Alternating copolymers of carbon dioxide with glycidyl ethers for novel ion-conductive polymer electrolytes

Article abstract of DOI:10.1016/j.polymer.2010.07.037

To overcome the low ionic conduction of existing poly(ethylene oxide)-based polymer electrolytes, we consider polycarbonates obtained from the copolymerization of CO₂ and epoxy monomers. We synthesized four types of polycarbonates possessing phenyl, n-butyl, t-butyl and methoxyethyl side groups using zinc glutarate, and measured the ionic conductivity of their electrolytes, including 10 mol% of LiTFSI. The electrolyte possessing methoxyethyl side groups had the highest conductivity, of the order of 10^-6 S cm^-1 at room temperature. The activation energy (E_a) for ionic conduction in the polycarbonate electrolytes was estimated from the VTF equation, and the E_a of the electrolyte possessing n-butyl side groups was almost the same with the polyether-based electrolytes. An interesting feature of our study is that the polycarbonate is a unique candidate for ion-conductive polymers because of its flexible and hydrophobic properties.

Full text of DOI:10.1016/j.polymer.2010.07.037

Polymer 51 (2010) 4295e4298
Contents lists available at ScienceDirect
Polymer
journal homepage: www.elsevier.com/locate/polymer
Polymer Communication
Alternating copolymers of carbon dioxide with glycidyl ethers for novel
ion-conductive polymer electrolytes
*
Yoichi Tominaga , Tomoki Shimomura, Mizuki Nakamura
Department of Organic and Polymer Materials Chemistry, Tokyo University of Agriculture and Technology, Koganei, Tokyo 184-8588, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 30 April 2010
Received in revised form
To overcome the low ionic conduction of existing poly(ethylene oxide)-based polymer electrolytes, we
2
consider polycarbonates obtained from the copolymerization of CO and epoxy monomers. We
synthesized four types of polycarbonates possessing phenyl, n-butyl, t-butyl and methoxyethyl side
groups using zinc glutarate, and measured the ionic conductivity of their electrolytes, including 10 mol%
of LiTFSI. The electrolyte possessing methoxyethyl side groups had the highest conductivity, of the order
22 July 2010
Accepted 24 July 2010
Available online 30 July 2010
ꢀ
6
ꢀ1
of 10 S cm at room temperature. The activation energy (E
electrolytes was estimated from the VTF equation, and the E
a
) for ionic conduction in the polycarbonate
of the electrolyte possessing n-butyl side
a
Keywords:
Polymer electrolyte
Ionic conductivity
groups was almost the same with the polyether-based electrolytes. An interesting feature of our study is
that the polycarbonate is a unique candidate for ion-conductive polymers because of its flexible and
hydrophobic properties.
2
CO /epoxide copolymer
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
As one way to improve the conductivity, processing with CO
2
under subcritical and supercritical conditions has been reported for
simple polyether-salt mixtures [8], organically modified ceramics
as polymer electrolytes [9], and polyether-clay composites [10]. Our
previous reports concluded that CO molecules which permeated
2
into the treated sample can promote the dissociation of ions in the
local structure, and increase ionic mobility. In other words, intro-
Polymer electrolytes are soft materials that are notable as ionic
conductors, since they have good safety characteristics (not flam-
mable, no leakage), and their flexibility and light weight are
particularly useful in solid-state lithium-ion secondary batteries
[1,2]. Ionic conduction in poly(ethylene oxide) (PEO)-metal salt
complexes was first discovered in the 1970s [3,4], and there have
since been many studies of PEO-based electrolytes with a reduced
degree of crystallinity and improved salt solubility [5e7]. Unfor-
tunately, these electrolytes have relatively low conductivity, not
exceeding 10 S cm at room temperature. It is well known that
the migration of ions in PEO can be realized by the local motion of
oxyethylene chains in the amorphous region. The local structure
that is crucial in the migration is believed to facilitate cationedipole
interaction [5]. In fact, this interaction sometimes inhibits fast
migration, because of the strong cohesion of cations and dipoles,
duction of CO
the conductivity of polymer electrolytes, so we accordingly focused
on a CO /epoxide copolymer. CO /epoxide alternating copolymer-
2
as a raw material into the base polymer can improve
2
2
ization was first carried out in 1969 [11,12], and today there are
numerous reports and reviews, mainly of the development of
highly active catalysts [13,14] so as to yield the corresponding
polymer efficiently. The copolymerization method is promising not
only for the novel polymerization reaction, but also in view of the
potential carbon source, in environmental terms. Studies involving
polycarbonates have recently been carried out for novel functional
materials, including biodegradable polymers [15], nanocomposites
[16], and liquid crystalline complexes [17]. In the present study we
ꢀ5
ꢀ1
g
which increases the glass transition temperature, T . To overcome
these problems, another candidate for the base matrix is needed,
without an oxyethylene framework.
