Highly branched poly(N-isopropylacrylamide)s with imidazole end groups prepared by radical polymerization in the presence of a styryl monomer containing a dithioester group

Article abstract of DOI:10.1021/ma047742n

Highly branched poly(N-isopropylacrylamide) (PNIPAM) compounds were prepared by copolymerization of 3H-imidazole-4-carbodithioic acid 4-vinylbenzyl ester, 1, with N-isopropylacrylaniide (NIPAM) using reversible addition-fragmentation chain transfer (RAFT)

Full text of DOI:10.1021/ma047742n

Macromolecules 2005, 38, 4595-4603
4595
Highly Branched Poly(N-isopropylacrylamide)s with Imidazole End
Groups Prepared by Radical Polymerization in the Presence of a Styryl
Monomer Containing a Dithioester Group
Steven Carter, Barry Hunt, and Stephen Rimmer*
The Polymer and Biomaterials Chemistry Laboratories, Department of Chemistry (Polymer Centre),
University of Sheffield, Sheffield, South Yorkshire, UK S3 7HF
Received November 3, 2004; Revised Manuscript Received March 23, 2005
ABSTRACT: Highly branched poly(N-isopropylacrylamide) (PNIPAM) compounds were prepared by
copolymerization of 3H-imidazole-4-carbodithioic acid 4-vinylbenzyl ester, 1, with N-isopropylacrylamide
(NIPAM) using reversible addition-fragmentation chain transfer (RAFT) polymerization. The polymeriza-
tions proceeded well with few side reactions. An increase in the content of 1 in the monomer feed appears
to increase the number of branch chains, and at the same time no evidence was obtained for the presence
of substituted acrylamide chain ends that may potentially result via elimination of the dithioate group.
The polymer products show a clear tendency to increased molecular weight as the extent of conversion
of monomer increases, while size exclusion chromatography (SEC) profiles indicate a complex distribution
of molecular weights compared to linear polymers obtained with a non-RAFT carboxylate monomer. Both
NMR and viscometry indicate that, as expected, increasing the amount of 1 in the feed has the effect of
increasing the degree of branching in the final product. This increase in branching reduces the intrinsic
viscosity of the solutions of the highly branched polymers compared to similar linear polymers. Poly(N-
isopropylacrylamide) displays a lower critical solution temperature (LCST) in aqueous solutions, and
cloud point data indicate a clear effect of chain architecture on the temperature at which this transition
occurs. Thus, a set of linear analogous copolymers have LCST’s that, for equivalent mole fractions of
imidazole content, are higher than the similar highly branched polymers. However, on complexation of
copper by the imidazole groups the LCSTs of the linear and highly branched sets cannot be differentiated.
Introduction
polymeric dithioate compound via the formation of an
intermediate radical. In this way polymers of low
polydispersity can be obtained and a variety of R and Z
groups exploited to give polymer chain-end functional-
ity.
Branched polymers owe much of their utility to the
presence of a large number of chain ends per molecule
and their chain architecture. The latter can have a
profound effect on materials properties, such as rheology
and solubility.1,2 On the other hand, the large number
of chain ends can be used to add useful chemical
functionality, which may differ from similar functional-
ity added along the main chain.2-5 These polymers can
be produced in chain growth polymerization by using
branching monomers, which act as both monomers and
transfer agents or as monomers and initiators.6 Both
of these functions are combined in approaches that use
addition-fragmentation as the branch forming reac-
tion,7 and of these, the reversible addition-fragmenta-
tion chain transfer (RAFT) methodology introduced by
Thang et al.8,9 also offers the opportunity to modify the
end groups. This method has proved to be a useful
technique for the synthesis of reactive polymers and
polymers with well-defined architectures. The currently
accepted mechanism of RAFT polymerization, as docu-
mented by other authors, has been formulated so as to
proceed by a series of reversible addition-fragmentation
steps, as shown in Scheme 1. The process is initiated
by the formation of a propagating radical Pn• which adds
to the thiocarbonyl group of the RAFT agent to give a
radical adduct which fragments to a polymeric dithioate
species and a new radical adduct, R•. This radical then
reinitiates polymerization to give a new propagating
radical, Pm•. An equilibrium is thus set up between the
RAFT polymerization has been used to prepare block
copolymers10 and star-shaped polymers.11,12 Yang et al.13
prepared highly branched polystyrene using a RAFT
polymerization by incorporating a polymerizable dithio-
ate ester as the RAFT agent. These branching mono-
mers based on vinyl dithioate esters place the dithioate
ester group at the chain ends. Therefore, if the RAFT
agent contains a specific chemical functionality, this will
also be located at the chain ends. More recently, a
dithioate ester was used as a transfer agent aimed at
preventing the gelation of a radical polymerization
containing a difunctional monomer.14 Alternatively, the
dithioate ester chain ends can be further modified to
provide the target functionality.15
In the work reported here we describe the use of
an imidazole functional RAFT agent, 1, used to pre-
pare stimuli responsive poly(N-isopropylacrylamide)
(PNIPAM), highly, branched polymers with imidazole
groups at the chain ends. The latter function finds
application as a ligand in protein purification16-20 and
potentially in gene and drug delivery.21-23 The lower
critical solution temperature of these polymers was then
compared to those of linear analogues produced by
copolymerization of NIPAM with 2. We also describe an
initial study using the RAFT agent 3, which contains
the same imidazole functionality as 1 but lacks the
styryl double bond. Structures 1-3 are shown in Figure
1. PNIPAM is probably the most well-studied stimuli
responsive polymer, and it has attracted considerable
•
•
propagating radicals Pn and Pm and the dormant
* Corresponding author: Tel +44 114 222 5965; Fax +44 114
222 9346, e-mail S.Rimmer@Sheffield.ac.uk.
