Li deficiencies in LiNbO
3
films prepared by pulsed laser deposition
in a buffer gas
J. Gonzalo,
a)
C. N. Afonso,
b)
and J. M. Ballesteros
Instituto de Optica, CSIC, Serrano 121, 28006 Madrid, Spain
A. Grosman and C. Ortega
Groupe de Physique des Solides, Universite
´
s de Paris VII et VI, URA 17 du CNRS, 2 Place Jussieu.
75251 Paris Cedex 05, France
~Received 1 April 1997; accepted for publication 18 June 1997!
The origin of Li deficiency in films grown by laser ablation of single-crystal LiNbO
3
targets in a
buffer gas has been investigated by analyzing the stoichiometry of the deposited films as a function
of the following parameters: the distance target-substrate, the nature of the buffer gas ~He, O
2
, and
Ar! and the deposition configuration. The results show that significant Li losses are related to
scattering processes during the expansion regime which are higher the higher the mass of the gas
species. The results show that the Li content of the films can be enhanced by setting the substrate
either at distances larger than the plume length or in a configuration in which the substrate is not
facing the target. © 1997 American Institute of Physics. @S0021-8979~97!06418-9#
I. INTRODUCTION
Lithium niobate (LiNbO
3
) is known to have excellent
nonlinear properties, including piezo-electric, acousto-optic,
and electro-optic properties.
1
These properties make LiNbO
3
a promising candidate for many opto-electronic applications
such as modulators, frequency converters, or storage
media
2,3
if the production of high quality films on substrates
with lower refractive index can be achieved.
Several deposition methods have been attempted to grow
high quality LiNbO
3
films, including molecular beam
epitaxy,
4
liquid phase epitaxy,
5
sol-gel process,
6
rf
sputtering,
7
and more recently, pulsed laser deposition
~PLD!.
8–12
The latter technique has been very successful in
the growth of high T
c
superconducting films and other com-
plex oxides.
13,14
The main limitation when producing
LiNbO
3
films by PLD is the loss of Li in the films, which
promotes the growth of a Li deficient phase, such as
LiNb
3
O
8
.
8,10,12
The use of Li-enriched sintered targets
10,11
or
mixed O
2
/Ar gas ambients
8
can partially overcome this prob-
lem and lead to stoichiometric LiNbO
3
films; nevertheless,
the conditions for producing high quality films are still un-
clear, and in the literature some controversial results can be
found. Some authors
8
claim that no single phase films can be
obtained by ablation of stoichiometric LiNbO
3
in a pure oxy-
gen environment, whereas others succeeded.
12
In other
cases,
10,11
no stoichiometric films were obtained by ablation
of stoichiometric targets, this result being attributed to the
re-evaporation of the adsorbed Li on the film surface upon
rising the temperature.
10
In addition, the use of single-
crystalline stoichiometric targets to grow stoichiometric
films might be desirable, because their use has been reported
to lead to the growth of particulate-free films.
9
Therefore,
further studies of the influence of the deposition parameters
on the PLD process have to be performed in order to eluci-
date these points.
In an earlier work, we combined Rutherford backscatter-
ing spectrometer ~RBS! and nuclear reaction analysis ~NRA!
to study the influence of the target composition, the laser
fluence and the oxygen pressure applied during deposition on
the stoichiometry of films grown by ablation of LiNbO
3
.
11
The results showed that a Li/Nb molar fraction of 1.5–2.0
was required to obtain nearly stoichiometric films in agree-
ment with other reports,
10
and this result depended very little
on both laser fluence and oxygen pressure. The aim of this
work is to investigate the influence of the target-substrate
distance, the nature of the buffer gas and the substrate con-
figuration, on the stoichiometry of films grown by ablation of
single-crystalline targets. The results are analyzed in terms of
the expansion dynamics of the plasma and the role of colli-
sional processes leading to a preferential scattering of the
lighter species.
