Preparation of phosphine-functionalized polystyrene stars by metal catalyzed controlled radical copolymerization and their application to hydroformylation catalysis

Article abstract of DOI:10.1039/c3dt33082f

Well defined star copolymers have been prepared by copper-catalyzed atom transfer radical copolymerization of styrene and styryldiphenylphosphine starting from a modified Boltorn H30 multifunctional initiator. These polymers and an analogue obtained by debromination of the arm ends with nBu ₃SnH have been used in combination with [Rh(acac)(CO)₂] for the homogeneous phase hydroformylation of 1-octene.

Full text of DOI:10.1039/c3dt33082f

Dalton
Transactions
View Article Online
View Journal | View Issue
PAPER
Preparation of phosphine-functionalized polystyrene
stars by metal catalyzed controlled radical
copolymerization and their application to
hydroformylation catalysis†
Cite this: Dalton Trans., 2013, 42, 9148
Andrés F. Cardozo,a,b,c Eric Manoury,a,c Carine Julcour,b,c Jean-François Blanco,b,c
Henri Delmas,b,c Florence Gayeta,c and Rinaldo Poli*a,c,d
Well defined star copolymers have been prepared by copper-catalyzed atom transfer radical copolymeri-
zation of styrene and styryldiphenylphosphine starting from a modified Boltorn™ H30 multifunctional
initiator. These polymers and an analogue obtained by debromination of the arm ends with nBu3SnH
have been used in combination with [Rh(acac)(CO)2] for the homogeneous phase hydroformylation of
1-octene.
Received 24th December 2012,
Accepted 20th February 2013
DOI: 10.1039/c3dt33082f
www.rsc.org/dalton
Introduction
Polymer science and the desire to fabricate advanced materials
with specific functions have taken a giant step forward with
the development of controlled radical polymerization tech-
niques that combine functionality tolerance with precise
control of architecture, topology, composition, molecular
weight, mole fraction and location of specific functional
groups.1–6 A particularly successful protocol for controlled
radical polymerization is atom transfer radical polymerization
(ATRP).7,8 This is a metal-catalyzed process whereby multiple
controlled insertions of monomers (M) into the halogen-
capped chain end occur by reversible activation of the
dormant chain (Pn–X). It involves a halogen atom transfer
process to the catalyst, a complex of a transition metal Mt sup-
ported by a given coordination sphere L (Mtx/L) that changes
its formal oxidation state from x to x + 1 during the activation
process, and deactivation by the reverse process (Scheme 1).4,5
The position of the ATRP equilibrium is in great favour of the
dormant species over the active radicals, thereby greatly redu-
cing the incidence of bimolecular terminations.
Scheme 1 Mechanism of the ATRP process.
Phosphines have so far received little attention as func-
tional groups in the controlled radical polymerization
approach, in contrast with the wide interest in phosphine
functionalized polymers for catalytic applications.9,10 For
instance, rhodium catalysts anchored on polymer-grafted
phosphine ligands were first described by Manassen11 and
then developed in the 70s by Grubbs,12,13 Čapka14 and
Pittman15 as hydrogenation, hydrosilylation and hydroformyla-
tion catalysts. Many other polymer-supported phosphine
ligands have since been reported, mostly containing modified
triphenylphosphine.9,16–20 The strategies used to obtain poly-
meric ligands can be divided into two families, the first one
involving incorporation of the phosphine function by chemical
modification of a pre-existing phosphine-free polymer16,21,22
and the second one involving the use of a phosphine-contain-
ing comonomer during the polymerization. Notably, both
linear and cross-linked triphenylphosphine-containing poly-
styrenes have been obtained by copolymerization of 4-di-
phenylphosphinostyrene (or styryldiphenylphosphine, SDPP)
and regular styrene (S) by anionic or free radical methods,
possibly in the presence of p-divinylbenzene or 1,4-bis(4-vinyl-
aCNRS, LCC (Laboratoire de Chimie de Coordination), 205 route de Narbonne,
BP 44099, F-31077 Toulouse Cedex 4, France. E-mail: rinaldo.poli@lcc-toulouse.fr;
Fax: +33-561553003; Tel: +33-561333173
bCNRS, LGC (Laboratoire de Génie Chimique), 4 Allée Emile Monso, BP 84234,
31030 Toulouse Cedex 4, France
cUniversité de Toulouse, UPS, INPT, F-31077 Toulouse Cedex 4, France
dInstitut Universitaire de France, 103, bd Saint-Michel, 75005 Paris, France
†Electronic supplementary information (ESI) available: GPC characterization of
some of the isolated star polymers. See DOI: 10.1039/c3dt33082f
phenoxy)-butane as
a cross-linking comonomer. These
9148 | Dalton Trans., 2013, 42, 9148–9156
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
polymers are characterized, however, by broad molecular Aldrich), 2-bromoisobutyryl bromide (98%, Aldrich),
weight distributions and uncontrolled architectures.23,24
nBu3SnH (97%, ACROS), tin bis(2-ethylhexanoate) (96%, Alfa
Living/controlled polymerization techniques allow exerting Aesar), methanol (98%, Aldrich), hexane (>95%, Aldrich),
control over these parameters but examples involving phos- 1-octene (99+%, ACROS), n-nonanal (>97%, Alfa Aesar) and
phine containing monomers are limited. The anionic living dodecane (99%, Aldrich) were used as received. Boltorn™ H30
polymerization of SDPP has provided homopolymers with was dried under vacuum for 24 h at room temperature and
narrow MW distribution, subsequently extended to diblock analyzed by 1H NMR according to established protocols,33–35
copolymer structures by polymerization of regular styrene.25 confirming the specifications provided by the commercial
Terashima, Sawamoto et al. have demonstrated the power of source (Perstorp Chemicals; Mn = 3600 g mol−1, NOH = 32).
