Experimental and computational study of β-H transfer between cobalt (I) alkyl complexes and 1-alkenes

Article abstract of DOI:10.1021/om049581h

Bis(imino)pyridine cobalt(I) alkyl complexes react with 1-alkenes by β-hydrogen transfer, providing a model reaction for the study of a commonly encountered chain transfer process in polymerization and oligomerization catalysis. The influence of steric effects on reaction rates is described. The theoretical models largely agree with the experimentally determined structures, provide a more detailed view of the species involved, and are consistent with the observed reactivities. Both experiment and theory support a stepwise pathway involving a cobalt-hydride intermediate.

Full text of DOI:10.1021/om049581h

Organometallics 2004, 23, 5503-5513
5503
Experimental and Computational Study of â-H Transfer
between Cobalt(I) Alkyl Complexes and 1-Alkenes
Kilian P. Tellmann, Martin J. Humphries, Henry S. Rzepa,* and
Vernon C. Gibson*
Department of Chemistry, Imperial College London, Exhibition Road, London SW7 2AZ, U.K.
Received June 8, 2004
Bis(imino)pyridine cobalt(I) alkyl complexes react with 1-alkenes by â-hydrogen transfer,
providing a model reaction for the study of a commonly encountered chain transfer process
in polymerization and oligomerization catalysis. The influence of steric effects on reaction
rates is described. The theoretical models largely agree with the experimentally determined
structures, provide a more detailed view of the species involved, and are consistent with the
observed reactivities. Both experiment and theory support a stepwise pathway involving a
cobalt-hydride intermediate.
Introduction
former are totally inert toward ethene if the metal-
attached alkyl lacks â-hydrogens. Following success-
ful preparation of longer chain bis(imino)pyridine co-
balt(I) n-alkyls, we reported our initial findings on the
reactivity of these complexes with alkenes.6 A study on
â-diketiminate iron alkyl complexes has afforded closely
related results.7 Further experimental investigations on
the organocobalt complexes as well as supporting theo-
retical studies are the subject of this paper.
Some of the most highly active ethene polymerization
catalysts to have been reported in recent years are based
on the bis(imino)pyridine (L) ligand framework, par-
ticularly when attached to metals such as Fe and Co,1
V,2 and Cr.3 The considerable difficulty of synthesising
iron alkyl derivatives, the surprising inertness of second-
row congeners (Ru, Rh),4 and the paramagnetism ac-
companying the V and Cr systems2,3 have made detailed
study of the reactivity of long-chain metal alkyls in
LMRn species problematic.
Experimental Results and Discussion
Synthesis, Properties, and Structure. The highly
air- and moisture-sensitive cobalt(I) n-alkyls I-VII were
prepared by reaction of divalent precursors with a slight
excess of the appropriate Grignard reagent (R ) Et,
n
n
nPr, Bu, Hx, CH2CH2Ph, or “PhEt”) in diethyl ether
to afford the desired products in good yield (Scheme 1).
These compounds are assumed to adopt a square-
planar coordination geometry on the basis of their
diamagnetism (16VE, d8 configuration), the crystal
structures of stable analogues LCoCl5a,b and LCoCH2-
SiMe3,5c and their apparent C2v symmetry in solution
(NMR). The NOE data provided in the Supporting
Information is entirely consistent with this description.
Attempts to synthesize cobalt(I) n-alkyl derivatives
containing aryl substituents carrying only one ortho-
substituent, or those bearing a branch in the R- or
â-positions of the alkyl chain, were unsuccessful.
Noteworthy features of the complexes are distinctive
nuclear magnetic resonances attributable to the ketimine
methyl groups8 (-1.33 to -1.35 ppm I-IV, -1.22 ppm
V, -1.47 ppm VI, -1.53 ppm VII), the para-H of the
pyridine ring (+10.25 ppm to +10.18 ppm I-VII), and
the â-hydrogens of the alkyl chain (-1.18 ppm I, -0.73
ppm II, -0.85 ppm III, -0.83 ppm IV, +0.32 ppm V,
-1.35 ppm VI, -1.16 ppm VII). On the basis of the
following observations it was concluded that the high-
As reported previously, it is, however, possible to
isolate neutral5,6 and ionic5a,b diamagnetic derivatives
of bis(imino)pyridine cobalt complexes. Only the latter
act as initiators for ethene polymerization, whereas the
(1) (a) Britovsek, G. J. P.; Gibson, V. C.; Kimberley, B. S.; Solan, G.
A.; White, A. J. P.; Williams, D. J. Chem. Commun. 1998, 849. (b)
Small, B. L.; Brookhart M.; Bennett, A. M. A. J. Am. Chem. Soc. 1998,
120, 4049. (c) 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.
(2) Reardon, D.; Conan, F.; Gambarotta, S.; Yap, G.; Yang, Q. Y. J.
Am. Chem. Soc. 1999, 121, 9318.
(3) Esteruelas, M. A.; Lo´pez, A. M.; Me´ndez, L.; Oliva´n, M.; On˜ate,
E. Organometallics 2003, 22, 395. Small, B. L.; Carney, M. J.; Holman,
D. H.; O’Rourke, C. E.; Halfen, J. A.; Macromolecules 2004, 37, 4375-
4386.
(4) Dias, E. L.; Brookhart, M.; White, P. S. Organometallics 1999,
19, 4995.
(5) (a) Humphries, M. J.; Clentsmith, G. K. B.; Gibson, V. C.;
Tellmann, K. P. Abstracts of Papers, 222nd ACS National Meeting,
Chicago, IL, August 26-30, 2001. (b) Gibson, V. C.; Humphries, M.
J.; Tellmann, K. P.; Wass, D. F.; White, A. J. P.; Williams, D. J. Chem.
Commun. 2001, 2252. (c) 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.
(7) Vela, J.; Smith, J. M.; Lachicotte, R. J.; Holland, P. L. Chem.
Commun. 2002, 2886.
(6) (a) Tellmann, K. P.; Humphries, M. J.; Gibson, V. C. Abstracts
of Papers, 224th ACS National Meeting, Boston, MA, August 18-22,
2002. (b) Gibson, V. C.; Tellmann, K. P.; Humphries, M. J.; Wass, D.
F. Chem. Commun. 2002, 2316.
(8) Similar high-field chemical shifts have also been observed in
diamagnetic R-diimine complexes of nickel. (a) Svodoba, M.; tom Dieck,
H. J. Organomet. Chem. 1980, 191, 321. (b) tom Dieck, H.; Svodoba,
M.; Greiser, T. Z. Naturforschung. 1981, 36b, 823.
10.1021/om049581h CCC: $27.50 © 2004 American Chemical Society
Publication on Web 10/08/2004

5504 Organometallics, Vol. 23, No. 23, 2004
Scheme 1. Synthesis of Bis(imino)pyridine Cobalt(I) n-Alkyl Complexes
Tellmann et al.
Scheme 2. Reaction of Cobalt(I) n-Alkyl
field shifts for the latter hydrogen nuclei are not caused
by â-agostic interactions:9 (a) the one-bond 1H-13C
coupling constants for I and II are in the range expected
for non-agostic alkyl groups (123-125 Hz), and there
is no pronounced change in the size of the coupling
constant on moving from a â-CH3 (I) to a â-CH2 (II)
group; (b) no appreciable change occurred in the chemi-
cal shift of the â-hydrogens in III on cooling a sample
in d8-toluene to 193 K nor on addition of a donor
molecule such as THF or diethyl ether, which usually
leads to displacement of agostic interactions. For these
reasons, and in view of the approximately orthogonal
orientation of the aryl ring systems to the chelate plane
usually encountered in these complexes, the high-field
resonances of the â-hydrogens are attributed to ring
current effects.10 The origin of the distinctive chemical
shifts for the ketimine methyl and pyridyl hydrogens
is less clear, although it has recently been suggested
that interpretation of the molecular ground states as
low-spin Co(II) antiferromagnetically coupled to a ligand
radical anion, in combination with low-lying exited
triplet states, may provide an explanation.11
Complexes with Ethene via â-H Transfer
Table 1. Selected Kinetic Data for â-H Transfer
from Complexes II-V and VII to Ethene
relative reaction rates
starting
material T (K)
reaction half-life
vs VII
(at 294 K)
vs V
(at 300 K)
(t1/2) (min)
II
294
294
300
300
294
300
294
40.0
179.2
92.5
66.1
68.3
15.4
11.0
3.64
16.3
III
III
IV
V
6.01
4.29
6.21
1.00
V
1.00
VII
Organocobalt complexes I-VII are not stable indefi-
nitely. However, their decomposition is sufficiently slow
to allow their isolation and manipulation for a few hours
at room temperature. Storage in the presence of either
an ethereal donor or an 1-alkene of carbon number iden-
tical to the alkyl chain suppresses the rate of decom-
position markedly, but not altogether. As mentioned
above, no change is observed in the NMR spectra of the
cobalt(I) alkyls I-VII on exposure to ethereal donors.
