Novel malachite green- and rhodamine B-labeled cationic chain transfer agents for RAFT polymerization

Article abstract of DOI:10.1016/j.polymer.2011.10.041

Two novel cationic RAFT agents have been synthesized, one labeled with a Malachite Green (MG) dye and another with a Rhodamine B (RhoB) dye. MG-labeled dithiobenzoate (MGEDBA) was prepared in a straightforward manner after synthesis of MG-ethylammonium chloride that reacted with a precursor dithiobenzoate bearing an activated ester function. However, the analogous reaction with RhoB amino derivative led to a mixture of dithiobenzoate and thioamide derivatives. An alternative approach yielded the RhoB-labeled RAFT agent (RhoBEDBA) with complete conversion. The purification of these dye-labeled RAFT agents was very challenging because of their dual nature (aromatic and ionic). Both MGEDBA and RhoBEDBA were efficient RAFT chain transfer agents to control the polymerization of N,N-dimethylacrylamide (DMA). The resulting α-end-labeled MG- and RhoB-PDMA samples presented low dispersities (<1.2) and both chain-ends were preserved. Finally, we showed that the attachment of RhoB and MG to the PDMA polymer chain-end did not influence the photophysical properties of these dyes. Therefore, these new dye-labeled RAFT agents can be used to prepare various labeled polymers and especially water-soluble ones, to study their conformation and dynamics in solution or at interfaces using fluorescence methods, or as labeled probes for imaging and/or diagnosis purposes.

Full text of DOI:10.1016/j.polymer.2011.10.041

Polymer 52 (2011) 5933e5946
Contents lists available at SciVerse ScienceDirect
Polymer
journal homepage: www.elsevier.com/locate/polymer
Novel Malachite Green- and Rhodamine B-labeled cationic chain transfer agents
for RAFT polymerization
,
b
,
, ,
, ,1
Mariana Beija a 2, Carlos A.M. Afonso b 3, José Paulo S. Farinha b, Marie-Thérèse Charreyre a
,
*
*
José M.G. Martinho b
,
*
a Unité Mixte CNRS-bioMérieux, École Normale Supérieure de Lyon, 46 Allée d’Italie, 69364 Lyon, Cedex 07, France
b Centro de Química-Física Molecular and IN - Institute of Nanoscience and Nanotechnology, Instituto Superior Técnico, Av. Rovisco Pais 1, 1049-001 Lisbon, Portugal
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 7 July 2011
Received in revised form
23 September 2011
Accepted 20 October 2011
Available online 25 October 2011
Two novel cationic RAFT agents have been synthesized, one labeled with a Malachite Green (MG) dye and
another with a Rhodamine B (RhoB) dye. MG-labeled dithiobenzoate (MGEDBA) was prepared in
a straightforward manner after synthesis of MG-ethylammonium chloride that reacted with a precursor
dithiobenzoate bearing an activated ester function. However, the analogous reaction with RhoB amino
derivative led to a mixture of dithiobenzoate and thioamide derivatives. An alternative approach yielded
the RhoB-labeled RAFT agent (RhoBEDBA) with complete conversion. The purification of these dye-
labeled RAFT agents was very challenging because of their dual nature (aromatic and ionic). Both
MGEDBA and RhoBEDBA were efficient RAFT chain transfer agents to control the polymerization of N,N-
Keywords:
RAFT polymerization
Dye-labeled polymers
N,N-dimethylacrylamide
dimethylacrylamide (DMA). The resulting a-end-labeled MG- and RhoB-PDMA samples presented low
dispersities (Ð<1.2) and both chain-ends were preserved. Finally, we showed that the attachment of
RhoB and MG to the PDMA polymer chain-end did not influence the photophysical properties of these
dyes. Therefore, these new dye-labeled RAFT agents can be used to prepare various labeled polymers and
especially water-soluble ones, to study their conformation and dynamics in solution or at interfaces using
fluorescence methods, or as labeled probes for imaging and/or diagnosis purposes.
Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction
is highly desirable[10], especially for the determination of distances at
the macromolecular level using fluorescence techniques such as non-
Dye-labeling of polymers is a useful strategy in macromolecular
science, both to study chain dynamics including coil-to-globule
transitions [1], intra- and inter-molecular associations [2,3], end-to-
end cyclizations [4] or conformations of polymer chains in thin films
[5], and for applications such as imaging [6,7], signal amplification for
diagnostic tests [8] and light-harvesting or antennae [9]. Knowing the
exact location of the dye along the polymer chain is a key point in
several of these studies. Therefore, dye-labeling at specific positions
radiative Förster Resonance Energy Transfer (FRET) [11].
Recently, with the advances in controlled radical polymerization
(CRP) methods, it has been possible to prepare a large range of
polymerswithnotonlycontrolledmolecularweightandarchitecture
but also precise end-group functionality. Among these techniques,
the reversible addition-fragmentation chain transfer (RAFT) [12,13]
polymerization is one of the most versatile. RAFT polymerization
can be described as a conventional free radical polymerization that is
conducted in the presence of a suitable chain transfer agent (CTA)
from the family of thiocarbonylthio compounds (with the general
formula Z-C(¼S)eSR). It operates on the principle of degenerative
chaintransfer, inwhicha rapidequilibrium betweenpropagatingand
dormant chains is established, leading to the same probability of
growth of all chains, hence yielding well-defined polymers. An
important feature of polymer chains synthesized by RAFT polymer-
ization is the presence of the R and Z groups from the original CTA at
suchasthepolymer chain-end (a or u) or the junction between blocks
* Corresponding authors. Tel.: þ33 (0) 4 72 72 89 38; fax: þ33 (0) 4 72 72 87 87.
E-mail addresses: carlosafonso@ff.ul.pt (C.A.M. Afonso), marie-therese.
charreyre@ens-lyon.fr (M.-T. Charreyre), jgmartinho@ist.utl.pt (J.M.G. Martinho).
1
Present address: Laboratoire Joliot-Curie, USR CNRS3010, ENS de Lyon, 46 Allée
d’Italie, 69364 Lyon, Cedex 07, France and Laboratoire Ingénierie des Matériaux
Polymères, UMR CNRS5223, INSA-Lyon, 69621 Villeurbanne Cedex, France.
their a- and u-ends, respectively (with the exception of symmetric
2
Present address: Centre for Advanced Macromolecular Design (CAMD), School
trithiocarbonates, in which Z is in the middle of the chain).
A widespread methodology for labeling RAFT polymers takes
advantage of the terminal thiocarbonylthio moiety [14], that can be
of Chemical Engineering, UNSW, Sydney, NSW 2052, Australia.
3
Present address: Faculty of Pharmacy, University of Lisbon, Av. Prof. Gama Pinto,
1649-019 Lisbon, Portugal.
0032-3861/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.polymer.2011.10.041

5934
M. Beija et al. / Polymer 52 (2011) 5933e5946
converted into a thiol through aminolysis [15,16] or reduction
[17,18]. Then, reaction with a dye derivative bearing a complemen-
tary function (such as maleimide [15,17] or iodide [18]) is carried
out, leading to dye-end-labeled polymers. However, large amounts
of dye derivatives have usually to be used, especially for long
polymer chains (otherwise low labeling yields are expected),
implying large costs.
Another common strategy consists in the use of dye-labeled
CTAs (modified either at R or Z). Several kinds of CTAs bearing
dyes, such as naphthalene [19,20], acenaphtylene [21], anthracene
[9,22e24], phenanthrene [25,26], pyrene [27], diphenylanthracene
[28], carbazole [29e31], indole [31], spirooxazine [32], naph-
talenesulfonate [33], coumarin [34] or Ru(II) polypyridine [35,36],
can be found in the literature as recently reviewed by Beija et al.
[10]. In order to limit the possibility of dye loss in the final polymer,
the link between the dye and the polymer has to be sufficiently
robust. However, many of the above examples correspond to Z-
modified CTAs that are very sensitive since the dithioester linkage
can be easily degraded by hydrolysis, nucleophilic attack and
radical mechanisms. In addition, some of the R-modified CTAs
present an ester linkage, fragile in basic aqueous media.
Moreover, in all these examples (with the exception of naph-
talenesulfonate), only hydrophobic dyes have been employed
(mainly from the family of polycyclic aromatic hydrocarbons). This
is a limitation for water-soluble polymers, since it is known that the
presence of a dye attached to a polymer chain can induce important
changes in polymer chain conformation and properties [37] espe-
cially in aqueous solutions. In addition, hydrophobic dyes tend to
form dimers, causing significant changes on the photophysical
properties of the labeled polymer chains. Although desirable for
labeling water-soluble polymers, the synthesis of hydrophilic dye-
labeled RAFT agents is indeed very challenging. Usually, water-
soluble dyes are ionic aromatic molecules, thus amphiphilic in
nature, which renders their purification extremely difficult.
From a photophysical point of view, most of the dyes used for
the synthesis of RAFT agents absorb in the UV. Nevertheless, for
biomedical applications such as imaging or fluorescence detection,
it is desirable to have dyes that absorb in the visible or near-infrared
(NIR) wavelength range, where light penetration in tissues is
deeper while being less harmful to cells.
solution or on surfaces and interfaces. Indeed, we used the RhoB-
labeled CTA to synthesize RhoB -end-labeled thermosensitive
a
block copolymers via the RAFT process. Then, we studied the phase
separation in LangmuireBlodgett films of these copolymers. The
changes in the film structure could be very well characterized by
laser scanning confocal fluorescence microscopy which had not
been possible using atomic force microscopy [46]. In the present
paper, we focus on the synthesis of the MG- and RhoB-labeled CTAs,
their behavior in RAFT polymerization and the photophysical
characteristics of both the dye-labeled CTAs and the resulting
end-labeled PDMA homopolymers.
a-
2. Experimental section
2.1. Materials
Rhodamine B (95%, Sigma) and 4,40-bis(dimethylamino)benzo-
phenone (Michler’s ketone; 98%, Aldrich) were used as received.
Before each use, n-butyllithium (n-BuLi) was titrated with menthol
and 2,20-dipyridyl in dry THF. N,N-dimethylacrylamide (DMA) was
obtained from Aldrich (99%) and purified by distillation at 24 ꢀC and
0.4 mbar in the presence of hydroquinone. The initiator 2,20-azo-
bis(isobutyronitrile) (AIBN) (Fluka, 98%) was purified by recrystal-
lization from ethanol. All other reactants were purchased from
Aldrich and used without further purification. Synthesis grade
solvents were distilled before use. Tetrahydrofuran (THF),
dichloromethane and N,N-diisopropylethylamine (DIPEA) were
dried by distillation under argon in the presence of CaH2. Dime-
thylsulfoxide (DMSO) (Aldrich, anhydrous, 99.9%), N,N-dime-
thylformamide (DMF) (Fluka, 99.9%), chloroform-d (CDCl3; þ1%
TMS, SDS), acetone-d6 (SDS), methanol-d4 (Aldrich), deuterium
oxide (D2O; Aldrich, 99.9%, 0.75 wt% TSP), DMSO-d6 (Aldrich,
99.96%), lithium bromide (Merck) were used as received. The CTAs
succinimidoxycarbonyl ethyldithiobenzoate (SEDB, 1) [47] and tert-
butyl dithiobenzoate (tBDB) [48] were synthesized according to
reported procedures.
Flash column chromatography was performed with silica gel 60
(0.040e0.063 mm) purchased from Merck and MN AlugramÒ SIL G/
UV254 plates were used for analytical thin layer chromatography.
The plates were visualized at 254 nm or at 430 nm (for rhodamine
derivatives). All reactions were carried out under argon or nitrogen
atmosphere and glass material was oven dried at 120 ꢀC prior to use.
Here, we present the synthesis of two novel R-modified CTAs
functionalized with the dyes Rhodamine B (RhoB) and Malachite
Gren (MG) through a robust amide bond, and their application for
the synthesis of
a-dye-labeled poly(N,N-dimethyl acrylamide)
2.2. Synthesis of dye-labeled CTAs
(PDMA) polymer chains. Both of these dyes are cationic, thus water-
max
soluble. RhoB absorbs and emits in the red (
l
¼ 572 nm), pre-
2.2.1. Malachite Green-labeled CTA
2.2.1.1. Malachite Green-ethylammonium chloride (2). p-bromo-
phenylethyl-1-amine (100 mL, 129 mg, 0.65 mmol) was dissolved in
em
senting a high fluorescence quantum yield (4F ¼ 0.53) [38]. MG is
a non-fluorescent dye under common experimental conditions that
max
abs
shows a high molar absorption coefficient (
l
¼ 616.5 nm;
dry THF (4 mL) and cooled to ꢁ78 ꢀC. A solution of n-BuLi
(1.3 mmol) was then added dropwise and after ca. 60% of addition,
a white precipitate was formed. The reaction mixture was stirred
for 2 h. Subsequently, a solution of Michler’s ketone (86.6 mg,
0.32 mmol) in dry THF (4 mL) was added dropwise and the reaction
mixture was let to react overnight. The reaction was quenched by
addition of concentrated HCl aqueous solution. The aqueous layer
was washed with hexane (3 ꢂ 20 mL) and the water was further
removed by evaporation under vacuum. Precipitation from ethanol
in diethyl ether yielded 146 mg of 2 as a highly hygroscopic green
solid compound.
max
3
¼ 148 900) [39]. Besides, rhodamine and MG dyes form an
eaxbcsellent donor/acceptor pair for FRET studies, presenting a high
Förster critical radius (around 61 Å) [40]. Another advantage of this
pair is that MG is a non-fluorescent probe, avoiding the contami-
nation of rhodamine emission and leading to a dark background,
which is an ideal situation for both photophysical studies and
imaging purposes. Thus, this pair has been used for the surface
characterization of polymer nanoparticles [41,42] and to study DNA
hybridization on the surface of thermoresponsive polymer nano-
particles [43e45]. To the best of our knowledge, this is the first time
that the synthesis of cationic dye-labeled CTAs suitable for RAFT
polymerization, absorbing in the visible region of electromagnetic
spectrum and forming a FRET donor/acceptor pair is described.
These new dye-labeled RAFT agents open up the possibility to
design several polymer architectures and to further investigate by
FRET (and other fluorescence techniques) their dynamics in
(See Supporting Information, Table S1, for assignment)
1H NMR (300 MHz, methanol-d4):
d
(ppm) ¼ 3.00 (t, 2H6,
J ¼ 7.1 Hz), 3.14 (t, 2H7, J ¼ 7.4 Hz), 3.28 (s, 12H1), 6.97e7.82 (m, 4H2,
4H3, 2H4, 2H5).
13C NMR (75 MHz, methanol-d4):
d
(ppm) ¼ 33.79 (C11), 34.36
(C12), 41.09 (C1), 114.86 (C3), 116.93 (not assigned), 121.67, 128.18,

