Controlled radical polymerization of styrene in the presence of lithium molybdate(V) complexes and benzylic halides

Article abstract of DOI:10.1016/S0022-328X(99)00149-7

The new lithium molybdate(V) complexes [LiMo(NAr)₂(C-N)R] (C-N=C₆H₄(CH₂NMe₂)-2; R=(C-N) (5), Me (6), CH₂SiMe₃ (7), p-tolyl (8)), have been generated in situ from reaction of the corresponding molybdenum(VI) complexes [Mo(NAr)₂(C-N)R] (C-N=C₆H₄(CH₂NMe₂)-2; R=(C-N) (1), Me (2), CH₂SiMe₃ (3), p-tolyl (4)) with n-BuLi. The nature of these radical anions was studied by EPR spectroscopy. The spectra of toluene solutions of in situ prepared complexes 5-8 revealed the presence of two different paramagnetic species, i.e. a molybdenum compound with distinct g_iso- and A_iso-values and an unidentified radical with a sometimes strong signal at g=1.986±0.001, lacking any hyperfine coupling. Extended Hueckel calculations on the crystal structure of 5 showed that the single electron occupies a molybdenum centered orbital, merely d_x²-y² in character. In situ prepared complexes 5-8 were successfully applied in the atom transfer radical polymerization (ATRP) of styrene using benzyl chloride as the initiator. The efficiency of the benzyl chloride initiator is rather poor (6-18%). Reaction of the lithium molybdate(V) complex 5 with (α-chloroethyl)benzene and (α-bromoethyl)benzene resulted in the formation of 1, LiCl and LiBr, respectively. The molecular weights as well as the molecular weight distributions show that the catalytic system, BzCl/5-8, catalyses styrene polymerization successfully but does not exercise much control over the polymerization reaction due to the poor initiator efficiency of benzyl chloride and probably the extreme air-sensitivity of the lithium molybdate(V) compounds. The unidentified radical (g=1.986±0.001) is unable to initiate radical polymerization but possibly influences the ATRP activity.

Full text of DOI:10.1016/S0022-328X(99)00149-7

www.elsevier.nl/locate/jorganchem
Journal of Organometallic Chemistry 584 (1999) 246–253
Controlled radical polymerization of styrene in the presence of
lithium molybdate(V) complexes and benzylic halides
Jim A.M. Brandts a, Patrick van de Geijn a, Ernst E. van Faassen b, Jaap Boersma a,
Gerard van Koten a,*
a Debye Institute, Department of Metal-Mediated Synthesis, Utrecht Uni6ersity, Padualaan 8, NL-3584 CH, Utrecht, The Netherlands
b Debye Institute, Department of Atomic and Interface Physics, Utrecht Uni6ersity, Princetonlaan 4, NL-3584 CC, Utrecht, The Netherlands
Received 1 January 1999; received in revised form 16 February 1999
Abstract
The new lithium molybdate(V) complexes [LiMo(NAr)2(C–N)R] (C–N=C6H4(CH2NMe2)-2; R=(C–N) (5), Me (6),
CH2SiMe3 (7), p-tolyl (8)), have been generated in situ from reaction of the corresponding molybdenum(VI) complexes
[Mo(NAr)2(C–N)R] (C–N=C6H4(CH2NMe2)-2; R=(C–N) (1), Me (2), CH2SiMe3 (3), p-tolyl (4)) with n-BuLi. The nature of
these radical anions was studied by EPR spectroscopy. The spectra of toluene solutions of in situ prepared complexes 5–8
revealed the presence of two different paramagnetic species, i.e. a molybdenum compound with distinct giso- and Aiso-values and
an unidentified radical with a sometimes strong signal at g=1.98690.001, lacking any hyperfine coupling. Extended Hu¨ckel
calculations on the crystal structure of 5 showed that the single electron occupies a molybdenum centered orbital, merely dx
2
2
−y
in character. In situ prepared complexes 5–8 were successfully applied in the atom transfer radical polymerization (ATRP) of
styrene using benzyl chloride as the initiator. The efficiency of the benzyl chloride initiator is rather poor (6–18%). Reaction of
the lithium molybdate(V) complex 5 with (a-chloroethyl)benzene and (a-bromoethyl)benzene resulted in the formation of 1, LiCl
and LiBr, respectively. The molecular weights as well as the molecular weight distributions show that the catalytic system,
BzCl/5–8, catalyses styrene polymerization successfully but does not exercise much control over the polymerization reaction due
to the poor initiator efficiency of benzyl chloride and probably the extreme air-sensitivity of the lithium molybdate(V) compounds.
