Philosophical Magazine

Vol. 88, No. 26, 11 September 2008, 3129–3143

Dielectric spectroscopy characteristics of ferroelectric

Pb0.77K0.26Li0.2Ti0.25Nb1.8O6 ceramics
K.S. Raoa*, P.M. Krishnab, D.M. Prasadc,
T.S. Lathaa and C. Satyanarayanaa
a
Department of Physics, Centre for Piezoelectric Transducer Materials,
Andhra University, Visakhapatnam, India; bDepartment of Physics,
GVP College of Engineering, Visakhapatnam, India; cDepartment of Basic
Science & Humanities, GIT, GITAM University, Visakhapatnam, India
(Received 18 December 2007; final version received 1 September 2008)

Temperature and frequency dependence of dielectric constant and conductivity

properties of Pb0.77K0.26Li0.2Ti0.25Nb1.8O6 (PKLTN) ceramics are modelled
through the universal dielectric response (UDR). Partial substitution of Ti4þ
for Nb5þ was compensated by charge and the creation of oxygen vacancies
according to the Kroger-Vink notation. The electrons released by this reaction are
captured by Nb5þ and Ti4þ to generate Nb3þ and/or Ti3þ. The hopping
of electrons between Nb5þ–Nb3þ and Ti4þ–Ti3þ are believed to participate in
conductivity. Characterization of the dielectric constant has been performed
from room temperature to 590 C in the frequency range from 45 Hz to 5 MHz.
The measured dielectric constant obeyed Jonscher’s dielectric dispersion
relations: "l ¼ "1 þ sin(n(T )/2)(a(T )/"0)(!n(T)1) and "ll ¼ /i"0! þ cos(n(T )/2)
(a(T )/"0)(!n(T)1). Cole-Cole plots inclined at an angle (1  n(T ))/2
and followed the trend of universal material behavior "1 þ A(T )(!n(T)1).
The exponent n(T ) and coefficient A(T ) ¼ (a(T )S/L) exhibited a minimum and
maximum at Tc ¼ 425 C, respectively. The conductivity studies show the
contribution of the hopping of bound charge carriers to conduction in PKLTN.
Keywords: niobate ceramics; dielectric constant; electrical conductivity;
relaxation

1. Introduction
Dielectric spectroscopy (DS) is an experimental tool, applicable in a wide range of
materials and currently covers an extraordinary spectral range from 106 to 1012 Hz.
Dielectric studies can reveal a lot of information on a solid, such as phase transition,
defect, and transport properties [1]. A detailed analysis of the frequency and temperature
dependence of ac conductivity and permittivity is necessary to characterize the microscopic
mechanisms and associated relaxation phenomena of charge carrier transport. DS is
a non-invasive, very sensitive technique to investigate a variety of complex systems [2,3],
including biological systems [4,5].
Dielectric relaxation spectroscopy (DRS) is another powerful technique to acquire
data on the molecular dynamics of almost any kind of materials, including liquid crystals.

*Corresponding author. Email: konapala@sify.com

ISSN 1478–6435 print/ISSN 1478–6443 online

ß 2008 Taylor & Francis
DOI: 10.1080/14786430802464222
http://www.informaworld.com
3130 K.S. Rao et al.

