquidus line between the peritectic and eutectic temperatures. The real YBCO growth
system is not a simple binary system, but a complex system possibly with a wide
coexistence range of 123 + 211 + L.
Table 20. Flux growth of YBCO crystals by top-seeding
No. 1 2 3 4
Composition(Mol.):
( 1/6)YBCO 0.17 0.30 0.17 0.25
Flux 0.83 0.70 0.83 0.75
BaO 0.28 0.28 0.30 0.375
CuO 0.72 0.72 0.70 0.625
Crucible Al203 Y203 AI203 Y203:Ti

seed Alumina YBCO YBCO Sm- 123
Nutrition None None YBCO Y-211
Rotation (rpm) Seed, 20 0 Crucible Seed, 120

Temp.grad. (°C) 7 - 10 Yes 15
Pull rate (mm/day) 1 - 3 0 0 4.8
Cooling range (°C) 980-960 990 for 10 days 930 for days 1005-1000
990-970
Cooling rate (°Ch- 1) O. 1-0.2 0.5 0 0
Crystal
Max. size (mm) 6x3xl 6.4 thick 8x6x5 8x8
Tc (K) ~ 70 91 90 89
Reference 155 156 157 158

66 Chen Changkang
Several typical methods for top-seeding have been attempted and they are summarized in
Table 20 from No.l to No.4 [155-158] and with the four pesudo-binary phase diagrams
in figs. 48 (1) to (4) [59, 76, 131 ] that have the same flux compositions as Table 20 No. 1
to No.4 respectively.
T(C°)
~020
~000
o
~80 (Y,P, IBCO*L

0% I iOC !
960
Y~BoCuO~ ~ t t ;,
9~0
(Y,F, ]BCO* C u O ' L
920
SO0
IY, P~ )BCO-B~CuO~*CuO'L I . . . . . .
'~' YBoeOJ~37, ,k _ '~tS'/-~
860
8aO
820
eoo ~, 2'0 40 60 60 100

~z15'r~oz C u 3 0 r r ¢aol % 1 / ~ 3Ce,~0
(3 J
~10() F L
Y~zCu.O~-, . t_-
1200
L
i 2114 o 2"L
l|O~
YI3c,eCu,O , - ~.'C~-32 •
t~16 - -
~---- ' JII*L
9oo %o'-., ~o / " II c.c. L
[
_ 7
I (o3 8O 6,q ~ 20 o C, 20 "-2 4' #.. :I'C
"~Ba:C u ~(9~ toni'* 7Bz( uO2-1 i CuO
(2) (4-)
Fig. 48. Growth processes related to phase diagrams used for top-seeding
The first (No. 1) experiment has low solute concentration with a eutectic flux system in an
alumina crucible. Crystal growth by slow cooling follows the liquidus line for YBCO as
the primary phase. A recrystallized AtzO 3 tube with a small bore was used as a source of
nucleation. The high temperature solution was pulled up by capillary action. As a result, a
duster of crystals was grown on the end of the tube, but the crystals were contaminated
by A1 and showed a relatively low Tc. In experiment No.2, a high concentration system
lifts the growth temperature into peritectic range with coexistence of 123+211 +L phases.
A YBCO crystal as a seed was introduced into solution. Under the influence of a
temperature gradient the seed was grown by the process of Ostwald ripening at 990 °C
for ten days. Then the crystal growth was continued by slow cooling along the liqidus line
of 123+L. Crystals grown in yttria crucibles always exhibit a high superconducting
transition temperature after annealing in oxygen. A low solute concentration flux system
with a composition near to the eutectic was used in batch No.3. Nutrient for crystal
growth was made firstly by crystallization of YBCO on the crucible base by cooling down
to 930 °C under a negative temperature gradient (cooler on bottom than on melt surface).
Then a YBCO seed was put in the solution and the temperature gradient was reversed.
Under the temperature gradient the nutrient was gradually dissolved and crystal growth
was driven by mass-transport. The lower growth temperature results in less corrosion of
the A120 3 crucible [104]. The higher BaO concentration, above that of the eutectic
composition, decreases the distribution coefficient of aluminium between the solid and
liquid phases, which may responsible to the high Tc. Experiment No.4 utilized a growth
system with an intermediate concentration of solute and solvent composition far away
from the eutectic point, (3BaO+5CuO)/8. It was selected by extrapolation from the "211"
-"123" to the side edge of BaO-CuO in the ternary system of YO 1.5 - BaO - CuO (fig.
36). Therefore the growth temperature is quite high, located around the peritectic range
like experiment No. 2. Instead of using YBCO as a seed in experiment No. 2 and YBCO
precipitation as nutrient in experiment No. 3, Sm-123 crystal was introduced in the
solution surface and 211 powder was placed at the bottom of the crucible. Under a
temperature gradient of 15 °C, the seed was grown in the liquidus range of 123+L by
mass -transport and the nutrient was gradually dissolved in the peritectic range:
[21 I] (S) ......... > (21 I) (L) (33)
(211) (L) + 3 (010) (L) + 5 (001) (L) . . . . . . . . . > 2 (123) (L) (34)
(123) (L) . . . . . . . . . . > [123] (S) growth on the seed (35)
where (010) and (001) represent BaO and CuO respectively and (S) and (L) the solid and
liquid. During crystal growth, the composition and the temperature are kept unchanged by
using a flux composition of BaO:CuO = 3:5 (Mol.). A high rotation rate of the seed crystal
68 Chen Changkang
was applied to speed up the growth rate ( The maximum growth rate was 4.8 mm/day in
air [ 176] ) and to drive the floating particles away from the growing crystal. The growth
rate depends significantly on the atmosphere [177]. Under one atmosphere of oxygen, the
growth rate could be 1.5-2.5 times higher than that in air [180]. It was reported [159]
recently that large YBCO single crystals with a size of more than one cm could be grown
along the c-axis reproducibly. An annealed sample exhibited Tc ~ 93.6 K with a transition
width of 0.9 K.
(3) Crystal growth of Lil+xTi2_xO 4 and electric crystallization
Lil+xTi2_xO4 ( 0 < x <_ 1/3 ) is a non-copper compound with spinal structure. When x=0,
LiTi204 was discovered in 1973 as being a high Tc oxide superconductor with Tc~12 K,
in which Titanium has two valences of Ti+3 and Ti +4. When x = 1/3, Li4/3Ti5/304 is an
insulating compound with the single valency of Ti +4. Figure 49 shows the binary phase
diagram of L i 2 0 - TiO 2 [160]. Li4/3Ti5/304 appears in the diagram with its
decomposition temperature at 1015 °C in the solid state. No liquidus range for crystal
growth exists by using L i 2 0 as flux alone. The author tried a binary flux system for the
crystal growth of Li4/3Ti5/304. Figure 50 (a) [163] shows the system of LiF - LiBO2, in
which the flux composition was selected near to the eutectic point of 0.36LiF+0.64LiBO 2.
The mixture of Li2CO 3 : TiO 2 : T i 2 0 3 : LiF : B 2 0 3 = 3.38 : 1 : 0.5 : 3.24 : 2.88 (Mol.)
was placed in a molybdenum crucible and heated up to 1000 °C in an Ar atmosphere.
After soaking for 5 hs, it was cooled to 750 °C at a rate of 1 °Ch-1. Then the flux was
poured out and crystals of Li4/3Ti5/304 with a size up to 2xl.5xl mm 3 were left in the
crucible. However, no superconducting LiTi204 phase has been found in the phase
diagram of L i 2 0 - TiO 2 (Fig. 49) because Ti +4 is more stable than Ti +3. Different
methods have been used for the flux growth of LiTi204. One is the chemical reaction of
TiO 2 with Li metal in a sealed ion vessel [161]:
Li + 2TiO 2 . . . . . . > LiTi204 (36)
in which the crystal is grown from high temperature solution with liquid lithium in the
presence of a high vapour pressure of lithium. Another [162] is the use of a chemical
reducing agent, such as Zr metal, as in the reaction:
Zr + 4LiBO 2 + 4TiO 2 + 0.502 . . . . . . > 2LiTI204 + L i 2 B 4 0 7 + ZrO 2 (37)
with excess of LiBO 2 as flux in an ambient atmosphere. The third one is electric
crystallization in molten salt system, in which the reduction takes place at the cathode:
Ti +4 + e . . . . . . > Ti +3 (38)
"-1 ~ ~ ~ I--T~
1900I
Liquid
i'
1700[~ IJ_
I 1547o i i /
~500
/ / \ ',.. ,,/
~_,~ ii -- -'[l,_
LI2Ti(~Xl \ !
\
k
1
I
SS \
-.
%\ /
~3OO
- \ / I 1295i 1~ [ ] { 7 6 7 . )
" ~'12o0o 1(477,) / L.ilb
', ' - T - ~ ~', / <'"-'tl L~T~oT,~
1100
900
i G
70( .... 4, a :t
i
q -I
20 40 60 80 100
Li20 1102
Fig. 4 9 . T h e s y s t e m of L i 2 0 - T i O 2
, 1 , ' l - -
845* 845*
~,~ Liquid
.#,>d~8~° / 8~2,- 6
~,
~
710° V
7S5° ~
_ 6eel
762" ~
-
II00 [
I0O0
60C #" 585 ° 900
?. . . . . . . .
tsf
_~_~£ ..........
8ooi ~
700 -
6OO
m
40C "~ 500
4O0
L3
_ ,~ 500 i._ L
2O 40
-~ 6'6 -8' -0- - - ~ 0 Io 20 &o 4c 50 60 20 80 90100
LiBO~, LiF Li-eO BzOs N(lzO5 z:,Oz
Fig. 5 0 . T h e s y s t e m (a) L i B O 2 - LiF (b} L i 2 0 " B 2 0 3 - Na20"B203