ATRP with a ruthenium catalyst to produce polymers incorpor- N,N,N′,N′,N″,N″-Hexamethyltris(2-aminoethyl)amine (Me6TREN)
ating SDPP. More specifically, core–shell polymers with the was prepared according to a literature protocol.36 The macro-
phosphine functions confined in the cross-linked nanogel core initiator (CH3)2C(COOEt)-(Sx-co-SDPPy)-Br (x ∼ 40, y ∼ 12;
were produced starting from a halogen-terminated polyacrylate Mn,SEC = 7520 g mol−1; Đ = 1.38) was prepared as described in
macroinitiator in the presence of SDPP and a cross-linker our previous contribution.32 Carbon monoxide and dihydro-
(ethylene glycol dimethyl dimethacrylate) as comonomers.26–28 gen were obtained from Linde Gas. Syngas was prepared by
These polymers have subsequently been used as macroligands introducing equimolar amounts of CO and H2 into a moni-
to catalyze a variety of reactions.29–31
tored gas reservoir feeding the autoclave reactor at constant
In our own laboratory, we wish to develop macroligands pressure.
where the phosphine functions are contained in flexible linear
Instrumentation
arms rather than in a cross-linked core, thereby being poten-
tially more accessible to substrates and reagents in catalysis. The 1H NMR spectra were recorded at 25.0 °C on Bruker
We have recently reported that the copper-catalyzed atom Avance 300 (5 mm TXO or QNP probes) or Bruker Avance 400
transfer radical copolymerization of S and SDPP yields well (5 mm Atma BBFO probe) spectrometers operating at 300 and
controlled linear polymers when the amount of SDPP in the 400 MHz, respectively. The chemical shifts were calculated
feed is up to 25%.32 In the present contribution, we extend relative to the residual protonated solvent peaks as internal
this approach to the synthesis of well-defined phosphine-func- standards and reported in ppm with positive values upfield
tionalized polystyrene stars. The potential of the resulting from SiMe4. The 31P{1H} NMR spectra were obtained at 25.0 °C
macroligands in catalysis is shown using the hydroformylation on a Bruker Avance 300 spectrometer operating at 121.5 MHz
of 1-octene as a benchmark reaction.
and the chemical shifts are given in ppm relative to the signal
of an 85% H3PO4 solution in D2O, which was used as an exter-
nal standard. The SDPP conversion during the polymerization
was followed by 31P{1H} NMR using a 40 s relaxation delay to
ensure the complete recovery of the magnetization vector
between consecutive scans and the relative areas of the free
Experimental section
General
All glassware was dried at 125 °C overnight. The polymeriz- and polymerized SDPP resonances were quantified by deconvo-
ation reactions were carried out under argon in round bottom lution of the signal with the help of the MestRe-C program
Schlenk tubes connected to three-way stopcocks. All solvents using Lorentzian lineshapes. The amount of phosphorus in
(toluene, 99.7%, Aldrich; anisole, 99.0%, Fluka; dichloro- the purified polymer was measured by integration of the
methane, >99.5%, Aldrich; and THF, >99.9%, Aldrich) were 31P{1H} NMR resonance using a 50 s relaxation delay between
dried using conventional methods and then distilled under an pulses, against a known amount of Ph2EtPvO as the internal
argon atmosphere prior to use. All the purification operations standard.
concerning the SDPP-containing polymers were equally con-
ducted under argon.
The size exclusion chromatographic (SEC) analyses were
carried out at 35 °C on an instrument equipped with a 5 μm
Styrene (99%) and divinylbenzene (80%) (Sigma-Aldrich) gel PL pre-column (50 × 7.5 mm) and a 5 μm mixed gel PL-D
were dried on CaH2 (99%, Aldrich), then distilled under separation column (300 × 7.5 mm) from Polymer Laboratories,
reduced pressure and kept under an inert atmosphere at using THF (1 mL min−1) pre-filtered through 200 nm mem-
−25 °C prior to use. The deuterated solvents for the NMR ana- branes as the mobile phase. The samples were prepared by dis-
lyses were purchased from Eurisotop. 4-Styryldiphenylphos- solution in THF at concentrations of 4 mg mL−1, then
phine (97%, Aldrich), PPh3 (99%, Aldrich), CuBr (99.999%, immersed in an ultrasound bath for 1–2 min, followed by
Aldrich), CuBr2 (99.0%, Fluka), CuCl (99.999%, Aldrich), standing for at least 4 h for the linear polymers (24 h for the
K2CO3 (99%, Aldrich), tris(2-pyridylmethyl)amine (97%, star polymers) and filtration through 200 nm Teflon filters.
Aldrich), AIBN (>98%, ACROS), [Rh(acac)(CO)2] (99%, Alfa Molar masses were measured with a multiangle light scatter-
Aesar), 4-(dimethylamino)pyridine (99%, Fluka), pentamethyl- ing detector (MALLS; minidawn Tristar Wyatt Technology Cor-
diethylenetriamine (99%, Aldrich), 1-bromo ethylbenzene poration), coupled to a differential refractometer (RI2000
(97%, Aldrich), methyl 2-bromopropionate (99.0%, Fluka), Sopares). An operating protocol of total recovery of the injected
ethyl 2-bromoisobutyrate (98%, Aldrich), triethylamine (96%, mass was employed for the variable composition copolymers,
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 9148–9156 | 9149

View Article Online
Paper
Dalton Transactions
allowing the simultaneous determination of the dn/dc para- powder (0.205 g, 25% yield). Mn,SEC = 1.04 × 106 g mol−1 (dn/dc
meter. This method gave a dn/dc value in agreement with the = 0.028), Đ = 1.32. Cross-linking yield = 81%.
literature (0.184) for the PS homopolymer.37 The SEC traces of
the star polymers prepared by the convergent route were
deconvoluted as the sum of two Gaussian functions.