The unidentified radical (g=1.98690.001) is unable to initiate radical polymerization but possibly influences the ATRP activity.
© 1999 Elsevier Science S.A. All rights reserved.
Keywords: Molybdate; Alkylimido; EPR spectroscopy; MO calculations; Styrene polymerization
1. Introduction
reactions the problem of radical coupling and dispro-
portionation is circumvented by the reversible reac-
tion of the polymer radicals with a persistent radical
to a dormant species, see Eq. (1). The high degree of
control arises from a permanent low concentration of
radicals, which suppresses bimolecular termination
processes. Thermodynamically the equilibrium in Eq.
(1) must be on the side of the dormant species (left)
while kinetically the equilibrium reactions must be
faster than the rate of propagation in order to get
polymer chains with comparable length [3].
The ‘living/controlled’ [1] character of atom transfer
radical polymerization (ATRP) makes it possible to
synthesize well-defined polymers [2] because irre-
versible side reactions like chain transfer or chain ter-
mination are suppressed. In these polymerization
* Corresponding author. Tel.: +31-30-2533120; fax: +31-30-
2523615.
E-mail address: g.vankoten@chem.uu.nl (G. van Koten)
0022-328X/99/$ - see front matter © 1999 Elsevier Science S.A. All rights reserved.
PII: S0022-328X(99)00149-7

J.A.M. Brandts et al. / Journal of Organometallic Chemistry 584 (1999) 246–253
247
Scheme 1. Synthesis of the paramagnetic lithium molybdate(V) complexes 5–8.
halide donors and are able to reform the dormant
species (see Eq. (1)), deactivation step with kdeact
.
2. Results
2.1. Synthesis of catalysts
The molybdenum(VI) complexes [Mo(NAr)2(C–
N)R] (C–N=C6H4(CH2NMe2)-2; R=(C–N) (1), Me
(2), CH2SiMe3 (3), p-tolyl (4)) were synthesized via the
reaction of [Mo(NAr)2Cl2(DME)] with half an equiva-
lent of [Zn(C–N)2] and subsequent transmetallation
reactions with RLi or RMgCl, according to literature
procedures [9]. These complexes react with n-BuLi in
toluene and eventually form the paramagnetic Mo(V)
molybdate species with general formula [Li-
Mo(NAr)2(C–N)R] (C–N=C6H4(CH2NMe2)-2; R=
(C–N) (5), Me (6), CH2SiMe3 (7), p-tolyl (8)) (see
Scheme 1).
Compound 5 can be isolated as a dark crystalline
material while complexes 6–8 are brown, very soluble
oils. The solid state structure of 5 has been elucidated
by single-crystal X-ray analysis and reveals two bridg-
ing (C–N) ligands (via Cipso of the C–N-ligands) be-
tween the molybdenum and lithium centers while the
two imidoaryl groups on the molybdenum center are
bonded perpendicularly to the Cipso–Mo–Cipso plane
[9].
(1)
Although bimolecular decomposition always occurs,
it becomes negligible because of the persistent radical
effect [3a,4], if the persistent radical species is stable.
The final molecular weight of the polymer is determined
by the amount of initiator.
A number of low oxidation state metal complexes
like ruthenium(II) [5], copper(I) [6], iron(II) [7] and
nickel(II) [8] are known to be effective ATRP catalysts.
These catalysts influence the equilibrium between dor-
mant and active species and all have an accessible
one-electron redox couple that promotes atom transfer.