This is also the case for 1-alkene partners of identical
carbon number, with each compound appearing unper-
turbed by the presence of the other, even by the methods
of polarization transfer (NOE, EXSY). Despite this,
isotopic scrambling is observed between LCoEt (I) and
C2D4 suggesting that, while exchange is slow on the
NMR time scale, it is frequent enough to significantly
inhibit decomposition. Lower temperatures also have a
beneficial impact on the lifetime of complexes I-VII.
On decomposition, several diamagnetic products are
formed, the only unambiguously identified byproduct
being the alkane R-H (ethane, propane, and butane for
I-III, respectively). The other, unidentified decomposi-
tion products appear as complexes of lower symmetry
than the starting material. Presumably, irreversible
elimination of the hydrocarbon moiety R-H from the
organocobalt complexes occurs concomitant to, or as a
consequence of, ligand C-H activation.
Kinetics and Mechanism of â-H Transfer. Expo-
sure of II-V or VII to ethene leads to formation of I or
VI and CH2dCHR (R ) H, Me, Et, nBu, Ph), thus
offering the opportunity to gain insight into the steric
and electronic influences of bis(imino)pyridine ligands
on the reactivity of their associated metal alkyls.
The conversion of the longer chain cobalt-alkyl species
II-V and VII to the cobalt ethyl complexes I and VI
(Scheme 2) may be conveniently followed by NMR
spectroscopy, revealing a first-order dependence of the
reaction rate on [LCoCH2CH2R] and a zero-order de-
pendence on [ethene] in each case. The summary of the
kinetic data provided in Table 1 gives an indication of
the effects influencing the overall reaction rate, chiefly
the length of the n-alkyl chain and the steric constraints
of the ligand. For instance at 294 K, â-H transfer for
the n-propyl derivative II proceeds 4.48 times faster
than for the n-butyl derivative III, due, it is presumed,
to increased steric congestion in the transition state for
the n-butyl species. Similarly, the reaction of the 2-phe-
nylethyl derivative V with ethene proceeds substantially
(9) Brookhart, M.; Green, M. L. H.; Wong, L.-L. Progr. Inorg. Chem.
1988, 36, 1.
(10) The “quasi-aromatic nature” of transition metal containing
chelate rings may provide a rationalization for some of the observed
effects. The properties of five-membered, 2,4-unsaturated metal-
containing rings have been described by: Bayer, E. Angew. Chem.
1964, 76, 76.
(11) Knijnenburg, Q.; Hetterscheid, D.; Kooistra, T. M.; Budzelaar,
P. H. M. Eur. J. Inorg. Chem. 2004, 1204.

â-H Transfer between Co Complexes and 1-Alkenes
Organometallics, Vol. 23, No. 23, 2004 5505
Figure 1. Eyring plot for the reaction of LConBu III with ethene in C6D6.
Scheme 3. Plausible Mechanistic Pathways for
faster at 300 K than for either complex III or IV, but
slower than for II. Furthermore, the rate of â-H transfer
to ethene takes place faster in complexes with smaller
ortho-substituents in the bis(imino)pyridine ligand:
Reaction of the n-butyl complex VII (Ar ) 2,4,6-
Me3C6H2) with ethene proceeds 16.3 times faster than
for derivative III (Ar ) 2,6-iPr2C6H3), indicative of a
significantly lower barrier in the transition state (∆∆Gq
â-Hydrogen Transfera
) ∆Gq
- ∆Gq
) 1.63 kcal mol-1).
VIIfI
IIIfI
In addition to this, C-H bond-breaking was inferred
in the transition state of the rate-determining step in
view of a kinetic isotope effect (KIE) of 3.32 (0.04) for
the conversion of d9-III into d5-I by d4-ethene at 300 K.
The alkyl chain and the 1-alkene had to be fully
deuterated in order to avoid complications arising from
H/D scrambling. The magnitude of the observed KIE is
about half that of the usually anticipated primary KIE
for C-H bonds (but much larger than a secondary KIE),
in accordance with expected deviations arising from
nonlinear movement through the transition state,12 as
would be anticipated for this type of hydrogen transfer.
To obtain further insight into the nature of the
transition state, the reaction of III with ethene was also
followed over the temperature range 283-315 K. With
clean transformations occurring in each case, an Eyring
plot (Figure 1) was obtained and the activation param-
eters were determined as ∆Hq ) 24.7 ( 0.9 kcal mol-1
a
Path A is favored on the basis of experimental observa-
tions.
LConBu III) with dihydrogen in situ and, while we were
unable to locate the hydride resonance by NMR (ICo
)
7/2), a strong absorption in the IR spectrum at 2092
cm-1 was attributed to the Co-H stretch. Furthermore,
the description of VIII as LCoH is supported by the
observation of either methane, ethane, or butane as the
reaction by-product, depending on the alkyl precursor
employed. Isolation of VIII has not proven possible
because of the compound’s pronounced sensitivity and
instability toward elimination of free ligand. No further
stability is imparted by isotopic substitution to give
LCoD due to scrambling of the isotopic label into the
ortho-iPr groups of the ligand to afford (d-L)CoH, as
and ∆Sq ) 6 ( 3 cal mol-1 K-1 (∆Gq
) 22.9 ( 0.9
298
demonstrated by the H and 13C NMR data of complex
dn-VIII, as well as the mass spectra of the free ligand
recovered after hydrolysis of the complex.
1
kcal mol-1). The two most plausible reaction profiles as
limiting cases for â-H transfer are shown in Scheme 3:
Path A proceeds via a stepwise, dissociative mechanism
with four-centered transition states, whereas path B
follows a concerted, associative mechanism having a six-
centered transition state. The relatively small entropy
of activation along with the independence of the reaction
rate on ethene pressure is more consistent with path
A, since the involvement of ethene in the transition state
for path B would be anticipated to afford a first-order
dependence on [C2H4] and a strongly negative entropy
of activation.
The reaction of VIII with 1-alkenes (e.g., ethene,
propene, 1-butene) yields the anticipated cobalt(I) n-
alkyls I-III, lending further weight to the description
of VIII as LCoH (Scheme 4). The alkene insertion into
the cobalt-hydride bond is very rapid at room temper-
ature (,5 min) and therefore does not allow for a
determination of the relative rates of alkene insertion.
Crucially, though, this insertion is much faster than the
typical rates of â-H transfer, providing additional sup-
port for the mechanism to proceed in stepwise fashion
The postulated reaction intermediate of path A is
LCoH VIII. This compound is generally accessible by
the reaction of a cobalt(I) alkyl (LCoMe,5 LCoEt I,
(12) More O’Ferrall, R. A. J. Chem. Soc. (B) 1970, 785.

5506 Organometallics, Vol. 23, No. 23, 2004
Tellmann et al.
Scheme 4. Preparation of the Cobalt-Hydride
Complex VIII and Its Reactivity toward 1-Alkenes
Scheme 5. Reversible Displacement of 1-Alkenes
from LCoR Complexes
mixture are isomerized to internal alkenes at rates not
markedly slower than those of the exchange reactions.
Since internal alkenes do not appear to participate in
â-H transfer (consistent with failure to isolate R- or
â-branched cobalt(I) alkyls), the equilibrium is observed
to drift over time. For this reason some uncertainty
remains over the extent of adaption of the ratio of cobalt
n-alkyls to the varying concentration of 1-alkenes at any
stage and, hence, the accuracy of the calculated values
of Keq. In general, though, it can be stated that the
longer-chain cobalt n-alkyl complexes II-IV are all very
similar in energy from a thermodynamic perspective
(∆G ca. 1 kcal mol-1), but higher in energy than the
cobalt ethyl I by around 3.5-4.0 kcal mol-1 at room
temperature.
(path A), rate-limited by elimination of alkene to the
high-energy intermediate VIII.
When using higher 1-alkenes instead of ethene in the
alkene exchange reaction, an equilibrium not far from
unity is attained. From a kinetic perspective, for in-
stance, the initial rate of reaction between LConBu III
and propene (to give II and 1-butene) is virtually iden-
tical to that of â-H transfer from III to ethene up to
95% of the first half-life. If the concerted path B were
followed, it would be expected that incorporation of pro-
pene into the transition state would lead to a significant
increase in steric congestion and thereby reduce the rate
of the reaction, whereas the transition state for path A
does not involve the incoming monomer. Also, the
equilibrium between II and III is reached by reversible
first-order kinetics (first order in either LCoR species).
