Novel poly(diphenylacetylene)s with both alkyl and silyl groups as gas permeable membranes: Synthesis, desilylation, and gas permeability

Article abstract of DOI:10.1016/j.polymer.2010.01.022

Diphenylacetylenes having a dimethyloctylsilyl group and an alkyl group at para positions [Me₂n-C₈H₁₇SiC₆H₄C{triple bond, long}CC₆H₄R; R = H (1a), i-Pr (1b), t-Bu (1c), n-Bu (1d)

Full text of DOI:10.1016/j.polymer.2010.01.022

Polymer 51 (2010) 1279–1284
Contents lists available at ScienceDirect
Polymer
journal homepage: www.elsevier.com/locate/polymer
Novel poly(diphenylacetylene)s with both alkyl and silyl groups as gas
permeable membranes: Synthesis, desilylation, and gas permeability
*
Aiko Takeda, Toshikazu Sakaguchi , Tamotsu Hashimoto
Department of Materials Science and Engineering, Graduate School of Engineering, University of Fukui, Bunkyo 3-9-1, Fukui 910-8507, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Diphenylacetylenes having a dimethyloctylsilyl group and an alkyl group at para positions [Me2n-
C8H17SiC6H4C^CC6H4R; R ¼ H (1a), i-Pr (1b), t-Bu (1c), n-Bu (1d)] and having only an alkyl group
[PhC^CC6H4R; R ¼ i-Pr (1B), t-Bu (1C)] were synthesized and then polymerized with TaCl5/n-Bu4Sn catalyst
to provide the corresponding poly(diphenylacetylene)s (2a, 2b, 2c, 2d, 2B, and 2C). The formed polymers
afforded tough free-standing membranes by casting from toluene solutions. Desilylation reaction of
Si-containing membranes (2a–d) was carried out with trifluoroacetic acid to give the desilylated
membranes (3a–d). The permeability of these membranes to O2, N2, and CO2 were determined. All the Si-
containing membranes exhibited almost the same gas permeability. The desilylation of Si-containing
membranes of 2a–c resulted in large increase of gas permeability. No apparent increasing of gas perme-
ability was observed in the desilylation of 2d. To clarify the effects of desilylation, CO2 diffusivity (D(CO2)),
CO2 solubility (S(CO2)), and fractional free volume (FFV) of the polymer membranes were investigated. The
S(CO2) values of desilylated membranes were much larger than that of Si-containing counterparts. The
D(CO2) and FFV of membranes of 2a–c increased through desilylation. The desilylated membrane of 3d had
small D(CO2) value and almost the same FFV compared with 2d. Further, the comparison of the permeability
between three types of membranes with the same chemical structure revealed that the microvoids were not
generated by the desilylation of membranes of poly(diphenylacetylene)s containing alkyl groups.
Ó 2010 Elsevier Ltd. All rights reserved.
Received 20 November 2009
Received in revised form
6 January 2010
Accepted 8 January 2010
Available online 18 January 2010
Keywords:
Poly(diphenylacetylene)
Gas permeability
Desilylation
1. Introduction
mobility of polymer chain is restrained in a solid state [4–9].
However, the details about the effect of desilylation in a solid state
Polyacetylenes with bulky spherical substituents show
extremely high gas permeability. This is because both their stiff
main chain composed of alternating double bonds and the steric
repulsion of the bulky substituents make membranes sparse [1].
Therefore, poly(substituted acetylene)s are promising materials for
gas separation membranes. A variety of poly(substituted acety-
lene)s have been synthesized so far, and the gas permeability of
their membranes has been investigated [1–3]. Since a polymer
membrane is usually prepared by casting its solution, the
membrane of a solvent-insoluble polymer is not prepared directly
by solution-casting. However, several membranes of insoluble
poly(substituted acetylene)s have been prepared by a indirect
method, namely desilylation of solvent-soluble membrane con-
taining silyl groups [4]. In the desilylation reaction of polymer
membrane, it has been predicted that the spaces occupied by silyl
groups are maintained in some level as microvoids because the
upon gas permeability have not been known yet.
In the previous paper [10], we investigated the desilylation of
membranes of poly(diphenylacetylene)s containing both trime-
thylsilyl and linear alkyl groups, and reported that additional
microvoids in their membranes were not generated through desily-
lation. Linear alkyl groups are flexible pendant chains, and thus they
are unsuitable to maintain
a
microvoid. However, poly(-
diphenylacetylene)s having both trimethylsilyl and branched alkyl
groups such as t-butyl groups were insoluble in any solvents, and
their membranes could not be prepared. Therefore, the effect of the
branched alkyl groups versus the linear alkyl groups remains
unknown. In order to solve the problem of insolubility, dimethy-
loctylsilyl groups were introduced to poly(diphenylacetylene)s
instead of trimethylsilyl groups. This enabled us to investigate the
effect of desilylation on gas permeability of poly(diphenylacetylene)s
having branched alkyl groups for the first time.
