Synthesis of poly(2-methyl-2-oxazoline) star polymers with a β-cyclodextrin core

Article abstract of DOI:10.1071/CH12232

Synthesis of star polymers with a-cyclodextrin (CD) core was undertaken using the arm-first, then the core-first strategy. Cationic ring opening polymerisation (CROP) of 2-methyl-2-oxazoline (MeOx) was first initiated by allyl bromide, and then quenched with heptakis(6-deoxy-6-amino)-CD in order to get a 7-arm star polymer. Then heptakis(6-deoxy-6-iodo-2,3-di-O-acetyl)-CD was synthesised in order to get an initiator for the CROP of MeOx. Initiation and propagation kinetic measurements were undertaken and the ratio kp/ki was found to be too high to provide a controlled polymerisation. Using iodine as co-initiator allowed a decrease of the kp/ki ratio that gave better control of the polymerisation. DOSY NMR and viscosity characterisations were undertaken, and both techniques lead to the demonstration of a lower hydrodynamic volume of the star polymers versus the linear counterparts, for compounds of the same molecular weight.

Full text of DOI:10.1071/CH12232

RESEARCH FRONT
CSIRO PUBLISHING
Aust. J. Chem. 2012, 65, 1145–1155
Full Paper
http://dx.doi.org/10.1071/CH12232
Synthesis of Poly(2-methyl-2-oxazoline) Star Polymers
with a b-Cyclodextrin Core
A
A
B
Guillaume Pereira, Ce´cile Huin, Simona Morariu,
A C
,
A C
,
Ve´ronique Bennevault-Celton,
and Philippe Gue´gan
A
LAMBE, UMR8587, University of Evry, 91025 Evry Cedex, France.
B
‘Petru Poni’ Institute of Macromolecular Chemistry, 700487 Iasi, Romania.
C
Corresponding authors. Email: veronique.celton@univ-evry.fr;
philippe.guegan@univ-evry.fr
Synthesis of star polymers with a b-cyclodextrin (CD) core was undertaken using the arm-first, then the core-first strategy.
Cationic ring opening polymerisation (CROP) of 2-methyl-2-oxazoline (MeOx) was first initiated by allyl bromide, and
then quenched with heptakis(6-deoxy-6-amino)b-CD in order to get a 7-arm star polymer. Then heptakis(6-deoxy-6-iodo-
2,3-di-O-acetyl)b-CD was synthesised in order to get an initiator for the CROP of MeOx. Initiation and propagation
kinetic measurements were undertaken and the ratio kp/ki was found to be too high to provide a controlled polymerisation.
Using iodine as co-initiator allowed a decrease of the kp/ki ratio that gave better control of the polymerisation. DOSY NMR
and viscosity characterisations were undertaken, and both techniques lead to the demonstration of a lower hydrodynamic
volume of the star polymers versus the linear counterparts, for compounds of the same molecular weight.
Manuscript received: 9 May 2012.
Manuscript accepted: 22 June 2012.
Published online: 24 July 2012.
TEMPO-mediated living radical,[12] reversible addition-
fragmentation chain transfer (RAFT),[13] and ring opening
polymerisation[14–16] using numerous monomers have been
investigated in order to synthesise well defined star polymers by
the core-first method. The drawback of ‘the arm-first method’ is
the necessity to use an excess of polymer chains in comparison
with the number of linking CD functions in order to ascertain a
quantitative functionalisation, requiring a further purification
step to eliminate the excess of linear polymer. Moreover, the
linking yield depends on the molecular weight of the linear
polymer. Indeed, low linking efficiency was obtained when
using high molecular weight polymer chains. In spite of these
difficulties, different examples of star polymer syntheses using
this method have been described in the literature.[17–19]
Star polymers with a cyclodextrin core and poly(2-
oxazoline) arms were recently reported using both synthesis
strategies[16,19] for their various potential applications.[20] In the
arm-first method, a two-step procedure using click chemistry
was used, while in the core-first method, a CD derivative with
tosyl functions was synthesised to subsequently initiate the
polymerisation of the 2-methyl-2-oxazoline (MeOx). Star poly-
mers with PMeOx arms were already reported with other core
molecules[21–23] and in some cases, steric hindrance was found
to decrease the kinetics of the initiation step, leading to ill
defined polymers as suggested by Hoogenboom et al.[23]
In this paper, we focus on the synthesis of 7-arm star
polymers with poly(2-methyl-2-oxazoline) (PMeOx) arms, a
good precursor of vectors for gene therapy applications,[24] and a
b-CD core through both strategies in a one-step procedure. The
Introduction
Development of star polymers with controlled structures in
terms of molecular weight and functionality presents a great
interest due to their attractive rheological properties.[1,2] The
main advantage of a high molecular weight star architecture is
that it does not significantly affect the viscosity of the solution,
compared with the corresponding linear polymer.[3] Star
polymers also have lower melting points, improved mechanical
properties, and are more processable.[4] Star polymers can be
described as structures in which several arms radiate from a low
molecular weight core. Cyclodextrins were widely used as core
structures[5] because of their biocompatibility, biodegradability,
and their ability to form inclusion complexes with many
hydrophobic compounds.[6] Indeed, cyclodextrins (CDs) which
are a series of cyclic oligosaccharides composed of six, seven, or
eight glucopyranose units (respectively named a-, b-, or g-CD)
linked by a-1,4-linkages possess a structure of truncated conical
shape with a non-polar cavity. These molecules have secondary
hydroxyl groups (12 to 16) and primary hydroxyl groups (6 to 8),
which can be selectively modified.[7,8] Therefore, they are
excellent candidates to synthesise star polymers with discrete
numbers and various natures of the arms. Two different strate-
gies were used to synthesise these products. The first one, named
‘the arm-first method’, consists of linking preformed living
polymer chains to a CD core, while the second method, named
‘the core-first method’, consists of initiating the polymerisation
from the initial CD hydroxyl functions or from the new functions
generated on the CD. Radical polymerisation methods such
as atom transfer radical polymerisation (ATRP),[5,9–11]
Journal compilation Ó CSIRO 2012
www.publish.csiro.au/journals/ajc

1146
G. Pereira et al.
N
CH2 CH2
N
Br
n
ꢀ
n
Br
H3C
O
H3C
O
Scheme 1.
1
3
5
6
8
9
synthetic pathway allows for the modification of only the
primary rim of the cyclodextrins. The arm-first strategy is based
on the coupling reaction of the bromoalkyl end-chains of living
PMeOx and amino groups borne by the heptakis(6-deoxy-6-
amino)b-CD (NH2-b-CD), while the core-first strategy consists
of polymerisation of 2-methyl-2-oxazoline initiated with
N
CH2 CH2
N
CH2 CH2 Br
n
2
4
7
H3C
5; 6
4; 7
O
O
H3C
3
8
9
H2O
1
2
heptakis(6-deoxy-6-iodo-2,3-di-O-acetyl)b-CD
(I-OAc-b-
ppm
CD). Kinetic measurements of the initiation and propagation
reactions are performed and compared with monofunctional
compounds to shed more light on the final architecture of the
polymers. The influence of initiation kinetics on the structure of
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5
6.0
the final compound was demonstrated. 1H and quantitative 13
C
NMR and 2D NMR were used to give information on the
structure of the star polymers. The behaviour of the polymers
in solution was then investigated by measuring the diffusion
coefficients of the polymers by DOSY NMR and by performing
viscosity experiments.
