-
nNb
i b
Camndiamat
iNbe p
indicate that single phase was formed for pure KNN while a small
amount of second phase (K Li Nb O ,
area of.5Na0.5
However, a few drawbacks still prevent wide scale industrial
useof KNN-based materials owing to the volatility of the alkali
speciesat high temperatures [2] and hygroscopic nature of the
reactantpowders [3] leading to non-stoichiometry and
inhomogeneouscompositions. All these problems require carefully
controlled
2 3ceramics in order to improve themechanical quality factor Qm
[10e12]. Also, numerous compositional engineering approaches
havebeen explored to optimize the piezoelectric properties of
KNNmaterials like adding BaTiO3 [13], SrTiO3 [14], LiNbO3 [15],
LiTaO3[16], LiSbO3 [17], and ZnO [18] to form new solid solutions.
Amongthese (K0.5Na0.5)NbO3 LiNbO3 (abbreviated as
KNLN)-basedceramics was considered as an excellent candidate for
lead-freepiezoelectric ceramics because of high piezoelectric
propertiessuch as d33 210e300 pC/N and high Curie temperature (450
C).
* Corresponding author.
Contents lists available at
Current Appl
.e
Current Applied Physics 13 (2013) 430e440E-mail address:
[email protected] (R. Rai).a solid solution of ferroelectric
potassium niobate (KNbO3) andantiferroelectric sodium niobate
(NaNbO3). KNN is the only knownlead-free system with a perovskite
structure, which has a higherCurie temperature than the commonly
used lead zirconium titanate(PZT). The high Curie temperature,
relatively strong electrome-chanical coupling factor, presence of
morphotropic phase boundary(MPB) and environmentally friendly
constituents are severaladvantages of KNN system, thus making it a
promising lead-freepiezoelectric material. Moreover, following the
work of Saito et al.[1], KNN-based materials in fact outperformed
other important Pb-free bismuth sodium titanate (BNT)-based
family.
[4,5]. Various methods to overcome the drawbacks pertinent toKNN
have been undertaken targeting mainly the sintering anddensication
behavior of KNN-based ceramics. Densication ofKNN ceramics
signicantly improved by hot-pressing, with sin-tered ceramics
reaching w99% of the theoretical density and theresulting
piezoelectric constant nearly twice the value of conven-tionally
sintered KNN [6,7]. Dense ceramics with a high piezo-electric
constant of 148 pC/N were achieved by Li et al. [8] usingadvanced
processing methods compared to 90 pC/N for undopedKNN prepared
using conventional methods [9]. Sintering aids likeCuO and Bi O
have also been used to modify pure and doped
KNNKeywords:CeramicsImpedance spectroscopyX-ray
diffractionDielectric propertiesPhase transactions
1. Introduction
Recently, much attention in theceramics has been focused on
(K01567-1739/$ e see front matter 2012 Elsevier
B.V.http://dx.doi.org/10.1016/j.cap.2012.09.009analysis was
performed using an equivalent circuit model. The impedance response
in pure KNN andKNLN ceramics could be deconvoluted into two
contributions, associated with the bulk (grains) and thegrain
boundaries. Activation energies for conductivity were found to be
strongly frequency dependent.The activation energy obtained from
dielectric relaxation data was attributed to oxygen vacancies.
FromPFM we found that the composition with 6.5 wt.% LN displays
stronger piezocontrast as compared topure KNN implying an evidence
of a pronounced piezoelectric coefcient.
2012 Elsevier B.V. All rights reserved.
lead-free piezoelectric)NbO3 (KNN) which is
manufacturing conditions and low reaction temperatures
causingpoor densication behavior. The high dielectric losses
related to theA-site vacancies and oxygen deciencies resulting from
the evap-oration loss of alkali ions are also detrimental to the
KNN ceramicsAccepted 12 September 2012Available online 18 September
2012 structure from orthorhombic to tetragonal with increase in LN
content. The electrical behavior of the
ceramics was studied by impedance spectroscopy technique in the
high temperature range. Impedance10 September 20126 4 10 30
w3%) was present in LN doped KNN ceramics. Phase analysis
indicated the change in the crystalImpedance spectroscopy and
piezorespoof lead-free (1 x) K0.5Na0.5NbO3 xLiRadheshyam Rai a,*,
Indrani Coondoo a, Rashmi RanaDepartment of Materials and Ceramics
Engineering and CICECO, University of Aveiro,b Ferroelectric
Research Laboratory, Department of Physics, AN College, Patna
800013, IcDepartment of Mechanical Engineering and Centre for
Mechanical Technology & Auto
a r t i c l e i n f o
Article history:Received 16 May 2012Received in revised form
a b s t r a c t
(1 x) K0.5Na0.5NbO3 xLferroelectric ceramics wer
journal homepage: wwwAll rights reserved.se force microscopy
analysisO3 ceramics
, Igor Bdikin c, Seema Sharma b, Andrei L. Kholkin a
pus Universitario de Santiago, 3810-193 Aveiro, Portugal
ion, University of Aveiro, 3810-193 Aveiro, Portugal
O3 (where x 0.0, 5.0, 5.5, 6.0, and 6.5 wt.%) (KNLN) perovskite
structuredrepared by the solid-state reaction method. X-ray
diffraction patterns
SciVerse ScienceDirect
ied Physics
lsevier .com/locate/cap
-
Until now, however, there is no report on detailed impedance
andpiezoresponse force microscopy (PFM) analysis in KNLN
ceramics.Consequently, the purpose of this paper is to investigate
the effectof the amount of LiNbO3 onmicrostructure, dielectric,
electrical andlocal piezoelectric properties of KNN ceramics.
