(Py)2Co(CH2SiMe3)2 as an easily accessible source of "coR2"

Article abstract of DOI:10.1021/om901045s

(Py)₂CoR₂ (R = CH₂SiMe₃) is easily prepared from (Py)₄CoCl₂ and RLi. It is fairly stable at room temperature and serves as a convenient source of CoR₂ for transfer to other ligands. Unfortunately, (Py)₂CoR₂ was obtained only as an oil, but the structure of the related complex (Py) ₂CoR-₂ (R- = CH₂CMe₂Ph) could be confirmed by a single-crystal X-ray diffraction study. Transfer of the CoR ₂ fragment from (Py)₂CoR₂ or (TMEDA)CoR ₂ to diiminepyridine-type ligands (1-6) was studied as a function of ligand steric and electronic properties. Reaction with N-2,6-dimethylphenyl (1) and N-2,4,6-trimethylphenyl (2) ligands produced diamagnetic monoalkyl complexes; the structure of (1)CoR was confirmed by X-ray diffraction. With the less shielding N-phenyl (3) and N-benzyl (4) ligands, ¹H NMR indicated formation of diamagnetic Co^I alkyl species, but they were not stable enough to allow isolation. Fluorinated ligand 5 appears to be less reactive and-despite its supposedly stronger φ-acceptor character-also does not lead to formation of a stable Co^I alkyl complex. With PyBOX ligand 6, high-spin dialkyl complex (6)CoR₂ was observed by ¹H NMR. Based on these observations and DFT calculations, a mechanism is proposed for formation of diiminepyridine Co^I alkyls that involves formation of a high-spin κ² complex, spin flip to give a low-spin κ³ complex, and irreversible loss of an alkyl radical.

Full text of DOI:10.1021/om901045s

Organometallics 2010, 29, 1897–1908 1897
DOI: 10.1021/om901045s
(Py)2Co(CH2SiMe3)2 As an Easily Accessible Source of “CoR2”
Di Zhu,† Femke F. B. J. Janssen,‡ and Peter H. M. Budzelaar*,†
†Department of Chemistry, University of Manitoba, Winnipeg, MB R3T 2N2, Canada, and
‡IMM Solid State Chemistry, Radboud University, Heyendaalseweg 135, 6525 AJ Nijmegen,
The Netherlands
Received December 5, 2009
(Py)2CoR2 (R = CH2SiMe3) is easily prepared from (Py)4CoCl2 and RLi. It is fairly stable at room
temperature and serves as a convenient source of CoR2 for transfer to other ligands. Unfortunately,
(Py)2CoR2 was obtained only as an oil, but the structure of the related complex (Py)2CoR02 (R0 =
CH2CMe2Ph) could be confirmed by a single-crystal X-ray diffraction study. Transfer of the CoR2
fragment from (Py)2CoR2 or (TMEDA)CoR2 to diiminepyridine-type ligands (1-6) was studied as a
function of ligand steric and electronic properties. Reaction with N-2,6-dimethylphenyl (1) and N-
2,4,6-trimethylphenyl (2) ligands produced diamagnetic monoalkyl complexes; the structure of
(1)CoR was confirmed by X-ray diffraction. With the less shielding N-phenyl (3) and N-benzyl (4)
1
ligands, H NMR indicated formation of diamagnetic CoI alkyl species, but they were not stable
enough to allow isolation. Fluorinated ligand 5 appears to be less reactive and-despite its supposedly
stronger π-acceptor character-also does not lead to formation of a stable CoI alkyl complex. With
1
PyBOX ligand 6, high-spin dialkyl complex (6)CoR2 was observed by H NMR. Based on these
observations and DFT calculations, a mechanism is proposed for formation of diiminepyridine CoI
alkyls that involves formation of a high-spin κ2 complex, spin flip to give a low-spin κ3 complex, and
irreversible loss of an alkyl radical.
Introduction
synthetic route of treating a DIMPY metal dihalide precur-
sor with an alkyllithium reagent works well in selected cases,2
but frequently (in particular for less shielding ligands) com-
plications such as reduction, ligand alkylation, and depro-
tonation occur.2,3 An alternative approach is to use a metal
dialkyl with labile ligands and displace these by the desired
Transition metal alkyl complexes are among the more
reactive-and therefore more interesting-species in orga-
nometallic chemistry. Precisely because of this reactivity,
preparing and isolating such species can be a challenge.
Conventional wisdom holds that metal alkyls tend to be
stable if they (a) are coordinatively saturated (closed-shell
16-e or 18-e) and (b) have no β-hydrogens. The last decades
have seen a surge in interest in the chemistry of paramagnetic
first-row transition metal complexes bearing less traditional
ligands such as imine, pyridine, and carbene instead of the
more common phosphine, cyclopentadienyl, and carbon
monoxide. This interest was for a large part fueled by the
discovery of Fe and Co Brookhart/Gibson polymerization
catalysis using diiminepyridine (DIMPY) ligands.1
DIMPY ligand. A very convenient FeR2 “synthon”, re-
ꢀ
4,5
ported a few years ago by Campora, is (Py)2FeR2 (R =
CH2SiMe3, CH2Ph, and CH2CMe2Ph), prepared by treating
(Py)4FeCl2 with RLi or RMgX; its utility was clearly illu-
strated by the synthesis of [2,6-(2,4,6-Me3C6H2NdCMe)2-
C5H3N]Fe(CH2SiMe3)2. The group of Chirik has investi-
gated the scope of this transfer reaction for the synthesis of
(DIMPY)Fe dialkyl complexes and found that it is more
robust than the FeCl2/RLi approach, although some side
reactions were still observed.2c
For Co, reaction of (DIMPY)CoCl2 with metal alkyls
leads exclusively to reduction, giving either (DIMPY)CoCl
(2) (a) Bouwkamp, M. W.; Bart, S. C.; Hawrelak, E. J.; Trovitch,
R. J.; Lobkovsky, E.; Chirik, P. J. Chem. Commun. 2005, 3406.
(b) Bouwkamp, M. W.; Lobkovsky, E.; Chirik, P. J. J. Am. Chem. Soc.
ꢀ
2005, 127, 9660. (c) Fernandez, I.; Trovitch, R. J.; Lobkovsky, E.; Chirik, P. J.
Organometallics 2008, 27, 109.
(3) See e.g.: Knijnenburg, Q.; Gambarotta, S.; Budzelaar P. H. M.
Metal alkyls are likely the active species here, but isolating
such alkyls turns out to be nontrivial. For Fe, the standard
Dalton Trans. 2006, 5442, and references there.
ꢀ
(4) (a) Campora, J.; Marcos Naz, A.; Palma, P.; Alvarez, E. Orga-
ꢀ
nometallics 2005, 24, 4878. (b) (Py)2Fe(CH2SiMe3)2 was prepared by us
according to Campora's method and characterized by H NMR.
(5) For related labile-ligand iron alkyls, see: (a) Hill, D. H.; Sen, A.
1
ꢀ
*Corresponding author. E-mail: Peter_Budzelaar@umanitoba.ca.
(1) (a) Small, B. L.; Brookhart, M.; Bennett, A. M. 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.
ꢀ
J. Am. Chem. Soc. 1988, 110, 1650. (b) Albores, P.; Carrella, L. M.; Clegg,
ꢀ
W.; García-Alvarez, P.; Kennedy, A. R.; Klett, J.; Mulvey, R. E.; Rentschler,
E.; Russo, L. Angew. Chem., Int. Ed. 2009, 48, 3317.
r
2010 American Chemical Society
Published on Web 03/19/2010
pubs.acs.org/Organometallics