All are able to accept a halide ligand, have low affinity
for alkyl radicals and contain a metal with low Lewis
acidity. For effective initiation the initiation rate must
be faster or comparable to the propagation rate and an
alkyl halide must be used that structurally resembles the
monomer [2a].
High oxidation state transition metal complexes, al-
though often highly Lewis acidic, might result in faster
polymerization reactions and suppressed side reactions
by influencing the equilibrium mentioned in Eq. (1).
Therefore, we studied the application of the newly
developed lithium molybdate(V) complexes of general
formula [LiMo(NAr)2(C–N)R] (C–N=C6H4(CH2-
NMe2)-2; R=(C–N) (5), Me (6), CH2SiMe3 (7), p-tolyl
(8)) as catalysts for the ATRP of styrene using benzyl
chloride as the initiator. Although the lithium molyb-
date(VI) halide compounds (formed after one electron
oxidation of the corresponding lithium molybdate(V)
complexes) are no odd electron species (as in the other
metal containing persistent radical species) [5–8], the
lithium molybdate(VI) halide compounds are good
Spectroscopic characterization of the clear brown
toluene solutions of 5–8 by NMR only showed very
broad residual solvents peaks but lacked any reso-
nances related to 5–8, corroborating the presence of a
paramagnetic molybdate(V) species. This conclusion
was supported by the results of the EPR measurements
of toluene solutions of these compounds which revealed
the presence of two different paramagnetic species, i.e.
one molybdenum compound showing the characteristic
six line splitting pattern [10] due to 95Mo+97Mo (both
isotopes having I=5/2 and combined natural abun-
dance of 25%) around a central line arising from 98Mo
(I=0, 75% natural abundance), with various giso- and
dance of 25%) around a central line arising from 98Mo
(I=0, 75% natural abundance), with various giso- and

248
J.A.M. Brandts et al. / Journal of Organometallic Chemistry 584 (1999) 246–253
Table 1
Aiso-values (vide infra) and one, sometimes, strong sig-
nal at g=1.98690.001 which shows no hyperfine
coupling.
Experimentally determined EPR parameters for lithium molyb-
date(V) complexes (5–8) in toluene
During the reaction of 1–4 with n-BuLi the forma-
tion of 1-butene, which must result from b-H elimina-
tion reactions, is detected. This indicates that at least
part of the n-butyl nucleophiles were bonded to the
Mo-center. In fact, in the presence of 1,2-
dimethoxyethane (DME), the intermediate lithium
5
6
7
8
Conc. (mM)
giso
Aiso (G)
8
10
8
15
1.976
a
a
1.972
1.977
34.6590.05
30.1690.08 30.6590.05
a Hyperfine coupling unresolved because the Mo(V) spectrum was
organomolybdate(VI)
complexes
[Li(DME)nMo-
coinciding with a strong resonance peak at g=1.98690.001.
(NAr)2(C–N)(n-Bu)(R)] (R=Me, CH2SiMe3, p-tolyl;
n=1–2), formed by the reactions of 2–4 with n-BuLi,
can be isolated and spectroscopically characterized at
temperatures below −10°C1. Because the amounts of
1-butene formed during the reactions in toluene were
not quantitative, b-H elimination may not be the only
pathway that eventually leads to the formation of 5–8.
The Mo–H complexes2 which must have been formed
after b-H elimination, are presumably unstable and will
form 5–8 after H-elimination. Alternative routes in-
volve direct butyl radical formation3 after Mo–C bond
cleavage or innersphere single electron transfer between
1–4 and n-BuLi clusters [11].
unidentified compound(s) is stable for days and is
formed independently of the solvent used. The radical
is not molybdenum centered since the characteristic six
line hyperfine splitting is absent. Because the reactions
of 1–4 with n-BuLi result in products with similar
spectroscopic properties we believe that 6–8 are struc-
turally related to 5.
2.2. EPR and extended Hu¨ckel calculations
Fig. 1 shows the EPR spectra of in situ generated 5,
7 and 8 in toluene. These spectra clearly show the
characteristic six line spectra around a central line
caused by the interaction of the Mo nucleus (I=0 and
I=5/2) with the unpaired electron while hyperfine in-
teraction of the nitrogens is not visible [12].
