Characteristics of Ce-Doped ZrO2 Dielectric Films Prepared by a Solution Deposition Process

Article abstract of DOI:10.1149/1.3367749

The microstructural and electrical properties of sol-gel deposited ultrathin Zr_1-x Ce_x O₂ films with different Ce contents (x=0, 0.1, 0.3, and 0.5) were studied using various characterization tools. Ce doping reduced the crystallization (densification) temperature and increased the dielectric constant of the Zr_1-x Ce_x O ₂ film. There was no degradation of the hysteresis characteristics, and a systematic negative shift in the flatband voltage was observed by incorporating Ce atoms. Leakage current measurements showed no detrimental effects of Ce doping up to x=0.5, and the conduction mechanism analyses revealed that the Zr_1-x Ce_x O₂ films follow a Poole-Frenkel (PF) conduction, exhibiting a systematic increase in the linear slope of the PF plot possibly due to the decrease in dielectric trap sites with increasing Ce content.

Full text of DOI:10.1149/1.3367749

G142
Journal of The Electrochemical Society, 157 ͑6͒ G142-G146 ͑2010͒
0
013-4651/2010/157͑6͒/G142/5/$28.00 © The Electrochemical Society
Characteristics of Ce-Doped ZrO Dielectric Films Prepared by
2
a Solution Deposition Process
a
a
b
a
a
Myung Soo Lee, Chee-Hong An, Jun Hyung Lim, Jin-Ho Joo, Hoo-Jeong Lee,
a,z
and Hyoungsub Kim
a
School of Advanced Materials Science and Engineering, Sungkyunkwan University, Suwon 440-746,
Korea
b
Department of Materials Science and Engineering, Stanford University, Stanford, California 94305, USA
The microstructural and electrical properties of sol–gel deposited ultrathin Zr1−xCe O films with different Ce contents ͑x = 0, 0.1,
x
2
0
.3, and 0.5͒ were studied using various characterization tools. Ce doping reduced the crystallization ͑densification͒ temperature
and increased the dielectric constant of the Zr1−xCe O film. There was no degradation of the hysteresis characteristics, and a
x
2
systematic negative shift in the flatband voltage was observed by incorporating Ce atoms. Leakage current measurements showed
no detrimental effects of Ce doping up to x = 0.5, and the conduction mechanism analyses revealed that the Zr1−xCe O films
x
2
follow a Poole–Frenkel ͑PF͒ conduction, exhibiting a systematic increase in the linear slope of the PF plot possibly due to the
decrease in dielectric trap sites with increasing Ce content.
©
2010 The Electrochemical Society. ͓DOI: 10.1149/1.3367749͔ All rights reserved.
Manuscript submitted November 27, 2009; revised manuscript received February 24, 2010. Published April 26, 2010.
8
In an effort to find a suitable gate dielectric material to replace
conventional SiO2 for application in complementary metal oxide
semiconductor field-effect transistors ͑CMOSFETs͒, high-␬ dielec-
tric materials have been intensively studied for many years, and
current and requires a high postdeposition annealing temperature.
For the ZrO system, it has been theoretically predicted that mono-
clinic ZrO2 can be easily stabilized as a tetragonal phase having a
high dielectric constant by incorporating several stabilizing agents,
such as group IV and rare-earth elements. Using ultrathin HfO2
films having crystal structures and chemical properties similar to
2
1
9
some have already been used in early manufacturing stages. During
the material selection process from among the huge pool of high-␬
metal oxides, various vacuum-based deposition techniques have
been adopted, such as chemical vapor deposition and physical vapor
deposition, and the corresponding microstructural and electrical
properties of many metal oxide systems have been widely
ZrO2, many researchers also experimentally verified the effective-
1
0,11
ness of doping using rare-earth elements.
