Studies on the atropisomerism of Fe(II) 2,6-Bis(N-arylimino)pyridine complexes

Article abstract of DOI:10.1021/ic802271y

NMR spectra of free 2,6-bis(N-arylimino)pyridine (PDI) ligands displaying different substituents at the ortho and ortho′ positions of the two N-aryl rings indicate that they can exist in syn (meso) and anti (chiral) configurations. These interconvert in solution at room temperature, via rotation of the aryl group. The corresponding paramagnetic FeX₂(PDI) complexes exhibit the same kind of isomerism, a property that is thought to be important for their activity as R-olefin polymerization catalysts. For the first time, this has been detected by ¹H NMR and studied in solution. Although the conformational stability of the diastereoisomeric complexes varies widely (depending on the size of the substituents at the imine and the aromatic rings), a moderate degree of steric hindrance suffices to allow their chemical separation. A simple procedure is developed for the preparation of these complexes in diastereoisomerically pure form. In addition, introduction of a prochiral substituent in the pyridine ring enables positive assignment of the stereoisomers. Isomerization rate measurements of the Fe(II) complexes in solution suggest that isomerization very likely involves the dissociation of the corresponding Fe-N(imino) bond prior to the rotation of N-aryl groups. DFT calculations provide additional support to the conformational assignment as well as the dissociative isomerization mechanism.

Full text of DOI:10.1021/ic802271y

Inorg. Chem. 2009, 48, 3679-3691
Studies on the Atropisomerism of Fe(II) 2,6-Bis(N-arylimino)pyridine
Complexes†
´
Juan Ca´mpora,* M. Angeles Cartes, Antonio Rodr´ıguez-Delgado, A. Marcos Naz, Pilar Palma,
and Carmen M. Pe´rez
Instituto de InVestigaciones Qu´ımicas, CSIC, UniVersidad de SeVilla, c/ Ame´rico Vespucio, 49,
41092, SeVilla, Spain
Diego del Rio
SRI International, 333 RaVenwsood AVe., Menlo Park, California 94065
Received November 26, 2008
NMR spectra of free 2,6-bis(N-arylimino)pyridine (PDI) ligands displaying different substituents at the ortho and
ortho′ positions of the two N-aryl rings indicate that they can exist in syn (meso) and anti (chiral) configurations.
These interconvert in solution at room temperature, via rotation of the aryl group. The corresponding paramagnetic
FeX2(PDI) complexes exhibit the same kind of isomerism, a property that is thought to be important for their
1
activity as R-olefin polymerization catalysts. For the first time, this has been detected by H NMR and studied in
solution. Although the conformational stability of the diastereoisomeric complexes varies widely (depending on the
size of the substituents at the imine and the aromatic rings), a moderate degree of steric hindrance suffices to
allow their chemical separation. A simple procedure is developed for the preparation of these complexes in
diastereoisomerically pure form. In addition, introduction of a prochiral substituent in the pyridine ring enables
positive assignment of the stereoisomers. Isomerization rate measurements of the Fe(II) complexes in solution
suggest that isomerization very likely involves the dissociation of the corresponding Fe-N(imino) bond prior to the
rotation of N-aryl groups. DFT calculations provide additional support to the conformational assignment as well as
the dissociative isomerization mechanism.
Introduction
different kinds of applications: not only polymerization of
ethylene and propylene1,3 but also R-olefin oligomerization4
and polymerization of polar monomers.5 The partially
modular design of the PDI ligands allows the variation of
their stereoelectronic properties by relatively simple synthetic
means, and appropriate modifications enable the introduction
of binding sites in the molecule for catalyst immobilization.6
Due to potential practical applications, recent work in our
laboratory has focused on the synthesis and study of a
number of iron PDI complexes,7 and on the development of
methods to vary the substitution pattern of the organic
ligand.7b,c In the course of this work, we observed that the
1H NMR spectra of FeX2(PDI) displaying different substit-
uents at the ortho and ortho′ positions of the N-aryl groups
Since the discovery of their catalytic properties nearly a
decade ago,1 iron and cobalt olefin complexes of 2,6-
bis(imino)pyridine (PDI) ligands have become a paradigmatic
group among homogeneous catalysts for polymerization or
oligomerization. The high activity and the low toxicity of
the iron catalysts2 have contributed to the attraction of interest
of both industry and academic research groups. A large
number of FeX2(PDI) catalysts have been prepared, covering
†
Dedicated to Prof. Ernesto Carmona on the occasion of his 60th
birthday.
* To whom correspondence should be addressed. E-mail: campora@
iiq.csic.es.
(1) (a) Small, B. L.; Brookhart, M.; Bennett, A. A. J. Am. Chem. Soc.
1998, 120, 4049. (b) 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. (c) Small, B. L.; Brookhart,
M. J. Am. Chem. Soc. 1998, 120, 7143. (d) Gibson, V. C.; Spitzmesser,
S. K. Chem. ReV. 2003, 103, 283.
(2) (a) Chem. Eng. News 2008, July 28, 53. (b) Bart, S. C.; Lobkovsky,
E.; Chirik, P. J. J. Am. Chem. Soc. 2004, 126, 13794. (c) Correa, A.;
Garc´ıa-Manchen˜o, O.; Carsten Bolm, C. Chem. Soc. ReV. 2008, 37,
1108.
10.1021/ic802271y CCC: $40.75  2009 American Chemical Society
Inorganic Chemistry, Vol. 48, No. 8, 2009 3679
Published on Web 03/16/2009

