Synthesis of pH-sensitive micelles from linseed oil using atom transfer radical polymerisation (ATRP)

Article abstract of DOI:10.1016/j.polymer.2012.07.041

This paper reports the synthesis of an amphiphilic copolymer from linseed oils and its successive auto-association in water into pH-sensitive micelles. An original ATRP lipoinitiator is first designed from linseed oil in two steps. tert-butyl acrylate (tBA) polymerization is consequently initiated from this original initiator and amphiphilic copolymers are obtained after subsequent acidolysis of the PtBA block into poly(acrylic acid) (PAA). The ability of a lipid-b-PAA copolymer to auto-associate in water is finally investigated through different techniques (Fluorescence, Surface Tension, QELS). This copolymer forms well-defined micelles in acidic media with a low critical micellar concentration (cmc) of 7.6 mg L^-1 and dissociates when the pH is raised above 7.

Full text of DOI:10.1016/j.polymer.2012.07.041

Polymer 53 (2012) 4344e4352
Contents lists available at SciVerse ScienceDirect
Polymer
journal homepage: www.elsevier.com/locate/polymer
Synthesis of pH-sensitive micelles from linseed oil using atom transfer radical
polymerisation (ATRP)
Louise Hespel , Elise Kaifas , Laurence Lecamp , Luc Picton , Gaëlle Morandi ,*, Fabrice Burel a
a
a
a
b
a
a
INSA de Rouen, Polymères Biopolymères Surfaces CNRS UMR 6270 FR3038, Avenue de l’université BP08, 76801 Saint Etienne du Rouvray Cedex, France
Université de Rouen, Polymères Biopolymères Surfaces CNRS UMR 6270 FR3038, UFR des Sciences et Techniques, bâtiment Pierre-Louis Dulong, Bd Maurice de Broglie,
b
7
6821 Mont Saint Aignan Cedex, France
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 15 March 2012
Received in revised form
This paper reports the synthesis of an amphiphilic copolymer from linseed oils and its successive auto-
association in water into pH-sensitive micelles. An original ATRP lipoinitiator is first designed from
linseed oil in two steps. tert-butyl acrylate (tBA) polymerization is consequently initiated from this
original initiator and amphiphilic copolymers are obtained after subsequent acidolysis of the PtBA block
into poly(acrylic acid) (PAA). The ability of a lipid-b-PAA copolymer to auto-associate in water is finally
investigated through different techniques (Fluorescence, Surface Tension, QELS). This copolymer forms
13 July 2012
Accepted 16 July 2012
Available online 13 August 2012
ꢀ
1
well-defined micelles in acidic media with a low critical micellar concentration (cmc) of 7.6 mg L and
dissociates when the pH is raised above 7.
Keywords:
Amphiphilic copolymer
Stimuli-sensitive micelles
Vegetable oil
Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction
drug delivery system [5]. Developing new vectorisation systems of
smaller size, typically micelles, but having the same affinity with
By using nanoscale delivery vehicles, the pharmacological
properties (e.g. solubility, circulating half-life) of drugs can be
drastically improved, leading to safe and effective systems. Nano-
technologies, by addressing some of the shortcomings associated
with potential drugs, have made a significant impact on the
development of drug delivery systems [1]. Since liposomes were
first described in the 1960s and proposed as carriers of proteins and
drugs for disease treatment, they have been extensively studied
and their efficiency as drug delivery systems has been demon-
strated [2]. Liposomes can efficiently encapsulate both hydrophilic
and hydrophobic drugs and their similarity with the constitution of
cell membranes improve interactions between cells and the
delivery device. There has been considerable interest in modifica-
tion of liposomes surfaces to enhance their blood circulation time,
obtain a controlled release [3], and introduce homing devices
cells membrane is then of great interest and we are consequently
interested in amphiphilic block copolymers containing a hydro-
phobic block based on a lipid chain.
One simple strategy to obtain lipid-b-polymer architecture is to
post-modify a preformed polymer chain. According to this strategy,
the synthesis of poly(ethylene glycol)-b-phospholipids was re-
ported in the literature [6]. These copolymers form micelles with
ꢀ5
a low critical micellar concentration (cmc) of 10 M and present
interesting loading and release properties. Another synthetic
approach is to use directly a lipid derivative as the initiator (or the
transfer agent) to carry out the polymerization. This approach
allows a better control of the final structure and avoids a subse-
quent purification step to eliminate unreacted polymeric chains.
According to this strategy, the synthesis of lipid-based initiator/
transfer agent for free radical polymerization [7e9], controlled
radical polymerization (RAFT [10], NMP [11,12], ATRP [13e15]) and
ring-opening polymerization [16] were reported. However, in all
these studies, the so-called “lipid” block was issued from
petroleum-based reactants (mainly aminoalkyl chains) and not
from renewable resources. Furthermore, these studies were mainly
developed by the lipid-membrane community [17,18], conse-
quently the final copolymers were involved in liposomes formation
or vesicles adsorption, and their ability to form micelles was not
studied. Development of polymers from renewable resources is
(
antibody, receptor, ligand). Liposomes are now in the marketplace
as pharmaceuticals, however liposomology still faces major defi-
ciencies [4]. The relatively large size of liposomes is an inconvenient
to access tumour with a small vasculature cutoff and due to their
smaller size, micelles may have additional advantages as a tumour
*
Corresponding author. Tel.: þ332 32 95 66 49; fax: þ332 32 95 66 44.
