Metal versus ligand alkylation in the reactivity of the (bis-iminopyridinato)Fe catalyst

Article abstract of DOI:10.1021/ja054152b

The alkylation of the Brookhart-Gibson {2,6-[2,6-(i-Pr)₂PhN= C(CH₃)]₂(C₅H₃N)} FeCl₂ precatalyst with 2 equiv of LiCH₂Si(CH₃)₃ led to the isolation of several catalytically very active products depending on the reaction conditions. The expected dialkylated species {2,6-[2,6-(i-Pr) ₂PhN=C(CH₃)]₂}(C₅H ₃N)Fe(CH₂-SiMe₃)₂ (2) was indeed the major component of the reaction mixture. However, other species in which alkylation occurred at the pyridine ring ortho position, {2,6-[2,6-(i-Pr) ₂PhN=C(CH₃)]₂-2-CH₂SiMe ₃}-(C₅H₃N)Fe(CH₂SiMe₃) (1), and at the imine C atom, {2-[2,6-(i-Pr)₂PhN=C(CH ₃)]-6-[2,6-(i-Pr)₂PhNC(CH₃)-(CH₂ SiMe₃)](C₅H₃N)}Fe(CH₂SiMe ₃) (3), have also been isolated and fully characterized. In addition, deprotonation of the methyl-imino functions and formation of a new divalent Fe catalyst {[2,6-[2,6-(i-Pr)₂PhN-C=(CH₂)]₂(C ₅H₃N)}Fe(μ-Cl)Li(THF)₃ (4) also occurred depending on the reaction conditions. In turn, the formation of 4 might trigger the reductive coupling of two units through the methyl-carbon wings. This process resulted in the one-electron reduction of the metal center, affording a dinuclear Fe(I) alkyl catalyst {[{[2,6-(i-Pr)₂C₆H ₅]N=C(CH₃)}(C₅H₃N){[2,6-(i-Pr) ₂C₆H₅]N=CCH₂}Fe(CH ₂SiMe₃)]}₂ (5). Different from other metal derivatives, complex 5 could not be prepared from the monodeprotonated version of the ligand. Its reaction with a mixture of FeCl₂ and RLi afforded instead [{2,6-[2,6-(i-Pr)₂PhN-C=(CH₂)]₂(C ₅H₃N)}FeCH₂Si(CH₃) ₃]-[Li(THF)₄] (6) which is also catalytically active. All of these high-spin species have been shown to have high catalytic activity for olefin polymerization, producing polymers of two distinct natures, depending on the formal oxidation state of the metal center.

Full text of DOI:10.1021/ja054152b

Published on Web 08/24/2005
Metal versus Ligand Alkylation in the Reactivity of the
(Bis-iminopyridinato)Fe Catalyst
Jennifer Scott,† Sandro Gambarotta,*,† Ilia Korobkov,† and Peter H. M. Budzelaar*,‡
Contribution from the Department of Chemistry, UniVersity of Ottawa,
Ottawa, Ontario K1N 6N5, Canada, and Department of Inorganic Chemistry,
Radboud UniVersity Nijmegen, ToernooiVeld 1, 6525 ED Nijmegen, The Netherlands
Received June 23, 2005; E-mail: sgambaro@science.uottawa.ca; budzelaa@cc.umanitoba.ca
Abstract: The alkylation of the Brookhart-Gibson {2,6-[2,6-(i-Pr)2PhNdC(CH3)]2(C5H3N)} FeCl2 precatalyst
with 2 equiv of LiCH2Si(CH3)3 led to the isolation of several catalytically very active products depending on
the reaction conditions. The expected dialkylated species {2,6-[2,6-(i-Pr)2PhNdC(CH3)]2}(C5H3N)Fe(CH2-
SiMe3)2 (2) was indeed the major component of the reaction mixture. However, other species in which
alkylation occurred at the pyridine ring ortho position, {2,6-[2,6-(i-Pr)2PhNdC(CH3)]2-2-CH2SiMe3}-
(C5H3N)Fe(CH2SiMe3) (1), and at the imine C atom, {2-[2,6-(i-Pr)2PhNdC(CH3)]-6-[2,6-(i-Pr)2PhNC(CH3)-
(CH2 SiMe3)](C5H3N)}Fe(CH2SiMe3) (3), have also been isolated and fully characterized. In addition,
deprotonation of the methyl-imino functions and formation of a new divalent Fe catalyst {[2,6-[2,6-(i-Pr)2PhNs
Cd(CH2)]2(C5H3N)}Fe(µ-Cl)Li(THF)3 (4) also occurred depending on the reaction conditions. In turn, the
formation of 4 might trigger the reductive coupling of two units through the methyl-carbon wings. This process
resulted in the one-electron reduction of the metal center, affording a dinuclear Fe(I) alkyl catalyst {[{[2,6-
(i-Pr)2C6H5]NdC(CH3)}(C5H3N){[2,6-(i-Pr)2C6H5]NdCCH2}Fe(CH2SiMe3)]}2 (5). Different from other metal
derivatives, complex 5 could not be prepared from the monodeprotonated version of the ligand. Its reaction
with a mixture of FeCl2 and RLi afforded instead [{2,6-[2,6-(i-Pr)2PhNsCd(CH2)]2(C5H3N)}FeCH2Si(CH3)3]-
[Li(THF)4] (6) which is also catalytically active. All of these high-spin species have been shown to have
high catalytic activity for olefin polymerization, producing polymers of two distinct natures, depending on
the formal oxidation state of the metal center.
Introduction
toward a nonanticipated ability of the diimine ligand to be
involved in the reactivity of the M-C bond. The ligand is
The tremendous activity displayed by diimine complexes of
late transition metals as olefin polymerization catalysts has
increasingly attracted interest since the initial discoveries by
Brookhart and Gibson1 and fueled research aimed at under-
standing these fascinating systems. The bis-iminopyridine
complexes, and the Fe derivative {2,6-[2,6-(i-Pr)2PhNdC(CH3)]2-
(C5H3 N)}FeCl2 in particular, are especially prominent in this
family of Ziegler-Natta catalysts, given their outstanding
activity that rivals even that of the early metal-based commercial
catalysts. The three key issues at the basis of the exceptional
catalytic performance of these systems, which still remain not
well understood, are (1) the role played by the ligand system,
(2) the oxidation state and electronic configuration of the metal
in the catalytically active species, and (3) the stability of the
M-C bond.
definitely noninnocent and has been shown to engage in a series
of transformations. These involve alkylation at any position of
the pyridine ring,2,3 including the N atom,7 deprotonation of
the methyl sidearms,3,7b,8,9 alkylation of the imine functions,10
and even dimerization via either C-C bond formation through
the deprotonated alkyl substituents3,5,6 or pyridine ring cyclo-
addition.3 Some of these processes can be reversed and may or
may not involve redox of the metal center.
The second point concerns the actual oxidation state and
electronic configuration of the metal center. The bis-imino-
(2) Reardon, D.; Conan, F.; Gambarotta, S.; Yap, G. P. A.; Wang, Q. J. Am.
Chem. Soc. 1999, 121, 9318.
(3) Sugiyama, H.; Aharonian, G.; Gambarotta, S.; Yap, G. P. A.; Budzelaar,
P. H. M. J. Am. Chem. Soc. 2002, 124, 12268.
(4) Reardon, D.; Aharonian, G.; Gambarotta, S.; Yap, G. P. A. Organometallics
2002, 21, 786.
(5) Scott, J.; Gambarotta, S.; Korobkov, I. Can. J. Chem. 2005, 83, 279.
(6) Sugiyama, H.; Korobkov, I.; Gambarotta, S.; Moeller, A.; Budzelaar, P.
H. M. Inorg. Chem. 2004, 43, 5771.
Recent work in the chemistry of vanadium,2 chromium,3
manganese,3,4 cobalt,5 and lanthanides6 has clearly pointed
(7) (a) Clentsmith, G. K. B.; Gibson, V. C.; Hitchcock, P. B.; Kimberley, B.
S.; Rees, C. W. Chem. Commun. 2002, 1498. (b) Khorobkov, I.; Gam-
barotta, S.; Yap, G. P. A. Organometallics 2002, 21, 3088. (c) Blackmore,
I. J.; Gibson, V. C.; Hitchcock, P. B.; Rees, C. W.; Williams, D. J.; White,
A. J. P. J. Am. Chem. Soc. 2005, 127, 6012.
† University of Ottawa.
‡ Radboud University of Nijmegen.