2. Experimental procedure
Polycrystalline samples of (1 x) K0.5Na0.5NbO3 xLiNbO3(where x
0.0, 5.0, 5.5, 6.0, and 6.5 wt.%) (KNLN) ceramics wereprepared by
the conventional mixed oxide route. The startingmaterials were high
purity metal oxide or carbonate powders,K2CO3 (>99.8%), Na2CO3
(>99.8%), Li2CO3 (>99.99%), and Nb2O5(>99.9%). The powders
were weighed according to the stoichio-metric proportion for the
required compositions andmixed for 48 husing propan-2-ol and
zirconia media. The powders were calcinedat 920 C for 4 h, after
that the resulting powders were mixedthoroughly with a PVA binder
solution. Cylindrical pellets wereprepared in a uniaxial press with
a 10 mm diameter die at 100 MPaand then sintered at 1000 C for 4 h
and nally cooled to roomtemperature at the rate of 180 C/h. Silver
paste as electrodes wereapplied on the top and bottom surfaces on
the samples for theelectrical measurements.
The formation and quality of compounds were veried with X-ray
diffraction (XRD) technique. The XRD patterns of the
compounds were recorded at room temperature using X-raypowder
diffractometer (Rigaku Miniex, Japan) with CuKa radia-tion (l
1.5405 A) in a wide range of Bragg angles 2q(20 2q 60) at a
scanning rate of 2 min1. SEMmicrographs ofthe fractured surface of
the sintered samples were obtained withthe help of JEM-2000FX (JEOL
Ltd) scanning electron microscopeoperated at 20 keV. The dielectric
properties of the samples weredetermined from RT (room temperature)
to 450 C using animpedance analyzer (PSM Impedance Analyzer 1734)
from 100 Hzto 1 MHz. A commercial Atomic Force Microscope (AFM,
Multi-mode, Nanoscope IIIA) was used for the ferroelectric
domainimaging. The microscope was equipped with an external
lock-inamplier (SR-830, Stanford Research) and a function
generator(FG120, Yokogawa). Stiff conducting Si cantilevers (42
Nm1, PPP-NCHR, Nanosensors) were used for the measurements
performedin ambient environment. The tip had the shape of a
polygon-basedpyramidwith the height of 10e15 mmand the radius rtipz
10 nm atthe apex. The PFM studies were carried out with Uac 5 V
andf 50 kHz applied voltage on the tip.
3. Results and discussion
Fig. 1 shows XRD of the KNN LN ceramics for
differentcompositions with x 0.0, 5.0, 5.5, 6.0, and 6.5 wt.%. The
crystalstructure changes from orthorhombic to tetragonal with
increasing
In
te
nsity ( a. u
. )
KNN
KNN - 5.0LN
KNN - 5.5LN
KNN - 6.0LN
KNN - 6.5LN
(1
01
)
(0
10
)
(0
02
)
(1
11
)
(2
02)
(0
20
)
(2
12
)
(1
21
)
(1
13
)
(3
11)
2)
In
te
nsity ( a. u
. )
KNN
KNN-5LN
KNN-5.5LN
KNN-6.0LN
KNN-6.5LN
Tetragonal phase appears
3. LN
a b
R. Rai et al. / Current Applied Physics 13 (2013) 430e440 43120
30 40 50 60
2 (degrees)
3.94
3.96
3.98
4.00
4.02
4.04
0 1 2
Latti
ce Pa
ram
eter
((
Conc
a b c
densitycFig. 1. (a) XRD of the (1 x) K0.5Na0.5NbO3 xLiNbO3
ceramics and (b) enlarged XRD pacompositions x 0.0, 5.0, 5.5, 6.0
and 6.5 wt.% respectively.44 45 46 47
(0
20)
(2
0
2 (degrees)
4 5 6
1.6
1.8
2.0
2.2
2.4
2.6
2.8
3.0
(wt%)
gm/c
m3
tterns of the ceramics for 2q 44e48 (c) lattice parameter and
density of different
-
x. Since the crystal structure of KNN (perovskite structure) is
verydifferent from LiNbO3 (ilmenite structure), the single
phaseformation suggests that Li and Nb3 have diffused into the
KNN
angles implying an increase in the lattice parameters
withincreasing x. This may be attributed to the smaller ionic radii
of Li
as compared to those of Na (1.16 A) and K (1.52 A). In the
Fig. 2. SEM micrograph illustrating typical microstructure for
(1 x) K0.5Na0.5NbO3 xLiNbO3 ceramics of different compositions x
0.0, 5.0, 5.5, 6.0 and 6.5 wt.%.