70 Chen Changkang
Electrolysis as in slow cooling provides another driving force for producing

supersaturation for crystal growth, but here it is controlled by varying the electric current.
The size of crystals grown using electrolysis are normally larger than those by either of
the other two methods. Crystals of LiTi204 are preferentially nucleated and grown on the
electrode. LiBO 2 [162], NaBO2+LiBO2+Na F [42] were used as flux for the
electro-crystallization. Figure 50(b) shows the phase diagram of LiBO 2 - NaBO 2, which
is a simple eutectic system. Obviously, the dissolution temperature is decreased by using
the more complex flux of LiBO2+NaBO 2 compared to the single flux of LiBO 2. Also the
viscosity is lowered by adding NaF which breaks the network structure in the solution.
This results in one order of magnitude larger size of single crystal (1-2 mm in size for a
NaBO2+LiBO2+NaF flux in a graphite crucible and 0.1 mm in size for a LiBO 2 flux in
an alumina crucible). Similarly by using the flux of Li20+LiF+Na2B407, we obtained
crystals in a molybdenum crucible under an Ar atmosphere. AC susceptibility
measurement indicates the as-grown crystals of LiTi204 are superconducting with Tc= 12
K and a transition width of 0.5 K as shown in fig. 51.
4.~
°r..~
~6
+
°r,.~
°r..~ t'
+
.~..~
I I I I I I I I
4 6 8 10 12 14 16 18 20 T(K)
Fig. 51. AC susceptibility of a LiTi204 single crystal
(4) Euteetie temperature and crystal separation
The separation of crystals from a flux is the last important stage of crystal growth. Since
no suitable inorganic or organic solvents have been found to dissolve the flux without
damaging crystals for the most of oxide superconductors, the crystal has to be separated
from the flux above the eutectic temperature. The selection of the termination temperature
of crystal growth depends on the phase diagram of the growth system. On one hand the
temperature should be close to the eutectic temperature in order to have a high yield of
crystals. On the other hand, the separation is not sufficient due to the high solution
viscosity near to the eutectic temperature. Most of growth systems have multiple
compounds. Their pesudo-binary phase diagrams do not obey the binary phase rule. For
the crystal growth of 123 phase from a self-flux system, the coexistence of YBCO with
CuO and BaCuO 2 can appear in the solid-liquid range and the lowest eutectic
temperature should be estimated by experiments.
Many obvious factors influence the eutectic temperature Te of the growth system.
(1) Binary flux systems are often used for crystal growth, in which Te depends on the flux
composition. For the dopant superconductors, various doping concentrations might change
Te. Figure 52 (a) and (b) show phase diagrams of Nd2CuO 4 - CuO and NCCO - CuO in
air [54], which were estimated by DTA and by direct observation using a Pt sheet. The
two systems Nd2CuO 4 - CuO and Ndl.85Ce0.15CuO4 - CuO have nearly the same Te =
1040 °C + 10 °C. Figure 53 shows phase diagrams for N d 2 0 3 - CuO and
(0.85Nd203-O.3CeO2) - CuO, which were determined by visual inspection using an
A120 3 crucible [56]. Te in (Nd 1.7Ce0.3)CuO4 - CuO is much higher than in Nd2CuO 4 -
CuO. It has been reported that the solid solubility of Ce in NCCO lies in the range 0 < x
< 0.18 [93].
(2) As described above, additives were used to improve the size and the quality of YBCO
crystals. They have relative low melting points and reduce the eutectic temperatures of the
growth systems. As the concentration of BaF 2 additive increases to 5 wt.% in a yttria
crucible, the termination temperature for the crystal growth of YBCO from the eutectic
flux drops to 850 °C.
(3) The corrosion of crucibles can also move the eutectic temperature up or down. The
formation of BaZrO 3 by reaction between flux and a zirconia crucible increases the
solution viscosity rapidly, so that the solidification temperature of the YBCO growth
system gets higher and the time for crystal growth is shortened. The dissolution of A1203
or MgO from crucibles plays a role as flux and lowers Te, while the dissolved Y 2 0 3 from
a yttria crucible is a source of the starting oxide for YBCO crystal growth and enhances
Te. The extent of reaction or dissolution is time and temperature dependent and affects Te
correspondingly.
(4) Atmosphere also affects the eutectic temperature. Te in self-fluxe systems for crystal
growth of YBCO and LSCO are lower in oxygen than in a reducing atmosphere. In order
to lower Te, crystal growth might be carried out under a N 2 or an Ar atmosphere. After
the separation of crystals from a flux, one can switch to an oxygen atmosphere for
oxidation to create superconductivity.
72 Chen Changkang
-~" ' 1 I [. . . .
(c)
1400 Nd203 + Liq. Liquid 1400 NdtBsCe.1503 + Liq Liquid
L) CuO
o \\ + 1290 ° ~ I0 ',\
o 1240 ° ± 10 ~ Cu20
1200 ~- \\\ t
1200
Nd2CuO 4 4- L l q
E
o
1040 ° _~10 ~11040°~10 ~,,' ]

1000 ~ooo;I <- NdIs",~Ce!I~C;O~.~ Cut,
•
J .............................. - -
0 50 60 70 80 90 109 0 50 60 70 89 90 f00
Nd203 CuO (mol %) NdlesCe~sOs CuC me %',
Fig. 52. S y s t e m s (a) N d 2 C u O 4 - CuO (b) N d l . 8 5 C e 0 . 1 5 C u O 4 - CuO
rO I ! !
Nd Ce CuO4 (0.85N~20
~- O.3CeO2)-CuOo
2- X Nd 203-Cu 0 •
,o
0
0
I 1 _ I
o 2o 40 60 80 100
( Nd016* Ce02) (%) Cu©
Fig. 53. S y s t e m s (a) N d 2 0 3 - CuO and (b) ( 0 . 8 5 N d 2 0 3 - 0 . 3 C e O 2 ) - CuO

Figure 54 shows schematically six methods for separating crystals: (a) Crystals are grown
in cavities of a partially melting system. The solidified melt is broken mechanically at
room temperature and crystals are picked out with some damage; (b) High temperature
solution is drained through holes bored in the bottom of the crucible. It may be difficult to
completely separate crystals from the residual solution because of the high viscosity of the
solution near Te. Also, crystals grow in the drainage holes and block them up; (c) For
top-seeding ITGA or pulling techniques, the rod or wire with crystals are pulled out off
the solution above Te. But crystals that grow spontaneously on the wall or base remain
along with the residual solution; (d) A porous ceramic bar dips into and absorbs the
solution. However the dipping process may damage or contaminate some crystals and
some residue flux may remain in the crucible; (e) This method is called "self-decanting"
and has been frequently used for crystal growth in zirconia crucibles. An appropriate
temperature gradient can be applied to induce creeping; (f) This method is simplest and
most effective. The crucible is upturned for a few hours in high temperature furnace in
order to decant the solution into another container or brick. Some crystals poured out
with flux into the brick can even be separated after the absorption of the flux by the brick.
(=) (d) picx=
0n
(e)
AT
(f)
~
Fig. 54. Six methods for separating YBCO crystals from the flux
Single crystals
Resid al solution
74 Chen Changkang
6. P h a s e d i a g r a m a n d s u p e r c o n d u c t i v i t y
The superconductivity of oxide materials is a highly sensitive function of its structure and
composition. The relation between structure, composition and superconductivity can be
expressed using the T-x phase diagram at low temperature.
(1) S t r u c t u r e t r a n s i t i o n a n d s u p e r c o n d u c t i v i t y in Y B C O
After crystal growth YBCO undergoes a structural transition from the tetragonal phase
(P4/mmm) to the orthorhombic phase (Pmmm) during cooling to room temperature,
accompanied by incorporation of oxygen from the atmosphere. There are at least three
phases with different structural features and their transitions have a continuous nature.
They are (a) orth-I structure for YBa2Cu307, (b) orth-II structure in YBa2Cu306. 5 and
(c) tetragonal structure for YBa2Cu306, which are shown in fig. 55 (a-c). Their
distinguishable differences are in the Cu-O chain layer: Cu-O-Cu chain for orth-I (Tc~90
K), alternating Cu-O-Cu and Cu-Cu chains for orth-II (Tc~60 K) and only a Cu-Cu chains
for T-structure (semiconductor).
• . .......:
- '"-U ~ U _y
YtBa2Cu~O 7 Tc--90K YIBa2Cu306 To---0
a b c
Fig. 55. Structure of YBCO: (a) Orth-I phase ( Tc~90 K )

(b) Orth-II phase ( Tc~ 60 K )
(c) Tetragonal (T) phase ( Tc~ 0 K )
Figure. 56 (a) and (b) show T-x phase diagrams. (a) expresses the variation of Tc with
oxygen loss for YBCO. (b) is the temperature dependence of oxygen loss in air. The
oxygen content remains nearly unchanged up to ~400 °C when the first Tc plateau
appears for the sample with the orth-I structure. Then the oxygen is gradually lost from
the Cu-O chain by the first decomposition reaction during heating:
YBa2Cu30 7 . . . . . . . . > YBa2Cu307_y + (y/2) 0 2 (39)
At ~600 °C the orth-II phase forms with the second superconducting transition plateau
(Tc~60 K). As the temperature reaches ~800 °C the oxygen loss exceeds 0.6 and Tc takes
a sharp drop with the formation of a non-superconducting tetragonal phase. The
transformation is an order-disorder structure transition, but also a decomposition chemical
reaction occurs as well, accompanied by a small thermal effect [67]. At 1000 °C, the
whole tetragonal frame collapses as a result of the second decomposition reaction (5).
1000
I00 1 I I
a
800
80 Wc
600
60
l
40 I '~00
i
4 Odh-- I ONh-II 7etrc
i
20 YBa 2Cu307_ x I 200

t
I
I ! 2o. I 0 I I I I
0 0.2 0.4 0.6 0.8 1C
0.0 0.2 0.4 0.6 0.8 1.0
y ,n YBo2Cu~OT_y
Fig. 56. (a) The variation of Tc with oxygen loss