The observation that many systems display similar frequency dependence conductivity
and permittivity, namely Jonscher’s ‘universal dielectric response’ (UDR), is now a topic
of intense experimental and theoretical interest. At low frequencies, bulk (real) ac
conductivity (0) is frequency-independent but, at higher frequencies, ac conductivity
increases, following a power-law behavior such that (!)  !n, where ! is angular
frequency, and n (51.0) is a temperature-dependent parameter. The universality is
considered with conventional Debye’s behavior of relaxation times and a stretched
exponential parameter. The power laws are observed in a range of single crystals,
polycrystalline and amorphous materials, including ceramics, polymers, composites, wet
cement, electronic and ionic conductors, as it is important that any model must be able to
account for the ubiquity of the UDR.
The frequency dependence of the measured dielectric constant and dielectric loss
reflects the intrinsic property of the bulk material, effects from electrodes and any internal
interfacial barriers. The results obtained using DS are basically unambiguous and provide
a true picture of electrical properties of the sample, which can then be related to the
intrinsic properties of the material and are independent of sample geometrical factors.
For this reason, DS was selected as the characterization technique in the present study.
Niobates with a tungsten–bronze structure have attracted considerable scientific and
commercial attention owing to their wide potential applications as ferroelectric,
piezoelectric, pyroelectric and nonlinear optical materials [6–10]. The tetragonal tungsten
bronze (TTB) unit cell can accommodate metal ions in five different sites, designated A1,
A2, C, B1 and B2. The structure basically consists of a complex array of distorted BO6
octahedra sharing corners in such a way that there are three different types of interstices
between them, which are available for cation occupation. The unit cell may be described as
(A1)4 (A2)2 C4 (B1)2 (B2)8O30, where the A1 and A2 sites may be substituted by Na, K, Cs,
Ca, Sr, Ba, Pb, Bi, rare earths, etc. the much smaller C site by Li, Be, Mg and Al, and the
B1 and B2 sites by Mg, Fe, Ti, Zr, Sn, Nb, Ta, W, etc. However, they must be made in such
a way that the valency charge of the ions occupying the A, C and B sites balance those of
the O2 ions, maintaining overall electrical neutrality.
Pb(1x)K2xNb2O6 (PKN) (for x ¼ 0.2) has a large electromechanical coupling factor of
bulk waves and a small temperature dependence of small fundamental frequencies, which
was first discovered by Yamada et al. [12,13]. This material is an important material for
surface acoustic wave (SAW) devices due to its non-linear optical properties [12,14–20].
However, the utilization of PKN single crystals is largely limited due to PbO loss resulting
from the high-temperature growth and to the formation of cracks when cooling through
the Curie temperature [21]. PKN belongs to space group C2v14–Cm2m and its unit cell
contains four formula units.
Elaboration, control of stoichiometry and dielectric spectroscopy analysis of PKN
(x ¼ 0.2) has been reported by Zhigao et al. [22]. Impedance spectroscopic analysis and
piezoelectric properties on Pb0.8K0.4Nb2O6 has been reported by Rao et al. [23]. Influence
of Sm3þ substitution on impedance, dielectric and electromechanical properties in
Pb(1x)K2xNb2O6 (x ¼ 0.23) has also been reported [24]. Furthermore, the effect of
simultaneous substitution of Liþ and Ti4þ in Pb2KNb5O15 ceramics on structure,
dielectric, modulus, impedance and conductivity properties have been characterized [25].
Structure, dielectric and piezoelectric strain coefficient of Pb(12x)Kx M3þ
x Nb2O6 ceramics,
where, M ¼ La or Bi and 0.15 x 50.4 has been reported by Neurgaonkar et al. [26].
The present material, Pb0.77K0.26Li0.2Ti0.25Nb1.8O6 (PKLTN) has been derived from PKN
 Philosophical Magazine 3131

(for x ¼ 0.23) belongs to the formula-type A6C2B10O30, where all A- and C-sites are filled –
known as a completely filled structure.
To the best of our knowledge, the dielectric spectroscopy characterization of
a substitutional solid solution of ferroelectric Pb0.77K0.46Nb5O15 ceramics, in which Kþ
is modified with Liþ and Nb5þ with tetravalent Ti4þ, has not been reported so far. The
present study illustrates the preparative conditions, structural characterization, dielectric
0
dispersion, charge carrier-induced dielectric constant ("carr ) and coupling between the
carriers and phonons in completely filled (i.e. Pb0.77K0.26Li0.2Ti0.25Nb1.8O6 (PKLTN)),
ceramics. The strength of polarizability and degree of interaction between the mobile ions
are also analyzed by pre-factor A(T), exponent n(T) of Jonscher’s ‘universal law’ of
conductivity.

2. Experimental procedure
2.1. Material
Pb0.77K0.26Li0.2Ti0.25Nb1.8O6 (PKLTN) was synthesized by a solid-state reaction as
follows:

0:77PbO þ 0:13ðK2 CO3 Þ þ 0:1ðLi2 CO3 Þ þ 0:25ðTiO2 Þ þ 0:9ðNb2 O5 Þ

! Pb0:77 K0:26 Li0:2 Ti0:25 Nb1:8 O6 þ 0:23CO2 :
Initially, a stoichiometric mixture of AR grade powders of oxides, carbonates and titanates
of PbO, Li2CO3, Nb2O5, TiO2 and K2CO3 were weighed and ground mechanically in an
agate mortar in a methanol medium. The K2CO3 was heated at 300 C before weighing to
reduce its hygroscopic nature and the ground mixture subsequently taken for thermal
treatment. The ground powder was calcined three times at 900 C for 4 h to achieve a single
phase with grinding at each stage. A weight check after calcination confirmed the
decomposed value of the carbonates.
Room temperature X-ray diffractogram (XRD) (Figure 1) were taken with a Philips
X-ray diffractometer (Cu K radiation  ¼ 1.5406 Å) in 10  2  80 range at 2 /min
scanning rate. The 100% intensity peak of PKLTN was observed at 2 ¼ 32.80 . From
inter-planar spacing (dobs) values, the Miller indices (hkl) and unit cell parameters were
estimated using a standard computer program package (POWD, Interpretation and
Indexing Program from E. Wu, School of Physical Sciences, Flinders University,
Adelaide, South Australia). The preliminary structural analysis of PKLTN indicated
a unit cell with orthorhombic structure. This crystal system was confirmed by POWD,
as the
P difference
P between the dobs and calculated inter-planar spacing (dcal) is minimal;
i.e.
d ¼ (dobs  dcal) ¼ min.
The lattice parameters computed in PKLTN are a ¼ 18.2450 , b ¼ 18.2970 and c ¼ 3.8940 .
In PKN (x ¼ 0.23), the lattice reported parameters [24] are: a ¼ 17.7870 , b ¼ 18.0610 and
c ¼ 3.8600 . The orthorhombic distortion (b/a) ¼ 1.008 in PKLTN and 1.015 in PKN
(x ¼ 0.23) reveals that the simultaneous substitution of Liþ and Ti4þ does not distort the
orthorhombic structure. However, an increase in unit cell volume V ¼ 1300.24 Å3 has been
observed compared to PKN (x ¼ 0.23). This may be due to partial substitution of the
smaller ionic radii element Ti4þ (0.64 Å) for Nbþ5 ions (1.000 ) in the B-site and also Liþ
ions entering the C sites, which are empty in PKN. The theoretical density and porosity of
3132 K.S. Rao et al.

