Potential of polymethacrylate pseudo crown ethers as solid state polymer electrolytes

Article abstract of DOI:10.1039/c7cc02385e

The association of kinetic studies, DFT calculations and ¹H-⁷Li NMR analyses allowed the control of the cyclo-ATRP of PEG₉DMA and the production of polymethacrylate pseudo crown-ethers of various molar masses. Their potential to act as a solid-state polymer electrolyte in Li-ion batteries has been highlighted and may come from the supramolecular organization of the cyclo-PEG forming a Li⁺ diffusion channel.

Full text of DOI:10.1039/c7cc02385e

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
Potential of polymethacrylate pseudo crown
ethers as solid state polymer electrolytes†
Cite this: Chem. Commun., 2017,
53, 6899
b
d
S. Moins,a J. C. Martins,
A. Krumpmann,c V. Lemaur,
J. Cornil,d N. Delbosc,a
a
e
a
A. Decroly,c Ph. Dubois,
R. Lazzaroni,d J.-F. Gohy
and O. Coulembier
*
Received 28th March 2017,
Accepted 5th June 2017
DOI: 10.1039/c7cc02385e
rsc.li/chemcomm
The association of kinetic studies, DFT calculations and 1H–7Li NMR experimental studies indicate that the Li ion diffusion results
analyses allowed the control of the cyclo-ATRP of PEG9DMA from a subtle interplay between the PEO chain dynamics and
and the production of polymethacrylate pseudo crown-ethers of the Coulomb interactions between cation–anion, cation–chain,
various molar masses. Their potential to act as a solid-state polymer and anion–chain.6 The Li+ migration is typically described via
electrolyte in Li-ion batteries has been highlighted and may come a hopping mechanism in which the charge hops between
from the supramolecular organization of the cyclo-PEG forming a trapping centers with the help of the segmental dynamics in
Li+ diffusion channel.
the amorphous phase of the PEO-based complex.7 Interestingly,
ionic conductivities have also been measured in crystalline
Lithium-ion batteries (LIBs) operate by transferring electrons polymer matrices yielding a directional motion of Li+ ions well
from a lithium insertion compound to a host intercalation separated from the anionic counterparts.8
material upon charging, with a concomitant migration of Li
Macrocyclic polyethers, or crown-ethers (CEs), are known to
ions between the two electrodes via an electrolytic medium, and solvate cations by expelling the associated anions out of the
vice versa for discharging.1 The nature of the electrolyte is a coordination sphere. In 1981, Newman already raised the
critical component to improve the LIB performances. Solid possibility of using CEs to improve the lithium conductivity,
polymer electrolytes (SPEs) are becoming increasingly attractive stating that weakly associated CE/Li+XÀ complexes may exhibit
for LIBs due to their excellent properties such as safety and fast cation conduction.9 Moreover, the ability of the crown ether
flexibility.2 Since Wright’s discovery of ionic conductivity in rings to stack, forming diffusion channels, may considerably
poly(ethylene oxide) (PEO) complexes with alkali metal salts3,4 improve the cation motion rendering CEs even more attractive
and Armand’s suggestion of a PEO conductor as a material for the development of ionic conductive materials.9,10 The
of interest for the development of electrochemical devices,5 limited research made on poly(CE) points to very poor ion-
polymer electrolytes have been extensively investigated as conductivities (s) at r.t. (10À10 S cmÀ1 o s o 10À8 S cmÀ1 11
)
promising candidates to prepare thinner, lighter and safer which may arise from the architecture of the poly(CE), with
LIBs. Optimizing the lithium diffusion into a polymer matrix either an inappropriate CE cavity or an inherent incapacity for
requires a detailed description at the microscopic level of the CEs to form stable diffusive channels. Considering that the
the structure of the polymer–ion complexes. Theoretical and inability of CE rings to self-assemble might come from the presence
of the flexible arm linking the CEs to the poly(meth)acrylate
a Laboratory of Polymeric and Composite Materials, Center of Innovation and
Research in Materials and Polymers (CIRMAP), University of Mons (UMons),
Place du Parc 23, 7000 Mons, Belgium. E-mail: olivier.coulembier@umons.ac.be
b NMR and Structure Analysis Unit, Department of Organic and Macromolecular
Chemistry, Faculty of Sciences, Ghent University, 9000 Ghent, Belgium
c Research Center in Materials Engineering, Department of Materials Science,
University of Mons (UMons), Place du Parc 23, 7000 Mons, Belgium
backbone, a series of ‘‘arm-free’’ polymethacrylate pseudo crown
ethers has been prepared to favor the formation of ion-channels
by stacking of macrocycles.10 To that end, we revisit here the
cation template-assisted polymerization method developed
by Sawamoto12 by adapting it to a controlled radical cyclo-
polymerization strategy (Scheme 1).