R. Rai et al. / Current Applied Physics 13 (2013)
430e440432lattice, with Li entering the (Na0.5K0.5) site and Nb5
occupyingthe B site, to form a homogeneous solid solution. As x
increases,a tetragonal phase appears and increases continuously. It
is alsonoted that the diffraction peaks shift slightly toward low
diffractionFig. 3. (a) Variation of dielectric constant, (b) loss
(tan d) with temperature (c) plots of peak(1 x) K0.5Na0.5NbO3
xLiNbO3 ceramics of different compositions x 0.0, 5.0, 5.5, 6.0
anorthorhombic zone, the lattice constants c and a have very
closevalues, which explains the two peak splitting of (200)
reections ata 2q 45.5 for an orthorhombic structure, rather than
three peaksplitting. Lattice constants and density of KNN-LN
ceramicsr and Tc and (d) r and transition temperature vs. doping
concentration at 100 kHz ford 6.5 wt.%.
-
depending on LiNbO3 concentration were calculated as shown
inFig. 1(c). Lattice constants of KNN ceramics in
orthorhombicsymmetry were calculated as a 3.9551 A, b 3.95765 A
andc 4.0502 A, which are similar to the bulk KNN. As x changes
thecrystal structures changed from orthorhombic to
tetragonalsymmetry, which indicated that MPB coexisted orthorhombic
andtetragonal phase. The small XRD peaks at 33 and 52 for KNNdoped
with for x 5.5 and x 6.0% are attributed to an ortho-rhombic phase
of Nb2O5 (O-Nb2O5) corresponding to JCPDS lenumber 527-1313. These
peaks are probably due to the formation ofNb-rich second phase
leading to the creation of A-site vacancies inthe ABO3 perovskite
structure.
Fig. 2 presents the scanning electron micrographs for
fracturedsurfaces of KNN xLN with 0.0, 5.0, 5.5, 6.0, and 6.5 wt.%.
Thefracture surfaces show homogeneous grains with dense
structure.The grain size of all samples is of the order of 2e4
mm.
The temperature dependence of the permittivity and loss shownin
Fig. 3(a and b) conrms the presence of a phase transition
fromorthorhombic to tetragonal at w200 C for pure KNN and
tetrag-onal to cubic at w380 C. However, the composition
containing6.5 wt.% LN, the peak related to phase transition
temperature fromthe orthorhombic to the tetragonal is not observed.
It is probablylocated below room temperature as has been observed
in the caseof 7% lithium doped KNN ceramics [19]. It is also
observed that with6.5 wt.% LN in KNN, the dielectric permittivity
is reduced while theCurie temperature increased (Fig. 3c and d).
The loss tangent (tan d)is also found to decrease with LN addition.
A-site vacancy createdby the Nb rich phase (as is evident from XRD
spectra), increases the
dielectric loss factor for x > 5.5% while the room
temperaturedielectric constant value and the dielectric loss value
are almostunaffected by the introduction of vacancies. The A- and
B-siteelements have valence states of 1 and 5, respectively,
whileoxygen has 2. In KNN ceramics, Nb is believed to be
incorporatedat the B-site of the perovskite lattice and therefore
acts as anacceptor. Oxygen vacancies are formed and when they
combinewith Nb ions, defect dipoles are created. These defect
dipoles movealong the direction of polarization and are based on
the principle ofsymmetry conforming of point defects and lead to
hard dopingeffects. It is well known that piezoelectrics with low
permittivityare useful for energy harvesting because the gure of
meritdescribes the operation of the energy harvesting system
[20].According to this, a material with low permittivity will have
highvoltage induced by strain. In this regard, the composition
with6.5 wt.% LN is expected to exhibit a higher hij.
Fig. 4(a and c) shows variation of the real part of
impedance(Z0) as a function of frequency (103e106 Hz) for
differenttemperatures and compositions. The pattern shows a
sigmoidalvariation as a function of frequency in the low frequency
regionfollowed by a saturation region in the high frequency region.
