Hyperbranched poly(phenylenevinylene) bearing pendant phenoxys for a high-spin alignment

Article abstract of DOI:10.1039/b203891a

2,6-Dibromo-4-(3′,5′-di-tert-butyl-4′-acetoxyphenyl)styrene was polymerized in a one-pot reaction using a palladium catalyst, and subsequent hydrolysis and oxidation yielded the hyperbranched poly[(4-(3′,5′-di-tert-butyl-4′-ylooxyphenyl)-1,2,(6)- phenylen

Full text of DOI:10.1039/b203891a

View Article Online / Journal Homepage / Table of Contents for this issue
Hyperbranched poly(phenylenevinylene) bearing pendant phenoxys
for a high-spin alignment
Hiroyuki Nishide,* Mitsutaka Nambo and Makoto Miyasaka
Department of Applied Chemistry, Waseda University, Tokyo 169-8555, Japan
Received 22nd April 2002, Accepted 27th August 2002
First published as an Advance Article on the web 27th September 2002
2,6-Dibromo-4-(3’,5’-di-tert-butyl-4’-acetoxyphenyl)styrene was polymerized in a one-pot reaction using a
palladium catalyst, and subsequent hydrolysis and oxidation yielded the hyperbranched poly[(4-(3’,5’-di-tert-
butyl-4’-ylooxyphenyl)-1,2,(6)-phenylenevinylene)], which three-directionally satisfies both an alternant but
non-Kekule´-type p-conjugation and the ferromagnetic connectivity of the unpaired electrons of the pendant
phenoxyl. The AFM image, unusually low solution viscosity, and low chemical reactivity of the phenol group
indicated the polymer to have a highly branched and globular structure. In spite of the extremely crowded
branching, p-conjugation in the polymer skeleton was maintained, as indicated by photoelectron and EPR
spectroscopies. The phenoxyl polymer even with a spin concentration of only 0.4 spin per monomer unit
displayed an average S value of 3.
arrows in Chart 1, three-directionally requires the ferromag-
Introduction
netic (up) alignment of all of the phenoxyls’ spin (represented
by a big arrow) within the polymer. (In Chart 1, the big arrow
and the small up and down arrows represent the spin on the
phenoxyl radical and the direction of the temporarily induced
spin density on the carbons, respectively.)
There has been no report on such a hyperbranched, head-
to-tail linked, and 1,2,4,6-substituted phenylene polymer,
although the dendrimer of a simple poly(1,3,5-phenyleneviny-
lene) with fifth-generation26–30 and hyperbranched copolymers
of phenylene and 1,1-vinylene31 have been recently synthesized
as a new class of p-conjugated polymers. In this paper, we
synthesized an asymmetric AB2 type monomer (2,6-dibromo-4-
phenoxyl-precursor substituted styrene) and polymerized the
monomer in one-pot via the palladium-catalyzed Heck reaction
to form the hyperbranched and head-to-tail linked poly(phe-
nylenevinylene). The p-conjugated skeleton of the hyper-
branched nano-size polymer is discussed.
Synthetic research into high-spin organic molecules using
intramolecular through-bond ferromagnetic spin alignment has
been exhaustively continued to realize pure organic-derived
compounds of unknown magnetism.1–4 Some of the p-conjugated
and alternant, but non-Kekule´-type, organic polymers bearing
multiple radical groups (unpaired electrons) as the spin source
display ferromagnetic or high-spin ordering, owing to intra-
macromolecular spin-exchange interaction through the
p-conjugated skeleton.5–13 Their spin-alignment numbers or
the spin quantum numbers (S) in the ground state have long been
expected to be proportional to the degree of polymerization.14
In addition, the stability of the high-spin state is also theore-
tically predicted to be enhanced exponentially by the dimen-
sions of the ferromagnetically spin-connecting p-conjugated
skeleton.15 For example, Rajca et al.16 recently reported the
highest recorded value of spin-alignment for pure organic
molecules, with the average S over 5000 at low temperatures,
by synthesizing two-dimensionally dendric-macrocyclic poly(1,3-
phenylenephenylmethine)s, but they lacked chemical stability
at room temperature. Bushby et al.5 focused on a triphenyl-
amine radical cation that is stable in ambient conditions and
synthesized the networked poly(phenylenephenylaminium
cationic radical) with S ~ 3. However, an increase in the
molecular size and/or the dimensionality of such radical
polymers always accompanies a structural defect or distorted
and ineffective p-conjugation. That is, a tradeoff problem
remains between the extended p-conjugation and the higher
dimensional structure in the polymer framework. (A study of
electroconductive and optoelectrical polymers encounter a
similar tradeoff.17,18
possibility of a hyperbranched polymer obtained by a one-
pot reaction of an AB2 type monomer,19,20 as a p-conjugated
skeleton to ferromagnetically connect spins.
For this purpose, we have designed a hyperbranched
phenoxyl polymer, poly[4-(3’,5’-di-tert-butyl-4’-ylooxyphe-
nyl)-1,2,(6)-phenylenevinylene] (Chart 1), by extending our
previous results21–25 on the corresponding linear poly(1,2-
phenylenevinylene)-based phenoxyl polymers that displayed
partial spin-alignment (S ~ 2–4) and sufficient chemical
stability. The hyperbranched phenoxyl polymer satisfies both
an alternant and non-Kekule´ p-conjugated structure, and its
ferromagnetic connectivity represented by up–down small
Results and discussion
Polymer synthesis
2,6-Dibromo-4-(3’,5’-di-tert-butyl-4’-acetoxyphenyl)styrene
6
was synthesized as a three-functional monomer to be linked
with head-to-tail bonds via the b-phenylation of the styrenic
vinyl group with phenyl bromide by the Heck reaction,32
according to Scheme 1. The 3,5-di-tert-butyl-4-acetoxyphenyl
group was introduced onto the 4-bromo substituent of 2,4,6-
tribromotoluene to produce 5c via preferential coupling with
[3,5-di-tert-butyl-4-(trimethylsiloxy)phenyl]magnesium bromide
using a nickel catalyst33 to yield 5a, followed by hydrolysis and
acetylation. The methyl group of 5c was converted to a vinyl
group via the Wittig reaction34 to yield 6 (for details see
the Experimental section). 4,6-Dibromo-2-(3’,5’-di-tert-butyl-
4’-acetoxyphenyl)styrene (10), the isomer of 6, was also syn-
thesized via the same synthetic procedure.