In this paper, the desilylation of membranes of
poly(diphenylacetylene)s having dimethyloctylsilyl and various
alkyl groups, which contain linear and branched alkyl groups, at
para position of phenyl groups was performed. The effect of
* Corresponding author. Tel.: þ81 776 27 9779; fax: þ81 776 27 8767.
E-mail address: sakaguchi@matse.u-fukui.ac.jp (T. Sakaguchi).
0032-3861/$ – see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.polymer.2010.01.022

1280
A. Takeda et al. / Polymer 51 (2010) 1279–1284
desilylation on gas permeability was investigated in detail. The
desilylated polymers in this study dissolve in common solvents,
and the polymer membranes with the same chemical structure
could be prepared by three different routes (Scheme 1). First
method is the desilylation of Si-containing membranes, and second
is solvent-casting method using the desilylated polymers. Final
route is solvent-casting method using polymers synthesized
directly by the polymerization of monomers without silyl group.
The comparison of the permeability between such three types of
membranes can reveal the effect of desilylation of the membrane
on gas permeability. The membranes which were prepared by
desilylation in a solid state exhibited the same or low permeability
compared with the other two types of membranes without silyl
group. This indicates that the microvoids were not generated by the
desilylation of the membranes.
P
expressed in barrer unit (1 barrer ¼ 10ꢂ10 cm3(STP) cm
cmꢂ2 sꢂ1 cm Hgꢂ1) was calculated from the slope of the steady-
state line. The D value was determined by the time lag method
using the following equation:
D ¼ l2=6
q
here, l is the membrane thickness and q is the time lag, which is given
by the intercept of the asymptotic line of time-pressure curve to the
time axis. The S value was calculated by using equation S ¼ P/D.
2.2. Materials
Toluene as polymerization solvent was purified by distillation
over calcium hydride. TaCl5 as main catalyst was commercially
supplied by Aldrich and used without further purification, while
n-Bu4Sn as cocatalyst was purified by distillation. Phenylacetylene, p-
iodo-i-propylbenzene, p-t-butyliodobenzene, p-n-butyliodobenzene
and common organic solvents were commercially obtained and used
without further purification. p-Dimethyloctylsilylphenylacetylene
were synthesized referring to the literature [11]. 1-(p-Dimethy-
loctylsilyl)phenyl-2-phenylacetylene (1a), 1-(p-i-propyl)phenyl-2-
phenylacetylene (1B), and 1-(p-t-butyl)phenyl-2-phenylacetylene
(1C) were synthesized according to the literatures [5,12]. Synthesis
and properties of poly[1-(p-n-butyl)phenyl-2-phenylacetylene] (2D)
have been reported in our previous paper [10], and the data of 2D
were used in this study to compared with 3d.
2. Experimental
2.1. Measurements
The molecular weights and polydispersity ratios of polymers
were estimated by gel permeation chromatography (tetrahydro-
furan (THF) as eluent, polystyrene calibration) at 40 ꢀC on
a Shimadzu LC-10AD chromatograph equipped with three poly-
styrene gel columns (Shodex KF-802.5 ꢁ1 and A-80M ꢁ 2) and
a Shimadzu RID-6A refractive index detector. IR spectra were
recorded on a Nicolet MAGNA 560 spectrometer. NMR spectra were
obtained on a Jeol LA-500 spectrometer. Elemental analyses of
monomers were performed at the Microanalytical Center of Kyoto
University.
2.3. Monomer synthesis
Gas permeability coefficients of polymer membranes were
measured with a Rikaseiki K-315-N gas permeability apparatus at
25 ꢀC under 1 atm upstream pressure. The permeability coefficient
Monomers were synthesized according to Scheme 2 with
reference to the literature [13]. The synthesis procedures and
analytical data of monomers are as follows.
Scheme 1. Synthesis of polymers and preparation of polymer membranes.

A. Takeda et al. / Polymer 51 (2010) 1279–1284
1281
Scheme 2. Synthesis of monomers.
