Organometallic intermediates in the controlled radical polymerization of styrene by α-diimine iron catalysts

Article abstract of DOI:10.1021/om700395h

The α-diimine iron complexes ^R′,R″[N,N] FeCl₂ (^R′,R″[N,N] = R′-N=CR″- CR =N-R′) are efficient catalysts for the atom-transfer radical polymerization (ATRP) of styrene when R′ and R″ are electrondonating substituents and favor catalytic chain transfer (CCT) when electron-withdrawing substituents are employed. An organometallic pathway, alongside a halogen-atom-transfer equilibrium, is proposed to mediate the observed reactivity. The model alkyl complexes, ^R′,R″[N,N] FeCl₂(R), where R = PhCH₂ and Ph(Me)CH, were generated via treatment of ^R′,R″[N, N]FeCl₃ with RMgCl. The alkyl derivatives obtained from intermediate spin-state ^R′,R″[N,N]FeCl₃ complexes were found to be stable to ca. -30 °C and favor CCT, whereas the alkyl derivatives derived from high-spin-state trichloride precursors are unstable above -78 °C and favor ATRP. Azo-initiated polymerizations of styrene are moderately controlled by a-diimine iron catalysts. The role of organometallic-mediated radical polymerization (OMRP) in the controlled polymerization of styrene is discussed: an analysis of the radical concentrations generated by the competing OMRP and ATRP equilibria indicates that the halogenophilicity of the Fe(II) catalyst dominates the carbophilic alkyl radical-trapping capacity of the Fe(II) species in this a-diimine catalyst system.

Full text of DOI:10.1021/om700395h

Organometallics 2007, 26, 4725-4730
4725
Organometallic Intermediates in the Controlled Radical
Polymerization of Styrene by r-Diimine Iron Catalysts
Michael P. Shaver, Laura E. N. Allan, and Vernon C. Gibson*
Department of Chemistry, Imperial College, Exhibition Road, South Kensington, London, SW7 2AZ, U.K.
ReceiVed April 24, 2007
The R-diimine iron complexes R′,R′′[N,N]FeCl2 (R′,R′′[N,N] ) R′-NdCR′′-CR′′dN-R′) are efficient
catalysts for the atom-transfer radical polymerization (ATRP) of styrene when R′ and R′′ are electron-
donating substituents and favor catalytic chain transfer (CCT) when electron-withdrawing substituents
are employed. An organometallic pathway, alongside a halogen-atom-transfer equilibrium, is proposed
to mediate the observed reactivity. The model alkyl complexes, R′,R′′[N,N]FeCl2(R), where R ) PhCH2
and Ph(Me)CH, were generated via treatment of R′,R′′[N,N]FeCl3 with RMgCl. The alkyl derivatives obtained
from intermediate spin-state R′,R′′[N,N]FeCl3 complexes were found to be stable to ca. -30 °C and favor
CCT, whereas the alkyl derivatives derived from high-spin-state trichloride precursors are unstable above
-78 °C and favor ATRP. Azo-initiated polymerizations of styrene are moderately controlled by R-diimine
iron catalysts. The role of organometallic-mediated radical polymerization (OMRP) in the controlled
polymerization of styrene is discussed: an analysis of the radical concentrations generated by the competing
OMRP and ATRP equilibria indicates that the halogenophilicity of the Fe(II) catalyst dominates the
carbophilic alkyl radical-trapping capacity of the Fe(II) species in this R-diimine catalyst system.
Scheme 1. Possible Pathways for r-Diimine Iron-Mediated
Introduction
Radical Polymerization of Styrene
The development of controlled/“living” radical polymerization
reactions has provided a technologically important advance in
synthetic polymer chemistry. Key milestones include the
introduction of atom-transfer radical polymerization,1,2 nitroxide-
mediated radical polymerization,3 and reversible addition-
fragmentation chain transfer.4,5 In each of these reactions control
is achieved by the reversible deactivation of growing polymer
chains, maintaining a low concentration of radicals.
In our own studies we have shown that R-diimine iron
complexes R′,R′′[N,N]FeCl2 (R′,R′′[N,N] ) R′-NdCR′′-CR′′d
N-R′) are able to mediate the polymerization of styrene either
through an atom-transfer radical polymerization (ATRP) mech-
anism or via a catalytic chain-transfer (CCT) process.6 Inves-
tigations into substituted R-diimine complexes through com-
putational, reactivity, and mechanistic studies have indicated
that this switch in polymerization mechanism correlates with a
change in metal spin-state in the corresponding R′,R′′[N,N]FeCl3
complexes, with electron-withdrawing groups stabilizing inter-
mediate spin-state (S ) 3/2) complexes and electron-donating
groups supporting high-spin (S ) 5/2) Fe(III) species. It is
proposed that these divergent reactivities arise as a consequence
of the interplay of the ATRP and organometallic-mediated
radical polymerization (OMRP) equilibria shown in Scheme 1.
centration of radicals in solution and thus reducing bimolecular
termination reactions. The degree of control via ATRP depends
upon the propensity of the Fe(II) species to abstract a halogen
atom from an alkyl halide initiator or propagating chain end,
which can be viewed to be a function of the Fe(II) center’s
halogenophilicity. The radicals present in solution can also be
trapped by the Fe(II) complex to form a new Fe(III) organo-
metallic species, establishing an OMRP equilibrium; the analo-
gous idiom for the predilection of the Fe(II) species for a carbon-
based radical can be termed its carbophilicity. Olefin-terminated
polymer chains can then be generated via â-hydrogen elimina-
tion from this iron-alkyl species or, alternatively, through direct
hydrogen transfer from the organoradical to the Fe(II) catalyst.