M. Beija et al. / Polymer 52 (2011) 5933e5946
5935
128.45 (Cq), 129.80, 129.93, 130.30, 130.82, 131.55, 131.87, 132.96,
133.17, 136.32, 137.98 (Cq), 139.92 (Cq) 141.88, 158.48 (C6).
MS (ESI): m/z (%) ¼ 424.5 (15.7), 372.2 (54.1) [(M-H)þ], 333.08
(16.7), 295.2 (58.2), 269.2 (100), 253.2 (31.3).
obtained solid was partitioned between EtOAc and a diluted solu-
tion of NaHCO3. The aqueous layers were combined and washed
three times with EtOAc to remove the remaining Rhodamine B
lactone, then saturated with NaCl and acidified with 1 mol Lꢁ1 HCl.
Rhodamine B piperazine amide was extracted several times with
a mixture of dichloromethane/i-propanol (2:1) until a faint pink
color persisted in the aqueous layer. The combined organic layers
were dried over Na2SO4, filtered and concentrated under reduced
pressure. After precipitation in diethyl ether from a methanol
solution, 617.5 mg of 4 was recovered by filtration as a dark purple
solid (52%).
HRMS (ESI): m/z calcd for C25H30Nþ3 : 372.2434, found: 372.2442.
2.2.1.2. Malachite Green ethyl dithiobenzoate amide (MGEDBA,
3). To a stirred solution of SEDB (1, 206.9 mg, 0.64 mmol) and
DIPEA (230 mL, 171.2 mg, 1.32 mmol) in anhydrous DMF (30 mL),
a solution of MG-ethylammonium chloride (279.7 mg, 0.63 mmol)
in DMF was added dropwise at 30 ꢀC. After 2 h 30 min, additional
DIPEA (150
mL, 111.3 mg, 0.86 mmol) was added and the reaction
Spectral data similar to those reported (1H NMR, 500 MHz,
methanol-d4 and 13C NMR, 100 MHz, methanol-d4) [49]. (See Sup-
porting Information, Table S1, for assignment.)
mixture was stirred overnight. After solvent removal, the mixture
was chromatographed on a silica gel column (eluent: CH3CN/H2O
10:0 to 9:1) to afford 199 mg of 3 as a green product.
1H NMR (300 MHz, methanol-d4):
d
(ppm) ¼ 1.34 (t, 12H1,
1H NMR (200 MHz, DMSO-d6; see supporting information,
J ¼ 7.0 Hz), 3.15 (br s, 4H11), 3.69-3.74 (m, 8H2 and 4H10), 7.00 (d,
2H3, J ¼ 1.9 Hz), 7.12 (dd, 2H4, J ¼ 1.8, 9.5 Hz), 7.29 (d, 2H5,
J ¼ 9.5 Hz), 7.54-7.57 (m, 1H9), 7.77-7.84 (m, 1H6, 1H7, 1H8).
Figure S1):
d
(ppm) ¼ 1.51 (d, 3H5’, J ¼ 7.1 Hz); 3.11-3.15 (m, 2H6);
3.29 (s, 12H1); 3.53-3.63 (m, 2H7); 4.59 (q, 1H4’, J ¼ 7.0 Hz); 7.03-
7.66 (m), 7.87-7.91 (m) (4H2, 4H3, 2H4, 2H5, 1H1’, 2H2’, 2H3’).
13C NMR (50 MHz, CDCl3; see supporting information,
13C NMR (75 MHz, methanol-d4):
d
(ppm) ¼ 12.93 (C1), 44.31,
45.70 (C17, C18), 46.98 (C2), 97.43 (C4), 114.85 (C6), 115.55 (C8),
128.93, 131.45, 131.65, 131.92, 132.48, 133.06, 135.63, 156.73, 157.26,
159.26, 169.52 (C16).
Figure S2):
d
(ppm) ¼ 16.76 (C7’); 38.73 (C11); 40.55 (C12); 41.04 (C1);
111.71, 113.58 (C3); 126.97, 127.09, 127.24, 128.03, 128.31, 128.39,
128.44, 128.76, 129.43, 130.88, 135.22, 137.38 (C4, C5, C7-C10, C1’-C3’);
144.31 (C4’); 146.17, 149.44 (C2); 156.81 (C6); 170.78 (C13); 227.08
(C16). Note: It is probable that C4 and C5 present more than one
peak.
MS (ESI): m/z (%) ¼ 511.2 (100) [(M - H)þ].
2.2.2.3. Rhodamine
5). Method 1: A solution of Rhodamine B piperazine amide
(91.5 mg, 0.16 mmol) and DIPEA (55 L, 0.32 mmol) in dichloro-
B ethyl dithiobenzoate amide (RhoBEDBA,
MS (ESI; see supporting information, Figure S3): m/z (%) ¼ 580.2
m
(100) [Mþ], 564.2 (17) [(M-SþO)þ], 476.2 (16).
methane was added dropwise to a solution of SEDB (1, 51.1 mg,
0.16 mmol) in dichloromethane (20 mL) and the mixture stirred for
18 h at 30 ꢀC. Then, it was washed with acidified water (3 ꢂ 40 mL),
dried over Na2SO4, filtered and the solvent was evaporated. The
obtained solid was redissolved in methanol and precipitated by the
addition of diethyl ether. After filtration, 71.8 mg of a dark purple
solid were recovered. Characterization techniques showed also the
presence of RhoB piperazine thioamide (6) with a molar fraction of
50% (yield of 5: 31%; see spectral data of pure 5 in method 2).
MS (ESI) of the mixture of 5 and 6: m/z (%) ¼ 719.2 (100) [5, Mþ],
703.1 (16) [6, (M-SþO)þ], 631.2 (67) [6, Mþ]
2.2.2. Rhodamine B-labeled CTA
2.2.2.1. Rhodamine B lactone. Followed reported procedure [49].
Rhodamine B (1.5 g, 3.1 mmol) was dissolved in 1 mol Lꢁ1
aqueous NaOH solution (100 mL) and stirred for 2 h. Thereafter, it
was partitioned with ethyl acetate (75 mL). The organic layer was
isolated and the aqueous layer was extracted twice with ethyl
acetate (EtOAc). Combined organic layers were washed once with
1 mol Lꢁ1 NaOH and then brine. The organic solution was dried
with Na2SO4, filtered, concentrated under reduced pressure and
dried to yield 1.3 g of Rhodamine B lactone as a pink product (94%).
Spectral data similar to those reported (1H NMR, 300 MHz,
methanol-d4) [49].
Method 2: Rhodamine
0.34 mmol) and DIPEA (150
in dry dichloromethane (5 mL). Then,
B
piperazine amide (198.5 mg,
L, 0.86 mmol, 2.5 eq) were dissolved
-bromopropanoyl bromide
m
a
(See Supporting Information, Table S1, for assignment)
was added dropwise to previous solution at 0 ꢀC during 10 min.
After reacting for 2 h 30 min, the reaction mixture was diluted with
dichloromethane (20 mL), washed with saturated Na2CO3 solution
(3 ꢂ 40 mL, pHw12), then with saturated NH4Cl solution (40 mL),
dried (Na2SO4), filtered and concentrated under vacuum to afford
185 mg of Rhodamine B halopropanamide (7) as a dark purple solid
(mixture 1:1 of bromo- and chloro- derivatives calculated by 1H
NMR; 82% yield).
1H NMR (300 MHz, acetone-d6):
d(ppm) ¼ 1.15 (t, 12H1,
J ¼ 7.0 Hz), 3.41 (q, 8H2, J ¼ 7.0 Hz), 6.43 (dd, 2H4, J ¼ 2.6, 8.7 Hz),
6.46 (d, 2H3, J ¼ 2.3 Hz), 6.54 (d, 2H5, J ¼ 8.7 Hz), 7.23 (br d, 1H9,
J ¼ 7.5 Hz), 7.68 (ddd, 1H7, J ¼ 1.0, 7.4, 7.4 Hz), 7.76 (ddd, 1H8, J ¼ 1.2,
7.4, 7.4 Hz), 7.96 (br d, 1H6, J ¼ 7.4 Hz).
13C NMR (75 MHz, acetone-d6):
86.46 (C9), 99.15 (C4), 107.91 (C6), 109.95 (C8), 125.91, 126.01, 129.57,
130.56, 131.29, 136.54, 151.27, 154.96, 155.92, 170.66 (C16).
d
(ppm) ¼ 13.78 (C1), 45.92 (C2),
(See supporting information, Table S1, for assignment).
1H NMR (200 MHz, acetone-d6):
d(ppm) ¼ 1.26 (t, 12H1,
2.2.2.2. Rhodamine
B
piperazine amide (4). Followed reported
J ¼ 5.5 Hz), 1.51 (d, 3H5’, X ¼ Cl, J ¼ 6.4 Hz), 1.62 (d, 3H5’, X ¼ Br,
J ¼ 6.9 Hz), 3.16 (br s, 4H11), 3.38-3.54 (m, 4H10), 3.73-3.77 (br m,
8H2), 4.18 (q,1H4’, X ¼ Br, J ¼ 7.1 Hz), 4.55 (q,1H4’, X ¼ Cl, J ¼ 6.9 Hz),
6.91 (br, 2H3), 7.28 (br m, 2H4 and 2H5), 7.50 (br m, 1H9), 7.75 (br m,
1H6, 1H7, 1H8).
procedure [49].
A 2 mol Lꢁ1 solution of trimethyl aluminium in hexanes (2 mL,
4.08 mmol) was added dropwise to a solution of piperazine
(707.9 mg, 8.2 mmol) in dry dichloromethane (4.5 mL). The reaction
mixture was stirred for 1 h until formation of a white precipitate.
Then, a solution of Rhodamine B lactone (903.5 mg, 2.04 mmol) was
added dropwise to the previous mixture. After reacting under
reflux for 24 h, the mixture was quenched by addition of
a 0.1 mol Lꢁ1 HCl aqueous solution. The heterogeneous solution was
filtered and the retained solids were washed with dichloromethane
and then with a mixture dichloromethane/methanol (4:1). The
combined filtrated was concentrated, resolubilized in dichloro-
methane, filtered to remove exceeding salt and reconcentrated. The
13C NMR (50 MHz, acetone-d6):
d
(ppm) ¼ 13.98 (C1), 22.26 (C21),
42.92, 43.72 (C17, C18), 45.79 (C4’, X ¼ Br), 47.57 (C2), 56.16 (C4’,
X ¼ Cl), 97.73 (C4), 115.16, 115.29 (C6), 116.32 (C8), 129.53, 130.37,
131.47, 131.59, 132.08, 132.96, 133.78, 157.36, 157.52, 157.71, 168.77
(C16), 169.17 (C8’, X ¼ Cl), 170.84 (C8’, X ¼ Br).
MS (ESI): m/z (%) ¼ 601.3 (100) [M(X ¼ 35Cl)þ], 603.3 (33)
[M(X ¼ 37Cl)þ], 645.1 (20) [M(X ¼ 79Br)þ], 647.1 (20) [M(X ¼ 81Br)þ].
Carbon disulfide (220 mL, 277 mg, 3.6 mmol) was added drop-
wise to a solution of phenylmagnesium bromide (3.6 mmol) in dry

