Available online at [Link].

com

Solid State Ionics 179 (2008) 689 – 696

[Link]/locate/ssi

Solid polymer electrolytes based on polyethylene oxide and lithium

trifluoro- methane sulfonate (PEO–LiCF3SO3): Ionic conductivity
and dielectric relaxation
N.K. Karan a , D.K. Pradhan a , R. Thomas a , B. Natesan b , R.S. Katiyar a,⁎
a
Department of Physics and Institute for Functional Nanomaterials, University of Puerto Rico, PO Box 23343, San Juan, PR-00931, USA
b
Department of Physics, National Institute of Technology, Trichy 620 015, India
Received 21 August 2007; received in revised form 19 March 2008; accepted 29 April 2008

Abstract

Polymer electrolytes consisting of polyethylene oxide (PEO) and LiCF3SO3 were synthesized by solution casting method as a function of
EO/Li ratio. An increase in the glass transition temperature of the polymer electrolytes with increasing Li salt content suggested the coordination of
the Li ions to the oxygen atoms of polymer backbone. Dielectric spectroscopic studies were performed to understand the ion transport process
in polymer electrolytes. The dc conductivity showed a maximum for EO/Li ≈ 24. The ac conductivity analysis revealed the existence of nearly
constant loss (NCL) contribution at lower temperatures. The dielectric loss spectra showed the presence of one relaxation for all compositions, which
is associated with the motions of the Li ion coordinated polymer segments. The relaxation has been characterized by the empirical Havriliak–Negami
(H–N) equation. The temperature dependence of the relaxation times and the conductivity followed the Vogel–Tamman–Fulcher (VTF) equation
yielding qualitatively similar pseudoactivation energies, which suggested strong coupling between the ionic conductivity and the segmental relaxation
in the polymer electrolytes.
© 2008 Elsevier B.V. All rights reserved.

Keywords: Polymer electrolytes; Ionic conductivity; Dielectric relaxations; VTF behavior

1. Introduction where pure crystallinity has been reported [2]. Driven by their
potential applications in devices, such as in rechargeable Li ion
Polymer electrolytes represent a fascinating class of solid-state batteries, electrochromic displays, etc. there has been much
coordination compounds, which support ionic conductivity in a interest in the development and understanding of the mechan-
flexible, yet solid membrane. They consist of salts (e.g. LiClO4) ism of the ion conduction in solid polymer electrolytes [3–5].
dissolved in solid polymers (e.g. polyethylene oxide, PEO). The It is currently admitted that the ionic conductivity in polymer
polymer must contain a Lewis base (usually an ether oxygen), electrolytes takes place in the amorphous phase, where the ion
which serves to coordinate the cations, thus promoting dissolution conduction is mediated by local motion of the polymer chain
of the salt. Among all polymers, PEO has been studied most segments above the glass transition temperature, Tg. For example,
extensively so far because of its efficiency in coordinating metal nuclear magnetic resonance (NMR) studies of ionic motion
ions, due to the optimal distance and orientation of the ether within the PEO: LiCF3SO3 systems, at compositions where the
oxygen atoms in polymer chains [1]. In general, polymer crystalline 3:1 complex and the amorphous phase coexisted above
electrolytes based on PEO are semicrystalline in nature except Tg, demonstrated that ion transport occurs in the amorphous
in a few particular salt concentrations (depending upon the salt) phase [6]. However, in recent years it has been shown that ion
conduction might occur in the crystalline phase of polymer
⁎ Corresponding author. Tel.: +1 787 751 4210; fax: +1 787 764 2571. electrolytes as well [2,7].
E-mail addresses: rkatiyar@[Link], rkatiyar@[Link] Solid polymer electrolytes differ from conventional glassy ion
(R.S. Katiyar). conductors in that the glass transition temperature (Tg) of the
0167-2738/$ - see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/[Link].2008.04.034
690 N.K. Karan et al. / Solid State Ionics 179 (2008) 689–696

