Correlation of metal spin-state in α-diimine iron catalysts with polymerization mechanism

Article abstract of DOI:10.1021/ic701500y

The α-diimine iron complexes, R′,R″[N,M]FeCl₂ (R′R″[N,M] = R′-N=CR″-CR″=N-R′, where R′ = tert-butyl (tBu), cyclohexyl (Cy) and R′ = phenyl (Ph), para-fluorophenyl (F-Ph), para-bromophenyl (Br-Ph), para-methylphenyl (Me-Ph), or para-methoxyphenyl (MeO-Ph)), are found to polymerize styrene through a catalytic chain transfer (CCT) mechanism. Magnetic moment measurements indicate that Fe(III) complexes containing these ligands possess intermediate (S = 3/2) spin-state iron centers. In contrast, Fe(III) complexes bearing proton (R″ = H) and para-dimethylaminophenyl (R″ = NMe₂-Ph) substituents are high-spin and are efficient atom transfer radical polymerization (ATRP) catalysts. Hammett plots show a linear correlation of the substituent constant, σ, with polymerization rate and polymer molecular weight, respectively.

Full text of DOI:10.1021/ic701500y

Inorg. Chem. 2007, 46, 8963−8970
Correlation of Metal Spin-State in
Polymerization Mechanism
r-Diimine Iron Catalysts with
Laura E. N. Allan, Michael P. Shaver, Andrew J. P. White, and Vernon C. Gibson*
Contribution from the Department of Chemistry, Imperial College London,
South Kensington, London, SW7 2AZ, UK
Received July 27, 2007
R′
,R
′,R
The
cyclohexyl (Cy) and R′′ ) phenyl (Ph), para-fluorophenyl (F
(Me Ph), or para-methoxyphenyl (MeO Ph)), are found to polymerize styrene through a catalytic chain transfer
(CCT) mechanism. Magnetic moment measurements indicate that Fe(III) complexes containing these ligands possess
intermediate (S
3/2) spin-state iron centers. In contrast, Fe(III) complexes bearing proton (R′′ ) H) and para-
dimethylaminophenyl (R′′ ) NMe2 Ph) substituents are high-spin and are efficient atom transfer radical polymerization
R
-diimine iron complexes,
′′[N,N]FeCl2 (R ′′[N,N]
)
R
−
′−
N
d
CR′′−CR′′d
N
−
R
′
, where R′ ) tert-butyl (tBu),
Ph), para-bromophenyl (Br Ph), para-methylphenyl
−
−
−
)
−
(ATRP) catalysts. Hammett plots show a linear correlation of the substituent constant,
and polymer molecular weight, respectively.
σ, with polymerization rate
Introduction
a record of the individual catalytic events, with molecular
weights and polymer end-groups providing a signature of
the prevalent mechanism.
Despite the widely recognized and important influence of
metal spin-state on the metal-centered reactivity of enzymes,1-5
the effect of metal spin-state on metal-based reactivity in
nonbiological systems remains largely undocumented. Re-
cently, we identified and reported a correlation between metal
spin-state and the nature of the resultant polymer in a metal-
catalyzed radical polymerization system. We concluded that
differing mechanistic pathways had been followed, depending
upon the spin-state adopted by the catalyst system, the latter
being determined by the ligand substituents.6
Polymerizations mediated by R-diimine Fe(II) complexes
bearing alkylimine substituents, in which the parent Fe(III)
complex is in a high spin (S ) 5/2) state, afforded controlled
polymerizations of styrene via an atom transfer radical
polymerization (ATRP) mechanism, whereas arylimine-based
systems, with the corresponding Fe(III) species in an
ATRP is a facile, pseudo-living radical polymerization,
which uses metal-mediated halogen exchange to control the
equilibrium between active and dormant polymer chains.10,11
By ensuring that the majority of species lie on the dormant
side of this dynamic equilibrium, the radical concentration
remains low. This minimizes bimolecular termination reac-
tions and generates halogen-terminated polymer chains with
molecular weights very close to theoretical values, increasing
linearly with conversion and displaying low polydispersities.
Whereas copper12 and ruthenium13 are the most extensively
studied metals in ATRP catalysis, a wide range of other
metals have also been shown to be active, including
nickel,14-20 palladium,21,22 rhodium,23,24 molybdenum,25-32
3
(7) Gibson, V. C.; O’Reilly, R. K.; Reed, W.; Wass, D.; White, A. J. P.;
Williams, D. J. Chem. Commun. 2002, 1850.
(8) Gibson, V. C.; O’Reilly, R. K.; Wass, D. F.; White, A. J. P.; Williams,
D. J. Macromolecules 2003, 36, 2591.
intermediate (S ) /2) spin-state, followed a catalytic chain
transfer (CCT) mechanism.7-9 The isolated polymer provides
* To whom correspondence should be addressed. E-mail: v.gibson@
imperial.ac.uk.
(1) Williams, R. J. P. Fed. Proc. Fed. Am. Soc. Exp. Biol. 1961, 20, 5.
(2) Hoard, J. L.; Hoard, M. J.; Hamor, T. A.; Caughey, W. S. J. Am.
Chem. Soc. 1965, 87, 2312.
(3) Scheidt, W. R.; Reed, C. A. Chem. ReV. 1981, 81, 543.
(4) Perutz, M. F.; Wilkinson, A. J.; Paoli, M.; Dodson, G. G. Ann. ReV.
Biophys. Biomol. Struct. 1998, 27, 1 and references therein.