II. EXPERIMENT
A LiNbO
3
single-crystal ~@Li#/@Nb#50.94 in the melt!
has been ablated using an ArF excimer laser ~l5193 nm,
t
512 ns FWHM! focused on the target surface. The target
was mounted in a rotating holder and placed in a vacuum
chamber evacuated to a residual pressure of 2.0310
2 7
mbar,
which will be referred to hereafter as vacuum. The angle of
incidence of the laser beam was 45°, and the laser energy
density at the target surface was '2 J/cm
2
. The films were
grown on Si~100! substrates held at room temperature by
ablating the target during 60 minutes after 5 minutes of pre-
ablation, both at 5 Hz. Although the target composition
might be modified by the laser ablation process
15
and thus
lead to non-stoichiometric films, this cleaning procedure to-
gether with the long deposition time guarantees that a steady
state in the target composition is reached and therefore all the
films are deposited under similar target conditions.
The influence of the target-substrate distance (d) on the
composition of the films has been analyzed by growing films
a!
Present address: Department of Applied Physics, University of Hull, Hull,
HU6 7RX, United Kingdom.
b!
Electronic mail: cnafonso@pinar1.csic.es
3129J. Appl. Phys. 82 (6), 15 September 1997 0021-8979/97/82(6)/3129/5/$10.00 © 1997 American Institute of Physics
Downloaded 18 Feb 2010 to 161.111.180.191. Redistribution subject to AIP license or copyright; see http://jap.aip.org/jap/copyright.jsp
on substrates located at d5 20, 33, and 40 mm from the
target in 7.03 10
2 2
mbar of oxygen. The role of the buffer
gas was studied by growing films at a fixed distance (d
5 33 mm! in four different environments: vacuum and 7.0
3 10
2 2
mbar of He, O
2
, and Ar. Finally, since the scattering
of the ejected species by the gas atoms and/or molecules is
known to play a crucial role on the stoichiometry of the films
grown by PLD
16–18
an experiment involving simultaneous
deposition on two substrates was performed. The substrates
were located at two sites as described in detail elsewhere.
18
One was along the target normal ~S1! at a fixed distance
(d5 33 mm! and facing the target, and the second one ~S2!
was shifted upwards and located such as the deposit was
non-facing the target. Two different configurations were con-
sidered for substrate S2 during the deposition process: ~i! the
re-emission ~S2RE! configuration, in which both S1 and S2
substrates were present and ~ii! the gas scattering ~S2GS!
configuration in which only the S2 substrate was present, and
therefore, the material deposited could only come from the
scattering of the ablated species by the buffer gas.
The composition of the films was determined at their
thicker region by simultaneous RBS and NRA using a deu-
teron beam at 0.85 MeV.
11
The Nb content was detected by
RBS, while the Li and the O contents were determined by
NRA through the
6
Li(d,
a
0
)
4
He and
16
O(d,p)
17
O
*
reac-
tions, respectively. We have considered as representative ra-
tios to describe the composition of the films, the lithium and
oxygen to niobium atomic ratios (N
Li
/N
Nb
& N
O
/N
Nb
). The
absolute values of the film composition have been deter-
mined by comparison with references known within 3% ~O
and Nb! and 20% ~Li!. The reduced precision of the Li ref-
erence leads, therefore, to a low precision in the determina-
tion of the absolute values. Nevertheless, the changes of rela-
tive composition of films grown in the different experimental
conditions remain unaffected by this imprecision and are
only affected by the statistical errors ~root-mean-square er-
rors! that are in all cases in the range 4–7.5%. Further details
of the experimental configuration used for analysis and the
determination of the experimental errors can be found
elsewhere.
11
III. RESULTS
Figure 1~a! shows the total amount of atoms (N
tot
5 N
Li
1 N
Nb
1 N
O
) measured in films deposited on the facing sub-
strate ~S1! at different distances from the target in 7.0
3 10
2 2
mbar of O
2
. It is seen that the amount of material
deposited on the substrate decreases when the distance from
the target increases, this decrease being sharper for d. 33
mm. Figure 1~b! shows the dependence of the N
Li
/N
Nb
and
N
O
/N
Nb
ratios measured in the same films. Both ratios de-
crease as d is increased to 33 mm; nevertheless, they in-
crease for greater distances. Figure 1~b! also shows that the
films are in all cases deficient in Li, with N
Li
/N
Nb
ratios well
below the desired value for LiNbO
3
(N
Li
/N
Nb
5 1). Finally,
both Figs. 1~a! and 1~b! clearly show that there is a change in
behavior near d5 33 mm.