The sterically very crowded structure of the polymer in Chart
1 first suggested a strong retardation of the polymer formation.
The hyperbranched (precursor acetoxy) polymer was synthe-
sized through a simple one-pot reaction of the starting AB2
type monomer 6, in the presence of a (tri-o-tolylphosphine)-
palladium catalyst at 90 uC. Although the polymerization of
6 proceeded more slowly, the molecular weight of the
) This motivated us to explore the
3578
J. Mater. Chem., 2002, 12, 3578–3584
This journal is # The Royal Society of Chemistry 2002
DOI: 10.1039/b203891a

View Article Online
Chart 1
to the high solvent–solubility of the hyperbranched polymer 2a.
The GPC profile of the polymer was unimodal. Further reac-
tion of the isolated polymer with the Pd catalyst resulted in
the recovery of the starting polymer. These results exclude a
coupling reaction between the oligomers during the polymeri-
zation because the surface of the oligomer is occupied by the B
group of the AB2 monomer (the Br group of 6).
The head-to-tail and trans-vinylene linkage structures of
polymer 2a were confirmed by NMR, IR, and fluorescence
spectroscopies. Head-to-tail linkage in 2a was supported by a
model reaction of 5a and styrene using the same Pd catalyst and
reaction conditions, which quantitatively yielded 2,6-distyryl-4-
(3’,5’-tert-butyl-4’-acetoxyphenyl)toluene (for details see the
Experimental section). Polymer 2a was converted to the
corresponding hydroxy polymer 2b after complete elimination
of the protecting acetyl group in alkaline solution. The
sterically crowded structure of the hyperbranched 2a retarded
the hydrolysis: hydrolysis of 2a took a long time (24 h) even
with a large excess of alkaline solution, while hydrolysis linear
3a was completed for the with a small excess of alkaline
solution after 12 h ([KOH]/[the acetoxy unit] w 50 and ~ 12
for 2a and 3a, respectively).
Hyperbranched structure and p-conjugation
The precursor acetoxy polymer 2a was obtained as a yellow
powder. In spite of its poly(phenylenevinylene) skeleton, the
Scheme 1
hyperbranched polymer 2a (Chart 2) was unexpectedly somewhat
higher in comparison with the polymerization of the corre-
sponding two-functional 2-bromo-4-(3’,5’-di-tert-butyl-4’-acet-
oxyphenyl)styrene (7) which yields the linear polymer 3a (e.g.,
4
¯
Mw ~ 2.4 6 10 (degree of polymerization ~ 56) and ~ 4.9 6
103 (degree of polymerization ~ 14) for the hyperbranched and
linear polymer, respectively, under the same reaction condi-
tions, see Table 1). The higher molecular weight can probably
be ascribed to the three-functionality of the monomer 6 and/or
Chart 2
J. Mater. Chem., 2002, 12, 3578–3584
3579

View Article Online
Table 1 Examples of the polymerization,a molecular weight, and hydrolysisb of the polymers
Acetoxy polymers
Hydroxy polymers
3
3
¯
Mw/10
¯
Mw/Mn
¯
¯
Mw/10
¯ ¯
Mw/Mn
Monomer/polymers
[M]o/M
Yield (%)
6/2a/2b
0.1
0.5
0.5
0.5
0.1
16
50
68
34
16
3.1
24
4.9
3.8
4.2
1.6
1.7
1.5
1.6
1.3
3.2
24
4.9
—
1.5
1.6
1.5
—
7/3a/3b
8/4
10/11a/11b
5.0
1.3
a[Pd(OAc)2]/[M]o ~ 0.1, [P(C6H4CH3)3]/[Pd(OAc)2] ~ 2, [N(C2H5)3]/[M]o ~ 5. Temp. ~ 90 uC. Time ~ 24 h. Yield ~ methanol insoluble
fraction. b[KOH]/[acetoxy unit] ~ 50, time ~ 24 h (except for 3a/3b ~ 12 h), temp. ~ 50 uC.
Table 2 Solution viscosity,a UV-vis absorption,b fluorescence,c and ionization threshold (Ith)d of the polymers
UV-vis abs
max/nm
Fluorescence
em/nm
Polymer
Mw/103
Viscosity [g]/dl g21
l
e
max/104 dm3 mol21 cm21
l
W
Ith/eV
2a
2b
3a
24
24
24
4.9
24
0.031
357
375
304
302
322
0.80
0.82
1.1
1.1
1.3
470
490
448
0.12
0.12
0.61
5.88
5.86
5.85
0.39
0.12
3b
450
0.63
5.84
c
aIntrinsic viscosity [g], toluene solution at 25 uC. bCHCl3 solution. Benzene solution, quantum yield (W) normalized with 9,10-diphenylanthra-
cene (W ~ 0.84). dPoly(1,2-phenylenevinylene), 4, and 1 ~ 6.17, 5.67 and 5.89 eV, respectively. Reference data measured using the same proce-
dure with the same UPS spectrometer for poly(1,4-phenylenevinylene), poly(acetylene) and poly(1,4-phenylene): Ith ~ 5.29, 5.25, and 5.83 eV,
respectively (from previously reported data:35, 36 Ith ~ 5.3, 5.2, and 5.8 eV, respectively).
hyperbranched polymer was quite soluble in common solvents
such as benzene, THF, and CHCl3. In addition, the solution
viscosity of 2a was one order of magnitude lower in comparison
with those of the corresponding linear polymer 3a (Table 2).