The origin of the singlet resonance with g=1.9869
0.001 is still unknown. The EPR signal from the(se)
The line widths of the spectra vary between 7 and 8
G and further experiments showed that these line
broadenings are not caused by spin–spin interactions.
An increase of the concentration from approximately 8
mM to 20 mM does not lead to a change in the
spectrum. The hyperfine coupling constant, A, is given
by the splitting between the sequential lines. These lines
are not equally spaced, e.g. splittings are typically rang-
ing between 30 and 40 G, and a second order perturba-
tion expansion in the hyperfine constant A is required
to determine the accurate values of the g-factor and A
[13]. From the experimental spectra the widths and line
positions were taken and fitted (see Table 1). The sign
of A can not be determined from the spectra.
Addition of air to the toluene solutions of 8 reduces
the intensity of the singlet resonance (g=1.98690.001)
with 30%, whereas the signal of the molybdenum com-
pound is reduced by almost 90%. Finally, a strong
solvent effect on the hyperfine coupling of 5 was
observed.
These EPR results show that the unpaired electron in
the molybdenum complexes is strongly interacting with
the molybdenum nucleus. The location of this electron
is supported by the results of the extended Hu¨ckel
calculations4 using the X-ray data of the structure of 5.
These calculations showed the HOMO of the complex
Fig. 1. EPR spectra of in situ prepared 5 (b) and 8 (a) in toluene.
1 The complexes [Li(DME)nMo(NAr)2(C–N)(n-Bu)(R)] (R=Me,
CH2SiMe3, p-tolyl; n=1–2) can be isolated as yellow solids at
temperatures below 0°C, but decompose in deuterated solvents at
ambient temperature.
2 We were unable to detect traces of the ‘Mo–H’ species during the
process of b-H elimination (monitored by NMR), probably due to
the instability of the Mo–H bond.
3 After the reaction all volatiles were collected in a cold trap and
the mixture was analyzed by NMR and GC–MS. We were unable to
detect any products arising from butyl radicals (C8H18 fragments or
’
products formed after reaction of Bu with toluene), even when the
4 Extended Hu¨ckel calculations were performed with a Wolfberg–
Helmholtz constant of 1.75 and STO-6G basis set and Alvarez
Collected Parameters containing a double zeta basis set.
reaction was carried out in the presence of a radical trap like
TEMPO. We only observed the formation of 1-butene.

J.A.M. Brandts et al. / Journal of Organometallic Chemistry 584 (1999) 246–253
249
Scheme 3. Initiation, propagation and dormant species formation.
Fig. 2. The single electron in 5 occupies an orbital that is basically
dx in character as calculated by extended Hu¨ckel calculations.
2
2
−y
Hydrogens are omitted for clarity.
The reaction of I and II with compound 5 resulted in
the gradual formation of 1 and LiBr and LiCl, respec-
tively. For this reaction various reaction pathways can
be envisaged. Firstly, nucleophilic substitution could be
expected leading to the formation of the C–C coupling
product (C–N)-benzyl and a molybdate with an X
(X=Cl, Br) ligand. Secondly, an electron transfer reac-
tion could occur as shown in Scheme 2 in which the
molybdate(V) ([MoV]−) is oxidized to a molybdate(VI)
species ([MoVIX]−) with concomitant formation of a
benzyl radical. This reaction is typical for the initiation
step in ATRP reactions. The benzyl radicals formed
react with the alkene producing product radicals. Both
radicals can react back with [MoVIX]− to give 5 and
the corresponding alkyl halide (dormant species,
Scheme 3).
to be a molybdenum centered, non-bonding orbital of
dx
character which is situated on and perpendicu-
2
2
lar −toy the Mo···Li axis (see Fig. 2).
Complexes 5–8 appear to be very sensitive towards
electrophilic reagents like O2 and vide infra (a-bro-
moethyl)benzene and (a-chloroethyl)benzene.