E-mail address: gaelle.morandi@insa-rouen.fr (G. Morandi).
0
032-3861/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.polymer.2012.07.041

L. Hespel et al. / Polymer 53 (2012) 4344e4352
4345
a growing field of interest [19,20], and a major concern of the
2.3. Linseed oil reduction
scientific community. Our group aims at developing original bio-
sourced lipopolymers directly from naturally occurring vegetable
oils using ATRP. In that purpose, an original synthetic pathway in
two main steps was chosen: (1) the vegetable oil is first chemically
modified to introduce an initiating site for Atom Transfer Radical
Polymerization (ATRP) leading to a lipoinitiator and (2) a synthetic
poly(acrylic acid) hydrophilic block is added to the lipid block using
ATRP of tert-butyl acrylate followed by subsequent acidolysis. If
modified fatty acids have already been polymerized by ATRP [21e
A solution of linseed oil (2.9 g) in 45 mL of anhydrous THF was
added dropwise to a mixture of lithium aluminohydride (2.1 g) in
45 mL of THF under inert atmosphere. The reaction media was
stirred at room temperature overnight. After hydrolysis of the
4
unreacted LiAlH through water dropwise addition (w20 mL), the
reaction media was filtered and THF was removed by evaporation.
The resulting oil was consequently solubilized in dichloromethane
and washed with a saturated sodium chloride aqueous solution.
The organic layer was dried over anhydrous magnesium sulphate,
2
4], this is, to the best of our knowledge, the first report of an oil
derivative used as an ATRP initiator.
Linseed oil was chosen as the starting material because of its
large local bioavailability added to its low use in the food industry
and the solvent was removed by evaporation. The final alcohol was
1
obtained as a yellowish oil with a yield of 89%. H NMR (CDCl
3
,
300 MHz):
(m, eHC ¼ CHeCH
(m, CH eCH OH), 1.25 (m, CH
alcohol), 0.8 (m, CH other fatty alcohols). C NMR (CDCl
75 MHz): eOH),
(ppm) 131.9e127.1 (eHC ¼ CHe), 63.0 (eCH
32.7e20.5 (CH aliphatic), 14.2 (CH alkyl chain). MS [M-Li ]:
m/z ¼ 271 (linolenic) 273 (linoleic), 275 (oleic), 277 (stearic).
d
(ppm) 5.3 (m, eHC ¼ CHe), 3.6 (t, eCH
eHC ¼ CHe), 2.0 (m, eHC ¼ CHeCH
aliphatic), 0.9 (t, CH
2
eOH), 2.75
eCH ), 1.5
linolenic
(
indeed avoiding any market competition). Vegetable oils are
2
2
2
constituted of triglycerides with a statistical mixture of fatty acids
which are highly polyunsaturated ones in the case of linseed oil
2
2
2
3
13
3
3
,
[25]. Obtaining well-defined systems from heterogeneous starting
d
2
þ
material is an interesting challenge to take up. Amongst the
different possible synthetic approaches, fatty acids chains were
chosen to constitute the hydrophobic block in order to reduce the
heterogeneity of the final material (compared to a triglycerides-
based copolymer) and to obtain small-sized micelles.
2
3
2.4. Lipoinitiator 1 synthesis
Fatty alcohol 1 (1.20 g) was mixed with triethylamine (1.9 mL, 3
eq.) and anhydrous dichloromethane (30 mL) under argon atmo-
sphere. The mixture was cooled by an ice-water bath and the 2-
bromoisobutyryl bromide (1,1 mL, 2 eq.) was added dropwise
under stirring. The mixture was stirred at room temperature during
2
. Experimental section
2.1. Materials
4
h. Hydrochloric acid aqueous solution (10%, 15 mL), sodium
THF was distilled over sodium/benzophenone and dichloro-
methane over CaH
2
prior to use. Linseed oil was kindly provided by
hydrogenocarbonate aqueous solution (10%, 15 mL) and saturated
sodium chloride aqueous solution (15 mL) were consequently used
to wash the dichloromethane solution. The organic layer was dried
over anhydrous magnesium sulphate, and filtered. After removing
the solvent by evaporation, a column chromatography was per-
formed to purify the product (diethylether/cyclohexane: 5/95). The
Novance (Compiègne, France). tert-butyl acrylate (tBA, 99%) from
ꢁ
Acros was distilled under vacuum and stored at 4 C after purifi-
0
0
00
cation. 2-bromo-2-methylpropionylbromide (98%), N,N,N ,N ,N -
pentamethylethylenetriamine (PMDETA, 99þ%), anisole (99%),
toluene (99%), Silica Gel (60e200 mesh), lithium aluminohydride,
and triethylamine were purchased from Acros. Copper (I) bromide
final product was obtained as a colourless liquid with a yield of 65%.
1
H NMR (CDCl
eOCO), 2.75 (m, eHC ¼ CHeCH
m, eHC ¼ CHeCH eCH ), 1.85 (s, Br(CH Ce), 1.6 (m, CH
OCO), 1.25 (m, CH aliphatic), 0.9 (t, CH linolenic chain), 0.8
, 75 MHz): (ppm) 171.7
C]O), 131.9e127.0 (eHC ¼ CHe), 66.1 (eCH eOCO), 55.9 (CeBr),
1.8e22.6 (CH aliphatic), 30.7 (Br(CH Ce), 14.1 (CH alkyl
chain). MS [M-Liþ]: m/z ¼ 421 (linolenic), 423 (linoleic), 425 (oleic),
27 (stearic).
