Synthesis and gas permeation properties of various Si-containing poly(diarylacetylene)s and their desilylated membranes

Article abstract of DOI:10.1016/j.polymer.2010.02.015

Diarylacetylene monomers having trimethylsilyl groups and other substituents (substituted biphenyl, 1a and 1b; trimethylsilylmethylphenyl, 1c-e) were synthesized and polymerized with TaCl₅-n-Bu₄Sn catalyst to produce the correspondin

Full text of DOI:10.1016/j.polymer.2010.02.015

Polymer 51 (2010) 1548e1554
Contents lists available at ScienceDirect
Polymer
journal homepage: www.elsevier.com/locate/polymer
Synthesis and gas permeation properties of various Si-containing
poly(diarylacetylene)s and their desilylated membranes
,
,
*
Yanming Hu a b, Kyohei Hattori a, Akito Fukui a, Masashi Shiotsuki a, Fumio Sanda a, Toshio Masuda a
a Department of Polymer Chemistry, Graduate School of Engineering, Kyoto University, Katsura Campus, Nishikyo-ku, Kyoto 615-8510, Japan
b State Key Laboratory of Fine Chemicals, Department of Polymer Science and Engineering, School of Chemical Engineering, Dalian University of Technology, 158 Zhongshan Road,
Dalian 116012, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Diarylacetylene monomers having trimethylsilyl groups and other substituents (substituted biphenyl, 1a
and 1b; trimethylsilylmethylphenyl, 1cee) were synthesized and polymerized with TaCl5-n-Bu4Sn
catalyst to produce the corresponding poly(diarylacetylene)s (2aed). Polymers 2aec had high molecular
weights and were soluble in common organic solvents. Free-standing membranes of 2aec as well as
previously reported 2feh were prepared by the solution-casting method. Desilylation of these Si-con-
taining polymer membranes was carried out with trifluoroacetic acid to afford 3a, 3b, and 3feh. Upon
desilylation, biphenyl-containing membranes became less permeable (3a, b), whereas fluorene-con-
taining membranes became more permeable (3feh). In particular, 3h exhibited extremely high gas
permeability (PO2 ¼ 9800 barrers), which is about the same as that of poly(1-trimethylsilyl-1-propyne).
Desilylated membranes 3a and 3feh showed different gas permeability from that of polymers 2iek
which have the identical chemical structures and obtained directly by the polymerization of the cor-
responding monomers.
Received 5 January 2010
Received in revised form
5 February 2010
Accepted 7 February 2010
Available online 12 February 2010
Keywords:
Gas permeability
Membranes
Polyacetylene
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
properties have been reported to date [13e16]. Poly[1-phenyl-2-
(p-trimethylsilyl)phenylacetylene] [poly(TMSDPA)] exhibits both
Substituted polyacetylenes are attractive polymers from the
viewpoints of their unique physical and chemical characteristics as
well as their potential applications as functional materials in opto-
electronics, sensing based on stimuli-responsiveness, membrane
separation of gases, and other fields [1e6]. Membrane-based sepa-
ration technology has attracted considerable attention in the past few
decades. As an important category of gas separation membrane
materials, substituted polyacetylenes, especially polymers derived
from disubstituted acetylenes, are characterized by high gas perme-
ability and high vapor/gas selectivity [7e10]. Unique permeation
properties of these polymers are mainly attributed to their large
excessfree volume, whichoriginates fromtheirstiff mainchain, bulky
substituents, and low cohesive energy structure. For example,
regarding poly(1-trimethylsilyl-1-propyne) [poly(TMSP)] which is
one of the most gas-permeable polymers among all the existing
polymers [11,12], many studies concerning its gas permeation
high gas permeability and excellent thermal stability; its oxygen
permeabilitycoefficient (PO2) isaslarge as1100 barrers, and theonset
temperature of weight loss reaches 420 ꢀC [17,18]. More recently we
have developed indan-containing polyacetylenes that display gas
permeability even higher than that of poly(TMSP) [19].
The design and synthesis of novel polymers as membrane
materials are of great importance not only because some limita-
tions of the existing polymers may be overcome but also because
valuable information will be provided for membrane science and
technology. We have previously reported that solvent-insoluble
membranes of poly(diphenylacetylene) can be obtained through
the desilylation of various silylated poly(diphenylacetylene)s,
which cannot be obtained directly by the solvent-casting method
owing to its insolubility [20,21]. Desilylation not only imparts
chemical stability to the polymers but also affects their gas
permeability. For example, the PO2 of poly[1-b-naphthyl-2-(p-tri-
methylsilylphenyl)acetylene] is 3500 barrers, while the value of its
desilylated derivative increases to 4300 barrers [22]. In contrast,
the PO2 of poly(TMSDPA) decreases from 1500 barrers to 910 bar-
rers after desilylation [21]. It is presumed that molecular-scale
voids are formed by the desilylation reaction in the solid state, and
that the number/size of the formed microvoids is affected by the
* Corresponding author. Department of Environmental and Biological Chemistry,
Fukui University of Technology, 3-6-1 Gakuen, Fukui 910-8505, Japan. Tel./fax: þ81
776 29 2714.