In order to promote an efficient diffusion of the lithium cation
inside the channels formed by the association of the adjacent
crown-ether macrostructures, the size of the poly(ethylene glycol)
dimethacrylate (PEGDMA) used as a macromonomer is of primary
importance.13 Accordingly, PEG9DMA (carrying 9 –CH2CH2O–
repeating units) has been selected since the crown-ether
d Laboratory for Chemistry of Novel Materials, Center of Innovation and Research
in Materials and Polymers (CIRMAP), University of Mons (UMons),
Place du Parc 23, 7000 Mons, Belgium
e
´
BSMA/IMCN, Universite catholique de Louvain, Place Louis Pasteur,
1 1348 Louvain-la-Neuve, Belgium
† Electronic supplementary information (ESI) available: Materials, techniques,
NMR and DSC analyses. See DOI: 10.1039/c7cc02385e
This journal is ©The Royal Society of Chemistry 2017
Chem. Commun., 2017, 53, 6899--6902 | 6899

View Article Online
Communication
ChemComm
Scheme 1 Cation template-assisted cyclopolymerization of poly(ethylene
glycol) dimethacrylate (the LiTFSI salt is omitted in the final polymer
structure).
generated allows an electrostatic interaction with Li+ but pre-
sents a low efficiency of complexation.13 Prior to the cation
template-assisted cyclopolymerization, the ability of the mono-
mer to interact with the lithium cation was evaluated by 1H and
7Li NMR using bis(trifluoromethane)sulfonimide lithium
(LiTFSI). Acetonitrile (ACN) was selected due to its high dielec-
Fig. 1 Influence of the LiTFSI initial content (here in equivalents with
respect to [PEG9DMA]0) on conversion (top) and molar mass (bottom).
Dispersities of the crude final polymers are shown in parentheses.
tric constant14 and its ability to dissolve electrolyte salts. Using other parameters. The targeted degree of polymerization (DP =
the residual ACN signal as an internal reference, a single set [PEG9DMA]0/[EBi]0), representing the number of cyclo-PEG
of resonances is observed upon the addition of LiTFSI, with the chains per polymethacrylate backbone, was initially set at 10.
1H chemical shifts of PEG9DMA moving all downfield with The resulting molecular weight and conversion data indicate
respect to their original positions as the LiTFSI:PEG9DMA that the salt only slightly affects the kinetics, but it has a
ratio is increased from 0.1 to 2.0 (Fig. S2 in the ESI†). The tremendous impact on the control of the reaction (Fig. 1).
1H resonances shift to different extents, ranging from 0.003 to Using a default of LiTFSI clearly leads to cross-linking reactions
0.093 ppm. Simultaneously, the 7Li resonance moves downfield (gelation) which increase considerably both the molar mass and
by 0.48 ppm in the same titration range (Fig. S3, ESI†). Within the dispersity (2 o ÐM = Mw/Mn o 6.8). By using a minimum of
this range, the chemical shift changes appear to vary linearly as 1 eq. of lithium salt per PEG9DMA, the experimental molar
a function of the LiTFSI : PEG9DMA ratio. The presence of a mass (determined by SEC) increases proportionally to the
direct interaction between the lithium ion and the PEG chain degree of conversion and the polymerization proceeds homo-
was demonstrated by heteronuclear NOE spectroscopy (HOESY) genously with no gelation. When 1 eq. of lithium salt is used,
on the system with a 2.0 LiTFSI : PEG9DMA ratio (Fig. S1, ESI†). the relatively good control of the process is deduced from the
In this spectrum, clear 1H–7Li NOE contacts can be found fact that the experimental molar mass (MnSEC = 4300 g molÀ1
,
between the lithium ion and the various CH2 groups of the ÐM = 1.62) is in close agreement with the theoretical one
PEG chain but not with the methacrylate signals, thus indicat- (conv.B 0.7, Mnth = 4100 g molÀ1).
