Vapor-phase transport as a novel route to hyperbranched polyamine-oxide hybrid materials

Article abstract of DOI:10.1021/cm303931q

A new method to prepare hyperbranched polyamine-oxide hybrid materials by means of a vapor-phase transport is developed. In this method, hybrid materials having hyperbranched amine polymers covalently bound to an oxide support are formed by exposing the oxide support to the vapor of small nitrogen-containing heterocyclic monomers, in contrast to the conventional liquid-phase method, in which the support is dispersed in an organic solution containing monomer species. The aziridine and azetidine monomers are polymerized on the surface of the oxide supports (i.e., silica and alumina), resulting in poly(ethylenimine) or poly(propylenimine) chains attached to the porous solid support. The results suggest that the hybrid materials can be prepared over a wide range of preparation conditions with organic contents comparable to or even higher than those obtained from the standard liquid-phase method. It is demonstrated that supports with more acidity result in the hybrid materials with higher organic content. Interestingly, the resulting supported polyamines have lower molecular weights than the previously reported materials prepared by the liquid-phase method. It is anticipated that the vapor-phase synthesis can be applied for the efficient introduction of polyamines into structural forms of supports such as fibers, membranes, and monoliths, for which the liquid-phase method may be inappropriate or inefficient.

Full text of DOI:10.1021/cm303931q

Article
pubs.acs.org/cm
Vapor-Phase Transport as A Novel Route to Hyperbranched
Polyamine-Oxide Hybrid Materials
†
Watcharop Chaikittisilp, Stephanie A. Didas, Hyung-Ju Kim, and Christopher W. Jones*
School of Chemical & Biomolecular Engineering, Georgia Institute of Technology, 311 Ferst Drive, Atlanta, Georgia 30332-0100,
United States
*
S Supporting Information
ABSTRACT: A new method to prepare hyperbranched polyamine-oxide
hybrid materials by means of a vapor-phase transport is developed. In this
method, hybrid materials having hyperbranched amine polymers covalently
bound to an oxide support are formed by exposing the oxide support to the
vapor of small nitrogen-containing heterocyclic monomers, in contrast to
the conventional liquid-phase method, in which the support is dispersed in
an organic solution containing monomer species. The aziridine and
azetidine monomers are polymerized on the surface of the oxide supports
(
(
i.e., silica and alumina), resulting in poly(ethylenimine) or poly-
propylenimine) chains attached to the porous solid support. The results
suggest that the hybrid materials can be prepared over a wide range of preparation conditions with organic contents comparable
to or even higher than those obtained from the standard liquid-phase method. It is demonstrated that supports with more acidity
result in the hybrid materials with higher organic content. Interestingly, the resulting supported polyamines have lower molecular
weights than the previously reported materials prepared by the liquid-phase method. It is anticipated that the vapor-phase
synthesis can be applied for the efficient introduction of polyamines into structural forms of supports such as fibers, membranes,
and monoliths, for which the liquid-phase method may be inappropriate or inefficient.
KEYWORDS: hyperbranched aminosilica, supported polyamine, poly(ethylenimine), poly(propylenimine), organic−inorganic hydrid,
ring-opening polymerization, class 3 CO adsorbent, PEI, PPI
2
INTRODUCTION
between the amines and the supports while providing high
loadings of amine groups, although very large polymer loadings
can sometimes remove or block the porosity. Ring-opening
polymerization of cyclic monomers onto the support surface6s
can be applied for various polymers including polypeptides
■
Porous organic−inorganic hybrid materials are of growing
interest because their applications can be expanded beyond the
traditional scope of organic or inorganic components alone.
The organic functionalization of porous oxide frameworks,
mostly demonstrated to date on silica-based materials, can offer
not only new functions but also additional synergistic
properties created by specific interactions at the hybrid
7
,8
and poly(ethylenimine). Small nitrogen-containing hetero-
cycles having high ring strain such as aziridine (three-
membered ring) and azetidine (four-membered ring) can
undergo nucleophilic ring-opening polymerization to generate
polyamines on the surface or in the pores of porous oxides with
1
interfaces. In particular, functionalization of mesoporous silica
with amino groups has been intensively studied because these
materials can be applied in a wide variety of applications
including catalysis, adsorption of heavy metal cations,
9
very high yields (see Scheme 1).
Among others, our group has developed a simple way to
prepare supported poly(ethylenimine) materials by ring-
opening polymerization of aziridine in the presence of porous
2
−5
immobilization of drugs and biomolecules, and CO capture.
2
Incorporation of amine moieties into/onto the support
frameworks has been achieved mostly via four liquid-phase
synthetic routes: (i) physical impregnation of monomeric or
polymeric amines into/onto the porous supports, (ii) covalent
grafting of amines, most often organosilanes with pendant
amino groups, onto the support surfaces, (iii) direct co-
condensation of amine-containing molecules and conventional
precursors during materials syntheses, and (iv) in situ
polymerization of amine-containing monomers in the pores
of supports, with the latter three methods resulting in amines or
7,8,10
silicates as supports in organic solvent media (Scheme 1).
These materials, which have been referred to as hyperbranched
aminosilica (HAS) materials, have been used as adsorbents for
11
the selective removal of oxygenates from bio-oil and for CO2
8
,12
capture from flue gas and ambient air.
Although the HAS
materials showed excellent performance in CO capture from
2
humid gas streams under a lab-scale, packed bed configuration,
heat and mass transfer resistances associated with the
1
aminopolymers covalently bound to the supports.
Among the aforementioned methods, in situ polymerization
to produce polyamines can provide strong covalent links
Received: December 7, 2012
Revised: January 22, 2013
Published: January 25, 2013
©
2013 American Chemical Society
613
dx.doi.org/10.1021/cm303931q | Chem. Mater. 2013, 25, 613−622

Chemistry of Materials
Article
Scheme 1. Ring-Opening Polymerization of Aziridine and
Azetidine onto the Oxide Support
EXPERIMENTAL SECTION
■
Chemicals. The following reagents were obtained from the
commercial suppliers and used as received without further purification:
2
(
-chloroethylamine hydrochloride (Aldrich, 99%), sodium hydroxide
EMD, ACS grade, 97%), azetidine hydrochloride (Aldrich, 97%),
potassium hydroxide (Fluka, purum ≥85%), Pluronic P123 EO -
2
0
PO -EO triblock copolymer (Aldrich), concentrated hydrochloric
70
20
acid (Sigma-Aldrich), tetraethyl orthosilicate (Aldrich, 98%), alumi-
num isopropoxide (Aldrich), 2 N hydrochloric acid (BDH), methanol
(BDH), acetone (BDH), activated acidic aluminum oxide (Sigma-
Aldrich, Brockmann I), activated basic aluminum oxide (Sigma-
Aldrich, Brockmann I), activated neutral aluminum oxide (Sigma-
Aldrich, Brockmann I), water (Fluka, TraceSELECT), sodium sulfate
(Sigma-Aldrich, 99+%), acetic acid (Sigma-Aldrich, glacial), and
poly(ethylenimine)s (Aldrich, Mw = 800, 1300, 2000, 25000; M =
n
6
00, 1200, 1800, 10000; Polysciences, M = 1800).
w
Synthesis of Aziridine. Aziridine was synthesized from 2-
8a
chloroethylamine according to the previous report with some
modifications in the purification step. To a 250-mL round-bottom
flask containing 2-chloroethylamine hydrochloride (30 g) was added a
sodium hydroxide aqueous solution (25.8 g of sodium hydroxide in
exothermic chemisorption of CO on polyamines and hindered
2
1
70 g of deionized water). The resultant solution was heated to 50 °C
diffusion in polymer-filled pores may prohibit packed bed
5a,13
and stirred at this temperature for 2 h. Aziridine was then recovered by
a partial static vacuum distillation at 75 °C. The collected distillate was
dried over sodium hydroxide pellets and kept in a freezer overnight.
