scispace - formally typeset
Open AccessJournal ArticleDOI

Dielectric functions and electronic band structure of lead zirconate titanate thin films

TLDR
In this article, the authors measured pseudodielectric functions in the visible-deep ultraviolet spectral range of Pb(ZrxTi1−x)O3 (x=0.56, 0.82) (PZT), Pb0.96O3, Pb 0.91La0.09(Zr0.65Ti0.35)0.98Nb0
Abstract
We measure pseudodielectric functions in the visible-deep ultraviolet spectral range of Pb(ZrxTi1−x)O3 (x=0.2,0.56,0.82) (PZT), Pb0.98Nb0.04(Zr0.2Ti0.8)0.96O3, Pb0.91La0.09(Zr0.65Ti0.35)0.98O3, and Pb0.85La0.15Ti0.96O3 films grown on platinized silicon substrates using a sol-gel method and on (0001) sapphire using a radio-frequency sputtering method. Using a parametric optical constant model, we estimate the dielectric functions (ϵ) of the perovskite oxide thin films. Taking the second derivative of the fitted layer dielectric functions and using the standard critical-point model, we determine the parameters of the critical points. In the second derivative spectra, the lowest band-gap energy peak near 4eV is fitted as a double peak for annealed PZTs due to the perovskite phase. As-grown PZTs have mainly pyrochlore phase and the lowest band-gap peak is fitted as a single peak. We also examine the effect of dopants La and Nb, which substitute at Pb and Zr (Ti) sites, respectively. We found three band gaps E...

read more

Content maybe subject to copyright    Report

Virginia Commonwealth University
VCU Scholars Compass
3-+:81+)3)5,647;:-85/15--815/ ;*31+):1659 -7:6.3-+:81+)3)5,647;:-85/15--815/

Dielectric functions and electronic band structure
of lead zirconate titanate thin lms
Hosun Lee
Virginia Commonwealth University
Youn Seon Kang
Virginia Commonwealth University
Sang-Jun Cho
Virginia Commonwealth University
See next page for additional authors
6336=:019)5,),,1:165)3=6829): 0D79+063)89+647)99<+;-,;-/8-(7;*9
)8:6.:0- 3-+:81+)3)5,647;:-85/15--815/644659
--)5/'!06!-:)31-3-+:81+.;5+:1659)5,-3-+:8651+*)5,9:8;+:;8-6.3-),?18+65):-:1:)5):-:015C349
6;85)36.7731-, 0>91+967>81/0:@  ;*319015/
B198:1+3-19*86;/0::6>6;.68.8--)5,67-5)++-99*>:0--7:6.3-+:81+)3)5,647;:-85/15--815/):$#!+063)89647)99:0)9*--5
)++-7:-,.6815+3;9165153-+:81+)3)5,647;:-85/15--815/ ;*31+):1659*>)5);:0681?-,),41519:8):686.$#!+063)89647)9968468-
15.684):16573-)9-+65:)+: 31*+647)99<+;-,;
6=536),-,.864
0D79+063)89+647)99<+;-,;-/8-(7;*9

Authors
69;5--'6;5!-65)5/!)5/;5066&1)6),196826A")-65/)5/013!66--15/*6
1!;;)1%-1 !5>,-8)5,"<)59
B19)8:1+3-19)<)13)*3-):$#!+063)89647)99 0D79+063)89+647)99<+;-,;-/8-(7;*9

