Activated CO2 sorption in mesoporous imidazolium-type poly(ionic liquid)-based polyampholytes

Article abstract of DOI:10.1021/cm4009128

In this contribution, we present the formation of mesoporous polyampholyte networks via self-complexation (inter-and intrapolyelectrolyte complexation) of copolymers bearing both the imidazolium cations and the carboxylic acid units. The copolymers were p

Full text of DOI:10.1021/cm4009128

Article
pubs.acs.org/cm
Activated CO2 Sorption in Mesoporous Imidazolium-Type Poly(ionic
liquid)-Based Polyampholytes
Sebastian Soll, Qiang Zhao, Jens Weber,* and Jiayin Yuan*
Department of Colloid Chemistry, Max Planck Institute of Colloids and Interfaces, Research Campus Golm, D-14476 Potsdam,
Germany
S
* Supporting Information
ABSTRACT: In this contribution, we present the formation of
mesoporous polyampholyte networks via self-complexation (inter-
and intrapolyelectrolyte complexation) of copolymers bearing both
the imidazolium cations and the carboxylic acid units. The copolymers
were prepared via straightforward free radical copolymerization of
acrylic acid and vinylimidazolium-based ionic liquid monomers
possessing different alkyl substituents in DMSO at 80 °C. Nitrogen
adsorption measurements and electron microscopy were used to
examine the porous structures. The Brunauer−Emmett−Teller (BET) specific surface areas of the resulting mesoporous
complexes were measured to be up to 260 m2/g and varied in terms of the complexation solvent quality, the copolymer
composition, and the precipitation concentration of copolymers as well as the chemical structure of the employed ionic liquid
monomers. The CO2 sorption behavior of a selected mesoporous polyampholyte was studied in detail. It could be shown that
two processes are effective: adsorption at the external surface and absorption into the polymer matrix. Fourier transform infrared
(FTIR) spectroscopy gave hints that the absorption process comes along with the formation of imidazolium-carboxylates,
presumably via a transient N-heterocyclic carbene intermediate.
KEYWORDS: polyampholyte, poly(ionic liquid), polyelectrolyte complex, CO2 sorption, mesoporous materials, N-heterocyclic carbenes
chemistry, and catalysis. Their application spectrum covers
model systems for electrostatic interactions and structural
organization in proteins, secondary oil recovery, water
desalination, recovery of metal ions, and drag reduction. So
far, the research scope on polyampholytes focuses majorly on
their particular physics in aqueous solutions and the
corresponding theory and simulation.18−20 In comparison, the
solid state properties and function of polyampholytes attract
somehow rare attention, not to mention mesoporous solid
polyampholytes. Introducing mesopores (pore sizes between 2
and 50 nm) would provide sufficiently high surface area and
satisfactory mass/energy transfer at the same time.21,22 So far,
mesoporous poly(ionic-liquid) copolymers, containing alter-
nate charges along their single polymeric chains, i.e.
mesoporous polyampholytes, have not been reported to the
best of our knowledge. From a practical application point of
view, mesoporous polyampholytes represent a logic pathway to
improve and amplify their material functions.
1. INTRODUCTION
Recently poly(ionic liquid)s or polymerized ionic liquids (PILs)
as a subclass of polyelectrolytes have attracted rapidly
increasing attention in the field of polymer and materials
science.1−5 They are commonly prepared by straightforward
polymerization of monomeric ionic liquids (ILs). The presence
of an IL moiety in the polymer repeating unit integrates some
IL properties with the polymeric architecture, therefore
expanding the traditional application spectrum of conventional
polyelectrolytes studied so far. Some application examples of
PILs have already been demonstrated very recently, such as
solid polymeric electrolytes, stationary phases in chromatog-
raphy, carbon precursors, catalysis, etc.6−13
Polyampholytes are charged macromolecules containing both
positive and negative charges along their polymeric chains.14−17
Three different categories can be distinguished in polyampho-
lytes generally: quenched, annealed, and zwitterionic ones. The
first are permanently charged regardless of the solution pH,
while in the second class the net charge and the charge
distribution depend on the pH value of the solution.
Zwitterionic polyampholytes have oppositely charged ion
pairs localized on one pendant substituent. Typical poly-
ampholytes were prepared from charged anionic or cationic
monomers, acidic and basic vinyl monomers, zwitterionic
monomers, or ion-pair comonomers. Since the first study of
synthetic polyampholytes in 1950, they have drawn widespread
attention and interest in the fields of polymer chemistry and
physics, molecular biology, colloid chemistry, coordination
Herein, we describe a facile template-free synthetic route to
mesoporous polyampholytes from PIL-based copolymers as a
single structural building component via self-complexation in
organic solvents, in contrast to common polycation-polyanion
binary systems in aqueous solutions. The polyampholyte
copolymers were prepared via conventional free radical
Received: March 20, 2013
Revised: July 1, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/cm4009128 | Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
Figure 1. (A) The synthetic scheme of polyampholytes P(XVImBry-co-AA) from copolymerization of acrylic acid and four vinylimidazolium-based
1
ionic liquid monomers in DMSO; (B) H NMR of P(CMVImBr1.03-co-AA) in DMSO-d6. The peak at 2.5 ppm is the residue DMSO peak.
