ORIGINAL PAPER

High performance PEO-based polymer electrolytes

and their application in rechargeable lithium polymer batteries

H. H. Sumathipala & J. Hassoun & S. Panero & B. Scrosati

Received: 14 May 2007 /Accepted: 4 July 2007 /Published online: 15 August 2007# Springer-Verlag 2007

Abstract Solvent-free, lithium-ion-conducting, composite

polymer electrolytes have been prepared by a double

dispersion of an anion trapping compound, i.e., calyx(6)

pyrrole, CP and a ceramic filler, i.e., super acid zirconia, S-

ZrO2 in a poly(ethylene oxide)-lithium bis(oxalate) borate,

PEO–LiBOB matrix. The characterization, based on differ-

ential thermal analysis and electrochemical analysis,

showed that while the addition of the S-ZrO2 has scarce

influence on the transport properties of the composite

electrolyte, the unique combination of the anion-trapping

compound, CP, with the large anion lithium salt, LiBOB,

greatly enhances the value of the lithium transference

number without depressing the overall ionic conductivity.

These unique properties make polymer electrolytes, such as

PEO20LiBOB(CP)0.125, of practical interest, as in fact

confirmed by tests carried out on lithium battery prototypes.

Keywords Polymer . Electrolyte . Lithium . Battery

Introduction

Great attention is presently devoted to lithium batteries, as

they may greatly contribute to achieve important goals such

as energy renewal and environmental control. In fact,

lithium batteries are expected to play a key role in the

storage of intermittent energy sources, e.g., solar or wind,

as in powering controlled emission cars, e.g., electric or

hybrid vehicles.

However, the present lithium battery technology is not

yet totally adequate for these applications, and its progress

requires the innovation of the battery structure with the use

of electrode and electrolyte materials which are more

energetic, safer, and more economic than the conventional

ones.

For instance, it is advisable to move from the present

liquid electrolyte to a solid polymer electrolyte, as this is

expected to increase the reliability of the battery as well as

to improve its environmental compatibility and diversifies

its geometry and design [1]. In this respect, polymer

electrolytes made by the combination of poly(ethylene

oxide) and a lithium salt, PEO–LiX, have drawn much

attention due to their unique features, which include high

conductivity and good chemical compatibility [2, 3].

These PEO–LiX membranes, which are true solid

solvent-free polymer electrolytes, have so far been scarcely

used due to some intrinsic still to be solved problems. One

of them is the thermal dependence of the conductivity,

which in fact assumes high values only at temperatures

above 70 °C [3]. However, this problem is not really

crucial, and it may be faced in the type of applications in

which the battery is called. For instance, operational

temperatures in the 80–90 °C range are not in contrast with

the requirements for batteries designed for power cars.

Much more serious, in terms of practical applications, is

the low lithium ion transference number, which, for

standard PEO–LiX electrolytes, averages on 0.2–0.3 [3].

As the electrochemical process of lithium batteries involves

the lithium ion intercalating process to and from the

cathode [4], a low transference number results in concen-

Ionics (2007) 13:281–286

DOI 10.1007/s11581-007-0137-4

H. H. Sumathipala : J. Hassoun : S. Panero :B. Scrosati (*)

Dipartimento di Chimica, Università di Roma,

“La Sapienza,

00185 Rome, Italy

e-mail: bruno.scrosati@uniroma1.it

H. H. Sumathipala

Department of Physics, University of Kelaniya,

Kelaniya, Sri Lanka

tration polarizations, which, in turn, greatly limit the power

capability of the battery.

Accordingly, great efforts are presently directed to solve

this problem. Within a research project involving three

academic laboratories, including ours, a new approach has

been introduced which resulted in an enhancement of the

lithium ion transference number of the PEO–LiX polymer

electrolyte [5]. The approach consists of dispersing an

“anion trapping” additive, i.e., calyx(6)pyrrole, CP, in the

polymer matrix. This led to a composite polymer electrolyte

with an outstanding value of the lithium transference

number, which passes from 0.3 to almost unity [6].

However, this enhancement was somehow contrasted by a

certain decay in conductivity [5].

More recently, we have refined the concept by dispersing

in the polymer matrix two complementary additives, chosen

in such a way that one of them, CP, was expected to

enhance the lithium ion transference number and the other,

super-acid zirconia, SZ, to keep the conductivity at high

levels. The results indeed demonstrated that this “dual

composite” approach, applied to a PEO–LiCF3SO3 polymer

electrolyte, is conceptually correct and very promising [7].