The presence of a buffer gas has also a strong influence
on the relative composition of the films as it can be seen in
Fig. 2 for films grown at d533 mm in different gas environ-
ments: vacuum and 7.03 10
2 2
mbar of He, O
2
, or Ar. It is
clearly seen that, although the N
Li
/N
Nb
and N
O
/N
Nb
ratios
are always below the stoichiometric content, they generally
follow opposite trends: N
Li
/N
Nb
decreases and N
O
/N
Nb
increases as the mass and radius of the gas species are
increased.
The total amount of atoms (N
tot
) ~Fig. 3! and the relative
composition of the films ~Fig. 4! also have a strong depen-
dence on the substrate configuration. Results obtained for
films grown on the facing ~S1! and non-facing ~S2! sub-
strates either in vacuum or in a buffer gas ~He, O
2
and Ar!
are included in both figures. It is clearly seen in Fig. 3 that
the value of N
tot
for films grown on the S1 and S2 substrates
follow opposite trends: the former decreases and the latter
increases as the mass and radius of the buffer gas increases.
FIG. 1. ~a! Total amount of atoms (N
tot
), and ~b! Li and O to Nb atomic
ratios (N
Li
/N
Nb
, N
O
/N
Nb
) of films deposited on substrate S1 at different
distances ~ d520 mm, 33 mm and 40 mm! from the target surface in 7.0
3 10
2 2
mbar of O
2
. The dashed lines correspond to the estimation of the
plume length according to the adiabatic expansion model.
FIG. 2. Li and O to Nb atomic ratios (N
Li
/N
Nb
, N
O
/N
Nb
! for films depos-
ited on substrate S1 at d533 mm from the target surface in the different gas
environments considered in this work ~vacuum and 7.0310
2 2
mbar of He,
O
2
,orAr!. The dashed lines indicate the composition of the target.
3130 J. Appl. Phys., Vol. 82, No. 6, 15 September 1997 Gonzalo
et al.
Downloaded 18 Feb 2010 to 161.111.180.191. Redistribution subject to AIP license or copyright; see http://jap.aip.org/jap/copyright.jsp
In vacuum, the amount of deposited material on S2 sub-
strates is below our experimental resolution either in S2GS
or S2RE configuration; whereas in the other buffer gases
considered, the amount of deposited material in the S2GS
configuration is higher than in the S2RE, similarly to what
has been earlier reported for BiSrCaCuO films.
18
Figure 4 shows the N
Li
/N
Nb
and N
O
/N
Nb
ratios for films
grown in the different configurations and buffer gases con-
sidered. Both ratios ar always higher for films grown on S2
substrates than for the corresponding films deposited on the
S1 substrates. Films deposited on S2 substrates in the pres-
ence of a buffer gas shown a N
Li
/N
Nb
ratio close to the
stoichiometric value, or even much higher when using the
lightest gas ~He!. The oxygen relative content (N
O
/N
Nb
! of
films grown on S2 substrates is in all cases above the sto-
ichiometric value, although the exact value depends on the
particular buffer gas, and similarly to the N
Li
/N
Nb
ratio, it
decreases for films grown in heavier buffer gases ~i.e., O
2
or
Ar!.
IV. DISCUSSION
The expansion of the plasma produced by laser ablation
in the presence of a buffer gas is known to be dominated by
the interaction of the ejected species and the background gas
atoms or molecules.
16,17,19–22
In the pressure range consid-
ered (P, 0.1 mbar!, this interaction has been modeled
16,17
in
terms of the scattering of the ejected species by the buffer
gas atoms and/or molecules, which alter the kinetic energy of
the ejected species and eventually lead to changes in their
angular distribution and in the composition of the grown
films. The changes depend not only on the nature and pres-
sure of the buffer gas, but also on the target-substrate
distance.
