Microporous polyimides with uniform pores for adsorption and separation of CO2 gas and organic vapors

Article abstract of DOI:10.1021/ma400496q

Three microporous polyimides, MPI-1, MPI-2, and MPI-3, with uniform pores were synthesized via one-pot polycondensation from tetrakis(4-aminophenyl) methane, tris(4-aminophenyl)amine and 1,3,5-tris(4-aminophenyl)benzene with pyromellitic dianhydride, respectively. The amorphous networks exhibit excellent thermal stability, large BET surface areas up to 1454 m² g ^-1, and narrow pore size distribution in the range from 5 to 6 A. Their adsorption-desorption isotherms of CO₂ are reversible, and the CO₂ uptakes at 273 K and 1 bar are up to 16.8 wt %. Moreover, based on the ratios of initial slopes of isotherms, the CO ₂/N₂ and CO₂/CH₄ separation factors are as high as 102 and 12, respectively. The above CO₂ adsorption and separation properties are attributed to the presence of abundant electron-rich heteroatoms in the polyimide networks and the unifrom ultralmicroporous structures. In addition, for MPI-1, the adsorption capacity of benzene vapor is 119.8 wt %, while the separation factors of benzene over nitrogen and water reach 342 and 28, respectively. The outstanding selective adsorptions of CO ₂ gas and benzene vapor endow the microporous polyimides with promising potential in CO₂ capture and separation as well as air- and water-cleaning applications.

Full text of DOI:10.1021/ma400496q

Article
pubs.acs.org/Macromolecules
Microporous Polyimides with Uniform Pores for Adsorption and
Separation of CO2 Gas and Organic Vapors
Guiyang Li and Zhonggang Wang*
State Key Laboratory of Fine Chemicals, Department of Polymer Science and Materials, School of Chemical Engineering, Dalian
University of Technology, Dalian 116024, China
S
* Supporting Information
ABSTRACT: Three microporous polyimides, MPI-1, MPI-2, and MPI-3,
with uniform pores were synthesized via one-pot polycondensation from
tetrakis(4-aminophenyl)methane, tris(4-aminophenyl)amine and 1,3,5-
tris(4-aminophenyl)benzene with pyromellitic dianhydride, respectively.
The amorphous networks exhibit excellent thermal stability, large BET
surface areas up to 1454 m2 g−1, and narrow pore size distribution in the
range from 5 to 6 Å. Their adsorption−desorption isotherms of CO2 are
reversible, and the CO2 uptakes at 273 K and 1 bar are up to 16.8 wt %.
Moreover, based on the ratios of initial slopes of isotherms, the CO2/N2
and CO2/CH4 separation factors are as high as 102 and 12, respectively.
The above CO2 adsorption and separation properties are attributed to the presence of abundant electron-rich heteroatoms in the
polyimide networks and the unifrom ultralmicroporous structures. In addition, for MPI-1, the adsorption capacity of benzene
vapor is 119.8 wt %, while the separation factors of benzene over nitrogen and water reach 342 and 28, respectively. The
outstanding selective adsorptions of CO2 gas and benzene vapor endow the microporous polyimides with promising potential in
CO2 capture and separation as well as air- and water-cleaning applications.
reach high level. However, at low pressure region, e.g., in the
case of practical CO2 capture for post-combustion flue gas, the
CO2 capacities are not satisfactory because of the lack of strong
affinity toward CO2 molecule. Therefore, the current efforts are
made to incorporate polar groups onto microporous frame-
works through post-modification method or directly synthesize
micoroporous polymers using the nitrogen- and oxygen-rich
monomers. In this way, CO2 adsorption capability and isosteric
heat of adsorption can be enhanced by means of the dipole−
quadrupole interactions or hydrogen-bonding effect between
pore wall and CO2 molecule. Cooper and co-workers found
that, at 273 K and 1 bar, the network-A only composed of
aromatic benzene rings could uptake CO2 of 2.65 mmol g−1.
After inserting a triazole ring between two phenyls, the
network-C exhibited remarkably improved CO2 uptake up to
3.86 mmol g−1.12 The other successful examples are polyamine
and sulfonate-grafted PPN-6s,13 oxygen-rich phloroglucinol-
based porous polymers,14 nitrogen-rich microporous poly-
benzimidazoles,15 nitrogen- and oxygen-rich microporous
cyanate resins,16 etc.
INTRODUCTION
■
Carbon dioxide (CO2) capture and storage is a newly emerged
technology with promising potentials in the reduction of
greenhouse gas and purification of nature gas. CO2 gas, mainly
produced from the burning of fossil fuels and biomass, is
regarded as a culprit for global-warming. Usually, the flue gas of
post-combustion from coal-fired power plants is composed of
15−16% CO2, 70−75% N2, 5−7% water vapor, and some other
impurities,1 whereas the pre-combustion gases, e.g., natural gas
and landfill gas, contain as high as 5−20% and 40−60% CO2
contaminant, respectively, which has to be removed to facilitate
sufficient burning.2 On the other hand, as an important
chemical material and reaction medium, the recovered CO2 can
be utilized in the syntheses of numerous polymers and fine
chemicals such as polycarbonates,3 cyclic carbonate,4 dimethyl
ether,5 and methylacrylic acid.6 In this regard, developing
porous adsorbents with large surface area, high CO2 adsorption
capacity, excellent CO2/N2 and CO2/CH4 selectivities is of
significance for both academic study and industrial applications.
