Full text of DOI:10.1557/JMR.2000.0090

ARTICLES
Growth mechanism of YBa2Cu3O7−␦ thin films and
precipitates on planar and vicinal SrTiO3 substrates
J. Kim
Institute for Materials Science, George Washington University,
Washington, District of Columbia 20006
D.B. Chrisey, J.S. Horwitz, M.M. Miller, and C.M. Gilmore
Naval Research Laboratory, Washington, District of Columbia 20375
(Received 17 August 1999; accepted 22 December 1999)
Effects of the vicinal angle, film thickness, and temperature on the growth modes,
microstructures, and electrical properties of YBa2Cu3O7−␦ on SrTiO3 were studied.
Island growth transition between the initial nucleation and the later coalescence stages
was observed with film thickness on a planar SrTiO3, while no islands were observed
at the later stage due to the step-flow mode. As the growth temperature increased,
a-axis precipitates were transformed to c-axis precipitates (islands), while no islands
formed on vicinal SrTiO3. The supercurrent critical temperature was strongly related to
the substrate vicinal angle due to the step-flow mode.
I. INTRODUCTION
Although these different approaches improved the
smoothness of the YBCO film surfaces, more research is
still needed to understand the origin and growth mecha-
nism of precipitates and the relationship between the film
and precipitate nucleation and growth.
The discovery of high-temperature superconductivity
set off an explosion of research on the development of
high-temperature superconducting (HTS) devices based
on YBa2Cu3O7−␦ (YBCO) thin films, including passive
microwave devices, Josephson junctions, and supercon-
ducting transistors.1 It was the potential of these devices
that catalyzed the development of new and improved
thin-film processing techniques to address the compli-
cated chemical composition and structure of these
materials. In general, junction devices will require high-
quality thin films of YBCO with classic smooth surfaces
for maximum performance. However, the nucleation and
growth of YBCO films can yield examples of crystalline
defects in thin-film ceramics, such as screw dislocations,
precipitates, or other phases. These defects limit the suc-
cessful application of YBCO thin film devices, especially
in multilayer and vertical transport systems requiring
smooth YBCO surfaces with bulklike properties. YBCO
precipitates exist within and on the surface of c-axis
YBCO films,2 and they even grew omnidirectionally cre-
ating holes through, for example, insulating oxide layers
in junction devices.3 There have been several approaches
in the conventional pulsed laser deposition (PLD) proc-
ess, such as off-axis deposition,4,5 controlling deposition
conditions,6,7 or pre-annealing the substrates,8 to mini-
mize the size and areal density of precipitates. A funda-
mentally different approach utilized bombardment with a
large ionized Ar cluster beam to remove rough surface
morphology and precipitates on YBCO film surfaces.9
The initial growth at the film–substrate interface de-
termines the subsequent growth and surface morphology
as well as electronic properties of thicker films. Many
publications proved the existence of island nuclei even
for monolayer c-axis YBCO films, and the islands grew
larger with the film thickness.10–12 It was indicated that
the nucleation and growth mode for the island formation
took place from the initial stage, and the island size de-
pended on the film thickness. Growth modes reported for
c-axis YBCO films were the island growth mode,13–15
spiral growth mode,16–18 or layer-by-layer growth mode19
above a certain film thickness, while the layer-by-layer
growth mode did not occur for the initial growth stage of
YBCO.20,21 The chemical compositions of precipitates
have been reported as Y-rich or Cu-rich phases,6,7,10,22
while a similar composition as the YBCO film matrix has
been frequently observed for island-shaped precipi-
tates.6,23 A needlelike or rectangular precipitate was ob-
served in c-axis YBCO films containing an a- or b-axis
oriented phase.24 The a-axis YBCO films grown at low
temperatures (<700 °C) resulted in smooth film surfaces
without precipitates.25,26 Since the shape and growth of
precipitates was strongly related to the growth of the film
matrix, especially from the very early growth stage, there
might be a relationship between the film and precipitate
growth mechanism. Therefore, a fundamental research
596
J. Mater. Res., Vol. 15, No. 3, Mar 2000
© 2000 Materials Research Society
http://journals.cambridge.org
Downloaded: 15 Mar 2015
IP address: 165.95.226.100

J. Kim et al.: Growth mechanism of YBa2Cu3O7−␦ thin films and precipitates on planar and vicinal SrTiO3 substrates
effort on ultrathin YBCO films is required to understand
not only the film precipitate growth mode but also espe-
cially how to modify it in the very early stages.
YBCO films grew with step bunching and pinning
holes.22,28,31 The areal density of a-axis precipitates was
reduced for the films on vicinal LAO substrates, empha-
sizing that substrate-mediated control of the surface
structure of YBCO films could be achieved by the vicinal
substrate.30,32
Interesting observations have emerged for the YBCO
film growth mechanism on vicinal substrates that have
terraces of low index planes.27 A schematic diagram of
the surface morphology of a vicinal SrTiO3 (STO) sub-
strate is shown in Fig. 1. Vicinal substrates have terrace
structures on the surface, providing preferable attach-
ment sites for the incoming adatoms or particles. As the
vicinal angle increases from 2° to 5°, the length between
two terraces decreases from 111.8 to 44.6 Å. Assuming
the diffusion length of the incoming adatoms or particles
is constant, i.e., the same for 0°, 2°, and 5° miscut sub-
strates, the ultimate location of the incoming materials
could be strongly influenced by the terrace distance, de-
termining the growth mode. Since the surface morphol-
ogy of substrates influences the growth mode and the
resultant microstructure of growing films, the vicinal
substrates have a significant effect on the structure and
resultant electrical transport properties of YBCO films.28
Higher superconducting critical temperature and critical
current density were frequently observed for YBCO
films on vicinal LaAlO3 (LAO) substrates,29 where the
nucleation and growth of YBCO films was aided by the
morphological configuration of the vicinal LAO sub-
strate surface.30 The growth direction of c-axis YBCO
films on vicinal MgO substrates was normal to the sub-
strate surface, while it was normal to the crystallographic
axes of substrates on vicinal STO or LAO substrates.22,30,31
The growth mechanism of YBCO films was changed
from the island or spiral growth mode on planar STO
substrates to the step-flow growth mode on vicinal STO
substrates.22,31 The competition between the step-flow
and island growth was strongly affected by the diffusion
length of arriving particles and the terrace width.30 Due
to the step structure of the vicinal STO substrate surface,
No systematic study has been conducted yet, however,
for the growth mechanism of YBCO precipitates. In par-
ticular, the relation between the film growth mode and
the precipitate in the initial film growth stage has not
been explored. In this study, YBCO films deposited on
(100) STO substrates were studied to determine the re-
lation between the growth mechanism of YBCO films
and the precipitates. The 2° and 5° miscut STO substrates
were used to study the effect of the STO vicinal angle on
the film growth, surface microstructures, and supercur-
rent properties. The growth mode transition of the film
and precipitates was studied at the early growth stage.
Especially, the growth mode relationship between the
films and precipitates was systematically discussed with
various film thicknesses. We also proposed a unique
transition characteristic of YBCO precipitates with in-
crease of the growth temperature. Finally, a possible sub-
strate meditation was suggested to eliminate YBCO
precipitates without degradation of the film electrical
properties.
II. EXPERIMENTAL PROCEDURE
YBCO films were deposited on planar and vicinal
STO substrates by PLD. A commercially fabricated stoi-
chiometric YBCO target (5-cm diameter × 0.5-cm thick-
ness) was used to deposit YBCO on (100)-oriented STO
substrates with 0°, 2°, and 5° miscut along the [010]
direction. The target was located 7 cm away from the
substrate. Each substrate was cleaned in acetone using an
ultrasonic cleaner for 15 min and rinsed in methyl alco-
hol. Silver paint was used to attach the substrates to a
FIG. 1. Schematic diagram of a SrTiO3 vicinal substrate: [hkl]C is a crystallographic direction; [hkl]G is a geometric direction; and ␪ is a
vic
vicinal angle.
J. Mater. Res., Vol. 15, No. 3, Mar 2000
597
http://journals.cambridge.org
Downloaded: 15 Mar 2015
IP address: 165.95.226.100