Dielectric functions and electronic band structure of lead zirconate titanate
thin films
Hosun Lee,
a
Youn Seon Kang, Sang-Jun Cho, Bo Xiao, and Hadis Morkoç
Department of Electrical Engineering, Virginia Commonwealth University, 601 W. Main Street, Richmond,
Virginia 23284
Tae Dong Kang and Ghil Soo Lee
Department of Physics, Kyung Hee University, Suwon 449-701, South Korea
Jingbo Li and Su-Huai Wei
National Renewable Energy Laboratory, Golden, Colorado 80401
P. G. Snyder
Department of Electrical Engineering, University of Nebraska-Lincoln, Lincoln, Nebraska 68588
J. T. Evans
Radiant Technology Inc., 2021 Girard Blvd, SE3. Albuquerque, New Mexico 87106
Received 2 June 2005; accepted 3 October 2005; published online 9 November 2005
We measure pseudodielectric functions in the visible-deep ultraviolet spectral range of
PbZr
x
Ti
1−x
O
3
x =0.2,0.56,0.82兲共PZT,Pb
0.98
Nb
0.04
Zr
0.2
Ti
0.8
0.96
O
3
,Pb
0.91
La
0.09
Zr
0.65
Ti
0.35
0.98
O
3
, and Pb
0.85
La
0.15
Ti
0.96
O
3
films grown on platinized silicon substrates using a
sol-gel method and on 0001 sapphire using a radio-frequency sputtering method. Using a
parametric optical constant model, we estimate the dielectric functions
of the perovskite oxide
thin films. Taking the second derivative of the fitted layer dielectric functions and using the standard
critical-point model, we determine the parameters of the critical points. In the second derivative
spectra, the lowest band-gap energy peak near 4 eV is fitted as a double peak for annealed PZTs due
to the perovskite phase. As-grown PZTs have mainly pyrochlore phase and the lowest band-gap
peak is fitted as a single peak. We also examine the effect of dopants La and Nb, which substitute
at Pb and Zr Ti sites, respectively. We found three band gaps E
a
共⬃3.9 eV, E
b
共⬃4.5 eV, and
E
c
共⬃6.5 eV in the order of increasing energy. The E
a
and E
b
band-gap energies were not sensitive
to Zr composition. We discuss the change of critical-point parameters for PZTs in comparison to the
band-structure calculations based on local-density approximation. The near constancy of the lowest
band-gap energy independent of Zr composition is consistent with the band-structure
calculations. © 2005 American Institute of Physics. DOI: 10.1063/1.2128043
I. INTRODUCTION
Perovskite oxide materials have very wide applications:
various sensors, nonvolatile and dynamic random access
memories, tunable capacitors for high-frequency microwave
applications, electro-optic modulators, infrared detectors,
and microelectromechanical systems.
1,2
Hybridized semicon-
ductors including perovskite oxide materials are being inves-
tigated intensively.
3
Among various perovskite oxide materi-
als, PbZr
x
Ti
1−x
O
3
PZT is the most widely
commercialized due to its excellent ferroelectric and piezo-
electric material properties notwithstanding fatigue.
4
The in-
terface degradation between PZT layer and metal electrodes
is known to be the main cause for fatigue of PZT.
5
Other
lead-free perovskite oxide materials such as Ba
1−x
Sr
x
TiO
3
,
SrBi
2
Ta
2
O
9
, and Bi
3.25
La
0.75
Ti
3
O
12
are being investigated as
alternatives because of health and environmental concerns as
well as fatigue-free properties.
Electrical properties of PZTs have been investigated
intensively.
5
However, optical properties of PZT are rela-
tively less investigated.
6,7
Spectroscopic ellipsometry can
measure dielectric functions
of ferroelectric thin films
and subsequently can provide electronic band-structure infor-
mation as well as thickness and microstructure. Several el-
lipsometric investigations have been reported.
7,8
However,
their spectral range was limited to the region below and near
the lowest band-gap region, that is, between 1.5 and 5 eV.
7,8
Ellipsometric studies have not been done which cover the
near-infrared NIR to deep ultraviolet DUV spectral re-
gion.
PZT-based devices show excellent ferroelectric proper-
ties such as high remanent polarization and low coercive
electric field. Semiconducting properties of PZT perovskite
oxides have also been reported by using current-voltage
measurements by Pintile et al.
9
and Boerasu et al.
10
They
noted that a large concentration of defects can act as trapping
centers and that Schottky barriers can be present at contacts
instead.
In order to obtain superior ferroelectric and piezoelectric
properties, perovskite oxides ABO
3
must be annealed after
deposition to increase the perovskite phase. However, the
a
Permanent address: Department of Physics, Kyung Hee University, Suwon
449-701, South Korea, on sabbatical leave; electronic mail:
hlee@khu.ac.kr
JOURNAL OF APPLIED PHYSICS 98, 094108 2005
0021-8979/2005/989/094108/10/$22.50 © 2005 American Institute of Physics98, 094108-1
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
128.172.48.58 On: Tue, 20 Oct 2015 14:18:46