Results and Discussion
Synthesis of Star PMeOx Using the Arm-First Method
The arm-first method requires the use of preformed polymer
with controlled structure, achieved for the polymerisation of 2-
oxazoline.[25,26] Initial polymerisation was performed by initi-
ating the polymerisation of MeOx (3.0 mol Lꢀ1) with allyl
bromide (0.2 mol Lꢀ1) in acetonitrile at 808C (Scheme 1). After
24 h of reaction, no quench was performed, in order to observe
the bromide end-chains of the polymer. The COSY NMR
spectrum of the polymer obtained after precipitation from die-
thyl ether is shown in Fig. 1. In addition to the methylene and the
methyl protons of the polymer units at 3.4 and 2.1 ppm,
respectively, the protons belonging to the allyl end-chains
appeared at 5.7, 5.2, and 3.8 ppm and the methylene protons
witnessed to the covalent bromide end-chains at 4.4 and
3.3 ppm. These assignments were confirmed by the intensity of
the peaks. It can also be observed at ,4.7 ppm a very low
fraction of remaining ionic bromide species (,5 %). Based on
the polymer units and the intensity of the allyl group, the average
number degree of polymerisation was determined and found
equal to 15.2, close to the theoretical value (15.0). The intensity
of the allyl end-chains was in agreement with the intensity of the
bromide end-chains (ionic þ covalent), showing the control of
the polymerisation.
This polymerisation was then repeated with a low monomer/
initiator ratio ([MeOx]/[allyl bromide] ¼ 5.0) to facilitate the
star polymer characterisation. A part of the living chain ends was
quenched by a heptakis(6-deoxy-6-amino)b-CD solution in
DMSO (Scheme 2), the other part being not quenched and
characterised by NMR to determine the molar mass of the linear
chains. The polymerisation medium with the added NH2-b-CD
solution was stirred at 808C for 4 days in order to obtain high
linking yield. The polymer chains were in slight excess in
comparison with the amino functions of the CD ([PMeOx]/
[NH2] ¼ 1.1) to obtain a star polymer with seven arms. At the
end of the reaction, the star polymer was precipitated in Et2O
and collected by filtration, then dried. The excess of linear
6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0
ppm
Fig. 1. COSY NMR of a linear PMeOx after precipitation in Et2O (CDCl3,
258C).
PMeOx was removed by dialysis and the purified star PMeOx
was then analysed by NMR. First, the 1H NMR spectrum
showed that the anomeric proton was overlapped with the
unsaturated methylene protons of the allylic moieties at
,5.1 ppm, the intensity of these signals being 3 if we consider
an intensity of 1 for the other unsaturated proton at 5.7 ppm
(Fig. 2). This result suggests that the star polymer was composed
of 7 arms. The 1H NMR analysis of the linear chains gave a DPn
equal to 5.6, in agreement with the theoretical value, leading to a
star PMeOx with a Mn close to 4700 g molꢀ1. The 13C NMR of
the star polymer (Figure S1 in the Supplementary Material)
showed that the C6 peaks of the NH2-b-CD (Figure S2 in the
Supplementary Material) totally disappeared at 42.9 ppm while
the signals attributed to C1 to C5 still appeared. C6 of the star
polymer might shift under the peaks of the methylene groups of
the polymer. Moreover, obtaining a sharp single peak for the
anomeric carbon suggested that multiple additions of cationic
chains to one amine group did not occur, but only one reaction
on each glucopyranose unit as described in Scheme 2.
In order to demonstrate the formation of the star structure,
diffusion-ordered spectroscopy (DOSY) experiments were per-
formed on the linear and star PMeOx and on the cyclodextrin
derivative used for quenching. The DOSY method invented by
Morris and Johnson[27] allows measurement of translational
self-diffusion of molecules in solution and was often used to
analyse mixtures or to characterise the structure of the mole-
cules.[28–30] The different diffusion coefficients (D) determined

Star Polymers with Cyclodextrin Core and Poly(2-methyl-2-oxazoline) Arms
1147
NH CH2 CH2
O
N
C
CH2
C
H
CH2
NH2
O
n
H3C
O
CH2 CH2
N
Br
ꢀ
n
HO
O
O
HO
H3C
OH
O
7
7
OH
1.1 eq/NH2
Scheme 2.
7
8
10 11 12
Table 1. Molecule diffusion coefficients determined in D2O at 258C
6
NH CH2 CH2
N
C
CH2
C
H
CH2
n
4
O
O
H C
Run
Structure
D ꢂ 1010 [m2 sꢀ1
]
Molar mass [g molꢀ1
]
9 3
2
5
O
HO
OH
3
7
1
2
3
NH2-b-CD
Linear PMeOx
Star PMeOx
2.45 ꢃ 0.01
3.27 ꢃ 0.10
1.86 ꢃ 0.10
1128
600
1
9
4740
D2O
DMSO
log(m2 sꢁ1
)
2; 3; 4; 5; 6; 7; 8; 10
1; 12
11
ꢁ10.0
ꢁ9.5
ꢁ9.0
ꢁ8.5
6.0
5.5
5.0
4.5
4.0
3.5
3.0
2.5
2.0
Fig. 2. 1H NMR of the star PMeOx in D2O.
are reported in Table 1. All the experiments were performed in
D2O at 258C at a concentration lower than the overlap concen-
trations of the polymers. These results clearly showed that the
linear polymer chains had a lower hydrodynamic volume than
the star polymer and the NH2-b-CD presented an intermediate
hydrodynamic volume, in agreement with their respective molar
masses. It means that, in the case of run 3, the linear PMeOx
chains were linked to the CD core to form a star structure. Fig. 3
shows the DOSY NMR of this star polymer, one dimension
accounting for conventional chemical shifts and the other for
diffusion coefficients. This figure underlines that all the proton
signals diffused at the same rate, proving that all arms were
grafted on the CD core and the synthesis of a well defined 7-arm
star polymer. To conclude, the arm-first method allowed the
synthesis of star PMeOx with controlled structure.
6.0
5.5
5.0
4.5
4.0
3.5
3.0
2.5
2.0
1.5
ppm
Fig. 3. DOSY NMR of the star PMeOx (after dialysis) at 258C in D2O.
by the hydroxyl functions. Kinetics were first carried out in
order to determine the initiation (ki) and propagation (kp) con-
stants using I-OAc-b-CD as initiator. The polymerisations were
conducted at 808C in acetonitrile and aliquots were taken versus
time. The time conversion curve of the MeOx polymerisation
showed that the polymerisation was very slow: a yield of 86 %
needed 90 h of reaction time (Fig. 4).
The kinetic values were determined considering the follow-
ing equations:
For the initiation step:
Synthesis of Star PMeOx Using the Core-First Method
2-Alkyl-2-oxazoline monomers can be polymerised using alkyl
iodide as initiator and control of the polymerisation was
described in the literature.[31–33] Based on these results, the
polymerisation of MeOx initiated by the heptakis(6-deoxy-6-
iodo-2,3-di-O-acetyl)b-CD (Scheme 3, Figure S3 in the
Supplementary Material) was investigated. The iodine deriva-
tives allow the formation of ionic species and the hydroxyl
functions of the secondary rim were protected because the cat-
ionic polymerisation is usually sensitive to nucleophilic attack.
It has to be mentioned that Kabiri and colleagues[16] reported the
polymerisation of 2-ethyl-2-oxazoline initiated by heptakis(6-
deoxy-6-tosyl)b-CD, without looking for side reactions induced
ꢀd½CH2Iꢁ=dt ¼ ki½CH2Iꢁ½Mꢁ
ð1Þ
where [CH2I], [M] respectively represent the concentrations of
CH2I group and monomer at time t.
The integration of this equation gives:
t
Z
ln ½CH2Iꢁ0=½CH2Iꢁ ¼ ki ½Mꢁdt
ð2Þ
0

1148
G. Pereira et al.
OH
I
I
Ac2O, pyridine
72 h, 20°C
PPh3, I2
O
O
O
HO
HO
AcO
O
O
O
DMF, 70°C
18 h
7
7
OH
OH
OAc
7
Scheme 3.