Ca´mpora et al.
polymerization catalysts is dominated by a chain-end control
mechanism, the isotacticity of polypropylene produced by a
complex displaying the chiral anti configuration is signifi-
cantly improved as compared to that obtained with the syn
isomer. This points to the existence of a concurrent enan-
tiomorphic site stereoregulation mechanism.
A number of authors preceded us by proposing that
atropisomerism of FeX2(PDI) complexes has important
effects on their catalytic perfomance,3a,k,12 but without
providing direct evidence of its existence. Not only PDI-
based catalysts but related systems, such as group 10
R-diimine complexes, might also be influenced by analogous
sterochemical phenomena.13 A number of X-ray diffraction
structures of FeX2(PDI) complexes with unsymmetrically
substituted aryl rings have been reported. These often display
a single diastereoisomer, either the syn14 or the chiral anti.15
Sometimes, both forms coexist in a single crystalline phase,16
suggesting that these may also exist together in solution.
Therefore, it is somewhat surprising that, in spite of their
potential applications in stereoselective catalysis, there is
virtually no information on the stereochemical stability of
such isomers in solution. This might be due to the fact that,
although these paramagnetic complexes give rise to useful
1H NMR spectra, these may be difficult to interpret, and
therefore they are not systematically recorded. In this article,
we show that conformational isomerism of iron PDI com-
plexes is readily detected in their solution NMR spectra. The
stereochemical stability of these compounds varies widely,
ranging from fast exchange on the NMR time scale to
chemically separable isomers depending on the size of the
substituents. This enables the study of the isomer exchange
rates, which point to a dissociative mechanism for the mutual
interconversion process.
Figure 1. Atropisomerism in FeX2(PDI).
indicate that they exist as a mixture of geometric isomers.
This phenomenon is due to the restricted rotation of the
N-aryl groups, which can adopt one of two possible disposi-
tions, syn or anti, that is, a case of atropisomerism (Figure
1). While the syn configuration gives rise to a meso
diastereoisomer (Cs symmetry), the anti results in a racemic
mixture of enantiomeric isomers with C2 symmetry. This
traces an interesting parallelism with the well-known Zr ansa-
metallocene complexes pioneered by Ewen and Brintzinger
et al. as catalysts for stereoselective R-olefin polymeriza-
tion.8-10 Recently, we have shown that this analogy can be
successfully extended to the domain of catalysis.11 Although
the stereoselectivity of FeX2(PDI) complexes as propylene
(3) See, for example: (a) Small, B. L.; Brookhart, M. Macromolecules
1999, 32, 2120. (b) Britovsek, G. J. P.; Bruce, M.; Gibson, V. C.;
Kimberley, B. S.; Maddox, P. J.; Mastroianni, S.; McTavish, S. J.;
Redshaw, C.; Solan, G. A.; Stro¨mberg, S.; White, A. J. P.; Williams,
D. J. J. Am. Chem. Soc. 1999, 121, 8728. (c) Britovsek, G. J. P.; Baugh,
S. P. D.; Hoarau, O.; Gibson, V. C.; Wass, D. F.; White, A. J. P.;
Williams, D. J. Inorg. Chim. Acta 2003, 345, 279. (d) Armspach, D.;
Matt, D.; Peruch, F.; Lutz, P. Eur. J. Inorg. Chem. 2003, 5, 805. (e)
Chen, J.; Huang, Y.; Li, Z.; Zhang, Z.; Wei, C.; Lan, T.; Zhang, W.
J. Mol. Catal. A: Chem. 2006, 259, 133. (f) Amort, C.; Malaun, M.;
Krajete, A.; Kopacka, H.; Wurst, K.; Christ, M.; Lilge, D.; Kristen,
Marc, O.; Bildstein, B. Appl. Organomet. Chem. 2002, 16, 506. (g)
Kaul, F. A. R.; Puchta, G. T.; Frey, G. D.; Herdtweck, E.; Herrmann,
W. A. Organometallics 2007, 26, 988. (h) Boca, M.; Boca, R.;
Kikelbick, G.; Linert, W.; Svoboda, I.; Fuess, H. Inorg. Chim. Acta
2002, 36, 338. (i) Smit, T. M.; Tomov, A. K.; Gibson, V. C.; White,
A. J. P.; Williams, D. J. Inorg. Chem. 2004, 43, 6511. (j) Chen, Y.;
Chen, R.; Qian, C.; Dong, X.; Sun, J. Organometallics 2003, 22, 4312.
(k) Ionkin, A. S.; Marshall, W. J.; Adelman, D. J.; Fones, B. B.; Fish,
B. M.; Schiffhauer, M. F. Organometallics 2006, 25, 2978.
(4) (a) Small, B. L.; Brookhart, M. J. Am. Chem. Soc. 1998, 120, 7143.
(b) Britovsek, G. J. P.; Mastroianni, S.; Solan, G. A.; Baugh, S. P. D.;
Redshaw, C.; Gibson, V. C.; White, A. J. P.; Williams, D. J.; Elsegood,
M. R. J. Chem.sEur. J. 2000, 6, 2221. (c) Small, A, B. L.; Marcucci,
J. Organometallics 2001, 20, 5738. (d) Chen, Y.; Qian, C.; Sun, J.
Organometallics 2003, 22, 1231. (e) Bianchini, C.; Giambastiani, G.;
Rios Guerrero, I.; Meli, A.; Passaglia, E.; Gragnoli, T. Organometallics
2004, 23, 6087. (f) Tellmann, K. P.; Gibson, V. C.; White, A. J. P.;
Williams, D. J. Organometallics 2005, 24, 280.
Results and Discussion
Synthesis of Ligands and Complexes. Ligands and
complexes employed in this work are shown in Chart 1. Most
of the ligands were prepared by standard procedures,
involving the condensation of 2,6-diacylpyridines or 2,6-
diformylpyridines with the corresponding anilines under
azeotropic water-removal conditions. Ligand L9 was pre-
pared from L3 using a selective alkylation procedure recently
reported by us.7b
(5) Castro, P. M.; Lappalainen, K.; Ahlgre´n, M.; Leskela¨, M.; Repo, T.
J. Polym. Sci., Part A: Polym. Chem. 2003, 41, 1380.
Iron complexes were prepared by reacting the correspond-
ing ligand with a suitable metal precursor (FeCl2 · 4H2O,
FeCl2(dme), or anhydrous FeBr2; dme ) dimethoxyethane),
(6) (a) Kaul, F. A. R.; Puchta, G. T.; Schneider, H.; Bielert, F.; Mihalios,
D.; Herrmann, W. A. Organometallics 2002, 21, 74. (b) Kim, I.; Han,
B. H.; Ha, C.-S.; Kim, J.-K.; Suh, H. Macromolecules 2003, 36, 6689.
(c) Jin., G.-X.; Zhang, D. J. Polym. Sci. Polym. Chem. 2004, 42, 1018.
´
(7) (a) Ca´mpora, J.; Naz, A. M.; Palma, P.; Alvarez, E.; Reyes, M. L.
Organometallics 2005, 24, 4878. (b) Ca´mpora, J.; Pe´rez, C. M.;
(12) Pellechia, C.; Pappalardo, D.; Mazzeo, M. Macromol. Rapid Commun.
1998, 19, 651.
´
Rodr´ıguez-Delgado, A.; Naz, A. M.; Palma, P.; Alvarez, E. Organo-
metallics 2007, 26, 2104. (c) Ca´mpora, J.; Naz, A. M.; Rodr´ıguez-
(13) (a) Gates, D. P.; Svejda, S. A.; On˜ate, E.; Killian, C. M.; Johnson,
L. K.; White, P. S.; Brookhart, M. Macromolecules 2000, 33, 2320.
(b) Pappalardo, D.; Mazzeo, M.; Antinucci, S.; Pellechia, C.
Macromolecules 2000, 33, 9483. (c) Zou, H.; Ming Zhu, F.; Wu,
Q.; Yan al, J.; An Lin, S. J. Polym. Sci., Part A: Polym. Chem.
2005, 43, 1325.
(14) Crystal structures of FePDI complexes displaying syn configuration
(Cambridge Structural Database codes): FEQLUJ, FEQMAQ4f;
MAZWEP, MAZWIT, MAZWOZ4b; and XALLEB3a.
(15) Crystal structures of FePDI complexes displaying anti configuration
(CSD codes): EKINAN4d, QELJAT, QELJEX, and QELJIB3k.
(16) Disordered FePDI structures (CSD codes): EMEHOT3a;
QELJUN3k.
´
Delgado, A.; Alvarez, E.; Tritto, I.; Boggioni, L. Eur. J. Inorg. Chem.
2008, 1871.
(8) Razavi, A.; Thewalt, U. J. Organomet. Chem. 1993, 445, 111.
(9) (a) Ewen, J. A. J. Am. Chem. Soc. 1984, 106, 6355. (b) Ewen, J. A.
In Catalytic Polymerisation of Olefins, Studies in Surface Science and
Catalysis; Keii, T., Soga, K., Eds.; Elsevier: New York, 1986; p 271.
(10) (a) Kaminsky, W.; Kulper, K.; Brintzinger, H.-H.; Wild, F. R. W. P.
Angew. Chem., Int. Ed. Engl. 1985, 24, 507. (b) Ro¨ll, W.; Brintzinger,
H.-H.; Rieger, B.; Zolk, R. Angew. Chem., Int. Ed. Engl. 1990, 29,
279.
(11) Rodr´ıguez-Delgado, A.; Ca´mpora, J.; Naz, A. M.; Palma, P.; Reyes,
M. L. Chem. Commun. 2008, 5230.
3680 Inorganic Chemistry, Vol. 48, No. 8, 2009

Fe(II) 2,6-Bis(N-arylimino)pyridine Complexes
Chart 1
Scheme 1
usually in THF. By selecting an appropriate combination of
solvent and precursor, the products precipitated directly from
the reaction mixture. All of them are dark-blue paramagnetic
complexes, with effective magnetic moments in the range
of 5.0-5.8 µB in the solid state (magnetic susceptibility bal-
ance), similar to the values reported by other authors.3-5,12-16
The methodology used for ligand synthesis did not provide
satisfactory yields for L1, probably due to the low nucleo-
philicity of 2-chloro-4,6-dimethylaniline. In order to circum-
vent this problem, we avoided using this ligand in the
synthesis of its Fe complex, C1. The alternative procedure
(Scheme 1, top) involved the reaction of the above-mentioned
aniline with the complex [FeCl2(2,6-diacetylpyridine)]. The
latter was obtained as a magenta paramagnetic solid when a
suspension of FeCl2(dme) was treated with 2,6-diacetylpy-
ridine. The condensation reaction took place in refluxing
Inorganic Chemistry, Vol. 48, No. 8, 2009 3681