(b) The temperature dependence of oxygen loss in air
The structural transition from tetragonal to orthorhombic phases is controlled by oxygen

diffusion during cooling. Investigation has shown that a large anisotropy occurs in the
magnitude of oxygen diffusion in single crystals [ 166]: The diffusion coefficient Dab in the
ab-plane is far greater than Dc in the c-direction and Db>>Da, which implies that the
main route for oxygen diffusion is along Cu-O chains. The structural transition in single
crystals induces twin structure caused by crystallographic strain with an increase in the
76 Chen Changkang
b-axis and decreases in the a- and c-axes. Figure 57 shows the twin pattern in a large
YBCO crystal lxl cm 2 in size, which was as-grown in a yttria crucible. The picture
confirms that predominant oxygen diffusion takes place in the ab plane. Because of the
large size of the ab plane and the limitation of diffusion time for the as-grown crystal,
twin patterns appear only around four sides with no twin domain in the central area.
Fig. 57. Twin patterns on a large YBCO crystal
The diffusion process is influenced by the quality of the crystals. It is easy to oxidise high
quali~ crystals. Experiments show that a low level of contamination in the Cu-O chain by
A1+3 hinders the diffusion process even if it does not greatly depress Tc after 0 2 anneal.
Figure 58 shows AC susceptibility measurements on two YBCO crystals before and after
oxidation: (a) The crystal with a low level of A1 contamination exhibits two
superconducting transition steps (Tc~90 K for orth-I phase and Tc~60 K for orth-II
phase) in the as-grown state. After annealing in 0 2 for two days, only a single transition
of Tc~91 K appears. (b) The other crystal free from A1 contamination only exhibits a
single transition at Tc~92 K in the as-grown state, whereas only a small increase in Tc to
93 K occurs after annealing. Oxidation under a suitable high pressure of oxygen speeds
up the diffusion rate. However we should keep in mind that Y-123 is unstable at high
pressure and it may decompose into the Y-247 or the Y-124 phase. Figure 59 (a, b) is an
AC susceptibility measurement for a YBCO crystal (a) after annealing in 1 bar of oxygen
for two days and (b) after annealing in 40 bar of oxygen for 2 days. The appearance of
High T C Oxide Superconductors 77
two steps, at 92.9 K and 91.7 K respectively, is probably the result of the partial
decomposition reaction, in which the Y - 1 2 3 phase decomposes into the Y - 2 4 7 phase.
I I I I I I I I I [ _
_ 50K _~..__ : 100K
.
G" -
J5 -
~J
35
o--
(h
eJ
tO
(n
co
I I I I t l I I I L . ~ I I I I I I I I I ,. I__
E 50K ~2 1 00K 50K • 100K
ta} [b]
Fig. 58. AC susceptibility measurement showing two superconducting step transitions
90 92 94 96 98
I
Temperature (K)
..d
°,..~
,..¢2
C-,-
r~ 90 92 94 96 98
I
Temperat, ure (K)
Fig. 59. AC susceptibility measurement showing partial decomposition of Y - 1 2 3

78 Chen Changkang
(2) Crystal chemistry and superconductivity of Bi-2201
In the system Bi2Sr2Can_lCUnO2(n+2) there are three superconducting compounds with

one , two and three Cu-O layers (n=l, 2 and 3) respectively, which are sandwiched
between the Bi-O layer in the structure. Reference to the phase diagram shows that it is
very difficult to grow single crystals Bi-2223 with the highest Tc. Most of studies have
concentrated on the crystal growth of the two layer compound Bi-2212 with Tc=80-90 K.
Single crystals of Bi-220.1 with a single Cu-O layer structure offer an opportunity to study
the normal state properties down to a significantly low temperature (Tc<20 K).
In the ternary system of SrO-Bi203-CuO [167] shown in fig. 60, there exist two kinds of
Bi-2201 phases. One has small deviation from the chemical stoichiometry of
BiO1.5:SrO:CuO = 2:2:1. The other is a solid solution having a greater deviation from
stoichiometry. Measurements by magnetic susceptibility, electrical resistivity and electron
microprobe analysis have been carried out on the crystal grown from the KC1-KF flux
system [106, 168]. Table 21 summarizes the crystal growth and characterisation of
Bi-2201 in four batches. Average compositions of all batches deviate from chemical
stoichiometry, Bi atoms in the formula are in excess of 2 and Sr atoms are less than 2.
The four crystal compositions are marked in the ternary diagram of fig. 60. The crystal of
batch No.I, which is superconducting, locates near to the stoichiometric composition. But
the crystals of batches Nos.2-4 which do not exhibit superconductivity and are located in
the region of the solid solution phase. Figure 61 shows the linear dependence of Bi on Sr
in crystals, which suggests that Sr atoms are partially substituted by Bi. In the structure of
Bi-2201 the Sr atom is close to the Cu-O plane, which is considered to be the conductive
layer. The substitution of Sr by Bi, with a larger radius and higher charge than Sr, greatly
affects the superconducting properties. Only the crystal with a composition close to
stoichiometry is superconducting.
Table 21. Crystal growth and characterization of Bi-2201
Batch no. Starting composition Flux Av. composition SC*?

B i 2 0 3 SrCO 3 CuO Flux* evaporation of cryst. (at.)
( Mol. ) ( Wt.% ) ( wt.% ) Bi Sr
1 1 4 4 50 5.0 2.13 1.96 Yes

2 1 6 4 50 7.2 2.26 1.79 No
3 1 4 4 50 13.3 2.34 1.60 No
4 1 4 4 50 12.8 2.23 1.66 No
Flux*: 0.46KCl+O.54KF. SC*?: Superconducting ?

SrO
S B /,50°/o3rO
~,, /~.~:L
.~ ,,/ \ . °e~
$2B2 LOX~ $3B2 XTxLS O
# ~ /?k\. ~ ,. \ ...-s:~.s
SB2- , , , - ~ -
Rhss~
1/2(Bi2031 tool°l,CuO CuO
Fig. 60. The location of the BSCO crystal compositions in the ternary diagram of
Bi203-SrO-CuO
2..5
2.~ •
2.3
rr~
2,2
2.1
2.0 , , , l l I l I
1.6 1.7 1.8 1.9 2.0
Sr
Fig. 61. The Bi and Sr compositions of BSCO crystals

80 Chen Changkang
As-grown crystals of Bi2+xSr2_xCUO6_y are significantly oxygen deficient [168]. An

annealing atmosphere makes great influence to the superconductivity. Tc of one as-grown
crystal decreased from 5.5 K to bzVow 4.2 K after annealing in air at 400*(2 for 15 hours,
and then increased to 12.5 K by further treatment in vacuum for 8 hours. It is said that
the solid solution phase always forms first during sintering heat treatment and then the
composition gradually moves toward stoichiometry. Deficiencies of oxygen perhaps speed
up this process [167].
(3) Substitution and superconaductivity in BPBO crystals
BaPbl_xBixO3 (BPBO) (x = 0.05 - 0.35) is a non-cuprate oxide superconductor with no

rare-earth elements. Figure 62 shows the partial phase diagram of BaO - PbO [44]. Since
single crystals of BPBO melt incongruently they are grown either by hydrothermal
synthesis [17] or by the flux method. BaPbO 3 decomposes into Ba2PbO 4 at 1080 °C. By
using flux of PbO, one can use the existence of a liquidus range between 1080 °C, 67
mol.% PbO and 850 °C, 96 mol.% PbO for the crystal growth. BaPbO 3 exhibits metallic
behaviour electrically and becomes superconducting after partial substitution of Pb by Bi.
So the dopant of B i 2 0 3 is a necessary solute composition. Bi20 3 which has a relatively
low melting point of 860 °C is also a good solvent for crystal growth, since it reduces the
crystallization temperature and the eutectic temperature in the growth system. The
crystallisation temperature (Tk) and eutectic temperature (Te) in the flux system with
different concentrations of Bi20 3 are listed in Table 22. As the B i 2 0 3 concentration
increases, Tk decreases. The crystallisation temperature of batch No. 3 is higher than that
of No. 2 because a seed was used, which greatly reduced the supercooling.
(d)
- 1280°
.--..
1200
L) Q.
_.2280°
No
~. loo0
E
eo
8OO
"f 1 1 '
30 40 60 80 100
BaCO3 mol % PbO
Fig. 62. Phase diagram of Ba2PbO 4 - PbO

Table 22. Dependence of crystallization and eutectic temperatures on the
concentration of Bi20 3 in 2PbO-BaO-PbO2-yBi20 3
No. y Tk (°C) Te (°C) Method
1 0 1070 850 Direct observation

2 0.33 987 ITGA on wire
3 0.40 996 835* ITGA on seed
4 0.60 944 ITGA on seed
* exothermal effect recorded by the thermocouple underneath the ITGA crucible
Table 23 shows crystal growth data for BaPbO 3 and BPBO from the same system with
various dopant concentrations. Rectangular crystals of BaPbO 3 grow on the base of the
crucible. Nucleation appears predominantly on the melt surface of the growth system with
Bi20 3. Platelet crystals always cover the surface of the solution by evaporation of solvent
at high temperature. However by using large amounts of starting materials, crystals with
pyramid and rectangular habit appear on the bottom or the wall of the Pt crucible.
Table 23. Crystal growth of BPBO from 2PbO-BaO-PbO2-yBi203 with

various concentration of Bi20 3
No. y Crystal growth Crystal

Method Cooling range Cooling rate Habit Growth site
( °C ) (°Ch -1 )
! 0 Sp* 1080-879 0.5-2 Rectangular Base