former is, in general, below room temperature whereas that of the CuKα radiation from 5–60° in the θ–2θ scan. Differential
conventional glasses is way above room temperature. Thus, scanning calorimetry (DSC) measurements were done using
polymer electrolytes are, in general, in rubbery state whereas the Shimadzu DSC-50 with a low temperature measuring head and
conventional glass conductors are in “glassy” state, at room liquid nitrogen as coolant. Samples were crimped in aluminum
temperature [8]. Hopping of the mobile ions between different pans inside the globe box under argon atmosphere and
fixed vacant sites in a frozen disordered structure gives rise to the transferred to DSC cell for the measurements. Samples in alu-
conductivity in glassy ionic conductors. However, the situation is minum pans were stabilized by slow cooling to –100 °C and
somewhat different than simple ion hopping in polymer then heated to 150 °C at a rate of 10 °C/min. The data were
electrolytes, as they are in rubbery state at room temperature recorded during the heating scan. Al2O3 powder was used as a
and the hopping sites are not fixed in a dynamically evolving reference. The real (ε′) and imaginary (ε″) parts of complex
disordered structure. To incorporate this dynamically changing dielectric permitivity (ε⁎) were measured with a Novocontrol
environment in the case of polymer electrolytes, several models, GmbH Concept 40 broadband dielectric spectrometer. The
such as the dynamic bond percolation (DBP) [9], dynamically membrane was sandwiched between two spring-loaded gold
disordered hopping (DDH) [10,11] have been proposed to electrodes. The film was heated to 100 °C and then cooled to
describe the ion transport processes characterized by two time − 100 °C. Temperature control was accomplished with a Quatro
constants: one describes the local carrier hopping and the other Cryosystem with a stability of ± 0.1 °C. Dielectric data (ε′, ε″)
one characterizes the reorganization of the host. From the were collected during cooling as a function of frequency from
viewpoint of an individual hopper, this means that if the local 10− 2 to 106 Hz using an oscillation level of 10 mV, at every
environment at any given time does not permit a hop, that 10 °C interval (after 15 min equilibration at each tempera-
environment will evolve such that, after a certain average waiting ture). For conductivity data analysis, the complex conductivi-
time the hop will no longer be prohibited. In these so called ty, σ⁎ = σ′ + iσ″, was calculated from the complex permittivity
polymer electrolytes, long range ionic diffusion leading to ionic using the following relations,
conductivity only occurs in the presence of local segmental
motions of the polymer host [1,3,4,9–12]. Thus, ionic conductiv- rV ¼ e0 xeW; rW ¼ e0 xe V ð1Þ
ity in polymer electrolytes is very much coupled to the segmental
motion of the host polymer—more is the segmental mobility, where ε0 is the permittivity of free space. The dc conductivity
higher is the ionic conductivity. Since the promising candidates values were obtained by extrapolating the frequency dependent real
for the polymer electrolytes have polar atoms (for solvating part of the complex conductivity (σ′) to zero frequency. The errors
cations) in their backbone, one way to investigate their segmental involved in the obtained conductivity values were less than 3%.
mobility is dielectric spectroscopy, which has been used
previously to study the ion dynamics in the polymer electrolytes 3. Result and discussion
[13–15].
In this paper we have synthesized polymer electrolyte 3.1. Composition dependence of Tg and XRD
membranes based on PEO and LiCF3SO3 as a function of EO/Li
ratio. Dielectric properties in these films were studied using Glass transition temperature (Tg) is an important parameter for
broadband dielectric spectroscopy as a function of temperature. amorphous polymer electrolytes. Fig. 1a shows the variation of
Dielectric relaxations of the polymer electrolytes have been the glass transition temperature, as measured from DSC of the
characterized by the empirical Havriliak–Negami function and polymer salt complexes (PSC) as a function of the EO/Li ratio
an attempt has been made to correlate the ionic conductivity to (DSC thermograms are shown in the inset). The Tg of PSCs
the segmental relaxation of the polymer electrolytes. increased from that of pure PEO with increasing EO/Li ratio up to
EO/Li = 24. However, further increase in the salt content did not
2. Experimental have much influence on the Tg. Though the increase of Tg with
increasing amount of salt is common, there have been reports
Solid polymer electrolyte membranes, comprised of poly- where Tg variation was not very smooth and even Tg decreased
ethylene oxide (PEO, MW N 5 × 106) and lithium trifluoro- with increasing amount of salt content in the polymer electrolytes
methanesulfonate (LiCF3SO3), were prepared by solution [16]. In the present case, the observation can be explained by
casting method using acetonitrile as solvent inside a glove assuming that the elevation of Tg with salt addition arises from the
box (Mbraun Unilab 2000), which was under constant Ar flow transient cross-linking between the Li ions and the O atoms in the
(H2O b 0.1 ppm). Appropriate amounts of the polymer and the polymer chains. So, in a way the elevation of Tg could be
salt, corresponding to various EO/Li ratios were first dissolved correlated to the number of the free Li ions — more is the number
into the solvent and stirred at room temperature for ~ 24 h. The of the free Li ions, higher is the elevation of Tg for PSCs. Even
resulting viscous slurry was then casted on a Teflon plate and though the samples with EO/Li b 24 contained more Li salt, the
dried at room temperature for ~ 24 h, to get free standing number of the free Li ions might not have increased. In effect, the
electrolyte membranes having thickness of ~ 100 μm. To ion-pairing for the samples with EO/Li b 24 kept the number of
remove the solvent completely from these membranes, they the free Li ions, which were responsible for the transient cross-
were further vacuum dried at ~ 55 °C for one more day. X- linking to the polymer chains, more or less constant. Hence, for
ray diffraction (XRD) was carried out using (Siemens D5000) the compositions with EO/Li b 24, even though the salt content
 N.K. Karan et al. / Solid State Ionics 179 (2008) 689–696 691