10.1021/ma047742n CCC: $30.25 © 2005 American Chemical Society
Published on Web 05/04/2005

4596 Carter et al.
Macromolecules, Vol. 38, No. 11, 2005
Scheme 1. Currently Accepted Mechanism for RAFT Polymerization
Table 1. Triple Detection SEC Data (Eluant Was THF/
TBAB, with Triple Detection) for Linear PNIPAM
Generated by Polymerization in the Presence of 3
attention as the basis of many thermally responsive
polymer systems. The behaviors of both the homo-
polymer24-30 and its copolymers31,32 are very well stud-
ied due to the narrow temperature range over which
the lower critical solution transition occurs in aqueous
media.33 NIPAM has also been successfully polymerized
using the RAFT technique.34,35
mole ratio
NIPAM/3
temp (°C)
conv (mol %)
Mw (g mol-1
)
PDI
116:1
174:1
232:1
87:1
116:1
174:1
232:1
348:1
1394:1
60
60
60
80
80
80
80
80
80
20
51
45
35
67
56
34
48
56
7 600
12 000
22 500
31 200
38 500
51 000
58 800
62 600
69 200
1.12
1.35
1.26
1.43
1.41
1.21
1.21
1.26
1.29
The latter data (see Table 1) were obtained by triple
detection SEC in THF (0.1 wt % tetrabutylammonium
bromide). Table 1 shows that the polydispersity (PdI)
indexes ranged from 1.2 to 1.45 and were similar to
those obtained by Mu¨ller.35 The PdIs are a good indica-
tion of the control provided to the radical polymerization
by the inclusion of 3. The data thus suggest that the
RAFT mechanism outlined in Scheme 1 is occurring at
the two specified temperatures. Further characteriza-
tion using MALDI-TOF mass spectrometry (using
entry 1 from Table 1) showed, as predicted, the presence
of the benzyl and 4(5)-imidazole dithioate end groups
(see Supporting Information).
Highly Branched PNIPAM. The formation of the
PNIPAM highly branched polymers required the syn-
thesis of the imidazole-functionalized RAFT agent, 3H-
imidazole-4-carbodithioic acid 4-vinylbenzyl ester, 1.
The use of this styryl imidazole-dithioate allowed
copolymerization with NIPAM to proceed via the pres-
ence of a styryl double bond and chain branching to
proceed via RAFT polymerization with the imidazole
dithioate. RAFT agent 1 was synthesized in reasonable
Figure 1. Structures of imidazole species that were incorpo-
rated in polymerizations of NIPAM.
Results and Discussion
Linear PNIPAM. The use of RAFT polymerization
to synthesize low-polydispersity linear PNIPAM has
previously been carried out by other workers using
benzyl 1-pyrrolecarbodithioate as the chain-transfer
agent in 1,4-dioxane at 60 °C with AIBN as initiator.35
In a preliminary study designed to investigate the
feasibility of using 1 to produce highly branched
PNIPAMs, we carried out a similar series of RAFT
polymerization reactions using 3 and NIPAM. By reduc-
ing the amount of 3, we were able to selectively
synthesize a range of PNIPAMs of increasing molecular
weights.

Macromolecules, Vol. 38, No. 11, 2005
Highly Branched Poly(N-isopropylacrylamide)s 4597
Scheme 2. Highly Branched PNIPAM Prepared by RAFT Polymerization
yield via the cesium carbonate-assisted coupling of 4(5)-
imidazole dithioic acid to 4-vinylbenzyl chloride.
were observed, also as broad peaks, between 6.7 and
7.3 ppm. The results of the calculations derived from
these spectra are shown in Table 2. The data show that
it was possible to produce average values of degrees of
branching (DB) by comparison of the NIPAM isopropyl
methine protons at δ 4.0 to the aromatic protons
(imidazole + xylyl) at δ 6.7-8.0 to obtain the number
of repeat units per branch (RB).
The data also show that the monomer feed and
polymer compositions are very similar after polymeri-
zation for 24 h and that they are essentially within the
experimental error of the NMR technique, which was
calculated to be (5%. These analyses thus suggest that,
at least at these low levels of mole fraction of 1 in the
feed, the monomer feed can be used to approximate the
overall polymer composition.
All of the attempted RAFT polymerizations of NIPAM
in the presence of 1 were carried out for 24 h and
generated polymeric materials as hard free-flowing
powders after reprecipitation. Scheme 2 shows the
generalized scheme for the preparation of these highly
branched PNIPAMs with imidazole end groups. As can
be seen from these structures, the imidazole groups sit
at the chain ends and the styryl unit provides a
branching point. Evidence for the presence of the
imidazole groups at the termini of polymer branches
was obtained via 1H NMR as shown in Figure 2, in
which broad signals that were derived from the CH
imidazole resonances were apparent at δ 7.5-8.0 (see
the expanded region of Figure 2). The styryl residues
Figure 2. 1H NMR (400 MHz) spectrum of highly branched PNIPAM (υ1 ) 0.042, υ1′ ) 0.039) at 24 h reaction time.
Table 2. Results of Polymerization Reactions of NIPAM in the Presence of 1 at 60 °Ca
yield (%)
P
NIPAM/1 (feed)
υ1
υ1′
RB
DB
24 h
48 h
Mw (g mol-1
)
24 h
48 h
15:1
23:1
25:1
28:1
30:1
33:1
46:1
0.063
0.042
0.039
0.035
0.033
0.029
0.021
0.077
0.039
0.048
0.042
0.036
0.032
0.024
5.2
8.8
0.19a
0.12a
0.17a
0.16b
0.15b
0.08b
0.07b
18
17
19
23
19
22
15
52
48
47
42
45
52
39
61 300
60 900
19 748
25 090
24 672
22 960
53 400
0.6
0.8
0.9
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
5.8
6.0
6.5
12.5
13.8
a υ1 ) mole fraction of 1 in the monomer feed; υ1′ ) mole fraction of 1 in the polymer (from 1H NMR spectra); RB ) repeat units per
branch; DB ) degree of branching (1/RB) as determined by 1H NMR after a ) 48 h and b ) 24 h and where RB ) {area(CH-isopropyl)
+ [area(CH-Im) + area(CH-aromatic)]/6}/{[area(CH-Im) + area(CH-aromatic)]/6}. P ) proportion of chain transfer agent in highly
branched polymer; at 70 °C/28 h DB ) 0.04. The associated error in measuring integrated NMR signals for all polymer samples was
estimated to be (5%.