perovskite phase may not result because of the incomplete
transformation from pyrochlore phase A
2
B
2
O
7
, which lacks
ferro- and piezoelectric properties.
11
The properties of PZT films can be modified optically
and electrically by the addition of foreign ions, for example,
La and Nb, substituting for host atoms Pb, Zr Ti兲兴. The
addition of donor dopants reduces the concentration of in-
trinsic oxygen vacancies caused by PbO evaporation during
the growth and annealing of PZT thin films. Electrons from
donors compensate holes from acceptor impurities originat-
ing from Pb vacancies. The transparent
Pb
1−x
La
x
Zr
y
Ti
1−y
1−0.25x
O
3
PLZT films have distinguished
electro-optic effects and excellent relaxor properties, and
have improved various material characteristics: dielectric,
ferroelectric, piezoelectric, electro-optic, and pyroelectric
behaviors.
12
In general, relaxor behavior is characterized by
a broad maximum in the temperature dependence of dielec-
tric property, and a strong frequency dispersion of the per-
mittivity at temperatures around and below the transition
temperature. Nb dopant increases the electrical resistance of
Pb
1−0.5x
Nb
x
Zr
y
Ti
1−y
1−x
O
3
PNZT thin films and produces
polarization hysterisis loops with low coercive fields and
large remanent polarizations.
13
Pb
1−x
La
x
Ti
1−x/4
O
3
PLT also
has interesting properties of dielectric, ferroelectric, pyro-
electric, piezoelectric, and nonlinear electro-optic properties.
Particularly, relaxor behavior is observed.
14
In this work, we measured the pseudodielectric functions
of PZTs, PLT, and doped PZTs. Using a parametric optical
constant POC model, we estimated the dielectric functions
of PZT layers and determined the critical-point CP param-
eters by performing a standard critical-point SCP model
analysis. The determined band-gap values are consistent with
literature values. The lowest band gaps are fitted as a double
peak E
a
and E
b
, and the band-gap energies are almost con-
stant, independent of Zr composition. We also compared the
experimentally determined band-gap values with those of
band-structure calculations of local-density approximation
LDA. The calculated results showed that PZT has a direct
band gap at the X point and that the band-gap energy is
almost constant as a function of Zr composition.
II. EXPERIMENTS
We grew PbZr
x
Ti
1−x
O
3
x=0.56 and 0.82兲共abbreviated
as PZT56 and PZT82 samples on sapphire Al
2
O
3
using rf
magnetron sputtering methods. We also grew
PbZr
0.2
Ti
0.8
O
3
abbreviated as PZT20,
Pb
0.98
Nb
0.04
Zr
0.2
Ti
0.8
0.96
O
3
abbreviated as PNZT,
Pb
0.91
La
0.09
Zr
0.65
Ti
0.35
0.98
O
3
abbreviated as PLZT, and
Pb
0.85
La
0.15
Ti
0.96
O
3
abbreviated as PLT thin films using sol-
gel methods on platinized silicon Pt/TiO
2
/SiO
2
/Si at Ra-
diant Technologies.
We used rf sputtering to grow PZT56 and PZT82 on
sapphire. The growth temperature was measured to be
410 °C, and rf power was 200 W. The argon and oxygen gas
flow rates were 40 and 19 SCCM standard cubic centimeter
per minute, respectively. We used a nominal
Pb
1.2
Zr
0.5
Ti
0.5
O
3
target of 20% excess Pb, adopted a rotat-
ing substrate holder, and used additional Zr and Ti pellets to
control stoichiometry. The PZT samples were annealed at
800 °C in air to obtain the perovskite phase.
PZT, PLT, PNZT, and PLZT thin films with 270 nm
thickness were grown using sol-gel process on platinized
silicon wafer. The structure of the platinized silicon was
Pt150 nm/TiO
2
40 nm/SiO
2
500 nm/100 silicon sub-
strate. The Pt and TiO
2
layers were deposited by e-beam
deposition. PZT layers were crystallized by annealing in
oxygen ambient at a temperature of 650 °C. Ferroelectric
capacitors were fabricated from PZT, PLT, PNZT, and PLZT
samples and tested using ferroelectric testers at Radiant
Technologies. They showed excellent ferroelectric hysteresis
behavior in terms of low coercive bias voltage and large
remanent polarization. For example, the PNZT layer had a
coercive bias voltage of 2 V and remanent polarization of
18
C/cm
2
.
Microstructure and crystal orientation were determined
using a conventional x-ray diffractometer MacScience
Model M18XHF, maximum power 18 kW. The composi-
tions were determined by using energy-dispersive x-ray
spectrometry EDS. We estimated surface roughness with
atomic force microscopy AutoProbe CP Research System,
Thermo Microscope Inc.. Surface morphology of the
samples was diagnosed with scanning electron microscopy.
The x-ray and atomic force microscopy AFM results are
summarized in Table I.
Spectroscopic ellipsometric measurements were per-
formed using variable angle spectroscopic ellipsometry and
vacuum ultraviolet spectroscopic ellipsometry J. A. Wool-
lam Co. at incidence angles of 60°, 65°, 70°, and 75° using
an autoretarder. Multiangle capability increases the accuracy
of layer modeling.
III. RESULTS
A. X ray and microstructure
Figures 1a–1h show the x-ray data of the PZT, PLT,
and doped PZT thin films. PZT56 and PZT82 which were
sputter grown on sapphire and annealed showed higher crys-
tallinity than PZT20, PLT, PNZT, and PLZT samples sol-gel
grown on platinized silicon substrates because the x-ray in-
tensity from the perovskite phase was much higher for the
sputter-grown and annealed samples. As-grown PZTs using
sputter deposition have mainly the pyrochlore phase with a
small proportion of the perovskite phase as shown in Figs.
1a and 1c.
15
Table I lists the main x-ray peaks from the
perovskite phase of annealed PZTs and those of the pyro-
chlore phase from as-grown PZTs. PZTs grown on platinized
silicon wafer using the sol-gel method have mainly 001
perovskite phase, whereas PZTs grown using the sputtering
method have mainly the 111 phase for PZT56 and mainly
the 110 phase for PZT82. Sample microstructure and crys-
tal orientation can be affected by Zr composition, buffer and
substrate effects, annealing condition, and growth methods.
A morphotropic phase boundary of tetragonal to rhombohe-
dral phase transition occurs at Zr=0.53 for bulk PZT
094108-2 Lee et al. J. Appl. Phys. 98, 094108 2005
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
128.172.48.58 On: Tue, 20 Oct 2015 14:18:46