90
80
70
60
50
40
30
20
10
0
CDCl3
a
b
ꢁ
I
N
CH2 CH
2 n
N
CH3 of OAc;
c
ꢀ
d
O
H3C
H3C
O
c
CꢂO of OAc;
d
a; b
C1
C4
C6
160
140
120
100
80
60
40
20 ppm
0
20
40
60
80
100
Fig. 5. Quantitative 13C NMR of the PMeOx initiated by I-OAc-b-CD in
CDCl3.
Time [h]
Fig. 4. Time-conversion curve of the MeOx polymerisation initiated by
I-OAc-b-CD at 808C in acetonitrile; [MeOx] ¼ 3.28 mol Lꢀ1, [I-OAc-b-
0.40
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0
CD] ¼ 9.20 ꢂ 10ꢀ3 mol Lꢀ1
.
Assuming that kp1 (1st propagation step) is equal to kpn (nth
propagation step), the kinetic equation for the propagation is:
ꢀd½Mꢁ=dt ¼ ki½CH2Iꢁ½Mꢁ þ kp½Pꢄꢁ½Mꢁ
ð3Þ
where [P*] represents the concentration of active centres at t
and kp1 ¼ kpn ¼ kp. Then, considering that [P*] is equal to
[CH2I] - [CH2I]0 implies:
0
10
20
30
40
t
[M]dt
0
ln ½Mꢁ0=½Mꢁ
½CH2Iꢁ0t
t
¼ ðki ꢀ kpÞ þ kp
ð4Þ
t
Fig. 6. Plot of Eqn 2 in the MeOx polymerisation initiated by I-OAc-b-CD
at 808C in acetonitrile; [MeOx] ¼ 3.28 mol Lꢀ1
R
R
½CH2Iꢁdt
½CH2Iꢁdt
,
[I-OAc-b-CD] ¼
9.20 ꢂ 10ꢀ3 mol Lꢀ1
.
0
0
where [CH2I]0 is the initial concentration of CH2I group.
0.21
0.18
0.15
0.12
0.09
0.06
0.03
0
Eqns 2 and 4 were used to graphically determine the kinetic
constants ki and kp. The concentration of CH2I (C6 of the CD)
was determined by quantitative 13C NMR (Fig. 5), thus only the
samples at low monomer conversion were considered, i.e. for
reaction time lower than 11 h. The integrated values of [M] and
[CH2I] were then obtained respectively by graphical integration
of the [M]-time and [CH2I]-time curves (respectively Figures S4
and S5 in the Supplementary Material). The linear plot of
Eqn 2 is given in Fig. 6, whose slope corresponds to ki (ki ¼
2.5 ꢂ 10ꢀ6 L molꢀ1 sꢀ1). The kp value was determined from the
slope of Fig. 7 (kP ¼ 2.0 ꢂ 10ꢀ4 L molꢀ1 sꢀ1), but could be also
deduced from the intercept with the y-axis. A value of
1.9 ꢂ 10ꢀ4 L molꢀ1 sꢀ1 was obtained in this case, which is very
close to the previous value. Therefore, these results showed that
kp is much higher than ki with a magnitude order of ,100, which
is obviously unfavourable for control of polymerisation.
1.05
1.10
1.15
1.20
[CH2I]0 ꢃ t
t
[CH2I]dt
0
Fig. 7. Plot of Eqn 4 in the MeOx polymerisation initiated by I-OAc-b-CD
at 808C in acetonitrile; [MeOx] ¼ 3.28 mol/L, [I-OAc-b-CD] ¼ 9.20 ꢂ
10ꢀ3 mol/L.
In order to understand the kinetic results, obtained with the
CD-based initiator, a model polymerisation was investigated,

Star Polymers with Cyclodextrin Core and Poly(2-methyl-2-oxazoline) Arms
1149
ꢁ
CH3
H3C
I
CH3
H3C
N
CH
CH2
I
n
ꢀ
N
CH
N
C
CH2 CH2
CH2
ꢀ
n
H3C
O
H3C
H3C
O
O
Scheme 4.
100
90
80
70
60
50
40
30
20
10
0
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0
I-OAc-β-CD (I2 ꢂ 0%)
I-OAc-β-CD (I2 ꢂ 14%)
RI (I2 ꢂ 0%)
RI (I2 ꢂ 20%)
0.6
0.8
1.0
1.2
1.4
RI (I2 ꢂ 20%)
[CH2I]0 ꢃ t
RI (I2 ꢂ 100%)
t
[CH2I]dt
0
0
2
4
6
8
10
12
Time [h]
Fig. 10. Plot of Eqn 4 in the MeOx polymerisation initiated by RI at 808C
in acetonitrile; [MeOx] ¼ 3.28 mol Lꢀ1, [RI] ¼ 6.44 ꢂ 10ꢀ2 mol Lꢀ1
.
Fig. 8. Yield versus polymerisation time plot for MeOx polymerisation
initiated by I-OAc-b-CD (square) or by RI (triangle) at 808C in acetonitrile
with or without I2; [MeOx] ¼ 3.28 mol Lꢀ1, [CH2I] ¼ 6.44 ꢂ 10ꢀ2 mol Lꢀ1
.
The idea of activation of the iodo group by addition of iodine
was first investigated when using 1-iodo-2-methylpropane as
initiator. Different iodine amounts were examined: 20 and
100 % of the CH2I content. I2 was introduced into the polymeri-
sation medium at the same time as the solvent, monomer, and
initiator and the activation occurred during heating of the
solution until 808C and during the polymerisation (runs 4 and 6),
except for run 5 for which the solution was stirred at room
temperature for one night before the heating step. The corre-
sponding yield versus time plots and kinetic constant values are
given in Fig. 8 and Table 2. It appeared that the polymerisation
was accelerated in the presence of iodine and no major differ-
ence was observed between the two times of activation. The
polymerisation was so fast with 100 % of I2 that no kinetic
constant determination was possible due to the lack of data
points. The use of this co-initiator increased at the same time the
rates of the initiation and propagation witha factor equal to E1.3
for ki and E2.3 for kp for the experiments performed with 20 %
of I2. Hence, the kp/ki ratio could not be decreased when using
this iodine content. Considering these results, another percent-
age of iodine was investigated in the case of the polymerisation
initiated by I-OAc-b-CD. Indeed, it was published that the
triiodide ion can form an inclusion complex with the
b-CD.[34] So, if a molar ratio 1/1 between iodine and I-OAc-
b-CD, which corresponds to an iodine percentage of 14 % with
regards to the iodo functions, was used, it might be expected that
the initiation was preferentially activated in comparison with
the propagation. As can be seen in Table 2, the kinetic
study respectively gave, for ki and kp, 4.0 ꢂ 10ꢀ5 and 3.1 ꢂ
10ꢀ4 L molꢀ1 sꢀ1, giving a ratio kp/ki equal to 7.5. Therefore, in
presence of iodine the polymerisation of MeOx was accelerated
(Fig. 8) and the ratio kp/ki could be tremendously decreased. It
means that these experimental conditions strongly favoured the
initiation step. Nevertheless, the propagation step remained the
fastest step.