4598 Carter et al.
Macromolecules, Vol. 38, No. 11, 2005
Figure 3. Typical SEC chromatograms derived from (a) NIPAM polymerized in the presence of 1 and (b) linear polymer prepared
copolymerization of NIPAM and 2.
Table 3. Results of Copolymerization Reactions of 2 with
The NMR spectrum in Figure 2 can also be used to
NIPAMa
determine the proportion of chain transfer agent 1 that
υ2
υ2′
Mwb (g mol-1
)
PDI
conv (mol %)
has been incorporated into these polymers (P) by
1
examining the H NMR spectra at δ ) 5.15-5.65 ppm
0.021
0.026
0.029
0.033
0.036
0.039
0.040
0.046
0.053
0.019
0.022
0.027
0.026
0.028
0.034
0.044
0.049
0.049
148 000
119 000
103 000
73 000
72 300
59 600
85 500
84 600
76 900
5.1
4.1
3.3
3.5
3.9
5.0
4.0
4.6
4.1
74
72
52
51
58
39
47
34
44
(see expanded region) attributable to the CH2d (styryl)
groups. The ratio of these signals to those of the
imidazole end groups (δ ) 7.5-8.0 ppm) then gives the
proportion of unreacted styryl groups in the polymer.
The other vinyl proton (CH2dCH-) of the styryl group
is observed at 6.6 ppm and serves to unequivocally
assign these resonances as vinyl end groups. They are
the result of the incomplete conversion of styryl groups
that arise from molecules of 1 that have undergone
transfer and reinitiation but have not propagated by
copolymerization with NIPAM. The results of the chain
end functionality calculations are given in Table 2. As
can be seen from these data, significant amounts of
alkene (stryryl) end groups were only observed in
polymerizations in which the fraction of 1 in the
monomer feed exceeded 3.8 mol % and when the
reactions were run for 24 h. The resonances due to vinyl
peaks were absent from the products obtained from
polymerizations that were carried out for 48 h (Table
2). This observation confirms that the vinyl resonances
are due to incomplete conversion of the styryl group of
1. It should also be noted we could not assign any
resonances from alkene chain ends produced by base-
catalyzed or pyrolytic elimination of the dithioate end
group.
a υ2 ) mole fraction of 2 in the monomer feed; υ2′ ) mole fraction
of 2 in the polymer (from 1H NMR spectra). b Determined by SEC
(DMF, PEO standards).
in accordance with Table 2 which are exemplified by
the log(molecular weight) plots in Figure 4. All of the
polymers were characterized by high Mw values and
high polydispersities, which ranged from 6 to 14. The
SEC profiles do not appear to be due to simple mixtures
and illustrate the complex nature of these reactions.
We next repeated the polymerization carried out with
a feed mole fraction of 1, υ1 ) 0.063 (mole fraction 1 in
polymer from 1H NMR, υ1′ ) 0.077), which corresponded
to a NIPAM:1 feed ratio of 15:1. The reaction was
repeated and sampled in order to enable a study, using
SEC, of the development of molecular weight and its
distribution during the polymerization. Figure 5 shows
the molecular weight distributions determined at vari-
ous reaction times, as determined using a poly(ethylene
oxide) calibration.
To minimize adsorption phenomena, we used an
eluant composed of DMF containing 0.1% ammonium
acetate at 70 °C. The first observation to be made from
these data is that the molecular weight distributions
shift to higher values of log(molecular weight) as the
polymerization progresses. Unfortunately, the changes
in the form of the distribution with reaction time
preclude the formation of the molecular weight vs
monomer conversion plots that are often associated with
studies of living or living-radical chain growth polymer-
izations. However, it is clear qualitatively from the data
illustrated in Figure 5 that an increase in molecular
weight occurs as the extent of monomer conversion
increases. The branching process produces increasingly
broader molecular weight distributions as the polym-
erization progresses, and a clear shift to higher molec-
ular weights is observed. These observations are in
agreement with the predictions of Mu¨ller et al.,36 who
have shown that the molecular weight distributions
produced by related self-condensing vinyl polymeriza-
tions broaden exponentially with the degree of conver-
sion of the vinyl groups. We also noted that while the
Polymerizations in the presence of 1 produced high
molecular weight polymer, and as expected in a polym-
erization containing a branching monomer, the molec-
ular weight distributions were broad and multimodal.36
The multimodal chromatograms are exemplified, in
Figure 3, by the size exclusion chromatography (SEC)-
derived molecular weight distributions. The figure also
shows data derived from copolymerization of NIPAM
with 2, a nonbranching imidazole-functional comono-
mer. These SEC results illustrate that the molecular
weight distributions of these linear copolymers are
unimodal. Therefore, the multimodal distributions ob-
served in polymerizations incorporating 1 appear to be
a consequence of the branch producing transfer events.
The results from all of the copolymerizations incorporat-
ing 2 are given in Table 3. Two aspects are worthy of
note from these data. First, over these limited feed
ratios, the feed and final polymer compositions are
essentially the same. Second, the polydispersities range
from 3.3 to 5.1, i.e., values that are typical for radical
copolymerizations taken to moderate to high conver-
sions.
