Polymers of intrinsic microporosity with dinaphthyl and thianthrene segments

Article abstract of DOI:10.1021/ma101930x

Novel intrinsically microporous homopolymers and copolymers derived from PIM-1 monomers (5,5,6,6-tetrahydroxy-3,3,3,3-tetramethylspirobisindane and 2,3,5,6-tetrafluoroterephthalonitrile) with two additional monomers- tetrahydroxydinaphthyl and tetrafluorotetraoxide thianthrene-are reported as potential materials for membrane-based gas separations. The resulting copolymers prevent efficient space packing of the stiff polymer chains and consequently exhibit analogous behavior to that of PIM-1, the most widely reported polymer in this class of materials. In addition, the copolymerization provides high molecular weight copolymers and low polydispersity if the polymerization reactions were conducted at elevated temperature for an extended period of time. Detailed structural characterization of the new monomers and polymers was determined by ¹H and ¹⁹F nuclear magnetic resonance spectroscopy (NMR). The thermal properties were detected by differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA). Polymer free volume was calculated from the polymer density and specific van der Waals volume. Under the same testing conditions, the homopolymer containing thianthrene units and most of the analogous copolymers have an excellent combination of properties with good film-forming characteristics. The gas transport properties show higher selectivity for gas pairs such as O ₂/N₂, CO₂/N₂, and H ₂/N₂ with a corresponding decrease in permeability compared to PIM-1. This work also demonstrates that significant improvements in properties may be obtained through copolymers of intrinsic microporosity (CoPIM)s. Furthermore, this work extends the spectrum of high molecular weight soluble PIMs beyond those reported previously.

Full text of DOI:10.1021/ma101930x

8580 Macromolecules 2010, 43, 8580–8587
DOI: 10.1021/ma101930x
Polymers of Intrinsic Microporosity with Dinaphthyl and
Thianthrene Segments†
Naiying Du,‡ Gilles P. Robertson,‡ Ingo Pinnau,§ and Michael D. Guiver*,‡,^
‡Institute for Chemical Process and Environmental Technology, National Research Council of Canada,
Ottawa, Ontario K1A 0R6, Canada, §Advanced Membranes and Porous Materials Center,
King Abdullah University of Science and Technology, Thuwal 23955-6900, Kingdom of Saudi Arabia, and
^WCU Department of Energy Engineering, Hanyang University, Seoul 133-791, South Korea
Received August 24, 2010; Revised Manuscript Received September 11, 2010
ABSTRACT: Novel intrinsically microporous homopolymers and copolymers derived from PIM-1 mono-
mers (5,50,6,60-tetrahydroxy-3,3,30,30-tetramethylspirobisindane and 2,3,5,6-tetrafluoroterephthalonitrile)
with two additional monomers;tetrahydroxydinaphthyl and tetrafluorotetraoxide thianthrene;are reported
as potential materials for membrane-based gas separations. The resulting copolymers prevent efficient space
packing of the stiff polymer chains and consequently exhibit analogous behavior to that of PIM-1, the most
widely reported polymer in this class of materials. In addition, the copolymerization provides high molecular
weight copolymers and low polydispersity if the polymerization reactions were conducted at elevated
temperature for an extended period of time. Detailed structural characterization of the new monomers and
1
polymers was determined by H and 19F nuclear magnetic resonance spectroscopy (NMR). The thermal
properties were detected by differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA).
Polymer free volume was calculated from the polymer density and specific van der Waals volume. Under the
same testing conditions, the homopolymer containing thianthrene units and most of the analogous copolymers
have an excellent combination of properties with good film-forming characteristics. The gas transport
properties show higher selectivity for gas pairs such as O2/N2, CO2/N2, and H2/N2 with a corresponding
decrease in permeability compared to PIM-1. This work also demonstrates that significant improvements in
properties may be obtained through copolymers of intrinsic microporosity (CoPIM)s. Furthermore, this work
extends the spectrum of high molecular weight soluble PIMs beyond those reported previously.
Scheme 1. Structure of PIM-1
Introduction
Since 2004, a novel class of polymers, termed “polymers of
intrinsic microporosity” (PIMs), have been prepared that reveal
unusually high free volume and high surface areas.1-4 Different
from other polymeric microporous materials, such as solvent
swollen polymers,5,6 polymer networks with rigid units,7-11 or
polymers with bulky groups,12,13 PIMs have special ladder-type
main-chain structures with contorted or kinked centers which
prevent efficient space packing and pore collapse, resulting in
intrinsically microporosity materials. Even compared with tradi-
tional microporous materials, such as zeolites or activated carbons
which are widely used industrially, PIMs present a unique advan-
tage due to the combination of processability with a high degree of
microporosity, which are attractive for applications such as
adsorbents and catalysis. Solvent-soluble film-forming PIMs of
high molecular weight are particularly promising as gas separation
materials on account of their high gas permeability. A number of
published examples of PIMs have shown gas permeability data
that lie above the Robeson upper bound for the separation of
important gas pairs (O2/N2, CO2/N2), moderate to good selec-
tivity, and high permeability.1-4 In addition, PIMs also exhibit
excellent performance for hydrocarbon/methane and hydrocar-
bon/hydrogen separation.14 The most well-known polymer of this
type is referred to as PIM-1, which is derived by step-growth
polymerization of 5,50,6,60-tetrahydroxy-3,3,30,30-tetramethyl-
spirobisindane (TTSBI), and 2,3,5,6-tetrafluoroterephthalonitrile
(TFTPN). Although PIM-1 has a spiro-contorted center and a
rigid ladder-like structure (Scheme 1), it can be readily cast from
solution into mechanically strong free-standing films. To date, a
number of studies on PIMs have been reported by several research
groups;4,15-17 however, there are still a limited number of reported
monomers which include spiro/contorted sites or tetrafluoro
groups similar to PIM-1 monomers and which provide high
molecular weight solvent-soluble PIMs suitable for membrane
fabrication and gas permeability measurement.