5936
M. Beija et al. / Polymer 52 (2011) 5933e5946
THF (5.6 mL) at 0 ꢀC. After stirring for 1 h 30, a 2 mol Lꢁ1 HCl
aqueous solution (4 mL) was added to the reaction mixture and
extracted with diethyl ether (10 mL). Sodium dithiobenzoate was
extracted from organic layer with a 1 mol Lꢁ1 NaOH aqueous
solution. The dithiobenzoate concentration was adjusted to
0.2 mol Lꢁ1 and the solution pH was adjusted to 6. To a stirred
solution containing Rhodamine B propanoyl piperazine amide
(185 mg, 0.28 mmol) and a catalytic amount of NaI (15 mg,
0.1 mmol) in 5 mL of dichloromethane, 5 mL of sodium dithio-
benzoate solution was added and the mixture was vigorously stir-
red overnight. The reaction mixture was washed with a borate
buffer solution (pH ¼ 9), then brine. The organic layer was dried
over Na2SO4, filtered and concentrated under reduced pressure.
After purification by dissolution in methanol, followed by precipi-
tation in diethyl ether and silica gel chromatography (CH2Cl2/
methanol 10:0 to 8:2), 5 was obtained as a purple powder in 58%
yield.
reference) at initial polymerization time (t ¼ 0) and at polymeri-
zation time, t.
2.4. Purification of polymer samples
PDMA polymer chains were purified by dialysis using Slide-A-
LyzerÒ dialysis cassettes from Pierce with a molecular weight cut-
off (MWCO) of 3500 Da or 2000 Da (polymers with
Mn < 10 000 g molꢁ1). Before dialysis, samples were diluted 10
times with milliQ water in order to have a maximum of 10% v/v
DMSO content. Typically, after introduction of the sample, the
cassette was left to float in a water bath (500 times the volume of
the sample). After 2 h dialysis, water was changed and dialysis
proceeded for another 2 h. Afterwards, water was changed again, at
least once a day, until all low molecular weight molecules have
been removed (confirmation by 1H NMR and/or SEC analyses). The
sample was recovered and freeze-dried in a Lyoalfa 6-80 Telstar
(Varian DS 202) equipment.
1H NMR (300 MHz, acetone-d6, see supporting information,
Figure S4):
d
(ppm) ¼ 1.29 (t, 12H1, J ¼ 7.0 Hz), 1.54 (d, 3H5’,
J ¼ 6.4 Hz), 2.99 (br s, 4H11), 3.38-3.57 (m, 4H10), 3.76 (br s, 8H2),
5.09 (q, 1H4’, J ¼ 6.5 Hz), 6.96 (d, 2H3, J ¼ 2.2 Hz), 7.24 (dd, 2H4,
J ¼ 2.3, 9.5 Hz), 7.34 (d, 2H5, J ¼ 9.1 Hz), 7.47 (dd, 2H2’, J ¼ 7.6, 7.8 Hz),
2.5. Characterization
2.5.1. Nuclear magnetic resonance
7.54-7.58 (m, 1H9), 7.75-7.80 (m, H12, H13, H14), 7.96 (d, 2H24
,
1H and 13C NMR spectra were recorded on a Bruker AC 200
spectrometer (200 MHz and 50 MHz, respectively), on a Bruker
Avance II (300 MHz and 75 MHz, respectively) and, for polymer
characterization, on a Bruker Avance II (400 MHz, 40 ꢀC). Chemical
J ¼ 7.1 Hz).
13C NMR (75 MHz, acetone-d6, see supporting information,
Figure S5):
d(ppm) ¼ 13.92 (C1); 18.26 (C7’); 43.26 (C18); 47.62 (C2);
48.86, 49.22 (C17, C6’); 97.89 (C4); 115.45, 115.50 (C6); 116.33 (C8);
128.52, 129.65, 131.58, 131.74, 132.23, 133.07, 134.06, 134.77, 137.63,
157.59, 158.15, 159.55 (C3, C5, C7, C9-C15, C23-C26); 168.83 (C16);
170.19 (C8’); 228.69 (C5’).
shifts are reported as d values relative to tetramethylsilane
(
dH ¼ 0 ppm). Signal multiplicities are designed as follows:
s e singlet, d e doublet, t e triplet, q e quadruplet, m e multiplet,
br e broad.
MS (ESI): m/z (%) ¼ 719.2 (100) [Mþ]
HRMS (FAB, see supporting information, Figure S6): m/z calcd
for C42H47N4O3Sþ2 : 719.308960; found: 719.311351.
2.5.2. Mass spectrometry
Low resolution mass spectra were carried out at the Institut de
Biologie et Chimie de Protéines (IBCP, Lyon, France) using either
electrospray ionization (ESI) or matrix-assisted laser desorption/
ionization time-of-flight (MALDI-ToF) mass spectrometry tech-
niques with positive ions detection. ESI mass spectra were recorded
on an API 165 (Applied Biosystems) spectrometer (ion spray
voltage: 5 kV), using a quadrupole scan speed of 100 Da/s and
2.3. Polymerization procedures
Typically, monomer(s), initiator, solvent, CTA and trioxane (used
as internal reference for monomer conversion) were introduced
either in an NMR tube fitted with a Young’s valve or a Schlenk tube
equipped with a magnetic stirrer. The polymerization mixture was
then degassed by five freeze-pump-thaw cycles and set under
nitrogen atmosphere.
a flow rate of 5 mL/min; samples were dissolved in methanol/water/
formic acid (50/50/0.1). Otherwise, MALDI-ToF mass spectra were
recorded on a Voyager DE-PRO (Applied Biosystems) spectrometer
with an acceleration voltage of 20 kV and with an extraction delay
For experiments carried out in an NMR tube, a 1H NMR spec-
trum of the mixture was performed at 27 ꢀC prior to polymeri-
zation initiation. After removing the NMR tube, the cavity of the
magnet was heated at the required temperature and allowed to
stabilize for 30 min. The actual temperature was checked by the
of 140 ns;
100 L of acetonitrile/water/TFA (50/50/0.1) mixture was used as
matrix.
Alternatively, high- and low-resolution mass spectra (Electronic
a-cyano-4-hydroxycinnamic acid (0.5 mg) dissolved in
m
chemical shift difference of protons of
a
(80:20) ethyl-
Impact - EI, ESI or Fast Atom Bombardment - FAB) were carried out
by the mass spectrometry service of the University of Santiago de
Compostela (Spain).
eneglycol:DMSO mixture [50,51]. Afterwards, the NMR tube con-
taining the polymerization mixture was reintroduced and let to
equilibrate for 2 min, before the magnet was shimmed. Monomer
conversions were periodically (3e15 min) monitored by on-line 1H
NMR spectroscopy.
Alternatively, for experiments carried out in a Schlenk tube,
a first sample was withdrawn from the polymerization medium at
room temperature and then, polymerization was thermally initi-
ated by introducing the Schlenk tube in a thermostated oil bath at
the required temperature. Samples were periodically withdrawn
from the polymerization medium and quenched in liquid nitrogen
for determination of monomer conversion by 1H NMR. Usually,
2.5.3. Organic size exclusion chromatography
Polymer samples were characterized by SEC-MALS using
a Waters 510 pump fitted with three Phenogel columns (104, 103
and 102 Å; 5
m
m; 7.8 ꢂ 300 mm; columns temperature: 35 ꢀC)
connected in series and a miniDAWN-treos multi-angle laser light
scattering detector (three angles: 45.8ꢀ, 90ꢀ and 134.2ꢀ; laser
wavelength: 658 nm) from Wyatt Technologies and a Waters 2410
refractive index detector at 30 ꢀC. Simultaneously, detection of
either fluorescence (Waters 478 detector) or absorption (Waters
486 detector) of polymer samples was performed to confirm the
400
mL of chloroform-d or acetone-d6 were added to 200 mL of each
sample.
existence of the chromophore/fluorophore at the a-chain-end. MG-
Determination of monomer conversion was achieved by
comparison between the integral values of the monomer(s) vinyl
protons and the trioxane protons (singlet peak, used as internal
PDMA samples were observed at an absorption wavelength of
600 nm and RhoB-PDMA samples were detected by fluorescence
(lexc ¼ 560 nm; lem ¼ 575 nm). A solution of 0.05 M LiBr in DMF