From our studies to date, it has become apparent that ATRP or
CCT products can be favored depending upon the relative
halogenophilicity versus carbophilicity of these species which
is determined by the effect of the imino or ligand backbone
substituents on the spin-state of the iron centers.6-8
The halogen-atom transfer equilibrium established in ATRP
affords chlorine-capped polymer chains, minimizing the con-
* Corresponding author. Fax: 020 7594 5810. Tel: 020 7594 5830.
E-mail: v.gibson@imperial.ac.uk.
(1) Matyjaszewski, K.; Xia, J. Chem. ReV. 2001, 101, 2921.
(2) Kamigaito, M.; Ando, T.; Sawamoto, M. Chem. ReV. 2001, 101, 3689.
(3) Hawker, C. J.; Bosman, A. W.; Harth, E. Chem. ReV. 2001, 101,
3661.
(4) Favier, A.; Charreyre, M.-T. Macromol. Rapid Commun. 2006, 27,
653.
(5) Moad, G.; Rizzardo, E.; Thang, S. H. Aust. J. Chem. 2005, 58, 379.
(6) Shaver, M. P.; Allan, L. E. N.; Rzepa, H. S.; Gibson, V. C. Angew.
Chem., Int. Ed. 2006, 45, 1241.
The role of OMRP in controlling the polymerization of
styrene in this system is rather less well understood than the
ATRP process. Transition metals, especially cobalt, have
previously shown efficacy in controlling radical polymerizations
10.1021/om700395h CCC: $37.00 © 2007 American Chemical Society
Publication on Web 08/11/2007

4726 Organometallics, Vol. 26, No. 19, 2007
ShaVer et al.
Scheme 2. Decomposition Pathways for R′,R′′[N,N]FeCl2Bn
by an entirely organometallic pathway. This mechanism works
on the same principles as a reversible addition-cleavage
mechanism, with the transition metal acting as a reversible spin-
trap. OMRP offers control over polymerizations based on the
reversible homolytic cleavage of the weak bond between an
alkyl group and a metal catalyst. Cobalt systems supported by
judiciously chosen ligands have been shown to yield high
molecular weight acrylate polymers with very low polydisper-
sities (Mw/Mn ≈ 1.1).9,10 Recently this work has expanded to
include the controlled polymerization of vinyl acetate. While
Co(acac)2 is unable to mediate the polymerization of acrylate
monomers, it can effectively control the polymerization of vinyl
acetate in bulk11 and suspension12 and has led to the develop-
ment of end-functionalized13 and block copolymers14 of poly-
(vinyl acetate) and poly(vinyl alcohol).
A few systems have been reported to display both ATRP and
OMRP behavior. The polymerization of styrene by the half-
sandwich molybdenum compounds15,16 CpMoCl2(PMe3)2 and
CpMoCl2(dppe) proceed via ATRP, generating polystyrene with
moderate polydispersities (PDIs) (ca. 1.5) and molecular weights
that increase linearly with conversion. The same catalysts can
also function in the OMRP of styrene, where the polymer
molecular weights, though higher than those predicted on the
basis of monomer to initiator ratios, increase linearly with
conversion and have PDIs (1.3-1.7) indicative of controlled
behavior. Other examples of catalysts that are capable of
Complexes
favor Fe(III) complexes of high spin-state and generate halogen
end-capped polymers via ATRP. Here, we report the results of
our investigations into the capacity of R′,R′′[N,N]FeCl2 catalysts
to engage in OMRP and a study of the factors that influence
the mechanistic pathways, ATRP versus OMRP, followed in
these reactions.
Results
Formation and Stability of Fe(III)-Alkyl Species. Our
earlier studies have shown that incorporating electron-withdraw-
ing groups at the 2,3 positions of R-diimine ligands promotes a
switch in polymerization mechanism from ATRP to CCT.6,8 This
switch also correlates with a change in the spin-state of the
parent Fe(III) complexes, from S ) 5/2 to S ) 3/2. But what
effect does this spin-state change have on the stability and
function of R′,R′′[N,N]FeCl2(P) intermediates in the polymeri-
zation?
In order to gain further understanding of the role of organo-
metallic intermediates in these polymerization systems, the
reactivity of a family of R′,R′′[N,N]FeCl3 (Scheme 2) complexes
toward Grignard reagents was investigated. The stability of the
product R′,R′′[N,N]FeCl2R complexes can provide an indication
of the relative carbophilicity of the parent R′,R′′[N,N]FeCl2
catalysts. Initially, the mono-benzyl Fe(III) catalysts were
targeted via treatment of R′,R′′[N,N]FeCl3 with BnMgCl (Bn )
C6H5CH2) at -78 °C. There are two possible decomposition
pathways for the alkyl complexes so generated, as shown in
Scheme 2. If the R′,R′′[N,N]FeCl2Bn is thermally unstable, then
it would be anticipated to decompose by homolytic Fe-C bond
cleavage to generate benzyl radicals, which can couple to give
bibenzyl (Bn2, PhCH2CH2Ph). If there is enhanced stability for
the iron alkyl species, it will decompose at a higher temperature.