3
, 300 MHz):
d
(ppm) 5.3 (m, 2H, eHC ¼ CHe), 4.1
(
99.99%) was purchased from Aldrich. Copper (II) bromide (99%)
was purchased from Alfa Aesar.
(t, 2H, eCH
2
2
eHC ¼ CHe), 2.0
(
CH
2
2
3
)
2
2
e
2
2
3
13
(
(
m, CH
3
other chains). C NMR (CDCl
3
d
2
.2. Instrumentation
2
3
2
3
)
2
3
1H and 13C NMR spectra were recorded on a Bruker AC-P
00 MHz spectrometer. Chemical shifts are reported in ppm rela-
3
4
tive to the deuterated solvent resonances. Molecular weights and
molecular weight distributions were measured using size exclusion
chromatography (SEC) on a Varian PL-GPC50 device equipped with
two mixed packed columns (PL gel mixed type C). The eluent used
2.5. tBA ATRP
ꢀ
1
is dichloromethane at a flow rate of 1 mL min at room temper-
ature and poly(methyl methacrylate) standards were used for
calibration. Fourier Transform InfraRed (FTIR) spectra were recor-
ded on a Perkin Elmer Spectrum 2000 FTIR, equipped with a dia-
mond ATR (Attenuated Total Reflection) device. Fluorescence
measurements were recorded on a Cary Eclipse Varian spectro-
photometer. Surface tension was measured using Krüss Tensiom-
eter K12 with the SFT Wilhelmy plate method and the following
A Schlenk tube was loaded with copper (I) and copper (II)
bromide, capped with a rubber septum, and cycled three times
between vacuum and argon to remove oxygen. In another Schlenk
tube, toluene, lipoinitiator 1, anisole (internal reference) and tBA
were introduced according to the following ratios: [tBA]:[(1)]:[Cu(I)
Br]:[Cu(II)Br
2
]:[PMDETA] ¼ 50 : 1 : 0.5 : 0.025 : 0.525, toluene: 75%
ꢁ
v/v, anisole: 5% v/v, 60 C. The resulting solution was degassed by
three freezeepumpethaw cycled and was added to the Cu(I)Br and
parameters:
detection
speed:
6
mm/min,
detection
2
Cu(II)Br contained in the first Schlenk tube via a cannula. The
sensitivity ¼ 0.01 g, immersion depth ¼ 2 mm, values ¼ 10,
Acquistion ¼ linear, times ¼ 1000 s. The terminal conditions are
values for mean ¼ 5 values and standard deviation ¼ 0 mN/m. QELS
measurements were recorded on a Malvern Zetasizer. ESI-MS
analyses were performed on an Esquire-LC ion-trap mass spec-
trometer equipped with an ESI source and the Esquire control 6.16
data system (Bruker Daltonics, Bremen, Germany).
Schlenk tube was placed in an oil bath thermostated at the poly-
merization temperature. At t ¼ 0, the ligand was added. Aliquots
were taken periodically via a degassed syringe to follow the kinetic
of the polymerization process. The aliquots were diluted with THF
followed by filtration through a basic alumina column prior to
analysis by SEC. The final polymer 2 was isolated by filtration
through a basic alumina column followed by solvent evaporation.

4346
L. Hespel et al. / Polymer 53 (2012) 4344e4352
2
.6. Lipid-b-poly(tert-butyl acrylate) acidolysis
was used to prepare the copolymer 3b solutions ranging from
ꢀ1 ꢀ1
1
mg L to 1 g L . The solution, prepared by successive dilution
ꢀ1
Trifluoroacetic acid (TFA) (2 equiv. according to tert-butyl acry-
from a stock solution of 1 g L were stirred overnight prior to
measurement. Excitation spectra of pyrene were recorded between
330 and 360 nm (bandwidth ¼ 1.5 nm) at a fixed emission wave-
late units) was added to a solution of copolymer 2 in dichloro-
methane. The reaction mixture was stirred overnight until the de-
tert-butylated polymer precipitated. Polymer 3 was isolated by
filtration and washed with dichloromethane prior to drying.
length (
.10. QELS measurements
A copolymer 3 solution (1 g L ) was prepared in a buffer
l
Em ¼ 374 nm, bandwidth ¼ 5 nm).
2
2
.7. Conductimetry e determination of the number of repeating unit
ꢀ1
mL of an hydrochloric acid solution (0.1 mol Lꢀ1) was added to
solution containing 0.1 M of NaCl (pH 4.6: 10 M sodium acetate/
acetic acid, pH 7: 10 M sodium phosphate/disodium phosphate,
pH 10: 10 M sodium bicarbonate/sodium carbonate), filtered on
a 0.45 m cellulose filter and introduced in the Zetasizer device.
ꢀ2
1
ꢀ1
ꢀ2
the solution of amphiphilic copolymer 3 (1 g L ) in order to stress
the beginning of the carboxylic acid units titration (through the
slope gradient change). The solution was then titrated using
a solution of sodium hydroxide (0.1 mol L ) and the titration is
followed by conductimetry. Calculation details are provided in the
supporting information (S4).
ꢀ
2
m
ꢀ
1
2.11. Surface tension
ꢀ
1
ꢀ1
Copolymer 3 solutions (1 mg L to 1 g L ) were prepared by
ꢀ1
2.8. Determination of the iodine value
successive dilution from a copolymer stock solution (1 g L ) in
a buffer solution containing 0.1 M of NaCl (pH 4.6: 10 M sodium
M sodium phosphate/disodium
phosphate, pH 10: 10 M sodium bicarbonate/sodium carbonate).