Extremely low solution viscosity is one of the characteristics
of dendric macromolecules, which is ascribed to a highly
branched and globular structure.
The dilute solution of the hyperbranched polymer 2a was
then transferred to a mica surface and subjected to atomic force
microscopy (AFM). Although the polymer collapsed on a
graphite surface, globular polymers were detectable on the
mica (Fig. 1). The globular size corresponds to the molecular
weight of the hyperbranched polymer: a horizontal distance
of ca. 50 nm and a vertical distance of ca. 8 nm were estimated
for 2a with a molecular weight of 2.4 6 104, by taking into
account the width of the cantilever tip in tapping mode.
UV-vis absorptions of the hyperbranched polymers 2 were
broadened but significantly shifted bathochromically, relative
to those of the linear polymers 3 (Table 2). The absorptions
were not influenced by the solvent species, such as benzene,
CHCl3, and acetone. The emission maxima of the fluorescence
ascribed to the poly(phenylenevinylene) skeleton were also
shifted to longer wavelengths for 2 with a decrease in the
quantum yield of the fluorescence in comparison with those for
3. A cast (solvent-free) film of 3b was prepared: UV-vis
absorption and fluorescence maxima were shifted bathochro-
mically to 390 and 470 nm, respectively, which nearly agreed
with those of 2b. This result suggests that an excitonic
interaction causes the bathochromic shifts, that the aromatic
rings are closely spaced together in the hyperbranched polymer,
and that the p-conjugation of the hyperbranched 2 is similar to
that of linear 3.
The p-conjugated electronic structure of poly(phenylene-
vinylene)s has been investigated with ultraviolet photoelectron
spectroscopy (UPS).35,36 The ionization threshold (Ith) of the
hyperbranched polymers 2 was estimated by UPS (Table 2).
The Ith values of 2 including the phenoxyl polymer 1 were
comparable to those of the linear polymers 3. The Ith value of
the hyperbranched poly(1,2,(6)-phenylenevinylene)
4 was
considerably smaller than that of the linear poly(1,2-phenyle-
nevinylene). These results indicate that the p-conjugation in the
poly(phenylenevinylene) skeleton in the hyperbranched poly-
mers 2 and 1 is surprisingly maintained despite the extremely
crowded branching structure.
Radical generation and EPR spectra
The hydroxy polymer 2b was converted to a deeply red-colored
phenolate anion of the polymer with a small excess of
(C4H9)4NOH in CH2Cl2. Quantitative formation of the
phenolate anion was detected by complete disappearance of
the IR absorption at 3638 cm21 (nO–H). The anionic solution of
2b was heterogeneously treated with an aqueous potassium
ferricyanide phase to yield phenoxyl polymer 1. Polymer 1 was
also soluble in common solvents. GPC elution curves of the
polymers 2b and 1, i.e., before and after radical generation,
were coincident with each other. Radical generation does not
bring about oxidative degradation or cross-linking of the
hyperbranched polymer. Polymer 1 was isolated as a deep
brownish powder which was chemically persistent at room
temperature even in air. For example, the half-life of the radical
Fig. 1 AFM images (tapping mode) of the hyperbranched polymer 2a
with a moleculer weight of 4.9 6 103 (a) and 2.4 6 104 (b).
3580
J. Mater. Chem., 2002, 12, 3578–3584

View Article Online
Fig. 2 Controlled potential coulometric oxidation of the phenolate
anions of the hyperbranched polymer 2b (#) and the linear polymer
3b (%). Inset: cyclic voltammogram of the phenolate anions of
the hyperbranched polymer 2b (a) and 5b (b), in CH2Cl2 with
(n-C4H9)4NBF4 and (CH3)4NOH.
1, estimated by monitoring the EPR signal intensity, was 1.6
days at room temperature.
The phenoxyl polymer 1 was also generated by electro-
chemical oxidation. The cyclic voltammogram of the phenolate
anion of 2b was recorded reversibly in repeated potential
sweeps (inset in Fig. 2). This result means that the phenoxyl
polymer is chemically persistent even in solution at room
temperature and that radical generation is not accompanied by
a
subsequent chemical side reaction. Differential pulse
voltammetry gave a simple unimodal response even though
the polymer involved multiple redox sites. The redox potential
(20.13 V vs. Ag/AgCl) almost agreed with those of the
monomeric phenolate derivatives (20.06 and 20.19 V for 5b
and 3b, respectively).
Coulometric oxidation of the phenolate anion of linear 3b on
a carbon felt electrode indicated approximately stoichiometric
oxidation (Fig. 2). Electrolytic EPR spectroscopy supported
the formation of the phenoxyl radical. Under the same
conditions, oxidation of the phenolate anion of the hyper-
branched 2b was slower than that of 3b and ceased at a yield of
ca. 60%. It is believed that the hyperbranched structure retards
the oxidation of the inner-embedded phenolate groups.
Fig. 3a shows the EPR spectrum of the hyperbranched
phenoxyl polymer 1 at a low spin concentration; the broad but
hyperfine spectrum would be attributed to the 5–7 protons of
the phenoxyl pendant ring and phenylenevinylene skeleton,
which is in contrast to the three-line hyperfine structure of
2,4,6-tri-tert-butylphenoxyl ascribed to an unpaired electron
localized in the phenoxy ring. The proton hyperfine structure
was more clearly observed for the model compound 9 (Fig. 3b),
which is simulated by the dashed line by taking into account the
skeleton’s phenylene and vinylene protons. These EPR results
suggest an effectively delocalized spin distribution over the
branched skeleton, most likely because of its still p-conjugated
structure. The data mentioned above indicate a developed
p-conjugation in the poly(phenylenevinylene) skeleton in spite
of the sterically very crowded structure of the hyperbranched
polymer.