Ca´mpora et al.
Chart 2
ethanol, cleanly affording C1 in high yield. As shown in
Scheme 1 (bottom), a similar approach also proved conve-
nient for the synthesis of complexes displaying two different
N-aryl groups, like C8. Many complexes of this kind have
been reported in the literature, but the synthesis of the
corresponding ligands by consecutive condensation reactions
is often unselective and their isolation in pure form can be
difficult. Thus, we prepared L8 by the condensation of 2,6-
diacetylpyridine with 1 equiv of 2,4,6-trimethylaniline to
afford the corresponding monoimine, which was subse-
quently reacted with 2,6-diisopropyl aniline. While the
selectivity of the first step could be controlled by using an
excess of the diketone, the second one afforded a mixture
of the desired ligand and the two symmetric diimines. The
alternative route to complex C8 involved, first, the prepara-
tion of the corresponding Fe complex of the monoimine.
Similar N,N,O chelated complexes have been recently
reported by Herrmann et al.3g The reaction of this complex
with 2-isopropropylaniline under mild conditions selectively
afforded the desired product.
Figure 2. Experimental (left) and simulated (right) methyl region of the
1H NMR signals (δ 1.90 ppm, o-Me; 1.05-1.15 ppm, iPr) of L3 at different
temperatures. Temperatures and exchange rate constants are shown at the
right-most part of the figure.
doublets split each of them in a pair of equally intense
signals, revealing that this compound exists as a 1:1 mixture
of isomers in solution. The magnitude of the splitting is very
small (just a few hertz), and it is likely that in some solvents
like CDCl3 it simply cannot be resolved. The fact that this
splitting only becomes noticeable for signals belonging to
the ortho substituents clearly points to the rotation of the
aryl groups as the source of isomerism. The 1H NMR
spectrum of L6 in C6D6 or toluene-d8 displays similar
splitting of the ortho Me and t-Bu signals. In this case, we
noticed that one of the two isomers prevails in freshly
prepared samples, but the intensity ratio of each pair of
signals gradually evolves until the 1:1 equilibrium ratio is
attained. Measurements of the isomerization rate at different
temperatures between 288 and 343 K give a free activation
energy ∆G‡ of 20.2(2) kcal mol-1 at 298 K. In contrast, the
spectra of L3 always display the same 1:1 equilibrium ratio,
preventing direct measurement of the isomerization rate.
However, its spectrum shows some line broadening above
room temperature, as a consequence of the rapid rotation of
the N-aryl groups. As can be seen in Figure 2, individual
signals of the isomers coalesce into a single set of signals at
378 K. Above this temperature, the iPr methyl signals
continue to broaden and merge into a single broad resonance.
The fast exchange limit, at which the diastereotopic character
of the isopropyl methyl groups would be fully lost, could
not be reached, but at the highest temperature studied (443
Restricted Aryl Rotation in Free PDI Ligands. Before
examining the isomerism of Fe-PDI complexes, it is
important first to consider the ligands themselves. Although
they are not affected by any of the complications associated
with paramagnetic compounds, to the best of our knowledge,
there are no literature references regarding their atropisom-
erism. Brookhart and Small observed that the rotation of the
aryl groups is hindered by bulky substituents in the ortho
positions.3a They mentioned that in the H and 13C NMR
1
spectra of derivatives I and II (Chart 2) the Me’s of the iPr
groups have diastereotopic character, each of them giving
rise to one distinct signal. This indicates that the rotation of
the 2,6-diisopropylphenyl groups is slow on the NMR time
scale. The same was observed for the unsymmetrically
substituted ligand III, albeit the presence of only one set of
iPr signals indicates that the 2-t-butylphenyl moiety rotates
freely. These observations exemplify how isopropyl groups
can be used to probe the rotation of aryl groups, even when
the molecule is not susceptible to rotational isomerism.
Although the 1H NMR spectrum of L3 in CDCl3 contains
only one set of signals, it displays two distinct doublets for
the isopropyl methyls, which indicate that the rotation of
o-isopropyl-o′-methylphenyl rings is slow at room temper-
ature. According to Brookhart and Small, this observation
suggests that the compound exists as a single isomer in
solution. However, we found that, on changing the solvent
to C6D6, the o-methyl singlet and the two isopropyl methyl
3682 Inorganic Chemistry, Vol. 48, No. 8, 2009

Fe(II) 2,6-Bis(N-arylimino)pyridine Complexes
Table 1. Activation Parameters for Isomerization Processes
compound
R1
R2
R4
X
T range (K)
equil. ratio
∆H‡ kcal/mol
∆S‡ cal/mol K
∆G‡
kcal/mol
(298K)
C3
C4
C6
L3a
L6
Me
Me
Me
Me
Me
iPr
iPr
Me
nBu
Me
Me
Me
Cl
Cl
Cl
296-326
296-326
263-306
298-413
288-319
1:1
1:1
2:8
1:1
1:1
25.5(8)
25(3)
24(2)
14(2)
16(2)
+5(3)
+5(10)
+5(6)
-21(4)
-13(7)
24.0(1)
23.7(2)
22.3(2)
20.7(1)
20.2(2)
tBu
iPr
tBu
a From NMR spectral simulation data.
K), there are clear signs of such coalescence. Also shown in
Figure 2 is a full line shape simulation of the spectra, which
provides the isomer exchange rates at each temperature. The
activation parameters deduced from these data (see below,
Table 1) are remarkably similar to those calculated for L6,
in spite of the different techniques employed in the measure-
ments. Thus, the value of the activation barrier, ∆G‡298, is
the same for the two compounds within the experimental
error.
Figure 3. Isomerism due to the configuration of the metal center (not
observed).
1
Atropisomerism of Fe-PDI Complexes. The H NMR
spectra of the paramagnetic FeX2(PDI) complexes C1-C9
have been fully assigned on the basis of signal intensities
and bandwidth. In spite of their wide breadth, and the rather
unpredictable signal positions, they bear reasonable analogy,
which facilitates the assignment task. Chemical shifts display
the expected Curie dependency on T-1 and are reported at
20 °C unless otherwise stated. The resonances of the PDI
pyridine and imine fragments display large isotropic shifts,
due to the efficient spin delocalization along the ligand π
orbitals. Thus, signals corresponding to the pyridine H3 and
H4 atoms are invariably found in the low-field extreme of
the spectrum (30-90 ppm), while the acetaldimine methyl
resonates in the high-field side (-20 to -35 ppm). Since
the orthogonal disposition of the ligand plane and the N-aryl
groups hamper spin delocalization, the corresponding signals
tend to appear closer to the diamagnetic region.
NMR spectra of complexes having two different ortho aryl
substituents usually display two complete sets of signals,
indicating that they exist as isomeric mixtures. In contrast
with the free ligands, the large isotropic shifts characteristic
of paramagnetic complexes facilitate the observation of the
splitting of most of their signals. However, the separation
of each signal pair is not large, in general less than a few
parts per million, which prevents the observation of stere-
ochemically significant trends. Isomeric splitting is never
seen in complexes having one or two symmetrically substi-
tuted aryl groups. For example, the spectrum of the “mixed”
complex C8, with a symmetric mesityl group, shows a single
set of signals. The overall number of methyl signals (seven,
including the two R-imino groups) can only be accounted
for if the rotation of both aryl groups is slow on the NMR
time scale. This rules out the square-pyramidal configuration
of the iron center as the source of the observed isomerism,
since this would lead to two possible isomers, as shown in
Figure 3.17
3, three qualitative situations can be distinguished. The first
is exemplified by complexes C3, C3′, C4, and C6, possessing
relatively bulky substituents of the size of or larger than
methyl at both aryl ortho positions and at the imine R carbon.
Their interconversion is slow in solution at room temperature,
and they can be fully or partially separated, as discussed
below. The second group is formed by complexes having
one ortho substituent smaller than a methyl group but larger
than a hydrogen atom, for example, Cl (C1) or a fused benzo
ring (C7). The 1H-NMR spectra of these complexes display
two distinct sets of signals, but the isomer exchange rate is
too fast to allow their chemical separation. As the equilibrium
situation is rapidly attained in solution, the NMR signals of
the isomers always show the same intensity ratio, generally
1:1. Complexes that exhibit free rotation of the aryl groups
at room temperature compose the third group. This is the
case for C2 and C5, with a hydrogen atom at either the imine
R-carbon or at one of the aryl ortho positions. In these
relatively unhindered molecules, free rotation of the N-aryl
groups is responsible for the rapid exchange between the
syn and anti configurations at room temperature. This is
shown by the apparent loss of the diastereotopic character
of the isopropyl methyl groups in C2, whose spectrum
displays a single methyl resonance for 12 H at δ 1.12 ppm.
1
The room-temperature H NMR spectrum of C5 provides
no indication of whether the compound exists as two rapidly
exchanging isomers or as a single, highly favored configu-
ration. Seeking to clarify this point, we recorded the spectrum
of C5 at lower temperatures. A second set of signals with
ca. 10% of the intensity of those of the major isomer can be
observed below 263 K, the slow exchange limit being
reached at 223 K. Although the paramagnetic line broadening
and the low concentration of the minor isomer complicate
the use of spectral line shape simulation to deduce the
exchange rate constants, an estimation provides ∆G‡ values
of ca. 13 kcal mol-1 between 298 and 263 K. The size of
this barrier indicates that the imine R-methyl still poses some
hindrance to the rotation of the aryl group, even when one
of the ortho positions remains unsubstituted. Note the
difference with C2: in this case, apparent equivalent isopro-
As expected for atropisomerism, steric effects control the
conformational stability of the isomers. As shown in Chart
(17) The iron center is expected to invert its configuration very rapidly at
room temperature. An energy barrier of ca. 10 kcal mol-1 has been
measured for the organometallic complex Fe(CH2SiMe3)2(PDImes
)
(PDImes ) 2,6-bis(N-mesitylacetaldimino)pyridine (see ref 7a).
Inorganic Chemistry, Vol. 48, No. 8, 2009 3683