2.3.1. 1-(p-dimethyloctylsilyl)phenyl-2-(p-i-propyl)phenylacetylene
(1b)
1.29 (m, 12H, SiCH2(CH2)6), 0.95 (t, J ¼ 7.3 Hz, 3H, Ar(CH2)3CH3),
0.90 (t, J ¼ 6.8 Hz, SiCH2(CH2)6CH3), 0.76 (t, J ¼ 7.9 Hz, 2H, SiCH2),
0.28 (s, 6H, SiCH3). 13C NMR (CDCl3, ppm): 143.3, 140.0, 133.4, 131.5,
130.5, 128.4, 123.7, 120.4, 89.9, 88.9, 35.6, 33.6, 33.4, 31.9, 29.3, 23.8,
22.7, 22.3, 15.6, 14.1, 14.0, ꢂ3.1. Anal. Calcd for C28H40Si: C, 83.1; H,
10.0; Si, 6.9. Found: C, 82.9; H, 9.8.
A 500 mL three-necked flask was equipped with a reflux
condenser, a three-way stopcock, and a magnetic stirring bar.
Dichlorobis(triphenylphosphine) palladium (0.030 g, 0.043 mmol),
cuprous iodide (0.065 g, 0.34 mmol), and triphenylphosphine
(0.052 g, 0.20 mmol) were placed in the flask. After the flask was
flushed with nitrogen, p-iodoisopropylbenzene (5.0 g, 20 mmol) and
triethylamine (200 mL) were added, and then a solution of p-dime-
thyloctylsilylphenylacetylene (5.5 g, 20 mmol) in triethylamine
(50 mL) was applied. The mixture was stirred for 2 h at room
temperature. After the triethylamine was evaporated, ether
(ca. 200 mL) was added, and the insoluble salt was filtered off. The
solution was washed with HCl aq. (1.0 M) three times. The ethereal
solution was dried over anhydrous sodium sulfate. After filtration,
ether was evaporated, and the crude product was purified by silicagel
column chromatography (eluent: hexane) to give the desired product
(5.6 g, 71%) as colorless liquid.1H NMR (CDCl3, ppm): 7.47 (m, 6H, Ar),
7.19 (d, J ¼ 8.0 Hz, 2H, Ar), 2.90 (sept, J ¼ 6.9 Hz, 1H, ArCH), 1.25 (m,
12H, SiCH2(CH2)6), 1.24 (d, J ¼ 6.9 Hz, 6H, ArCH(CH3)2), 0.87
(t, J ¼ 7.1 Hz, 3H, SiCH2(CH2)6CH3), 0.73 (t, J ¼ 7.8 Hz, 2H, SiCH2), 0.25
(s, 6H, SiCH3). 13C NMR (CDCl3, ppm): 149.2, 140.1, 133.4, 131.6, 130.6,
126.4, 123.7, 120.6, 90.0, 88.9, 34.1, 33.6, 31.9, 29.3, 23.8, 23.8, 22.7,
15.6, 14.1, ꢂ3.1. Anal. Calcd for C27H38Si: C, 83.0; H, 9.8; Si, 7.2. Found:
C, 83.0; H, 9.9.
2.4. Polymerization
Polymerization was carried out in a glass tube equipped with
a three-way stopcock under dry nitrogen. Unless otherwise specified,
the reaction was carried out at 80 ꢀC for 24 h under the following
conditions: [M]0 ¼ 0.20 M, [TaCl5] ¼ 20 mM, and [n-Bu4Sn] ¼ 80 mM.
A detailed procedure of polymerization is as follows: The monomer
solutionwas prepared in a glass tube. Another glass tube was charged
with TaCl5, n-Bu4Sn, and toluene; this catalyst solution was aged at
80 ꢀC for 10 min, and then monomer solution was added to it. Poly-
merization was run at 80 ꢀC for 24 h, which was quenched with
a small amount of methanol. The resulting polymer was isolated by
precipitation into a large excess of methanol, and its yield was
determined gravimetrically.
2.5. Membrane fabrication and desilylation
Membranes (thickness ca. 20–100 mm) of polymers (2a, 2b, 2c,
2d, 2B, and 2C) were fabricated by casting their toluene solutions
(conc. 0.30–0.80 wt%) into Petri dishes at room temperature. The
dish was covered with a glass vessel to slow solvent evaporation
(3–5 days). After a membrane was formed, the membrane was
pealed off, and it was immersed in methanol for 24 h and dried to
constant weight at room temperature. As shown in the literature,
the desilylation of membranes of 2a–d was carried out using
trifluoroacetic acid [4]. A detailed procedure is as follows: The
polymer membrane was immersed in trifluoroacetic acid at room
temperature for 24 h. To remove residual impurities in polymer
matrix, the membrane was immersed in acetone followed by
methanol at room temperature for 24 h. The membrane was dried
at room temperature under atmospheric pressure for 24 h. The
completion of desilylation was confirmed by the comparison
between IR spectra of membranes before and after the reaction.