Aziridine phase-separated as an upper layer of liquid and was
recovered. The purified aziridine was obtained as a colorless oil in
operation at a commercial scale.
For HAS and related
adsorbents to be utilized for large-scale CO capture, the use of
2
structured contactors such as hollow fibers, nanofibers, and
honeycomb monoliths as support materials has been proposed
to resolve some of these problematic issues associated with
1
70−80% yield and stored in a freezer. H NMR (400.0 MHz,
13
flowing large volumes of CO -containing gas over these
(CD ) SO, TMS): δ (ppm) 1.17, 1.53; C NMR (100.6 MHz,
2
3
2
13a,14
materials.
However, preparation of aminosilica hybrid
(CD
3
)
2
SO, TMS): δ (ppm) 17.4.
Preparation of Azetidine. Azetidine was prepared by distillation
materials, including HAS materials, typically requires reaction in
a condensed phase using liquid reagents or solvents.
Functionalization of the structured contactors, such as
monoliths, via such liquid-phase polymerization techniques is
difficult to carry out on a large scale.
1
8
of azetidine hydrochloride over potassium hydroxide. To a 100-mL
round-bottom flask was added potassium hydroxide (7.1 g) and
deionized water (4 mL). The mixture was stirred until potassium
hydroxide was completely dissolved. Then, 5 g of azetidine
hydrochloride was added to the stirred potassium hydroxide solution.
As is widely known, small nitrogen-containing heterocycles
can be vaporized even at ambient conditions because of their
1
Azetidine was isolated by distillation and stored in a freezer. H NMR
13
(
400.0 MHz, CDCl , TMS): δ (ppm) 1.99, 2.25, 3.55; C NMR
9
3
low boiling points. We have thus considered that vapors of
(
100.6 MHz, CDCl , TMS): δ (ppm) 22.1, 48.2.
3
such heterocyclic molecules should effectively transport or
diffuse into the interiors of the structured contactors, and that
vapor-phase polymerization techniques may therefore be
possible. If the internal surface of contactors contains suitable
polymerization initiators, ring-opening polymerization should
spontaneously occur under appropriate conditions, resulting in
contactor-supported polymers. Related approaches have been
Caution! Aziridine and azetidine are highly reactive and toxic. Use
extreme caution when handling these compounds. Only handle in a well-
ventilated fume hood and always wear proper personal protective
equipment.
Synthesis of SBA-15 Mesoporous Silica. SBA-15 silica supports
19
were synthesized by using a modified, previously reported method,
scaling the ingredients accordingly. In a typical procedure, pluronic
P123 triblock copolymer (12 g), concentrated hydrochloric acid (72
g), and deionized water (328 g) were mixed in a 1-L Erlenmeyer flask.
After being stirred for 5 h at 40 °C, 25.4 g of tetraethyl orthosilicate
was added to the stirred solution. The resultant mixture was stirred for
15
applied to prepare thin films of conducting polymers,
16
17
mesoporous silicas, and mesoporous metals, in which
monomers or reductants were transported and infiltrated
through the vapor phase onto the substrates.
2
1 h at 40 °C, producing a cloudy solution with a white precipitate.
We describe herein a novel synthetic route to prepare
hyperbranched polyamine-oxide hybrid materials through
vapor-phase transport. In this method, the small nitrogen-
containing heterocyclic monomers are transported into/onto
the solid supports in the vapor phase and subsequently
polymerization is initiated at the support surface. The amount
of polyamines formed on the supports can be affected by
several synthesis parameters including temperature, reaction
time, and the acidity of supports. In comparison to the
conventional liquid-phase method, this novel vapor-phase
method is anticipated to be more efficient in introduction of
polyamines into the structured contactors, and may prove to be
Subsequently, this mixture was heated statically at 100 °C for 24 h.
SBA-15 material was isolated by filtration and washed with a copious
amount of deionized water. The obtained white solid was dried at 75
°C overnight. The organic template was removed by calcination at 550
°C for 6 h with an intermediate step at 150 °C for 2 h (a heating rate
−1
of 1 °C min ).
Incorporation of Aluminum into SBA-15. The acidity of SBA-
1
5 material was modified by postsynthetic grafting of aluminum
20
species onto the SBA-15 surface. Briefly, the aluminate solution was
prepared by dissolving 0.26 g of aluminum isopropoxide in 60 mL of
0
6
.03 M hydrochloric acid. After being stirred at room temperature for
h, to the aluminate solution was added 1.5 g of the calcined SBA-15.
The suspension was stirred at room temperature for 18 h. The
suspension was filtered and washed with deionized water. The
recovered solid was dried at 75 °C overnight and then calcined at 550
°C for 6 h with an intermediate step at 150 °C for 2 h (a heating rate
a more scalable technique to create important CO adsorbents.
2
To the best of our knowledge, polyamine-oxide hybrid
materials are successfully prepared here via the vapor-phase
method for the first time.
−
1
of 1 °C min ).
6
14
dx.doi.org/10.1021/cm303931q | Chem. Mater. 2013, 25, 613−622

Chemistry of Materials
Article
Table 1. Physical and Chemical Characteristics of the Hyperbranched Poly(ethylenimine)-Oxide Hybrids Prepared on SBA-15
amine content
(mmol g−
1
)
synthesis condition
temp.
(°C)
time
(h)
organic loading
S
V
percent occupancy
remaining V
p
3
g
BET
Total
a
b
b
c
(m g−1)d
2
(cm g−1)e
3
f
−1
sample
ratio
(wt %)
NTGA
NEA
(%)
(cm g
)
support
S-70C-1-24h
S-70C-2-24h
S-70C-3-24h
S-70C-4-24h
S-23C-4-24h
S-23C-4-168h
S-50C-4-24h
S-50C-4-48h
S-60C-4-24h
S-80C-4-24h
70
70
70
70
23
23
50
50
60
80
1
2
3
4
4
4
4
4
4
4
24
24
26.6
31.1
38.6
41.5
15.7
27.2
31.7
39.9
35.0
42.3
6.19
7.23
8.98
9.66
3.66
6.33
7.38
9.27
8.15
9.83
5.51
6.61
8.43
8.97
3.32
5.74
6.98
8.55
7.60
9.19
214
182
40
0.42
0.33
0.11
0.07
0.62
0.42
0.32
0.08
0.15
32.6
40.6
56.5
63.8
16.8
33.6
41.7
59.6
48.5
65.8
0.04
0.05
0.19
0.17
0.01
0.03
0.04
0.20
0.22
0.14
24
24
22
24
372
233
174
30
168
24
48
24
67
24
21
0.07
a
b
c
d
e
The mass ratio of aziridine/support. Determined by TGA. Determined by elemental analysis. Surface area calculated by the BET method. Total
f
3
pore volume at P/P = 0.99. Calculated by dividing the estimated volume of polyamine (in cm of polymer per gram of support) by the total pore
volume of bare support, assuming a polyamine density of 1.05 g cm . Remaining pore volume = [pore volume of bare support] − [pore volume of
hybrid material] − [volume of polyamine].
o
−3
g
5c,8b
Preparation of Hyperbranched Polyamine-Oxide Hybrids
via Vapor-Phase Transport. For the laboratory scale experiment,
polymerization of aziridine on the oxide supports via vapor-phase
transport was carried out in a 15-mL glass pressure tube. Typically, the
support materials were hand-ground by a mortar and pestle and then
dried at 105 °C for at least 48 h. About 0.15 g of the well ground and
dried supports was added into the pressure tube. A small glass test
tube (12 × 75 mm, VWR) containing a different amount of aziridine
was then placed inside the pressure tube (see Supporting Information,
Figure S1). The pressure tube was closed tightly and heated to a
desired temperature for a specified period of reaction time. The
reaction was quenched by adding the pressure tube into an ice bath.