2 0.33 ITGA 987-967 0.5-0.1 Platelet Surface
3 0.40 TGA-seed 996-852 0.4 Platelet Surface
4 0.50 Sp* 986-860 l Platelet Surface
Pyramidal Base, wall
Rectangular Base, wall
5 0.60 ITGA 944-855 5-0.3 Platelet Surface
*Sp: Spontaneous nucleation
Table 24 shows chemical compositions and superconducting transitions for these three
kinds of crystals from the same batch. Although they are grown in different positions of
the crucible, the chemical composition and Tc are nearly the same, indicating high
82 Chen Changkang
homogeneity both in the solution and crystals. The content of Bi in the platelet crystal on
the melt surface is slightly higher than that in the pyramidal or rectangular crystals inside
the solution because of the evaporation of PbO during crystal growth, so that Tc of
platelet crystals is also a little higher.
Table 24. Composition and superconductivity of BPBO crystals with different

habits in the same batch
No. Crystal habit Crystal composition Superconducting transition

Ba Pb Bi Tc (K) Width (K)
1 Platelet 1.04 0.78 0.22 11.6 1.4

2 Pyramidal 1.04 0.79 0.21 11.2 1.2
3 Rectangular 1.05 0.79 0.21 11.1 1.0
Depending on the doping concentration, the structure of BPBO can be monoclinic,

tetragonal or orthorhombic [169]. The superconductivity is also regulated by the doping
level of Bi in the crystals, which could be controlled by the concentration of Bi20 3 in the
starting materials. Table 25 shows the dependence of the crystal composition on
concentrations of Bi203 in the starting composition of BaO-2PbO-PbO 2. As Bi203 in
the starting materials increases, the doping concentration in the crystal increases with the
decrease of Pb because of the partial replacement of Pb by Bi. The atomic ratio of
Ba/(Pb+Bi) remains approximately 1: 1.
Table 25. Dependence of the crystal composition on Bi203 concentration

in the starting materials
No. Starting composition (mol.) Crystal composition (at.)

BaCO 3 PbO PbO 2 Bi203 Ba Pb Bi
1 1 2 1 0.33 1.06 0.84 0.16

2 1 2 1 0.40 1.04 0.81 O. 19
3 I 2 1 0.50 1.04 0.78 0.22
4 1 2 1 0.60 1.05 0.77 0.23
Combining data from the author's laboratory [170] and ref. [171], a schematic T-XBi
phase diagram for single crystal BPBO at low temperature is drawn in fig. 63, which
generally agrees with the data for the sintered material. For the crystal of BaPbO 3 (XBi =
0), the dependence of electrical resistance on the temperature shows metallic character.
As the crystal is doped with Bi ions, it becomes superconducting. The Tc firstly increases
with the enhancement of doping concentration and reaches the highest Tc~ 13 K at XBi ~
0.25. Single crystals of Bal.05Pb0.77Bi0.230 3 with Tc (on set) = 12.6 K and a the
transition width within 1 K have been obtained. When XBi increases further, Tc falls and
finally enters the semiconducting region.
70
/
/
60 /
/
/
50 /
/
l+O Metallic / Semiconducting
Bi TM / Bi m ÷ Bi v
T~(K) /
30 /
/
/
20
10 j 00~ ~O
/ 041 \
% Superconducting
0 -- "~ I I ~. I I I
0 0.2 0.t, 0.6 0.8 1.0

Ba Pb 0 3 XBi Ba Bi 0 3
Fig. 63. Schematic phase diagram for BPBO crystals
(4) Cross substitution and superconductivity in CLBCO crystals
(CaLaBa)Cu307_y (CLBCO) is a rare "123" oxide superconductor with a tetragonal

structure (Tc~80 K), which is free from micro-twins. There is a complex cationic
distribution between yttrium and barium sites in the "123" compound [ 172]. The variation
of ionic charge, ionic radius and coordination in the 123 structure can give rise to
complicated cross substitution in the Y and Ba sites for the Ca 2+, La +3 and Ba +2 ions.
The superconductivity is strongly dependent on the substitution state and can disappear in
some cases.
Research has been carried out on the phase diagram and crystal growth of CLBCO from
the self-eutectic flux of 0.28BaO + 0.72 CuO in platinum crucible in the author's
laboratory. Figure 64 shows the pesudo-binary phase diagram of the growth system
determined using DTA in air. Compared to the YBCO system, the decomposition
temperature of CLBCO is about 70 °C higher than YBCO and its solubility in the eutectic
84 Chen Changkang
flux is also higher. The data provides useful information for the selection of the starting
composition and the temperature program. As a result, single crystals of CLBCO with
sizes up to 7x6.5x0.4 m m 3 were grown from the starting composition of 0.2(1/6)CLBCO
0.8(0.28BaO+0.72CuO) with Ca:La = 1.25:0.75 (at.) and lwt.% BaF 2 as additive. As
with the crystal growth of YBCO, a small amount of BaF 2 additive greatly improves the
superconductivity of the crystal.
1100
/
/-
1050 /
/
CLBCO 4- L
1
1000 - /
T/[ c C] i i ~
950 ~ /
/
.. 'k/ _
9{)
85O
---- = c
I I I I I I I
8001 , ,
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
1 (7BaO + 18CuO) 1 CLBC 0
25 6
Fig. 64. Phase diagram of the growth system for CLBCO
Table 26. The dependence of Tcm of CLBCO crystals on BaF 2 concentration in the
starting materials
No. 1 2 3 4
BaF 2 ( w t . % ) 0 0.5 2.0 2.5

Tcm ( K ) No S.C ~30 43 51
/~Tc m ( K ) very weak 8 17
Table 26 shows the relation between the main superconducting transition temperature
(Tc m) and the concentration of BaF 2 in the starting composition in the Pt crucible. Tc m
for oxygenated crystals increases from non-superconducting to a Tc m of 51 K by the

addition of BaF 2. The function of the additive can again be understood by the cluster
model as described previously for the crystal growth of YBCO. The occupation of oxygen
sites in the Cu-O chain by F- depresses the contamination of the chain by Pt from the
crucible and reduces the structure distortion in the nearby Ba-O layer.
Figure 65 shows the superconductivity diagram of CLBCO crystals, indicating the
dependence of the main transition temperature on the Ca/La ratio in the starting
composition. The AC susceptibility measurement was performed on crystals after
oxidation at 600 - 400 °(2 for two days. Only crystals grown from starting materials with
a Ca content between 0.6 and 1.6 are superconducting and the Tc peak is located around
a Ca content of 1.2. CLBCO crystals always exhibit a lower superconducting transition
temperature than its sintered powder, which may be caused by the Pt contamination from
the crucible. The cross cationic substitution between Y and Ba sites exists in the
conductive layers (CuO 2 plane - Y - CuO 2 plane) and charge reservoir layer (BaO - CuO
chain - BaO), which leads to a complex relationship between composition, structure and
properties.
80
70
"i
I •
60
50-
40- / .-!
30- /'!
20 ~' 8.C I
10
0
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Ca in (Ca + La) = 2 of starting materials
Fig. 65. The dependence of Tcm of CLBCO crystals on the Ca/La ratio in the starting
materials
86 Chen Changkang
The partial cros~ substitutions of Ba sites by Nd or Pr in crystals of Nd 1+xBa2-xCu307-y

[185, 186] or Prl+xBa2_xCu307_y [187] also have a great effect on their
superconductivity. The effects depend on the atmosphe,.~. and the flux composition used
for the crystal growth [ 184]. An atmosphere of low oxygen partial pressure depresses the
substitution in Ba sites by Nd +3 and this can cause increases in Tc up to 96 K [185]. A
high Ba content in the self-flux system BaO - CuO also reduces such substitution and
improves the superconductivity. The improvement can be understood with the aid of the
quasi-ternary phase diagram of NdO1. 5 - BaO - CuO system at 1073 °C [186]. In the
phase diagram solid solutions of Ndl+xBa2_xCu307_y are connected to the liquidus
compositions through tie lines. As the Ba/Cu ratio in the liquldus approaches 0.8, the
composition of NdBCO crystals is close to stoichiometry for the 123 phase. Large single
crystals NdBCO having a Tc up to 95-96 K were successfully grown in Nd203 crucibles
either in a low pressure atmosphere or by using a high Ba content flux in air. The Pr/Ba
substitution ratio of Pr 1+xBa2 xCu307 - is also expected by tie-lines in PrOy -. BaO -
CuO ternary phase diagram [18-8]. -x
V. Conclusions
High Tc cuprate superconductors with a layer structure are incongruent melting

compounds with multiple oxide components. Due to the multiple components and a
variety of various external conditions it is very difficult to represent crystal growth with
the aid of phase diagrams. Pseudo-binary phase diagrams are simpler and more widely
used. Flux growth is the main method for obtaining single crystals of high Tc oxide
superconductors. The pesudo-binary phase diagram for the growth system belongs to the
peritectic-eutectic type.
A whole range of experimental conditions, such as slow equilibration, crucible corrosion

