Effects of alkyl substitution on the physical properties and gas transport behavior in selected poly(R-phenoxyphosphazenes)

Article abstract of DOI:10.1016/j.polymer.2011.07.010

A systematic preparation of alkyl substituted phenoxyphosphazene polymers was performed and their gas transport properties determined. In this study, phosphazenes substituted with 4-methylphenol, 4-ethylphenol, and 4-isopropylphenol are reported. An additional polymer substituted with 4-tert-butylphenoxy-1-ethanol also was synthesized in this work. Data derived for these materials, including chemical, thermal and gas transport characterization, were compared to previous reports discussing poly[bis-phenoxyphosphazene] and its analog with tert-butyl substitution: poly[bis-(4-tert-butylphenoxy)phosphazene]. The tert-butyl moiety influences orderly chain packing, presumably through steric hindrance that can influence aromatic π-stacking. For the new poly[(alkylphenoxy)phosphazenes], semicrystallinity is maintained and the added steric bulk serves to decrease the polymer glass transition temperature (Tg) and increase both permeability and selectivity for the gas pairs: O₂/N₂ and CO ₂/CH₄. Removal of the tert-butyl moiety from the immediate vicinity of the backbone through a flexible spacer serves to depress the Tg as compared to poly[bis-(4-tert-butylphenoxy)phosphazene], but provides no performance enhancement for gas transport.

Full text of DOI:10.1016/j.polymer.2011.07.010

Polymer 52 (2011) 3879e3886
Contents lists available at ScienceDirect
Polymer
journal homepage: www.elsevier.com/locate/polymer
Effects of alkyl substitution on the physical properties and gas transport behavior
in selected poly(R-phenoxyphosphazenes)
*
Mark D. Ogden, Christopher J. Orme, Frederick F. Stewart
Idaho National Laboratory, P.O. Box 1625, Idaho Falls, ID 83415-2208, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A systematic preparation of alkyl substituted phenoxyphosphazene polymers was performed and their
gas transport properties determined. In this study, phosphazenes substituted with 4-methylphenol,
4-ethylphenol, and 4-isopropylphenol are reported. An additional polymer substituted with 4-tert-
butylphenoxy-1-ethanol also was synthesized in this work. Data derived for these materials, including
chemical, thermal and gas transport characterization, were compared to previous reports discussing poly
[bis-phenoxyphosphazene] and its analog with tert-butyl substitution: poly[bis-(4-tert-butylphenoxy)
phosphazene]. The tert-butyl moiety influences orderly chain packing, presumably through steric
Received 24 May 2011
Received in revised form
8 July 2011
Accepted 11 July 2011
Available online 20 July 2011
Keywords:
Polyphosphazene
Membranes
hindrance that can influence aromatic
p-stacking. For the new poly[(alkylphenoxy)phosphazenes],
semicrystallinity is maintained and the added steric bulk serves to decrease the polymer glass transition
temperature (Tg) and increase both permeability and selectivity for the gas pairs: O2/N2 and CO2/CH4.
Removal of the tert-butyl moiety from the immediate vicinity of the backbone through a flexible spacer
serves to depress the Tg as compared to poly[bis-(4-tert-butylphenoxy)phosphazene], but provides no
performance enhancement for gas transport.
Gas separations
Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction
formation of phosphorus-oxygen bonds (PeOH, P]O), which
themselves result in backbone cleavage through hydrolysis of PeN
Poly(organophosphazenes) are unique materials that are
composed of an alternating phosphorus and nitrogen backbone.
Tethered to the phosphorus are pendant groups that serve to
stabilize the polymer to hydrolysis and provide variation in the
polymers’ chemical and physical properties [1]. Phosphazenes are of
interest due to their high degree of chemical stability. Many phos-
phazenes can be considered high performance polymers in that they
are stable at temperatures as high as 300 ꢀC [2]. The synthesis of
poly(organophosphazenes) differs from that of other polymers in
that the backbone is formed initially, followed by a substitution
process to yield the final material from which membranes can be
formed [reference needed]. In our laboratories, the polymer back-
bone is formed by ring-opening polymerization of commercially
available hexachlorocyclotriphosphazene,1, to yield linear poly[bis-
chlorophosphazene], 2, Fig. 1. Poly(bis-chlorophosphazene) is an
unstable material when exposed to ambient moisture due to the
lability of chlorine. Exposure to moisture leads to nucleophilic attack
by water at the phosphorus resulting in evolved HCl and the
bonds. The key to stabilization of the polymer is removal of chlorine
and replacement with nucleophiles, such as alcohols [3], amines [4],
or organometallic reagents [5], as represented by structure 3.
Numerous substitution chemistries have been reported [6].
The unique chemistry of polyphosphazenes allows for
a systematic study of structure using standard analytical tech-
niques, and through applications, such as gas transport. Through
controlled pendant group substitution, a strategic approach to
polymer composition can be made. An example of this concept is
the balance between hydrophilic and hydrophobic groups and the
resultant effect on gas permeation [7]. Hydrophilicity was provided
by 2-(2-methoxyethoxy)ethanol (MEE) and the hydrophobic
component was 4-methoxyphenol. Increased permeability for CO2
was observed with increased MEE substitution, which was attrib-
uted to an affinity between MEE and the permeant gas. A refine-
ment of this study showed that the relationship was more likely
due to a correlation with the polymer glass transition temperature
(Tg), where lower Tg polymers, regardless of chemical functionality,
yielded higher CO2 permeability [8]. Another example examined
polyphosphazenes formed with phenol and 4-phenylphenol
pendant groups at various loadings of each [9]. In this series, the
range of Tg values obtained was not as wide; however, the same
* Corresponding author. Tel.: þ1 208 526 8594; fax þ1 208 526 8541.