Counts
100 g.CAF

50

0
20 30 40 50 60 70
Position [*2Theta]

Figure 1. X-ray diffractogram of PKLTN (Colour online).

the PKLTN were estimated using XRD analysis and found to be 5.70 g/ml and 0.013,
respectively.
A few drops of polyvinyl alcohol (PVA) were added to the fine homogeneous calcined
powder as a binder to strengthen and improve granular flow, and reduce brittleness of the
sample after sintering. This powder was compacted into discs of 11 mm in diameter and
3 mm thick with a hydraulic press under a 580-MPa uniaxial pressure. The green pellets
were placed on a platinum foil and some calcined powder sprinkled on both surfaces of
the sample to compensate for the loss of lead and sealed in an inverted alumina crucible.
The pellets are initially fired at 600 C for 1 h in a microprocessor-programmable furnace
to eliminate the binder and, finally, sintering at 1070 C for 60 min. The experimental
density of the PKLTN ceramic was calculated using Archimedes’ principle and found to be
5.63 g/ml, indicating that the sintered samples are dense and crack-free.
As discussed previously, the total number of cations available in octahedral positions
of the B-site is 10. However, in the present composition, partial substitution of Ti4þ for
Nb5þ increased the B sites by 0.25. These excess B-cations are associated with charge
compensation of the compound. Furthermore, with the replacement of Nb5þ ions with
Ti4þ ions at lattice site B, there is a strong possibility of creating oxygen vacancies, which
maintain localized charge neutrality [25,27]. Since the ceramic samples were sintered at
high temperature 41000 C, a slight amount of oxygen loss can occurs according to the
Kroger-Vink notation:
V0 ! V00 þ e0
V00 ! V000 þ e0 ,

and these electrons may bond to Ti4þ in the form:

Ti4þ þ e0 ! Ti3þ ;
where V0 is the oxygen vacancy, V00 and V000 are single and doubly ionized oxygen
vacancies, respectively, and e0 is the electron released or captured. Since Nb can exist in the
 Philosophical Magazine 3133

þ3 and þ5 states and Ti can exist in the þ3 and þ4 states, electrons may be released in the
above reaction and captured by Nb5þ and Ti4þ to generate Nb3þ and or Ti3þ. Thus,
conduction may occur by hopping of electrons between Nb5þ and Nb3þ and Ti4þ and
Ti3þ. In the low-frequency region, conduction arises predominantly by short-range
translational hopping of electrons between Nb5þ and Nb3þ and Ti4þ and Ti3þ. Whereas at
high frequencies, conductivity is due to localized orientational hopping (back and forth
hopping) between the Nb5þ and Nb3þ and Ti4þ and Ti3þ states.

2.2. Low frequency spectroscopy measurements

The well-sintered cylindrical samples were carefully polished to eliminate possible
inhomogeneity due to exchanges with the sintering atmosphere. A silver paste was then
used to form an electrode on the sample surfaces. Dielectric constant, conductivity as
a function of temperature (from room temperature (RT) to 600 C) and frequency (45 Hz–
5 MHz), recorded simultaneously at 67 frequency-points using a computer-interfaced
Hioki 3532-50 LCR Hi-Tester (Ueda, Japan), were measured. A constant heating rate of
5 C/min was maintained, with an offset temperature 0.2 C and time period of 0.5 min, for
the property measurements.
The capacitance Cp and loss tangent tan  were used in calculating the real dielectric
constant ("l) and imaginary dielectric constant or dielectric loss ("II) according to the
following equations:

"1 ¼ Cpd=ð"0 AÞ
"11 ¼ "1 tan ;
where d and A are the thickness and electrode area of the ceramic disc, respectively, and
"0 is the free space permittivity (8.854  1012 F/m). Studies on the microstructure,
impedance and electrical modulus characterization of the present material are in progress.

