Full text of DOI:10.1557/JMR.2002.0257

Effects of IrO2/Pt hybrid electrodes on the crystallization and
ferroelectric performances of sol-gel-derived Pb(Zr,Ti)O3 thin
film capacitors
Seung-Hyun Kim,a) Dong-Yeon Park, Hyun-Jung Woo, Dong-Soo Lee, and Jowoong Ha
Inostek Inc., 356-1 Gasan-dong, Keumchun-gun, Seoul 153-023, Korea
Cheol Song Hwangb)
School of Materials Science and Engineering, Seoul National University, Seoul 151-743, Korea
(Received 8 March 2001; accepted 18 April 2002)
The effects of IrO2/Pt layered hybrid bottom and/or top electrode structures on the
leakage current density versus voltage (J–V), polarization versus voltage (P–V),
ferroelectric imprint, and fatigue properties of chemical-solution-derived Pb(ZrxTi1-x)O3
(PZT, Zr/Ti ס 35/65) thin films were investigated. The best P–V and J–V
performances were obtained from a capacitor with nonhybrid electrodes (Pt/PZT/Pt
capacitor). However, the poor fatigue performance of the capacitor required the
adoption of hybrid electrode structures. A thin IrO2 layer, as thin as 6 nm, which was
inserted between top Pt electrode and PZT layer was sufficient for improving the
fatigue performance without any degradation of the other ferroelectric properties.
However, the same layer adopted on the bottom Pt electrode was not effective in
improving the fatigue performance with degradation in P–V and J–V properties. This
was ascribed to IrO2 layer dissolution into the PZT layer during the crystallization
annealing of the PZT thin film. A thicker IrO2 layer resulted in more serious
degradation.
PZT.4–7 However, these electrodes have generally re-
I. INTRODUCTION
sulted in larger leakage currents and greater susceptibil-
ity to dielectric breakdown than noble metal electrodes.
In our early paper,2 it was demonstrated that when using
such a hybrid electrode system, the film properties are
influenced by the thickness/partial coverage of the oxide
electrode layers. Therefore, optimization research on the
thickness of the IrO2 electrode layer, which allows ex-
cellent fatigue and minimal leakage current values, was
performed and the results are presented here.
Although a number of experimental and theoretical
contributions have been made concerning issues of po-
larization fatigue, leakage current, imprint, and degrada-
tion by forming gas annealing of ferroelectric thin-film
devices,1–3 no clear consensus has emerged regarding
their origins or potential solutions. The purpose of this
research was to investigate aspects of these degradations
which were observed in recent experimental results and
to define and suggest various phenomenological expla-
nations and possible solutions. In an approach to address
these issues, an extensive series of experiments on the
effects of IrO2/Pt layered hybrid bottom and/or top elec-
trode heterostructures on those degradations were per-
formed. It is well known that the choice of electrode
materials used for the lead zirconate titanate (PZT) film
has an impact on ferroelectric film quality and device
performance. In general, the use of conductive oxide
electrodes such as RuO2, SrRuO3, IrO2, or La1-xSrxCoO3
helps to improve the resistance to polarization fatigue for
II. EXPERIMENTAL PROCEDURE
The choice of starting materials and solvent is very
important for producing good-quality thin films using the
chemical-solution-deposition (CSD) technique. A PZT
solution (Zr:Ti ס 35:65) was prepared using lead acetate
trihydrate [Pb(OAc)2.3H2O], titanium isopropoxide
[Ti(O-i-Pr)4], and zirconium n-butoxide [Zr(O-n-Bu)4].
Methanol was used as the solvent for all precursors. A
12% excess of lead was added to the precursor solutions
to compensate for lead loss during the final annealing
process. In the CSD method, the hydrolysis and conden-
sation rate of the solution are important factors that must
be controlled to maximize solution stability. In general,
to reduce the reactivity of hydrolysis and condensation, a
a)e-mail: shkim@inostek.com
b)e-mail: cheolsh@plaza.snu.ac.kr
J. Mater. Res., Vol. 17, No. 7, Jul 2002
© 2002 Materials Research Society
1735
http://journals.cambridge.org
Downloaded: 13 Mar 2015
IP address: 130.126.162.126

S-H. Kim et al.: Effects of IrO2/Pt hybrid electrodes on the crystallization and ferroelectric performances of sol-gel-derived Pb(Zr,Ti)O3
complexing or chelating agent is required.8 In this study,
we used modified alkanolamine as a complexing agent
for the PZT solution.