ing a proximity only with the former. Taken together, these data
Interestingly, the control of the process is refined for a
reveal a dynamic interaction around the equilibrium character- [LiTFSI]0-to-[PEG9DMA]0 initial ratio of 2, as attested by the
ized by a fast exchange on the 1H and 7Li NMR time scales decrease in the final dispersity (ÐM B 1.5). To better understand
whereby lithium ions selectively interact with the crown ether the role of the amount of LiTFSI on the dispersity, molecular
part of the PEG9DMA macromonomer.
dynamics simulations (see the ESI,† for details) have been per-
The cyclopolymerization of PEG9DMA has been realized by formed on two extreme cases, i.e., a PEG9DMA macromonomer
atom transfer radical polymerization (ATRP) in ACN at 70 1C in the absence or presence of 2 lithium cations.
using ethyl a-bromoisobutyrate (EBi) as the initiator and a
catalytic complex of CuBr and HMTETA ([EBi]0 = [CuBr]0
Analysis of the trajectories of molecular dynamics runs at
room temperature (Fig. 2) reveals that, in the absence of the
=
2[HMTETA]0). Polymerizations were performed for a [PEG9DMA]0 salt, the monomer molecule is quite flexible, as demonstrated
of ca. 0.15 M in the presence of LiTFSI. To assess the impact of by the large fluctuations (from 4 Å to 36 Å) of the distance
the lithium salt content on both control and kinetics, a series of between the reactive methacrylate groups, i.e., between the
reactions was performed by varying the initial amount of LiTFSI methylene carbons of the methacrylate functions. Introducing
(0 r [LiTFSI]0/[PEG9DMA]0 r 2) while keeping constant all two equivalents of lithium cations induces strong changes in
6900 | Chem. Commun., 2017, 53, 6899--6902
This journal is ©The Royal Society of Chemistry 2017

View Article Online
ChemComm
Communication
It is generally agreed that ion-transport above the glass
transition temperature (Tg) is dependent on the segmental
motion of polymer chains. To maximize the ionic conductivity,
the polymer hosts should be flexible and amorphous and
should present a low Tg value.15 Unlike P(CE)–salt mixtures,11
the addition of LiTFSI on P(cycloPEG9DMA) ([EO]0/[Li+]0 = 20) is
accompanied by a variation of the Tg value depending on length
of the polymer (Table 1 and Fig. S5–S10, ESI†). Generally, the
incorporation of a lithium salt into a PEG-based macrostruc-
ture is accompanied by an increase of the Tg value due to the
good complexation of the cation by the ethylene oxide units and
the increase of rigidity. In the present case, it seems that the
separation of both the lithium cation and the TFSI anion (by
the effective complexation of the Li+ by the pseudo crown ether)
is compensated by a plasticization of the mixture from the TFSI
anions. If the plasticization is important for a low molar mass
P(cycloPEG9DMA), its effect is less and less pronounced for
higher molar mass structures. Both the low Tg values and the
absence of crystallinity prompted us to assess the conductive
character of the three samples. Fig. 3a shows the Nyquist plot
from a representative electrochemical impedance spectroscopy
(EIS) experiment realized on the P(cycloPEG9DMA)8 complex at
293 K; the response is typical of an electrolyte behavior. The
measurements on the three investigated systems at different
Fig. 2 Distribution profiles of the distance between the two reactive
groups of a PEG9DMA molecule, isolated (top) and complexed with
two (bottom) lithium cations together with illustrations of characteristic
structures, as extracted from the molecular dynamics trajectories. The size
of the lithium cations (in purple) has been adapted for clarity.
the three-dimensional structure of the complex. To avoid
electrostatic repulsions between the lithium atoms, the PEG
segment surrounds them through interactions with its ether
1
groups, in agreement with the absence of H–7Li NOE signals.
Interestingly, in these conformations, the reactive methacrylate
end groups get closer to each other (from 4 Å to 20Å) and
therefore should favor the cyclisation processes.