1898 Organometallics, Vol. 29, No. 8, 2010
Zhu et al.
Figure 1. 1H NMR spectra of (A) (Py)2CoR2 and (B) (TMEDA)CoR2.
Table 1. Comparison of 1H NMR Shifts (ppm) for (Py)2MR2
Complexes (M = Co, Fe)
metal complex in solution. After filtration and evaporation
of the solvent, the product was obtained as a bright green oil;
1
all attempts at crystallization failed.10 The H NMR data
4
4
(Py)2CoR2
(Py)2FeR2
(Py)2CoR02
(Py)2FeR02
(see Figure 1A) indicate the overall composition of the
complex and are rather similar to those of the Fe analogue
(see Table 1) except for the pyridine H4 proton (Co: -8.5
ppm; Fe: þ17.4 ppm4). Use of an internal standard
(hexamethylbenzene) allowed determination of the yield
(∼75%) by NMR. The complex appears to be indefinitely
stable under an inert atmosphere at -35 ꢀC and survives for
at least several days at room temperature. It is, however,
extremely sensitive to air and moisture. A few aspects of the
synthesis should be noted:
Me
10.3
114
38.4
-8.5
11.2
129a
35.5
17.4
21.6
108
25.2
Py H2
Py H3
Py H4
Ph Ho
Ph Hm
Ph Hp
32.7
-8.3
10.2b
7.4b
3.9
35.3
12.9
120
5.4
11.8
a This work (not reported by Campora4). b The assignments of Ho and
Hm could be reversed.
or (DIMPY)CoR, even for hindered DIMPY ligands.6 It is
not clear at present whether the metal alkyl directly reduces
Co or first transfers an alkyl group, which could then be lost
as a radical. Unlike for Fe, no convenient CoR2 equivalent is
available that might allow one to circumvent this problem.
Hay-Motherwell reported the synthesis and X-ray charac-
terization of (TMEDA)Co(CH2SiMe3)2 nearly 20 years ago,
but the complex was obtained in only 11% yield and its
isolation is awkward;7 as a result, this precursor has not been
used frequently. In addition, the group of Theopold has
reported the synthesis of ligand-free diaryl complex
[Co(2,4,6-Me3C6H2)2]2, but again synthesis is far from tri-
vial.8 The synthesis of the more hindered Co[2,6-(2,4,6-
Me3C6H2)2-C6H3]2 seems to be simpler,9 but this complex
is so hindered that complexation with ligands like DIMPY is
• The quality of the RLi used is pivotal. We crystallize
freshly received RLi from pentane and store it as a solid in a
glovebox to maintain its quality.
• (Py)4CoCl2 was prepared from anhydrous CoCl2 rather
than from the commonly used hydrated dichloride,11 to
avoid potential problems in the reaction with RLi. The
complex easily loses part of its coordinated pyridine under
vacuum, changing from pink to purple-blue.12 Ligand loss
also occurred when the pink solid was suspended in pentane
for the synthesis of (Py)2CoR2, but this dissociation did not
affect the outcome of the reaction.
• Chirik reported that reacting (Py)4FeCl2 with only 1
equiv of RLi resulted in formation of (Py)2FeRCl.2c For
Co, we found that use of 1 equiv of RLi produced only
(Py)2CoR2 and unreacted (Py)nCoCl2.
ꢀ
unlikely. Therefore, we decided to check whether Campora’s
approach would also work for Co.
Since (Py)2CoR2 could not be obtained in crystalline form,
obtaining a satisfactory elemental analysis for this complex
was not possible. Therefore, we employed several alternative
methods to establish its constitution. First, the purity was
Results and Discussion
1
estimated at 91% from the H NMR spectrum using an
Synthesis of (Py)2Co(CH2SiMe3)2. Treatment of
“(Py)4CoCl2” with RLi (R = CH2SiMe3) in pentane at low
temperature, followed by slow warming to room tempera-
ture, resulted in clean formation of (Py)2CoR2 as the only
internal standard (hexamethylbenzene). Second, using the
same internal standard, the magnetic moment was deter-
mined by the Evans NMR method13 as 4.8(3) μB (average of
ꢀ
(10) Campora similarly found (Py)2FeR2 to be too soluble to allow
(6) (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.
(7) Hay-Motherwell, R. S.; Wilkinson, G. Polyhedron 1990, 9, 931.
(8) Theopold, K. H.; SiIvestre, J.; Byrne, E. K.; Richeson, D. S.
Organometallics 1989, 8, 2001.
crystallization.4
(11) Long, G. J.; Clarke, P. J. Inorg. Chem. 1978, 17, 1394.
(12) (a) Howard, G. D.; Marianelli, R. S. Inorg. Chem. 1970, 9, 1738.
(b) Katzin, L. I. J. Chem. Phys. 1961, 35, 467.
€
(13) (a) Loliger, J.; Scheffold, R. J. Chem. Educ. 1972, 49, 646. (b) De
Buysser, K.; Herman, G. G.; Bruneel, E.; Hoste, S.; Van Driessche, I. Chem.
Phys. 2005, 315, 286.
(9) Kays, D. L.; Cowley, A. R. Chem. Commun. 2007, 1053.

Article
Organometallics, Vol. 29, No. 8, 2010 1899
Scheme 1
Figure S3), but its formation is relatively slow and the
product decomposes in a few hours at room temperature in
solution. Also, the presence of pyridine induces dispropor-
tionation of the Grignard reagent to give a complex we
assume to be (Py)2MgR02, with solubility properties very
similar to those of (Py)2CoR02. Several crystallization at-
tempts therefore produced mixtures of near-colorless
(Py)2MgR02 crystals and green-black blocks of (Py)2CoR02.
We eventually found conditions that lead to formation of
1
fairly pure (by H NMR) (Py)2CoR02, albeit in low yield
Figure 2. X-ray structure of (Py)2CoR02 (thermal ellipsoids
drawn at 30% probability, hydrogen atoms omitted for clarity).
Selected bond distances (A) and angles (deg): Co(1)-C(11):
2.066(3); Co(1)-C(31): 2.075(4); Co(1)-N(1): 2.117(3); Co(1)-
N(2): 2.108(3); N(1)-Co(1)-N(2): 127.32(15); C(11)-Co(1)-
C(31): 96.20(11).
(20%). Even in the solid state, (Py)2CoR02 was found to
decompose in about a day at room temperature. A fragment
broken from a green-black (Py)2CoR02 crystal was used for a
single-crystal structure determination. The quality of the
data is not very high due to decay of the crystal during the
measurement (see Experimental Section for details). Never-
theless, the results unambiguously show the connectivity
expected for (Py)2CoR02 (Figure 2). Both the unit cell
˚
three measurements at different concentrations). Despite the
relatively large error margin, this falls squarely in the range
expected for high-spin CoII (4.3-5.214). Third, hydrolysis of
an NMR sample containing the internal standard showed
formation of Me4Si and Py in the amounts expected for
(Py)2CoR2. Finally, the presence of a transferable CoR2
fragment was demonstrated by the reaction with TMEDA
to give the known7 complex (TMEDA)CoR2. On treating
(Py)2CoR2 with excess TMEDA in pentane, the reaction did
not immediately go to completion, but evaporation of all
solvents and crystallization from pentane containing a small
amount of extra TMEDA resulted in formation of bluish-
purple crystals of (TMEDA)CoR2; the structure was con-
firmed by single-crystal X-ray diffraction (see Figure S1).
This two-step synthesis has a considerably higher yield (57%
based on (Py)4CoCl2) than the procedure described by Hay-
Motherwell.7 The 1H NMR resonances of the TMEDA CH2
and CH3 groups overlap (Figure 1B: b,c), but the signals
have very different linewidths and can be separated by
deconvolution.
dimensions and the molecular conformation are very similar
ꢀ
4
to those reported by Campora for the Fe analogue; the most
notable differences between the two structures are in the
CMC angle (about 5ꢀ smaller for Co) and the NMN angle
(about 5ꢀ larger for Co). We do not at present have a good
explanation for the lower stability of (Py)2CoR02 compared
to (Py)2CoR2, but the analogous Fe complexes seem to
follow the same stability order.
Exchange with DIMPY and Related Ligands. In the con-
text of our study of ligand effects,15 we are interested in
preparing variations on the DIMPY theme with varying
steric and σ-donor/π-acceptor properties. To assess the
potential of (Py)2CoR2 and (TMEDA)CoR2 as sources of
the CoR2 fragment, we studied transfer to ligands 1-6
(Scheme 1). Ligands 1-4 are electronically similar to “stan-
dard” DIMPY ligand 7 but are less effective at shielding the
metal center. Of the remaining two variations, fluorinated
ligand 5 emerged as being a clearly better π-acceptor than 1,
but also a weaker σ-donor,15 whereas PyBOX ligand 616 is a
much poorer π-acceptor17 than 1-4.
Structure of (Py)2Co(CH2CMe2Ph)2. While the above-
mentioned characterization data leave little doubt about
the constitution of (Py)2CoR2, the lack of structural data is
Both (Py)2CoR2 and (TMEDA)CoR2 reacted rapidly with
both 1 and 2 to give deep purple, diamagnetic monoalkyl
complexes as the sole observed products. They were easily
ꢀ
unsatisfactory. Since Campora did succeed in obtaining
crystals of (Py)2FeR02 (R0 = CH2CMe2Ph),4 we decided to
attempt synthesis of the Co analogue from (Py)4CoCl2
and R0MgCl. This reaction was found to suffer from a
few complications. According to 1H NMR, the hoped-
for product (Py)2CoR02 is indeed formed (see Table 1 and
(15) Zhu, D.; Budzelaar, P. H. M. Organometallics 2008, 27, 2699.
(16) (a) Nishiyama, H.; Kondo, M.; Nakamura, T.; Itoh, K. Orga-
nometallics 1991, 10, 500.
(17) Calculated ligand parameters for 6 (following the procedure of
ref 15): σ = þ1.83, π = þ12.21 (Ereorg = -10.01). These values are
relative to ligand 1; positive values indicate poorer donor or acceptor
properties.
(14) Day, M. C.; Selbin, J. In Theoretical Inorganic Chemistry, 2nd
ed.; Reinhold: New York, 1969; p 492.