In this study, several ultrathin Zr1−xCexO2 films having a thick-
ness range of 3.5–4 nm were prepared with various Ce doping con-
centrations by a sol–gel method, and the Ce/͑Zr + Ce͒ atomic ratio
x was carefully controlled by varying the molar ratios of each pre-
cursor at the initial solution mixing stage. The effects of the Ce
content on the microstructural and chemical properties of
Zr1−xCe O films were investigated using several characterization
tools, such as high resolution transmission electron microscope
͑HRTEM͒ and X-ray photoelectron spectroscope ͑XPS͒. Further,
simple metal-oxide-semiconductor ͑MOS͒ capacitors were fabri-
cated to extract various electrical parameters.
2
investigated. Among these deposition techniques, the most promis-
ing one is atomic layer deposition due to its excellent resulting film
quality and precise thickness controllability required for the subna-
3
nometer thick gate dielectric applications. However, it requires a
suitable precursor combination and even a chemical development,
which hinders a systematic combinatorial study of new or mixed
metal oxide systems. Additionally, process complexity and high
equipment cost problems always arise due to the requirement of the
expensive vacuum-related components.
x
2
Recently, to overcome these technical hurdles and to understand
the material properties of various new metal oxide or mixed sys-
tems, Van Elshocht et al. suggested an alternative to the vacuum-
based deposition technique and studied several metal oxides synthe-
Experimental
The initial precursor solution was prepared by dissolving zirco-
nium͑IV͒ acetylacetonate ͓Zr͑acac͒ , Aldrich͔ and cerium nitrate
4
4
sized by using a chemical solution deposition technique. The main
hexahydrate ͓Ce͑NO ͒ ·6H O, Aldrich͔ in n-dimethylformamide
3
3
2
advantages of this solution-based deposition are easy precursor se-
lection, versatile composition controllability, and low cost manufac-
turability. Another possible application area of this solution-based
deposition process for high-␬ dielectric materials could be flexible
and transparent electronic devices, which require low synthesizing
͓
HCON͑CH ͒ , Aldrich͔ to a concentration of 0.1 M. As a stabiliz-
3 2
ing agent, ethanolamine ͑C H NO, Aldrich͒ was added with a molar
2
2
ratio of 2:1 relative to Zr and Ce solutes. The Ce doping concentra-
tion x was 0, 0.1, 0.3, or 0.5 and was carefully controlled by adjust-
ing the amount of Zr and Ce precursors. Ce doping concentrations of
the prepared solutions were measured to be within an experimental
error range of ⌬x Յ 0.01 according to inductively coupled plasma
atomic emission spectroscopy analyses. The mixture was then mag-
netically stirred at 100°C for 90 min to achieve a chemically homo-
geneous solution. To characterize the initial precursor solution, ther-
mal analysis was carried out in an air ambient using a differential
thermal analyzer ͑DTA, Seiko Exstar 7300͒ with a ramping rate of
5
temperature, low processing cost, and high volume productivity. In
several active and passive devices, such as thin film transistors, sen-
sors, and flexible CMOSFET devices, high quality dielectric films
having a high dielectric constant produced by a low cost solution-
based deposition technique may be needed in the near future, even
for printed electronic devices. However, the most challenging ob-
stacle for the application of the solution-based deposition technique
is to reduce the postdeposition densification temperature in achiev-
ing high quality dielectric films.
1
0°C/min.
Before the deposition of Zr1−xCe O films with different Ce con-
x
2
Among many possible high-␬ material systems studied so far,
tents, small pieces ͑1 ϫ 1 cm͒ of p-type Si͑100͒ wafer were
ZrO is one of the most promising candidates due to its high dielec-
2
cleaned by immersing in a 1% HF/H O solution for 90 s. The
6
,7
2
tric constant ͑␬ = 25–30͒ and relatively large bandgap ͑5.8 eV͒.
Zr Ce O films were deposited by spin-coating a single layer of
1
−x
x
2
However, because the typical solution-based deposition process,
e.g., the sol–gel method, often results in a low density of the depos-
the precursor solution and dried on a hot plate at 100°C for 10 min.