22
During the initial expansion, which extends along
as a few mm from the target surface, the driver mass ~abla-
tion products! is very large compared with the mass of the
driven gas; and thus, the buffer gas is pushed on by the
leading edge of the plume, the plume expansion remaining
unaffected by the presence of the buffer gas.
19–21
For higher distances, and once the ejected species inter-
act with the buffer gas atoms and/or molecules, the decrease
of the thickness and the compositional changes observed in
Fig. 1 as a function of the target-substrate distance should be
understood in terms of the collision probability and the an-
gular dispersion. Assuming a simple model based on hard
spheres
23
and taking into account that the mean velocity of
the ablated species is much higher than the thermal mean
velocity of the buffer gas species, the meanfree path l
1
of
species 1 in buffer gas 2 is given by
l
1
5 kT/
p
Pd
12
2
, ~1!
where k is the Boltzmann constant, T is the temperature of
the gas, P is the gas pressure and d
12
5 (d
1
1 d
2
)/2isthe
impact parameter ~d
1
and d
2
being the diameters of the spe-
cies 1 and 2!. Taking into account the experimental condi-
tions and the diameter of the different species and gases,
showed in Table I,
24
we have estimated the mean free path of
the ejected atomic species Li (l
Li
), Nb (l
Nb
), and O ~l
O
)in
O
2
. The results are in all cases of the order of a few mm, the
mean free path of Li being approximating l
Li
'1.3 mm.
Since the ratio l
Li
/l
Nb
'1, both species have a similar col-
lisional probability; nevertheless m
Li
! m
Nb
and thus the an-
gular dispersion in laboratory coordinates of Li per collision
is much higher than that of Nb. Consequently, the N
Li
/N
Nb
FIG. 3. Total amount of atoms (N
tot
) for films deposited on substrates S1
and S2 in the two configurations considered ~S2RE & S2GS! at a distance
d5 33 mm from the target surface in the different gas environments consid-
ered in this work ~vacuum and 7.03 10
2 2
mbar of He, O
2
,orAr!. The figure
also shows the experimental configurations considered in this work: Re-
emission ~S2RE! and gas scattering ~S2GS! configurations.
FIG. 4. Li and O to Nb atomic ratios (N
Li
/N
Nb
, N
O
/N
Nb
) for films depos-
ited in 7.0310
2 2
mbar of He, O
2
, and Ar respectively on substrates S1 and
S2 in the two configurations considered ~S2RE & S2GS! at d5 33 mm from
the target surface. The dashed lines indicate the composition of the target.
3131J. Appl. Phys., Vol. 82, No. 6, 15 September 1997 Gonzalo
et al.
Downloaded 18 Feb 2010 to 161.111.180.191. Redistribution subject to AIP license or copyright; see http://jap.aip.org/jap/copyright.jsp
ratio should decrease as the distance d to the target is in-
creased. This reasoning agrees with the results shown in Fig.
1~b! for distances shorter than d5 33 mm. In order to apply
this reasoning to the N
O
/N
Nb
ratio, one has to take into ac-
count the mean free path ratio (l
O
/l
Nb
'2.5, in O
2
! and the
mass relationship (m
O
! m
Nb
). Thus, although the collision
probability of O in O
2
is a 40% of that of Nb, the angular
dispersion per collision is much higher for O. Therefore, it is
not straightforward to predict accurately the behavior of the
N
O
/N
Nb
ratio when increasing the distance. However, since
l
O
and
s
Nb
in O
2
are much smaller than 33 mm, the O and
Nb species will suffer several collisions before reaching the
substrate. Then the N
O
/N
Nb
ratio at this distance will be
most likely dominated by the effect of the angular dispersion
and therefore it should decrease, in agreement with what is
observed experimentally.