1900 Organometallics, Vol. 29, No. 8, 2010
Zhu et al.
Figure 3. X-ray structure of (1)CoR (thermal ellipsoids drawn at 30% probability, hydrogen atoms omitted for clarity). Selected bond
˚
distances (A) and angles (deg): Co(1)-N(1): 1.834(2); Co(1)-N(2): 1.9165(15); C(1)-N(2): 1.330(2); C(1)-C(2): 1.434(3); Co(1)-C(14):
1.964(3); N(1)-Co(1)-N(2): 80.63(5); Si-C(14)-Co(1): 128.55(15); N(1)-Co(1)-C(14): 165.07(10).
identified by their characteristic 1H NMR signals (Py H4 at
10.1 ppm, NdCMe at -1.2 ppm18). Since synthesis of
(TMEDA)CoR2 requires an additional step, (Py)2CoR2 is
the reagent of choice here. The complex formed from 2 could
not be obtained in crystalline form, but (1)CoR crystallizes
well; the X-ray structure is shown in Figure 3. It is rather
similar to the previously reported complex of 7:6a the
Co1-N1, C1-C2, and C1-N2 bond lengths in (1)CoR are
virtually identical to the corresponding distances in (7)CoR,
tration. We speculate that what is formed in this case might
be a labile pyridine adduct (3)(Py)CoR in dynamic equilib-
rium with (3)CoR (vide infra). In any case, reaction mixtures
obtained from (Py)2CoR2 and 3 slowly deposit black inso-
luble material and no products could be isolated from them.
As for 3, attempts to isolate well-defined complexes of 4
failed. Therefore, these reactions were monitored by 1H
NMR. The cleanest spectra were obtained using an excess
of the cobalt dialkyl precursor. Reaction of (TMEDA)CoR2
with 4 produced the signals expected for (4)CoR: pyridine
H4 at 9.74 ppm, imine Me at -0.64 ppm. Like for 3, reaction
of 4 with excess (Py)2CoR2 gave signals consistent with a
single CoI alkyl, but at different chemical shifts than for the
TMEDA reaction and dependent on reactant ratio and
concentration (Figure S5B,C). In support of our hypothesis
of pyridine complexation to the product, addition of excess
pyridine to the (TMEDA)CoR2 þ 4 reaction mixture re-
sulted in signals similar to those observed for (Py)2CoR2 þ 4
(Figure 5D). Thus, it appears that the decreased steric
shielding of ligands 3 and 4 (compared to 1 and 2) does not
hinder formation of low-spin CoI complexes but leads to a
significant decrease of their stability.
For fluorinated ligand 5, we expected easy formation of
a highly stabilized CoI alkyl. A description of the
(previously unpublished) synthesis of 5 was kindly pro-
vided by Jon M. Malinoski (Brookhart group, UNC
Chapel Hill)19 and is included in the Experimental Section;
the structure of the ligand was confirmed by an X-ray
structure determination (Figure 4). Unfortunately, ligand
exchange was not very successful. With (TMEDA)CoR2,
not much reaction was observed within 2 h of mixing; on
longer standing, the spectra broadened and a black solid
precipitated. With excess (Py)2CoR2, 19F NMR spectra
˚
while the Co-C14 bond in (1)CoR is about 0.02 A shorter,
perhaps reflecting the lower steric pressure in the dimethyl-
phenyl-substituted ligand. In both complexes, the alkyl
carbon is bent out of the CoN3 plane by a very similar
amount (trans NCoC angle close to 165ꢀ). This is most likely
to mitigate repulsion between the bulky trimethylsilyl group
and the substituent aryl groups since (7)CoMe has a virtually
linear NCoC arrangement.6b
Encouraged by the clean reactions with 1 and 2, we also
attempted transfer to the much more open ligands 3 and 4.
(TMEDA)CoR2 reacts with 3, and 1H NMR spectra of the
reaction mixtures (Figure S4A) indicated fairly clean forma-
tion of (3)CoR, with the expected characteristic signals for
the pyridine H4 (9.79 ppm) and imine Me (-1.10 ppm).18
The solutions show some stability (days), but all attempts at
workup led to decomposition. 1H NMR spectra obtained for
the reaction of (Py)2CoR2 with 3 also indicated formation of
a single diamagnetic CoI alkyl complex, but interestingly
enough the chemical shifts are quite different from those
obtained using (TMEDA)CoR2; in particular, the character-
istic signals at 9.79 and -1.10 ppm are absent (Figure S4B,C).
Also, the chemical shifts observed this way vary depending
on the (Py)2CoR2:3 ratio used and on the absolute concen-
(18) Knijnenburg, Q.; Hetterscheid, D. G. H.; Kooistra, T. M.;
Budzelaar, P. H. M. Eur. J. Inorg. Chem. 2004, 1204.
(19) Malinoski, J. M.; Brookhart, M. Personal communication.

Article
Organometallics, Vol. 29, No. 8, 2010 1901
Figure 4. X-ray structure of 5 (30% thermal ellipsoids; hydrogen atoms omitted for clarity).
Figure 5. 1H NMR spectrum of “(6)CoR2”, still containing some (TMEDA)CoR2 (see text).
obtained immediately after mixing showed formation of a
single main diamagnetic product as well as remaining free
ligand (Figure S6B), and the corresponding 1H NMR
spectra (Figure S7B) showed upfield peaks compatible
with formation of a diamagnetic CoI alkyl species and
Me4Si. Even on standing, the free ligand did not comple-
tely disappear, but spectra became broader as a black solid
precipitated (Figure S6C). On hydrolysis of the reaction
mixture, the diamagnetic product was mostly converted
back into the free ligand (Figure S6D). It seems likely
that this diamagnetic product is (5)CoR or a pyridine
adduct of it. In any case, the better π-acceptor ability of
ligand 5 does not seem to translate into easier formation of
a stable CoI alkyl.
reaction was observed with (TMEDA)CoR2, although this
turned out to be an equilibrium with an equilibrium con-
stant of 0.9(1).20 In the equilibrium mixture, the 1H NMR
signals of free 6 are sharp, but those of free TMEDA are
broadened, presumably due to exchange with (TMEDA)-
CoR2 (Figure 5).21 The equilibrium mixture was fairly
stable at -35 ꢀC but decomposed in hours at room tem-
perature.
Choice of Computational Method. DFT is the only prac-
tical method for studying transition metal complexes of the
type described here. The main choices to be made are basis
set and functional. We decided to use the modest SV(P) basis
for initial geometry optimizations and vibrational analyses,
and the larger TZVP basis for reoptimization and improved
For weak π-acceptor ligand 6, formation of CoI should
be more difficult, and we hoped this might allow isolation
of stable dialkyl complexes, the presumed intermediates
before loss of an alkyl group. Unfortunately, the reaction of
6 with (Py)2CoR2 was found to be far from clean. A cleaner
(20) See the Supporting Information for estimation of the K value.
(21) Exchange broadening, if any, of (TMEDA)CoR2 would not be
detectable because its line width is completely dominated by paramag-
netic broadening. The broadening of free TMEDA may have contribu-
tions from both standard exchange and paramagnetic relaxation.

1902 Organometallics, Vol. 29, No. 8, 2010
Zhu et al.
energies. Since the reactions we are studying involve changes
in spin state, different functionals can be expected to produce
rather different results.22 In particular, “pure” functionals
such as BP86 tend to favor low-spin states, while hybrid
functionals such as B3LYP often favor high-spin states.
There is, in general, no functional that can be expected to
always produce accurate spin-state energy differences. For a
series of cyclopentadienylcobalt carbonyl complexes, Gan-
don et al recently found that BP86 gave results close to
experiment, whereas B3LYP incorrectly predicted the triplet
state to be much too stable.23 In contrast, in a study of
(DIMPY)FeR2 complexes24 experimentally found to be
high-spin, even b3-lyp put the high-spin state barely above
the intermediate-spin state, and pure functionals gave a very
definite preference for lower spin states.
To cover our bases, we explored the chemistry of our CoI
and CoII alkyls with Turbomole25 using both pure (b-p) and
hybrid (b3-lyp) functionals (b-p and b3-lyp are similar to the
Gaussian BP86 and B3LYP functionals, respectively, but use
VWN(V) instead of VWN(III) for the correlation part of the
functional26,27). The results showed that, within the high-
spin state, the two functionals behave rather similarly.
However, the predictions diverge for low-spin (DIMPY)-
CoR2 and (PyBOX)CoR2. At the b3-lyp level, these are
somewhat higher in energy than the high-spin states, which
leads to predicted ligand exchange equilibria in agreement
with experiment (vide infra). With the b-p functional, how-
ever, the low-spin states of both (DIMPY)CoR2 and
(PyBOX)CoR2 are much lower than the high-spin states,
leading to equilibrium predictions incompatible with experi-
ment. In addition, paramagnetic NMR shifts using b3-lyp
were found to agree fairly well with experiment for these
systems, while the b-p predictions are consistently further
off. For these reasons, we only present the b3-lyp results here
(the b-p results are included in the ESI), without implying
that b3-lyp would in general be better for predicting spin-
state energy differences.
Figure 6. Comparison of calculated (B3LYP: δorb þ δFC only)
and observed 1H NMR chemical shifts (the solid line represents
the ideal δobs = δcalc relation).
The reference shift δorb was calculated using Gaussian 0329
(B3LYP/TZVP), the pseudocontact term δPC was neglected,
and the Fermi contact term δFC was calculated according to
eq 2,30 where the isotropic hyperfine coupling Aiso was
calculated with ORCA31 (B3LYP/TZVP).
geβeSðS þ 1Þ
δFC ¼ Aiso
ð2Þ
gNβN3kT
For (Py)2CoR2, (Py)2CoR02, and (TMEDA)CoR2 the
agreement (Figure 6) is seen to be quite good (correlation
coefficient of 0.9927) and supports our assignment in Table 1
(although the assignments for the phenyl o and m hydrogens
of (Py)2CoR02 might well be switched). Encouraged by this,
we also estimated the shifts for (6)CoR2, for which we do not
know the structure. It is not a priori obvious what hapticity
(κ2 or κ3) and spin state should be expected for this complex.
Calculations using the pure functional b-p put the low-spin
(LS) κ3 structure about 10 kcal/mol below the high-spin (HS)
κ2 structure, while the hybrid functional b3-lyp predicts the
high-spin κ2 structure to be lowest, with LS κ3 about 10 kcal/
mol higher in energy. The calculated NMR data (Table S2)
are much more consistent with a HS ground state and fast
Paramagnetic NMR Shifts. The NMR signals of CoII
dialkyls are all broadened and show strong paramagnetic
shifts. The only “tool” for assignment is the integral, and that
is not always enough. Therefore, we used computational
methods to predict the paramagnetically shifted signals and
support the assignments. The observed chemical shift can be
written as28
δobs ¼ δorb þ δFC þ δPC
ð1Þ
(22) See e.g.: Harvey, J. N. Annu. Rep. Prog. Chem. C 2006, 102, 203.
(23) Gandon, V.; Agenet, N.; Vollhardt, K. P. C.; Malacria, M.;
Aubert, C. J. Am. Chem. Soc. 2009, 131, 3007.
(24) Scott, J.; Gambarotta, S.; Korobkov, I.; Budzelaar, P. H. M. J.
Am. Chem. Soc. 2005, 127, 13019.
(29) 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.; Bakken,
V.; 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.; Zakr-
zewski, 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.
€
(25) (a) Ahlrichs, R.; Bar, M.; Baron, H.-P.; Bauernschmitt, R.;
Bocker, S.; Ehrig, M.; Eichkorn, K.; Elliott, S.; Furche, F.; Haase, F.;
€
€
€
Haser, M.; Hattig, C.; Horn, H.; Huber, C.; Huniar, U.; Kattannek, M.;
€
€
Kohn, A.; Kolmel, C.; Kollwitz, M.; May, K.; Ochsenfeld, C.; Ohm, H.;
Schafer, A.; Schneider, U.; Treutler, O.; Tsereteli, K.; Unterreiner, B.;
€
Von Arnim, M.; Weigend, F.; Weis, P.; Weiss, H. Turbomole Version 5;
Theoretical Chemistry Group, University of Karlsruhe, 2002. (b) Treutler, O.;
Ahlrichs, R. J. Chem. Phys. 1995, 102, 346.
(26) Vosko, S.; Wilk, L.; Nussair, M. Can. J. Phys. 1980, 58, 1200.
(27) Ahlrichs, R.; Furche, F.; Grimme., S. Chem. Phys. Lett. 2000,
325, 317.
(28) See e.g.: (a) Moon, S.; Patchkovskii, S. First Principles Calcula-
tions of Paramagnetic NMR Shifts. In Calculations of NMR and EPR
Parameters; Kaupp, M.; B€uhl, M.; Malkin, V. G., Eds.; Wiley-VCH:
Weinheim, 2004. (b) Pennanen, T. O.; Vaara, J. Phys. Rev. Lett. 2008,
100, 133002.
(30) McConnell, H. M.; Chesnut, D. B. J. Chem. Phys. 1958, 28, 107.
(31) Neese, F. ORCA, version 2.6, Revision 4; University of Bonn:
Germany, July 2007