The spin-coated dielectric films were sintered and densified at
500°C for 60 min in an air ambient before characterization and
device fabrication. The final Ce doping concentration in the densi-
fied Zr1−xCe O film was confirmed by Rutherford backscattering
ited film, the sol–gel deposited ZrO film exhibits a high leakage
2
x
2
z
E-mail: hsubkim@skku.edu
spectroscopy and was measured to be close to the initial molar mix-
Downloaded on 2015-05-09 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 157 ͑6͒ G142-G146 ͑2010͒
G143
Figure 2. Cross-sectional HRTEM images of sol–gel deposited Zr1−xCe O
x
2
films after the densification process ͑500°C, 1 h in air ambient͒: ͑a͒ x = 0,
͑
b͒ x = 0.1, ͑c͒ x = 0.3, and ͑d͒ x = 0.5. Fourier-transformed images from the
Figure 1. ͑Color online͒ DTA curves of the prepared precursor solutions as
a function of the Ce content, x = Ce/͑Zr + Ce͒, used for the sol–gel deposi-
tion of Zr Ce O films.
dotted-circle regions are also depicted beside each HRTEM picture.
1
−x
x
2
not be substantiated at this point, one of the possibilities could be
the corresponding increase in the viscosity of the precursor solution,
ing ratios of the prepared precursor solutions within an experimental
error range of ⌬x Յ 0.02. For convenience, in the rest of this paper,
the doping concentration is expressed using a molar mixing ratio of
Zr and Ce precursors during the precursor preparation, i.e., x = 0,
and more study is under way. Also, a 1.8 nm thick SiO or silicate
2
layer, an interfacial oxide, was formed for the pure ZrO sample
2
possibly during the densification process at 500°C in an oxygen-rich
air ambient, and its thickness slightly increased to 2.5 nm for all the
0
.1, 0.3, and 0.5. The carbon content of the film was measured to be
Ce-doped ZrO samples, irrespective of the Ce content. Although
2
less than the detection limit of XPS after removing the surface car-
bon layer by Ar etching.
the origin of this thickness difference of the interfacial layer ͑IL͒
between ZrO and Ce-doped ZrO samples is not fully understood,
2
2
The chemical bonding status of the Zr1−xCe O films was mea-
x
2
we suggest that it is related to the different extents of the catalytic or
sured using an XPS ͑Multilab 2000͒ with an Al K␣ ͑h␯
the oxidation/reduction effect between ZrO and CeO in instigating
2
2
=
1486.6 eV͒ X-ray source, whereas the microstructural character-
interfacial oxide growth. According to the aforementioned DTA
analysis results, the crystallization/densification of the ZrO2 film
would occur at a much lower temperature than the densification
temperature of 500°C performed on all the samples after the sol–gel
deposition. However, only an amorphous image without any notice-
able crystallite or local ordering was identified from the cross-
istics were studied on cross-sectional transmission electron micros-
copy samples using an HRTEM ͑JEM2100F͒ operating at an accel-
eration voltage of 200 kV. For the fabrication of simple MOS
capacitor structures, an Al ͑10 nm͒/TaN ͑50 nm͒ stacked structure
2
was formed as a top gate electrode having a size of 7850 ␮m by
using a sputtering technique and patterning via a simple lift-off pro-
cess. The electrical properties of the fabricated capacitor samples
were measured after forming gas annealing ͑5% H + 95% N ͒ at
sectional HRTEM analysis of the ZrO sample, as shown in Fig. 2a.
2
This phenomenon can be understood by the thickness effect on the
2
2
retardation of the crystallization usually observed in ZrO films pre-
2
1
3
400°C for 30 min. Capacitance–voltage ͑C-V͒ characteristics were
pared by vacuum-based techniques. With the Ce content increas-
measured at 100 kHz using an Agilent E4980A LCR meter to extract
various MOS capacitor parameters. The leakage current density was
also measured using a Keithley 230 programmable voltage source
and a 6514 programmable electrometer.
ing, some local ordering in the Zr1−xCe O film started to appear
x
2
based on HRTEM and locally Fourier-transformed images appeared
on selected regions, as shown in Fig. 2b-d. These observations agree
well with the previously described DTA results, i.e., the reduction in
the crystallization/densification temperature of the ZrO film by add-
ing Ce atoms.