The former collision-based reasoning does not explain
the change of behavior observed for N
Li
/N
Nb
and N
O
/N
Nb
ratios for films grown at d5 40 mm shown in Fig. 1~b!.In
the presence of a buffer gas and after the initial free expan-
sion region, it has been suggested that the plasma expansion
occurs in two consecutive stages:
20,21
~i! an expansive stage
in the region where the plume pressure is still higher than the
gas pressure and in which the deceleration of the ejected
species is observed and ~ii! a diffusive stage that starts once
the plume and gas pressures equilibrate, the ejected species
being then transported by diffusion to the substrate. The dis-
tance at which the change of stage occurs ~plume length!
represents a distance related pressure threshold that has
strong influence on the physical properties of the films like
crystallinity,
12
refractive index,
9,25
or roughness.
26
The
plume length can be estimated
20,25
within the model of adia-
batic expansion as
L
P
5 A
@~
g
21
!
E
#
1/
~
3
g
!
P
2 /1
~
3
g
!
V
~
g
21
!
/
~
g
23
!
, ~2!
where A is a geometrical factor related to the shape of the
laser spot at the target surface,
g
is the ratio of specifics heats
(C
P
/C
v
), E is the laser energy per pulse, P is the gas pres-
sure, and V is the initial volume of the plasma (V'
v
0
3
t
3 spot size,
v
0
being the initial species velocity, and
t
the
laser pulse duration!. Taking into account the experimental
parameters used in this work ~ A51.6,
t
512 ns and a spot
size of approx. 1 mm
2
! together with the expansion velocities
of the ejected species measured in an earlier work
9
(
v
0
5 0.8–0.93 10
4
m/s! and a value of
g
51.3–1.4,
20
the
above formula leads to a plume length L
p
'346 4 mm. This
result suggests that at d5 20 mm the films were grown in the
expansion stage, whereas when d5 40 mm the films were
grown in the diffusion one.
As a consequence of the scattering processes which oc-
cur for distances shorter than the plume length, the angular
distribution of the species at d'L
p
at which the diffusion
regime starts to dominate is not the same for the different
species: the lighter the species ~Li and O! the broader the
angular distribution. Since in the diffusion regime the species
are thermalized and their movement becomes random, the
species having initially the broader distribution will provide
a higher relative amount of species in the forward direction,
i.e., the center of the deposit in the facing substrate, as they
simulate planar diffusion to a slightly higher degree. Accord-
ing to this reasoning, the films grown at d5 40 mm should
present an enrichment in Li and O ~in respect to Nb! when
compared to those grown at d5 33 mm. This reasoning is in
excellent agreement with the experimentally measured ratios
shown in Fig. 1~b!; therefore, it can be concluded that the
composition of the films is collision or diffusion controlled
for d, L
p
and d. L
p
respectively and that the Li and O
content of the films is enhanced at larger distances by a dif-
fusion controlled process. This interpretation is also in agree-
ment with the sharp decrease in the film thickness of the
films grown at d5 40 mm @Fig. 1~a!# since for distances
larger than L
p
the total amount of material reaching the sub-
strate should be significantly decreased as a consequence of
the random movement of the species.
Equation ~1! also indicates that the mean free path of the
ejected species depends on the nature of the buffer gas and it
should decrease as the radius of the gas species increases,
thus increasing the collision probability. In addition, the an-
gular dispersion per collision increases as the mass of the gas
species increases. Therefore as the mass and the radius of the
gas species increase, as it is our case, a larger fraction of
species are deviated from their original trajectories, leading
to less species reaching the substrate located in front of the
target at d5 32 mm. This reasoning is in very good agree-
ment with the experimental results plotted in Fig. 3 which
show that the thickness of the films grown on the substrate
S1 ~facing the target and along its normal! is lower in a gas
environment than in vacuum and it decreases as the mass and
radius of the gas species increases ~He→O
2
,Ar!. In order to
fully understand the composition changes occurring when
varying the nature of the gas ~Fig. 2!, the dependence of the
scattering probability on the mass and radius of both the
ejected and gas species has to be taken into account. Since
the mean free path ratio l
Li
/l
Nb
'1 in any of the considered
atmospheres, the N
Li
/N
Nb
ratio reaching the substrate will be
mainly controlled by the mass of the gas species, and a de-
crease of the value of N
Li
/N
Nb
should be expected in heavier
environments in reasonably good agreement with the experi-
mental results shown in Fig. 2.