Article
Table 2. Calculated Free Energy Differencesa for Ligand
Exchange Reactions (kcal/mol)
Organometallics, Vol. 29, No. 8, 2010 1903
sumably for steric reasons. Loss of an alkyl is favorable for
(κ3-1)Co(CH2SiMe3)2 because it relieves the steric pressure.
In the reaction with 6, formation of the κ3 complex is even
less favorable than with 1. This may in part be due to steric
constraints of the PyBOX ligand backbone. In any case, the
net barrier to alkyl loss is seen to be significantly higher for 6
than for 1, in agreement with the fact that we can observe a
cobalt dialkyl complex of 6 but not of 1.
R
Me
CH2SiMe3
(Py)2CoR2 þ TMEDA f
-3.5
þ7.0
-1.4
-1.4
(TMEDA)CoR2 þ 2 Py
(TMEDA)CoR2 þ 1 f
(1)CoR2 þ TMEDA
þ12.7
þ2.8
The overall reaction HS-(κ2-L)CoR2 f LS-(κ3-L)CoR þ R
3
(TMEDA)CoR2 þ 6 f
(6)CoR2 þ TMEDA
is endergonic in all cases. However, alkyl radicals are
extremely reactive and will react with anything they encoun-
ter, making alkyl dissociation effectively irreversible.
Ligand Effects on CoIR Formation. The reactions of
(Py)2CoR2 and (TMEDA)CoR2 with DIMPY and PyBOX
shed some light on the mechanism of CoI alkyl formation for
DIMPY complexes. The original DIMPY ligands 1 and 2 are
strong enough σ-donors to displace pyridine from
(Py)2CoR2 (at least in equilibrium) and strong enough π-
acceptors to then promote homolysis of one Co-C bond.36
With less shielding ligands such as 3 and 4, the same reactions
probably occur, but the initially formed LCoR complexes
decompose. With weaker σ-donor ligand 5, displacement of
pyridine or TMEDA is slower, but the better π-acceptor
qualities of this ligand do not lead to a stable, isolable CoI
alkyl. With the weaker π-acceptor ligand 6, CoI is less
stabilized and loss of alkyl does not occur immediately,
although the observed decomposition of (6)CoR2 may still
involve CoI intermediates.
a Electronic energies calculated using b3-lyp/TZVP; ZPE, enthalpy
and entropy corrections for 298 K, 1 bar, gas phase, taken from SV(P)
calculations.
exchange between the two alternative κ2 structures, i.e., with
the b3-lyp predictions. In particular, the predicted chemical
shift for the pyridine H4 proton is -134 ppm for HS κ2 and
þ163 for LS κ3 (observed: -66.5 ppm).
Ligand Exchange and Formation of CoI Complexes.
(Py)2CoR2 appears to be useful as a source of “CoR2” in
ligand exchange reactions, although some of these appear to
be equilibria: the chelate effect (with TMEDA and 6) does
not seem to be enough to drive reactions to completion.
Calculations (Table 2) confirm that the Py/TMEDA ligand
exchange reaction is close to thermoneutral. The b3-lyp
results also predict the TMEDA/6 ligand exchange equilib-
rium to be close to thermoneutral (ΔGcalc = þ2.8 kcal/mol),
again in agreement with experiment (ΔGobs ≈ þ0.1 kcal/
mol),32 whereas exchange with 1 is considerably uphill
(ΔGcalc = þ12.7 kcal/mol).
On the basis of the experimental and computational results
described above, we arrive at a mechanistic proposal for the
reaction of DIMPY ligands with “CoR2” as summarized in
Scheme 2, involving a ligand exchange equilibrium, easy spin
flip, and finally irreversible loss of alkyl. Of course, variations
where alkyl loss happens at the LS-κ2 stage cannot be excluded
at this point. As well, in reactions of (Py)2CoR2 the decom-
position might involve pyridine adducts (L)(Py)CoR.
Experimentally, ligand exchange leads to formation of CoI
complexes for 1 and 2, but not as easily for 6. A reasonable
mechanism, compatible with our observations, would be
initial formation of a (κ2-L)CoR2 complex, conversion to
(κ3-L)CoR2, and loss of an alkyl radical; this path was
examined in more detail for ligands 1 and 6 using DFT.
Since (Py)2CoR2 and (TMEDA)CoR2 both have S = 3/2 but
the final CoI complex is diamagnetic, a spin flip has to occur
at some point, and this might be relevant to the reaction
profile. Therefore, we located local minima for the various
possible coordination geometries and spin states and also
located a number of minimum-energy crossing points
(MECP)33,34 between the S = 3/2 and S = 1/2 states starting
from points close to the κ2 and κ3 geometries of the DIMPY
complexes. Spin flips turn out to be quite easy in these
systems and do not represent significant barriers in the
energy profile. The final calculated free energy profiles for
ligands 1 and 6 in combination with CoMe2 and CoR2
fragments are shown in Figure 7. In all cases, the lowest
energy path involves a spin flip at the (κ2-L)CoR2 stage.
Forming LS (κ3-1)CoMe2 is then favorable, but formation of
LS (κ3-1)Co(CH2SiMe3)2 is 10 kcal/mol less favorable, pre-
Conclusions
The only literature mention of a pyridine cobalt dialkyl
complex dates back to 1967, when Matsuzaki and Sukawa
reported (Py)2CoMe2 to be extremely unstable;37 it was
characterized only by EPR. We were, therefore, pleasantly
surprised by the stability of the new complex (Py)2Co-
(CH2SiMe3)2, which at this point is by far the most easily
accessible CoII dialkyl. Its availability will make it easier to
do systematic comparisons of dialkyl derivatives along the
series Mn,38 Fe,4 Co, Ni,39 and Zn.40
(35) Thermal corrections, including zero-point energy corrections,
are not simple for crossing points. The MECP “free energies” in the
profile were obtained by combining their electronic energies with the
average of the thermal corrections for the HS and LS minima connected
to the MECP.
(32) In contrast, b-p predicts the equilibrium to be strongly on the
side of low-spin (κ3-6)CoR2. If comparisons are restricted to the high-
spin state, b3-lyp and b-p predictions are rather close.
(33) See e.g.: Harvey, J. N. Spin-Forbidden Reactions in Transition
Metal Chemistry. In Computational Organometallic Chemistry; Cundari,
T. R., Ed.; Marcel Dekker: New York, 2001.
(36) For the analogous Co-C homolysis in CoIII complexes, see e.g.:
(a) De Bruin, B.; Dzik, W. I.; Li, S.; Wayland, B. B. Chem.;Eur. J. 2009,
15, 4312. (b) Li, S.; De Bruin, B; Peng, C.-H.; Fryd, M.; Wayland, B. B. J.
Am. Chem. Soc. 2008, 130, 13373. Note that because of the strong p-
acceptor character of the DIMPY ligand, a picture involving electron transfer
to the ligand prior to homolysis, i.e., involving CoIII, might actually have
some validity.
(37) (a) Matsuzaki, K.; Yasukawa, T. J. Phys. Chem. 1967, 71, 1160.
(b) We checked that our procedure does not work for the synthesis of
(Py)2CoMe2, although color changes observed during the reaction suggest
a transient cobalt dimethyl species may be formed at some stage.
(38) Alberola, A.; Blair, V. L.; Carrella, L. M.; Clegg, W.; Kennedy,
A. R.; Klett, J.; Mulvey, R. E.; Newton, S.; Rentschler, E.; Russo, L.
Organometallics 2009, 28, 2112.
(34) MECP were located between DFT configurations having Sz =
1
/
and 3/2. All are single-determinant functions and do not represent
2
pure spin states. Some “low-spin states”, in particular with the b3-lyp
functional, show significant spin contamination (ÆS2æ values between up
to 1.8). The MECP, therefore, correspond to switches between virtually
pure quartet states and mixtures of doublet and quartet, and the true
crossing barriers might be higher. But even if the true crossing barriers
were twice as high as the ones we calculate, spin flip would still be easy.