(1) (a) Johnson, L. K.; Killian, C. M.; Brookhart, M. S. J. Am. Chem. Soc.
1995, 117, 6414. (b) Killian, C. M.; Tempel, D. J.; Johnson, L. K.;
Brookhart, M. S. J. Am. Chem. Soc. 1996, 118, 11664. (c) Small, B. L.;
Brookhart, M.; Bennett, A. M. A. J. Am. Chem. Soc. 1998, 120, 4049. (d)
Britovsek, G. J. P.; Gibson, V. C.; Kimberley, B. S.; Maddox, P. J.;
McTavish, S. J.; Solan, G. A.; White, A. J. P.; Williams, D. J. Chem.
Commun. 1998, 849.
(8) Enright, D.; Gambarotta, S.; Yap, G. P. A.; Budzelaar, P. H. M. Angew.
Chem., Int. Ed. 2002, 41, 3873.
(9) Sugiyama, H.; Gambarotta, S.; Yap, G. P. A.; Wilson, D. R.; Thiele, S.
K.-H. Organometallics 2004, 23, 5054.
(10) Bruce, M.; Gibson, V. C.; Redshaw, C.; Solan, G. A.; White, A. J. P.;
Williams, D. J. Chem. Commun. 1998, 2523.
9
10.1021/ja054152b CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 13019-13029
13019

A R T I C L E S
Scott et al.
pyridine ligand system is well-known to act as a potent stabilizer
of the metal center and has allowed the isolation of species
which, formula-wise, have the appearance of carrying the metal
in low or exotic oxidation states.6,11-13 However, the low formal
oxidation state of some of these complexes may be deceiving
since the ligand can accommodate up to three electrons in the
delocalized p-system.8 Therefore, an internal charge-transfer
process may occur which results in a higher actual metal
oxidation state.6 Remarkably however, the reactivity of these
“low-valent” metal centers is not quenched, given that dinitrogen
complexes of this particular ligand system have been isolated
for both vanadium12 and iron.13 This is an indication that
intermediate reduced species can be stabilized while preserving
their high reactivity. In the case of the bis-iminopyridine Fe
catalyst, the oxidation state of the metal in the catalytically active
species has not been conclusively determined. Based on
Mossbauer and EPR studies of the pre-catalyst, it was argued
that the active species may involve the trivalent oxidation state
of Fe.14 This conclusion was debated by Talsi et al., who
suggested, on the basis of 1H NMR, 2H NMR, and EPR studies,
the formation of a paramagnetic Fe(II) alkyl bridging the Al
cocatalyst.15 In addition to oxidation state, the spin multiplicity,
in relation to the coordination number, also seems to play an
important role in determining high activity. Calculations on the
catalytically active species in the Fe catalyst, assumed to be
either a cationic [LFeR]+ complex16 or a neutral LFeMe-
(µMe)2AlMe2 species,17 have led to a debate concerning the
spin state of the metal center. Recent observations by Chirik
on the catalytic activity of (diimine)FeR2 complexes18 lead to
the tentative conclusion that the combination of high-spin state
and tetracoordination may cause divalent Fe alkyls to be poorly
active catalyst precursors due to the absence of empty orbitals
available for olefin binding. Conversely, the recent isolation of
cationic and divalent organoiron species has established the
competency of cationic Fe(II) alkyls as polymerization catalysts.18c
The last point is the stability of the M-C bond. This issue,
which is obviously central to catalyst behavior and activity, is
particularly relevant in the case of the Fe catalyst. Organo-iron
compounds are reasonably well established, including complexes
of diamines and diimines, the latter being closely related to this
work.18 These species, however, display either poor stability
toward reduction to the mono-18 or zerovalent state or substan-
tially diminished catalytic activity.19 Yet, organo-iron complexes
produce polymers when activated by MAO and Lewis acids.
Although the stability of the Fe-C bond may be greatly
increased by using ancillary ligands such as phosphines19m-o
or cyclopentadienyl and CO,19a this practice usually quenches
the reactivity of the metal center. It is therefore conceivable
that, given the outstanding activity of the bis-iminopyridine Fe
catalyst, the ligand provides the M-C bond with the appropriate
stability to disfavor termination pathways while maintaining the
high reactivity necessary for catalytic efficiency.
The aim of this study was to understand the chemistry of
the Fe-C functionality in the bis-iminopyridine Fe catalyst
{2,6-[2,6-(i-Pr)2PhNdC(CH3)]2(C5H3N)}FeCl2. Given the com-
plexity of the behavior of this noninnocent ligand and its
frequent involvement in the reactivity of the metal carbon bond,
it is nearly impossible at this stage to reasonably argue about
the true nature of the catalytically active species. Contrary to
the other metal complexes mentioned above, in which alkylation
processes have allowed the isolation of species that provide
important mechanistic hints and partially unveil the complexity
of this chemistry,2-4,9,10,20 a crystallographically characterized
organo-iron active intermediate has been isolated in only one
case.18c
Herein, we describe the results of the alkylation of the bis-
iminopyridine-FeCl2 catalyst with R-Li [R ) CH2Si(CH3)3].
The choice of this particular alkylating agent was advised by
the well-known stability of its organometallic derivatives.
Experimental Section
All operations were performed either under a nitrogen atmosphere
using standard Schlenk techniques or in a purified nitrogen-filled
drybox. The THF complex of FeCl2 was prepared according to the
standard procedure. The ligand 2,6-[2,6-(i-Pr)2PhNdC(CH3)]2(C5 H3N),1
the mono-7 and di-deprotonated6,9 derivatives, and LiCH2Si(CH3)3
21
were prepared following published procedures. Infrared spectra were
recorded on a Mattson 9000 and Nicolet 750-Magna FT-IR instrument
from Nujol mulls prepared in a drybox. Samples for magnetic
susceptibility measurements were weighed inside a drybox equipped
with an analytical balance and sealed into calibrated tubes, and the
measurements were carried out at room temperature with a Gouy
balance (Johnson Matthey). Magnetic moments were calculated fol-
lowing standard methods, and corrections for underlying diamagnetism
were applied to the data. Elemental analyses were performed on a
Perkin-Elmer 2400 CHN analyzer. Data for X-ray crystal structure
determinations were obtained with a Bruker diffractometer equipped
with a Smart CCD area detector.
(19) (a) Kerber, R. C. ComprehensiVe Organometallic Chemistry II; Abel, E.
W., Stone, F. G. A., Wilkinson, G., Eds.; Pergamon, New York, 1995. (b)
Machelett, B. Z. Chem. 1976, 16, 116. (c) Mueller, H.; Seidel, W.; Goerls,
H. J. Organomet. Chem. 1993, 445, 133. (d) Klose, A.; Solari, E.; Floriani,
C.; Chiesi-Villa, A.; Rizzoli, C.; Re, N. J. Am. Chem. Soc. 1994, 116, 9123.
(e) Seidel, W.; Mueller, H.; Goerls, H. Angew. Chem., Int. Ed. 1995, 34,
325. (f) Wehmschulte, R. J.; Power, P. P. Organometallics 1995, 14, 3264.
(g) Viefhaus, T.; Schwarz, W.; Hubler, K.; Locke, K.; Weidlein, J. Z. Anorg.
Allg. Chem. 2001, 627, 715. (h) Holland, P. L.; Cundari, T. R.; Perez, L.
L.; Eckert, N. A.; Lachicotte, R. J. J. Am. Chem. Soc. 2002, 124, 14416.
(i) Smith, J. M.; Lachicotte, R. J.; Holland, P. L. Organometallics 2002,
21, 4808. (j) Sciarone, T. J. J.; Meetsma, A.; Hessen, B.; Teuben, J. H.
Chem. Commun. 2002, 1580. (k) Kisko, J. L.; Hascall, T.; Parkin, G. J.
Am. Chem. Soc. 1998, 120, 10561. (l) Fryzuk, M. D.; Leznoff, D. B.; Ma,
E. S. F.; Rettig, R. J.; Young, V. G. Organometallics 1998, 17, 2313. (m)
Hermes, A. R.; Girolami, G. S. Organometallics, 1987, 6, 763. (n) Chatt,
J.; Shaw, B. L. J. Chem. Soc. 1961, 285. (o) Seidel, V. W.; Lattermann,
K.-J. Z. Anorg. Allg. Chem. 1982, 488, 69. (p) Spencer, L. P.; Altwer, R.;
Wei, P.; Gelmini, L.; Gauld, J.; Stephan, D. W. Organometallics 2003,
22, 3841. (q) Mueller, H.; Seidel, W.; Goerls, H. Z. Anorg. Allg. Chem.