E-mail address: Frederick.Stewart@INL.gov (F.F. Stewart).
0032-3861/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.polymer.2011.07.010

3880
M.D. Ogden et al. / Polymer 52 (2011) 3879e3886
2.2. Synthesis of poly[bis-chlorophosphazene] (2)
Cl
Cl
P
Cl
Cl
Synthesis of poly(bis-chlorophosphazene), 2, which was
synthesized by ring-opening melt polymerization at 250 ꢀC using
a modified method of Singler [11]. Into a thick-walled glass tube
was added hexachlorocyclotriphosphazene, 1 (60 g, 172 mmol). The
tube was sealed under vacuum and heated in an oven at 250 ꢀC
until flow was nearly absent. Time varies depending on the source
of 1, which can vary between 12 and 60 h. Crude polymer was
separated by dissolution of the resulting opaque white rubber into
75 ml of toluene followed by precipitation into 800 ml of hexane.
The desired rubber was obtained by decanting off unreacted 1.
Polymer 2 was not isolated. To minimize decomposition or
unwanted cross-linking of this material, it was immediately dis-
solved into dry toluene and used within a day.
250 oC
P
Cl
Cl
Cl
Cl
N
P
P
Sealed
Tube
n
(2)
(1)
G
G
G = RO-, RN-, or R-
P
(3)
N
n
Fig. 1. Synthesis of stable poly(organophosphazenes).
2.3. Poly(bis- phenoxy)phosphazene (4) and poly[bis-(4-tert-
butylphenoxy)phosphazene] (5)
general relationship between increased gas permeation with
decreasing Tg was observed.
The synthesis and characterization of polymers 4 and 5 have
been reported elsewhere [10].
In this paper, a series of polymers were formed with potential
constraint of backbone motions, through the systematic attachment
of bulky groups to phenoxy-substituted polyphosphazenes, and the
effect of this substitution on gas permeability was explored. Dis-
cussed are phosphazenes formed with 4-methylphenol, 4-
ethylphenol, and 4-isopropylphenol pendant groups. Selection of
these polymers builds upon previous work characterizing the gas
permeability of poly[bis-phenoxyphosphazene], 4 and poly[bis-(4-
tert-butylphenoxy)phosphazene], 5 [10]. The new alkylphenox-
yphosphazene polymers in this work were selected to span the
range defined by 4 and 5 in terms of steric bulk at the paraposition of
thearomatic ring. Anadditional polymer was synthesizedasa partof
this study that displaced the substituted alkylphenoxy moiety
farther from the polymer backbone through an ethyleneoxy spacer,
11, Fig. 3. All polymers will be discussed in terms of how the alkyl
substitution at the 4-position (para) affects properties, such as Tg
and thegas permeability for selected gases (H2, CO2, O2, N2, and CH4.)
2.4. Synthesis of the poly[(4-alkylphenoxy)phosphazenes] (6e8)
Polymers 6e8 were synthesized using a similar method. The
synthesis of polymer 6 is shown here. In a three neck 2 L round
bottom flask was added
a thermometer, mechanical stirrer,
nitrogen purge, a heating mantle, and a condenser. To this was
added 34.67 g (321 mmol) of 4-methylphenol followed by
w300 ml of anhydrous THF. To this mixture was added sodium
hydride (12.84 g, 321 mmol) in small increments under a nitrogen
purge. Once all of the hydride was added, the reaction was allowed
to stir for 1 h to ensure complete reaction. Poly[bis-
chlorophosphazene] (9.31 g, 80 mmol), as a solution in 75 ml of
anhydrous toluene, was added to the reaction mixture. To this
mixture was added w200 ml of anhydrous 2-methoxyethylether
and the temperature was increased to 105 ꢀC. To attain this
temperature, a DeaneStark adapter was added to the experimental
apparatus to remove lower boiling THF until the temperature
reached 105 ꢀC. This reaction was monitored by 31P NMR until no
further changes in the spectra were noted, approximately 4 days.
The reaction was cooled to room temperature. Isolation of the
polymer was accomplished by successive precipitations from THF
solution (300e500 ml). Initially, the mother liquor was poured into
a mixture of ethanol/water (2250 ml/250 ml) and the polymer was
collected. The second precipitation was into deionized water (3 L)
and the third was into hexane (1 L). The final polymer mass was
collected and dried in a vacuum oven at 65 ꢀC overnight. Polymer 6:
yield 86%. Polymer 7: yield 68%. Polymer 8: yield 43%.