E-mail address: masuda@fukui-ut.ac.jp (T. Masuda).
0032-3861/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.polymer.2010.02.015

Y. Hu et al. / Polymer 51 (2010) 1548e1554
1549
substituents. Further, the decrease of local mobility upon the loss of
silyl group must also be considered. However, detailed studies of
the effects of substituents on the gas permeability of desilylated
polymer membranes are rather limited. Thus it is quite interesting
to compare in detail the gas permeability of desilylated poly(di-
arylacetylenes) with that of silyl-containing precursor polymers
and polymers having identical chemical structures which are
synthesized directly by the polymerization of the corresponding
monomers.
Co. Ltd. Phenylacetylene (Aldrich), 4-bromo-40-tert-butylbiphenyl
(TCI), 4,40-dibromobiphenyl, benzyltrimethylsilane, 1,3-dibromo-
benzene, trifluoroacetic acid (TFA), and common solvents (Wako
Pure Chemical) were used without further purification except
toluene as a polymerization solvent which was purified by the
standard method.
2.3. Monomer synthesis
In the present study, several Si-containing poly(diarylacetylene)s
were synthesized Scheme 1. Free-standing membranes were fabri-
cated from several resultant polymers, and desilylation of the
membranes was carried out with trifluoroacetic acid. Then, the solu-
bility characteristics and thermal properties of the polymers were
elucidated. Further, the gas permeation parameters of the polymer
membranes were determined, and their diffusion and solubility
coefficients were revealed.
(4-Iodobenzyl)trimethylsilane,
4-bromo-40-trimethylsilylbi-
phenyl, and 3-(trimethylsilyl)phenylacetylene were prepared
according to the literature methods [23,24]. Monomers 1aed were
synthesized referring to the literature of ethynylation reactions
(Scheme 2) [25]. Monomers 1feh were synthesized as previously
described [26].
2.3.1. 1-(4-Trimethylsilylphenyl)-2-(40-tert-butylbiphenyl-4-yl)
acetylene (1a)
2. Experimental section
A 500 mL three-necked flask was equipped with a reflux
condenser, a three-way stopcock, and a magnetic stirring bar. After
the flask was flushed with nitrogen, 4-bromo-40-tert-butylbiphenyl
(5.0 g,17 mmol), bis(triphenylphosphine)palladium dichloride (250
mg, 0.35 mmol), cuprous iodide (390 mg, 2.1 mmol) and triphe-
nylphosphine (360 mg, 1.4 mmol) were placed in the flask and
dissolved in triethylamine (90 mL) at room temperature. Then,
a solution of 4-(trimethylsilyl)phenylacetylene (3.0 g, 17.3 mmol) in
triethylamine (20 mL) was added, and the reaction mixture was
heated at reflux temperature for 5 h. Triethylamine in the reaction
mixture was evaporated off, and then diethyl ether (200 mL) was
added to the residual mass. Solvent-insoluble solid was filtered off,
and the filtrate was washed with 2 M HCl aq. and then with water.
The ethereal solution was dried over anhydrous magnesium sulfate.
After evaporating the solvent, the crude product was purified by
silica gel column chromatography (eluent: hexane) to give a white
2.1. Measurements
Molecular weights of polymers were estimated by gel perme-
ation chromatography (GPC) on a Shimadzu PU/SPD-6A/UV-975
chromatograph equipped with polystyrene columns (Showa Denko
Shodex K-805, K-806, and K-807) using CHCl3 as an eluent at a flow
rate of 1.0 mL/min, calibrated with polystyrene standard samples.
NMR spectra were recorded on a JEOL EX-400 spectrometer.
Elemental analysis of monomers was carried out at the Microana-
lytical Center of Kyoto University. Thermogravimetric analysis
(TGA) was conducted in air with a Shimadzu TGA-50 thermal
analyzer.
2.2. Materials
Tantalum (V) chloride (TaCl5, Aldrich) was used without further
purification. Tetra-n-butyltin (n-Bu4Sn, Wako) was used after
distillation. 4-(trimethylsilyl)phenylacetylene was donated by NOF
solid; yield 29%. 1H NMR (400 MHz,
Ar), 1.35 (s, 9H, CCH3), 0.27(s, 9H, SiCH3). 13C NMR (100 MHz,
CDCl3): 150.7, 141.0, 140.8, 137.4, 133.2, 132.0, 130.6, 126.8, 126.6,
d, CDCl3): 7.65e7.40 (m, 12H,
d
,
Scheme 1. Synthesis of poly(diarylacetylene)s.