We also obtained the SEC profiles of the highly
branched polymers prepared at various NIPAM:1 ratios

Macromolecules, Vol. 38, No. 11, 2005
Highly Branched Poly(N-isopropylacrylamide)s 4599
Figure 4. SEC (eluant 0.1% ammonium acetate in DMF at 70 °C) profiles for highly branched polymers (24 h reaction time)
showing the effect of NIPAM:1 feed ratios.
Figure 5. Molecular weight distributions for polymerizations of NIPAM in the presence of 1, obtained at various reaction times
(υ1 ) 0.063, υ1′ ) 0.077).
Figure 7. Mark-Houwink plots for linear (b) and branched
polymer (a) (υ1 ) 0.063, υ1′ ) 0.077) generated at 60 °C.
CH signals at δ ) 5.15-5.65 ppm that may have been
attributable to end-chain or styryl double bonds. It was
Figure 6. Time-conversion plots for polymerizations of
NIPAM/1 at 60 (9), 70 (b), and 80 (2) °C using a monomer
feed ratio of υ1 ) 0.063.
also found that the DB value was significantly lower
(see Table 2). Thus, it would appear that at these
elevated temperatures the copolymerization reaction
between the styryl groups of 1 and NIPAM occurs to a
greater extent than at 60 °C. SEC results also revealed
the presence of a relatively large amount of high
molecular weight polymer at 70 °C, and this would
suggest that at elevated temperatures highly branched
polymer is produced via the RAFT mechanism, albeit
after a shorter induction period.
yield of polymer could be increased beyond 40% using
an extended reaction time of 48 h, the SEC profile did
not appear to notably change beyond 24 h.
Next several reactions were performed at a range of
temperatures (at 60, 70, and 80 °C).The reaction time
vs monomer conversion plots were produced gravimetri-
cally and are shown in Figure 6.
SEC Viscometry. Ideally, we would have compared
data from SEC with triple detection of the linear co-
polymers derived from copolymerization of NIPAM and
2 with the branched polymers derived from polymeri-
zation of NIPAM in the presence of 1. However, the
latter were only partially soluble in THF and gave
generally a two-phase mixture of solution and dispersed
polymer. We therefore carried out reversible addition-
fragmentation chain transfer (RAFT) polymerization
studies using the polymers prepared from polymeriza-
tion of NIPAM in the presence of 3. However, these
studies were limited by the decrease in solubility of the
The polymerization carried out at 60 °C gave a strong
indication of the induction period that is often associated
with RAFT polymerizations.35,37 In these cases the
induction period appears to be completed by ≈16 h
reaction time. At elevated temperatures, however, this
long induction period is not evident, and to investigate
1
this effect further, we chose to examine the H NMR
spectrum of the polymer produced at a NIPAM:1 feed
ratio of 15:1 after reaction at 70 °C for 28 h. In this case
we found that the imidazole CH signals were still
present while there was no evidence for the presence of

4600 Carter et al.
Macromolecules, Vol. 38, No. 11, 2005
Figure 8. Mark-Houwink plots for linear (b) and branched polymer (a) (υ1 ) 0.021, υ1′ ) 0.024) generated at 60 °C.
polymers in THF as the amount of 1 was increased. It
was not possible to use DMF as the eluant with the
viscometric detector because of the high viscosity of this
solvent. Within these limitations we were able to extract
some useful information using this technique. Mark-
Houwink plots were obtained by plotting log[η] vs
log(molecular weight) for selected branched polymers
generated from the polymerization of NIPAM in the
presence of 1 and for the linear PNIPAMs prepared in
the presence of 3. The linear copolymers were produced
with the aim of comparing the behavior of the highly
branched polymers to linear polymers with imidazole
functionality. A highly branched polymer generated
from a feed ratio of υ1 ) 0.063 (υ1′ ) 0.077) at 60 °C
gave the Mark-Houwink plot shown in Figure 7a, while
Figure 9. Plots of cloud point vs mole fraction of imidazole
a linear polymer (mole ratio NIPAM:3 ) 1394:1), which
for branched (9) and linear PNIPAMs (b) and the cloud points
had a molecular weight distribution that lay within the
obtained from the branched (1) and linear ([) PNIPAMs after
same range, gave the plot in Figure 7b. It is clear from
these data that at all molecular weights the highly
branched polymer gave lower values of log[η] and that
plot (a) had a considerably lower gradient. The branched
polymer was characterized by R ) 0.166 compared to
its linear counterpart, which gave R ) 0.612.
the addition of Cu(II) ions.
cloud points for the linear polymers were significantly
higher than the cloud points of their branched equiva-
lents. This change in cloud point was unexpected since,
if we consider that the system’s free volume would be
expected to increase as the average number of end
groups per chain increases, it might reasonably be
expected that the cloud point would increase with
branching. These results can only be rationalized if we
assume that, in the absence of an added metal ion, the
imidazole end groups can aggregate more effectively
than the pendant imidazole groups on the linear poly-
mers. The addition of Cu(II) ions substantially raised
and sharpened the cloud points for each polymer to an
upper limit of 32-36 °C for the branched polymers and
33-36 °C for their linear counterparts. The cloud points
of the branched and linear polymers in the presence of
copper ions thus appear to be similar. A tentative
explanation for this effect would be that in the native
form both the linear and branched polymers are present
as aggregates in solution, and this may be due to the
presence of extensive intermolecular binding interac-
tions occurring between the imidazole groups. The
addition of Cu(II) upon binding to the imidazoles has
the effect of rendering the chain ends more hydrophilic
and at the same time disrupting these interactions so
that aggregation is less likely to occur unless the
temperature is further raised. Further work is being
carried out in our laboratories to study these effects in
detail using light scattering, NMR and luminescence
techniques and will be reported in due course.
