Substituent exchange reactions of trimeric and tetrameric aryloxycyclophosphazenes with sodium 2,2,2-trifluoroethoxide

Article abstract of DOI:10.1039/c1dt11606a

Substituent exchange reactions of sodium 2,2,2-trifluoroethoxide with trimeric and tetrameric aryloxycyclophosphazenes with phenoxy, 4-formylphenoxy, 4-cyanophenoxy and 4-nitrophenoxy side groups were conducted at 66°C in THF and monitored by ³¹P NMR and mass spectrometry. These are model reactions for their counterparts with high polymeric linear organophosphazenes. The ease of displacement of OAr in cyclic trimeric and tetrameric molecules by CF₃CH₂O increased significantly with the presence of electron-withdrawing substituents in the polyphosphazene in the order, phenoxy ? 4-formylphenoxy < 4-cyanophenoxy ≈ 4-nitrophenoxy. Fully substituted 2,2,2-trifluoroethoxyphosphazene trimer and tetramer were formed by side group exchange, but these reactions were followed by an attack by the nucleophile on the α-carbon of the 2,2,2-trifluoroethoxy groups linked to phosphorus to give a species in which one trifluoroethoxy group had been replaced by an ONa unit, and bis(trifluoroethyl) ether was formed as a side product. On the other hand, only partly exchanged species were formed when sodium phenoxide reacted with the trifluoroethoxy phosphazene trimer and tetramer, but again a product with an ONa side group was formed eventually together with phenyltrifluoroethyl ether generated via alpha-carbon attack. The relative sensitivity of 2,2,2-trifluoroethoxy and phenoxyphosphazene cyclic trimers and tetramers to the presence of trifluoroethoxide was established.

Full text of DOI:10.1039/c1dt11606a

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Dalton
Transactions
Cite this: Dalton Trans., 2012, 41, 2100
www.rsc.org/dalton
PAPER
Substituent exchange reactions of trimeric and tetrameric
aryloxycyclophosphazenes with sodium 2,2,2-trifluoroethoxide
Xiao Liu, Jonathan P. Breon, Chen Chen and Harry R. Allcock
Received 25th August 2011, Accepted 10th October 2011
DOI: 10.1039/c1dt11606a
Substituent exchange reactions of sodium 2,2,2-trifluoroethoxide with trimeric and tetrameric
aryloxycyclophosphazenes with phenoxy, 4-formylphenoxy, 4-cyanophenoxy and 4-nitrophenoxy side
groups were conducted at 66 ◦C in THF and monitored by 31P NMR and mass spectrometry. These are
model reactions for their counterparts with high polymeric linear organophosphazenes. The ease of
displacement of OAr in cyclic trimeric and tetrameric molecules by CF3CH2O increased significantly
with the presence of electron-withdrawing substituents in the polyphosphazene in the order, phenoxy ꢀ
4-formylphenoxy < 4-cyanophenoxy ª 4-nitrophenoxy. Fully substituted 2,2,2-trifluoro-
ethoxyphosphazene trimer and tetramer were formed by side group exchange, but these reactions were
followed by an attack by the nucleophile on the a-carbon of the 2,2,2-trifluoroethoxy groups linked to
phosphorus to give a species in which one trifluoroethoxy group had been replaced by an ONa unit,
and bis(trifluoroethyl) ether was formed as a side product. On the other hand, only partly exchanged
species were formed when sodium phenoxide reacted with the trifluoroethoxy phosphazene trimer and
tetramer, but again a product with an ONa side group was formed eventually together with
phenyltrifluoroethyl ether generated via alpha-carbon attack. The relative sensitivity of
2,2,2-trifluoroethoxy and phenoxyphosphazene cyclic trimers and tetramers to the presence of
trifluoroethoxide was established.
in the polymer by another using a second organic nucle-
Introduction
ophile (Scheme 1). This second approach is an appealing al-
ternative, especially for the high polymers since it offers the
prospect that single-substitutent poly(organophosphazene)s can
be converted readily to mixed-substituent materials which are
of broad technological interest. It also raises the possibility
that a poly(organophosphazene) that is stable for long peri-
ods of time in the atmosphere might be employed as a gen-
eral macromolecular intermediate for the preparation of other
poly(organophosphazene)s. The chloro-derivative intermediate is
sensitive to moisture and must be stored under carefully controlled
inert conditions. Also, the substitution reactions of the chloro
intermediate can only be conducted in a limited number of organic
solvents, such as tetrahydrofuran or dioxane.
In addition to its utility for the preparation of new polymers,
a systematic study of organic side group exchange reactions can
also provide useful information as a practical guide to optimize
the classical synthesis of cosubstituted poly(organophosphazenes)
prepared by the sequential or simultaneous reactions of two or
more nucleophiles with (NPCl2)n.18–23 Earlier preliminary work
showed that fluorinated alkoxy units24 and phenoxy side groups
can be displaced from polyphosphazenes.15 For example, for
polyphosphazenes that bear both trifluoroethoxy and phenoxy
side groups, replacement of phenoxy side groups by 2,2,2-
trifluoroethoxide occurs at non-geminally substituted phosphorus
atoms. The reverse reaction, exchange of the trifluoroethoxy group
Polyphosphazenes are hybrid macromolecules with an essentially
linear backbone of alternating phosphorus and nitrogen atoms
and with two organic side groups linked to each phosphorus.1
These polymers have been widely studied for various applications
such as bone regeneration scaffolds,2–4 fire retardants,5,6 low-
temperature elastomers,7 fuel cell membranes8–10 and solid or
gel polymer lithium ion conductors11,12 A distinctive feature of
polyphosphazenes is the ease with which the polymer properties
can be tuned through changes in the side groups.