(5) Meunier, B.; deVisser, S. P.; Shaik, S. Chem. ReV. 2004, 104, 3947.
(6) Shaver, M. P.; Allan, L. E. N.; Rzepa, H. S.; Gibson, V. C. Angew.
Chem., Int. Ed. 2006, 45, 1241.
(9) O’Reilly, R. K. Synthesis of Novel Iron Complexes and their use in
Atom Transfer Radical Polymerisation. Ph.D. Thesis, University of
London, 2003.
(10) Matyjaszewski, K.; Xia, J. Chem. ReV. 2001, 101, 2921.
(11) Kamigaito, M.; Ando, T.; Sawamoto, M. Chem. ReV. 2001, 101, 3689.
(12) Pintauer, T.; Matyjaszewski, K. Coord. Chem. ReV. 2005, 249, 1155.
(13) Kamigaito, M.; Ando, T.; Sawamoto, M. The Chemical Record 2004,
4, 159.
(14) Granel, C.; Dubois, P.; Je´rome, R.; Teyssie´, P. Macromolecules 1996,
29, 8576.
(15) Li, P.; Qiu, K.-Y. Polymer 2002, 43, 5873.
10.1021/ic701500y CCC: $37.00
Published on Web 09/14/2007
© 2007 American Chemical Society
Inorganic Chemistry, Vol. 46, No. 21, 2007 8963

Allan et al.
iron,7,33-39 and rhenium.40,41 More recently, systems based
on titanium,42,43 cobalt,44 and osmium45 have been added.
CCT is an efficient method for synthesizing low-molec-
ular-weight polymers via a free-radical mechanism.46,47 It is
generally thought that the mechanism proceeds via a two-
step process, involving hydrogen abstraction by the metal
center and then subsequent reinitiation through monomer
insertion into the metal-hydrogen bond.48-50 It is also
possible for the metal hydride species to be formed via
trapping of the alkyl radical to give initially an organometallic
Scheme 1
(16) Moineau, G.; Minet, M.; Dubois, P.; Teyssie´, P.; Senninger, T.; Je´roˆme,
R. Macromolecules 1999, 32, 27.
(17) Uegaki, H.; Kotani, Y.; Kamigaito, M.; Sawamoto, M. Macromolecules
1997, 30, 2249.
(18) Uegaki, H.; Kotani, Y.; Kamigaito, M.; Sawamoto, M. Macromolecules
1998, 31, 6756.
(19) Uegaki, H.; Kamigaito, M.; Sawamoto, M. J. Polym. Sci., Part A:
Polym. Chem. 1999, 37, 3003.
(20) O’Reilly, R. K.; Shaver, M. P.; Gibson, V. C. Inorg. Chim. Acta 2006,
359, 4417.
(21) Lecomte, P.; Drapier, I.; Dubois, P.; Tessiye´, P.; Je´rome, R. Macro-
molecules 1997, 30, 7631.
(22) Hong, S. C.; Jia, S.; Teodorescu, M.; Kowalewski, T.; Matyjaszewski,
K.; Gottfried, A.; Brookhart, M. J. Polym. Sci., Part A: Polym. Chem.
2002, 40, 2736.
(23) Percec, V.; Barboiu, B.; Neumann, A.; Ronda, J. C.; Zhao, M.
Macromolecules 1996, 29, 3665.
(24) Moineau, G.; Granel, C.; Dubois, P.; Teyssie´, P.; Je´roˆme, R.
Macromolecules 1998, 31, 542.
(25) Grognec, E. L.; Claverie, J.; Poli, R. J. Am. Chem. Soc. 2001, 123,
9513.
(26) Stoffelbach, F.; Poli, R.; Richard, P. J. Organomet. Chem. 2002, 663,
269.
(27) Stoffelbach, F.; Claverie, J.; Poli, R. C.R. Acad. Sci., Ser. IIc: Chim.
2002, 5, 37.
(28) Stoffelbach, F.; Haddleton, D. M.; Poli, R. Eur. Polym. J. 2003, 39,
2099.
(29) Poli, R.; Stoffelbach, F.; Maria, S.; Mata, J. Chem.sEur. J. 2005, 11,
2537.
(30) Maria, S.; Stoffelbach, F.; Mata, J.; Daran, J.-C.; Richard, P.; Poli, R.
J. Am. Chem. Soc. 2005, 127, 2537.
(31) Mata, J.; Maria, S.; Daran, J.-C.; Poli, R. Eur. J. Inorg. Chem. 2006,
13, 2624.
(32) Stoffelbach, F.; Richard, P.; Poli, R.; Jenny, T.; Savary, C. Inorg. Chim.
Acta 2006, 359, 4447.
(33) Ando, T.; Kamigaito, M.; Sawamoto, M. Macromolecules 1997, 30,
4507.
(34) Fuji, Y.; Ando, T.; Kamigaito, M.; Sawamoto, M. Macromolecules
2002, 35, 2949.
(35) Gibson, V. C.; O’Reilly, R. K.; White, A. J. P.; Williams, D. J. J.
Am. Chem. Soc. 2003, 125, 8450.
(36) Ibrahim, K.; Lofgren, B.; Seppala, J. Eur. Polym. J. 2003, 39, 939.
(37) Louie, J.; Grubbs, R. H. Chem. Commun. 2000, 16, 1479.
(38) Kotani, Y.; Kamigaito, M.; Sawamoto, M. Macromolecules 1999, 32,
6877.
(39) Matyjaszewski, K.; Wei, M.; Xia, J.; McDermott, N. E. Macromol-
ecules 1997, 30, 8161.