It is well-known that gas transport properties of polymeric
membrane can be tuned by modifying or tailoring the polymer
structure.12,19-21 In our previous work, we reported several
soluble high molecular weight PIMs derived from TTSBI and
new tetrafluoro monomers which contain trifluoromethyl or
sulfone groups as full or partial replacement of the tetrafluor-
odinitrile monomer.4,18 Although the distance between the spiro
centers on the chains was the same as PIM-1, the gas permeability
and selectivity of the new PIMs could be tuned by the type of
pendent groups as a consequence of space filling and rigidity
† NRCC Publication No. 52231.
*Corresponding author. E-mail: michael.guiver@nrc-cnrc.gc.ca.
pubs.acs.org/Macromolecules
Published on Web 10/01/2010
r 2010 American Chemical Society

Article
Macromolecules, Vol. 43, No. 20, 2010 8581
Scheme 2. Synthetic Route to Copolymers (DNPIMs, TOTPIMs, and DNTOTPIMs) Based on PIM-1 Monomers (B, C) and THDN (A) and TFTOT (D)
increase. In the present work, we focus on the design of new
ladder-type PIM structures which contain other highly kinked
units and/or those having a longer distance between the twist or
spiro centers to improve the understanding of structure-prop-
erty relationships. The new PIMs are derived from TTSBI,
TFTPN, 2,203,30-tetrahydroxy-1,10-dinaphthyl (THDN), and a
newly reported tricyclic monomer: 2,3,7,8-tetrafluoro-5,50,10,100-
tetraoxide thianthrene (TFTOT). In an earlier communication,
we reported a single copolymer composition based on THDN
(DNPIM-33).22 Attempts will be made to correlate the structure
of these new PIMs with the degree of microporosity achieved, as
assessed by gas adsorption, and gas permeability. The spatially
twisted dinaphthyl unit has a different contortion angle com-
pared with the spirobisindane unit, whereas the tricyclic TFTOT
unit provides a longer distance between the centers of contortion
than the TFTPN unit.
aluminum chloride, chromium oxide (CrO3), anhydrous potassium
carbonate (K2CO3), tetrahydrofuran (THF), acetic acid, and
toluene were all obtained from Sigma-Aldrich and were used as
received. The monomer 5,50,6,60-tetrahydroxy-3,3,30,30-tetramethyl-
spirobisindane (Sigma-Aldrich) was purified by crystallization from
methanol. The monomer 2,3,5,6-tetrafluoroterephthalonitrile
(Matrix scientific) was purified by vacuum sublimation at 150 ꢀC
under an inert atmosphere.
Characterization Methods. The structures of the two synthe-
sized monomers and PIMs were fully characterized using NMR
spectroscopy. NMR analyses were recorded on a Varian Unity
Inova spectrometer at a resonance frequency of 399.961 MHz
for 1H and 376.276 MHz for 19F. 1H and 19F NMR spectra were
obtained from samples dissolved in CDCl3 or DMSO-d6 using a
5 mm pulsed field gradient indirect detection probe. The solvent
signals (CDCl3 1H 7.25 ppm; DMSO-d6 1H 2.50 ppm) were used
as the internal references. An external reference was used for 19
NMR: CFCl3 0 ppm.
F
Experimental Section
Elemental analysis was carried out with a Thermoquest
CHNS-O elemental analyzer.
Molecular weight and molecular weight distributions were
measured by GPC using Ultrastyragel columns and THF as the
Materials. 2,3-Naphthalenediol (Acros), ferric chloride hexa-
hydrate (FeCl3 6H2O, Anachemia), and 1,2-difluorobenzene
(Oakwood) were used as received. Sulfur monochloride, anhydrous
3

8582 Macromolecules, Vol. 43, No. 20, 2010
Table 1. Properties of the CoPIMs Series and PIM-1 Polymerized at 160 ꢀC for 120 min
Du et al.
polymers
THDN (mol ratio)
TTSBI (mol ratio)
TFTPN (mol ratio)
TFTOT (mol ratio)
Mn
Mw
Mw/Mn
DNPIM-100a
DNPIM-66a
DNPIM-50
DNPIM-33
DNPIM-25
1
2
1
1
1
0
0
0
0
0
0
1
1
1
1
0
0
1
1
2
3
1
3
2
3
4
5
1
2
3
4
1
1
3
2
3
4
0
1
1
2
3
4
1
2
3
4
1
0
0
0
0
0
1
2
1
1
1
1
1
1
1
1
0
32 000
45 000
47 000
56 000
65 000
99 000
135 000
118 000
129 000
137 000
3.1
3.0
2.5
2.3
2.1
TOTPIM-100b
TOTPIM-66
TOTPIM-50
TOTPIM-33
TOTPIM-25
TOTPIM-20
DNTOTPIM-50
DNTOTPIM-33
DNTOTPIM-25
DNTOTPIM-20
PIM-1 (40 min)4
35 000
41 000
30 000
42 000
40 000
14 000
38 000
56 000
42 000
55 000
74 000
89 000
70 000
84 000
81 000
32 000
74 000
113 000
110 000
85 000
2.1
2.2
2.3
2.0
2.1
2.3
1.9
2.0
2.6
1.6
a After purification. b The molecular weight of TOTPIM-100 was not tested by GPC, since the polymer was not soluble in the THF eluent.
eluent at a flow rate of 1 mL/min. The values obtained were
determined by comparison with a series of polystyrene standards.
Polymer thermal degradation curves were obtained from
TGA (TA Instruments model 2950). Polymer samples for
TGA were initially heated to 120 ꢀC under nitrogen gas and
maintained at that temperature for 1 h for moisture removal and
then heated to 600 at 10 ꢀC/min for degradation temperature
measurement. The polymers were tested for the occurrence of
glass transition temperatures (Tg) from DSC measurements (TA
Instruments model 2920); the samples were heated at 10 ꢀC/min
under a nitrogen flow of 50 mL/min, then quenched with liquid
nitrogen, and reheated at 10 ꢀC/min for observation of the
presence of Tg.