Divergent preparation of poly(S-co-SDPP) stars
The gas chromatographic analyses were carried out on a
(a) Preparation of the H30Br macroinitiator. This pro-
Thermo Fisher Trace GC 2000 chromatograph equipped with a cedure follows a protocol that was previously described for a
CB-CP WAX 52 capillary column (25 m × 0.25 mm; 0.2 μm film closely related multifunctional microgel (Boltorn™ H40).38 In
thickness) and a flame ionization detector, using helium as a Schlenk tube was prepared a solution of Boltorn™ H30
the carrier gas.
(2.25 g, 0.625 mmol) in THF (70 mL). After stirring overnight
Dynamic light scattering measurements were recorded at to ensure complete dissolution, a solution of 4-(dimethyl-
25 °C on a Malvern Instruments Nano-ZS Zetasizer for the amino)pyridine (4.85 g, 39.3 mmol) and Et3N (3.52 mL,
cross-linked polymers using toluene solutions at concen- 25.0 mmol) in THF (20 mL) was added. The reaction vessel
trations of ca. 4 mg mL−1. The dissolution was accomplished was equipped with a dropping funnel and placed in an ice
with the help of an ultrasound bath for 1 min, followed by bath. A solution of 2-bromoisobutyryl bromide (∼10 mL,
standing for 24 h and filtration through 450 nm filters. The 75.0 mmol) in 10 mL of THF was placed in the dropping
hydrodynamic radius was obtained from the measured funnel and then added dropwise over 45 min to the Boltorn™
diffusion coefficient through the Stokes–Einstein equation.
H30 solution. After the end of the addition, the reaction
mixture was stirred for 2 h in the ice bath and then for an
additional 96 h at room temperature. The precipitated salts
Convergent preparation of poly(S-co-SDPP) stars
(a) ATRP conditions. In a Schlenk tube equipped with a were filtered off and the solvent was partially evaporated to ca.
magnetic stirrer bar was prepared a solution of CuBr (22 mg, 20 mL under reduced pressure. The functionalized Boltorn™
0.15 mmol), CuBr2 (4 mg, 0.02 mmol) and PMDETA (40 μL, product (H30Br) was precipitated by adding the mixture to
0.17 mmol) in toluene (5 mL). This was then mixed with a 150 mL of cold methanol. The viscous coagulated product was
separate solution containing the (CH3)2C(COOEt)-(S40-co- washed three times with methanol, letting the mixture settle
SDPP12)-Br macroinitiator (0.574 g, 0.076 mmol) and DVB overnight each time. The viscous material was dried under
(0.3 mL, 1.7 mmol) in 4 mL of toluene. The mixture was then vacuum at 35 °C for 3 days, yielding the product in the form of
immersed in an oil bath thermostated at 90 °C for 60 h, then an expanded foam, which turned into a viscous resin after a
cooled to room temperature, diluted with 50 mL of toluene few days (1.85 g, 35% yield). Mn,SEC = 17 800 g mol−1; dn/dc =
and filtered through an activated neutral alumina column. The 0.068; Đ = 1.46. The 1H NMR of the product confirmed the
polymer was precipitated by addition of cold hexane (400 mL) complete esterification of the Boltorn™ OH groups (see ESI
and the mother liquor was removed by a filter-cannula. The Fig. S1†).
polymer was dried under vacuum until complete removal of
(b) Preparation of the poly(S-co-SDPP) stars from the
the residual solvent (ca. 2 days) to yield a white powder (0.47 g, H30Br core. Only the preparation of the B-3 product will be
64% yield). Mn,SEC = 130 600 g mol−1 (dn/dc = 0.117), Đ = 1.52. described in detail. In a Schlenk tube, CuBr (258 mg,
Cross-linking yield = 41%.
1.76 mmol), CuBr2 (56 mg, 0.25 mmol), TPMA (597 mg,
Additional experiments carried out with a DVB–styrene 2.01 mmol) and anisole (4 mL) were added to styrene (24 mL,
mixture, or using CuCl in place of CuBr, or other macroinitia- 201.5 mmol). After stirring for 5 min at room temperature, a
tors having shorter or longer chains yielded lower cross-linking solution containing the H30Br macroinitiator (527 mg,
yields. Experiments carried out with EGDMA as the cross- 0.063 mmol) and SDPP (11.62 g, 40.3 mmol) in toluene
linker yielded a macrogel.
(33 mL) was added to the reactor through a Teflon cannula.
(b) ARGET-ATRP conditions. In a Schlenk tube equipped The reactor was then immersed in a preheated thermostatic oil
with a magnetic stirrer bar was prepared a solution of CuBr2 bath at 80 °C. Samples were periodically withdrawn with an
(0.5 mg, 0.002 mmol), TPMA (6 mg, 0.02 mmol) and DVB argon purged syringe, diluted with toluene and filtered
(210 μL, 1.2 mmol) in toluene (2 mL). This was then mixed through an activated neutral alumina column. Three separate
with a separate solution containing the (CH3)2C(COOEt)-(S40- fractions of each sample were used respectively for the gas
co-SDPP12)-Br macroinitiator (0.625 g, 0.08 mmol) in 3 mL of chromatographic determination of the styrene conversion, for
toluene. A solution of tin bis(2-ethylhexanoate) (11 μL, 14 mg, the 31P NMR analysis of the SDPP conversion and for the
0.032 mmol) in toluene (1 mL) was prepared separately. A polymer analysis by SEC. The reaction was stopped after 29.5 h
70 μL sample of this solution was added to the mixture before and the mixture was diluted with 15 mL of toluene and filtered
immersion in an oil bath thermostated at 90 °C. Additional ali- through an activated neutral alumina column. The polymer
quots of the tin reagent solution were added at later times was precipitated in 400 mL of cold hexane and separated from
(120 μL after 1 h, 310 μL after 2h30′ and 500 μL after 5 h). After the mother liquor by filtration through a Teflon filter cannula,
a total of 64 h, the reaction was stopped by cooling to room then dried under vacuum (ca. 3 days) to yield a white powder
temperature and the polymer was recovered and dried by the (1.25 g, 14% yield). Mn,SEC = 490 600 g mol−1 (dn/dc = 0.059),
same procedure described above in part (a) to yield a white Đ = 1.13.