Fig. 2. Temperature dependence plots of the conductivity of polymer electrolytes

with various EO/Li ratios. The continuous lines show the fit of experimental data
to the VTF equation, σ = σo T − 1/2 exp(−B / T − To). Inset shows the conductivity
as a function of EO/Li ratio at different temperatures.

fitted to the Vogel–Tamman–Fulcher (VTF) equation [3,20] of

the form:

r ¼ r0 T 1=2 eEv =k ðT T0 Þ ð2Þ

where σ0 is the prefactor, T is the absolute temperature, k is

the Boltzmann constant, Ev is the pseudoactivation energy and T0
is the equilibrium glass transition temperature at which the “free”
volume disappears or at which configuration free entropy
becomes zero (i.e., molecular motions cease). The continuous
lines in Fig. 2 show the results of the non-linear least square fitting
of the conductivity data to Eq. (2) assuming a starting T0 value on
Fig. 1. (a) Variation of the glass transition temperature (Tg), (b) XRD patterns of
the polymer electrolytes with EO/Li ratios. Inset of (a) shows the DSC
the basis of the boundary condition (Tg-60 K) ≤T0 ≤ (Tg-40 K) as
thermograms for various compositions. described in the literature [3,21–23].The best fitted parameters,
σ0, Ev and T0 are listed in Table 1. Although free volume model
was originally adopted for explaining viscoelastic properties of
increased, the Tg remained almost the same. Fig. 1b shows the polymers, the reasonably good fit of σ to VTF equation over a
XRD pattern of the electrolyte membranes with various EO/Li wide range of temperature demonstrates the close coupling
along with that of the pure polymer membrane. All of the between the conductivity and the polymer chain segment mobility
diffraction peaks corresponding to the crystalline PEO were of the PSCs [24]. Inset of Fig. 2 shows the variation of
present in the XRD pattern of the electrolyte samples [17]. More- conductivity at different temperatures as a function of EO/Li ratio.
over, there were no extra peaks in the XRD spectra of polymer salt The conductivity increased gradually with increasing salt
complexes, which suggested that the electrolyte samples were concentration and reached a maximum of ~1.3 × 10− 6 S/cm at
semicrystalline in nature having PEO crystallites dispersed into room temperature for EO/Li ~ 24. Conductivity decreased with
the amorphous polymer salt complex matrix. further addition of salt. In a qualitative way, the conductivity

3.2. dc conductivity behavior Table 1

The glass transition temperature (Tg) and the VTF fitted parameters from the
The obtained conductivity (outlined in the Experimental temperature dependence of the conductivity and the dielectric relaxation times
section) variation of the PSCs as a function of temperature is of the polymer electrolyte for various salt content
shown in Fig. 2, where conductivity has been plotted as a EO/Li Tg (K) σ VTF fit VTF fit
function of 1000/T. It is clear from the plots that the tempera- σ0 Ev (eV) T0 (K) τ0 (s) Ev (eV) T0 (K)
ture dependence of conductivity did not follow the Arrhenius 200 212.31 45 0.207 155.50 2.47 × 10− 10 0.090 187
behavior. The bend in the curve has been observed in ionically 100 212.60 70 0.202 162.07 5.00 × 10− 12 0.119 190
conducting polymers and has been explained invoking the 30 218.85 125 0.172 172.75 1.31 × 10− 11 0.110 186
concept of free volume [18,19]. To have better insight into the 24 219.14 145 0.194 163.00 3.26 × 10− 11 0.110 182
8 216.44 185 0.207 161.00 1.58×10−12 0.125 182
temperature dependence of σ, the conductivity data have been
692 N.K. Karan et al. / Solid State Ionics 179 (2008) 689–696