2.3. Reaction of the lithium molybdate 5 with benzyl
halides
Compound 5 was reacted with different alkyl halides
(I: (a-bromoethyl)benzene; II: (a-chloroethyl)benzene;
III: benzyl chloride (BzCl)) and the reaction products
were analyzed (see Scheme 2).
Scheme 2. Reaction of 5 with various alkyl halides.

250
J.A.M. Brandts et al. / Journal of Organometallic Chemistry 584 (1999) 246–253
Any dimerization reaction of the benzyl or product
radicals would lead to irreversible formation of
[MoVIX]− (persistent radical effect). However, we
found that the reaction of the a-substituted benzyl
halides I and II, respectively, with 5 resulted in the
gradual formation of 1 and LiX. This forward reaction
produces the neutral Mo(VI) compound (1) after LiX
elimination. When benzyl chloride (III) was used, nei-
ther 1 nor LiCl was formed. Here, the equilibrium
k1/k−1 is on the left side and assures that the concen-
tration of [MoVIX]− remains low and that the reaction
in which 1 and LiCl are formed, is suppressed.
The
reactions
[MoV]− “[MoVIX]−
and
Fig. 4. Semilogarithmic plot of styrene conversion vs. time. (i) With-
out BzCl, (ii) without molybdenum compound (5), (iii) without BzCl
nor 5, and (iv) average in the presence of BzCl and 5.
[MoVIX]− “1 have different rates. It is known that the
reactivity of the alkyl halides, and thus the position of
equilibrium k1/k−1, is reflected in the bond dissociation
energy (BDE) of these compounds [6c]. The order of
BDE in I–III is IBIIBIII (for BzCl the BDE amounts
to 70 kcal mol−1) [14]. The higher the BDE, the more
the equilibrium will be on the left side: this implies that
in the case of I and II the equilibrium k1/k−1 is so far
to the right side that the reaction to the MoVI species 1
is possible. For effective ATRP the equilibrium needed,
has to be on the side of the dormant species. Appar-
ently, the use of III realizes this and the permanent low
concentration of [MoVIX]− assures that the forward
reaction is suppressed. Reactions of I–III with com-
pounds 6–8 gave similar results.
M in toluene at 80°C), polymerization of styrene is still
observed after 22 h (vide infra).
2.4. Atom transfer radical polymerization with 5/benzyl
chloride
In a typical experiment a mixture of in situ prepared
5 and 1 equivalent of BzCl, was added to a mixture of
30 equivalents of mesitylene (internal standard) and 100
equivalents of styrene ([St]0=3 M, [5]=0.03 M,
[BzCl]0=0.03 M in toluene at 80°C). The polymeriza-
tion experiments were followed by GC using mesitylene
as internal standard. After 22 h the reaction was
stopped and the resulting polystyrene (PS) polymers
were isolated and characterized by GPC.
In order to examine the influence of the BzCl initia-
tor and the catalyst, reactions were performed in
toluene at 80°C under various experimental conditions.
The semilogarithmic plot of ln([St]0/[St]) versus time is
shown in Fig. 4 and the results are summarized in
Table 2.
When styrene (St) was added to the reaction mixture
([St]0=3 M, [5]=0.03 M, [XC(Ph)HR%]0=0.03 M
(X=Br, Cl) in toluene at 80°C) and the polymerization
was monitored, fast initial polymerization was observed
but this polymerization activity ceased within 1 h,
probably because recombination of the reactive radical
chains occurred (see Fig. 3).
However, when St was added to the reaction mixture
of 5 and BzCl ([St]0=3 M, [5]=0.03 M, [BzCl]0=0.03
PS yields are low if polymerizations are performed
without 5 (Entry 2), without BzCl (Entry 3) or without
both catalyst components (Entry 1). However, with the
homogeneous initiating system, 5/BzCl, conversion
reached 38% after 22 h (Entry 4). The molecular
weights, determined during this experiment (Entry 4)
increased during the polymerization reaction, although
not linearly. The use of half the amount of [BzCl]0
afforded a less active system but yielded PS with a
higher molecular weight. These findings show a trend
which is consistent with the fact that the initial amount
of alkyl halide [RX]0 determines the number of initiated
chains: the polymerization rate is first order in [RX]0
while the molecular weights are reciprocally dependent
upon [RX]0. Because the initiator efficiency is rather
low (99%) the influence of variations in [BzCl]0 is
limited.