9150 | Dalton Trans., 2013, 42, 9148–9156
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
The same reaction was repeated to produce the following (1-octene, n-nonanal, 2-methyloctanal) was confirmed by GC/
polymers (the reaction conditions and the amount of reagents MS analysis.
were the same unless otherwise stated):
B-1. Reaction time = 25.5 h; 1.7 g (15% yield). Mn,SEC
=
387 200 g mol−1 (dn/dc = 0.067), Đ = 1.10. DDHLS = 94.2 nm in
Results and discussion
toluene.
B-2. Reaction time = 24.5 h, CuBr (86 mg, 0.59 mmol), Multiarm star polymers may be accessed by two complemen-
CuBr2 (19 mg, 0.08 mmol), SDPP (1.937 g, 6.73 mmol), TPMA tary, convergent (or “arm first”) and divergent (or “core first”)
(199 mg, 0.67 mmol), styrene (8.0 mL, 67.3 mmol), anisole approaches. In addition to using well defined multifunctional
(1.0 mL), H30Br (176 mg, 0.021 mmol) in 11 mL of toluene. molecular systems as terminating agents (for the arm first
5.7 g (17% yield). Mn = 235 800 g mol−1 (dn/dc = 0.081), approach) or initiators (for the core first approach), it is also
Đ = 1.43.
possible to use diolefins as cross-linkers yielding microgel
cores, with the advantage of simpler operating procedures and
a more moderate cost but with the inconvenience of yielding
polymers with a variable number of arms.39
Dehalogenation of polymer B-3
In a Schlenk tube equipped with a magnetic stirrer bar was
prepared a solution of B-3 (600 mg) and nBu3SnH (0.11 mL,
0.39 mmol) in 15 mL of toluene. After homogenization, AIBN
(1 mg, 0.007 mmol) was added and the reaction vessel was
immersed in a preheated thermostatic oil bath at 60 °C. After
26 h of reaction and cooling, the polymer was recovered by pre-
cipitation in 300 mL of cold hexane. The polymer (556 mg) was
thus recovered after filtration and vacuum drying at 50 °C for
(a) Convergent polymer synthesis
Since we have recently synthesized linear P(S-co-SDPP)-Br poly-
mers by ATRP,32 we first attempted the synthesis of star poly-
mers from these macroinitiators by the convergent approach
and a diolefin cross-linker. Cross-linking indeed took place
when using 1,4-divinylbenzene (DVB) to form well defined
stars (Scheme 2) as shown by the polymer SEC analysis, but
SEC also revealed that a large fraction of the linear macroini-
tiators was not incorporated into the star product. This
phenomenon is well known in the convergent synthesis of star
polymers39 and is not exclusively caused by the loss of
bromide functionality from the linear macroinitiators: the
degree of arm chain end functionality revealed by quantitative
NMR analysis (>90%)32 is indeed much larger than the cross-
linking yield, at best 41% (see Experimental section and ESI
Fig. S1†). A number of attempts to optimize the cross-linking
yield (use of a DVB–styrene mixture, or EGDMA as the cross-
linker) did not give better results. A final attempt involved the
use of ARGET (Activator ReGenerated by Electron Transfer)40
ATRP conditions to accomplish the cross-linking reaction with
DVB, since recent reports have demonstrated the advantage of
this approach for the convergent synthesis of star polymers.41
Indeed, the cross-linking yield was significantly improved, but
it still remained unacceptably low (19% of free arms remain-
ing, see the SEC trace in Fig. S2†). Therefore, the convergent
synthetic strategy was abandoned and we turned to the diver-
gent synthesis from a preformed microgel core.
24 h. Mn,SEC = 465 800 g mol−1 (dn/dc = 0.083), Ð = 1.35, DDHLS
=
21 nm in toluene.
1-Octene hydroformylation
The Rh(CO)2(acac) precatalyst (53.2 mg, 0.2 mmol) and the
desired quantity of ligand (PPh3 or functionalized polymer)
were introduced into a Schlenk tube. Then, a separate mixture
containing 1-octene (16 mL, 102 mmol), dodecane (inert refer-
ence, 2 mL) and toluene (84 mL) was added through a Teflon
cannula and the resulting mixture was stirred for 20 min. It
was then transferred into a Hastelloy C276 autoclave equipped
with a gas inducing stirrer. A first sample was withdrawn, then
the system was purged three times with 15 bar of dinitrogen,
then four times with 15 bar of syngas. The reactor was sub-
sequently heated under low syngas pressure (2 bar) and slow
stirring speed (140 rpm, well under self gas induction) to
hinder significant gas dissolution and the start of the reaction.
When the desired reaction temperature (90 °C) was achieved,
stirring was stopped and the autoclave was pressurized and
constantly fed with syngas at the desired pressure (20 bar). A
sample was withdrawn to evaluate the amount of products
formed during the heating procedure. Then, the data acqui-
sition was started and the stirring speed was set to 1200 rpm.
Both temperature and pressure of the reactor and the gas
ballast were recorded online on a computer at a rate of 1 Hz.