Dense polymer films for gas permeability measurements were
prepared from 1 to 2 wt % polymer solutions in chloroform.
Polymer solutions were filtered through 0.45 μm poly-
(tetrafluoroethylene) filters and then cast into Teflon Petri
dishes in a glovebox and slowly evaporated for 1 day. The films
were boiled in water for 1 h, then soaked in methanol, and dried
in the vacuum oven at 100 ꢀC for 24 h. The resulting membranes
with thickness in the range of 60-80 μm were flexible and bright
yellow. The absence of residual solvent in the films was con-
firmed by a weight loss test using TGA.
Surface adsorption was evaluated by Brunauer-Emmett-
Teller measurements (BET) using an ASAP 2020 (Accelerated
Surface Area and Porosimetry System) instrument from Micro-
meritics Instrument Corp., and data analysis was performed
using ASAP 2020 version 3 software Rev H. Samples were
initially degassed conditions under vacuum at 80 ꢀC for 24 h.
Density measurements of polymeric membranes were by a
standard gravimetric method in ethanol. The polymers mem-
branes having thickness in the range of 70-90 μm were initially
dried under vacuum 100 ꢀC for 24 h.
Preparation of 2,20,3,30-Tetrahydroxy-1,10-dinaphthyl. A lite-
rature procedure was employed,23 whereby a mixture of 2,3-
naphthalenediol (16 g, 0.1 mol) and FeC13 6H2O (27 g, 0.2 mol)
3
was ground to a finely divided powder using an agate mortar
and pestle. The mixture was placed in a test tube and sonicated
with ultrasound at 50 ꢀC for 4 h. Decomposition of the reaction
product with dilute HC1 gave crude 2,203,30-tetrahydroxy-1,10-
dinaphthyl in 85% yield. The tetrol was recrystallized three
times from THF to give white needle crystals in 53% yield; mp
1
> 300 ꢀC. H NMR (DMSO-d6) δ: 6.80-6.82 (d, J = 8.0 Hz,
2H), 6.94-6.98 (t, J = 8.0 Hz, 2H), 7.14-7.18 (t J = 8.0 Hz,
2H), 7.24 (s, 2H), 7.64-7.62 (d, J = 8.0 Hz, 2H), 8.41 (s, OH),
10.07 (s, OH). Elem. Anal. Calcd for C20H14O4 (318.32 g/mol):
C, 75.46%; H, 4.43%. Found: C, 75.41%; H, 4.56%.
Preparation of 2,3,7,8-Tetrafluoro-5,50,10,100-tetraoxide Thia-
nthrene. 2,3,7,8-Tetrafluorothianthrene was synthesized accord-
ing to a modified procedure.24 Into a 250 mL three-neck flask
equipped with a magnetic stirrer, an argon inlet and a condenser,
difluorobenzene (20 mmol), AlCl3 (60 mmol), and dry dichloro-
methane (50 mL) were added. The reaction mixture was cooled to
0 to -20 ꢀC in an ice salt bath (NaCl/ice = 3:1, w/w) and stirred
for 1 h. Thereafter, the reaction mixture was added dropwise into
S2Cl2 (20 mmol), and the temperature was gradually increased to
room temperature. After stirring at room temperature for 2 h, the
reaction mixture was refluxed for 20 min and then poured into
water. The crude product was washed with 8 N hydrochloric acid
and extracted with dichloromethane and dried over MgSO4. After
removal of the dichloromethane, the 2,3,7,8-tetrafluorothian-
threne (5 g) was recrystallized from hexane to give white needle
crystals in a yield of 49%; mp =108 ꢀC. 1H NMR (chloroform-d)
δ: 7.309 (t, J = 8.0 Hz, 4H). 19F NMR (chloroform-d) δ: -136.4
(t, J = 8.0 Hz, 4F).
20 g of 2,3,7,8-tetrafluorothianthrene was oxidized using
CrO3 (excess) in acetic acid (100 mL) at a reaction temperature
maintained at 100 ꢀC for 24 h. The resulting white 2,3,7,8-
tetrafluoro-5,50,10,100-tetraoxide thianthrene was collected and
recrystallized from DMF to give white flake crystals in a yield
The mechanical properties of polymers were tested on a
model 5565 Instron materials tester. Film samples having
thickness in the range of 70-90 μm were cut by a standard
dumbbell-shaped die.
Permeability coefficients (P) of N2, O2, H2, and CO2 were
determined at 25 ꢀC with a feed pressure of 50 psig and atmo-
spheric permeate pressure (0 psig) using the constant-pressure/
variable-volume method. The permeate gas flow rate was mea-
sured by a mass flow meter (Agilent ADM 2000) or a bubble
flowmeter. The permeability (P) was calculated by using a
following equation:
1
of 92%; mp > 300 ꢀC. H NMR (DMSO-d6) δ: 8.087 (t, J =
8.0 Hz, 4H). 19FNMR (DMSO-d6) δ: -123.185 (t, J= 8.0Hz, 4F).
Elem. Anal. Calcd for C12H4F4O4S2 (352.28 g/mol): C, 40.91%; H,
1.14%, S 18.20%. Found: C, 41.09%; H, 1.14%; S 18.36%.
Preparation of PIM-1. PIM-1 was synthesized by polycon-
densation of TTSBI and TFTPN using the procedure reported
previously.4 The reactions were run for 40 min.
Preparation of DNPIMs, TOTPIMs, and DNTOTPIMs.
Two homopolymers and three series of CoPIMs were synthe-
sized by polycondensation of TTSBI, TFTPN, THDN, or
TFTOT (with different molar ratios) using a procedure similar
to that of PIM-1,4 but for a longer reaction time (120 min). The
properties of the homopolymers and CoPIMs are summarized
in Table 1.