1996, 622, 756. (r) Hill, D. H.; Sen, A. J. Am. Chem. Soc. 1988, 110,
1651.
(11) de Bruin, B.; Bill, E.; Bothe, E.; Weyhermueller, T.; Wieghardt, K. Inorg.
Chem. 2000, 39, 2936.
(12) Vidyaratne, I.; Gambarotta, S.; Korobkov, I.; Budzelaar, P. H. M. Inorg.
Chem. 2005, 44, 1187.
(13) Bart, S. C.; Lobkovsky, E.; Chirik, P. J. J. Am. Chem. Soc. 2004, 126,
13794.
(14) Britovsek, G. J. P.; Clentsmith, G. K. B.; Gibson, V. C.; Goodgame, D.
M. L.; McTavish, S. J.; Pankhurst, Q. A. Catal. Commun. 2002, 3, 207.
(15) (a) Bryliakov, K. P.; Semikolenova, N. V.; Zudin, V. N.; Zakharov, V. A.;
Talsi, E. P. Catal. Commun. 2004, 5, 45. (b) Bryliakov, K. P.; Semikole-
nova, N. V.; Zakharov, V. A.; Talsi, E. P. Organometallics 2004, 23, 5375.
(16) (a) Deng, L.; Margl, P.; Ziegler, T. J. Am. Chem. Soc. 1999, 121, 6479.
(b) Griffiths, E. A. H.; Britovsek, G. J. P.; Gibson, V. C.; Gould, I. R.
Chem. Commun. 1999, 1333. (c) Khoroshun, D. V.; Musaev, D. G.; Vreven,
T.; Morokuma, K. Organometallics 2001, 20, 2007.
(20) (a) Kooistra, T. M.; Knijnenburg, Q.; Smits, J. M. M.; Horton, A. D.;
Budzelaar, P. H. M.; Gal, A. W. Angew. Chem., Int. Ed. 2001, 40, 4719.
(b) Gibson, V. C.; Humphries, M. J.; Tellmann, K. P.; Wass, D. F.; White,
A. J. P.; Williams, D. J. Chem. Commun. 2001, 2252. (c) Humphries, M.
J.; Tellmann, K. P.; Gibson, V. C.; White, A. J. P.; Williams, D. J.
Organometallics 2005, 24, 2039.
(17) Zakharov, I. I.; Zakharov, V. A. Macromol. Theory Simul. 2004, 13, 583.
(18) (a) Bart, S. C.; Hawrelak, E. J.; Schmisseur, A. K.; Lobkovsky, E.; Chirik,
P. J. Organometallics 2004, 23, 237. (b) Bouwkamp, M. W.; Bart, S. C.;
Hawrelak, E. J.; Trovitch, R. J.; Lobkovsky, E.; Chirik, P. J. Chem.
Commun. 2005, 3406. (c) Bowkamp, M. W.; Lobkovsky, E.; Chirik, P. J.
J. Am. Chem. Soc. 2005, 127, 9660.
(21) Tessier-Youngs, C.; Beachley, O. T., Jr. Inorg. Synth. 1986, 24, 95.
9
13020 J. AM. CHEM. SOC. VOL. 127, NO. 37, 2005

Alkylation in (Bis-iminopyridinato)Fe Catalyst Reactivity
A R T I C L E S
Preparation of {2,6-[2,6-(i-Pr)2PhNdC(CH3)]2(C5H3N)}FeCl2. A
suspension of 2,6-[2,6-(i-Pr)2PhNdC(CH3)]2(C5H3N) (2.05 g, 4.26
mmol) in THF was added to a suspension of FeCl2(THF)1.5 (1.00 g,
4.26 mmol) in THF, and the resulting dark blue mixture was stirred
overnight. After evaporating the THF, the residue was dissolved in an
appropriate amount of CH2Cl2, concentrated, and layered with hexane.
Dark blue crystals of the Fe-bis-iminopyridine-dichloride starting
complex were grown while sitting at room temperature for 2 days.
Although this procedure is very similar to published procedures,1 this
synthesis strictly avoids the use of water or alcohols and, as such,
ensures a “dry” starting complex for further reactions.
centrifuged, and a dark olive-green solution (from which 4 crystallized)
was separated from the dark precipitates. Diethyl ether was added to
the precipitates, and the solution was centrifuged to separate a dark
purple solution from white precipitates. Concentration of the solution
and standing in the freezer for 2 days afforded crystals of 2 (0.304 g,
0.427 mmol, 52% yield).
Preparation of {2-[2,6-(i-Pr)2PhNdC(CH3)]-6-[2,6-(i-Pr)2PhNC-
(CH3)(CH2SiMe3)](C5H3N)}Fe(CH2SiMe3) (3). A procedure identical
to that described for complex 1 was followed. Crystals of 3 grew out
of the same hexane solution as 1 but could be physically separated
with the help of a stereomicroscope due to their very distinct rectangular
shape (0.055 g, 0.08 mmol, 10% yield). Anal. Calcd (found) for C41H65-
FeN3Si2: C, 69.16 (69.10); H, 9.20 (9.18); N, 5.90 (5.87).
Preparation of {2,6-[2,6-(i-Pr)2PhNdC(CH3)]2-(2-CH2SiMe3)}-
(C5H3N)Fe(CH2SiMe3) (1). A suspension of FeCl2(THF)1.5 (0.200 g,
0.85 mmol) and 2,6-[2,6-(i-Pr)2PhNdC(CH3)]2(C5H3N) (0.410 g, 0.85
mmol) in approximately 10 mL of THF was stirred for 4 h affording
the usual dark-blue color. The suspension was cooled to -35 °C and
mixed with a cooled solution of LiCH2SiMe3 (0.168 g, 1.78 mmol) in
THF. The color of the solution instantly changed from royal blue to
dark reddish-orange upon mixing. The solution was stirred for
approximately 1 min to ensure complete mixing and evaporated to
dryness, maintaining a cold temperature throughout the procedure. Cold
hexane was added to the dark brown residue, and a dark reddish-orange
suspension was centrifuged to remove a substantial amount of insoluble
material. The resulting solution was allowed to stand at -35 °C for 1
day, upon which time the mother liquor was removed from small dark
crystals (recognized later as a small amount of 2), concentrated, and
put back in the freezer for 2 days, resulting in the crystallization of 1
as dark red crystals (0.212 g, 0.30 mmol, 35% yield). Anal. Calcd
(found) for C41H65FeN3Si2: C, 69.16 (69.05); H, 9.20 (9.18); N, 5.90
(5.87). IR (Nujol mull, cm-1): ν 3066 (w), 2929 (s), 1614 (m), 1568
(s), 1498 (s), 1363 (m), 1352 (m), 1321 (s), 1288 (m), 1242 (m), 1195
(m), 1143 (m), 1099 (w), 1056 (w), 1043 (w), 970 (w), 946 (w), 937
Transformation of 3 to 2. Addition of diethyl ether to crystals of
3 (0.055 g, 0.08 mmol) afforded a dark purple solution from which
dark purple crystals of 2 were isolated after freezing at -35 °C for 2
days (0.035 g, 0.05 mmol, 62% yield).
Preparation of {2,6-[2,6-(i-Pr)2PhNsCd(CH2)]2(C5H3N)}Fe(µ-
Cl)Li(THF)3 (4). Mehod A. To a suspension of {2,6-[2,6-(i-Pr)2PhNd
C(CH3)]2(C5H3N)}FeCl2 (0.500 g, 0.82 mmol) in THF (10 mL) a
solution of LiCH2SiMe3 (0.170 g, 1.80 mmol) also in THF (5 mL)
was added at -35 °C. The color slowly changed from dark blue to
dark brownish-red, and the solids slowly went into solution with stirring.
After stirring for approximately 30 min, the solvent was evaporated
and cold hexane was added to the brown residue. The resulting
suspension was centrifuged, and a dark olive-green solution was
separated from dark precipitates (identified as 2 upon recrystallization
in ether 65% yield). The solution was placed in the freezer overnight,
and then the mother liquor was removed from a small amount of
crystallized 2 and placed back in the freezer. Dark orange crystals of
4 were grown from the hexane solution after standing at -35 °C for 2
days (0.098 g, 0.12 mmol, 15% yield). Anal. Calcd (found) for C45H66-
ClFeLiN3O3: C, 67.96 (67.93); H, 8.37 (8.31); N, 5.28 (5.24). IR (Nujol
mull, cm-1): ν 2918 (s), 2854 (s), 1644 (m), 1572 (s), 1322 (m), 1279
(s), 1249 (m), 1236 (s), 1210 (w), 1193 (w), 1154 (m), 1134 (m), 1107
(m,b), 1047 (s), 1005 (m), 972 (s), 890 (m), 848 (s,b), 822 (s), 773 (s),
727 (s) [µeff ) 5.8 µB].
