PHYSICAL REVIEW
B
VOLUME
38,
NUMBER
13
1 NOVEMBER
1988
Preparation,
structure,
and
properties
of
the
superconducting
compound
series
Bi2Sr2Ca„—
1Cu,
o»
with
n
1,
2,
and
3
J. M.
Tarascon
Bellcore, 331
Newman
Springs
Road,
Red
Bank,
New
Jersey
07701
W. R. McKinnon
National
Research
Council,
Division
of
Chemistry,
Ottawa,
Canada
KIAOR9
P.
Barboux,
D.
M.
Hwang,
B.
G.
Bagley,
L. H. Greene, and
G.
W.
Hull
Bellcore,
331
Newman
Springs
Road,
Red
Bank,
New
Jersey
07701
Y.
LePage
National
Research
Council,
Division
of
Chemistry,
Ottawa,
Canada KIAOR9
N.
Stoffel
and
M. Giroud
Bellcore, 331
Ne~man
Springs
Road,
Red
Bank,
New
Jersey
07701
(Received
18
May
1988;
revised
manuscript
received
26
August
1988)
Crystals
of the
three
Bi-based
cuprates
of
general
formula
Bi2Sr2Ca,
-&Cu„O~
with
n
1,
2,
and
3
have
been isolated and their
structural
and
physical properties
investigated.
The structures
are
similar,
diR'ering
only
in
the number of
Cu02-Ca-Cu02
slabs
packed
along
the c axis. The
inser-
tion
of
one and two
slabs increases
c from 24.6 to
30.
6 and
37.
1 A.
Transmission
electron
micros-
copy
shows there
are
stacking
faults within
the
crystals
in
agreement
with
our
x-ray
data and
its
analysis.
Resistivity, ac
susceptibility,
and dc
magnetization
measurements
demonstrate
supercon-
ductivity
in
the n
1, 2,
and 3
phases
at
10, 85,
and 110
K,
respectively.
The observed
transition
temperatures and the
stacking
fault
densities are
dependent
upon
sample
processing,
in
particular,
the
annealing
temperatures
and
cooling
rates. The
transition
temperature
is,
within
the
accuracy
of
our chemical
titration,
independent of the
average
copper
valency
that
was determined
to be
2.
15+'0.
03 for
each of the three
compounds.
I. INTRODUCTION
Since
the
discovery
by
Bednorz and
Miiller'
of
high
transition
temperature
(T,
) superconductivity
in the
ox-
ides and the
achievement of
T,
90 K
by
Wu
et al.
,
these
new
materials have
all
contained either
La or a rare
earth as a
principal
constituent.
Recently,
superconduc-
tivity
at 20
K
was
reported
by
Michel
et
a/. for the
Bi-Sr-
Cu-0
system.
3
The
addition
of Ca to this
ternary
led
Maeda, Tanaka,
Fukutumi,
and
Asano to
the
discovery
of bulk
superconductivity
at 85 K and evidence
of
super-
conductivity at 110 K in the
Bi-Sr-Ca-Cu-0
system.
Since this initial
discovery,
there
have
been
many
publica-
tions
detailing
the existence
of
superconductivity
at
T=110
K,
and
discussing
the structure of
compounds
of
the form
Bi-Sr-Ca-Cu-O.
"
The
compound
of formula
Bi4Sr3Ca3CuqO|s
(hereafter denoted 4:3:3:4)was found
to be
responsible
for
superconductivity
at
85 K in the
Bi
system
and its
structure established.
'
'
The
crystal
sub-
structure can
be vie~ed as a
three-dimensional
packing
of
Bi2Sr2CaiCu208
slabs
along
the
c
axis,
with
the
main
feature
being
the
presence
of two
Bi-0
layers separated
by
3.
0
A and
shifted, with
respect
to
each
other
(crystallo-
graphic
shear)
along
the
diagonal
direction
of
the
perovskite
subcell.
Within
this
class of
compounds,
Bi
can
be
replaced
by
thalium
and
new
phases
of
general
formula
T12Ba2Ca„—
~Cu„O»
(n
1, 2,
and
3)
have been
isolat-
ed.
'4 's
The
thallium
substitution
does
not
significantly
affect the structure,
which is
tetragonal
with a double lay-
er of
Tl-0
se arated
by
2.