The most widely explored method for the synthesis of
poly(organophosphazenes) is based on the thermal ring-opening
polymerization of hexachlorocyclotriphosphazene, (NPCl2)3, to
high polymeric poly(dichlorophosphazene), (NPCl2)n, followed by
the replacement of the labile chlorine atoms in this macromolec-
ular intermediate by organic groups, such as alkoxy, aryloxy, or
amino units13,14 as shown in Scheme 1. In addition, cosubstituted
poly(organophosphazenes) can be synthesized by sequential or
simultaneous addition of different nucleophiles to conveniently
obtain various materials with tunable properties.
However, as mentioned in earlier publicatons,15–17 an alterna-
tive synthesis approach is to replace one organic substituent
Department of Chemistry, The Pennsylvania State University, University
Park, Pennsylvania, 16802. E-mail: hra@chem.psu.edu; Fax: 814 865 3314;
Tel: 814 865 3527
2100 | Dalton Trans., 2012, 41, 2100–2109
This journal is
The Royal Society of Chemistry 2012
©

View Article Online
Scheme 1 Synthesis of mixed-substituent poly(organophosphazenes) by Sequential or Simultaneous Addition of Nucleophiles to (NPCl2)n or by Side
Group Exchange Reactions.
by phenoxide ions, was not detected.15 Therefore, the order of the
sequential addition of different side groups must be considered in
order to obtain the desired materials with targeted compositions.
The work presented here is an attempt to examine these pro-
cesses in more detail. Because the reactions of small molecule cyclic
phosphazenes often mimic the behavior of the high polymers, and
are easier to analyze, this study is focused on the behavior of cyclic
trimeric and tetrameric species. The comparable processes for high
polymers will be addressed in a later study.
7.16, 7.23 (5H, m, OC6H5) m/z = 694 ([M+H]+) m/z was calculated
for C36H30N3O6P3.
Hexakis(4-formylphenoxy)cyclotriphosphazene (3). The synthe-
sis procedure was reported previously.26 Yield: 91.2%. 31P NMR
1
(CDCl3), ppm: d +7.67 (3P, s). H NMR (CDCl3), ppm: d 7.13,
7.72 (4H, d, OC6H4CHO), 9.92 (1H, s, OC6H4CHO) m/z = 862.4
([M+H]+) m/z was calculated for C42H30N3O12P3.
Hexakis(4-cyanophenoxy)cyclotriphosphazene (4). Prepared as
described previously.27 Yield: 64.2%. 31P NMR (d-DMSO), ppm:
1
d +7.41 (3P, s). H NMR (d-DMSO), ppm: d 7.17, 7.83 (4H, d,
OC6H4CN).
Experimental
Hexakis(4-nitrophenoxy)cyclotriphosphazene (5). Synthesized
via a process described earlier.28 Yield: 82.7%. 31P NMR (d-
DMSO), ppm: d +7.30 (3P, s). 1H NMR (d-DMSO), ppm: d
7.31,8.16 (4H, d, OC6H4NO2).
Materials and equipment
All reactions were carried out under an atmosphere of dry
argon using standard Schlenk line techniques. Tetrahydrofuran
(EMD) was dried using solvent purification columns.25 2,2,2-
Trifluoroethanol (Aldrich) was purified by vacuum distillation
from CaH2 (Aldrich). Phenol (Aldrich) was purified by sub-
limation. 4-Nitrophenol was recrystallized twice from toluene.
Hexachlorocyclotriphosphazene, (NPCl2)3, (from various sources
including Fushimi Pharmaceutical Co., Japan, and Ningbo Chem-
ical, China) was purified by recrystallization from hexane and by
vacuum sublimation at 50 ◦C. Octachlorocyclotetraphosphazene,
obtained from the sublimation residues after the trimer sublima-
tion, was recrystallized twice from toluene and was then vacuum
sublimed. Other reagents were used as received. 31P and 1H NMR
spectra were obtained with a Bruker 360 WM instrument operated
at 145 and 360 MHz, respectively. Spectrometric analysis data were
collected using the turbospray ionization technique with use of an
Applied Biosystems API 150EX LC/MS mass spectrometer.
Synthesis of cyclic tetrameric compounds
Octakis(2,2,2-trifluoroethoxy)cyclotetraphosphazene (6). Prepared
as described previously.14 Yield: 36.1%. 31P NMR (CDCl3), ppm:
d -1.44 (4P, s). 1H NMR (CDCl3), ppm: d 4.27 (2H, t, OCH2CF3).
m/z = 973 ([M+H]+) m/z was calculated for C16H16F24N4O8P4.
Octaphenoxycyclotetraphosphazene (7).29 Yield: 36.1%. 31P
1
NMR (CDCl3), ppm: d -11.79 (4P, s). H NMR (CDCl3), ppm:
d 6.96, 7.10, 7.17 (5H, m, OC6H5). m/z = 925 ([M+H]+) m/z was
calculated for C48H40N4O8P4.
Octakis(4-formylphenoxy)cyclotetraphosphazene (8).30 Yield:
1
57.1%. 31P NMR (d-DMSO), ppm: d -13.96 (4P, s). H NMR
(d-DMSO), ppm: d 7.10, 7.71 (4H, d, OC6H4CHO), 9.92 (1H,
s, OC6H4CHO). m/z = 1149.3 ([M+H]+) m/z was calculated for
C56H40N4O16P4.