M. Beija et al. / Polymer 52 (2011) 5933e5946
5937
was used as eluent at a flow rate of 0.3 mL minꢁ1. Typically, 50
m
L of
the amino group and the dye moiety (at least a C2 group) in order to
decrease steric hindrance that could influence the control of the
polymerization: indeed, during the pre-equilibrium, while steric
hindrance on the R group favors the fragmentation step, it disfavors
the addition step (limited access to the thiocarbonylthio group) and
the re-initiation step; and 3) the bond between the linker and the
dye should be strong, such as a C-C or an amide bond, in order to
resist to hydrolysis.
a sample solution (10 g Lꢁ1) was injected in the system. Prior to
injection, these solutions were centrifuged at 15 000 rpm for
15 min. Relative molecular weights were obtained using a calibra-
tion curve with N-succinimidyl propionate-PDMA and t-butyl-
PDMA samples as PDMA standards.
2.5.4. Determination of specific refractive index increments
The specific refractive index increment (dn/dC) for PDMA was
determined in 0.05 M LiBr DMF solution using a WYATT Optilab rEX
detector operating at 658 nm and 30 ꢀC (0.0813 ꢃ 0.0005 mL gꢁ1).
3.1.1. Synthesis of MG-labeled CTA
In the literature, MG derivatives have mainly been synthesized
using three methods: Baeyer condensation [54e57], addition of
organolithium compounds [55,58e62] or addition of Grignard
reagents [55,63]. Often, the second method has been chosen since
softer experimental conditions can be employed compared to the
first method and higher yields are obtained compared to the latter.
Moreover, forour purpose, the introductionof a linker should impart
the lowest modifications to the MG spectral properties, which can be
achieved with a C-C bond [55]. Then, we decided to synthesize the
MG amino derivative by the reaction of Michler’s ketone (a
2.5.5. Absorption and fluorescence
UVeVisible absorption measurements were performed in
a Shimadzu Model 3101 spectrometer and fluorescence spectra
were recorded on a SPEX Fluorolog F112A fluorometer (bandwidth
of 0.9 nm (excitation) and 0.45 nm (emission)). Emission spectra of
RhoB derivatives were obtained between 535 nm and 750 nm,
using 530 nm as excitation wavelength. All spectra were obtained
with right angle geometry, at room temperature and using quartz
cells of 0.5 cm ꢂ 0.5 cm.
commercial compound) and
compound obtained from a halogen-lithium exchange reaction on
a commercial p-bromophenylethyl-1-amine (Scheme 2).
a p-functionalized aryllithium
Fluorescence intensity decay curves, with picosecond resolu-
tion, were obtained by the single-photon timing technique using
laser excitation of 570 nm. The system consists of a mode-locked
Coherent Inova 440-10 argon ion laser synchronously pumping
a cavity dumped Coherent 701-2 dye laser using Rhodamine 6G,
which delivers 3e4 ps pulses at a repetition rate of 4 MHz. The
fluorescence was observed at 585 nm using a polarizer set at the
magic angle and a cut-off filter to effectively eliminate the scattered
light. The fluorescence was selected by a Jobin-Yvon HR320
monochromator with a grating of 100 lines/mm and detected by
The main advantage of using p-bromophenylethyl-1-amine as
a reactant was the possibility of preparing directly an amino
derivative of MG. Conversely, the presence of a free amino group
could favor some side reactions such as the nucleophilic attack of
the amine to Michler’s ketone. A prior protection of the amino
group with a silicon-based protecting group developed by Djuric
et al. [64] or with a trityl group was attempted but the MG amino
derivative could not be obtained in significant yield. Nonetheless,
the MG-ethylammonium chloride 2 was successfully synthesized in
a one-pot procedure, without the need to protect the amino group
when 2 eq. of n-BuLi was added during the halogen-metal
exchange step. Indeed, despite the risk of competitive reaction of
the amine group with Michler’s ketone, the carbanion is much
more reactive and the formation of the desired compound is
preferred.
a
Hamamatsu 2809U-01 microchannel plate photomultiplier.
Time-scales of 7e12 ps/channel were used. The decay curves were
analyzed using home-made non-linear least-square reconvolution
software based on the Marquard algorithm [52]. All measurements
were performed at room temperature.
3. Results and discussion
The purification of compound 2 was very challenging either
when it was isolated as a carbinol (by a neutral work-up) or in
chromatic form (by an acid work-up). In the first case, we tried to
purify the carbinol using silica gel or reverse-phase chromatog-
raphy, but it led to degradation of the product. Instead, when the
product was precipitated and filtered, we were able to eliminate the
residual Michler’s ketone, but there were still unidentified impu-
rities in the final product. In the second case, Michler’s ketone could
be removed by extraction with hexane or diethyl ether from an
aqueous solution of 2. However, it was not possible to find a suffi-
ciently polar eluent to perform a chromatographic purification and
recrystallization was unable to eliminate all impurities. It is note-
worthy that non-functionalized MG is sold with only 85% purity.
Hence, although compound 2 was not completely pure, we decided
to perform the reaction with 1 and to purify the resulting MG-
labeled CTA.
3.1. Synthesis and characterization of dye-labeled CTAs
In order to synthesize the new dye-labeled CTAs, a convergent
strategy was chosen involving a coupling reaction between an
amino derivative of the dye and a precursor dithioester 1, according
to Scheme 1. This strategy was first developed for the synthesis of
bio-related dithiobenzoates carrying a sugar [47], a biotin [47] or
a phospholipid [53] and has already been used to get a fluorescent
phenanthrene-labeled CTA [26]. However, this is the first time that
it is applied to synthesize cationic dye-labeled CTAs, which repre-
sents a major challenge. Indeed, very few CTAs of this kind are
described in the literature. Prior to the coupling reaction, both MG
and RhoB amino derivatives were synthesized since they were not
commercially available or analogous were too expensive. Several
aspects had to be taken into consideration for the choice of an
appropriate synthetic pathway for these dye derivatives: 1) the dye
should present a free amino group for further reaction with the
precursor dithiobenzoate 1; 2) a linker should be present between
The coupling between precursor dithioester
1 and MG-
ethylammonium chloride 2 was carried out using similar experi-
mental conditions as previously reported (Scheme 3) [53]. Since the
MG amino derivative (2) was isolated as an ammonium salt,
a tertiary amine (N,N-diisopropyl ethylamine, DIPEA) was added for
deprotonation. Contrary to the previous cases, the reaction was
performed in DMF since MG-ethylammonium chloride was only
soluble in highly polar solvents (1 was also very soluble in DMF).
Within the first hours of reaction, the formation of a new product
could be detected by TLC (new green spot at Rf ¼ 0.32, eluent:
CH3CN/H2O 95:5) and by 1H NMR [two new peaks could be
Scheme 1. Synthesis of dye-labeled CTAs from
a precursor dithioester 1.NHS:N-
hydroxysuccinimide.

5938
M. Beija et al. / Polymer 52 (2011) 5933e5946
Scheme 2. Synthesis of MG-ethylammonium chloride (2).
observed: at 1.51 ppm (d) and at 4.59 ppm (q), respectively corre-
sponding to CH3 and CH of compound 3].
reaction was fast, with formation of products within the first
minutes as observed on TLC plate (two new spots at Rf ¼ 0.27 and
0.33; 4 does not migrate and Rf (1) ¼ 0.95; eluent: CH2Cl2/Acetone
8:2). DIPEA ammonium salt and NHS as well as remaining 4 (more
soluble in water than RhoBEDBA 5 due to the double cationic
charge) could be eliminated by washings with acidified water. The
final product was then purified by precipitation in diethyl ether
from a methanol solution in order to remove the remaining
precursor dithioester 1. However, the purified product contained
two rhodamine derivatives: RhoBEDBA (5) and a by-product
identified as a RhoB piperazine thioamide (6) (Scheme 4A) as
proved by several evidences: a) 1H NMR spectrum showed a peak at
5.09 ppm (q; by coincidence, at the same chemical shift than the
corresponding CH of 1) and a peak at 1.54 ppm (d; corresponding to
protons CH3 of RhoBEDBA), but the integrals corresponding to
aromatic protons and to protons of N,N-diethylamino and pipera-
zine groups were two-fold higher than expected; b) 13C NMR
spectrum showed, in addition to the peak at 228.7 ppm (charac-
teristic of thiocarbonyl carbon of RhoBEDBA), another peak at
202.6 ppm, which might correspond to a carbon of a C]S bond; c)
on the ESI-MS spectrum, two peaks were present, one at m/
z ¼ 719.2 corresponding to the desired product and another at m/
z ¼ 631.2 corresponding to the RhoB piperazine thioamide by-
product (6). 6 was formed upon nucleophilic attack to the
dithioester moiety by the amino group of the dye derivative. It was
previously demonstrated that this side-reaction did not occur with
primary amines if optimal reaction conditions were employed
(slow addition of a maximum of one equivalent of amine) [47]. In
our case, the dye derivative is a secondary amine and it was esti-
mated by the above three techniques that the final product con-
tained 44e50% of thioamide [67].
During purification of MGEDBA (3), several difficulties have
been encountered due to the strong polarity of this molecule and to
the presence of residual impurities from the amino-functionalized
dye. Since 3 was soluble in water, elimination of N-hydrox-
ysuccinimide (NHS) with water washings was not possible.
Therefore, a very polar mixture of solvents (acetonitrile/water) had
to be used as eluent for purification by silica gel chromatography to
promote migration of MGEDBA (3). However, due to the strong
polarity of the eluent, some impurities remained in the final
product (1H and 13C NMR spectra showed some peaks at high
fields). The structure of 3 was confirmed by NMR and ESI-MS
analyses (cf. Supporting Information, Figures S1 to S3).
3.1.2. Synthesis of RhoB-labeled CTA
As recently reviewed [65], rhodamine derivatives can be ob-
tained by modification of the amino groups of the xanthenic moiety
(position 3 and 6) as well as modification of the carboxyphenyl ring
at positions 40 and/or 50 or modification of the carboxylic acid group
(position 20). The first strategy usually leads to loss of fluorescence
of the modified dye. With the second strategy, a mixture of prod-
ucts (whose separation is very difficult and time-consuming) is
generally obtained. Therefore, we decided to use the third strategy
for the synthesis of a rhodamine B (RhoB) amino derivative. As in
the case of MG, the RhoB derivative should also have a strong bond
(such as a CeC or an amide bond) between the RhoB moiety and the
amino group. In addition, the modification should impart no or very
little changes on the photophysical properties of the dye. So, we
decided to prepare a RhoB piperazine amide derivative using the
reaction conditions proposed by Nguyen and Francis [49]. Since this
molecule is a tertiary amide, the cyclization to spirolactam [66]
(common for secondary amides of rhodamines) is avoided and
the fluorescence properties of RhoB are maintained even at basic
conditions. The compound 4 (Scheme 4) was successfully obtained
with a yield of 52%. After this purification step, the structure of 4
was confirmed by 1H and 13C NMR as well as ESI mass spectrometry
(see experimental part) and these analyses indicated a high purity
of the product.
The specificity of the nucleophilic attack depends on the relative
electrophilicities of each reactive center and the steric hindrance of
reactants. The activated ester moiety is more electrophilic than the
dithioester moiety. However, secondary amines are more sterically
hindered than primary amines and the presence of the methyl
group at the b-position relatively to carbonyl may encumber the
approach between the secondary amine and the activated ester
moiety, decreasing the reactivity of the amine towards this acti-
vated ester center.
So, in order to get a pure product, it would be necessary to
separate 5 and 6. However, due to the great similarity between
these two structures, it was not possible to find an appropriate
eluent for chromatography. Likewise, neither extractions nor
recrystallization would be efficient since their solubilities are
identical. Hence, a different synthetic strategy was envisioned to
obtain a pure RhoBEDBA.
Then, the synthesis of RhoB-labeled CTA was first investigated
using the same strategy (convergent approach) as for MGEDBA (in
dichloromethane with DIPEA as a base for deprotonation). The
We decided to employ a linear approach (Scheme 4B) similar to
the one described by Mertoglu et al. for the synthesis of a naph-
talenesulfonate-labeled CTA [33]. In a first step, RhoB bromopro-
panoylpiperazine amide (7; X ¼ Br) was prepared by the reaction of
Scheme 3. Synthesis of Malachite Green ethyl dithiobenzoate amide (MGEDBA, 3).
DIPEA: N,N-diisopropyl ethylamine; NHS: N-hydroxysuccinimide.