J. Kim et al.: Growth mechanism of YBa2Cu3O7−␦ thin films and precipitates on planar and vicinal SrTiO3 substrates
III. RESULTS AND DISCUSSION
heater block. Since the growth temperature was meas-
ured under the heater block surface, the real substrate
temperature might be lower than the recorded growth
temperature. For the study of the temperature depend-
ence of the film characteristics, the growth temperature
was varied between 600 and 850 °C by 50 °C. Laser
ablation took places with a KrF (␭ ס 248 nm) excimer
laser (Lambda Physik LPX 300, Go¨ttinger, Germany) at
1.1 J/cm2 with 0.05 cm2 spot size in a growth chamber
containing a background pressure of 400 mtorr of oxy-
gen. After deposition the films were immediately cooled
in the chamber under 1 atm of oxygen pressure. The
average film thickness after 7000 laser shots at 5 Hz was
measured by a profilometer, and the resultant thickness
was 1700 Å. The average thickness of each film was
calculated with the ratio of 0.243 Å/shot, resulting in the
thickness ranging from 7 to 1700 Å. For convenience,
YBCO films on 0°, 2°, and 5° miscut STO substrates
were named as Y0S, Y2S, and Y5S, respectively, and
this notation will be used to represent YBCO films
grown on planar and vicinal surfaces of STO, in general.
The structure of the YBCO films was studied by x-ray
diffraction (XRD) using a Rigaku Rotaflex RU-200B
(Tokyo, Japan) x-ray diffractometer equipped with a Cu
K␣ source. For films grown on vicinal substrates, com-
pensation of the miscut angles, 2° and 5°, is needed in
order to satisfy Bragg conditions for the film planes.
After the initial ␻ scan to obtain the STO substrate (200)
peak at 2␪ angle of 46.483°, the second ␻ scan was
performed to optimize the intensity from YBCO
(005) peak. Then a ␪–2␪ scan was carried out for 1700-
Å-thick films grown at different growth temperatures and
for 7- to 480-Å-thick films grown at 850 °C.
XRD patterns of ␻ scans for the (005) YBCO peak of
1200-Å films are shown in Fig. 2. Peak positions were
observed at 19.22°, 17.36°, and 14.41°, and their peak
widths were 0.19°, 0.48°, and 0.69° for the Y0S, Y2S,
and Y5S films, respectively. The (005) peak of the bulk
YBCO is at ␪ ס 19.256°; thus, these peak positions were
shifted by about 0.036°, 1.896°, and 4.846° for the Y0S,
Y2S, and Y5S films, respectively. It indicated that the
c-axis YBCO films grew along the [001] direction of
STO substrates, even on 2° and 5° miscut vicinal STO
substrates. The full width at half-maximum (FWHM) in-
creased, and the peak intensity decreased with increase of
the vicinal angle. This is in agreement with previous
observations that the FWHM of 1000-Å-thick YBCO
films grown on 0°, 2°, 4°, and 6° miscut (001) MgO
substrates increased with the vicinal angles.34 It is
generally accepted that the defect-free surface of a
single-crystal substrate would provide the best epitaxial
quality of thin films, especially when the film–substrate
lattice mismatch is less than 5%, such as for YBCO
(a ס 3.816 Å, b ס 3.886 Å) on STO (a ס 3.905 Å). The
substrate with higher vicinal angles contains more kink
or terrace sites, providing more defects on its surface.
The more defects the substrate surface contains, the more
defects, such as grain boundaries, pin holes, precipitates,
or secondary phases, the film will have, resulting in the
poorer electrical properties. In our present experiment,
the ␻ scan FWHM of YBCO (005) peaks was propor-
tional to the vicinal angle for 1200-Å-thick films.
A. Growth temperature effect
The surface morphology of the films was studied by
scanning electron microscopy (SEM) using a LEO-
1500 SEM. All micrographs were taken with 30° of
specimen tilt angle to observe the three-dimensional
morphology. To study the growth modes and microstruc-
tures for the early stage film growth, 30- and 60-Å-thick
films were studied by atomic force microscopy (AFM)
using a Digital Instrument Dimension 3000.
The ␪–2␪ XRD patterns for 1700-Å-thick Y0S, Y2S,
and Y5S films with various substrate temperatures rang-
ing from 600 through 850 °C in increments of 50 °C are
The superconducting critical temperature (Tc), the
width of critical temperature transition (⌬Tc), and the
critical current density (Jc) for 1200-Å-thick films were
measured inductively using a contactless method.33 Coil
inductance was measured with temperature in a liquid-
nitrogen tank. To read accurate values of Tc and ⌬Tc, the
second derivatives of coil inductance with respect to the
temperature was taken. Tc was measured from the x in-
tercept of the second derivative, and ⌬Tc was taken from
the width of the second derivative curve at zero. The Jc
measurement was illustrated elsewhere.33 To study the
temperature effect on the supercurrent transport property,
Tc for 1700-Å-thick films was measured as a function of
the growth temperature.
FIG. 2. XRD patterns of ␻ scans at the YBCO (005) peak for the
1200-Å-thick films deposited at 850 °C.
598
J. Mater. Res., Vol. 15, No. 3, Mar 2000
Downloaded: 15 Mar 2015
http://journals.cambridge.org
IP address: 165.95.226.100

J. Kim et al.: Growth mechanism of YBa2Cu3O7−␦ thin films and precipitates on planar and vicinal SrTiO3 substrates
FIG. 3. XRD patterns of ␪–2␪ scans with various growth temperatures for the 1700 Å thick films: (a) Y0S, (b) Y2S, (c) Y5S. (Y, YBa2Cu3O7−␦
;
S, SrTiO3.)
shown in Figs. 3(a)–3(c), respectively. For the Y0S and
Y2S films grown at 600 and 650 °C, the YBCO (005)
and (007) peaks were very weak, while at 650 °C the
YBCO (200) peak appeared strongly, and the (020) peak
existed as a shoulder along the STO (200) peak, indicat-
ing that a-axis-oriented growth was dominant over c-axis
growth. It was well known that a-axis-oriented YBCO
film growth was observed at temperatures ranging from
600 to 650 °C.6,11,35,36 For the Y5S films, YBCO (005)
and (007) peaks were weak for 600 °C. However, as the
growth temperature increased above 600 °C, these peaks
were strongly intensified and no YBCO (200) peak was
observed. It indicates that the 5° miscut vicinal STO sub-
strate would favor c-axis YBCO growth instead of a-axis
growth even at low growth temperatures, such as 650 °C
in this presented deposition conditions.
As the substrate temperature increased to 700 and
750 °C, the intensity of YBCO (005) and (007) peaks
increased and their peak widths decreased for all films. It
was reported that the intensity ratio of the YBCO (00l) to
(h00) peaks, such as (005) to (100), increased with in-
creasing substrate temperature,35 indicating that increas-
ing the growth temperature above 650 °C enhanced the
c-axis orientation and reduced the a-axis orientation.25
The YBCO (200) peak intensity increased with tempera-
ture up to 750 °C and decreased as the temperature in-
creased further to 850 °C. Shi et al.37 observed that
YBCO (200) and (020) peaks strongly appeared at
700 °C, and then they disappeared as the temperature
increased above 760 °C. The a-axis oriented YBCO
films have been fabricated at 750 °C to obtain high-
quality films, and the quality of a-axis-oriented films
degraded due to increase of the c-axis orientation with
the increase of the growth temperature.26,35,38
For the Y0S films grown at 700 and 750 °C, the
YBCO (200) and (005) peaks were narrow and intensi-
fied, indicating that both a-axis- and c-axis-oriented
YBCO structures were equally distributed. As the vicinal
angle increased, however, the intensity of the YBCO
(200) peak decreased but the intensity of YBCO (00l)
peaks were not changed, implying that vicinal STO sub-
strates favor c-axis growth of YBCO rather than a-axis
growth under these deposition conditions. Me´chin et al.22
also observed that all YBCO (h00) peaks disappeared
when YBCO films were deposited on vicinal STO sub-
strates. For YBCO films deposited on (110) NdGaO3
substrates at 760 °C, the intensity ratio of the YBCO
(200) peak to the YBCO (006) peak dramatically de-
creased for a 2° miscut vicinal substrate relative to 0°.24
Therefore, the intensity ratio of the YBCO (00l) peak to
(h00) peak increased as the vicinal angle was increased
in the present deposition conditions. This orientation
change with the vicinal angle was still observed when the
growth temperature was increased up to 850 °C. Espe-
cially, no YBCO (h00) peak was observed on the Y5S
films above 600 °C of the growth temperatures. How-
ever, an unidentified peak at 2␪ ס 43° appeared only
on YBCO films grown on vicinal substrates at 850 °C.
J. Mater. Res., Vol. 15, No. 3, Mar 2000
Downloaded: 15 Mar 2015
599
http://journals.cambridge.org
IP address: 165.95.226.100

J. Kim et al.: Growth mechanism of YBa2Cu3O7−␦ thin films and precipitates on planar and vicinal SrTiO3 substrates
FIG. 4. SEM pictures of the surface morphology with various growth temperatures for the 1700-Å-thick films.
This suggested that the optimum growth temperature
for Y2S and Y5S films could be lower than for the Y0S
film. This will be discussed later in Sec. III. C.
SEM micrographs are shown in Fig. 4 for 1700-Å-
thick Y0S, Y2S, and Y5S films grown at various sub-
strate temperatures. At low temperatures, 600 and
650 °C, all films showed a smooth surface with a few
irregular-shaped precipitates (region A). The size of
these precipitates was smaller than 1000 Å in length, and
the areal density was less than 3.0 × 108/cm2. It was
reported that YBCO films grown at low temperatures,
such as 600 °C, with a-axis orientation had a smoother
600
J. Mater. Res., Vol. 15, No. 3, Mar 2000
Downloaded: 15 Mar 2015
http://journals.cambridge.org
IP address: 165.95.226.100