2.3.2. 1-(p-t-butyl)phenyl-2-(p-dimethyloctylsilyl)phenylacetylene
(1c)
This monomer was prepared by the same method as for 1b
using p-t-butyliodobenzene instead of p-iodoisopropylbenzene.
Yield 54%, colorless liquid. 1H NMR (CDCl3, ppm): 7.50 (m, 6H, Ar),
7.38 (d, J ¼ 8.3 Hz, 2H, Ar), 1.34 (s, 9H, CCH3), 1.33 (m, 12H,
SiCH2(CH2)6), 0.90 (t, J ¼ 6.7 Hz, 3H, SiCH2(CH2)6CH3), 0.76
(t, J ¼ 7.8 Hz, 2H, SiCH2), 0.27 (s, 6H, SiCH3). 13C NMR (CDCl3, ppm):
151.5, 140.1, 133.4, 131.3, 130.6, 125.3, 123.7, 120.3, 89.9, 88.9, 34.8,
33.6, 31.9, 31.2, 29.2, 23.8, 22.7, 15.6, 14.1, ꢂ3.1. Anal. Calcd for
C28H40Si: C, 83.1; H, 10.0; Si, 6.9. Found: C, 82.9; H, 10.2.
2.3.3. 1-(p-n-butyl)phenyl-2-(p-dimethyloctylsilyl)phenylacetylene
(1d)
This monomer was prepared by the same method as for 1b
using p-n-butyliodobenzene instead of p-iodoisopropylbenzene.
Yield 79%, colorless liquid. 1H NMR (CDCl3, ppm): 7.50 (m, 6H, Ar),
7.18 (d, J ¼ 8.2 Hz, 2H, Ar), 2.64 (t, J ¼ 7.6 Hz, 2H, ArCH2), 1.62 (quint,
J ¼ 7.7 Hz, 2H, ArCH2CH2), 1.38 (sext, J ¼ 7.5 Hz, 2H, ArCH2CH2CH2),
2.6. Fractional free volume (FFV) of polymer membranes
The densities of membranes were determined by hydrostatic
weighing using a Mettler Toledo balance and a density determination

1282
A. Takeda et al. / Polymer 51 (2010) 1279–1284
kit. In this method, a liquid with known density (r0) is needed, and
the membrane density ( ) is given by the following equation:
2C have enough high molecular weights to fabricate free-standing
membranes.
r
r
¼
r0 ꢁ MA=ðMA ꢂ MLÞ
3.2. Solvent solubility, preparation, and desilylation of membranes
where MA is membrane weight in air and ML is membrane weight in
the auxiliary liquid. An aqueous sodium nitrate was used as the
auxiliary liquid. FFV is calculated by the following equation:
The solubility of the polymers is summarized in Table 2. All the Si-
containing polymers 2a–d completely dissolved in common organic
solvents such as hexane, CCl4, toluene, diethyl ether, chloroform, and
THF. The polymers without silyl groups (2B–D) also exhibited good
solubility. Tough free-standing membranes could be fabricated by
casting polymers from their toluene solutions. The desilylation of
membranes of 2a–d was carried out in trifluoroacetic acid at room
temperature for 24 h to give the corresponding desilylated
membranes (3a–d). The completion of desilylation was confirmed by
IR spectra of the polymer membranes. Fig. 1 shows the IR spectra of
membranes of 2b, 3b, and 2B. The absorptions at 1250 cmꢂ1 derived
from stretching of SiC–H bonds and at 1120 cmꢂ1 derived from
vibration of Si–C completely disappeared in the spectrum of
membrane of 3b. The spectrum of 3b agreed well with that of 2B,
which was obtained directly by the polymerization of the monomer
without a dimethyloctylsilyl group. After desilylation, weights of the
polymer membranes decreased to the values anticipated for desily-
lation, indicating the completion of the reaction.
The solubility of the desilylated polymers 3a–d is shown in Table
2. The desilylated polymers 3a–d showed less solubility than Si-
containing polymers 2a–d. The polymer 3a, which has no alkyl
substituent on phenyl groups, was insoluble in any solvents, while
2a was soluble in various organic solvents. The desilylated poly-
mers 3b–d have the identical chemical structures to 2B–D,
respectively. However, the desilylated polymers 3b and 3c exhibi-
ted somewhat poor solubility than 2B and 2C, respectively. The
desilylated polymer 3b was partly soluble in CCl4, toluene, CHCl3,
and THF, although 2B was completely dissolved in these solvents.