Alkanolamine is used to promote the formation of ho-
mogeneous products and improve stability of the solu-
tion, by reducing the hydrolysis reaction rate.8,9 It does
not lead to ester reactions and improves resistance to
aging of the solution. In addition, the alcohol–
alkanolamine system has excellent dissolving power for
many inorganic salts and alkoxides. For example, most
of the precursors for Pb-based thin films can be dissolved
within 1–5 min in a solution of alcohol–alkanolamine.
Films were spin coated at 4000 rpm for 30 s and sub-
sequently dried at 450 °C for 10 min and 650 °C for 2
min for each layer. The thickness per layer was 125 nm.
After two layers were deposited, the coated films were
annealed at 650 °C for 30 min in a furnace through a
direct insertion method. The final film thickness was
FIG. 1. XRD patterns of the PZT (Zr:Ti ס 35:65) thin films grown on
IrO2/Pt hybrid electrodes having different IrO2 interlayer thicknesses
from 0 to 30 nm.
250 nm as measured by scanning electron microscopy.
For hybrid electrodes, thin IrO2 was reactively deposited
to three nominal thicknesses, 6, 15, or 30 nm, by reactive
sputter deposition. For the electrical measurements, Pt
structure. Therefore, it is well known that PZT films on
oxide electrodes tend to have a large amount of a non-
ferroelectric second phase and display a rosette micro-
structure.4 Kwok and Desu reported that whenever the
pyrochlore intermediate phase was formed, the pyro-
chlore to perovskite conversion occurred through an in-
terface controlled transformation and the resultant films
were randomly oriented.13 As a consistent trend, the
XRD patterns of the PZT films grown on IrO2 show a
less textured structure, as shown in Fig. 1, even though
peaks from a second phase other than PZT are not de-
tected due to its nanoscale grain size.
Figure 2 shows the FESEM micrographs of the PZT
film surfaces having IrO2 bottom layers with thicknesses
of (a) 0 nm, (b) 10 nm, and (c) 30 nm, respectively. The
figures clearly show a more non-uniform and rosettelike
surface morphology with increasing bottom-IrO2 layer
thickness confirming the XRD results.
More detailed information on the texture evolution and
ferroelectric properties of the PZT films depending on
the types of the hybrid electrodes can be obtained from
composition analysis along the depth direction by Auger
electron spectroscopy (AES).
Figures 3(a) and 3(b) show the AES depth profiles of
the samples having 6-nm-thick bottom IrO2 and 30-nm-
thick IrO2 layers, respectively. The sputter-etching rate
during the analysis was reduced to <10 Å/min consider-
ing the small thickness of the IrO2 layer. Interestingly,
the sample that had a 6-nm IrO2 layer does not show any
discernible Ir signal even at the bottom interface whereas
the sample having a 30-nm-thick IrO2 layer clearly
shows the existence of the IrO2 layer between the PZT
and the bottom Pt electrode. In the Auger electron energy
top electrodes have a diameter of 0.2 mm were fabricated
by sputtering using a stainless-steel shadow mask. After
top electrode fabrication, the capacitors were postan-
nealed at 650 °C for 5 min in air to cure the sputtering
damage. Crystallinity, surface microstructure, leakage
current, P–V hysteresis, and 1 imprint and fatigue char-
acteristics of the films were investigated using x-ray dif-
fraction (XRD), field emission scanning electron
microscopy (FESEM, JEOL, Tokyo, Japan), a Keithley
617 programmable electrometer (Keithley, OH), and an
RT 6000S ferroelectric tester (Radiant, NM), respec-
tively. For the measurements of leakage current, a delay
time of 3 s was adopted.