1550
Y. Hu et al. / Polymer 51 (2010) 1548e1554
Scheme 2. Synthesis of diarylacetylene monomers having trimethylsilyl moieties.
125.8, 123.6, 121.9, 90.1, 89.8, 34.6, 31.3, ꢁ1.2. Anal. Calcd. for
ether. The organic phase was washed first with 1 M Na2S2O3$5H2O
aq (200 mL ꢂ 5) and then with H2O (200 mL ꢂ 3). The ethereal
solution was dried over anhydrous magnesium sulfate. After
evaporating the solvent, the crude product was purified by silica gel
column chromatography (eluent: hexane) to give (4-iodobenzyl)
C27H30Si: C, 84.76; H, 7.90; Si, 7.34. Found: C, 84.76; H, 7.90.
2.3.2. 1-Phenyl-2-(40-trimethylsilylbiphenyl-4-yl)acetylene (1b)
A 500 mL three-necked flask was equipped with a dropping
funnel, a three-way stopcock, and a magnetic stirring bar. After the
flask was flushed with nitrogen, 4,40-dibromobiphenyl (20 g, 64
mmol) and diethyl ether (100 mL) were added and cooled at 0 ꢀC. At
the same temperature, a hexane solution of n-butyllithium (42 mL,
1.58 M, 66 mmol) was added dropwise, and then the mixture was
stirred for 1 h at room temperature. Then, a solution of trimethyl-
chlorosilane (7.0 g, 64 mmol) in diethyl ether (50 mL) was added
dropwise at 0 ꢀC, and stirring was continued overnight at room
temperature. A small amount of water was added at 0 ꢀC, and the
reaction mixture was extracted with diethyl ether. The organic
phase was washed with water, and dried over anhydrous magne-
sium sulfate. After diethyl ether was evaporated, the crude product
was purified by silica gel column chromatography (eluent: hexane)
to give 4-bromo-40-trimethylsilylbiphenyl.
trimethylsilane as a colorless liquid; yield 38%. 1H NMR (400 MHz,
d,
CDCl3): 7.49 (d, J ¼ 8.0 Hz, 2H, Ar), 6.73 (d, J ¼ 7.6 Hz, 2H, Ar), 2.01
(s, 2H, CH2Si), ꢁ0.03 (s, 9H, SiCH3). 13C NMR (100 MHz,
140.3, 137.0, 130.1, 88.2, 26.8, ꢁ2.0.
d, CDCl3):
Monomer 1c was prepared by the same method as for monomer
1a using (4-iodobenzyl)trimethylsilane and phenylacetylene
instead of 4-bromo-40-tert-butylbiphenyl and 4-(trimethylsilyl)
phenylacetylene to give a white solid; yield 80%. 1H NMR (400 MHz,
d
, CDCl3): 7.55e7.43 (m, 2H, Ar), 7.40e7.25 (m, 5H, Ar), 6.96 (d, J ¼
8.0 Hz, 2H, Ar), 2.10 (s, 2H, CCH2Si), ꢁ0.01 (s, 9H, SiCH3). 13C NMR
(100 MHz, d, CDCl3): 141.4, 131.5, 131.4, 128.2, 128.0, 127.9, 123.6,
118.5, 89.8, 88.5, 27.5, ꢁ1.9. Anal. Calcd. for C18H20Si: C, 81.76; H,
7.62; Si, 10.62. Found: C, 81.65; H, 7.62.
Monomer 1b was prepared by the same method as for monomer
1a using 4-bromo-40-trimethylsilylbiphenyl and phenylacetylene
instead of 4-bromo-40-tert-butylbiphenyl and 4-(trimethylsilyl)
phenylacetylene to give a white solid; yield 36%. 1H NMR (400 MHz,
2.3.4. 1-(4-Trimethylsilylphenyl)-2-(4-trimethylsilylmethylphenyl)-
acetylene (1d)
This monomer was prepared by the same method as for
monomer 1a using (4-iodobenzyl)trimethylsilane instead of 4-
bromo-40-tert-butylbiphenyl to give a white solid; yield 29%. 1H
d
, CDCl3): 7.75e7.55 (m, 10H, Ar), 7.45e7.30 (m, 3H, Ar), 0.30 (s, 9H,
SiCH3). 13C NMR (100 MHz,
d, CDCl3): 140.9, 140.7, 139.8, 133.9,
NMR (400 MHz,
d
, CDCl3): 7.47 (s, 4H, Ar), 7.38 (d, J ¼ 8.0 Hz, 2H, Ar),
132.0, 131.6, 128.4, 128.3, 127.0, 126.3, 123.3, 122.2, 90.1, 89.3, ꢁ1.1.