Then, quenching the reaction with H2O below its decomposition
temperature would be expected to generate toluene via proto-
nolysis of the Fe-C bond.
controlling polymerization through both ATRP and OMRP
17,18
mechanisms include Cp2TiCl2
and OsCl2(PPh3)3.19 These
findings suggest that, under ATRP conditions, the radical
concentration is regulated not only by the atom-transfer equi-
librium but also via the trapping of radicals by the lower
oxidation state metal species. The interplay of these various one-
electron processes has recently been reviewed.20
In an earlier study we detailed the synthesis, characterization,
and styrene polymerization behavior of a family of 2,3-
substituted R-diimine iron catalysts.8 Electron-withdrawing
substituents favor Fe(III) complexes with intermediate spin-
states and generate low molecular weight vinylene-terminated
products characteristic of a CCT mechanism. Electron-donating
substituents, such as alkyl groups or 4-(dimethylamino)phenyl,
(7) While the term halogenophilicity has been previously employed to
describe the reaction of a metal complex with a halide ion (see Macro-
molecules 2004, 9768), the terms halogenophilicity and carbophilicity are
used here to describe the relative predilection for a complex to form metal-
halogen and metal-carbon bonds, respectively.
(8) Allan, L. E. N.; Shaver, M. P.; White, A. J. P.; Gibson, V. C. Inorg.
Chem., in press.
(9) Wayland, B. B.; Poszmik, G.; Mukerjee, S. L.; Fryd, M. J. Am. Chem.
Soc. 1994, 116, 7943.
Addition of BnMgCl at -78 °C to the high-spin complex 1
produced an immediate color change from yellow to deep
purple, and Cy,H[N,N]FeCl2 was isolated from the reaction
mixture. GC analysis of the reaction products indicated the
formation of Bn2 via the reductive alkylation of 1. Similarly,
alkylation of 2, also a high-spin complex, produced Bn2 at -78
°C. In both of these cases, the iron-alkyl species is unstable at
-78 °C; the Fe-C bond cleaves homolytically to generate
carbon radicals in solution, which couple to form Bn2. The
reaction of 3 and 4 with BnMgCl at -78 °C produced very
different results. GC analysis of quenched aliquots showed that
toluene (T) was the major product, not bibenzyl. As the solution
was warmed to -30 °C, a red-brown to purple color change
coincided with Bn2 becoming the major decomposition product
(Figure 1). Thus, it is apparent that radical coupling occurs at
a higher temperature, implying that the Fe-alkyl complexes
derived from 3 and 4 have improved stability. This corresponds
(10) Davis, T. P.; Kukulj, D.; Haddleton, D. M.; Maloney, D. R. Trends
Polym. Sci. 1995, 3, 7943.
(11) Debuigne, A.; Caille, J.-R.; Je´roˆme, R. Angew. Chem., Int. Ed. 2005,
44, 1101.
(12) Debuigne, A.; Caille, J.-R.; Detrembleur, C.; Je´roˆme, R. Angew.
Chem., Int. Ed. 2005, 44, 3439.
(13) Debuigne, A.; Caille, J.-R.; Willet, N.; Je´roˆme, R. Macromolecules
2005, 38, 5452.
(14) Debuigne, A.; Caille, J.-R.; Willet, N.; Je´roˆme, R. Macromolecules
2005, 38, 9488.
(15) Grognec, E. L.; Claverie, J.; Poli, R. J. Am. Chem. Soc. 2001, 123,
9513.
(16) Stoffelbach, F.; Poli, R.; Richard, P. J. Organomet. Chem. 2002,
663, 269.
(17) Grishin, D. F.; Semyonycheva, L. L.; Telegina, E. V.; Smirnov, A.
S.; Nevodchikovc, V. I. Russ. Chem. Bull. Int. Ed. 2003, 52, 505.
(18) Grishin, D. F.; Ignatov, S. K.; Shchepalov, A. A.; Razuvaev, A. G.
Appl. Organomet. Chem. 2004, 18, 271.
(19) Braunecker, W. A.; Itami, Y.; Matyjaszewski, K. Macromolecules
2005, 38, 9402.
(20) Poli, R. Angew. Chem., Int. Ed. 2006, 45, 5058.

Controlled Polymerization of Styrene by Iron
Organometallics, Vol. 26, No. 19, 2007 4727
Scheme 4. Trapping Alkyl Radicals Generated by Fe(III)
Complex Decomposition and ATRP Initiation Reactions with
TEMPO
Figure 1. Ratio of integrated GC peaks for the formation of
bibenzyl (Bn2) and toluene (T) arising from the decomposition of
R′,R′′[N,N]FeCl2Bn at different temperatures. Higher ratios indicate
bibenzyl is the primary decomposition product. Ratios below 1
indicate toluene is the primary decomposition product.