ꢀ
ꢁꢀ ꢁꢀ
ꢁ
273 dV
l
P ¼
T
dt
ΔpA
where dV/dt is the permeate-side flow rate (cm3/s) and T is the
operation temperature (K). The membrane effective area was
9.6 cm2.

Article
DNPIM-100, H NMR (chloroform-d) δ (ppm): 7.88-6.67
Macromolecules, Vol. 43, No. 20, 2010 8583
1
(10H).
TOTPIM-100, 1H NMR (chloroform-d) δ (ppm): 7.54 (s, 2H),
7.45 (s, 2H), 6.67 (s, 2H), 6.31 (s, 2H), 2.33 (s, 2H), 2.12 (s, 2H), 1.33
(s, 6H), 1.23 (s, 6H).
Results and Discussion
The synthesis of THDN reported by Toda in 1989 is unex-
pectedly simple. It was carried out with 2,3-naphthalenediol in
the presence of a catalyst23 at 50 ꢀC under ultrasonic conditions
for 4 h. In the present work, it was found the optimum
temperature for this reaction is in the range of 50-70 ꢀC. Higher
temperatures led to more insoluble side products.
To the best of our knowledge, there is only one article that
mentions the synthesis of the monomer precursor 2,3,7,8-tetra-
fluorothianthrene; however, structural characterization of the
monomer is lacking.24 Although Budd et al. reported 2,3,7,8-
tetrachloro-5,50,10,100-tetraoxide thianthrene as a potential PIM
monomer in a patent,15 the reactivity of this monomer is
considerably lower than the tetrafluoro analogue. As a result,
the tetrachlorotetraoxide thianthrene monomer is unsuitable for
preparing PIMs with sufficiently high molecular weight. The
2,3,7,8-tetrafluorothianthrene precursor was oxidized using
CrO3/acetic acid instead of conventional oxidation reagents, such
as H2O2/acetic acid, because the solubility of the resulting 2,3,7,8-
tetrafluoro-5,50,10,100-tetraoxide thianthrene is poor, and it was
found that H2O2/acetic acid is too weak to effect complete
oxidation.
Figure 1. GPC curves for the DNPIM polymer series and PIM-1.
Two homopolymers and three series of copolymers were
synthesized by polymerization of combinations of TTSBI,
TFTPN, THDN, or TFTOT, using a modified procedure similar
to the synthesis of PIM-1. The homopolymers are referred to as
DNPIM-100 or TOTPIM-100, where PIM denotes polymer of
intrinsic microporosity, DN denotes dinaphthyl groups, and
TOT denotes tetraoxide thianthrene groups. The copolymers
are identified as DNPIM, TOTPIM, or DNTOTPIM with
suffixes, -66, -50, and -33, -25, -20, which refer to dinaphthyl
groups to TFTPN or tetraoxide thianthrene groups to TTSBI
ratio in the copolymers (see Table 1).
We previously reported in a communication a single CoPIM
(DNPIM-33) based on TTSBI, TFTPN, and THDN.22 The
reaction was carried out at 160 ꢀC for 120 min in contrast with
the synthesis of PIM-1 under the same conditions; copolycon-
densation led to high molecular weight and low polydispersity. It
is interesting to note that the same phenomena were observed in
other copolymerizations based on TTSBI, TFTPN, and TFTOT,
with or without THDN. It is shown in Table 1 and Figure 1 that
for all the copolymers a high Mn can be obtained, with a narrow
molecular weight distribution. There is no evidence of cross-
linking or the formation of cyclics. Thus, a conclusion can be
made that under the same reaction conditions (160 ꢀC, 120 min)
molecular weight distribution broadening (cyclics or cross-
linking) is effectively reduced by introducing a certain ratio of
other comonomers into the PIM chain, leading to high molecular
weight CoPIMs. PIM-1 prepared under similar conditions
(extended reaction time) results in a cross-linked product. One
conceivable explanation is that the comonomers disrupt chain
uniformity and change the ladder structures; a more unfolded
structure could result in less cross-linking and cyclics. Another
explanation is that they have different reactivity, and the como-
nomers with high reactivity in the polymerization system were
changed in concentration by introducing other comonomers with
low reactivity.
Figure 2. MALDI-TOF mass spectrum of DNPIM-100.
molecular weight material having Mn of 3000 Da.17 A higher
molecular weight of DNPIM-100 homopolymer (after purifica-
tion, Mn = 32 000 Da) was obtained under reaction conditions of
160 ꢀC and 2 h with polar solvents, but there were still numerous
low molecular weight cyclics and oligomers present. Although
Mn of DNPIM-100 would appear to be sufficiently high molec-
ular weight, strong free-standing films could not be obtained due
to the extremely rigid and contorted polymer chain. It is likely
that the steric hindrance over short segmental chain lengths of
dinaphthyl leads to difficulty in two such monomer segments
accessing a single TFTPN monomer and excessive spatial con-
tortion increases the possibility of cyclic formation from spatially
close reactive groups. As shown in Figure 1, DNPIM-100
(without purification) and DNPIM-66 (without purification)
have a significant amount of low molecular weight fractions.
From the MALDI-TOF mass spectrum (Figure 2), it can be seen
that DNPIM-100 contains a high level of cyclics and oligomers,
which have two -F and two -OH groups at the polymer chain
terminus, which is in good agreement with the explanation. In the
case of TOTPIM-100 homopolymer, cross-linking occurs easily
during the reaction. According to the different electronic activation
power of electron withdrawing groups (-SO2R > -CF3 >
-CN),25 the reactivity of -F activated by -SO2- is much higher
than by -CN. The polymerization to give TOTPIM-100 proceeded
rapidly in N-methylpyrrolidone (NMP), providing high molecular
weight polymer within 2 h, and the resulting polymer could be cast
into free-standing films from chloroform solution. However, the
same polymerization for TOTPIM-100 run in DMAc resulted in
cross-linked white insoluble product within a short time.