J. Kim et al.: Growth mechanism of YBa2Cu3O7−␦ thin films and precipitates on planar and vicinal SrTiO3 substrates
surface and fewer precipitates than c-axis oriented films
grown at higher temperatures.11 The irregular- or plate-
shaped precipitates have been observed in a-axis YBCO
films grown at low temperatures, such as 650 °C,25 while
rectangular precipitates were found in a-axis films grown
at higher temperatures, such as 750 °C.38 The Y2S and
Y5S films grown at 600 °C had fewer precipitates than
Y0S films and the areal density was reduced to
1.1 × 108/cm2 for the Y5S film, implying that vicinal
substrates would be useful to reduce the areal density of
the a-axis precipitate.
for the Y5S film. Instead, island-shaped precipitates (is-
lands) appeared on Y0S and Y2S films (region D). The
islands had no directional growth reference along the a or
b axis of YBCO, but they grew along the c axis of the
YBCO film.6,12,23,38 However, it has been also reported
that the islands were crystalline compounds combined of
both epitaxial and nonepitaxial phases as well as of
YBCO a-axis precipitates.23 It was often observed that
one or two acicular precipitates were located at the center
of the island, implying that islands might be transformed
from acicular precipitates that were a-axis directionally
grown and combined with other acicular precipitates (re-
gion E). Since no precipitates were observed on the Y5S
film grown at 800 °C, the high vicinal angle of the STO
substrates provided the potential to prevent the formation
of the island precipitate. Gone et al.11 reported that vici-
nal STO substrates could be useful to suppress the nu-
cleation of precipitates in YBCO thin films.
At 850 °C, acicular precipitates disappeared from all
films, which agreed well with XRD results in Fig. 3.
Both Y2S and Y5S films were completely free of all
types of precipitates, while the Y0S film had more is-
lands. The areal density of islands on Y0S films in-
creased with the temperature, but the size was relatively
constant. Islands have been characterized as Cu-enriched
phases, and the areal density was increased with the
growth temperature.6 On the Y2S film, periodically
aligned wavelike terraces were observed with less than
800 Å of separation, while much rougher and larger
ridges appeared on the Y5S film with around 4000 Å of
separation. The wavelike terraces and ridges have been
reported to be due to the step-flow growth of the films on
the vicinal substrates.22,28,31,32,39 The ridges generally
consisted of several wavelike terraces, and the wave
height represented the height of one unit cell of the
YBCO c axis ( 12 Å).22,28,31,32 It was reported that
the ridges were caused by terraces of the vicinal sub-
strates, and wavelike terraces originated from the
stacking fault due to the mismatch between the STO
unit cell (a ס 3.905 Å) and the YBCO c-axis unit cell
(c ס 11.682 Å).22,28,32 Besides stacking faults, other
secondary phases or impurities could cause the step
bunching, resulting in wavelike terraces on the
films.22,31,39
For the intermediate temperature range, 700 and
750 °C, rectangular precipitates (region B) appeared on
all films. Rectangular precipitates commonly appeared
on YBCO films grown at temperatures ranging from
700 to 800 °C,6,22–24 and they were preferentially ori-
ented along the a axis of the YBCO film.6,23 Me´chin
et al.22 observed similar rectangular precipitates in YBCO
films grown on planar STO substrates at 750 through
800 °C, which were associated with the presence of
a-axis oriented phases confirmed by XRD. From the
XRD data in Fig. 3, the YBCO (200) peak was observed
at temperatures of 700 and 750 °C, suggesting that the
YBCO (200) orientation was likely due, in part, to
the rectangular precipitate. The size of precipitates was in
the range of 1000 through 1300 Å for films grown at
700 °C, but the areal density depended on the vicinal
angle. Precipitates covered the entire surface of the Y0S
film (>3.5 × 1010/cm2), but the areal density was reduced
to 1.2 × 109/cm2 and 9.0 × 108/cm2 for the Y2S and Y5S
films, respectively. The reduction in the areal density of
rectangular precipitates with the vicinal angle was also
noticeable for the films grown at 750 °C. The areal den-
sity rectangular precipitates on the Y0S film grown at
750 °C was higher than 2.2 × 1010/cm2, and it was re-
duced to 9.3 × 109/cm2 and 4.6 × 109/cm2 on the Y2S and
Y5S films, respectively. Furthermore, Y5S films showed
the smallest precipitate size for 700 and 750 °C tempera-
tures, implying that the higher the substrate vicinal angle,
the smaller and fewer the rectangular precipitate. Bo´ıkov
et al.24 observed the same reduction of rectangular pre-
cipitates on YBCO films grown on 2° miscut vicinal
NdGaO3 substrates. For YBCO films grown on planar
and 4° miscut vicinal STO substrates,22 rectangular
precipitates appeared on both films, but the areal density
of precipitates was reduced on the vicinal substrates.
The reduction of rectangular precipitates was also ob-
served for a-axis YBCO films grown on vicinal LAO
substrates.30
The first step-flow growth morphology appeared on
the Y2S film at 700 °C and on the Y5S film at 650 °C,
revealing aligned wavelike morphology. At these low
temperatures, the ridges were not well developed and the
YBCO matrix surface, except for precipitates, was rela-
tively smooth. As the growth temperature increased, the
wavelike morphology became rough ridges, resulting in
rougher film surfaces. The noticeable morphology differ-
ence between films grown on planar and vicinal sub-
strates was that precipitates disappeared only when the
temperature was high (850 °C for Y2S films and 800 °C
At 800 °C, rectangular precipitates became more
acicular for Y0S and Y2S films (region C in Fig. 4).
These acicular precipitates were also reported as a-axis-
oriented and Cu-enriched phases.6,39 The size and areal
density of acicular precipitates were reduced as the vici-
nal angle increased, and no precipitates were observed
J. Mater. Res., Vol. 15, No. 3, Mar 2000
Downloaded: 15 Mar 2015
601
http://journals.cambridge.org
IP address: 165.95.226.100

J. Kim et al.: Growth mechanism of YBa2Cu3O7−␦ thin films and precipitates on planar and vicinal SrTiO3 substrates
for Y5S films) enough to create ridges. It indicated that
a certain thermal energy might be required to fully acti-
vate the step-flow growth, such as 850 °C for the Y2S
film and 800 °C for Y5S, even though the YBCO film
followed the step-flow growth mode on vicinal STO sub-
strates above 650 °C. The presence of rectangular pre-
cipitates on YBCO films on 4° miscut STO substrates
was due to the low temperature,22 and no islands were
observed when the step-flow growth was completely
dominant at high temperature, indicating that the island
morphology did not form when the step-flow growth was
the dominant mode. A detailed explanation will be given
in Sec. III. C.
In this section, the temperature effects on the film
structure and the precipitate morphology on planar and
vicinal STO were discussed. Y0S precipitate morphology
was transformed from the irregular shape, through the
rectangular and acicular shapes, to the island shape with
the growth temperature. The precipitate transition ten-
dency was completely matched with the film structures.
The transition between rectangular and island precipi-
tates was 800 °C where the acicular precipitates were
appeared. The rectangular precipitates, which were well
known as a-axis precipitates, were elongated longer and
became acicular shaped. Two or more acicular precipi-
tates were cross-combined so that the one directional
elongation stopped and the island growth started. It pro-
posed that the acicular precipitates were the intermediate
precipitate shape between the a-axis rectangular precipi-
tate and the c-axis island precipitate. For the films
deposited on the vicinal substrates, a few irregular pre-
cipitates were also observed below 700 °C. For Y2S
films at 700 and 750 °C, the rectangular precipitate areal
density were lower than for the Y0S film precipitates,
indicating that the vicinal STO substrate suppressed the
a-axis rectangular precipitate growth. The acicular pre-
cipitates were observed on the Y2S film at 800 °C, but no
islands formed at 850 °C due to the fully developed step-
flow growth. For the Y5S films, the size and areal density
of rectangular precipitates were very small at 700 °C and
the size became smaller at 750 °C. Then, the rectangular
precipitates could not be combined together due to the
same size and wide separation even after precipitates
were elongated. Therefore, rectangular precipitates of the
Y5S film could not be transformed to islands, so that
there were no islands on Y5S at 800 and 850 °C.
B. Film thickness effect
The ␪–2␪ XRD patterns, in the range of 2␪ from 35° to
57°, for 7- to 480-Å-thick Y0S, Y2S, and Y5S films
grown at 850 °C are shown in Figs. 5(a)–5(c), respec-
tively. No YBCO peaks were observed for any films up
to 30-Å thickness. Ultrathin (<40 Å) YBCO films grown
on STO substrates by sputtering deposition revealed a
pseudomorphic structure, a perovskite-like structure, but
not the exact YBCO structure.40 Zegenhagen et al.41 also
observed that a 40-Å-thick YBCO film on a STO sub-
strate was heavily strained, ⑀ ס −2.0%, due to the pseu-
domorphic growth, while the strain reduced to −0.8% for
FIG. 5. XRD patterns of ␪–2␪ scans with various film thicknesses for the films deposited at 850 °C: (a) Y0S, (b) Y2S, (c) Y5S. (Y, YBa2Cu3O7−␦
;
S, SrTiO3.)
602
J. Mater. Res., Vol. 15, No. 3, Mar 2000
http://journals.cambridge.org
Downloaded: 15 Mar 2015
IP address: 165.95.226.100