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0
9
11
13
t
15
17
[M]dt
0
Fig. 9. Plot of Eqn 2 in the MeOx polymerisation initiated by RI at 808C in
acetonitrile; [MeOx] ¼ 3.28 mol Lꢀ1, [RI] ¼ 6.44 ꢂ 10ꢀ2 mol Lꢀ1
.
using 1-iodo-2-methylpropane (named RI) as initiator
(Scheme 4). This iodine derivative has an aliphatic chemical
structure close to the one of I-OAc-b-CD without the steric
hindrance. The remaining initiator was determined from the
1H NMR spectrum of the cryodistilled fraction of the sample
(Figure S6 in the Supplementary Material), due to its low boiling
point. As shown by Fig. 8, the polymerisation initiated by the
monofunctional initiator occurred much more quickly than in
the case of I-OAc-b-CD. The corresponding kinetic constants
using Eqns 2 and 4 (Figs 9 and 10) are reported in Table 2. The ki
value obtained in this experiment (run 3) is ,20 times higher
than the ki value of the previous polymerisation (run 1),
demonstrating a faster initiation step, that could be attributed
at first glance to an easier access of the active centres, compared
with the iodomethyl moieties of the I-OAc-b-CD. However, ki
still remained lower than kp.
Fig. 11 shows the increase of the CH2I conversion induced by
the addition of iodine (14 % of iodo functions of I-OAc-b-CD)

1150
G. Pereira et al.
Table 2. Determination of the kinetic constants
Run
InitiatorA
I2 [%]
ki ꢂ 105 [L molꢀ1 sꢀ1
]
kp ꢂ 103 [L molꢀ1 sꢀ1
]
kp/ki
1
I-AcO-b-CD
I-AcO-b-CD
0
14
0
0.25
4.0
4.6
5.8
6.1
nd
0.2
0.3
1.6
3.9
3.6
nd
79
7.5
34
67
60
nd
2
3
RI
RI
RI
RI
4
5B
20
20
100
6
AI-AcO-b-CD and RI respectively correspond to heptakis(6-deoxy-6-iodo-2,3-di-O-acetyl)b-CD and 1-iodo-2-
methylpropane.
BRun 5: activation by I2 for one night.
45
40
35
30
25
20
15
10
5
35000
I-OAc-β-CD (I2ꢂ14%)
MnNMR
Mnth
I-OAc-β-CD (I2ꢂ0%)
30000
25000
20000
15000
10000
5000
0
0
20
40
60
80
100
0
Monomer conversion [%]
0
2
4
6
8
10
Time [h]
Fig. 13. Molar mass versus monomer conversion plot for the poly-
merisations initiated by I-OAc-b-CD with 14 % of iodine in acetonitrile
at 808C; [MeOx]¼ 3.28mol Lꢀ1, [CH2I]¼ 6.44 ꢂ 10ꢀ2 mol Lꢀ1, [I2] ¼ 9.2 ꢂ
10ꢀ3 mol Lꢀ1. MnNMR and Mnth respectively represent the molar mass
Fig. 11. CH2I conversion versus time plots for polymerisations initiated
by I-OAc-b-CD with or without iodine in acetonitrile at 808C; [MeOx] ¼
3.28 mol Lꢀ1, [CH2I] ¼ 6.44 ꢂ 10ꢀ2 mol Lꢀ1
.
calculated from 1H NMR spectrum and theoretical value. Mnth ¼ MCD core
þ
([M]ꢂ yield ꢂ 7ꢂ Mpolymer unit)/[CH2I].
2.5
2.0
1.5
1.0
0.5
0
concentration[31,33]), the oxazolinium iodine moieties, and the
oxazolinium triiodine moieties. Taking into account the kinetics
results obtained without I2 and with I2, the triiodine oxazolinium
are much more reactive than both other species. Using RI or
I-OAc-b-CD, one noticed that the ki are very close (run 2 and
run 4 in Table 2) suggesting that the cyclodextrin cavity plays a
negligible role in the initiation of oxazoline using such systems.
However, the initiation kinetics of I-OAc-b-CD is strongly
enhanced when I2 is used (run 2 compared with run 1) while it
is kept grossly constant for RI (run 3 and 4). This observation can
be explained by a cooperative effect of the interaction of the
seven iodine moieties borne by the CD with one molecule of
iodine. This interpretation is supported by the demonstration by
13C NMR of a higher interaction of I-OAc-b-CD with iodine
than with RI (not presented).
0
1
2
3
4
5
6
Time [h]
Fig. 12. ln [M]0/[M] versus time plot for the polymerisation initiated by
I-OAc-b-CD with 14 % of iodine in acetonitrile at 808C; [MeOx] ¼
3.28 mol Lꢀ1, [CH2I] ¼ 6.44 ꢂ 10ꢀ2 mol Lꢀ1, [I2] ¼ 9.2 ꢂ 10ꢀ3 mol Lꢀ1
.
Concerning the molar mass of the polymers, Fig. 13 shows
that the molar mass calculated by NMR linearly increased with
the monomer conversion until a yield of ,50 %, in reasonable
agreement with the theoretical value. The low difference could
be explained by the high hygroscopicity of the polymer, leading
to overestimation of the yield when the polymer was weighed.
The intercept with the y-axis represents the molar mass of the
cyclodextrin core. After 50 % yield, the values obtained by
NMR are lower than the theoretical values. This discrepancy
might come from the inaccuracy of the NMR when analysing
high molecular weight chains. It is difficult to evaluate the
number of the arms of the polymers at the end of the
and Fig. 12 represents the plot ln [M]0/[M] versus time for this
polymerisation. Surprisingly, the polymerisation kinetics of
Fig. 12 is fitted by a linear relationship, suggesting a constant
active species concentration versus time, even at the beginning
of the polymerisation. It has to be pointed out that for the first
point on both figures, 15 % of iodoalkyl are converted into
oxazolinium moieties, and that is to say that one glucopyranose
residue per cyclodextrin initiated the MeOx polymerisation.
These results show that the chain growing results from a fast
equilibrium between the covalent species (in low

Star Polymers with Cyclodextrin Core and Poly(2-methyl-2-oxazoline) Arms
1151
Ref.
Table 3. Kinetic constants mentioned in the literature for polymerisations initiated by aryl and alkyl iodides
Initiator
Solvent
Monomer
T [8C]
ki ꢂ 105 [L molꢀ1 sꢀ1
]
kp ꢂ 103 [L molꢀ1 sꢀ1
]
kp/ki
CH3(CH2)17
CH3I
I
CH3CN
CH3CN
CH3CN
CH3CN
PhCl
MeOx
MeOx
MeOx
MeOx
EtOx
80
30
nd
9.8
1.33
–
0.5
0.4
–
32
33
33
33
31
31
31
0.047
0.114
0.285
0.262
0.788
6.33
CH3I
40
26.8
nd
CH3I
50
1-iodobutane
PhCH2I
80
0.95
32.6
374
27.6
2.4
1.7
PhCl
EtOx
80
PhCH2I
PhCl
EtOx
110
EtOx: 2-ethyl-2-oxazoline; PhCl: chlorobenzene.
polymerisation. It has to be kept in mind that this information
was obtained from quantitative 13C NMR from the peak of the
residual C6 carbon. As previously indicated, this calculation can
be done only when the yield is not too high.