The desilylated polymer 3c partially dissolved in Et2O, which was
a good solvent to 2C. This may be accounted for by the idea that the
geometric structures of 3b and 3c are somewhat different from
those of 2B and 2C, respectively. The main chains of 3b and 3c are
generated by the polymerization of 1b and 1c which have bulky
silyl group, so the structures of 3b and 3c would have high
stereoregularity compared to the polymers 2B and 2C. The
metathesis polymerization of substituted acetylene produces the
polyacetylene main chain composed of a mixture of cis-form and
trans-form [17]. Khotimsky et al. reported that the geometric
structure of the substituted acetylene polymer is defined by
a combination of the size of substituent, kind of transition metal,
polymerization solvent, and temperature [18]. They determined the
ratios of cis and trans structures of poly[1-(trimethylsilyl)-1-pro-
pyne] and poly[1-(trimethylgermyl)-1-propyne] by using 13C NMR.
In order to examine the structures of main chains of 2B and 3b, 13C
NMR spectra of the polymers were measured. However, no peaks
À
Áꢀ
À
Áꢀ
FFV ¼ vsp ꢂ v0 vspz vsp ꢂ 1:3vw vsp
where vsp is the polymer specific volume, and v0 is the occupied
volume of the polymer. The occupied volume is typically estimated
as 1.3 times the van der Waals volume (vw), which is calculated
using the group contribution method [14].
3. Results and discussion
3.1. Polymerization
The polymerizations of monomers (1a–d and 1B–D) were
carried out by using TaCl5/n-Bu4Sn catalyst in toluene at 80 ꢀC. It is
well known that TaCl5/n-Bu4Sn catalyst achieves good yields of
disubstituted acetylene polymers with high molecular weights
[15–17], which is essential for fabrication of tough free-standing
membranes. The results of polymerizations are summarized in
Table 1.
The polymerization of 1a having dimethyloctylsilyl group
produced a polymer 2a with high molecular weight in good yield
(Mw 2.68 ꢁ 106, yield 84%) (Run 1). Monomer 1b having both
dimethyloctylsilyl and i-propyl groups polymerized under the same
condition to give a polymer 2b with pretty high molecular weight.
The formed polymer 2b contained insoluble parts in any solvents,
and the Mw of THF-soluble product exceeded three million (Run 2).
The high polymerizability of 1b may be explained by the high
electron density of triple bond. Monomer 1b have both a silyl group
and an alkyl groups at para position of benzene rings as electron-
donating groups. When the initial monomer concentration was
lowered to 0.10 M, the polymerization of 1b gave a high molecular
weight polymer 2b totally soluble in common solvents (Run 3).
Membrane preparation and properties of 2b were examined using
the sample obtained by the polymerization with the lower initial
monomer concentration (Run 3). The polymerizations of 1c and 1d
produced polymers 2c and 2d, respectively, in good yields. The
weight-average molecular weights of 2c and 2d were 1.40 ꢁ 106
and 1.91 ꢁ106, respectively. Although 1c and 1d also have both silyl
group and an alkyl group, the bulkier alkyl groups may decrease
their polymerizability. The polymerizations of 1B and 1C without
dimethyloctylsilyl groups were carried out in the same method as
1a to give polymers 2B and 2C, respectively. The polymers 2B and
Table 1
Polymerizations of monomers.a
Table 2
Solubility of polymers.a
Yieldb [%]
Mw ꢁ 10ꢂ6c
Mw/Mn
c
Run
Monomer
[M]0
solvent
2a
2b
2c
2d
2B
2C
2Db
3a
3b
3c
3d
1
2
3
4
5
6
7
1a
1b
1b
1c
1d
1B
1C
0.20
0.20
0.10
0.20
0.20
0.20
0.20
84
76
77
59
2.68
3.13
2.08
1.40
1.91
1.76
0.836
8.60
8.39
5.19
5.10
4.90
10.4
4.49
hexane
CCl4
toluene
Et2O
CHCl3
THF
acetone
DMF
þ
þ
þ
þ
þ
þ
ꢂ
ꢂ
ꢂ
þ
þ
þ
þ
þ
þ
ꢂ
ꢂ
ꢂ
þ
þ
þ
þ
þ
þ
ꢂ
ꢂ
ꢂ
þ
þ
þ
þ
þ
þ
ꢂ
ꢂ
ꢂ
ꢂ
þ
þ
ꢂ
þ
þ
ꢂ
ꢂ
ꢂ
ꢂ
þ
þ
þ
þ
þ
ꢂ
ꢂ
ꢂ
ꢂ
þ
þ
ꢃ
þ
þ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢃ
ꢃ
ꢂ
ꢃ
ꢃ
ꢂ
ꢂ
ꢂ
ꢂ
þ
þ
ꢃ
þ
þ
ꢂ
ꢂ
ꢂ
ꢂ
þ
þ
ꢃ
þ
þ
ꢂ
ꢂ
ꢂ
76
84d
73d
a
Polymerized with TaCl5–n-Bu4Sn in toluene at 80 ꢀC for 24 h; [TaCl5]0 ¼ 20 mM,
[n-Bu4Sn]0 ¼ 80 mM.