Synthesis of Copolymers of Acrylic Acid and Ionic Liquid
Monomers. The copolymers were prepared via conventional free
radical copolymerization of the corresponding ionic liquid monomer
and acrylic acid (AA) in a feeding ratio close to 1.5:1. For the
copolymerization made up of CMVImBr (see Figure 1 for chemical
structure) and AA, several additional feeding molar ratios (5:1, 3.5:1,
2:1, 1:1.5, 1:2, and 1:3) were tested to create a copolymer toolbox with
the same building units yet a series of compositions. Here, the
procedure to prepare the copolymer P(CMVImBr-co-AA) from
CMVImBr and AA at a feeding molar ratio of 1.5:1 is introduced as
an example. 4.05 g (0.019 mol) of CMVImBr, 0.87 mL (0.0127 mol)
of acrylic acid, 100 mg (0.6 mmol) of AIBN, and 50 mL of DMSO
were added in a 100 mL flask. After complete dissolution of the solid,
the mixture was deoxygenated by four cycles of freeze−pump−thaw
process. After being backfilled with argon, the flask was stirred in an oil
bath thermostatted at 80 °C for 15 h. After cooling down, the reaction
mixture was added dropwise into an excess of diethyl ether. The
precipitate was redissolved in DMSO and precipitated again in
acetone. After filtering off, the product was dried under vacuum at 60
°C till constant weight. The chemical structures were confirmed by 1H
NMR spectra in Figure 1. The composition was determined by
elemental analysis.
polymerization of acrylic acid with IL monomers of different
structure types in DMSO at 80 °C. The as-synthesized
polyampholytes formed a solid mesoporous matrix via inter/
intrapolyelectrolyte complexation when precipitated in ammo-
nia-containing organic solvents. The process is triggered by in
situ deprotonation of the carboxylic acid groups in the same
copolymer chain. The copolymer composition, the concen-
tration of polymer solutions, and the quality of the complex-
ation solvents (defined as different organic solvents in which
complexation took place) were systematically varied to reveal
the influence of the experimental parameters on the pore
structures. It should be mentioned that many porous PIL-based
materials reported to date are macroporous (pore size >50 nm)
with specific surface area up to 37 m2 g−1.23−25 Mesoporous
PIL-based systems with specific surface area up to 310 m2 g−1
were reported only recently via a hard-templating synthetic
approach or via an interpolyelectrolyte complex mecha-
nism.26−28 Mesoporous materials based on layer-by-layer
assembly of polyelectrolytes have been reported before as
well, but templates were necessarily required in order to
imprint substantial mesoporosity in most cases.29,30
Additionally and also within the focus of the present paper,
there has been a large interest in the CO2 adsorption properties
of (polymerized) ionic liquids, mostly in the bulk.9 The present
system combines a straightforward synthesis with high surface
areas in the final product, which makes it indeed a suitable
candidate for CO2 utilization. Previous reports have shown that,
although not competitive in terms of capacity, PILs seem to
have some kind of intrinsic CO2-philicity. This feature makes
them interesting as it provides some added benefit over purely
adsorptive systems such as activated carbons. Yet, the
adsorption properties are not fully understood, and it is
hence the aim of the present paper to provide insight into the
finer details of CO2 sorption in PIL-based porous materials.
Anion Exchange of the Synthesized Copolymer. In a typical
anion exchange process (taking Tf2N− as an example), copolymers
with Br− anion (10 g/L) were dissolved in deionized water. An
aqueous solution of LiTf2N (50 g/L) was added dropwise into the
copolymer solution to reach a molar ratio of [Tf2N]/[Br] = 1.15. After
addition, the stirring continued for 3 h. The solid precipitate was
separated by filtration and washed several times with deionized water
and dried at 90 °C by high vacuum overnight.
Preparation of Mesoporous Polyampholytes. In a typical run,
P(CMVImBr-co-AA) was dissolved in DMSO solvent at a designed
concentration at room temperature to form a transparent solution.
Thirty mL of diethyl ether (complexing solvent) containing 0.5 wt %
of NH3 (prepared by using 3.5 g of 2 M NH3 in isopropanol) was
placed in a glass beaker, into which 2 mL of the copolymer solution
was dropped (speed: 3 mL/min) under a stirring rate of 900 rpm and
sonication (40% sonication amplitude of a Bandelin Electronics
Sonorex RK 100 ultrasonic bath). The sonication was kept for 1 min
after the addition was finished. The upper liquid phase was discarded,
and the solid content was collected, washed by diethyl ether four
times, and dried under vacuum at 50 °C for 12 h.
2. EXPERIMENTAL SECTION
Materials. Lithium bis(trifluoromethanesulfonyl)imide (LiTf2N,
99.95%), potassium hexafluorophosphate (KPF6, 97%), 1-vinyl-
imidazole (99%), bromoethane (98%), bromoacetonitrile (97%), 1-
bromobutane (99%), benzyl bromide (98%) and acrylic acid (AA,
99%) were purchased from Sigma-Aldrich and used without further
purification. 2,2′-Azobis(2-methylpropionitrile) (AIBN, 98%) was
recrystallized from methanol. 1-Ethyl-3-vinylimidazolium bromide
(EVImBr), 1-butyl-3-vinylimidazolium bromide (BMVImBr), 1-
cyanomethy-3-vinylimidazolium bromide (CMVImBr), and 1-benzyl-
3-vinylimidazolium bromide (BzVImBr) were synthesized according
to the previous literature.31−33 All the solvents used in this study were
of analytic grade.
1
Characterization Methods. H nuclear magnetic resonance (1H
NMR) measurements were carried out at room temperature using a
Bruker DPX-400 spectrometer operating at 400.1 MHz. DMSO-d6 and
D2O were used as solvents. Elemental analysis was performed for
carbon, hydrogen, and nitrogen using a Vario EL Elementar. Field
emission scanning electron microscopy (FESEM) observations were
performed on a LEO 1550-Gemini instrument. The samples were
loaded on carbon coated stubs and coated by sputtering a Au/Pd alloy
prior to imaging. X-ray diffraction experiments were done with a
Bruker D8 diffractometer using Cu Kα radiation (λ = 0.154 nm) and a
B
dx.doi.org/10.1021/cm4009128 | Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
Figure 2. Nitrogen adsorption/desorption isotherm (A), pore size distribution curve obtained by QSDFT analysis of the adsorption branch (B), and
SEM image (C) of a mesoporous polyampholyte product prepared from P(CMVImBr1.03-co-AA).
scintillation counter. FT-IR spectra were performed on a BioRad 6000
FT-IR spectrometer; samples were measured in solid state using a
Single Reflection Diamond ATR.
in a protonated form, thus the overall net charge is positive and
governed only by the imidazolium fraction. In contrast in basic
organic media deprotonation of the COOH groups takes place
and the [imidazolium]/[COOH] ratio determines the final
charge of the polyampholytes. In this study we made use of the
unique combination of the strong imidazolium cation and the
pH dependent carboxylate electrolyte in exactly the same
copolymer chain for the preparation of porous polyampholytes.