Fig. 3 EPR spectra of the hyperbranched phenoxyl polymer 1. (a) The
solid line at g ~ 2.0045 for 1 with a spin concentration of 0.03 spin per
unit in CH2Cl2 solution at room temperature. (b) Monoradical of 9
with a spin concentration of 0.10 in benzene solution at g ~ 2.0043;
dashed line for simulation with aH(Ha, Hb) ~ 0.19, and aH(Hc, Hd) ~
0.04 mT. (c) DMs ~ ¡2 spectrum for 1 with a spin concentration of
0.35 in CH2Cl2 at 10 K. (d) Curie–Weiss plots (EPR signal intensity vs.
1/T) for 1.
from linearity in the lower temperature (v30 K) region. This
upward deviation supported the presence of a multiplet state
for the phenoxyl polymer 1.
Magnetization
The magnetization of the phenoxyl polymer 1 in a frozen
solution was measured with a SQUID magnetometer.37 The
magnetization (M) plots normalized with the saturated
magnetization (Ms) of 1 with a spin concentration of 0.38
are shown to be close to the Brillouin curve for S ~ 3 (closed
symbol in Fig. 4). Unfortunately, only ca. 40% of the potential
spin-sites were successfully oxidized to the phenoxyl radical
even after applying heterogeneous oxidation and a variety of
conditions in the radical generation step.38 Coulometric
oxidation has suggested that spin concentration may be limited
in the radical generation of the hyperbranched polymer
(Fig. 2). The average S ~ 3 value for 1 is lower than the
degree of polymerization of 63 for this polymer, but it is
reasonable considering the spin defect of ca. 60%. Fig. 4 also
gives the M/Ms plots for the linear phenoxyl polymer 3 with
spin concentrations of 0.36 and 0.54: they lie almost on the
The EPR spectrum of 1 showed a sharp and unimodal signal
with increasing spin concentration. g ~ 2.0045 for the signal
was ascribed to an oxygen-centered radical. Fig. 3(c) shows a
DMs ~ ¡2 forbidden transition assigned to a triplet species at
g ~ 4 for 1 with a spin concentration of 0.35 spin per monomer
unit. The EPR signal in the DMs ~ ¡2 region was doubly
integrated to give Curie plots (100–6 K, Fig. 3(d)). Although
the signal intensity was proportional to the reciprocal of the
temperature at higher temperature, the plots deviated upward
J. Mater. Chem., 2002, 12, 3578–3584
3581

View Article Online
2,6-Dibromo-4-(3’,5’-tert-butyl-4’-hydroxyphenyl)toluene (5b).
Methanol (53 cm3) and 10 M HCl (19 cm3) were added to
the THF solution (105 cm3) of 5a (17.4 g, 33 mmol), and the
resulting solution was stirred at room temperature for 3 h.
After the removal of methanol, the solution was extracted
with ether. The crude product was purified using a silica gel
column with hexane–CHCl3 (1 : 1) eluent. Recrystallization
from hexane gave a needle crystal of 2,6-dibromo-4-(3’,5’-di-
tert-butyl-4’-hydroxyphenyl)toluene (5b) (9.0 g): yield 60%.
Mp 178 uC; IR (KBr pellet/ cm21) 3626 (nO–H); dH (CDCl3)
1.5(s, 18H, tert-butyl), 2.6 (s, 3H, CH3), 5.3 (s, 1H, OH), 7.3
(s, 2H, phenyl), 7.7 (s, 2H, phenyl); dC (CDCl3) 23.32, 31.27,
34.50, 123.80, 125.50, 129.47, 130.55, 134.95, 136.55, 142.73;
MS (m/z) 452, 454, 456 (M1, M1 1 2, M1 1 4), calcd for M
~ 454.2. Anal. Calcd for (C21H26OBr2): C, 55.5; H, 5.8; Br,
35.2. Found: C, 55.4; H, 5.6; Br, 35.1%.
Fig. 4 Normalized plots of magnetization (M/Ms) vs. the ratio of
magnetic field and temperature (H/(T2h)) for the hyperbranched
phenoxyl polymer 1 with a spin concentration 0.38 spin per unit ($),
and for the linear polyradical 3 with a spin concentration 0.36 (#) and
0.54 (%) in frozen CH2Cl2 at 1.8–5.0 K. h is the coefficient of a weak
antiferromagnetic interaction. The theoretical curves corresponding to
S ~ 0.5 (—), 1, 1.5, 2, 2.5, 3, and 3.5 ( - - - ) Brillouin functions are also
depicted.
2,6-Dibromo-4-(3’,5’-tert-butyl-4’-acetoxyphenyl)toluene (5c).
Compound 5b (9.0 g, 20 mmol) was suspended in acetic
anhydride (70 cm3) and was stirred in the presence of a few
drops of perchloric acid at room temperature for 15 h. After the
addition of excess water, the product was filtered off and
extracted with ether and purified by column separation (hexane–
CHCl3 (1 : 1) eluent). Recrystallization from hexane gave
2,6-dibromo-4-(3’,5’-di-tert-butyl-4’-acetoxyphenyl)toluene (5c)
(7.5 g): yield 76%. Mp 154 uC; IR (KBr pellet/cm21) 1763
(nCLO); dH (CDCl3) 1.39 (s, 18H, tert-butyl), 2.17 (s, 3H,
CH3), 2.37 (s, 3H, COCH3), 7.41 (s, 2H, phenyl), 7.67 (s, 2H,
phenyl); dC (CDCl3) 22.63, 23.36, 31.47, 35.60, 125.18,
125.43, 130.49, 135.35, 135.81, 142.03, 143.14, 143.19, 170.94;
MS (m/z) 494, 496, 498 (M1, M1 1 2, M1 1 4), calcd for
M ~ 496.3. Anal. Calcd for (C23H28O2Br2): C, 55.7; H, 5.7;
Br, 32.2. Found: C, 55.6; H, 5.6; Br, 32.%1.