6
A.
,
but does
raise
T,
to 125 K
when n
3.
'
In contrast
to the
thallium
phases,
which
are
easy
to
form
independent
of
n,
difficulties
have been
encountered in
the
preparation
of the homologue
Bi
phases,
particularly
the
110-K
phase.
We
previously'
measured
zero resistance
at
107
K
on
Bi
pseudomorphs
of
composition 2:2:2:3,
but
were
not
able to
determine the
structure
because
of the absence of long-range
structural
order.
However, the
presence
of
a
larger c
axis
for
the
110
K
was
evidenced
by
the
presence
of
a
Bragg
peak
located
at 28 4.6
in the
x-ray
powder
pattern
of
a multiphase
sample
showing
both
85-
and
110-K
superconducting
transitions. We
report
here the
synthesis
of the
Bi2SrqCa„—
~Cu„O»
phases
with
n-l,
2,
and
3
and on the
relationship
between their
structural and
physical
proper-
ties. We
show evidence
that the
110-K
phase
has the
2:2:2:3
formula
and,
furthermore,
that its
structure
con-
tains three
Cu02 layers
with the
possibility
for stacking
faults.
For
clarity,
we
present
the synthesis
and
physical
measurements
for each
phase separately
and
comment on
their interrelationships
in
the discussion
section.
38 88&5
1988
The
American
Physical Society
8886
J.
M.
TARASCON
et
al.
38
II.
SYNTHESIS AND STRUCTURE OF THE
Bi2Sr2Ca,
—
~Cu„O„PHASES
-
Bi2Sr2Can
&CunOy
(a)
A.
n
Michel et al.
reported superconducting
critical
tem-
peratures
ranging
from 10
to 20
K
for
the
Bi-Sr-Cu-0
system
and attributed
superconductivity
to
the Bi2-
Sr2Cu207 phase,
whereas
a different
phase (Bi2Sr2CuOs)
was
proposed
by
two
other
groups.
'
's
However, all
agree
that
the
T,
of
the
resulting
material
is
strongly
dependent
on
sample
processing.
We
observe,
for
exam-
ple,
that
depending
upon
whether the
sample
is
heated
below
or
above its
melting
point
produces
marked
differences in
the
resulting
physical
properties.
Bi2Sr2CuOs
(2:2:0:1)was
prepared
by
firing,
at
high
temperatures,
stoichiometric
amounts
of
Bi203,
SrCO3,
or
Sr02,
and CuO
powders
(each 99.999%
pure).
Three
heating processes
were
investigated:
(1)
the mixed
pow-
ders
(using
SrCOs)
were
fired at
temperatures
ranging
from
840-880'C
for several
days
and
then furnace cooled
to
room
temperature
(samples
denoted as
P);
(2)
the
mixed
powders (using SrCO3)
were
melted
at
1100'C
for
2
h,
the
temperature
was lowered
to
890'C
and held there
for
several
days,
then the
samples
were furnace cooled
(samples
denoted
as
M);
(3) Sr02
was
used as the Sr
ox-
ide
precursor,
and
pressed
pellets
were
placed
on
alumina,
sealed in
quartz
tubes,
and reacted for
15
h
at
870'C
(samples
denoted
as
Q).
The M
samples, depending
on
the
cooling
rate from 1100
to
890'C,
may
exhibit either
large
plateletlike
crystals radiating
out of the bulk
(4 h
cool) or
an acicular structure
(furnace cooled).
This is
common
for
layered-type
materials.
For
example
graph-
ite,
which has a lamellar
structure,
can
also be made in a
fiber
form
(graphite
fibers). Both M and P
samples
con-
tain
a
few
percent
of
a second
phase,
but the
Q
samples
are
single phase.
In
all
three
cases,
the
Bragg
peaks
of the
majority
phase
can be
completely
indexed on
a
pseu-
dotetragonal
unit cell with lattice
parameters a 5.
39
A.
and c 24.
6
A.
The
experimental agreement
in the lattice
parameters
for
samples
Q
(prepared
in sealed
tubes)
with
those
of
samples
prepared
in
open
containers
shows
that
there is little
or
no
Bi lost
from the
open
containers
during
the reaction.