While the following part will present the synthesis of the
various polymers, only the cyanomethyl system will be
discussed and analyzed in detail within the second part of the
manuscript.
Nitrogen (N2) sorption experiments were performed with a
Quantachrome Quadrasorb machine at liquid nitrogen temperature
(77.4 K), and initial data analysis was performed using the Quadrasorb
5.05 software package (Quantachrome Instruments). The surface area
and pore volume were calculated using the Brunauer−Emmett−Teller
(BET) equation and the Barrett−Joyner−Halenda (BJH) method,
respectively. CO2 adsorption was measured at 273 or 283 K using a
Quantachrome Autosorb-MP1 machine. Different measurement
modes were used with respect to the allowed equilibration times.
Measurements using an equilibration time of 3 min (the time interval,
in which the pressure change must be below the allowed difference,
typically Δp < 0.0008 atm) are designated as “fast” CO2 adsorption.
An equilibration time of 3 min is a commonly used “default” value of
Quantachrome Instruments. We also used the maximum setting of the
equilibration time (99 min). Those measurements are designated as
“slow” measurements. A detailed explanation can be found in the
Supporting Information. High-purity gases were used in all cases. All
samples were degassed at 80 °C for 20 h before gas adsorption
measurements.
The success in the synthesis of each copolymer was
confirmed by elemental analysis and proton nuclear magnetic
resonance spectra (1H NMR, see the SI for full characterization
details). Elemental analysis was conducted here to each
copolymer to access the realistic composition, because the
severe overlapping of the copolymer backbone protons in
poly(acrylic acid) (PAA) and PIL components makes NMR
characterization alone inaccurate to estimate the exact ratio.
After some fine-tuning in the experiment parameters, a close-to
1:1 molar ratio of [imidazolium]/[COOH] was reached in all
four types (Tables S1 and S2). Figure 1B shows the
3. RESULTS AND DISCUSSION
Figure 1A shows the chemical structures and the general
synthetic route to several polyampholyte copolymers, which
contain both cationic vinylimidazolium units and carboxylic
acid groups along each polymer chain. Four ionic liquid (IL)
monomers were chosen in this study to prove the general
concept of mesoporous polyampholytes. They were synthesized
via quaternization of 1-vinylimidazole with corresponding alkyl
bromides. Here, the R group on the vinylimidazolium cation
varies from short cyanomethyl and ethyl chains to long and
large n-butyl and benzyl species.
The copolymers were readily prepared in one step via free
radical copolymerization of acrylic acid and aforementioned IL
monomers in DMSO at 80 °C. For reasons of simplicity, the IL
monomers were denoted as XVImBr, and their copolymers as
P(XVImBry-co-AA), where X = E, CM, B, and Bz refer to the
ethyl, cyanomethyl, n-butyl, and benzyl substituents, while y is
the molar ratio of [imidazolium]/[COOH] in the copolymer
chains.
The use of DMSO as reaction medium is very crucial as it
provides good solubility to the formed copolymers and at the
same time keeps the carboxylic acid groups in an associated
state. In dependence of the solvent quality and the
[imidazolium]/[COOH] ratio, these copolymers could be
either positively or negatively charged, i.e. they are annealed
polyampholytes. For example, in organic solvents or in neutral
or acidic aqueous solutions the acrylic acid units are dominantly
1
characteristic H NMR spectrum of P(CMVImBr1.03-co-AA).
The broad multisignals in the range of 1.0 to 2.7 ppm are
assigned to the backbone protons of both PAA and PIL
components except the methine [−CH2-CH(N)-CH2−] one
close to the nitrogen atom. This proton signal shifts to the low
field as two splitted sets at 3.3 and 4.4 ppm. The [-CH2CN]
methylene protons appear at 5.6 ppm. Signals at 9.0 and 8.5
ppm correspond to the imidazolium ring protons of
[>NCHN<] and [>NCHCHN<], respectively. All in all, the
1H NMR spectra and EA results verify the chemical structure
and composition of the copolymers.
In our attempt to construct the porous polyampholyte
networks, we first tested P(CMVImBr1.03-co-AA) in various
experimental conditions. The copolymer was initially dissolved
at a concentration of 0.9 wt % in DMSO, where the COOH
units are existent in an associated status. At this stage, no or
only weak complexation occurred, and the solution appeared
transparent and homogeneous. This solution was dropwise
added into an excess of diethyl ether containing 0.5 wt % of
ammonia under ultrasound treatment. Under this basic
condition the COOH groups in the copolymer chains were
deprotonated into COO−, which in turn brings to an in situ
ionic complexation between the cationic imidazolium units and
the COO− anions. Note that both the cation and the anion
precursor (the acid form of COOH) are present in the same
C
dx.doi.org/10.1021/cm4009128 | Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
Figure 3. (A) Plot of SBET of mesoporous polyampholytes from P(CMVImBr1.03-co-AA) vs the copolymer concentration in DMSO used for sample
preparation. The dotted red lines are a guide to the eye. (B) Plot of SBET vs the copolymer composition. (C) Plot of SBET of mesoporous
polyampholytes from P(CMVImBr1.03-co-AA) vs solvent polarity index.