Brillouin curves for S ~ 1 and 2, respectively. By comparing
these magnetization data, it can be concluded that the hyper-
branched skeleton acts as an effective p-conjugated coupler of
the pendant spins.
Conclusion
Hyperbranched poly(1,2,(6)-phenylenevinylene) bearing 4-sub-
stituted phenol groups have been synthesized. p-Conjugation
was maintained in spite of the highly branched polymer
skeleton. The phenoxyl polymer 1 with only a spin concentra-
tion of 0.4 gave an average spin state of S ~ 3, which suggests
that a higher spin alignment could be realized by optimization
of the radical generation step. The molecular approach to the
study of mesoscopic phenomena such as molecular recognition
and magnetism is now at the stage that it can be applied to large
molecules with multiple interacting entities through to a nano-
size molecules with added complexities. Our results suggest that
a hyperbranched but still p-conjugated poly(phenyleneviny-
lene) is an effective backbone structure that could be used for
the preparation of a purely organic-based electronic and/or
magnetic material with a high dimensional structure that could
display nanoscopic properties.
2,6-Dibromo-4-(3’,5’-di-tert-butyl-4’-acetoxyphenyl)styrene (6).
N-Bromosuccinimide (3.4 g, 20 mmol) and a,a’-azobisiso-
butyronitrile (a few mg) were suspended in a CCl4 solution
(28 cm3) of 5c (7.5 g, 15 mmol) and refluxed until succinimide
floated on the solution. The mixture was cooled and filtered off.
After evaporation, benzene (32 cm3) and triphenylphosphine
(4.2 g, 16 mmol) were added to it and stirred at 50 uC for 5 h.
Excess reactants were removed on a silica gel column with
hexane–CHCl3 (1 : 1) eluent, and the residue was eluted
with methanol. Freeze-drying of the methanol solution gave
the phosphonium salt (11.2 g): yield 88%.
The phosphonium salt (11.2 g, 13 mmol) was suspended in
25% formaldehyde (190 cm3) and 5 M NaOH (20 cm3) was
added dropwise. The mixture was stirred for 1 h and extracted
with ether. The crude product was purified using a silica gel
column with hexane–CHCl3 (1 : 1) eluent. Recrystallization
from hexane gave 2,6-dibromo-4-(3’,5’-di-tert-butyl-4’-acetoxy-
phenyl)styrene (6) (4.9 g): yield 72%. Mp 139 uC; IR (KBr
pellet/cm21) 1765 (nCLO), 1631 (nCLC); dH (CDCl3) 1.36 (s, 18H,
t-butyl), 2.35 (s, 3H, COCH3), 5.19, 5.38 (d, 2H, CHCH2), 6.51
(dd, 1H, CHCH2), 7.41 (s, 2H, phenyl), 7.75 (s, 2H, phenyl); dC
(CDCl3) 22.67, 31.58, 35.57, 120.70, 122.39, 124.32, 127.85,
132.41, 134.17, 135.88, 142.27, 144.53, 147.46, 170.87; MS (m/z)
506, 508, 510 (M1, M1 1 2, M1 1 4), calcd for M ~ 508.3.
Anal. Calcd for (C24H28O2Br2): C, 56.7; H, 5.6; Br, 31.4.
Found: C, 56.7; H, 5.5; Br, 31.2%.
Experimental
Synthesis
2,6-Dibromo-4-(3’,5’-di-tert-butyl-4’-trimethylsiloxyphenyl)-
toluene (5a). TheGrignardsolution(THF,156cm3)of(4-bromo-
2,6-di-tert-butylphenoxy)trimethylsilane (29 g, 81 mmol) was
added to the cooled THF solution (60 cm3) of 2,4,6-tribromo-
toluene39 (22.2 g, 68 mmol) and [1,3-bis(diphenylphosphino)-
propane]nickel(II) dichloride (81.5 mg), and the mixture was
refluxed for 3 h. The reaction was quenched with 2 M HCl
(84 cm3) and then extracted with ether. The extract was washed
with water, dried, evaporated, and then purified by silica gel
column separation with hexane elution. It was recrystallized
from hexane to give a pale yellowish needle crystal of 2,6-
dibromo-4-(3’,5’-di-tert-butyl-4’-trimethylsiloxyphenyl)toluene
(5a) (17.4 g): yield 49%. Mp 145 uC; IR (KBr pellet/cm21
)
2,6-Dibromostyrene (8). N-Bromosuccinimide (5.0 g, 28 mmol)
and a,a’-azobisisobutyronitrile (a few mg) were suspended
in a CCl4 solution (36 cm3) of 2,6-dibromotoluene (5.0 g,
20 mmol) and refluxed until succinimide floated on the
solution. The mixture was cooled and filtered off. After
evaporation, benzene (40 cm3) and triphenylphosphine (5.0 g,
19 mmol) were added, and stirred at 50 uC for 3 h.
1255 (nSi–C), 914 (nSi–O); dH (CDCl3) 0.40 (s, 9H, Si-CH3), 1.23
(s, 18H, tert-butyl), 2.60 (s, 3H, CH3), 7.48 (s, 2H, phenyl),
7.84 (s, 2H, phenyl); dC (CDCl3) 3.87, 18.35, 23.38, 35.10,
120.54, 125.68, 125.77, 134.79, 139.64, 140.79, 143.61, 154.36;
MS (m/z) 524, 526, 528 (M1, M1 1 2, M1 1 4), calcd for M ~
526.4. Anal. Calcd for C24H34OBr2Si: C, 54.8; H, 6.5; Br, 30.4.
Found: C, 54.6; H, 6.3; Br, 30.2%.