J. Kim et al.: Growth mechanism of YBa2Cu3O7−␦ thin films and precipitates on planar and vicinal SrTiO3 substrates
the 80-Å-thick film. They concluded that the critical film
thickness, to form a stable c-axis YBCO film, would be
between 40 and 80 Å. We believed that the film struc-
tures up to 30-Å thickness were not crystallized YBCO
but pseudomorphic structures. Therefore, no YBCO
peaks were observed for films thinner than 60 Å, due to
the initial pseudomorphic film structure and the weak
diffraction signal from the ultrathin layer.
structure became stable without heavy strains as the ini-
tially induced heavy strain was relaxed above a certain
film thickness. Bro¨tz et al.46 also observed the same
strain relaxation for YBCO films on vicinal STO sub-
strates as the film thickness increased to 240 Å. There-
fore, above 110 Å of the film thickness, the effect of the
film thickness was not considerable any more because
most heavy strains were relaxed, so that YBCO XRD
peaks became sharper and intensified with the film
thickness.
The YBCO (005) peak of the 60-Å-thick films was
wider than those of thicker films. The film strain can be
interpreted from the XRD peak line broadening.42–45
Since there were no measurable YBCO peaks up to 30 Å
of the film thickness, the film strain from 60 Å of the film
thickness was calculated with an equation reported by
Hall.43 The calculated strains were 0.081, 0.179, and
0.251 for 60-Å-thick Y0S, Y2S, and Y5S films, respec-
tively. The positive strains indicated that all 60-Å-thick
films were strained by tension. Maurice et al.45 reported
that the heavy strain in a 100-Å-thick YBCO film growth
on a STO substrate was strongly due to alternative mis-
match resulted from both twins and dislocations at the
film–substrate interface during the pseudomorphic initial
growth. Up to 30 Å of film thickness no XRD peaks of
YBCO were observed due to the pseudomorphic struc-
ture, and the crystallized YBCO structure appeared with
very wide XRD peaks from 60 Å. This proposed that
60-Å-thick YBCO films might be still heavily strained.
Therefore, the wide YBCO (005) peaks at 60-Å thickness
might be due to the remaining tensile strain. The (005)
peak intensity of 60-Å-thick films decreased as the vici-
nal angle increased in Fig. 5. The 60-Å-thick film strain
was increased with increase of the vicinal angle, indicat-
ing the YBCO film strain was a function of the STO
vicinal angle. Maurice et al.45 observed that the YBCO
film strain on a STO substrate was due to the twins and
dislocations at STO surface step structures. For YBCO
films on vicinal STO substrates, Bro¨tz et al.46 reported
that the YBCO films on vicinal STO were more strained
by tension than films on a planar STO. Therefore, in the
present experiment, the YBCO film strain at a given film
thickness was proportional to the STO vicinal angle due
to increase of the STO surface step structure.
In Fig. 5, the intensity ratio of the YBCO (005) peak to
the STO (200) peak depended on the film thickness as
well as the vicinal angle for the films thicker than 110 Å.
The ratio increased as the film thickness increased for all
vicinal angles. However, Y5S films showed the highest
intensity ratio for 240- and 480-Å-thick films, while Y0S
films showed the highest ratio for 60- and 110-Å-thick
films. The YBCO growth mode on vicinal STO sub-
strates changed from the island growth to the step-flow
growth mode22,28,31,32,39 above 30 Å of the film thick-
ness (discussed in Fig. 6). The volume fraction of c-axis
in-plane domains in YBCO films thicker than 700 Å in-
creased by deposition on 2.5° miscut vicinal STO sub-
strates,49 and the c-axis growth was preferred to the
vicinal STO substrates for 1700-Å-thick films. It indi-
cates that vicinal STO substrates favor the c-axis YBCO
growth above a certain film thickness providing fully
dominant step-flow growth. Therefore, as films were
thicker than 110 Å, the STO substrate having a higher
vicinal angle provided the stronger YBCO (005) peak.
SEM micrographs for the film thickness ranging from
7 to 480 Å are presented in Fig. 6 for Y0S, Y2S, and Y5S
films grown at 850 °C. Islands nucleated even on the
7-Å-thick films for all YBCO films (region F). Island
nuclei were often observed even from the very first
monolayer (12 Å) of YBCO films.10,40,41,50 A pseudo-
morphic island growth was observed on ultrathin (<40 Å)
YBCO films grown on STO and MgO substrates by sput-
tering deposition.40 CuO, Cu2O, or Y2O3 islands nu-
cleated before YBCO growth in ultrathin films,14,40 and
the precipitation of Cu-rich phases was observed on the
very first unit cell of YBCO films.10 The matrix sur-
rounding precipitates was not the YBa2Cu3O7−␦, but the
pseudomorphic phase in the early stage.10,40,41 Since no
YBCO peaks were observed up to 30 Å of the thickness
by XRD, and since island nuclei with the size of at least
230 Å in diameter were observed by SEM, island nuclei
prevailed against the YBCO crystal structure, leading the
embryonic film structure in the very early stages of
YBCO nucleation and growth.
When the film thickness was 110 Å, both (005) and
(007) peaks were clearly observed in Fig. 5. These peaks
became sharper and more intensified as the film thick-
ness increased and were strongly developed when the
thickness reached 480 Å. As the film thickness reached
110 Å, the film strains were relaxed to 0.006, 0.015, and
0.018 for Y0S, Y2S, and Y5S films. Above 110 Å of
YBCO film thickness, the film strains did not change
significantly but continuously decreased with increase of
the film thickness, so that strains were less than 0.007 for
all films above 240 Å of the film thickness. The strain
relaxation was strongly dependent on the film thickness
in the initial growth stage,45,47,48 implying that the film
It was also noteworthy that the areal density of island
nuclei increased with the thickness up to 30 Å for the
Y0S films, but the size was relatively constant, indicating
that only nucleation of islands took places in this thick-
ness range. There was no significant change in the areal
J. Mater. Res., Vol. 15, No. 3, Mar 2000
Downloaded: 15 Mar 2015
603
http://journals.cambridge.org
IP address: 165.95.226.100

J. Kim et al.: Growth mechanism of YBa2Cu3O7−␦ thin films and precipitates on planar and vicinal SrTiO3 substrates
FIG. 6. SEM pictures of the surface morphology with various film thicknesses for the films deposited at 850 °C.
density or size of island nuclei for Y2S and Y5S films,
indicating that the nucleation of island possibly occurred
on YBCO films grown on vicinal STO substrates in the
very early stage, but somehow it was restricted in a cer-
tain nucleus density. This will be discussed in more
length in Sec. III. C.
As the film thickness increased above 30 Å, the ten-
dency of the areal density and size of the island nuclei
was completely opposite to thinner Y0S films. The areal
density of precipitates no longer increased above 30 Å in
thickness, but it was suddenly reduced and remained
steady as the thickness increased up to 1700 Å (see the
604
J. Mater. Res., Vol. 15, No. 3, Mar 2000
Downloaded: 15 Mar 2015
http://journals.cambridge.org
IP address: 165.95.226.100

J. Kim et al.: Growth mechanism of YBa2Cu3O7−␦ thin films and precipitates on planar and vicinal SrTiO3 substrates
picture of Y0S grown at 850 °C in Fig. 4). The size of
Y0S islands, however, increased continuously with the
thickness above 30 Å. In Fig. 4, only islands appeared on
the Y0S film grown at 850 °C. Since similar precipitates
formed from the very early stage10,14,40,50 and since the
size increased with the film thickness,10,14,50 the small
precipitates in Y0S films might be nuclei of the islands.
The size of Cu, Cu2O, or Y2O3 islands in ultrathin YBCO
films increased with the YBCO film thickness.50 Gong
et al.11 observed a similar result that islands nucleated at
the initial stage of the film growth and grew larger with
the film thickness, keeping the same areal density for the
films thicker than 500 Å. Kanda et al.10 also observed the
decrease of the precipitate areal density from AFM im-
ages. The areal density of the precipitates was greatly
reduced from 2.8 × 108/cm2 to 2.0 × 107/cm2 as the film
thickness reached 3 unit cells ( 36 Å) and kept the con-
stant areal density above 36 Å, while the size increased
linearly.
An interesting point for the Y0S films was that the
sudden change in the precipitate areal density occurs si-
multaneously with the change in size. The areal density
and size of the precipitates versus the film thickness of
Y0S films is shown in Fig. 7. The areal density increased
significantly from 7.0 × 107/cm2 to 3.8 × 109/cm2, but
the average diameter was relatively constant, keeping
0.8 ␮m for the thickness up to 30 Å. As the film thick-
ness increased further, however, the areal density de-
creased significantly to 3 × 108/cm2 and the average
diameter increased up to 0.39 ␮m in diameter for the
films thicker than 30 Å. It was also clearly noticeable
that there was the structural change between 30- and
60-Å-thick films from the XRD patterns of Fig. 5, indi-
cating that the Y0S film structure change was accompa-
nied with the surface morphology change.
For the Y2S and Y5S films, the areal density of the
precipitate nuclei was less than for Y0S films, and there
was no noticeable increment in areal density and size of
the nuclei with thickness up to 30 Å. Besides, there was
no noticeable difference in the film morphology. There-
fore, the growth mode of YBCO films grown on vicinal
substrates might involve the island growth mode at the
very beginning growth stage, at least up to 30 Å of the
thickness in this experiment (see the detail discussion in
Sec. III. C. The only difference of the island nuclei on
vicinal substrates from the Y0S films was that the areal
density of small precipitates was relatively constant with
film thickness. As seen in the previous section, islands
grew preferentially on the planar substrates than on the
vicinal substrates, implying that vicinal STO substrates
would prevent islands from growing further due to the
step-flow growth of the film matrix. Therefore, for the
films grown on vicinal STO substrates, nucleated islands
could not grow any more but just kept the same areal
density and size.
The 60-Å-thick Y2S and Y5S films showed the ridge
microstructures, and ridges increased in size with film
thickness. The ridge morphology was formed normal to
the direction of the STO vicinal angle due to the step-
flow growth mode.22,28,32,51 For thick YBCO films
(>1000 Å), the steps appeared as straight lines on an-
nealed STO vicinal substrates28,31,32 and irregularly on
unannealed STO vicinal substrates.22 Even on the an-
nealed vicinal substrates, irregular ridges were found on
thinner, 170-Å-thick, YBCO films.51 Especially for the
Y2S films thicker than 30 Å, tiny exclusions were ob-
served in front of the ridges. Mukaida et al.51 also ob-
served same exclusions at ridge edges of YBCO films
grown on a 1.7° miscut NdGaO3 vicinal substrate. They
reported that these exclusions might be caused by Cu-
rich phases and segregated from the step-flow growth,
which disturbed the further motion of the step-flow
growth, resulting in the step bunching. However, for the
thick films, such as 1700-Å-thick Y2S and Y5S films
grown at 850 °C in Fig. 4, no exclusions were observed
while the step bunching was revealed more clearly. The
most interesting observation in the film thickness study
was that no precipitates were observed on Y2S or Y5S
films thicker than 30 Å. Their disappearance occurred
when the step-flow growth was fully dominant. As the
film thickness increased up to 30 Å on vicinal STO sub-
strates, island nuclei existed without significant change
in size or areal density, implying that the island growth
mode was also involved in the early growth stage. When
the film thickness increased above 30 Å, however, the
film growth was changed to the step-flow growth mode
and no islands were observed. Therefore, islands existed
only when the YBCO film followed the island growth
mode and could not grow any further when the step-flow
growth mode was dominant. This conclusion is consis-
FIG. 7. Areal density and size of islands with various film thicknesses
for the Y0S films deposited at 850 °C.
J. Mater. Res., Vol. 15, No. 3, Mar 2000
Downloaded: 15 Mar 2015
605
http://journals.cambridge.org
IP address: 165.95.226.100