Thissuggests of the manifestation of the mixed nature of
polarizationbehavior in the material. The merger of the real part
of impedance(Z0) in the frequency domain for all temperatures
indicatesa possibility of the release of the space charge as a
result oflowering in the barrier properties of the material. These
resultsindicate that the electrical conduction will increase with
rise intemperature and the phenomenon is dependent on the release
of
60
80 5LN 5.5LN 6LN 6.5LN
RT
300
400 5LN 5.5LN 6LN 6.5LN
RT
0
a b
R. Rai et al. / Current Applied Physics 13 (2013) 430e440 43310
100 1000
0
20
40
Frequency (kHz)
Z'(k
)
10 100 1000
5
10
15
20
Frequency (kHz)
Z'(k
)
5LN 5.5LN 6LN 6.5LN
400 0Cc
0Fig. 4. (a, c) Variation of real part impedance (Z ) and (b, d)
imaginary part of the impedanceramics with frequency at different
temperatures.10 100 1000
0
100
200
Frequency (kHz)
z"(k
)
10 100 10000
2
4
6
8
Frequency (kHz)
z"(k
)
5LN 5.5LN 6LN 6.5LN
At 400 0Cd
00ce (Z ) of (1 x) K0.5Na0.5NbO3 xLiNbO3 (where x 0.0, 5.0, 5.5,
6.0, and 6.5 wt.%)
-
the space charge. The extent of steepness in the low
frequencyregion is observed to have a very strong dependence of Z0
on thecomposition irrespectively of temperature. Further, the
imped-ance value is observed to increase rst in the low frequency
regionwith rise in temperature followed by a relative decrease
atsubsequently higher temperatures. The result may be related toa
change in the charge-ordering pattern with temperature; theeffect
showing clearly perceptible change in its behavior in thelow
frequency region. A decreasing trend of Z0 with temperaturerise
suggests the presence of negative temperature coefcient
ofresistance (NTCR) in the material in the low frequency region
buttends to merge in the high frequency region at almost
alltemperatures. These results indicate a possibility of increase
in acconductivity with increasing temperature in the high
frequencyregion, possibly due to the release of space charge. Fig.
4(b and d)exhibits frequency and temperature dependence of
imaginarypart of impedance (Z00) for KNLN solid-solutions. The
trend of thevariation of Z00 with frequency is the typical of the
presence of theelectrical relaxation in the materials which is
temperature-dependent. The asymmetric broadening of the peaks
suggeststhe presence of the electrical process in the material with
a spreadof relaxation time. The shifting of the peaks indicates
that the netrelaxation time is decreasing with the increase in
temperature. Inlow temperature/frequency range, Z00 has very high
value for allthe compositions which decreases with increasing
frequency andattains low values. This is an interesting (unusual)
trend in thevariation of Z00 with frequency/temperature for KNLN
ceramicsamples. The appearance and nature of peaks at a
characteristicangular frequency umax ( 2pfmax) provides information
on thetype and strength of the dielectric relaxation phenomenon
occurring in the material. The nature of this pattern suggests
thepresence of a weak dielectrical relaxation above 400 C in
KNLNsystem. At low temperatures (400 C), the spectrum showsa
monotonous decrease having dispersive nature in the lowfrequency
region followed by a plateau at higher frequencies. Thisobservation
clearly testies the presence of electronic, dipolar andspace charge
polarizations in the material, and the appearance ofthe peak(s)
being a result of dipolar contribution.
Fig. 5(a and c) shows the variation of real part of modulus asa
function of frequency at different temperatures for all
composi-tions of KNLN. The variation of M0 with frequency for all
thesamples shows a dispersion tending toward MN (the
asymptoticvalue of M0 at higher frequencies) and dispersion shifts
towardhigher frequency side with increasing temperature.
Monotonousdispersionwith increasing frequency at lower
temperaturesmay bedue to the short range mobility of charge
carriers. It comprises ofthe features such as a very low value
(approaching to zero) ofM0 inthe low frequency region and a
continuous dispersion withfrequency. Such results may possibly be
related to a lack ofrestoring force governing the mobility of the
charge carriers underthe action of an induced electric eld. This
behavior supports thatas the frequency increases; each ion moves a
shorter and shorterpath of electric eld, until the electric eld
changes so rapidly thatthe ions only rattle within the connement of
their potentialenergy wells. The value of M0 decreases with
increasing tempera-ture in the studied frequency range. Fig. 5(b
and d) depicts variationof imaginary part of modulus (M00) with
frequency. High frequencyrange side represents that ions are
separately conned to theirpotential wells, and the ion can make
only the localized motionwithin the wells [21,22].