Anal. Calcd. for C23H22Si: C, 84.61; H, 6.79; Si, 8.60. Found: C, 84.76;
H, 7.10.
6.96 (d, J ¼ 8.0 Hz, 2H, Ar), 2.10 (s, 2H, CCH2Si), 0.26 (s, 9H, SiCH3),
ꢁ0.02 (s, 9H, SiCH3). 13C NMR (100 MHz,
d, CDCl3): 141.4, 140.5,
133.1, 131.4, 130.5, 127.9, 123.9, 118.5, 90.1, 88.6, 27.5, ꢁ1.2, ꢁ1.8.
Anal. Calcd. for C21H28Si2: C, 74.93; H, 8.38; Si, 16.69. Found: C,
74.64; H, 8.38.
2.3.3. 1-Phenyl-2-(4-trimethylsilylmethylphenyl)acetylene (1c)
A 300 mL three-necked flask was equipped with a three-way
stopcock and a magnetic stirring bar. After the flask was flushed
with nitrogen, benzyltrimethylsilane (8.7 g, 53 mmol), iodine (5.3 g,
21 mmol), and periodic acid (2.6 g, 11 mmol) were dissolved in
aqueous acetic acid solution (52 mL, 80%) containing sulfuric acid
(1.6 mL). The reaction mixture was stirred at 100 ꢀC for 15 h. After
H2O (200 mL) was added, the solution was extracted with diethyl
2.3.5. 1-(3-Trimethylsilylphenyl)-2-(4-trimethylsilylmethylphenyl)-
acetylene (1e)
This monomer was prepared by the same method as for
monomer 1a using (4-iodobenzyl)trimethylsilane and 3-(trime-
thylsilyl)phenylacetylene instead of 4-bromo-40-tert-butylbiphenyl
and 4-(trimethylsilyl)phenylacetylene to give a colorless liquid;

Y. Hu et al. / Polymer 51 (2010) 1548e1554
1551
yield 63%. 1H NMR (400 MHz,
d, CDCl3): 7.65 (s, 1H, Ar), 7.50e7.35
summarized in Table 1. It has been reported that TaCl5-n-Bu4Sn is
an effective catalyst for the polymerization of sterically crowded
disubstituted acetylenes including diphenylacetylene derivatives
to provide polymers having high molecular weights [4,5], which is
essential for fabrication of free-standing membranes. The poly-
merization of monomers 1a and 1b having both biphenyl and
trimethylsilyl groups afforded polymers 2a and 2b in high yields
(65% and 78%, respectively), whose weight-average molecular
weights (Mw) exceeded 6.0 ꢂ 106. Monomer 1c containing a tri-
methylsilylmethyl group also polymerized to give a polymer with
somewhat lower Mw (6.6 ꢂ 105) in 55% yield. When cyclohexane
was used as polymerization solvent, both Mw and yield of the
polymer increased. Similar results have been observed in the
polymerization of other diarylacetylenes in our previous studies
[22]. On the other hand, polymer 2d obtained from monomer 1d,
which has a trimethylsilyl substituent on the para-position of
a phenyl ring, was insoluble in any organic solvents. The meta-
counterpart 2e did not polymerize under the same conditions,
which should be due to steric hindrance. Thus, the monomer
structure significantly affects polymerization behavior and poly-
mer solubility.
(m, 4H, Ar), 7.31 (t, J ¼ 7.6 Hz,1H, Ar), 6.96 (d, J ¼ 8.0 Hz, 2H, Ar), 2.10
(s, 2H, CCH2Si), 0.27 (s, 9H, SiCH3), ꢁ0.02 (s, 9H, SiCH3). 13C NMR
(100 MHz, d, CDCl3): 141.4, 140.7, 136.4, 132.7, 131.7, 131.5, 128.0,
127.6, 123.0, 118.6, 89.7, 88.8, 27.5, ꢁ1.2, ꢁ2.0. Anal. Calcd. for
C21H28Si2: C, 74.93; H, 8.38; Si, 16.69. Found: C, 75.12; H, 8.52.
2.4. Polymerization procedure
Polymerizations were performed using toluene as solvent in
a Schlenk tube equipped with a three-way stopcock at 80 ꢀC for 24
h under dry nitrogen at the following reagent concentrations:
[TaCl5] ¼ 20 mM, [n-Bu4Sn] ¼ 40 mM. The formed polymers were
isolated by precipitation into a large amount of methanol, and dried
to constant weight. The polymer yields were determined by
gravimetry.