1904 Organometallics, Vol. 29, No. 8, 2010
Zhu et al.
Figure 7. Reaction profiles (ΔG, b3-lyp/TZVP, kcal/mol) for CoMe2 and Co(CH2SiMe3)2 fragments bound to ligands 1 and 6. Points
marked ꢀ are minimum-energy crossing points (MECP) between HS and LS states; the others are local minima. MECP free energies
are estimated.35 For ligand 6, points marked κ2* have the oxazoline ring with its oxygen atom toward Co.
Scheme 2. Proposed Mechanism for Formation of
(DIMPY)CoR Complexes
course of the reaction. DFT results indicate that high-spin
states are preferred for all CoII dialkyl complexes; spin flip,
required at some point to eliminate R and form a low-spin
CoI monoalkyl, is predicted to be easy. Comparison of
calculated and observed paramagnetic H NMR shifts ap-
pears to be a useful tool in assigning spin states of complexes
like (6)CoR2 that cannot be obtained in pure form.
1
Experimental Section
General Procedures. All experiments were performed under
an argon atmosphere using standard Schlenk techniques or in a
nitrogen-filled drybox. Pyridine and tetramethylethylenedia-
mine were obtained from Aldrich and dried by distillation from
calcium hydride. Pentane, hexane, toluene, diethyl ether, tetra-
hydrofuran, benzene, and benzene-d6 were distilled from so-
dium/benzophenone. LiCH2SiMe3 was purchased from Aldrich
and crystallized from pentane at -35 ꢀC. Anhydrous CoCl2 and
2-methyl-2-phenylpropylmagnesium chloride solution (0.5 M in
diethyl ether) were purchased from Aldrich and used as received.
Ligands 1-3,41 4,42 and 616 were prepared according to pub-
lished procedures.
1H and 13C NMR spectra were recorded on a Bruker Avance
300 MHz spectrometer. All NMR shifts (δ, ppm) were refer-
enced to the solvent (benzene-d6, 1H NMR: C6D5H δ 7.16 ppm;
13C NMR: C6D6 δ 128.0 ppm; CDCl3, 1H NMR: CHCl3 δ 7.26
ppm; 13C NMR: CHCl3 δ 77.0 ppm). Data were collected at
room temperature unless otherwise noted. Deconvolution and
line width determination for broad peaks were done with the
SpinWorks package.43 Elemental analyses were done by Guelph
Chemical Laboratories Ltd.
(Py)2CoR2, is a new, easily accessible source of transfer-
able “CoR2”. If its noncrystalline nature is an issue, or if the
liberated Py can interfere with further chemistry, (Py)2CoR2
can be converted into crystalline (TMEDA)CoR2, which is
equally useful as a CoR2 synthon. The structure of the related
but less stable complex (Py)2Co(CH2CMe2Ph)2 was con-
firmed by X-ray diffraction. Transfer of the CoR2 fragment
to DIMPY-type ligands is not straightforward and illustrates
the delicate interplay between σ-donation, π-back-donation,
steric effects, and the choice of labile ligand in affecting the
2,6-(CF3CO)2C5H3N. In a 250 mL three-necked flask with Ar
connection, dropping funnel, and septum equipped with an
internal temperature sensor was placed 5 g (21.1 mmol) of
ꢀ
(39) (a) Carmona, E.; Gonzalez, F.; Poveda, M. L.; Atwood, J. L.;
ꢀ
Roger, R. D. J. Chem. Soc., Dalton Trans. 1981, 777. (b) Campora, J.; del
(41) Fan, R.-Q.; Zhu, D.-S.; Mu, Y.; Li, G.-H.; Yang, Y.-L.; Su, Q.;
Feng, A.-H. Eur. J. Inorg. Chem. 2004, 4891.
ꢀ
Mar Conejo, M.; Mereiter, K.; Palma, P.; Perez, C.; Reyes, M. L.; Ruiz, C. J.
€
(42) Lappalainen, K.; Yliheikkila, K.; Abu-Surrah, A. S.; Polamo,
Organomet. Chem. 2003, 683, 220.
(40) Moorhouse, S.; Wilkinson, G. J. Chem. Soc., Dalton Trans. 1974,
€
M.; Leskela, M.; Repo, T. Z. Anorg. Allg. Chem. 2005, 631, 763.
2187.
(43) Marat, K. SpinWorks 3.1; University of Manitoba, 2009.

Article
Organometallics, Vol. 29, No. 8, 2010 1905
2,6-dibromopyridine in 56 mL of diethyl ether and 44 mL of
THF. The magnetically stirred mixture was cooled in a methanol/
liquid nitrogen cooling bath to an internal temperature of -95 ꢀC
(dibromopyridine starts to precipitate). Then 13.6 mL of 1.6 M n-
BuLi in hexanes (21.6 mmol) was added in about 10 min, keeping
the temperature of the mixture below -90 ꢀC. A 3 mL amount of
diethyl etherwas usedtorinsethe wallofthe droppingfunnel, and
this was added to the mixture. Warming of the mixture to -90 ꢀC
resulted in dissolution of the precipitated dibromopyridine, giv-
ing a clear yellow solution. After an additional 5 min of stirring
at -90 ꢀC, the mixture was cooled to -100 ꢀC and 2.74 g
(21.4 mmol) of methyl trifluoroacetate was added over 10 min,
keeping the temperature below -90 ꢀC. Some ether was used to
rinse the funnel. The mixture was stirred for 20 min at -90 ꢀC,
then cooled to -113 ꢀC, and 13.6 mL of 1.6 M n-BuLi was added
while the temperature was kept below -100 ꢀC. After 5 min of
stirring, 2.8 g (21.8 mmol) of methyl trifluoroacetate was added
while keeping the temperature below -100 ꢀC.
(Py)2CoR2. Pink (Py)4CoCl2 (0.32 g, 0.72 mmol) was trans-
ferred into a 50 mL Schlenk tube, and 10 mL of pentane was
added to form a blue suspension. The resulting mixture was
cooled to -70 ꢀC and kept at this temperature for 20 min.
LiCH2SiMe3 (0.136 g, 1.44 mmol) was weighed and dissolved in
10 mL of hexane; this solution was added dropwise to the above
blue suspension at -70 ꢀC. The color of the mixture changed to
blue-green. It was kept at -70 ꢀC for another 20 min and was
then allowed to slowly (3 h) warm to room temperature. During
this, the color first changed to red (around -10 ꢀC), and then at
around 0 ꢀC most of the solid dissolved to form a yellow-green
solution with some suspended white solid. This was stirred for
another hour at room temperature and filtered through a glass
frit. Solvents were removed in vacuo, leaving a thick bright green
oil. Addition of a small amount of hexamethylbenzene as an
internal standard allowed determination of the yield by NMR:
75%. For most subsequent experiments, a single portion of
(Py)2CoR2 was prepared and used assuming this 75% yield.
1H NMR (25 ꢀC, benzene-d6, 300 MHz): δ 114 (4H, Δν1/2
1600 Hz, Py H2), 38.4 (4H, Δν1/2 500 Hz, Py H3), 10.3 (18H,
Δν1/2 80 Hz, SiMe3), -8.5 (2H, Δν1/2 190 Hz, Py H4); CH2(Co)
not observed.
The mixture was allowed to warm, and when it reached
-60 ꢀC, a mixture of 7 mL of 12 M HCl and 5 mL of water was
added. The reaction mixture was then poured into a mixture of
25 mL of 12 M HCl and 150 mL of water. The organic layer
was separated and dried with magnesium sulfate, and the
solvent was evaporated in vacuo. The crude product was
refluxed in rotary vane pump vacuum for 1 h (bath 120 ꢀC,
small flask, short condensor) and then solidified on cooling.
Short-path distillation (the product solidifies in the receiver)
gave 4.47 g (82%) of the colorless product, which according to
NMR still contained ca. 14% of the monohydrate. This
mixture was used without further purification for the synthesis
of 5.
Evans Method for Determination of Magnetic Moment. 13 In a
drybox, one drop of (Py)2CoR2 was dissolved in about 0.5 mL of
C6D6 in a small vial, and 0.0253 g of C6Me6 (reference and internal
standard) was added. The resulting green solution was diluted
to 3.1 mL by adding C6D6. Integration of the 1H NMR spectrum
showed the concentration of (Py)2CoR2 to be 0.22 mol/L; the
chemical shift difference between C6Me6 in the paramagnetic
solution and in pure C6D6 was found to be 9.14 ppm, implying a
magnetic moment of 4.9 μB Two further experiments at different
concentrations produced similar values (0.66 mol/L: 4.5 μB;
0.052 mol/L: 5.0 μB); average 4.8(3) μB.
Hydrolysis. In a drybox, 0.158 g of (Py)2CoR2 and 0.053 g of
C6Me6 were weighed into a small vial. Part of this sample was
used for a 1H NMR spectrum in C6D6. From the observed molar
ratio of (Py)2CoR2 and C6Me6 (1.13:1, by integration), the
purity was calculated to be 91%. To the NMR sample was
added 0.05 mL of water, the resulting suspension was quickly
filtered, and the 1H NMR spectrum of the filtrate was recorded.
The only visible signals were due to pyridine, tetramethylsilane,
and hexamethylbenzene. The observed molar ratio of the Py,
Me4Si, and C6Me6 species in solution (1.97:2.12:1, by integra-
tion; expected 2.26:2.26:1) agreed reasonably well with the
expected values for (Py)2CoR2, considering that some Py will
remain coordinated to Co and/or stay in the water droplet and
some Me4Si will evaporate during filtration.
1H NMR (300.1 MHz, CDCl3, 25 ꢀC): δ 8.43 (2H, d, 3J
7.9 Hz, Py H3), 8.27 (1H, t, 3J7.9 Hz, Py H4). 13CNMR(75.5MHz,
2
CDCl3, 25 ꢀC): δ 179.7 (q, JCF 35 Hz, CdO), 147.9 (Py C2),
139.2 (Py C4), 129.1 (Py C3), 116.4 (q, 1JCF 290 Hz, CF3). 19
NMR (282.4 MHz, CDCl3, 25 ꢀC): δ -73.1.
F
Before refluxing and distillation, the crude mixture contained
only a small amount of 2,6-(CF3CO)2C5H3N; the main compo-
nents were its bis(hemiketal) (mixture of two diastereomers),
hemiketal/hydrate, bis(hydrate), monohemiketal, and monohy-
drate, causing the NMR spectra to look messy. Crystals of the
bis(hydrate), suitable for X-ray diffraction, crystallized sponta-
neously from the THF solution of an NMR sample.
2,6-[2,4,6-Me3C6H2NdC(CF3)]2C5H3N (5). (This synthesis
does not require an inert atmosphere.) The crude bis-
(trifluoroacetyl)pyridine mixture described above was dissolved
in 50 mL of toluene, 4.3 g of 2,4,6-trimethylaniline and a small
quantity of p-toluenesulphonic acid were added, and the orange
mixture was refluxed in a Dean-Stark apparatus for 72 h. The
toluene was removed using a rotary evaporator. The residue was
crystallized from hexane/toluene (50 mL, ca. 2:1, 85 ꢀC/RT) to
give 6.23 g (78%) of yellow crystals. A crystal from this batch
was used for X-ray structure determination.
(TMEDA)CoR2. (Py)2CoR2 (0.82 mmol, assuming 75% yield
from (Py)4CoCl2; see above) was dissolved in 5 mL of pentane,
and the green solution was cooled to -30 ꢀC. TMEDA (10 equiv,
1.25 mL, 8.2 mmol) was slowly added at -30 ꢀC, resulting in a
color change to first red and then brown. After filtering, the
solution was evaporated to dryness. The residue was dissolved
in 2 mL of pentane, 1 drop of TMEDA was added, and the
solution was cooled to -35 ꢀC, depositing a dark solid. The
cold liquid was pipetted off, and the solid was dried to give a
purple-blue solid (0.22 g, 57% yield based on (Py)4CoCl2 or
76% yield based on (Py)2CoR2). Deep bluish-purple crystals
suitable for X-ray diffraction were grown by cooling the
saturated pentane solution at -35 ꢀC overnight.
1H NMR (300.1 MHz, CDCl3, 25 ꢀC): δ 7.42 (1H, t, 3J 7.9 Hz,
Py H4), 6.93 (2H, d, 3J 7.9 Hz, Py H3), 6.75 (4H, s, Ar m), 2.22
(6H, s, Ar p-Me), 1.90 (12H, s, Ar o-Me). 13C NMR (75.5 MHz,
2
CDCl3, 25 ꢀC): δ 154.6 (q, JCF 33 Hz, CdN), 149.1 (Py C2),
142.3 (Ar i), 136.6 (Py C4), 134.1 (Ar p), 128.8 (Ar m), 124.8 (Ar o),
1
124.2 (Py C3), 119.2 (q, JCF 280 Hz, CF3), 20.6 (Ar p-Me),
17.5 (Ar o-Me). 19F NMR (282.4 MHz, CDCl3, 25 ꢀC): δ -68.6.
Anal. Calcd for C27H25F6N3 (505.50): C, 64.15; H, 4.98; N, 8.31.
Found: C, 64.34; H, 5.30; N 8.40.
1H NMR (25 ꢀC, benzene-d6, 300 MHz): δ 80 (4H, Δν1/2
600 Hz, CH2N), 78 (12H, Δν1/2 350 Hz, NMe), 9.6 (18H, Δν1/2
70 Hz, SiMe3); CH2(Co) not observed.
(Py)2CoR02. A number of procedures were tried for the
synthesis of this complex, but none gave a completely pure
product. Procedure A below was used to generate the mixture of
(Py)2MgR02 and (Py)2CoR02, from which a crystal was used for
the X-ray structure determination mentioned in the text. Pro-
cedure B gave a better quality product for which the NMR
spectrum could be assigned with some confidence.
(Py)4CoCl2. Anhydrous CoCl2 (1.36 g, 10.4 mmol) was
transferred into a 100 mL Schlenk tube, and 15 mL of pyridine
was added. The resulting suspension of initially blue solid in a
pink solution was stirred overnight at room temperature, during
which the solid became pink. The solid was filtered off and dried
in vacuo, giving 3.65 g (78%) of pink crude (Py)4CoCl2. Anal.
Calcd for C20H20Cl2CoN4 (446.24): C, 53.83; H, 4.52; N, 12.56;
Cl, 15.89. Found: C, 53.67; H, 4.85; N, 12.81; Cl, 16.10.