R. Rai et al. / Current Applied Physics 13 (2013) 430e4404340
00Fig. 5. (a, c) Plot of real part of modulus (M ) and (b, d)
imaginary part of modulus (M ) withceramic.frequency of (1 x)
K0.5Na0.5NbO3 xLiNbO3 (where x 0.0, 5.0, 5.5, 6.0, and 6.5
wt.%)
-
Fig. 6(aee) shows complex impedance spectrum (Nyquist plot)of
samples of different compositions at 300, 350, 400, 425 and450 C,
respectively. We have tted the experimental data (Fig. 6f)for all
the compositions with the equivalent circuit usingcommercially
available software Z Simp Win Version 2. There isa close agreement
observed between experimental and tted data.It is observed that at
lower temperatures a single semi-circular arcappears. This single
semicircular arc suggests the presence of graininterior (bulk)
property of the material [23]. However, at highertemperatures,
another arc is present, and the spectrum comprises
two semicircular arcswith their centers lying below the real
axis forall compositions (temperatures above 400 C). The high
frequencysemi-circle (rst arc) can be attributed to the bulk (grain
interior)properties of the material arising due to a parallel
combination ofbulk resistance (Rb) and bulk capacitance (Cb). The
low frequency(second) semicircular arc of the impedance spectrum at
elevatedtemperatures (pattern at 425 and 450 C, respectively) has
beenattributed to the presence of grain boundary arising due toa
parallel connection of grain boundary resistance (Rgb)
andcapacitance (Cgb). Two semi-circular arcs of the impedance
0 200 400 6000
100
200
300
400
500
600
z'(k)
z"(k
)
5LN 5.5LN 6LN 6.5LN
300 0C
0 50 100 150 200 2500
50
100
150
200
250
z'(k)
z"(k
)
5LN 5.5LN 6LN 6.5LN
350 0C
0 5 10 15 20 250
5
10
15
20
25
z'(k)
z"(k
)
5LN 5.5LN 6LN 6.5LN
400 0C
0 2 4 60
2
4
6
z'(k)
z"(k
)
5LN 5.5LN 6LN 6.5LN
425 0C
4
2
3
4
"(k
)
5LN 5.5LN 6LN 6.5LN
450 0C
15
20Z measured- 5LN Z calculated- 5LN Z measured- 5.5LN Z
calculated- 5.5LN Z measured- 6LN Z calculated- 6LN Z measured-
6.5LN Z calculated- 6.5LN
)
(Q)
(R)
(Q)
(R)
a
c
e f
d
b
R. Rai et al. / Current Applied Physics 13 (2013) 430e440 4350 1
2 30
1
z'(k)
zFig. 6. (aef). Variation of real and imaginary part of
impedance of different (1 x) K0.5Na0.5Nand 450 C respectively.0 5
10 15 200
5
10
' ()
'' (bO3 xLiNbO3 (where x 0.0, 5.0, 5.5, 6.0, and 6.5 wt.%)
ceramics at 300, 350, 400, 425
-
spectrum may be due to grain bulk and grain boundary
contribu-tions, which is consistent with the brick-layer model for
a poly-crystalline material [24]. These semicircular arcs can be
expressedin terms of an equivalent circuit (Fig. 6f) built up by
cascading oftwo parallel resistance and capacitance connected in
series (RCcircuits), each being responsible for a semi-circle. As a
result, boththe grain bulk and the grain boundary phenomena occur
simulta-neously with increasing temperature. This effect may be
related tothe change in microstructure governed predominantly by
thethermal state of the material, which throws a signicant light
onthe relationship of temperature-dependent microstructure
andelectrical properties of the materials. The intercept of the
semi-circular arc on the real axis gives the dc resistance of the
material. Itis seen from Fig. 6 that the dc resistance decreases
with the increasein temperature as well as the LN content in the
KNN ceramics. The
Table 1The value of bulk resistance (Rb), grain boundary
resistance (Rgb), bulk capacitance(Cg) and grain boundary
capacitance (Cgb) of LN doped KNN ceramics of differentcompositions
x 0.0, 5.0, 5.5, 6.0 and 6.5 wt.%.
x Temp (C) Rg (kU) Rgb (kU) Cg (nF) Cgb (nF)
5 400 4.668 7.087 3.683 2.426425 2.174 2.978 5.765 3.732450
1.539 2.279 2.346 1.921
5.5 400 13.133 22.459 2.859 1.672425 2.029 2.978 4.536 3.089450
1.468 1.879 2.102 1.642
6 400 3.680 4.885 4.006 3.081425 2.249 3.228 3.504 2.441450
1.942 3.667 2.530 1.334
6.5 400 9.348 14.596 1.839 1.178425 3.845 6.438 4.471 2.670450
1.970 3.139 2.144 1.345
Fig. 7. (aee) Variation of real and imaginary part of modulus
with temperature of (1
R. Rai et al. / Current Applied Physics 13 (2013) 430e440436x)
K0.5Na0.5NbO3 xLiNbO3 (where x 0.0, 5.0, 5.5, 6.0, and 6.5 wt.%)
ceramics.