Given the similarity of propene and 1-butene, in com-
bination with the extra steric requirements of the
additional substituent on the alkene in an associative
transition state, a change from first-order to second-
order kinetics might be anticipated if path B were the
predominating mechanism. The equilibrium however is
attained by reversible first-order kinetics, which is fully
consistent with observations made for â-H transfer to
ethene and the stepwise reaction path A in Scheme 3.
Thermodynamic Aspects. As alluded to above,
alkyl exchange via â-H transfer between the longer
chain cobalt n-alkyl complexes and higher 1-alkenes
(gC3) occurs readily, as well as reversibly, at stoichio-
metric concentrations, with Keq near unity. This was
exemplified for equilibria between complexes II/III,
II/IV, and III/IV in combination with the appropriate
1-alkene pairs and is hence assumed to be general for
LCoR complexes.
Consideration of the isodesmic equation
ethene + n-butane ) ethane + 1-butene
illustrates that the thermodynamic stability of a sub-
stituted alkene (such as 1-butene) is greater than that
of an unsubstituted alkene (such as ethene) by 3.2 kcal
mol-1 at 298 K.13 Evidently, the exothermicity of alkyl
exchange reactions that afford LCoEt I plus 1-alkene
(∆G about 3.5-4.0 kcal mol-1) is predominantly ac-
counted for by the greater stability of the alkene moiety,
with a further, modest contribution due to the increased
stability of the organometallic complex.
Experimental Conclusions
In summary, the experimental observations strongly
indicate that â-H chain transfer proceeds via a stepwise
mechanism involving a cobalt-hydride intermediate
rather than a concerted process involving direct â-H
transfer to monomer. However, further investigation of
this reaction was warranted, particularly in view of the
apparent differences between the reactivity of the
cobalt(I) n-alkyls I-VII and the highly active bis(imino)-
pyridine cobalt polymerization system. In the catalyti-
cally active system chain termination by â-H transfer
is first order in ethene concentration (based on the
observation that PE molecular weight is independent
of ethylene pressure), implying involvement of an ethene
molecule in the process. Also, the reactivity described
in this article represents a chain shortening process, in
stark contrast to the highly efficient chain growth
process of the polymerization system.
Indeed, a reversal of the alkyl exchange reactions that
afford thermodynamically more stable LCoEt I is pos-
sible when using large excesses of a 1-alkene (Scheme
5): Addition of excess 1-hexene to a solution of LCoEt
I in d6-benzene leads to the formation of detectable
quantities of the targeted cobalt n-alkyl IV together
with an equimolar quantity of ethene.
Unfortunately, however, an accurate determination
of the equilibrium constants for â-H transfer between
the higher cobalt n-alkyls is severely complicated by a
side reaction, in which the 1-alkenes present in the
We thus decided to explore the computed potential
energy surface for the compounds central to the mech-
(13) Robert C. Weast, Ed.-in-chief. Handbook of Chemistry and
Physics, 58th ed.; CRC Press: Cleveland, OH, 1977.

â-H Transfer between Co Complexes and 1-Alkenes
Organometallics, Vol. 23, No. 23, 2004 5507
Scheme 6. Schematic for the Free Energy Surface for Stationary Points 1-16 (Ar ) R ) H)a
a Stationary points 1-6 relate to ethene exchange, 7, 8, and 12-14 relate to alkene polymerization, and 9-11 relate to alkene
exchange. The existence of the point marked with the asterisk (*) is inferred on the free energy rather than potential energy
surface (see main text).
anism, as well as the energies of the transition states
connecting them. To differentiate unambiguously be-
tween computational results and experimentally char-
acterized species, separate numbering schemes have
been adopted for each section, using I (or VI) ) 4, III
(or VII) ) 8, IV ) 14, and VIII ) 1 (cf. Schemes 1, 6).
The nomenclature employed in the computational sec-
tion does not distinguish explicitly between the more
subtle ligand substitution patterns (Ar ) 2,6-iPr2C6H3
versus Ar ) 2,4,6-Me3C6H2). As a general guideline,
even arabic numerals correspond to minima in the
potential surface, whereas odd arabic numerals denote
transition states.
the 6-31G(d,p)/Co:6-311G(d) combination as suitable for
further studies.
Overview of Potential Surface. The reactions
shown in Scheme 6 involve both associative and dis-
sociative pathways, with accompanying changes in
entropy. To compare the energies on an equal scale, we
calculated the zero-point, thermal, and entropy cor-
rected free energies based on (unscaled) vibrational fre-
quencies, evaluated for optimized geometries using the
Gaussian programs14 and the 6-31G(d,p)/Co:6-311G(d)
basis. For Ar ) R ) H, the following general surface
characteristics are revealed (Scheme 6). The initial
reaction in this system involves coordination of ethene
to 1 to form a π complex 2 (-14.5 kcal mol-1), followed
by insertion across the Co-H bond to create the Co-Et
alkyl 4, via a transition state 3, exhibiting only a small
barrier (1.6 kcal mol-1 from 2). The overall (total energy)
Computational Discussion
Basis Set Quality. It was thought imperative to
establish the appropriate basis set level to use in
exploring the overall potential surface using the B3LYP
hybrid density functional procedure. The key intermedi-
ates in the overall scheme (Scheme 6) are the cobalt
alkyls 4, 8, and 14 (and the parent hydride 1) and the
effects of ligand substitution (at Ar and R) upon these
species. At an initial 6-31(d) basis set level on all atoms,
including the cobalt, we noted that the hydrogen of the
Co-H bond of 1 was computed to lie significantly above
the plane of the ligand chelate. We felt this result might
be basis set dependent, and indeed increasing the basis
to 6-31G(d,p) on C, H, and N and to 6-311G(d) for Co
[subsequently simply referred to as 6-31G(d,p)] pla-
narizes this metal center. At this enlarged basis, and
for Ar ) R ) H, the geometries of the cobalt n-alkyls 4
and 8 still reveal significant nonplanarity at the Co
center, due to agostic interactions with a â-hydrogen of
the alkyl chain. The Co‚‚‚H bond lengths in 4 are 1.719/
1.725 Å [6-31G(d)/6-31(d,p)]. An alternative R-agostic
conformation was also located for 4, which is higher
than the â form by 0.6 kcal mol-1 in free energy. Further
expansion of the basis set to 6-311(d,p) or 6-311(3d,p)
on all atoms has an insignificant effect on the metal
geometries in these various species, and we considered
exothermicity from 1 + ethene to 4 is -19.0 kcal mol-1
.
Repeating this process from 4 to 8 via 6 but with Co-H
insertion replaced by Co-C insertion reveals an exo-
thermicity of -11.2 kcal mol-1, but with the barrier 7
(21.8 kcal mol-1 from 6) being much higher than for
Co-H insertion. A third insertion step from 8 to 14 via
12 shows little change from the second step, being
exothermic by -10.8 kcal mol-1
.
The two alternative alkene exchange mechanisms
bifurcate from the C-C bond formation (polymerization)
route at structure 8. In the first of these, 8 could instead
undergo â-H abstraction by Co via transition state 9 to
give 10, followed by dissociation of butene to give 11
()1 + butene). These points are +5.6, +3.7, and +14.5
kcal mol-1 relative to 8 in free energy. The second ex-
change mechanism bifurcates from either 6 or its homo-
logue 12 via the associative transition states 5 or 15,
representing concerted â-hydrogen transfer to alkene.
This route has barriers of 44.8/43.2 kcal mol-1 from 6/12.
An equivalent, but stepwise, route for alkene exchange
can also be accomplished by following the pathway
6-4-3-2-1 (overall barrier 23.4 kcal mol-1) or the
homologous 12-8-9-10-11 (barrier 18.9 kcal mol-1).

5508 Organometallics, Vol. 23, No. 23, 2004
Tellmann et al.
Effect of Ligand Bulk on the Cobalt Alkyl Com-
plexes. We next addressed the real, i.e., fully substi-
tuted, system by extending the basis set layer model to
include the Ar ) 2,6-iPr2C6H3 and R ) Me groups,
described with the minimal STO-3G basis. To establish
at least approximately the extent to which this basis
adequately reproduces the steric interactions exerted by
the bulky aryl groups, we compared the B3LYP/STO-
3G energies for cis- and trans-1,2-dimethyl-1,2-di(tert-
butyl)ethene,15 for which the latter is known to have
the lower enthalpy of formation by -8.6 kcal mol-1. The
B3LYP/STO-3G energies give a difference of -6.6 kcal
mol-1. This suggests that steric effects modeled this way
will be only slightly underestimated.