and various dopants, additives and atmospheres et al, can have a great influence on the
parameters of the phase diagram. Therefore, it is necessary to have in-situ determination
of some basic growth parameters, for instance the crystallization temperature and the
solute solubility in high temperature solution. ITGA is a good method for in-situ
calibration of a system. Once the key parameters have been determined crystal growth
can be carried out by top-seeding. At the termination of crystal growth the crystals can be
lifted from'the solution above the eutectic temperature. Thus the difficulty of separating
cuprate crystals from the flux is easily surmounted.
The phase diagram is the key source of knowledge in deciding how to grow high Tc oxide
superconducting materials. The investigation of the phase diagram and its application
enables one to better understand the relationship between growth conditions and
properties, structures and compositions of crystals, which lead to improve crystal size ag
well as superconductivity. The Phase diagram can also be used in providing basic
information in order to explore new materials and to predict the difficulties of crystal
growth and possible routes to overcome such problems.
Extensive research has been focused on the three most important superconducting
families, La-Sr-Cu-O, Y-Ba-Cu-O and Bi-Sr-Ca-Cu-O, which can be considered as binary,
ternary and quaternary oxide systems respectively. Narrow stable ranges of Bi-2201, 2212
and 2223 in the phase diagram occasionally cause intergrowth in the basic phase. Great
difficulty is found in growing 2223 single crystals from the self-flux system. The
superconductivity of Bi-2201 strongly depends on the Sr53i ratio. Only crystals with a
composition close to stoichiometry are superconductings. The self-flux system with
eutectic composition 0.72 CuO + 0.28 BaO is commonly used for the crystal growth of
the 123 phase due to its high reactivity and the low solubility of the phase. The
pesudo-binary diagram (YBCO - Flux) indicates a suitable starting composition and
temperature program for the crystal growth.
The author's experiments show that small amounts of BaF2 dramatically improve not only
the liquidus characteristics for crystal growth, but also the superconductivity. F- ions
reduce the A1 contamination in Cu-O chains from Alumina crucibles and increase the
structural orthorhombicity. This increases the superconducting transition temperature and
reduces the transition width. When Yttria crucibles are used, the orthorhombicity is
further increased. A suitable amount of the additive greatly reduces the transition width
and increases the activation energy for flux creep as well as the critical cau'rent density.
A cluster model has been proposed to understand the role of the additive. The surface
observation has been made using the optical microscope and nanostructural morphology
and atomically resolved images have been observed using the scanning tunnelling
microscope. The conductivities of the Cu-O chains and Cu-O planes have been
investigated with the scanning tunnelling spectroscope. Composition measurements have
been carried out using electron probe microanalysis and structural refinement has been
performed using X-ray diffraction technique. All those studies have provided significant
evidence about the relationship between the growth conditions and the structure,
composition and superconductivity of the crystals.
The author's investigation on the crystal growth of LSCO using SrF 2 as an additive has
also been very successful. The flux gown crystals exhibit the highest values of Tc and the
narrowest transition widths up to date. The improvement can again be explained using the
author's cluster model. Bond-breaking effects by F- ions increase the concentration of
growth units and reduce the solution viscosity. The improvement of superconductivity is
possibly achieved by introduction of interstitial F- between two La-O layers in the
structure, which modifies the charge carrier concentration.
A series of phase diagrams have been applied to crystal growth of YBCO, (Y,Pr)BCO,
PrBCO, NdBCO and (Nd,Ce)2CuO 4 and revealed the influence of rare earth substitution
on the parameters of the phase diagrams. The peritectic temperature of PrBCO crystals i~
88 Chen Changkang
70 °C higher than that of the powder. Like NdBCO and SmBCO, PrBCO crystals can be
used as seeds for melt-texturing growth. The growth of tetragonal superconducting
CLBCO crystals has revealed a complicated cross substitution between Ca, La and Ba.
The superconducting transition temperature depends on the Ca:La ratio in the starting
materials. The phase diagrams of PSYCCO - PbO and PSYCCO - (PbO+NaCI) show that
NaCI is not only an air-insulating layer, but also an excellent flux for crystal growth.
Large single crystals have been grown from the immiscible flux system in an ambient
atmosphere. For non-copper oxide superconductors, using external electric field
completely changes the nature of phase diagram and its products. This interesting effect
needs further detailed clarification. Superconducting crystals of BKBO and LiTi204 were
grown by electro-crystallisation although no superconducting crystals of Li4/3Ti5/304
were obtained using the normal flux method. The superconducting diagram of BaPbO 3 -
BaBiO 3 enables one to control Tc by selecting the starting composition. Single crystals a
centimetre in size have been produced. The composition analysis and susceptibility
measurement indicate the very high homogeneity of the crystals.
As is well known, the phase diagram provides thermodynamical equilibrium data. But
crystal growth is a process which is not strictly in equilibrium. Kinetic factors should then
be taken into account when phase diagrams are applied to crystal growth. For instance,
the "peritectic jump" between crystalline compositions can be avoided by mass
transportation under a temperature gradient. With aid of this technique YBCO single
crystals a cubic centimetre in size were grown by pulling using a seed. Thus a kinetic
reason for the "eutectic jump" between two kinds of crystal phase along nearby liquidus
lines can account for their appearance in a "single" composition system. This resulted in
large BaCuO2 crystals sitting on the surface of a large CuO crystal. Both CuO and
BaCuO 2 single crystals are the important parent materials for high Tc cuprate
superconductors. Another relative large single crystals CuWO 4 have also been grown by
using relevant phase diagram [ 190] in the author's laboratory [ 191].
Acknowledgement
The author is deeply grateful to Prof. J. B. Mullin for his valuable advice and many
corrections, which have greatly improved this long review paper.
The research was carried out in the Clarendon Laboratory, University of Oxford and the
author wishes to acknowledge the Engineering and Physical Science Research Council
(EPSRC) for funding the National Crystal Growth Facility for Superconducting Oxides in
the UK.
References
1. H. K. Onnes, Commun. Phys Lab., Univ. Leiden, 122b, 225 (1911 )

2. J. G. Bednorz and K. A. Muller, Z. Physik B64, 189 (1986)
3. R. J. Cava, R. B. Van Dover, B. Batlogg and E. A. Rietman, Phys. Rev. Let. 58, 408
(1987)
4. H. Takagi, S. Uchida, K. Kishio, K. Kitazawa, K. Fueki and S. Tanaka, Jpn. J. Appl.
Phys. 26, L320, (1987)
5. M. K. Wu, J. R. Ashburn, C. J. Torng, P. H. Hor, R. L. Gao, Z. N. Huang, Y. Q. Wang
and C. W. Chu, Phys. Rev. Let. 58,908 (1987)
6. H. Maeda, Y. Tanaka, M. Fukutomi and T. Asano, Jpn, J. Appl. Phys. 2,_7_,L209 (1988)
7. Z. Z. Sheng and A. M. Herman, Nature 332, 138 (1988)
8. S. N. Putilin, E. V. Antipov, O. Chmaissen and M. Marezio, Nature 362, 226 (1993)
9. D. C. Johnson, H. Prakash, W. H. Zachariasen and R. Viswanathan, Mater. Res. Bull.
8,777 (1973)
10. A. I. V. Sleight, J. L. Gillson and P. E. Bierstedt, Solid State Commun. 17, 27 (1975)
11 L. R. Matheiss, E. M. Gyorgy and D. W. Johnson Jr, Plays Rev. B37, 3745 (1988)
! 2. S. Bosi, T. Puzzer, B. Hunter, G. R. Rusell, S. L. Town and K. N. R. Taylor, J. Mat.
Sci. Letts. 8, 497 (1989)
13. K. N. R. Taylor, "Studies of high temperature superconductors", ed by A. V. Narlikar,
Nova Science Publishers, New York, Vol. 6, 1 (1990)
14. J. S. Abell, C. N. W. Darlington, A. Drake, C. A. Hollin, E. M. Forgan, D. A. O'connor
and S. D. Sutton, Physica C 162-164. 909 (1989)
15. D. Colson, A. Bertinotti, J. Hammann, J. F. Marucco, A. Pinatel, Physica C 233. 231
(1994)
16. L. W. Finger, R. M. Hazen, R. T. Down, R. L. Meng, C. W. Chu, Physica C 226, 216
(1994)
17. S. Hirano and S. Takahashi, JJAP Series I. Superconducting materials, (1988) pp. 90
18. R. A. Laudise, L. F. Schneemeyer and R. L. Barns, J. Crystal Growth 8_5, 569 (1987)
19. T. Inoue, S. Hayashi, H. Komatsu and M. Shimuzu, Jpn. J. Appl. Phys. 26, L732
(1987)
20. Y. Hidaka, Y. Enomoto, M. Suzuki, M. Oda and T. Murakaru, Jpn. J. Appl. Phys. _2__6.
L377 (1987)
21. Chen Changkang, B. E. Watts, B. M. Wanklyn and P. Thomas, Solid State Commun.
_6_6,611 (1988)
22. P. J. Picone, H. P. Jenssen and D. R. Gabbe, J. Crystal Growth 85, 576 (1987)
23. K. Oka and H. Unoki, Jpn. J. Appl. Phys. 28, L937 (1989)
24. S. Hayashi, H. Komatsu, T. Ono, K. Sasaki, Y. Koike and T. Fukase, Jpn. J. Appl.
Phys. 26, L 1197 (l 987)
90 Chen Changkang
25. D. L. Kaiser, F. Holtzberg, B. A. Scott and T. R. McGuire, Appl. Phys. Lett. 51, 1040
(1987)
26. B. M. Wanklyn, Chen Changkang, B. E. Watts, P. Haycock and F. Pratt, Solid State
Commun. 66, 441 (1988)
27. A. Ono, H. Nozaki and Y. Ishizawa, Jpn. J. Appl. Phys. 27, L340 (1988)
28. J. Karpinski, S. Rusechi, E. Kaldi and E. Jilek, J. Less-common Metals 164&165, 3
(1990)
29. J. J. Capponi, O. Chmaissem, T. Fournier, M. F. Gorius, S. Korczak, M. Marezio and
J. L. Tholence, J. Less-common Metals 164&165, 808 (1990)
30. Z. Korczak, W. Korczak, S. Kolesnik, T. Skoskiewicz and J. Igalson, Supercond. Sci.
Technol, 3, 370 (1990)
31. J. S. Xue, M. Readyk, Y. P. Lin, C. V. Stager and J. E. Greedan, Physica C 166, 29
(1990)
32. Chen Changkang, B. M. Wanklyn, E. Diegues, A. J. Cook, J. W. Hodby, A. Dabkowski
and H. Dabkowska, J. Crystal Growth 118, 101 (1992)
33. R. J. Cava, B. Batlogg, J. J. Krajiewski, L. W. Rupp, L. F. Schneemeyer, T. Siegrist, R.
B. Van Dover, P. Harsh, W. F. Peck Jr, P. K. GaUagher, S. H. Glarum, J. H. Marshall, R.
C. Farrow, J. V. Waszczak, R. Hull and P. Trevor, Nature 336.211 (1988)
34. T. Hibiya, S. Imoto, Y. Nakabagashi and M. Sakonjyu, J. Crystal Growth 102, 862
(1990)
35. J. S. Xue, M. Reedyk, A. Dabkowski, H. Dabkowska, J. E. Greedan and Chen
Changkang, J. Crystal Growth 113, 371 (1991)
36. A. Mackenzie, E. Marseglia, I. Marsden, G. Lonzarich, Chen Changkang and B. M.
Wanklyn, Physica C 162-164, 1029 (1989)
37. P. A. J de Groot, S. B. Ota, P. C. Lanchester, D. J. Godfrey, B. M. Wanklyn, Chen
Changkang and D. J. Steel, J. Phys: Condens. Matter, 1, 5817 (1989)
38. L. F. Schnnemeyer, R. B. Van Dover, S. H. Glarum, S. A. Sunshine, R. M. Fleming, B.
Batlogy, T. Siegrist, J. H. Marshll, J. V. Waszczak and L. M. Rupp, Nature 332, 422
(1988)
39. H. Takeshi, A. Katsumi, I. Shozo, N. Keikichi and O. Keiichi, Jpn. J. Appl. Phys. 22,
L2055 (1988)
40. P. Strebel and T. Fournier, J. Less-common Metals, 164&165. 519 (1990)
41. T. Kotani, T. Kaneko, H. Takei and K. Tada, Jpn. J. Appl. Phys. 28, L1378 (1989)
42. O. Ourmeyer, T. P. Kappler, A. Derory, M. Drillon and J. J. Capponi, So~d State
Commun. 7__44 621 (1990)
43. J. Akimoto, Y. Gotoh, K. Kawaguchi and Y. Oosawa, J. Solid State Chem. 96, 446
(1992)
45. A. Katsui, Jpn. J. Appl. Phys. 2_1, 553 (1982)
46. B. M. Wanklyn, Chen Changkang, J. W. Hodby and C. J Salter, Less-common Metals,