(40) Kotani, Y.; Kamigaito, M.; Sawamoto, M. Macromolecules 1999, 32,
2420.
(41) Kotani, Y.; Kamigaito, M.; Sawamoto, M. Macromolecules 2000, 33,
6746.
intermediate, which then undergoes a â-H elimination
process to release the olefin.51,52 The first part of this latter
process forms the basis of the organometallic-mediated
radical polymerization (OMRP) reaction, which, when
combined with the â-H elimination step, results overall in
the CCT pathway (Scheme 1). A separate study in our
laboratory has focused on the organometallic intermediates
in the R-diimine iron system and their use in OMRP.53
ATRP and CCT polymerization mechanisms are not
mutually exclusive; they are intimately linked by the
equilibria illustrated in Scheme 1. Various factors can play
a role in determining whether a polymerization catalyzed
by a particular complex will occur via the predominantly
halogenophilic ATRP route, whether the OMRP equilibrium
will dominate, or whether hydrogen transfer will occur to
give CCT products. Such competing mechanisms have been
observed by Poli and co-workers with the half-sandwich
molybdenum compounds, CpMoCl2(PMe3)2 and CpMoCl2-
(dppe), which are active catalysts for the polymerization of
styrene under both ATRP and OMRP conditions.25,26
CpMoCl2(C4H6) and CpMo(SiMe3)2(C4H6), however, afford
short-chain, olefin-terminated oligomers via an overall CCT
process. A recent minireview discusses the interplay between
these one-electron processes.54
The correlation of the metal spin-state with the mechanistic
outcome in catalytic polymerization using R-diimine Fe(II)
complexes remains the first reported example of its kind.
Initially, we found that the spin-state is strongly influenced
by the imine N-donor substituents, with alkylimine donors
affording high-spin Fe(III) species, whereas arylimine de-
rivatives gave intermediate-spin Fe(III) species.7,8 Subse-
quently, we found that aryl substituents attached to the carbon
atoms of the imine backbone can also affect the spin-state
of the iron centers, as a result of extensive delocalization
throughout the ligand backbone.6 Here, we present the results
of our studies on Fe(II) and Fe(III) complexes that bear
R-diimine ligands with differing para-substituted aryl groups
attached to the backbone carbon atoms, along with a study
of their behavior in catalytic polymerization.
(42) Kabachii, Y. A.; Kochev, S. Y.; Bronsteinl, L. M.; Blagodatskikh, I.
B.; Valetsky, P. M. Polym. Bull. 2003, 50, 271.
(43) Grishin, D. F.; Semyonycheva, L. L.; Telegina, E. V.; Smirnov, A.
S.; Nevodchikovc, V. I. Russ. Chem. Bull., Int. Ed. 2003, 52, 505.
(44) Weiser, M. S.; Lhaupt, R. M. J. Polym. Sci., Part A: Polym. Chem.
2005, 43, 3804.
(45) Braunecker, W. A.; Itami, Y.; Matyjaszewski, K. Macromolecules
2005, 38, 9402.
(46) Gridnev, A. J. Polym. Sci., Polym. Chem. Ed. 2000, 38, 1753.
(47) Gridnev, A.; Ittel, D. S. Chem. ReV. 2001, 101, 3611.
(48) Enikolopyan, N. S.; Smirnov, B. R.; Ponomarev, G. V.; Bel’govskii,
I. M. J. Polym. Sci., Polym. Chem. Ed. 1981, 19, 879.
(49) Burczyk, A. F.; O’Driscoll, K. F.; Rempel, G. L. J. Polym. Sci., Polym.
Chem. Ed. 1984, 22, 3255.
(51) Gridnev, A. A.; Ittel, S. D.; Wayland, B. B.; Fryd, M. Organometallics
1996, 15, 5116.
(52) Heuts, J. P. A.; Forster, D. J.; Davis, T. P. Macromolecules 1999, 32,
2511.
(53) Shaver, M. P.; Allan, L. E. N.; Gibson, V. C. Organometallics 2007,
26, 4725.
(54) Poli, R. Angew. Chem., Int. Ed. 2006, 45, 5058.
(50) Gridnev, A. A.; Ittel, S. D.; Fryd, M.; Wayland, B. B. Organometallics
1993, 12, 4871.
8964 Inorganic Chemistry, Vol. 46, No. 21, 2007

Metal Spin-State with Polymerization Mechanism
Scheme 2
Scheme 3
Figure 1. The molecular structure of the CS-symmetric 13. Selected bond
lengths (Å): Fe-Cl(1), 2.2551(16); Fe-Cl(2), 2.2183(15); Fe-N(1), 2.077-
(3); Fe-N(1A), 2.077(3); N(1)-C(1), 1.284(4); C(1)-C(1A), 1.521(6).
Selected bond angles (°): Cl(1)-Fe-Cl(2), 125.62(7); Cl(1)-Fe-N(1),
101.79(9); Cl(1)-Fe-N(1A), 101.79(9); Cl(2)-Fe-N(1), 119.24(9); Cl-
(2)-Fe-N(1A), 119.24(9); N(1)-Fe-N(1A), 119.24(9).