crystals.
16
Note that annealed PZT56 shows a small amount
of
-PbO phase, which formed during annealing, as shown
in Fig. 1b.
17
B. Spectroscopic ellipsometry
We modeled the sample structures as surface roughness
layer, main layer, and substrate. To model the surface rough-
ness layer, we used an effective-medium approximation with
a mixture of the main layer and voids. The effective dielec-
tric function was calculated by the Bruggeman effective-
medium approximation EMA,
18
f
pzt
pzt
eff
pzt
+2
eff
+ f
void
void
eff
void
+
eff
=0, 1
where the effective dielectric function of the rms roughness
layer was denoted by
eff
, and dielectric functions volume
fractions of PZT phase and voids are denoted, respectively,
as
pzt
f
pzt
and
void
f
void
. Details are shown in Table II.
The dielectric function of the main PZT layer was fitted
using the POC model, in which the dielectric function is
written as the summation of m energy-bounded, Gaussian-
broadened polynomials and P poles accounting for the index
effects due to absorption outside the model region.
19
The
advantage of the POC model is that we can determine simul-
taneously both the band-gap parameters and the dielectric
function from the pseudodielectric function. The POC model
equation is Kramers-Kronig consistent, and is given by
=
1
+ i
2
=1
+ i
j=1
m
Emin
Emax
W
j
E
,E,
j
dE
+
j=m+1
m+P+1
A
j
2
E
j
2
, 2
where
,E,
=
0
e
i
E+i2
2
ss
ds
0
e
i
+E+i2
2
ss
ds
=
8
2
e
y
1
2
+ e
y
1
2
erfiy
1
e
y
2
2
e
y
2
2
erfiy
2
兲兴, 3a
y
1
=
E
2
2
, y
2
=
+ E
2
2
, 3b
W
j
E =
k=0
N
P
j,k
E
k
uE a
j
ub
j
E, 4
where ux the unit step function. Here E
j
,
j
, and A
j
are the
energy threshold, broadening, and amplitude, respectively,
for the jth band-gap structure.
The SCP model assumes simple parabolic dispersion re-
lations for the valence and conduction bands and was devel-
oped by Cardona
20
and Aspnes.
21
This model provides accu-
rate CP parameters such as energy threshold, broadening,
amplitude, and excitonic phase angle. The SCP line-shape
equation is given by
22
= C Ae
i
E + i
n
, 5
where the CP is described by the amplitude A, threshold
energy E, broadening , and the excitonic phase angle .
The exponent n takes the values of 1/2 for one-dimensional
1D,0logarithmic, i.e., ln
E +i兲兴 for two-
dimensional 2D, and 1/2 for three-dimensional 3D CP’s.
Discrete excitons are represented by n =−1. Here the exci-
tonic phase angle represents a coupling between the dis-
crete exciton states and continuum band states. To remove
the background contribution, we fit the second derivative of
the dielectric function with respect to energy d
2
/d
2
using the SCP model.
Figure 2 shows the pseudodielectric functions of a
PZTs and b PLT and doped PZTs. The three band-gap
peaks are identified as E
a
, E
b
, and E
c
, respectively. The band-
TABLE I. Thickness and volume fraction of void in surface roughness layer. The thickness of the main layer
was determined by ellipsometry. AFM-measured roughness and the main x-ray peaks are listed with crystal
phase Pv: perovskite, Pr: pyrochlore, substrate, and growth method.
Main
layer
thickness
nm
Roughness
layer
thickness
nm, void
fraction
%
AFM rms
roughness
nm
Main
x-ray
peak
Substrate, growth
method
PbZr
0.2
Ti
0.8
O
3
204.0 9.1, 29.8 2.05 Pv 100 Platinized Si, sol gel
PbZr
0.56
Ti
0.44
O
3
, as grown 367.9 6.2, 22.1 Pr 400 Al
2
O
3
, sputter
PbZr
0.56
Ti
0.44
O
3
339.9 16.7, 28.8 5.27 Pv111 Al
2
O
3
, sputter
PbZr
0.82
Ti
0.18
O
3
, as grown 188.8 9.9, 19.5 Pr 222 Al
2
O
3
, sputter
PbZr
0.82
Ti
0.18
O
3
177.2 13.5, 25.2 11.43 Pv 110 Al
2
O
3
, sputter
Pb
0.98
Nb
0.04
Zr
0.2
Ti
0.8
0.96
O
3
177.2 11.4, 23.3 2.57 Pv 100 Platinized Si, sol gel
Pb
0.91
La
0.09
Zr
0.65
Ti
0.35
0.98
O
3
291.0 10.0, 12 1.07 Pv 100 Platinized Si, sol gel
Pb
0.85
La
0.15
Ti
0.96
O
3
308.2 0.0, 0.0 5.11 Pv100 Platinized Si, sol gel
094108-3 Lee et al. J. Appl. Phys. 98, 094108 2005
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
128.172.48.58 On: Tue, 20 Oct 2015 14:18:46