Fig. 3. Semilogarithmic plot of styrene conversion vs. time using
different initiators ([St]0=3 M, [5]=0.03 M, [XC(Ph)HR%]0=0.03 M
in toluene at 80°C).

J.A.M. Brandts et al. / Journal of Organometallic Chemistry 584 (1999) 246–253
251
Table 2
Styrene polymerizations in toluene with various components at 80°C
Entry
[BzCl]0 (×102 mol l−1
)
[5] (×102 mol l−1
)
% yield PS after 22 h
Mw (PDI)×10−3 (g mol−1
)
1
2
3
4
5
0
3
0
3
0
0
3
3
3
3
5
11
38
33
n.d.a
n.d.
102 (1.89)
44 (1.57)
51 (1.48)
1.5
a Not determined.
influence of this unidentified radical on the BzCl/5–8
system and the styrene polymerization is subject of
future investigations.
Molybdenum(V) compounds 6, 7 and 8 were also
tested for their ATRP activity with benzyl chloride as
the initiator. The kinetic plots are shown in Figs. 5–7,
respectively, and the results are outlined in Table 3.
The combination of BzCl and molybdenum com-
pounds accelerates the polymerization reaction consid-
erably. Yields of PS are in the range 38–60% and the
molecular weights are high (32 000–85 000 g mol−1).
The PDIs are rather high (1.57–1.71) but remain con-
stant for the different experiments. With conversions
between 38 and 60%, molecular weights of the resulting
PS were expected to be between 4000 and 6200 g mol−1
but the actual values are much higher which indicates
that the initiator efficiency is poor. Only a small per-
centage (6–18%) of the benzyl chloride initiator is
participating in the polymerization reaction. Indeed,
considerable amounts of unreacted BzCl were found
(GC) during and after the polymerization process5.
The molecular weights and particularly the PDIs of
the resulting polymers, shown in Table 3, show that the
catalytic systems used here obey the theoretical rules
that are found in other ATRP systems [2a], but poor
initiator efficiency results in poor control during the
polymerization reaction. Moreover, the ATRP experi-
ments were difficult to reproduce and therefore the
standard deviations of both the styrene conversions and
the molecular weights are large. As 5–8 are very sensi-
tive towards air, any possible air-leakage in the system
may be a reason for the observed differences in reactiv-
ity of the same catalytic system. Furthermore, decom-
position of the molybdenum compounds in time may be
an additional problem.
3. Experimental
3.1. General methods
All reactions were carried out under an atmosphere
of dry N2, using standard Schlenk techniques. All sol-
vents were distilled from sodium under nitrogen, prior
to use, except CH2Cl2, which was distilled from CaH2.
Commercially available reagents were distilled prior to
use. [Mo(NAr)2(C–N)R] (C–N=C6H4(CH2NMe2)-2;
R=(C–N) (1), Me (2), CH2SiMe3 (3), p-tolyl (4)) were
synthesized according to literature procedures [9]. The
1H- and 13C{1H}-NMR spectra were recorded on a
Bruker AC 300 spectrometer, in benzene-d6 at room
temperature (r.t.) unless otherwise indicated. GC mea-
surements were performed on a Philips PU 4600 Gas
Chromatograph with a J&W Scientific liquid-phase
DB-5 column (30 m×0.320 mm, 0.25 mm film thick-
ness). GPC measurements were taken on a Jordi-Gel
DVB Mixed Bed, 300 mm and I.D. 7.8 mm column,
using a Thermo Separation Products P200 pump and
UV2000 and Shodex RI-71 detectors.
The influence of the unidentified radical (with g=
1.98690.001) on the polymerization appears to be
modest. In polymerization experiments in which the in
situ prepared molybdenum complexes are used without
the BzCl initiator, only low PS yields are obtained.