In this work, we have further expanded the approach by

studying the role of the lithium salt in the PEO–LiX

combination, i.e., by passing from LiCF3SO3 to lithium bis

(oxalate) borate, LiBOB. The use of a salt with a large

anion with a delocalized charge, such as BOB, is known to

enhance the amorphous content in the PEO structure, and

thus, to promote ionic conductivity [8, 9]. An extra bonus

of LiBOB is in the environmental compatibility and safety

which is higher than that of more conventional salts, such

as LiClO4 or even LiCF3SO3. It is then expected that the

PEO–LiBOB combination with dispersed CP and SZ

additives may give rise to a highly lithium-conducting,

safe, and sustainable polymer electrolyte.

To verify this assumption, we have determined the

electrochemical properties of the CP/SZ-added PEO–

LiBOB dual composite electrolyte and evaluated its

practical relevance as separator in rechargeable lithium

batteries. The results are reported in this work.

Experimental

Three polymer electrolyte samples were prepared and tested

in this work, all with the same PEO to LiBOB and PEO–

LiBOB to CP ratios. The composition of these samples and

their acronyms are shown in Table 1.

Calix(6)pyrrole was prepared adopting the procedure

described in [7]. PEO (600,000 MW, Aldrich regent grade)

and lithium bis(oxalate)borate (Libby) were dried under

vacuum at 40 and 150 °C, respectively, for at least 36

h before use. Zirconia, obtained by Mel Chemical in the

form of super-acid Zr(OH)2 was heat-treated at 500 °C for 2

h to turn it into the desired super-acid version, S-ZrO2 (SZ)

and vacuum-dried at for 6 h before use. The electrolyte

components were carefully sieved and then introduced in

their correct proportion inside sealed polyethylene bottles

and thoroughly mixed by soft ball-milling for at least 24

h to obtain homogeneous mixture of powders.

Composite, PEO20LiBOB (CP)0.125 polymer electrolyte

membranes were prepared varying the amount of dispersed

SZ. The preparation was carried out following a procedure

established in our laboratory [7]. To avoid any contamina-

tion with external ambient, all the samples were prepared in

a controlled argon atmosphere having a humidity content

below 10 ppm and were removed from the dry box for any

further test only after housing them in sealed coffee bag

envelopes.

Homogeneous rigid membrane samples having thickness

ranging from 50 to 300 μm were obtained after hot

pressing. They were stored in an argon-filled dry box for

subsequent measurements.

The differential scanning calorimetric (DSC) study was

performed using a Mettler DSC calorimeter. The samples

were prepared in an argon-filled dry box by sealing

polymer electrolyte pieces (about 2 mg) in aluminum pans.

They were heat-treated at 100 °C for 1 h and stored for 7

days at room temperature before scanning. Care was taken

to use the same thermal history for all samples. The

scanning tests were done under a nitrogen flow at 5 °C/min

heating rate in the 25–120 °C temperature range.

The ionic conductivity measurements were performed by

AC impedance spectroscopy. The conductivity cells were

prepared by housing electrolyte membrane samples having

an 8-mm diameter in a Teflon O-ring and sandwiching them

between two blocking stainless steel electrodes. The O-ring

circles the membrane sample so as to maintain fix sample

thickness throughout the measurement. The cells were

heated to about 100 °C and kept at this temperature for

24 h to reach the thermal equilibrium. The measurements

were made in the 100–40 °C temperature range on both

cooling and heating scans. The cells were maintained at a

given temperature for 24 h to thermally equilibrate them

before each measurement.

The lithium ion transference number was determined by

the technique introduced by Bruce et al. [10]. A constant

DC voltage was applied cross a symmetrical Li/polymer

Table 1 Composition and acronym of the polymer electrolyte

samples prepared in this work

Membrane sample Composition

PEO (PEO)20LiBOB(CP)0.125PEO-5% (PEO)20LiBOB(CP)0.125+5%SZ

PEO-10% (PEO)20LiBOB(CP)0.125+10%SZ

282 Ionics (2007) 13:281–286

electrolyte/Li cell, and the current through the cell was

monitored until a steady-state value was reached. An

impedance analysis was carried out immediately before

and after the application of the DC voltage to estimate the

effect of changes of the passive layer resistance at the Li

electrode interface. The impedance spectra were analysed

by the non-linear fit method developed by Boukamp [11,

12]. The cell was left to thermally equilibrate for at least 1

day before each measurement.

The lithium ion transference number TLi+ was calculated

by the following equation:

TLiþ ¼Iss $V � RoIoð Þ

Io $V � RssIssð Þð1Þ

where Io is the initial current, Iss is the steady state current,

ΔV is the applied voltage, Ro and Rss are the initial and

final, respectively, resistances of the passive layer onto

lithium metal electrodes.