Figure 2. Space-filling representations for 8, where Ar )
2,6-iPr2C6H3 and R ) Me, showing (a) the snug fit of the
n-butyl group (yellow) into the coordination site on the
cobalt center, accompanied by the tilting of the aryl groups
for the agostic conformation; the agostic hydrogen is shown
in orange; (b and c) two lower energy nonagostic conforma-
tions, for which the C-H‚‚‚π bonded hydrogens are de-
picted in orange.
The first task was to analyze the effect of the bulky
ligands on the agostic interactions previously calculated
for 4 and 8. With Ar ) 2,6-iPr2C6H3 and R ) H, the
Co‚‚‚H agostic length compresses from 1.744 to a shorter
1.707 Å. In contrast, for the larger n-butyl group in 8,
the Co‚‚‚H agostic distance lengthens to 1.768 Å (Ar )
2,6-iPr2C6H3, R ) H), due to induced twisting of the
agostic conformation by interaction between the termi-
nal ethyl group of the butyl side chain and the aryl
substituents. This steric compression induces a further
new nonagostic conformation (Figure 2) not found with
the unsubstituted systems. For 4 (Ar ) 2,6-iPr2C6H3, R
) H) this is still higher than the agostic form, but for 8
(Ar ) 2,6-iPr2C6H3, R ) H) it becomes lower in free
energy by 0.4 kcal mol-1
.
Figure 3. Space-filling representation for 6, where Ar )
2,6-iPr2C6H3 and R ) Me, showing coordinated ethene
(magenta) and Co-ethyl (orange), separated by the bulk of
the isopropyl groups. The distance between the nearest two,
nonbonded carbon atoms of the ethyl chain and the
coordinated ethene is 2.708 Å.
Buttressing the aryl group by addition of the methyl
substituent (R ) Me) to the ligand amplifies these
effects. In the case of 4 (Ar ) 2,6-iPr2C6H3, R ) Me) the
Co‚‚‚H agostic length is now compressed to 1.705 Å, and
chain twisting in 8 (Ar ) 2,6-iPr2C6H3, R ) Me) further
increases Co‚‚‚H to 1.809 Å (Figure 2). The additional
R ) Me group completes inhibition of the agostic
conformations, both LCoEt 4 and LConBu 8 now being
predicted to prefer nonagostic conformations by 1.7 and
3.3 kcal mol-1 in the free energies. This is in agreement
with experimental observations made for 8, which show
no significant thermal population of such conformations.
The nonagostic conformation of 8 (Ar ) 2,6-iPr2C6H3,
R ) Me) has some further interesting features. An all-
anti orientation of the n-butyl side chain allows parallel
alignment with the face of one of the aryl groups,
thereby maximizing the C-H‚‚‚π interactions (Figure
2c); an even lower energy conformation (by 1.4 kcal
mol-1) adopts a gauche orientation in the n-butyl chain,
again with the effect of enhancing the C-H‚‚‚π bonding
(Figure 2b). If the aryl substituents are changed to Ar
) 2,4,6-Me3C6H2, the greater steric bulk at the 4-posi-
tion equalizes the free energies of the two C-H‚‚‚π
bonded conformations.
(14) (a) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G.
E.; Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J.
A., Jr.; Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.;
Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.;
Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo,
C.; Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.;
Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.;
Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Stefanov, B. B.; Liu, G.;
Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.;
Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.;
Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; Johnson, B. G.; Chen,
W.; Wong, M. W.; Andres, J. L.; Head-Gordon, M.; Replogle, E. S.;
Pople, J. A. Gaussian 98 (revision A.11); Gaussian, Inc.: Pittsburgh,
PA, 1998. (b) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria,
E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. R., 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.; Challa-
combe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.;
Gonzalez, C.; Pople, J. A. Gaussian 03 (revision B.04); Gaussian, Inc.:
Pittsburgh, PA, 2003.
Effect of Ligand Bulk on Co-Alkene Complexes.
It is evident from the preceding section that the steric
bulk of the ligand has a pronounced effect on geometries,
conformations, and energies of the cobalt n-alkyls (4,
8, 14). The effect on the cobalt complexes containing
π-bonded alkenes (2, 6, 12) may be expected to be even
larger because of the greater steric requirements of
these in the metal coordination sphere (Figure 3). Thus
∆G for the equilibria 1, 4, 8 + ethene ) 2, 6, 12 are
-14.5, -4.5, and -4.4 kcal mol-1 for the unsubstituted
ligand system of low bulk (Ar ) R ) H), but -14.8, +2.6,
and +7.0 kcal mol-1 for the fully substituted ligand
framework (Ar ) 2,6-iPr2C6H3, R ) Me). This reveals
significant steric and entropic inhibition of π complex
formation, which increases with the size of the ac-
(15) Gano, J.; Lenoir, D.; Park, B.-S.; Roesner, R. A. J. Org. Chem.
1987, 52, 5636.

â-H Transfer between Co Complexes and 1-Alkenes
Organometallics, Vol. 23, No. 23, 2004 5509
Scheme 7. Scheme for Alkene Exchangea
a Free energies (kcal mol-1) for Ar ) R ) H (bracketed) and Ar ) 2,6-iPr2C6H3, R ) Me (italics) are shown relative to 8 +
ethene.
companying Co-alkyl side chain. Such inhibition of π
complex formation agrees with the lack of detection of
such π complexes experimentally.
Transition States. Having established the basic
features of the potential surface, and the effect of steric
compression upon it with the Ar and R groups, we turn
to discussing in more detail the free energy barriers for
both alkene exchange and alkene insertion/polymeri-
zation (Scheme 7).
Steric Effects on the Alkene Polymerization. We
discuss first the reaction that is observed not to occur,
namely, the C-C bond formation between the cobalt
n-alkyl group and the coordinated alkene, which would
ultimately result in polymerization (rather than alkene
exchange by â-H transfer). The key transition states are
7 (ethyl/ethene C-C formation) and 13 (n-butyl/ethene
C-C formation). The free energy barrier of 7 relative
to 6 (Ar ) R ) H) is 21.8 kcal mol-1, rising to 32.6 kcal
mol-1 for the fully substituted ligand (Ar ) 2,6-iPr2C6H3,
R ) Me). Inspection of the geometry of 7 reveals that
the least hindered orientation of the methyl group (Fig-
ure 4a) encounters steric congestion from the aryl
groups, rationalizing the 10.8 kcal mol-1 increase in the
barrier induced by increasing the ligand bulk. The eth-
ene dimerization reaction is therefore sterically inhib-
ited to a significant degree. Likewise, formation of the
cobalt n-hexyl product 14 via transition state 13 from
reactant 12 has free energy barriers of 21.8 (for Ar ) R
) H) and 37.4 kcal mol-1 (for Ar ) 2,6-iPr2C6H3, R )
Me), again indicating little probability of reaction by this
route.
Figure 4. Space-filling representations for (a) 7 with Ar
) R ) H, showing C-C forming bond (yellow) and methyl
(orange) substituent; the alternative endo-facing conforma-
tion for the methyl group (replacing the magenta hydrogen)
encounters hindrance from the ligand face; and (b) 7 with
Ar ) 2,6-iPr2C6H3, R ) Me.
of 44.8 kcal mol-1 (computed relative to 6 to facilitate
comparison with the C-C polymerization route). This
increases to 48.8 kcal mol-1 with Ar ) 2,6-iPr2C6H3, R
) Me (Figure 5), revealing a steric inhibition of this
associative route of +4.0 kcal mol-1. The analogous
alkene exchange mechanism between the cobalt n-butyl
complex and ethene involves the equilibrium between
12 and 16 connected via transition state 15. This point
has a free energy (relative now to 12) of 43.2 kcal mol-1
(Ar ) R ) H) and 44.5 kcal mol-1 (Ar ) 2,6-iPr2C6H3, R
) Me), this time revealing a rather smaller steric
inhibition. In general, however, these all represent very
high thermal barriers, making it highly unlikely that
alkene exchange reactions could proceed by this route
at typical (ambient) temperatures.
Steric Effects on Alkene Exchange Mechanism.
We now turn to the reaction that is observed. Transition
point 5, Ar ) R ) H, which corresponds to associative
ethene exchange, has a computed free energy barrier
Seeking alternative lower energy routes for alkene
exchange, we turned to the dissociative routes from 6
to 1, or the homologous pathway from 12 to 11. The
overall free energies for the route from 6 to 1 (+2

5510 Organometallics, Vol. 23, No. 23, 2004
Tellmann et al.
relative to species 8 + ethene (the left-hand side of the
equilibrium). We have already noted that formation of
the two alkene complexes 12 and 16 is actually inhibited
by the sterically demanding ligand (Ar ) 2,6-iPr2C6H3,
R ) Me). The overall free energy of reaction is rela-
tively unaffected by these steric effects, being -4.4 kcal
mol-1 for Ar ) R ) H and -2.6 kcal mol-1 for Ar )
2,6-iPr2C6H3, R ) Me, and correctly favoring the right-
hand side of the equilibrium.