Apparently, the unidentified radical is not capable to
initiate radical polymerization. However, because the
binary systems BzCl/5–8 all show similar polymeriza-
tion behavior with features characteristic for ATRP
[2a], the unidentified radical possibly does play a role in
the ATRP activity observed. The identification and the
Fig. 5. Semilogarithmic plot of ln([St]0/[St]) vs. time with [St]0=3 M
using [LiMo(NAr)2(C–N)Me] (6) (=0.03 mM).

252
J.A.M. Brandts et al. / Journal of Organometallic Chemistry 584 (1999) 246–253
Fig. 7. Semilogarithmic plot of ln([St]0/[St]) vs. time with [St]0=3 M
using LiMo(NAr)2(C–N)(p-tolyl)] (8) (=0.03 mM).
Fig. 6. Semilogarithmic plot of ln([St]0/[St]) vs. time with [St]0=3 M
using [LiMo(NAr)2(C–N)(CH2SiMe3)] (7) (=0.03 mM).
[Mo(NAr)2(C–N)R] (C–N=C6H4(CH2NMe2)-2; R=
(C–N) (1), Me (2), CH2SiMe3 (3), p-tolyl (4)) (0.50
mmol), respectively, in toluene (8 ml) to which 0.77 ml
of a 1.60 M solution of n-BuLi (0.48 mmol) was added
at −78°C. The reaction mixture was allowed to warm
to r.t. and was stirred overnight. After 18 h, 0.063 g of
benzyl chloride (0.50 mmol) was added and the reaction
mixture was stirred for 15 min.
In a Schlenk vessel was mixed 5.21 g of styrene (50
mmol) and 1.80 g mesitylene (15 mmol) as an internal
standard. The solution was warmed to 80°C and the
catalyst solution was quickly added to this mixture. The
polymerization reaction was followed from this point
on. Samples of 0.5 ml were taken from the reaction
mixture at 1 h intervals and were quenched with 5 ml of
methanol. The white precipitate was removed by filtra-
tion and the clear solution was measured by GC. The
reaction was stopped after 22 h by quenching with
methanol. The polymer was collected, dried in vacuo
and analyzed by NMR and GPC.
3.2. EPR spectroscopy
The EPR spectra were recorded at r.t. on a modified
Bruker ESP 300 spectrometer operating near 9.4 GHz.
The samples were enclosed in carefully sealed quartz
tubes (¥=2 mm) developed for manipulation of liquid
samples under conditions free of oxygen and water. The
tubes were placed in the center of a Bruker ST1 res-
onator (operating in the TE102 mode with unloaded
Q=3800). The magnetic field was centered around
3380 G, with a sweep width of 500 G. The magnetic
field was modulated with a frequency of 100 kHz and
amplitude of 2.5 G. The microwave power was 20–40
mW, gain=2×104, time constant=5 ms, and ADC
conversion time=10 ms. The EPR spectra contain 8k
field points in order to achieve good field resolution.
The concentrations contained 8–15 mM of molybde-
num and toluene were used as a solvent throughout.
The values of the g-factor and hyperfine splitting A
were extracted from the experimental EPR spectra by
fitting the experimental line positions to those obtained
from a perturbation expansion up to second order in
the hyperfine coupling [13]. This procedure leaves the
sign of the hyperfine coupling undetermined.
Acknowledgements
J.A.M.B., J.B. and G.v.K. gratefully acknowledge Dr
G.J.M. Gruter (DSM Research) and Dr F.H. van der
Steen (Shell, Amsterdam) for fruitful discussions. Fi-
nancial support was granted by the Dutch Technology
3.3. General polymerization procedure
Compounds 5–8 were prepared in situ by dissolving
Table 3
Results of ATRP of styrene using 5–8
b
Compound
Average styrene conversion (%)a
Mw (g mol−1) expected
Mw×10−3 (g mol−1 a
)
PDIa
BzCleff (%)
5
6
7
8
38 (7)
55 (7)
50 (5)
60 (10)
3958
5728
5207
6249
44 (7)
1.57 (0.02)
1.69 (0.04)
1.62 (0.07)
1.71 (0.01)
9
18
6
32 (3.5)
85 (15)
48 (7)
13
a The estimated standard deviations are shown in parentheses.
b Benzyl chloride initiator efficiency=Mw/Mw expected.