Octakis(4-cyanophenoxy)cyclotetraphosphazene (9). A mixture
of (NPC12)4 (1.5 g, 3.2 mmol), 4-cyanophenol (3.6 g, 29.1 mmol)
and Cs2CO3 (9.5 g, 29.1 mmol) in THF (150 ml), was stirred at
reflux for 1 day. The reaction mixture was filtered, and the filtrate
was concentrated and precipitated into water. After filtration, the
crude 4b was reprecipitated twice from DMSO into water and
twice from DMSO into THF. The product was dried under vacuum
to yield 2.96 g of a white solid. Yield: 82.7%. 31P NMR (d-DMSO),
ppm: d -14.26 (4P, s). 1H NMR (d-DMSO), ppm: d 7.14, 7.83 (4H,
d, OC6H4CN).
Synthesis of cyclic trimeric compounds
Hexakis(2,2,2-trifluoroethoxy)cyclotriphosphazene (1). This com-
pound was prepared from (NPCl2)3 and sodium trifluoroethoxide
according to a procedure described previously.1 Yield: 75.8%. 31
P
NMR (CDCl3), ppm: d +17.45 (3P, s). 1H NMR (CDCl3), ppm: d
4.28 (2H, m, OCH2CF3) m/z = 730 ([M+H]+) m/z was calculated
for C12H12F18N3O6P3.
Hexaphenoxycyclotriphosphazene (2). This compound was pre-
pared via a procedure reported earlier.1 Yield: 81.4%. 31P NMR
Octakis(4-nitrophenoxy)cyclotetraphosphazene (10), was pre-
1
(CDCl3), ppm: d +9.35 (3P, s). H NMR (CDCl3), ppm: d 6.98,
pared via an earlier procedure.31 Yield: 70.5%. 31P NMR
This journal is
The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 2100–2109 | 2101
©

View Article Online
1
(d-DMSO), ppm: d -14.58 (4P, s). H NMR (d-DMSO), ppm:
Reactions of hexaphenoxycyclophosphazene (2) with sodium
d 8.13, 7.26 (4H, d, OC6H4NO2)
trifluoroethoxide
Purification of N3P3(ONa)(OCH2CF3)5(11). The crude prod-
ucts from the ligand exchange reaction of hexakis(2,2,2-
trifluoroethoxy)cyclophosphazene with sodium trifluoroethoxide
in THF were dried by rotary evaporation and were redissolved
in dichloromethane. After extraction with ammonium chloride
(10 wt%) (¥3) and deionized water (¥3), the organic layer was
dried over MgSO4 and filtered, and the solvent was removed by
rotary evaporation. A solution of the crude products in a 2 : 1
mixture of dichloromethane and ethyl acetate was passed through
a silica gel column prepared using the same solvent. The product
was dried under vacuum to give 11, a yellow liquid. 31P NMR
Sodium trifluoroethoxide can induce substituent exchange
reactions with various aryloxycyclotriphosphazenes, such
as hexakis(p-nitrophenoxy)cyclotriphosphazene, hexakis(o-
nitrophenoxy) cyclotriphosphazene, hexakis(p-chlorophenoxy)-
cyclotriphosphazene and hexaphenoxycyclotriphosphazene
according to an earlier study.32 It was also reported that, when
sodium trifluoroethoxide reacts with 2, significant amounts
of 1 and a large amount of N3P3(OC6H5)3(OCH2CF3)3 were
formed. Very little N3P3(OC6H5)2(OCH2CF3)4 and almost no
N3P3(OC6H5)(OCH2CF3)5 were obtained. However, in this earlier
work, the reaction was studied over a relatively short period
of time (3 days), and the progress of the entire reaction and
identification of the ultimate products were not monitored
or examined in detail. Such information is crucial for fully
understanding the processes involved in exchange reactions. This
is the major issue addressed in this paper.
1
(CDCl3), ppm: d +18.99, 19.58 (3P, d), 11.71 (2P, t). H NMR
(CDCl3), ppm: d 4.28 (2H, m, OCH2CF3)
Substituent exchange reactions for both cyclic trimer and tetramer
derivatives
In this investigation, when hexaphenoxycyclotriphosphazene
2
reacted with sodium trifluoroethoxide in a ratio of
All the substituent exchange reactions were carried out in a similar
manner. The following is a typical procedure. A solution of 1
(1 g, 1.37 mmol) in THF (10 ml) was added dropwise to a stirred
solution of sodium phenoxide (1.91 g, 16.5 mmol) in THF (90 ml).
The mixture was stirred at reflux in THF. Typically, reactions were
allowed to proceed for up to 50 days. At timed intervals, starting
after one day, samples were taken and the reaction progress was
monitored by 31P NMR and mass spectrometry. The presence of
the etheric side products was established by mass spectrometric
analysis of the reaction mixtures.
1 : 12 at reflux temperature (66 ◦C) in THF, hexakis(2,2,2-
trifluoroethoxy)cyclophosphazene (1) is generated by a substituent
exchange reaction. This product was detected by 31P NMR
spectrometry after two days of reaction (Fig. 2b) together with
a second phosphazene product (11) (doublet-triplet, 19.9 ppm
and 11.1 ppm). This second product was also formed in other
substituent exchange reactions and will be discussed later. The
amounts of both products 1 and 11 increased as the reaction
progressed and as the starting material, 2, was consumed. After
50 days reaction, no further change of the ratio in the components
in the reaction was detected, which indicated that the interaction
had reached a steady state. At this stage the percentage of each
phosphazene component in the reaction mixture was 2 16.9%, 1
10.8% and 11 72.3% (Fig. 2d).