164 &165,926 (1990)
47. A. Katsui and M. Suzuki, Jpn. J. Appl. Phys. 21, 157 (1982)
48. J. P. Wignacourt, J. S. Swinnea, H. Steifiuk and J. B. Goodenough, Appl. Phys. Lett.
53, 1753 (1988)
49. M. L. Norton, Mat. Res. Bull. 24, 1391 (1989)
50. H. R. Shanks, J. Crystal Growth 13/14,433 (1972)
51. P. Caillet, Bull. Soc. Chim. Fr. 12, 4752 (1967)
52. J. Karpinski, K. Conder, Ch. Kruger, H. Schwer, I. Mangelschots, E. Jilek, E. Kaldis
(ed.), Materials and crystallographic aspects of high Tc superconductivity, 555-583,
Khewer Academic Publisher, Netherlands (1994)
53. J. Karpinski, J. Beeli, E. Kaldis, A. Wisard and E. Jilek, Physica C 153,830 (1988)
56. S. Pinol, J. Fontcuberta, C. MiravitUes and D. McK. Paul, Physica C 165, 265 (1990)
57. H. Murakami, T. Suga, T. Noda, Y. Shichara, S. Tanaka, Jpn. J. Appl. Phys. 29,
2720(1990)
58. Zhang Jishan, Jiang Xiaoping, Huang Jianquo, Jiang Ming, Yu Yang, Zhang Zhiya,
Ge Yunlong, Tang Yajun, Qiao Guiwen and Hu Zhuangqi, Supercond. Sci. Technol. 1,
107 (1988)
59. K. Oka, K. Nakane, M. Ito, M. Salto and H. Unoki, Jpn. J. Appl. Phys. 27, L1065
(1988)
60. A. Erb, T. Traulsen, G. Muller-Vogt, J. Crystal Growth 137,487 (1994)
61 .R.L. Meng, L. Beauvais, X. N. Zhang, Z. J. Huang, Y. Y. Sun, Y. Y. Xue and C. W.
Chu, Physica C 216. 21 (1993)
62. S. Adachi, A. Tokiwa-Yamamoto, M. Itoh, K. Isawa and H. Yamauchi, Physica C214,
313 (1993)
63. M. Paran Thaman, Physica C 222, 7 (1994)
64. A. B. Bykov, L. N. Demianets, I. P. Zibrov, G. V. Kanunnikov, O. K. Melnikov and S.
M. Stishov, J. Crystal Growth 91,302 (1988)
65. A. M. M. Gadalla, W. F. Ford and J. White, Trans. Brit. Ceram. Soc. 62, 57 (1963)
67. T. Sugai, G. Oya and S. Imai, Jpn. J. Appl. Phys. 28, 341 (1989)
68. M. Nevriva, E. Pollert, L. Matejkova and A. Triska, :1. Crystal Growth 9_!_,434 (1988)
69. Y. Ueda, A. Mitushima, H. Toda, N. Kojima, M. Yoshikawa and K. Kosuge, Mat. Res.
Bull. 23, 1409 (1988)
70. I. N. Belyaev and M. L. Sholokhovich, Shornik Statei Obshchei. Khim. Akad. Nauk
SSSR, 1, 141 (1983)
71. S. H. Smith and D. Elwell, J. Mater. Sci. 2, 297 (1967)
92 Chen Changkang
72. D. Elwell and H. Scheel, "Crystal growth from high temperature solution", Acad.
Press, London, New York, San Francisco, (1975) P119
73. S. H. Smith, V. V. Fenin, J. Mater. Sci. 20, 2823 (1985)
74. Chen Changkang and Zhang Gaofang, J. Chinese Silicate Society, 14, 219 (1986)
75. Chen Changkang and B. M. Wanklyn, Supercond. Sci. Technol. 3,540 (1990)
76. K. Dembinski, M. Gervais, P. Odier and J. P. Coutures, J. Less-common Metals
164&!65, 177 (1990)
77. H. Hoshizaki, S. Kawabata, N. Kawahara, H. Enami, T. Shinokara and T. Imura, Jpn.
J. Appl. Phys. 29, pp1444 (1990)
78. H. Hoshizaki, S. Kawabata, N. Kawahara, H. Enami, T. Shinokara and T. Imura, Jpn.
J. Appl. Phys. 28, pp975 (1990)
79. M. Nevriva, E. Pollert and P. Honskus, Physica C 199, 328 (1992)
80. N. Nevriva, P. Holba, S. Durcok, D. Zemanova, E. Polert and A. Triska, Physica C
157, 334 (1989)
81. L. N. Dem'yanets, Soviet Phys. USPEKHI, 34, 36 (1991)
82. Th. Wolf, W. Goldacker, B. Obst, G. Roth and R. Flukiger, J. Crystal Growth 96,
1010 (1989)
83. P. D. Han, L. Chang and D. A. Payne, J. Crystal Growth 128. 798 (1993)
84. W. D. Mosley, J. Z. Lin, A. Matsushita, Y. P. Lee, P. Klavins and R. N. Shelton, J.
Crystal Growth 128, 804 (1993)
85. Chen Changkang, B. E. Watts, B. M. Wanklyn, P. A. Thomas and P. W. Haycock, J.

Crystal Growth 91,659 (1988)
86. A. Sen, A. Das Sharma and H. S. Maiti, Mat. Lett. 7, 433 (1989)
87. F. Holtzberg and C. Feild, Eur. J. Solid State Inorg. Chem. 27, 107 (1990)
88. F. Licci, H. J. Scheel and P. Tissot, J. Crystal Growth 112, 600 ( 199 i )
89. J. Wojcik, M. Rosochowska, H. Niculescu and Pajaczkowsky, J. Crystal Growth 91,
(1988) 255
90. T. Graf, J. L. Jorda and J. Muller, J. Less-common Metals 146, 49 (1989)
91. K. Dembisky, M. Gervais, P. Odier, J. Coutures and J. P. Coutures, Mater. Sci. & Eng.
B5, 345 (1990)
92. Chen Changkang, Hu Yongle, B. M. Wanklyn and J. W. Hodby, J. Crystal Growth
129, 239 (1993)
93. C. N. Pieczulewski, et at: J. Am. Ceram. Soc. 73, 2141 (1990)
94. R. J. Cava, et al. Nature 339, 291 (1989)
95. E. M. Levin and R. S. Roth, J. Research Natl. Bur. Standards 68A. 199 (1964)
96. T. Onoda, et al. Jpn. J. Appl. Phys. Part 2, 26, L771 (1987)
97. A. M. M. Gadalla, "Science of Ceramics", Vol. 3 Ed. by G. H. Stewart, Academic
Press Inc (London) Ltd. 1967
98. E. Gebhardt and W. Obrowski, Z. Metallk. 45, 333 (1954)
99. C R. Veakateswara Rao, P. Murugakoethan, R. Jayavel, C. Subramanian and P.