Aqueous solutions of amines such as ethanolamine have been
widely employed to absorb CO2, but the subsequent process
for the regeneration of amine and recovery of CO2 is rather
energy-consuming.7 As an alternative method, the physical
sorption of CO2 by porous solid materials with large surface
area are receiving attention, including porous zeolites,8 metal−
organic frameworks,9 and microporous organic polymers.10
Recently, some results for CO2 adsorption in microporous
organic polymers have been reported.11 At high pressure, the
saturation uptakes of CO2 in porous polymer networks can
In parallel to CO2 capture, the recovery and selective removal
of toxic organic vapors from air and water are equally important
since the chemical pollutants in working and life environments,
emitted from synthetic industry and chemical products such as
paints, varnishes, and chemical cleaners, are seriously harmful
Received: March 8, 2013
Revised: March 22, 2013
Published: April 1, 2013
© 2013 American Chemical Society
3058
dx.doi.org/10.1021/ma400496q | Macromolecules 2013, 46, 3058−3066

Macromolecules
Article
mA) with a copper target at a scanning rate of 2°/min. Field-emission
scanning electron microscopy (FE-SEM) experiments were carried on
a Nova NanoSEM 450. Before measurement, the samples were
sputter-coated with chromium film to facilitate conduction. Trans-
mission electron microscopy (TEM) images were obtained on a
Hitachi HT-7700 operated at 100 kV. Thermogravimetric analysis
curves were recorded on a NETZSCH TG 209 thermal analyzer by
heating the samples (∼8 mg) up to 800 °C with ramping rate of 10 °C
min−1 under nitrogen flow. Adsorption and desorption measurements
for all the gases and vapors were conducted on an Autosorb iQ
(Quantachrome) analyzer. Prior to measurements, the samples were
degassed at 150 °C under high vacuum overnight. Adsorption and
desorption isotherms of nitrogen were measured at 77 K. The surface
areas were calculated according to the Brunauer−Emmett−Teller
(BET) model in the relative pressure (P/P0) range from 0.07 to 0.15
for MPI-1, from 0.12 to 0.20 for MPI-2, and from 0.05 to 0.12 for
MPI-3. CO2 and CH4 adsorption isotherms were measured at 273 K
up to 1.0 bar. The adsorptions of benzene, cyclohexane and water
vapors were measured with the pressure up to the saturated vapor
pressure at 298 K. N2 adsorption isotherms at 273 and 298 K were
measured in order to evaluate the adsorption selectivities of CO2/N2,
CO2/CH4, benzene/N2, and benzene/water, which were calculated
from the ratios of initial slopes of pure-component sorption isotherms
of gases and vapors.
to human health even though at a low concentration.
Nevertheless, relative to the studies on CO2 capture, less
attention has been paid to the adsorption of organic vapors.
Only very recently, its significance is being emphasized owing
to the growing demands for air and water-cleaning as well as
the recovery of organic chemicals.15d,16,17 Langsam and
Robeson reported that poly(TMSP) with large free volume
were effective to remove chloroform and toluene from water.18
The wholly aromatic microporous polymers also have
promising potential for this purpose, in particular for aromatic
compounds, due to the π−π interaction between phenyls and
benzene rings.
In the past years, linear polyimides with intrinsic micro-
porosity and hypercrosslinked microporous polyimide networks
have been reported.19 Among them, the microporous
polyimides with three-dimensional network structure have
aroused great interest because of the extremely rigid
heteroaromatic skeleton and excellent resistance to high
temperature and aggressive chemicals. Moreover, the presence
of abundant nitrogen and oxygen atoms from the imide rings
may lead to polyimide networks the favorable interaction with
CO2 gas. Some interesting data of gas and vapor adsorptions in
microporous polyimides have appeared in the literature. For
example, the tetraphenyladmantane-based microporous poly-
imide can uptake 14.6 wt % CO2 at 273 K and 1 bar, 99.2 wt %
benzene and 59.7 wt % cyclohexane at 298 K and 0.9 bar.19g
However, up to now, systematical studies on synthesis, pore
structure, chemical composition and their correlations with
adsorptions and separations of CO2 gas and organic vapors for
microporous polyimides are rarely reported. In addition, the
pore sizes in many previous materials are relatively wide, which
are disadvantageous for the separation of gases. The motivation
of the present work, therefore, is to prepare series of
microporous polyimides with uniform pore size through
elaborately optimizing the polymerization process. Their
adsorptions of CO2 gas and organic vapors as well as the
separations of CO2/N2, CO2/CH4, CO2/H2O, benzene/N2,
and benzene/H2O were evaluated and investigated in term of
the porosity parameters, chemical structure and composition of
the polymer networks.
Synthesis of Tetrakis(4-nitrophenyl)methane (TNPM). Tetra-
phenylmethane (7.4 g, 23.09 mmol) was added into fuming nitric acid
(40 mL) at −40 °C under vigorous stirring. Then, acetic anhydride
(12.5 mL) and acetic acid (25 mL) were slowly added and stirred for 1
h. After filtration, the resulted yellow solid was recrystallized with THF
to give yellow crystals. Yield: 65%. Mp: 339 °C; 1H NMR (400 MHz,
CDCl3): δ (ppm) 8.22−8.25 (d, 8H, Ar−H), 7.60−7.62 (d, 8H, Ar−
H). FTIR (KBr, cm−1): 3070, 3100, 1605, 1591, 1519, 1493, 1347,
840, 757, 744, 711.
Synthesis of Tetrakis(4-aminophenyl)methane (TAPM).
TAPM was prepared in reference to the procedure in ref 20 with
some modifications. In a hydrogenator, dry tetrahydrofuran (40 mL),
TNPM (1.6 g, 3.19 mmol) and catalysis amount of Pd/C was added
and purged with nitrogen. Then, the mixture was stirred under 1.2
MPa hydrogen pressure at room temperature for 3 days. After
removing the Pd/C powder and solvent, the white product was
1
purified by recrystallization. Yield: 95%. Mp: 262 °C; H NMR (400
MHz, DMSO-d6): δ (ppm) 6.66−6.68 (d, 8H, Ar−H), 6.38−6.40 (d,
8H, Ar−H), 4.85 (s, 8H, −NH2). FTIR (KBr, cm−1): 3442, 3400,
3363, 3333, 3027, 3100, 1620, 1508, 1281, 1216, 1184, 820, 754, 577;
HRMS calculated for C25H24N4, 380.2001; found, 380.2004.
EXPERIMENTAL SECTION
Synthesis of Tris(4-aminophenyl)amine (TAPA). TAPA was
prepared by a procedure similar to that for TAPM except that the
precursor compound is tris(4-nitrophenyl)amine instead of tetrakis(4-
■
Materials. Tetraphenylmethane, tris(4-nitrophenyl)amine, 4-nitro-
acetophenone, and pyromellitic dianhydride (PMDA) were purchased
from J&K Chemical Co., Ltd. PMDA was purified by sublimation prior
to use. m-Cresol, isoquinoline, and other reagents were purchased
from Shanghai Chemical Reagent Co. m-Cresol was purified by
distillation under reduced pressure and dehydrated with 4A molecular
sieves. Tetrahydrofuran (THF) was purified by refluxing over sodium
with the indicator benzophenone complex. Isoquinoline, fuming nitric
acid, acetic acid, acetic anhydride, and other reagents were of reagent
grade and used as received.