Ca´mpora et al.
Chart 3
pyl methyl signals implies essentially unrestricted rotation
of the aryl groups.18
with diethyl ether and to use the more reactive reagent
FeCl2(dme), instead of simple iron salts. The NMR spectrum
in CD2Cl2 of the precipitate displays two isomers in a 2.3:1
ratio. When this solution is left at room temperature, the
isomer ratio gradually changes, eventually inverting its value
to a 1:4 equilibrium composition. This indicates a slight but
noticeable thermodynamic preference (ca. 0.8 kcal mol-1)
for one of the isomers, in contrast with the essentially
degenerate equilibria observed in the preceding complexes.
In spite of our efforts, we have been unable to grow quality
crystals of the former complexes for X-ray diffraction studies.
However, NMR spectra of microcrystalline samples obtained
by the slow evaporation of C3 solutions in CH2Cl2 indicated
the presence of a single species (C3s). Interestingly, this is
not the same isomer that precipitates during the synthesis
(C3a), but the second one observed after solution isomer-
ization (see Figure 4). This behavior indicates that C3s is
less soluble in CH2Cl2 and crystallizes selectively while the
slow exchange process maintains the equilibrium in solution.
Since the two isomers are equally stable in solution, but one
is appreciably less soluble than the other, when a solid C3a
and C3s mixture is stirred in a hexane/dichloromethane
suspension (where they are only sparingly soluble), a
continuous dissolution-crystallization process gradually
shifts the isomer ratio in the solid phase until it is turned
into essentially pure C3s (spectrum C, Figure 4). Reasoning
that the cause of the selective formation of C3a during
synthesis could also be a solubility effect, we stirred a
suspension of C3a and C3s in THF for 24 h, and the initial
composition of the solid (i.e., > 90% C3a) was restored.
These properties provide a simple and very convenient
method to pertub isomeric mixtures of this compound,
shifting their composition toward one or the other isomer
Complexes with the highest degree of steric hindrance,
1
such as C3, show interesting solubility behavior. H NMR
spectra of fresh samples of this product in CD2Cl2, prepared
with the solid that precipitated from the reaction mixtures
(THF), display one species as the prevalent (>90%) or even
the only component. However, when these samples are left
at room temperature over several hours or are gently warmed
(40 °C), less intense signals grow at the expense of the other
until a 1:1 ratio is attained. This is illustrated in Figure 4
(spectra A and B). Rapid evaporation of these samples to
dryness does not restore the initial situation, but affords solid
mixtures which retain the solution isomer ratio. Analytical
and magnetic data gathered for these solid mixtures are
identical to those of the original sample, confirming that the
two species are indeed isomers. These observations indicate
that, during its synthesis, C3 selectively precipitates es-
sentially as a single isomer, which becomes trapped in the
solid state. Since in solution this isomer slowly equilibrates
with the second one, and both are equally stable, the
insolubility of the product in the reaction solvent (THF) is
essential in order to generate isomerically pure samples.
Complexes C3′ and C4 behave similarly and they can be
prepared as solid samples containing >90% of a single
isomer, using FeBr2 or hydrated iron chloride as precursors
in THF. Equilibration of these compounds in CD2Cl2 at room
temperature also affords 1:1 isomer mixtures. In the case of
the more soluble complex C6, it is necessary to replace THF
(18) The separation of the two diastereotopic isopropyl Me signals is usually
very large in these paragmagnetic complexes, for example, 10.5 ppm
in C8. Coalescence of such signals involves very high exchange rates
(ca. 5 × 104 s-1 in a 400 MHz spectrometer), i. e., an energy barrier
below 10 kcal mol-1
.
3684 Inorganic Chemistry, Vol. 48, No. 8, 2009

Fe(II) 2,6-Bis(N-arylimino)pyridine Complexes
1
Figure 4. H NMR spectra of complex C3 in CD2Cl2 (selected signals assigned; CD2Cl2 residual signal marked with an asterisk). (A) Freshly prepared
sample directly from the material precipitated during the synthesis (C3a). (B) Spectrum of the same sample, after 24 h at room temperature, showing a 1:1
isomeric mixture. (C) Freshly prepared sample from recrystallized material (C3s).
depending on the solvent used. This procedure can be applied
to compounds C3′ and C4, with the same results. The higher
solubility of C6 prevented the carrying out of this procedure
in THF, but stirring in hexane/CH2Cl2 shifted the C6a/C6s
ratio of a solid sample from 1:4 (solution equilibrium value)
to 1:10.
The availability of isomerically pure or enriched samples
of some FeX2(PDI) complexes, and their relatively slow
interconversion in solution, renders this system suitable for
kinetic study. Hence, we monitored the equilibration of the
1
“a” isomers in CD2Cl2 solution with H NMR at different
temperatures. In all cases, first-order relaxation kinetics were
observed. Figure 5 shows a collective Eyring plot of the
direct rate constants, while values for the corresponding
activation parameters are collected in Table 1. These are
compared to the isomer exchange rate data for ligands L3
Figure 5. Eyring plot for isomerization processes in solution.
and L6. Although the limited range of temperatures acces-
sible for these kinetic studies by NMR (30-40 K) limits
the accuracy of their values, they have some useful mecha-
nistic meaning at a qualitative level (as discussed later). At
room temperature (298 K), the values of ∆G‡ for the
complexes are similar, within a 3 kcal mol-1 range, and 1-5
kcal mol-1 higher than those found for the free ligands L3
and L6 (ca. 20 kcal mol-1).
Note that when using the a/s terminology we refer only
to solubility properties, and so far we have made no
assumptions on their actual configurations, since NMR
spectra provide no information in this regard. Nevertheless,
it seems reasonable to assume that there is a relationship
between solubility properties and the structure of the isomers,
especially for complexes with such similar behavior as C3,
C3′, and C4. Supporting this assumption, it is often found
that FeX2(PDI) complexes susceptible to isomerism display
the same configuration in their crystal structures if the
substitution pattern of the N-aryl groups is not very different
and the samples were crystallized from similar solvents.
Thus, complexes containing 2-Me, 2-CF3, and 2-Ph groups
on the N-aryl rings are invariably found in the syn
configuration,3a,b,4f while others with large aryl groups in
the meta position are prone to crystallization as the anti
isomers.3k Confident that positive structural assignment of
the isomeric configurations could be extended at least to the
structurally related congeners, we synthesized complex C9,
Inorganic Chemistry, Vol. 48, No. 8, 2009 3685

Ca´mpora et al.
Figure 6. Expansion of the 1H NMR spectrum displaying the neophyl methyl signals (A) of a freshly prepared solution of C9 in CD2Cl2 and (B) the same
sample after gentle warming.
an analogue of C3 displaying a prochiral neophyl group at
position 4 of the pyridine ring. We have recently reported
that alkyl substitution of this remote position has a negligible
influence on the spectroscopic or catalytic properties of
FeX2(PDI) complexes.7c Hence, it is not expected to have
any effect on the conformational preference of the N-aryl
groups. The prochiral group reveals the presence or absence
of a symmetry mirror in the molecule, as the methyl groups
and methylene hydrogen atoms are chemically equivalent
in the symmetric syn isomer but become inequivalent in the
anti. Since C9 is quite soluble even in diethyl ether, we found
it difficult to prepare isomerically pure samples, but careful
recrystallization from CH2Cl2/hexane produced essentially
one isomer (C9s), and only a small amount (≈10%) of the
low-field signals are the same for the C3 and C9 mixtures.
For example, the two m-H’s of the C3a N-aryl groups give
rise to two very close signals at δ 15.10 and 15.96 ppm. In
the mixture spectrum, the two signals corresponding to C3s
(δ 16.59, 14.37 ppm) are flanking those of C3a, giving rise
to a conspicuous feature (see Figure 4). The same is observed
in the C9 system, where the two m-H signals of the anti
diastereoisomer (δ 16.14, 14.86 ppm) are also flanked by
the more separated ones of the syn (δ 16.29, 14.76 ppm).
Similar relationships are established for the signals corre-
sponding to the imino methyls, the isopropyl CH and Me
groups, and the H3 atoms of the pyridine ring. Such subtle
spectral analogies cannot be carried to structurally more
distant complexes like C3′ and C4, but it remains reason-
able extending the anti, syn configuration assignments to the
corresponding a and s isomers, respectively. The case of
complex C6 is less certain, since this compound displays a
slightly different behavior. However, theoretical assessment
of the relative energy of the syn and anti configurations
suggests that the former relationships hold for C6 (see
below).
Mechanism of the Isomerization Process. Assuming that
the source of isomerism is the rotation of the aromatic rings
attached to the imine functionality, the energy barrier of the
isomer interconversion is caused by the encounter of the
R-imino group and the ortho substituents along the least
hindered rotation path. Restrictions to rotation essentially
disappear in bis(formaldimino)pyridine complexes (R-imine
) H), but for complexes with an R-Me group in the imine
functionality, the rotation barrier is largely determined by
the size of the smaller N-aryl ortho substituent. For this
reason the energy barrier is very similar in all cases where
this substituent is methyl, independently of the size of the
1
second isomer (C9a). As expected, the H spectrum of C9
resembles that of C3, and this facilitates the identification
of several sharp signals (some with almost typically dia-
magnetic linewidths) corresponding to the remote neophyl
group. A narrow signal (∆1/2 ) 8 Hz, relative intensity 6H)
at δ 2.98 ppm can be unambiguously assigned to the neophyl
CMe2 group, indicating that C9s has the symmetric syn
structure. When this sample is warmed at 40 °C for 2 h, the
expected isomerization process ensues, until an approxi-
mately 1:1 mixture is obtained (Figure 6). These changes
are clearly observed in Figure 6, an expansion of the central
1
region of the H NMR spectrum of C9, before and after
isomer equilibration. The splitting of the CMe2 signal in two
at δ 3.04 and 2.69 ppm clearly discloses the asymmetry of
the C9a diastereoisomer. Comparing the spectra of the C9
and C3 (as 1:1 diasteromer mixtures) reveals some common
trends, which further support the identification of the isomers.
In general, whenever the analogous signals of both spectra
are well-resolved a/s pairs, the identities of the high- and
3686 Inorganic Chemistry, Vol. 48, No. 8, 2009