Results and discussion
The compounds to be discussed are shown in Fig. 1. In the
following sections the substituent exchange reactions of 1 and
2 with the nucleophiles, sodium phenoxide and sodium triflu-
oroethoxide will be discussed first, together with the reactant
ratio effects in this system, followed by similar reactions at the
cyclic tetramer level (6 and 7). In the following section, the use of
sodium trifluoroethoxide as the exchange reagent for substituted
aryloxycyclophosphazenes will then be described both for the
cyclic trimeric (3–5) and tetrameric systems (8–10). Finally, the
stability of 1, 2, 6 and 7 in the presence of excess nucleophile in
phosphazene syntheses will be discussed. A summary of the results
is given in Table 1.
The reaction rate and ratios of the products were significantly
dependent on the concentration of the nucleophile. The rate of
consumption of 2 under reaction conditions b (Table 1) was much
faster than under conditions a as the amount of trifluoroethoxide
was doubled. Thus, after 50 days reaction, the starting material
2 was almost completely consumed (reaction conditions b and
Fig. 3d). Furthermore, the higher concentration of trifluoroethox-
ide caused a greater tendency for 1 to be transformed to side
product 11 (Fig. 4b).
The absence of partly exchanged products from reactions
carried out under these conditions suggests that the reaction
Fig. 1 Structures of cyclic trimeric and tetrameric phosphazenes.
2102 | Dalton Trans., 2012, 41, 2100–2109 This journal is
The Royal Society of Chemistry 2012
©

View Article Online
Table 1 Reaction conditions and major products
Reaction
a
Initial phosphazene
Nucleophile
Molar ratio of phosphazene : nucleophile
1 : 12
Major products
2
1, 11b
b
c
2
1
1 : 24
1 : 12
1, 11
11, 12
d
1
1 : 24
11, 12
e
f
3
4
5
7
7
6
1 : 12
1 : 12
1 : 12
1 : 12
1 : 24
1 : 12
1, 11
1, 11
g
h
i
1, 11
6, 13, 14
6, 13, 14
13, 14, 15
j
k
6
1 : 24
13, 14, 15
l
8
1 : 12
1 : 12
1 : 12
1 : 12
1 : 12
1 : 12
6, 13, 14
6, 13, 14
6, 13, 14
11
m
n
o
p
q
9
10
1
6
13
2
—
r
7
1 : 12
—
a1 = N3P3(OCH2CF3)6; 2 = N3P3(OC6H5)6; 3 = N3P3(OC6H4CHO-p)6; 4 = N3P3(OC6H4CN-p)6; 5 = N3P3(OC6H4NO2-p)6; 6 = N4P4(OCH2CF3)8;
7 = N4P4(OC6H5)8; 8 = N4P4(OC6H4CHO-p)8; 9 = N3P3(OC6H4CN-p)6; 10 = N3P3(OC6H4NO2-p)6; b11 = N3P3(OCH2CF3)5(O-Na+); 12 =
N3P3(OCH2CF3)5(OC6H5); 13 = N4P4(OCH2CF3)7(O-Na+); 14 = N4P4(OCH2CF3)6(O-Na+)2; 15 = N4P4(OCH2CF3)8-x(OC6H5)x (x = 1 ~ 4).
proceeds as a self-accelerating process due to the strong electron-
withdrawing effect of trifluoroethoxy side groups. This electron-
withdrawal renders the cyclophosphazene more susceptible to
further attack by the nucleophile.
This product is formed from 1, itself generated by side group
exchange from 2. Evidence in favor of this interpretation is
as follows. (a) No partially exchanged phenoxy-trifluoroethoxy
cyclic phosphazenes were detected by mass spectrometry for these
reaction conditions: only the fully exchanged cyclophosphazene 1
was present. (b) There is a parallel increase in the formation of 1
and 11 as the reaction proceeds. (c) No products similar to 11 but
containing phenoxy groups were detected in any of these reactions.
The mechanism shown in Scheme 2 supposes that trifluo-
roethoxide ion participates in a nucleophilic attack on the alpha-
carbon atom of a side group in 1. Two factors favor the view
that the main attacking species is trifluoroethoxide rather than
Side product 11
Side product 11 was shown by 31P NMR and mass spectrometry
to have the structure shown in Scheme 2. The mass spectral
identification was accomplished using a negative model mass
spectrometer under a neutral mobile phase (acetonitrile) to give a
mass of 645.9, which matches the structure shown in Scheme 2.
This journal is
The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 2100–2109 | 2103
©

View Article Online
Fig. 2 31P NMR spectra for the reaction between 2 and sodium
trifluoroethoxide (molar ratio 1 : 12) a) 2 h; b) 2 days; c) 16 days; d) 50
days.
Fig. 4 Reactions of hexaphenoxycyclotriphosphazene (2) with sodium
trifluoroethoxide: molar ratio a) 1 : 12; b) 1 : 24.