ꢀ
2
ꢀ
2
This iodine value was determined according to the Wijs method.
acetate/acetic acid, pH 7: 10
ꢀ2
A known amount of lipoinitiator 2 (0.1 g) is introduced in 20 mL of
dichloromethane and 20 mL of Wijs solution is added. The solution
is stirred during 1 h in the dark. Then 100 mL of water and 20 mL of
potassium iodide solution (C ¼ 10 g/L) are introduced. This mixture
is titrated with a solution of sodium thiosulfate (C ¼ 0.2 mol/L).
Details are provided in the supporting information (S3).
3
. Results and discussion
3.1. Lipoinitiator synthesis
As a natural product, the composition of linseed oil varies with the
2
.9. Fluorescence measurements
source. The starting linseed oil was characterized by Mass Spec-
trometryand H NMR (Sup. inf. S1) to determine theratioofeach fatty
1
ꢀ3
A pyrene stock solution (4 ꢂ 10 M e 50 mL) in acetone was
prepared and 50 L were introduced in a 500 mL volumetric flask
to obtain a final pyrene concentration of 4.10 M after dilution).
Acetone was allowed to evaporate and a buffer solution containing
acid accordingto a previously reported method [25]. Thislinseed oilis
composed of 54.0% linolenic acid, 9.5% linoleic acid, 23.4% oleic acid
and 13.1% of saturated acid. Linseed oil was first reacted with lithium
aluminohydride to reduce the triglyceride ester functions and obtain
fatty alcohols (Scheme 1 e forclarity purpose, the startingtriglyceride
has been depicted as a 1:1:1 mixture of linolenic, linoleic and oleic
acids, butonehastokeepinmindthatitisinfactacomplexmixtureof
m
ꢀ7
(
ꢀ
2
0
.1 M of NaCl was added (pH 4.6: 10 M sodium acetate/acetic
ꢀ2
acid, pH 7: 10 M sodium phosphate/disodium phosphate, pH 10:
10
ꢀ
2
M sodium bicarbonate/sodium carbonate). This stock solution
Scheme 1. Lipoinitiator(s) synthesis from linseed oil (for clarity purpose, the starting triglyceride has been depicted as a 1:1:1 mixture of linolenic, linoleic and oleic acids, but one
has to keep in mind that it is in fact a complex mixture of different fatty acids).

L. Hespel et al. / Polymer 53 (2012) 4344e4352
4347
A
Glycerol CH2
Hβ
Hα
B
Hα
C
Hβ
5.6 5.4 5.2 5.0 4.8 4.6 4.4 4.2 4.0 3.8 3.6 3.4 3.2 3.0 2.8 2.6 2.4 2.2 2.0 1.8 1.6 1.4 1.2 1.0 0.8
Fig. 1. NMR spectra of A) linseed oil, B) fatty alcohol and C) lipoinitiator 1.
different fatty acids as mentioned previously). NMR analysis (Fig. 1B)
unsaturations/molecule for fatty alcohols and lipoinitiator 1, proving
the double bond conservation during the lipoinitiator synthesis.
Synthesis of lipoinitiator 1 with a good yield in two easy steps is
then demonstrated. As previously mentioned, the alkyl chains
nature is not homogeneous so that lipoinitiator 1 is a mixture of C18
chains containing various numbers of unsaturations (between
0 and 3), the average value being 2 unsaturations/chain. This is, to
the best of our knowledge, the first reported synthesis of a lip-
oinitiator directly from renewable resources.
of the fatty alcohol shows the disappearance of the CH
moiety signals (
¼ 4.1 ppm) and the appearance of signals corre-
sponding to the protons in and positions of the hydroxyl function
¼ 3.6 ppm and ¼ 1.5 ppm). FT-IR Analysis (Sup. inf. S2) shows
the disappearance of the
appearance of a (OeH) band at 3320 cm
2
glycerol
d
a
b
(d
Ηa
d
Ηb
ꢀ
1
n
(C]O) signal at 1730 cm
and the
ꢀ1
n
.
The hydroxyl extremity was subsequently reacted with 2-bro-
moisobutyryl bromide to introduce the ATRP initiating moiety. The
1
H NMR spectrum (Fig. 1C) of the final product shows the shift of
the signals corresponding to the protons in
a
and
b
positions of the
3.2. Amphiphilic copolymers synthesis
introduced ester function (
d
Η
a
¼ 4.1 ppm and
d
Η
b
¼ 1.6 ppm) and
the appearance of the gem-dimethyl signal at 1.85 ppm. This
Lipoinitiator 1was then engaged in ATRP to synthesize a hydrophilic
block and obtain amphiphilic copolymers. tert-Butyl acrylate (tBA) was
chosen as the monomer because it’s an hydrophobic monomer
allowing to conduct the polymerization in an organic media compat-
ible with the initiator 1 lipo-affinity. This choice avoids then any
compatibility issues between the initiator and the monomer. Further-
more, hydrophobic poly(tert-butyl acrylate) easily give access through
acidolysis to hydrophilic PAA, commonly accepted as a bioadhesive and
safe polymer [26]. Lipoinitiator 1 was first engaged in copper mediated
1
3
structure is confirmed by C NMR. The appearance of a
n
(C]O)
signal at 1735 cm is also observed in FTIR (sup. inf. S2) simulta-
neously to the (OeH) signal disappearance.