M. Beija et al. / Polymer 52 (2011) 5933e5946
5939
Scheme 4. Synthesis of Rhodamine B ethyl dithiobenzoate amide (RhoBEDBA, 5) by convergent (A) or linear (B) approaches. DIPEA: N,N-diisopropyl ethylamine; NHS: N-
hydroxysuccinimide.
2-bromopropanoyl bromide and 4, using DIPEA as a base (Scheme
B-1). The initial RhoB amino derivative was fully converted within
only 1 h 30 min. 1H NMR analysis showed two quadruplet peaks at
4.18 and 4.55 ppm and two doublet peaks at 1.51 and 1.52 ppm,
indicating that two products were formed. ESI-MS analysis
confirmed the presence of the desired product (peaks of equal
intensities at m/z ¼ 645.1 and 647.1 expected for this molecule
containing a Br atom). However, a very intense peak at m/z ¼ 601.3
and a peak at m/z ¼ 603.3 (1/3 of the intensity of the former peak)
were also present, corresponding to the same molecule with
a chlorine atom (i.e. 7 with X ¼ Cl) instead of a bromine one. We
suppose that the formation of this product occurred by the
homopolymerization of DMA, a bisubstituted acrylamide derivative
leading to water-soluble and biocompatible polymers [68,69]. The
RAFT polymerization of DMA had been previously studied in our
group using tert-butyl dithiobenzoate (tBDB) as CTA [70,71]. This
CTA was very efficient to control the polymerization, giving rise to
PDMA chains with experimental number average molecular
weights (Mn) close to theoretical values and narrow molecular
weight distributions. Moreover, it was previously observed with
another acrylamide (N-acryloylmorpholine), that no significant
differences in kinetics existed when tBDB or dithiobenzoates
bearing a propanamide derivative as R group were employed [72].
Here, the polymerizations were performed in dimethylsulfoxide
(DMSO) since it is a good solvent of DMA, PDMA and both MGEDBA
and RhoBEDBA. The dye-labeled CTAs were not soluble enough in
water to perform the polymerization in aqueous solution. As
mentioned in the introduction, since these dyes are simultaneously
aromatic and ionic, they have an amphiphilic nature. So, while they
are completely soluble in water at the usual concentrations for
fluorescence studies (10ꢁ7 to 10ꢁ4 mol Lꢁ1), it is not possible to get
aqueous solutions at the required concentrations for RAFT poly-
merization (10ꢁ3 to 10ꢁ1 mol Lꢁ1). The polymerizations were
carried out at 75 ꢀC using AIBN as initiator, with
[DMA]0 ¼ 1.6 mol Lꢁ1 and [DMA]0:[CTA]0:[AIBN]0 ¼ 350:1:0.2
(Table 1). Several runs were performed in NMR tubes equipped
with a Young’s valve in order to monitor the polymerization
kinetics by on-line 1H NMR (in DMSO-d6). Other runs were per-
formed in Schlenk tubes to obtain greater amounts of polymer
samples.
concomitant attack of the counter-ion of 4 (Clꢁ) to the carbon at
a-
position relative to the carbonyl through a SN2 mechanism. Since
Brꢁ is a better leaving group than Clꢁ, the so-formed product would
not further react with Brꢁ, leading to formation of a RhoB chlor-
opropanoylpiperazine amide (7; X ¼ Cl) as the main product. Since
both molecules are reactive, such mixture of products was used in
the second step of RhoBEDBA synthesis. Moreover, since all reac-
tants and other by-products could be eliminated during washings,
no chromatographic purification was carried out at this step. Yields
between 80 and 87% were reached.
In the second step, the reaction between 7 and the dithio-
benzoate anion is carried out in order to form RhoBEDBA. A
biphasic system was used, with solubilization of 7 in dichloro-
methane (a very good solvent for rhodamine derivatives) and later
addition of an aqueous solution (pH ¼ 6) of PhCS2Na salt (3.5 eq), in
the presence of a catalytic amount of NaI to increase the reactivity
of the rhodamine halopropanamide derivatives. Using these
experimental conditions, it was possible to quantitatively obtain
the desired product. RhoBEDBA was first purified by precipitation
in diethyl ether from a methanol solution (99% yield). RhoBEDBA
was further purified by silica gel chromatography to obtain a highly
pure product. However, the yield decreased to 58% since much of
the product was lost by adsorption onto the silica (the stationary
phase was pinkish at the end of purification). This yield is very close
to that obtained by Mertoglu et al. for the synthesis of the
First, a reference experiment (run 1) was carried out using tBDB
in DMSO (previous polymerizations of DMA mediated by tBDB had
been conducted in 1,4-dioxane [70,71]). Then, under similar
experimental conditions, DMA polymerization was undertaken in
the presence of MGEDBA (runs 2, 3 and 4) and three different
samples of RhoBEDBA (runs 5e9): sample RhoBEDBA-1 synthe-
sized by method A (Scheme 4) with a purity of 53% due to the
presence of a thioamide by-product; sample RhoBEDBA-2 synthe-
sized by method B (Scheme 4) and purified by precipitation from
methanol in diethyl ether (no impurities were detected in this
naphtalenesulfonate-labeled dithiobenzoate using
a
linear
approach (55%) [33]. The RhoB-labeled CTA was then fully charac-
terized by 1H NMR, 13C NMR and high resolution fast atom
bombardment (FAB) mass spectrometry (cf. Experimental Section).
Those analyses confirmed that a very pure rhodamine-labeled CTA
was obtained.
sample by standard characterization techniques such as 1H and 13
C
NMR, ESI-MS); and sample RhoBEDBA-3, obtained by further
purification by silica gel chromatography of sample RhoBEDBA-2.
Although sample RhoBEDBA-1 presented a thioamide by-product,
we expected that it should not interfere with the RAFT mecha-
nism since it was previously observed that the polymerization of n-
butyl acrylate in the presence of an oligopeptide-modified dithio-
benzoate containing 24% of a thioamide by-product did not cause
any inhibition nor additional retardation [73]. Two runs were
carried out with a smaller [DMA]0:[CTA]0 ratio (runs 4 and 9, with
3.2. Homopolymerization of DMA mediated by dye-labeled CTAs
The efficiency of the novel dye-labeled CTAs, MGEDBA (3) and
RhoBEDBA (5), was investigated in the case of the RAFT

5940
M. Beija et al. / Polymer 52 (2011) 5933e5946
Table 1
Experimental conditions of RAFT polymerization of DMA. T ¼ 75 ꢀC; [CTA]0/[AIBN]0 ¼ 5.
Run
CTA
[DMA]0/[CTA]0
[DMA]0 (mol Lꢁ1
)
Solvent
Name
Structure
1
2
tBDB
356
351
1.6
1.6
DMSO-d6
DMSO-d6
MGEDBA
3a
4a
MGEDBA
MGEDBA
343
51
1.6
1.0
DMSO
DMSO
5
RhoBEDBA-1b
354
1.6
DMSO-d6
6
RhoBEDBA-2
RhoBEDBA-3
RhoBEDBA-3
RhoBEDBA-3
357
351
302
126
1.4
1.6
1.6
1.6
DMSO-d6
DMSO-d6
DMSO
7
8a
9a
DMSO
a
Carried out in a Schlenk tube.
[RhoBEDBA-1]0 was corrected taking into account its purity (53%).
b
MGEDBA and RhoBEDBA-3, respectively) in order to get polymer
chains with a lower Mn to facilitate characterization of the a-end
group.
The runs that were carried out using RhoBEDBA samples that
were not further purified by silica gel chromatography (Rho-
BEDBA-1 and RhoBEDBA-2) showed an apparent inhibition during
the full polymerization time (90 min; Fig. 1; run 5: open circles
and run 6: grey circles, respectively). MALDI-ToF MS analyses of
the polymerization media demonstrated that this inhibition
period was indeed a very long induction period since dormant
polymerization of NAM in the presence of non purified CTAs
(w25e39% of impurities) showed additional retardation, which
was partially suppressed after meticulous purification of these CTAs
(98% purity), with the conversion plateau increasing from 55% to
A
60
40
20
0
oligomers bearing the RhoB moiety at the
a-chain-end were
detected (see supporting information; Figure S7). This long
induction period might be caused by the presence of some
impurities in the RhoBEDBA-1 and RhoBEDBA-2 samples acting as
radical traps. The exact structure of these impurities remains
unknown since they were present in very small amounts (not
detectable by NMR or MS) but they probably have an ionic
aromatic structure close to that of Rhodamine B derivatives which
would explain that they were extremely difficult to eliminate.
Since this long induction period was also observed in run 6, the
thioamide side-product (only present in run 5) was probably not
responsible for this effect.
A huge difference was observed when using the RhoBEDBA-3
sample (further purified by silica gel chromatography) in run 7
(Fig. 1; black circles). These results demonstrate that very low
amounts of impurities in the CTA sample (run 6, below the detec-
tion limit of the characterization techniques) may strongly influ-
ence the polymerization kinetics (run 6, 90 min induction period in
comparison with run 7, 20 min induction period). Then, when
evaluating a new RAFT agent, the occurrence of a long induction
period should not only be related to the CTA structure but also to
possible traces of impurities in the CTA sample (unless it has been
checked carefully). Plummer and co-workers [74] have studied the
effect of the presence of impurities in cumyl dithiobenzoate on the
polymerization kinetics of 2-hydroxyethyl methacrylate, styrene,
and methyl acrylate. They have shown that even the presence of
a very low amount of impurity (w3%) may induce partial inhibition
or retardation of the polymerization. Moreover, in our group, the
0
50
100
150
Time (min)
1.0
0.8
0.6
0.4
0.2
0.0
B
0
50
100
150
Time (min)
Fig. 1. Time evolution of monomer conversion (A) and pseudo-first order kinetic plots
(B) for RAFT polymerization of DMA in DMSO-d6 at 75 ꢀC using tBDB (run 1; solid line),
MGEDBA (run 2; ), RhoBEDBA-1 (run 5; B), RhoBEDBA-2 (run 6; ) and RhoBEDBA-
3 (run 7; C) as CTA. [CTA]0/[AIBN]0 ¼ 5; [DMA]0/[CTA]0 z 350; [DMA]0 ¼ 1.6 mol Lꢁ1
(except for run 6: [DMA]0 ¼ 1.4 mol Lꢁ1).