Nature of the Transition State for Alkene Ex-
change. The remaining issues relate to the nature of
the transition state for the alkene exchange mechanism.
The conventional conclusion from the experimental
Figure 5. Space-filling representation for synchronous
alkene exchange transition states (a) 5 and (b) 15, both
with Ar ) 2,6-iPr2C6H3 and R ) Me, showing the ethene/
ethyl and butene/butyl groups (both sets in yellow), with
particular emphasis on the transferring hydrogen atom
(orange).
observation of a kinetic isotope effect (kH/kD
(d9) )
298
3.3) for the exchange of d9-8 (d9-III) with C2D4 is that
the transition state must clearly involve C-H cleavage,
and the temperature dependence of the reaction rate
gives ∆Gq ) 22.9 ( 0.9 kcal mol-1 and ∆Sq ) +6 ( 3
cal mol-1 K-1. These experimental values clearly ex-
clude the synchronous route via 15 on the grounds of
both the computed free energy barrier and its computed
entropy (∆Sq ) -50.2 cal mol-1 K-1). Of the two
alternative species that might constitute the rate-
limiting step, species 9 (Ar ) 2,6-iPr2C6H3, R ) Me) has
a small negative entropy (∆Sq ) -5.6 cal mol-1 K-1
)
and directly involves C-H cleavage. This latter char-
Figure 6. (a) Space-filling representation for the â-H
transfer/elimination transition state 9 (Ar ) 2,6-iPr2C6H3,
R ) Me) and (b) ball-and-stick model of transition state 9.
The displacement vectors for the calculated imaginary
normal mode (688i cm-1) clearly identify it as a â-H
elimination step.
acteristic can be probed by calculating the value of
kH/kD
(d9) using the Bigeleisen equations16 and the
298
rigid-rotor-harmonic-oscillator approximation. The value
obtained of 3.8 is in reasonable agreement with the
experimental value; its reduced value (from a normal
value of around 7 for linear hydrogen transfers) is in
accord with the nonlinearity of the Co-H-C transfer
in 9. The predicted free energy of activation (∆Gq ) 10.6
kcal mol-1) is however significantly lower than mea-
sured, which raises doubts as to whether 9 is truly the
structure of the rate-limiting species. Evidence that does
support 9 is the substituent effect induced by replacing
Ar ) 2,6-iPr2C6H3 by the sterically less demanding Ar
) 2,4,6-Me3C6H2. This experimentally is known to
accelerate the rate of exchange with ethene by a factor
of 16.3, which corresponds to a reduction in the free
energy activation barrier of 1.6 kcal mol-1. The calcu-
lated reduction of 2.5 kcal mol-1 tends to support this
steric argument.
ethenes) are 23.4 and 16.1 kcal mol-1 for Ar ) R ) H
and Ar ) 2,6-iPr2C6H3, R ) Me, respectively. Further-
more, a steric acceleration for the dissociative route of
-7.3 kcal mol-1 is revealed, opposite of the steric
inhibition anticipated for the associative mechanism.
The transition state for the â-H transfer 3, which
precedes 1, has barriers of 10.4 and 3.3 kcal mol-1 (Ar
) R ) H and Ar ) 2,6-iPr2C6H3, R ) Me, respectively)
relative to 6, revealing a slightly smaller steric accelera-
tion (of 6.3 kcal mol-1). The homologous dissociative
stepwise route with n-butyl replacing ethyl (12 to 11)
has corresponding free energy barriers of 18.9 kcal
mol-1 (Ar ) R ) H) and 9.0 kcal mol-1 (Ar ) 2,6-i-
Pr2C6H3, R ) Me), with a steric acceleration of -9.9 kcal
mol-1. The transition state 9 for â-H transfer on the
dissociative pathway has a free energy barrier from 12
of 10.0 and 3.6 kcal mol-1, respectively, again with a
System 11, which represents the dissociative endpoint
and (total) energy maximum for the elimination of
butene, is in better agreement with experiment than 9
in terms of overall free energy barrier (∆Gq ) 16.0 kcal
mol-1), but it deviates significantly in its predicted
entropy of activation (∆Sq ) +52.0 cal mol-1 K-1), which
is nevertheless a value typical of a fully dissociative
mechanism. Although no direct C-H cleavage is in-
volved in 11, the relatively large loss of zero-point
energy as represented by the Co-H/D bond means that
11 also leads to a large predicted (in this case equilib-
somewhat reduced steric acceleration (-6.4 kcal mol-1
relative to ethyl/ethene exchange.
)
Alkene Exchange Equilibrium. The preceding
arguments have established that the lowest energy
mechanism for alkene exchange proceeds via transition
state 9, leading to a dissociative endpoint 11 () 1 +
butene), along with sterically induced lowering of the
reaction barriers. It becomes convenient at this point
to re-normalize the potential surface appropriate for
considering the actually observed equilibrium:
rium) isotope effect of kH/kD
) 5.4 (d9). The isotope
298
effect can be decomposed to kH/kD
) 3.8 (d1) when
298
only the abstracted â-H is deuterated, and kH/kD
)
298
1.4 (d8) when the remainder of the n-butyl chain is
deuterated. From this we conclude that the replacement
of a C-H bond with a Co-H bond during this reaction
will result in a large observed isotope effect regardless
8 + ethene ) 4 + 1-butene
This is known experimentally to proceed from 8 to give
4, and so we present the computed free energies
summarizing these reactions in Scheme 7, normalized
(16) See: Maccoll, A. Ann. Rep. Prog. Chem. 1974, 71, 77-101.

â-H Transfer between Co Complexes and 1-Alkenes
Organometallics, Vol. 23, No. 23, 2004 5511
of whether C-H cleavage is directly involved in the
transition state. Observed isotope effects are therefore
not necessarily an unambiguous indicator of the nature
of the transition state.
barrier than predicted for the formation of 11 alone and
a much smaller entropy.
Overall Conclusions
The measurement of a small but positive entropy of
activation implies that the kinetic transition state is
looser in structure than 9, but is more bound than the
fully dissociated 11. This is likely to be achieved
somewhere in the potential region between species 10
and 11. An optimized reaction path scan using the
distance between the Co and the midpoint of the CdC
bond of the butene as a reaction coordinate reveals there
to be no obvious energy maximum other than the fully
dissociated system 11, and hence no defined transition
state on this region of the total energy surface. This
leads us to consider the possibility that this represents
an example of a so-called barrierless reaction, in which
the kinetics are determined by a transition state defined
by the free energy and not the potential energy surface.
Examples of such reactions are known, and a particu-
larly relevant one relates to the reaction between
substituted ethenes and a Cu catalyst.17 No transition
state in the potential energy surface was reported (at a
level of theory similar to that used here), but ap-
proximate estimates of the free energy17 suggested that
the free energy surface does have a maximum higher
than either reactant or product. Related observations
have previously been made for the addition of carbenes
to alkenes.18 Such behavior for the reaction of 8 to 11
would imply a free energy barrier (marked with an
asterisk (*) in Scheme 6) higher than that computed for
11. This would reconcile the computed barrier by
bringing it into closer agreement with experiment, and
very probably both the entropy and the isotope effects
would also match the experimental values.
Modeling Conclusions. We conclude that the com-
putational modeling of this reaction has firmly estab-
lished the following features: (a) modeling of the full
system, i.e., including the bulky aryl substituents, is
mechanistically essential. The steric effects arising from
both the ketimine methyl (R) and the aryl (Ar) groups
mediate the agostic interactions, inhibit associative
pathways, accelerate dissociative routes, and play a vital
role in one key transition state for â-H transfer; (b)
calculating free energies clearly establishes that as-
sociative pathways are disfavored on entropic grounds;
(c) pathways for C-C bond formation/alkene polymer-
ization are clearly shown to be higher than those for
alkene exchange, in accord with experiment; (d) hydro-
gen isotope effects for the alkene exchange mechanism
are not necessarily diagnostic of an explicit C-H cleav-
age in the rate-determining step, but could also arise
because of loss of zero-point energy in the formation of
a Co-H bond; (e) no transition state characterized in
the potential energy surface has the correct predicted
combination of free energy barrier and entropy, and we
conclude that the rate-limiting step is probably a
barrierless reaction corresponding to a free energy
maximum on the pathway for dissociation of butene
from the complex 10, which may have both a higher
A good agreement between experimental and compu-
tational studies was found in the description of â-hy-
drogen transfer in bis(imino)pyridine cobalt(I) n-alkyl
complexes, particularly when entropic contributions
were considered in the theoretical approach. Both stu-
dies independently conclude that alkene exchange oc-
curs via a dissociative mechanism centered on a cobalt-
hydride species. In view of this mutual overlap in the
elucidation of alkene exchange, the mechanism may be
considered to be reasonably well understood. Associative
mechanisms for either alkene exchange or alkene inser-
tion are evidently disfavored in the neutral cobalt(I)
complexes considered.