3. Results and discussion

3.1. Influence of temperature and frequency on dielectric constant
Temperature dependence of the real dielectric constant ("l) at various frequencies
(Figure 1) allowed us to determine the Curie temperature (Tc) at 425 C. The increase in the
"l value with decreasing frequency is a characteristic feature of polar dielectric materials.
The following are the observations made from Figure 2.
(1) The value of "l increases gradually with increasing temperature due to interfacial
polarization becoming more dominant than dipolar polarization. It passes through
Tc and then decreases due to phase transition from a ferroelectric to a paraelectric
phase.
(2) Dispersion of "l in the paraelectric phase is mainly attributed to ionic conductivity.
At high temperatures and low frequencies, another maximum appeared. This
phenomenon is related to space-charge effects and not another phase transition.
The frequency-independence of Tc in PKLTN shows that the material is classical
ferroelectric, but not of the relaxor type. This is in good agreement with the fact that
3134 K.S. Rao et al.

165

150 Pb0.77K0.26Li0.2Ti0.25Nb1.8O6

135 500Hz
1KHz
120 10KKHz
εl 20KHz
105

90

75

60

100 200 300 400 500 600

Temperature °C

Figure 2. Temperature dependence of the real dielectric constant at different frequencies.

Table 1. Dielectric data.

Composition "1RT (1 kHz) Tc ( C) "1TC (1 kHz)  RT (S/cm  109) Curie constant, K (105)

PKLTN 64 425 123 1.77 1.25

each crystallographic site in the structure is occupied by only one type of ion [28].
The broadness in the dielectric constant versus temperature response plot is attributed
to compositional fluctuations caused by the simultaneous substitution of Liþ, Ti4þ and
analogues by other compounds with a pervoskite structure [29]. The sharp decrease in
dielectric constant with increasing frequency can also be explained in terms of interfacial
polarization.
The Curie-Weiss law has been fitted in the paraelectric region and the value of Curie
constant (K ¼ 1.25  105) estimated, confirming that PKLTN belongs to ferroelectrics of
oxygen-octahedral type [30–32]. The values for room temperature dielectric constant ("1RT ),
Tc ( C), dielectric constant at transition temperature ("lTc ), room temperature conductivity
( RT) and K are given in Table 1.
The Arrhenius plot of the influence of temperature on the imaginary dielectric constant
"ll is shown in Figure 3. An anomaly was observed at a particular temperature, which
almost coincides with the Tc of the material. In the studied frequency region, the dispersion
of "ll at higher temperatures is stronger than that of "l, revealing that it is influenced by
dc conductivity. It is directly related to ac conductivity by the equation: "ll ¼  ac/"0!.
The log  ac  1/T curves should have the same shape as log "ll1/T. It is clear from
Figure 3 that the high-temperature behavior of "ll may be described by the Arrhenius law,
suggesting that the conduction mechanism is thermally activated. The typical complex
dielectric constant (Cole-Cole) plots at three different temperatures (paraelectric region)
are shown in Figure 4. In describing dielectric relaxation, a generalized Debye model based
 Philosophical Magazine 3135

3.5
Pb0.77K0.26Li0.2Ti0.25Nb1.8O6
3.0

2.5 500Hz
1kHz
2.0 10kHz

Log εll
20kHz
1.5

1.0

0.5

0.0 TC

–0.5
1.0 1.5 2.0 2.5 3.0 3.5

1000/T (K–1)
Figure 3. Arrhenius plot of imaginary part of the dielectric constant at different frequencies.

2000
Pb0.77K0.26Li0.2Ti0.25Nb1.8O6

1500 440°C

500°C

580°C
εll

1000

500

0
0 200 400 600 800 1000
εl

Figure 4. Cole-Cole plots of the complex dielectric constant at different temperatures.

on Cole-Cole plots is used, with frequency-dependent complex permittivity "* represented

by the equation [11]
   
" ð!Þ ¼"1 þ "s1  "s2 =ð1 þ !1 Þ11 þ ð"s2  "1 =ð1 þ !2 Þ12 ;

where ! is the angular frequency,  1 and  2 are the relaxation frequencies, 1 and 2 are
the characteristics of relaxation time without unit (1    0). This represents the width of
the distribution of the relaxation time.
The value of  has been estimated from the Cole-Cole plots by measuring the angle
between the real dielectric constant and radius of the circle. If the centre of the semicircle
3136 K.S. Rao et al.