Although
the M
samples
contain
a
second
phase,
that
phase
is
not
present
in
plateletlike
crystals picked
from
the
M
samples.
Figure
1(a)
shows the x-ray
diffraction
pat-
tern of
such
crystals.
Because the
crystals
are oriented
(as
we
found
before'
),
only
the
[001]
reflections
appear.
The
plateletlike
crystals
have
a
composition 2:2:0:1,as
deter-
mined
by
Rutherford
backscattering
spectrometry
(RBS)
(see
Fig.
2).
Transmission electron
microscopy
(TEM) on
crushed
(2:2:0:1)
crystals
reveals lattice
parameters
similar to
those determined
by
x-ray diffraction
with,
in
addition,
an
incommensurate
superstructure
along
the
[Ok
0]
direction.
A TEM
micrograph
is shown in
Fig.
3(a). Note that in
the
[001]
direction there
is
a
Bi-0
double
layer repeat
dis-
tance of 12.3 A.
Because of
the
crystallographic
shear
in
Bi
layers,
this
leads to
a c parameter of
2x12.
3
24.
6
A.
These
results
agree
with
those
reported
by
Torardi
et
al.
'
for the
Bi2Sr2Cu06
phase
containing one
Cu02
layer.
n=1
Tc
—
—
10K
c
=24.
65A
002
006
008
I
g)
'
I
,
"v'
t
008
0010
LA—
ol
0012
g)
v-
O
NO
Fl=2
Tc
-—
85K
c
=30.7A
L
O
c
002
006
0010
0012
0016
0012
(c)
002
0010
fl=3
Tc
=110K
c
=37.
1A
I
QQ14
0024
002Q
o
1017
4
14
24
34
44
54
Diffraction
Angle
28
(deg)
FIG. 1.
The x-ray
diffraction
patterns
for
highly
c-oriented
multicrystalline
samples
of the three Bi
phases
are
shown
with
the
Miller
indices
above
each
peak.
Note
that
only
the
[00ll
reflections with
l
even
are observed.
Resolution
has been
de-
creased in order
to increase the
intensity
of the
Bragg
peaks
in
(a)
and
(c).
2.
0
I
2.
2
I
Energy
(MeV)
2.4
2.
6 2.
8
I
3.0
I
80—
3.08
MeV He RBS
a.
O
20—
0
350
400
Cu
450
Sr
500
550
Channej
FIG.
2. RBS measurement
of a
Bi4Sr4Cu20~
crystal
prepared
from the
melt (see
text).
The
simulation
for
one,
two,
and
three
Cu
are shown
by
curves
a, b,
and
c,
respectively.
38
PREPARATION, STRUCTURE, AND
PROPERTIES
OF THE. . . 8887
B.
n 2
6~
*
The
preparation
of the
Bi2Sr2Ca~Cu20s
phase
from
stoichiometric
amounts
of
the oxides and carbonates
re-
sults in
a
multiphase
sample.
'
However, single-phase
materials
are obtained
by
firing
(at
850'C
for two
days)
a
mixture
of nominal
composition
4:3:3:4.
Crystals
of
this
phase
have
been
grown using
an
excess
of
Bi203
and CuO
which,
at these
temperatures,
act as a flux. The x-ray
diffraction
pattern
of
a plateletlike
crystal
is
shown in
Fig.
1(b). The
[00l] reflections,
which can be
used as
a signa-
ture,
are shifted towards lower
angles
indicating
a
larger
c
axis
than for the
n 1
phase.
Single-crystal
x-ray
stud-
ies'
show
a
pseudotetragonal
substructure with
a
5.39
A and c 30.6 A. The TEM
image
in
Fig.
3(b)
shows
the
c
axis
repeat
distance
of
c/2
15.
0
A.
The structure
con-
tains
one
Ca
layer
sandwiched between two
Cu-02,
two
Sr-o,
and the two
Bi-0
layers
resulting
in
n
2.
As
with
the n 1
material,
there is also a superstructure
along
the
[OkO]
direction.
Single-crystal
x-ray
studies
suggest
that
the n 2
phase
has the chemical
formula
4:4:2:4,
whereas
only
compounds
close
to
the
nominal
composition
4:3:3:4
are
single
phase.
This result
remains an
enigma.