Although the qualitative agreement between the ex-
perimental and computational findings is good, quan-
titative differences remain, which would probably be
resolved in a more extensive theoretical scrutiny. The
calculated kinetic and thermodynamic values for
the relationship between LConBu (III/8 with Ar )
2,6-iPr2C6H3, R ) Me) and LCoEt (I/4) of ∆Gq ) 16.0
kcal mol-1 (via VIII/1) and ∆G298 ) -1.8 kcal mol-1 fall
somewhat short of the experimentally determined ones
(∆Gq ) 22.9 kcal mol-1 and ∆G298 about -3.5 to -4.0
kcal mol-1). Despite the fact that the theoretical model
provides a more refined understanding of the origins of
isotope effect, there is some disagreement over the
magnitude of the isotope effect (3.8-5.4 computed, 3.3
measured).
The fact that the stable species LCoH 1 ()VIII)
represents the highest point on the calculated PES for
dissociative mechanisms, with no kinetic stabilization
from surrounding transition states, is possibly the most
surprising result of the theoretical study. In view of
other studies17,18 it does not, however, appear unreason-
able to postulate that 1 is probably surrounded by free
energy barriers on either side for alkene association/
dissociation, which were not located on the standard
total energy surface. Further resolution of this issue
must await the development of computational method-
ologies for the location and quantification of these
kinetically stabilizing barriers and which are capable
of handling systems of this size and complexity.
Experimental Methods
Computational Details. Calculations were performed
using Gaussian 98 (revision A.11) and Gaussian 03 (revision
B.04). Geometries were optimized at two levels. Exploratory
studies were specified as B3LYP employing a 6-311G(d,p) basis
set for Co and 6-31G(d,p) (Ar ) R ) H). The substituted
systems (Ar ) 2,6-iPr2C6H3 or 2,4,6-Me3C6H2, R ) Me) were
modeled at the same level, but using a STO-3G basis for the
Ar and R ) Me groups. Corrections for free energies were
derived from a frequency calculation at these levels (including
symmetry number corrections for ethene and species 1).
Transition states were characterized by the single negative
root in the Hessian matrix exhibiting the appropriate normal
mode. Coordinates and energies for computed species are
available in the Supporting Information.
(17) Rasmussen, T.; Jensen, J. F.; Østergaard, N.; Tanner, D.;
Ziegler, T.; Norrby, P.-O. Chem. Eur. J. 2002, 8, 177.
(18) (a) Houk, K. N.; Rondan, N. G.; Mareda, J. J. Am. Chem. Soc.
1984, 106, 4291. (b) Houk, K. N.; Rondan, N. G. J. Am. Chem. Soc.
1984, 106, 4293.
Experimental Details. With the exception of the ligand
preparations, all reactions were carried out under an atmo-
sphere of dry dinitrogen gas. Standard Schlenk-line and

5512 Organometallics, Vol. 23, No. 23, 2004
Tellmann et al.
associated techniques were employed for the manipulation of
air-sensitive compounds. Conventional inert atmosphere dry-
boxes were used for the preparation of the analytical and
spectroscopic samples, as well as for weighing and storage of
air-sensitive compounds.
All starting materials were obtained from commercial
sources. Ligands and cobalt(II) and cobalt(I) complexes were
prepared as described previously.1 With the exception of
1-butanol, pentane, and toluene, all solvents were dried by
refluxing over an appropriate drying agent,19 distilled, and
degassed prior to use. Pentane, heptane, and toluene were
dried according to a published procedure.20 1-Butanol was
purchased in anhydrous grade and used as such. Pentane,
heptane, toluene, and diethyl ether were stored over a K
mirror, and THF was stored over molecular sieves.
All spectra were recorded at ambient temperature unless
otherwise stated. All NMR spectra were recorded on a Bruker
AC-250 (1H and 13C spectra at 250.133 and 62.896 MHz,
respectively). On the occasions that higher field instruments
were employed, spectra were recorded on either a Bruker DRX-
400 (1H and 13C spectra at 400.129 and 100.613 MHz,
respectively) or a Bruker AM-500 (1H resonance at 500.133
MHz). 1H and 13C spectra were referenced internally (the
former to the residual protium signal of the deuterated
solvents, the latter directly to the 13C multiplets due the
resonance signals of the NMR solvent) and are reported
relative to the chemical shift of tetramethylsilane.
(CoCH2CH2CH3), 16.2 (CoCH2CH2CH3), 1.5 (CoCH2CH2CH3).
EI-MS (m/z): 539 [Co{N3(dipp)2}], 82%; 466, 46%; 308, 78%;
90, 100%.
Preparation of [Co(CH2CH2CH2CH3){N 3(dipp)2}], III.
Using a procedure similar to that employed for I, the title
compound was isolated in yields of 120-150 mg (40-50%) from
0.500 mmol of [CoCl2{N3(dipp)2}] (306 mg) and 1.50 mmol of
n-butylmagnesium chloride (0.75 cm3 of a 2.0 M solution in
Et2O).
3
1H NMR (C6D6): δ 10.24 (1H, t, JH-H ) 7.6 Hz; p-H, py),
3
7.99 (2H, d, JH-H ) 7.6 Hz; m-H, py), 7.52 (2H, distorted
3
triplet, JH-H ) 7.7 Hz; p-H, Ar), 7.40 (4H, distorted doublet,
3
3JH-H ) 7.7 Hz; m-H, Ar), 3.11 (4H, septet, JH-H ) 6.8 Hz;
(CH3)CH(CH3)), 1.28 (2H, m; CoCH2CH2CH2CH3), 1.17 (12H,
d, 3JH-H ) 6.8 Hz; (CH3)CH(CH3)), 0.87 (2H, m; CoCH2CH2CH2-
CH3), 0.79 (12H, d, 3JH-H ) 6.8 Hz; (CH3)CH(CH3)), 0.61 (3H,
3
t, JH-H ) 7.3 Hz; CoCH2CH2CH2CH3), -0.85 (2H, m;
CoCH2CH2CH2CH3), -1.31 (6H, s; ArNdCCH3). 13C{1H} NMR
(C6D6): δ 165.2 (ArNdCCH3), 157.7 (o-C, py), 155.0 (i-C, Ar),
140.9 (o-C, Ar), 126.4 (p-C, Ar), 124.0 (m-C, Ar), 122.6 (m-C,
py), 117.6 (p-C, py), 30.5 (CoCH2CH2CH2CH3), 28.5 ((CH3)CH-
(CH3)), 26.4 (ArNdCCH3), 24.5 ((CH3)CH(CH3)), 23.4 ((CH3)-
CH(CH3)), 14.2 (CoCH2CH2CH2CH3), 13.7 (CoCH2CH2CH2CH3),
-2.2 (CoCH2CH2CH2CH3).
Preparation
of
[Co(CH2CH2CH2CH2CH2CH3)-
{N3(dipp)2}], IV. Using a procedure similar to that employed
for I, the title compound was isolated in yields of 150-190
mg (48-61%) from 0.500 mmol of [CoCl2{N3(dipp)2}] (306 mg)
and 1.50 mmol of n-hexylmagnesium bromide (0.75 cm3 of a
2.0 M solution in Et2O).
Preparation
of
Complexes.
Preparation
of
[Co(CH2CH3){N3(dipp)2}], I. The precursor complex [CoCl2-
{N3(dipp)2}] (306 mg, 0.500 mmol) was treated with ethyl-
magnesium chloride (0.50 cm3 of a 3.0 M solution in Et2O, 1.50
mmol) in diethyl ether (20 cm3) at -78 °C and allowed to warm
to 0 °C over 24 h, followed by removal of all volatiles under
reduced pressure. Extraction with pentane (30 cm3), followed
by filtration and solvent removal from the supernatant,
afforded an intensely colored solid that was dried in vacuo at
30 °C for 30 min. Isolated yield: typically 120-150 mg,
42-53% based on [CoCl2{N3(dipp)2}].