Fig. 3 31P NMR spectra for reaction between 2 and sodium trifluo-
roethoixde (molar ratio 1 : 24) a) 2 h; b) 1 day; c) 30 days; d) 50 days.
phenoxide. First, in a separate reaction between 1 and sodium
trifluoroethoxide, the etheric side product CF3CH2OCH2CF3
was identified by mass spectrometry (mass = 163.01). Second,
in the reaction between 2 and sodium trifluoroethoxide, the
concentration of trifluoroethoxide is always higher than phenoxide
even at the end of the side group exchange process.
Reactions of hexakis(2,2,2-trifluoroethoxy)phosphazene (1) with
sodium phenoxide
The reverse reaction is less facile than the one discussed above. No
hexaphenoxycyclophosphazene (2) was detected in the substituent
exchange reactions between 1 and sodium phenoxide at a ratio 1
to 12 (reaction conditions c) or 1 to 24 (conditions d). Only the
mono-exchanged derivative, N3P3(OCH2CF3)5OPh (12), and side
product 11 (Fig. 5) were detected. Both reactions reached a steady
state after 10 and 25 days respectively. (Fig. 6) From the mass
spectra, the final products from reaction conditions c and d were
compounds 11 and 12 (Scheme 3), which suggests that the two
products were formed through different pathways.
Fig. 5 31P NMR spectra for substituent exchange reaction between 1 and
sodium phenoxide for 2 days: molar ratio a) 1 : 12; b) 1 : 24.
nucleophilicity which make it difficult to achieve full replacement
of trifluoroethoxy groups.
Thus, the smaller and more nucleophilic trifluoroethoxide ion
can replace larger phenoxy groups to lower the overall steric
restrictions around phosphorus, but the reverse process is more
restricted. Therefore, although this exchange reaction is reversible,
it relies significantly on the preferred production of 1. Meanwhile,
1 is still liable to be attacked by a nucleophile (either phenoxide
or trifluoroethoxide) at the a-carbon of the alkoxy side unit to
generate side product 11. The mass spectrometric identification of
The reason why no fully exchanged product, hexaphenoxycy-
clophosphazene, (2), was detected is probably due to the higher
steric hindrance characteristic of phenoxide units and the lower
2104 | Dalton Trans., 2012, 41, 2100–2109
This journal is
The Royal Society of Chemistry 2012
©

View Article Online
Scheme 2 Two step pathway involved in the formation of compound 11.
Scheme
3
Two separate processes leading to the formation of 11 and 12 during the reaction of sodium phenoxide with hex-
akis(2,2,2-trifluoroethoxy)cyclotriphosphazene (1).
trifluoroethylphenyl ether (mass = 176.02) as a reaction product
from the interaction of 1 with sodium phenoxide is further
evidence for the process shown.
On the other hand, the final products from reaction conditions
c and d also depended on concentration as with conditions a and
b. In the initial stage, condition d induced a much higher reactivity
than did condition c. For example, 23.3% of the starting material
(1) was consumed under conditions c after two days (Fig. 6a).
By contrast, 58.3% of 1 had reacted with sodium phenoxide
under conditions d (Fig. 5b). Furthermore, when the interactions
reached their steady state, only small amounts of compounds 11
and 12 were generated under conditions c. Most of the starting
material 1 remained intact (Fig. 6a). However, when twice the
amount of sodium phenoxide was used (condition d), a significant
This journal is
The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 2100–2109 | 2105
©

View Article Online
follows the same tendency as the results obtained for conditions a
and b.
Reactions at the cyclic tetrameric level:
octaphenoxycyclophosphzene (7) with sodium trifluoroethoxide
As in the cyclic trimer reactions, in the tetrameric sys-
tem, octakis(2,2,2-trifluoroethoxy)cyclophosphazene (6), was pro-
duced through substituent exchange reactions between octaphe-
noxycyclophosphazene (7) and sodium trifluoroethoxide, but at a
much faster rate. The side products 13 and 14 were detected by
31P NMR and confirmed by mass spectra. Again, reactions took
place through the pathways shown in Scheme 4. However, the
faster reaction rate at the cyclic tetramer level was illustrated by
the following observation: in the cyclic trimer system, the starting
material 2 could still be detected by 31P NMR spectrometry after
one month (reaction conditions a and b), but in the cyclic tetramer
system, most of the starting material 7 had been converted to 6,
together with compounds 13 and 14 within 3 days with only 7.7%
and 1.8% of starting material 7 remaining (conditions h and i and
Fig. 7). These effects may reflect the greater flexibility of the larger
ring and the resultant greater exposure of the phosphorus atoms
to nucleophilic attack, due to the more open structure and lower
steric hindrance compared to the cyclic trimer.1 It also suggests
that this effect could be even more evident at the high polymer
level. In addition, recent calculations show that the formal positive
charge on phosphorus is greater in the tetramer vs. the trimer
which is another likely source of the greater rate of reaction in the
tetramer. Actually, the trimer/tetramer results presented here are
consistent with the effects found for other related reactions. Thus,
the rate of exchange of chloride ion in (NPCl2)4, is faster than
in (NPCl2)3.33 Hydrolysis of (NPCl2)4 or (NPF2)4, proceeds faster
than the hydrolysis of (NPC12)3 or (NPF2)3.34,35 Also, aminolysis of
(NPCl2)4 takes place more rapidly than the reaction of (NPC12)3,36
and the degradation of cyclophosphazenes to phosphoranes in the
Fig. 6 Reactions of hexakis(2,2,2-trifluoroethoxy)cyclotriphosphazene
(1) with sodium phenoxide: molar ratio a) 1 : 12; b) 1 : 24.
amount of compound 11 was generated, with roughly 75% of
1 consumed in the end (Fig. 6b). This showed that hexakis(2,2,
2-trifluoroethoxy)cyclophosphazene is more prone to attack by the
nucleophile as the concentration of the phenoxide increases, which
Scheme 4 Products from the reactions between aryloxy tetramers (7–10) and sodium trifluoroethoxide.