temperatures has been analyzed for all the samples. Fig. 3a shows
the frequency dependence of the ac conductivity at various
temperatures for EO/Li = 30. A general trend in the frequency
dependence of the conductivity was observed at all temperatures
for all samples. An almost frequency independent conductivity
plateau at lower frequencies was followed by the conductivity
dispersion region at the higher frequencies. A similar kind of
frequency dispersion was observed for all other compositions as
well. The general trend of conductivity as a function of frequency
is very similar to the previous reports in the literature for ionically
conducting ceramics, glasses, polymers [14,22–26]. The switch-
over from the frequency independent to frequency dependent
region marks the onset of conductivity relaxation that gradually
shifts to higher frequencies as temperature increases (Fig. 3a). The
frequency independent conductivity is, in general, identified with
the dc conductivity (σdc). To quantitatively describe the frequency
dependence of the conductivity, σac has been fitted to the well-
known power law [27,28]

rac ¼ rdc þ Af n ð3Þ

where, σdc is the dc conductivity, A is constant (weakly

temperature-dependent), f is the frequency and n is the power
law exponent with 0 b n b 1. The conductivity data were fitted to
the above equation (continuous lines) and are shown in Fig. 3(a)
for EO/Li = 30 along with the experimental points. The con-
ductivity showed a hump in the intermediate frequencies that
moved to higher frequencies with increasing temperature due to
inherent inhomogeneous nature of the samples [29]. The fitting
was indeed reasonably good to the above equation for higher
temperatures (region I, 313 K to 253 K). However, it should be
mentioned here that the above two-term equation did not repro-
duce the experimental data at lower temperatures (below 253 K)
satisfactorily. In this low temperature region the conductivity
was best fitted by adding a third term, which was nearly linear in
frequency, to the above two-term equation. Therefore, in the low
temperature region (region II) the conductivity was fitted to the
equation

rac ¼ rdc þ Af n þ Bf ð4Þ

where, B is a constant. This nearly linear frequency dependent

term in the conductivity leads to the frequency independent di-
electric loss (ε″ = σ′/2πε0 f ).
Fig. 3. (a) Frequency dependence of the conductivity at various temperatures for
EO/Li = 30. The dotted and continuous lines are the fits of the experimental ac Generalized Universal Power Law (G-UPL) (Eq. (3)) has been
conductivity data to Eqs. (3) and (4), respectively. (b) Temperature dependence very successful in explaining the frequency dependence of σ′ in
of the dielectric loss at different frequencies for EO/Li = 30. (c) Variation of the disordered materials (ionically conducting glasses, conducting
power law exponent, n with temperature for various EO/Li. Inset shows the polymers etc) within the framework of the physical models [25,26]
variation of n as a function of EO/Li at different temperatures.
that has been proposed for explaining the conductivity mechanism
variation was similar to the variation of Tg with EO/Li, suggesting in structurally disordered ionic conductors. However, Nowick
the decrease in the conductivity was due to the ion-pairing at et al. [30,31] have shown that the two-term equation (σ′ =σ0 +Af n)
higher salt concentrations. does not sufficiently reproduce the experimental results and there
was a need to bring in a third term, which was almost independent
3.3. ac conductivity behavior of frequency. This procedure has been adopted widely in the
literature and the existence of NCL is now well recognized for
In order to understand the ion dynamics in the polymer elec- ionically conducting glasses, ceramics, etc [30–38]. A contribu-
trolytes, the frequency dependence of the conductivity at various tion from thermally activated ionic motion is seen in any
 N.K. Karan et al. / Solid State Ionics 179 (2008) 689–696 693

equivalent representation of the electrical data (σ⁎ or ε⁎ or M⁎).