copolymer in a quasi-equivalent molar ratio, so a homogeneous
mixing of these two species has been reached at a molecular
level already before the complexation took place. A diffusion
process, which is usually quite important in a common
polycation-polyanion binary interpolyelectrolyte complex sys-
tem, is thus minimized in the current system. The driving force
for the ionic complexation is the entropy gain resulting from
the release of the low molecular ammonium bromide salt, in
which the bromine anion and ammonium cation were
previously associated with the imidazolium cation and the
carboxylate anion in the copolymer chains. The ammonium
bromide salt was however not detected in the product, as the X-
ray diffraction analysis excludes its existence. In our experiment,
the diethyl ether was mixed with 10% of isopropyl alcohol
(from the ammonia solution) and 7% of DMSO (from the
copolymer solution), which seems to dissolve the in situ
formed salt. The complex precipitated still in this solvent
mixture, which was separated and washed 5 times with diethyl
ether before drying in a vacuum oven at 50 °C for 12 h.
Figure 2A shows the nitrogen adsorption/desorption
isotherm of the as-synthesized porous polyampholytes based
on P(CMVImBr1.03-co-AA). It exhibits ambiguously a type-IV
isotherm with a hysteresis loop in the P/P0 range from 0.6 to
0.9. The pore size distribution curve (determined by quenched
solid density functional theory, QSDFT, adsorption branch,
cylindrical pores, Figure 2B) indicates that the porosity of the
material is essentially made up of mesopores having radii from
6 to 12 nm. The specific surface area calculated from
Brunauer−Emmett−Teller equation (SBET) is 260 m2 g−1 and
the pore volume of this material is 0.49 cm3 g−1. It must be
stated that ammonia controlled strictly the complexation
process. Without ammonia activation, the ion pairing of the
imidazolium and carboxylate did not occur, and the product
was essentially nonporous (Figure 2A) and remained soluble in
DMSO.
The pore generation in this approach coincides with a series
of events, i.e. the deprotonation of the COOH groups, followed
by complexation and precipitation. It is not surprising that the
porous structure and the SBET values of the complex are
dependent on the exact preparation conditions, for example,
the copolymer concentration, the copolymer composition, the
complexation solvent, etc. By variation of the P(CMVImBr1.03
-
co-AA) copolymer concentration in DMSO, the SBET of the
resulting material could be tuned with a maximal value of 260
m2 g−1 at a concentration of 0.9 wt % (Figure 3A). The SBET
value is comparable with the previously reported ones for
porous complexes formed between a PIL and PAA.26 Above
this concentration, the SBET value drops smoothly. Never-
theless, the volumetric yield increases, as a concentrated
solution holds more polymers. To study the composition effect,
seven P(CMVImBr-co-AA) copolymers of the [imidazolium]/
[COOH] molar ratios in the range of 3.80 to 0.26 were tested
at a fixed copolymer concentration of 7 wt % in the same
procedure. The SBET values of the porous products were plotted
against the [imidazolium]/[COOH] ratio. As shown in Figure
3B, the highest SBET value peaks at a composition of
[imidazolium]/[COOH] = 1. This is understandable, because
at this optimal condition the population of the cations and
anions is in fact identical, facilitating a maximum ion
complexation. The effect of different complexation solvents is
shown in Figure 3C. Hereby the solvent polarity index could be
considered as an important parameter for the formation of the
mesoporous structure. Though the complex products are
nonporous in polar solvents, like methanol or ethanol, the
SBET value increases in less polar or nonpolar solvents with
decreasing polarity index from acetone (95 m2/g) to chloro-
form (135 m2/g) and THF (130 m2/g) and further to diethyl
ether (230 m2/g). A plateau is reached below a polarity index of
3, a typical value for nonpolar solvents like toluene and hexane.
These are poor solvents for the copolymer, which are beneficial
to develop the mesoporous network and minimize the chain
motion. The chemical structure is another important factor that
affects the pore characteristics. Table 1 summarizes the
influence of the alkyl substituent on the imidazolium ring and
various anions on the SBET values of the formed products. SBET
decreases with increasing steric hindrance of group R (Entries
1−4), when keeping the same anion (Br−). It seems that the
ionic complexation is partially screened when bulky groups are
attached to the imidazolium ring. The highest values are
obtained in the case of the ethyl group, while a longer aliphatic
butyl chain or an aromatic benzyl group leads to a reduced
specific surface area. On the contrary exchange of the
counteranion has little to no effect.
Finally, it should be noted that the preparation of the
mesoporous polyampholytes is also possible in the absence of
ultrasonic irradiation. The porosity is however reduced by
∼15% (see Figure S7 for exemplary results).
An image recorded from the scanning electron microscopy
(SEM) of the porous complex is shown in Figure 2C. Primary
particles of sizes around 50 nm are observed, which seem to be
agglomerated into large species. The observed porosity can
consequently be attributed to the void spaces in-between the
formed nanoparticles. These particles are physically cross-
linked and stabilized by strong electrostatic interaction between
the cationic imidazolium units and the anionic COO− groups in
the copolymer.
D
dx.doi.org/10.1021/cm4009128 | Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
imidazolium and the carboxylate on the adsorption, some
analogous systems such as nonporous P(CMVImBr) and
P(CMVIm-OAc) homopolymers and a mesoporous polyelec-
trolyte complex prepared from P(CMVIm-Tf2N) and PAA
were also investigated (see chemical structures in Scheme
S1).26
Table 1. Specific Surface Area in Dependence of Counter
Anions and Alkyl Substituents in the Polyampholyte
(Prepared via Precipitating a Copolymer Concentration of
1.8 wt % in DMSO into Ammonia-Containing Diethyl
a
Ether)
entry
copolymer
SBET (m2/g)
From previous experiments on mesoporous PILs it is known
that the equilibration time can have an impact on the observed
CO2 uptake. Hence, all adsorption experiments presented in
the following were made using two different measurement
modes of the volumetric adsorption apparatus (see the
Experimental Section). First, the measurements were con-
ducted using the default settings of the machine. Those
measurements are designated as fast measurements in the
following. Second, the equilibration settings were set to the
maximum in order to access a potentially underlying slow
absorption process. Those measurements are designated as slow
measurements in the following.