III. RESULTS AND DISCUSSION
Figure 1 shows the XRD patterns of the PZT (Zr:Ti ס
35:65) thin films as a function of IrO2 thickness on the Pt
electrode. The PZT thin film directly deposited on 111-
oriented Pt shows a highly textured 111 orientation be-
cause the activation energy for PZT nucleation is the
lowest for heterogeneous nucleation on lattice matched
substrates.10,11 The film on the 6-nm IrO2/Pt electrode
shows some 100 orientation but has a strong 111 pre-
ferred orientation, while the film on the thicker IrO2/Pt
reveals random orientation. It is postulated that the acti-
vation energy for nucleation of perovskite PZT is con-
siderably higher than that of growth. Thus, nucleation
and crystallization of a single-phase perovskite PZT film
is much more difficult to achieve on an oxide electrode
than it is on Pt.12 A large difference between the activa-
tion energy for nucleation and growth leads to the for-
mation of rosettes and a two-phase pyrochlore/perovskite
1736
J. Mater. Res., Vol. 17, No. 7, Jul 2002
http://journals.cambridge.org
Downloaded: 13 Mar 2015
IP address: 130.126.162.126

S-H. Kim et al.: Effects of IrO2/Pt hybrid electrodes on the crystallization and ferroelectric performances of sol-gel-derived Pb(Zr,Ti)O3
FIG. 3. AES depth profiles of the samples having (a) 6-nm-thick and
(b) 30-nm-thick bottom IrO2 layers.
in the PZT layer must be approximately 2% from a
simple calculation on the basis of the film thickness ratio.
For these overlapping Ir and Pt peaks, the 2% Ir is almost
FIG. 2. FESEM micrographs of the PZT film surfaces having IrO2
bottom layers with the thickness of (a) 0 nm, (b) 10 nm, and (c) 30 nm,
respectively.
undetectable.
The bottom IrO2 thickness-dependent phase evolution
behavior shown in Fig. 1 can now be more clearly un-
derstood. For the case of the PZT layer on the 6-nm-thick
spectrum, the Ir signal slightly overlaps the Pt signal, so
that the Ir signal was not discernible from the background
if the intensity is not large enough.
IrO2 layer, the initial crystallization must occur on the IrO2
layer, which results in a non-highly 111-textured struc-
ture. However, as crystallization proceeds, the IrO2 cov-
erage on the Pt layer decreases due to dissolution of the
IrO2 layer and thus crystallization occurs on the Pt sur-
face, which produces a highly 111-textured film. How-
ever, for the cases of 15- and 30-nm-thick IrO2 layers, the
full coverage of the layer on the Pt electrode precludes
direct contact of the PZT layer with the Pt resulting in a
nontextured structure. The more random nature of the
PZT film on the 30-nm-thick IrO2 layer than that on
It appears that the IrO2 layer dissolves into the PZT
layer to a certain extent during the crystallization anneal-
ing of the PZT layer at 650 °C, so that the 6-nm-thick
IrO2 layer does not exist at the PZT/bottom Pt interface.
However, a 30-nm-thick IrO2 layer remains at that inter-
face due to the larger thickness after the partial dissolu-
tion. If all of the 6-nm-thick IrO2 layer dissolves into the
250-nm-thick PZT layer, the atomic concentration of Ir
J. Mater. Res., Vol. 17, No. 7, Jul 2002
1737
http://journals.cambridge.org
Downloaded: 13 Mar 2015
IP address: 130.126.162.126

S-H. Kim et al.: Effects of IrO2/Pt hybrid electrodes on the crystallization and ferroelectric performances of sol-gel-derived Pb(Zr,Ti)O3
the 15-nm-thick IrO2 layer suggests that the absolute
amount of dissolved Ir in the PZT also have some influ-
ence on the crystallization behavior.
This thickness-dependent presence of the IrO2 layer
greatly influences the P–V as well as the J–V and fatigue
properties.
Figure 4 shows the J–V curves of PZT thin films with
different IrO2 layer thicknesses at the bottom electrode
interface. The PZT thin films on the Pt and 6-nm-IrO2/Pt
hybrid electrode show an excellent leakage current
density, which is in the order of 10−7–10−8 A/cm2 up to
10 V, while the films on the thicker IrO2 (15 and 30 nm)
bottom hybrid electrode reveal a higher leakage current
density (10−5–10−6 A/cm2) with increasing IrO2 layer
thickness. Even though a rigorous conduction mecha-
nism study was not performed in this study, from our
previous study,14 the conduction of this 250-nm-thick
PZT film must be a interface limited one (such as Schott-
ky emission and tunneling) combined with bulk limited
one (such as space charge limited). Especially, the PZT/
IrO2 contact must have a lower potential barrier height
than that of the PZT/Pt contact due to the smaller work
function of IrO2, resulting in a more bulk limited con-
duction behavior. Therefore, the increase in the leakage
current with increasing IrO2 thickness for the cases of
15- and 30-nm-thick film appears to be either due to an
increase in the defect charge density by larger dissolu-
tion of Ir or to the formation of a conducting second
phase. In general, PZT on an oxide electrode has two
known pyrochlore-type secondary phases.15–17 The first
FIG. 4. J–V curves of PZT thin films having different IrO2 layer
thicknesses at the bottom electrode interface.