Another set of
samples (M)
of nominal
composition
4:3:3:4was
prepared
by
melting
the mixture
(oxides
plus
carbonates) at
1140'C
and then
cooling
to
890'C,
whereupon
it
was
maintained
at
this
temperature
for a
few
days
before
cooling
to room temperature.
The
result-
ing
samples
contain
plateletlike crystals
(corresponding
to
the
n
2
phase)
embedded
in a material
which
contains a
few
precent
of
second
phase
as
determined
by
x-ray
diffraction. These
crystals
have
lattice
parameters
which
are
identical
to
those
of
the
powdered
samples,
but
they
exhibit a different
T„as
discussed in Sec.
III.
C. n 3
FIG.
3.
Crushed
powders
of the three
types
of
samples
were
examined
using
a JEOL 4000FX
transmission
electron
micro-
scope. (a) TEM lattice
image
of the
10-K
material
along
the
[100]direction.
The double
dark
layers
are
assigned
as
the
dou-
ble
Bi-0
layers.
The
doubling
of the unit cell
due
to the
shear-
ing
of the Bi
layers
is delineated
by
the
interweaving
appearance
of this
micrograph. (b)
Bright-field
TEM
image
of
the
85-K
material
showing uniform
c-layer
spacing
of
15.4
A.
(c)
Bright-field
TEM
image
of
the
110-K
material
showing
imper-
fect
order
in the
c direction. The
observed
c-layer
spacing
is
usually
19
A. However, about
every
5
to 10
layers
a c-layer
spacing
of
15 A
is
observed.
(b)
and
(c)
were
taken with the
c-
axis
normal
to the
electron
beam. The resolution
is limited
by
sample
thickness.
Compared
to the
10-K
material,
the
85-
and
110-K
materials
are much more
anisotropic,
and
sections
with
dimensions
perpendicular
to the
c direction
thin
enough
for
high-resolution
imaging
are difficult to
find in
the crushed
powders.
Mixtures of
nominal
composition
4:3:3:4were
rapidly
raised
to
temperatures close
to,
but
below,
the
melting
point
of
the
4:3:3:4
phase
(885'C)
and
maintained at this
temperature
for
about
a week,
then
furnace cooled
(6
h).
After
such
a heat
treatment,
the
alumina
crucible
con-
tainer
had,
at
its
center,
a black
sintered
phase
on
the
top
of
which
were
crystals
of
golden
color
surrounded
by
a
yellowish
phase,
most
likely
resulting
from the
decomposi-
tion
of
the 4:3:3:4
phase.
The bulk of the
material
had
"crystals"
with metallic luster.
These
crystals
exhibit
only
intermediate-range order
(cryptocrystalline).
The
degree
of long-range
order of the
crystals
changes
with
the
processing
treatment and
may
also
differ
between
crystals
belonging
to
the same
batch.
Polycrystals
with
long-range
order and a
110-K
transition have
been
isolat-
ed,
chemically
analyzed,
and studied
by
x-ray
diffraction
and TEM. A cation
concentration
of
2:2:2:3,
similar
to
the
125-K
T,
thallium
phase,
was obtained
by
wet
chemi-
cal
analysis.
The x-ray
pattern
for one
of these
"crystals"
pressed
onto
a
glass
slide
is
shown in
Fig.
1(c).
Because
of
the
strong
preferential
orientation,
only
the [OOI]
lines
diffract.
Figure
1
compares
this
diffraction
pattern
to
those
obtained
using
plateletlike crystals
of
the
n 1 and
n 2
phases.
Note
that the low-angle
Bragg
peak
is
again
8888
J.M. TARASCON
et al.
38
shifted
to
lower
angles
thus
indicating
an
even
larger
c
axis.
From
a
least-squares
fit
of
the
Bragg
peaks
the
unit-cell
parameters
a 5.
39
A
and c
37.
1 A are
ob-
tained.
By
analogy
to
the
thallium
phase
the increase in
the
c
axis
results from
an extra
CuOz layer (n
3).
Moreover,
the diffraction
lines for
the
2:2:2:3
phase,
as
shown in
Fig.
1(c),
are broad and
they
are not
exactly
at
[001] positions.
This
line distortion could
result
from
the
presence
of
stacking
faults.