Figure 4 shows the CO2 adsorption/desorption isotherms of
P(CMVImBr1.03-co-AA) (Entry 2 in Table 1) obtained at 273
and 283 K. The difference between the measurement modes is
quite obvious. The fast adsorption/desorption isotherms do not
show any pronounced hysteresis, while a very strong hysteresis
is found for the slow measurements (Figure 4A). The isosteric
heats of adsorption/desorption of the fast measurements are in
the range of 25−30 kJ mol−1, which is a common range for
physisorption of CO2 at polar surfaces. Hence, it can be stated
that the fast measurement does only represent the CO2
adsorption at the outer surface of the mesoporous poly-
ampholyte material.
1
2
3
4
5
6
P(EVImBr1.04-co-AA)
230
190
185
130
180
195
P(CMVImBr1.03-co-AA)
P(BVImBr0.99-co-AA)
P(BzVImBr0.96-co-AA)
P(CMVIm[PF6]1.03-co-AA)
P(CMVIm[Tf2N]1.03-co-AA)
a
PF6: hexafluorophosphate; Tf2N: bis(trifluoro-methanesulfonyl)-
amide.
Porous polymeric materials as well as IL based materials have
shown high potential in CO2 adsorption and/or utilization
approaches.34,35 Previously, we studied the CO2 adsorption
properties of a mesoporous PIL, which has been prepared by
hard-templating. Some evidence was found for CO2−PIL
interaction beyond pure surface adsorption.27 Hence, there is
strong interest to study the CO2 sorption properties of PIL-
type complexes. CO2 sorption in PILs has been discussed
heavily in the past years, and there is a general agreement on
the fact that PILs seem to have a significant CO2-philicity,
which is just about to be understood.36−38
P(CMVImBr1.03-co-AA) was chosen for a more thorough
investigation of its CO2 adsorption properties. In order to
understand also the effect of the interplay between the
Figure 4. a) CO2 adsorption/desorption isotherms of P(CMVImBr1.03-co-AA) (Entry 2 in Table 1) obtained at different temperatures and
equilibration settings (closed symbols: fast; open symbols: slow); b) isosteric heats of CO2 sorption; c) difference isotherm obtained by subtraction
of the fast isotherm from the slow isotherms; d) CO2 adsorption isotherms of bulk P(CMVIm-OAc) at different conditions (see the SI for isotherm
plots of nCO2 vs pressure).
E
dx.doi.org/10.1021/cm4009128 | Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
In contrast, the slow measurements give a different picture.
The observed hysteresis upon desorption does not close even at
very low pressures, which indicates that it is not due to simple
physisorption at the surface. Furthermore, the total amount
adsorbed at 760 mmHg depends on the measurement
temperature, while the amount of CO2, which could not be
desorbed under the chosen equilibrium conditions, is the same
for both temperatures. This finds its reflection in the isosteric
heats of adsorption/desorption, which were calculated from the
isotherms. The isosteric heat of slow adsorption declines from
20 kJ mol−1 at low coverage to 11 kJ mol−1 at high coverage,
which is significantly (∼5 to 10 kJ mol−1) lower compared to
the fast adsorption process. Such a low enthalpy could be
explained by two overlaying contributions, namely gas binding
and matrix reorganization. Although those contributions cannot
be distinguished experimentally, it was shown theoretically that
both are present upon dissolution of gases in rubbery
polymers.39 Van der Vegt et al. estimated the binding enthalpy
of CO2 in siloxane polymers to ∼−24 kJ mol−1, while the
reorganization enthalpy was calculated to ∼12 kJ mol−1.39
Thus, the overall CO2 heat of adsorption (qst = −ΔHads) is ∼12
kJ mol−1, which is very close to the values calculated here, and
indicates that a true carbon dioxide solubility (including
polymer rearrangements) is observed here. Once the CO2
made its way into the PIL matrix, it seems to be bound rather
strongly, which is evidenced by the fact that both isotherms do
not show closure of the hysteresis.
Interestingly, the heat of slow desorption was very much
comparable to the fast process until the low coverage region is
reached. At ∼4 cm3 g−1 STP the calculated heats showed a
sharp increase, indicative of a strong binding. The determi-
nation of the heat of desorption was finally not possible below
3.5 cm3 g−1 STP as CO2 did not desorb within the given
pressure regime and equilibration settings (CO2 did also not
desorb at 273 K down to pressures of 0.4 mbar within a time
frame of >2 weeks, see the SI). Overall, the data discussed so far
reveal two processes to be present: adsorption at the surface as a
fast process and absorption/chemisorption within the PIL-based
material as the slow process. While adsorption at the surface is
quite well understood, the absorption into the polymeric matrix
is of more interest in the following.
An attempt was made to separate the contributions by
subtraction of the fast isotherm from the slow one (Figure 4c).