The instantaneous reaction rate was measured from the syngas
consumption. Samples were withdrawn periodically for the
kinetic monitoring by GC/FID (after 10, 20, 40 and 60 min for
PPh3, every 30 min until 2 h and then every hour for the
macroligands). For the gas-chromatographic analyses,
a
precise quantity of anisole (internal standard) was added
before dilution with diethyl ether and injection into the
GC/FID. The identification of the reaction components
Scheme 2
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 9148–9156 | 9151

View Article Online
Paper
Dalton Transactions
(b) Divergent polymer synthesis
the benzylic bromide chain ends under the polymerization
conditions.32 Three polymers were synthesized, two with a feed
molar SDPP fraction fSDPP = 0.17 and one with fSDPP = 0.08. A
low SDPP/S fraction in the feed was selected, because our pre-
vious work on the linear P(S-co-SDPP)-Br polymers has shown
that good control is achieved only for low fSDPP values (≤25%).
Given the very slow polymerization and in order to obtain stars
of relatively small size, the polymerizations were stopped at
low conversion. Use of a more active catalyst such as CuBr/
Me6TREN or CuBr/PMDETA resulted in gelification even when
operating under very dilute conditions. It is worth mentioning
that for the synthesis of the linear P(S-co-SDPP)-Br copolymers,
although the CuBr/TPMA could ensure slower reactions and
better control, the CuBr/Me6TREN system was also able to
yield products with well controlled molecular weights and low
dispersities.
Among various possible multifunctional initiators, we have
focused our attention on Boltorn™, a commercially available
dendritic polyester obtained by polycondensation of 2,2-bis-
(hydroxymethyl)propionic acid on a central polyol core that
offers a large number of hydroxo functionalities. For our pur-
poses we have selected Boltorn™ H30 that contains 32 such
functional units on average. Therefore, its use as a macroinitia-
tor leads to stars with an average of 32 arms. Before divergent
arm growth, Boltorn™ H30 was converted into a suitable
macroinitiator for ATRP (henceforth abbreviated as H30Br) by
reaction with 2-bromoisobutyryl bromide, see Scheme 3. The
same transformation has been previously described for
Boltorn™ H40.38 The completeness of the transformation was
1
confirmed by the H NMR analysis (Fig. S4†). The 2-bromoiso-
butyrate function introduced on H30Br is a suitable initiator
for the ATRP of a variety of monomers, including styrenics. It
has been used to functionalize a number of substrates (silicon
wafers, cellulose nanocrystals, PEO, etc.) for the initiation of
styrene polymerization.42–44 The macroinitiator H30Br was
then used to grow the copolymer arms as shown in Scheme 3.
We selected the CuBr/TPMA couple in toluene at 80 °C as
the catalytic system and operating conditions for the sub-
sequent star growth (conditions that were optimized in pre-
liminary experiments on the corresponding star PS
homopolymers), which resulted in very slow but well con-
trolled polymerization (the results are collected in Table 1). It
is important to point out that, as shown by previously reported
controlled experiments, the phosphine groups do not attack
Polymers B-1 and B-3 were obtained under the same operat-
ing conditions (except on a larger scale and with a slightly
longer reaction time for B-3). The similarity of the monomer
conversions and polymer parameters for these three polymer-
izations shows good reproducibility. Polymer B-2 was obtained
with a lower molar SDPP fraction in the feed. The monomer
conversions were measured by GC for styrene and by 31P NMR
for SDPP (see Experimental section for details). The 31P NMR
indicated no significant oxidation (<1%) of the phosphine
functionalities. All polymerizations show remarkably good
control of the molar mass and low dispersities, especially
those with fSDPP = 0.17. The measured molecular weights
(which were obtained by the complete recovery method given
the unavailability of the dn/dc value) are always greater than
the theoretical values, based on the measured monomer con-
version and H30Br composition (i.e. the known average arm
number).
The polymerization leading to B-2 was monitored in terms
of kinetics and mass evolution with conversion. The results
are shown in Fig. 1. The evolution of the monomer conversion
with time in the semilogarithmic plot of part (a) shows a rela-
tively good fit to the first order law for both monomers. The
faster incorporation of SDPP is in line with the known reactiv-
ity ratios for the free radical copolymerization of these two
monomers (rSDPP = 1.43, rS = 0.52).23 Since the reactions were
stopped at relatively low conversions, the feed composition
drift is limited and therefore the arms should not have a
marked compositional gradient. The molecular weight grows
linearly with conversion, and the dispersity remains relatively
constant around 1.4. These phenomena are characteristic of a
Scheme 3
Table 1 H30-[P(S-co-SDPP)-Br]32 stars synthesized by ATRPa
Product
S/SDPP/Br
fSDPP
t/h
S conv/%
SDPP conv/%
Mn,SEC/g mol−1
Mth/g mol−1
Đ
dn/dc
xb
yb
B-1
B-2
B-3
100 : 20 : 1
100 : 10 : 1
100 : 20 : 1
0.17
0.08
0.17
25.5
24.5
29.5
17
13
—
21
32
23
3.87 × 105
2.36 × 105
4.91 × 105
1.04 × 105
0.81 × 105
—
1.10
1.43
1.13
0.067
0.081
0.059
17.0
13.0
4.2
3.2
c
a Conditions: [H30Br] = 0.01 M; [TPMA] = 0.032 M, CuBr/CuBr2/TPMA = 0.88 : 0.12 : 1; solvent = toluene; T = 80 °C. b Average number of S (for x)
and SDPP (for y) units in each arm (Scheme 3), calculated on the basis of the monomer conversions. c Not available.
9152 | Dalton Trans., 2013, 42, 9148–9156
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
Fig. 3 1H NMR spectrum of polymer B-3-H in CDCl3. The starred resonances
are due to the solvent water impurity.
and also the benzylic CHBr chain end resonance at the charac-
teristic position (δ 4.5). The small and broad signal in the
region between δ 3 and 4.5 is assigned to the protons of the
central Boltorn™ H30 core. The large width of this signal is
probably caused by the reduced mobility (long correlation
time) of this less flexible part of the molecule, even in the
good CD2Cl2 solvent used for the NMR measurement.