2
synthesized alternating copolymers of CO with glycidyl ether
type-epoxide monomers, and used the polymers in preparing
ion-conductive polymer electrolytes. Polycarbonate is a good
candidate for novel ion-conductive polymers, because it has
moderate polar groups on the main chain and is flexible, with
*
Corresponding author.
E-mail address: ytominag@cc.tuat.ac.jp (Y. Tominaga).
g
a relatively low T . The polar carbonate groups are likely to dissolve
0
032-3861/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.polymer.2010.07.037

4296
Y. Tominaga et al. / Polymer 51 (2010) 4295e4298
salts without strong cohesion of cations, and carrier ions may
migrate rapidly in the polycarbonates.
2.3. Measurements
Polycarbonate electrolytes were prepared using the simple
casting method. The polycarbonate was dissolved in chloroform
with lithium bis-(trifluoromethane sulfonyl) imide (LiTFSI, donated
from Daiso Co.) at room temperature. LiTFSI was able to dissolve in
chloroform, and the polymer/salt mixed solution was completely
transparent. The LiTFSI content in the electrolyte was chosen to be
10 mol% to a monomer unit of each polycarbonate. The solution
2
. Experimental section
2.1. Preparation of monomers and catalyst
Zinc oxide (ZnO, 99%), glutaric acid (GA, 98%) and CO
2
(99.99%)
ꢁ
were all used as received. Glycidyl ether (GE) monomers possessing
phenyl (Phe, 98%), tert-butyl (tBu, 99%) and n-butyl (nBu, 95%)
groups were purchased and were stored using 4 Å molecular sieves
prior to copolymerization. A GE monomer possessing methoxyethyl
was cast onto the plastic dish and dried under vacuum at 60 C for
24 h. Differential scanning calorimetry (DSC) measurements of all
ꢁ
samples were made using a DSC120 (Seiko Inst.) from ꢀ100 C to
ꢁ
ꢁ
ꢀ1
2
300 C at a heating rate of 10 C min under dry N gas. The ionic
(
MeEt) group was synthesized from epichlorohydrin and 2-
conductivities of all electrolytes were measured by the complex
impedance method, using an impedance/gain-phase analyzer
4194A (HP) in the frequency range from 100 Hz to 15 MHz. The
methoxyethanol in the presence of NaOH. As polymerization
catalyst, zinc glutarate (ZnGA) was synthesized from ZnO and GA
ꢁ
[18,19]. GA (0.99 mol) was dissolved in toluene (90 mL) in a flask
temperature was reduced from 100 to 30 C and the cell was held
ꢁ
ꢁ
equipped with a DeaneStark trap with a reflex condenser and
a drying tube. ZnO (1.00 mol) was added as a fine powder into the
solution and stirred vigorously at 55 C for 4 h, and the solution was
constant at 10 C or 20 C intervals for at least 30 min, after which
each impedance measurement was carried out.
ꢁ
then refluxed for 24 h. After cooling to room temperature, the
mixture was filtered, washed three times with acetone, and dried
under vacuum at 120 C.
3. Results and discussion
ꢁ
3.1. Copolymerization of CO with glycidyl ethers
2
The four polycarbonates were obtained as high molecular weight
polymers (Table 1), but they differed in color and stiffness. P(Phe-
GEC), P(nBu-GEC) and P(MeEt-GEC) were transparent polymers, and
P(nBu-GEC) and P(MeEt-GEC) were jellylike rubbers and were much
softer than P(Phe-GEC) at room temperature. P(tBu-GEC) was
awhite fibrous solid. These electrolyteswith 10 mol% LiTFSI included
were all slightly opaque and were rubbery solids; no precipitation of
the salt was observed. NMR measurements do not showany peaks in
2
.2. Copolymerization and characterization
Alternating copolymerization of CO with GE monomer was
2
undertaken in a stainless reactor (Taiatsu Techno Co.). The GE
monomer was added with ZnGA (appl. 5 mol% to monomer) to the
reactor in a dry Ar-filled glove box. The reaction conditions were
fixed at 8.2 MPa and 60 C for 24 h. In the case of polymerization
using MeEt-GE monomer, the conditions were 5.0 MPa and 60
ꢁ
ꢁ
1
C
the range 3.4e3.9 ppm in the H NMR spectra of the original poly-
for 7 days. The polymerization process is summarized in Scheme 1.