Method B. {[2,6-{[2,6-(i-Pr)2C6H3]NsCd(CH2)}2(C5H3N)]Li(THF)}-
{Li(THF)4} (0.727 g, 0.85 mmol) was added to a suspension of FeCl2-
(THF)1.5 (0.200 g, 0.85 mmol) in THF (35 mL) at -35 °C. The color
of the suspension turned instantly dark orange-brown. After stirring
for approximately 30 min, the solvent was evaporated to dryness and
the resulting mass was washed with ether to remove a dark green
solution. The residual solid wase dissolved in toluene, centrifuged to a
dark greenish-brown solution, concentrated, and placed in the freezer.
Dark orange crystals of 4 were obtained upon standing at -35 °C for
2 days (0.379 g, 0.48 mmol, 56% yield).
(w), 881 (s), 850 (s,b), 813 (m,b), 773 (m), 721 (m,b), 700 (m) [µeff
5.6 µB].
)
Preparation of {2,6-[2,6-(i-Pr)2PhNdC(CH3)]2}(C5H3N)Fe(CH2-
SiMe3)2 (2). Method A. The preparation was carried out as above except
that the remaining hexane-insoluble solids were dissolved in ether. The
color became dark purple, and the mixture was centrifuged to separate
a dark purple solution from a small amount of colorless solids. The
ether solution was concentrated and kept at -35 °C for 2 days, upon
which crystals of 2 suitable for X-ray analysis were isolated (0.180 g,
0.25 mmol, 30% yield). Anal. Calcd (found) for C41H65FeN3Si2: C,
69.16(69.13); H, 9.20(9.15); N, 5.90(5.83). IR (Nujol mull, cm-1): ν
2931 (s), 2865 (s), 1645 (w), 1585 (w), 1548 (w), 1282 (s), 1259 (m),
1263 (s), 1193 (w), 1155 (m), 1135 (m), 1110 (w), 1097 (w), 973 (s),
916 (w), 894 (w), 850 (s,b), 823 (s), 775 (m), 727 (m,b) [µeff
5.6 µB].
)
Method B. A solution of LiCH2SiMe3 (0.168 g, 1.78 mmol) in ether
(5 mL) was added to a suspension of FeCl2(THF)1.5 (0.200 g, 0.85
mmol) in ether (5 mL) at -35 °C. The mixture began to turn yellow
with swirling. The mixture was kept at -35 °C with periodic swirling
for approximately 2 min. The resulting brownish-yellow suspension
was added to a cooled suspension of 2,6-[2,6-(i-Pr)2PhNdC(CH3)]2-
(C5H3N) (0.410 g, 0.85 mmol) in ether (10 mL). The color instantly
became dark purple, and the mixture was stirred for 2 min to ensure
complete mixing. The reaction was then centrifuged to remove a dark
purple solution from white precipitates. Upon concentrating and freezing
at -35 °C, dark purple block crystals of 2 were obtained (0.357 g,
0.50 mmol, 59% yield).
Preparation of [{2,6-[2,6-(i-Pr)2PhNsCd(CH2)]2(C5H3N)}FeCH2Si-
(CH3)3][Li(THF)4] (6). A solution of LiCH2SiMe3 (0.168 g, 1.78 mmol)
in ether was added to a suspension of FeCl2(THF)1.5 (0.200 g, 0.85
mmol) in ether at -35 °C. The mixture was stirred for approximately
2 min, upon which time the color gradually became darker yellow-
brown. At this time, a solution of {2-[2,6-(i-Pr)2PhNsCd(CH2)]-6-
[2,6-(i-Pr)2PhN-CCH3](C5H3N)}Li(THF) (0.476 g, 0.85 mmol) in
ether, also kept at -35 °C, was added. The color of the mixture became
dark royal blue, and stirring was continued for 1 h. The solvent was
removed in vacuo, and the residual mass was dissolved in THF. The
solution was centrifuged and separated from a small amount of dark
precipitate. Concentration and layering with hexane afforded, after
standing for 2 days, reddish-brown crystals of 6 (0.430 g, 0.47 mmol,
55% yield). Anal. Calcd (found) for C53H86SiN3FeLiO4: C, 69.18
(68.77); H, 9.42 (9.68); N, 4.57 (5.01). IR (Nujol mull, cm-1): ν 2952
(s), 2855 (s), 1561 (s), 1523 (w), 1468 (s), 1434 (m), 1377 (m), 1367
(m), 1356 (m), 1320 (m), 1309 (w), 1283 (w), 1252 (m), 1228 (w),
1209 (w), 1176 (w), 1125 (w), 1101 (w), 1083 (w), 1040 (s), 1003
Method C. A solution of LiCH2SiMe3 (0.170 g, 1.80 mmol) in THF
(5 mL) was added to a suspension of {2,6-[2,6-(i-Pr)2PhNdC(CH3)]2-
(C5H3N)}FeCl2 (0.500 g, 0.82 mmol) in 10 mL of THF at -35 °C.
The color slowly changed from dark blue to dark brownish-red, and
the solids slowly went into solution upon stirring. After stirring for
approximately 30 min, the solvent was evaporated and cold hexane
was added to the brown residue. The resulting suspension was
9
J. AM. CHEM. SOC. VOL. 127, NO. 37, 2005 13021

A R T I C L E S
Scott et al.
Table 1. Crystal Data
1
3
4
6
formula
Mw
C41H65FeN3Si2
711.99
C41H65FeN3Si2
711.99
C45H66ClFeLiN3O3
795.25
C53H86N3FeLiO4Si
920.13
crystal system
space group
a (Å)
b (Å)
c (Å)
R (deg)
â (deg)
γ (deg)
V (Å3)
monoclinic
P2(1)/n
9.8426(12)
18.821(2)
22.748(3)
90
91.684(2)
90
4212.1(9)
4
triclinic
P1h
monoclinic
P2(1)/n
11.574(3)
18.331(5)
21.787(5)
90
96.094(5)
90
4596.3(19)
4
monoclinic
P2(1)/c
15.998(19)
27.61(3)
24.90(3)
90
97.75(2)
90
10 902(22)
4
9.610(3)
10.865(3)
22.115(7)
89.135(5)
85.805(6)
66.610(5)
2113.5(12)
2
Z
radiation (KR, Å)
0.710 73
206(2)
1.124
0.445
1548
0.710 73
208(2)
1.119
0.443
772
0.710 73
208(2)
1.149
0.425
1708
0.710 73
213
1.121
0.342
4000
T (K)
D
µ
calcd (g cm-3
)
calcd (mm-1
)
F000
R, Rw
GOF
2
0.0623, 0.1399
1.036
0.0721, 0.1272
1.011
0.0783, 0.1540
1.065
0.0688, 0.1518
1.008
Table 2. Selected Bond Distances and Angles
1
3
4
5
6
Fe-N(1) ) 2.224(3)
Fe-N(1) ) 2.433(3)
Fe-N(1) ) 2.013(6)
Fe-C(34) ) 2.05
N(1)-C(2) ) 1.31
C(1)-C(2) ) 1.50
N(3)-C(8) ) 1.31
C(8)-C(9) ) 1.49
Fe-N(1) ) 2.094(4)
Fe-N(2) ) 2.015(2)
Fe-N(2) ) 2.130(3)
Fe-N(2) ) 2.095(6)
Fe-N(2) ) 2.138(4)
Fe-N(3) ) 2.160(3)
Fe-C(41) ) 2.036(3)
Fe-N(3) ) 1.960(3)
Fe-N(3) ) 2.022(6)
Fe-Cl ) 2.318(2)
Fe-N(3) ) 2.078(4)
Fe-C(38) ) 2.045(4)
Fe-C(34) ) 2.068(5)
C(1)-C(2) ) 1.499(5)
N(2)-C(3) ) 1.473(4)
C(3)-C(4) ) 1.516(4)
C(4)-C(5) ) 1.338(5)
C(5)-C(6) ) 1.429(5)
C(6)-C(7) ) 1.371(4)
C(7)-N(2) ) 1.366(4)
C(8)-C(9) ) 1.489(5)
N(1)-Fe-N(2) ) 75.94(10)
N(1)-Fe-N(3) ) 121.88(10)
N(1)-Fe-C(41) ) 115.74(13)
N(2)-Fe-N(3) ) 78.37(11)
N(2)-Fe-C(41) ) 140.44(13)
N(3)-Fe-C(41) ) 117.27(13)
C(1)-C(2) ) 1.499(6)
N(1)-C(2) ) 1.380(9)
C(1)-C(2) ) 1.362(10)
N(3)-C(8) ) 1.375(9)
C(8)-C(9) ) 1.344(10)
Cl-Li ) 2.362(13)
N(1)-Fe-N(2) ) 75.7(3)
N(1)-Fe-N(3) ) 151.3(3)
N(1)-Fe-Cl ) 104.25(19)
N(2)-Fe-N(3) ) 75.7(3)
N(2)-Fe-Cl ) 177.48(18)
N(3)-Fe-Cl ) 104.44(19)
Fe-Cl-Li ) 171.0(4)
N(1)-C(2) ) 1.360(6)
N(3)-C(8) ) 1.477(5)
N(3)-C(8) ) 1.354(6)
C(8)-C(9) ) 1.533(5)
C(1)-C(2) ) 1.337(7)
N(1)-Fe-N(2) ) 69.27(12)
N(1)-Fe-N(3) ) 145.82(13)
N(1)-Fe-C(38) ) 99.55(15)
N(2)-Fe-N(3) ) 78.49(13)
N(2)-Fe-C(38) ) 156.75(16)
N(3)-Fe-C(38) ) 114.62(16)
C(8)-C(9) ) 1.356(7)
N(1)-Fe-N(3) ) 145.67(15)
N(2)-Fe(1)-C(34) ) 155.4(2)
N(1)-Fe-C(34) ) 107.1(2)
N(2)-Fe-N(3) ) 74.08(16)
N(1)-C(2)-C(1) ) 127.6(5)
Fe-C(34)-Si(1) ) 128.6(3)
(m), 968 (w), 935 (w), 916 (w), 887 (s), 868 (s), 812 (m), 771 (m),
760 (m), 746 (m), 722 (m) [µeff ) 5.6 µB].