Instrumention. Fourier transform infrared spectra (FTIR) of
synthesized products were recorded using a Nicolet 20XB FT-IR
spectrophotometer in 400−4000 cm−1. Samples were prepared by
dispersing the complexes in KBr to form disks. 1H NMR spectra were
recorded on a 400 MHz Varian INOVA NMR spectrometer, using
tetramethylsilane as an internal standard. Solid-state 13C CP/MAS
(cross-polarization with magic angle spinning) spectra were measured
on a Varian Infinity-Plus 400 spectrometer at 100.61 MHz at an MAS
rate of 10.0 kHz using zirconia rotors 4 mm in diameter using a
contact time of 4.0 ms and a relaxation delay of 2.0 s. Elemental
analyses were determined with an Elementar Vario EL III elemental
analyzer. Wide-angle X-ray diffractions (WAXD) from 5° to 60° were
performed on Rigku D/max-2400 X-ray diffractometer (40 kV, 200
1
nitrophenyl)methane. Yield: 90%. Mp: 241 °C; H NMR (400 MHz,
DMSO-d6): δ (ppm) 6.60−6.62 (d, 6H, Ar−H), 6.46−6.44 (d, 6H,
Ar−H), 4.68 (s, 6H, −NH2). FTIR (KBr, cm−1): 3411, 3340, 3029,
1622, 1504, 1262.
Synthesis of 1,3,5-Tris(4-nitrophenyl)benzene (TNPB). TNPB
was prepared in reference to the literature21 with some modifications.
4-Nitroacetophenone (25.1 g, 0.15 mol) and anhydrous ethanol (43.7
mL, 0.75 mmol) were charged in a three-necked flask equipped with a
magnetic stirrer. Thionyl chloride (18.2 mL, 0.25 mol) was added
dropwise. Then the mixture was reacted at the refluxing temperature
for additional 1 h. The solid was neutralized and washed successively
with deionized water, ethyl ether, and ethanol. The crude product was
recrystallized with DMF to give yellow needle-like solid. Yield: 77%.
Mp: 325 °C. 1H NMR (400 MHz, DMSO-d6): δ (ppm) 8.38−8.36 (d,
6H, Ar−H), 8.27−8.29 (6H, Ar−H), 8.26 (3H, Ar−H). FTIR (KBr,
cm−1): 1594, 1511, 1349, 862, 843, 629.
Synthesis of 1,3,5-Tris(4-aminophenyl)benzene (TAPB).
TAPB was prepared by a similar procedure to TAPM. Yield 95%.
1
Mp: 262 °C; H NMR(400 MHz, acetone-d6): δ (ppm) 7.59 (s, 3H,
Ar−H), 7.51−7.53 (d, 6H, Ar−H), 6.77−6.79 (d, 6H, Ar−H), 4.76
(6H, -NH2); 13C NMR (100 MHz, acetone-d6): δ (ppm) 149.0, 143.1,
3059
dx.doi.org/10.1021/ma400496q | Macromolecules 2013, 46, 3058−3066

Macromolecules
Article
Scheme 1. Synthesis Routes to the Three Polyimide Networks
Figure 1. SEM images of MPI-1 (a), MPI-2 (b), and MPI-3 (c); HR-TEM images of MPI-1 (a′), MPI-2 (b′) and MPI-3 (c′).
130.7, 128.9, 122.3, 115.5. FTIR (KBr, cm−1): 3442, 3360, 1620, 1593,
1516, 1283, 827.
mL m-cresol were added and stirred for 2 h. After the reaction was
slowly heated to room temperature, a catalytical amount of
isoquinoline was added. Then the temperature of polymerization
was raised in the heating schedule: 30 °C for 8 h, 80 °C for 4 h, 160
°C for 4 h, 200 °C for 10 h, and 220 °C for 4 h. Finally, the system was
cooled down, and the solid was isolated and washed successively with
DMF, methanol, and THF. The resultant product was extracted with
THF in a Soxhlet apparatus for 24 h, and dried at 120 °C under
Preparation of Polyimide Networks. The polymerizations of
MPI-1, MPI-2, and MPI-3 were carried out in similar procedures. Only
the polymerization of MIP-1 is described here as an example. A dry
Schlenk flask equipped with a stirrer, a condenser, and an ice-bath was
degassed using two evacuation−argon backfill cycles. Under argon
flow, TAPM (0.40 g, 1.38 mmol), PMDA (0.45 g, 2.07 mmol), and 10
3060
dx.doi.org/10.1021/ma400496q | Macromolecules 2013, 46, 3058−3066

Macromolecules
Article
Figure 2. (a) Adsorption (filled) and desorption (empty) isotherms of N2 for MPI-1 (+600), MPI-2 (+300), and MPI-3. (b) Pore size distribution
curves for MPI-1, MPI-2, and MPI-3.
vacuum to constant weight. Quantitative yield was achieved in every
case.
weight-losses at around 240 °C are possibly caused by the
solvent and catalyst entangled in the microporous networks.
The surface morphologies of the products were observed by
field-emission scanning electron microscopy (Figure 1a−c).
Similar to other microporous polymers,22 the samples are
homogeneous, composing of loose agglomerates of tiny
particles with rough surface and irregular shape. The pore
morphology and pore size of the three samples were examined
by a high-resolution transmittance electron microscopy (TEM
in Figure 1a′-1c′). For the sake of clarity, the enlarged TEM
images are presented in Figure S4−S6, Supporting Information.
The white region corresponds to the nanopore channels, while
the dark region is attributed to the polyimide skeleton. As can
be seen, the pores in all the three polyimides are uniform with
the width of pore channel at around 0.5 nm.
The porosity parameters of polyimide networks were
investigated by physical sorption of nitrogen at 77 K (Figure
2a). The isotherms displayed rapid nitrogen uptake at the very
low relative pressure (<0.01), indicative of the characteristics of
permanent micropores. In this study, three monomers with
different geometry configurations were selected as the building
blocks. TAPM has a tetrahedral shape with four phenyls
stretching outward from a sp3-hybridized carbon core. The
three-dimensional network with tetraphenylmethanes linked by
rigid imide rings is liable to generate large amount of pores.