FIG. 5. P–V hysteresis loops of PZT filsm with four different electrode stacks (a) Pt/PZT/Pt, (b) Pt top/6-nm-IrO2/PZT/Pt, (c) Pt/PZT/6-nm-
IrO2/Pt bottom, and (d) Pt top/6-nm-IrO2/PT/6-nm-IrO2/Pt bottom as a function of maximum applied voltage from 2 to 5 V.
1738
J. Mater. Res., Vol. 17, No. 7, Jul 2002
http://journals.cambridge.org
Downloaded: 13 Mar 2015
IP address: 130.126.162.126

S-H. Kim et al.: Effects of IrO2/Pt hybrid electrodes on the crystallization and ferroelectric performances of sol-gel-derived Pb(Zr,Ti)O3
FIG. 6. P–V hysteresis loops of PZT thin films with four different electrode stacks (a) Pt/PZT/Pt, (b) Pt/PZT/15-nm-IrO2/Pt bottom, (c) Pt
top/15-nm-IrO2/PZT/15-nm-IrO2/Pt bottom, and (d) Pt/PZT/30-nm-IrO2/Pt bottom as a function of maximum applied voltage from 2 to 5 V.
is an insulating PZT pyrochlore [Pb2(Zr,Ti)2O7-x].
However, this type of pyrochlore would not cause the
high leakage currents, because it has a high resistivity of
about 1010–1011 ⍀-cm The second pyrochlore-type phase
is a PZT pyrochlore-related conducting phase (specifically
for RuO2, PZT pyrochlore–ruthenate [Pb2(Ru,Zr,Ti)2O7-x]
or lead ruthenate (Pb2Ru2O7-x). This would produce a
high conduction path, and its lattice constant is very simi-
lar to PZT pyrochlore. It is, therefore, more likely that
this phase is responsible for the high leakage current
of the PZT films on the oxide electrode/Pt hybrid struc-
ture. This is consistent in that the films on thicker IrO2/Pt
would produce much more of the PZT pyrochlore-related
conducting phase. It is assumed that PZT pyrochlore or
PbO is more reactive with IrO2 than it is with PZT
perovskite and therefore forms a conducting phase very
easily like in the case of RuO2. Furthermore, the in-
creased surface roughness (Fig. 2) with the increasing
bottom-IrO2 thickness also contributes to the increased
leakage current density.
the sample without the IrO2 interlayer. This is be-
cause the contact barrier height increases to that of
Pt/PZT and the interface suppresses the leakage cur-
rent flow.
Figure 5 shows the P–V hysteresis loops of PZT films
with four different electrode stacks (Pt/PZT/Pt, Pt
top/6-nm-IrO2/PZT/Pt, Pt/PZT/6-nm-IrO2/Pt bottom,
and Pt top/6-nm IrO2/PZT/6-nm-IrO2/Pt bottom) as
a function of the maximum applied voltage from 2 to
5 V. It can be seen that the loop shapes for the films on
the four different electrode stacks are almost same,
while the remanent polarization (Pr) values are slightly
different. The measured Pr values of the PZT thin films
crystallized on bare Pt electrode are similar (2Pr of
60 ␮C/cm2 at 5 V), regardless of the top electrode struc-
tures. The PZT thin films grown on IrO2/Pt hybrid
electrode reveal a somewhat smaller Pr value (2Pr
of 48 ␮C/cm2 at 5 V) when compared to the films on Pt.