Table 1. OMRP of Styrene by r-Diimine Iron Catalystsa
catalyst
equiv
% conv
Mn
Mn,theo
PDI
8.4
Scheme 3. Decomposition of Cy,R′′[N,N]FeCl2(CHMePh)
none
0
100
77
51
7104
31 245
5
5
5
1.0
2.0
4.0
bimodal (shoulder)
bimodal (shoulder)
Mpk1 ∼50 000
Mpk2 ∼21 000
bimodal (shoulder)
16 117
Complexes
36
5
6
6
6
6
8.0
1.0
2.0
4.0
8.0
24
48
31
23
7.5
14 998
9686
7186
2333
3.7
2.1
1.8
1.4
12 001
8268
2190
a AIBN, styrene, and catalyst equivalents of 0.5:300:X. Polymerizations
conducted in sealed ampules at 120 °C in 1.0 mL of toluene for 48 h.
Conversion of monomer determined by 1H NMR. Molecular weights
calculated by GPC versus styrene standards.
Cy,F-Ph[N,N]FeCl2 and 1-phenylethyl-substituted TEMPO, rather
than 2,3-diphenylbutane, as shown in Scheme 4. The same
product can alternatively be generated by mimicking the ATRP
initiation step and reacting the Fe(II) complex with 1-phenylethyl
chloride in the presence of TEMPO. This initiation reaction
generates phenylethyl radicals with which the persistent TEMPO
radical reacts.
Organometallic-Mediated Radical Polymerizations (OM-
RP). If R′,R′′[N,N]FeCl2(P) species play a significant role in
defining the chemistry of this system, initiation of a polymer-
ization by a radical source such as AIBN (azobisisobutyronitrile)
in the presence of R′,R′′[N,N]FeCl2 would be expected to establish
the OMRP equilibrium independently of the ATRP equilibrium
(see Scheme 1). We thus investigated a series of azo-initiated
polymerizations using Cy,H[N,N]FeCl2, 5, and DiPP,H[N,N]FeCl2,
6, as catalyst precursors; the reactions were conducted at 120
°C in order to mimic the conditions of atom-transfer radical
polymerization and to achieve efficient initiation of the polym-
erization. The results are collected in Table 1.
The polystyrene products generated using the high-spin
system, 5, were bimodal (Figure S2), appearing as shoulders
for 1.0. 2.0, and 8.0 catalyst equivalents, but just resolved for
4.0 equiv. The latter revealed a Mpk for the higher molecular
weight fraction of ca. 50 000 Da and ca. 21 000 Da for the lower
molecular weight component. It is likely that the high molecular
weight fraction is generated by bimolecular coupling of the
radical chains, a termination process that would lead to doubling
of the molecular weight. As the loading of 5 is increased relative
to the radical initiator concentration, the lower molecular weight
fraction increases proportionately, a consequence of more
efficient trapping by the Fe(II) species; the low molecular weight
fraction becomes dominant at 8 equiv of catalyst. The lower
spin catalyst 6 gave monomodal distributions, and a decrease
in polydispersity was observed with increasing catalyst loading,
well with previously reported behavior for the intermediate spin-
state complex DiPP,H[N,N]FeCl3.6
Decomposition of these benzyl complexes does not reproduce
the olefinic end-groups generated by catalytic chain transfer,
however, since the R′,R′′[N,N]FeCl2Bn intermediates lack â-hy-
drogens. To more accurately reproduce the environment of a
growing polystyrene chain, the study was extended to 1-phe-
nylethyl iron complexes. Reaction of 1-phenylethyl magnesium
chloride21 with 1 or 3 in THF at -78 °C produced a slow color
change to purple-brown, indicative of decomposition.
For both 1 and 3, the major product was found to be 2,3-
diphenylbutane, resulting from the coupling of two 1-phenyl-
ethyl radicals (1, 89%; 3, 86%). In addition, styrene (1, 6%; 3,
9%), ethylbenzene (1, 3, 4%), and 1-phenylethyl chloride (1,
3, 1%) were observed (Scheme 3). The formation of styrene
confirms that olefin end-groups can be generated from Fe(III)-
alkyl intermediates. The more carbophilic derivative 3 produces
50% more styrene than 1 during its decomposition. Increasing
the concentration of Fe(II) species by “spiking” the reaction
mixture with varying amounts of Cy,R′′[N,N]FeCl2 during de-
composition had no observable effect on the product distribution.
Trapping Experiments Using TEMPO. The organic radicals
that form during the decomposition of R′,R′′[N,N]FeCl2R can be
potentially trapped by other organic radical species. Decomposi-
tion of Cy,F-Ph[N,N]FeCl2(CHMePh) in the presence of tetram-
ethylpiperidinyl-1-oxy (TEMPO) results in the formation of
(21) Baker, K. V.; Brown, J. M.; Hughes, N.; Skarnulis, A. J.; Sexton,
A. J. Org. Chem. 1991, 56, 4585.

4728 Organometallics, Vol. 26, No. 19, 2007
ShaVer et al.
R′,R′′
Table 2. OMRP of Styrene Catalyzed by r-Diimine Iron
erization studies highlight the potential involvement of
-
[N,N]FeCl2(P) species during atom-transfer radical polymerizations
of styrene. The Fe(III)-alkyl stability studies clearly demonstrate
a correlation between iron-carbon bond strength and the spin-
state of the Fe(III) complexes. High-spin (S ) 5/2) Fe(III)-alkyl
species are unstable at -78 °C, while the substitution of an
electron-withdrawing aryl group at the 2,3 position gives
improved thermal stability to the alkyl species and correlates
with a lowering of the Fe(III) spin-state. These intermediate
spin-state complexes are more efficient alkyl radical traps and
correlate with a switch to catalytic chain transfer as the observed
chain termination mechanism. To date, despite repeated at-
tempts, it has not proven possible to isolate and characterize an
Fe(III)-alkyl species.