However, there is always present another contribution that is
almost frequency independent in the ε⁎ representation and leads to
the addition of the third term to the UPL function. At low
temperatures the thermally activated contribution from ionic
motion is moved to low frequencies outside the experimental
window and the constant loss is the only contribution that remains
[37]. Even though existence of “ubiquitous” NCL is now well
accepted, its origin is still not completely understood and remains
controversial. Jain et al. [36] proposed that the NCL arise from the
motion of group of atoms in the asymmetric double well potential
configurations. Recently, Ngai et al. [32,37] have pointed out that
the NCL originates from the slowing down of ionic motions in the
cages due to the ion–ion interactions and correlations in the short
time regime. Similar studies in various glassy, ceramic and
polymers with different compositions are needed in order to gain
further insights about the origin of NCL phenomenon. Polymer
electrolytes (with glass transition temperature well below room
temperature) differ from conventional disordered systems in that
their structures are not fixed (in the timescale of experiments),
rather it is dynamic in nature, i.e. it evolves with time, which in no Fig. 4. Frequency dependence of ε′ and ε″ at 293 K for various EO/Li.
way should prohibit the manifestation of NCL behavior in the
polymer electrolytes. The observation of NCL behavior in polymer
electrolytes along with ion conducting glasses and ceramics points
to the fact that the NCL is a universal phenomenon in the case of PSCs along with pure PEO. The real part of the permittivity, ε′
ionic conductors. showed two relaxation processes for both pure PEO and PSCs
As a support to the addition of a nearly linear frequency (Fig. 4a). The appearance of these processes is more clearly
dependent term to the two-term equation (Eq. (4)) for describing seen in ε″ vs. frequency plot (Fig. 4b). The relaxation in the low
the frequency dependence of conductivity at lower tempera- frequency region is the electrode polarization phenomenon
tures, the imaginary part of permittivity, ε″ for various occurring as a result of an accumulation of ions near the
frequencies has been plotted against temperature in Fig. 3b electrodes. To further justify the point the imaginary part of the
for EO/Li = 30. As can be seen from Fig. 3b, below ~ 243 K, ε″ conductivity, σ″ has been plotted along with ε″ as a function of
becomes independent of frequency, which in turn supports the frequency at various temperatures in Fig. 5a. It has been
incorporation of the linear frequency dependent term in the reported that the peaking of σ″ spectra in the low frequency
above equation [34]. So, clearly there exists a cross over from region manifests the electrode polarization [15]. In the present
NCL dominated ac conductivity at lower temperatures to ion case, the low frequency relaxation is accompanied by the
hopping mediated ac conductivity at higher temperatures in peaking of σ″ at the same frequency (marked by horizontal
polymer electrolytes. The combination of NCL and fractional lines in Fig. 5a), which clearly suggests that the low frequency
power law describes well the dynamics of ions in the polymer relaxation is indeed due to electrode polarization. The relaxation
electrolytes. at the high frequency arises from the motion of the polymer
The temperature dependence of the extracted power law chains. The addition of salt had a profound effect on the re-
exponent, n is shown in Fig. 3c for all PSCs. The n values lied in laxation rate for this process and the relaxation could be, ten-
the range 0.7–0.4 over the temperature range studied. Values of tatively, attributed to the motions of the Li ion coordinated
n in this range are common for the polymer electrolytes reported polymer chain segments [41].
in the literature [14,21,39]. At lower temperatures, n remains In order to understand the dynamics of the relaxation process
almost constant and it increases with increasing temperature for in polymer electrolytes, the experimental complex dielectric
all PSCs. The inset of Fig. 3c shows the composition depen- permittivity, ε⁎ (ω) data were fitted to the Havriliak–Negami
dence of n at selected temperatures. At constant temperature, n function [42]:
decreases as salt concentration increases. This agrees well with Xn o
b
the idea that n physically represents the strength of the ion–ion e4ðxÞ ¼ e V  jeW ¼ el þ ðel  es Þ=½ð1 þ ixsÞa 
interactions and as the salt concentration increases the ion–ion ð5Þ
interactions are expected to increase [40]. þ ðr=e0 xÞN

3.4. Dielectric relaxation where, the second term accounts for the conductivity, which
affects the low frequency loss. The summation symbol in the
Fig. 4 shows the real (ε′) and imaginary (ε″) parts of first term is to indicate the occurrence of more than one re-
permittivity as a function of frequency at 293 K for various laxation process including the electrode polarization. The
694 N.K. Karan et al. / Solid State Ionics 179 (2008) 689–696

tion and the high frequency relaxation did not follow the
Arrhenius behavior. Similar type of temperature dependence of
the relaxation time for the electrode polarization has been
observed elsewhere for polymer electrolyte [15]. For quantita-
tive comparison they have been fitted to the VTF function of the
following form

s ¼ s0 eEv =k ðT T0 Þ ð6Þ

where, τ is relaxation time, τ0 is relaxation time pre-exponential

factor, T is absolute temperature, Ev is pseudoactivation energy
and T0 is the equilibrium glass transition temperature. The fitted
parameters, τ0, Ev, T0 are listed in Table 1.
The temperature dependences of the fitting parameters (α
and β) are shown in Fig. 7 for the high frequency relaxation
process. As can be seen from Fig. 7, the parameter α decreased
with decreasing temperature, implying the broadening of the
distribution of relaxation times as temperature decreases. The
parameter β remained almost constant at 1 in the whole tem-
perature range.