Indeed, MPI-1 exhibits the BET surface area, Langmuir surface
area and total pore volume of 1454 m2 g−1, 2116 m2 g−1, and
1.07 cm3 g−1, respectively, which are the highest among the
three polymers (Table 1). In contrast, the planar-triangular
shape of TAPB makes the polymer incline to form two-
dimensional network. The π−π stacking between layers may
RESULTS AND DISCUSSION
■
Synthesis and Porous Structure of Polyimide Net-
works. Using m-cresol as reaction medium and isoquinoline as
catalyst, the condensation polymerizations of pyromellitic
dianhydride (PMDA) with tetra(4-aminophenyl)methane
(TAPM), 1,3,5-tris(4-aminophenyl)benzene (TAPB) and tris-
(4-aminophenyl)amine (TAPA) led to three polyimide net-
works MPI-1, MPI-2 and MPI-3, respectively (Scheme 1).
In FTIR spectra (Figure S1, Supporting Information), the
structures of polyimides are confirmed by the characteristic
bands at 1779 and 1726 cm−1 due to the symmetric and
asymmetric vibrations of the CO group, and that at 1349
cm−1 corresponding to the stretching vibration of the C−N−C
moiety of five-member imide ring. Meanwhile, the absorptions
of amino and anhydride in monomers are absent, indicating
that the reactive groups have been completely converted under
the polymerization conditions. In solid-state 13C CP/MAS
NMR spectra (Figure S2, Supporting Information), the
resonance of carbonyl carbon in imide ring appears at 165
ppm. The quaternary carbon in MPI-1 locates at 20 ppm, and
the N-substituted phenyl carbon for MPI-2 is at 146 ppm. The
overlapping signals from 121 to 137 ppm belong to the other
aromatic carbons in backbone. Elemental analyses show that
the measured chemical compositions are consistent with the
calculated values (Table S1, Supporting Information) except for
hydrogen and nitrogen. The slightly higher hydrogen and lower
nitrogen contents than the theoretical values may be caused by
the absorbed moisture and CO2 from air.
For each sample, the resultant brown solid can not dissolve
in any solvents, suggesting the hypercrosslinked structure. The
thermal stabilities of three samples were evaluated by
thermogravimetric analysis (TGA) under a nitrogen atmos-
phere (Figure S3, Supporting Information). The weight losses
corresponding to the decomposition of skeleton occur at over
530 °C, and the char yields at 800 °C are 57.4, 53.0, and 59.3
wt % for MPI-1, MPI-2, and MPI-3, respectively. The small
Table 1. Porosity Parameters of the Polyimide Networks
Obtained by N2 Adsorption
SBET
SLangmuir
Vtotal
pore size
(nm)
samples
monomers
(m2 g−1
)
(m2 g−1
)
(cm3 g−1
)
MPI-1
MPI-2
MPI-3
TAPM + PMDA
TAPA + PMDA
TAPB + PMDA
1454
814
2116
1185
842
1.07
0.59
0.31
0.59
0.52
0.55
586
3061
dx.doi.org/10.1021/ma400496q | Macromolecules 2013, 46, 3058−3066

Macromolecules
Article
Figure 3. Energy-optimized molecule models for MPI-1 (left), MPI-2 (middle), and MPI-3 (right): (a) obtained from an ideal linking mode; (b)
obtained from a randomly interpenetrating mode.
lead to the loss of pore. As a result, MPI-3 shows apparently
lower BET surface area of 586 m2 g−1 and pore volume of 0.31
cm3 g−1. The topological structure of MPI-2 constructed from
the nitrogen-centered TAPA monomer with triangular-
pyramidal shape lies between MPI-1 and MPI-3, and
consequently, MPI-2 has a modest BET surface area of 814
m2 g−1.
into ultramicropores (<7 Å) by the interpenetrated segments as
depicted in Figure 3b.
Adsorption of Organic and Water Vapors in Poly-
imide Networks. The adsorption isotherms of benzene,
cyclohexane and water vapors measured at 298 K are presented
in Figure 4. In contrast to cyclohexane and water, the three
The pore size distributions obtained by the nonlocal density
functional theory (NLDFT) are illustrated in Figure 2b. It is
interesting to observe that, although the WAXD curves indicate
that the three polymer networks are amorphous (Figure S7,
Supporting Information), their pore sizes are quite uniform,
centering at around 0.59, 0.52, and 0.55 nm for MPI-1, MPI-2,
and MPI-3, respectively, which are well consistent with the
observations in TEM images. In order to elucidate the possible
pore-forming mechanism, molecular simulations were carried
out using Materials Studio software. At first, the molecular
models were built assuming that the cross-linking reactions are
complete and all the reactive groups are linked in an ideal mode
according to the geometrical structure of monomer. After being
optimized by geometry and molecular mechanics successively,
and then relaxed through dynamic calculation at 298 K for 200
ps based on the COMPASS force field, the polymer networks
present the network topologies as illustrated in Figure 3a, in
which the pore sizes of MPI-1, MPI-2 and MPI-3 are 3.4, 2.5,
and 3.5 nm, respectively, far larger than the experimental values.
The above results indicate that the interpenetration of
networks during the synthesis of polyimides can not be ignored
since the polycondensations between polyamine and dianhy-
dride monomers are kinetically controlled, following the step-
polymerization mechanism. The whole process successively
undergoes several stages including oligomerization, branching,
gelation and the eventual formation of hypercrosslinked
network. Similar to the case of microporous polyarylates,23
under the controlled conditions, the polyamine monomer and
PMDA first react to yield hyperbranched precursor. Sub-
sequently, upon gradually heating the system to a higher
temperature, the long branches of the precursors start to
randomly cross-link with each other. Thus, the larger
mesopores shown in Figure 3a are homogeneously divided
Figure 4. Organic and water vapor adsorption isotherms at 298 K for
three polyimide networks (benzene +450, cyclohexane +250, water).