Table 4. Diffusion coefficients determined in DMSO at 258C
Sample
D ꢂ 1011 [m2 sꢀ1
]
MAn [g molꢀ1
]
Structure
I-OAc-b-CD
16.3 ꢃ 0.1
10.4 ꢃ 0.03
7.5 ꢃ 0.3
2493
5200
Initiator
Star
The issue of kp/ki for the polymerisation of 2-alkyl-2-
oxazoline using aryl and alkyl iodide as initiator has already
been addressed by others (Table 3). Only the polymerisations
initiated by CH3I allowed a propagation constant lower than the
initiation one, although the temperatures studied by Saegusa
et al.[33] were different from our conditions. From the
published temperature and constant values, the activation
energy for the propagation and initiation steps can be
estimated and the corresponding constants calculated at 808C:
S0
S1
S2
S3
S4
S5
S6
S7
L1
L2
L3
6700
Star
6.5 ꢃ 0.2
8400
Star
6.1 ꢃ 0.03
5.4 ꢃ 0.08
4.9 ꢃ 0.07
4.1 ꢃ 0.06
3.0 ꢃ 0.05
8.7 ꢃ 0.1
11600
13500
16600
21700
39400
2100
Star
Star
Star
Star
Star
Linear
Linear
Linear
ki ¼ 848 ꢂ 10ꢀ5 L molꢀ1 sꢀ1
,
kP ¼ 2.85 ꢂ 10ꢀ3 L molꢀ1 sꢀ1
,
7.2 ꢃ 0.08
4100
giving a ratio kp/ki E 0.3. Volet et al.[32] conducted MeOx
polymerisation in acetonitrile at 808C using two kinds of alkyl
iodides (CH3(CH2)nI with n ¼ 11 and 17). They showed for the
larger initiator the linearity of the ln [M]/[M]0 versus time plot
and the molar mass versus conversion plot, with experimental
Mn close to theoretical values, which demonstrates the control of
polymerisation. The initiation constant was not determined, but
their kp value was close to our kp value obtained in the case of
1-iodo-2-methylpropane. Liu et al.[31] studied the 2-ethyl-2-
oxazoline polymerisation initiated by benzyl iodide and
1-iodobutane in chlorobenzene at 808C and kinetic experiments
found a ratio kp/ki equal to 2.4 for the activated aryl iodide and
27.6 for the unactivated alkyl iodide. Moreover, they mentioned
in the case of 1-iodobutane an incubation period at the beginning
of the polymerisation and a substantial remaining amount of
initiator at the end of the polymerisation. To conclude, the
initiation constants obtained in the polymerisations using
I-OAc-b-CD as initiator are smaller than any monofunctional
counterpart, which could be explained by a thermodynamic
penalty in forming multiple cationic species at the same time on
the primary face of the CD. Indeed, reacting each -CH2I with one
oxazoline leads to the formation of a cyclodextrin derivative
with a net charge of þ7, which is thermodynamically highly
unfavourable.
4.1 ꢃ 0.03
5400
ADetermined by 1H NMR spectroscopy.
polymers as well as for the initiator. The diffusion coefficients
are listed in Table 4. As it can be seen, the initiator had the
highest diffusion coefficient, which proved its smallest hydro-
dynamic volume in agreement with its molar mass. If one
compared a linear PMeOx and a star PMeOx with the same
molar mass (samples S0 and L3), the D value of the star polymer,
correlated with its hydrodynamic volume, is clearly higher than
the linear polymer as expected.[2] Considering either only the
star polymers or only the linear polymers, a decrease of the
diffusion coefficient was observed with increasing molar mass
of the product. This result is consistent with the fact that the
hydrodynamic volume increased with Mn. It can also be noted
that the same D (4.1 ꢂ 10ꢀ11 m2 sꢀ1) value was obtained for a
linear PMeOx of 5400 g molꢀ1 (sample L3) and for a star
PMeOx of 21700 g molꢀ1 (sample S6). Size exclusion chroma-
tography was also used for the determination of the molar mass
of the star samples (Figure S7 in the Supplementary Material)
and the results are reported in Table S1 in the Supplementary
Material. The Mn values in this case are only indicative because
of the polystyrene calibration and the star shape of the polymers.
However, good polydispersity indices (PDI) were reported for
the various synthesised polymers.
To determine the physico-chemical properties of these poly-
mers, a series of star PMeOx was synthesised using the previous
process, i.e. an iodine percentage of 14 % with regards to the
iodo functions of I-OAc-b-CD, as well as linear PMeOx for
comparison. The molar mass of the star polymer varied from
5200 to 39400 g molꢀ1 and the molar mass of the linear poly-
mers was between 2100 and 5400 g molꢀ1 (Table 4). These
syntheses pointed out the reproducibility of the polymerisations.
The diffusion coefficients were then determined by performing
DOSY NMR experiments. The latter were conducted in DMSO
at 258C, since this solvent was common for the linear and star
Viscometric measurements were also realised on these
polymers in chloroform, except for the samples S0 and S2,
which were obtained in insufficient amounts. The intrinsic
viscosity, ½Zꢁ, is usually evaluated from the viscosity measure-
ments at finite concentration by using the extrapolation methods
to infinite dilution. The intrinsic viscosities of the linear and star
PMeOx in chloroform at 258C were determined by using the
Huggins method which involves the extrapolation of the

1152
G. Pereira et al.
Table 5. Molecular weights and viscometric parameters for linear and
ꢁ0.8
ꢁ0.9
ꢁ1.0
ꢁ1.1
ꢁ1.2
ꢁ1.3
ꢁ1.4
ꢁ1.5
star PMeOx determined in chloroform at 258C
Sample
MAn [g molꢀ1
]
½Zꢁ [dL gꢀ1
]
c* [g dLꢀ1
]
L1
L2
L3
S1
S3
S4
S5
S6
S7
2100
4100
0.053
0.080
0.107
0.044
0.059
0.063
0.072
0.095
0.146
19.01
12.45
9.38
5400
6700
22.83
17.09
15.92
13.93
10.54
6.86
11600
13500
16600
21700
39400
Linear
Star
ADetermined by 1H NMR spectroscopy.
3.2
3.6
4.0
4.4
4.8
Log [Mn]
reduced viscosity (Zsp=c) at zero concentration according to
Eqn 5:
Fig. 14. Plots of the Mark-Houwink relationships for linear and star
PMeOx in chloroform at 258C.
Zsp
2
¼ ½Zꢁ þ kH ꢅ ½Zꢁ ꢅ c
ð5Þ
c
is the concentration which separates the dilute and the semi-
dilute regimes. In concentrated domains the degree of coil
overlapping are high and the macromolecules lose their indi-
viduality due to the formation of an infinitely large network of
overlapping molecules. The overlap concentration (c*) is
defined as the ratio between the mass of the macromolecule
where kH represents the Huggins constant which offers infor-
mation about the polymer interactions and the solvent quality
and c is the polymer concentration. Zsp represents the specific
viscosity (Zrel ꢀ 1), which reflects the fractional change in
viscosity produced by the addition of the solute and Zrel is the
relative viscosity showing the change in viscosity, usually
expressed as a ratio between the viscosity of the polymer
solution and the viscosity of the solvent. The Zsp=c versus c
dependencies for the linear and star PMeOx in chloroform at
258C are shown in Figure S8 in the Supplementary Material. The
values of ½Zꢁ obtained using the Huggins relationship (Eqn 5) for
all investigated polymers in the dilute region (for which
1.15 , Zrel ,1.9.) are given in Table 5.
and the volume that it occupies in solution[36,37]
:
M
cn ¼
ð8Þ
NA ꢅ VM
where M and VM are the molar mass of the particle and its
volume, respectively, and NA represents Avogadro’s number.