Fig. 5. (a) Frequency dependence of ε″ and σ″ at different temperatures of

polymer electrolyte with EO/Li = 30. (b) HN fitting of the frequency dependence
of ε″ at different temperatures of EO/Li = 30. The symbols are the experimental
points and the continuous lines are fits to the H–N equation.

symbols ε∞ and εs are the limiting dielectric constants at

infinitely high and low frequencies, respectively; ω is the
angular frequency (ω = 2πf ); τ is the characteristic average
relaxation time of the process; σ is the conductivity; εo is the
vacuum permittivity and N is an exponent that characterizes the
conduction process. N was found to lie in between 0.5 and 0.8
depending upon salt concentration and temperature. The
exponents α and β (0 b α, β b 1) are the HN fitting parameters
describing the distribution of the relaxation times. The dielectric
relaxation strength, ε∞ − εs, is represented as Δε.
The frequency dependence of the dielectric loss for EO/
Li = 30, at several temperatures along with the corresponding
fits to Eq. (5), are shown in Fig. 5b, where the symbols represent
the experimental data and the continuous lines are fitted values.
The relaxations became increasingly slower with decreasing
temperature. The temperature dependences of the relaxation
time, for both the electrode polarization and the high frequency
relaxation, obtained from the fitting are shown in Fig. 6, Fig. 6. Temperature dependence of the relaxation times: (a) the electrode
respectively. It is clearly seen from Fig. 6 that the temperature polarization; (b) the dielectric relaxation of the polymer electrolytes. The
dependence of relaxation times for both the electrode polariza- continuous lines are the fits to the VTF equation.
 N.K. Karan et al. / Solid State Ionics 179 (2008) 689–696 695

the ionically conducting polymers can be expressed in terms of

the mobility μ by the relationship

r ¼ μnq ð8Þ

where n and μ are, respectively, the number density and

mobility of the charge carriers and q is the charge. So, a
complete description of temperature dependence of conductiv-
ity should take temperature dependence of both the mobility and
ion concentration into account. Since the mobility of free ions is
expected to be controlled by the segmental motion of polymer
matrix, it should have a VTF temperature dependence [13]
μ ¼ μ0 exp ½E1 =k ðT  T0 Þ ð9Þ

where, E1 is the activation energy for mobility, μ0 is the infinite

Fig. 7. Temperature dependence of the H–N fitted parameters α and β for the temperature mobility. Moreover, the mobile ion concentration is
dielectric relaxation of the polymer electrolytes. assumed to follow an Arrhenius behavior [46],
n ¼ n0 exp ½E2 =kT  ð10Þ
3.5. Coupling between ionic conductivity and dielectric
where, E2 is the coulomb energy of cation anion pair and n0 is
relaxation
total ion concentration as T → ∞. So, the temperature depen-
dence of conductivity can be expressed by combining Eqs. (9)
In polymer electrolytes the ion transport is assisted by
segmental movement of the polymer chains and in the literature and (10) as
the relaxation mode responsible for ionic conduction has been r ¼ μ0 n0 q exp ½E2 =kT  E1 =k ðT  T0 Þ ð11Þ
correlated to α relaxation [10,43]. Several workers have noted
correlation between the conductivity of PPO (polypropylene Consequently, this equation does not predict a simple VTF
oxide) salt complexes and the characteristic frequency of the behavior for the conductivity. Furthermore, it shows that the
dielectric relaxation corresponding to the glass transition in pure activation energy for the conductivity is not simply E2 but also
host polymer [44,45]. The temperature-dependent response of includes an Arrhenius activation energy term E1. This might be
both these quantities follows the empirical VTF equation: the reason behind the apparent discrepancy between the activation
energy values obtained from the temperature dependence of the
f ðT Þ ¼ AeðB=ðT T0 ÞÞ ð7Þ conductivity fitting and the dielectric relaxation fitting. It should
be pointed out here that the best fitted T0 values obtained from the
where, A, B and T0 are the fitting parameters and f(T) is either dielectric relaxation fitting are approximately 20 K higher than
the characteristic frequency of the relaxation or the dc those obtained by fitting the conductivity with a simple VTF
conductivity of the salt complexes. Typically, the fits to the relation (Table 1), which is in good agreement with the previous
VTF equation for the conductivity data of PSCs and the report [46]. Since the VTF dependence is characteristic for an
dielectric data on relaxation rate of the glass transition amorphous material above the glass transition temperature, it
relaxation result in similar values for the parameter B [44,45]. shows that the flexibility of the polymer chains, characteristic
Based on this similarity of B values obtained from two different for amorphous regions of semicrystalline polymer electrolytes,
aspects, it has been suggested that the relaxation corresponding determines the ionic conductivity. This observation clearly de-
to the glass transition is indeed very much coupled to the ion monstrates that the segmental motions of the polymer salt
transport in polymer electrolytes. However, the preceding complex rather than that of the host polymer alone control the ion
discussion focused on the relation between the relaxation conduction significantly in polymer electrolytes.
processes in the pure polymer host and the conductivity of the
corresponding PSCs. On the contrary, the knowledge of the 4. Conclusions
relaxation rate in the PSCs would be of greater interest rather
than that in pure host polymer, while correlating ionic Polymer electrolytes consisting of polyethylene oxide (PEO)
conductivity to segmental motions. In the present case, it is and LiCF3SO3 were synthesized by a solution casting method
seen from Table 1 that the VTF fitting parameter, B, is not with different EO/Li ratio. The segmental relaxation process
comparable in a quantitative way for the conductivity measure- and its role in controlling ion transport in these solid polymer
ments and for the dielectric relaxation of the polymer salt electrolytes have been investigated using dielectric spectro-
complexes. This apparent discrepancy of B value obtained from scopy and thermal analysis. DSC results revealed that up to EO/
the conductivity fitting and the dielectric fitting can be Li = 24, Tg increases with increasing salt content implying the
accounted by considering the fact that the conductivity, σ in coordination of Li ions to the polymer backbone oxygens.
696 N.K. Karan et al. / Solid State Ionics 179 (2008) 689–696