2. Experimental section
2.1. Methods and materials
Hexachlorocyclotriphosphazene was purchased from Aldrich
Chemical Company and Esprit Chemical Company and sublimed in
vacuo at 85 ꢀC prior to use. Anhydrous solvents (toluene, tetrahy-
drofuran (THF), and 2-methoxyethylether), 4-methylphenol, 4-
ethylphenol, 4-isopropylphenol, ethylene glycol, p-toluene
sulfonyl chloride, silica gel (200e400 mesh), and sodium hydride
(as a 60% dispersion in mineral oil), also were acquired from Aldrich
Chemical Co. and used as received. Methanol, ethyl acetate, and
hexane were purchased from VWR and used without purification.
Polymer 4 was acquired from Elf-Atochem and used as received.
Nuclear magnetic resonance (NMR) spectra were collected using
a Bruker DMX 300 WB operating at 7.04 T field strength: 300 MHz
(1H) and 121 MHz (31P). 31P chemical shifts were referenced to an
external H3PO4 standard. Thermal characterization data were
acquired using a TA Instruments model Q200 differential scanning
calorimeter (DSC) and a model Q500 thermogravimetric analyzer
(TGA). Polymer densities were obtained by helium pycnometry
using a Micromeritics Accupyc Model 1330 pycnometer. Molecular
weight determinations were performed using Gel Permeation
Chromatography against polystyrene standards by Polyhedron
Laboratories, Inc. (Houston, TX).
2.5. Ethylene glycol monotosylate, 9
A dry 2 L 3-necked round bottom flask was charged with
ethylene glycol (853 g, 13.75 mol), triethylamine (75.7 g,
550 mmol) and 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU), (1 ml,
6.7 mmol). The resulting mixture was stirred at 0 ꢀC in a water ice
bath. The flask was then equipped with an addition funnel charged
with p-toluenesulfonyl chloride (53.8 g, 275 mmol) in methylene
chloride (400 ml), which was added to the reaction flask slowly
over 2 h. The final reaction mixture was allowed to come to room
temperature and was stirred overnight. The reaction mixture was
then transferred to a 2-L separatory funnel and was washed with
two 200 ml portions of saturated NaHCO3, two 200 ml portions of
saturated citric acid, and two 200 ml portions of saturated NaCl.
The remaining organic was stripped of methylene chloride to yield

M.D. Ogden et al. / Polymer 52 (2011) 3879e3886
3881
a pale yellow oil, 34.7 g. The oil was characterized by NMR spec-
troscopy to be a mixture of the desired product, ethylene glycol
monotosylate, 91%, with remainder consisting of ethylene glycol di-
tosylate. No separation of this mixture was performed.
phosphazenes (>100 kD) result in no significant intrusion of the
polymer into the support. Characterization of thickness was
determined using two methods. First, thickness was calculated by
measurement of the mass of polymer applied to the support and
the polymer density. Second, measurement of thickness was also
performed using a micrometer, which upon subtraction of the
support thickness, yielded values consistent with the calculation
2.6. 2-(4-Tert-butylphenoxy)-1-ethanol, 10
A dry 1 L 3-neck round bottom flask was charged with tert-
butylphenol (26.0 g, 173 mmol), anhydrous toluene (100 ml) and
anhydrous 1,4-dioxane (600 ml). To this was added a magnetic
stirbar and a dry N2 sparge gas. Sodium hydride (3.98 g, 173 mmol)
was added slowly and the resulting mixture was stirred until no
evidence of hydrogen evolution was observed. In a second flask, an
oven dried 2 L 3-neck, a solution was formed from 9 in toluene
(50 ml) and anhydrous 1,4-dioxane (500 ml). The contents of the
first flask were then added slowly over 45 min to the second. The
reaction was then heated to reflux for 3 h upon which a large
amount of white precipitate formed. The reaction was cooled and
filtered. The filtrate was washed with two 200 ml portions of water
and then the organic layer was dried with anhydrous MgSO4.
Stripping of the solvent by rotary evaporation yielded 27.8 g of
crude product. NMR analysis showed the desired 2-(4-tert-butyl-
phenoxy)-1-ethanol, 10, along with ethylene glycol di-tert-
butylphenoxide, and a small amount of p-toluenesulfonic acid.
The acid was removed by dissolution of the crude product in
hexane, followed by filtration. The filtrate was then stripped of
method. Thicknesses ranged from 20 to 80 mm. The strong corre-
lation between these two methods suggested that intrusion of the
polymer into the pores of the support was minimal.
2.9. Pure gas permeability analysis
Permeabilities were determined using a literature barometric
method [12,13] where the permeate volume was 1021.5 ml, the
membrane area was 13.2 cm2, and the initial feed gas pressure was
30 psi. Gases studied included H2, CO2, O2, N2, and CH4.
3. Results and discussion
3.1. Synthesis of alkoxyphenoxy phosphazenes
Synthesis of polymers 6e8 was performed by exposure of
previously prepared poly[bis-chlorophosphazene], 2, with nucleo-
philes, themselves formed from the reaction of commercially
available phenols with sodium hydride. The reactions were moni-
tored regularly using 31P NMR spectroscopy performed on small
aliquots removed from solution. In general, at the beginning of the
reaction, polymer 2 gives a strong narrow singlet at approx-
imately ꢁ19 ppm. Upon the addition of aromatic nucleophiles,
multiple peaks are seen to form approximately ꢁ15 ppm. As the
reaction progresses, a peak corresponding to the desired homoge-
neously substituted structures grows in at approximately ꢁ17 ppm.