Pb0.77K0.26Li0.2Ti0.25Nb1.8O6 300°C
(a) 3.0 Pb0.77K0.26Li0.2Ti0.25Nb1.8O6 300°C (b)
340°C
340°C 4
2.8 380°C
380°C
400°C
400°C
2.6 425°C
3 425°C
480°C
480°C
2.4
εl

εll
500°C
500°C 2 540°C
2.2 540°C
580°C
580°C
2.0 1
1.8
1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 0
2 3 4 5 6
log f Hz
log f Hz
Figure 5. Frequency dependence of (a) "0 and (b) "00 at log–log scale on frequency.

lies on the "l axis (i.e.  is zero) it is Debye type. Otherwise, the centre is below "l axis and
 6¼ 0, it represents a non-Debye type. 1 and 2 are empirical ‘broadness’ parameters
showing the degree of departure from the Debye model. "s1 and "s2 are the limiting values
of permittivity. "1 is the high frequency limiting the dielectric constant.
From the Cole-Cole plots (Figure 4), the universal material behavior, represented by
"1 þ A(T )(!n(T)1), is obeyed by forming a straight line inclined at an angle (1  n(T ))/2
and an intercept at "1 with the horizontal axis [33]. Figures 5a and b show the log–log
plots of "l and "ll as a function of frequency at different temperatures in PKLTN.
The observed variation in "l with frequency is ascribed to the formation of a space-charge
region at the electrode/sample interface, which is known as !(n1) variation or non-Debye
type of behavior, where the space-charge regions with respect to frequency is explained in
terms of ion diffusion [33]. The low frequency region is attributed to the contribution of
charge accumulation at the interface. At high frequencies, due to high periodic reversal of
the field at the interface, the contribution of charge carriers (ions) towards the dielectric
constant decreases with increasing frequency. Hence, "l decreases with increasing
frequency. A typical variation in "ll is shown in Figure 5b. In the present case, conduction
losses predominate at lower frequencies and, hence at all temperatures, "ll shows 1/!
dependence on frequency. "ll is also found to increase with increasing temperature and
dielectric peak loss has not been observed. In ionically conducting materials at low
frequencies, there are unavoidable electrode polarization effects and the dielectric constant
is high. The low frequency slope of the curve of log "ll versus log f is close to 1, indicating
the dominance of dc conduction in the frequency region.
Figure 6 shows the variation in the real part of ac conductivity as a function of
frequency at different temperatures in the ferro- and paraelectric region. The frequency
dependence of conductivity in different materials is described by the power relation
proposed by Jonscher [33]:

ð!Þ ¼ dc þ A!n ;

where (!) is total conductivity,  dc is frequency-independent conductivity and

the coefficient A and the frequency exponent n are thermally activated, material-
dependent quantities. The term A!n contains ac dependence and characterizes all
 Philosophical Magazine 3137

Pb0.77K0.26Li0.2Ti0.25Nb1.8O6
1E-5 σ (ω) αωn

Conductivity (S/cm)
Frequency independent conductivity,
1E-6 σ (0)

380°C
1E-7
420°C
480°C
520°C
1E-8
540°C
580°C

1E-9
102 103 104 105 106
f (Hz)

Figure 6. Variation of  ac versus f Hz at different temperatures.

dispersion phenomena. A plateau exists in Figure 6, like region ( dc), at the low-frequency
region and, consequently, conductivity increases with increasing frequency, varying
approximately as a power of frequency !n (where n is a function of temperature as well as
frequency) at all temperatures. The exponent n has been found to behave [34–36] in
a variety of forms, like a constant, decreasing with temperature, increasing with
temperature, etc., but always within 0 5 n 5 1. Funke [37] explained that the value of n
might have a physical meaning (i.e. n  1 would mean that the hopping motion involved is
a translational motion with a sudden hopping). On the other hand, value of n 4 1 would
mean that the motion involved is a localized hopping of the species with a small hopping
without leaving the neighborhood [38].
From Figure 6, the switch from the frequency-independent  dc to frequency-dependent
(!) regions shows the onset of the conductivity relaxation phenomenon and the
translation from long-range hopping to short-range ion motion [39]. The dispersion in
conductivity at low frequencies may be due to electrode polarization. The frequency
dependence of ac conductivity may arise due to free as well as bound carriers. In the
present case, ac conductivity increases with increasing frequency and, hence, must be
related to bound carriers trapped in the sample. In PKLTN, the Liþ ion mainly exhibits
cationic conductivity due to the mobility of Liþ in the partially filled C-site.