In order to establish whether the residual benzylic bromide
chain end functionalities interfere with the catalytic appli-
cation of the macroligands, it was necessary to prepare a de-
halogenated version. This was accomplished for polymer B-3,
following an established protocol,45 by reaction with tributyl
tin hydride, as shown in Scheme 3, to give polymer B-3-H. The
complete disappearance of the 1H NMR resonance of the term-
inal CHBr functions at δ 4.5 (Fig. 3) confirms the total elimin-
ation of the benzylic bromide chain ends, whereas the 31P
NMR spectrum indicates that the polymer-bonded –PPh2 units
have been preserved intact (resonance at δ −6.2). The SEC
analysis shows a monomodal molecular weight distribution
(Fig. S7†) with parameters close to those of the precursor B-3
polymer (Mn = 4.66 × 105 g mol−1, Đ = 1.35).
Fig. 1 First-order kinetic plot of the monomer conversion (a) and evolution
with conversion of Đ (b) and Mn,SEC (c) for the star copolymer B-2.
(c) Hydroformylation of 1-octene
The polymers B-1, B-2, B-3 and B-3-H were then used as macro-
ligands for Rh(acac)(CO)2 in the catalytic hydroformylation of
1-octene under homogeneous conditions, following the same
protocol recently used by us for the same process with the cor-
responding linear polymers, P(S-co-SDPP)-Br,32 and earlier for
the catalyst based on PPh3.46 The catalytic results are shown in
Table 2. From previous work, it is known that the maximum
linear/branched (or l/b) ratio is achieved with a P/Rh ratio of 8
or higher,46 whereas the catalytic activity is reduced in the
Fig. 2 1H NMR spectrum of polymer B-3 in CD2Cl2. The sharp starred reson-
ance (δ 1.56) is due to the solvent water impurity.
controlled system. The SEC traces of these polymers show the presence of excess phosphine ligand. Indeed, the comparison
uniform progression of the monomodal molecular weight dis- of the results obtained with a P/Rh ratio of 4 and 8, carried out
tribution, see Fig. S6.†
for the macroligand B-1 (runs 3 and 4), confirms these trends.
In addition to the SEC characterization, polymer B-1 has The comparison with the hydroformylation results in the pres-
also been analyzed by dynamic light scattering, confirming the ence of PPh3 (run 1) or linear copolymer CP-5 (run 2)32 shows
monomodal distribution and narrow size dispersion (see that on increasing ligand complexity (on going from PPh3 to
Fig. S5†), with an average hydrodynamic diameter of 22 nm. the linear copolymer to the star copolymer) the catalytic
The 1H NMR spectrum of the B-3 copolymer, see Fig. 2, clearly activity decreases whereas the l/b ratio increases reaching the
shows the aliphatic skeleton and aromatic substituent signals, maximum value of 5.6 for copolymer B-1 (run 4). Copolymer
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 9148–9156 | 9153

View Article Online
Paper
Dalton Transactions
Table 2 Homogeneous hydroformylation of 1-octene in toluene solutiona
observation is at variance with the behavior of the correspond-
ing linear polymers, where the presence of the Br chain end
was found to be totally neutral in terms of both rate and selec-
tivity.32 Hence, the terminal C-Br functionalities do not appear
to chemically inhibit the catalysis. A possible interpretation of
this difference is based on the different average conformation
for the two polymeric architectures, resulting in a steric hin-
drance of the 32 Br atoms in the star polymer toward access of
the olefin substrate to the catalytic sites, whereas the micro-
structure of the sites, which controls the stereoselectivity, would
not be affected. The linear polymers will tend to organize
as random coils that will be essentially independent (no
entanglements) in dilute solution in a good solvent, as is the
case here (1 mM solution in toluene). Under these conditions,
the presence of only one Br atom at the chain end will not
significantly perturb the substrate access to the catalytic centers
distributed in the center of the coil. For the star polymers, on
the other hand, each of the 32 arms will propel radially away
from the core and therefore place the Br atoms on or near the
polymeric particle surface. Incidentally, this different macro-
molecular organization of linear and star polymers may also be
responsible for the different behavior with respect to the trend of
the l/b ratio with the molecular weight (chain length for the
linear polymers or arm length for the star polymers).
Thus, several factors are at play (density of phosphine func-
tionalities, polymer size, architecture, P/Rh ratio). Understand-
ing how each one of them independently acts on the l/b
selectivity requires more extensive investigations that are
outside the scope of the present preliminary application of our
star polymers to catalysis. The important point, however, is
that controlled radical polymerization and specifically ATRP
under the conditions outlined in the present contribution has
the possibility to access polymeric materials where each of
these parameters can be individually changed by design. Thus,
tailored polymeric macroligands with an optimized structure
for each specific application may be developed.
Residual
Initial rate, r0 ×
Run
Ligand
P/Rh
t/h
octeneb/%
l/b
104 c/mol l−1 s−1
1
2
3
4
5
6
7
PPh3
CP-5d
B-1
B-1
B-2
B-3
B-3-H
8
8
4
8
8
4
4
0.33
2
4.5
5.5
4.5
4.5
2.5
7
14
25
59
56
28
20
2.8
3.9
4.5
5.6
4.2
4.0
4.0
28.5
7.5
3.1
e
—
2.9
3.8
5.7
a Conditions: [Rh(acac)(CO)2] = 2.0 × 10−3 M, [1-octene] = 1.0 M, T =
90 °C, Psyngas = 20 bar (CO/H2 = 1). b From GC/FID analysis. c From
syngas consumption. d CP-5 = (CH3)2C(COOEt)-(SDPPx-s-Sy)-Br with x =
103 and y = 26; data from ref. 32. e Not available.
B-2, containing a lower density of phosphine functionalities,
yields a lower l/b ratio than copolymer B-1 at similar conver-
sions under identical operating conditions (cf. runs 4 and 5).