Near completion, only one peak is observed and the width narrows
as the remaining chlorines are displaced. Once complete, the
product polymer is isolated and purified through a series of
precipitations into selected solvents. Removal of entrained solvent
yields solid materials whose physical characteristics are largely
dictated by the pendant group. Poly[bis-phenoxyphosphazene], 4,
is a fibrous solid. Use of 4-methylphenol as a pendant group results
in little gross change to the polymer, as does the use of 4-
ethylphenol. Polymer 8 with 4-isopropylphenol substitution does
yield a significant change. This polymer was dense rubbery solid.
Interestingly, polymer 5, poly[bis(4-tert-butylphenoxy)phospha-
zene] tends to be somewhat fibrous.
solvent and chromatographed on
a silica gel column using
a mixture of 70% hexane/30% ethyl acetate as the eluent. Collection
of the desired fraction from the column yielded 10 as an oil (7.8 g,
15% yield from p-toluenesulfonyl chloride).
2.7. Poly(bis-(2-(4-tert-butylphenoxy)-1-ethoxy))phosphazene, 11
To a dry 1 L 3-necked flask, equipped with a mechanical stirrer,
thermometer, and nitrogen purge, was added 10 (15.3 g,
78.9 mmol) and anhydrous THF (300 ml). Sodium hydride (2.87 g,
74.9 mmol) was added over 10 min and the resulting solution was
stirred for 1 h at room temperature. At this time, a solution of
polymer 2 in anhydrous toluene (70 ml) was added to the reaction
mixture and the resulting solution was stirred at room temperature
for 20 h, at which time it was determined to be complete by P-31
NMR spectroscopy. Isolation of the product polymer was performed
by pouring the mother liquor into 2.5 L of 80% isopropanol in water.
The polymer floated on top of the solution and was mechanically
collected. This material was then dissolved into THF (200 ml) and
precipitated into water (1.5 L). The collected swollen rubber was re-
dissolved into THF (200 ml) and centrifuged to remove any undis-
solved matter. The clarified solutions were then poured into
methanol (1 L) and the collected precipitate was dried in a vacuum
oven at 60 ꢀC. Drying yielded 6.5 g of an off-white rubber in 76%
yield.
As a probe of the dependence of the location of the tert-butyl
group with respect to the polymer backbone, polymer 11, Fig. 3,
was prepared. The pendant group for this polymer is not available
commercially and was prepared in our laboratories, Scheme 1.
Initially, one hydroxyl moiety on ethylene glycol was converted to
the tosylate using a large excess of ethylene glycol to minimize
attachment of tosylate at both hydroxyl positions. Selectivity of up
to 92% monotosylate addition has been achieved. Attachment of 4-
2.8. Membrane formation
tert-butylphenol was achieved through
a Williamson ether
Membranes were formed using a solution casting method from
THF at polymer concentrations ranging from 2 to 8% by weight.
After dissolution, the solutions were clarified by centrifugation to
remove insoluble material, if necessary. In a fume hood, stainless
steel supports with a 0.5 pore diameter were placed on a level
surface and the polymer solutions were cast directly using a Pasteur
pipette. The membranes were covered with a beaker to slow the
evaporation rate for the best results in forming defect free
membrane films. After several hours, additional membrane drying
was performed in an oven at 60 ꢀC. Previous work has shown that
membranes formed in this method using high molecular weight
synthesis by reaction of the phenol with sodium hydride, followed
by exposure to the ethylene glycol monotosylate. Please note that
the di-substituted tosylate was not removed before 4-tert-butyl-
phenol attachment, so allowances were made in the addition of the
phenoxide to account for the equally reactive and undesired di-
substituted adduct.
Purification of the desired compound, 10, was performed using
silica gel column chromatography. Attachment of 10 to polymer 2
proceeded similarly to the previously discussed polymers 6e8,
yielding polymer 11. The net effect of the chemistries was to insert
an ethyleneoxy (eCH2eCH2eOe) linkage between the polymer

3882
M.D. Ogden et al. / Polymer 52 (2011) 3879e3886
O
O
S
O
CH2CH2 OH 92 %
S
Cl
O
(9)
O
HO CH2CH2 OH
50 fold molar excess
O
S
O
8 %
O
CH2CH2
O
S
O
O
O
CH2CH2 OH
92 %
O- Na+
(10)
O
CH2CH2 O
8 %
Chromatographic Separation
70/30 Ethylacetate/Hexane
O
CH2CH2 OH (10)
15 % yield with respect to p-toluenesulfonyl chloride
Scheme 1. Synthesis of 2-(4-tert-butylphenoxy)-1-ethanol.
backbone and the 4-tert-butylphenol. The effect on gross proper-
ties was substantial, where the product was a dense rubber.
this work, a Tg of 48 ꢀC was measured for polymer 5. Most
important is the observation of a T1 transition associated with
crystalline rearrangement at 118 ꢀC, somewhat lower than recor-
ded by Bowmer [17] (140 ꢀC). The higher Tg of 5 also suggests the
substituent has a hindering influence on backbone chain flexibility.