Beyond this concentration ion-pairing limits the number of free [11] A. Nitzan, M.A. Ratner, J. Phys. Chem. 98 (1994) 1765.
Li ions and the Tg remains almost constant. The temperature [12] D.F. Shriver, Chem. Rev. 88 (1988) 109.
[13] J.P. Runt, J.J. Fitzgerald (Eds.), Dielectric Spectroscopy of Polymeric
dependence of ac conductivity revealed the existence of near- Materials: Fundamentals and Applications, American Chemical Society,
ly constant loss (NCL) contribution at low temperatures in Washington, DC, 1997.
polymer electrolytes and there is a crossover from NCL at low [14] V. Di Noto, J. Phys. Chem. B 106 (2002) 11139.
temperatures to ion hopping ac conductivity at higher tem- [15] T. Furukawa, M. Imura, H. Yuruzume, Jpn.J. Appl. Phys. 36 (1997) 1119.
[16] J.S. Thokchom, C. Chen, K.M. Abraham, B. Kumar, Solid state Ion. 176
peratures. The observed dielectric relaxation in the loss (ε″)
(2005) 1887.
spectrum is attributed to the motion of the cation coordinated [17] C.P. Rhodes, R. Frech, Macromolecules 34 (2001) 2660.
PEO chain. Compared to pure PEO, the strength and the lo- [18] M.H. Cohen, D. Turnball, J. Chem. Phys. 31 (1959) 1164.
cation of the relaxation frequency showed a drastic change [19] G.S. Grest, M.H. Cohen, Phys. Rev. B 21 (1980) 4113.
in the PSCs. The temperature dependence of H–N fitting [20] H. Vogel, Z. Phys. 22 (1921) 645.
parameters showed that the symmetric broadening of relaxation [21] V. Di Noto, V. Zago, G. Pace, M. Fauri, J. Electrochem. Soc. 151 (2004)
A224.
curves decreased with increasing temperature, whereas the [22] V. Di Noto, M. Vittadello, S. Lavina, M. Fauri, S. Biscazzo, J. Chem. Phys.
asymmetric broadening and dielectric relaxation strength B 105 (2001) 4584.
remained constant with temperature. The temperature depen- [23] V. Di Noto, M.D. Longo, V. Munchow, J. Chem. Phys. B 103 (1999) 2636.
dence of conductivity and relaxation times associated with the [24] T. Miyamoto, K.J. Shibayama, K. J. Appl. Phys. 44 (1973) 5372.
relaxation of polymer electrolytes followed VTF behavior [25] B. Roling, C. Martiny, K. Funke, J. Non-Cryst. Sol. 249 (1999) 201.
[26] B. Roling, J. Non-Cryst. Sol. 244 (1999) 34.
yielding qualitatively similar pseudoactivation energy values. [27] D.P. Almond, A.R. West, R.J. Grant, J. Solid State Commun. 44 (1982)
These observations clearly shows that the ion transport is 1277.
closely coupled to the ion coordinated segmental motions, [28] D.P. Almond, G.K. Duncan, A.R. West, Solid State Ion. 8 (1983) 159.
which further suggests that the polymer segmental dynamics [29] J.R. Dygas, B.M. Faraj, Z. Florjan'czyk, F. Krok, M. Marzantowicz, E.Z.