ꢀ1
n
The average number of unsaturations per molecules was investi-
gated at each step by NMR and by titration using the Wijs method.
Detailsareprovidedinthesupportinginformation(S3).Bothmethods
give consistent results with average values of 6 unsaturations/mole-
cule for linseed oil (containing
3 fatty acid chains) and 2
Scheme 2. Synthesis of amphiphilic lipid-b-poly(acrylic acid) copolymers from lipoinitiator 1 (for clarity purpose only the major aliphatic chain is represented).

4348
L. Hespel et al. / Polymer 53 (2012) 4344e4352
A
2
2
1
1
0
0
.50
.00
.50
.00
.50
.00
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0
10
20
30
40
50
60
time (h)
5
5
4
4
3
3
2
2
1
1
500
000
500
000
500
000
500
000
500
000
1.45
B
1
1
1
1
1
1
.4
.35
.3
.25
.2
.15
1.1
0.85
0
.05
0.15
0.25
0.35
0.45
0.55
0.65
0.75
Conversion
23h 8h 5h
54h
2.5h
C
(
min)
ꢁ
Fig. 2. Kinetic study of tBA ATRP initiated by lipoinitiator 1:[tBA]:[1]:[Cu(I)Br]:[Cu(II)Br2]:[PMDETA] ¼ 50:1:0.5:0.025:0.525, toluene (75% v:v), 60 C. (A) Kinetic plot, (B) Evolution
of the number average molecular weights (– theoretical, - experimental) and the polydispersity indexes (
6) vs. monomer conversion and (C) GPC traces.

L. Hespel et al. / Polymer 53 (2012) 4344e4352
4349
0
0
ATRP of tBA with a Cu(I)Br/Cu(II)Br
2
/PMDETA (N,N,N’,N’,N -pentam-
dichloromethane leads after shaking to a fine dispersion, stable
over days (Fig. 3B). The surfactant properties of our original bio-
sourced copolymer 3 are then qualitatively observable.
In order to better characterize the auto-association properties of
copolymer 3c in water, excitation fluorescence measurements
using pyrene as a probe were conducted (the pyrene was excited
ꢁ
ethyldiethylenetriamine) catalytic system at 60
Scheme 2).
The ln([M]
0 h showing a first order kinetic compatible with a constant
C in toluene
(
0
/[M]) vs. time plot (Fig. 2A) is linear during the first
1
concentration of active species until 50% conversion. The number
average molecular weight vs. conversion plot provides good
agreement between experimental and theoretical molecular
weights (even above 50% conversion) and the polydispersity index
progressively decreases with the conversion (Fig. 2B).
with
l
Ex ¼ 330e360 nm and its emission was then recorded at
l
Em ¼ 374 nm). Depending on the pyrene surroundings, the
maximum emission is obtained for different excitation wave-
lengths, shifting from 333 nm to 338 nm when going from hydro-
philic to hydrophobic surroundings. The formation of micelles is
then easily observable from a shift of the excitation wavelength
corresponding to the maximum emission of pyrene [29]. Pyrene
excitation spectra were recorded for a range of copolymer
0
The non-linear plot for ln([M] /[M]) vs. time above 50% conver-
sion indicates a decrease in the propagating species concentration
which could be attributed to termination reactions involving the
lipoinitiator unsaturation sites (bychainechain coupling). However,
termination reactions should be accompanied by a deviation
between the experimental and the theoretical molecular weights,
a PDI increase and the appearance of shoulders on the GPC traces
ꢀ1
ꢀ1
concentration going from 1 mg L to 1 g L and the ratios of the
emission intensity for
was plotted vs. the copolymer concentration (Fig. 4). The critical
micellar concentration (cmc) corresponds to the point where the
maximum wavelength begins to change due to pyrene incorpora-
tion in the micelles hydrophobic core i.e. a I338/I333 ratio change.
This experiment was carried out in buffer solutions (10
different pHs (4.6, 7 and 10) to investigate the pH influence on the
copolymer auto-association ability, and in presence of 0.1 of NaCl
lEx ¼ 338 nm (I338) over
lEx ¼ 333 nm (I333
)
(Fig. 2C), and none of these phenomena was observed.
Thus, the non-living character of the polymerization is appar-
ently not accompanied by a loss of molecular weights control and
lipoinitiator 1 enables a good control of the final copolymer
composition and architecture.
A range of lipid-b-PtBA copolymers 2 with different lengths of
the PtBA block were then synthesized by varying the initial
monomer]/[initiator] ratio and the final conversion (Table 1). The
theoretical polymerization degree of the PtBA block is ranging from
to 47 and there is a good agreement between theoretical and
experimental number average molecular weights further proving
the interest of ATRP to control the hydrophilic block length.
To obtain the final amphiphilic copolymers, an acidolysis using
trifluoroacetic acid (TFA) is performed to deprotect the tert-butyl
moieties. We focused our attention on copolymer 3c (issued from 2c
ꢀ
2
M) at 3
M
to shield the electrostatic repulsion between the carboxylate
moieties. At pH 4.6 the formation of hydrophobic clusters is
[
ꢀ1
ꢀ6
ꢀ1
detected at 8 mg L (i.e. 5.3 ꢂ 10 mol L ). As reported in the
literature, pyrene fluorescence spectrum is not significantly modi-
fied in presence of pure PAA even for low ionization degree [30,31].