2.5. Membrane fabrication and desilylation
The membranes (thickness ca. 80e120 mm) of polymers 2aec
and 2feg were fabricated by casting from toluene solution of the
polymer (concentration ca. 0.50e1.0 wt%) onto a flat-bottomed
Petri dish. Then, the dish was covered with a glass vessel to slow
solvent evaporation (ca. 3e5 days). After a membrane was formed,
the membrane was peeled off, and it was immersed in methanol for
24 h and dried to constant weight at room temperature for 24 h.
With reference to the method described in the literature [20,27],
the desilylation reaction of polymer membranes was carried out
using trifluoroacetic acid. A typical procedure of the desilylation
reaction is as follows; a polymer membrane was immersed in
a mixture of TFA and water (volume ratio 9:1) at room temperature
for 24 h. The membrane was immersed in water for 24 h, then
washed with water to remove residual impurities, and dried at
room temperature for 24 h and in vacuo for 5 h to constant weight.
3.2. Fabrication and desilylation of polymer membranes
Free-standing membranes could be fabricated by casting
polymers 2aec from their toluene solution, whereas the
membrane of 2d could not be prepared due to its insolubility.
The formed membranes were sufficiently tough and transparent.
The membranes from 2feg were also fabricated to prepare their
desilylated membranes. The desilylation reaction of polymer
membranes 2a, 2b, and 2feh was carried out in a trifluoroacetic
acid/H2O mixture (9:1 volume ratio) at room temperature for 24 h.
The completion of desilylation of 2b, 2f and 2h was confirmed by
the IR spectra and TGA curves of the polymers. Fig. 1 shows the IR
spectra of membranes 2b and 3b. The absorption peaks at 1250,
1120, and 860 cmꢁ1 indicating the presence of the trimethylsilyl
groups in 2b disappeared in 3b. Just the same changes were
observed in the cases of 2f, 2h, 3f, and 3h. Because of overlapping
of the absorption peaks between t-butyl and trimethylsilyl groups,
the complete removal of trimethylsilyl groups could not be
confirmed by IR spectroscopy in the case of 3a, for which TGA was
used to confirm the practically complete desilylation (Fig. 2). In the
cases of 3a, 3b, 3f, and 3h, no SiO2 residue was detected when
the polymers were heated above 700 ꢀC, which confirms that the
desilylation reactions proceeded quantitatively. On the other hand,
desilylation of polymer 2g did not reach completion under the
same reaction conditions; the result of TGA indicates that about
85% of the silyl groups of 2g were removed under the above-stated
conditions.
2.6. Measurement of gas permeability
Gas permeability coefficients (P) of polymer membranes were
measured with a Rikaseiki K-315-N gas permeability apparatus
equipped with an MKS Baratron detector at 25 ꢀC. The downstream
side of the membrane was evacuated to 0.3 Pa, while the upstream
side was filled with a gas at about 1 atm (105 Pa), and the increase of
pressure in a downstream receiving vessel was measured. The P
values were calculated from the slopes of timeepressure curves in
the steady state where Fick's law held.
The gas diffusion coefficients (D) were determined by the time
lag method using the following equation:
D ¼ l2=6
q
Here, l is membrane thickness, and
q is time lag, which is given by
the intercept of the asymptotic line of the timeepressure curve to
the time axis. The membrane thickness was controlled so that the
time lag would be in the range 10e300 s, preferably 30e150 s.
When the time lag was <10 s, the error of measurement became
relatively large. If the time lag was, on the contrary, >300 s, the
error based on baseline drift became significant. The gas solubility
coefficients (S) were calculated by using equation S ¼ P/D.
Table 1
Polymerization of monomers 1aee by TaCl5-n-Bu4Sn catalysta.
Monomer
Solvent
Polymerb
Yield (%)
Mw ꢂ 10ꢁ3c
Mw/Mn
c
1a
1b
1c
1c
1d
1e
toluene
toluene
toluene
cyclohexane
toluene
65
78
55
74
52
0
>6000
>6000
660
e
e
4.7
5.3
ed
e
1000
3. Results and discussion
ed
toluene
e
3.1. Polymerization
a
Polymerization at 80 ꢀC for 24 h; [TaCl5] ¼ 20 mM, [n-Bu4Sn] ¼ 40 mM.
Methanol-insoluble product.
Determined by GPC eluted with CHCl3 (polystyrenes as standards).
Insoluble in any solvents.
b
c
The polymerization of diarylacetylene monomers (1aee) was
examined by using TaCl5-n-Bu4Sn catalyst, whose results are
d

1552
Y. Hu et al. / Polymer 51 (2010) 1548e1554
Table 2
Solubility of polymersa.