M. Beija et al. / Polymer 52 (2011) 5933e5946
5941
80%. However, besides the presence of these impurities, the
molecular weights were generally not affected (dispersities below
1.2, experimental molecular weights very close to the calculated
ones taking into account the amount of impurities) [72].
In comparison with the reference experiment with tBDB (Fig. 1;
run 1: solid line), the rate of polymerization of DMA mediated by
RhoBEDBA-3 (Fig. 1; run 7: black circles) was very similar, with ca.
60% DMA conversion reached within 2 h, suggesting that this
dithiobenzoate caused a comparable retardation of DMA poly-
merization, even if a slightly longer induction period was observed
(20 min vs. 10e15 min in the case of tBDB). This behavior is in
agreement with that previously observed with N-acryloyl mor-
pholine in the presence of tBDB or dithiobenzoates bearing
a propanamide derivative as R group [72].
As explained in the previous part, MGEDBA sample contains an
important amount of impurities (very difficult to eliminate). Then,
the real [CTA]0 was lower than calculated. If we assume that
MGEDBA has a similar influence as RhoBEDBA on the RAFT poly-
merization kinetics (which is a reasonable hypothesis considering
that they have the same Z group and a leaving R group of the same
nature), and that these impurities are perfectly inert, faster poly-
merization kinetics (and shorter induction periods) should be
observed (Fig. 1, run 2: open squares compared to run 7: black
circles). However, since the polymerization rate was almost the
same, it is likely that the impurities were not inert. Then, the
increase in polymerization rate due to a lower real [MGEDBA]0
might have been balanced by the decrease in polymerization rate
caused by the presence of impurities, leading to an apparently
similar polymerization rate.
Fig. 2. (A) RI (black line) and UVeVisible (grey line; labs ¼ 600 nm) traces of MG-
PDMA sample obtained in run 4 and (B) RI (black line) and fluorescence (grey line;
lexc ¼ 560 nm; lem ¼ 575 nm) traces of RhoB-PDMA sample obtained in run 9.
3.3. Characterization of dye-labeled polymer chains and end-group
analysis
which seems to correspond to chains terminated by bimolecular
combination according to the molecular weight value at the peak
p, main peak
The
a-dye-labeled PDMA chains were characterized by size
maximum (Mp) (A:
M
¼
9700 g , Mp,
molꢁ1
exclusion chromatography (SEC) in DMF, since this eluent is
appropriate for characterization of PDMA samples [75e81] while
being a good solvent for both MG and RhoB dyes. Although our
apparatus was equipped with a multi-angle light scattering (MALS)
detector, it was not possible to obtain the absolute molecular
weights of the polymer samples due to optical interference with
MG (that absorbs at the laser wavelength with a very high molar
absorption coefficient) and RhoB. Although this later should not
absorb light at the laser wavelength, a significant response of the
MALS detector was observed when RhoBEDBA was analyzed alone.
The real reasons for this phenomenon remain unclear but might be
related to fluorescence of the RhoB (eventually due to excitation by
daylight), which would perturb the light scattering analysis.
Therefore, it was necessary to use a calibration curve, obtained with
narrow molecular weight distribution PDMA samples synthesized
in our laboratory by RAFT polymerization mediated by tert-butyl or
N-succinimidyl dithiobenzoates as CTAs. However, although these
PDMA standards have a backbone of the same nature as the MG-
¼ 18 500 g molꢁ1; B: Mp,
¼ 8500 g molꢁ1, Mp,
shoulder
main peak
¼ 17 800 g molꢁ1). Such longer chains cannot result from the
shoulder
binding of two SH-ended chains (by formation of a disulfide bond)
since hydrolysis of the dithiobenzoate moiety does not occur in
DMF (anhydrous DMF was used). However, this shoulder repre-
sents only ca. 10 wt.% of the sample (i.e. 5% of the polymer chains). A
similar result was obtained by Donovan et al. [77] for the poly-
merization of DMA mediated by a dithiobenzoate with an analo-
gous structure to our CTAs, bearing instead a N,N-dimethyl
propionamide moiety as R group (with an identical [CTA]0/[AIBN]0
ratio of 5). The authors demonstrated that if the [CTA]0/[AIBN]0
ratio was increased to 80, the shoulder was completely eliminated
(though, the polymerization rate strongly decreased and several
Table 2
Theoretical and experimental Mn and dispersity (Ð) values for dye-labeled PDMA
samples.
a
b
c
d
e
Run Polymer
Conv. Mn,theor
(%)
Mn, SEC
Ðb
Mn,NMR
(g.molꢁ1
)
DIexp DItheo
and RhoB-labeled PDMA samples, their
different. Hence, the -MG- and -RhoB-labeled PDMA samples
a-chain-ends are very
(g.molꢁ1
)
(g.molꢁ1
)
a
a
2
3
7
8
9
MG-PDMA 42
MG-PDMA 30
RhoB-PDMA 61
RhoB-PDMA 39
RhoB-PDMA 33
15,200
2100
21,900
12,500
4900
41,600
9000
n.d.
18,900
8600
1.06 n.d.
e
e
and the PDMA standards may have slightly different conformation
in the eluent, thus different hydrodynamic volumes and, accord-
ingly, different retention volumes for polymer chains of the same
molecular weight. In addition,
stationary phase as previously observed when analyzing
functionalized PNAM samples by SEC [53]. Therefore, the experi-
mental Mn values of the a-dye-labeled PDMA samples have to be
interpreted with caution.
Size exclusion chromatograms of MG-PDMA and RhoB-PDMA
samples (runs 4 and 9, Fig. 2A and B, respectively) are symmetric
with the exception of a small shoulder at low retention volumes,
1.06 8900
n.d. 21,200
1.07 17,900
1.11 7800
1.0
e
0.93
e
1.0
1.1
0.91
0.90
a-chain-ends may interact with the
a
-lipid-
½DMAꢄ0 ꢂ MWDMA ꢂ conversion
½CTAꢄ0
a
Mn;theor
¼
þ MWCTA
b
Obtained by SEC using 0.05 mol Lꢁ1 LiBr in DMF as eluent (calibration with
PDMA standards).
c
Average value between Mn,NMR values obtained using region B or C (cf. Figs. 4
and 5 and equation (1)).
d
DIexp ¼ Mn,SEC/Mn,NMR
.
t
d
e
DItheor ¼ [CTA]0/{[CTA]0 þ [initiator]0 ꢂ ð1 ꢁ eꢁk Þ}.

5942
M. Beija et al. / Polymer 52 (2011) 5933e5946
indeed controlled by [CTA]0 (as expected since RhoB-labeled CTA
used in runs 7 to 9 was pure).
After purification by dialysis, the dye-labeled PDMA samples
were characterized by 1H NMR (MG-PDMA: Fig. 4; RhoB-PDMA:
Fig. 5). By comparison of the 1H NMR spectra of the dye-labeled
PDMA samples and the corresponding dye-labeled CTA, the pres-
ence of MG ethyl propanoyl amide (Fig. 4: aromatic protons, at
7.0e7.8 ppm and protons 1, at w3.2 ppm) and RhoB propanoyl
piperazine amide moieties (Fig. 5: protons 3-9, at 6.8e7.8 ppm;
protons 1, at w1 ppm and protons 2 and 10) was confirmed at the
a-
end of the chains, as well as the dithiobenzoate moiety at the -end
u
(Fig. 4: MG-PDMA; proton 3’ at 7.9 ppm and Fig. 5: RhoB-PDMA;
protons 1’-2’, between 7.2 and 7.8 ppm and protons 3’, at 8 ppm),
showing that both chain-ends are preserved after purification by
dialysis.
Fig. 3. Comparison of RI traces of SEC chromatograms of RhoB-PDMA samples from
run 8 (black line) and run 9 (dotted line).
If we consider that all chains are dormant and bear the dye
moiety at the a-chain-end, an estimation of Mn of PDMA samples
can be obtained by comparison of the integral corresponding to
region A (6.8e8.0 ppm) with integrals of region B (2.2e3.4 ppm) or
C (1.0e2.0 ppm) using equation (1):
days were necessary to achieve 67% conversion). In addition, these
SEC chromatograms confirmed the presence of MG and RhoB at the
polymer chain-end. Indeed, the RI and the UVevisible traces (in the
case of MG-PDMA samples) or the RI and the fluorescence traces (in
the case of RhoB-PDMA samples) perfectly overlap.
The control of the molecular weights of the MG- and RhoB-
labeled PDMA samples was investigated from the SEC analyses
and the 1H NMR spectra. The dispersity determined by SEC was
very low (Ð<1.2, Table 2). In the case of MG-PDMA samples, the
experimental Mn values were much higher than the calculated
Mn,theor (Table 2) which was expected considering the significant
amount of impurities in the CTA. In the case of RhoB-PDMA
samples, the experimental Mn values were somewhat higher than
ꢀ
ꢁ
IB
C ꢁ nB
IA
nA; CTA
or
or C; CTA
Mn;NMR
¼
=
ꢂ MWDMA
nB
or C; DMA
þ MWCTA
(1)
where I is the integral value, nCTA and nDMA are respectively the
number of CTA and DMA protons in the corresponding region,
MWDMA and MWCTA are respectively the molecular weight of DMA
and CTA.
The values of Mn,NMR are slightly overestimated since we
considered that all chains bear both fragments from the CTA, which
is not completely correct due to the presence of some dead chains
and some chains initiated by primary radicals from AIBN. However,
according to our experimental conditions, we should expect less
than 10% of dead chains. Generally, the ratio between absolute Mn
values and Mn,NMR values should provide an estimation of dye
incorporation (DI). In our case, since it was not possible to get
Mn,theor (Table 2). In fact, the presence of a a-RhoB-chain-end can
influence the retention volume as mentioned above [72]. However,
the relative comparison between runs 8 and 9 (where Mn,theor is 2.5
times higher in run 8) indicates that experimental Mn values are
2.2e2.3 times higher in run 8 (Table 2 and Fig. 3 that shows a clear
shift between the corresponding peaks). These observations
confirm that the molecular weights of the RhoB-PDMA samples are
1
Fig. 4. Comparison of H NMR spectra of MGEDBA (DMSO-d6, 300 MHz, 27 ꢀC) and MG-PDMA (run 4;Mn,
¼ 9000 g molꢁ1; D2O, 400 MHz, 40 ꢀC).
SEC