2106 | Dalton Trans., 2012, 41, 2100–2109 This journal is The Royal Society of Chemistry 2012
©

View Article Online
Fig. 8 31P NMR spectrum for substituent exchange reaction between 4
Fig. 7 31P NMR spectra for substituent exchange reaction between 7 and
and sodium trifluoroethoxide for: a) 0 h; b) 0.5 h; c) 4 h; d) 1 day.
sodium trifluoroethoxide for 3 days: molar ratio a) 1 : 16; b) 1 : 32.
presence of catechol or o-phenylenediamine occurs more readily
with cyclic tetramers than with trimers.37,38
Reactions of octakis(2,2,2-trifluoroethoxy)cyclophosphazene (6)
with sodium phenoxide
The reverse exchange reaction at the cyclic tetramer level (reactions
j and k) gave results similar to their cyclic trimeric coun-
terparts. No octaphenoxycyclophosphazene (7) was generated
in either reaction after 22 days, and only partially-substituted
N4P4(OCH2CF3)8-x(OC6H5)x (15) species were identified by mass
spectrometry. The value of x can be from 1 to 4. Furthermore, the
side product 13 appeared after 5 days reaction, showing that the
a-carbon attack on the side group by the nucleophile also took
place in these reactions.
Fig. 9 31P NMR spectra for substituent exchange reaction between 9 and
sodium trifluoroethoxide for: a) 0 h; b) 0.5 h; c) 4 h; d) 1 day.
Reactions of 3, 4, 5 and 8, 9, 10 with sodium trifluoroethoxide
as an example, as shown in Fig. 8, all of the starting material
(hexakis(4-cyanophenoxy)cyclophosphazene, 4) was converted to
1 within 30 min without the appearance of any detectable
side product 11. However, after 4 h, compound 11 was de-
tected and increased in concentration as the reaction proceeded.
Similar results at the cyclic tetramer level between octakis(4-
cyanophenoxy)cyclophosphazene (9) and sodium trifluoroethox-
ide were also obtained, as seen in Fig. 9. This process is quite
similar to that of its cyclic trimeric counterpart (4). Therefore,
the assumption proposed at the beginning of this section was
reinforced by the results obtained in these reactions. Thus, the
replacement reactivity was dramatically increased due to the
electron-deficiency within the cyclic trimeric or tetrameric rings.
In all of these side group exchange reactions fully-
substituted hexakis(2,2,2-trifluoroethoxy)cyclophosphazene (1)
or octakis(2,2,2-trifluoroethoxy)cyclophosphazene (6) were gener-
ated by side group exchange. This is similar to the results discussed
above for reactions a and h. However, a much faster exchange
reaction rate occurred with these substituted aryloxycyclote-
traphosphazene than with the parent phenoxyphosphazenes. The
replacement reaction proceeds much faster because of the electron-
withdrawing substituent groups on the aryl rings. This renders
the skeleton more electron-deficient and more liable to attack by
nucleophiles. This assumption was confirmed by the reactions of
cyclic trimers and tetramers with 4-formylphenoxy (3 and 8), 4-
cyanophenoxy (4 and 9) and 4-nitrophenoxy (5 and 10) side groups
with sodium trifluoroethoxide. As shown in Table 2, within half
an hour, hexakis(2,2,2-trifluoroethoxy)cyclophosphazene (1) or
octakis(2,2,2-trifluoroethoxy)cyclophosphazene (6) were formed
by substituent exchange reactions and were detected by 31P
NMR spectrometry. This same process would require 1 or 2
days for hexaphenoxycyclophosphazene (2) and octaphenoxy-
cyclophosphazene (7). Moreover, all of the starting materials
were consumed within half an hour, except in the case of 4-
formylphenoxy trimer (3) (2 h). Again, this is much faster
than the analogous reactions of hexaphenoxycyclophosphazene
(2) and octaphenoxycyclophosphazene (7) with sodium trifluo-
roethoxide (reactions a and b). Taking 4-cyanophenoxy trimer
Stability of 1, 2, 6 and 7 in the presence of nucleophiles
Compounds 11, 13 and 14 always appeared during different
substituent exchange reactions, which indicated a relatively
high sensitivity of both hexakis(2,2,2-trifluoroethoxy)-
cyclophosphazene (1) and octakis(2,2,2-trifluoroethoxy)-
cyclophosphazene (6) under these reaction conditions. Actually,
it has been shown that fluoroalkoxycyclophosphazenes,39
aryloxy- and spiroarylenedioxycyclophosphazenes28 undergo
hydrolysis
in
basic
aqueous
media.