Results
Synthesis of Fe(II) Compounds. The R-diimine ligands,
R′,R′′[N,N], (where R′,R′′[N,N] ) R′-NdCR′′-CR′′dN-R′,
R′ ) tert-butyl (tBu), cyclohexyl (Cy), and R′′ ) phenyl
(Ph), para-fluorophenyl (F-Ph), para-bromophenyl (Br-
Ph), para-methylphenyl (Me-Ph), para-methoxyphenyl
(MeO-Ph), para-dimethylaminophenyl (NMe2-Ph)) were
prepared via the modification of a literature procedure,55
using TiCl4 as an activating agent (Scheme 2). After aqueous
workup, the crude residues were recrystallized from hot
alcohol to give the desired products in good isolated yields
(65-85%).
Figure 2. Plot of ln(M0/Mt) versus time for the bulk polymerization of
styrene (200 equiv, 1-PECl, 120 °C) by Cy,R′′[N,N]FeCl2 complexes (R′′ )
p-F-Ph, 9; R′′) Ph, b; R′′ ) p-Br-Ph, (). Cy,H[N,N]FeCl2 rate (purple
line) included for reference purposes.7
For the Fe(II) chloride complexes, 13-24, the addition
of CH2Cl2 to an intimate mixture of the pro-ligand and FeCl2-
(THF)1.5 produced deep blue solutions, except in the case of
18 and 24, which were red-brown (Scheme 3). After stirring
overnight, the solutions were filtered and concentrated before
the desired product was obtained as a purple or red-brown
solid, following precipitation with pentane. Magnetic moment
measurements (Evans’ NMR method) on 13-24 revealed
high-spin Fe(II) centers with µeff in the range 4.93-5.28 µB.
Despite the paramagnetism of the complexes, contact-shifted
1H NMR spectra could be obtained for each compound and
assigned on the basis of integrated signal intensities and
chemical shifts. The imine νCdN stretch was visible between
1600 and 1611 cm-1, and FAB-MS showed the expected
[M+H]+ peaks as well as common fragments, such as
[M-Cl]+. In addition, crystals suitable for X-ray structure
analysis of 13, 14, 18-20, and 24 were obtained. The
complexes were found to possess closely related monomeric
solid-state structures with four-coordinate metal centers.
As an example, crystals suitable for X-ray analysis were
obtained through slow cooling of a saturated solution of 13
Figure 3. Plot of molecular weight versus conversion for the bulk
polymerization of styrene (200 equiv, 1-PECl, 120 °C) by Cy,R′′[N,N]FeCl2
complexes (R′′ ) p-F-Ph, 9; R′′ ) Ph, b; R′′ ) p-Br-Ph, ().
in toluene. The X-ray crystal structure of 13 revealed a CS-
symmetric complex with a highly distorted tetrahedral iron
center, where the cis angles are in the range of 78.43(15)-
125.62(7)°, with the most acute being the bite angle of the
chelating N,N′ ligand (Figure 1). The crystallographic mirror
plane includes the iron center and the two chlorines and
bisects the C(1)-C(1A) single bond [1.521(6) Å] between
(55) Kimpe, N. D.; D’Hondt, L.; Stanoeva, E. Tetrahedron Lett. 1991, 32,
3879.
Inorganic Chemistry, Vol. 46, No. 21, 2007 8965

Allan et al.
Figure 4. Plot of ln(M0/Mt) versus time for the bulk polymerization of
styrene (200 equiv, 1-PECl, 120 °C) by Cy,R′′[N,N]FeCl2 complexes (R′′ )
p-Me-Ph, 9; R′′ ) p-MeO-Ph, b; R′′ ) p-NMe2-Ph, (). Cy,H[N,N]FeCl2
rate (purple line) included for reference.7
Figure 6. Molecular structure of one (I) of the two crystallographically
independent CS-symmetric complexes present in the crystals of 31.
the top [Cl(1) side] of the phenyl rings is rotated away from
the mirror plane. The closely related structures, 14, 18-20,
and 24 can be found in the Supporting Information.
Polymerization Studies. 13-24 were tested for the ATRP
of styrene (200 equiv, bulk), under an inert atmosphere at
120 °C, using 1-phenylethyl chloride (1-PECl) as the initiator.
As seen previously with the fluorophenyl derivatives,6
complexes bearing electron withdrawing groups in the para
position of the phenyl ring (13, 14, 19, and 20) resulted in
polymerizations that proceeded slowly. Figure 2 shows the
rate data for 13-15, along with Cy,H[N,N]FeCl2 for compari-
son. They required ca. 70 h to reach 80% conversion, with
pseudo-first-order rate constants of 0.01-0.03 h-1. This is
significantly slower than the aldimine derivatives, which
reached high conversions within 24 h, with rate constants in
the range of 0.25-0.27 h-1.7
Figure 5. Plot of molecular weight versus conversion for the bulk
polymerization of styrene (200 eq., 1-PECl, 120 °C) by Cy,R′′[N,N]FeCl2
complexes (R′′ ) p-Me-Ph, 9; R′′ ) p-MeO-Ph, b; R′′ ) p-NMe2-Ph,
().
Scheme 4
The molecular weight data on polystyrene samples gener-
ated using these catalysts revealed that the predominant
mechanism of polymerization was not ATRP. Low-molec-
ular-weight polymers were isolated, with molecular weights
independent of conversion (Figure 3) and consistent with
the operation of a CCT mechanism. End-group analyses by
1H NMR spectroscopy revealed the presence of olefinic end-
groups at 6.1-6.4 ppm, with no evidence for the chlorine-
terminated chains expected from an ATRP mechanism
(∼4.6).