Citations
More filters
Journal ArticleDOI

Processing, Structure, Properties, and Applications of PZT Thin Films

TL;DR: Ferroelectrics are dielectric materials that have spontaneous polarization in certain temperature range and show nonlinear polarization-electric field dependence called a hysteresis loop as discussed by the authors.
Journal ArticleDOI

Wide bandgap tunability in complex transition metal oxides by site-specific substitution.

TL;DR: The layered ferroelectric bismuth titanate is examined and it is demonstrated that, by site-specific substitution with the Mott insulator lanthanum cobaltite, its bandgap can be narrowed by as much as 1 eV, while remaining strongly ferroElectric.
Journal ArticleDOI

Short-circuit photocurrent in epitaxial lead zirconate-titanate thin films

TL;DR: In this article, the spectral properties of the short-circuit photocurrent (SC-PHC) in the 200-800nm wavelength domain were investigated and it was shown that the sign and magnitude of the signal depend on the internal bias and on the spontaneous polarization direction and value.
Journal ArticleDOI

Post density functional theoretical studies of highly polar semiconductive Pb(Ti 1 − x Ni x )O 3 − x solid solutions: Effects of cation arrangement on band gap

TL;DR: In this paper, a combination of conventional density functional theory (DFT) and post-DFT methods, which include the local density approximation plus Hubbard $U$ ($\text{LDA}+U$), PBE0, and self-consistent $\mathit{GW}$, was used to study the electronic properties of Ni-substituted PbTiO${}_{3}$ (Ni-PTO) solid solutions.
Journal ArticleDOI

Switchable charge-transfer in the photoelectrochemical energy-conversion process of ferroelectric BiFeO₃ photoelectrodes.

TL;DR: BiFeO3 ferroelectric photoelectrodes are focused on to break the limits imposed by common semiconductors and are able to tune the transfer of photo-excited charges generated either in Bi FeO3 or the surface modifiers by manipulating the poling conditions of the ferro electric domains.
References
More filters
Journal ArticleDOI

Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set.