J. Kim et al.: Growth mechanism of YBa2Cu3O7−␦ thin films and precipitates on planar and vicinal SrTiO3 substrates
tent with the discussion of the previous section, the tem-
perature dependency, in that all precipitates disappeared
when the step-flow growth mode was completely domi-
nant at 850 °C. In addition, the results in Fig. 5 indicated
that islands grew easily on the planar substrate that pro-
vided the island growth mode. Further detailed discus-
sion will be given in the next section.
While two growth modes mentioned above were
widely studied for YBCO films with a certain thickness,
many publications have been focused on the growth
mode for the very early stage of deposition.15,16,20,55,56
Baronnet56 reported that screw dislocations could occur
during the initial stage and the coalescence stage, imply-
ing the possibility of the spiral growth in the initial stage.
However, since screw dislocations would be formed dur-
ing island coalescence generated by the thermal energy,
i.e., the growth temperature,17,57 the island growth mode
might be followed by the spiral growth, even in the ini-
tial stage. In the initial stage of the YBCO film growth,
the island growth mode has been more frequently
observed than the spiral growth.14–16,20,21,55 Most
observations were the 2D island mode, except two ref-
erences.14,21 Using reflection high-energy electron-
diffraction (RHEED), a fundamentally different mode,
layer-by-layer growth mode, was inferred for YBCO
films thinner than 150 Å grown on (100) STO and (100)
MgO substrates by reactive evaporation.19 Chan-
drasekhar et al.20 used similar reactive evaporation to
deposit a YBCO film on a (100) STO substrate and ana-
lyzed with RHEED, but the layer-by-layer model was not
observed. The layer-by-layer mode may not be suitable,
because YBCO itself is a complicated four-component
compound and the unit cell is anisotropic and atomisti-
cally inhomogeneous.21 In order to study the growth
mode of nuclei in the initial stage of the present experi-
ment, AFM was used for 30- and 60-Å-thick films
(Fig. 8). If it was a 2D island, 2-dimensional steps should
C. Growth mode
1. On planar STO substrates
The growth mode of the c-axis-oriented YBCO films
on planar substrates has been routinely studied. The is-
land growth13,15,18 and the spiral growth16–18 are the gen-
eral growth modes for YBCO films above a certain
thickness (>50 Å). The island growth mode can be di-
vided into two types, a 2-dimensional nucleation and
growth (2D island growth)15 and an irregular-shaped or
round-shaped island growth (3D island growth).13 Even-
tually these two types are similar except slightly different
shapes; the former has the step-shaped islands and the
latter has no steps. The spiral growth mode was caused
by screw dislocations in YBCO films, therefore some-
times named as the screw dislocation growth.16,17,52 In
many cases, the 2D island growth has been easily ob-
served on YBCO films grown by PLD,13,15,53 and the
spiral growth noticed occasionally from liquid-phase epi-
taxy,52 thermal coevaporation,54 or magnetron sputter-
ing.17,21 However, there is no strict relation between the
growth technique and growth mode.
FIG. 8. Three-dimensional AFM images for the 30- and 60-Å-thick films deposited at 850 °C: (a) 30-Å-thick Y0S, (b) 30-Å-thick Y2S, (c)
30-Å-thick Y5S, (d) 60-Å-thick Y0S, (e) 60-Å-thick Y2S, (f) 60-Å-thick Y5S.
606
J. Mater. Res., Vol. 15, No. 3, Mar 2000
http://journals.cambridge.org
Downloaded: 15 Mar 2015
IP address: 165.95.226.100

J. Kim et al.: Growth mechanism of YBa2Cu3O7−␦ thin films and precipitates on planar and vicinal SrTiO3 substrates
appear around each nucleus. However, no steps around
nuclei were observed on the 30- or 60-Å-thick Y0S film,
indicating that the island nucleus growth in the initial
stage was the 3D island mode.
island, the secondary nucleus coalesced rapidly, forming
a much larger island. Furthermore, the broad (005) and
(006) XRD peaks indicated that the 60-Å-thick Y0S film
turned into the proper YBCO characteristics. Therefore,
the morphology change of 30- and 60-Å-thick Y0S films
might be due to the island coalescence and the transfor-
mation from the embryonic structure into the matured
YBCO film structure.
In general, the film growth stages could be divided
into four steps: the island stage; the coalescence stage;
the channel stage; the continuous growth stage.58 During
the island stage, the initial stage, nucleation and nuclei
growth occurred. As the size of subcritical embryos
reached the critical size for the nuclei growth, nuclei grew
larger and the coalescence stage would start. For the Y0S
films, until the film thickness reached up to 30 Å, the
areal density of the island nuclei increased, but the size
was relatively constant. A 40-Å-thick YBCO film depos-
ited on a (001) STO substrate was a pseudomorphic
structure and highly strained (⑀perpendicular ס −2.0%).41
As the film thickness reached 80 Å, the strain was
dramatically relaxed (⑀perpendicular ס −0.8%), indicating
that the critical thickness between the initial and coales-
cence stages could be below 80 Å. Let us apply the same
concept used for the initial nucleation and grain growth
mechanism. Assuming the island as a grain in the present
experiment, the critical YBCO film thickness might be
between 30 and 60 Å and the initial grain growth (the
initial island growth in the present experiment) started
above 30 Å. Therefore, during the initial stage, below
60 Å of the film thickness, island nuclei were formed and
distributed over the Y0S films.
During the initial stage, no YBCO peaks appeared due
to the embryonic YBCO structure (see XRD data in
Fig. 5). Once the nucleus reached the critical size at the
certain film thickness (>30 Å), the island nucleus grew
larger and YBCO (005) and (006) peaks were detected. It
indicated that the embryonic film structure transformed
into the proper YBCO structure. As the XRD data indi-
cated the structure change between 30- and 60-Å-thick
films, the sudden change of the areal density, size, and
shape of the island nuclei also confirmed the film struc-
ture change with the film thickness. From the AFM pic-
ture for the 60-Å-thick Y0S film, most initial nuclei
disappeared and another type of island nucleation
emerged. The first type island nucleus was a sphere
shape, but the second type was irregular. The second type
nuclei were much larger in size but lower in areal density
than the first type nuclei, while the height did not change.
Once initial nuclei distributed up to a certain nucleus
areal density, the coalescence might occur due to the
shortened distance between two neighbor nuclei provid-
ing easier diffusion into each other.58 In general, initial
nuclei were spherical during the event. After the coa-
lescence took place, the spherical shape became irregular
due to the rearrangement of crystallographic structure
and the coalescence of diffused neighbor islands. During
the coalescence stage, a secondary nucleus grew until it
touched a neighbor island, and if it diffused to a larger
Since the theory of the nucleation and coalescence is
suitable for the crystal or grain growth mechanism, the
question in this issue is whether the YBCO island would
follow the grain growth mechanism or not. In many pub-
lications, the island morphology has been named as a
precipitate or an outgrowth. Unlike rectangular precipi-
tates on films grown at 750 °C (Fig. 4), island precipi-
tates usually were accompanied on c-directionally grown
YBCO films.6,23,38 Another particular c-axis precipitate
was the spiral precipitate caused by the screw disloca-
tion, and the spiral precipitate was frequently reported as
the origin of the YBCO film growth.16–18,21,52 The
misorientation between precipitates and the YBCO film
matrix caused the nucleation of screw dislocations in
the growth of synthetic micas, resulting in the spiral
growth.56 Catana et al.23 reported that island precipitates
were located on and correlated with the intersection of
screw dislocations, generating the spiral growth. Interfa-
cial energy and stresses between island precipitates and
the YBCO film matrix energetically provided the driving
force for the nucleation of the spiral growth. Surface
precipitates contained compressive stress that could be
relaxed by diffusing atoms to unstressed matrix area. The
2D island was also pointed as the growth origin of YBCO
films grown by PLD and direct-current (dc) sputtering
deposition.53 Most substrates, such as a STO substrate
having a lattice mismatch of greater than 1% with respect
to the bulk YBCO, would provide the 2D island nuclea-
tion followed by the island coalescence.30 The surface
morphology of the 1700-Å-thick Y0S film [Fig. 9(a)]
showed circular ripples around the island which might be
evidence of diffused materials between the island and the
YBCO film matrix. Therefore, the island morphology
was the origin of the island growth in Y0S films and
followed the conventional grain growth theory, the nu-
cleation and coalescence.
As the Y0S film thickness increased above 30 Å, the
size of islands became larger and the areal density de-
creased due to the continuous island coalescence. Most
of small secondary nuclei on the 60- and 110-Å-thick
Y0S films disappeared as the film thickness increased,
due to the continuous coalescence process. The size of
the islands increased continuously with film thickness
while the areal density decreased slightly. However,
we could not completely eliminate the islands up to
1700 Å of the Y0S film thickness and even for the much
thicker films.
J. Mater. Res., Vol. 15, No. 3, Mar 2000
Downloaded: 15 Mar 2015
607
http://journals.cambridge.org
IP address: 165.95.226.100