Complexes Using V-70 as Initiatora
catalyst
equiv
% conv
Mn
Mn, theo
PDI
none
0
100
54
26
21
38
20
12
8852
11 725
7297
5953
9681
5502
3586
31 245
16 872
8124
6561
11 873
6249
4.2
2.4
1.9
1.5
2.0
1.5
1.3
5
5
5
6
6
6
0.5
1.0
2.0
0.5
1.0
2.0
3749
a
V-70, styrene, and catalyst equivalents of 0.5:300:X. Polymerizations
conducted in sealed ampules at 70 °C in 1.0 mL of toluene for 48 h.
Conversion of monomer determined by 1H NMR. Molecular weights
calculated by GPC versus styrene standards.
with moderate control (PDI ) 1.4, Mn ) 2190, Mn,theo ) 2333)
obtained using 8 equiv of catalyst. At these high temperatures
the Fe(III)-R species is highly unstable and excessive catalyst
loadings are required to obtain even moderate control. The
reaction is extremely slow, with less than 10% conversion in
48h. It is apparent that the R-diimine iron complexes can operate
in OMRP equilibria to control the polymerization, but under
these high-temperature conditions they are not particularly
effective catalysts.
Since the iron-alkyl species should be more stable at lower
temperatures, 2,2′-azobis(4-methoxy-2,4-dimethylvaleronitrile),
V-70, with a half-life of 10 h at 30 °C, was investigated. The
results of screening reactions at 70 °C are shown in Table 2.
Control over the polymerization is now achievable using 2 equiv
of catalyst, with lower spin 6 displaying narrower molecular
weight distributions than for its high-spin relative 5.
Further Studies. A series of further experiments was carried
out to help elucidate the interplay of the ATRP and OMRP
equilibria.
Mixed Catalyst Systems. The effect of mixing an efficient
ATRP catalyst with a catalyst that promotes termination through
CCT should allow control over the rate of polymerization and
molecular weight of the polymer produced. Conducting styrene
polymerizations with catalyst mixtures comprising the ATRP-
directing catalyst, Cy,H[N,N]FeCl2 5, and the CCT-directing
catalyst, Cy,F-Ph[N,N]FeCl2 7, in ratios of 9:1, 3:1, and 1:1
showed that an increased concentration of 7 both decreases the
rate (10% 7, 0.20 h-1; 25% 7, 0.09 h-1; 50% 7, 0.06 h-1; Figure
S7 in the SI) and lowers the molecular weight of the resultant
polymer. As the concentration of CCT catalyst increases,
molecular weight distributions develop a pronounced tailing
toward lower molecular weights (Figure S8 in the SI) and can
be attributed to an increase in the proportion of chain-transfer
events. Consistently, 1H NMR studies showed a 5-fold increase
in the number of olefin-terminated polymer chains as the catalyst
ratio is changed from 9:1 to 1:1.
Studies under OMRP conditions show that catalysts 5-7 can
engage in OMRP equilibria, but that they are not particularly
effective catalysts for the OMRP of styrene due to the low
stability of the Fe(III)-alkyl species under the high-temperature
conditions of the polymerization. Employing AIBN as an
initiating system at 120 °C, 8 equiv of catalyst are required to
afford moderate control. Nevertheless, a spin-state effect is
observed, with ligands favoring lower spin-state Fe(III) centers
acting as superior catalysts. With V-70, the polymerizations can
be performed at lower temperatures and with lower catalyst
loadings.
Organoradical-Trapping/â-Hydrogen Elimination versus
Direct Hydrogen Transfer. An issue for the OMRP process
is the mechanism of chain transfer, which may proceed either
via direct H atom abstraction from the organoradical or via initial
formation of an Fe(III)-alkyl species followed by â-hydrogen
elimination to give the unsaturated polymer product. In the
decomposition studies of 1-phenylethyliron complexes, the lower
spin carbophilic catalyst, Cy,F-Ph[N,N]FeCl3, generates 50% more
styrene than its high-spin relative. If the R′,R′′[N,N]FeCl2R
complex has a longer lifetime due to increased carbophilicity
of the parent catalyst, decomposition through â-H elimination
would be expected to generate more styrene. Conversely, a direct
hydride-transfer mechanism should not depend upon Fe-R
stability, but rather on Fe(II) concentration. Addition of
Cy,F-Ph[N,N]FeCl2, 7, to the reaction mixture prior to complex
decomposition did not affect the amount of styrene formed,
suggesting that the organometallic pathway is favored. Further,
the doubling of the rate of polymerization upon doubling of
catalyst concentration but with no change in the polymer
molecular weight suggests that this system does not operate via
a traditional H-transfer CCT mechanism. In our case, the rate
of reaction would increase due to a shift in the ATRP
equilibrium and an increase in radicals in solution, while the
organometallic complexes formed would also double in con-
centration. Since both monomer consumption and termination
rates are doubled, the observed molecular weight remains the
same. Together, these observations lend support to an organo-
radical-trapping/â-hydrogen elimination pathway.