TL;DR: An efficient scheme for calculating the Kohn-Sham ground state of metallic systems using pseudopotentials and a plane-wave basis set is presented and the application of Pulay's DIIS method to the iterative diagonalization of large matrices will be discussed.
Journal ArticleDOI

Ab initio molecular dynamics for liquid metals.

TL;DR: In this paper, the authors present an ab initio quantum-mechanical molecular-dynamics calculations based on the calculation of the electronic ground state and of the Hellmann-Feynman forces in the local density approximation.
Journal ArticleDOI

Ferroelectric ceramics : History and technology

TL;DR: Ferroelectric ceramics have been the heart and soul of several multibillion dollar industries, ranging from high-dielectric-constant capacitors to later developments in piezoelectric transducers, positive temperature coefficient devices, and electrooptic light valves as mentioned in this paper.
Book

Handbook on semiconductors

T. S. Moss
Book

Ferroelectric materials and their applications

TL;DR: Ferroelectric tungsten - bronze type niobate crystals as discussed by the authors is a type of ferroelectric crystal material that can be used to construct piezoelectric composites.
Related Papers (5)
Frequently Asked Questions (16)
Q1. What have the authors contributed in "Dielectric functions and electronic band structure of lead zirconate titanate thin films" ?

In this paper, the pseudodielectric functions in the visible-deep ultraviolet spectral range of Pb ZrxTi1−x O3 x=0.2,0.56,0.,0.8 0. 

The modeled pseudodielectric function mod is a function of the thickness of the thin film, the volume fraction, and the dielectric functions of the void and PZT materials. 

PZTs grown on platinized silicon wafer using the sol-gel method have mainly 001 perovskite phase, whereas PZTs grown using the sputtering method have mainly the 111 phase for PZT56 and mainly the 110 phase for PZT82. 

They attributed the increase of the band-gap energy to the increase of the X1c conduction-band energy due to an increase of the lattice constant as Zr concentration increases. 

The authors suggest that the measured Eb band gap can be due to transitions along the X- and X-M lines originating from these two transitions. 

The fitting parameters were the thickness of the thin film and the volume fraction of each constituent as well as the dielectric functions of PZTs. 

The reason why the band-gap energy values of this work are higher than those available in literature may be attributed to the fact that the band-gap energies were determined accurately using SCP fitting rather than the band edge 2 Ref. 6 or optical band-gap E 2 Ref. 7 values determined from optical absorption spectra. 

The authors note that the fitting in the UV-DUV spectral range is sensitive to the surface roughness layer, but insensitive to the thickness of the main layer because of very small penetration depth of the light. 

The authors assumed a surface roughness layer with a mixture of 50% void and 50% main layer for the layer modeling in the UV-DUV spectral range. 

Because the Zr 4d atomic energy level is about 0.73 eV higher than the Ti 3d level, when the Zr concentration increases, the X3c level also increases as shown in Fig. 

Even though the band structures of the thin films will be affected by strain and electric-field effects, the band calculations will give qualitative information on the Zr composition dependence of PZT band-gap energies. 

In Table II, sputter-grown PZT56 and PZT82 thin films without annealing have essentially single peaks near 4 eV probably because the pyrochlore phase is dominant in the films as shown in x-ray data of Figs. 1 a and 1 c .26 In Table II and Fig. 5, small amounts of Nb and La dopants do not cause an appreciable shift of band-gap energies for Ea, Eb, and Ec. 

IP:128.172.48.58 On: Tue, 20 Oct 2015 14:18:46gap peaks are more clearly discerned in the derivative spectra of the fitted PZT dielectric functions as shown in Figs. 5 and 6. Pseudodielectric functions in the spectral range below 4 eV were dominated by very strong interference patterns for sol-gel-grown PZT thin films on platinized silicon wafers, suggesting very abrupt interfaces. 

The fitting in the DUV spectral range was not satisfactory possibly because the interference pattern in the NIR-visible VIS range was dominant. 

The slight increase or near constancy of the band-gap energies, Ea and Eb, of PZTs as a function of the Zr content suggests that the substitution of Ti by Zr does not change appreciably the electronic band structure of PZT materials. 

The fitted band-edge energy values were smaller by about 0.2 and 0.1 eV, respectively, than Ea, i.e., the ellipsometrically determined CP energy values.