The Li/electrolyte/LiFePO4 cells were assembled by

sandwiching a lithium metal foil, a polymer electrolyte

membrane, and the LiFePO4 composite cathode film. The

latter was prepared by blending the LiFePO4 active material

(80%) with Super-P carbon as the electronic conductor (10

mass%) and PVdF 6020 Solay Solef(10 mass%) as a

binder. The components of the cell were placed in a Teflon

container having two stainless steel plates as the current

collectors. Care was taken to avoid direct contact between

the Li anode and the Teflon container. All the assembling

and the testing procedures were done in a controlled argon

atmosphere dry-box having both humidity and oxygen

content below 10 ppm.

The cells were characterized by galvanostatic cycling in

the 3.0- to 3.8-V range at different current densities and at

different temperatures (80–100 °C). The performance of the

cells was evaluated in terms of specific capacity, charge/

discharge efficiency, and cycle life. Before the tests, the

cells were kept at the highest testing temperature, i.e., about

100 °C, for at least 36 h to reach the thermal equilibrium as

well as to allow the diffusion lithium salt inside the cathode

film. The operating temperature of the cell was controlled

by a Buchi oven, and the data acquisition was done using

the Maccor 1400 battery tester.

Results and discussion

Three different samples, varying by their SZ content, have

been prepared and tested in this work. Figure 1 shows the

surface appearance of one of them chosen as a typical

example, i.e., the (PEO)20LiBOB(CP)0.125. The optical

picture clearly shows the homogeneity of the membrane,

which is maintained in all the examined samples.

Figure 2 shows the DSC response of the three polymer

electrolyte samples prepared in this work, see Table 1. The

figure provides a useful comparison between the response

of the SZ filler-free membrane and those containing

different amounts of filler. A double-peak feature is noticed

for the curves associated with the ceramic-added samples.

As already discussed in previous works [8, 9], this may be

associated to the influence of the ceramic filler on the

melting process. However, overall, the difference between

the various samples is minor, and this suggests that the

dispersion of the SZ filler does not significantly influence

the level of the amorphous content of the (PEO)20LiBOB

Fig. 1 Surface appearance of a P(EO)20LiBOB (CP)0.125 polymer

electrolyte membrane

20 40 60 80 100 120

-0.5

-0.4

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

En

do

He

at F

low

(W/g

)

Temperature / °C

PEO

PEO + 5%

PEO + 10%

Fig. 2 DSC traces of P(EO)20LiBOB (CP)0.125 polymer electrolytes

with and without addition of the SZ filler. Scan rate: 5 °C/min

Ionics (2007) 13:281–286 283

(CP)0.125 polymer electrolyte. This is not surprising, as the

amorphicity is already promoted by the large lithium salt

anion, and thus, a low effect by the filler is expected.

Table 2 summarizes the thermal data derived from the

DSC analysis. For comparison purpose, also the data

related to a CP-free PEO–LiBOB membrane, which

underwent comparable thermal history and derived from

reference [9], are reported. The data of Table 2 confirm the

above-discussed scarce influence of the ceramic filler on

the thermal properties of the electrolytes. Indeed, their

melting temperature, Tm, does not significantly change

passing from SZ-free to SZ-added membranes.

However, the presence of calixpyrrole significantly

modifies both Tm and ΔHm. Therefore, it is reasonable to

assume that, in addition to the large BOB anion, also the

CP macromolecule contributes to enhance the amorphous

content of the polymer. This further explains why the effect

of SZ cannot be relevant in this respect.

The lithium transference number, TþLi , of the CP-added

polymer electrolyte sample, i.e., the PEO20LiBOB(CP)0.125membrane, was determined by DC polarization test

combined by impedance spectroscopy, see “Experimental”

part. A value of TþLi in the range of 0.50–0.60 was found.

Previous work carried out in our laboratory showed that the

TþLi of the pristine PEO20LiBOB electrolyte is of the order

of 0.25–0.30 [8, 9]. Thus, it is clear that the addition of CP

has significantly increased the lithium transport.

Figure 3 shows the Arrhenius plots of the ionic

conductivity of the membranes with and without addition

of CP. The comparison between the two plots shows that

the addition of CP has not depressed the conductivity of the

electrolyte, but even partly raised it. This is probably

associated to the already discussed effect of CP in

promoting the amorphous content of the polymer. The

results of Fig. 3 is important, as it shows that the unique

combination of an anion trapping agent, CP, with a large

anion lithium salt, LiBOB, leads to composite polymer

electrolytes having a high lithium transference number

without depression in ionic conductivity. This is a rare

event, which makes the PEO20LiBOB(CP)0.125 membrane

of considerable interest for practical applications.