2aec 2d 2f 2g 2h 2i 2j 2k 3a 3b 3f 3g 3h
Hexane
Cyclohexane
Toluene
CHCl3
THF
Methanol
DMF
ꢁ
ꢃ
þ
þ
þ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
þ
þ
þ
þ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
þ
þ
þ
ꢁ
ꢁ
ꢁ
þ
þ
þ
þ
þ
ꢁ
ꢁ
ꢁ
ꢁ
ꢃ
þ
þ
þ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
þ
þ
þ
ꢁ
þ
ꢁ
ꢁ
ꢁ
þ
þ
þ
ꢁ
þ
ꢁ
ꢁ
ꢁ
þ
þ
þ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
þ
þ
þ
ꢁ
þ
ꢁ
ꢁ
ꢃ
þ
þ
þ
ꢁ
þ
ꢁ
ꢁ
ꢁ
þ
þ
þ
ꢁ
þ
ꢁ
DMSO
a
Symbols: þ: soluble; ꢃ: partly soluble; ꢁ: insoluble.
3.4. Gas permeation properties
The gas permeability of membranes 2aec, and desilylated
membranes 3a, 3b, and 3feh to various gases was measured at
25 ꢀC. As seen from Table 3, 2a having both trimethylsilyl and
p-tert-butylbiphenyl groups showed the oxygen permeability
coefficients (PO2) of 620 barrers, which is obviously lower than that
of poly(TMSDPA) (1500 barrers). This result may be due to the
presence of the planar biphenyl group which may stack with each
other, leading to the decrease of microvoids. When spacers (phenyl
and methylene for 2b and 2c, respectively) were inserted between
the silicon atom and the phenyl ring of poly(TMSDPA), the gas
permeability decreased. Thus, the PO2 value of 2b was 390 barrers
which is about a quarter that of poly(TMSDPA), and polymer 2c
having trimethylsilylmethyl groups exhibited the lowest perme-
ability among 2aec (PO2 ¼ 140 barrers). It is considered that
incorporation of flexible or planar spacers reduces the stiffness of
polymer chain, resulting in the decrease of microvoids, and in turn,
gas permeability. Similar results have been reported in previous
studies; namely, the introduction of flexible groups in polymer
results in a decrease of gas permeability; e.g., the PO2 value of poly
[1-(4-methylphenyl)-2-(4-tert-butyldimethylsiloxylphenyl)acety-
lene] is 230 barrers, while that of poly[1-(4-ethylphenyl)-2-(4-tert-
butyldimethylsiloxylphenyl)acetylene] is 140 barrers [29].
Fig. 1. IR spectra of polymer 2b and its desilylated product 3b (films).
3.3. Solubility and thermal stability of the polymers
The solubility of the formed polymers was examined for various
solvents (Table 2). Si-containing polymers 2aec and 2feh are
readily soluble in common organic solvents including toluene,
CHCl3, and THF. In contrast, polymer 2d having two similar side
groups in size are insoluble, which has also been observed with
other polyacetylene derivatives [13]. Desilylated polymers 3a, 3f,
and 3h showed practically the same solubility as that of 2i [28], 2j
[26], and 2k [26] (Chart 1), respectively, which have the identical
chemical structures and were obtained directly by the polymeri-
zation of the corresponding monomers. These results indicate that
the desilylation reaction did not affect the solubility of the poly-
mers. The thermal stability of polymers 2aec, 3a, 3b, and 3feh was
examined by TGA in air (Fig. 2). Except 2c, the onset temperatures
of weight loss of the present polymers were above 400 ꢀC, indi-
cating excellent thermal stability.
The desilylation reaction affected the gas permeability of the
polymers in different ways when they have different substituents.
For example, the PO2 values of the desilylated membranes 3a and
3b having biphenyl groups were 260 and 180 barrers, respectively,
which are lower than their Si-containing precursors 2a and 2b. On
the other hand, the gas permeability of the fluorenyl-containing
polymers increased to 1.5e2 times after desilylation. Thus, the PO2
Table 3
Gas permeability coefficients (P) of polymer membranes.
Polymer
P (barrera)
He H2
560
PO2/PN2
O2
620
390
140
1000
440
5400
260
180
1300
830
9800
130
660
N2
250
160
44
540
190
3400
95
CO2
3300
2200
640
4100
1900
18000
1600
1300
4900
3400
24000
840
CH4
730
450
120
1500
500
10000
260
170
2300
940
19000
120
810
2a
2b
2c
1200
810
2.5
2.4
3.2
1.9
2.3
1.6
2.7
2.8
1.5
2.0
1.2
2.8
2.2
1.5
370
240
420
2fb
2gb
2hb
3a
840
480
1900
980
3400
260
8500
580
3b
3f
3g
3h
2ic
2jb
2kb
200
440
65
1100
890
8800
170
2500
2000
19000
330
870
420
8100
46
300
3300
600
3200
1400
8100
3200
17000
4800
9100
a
At 25 ꢀC in the units of 1 ꢂ 10ꢁ10 cm3 (STP) cm/(cm2 s cmHg) (¼1 barrer).
b
c
Data from Ref. [26].