1906 Organometallics, Vol. 29, No. 8, 2010
Zhu et al.
Procedure A. Pink (Py)4CoCl2 (1.24 g, 2.78 mmol) was
transferred into a 50 mL Schlenk tube, and 26 mL of dry diethyl
ether was added to form a blue suspension. The resulting
mixture was cooled to -50 ꢀC. R0MgCl (11.8 mL of a 0.5 M
solution in diethyl ether, 5.90 mmol) was added dropwise. The
solution quickly turned green, and a white solid was formed. The
stirred mixture was slowly (in 30 min) warmed to room tem-
perature. Then 1.5 mL of dry pyridine was added, and the
resulting green suspension was stirred at room temperature for
another 30 min and filtered through a glass frit. The solvents
were removed in vacuo, the residue was transferred into a
drybox, dry pentane was added, and the suspension was filtered
through glass wool. On standing overnight at -35 ꢀC, the
filtrate deposited whitish crystals of (presumably) (Py)2MgR02
(see below), dark green blocks of (Py)2CoR02, and some sticky
dark oily droplets. A fragment of one of the dark green blocks
hour, the deep purple solution was filtered and the solvent was
evaporated in vacuo. The residue was dissolved in 4 mL of
diethyl ether/pentane (1:1), the solution was concentrated to
2 mL, 1 mL of hexane was added, and the mixture was cooled to
-35 ꢀC overnight. The purple mother liquor was decanted, and
the residue was recrystallized from toluene (6 drops)/hexane
(5 mL) to give shiny dark crystalline (1)CoR (0.23 g, 42%)
suitable for X-ray diffraction.
1H NMR (25 ꢀC, benzene-d6, 300 MHz): δ 10.08 (t, 1H, J
7.7 Hz, Py H4), 7.74 (d, 2H, J 7.7 Hz, Py H3), 7.36 (t, 2H, J 7.4
Hz, Ar p), 7.27 (d, 4H, J 7.4 Hz, Ar m), 2.12 (s, 12H, Ar Me), 0.81
(s, 2H, CoCH2), -0.62 (s, 9H, SiMe3), -1.26 (s, 6H, NdCMe).
13C NMR (25 ꢀC, benzene-d6, 75 MHz): δ 165.6 (NdC), 158.4
(Py C2), 156.0, 130.2, 129.1, 125.9, 123.3, 116.7, 23.8 (NdCMe),
19.5 (Ar Me), 3.8 (SiMe3), -10.7 (br, CoCH2). Anal. Calcd for
C29H38CoN3Si (515.65): C, 67.55; H, 7.43; N, 8.15. Found: C,
67.32; H, 7.14; N, 7.70.
1
was used for single-crystal X-ray structure determination. H
NMR of the deposited solids in C6D6 indicated the presence of
both (Py)2MgR02 and (Py)2CoR02 (see below).
(2)CoR. This reaction was carried out as described for 1, using
0.9 mmol of (Py)2CoR2 and 0.35 g (0.88 mmol) of 2. After
evaporation of the solvents, the resulting thick purple oil was
dissolved in pentane and filtered through glass wool. Slow
evaporation produced a sticky product (0.26 g, crude yield:
56%). All attempts at crystallization failed.
Procedure B. Pink (Py)4CoCl2 (0.69 g, 1.55 mmol) was
transferred into a 100 mL Schlenk tube, and 26 mL of dry
THF was added to form a blue clear solution. This was cooled to
-50 ꢀC, during which a pink solid precipitate formed. R0MgCl
(2.8 mL of a 0.5 M solution in diethyl ether, 1.4 mmol) was
added dropwise. The resulting mixture was warmed to around
0 ꢀC in 1 h, during which it turned purplish and a white solid was
formed. The solution was again cooled to -50 ꢀC, and R0MgCl
(3.2 mL of a 0.5 M solution in diethyl ether, 1.6 mmol) was
added dropwise. The resulting suspension was stirred at -50 ꢀC
for 20 min and warmed to room temperature in 1 h, during
which it turned green. Then 0.6 mL of dry pyridine was added,
and the green mixture was cooled to -30 ꢀC and concentrated to
about 2 mL. A 20 mL portion of dry pentane was added, the
solution was filtered through a glass frit, and the filtrate was
cooled to -35 ꢀC. Overnight, a dark and shiny solid crystallized;
the mother liquid was pipetted off, leaving 0.15 g (20%) of
reasonably pure (Py)2CoR02. NMR samples always show some
decomposition to diamagnetic compounds (pyridine, PhCMe3),
and within 4 h all signals due to (Py)2CoR02 disappear.
1H NMR (25 ꢀC, benzene-d6, 300 MHz): δ 10.13 (t, 1H, J
7.0 Hz, Py H4), 7.79 (d, 2H, J 7.0 Hz, Py H3), 7.11 (s, 4H, Ar m),
2.37 (s, 6H, Ar p-Me), 2.16 (s, 12H, Ar o-Me), -0.60 (s, 9H,
SiMe3), -1.14 (s, 6H, NdCMe). The CoCH2 signal could not be
unambiguously assigned.
Reaction of (TMEDA)CoR2 with 3. In a drybox, 5.2 mg (17
μmol) of 3 was dissolved in benzene-d6 and transferred into an
NMR tube, and a solution of 6.2 mg of (TMEDA)CoR2 (18
1
μmol, 1.1 equiv) in benzene-d6 was added. For the H NMR
spectrum, see Figure S4A.
1H NMR (25 ꢀC, benzene-d6, 300 MHz; tentative assign-
ments): δ 9.79 (t, 1H, J 7.4 Hz, Py H4), 8.17 (d, 4H, J 6.6 Hz, Ar
o), 7.76 (d, 2H, J 7.2 Hz, Py H3), 7.35-7.43(m, 6H, Ar m and Ar
p), 0.35 (s, 2H, CoCH2), -0.64 (s, 9H, SiMe3), -1.10 (s, 6H,
NdCMe).
Reaction of (Py)2CoR2 with 3. In a drybox, about 59 μmol of
(Py)2CoR2 was dissolved in benzene-d6 and 0.015 g (48 μmol;
0.81 equiv) of 3 was added. A 1H NMR spectrum recorded
immediately afterward (Figure S4B,C) showed peaks that we
tentatively assign in terms of formation of 1 equiv of Me4Si and 1
equiv of (3)CoR or (3)(Py)CoR.
1H NMR (25 ꢀC, benzene-d6, 300 MHz; assignment
tentative): δ 108 (4H, Δν1/2 5300 Hz, Py H2), 32.7 (4H, Δν1/2
1700 Hz, Py H3), 21.6 (12H, Δν1/2 270 Hz, CMe2), 10.2 (4H,
Δν1/2 160 Hz, Ar o), 7.4 (4H, Δν1/2 27 Hz, Ar m), 3.9 (2H,
Δν1/2 17 Hz, Ar p), -8.1 (2H, Δν1/2 590 Hz, Py H4).
Formation of (Py)2Mg(CH2CMe2Ph)2. Under a nitrogen
atmosphere, 5 mL of dry pyridine was dissolved in 15 mL of
dry pentane, and 5 mL of 2-methyl-2-phenylpropylmagnesium
chloride solution (0.5 M in diethyl ether) was added, resulting in
precipitation of a white powder. The resulting yellow solution
with white suspended solid was stirred for another 2 h at room
temperature and filtered through a glass frit. The filtrate was
evaporated to dryness, and the resulting yellow sticky oil was
dissolved in toluene and layered with pentane at -35 ꢀC over-
night, forming a yellow crystalline solid (0.45 g, 80%).
1H NMR (25 ꢀC, benzene-d6, 300 MHz; tentative assign-
ments): δ 8.71 (br, 1H, Py H4), 8.37 (br, Py H3), 7.00 (4H, Ph o),
6.91 (4H, Ph m), 6.82 (2H, Ph p), 0.85 (s, 6H, MeCdN), 0.45 (s,
2H, CoCH2), 0.05 (s, 12H, Me4Si), -0.50 (s, 9H, SiMe3).
Reaction of (TMEDA)CoR2 with 4. In a drybox, 6.8 mg of 4
(20 μmol) was weighed and dissolved in benzene-d6, followed by
addition of around 9.2 mg of (TMEDA)CoR2 (26 μmol, 1.3
1
equiv) in benzene-d6. A H NMR spectrum recorded immedi-
ately (Figure S5A) showed peaks that could be tentatively
assigned to (4)CoR.
1H NMR (25 ꢀC, benzene-d6, 300 MHz): δ 7.90 (d, 4H, J
4.0 Hz, Py H2), 7.78 (d, 4H, J 7.7 Hz, Ar o), 7.12 (t, 4H, J 7.7 Hz,
Ar m), 6.99 (t, 2H, J 7.2 Hz, Ar p), 6.78 (t, 2H, J 7.4 Hz, Py H4),
6.41 (t, 4H, J 6.4 Hz, Py H3), 1.83 (s, 12H, Me), 0.77 (s, 4H,
CH2). 13C NMR (25 ꢀC, benzene-d6, 75 MHz): δ 158.3 (Ar i),
148.82 (Py C2), 137.69 (Py C4), 127.80 (Ar m), 125.92 (Ar o),
124.21 (Py C4), 123.91 (Ar, p), 40.41 (CMe2), 36.62 (CMe2),
34.48 (CH2).
1H NMR (25 ꢀC, benzene-d6, 300 MHz; tentative assign-
ments): δ 9.74 (t, 1H, J 7.4 Hz, Py H4), 7.40 (d, 2H, J 7.4 Hz, Py
H3), 6.61 (s, 4H, PhCH2), -0.16 (s, 9H, SiMe3), -0.64 (s, 6H,
NdCMe).
Pyridine was added to this sample in small increments; this
resulted in changes in shifts of the 1H NMR signals attributed to
the CoI alkyl. Figure S5D shows the spectrum after addition of
3 μL of Py.
(1)CoR. (Py)2CoR2 (1.14 mmol assuming 75% yield from
(Py)4CoCl2; see above) was dissolved in 8 mL of diethyl ether,
and the green solution was cooled to -30 ꢀC. Ligand 1 (0.39 g,
1.06 mmol) was dissolved in 20 mL of diethyl ether, and the clear
yellow solution was slowly added into the above (Py)2CoR2
solution at -30 ꢀC. After the addition, the solution (which had
turned red) was allowed to warm to room temperature (around
0 ꢀC the color changed to purple). After stirring for another
Reaction of (Py)2CoR2 with 4. In a drybox, about 6.8 mg
(20 μmol) of 4 was added to a solution of about 40 μmol of
(Py)2CoR2 in benzene-d6. A 1H NMR spectrum recorded
immediately afterward (Figure S5C) showed peaks that could
be interpreted in terms of formation of 1 equiv of Me4Si and
1 equiv of (4)CoR or (4)(Py)CoR.
1H NMR (25 ꢀC, benzene-d6, 300 MHz; tentative assign-
ments): δ 9.32 (t, 1H, J 6.6 Hz, Py H4), 7.54 (d, 2H, J 6.6 Hz, Py