The effect of decreasing the amount of 1 in the feed
is illustrated in Figure 8, which shows the Mark-
Houwink plot for a branched polymer generated from a
feed of υ1 ) 0.021 (υ1′ ) 0.024) at 60 °C (a) and the linear
analogue (mole ratio NIPAM:3 ) 87:1) (b). These data
are similar to those of Figure 7, in that, for all values
of log M, log[η] is lower for the branched material than
it is in the linear polymer. However, in Figure 8 the
slope of the plot for the highly branched polymer,
characterized by R ) 0.253, more closely approaches
that of the linear material where R ) 0.555. Thus, these
results support the expectation that increasing the
amount of the branching RAFT agent, 1, increases the
degree of branching and that this in turn further
decreases solution viscosity.
Cloud Point Analysis of Highly Branched Poly-
mers and Linear Analogues. The highly branched
polymers formed from the polymerization of NIPAM in
the presence of 1 were ultrafiltered through a cellulose
membrane (MWCO 10 000), dissolved in water (10 mg/
mL) at 0-1 °C, and the cloud points were determined
visually. The range of linear analogous polymers formed
from NIPAM and 2 with an equivalent hydrophobe
content as the branched versions were also dissolved
in water at low temperature and the cloud points
determined. For each polymer solution the cloud point
was also recorded in the presence of copper(II) ions by
the addition of an aliquot of copper(II) sulfate. In Figure
9, the cloud points of all of these polymers are plotted
against proportion of imidazole functional polymer. The
Summary
In summary we have shown that RAFT controlled
radical polymerization can be used to produce highly

Macromolecules, Vol. 38, No. 11, 2005
Highly Branched Poly(N-isopropylacrylamide)s 4601
Synthesis of 3H-Imidazole-4-carbodithioic Acid 4-
Vinylbenzyl Ester, 1. 4(5)-Imidazoledithiocarboxylic acid
(22.05 g, 70%, 107 mmol) was dissolved in DMF (450 mL), in
which an excess of cesium carbonate (154 g, 471 mmol, 4.4
equiv) was suspended and the mixture stirred rapidly at room
temperature under a nitrogen atmosphere for 30 min to form
a deep orange coloration. 4-Vinylbenzyl chloride (18.33 g, 16.93
mL, 120 mmol) was added to the suspension in one portion,
and stirring continued for 70 h. The suspension was filtered
to remove cesium carbonate, and the solids were washed with
dichloromethane. The DMF was removed by rotary evapora-
tion (50 °C, high vacuum), and the remaining oil was dissolved
in dichloromethane (800 mL) and then washed with saturated
sodium hydrogen carbonate solution (2 × 500 mL), dried over
anhydrous magnesium sulfate, filtered, and evaporated to give
a dense orange-brown oil.
An initial purification was carried out using flash silica
column chromatography (2.5% MeOH/CH2Cl2) to give a semi-
pure oily solid. A final purification was carried out by elution
down an alumina column (2% MeOH/CH2Cl2), and after
evaporation of the appropriate fractions the title compound
was isolated as bright orange crystals (7.47 g, 26.8%), Rf (silica,
2.5% MeOH/CH2Cl2) 0.19. 1H NMR (CDCl3, ca. 5% CD3OD,
RT, 250 MHz): δ/ppm 4.52 (s, 2H, -S-CH2-), 5.15 (d, 1H,
dCH2 vinyl, J ) 10.63 Hz), 5.65 (d, 1H, dCH2 vinyl, J ) 17.51
Hz), 6.62 (dd, 1H, dCH vinyl, J1 ) 16.26 Hz, J2 ) 9.38 Hz),
7.25 (s, 4H, -C6H4-)7.60 (s, 1H, CH-imidazole), 7.78 (s, 1H,
CH-imidazole). LC-MS (TOF, ES+): 12.67 min [50% AcN (0.1%
formic acid)/50% H2O (0.1% formic acid)-95% AcN, 20 min],
261 (M+), 262 (MH+) [desolvation gas 462 l h-1, capillary 3129
V, sample cone 36 V, extraction cone 3 V, desolvation temper-
ature 150 °C, source temperature 100 °C]. Calculated for
C13H12N2S2: C, 59.97; H, 4.65; N, 10.76; S, 24.63. Found: C,
59.78; H, 4.68; N, 10.72; S, 24.55.
Synthesis of 4(5)-Imidazole-(4-vinylbenzyl carbox-
ylate), 2. 4-Imidazolecarboxylic acid (5.40 g, 48.2 mmol) was
suspended in DMF (160 mL) and stirred under a nitrogen
atmosphere. 4-Vinylbenzyl chloride (7.36 g, 48.2 mmol) was
added to the suspension followed by triethylamine (4.87 g, 48.2
mmol) and water (0.867 g, 48.2 mmol), and then stirring was
continued for 24 h at room temperature. The suspension was
partitioned between water (150 mL) and ethyl acetate (250
mL) and then separated, and the aqueous layer was further
extracted with ethyl acetate (2 × 250 mL). The organic extracts
were combined, dried over magnesium sulfate, filtered, and
evaporated to give an oil in residual DMF. The excess solvent
was removed in vacuo (high vacuum) at 50 °C. The resultant
oil was separated using flash column chromatography with
95:5 dichloromethane/methanol on silica to give 1.134 g (10.3%)
of the title compound as a white solid, Rf (silica, 95:5 di-
chloromethane/methanol) 0.15. 1H NMR (CDCl3/ca. 5% CD3OD,
RT, 250 MHz): δ/ppm 5.21 (s, 2H, -O-CH2-), 5.18 (d, 1H,
dCH2, vinyl, J ) 10.63 Hz), 5.68 (d, 1H, dCH2 vinyl, J ) 17.51
Hz), 6.63 (dd, 1H, dCH vinyl, J1 ) 16.26 Hz, J2 ) 9.38 Hz),
7.32 (s, 4H, -C6H4-, 7.60 (s, 1H, CH-imidazole), 7.65 (s, 1H,
CH-imidazole). LC-MS (TOF, ES+): 5.70 min [50% AcN (0.1%
formic acid)/50% H2O (0.1% formic acid)-95% AcN, 20 min],
229 (M+), 230 (MH+) [desolvation gas 462 l h-1, capillary 3129
V, sample cone 36 V, extraction cone 3 V, desolvation temper-
ature 150 °C, source temperature 100 °C]. Calculated for
C13H12N2O2: C, 68.41; H, 5.30; N, 12.27; O, 14.02. Found: C,
68.52; H, 5.05; N, 12.20; O, 14.23.