Methyl substituted polymer 6 has nearly the same Tg as 4
at ꢁ0.6 ꢀC; however, the T1 transitions are significantly higher in
temperature. It can be concluded that a methyl group is not large
enough to significantly influence the backbone mobility, nor does it
significantly disrupt crystallinity, which is not surprising consid-
ering the data for 5. The slightly larger ethyl substituent on polymer
7 does influence the Tg as shown in a decrease to ꢁ18 ꢀC, consistent
with Bowmer [17], which is due to the flexible nature of an ethyl
group and its ability to self-plasticize the polymer structure. Crys-
talline transitions are also seen with this polymer, but the
temperatures at which they are observed are lower than polymer 4.
Isopropyl substitution, polymer 8, does not result in a further loss in
Tg, where it was measured at ꢁ19 ꢀC and the semicrystalline nature
remained; however, the transitions were significantly lower in
temperature than what was observed for polymer 7.
Placement of the ethyleneoxy spacer moves the tert-butyl group
farther from the polymer backbone, thus lessening its influence on
backbone motion. The Tg for polymer 11 (Fig. 3) was 23 ꢀC which
suggests that the backbone motional freedom is influenced by the
highly flexible ethyleneoxy structures, which counteracts the
hindering influences of the tert-butyl groups. A complete absence
of semicrystallinity is observed for this polymer.
TGA of the polymers using air as the purge gas suggests little
effect on the overall polymer stability with respect to temperature.
The four polymers with aromatic moieties directly attached to P
had decomposition temperature (Td) values similar to each other.
The outlier was polymer 11, with a Td substantially lower, which is
3.2. Thermal characterization and pycnometry
Differential Scanning Calorimetry (DSC) and Thermogravimetric
Analysis (TGA) were employed to probe the morphology and
stability of the polymers. Data is shown in Table 1. A previous report
from our laboratory discussed polyphosphazenes substituted with
phenol, 4 and 4-tert-butylphenol, 5 [10]. The difference between
these polymers was the substitution at the para position, proton
versus a tert-butyl group, Fig. 2. Poly[bis-phenoxyphosphazene], 4,
is a semicrystalline polymer [14,15]. Cast directly from solvent
yields a membrane that is approximately 45% crystalline; although
the degree of crystallinity is heavily influenced by thermal history
[16]. Crystalline domains may be attributable to aromatic
pep
interactions that serve to create order in the system. Attachment of
the tert-butyl groups at the para position can somewhat, but not
completely, disrupt the
pep interactions, therefore changing the
morphology of the polymer. The Tg values for polymers 4 and 5 are
available in the literature at ꢁ1 ꢀC and 43 ꢀC, respectively [10]. In
Table 1
Thermal analysis and pycnometry data.
Polymer
Tg (ꢀC)
T1 (ꢀC)
Td (ꢀC)
Density (g/cm3)
Ref
4
5
6
7
8
11
ꢁ1
48
ꢁ0.6
ꢁ18
ꢁ19
23
99
118
134
54
47
None
126
351
386
388
334
398
327
1.332
1.160
1.223
1.202
1.140
1.126
[10]
This work
This work
This work
This work
This work
156
101
69

M.D. Ogden et al. / Polymer 52 (2011) 3879e3886
3883
(4)
(5)
O
O
O
O
P
P
N
N
n
n
(6)
(7)
O
O
O
O
P
P
N
N
n
n
(8)
O
O
P
N
n
Fig. 2. Structures for (poly)organophosphazenes 4e8.
attributable to the lower thermal stability seen in polymers with
ethereal linkages attached to P. This conclusion is consistent with
that this polymer has the highest organic fraction as a function of
overall polymer structure.
poly[(bis-(2-(2-methoxyethoxy)ethoxy))phosphazene],
formed
with 100% MEE substitution, which decomposes at 280 ꢀC [18].
Pycnometry was used to measure the densities of the various
polymers. In our previous report, we were able to reflect upon the
observation that the attachment of tert-butyl groups resulted in
lower polymer density [10]. This seemed reasonable due to the
steric influences of these groups. In this study, the incremental
attachment of alkyl groups serves to incrementally lower the
polymer densities giving the following order: 4 > 6 > 7 > 8 z 5.
Please note that the data derived from isopropyl and tert-butyl
attachments result in little difference. The density for 11 was the
lowest of the set in this paper, which may be a reflection of the fact
3.3. Dilute solution characterization
Inanattempttofurtherunderstandthegrossphysicalpropertiesof
the polymers, dilute solution Gel Permeation Chromatographic anal-
ysis was performed. The lack of control in the polymerization of
hexachlorocyclotriphosphazene, 1, has been discussed extensively
[19]. The rate and degree of polymerization of 1 are susceptible to
multiple influences including inorganic impurities, water, oxygen, as
well as the presence of Lewis acids such as the boron contained in the
borosilicate glass polymerization tubes. Table 2 shows the data for the
sixpolymersinthisstudy. Directcomparisonofthedatamustconsider
the use of differing supplies of hexachlorocyclotriphosphazene,1;poly
[bis-chlorophosphazene], 2; and analysis methods. Data for polymer 4
was provided by the supplier and 5 comes from a previous report [10].
Polymers 6e8 and 11 discussed in this work were studied by an
outside vendor.
Table 2
Gel permeation chromatography characterization data.