To do so, the experimental isotherms were fitted using
Langmuir or dual-site approaches (see the SI), and the
difference isotherm ΔV was calculated by subtracting the
isotherms calculated based on the fitting parameters. The
corresponding isotherms look very much the same at 273 and
283 K. Both are characterized by an induction period at low
pressure followed by a linear uptake of gas, which cannot be
desorbed anymore. The shape of the difference isotherm is
comparable to that of an adsorption/desorption isotherm of a
nonporous (SBET < 10 m2 g−1) P(CMVIm-OAc) polymer
(Figure 4d). This polymer, which has loose acetate counterions
instead of the polymeric acrylic acid carboxylate of the
mesoporous polyampholyte material, does also show a large
desorption hysteresis as well as a slow absorption mode
(evidenced by the difference of the adsorption isotherms
depending on the equilibration settings). Interestingly, the
adsorption power of this polymer was much higher than that of
the corresponding bulk P(CMVIm-Br) polymer. The unusual
adsorption behavior of imidazolium-acetate ionic liquids with
regard to CO2 is well-known and was subject to intense
investigations.40−43 Basically, the acetate counterion can act as a
base and deprotonate the imidazolium ring to form a transient
N-heterocyclic carbene, which can react with CO2 to form a
carboxylate-imidazolium zwitterion (see the upper part of
Scheme 1).
Scheme 1. Supposed Mechanism of CO2 Fixation in
Imidazolium-Carboxylate Based Poly(ionic liquid) Systems,
Note the Necessity of Chain Rearrangements upon CO2
Binding
Usually, high CO2 pressures are used and interest arose,
whether the observed hysteresis could still be due to the above-
described mechanism. A simple ATR-FTIR experiment was
conducted to answer the question. First, the nonporous
polymer P(CMVIm-OAc), which showed the stronger binding,
was degassed under vacuum at 80 °C and subjected to a CO2
atmosphere for ∼20 h. The IR characteristics of the sample
were measured using an ATR-FTIR setup. The spectrum was
dominated by free CO2, although there were already indications
of bound CO2. Accordingly, the sample was shortly evacuated
to remove excess CO2 and subjected again to the ATR-FTIR
measurement. While being kept in the ATR setup, the sample
was purged with N2 flow using a funnel. IR spectra were
recorded after 2 and 4 min of purge. All IR data are presented
in Figure 5 together with the spectra of the degassed and N2
loaded P(CMVIm-OAc).
Free CO2 is known to show two major bands, antisymmetric
stretching (ν3) at roughly 2335 cm−1 and bending (ν2) at
around 669 cm−1.44 The ATR setup used here is not very
convenient for the discussion of the ν3 band and this band is
hence just an indicator, whether free (noninteracting) CO2 is
present or not. In contrast, the ν2 band is sensitive to changes
of the environment, and it was shown previously that CO2,
which undergoes specific interaction with polar groups of
polymers, exhibits two additional bands at lower wavenumbers
as a consequence of the changed symmetry.44 We observed the
presence of ν2 bands, characteristic of interacting CO2 at 650
cm−1 next to the ν2 band of free CO2. A more detailed analysis
is however not possible, as the parent polymer shows IR
absorption in this range as well. Nevertheless, this gives
evidence of CO2 absorbed within the polymer matrix and does
backup the gas adsorption results. The CO2 bands are generally
decreasing in intensity upon purging with N2, which is in
accordance with a physisorption model.
Analysis of the carbonyl region showed a new band at 1654
cm−1, which cannot be related to free CO2 but to CO2 bound
as carboxylate to the imidazolium ring.40 This provides
evidence for the above speculation. Usually, the reaction can
easily be reversed (decarboxylation) by humidity or thermal
decomposition, which is in accordance with the fact that the
polymers can be regenerated by degassing at 80 °C under high
vacuum to allow for full regain of adsorption capacity.
F
dx.doi.org/10.1021/cm4009128 | Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
Figure 5. FTIR spectra of pure and CO2 loaded P(CMVIm-OAc), left-hand side: characteristic region of free CO2 (ν3 band, ATR setup); center:
carbonyl region, the arrow indicated the new band at 1654 cm−1 (cf. ref 40); right-hand side: ν2 region of CO2; the gray line indicates the ν2 band of
CO2 absorbed in the matrix (cf. ref 44).
Figure 6. Left-hand side: characteristic region of free CO2 (ν3 band, ATR setup); right-hand side: detail of the carbonyl region showing the weak
peak at 1653 cm−1 (cf. ref 40).
Figure 6 shows the corresponding ATR-FTIR spectra of
porous P(CMVimBr1.03-co-AA) loaded with CO2 and subjected
to N2 purge together with the spectrum of the pure polymer.
Again, a weak but reproducible peak can be seen at 1653 cm−1.
The peak is of lower intensity compared to the acetate
containing PIL, which is in accordance with the higher CO2
uptake of the acetate homopolymer.
Finally, it is worth noting that a porous poly(ionic liquid)
based material that was prepared via interpolyelectrolyte
complexation using cationic PIL PCMVImTf2N (Tf2N stands
for bis(trifluoro-methanesulfonyl)amide anion) and PAA26
showed a CO2 adsorption/desorption behavior (including
nonclosing hysteresis), which was very much the same as that
of the materials described here (see the SI). This provides
evidence that the underlying CO2 binding/activation process is
indeed due to the molecular characteristics of the ion pair
imidazolium/carboxylate fixed in the polymer matrix and the
associated intermediate NHC.
The CO2 adsorption behavior of such permanently porous
materials prepared from a single copolymer were analyzed in
detail using one of them as a model compound. Next to fast
CO2 adsorption to the surface (which is the predominant
mechanism in typical microporous organic polymers, such as
CMPs, PIMs, etc.), additional absorption into the polymeric
matrix was observed. This process is slow compared to the
surface adsorption and comes with an energetic penalty, which
is most probably due to necessary chain reorganizations. It was
observed that such absorbed CO2 could not be desorbed easily,
which lead us to speculate about chemical fixation of CO2.