2
Results and Discussion
Figure 1 shows the measured DTA curves of the ZrO solution
The change in the chemical bonding status of the Zr1−xCe O
2
x
2
mixed with various Ce amounts ͑x = 0, 0.1, 0.3, and 0.5͒, which
films with varying Ce contents was analyzed by performing an XPS
provides valuable information on the thermal behavior of the pre-
measurement, and the scanned Zr 3d, O 1s, and Ce 3d3 core spec-
tra are shown in Fig. 3. The binding energies of the core levels were
/2
cursors and solvents used in this experiment. In the pure ZrO so-
2
lution, a clear exothermic peak located near 372°C was observed,
calibrated by setting the Si 2p peak coming from the Si substrate at
1
2
14
which appears associated with ZrO crystallization and/or densifi-
99.3 eV. For the ZrO sample, the Zr 3d peak originating from
2
2
5/2
cation via a loss of residual organic elements. With increasing Ce
the Zr atom bound to oxygen was located at 182.9 eV, separated by
content in the ZrO solution, a noticeable shift of this exothermic
2.4 eV from the Zr 3d peak. In the Ce-doped ZrO samples, the
2
3/2
2
peak to a lower temperature was induced ͑a shift of ϳ70°C for the
complex Ce 3d3 spectrum was observed ͑shown in Fig. 3c͒ due to
/2
Zr Ce O sample with x = 0.5͒, implying the possible reduction in
the existence of several satellite peaks originating from shakedown
1
−x
x
2
1
5
the crystallization/densification temperature. These results were also
verified by the cross-sectional HRTEM analysis, which is discussed
below.
peaks. With increasing Ce content, we observed a systematic shift
in the position of the Zr core peaks to the lower binding energy by
an amount of ϳ0.4 eV, as shown in Fig. 3a. A similar shift of the
To measure the thicknesses and to investigate the detailed micro-
Ce 3d core peak to the same direction with the Zr 3d core peak
3/2
structural evolution of the sol–gel deposited Zr1−xCe O films with
was also observed ͑see Fig. 3c͒, which can be explained by a
x
2
1
6
various Ce concentrations, cross-sectional HRTEM pictures were
taken, as shown in Fig. 2. The thicknesses of the Zr1−xCe O films
charge-transfer model applied to the ZrSiO system. Because the
4
1
7
Ce–O bond in CeO is more ionic than the Zr–O bond in ZrO , the
x
2
2
2
were measured to be 3.5–4.0 nm for the samples with x values of 0,
.1, 0.3, and 0.5, respectively. Although the processing conditions
were kept the same, a slight increase in the final film thickness with
increasing Ce content was observed. Although its exact origin could
Ce–O and Zr–O bonds in the Zr–Ce–O matrix become more ionic
and covalent, respectively, than in a pure oxide state. In O 1s, two
peaks were identified, one with a low binding energy corresponding
to the metal oxide bonds ͑Zr–O or Ce–O͒ and the other with a high
0
Downloaded on 2015-05-09 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

G144
Journal of The Electrochemical Society, 157 ͑6͒ G142-G146 ͑2010͒
Figure 3. XPS core spectrum of the
Zr1−xCe O films with different Ce con-
x
2
centrations: ͑a͒ Zr 3d, ͑b͒ O 1s, and ͑c͒
Ce 3d3/2. S indicates a satellite peak in
c͒.
͑
binding energy corresponding to the interfacial oxide ͑SiO or sili-
our sol–gel deposited ZrO2 film showed a much lower dielectric
constant of around 10 possibly because of its amorphous nature
having a low atomic density. With increasing Ce doping content, the
extracted dielectric constant of the Zr1−xCe O film increased, as
2
cate͒. The O 1s peak from the interfacial oxide did not show any
significant peak shift with Ce content. However, the peak associated
with the high-␬ layer exhibited a systematic shift to the low binding
energy side with increasing Ce content, as explained by the charge-
x
2
shown in Fig. 4b. Because the dielectric constant of CeO was
1
6
2
transfer model, implying an increase in Ce–O bond content.