M. Beija et al. / Polymer 52 (2011) 5933e5946
5943
1
Fig. 5. Comparison of H NMR spectra of RhoBEDBA-3 (acetone-d6, 300 MHz, 27 ꢀC) and RhoB-PDMA (run 9;Mn,
¼ 8600 g molꢁ1; D2O, 400 MHz, 40 ꢀC).
SEC
absolute Mn values, the relative Mn,SEC values were used. However,
since both Mn,NMR and Mn,SECvalues are overestimated, the dye
incorporation values have to be considered with caution (DIexp in
Table 2). Although there are some non-labeled chains (less than 10%
considering the used [CTA]0/[AIBN]0 ratio, the temperature and the
polymerization times), for most applications these non-labeled
chains do not interfere with the results (it is often necessary to
mix labeled with non-labeled chains otherwise fluorescence is too
high).
second highest occupied bonding orbital to the lowest vacant
orbital and involves migration from the phenyl ring to the rest of
max
the system [82] appears at around
l
¼ 425 nm. Only a slight
abs
difference of a few nm in the maximum absorbance wavelength can
be detected between the various derivatives (Table 3) and it is very
max
close to the value of non-modified MG (
l
¼ 616.5 nm [39]),
abs
showing that the derivatization of MG led to almost no changes in
the molecular orbital energy levels. Molar absorption coefficients of
These results confirm that the novel high valued water-soluble
dye-labeled dithiobenzoates, MGEDBA and RhoBEDBA (with
a very polar and charged structure), are as efficient as tBDB
(a simple CTA) to control the RAFT polymerization of DMA.
Although CTA purity has a strong influence on RAFT polymerization
kinetics, we got a very satisfying control over molecular weights
and chain-end functionalization of the RhoB-labeled PDMA
samples.
3.4. Photophysical properties of dye-labeled CTAs and
corresponding polymers
The absorption spectra of MG-ethylammonium chloride (2),
MGEDBA (3) and MG-PDMA (run 4) were traced in acidified ethanol
(Fig. 6). Being a planar or nearly planar molecule, MG presents two
types of absorption bands, one polarized along the x direction and
other along the y direction. So, as expected, the maximum
absorption band of the MG derivatives and MG-labeled polymer is
located at ca. 620 nm, which corresponds to the promotion of an
electron from the nonbonding molecular orbital (NBMO) of one of
the nitrogen atoms to the lowest antibonding orbital, producing an
excited state with a high electron density on the central carbon
Fig. 6. Absorption spectra of MG-ethylammonium chloride (2, black, Capp ¼ 4.04 ꢂ
10ꢁ5 mol Lꢁ1), MGEDBA (3, grey, Capp ¼ 9.67 ꢂ 10ꢁ5 mol Lꢁ1) and MG-PDMA (run 4,
red, [polymer] ¼ 1.58 ꢂ 10ꢁ5 mol Lꢁ1). Insert: normalized spectra indicating that the
maximum absorption wavelength remains the same. (For interpretation of the refer-
ences to colour in this figure legend, the reader is referred to the web version of this
article.)
max
abs
max
atom (x-band transition) (
l
¼ 616.5 nm;
3
¼ 148 900) [39].
The y band that arises from excitation of aanbselectron from the

5944
M. Beija et al. / Polymer 52 (2011) 5933e5946
Table 3
Photophysical properties of dye derivatives and dye-labeled polymers.
max
em
Derivative
lmabasx/nm
3
max/104 molꢁ1 L cmꢁ1
l
/nm
4F
Decay parameters
abs
b
a1
s
1/ns
a2
s
2/ns
a3
s
3/ns
s
h i
X2
2
622
620
619
562
563
562
2.8a
e
e
e
e
e
e
e
e
e
e
e
e
MGEDBA
MG-PDMA
4
RhoBEDBA
RhoB-PDMA
0.41a
e
e
e
e
e
e
e
e
e
8.4
e
e
e
e
e
e
e
e
e
8.7 ꢃ 0.9
9.3 ꢃ 0.4
9.5 ꢃ 0.2
583
586
582
0.54
0.21
0.58
1
5.27
1
1.97
2.00
2.12
e
e
e
e
1.97
0.76
2.12
1.02
0.91
1.03
11.6
e
1.03
e
22.4
e
0.32
e
a
Apparent value, since the real concentration in chromophore is unknown.
P
i aisi
i ai
b
P
Calculated by h
s
i ¼
.
MG-ethylammonium chloride (2) and MGEDBA (3) were obtained
from the plots of maximum absorbance as a function of concen-
tration using the BeereLambert law (Table 3). However, since
impurities remained in these products, the values were much lower
RhoB-PDMA were determined using BeereLambert law. Those
values are slightly lower than of the parent RhoB
max
(
3
¼ 11.1 ꢂ 104 molꢁ1 L cmꢁ1) [38], however they are in agree-
abs
ment with previously reported values [49]. As in the case of
MGEDBA and MG-PDMA, the band corresponding to the dithio-
benzoate moiety cannot be detected due to its much lower molar
absorption coefficient value.
than
expected
for
this
kind
of
compounds
max
(
3
¼ 10.6 ꢂ 104 molꢁ1 L cmꢁ1) [83], because the apparent
abs
concentration of the compound did not correspond to its real
concentration. On the other hand, the molar absorption coefficient
value obtained for MG-PDMA is much closer to that of MG, showing
that after purification by dialysis most of the impurities were
removed. It is noteworthy that in the case of both MGEDBA and
The fluorescence quantum yields of these RhoB derivatives and
of RhoB-PDMA were determined by comparing their fluorescence
emission spectra with that of Rhodamine 101 in acidified ethanol
using equation (2):
ꢀ
ꢁ
MG-PDMA, the band corresponding to the absorption of the
ARho101
F
nsolvent
nEtOH
max
fF
¼
ꢂ
ꢂ
2ꢂfF; Rho 101
(2)
dithiobenzoate moiety
(
l
¼
500e530 nm) could not be
abs
A
FRho101
observed on the spectrum, which is due to the fact that this chro-
mophore has a much lower molar absorption coefficient value (
max
3
where 4F is the fluorescence quantum yield, A is the absorbance
at the excitation wavelength (Aꢅ0.1), F is the area under the
abs
<100 molꢁ1 L cmꢁ1) [84].
Fig. 7 shows absorption and fluorescence emission spectra of
RhoB piperazine amide (4), RhoBEDBA (5) and RhoB-PDMA (run 9)
solutions in acidified ethanol (for simplicity, the absorption and
fluorescence emission maxima were normalized to unity). As it can
be observed, the two RhoB derivatives and RhoB-PDMA present
very similar absorption and fluorescence maximum wavelengths
(Table 3), showing that both the modification of the RhoB dye and
the attachment to a polymer chain do not cause any significant
change in its photophysical properties.
The absorption spectra of RhoB piperazine amide (4), RhoBEDBA
(5) were traced at several dye concentrations (10ꢁ6 to 10ꢁ5 mol Lꢁ1
)
and the plots of maximum absorbance as a function of concentra-
tion were linear for both derivatives (results not shown). Hence,
molar absorption coefficients of these derivatives as well as of
Fig. 8. Fluorescence decay curves for RhoB-PDMA (run 9, black), RhoBPA (blue) and
RhoBEDBA (red) solutions in ethanol (10ꢁ6 mol Lꢁ1). The solutions were excited at
570 nm and emission was monitored at 585 nm. The solid black line represents the
best fit to the data. The lamp profile is also shown (light grey line). The weighted
residuals are shown below the decay curves. (For interpretation of the references to
colour in this figure legend, the reader is referred to the web version of this article.)
Fig. 7. Normalized absorption and emission (lexc ¼ 530 nm) spectra of RhoB pipera-
zine amide (4, black lines), RhoBEDBA (5, grey lines) RhoB-PDMA (run 9, dotted lines)
solutions in ethanol at OD ¼ 0.1.