Catalyst Loading. In classic CCT systems, doubling the
concentration of catalyst maintains a similar rate but decreases
the molecular weight of the short-chain oligomers.22-24 Doubling
the concentration of Cy,F-Ph[N,N]FeCl2, 7, doubled the rate of
the reaction but did not change the molecular weight of the
resultant polymer.
Effect of the OMRP Equilibrium on the Overall Outcome
of the Polymerization. The moderate control over styrene
polymerization under OMRP conditions confirms the likely
involvement of trapped alkyl radicals in ATRP equilibria. In
all cases, however, control of the polymerization by the
carbophilic regime is mediocre at best. In many ways, this is
expected since, regardless of their substituents, Fe(III)-alkyl
complexes decompose at low temperatures (-30 °C) and
therefore will have a short half-life under ATRP reaction
conditions (120 °C). With the OMRP equilibrium shifted toward
the Fe(II) species, the radical concentration remains high and
Discussion
Organometallic Intermediates. The alkylation/decomposi-
tion experiments and organometallic-mediated radical polym-
(22) Enikolopyan, N. S.; Smirnov, B. R.; Ponomarev, G. V.; Bel’govskii,
I. M. J. Polym. Sci. Polym. Chem. Ed. 1981, 19, 879.
(23) Burczyk, A. F.; O’Driscoll, K. F.; Rempel, G. L. J. Polym. Sci.
Polym. Chem. Ed. 1984, 22, 3255.
(24) Gridnev, A. A.; Ittel, S. D.; Fryd, M.; Wayland, B. B. Organome-
tallics 1993, 12, 4871.

Controlled Polymerization of Styrene by Iron
Organometallics, Vol. 26, No. 19, 2007 4729
Table 3. Radical Concentrations for ATRP and OMRP of
Styrene with Various Catalysts
conditions (entries 9, 10, respectively), compared to an 8-fold
increase in [R•] under ATRP/CCT conditions (entries 2 versus
5). Thus, the difference in carbophilicity between these two
catalysts, though discernible, is not all that significant, and
therefore we conclude that carbophilicity per se does not account
for the observed change in mechanism. Rather, the change in
spin-state of the parent Fe(III) complexes must induce a more
drastic change in halogenophilicity, and it is this effect that leads
to the more profound influence on mechanistic pathway.
The organometallic (OMRP) pathway alone is unable to
control the polymerization because most of the control is derived
from the atom-transfer equilibrium. As illustrated in Scheme
1, the radical-trapping equilibrium disfavors the unstable Fe-
(III)-alkyl complex, keeping the radical concentration high
unless excessive catalyst loadings are used. This organometallic
pathway, though not dominant, provides a route to catalytic
chain transfer. Decreased halogenophilicity is provided by
electron-withdrawing groups on the catalyst backbone and is
“switched on” by a change in spin-state of the parent Fe(III)
complexes; under ATRP conditions, the polymerization slows
and causes the termination event to be kinetically competitive,
generating CCT products. The R-diimine iron catalysts ef-
fectively control the polymerization through the ATRP equi-
librium, while the OMRP equilibrium provides a link between
halogen-terminated and olefin-terminated polymer products.
entry
R′
R′′
kapp (s-1
)
[R•]
mechanism
1
2
3
4
5
6
7
8
9
10
Cy
Cy
Cy
Cy
DiPP
Cy
Cy
Cy
Cy
NMe2-Ph 2.00 × 10-4 9.80 × 10-8 ATRP
H
6.83 × 10-5 3.34 × 10-8 ATRP
4.67 × 10-5 2.28 × 10-8 ATRP/CCT
2.83 × 10-5 1.39 × 10-8 ATRP/CCT
1.39 × 10-5 6.79 × 10-9 CCT
1.17 × 10-5 5.71 × 10-9 CCT
9.44 × 10-6 4.62 × 10-9 CCT
6.11 × 10-6 2.99 × 10-9 CCT
1.67 × 10-6 3.47 × 10-9 OMRP
1.53 × 10-6 3.18 × 10-9 OMRP
Me
MeO-Ph
H
Me-Ph
Ph
F-Ph
H
DiPP
H
control is lost without very high levels of catalyst loading and
lower reaction temperatures. But if poor carbophilicity persists
across the substitution patterns studied, why does such a drastic
switch in mechanism occur?
Under ATRP conditions, halogen- and olefin-terminated
polymer products are formed because of differences in the
relatiVe halogeno- versus carbophilicities of the various R-di-
imine iron catalysts. These effects become apparent upon
inspection of the radical concentrations, which are controlled
by both the halogenophilic and carbophilic equilibria operating
under ATRP conditions. The concentration of radicals is related
to the apparent rate constant by eq 1, where kapp is the apparent
rate constant for the polymerization and kp is the propagation
rate constant for styrene at a temperature T. Values for [R•]
with different R-diimines under ATRP and OMRP conditions
are tabulated in Table 3.