Figure 4 compares the conductivity Arrhenius plots of

the CP-added polymer electrolytes with and without

addition of the SZ ceramic filler. We notice that the filler

has no positive effects on the conductivity but rather

depresses it when added at high concentration. This is not

surprising in view of the already discussed competition

with the lithium salt:the large anion promotes by itself the

amorphous character of the polymer, this vanishing any

possible contribution by the filler which then is present as

inert component and, under this condition, may even

depress the conductivity by blocking the percolation

channels.

At the light of the above-discussed results, the PEO20

LiBOB(CP)0.125 membrane has been selected as the most

Table 2 Thermal data of P(EO)20LiBOB (CP)0.125 polymer electro-

lytes with and without addition of the SZ filler

Sample Tm (°C) ΔHm(Jg−1)

P(EO)20LiBOB(CP)0.125 56 61

P(EO)20LiBOB(CP)0.125+5% SZ 55 54

P(EO)20LiBOB(CP)0.125+10% SZ 53 58

P(EO)20LiBOBa 45.5 95

a From reference [9]

2.6 2.7 2.8 2.9 3.0 3.1 3.2

1E-5

1E-4

1E-3

0.01

Co

nd

uctivity/ S

cm

-1

1000/T / K-1

P(EO)20

LiBOB

P(EO)20

LiBOB(CP)0.125

Fig. 3 Arrhenius plots of the ionic conductivity of the P(EO)20LiBOB

and of P(EO)20LiBOB (CP)0.125 polymer electrolyte membranes

2.6 2.7 2.8 2.9 3.0 3.1 3.2

1E-5

1E-4

1E-3

0.01

Co

nd

uctivity/ S

cm

-1

1000/T / K-1

PEO

PEO-5%

PEO-10%

Fig. 4 Arrhenius plots of the ionic conductivity of the P(EO)20LiBOB

and of P(EO)20LiBOB (CP)0.125 polymer electrolytes with and without

addition of the SZ filler

284 Ionics (2007) 13:281–286

promising electrolyte for battery applications. To confirm

this, we have tested this electrolyte in a prototype lithium

cell using lithium iron phosphate as cathode:

Li=PEO20LiBOB CPð Þ0:125�

LiFePO4 ð2Þ

The electrochemical process of this cell is expected to be

the cycling lithium de-insertion/insertion in lithium iron

phosphate:

xLiþ LiFePO4$ 1� xð ÞLiþ Li 1�xð ÞFePO4 ð3Þ

to which is associated a theoretical specific capacity of 170

mAhg−1 [13]. This is also the maximum capacity of the battery

which, using an excess of lithium anode, is cathode-limited.

Figure 5 shows some typical voltage profiles obtained at

various C rates (Fig. 5a) and at various temperatures (Fig.

5b) . The two-phase flat voltage plateaus evolving around

3.4 V which are typical of the lithium iron phosphate

(process [2]) are easily distinguishable. At C/10 rate, a

specific capacity of the order of 120 mAhg−1, i.e., 70% of

the theoretical, is obtained. As expected, the capacity

decays at higher rates, however still remaining at apprecia-

ble values.

Figure 6 shows the effect of the temperature on the

capacity of the battery. At temperatures above 83 °C, the

battery is capable of delivering a high capacity level, i.e.,

>80 mAhg−1 at current densities lower than 0.25 mAcm−2,

corresponding to C/5, C/8, and C/10 rates. The decay in

capacity with increasing rate and decreasing temperature is,

as expected, associated to the increase in ohmic drop.

However, it is important to remark that the current density

that the battery is able to sustain is quite higher than the

limiting value of conventional PEO-based lithium batteries,

reported as 0.1 mAcm−2 [14, 15]. In the battery here

discussed, this value rises to 0.25 mAcm−2, and this

important increase is associated with the enhancement in

the lithium transference number promoted by the addition

0 20 40 60 80 100 120 140 160

3.0

3.2

3.4

3.6

3.8

Ce

ll vo

lta

ge

/ V

Specific Capacity / mAhg-1

C/10

C/8

C/5

C/3

0 20 40 60 80 100 120 140 160

3.0

3.2

3.4

3.6

3.8

Ce

ll vo

lta

ge

/ V

Specfic Capacity / mAhg-1

89 °C

94 °C

98 °C

a

b

Fig. 5 Voltage versus specific capacity profiles of consecutive cycles

of the Li/ P(EO)20LiBOB (CP)0.125/LiFePO4 battery at different rates

(a) and at different temperatures (b)