Substituting the para position of the phenyl ring with a
more electron-donating group gave complexes that showed
a trend back toward typical ATRP behavior. In the case of
p-Me-Ph substituted complexes, 16 and 22, rates are still
relatively slow (0.04-0.06 h-1), and the molecular weight
data do not increase linearly with conversion. End-group
analyses confirmed the presence of vinylene end-groups,
consistent with CCT. Replacing the methyl group by a
methoxy substituent increased the rate to 0.10 for 17 and
0.07 h-1 for 23, and molecular weights showed an initial
linear increase with conversion. A limit is reached, at around
8000 Da, and the polymer showed only olefin end-groups,
with no signals in the 1H NMR spectrum that were
characteristic of a halogen end-group. The most electron-
the two imine units [C(1)-N(1) 1.284(4) Å]. The two imines
are perfectly coplanar (a consequence of the mirror sym-
metry), and the iron perches above this plane by ca. 0.59 Å
in the direction of Cl(1). The FeCl2 unit adopts a conforma-
tion that puts Cl(1) in a pseudo-apical position and Cl(2) in
a pseudo-basal position, such that Cl(1) is ca. 2.83 Å above
the diimine plane, whereas Cl(2) is only ca. 0.50 Å below
it. Associated with this is an elongation of the Fe-Cl(1) bond
length [2.2551(16)] cf. that to Cl(2) [2.2183(15) Å]; the angle
between the two chlorine atoms is substantially enlarged from
ideal tetrahedral, being 125.62(7)°. The aryl rings on the
backbone are slightly twisted away from orthogonal; the
torsion angle about the C(1)-C(Ph) bond is ca. 102°, and
8966 Inorganic Chemistry, Vol. 46, No. 21, 2007

Metal Spin-State with Polymerization Mechanism
Table 1. Selected Bond Lengths (Angstroms) and Angles (Degrees) for the Two Crystallographically Independent CS-Symmetric Complexes (I and II)
Present in the Crystals of 31
Mol I
Mol II
Mol I
Mol II
Fe-Cl(1)
2.285(3)
2.145(6)
2.294(3)
2.141(6)
2.2109(19)
1.456(12)
94.68(7)
Fe-Cl(2)
2.2053(19)
2.243(7)
1.272(10)
1.247(9)
100.65(19)
94.13(6)
86.82(10)
77.2(2)
2.2109(19)
2.275(7)
1.276(9)
1.265(9)
98.7(2)
94.68(7)
87.03(11)
77.7(2)
Fe-N(1)
Fe-N(2)
Fe-Cl(2A)
C(1)-C(2)
Cl(1)-Fe-Cl(2)
Cl(1)-Fe-N(2)
Cl(2)-Fe-N(1)
Cl(2)-Fe-Cl(2A)
N(1)-Fe-Cl(2A)
2.2053(19)
1.455(12)
94.13(6)
177.83(19)
114.76(7)
127.08(14)
114.76(7)
N(1)-C(1)
N(2)-C(2)
Cl(1)-Fe-N(1)
Cl(1)-Fe-Cl(2A)
Cl(2)-Fe-N(2)
N(1)-Fe-N(2)
N(2)-Fe-Cl(2A)
176.4(2)
117.06(7)
122.69(13)
117.06(7)
86.82(10)
87.03(11)
donating substituent, p-NMe2-Ph, results in catalysts with
extremely fast polymerization rates (0.72 for 18 and 0.36
h-1 for 24), and the molecular weight now increases linearly
with conversion. End-group analyses on precipitated poly-
styrene samples revealed that the polymer is halogen
terminated, with no olefin peaks detectable. Rate data and
plots of Mn versus conversion for 16-18 are shown in
Figures 4 and 5, respectively.
The rate of polymerization for 18, at 0.72 h-1, is the
highest found to date in the R-diimine system, reaching 85%
conversion within 3 h, some three times faster than the
aldimine R-diimine catalyst, Cy,H[N,N]FeCl2, (0.25 h-1).
Molecular weights for 18 do not match-up perfectly with
theoretical values, being consistently ca. 10-20% lower than
expected. This is attributed to a small amount of termination
through CCT, giving olefin-terminated oligomers, which are
not reincorporated into the polymer. The PDIs range from
1.2 to 1.3 until around 75% conversion, where they broaden
to 1.4, and it is notable that the GPC traces show an
asymmetry, tailing toward low molecular weight and sup-
porting the occurrence of a small amount of chain transfer
(Figure S44 in the Supporting Information). GC-MS and 1H
NMR spectroscopy on the hydrocarbon-soluble fraction of
the crude polymer confirmed the presence of small amounts
of olefin-terminated products.
revealed µeff values for 25-29 and 32-36 in the range of
3.9-4.2 µB, close to the intermediate (S ) 3/2) spin value of
3.87,56 whereas 30, 31, and 37 have values in the range of
5.6-5.8 µB, consistent with high-spin Fe(III) centers. More
detailed studies on 25-31 are in progress using SQUID,
EPR, and Mo¨ssbauer spectroscopy and will be reported in
due course.
X-ray structures of LnFeCl3 species are surprisingly rare,
and there are only a few examples of crystallographically
characterized five-coordinate iron trichloride complexes
incorporating nitrogen donors.57,58 Crystals of 31 suitable for
X-ray analysis were obtained through slow evaporation of a
solution of tBu,H[N,N]FeCl3 in CDCl3/cyclohexane (95:5 v/v)
and comprised two crystallographically independent Cs
symmetric molecules; molecule I is shown in Figure 6 and
molecule II in Figure S26 in the Supporting Information.
The two molecules have essentially identical geometries
(Table 1), the rms fit of the non-hydrogen atoms being ca.