Ramasamy, Supercond. Sci. Technol. 6, 443 (1993)
100. K. N. R. Taylor, "Studies of high temperature superconductors", ed. by A. V.
Narlikar, Nova Science Publishers, New York, Vol. 6, I (1990)
101. H. J. Scheel, MRS Bulletinl Sept. p26 (1994)
102. W. Assmus and W. Schmidbauer, Supercond. Sci. Technol. 6, 555 (1993)
103. Chen Changkang, Prog. Cryst. Growth & Charact. Mat. 24, 213-267 (1992)
104. Chen Changkang, "Studies of high temperature superconductors", ed. by A. V.
Narlikar, Nova Science Publishers, New York, Vol. 16, 101-155 (1995)
105. L. F. Schneemeyer, R. B. Van Dover, S. H. Glarun, S. A. Sunshine, R. M. Fleming, B.
Batlogy, T. Siegrist, J. H. Marshall, J. V. Waszczak and L. W. Rupp, Nature 332, 422
(1988)
106. Chen Changkang, B. M. Wanklyn, D. T. Smith and F. R. Wondre, J. Mater. Sci. 2_6,
5323 (1991)
107. J. Cassedanne and C. P. Campelo, An. Acad. Brasil. Cienc. 38, 36 (1966)
108. Chen Changkang, Hu Yongle, B. M. Wanklyn, J. W. Hodby and F. R. Wondre, J.
Mater. Sci. 28, 5045 (1993)
109. B. G. Kaldaan, V. B. Lazarev and I. S. Shaplygin, Zh. Neorg. K_him. 24, 1663 (1979)
110. R. Guillermo, P. Conflant, J. C. Boivin and D. Thomas, Rev. Chem. Miner. 15, 153
(1978)
111. P. Conflant, J. C. Boivin and D. Thomas, J. Solid State Chem. 18, 133 (1976)
112. D. Elwell and H. Scheel,"Crystal growth from high temperature solution", Acad.
Press, London, New York, San Francisco (1975)
113. U. Kuxmann and P. Fischer, ErzmetaU, 27,533 (1974)
114. R. S. Roth, C. J. Rawn, J. J. Ritter and B. P. Burton, J. Am. Ceram. Soc, 72, 1545
(1989)
115. B. M. Wanklyn, E. Dieguez, Chen Changkang, A. K. Pradhan, J. W. Hodby, Hu
Yongle, D. Smith and F. W. Wondre, J. Crystal Growth, 128,738 (1993)
116. Chen Changkang and B. M. Wanklyn, J. Crystal Growth, 96, 547 (1989)
117. E. Sceil and H. Stadebnair, Z. Metallk, 43,227 (1952)
118. W. Plato, Z. Physik. Chem, 58, 364 (1907)
119. A. Erb, E. Walker, R. Flukiger. Physica C 243, 245 (1995)
120. Wai Lo, D. A. Cardwell, S. L. Dung, R. G. Barter, J. Mater. Sci. 30, 3995 (1995)
121. C. J. Kim, K. B. Kim, H. W. Park, G. W. Hong, Physoca C 2~0, 153 (1995)
122. V. R. Todt, S. Sengupta, D. Shi, P. R. Sakm, P. J. McGinn, R. B. Poeppel and J. R.
Hull, J. Electron. Mater. 22, 1285 (1993)
123. Y. L. Chen, L. Zhang, H. M. Chen and M. P. Harmer, J. Mater. Res. 8, 1 (1993)
124. M. Sumida, M. Tagami, Ch. Krauns, Y. Shiobara, T. Umeda, Physica C 249, 47
(1995)
125. K. Osamura, S. Sato, W. Zhang and T. Kizu, Physica C 185-189. 589 (1991)
94 Chen Changkang
126. Chen Changkang, A. T. Boothroyd, Hu Yongle, F. R. Wondre, B. M. Wanklyn and J.
W. Hodby, Physica C 214. 231 (1993)
127. S. Kambe, K. Ydshio, N. Ooba, N. Sugii, K. Kitazawa and K. Fueld, Superconducting
materials, Jpn. J. Appl. Phys. Series, eds. S. Nakajima and H. Fukuyama (1988) pp.l 1-14
t28. Chert Changkang, Hu Yongle, B. M. Wanldyn, S. Hazell, A. K. Pradhan, J. W.
Hodby, A. Boothroyd and F. R. Wondre, J. Crystal Growth, 128, 767 (1993)
129. G. L. Olcese, Proceeding of 3rd national meeting of high Tc superconductivity,
Genoa, Italy, 12-14 Feb. (1990)
130. K. Oka, M. Saito, M. Ito, K. Nakane, K. Murata, Y. Nishihara and H. Unoki, Jpn. J.
Appl. Phys, 28, L219 (1989)
131. Chen Changkang, J. W. Hodby, Hu Yongle and B. M. Wanldyn, J. Mater. Chem, 4,
464 (1994)
132. Chen Changkang, J. W. Hodby, Hu Yongle and B. M. Wanldyn, to be published
133. Chen Changkang, J. W. Hodby, Hu Yongle and B. M. Wanldyn, to be published
134. A. Katsui, Y. Hidaka and H. Ohtsuka, Jpn. J. Appl. Phys, 26, L1521 (1987)
135. H. Takei, F. Sakai, H. Hasegawa, S. Nakajima and M. Kikuchi, Jpn. J. Appl. Plays, 32,
L1403 (1993)
136. T. Manako, Y. Shimakawa, Y. Kuboand, H. Igarashi, Physica C 190, 62 (1991)
137. Ch. Kranns, M. Tagami, M. Sumida, Y. Yamada and Y. Shiohara, Proc. 6th. symp.
on Superconductivity (spring-verlag, Tokyo, (1994) p767
138. E. M. Levin et al, Phase diagrams for ceramists, l, 97 (1979)
139. A. N. Maljuk, A. A. Zhokhov, G. A. Emel'chenko, I. I. Zverkova, A. N. Turanov and
V. Sh. Shekhatman, Physica C 214, 93 (1993)
140. K. Watanabe, J. Crystal Growth, 114, 269 ( 1991 )
141. S. N. Barilo, A. P. Ges, S. A. Guretskii, D. I. Zhigunov, A. A. Ignatenko, I. D.
Lomako, A. M. Luginets and V. V. Novokhrost, Bull. Mater. Sci, 14, 387 (1991)
142. H. J. Scheel and F. Licci, J. Crystal Growth, 85, 607 (1987)
143. R. S. Roth, J. R. Dennis and K. C. Davis, Advan. Ceram. Mater. _1,303 (1987)
144. J. Sestak, "Studies of High Temperature Superconductors", ed. A. Narlikar, Nova
Science publishers, New York, Vol. 7_, 23 (1990)
145. K. Fischer, R. Hergt, D. Linzen, Cryst. IRes. Technol. 2_~, 1169 (1988)
146. Chert Changkang, A. J. S. Chowdhury, Hu Yongle, M. Spears, J. W. Hodby, B. M.
Wanklyn, Supercond. Sci. Technol. 7, 795 (1994)
147. A. V. Narlikar, S. B. Samanta, Chen Changkang, Hu Yongle, J. W. Hodby. B. M.
Wanldyn, J. Crystal Growth, !58,248 (1996)
148. T. Siegrist, L. F. Schneemeyer, J. V. Waszczak, N. P. Singh, R. L. Opila, B. Batlogy,
L. W. Rupp and D. W. Murphy, Phys. Rev. 13, 8365 (1987)
149. A. V. Narlikar, P. K. Dutta, S. B. Samanta and O. N. Srivastava, J. Crystal Growth,
121,527 (1992)
150. A. V. Narlikar, P. K. Dutta, S. B. Samanta, O. N. Srivastava, P. Ramasarny, S. C.

Sabarwal, M. K. Gupta and B. D. Padalia, J. Crystal Growth, 116, 37 (1992)
151. B. N. Sun and H. Schmid, J. Crystal Growth, l O0, 297 (1990)
152. S. R. Ovshinsky, R. T. Young, D. D. Allred, G. De Maggio and G. A. Van der
Leeden, Phys. Rev. Lett. 58, 2579 (1987)
153. Meng Xian-Ren, Ren Yan-Ru, Fu Min-Hua, et al, Solid State Commun. 64, 325
(1987)
154. P. K. Dutta, S. B. Samanta and A. V. Narlikar, Chen Changkang, J. W. Hodby and B.
M. Wanldyn, Phil. Mag. A, 66, 507 (1992)
155. S. M. Rao, B. H. Loo, N. P. Wang and R. J. Kelley, J. Crystal Growth, 110, 989
(1991)
156. H. Asaoka, H. Takei, Y. Iye, M. Tamura, M. gdnoshita and H. Takeya, Jpn. J. Appl.
Phys, 32, 1091 (1993)
157. A. A. Zhokhov and G. A. Emel'chenko, J. Crystal Growth, 129, 786 (1993)
158. Y. Yamada and Y. Shiohara, Physica C 217, 182 (1993)
159. Y. Namikawa, M. Egami, S. Koyama, Y. Shiohara, H. Kutami, MRS Bulletin,
February (1996) p79
160. G. Izquierdo and A. R. West, Mater. Res. Bull. 15, 1655 (1980)
161. J. Akimeto, Y. Gotoh, K. Kawaguchi and Y. Oosawa, J. Solid State Chem. 96, 446
(1992)
162. J. A. Campa, M. Velez, C. Cascales, E. G. Puebla, M. A. Monge, I. Rasines, Private
commun.
163. I. I. Kitaigorodskii, T. A. Popova and O. K. Botvinkin, Neorgan. Khim. Akad. Nauk
SSSR, 6, 139 (1933)
164. B. M. Wanldyn, Chen Changkang, P. A. J. de Groot and P. Rapson, Supercond. Sci.
Technol. 2, 129 (1989)
165. A. W. Hewat, "Materials and crystallographic aspects of HTc-superconductors', ed.
E. Kaldis, Kluwer acadamic publishers, The Netherland, p l7 (1994)
166. M. B. Brodsky, MRS Bulletin, May 1991, 31
167. R. S. Roth, C. J. Rawn and L. A. Bendersky, J. Mater. Res. 5, 46 (1990)
168. A. P. Mackenzie, S. D. Hughes, J. R. Cooper, A. Carrington, Chen Changkang, B. M.
Wanldyn, Phys. Rev. B 45, 527 (1992)
169. D. T. Marx, P. G. Radaell, J. D. Jorgensen, R. L. Hitterman, D. G. Hinks, S. Pei and
B. Dabrowski, Phys. Rev. B 46, 1144 (1992)
170. Chen Changkang, M. Spears, J. W. Hodby, B. M. Wanldyn, O. Stoppard and J. A.
Wilso, J. Crystal Growth, 1~ l, 295 (1995)
171. A. Katsui and Y. Hidaka, J. Crystal Growth, 66, 228 (1984)
172. U. Amador-Elizondo, M. A. Senari, E. Moran, M. A. Alario-Franco, A. Vegas and M.
Martinez-Ripoll, J. Less-Common Matels, _L~ &! 65, 900 (1990)
96 Chen Changkang
173. S. M. Rao, R. H. Chang, K. S. Law, M. K. Wu, C. P. Khattak, J. Crystal Growth,