12 10 8 6 4 2 0

0

20

40

60

80

100

120 83 °C

89 °C

94 °C

Current density / mAcm-2

Sp

ecific

Ca

pa

city/m

Ah

g-1

1/C- rate

0.125 0.50.250.170.1

Fig. 6 Capacity versus 1/C rate at different temperatures of the Li/P

(EO)20LiBOB (CP)0.125/LiFePO4 battery. Legends show the

corresponding temperatures

0 5 10 15 20 25 30 35 40 45 50

0

20

40

60

80

100

120

140

160

Cycle Number

Sp

ecific

Ca

pa

city/ m

Ah

g-1

charging

discharging

Fig. 7 Capacity versus number of cycles of the Li/ P(EO)20LiBOB

(CP)0.125/LiFePO4 battery at 98 °C. C-rate is C/10 corresponding to a

current density of 0.1 mAcm−2

Ionics (2007) 13:281–286 285

of CP. This is another evidence of the practical relevance of

the electrolyte developed in this work and, indirectly, of the

validity of the concept adopted in our laboratory for

preparing them.

Figure 7, which shows the capacity versus cycle number,

evidence the cycling performance of the Li/PEO20LiBOB

(CP)0.125/LiFePO4 battery. With the exception of the first

cycle, the charge–discharge efficiency is over 98%, this

indicating the high reversibility of the electrochemical

process [2]. The low value in the first cycle is typical of

lithium intercalation processes, being associated to a

rearrangement of the hosting structure upon initial uptake

or release of the lithium ions.

Conclusion

In this work, we have shown that the addition of a anion-

trapping compound, such as calyx(6)pyrrole, CP, combined

with the use of a large anion lithium salt, such as lithium bis

(oxalate) borate, LiBOB, gives rise to PEO-based polymer

electrolytes having excellent transport properties. Polymer

electrolytes of the PEO20LiBOB(CP)0.125 composition have

in the 80 °C temperature range a lithium transference

number of about 0.5 (vs the 0.3 value of conventional PEO

systems) and an ionic conductivity of the order of 0.001

Scm−1. These unique properties make the electrolyte of

interest for practical applications. Indeed, a lithium battery

using PEO20LiBOB(CP)0.125 as electrolyte and a lithium

iron phosphate as cathode, has shown promising perfor-

mance in terms of cycle life and charge–discharge

efficiency.

Acknowledgements This work has been carried out with the

financial support of the European Office of Aerospace Research and

Developement, London, UK. One of us, (H.H.S.) wishes to acknowl-

edge the TRIL Fellowship offered by the International Centre of

Theoretical Physics (ICTP), Trieste.

References

1. Scrosati B (2005) The Chemical Record 5:286

2. Scrosati B (ed) (1993) Applications of electroactive polymers.

Chapman & Hall, London

3. Gray F, Armand M. In: Besenhard JO (ed) Handbook of battery

materials. Wiley-VCH, Wenheim

4. Scrosati B, Neat R (1993) In: Scrosati B (ed) Applications of

electroactive polymers. Chapman & Hall, London, p 182

5. Blazejczyk A, Wieczorek W, Kowarski R, Peled E, Golodnisky D,

Scanlon LG, Appetecchi GB, Scrosati B (2004) J Electrochem

Soc 151:A1762

6. Kalita M, Bukat M, Ciosek M, Siekierski M, Rodriguez T,

Greenbaum SG, Kovarski R, Golodnisky D, Peled E, Scrosati B,

Wieckzorek W (2005) Electrochim Acta 50:3942

7. Panero S, Scrosati B, Sumathipala HH, Wieckzorek W (2007) J

Power Sources 167:510

8. Appetecchi GB, Zane D, Scrosati B (2004) J Electrochem Soc

151:A1369

9. Croce F, Settimi L, Scrosati B, Zane D (2006) J New Materials

9:3

10. Bruce PG, Evans J, Vincent CA (1988) Solid State Ionics 28–

30:918

11. Boukamp BA (1986) Solid State Ionics 18:136

12. Boukamp BA (1986) Solid State Ionics 20:31

13. Padhi AK, Nanjundaswarmy KS, Masquelier C, Okada S,

Goodenough JB (1997) J Electrochem Soc 144:1609

14. Prosini PP, Passerini S, Vellone R, Smyrl WH (1988) J Power

Sources 75:73

15. Croce F, Fiory FS, Persi L, Scrosati B (2001) Electrchem Solid

State Lett 4:A121

286 Ionics (2007) 13:281–286