The solid sample was washed with excess amounts of methanol (100
mL × 3) and acetone (100 mL × 3) to remove the physisorbed
aziridine and polyamines from the sample surface and then recovered
by filtration. The resulting solid was dried under high vacuum at room
temperature overnight. The resulting hyperbranched poly-
1530 instrument. Organic loadings were determined by thermogravi-
metric analysis (TGA) using a Netzsch STA409 instrument. Samples
−1
were heated under a mixed gas stream of air (90 mL min ) and
−1
−1
nitrogen (30 mL min ) with a heating rate of 10 °C min . The
amine loading was estimated from relative weight loss in the range of
22
1
50−600 °C. The organic contents were first estimated by TGA, and
these values were then verified by elemental analyses (EA). Inductively
Coupled Plasma-Optical Emission Spectroscopy (ICP-OES) was used
to determine the aluminum content in the supports. EA and ICP-OES
experiments were performed by Columbia Analytical Services
(
Tucson, AZ). Samples were dried under vacuum prior to analysis
2
7
at 80 °C. Al solid-state magic-angle spinning (MAS) nuclear
magnetic resonance (NMR) experiments were carried out on a Bruker
DSX 300 spectrometer at a frequency of 78.2 MHz. The sample was
spun at 5 kHz with a single pulse of π/6 and a recycle delay of 0.5 s.
Solution-state NMR spectra were recorded on a Varian Mercury
Vx400 spectrometer. Inversely gated solution-state 13C NMR was used
to determine the ratios of primary, secondary, and tertiary amines. A
proton decoupled technique with a relaxation time of 4 s allows for
(
ethylenimine)-oxide hybrids were denoted as W-XC-Y-Zh, where W
represents the oxide supports, X represents the synthesis temperature,
Y represents the aziridine-to-support mass ratio, and Z represents the
synthesis time. The oxide supports included SBA-15, Al-grafted SBA-
23
quantitative analysis of the cleaved polymers. Samples were prepared
1
5, acidic alumina, basic alumina, and neutral alumina, which are
referenced as S, AS, AA, BA, and NA, respectively. For example, S-
0C-4-24h refers to the hybrid material prepared on the SBA-15
in D O with a drop of chromium(III) acetylacetonate in deuterated
2
dimethylsulfoxide (DMSO) to aid relaxation and minimize interfer-
ence from residual silica or base after cleavage. Spectra were recorded
using 12,000 scans on a Bruker DRX 500 spectrometer. Molecular
weights of cleaved polyamines were estimated by aqueous gel
permeation chromatography with a system composed of a Shimadzu
LC-20AD pump, a Shimadzu RID-10A RI detector, a Shimadzu SPD-
20A UV detector, a Shimadzu CTO-20A column oven, and Tosoh
Bioscience TSKgel PWXL Guard, Viscotek Viscogel G6000 and
G4000 columns mounted in series. The mobile phase consisted of 0.5
M acetic acid and 0.3 M Na SO , and the flow rate was maintained at
5
support with the aziridine/SBA-15 mass ratio of 4 at 50 °C for 24 h.
The hyperbranched poly(propylenimine)-oxide hybrids were also
prepared using the same procedures from azetidine monomers.
Removal of Polyamines from Silica Supports. The supported
polyamines were cleaved from the solid support by alkali treatment.
This removal procedure would be universal for any supports that can
be dissolved in alkaline media. About 0.5 g of supported polyamines
was dispersed in 100 mL of deionized water. Then, 35 g of potassium
hydroxide was added to the dispersion. The resulting mixture was
stirred at 50 °C for 24 h, after which the support was degraded into
soluble species. At least 70 g of water was removed by rotary
evaporation at about 60 °C. The remaining solution was kept in a
freezer overnight. The polyamines appearing as viscous liquid were
phase-separated as an upper layer of the mixture and mechanically
recovered by a spatula for further analyses.
2
4
−1
0
.4 mL min .
CO Adsorption Measurement. A TA Q500 thermogravimetric
2
analyzer was used to assess CO2 adsorption capacities under
anhydrous conditions with 10% CO2 balanced with helium. The
hybrid materials were loaded in a platinum pan and pretreated under a
helium atmosphere at 110 °C for 3 h with a heating rate of 5 °C min−
1
Characterization. Powder X-ray diffraction (XRD) patterns were
collected on a Philips X′pert diffractometer using Cu Kα radiation.
Nitrogen adsorption−desorption was performed on a Micromeritics
TriStar II 3020 at −196 °C. Before the measurement, the samples
were degassed at 110 °C under vacuum for at least 8 h. Pore size
distribution was calculated by a nonlocal density functional theory
to remove any preadsorbed CO and water, following by cooling to 25
2
−1
°
C with a rate of 5 °C min where they were held at this temperature
for 1 h to stabilize the sample weight and temperature before
introducing the CO -containing gas. Adsorption experiments were
2
started by exposing the samples to a dry CO gas mixture for 6 h. The
2
2
1
(
NL-DFT) by using the silica cylindrical pore, equilibrium model.
adsorption capacities were calculated from the weight gain after the
Scanning electron microscope (SEM) images were taken from an LEO
CO exposure.
2
6
15
dx.doi.org/10.1021/cm303931q | Chem. Mater. 2013, 25, 613−622

Chemistry of Materials
Article
RESULTS AND DISCUSSION
■
Effects of Synthesis Parameters. The solid support and
liquid aziridine were put separately inside the reaction vessel,
which was heated at a prescribed temperature for a fixed time.
After synthesis, no accumulated liquid aziridine was found in
the solid support container, suggesting that the solid support
was contacted directly with aziridine in the vapor phase only,
not liquid aziridine. The recovered solids were washed
extensively with organic solvents to remove physisorbed
monomer or polymer species (600 mL per 0.15 g of initial
solid support). To study the feasibility of this novel method,
poly(ethylenimine)s were prepared on SBA-15 mesoporous
silica as the solid support (see Supporting Information, Figure
24
S2 for the detailed characterization of SBA-15) under varied
synthesis parameters, as summarized in Table 1. We first
studied the effects of the aziridine-to-support mass ratio by
fixing the synthesis temperature and time at 70 °C and 24 h,
respectively. This temperature is higher than the boiling point
9
b
of aziridine (56.7 °C); therefore, aziridine should be
vaporized inside the vessel efficiently. The mass of poly-
(
ethylenimine) formed increased as the mass ratio was
increased, with the highest organic loading reaching 41.5 wt
at an aziridine/support mass ratio of 4 (see Table 1). Powder
%
XRD patterns of the obtained hybrids are shown in Figure 1a.
Figure 2. (a,c) Nitrogen sorption isotherms (filled and empty symbols
representing adsorption and desorption branches, respectively) and
(b,d) the corresponding histograms of NL-DFT pore size distribution
of the resulting hybrids. In (a), the isotherms of S-70C-2-24h, S-70C-
3
−1
3
-24h, and S-70C-4-24h were offset by 300, 550, and 650 cm g STP,
respectively; while in (c), the isotherms of S-50C-4-24h, S-60C-4-24h,
S-70C-4-24h, and S-80C-4-24h were offset by 400, 650, 750, and 850
3
−1
cm g STP, respectively.
2
−1
decreased from 214 m g for the S-70C-1-24h sample to 22
2
−1
m g for the S-70C-4-24h sample, while, the total pore
3
−1
volumes (at P/P = 0.99) decreased from 0.42 to 0.07 cm g
o
for the S-70C-1-24h and S-70C-4-24h samples, respectively.