Hexakis(2,2,2-
trifluoroethoxy)cyclophosphazene can be hydrolyzed in solution
by aqueous alkali, undergoing nucleophilic attack by hydroxide
This journal is
The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 2100–2109 | 2107
©

View Article Online
Table 2 Reaction time of aryloxy trimer and tetramer derivatives with sodium trifluoroethoxide
Trimers
Tetramers
2
3
4
5
7
8
9
10
a
b
c
t1
t2
t3
2 dd
—
3 d
0.5 hd
2 h
1 d
0.5 h
0.5 h
4 h
0.5 h
0.5 h
1 h
1 d
7 d
2 d
0.5 h
0.5 h
3 d
0.5 h
0.5 h
1 d
0.5 h
0.5 h
18 h
a Time for the appearance of 1 or 6. b Time for complete consumption of starting material. c Time for the appearance of 11 or 13 and 14. d d: day, h: hour.
ion at phosphorus via an SN2-type mechanism through a
nongeminal pathway.39 Such a mechanism allows retention of
the ~120◦ N–P–N ring angle during formation of a trigonal-
bipyramidal transition state. Moreover, the initial product formed
during the hydrolysis of 1 is believed to be 11. Cyclic tetramers
behave similarly, but with two- to four-fold faster hydrolysis rates,
which can be explained by the reasons discussed above. Care was
taken to exclude water from all the reactions described here, and
it can be assumed that the attacking species is trifluoroethoxide
rather than hydroxide. A similar cleavage mechanism and
products have also been detected from the reaction of excess
nucleophile with heterophosphazenes by Manners et al.40
exposure time of the final products to an excess of the nucleophile
to prevent the generation of the undesired side product.
By contrast, no side product was detected from the reactions
of hexaphenoxycyclophosphazene (2) or octaphenoxycyclophos-
phazene (7) with sodium phenoxide under exchange reaction
conditions after one week. This illustrates the ability of 2 and
7 to resist nucleophilic attack, and may reflect the excellent
protection of the backbone by bulky phenoxy groups especially
to nucleophilic attack by sodium phenoxide. It may also be
connected with the absence of a bridging CH2 group, which would
enhance the conformational flexibility of both the polymer and
the nucleophile and deprive the system of a site for alpha-carbon
attack.
Nevertheless,
a key reaction to supplement the side
group exchange studies is the interaction of hexakis(2,2,2-
trifluoroethoxy)cyclophosphazene (1) with sodium trifluoroethox-
ide (molar ratio 1 : 12) under the same reaction condition as in
all the other reactions. As shown in Fig. 10, the side product
11, appeared after one day of reaction, even though initially it
constituted less than 2% of the product mixture. As the reaction
proceeded, 1 was gradually converted to compound 11 and no
other products were detected. After 41 days, approximately 90%
of 1 had been converted into 11. Moreover, as mentioned earlier,
chemical ionization mass spectrometric analysis of the reaction
mixture revealed the presence of bis(trifluoroethyl) ether (mass
163.01), which is the other side product consistent with the
proposed mechanism. These experiments indicate the sensitivity
of hexakis(2,2,2-trifluoroethoxy)cyclophosphazene (1) when ex-
posed to a large excess of a nucleophile under normal synthetic
conditions. Thus, attention should be paid to this aspect when
synthesizing small molecules or even high polymers to limit the
Conclusions
The side group exchange reactions of cyclic trimeric and tetrameric
aryloxycyclophosphazenes with sodium 2,2,2-trifluoroethoxide
are dependent on the electron-withdrawing effects of different side
groups. The ease of displacement of OAr in cyclic trimeric and
tetrameric molecules by CF3CH2O- increased significantly in the
order, OAr = phenoxy ꢀ 4-formylphenoxy < 4-cyanophenoxy ª 4-
nitrophenoxy derivatives. In addition, the side product generated
via the attack by the nucleophile on the a-carbon of the 2,2,2-
trifluoroethoxy group was detected in all the exchange reactions
after 2,2,2-trifluoroethoxy phosphazene trimer or tetramer had
been formed. The cyclic phenoxyphosphazene trimer and tetramer
showed significantly better stability in the presence of sodium
phenoxide than did the 2,2,2-trifluoroethoxy trimer or tetramer.
The information obtained here is valuable as a model study
for the synthesis of phosphazene high polymers. As in the
cyclic trimer or tetramer, the a-carbon of the fluoroalkoxy
side groups in the corresponding high polymeric phosphazenes
may also be liable to attack by excess nucleophilic reagents in
the reaction solution, ultimately generating products with low
molecular weights. These results are also significant in view of
the report by Ferrar et al. that the reaction of excess sodium
trifluoroethoxide with (NPCl2)n leads to lower molecular weight
trifluoroethoxyphosphazene polymers.41 Furthermore, reaction of
poly[bis(trifluoroethoxy)phosphazene] with excess sodium triflu-
oroethoxide also resulted in the transformation of a crystalline
polymer to an amorphous one. This result coincides with changes
in DSC thermograms, optical micrographs, and 31P NMR spectra,
that suggest a secondary reaction by sodium trifluoroethoxide on
the polymer.
The work reported here is also important with respect to the
order in which nucleophiles should be added to high polymeric
(NPCl2)n to obtain specific mixed-substituent side group ratios and
Fig. 10 Formation of species 11 from the reaction of 1 with sodium
trifluoroethoxide: molar ratio 1 : 12.
2108 | Dalton Trans., 2012, 41, 2100–2109
This journal is
The Royal Society of Chemistry 2012
©

View Article Online
avoid both side group exchange and the formation of side products
similar to 11. In order to understand these aspects in more detail,
comparable reactions of short chain linear phosphazenes and high
polymers are currently underway.
16 H. R. Allcock, A. E. Maher and C. M. Ambler, Macromolecules, 2003,
36, 5566.
17 H. A. Bates, J. Am. Chem. Soc., 1982, 104, 2490.
18 G. A. Carriedo, J. Jimenez, P. Gomez-Elipe and F. J. G. Alonso,
Macromol. Rapid Comm., 2001, 22, 444.