Polymer
Mw (Daltons)
Polydispersity Index (PDI)
Ref
4
5
6
7
8
11
6.00 ꢂ 105
1.15 ꢂ 106
3.68 ꢂ 105
3.77 ꢂ 105
3.73 ꢂ 105
7.79 ꢂ 105
NA
[10]
[10]
This work
This work
This work
This work
1.14
38.8
16.6
10.0
22.4
Fig. 3. Structure of polymer 11.

3884
M.D. Ogden et al. / Polymer 52 (2011) 3879e3886
Table 4
Other differences include the fact that polymer 11 and polymers
Pure gas permeability data. Permeability given in Barrers (1 Bar-
6e8 were formed using differing batches of polymer 2, and these
batches were formed from hexachlorocyclotriphosphazene
supplied by differing vendors. The importance of these molecular
weight and PDI differences on characteristics such as thermal
transitions and gas transport is a fact that should be examined.
Polymer 4 offers the best opportunity to study these effects due to
the wealth of information available in the literature [10,16,20,21]. A
comparison of data reveals little deviation in permanent gas
permeability and Tg, although it could be argued that molecular
weight would critically influence gross mechanical behaviors, such
as stress/strain and tensile strength.
rer ¼ 1.33 ꢂ104 m2(STP) m mꢁ2 sꢁ1 kPaꢁ1).
Polymer
Permeability (Barrers)
H2 CO2
7.5 4.8
Ref
O2
2.1
8.2
3.4
N2
1.3
2.4
1.1
CH4
4
5
6
1.2
1.7
2.8
[10]
[10]
This work
This work
This work
This work
23.0
9.3
17.0
10.2
7
8
11
21.3
98.2
16.2
32.3
135.1
13.8
15.0
41.6
3.5
4.9
14.2
0.9
8.9
30.3
1.7
gases i and j. Further, this ideal selectivity can be described as
a function of the solubility selectivity and the diffusivity selectivity,
Eq. (2).
3.4. NMR characterization
NMR characterization of phosphazenes typically employs the
commonly observed 31P, 1H, and 13C nuclei. Observed atoms are
more sensitive to changes in their immediate vicinities than they
are with more remote structures. For example, single 31P peaks
were observed for all four new polymers, Table 3. Polymers 6, 7, and
8 all have roughly similar chemical shifts (approximately ꢁ17
to ꢁ8 ppm), which is not unreasonable. In fact, this compares well
to polymers 4 and 5. Modifications in 6e8 were performed rela-
tively far from the polymer backbone, thus the influence of the alkyl
substitution on the aromatic rings is minimal as seen at phos-
phorus. Polymer 11 also supports this assertion. Insertion of the
ethyleneoxy spacer group at phosphorus directly influences the
phosphorus chemical shift. A singlet is seen at ꢁ6.2 ppm, which is
consistent with other phosphazenes that have ethyleneoxy
substitution at phosphorus [22]. Proton and 13C NMR spectroscopy
were used to confirm the presence and structure of the organic
components of the polymers, Table 3.
À
Á
À
Á
À
Á
a
¼
Pi=Pj
¼
Di=Dj ꢂ Si=Sj
(2)
ði;jÞ
Hydrogen is a small gas whose transport is largely diffusive. A
previous report has discussed H2 permeation through 4 and 5 [10].
Upon attachment of the tert-butyl group, the H2 permeability
modestly, but significantly, increased from 7.5 Barrers for 4 to
23 Barrers for 5. If we assume that 4 and 5 represent extremes and
that 6e8 are incremental changes between these extremes, then
their H2 permeabilities would be expected to be between 7.5 and
23 Barrers. Polymer 6 permeability measured 9.3 Barrers, consis-
tent with the expectation. However, polymers 7 and 8 did not
follow the expected trend. H2 permeability for ethyl substituted 7
was significantly larger than 6, and almost as large as 5, at
21.3 Barrers. Isopropyl substituted polymer 8 gave an even more
unexpected H2 permeability of 98.2 Barrers.
Previous reports have demonstrated the relationship between
polymer Tg and gas permeability [8]. In general, lower Tg values
result in higher gas permeabilities. This relationship has been
demonstrated with CO2; however, it remains applicable, although
to a lesser degree, to the other gases in this study. The isopropyl
group performs an internal plasticization role, resulting in a far
lower Tg than expected. The lower Tg, which correlates to higher
polymer segmental chain motions, may also result in higher gas
permeability. However, this rather simple correlation is not
universal as shown by the comparison of 5 and 11. In 11, the
sterically bulky tert-butyl group is more remote from the backbone,
thus exhibiting less hindrance on chain motions as reflected by
a lower Tg than 5. However, the lower Tg does not result in higher
H2 permeability. It is significantly less suggesting other influences
such as gas solubility may play a role.
Highly condensable gases, such as CO2, exhibit more solubility
driven permeability in many membrane separations [26]. Polymer
4 shows the lowest CO2 permeability at 4.8 Barrers. The tert-butyl
group raises the permeability to 17.0 Barrers. Once again, the
methyl group substituted polymer 6 was found inside the limits
established between 4 and 5 at 10.2 Barrers. However, 7 and 8 were
substantially higher at 32.3 and 135.1 Barrers, respectively. This
data supports both diffusivity and solubility contributions to
transport. Using a diffusive argument, the Tg values for both 7 and 8
are much lower than 6, which explains significantly higher
permeability values with respect to the unsubstituted polymer 4.