0.05 Å. The mirror plane contains the iron center, the two
imine units, and Cl(1). The geometry at the iron is distorted
trigonal bipyramidal, with Cl(1) and N(2) in the axial
positions; the metal lies ca. 0.23 [0.23 Å] out of the equatorial
{N(1),Cl(2),Cl(2A)} plane (the value in square parentheses
refers to molecule II). The bite of the N,N′ chelating ligand
is slightly contracted compared to that seen in 13, 77.2(2)°
[77.7(2)°] in the case of 31, cf. 78.43(15)° in 13. For both
the chlorines and the nitrogens, it is noticeable that the axial
bonds, (Fe-Cl(1) 2.285(3) [2.294(3) Å], Fe-N(2) 2.243(7)
[2.275(7) Å]) are longer than their equatorial counterparts
(Fe-Cl(2) 2.2053(19) [2.2109(19) Å], Fe-N(1) 2.145(6)
[2.141(6) Å]). The C-C bond between the imine units is
noticeably shorter here (1.455(12) [1.456(12) Å]) than in 13
(1.521(6) Å), possibly due to a lesser steric repulsion between
the carbon substituents (protons in 31 compared to phenyls
in 13).
Similar trends to those described above are seen in the
rates and molecular weight data for the tBu derivatives, 19-
24. Incorporation of electron-withdrawing groups into the
ligand backbone results in a low-molecular-weight, olefin-
terminated polymer, whereas electron-donating groups tend
back toward ATRP behavior. These data are included in the
Supporting Information.
Synthesis of Fe(III) Compounds. In our preliminary
report,6 we showed that a key factor in determining the
differing radical polymerization mechanisms catalyzed by
R-diimine iron catalysts is the spin-state of the metal in the
oxidized Fe(III) species present in the ATRP equilibrium,
high-spin centers, giving rise to ATRP, whereas intermediate-
spin centers afforded CCT behavior. To further explore the
origin of the differing polymerization mechanisms observed
using aryl-substituted Fe(R-diimine) catalysts, Fe(III) deriva-
tives were targeted and prepared according to Scheme 4.
Careful control over the reaction conditions was required to
ensure the desired products were obtained in high purity.
Unlike their Fe(II) counterparts, 25-37 did not give
The intermediate-spin complexes 25-29 and 32-36 could
not be characterized crystallographically because of their
instability in both the solid and solution states. These
compounds were found to be sensitive to light, air, and
moisture, and control over the concentration, time, and
(56) The lower µeff values observed for these Fe(III) complexes could also
be due to a mixed spin system or spin admixture, rather than a pure
S ) 3/2 spin-state. However, the presence of the lower spin-state species
undoubtedly correlates with the differing polymerization behavior.
(57) Daran, J.-C.; Jeannin, Y.; Martin, L. M. Inorg. Chem. 1980, 19, 2935.
(58) Millington, K. R.; Wade, S. R.; Willey, G. R. Inorg. Chim. Acta 1984,
89, 185.
1
meaningful H NMR spectra. Evans’ NMR measurements
Inorganic Chemistry, Vol. 46, No. 21, 2007 8967

Allan et al.
Figure 7. Hammett plot showing the correlation between the substituent
constant and the rate of styrene polymerization for Cy,R′′[N,N]FeCl2 catalysts.
Figure 8. Hammett plot showing the correlation between the substituent
constant and the molecular weight of polystyrene at 50% conversion for
Cy,R′′[N,N]FeCl2 catalysts.
temperature of a reaction was necessary to isolate pure
complexes. This is not unexpected because a decrease in the
halogenophilicity of the parent Fe(II) complexes would
potentially promote the decomposition (via halogen loss) and
disproportionation reactions of the intermediate-spin Fe(III)
species. Attempts to crystallize the dark-purple complex
tBu,FPh[N,N]FeCl3, 33, resulted in the formation of red crystals,
whose X-ray structures showed that two molecules of
tBu,F-Ph[N,N]FeCl3 had disproportionated to form tBu,F-Ph[N,N]-
FeCl2, FeCl4- and tBu,F-Ph[N,N]H+, 38 (Figure S28, Support-
ing Information).
calculations predict an increase in the Cl(2)-Fe-Cl(2)*
angle by approximately 10° in the quartet system and also
show a significant difference in the Fe-N(2) bond lengths
between the congeners, with the 2.31 Å sextet state being
0.24 Å longer than the 2.07 Å quartet state. Complexes with
5
S ) /2, sextet, structures are predicted to exhibit trigonal
bipyramidal structures, whereas S ) 3/2, quartet, complexes
should favor a more square-pyramidal geometry. The Fe-
N(2) bond length obtained from the solid-state structure is
2.24 Å, correlating well with the sextet spin-state, and the
geometry at the iron center is best described as trigonal
bipyramidal.
Discussion
Correlation of the Metal Spin-State with the Polym-
erization Mechanism. Previous work showed that Cy,H[N,N]-
FeCl3 exhibited a solution magnetic moment of 5.97 µB,
corresponding well with the spin-only value of 5.92 µB for
The Effect of Para-Substituted Phenyl Substituents on
the Observed Polymerization Mechanism. Incorporating
electron-withdrawing aryl groups into the diimine backbone
results in a switch of the polymerization mechanism from
ATRP toward CCT, despite the presence of the ATRP-
directing N-alkyl donor. When the more electron-withdraw-
ing para-fluorophenyl substituents are incorporated, there is
a greater prevalence of CCT, the rate is decreased, and the
polymer obtained is of even lower molecular weight (Mn
∼1800 for Cy,F-Ph[N,N]FeCl2, cf. Mn ∼3400 for Cy,Ph[N,N]-
FeCl2, corresponding to approximately 17 versus 33 mono-
mer insertions, respectively). Because the radical concen-
tration is determined by the interplay of the ATRP and CCT
equilibria, it is not possible, without extensive modeling, to
employ Mayo plots to determine the chain-transfer constants.