J. Kim et al.: Growth mechanism of YBa2Cu3O7−␦ thin films and precipitates on planar and vicinal SrTiO3 substrates
FIG. 9. Surface morphology of the 1700-Å-thick films deposited at 850 °C: (a) Y0S, (b) Y2S, (c) Y5S.
2. On vicinal substrates
step, might migrate to the steps rather than nucleate is-
lands.60 Only other particles located far away from the
steps would form their own islands by themselves be-
cause the distance between steps and particles was longer
than the diffusion length of particles. It indicated that,
during the very first growth stage of the films on vicinal
substrates, the initial step-flow growth might be accom-
panied with the island nucleation. Biggers et al.29 also
observed islands as well as step-flow growth images of a
YBCO film grown on a vicinal LAO substrate. As par-
ticles migrate to one step, neighbor steps will also com-
bine with other particles so that the initial distance
between steps and particles will be conserved. Unless the
distance increases, the island nucleus areal density will
not increase. Also the step-flow growth would be pre-
ferred with the film thickness due to the reduction of the
remained small lattice mismatch.59 Therefore, the areal
density of island nuclei of Y2S and Y5S films was less
than that of Y0S films and relatively constant with the
film thickness.
The vicinal substrates provide preferred nucleation
sites of surface steps such as terraces or kink sites having
higher binding energy due to a higher coordination.59 If
arriving particles have enough surface mobility because
of their impinging kinetic energy and growth tempera-
ture, and if the diffusion length is longer than about
one-half distance between the surface steps of vicinal
substrates, those particles will likely move to the step
sites which promote the step-flow growth mode. Once a
proper film thickness is built up over the substrate sur-
face, the step-flow growth occurs preferably because the
remaining small lattice mismatch might be completely
eliminated.59 Other particles farther away from the step
will follow the island growth mode by themselves. In this
case, the island nucleation would occur between steps of
a vicinal substrate due to the suppression of surface dif-
fusion, the increase of supersaturation of adsorbing par-
ticles, or the wide step distance.60
In present paper, up to 30 Å of the film thickness, the
island nuclei appeared on all films, even the Y2S and
Y5S films (Fig. 4). However, the island areal density of
Y2S and Y5S films was significantly different from the
areal density of Y0S films. In order to study the effect of
vicinal angles on the island nucleation, the areal density
of the island nuclei with vicinal angles was calculated
from the SEM pictures and plotted as a function of the
film thickness in Fig. 10. The island densities of Y2S and
Y5S films were relatively constant with the film thick-
ness while the areal density on Y0S films significantly
increased from 3.8 × 108/cm2 to 2.5 × 109/cm2.
As discussed previously, the areal density of island
nuclei on Y2S films increased due to the continuous
formation of the island nuclei. In the case of Y2S and
Y5S films grown on vicinal substrates, however, par-
ticles located closer to steps, so that their diffusion length
was longer than the distance between the particle and the
The lowest island areal density was always observed
from Y5S films: the average densities as the function of
film thickness from 7 through 30 Å were 1.5 × 109/cm2,
2.6 × 109/cm2, and 2.3 × 109/cm2 for the Y0S, Y2S, and
Y5S films, respectively. Since a different vicinal angle
provides a different step distance, and since the diffusion
length of particles should be identical under same depo-
sition conditions, the areal density of the island nuclei
might be a function of the vicinal angle. Haage
et al.28,31,32 reported that the surface image of YBCO
films grown on vicinal substrates was a function of the
substrate vicinal angles, miscut direction, and growth
temperature. As the vicinal angle increased, the step dis-
tance decreased so that the probability to nucleate islands
was reduced due to the shortened distance between steps
and particles. For 30-Å-thick Y2S and Y5S films, island
nuclei were aligned due to the direction of the terrace on
vicinal substrates [see Figs. 8(b) and 8(c)]. The row dis-
608
J. Mater. Res., Vol. 15, No. 3, Mar 2000
Downloaded: 15 Mar 2015
http://journals.cambridge.org
IP address: 165.95.226.100

J. Kim et al.: Growth mechanism of YBa2Cu3O7−␦ thin films and precipitates on planar and vicinal SrTiO3 substrates
length of incoming YBCO particles was constant at
given deposition conditions, the step-flow growth would
occur on the vicinal STO substrates having the one-half
terrace distance shorter than the YBCO particle diffusion
length. As shown in Fig. 1, the average terrace distance
of 2° miscut STO substrates is about 112 Å. Therefore,
the estimated diffusion length of YBCO particles might
be longer than 56 Å at given deposition conditions in this
experiment.
As previously discussed in Fig. 4, the step-flow
growth did not occur on the Y2S and Y5S films grown at
600 °C. The aligned wavelike terraces appeared on the
Y2S film at 700 °C and on the Y5S film at 650 °C, while
fully developed step-flow morphology was observed at
850 °C for the Y2S film and 800 °C for the Y5S film.
The certain kinetic energy and thermal energy are re-
FIG. 10. Areal density of island nuclei with various film thicknesses
quired for the incoming YBCO particles to migrate to-
in the initial growth stage.
ward their preferable sites.59 The kinetic energy was
constantly given by the laser energy density (1.1 J/cm2)
and the laser spot size (0.05 cm2) in this experiment. The
tance of aligned islands of the Y5S film was less than that
thermal energy was varied from 600 to 850 °C by 50 °C
for the growth temperatures, indicating that the total en-
ergy for the YBCO particle migration was determined by
the growth temperatures. At 600 °C, the step-flow did not
occur due to the lack of the applied thermal energy. Al-
though the first step-flow morphology appeared on the
Y2S and Y5S films at 700 and 650 °C, respectively,
these growth temperatures were not enough to fully pro-
ceed the step-flow growth. As the growth temperatures
increased, the film surface morphology became rougher,
indicating that ridges became larger due to the activated
step-flow growth. At high temperatures, 850 °C for the
Y2S film and 800 °C for the Y5S film, the ridges were
shown very clearly with many wavelike terraces. There-
fore, a certain growth temperature was required for the
fully-developed step-flow growth.
Kink or terrace sites can be easily created by a small
additional vicinal angle, producing finite and discontin-
ued terraces, resulting in structural defects in the films.
Haage et al.28,31 reported that the elongated cigarlike
growth of YBCO films on vicinal STO substrates was
observed on kink sites and terraces of vicinal substrates.
For YBCO films grown on a pre-annealed 10° miscut
(100) STO substrate, the width of the film terrace was
six times wider than the substrate terrace. Since the one
unit cell height of STO (3.905 Å) is around one-third of
the YBCO c axis (11.682 Å), the film itself might have
defects, such as stacking faults.31 The cross-sectional
TEM analysis revealed antiphase boundaries inside the
YBCO film, due to stacking faults. The YBCO lattice
was strongly disrupted: CuO planes were frequently dis-
continued, overlapped, bent, or absent. These defects
were aligned transversely to the direction of STO ter-
races, and they were well-continued longitudinally, re-
sulting in ridges on the films.
of the Y2S film due to the reduced terrace width of the
vicinal substrates. Therefore, the areal density of island
nuclei was inversely proportional to the vicinal angle of
substrates.
As the thickness of Y2S and Y5S films increased
above 30 Å, all island nuclei disappeared and typical
step-flow images were observed (Fig. 6). Assuming that
the island nucleus was sphere-shaped and the height
above the film surface was the half of the diameter, each
nucleus emerges up to 500 Å of the height above the film
surface. Also from the AFM images (Fig. 8), the height
of islands was around 450 Å for 30-Å-thick Y2S and
Y5S films. As the film thickness increased to 60 Å by an
additional 30 Å, all islands disappeared and the peak to
valley height of the film ridges was less than 200 Å.
Therefore, the film surface roughness decreased and the
step-flow image on films developed as the film thickness
increased to 60 Å, indicating that island nuclei diffused
into neighbor steps and joined the step-flow growth.
In our experiment, the step-flow growth was accom-
panied with island nucleation up to 30 Å of film thick-
nesses for Y2S and Y5S films. However, above the 30 Å,
only the step-flow growth was observed on the films and
all islands defused into the film matrix and follow the
step-flow growth mode. The step-flow growth would
occur much preferably above a certain film thickness
because the remaining lattice mismatch might be com-
pletely eliminated,59 indicating that a certain film thick-
ness was required for the fully developed step-flow growth
even though it could start from the very early stage.
Since the criterion, whether the step-flow growth
would occur or not, is determined by the distance be-
tween two terraces of vicinal substrates, the diffusion
length of incoming YBCO particles can be estimated by
the one-half terrace distance. Assuming that the diffusion
J. Mater. Res., Vol. 15, No. 3, Mar 2000
Downloaded: 15 Mar 2015
609
http://journals.cambridge.org
IP address: 165.95.226.100