Indeed, the analysis of the FTIR spectra of CO2 loaded
polymers provided evidence that the trapped CO2 can be
activated to form imidazolium-carboxylate zwitterions even at
low temperature and pressure. We believe the mechanism to be
comparable with the previously reported formation of transient
N-heterocyclic carbene species within low-molecular ionic
liquids of the imidazolium-carboxylic anion type. The present
system hence might provide a new way to generate
heterogeneous N-heterocyclic carbenes (NHCs)45 and can
find use in a variety of applications such as catalysis. Indeed, the
synthesis of heterogeneous (immobilized) NHCs is a topic of
current interest and various; however, more complex ways have
already been reported.46−48 Due to the fact that the carboxylate
anion is fixed to the imidazolium within the same polymer
chain, we furthermore expect that leaching or ion exchange
problems could be reduced. The possibility of the interaction of
the carboxylate anions with CO2 in nonimidazolium based
systems should also not be discarded in future studies, as there
are additional hints on such interactions.49 They are however
4. CONCLUSIONS
In conclusion, we presented a unique type of mesoporous
polymers. The intra/intermolecular complexation of imidazo-
lium and carboxylic acid groups, which are combined into a
copolymer, leads to a fixed nanoparticulate structure upon
precipitation in basic organic medium. Interstitial voids are left
open as mesopores of 6−12 nm in diameter and specific surface
areas up to 260 m2 g−1 can be achieved. The fine chemical
details of the polymer backbone (ratio imidazolium/COOH) as
well as the conditions of the precipitation process are shown to
control the overall porosity.
G
dx.doi.org/10.1021/cm4009128 | Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
(23) Yan, F.; Texter, J. Angew. Chem., Int. Ed. 2007, 119, 2492.
less explored yet and await more thorough investigation in the
future. Finally, the permanent porosity does finally provide fast
mass transfer, which is an additional benefit.
(24) Yan, F.; Texter, J. Chem. Commun. 2006, 2696.
(25) Huang, J.; Tao, C.-a.; An, Q.; Zhang, W.; Wu, Y.; Li, X.; Shen,
D.; Li, G. Chem. Commun. 2010, 46, 967.
(26) Zhao, Q.; Zhang, P.; Antonietti, M.; Yuan, J. J. Am. Chem. Soc.
2012, 134, 11852.
(27) Wilke, A.; Yuan, J.; Antonietti, M.; Weber, J. ACS Macro Lett.
2012, 1028.
(28) Zhao, Q.; Soll, S.; Antonietti, M.; Yuan, J. Polym. Chem. 2013, 4,
2432−2435.
(29) Wang, Y.; Hosta-Rigau, L.; Lomas, H.; Caruso, F. Phys. Chem.
Chem. Phys. 2011, 13, 4782.
ASSOCIATED CONTENT
* Supporting Information
Additional characterization data (NMR, elemental analysis, gas
adsorption and IR data, data evaluation) are presented. This
material is available free of charge via the Internet at http://
pubs.acs.org.
■
S
(30) Wang, Y.; Yu, A.; Caruso, F. Angew. Chem., Int. Ed. 2005, 44,
2888.
(31) Yuan, J.; Giordano, C.; Antonietti, M. Chem. Mater. 2010, 22,
5003.
AUTHOR INFORMATION
Corresponding Author
*Fax: +49-(0)331-5679502. E-mail: jens.weber@mpikg.mpg.de
(J.W.), jiayin.yuan@mpikg.mpg.de (J.Y.).
■
(32) Marcilla, R.; Alberto Blazquez, J.; Rodriguez, J.; Pomposo, J. A.;
Mecerreyes, D. J. Polym. Sci., Part A: Polym. Chem. 2004, 42, 208.
(33) Dickmeis, M.; Ritter, H. Macromol. Chem. Phys. 2009, 210, 776.
(34) See, for example: (a) Chen, Q.; Luo, M.; Hammershøj, P.;
Zhou, D.; Han, Y.; Laursen, B. W.; Yan, C.-G.; Han, B.-H. J. Am. Chem.
Soc. 2012, 134, 6084. (b) Dawson, R.; Ratvijitvech, T.; Corker, M.;
Laybourn, A.; Khimyak, Y. Z.; Cooper, A. I.; Adams, D. J. Polym. Chem.
2012, 3, 2034. (c) Kiskan, B.; Antonietti, M.; Weber, J. Macromolecules
2012, 45, 1356. (d) Wilke, A.; Weber, J. J. Mater. Chem. 2011, 21,
5226. (e) Du, N.; Park, H. B.; Robertson, G. P.; Dal-Cin, M. M.;
Visser, T.; Scoles, L.; Guiver, M. D. Nat. Mater. 2011, 10, 372. (f) Luo,
Y.; Li, B.; Wang, W.; Wu, K.; Tan, B. Adv. Mater. 2012, 24, 5703.
(35) (a) Bara, J. E.; Hatakeyama, E. S.; Gabriel, C. J.; Zeng, X.;
Lessmann, S.; Gin, D. L.; Noble, R. D. J. Membr. Sci. 2008, 316, 186.
(b) Zhu, X.; Chai, S.; Tian, C.; Fulvio, P. F.; Han, K. S.; Hagaman, E.
W.; Veith, G. M.; Mahurin, S. M.; Brown, S.; Liu, H.; Dai, S. Macromol.
Rapid Commun. 2013, 34, 452. (c) Zhu, X.; Hillesheim, P. C.;
Mahurin, S. M.; Wang, C.; Tian, C.; Brown, S.; Luo, H.; Veith, G. M.;
Han, K. S.; Hagaman, E. W.; Liu, H.; Dai, S. ChemSusChem 2012, 5,
1912.
(36) Tang, J.; Sun, W.; Tang, H.; Radosz, M.; Shen, Y.
Macromolecules 2005, 38, 2037.
(37) Tang, J.; Tang, H.; Sun, W.; Radosz, M.; Shen, Y. Polymer 2005,
46, 12460.