Fe(II) 2,6-Bis(N-arylimino)pyridine Complexes
Scheme 2
other ortho group. Surprisingly, the energy barrier does not
increase appreciably for compound C4, which displays n-Bu
as an R-imino substituent. This indicates that the steric
hindrance posed by the primary alkyl in its preferred
configuration is almost identical to that caused by a methyl.
The interconversion of syn and anti isomers of the
FeX2(PDI) at room temperature is somewhat surprising when
their rigid molecular structure is compared to those of organic
compounds that give rise to very stable atropisomers, such
as biphenyl derivatives.19 In spite of the apparently rigid
structure of the complexes, data in Table 1 show that
coordination of the ligands L3 or L6 raises the rotation
barrier only 1-5 kcal mol-1. This strongly suggests that aryl
group rotation is preceded by the dissociation of the
corresponding Fe-N(imine) bond. This dissociative isomer-
ization mechanism is shown in Scheme 2, along with a
qualitative energy profile. Assuming that the Fe-N dissocia-
tion step is relatively rapid, and the configuration inversion
rate is determined by the rotation of the aryl fragment, the
overall energy barrier, ∆G‡, can be composed of two terms:
the energy required to dissociate the Fe-N bond, ∆G1, and
surprising for a dissociative mechanism. In contrast, activa-
tion entropies are negative for the free ligands. Negative ∆S‡
values are often associated with the rotation of rigid
molecular fragments,20 and this has been explained as a result
of the restriction of some vibrational modes as the molecule
approaches an encumbered transition state.21 As for the free
activation energy, activation entropies of complexes are the
sum of the dissociation step thermodynamic entropy (∆S1)
and the activation parameter linked to the rotation transition
step (∆S‡ ). Assuming that, as found for the free ligands,
2
∆S‡ is negative, it seems plausible that its approximate
2
cancelation by the positive ∆S1 term leads to near-zero
activation entropy values measured for the isomerization of
C3, C4, and C6.
Computational Models of the syn-anti Isomerism.
With the aim of gaining further understanding of the
isomerization mechanism and, at the same time, of providing
some additional support to our ideas, we performed DFT
calculations on molecular models of the Fe(II)PDI com-
plexes, using the hybrid functional B3LYP. Geometries were
optimized using triple-ꢀ basis functions on the metal center
and the atoms directly bound to it (N and halogen), and
double-ꢀ for C and H. Triple-ꢀ functions were applied to all
atoms for energy calculations.
First, we addressed the feasibility of the dissociative step
that, according to our proposal, facilitates the rotation
process. We decided to investigate the simplified model
shown in Figure 7, in order to study the influence of the
R-imine substituent (R) and the halogen bound to iron (X).
Calculations show that the dissociation energy of the
Fe-N bond of VI to afford the tetracoordinated species VII
is between 0.6 and 6.3 kcal mol-1 (Table 2). This is in
excellent agreement with the 1-5 kcal mol-1 range deduced
for the experimental system. When R ) H, the dissociated
imine group becomes coplanar with the heterocyclic ring,
rendering VII almost as stable as the pentacoordinated
the activation energy for the rotation step itself, ∆G‡ . The
2
second term is similar for C3, C6, and also C4, considering
the likeliness of their experimental ∆G‡ values.
Since the rotation of the aryl group takes place in the free
imine arm of intermediate V, the ∆G‡ term is expected to
2
be similar to the isomerization energy barrier measured for
the free ligands L3 and L6, that is, 20-21 kcal mol-1.
Therefore, the Fe-N dissociation energies, ∆G1, can be
estimated as the difference between the rotation barrier of
the complexes and that of the free ligands, that is, between
1 and 5 kcal mol-1. The value of ∆G1 corresponding to the
C6 complex would be the lowest (ca. 1 kcal mol-1). A
plausible explanation for this is that the dissociative step is
favored by the strain caused by the bulk of the t-Bu
substituents.
At first glance, the nearly zero activation entropies
measured for Fe(PDI) complexes might appear somewhat
(20) Some examples: (a) Lafon, O.; Lesot, P.; Fan, C. A.; Kagan, H. B.
Chem.sEur. J. 2006, 12, 3772. (b) Pontes, R. M.; Basso, E. A.; dos
Santos, F. P. J. Org. Chem. 2007, 72, 1901. (c) Matchett, S. A.; Zhang,
G.; Frattarelli, D. Organometallics 2004, 23, 5440. (d) Belyakov, P. A.;
Shastin, A. V.; Strelenko, Yu. A. Russ. Chem. Bull., Int. Ed. 2001,
50, 2473. (e) Albe´niz, A. C.; Calle, V.; Espinet, P.; Go´mez, S. Inorg.
Chem. 2001, 40, 4211.
(19) (a) Carey, A. F.; Sundberg, R. J. AdVanced Organic Chemistry, 3rd
Edition. Part A; Plenum Press: New York, 1990; pp 97-98. (b)
Preliminary assessment of the isomerization mechanism using a
molecular mechanics model indicated that simple aryl rotation is
characterized by high-energy barriers for the acetaldimino deriva-
tives.
(21) Li, J. H.; Allinger, N. L. J. Am. Chem. Soc. 1989, 111, 8566.
Inorganic Chemistry, Vol. 48, No. 8, 2009 3687

Ca´mpora et al.
Figure 7. Simplified model used to study the influence of the R-imine
substituent (R) and the halogen bound to iron (X).
Table 2. Calculated Energies and Imine-Pyridine Dihedral Angles for
the Fe-N Dissociation Step
dihedrala (deg)
energyb (kcal mol-1
)
R
X
TS
VII
TS
VII
H
Me
Me
Cl
Cl
Br
86.3
92.8
90.1
180.0
149.9
149.1
10.2
7.1
8.3
0.6
4.9
6.3
a Npy-C-CdNim dihedral angle. b Relative to VI.
complex VI. However, where R ) CH3, the significant steric
bulk of the R-Me prevents this from happening, increasing
the energy of VII relative to VI. This increase is not too
large but becomes significant for X ) Br. It must be noted
that the calculation could be somewhat overestimating the
Fe-N dissociation energy, since the methyl substitution at
nitrogen renders the imine a better ligand than in the real
system.
Figure 8. Optimized structures of three molecular models displaying
different substituents at the ortho positions of the aryl groups: VIII, i-Pr
and Me; IX, t-Bu and Me; and X, t-Bu and H. The syn isomer is shown at
the right and the anti at the left.
Table 3. Calculated and Experimental Energy Difference of Calculated
Models and Complexes
In the calculations, the transition state corresponds to an
approximately orthogonal disposition of the free imine arm
with regard to the ligand plane. For R ) Me, its energy is
7-8 kcal mol-1 (Table 2), clearly lower than the experi-
mental ∆G‡ of the real system, which confirms that the
dissociative step is rapid in relation to the rotation of the
aryl groups. However, for R ) H, the Fe-N dissociation
barrier is of the same order or slightly higher than that of
the facile aryl rotation. This suggests that for formaldimino
complexes, isomerization could take place without a dis-
sociative step.
Next, we optimized the structures of three molecular
models displaying full substitution paterns in their syn and
anti configurations. Models VIII and X reproduce the
structure of complexes C3 and C5, while model IX is
analogous to C6 but without the p-tBu substituent in the aryl
rings. The geometries of the six molecules are shown in
Figure 8, and Table 3 lists the computed relative energies
for the isomers together with those deduced from solution
NMR data. As can be seen in Figure 8, the iron centers tend
to display square-pyramidal coordination, which is usually
preferred over a trigonal-bipyramidal environment in this
kind of complexes. Although the energy differences between
the syn and anti configurations are at the limit of the
calculation accuracy, the experimental trend is well repro-
duced by the theoretical calculation. It shows nearly equal
stabilities for the isomeric pair VIII (describing C3), and
more significant differences for the isomers IX (C5) and X
(like C6). In the latter two cases, the energy balance favors
the syn configuration, more notably for X, where the size
difference of the ortho substituents (H and tBu) is the largest.
In this case, the syn and anti structures have important
differences that become noticeable in Figure 8. The syn
isomer readily accommodates the iron center in the usually
model
ortho
∆E(antifsyn)
∆Gexp
complex Keq (kcal mol-1
a
b
d
molecule substituents (kcal mol-1
)
)
A
B
C
Me, i-Pr
Me, t-Bu,
H, t-Bu
+0.33
-0.61
-2.84
C3
C6
C5
1
0.00
-0.82
-1.34
4
10c
a Negative values favor the syn isomer. Keq ) [major isomer]/[minor
isomer], CD2Cl2 at RT unless otherwise stated. c At 233 K. d Calculated
from Keq.
b
preferred square-pyramidal coordination environment, al-
lowing the apical Cl to lie in the least hindered face of the
complex. This is not possible in the anti configuration, which
t
bears one bulky Bu group on each side ot the molecule,
forcing the iron center to become trigonal bipyramidal in
order to minimize unfavorable steric interactions with the
halide ligands. The same steric effects probably play a similar
role in the destabilization of the anti isomer of IX. Thus,
the calculations suggest that, for C5 and C6, the syn isomer
should be preferred in solution. Recalling that, experimen-
tally, C5s is the most stable isomer in solution, the calculation
supports that this has the syn configuration, following the
same trend as C3, C3′, and C4.
Conclusions
The 1H NMR spectra of paramagnetic Fe(II)-2,6-bisimi-
nopyridine complexes having N-aryl groups with two dif-
ferent substituents at the ortho positions revealed that in
solution they exist as geometric isomers, or atropisomers,
due to the slow rotation of the aromatic rings. This property
is important for their performance as catalysts for R-olefin
polymerization or other catalytic processes susceptible to
stereochemical control. We synthesized a series of complexes
exhibiting different degrees of steric hindrance, and we have
shown that the stereochemical stability of such complexes
3688 Inorganic Chemistry, Vol. 48, No. 8, 2009