162. 48 (1996)
174. Chen Changkang, Hu Yongle, M. Spears, J. W. Hodby, B. M. Wanklyn, A. V.
Narlikar and S. B. Samanta, J. Mater Sci. Lett. 15, 886 (1996)
175. C. Kriebel, M. Banazadeh, G. Winkel, N. Knauf. B. Roden, M. Braden, B. Freitag, J.
Crystal Growth, 141, 124 (1994)
176. M. Egami, Y. Yamada, Y. Namikawa and Y. Shiohara, J. Electronic Materials, 24,
1907 (1995)
177. X. Yao, T. Mizukoshi, M. Egami, H. Zama, M. Nakamura and Y. Shiohara, Jpn. J.
Appl. Phys. 35, 2126 (1966)
178. R. S. Roth, C. J. Rawn, J. D. Whitler, C. K. Chiang and W. K. Wong-Ng, J. Am.
Ceram. Soc. 72, 395 (1989)
179. A. N. Maljak, G. A. Emel'chenko, I. I. Zver'kova and A. V. Kosenko, Supercond. Sci.
Technol. 7, 596 (1994)
180. X. Yao, M. Egami, Y. Namikawa, T. Mizukoski, Y. Shiohara, S. Tanaka, J. Crystal
Growth 165, 198 (1996)
181. Chen Changkang, J. W. Hodby, M. Spears, F. Wondre, B. M. Wanklyn, M. Slaski
and C. Gough, Physica C 282-287, 537 (1997)
182. I. Tanaka, H. Kojima, Nature 337.21 (1989)
183. Chen Changkang, I. Kosseva, T. Hargreave, F. Wondre, J. W. Hodby and B. M.
Wanklyn, Physica C 282-287, 445 (1997)
184. X. Yao and Y. Shiohara, Supercond. Sci Technol. 10, 249 (1997)
185. M. Nakamura, H. Kutami, Y. Shiohara, Physica C 260. 297 (1996)
186. X. Yao, M. Kambara, T. Umeda and Y. Shiohara, Jpn. J. Appl. Phys. 36, L400
(1997)
187. M. Tagami, Y. Shiohara, J. Crystal Growth 171,409 (1997)
188. M. Tagami, M. Kambara, T. Umeda and Y, Shiohara, J. Mater. Res. 13 (1998) 118
189. N. M. HWang, R. S. Roth and C. J. Rawn, J. Am. Ceram. Soc., 73, 2531 (1990)
190. V. A. Efremov and V. K. Trunov, Zh. Neorg. Khim., 19, 501 (1974)
191. B. Lake, D. A. Tennant, R A. Cowley, J. D. Axe and C. K. Chen, J. Phys.: Condens.
Matter, 8, 8613 (1996)
Chen Changkang graduated from the University of Science and technology of China in
1965 and began his scientific career in Fujian Institute of Research on the Structure of
Matter, Chinese Academy of Sciences. He was engaged in physico-chemistry investigation
and crystal growth of Laser and non-linear optical materials for 23 years. He become the
group leader of material physico-chemistry at the institute in 1978. In 1987 he was
rewarded a Fellowship of the Royal Society and carried out a collaboration research on
high Tc superconducting crystals in Clarendon Laboratory, the University of Oxford. Since
1989 he has been working as a research staff in the UK National Crystal Growth Facility
for Superconducting Oxides in Oxford. He was selected to be a senior research association
in Balliol College in 1994.
Dr. C. Chen is the author of over 100 research papers. He translated a book "Laser
Crystals" from Russian into Chinese (Science Press of China, 1981 ). Two large reviews on
high Tc superconducting crystals were published in the international review journal of
PCGCM and in a book in USA. During working in the National Crystal Growth Facility
for Superconducting Oxides, he has grown more than 40 kinds of single crystals, which
cover nearly all high Tc families and relative compounds. More than two thousand
specimens have been supplied to the national and international commtmities for wide
range researches. He also published over 10 phase diagrams on superconductor and
related systems. He was a member of Chemistry Society of China and is a member of
British Association for Crystal Growth.
Dr Chen Changkang
Clarendon Laboratory
University of Oxford
Parks Road
Oxford. OXI 3PU

Diagarama de Fase de Los Cupratos

Uploaded by

Copyright:

Available Formats

Diagarama de Fase de Los Cupratos

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Diagarama de Fase de Los Cupratos

Uploaded by

Copyright:

Available Formats

Prog. Crystal Growth and Charact. Vol. 36, Nos. 1-2. pp.

PHASE DIAGRAM AND ITS APPLICATION

Clarendon Laboratory, Oxford University. E-mail: c.chenl @physics.ox.ac.uk

I. High Tc oxide superconducting phase and crystal growth 2

PHASE DIAGRAM AND ITS APPLICATION TO THE CRYSTAL GROWTH

I. HIGH Te OXIDE SUPERCONDUCTING PHASE AND CRYSTAL GROWTH

Since the great discovery of the physical phenomenon of superconductivity in mercury by

191C 1930 t950 lq7p 1QQO ;' ~f!c "loar

Fig. I. The progress of Tc for superconducdn.~, matcrialx

--- I~KB'] ~ '

0 2(1 /10 1) 1 8() I(l() 1'21} ]:l(I

Fig. 2. Superconducting transition temperature (Tc) and deconlposition (Td) or melting

( u oxidation stale . . . . . . BaPbO3., X -- ~'~ t3a't~iO

Fig. 3. Schematic phase diagrams of oxide superconductors

(4) Phase transition in solid state like

YBa2Cu306(Tet. ) + 1/402 = YBa2Cu306.5(orth-II ) + I / 4 0 2 = YBa2Cu307(orth-I ) (1)

2YBa2Cu307 + 3 H 2 0 ===3Ba(OH) 2 + 5CuO + Y2BaCuO5 + 0 . 5 0 2 (2)

no YBCO phase was obtained in the hydrothermal experiments [18].

Table 1. Flux growth of oxide superconductors

Crystal Flux Note Ref.

S-F: self-flux, P.S-F: Partially self-flux

I1. THE PttASE RULE, PHASE DIAGRAM AND CRYSTAL G R O W T H

Nla= Nlb= NlC . . . . . . . . NlP "1

Nca= Ncb= NcC . . . . . . . . . . NcP

Therefl)re the number of indcpcndenl parameters is P(C-1)-C(P-1) = C-P, With the

2YBa2Cu307_y ===Y2BaCuO5 + 3BaCuO 2 + 2Cud + (0.5-x)O 2 (5)

3YBa2Cu307_y ===Y2BaCuO5 +YBa2Cu408+3BaCuO2+CuO+(O.5-1.5x)O 2 (6)

4YBa2Cu307-y===Y2BaCuO5+Y2Ba4Cu7014+x+3BaCuO2 +Cud+( 1-2y-x/2)O 2 (7)

~4(:-:,c, " "°

Fig. 4. The system of Na2WO 4 - WO 3

Table 2. Phases of NaxWO 4 grown from the flux system of N a 2 W O 4 - W O 3 by

W O 3 in melt (mol.%) <5(:) 50-62 62-70 >70

Temp. range ( C ) 700-900 740-790 790-890 >890

2. Binary system -- X-T diagram:

D + L ......... > S C (8)

Fig. 7. Triangle prism for ternary Y2BaCuO5(211) - Ba2Cu305(023) - CuO(O01)

The three phases are located at:

(1 2 3 ) = 0.8( 1 2 1 1 / 4 ) - 0.2(0 0 1) (9)

0.5(24 7)=0.571(1 2 11/4)-0.429(00 1) (1o)

(1 2 4) = 0.444( 1 2 11/4) - 0.556(0 0 1) (11)

Y2Ba4Cu7014+y .... > 2YBa2Cu307_ x + CuO + 0.5(2x+y-1)O 2 (12)

Y2Ba4Cu7014+y ----> YBa2Cu307_ x + YBa2Cu408 + 0.5(x+y-1)O 2 (13)

yO U 2," 4-[/ h 2 [,~,p8 5 ~0 LuO

(211 ).( I 2~ )Te " (125)Te*(0 ~I*

(12~} ~.(Ll'i 1 *"

: :.:;~ I t0 ..... '* I '~ s:o..,lo 1

~,~, ~,~ A (~,,: ) I, L ,~1,( n ' ~ tool % Cu()

Fig. g (l).Ternary system of Y O I . 5 - BaCuO 2 - CuO at 9 5 0 - 1 0 0 0 °C.

\ ~4(} l J .I J d) 4 (-~( u-(~AP~_ ~*Ca('u

Fig. 9(a). Bi 1.5-SrO-CaO-CuO quaternary system

Fig. 10. Pseudo-binary system of TI2Ba2CuO6-(CaO+CuO )

HI. EXPERIMENTAL DETERMINATION OF PHASE DIAGRAM

Table 3. Thermogravimetr~c analysis of the starting mixture of 2.28 BaCO 3 + 2 CaCO 3 +

T (°C) 740 940 960 1160 1230

T ( °C ) 660 730 780 830 880

PbO 2 . . . . > PbO + 0.5 0 2 (14)

BaCO 3 + PbO + 0 . 5 0 2 .... > BaPbO 3 + CO 2 (15)

which is completed at around 700 °C with a loss of CO 2 comparable to the calculated

Table 5. Crystal growth of YBCO with and without precursors

(211 ).( I 2~ )Te " (125)Te(0 ~I

_ I ul ~ _ 1 2 . i . , ......J_ /1 ir ~L±.d _ _ __[ I