3
1H NMR (C6D6): δ 10.23 (1H, t, JH-H ) 7.6 Hz; p-H, py),
3
8.00 (2H, d, JH-H ) 7.6 Hz; m-H, py), 7.52 (2H, distorted
3
triplet, JH-H ) 7.7 Hz; p-H, Ar), 7.40 (4H, distorted doublet
3
3JH-H ) 7.7 Hz; m-H, Ar), 3.11 (4H, septet, JH-H ) 6.8 Hz;
(CH3)CH(CH3)), 1.28 (2H, m; CoCH2CH2CH2CH2CH2CH3),
3
1.17 (12H, d, JH-H ) 6.8 Hz; (CH3)CH(CH3)), 1.04 (2H,
m; CoCH2CH2CH2CH2C H2CH3), 0.96 (2H, m; CoCH2CH2CH2-
CH2CH2CH3), 0.83 (∼2H, multiplet partially overlapping with
triplet from Co(CH2)5CH3; CoCH2CH2CH2CH2CH CH3), 0.81
2
3
1H NMR (C6D6): δ 10.25 (1H, t, JH-H ) 7.6 Hz; p-H, py),
(∼3H, triplet partially overlapping with doublet from (CH3)-
CH(CH3), 3JH-H ) 7.0 Hz; Co(CH2)5CH3), 0.79 (∼12H, d, 3JH-H
) 6.8 Hz; (CH3)CH(CH3)), -0.83 (2H, m; CoCH2CH2CH2CH2-
CH2CH3), -1.31 (6H, s; ArNdCCH3). 13C{1H} NMR (C6D6):
165.2 (ArNdCCH3), 157.7 (o-C, py), 155.0 (i-C, Ar), 140.9 (o-
C, Ar), 126.5 (p-C, Ar), 124.0 (m-C, Ar), 122.5 (m-C, py), 117.6
(p-C, py), 31.9 (CoCH2CH2CH2CH2CH2CH3), 31.5 (CoCH2CH2-
CH2CH2CH 2CH3), 28.5 ((CH3)CH(CH3)), 27.9 (CoCH2CH2CH2-
3
7.99 (2H, d, JH-H ) 7.6 Hz; m-H, py), 7.51 (2H, distorted
3
triplet, JH-H ) 7.7 Hz; p-H, Ar), 7.39 (4H, distorted doublet,
3
3JH-H ) 7.7 Hz; m-H, Ar), 3.14 (4H, septet, JH-H ) 6.8 Hz;
(CH3)CH(CH3)), 1.51 (2H, q, 3JH-H ) 7.9 Hz; CoCH2CH3), 1.17
3
3
(12H, d, JH-H ) 6.8 Hz; (CH3)CH(CH3)), 0.74 (12H, d, JH-H
3
) 6.8 Hz; (CH3)CH(CH3)), -1.18 (3H, t, JH-H ) 7.9 Hz;
1
CoCH2CH3; JH-C ) 123 Hz), -1.33 (6H, s; ArNdCCH3).
13C{1H} NMR δ (C6D6): 165.2 (ArNdCCH3), 157.8 (o-C, py),
154.9 (i-C, Ar), 140.8 (o-C, Ar), 126.6 (p-C, Ar), 124.1 (m-C,
Ar), 122.6 (m-C, py), 117.6 (p-C, py), 28.5 ((CH3)CH(CH3)), 26.4
(ArNdCCH3), 24.3 ((CH3)CH(CH3)), 23.3 ((CH3)CH(CH3)), 13.0
(CoCH2CH3), -0.5 (CoCH2CH3).
CH2CH CH3), 26.4 (ArNdCCH3), 24.5 ((CH3)CH(CH3)), 23.4
2
((CH3)CH(CH3)), 22.9 (CoCH2CH2CH2CH2 CH2CH3), 14.5
(Co(CH2)5CH3), -1.7 (CoCH2CH2CH2CH2CH CH3).
2
Preparation of [Co(CH2CH2Ph){N3(dipp)2}], V. The
synthesis was very similar to that of I, as described above.
[CoCl2{N3(dipp)2}] (489 mg, 0.750 mmol) was treated with
2-phenylethylmagnesium chloride (2.25 cm3 of a 1.0 M solution
in THF, 2.25 mmol) in diethyl ether (30 cm3) at -78 °C,
allowed to warm to 0 °C over 24 h, and subsequently stirred
for half an hour at ambient temperature, followed by removal
of all volatiles under reduced pressure. Extraction with a
mixture of pentane (30 cm3) and toluene (15 cm3), filtration,
and solvent removal from the supernatant afforded a solid,
which was dried in vacuo at 30 °C for 30 min. Isolated yield:
typically 180-240 mg, 37-50% based on [CoCl2{N3(dipp)2}].
Preparation of [Co(CH2CH2CH3){N3(dipp)2}], II. Using
a procedure similar to that employed for I, the title compound
was isolated in yields of 120-150 mg (41-51%) from 0.500
mmol of [CoCl2{N3(dipp)2}] (306 mg) and 1.50 mmol of n-
propylmagnesium chloride (0.75 cm3 of a 2.0 M solution in
Et2O.
3
1H NMR (C6D6): δ 10.25 (1H, t, JH-H ) 7.6 Hz; p-H, py),
3
7.98 (2H, d, JH-H ) 7.6 Hz; m-H, py), 7.52 (2H, distorted
3
triplet, JH-H ) 7.8 Hz; p-H, Ar), 7.40 (4H, distorted doublet,
3
3JH-H ) 7.8 Hz; m-H, Ar), 3.12 (4H, septet, JH-H ) 6.8 Hz;
1
(CH3)CH(CH3)), 1.31 (2H, m; CoCH2CH2CH3; JH-C ) 119
3
1H NMR (C6D6): δ 10.21 (1H, t, JH-H ) 7.6 Hz; p-H, py),
3
Hz), 1.17 (12H, d, JH-H ) 6.8 Hz; (CH3)CH(CH3)), 0.78
3
7.90 (2H, d, JH-H ) 7.6 Hz; m-H, py), 7.60 (2H, distorted
(12H, d, 3JH-H ) 6.8 Hz; (CH3)CH(CH3)), 0.50 (3H, t, 3JH-H
)
3
triplet, JH-H ) 7.8 Hz; p-H, Ar), 7.45 (4H, distorted doublet,
1
7.1 Hz; CoCH2CH2CH3; JH-C ) 123 Hz), -0.73 (2H, m;
3JH-H ) 7.8 Hz; m-H, Ar), 7.11 (2H, m; m-H, CoCH2CH2Ph),
1
CoCH2CH2CH3; JH-C ) 125 Hz), -1.31 (6H, s; ArNdCCH3).
13C{1H} NMR (C6D6): δ 165.3 (ArNdCCH3), 157.7 (o-C, py),
155.1 (i-C, Ar), 140.8 (o-C, Ar), 126.5 (p-C, Ar), 124.0 (m-C,
Ar), 122.6 (m-C, py), 117.6 (p-C, py), 28.5 ((CH3)CH(CH3)), 26.4
(ArNdCCH3), 24.4 ((CH3)CH(CH3)), 23.4 ((CH3)CH(CH3)), 21.8
(19) Armarego, W. L. F.; Perrin, D. D. Purification of Laboratory
Chemicals, 4th ed.; Butterworth/Heinemann: Oxford, 1998.
(20) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.;
Timmers, F. J. Organometallics 1996, 15, 1518.

â-H Transfer between Co Complexes and 1-Alkenes
Organometallics, Vol. 23, No. 23, 2004 5513
3
6.92 (1H, broadened triplet, JH-H ca. 7.3 Hz; p-H, CoCH2-
quantity of room-temperature hydrogen gas was then rapidly
admitted into the vacuum above the solution, frozen at 77 K.
The solution was allowed to thaw and mixed with the gas by
repeated shaking. Other cobalt(I) n-alkyl reagents may be used
as reaction precursors in place of [CoMe{N3(dipp)2}]. Yield:
>98% by NMR; only byproduct (<2%) is uncomplexed ligand.
The volatiles may be trapped into another NMR tube for
further characterization.
3
CH2Ph), 6.61 (2H, broadened doublet, JH-H ca. 7.1 Hz; o-H,
CoCH2CH2Ph), 3.11 (4H, septet, JH-H ) 6.8 Hz; (CH3)-
3
CH(CH3)), 1.19 (∼2H, multiplet partially overlapping with
doublet from (CH3)CH(CH3); CoCH2CH2Ph), 1.17 (∼12H, d,
3JH-H ) 6.8 Hz; (CH3)CH(CH3)), 0.72 (12H, d, 3JH-H ) 6.8 Hz;
(CH3)CH(CH3)), 0.32 (2H, m; CoCH2CH2Ph), -1.22 (6H, s;
ArNdCCH3). 13C{1H} NMR (C6D6): δ 165.6 (ArNdCCH3),
157.1 (o-C, py), 154.9 (i-C, Ar), 147.6 (i-C, CoCH2CH2Ph), 141.2
(o-C, Ar), 128.1 (o-C, CoCH2CH2Ph), 127.9 (m-C, CoCH2-
CH2Ph), 126.6 (p-C, Ar), 124.3 (m-C, Ar), 124.1 (p-C, CoCH2-
CH2Ph), 122.8 (m-C, py), 117.6 (p-C, py), 36.7 (CoCH2CH2Ph),
28.6 ((CH3)CH(CH3)), 26.2 (ArNdCCH3), 24.5 ((CH3)CH(CH3)),
23.5 ((CH3)CH(CH3)), 1.2 (CoCH2CH2Ph).