After polymerization the reactor was cooled to room temperature,
and the resulting mixture was dissolved in chloroform. The chlo-
roform solution was filtered in order to remove ZnGA, and was then
concentrated to a proper volume using a rotary evaporator. The
solution was dropped into excess methanol; this dropping process
was carried out at least three times. The precipitated polymer,
carbonates (eCH CHOe main chain) which depend on the glycidyl
ether homopolymer [19]. It follows that all polycarbonates are
alternating copolymers of CO2 and GE monomer. In addition, no
NMR peaks of unreacted GE monomers were observed, only
polycarbonate.
2
3.2. Thermal analysis
ꢁ
which is abbreviated as P(R-GEC), was dried under vacuum at 60 C
for 24 h. The 1H and C NMR spectra of all of the synthesized
polycarbonates (see Figs. S1eS4 of Electronic Supplementary
Material) were observed using a JEOL EX-400. Molecular weights
and polydispersities of polycarbonates were estimated using a gel
permeation chromatography (GPC) system (JASCO Co.), with two
columns (TOSOH TSKgel GMHHR-H) and chloroform (HPLC grade) as
13
Fig. 1 shows DSC curves of the original polycarbonates and
the electrolytes. The values of T for all samples are summarized in
g
Table 1. For the original polymers, the values of T were significantly
g
different because of the structure of their side groups (denoted by
ꢁ
R in Scheme 1). For P(nBu-GEC), Tg was 33 C lower than for P
(tBu-GEC). This is due to the difference in mobility of the side
groups, n- and tert-butyl, in the polycarbonates. Steric hindrance of
the tert-butyl group should be very different from that of the
n-butyl group, even though these groups have the same formula
weight. For P(tBu-GEC) there was a very weak glass transition,
because of coexistence in small amounts of crystalline domains,
ꢀ
1
an eluent at a flow rate of 1.0 mL min (calibrated by polystyrene
standards).
CH
2
O
CH
2
O
O
ꢁ
which are related to the endothermic peak at 141 C. P(Phe-GEC) is
HC
ZnGA (cat.)
CH
CH2
O
C
O
a glassy polymer because of its rigid side groups, with the highest T
of all the polymers. P(MeEt-GEC) had the lowest T value of all the
g
CH2
CO
2
g
O
Table 1
R
x
R
Characterization of synthesized polycarbonates.
P(R-GEC)
Yielda
M
n
ꢂ 10
4
M
w
/M
n
T
g
( C)
ꢁ
’ ( C)b
ꢁ
T
g
CH3
R ¼ Phe
tBu
nBu
0.44
1.21
1.35
0.30
1.5
4.4
19
1.7
4.5
1.5
2.3
2.1
45
9
41
18
R=
,
C
CH
3
,
CH CH CH CH
,
CH CH OCH
3
2
2
2
3
2
2
ꢀ24
ꢀ20
CH3
MeEt
ꢀ45
ꢀ55
a
Scheme 1. Copolymerization of CO
2
with GE monomers (side groups R ¼ Phe, tBu, nBu
Yield (g/g of cat.) of P(R-GEC) copolymers insoluble in methanol.
T of polycarbonate/LiTFSI (10 mol%) electrolytes.
g
b
and MeEt).

Y. Tominaga et al. / Polymer 51 (2010) 4295e4298
4297
-
-
-
-
-
-
-
3
4
5
6
7
8
9
R= MeEt
nBu
Phe
tBu
2
.6
2.8
3.0
000 T-1 (K-1)
3.2
3.4
1
Fig. 2. Temperature dependence of ionic conductivity for P(R-GEC)10LiTFSI electro-
lytes. (side groups R ¼ Phe, tBu, nBu and MeEt).
activation energy for ionic conduction, between the polycarbonate
and PMEO. The temperature dependence of amorphous polymer
electrolytes is known to follow the VTF empirical equation,
1=2
s
¼ A=T $exp½ ꢀ E
a
=RðT ꢀ T Þꢃ
(1)
0
Fig. 1. DSC curves of P(R-GEC) originals (dashed lines) and P(R-GEC)10LiTFSI electro-
lytes (solid lines). (side groups R ¼ Phe, tBu, nBu and MeEt).