J. Kim et al.: Growth mechanism of YBa2Cu3O7−␦ thin films and precipitates on planar and vicinal SrTiO3 substrates
The typical step-flow microstructures of 60-Å-thick
Y2S and Y5S films increased in size as the thickness
increased, aligning normal to the direction of the STO
vicinal angles (Fig. 6). For 1700-Å-thick Y2S and Y5S
films grown in Fig. 9, each ridge contained a number
of small and narrow wavelike terraces. Similar micro-
structure was also found in YBCO films grown on non-
annealed 4° and 6° miscut STO substrates.22 Each lateral
ridge contained several individual wavelike terraces of
which the width was a function of the vicinal angle of
STO substrates. The height of the individual wavelike
terraces was close to the YBCO c-axis unit cell length,
implying that the growth mode of the YBCO films was
changed from the island growth to step-flow growth mode.
In this experiment, therefore, the step-flow growth
mode was accompanied with the island nucleation and
growth mode in the very early stage, up to 30 Å of the
film thickness. As the films are thicker than 30 Å, all
island nuclei diffused into the film matrix and joined the
step-flow growth so that the film growth mode was com-
pletely governed by the step-flow mode. As the vicinal
angle increased, the step-flow growth occurred more
preferably. In addition, a certain growth temperature,
850 °C for the Y2S film and 800 °C for the Y5S film,
was also required for the thermal energy to fully proceed
the step-flow growth.
Y5S transport properties was due to the degraded film
structure that was indicated by the wide FWHM for the
YBCO (005) ␻-scan peaks in Fig. 2.
Since the growth temperature is the major factor to
influence the supercurrent transport properties, the criti-
cal transition temperatures of 1700-Å-thick films are
plotted in Fig. 11 as a function of growth temperatures in
the range of 700–850 °C. At 700 °C, films showed the
lowest Tc (83.3, 81.5, and 84.1 K for Y0S, Y2S, and
Y5S, respectively). The highest Tc appeared at 850 °C for
Y0S and Y2S films (89.9 and 89.9 K, respectively) and
at 800 °C for the Y5S film (89.1 K). As the temperature
increased, Tc increased continuously, except Y5S films.
The higher growth temperature provided the higher Tc
value up to 840 °C due to the reduction of the degree of
random orientation with increase of the growth tempera-
ture.37 Tc of the Y2S film was always lower than that of
the Y0S film through all range of growth temperatures,
and the difference between two Tc values of Y0S and
Y2S films decreased from 1.9 K at 700 °C to <0.1 K at
850 °C. The lower Tc of Y2S films might be due to the
disordered film structure caused by the surface defects,
such as terraces or kink sites. As the growth temperature
increased, however, the thermal energy becomes high
enough to grow YBCO films c directionally even on the
2° miscut substrate. Even on (110) and (111) ZrO2 sub-
strates with large lattice mismatch, YBCO films grew
preferably c-axis oriented with high Tc when the growth
temperature was high enough.37 Therefore, Tc of Y0S
and Y2S films at 850 °C was almost identical because the
thermal energy was high enough to overcome the in-
creased lattice mismatch by 2° of the miscut angle.
A 6° miscut LAO substrate resulted in higher Tc and Jc
of YBCO films than a planar LAO substrate29 while
there was no relation between Tc and vicinal angles for
YBCO films on 0°, 2°, 4°, and 6° miscut STO22 or
D. Supercurrent transport property
Supercurrent transport properties, Tc, ⌬Tc, and Jc, for
1200-Å-thick films grown at 850 °C are listed in Table I.
The Y0S film showed highest Tc (89.7 K) and Jc
(1.4 MA/cm2) with narrowest ⌬Tc (0.5 K), whereas the
Y5S film showed lowest Tc (88.9 K) and Jc (0.5 MA/cm2)
with widest ⌬Tc (0.6 K). However, the properties of the
Y2S film were very close to those of the Y0S film. For
2000-Å-thick YBCO films grown on vicinal (001) MgO
substrates, there was no significant effect of vicinal angle
up to 2° on Tc and Jc.34 As the vicinal angle reached 4°,
Jc dropped dramatically. Tc also decreased when the vici-
nal angle was above 4°. It resulted from the degraded
c-axis alignment that was noticeable from the FWHM of
YBCO (005) peak. The 1600-Å-thick YBCO films on
vicinal MgO substrates also showed the significant de-
grade of supercurrent transport properties above 4° of the
substrates vicinal angle.61 Therefore, the degradation of
TABLE I. Supercurrent transport properties for the 1200-Å-thick
films deposited at 850 °C.
Film
Tc (K)
⌬Tc (K)
Jc (MA/cm2)
Y0S
Y2S
Y5S
89.7
89.7
88.9
0.5
0.5
0.6
1.4
1.5
0.5
FIG. 11. Supercurrent transition temperatures with various growth
temperatures of the 1700-Å-thick films.
610
J. Mater. Res., Vol. 15, No. 3, Mar 2000
http://journals.cambridge.org
Downloaded: 15 Mar 2015
IP address: 165.95.226.100

J. Kim et al.: Growth mechanism of YBa2Cu3O7−␦ thin films and precipitates on planar and vicinal SrTiO3 substrates
MgO34 substrates. Lower Tc of films on vicinal sub-
TABLE III. Effects of the film thickness on YBCO growth mode.
strates than on planar substrates was also observed else-
Film thickness
where.31 Therefore, no strict relationship between the
Film
Y5S
7–30 Å
60–1700 Å
vicinal angle and Tc has been proven yet. In this pres-
ent experiment, surprisingly, the Y5S film provided
higher Tc than other films in the temperature range of
700 through 800 °C, while Y2S films showed lower Tc
than Y0S films. The growth temperature proving the
highest Tc of Y5S films was 800 °C instead of 850 °C,
indicating that the optimum growth conditions for the
Y5S film, at least the growth temperature, could be dif-
ferent from those of the Y0S film. It might be due to the
change of the film growth mode on vicinal substrates.
There is no clue for the Tc versus vicinal angle relation-
ship, because the highest Tc of the Y5S film grown at
800 °C was still lower than that of the Y0S or Y2S film
grown at 850 °C. It also implied that the optimum growth
temperature of YBCO films could be reduced by the
vicinal substrates. This typical Tc characteristic on vici-
nal STO substrates in this experiment may lead the ex-
planation why c-axis XRD peaks were stronger and
a- and b-axis peaks were weaker for YBCO films on
vicinal substrates than on planar substrates in Fig. 3.
Island nucleation &
step-flow nucleation
Island nucleation &
step-flow nucleation
Island nucleation
Step-flow growth
Y2S
Y0S
Step-flow growth
Island growth
growth mode transition was summarized in Table III. A
critical transition film thickness between the initial stage
and the later stage was between 30 and 60 Å in presented
growth conditions. For the Y0S films, only the island
growth mode was observed from the initial stage. The
areal density of island nuclei increased with the film
thickness, but the size was conserved in the initial stage
(the island nucleation stage), while the areal density was
conserved and the size increased with the film thickness
in the later stage (the island coalescence and growth
stage). The island was the source of the island growth of
the YBCO film. For the Y2S and Y5S films, the step-
flow growth and island growth modes competed each
other in the initial stage, resulting in low areal density of
the island nucleus without further distribution. In the
later stage, above 30 Å of the film thickness, the step-
flow growth was dominant and the island disappeared,
indicating that the island could not grow further under
the step-flow growth mode. The higher critical supercur-
rent transition temperature (Tc) and critical supercurrent
density (Jc) and the narrow transition temperature width
(⌬Tc) were observed from the film grown on planar sub-
strate. Below 850 °C growth temperatures, however, the
film grown on the 5° miscut substrate provided higher Tc
than others, implying that the growth condition was dif-
ferent for the film on the 5° miscut substrate.
E. Summary
The effects of the vicinal angle, growth temperature,
and film thickness on the film growth mode, precipitate
growth mode, and electrical transport properties of
YBa2Cu3O7−␦ films deposited on (100) SrTiO3 sub-
strates by pulsed laser deposition have been demon-
strated. For the 1700-Å-thick films, the c-axis growth
tendency was proportional to the vicinal angle and a-axis
precipitates reduced as the vicinal angle increased at
given growth temperature, indicating that the step-flow
growth might reduce the a- or b-oriented phases and
precipitates. The precipitate shape transition for the
1700-Å-thick films with the growth temperature was
summarized in Table II. The transition precipitate be-
tween a-axis rectangular and c-axis island precipitates
was the acicular precipitates that were formed by elon-
gated rectangular precipitates at 800 °C. Due to the fully
activated step-flow growth, no precipitates were ob-
served on the Y2S film T 850 °C and on the Y5S film
above 750 °C. The film thickness effect on the film
IV. CONCLUSION
The most interesting issue discussed in this paper was
the existence of a clear critical transition film thickness
between two island growth stages, the initial nucleation
and the later coalescence and growth stage for the
c-oriented YBa2Cu3O7−␦ films grown on planar sub-
strates. The growth stage transition was the same point
for the YBa2Cu3O7−␦ films grown on vicinal SrTiO3 sub-
strates whether the island growth continues or stops fur-
ther growth due to the terrace structure of the substrate
surface. Especially, two growth modes, the island and
step-flow modes, co-existed in the initial stage for the
c-oriented films grown on vicinal substrates, and only the
step-flow mode remained in the later stage, resulting in
no island on the film surface. Little imperfections on the
substrate surface, e.g., steps or terraces, ultimately influ-
TABLE II. Effect of the growth temperature on precipitate shape.
Growth temperature
Film
600–650 °C
700–750 °C
800 °C
850 °C
Y5S
Y2S
Irregular
Irregular
Rectangular
Irregular
No precipitate
Acicular &
island
No precipitate
No precipitate
Y0S
Irregular
Irregular
Acicular &
island
Island
J. Mater. Res., Vol. 15, No. 3, Mar 2000
611
http://journals.cambridge.org
Downloaded: 15 Mar 2015
IP address: 165.95.226.100