Nonlocal density functional theory (NL-DFT) was employed
to estimate the pore size distributions of the materials (Figure
2b). It can be seen that the pore diameters of the obtained
hybrid materials shifted to smaller sizes, and the pore size
distributions became broader when the organic loadings
increased. Comparing with the parent SBA-15 support, the
reductions of the BET surface area, total pore volume, and pore
diameter indicate that at least some portions, if not all of the
polyamines, were occluded inside the pore space of the SBA-15
support. SEM images of the parent SBA-15, and the resulting
hybrid materials (Supporting Information, Figure S3) reveal
that the morphologies and external surfaces of all the samples
were similar, thereby further supporting the notion that the
majority of polyamines were in the particle interior and not on
the external surface of the hybrid materials.
In the next set of experiments, the hybrid materials were
prepared with an aziridine/support ratio of 4 at different
temperatures for 24 h. As listed in Table 1, the synthesis
temperature influenced the organic content of the obtained
materials; a higher temperature resulted in the higher organic
loadings. The highest loading obtained at the synthesis
temperature of 80 °C is slightly higher than the highest loading
of the materials synthesized by the conventional liquid-phase
Figure 1. Powder XRD patterns of the resulting hyperbranched
poly(ethylenimine)-oxide hybrids synthesized at different aziridine-to-
support ratios (a) and at different synthesis temperatures (b).
The S-70C-1-24h sample clearly exhibits XRD peaks with
(
100), (110), and (200) reflections, which are characteristic of
the two-dimensional (2D) hexagonal mesostructure of the
SBA-15 support. The intensity of the XRD peaks decreased as
the ratios increased, consistent with the increasing amount of
polyamines on the support surface. This decrease in peak
intensity is a result of a reduction in electron density contrast
between the mesopore spaces and the silica pore walls after
polymerization, caused by the growth of polyamines in the pore
interior. These results also suggest that the ordered
mesostructure of the SBA-15 supports was preserved after
functionalization.
The porous characteristics of the resulting hybrids were
studied by nitrogen adsorption measurements. Figure 2a
depicts nitrogen adsorption−desorption isotherms of the
hybrid materials, clearly showing IUPAC type IV isotherms
with hysteresis indicative of mesoporous materials, especially
for the samples with lower organic loadings. The apparent
Brunauer−Emmett−Teller (BET) specific surface areas
6
16
dx.doi.org/10.1021/cm303931q | Chem. Mater. 2013, 25, 613−622

Chemistry of Materials
Article
Table 2. Physical and Chemical Characteristics of the Hyperbranched Poly(ethylenimine)-Oxide Hybrids Prepared on the Al-
a
Grafted SBA-15 and Alumina Supports
amine content
mmol g−
1
)
(
synthesis temp.
organic loading
S
V
percent occupancy
remaining V
p
3
g
BET
Total
b
b
c
(m g−1)d
2
(cm g−1)e
3
f
−1
sample
(°C)
(wt %)
NTGA
NEA
(%)
(cm g
)
support
Samples Prepared on the Al-Grafted SBA-15 Supports
AS-23C-4-24h
AS-50C-4-24h
AS-60C-4-24h
AS-70C-4-24h
AS-80C-4-24h
23
50
60
70
80
16.0
32.3
36.8
43.5
46.8
3.73
7.52
3.36
7.24
307
57
19
17
16
0.54
0.13
0.06
0.05
0.05
20.4
51.1
62.4
82.4
94.2
0.00
0.18
0.18
0.00
8.57
8.69
10.12
10.89
9.53
10.36
−0.12
Samples Prepared on the Alumina Supports
h
AA-50C-4-24h
NA-50C-4-24h
BA-50C-4-24h
50
50
50
15.2
4.8
3.54
1.12
0.91
n.d.
n.d.
n.d.
24
128
139
0.06
0.21
0.21
74.4
20.1
15.4
c
−0.01
−0.03
−0.02
h
h
3.9
a
b
The aziridine/support mass ratio and synthesis time were set at 4 and 24 h, repectively. Determined by TGA. Determined by elemental analysis.
Surface area calculated by the BET method. Total pore volume at P/P = 0.99. Calculated by dividing the estimated volume of polyamine (in cm
of polymer per gram of support) by the total pore volume of bare support, assuming the polyamine density of 1.05 g cm . Remaining pore volume
[pore volume of bare support] − [pore volume of hybrid material] − [volume of polyamine].
d
e
f
3
o
−3
g
5c,8b h
=
Not determined.
7
,8
method reported previously, suggesting that this vapor-phase
method can be as useful as the liquid-phase method for the
introduction of hyperbranched polyamines into the mesopo-
rous supports. Interestingly, materials with a significant organic
content can also be prepared even at room temperature. The
synthesis time also affected the organic loading, with a longer
reaction time resulting in higher organic loading (e.g., S23C-4-
polyamine loading. This is in contrast with the previously
8b
reported materials prepared by liquid-phase reaction, where
some microporosity remained at low amine loadings but
−1
disappeared at amine loadings higher than 5.3 mmol N g .
Under the present conditions, polyamines prepared via the
vapor-phase method seem to be reactively grown on the oxide
surface, and the micropores of the support can accommodate
only a small amount of polyamines before the micropores are
closed or blocked by the formed polyamines.
24h versus S23C-4-168h).
Powder XRD patterns of the obtained hybrids shown in
Figure 1b reveal a 2D hexagonal mesostructure, suggesting that
the ordered mesostructure of the supports was retained after
the functionalization. Nitrogen adsorption−desorption iso-
therms and the corresponding NL-DFT pore size distributions
of the obtained hybrid materials prepared at different
temperatures are shown in Figure 2c and d, respectively. The
BET surface areas and total pore volumes were calculated to
The “remaining pore volume” parameter was calculated for
the simple assessment of pore closing or blocking as the
normalized volume per gram of support by subtracting the pore
volume of hybrid material and the estimated volume of
polyamine from the pore volume of bare oxide support (i.e.,
[remaining pore volume] = [pore volume of bare support] −
[pore volume of hybrid material] − [volume of grown
polyamine]). It was assumed that the polyamines were not
grown in the micropores of support and that the polyamine
2
−1
3
−1
vary from 372 to 21 m g and from 0.62 to 0.07 cm g ,
respectively, depending on the polyamine loadings. Again, the
NL-DFT pore diameters shifted to smaller sizes and the pore
size distributions became broader when the organic loadings
increased. The decreases in BET surface area, total pore
volume, and pore diameter suggest that the polyamines are
formed in the interiors of the SBA-15 support, as noted above.
In the hybrid materials with low organic loadings (up to
about 30 wt %), the hystereses were essentially parallel for the
adsorption and desorption branches. In addition, the NL-DFT
pore size distributions were relatively narrow. These data
suggest that most of polyamines were well-dispersed inside the
pores of the supports and likely grown via a layer-by-layer
fashion. On the contrary, in the hybrid materials with higher
organic loadings, the broad NL-DFT pore size distributions
suggest that polyamines were likely ill-dispersed inside the
pores and some polymers likely formed near/in pore windows,
leading to some pore blocking. This phenomenon is also
supported by the fact that the total pore volume approached
zero at higher organic loading but the occupancy of polyamines
reached only about 70% of the total porosity of the bare
support (see Table 1).
5c,8b
density is equal to the bulk density of poly(ethylenimine).
In an ideal case where all the polyamines are occluded inside
the pores, the remaining pore volume should equal zero. A
positive value of this parameter means there are some pore
spaces inaccessible to nitrogen adsorption at −196 °C, while a
negative value suggests that parts of polyamine are likely
located on the particle exteriors. It is also possible to have both
inaccessible pore spaces and polyamines located on the particle
exteriors that can result in positive or negative values,
depending on the relative magnitudes of each effect. As
shown in Table 1, it is likely that at low organic loadings most
of the polyamines were located inside the mesopores of the
hybrid materials without the pore blocking, as the remaining
pore volumes were close to zero, while at higher organic
loadings, the remaining pore volumes became more positive,
suggesting some pore blocking. In both cases, the remaining
pore volumes suggest that only a small amount, if any, of
polyamines were grown on the particle exteriors, consistent
with SEM observations.