Polymerizations were then realized for higher polymerization
degrees (DP = 50 and 100, [LiTFSI]0/[PEG9DMA]0 B 1.5), followed
by 1H-NMR and SEC analyses. Conversions were limited to circa
70–75% to prevent deleterious reactions. The number-average
molecular weights increased with the targeted DPs while dis-
persities were kept below 1.7. Table 1 summarizes the molecular
characterizations of samples purified from residual PEG9DMA
(Fig. S4, ESI†).
Table 1 Molecular characterizations of P(cycloPEG9DMA) produced by
Li+ template-assisted cyclopolymerization of PEG9DMA
MnSECb (g molÀ1
)
ÐM
Tg (1C)
Tg-Lid (1C)
a
b
c
Entry
DPexp.
1
2
3
8
35
70
4300
7200
9600
1.5
1.6
1.7
À14
À14
À15
À22
À19
À10
a
b
Determined from the conversion. Determined by SEC analysis in
c
THF at 35 1C with PEO calibration. Measured by DSC analysis on a
pristine polymer. Measured by DSC analysis on the polymer–LiTFSI
d
Fig. 3 (a) Nyquist plot P(cycloPEG9DMA)8 complex at 293 K (equivalent
circuit modeling in the inset). (b) Arrhenius plots of the conductivities.
complex ([EO]0/[Li]0 = 20).
This journal is ©The Royal Society of Chemistry 2017
Chem. Commun., 2017, 53, 6899--6902 | 6901

View Article Online
Communication
ChemComm
temperatures are presented as an Arrhenius plot in Fig. 3b. As Equipements de Calcul Intensif – CECI). J. C. and O. C. are
expected, the ionic conductivities increase as the temperature FNRS research fellows.
increases due to the increased mobility of the polymer chain
segments in the solid polymer electrolytes. As compared
to both DPs 35 and 70, the ionic conductivity of the DP 8
presents a curved evolution. This peculiar behavior (out of the
scope of the present communication) might come from the
Notes and references
1 G. P. Gholam-Abbas Nazri, Lithium Batteries, Science and Technol-
ogy, Springer, 2009.
2 (a) Z. Xue, D. He and X. Xie, J. Mater. Chem. A, 2015, 3, 19218;
viscosity of oligoPEG which follows a Vogel–Tamman–Fulcher
(VTF) model. As reported, oligoPEG of 400 g molÀ1 (quite similar
to the one used here) presents a viscosity-to-temperature
curve characterized by a kink for a 1000/T of 3–3.1 KÀ1
(Fig. S11, ESI†).16
(b) K. L. Mathews, A. M. Budgin, S. Beeram, A. T. Joenathan, B. D.
Stein, U. Werner-Zwanziger, M. Pink, L. A. Baker, W. E. Mahmoud,
J. P. Carini and L. M. Bronstein, J. Mater. Chem. A, 2013, 1, 1108;
(c) J. F. Van Humbeck, M. L. Aubrey, A. Alsbaiee, R. Ameloot,
G. W. Coates, W. R. Dichtel and J. R. Long, Chem. Sci., 2015, 6, 5499.
3 D. E. Fenton, et al., Polymer, 1973, 14, 589.
4 P. V. Wright, Br. Polym. J., 1975, 7, 319.
The measured values are in the range of the conductivities
obtained for efficient solid polymer electrolytes,17 and are, as
expected, higher than the ionic conductivities of the corresponding
doped polymers carrying linear ethylene oxide dandling groups.18
For example, doped P(MMA-r-OEM) presents a conductivity of
2.8 Â 10À8 S cmÀ1 at r.t., which is significantly lower than that
in the present case. To prove that this difference is due to the
organization of the crown-ether functions which are sufficiently
large to welcome the lithium cation and let it diffuse, a compar-
ison with a P(cycloPEG4DMA) complex has been realized. Since the
5 M. Armand, Solid State Ionics, 1983, 9–10(Part 2), 745.
6 O. Borodin, et al., Macromolecules, 1998, 31, 8396.
7 P. Johansson, Polymer, 2001, 42, 4367.
8 (a) A. M. Christie, S. J. Lilley, E. Staunton, Y. G. Andreev and P. G. Bruce,
Nature, 2005, 433, 50; (b) Z. Gadjourova, Y. G. Andreev, D. P. Tunstall and
P. G. Bruce, Nature, 2001, 412, 520; (c) G. S. MacGlashan, Y. G. Andreev
and P. G. Bruce, Nature, 1999, 398, 792; (d) C. Zhang, S. Gamble,
D. Ainsworth, A. M. Z. Slawin, Y. G. Andreev and P. G. Bruce, Nat.
Mater., 2009, 8, 580; (e) S. J. Lilley, Y. G. Andreev and P. G. Bruce, J. Am.