3582
J. Mater. Chem., 2002, 12, 3578–3584

View Article Online
Excess reactants were removed on a silica gel column with
hexane–CHCl3 (1 : 2) eluent, and the residue was eluted
with methanol. Freeze-drying of the methanol solution gave
the phosphonium salt (5.1 g): yield 43%.
The phosphonium salt (5.1 g, 8.7 mmol) was suspended in
25% formaldehyde (44 mL) and 5 M NaOH (13 cm3) was
added dropwise. The mixture was stirred for 1 h and extracted
with ether. The crude product was purified using a silica gel
column with hexane eluent to give 2,6-dibromostyrene as a
yellowish viscous oil (0.9 g): yield 36%. IR (cm21) 1631 (nCLC);
dH (CDCl3) 5.66, 5.68 (d, 2H, CHCH2), 6.63 (dd, 1H, CHCH2),
6.95 (dd, 1H, phenyl), 7.54 (d, 2H, phenyl); dC (CDCl3) 122.67,
123.59, 129.13, 131.80, 132.22, 135.22; MS (m/z) 260, 262, 264
(M1, M1 1 2, M1 1 4), calcd for M ~ 261.9.
Chart 3
(0.23 g, 2.2 mmol) using the Pd catalyst in completely the same
manner as for the polymerization of 6. The product was
purified using a silica gel column with hexane–CHCl3 (1 : 1)
eluent: yield 85%. IR (KBr pellet/cm21) 1765 (nCLO), 960
(dtransHCLCH); dH (CDCl3) 1.42 (s, 18H, tert-butyl), 2.18 (s, 3H,
CH3), 2.36 (s, 3H, COCH3), 7.19–7.78 (m, 18H, Ar); dC
(CDCl3) 22.70 (CH3), 31.44, 35.45 (tert-butyl), 124.60, 126.46,
126.64, 127.30, 127.46, 128.13, 128.27, 128.65, 128.72, 128.78,
128.96, 134.31, 137.48, 142.67 (Ar), 171.10 (Ac). The 14 peaks
based on aromatic carbons were assigned to the small letter
numbers 3, 6, 9, 10, 12, 14, 13, 8, 4, 5, 11, 7, 2 and 1,
respectively, in structure 9a in Chart 3. MS (m/z) : 542 (M1),
calcd for M ~ 542.8.
4,6-Dibromo-2-(3’,5’-di-tert-butyl-4’-acetoxyphenyl)styrene(10).
Compound 10 is an isomer of 6 and was synthesized via the
same synthesis procedure and polymerized to yield the
hyperbranched poly(2-substituted-1,4,(6)-phenylenevinylene)
(11). 10: Mp 117 uC; IR (KBr pellet/cm21) 1765 (nCLO),
1630 (nO–H); dH (CDCl3) 1.4 (s, 18H, tert-butyl), 2.4 (s, 3H,
CH3), 5.6 (d, 1H, CHCH2), 5.7 (d, 1H, CHCH2), 6.7 (dd,
1H, CHCH2), 7.2 (s, 2H, phenyl), 7.4 (s, 1H, phenyl), 7.8 (s,
1H, phenyl); dC (CDCl3) 22.7, 30.9, 31.5, 35.6, 120.7, 122.4,
132.4, 134.2, 134.3, 135.8, 136.5, 142.3, 144.5, 147.5, 170.9;
MS (m/z): 506, 508, 510 (M1, M1 1 2, M1 1 4), calcd for
M ~ 508.3. Anal. Calcd for (C24H28O2Br2): C, 56.7; H, 5.6;
Br, 31.4. Found: C, 56.9; H, 5.5; Br, 31.2%.
Linear poly[4-(3’,5’-di-tert-butyl-4’-hydroxyphenyl)-1,2-phe-
nylenevinylene] (3b) was synthesized as in the previous
papers.21,22
Polymerization
Palladium acetate (66 mg, 0.30 mmol) and tri-o-tolylphosphine
(180 mg, 0.59 mmol) were added to a DMF solution (3.9 cm3)
of 6 (1.0 g, 3.4 mmol) and triethylamine (1.0 g, 9.9 mmol). The
solution was heated at 90 uC for 24 h. The mixture was
separated using a polystyrene gel column with CHCl3 as eluent
and was purified by reprecipitation from CHCl3 in methanol to
yield the polymer (2a) as a yellow powder (0.50 g): yield 50%.
The molecular weight of the polymer was measured by light-
scattering molecular weight analysis (Tosoh LS-8000). IR (KBr
pellet/cm21) 1765(nCLO), 968 (dtransHCLCH); dH (CDCl3) 1.35 (s,
18H, tert-butyl), 2.35 (s, 3H, -OCOCH3), 6.72–7.74 (m, 4H,
Ar). Anal. Calcd for C24nH27n 1 1Brn11O2n (n ~ 56): C, 67.2;
H, 6.3; Br, 19.0. Found: C, 67.3; H, 6.3; Br, 18.6%. Examples of
the yield and the molecular weight under different polymeriza-
tion conditions are given in Table 1.
Poly(1,2,(6)-phenylenevinylene) (4). 2,6-Dibromostyrene was
polymerized with the Pd catalyst, and the polymer was purified
in the same manner as described above: yield 45%. IR (KBr
pellet/cm21) 968 (dtransHCLCH); dH (CDCl3) 6.78–7.60 (m, 5H,
phenyl). Anal. Calcd for C8nH5n11Brn11 (n ~ 21): C, 51.9; H,
2.7; Br, 45.3. Found: C, 51.9; H, 2.6; Br, 45.0%.
AFM
5 mL of the hyperbranched polymer 2a solution (10 mM) was
placed on freshly cleaved mica. After 1 min, excess fluid was
carefully blotted off using filter paper and air-dried. AFM
measurements were carried out using a Digital Instruments
Nanoscope III in the tapping mode under ambient conditions.