A decrease of reaction rate for a polymer supported catalyst
is generally explained by a reduced accessibility of the catalytic
sites to the olefin, a phenomenon that should be accentuated
in bad solvents where the polymer prefers a less expanded con-
formation and the functionalities are therefore less accessible
to the solvent and solute molecules.47 The reduced mobility
and/or accessibility of the phosphine functions in the more
complex star architecture is probably a key factor in the hydro-
formylation kinetics. The toluene solvent used in these cataly-
tic experiments is actually a relatively good solvent for the PS-
type polymer, but the Rh atom may act as an intra-chain or
inter-chain cross-linker, leading to an additional diffusional
limitation. This phenomenon should have a greater effect in
the star polymers because of the possible inter-arm cross-
linking.
From the point of view of selectivity, an improvement of the
l/b ratio for polymer supported phosphine is well known in the
literature.48–52 For instance, the l/b ratio for the hydroformyla-
tion of propene at 85 °C approximately doubles from 3 to 6 on
going from PPh3 to polypropylene-supported phosphine.50 The
major effect has been observed, as detailed in our previous
contribution,32 on going from PPh3 (run 1) to the linear copo-
lymer (run 2). There is, however, a minor but definite further
increase for the l/b ratio upon going from the linear to the star
polymer when comparing the results obtained with the same
P/Rh ratio (cf. run 2 and runs 4 and 5). Polymer B-2 (run 5),
containing ca. half the density of –PPh2 functionalities relative
to B-1, gives a somewhat lower l/b ratio. In our previous study
on the linear polymers, longer chains were shown to increase
the l/b ratio for equivalent phosphine density.32 For the
present star polymers, on the other hand, the l/b ratio
decreases when going from the lower MW polymer B-1 to the
higher MW polymer B-3 (cf. runs 3 and 6).
Conclusions
In this paper, we have taken the atom transfer radical copoly-
merization of styrene and styryldiphenylphosphine one step
further, going from linear copolymers to star polymers with
controlled size and phosphine density. The convergent
method consisting of cross-linking pre-made chain-end func-
tionalized macroinitiators with a diolefin (divinylbenzene) did
not lead to a satisfactory cross-linking yield, although star
polymers with controlled size could be obtained. The divergent
method starting from a multifunctional initiator, in the
example shown here
a hyperbranched polyester of the
Boltorn™ family with ca. 32 external OH groups, leads to well
defined stars with small molecular weight distributions. Their
preliminary application to a catalytic process of industrial
interest, the hydroformylation of higher olefins, exemplified in
the present case by the model 1-octene substrate, shows the
expected performance for a polymer-supported phosphine and
We have also carried out a catalytic test with the debromi-
nated macroligand B-3-H (run 7) under the same conditions as
run 6, namely with a P/Rh ratio of 4. A comparison of the two
runs shows that the catalysis is faster with B-3-H than with
B-3, whereas the selectivity is identical for the two runs. This
9154 | Dalton Trans., 2013, 42, 9148–9156
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
also demonstrates, in comparison with the previously reported 19 F. Shibahara, K. Nozaki, T. Matsuo and T. Hiyama, Bioorg.
performance of the linear polymers,32 that the polymer archi-
Med. Chem. Lett., 2002, 12, 1825–1827.
tecture has an effect on the catalytic performance in addition 20 P. W. N. M. van Leeuwen, T. Jongsma and G. Challa, Macro-
to polymer size, phosphine density, and P/Rh ratio. The mol. Symp., 1994, 80, 241–256.
present results open several new perspectives in terms of 21 K. G. Allum, R. D. Hancock, I. V. Howell, R. C. Pitkethly
polymer structure optimization, application to other catalytic
reactions, catalyst recovery and recycling by ultrafiltration, and
and P. J. Robinson, J. Organomet. Chem., 1975, 87,
189–201.
development of other polymers with different grafted ligands 22 N. A. Demunck, M. W. Verbruggen and J. J. F. Scholten,
for an even greater number of catalytic applications.
J. Mol. Catal., 1981, 10, 313–330.
23 R. Rabinowitz, R. Marcus and J. Pellon, J. Polym. Sci.,
Polym. Chem., 1964, 2, 1241–1249.
24 M. Kwok, W. Choi, H. S. He and P. H. Toy, J. Org. Chem.,
2003, 68, 9831–9834.
25 H. G. Borner and W. Heitz, Macromol. Chem. Phys., 2000,
201, 740–746.
26 T. Terashima, M. Kamigaito, K.-Y. Baek, T. Ando and
M. Sawamoto, J. Am. Chem. Soc., 2003, 125, 5288–5289.
27 T. Terashima, M. Ouchi, T. Ando, M. Kamigaito and
M. Sawamoto, J. Polym. Sci., Polym. Chem., 2006, 44,
4966–4980.
Acknowledgements
We are grateful to the Agence Nationale de la Recherche (pro-
gramme BLANC, project BIPHASNANOCAT, Grant No. ANR-11-
BS07-025-01), to the Centre National de la Recherche Scientifi-
que, and to the Institut Universitaire de France for financial
support. AFC thanks the PRES Toulouse for a Ph.D. fellowship.
We thank Professor Sergio Castillon (Université Rovira i Virgili,
Tarragone) for lending us a generous amount of Boltorn™ H30.
28 T. Terashima, M. Ouchi, T. Ando, M. Kamigaito and
M. Sawamoto, Macromolecules, 2007, 40, 3581–3588.
29 T. Terashima, M. Ouchi, T. Ando and M. Sawamoto,
J. Polym. Sci., Part A, 2010, 48, 373–379.
Notes and references
1 K. Matyjaszewski, Y. Gnanou and L. Leibler, Macromolecu-
lar Engineering: Precise Synthesis, Materials Properties, Appli- 30 T. Terashima, M. Ouchi, T. Ando and M. Sawamoto, Polym.
cations, Wiley-VCH Verlag GmbH, 2007. J., 2011, 43, 770–777.