polyimides exhibit a rapid rise of benzene uptake at the low
relative pressure (P/P0 < 0.1), indicating that the wholly
aromatic skeleton have a strong affinity toward benzene
molecules. The data in Table 2 show that, at P/P0 = 0.8, the
sorption capacity of benzene for MPI-1 can reach 119.8 wt %,
which is much higher than carbon material F42C (39.5 wt %),24
metal azolate framework MAF-2 (20.6 wt %),25 and micro-
porous cyanate resin CE-1 (58.5 wt %).16a Cyclohexane is also
an organic vapor, but the lack of π−π interaction leads to
cyclohexane apparently lower uptake in comparison with
benzene vapor. In addition, the isotherms for water vapor are
typical type III sorption, and its uptake in MPI-1 is only 16.7 wt
3062
dx.doi.org/10.1021/ma400496q | Macromolecules 2013, 46, 3058−3066

Macromolecules
Article
Table 2. Uptakes of CO2 Gas, Organic, and Water Vapors in the Polyimide Networks
a
b
gas adsorption and selectivity
vapor adsorption and selectivity
sample
CO2 (wt %)
CO2/N2
CO2/CH4
CO2/H2O
C6H6 (wt %)
c-C6H12 (wt %)
H2O (wt %)
C6H6/N2
C6H6/H2O
MPI-1
MPI-2
MPI-3
16.8
13.8
9.9
102
71
8
12
10
10.5
5.5
119.8
76.6
54.9
50.1
44.8
41.5
16.7
9.9
342
310
167
28
13
14
41
10.9
9.6
a
Uptakes for CO2, N2, and CH4 were determined at 273 K and 1 bar. The CO2/N2 and CO2/CH4 selectivities were calculated from initial slopes of
pure-component sorption isotherms. The CO2/H2O selectivities were calculated from the ratio of uptake of CO2 at 0.15 bar and 298 K to that of
b
water vapor at 0.05 bar at 298 K. Uptakes for C6H6 (benzene vapor), c-C6H12 (cyclohexane vapor) and H2O (water vapor) were determined at P/
P0 = 0.8 and 298 K. The C6H6/N2 and C6H6/H2O selectivities were calculated from initial slopes of pure-component sorption isotherms.
Figure 5. Adsorption (filled) and desorption (empty) isotherms of CO2 at 273 and 298 K for MPI-1, MPI-2, and MPI-3.
%, indicative of the hydrophobic nature of the polyimide
networks.
isotherms at different temperatures in term of Clausius−
Clapeyron equation:29
The separations of benzene vapor from N2 and H2O vapor
were examined by comparing their initial slopes (Figure S8,
Supporting Information). The results reveal that the benzene/
N2 and benzene/H2O selectivities are as high as 342 and 28,
respectively. The high adsorption capacity of benzene and its
excellent selectivities over N2 and water are of significant
importance in air and water-cleaning in the environmental
protection field.
Q st
ln p =
+ C
(1)
RT
where P, T, R, and C are the pressure, temperature at the
equilibrium state, the gas constant, and equation constant,
respectively.
The dependencies of CO2 isosteric enthalpies (Qst) on the
adsorbed amount are shown in Figure 6. For each sample, the
CO2 Adsorption and Selectivity of CO2/N2 and CO2/
CH4. The large surface area, ultramicroporous structure and
narrow pore size distribution of the polyimides arouse our
interest to further explore their property of CO2 capture and
separation from other gases. The sorption isotherms of CO2 at
273 and 298 K are presented in Figure 5. For all the three
samples, the CO2 uptakes appear a rapid rise in the initial stage,
implying that CO2 molecule has favorable interaction with the
polymer skeleton. Moreover, the adsorption−desorption curves
of CO2 are roughly reversible. Thus, the samples can absorb
CO2 gas even on exposure to atmosphere and can release the
equivalent amount of CO2 from their cavity while releasing the
pressure, which characteristic is quite desirable for the practical
application in CO2 capture and regeneration of porous
adsorbent. Furthermore, it is noted that the uptakes of CO2
continually increase with the pressure, and have not reached
saturation in the experimental range, indicating that the higher
adsorption capacity can be achieved at the increased pressure.
MPI-1 has the CO2 uptake of as high as 16.8 wt % at 1 bar
and 273 K (Table 2), which are comparable to that of COFs
(5.7−16.7 wt %),11b,26 CMPs (4.0−17.0 wt %),22,27 and other
microporous polymers.28 In addition to the large surface area,
the excellent CO2 adsorption performance is probably resulted
from the high affinity of skeleton toward CO2 since polyimide
networks contain abundant electron-rich nitrogen and oxygen
atoms. To support the above deduction, the isosteric enthalpies
(Qst) of MPIs toward CO2 were calculated from the adsorption
Figure 6. Variation of CO2 isosteric enthalpies with the adsorbed
amount.
Qst values significantly decrease with the adsorbed amount,
meaning that the interaction between CO2 and pore wall is
stronger than that between CO2 molecules. Interestingly, the
three polymers have the similar chemical structure but
apparently different Qst values. Moreover, the Qst value has
the same ranking order as the surface area, pore volume and
CO2 adsorption capacity, implying that, besides the intrinsic
affinity of polymer skeleton, the CO2 capture ability of
microporous polymers may also be highly related to the
3063
dx.doi.org/10.1021/ma400496q | Macromolecules 2013, 46, 3058−3066

Macromolecules
Article
porosity, topological structure and complexity of pore channels
in the network. The combination of the pore structure and
affinity between CO2 and pore wall results in MPI-1 the highest
CO2 uptake among the three polymers.
those of other porous materials like HCPs (20−24 kJmol−1),31
BILPs (26.5−28.8 kJmol−1),15c,32 BLPs (20.2−28.3 kJmol−1),33
and comparable to microporous polymers with organic
ammonium ions or Lewis basic amine in pores (30−55
kJmol−1).34 The exceptionally high Q0 values can be due to the
enhanced affinity for CO2 molecule because of the abundant N
and O atoms as well as the sorption selectivity effects of the
ultramicropores35 in this series of ultramicroporous polyimides.
In order to evaluate the separation property of the
polyimides, the sorption isotherms of CH4 and N2 at 273 K
were measured and are compared with those of CO2 in Figure
8. The curves show that CO2 has the considerably higher
uptake than N2 and CH4 in the whole pressure range. The
initial slopes of the CO2, N2 and CH4 based on the adsorption
isotherms at 273 K were calculated (Figure S9, Supporting
Information), and their ratios of CO2 over N2 and CH4 are
used as the separation factors of CO2/N2 and CO2/CH4 gas
pairs.15c,36 It is known that the kinetic diameter of CO2 (3.30
Å) is much smaller than that of N2 (3.64 Å),37 while the
polyimides in this study have very small pores (ca. 5.5 Å) and
quite narrow pore size distributions. The sorption selectivity
effects are advantageous for the recognition of small CO2 from
the large N2 molecule. For MPI-1, the separation factor of
CO2/N2 is up to 102, surpassing those for zeolitic imidazole
frameworks (ZIFs, 20−50),8b Bio-MOF-11 (81),9c and
comparable to the amine-decorated 12-connected porous
The physical sorption of CO2 gas on the polyimide network
can also be investigated though the virial equation30 as below:
ln(n/P) = A0 + A1n + A2n2 + ···
(2)
where n is the adsorbed amount at pressure P, KH refers to
Henry’s constant, and A0, A1, A2, etc. are virial coefficients.