20
21
theoretically and experimentally found to be 20–24, i.e., similar
Figure 4a shows the C-V characteristics of the Zr1−xCe O films
x
2
to or lower than that of ZrO , there should be no increase in the
2
with different Ce doping concentrations measured at a frequency of
dielectric constant by a physical mixing of ZrO and CeO . Accord-
100 kHz after forming gas annealing. The measured C-V curves
2
2
ing to the observed HRTEM results ͑see Fig. 2͒, local atomic order-
ing and crystallization of the Zr1−xCexO2 film occurred by adding
CeO2 into the ZrO2 film, and it is believed that this partial
crystallization/densification caused the increase in the dielectric con-
stant. A similar enhancement of the dielectric constant through a
microstructural stabilization process by doping with group IV and
rare-earth elements in ZrO and HfO systems were also reported
exhibited an excellent transition characteristic of the accumulation
to the inversion region. The measured amount of hysteresis during
the bidirectional gate voltage sweep was less than ϳ90 mV irre-
spective of the Ce content, which implies the absence of any Ce-
induced degradation of the charge trapping characteristics. One in-
teresting observation is a systematic linear negative shift of the
flatband voltage ͑VFB͒ by 0.3 V with increasing Ce content up to
x = 0.5 ͑VFB = −0.37 V for pure ZrO2 and VFB = −0.67 V for
2
2
9-11
both by a first-principles calculation and by several experiments.
Zr0.5Ce0.5O ͒, as shown in Fig. 4a and b. A similar phenomenon has
2
1
8
Several works reported that a moderate doping concentration be-
tween 10 and 20 cation atom % is required for the maximum in-
also been observed in the Hf oxide system by adding La atoms,
and our experimental results could suggest a possible application of
9
,10
crease in the dielectric constant.
However, in our system, a con-
Ce as a dopant in ZrO or HfO gate dielectrics to overcome the
2
2
tinuous increase in the dielectric constant with increasing Ce content
was observed. This is believed to be related to the initially poor
Fermi-level pinning effect and to achieve an easy tunability of the
threshold voltage in CMOSFET devices. Another possible origin of
the flatband shift could be either an increase in the positive fixed
charge or a decrease in the negative fixed charge with increased Ce
doping, depending on the initial polarity of the fixed charge existing
atomic density of the ZrO film and to be caused by the subsequent
2
densification effect of the Ce in parallel with the microstructural
stabilization process.
in ZrO2.
The leakage current characteristics were also measured using a
gate injection method, as shown in Fig. 5a. The lower leakage cur-
rent level of the Ce-doped ZrO2 films relative to pure ZrO2 is be-
lieved to be caused by the physical thickness difference of the
high-␬ film and the interfacial oxide, as confirmed by the previous
HRTEM analysis ͑see Fig. 2͒. Although an exact comparison was
not possible due to the thickness change during spin-coating and the
following densification process, Ce doping into ZrO2 film at least
does not deteriorate the leakage current characteristics. In many
In high-␬ metal oxides on Si, the IL is typically formed as a
metal-silicate film having a higher dielectric constant than that of
1
9
pure SiO2. Because a series of Zr1−xCe O samples having differ-
x
2
ent thicknesses is not available at this point, we approximated the
dielectric constant of the interfacial oxide with that of SiO to in-
2
vestigate the overall behavior of the dielectric constant of the
Zr1−xCe O films as a function of Ce doping content. Based on the
x
2
HRTEM and C-V measurement results, the dielectric constants of
the prepared Zr1−xCe O films were calculated as a function of Ce
x
2
doping concentration, and the results are plotted in Fig. 4b. Al-
high-␬ dielectric systems including ZrO films, the Poole–Frenkel
2
though the dielectric constant of the ZrO film depends on the crys-
2
͑PF͒ conduction is usually observed at a moderately high electric
1
9,22
22,23
talline structure and deposition methods, a typical tetragonal ZrO
film was reported to have a dielectric constant of 25–30. However,
2
field,
and it is theoretically expressed by
6
,7
Downloaded on 2015-05-09 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 157 ͑6͒ G142-G146 ͑2010͒
G145
Figure 4. ͑Color online͒ ͑a͒ C-V characteristics of the Zr1−xCe O films with
x
2
different Ce concentrations measured at a frequency of 100 kHz and ͑b͒
variation in the extracted dielectric constant of the Zr1−xCe O films and the
x
2
flatband voltage as a function of the Ce doping concentrations. The interfa-
cial oxide is assumed to be a pure SiO2 film in calculating the dielectric
constant of the Zr1−xCe O2 films. The error bars on the dielectric constant
x
were determined by measuring several MOS devices, and the flatband volt-
age was read from the C-V curve swept from inversion to accumulation after
calculating the flatband capacitance.