Fe(II) 2,6-Bis(N-arylimino)pyridine Complexes
product was dried under vacuum (1.81 g, 81%). The blue powdery
solid was analyzed by H NMR, observing a signal pattern that
varies widely. When the ligand substituents are small, isomer
interconversion is fast, even on the NMR time scale.
However, a moderate degree of steric hindrance suffices to
allow the chemical separation of the syn and anti diastereo-
somers. We developed a simple procedure based on solubility
differences for the preparation of diastereoisomerically pure
or enriched samples of these complexes. The actual config-
uration of these was assigned by the introduction of a
prochiral group in the pyridine ring of one of the ligands
and was generalized on the basis of the isomer solubility
properties.
The availability of these complexes as single diastereoi-
somers allowed kinetic measurements of the isomerization
rates, and the determination of the corresponding activation
parameters. Isomerization free activation energies for the
complexes are only slightly larger than in the corresponding
free ligands, suggesting that the aryl rotation requires the
previous dissociation of the Fe-N(imine) bond. This mech-
anism provides a framework for a consistent interpretation
of the isomerization ∆H‡ and ∆S‡ parameters for both
complexes and free ligands. This interpretation is further
supported by DFT calcultations, whose predictions on the
relative stabilities of the isomers and the feasibility of the
Fe-N(imine) dissociation closely match the experimental
data.
1
corresponds with that of the isomer C3a. Anal. calcd for
C29H35N3Cl2Fe: C, 63.06; H, 6.39; N, 7.61. Found: C, 63.48; H,
7.62; N, 7.45. µeff ) 5.7 µB (magnetic balance, 25 °C). 1H NMR of
C3a (400 MHz, CD2Cl2, 298 K): δ 81.29 (∆ν1/2 ) 41 Hz, 2H,
3,5-CHpy), 60.87 (∆ν1/2 ) 44 Hz, 1H, 4-CHpy), 15.96 (∆ν1/2 ) 19
Hz, 2H, m-CH(ar)), 15.10 (∆ν1/2 ) 19 Hz, 2H, m-CH(ar)), 8.19 (∆ν1/2
) 99 Hz, 6H, o-CH3), -2.24 (∆ν1/2 ) 54 Hz, 6H, CH(CH3)2),
-4.91 (∆ν1/2 ) 18 Hz, 6H, CH(CH3)2), -11.03 (∆ν1/2 ) 25 Hz,
2H, p-CH(ar)), -16.23 (∆ν1/2 ) 280 Hz, 2H, CH(CH3)2), -29.87
(∆ν1/2 ) 34 Hz, 16H, CH3-CdN). IR (Nujol mull, cm-1): 1625,
1587 (ν(CdN)). UV-vis (CH2Cl2): λmax 696 nm (ε ) 3238).
Isomerization of C3a. C3a in CD2Cl2 was allowed to stand at
1
room temperature under inert atmosphere. After 24 h, a new H
NMR analysis showed the signals of C3a together with a new set
of signals corresponding to C3s (relative ratio of 1:1). The
equilibrium could also be achieved by heating the solution at 40
1
°C for 3 h. H NMR of C3s (400 MHz, CD2Cl2, 298 K): δ 81.04
(∆ν1/2 ) 47 Hz, 2H, 3,5-CHpy), 60.34 (∆ν1/2c) 28 Hz, 1H, 4-CHpy),
16.59 (∆ν1/2 ) 20 Hz, 2H, m-CH(ar)), 14.37 (∆ν1/2 ) 21 Hz, 2H,
m-CH(ar)), 5.03 (∆ν1/2 ) 104 Hz, 6H, o-CH3), 4.58 (∆ν1/2 ) 52
Hz, 6H, CH(CH3)2), -5.37 (∆ν1/2 ) 18 Hz, 6H, CH(CH3)2), -10.96
(∆ν1/2 ) 19 Hz, 2H, p-CH(ar)), -11.61 (∆ν1/2 ) 280 Hz, 2H,
CH(CH3)2), -29.17 (∆ν1/2 ) 34 Hz, 6H, CH3-CdN).
Rectification of the Isomeric Mixtures of C3. A sample of the
mixture C3a/C3s (1:1; 0.1 g, 0.18 mmol) was suspended in hexane/
dichloromethane (20 mL). The ratio of the solvents (ca. 1:1) was
adjusted to allow a slight solubility of the solid (the clear solution
has a faint blue color). The suspension was stirred at ambient
temperature for 3 days. Upon filtration and drying under a vacuum,
Protocols for the separation of diastereoisomers and the
mechanistic insights described in this paper may prove
important for the applications of FeX2(PDI) complexes in
stereoselective catalytic processes. We are currently applying
these ideas to further improve the stereochemical stability
of these complexes.
1
the blue isolated solid was identified by H NMR as essentially
pure C3s. In a similar experiment, the isomeric mixture (0.1 g,
0.18 mmol) was suspended in THF and stirred at room temperature
for 3 days. Upon filtration and drying under a vacuum, the blue
Experimental Section
1
isolated solid was identified by H NMR as essentially pure C3a.
Most preparations and other operations were carried out under
an oxygen-free nitrogen or argon atmosphere using conventional
Schlenk techniques or in a nitrogen-filled glovebox. Conventional
organic synthesis procedures were performed under normal air
atmosphere. NMR spectra were recorded on a Bruker DRX 300,
Complex C3′. FeBr2 (0.43 g, 2.0 mmol) was suspended in
THF (15 mL), and a solution of L3′ (0.89 g, 2.1 mmol) in THF
(15 mL) was added dropwise at room temperature. The color of
the resultant yellow mixture slowly changed to deep-blue. The
suspension was stirred for 24 h and was allowed to settle. The
liquid phase was removed by filtration, and the blue solid was
washed with diethylether (20 mL) and hexane (20 mL). The blue
1
400, or 500 MHz spectrometers. H and 13C{1H} resonances of
the solvents were used as internal standards, but the chemical shifts
were reported with respect to TMS. Mycroanalyses were performed
by the Mycroanalytical Service of the Instituto de Investigaciones
Qu´ımicas (Seville, Spain). Infrared and UV-vis spectra were
recorded on Bruker Vector 22 and Perkin-Elmer Lambda 12
spectrophotomers, respectively. Magnetic susceptibility measure-
ments were made on a Sherwood Scientific balance, model MSB-
Auto. HPLC-grade solvents were freshly distilled prior to use. THF
and diethyl ether were distilled under nitrogen from sodium
benzophenone; CH2Cl2 and chlorobenzene from CaH2; and hexane,
toluene, and benzene from sodium.
1
residue was dried under a vacuum (0.91 g, 71%). The H NMR
spectra of the solid show signals which correspond to the mixture
of isomers in a relative ratio of 9:1, for the major product, C3′a,
relative to the minor, C3′s. Anal. calcd for C29H35N3Br2Fe: C,
54.32; H, 5.50; N, 6.55. Found: C, 53.95; H, 5.27; N, 6.60. µeff
) 5.5 µB (magnetic balance, 25 °C). IR (Nujol mull, cm-1): 1613,
1584 (ν(CdN)). UV-vis (CH2Cl2): λmax 686 nm (ε ) 2643).
Equilibration and rectification procedures analogous to those
described for C3 allowed the preparation of pure samples of
1
the two isomers. H NMR of C3′a (400 MHz, CD2Cl2, 298 K):
Details of the synthesis of ligands, inorganic precursors, and some
metal complexes (C1, C2, C5, and C7) can be found in the
Supporting Information.
δ 76.09 (∆ν1/2 ) 51 Hz, 2H, 3,5-CHpy), 71.06 (∆ν1/2 ) 29 Hz,
1H, 4-CHpy), 15.91 (∆ν1/2 ) 20 Hz, 2H, m-CH(ar)), 15.16 (∆ν1/2
) 18 Hz, 2H, m-CH(ar)), 8.94 (∆ν1/2 ) 107 Hz, 6H, o-CH3),
2.02 (∆ν1/2 ) 50 Hz, 6H, CH(CH3)2), -4.09 (∆ν1/2 ) 19 Hz,
6H, CH(CH3)2), -11.52 (∆ν1/2 ) 17 Hz, 2H, p-CH(ar)), -26.13
Synthesis of the Complexes and Isomer Purification
Procedures. Complex C3.3a To a stirred suspension of FeCl2 ·4H2O
(0.795 g, 4.0 mmol) in THF, L3 (1.0 g, 4.4 mmol, 1.1 equiv) was
added at room temperature, and the mixture was stirred for 16 h.
The color slowly changed from yellow to deep blue. The suspension
was allowed to settle and was filtrated, and the solid was washed
with diethyl ether (2 × 20 mL) and hexane (2 × 20 mL). The
1
(∆ν1/2 ) 29 Hz, 6H, CH3-CdN). H NMR of C3′s (400 MHz,
CD2Cl2, 298 K): δ 75.87 (∆ν1/2 ) 39 Hz, 2H, 3,5-CHpy), 72.48
(∆ν1/2 ) 28 Hz, 1H, 4-CHpy), 15.81 (∆ν1/2 ) 20 Hz, 2H,
m-CH(ar)), 15.49 (∆ν1/2 ) 18 Hz, 2H, m-CH(ar)), 10.60 (∆ν1/2
)
Inorganic Chemistry, Vol. 48, No. 8, 2009 3689