19 G. A. Carriedo, J. I. F. Martinez, F. J. G. Alonso, E. R. Gonzalez and
A. P. Soto, Eur. J. Inorg,. Chem., 2002, 1502.
20 A. E. Maher and H. R. Allcock, Macromolecules, 2005, 38, 641.
21 H. R. Allcock and W. E. Krause, Macromolecules, 1997, 30, 5683.
22 H. R. Allcock, P. E. Austin and T. F. Rakowsky, Macromolecules, 1981,
14, 1622.
23 H. R. Allcock, M. S. Connolly and P. J. Harris, J. Am. Chem. Soc.,
1982, 104, 2482.
24 H. R. Allcock and G. Y. Moore, Macromolecules, 1972, 5, 231.
25 A. B. Pangborn, M. A. Giardello, R. H. Grubbs, R. K. Rosen and F. J.
Timmers, Organometallics, 1996, 15, 1518.
26 H. R. Allcock and P. E. Austin, Macromolecules, 1981, 14, 1616.
27 G. A. Carriedo, L. FernandezCatuxo, F. J. G. Alonso, P. G. Elipe, P. A.
Gonzalez and G. Sanchez, J. Appl. Polym. Sci., 1996, 59, 1879.
28 H. R. Allcock and E. J. Walsh, J. Am. Chem. Soc., 1972, 94, 4538.
29 J. P. Obrien, W. T. Ferrar and H. R. Allcock, Macromolecules, 1979, 12,
108.
Notes and references
1 H. R. Allcock, Chemistry and Applications of Polyphosphazenes, Wiley-
Interscience: Hoboken, NJ, 2003.
2 A. L. Weikel, S. Y. Cho, N. L. Morozowich, L. S. Nair,
C. T. Laurencin and H. R. Allcock, Polym. Chem., 2010, 1,
1459.
3 M. Deng, L. S. Nair, S. P. Nukavarapu, S. G. Kumbar, T. Jiang, A.
L. Weikel, N. R. Krogman, H. R. Allcock and C. T. Laurencin, Adv.
Funct. Mater., 2010, 20, 2794.
4 M. Deng, L. S. Nair, S. R. Nukavarapu, T. Jiang, W. A. Kanner, X. D.
Li, S. G. Kumbar, A. L. Weikel, N. R. Krogman, H. R. Allcock and C.
T. Laurencin, Biomaterials, 2010, 31, 4898.
5 C. S. Reed, J. P. Taylor, K. S. Guigley, M. M. Coleman and H. R.
Allcock, Polym. Eng. Sci., 2000, 40, 465.
6 H. R. Allcock and J. P. Taylor, Polym. Eng. Sci., 2000, 40, 1177.
7 H. R. Allcock, Adv. Mater., 1994, 6, 106.
8 S. T. Fei, R. M. Wood, D. K. Lee, D. A. Stone, H. L. Chang and H. R.
Allcock, J. Membrane Sci., 2008, 320, 206.
9 J. W. Fu, Q. Xu, J. F. Chen, Z. M. Chen, X. B. Huang and X. Z. Tang,
Chem. Commun., 2010, 46, 6563.
30 K. Inoue, T. Yamauchi, T. Itoh and E. Ihara, J. Inorg. Organomet.
Polym. Mater., 2007, 17, 367.
31 E. Kober, H. Lederle and G. Ottmann, Inorg. Chem., 1966, 5, 2239.
32 H. R. Allcock and L. A. Smeltz, J. Am. Chem. Soc., 1976, 98, 4143.
33 D. B. Sowerby, J. Chem,. Soc., 1965, 1396.
34 F. Seel and J. Langer, Z. Anorg. Allg. Chem., 1958, 295, 316.
35 H. N. Stokes, American Chemical Journal, 1896, 18, 780–788.
36 T. Moeller and S. G. Kokalis, J. Inorg. Nucl. Chem., 1963, 25, 1397.
37 H. R. Allcock and R. L. Kugel, J. Am. Chem. Soc., 1969, 91, 5452.
38 H. R. Allcock and E. J. Walsh, Inorg. Chem., 1971, 10, 1643.
39 H. R. Allcock and E. J. Walsh, J. Am. Chem. Soc., 1972, 94, 119.
40 D. P. Gates, P. Park, M. Liang, M. Edwards, C. Angelakos, L. M.
LiableSands, A. L. Rheingold and I. Manners, Inorg. Chem., 1996, 35,
4301.
10 E. M. W. Tsang, Z. B. Zhang, A. C. C. Yang, Z. Q. Shi, T. J. Peckham,
R. Narimani, B. J. Frisken and S. Holdcroft, Macromolecules, 2009, 42,
9467.
11 D. K. Lee and H. R. Allcock, Solid State Ionics, 2010, 181, 1721.
12 J. Thielen, W. H. Meyer and K. Landfester, Chem. Mater., 2011, 23,
2120.
13 H. R. Allcock, R. L. Kugel and K. J. Valan, Inorg. Chem., 1966, 5,
1709.
41 W. T. Ferrar, A. S. Marshall and J. Whitefield, Macromolecules, 1987,
14 H. R. Allcock and R. L. Kugel, Inorg. Chem., 1966, 5, 1716.
15 H. R. Allcock and Y. B. Kim, Macromolecules, 1994, 27, 3933.
20, 317.
This journal is
The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 2100–2109 | 2109
©