The thermal stability of the supported polyamines was
examined by TGA in the presence of gaseous oxygen. Thermal
decomposition or oxidation of the polyamines initiated at about
200 °C with the peak of the first-derivative TG weight (DTG)
It is noteworthy that the micropore volumes of all the hybrid
materials determined by the t-plot method became zero or
negligible after vapor-phase functionalization, irrespective of the
6
17
dx.doi.org/10.1021/cm303931q | Chem. Mater. 2013, 25, 613−622

Chemistry of Materials
Article
Figure 3. (a) Powder XRD patterns, (b) nitrogen adsorption−desorption isotherms (filled and empty symbols representing adsorption and
desorption branches, respectively), and (c) the corresponding histograms of NL-DFT pore size distribution of the hyperbranched
poly(ethylenimine)-oxide hybrids prepared on the Al-grafted SBA-15 support. The isotherms of AS-50C-4-24h, AS-60C-4-24h, AS-70C-4-24h,
3
−1
and AS-80C-4-24h were offset by 400, 500, 600, and 700 cm g STP, respectively.
observed at 220 °C (see Supporting Information, Figure S4).
These peaks are about 40 °C higher than those of materials
prepared by impregnation of poly(ethylenimine) into SBA-
Al-grafted SBA-15 is more acidic than the pure silica parent
material.
Powder XRD patterns of the obtained samples are shown in
5
g
15, suggesting that the resulting supported polyamines are
Figure 3a. Similar to the case of SBA-15, the patterns arising
from the 2D hexagonal mesostructure were observed, indicating
the ordered mesostructure remained in the hybrid materials
after the polymerization. Figure 3b and 3c depict the nitrogen
adsorption−desorption isotherms and the corresponding NL-
DFT pore size distributions of these materials, respectively. At
low loadings, parallel hysteresis loops were observed, suggesting
the grown polyamines were homogeneously dispersed over the
interior of the supports. As the polyamine loadings increased,
the NL-DFT pore diameters shifted to smaller pore sizes and
the pore size distributions became broader, akin to the silica
SBA-15-supported hybrid materials. Quantification of micro-
porosity by the t-plot method suggests that the micropores
disappear or are inaccessible in the resulting hybrids. The
calculated BET surface areas and total pore volumes ranged
more thermally stable than in the case of physically bound
poly(ethylenimine)s. As is generally known, Si−O−C linkages
are easily hydrolyzed, particularly under acidic and basic
2
5
conditions. However, the oxide-supported, hyperbranched
polyamine materials have been previously demonstrated to be
7
,8
stable under several conditions containing water vapor,
although the Si−O−C linkages may be cleaved under harsher
environments. The stability against hydrolysis of these materials
under hydrous conditions is likely because the Si−O−C
linkages are protected from water by bulky hyperbranched
25
polymers (steric effect).
Effects of Support Acidity. SBA-15 mesoporous silica
supports were functionalized with aluminum by a postgrafting
2
0
method to investigate the effects of support acidity on
supported polyamine synthesis via the vapor-phase transport.
The detailed characterization of the Al-grafted SBA-15 support
2
−1
3
−1
from 307 to 16 m g and from 0.54 to 0.05 cm g ,
respectively. Interestingly, the maximum occupancy of poly-
amines reached 94% of the total porosity of the supports, which
is much higher than in the case of SBA-15 supports (vide
supra).
26
is shown in the Supporting Information, Figure S5. The
incorporation of aluminum was confirmed by ICP-OES, with
the material having a Si/Al atomic ratio of 29. The nature of the
2
7
aluminum species was determined by solid-state Al MAS
NMR. The spectrum (see Supporting Information, Figure S5b)
reveals that Al atoms are located primarily in tetrahedral
positions, although nonframework, octahedral Al species also
exist.
Plots of total pore volume versus amine content and
polyamine occupancy are depicted in Figure 4a and b,
respectively. There are clearly two different regions of the
plots. At low polyamine loadings, the pore volumes decreased
linearly as the content and occupancy of polyamines increased
The hybrid materials prepared on the Al-grafted SBA-15
supports with an aziridine/support ratio of 4 at different
temperatures for 24 h are summarized in Table 2. Similar to the
hybrids prepared on SBA-15 silica supports, the materials
obtained at higher synthesis temperature possessed higher
organic loadings, with the highest value obtained being 46.8 wt
%
, which is to our knowledge the highest loading reported thus
7
,8,10
far for this class of material.
Under the same synthesis
conditions, the hybrid materials prepared on the Al-grafted
SBA-15 supports have higher polyamine loadings than those
prepared on the pure silica SBA-15, with the different loadings
being more pronounced at higher synthesis temperatures,
suggesting that the acidity of the supports can enhance the
polymerization of the cyclic amines. This was expected, as the
polymerization reaction is known to be acid-catalyzed and the
Figure 4. Total pore volume of the hyperbranched poly-
(ethylenimine)-oxide hybrids prepared on the SBA-15 (filled symbols)
and the Al-grafted SBA-15 (empty symbols) supports as function of
(a) amine contents and (b) percent occupancy.
6
18
dx.doi.org/10.1021/cm303931q | Chem. Mater. 2013, 25, 613−622

Chemistry of Materials
Article
up to certain values (about 9 mmol N g−1 for amine content
and 60% occupancy) before the pore spaces (or pore openings)
became saturated. After passing this point, decreases in pore
volume were very small as additional polymers were added. In
the former region, most polyamines seem to be grown quasi-
homogeneously and none of pores appear blocked or closed. In
the latter part where the pore apertures were almost closed,
additional polyamines that were formed likely attached near the
pore apertures or outer surface and accordingly some pore
blocking occurred.
polymerization of aziridines, as Lewis acids are known to
catalyze nucleophilic addition reactions, especially for sub-
stituted aziridines. For unsubstituted aziridine used here,
9
however, the initiation reaction for the ring-opening polymer-
ization is likely the formation of the cationic aziridinium
intermediates that are more easily formed with Brønsted acids.
At this stage, it remains unclear how supported Lewis acid sites
affect the ring-opening polymerization of aziridines under the
vapor-phase conditions investigated here.
Characteristics of the Supported Poly(ethylenimine).
The degree of polymerization and branching of the supported
polyamines were evaluated. The supported polyamines were
cleaved from their supports by dissolution of the silica
framework in concentrated alkali solution. The molecular
weights of the cleaved polyamines were determined by aqueous
gel permeation chromatography (size exclusion chromatog-
raphy) using commercial poly(ethylenimine)s as standards. As
reported in Table 3, the obtained polyamines possessed low
The effects of support acidity were further studied on
mesoporous activated alumina supports. The textural properties
of the acidic, neutral, and basic alumina supports used here are
2
7
nearly identical; therefore, the effect of surface acid/base
characteristics can be clearly elucidated. The poly-
(
ethylenimine)−alumina hybrid materials were synthesized
with an aziridine/support mass ratio of 4 at 50 °C for 24 h.
As summarized in Table 2, the polyamines were formed on the
acidic alumina support with about 75% occupancy whereas the
neutral and basic alumina could accommodate only small
amounts of polyamines, with the basic alumina having the
lowest loading. This also indicates that besides silica supports,
polyamines can be grown on other oxides such as alumina. In
general, the polymerization of these cyclic amines by vapor-
phase transport appears applicable to any support containing
acidic hydroxyl groups on their surface that can facilitate the
ring-opening reaction.