-
values of the electrical parameters by tting the measured
datawith the equivalent circuit at 400 C are given in Table 1.
Theobserved depressed semicircular arcs having centers lying
belowthe real impedance (Z0) axis is due to the presence of
distributedphase elements. The relaxation process associated with
thisobservation is non-ideal in nature. The non-ideal behavior
origi-nated from several factors grain orientations, like grain
sizedistributions, grain boundaries, atomic defect distribution
andstressestrain phenomena.
A relative comparison of the complex impedance spectrum
andmodulus spectrum for all the concentrations of LN in KNN
showsone arc in the modulus pattern with different radii (Fig. 7)
whereastwo arcs in the impedance spectrum. This result may be
explainedin term of the capacitance effect by the fact that in an
equivalentcircuit comprising of two parallel RC circuit elements
(connected inseries), the resistance of one may be much lower than
that of theother, but the two capacitance values may not be too
differentleading to a well resolved arc pattern in the modulus
spectrumunlike that in the impedance pattern [23, 24]. Furthermore,
themodulus spectrum shows a notable change in the shape
withincreasing temperature suggesting a probable change in
thecapacitance value of the KNLN material as a function of
tempera-ture as well as concentration of LiNbO3. This suggests that
thedielectric behavior of the solid-solution is controlled by the
LiNbO3addition.
The variation of normalized complex electric impedance
withnormalized frequency is shown in Fig. 8. This is known as a
mastercurve, which enables us to shed light on the dielectric
behavior ofthe material as a function of temperature. The overlap
of the curves
at different temperatures into a single master curve indicates
thatthe relaxation describes the same mechanism at
differenttemperatures. This behavior is represented by the
non-exponentialtype of conductivity relaxation [25]. A
non-exponential typeconductivity relaxation suggests the
possibility that the ionmigration takes place via hopping mechanism
accompanied bya consequential time dependent mobility of other
charge carriers ofthe same type in the vicinity. The master modulus
shows (1) theslight shift in the impedance peak pattern, (2) the
similar shape andpattern with slight variation in full width at
half maximum(FWHM) with increasing temperature, and (3) the
asymmetricnature of the curves which is a characteristic of the
non-exponential behavior of conductivity. The FWHM of the peaks
iswider than that of the Debye peak, which also suggests the
pres-ence of non-Debye-type relaxation phenomena.
Fig. 9 shows the variation of AC conductivity (sac)
withtemperature at 10 kHz frequencies for all samples. The
nanoappreciable variation of sac with temperature indicates
themultiple relaxations. The ac conductivity is dependent on
thecapacitance of the sample given by the expression.
sac 6000
where 0 is the vacuum dielectric constant, 6 is the
angularfrequency, 0 and 00 are the real and imaginary part (the
dielectricloss) of the dielectric permittivity. The AC conductivity
increaseswith increase in temperature showing a negative
temperaturecoefcient of resistance (NTCR) behavior. Arrhenius
equation wasused for the calculation of activation energy for the
ceramics,
R. Rai et al. / Current Applied Physics 13 (2013) 430e440
437Fig. 8. (aed) Impedance scaling behavior of (1 x) K0.5Na0.5NbO3
xLiNbO3 (where x 0.0, 5.0, 5.5, 6.0, and 6.5 wt.%) ceramics in the
master curves.
-
bO3
R. Rai et al. / Current Applied Physics 13 (2013) 430e440438s
s0Ea=kBT
sac is the AC conductivity, Ea is the activation energy, kB is
theBoltzmann constant, T is the temperature in Kelvin scale and s0
isthe dc conductivity.
The activation energy calculated at 10 kHz in the
temperaturerange (390e330 C) and (200e150 C) is between 1.30 eV
and0.55 eV respectively. The low value of activation energy
obtainedcould be attributed to the inuence of electronic
contribution to theconductivity; The increase in conductivity with
temperature may
Fig. 9. Temperature dependence of the sac of (1 x) K0.5Na0.5NbO3
xLiNbe attributed to the oxygen vacancies due to the loss of
oxygenduring the sintering process and leave behind free
electronsmaking them n type semiconductor [26e28]. In oxide
ferroelec-trics, doubly charged oxygen vacancies are the most
mobile chargesand play an important role in the conduction process.
The volatil-ization of A-site elements (Na and K) will result in
the deciency inA-site cations, and thereafter the generation of
oxygen vacanciesdue to the valence balance, which may alter the
lattice distortionand thus inuence the dielectric behavior of these
ceramics.