However, the significant permeability difference between 7 and 8
cannot be explained by a Tg argument alone due to their similarity.
However, the difference in CO2 permeabilities may be a function of
solubility influences induced by the additional alkyl content of 8.
Ideal gas selectivity is a relative measure of the membrane’s
ability to perform an actual gas separation. In the determination of
an ideal separation factor, the nature of each gas and the polymer
membrane dictate the behavior and provides a probe of structure
3.5. Pure gas permeability and ideal selectivity
Phosphazenes have been studied as membranes for gas sepa-
rations [23,24]. To characterize this set of phosphazenes, pure gas
permeability measurements were obtained from flat sheet
membranes at 30 ꢀC. Gases used in this study included H2, CO2, O2,
N2, and CH4 and a data summary for these gases is shown in Table 4.
This selection of gases represents an envelope of gas properties,
which dictate the relative contributions of diffusion and solubility
to the mechanism of gas transport. The envelope spans from small
highly penetrating gases (H2) to larger more condensable gases
(CO2).
Permeability of any gas i through a membrane has been
described as a function of the gases’ solubility (S) in the membrane
and it ability to diffuse (D) through the membrane, Eq. (1) [25].
Pi ¼ Di ꢂ Si
An equally valuable term is the ideal gas selectivity,
(1)
a
, which is
defined as a ratio of the pure gas permeability measurements for
Table 3
NMR characterization data.
Polymer 31P
(ppm)
1H (ppm)
13C (ppm)
6
ꢁ18.1 6.74 (d), 6.58 (d), 2.10 (s)
150.1, 132.7, 129.6, 121.5,
21.1
7
ꢁ17.5 6.79 (d), 6.60 (d), 2.38 (dd), 1.05 150.2, 139.1, 128.4, 121.7,
(dd) 28.6, 16.0
ꢁ17.4 6.79 (d), 6.63 (d), 2.61 (m), 1.03 150.3, 143.7, 126.9, 121.8,
(d) 33.8, 24.5
ꢁ6.2 6.99 (d), 6.56 (d), 4.34 (brs), 3.89 156.9, 143.3, 126.5, 114.5,
(brs), 1.14 (s) 67.7, 65.2, 34.4, 32.0
8
11

M.D. Ogden et al. / Polymer 52 (2011) 3879e3886
3885
10
and processes that increased the upper bound to higher levels for
many gas pairs [28].
Application of the upper bound analysis to the O2/N2 gas pair for
all polymers is shown in Fig. 4. In general, the performance of these
polymers is typical of rubbery polymers; however, this type of
analysis does provide insight into the effects of the synthetic
modifications. Poly[bis-phenoxyphosphazene], 4, shows the lowest
O2 permeability and selectivity over N2. Taken together, polymers
6e8 show a steady increase in permeability while exhibiting an
insignificant loss in selectivity. Thus, increasing the alkyl content on
the polymers serves to increase both permeability and selectivity
for the O2/N2 gas pair when compared to the 4. However, the trend
shown for 6e8 does not apply to polymer 5. Higher selectivity is
noted for 5 at the expense of O2 permeability. Polymer 11 exhibits
a trade-off in performance with a loss of permeability and an
increase in selectivity with respect to 5.
Upper Bound
(11)
(6)
(5)
(7)
(8)
(4)
CO2/CH4 is an industrially relevant gas separation and thus
worthy of examination. The highest selectivity polymers were 5
and 11 and both offered increased performance in terms of both
higher CO2 permeability and ideal separation factor as compared to
4, Fig. 5. The series of 6e8 displayed little significant change in
separation factor; however, the CO2 permeability was sequentially
increased by the modification of the phenoxy pendant group.
1
0.1
1
10
100
1000
104
105
O2 Permeability (Barrers)
Fig. 4. Permeability/selectivity plot for O2/N2.
necessary to create membrane materials with both higher perme-
ability and selectivity. Robeson reported in 1991 an empirical
correlation between the permeability and ideal selectivity of
a membrane towards selected gas pairs [27]. The result of this work
is an easily employed method to assess membrane performance. An
inverse correlation, or trade-off, in membrane performance is
observed for many membranes. This observation reflected the
common occurrence in which an increase in the permeability of
a selected gas component often results in a loss in selectivity.
Graphically, this observation was shown as an empirical upper
bound on membrane performance. The Robeson upper bound has
become a useful measure to compare membrane performance and
to judge the results of either material or process modifications to
membrane systems. In 2008, Robeson updated the empirical
correlation to reflect the abundance of new membrane materials
4. Conclusion
In this report, a systematic synthetic modification to phenoxy-
substituted polyphosphazenes is described along with the effect
that this substitution has on physical and chemical properties as
well as gas permeation. The attachment of alkyl groups, up to but
not including tert-butyl, depresses the glass transition temperature
(Tg); which results in an increase in the permeability for a selection
of gases. The O2/N2 and CO2/CH4 gas pairs show that increases in
permeability, at the expense of selectivity, can be attained through
a slight increase in the steric bulk of the alkyl groups. Once the
groups become large, as represented by the tert-butyl moiety, some
permeability for both O2 and CO2 is lost; however, the membranes
become more selective. Furthermore, movement of the tert-butyl
group farther from the backbone using a flexible ethyleneoxy
spacer results in a completely amorphous structure, which is
a reflection of the reduced influence of the aromatic moiety and its
tert-butyl substituent on the backbone. This synthetic modification
fails to provide any significant increase in gas transport
performance.