The volume can be estimated from the following equation
which defines the intrinsic viscosity:
The representation of the experimental intrinsic viscosities as
a function of the polymer molecular weight (Fig. 14) allowed
determination of the Mark-Houwink equations (½Zꢁ ¼ K ꢅ Man)
for linear and star PMeOx in chloroform at 258C:
NA ꢅ V
½Zꢁ ¼
ð9Þ
M
From Eqns 8 and 9 results the following equation which
allows calculation of the overlap concentration from viscomet-
ric data:
0:718
½Zꢁ ¼ 2 ꢅ 10ꢀ4 ꢅ Mn
ðdLgꢀ1Þ linear polymer
ð6Þ
ð7Þ
0:693
½Zꢁ ¼ 0:9 ꢅ 10ꢀ4 ꢅ Mn
ðdLgꢀ1Þ star polymer
1
cn ¼
ð10Þ
½Zꢁ
The value of the exponent a typically varies from 0.5 (at the y
condition) to 0.8 (for flexible chains in a good solvent), 0.8 #
a # 1 for inherently stiff molecules (e.g. cellulose derivatives,
DNA) and 1 # a # 1.7 for highly extended chains (e.g. poly-
electrolytes in solutions of very low ionic strength). The a
exponent contains information about the shape of the polymer
chains in solution. The values of a obtained for both linear and
star PMeOx are around 0.7 showing that chloroform is a good
solvent for PMeOx and their chains are flexible. The parameters
K and a obtained for linear and star PMeOx are in agreement
with literature data.[32]
Generally, polymer solutions can be divided into three
regions: dilute, semidilute, and concentrated.[35] In dilute solu-
tion, the concentration is sufficiently low and the polymer chains
conserve their individuality and the intermolecular interactions
are ignored. When the concentration increases, the intermolec-
ular interactions progressively become predominant and at a
certain concentration c* the domain of the polymer molecules
begin to overlap and eventually entanglements may develop. c*
The overlap concentrations for the investigated linear and
star polymers are listed in Table 5. We observe that the c* values
for the star structures are higher than those corresponding to
linear structures of the same molecular weight. The increase of
the molecular weight of the polymer determines the decrease of
c* for both linear and star PMeOx. The increase of hydro-
dynamic dimension of the macromolecular coils and therefore
the increase of [Z] of the star structure has as a consequence the
decrease of c* values. Generally, c* varies as Mꢀa where a is the
Mark-Houwink exponent.[38–40]
Finally, the effect of branching on the intrinsic viscosity of a
star polymer can be expressed by gZ0 defined as the ratio of the
intrinsic viscosities of the branched polymer to that of its linear
equivalent of the same molecular weight[36,41]
:
½Zꢁstar
gZ0 ¼
ð11Þ
½Zꢁlinear

Star Polymers with Cyclodextrin Core and Poly(2-methyl-2-oxazoline) Arms
1153
The gZ0 values estimated from viscometric data for PMeOx
star polymers were between 0.38 and 0.49. The functionality of
the star polymer (arm number), f, can be indirectly determined
by comparing the experimental gZ0 value with the theoretical gZ0
(6.91 g, 26.4 mmol) was added. Next a fresh prepared solution of
CBr4 (8.76 g, 26.4 mmol) in DMF (10 mL) was rapidly added.
The orange-yellow solution was stirred at room temperature for
48 h under N2 pressure. The colour of reaction turns to tan or
brown. Methanol (5 mL) was added to quench the reaction and
the product was precipitated by addition of ethanol (300 mL).
The filter cake was then washed with ethanol followed by a 7 : 3
ethanol/water solution. The product was then dried at 608C
under high vacuum to yield a stable white powder (1.03 g, 90 %).
1H NMR ([D6]DMSO) ppm, d: 5.91, 5.77 (d, broad peak,
OH-2, OH-3, 14H), 4.91 (broad peak, H1, 7H), 4.0–3.1 (m, H2 to
H6, 42H) (Figure S9 in the Supplementary Material). 13C NMR
([D6]DMSO) ppm, d: 102.09 (C1), 83.24 (C4), 72.64, 72.07,
70.37 (C2, C3, C5), 51.37 (C6) (Figure S10 in the Supplemen-
tary Material).
value defined as[41]
:
gZ0 ¼ ð3f ꢀ 2Þ=f2
ð12Þ
This relationship is usually validated for the star polymers
with the same arm length under theta conditions and it is used in
the present paper only to provide the arm number.
The theoretical gZ0 value calculated by imposing the number
of arms f ¼ 7 was 0.39. The experimentally determined gZ0 value
is very close to the predicted one, in agreement with a 7 arm-star
polymer with a CD core.
Synthesis of Heptakis(6-deoxy-6-amino)b-CD
Conclusion
The heptakis(6-deoxy-6-azido)b-CD (1 g, 0.76 mmol) was dis-
solved in DMF (20 mL) then PPh3 (3.16 g, 12.04 mmol) was
added. After 1 h of reaction, a concentrated aqueous ammonium
hydroxide solution (3.1 mL, solution at 5 mol Lꢀ1) was added
dropwise to the solution. The reaction mixture was stirred at
room temperature for 20 h, then concentrated under reduced
pressure and precipitated in ethanol. The insoluble product was
washed with ethanol then dried under vacuum to yield the title
compound (0.84 g, 98 %).
Star polymers with a cyclodextrin core and poly(2-methyl-2-
oxazoline) arms were synthesised using arm-first and core-first
strategies. Investigation of the kinetics of initiation versus
propagation demonstrated that the control of the core-first
strategy is difficult to achieve, and that only the use of a
co-initiator favours a controlled polymerisation if the yields are
high enough. Physico-chemical properties of the linear versus
star polymers are provided, and the results are consistent with
previous papers, for the viscometric characterisation as well for
the DOSY experiments. These new star polymers may be a very
interesting precursor for poly(ethylenimine)-based polymers
that can find some applications in gene delivery. The core-first
strategy allows for the preparation of versatile gene carriers that
can be modified at each end function of the arms with targeting
or ‘stealthness’ agents.
13C NMR (D2O) ppm, d: 104.10 (C1), 84.89 (C4), 74.70,
74.18, 70.58 (C2, C3, C5), 42.64 (C6) (Figure S2 in the
Supplementary Material).
Synthesis of Heptakis(6-deoxy-6-iodo)b-CD
Triphenylphosphine (23.08 g, 88 mmol), DMF (90 mL), and
iodine (22.34 g, 88 mmol) were successively introduced in a
250 mL flask equipped with a condenser under nitrogen. The
diiodide was slowly added while increasing the temperature up
to 508C. Purified b-CD (5 g, 4.4 mmol) was then added. The
solution was stirred at 708C for 18 h under N2. The reaction
mixture was then cooled to room temperature. A solution of
NaOMe in methanol (5.6 g in 80 mL) was prepared under
nitrogen in a 100 mL flask in an ice bath, then carefully intro-
duced in the reaction vessel with cooling. The mixture was
stirred for a few minutes, then precipitated in 600 mL of meth-
anol. The yellow insoluble product was filtered and washed with
methanol until the solvent becomes colourless. The heptakis(6-
deoxy ꢀ6-iodo)b-CD was dried at 808C under vaccum over one
night. The yield of the reaction was 90 %.
1H NMR ([D6]DMSO) ppm, d: 6.04, 5.94 (d, d, OH-2, OH-3,
14H), 4.98 (d, H1, 7H), 3.90–3.10 (m, H2 to H6, 42H)
(Figure S11 in the Supplementary Material). 13C NMR
(DMSO-d6) ppm, d: 102.12 (C1), 85.95 (C4), 72.17, 71.91,
70.96 (C2, C3, C5), 9.54 (C6) (Figure S12 in the Supplementary
Material).
Experimental
Materials
2-Methyl-2-oxazoline (Sigma-Aldrich, 98 %), allyl bromide
(Sigma-Aldrich, 98 %), 1-iodo-2-methylpropane (Alfa Aesar,
98 %), acetonitrile (SDS, 99.8 %), heptane (SDS, 99 %) were
dried over CaH2 and distilled just before use. Diethyl ether was
˚
dried by storage over 4A molecular sieves. Iodine (Sigma-
Aldrich, 99.8 %), sodium methoxide (Fluka, 97 %), acetic
anhydride (Sigma-Aldrich, 99 %), NaN3 (Sigma-Aldrich,
99.5 %), triphenylphosphine (Fluka, 98.50 %), anhydrous
N,N-dimethylformamide (Alfa Aesar, 99.8 %), ammonium
hydroxide solution (5 mol Lꢀ1, Fluka), carbon tetrabromide
(Sigma-Aldrich, 99 %) were used as received without further
purification. The b-CD was purified before use. A saturated
solution of b-CD was prepared at 808C, then slowly cooled to
room temperature and left at 08C for one night. The crystallised
b-CD was then filtered and dried under vacuum at 808C over one
night.