At low adsorbed amount of gas, A2 and higher terms can be
neglected. As shown in Figure 7, the virial plots of CO2 for all
MOF CAU-1 (101)9 and the hierarchically porous electron-
e
Figure 7. Virial plots of CO2 adsorption in MPI-1, MPI-2, and MPI-3
at 273 and 298 K.
rich covalent organonitridic frameworks (PECONF-1,
109).38Among the three polymers, MPI-1 simultaneously
possesses the higher CO2 uptake and CO2/N2 selectively.
Relatively, the CO2/N2 selectivities of MPI-2 and MPI-3
apparently decrease. Considering that the three polymers have
the similar pore sizes and distributions, the main reason for
their difference in CO2/N2 selectivity may lie in the complexity
of pore channels arisen from the topological structures of the
building units for the three polyimides.
For CO2/CH4 gas pair, things are different. The critical
temperature of CH4 gas is 191 K, which is higher than that of
N2 (126 K).39 The previous reports have demonstrated that the
gas solubility coefficient in a polymer is positively correlated
with its critical temperature.40 Therefore, for each polymer,
compared to N2 gas, CH4 has the more strong adsorption
ability with polymer skeleton, leading to the significantly higher
uptake of CH4 than N2. Therefore, in each case, CO2/CH4
selectivity is considerably lower than CO2/N2 (Table 2). At 273
K, the CO2/CH4 selectivity for MPI-2 is 12, less than BILP-2
but comparable to other BILPs15c and some ZIFs.8b On the
other hand, the ranking order of CO2/CH4 selectivities for the
three polymers is different from that of CO2/N2. The presence
the three polymers exhibit good linear relationship. The data of
A0, A1, and KH are listed in Table S2, Supporting Information.
A0 is related to the gas-material interaction, whereas A1 reflects
gas-gas interaction. It is seen that MPI-1 has the largest A0
value, which is consistent with the ranking orders of CO2
capacities and the isosteric enthalpies obtained by the
Clausius−Clapeyron equation.
As the measure of the interaction between CO2 and pore
wall, the enthalpy of adsorption at zero coverage (Q0) can be
calculated using the Vant Hoff equation:
Q 0
RT
d[ln KH]
dT
=
(3)
where Q0 is the slope of the plot of ln(KH) vs 1/T, while KH is
Henry’s constant derived from the equation KH = exp(A0)
based on virial plot.
According to the isotherms at different temperatures, the
calculated Q0 values for MPI-1, MPI-2, and MPI-3 are 34.8,
30.4, and 31.4 kJ mol−1, respectively. The values are higher than
Figure 8. Adsorption isotherms of CO2, CH4 and N2 gases at 273 K for MPI-1 (a), MPI-2 (b), and MPI-3 (c).
3064
dx.doi.org/10.1021/ma400496q | Macromolecules 2013, 46, 3058−3066

Macromolecules
Article
of quaternary carbons may result in MPI-1 the favorable
adsorption for the aliphatic CH4 gas. As a result, the CO2/CH4
selectivity of MPI-1 is the lowest among the three polymers.
It is noteworthy that, in fact, the components of flue gas,
natural gas and landfill gas are rather complex. In the case of
flue gas, the influence of water on the CO2/N2 selectivity
should also be considered. For a typical flue gas of post-
combustion from coal-fired plants, there are about 15−16%
CO2 and 5−7% water vapor.1 In other word, the partial
pressure of CO2 in the flue gas is about 0.15 bar, whereas that
of water vapor is 0.05 bar. Therefore, from the adsorption
curves of CO2 and water vapor at 298 K (Figure 4 and Figure
5), the ratios of adsorption capacity of CO2 at 0.15 bar to that
of water vapor at 0.05 bar can be calculated. The data in Table
2 show that, for flue gas, the three samples are more
preferentially adsorbed by CO2 rather than water vapor. For
example, the CO2/H2O ratios for MPI-1, MPI-2, and MPI-3 are
10.5, 5.5, and 10.9, respectively. Even so, the competition of
adsorption between water vapor and CO2 can not be ignored.
Especially for a humidified flue gas or at the high feed pressure,
the actual values of CO2/N2 sorption selectivity may be lower
than those listed in Table 2 to some extent.
as the data of elemental analysis and virial parameters of the
polyimide networks. This material is available free of charge via
the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*(Z.G.W.) E-mail: zgwang@dlut.edu.cn.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank the National Science Foundation of China (Nos.
51073030 and 51273031) and the Program for New Century
Excellent Talents in University of China (No. NCET-06-0280)
for financial support of this research.
REFERENCES
■
(1) D’Alessandro, D.; Smit, B.; Long, J. Angew. Chem., Int. Ed. 2010,
49, 6058−6082.
(2) Bae, Y. S.; Snurr, R. Q. Angew. Chem., Int. Ed. 2011, 50, 11586−
11596.
(3) Buchard, A.; Jutz, A. F.; Kember, M. R.; White, A. J. P.; Rzepa, H.
S.; Williams, C. K. Macromolecules 2012, 45, 6781−6795.
(4) Lee, E. H.; Ahn, J. Y.; Dharman, M. M.; Park, D. W.; Kim, I. L.
Catal. Today 2008, 131, 130−134.
(5) Zhao, Y.; Zhang, J. J. Nat. Gas Chem. 2007, 16, 389−392.
(6) Cheng, Q.; Zhong, S. Chin. J. Catal. 2003, 24, 558−562.
(7) (a) Rochelle, G. T. Science 2009, 325, 1652−1654.
(b) MacDowell, N.; Florin, N.; Buchard, A.; Hallett, J.; Galindo, A.;
Jackson, G.; Adjiman, C. S.; Williams, C. K.; Shah, N.; Fennell, P.
Energy Environ. Sci. 2010, 3, 1645−1669.
(8) (a) Zheng, B. S.; Bai, J. F.; Duan, J. G.; Wojtas, L.; Zaworotko, M.
J. J. Am. Chem. Soc. 2011, 133, 748−751. (b) Banerjee, R.; Furukawa,
H.; Britt, D.; Knobler, C.; O’Keeffe, M.; Yaghi, O. M. J. Am. Chem. Soc.
2009, 131, 3875−3877. (c) Yang, S. T.; Kim, J. Y.; Kim, J.; Ahn, W. S.
Fuel 2012, 97, 435−442. (d) McEwen, J.; Hayman, J. D.; Yazaydin, A.