The ratio of the oxygen to niobium mean free paths
(l
O
/l
Nb
) decreases as the radius ~and mass! of the buffer
gas species increases (l
O
/l
Nb
'3.3, 2.4 and 2.3 in He, O
2
,
and Ar, respectively!. Following the above reasoning, a de-
crease of the value of N
O
/N
Nb
should be also observed as the
mass of the gas species increases. Nevertheless, this reason-
TABLE I. Atomic ~or molecular! masses and radius, taken from Ref. 23,
corresponding to the different elements present in LiNbO
3
and the gases
considered in this work. The estimated mean free path values in oxygen ~l!
for Li, O, and Nb are also included.
Element Mass Radius ~Å!l~mm!
Li 6.94 2.05 1.3
O 16.0 0.65 3.3
Nb 92.1 2.08 1.3
He 4.00 1.09 ---
O
2
32.0 1.80 ---
Ar 40.0 1.82 ---
3132 J. Appl. Phys., Vol. 82, No. 6, 15 September 1997 Gonzalo
et al.
Downloaded 18 Feb 2010 to 161.111.180.191. Redistribution subject to AIP license or copyright; see http://jap.aip.org/jap/copyright.jsp
ing is in contradiction with the experimental results shown in
Fig. 2. An alternative explanation can be given by taking into
account that there are be oxidized species present in the
plasma which also contribute to the oxygen content in addi-
tion to atomic oxygen species.
21
These oxidized species are
heavier and they should suffer less angular dispersion than
the atomic species; thus explaining why the N
O
/N
Nb
ratio
increases as the mass of the gas species increases.
The species which are preferentially scattered are the
lightest species. They cannot reach the substrate located fac-
ing the target along its normal ~S1! and are in part collected
by the non-facing substrate S2. This interpretation is fully
supported by the opposite trends of the thickness of the films
grown on substrates S1 and S2 as the mass of the gas is
increased ~Fig. 3!. The changes in film composition for those
grown on the substrate S2 ~either in S2GS or S2RE configu-
ration! in the different gas environments ~Fig. 4! should then
be also understood in terms of scattering processes. Accord-
ing to the collisional framework (l
Li
/l
Nb
'1 and m
He
'm
Li
! m
Nb
!, Nb atoms will undergo a much lower angular
dispersion than Li atoms when colliding with He atoms. This
fact leads to a very large N
Li
/N
Nb
ratio in films deposited on
S2 substrates in a He environment. As the mass and radius of
the gas species increases ~He→O
2
,Ar!the probability of
suffering high angular dispersion becomes important for Nb,
and then N
Li
/N
Nb
ratio decreases in good agreement with the
experimental results showed in Fig. 4. The N
O
/N
Nb
ratio of
films grown on S2 substrates in the different gas environ-
ments follows a similar trend than the N
Li
/N
Nb
ratio. This
result is not surprising since the oxygen atomic species
ejected from the target are the only ones experiencing a high
angular dispersion and thus contributing to the oxygen con-
tent of the films deposited onto S2 substrate. The heavier
oxidized species suffer negligible dispersion.
V. CONCLUSIONS
The stoichiometric changes observed in films deposited
at different target-substrate distances, in different buffer
gases and substrates configuration can be understood in
terms of a collisional scheme in which the lighter species are
preferentially scattered by the buffer gas. The composition of
the films is collision or diffusion controlled for distances
shorter and longer than the plume length respectively, the
relative Li content being enhanced in the diffusion regime.
Finally, the preferential scattering of the lighter species leads
to films enriched in Li on substrates located off the target
normal.
ACKNOWLEDGMENTS
Dr. E. Die
´
guez ~UAM, Madrid, Spain! is thanked for
providing the targets, Professor R. W. Dreyfuss ~I. de Optica,
Madrid, Spain! is thanked for helpful discussions and com-
ments. This work has been partially supported by CICYT
~Spain! under Project TIC96-0467.