J.A.M. Brandts et al. / Journal of Organometallic Chemistry 584 (1999) 246–253
253
[6] (a) T.E. Patten, J. Xia, T. Abernathy, K Matyjaszewski, Science
272 (1996) 866. (b) J.-S. Wang, K Matyjaszewski, J. Am. Chem.
Soc. 117 (1995) 5614. (c) J.-S. Wang, K Matyjaszewski, Macro-
molecules 28 (1995) 7901. (d) T. Grimaud, K Matyjaszewski,
Macromolecules 30 (1997) 2216.
Foundation (STW) with financial aid from the Council
for Chemical Science of the Netherlands Organization
for Scientific Research (CW/NWO).
[7] (a) T. Ando, M. Kato, M. Kamigaito, M. Sawamoto, Macro-
molecules 29 (1996) 1070. (b) T. Ando, M. Kamigaito, M.
Sawamoto, Macromolecules 30 (1997) 4507.
[8] C. Granel, Ph. Dubois, R. Je´roˆme, Ph. Teyssie´, Macromolecules
29 (1996) 8576.
References
[1] The term ‘living/controlled’ is used here as defined by Maty-
jaszewski, in: K. Matyjaszewski, A.H.E. Mu¨ller, Polym. Prepr.
Am Chem. Soc. Div. Polym. Chem. 38 (1997) 6.
[2] (a) T.E. Patten, K. Matyjaszewski, Adv. Mater. 10 (1998) 901.
(b) K. Matyjaszewski, J. Phys. Org. Chem. 8 (1995) 197. (c) K.
Matyjaszewski, S. Gaynor, D. Greszta, D. Mardare, T. Shige-
moto, Macromol. Symp. 98 (1995) 73. (d) C.J. Hawker, Acc.
Chem. Res. 30 (1997) 373. (e) C.J. Hawker, Trends Polym. Sci.
4 (1996) 183. (f) M. Sawamoto, M. Kamigaito, Trends Polym.
Sci. 4 (1996) 371.
[9] J.A.M. Brandts, M. van Leur, R.A. Gossage, J. Boersma, A.L.
Spek, G. van Koten, Organometallics, accepted for publication.
[10] (a) C. Balagopalakrishna, J.T. Kimbrough, T.D. Westmoreland,
Inorg. Chem. 35 (1996) 7758. (b) P. Legzdins, S.F. Sayers,
Organometallics 15 (1996) 3907. (c) K.R. Barnard, M. Bruck, S.
Huber, C. Grittini, J.H. Enemark, R.W. Gable, A.G. Wedd,
Inorg. Chem. 36 (1997) 637.
[11] (a) M. Tamura, J. K. Kochi, J. Am. Chem. Soc. 93 (1971) 1487.
(b) F.A. Cotton, G. Wilkinson, Advanced Inorganic Chemistry,
5th ed., 1988, pp. 1306–1315.
[3] (a) K. Matyjaszewski, T.E. Patten, J. Xia, J. Am. Chem. Soc.
119 (1997) 674. (b) K. Matyjaszewski, Macromol. Symp. 111
(1996) 47.
[12] We estimated the hyperfine constant to be about 2.8 G, see J.M.
Assour, J. Chem. Phys. 43 (1965) 2477.
[13] N. Atherton, Electron Spin Resonance, Wiley, New York, 1973
Ch. 5.
[4] H. Fischer, Macromolecules 30 (1997) 5666.
[5] (a) M. Kato, M. Kamigaito, M. Sawamoto, T. Higashimura,
Macromolecules 28 (1995) 1721. (b) J. Ueda, M. Matsuyama, M.
Kamigaito, M. Sawamoto, Macromolecules 31 (1998) 557.
[14] TRC Thermodynamic Tables Hydrocarbons & Non-Hydrocar-
bons, Thermodynamics Research Center, Texas A&M Univer-
sity System, College Station, TX 7783-3111, USA.
.
.