Consequently, the detected hydrophobic clusters can’t be attributed
to pyrene solubilization into the PAA segments of the unimers but
definitely corresponds to micelles formation. The cmc value is more
than tenfold increased at pH 7, whereas at pH 10 no cluster
4
ꢀ
1
ꢀ1
acidolysis), presenting a theoreticalmolecular weightof 1634 g mol
formation is detected anymore until 1 g L . This behaviour can be
(
Mn;theo ¼ M
þ X
ꢂ MAA) and a hydrophilic/hydro-
explained by the modification of the copolymer hydrophilic/
hydrophobic balance when the pH is increased. At pH 4.6, the
polymeric chain is mainly on the carboxylic acid form, and micelles
are obtained at low concentrations. When the proportion of
deprotonated moieties increases, the copolymer hydrophilicity also
increases shifting the cmc towards higher concentrations.
However, as pyrene has been reported to be inefficient to monitor
micelles formation at pH ¼ 11 in the case of fluorocarbon-modified
PAA [32], QELS measurements (size distribution by number) have
lipoinitiator1
n;theo
phobic balance of 75/25 (wt/wt). Consequently, this copolymer
should easily disperse in water and self-assemble into small and
relatively monodisperse micelles [27]. The final lipid-b-PAA copol-
ymer 3c was titrated with a sodium hydroxide solution followed by
pH-metry and conductimetry to determine the copolymer pKa and
the average number of acrylic acid (AA) unit per chain. A pKa value of
5
.7, in really good agreement with the literature value of 5.8 is
measured [28]. An average value of 19 AA unit/chain is also obtained
from this titration (see detailed calculation in supporting information
S4). This really good agreement between the experimental X
and the theoretical X
¼ 17 calculated fromthe monomer conversion,
further proves the good control of the polymerization.
n
¼ 19
n
3.3. Micelles characterization
The final copolymer 3c is soluble in water and lead to uncol-
oured solution that tends to foams when stirred (Fig. 3A). Addition
of lipid-b-PAA copolymer 3c to
a mixture of water and
Table 1
Characterization of the lipid-b-PtBA copolymers.
Xn;theoa
Mn;theo
b
M
n;exp
c
PDIc
Ref.
[Monomer]/
Conv.
(%)
(g molꢀ1)
ꢀ1
(g mol )
[
initiator]
2a
2b
2c
2d
2e
2f
50
100
50
50
100
100
8
14
34
44
37
47
4
14
17
22
37
47
926
2080
2590
3290
5146
6426
1100
2230
2700
3170
5100
7200
1.25
1.50
1.27
1.28
1.32
1.26
a
b
c
Xn;theo ¼ ½monomerꢃ=½initiatorꢃ ꢂ conversion.
Mn;theo ¼ M þ X ꢂ MtBA
.
Fig. 3. A) Lipid-b-PAA 3c in water, B) Emulsion dichloromethane/water stabilized
by 3c.
lipoinitiator1
n;theo
determined by SEC using poly(methyl methacrylate) standard calibration.

4350
L. Hespel et al. / Polymer 53 (2012) 4344e4352
I338/I333 = f(c)
1.10
1.00
0.90
0.80
0.70
0.60
0.50
0.40
0.30
0.0001
0.001
0.01
C (g.L )
0.1
1
-
1
Fig. 4. I338/I333 vs. concentration for copolymer 3c at (:) pH ¼ 4.6, (-) pH ¼ 7 and (C) pH ¼ 10.
3
2
2
1
1
0
A
5
0
5
0
5
0
0
0
5
10
Diameter (nm)
15
20
3
2
2
1
1
B
5
0
5
0
5
0
0
0
5
10
Diameter (nm)
15
20
3
2
2
1
1
C
5
0
5
0
5
0
0
5
10
15
20
Diameter (nm)
Fig. 5. QELS analysis (size distribution by number) for copolymer 3c at A) pH ¼ 4.6, B) pH ¼ 7 and C) pH ¼ 10.

L. Hespel et al. / Polymer 53 (2012) 4344e4352
4351
55
50
45
40
35
0
0.2
0.4
0.6
0.8
1
-1
C (g.L )
Fig. 6. Surface tension vs. copolymer 3c concentration at (:) pH ¼ 4.6, (-) pH ¼ 7 and (C) pH ¼ 10.
been performed in identical conditions (buffered solutions þ 0.1
M
of
behaviour of amphiphilic copolymers in aqueous solution is
complex and the existence of two different critical concentrations
has already been reported for amphiphilic polysaccharides by
Duval-Terrié et al. [34]. Depending on the pH, the adsorption
behaviour of copolymer 3c at the air/water interface changes. At pH
NaCl) to investigate the object size vs. pH (Fig. 5). At pH ¼ 4.6, one
homogeneous population with an average diameter of 50e60 Å,
consistent with the size of micelles, is observed. Although never
detected by QELS analysis (size distribution in number) due to their
limited numbers, the simultaneous presence of unimers in equilib-
riumwith the micelles is highly probable [33]. At pH ¼ 10, the average
diameter is divided by 2 (20e30 Å) and could corresponds to the
presence of unimers in solution. At pH ¼ 7, the diameter distribution
dispersity is a bit higher, ranging from 20 to 40 Å, and consistent with
the simultaneous presence of small micelles and unimers in solution.