The synthesis of heptakis(6-deoxy-6-azido)b-CD was
adapted from ref. [42], heptakis(6-deoxy-6-amino)b-CD and
heptakis(6-deoxy-6-iodo)b-CD from ref. [43], and heptakis
(6-deoxy-6-iodo-2,3-di-O-acetyl)b-CD from ref. [44].
Synthesis of Heptakis(6-deoxy-6-iodo-2,3-di-O-acetyl)
b-CD
The heptakis(6-deoxy ꢀ6-iodo)b-CD (1 g, 0.525 mmol), acetic
anhydride (77.5 mL, 9.23 mmol), pyridine (5 mL, 62 mmol), and
4-dimethylaminopyridine (6.8 mg, 0.11 mmol) were succes-
sively introduced in a 100 mL flask under nitrogen. The reaction
mixture was stirred at room temperature under nitrogen for 72 h,
then mixed with 40 mL of ice water. The insoluble part was
filtered then crystallised in cold methanol. After filtration, the
Synthesis of Heptakis(6-deoxy-6-azido)b-CD
To a solution of b-cyclodextrin (1 g, 0.88 mmol) in dry DMF
(90 mL), was added NaN3 (6.81 g, 88 mmol) and the resulting
suspension was stirred at 808C for 2 h, under N2 pressure. The
reaction solution was cooled to room temperature before PPh3

1154
G. Pereira et al.
heptakis(6-deoxy-6-iodo-2,3-di-O-acetyl)b-CD was recovered
and dried under vacuum for one night. The yield of the reaction
was 83 %.
1H NMR ([D6]DMSO) ppm, d: 5.35–5.05 (d, dd, H1, H3,
14H), 4.80–4.60 (dd, H2, 7H), 3.95–3.45 (m, H4, H5, H6, 28H),
2.02 (s, CH3, 42). 13C NMR ([D6]DMSO) ppm, d: 170.00,
169.30 (C ¼ O), 95.99 (C1), 79.51 (C4), 70.02, 69.38 (C2, C3,
C5), 20.54 (CH3), 8.70 (C6) (Figure S3 in the Supplementary
Material).
calibrations was checked by measuring the self-diffusion coef-
ficient of a mixture H2O/D2O (10/90 in moles) at 258C.[46] The
DOSY experiments were carried out using the stegp1s pulse
sequence with a linear gradient of 16 steps between 2 and 95 %.
Before each diffusion experiment, the proton relaxation times
were determined in order to correctly set the D1 parameter of the
DOSY sequence and the length of the gradient d and the diffu-
sion time D were optimised for each analysed product. All the
DOSY experiments were carried out at 258C with a concentra-
tion of 16 mg mLꢀ1. The mathematical treatment of the data was
performed as previously described.[45]
The viscometric measurements were carried out in chloro-
form at 258C (ꢃ0.018C) using an Ubbelohde suspended-level
viscometer with a capillary diameter of 0.64 mm (Type 531 10,
Schott-Gera¨te) and an automatic viscometer (Instrument
LAUDA LMV 830) with a water bath (ECO ET 155). The time
measurement error was 1 ppm and the flow volume of the
viscometer was greater than 5 mL, making drainage errors
unimportant. The intrinsic viscosity values provided by viscom-
eter represent a mean of five measurements.
Star Poly(2-methyl-2-oxazoline) Synthesis Using
the Arm-First Method
The monomer 2-methyl-2-oxazoline was polymerised in dried
acetonitrile at 808C under nitrogen. To 23 mL of acetonitrile,
and 0.56 mL of allyl bromide (0.24 mol Lꢀ1) were added 2.7 mL
of 2-methyl-2-oxazoline (1.21 mol Lꢀ1). After stirring for 24 h,
half of the total volume was removed from the reaction vessel
with a syringe. This part was concentrated and dried under
vacuum until a constant weight was obtained. The other part of
the reaction was quenched with the heptakis(6-deoxy-6-amino)
b-CD. So 0.474 g of cyclodextrin derivative (0.42 mmol) dried
under secondary vacuum during one night was dissolved in dried
DMSO (20 mL) and introduced into the reactor. The polymer-
isation medium was left at 808C for 4 days with stirring. The
mixture was then concentrated, dissolved in methylene chloride
and precipitated in Et2O. The insoluble polymer was filtered
then dried under vacuum.
Supplementary Material
Spectra, tables of results and schemes are available as Supple-
mentary Material on the Journal’s website.
Acknowledgements
The research leading to these results has received funding from the European
Union’s Seventh Framework Program (FP7/2007–2013) under grant
agreement no. 264115-STREAM.
Star Poly(2-methyl-2-oxazoline) Synthesis Using
the Core-First Method
The heptakis(6-deoxy-6-iodo-2,3-di-O-acetyl)b-CD was dried
under dynamic secondary vacuum for several hours, then dried
heptane was vacuum-distilled into the vessel containing the
initiator. The mixture was stirred under secondary vacuum for
one night. An azeotropic distillation of heptane/H2O was then
conducted in order to eliminate the maximum amount of
residual water from the CD derivative. I-OAc-b-CD was again
dried under dynamic secondary vacuum for several hours before
use. Under inert atmosphere, dried acetonitrile, monomer, and
iodine solution in acetonitrile if necessary, were added. The
polymerisation medium was heated at 808C. The samples col-
lected versus time were concentrated then dried under vacuum
until a constant weight was obtained. The yield was calculated
removing the weight of the cyclodextrin derivative. The peaks of
the carbons C1 to C5 of the cyclodextrin core were used as
reference to calculate the residual CH2I concentration on the
13C NMR of the polymer.
References
[1] S. B. Kharchenko, R. M. Kannan, J. J. Cernohous, S. Venkataramani,
Macromolecules 2003, 36, 399. doi:10.1021/MA0256486
[2] N. Fazeli, N. Mohammadi, F. Afshar Taromi, Polym. Test. 2004, 23,
431. doi:10.1016/J.POLYMERTESTING.2003.09.014
[3] K. Inoue, Prog. Polym. Sci. 2000, 25, 453. doi:10.1016/S0079-6700
(00)00011-3
[4] T. He, T. Hu, X. Zhang, G. Zhong, H. Zhang, J. Appl. Polym. Sci. 2009,
112, 2120. doi:10.1002/APP.29657
[5] K. Ohno, B. Wong, D. M. Haddleton, J. Polym. Sci. A Polym. Chem.