Table 3. Molecular Weights and Degrees of Branching of the
Cleaved Poly(ethylenimine)s
b
NTGA
molecular weight
(Da)
amine ratio
1°:2°:3°
(mmol g−1)
sample
a
S-23C-4-24h
S-50C-4-24h
S-60C-4-24h
S-70C-4-24h
S-80C-4-24h
AS-70C-4-24h
3.66
7.38
8.15
9.66
9.83
10.12
(100)
31:42:27
31:40:29
32:39:29
32:39:29
32:39:29
30:43:27
530
620
950
Figure 5 shows nitrogen adsorption−desorption isotherms
and the corresponding NL-DFT pore size distributions of the
1060
1960
a
This calculated value was much lower than those for the standard
polymers and should be considered only in a qualitative manner. Also,
this value seems to be lower than the detection limit in our
b
chromatography instrument. Represented as a percentage ratio of
primary: secondary: tertiary amines.
molecular weights, up to 1960 Da. These values are in sharp
contrast to the hyperbranched polyamines polymerized via the
standard liquid-phase reaction, which were previously reported
8
b,c
to have the molecular weights of 4500−6500 Da. Although
the molecular level details remain unclear, the polymerization
and termination rates of poly(ethylenimine)s grown via the
vapor-phase transport method are likely different from those
that occur in the liquid-phase polymerization. One possible
explanation is that the swelling of polymer may strongly affect
the growth of polymer, as the swelling would enhance the
exposure of polymer ends to attack the heterocyclic monomers,
thereby resulting in the elongation of the polymer chains. The
swelling may be suppressed under the vapor-phase conditions,
giving the lower molecular weight polyamines.
Figure 5. (a) Nitrogen adsorption−desorption isotherms (filled and
empty symbols representing adsorption and desorption branches,
respectively) and (b) the corresponding histograms of NL-DFT pore
size distribution of the resulting hybrids prepared on alumina supports.
The isotherms of NA-50C-4-24h and BA-50C-4-24h were offset by
3
−1
200 and 400 cm g STP, respectively.
Figure 6 shows the molecular weights of the cleaved
polyamines as a function of the amine content of the hybrid
materials. This set of hybrid materials was synthesized at the
same aziridine/support ratio and synthesis time. As discussed
above, hybrid materials with higher polyamine loadings were
obtained from experiments at higher synthesis temperatures.
Therefore, the observed parabolic increase in molecular weights
as the amine content and synthesis temperature increased
suggests that the higher synthesis temperature enhanced the
degree of polymerization. Moreover, the material prepared on
the Al-grafted SBA-15 support (i.e., AS-70C-4-24h) had a much
higher molecular weight than those prepared on the SBA-15
support, again indicating that the acidity of support promotes
resulting hybrid materials. The isotherms and pore size
distributions of the hybrid materials prepared on the neutral
and basic alumina supports were almost identical to those of
the bare supports (see Supporting Information, Figure S6). In
contrast, nitrogen adsorption drastically decreased for the
hybrid prepared on the acidic alumina support, with the NL-
DFT pore diameter being smaller than that of the parent
support. These results additionally support the notion that
support acidity strongly influences the formation of supported
polyamines via ring-opening polymerization. It should be noted
that the coexistence of Lewis acid sites in the Al-grafted SBA-15
and alumina supports may also affect the ring-opening
6
19
dx.doi.org/10.1021/cm303931q | Chem. Mater. 2013, 25, 613−622

Chemistry of Materials
Article
peak areas and are summarized in Table 3. The results are
slightly altered from the average ratio of 1°:2°:3° = 28:46:26 for
the supported polyamines prepared by the liquid-phase
8b
methods reported previously, with the hybrid materials
prepared here having slightly higher and lower fractions of
primary and secondary amines, respectively. The ratios persist
across various samples, indicating that the degree of polyamine
branching was not significantly impacted by the synthesis
conditions and supports.
CO Adsorption. CO adsorption measurement using the
2
2
resulting hyperbranched poly(ethylenimine)-oxide hybrids was
performed under dry conditions using a simulated flue gas
(
10% CO balanced with helium) to evaluate their performance
2
in comparison with similar materials prepared in liquid media
Figure 6. Relation between molecular weights of the cleaved
poly(ethylenimine)s and amine contents of the corresponding hybrid
materials prepared on the SBA-15 (filled symbols) and the Al-grafted
SBA-15 (empty symbols) supports.
reported previously. The pseudoequilibrium CO adsorption
2
capacities and amine efficiencies, defined as the number of
moles of CO captured per mole of active amines, of the
2
resulting materials are summarized in Table 4. It is noteworthy
the polymerization and subsequently leads to a higher degree of
polymerization.
Solution-state C NMR with inversely gated decoupling was
employed to quantify the degree of polymer branching. As
illustrated in Figure 7a, there are eight different carbon atom
Table 4. CO Adsoprtion Capacities under Dry Conditions
Using Simulated Flue Gas (10% CO2)
2
13
NTGA
capacity
amine efficiency
sample
(mmol g−1) (mmol CO g−1) (mol CO per mol N)
2
2
S-70C-1-24h
S-23C-4-168h
S-70C-2-24h
AS-50C-4-24h
S-70C-4-24h
6.19
6.33
7.23
7.52
9.66
0.69
0.73
0.93
0.70
0.43
0.11
0.12
0.13
0.09
0.04
Hyperbranched Poly(ethylenimine)-SBA-15 Hybrid Prepared by the Standard
8
Liquid-Phase Method
liquid-HAS
5.44
Poly(ethylenimine)-Impregnated SBA-15
7.40 0.65
0.58
0.11
0.09
5g
PEI/SBA-15
that under dry conditions only primary and secondary amines
are expected to react with CO to produce carbamates through
2
5
the formation of zwitterionic intermediates. In this mecha-
nism, an additional free base is required per mole of CO2
captured for deprotonating the zwitterions to form the
carbamates. As a result, the theoretical maximum amine
efficiency for poly(ethylenimine)-based materials under dry
conditions is varied from 0.35 to 0.5, depending on the amount
5
,8,12
Figure 7. (a) Eight different environments of carbon atoms in
of tertiary amines.
At low amine loadings, the CO2
13
poly(ethylenimine). (b) Inversely gated solution-state C NMR
spectra of the cleaved poly(ethylenimine)s from various hybrid
materials.
adsorption capacities and amine efficiencies of the resulting
hybrids increased as the amine content increased. At the higher
amine loadings in which some pore spaces were blocked or
closed (vide supra), however, the capacities and amine
efficiencies were lowered. This is likely because the CO2
molecules cannot easily diffuse to access all the amines in the
interior of these materials because of the presence of
polyamines grown near/in the pore mouths and the steric
hindrances caused by the previously formed (cross-linking)
environments in the poly(ethylenimine). Figure 7b shows the
inversely gated 1 C NMR spectra of the cleaved polyamines.
Note that Cr(acac) in DMSO-d was added to the sample
3
3
6
solution as a paramagnetic relaxation agent to attain high
resolution spectra, as it aids relaxation by shortening the carbon
T1 decay times and minimizes interference from residual silica
or base after the polymer cleavage from the support.
Unfortunately, the signal of the DMSO-d6 NMR solvent
somewhat overlaps with the signal from the carbon at the h
position. To determine the amine ratio, the peak area of the h-
carbon was assumed to be equal to that of the a-carbon,
because both carbons are present in the same fragment of the
5
carbamates. As can be seen, the poly(ethylenimine)-oxide
hybrid materials prepared here are as efficient in CO2
adsorption as the related materials prepared by the standard
liquid-phase method, as well as materials made by physically
impregnating SBA-15 with poly(ethylenimine), suggesting that
the vapor-phase method is an effective way to produce CO2-
adsorbing polyamines. It should be noted that the CO2
adsorption performance of this class of materials can be
substantially improved under wet conditions, as reported
23a
polyamine and their peak areas should then be identical. The
molar ratios of primary, secondary, and tertiary amines of the
cleaved polyamines were quantitatively analyzed from the NMR
5a,8
previously.