Because of the
possible
im-
portance
of
such faults on
the
superconducting
properties,
we
present
the details
of our
analysis
characterizing
the
stacking
faults. A least-squares fit
of
the lattice parame-
ter
c to
the
[001]
peaks
of
Fig.
1(c)
gives
a
root-mean-
squared
deviation
of 0.
2'
between
the
calculated
and
measured values
of the
Bragg
angle
28,
as
compared to
a
value
of
0.
02'
typical
for
our diffractometer.
This shift
and some
of
the
broadening
can
be
explained
by
stacking
faults in
the 2:2:2:3
structure,
faults in which
same
layers
of
Cu3Ca206
are
replaced
by
layers
of
Cu2Cai04
from the
2:2:1:2
structure. The effects
of
such faults
on
the x-ray
powder pattern
can be calculated
using
the
analysis
of
Hendricks and Teller.
~
In
these
Bi-based
oxides,
most of
the x-ray
scattering
is
produced
by
the sandwiches of
I
3
(haiku)
g
p
eikuj
g~2
(2)
where
pl
is
the
probability
that
a
given
adjacent
pair
of
Bi2Sr204
sandwiches
is
separated
by
j
Cu
layers.
(For
simplicity,
we consider
only
j
2 and
j
3.
)
Then
Eq.
(1)
becomes
Bi2Sr204,
the Cu
and
Ca
layers
in
between determine
the
spacing
between these sandwiches,
but add little extra
scattering. Thus,
we
can use the
simplest
case
in
Ref.
20,
scattering
from identical
layers
having
random
spacings.
Let
F
be the
structure factor for the
BiqSr204
sandwich,
and let u
2
and
u
3
be
the
spacing
between the
centers of
Bi2Sr204
sandwiches
separated
by
2
or
3 Cu
layers
(and
1
or 2
Ca
layers).
The
x-ray
intensity
I
per
sandwich is
given
by
I-fF/'
(haiku)(e
ik-u)
(haiku) (e
ik—
u)+(haiku)(e
—
iku)
where
k is
the
scattering
vector,
k
(4z/X)sin8,
and
8
is
the
Bragg
angle.
The
average
of a
quantity
such
as
(e'
")
is
defined as
I
IF
I
2
1
—
p2
—
p3
—
2p2p3cos[k(u2
—
u
3)]
I+p$+
p3+2p2p3cos[k(u2
—
u3)]
—
2p2cos(ku2)
—
2p3cos(ku3)
(3)
D
-Ay'+By+C,
where
the
coefficients are
given
by
A
~1
—
p2
—
p2(1
p2)cos(—
ku2),
8
2p2(1
—
p2)sin(ku2),
C~2pq(l
—
cos(ku2)
.
This describes
a
Lorentzian,
centered at
yo-
8/2,
—
and
with
a
full width
at half maximum
given
by
hy
-44AC
—
8'.
(5)
(7)
(8)
Both the width
and
position
of the
peaks depend
on the
amount of
disorder
p2
and
(through
ku2)
on
the
position
of
the
peaks
of
the
2:2:1:2structure.
Although
we need
We
consider
the
case
where
the
compound
is
largely
the
2:2:2:3
phase,
so
p3
&
p2.
When
p2
is
zero,
I
consists
of
b
functions located where
the
denominator
of
Eq.
(3)
goes
to
zero,
at
ku3
lx,
with
1 an even
integer.
When
p2
be-
comes
nonzero,
the
peaks
broaden and shift. The
numera-
tor
varies
slowly
over
a
given
peak,
and
so determines
only
the
peak's
amplitude. The
denominator D
determines
the
shape
and
position
of a
given
[001]
peak.
Let
p
measure
the difference
of
ku3
from an
even
integer multiple
of
ir,
according
to
ku
3
—
lx.
(4)
Then
if
p
is small
(as it is in
our
case),
the
denominator
D
becomes
the
full expressions
above
for some
of the lines,
it
is
in-
structive to
consider
the
limit where
pq
is small and
where
the
[001]
peak
of 2:2:2:3is
near
a
[001']
peak
of
2:2:1:2.
If the nearest
2:2:1:2
peak
is
8
away
from
an integer
value
of
ir,
so
that
ku2
1'ir+b,
then
from
Eqs.