Data from Ref. [28].
Fig. 2. TGA curves of the polymers (in air, heating rate 10 ꢀC/min).

Y. Hu et al. / Polymer 51 (2010) 1548e1554
1553
Chart 1. Poly(diarylacetylene)s having the identical chemical structures to those of
desilylated polymers (3a, 3f, and 3h) and obtained directly by the polymerization of
the corresponding monomers.
into the gas diffusion coefficients (D) and gas solubility coefficients
(S). The calculation method was explained in the experimental
section. Fig. 3 shows the plot of D and S values of the methane
permeability of the present polymers. The time lag in the perme-
ation of 3h was too small to calculate the D value.
Fig. 3. Plot of diffusion coefficient (D) vs solubility coefficient (S) of the polymers
The SCH4 values of the polymers hardly changed upon desily-
lation, while the DCH4 values of the polymers with different
substituents showed different tendencies. The gas diffusivity of
polymer membranes should depend not only on the free volume of
membrane but also on the local mobility of substituents. For
instance, the DCH4 significantly decreased in the biphenyl-con-
taining membranes 3a and 3b by desilylation, which suggests the
strong effect of the decrease of local mobility. On the other hand,
the D values of 3f and 3g were larger than those of 2f and 2g, which
is attributable to the formation of microvoids in 3f and 3g at the
location where silyl groups were present in 2f and 2g. A more
detailed study is underway.
before and after desilylation in methane permeation.
values of 3f and 3g were 1300 and 830 barrers, respectively, and
higher than those of 2f and 2g (1000 and 400 barrers). It is quite
interesting that membrane 3h exhibited extremely high gas
permeability; for instance, its PO2 value was 9800 barrers, which is
much higher than that of 2h (5400 barrers) and is about the same
as that of poly(TMSP) (w10,000 barrers). These results indicate or
imply the following things: (i) the desilylation of diarylacetylene
derivatives increases or decreases gas permeability depending on
the chemical structure of the starting polymer, (ii) the variation of
gas permeability upon desilylation in the solid state is considered
to be mainly governed by the increase or decrease of microvoids
and the decrease of local mobility, and (iii) the increase or decrease
of microvoids should lead to the increase or decrease of excess free
volume, and in turn, gas permeability, while the decrease of local
mobility based on the loss of the silyl group will lead to the
decrease of permeability [30,31].
The PO2 values of desilylated membranes 3a, 3f, and 3h were
about twice those of 2i [28], 2j [26], and 2k [26] (130 barrers, 660
barrers, and 4800 barrers, respectively; Table 3), which have the
identical chemical structures and obtained by solution-casting
method using the polymers synthesized directly by the polymeri-
zation of the corresponding monomers. In contrast, 3g obtained by
desilylation showed much lower gas permeability than 2k,
although they have the identical chemical structure to each other.
Sakaguchi and coworker have reported that desilylation of the poly
[1-p-n-alkylphenyl-2-(p-trimethylsilyl)phenylacetylene] membranes
reduce gas permeability, while the desilylated polymers show gas
permeability similar to that of the same polymers obtained by
direct polymerization [32]. The following reasons are possible for
the finding that desilylated polymers and the corresponding poly-
mers obtained directly by polymerization show different gas
permeability: (i) they may have different geometric and/or regio
structures, (ii) the polymers desilylated in the solid state may have
less relaxed structures, and so forth.
4. Conclusion
In the present study, poly(diarylacetylene) derivatives having
trimethylsilyl groups were synthesized by the polymerization of
the corresponding monomers with TaCl5-n-Bu4Sn catalyst. Poly-
mers 2aec had high molecular weights, and showed good solu-
bility in common organic solvents and film-forming ability.
Desilylation of trimethylsilyl-containing polymer membranes 2a
and 2b, along with previously reported 2feh, was achieved with
trifluoroacetic acid to give membranes 3a, 3b, and 3feh. Biphenyl-
containing membranes 3a, 3b showed decreases in gas perme-
ability upon desilylation, whereas fluorene-bearing counterparts
3feh exhibited increases in gas permeability. These desilylated
membranes showed different gas permeability from that of
membranes prepared by solution-casting of the polymers having
the identical chemical structures and obtained directly by the
polymerization of the corresponding monomers.
References
[1] Liu J, Lam JWY, Tang BZ. Chem Rev 2009;100:1645e81.