DMSO
b
Acetone-insoluble product.
Measured by GPC.
Methanol-insoluble product.
c
a
Symbols: þ; soluble, ꢃ; partly soluble, ꢂ; insoluble.
d
b
Data from Ref. [10].

A. Takeda et al. / Polymer 51 (2010) 1279–1284
1283
polymer membranes exhibited almost the same oxygen perme-
ability despite the fact that they each have different kinds of alkyl
groups. With regard to nitrogen and carbon dioxide, no obvious
difference of permeability was observed between 2a–d. Gas
permeability of Si-containing membranes seems to be controlled
mainly by dimethyloctylsilyl groups. The densities and FFV of
Si-containing membranes were also nearly the same. Their FFV
values ranged from 0.154 to 0.182, which are small compared to
poly(diphenylacetylene)s without long alkyl groups. For instance,
the FFV values of poly[1-(p-trimethylsilyl)phenyl-2-phenyl-
acetylene] and poly[1-(p-t-butyl)phenyl-2-phenylacetylene] are
reported to be 0.26 [19] and 0.27 [20], respectively. The long alkyl
chains of dimethyloctylsilyl groups would occupy the free volume
in the polymer matrix because of the flexibility of octyl groups.
The desilylated membranes of 3a–c showed larger oxygen
permeability than Si-containing membranes. The P(O2) values of
3a–c were 1200, 300, and 820 barrers, respectively. In addition, the
desilylation lead to large increases in the FFV of membranes. The
long alkyl chains of dimethyloctylsilyl groups were disadvanta-
geous to gas permeation because linear alkyl chains occupy the free
volume in Si-containing membranes. The gas permeability coeffi-
cient and FFV of desilylated membrane of 3d having n-butyl groups
was 37 barrers and 0.152, respectively, which are similar to those of
2d. This is because the free volume is still occupied by the flexible
n-butyl groups even after elimination of dimethyloctylsilyl groups.
3.4. CO2 diffusivity and solubility
The CO2 diffusion coefficients (D(CO2)) were measured by time
lag method, and the CO2 solubility coefficients (S(CO2)) were
calculated using P and D values (Table 3). Unfortunately, the time
lags on oxygen and nitrogen permeability measurement were so
small that D(O2) and D(N2) values could not be determined. The
desilylated membranes of 3a–c showed much higher CO2 diffu-
sivity than Si-containing counterparts. The CO2 solubility of these
membranes greatly increased through desilylation. The increases of
diffusivity and solubility would be due to the increment of FFV. The
elimination of dimethyloctylsilyl groups of membranes ought to
affect gas solubility more effectively than gas diffusivity. The CO2
solubility coefficient of 3a, for instance, was approximately 15 times
as large as that of 2a, while the diffusivity coefficient of 3a was
approximately 1.7 times of that of 2a.
However, membrane of 3d exhibited unanticipated behavior of
variation in CO2 diffusivity and solubility. Desilylation of membrane
of 2d caused decrease of diffusivity and increase of solubility even
though the FFV scarcely changed by desilylation. Gas diffusivity of
polymer membranes is determined by local mobility of substitu-
ents as well as free volume of membranes [21]. It is reported that
flexible alkyl groups such as an octyl group exhibit relatively large
local mobility [22]. Therefore, the decrease of CO2 diffusivity
through desilylation can be explained by a lack of dimethyloctylsilyl
Fig. 1. IR spectra of membranes before and after desilylation (2b, 3b, and 2B).
were observed because of too high molecular weight of the poly-
mers. The desilylated polymer 3d having n-butyl groups showed
the same solubility as 2D, and they were soluble in CCl4, toluene,
CHCl3, and THF, and partly soluble in Et2O. Polymers 3d and 2D
might have different geometric structures, but they have linear
alkyl (n-butyl) groups, which improve the solubility of polymers.
Therefore, a conspicuous difference of solubility between 3d and
2D was not observed.