Article
Organometallics, Vol. 29, No. 8, 2010 1907
Table 3. Details of Crystal Structure Determinations
2,6-[CF3C(OH)2]2C5H3N
3
THF
5
(TMEDA)CoR2
(Py)2CoR02
(1)CoR
formula
mol wt
C9H7NO4F6 C4H8O
C27H25F6N3
505.50
tetragonal
I41
14.9916(10)
14.9916(10)
11.1077(5)
90
90
90
2496.4(3)
4
1.345
C14H38CoN2Si2
349.57
monoclinic
C2/c
25.565(3)
10.1487(10)
17.3758(18)
90
100.104(6)
90
4438.3(8)
8
1.046
0.875
1528
C30H36CoN2
483.54
triclinic
P1
11.8183(11)
11.8294(12)
12.1892(12)
105.652(2)
109.119(2)
108.825(2)
1383.1(2)
2
C29H38CoN3Si
515.65
3
379.26
orthorhombic
Pcmn
6.6689(3)
14.6578(8)
17.1651(12)
90
cryst syst
space group
monoclinic
P21/m (No. 11)
7.6802(3)
21.2770(10)
8.5207(4)
90
99.1876(11)
90
1374.52(11)
2
1.246
˚
a/A
˚
b/A
˚
c/A
R/deg
β/deg
γ/deg
90
90
3
˚
V/A
1677.91(17)
4
1.501
0.155
776
Z
Dc/g cm-3
abs coeff/mm-1
F000
1.161
0.639
514
0.111
1048
0.689
548
index ranges
-7 e h e 7
-17 e k e 17
-20 e l e 16
50
21 981
1531
-17 e h e 17
-17 e k e 17
-13 e l e 13
50
14 675
2193
-30 e h e 30
-12 e k e 12
-20 e l e 21
51
13 479
4140
-14 e h e 14
-14 e k e 14
-14 e l e 14
51
10 195
5166
-9 e h e 9
-25 e k e 25
-10 e l e 10
51
9173
2634
2θmax/deg
no. reflns
no. unique
no. > 2σ
GOF
no. params
R (Fo > 4σ(Fo)
R (all data)
wR2 (all data)
1116
1.086
126
0.0623
0.0916
0.1446
0.479, -0.272
1925
1.074
167
0.0358
0.0455
0.0853
0.111, -0.157
3163
1.054
186
0.0374
0.0496
0.1090
0.266, -0.171
4314
1.088
324
0.0685
0.0772
0.2060
0.555, -0.241
2459
1.075
174
0.0379
0.0398
0.1008
0.393, -0.174
-3
˚
largest peak, hole/e A
H3), 7.05-7.08 (br, Ar o), 6.90-6.93 (m, Ar p and Ar m), 6.32
(s, 4H, PhCH2), -0.15 (s, 6H, NdCMe), -0.24 (s, 9H, SiMe3).
CoCH2 could not be unambiguously assigned.
X-ray Structure Determinations (see also Table 3). General
Procedures. All measurements were done using Mo KR radia-
˚
tion (0.71073 A). Semiempirical absorption was done using
SADABS.44 Structures were refined using full-matrix least-
squares refinement on F2 with SHELXL97;45 hydrogen atoms
were placed at calculated positions and refined in riding
mode. Structures were checked for solvent-accessible voids
with PLATON.46
Figure S5B shows the results of a similar experiment using
different concentrations of 4 and (Py)2CoR2, resulting in sig-
nificantly different chemical shifts of the CoI alkyl.
1H NMR (25 ꢀC, benzene-d6, 300 MHz; tentative assign-
ments): δ 8.94 (br, 1H, Py H4), 7.63 (br, Py H3), 6.95, 6.89 (br,
10H, Ph), 6.08 (4H, NCH2), 0.25 (s, 6H, MeCdN), 0.12 (s, 2H,
CoCH2), 0.02 (s, 12H, Me4Si), -0.32 (s, 9H, SiMe3).
2,6-[CF3C(OH)2]2C5H3N. A transparent regular platelet
(approximately 0.3 ꢀ 0.2 ꢀ 0.1 mm) was glued on a thin glass
fiber. Data were collected at 208 K on a Nonius KappaCCD
diffractometer with area detector j and ω scans. The struc-
ture was solved using CRUNCH.47 There is some minor rota-
tional disorder in the CF3 groups, resulting in somewhat
larger anisotropic thermal displacement parameters. A cocrys-
tallized THF solvent molecule is situated in a mirror plane,
implying the presence of some disorder, which is reflected in
large thermal parameters perpendicular to the plane of the
molecule. This disorder could not be described adequately,
and attempts to do so resulted in physically unacceptable
geometrical parameters. A thermal ellipsoid plot is provided in
the Supporting Information.
2,6-[2,4,6-Me3C6H2NdC(CF3)]2C5H3N (5). A transparent
light yellow crystal fragment (approximately 0.2 mm in all
directions) was glued on a thin glass fiber. Data were collected
at room temperature on a Nonius KappaCCD diffractometer
with area detector j and ω scans. The structure was solved using
CRUNCH.47
Reaction of (Py)2CoR2 with 5. In a drybox, 6.9 mg (17 μmol) of
5 was weighed and dissolved in benzene-d6, followed by addition
about 28 μmol of (Py)2CoR2 in benzene-d6. 1H and 19F NMR
spectra were recorded immediately (Figures S6B, S7B) and
showed the formation of Me4Si and a small amount of a new
diamagnetic compound, possibly (5)CoR or its Py adduct, as
well as a large amount of both starting materials.
1HNMR(25ꢀC, benzene-d6, 300 MHz; tentative assignments): δ
8.56 (d, 2H, J 7.9, Py H3), 2.36 (s, 6H), 0.34 (s, 2H, CoCH2), -0.35
(s, 9H, SiMe3). 19F NMR (282.4 MHz, benzene-d6, 25 ꢀC): δ-54.1.
On standing, a black solid precipitated from the sample. A
spectrum (Figure S6C) recorded after 2 h still showed the
presence of unreacted starting materials, although the amount
of free 5 had decreased. Then 0.5 mL of water was injected, and
after separation of layers spectra taken of the benzene-d6 layer
showed mainly free 5 (Figure S6D).
Formation of (6)CoR2. (TMEDA)CoR2 (0.082 g, 0.23 mmol)
was dissolved in 2 mL of benzene and cooled to -35 ꢀC, and
0.06 g (0.23 mmol) of 6 in 2 mL of benzene was added. After
warming to room temperature, the resulting blue solution was
stirred for 10 min and filtered through glass wool, solvents were
evaporated, 2 mL of pentane was added, and the solution was
again filtered through glass wool. Cooling failed to produce
crystals, so the solvents were evaporated and the crude product
(a mixture of (6)CoR2 and (TMEDA)CoR2, see text and
Figure 5) was characterized by 1H NMR.
(1)CoR. A deep purple crystal fragment (ca. 0.6 ꢀ 0.2 ꢀ
0.1 mm) was mounted in a thin-walled glass capillary. Data were
collected at 293 K on a Bruker 4-circle diffractometer with
APEX detector. The crystal system and space group were
(44) Sheldrick, G. M. SADABS, Program for Empirical Absorption
€
Correction; University of Gottingen: Germany, 1996.
(45) Sheldrick, G. M. SHELXL-97, Program for the Refinement of
Crystal Sturctures; University of G€ottingen: Germany, 1997.
(46) Spek, A. L. PLATON. A Multipurpose Crystallographic Tool;
Utrecht University: Utrecht, The Netherlands, 2003.
(47) De Gelder, R.; De Graaff, R. A. G.; Schenk, H. Acta Crystallogr.
1993, A49, 287.
1H NMR (25 ꢀC, benzene-d6, 300 MHz): δ 44.4 (2H, br, Δν1/2
80 Hz, Py H3), 21.8 (18H, br, Δν1/2 120 Hz, SiMe3), -5.6 (4H,
br, CH2O), -16.9 (12H, br, Δν1/2 110 Hz, CMe2), -66.5 (1H, br,
Py H4).