The data presented above are entirely consistent with the
fact that the Pr values of the PZT films depend on
the film orientation and microstructure, which in turn
appear to be substantially affected by the bottom elec-
trode structure and materials. Other possible factors
which can affect the polarization behavior of PZT
On the other hand, for the case of the 6-nm-thick IrO2
layer, complete dissolution of the layer causes the contact
property to be similar to that of PZT/Pt resulting in al-
most the same low leakage current level as that of the
J. Mater. Res., Vol. 17, No. 7, Jul 2002
1739
http://journals.cambridge.org
Downloaded: 13 Mar 2015
IP address: 130.126.162.126

S-H. Kim et al.: Effects of IrO2/Pt hybrid electrodes on the crystallization and ferroelectric performances of sol-gel-derived Pb(Zr,Ti)O3
capacitors include the presence of second phases or lo-
calized inhomogeneous crystallization of the PZT films
as discussed above.
related to the internal bias field effects previously re-
ported in bulk ceramics. It is reported that this voltage
shift arises from asymmetric behavior of ferroelectrics
induced by an the asymmetric distribution of charge
trapping and/or charged defect and/or defect com-
plexes.18–21 Possible factors affecting the voltage shift
are thermally generated or field injected carriers and
the alignment of defect dipoles, all of which appear to
be thermally activated. In an integrated device, if the
shift is large enough, the coercive voltage becomes too
large in one direction to be switched by the programm-
ing voltage.
It is observed that the voltage shifts of the PZT thin
films slightly increases with increasing IrO2 thickness.
The measured ⌬V values of the PZT thin films are be-
tween 0.52 and 0.57 V, as shown in Fig. 6. This trend of
voltage shifts can be accounted for if the trapped charge
is not confined to the film/electrode interface but is dis-
tributed over a wider area near the interface region.22 The
slight increase in ⌬V with IrO2 thickness might be due to
the increase in the amount of injected charge during the
imprint test or the increased defect density as a result of
IrO2 dissolution.
Figure 6 shows the P–V hysteresis loops of PZT thin
films with thicker IrO2 (15 and 30 nm)/Pt top and bottom
hybrid electrodes for further evidence. It is observed that
the hysteresis loop shapes and Pr values are affected only
by the bottom electrode. The Pr values become smaller
with increasing the IrO2 bottom electrode thickness,
whereas the Pr values are not affected by the top elec-
trode structure, clarifying the above hypothesis. Regard-
less of the top IrO2 thickness, both the Pr values and loop
shapes of these films on Pt are almost consistent with
those of the films with the Pt top and bottom electrodes.
Figure 7 shows the imprint behavior or thermally in-
duced voltage shifts (⌬V) of the PZT thin films with four
different electrode stacks. The capacitors were stored un-
der the −Pr state at 150 °C for 5000 s. In general, the
evaluation standard for imprint is the voltage shift in
the hysteresis loop. The development of a voltage shift
in thin films left in a given polarization state for an
extended period of time can be induced by heating the
capacitor after poling to a positive/negative remanence
( Pr) or by heating the capacitor under an applied
saturating voltage (dc bias). This method is very ef-
fective in understanding the phenomena because it is
Figure 8 shows the AES depth profile results of the
top-Pt/6-nm-IrO2/PZT/Pt sample. It is clearly observed
that the 6-nm-thick IrO2 layer remained at the top interface
FIG. 7. Voltage shifts (⌬V) of the PZT thin films with four different electrode stacks (a) Pt/PZT/Pt, (b) Pt top/6-nm-IrO2/PZT/Pt, (c) Pt/PZT/
6-nm-IrO2/Pt bottom, and (d) Pt top/6-nm-IrO2/PZT/6-nm-IrO2/Pt bottom under the remanence and 150 °C for 5000 s.
1740
J. Mater. Res., Vol. 17, No. 7, Jul 2002
http://journals.cambridge.org
Downloaded: 13 Mar 2015
IP address: 130.126.162.126

S-H. Kim et al.: Effects of IrO2/Pt hybrid electrodes on the crystallization and ferroelectric performances of sol-gel-derived Pb(Zr,Ti)O3
even after the top electrode annealing. This is because the
top 6-nm-thick IrO2 layer was formed after the PZT crys-
tallization annealing was completed and received only
short heat treatment. It appears that the noncrystalline,
gel-state PZT more easily reacts with IrO2 than the well-
crystallized PZT film.