3.3. Gas permeability
The permeability of Si-containing polymer membranes (2a–d)
and desilylated membranes (3a–d) to various gases was examined
at 25 ꢀC (Table 3). Oxygen permeability coefficients (P(O2)) of 2a–d
were 34, 29, 26, and 25 barrers, respectively. The Si-containing
Table 3
Membranes properties of Si-containing and Desilylated membranes.
Membrane
2
3
P(O2)a P(N2)a P(CO2)a D(CO2)b ꢁ 108 S(CO2)c ꢁ 103 Density
FFV
P(O2)a P(N2)a P(CO2)a D(CO2)b ꢁ 108 S(CO2)c ꢁ 103 Density
FFV
(g cmꢂ3
)
[g cmꢂ3
]
a
b
c
34
29
26
25
11
9.1
7.5
7.9
170
120
100
110
490
190
430
550
3.4
6.1
2.8
2.6
0.975
0.995
0.973
0.979
0.182 1200
740
150
360
11
4200
1100
3000
160
810
400
770
100
52
27
39
15
0.966
1.04
0.890
1.07
0.283
0.191
0.296
0.152
0.154
0.169
0.164
300
820
37
d
a
Measured at 25 ꢀC. In the unit of barrer [1 barrer ¼ 1 ꢁ 10ꢂ10 cm3(STP) cm cmꢂ2 sꢂ1 cm Hgꢂ1].
b
c
Measured at 25 ꢀC by the ‘time lag’ method. In the units of cm2 sꢂ1
.
Calculated from the equation S ¼ P/D. In the units of cm3(STP) cmꢂ3 cm Hg.

1284
A. Takeda et al. / Polymer 51 (2010) 1279–1284
Table 4
4. Conclusions
Comparison of membrane properties of polymers without Si-groups.
P(O2)a
P(O2)/P(N2)
D(CO2)b ꢁ 108
S(CO2)c ꢁ 103
FFV
Novel poly(diphenylacetylene)s having both dimethyloctylsilyl
and various alkyl groups were synthesized, and free-standing
membranes of the produced polymers could be fabricated by
solution-casting. The desilylation of these membranes proceeded
quantitatively. All the polymer membranes with dimethyloctylsilyl
groups exhibited almost the same gas permeability. The dimethy-
loctylsilyl groups would control gas permeability of the polymers.
The desilylation of polymer membranes containing no alkyl group
(2a) and branched alkyl groups (i-propyl (2b) and t-butyl (2c))
resulted in increases of gas permeability and FFV. The CO2 solubility
increased more significantly than the CO2 diffusivity through
desilylation of these membranes. The desilylated membrane having
n-butyl groups (3d) exhibited nearly the same gas permeability
coefficient as Si-containing counterpart (2d) did. The desilylation of
2d increased CO2 solubility, and decreased CO2 diffusivity. The
desilylated membranes showed almost the same gas permeability
as the membranes prepared by re-casting of the desilylated poly-
mers. This indicates that microvoid was not generated by the
elimination of silyl groups in a solid state on poly[1-(p-dimethy-
loctylsilyl)phenyl-2-phenylacetylene] with linear or branched alkyl
groups.
3b
4b
2B
3c
300
320
350
820
910
780
37
2.0
2.1
2.2
2.3
2.1
2.3
3.4
3.5
3.0
400
240
490
770
780
490
100
120
200
27
63
31
39
46
65
15
10
22
0.191
0.251
0.261
0.296
0.277
0.272
0.152
0.146
0.202
4c
2C
3d
4d
2Dd
28
89
Measured at 25 ꢀC. In the unit of barrer [1 barrer ¼ 1 ꢁ 10ꢂ10 cm3
a
(STP) cm cmꢂ2 sꢂ1 cm Hgꢂ1].
Measured at 25 ꢀC by the ‘time lag’ method. In the units of cm2 sꢂ1
.
b
Calculated from the equation S ¼ P/D. In the units of cm3(STP) cmꢂ3 cm Hgꢂ1
.
c
d
Data from Ref. [10].
groups with large local mobility. The S(CO2) of 3d was much larger
than that of 2d, as is the case for 2a–c. This implies that dimethy-
loctylsilyl groups lower gas solubility performance of polymer
membrane.
3.5. Comparison of three types of membranes
Polymer membranes of 4b–d were fabricated by re-casting
using toluene solutions of the desilylated polymers (3b–d,
respectively). Polymer 3b did not dissolve completely in toluene,
and hence membrane of 4b was prepared by using toluene-soluble
parts of 3b. The other kind of membranes without dimethy-
loctylsilyl groups were fabricated from 2B and 2C, which were
obtained directly by the polymerization of the corresponding
monomers 1B and 1C. The desilylated membrane of 3b has an
identical chemical structure with 4b and 2B, but they were
prepared by different methods. The oxygen permeability coeffi-
cients, permselectivity, CO2 diffusivity, CO2 solubility, and FFV of
membranes without dimethyloctylsilyl group (3b–d, 4b–d, and
2B–D) are listed in Table 4.