4. Dielectric dispersion relations and interpretation of the experimental results

Both "l and "ll shows strong dispersions in the low frequency region (Figure 5). Such
a strong dispersions in both components of the complex dielectric constant is a common
feature in ferroelectrics associated with non-negligible ionic conductivity and is referred
to as low-frequency dielectric dispersion (LFDD) [40,41]. Detailed studies of this
phenomenon were carried by Jonscher et al. [42]. According to Jonscher’s power law,
the complex dielectric constant as a function of the frequency ! can be expressed as [40]
" ¼ "1  i"11 ¼ "1 þ ð=i"0 !Þ þ ððaT Þ="0 Þði!nðT Þ1 Þ;
3138 K.S. Rao et al.

where "1 is the ‘high frequency’ value of the dielectric constant, n(T ) is the temperature-
dependent exponent and a(T ) determines the strength of the polarizability arising from the
universal mechanism. The real and imaginary parts of the complex dielectric constant are
given by the following relations [27,40]:
"l ¼ "1 þ sinðnðT Þ=2ÞððaT Þ="0 Þð!nðT Þ1 Þ ð1Þ

"ll ¼ =i"0 ! þ cosðnðT Þ=2ÞððaT Þ="0 Þð!nðT Þ1 Þ; ð2Þ

where the first term in Equation (1) determines the lattice response and corresponding
dc conduction part, respectively. The second term in both equations reflects the charge-
carrier contribution to the dielectric constant. The temperature and frequency dependencies
of the dielectric constant "0 (Figure 5a) could be explained by Equation (1). The charge-
carrier term (sin(n(T )/2)(a(T )/"0)(!n(T)1)) dominates at low frequency and "1 is
negligible. Therefore, a measured value of n in Equation (1) yields a straight line with
a slope equal to (n  1) in the double logarithmic plot of "l and frequency. At high
frequencies, the charge carriers fail to respond to the external field; therefore, the measured
dielectric constant is due to a contribution from lattice polarization. This accounts for the
linear decrease in the low frequency region and a plateau region at high frequencies
(Figure 5). As A(T ) (¼a(T )s/L) increases with increase in temperature [43], the charge-
carrier term become more prominent at high temperatures, resulting in the low frequency
dielectric dispersion. The exponent n(T ) and coefficient A(T ) have been determined from
plots shown in Figure 5a.
The interaction between the charge carriers participating in the polarization process is
characterized by the parameter n and value of n ¼ 1 (Debye case), which is attainable [40]
at very low temperatures. Therefore, as the temperature increases, the interaction
increases, leading to a decrease in n [44]. The value of n calculated from the high frequency
region decreases as the temperature increases and attains a minimum at Tc and it
subsequently increases with a further increase in temperature. Careful analysis of
Figure 5b indicates the existence of slopes corresponding to 1 in the low frequency region
and  (1  n) in the high frequency region. As the dc conductivity term increases with
increasing temperature, the second term in Equation (2) is totally overshadowed by the
first term. Thus, at low frequencies and high temperatures, the dc conductivity term is
dominant and yields a slope of 1, which is consistent with the data shown in Figure 5b.

4.1. Qualitative interpretation of the experimental results

The theoretical values of "l has been evaluated using Equation (1) (by substituting n, a
and "1) and are fitted with the experimental data. Figure 7 represents such a fitted curve
for log "l as a function of frequency at 380 and 540 C. A good agreement between the
experimental and theoretical values of "l has been noticed in the high temperature
(4500 C) region than at low temperatures. This may be due to the contribution of
long-range electrostatic forces resulting from the spontaneous polarization of the sample,
mobility of Liþ and Ti4þ ions and tendency to cancel the short-range forces which arises
from attraction and repulsion between the neighboring ions [25].
The "1 value was chosen as the dielectric constant at 1 MHz, as the dispersion is
negligible around this frequency range. The temperature dependence of n(T ) and A(T )
 Philosophical Magazine 3139

103 103
540°C
380°C
Experimental
Experimental
Theory
Theory

Log εl
Log εl

102 102

Pb0.77K0.26Li0.2Ti0.25Nb1.8O6 Pb0.77K0.26Li0.2Ti0.25Nb1.8O6

101 101
103 104 105 106 102 103 104 105 106
Log f (Hz)
Log f (Hz)
Figure 7. Fitting results for dielectric constant at 380 and 540 C.

90

Pb0.77K0.26Li0.2Ti0.25Nb1.8O6
85

80

75
εα

70

65

60

0 100 200 300 400 500 600

Temperature °C

Figure 8. Temperature dependence of "1.

have been determined from the curve fitting (Figure 5a). Variations in "1, n(T ) and A(T )
are shown in Figures 8–10, respectively.
The value of "1 indicates the lattice contribution and exhibits a peak at TC.
An interesting feature of Figure 8 is two linear regions, corresponding to the ferro- and
paraelectric states. A linear relation of the exponent with temperature has also been found
in some ion conducting systems [45–47]. Near TC, a minimum n is observed. This
trend indicates the soft mode coupling mechanism and longer relaxation of charge carriers
[48–51]. Similar results have been obtained in PKN ceramics [25,52]. A maximum value of
A is observed near the TC. The critical behavior of the pre-factor A, which determines the
strength of polarizability (or the non-ideal conductivity), arises from diffusive motion of
carriers. The results show the strong temperature dependence of these parameters.
3140 K.S. Rao et al.