(38) Tang, J.; Tang, H.; Sun, W.; Radosz, M.; Shen, Y. J. Polym. Sci.,
Part A: Polym. Chem. 2005, 43, 5477.
(39) van der Vegt, N. F. A.; Kusuma, V. A.; Freeman, B. D.
Macromolecules 2010, 43, 1473.
(40) Van Ausdall, B. R.; Glass, J. L.; Wiggins, K. M.; Aarif, A. M.;
Louie, J. J. Org. Chem. 2009, 74, 7935.
Author Contributions
The manuscript was written through contributions of all
authors. All authors have given approval to the final version of
the manuscript.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
The authors thank the Max Planck Society for financial support.
■
ABBREVIATIONS
NHC: N-heterocyclic carbene; PILs: polymerized ionic liquids
or poly(ionic liquid)
■
REFERENCES
■
(1) Yuan, J.; Antonietti, M. Polymer 2011, 52, 1469.
(2) Mecerreyes, D. Prog. Polym. Sci. 2011, 36, 1629.
(3) Lu, J.; Yan, F.; Texter, J. Prog. Polym. Sci. 2009, 34, 431.
(4) Green, M. D.; Long, T. E. Polym. Rev. 2009, 49, 291.
(5) Green, O.; Grubjesic, S.; Lee, S.; Firestone, M. A. Polym. Rev.
2009, 49, 339.
(6) Yuan, J.; Schlaad, H.; Giordano, C.; Antonietti, M. Eur. Polym. J.
2011, 47, 772.
(7) Zhao, Q.; Fellinger, T.-P.; Antonietti, M.; Yuan, J. Macromol.
Rapid Commun. 2012, 33, 1149.
(8) Ohno, H. Macromol. Symp. 2007, 249−250, 551.
(9) Carlisle, T. K.; Bara, J. E.; Lafrate, A. L.; Gin, D. L.; Noble, R. D.
J. Membr. Sci. 2010, 359, 37.
(10) Zhao, F.; Meng, Y.; Anderson, J. L. J. Chromatogr., A 2008,
1208, 1.
(11) Lambertus, G. R.; Crank, J. A.; McGuigan, M. E.; Kendler, S.;
Armstrong, D. W.; Sacks, R. D. J. Chromatogr., A 2006, 1135, 230.
(12) Vijayakrishna, K.; Mecerreyes, D.; Gnanou, Y.; Taton, D.
Macromolecules 2009, 42, 5167.
(13) Vijayakrishna, K.; Jewrajka, S. K.; Ruiz, A.; Marcilla, R.;
Pomposo, J. A.; Mecerreyes, D.; Taton, D.; Gnanou, Y. Macromolecules
2008, 41, 6299.
(14) Ciferri, A.; Kudaibergenov, S. Macromol. Rapid Commun. 2007,
28, 1953.
(15) Dobrynin, A. V.; Colby, R. H.; Rubinstein, M. J. Polym. Sci., Part
B: Polym. Phys. 2004, 42, 3513.
(16) Lowe, A. B.; McCormick, C. L. Chem. Rev. 2002, 102, 4177.
(17) Yamakov, V.; Milchev, A.; Limbach, H. J.; Dunweg, B.; Everaers,
R. Phys. Rev. Lett. 2000, 85, 4305.
(18) Paul, G. H.; Jean-Francois, J. J. Chem. Phys. 1991, 94, 1543.
(19) Andrey, V. D.; Michael, R. J. Phys. II 1995, 5, 677.
(20) English, A. E.; Mafe, S.; Manzanares, J. A.; Yu, X.; Grosberg, A.
Y.; Tanaka, T. J. Chem. Phys. 1996, 104, 8713.
(41) Blath, J.; Deubler, N.; Hirth, T.; Schiestel, T. Chem. Eng. J. 2012,
181−182, 152.
(42) Cabaco
̧ , M. I.; Besnard, M.; Danten, Y.; Coutinho, J. A. P. J.
Phys. Chem. A 2012, 116, 1605.
(43) Carvalho, P. J.; Alvarez, V. H.; Schroder, B.; Gil, A. M.;
Marrucho, I. M.; Aznar, M.; Santos, L. M. N. B. F.; Coutinho, J. A. P. J.
́
̈
Phys. Chem. B 2009, 113, 6803.
(44) Kazarian, S. G.; Vincent, M. F.; Bright, F. V.; Liotta, C. L.;
Eckert, C. A. J. Am. Chem. Soc. 1996, 118, 1729.
(45) Fevre, M.; Pinaud, J.; Leteneur, A.; Gnanou, Y.; Vignolle, J.;
Taton, D.; Miqueu, K.; Sotiropoulos, J.-M. J. Am. Chem. Soc. 2012,
134, 6776.
(46) Pinaud, J.; Vignolle, J.; Gnanou, Y.; Taton, D. Macromolecules
2011, 44, 1900.
(47) Thiel, K.; Zehbe, R.; Roeser, J.; Strauch, P.; Enthaler, S.;
Thomas, A. Polym. Chem. 2013, 4, 1848.
(48) Zeitler, K.; Mager, I. Adv. Syn. Catal. 2007, 349, 1851.
(49) (a) Kim, N.-S.; Jeong, S.-K.; Yoon, S.-H.; Park, G.-S. Bull.
Korean Chem. Soc. 2011, 32, 4441. (b) Steckel, J. A. J. Phys. Chem. A
2012, 116, 11643.
̈
(21) Schuth, F. Chem. Mater. 2001, 13, 3184.
̈
(22) Schuth, F.; Schmidt, W. Adv. Eng. Mater. 2002, 4, 269.
̈
H
dx.doi.org/10.1021/cm4009128 | Chem. Mater. XXXX, XXX, XXX−XXX