(7)
and
(8)
the
peak
of
the
2:2:2:3
structure
will be
shifted
by
p2b
and
be
broadened
to a
full width at
half maximum
p2b
.
Thus
the
peaks
that
are
closest to
[001']
peaks
of
the
2:2:1:2
phase
are shifted
and
broadened
the least,
and
the
broadening
should
vary
more
strongly
from
one
peak
to
another
than the
shift does.
Of
the observed [001]
lines
in
Fig.
1(c),
the
lines
that
are
closest
to lines
from
2:2:1:2
are the
[0012]
and
[0024];
these
should
be the
least
broadened.
The [0014]
line should be most
broadened.
The broadening
does
not
vary
as
dramatically
as
the above
results
suggest,
imply-
ing
other
contributions
to
the
widths
of a
peak.
Thus
we
used
the
peak
positions
rather
than
their
widths to
esti-
mate
p2.
We
used
Eq.
(7)
for
po
to
correct
the
observed
peak
positions
for
different
values of
p2,
and
found that
for
p2
0.25
the
least-squares
refinement
of
the
c-lattice
parameter
gives
the
lowest standard
deviation.
(The
rms
deviation between
calculated and
observed angles
drops
to
0.
026'
from
the
value
of
0.
20'
found for
p2
0.
)
The
value of c
for
the
pure
2:2:2:3
structure
found
this
way
is
c
37.
08+
0.02
A.
Figure
4
compares
the
observed
and
calculated
peaks
using
our model.
The
arrows show
where
the
peaks
should be
for
p2
0.
The
dotted
lines are
Lorentzians
cal-
culated from
Eqs.
(7)
and
(8).
Note
that
the
calculated
peaks
are
in
the
correct
positions,
but are
all
too
narrow.
This can be
mostly
explained
by
the
extra
instrumental
PREPARATION, STRUCTURE,
AND PROPERTIES
OF THE
I
&
I
I I I
[oo]o]
lf
I
I
[ootz]
I l
1
I
I
I I
I I
I
too]o]
VJ
EP
a
I
1
l
23 24
28.
5
29.5
Oiffraction
Angle
28
{deg)
FIG.
4.
Line
shapes
of
the
2:2:2:3
compound,
determined
by
x-ray
powder
diffraction,
for
three
[00l]
lines taken from
the
pattern
shown
in
Fig.
1(c).
The
solid
curves are
the
data
(inten-
sity
measured
in
arbitrary
units) with
the baseline
removed
and
with
angles
corrected
for the
displacement
of the
sample
oA'
the
axis
of the
diff'ractometer.
(This correction was determined
dur-
ing
the least-squares
refinement.
)
The arrows indicate where
the
Bragg
peaks
for
a
perfect
2:2:2:3
structure
with c
37.08 A
would be. The dotted curves
are
Lorentzians whose
positions
and
widths
are
calculated from
Eqs.
(7)
and
(8).
The dashed
curves are convolutions
of the dotted curves with
a
parabolic
line
shape
(see
text).
The
extra
peaks
to
the
left
of
[0010]
and
to
the
right
of [0014]are the
[008]
and
[0012]
peaks
of
the 2:2:1:2
structure.
broadening
since we
reduced
the resolution
of the
instru-
ment
to
give
more
intensity
for this
small
sample.
Thus,
to
account for this
extra
broadening,
we convoluted the
Lorenzian
line
shape
with
an
arbitrary
chosen
parabolic
line
shape
(defined as
I(8)
1
—
[(8
—
80)/crl
for
[8
—
80[
&
e
and
0
for
[8
—
80[
&
cr,
where
Iis the
inten-
sity,
8 is the
Bragg
angle,
80
is the location of
peak,
and o
determines the width).
The result of this
simulation is
displayed
in
Fig.
4 (dashed
curve). In
this
figure,
o 0.
2
for all the
peaks.
Part
of the
crystal
used
to
perform
the
x-ray
measure-
ments
was also used for TEM.
Within the basal
plane,
simple
square patterns
without or with
extra
spots
corre-
sponding
to
the
modulation
were observed.
Thus,
it
is
difficult
to
make a definitive statement
about whether
or
not the
modulation
exists for the 2:2:2:3
phase.
A
TEM
micrograph
[Fig.