Ca´mpora et al.
117 Hz, 6H, o-CH3), -0.03 (∆ν1/2 ) 47 Hz, 6H, CH(CH3)2),
-4.17 (∆ν1/2 ) 22 Hz, 6H, CH(CH3)2), -11.52 (∆ν1/2 ) 17 Hz,
2H, p-CH(ar)), -26.65 (∆ν1/2 ) 34 Hz, 6H, CH3-CdN).
Complex C4. A solution of FeCl2 ·4H2O (0.30 g, 1.5 mmol) in
THF (15 mL) was treated with L4 (0.82 g, 1.6 mmol) in THF,
added dropwise at room temperature. The color and nature of the
mixture changed as described before. The reaction mixture was
stirred for 24 h and the solvent removed by filtration. The solid
was washed with diethyl ether (10 mL) and hexane (10 mL). The
resultant product was dried under a vacuum (0.736 g, 77%). The
1H NMR spectrum showed signals of essentially pure C4a. Anal.
calcd for C35H47N3Cl2Fe: C, 66.04; H, 7.44; N, 6.60. Found: C,
66.40; H, 7.35; N, 6.42. µeff ) 5.48 µB (magnetic balance, 25 °C).
IR (Nujol mull, cm-1): 1575 (ν(CdN)). UV-vis (CH2 Cl2): λmax
697 nm (ε ) 2400). Equilibration and rectification procedures
analogous to those described for C3 allowed the prepation of pure
samples of the two isomers. 1H NMR of C4a (300 MHz, CD2Cl2,
298 K): δ 83.27 (∆ν1/2 ) 55 Hz, 2H, 3,5-CHpy), 77.16 (∆ν1/2 ) 27
Hz, 1H, 4-CHpy), 16.24, (∆ν1/2 ) 27 Hz, 2H, m-CH(ar)), 14.21 (∆ν1/2
) 23 Hz, 2H, m-CH(ar)), 4.77 (∆ν1/2 ) 128 Hz, 6H, o-CH3), -1.46
(∆ν1/2 ) 18 Hz, 6H, CH3-C), -3.79 (∆ν1/2 ) 23 Hz, 6H,
CH(CH3)2), -11.96 (∆ν1/2 ) 37 Hz, 2H, p-CH(ar)), -13.77 (∆ν1/2
) 69 Hz, 6H, CH(CH3)2), -15.30 (∆ν1/2 ) 275 Hz, 2H, CH(CH3)2).
1H NMR of C4s (300 MHz, CD2Cl2, 298 K): δ 82.94 (∆ν1/2 ) 50
Hz, 2H, 3,5-CHpy), 79.46 (∆ν1/2 ) 27 Hz, 1H, 4-CHpy), 16.36 (∆ν1/2
) 23 Hz, 2H, m-CH(ar)), 14.21 (∆ν1/2 ) 23 Hz, 2H, m-CH(ar)),
-1.19 (∆ν1/2 ) 23 Hz, 6H, CH3-C), -4.30 (∆ν1/2 ) 18 Hz, 6H,
CH(CH3)2).
298 K): δ 81.29, (∆ν1/2 ) 85 Hz, 1H, 3- or 5-CHpy), 80.12 (∆ν1/2
) 85 Hz, 1H, 3 or 5-CHpy), 30.75 (∆ν1/2 ) 91 Hz, 1H, 4-CHpy),
24.11 (∆ν1/2 ) 207 Hz, 3H, o-CH3), 21.61 (∆ν1/2 ) 54 Hz, 3H,
p-CH3), 19.43 (∆ν1/2 ) 140 Hz, 1H, m-CH(ar2)), 15.98 (∆ν1/2 ) 99
Hz, 1H, m-CH(ar1)), 12.33 (∆ν1/2 ) 55 Hz, 1H, m-CH(ar1)), 11.85
(∆ν1/2 ) 146 Hz, 1H, m-CH(ar2)), 6.22 (∆ν1/2 ) 244 Hz, 6H,
CH(CH3)2), 5.03 (∆ν1/2 ) 244 Hz, 3H, o-CH3), -4.35 (∆ν1/2
)
140 Hz, 6H, CH(CH3)2), -12.50 (∆ν1/2 ) 54 Hz, 2H, p-CH(ar2)),
-12.50 (∆ν1/2 ) 54 Hz, 6H, CH3-CdN), -19.54 (∆ν1/2 ) 158
Hz, 1H, o-CH(ar2)), -25.30 (∆ν1/2 ) 73 Hz, 6H, CH3-CdN). IR
(Nujol mull, cm-1): 1621, 1587 (ν(CdN)). UV-vis (CH2Cl2): λmax
708 nm (ε ) 1872).
Complex C9. FeCl2(dme) (0.96 g, 0.57 mmol) in diethylether
(20 mL) was treated with L9 (0.38 g, 0.63 mmol) and stirred at
room temperature for 16 h. The color of the mixture changed to
blue, but no precipitate was observed. The solution was evaporated
to dryness and the residue extracted with CH2Cl2 (10 mL); traces
of unreacted FeCl2(dme) were separated by filtration. The solution
was taken to dryness and the residue dried, washed with diethyl
ether (2 × 20 mL) and hexane (2 × 20 mL), and dried under a
vacuum (191 mg, 50%). The solid was analyzed by 1H NMR; two
set of signals were observed, which corresponded to the isomeric
mixture anti/syn C9 in a relative ratio of 1:1. Careful recrystalli-
zation of the mixture with CH2Cl2/hexane (1:1) at -20 °C for 15
days afforded a blue microcrystaline solid, which was identified
by1HNMRasthesyn-C9(C9s)isomer.Anal.calcdforC39H47Cl2FeN3 ·1/
3CH2Cl2: C, 66.27; H, 6.74; N, 5.89. Found: C, 66.31; H, 6.99; N,
1
5.85. H NMR of C9s (400 MHz, CD2Cl2, 298 K): δ 82.91(∆ν1/2
) 53 Hz, 2H, 3,5-CHpy), 16.29, 14.76 (∆ν1/2 ) 19 Hz, 4H,
m-CH(ar)), 8.77 (∆ν1/2 ) 15 Hz, 2H, m-CH(Neo)), 7.41 (∆ν1/2 ) 18
Hz, 2H, o-CH(Neo)), 7.05 (∆ν1/2 ) 17 Hz, 2H, p-CH(Neo)), 6.21 (∆ν1/2
) 93 Hz, 6H, o-CH3), 3.74 (∆ν1/2 ) 46 Hz, 6H, CH(CH3)2), 2.98
(∆ν1/2 ) 8 Hz, 6H, C(Me)2), -3.77 (∆ν1/2 ) 19 Hz, 6H, CH(CH3)2),
-4.10 (∆ν1/2 ) 23 Hz, 2H, CH2), -8.26 (∆ν1/2 ) 267 Hz, 2H,
CH(CH3)2), -9.43 (∆ν1/2 ) 20 Hz, 2H, p-CH(ar)), -17.48 (∆ν1/2
) 45 Hz, 6H, CH3-CdN). IR (Nujol mull, cm-1): 1589 (ν(CdN)).
Complex C6. A solution of L6 (0.39 g, 0.7 mmol) in diethylether
(10 mL) was added to a stirred suspension of FeCl2(dme) (140 mg,
0.64 mmol) in the same solvent (10 mL). The color of the
suspension changed to blue. The mixture was kept under vigorously
stirring for 40 min. A blue solid precipitated and was isolated by
filtration, washed with cold (0 °C) diethylether (3 × 20 mL), and
dried under a vacuum (0.23 g, 47%). 1H NMR showed essentially
pure C6a, although signals corresponding to ca. 10% of isomer
C6s were also detected. µeff ) 5.33 µB (magnetic balance, 25 °C).
UV-vis (CH2 Cl2): λmax, 674 nm (ε ) 1298). IR (Nujol mull, cm-1):
1605, 1579 (ν(CdN)). Equilibration in CD2Cl2 at room temperature
leads to a 1:4 mixture of C6a and C6s, and this proportion changed
to 10:1 by stirring in hexane with a small amount of dichlorome-
thane. 1H NMR of C6a (400 MHz, CD2Cl2, 298 K): δ 88.42 (∆ν1/2
) 120 Hz, 2H, 3,5-CHpy), 45.62 (∆ν1/2 ) 127 Hz, 1H, 4-CHpy),
13.54 (∆ν1/2 ) 20 Hz, 2H, m-CH(ar)), 12.55 (∆ν1/2 ) 20 Hz, 2H,
m-CH(ar)), 3.86 (∆ν1/2 ) 85 Hz, 6H, o-CH3), 2.35 (∆ν1/2 ) 7 Hz,
18H, o-tBu), -13.04 (∆ν1/2 ) 115 Hz, 18H, p-tBu), -15.54 (∆ν1/2
1
UV-vis (CH2 Cl2): λmax 685 nm (ε ) 1421). H NMR of C9a
(400 MHz, CD2Cl2, 298 K): δ 83.18 (∆ν1/2 ) 53 Hz, 2H, m-CHpy),
16.14, 14.86 (∆ν1/2 ) 19 Hz, 4H, m-CH(ar)), 9.60 (∆ν1/2 ) 93 Hz,
6H, o-CH3), 8.71 (∆ν1/2 ) 15 Hz, 2H, m-CH(Neo)), 7.37 (∆ν1/2
)
18 Hz, 2H, o-CH(Neo)), 7.05 (∆ν1/2 ) 17 Hz, 2H, p-CH(Neo)), 3.04
and 2.69 (∆ν1/2 ) 8 Hz, 6H, C(Me)2), -1.66 (∆ν1/2 ) 46 Hz, 6H,
CH(CH3)2), -3.77 (∆ν1/2 ) 19 Hz, 6H, CH(CH3)2), -4.10 (∆ν1/2
) 23 Hz, 2H, CH2), -9.43 (∆ν1/2 ) 20 Hz, 2H, p-CH(ar)), -14.10
(∆ν1/2 ) 267 Hz, 2H, CH(CH3)2), -18.01 (∆ν1/2 ) 45 Hz, 6H,
CH3-CdN).
1
Kinetic Studies. These were done by NMR. Samples were
prepared in the glovebox by dissolving a known amount (4-6 mg)
of the corresponding complex (C3a, C3a′, C4a, or C6) or ligand
(L6) in CD2Cl2 (0.6 mL) and immediately placing the sample in
the spectrometer probe, which was previously stabilized at the
desired temperature. The sample was allowed to stabilize for 4-5
) 18 Hz, 6H, CH3-CdN). H NMR of C6s (400 MHz, CD2Cl2,
298 K): δ 83.61(∆ν1/2 ) 139 Hz, 2H, 3,5-CHpy), 52.22 (∆ν1/2
)
364 Hz, 1H, p-CHpy), 16.42 (∆ν1/2 ) 19 Hz, 2H, m-CH(ar)), 8.75
(∆ν1/2 ) 19 Hz, 2H, 4-CH(ar)), 1.64 (∆ν1/2 ) 10 Hz, 18H, o-tBu),
-5.96 (∆ν1/2 ) 202 Hz, 6H, o-CH3), -7.67 (∆ν1/2 ) 190 Hz, 18H,
p-tBu), -17.07 (∆ν1/2 ) 52 Hz, 6H, CH3-CdN).
1
min. H NMR spectra were recorded periodically until no further
Complex C8. The monoimine complex {FeCl2[2-(CH3CO)-6-
(CH3C(N-Mes)C5H3N]-κ-O,N,N} (see the Supporting Information;
20.8 g, 51 mmol) was suspended in chlorobenzene (150 mL), and
2-isopropylaniline (14.9 mL, 102 mmol) was added. The reaction
mixture was stirred for 6 days at 55 °C. The suspension was filtered,
and the resultant blue solid was washed with diethylether (2 × 50
mL) and hexane (1 × 50 mL) and dried under vacuum for 3 days
(26.04 g, 97%). The excess of aniline could be easily recovered by
evaporation of the filtrate to dryness. Anal. calcd for C27H31N3Cl2Fe:
C, 61.85; H, 5.96; N, 8.01. Found: C, 61.85; H, 5.81; N, 7.71. µeff
) 4.97 µB (magnetic balance, 25 °C). 1H NMR (400 MHz, CD2Cl2,
change was observed. Progress of the equilibration process was
monitored by integration of a well-resolved pair of signals, one
for each isomer. Probe temperatures were measured with a standard
CH3OH/CDCl3 calibration sample.
Computational Details. Electronic structures and geometries
of the model complexes were computed within the density
functional theory at the B3LYP level,22 using the 6-311G* basis
set for Fe, Cl, and N and 6-31G* for C and H atoms of the PDI
(22) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (b) Lee, C.; Wang,
Y.; Parr, R. G. Phys. ReV. B 1988, 37, 785.
3690 Inorganic Chemistry, Vol. 48, No. 8, 2009