The homopolymers derived from THDN and TFTPN or from
TFTOT and TTSBI can be obtained under the same reaction
conditions as the CoPIMs. The first reported synthesis of
DNPIM-100 at 80 ꢀC and 72 h reaction time gave only a low

8584 Macromolecules, Vol. 43, No. 20, 2010
Du et al.
Table 2. Thermal Properties of the DNPIMs, TOTPIMs,
DNTOTPIMs Series and PIM-1
polymers
Td (ꢀC)b
Td (ꢀC)c
Td5 (ꢀC)d
RW (%)e
DNPIM-100a
DNPIM-50
DNPIM-33
485.1
476.7
466.1
457.3
472.3
469.1
362.6
333.0
343.9
338.9
298.9
429.6
510.5
507.2
504.3
499.1
471.6
470.2
474.5
479.8
490.6
486.9
481.9
492.6
516.5
514.9
509.5
502.7
483.2
481.3
485.8
487.2
494.6
493.1
490.5
495.4
85.0
75.3
72.1
70.3
70.4
71.1
74.7
76.1
76.6
77.4
77.5
68.2
DNPIM-25
TOTPIM-100
TOTPIM-50
TOTPIM-33
TOTPIM-25
DNTOTPIM-50
DNTOTPIM-33
DNTOTPIM-25
PIM-1 (40 min)
a After purification. b Actual onset temperature of decomposition.
c Extrapolated onset temperature of decomposition measured by TGA.
d 5% weight loss temperature measured by TGA. e Residue weight at
600 ꢀC under N2.
The 1H NMR spectra of CoPIMs can be explained better
by three-dimensional representation. In the 3-D image it is
observed that the four methyl groups (H-1) are not equiva-
lent. Two of them are very close to H-5, and the electron
cloud of the -CH3 group is shielding this proton;hence its
very low chemical shift (6.4 ppm) for an aromatic proton.
Another two methyl groups are more distant from the H-4,
and a higher chemical shift (6.8 ppm) is observed due to
absence of shielding. This explains the reason that the four
methyl groups (H-1) do not appear as a singlet but as two
singlets. The same principle also applies to H-2 and H-3. For H-4
and H-5 protons in PIM-1 and DNPIM-50, respectively, they
appear at the same position because their environment is almost
the same. However, in TOTPIM-50 and DNTOTPIM-50,
different nearby electron-withdrawing groups (-CN and
-SO2-) cause different chemical shifts for H-4 and H-5
protons. The integral ratio of the four signals is 1:1:1:1,
which indicates 50% H-4 and H-5 are affected by -SO2-
groups and are shifted to low field. This is additional proof
indicating these two CoPIMs were obtained as the expected
structures. Although they are randomly polymerized, the
ratio of tetraoxide thianthrene to TFTPN groups is 1:1 in the
resulting copolymers.
Finally, the 19F NMR spectra (not shown) were collected
for all PIMs homo- and copolymers. It is worth mentioning
that no aromatic F signal was detected in all these PIM-1
homo- and copolymers.
Thermal Analysis. Thermal analyses results for the three
new PIMs series and PIM-1 are compared in Table 2. All the
polymers are amorphous, remaining glassy up to their
decomposition temperatures, and have excellent thermal
stabilities. No glass transitions were detected at temperatures
up to 350 ꢀC. Actual onset temperatures of decomposition in
nitrogen for DNPIMs, TOTPIMs, and DNTOTPIM were in
the range of 457-485, 333-472, and 298-344 ꢀC respec-
tively. The thermal stabilities of copolymers were slightly
reduced by tetraoxide thianthrene segments but improved
significantly by introducing dinaphthyl segment on the main
chain.
Mechanical Properties. The mechanical properties of three
CoPIMs-50 series and PIM-1 are listed in Table 3. Tensile
stress at break and tensile strain at break decreased obviously
due to the introduction of dinaphthyl groups into the poly-
mer chain. Compared with PIM-1 (tensile strain at break:
11.2%; tensile stress at break: 47.1 MPa), the tensile strain at
break of DNPIM-50 is only 5.67%, while the tensile stress at
break is 38.6 MPa. This could be possibly attributed to the
polymer being more contorted and less amenable to chain
Figure 3. 1H NMR spectra of PIM-1, DNPIM-50, TOTPIM-50, and
DNTOTPIM-50.
NMR Analysis. All three CoPIMs-50 were fully character-
1
1
ized by H and 19F NMR spectroscopy. The H spectra of
CoPIMs-50 were quite similar to those of PIM-1 at high field
due to their identical aliphatic protons occurring in the
TTSBI monomer segment. The additional signals at low
field due to aromatic protons in the different comonomer
segments were readily assigned in the 1H NMR spectra
(Figure 3). Furthermore, the experimental ratio of intensity
values for aromatic protons H-6 to H-11 compared with
aliphatic protons H-2,3 was found to be exactly as expected.
For example, the experimental ratio of intensity values for
protons on the aromatic rings of dinaphthyl segments in
DNPIM-50 compared with -CH3 is exactly 10H:12H, as
expected for one repeat unit of dinaphthyl group in copolymers.

Article
Macromolecules, Vol. 43, No. 20, 2010 8585
Table 3. Physical Properties of the DNPIM-50, TOTPIM-100, TOTPIM-50, DNTOTPIM-50, and PIM-1a
tensile stress
tensile strain
at break (MPa) at break (%)
polymers
Ρ (g/cm3) Vsp (cm3/g) M (g/mol) Vw (cm3/mol) Vf (cm3/g) FFV SBET (m2 g-1
)
DNPIM-50
TOTPIM-100
TOTPIM-50
DNTOTPIM-50
PIM-1
1.11
1.28
1.19
1.30
1.06
0.90
0.78
0.84
0.77
0.94
449
630
545
533
460
233
302
274
263
246
0.227
0.157
0.186
0.129
0.244
0.25
0.20
0.22
0.17
0.26
709
560
601
408
729
38.6
48.3
45.8
38.2
47.1
5.67
24.1
20.9
4.75
11.2
a The disruption in chain packing, validated by FFV, was calculated using the following relationship: Vf = (Vsp - 1.3Vw); FFV = Vf/Vsp, where Vf is
the free volume and Vsp is the specific volume. Film samples had a density in the range 1.06-1.21 g cm-3, as determined by measurements of their weight
in air and in ethanol. Vw is the specific van der Waals volume calculated using the group contribution method of Bondi.