6
20
dx.doi.org/10.1021/cm303931q | Chem. Mater. 2013, 25, 613−622

Chemistry of Materials
Article
Hyperbranched Poly(propylenimine)-Oxide Hybrid.
Polyamines with different lengths of methylene spacers
between the amine groups may affect the material stability
and performance in some applications such as catalysis and
CO2 capture. A SBA-15-supported poly(propylenimine)
material was also prepared via the vapor-phase transport
using azetidine monomer (see Scheme 1) with an azetidine/
SBA-15 mass ratio of 4 at 80 °C for 24 h. The polyamine
loading was calculated from the TGA weight loss to be about
ASSOCIATED CONTENT
Supporting Information
Experimental setup, detailed characterizations of the SBA-15,
Al-grafted SBA-15, and alumina supports, FE-SEM images and
differential TGA curves of the hybrids, and XRD pattern,
■
*
S
2
f
differential TGA curve, N sorption isotherms and NL-DFT
2
pore size distribution of the hyperbranched poly-
(propylenimine)-oxide hybrid material. This material is
available free of charge via the Internet at http://pubs.acs.org.
15 wt %. The powder XRD pattern of the resulting material
(
see Supporting Information, Figure S7a) indicates that the 2D
AUTHOR INFORMATION
Corresponding Author
E-mail: cjones@chbe.gatech.edu.
Present Address
World Premier International (WPI) Research Center for
Materials Nanoarchitectonics (MANA), National Institute for
Materials Science, 1-1 Namiki, Tsukuba, Ibaraki, 305-0044
Japan.
■
hexagonal mesostructure of the material was maintained, as in
the case of materials made with aziridine. The DTG curve
depicted in the Supporting Information, Figure S7b, shows a
peak at 245 °C, which is 25 °C higher than the peaks seen in
the supported poly(ethylenimine) materials, suggesting the
supported poly(propylenimine) is thermally more stable than
its poly(ethylenimine) counterparts. Nitrogen adsorption−
desorption isotherms and the corresponding NL-DFT pore
size distribution are shown in the Supporting Information,
Figures S7c and S7d, respectively. The BET surface area and
*
†
Notes
The authors declare the following competing financial
interest(s): C.W.J. has a financial interest in Global Thermostat,
LLC.
2
−1
total pore volume were calculated to be 334 m g and 0.68
cm g , respectively. The resulting poly(propylenimine)
hybrid material had lower organic loading than the poly-
3
−1
ACKNOWLEDGMENTS
■
(
ethylenimine) material prepared under the same conditions,
We thank Global Thermostat, LLC and School of Chemical &
Biomolecular Engineering at Georgia Tech (through the New-
Vision Professorship awarded to C.W.J) for financial support.
most likely because azetidine has a lower vapor pressure (i.e.,
higher boiling point) and less ring strain than aziridine. The
loading of the poly(propylenimine) may be improved by
preparing materials at harsher conditions. Optimization of
synthesis conditions to attain higher poly(propylenimine)
loadings merits investigation in future work.
9
REFERENCES
1) (a) Hoffmann, F.; Cornelius, M.; Morell, J.; Fro
■
(
̈
ba, M. Angew.
Chem., Int. Ed. 2006, 45, 3216−3251. (b) Soler-Illia, G. J. A. A.;
Innocenzi, P. Chem.Eur. J. 2006, 12, 4478−4494. (c) Fujita, S.;
Inagaki, S. Chem. Mater. 2008, 20, 891−908. (d) Yamamoto, K.;
Tatsumi, T. Chem. Mater. 2008, 20, 972−980. (e) Hu, L.-C.; Shea, K.
J. Chem. Soc. Rev. 2011, 40, 688−695. (f) Linares, N.; Serrano, E.;
Rico, M.; Balu, A. M.; Losada, E.; Luque, R.; García-Martínez, J. Chem.
Commun. 2011, 47, 9024−9035.
CONCLUSIONS
■
Preparation of supported hyperbranched polyamine-oxide
hybrid materials via vapor-phase transport was demonstrated
for the first time. In contrast to the conventional liquid-phase
7
,8,10
method reported previously,
in which the supports are
(2) (a) Margelefsky, E. L.; Zeidan, R. K.; Davis, M. E. Chem. Soc. Rev.
2008, 37, 1118−1126. (b) Wang, Q.; Guerrero, V. V.; Ghosh, A.; Yeu,
S.; Lunn, J. D.; Shantz, D. F. J. Catal. 2010, 269, 15−25. (c) Kubota,
Y.; Yamaguchi, H.; Yamada, T.; Inagaki, S.; Sugi, Y.; Tatsumi, T. Top.
Catal. 2010, 53, 492−499. (d) Das, S.; Asefa, T. ACS Catal. 2011, 1,
dispersed in an organic solvent (typically toluene) containing
the cyclic monomers, in the vapor-phase method, the liquid
monomers and the solid supports are placed separately in an
enclosed environment. The cyclic amine monomers were then
transported onto the support surface in the vapor-phase by
heating the vessel containing the monomer and support. Ring-
opening polymerization then occurs on the support surface,
resulting in polyamines covalently tethered to the supports.
The data shown here suggest that this method can be carried
out over a wide range of preparation conditions, with different
oxide supports, and with different cyclic amine monomers.
Compared to related materials reported thus far, a hybrid
material with the highest poly(ethylenimine) loading was
achieved by the vapor-phase method. Interestingly, the
resulting supported poly(ethylenimine)s possessed lower
molecular weights than the previously reported materials
prepared by the liquid-phase method. The acidity of supports
clearly promoted the polymerization, with more acidic supports
resulting in the hybrid materials with higher polyamine loadings
and degrees of polymerization. This vapor-phase method is
anticipated to effectively introduce covalently bound poly-
amines onto structured contactors such as tubular membranes
and monoliths, thereby facilitating possible use in large scale
applications in a scalable manner.
5
02−510. (e) Fihri, A.; Bouhrara, M.; Patil, U.; Cha, D.; Saih, Y.;
Polshettiwar, V. ACS Catal. 2012, 2, 1425−1431. (f) Brunelli, N. A.;
Didas, S. A.; Venkatasubbaiah, K.; Jones, C. W. J. Am. Chem. Soc. 2012,
1
34, 13950−13953.
(3) (a) Yoshitake, H.; Yokoi, T.; Tatsumi, T. Chem. Mater. 2003, 15,
1713−1721. (b) Yoo, S.; Lunn, J. D.; Gonzalez, S.; Ristich, J. A.;
Simanek, E. E.; Shantz, D. F. Chem. Mater. 2006, 18, 2935−2942.
(
(
(
1
c) Walcarius, A.; Mercier, L. J. Mater. Chem. 2010, 20, 4478−4511.
d) Yoshitake, H. J. Mater. Chem. 2010, 20, 4537−4550.
4) (a) Gu, J.; Fan, W.; Shimojima, A.; Okubo, T. Small 2007, 3,
740−1744. (b) Budi Hartono, S.; Qiao, S. Z.; Jack, K.; Ladewig, B. P.;
Hao, Z.; Lu, G. Q. Langmuir 2009, 25, 6413−6424. (c) Tran, D. N.;
Balkus, K. J., Jr. ACS Catal. 2011, 1, 956−968. (d) Suteewong, T.; Sai,
H.; Cohen, R.; Wang, S.; Bradbury, M.; Baird, B.; Gruner, S. M.;
Wiesner, U. J. Am. Chem. Soc. 2011, 133, 172−175.