2001, 39, 2206. doi:10.1002/POLA.1197
[6] M. V. Rekharsky, Y. Inoue, Chem. Rev. 1998, 98, 1875. doi:10.1021/
CR970015O
[7] N. Badi, P. Gue´gan, Carbohydr. Res. 2007, 342, 1989. doi:10.1016/
J.CARRES.2007.06.013
[8] A. R. Khan, P. Forgo, K. J. Stine, V. T. D’Souza, Chem. Rev. 1998, 98,
1977. doi:10.1021/CR970012B
[9] J. Li, Z. Guo, J. Xin, G. Zhao, H. Xiao, Carbohydr. Polym. 2010, 79,
277. doi:10.1016/J.CARBPOL.2009.08.006
[10] M. H. Stenzel-Rosenbaum, T. P. Davis, V. Chen, A. G. Fane, Macro-
Characterisations
NMR spectra were recorded on a Bruker Avance AM300
spectrometer (1H 300 MHz and 13C 75 MHz) using [D6]DMSO
or CDCl3 as solvents at 208C. For the Diffusion-Ordered
SpectroscopY (DOSY) experiments, the maximum field gradi-
ent strength was calibrated using a homemade Plexiglas phan-
tom (8 mm ꢃ 0.01 length and a width equal to the inner diameter
of the NMR tube) inserted in a H2O filled NMR tube and using
the pulse program calibgp. The linear plot of the obtained
gradient strengths against the gradient strength setting (GPZ1)
used gave a maximum field gradient strength equal to
molecules 2001, 34, 5433. doi:10.1021/MA0021803
[11] F. A. Plamper, H. Becker, M. Lanzendo¨rfer, M. Patel, A. Wittemann,
M. Ballauff, A. H. E. Mu¨ller, Macromol. Chem. Phys. 2005, 206, 1813.
doi:10.1002/MACP.200500238
[12] T. Kakuchi, A. Narumi, T. Matsuda, Y. Miura, N. Sugimoto, T. Satoh,
H. Kaga, Macromolecules 2003, 36, 3914. doi:10.1021/MA021296R
[13] M. H. Stenzel, T. P. Davis, J. Polym. Sci. A Polym. Chem. 2002, 40,
4498. doi:10.1002/POLA.10532
[14] P.-F. Gou, W.-P. Zhu, N. Xu, Z.-Q. Shen, J. Polym. Sci. A Polym.
Chem. 2008, 46, 6455. doi:10.1002/POLA.22955
[15] Y. Miao, C. Rousseau, A. Mortreux, P. Martin, P. Zinck, Polymer
2011, 52, 5018. doi:10.1016/J.POLYMER.2011.08.040
[16] M. Adeli, Z. Zarnegar, R. Kabiri, Eur. Polym. J. 2008, 44, 1921.
doi:10.1016/J.EURPOLYMJ.2008.03.028
56.8 G cm .
ꢀ1 [45] The temperature calibration of the spectrome-
ter was performed with a sample of 100 % CH3OH in the tem-
perature range between 298 and 313K. The accuracy of the

Star Polymers with Cyclodextrin Core and Poly(2-methyl-2-oxazoline) Arms
1155
[17] C. Yang, H. Li, S. H. Goh, J. Li, Biomaterials 2007, 28, 3245.
doi:10.1016/J.BIOMATERIALS.2007.03.033
[18] H. Huang, H. Yu, G. Tang, Q. Wang, J. Li, Biomaterials 2010, 31,
[32] G. Volet, V. Chanthavong, V. Wintgens, C. Amiel, Macromolecules
2005, 38, 5190. doi:10.1021/MA050407U
[33] T. Saegusa, T. Ikeda, Macromolecules 1973, 6, 808. doi:10.1021/
MA60036A004
1830. doi:10.1016/J.BIOMATERIALS.2009.11.012
[19] M. W. M. Fijten, C. Haensch, B. M. Van Lankvelt, R. Hoogenboom,
U. S. Schubert, Macromol. Chem. Phys. 2008, 209, 1887. doi:10.1002/
MACP.200800226
[20] R. Hoogenboom, Angew. Chem. Int. Ed. 2009, 48, 7978. doi:10.1002/
ANIE.200901607
[21] A. Kowalczuk, J. Kronek, K. Bosowska, B. Trzebicka, A. Dworak,
[34] K. Takeo, T. Kuge, Starch 1972, 24, 331. doi:10.1002/STAR.
19720241003
[35] S. Morariu, C. E. Brunchi, L. Ghimici, M. Bercea, Viscometric
Behaviour of Polymer Solutions, in Functional Polymeric
Materials Designed for High Tech Applications 2010, pp. 85–107
(Ed. M. Nechifor) (Research Signpost: Kerala, India).
[36] W. Burchard, Adv. Polym. Sci. 1999, 143, 113. doi:10.1007/3-540-
49780-3_3
[37] T. Aberle, W. Burchard, Starch 1997, 49, 215. doi:10.1002/STAR.
19970490602
[38] S. Morariu, M. Bercea, S. Ioan, C. Ioan, Eur. Polym. J. 1999, 35, 377.
doi:10.1016/S0014-3057(98)00144-X
Polym. Int. 2011, 60, 1001. doi:10.1002/PI.3103
[22] R. Luxenhofer, M. Bezen, R. Jordan, Macromol. Rapid Commun.
2008, 29, 1509. doi:10.1002/MARC.200800293
[23] R. Hoogenboom, M. W. M. Fijten, G. Kickelbick, U. S. Schubert,
Beilstein J. Org. Chem. 2010, 6, 773. doi:10.3762/BJOC.6.96
[24] E. Bertrand, C. Gonc¸alves, L. Billiet, J.-P. Gomez, C. Pichon,
H. Cheradame, P. Midoux, P. Gue´gan, Chem. Commun. 2011, 47,
12547. doi:10.1039/C1CC15716G
[39] M. Bercea, C. Ioan, C. Ioan, B. C. Simionescu, C. I. Simionescu, Prog.
Polym. Sci. 1999, 24, 379. doi:10.1016/S0079-6700(99)00007-6
[40] P. Gupta, C. Elkin, T. E. Long, G. L. Wilkes, Polymer 2005, 46, 4799.
doi:10.1016/J.POLYMER.2005.04.021
[25] K. Aoi, M. Okada, Prog. Polym. Sci. 1996, 21, 151. doi:10.1016/0079-
6700(95)00020-8
[26] S. Kobayashi, H. Uyama, J. Polym. Sci. A Polym. Chem. 2002, 40, 192.
doi:10.1002/POLA.10090
[41] B. H. Zimm, W. H. Stockmayer, J. Chem. Phys. 1949, 17, 1301.
doi:10.1063/1.1747157
[27] K. F. Morris, C. S. Johnson, Jr, J. Am. Chem. Soc. 1993, 115, 4291.
doi:10.1021/JA00063A053
[42] J. L. Jime´nez Blanco, J. M. Garcia Fernandez, A. Gadelle, J. Defaye,
Carbohydr. Res. 1997, 303, 367. doi:10.1016/S0008-6215(97)00176-6
[43] P. R. Ashton, R. Ko¨niger, J. Fraser Stoddart, J. Org. Chem. 1996, 61,
903. doi:10.1021/JO951396D
[44] F. Cao, Y. Ren, W. Hua, Carbohydr. Res. 2009, 344, 526. doi:10.1016/
J.CARRES.2008.10.001
[45] V. Bennevault-Celton, A. Urbach, O. Martin, C. Pichon, P. Gue´gan,
P. Midoux, Bioconjug. Chem. 2011, 22, 2404. doi:10.1021/
BC200167P
[28] R. Huo, R. Wehrens, J. van Duynhoven, L. M. C. Buydens, Anal. Chim.
Acta 2003, 490, 231. doi:10.1016/S0003-2670(03)00752-9
[29] Y. Oda, S. Matsuda, T. Yamanoi, A. Murota, K. Katsuraya, Supramol.
Chem. 2009, 21, 638. doi:10.1080/10610270802709345
[30] S. Viel, M. Mazarin, R. Giordanengo, T. N. T. Phan, L. Charles,
S. Caldarelli, D. Bertin, Anal. Chim. Acta 2009, 654, 45. doi:10.1016/
J.ACA.2009.06.049
[31] Q. Liu, M. Konas, J. S. Riffle, Macromolecules 1993, 26, 5572.
doi:10.1021/MA00073A007
[46] M. Holz, H. Weinga¨rtner, J. Magn. Reson. 1991, 92, 115.