1908 Organometallics, Vol. 29, No. 8, 2010
Zhu et al.
determined from the cell metric and systematic absences. Data
integration was performed using SAINT.48 The structure was
solved by the Patterson method with SHELXS97.49
basis sets and the functionals b3-lyp52 and b-p53 (the latter with
the RI approximation) in combination with an external optimi-
zer (PQS OPTIMIZE54 for minima, BOptimize55 for MECP33).
All calculations used the spin-unrestricted formalism, with Sz =
3/2 (CoII HS), 1/2 (CoII LS), or 0 (Co(0) LS). Vibrational analyses
were carried out for all stationary points to confirm their nature
and also to calculate thermal corrections (enthalpy and entropy,
gas phase, 298 K, 1 bar) and obtain free energies using the
standard formulas of statistical thermodynamics.56 For the
TZVP/b3-lyp calculations, calculation of analytical Hessians
proved to be prohibitively expensive, so we here combine the
TZVP optimized electronic energies with SV(P)-level thermal
corrections evaluated at the SV(P) optimized geometries.
The σ- and π-parameters of ligand 6 were generated according
to the procedure in our previous paper.15 Calculations of the
orbital contributions δorb to the chemical shift were done using
Gaussian 0329 (B3LYP functional, TZVP basis set, GIAO
method;57 same Sz as used for geometry optimizations). EPR
hyperfine couplings used to estimate Fermi contact shifts δFC
(eq 2) were calculated58 using the Orca program31 (B3LYP
functional, TZVP basis set).
(TMEDA)CoR2. The procedure was the same as for (1)CoR.
A deep bluish-purple crystal fragment (ca. 0.5 ꢀ 0.2 ꢀ 0.15 mm)
was mounted in a thin-walled glass capillary. Data were col-
lected at 293 K on a Bruker 4-circle diffractometer with APEX
detector. The crystal system and space group were determined
from the cell metric and systematic absences. Data integration
was performed using SAINT.48 The structure was solved by the
Patterson method with SHELXS97.49 A thermal ellipsoid plot is
provided in the Supporting Information. This is not a new
structure: the structure of the complex was originally deter-
mined by Hay-Motherwell,7 and we find the same unit cell and
coordinates. However, these authors did not report any char-
acterization data other than the X-ray structure, so we used our
structure determination as confirmation of the identity of the
complex.
(Py)2CoR02. A large irregular crystal fragment (ca. 0.6 ꢀ 0.5 ꢀ
0.4 mm) was broken off an even larger block of green-black
(Py)2CoR02, picked from the results of a crystallization attempt
(nearly colorless crystals of (Py)2MgR02 were also present). The
fragment was mounted in a thin-walled glass capillary. Data
were collected using a 0.30ꢀ scan width, a 15 s scan time, and
full-sphere coverage. The large scan width and fast scan time, as
well as the large size of the crystal, had been chosen to allow
Acknowledgment. We are grateful to J. M. Malinoski/
M. Brookhart (University of North Carolina at Chapel
Hill) for providing us with the synthetic procedure for
ligand 5 and to R. de Gelder, J. M. M. Smits (Radboud
University Nijmegen), and M. Cooper and F. Hawthorne
(University of Manitoba) for help with X-ray structure
determinations. This work was supported by NSERC,
CFI, and MRIF grants (to P.H.M.B. and F.H.) and a
University of Manitoba Graduate Fellowship (to D.Z.).
ꢀ
rapid data collection, since Campora reported that the iron
analogue decayed during data collection. Indeed, after about
75% of the full-sphere data were collected the crystal had
visibly changed shape and did not diffract any more. Analysis
of the diffraction data showed that at completion of the first
hemisphere of data decay was less than 15%. Therefore, only
this first hemisphere of data was processed and used in the
refinement. The data integration was performed using
SAINT48 with decay correction, and absorption correction
was performed by SADABS. The structure was solved by the
Patterson method using SHELXS97.49 Even though the final
thermal parameters and error margins look acceptable, the data
should be treated with caution because of the crystal decay
during data collection.
Supporting Information Available: Data for calculation of the
equilibrium constant involving TMEDA/6 exchange; thermal
ellipsoid plots for 2,6-[(HO)2C(CF3)]2C5H3N and (TMEDA)-
CoR2; coordinates (xyz files) and energies for calculated struc-
tures; X-ray crystallographic information (cif file) for the five
structure determinations described in the text. This material is
available free of charge via the Internet at http://pubs.acs.org.
Computational Methods. All geometry optimizations were
carried out with Turbomole25 using the SVP50 and TZVP51
(53) (a) Becke, A. D. Phys. Rev. A 1988, 38, 3089. (b) Perdew, J. P.
Phys. Rev. B 1986, 33, 8822.
(54) (a) PQS version 2.4; Parallel Quantum Solutions: Fayetteville, AR,
2001 (the Baker optimizer is available separately from PQS upon request).
(b) Baker, J. J. Comput. Chem. 1986, 7, 385.
(55) Budzelaar, P. H. M. J. Comput. Chem. 2007, 28, 2226.
(56) Frequencies below a threshold of 11 cm-1 were smoothly
adjusted to a value remaining above 10 cm-1 to avoid artificial “blow-
up” of the entropic correction for very low frequencies, for which the
harmonic oscillator model is no longer appropriate.
(48) SAINT, Bruker Smart program suite V6; Bruker.
(49) Sheldrick, G. M. SHELXS-97, Program for Crystal Structure
€
Solution; University of Gottingen: Germany, 1997.
€
(50) Schafer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97,
2571.
(51) Schafer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys. 1994, 100,
5829.
€
(52) (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. (d) All Turbomole calculations were performed using
the Turbomole functionals “b3-lyp” and “b-p”, which are close but not
identical to the Gaussian “B3LYP” and “BP86” functionals.
(57) (a) London, F. J. Phys. Radium, Paris 1937, 8, 397. (b) McWeeny,
R. Phys. Rev. 1962, 126, 1028. (c) Ditchfield, R. Mol. Phys. 1974, 27, 789.
(d) Dodds, J. L.; McWeeny, R.; Sadlej, A. J. Mol. Phys. 1980, 41, 1419.
(e) Wolinski, K.; Hilton, J. F.; Pulay, P. J. Am. Chem. Soc. 1990, 112, 8251.
(58) Neese, F. J. Chem. Phys. 2003, 117, 3939.