displacement coefficients. All hydrogen atoms were treated as idealized
contributions. All scattering factors are contained in several versions
of the SHELXTL program library, with the latest version used being
v.6.12.23 Crystallographic data and relevant bond distances and angles
are reported in Tables 1 and 2.
Calculations. Density functional calculations on model complexes
(see below) were performed with the TURBOMOLE program24 in
combination with the OPTIMIZE routine of Baker and co-workers.25
All relevant structures were fully optimized as minima or transition
states at the restricted or unrestricted b3-lyp26 level, employing the
standard SV(P) basis sets.27 All stationary points were characterized
by vibrational analyses using analytic or numerical second derivatives,
and thermal corrections (ZPE, enthalpy, entropy; 273K, 1 bar) were
calculated from these using standard formulas of statistical thermody-
namics. Improved electronic energies were calculated at the SV(P)
X-ray Crystallography. All of the compounds 1 to 6 consistently
yielded crystals that diffracted weakly, and the results presented are
the best of several trials. The crystals were mounted on thin glass fibers
using paraffin oil and cooled to the data collection temperature. Data
were collected on a Bruker AXS SMART 1k CCD diffractometer. Data
for the compounds 1 and 6 were collected with a sequence of 0.3° ω
scans at 0°, 120°, and 240° in æ. To obtain acceptable redundancy
data for compound 3, the sequence of 0.3° ω scans at 0°, 90°, 180°,
and 270° in æ was used. Initial unit cell parameters were determined
from 50 data frames collected at the different sections of the Ewald
sphere. Semiempirical absorption corrections based on equivalent
reflections were applied.22 Systematic absences in the diffraction data-
set and unit-cell parameters were consistent with monoclinic P21/n for
1, orthorhombic, triclinic P1h for 3, monoclinic P21/n for 4, and
monoclinic P21/c for 6. Solutions in centrosymmetric space groups for
all of the compounds yielded chemically reasonable and computationally
stable results of refinement. The structures were solved by direct
methods, completed with difference Fourier synthesis, and refined with
full-matrix least-squares procedures based on F2. The compound
molecules were located in common positions in the structures of all
the complexes. All non-hydrogen atoms were refined with anisotropic
(23) Sheldrick, G. M. Bruker AXS: Madison, WI, 2001.
(24) (a) Ahlrichs, R.; Baer, M.; Haeser, M.; Horn, H.; Koelmel, C. Chem. Phys.
Lett. 1989, 162, 165. (b) Treutler, O.; Ahlrichs, R. J. Chem. Phys. 1995,
102, 346. (c) Ahlrichs, R.; et al. Turbomole; version 5, January 2002.
Theoretical Chemistry Group, University of Karlsruhe.
(25) Baker, J. J. Comput. Chem. 1986, 7, 385; PQS, version 2.4; Parallel
Quantum Solutions, Fayetteville, Arkansas, USA, 2001; the Baker optimizer
is available separately from PQS upon request.
(26) (a) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785. (b) Becke,
A. D. J. Chem. Phys. 1993, 98, 1372. (c) Becke, A. D. J. Chem. Phys.
1993, 98, 5648. Note that the Turbomole functional “b3-lyp” is not identical
to the Gaussian “B3LYP” functional.
(22) Blessing, R. Acta Crystallogr. 1995, A51, 33.
(27) Schaefer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97, 2571.
9
13022 J. AM. CHEM. SOC. VOL. 127, NO. 37, 2005

Alkylation in (Bis-iminopyridinato)Fe Catalyst Reactivity
A R T I C L E S
Figure 1. Thermal ellipsoid plot for 1 with the ellipsoids drawn at the 30% probability level
Scheme 1
optimized geometries using the TZVPP basis set on all atoms;28 these
Results and Discussion
improved energies were combined with the (u)b3-lyp/sv(p) thermal
corrections to arrive at the final free energies.
The reaction of LiCH2SiMe3 with the dark-blue 2,6-[2,6-(i-
Pr)2PhNdC(CH3)]2(C5H3N)FeCl2, prepared in situ by reacting
the neutral 2,6-[2,6-(i-Pr)2PhNdC(CH3)]2(C5H3N) ligand with
FeCl2(THF)1.5, was carried out in THF at -35 °C (Scheme 1).
Dark-burgundy crystals of the divalent {2,6-[2,6-(i-Pr)2PhNd
C(CH3)]2-2-CH2SiMe3}(C5H3N)Fe(CH2SiMe3) (1) were isolated
after separation from the insoluble materials and fractional
crystallization in pure hexane. In turn, the hexane-insoluble
fraction was recrystallized from cold ether affording large, dark-
purple crystals of {2,6-[2,6-(i-Pr)2PhNdC(CH3)]2}(C5H3N)Fe-
(CH2SiMe3)2 (2) (Scheme 1). The formation of the two
complexes is a perfectly reproducible phenomenon under the
reaction conditions described in the Experimental Section. The
formulas and connectivity of both complexes have been
elucidated by X-ray crystal structures (Figures 1 and 2).
Extensive calculations were carried out on model complex L′FeMe2
(2a) and its alkyl-shifted isomers [L′-2-Me]FeMe (1a) and [L′-i-Me]-
FeMe (3a), where L′ is a model ligand bearing only Me groups at the
imine nitrogens. In addition, an isomer [L′-N-Me]FeMe (the product
of alkyl migration to the pyridine nitrogen) was included because this
might be an intermediate in alkyl transfers between metal and ligand.
Finally, the FeI complex L′FeMe (5a) was included as a model for
half of complex 5. Geometries for spin states from S ) 0 to 3 (for
1
5
FeII) or /2 to /2 (for FeI) were fully optimized for each of the above-
mentioned species. For the triplet and quintet states only, the geometries
of the full complexes 1, 2, 3, and [L-N-R]FeR (L-N-R ) pyridine
ring N-alkylated; R ) CH2SiMe3) were also fully optimized; larger
basis sets and thermal corrections were not feasible here.
(28) Schaefer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys. 1994, 100, 5829.
9
J. AM. CHEM. SOC. VOL. 127, NO. 37, 2005 13023

A R T I C L E S
Scott et al.
Figure 2. Thermal ellipsoid plot for 3 with the ellipsoids drawn at the 30% probability level.