O. Chem. Phys. 2013, 412, 72−76. (e) Pham, T. D.; Liu, Q. L.; Lobo,
R. F. Langmuir 2013, 29, 832−839.
(9) (a) Vaidhyanathan, R.; Iremonger, S. S.; Dawson, K. W.; Shimizu,
G. K. H. Chem. Commun. 2009, 5230−5232. (b) Couck, S.; Denayer, J.
́
F. M.; Baron, G. V.; Remy, T.; Gascon, J.; Kapteijn, F. J. Am. Chem.
Soc. 2009, 131, 6326−6327. (c) An, J. H.; Geib, S. J.; Rosi, N. L. J. Am.
Chem. Soc. 2010, 132, 38−39. (d) Si, X. L.; Jiao, C. L.; Li, F.; Zhang, J.;
Wang, S.; Liu, S.; Li, Z. B.; Sun, L. X.; Xu, F.; Gabelica, Z.; Schick, C.
Energy Environ. Sci. 2011, 4, 4522−4527.
CONCLUSIONS
■
Three microporous polyimide networks MPI-1, MPI-2, and
MPI-3 with BET surface areas in the range of 586−1454 m2 g−1
have been successfully synthesized through polycondensation
of pyromellitic dianhydride with tetra(4-aminophenyl)methane,
1,3,5-tris(4-aminophenyl)benzene), and tris(4-aminophenyl)-
amine, respectively. Their chemical structures were confirmed
by FTIR, solid-state 13C CP/MAS NMR spectra, and elemental
analyses. The analyses of N2 sorption isotherms reveal that
three polymers have quite narrow pore size distribution with
pores centered at ca. 5.5 Å, agreeing with the observations of
high-resolution TEM images. The small and uniform pore sizes
were resulted from the homogeneous and random inter-
penetration of network by means of elaborate control of the
polymerization conditions, which process was explained with
the assistance of computer modeling method. The presence of a
large amount of heteroatoms results in the strong affinity of
polyimide skeletons toward CO2 molecule. Consequently, all
the three polymers have high isosteric heat of adsorption above
30.0 kJ mol−1. At 1.0 bar and 273 K, the high CO2 uptake
capacities as high as 16.8 wt % have been achieved. The
selectivities for CO2/N2 and CO2/CH4 are up to 102, and 12,
respectively, superior to many other microporous organic
polymers. Furthermore, the polymers display excellent abilities
to adsorb organic pollutants like benzene vapor. For MPI-1, the
uptake for benzene vapor at room temperature reach 119.8 wt
%, and the benzene/N2 and benzene/H2O selectivities are 342
and 28, respectively, which exceed most silica, inorganic−
organic hybrid, and carbon materials. The above results indicate
that the resulted polyimides are promising candidates as
adsorbents for CO2 capture as well as air- and water-cleaning in
the environmental protection field.
(10) (a) Li, J. R.; Kuppler, R. J.; Zhou, H. C. Chem. Soc. Rev. 2009,
38, 1477−1504. (b) Panda, T.; Pachfule, P.; Chen, Y.; Jiang, J.;
Banerjee, R. Chem. Commun. 2011, 47, 2011−2013. (c) Weber, J.; Du,
N. Y.; Guiver, M. D. Macromolecules 2011, 44, 1763−1767. (d) Arstad,
B.; Fjellvag, H.; Kongshaug, K. O.; Swang, O.; Blom, R. Adsorption
2008, 14, 755−762.
(11) (a) Yuan, D. Q.; Lu, W. G.; Zhao, D.; Zhou, H. C. Adv. Mater.
2011, 23, 3723−3725. (b) Furukawa, H.; Yaghi, O. M. J. Am. Chem.
Soc. 2009, 131, 8875−8883.
(12) Dawson, R.; Stockel, E.; Holst, J. R.; Adams, D. J.; Cooper, A. I.
̈
Energy Environ. Sci. 2011, 4, 4239−4245.
(13) (a) Lu, W. G.; Yuan, D. Q.; Sculley, J.; Zhao, D.; Krishna, R.;
Zhou, H. C. J. Am. Chem. Soc. 2011, 133, 18126−18129. (b) Lu, W.
G.; J. Sculley, P.; Yuan, D.; Krishna, Q. R.; Wei, Z. G..; Zhou, H. C.
Angew. Chem., Int. Ed. 2012, 51, 7480−7484.
ASSOCIATED CONTENT
* Supporting Information
■
(14) (a) Katsoulidis, A. P.; Kanatzidis, M. G. Chem. Mater. 2011, 23,
1818−1824. (b) Katsoulidis, A. P.; Kanatzidis, M. G. Chem. Mater.
2012, 24, 471−479.
(15) (a) Rabbani, M. G.; El-Kaderi, H. M. Chem. Mater. 2012, 24,
1511−1517. (b) Rabbani, M. G.; Reich, T. E.; Kassab, R. M.; Jackson,
K. T.; El-Kaderi, H. M. Chem. Commun. 2012, 48, 1141−1143.
(c) Zhao, Y. C.; Cheng, Q. Y.; Zhou, D.; Wang, T.; Han, B. H. J.
S
This section contains nine figures and two tables, including the
FTIR and solid-state 13C CP/MAS NMR spectra, X-ray
diffraction patterns, TGA curves, enlarged TEM images,
adsorption selectivities of CO2 over CH4 and N2, and
adsorption selectivities of benzene over H2O and N2 as well
3065
dx.doi.org/10.1021/ma400496q | Macromolecules 2013, 46, 3058−3066

Macromolecules
Article
Mater. Chem. 2012, 22, 11509−11514. (d) Yu, H.; Tian, M. Z.; Shen,
C. J.; Wang, Z. G. Polym. Chem. 2013, 4, 961−968.
(38) Mohanty, P.; Kull, L. D.; Landskron, K. Nat. Commun. 2011, 2,
1−6.
(16) (a) Yu, H.; Shen, C. J.; Tian, M. Z.; Qu, J.; Wang, Z. G.
Macromolecules 2012, 45, 5140−5150. (b) Yu, H.; Shen, C. J.; Wang,
Z. G. ChemPlusChem 2013, DOI: 10.1002/cplu.201300090.
(17) Wang, X. S.; Liu, J.; Bonefont, J. M.; Yuan, D. Q.; Thallapally, P.
K.; Ma, S. Q. Chem. Commun. 2013, 49, 1533−1535.
(39) Jeans, J. An Introduction to the Kinetic Theory of Gases;
Cambridge University Press: London, 1982.