Figure 9 shows the fatigue properties of the PZT thin
films with four different electrode stacks. When cycled
with 5 V and a 500-kHz squared wave pulse, the Pt/PZT/
Pt capacitors show an abrupt polarization decrease after
the 107 cycles as expected, whereas the PZT capacitors
with both top and bottom 6-nm-IrO2/Pt hybrid electrodes
and only a top 6-nm-IrO2/Pt hybrid electrode exhibit an
almost flat fatigue profile up to 109 cycles. The fatigue
performance of the PZT capacitor having only top 6-nm-
IrO2/Pt hybrid electrode is slightly better than that having
the both top and bottom hybrid electrodes. It has been
shown by many researchers that fatigue of PZT thin films
can be substantially alleviated by use of conductive oxide
such as RuO2, IrO2, SrRuO3, or hybrid electrodes, incor-
porated only in both top and bottom electrodes.4–6 How-
ever, our results indicate that a high polarization
endurance behavior with fatigue cycles can be obtained
even on asymmetrical Pt top/6-nm-IrO2/PZT/Pt capaci-
tors as shown in Fig. 9. Recently similar results have
been presented for asymmetrical top SrRuO3/PZT/Pt
and Pt/SrRuO3/PLZT/Pt by Cross et al. and Stolichinov
et al.7,23 It was claimed that in an asymmetrical top hy-
brid electrode system, the physical mechanisms of po-
larization switching depend on the driving voltage
amplitude. The polarization switching endurance under a
high amplitude is governed only by the top SrRuO3 in-
terface, whereas the degradation properties of the bot-
tom Pt interface do not play an important role. On
the contrary, for driving voltage amplitudes lower than
80 kV/cm, the Pt-interface induced degradation is a lim-
iting factor for the fatigue performance. On the basis of
these results, they suggested that the fatigue of switching
polarization of PZT film capacitors is controlled by the
interfaces, and the most probable fatigue mechanism is
related to the inhibition of the near-interfacial nucleation
of the opposite domains by the entrapped mobile carriers.
This suggestion is well matched to our results presented
here. However, the use of only a bottom 6-nm-IrO2/Pt is
less effective in improving PZT fatigue resistance than
that of top 6-nm-IrO2/Pt hybrid electrode. This difference
can be accounted for by the fact that the bottom 6-nm-
thick IrO2 layer does not exist anymore after the crystal-
lization annealing of the PZT layer, as shown in Fig. 3,
whereas the top 6-nm-thick IrO2 layer is still there, as
shown in Fig. 8, and acts as a nucleation site for reverse
domain formation.
FIG. 8. AES depth profile results of the top-Pt/6-nm-IUrO2/PZT/Pt
sample.
IV. CONCLUSION
An optimization study on the layer sequence and
thickness of IrO2/Pt layered hybrid electrode structure on
the J–V, P–V, ferroelectric imprint, and fatigue properties
of chemical-solution-derived PZT thin films were per-
formed. Even though the best P–V and J–V performances
were obtained from capacitors having nonhybrid elec-
trodes (Pt/PZT/Pt capacitor), poor fatigue performance of
the capacitor required the adoption of hybrid electrode
structures. A thick IrO2 layer (>10 nm) badly affects the
crystallization behavior of the PZT layer generating a
nontextured structure with a rough surface morphology
and smaller Pr values. A thin IrO2 layer, as thin as 6 nm,
inserted between the top Pt electrode and the PZT layer
FIG. 9. Fatigue properties of the PZT thin films with four different
electrode stacks.
J. Mater. Res., Vol. 17, No. 7, Jul 2002
1741
http://journals.cambridge.org
Downloaded: 13 Mar 2015
IP address: 130.126.162.126

S-H. Kim et al.: Effects of IrO2/Pt hybrid electrodes on the crystallization and ferroelectric performances of sol-gel-derived Pb(Zr,Ti)O3
5. I.S. Chung, J.K. Lee, W.I. Lee, C.W. Chung, and S.B. Desu, in
was sufficient for improving the fatigue performance
without any degradation in the other ferroelectric prop-
erties. However, the same layer adopted on the bottom Pt
Ferroelectric Thin Films, edited by B.A. Tuttle, S.B. Desu,
R. Ramesh, and T. Shiosaki (Mater. Res. Soc. Proc. 361, Pitts-
burgh, PA, 1995), pp. 249–254.
electrode was not effective in improving the fatigue per-
formance with degradation in P–V and J–V properties
due to the dissolution of the IrO2 layer into the PZT layer
during the crystallization annealing of the PZT thin films.