Table 4. Gas Permeabilities and Ideal Selectivities of the DNPIM,
TOTPIM, DNTOTPIM Series and PIM-1
P (barrersa)
Rb
polymers
O2
N2 H2 CO2 O2/N2 CO2/N2 H2/N2
DNPIM-50
DNPIM-33
DNPIM-25
TOTPIM-100
TOTPIM-50
TOTPIM-33
TOTPIM-25
DNTOTPIM-50 621 172 1703 3065 3.61
DNTOTPIM-33 773 228 2077 3693 3.39
DNTOTPIM-25 973 287 2695 4814 3.30
522 132 1478 2627 3.96
907 242 2347 4646 3.75
1139 321 3049 5799 3.55
642 190 1368 3056 3.40
1198 413 2616 4756 2.90
1247 436 2854 5320 2.86
1596 570 3567 6441 2.80
19.9
19.2
18
11.2
9.7
9.5
7.2
6.3
6.5
6.3
10.0
9.1
9.4
6.1
8.6
16.1
11.5
12.2
11.2
18.0
16.2
16.8
11.3
15.2
PIM-1
1717 629 3847 7108 2.73
PIM-1 (40 min) 1133 353 3042 5366 3.20
a Permeability coefficients measured at 25 ꢀC and 50 psig pressure. 1
barrer =10-10 [cm3 (STP) cm]/(cm2 s cmHg). b Ideal selectivity R =
(Pa)/(Pb).
packing due to the incorporation of bulky dinaphthyl seg-
ments. In contrast, mechanical properties were enhanced
significantly by the incorporation of tetraoxide thianthrene
segments. The tensile strain at break for TOTPIM-50s is
more than double that of PIM-1, which indicates that
polymer chain assumed a more open-chain conformation
that allowed greater chain packing after incorporating the
tetraoxide thianthrene segments.
Figure 4. Relationship between O2 permeability and O2/N2 selectivity
of TOTPIM-100, CoPIMs, and PIM-1. PIM-1 (r) [1, reported by Budd
et al. (2005) at 200 mbar (2.90 psia) feed pressure at 30 ꢀC;15 2, reported
by Budd et al. (2008) at 1 atm (14.7 psig) feed pressure at 23 ꢀC;15 3,
reported by Staiger et al. at 4 atm (58.8 psia) feed pressure 35 ꢀC;16 4,
reported by Du et al. at 4.4 atm (50 psig) feed pressure 25 ꢀC; 5, the
membrane was prepared by the same procedure as other CoPIMs
and tested at 4.4 atm (50 psig) feed pressure 25 ꢀC]. TOTPIM-100
and CoPIMs at 4.4 atm (50 psig) feed pressure 25 ꢀC (0) [6, DNPIM-
50; 7, DNPIM-33; 8, DNPIM-25]. (b) [9, DNTOTPIM-50; 10,
DNTOTPIM-33; 11, DNTOTPIM-25;]. (]) [12, TOTPIM-100, 13,
TOTPIM-50; 14, TOTPIM-33; 15, TOTPIM-25].
Transport Properties of Pure Gases. The porosities of
TOTPIM-100 and CoPIMs-50 (DNPIM-50, TOTPIM-50,
and DNTOTPIM-50) were probed by nitrogen sorption BET
analysis at 77 K. The polymers were precipitated from chloro-
form into methanol, followed by extensive washing with metha-
nol prior to sorption measurements. PIM-1 was tested under the
same conditions and used as a reference material. Nitrogen
sorption measurements on these polymers revealed that the
samples were microporous. As shown in Table 3, compared
with PIM-1, BET surface area of TOTPIM-100 and CoPIM-50
decreased from 729 m2 g-1 for PIM-1 to 408 m2 g-1 for
DNTOTPIM-50, which is consistent with the calculated frac-
tional free volume (FFV).26-28 Table 3 also shows DNPIM-50
chain packing is the least efficient among the three CoPIMs,
which FFV is close to PIM-1. A detailed study on the compar-
ison of structures for DNPIMs and PIM-1 was reported
elsewhere.22 Dinaphthyl segments provide a shorter distance
between contorted centers and a smaller twist angle, resulting in
a more compact polymer chain. The molecular conformation is
more zigzag compared with the offset-linear structure of PIM-1
containing the spirobisindane segment. These two factors com-
bine to give a high FFV in dinaphthyl-containing PIM. In con-
trast, tetraoxide thianthrene segments provide a longer distance
between contorted centers and an offset-linear structure, which
results in a more efficient packing and less FFV.
of lower selectivity, and vice versa. An upper bound perfor-
mance for this trade-off relationship was proposed by
Robeson.29,30 Pure-gas permeability coefficients (P) were
measured on dense polymer films of PIM-1, TOTPIM-100,
and three series of CoPIMs for O2, N2, H2, and CO2. A
summary of these P values and ideal selectivities for various
gas pairs is shown in Table 4. According to numerous
reported gas separation data for PIM-1 and our own labora-
tory testing experience,15,16 it is evident that the gas transport
properties of PIMs are quite sensitive to the testing condi-
tions, such as ambient surroundings, pretreatment, and
different film fabrication techniques. However, all the PIM
data are above the Robeson upper bound (1991) with the
expected “trade-off” behavior between permeability and
selectivity, as shown in Figure 4. TOTPIM-100 and three
series of CoPIMs exhibit higher selectivity, coupled with some
reduction in gas permeabilities, compared with PIM-1 (no. 5
tested under the same conditions). The overall permeability-
selectivity performance combines to exceed the Robeson upper
bound (1991) for O2/N2. For all three series of CoPIMs, gas
permeabilities are observed to increase with decreasing ratios
of dinaphthyl or/and tetraoxide thianthrene groups in the
polymer chain. For the same ratio of dinaphthyl or/and
In general, there is a trade-off relationship in glassy or
rubbery polymers between gas permeability and selectivity
for common gases. Higher permeability is gained at the cost

8586 Macromolecules, Vol. 43, No. 20, 2010
Du et al.