Chem. Soc., 2006, 128, 12036; ( f ) L. Y. Yang, D.-X. Wei, M. Xu, Y.-F. Yao
and Q. Chen, Angew. Chem., Int. Ed., 2014, 53, 3631.
9 D. S. Newman, D. Hazlett and K. F. Mucker, Solid State Ionics, 1981,
3/4, 389.
crown ether cavities of P(cycloPEG4DMA) are much smaller than 10 (a) T. Nakamura, T. Akutawaga, K. Honda, A. E. Underhill, A. T. Coomber
and R. H. Friend, Nature, 1998, 394, 159; (b) C. D. Assouma, A. Crochet,
those of P(cycloPEG9DMA), the lithium cation is theoretically too
much stabilized and should not diffuse so well. To that end,
´ ´
Y. Cheremond, B. Giese and K. M. Fromm, Angew. Chem., Int. Ed., 2013,
52, 4682.
a P(cycloPEG4DMA)15 complex has been prepared, characterized 11 (a) D. Peramunage, J. E. Fernadez and L. H. Garcia-Rubio, Macromolecules,
(Mn = 1600 g molÀ1, ÐM = 1.32, Tg-Li B 0.9 1C) and compared to
1989, 22, 2845; (b) L. Collie, D. Parker, C. Tachon, H. V. S. A. Hubbard,
G. R. Davies, I. M. Ward and S. C. Wellings, Polymer, 1993, 34, 1541.
12 T. Terashima, M. Kawabe, Y. Miyabara, H. Yoda and M. Sawamoto,
P(cycloPEG9DMA) (Fig. S12, ESI†). As expected, the conductivities
are much lower proving indirectly that the supramolecular asso-
ciation of P(cycloPEG9DMA) is much better to produce a diffusion
pipe in which the lithium cation may diffuse. Note however that
Nat. Commun., 2013, 4, 2321.
13 R. Franski, Rapid Commun. Mass Spectrom., 2009, 23, 3488.
14 Organic Solvents: Physical Properties and Methods of Purification, ed. J. A.
Riddick, W. B. Bungh and T. K. Sakano, Willey, New York, 4th edn, 1986.
since the dispersion of micro- or nanoparticles (including SiO2,19 15 W. H. Meyer, Adv. Mater., 1998, 10, 439.
Al2O3,20 TiO2,19 zirconia particles21,22 as well as zeolites23) in PEO
based solid electrolytes is effective to enhance both ionic conduc-
16 T. Y. Wu, B. K. Chen, L. Hao, Y. C. Peng and I. W. Sun, Int. J. Mol.
Sci., 2011, 12, 2598.
17 M. Armand, Annu. Rev. Mater. Res., 1986, 16, 245.
tivity and mechanical properties, this could be a future option to 18 A. V. G. Ruzette, P. P. Soo, D. R. Sadoway and A. M. Mayes,
J. Electrochem. Soc., 2001, 148, A537.
19 S. Ketabi and K. Lian, Electrochim. Acta, 2015, 154, 404.
20 P. A. R. D. Jayathilaka, M. A. K. L. Dissanayake, I. Albinsson and
raise the ambient conductivities.
This work was supported by the European Commission/
Walloon Region (BATWAL & FEDER BIORGEL projects), by
the Interuniversity Attraction Pole program of the Belgian
Federal Science Policy Office (PAI 7/05), and by the Belgian
National Fund for Scientific Research (FNRS, Consortium des
B. E. Mellander, Electrochim. Acta, 2002, 47, 3257.
21 F. Croce, L. Settimi and B. Scrosati, Electrochem. Commun., 2006, 8, 364.
22 F. Croce, S. Sacchetti and B. Scrosati, J. Power Sources, 2006, 161, 560.
23 Y. X. Jiang, J. M. Xu, Q. C. Zhuang, L. Y. Jin and S. G. Sun, J. Solid
State Electrochem., 2008, 12, 353.
6902 | Chem. Commun., 2017, 53, 6899--6902
This journal is ©The Royal Society of Chemistry 2017