Therefore, it was determined that the thermal and chemi-
cal stabilities of the oxide electrode layer are critically
important in improving the fatigue performance by
adopting the oxide electrode layer. The adoption of the
conducting oxide layer in the top electrode stack, rather
than the bottom electrode, is much more effective elimi-
nating the degradation in crystallinity, texturing, P–V,
and J–V behaviors. The thickness of the top oxide layer
could be reduced to 6 nm, which is good for high-density
integrated devices.
6. R. Ramesh, H. Gilchrist, T. Sands, V.G. Keramidas, R. Haakenaasen,
and D.K. Fork, Appl. Phys. Lett. 63, 3592 (1993).
7. J. Cross, M. Fujiki, M. Tsukada, K. Matsuura, and S. Ontani, Jpn.
J. Appl. Phys. Lett. 38, L448 (1999).
8. S-H. Kim, C.E. Kim, and Y.J. Oh, Thin Solid Films 305, 321
(1997).
9. Y. Takahashi and J. Yamaguchi, J. Mater. Sci. 25, 3950 (1990).
10. G.J. Willems, D.J. Wouters, H.E. Maes, and R. Nouwen, Integr.
Ferroelectr. 15, 19 (1997).
11. C.J. Kim, D.S. Yoon, Z.T. Jiang, and K.S. No. J. Mater. Sci. 32,
1213 (1997).
12. H.N. Al-Shareef, K.R. Bellur, O. Auciello, and A.I. Kingon,
Ferroelectrics 152, 85 (1994).
13. C.K. Kwok and S.B. Desu, J. Mater. Res. 8(2), 339 (1993).
14. J.C. Shin, J. Park, C.S. Hwang, and H.J. Kim, J. Appl. Phys. 86,
506 (1999).
15. H.N. Al-Shareef, K.R. Bellur, O. Auciello, and A.I. Kingon, Thin
Solid Films 256, 73 (1995).
16. J.M. Longo, P.M. Raccah, and J.B. Goodenough, Mater. Res.
Bull. 4, 191 (1969).
ACKNOWLEDGMENT
This work was supported by National Research Labo-
ratories Program of the Ministry of Science and Tech-
nology of the Korean government.
17. H.N. Al-Shareef, K.D. Gifford, M.S. Ameen, S.H. Rou,
P.D. Hren, O. Auciello, and A.I. Kingon, Ceram. Trans. 25, 97
(1992).
18. G. Arlt and H. Neumann, Ferroelectrics 87, 109 (1988).
19. B.H. Park, S.J. Hyun, C.R. Moon, B.D. Choi, J. Lee, C.Y. Kim,
W. Jo, and T.W. Noh, J. Appl. Phys. 84, 4428 (1998).
20. W.L. Warren, D. Dimos, G.E. Pike, and K.V. Vanheusden, Appl.
Phys. Lett. 67, 1689 (1995).
REFERENCES
1. M. Dawber and J.F. Scott, Appl. Phys. Lett. 76, 1060 (2000);
M. Dawber and J.F. Scott, Appl. Phys. Lett. 76, 3655 (2000).
2. S-H. Kim, J.G. Hong, S.K. Streiffer, and A.I. Kingon, J. Mater.
Res. 14, 1018 (1999).
3. A. Tagantsev, Presented in the 12th International Symposium on
Integrated Ferroelectrics, Aachen, Germany, 12–15, 2000 (un-
published).
21. S-H. Kim, D-S. Lee, C.S. Hwang, D-J. Kim, and A.I. Kingon,
Appl. Phys. Lett. 77(19), 3036 (2000).
22. S-H. Kim, H-J. Woo, J. Ha, C.S. Hwang, and A.I. Kingon, Appl.
Phys. Lett. 78, 2885 (2001).
23. I. Stolichnov, A. Tagantsev, N. Setter, J. Cross, and M. Tsukada,
Appl. Phys. Lett. 74, 3552 (1999).
4. O. Auciello, K.D. Gifford, and A.I. Kingon, Appl. Phys. Lett. 64,
2873 (1994).
1742
J. Mater. Res., Vol. 17, No. 7, Jul 2002
http://journals.cambridge.org
Downloaded: 13 Mar 2015
IP address: 130.126.162.126