where A (K1/2 S cm ) is a constant that is proportional to the
ꢀ1
ꢀ1
ꢀ1
K ) is the
number of carrier ions, T
fundamental gas constant and E
for ionic transport via segmental motion [7]. We took T
lower than the value of T
0
is an ideal T
g
, R (J mol
original polymers, but the transition was also very weak (see Fig. S5
of Electronic Supplementary Material). For P(nBu-GEC) and P(tBu-
GEC), however, T clearly increased upon addition of LiTFSI. Morover
g
the small endothermic peak of P(tBu-GEC) disappeared, and these
ꢀ1
a
(kJ mol ) is an activation energy
ꢁ
0
to be 50 C
g
obtained from the DSC measurement
[
22]. Equation (1) can be rewritten by taking logarithms as
electrolytes became amorphous without any thermal peaks above
T
g
. These results indicate that the increase in T
g
is due to increased
ꢀ
$T1=2ꢁ
cross-linking structures in amorphous regions, based on the weak
interaction between cations or aggregated ions and polar groups in
the polycarbonate. In other words, the added salt can dissolve, and
the dissociated ions act as carrier ions in the polymer. Based on our
study of the polar groups in polycarbonate, we believe that both
ether oxygens of the side chain and the carbonate units of the main
ln
s
¼ ꢀE
a
=RðT ꢀ T
0
Þ þ lnA
(2)
The parameters A and E
the gradient of each linear plot based on equation (2). VTF plots for P
nBu-GEC), P(MeEt-GEC) and PMEO [21] electrolytes including LiTFSI
are shown in Fig. 3. It is clear that the parameter A was lower for P
a
can be estimated from the intercept and
(
1
/2
ꢀ1
1/
g
chain are involved in the interaction. The T values of P(Phe-GEC)
(
nBu-GEC)10LiTFSI (11.5 K S cm ) than for PMEO10LiTFSI (16.1 K
2
ꢀ1
and P(MeEt-GEC) changed in opposite senses upon addition of
LiTFSI. The decrease for P(Phe-GEC) and P(MeEt-GEC) may be due to
the plasticization effect of dissociated TFSI anions [20].
S cm ). This is because the non-polar (n-butyl) side group inhibits
the smooth dissociation of ions. However, the value of E for P(nBu-
GEC)10LiTFSI (11.6 kJ mol ) was almost the same as that for
a
ꢀ
1
ꢀ1
PMEO10LiTFSI (11.4 kJ mol ). This indicates thatthere is no difference
inthe potentialenergy for ionic transportinP(nBu-GEC)and inPMEO,
3.3. Impedance measurement
The temperature dependence of the conductivity of poly-
-2
carbonate/LiTFSI electrolytes is shown in Fig. 2. All polycarbonates
gave typical Arrhenius plots, similar to polyether-based amorphous
electrolytes, which are convex throughout the entire range of
measurement temperature. The conductivity of P(Phe-GEC)/LiTFSI
was the lowest of all polycarbonate electrolytes, approximately
PMEO10LiTFSI
-
-
-
4
6
8
(T : -51 oC)
g
ꢀ8 ꢀ1
ꢁ
1
0
S cm at 80 C, because its T was highest. The P(tBu-GEC)
g
ꢀ9
ꢀ1
electrolyte also had very lowconductivity, of the orderof 10 S cm
ꢁ
at 40 C. The P(nBu-GEC) electrolyte hadgoodconductivity relative to
R= nBu
P(Phe-GEC) and P(tBu-GEC) electrolytes, because its T
Moreover, the P(MeEt-GEC) electrolyte had the greatest conductivity
g
is low.
-
10
ꢀ
6
ꢀ1
ꢁ
MeEt
(
2.2ꢂ10 S cm at30 C)ofallpolycarbonate samples, anditsvalue
was only 10-times lower than typical polyether-based electrolytes
such as poly[oligo(oxyethylene glycol) methacrylate] (PMEO)/LiTFSI
-12
5
6
1
7
8
-1
9
10
ꢀ
5
ꢀ1
ꢁ
(
2.4 ꢂ 10 S cm at 30 C) [21]. The Arrhenius plots of P(nBu-GEC)
and P(MeEt-GEC) electrolytes in Fig. 2 have smaller gradient than for
P(tBu-GEC) and P(Phe-GEC), and have similar gradient to PMEO10
LiTFSI. This suggests that there is almost no difference, in the
-1
000 (T-T0) (K )
-
Fig. 3. VTF plots of P(R-GEC)10LiTFSI (side groups R ¼ MeEt and nBu) and PMEO10LiTFSI
electrolytes. (RMS of all plots >0.99).