M. Beija et al. / Polymer 52 (2011) 5933e5946
5945
corrected emission curve, and n is the refractive index of the used
solvents.
Solutions were prepared in order to have an optical density of
Acknowledgment
The authors thank the Portuguese NMR Network (IST-UTL
Center) for providing access to the NMR facilities, Dr. Michel Becchi
(CCMP, IBCP, SFR Biosciences Gerland-Lyon Sud UMS344/US8, Lyon,
France) for carrying out mass spectrometry, Mr. Vincent Martin
(Laboratoire des Matériaux Organiques à Propriétés Spécifiques,
Solaize, France) for the determination of specific refractive index
increment, Dr. Aleksander Fedorov for the time-resolved fluores-
0.1 at the excitation wavelength (lexc ¼ 530 nm). Although the
fluorescence quantum yield value obtained for 4 is very similar to
the reported value for RhoB in its cationic form, a much lower
value was observed for RhoBEDBA (Table 3). The main reason must
be the fluorescence quenching caused by the presence of
a dithiobenzoate group. [84] In fact, while the fluorescence decay
of 4 could be fitted with a single exponential function (with
c2 < 1.2 and homogeneously distributed weighted residuals and
autocorrelation of residuals), a sum of three exponentials was
necessary for the fitting of the fluorescence decay of RhoBEDBA
(Fig. 8). It is however remarkable that the fluorescence decay of
the RhoB-PDMA polymer sample is also monoexponential,
meaning that the polymer chain allows RhoB to be sufficiently far
away from the dithiobenzoate moiety to avoid intra-molecular
quenching (at these diluted polymer concentrations, the inter-
molecular quenching is very unlikely). The lifetime of RhoB-
PDMA is slightly longer than that of RhoBPA because attaching
the dye to a polymer chain reduces in some extent its radiationless
deactivation processes.
ꢀ
cence measurements and Dr. Zeljko Petrovski and Dr. Pedro Góis for
organic experimental help and advice. Mariana Beija thanks Fun-
dação para a Ciência e Tecnologia for her PhD grant (SFRH/BD/
18562/2004).
Supporting Information
Assignments of 1H and 13C NMR peaks of MG and RhoB deriv-
atives; 1H, 13C NMR and MS spectra of MGEDBA and RhoBEDBA,
MALDI-ToF MS spectra (reflectron mode) of final polymerization
mixtures of runs 5 and 6.
Appendix. Supplementary material
4. Conclusions
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.polymer.2011.10.041
We describe the synthesis of two novel cationic RAFT agents,
one labeled with a Malachite Green (MG) dye and another with
a Rhodamine B (RhoB) dye. First, MG and RhoB derivatives bearing
an amino functional group were synthesized. In the case of RhoB,
a secondary amino derivative was prepared and for MG, we pre-
sented an original one-pot synthetic route. Then, a MG-labeled
dithiobenzoate (MGEDBA) was prepared in a straightforward
References
[1] Piçarra S, Gomes PT, Martinho JMG. Macromolecules 2000;33:3947e50.
[2] Laukkanen A, Winnik FM, Tenhu H. Macromolecules 2005;38:2439e48.
[3] Relógio P, Martinho JMG, Farinha JPS. Macromolecules 2005;38:10799e811.
[4] Winnik MA. Acc Chem Res 1985;18:73e9.
[5] Beija M, Relógio P, Charreyre MT, da Silva AMG, Brogueira P, Farinha JPS, et al.
Langmuir 2005;21:3940e9.
[6] Callahan J, Kopecek J. Biomacromolecules 2006;7:2347e56.
[7] Kim K, Lee M, Park H, Kim JH, Kim S, Chung H, et al. J Am Chem Soc 2006;128:
3490e1.
[8] Charreyre MT, Mandrand B, Martinho JMG, Relógio P, Farinha JPS. Fluorescent
polymers soluble in an aqueous solution and a method for the production
thereof. Biomérieux S.A. Centre Nacional de la Recherche Scientifique; Insti-
tuto Superior Técnico; 2007. WO2007003781.
[9] Chen M, Ghiggino KP, Mau AWH, Rizzardo E, Thang SH, Wilson GJ. Chem
Commun; 2002:2276e7.
[10] Beija M, Charreyre M-T, Martinho JMG. Prog Polym Sci 2011;36:568e602.
[11] Farinha JPS, Martinho JMG. J Phys Chem C 2008;112:10591e601.
[12] Barner-Kowollik C. Handbook of RAFT polymerization. Weinheim: Wiley-
VCH; 2008.
[13] Chiefari J, Chong YK, Ercole F, Krstina J, Jeffery J, Le TPT, et al. Macromolecules
1998;31:5559e62.
[14] Willcock H, O’Reilly RK. Polym Chem 2010;1:149e57.
[15] Nakayama M, Okano T. Biomacromolecules 2005;6:2320e7.
[16] Favier A, Ladavière C, Charreyre M-T, Pichot C. Macromolecules 2004;37:
2026e34.
[17] Scales CW, Convertine AJ, McCormick CL. Biomacromolecules 2006;7:
1389e92.
[18] Segui F, Qiu XP, Winnik FM. J Polym Sci Part A Polym Chem 2008;46:314e26.
[19] Zhu J, Zhu XL, Cheng ZP, Liu F, Lu JM. Polymer 2002;43:7037e42.
[20] Li Y, Yang JW, Benicewicz BC. J Polym Sci Part A Polym Chem 2007;45:
4300e8.
[21] Chen M, Ghiggino KP, Thang SH, White J, Wilson GJ. J Org Chem 2005;70:
1844e52.
[22] Zhou NC, Lu LD, Zhu XL, Yang XJ, Wang X, Zhu J, et al. Polym Bull 2006;57:
491e8.
[23] Zhou NC, Zhu J, Zhang ZB, Zhu XL. E-Polymers; 2007.
[24] Fu J, Cheng ZP, Zhou N, Zhu J, Zhang W, Zhu X. Polymer 2008;49:5431e8.
[25] Zhu J, Zhu XL, Zhou D, Chen JY, Wang XY. Eur Polym J 2004;40:743e9.
[26] Prazeres TJV, Beija M, Farinha JPS, Charreyre MT, Martinho JMG. Polymer
2010;51:355e67.
[27] Zhou N, Lu L, Zhu J, Yang X, Wang X, Zhu X, et al. Polymer 2007;48:1255e60.
[28] Chen M, Ghiggino KP, Mau AWH, Sasse WHF, Thang SH, Wilson GJ. Macro-
molecules 2005;38:3475e81.
[29] Zhou D, Zhu XL, Zhu J, Hu LH, Cheng ZP. E-Polymers; 2006.
[30] Wan XM, Zhu XL, Zhu J, Zhang ZB, Cheng ZP. J Polym Sci Part A Polym Chem
2007;45:2886e96.
[31] Zhou D, Zhu X, Zhu J, Cheng ZP. J Polym Sci Part A Polym Chem 2008;46:
6198e205.
manner using
a convergent approach by reaction of MG-
ethylammonium chloride (2) with a precursor dithiobenzoate
bearing an activated ester function (1). However, the analogous
reaction with RhoB amino derivative led to a mixture of dithio-
benzoate and thioamide derivatives, suggesting that secondary
amines have almost the same reactivity with respect to the
dithiobenzoate moiety compared to the activated ester moiety. An
alternative linear approach was then used, yielding the RhoB-
labeled RAFT agent (RhoBEDBA) with complete conversion.
The purification of these dye-labeled CTAs was very challenging
because of their dual nature (aromatic and ionic). Indeed, elimi-
nation of all impurities from the MGEDBA sample was not possible.
On the contrary, it was possible to fully purify RhoBEDBA (>98% by
silica gel chromatography) keeping a satisfactory yield (58%).
Both MGEDBA and RhoBEDBA were efficient CTAs to control the
RAFT homopolymerization of DMA, since in all cases the resulting
polymer samples presented low dispersities (Ð<1.2) and both
chain-ends were preserved. In some cases, the presence of impu-
rities slowed down the polymerization kinetics. However, when
a very pure RhoBEDBA sample was employed, the time evolution of
monomer conversion was very similar to that observed when
a simple CTA,tBDB, was used.
Finally, we showed that the attachment of RhoB and MG to the
PDMA polymer chain-end did not influence the photophysical
properties of these dyes. Therefore, these new dye-labeled RAFT
agents can be used to prepare various polymers and especially
water-soluble ones, to study their conformation and dynamics in
solution or at interfaces using fluorescence methods, or to get
labeled probes for imaging and diagnosis purposes. Several ther-
mosensitive block copolymers have been prepared from the MG-
and RhoB-labeled PDMA homopolymers; their synthesis as well as
fluorescence studies in aqueous solutions will be reported in
a further publication.

5946
M. Beija et al. / Polymer 52 (2011) 5933e5946
[32] Such GK, Evans RA, Davis TP. Macromolecules 2006;39:9562e70.
[33] Mertoglu M, Laschewsky A, Skrabania K, Wieland C. Macromolecules 2005;38:
3601e14.
[60] Hallas G, Paskins KN, Waring DR, Humpston JR, Jones AM. J Chem Soc Perk
Trans 1977;2:450e6.
[61] Cho BP, Blankenship LR, Moody JD, Doerge DR, Beland FA, Culp SJ. Tetrahedron
[34] Chen M, Ghiggino KP, Mau AWH, Rizzardo E, Sasse WHF, Thang SH, et al.
Macromolecules 2004;37:5479e81.
[35] Chen M, Ghiggino KP, Thang SH, Wilson GJ. Angew Chem Int Ed 2005;44:
4368e72.
[36] Chen M, Ghiggino KP, Launikonis A, Mau AWH, Rizzardo E, Sasse WHF, et al.
J Mater Chem 2003;13:2696e700.
[37] Zana R. Fluorescence studies of amphiphilic block copolymers in solution. In:
Alexandridis P, Lindman B, editors. Amphiphilic block copolymers - self-
assembly and applications. Amsterdam: Elsevier; 2000.
[38] Arbeloa FL, Arbeloa TL, Estevez MJT, Arbeloa IL. J Phys Chem 1991;95:2203e8.
[39] Green FJ. The Sigma-Aldrich Handbook of stains, dyes and indicators. Aldrich
Chem Co Library; 1990.
2000;56:7379e88.
[62] Kimura K, Yokota G, Yokoyama M, Uda RM. Macromolecules 2001;34:2262e8.
[63] DeFina SC, Dieckmann TJ. Labelled Compd Radiopharm 2002;45:241e8.
[64] Djuric S, Venit J, Magnus P. Tetrahedron Lett 1981;22:1787e90.
[65] Beija M, Afonso CAM, Martinho JMG. Chem Soc Rev 2009;38:2410e33.
[66] Kim HN, Lee MH, Kim HJ, Kim JS, Yoon J. Chem Soc Rev 2008;37:1465e72.
[67] Albeit ESI-MS and 13C NMR techniques are not quantitative methods, since
the two molecular structures are very similar and the considered carbon is of
the same nature (C¼S), a similar ESI-MS response and a comparable relaxa-
tion for 13C NMR should be expected.
[68] Kopecek J, Sprincl L, Bazilova H, Vacik J. J Biomed Mater Res 1973;7:111e21.
[69] Chiari M, Cretich M, Consonni R. Electrophoresis 2002;23:536e41.
[70] Relógio P, Charreyre MT, Farinha JPS, Martinho JMG, Pichot C. Polymer 2004;
45:8639e49.
[40] Valeur B. Molecular fluorescence - principles and applications. Weinheim:
WileyVCH; 2002.
[41] Farinha JPS, Charreyre MT, Martinho JMG, Winnik MA, Pichot C. Langmuir
2001;17:2617e23.
[42] Fonseca T, Relógio P, Martinho JMG, Farinha JPS. Langmuir 2007;23:5727e34.
[43] Prazeres TJV, Farinha JPS, Martinho JMG. J Phys Chem C 2008;112:16331e9.
[44] Martinho JMG, Prazeres TJV, Moura L, Farinha JPS. Pure Appl Chem 2009;81:
1615e34.
[45] Moura LM, Martinho JMG, Farinha JPS. Chem PhysChem 2010;11:1749e56.
[46] Romão RIS, Beija M, Charreyre M-T, Farinha JPS, Gonçalves da Silva AMPS,
Martinho JMG. Langmuir 2010;26:1807e15.
[71] Relógio P. Water-soluble associative polymers: synthesis and photophysical
study of poly(N, N-dimethylacrylamide) and poly(N-acryloylmorpholine).
Química, vol. PhD. Lisbon: Universidade Técnica de Lisboa; 2006.
[72] Bathfield M. Synthèse d’agents de contrôle originaux pour l’obtention de
copolymères alpha-fonctionnels par le procédé RAFT. Application à l’élabo-
ration de particules à chevelure controlée. vol. Doctorat. Lyon: Université
Claude Bernard - Lyon 1; 2006.
[73] ten Cate MGJ, Rettig H, Bernhardt K, Borner HG. Macromolecules 2005;38:
10643e9.
[47] Bathfield M, D’Agosto F, Spitz R, Charreyre MT, Delair T. J Am Chem Soc 2006;
128:2546e7.
[74] Plummer R, Goh YK, Whittaker AK, Monteiro MJ. Macromolecules 2005;38:
5352e5.
[48] Favier A, Charreyre MT, Chaumont P, Pichot C. Macromolecules 2002;35:
8271e80.
[49] Nguyen T, Francis MB. Org Lett 2003;5:3245e8.
[75] Teodorescu M, Matyjaszewski K. Macromolecules 1999;32:4826e31.
[76] Kobayashi M, Ishizone T, Nakahama S. Macromolecules 2000;33:4411e6.
[77] Donovan MS, Lowe AB, Sumerlin BS, McCormick CL. Macromolecules 2002;35:
4123e32.
[50] Van Geet AL. Anal Chem 1968;40:2227e9.
[51] Hoffman RE, Becker ED. J Magn Reson 2005;176:87e98.
[52] Marquardt DW. J Soc Ind Appl Math 1963;11:431.
[53] Bathfield M, Daviot D, D’Agosto F, Spitz R, Ladavière C, Charreyre M-T, et al.
Macromolecules 2008;41:8346e53.
[54] Pratt EF, Green LQ. J Am Chem Soc 1953;75:275e8.
[55] Barker CC, Hallas G, Bride MH, Stamp A. J Chem Soc; 1961:1285.
[56] Ritchie CD, Sager WF, Lewis ES. J Am Chem Soc 1962;84:2349e56.
[57] Armstrong L, Jones AM. Dyes Pigm 1999;42:65e70.
[58] Hallas G, Grocock DE, Hepworth JD. J Soc Dyers Colourists 1970;86:200.
[59] Castelino RW, Hallas G. J Chem Soc B: Physical Organic; 1971:1471e3.
[78] Schierholz K, Givehchi M, Fabre P, Nallet F, Papon E, Guerret O, et al.
Macromolecules 2003;36:5995e9.
[79] Li YT, Lokitz BS, McCormick CL. Macromolecules 2006;39:81e9.
[80] Gondi SR, Vogt AP, Sumerlin BS. Macromolecules 2007;40:474e81.
[81] Kirkland SE, Hensarling RM, McConaughy SD, Guo Y, Jarrett WL,
McCormick CL. Biomacromolecules 2008;9:481e6.
[82] Duxbury DF. Chem Rev 1993;93:381e433.
[83] Herz ML. J Am Chem Soc 1975;97:6777e85.
[84] Farinha JPS, Relógio P, Charreyre MT, Prazeres TJV, Martinho JMG. Macro-
molecules 2007;40:4680e90.