Figure 5. Models of the segments based on TTSBI, TFTPN, THDN, and TFTOT as calculated with energy minimization.
tetraoxide thianthrene groups, the gas permeabilities are
observed to be in the sequence of TOTPIM>DNTOTPIM>
DNPIM. From a materials viewpoint, a shorter interchain
distance combined with the more contorted zigzag polymer
chain structure imparts higher selectivity but lower perme-
ability, and vice versa. Dinaphthyl has a shorter distance
between contorted centers, and the chain has a more compact
zigzag structure; hence, selectivity increased. Compared with
PIM-1 and TOTPIMs, the permeability is decreased because
of a shorter segmental length and smaller twist angle. For
TOTPIMs, tetraoxide thianthrene units have a longer seg-
mental length between spiro centers and a more offset-linear
structure, which leads to a more efficient chain packing and
reduction of the permeability but increasing selectivity.
Molecular Modeling. Molecular modeling results may help
to explain why, as compared to PIM-1, the coeffects of the
three CoPIMs exhibit improved gas selectivity without over-
all loss of performance relative to the upper bound. Con-
formational analysis of four segments comprising two repeat
unit lengths derived from TTSBI, TFTPN, THDN, and
TFTOT was modeled to study the effect and distribution
of dinaphthyl and tetraoxide thianthrene groups on chain
geometry and steric interaction. The calculation results of
geometry optimization with energy minimization using the
AMBER method provides a visualization of major confor-
mational changes occurring in the polymers, as shown in
Figure 5. The chains containing tetraoxide thianthrene units
have a relatively linear and long ladder structure, which
presumably leads to more efficient chain packing. The chains
containing dinaphthyl units resulting in a zigzag and short
ladder segment would lead to less efficient chain packing and
more contortion. For example, compared with PIM-1 (the
fragment including of TTSBI, TFTPN), the TTSBI þ TFTOT
segment shows a similarly unperturbed coil conformation, but
the distance between the spiro center is longer, resulting in a
more open-chain structure and increased chain-packing. On the
other hand, the TFTPN þ THDN segment is observed to be a
more contorted structure when viewed from the “x”, “y”, and
“z” axes perspective, but the angle at the spatially twisted
dinaphthyl center (∼60ꢀ) is considerably smaller than that at
the spirobisindane (∼90ꢀ); in addition, the distance between the
twist center is shorter. This is in good agreement with their
mechanical properties and gas permeation results of DNPIM,
which exhibits the lowest permeability and highest selectivity.
Conclusions
Novel PIMs homopolymers and copolymers derived from
TTSBI, TFTPN, TFTOT, and THDN extend the spectrum of
possible soluble polymeric structures having high molecular
weight that are suitable as film-forming membrane materials.
Reaction conditions of elevated temperature and reaction times
of 120 min led to polymers having high molecular weight and low
polydispersity. Incorporation of high molar ratios of TFTOT
and/or THDN monomers resulted in polymers exhibiting analo-
gous gas transport properties to that of conventional PIM-1. The
structures of the new monomers, homopolymers, andcopolymers
were confirmed by 1H and 19F NMR spectroscopy. The intrinsic
microporosities of these polymers are supported by FFV, BET
data, and gas permeability measurements. Thermal properties of
the new homopolymers and copolymers indicate they have high
thermal stability, with DNPIMs containing the dinaphthyl seg-
ment showing the highest stability. TOTPIM-100 and the Co-
PIMs have an excellent combination of properties and were more
gas/N2 selective than PIM-1. The increased selectivity coupled
with high permeability combines to exceed the Robeson upper-
bound line for O2/N2. These results suggest that these polymers
are potential candidates for gas separation membrane materials
and have a good combination of physical properties, similar to
PIM-1. The incorporation of the dinaphthyl and tetraoxide
thianthrene monomers into a copolymer system allows the gas
permeability, selectivity, and other properties to be tuned.
Acknowledgment. This work was partly supported by the Cli-
mate Change Technology and Innovation Initiative, Greenhouse

Article
Macromolecules, Vol. 43, No. 20, 2010 8587
Gas project (CCTII, GHG), Natural Resources Canada (NRCan).
M.D.G. acknowledges partial support from the WCU program
through the National Research Foundation of Korea funded by
the Ministry of Education, Science and Technology (No. R31-
2008-000-10092-0). The authors are grateful toMr. FloydTollfor
the elemental analysis, BET, and mechanical properties testing.
The authors are also grateful to Dr. Li Jian Jun and Mr. Jacek
Stupak, National Research Council, Institute for Biological
Sciences, for the MALDI-TOF MS measurements.
(13) (a) Nagai, K.; Masuda, T.; Nakagawa, T.; Freeman, B. D.; Pinnau,
I. Prog. Polym. Sci. 2001, 26, 721–798. (b) Pinnau, I.; Toy, L. G.
J. Membr. Sci. 1996, 116, 199–209.
(14) (a) Thomas, S.; Pinnau, I.; Du, N.; Guiver, M. D. J. Membr. Sci.
2009, 333, 125–131. (b) Thomas, S.; Pinnau, I.; Du, N.; Guiver, M. D.
J. Membr. Sci. 2009, 338, 1–4.
(15) (a) Budd, P. M.; Msayib, K. J.; Tattershall, C. E.; Reynolds, K. J.;
McKeown, N. B.; Fritsch., D. J. Membr. Sci. 2005, 251, 263–269.