QELS analyses then confirm the fluorescence results with a well-
defined system at pH ¼ 4.6 constituted of micelles, a transition state
atpH¼ 7, with simultaneouspresence ofmicelles andunimersand no
micelles anymore at pH ¼ 10.
ꢀ1
4.6, C w 0.12 g L , far above the cmc determined by fluorescence
s
ꢀ
1
measurements (10 mg L ). This difference demonstrates that the
copolymer self-associates to form micelles in solution before
saturation of the air/water interface and the gap between the two
concentrations (Cs and cmc) reflects the strong tendency of
copolymer 3c for self-assembly [34]. At pH 7, both values are close
ꢀ
1
ꢀ1
(C
s
w 0.15 g L , cmc ¼ 0.2 g L ) showing a preferential adsorption
of the copolymer at the air/water interface and hydrophobic clus-
ters formation only when the surface saturation is reached, simi-
larly to molecular surfactants. At pH 10, a really low adsorption is
observed and the surface saturation is not reached at 1 g L . This
result is in coherence with a previous study were the non-surface
activity of PAA at high ionization was reported [35]. As no hydro-
phobic cluster formation was detected either by fluorescence
measurements, the copolymer hydrophilicity seems to be high
enough to inhibit any self-association in bulk or interface adsorp-
tion in the range of the studied concentrations.
ꢀ
1
The air/water adsorption of copolymer 3c was also investigated
through surface tension measurements (buffered solutions þ 0.1
M of
NaCl) (Fig. 6). At pH ¼ 4.6, the surface tension of solutions of copol-
ymer 3c rapidly decreases proving the fast adsorption of the copol-
ymeratthe air/water interface. The surfacetension stabilizes oncethe
interfaceissaturated ata low value of
results in a slower copolymer adsorption and a decreased stabilizing
effect (
s
¼ 40mN/m. A pHincrease at 7
s
¼ 45 mN/m) after the surface saturation.
Finally, the influence of the hydrophilic block length on the
copolymer auto-association properties was investigated. Several
lipid-b-PAA copolymers with a hydrophilic/hydrophobic (wt/wt)
balance ranging from 38/62 to 80/20 were analysed by fluorescence
experiment to determine the cmc in acidic media (pH ¼ 4.6) (Table 2).
Another critical concentration can be defined from these plots,
at the intersection between the two linear portions of the curve,
and corresponding to the concentration of the air/water interface
saturation (C
interface is saturated, further increase in concentration leads to the
formation of micelles in bulk solution, C would thus correspond to
s
). For molecular surfactants, when the air/water
s
1
1
4
2
the cmc determined by fluorescence measurements. However the
Table 2
10
8
Characterization of a range of lipid-b-PAA copolymers (pH 4.6).
Xn;theoa
Mn;theo
(
b
Hydrophilic/
hydrophobic
balance (wt/wt)
Cmcc
(mg L
Cmcd
ꢀ1
mol L )
Ref.
g molꢀ1)
ꢀ1
)
(m
6
4
3
3
3
3
a
b
c
4
14
17
22
698
41/58
72/28
75/25
79/21
8.4
7.6
7.4
6
12.0
5.4
4.5
1418
1634
1994
2
d
3.0
0
a
b
c
Xn;theo ¼ ½monomerꢃ=½initiatorꢃ ꢂ conversion.
30
40
50
60
70
80
Mn;theo ¼ M
þ X
ꢂ MAA
lipoinitiator1
n;theo
Determined by fluorescence measurement using pyrene.
cmcmass
hydrophilic block weigth percentage
d
cmc ¼
.
Mn;theo
Fig. 7. Evolution of the cmc with the hydrophilic/hydrophobic balance.

4352
L. Hespel et al. / Polymer 53 (2012) 4344e4352
A linear dependence between the cmc and the hydrophilic/
Acknowledgements
hydrophobic balance was observed with a decrease in the cmc
when the hydrophilic block length increases (Fig. 7).
The authors thank the ERDF funding (ISCE e Chem & Interreg
IVa program 4061) for financial support.
The cmc evolution differs from the typical ionic and non-ionic
surfactants behaviour. The weak hydrophilicity of protonated PAA
probably results in a small effect of the block length on the
copolymer solubility as already described for PMMA-b-PDMAEMA
systems [36]. Simultaneously, longer PAA blocks will occupy more
space on the micelle surface decreasing the number of macro-
molecules required to form the micelle. Finally, increasing the PAA
block length leads to a predominance of the steric effect over the
solubility effect resulting in a cmc decrease. The synthetic pathway,
through a controlled polymerization technique, is thus of great
interest to accurately control the hydrophilic block length and
consequently adjust the auto-association properties.
Appendix A. Supplementary material
Supplementary data associated with this articlecan be found, in the
online version, at http://dx.doi.org/10.1016/j.polymer.2012.07.041.
References
[
[
1] Shi J, Votruba AR, Farokhzad OC, Langer R. Nano Lett 2010;10:3223e30.
2] Lasic DD, Papahadjopoulos D. Medical applications of liposomes. Amsterdam:
Elsevier Science B.V; 1998.
[3] Ta T, Convertine AJ, Reyes CR, Stayton PS, Porter TM. Biomacromolecules
010;11:1915e20.
2
[
[
[
[
[
[
4] Barenholz Y. Curr Opin Colloid Interface Sci 2001;6:66e77.
5] Weissig V, Whiteman K, Torchilin V. Pharm Res 1998;15:1552e6.