(40) (a) Amerongen, G. J. J. Appl. Phys. 1946, 17, 972−986.
(b) Suwandi, M. S.; Stern, S. A. J. Polym. Sci., Polym. Phys. Ed 1973, 11,
663−681. (c) Shah, V. M.; Hardy, B. J.; Stern, S. A. J. Polym. Sci.,
Polym. Phys. Ed. 1986, 24, 2033−2047.
(18) Robeson, L. M.; Langsam, M. Sep. Sci. Technol. 1992, 27, 1245−
1258.
(19) (a) Weber, J.; Su, Q.; Antonietti, M.; Thomas, A. Macromol.
Rapid Commun. 2007, 28, 1871−1876. (b) Ghanem, B. S.; McKeown,
N. B.; Budd, P. M.; Al-Harbi, N. M.; Fritsch, D.; Heinrich, K.;
Starannikova, L.; Tokarev, A.; Yampolskii, Y. Macromolecules 2009, 42,
7881−7888. (c) Wang, Z. G.; Zhang, B. F.; Yu, H.; Sun, L. X.; Jiao, C.
L.; Liu, W. S. Chem. Commun. 2010, 46, 7730−7732. (d) Wang, Z. G.;
Zhang, B. F.; Yu, H.; Li, G. Y.; Bao, Y. J. Soft Matter 2011, 7, 5723−
5730. (e) Luo, Y.; Li, B.; Liang, L.; Tan, B. Chem. Commun. 2011, 47,
7704−7706. (f) Rao, K. V.; Haldar, R.; Kulkarni, C.; Maji, T. K.;
George, S. J. Chem. Mater. 2012, 24, 969−971. (g) Shen, C. J.; Bao, Y.
J.; Wang, Z. G. Chem. Commun. 2013, 49, 3321−3323.
(20) Miller, R. D.; Burland, D. M.; Jurich, M.; Lee, V. Y.; Moylan, C.
R.; Thackara, J. I.; Twieg, R. J.; Verbiest, T.; Volksen, W.
Macromolecules 1995, 28, 4970−4974.
(21) (a) Hu, Z. G.; Liu, J.; Li, G. A.; Dong, Z. B.; Li, W. J. Chin.
Chem. Soc. 2004, 51, 581−583. (b) Yeh, R. M.; Xu, J.; Seeber, G.;
Raymond, K. N. Inorg. Chem. 2005, 44, 6228−6239.
(22) (a) Dawson, R.; Laybourn, A.; Clowes, R.; Khimyak, Y. Z.;
Adams, D. J.; Cooper, A. I. Macromolecules 2009, 42, 8809−8816.
(b) Stockel, E.; Wu, X. F.; Trewin, A.; Wood, C. D.; Clowes, R.;
̈
Campbell, N. L.; Jones, J. T. A.; Khimyak, Y. Z.; Adams, D. J.; Cooper,
A. I. Chem. Commun. 2009, 2, 212−214.
(23) Zhang, B. F.; Wang, Z. G. Chem. Mater. 2010, 22, 2780−2789.
́ ́
(24) Lillo-Rodenas, M. A.; Cazorla-Amoros, D.; Linares-Solano, A.
Adsorption 2011, 17, 473−481.
(25) Zhang, J. P.; Chen, X. M. J. Am. Chem. Soc. 2008, 130, 6010−
6017.
(26) Tilford, R.; Mugavero, W. S. J., III; Pellechia, P. J.; Lavigne, J. J.
Adv. Mater. 2008, 20, 2741−2746.
(27) Holst, J. R.; Stockel, E.; Adams, D. J.; Cooper, A. Macromolecules
̈
2010, 43, 8531−8538. (b) Dawson, R.; Adams, D. J.; Cooper, A.
Chem. Sci. 2011, 2, 1173−1177.
(28) (a) McKeown, N. B.; Budd, P. M. Macromolecules 2010, 43,
5163−5176. (b) Thomas, A. Angew. Chem., Int. Ed. 2010, 49, 8328−
8344. (c) Wu, D. C.; Xu, F.; Sun, B.; Fu, R. W.; He, H. K.;
Matyjaszewski, K. Chem. Rev. 2012, 112, 3959−4015.
(29) Krungleviciute, V.; Heroux, L.; Migone, A. D. J. Phys. Chem. B.
2005, 109, 9317−9320.
(30) Cole, J. H.; Everett, D. H.; Marshall, C. T.; Paniego, A. R.; Powl,
J. C.; Rodriguez-Reinoso, F. J. Chem. Soc., Faraday Trans. 1974, 70,
2154−2169.
(31) Martín, C. F.; Stockel, E.; Clowes, R.; Adams, D. J.; Cooper, A.;
̈
Pis, J. J.; Rubiera, F.; Pevida, C. J. Mater. Chem. 2011, 21, 5475−5483.
(32) Rabbani, M. G.; El-Kaderi, H. M. Chem. Mater. 2011, 23, 1650−
1653.
(33) (a) Jackson, K. T.; Rabbani, M. G.; Reich, T. E.; El-Kaderi, H.
M. Polym. Chem. 2011, 2, 2775−2777. (b) Reich, T. E.; Behera, S.;
Jackson, K. T.; Jena, P.; El-Kaderi, H. M. J. Mater. Chem. 2012, 22,
13524−13528.
(34) (a) Demessence, A.; D’Alessandro, D. M.; Foo, M. L.; Long, J.
R. J. Am. Chem .Soc. 2009, 131, 8784−8786. (b) Fu, Y. H.; Sun, D. R.;
Chen, Y. J.; Huang, R. K.; Ding, Z. X.; Fu, X. Z.; Li, Z. H. Angew.
Chem., Int. Ed. 2012, 51, 3364−3367.
(35) Yu, L.; Falco, C.; Weber, J.; White, R. J.; Howe, J. Y.; Titirici, M.
M. Langmuir 2012, 28, 12373−12383.
(36) (a) An, J.; Geib, S. J.; Rosi, N. L. J. Am. Chem. Soc. 2010, 132,
38−39. (b) Banerjee, R.; Furukawa, H.; Britt, D.; Knobler, C.;
O’Keeffe, M.; Yaghi, O. M. J. Am. Chem .Soc. 2009, 131, 3875−3877.
(37) Breck, D. W. Zeolite Molecular Sieves; John Wiley & Sons: New
York, 1994.
3066
dx.doi.org/10.1021/ma400496q | Macromolecules 2013, 46, 3058−3066