Complex 1 consists of an Fe atom coordinated to the three
nitrogen atoms of the ligand system [Fe-N(1) ) 2.224(3) Å,
Fe-N(2) ) 2.015(2) Å, Fe-N(3) ) 2.160(3) Å] which has
been alkylated at one of the two pyridine ring ortho-positions
(Figure 1). One silylated-alkyl group is also attached to Fe [Fe-
C(41) ) 2.036(3) Å], providing the metal center with a severely
distorted tetrahedral coordination geometry [N(1)-Fe-N(2) )
75.94(10)°, N(1)-Fe-N(3) ) 121.88(10)°, N(1)-Fe-C(41) )
115.74(13)°, N(2)-Fe-N(3) ) 78.37(11)°, N(2)-Fe-C(41) )
140.44(13)°, N(3)-Fe-C(41) ) 117.27(13)°]. The consequence
of the pyridine ring alkylation and resulting loss of aromaticity
can be noticed in the ring deviation from planarity and the
expected variation in bond lengths [N(2)-C(3) ) 1.473(4) Å,
C(3)-C(4) ) 1.516(4) Å, C(4)-C(5) ) 1.338(5) Å, C(5)-
C(6) ) 1.429(5) Å, C(6)-C(7) ) 1.371(4) Å, C(7)-N(2) )
1.366(4) Å]. The other geometrical parameters of the ligand do
not show significant features and are comparable to those of
the vanadium catalyst that underwent a similar fate upon
alkylation with MAO.2
good agreement with the original observations by Brookhart
and Gibson on the FeCl2 adduct.1c-d
At first glance, the formation of complex 2 seems to be the
result of a straightforward chlorine replacement of 2,6-[2,6-(i-
Pr)2PhNdC(CH3)]2(C5H3N)FeCl2. However, the yield of 2 can
be substantially improved (up to 30% increase), and the
formation of 1 virtually was eliminated (it was no longer present
among the products isolable from the reaction mixtures) when
the reaction was carried out by inverting the order of addition
of the reagents (beginning with a low-temperature reaction
between FeCl2 and RLi, followed by a low-temperature addition
of the ligand). This suggests, as an alternative possibility, that
the formation of 2 may also be the result of the direct reaction
of RLi with free FeCl2 followed by complexation of the diimine
ligand. The presumed presence of free FeCl2 in the reaction
mixture leading to the mixture of 1 and 2 may be caused by a
ligand dissociation equilibrium from 2,6-[2,6-(i-Pr)2PhNd
C(CH3)]2(C5H3N)FeCl2, as observed for example in the chem-
istry of the lanthanides.6,7c,9 This point is further substantiated
by the observation that transmetalation reactions are readily
obtainable with this ligand system. For example, simple mixing
at room temperature of 2,6-[2,6-(i-Pr)2PhNdC(CH3)]2(C5H3N)-
FeCl2 with CoCl2 in THF affords the corresponding CoCl2
adduct and FeCl2 after 2 h of stirring at room temperature.30
Another possibility is that the very poor solubility of both
unreacted FeCl2 and the complex could make the presence of
FeCl2 unavoidable during its in situ preparation. Even in this
case, the room-temperature magnetic moment of 2 [µeff ) 5.6
µB] compares well with that of both the starting 2,6-[2,6-
(i-Pr)2PhNdC(CH3)]2(C5H3N)FeCl21c-d and 1 thus confirming
that the metal has remained in its divalent state. Finally, it should
be observed that neither 1 nor 2 can be formed by reaction of
RLi with the ligand followed by reaction with FeCl2. The lithium
alkyl Me3SiCH2Li instantly reacts at either room or low
temperature to produce the dianionic form of the ligand
Single crystals of 2 displayed identical cell parameters and
structure to those recently reported by Chirik while this work
was in progress.18b The formation of 1 is the result of two
distinct processes. The first is the alkylation of the pyridine ring
ortho position. The ortho alkylation has a major impact on the
molecular structure since the consequent anionization of the
neutral ligand requires dissociation of one of the two chlorine
atoms attached to the Fe atom with a decrease in the metal
coordination number. The second process is a straightforward
replacement of the remaining chlorine atom by the second
alkylating agent. The connectivity, as summarized in Scheme
1, unambiguously assigns the divalent state to the Fe center.
The magnetic moment of 5.6 µB measured at room temperature
with the Gouy method is in the expected range29 for the four-
unpaired-electron configuration of high-spin d6 Fe(II) and is in
(29) Cotton, F. A.; Wilkinson, G. AdVanced Inorganic Chemistry, 5th ed.; Wiley
Interscience, New York, 1988.
(30) Scott, J. Gambarotta, S. Unpublished results.
9
13024 J. AM. CHEM. SOC. VOL. 127, NO. 37, 2005

Alkylation in (Bis-iminopyridinato)Fe Catalyst Reactivity
A R T I C L E S
Scheme 2
regardless of the stoichiometric ratio.6,9 The resulting dianion
forms a different product with FeCl2 (see below).
depending on the reaction conditions, 3 might be an intermediate
in the formation of 2, which apparently is thermodynamically
more stable but kinetically less favored. No evidence could be
found that 1 may be thermally transformed to 2 in either
presence or absence of ethers.
The reaction of 2,6-[2,6-(i-Pr)2PhNdC(CH3)]2(C5H3N)FeCl2
with 2 equiv of LiCH2SiMe3 followed a surprisingly different
course when it was carried out using the analytically pure
complex resuspended in THF (Scheme 2), rather than the in-
situ-generated species. In this case, we found no evidence for
the formation of either 1 or 3. The reaction, which is completely
reproducible, afforded a substantial yield of 2 and a smaller
amount of a new dark-orange byproduct {2,6-[2,6-(i-Pr)2PhNs
Cd(CH2)]2(C5H3 N)}Fe(µ-Cl)Li(THF)3 (4) which was isolated
and purified by fractional crystallization.
We have observed that, during the formation of 1, variable
amounts of well-formed crystals of a similar color but different
shape occasionally appeared, provided that the entire workup
is carried out at a temperature kept below -20 °C. Manual
separation of these crystals under microscope afforded the new
complex {2-[2,6-(i-Pr)2PhNdC(CH3)]-6-[2,6-(i-Pr)2PhNC(CH3)-
(CH2SiMe3)](C5H3N)}Fe(CH2SiMe3) (3) (Figure 2).
The structure of 3 consists of a tetracoordinate Fe atom
surrounded by the ligand system which has undergone alkylation
similar to the case of 1 (Figure 2). However, instead of being
attached to the pyridine ring-ortho position, the alkyl is found
connected to one of the two former imine C atoms. The
coordination sphere of the metal center is defined by the three
N atoms of the alkylated ligand system and the C atom of one
alkyl group [Fe-N(1) ) 2.433(3) Å, Fe-N(2) ) 2.130(3) Å,
Fe-N(3) ) 1.960(3) Å, Fe-C(38) ) 2.045(4) Å]. The
coordination geometry of Fe is once again distorted tetrahedral
[N(1)-Fe-N(2) ) 69.27(12)°, N(1)-Fe-N(3) ) 145.82(13)°,
N(1)-Fe-C(38) ) 99.55(15)°, N(2)-Fe-N(3) ) 78.49(13)°,
N(2)-Fe-C(38) ) 156.75(16)°, N(3)-Fe-C(38) ) 114.62-
(16)°]. The alkylation of the ligand Cimine atom causes a
substantial distortion and deviation from planarity in that region
of the molecule. The quaternization of the former imine C atom
reflects in bond distances and angles as expected for the
disappearance of the conjugation [N(3)-C(8) ) 1.477(5) Å,
N(3)-C(8)-C(7) ) 108.0(3)°]. The Fe-N distance is also
shorter than the others as a result of the acquisition of anionic
character of that particular N atom.
Complex 4 can also be conveniently prepared (56% isolated
crystalline material) via direct reaction of the dianionic form
of the ligand (obtained either in situ or in crystalline form from
the high yield reaction with 2 equiv of Me3SiCH2Li at either
Similar to the case of 1 and 2, the formation of this complex
also requires the intervention of two alkyl groups. However,
one has been attached to one of the two imine C atoms while
the metal center bears the second. This implies that, similar to
the case of 1, the ligand is monoanionic and that the metal center
is in the divalent state. During attempts to recrystallize 3 we
have observed that the dark-burgundy color becomes purple
when dissolved at room temperature in ether. Crystallization
from ether afforded a good yield of 2, thus indicating that simple
dissolution of 3 in ether is sufficient for the conversion, which
requires migration of the alkyl group from the imine carbon
atom toward the metal center. In turn, this also indicates that,
Figure 3. Thermal ellipsoid plot for 4 with the ellipsoids drawn at the
30% probability level
9
J. AM. CHEM. SOC. VOL. 127, NO. 37, 2005 13025

A R T I C L E S
Scott et al.