3(c)]
taken
perpendicular
to the
[001]
direction
shows
a
Bi-0
repeat
unit of
about
19
A
corre-
sponding
to
a
c
axis of
(2&
=
—
19)
=38
A because of
the
doubling
of the unit cell
due
to
the
crystallographic
shear
in
the
Bi-0
layers.
This
value
of
c
agrees
well with that
(37.
1
A)
determined
by
x
rays.
It
is also
important
to
note
that
some
repeat
units are shorter
because
of
stack-
ing
faults
(the double
Bi-0
layer repeat
unit
is
separated
by
2 instead
of
3
CuO~
layers).
This observation is
con-
sistent
with the
conclusions
from our
x-ray diffraction
analysis.
III.
PROPERTIES
OF THE
Si2Sr2Ca„—
~Cu„Oy
PHASES
We
report
here
physical
measurements
collected
only
on
crystals,
or
highly
crystalline
composites,
of the
Bi
phases
with
n=
1,
2,
and
3.
Cryptocrystalline
materials,
(:)
(:)
l—
n=2
re=85K
I
~t
n=3
I
!
Tc
—
—
110K
!
II. . . I
100 200
Temperature
(K)
300
FIG.
5.
The
resistivity
normalized
to
the
resistivity
at 300 K
for
crystals
of
Bi2Sr2Ca,
—
]Cu&Oy
with 7l
1 2
and
3 are
shown.
samples
of
the110-K
phase,
whose order
is
so
short
range
so as to
preclude a
proper
x-ray
characterization
are
not
discussed.
Resistivity measurements
as a function
of
temperature,
were made
at constant current in
a
standard four-probe
configuration.
Indium
(Ag)
solder or silver ink contacts
were used. The
crystals
examined were
usually
of
con-
stant thickness but
of
poorly
defined
shape
such that the
absolute
values
of the
resistivity have,
potentially,
a
large
error and
cannot
be used to
characterize
the
quality
of
the
crystals. Figure
5 shows
data for
the
n=1,
2,
and 3
Bi-
based
phases.
For the n 1
sample
the
resistance
de-
creases
linearly
from
room
temperature
down
to close to
T„whereas
for
n 2
the
range
of
temperature
over which
this
linear
behavior is observed is
considerably
smaller and
extends
only
from
300
down
to 150 K.
Finally,
this linear
resistivity
behavior in the
temperature
region
above
T,
is
not
obeyed
for
then 3
phase
up
to
200K.
The
negative
curvature
behavior above
T,
is unusual
and we
are not
aware
of
any
explanation
which would account for this
ob-
servation.
Thus,
we
believe that the microstructure
of
these materials
(including
the
propensity
to fault as
n
in-
creases)
may
be
responsible
for
this
curvature. This is
consistent
with
our data
since the
largest
curvature
is
ob-
served
for the n 3
phase
which contains
the most
stack-
ing
faults.
To
ensure that these
phases
are bulk
superconductors,
we
measured
magnetization
with
a dc
superconducting
quantum
interference
device
(SQUID)
magnetometer.
The measurements
were
performed
on
crystals
similar
to
those
used
for the resistivity
measurements.
Magnetic
data
are
shown in
Fig.
6.
From the magnetization
values,
without
taking
into account
the demagnetization
factor,
Meissner
values
greater
than
50% are
obtained
for
both
the
85-
and
110-K
phase.
Such
values
are
obtained
rou-
tinely
on
our bulk
polycrystalline
samples
as
well. This
is
in
contrast to
the
thallium
phases
where
the
Meissner
effect does
not exceed
25%-30%.
'
For
the
10-K
Bi-based
phase,
however,
we obtain a
much smaller
value
(10%).
The
properties
of
the three
Bi
phases
discussed thus far
do not
reflect the
difficulties encountered in
the
synthesis
of these materials, as
we
observe that
(for
all
of
these
phases)
T,
depends
upon
the thermal treatment
or
histo-
![FIG. 5. The resistivity normalized to the resistivity at 300 K for crystals of Bi2Sr2Ca, —]Cu&Oy with 7l 1 2 and 3 are shown.](/figures/fig-5-the-resistivity-normalized-to-the-resistivity-at-300-k-3m03wdsv.webp)