The P(O2) values of membranes of 3b, 4b, and 2B were 300, 320,
and 350 barrers, respectively, which were similar to each other. The
P(O2) value of 3c having t-butyl groups was also almost the same as
those of 4c and 2C. These results suggest that excess free volume
was not generated by desilylation in a solid state. The membranes
of 3d and 4d, which contain n-butyl groups, showed nearly the
same oxygen permeability, while the P(O2) value of 2D was obvi-
ously larger than those of 3d and 4d. The FFV of 2D was also much
larger than those of 3d and 4d. High oxygen permeability and FFV of
2D would be attributed to the difference of geometric structure of
polymer main chain. The polymer 2D was synthesized directly from
the monomer without Si-group. Therefore, the polymer 2D should
has different geometric structure from the polymers 3d and 4d. The
polymer 3b and 3c also would have different geometric structures
from 2B and 2C, respectively, as described in the section of solvent
solubility. However, they have large FFV owing to the steric repul-
sion of the branched alkyl groups. Hence the difference of
geometric structure may not affect their FFV and gas permeability
significantly.
References
[1] Nagai K, Masuda T, Nakagawa T, Freeman BD, Pinnau I. Prog Polym Sci
2001;26:721.
[2] Masuda TJ. Polym Sci Part A Polym Chem 2007;45:165.
[3] Sakaguchi T, Yumoto K, Shida Y, Shiotsuki M, Sanda F, Masuda T. J Polym Sci
Part A Polym Chem 2006;44:5028.
[4] Teraguchi M, Masuda T. Macromolecules 2002;35:1149.
[5] Sakaguchi T, Yumoto K, Shiotsuki M, Sanda F, Yoshikawa M, Masuda T.
Macromolecules 2005;38:2704.
[6] Sakaguchi T, Kwak G, Masuda T. Polymer 2002;43:3937.
[7] Sakaguchi T, Shiotsuki M, Masuda T. Macromolecules 2004;37:4104.
[8] Sakaguchi T, Shiotsuki M, Sanda F, Freeman BD, Masuda T. Macromolecules
2005;38:8327.
[9] Sakaguchi T, Shiotsuki M, Sanda F, Masuda T. J Membr Sci 2006;280:720.
[10] Takeda A, Sakaguchi T, Hashimoto T. Polymer 2009;50:5031.
[11] Aoki T, Nakahara H, Hayakawa Y, Kokai M, Oikawa E. J Polym Sci Part A Polym
Chem 1994;32:849.
[12] Kouzai H, Masuda T, Higashimura T.
1994;32:2523.
J Polym Sci Part A Polym Chem
[13] Sonogashira K. J Organomet Chem 2002;653:46.
[14] van Krevelen KW. Properties of polymer: their correlation with chemical
structure; their numerical estimation and prediction from additive group
contributions. 3rd ed. Amsterdam: Elsevier Science; 1990. p. 71–107.
[15] Masuda T, Sanda F, Shiotsuki M. In: Crabtree RH, Mingos DMP, editors.
Comprehensive organometallic chemistry III. Oxford: Elsevier; 2007 [chapter
11.16].
[16] Nomura R, Masuda T. In: Kroshritz JI, editor. Encyclopedia of polymer science
and technology. 3rd ed., vol. IA. Wiley; 2003. p. 1.
[17] Masuda T, Sanda F. In: Grubbs RH, editor. Handbook of metathesis, vol. 3.
Weinheim: Wiley-VCH; 2003 [chapter 3.11].
[18] Khotimsky VS, Tchirkova MV, Litvinova EG, Rebrov AI, Bondarenko GN. J Polym
Sci Part A 2003;41:2133.
[19] Nagai K, Freeman BD, Pinnau I, He Z, Masuda T, Teraguchi M, et al. Macro-
molecules 2000;33:2516.
[20] Pinnau I, He Z, Masuda T, Sakaguchi T. In: Pinnau I, Freeman BD, editors.
Advanced materials for membrane separation. ACS symposium series 876.
Washington, DC: American Chemical Society; 2004. p. 167–76.
[21] Pinnau I, Morisato A, He Z. Macromolecules 2004;37:2823.
[22] Kanaya T, Tsukushi I, Kaji K, Sakaguchi T, Kwak G, Masuda T. Macromolecules
2002;35:5559.