Conclusions
The contrasting ATRP versus CCT behavior of R′,R′′[N,N]-
FeCl2 catalysts correlates strongly with the spin-state of the
oxidized Fe(III) species and can be explained by the differing
halogeno- versus carbophilicities of the iron centers in these
Fe(II) complexes. Studying the decomposition of Fe(III) benzyl
and 1-phenylethyl complexes showed increased stability for
systems derived from Fe(III) complexes that are intermediate
spin in character, and these systems favor CCT. Azo-initiated
polymerizations of styrene (under OMRP conditions) are
moderately controlled by these R-diimine iron catalysts. While
the observed mechanism (ATRP versus CCT) strongly correlates
with metal spin-state of the Fe species, the rate of the
polymerization and nature of the polymer produced is dependent
upon the radical concentration generated by the competing
OMRP and ATRP equilibria. Halogenophilic catalysts, i.e., those
possessing electron-donating substituents, afford increased rates
of monomer consumption and increased control via ATRP; the
chain-transfer reaction is not kinetically relevant under these
circumstances. Conversely, catalysts bearing electron-withdraw-
ing groups have decreased halogenophilicity, resulting in slower
rates of monomer consumption and the observation of unsatur-
ated polymer chains arising from chain transfer.
kapp ) kp[R•] where kp ) 107.630 L mol-1 s-1 e(-32.51 kJ/mol/RT)
(1)
Entries 1-8 draw upon kinetic runs described in previous
papers in this series.8,25,26 A correlation between radical
concentration and polymerization mechanism is readily apparent.
Increased electron donation from the R-diimine ligand leads to
greater concentrations of radicals and, thus, markedly faster
polymerization rates. In entries 1 and 2, the consumption of
monomer is fast relative to chain-transfer reactions, and classic
ATRP behavior is observed. A decrease in [R•] and kapp slows
monomer consumption; in these cases, catalytic chain-transfer
events become kinetically competitive, and evidence for both
mechanisms is observed. When the polymerization rate is slowed
to a sufficient extent, catalytic chain-transfer events dominate
and short-chain, olefin-terminated polymers are observed. The
change from ATRP to CCT behavior is not abrupt, but rather a
consequence of a gradual decrease in kapp relative to the chain-
transfer reaction. This gradual switch in polymerization mech-
anism is reproduced by mixed catalyst systems consisting of
Cy,H[N,N]FeCl2, 5, and Cy,F-Ph[N,N]FeCl2, 7. Increasing the
proportion of 7, a catalyst with decreased halgenophilicity and
increased carbophilicity, leads to lower rates and lower radical
concentrations. As chain-transfer termination reactions increase,
the molecular weights decrease and the distributions show
pronounced tailing toward low molecular weights.
Experimental Section
General Experimental Procedures. All manipulations of air-
and/or moisture-sensitive compounds were carried out under
nitrogen using standard Schlenk and cannula techniques or in a
conventional nitrogen-filled glovebox. NMR spectra were recorded
on a Bruker AV-400 (1H, 400.3 MHz; 13C, 100.7 MHz) spectrom-
eter, at 293 K unless stated otherwise. Chemical shifts are reported
as δ (in ppm) and referenced to the residual proton signal and to
the 13C signal of the deuterated solvent, respectively; coupling
constants are reported in Hz. The following abbreviations have been
used for multiplicities: s (singlet), d (doublet), t (triplet), m
(unresolved multiplet), br (broad). GC analyses were conducted
on an Agilent 6890A gas chromatograph with a HP-5 column. GC/
CI (chemical ionization) or EI (electron impact) mass spectra were
Comparing the radical concentrations in Table 3 for catalysts
operating in the ATRP and OMRP regimes reveals that OMRP
radical concentrations are not significantly different from those
observed under ATRP conditions. Relatively minor differences
exist between radical concentrations for cyclohexylimino cata-
lysts and 2,6-diisopropylphenylimino catalysts under OMRP
(25) Gibson, V. C.; O’Reilly, R. K.; Reed, W.; Wass, D. F.; White, A.
J. P.; Williams, D. J. Chem. Commun. 2002, 1850.
(26) O’Reilly, R. K.; Shaver, M. P.; Gibson, V. C.; White, A. J. P.
Submitted to Macromolecules.

4730 Organometallics, Vol. 26, No. 19, 2007
ShaVer et al.
1
recorded on either a VG Autospec or a VG Platform II spectrometer
at Imperial College London. GPC data were collected using Cirrus
GPC/SEC software, version 1.11, connected to a Shodex RI-101
detector, and were referenced to polystyrene standards (Polymer-
Labs EasiCal, PS1). Solvents were dried by refluxing over an
appropriate drying agent and distilling, or passing through a cylinder
filled with commercially available Q-5 catalyst (13% Cu(I) oxide
on Al2O3) and activated Al2O3 (3 mm, pellets) in a stream of
nitrogen, and were degassed before use. NMR solvents were dried
over molecular sieves and degassed prior to use. Styrene was stirred
over calcium hydride for 24 h, vacuum-transferred, degassed, and
then stored in an inert atmosphere at -35 °C. R′,R′′[N,N]FeCl2
catalysts 5-7 were prepared according to literature procedures.6,25,27
R′,R′′[N,N]FeCl3 precursors 1-4 were prepared according to literature
procedures,6,8 stored in the absence of light, and used within 1 week
of synthesis. C6H5CHMeMgCl was synthesized according to a
literature procedure.21 AIBN was purchased from Aldrich Chemical
Co., stored at -35 °C, and used as received. V-70 was purchased
from Wako Chemicals Company, stored at -35 °C, and used as
received. All other reagents were purchased from ACROS Organics
or Aldrich Chemical Co. and used as received.