Synthesis of Benzyl-4(5)-imidazole Dithioate, 3. Ce-
sium carbonate (149 g, 0.458 mmol) was added to 4(5)-
imidazoledithiocarboxylic acid (21.43 g, 0.104 mmol) dissolved
in DMF (400 mL), and the mixture was rapidly stirred at room
temperature under a nitrogen atmosphere for 30 min. To the
suspension benzyl bromide (19.57 g, 0.114 mmol) was added
as a solution in DMF (50 mL), and stirring continued in a foil-
covered flask for 5 days. The suspension was filtered to remove
excess cesium carbonate, and the DMF was removed in vacuo
(50 °C, high vacuum). The resultant oil was dissolved in
dichloromethane (600 mL), then washed with saturated aque-
ous sodium hydrogen carbonate (2 × 500 mL), dried over
anhydrous magnesium sulfate, filtered, and evaporated to give
branched PNIPAMs by incorporation of a styryl dithio-
ate ester in the polymerization mixture. In this case the
RAFT agent also carried imidazole functionality, which
was placed at the chain ends of the final product.
Evidence for the livingness of the procedure was ob-
tained using a RAFT agent which did not have a
polymerizable styrenic double bond in its structure, but
which produced linear PNIPAMs of low polydispersity.
Mark-Houwink plots also showed the highly branched
nature of the polymers. Thus, when larger amounts of
the branching agent were incorporated in the structure,
the viscometry profiles displayed a greater divergence
in slope of the plot when compared to a linear PNIPAM.
Surprisingly, the highly branched polymers had cloud
points that were lower than equivalent linear polymers
that were produced by copolymerizing NIPAM with a
styryl monomer carrying an ester and imidazole func-
tionality. We tentatively suggest that the decrease in
cloud point as the degree of branching increases may
be due to aggregation of the imidazole end groups.
Experimental Section
Materials. N-Isopropylacrylamide (Aldrich, 97%) was re-
crystallized (3×) from hexane (Fisher, HPLC grade) via
dissolution at 45 °C and then cooling under refrigeration. 4(5)-
Imidazoledithiocarboxylic acid (Aldrich, 70%) and 4-imidazole-
carboxylic acid (Aldrich, 98%) were used without further
purification. Cesium carbonate (Aldrich, 99%) was ground to
a fine powder and dried at 200 °C. 4-Vinylbenzyl chloride
(Aldrich, 90%) was distilled under reduced pressure, and
benzyl bromide (Fluka, 98%) was used without further puri-
fication. Triethylamine (BDH, 99%) was distilled prior to use,
and R-azoisobutyronitrile (AIBN) (BDH, 97%) was recrystal-
lized from diethyl ether (Fisher, HPLC grade). N,N-Dimethyl-
formamide (DMF) (Aldrich, sure-seal) and dioxane (Aldrich,
sure-seal) were used as purchased; tetrahydrofuran and
dichloromethane were distilled from sodium or calcium hy-
dride, respectively, prior to use. Chromatographic media were
silica gel (BDH, 40-63 µm) and aluminum oxide (BDH,
Brockmann Grade 1). TLC was carried out using Merck silica
gel 60 F254 (aluminum sheet) plates and Macherey-Nagel
(Germany) aluminum oxide/UV254, 0.2 mm (precoated plastic
sheet) plates.
Instrumentation. Average molecular weights and molec-
ular weight distributions (measured relative to poly(ethylene
oxide) (PEO) standards) of polymers were measured by gel
permeation chromatography (GPC) with PL gel mixed B (10
µm particle size, 100-106 Å pore size, effective MW range 103-
10 × 106, 3 × 30 cm + guard columns) (Polymer Laboratories,
UK) and a refractive index detector. DMF was used as the
eluent at a flow rate of 1.0 cm3 min-1 at 70 °C. Sample
concentrations were approximately 2.0 mg cm-3 and were
filtered through tissue paper. Triple detection SEC was carried
out using a Viscotek VE1121 solvent pump (flow 1.0 cm3 min-1
)
and Rheodyne 7725i injector with 100 µL loop. Columns were
PL gel mixed C (5 µm particle size, 100-105 Å pore size,
effective MW range 102-2 × 106, 2 × 30 cm + guard column)
(Polymer Laboratories, UK) thermostated at 35 °C. Detectors
were Viscotek model 250 dual refractometer/viscometer plus
Viscotek RALLS detector. Data acquisition and analysis using
Polymer Laboratories multidetector software. Eluant was THF
(0.1% w/v tetrabutylammonium bromide). For triple detection
SEC ca. 10 mg of each polymer was accurately weighed and
dissolved in the THF (0.1% TBAB) eluant at ca. 2 mg/mL and
then filtered through tissue paper. 1H NMR (250 and 400 MHz)
and 13C NMR (62.5 and 100 MHz) spectra were recorded using
Bruker AC-250 and AMX2-400 instruments. CHN analysis
was performed on a Perkin-Elmer 2400 CHNS/0 series II
elemental analyzer. LC-MS analysis was carried out using a
Waters 2695 HPLC system coupled to a Waters LCT (TOF)
mass spectrometry detection system.