The effect of these equilibria on the radical concentration in
the polymerization medium is discussed in a separate study,
which also addresses the role of organometallic intermediates
in these processes.53
Switching to an electron-donating group at the para
position of the phenyl ring results in an increase both in rate
and in the molecular weight of the polymer obtained (the
peak molecular weight is ∼3500 for Cy,Me-Ph[N,N]FeCl2, cf.
Mn ∼7000 for Cy,MeO-Ph[N,N]FeCl2, corresponding to ap-
proximately 34 vs 67 monomer insertions, respectively).
When a strongly electron-donating group is installed in the
para position, as in the case of Cy,NMe2-Ph[N,N]FeCl2, the
mechanism switches back to ATRP, and the rate is further
increased to the highest value recorded within this (R-
diimine)Fe catalyst family. The equilibria between the two
5
a d5-high-spin Fe(III) center (S ) /2, sextet). In contrast,
the magnetic moment of the N-aryl derivative DiPP,H[N,N]-
FeCl3 (where DiPP is 2,6-diisopropylphenyl) was 3.99 µB,
correlating well to a d5-intermediate-spin Fe(III) center (3.87
µB, S ) 3/2, quartet).59 The magnetic moments for the
majority of the 2,3 aryl-substituted derivatives 25-29 and
32-36 were found to lie in the range 3.9-4.2 µB, consistent
with intermediate-spin Fe(III) centers, whereas the solely
alkyl-substituted complex, 31, was high-spin (µeff ) 5.81 µB).
Interestingly, the para-dimethylamino substituted complexes,
30 and 37, also exhibited solution magnetic moments
consistent with high-spin Fe(III) centers. This difference in
spin-state correlates well with the evident switch in polym-
erization mechanism, high-spin species catalyzing ATRP,
whereas the lower-spin species give rise to polymers via
CCT. Evans’ method determinations of µeff for 26 at elevated
temperatures confirmed that the lower spin-state is main-
tained at the polymerization temperature of 120 °C.
The expected geometric differences between the sextet and
quartet spin-states are reproduced in the solid-state structure
of 31. The Hartree-Fock calculations presented previously6
suggest that a change in spin-state should be visible in the
solid-state structures of the complexes, due to a distortion
of the antiperiplanar nitrogen donor bond parameters. The
(59) Gibson, V. C.; O’Reilly, R. K.; Rzepa, H. S.; Shaver, M. P. Polym.
Prepr. (Am. Chem. Soc., DiV. Polym. Chem.) 2005, 46, 160.
8968 Inorganic Chemistry, Vol. 46, No. 21, 2007

Metal Spin-State with Polymerization Mechanism
Table 2. Crystal Data, Data Collection, and Refinement Parameters for
13 and 31a
competing mechanisms are clearly finely balanced; small
changes to the electronic characteristics of the ligand
backbone dramatically affect the prevalence of ATRP versus
CCT polymerization pathways.
data
13
31
formula
solvent
fw
C26H32Cl2FeN2
0.5MeOH
515.31
C10H20Cl3FeN2
There is a clear correlation between the nature of the para
substituent of the phenyl ring in the 2,3-aryl substituted
complexes and both the rate of styrene polymerization and
molecular weight of polystyrene obtained. As the para
substituent becomes more electron-donating, the rate of
polymerization increases, and the peak molecular weight for
the complexes terminating through CCT is higher. Hammett
plots of substituent constant versus log(kX/kH) (Figure 7) and
of the polymer molecular weight at 50% conversion (Figure
8) show good linearity for Cy,R′′[N,N]FeCl2 catalysts.
330.48
color, habit
purple needles
orange needles
cryst size (mm)
T (K)
0.36 × 0.12 × 0.04 0.35 × 0.01 × 0.01
173
173
cryst syst
space group
a (Å)
b (Å)
c (Å)
tetragonal
P4h21m (no. 113)
20.3559(6)
orthorhombic
Cmc21 (no. 36)
10.2994(13)
19.845(3)
14.723(2)
6.5223(4)
R (deg)
â (deg)
γ (deg)
V (Å3)
2702.6(2)
4b
1.266
Mo KR
0.774
65
3009.2(7)
8c
1.459
Mo KR
1.513
66
Z
Dc (g cm-3
)
Conclusions
radiation used
µ (mm-1
2θ max (deg)
A series of new R-diimine ligands has been synthesized
through the TiCl4-activated condensation of the requisite
ketone and amine. Fe(II) and Fe(III) complexes of these
ligands have been prepared and characterized, including the
first R-diimine FeCl3 complex to be crystallographically
characterized. The catalytic reactivity of the Fe(II) dichloride
complexes has been investigated through styrene polymer-
ization studies, and the prevalent polymerization mechanism
has been shown to correlate with the spin-state of the Fe-
(III) trichloride species. Incorporating an electron-withdraw-
ing group, such as Ph or F-Ph, at the 2,3 positions of the
ligand backbone switches the polymerization mechanism
from ATRP to CCT, even in the presence of ATRP-directing
alkylimino donors. Substituting the para position of the aryl
group on the 2,3 position of the ligand backbone with
electron-donating substituents switches the polymerization
mechanism back to ATRP. The Hammett substituent con-
stant, σ, correlates well with both the rate of styrene
polymerization and the peak molecular weight of the isolated
polymer, illustrating a structure-reactivity relationship that
could prove useful in the future design of selective catalysts
based on metal spin-state.