Fig. 10(a) shows a representative AFM image of the
ceramicsample. Following the topography acquisition, the AFM
wasswitched to the piezoresponse force microscopy (PFM) regime
inwhich the conducting tip is scanned in contact mode while an
acvoltage (Vac) is applied between the tip and Ag electrode (Fig.
10b).
Fig. 10. AFM images of KN 6.5 LN ceramic samples: (a) topogrIn
these conditions, we could measure both out-of-plane (OOP)
andin-plane (IP) polarization components [29]. After the scan we
wereable to detect IP signal due to a shear component of the
piezotensor(d15), corresponding to the polarization parallel to the
surfacesample, and OOP that reects polarization perpendicular to
thesurface [normal component of the piezotensor (d33)].
Comparisonof the topography (Fig. 10a) and PFM (Fig. 10b and c)
images atteststhat the polishing scratches and other corrugations
of the samplesurface are not reected on the corresponding PFM
images thusgiving an evidence of the piezoelectric nature of the
contrast. ForOPP component (Fig. 10b), the contrast is roughly
proportional to
(where x 0.0, 5.0, 5.5, 6.0, and 6.5 wt.%) ceramics at 10 kHz
frequencies.the effective d33 coefcient and determined by the
projection of thepolarization vector P on the normal N to the
ceramic surface. Thedark and bright contrasts correspond to the
polarization headdirected to the sample bulk or to the surface,
respectively. IP pie-zoresponse image (Fig. 10c) is complementary
to the OPP pictureand reects the effective shear d15 piezoelectric
coefcientproportional to the corresponding IP component of
thepolarization.
Ferroelectric domain distributions distributions of the pure
and6.5 wt.% LN concentration samples are shown in Fig. 11(a and
b).The different grains of the respective materials are represented
bythe colored regions in the gure. The histograms of the OPP
PFMsignal from images are represented in Fig. 11c. It has a
clearasymmetric shape with a maximum slightly shifted toward
the
aphy, (b) OPP PFM, and (c) IP PFM. Uac 5 V, f 50 kHz.
-
ed PR. Rai et al. / Current Applinegative or positive values. It
means that the number of grains withthe polarization head
terminated at the free surface exceeds thenumber of oppositely
oriented dipoles. Under large magnication,a shoulder on the
positive slope was observed that may representa maximum in the
number of grains having the same effectivepiezoelectric coefcient.
Since the sign of the piezosignal is referredto the direction of
the polarization projection onto the samplesnormal, it can be
inferred that all investigated samples haveaverage non-zero
piezoresponse in the unpoled state. In our de-nition, it
corresponds to the polarization direction directed fromthe free
surface into sample (self-polarization effect). This corre-sponds
to the peak shift as observed in Fig.11c. The shifts to positiveand
negative values is related to different signs of
self-polarizationsand related to the preferred orientation of
surface domains. Theeffective piezoelectric coefcient, d33eff was
determined from thepeak half width (Fig. 11d). It is observed that
the composition with6.5 wt.% LN displays stronger piezocontrast as
compared to pureKNN implying an evidence of a pronounced
piezoelectric coef-cient. A broad response in Fig. 11(c) for 6.5
wt.% LN corresponds tothe trend shown in Fig. 11(d).
4. Conclusions
(1 x) K0.5Na0.5NbO3 xLiNbO3 (x 0e6.5 wt.%) lead
freepiezoelectric ceramic has been synthesized by conventional
solidstate reaction technique. The effect of different amounts of
Li andNb content on the phase structure and electric properties of
KNLNceramics were systematically investigated. XRD studies showed
thetransition from orthorhombic to tetragonal above x 5.5 wt.%
ofLN. Detailed studies of dielectric and electrical properties
indicatethat the material exhibits: (i) shift of Curie temperature
toward
Fig. 11. OPP PFM images of pure (a) and 6.5 wt.% LN (b), (c) and
(d) represent the LN conrepresenting piezoelectric activity.hysics
13 (2013) 430e440 439higher temperature side on increasing LN
concentration, (ii)decrease in dielectric constant and loss on
increase in LN content,(iii) follows Arrhenius relation in higher
temperature region, (iv)bulk or grain conduction up to 400 C, (v)
grain boundaryconduction up to 450 C, (vi) temperature dependent
relaxationphenomena. The activation energy values obtained for
different LNcontent suggest that the electrical conduction in KNLN
is mainlydue to the mobility of the ionized oxygen defects. We
observed thatthe compositionwith 6.5 wt.% LN displays stronger
piezocontrast ascompared to pure KNN implying an evidence of a
pronouncedpiezoelectric coefcient.