4298
Y. Tominaga et al. / Polymer 51 (2010) 4295e4298
showing that carrier ions can migrate even in the polycarbonate.
Shriver has argued that transport in a rigid polymer electrolyte pos-
sessing a carbonate unit is decoupled from the segmental motion in
References
[
[
1] Tarascon JM, Armand M. Nature 2001;414:359.
2] Scrosati B, Schalkwijk W. Advances in lithium-ion batteries. New York:
Plenum; 2002.
a
the polymer [23]. We believe that the decrease in the E of P(nBu-
GEC)10LiTFSI observed in the present study is a result of a similar
decoupling due to the weak ionedipole interaction. The dependence
for P(MeEt-GEC)10LiTFSI also followed the VTF equation, but the
[3] Fenton DE, Parker JM, Wright PV. Polymer 1973;14:589.
[
[
[
4] Wright PV. Brit Polym J 1975;7:319.
5] Ratner MA, Shriver DF. Chem Rev 1988;88:109.
6] Gray FM. Solid polymer electrolytes fundamentals and technological
applications. New York: VCH; 1991.
1
/2
ꢀ1
,
parameters were significantly different (A ¼ 109 K
S cm
¼ 16.7 kJ mol ). These unusual data are perhaps due to the
decrease inT caused by the addition of LiTFSI. We consider that ionic
ꢀ
1
[
[
7] Bruce PG. Solid state electrochemistry. Cambridge: Cambridge Univ.
Press; 1995.
8] Tominaga Y, Izumi Y, Kwak GH, Asai S, Sumita M. Macromolecules 2003;36:
8766.
E
a
g
migration in P(MeEt-GEC) is due not to the segmental motion of the
main chain but mainly to the fast local motion of short side chains,
such as g-relaxation. A study of this is under way, and other side
[9] Di Noto V, Vezzu K, Pace G, Vittadello M, Bertucco A. Electrochim Acta 2005;
0:3904.
5
[
[
10] Kitajima S, Tominaga Y. Macromolecules 2009;42:5422.
11] Inoue S, Koinuma H, Tsuruta TJ. Polym Chem Polym Lett 1969;7:287.
a
groups can be introduced to the polymer with high A and low E so as
to increase the conductivity.
[12] Inoue S, Koinuma H, Tsuruta T. Makromol Chem 1969;130:210.
[
[
[
13] Sugimoto Y, Inoue SJ. Polym Sci Part A Polym Chem 2004;42:5561.
14] Darensbourg DJ. Chem Rev 2007;107:2388.
15] qukaszczyk J, Jaszcz K, Kuran W, Listos T. Macromol Rapid Commun
Acknowledgement
2000;21:754.
[
[
16] Tan CS, Juan CC, Kuo TW. Polymer 1805;2004:45.
17] Yu T, Zhou Y, Liu K, Chen E, Wang D, Wang F. Macromolecules 2008;41:
This work was supported by Grant-in-Aid for Young Scientists (B)
No. 21750113)fromthe MinistryofEducation, Culture, Sport,Science
(
3175.
and Technology, Japan. One of authors (Y.T.) acknowledges financial
supports from the Sumitomo Foundation in 2008 and the Ogasawara
Foundation for Promotion of Science and Engineering in 2008.
[18] Motika SA, Pickering TL, Rokicki A, Stein BK. US Pat; 1991:5026676.
[
[
[
[
19] Ree M, Bae JY, Jung JH, Shin TJ. J Polym Sci Part A Polym Chem 1999;37:
863.
20] Armand M, Gorecki W, Andreani R. Proc. 2nd ISPE. London: Elsevier;
990. pp. 31.
21] Tominaga Y, Hirahara S, Asai S, SumitaPolym MJ. Sci Part B Polym Phys 2005;
3:5561.
1
1
Appendix. Supplementary data
4
22] Di Noto V, Vittadello M, Lavina S, Fauri M, Biscazzo SJ. Phys Chem B 2001;105:
4584.
Supplementary data associated with this article can be found in
the on-line version, at doi:10.1016/j.polymer.2010.07.037.
[23] Wei X, Shriver DF. Chem Mater 1998;10:2307.