0.98

n 0.96

0.94

Tc
0.92
Pb0.77K0.26Li0.2Ti0.25Nb1.8O6

300 350 400 450 500 550 600

Temperature °C

Figure 9. Variation in critical exponent n(T ) with temperature showing minimum at TC.

–9.55

–9.60 TC

–9.65

–9.70
log A

–9.75

–9.80

–9.85
Pb0.77K0.26Li0.2Ti0.25Nb1.8O6
–9.90
300 350 400 450 500 550 600
Temperature °C

Figure 10. Temperature dependence of the pre-factor A(T ) showing a peak at TC.

l
The charge carrier-induced dielectric constant "carr , calculated according to the second
term of Equation (1) for various frequencies, is shown in Figure 11.
Just as the intrinsic polarization demonstrates a prominent peak at the ferroelectric–
paraelectric phase transition (Figure 8), which is generally considered to be due to the
softening of one of the optical phonons, the observation of a similar peak for charge
carrier-induced polarization may imply the same origin. Therefore, a coupling between the
carriers and phonons seems to be involved [40]. Hence, the strong frequency dependence of
 Philosophical Magazine 3141

75

Pb0.77K0.26Li0.2Ti0.25Nb1.8O6
60

500Hz
45 1KHz
ε carrier 10KHz
30 20KHz
l

15

100 200 300 400 500 600

Temperature °C

Figure 11. Variation in carrier-induced dielectric constant "carr with temperature at various
frequencies.

the charge carrier-induced term is evident from Figure 11. Furthermore, according to
lattice dynamics theory, these carriers become mobile near TC, implying that the energy
loss is high and the stored energy is small.

5. Conclusions
1. Dielectric spectroscopy (DS) characterization has been carried out on dense
Pb0.77K0.26Li0.2Ti0.25Nb1.8O6 (PKLTN) ferroelectric ceramic sample (TC ¼ 425 C)
from 300 to 500 C in the 45-Hz to 5-MHz frequency range.
2. Dispersion of "l versus temperature above the TC reveals that ionic conductivity is
strongly related to space-charge effects. The "ll dispersion is stronger than the real
part, implying the dc conductivity influence in PKLTN.
3. The ac conductivity increases with increasing frequency, revealing the bound carriers
trapped in the sample. The dispersion in conductivity at low frequencies shows the
electrode polarization effect.
4. In Cole-Cole plots (non-Debye type), the universal material behavior was obeyed by
making a straight line inclined at an angle of (1  n(T ))/2 and an intercept at "1.
Using Jonscher’s dielectric dispersion relations, the lattice contribution ("1) and the
charge-carrier contribution ("carr) to the dielectric constant are evaluated. Good
agreement between the experimental and theoretical values of "0 has been found with
little deviation at low frequency or in the temperature range.
5. The true dielectric constant of PKLTN at various frequencies has been examined by
separating the influence of carrier-induced dielectric constant. The temperature-
dependent parameters n(T ) and A(T ) showed an anomaly at TC, revealing the soft
mode coupling mechanism, longer relaxation and strength of polarizability arising
3142 K.S. Rao et al.

from diffusive motion of carriers. The treatment adopted is in line to rationalize the
dielectric behavior of PKLTN ceramics.

Acknowledgements
KSR and PMK thank the Defence Research and Development Organization (DRDO) Government
of India for sanctioning the research project and providing fellowships.