Acknowledgments
The authors are grateful to the Foundation for Science
andTechnology of Portugal (FCT) for nancial support within
theprojects PTDC/CTM-CER/115085/2009 and
PTDC/FIS/108025/2008.Thework is partly supported by the EU network
Enermat.aa. SeemaSharma wish to acknowledge Department of Science
and Tech-nology, Govt of India (SR/S2/CMP-39/2006), India for the
nancialsupport under a research project scheme.
References
[1] Y. Saito, H. Takao, T. Tani, T. Nonoyama, K. Takatori, T.
Homma, T. Nagaya,M. Nakamura, Nature 432 (2004) 84e87.
[2] E. Hollenstein, M. Davis, D. Damjanovic, N. .Setter, Appl.
Phys. Lett. 87 (2005)182905.
[3] H.L. Du, Z.M. Li, F.S. Tang, S.B. Qu, Z.B. Pei, W.C. Zhou,
Mater. Sci. Eng. B 131(2006) 83.
[4] P.C. Goh, K. Yao, Z. Chen, J. Am. Ceram. Soc. 92 (2009)
1322.[5] N. Hagh, A. Safari, J. Electroceram. 18 (2007) 339.[6]
R.E. Jaeger, L. Egerton, J. Am. Ceram. Soc. 45 (1962) 209.
centration dependence of the piezoresponse histograms of
out-of-plane (OPP) signal
-
[7] R.P. Wang, R.J. Xie, T. Sekiya, Y. Shimojo, Mater. Res.
Bull. 39 (2004) 1709.[8] J.F. Li, K. Wang, B.P. Zhang, L.M. Zhang,
J. Am. Ceram. Soc. 89 (2006) 706.[9] J.G. Hao, Z.J. Xu, R.Q. Chu,
Mater. Chem. Phys. 118 (2009) 229.[10] M. Matsubara, T. Yamaguchi,
W. Sakamoto, J. Am. Ceram. Soc. 88 (2005) 1190.[11] J.G. Hao, Z.J.
Xu, R.Q. Chu, J. Mater. Sci. 44 (2009) 6162.[12] J.G. Hao, Z.J. Xu,
R.Q. Chu, J. Electron. Mater. 39 (2010) 347.[13] Y. Guo, K.
Kakimoto, H. Ohsato, Jpn. J. Appl. Phys. 43 (2004) 6662e6666.[14]
Y. Guo, K. Kakimoto, H. Ohsato, Solid State Commun. 129 (2004)
279e284.[15] Y. Guo, K. Kakimoto, H. Ohsato, Appl. Phys. Lett. 85
(2004) 4121e4123.[16] Y. Guo, K. Kakimoto, H. Ohsato, Mater. Lett.
59 (2005) 241e244.[17] S.J. Zhang, R. Xia, H. Hao, H.X. Liu, T.R.
Shrout, Appl. Phys. Lett. 92 (2008)
152904.[18] S.H. Park, C.W. Ahn, S. Nahm, Jpn. J. Appl. Phys. 43
(2004) 1072e1074.[19] E. Hollenstein, D. Damjanovic, N. Setter, J.
Eur. Ceram. Soc. 27 (2007) 4093.
[20] R. Elfrink, J. Micromech. Microeng. 19 (2009) 094005.[21]
R.N.P. Choudhary, Dillip K. Pradhan, C.M. Tirado, G.E. Bonilla,
R.S. Katiyar,
Phys. Status Solidi (b) 244 (2007) 2254e2266.[22] A. Shukla,
R.N.P. Choudhary, Physica B Condens. Matter 406 (2011) 2492e
2500.[23] D.C. Sinclair, A.R. West, J. Appl. Phys. 66 (1989)
3850.[24] A. Shukla, R.N.P. Choudhary, A.K. Thakur, J. Phys. Chem.
Solids 70 (2009) 1401.[25] A. Shukla, R.N.P. Choudhary, Curr. Appl.
Phys. 11 (2011) 414.[26] J.R. Macdonald, Impedance Spectroscopy,
Emphasizing Solid Materials and
Systems, Wiley Interscience Publ., New York, 1987.[27] S.
Poykko, D.J. Chadi, Appl. Phys. Lett. 76 (2000) 499e501.[28] S.
Steinsvik, R. Bugge, J. Gjonnes, J. Tafto, T. Norby, J. Phys. Chem.
Solids 58
(1997) 969e976.[29] N. Balke, I. Bdikin, S.V. Kalinin, A.L.
Kholkin, J. Am. Ceram. Soc. 92 (2009) 1629.
R. Rai et al. / Current Applied Physics 13 (2013) 430e440440
Impedance spectroscopy and piezoresponse force microscopy
analysis of lead-free (1 x) K0.5Na0.5NbO3 xLiNbO3 ceramics1.
Introduction2. Experimental procedure3. Results and discussion4.
ConclusionsAcknowledgmentsReferences