(b) Budd, P. M.; McKeown, N. B.; Ghanem, B. S.; Msayib, K. J.;
Fritsch, D.; Starannikova, L.; Belov, N.; Sanfirova, O.; Yampolskii, Y.;
Shantarovich., V. J. Membr. Sci. 2008, 325, 851–860.
(16) (a) Staiger, C. L.; Pas, S. J.; Hill, A. J.; Cornelius, C. J. Chem.
Mater. 2008, 20, 2606–2608. (b) Ghanem, B. S.; Hashem, M.; Harris,
K. D. M.; Msayib, K. J.; Xu, M.; Budd, P. M.; Chaukura, N.; Book, D.;
Tedds, S.; Walton, A.; McKeown, N. B. Macromolecules 2010, 43,
5287–5294. (c) McKeown, N. B.; Budd, P. M. Macromolecules 2010,
43, 5163–5176. (d) Ghanem, B. S.; McKeown, N. B.; Budd, P. M.;
Al-Harbi, N. M.; Fritsch, D.; Heinrich, K.; Starannikova, L.; Tokarev,
A.; Yampolskii, Y. Macromolecules 2009, 42, 7881–7888.
(17) McKeown, N. B.; Budd, P. M.; Msayib, K.; Ghanem, B. Micro-
porous Polymer. Material, WO 2005/012397 A2.
References and Notes
(1) Budd, P. M.; Ghanem, B. S.; Makhseed, S.; McKeown, N. B.;
Msayib, K. J.; Tattershall, C. E. Chem. Commun. 2004, 230–231.
(2) Budd, P. M.; Elabas, E. S.; Ghanem, B. S.; Makhseed, S.;
McKeown, N. B.; Msayib, K. J.; Tattershall, C. E.; Wong, D.
Adv. Mater. 2004, 16, 456–459.
(3) McKeown, N. B.; Budd, P. M.; Msayib, K. J.; Ghanem, B. S.;
Kingston, H. J.; Tattershall, C. E.; Makhseed, S.; Reynolds, K. J.;
Fritsch., D. Chem.;Eur. J. 2005, 11, 2610–2620.
(4) Du, N.; Robertson, G. P.; Song, J.; Pinnau, I.; Thomas, S.; Guiver,
M. D. Macromolecules 2008, 41, 9656–9662.
(18) Du, N.; Robertson, G. P.; Pinnau, I.; Guiver, M. D. Macromole-
cules 2009, 42, 6023–6030.
(5) Davankov, V. A.; Tsyurupa, M. P. React. Polym. 1990, 13, 27–42.
(6) Tsyurupa, M. P.; Davankov, V. A. React. Funct. Polym. 2002, 53,
193–203.
(7) Budd, P. M.; Ghanem, B.; Msayib, K.; McKeown, N. B.; Tattershall,
C. J. Mater. Chem. 2003, 13, 2721–2726.
(19) Aoki, T. Prog. Polym. Sci. 1999, 24, 951–993.
(20) George, S. C.; Thomas, S. Prog. Polym. Sci. 2001, 26, 985–1017.
(21) Du, N.; Robertson, G.; Song, J.; Pinnau, I.; Guiver, M. D.
Macromolecules 2009, 42, 6038–6043.
(22) Du, N.; Robertson, G. P.; Pinnau, I.; Thomas, S.; Guiver, M. D.
Macromol. Rapid Commun. 2009, 30, 584–588.
(23) Toda, F.; Tanaka, K.; Iwata, S. J. Org. Chem. 1989, 54, 3007–3009.
(24) Bock, V. H.; Stein, U.; Rittmeyer, P. Angew. Chem. 1982, 94, 540–541.
(25) March, J. Advanced Organic Chemistry; McGraw-Hill: New York,
1970; p 253.
(8) McKeown, N. B.; Gahnem, B.; Msayib, K. J.; Budd, P. M.;
Tattershall, C. E.; Mahmood, K.; Tan, S.; Book, D.; Langmi,
H. W.; Walton, A. Angew. Chem., Int. Ed. 2006, 45, 1804–1807.
(9) Webster, O. W.; Gentry, F. P.; Farlee, R. D.; Smart, B. E.
Makromol. Chem., Macromol. Symp. 1992, 54, 477–482.
(10) Urban, C.; McCord, E. F.; Webster, O. W.; Abrams, L.; Long, H. W.;
Gaede, H.; Tang, P.; Pines, A. Chem. Mater. 1995, 7, 1325–1332.
(11) Wood, C. D.; Tan, B.; Trewin, A.; Niu, H. J.; Bradshaw, D.;
Rosseinsky, M. J.; Khimyak, Y. Z.; Campbell, N. L.; Kirk, R.;
Stockel, E.; Cooper, A. I. Chem. Mater. 2007, 19, 2034–2048.
(12) (a) Dai, Y.; Guiver, M. D.; Robertson, G. P.; Kang, Y. S. Macro-
molecules 2005, 38, 9670–9678. (b) Dai, Y.; Guiver, M. D.; Robertson,
G. P.; Kang, Y. S.; Lee, K. J.; Jho, J. Y. Macromolecules 2004, 37,
1403–1410.
(26) Bondi, A. J. Phys. Chem. 1964, 68, 441–451.
(27) Van Krevelen, D. W. Properties of Polymers: Their Correlation with
Chemical Structure; Their Numerical Estimation and Prediction from
Additive Group Contributions; Elsevier: Amsterdam, The Netherlands,
1990.
(28) Chern, R. T.; Sheu, F. R.; Jia, L.; Stannett, V. T.; Hopfenberg,
H. B. J. Membr. Sci. 1987, 35, 103–115.
(29) Robeson, L. M. J. Membr. Sci. 1991, 62, 165–185.
(30) Robeson, L. M. J. Membr. Sci. 2008, 320, 390–400.