General Procedure for Alkyl Decomposition Experiments.
R′,R′′[N,N]FeCl3 (0.500 mmol) was dissolved in THF (10 mL) and
cooled to -78 °C. Dropwise addition of a solution of RMgCl (0.500
mmol) in THF resulted in the precipitation of a white solid (MgCl2).
An aliquot (1 mL) was removed from the reaction mixture and
quenched with distilled H2O. Subsequent aliquots were removed
and quenched over a range of temperatures. Depending on the
stability of the corresponding alkyl complex, vivid color changes
corresponding to the formation of R′,R′′[N,N]FeCl2 were observed
at various temperatures on warming. Aliquots were filtered through
basic alumina and analyzed by GC or GC/MS. GC area counts
were corrected against organic components present in unreacted
RMgCl.
as a white solid in 61% yield. H NMR (CDCl3): δ 1.02, 1.16,
1.29, 1.31 (4 × 3H, s, CH3), 1.46-1.56 (6H, m, CH2), 1.77 (3H,
3
3
d, JH-H 6.4, CH3-CH), 4.77 (1H, q, JH-H 6.4, CH-CH3), 7.22-
7.31 (5 H, m, Ar-H). 13C NMR (CDCl3): δ 15.9, 18.4, 23.2, 24.1,
25.6, 25.8, 34.1, 48.9, 78.1, 140.7, 126.5, 128.4, 129. MS (EI+,
m/z): 261 [M+].
Reaction of Cy,F-Ph[N,N]FeCl2 with 1-PECl and TEMPO. To
an ampule equipped with a magnetic stirrer bar were added toluene
(3 mL), Cy,H[N,N]FeCl2 (0.100 mmol), 1-phenylethyl chloride (0.100
mmol), and TEMPO (0.100 mmol). The ampule was then sealed
and heated in a sand bath at 120 °C for 1 h. The ampule was then
removed from heat and the contents were dissolved in THF, filtered
through alumina, dried in Vacuo, and recrystallized from Et2O to
afford 1-(1-phenylethoxy)-2,2,6,6-tetramethylpiperidine in 56%
yield. 1H NMR (CDCl3): δ 1.02, 1.16, 1.29, 1.31 (4 × 3H, s, CH3),
3
1.46-1.56 (6H, m, CH2), 1.77 (3H, d, JH-H 6.4, CH3-CH), 4.77
(1H, q, 3JH-H 6.4, CH-CH3), 7.22-7.31 (5 H, m, Ar-H). 13C NMR
(CDCl3): δ 15.9, 18.4, 23.2, 24.1, 25.6, 25.8, 34.1, 48.9, 78.1, 140.7,
126.5, 128.4, 129. MS (EI+, m/z): 261 [M+].
General Procedure for Polymerization Screening. All poly-
merizations were set up and performed under an atmosphere of
oxygen-free, dry nitrogen using standard Schlenk-line and glovebox
techniques. To an ampule equipped with a magnetic stirrer bar was
added a solution of monomer, solvent (where applicable) initiator,
and catalyst and then sealed. In most cases the catalyst was soluble
in the monomer solution. The ampules were heated in an oil or
sand bath at 70 or 120 °C for an allotted period of time. Aliquots
(0.3 mL) were removed at appropriate intervals. Conversion was
determined by integration of the monomer versus polymer backbone
resonances in the 1H NMR spectrum of the crude product in CDCl3.
The polystyrene was dissolved in THF and precipitated into
acidified methanol, filtered, and washed with methanol. The
precipitate was collected and dried for 24 h under vacuum. Samples
were then analyzed by GPC.
Reaction of Cy,F-Ph[N,N]FeCl2(CHMePh) with TEMPO.
Cy,F-Ph[N,N]FeCl3 (0.500 mmol) was dissolved in THF (10 mL)
and cooled to -78 °C. A solution of TEMPO (0.500 mmol) in
THF (5 mL) was added to this mixture dropwise via cannula and
an aliquot removed from the reaction. Addition of 1-PhEtMgCl
(0.500 mmol) in THF resulted in the precipitation of a white solid
(MgCl2). The reaction mixture was warmed to room temperature
and quenched with degassed H2O. The organic fraction was
separated, filtered through alumina, dried in Vacuo, and recrystal-
lized from Et2O to yield 1-(benzyloxy)-2,2,6,6-tetramethylpiperidine
Acknowledgment. The authors thank the Natural Sciences
and Engineering Research Council of Canada for a postdoctoral
fellowship (to M.P.S.) and the Engineering and Physical
Sciences Research Council (UK) for a studentship (to L.E.N.A.).
Supporting Information Available: Tables of GC product
distributions for alkylation experiments and tables and plots for
the organometallic-mediated radical polymerization of styrene. This
material is available free of charge via the Internet at http://
pubs.acs.org.
(27) Gibson, V. C.; O’Reilly, R. K.; Wass, D. F.; White, A. J. P.;
Williams, D. J. Macromolecules 2003, 36, 2591.
OM700395H