1000
Acknowledgements
100
This manuscript has been authored by Battelle Energy Alliance,
LLC under Contract No. DE-AC07-05ID14517 with the U.S. Depart-
ment of Energy. The United States Government retains and the
publisher, by accepting the article for publication, acknowledges
that the United States Government retains a nonexclusive, paid-up,
irrevocable, worldwide license to publish or reproduce the pub-
lished form of this manuscript, or allow others to do so, for United
States Government purposes.
(5)
10
(11)
(8)
(4)
(6) (7)
References
[1] Allcock HR, Kugel RL. J Am Chem Soc 1965;87(18):4216.
[2] Mark JE, Allcock HR, West R. Inorganic polymers. Englewood Cliffs: Prentice-
Hall; 1992.
1
0.01
0.1
1
10
100
1000
104
105
106
[3] Allcock HR, Kugel RL, Valan KJ. Inorg Chem 1966;5(10):1709e15.
[4] Allcock HR, Kugel RL. Inorg Chem 1966;5(10):1716e8.
[5] Wisian-Neilson P, Zhang CP, Koch KA, Gruneich JA. Phosphorus Sulfur Silicon
Relat Elem 1999;146:69e72.
CO2 Permeability (Barrers)
Fig. 5. Permeability/selectivity plot for CO2/CH4.

3886
M.D. Ogden et al. / Polymer 52 (2011) 3879e3886
[6] Gleria M, De Jaeger R. Polyphosphazenes: a review. In: Majoral J-P, editor.
New aspects in phosphorus chemistry V, vol. 250. Berlin, Germany: Springer-
Verlag; 2005. p. 165e251.
[7] Orme CJ, Harrup MK, Luther TA, Lash RP, Houston KS, Weinkauf DH, et al.
J Membr Sci 2001;186(2):249e56.
[18] Stewart FF, Harrup MK, Lash RP, Tsang MN. Polym Int 2000;49(1):57e62.
[19] Stewart FF, Peterson ES. Sulfur-nitrogen-phosphorus-containing polymers. In:
Dubois P, Coulembier O, Raquez JM, editors. Handbook of ring-opening
polymerization. Weinheim: Wiley-VCH Verlag GmbH & Co; 2009. p. 97e122.
[20] McCaffrey RR, Cummings DG. Sep Sci Technol 1988;23(12e13):1627e43.
[21] Allcock HR, Nelson CJ, Coggio WD, Manners I, Koros WJ, Walker DRB, et al.
Macromolecules 1993;26(7):1493e502.
[8] Orme CJ, Klaehn JR, Harrup MK, Luther TA, Peterson ES, Stewart FF. J Membr
Sci 2006;280(1e2):175e84.
[9] Muldoon JG, Pintauro PN, Wysick RJ, Lin J, Orme CJ, Stewart FF. J Membr Sci
2009;334(1e2):74e82.
[22] Blonsky PM, Shriver DF, Austin P, Allcock HR. J Am Chem Soc 1984;106(22):
6854e5.
[10] Orme CJ, Klaehn JR, Stewart FF. J Membr Sci 2004;238(1e2):47e55.
[11] Singler RE, Hagnauer GL, Schneider NS, Laliberti BR, Sacher RE, Matton RW.
J Polym Sci Pol Chem 1974;12(2):433e44.
[12] Amerongen GJV. J Appl Phys 1946;17(11):972e85.
[13] Barrer RM. Trans Faraday Soc 1939;35:628e43.
[14] Kojima M, Magill JH. Polymer 1989;30(10):1856e60.
[15] Kojima M, Satake H, Masuko T, Magill JH. J Mater Sci Lett 1987;6(7):775e8.
[16] Golemme G, Drioli E, Lufrano F. Vysokomol Soedin 1994;36(11):1945e52.
[17] Bowmer TN, Haddon RC, Chichester-Hicks S, Gomez MA, Marco C, Fatou JG.
Macromolecules 1991;24(17):4827.
[23] Gleria M, De Jaeger R. Applicative aspects of poly(organophosphazenes). New
York: Nova Science Publishers; 2004.
[24] Golemme G, Drioli EJ. Inorg Organomet Polym 1996;6(4):341e65.
[25] Wijmans JG, Baker RW. J Membr Sci 1995;107(1e2):1e21.
[26] Matteucci S, Yampolskii Y, Freeman BD, Pinnau I. Transport of gases and
vapors in glassy and rubbery polymers. In: Yampolskii Y, Pinnau I,
Freeman BD, editors. Materials science of membranes for gas and vapor
separation. West Sussex, England: John Wiley and Sons; 2006. p. 1e48.
[27] Robeson LM. J Membr Sci 1991;62(2):165e85.
[28] Robeson LM. J Membr Sci 2008;320(1e2):390e400.