4602 Carter et al.
Macromolecules, Vol. 38, No. 11, 2005
an orange-brown solid. An initial purification was carried out
using flash column chromatography on silica (5:4:1 40-60
petroleum ether/dichloromethane/methanol) to give, after
removal of solvent in vacuo, an orange-brown solid (ca. 13 g).
A second purification was carried out via column chromatog-
raphy on alumina (50:46:4 40-60 petroleum ether/dichoro-
methane/methanol) to give after removal of solvent the title
compound as a foul smelling bright orange solid (5.50 g,
22.6%). Rf (alumina, 50:46:4 40-60 petroleum ether/dichloro-
methane/methanol) 0.31. 1H NMR (CDCl3, ca. 5% CD3OD, RT,
250 MHz): δ/ppm 4.52 (s, 2H, -S-CH2-), 7.21 (m, 5H,
-C6H5-), 7.57 (s, 1H, CH-imidazole), 7.74 (s, 1H, CH-imid-
azole). LC-MS (TOF, ES+): 12.67 min [50% AcN (0.1% formic
acid)/50% H2O (0.1% formic acid)-95% AcN, 20 min], 234 (M+),
235 (MH+) [desolvation gas 462 l h-1, capillary 3129 V, sample
cone 36 V, extraction cone 3 V, desolvation temperature 150
°C, source temperature 100 °C]. Calculated for C11H10N2S2: C,
56.38; H, 4.30; N, 11.95; S, 27.37. Found: C, 55.89; H, 4.04;
N, 11.94; S, 27.43.
Synthesis of Highly Branched and Nonbranched Poly-
mers. Polymerization of NIPAM with 1 or 3. N-Isopropyl-
acrylamide (3.514 g, 30.9 mmol) was dissolved in dioxane (11
mL) and added to the required amount of 1 or 3 to give an
orange solution. The initiator AIBN (0.111 g, 0.68 mmol) was
dissolved in the solution, which was then transferred to a 25
mL glass ampule. The ampule was freeze-pump-thawed on
a vacuum line (10-4 mbar, three cycles), flame-sealed, and
heated in a thermostated water bath at 60 °C for 24 h. The
polymer was purified by reprecipitation as follows. The
polymer-dioxane solution was added dropwise to rapidly
stirring ether (500 mL) at room temperature over 20 min. The
ether was decanted from the solids, which were then vacuum-
oven dried at RT for 16 h and redissolved in dioxane (11 mL),
and the reprecipitation/drying procedure was repeated twice
more to give the polymer. The solids were dried (vacuum oven,
RT/16 h) to give a yellow-orange solid and finally ultrafiltered.
Polymerization of NIPAM with 2. N-Isopropylacrylamide
(1.00 g, 8.83 mmol) was dissolved in dioxane/DMF mixture (3.4
mL) under a nitrogen atmosphere, and an appropriate amount
of 2 was dissolved in the solution. The initiator AIBN was then
added, and full dissolution was effected by gentle shaking. The
solution was transferred to an ampule and freeze-pump-
thawed (10-4 mbar, three cycles); the ampule was flame-sealed
and then transferred to a water bath at 60 °C and heated for
24 h. The polymer solution was diluted with DMF (2.0 mL)
and then precipitated by adding dropwise to rapidly stirring
diethyl ether (80 mL). The ether was decanted off, and the
solids were washed with ether (3 × 20 mL) and then dried in
a vacuum oven at room temperature (16 h). The reprecipita-
tion/washing drying procedure was twice repeated, and then
the polymer was ultrafiltered.
Cloud Point Determination for Highly Branched and
Linear Polymers. The cloud points of the branched polymers
from the polymerization of NIPAM with 1 and of the linear
polymers from the polymerization of NIPAM with 2 were
determined by eye using a stirrer hot plate/temperature probe.
Each polymer (0.010 g) was dissolved in distilled/deionized
water (1.0 mL) at 0-1 °C (ice bath) and then immersed in a
magnetically stirred ice bath (100 mL) which was slowly
warmed from 0 °C to beyond 35 °C over ca. 15 min. The cloud
point range was determined as the temperature interval at
which the solution became cloudy and before which any
significant agglomerization occurred. Upon recooling to 0-1
°C (ice bath) a 100 µL aliquot of aqueous 0.1 M copper(II)
sulfate was added and the cloud point recorded by reimmersing
in a magnetically stirred ice bath and warming beyond 35 °C.
Purification of Highly Branched and Linear Polymers
by Ultrafiltration. The polymer (500 mg) was dissolved in
10% ethanol/acetone (300 mL) and concentrated by ultrafil-
tration through a 10 000 MWCO cellulose filter in a 350 mL
Millipore filtration unit at 4 atm nitrogen pressure. The
procedure was carried out over ca. 45 min to give a final
volume of ≈50 mL. The procedure was twice repeated by the
addition of more solvent (≈300 mL), and the solvent was
removed from the concentrate by rotary evaporation at 40 °C.
To allow for ease of handling, the solids were redissolved in
DMF (10 mL) and then reprecipitated into diethyl ether (400
mL). The ether was decanted off, the solids were washed with
ether (3 × 20 mL), and then the polymer was dried in a
vacuum oven at RT for 16 h. Final recoveries of polymer were
in the range 80-90%.
Supporting Information Available: MALDI TOF mass
spectrum of a PNIPAM produced in the presence of 2. This
material is available free of charge via the Internet at http://
pubs.acs.org.
References and Notes
(1) Magnusson, H.; Malmstro¨m, E.; Hult, A.; Johansson, M.
Polymer 2002, 43, 301-306.
(2) Polymer 2001, 42, 1931-1939.
(3) Karger-Kocsis, J.; Fro¨hlich, J.; Gryshchuk, O.; Kautz, H.;
Frey, H.; Mu¨lhaupt, R. Polymer 2004, 45, 1185-1195.