Silicon cantilevers (length 125 mm, width 30 mm, thickness 3–
5 mm) with a spring constant between 17 and 64 N m21 and a
resonance frequency in the range 240–400 kHz were used.
Resonance peaks in the frequency response of the cantilever
were selected in the range between 280 and 320 kHz for the
tapping mode oscillation. The scanning rate was usually 1.0 Hz.
Imaging was performed by displaying the amplitude signal and
the height signal, simultaneously.
Poly[4-(3’,5’-di-tert-butyl-4’-hydroxyphenyl)-1,2,(6)-phenyle-
nevinylene] (2b). Compound 2a (1.0 g) was dissolved in a small
amount of THF. To its suspension in DMSO (274 cm3) was
added 2.5 M KOH (18 cm3). The solution was stirred at 50 uC
for 24 h, and neutralized with 1 M HCl. The product was
extracted with CHCl3 and poured into methanol. Reprecipi-
tation from CHCl3 to methanol gave a yellow powder of 2b
(0.25 g): yield 25%, IR (KBr pellet/cm21) 3638 (nO–H),
960 (dtransHCLCH); dH (CDCl3) 1.38 (s, 18H, tert-butyl), 5.40
(s, 1H, OH), 6.72–7.61 (m, 4H, Ar). Anal. Calcd for
C22nH25n11OnBrn11 (n ~ 63): C, 68.5; H, 6.5; Br, 21.0.
Found: C, 71.1; H, 6.5; Br, 19.7%.
Electrochemical measurements
The voltammetric investigation was carried out in CH2Cl2 in
the presence of 0.1 M (n-C4H9)4NBF4 as the supporting
electrolyte and a small amount of (CH3)4NOH as the alkaline
solution with a platinum working microelectrode using a
function generator (Nikko Keisoku NFG-3) and a potentio-
galvanostat (NPGS-301). The differential pulse voltammetry
technique was also used since it permits the greatest resolution
of the signals associated with consecutive redox events, whose
potential is not very distant (an arbitary function generator
Hokuto Denko HB-105 and potentiostat/galvanostat HA-
501G). For coulometry, a large carbon felt electrode was used
with a digital coulomb meter (Nikko Keisoku NDCM-1) under
the application of 0.7 V (vs. Ag/AgCl). All the electrochemical
experiments were carried out in the absence of oxygen.
Poly(1,2,(6)-phenylenevinylene) (4). 2,6-Dibromostyrene was
polymerized with the Pd catalyst, and the polymer was purified
in the same manner as described above: yield 45%. IR (KBr
pellet/cm21) 968 (dtransHCLCH); dH (CDCl3) 6.78–7.60 (m, 5H,
phenyl). Anal. Calcd for C8nH5n11Brn11 (n ~ 21): C, 51.9; H,
2.7; Br, 45.3. Found: C, 51.9; H, 2.6; Br, 45.0%.
2,6-Distyryl-4-(3’,5’-tert-butyl-4’-acetoxyphenyl)toluene (9a).
Compound 5a (0.2203 g, 0.44 mmol) was reacted with styrene
J. Mater. Chem., 2002, 12, 3578–3584
3583

View Article Online
Oxidation
References and notes
(i) A small excess of (n-C4H9)4NOH was added to a CH2Cl2
solution (2 cm3) of 2b (12.2 mg, 20 monomer unit mmol L21
1
2
3
4
5
6
7
8
9
Magnetic Properties of Organic Materials, ed. P. M. Lahti, Marcel
Dekker Inc, New York, 1999.
)
Molecular Magnetism, ed. K. Ito and M. Kinoshita, Kodansha,
Tokyo and Gordon and Breach, Amsterdam, 2000.
Magnetism: Molecules to Materials II, ed. J. S. Miller and
M. D. Drillon, Wiley-VCH, Weinheim, 2001.
J. A. Crayston, J. N. Devine and J. C. Walton, Tetrahedron, 2000,
56, 7829.
R. J. Bushby, D. R. MaGill, K. M. Ng and N. Taylor, J. Mater.
Chem., 1997, 7, 2343.
L. Viadel, J. Carilla, E. Brillas, A. Labarta and L. Julia´, J. Mater.
Chem., 1998, 8, 1165.
H. Oka, T. Tamura, Y. Miura and Y. Teki, J. Mater. Chem., 1999,
9, 1227.
A. Rajca, J. Wongsriratanakul, S. Rajca and R. Cerny, Angew.
Chem., Int. Ed., 1998, 37, 1229.
K. K. Anderson and D. A. Dogherty, Adv. Mater, 1998, 10, 688.
and stirred for 1 h in a glove box. The solution was then
vigorously stirred with 1 cm3 of aqueous K3Fe(CN)6 (0.16 g,
12 equiv. to the phenolate) at room temperature. The solution
rapidly turned greenish brown after 10–20 min, which was
ascribed to phenoxyl radical formation. The organic layer was
washed with water and dried over anhydrous sodium sulfate to
give a solution of 1. (ii) All procedures were performed under
high vacuum. A small excess of (n-C4H9)4NOH was added to
a CH2Cl2 solution (2 cm3) of 2b (12.2 mg, 20 monomer unit
mmol L21) and solvent was removed after complete generation
of the polyanion. Subsequently a solution of ferrocenium
hexafluorophosphate (1.05 equiv. to the polymer) was added to
the solid polyanion at 195 K. The oxidation process was kept at
a low temperature for 6 h. The sample (solid or liquid) for
measurement was prepared by keeping the tube under vacuum.
10 C. Xie and P. M. Lahti, J. Polym. Sci., Part A: Polym. Chem.,
1999, 37, 779.
11 A. Rajca, S. Rajca and J. Wongsriratanakul, J. Am. Chem. Soc.,
1999, 121, 6308.
12 M. Miyasaka, T. Yamazaki, E. Tsuchida and H. Nishide,
Macromolecules, 2000, 33, 8211.