play a crucial role in controlling the ion transport processes in Monikowska, Solid State Ion. 157 (2003) 249.
[30] W.K. Lee, J.F. Liu, A.S. Nowick, Phys. Rev. Lett. 67 (1991) 1559.
the polymer electrolytes. [31] A.S. Nowick, B.S. Lim, A.V. Vaysleyb, J. Non-Cryst. Solids 172–174
(1994) 1243.
Acknowledgements [32] C. Leon, A. Rivera, A. Varez, J. Sanz, J. Santamaria, K.L. Ngai, Phys. Rev.
Lett. 86 (2001) 1279.
The financial support from DOE grant (DE-FGO2- [33] A. Rivera, J. Santamaria, C. Leon, J. Sanz, C.P.E. Varsamis, G.D.
Chryssikos, K.L. Ngai, J. Non-Cryst. Solids 310 (2002) 1024.
01ER45868) is gratefully acknowledged. One of us (N. K. [34] A. Rivera, J. Santamaria, C. Leon, K.L. Ngai, J. Phys., Condens. Matter 15
Karan) is grateful to the NSF RII project for the graduate (2003) S1633.
fellowship. [35] J.R. Macdonald, Phys. Rev. B 66 (2002) 064305.
[36] H. Jain, X. Lu, J. Non-Cryst. Solids 196 (1996) 285.
[37] K.L. Ngai, J. Chem. Phys. 110 (1999) 10576.
References
[38] B. Natesan, N.K. Karan, R.S. Katiyar, Phys. Rev. E 74 (2006) 042801.
[39] T. Winie, A.K. Arof, Ionics 12 (2006) 149.
[1] F.M. Gray, Solid Polymer Electrolytes, VCH, New York, 1991.
[40] A. Pan, A. Ghosh, Phys. Rev. B 62 (2000) 3190.
[2] Z. Gadjourova, Y.G. Andreev, D.P. Tunstall, P.G. Bruce, Nature 412 [41] R. Bergman, A. Brodin, D. Engberg, Q. Lu, C.A. Angell, L.M. Torell,
(2001) 520. Electrochim. Acta 40 (1995) 2049.
[3] J.R. MacCallum, C.A. Vincent, Polymer Electrolyte Reviews 1, Elsevier,
[42] S. Havriliak, S. Negami, Polymer 8 (1967) 161.
London, 1987. [43] R.M. Fuoss, J. Am. Chem. Soc. 63 (1941) 369.
[4] J.R. MacCallum, C.A. Vincent, Polymer Electrolyte Reviews 2, Elsevier, [44] J.J. Fontanella, M.C. Wintersgill, M.K. Smith, J. Semancik, C.G. Andeen,
London, 1989. J. App. Phys. 60 (1986) 2665.
[5] S.D. Druger, M.A. Ratner, A. Nitzan, Solid State Ion. 9/10 (1983) 1115.
[45] A.L. Tipton, M.C. Lonergan, M.A. Ratner, D.F. Shriver, T.T.Y. Wong, K.
[6] C. Berthier, W. Gorecki, M. Minier, M.B. Armand, J.M. Chabango, P. Han, J. Phys. Chem. 98 (1994) 4148.
Rigaud, Solid State Ion. 11 (1983) 91. [46] R.J. Klein, S. Zhang, S. Dou, B.H. Jones, R.H. Colby, J. Runt, J. Chem.
[7] Z. Stoeva, I.M. Litas, E. Staunton, Y.G. Andreev, P.G. Bruce, J. Am. Phys. 124 (2006) 144903.
Chem. Soc. 125 (2003) 4619.
[8] C.A. Angell, Solid State Ion. 18/19 (1986) 72.
[9] S.D. Druger, A. Nitzan, M.A. Ratner, J. Chem. Phys. 79 (1983) 3133.
[10] M.C. Lonergan, A. Nitzan, M.A. Ratner, D.F. Shriver, J. Chem. Phys. 103
(1995) 3253.