(4) Johansson, M.; Glauser, T.; Jansson, A.; Hult, A.; Malmstro¨m,
E. Prog. Org. Coat. 2003, 48, 194-200.
(5) Claesson, H.; Jikei, M.; Kakimoto, M. Prog. Polym. Sci. 2001,
26, 1233-1285.
(6) Coessens, V.; Pintauer, T.; Matyjaszewski, K. Prog. Polym.
Sci. 2001, 26, 337-377.
(7) Yamada, B.; Konosu, O.; Tanaka, K.; Oku, F. Polymer 2000,
41, 5625-5631.
(8) Chiefari, J.; Chang, Y. K.; Ercole, F.; Krstina, J.; Jeffery, J.;
Le, T. P. T.; Mayadunne, R. T. A.; Meijs, G. F.; Moad, C. L.;
Moad, G.; Rizzardo, E.; Thang, S. H. Macromolecules 1998,
31, 5559-5562.
(9) Maydunne, R. T. A.; Rizzardo, E.; Chiefari, J.; Chang, Y. K.;
Moad, G.; Thang, S. H. Macromolecules 1999, 32, 6977-6980.
(10) Chang, Y. K.; Le, T. P. T.; Moad, G.; Rizzardo, E.; Thang, S.
H. Macromolecules 1999, 32, 2071-2074.
(11) Mayadunne, R. T. A.; Jeffery, J.; Moad, G.; Rizzardo, E.
Macromolecules 2003, 36, 1505-1513.
(12) Moad, G.; Mayadunne, R. T. A.; Rizzardo, E.; Skidmore, M.;
Thang, S. H. Macromol. Symp. 2003, 192, 1-12.
(13) Wang, Z.; He, J.; Tao, Y.; Yang, L.; Jiang, H.; Yang, Y.
Macromolecules 2003, 36, 7446-7452.
(14) Liu, B.; Kazlauciunas, A.; Guthrie, J. T.; Perrier, S. Macro-
molecules 2005, 38, 2131-2136.
(15) Hadjout, S.; Lefesque, G.; Pham, T.-N.; Tran, H.-N. Polymer
1997, 38, 3691-3696.
(16) Kumar, A.; Galaev, I. Yu.; Mattiasson, B. Biotechnol. Bioeng.
1998, 59, 695-704.
(17) Kumar, A.; Galaev, I. Yu.; Mattiasson, B. Bioseparation 1999,
7, 185-194.
(18) Kumar, A.; Galaev, I. Yu.; Mattiasson, B. Bioseparation 1998,
7, 129-136.
(19) Kumar, A.; Khalil, A. A. M.; Galaev, I. Yu.; Mattiasson, B.
Enzymol. Microbol. Technol. 2003, 33, 113-117.
(20) Kumar, A.; Kamihira, M.; Galaev, I. Yu.; Mattiasson, B.
Langmuir 2003, 19, 865-871.
(21) . Gao, C.; Yan, D. Prog. Polym. Sci. 2004, 29, 183-275.
(22) Jeong, B.; Choi, Y. K.; Bae, Y. H.; Zentner, G.; Kim, S. W. J.
Controlled Release 1999, 62, 109-114.
(23) Stayton, P. S.; Hoffman, A. S.; Murthy, N.; Lackey, C.;
Cheung, C.; Tan, P.; Klumb, L. A.; Chilkoti, A.; Wilbur, F.
S.; Press, O. W. J. Controlled Release 2000, 65, 203-220.
(24) Schild, H. G. Prog. Polym. Sci. 1992, 17, 163-249.
(25) Kubota, K.; Fujishige, S.; Ando, I. Polym. J. 1990, 22, 15-
20.
(26) Kubota, K.; Hamano, K.; Kuwahara, N.; Fujishige, S.; Ando,
I. Polym. J. 1990, 22, 1051-1057.
(27) Ohta, H.; Ando, I.; Fujishige, S.; Kubota, K. J. Polym. Sci.,
Part B: Polym. Phys. 1991, 29, 963-968.
(28) Ohta, H.; Ando, I.; Fujishige, S.; Kubota, K. J. Mol. Struct.
1991, 245, 391-397.
(29) Terada, T.; Inaba, T.; Kitano, H.; Maeda, Y.; Tsukida, N.
Macromol. Chem. Phys. 1994, 195, 3261-3270.
(30) Zeng, F.; Tong, Z.; Feng, H. Polymer 1997, 38, 5539-5544.
(31) Chee, C. K.; Rimmer, S.; Shaw, D. A.; Soutar, I.; Swanson,
L. Macromolecules 2001, 34, 7544-7549.
(32) Barker, I. C.; Cowie, J. M. G.; Huckerby, T. N.; Shaw, D. A.;
Soutar, I.; Swanson, L. Macromolecules 2003, 36, 7765-
7770.

Macromolecules, Vol. 38, No. 11, 2005
Highly Branched Poly(N-isopropylacrylamide)s 4603
(33) Yamamoto, I.; Iwasaki, K.; Hirotsu, S. J. Phys. Soc. Jpn.
1989, 58, 210-215.
(36) Mu¨ller, A. H. E.; Yan, D.; Wulkow, M. Macromolecules 1997,
23, 7015-7023.
(37) Chernikova, E.; Morozov, A.; Leonova, E.; Garina, El.; Gol-
ubev, V.; Bui, C.; Charleux, B. Macromolecules 2004, 37,
6329-6339.
(34) Ganachaud, F.; Monteiro, M. J.; Gilbert, R. G.; Dourges, M.-
A.; Thang, S. H.; Rizzardo, E. Macromolecules 2000, 33,
6738-6745.
(35) Schilli, C.; Lanzendorfer, M. G.; Muller, A. H. E. Macromol-
ecules 2002, 35, 6819-6827.
MA047742N