Figure 5. ͑Color online͒ ͑a͒ Leakage current characteristics measured under
1/2
a gate injection condition and ͑b͒ ln J/V vs V ͑PF plot͒ relationships of the
Zr Ce O films as a function of the different Ce concentrations. The effec-
1
−x
x
2
tive bias applied to the gate stack is calibrated by subtracting the measured
flatband voltage ͑VFB͒ from the applied gate bias ͑VG͒.
ͱ
− q͓␾ − qV/͑t␲␧ ␧ ͔͒
t o r
AV
t
J =
exp
ͭ
ͮ
͓1͔
rkT
method, and their microstructural and electrical properties were sys-
tematically investigated. With increasing Ce doping content up to
x = 0.5, the crystallization and/or densification temperature is re-
duced according to the DTA analyses. Further, partial crystallization
͑local ordering͒ and possible densification were observed in the HR-
TEM images, which were attributed to one of the possible origins
for the increase in the effective dielectric constant of the Zr1−xCexO2
films measured from C-V characteristics. Ce incorporation into the
where A is the pre-exponential factor, t is the thickness of the di-
electric, ␾ is the trap energy level with respect to the conduction
t
band of the dielectric, ␧ is the relative dynamic dielectric constant
r
2
͑
the square of the refractive index, n ͒, and the coefficient r ranges
from 1 to 2 depending on the quantity of traps in the dielectric.
1
/2
Because the plots of current density ͑J͒/voltage ͑V͒ vs V show a
strong linear relationship ͑Fig. 5b͒, it can be concluded that the PF
conduction is the dominant conduction mechanism in most of the
voltage ranges for all the Zr1−xCe O samples. One interesting fea-
ZrO film caused a negative linear shift of the flatband voltage with
2
increasing Ce content, whereas no noticeable increase in the leakage
current was identified. All the Zr1−xCexO2 samples followed the PF
conduction mechanism and showed a systematic increase in the
slope of the PF plot with Ce doping increase possibly due to the
decrease in the trap sites in Zr1−xCexO2.
x
2
1
/2
ture in this graph is that the slope of the J/V vs V curves continu-
ously increases with increasing Ce content. Because the dynamic
dielectric constants of ZrO and CeO2 are quite similar, around
2
2
1,22
ϳ5.48,
the slope change in the PF plot can only be attributed to
the change in the coefficient r values according to Eq. 1. Yeargan
Acknowledgments
2
3
and Taylor reported that this r value increases from 1 to 2 with
increasing trap density in the dielectric film. Therefore, although the
direct derivation of this r parameter is not available in this experi-
ment, the increase in the PF plot slope may imply that the number of
trap sites in the Zr1−xCe O film decreases with Ce content increase,
This work was supported by the National Research Foundation
of Korea grant funded by the Korean government ͑2009-0066287͒.
Sungkyunkwan University assisted in meeting the publication costs of
this article.
x
2
which coincides well with the aforementioned HRTEM and C-V
analyses used to elucidate the crystallization/densification effect of
References
1
2
. K. Mistry, C. Allen, B. Beattie, D. Bergstrom, M. Bost, M. Brazier, M. Buehler, A.
Cappellani, R. Chau, et al., Tech. Dig. - Int. Electron Devices Meet., 2007, 247.
. G. D. Wilk, R. M. Wallace, and J. M. Anthony, J. Appl. Phys., 89, 5243 ͑2001͒.
3. T. Suntola, in Handbook of Crystal Growth, Vol. 3, Chap. 14, D. T. J. Hurle, Editor,
North-Holland, New York ͑1993͒.