Figure 4. Ball and stick drawing of 5.
low or room temperature) with FeCl2. In all the reactions
involving ligand deprotonation, it was always possible to detect
formation of tetramethylsilane in the reaction mixtures by GC-
MS experiments.
(5) (Figure 4). Unfortunately, no further purification was
possible due to the low yield and poor solubility of this species,
regrettably preventing full characterization. Only in one case
the quality of the crystals was just sufficient to allow a crystal
structure determination.31 Although the poor quality of the
diffraction data did not allow a full anisotropic refinement, the
connectivity and main features of this new compound were at
least demonstrated. The complex is dinuclear with the same
arrangement previously observed in Co,5 Ln,6 and Mn3 chem-
istry. The salient features are the tetracoordination of the Fe
atom, the dinuclear structure, and the nondeprotonation of the
residual carbon group attached to the imine function [C-C )
1.49 Å]. Given the presence of the alkyl group attached to the
Fe atom the oxidation state of the metal center can be regarded
as monovalent, from the formal point of view. The fact that a
monovalent Fe is generated at the expenses of the ligand system
is in line with the behavior of the Co derivatives.5
Further attempts to clarify the genesis of complex 5, and to
make this species available in larger scale, were carried out with
a number of experiments at both low and room temperature
and by changing the sequence of addition of the reagents. In
an attempt to follow synthetic pathways successful with other
metals, even reduction was probed by treating 4 with either NaH
or K(naphthalenide).3,5,6 However, only intractable materials
systematically resulted from those attempts. On the other hand,
given that to form complex 5 only one of the two imine Me
groups must undergo deprotonation, we have synthesized the
monodeprotonated lithium salt of the ligand [{[2,6-(i-Pr)2C6H5]Ns
Cd(CH2)}(C5H3N){[2,6-(i-Pr)2C6H5]NdCCH2}]Li.7b Low-tem-
perature addition of this anion to FeCl2, followed by addition
of 2 equiv of RLi afforded, again, only intractable materials.
However, the low-temperature treatment of FeCl2 with RLi,
where presumably a thermally labile FeR2 species might be
formed, followed by addition of the ligand monolithium salt
gave a new complex [{2,6-[2,6-(i-Pr)2PhNsCd(CH2)]2(C5H3
N)}Fe(CH2SiMe3)][Li(THF)] (6) in acceptable yield (Scheme
3). The magnetic moment in the solid state of this new species
[µeff ) 5.6 µB] is in line with those of the other compounds
reported in this work.
The molecule consists of a four-coordinate Fe atom (Figure
3) bonded to the three nitrogen atoms of the newly modified
ligand [Fe-N(1) ) 2.013(6) Å, Fe-N(2) ) 2.095(6) Å, Fe-
N(3) ) 2.022(6) Å] and a chlorine atom [Fe-Cl ) 2.318(2)
Å]. The chlorine atom bridges one Li cation solvated by three
molecules of THF [Cl-Li ) 2.362(13) Å]. The Fe center is
found in a slightly distorted square planar geometry [N(1)-
Fe-N(2) ) 75.7(3)°, N(2)-Fe-N(3) )75.7(3)°, N(3)-Fe-
Cl ) 104.44(19)°, Cl-Fe-N(1) ) 104.25(19)°, N(1)-Fe-N(3)
) 151.3(3)°, N(2)-Fe-Cl ) 177.48(18)°]. The ligand system
displays a very short distance between the Cimine and the C of
the substituent [C(1)-C(2) ) 1.362(10) Å, C(8)-C(9) ) 1.344-
(10) Å]. This typically indicates that the former Me group has
been deprotonated with consequent formation of a CdCH2
function, which in turn causes a parallel increase in the adjacent
CdNimine bond length [N(1)-C(2) ) 1.380(9) Å, N(3)-C(8)
) 1.375(9) Å].
The room-temperature magnetic moment of 4 in the solid
state [µeff ) 5.8 µB] compares well with the other complexes
reported in this work and is in agreement with the metal divalent
state. The connectivity of 4, as yielded by the X-ray structure
(Figure 4), clearly shows that the two former imine methyl
groups have been deprotonated to produce two ene-amido
functions (Scheme 2). There are precedents for this type of
behavior in the chemistry of this ligand system with several
metals including manganese,3 vanadium,11 lithium,9 and lan-
thanides.9 In the present case, the formation of 4 indicates the
existence of an alternative reaction pathway.
We have previously reported that, in some cases and
depending on the electronic configuration of the metal, the
formation of mono-deprotonated or doubly deprotonated species
similar to 4 leads to a curious dimerization reaction via radical
coupling of two CdCH2 moieties.3,5,6 This coupling (Scheme
2) requires electrons to be provided to the ligand system, which
can be obtained either from an external reducing agent (as in
the case of lanthanides)6 or at the expense of the CdCH2 bond,
in turn triggering a one-electron reduction of the metal center.3,5
In our preparation of complexes 2 and 4, we occasionally
observed the formation of a small, yet significant, amount of
dark crystals of a new dimeric species {[{[2,6-(i-Pr)2C6H5]Ns
Cd(CH2)}(C5H3N){[2,6-(i-Pr)2C6H5]NdC CH2}]Fe(CH2SiMe3)}2
Presumably, the reaction proceeds via formation of an
intermediate adduct of FeR2 with the monodeprotonated ligand.
The failure of this species to gives reductive dimerization toward
5, and its transformation instead to complex 6 is interesting. It
(31) Crystal data for 5: C74H106Fe2N6Si2, MW ) 1247.42, monoclinic, P21/c,
a ) 17.41 Å, b ) 12.07 Å, c ) 18.33 Å, â ) 111.1°, V ) 3851.8 Å3.
9
13026 J. AM. CHEM. SOC. VOL. 127, NO. 37, 2005

Alkylation in (Bis-iminopyridinato)Fe Catalyst Reactivity
A R T I C L E S
Scheme 3
Table 3. Polymerization Results
run
time
(min)
yield PE
activity
temp
C)
mass of
MAO
(g)
gPE/mmol
cat/hr/atm
complex
(
°
catalyst (g)
(equiv)
(
±
0.10)
1
2
4
5
23
23
23
23
23
23
0.008
0.010
0.005
0.007
0.006
0.007
500
500
500
500
500
500
30
30
30
30
30
30
2.80
3.32
4.30
3.60
3.30
3.90
509
474
1433
654
1015
678
6
L-FeCl2
high molecular weights (Figure 6, red line). In contrast, the
formally monovalent 5 produces only broadly dispersed high
molecular weight PE (Figure 6, blue line) while complex 4
shows an intermediate behavior in the sense that the two types
of PE are present in comparable amounts in the sample (Figure
6, green line). As expected, complex 6 displays catalytic activity
in the range of 4 and, more importantly, produces a similar
polymer. This suggests that complexes 1 and 2 are related in
the catalytic cycle and might lead to the same active species.
However, the PE generated by 5 is also formed by 4 and 6,
therefore indicating that somehow, during the catalytic cycle,
4 and 6 gave reduction of the metal center. The reason for the
remarkable difference between the samples of PE produced by
complexes in the two different oxidation states is unclear. It
may well be related to the different electronic configurations
of the two oxidation states, which in turn relates to the relative
stability of the intermediates involved in the chain propagation
and termination steps.32 At this stage, it is interesting to obserVe
that the monoValent species 5 is indeed catalytically Very actiVe
and that all the species generated by the alkylation of the FeCl2
adduct are catalytically actiVe, altogether concurring to the
polymerization process.
Figure 5. Thermal ellipsoid plot for 6 with the ellipsoids drawn at the
30% probability level. The THF-solvated cation has been removed for
clarity.
suggests that a residual chlorine atom attached to Fe is important
for the reductive dimerization and that the formation of the
Fe-C bond in 5 occurs after the reduction of the metal center.