J. Kim et al.: Growth mechanism of YBa2Cu3O7−␦ thin films and precipitates on planar and vicinal SrTiO3 substrates
12. M.G. Norton, R.R. Biggers, I. Maartense, E.K. Moser, and J.L.
Brown, Physica C 233, 321 (1994).
13. S.Y. Xu, C.K. Ong, and X. Zhang, Solid State Commun. 107, 273
enced the initial YBa2Cu3O7−␦ film growth mode so that
the characteristics of YBa2Cu3O7−␦ film properties and
surface morphology were controlled by the substrate sur-
(1998).
face structure. A precipitate morphology transition with
growth temperature was another unique observation
which had not been reported yet elsewhere. The acicular
precipitates were the transition morphology between
a-axis rectangular and c-axis island precipitates on a
YBa2Cu3O7−␦ film grown on a planar SrTiO3. The re-
duction of a-axis precipitates and disappearance of island
precipitates on vicinal substrates due to step-flow growth
were remarkable results for the study on the film surface
precipitates. It might be possible that there is an optimum
vicinal angle, film thickness, or growth temperature
which will eliminate precipitates without degrading su-
percurrent transport properties, possibly growing a high-
quality YBa2Cu3O7−␦ thin film which is needed in
many superconducting devices. One example from our
presented experimental results is the 1700-Å-thick
YBa2Cu3O7−␦ film on the 5° miscut (100) SrTiO3 sub-
strate deposited at 800 °C. The decrease of the optimum
growth temperature on the vicinal substrate was another
noteworthy point for manufacturers in terms of costs.
Additional research with various vicinal angles would be
useful to study the YBa2Cu3O7−␦ particle diffusion
length in order to develop an understanding of the growth
stage transition and the growth mode change. In addition,
it will be also interesting to grow films on planar sub-
strates whose surface has artificially generated step struc-
tures. More strains, such as TEM cross-section or in situ
growth mechanism analysis, are still needed for the more
fundamental research on the relationship between the
YBCO film growth and precipitates.
14. K. Hirata, F. Baudenbacher, and H. Kinder, Physica C 214, 272
(1993).
15. M.G. Norton and C.B. Carter, J. Cryst. Growth 110, 641 (1991).
16. J. Tsujino and Y. Shiohaara, Physica C 262, 236 (1996).
17. D.G. Schlom, D. Anselmetti, J.G. Bednorz, R.F. Broom, A. Ca-
tana, T. Frey, Ch. Gerber, H-J. Gu¨ntherodt, H.P. Lang, and
J. Mannhart, Z. Phys. B 86, 163 (1992).
18. B. Dam and B. Sta¨uble-Pu¨mpin, J. Mater. Sci. 9, 217 (1997).
19. T. Terashima, Y. Bando, K. Iijima, K. Yamamoto, K. Hirata,
K. Hayashi, K. Gamigaki, and H. Terauchi, Phys. Rev. Lett. 65,
2684 (1990).
20. N. Chandrasekhar, V.S. Achutharaman, V. Agrawal, and A.M.
Goldman, Phys. Rev. B 46, 8565 (1992).
21. M. Hawley, I.D. Raistrick, J.G. Berry, and R.J. Houlton, Science
251, 1587 (1991).
22. M. Me´chin, P. Berghuis, and J.E. Evetts, Physica C 302, 102 (1998).
23. A. Catana, J.G. Bednorz, Ch. Gerber, J. Mannhart, and D.G.
Schlom, Appl. Phys. Lett. 63, 553 (1993).
24. Y.A. Bo´ıkov, Z.G. Ivanov, E. Olson, V.A. Danilov, T. Claeson,
´
M. Shcheglov, and D. Erts, Phys. Solid State 37, 478 (1995).
25. T. Ushida, H. Higa, K. Higashiyama, I. Hirabayashi, and S. Ta-
naka, Appl. Phys. Lett. 59, 860 (1991).
26. N. Homma, S. Okayama, H. Takahashi, I. Yoshida, T. Morishita,
S. Tanaka, T. Haga, and K. Yamaya, Appl. Phys. Lett. 59, 1383
(1991).
27. P.W. Tasker and D.M. Duffy, Surf. Sci. 137, 91 (1989).
28. T. Haage, H-U. Habermeier, and J. Zegenhagen, Surf. Sci. 370,
L158 (1997).
29. R.R. Biggers, M.G. Norton, I. Martense, T.L. Peterson, E.K.
Moser, D. Dempsey, M.A. Capano, J. Talvacchino, J.L. Brown,
and J.D. Wolf, IEEE Trans. Appl. Supercond. 5, 1241 (1995).
30. M.G. Norton, B.H. Moeckly, C.B. Carter, and R.A. Buhrman,
J. Cryst. Growth 114, 258 (1990).
31. T. Haage, J. Zegebhagen, J.Q. Li, H-U. Habermeier, M. Cardona,
C. Jooss, R. Warthman, A. Forkl, and H. Kronmu¨ller, Phys. Rev.
B 56, 8404 (1997).
32. T. Haage, J. Zegebhagen, M. Cardona, and H-U. Habermeier,
J. Alloys Compd. 251, 23 (1997).
33. J.H. Claassen, M.E. Reeves, and R.J. Soulen, Jr., Rev. Sci. In-
strum. 62, 996 (1991).
REFERENCES
1. D.B. Chrisey, JOM 49(10), 28 (1997).
2. Y.J. Tian, L.P. Guo, L. Li, Y.Q. Zhou, Y. Tang, Z.X. Zhao,
S.F. Xu, H.B. Lu, Y.L. Zhou, Z.H. Chen, D.F. Cui, and G.Z.
Yang, Appl. Phys. Lett. 65, 234 (1994).
3. J. Kim, R.M. Stroud, R.C.Y. Auyeung, C.R. Eddy, D. Koller,
M.S. Osofsky, R.J. Soulen, Jr., J.S. Horwitz, and D.B. Chrisey,
(unpublished).
34. M. Matsui, K. Yamamoto, M. Nakajima, T. Shimano, and H. Mat-
suba, Supercond. Sci. Technol. 5, S427 (1992).
35. G.Y. Sung and J.D. Suh, Appl. Phys. Lett. 67, 1145 (1995).
36. A. Inam, C.T. Rogers, R. Ramesh, K. Remschnig, L. Farrow,
D. Hart, K. Venkatesan, and B. Wilkens, Appl. Phys. Lett. 57,
2484 (1990).
4. V. Boffa, T. Petrisor, L. Ciontea, U. Gambardella, and S. Bar-
banera, Physica C 276, 218 (1997).
37. W. Shi, J. Shi, J. Sun, W. Yao, and Z. Qi, Appl. Phys. Lett. 57,
822 (1990).
5. B. Holzapfel, B. Roas, L. Schultz, P. Bauer, and G. Saemaaa-
Ischenko, Appl. Phys. Lett. 61, 31178 (1992).
38. L. Luo, X.D. Wu, R.C. Dye, R.E. Muenchausen, S.R. Foltyn, and
Y. Coulter, Appl. Phys. Lett. 59, 2043 (1991).
6. S. Proyer, E. Stangle, M. Borz, B. Hellebrand, and D. Ba¨uerle,
Physica C 257, 1 (1996).
7. K. Verbist, A. Ku¨hle, and A.L. Vasiliev, Physica C 269, 131
(1996).
39. R. Feenstra, L.A. Boatner, J.D. Budai, D.K. Christen, M.D. Gal-
loway, and D.B. Poker, Appl. Phys. Lett. 54, 1063 (1989).
40. G. Linker, D. Huttner, O. Meyer, M. Ohkubo, and J. Reiner, J Al-
loys Compd. 251, 65 (1997).
8. M. Ye, M.P. Delplancke, J. Schroeder, R. Winand, and R. Del-
tour, Solid State Commun. 103, 645 (1997).
41. J. Zegenhagen, T. Siegrist, E. Fontes, L.E. Berman, and J.R. Patel,
Solid State Commun. 93, 763 (1995).
9. W.K. Chu, Y.P. Li, J.R. Liu, J.Z. Wu, S.C. Tidrow, N. Toyoda,
J. Matsuo, and I. Yamada, Appl. Phys. Lett. 72(2), 246 (1998).
10. N. Kanda, M. Kawasaki, T. Kitajima, and H. Koinuma, Phys.
Rev. B 56, 8419 (1997).
11. J.P. Gong, M. Kawasaki, K. Fujito, R. Tsuchiya, M. Yoshimoto,
and H. Koinuma, Phys. Rev. B 50, 3280 (1994).
42. E.J. Tarsa, E.A. Hachfeld, F.T. Quinlan, J.S. Speck, and E. Eddy,
Appl. Phys. Lett. 68, 490 (1996).
43. W.H. Hall, Proc. Phys. Soc. London Sec. A 62, 741 (1949).
44. G.K. Williamson and W.H. Hall, Acta Metall. 1, 22 (1953).
45. J-L. Maurice, O. Durand, M. Drouet, and J-P. Contour, Thin Solid
Films 319, 211 (1998).
612
J. Mater. Res., Vol. 15, No. 3, Mar 2000
Downloaded: 15 Mar 2015
http://journals.cambridge.org
IP address: 165.95.226.100

J. Kim et al.: Growth mechanism of YBa2Cu3O7−␦ thin films and precipitates on planar and vicinal SrTiO3 substrates
46. J. Bro¨tz, H. Fuess, T. Haage, and J. Zegenhagen, Phys. Rev. B 57,
3679 (1998).
47. S.B. Qadri, J.P. Yang, E.F. Skelton, and B.R. Ratna, Appl. Phys.
Lett. 70, 1020 (1997).
54. F. Baudenbacher, K. Hirata, P. Berberich, H. Kinder, W. Ass-
mann, and H.P. Lang, Physica C 185, 2177 (1991).
55. M.E. Bijlsma, D.H.A. Bank, H. Wormeester, A.V. Silfhout, and
H. Rogalla, J. Alloys Compd. 251, 15 (1997).
48. L. Ryen, E. Olsson, C.N.L. Edvardsson, and U. Helmersson, Phy-
sica C 304, 307 (1998).
49. P.C. McIntyre, M.J. Cima, and A. Roshko, J. Appl. Phys. 77, 5263
(1995).
50. M. Bauder, F. Baudenbacher, and H. Kinder, in Applied Super-
conductivity 1995 (Inst. Phys. Conf. Ser. 148, 1995), pp. 851–854.
51. M. Mukaida, S. Miyazawa, and M. Sasaura, Jpn. J. Appl. Phys.
30, L1474 (1991).
52. H.J. Scheel, C. Klemenz, F-K. Reinhart, H.P. Lang, and H-J.
Gu¨ntherodt, Appl. Phys. Lett. 65, 901 (1994).
53. B. Dam, N.J. Koeman, J.H. Rector, B. Sta¨uble-Pu¨mpin, U. Poppe,
and R. Griessen, Physica C 261, 1 (1996).
56. A. Baronnet, J. Cryst. Growth 19, 193 (1973).
57. S.J. Pennycook, M.F. Chicholm, D.E. Jesson, R. Feenstra, S. Zhu,
X.Y. Zheng, and D.J. Lowndes, Physica C 202, 1 (1992).
58. C.A. Neugebauer, in Handbook of Thin Film Technology, edited
by L.I. Meissel and R. Glang (McGraw-Hill, New York, 1970),
Chapter 8, pp. 8-5–8-37.
59. A.A. Chernov, in Modern Crystallography, edited by H.J. Queis-
ser (Springer, Berlin, Germany, 1984), Vol. 3, pp. 64–70.
60. T. Shitara and T. Nishinaga, Jpn. J. Appl. Phys. 28, 1212
(1989).
61. S.K. Streiffer, B.M. Lairson, and J.C. Bravmen, Appl. Phys. Lett.
57(23), 3 (1990).
J. Mater. Res., Vol. 15, No. 3, Mar 2000
Downloaded: 15 Mar 2015
613
http://journals.cambridge.org
IP address: 165.95.226.100