(
1
5) (a) Bollini, P.; Didas, S. A.; Jones, C. W. J. Mater. Chem. 2011, 21,
5100−15120. (b) Chaikittisilp, W.; Lunn, J. D.; Shantz, D. F.; Jones,
C. W. Chem.Eur. J. 2011, 17, 10556−10561. (c) Chaikittisilp, W.;
Kim, H.-J.; Jones, C. W. Energy Fuels 2011, 25, 5528−5537. (d) Wang,
L.; Yang, R. T. J. Phys. Chem. C 2011, 115, 21264−21272. (e) Didas, S.
A.; Kulkarni, A. R.; Sholl, D. S.; Jones, C. W. ChemSusChem 2012, 5,
2058−2064. (f) Sayari, A.; Heydari-Gorji, A.; Yang, Y. J. Am. Chem.
Soc. 2012, 134, 13834−13842. (g) Kuwahara, Y.; Kang, D.-Y.;
6
21
dx.doi.org/10.1021/cm303931q | Chem. Mater. 2013, 25, 613−622

Chemistry of Materials
Article
3
−1
Copeland, J. R.; Brunelli, N. A.; Didas, S. A.; Bollini, P.; Sievers, C.;
Kamegawa, T.; Yamashita, H.; Jones, C. W. J. Am. Chem. Soc. 2012,
(at P/P = 0.99) of 0.89 cm g , and micropore volume (determined
o
3
−1
by the t-plot method) of 0.07 cm g .
1
(
(
34, 10757−10760.
6) (a) Lunn, J. D.; Shantz, D. F. Chem. Mater. 2009, 21, 3638−3648.
b) Subra, G.; Mehdi, A.; Enjalbal, C.; Amblard, M.; Brunel, L.; Corriu,
R.; Martinez, J. J. Mater. Chem. 2011, 21, 6321−6326.
7) (a) Rosenholm, J. M.; Penninkangas, A.; Linden
Commun. 2006, 3909−3911. (b) Rosenholm, J. M.; Linden
Mater. 2007, 19, 5023−5034. (c) Rosenholm, J. M.; Duchanoy, A.;
Linden, M. Chem. Mater. 2008, 20, 1126−1133.
8) (a) Hicks, J. C.; Drese, J. H.; Fauth, D. J.; Gray, M. L.; Qi, G.;
(27) Acidic, neutral, and basic activated aluminas were obtained from
a commercial source. These materials have almost identical textural
2
−1
properties with BET specific surface areas of 143, 154, and 164 m g
2
−1
(the values in catalog are all 150 m g ) and total pore volumes (at P/
3
−1
(
́
, M. Chem.
, M. Chem.
P = 0.99) of 0.23, 0.24, and 0.25 cm g , respectively.
o
́
́
(
Jones, C. W. J. Am. Chem. Soc. 2008, 130, 2902−2903. (b) Drese, J. H.;
Choi, S.; Lively, R. P.; Koros, W. J.; Fauth, D. J.; Gray, M. L.; Jones, C.
W. Adv. Funct. Mater. 2009, 19, 3821−3832. (c) Drese, J. H.; Choi, S.;
Didas, S. A.; Bollini, P.; Gray, M. L.; Jones, C. W. Microporous
Mesoporous Mater. 2012, 151, 231−240.
(
9) (a) Padwa, A. In Comprehensive Heterocyclic Chemistry III;
Katritzky, A. R., Ramsden, C. A., Scriven, E. F. V., Taylor, R. J. K., Eds.;
Elsevier: Amsterdam, The Netherlands, 2008; Vol. 1, pp 1−105.
(
b) Dermer, O. C.; Ham, G. E. Ethylenimine and Other Aziridines;
Academic Press: New York, 1969. (c) Goethals, E. J.; Schacht, E. H.;
Bogaert, Y. E.; Ali, S. I.; Tezuka, Y. Polym. J. 1980, 12, 571−581.
(
d) Hu, X. E. Tetrahedron 2004, 60, 2701−2743.
(
10) (a) Kim, H. J.; Moon, J. H.; Park, J. W. J. Colloid Interface Sci.
000, 227, 247−249. (b) Kim, C. O.; Cho, S. J.; Park, J. W. J. Colloid
Interface Sci. 2003, 260, 374−378.
11) Drese, J. H.; Talley, A. D.; Jones, C. W. ChemSusChem 2011, 4,
79−385.
12) Choi, S.; Drese, J. H.; Eisenberger, P. M.; Jones, C. W. Environ.
Sci. Technol. 2011, 45, 2420−2427.
13) (a) Kulkarni, A. R.; Sholl, D. S. Ind. Eng. Chem. Res. 2012, 51,
631−8645. (b) Bollini, P.; Brunelli, N. A.; Didas, S. A.; Jones, C. W.
Ind. Eng. Chem. Res. 2012, 51, 15145−15152.
14) (a) Lively, R. P.; Chance, R. R.; Kelley, B. T.; Deckman, H. W.;
Drese, J. H.; Jones, C. W.; Koros, W. J. Ind. Eng. Chem. Res. 2009, 48,
314−7324. (b) Gebald, C.; Wurzbacher, J. A.; Tingaut, P.;
Zimmermann, T.; Steinfeld, A. Environ. Sci. Technol. 2011, 45,
101−9108.
15) (a) Winther-Jensen, B.; West, K. Macromolecules 2004, 37,
538−4543. (b) O’Connell, C. D.; Higgins, M. J.; Nakashima, H.;
Moulton, S. E.; Wallace, G. G. Langmuir 2012, 28, 9953−9960.
16) (a) Nishiyama, N.; Tanaka, S.; Egashira, Y.; Oku, Y.; Ueyama, K.
2
(
3
(
(
8
(
7
9
(
4
(
Chem. Mater. 2003, 15, 1006−1011. (b) Tanaka, S.; Nishiyama, N.;
Oku, Y.; Egashira, Y.; Ueyama, K. J. Am. Chem. Soc. 2004, 126, 4854−
4858.
(
17) Yamauchi, Y.; Takai, A.; Komatsu, M.; Sawada, M.; Ohsuna, T.;
Kuroda, K. Chem. Mater. 2008, 20, 1004−1011.
18) Causey, D. H.; Mays, R. P.; Shamblee, D. A.; Lo, Y. S. Synth.
Commun. 1988, 18, 205−211.
19) Zhao, D.; Huo, Q.; Feng, J.; Chmelka, B. F.; Stucky, G. D. J. Am.
Chem. Soc. 1998, 120, 6024−6036.
20) Li, Q.; Wu, Z.; Tu, B.; Park, S. S.; Ha, C.-S.; Zhao, D.
Microporous Mesoporous Mater. 2010, 135, 95−104.
21) Ravikovitch, P. I.; Neimark, A. V. J. Phys. Chem. B 2001, 105,
817−6823.
22) Brunelli, N. A.; Venkatasubbaiah, K.; Jones, C. W. Chem. Mater.
012, 24, 2433−2442.
23) (a) St Pierre, T.; Geckle, M. J. Macromol. Sci., Chem. 1985, 22,
77−887. (b) von Harpe, A.; Petersen, H.; Li, Y.; Kissel, T. J.
Controlled Release 2000, 69, 309−322.
24) The bare SBA-15 support used entirely in the present study has
(
(
(
(
6
(
2
(
8
(
2
−1
a BET surface area of 920 m g , total pore volume (at P/P = 0.99)
of 1.07 cm g , and micropore volume (determined by the t-plot
o
3
−1
3
−1
method) of 0.11 cm g .
25) Brinker, C. J.; Scherer, G. W. Sol−Gel Science: The Physics and
Chemistry of Sol−Gel Processing; Academic Press: San Diego, CA, 1990.
26) The Al-grafted SBA-15 support used in the present study
(
(
2
−1
possessed a BET specific surface area of 715 m g , total pore volume
6
22
dx.doi.org/10.1021/cm303931q | Chem. Mater. 2013, 25, 613−622