The crystal structure of 6 (Figure 5) is by all means
comparable to 4 except for the lithium cation, which is solvated
by four molecules of THF and unconnected to the Fe-containing
anion. The bond distances and angles of the distorted square
planar atom are very comparable [Fe(1)-N(1) ) 2.094(4) Å,
Fe(1)-N(2) ) 2.138(4) Å, Fe(1)-N(3) ) 2.078(4) Å]. The
Fe-C distance [Fe(1)-C(34) ) 2.068(5) Å] also compares well
with those of 1, 2, 3, and 5. The coordination geometry around
the Fe center displays a significant deviation from the planarity
[C(34)-Fe(1)-N(2) ) 155.41(18)°, N(1)-Fe(1)-N(3) ) 145.67-
(15)°, N(1)-Fe(1)-N(2) ) 73.34(17)°]. The dianionic character
of the ligand system is witnessed by the short C-C distances
[C(1)-C(2) ) 1.337(7) Å, C(8)-C(9) ) 1.356(7) Å] formed
by the C atom attached to the imine functions indicating a
definite C-C double bond character. By the same token, the
C-N distances are elongated [N(1)-C(2) ) 1.360(6) Å, N(3)-
C(8) ) 1.354(6) Å].
There are interesting implications as far as the catalytic
activity of these species is concerned. These complexes display
high catalytic activity despite the tetracoordination (except 2)
and the high spin state at room temperature and atmospheric
pressure and are very similar to that of the FeCl2 adduct.
Complexes 4 and 6 are twice as active comparably (Table 3).
However, the polyethylene samples display very different
natures. Complexes 1 and 2 form identical types of polyethylene
showing a reasonably narrow distribution at low molecular
weights and a minor amount of broadly dispersed PE at very
DFT calculations on simplified models 1a-3a (bearing just
Me groups at Fe and the imine nitrogens; see Supporting
Information) indicate that, for each of the FeII species consid-
ered, the IS (triplet, S ) 1) and HS (quintet, S ) 2) states are
close in energy, with the LS (singlet, S ) 0) state always being
significantly higher (Table 4). These results agree with our
experimental observation of a paramagnetic ground state for 1
and 2. They might also be relevant to the nature of the active
species in the Fe polymerization system. The spin state of the
catalytically active species has been at the center of a theoretical
debate. The active species in the catalytic cycle (generated by
mixing LFeCl2 with MAO) has commonly been assumed to be
LFeR+. Initial theoretical work by Ziegler et al. on this system16a
indicated that the HS state was too high in energy to be
accessible for LFeR2 and LFeR+ and would anyhow not be able
(32) Schmid, R.; Ziegler, T. Organometallics 2000, 19, 2756.
9
J. AM. CHEM. SOC. VOL. 127, NO. 37, 2005 13027

A R T I C L E S
Scott et al.
Figure 6. Gel permeation chromatogram of 1 and 2 (Mn 1860, Mw ) 66 100, Ms 741 700, PD ) 35.54), of 5 (Mn ) 3250, Mw ) 219 000, Mz ) 1 207 000,
PD ) 67.38) of 4 (Mn ) 2620, Mw ) 207 300, Mz ) 2 061 000, PD ) 79.12), and of 6 (Mn ) 3250, Mw ) 219 000, Mz ) 1 207 000, PD ) 67.38).
Table 4. Calculated Relative Free Energies (kcal/mol) of L′FeMe2
and Isomers in Different Spin States
to be all very close in energy (within a few kcal/mol). The S )
1/2 state is best described as containing IS FeII antiferromag-
complex
S
)
0 (LS)
S
)
1 (IS)
S
)
2 (HS)
S ) 3
netically coupled to a ligand radical anion, i.e., analogous to
the situation in the diamagnetic formally CoI alkyls LCoR.33
L′FeCl2
L′FeMe2 (2a)
[2-Me-L′]FeMe (1a)
[i-Me-L′]FeMe (3a)
[N-Me-L′]FeMea
L′FeMe+
23.58
13.56
36.20
16.64
37.15
24.13
15.56
9.33
15.33
0.00
16.28
8.49
0.00
5.79
11.66
5.31
10.48
0.00
12.10
15.35
33.25
21.97
43.24
32.70
It is interesting to observe that the extent of simplification
used for the calculation changes the relative order of stability
of the complexes. According to the calculations on the simplified
model systems, the IS state of imine adduct 3a is the most stable
of all species considered: it is lower than the corresponding
HS state (by 6 kcal/mol) and all other isomers considered (by
4-17 kcal/mol). This disagrees with the experimental observa-
tions and indicates that steric effects play an important role here.
Indeed, geometry optimization of the full systems 1-3 changes
the order of stability completely. The estimated free energies
of all HS states are now lower than those of the corresponding
IS states (Table 5, last column). The dialkyl 2 is lowest in
energy, followed by the two-adduct 1 at 4 kcal/mol and the
imine adduct 3 at 6 kcal/mol. This revised stability order agrees
well with experiment. However, it also indicates that if the bulky
alkyls were to be exchanged for smaller Me or n-alkyl groups
(as could happen during initiation with MAO), the Fe dialkyl
might no longer be the most stable isomer, with the imine adduct
becoming at least as favorable. In view of the remarkable ease
of alkyl shift between the imine moiety and the metal atom
observed in the present work, it seems that some form of ligand
alkylation might well be important in activation of the Fe
complex.
1
3
5
S
)
/
S
)
/
S ) /
2
2
2
L′FeMe
1.25
5.03
0.00
a Adduct with Me group at pyridine nitrogen.
to perform polymerization. Therefore, it was concluded that the
LS state was probably involved in the polymerization, although
participation of the IS state could not be excluded. The results
of separate calculations by Morokuma16c and Zakharov17 instead
indicated that the LS state is highest in energy, while IS and
HS states are close. Our calculations are more in line with these
results. Given that our experiments indicate that LFeR2 is already
HS, and that our calculations indicate an even larger preference
for the IS/HS states for LFeR+ than for LFeR2, we tentatively
conclude that any LFeR+ species involved in polymerization
will not be low-spin. Of course, it remains possible that the
catalytic cycle of this system does not involve LFeR+ at all.17
In the FeII complexes studied here, IS or HS states are preferred
in which the diiminopyridine ligand appears to be electronically
innocent; i.e., the complexes truly contain FeII. In contrast, for
the formally monovalent L′FeMe 5a (model for one-half of
dimer 5) the energies of S ) 1/2, 3/2, and 5/2 states are predicted
(33) Knijnenburg, Q.; Hetterscheid, D.; Kooistra, T. M.; Budzelaar, P. H. M.
Eur. J. Inorg. Chem. 2004, 1204.
9
13028 J. AM. CHEM. SOC. VOL. 127, NO. 37, 2005

Alkylation in (Bis-iminopyridinato)Fe Catalyst Reactivity
A R T I C L E S
Table 5. Effect of Steric Bulk on Alkyl Transfer Energetics (IS and
HS States Only; kcal/mol)
in the case of other metals, seems to be confined exclusively to
the dimerization pathway via coupling of two deprotonated
imine C atoms or to a reaction carried out at room temperature.
Otherwise, the combination of a divalent organo-iron species
with this ligand system seems to be reasonably robust, once
the complex is formed, at least with the particular alkyl used in
this study. The fact that the monovalent species also displays
very high catalytic activity and produces a distinctly different
PE making it unlikely that a trivalent Fe species (possibly
formed through disproportionative pathways) is involved in the
catalytic cycle.
system
∆Emodel
∆Gmodel
∆Efull
∆G
full (est)a
LFeR2 (2) IS
HS
[2-R-L]FeR (1) IS
HS
[i-R-L]FeR (3) IS
HS
10.75
12.10
16.94
13.34
0.00
9.33
5.79
15.33
11.66
0.00
4.28
1.42
14.10
0.00
4.85
0.17
8.16
0.00
17.81
4.08
11.89
6.16
6.79
5.31
[N-R-L]FeR IS
HS
16.69
11.07
16.28
10.48
18.30
5.13
25.20
12.35
a Obtained by combining the SV(P) energies for the full system with
thermal and basis set (SV(P)fTZVPP) corrections for the model system.
Acknowledgment. This work was supported by the Natural
Science and Engineering Council of Canada (NSERC). We are
indebted with NOVA (Calgary) for generous assistance with
GPC measurements.
Conclusions
In conclusion, with this study we have unveiled the remark-
ably complex behavior of the Gibson-Brookhart Fe catalyst
in the presence of alkylating agents. The noninnocent behavior
of this ligand system, as previously observed with several other
transition metals and different alkylating agents, also dominates
in the case of the Fe analogues. Different from the other cases,
however, the mobility of the alkyl groups and the variety of
transformations affect the redox chemistry only a little. The
reduction of the metal center, which is a rather common feature
Supporting Information Available: Complete crystallo-
graphic data (CIF) for complexes 1-4 and complete ref 24c.
This material is available free of charge via the Internet at
http://pubs.acs.org.
JA054152B
9
J. AM. CHEM. SOC. VOL. 127, NO. 37, 2005 13029