Ce doping in ZrO films.
2
Conclusion
In this paper, ultrathin Zr1−xCe O films having various Ce con-
x
2
4. S. Van Elshocht, A. Hardy, T. Witters, C. Adelmann, M. Caymax, T. Conard, S. De
Gendt, A. Franquet, O. Richard, M. K. Van Bael, et al., Electrochem. Solid-State
centrations ͑x = 0, 0.1, 0.3, and 0.5͒ were prepared by a sol–gel
Downloaded on 2015-05-09 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

G146
Journal of The Electrochemical Society, 157 ͑6͒ G142-G146 ͑2010͒
Lett., 10, G15 ͑2007͒.
14. J. F. Moulder, W. F. Stickle, P. E. Sobol, and K. D. Bomben, Handbook of X-Ray
Photoelectron Spectroscopy, p. 56, Physical Electronics, Eden Prairie, MN ͑1995͒.
15. K. B. Sundaram, P. F. Wahid, and O. Melendez, J. Vac. Sci. Technol. A, 15, 52
5
. I.-C. Cheng and S. Wagner, in Flexible Electronics: Materials and Applications,
Chap. 1, W. S. Song and A. Salleo, Editors, Springer, New York ͑2009͒.
. H. Kim, K. C. Saraswat, and P. C. McIntyre, J. Mater. Res., 20, 3125 ͑2005͒.
. J. Robertson, J. Vac. Sci. Technol. B, 18, 1785 ͑2000͒.
. A. Hardy, S. Van Elshocht, C. Adelmann, T. Conard, A. Franquet, O. Douhéret, I.
Haeldermans, J. D’Haen, S. De Gendt, M. Caymax, et al., Thin Solid Films, 516,
6
7
8
͑
1997͒.
6. M. J. Guittet, J. P. Crocombette, and M. Gautier-Soyer, Phys. Rev. B, 63, 125117
2001͒.
1
͑
1
1
7. S. Gennard, F. Cora, and C. R. A. Catlow, J. Phys. Chem. B, 103, 10158 ͑1999͒.
8. X. P. Wang, H. Y. Yu, M. F. Li, C. X. Zhu, S. Biesemans, Y. C. Yeo, G. Q. Lo, and
D. L. Kwong, IEEE Electron Device Lett., 28, 258 ͑2007͒.
8
343 ͑2008͒.
9
. D. Fischer and A. Kersch, Appl. Phys. Lett., 92, 012908 ͑2008͒.
0. S. Govindarajan, T. S. Boscke, P. Sivasubramani, P. D. Kirsch, B. H. Lee, H.-H.
Tseng, R. Jammy, U. Schroder, S. Ramanathan, and B. E. Gnade, Appl. Phys. Lett.,
1
1
2
9. J. P. Chang and Y.-S. Lin, Appl. Phys. Lett., 79, 3666 ͑2001͒.
0. T. Yamamoto, H. Momida, T. Hamada, T. Uda, and T. Ohno, Thin Solid Films,
9
1, 062906 ͑2007͒.
4
86, 136 ͑2005͒.
1
1. P. R. Chalker, M. Werner, S. Romani, R. J. Potter, K. Black, H. C. Aspinall, A. C.
2
1. F.-C. Chiu, Electrochem. Solid-State Lett., 11, H135 ͑2008͒.
Jones, C. Z. Zhao, S. Taylor, and P. N. Heys, Appl. Phys. Lett., 93, 182911 ͑2008͒.
2. S. Rossignol, Y. Madier, and D. Duprez, Catal. Today, 50, 261 ͑1999͒.
3. S. K. Dey, C. G. Wang, D. Tang, M. J. Kim, R. W. Carpenter, C. Werkhoven, and
E. Shero, J. Appl. Phys., 93, 4144 ͑2003͒.
1
1
22. M.-T. Wang, T.-H. Wang, and J. Y.-M. Lee, J. Electrochem. Soc., 152, G182
͑2005͒.
23. J. R. Yeargan and H. L. Taylor, J. Appl. Phys., 39, 5600 ͑1968͒.
Downloaded on 2015-05-09 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).