1
Properties of Lithium Niobate, EMIS Datareviews Series, No. 5 ~1989!.
2
E. Voger, in Electro-optic and Photorefractive Materials, edited by P.
Gunter ~Springer, Berlin, 1987!.
3
C. S. Tai, Proc. IEEE 84, 853 ~1996!.
4
R. A. Betts and C. W. Pitt, Electron. Lett. 21, 960 ~1985!.
5
A. Yamada, H. Tamada, and M. Saitoh, Appl. Phys. Lett. 62, 2848 ~1992!.
6
A. A. Wernberg, H. J. Gysling, A. J. Filo, and T. N. Blanton, Appl. Phys.
Lett. 62, 946 ~1993!.
7
R. C. Baumann, T. A. Roast, and T. A. Rabson, J. Appl. Phys. 68, 2989
~1990!.
8
S. B. Ogale, R. Nawathey-Dikshit, S. J. Dikshit, and S. M. Kanetkar, J.
Appl. Phys. 71, 5718 ~1992!.
9
C. N. Afonso, F. Vega, J. Gonzalo, and C. Zaldo, Appl. Surf. Sci. 69, 149
~1993!.
10
Y. Shibata, Y. Kanno, K. Kaya, M. Knai, and T. Kawai, Jpn. J. Appl.
Phys. 2 34, L320 ~1995!.
11
C. N. Afonso, J. Gonzalo, F. Vega, E. Die
´
guez, J. C. Cheang Wong, C.
Ortega, J. Siejka, and G. Amsel, Appl. Phys. Lett. 66, 1452 ~1995!.
12
S.-H. Lee, T.-K. Song, T. W. Noh, and J.-H. Lee, Appl. Phys. Lett. 67,43
~1995!.
13
K. L. Saenger, in Pulsed Laser Deposition of Thin Films, edited by D. B.
Chrisey and G. K. Hubler ~Wiley, New York, 1994!, p. 582, and refer-
ences therein.
14
C. N. Afonso, in Materials for Optoelectronics: New Developments, ed-
ited by F. Agullo
´
-Lo
´
pez ~World Scientific, Singapore 1995!, Chap. 1, and
references therein.
15
N. Omori and M. Inoue, Appl. Surf. Sci. 54, 232 ~1992!.
16
J. C. S. Kools, J. Appl. Phys. 74, 6401 ~1993!.
17
D. B. Geohegan, in Ref. 13, p. 115.
18
J. Gonzalo, C. N. Afonso, and J. Perrie
`
re, Appl. Phys. Lett. 67, 1325
~1995!.
19
J. Gonzalo, C. N. Afonso, and I. Madariaga, J. Appl. Phys. 81, 951 ~1997!.
20
P. E. Dyer, A. Issa, and P. H. Key, Appl. Surf. Sci. 46,89~1990!.
21
W. K. A. Kummuduni, Y. Nakayama, Y. Nakata, T. Okada, and M.
Maeda, J. Appl. Phys. 74, 7510 ~1993!.
22
H. S. Kwok, Appl. Surf. Sci. 109/110, 595 ~1997!.
23
S. Dushman, Scientific Foundations of Vacuum Technique ~Wiley, New
York, 1962!, Chap. 1.
24
Periodic Table of Elements ~SMI Corp., 1991!; R. A. Alberty and R. J.
Silbey, Physical Chemistry ~Wiley, New York, 1992!, Chap. 18.
25
J. Gonzalo, C. N. Afonso, and J. M. Ballesteros, Appl. Surf. Sci. 109/110,
606 ~1997!.
26
I. Zhang, D. Cui, Y. Zhou, L. Li, Z. Chen, M. Sazbadi, and P. Hess, Thin
Solid Films 287, 101 ~1996!.
3133J. Appl. Phys., Vol. 82, No. 6, 15 September 1997 Gonzalo
et al.
Downloaded 18 Feb 2010 to 161.111.180.191. Redistribution subject to AIP license or copyright; see http://jap.aip.org/jap/copyright.jsp