6] Lukyanov AN, Torchilin VP. Adv Drug Deliv Rev 2004;56:1273e89.
7] Kitano H, Akatsuka Y, Ise N. Macromolecules 1991;24:42e6.
8] Kitano H, Kuwayama M, Kanayama N, Ohno K. Langmuir 1998;14:165e70.
9] Kitano H, Ishino Y, Yabe K. Langmuir 2001;17:2312e6.
4
. Conclusion
To our best knowledge, this paper reports the first example of
ATRP initiator synthesized from naturally occurring vegetable oil.
The oil functionalization was conducted in two simple steps,
triglycerides reduction followed by esterification of the resulting
fatty alcohol to introduce the ATRP initiating site. Our original lip-
oinitiator leads to a well-controlled polymerization by ATRP and
a range of lipid-b-PtBA copolymers was synthesized with different
lengths of the synthetic block. Acidolysis of the PtBA block leads to
amphiphilic lipid-b-PAA copolymers. A copolymer with an hydro-
philic/hydrophobic ratio of 75/25 (wt/wt) was fully characterized
and interesting properties were detected. In acidic media, micelles
[10] Bathfield M, Daviot D, D’Agosto F, Spitz R, Ladaviere C, Charreyre M, et al.
Macromolecules 2008;41:8346e53.
11] Ohno K, Fukuda T, Kitano H. Macromol Chem Phys 1998;199:2193e7.
12] Gotz H, Harth E, Schiller S, Frank C, Knoll W, Hawker C. J Pol Sci Part A Pol
Chem 2002;40:3379e91.
[
[
[13] Théato P, Preis E, Brehmer M, Zentel R. Macromol Symp 2001;164:257e67.
[
[
[
14] Théato P, Zentel R. J Colloid Interface Sci 2003;268:258e62.
15] Théato P, Zentel R, Schwarz S. Macromol Biosci 2002;2:387e94.
16] Kitano H, Sumi Y, Tagawa K. Bioconjug Chem 2001;12:56e61.
[17] Janshoff A, Steinem C. Anal Bioanal Chem 2006;385:433e51.
18] Tanaka M, Sackmann E. Nature 2005;437:656e63.
19] Gandini A. Macromolecules 2008;41:9491e504.
20] Meier AR, Metzger JO, Schubert US. Chem Soc Rev 2007;36:1788e802.
[21] Chatterjee DP, Mandal BM. Polymer 2006;47:1812e9.
22] Radhunadh V, Baskaran D, Sivaram S. Polymer 2004;45:3149e55.
[23] Saikia PJ, Goswami A, Baruah SD. J Appl Polym Sci 2002;85:1236e45.
24] Xu W, Zhu X, Cheng Z, Chen JJ. Appl Polym Sci 2003;90:1117e25.
[
[
[
ꢀ1
ꢀ6
ꢀ1
form at a low cmc of 7.6 mg L (i.e. 5.3 ꢂ 10 mol L ) and a fast
air/water adsorption is observed leading to a good stabilization of
the interface. When the pH is raised, the cmc is highly increased
[
ꢀ1
ꢀ4
ꢀ1
(
w 0.2 g L , i.e. 1.2 ꢂ 10 mol L ) due to a modification of the
copolymer HLB and the air/water adsorption is slowed down. At pH
0, no self-association was detected anymore and the surface
[
[25] Loutelier-Bourhis C, Zovi O, Lecamp L, Bunel C, Lange C. Rapid Commun Mass
Spectrom 2009;23:3743e52.
1
[
[
26] Bromberg L, Temchenko M, Alakhov V, Hatton T. Int J Pharm 2004;282:45e60.
27] Letchford K, Burt H. Eur J Pharm Biopharm 2007;65:259e69.
[28] Gonnal LA, Li Q, Quinn JF, Tjipto E, Caruso F, Qiao GG. Macromolecules 2008;
adsorption was inhibited. Our original lipid-b-PAA copolymers
form well-defined and pH-sensitive micelles that start to dissociate
when the pH is raised above 7, which could be a stimuli to trigger
a drug release.
Finally it was demonstrated that the copolymer cmc can be
adjusted through modification of the hydrophilic/hydrophobic
balance further proving the interest of using a highly controlled
polymerization technique in the synthesis.
41:2620e6.
[
[
29] Wilhelm M, Zhao C, Wang Y, Xu R, Winnik MA. Macromolecules 1991;24:
033e40.
30] Bromberg LE, Barr DP. Macromolecules 1999;32:3649e57.
[31] Philippova OE, Hourdet D, Audebert R, Khokhlov AR. Macromolecules 1997;
0:8278e85.
1
3
[
[
32] Chen J, Jiang M, Zhang Y, Zhou H. Macromolecules 1999;32:4861.
33] Astafieva I, Zhong XF, Eisenberg A. Macromolecules 1993;26:7339e52.
This study demonstrates that well-defined micelles can be
synthesized from a heterogeneous natural product. Their ability to
encapsulate drugs as well as extension of this work to other
stimuli-responsive polymers are currently under investigation.
[34] Duval-Terrié C, Huguet J, Muller G. Colloids Surf A: Physicochem Eng Aspects
003;220:105e15.
2
[
35] Théodoly O, Jacquin M, Muller P, Chhun S. Langmuir 2009;25:781e93.
36] Xiao G, Hu Z, Zeng G, Wang Y, Huang Y, Hong X, et al. J Appl Polym Sci 2012;
124:202e8.
[