)
no. of unique reflns measured 4873
4282
obs, |Fo| > 4σ(|Fo|)
4722
164
0.077, 0.173
2870
175
0.097, 0.099
no. of variables
d
R1, wR2
a Details in common: graphite monochromated radiation, refinement
based on F 2. b The complex has crystallographic CS symmetry. c There are
two crystallographically independent CS-symmetric complexes in the
2
2
asymmetric unit. d R1 ) ∑||Fo| - |Fc||/∑|Fo|; wR2 ) {∑[w(Fo - Fc )2]/
2
2
∑[w(Fo )2]}1/2; w-1 ) σ2(Fo ) + (aP)2 + bP.
Platform spectrometers at Imperial College London. GPC data were
collected using Cirrus GPC/SEC software, ver. 1.11, connected to
a Shodex RI-101 detector and were referenced to polystyrene
standards (PolymerLabs 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.
Diazadiene ligands RNdCH-CHdNR, where R ) tBu or Cy, were
prepared in bulk through modification of the literature proce-
dures7,8,60 and purified through recrystallization. FeCl2(THF)1.5 was
synthesized according to literature procedures.61 FeCl3 was pur-
chased from Aldrich Chemical Co. and used as received. All of
the amines were freshly distilled before use.
Experimental Section
General Experimental Procedures. All of the 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. Elemental analyses were
performed by the microanalytical service of the chemistry depart-
ment of London Metropolitan University. Crystal data were
collected on Oxford Diffraction Xcalibur 3 and PX Ultra diffrac-
tometers. NMR spectra were recorded on Bruker AC250 (1H, 250.1
MHz; 13C, 60.9 MHz), DRX400 (1H, 400.1 MHz; 13C, 100.6 MHz),
and AV-400 (1H, 400.3 MHz; 13C, 100.7 MHz) spectrometers, at
293 K unless otherwise stated. 1H and 13C chemical shifts are
reported as δ and referenced to the residual proton signal and to
the 13C signal of the deuterated solvent, respectively. The following
abbreviations have been used for multiplicities: s (singlet), d
(doublet), t (triplet), m (unresolved multiplet); coupling constants
are reported in hertz. Infrared spectra were obtained as nujol mulls
on chemical ionization (CI) plates or as KBr discs on a PerkinElmer
1710X FTIR spectrometer. FAB, electron ionization, and CI mass
spectra were recorded on Micromass AutoSpec Premier and VG
Polymerization Procedures. All of the polymerizations were
set up and performed under an atmosphere of oxygen-free, dry
nitrogen. A solution of monomer, initiator, and catalyst (200:1:1;
[M]0 ) 8.73 M, [I]0 ) 0.044 M, [cat] ) 0.044 M) was added to a
30 mL ampule equipped with a magnetic stirrer bar. The ampules
were heated in a sand bath, at 120 °C, with magnetic stirring. After
stirring for the allotted period of time (15 min-2 h), an aliquot
(0.1 mL) was removed. Conversion was determined by integration
1
of the monomer versus polymer backbone resonances in the H
NMR spectrum of the crude product in CDCl3. Once the reaction
was completed, the contents of the ampules were dissolved in THF
and added dropwise to an approximately 20-fold excess of rapidly
stirred acidified methanol (1% HCl v/v). The precipitate that formed
(60) Hsieh, A. T. T.; West, B. O. J. Organomet. Chem. 1976, 112, 285.
(61) Cotton, F. A.; Luck, R. L.; Son, K.-A. Inorg. Chim. Acta 1991, 179,
11.
Inorganic Chemistry, Vol. 46, No. 21, 2007 8969

Allan et al.
was filtered off and washed with methanol. The precipitate was
dried for 24 h under a vacuum. Samples were analyzed by GPC.
X-ray Crystallographic Analyses. Table 2 provides a summary
of the crystallographic data for 13 and 31. Data were collected using
an Oxford Diffraction Xcalibur 3 diffractometer, and the structures
were refined based on F 2 using the SHELXTL and SHELX-97
program systems.62 The absolute structures of 13 and 31 were
unambiguously determined by a combination of R-factor tests [for
(3), x- ) +0.94(3); for 31: x+ ) +0.05(4), x- ) +0.95(4)].
CCDC 627620 (13) and 627626 (31).
Acknowledgment. The Engineering and Physical Sci-
ences Research Council, UK, and the Natural Sciences and
Engineering Research Council, Canada, are thanked for
financial support.
Supporting Information Available: Experimental details for
1-37, crystallographic data for all of the reported structures (CCDC
Nos. 627620-627629) and polymerization data for 19-24. This
material is available free of charge via the Internet at http:
//pubs.acs.org.
+
+
13: R1 ) 0.0773, R1- ) 0.0860; for 31: R1 ) 0.0969, R1- )
0.1032] and by use of the Flack parameter [for 13: x+ ) +0.06-
(62) SHELXTL PC version 5.1, B. A., Madison, WI, 1997; Sheldrick, G.
SHELX-97, Institut Anorg. Chemie, Tammannstr. 4, D37077: Go¨t-
tingen, Germany, 1998.
IC701500Y
8970 Inorganic Chemistry, Vol. 46, No. 21, 2007