Magnetic measurement
13 H. Oka, T. Tamura, Y. Miura and Y. Teki, J. Mater. Chem., 2001,
11, 1364.
14 N. Mataga, Theor. Chim. Acta, 1968, 10, 372.
15 N. N. Tyutyulkov and S. C. Karabunarliev, Chem. Phys., 1987,
112, 293.
16 A. Rajca, J. Wongsriratanakul and S. Rajca, Science, 2001, 294,
1503.
17 J. Roncali, Chem. Rev., 1997, 97, 173.
18 Q.-T. Zhang and J. M. Tour, J. Am. Chem. Soc., 1997, 119, 9624.
19 Y. H. Kim, J. Polym. Sci., Part A: Polym. Chem., 1998, 36, 1685.
20 M. Jikei and M. Kakimoto, Prog. Polym. Sci., 2001, 26, 1233.
21 H. Nishide, T. Kaneko, T. Nii, K. Katoh, E. Tsuchida and
K. Yamaguchi, J. Am. Chem. Soc., 1995, 117, 548.
22 H. Nishide, T. Kaneko, T. Nii, K. Katoh, E. Tsuchida and
P. M. Lahti, J. Am. Chem. Soc., 1996, 118, 9695.
23 H. Nishide, T. Ozawa, M. Miyasaka and E. Tsuchida, J. Am.
Chem.Soc., 2001, 123, 5942.
Prepared phenoxyl polymer 1 was immediately transferred to a
diamagnetic capsule or a quartz tube after the oxidation.
Magnetization and static magnetic susceptibility were mea-
sured with a Quantum Design MPMS-7 SQUID magnet-
ometer. The magnetization was measured to be from 0.1 to 7 T
at 1.8, 2.0, 2.5, 3, and 5 K. The static magnetic susceptibility
was measured from 2 to 200 K at a field of 0.5 T. Ferromagnetic
magnetization ascribed to impurities (v1 ppm) was deter-
mined by the Honda–Owen plots and subtracted from the
overall magnetization. Diamagnetic susceptibility (xdia) of the
sample solution and the capsule was estimated by Curie plots of
magnetic susceptibility.37 The corrected magnetization data
were fitted to Brillouin functions using a self-consistent version
of the mean field approximation.
24 H. Nishide, M. Miyasaka and E. Tsuchida, Angew. Chem., Int.
Ed., 1998, 37, 2400.
25 H. Nishide, M. Miyasaka and E. Tsuchida, J. Org. Chem., 1998,
63, 7399.
Spectroscopic measurements
EPR spectra were recorded on a JEOL JES-TE200 EPR
spectrometer with 100 kHz field modulation. Spin concentra-
tions of each sample were determined by double integration of
the EPR signal standardized with that of the 2,2,6,6-
tetramethylpiperidine 1-oxide solution. UPS spectra were
recorded on a UPS spectrometer (Riken Keiki AC-1). The
polymer thin films were prepared by casting the CH2Cl2
solutions on a glass plate. Ith was estimated in a manner
analogous to that described in ref. 35. NMR, IR, UV-vis,
fluorescence and mass spectra were measured on a JEOL NMR
JMM-LA500, a JASCO FT/IR-5300, a Shimadzu UV-2000,
Hitachi F-4500, and Shimadzu GC-MS QP-5050 spectrometer,
respectively. Intrinsic viscosities were determined at 25 uC in
toluene using an Ubbelohde viscometer.
26 H. Meier and M. Lehmann, Angew. Chem., Int. Ed., 1998, 37, 643.
27 S. K. Deb, T. M. Maddux and L. Yu, J. Am. Chem. Soc., 1997,
119, 9079.
28 E. D´ıez-Barra, J. C. Garc´ıa-Mart´ınez and J. Rodr´ıguez-Lo´pez,
Tetrahedron Lett., 1999, 40, 8181.
29 M. Halim, J. N. G. Pillow, I. D. W. Samuel and P. L. Burn, Adv.
Mater, 1999, 11, 371.
30 H. Meier, M. Lehmann and U. Kolb, Chem. Eur. J., 2000, 6, 2462.
31 T. M. Londergan, Y.-J. You, M. E. Thompson and W. P. Weber,
Macromolecules, 1998, 31, 2784.
32 R. F. Heck, Org. React., 1982, 27, 345.
33 K. Tamao, K. Sumitani, Y. Kiso, M. Zenbayashi and M. Kumada,
Bull. Chem. Soc. Jpn., 1976, 49, 1958.
34 G. Maerkl and A. Merz, Synthesis, 1973, 295.
35 K. Seki, Optical Techniques to Characterize Polymer Systems, ed.
H. Baessler, Elsevier Science, Amsterdam, 1989, p. 115.
36 T. Kaneko, E. Ito, K. Seki, E. Tsuchida and H. Nishide, Polym. J.,
1996, 28, 182.
37 For example, xdia ~ 21.12 6 1027 emu for 1 given in Fig. 4,
which almost agrees with the xdia value calculated by Pascal’s
solvent constant. xdia ~ 21.24 6 1028 emu for the diamagnetic
capsule to hold the sample solution.
Acknowledgements
This work was partially supported by Grants-in-Aid for COE
Research ‘‘Molecular Nano-Engineering’’ from MEXT, Japan.
We thank Ryo Numazaki for help with synthetic and chemical
oxidation studies. M. M. expresses his thanks for a Research
Fellowship from the Japanese Society for the Promotion of
Science for Young Scientists.
38 In order to improve the result, we also used homogeneous
oxidizing reagents such as ferrocenium hexafluorophosphate.
However, the SQUID measurement gave only S ~ 2.5 and a
spin concentration of 0.31.
39 R. Nevile and A. Winther, Ber., 1880, 13, 975.
3584
J. Mater. Chem., 2002, 12, 3578–3584