Fe(II) 2,6-Bis(N-arylimino)pyridine Complexes
ligands. Me, i-Pr, and t-Bu substituents on phenyl rings have
been described using the 3-21G* basis set. A single-point energy
calculation was performed on all optimized structures using the
6-311+G(2d,p) basis set. Vibrational frequency calculations were
done by diagonalization of the analytically computed Hessian
to ensure that the optimized structures were real minima (Nimag
) 0) or maxima (Nimag ) 1). All calculations were performed
using the Gaussian 03 package.23 Figures were drawn using
Molekel.24 XYZ coordinates of all optimized complexes are
available upon request.
Acknowledgment. Financial support from the DGI (Project
CTQ2006-05527/BQU), Junta de Andaluc´ıa (Project P06-
FQM-01704), and Repsol-YPF is gratefully acknowledged.
A.M.N. and M.A.C. are thankful for predoctoral grants from
CSIC (I3P and JAE programs), and C.M.P. is thankful for a
research contract from Repsol-YPF. A.R.-D. and D.R.
acknowdlege postdoctoral contracts from the Ministerio de
Educacio´n y Ciencia (Ramo´n y Cajal program) and the
European Union (sixth Framework Program Marie Curie
Outgoing International Fellowship), respectively. We thank
Dr. I. Urban (IIQ) for helpful discussions.
(23) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin,
K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone,
V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.;
Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa,
J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene,
M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Adamo, C.;
Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin,
A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.;
Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.;
Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas,
O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.;
Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.;
Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.;
Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen,
W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, revision
C.02; Gaussian, Inc.: Wallingford, CT, 2004.
Supporting Information Available: Details of the synthesis of
ligands, inorganic precursors, and some metal complexes (C1, C2,
C5, and C7). This material is available free of charge via the
Internet at http://pubs.acs.org.
IC802271Y
(24) Portmann, S.; Luthi, H. P. Chimia 2000, 54, 766.
Inorganic Chemistry, Vol. 48, No. 8, 2009 3691