2 C. J. Hawker, A. W. Bosman and E. Harth, Chem. Rev., 2001, 31 T. Terashima, M. Ouchi, T. Ando and M. Sawamoto,
101, 3661–3688. J. Polym. Sci., Polym. Chem., 2011, 49, 1061–1069.
3 W. A. Braunecker and K. Matyjaszewski, Prog. Polym. Sci., 32 A. F. Cardozo, E. Manoury, C. Julcour, J.-F. Blanco,
2007, 32, 93–146.
4 M. Ouchi, T. Terashima and M. Sawamoto, Chem. Rev.,
2009, 109, 4963–5050.
5 F. di Lena and K. Matyjaszewski, Prog. Polym. Sci., 2010, 35,
959–1021.
6 M. Destarac, Macromol. React. Eng., 2010, 4, 165–179.
7 J.-S. Wang and K. Matyjaszewski, J. Am. Chem. Soc., 1995,
117, 5614–5615.
8 M. Kato, M. Kamigaito, M. Sawamoto and T. Higashimura,
Macromolecules, 1995, 28, 1721–1723.
9 N. E. Leadbeater and M. Marco, Chem. Rev., 2002, 102,
3217–3273.
H. Delmas, F. Gayet and R. Poli, ChemCatChem, DOI:
10.1002/cctc.201200446.
33 A. Burgath, A. Sunder and H. Frey, Macromol. Chem. Phys.,
2000, 201, 782–791.
34 E. Zagar and M. Zigon, Macromolecules, 2002, 35, 9913–9925.
35 E. Zagar, M. Zigon and S. Podzimek, Polymer, 2006, 47,
166–175.
36 J. Queffelec, S. G. Gaynor and K. Matyjaszewski, Macromol-
ecules, 2000, 33, 8629–8639.
37 Polymer Handbook, ed. J. Brandrup, E. H. Immergut and
E. A. Grulke, John Wiley & Sons, Inc., Hoboke, NJ, 4th edn,
1999.
10 M. Guino and K. K. Hii, Chem. Soc. Rev., 2007, 36, 608–617. 38 G. Kreutzer, C. Ternat, T. Q. Nguyen, C. J. G. Plummer,
11 J. Manassen, Chim. Ind., 1969, 51, 1058–1062.
12 R. H. Grubbs and L. C. Kroll, J. Am. Chem. Soc., 1971, 93,
3062–3063.
13 R. H. Grubbs, L. C. Kroll and E. M. Sweet, J. Macromol. Sci.,
A, 1973, 7, 1047–1063.
14 M. Capka, P. Svoboda, M. Cerny and J. Hetflejs, Tetrahedron
Lett., 1971, 4787–4790.
15 G. O. Evans, C. U. Pittman, R. McMillan, R. T. Beach and
R. Jones, J. Organomet. Chem., 1974, 67, 295–314.
16 D. C. Bailey and S. H. Langer, Chem. Rev., 1981, 81, 109–148.
J. A. E. Manson, V. Castelletto, I. W. Hamley, F. Sun,
S. S. Sheiko, A. Herrmann, L. Ouali, H. Sommer, W. Fieber,
M. I. Velazco and H. A. Klok, Macromolecules, 2006, 39,
4507–4516.
39 H. F. Gao and K. Matyjaszewski, Prog. Polym. Sci., 2009, 34,
317–350.
40 K. Min, H. F. Gao and K. Matyjaszewski, Macromolecules,
2007, 40, 1789–1791.
41 J. Burdynska, H. Y. Cho, L. Mueller and K. Matyjaszewski,
Macromolecules, 2010, 43, 9227–9229.
17 G. Challa, J. Reedijk and P. vanLeeuwen, Polym. Adv. 42 K. Matyjaszewski, P. J. Miller, N. Shukla, B. Immaraporn,
Technol., 1996, 7, 625–633.
18 C. A. McNamara, M. J. Dixon and M. Bradley, Chem. Rev.,
2002, 102, 3275–3299.
A. Gelman, B. B. Luokala, T. M. Siclovan, G. Kickelbick,
T. Vallant, H. Hoffmann and T. Pakula, Macromolecules,
1999, 32, 8716–8724.
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 9148–9156 | 9155

View Article Online
Paper
Dalton Transactions
43 Z. Y. Li, P. P. Li and J. L. Huang, J. Polym. Sci., Polym. 48 C. U. Pittman and R. M. Hanes, Ann. N. Y. Acad. Sci., 1974,
Chem., 2006, 44, 4361–4371. 239, 76–87.
44 G. Morandi, L. Heath and W. Thielemans, Langmuir, 2009, 49 C. U. Pittman and R. M. Hanes, J. Am. Chem. Soc., 1976, 98,
25, 8280–8286. 5402–5405.
45 V. Coessens and K. Matyjaszewski, Macromol. Rapid 50 F. R. Hartley, S. G. Murray and P. N. Nicholson, J. Mol.
Commun., 1999, 20, 66–70. Catal., 1982, 16, 363–383.
46 R. M. Deshpande, B. M. Bhanage, S. S. Divekar and 51 P. Kalck, E. L. de Oliveira, R. Queau, B. Peyrille and
R. V. Chaudhari, J. Mol. Catal., 1993, 78, L37–L40. J. Molinier, J. Organomet. Chem., 1992, 433, C4–C8.
47 M. R. Buchmeiser, Polymeric Materials in Organic Synthesis 52 N. S. Imyanitov, V. A. Rybakov and S. B. Tupitsyn, Pet.
and Catalysis, Wiley-VCH, Weinheim, 2005.
Chem., 1992, 32, 408–412.
9156 | Dalton Trans., 2013, 42, 9148–9156
This journal is © The Royal Society of Chemistry 2013