References

[1] C.R. Bowen and D.P. Almond, Mater. Sci. Tech. 22 (2006) p.719.
[2] D.P. Almond and C.R. Bowen, Phys. Rev. Lett. 92 (2004) p.157601.
[3] M.A.L. Nobre and S. Lanfredi, J. Phys. Chem. Solid. 64 (2003) p.2457.
[4] J.E. Yardley, D.B. Kell, J. Barrett et al., Biotech. Genet. Eng. Rev. 17 (2000) p.3.
[5] K.C. Schuster, Adv. Biochem. Eng. Biotech. 66 (2000) p.185.
[6] R.R. Neurgankar, W.F. Hall, J.R. Oliver et al., Ferroelectrics 87 (1988) p.167.
[7] T. Yano, T. Ohta and A. Watanabe, Jpn. J. Appl. Phys. 9 (1970) p.1008.
[8] N. Wakiya, J.R. Wang, A. Saiki et al., J. Eur. Ceram. Soc. 19 (1990) p.1105.
[9] B. Tribotte, J.M. Haussonne and G. Desgardlin, J. Eur. Ceram. Soc 19 (1990) p.1105.
[10] G.L. Roberts, R.J. Cava, W.F. Peck et al., J. Mater. Res. 12 (1997) p.526.
[11] E. Senturk, Cryst. Res. Tech. 39 (2004) p.157.
[12] T.Y. Yamada, J. Appl. Phys. 46 (1975) p.2894.
[13] R.M. O’Connell, J. Appl. Phys. 46 (1980) p.530.
[14] A. Giess, B.A. Scott, G. Burns et al., J. Am. Ceram. Soc. 52 (1969) p.276.
[15] J. Ravez and B. Elouadi, Mater. Res. Bull. 19 (1975) p.1249.
[16] J. Thoret and J. Ravaz, Rev. Chim. Miner. T 24 (1987) p.288.
[17] A. Zegzout and M. Elaatmani, Sil. Ind. 62 (1997) p.149.
[18] T. Yamada, Appl. Phys. Lett. 23 (1973) p.213.
[19] H. Yamauchi, Appl. Phys. Lett. 32 (1978) p.599.
[20] J. Nakano and T. Yamada, J. Appl. Phys. 46 (1975) p.2361.
[21] R.R. Neurgankar, W.F. Hall, W.K. Cory et al., Mater. Res. Bull. 23 (1988) p.1459.
[22] L. Zhigao, J.P. Bonnet, J. Ravez et al., J. Eur. Ceram. Soc. 9 (1992) p.38.
[23] K.S. Rao, P.M. Krishna, T.S. Latha et al., Mater. Sci. Eng. B 131 (2006) p.127.
[24] K.S. Rao, P.M. Krishna, D.M. Prasadand, et al., J. Alloys Comp. (2008), in press.
[25] K.S. Rao, P.M. Krishna and D.M. Prasadand, J. Phys. Status Solidi (b) 244 (2007) p.2267.
[26] R.R. Neurgaonkar, J.R. Oliver, W.K. Cory et al., Mater. Res. Bull. 18 (1983) p.735.
[27] B.H. Venkataraman and K.B.R. Varma, J. Phys. Chem. Solid. 66 (2005) p.1640.
[28] T. Egami, S. Teslic, W. Dhowski et al., Ferroelectrics 99 (1997) p.103.
[29] K.S. Rao, A.S.V. Subramanyam and S.M.M. Rao, Ferroelectrics 196 (1997) p.109.
[30] J. Nakano and T. Yamada, J. Appl. Phys. 46 (1975) p.2361.
[31] T.R. Shrout, L.E. Cross and D.A. Hukin, Ferroelectr. Lett. 44 (1983) p.325.
[32] T.R. Shrout, H. Chenand and L.E. Cross, Ferroelectr. Lett. 74 (1987) p.317.
[33] A.K. Jonscher, Dielectric Relaxation in Solids, Chelsea Press, London, 1983.
[34] S.R. Elliot, Adv. Phys. 36 (1987) p.135.
[35] J.S. Kim and T.K. Song, J. Phys. Soc. Jpn. 70 (2001) p.3419.
[36] S. Upadhyaya, S.A. Kumar, D. Kumar et al., J. Appl. Phys. 84 (1998) p.828.
[37] K. Funke, Prog. Solid State Chem. 22 (1993) p.111.
[38] S. Sen and R.N.P. Choudhary, Mater. Chem. Phys. 87 (2004) p.256.
[39] R. Mizaras, M. Takashige, J. Banys et al., J. Phys. Soc. Jpn. 66 (1997) p.2881.
 Philosophical Magazine 3143

[40] Z. Lu, J.P. Bonnet, J. Ravez et al., Solid State Ionics 57 (1992) p.235.
[41] T.A. Nealon, Ferroelectrics 76 (1987) p.377.
[42] A.K. Jonscher, R.M. Hill and C. Pickup, J. Mater. Sci. 20 (1985) p.4431.
[43] Z. Lu, J.P. Bonnet, J. Ravez et al., Eur. J. Solid State Inorg. Chem. 28 (1991) p.363.
[44] D.M. Ginsberg (Editor), Physical Properties of High Temperature Super Conductors, World
Scientific, Singapore, 1989.
[45] F.E.G. Henn, J.C. Giuntini, J.V. Zanchetta et al., Solid State Ionics 42 (1990) p.29.
[46] J.C. Giuntini, B. Deroide, P. Belougne et al., Solid State Comm. 62 (1987) p.739.
[47] Y. Bensimon, J.C. Giuntini, P. Belougne et al., Solid State Comm. 68 (1988) p.189.
[48] Y. Wu, M.J. Forbess, S. Seraji et al., J. Appl. Phys. 89 (2001) p.5647.
[49] L.A. Dissado and R.H. Hill, Nature 279 (1979) p.685.
[50] L.A. Dissado and R.H. Hill, Phil. Mag. B 41 (1980) p.625.
[51] T. Mitsui, I. Tatsuzaki and E. Nakamura, An Introduction to Physics of Ferroelectrics, Science
Press, Beijing, 1983.
[52] M. Dong, J.M. Reau and J. Ravez, Solid State Ionics 91 (1996) p.183.