[2] Yashima E, Maeda K. Macromolecules 2008;41:3e12.
[3] Rudick JG, Percec V. Macromol Chem Phys 2008;209:1759e68.
[4] Masuda T. J Polym Sci Part A Polym Chem 2007;45:165e80.
[5] Masuda T, Sanda F, Shiotsuki M. In: Crabtree R, Mingos M, editors.
Comprehensive organometallic chemistry III, vol. 11. Oxford, UK: Elsevier;
2007. p. 557e93 [chapter 11.6.1].
[6] Nagai K, Lee YM, Masuda T. In: Matyjaszewsky K, Gnanou Y, Leibler L, editors.
Macromolecular engineering. Weinheim, Germany: Wiley-VCH; 2007 [Part 4,
chapter 19].
The permeability of the membranes to other gases such as He,
H2, N2, CO2, and CH4 showed similar tendencies. The separation
factors of oxygen against nitrogen (PO2/PN2) of the present poly-
mers were in a range of 1.2e3.2. A tradeoff relationship is observed
between permeability and permselectivity, namely, more perme-
able polymers are generally less permselective and vice versa [33].
The gas permeability coefficients (P) of the polymers before and
after desilylation were inspected in more detail by separating them
[7] Aoki T, Kaneko T, Teraguchi M. Polymer 2006;47:4867e92.
[8] Ulbricht M. Polymer 2006;47:2217e62.
[9] Freeman BD, Yampolskii Y, Pinnau I. Materials science of membranes for gas
and vapor separation. Chichester: Wiley; 2006.

1554
Y. Hu et al. / Polymer 51 (2010) 1548e1554
[10] Pinnau I, Freeman BD, editors. Advanced materials for membrane separation,
ACS Symposium series 876. Washington: American Chemical Society; 2004.
[11] Masuda T, Isobe E, Higashimura T. J Am Chem Soc 1983;105:7473e4.
[12] Masuda T, Isobe E, Higashimura T. Macromolecules 1985;18:841e5.
[13] Nagai K, Masuda T, Nakagawa T, Freeman BD, Pinnau I. Prog Polym Sci
2001;26:721e98.
[14] Thomas S, Pinnau I, Du NY, Guiver MD. J Membr Sci 2009;338:1e4.
[15] Volkov AV, Parashchuk VV, Stamatialis DF, Khotimsky VS, Volkov VV,
Wessling M. J Membr Sci 2009;333:88e93.
[22] Sakaguchi T, Kwak G, Masuda T. Polymer 2002;43:3937e42.
[23] Suzuki H, Nakamura K, Goto R. Bull Chem Soc Jpn 1966;39:128e31.
[24] Aoki T, Nakahara H, Hayakawa Y, Kokai M, Oikawa E. J Polym Sci Polym Chem
1994;32:849e58.
[25] Sonogashira K. J Organomet Chem 2002;653:46e9.
[26] Fukui A, Hattori K, Hu Y, Shiotsuki M, Sanda F, Masuda T. Polymer
2009;50:4159e65.
[27] Teraguchi M, Mottate K, Kim SY, Aoki T, Kaneko T, Hadano S, et al. Macro-
molecules 2005;38:6367e73.
[16] Kelman SD, Raharjo RD, Bielawski CW, Freeman BD. Polymer
2008;49:3029e41.
[28] Hu Y, Shimizu T, Hattori K, Shiotsuki M, Sanda F, Masuda T. J Polym Sci Part A
2010;48:861e8.
[17] Tsuchihara K, Masuda T, Higashimura T. J Am Chem Soc 1991;113:8548e9.
[18] Tsuchihara K, Masuda T, Higashimura T. Macromolecules 1992;25:5816e20.
[19] Hu Y, Shiotsuki M, Sanda F, Freeman BD, Masuda T. Macromolecules
2008;41:8525e32.
[20] Teraguchi M, Masuda T. Macromolecules 2002;35:1149e51.
[21] Sakaguchi T, Yumoto K, Shiotsuki M, Sanda F, Yoshikawa M, Masuda T.
Macromolecules 2005;38:2704e9.
[29] Hu Y, Sakaguchi T, Shiotsuki M, Sanda F, Masuda T. J Membr Sci
2006;285:412e9.
[30] Kanaya T, Teraguchi M, Masuda T, Kaji K. Polymer 1999;40:7157e61.
[31] Kanaya T, Tsukushi I, Kaji K, Sakaguchi T, Kwak G, Masuda T. Macromolecules
2002;35:5559e64.
[32] Takeda A, Sakaguchi T, Hashimoto T. Polymer 2009;50:5031e6.
[33] Robeson LM. J Membr Sci 1991;62:165e85.