3
1H NMR (C6D6): δ 10.83 (1H, t, JH-H ) 7.6 Hz; p-H, py),
7.60 (2H, distorted triplet, 3JH-H ) 7.7 Hz; p-H, Ar), 7.59 (2H,
d, 3JH-H ) 7.6 Hz; m-H, py), 7.46 (4H, distorted doublet, 3JH-H
3
) 7.7 Hz; m-H, Ar), 3.42 (4H, septet, JH-H ) 6.8 Hz; (CH3)-
CH(CH3)), 1.31 (12H, d, 3JH-H ) 6.8 Hz; (CH3)CH(CH3)), 0.28
3
(12H, d, JH-H ) 6.8 Hz; (CH3)CH(CH3)), -1.66 (6H, s;
Preparation of [Co(CH2CH3){N3(mes)2}], VI. Addition of
ethylmagnesium chloride (0.50 cm3 of a 3.0 M solution in Et2O,
1.50 mmol) to a suspension of [CoCl2{N3(mes)2}] (264 mg, 0.500
mmol) in diethyl ether (40 cm3) at -78 °C gave a mixture,
which was allowed to warm to 0 °C over 24 h, followed by
removal of all volatiles under reduced pressure. Extraction
with toluene (50 cm3), followed by filtration and solvent
removal from the supernatant, afforded an intensely colored
solid that was dried in vacuo at 25-30 °C for 30 min. Isolated
yield: 125 mg, 51% based on [CoCl2{N3(mes)2}].
ArNdCCH3). No resonance for Co-H is detected within +100
to -150 [I(59Co) ) 7/2]. 13C{1H} NMR (C6D6): δ 168.6
(ArNdCCH3), 160.1 (i-C, Ar), 156.5 (o-C, py), 140.2 (o-C, Ar),
126.3 (p-C, Ar), 124.2 (m-C, Ar), 123.7 (m-C, py), 118.0 (p-C,
py), 28.9 ((CH3)CH(CH3)), 25.6 (ArNdCCH3), 23.6 ((CH3)CH-
(CH3)), 22.7 ((CH3)CH(CH3)).
Reaction of [CoH{N3(dipp)2}] VIII with 1-Alkenes
(C2H4, C3H6, 1-C4H8). A solution of [CoH{N3(dipp)2}] VIII
(0.010 mmol prepared in situ) in C6D6 (0.65 cm3) was degassed
by two freeze-pump-thaw cycles, frozen again, and evacu-
ated. Subsequently, the desired quantity of 1-alkene (V ) 5.0
cm3, p ) 220 mbar; n ) 0.045 mmol) was condensed onto the
solution. The solution was thawed and gently shaken just prior
to examination by NMR. The principal reaction product
(>80%), either I, II, or III, was characterized by 1H and
13C{1H} NMR (spectroscopic details given above).
Alternatively, this compound may be prepared in situ by
reaction of VII with a small excess of ethylene in C6D6. Yield
1
quantitative by H NMR.
3
1H NMR (C6D6): δ 10.18 (1H, t, JH-H ) 7.6 Hz; p-H, py),
3
8.07 (2H, d, JH-H ) 7.6 Hz; m-H, py), 7.06 (4H, s; m-H, Ar),
2.30 (6H, s; p-CH3, Ar), 2.04 (12H, s; o-CH3, Ar), 1.31 (2H, q,
3JH-H ) 8.0 Hz; CoCH2CH3), -1.33 (3H, t, JH-H ) 8.0 Hz;
Attempted Preparation of [CoD{N3(dipp)2}], d1-VIII.
Attempted synthesis was analogous to that of VIII. A degassed
solution of [CoMe{N3(dipp)2}] (8.9 mg, 0.016 mmol) in C6D6
(0.7 cm3) was frozen (77 K), and an excess (3.5 equiv) of room-
temperature deuterium gas (2H2) was rapidly expanded into
the vacuum above the frozen solution. The solution was
subsequently allowed to thaw and mixed with the gas by
repeated shaking. Yield: 80-90% by NMR (about 10%-20%
uncomplexed ligand).
3
CoCH2CH3), -1.48 (6H, s; ArNdCCH3). 13C{1H} NMR (C6D6):
δ 164.3 (ArNdCCH3), 158.3 (o-C, py), 155.1 (i-C, Ar), 134.4
(p-C, Ar), 129.7 (o-C, Ar), 129.5 (m-C, Ar), 121.6 (m-C, py),
117.9 (p-C, py), 24.9 (ArNdCCH3), 21.3 (p-CH3, Ar), 19.1
(o-CH3, Ar), 10.8 (CoCH2CH3), -11.5 (CoCH2CH3).
Preparation of [Co(CH2CH2CH2CH3){N3(mes)2}], VII.
Using a procedure similar to that employed for I, the title
compound was isolated in yields of 120-150 mg (fw ) 513.6)
(47-58%) from 0.500 mmol of [CoCl2{N3(mes)2}] (264 mg) and
1.50 mmol of n-butylmagnesium chloride (0.75 cm3 of a 2.0 M
solution in Et2O).
i
The resonance signals of the Pr group (δ(1H) ) 3.42, 1.31,
0.28 ppm) are broadened and of diminished area (methine )
3.8/4.0H, sum of methyls ) 16.0/24.0H), the total reduction
in integral size correlating with the quantity of employed deu-
terium gas. In the 13C NMR, a complex arrangement of over-
lapping but slightly shifted multiplets (∆δ around a few tenths
3
1H NMR (C6D6): δ 10.25 (1H, t, JH-H ) 7.6 Hz; p-H, py),
3
8.05 (2H, d, JH-H ) 7.6 Hz; m-H, py), 7.08 (4H, s; m-H, Ar),
2.33 (6H, s; p-CH3, Ar), 2.04 (12H, s; o-CH3, Ar), 1.18 (2H, m;
CoCH2CH2CH2CH3), 0.71 (2H, m; CoCH2CH2CH2CH3), 0.58
i
of a ppm) is seen in the region of the Pr group’s signals. The
3
(3H, t, JH-H ) 7.2 Hz; CoCH2CH2CH2CH3), -1.16 (2H, m;
observed coupling patterns are indicative of extensive, non-
selective deuteration (I(2H) ) 1) in all three chemical shift
environments. All other resonance signals are as given for
VIII, above.
Hydrolysis of the reaction product with methanol and dilute
hydrochloric acid, followed by extraction with CH2Cl2 and
removal of solvent from the organic layer, afforded uncom-
plexed ligand. Analysis of the free ligand by mass spectroscopy
confirmed multiple deuteration of the ligand framework.
Attempts to rapidly trap LCoD with an alkene always afforded
bis(imino)pyridine cobalt complexes bearing nonlabeled n-alkyl
chains, but containing partially deuterated ligand.
CoCH2CH2CH2CH3), -1.53 (6H, s; ArNdCCH3). 13C{1H} NMR
(C6D6): δ 164.6 (ArNdCCH3), 158.4 (o-C, py), 155.4 (i-C, Ar),
134.4 (p-C, Ar), 129.7 (o-C, Ar), 129.4 (m-C, Ar), 121.7 (m-C,
py), 117.9 (p-C, py), 28.5 (CoCH2CH2CH2CH3), 25.4 (CoCH2-
CH2CH2CH3), 25.1 (ArNdCCH3), 21.3 (p-CH3, Ar), 19.2 (o-CH3,
Ar), 13.6 (CoCH2CH2CH2CH3), -3.3 (CoCH2CH2CH2CH3).
NMR-Scale Reactions. Reaction of [Co(CH2CH2R)-
{N3(Ar)2}] (II-V, VII) with a 1-Alkene. In an NMR tube, a
solution of [Co(CH2CH2R){N3(Ar)2}] (II-V, VII) (typically
0.030 mmol) in 0.6-0.7 cm3 C6D6 was degassed by two freeze-
pump-thaw cycles and frozen again, and the tube was
evacuated. Subsequently, the desired quantity of 1-alkene gas
was condensed (77 K) onto the frozen solution. The solution
was thawed and shaken just prior to examination by NMR.
Data characterizing [Co(CH2CH3){N3(Ar)2}] (I, VI), and
[Co(CH2CH2R){N3(Ar)2}] (II-V, VII) are as given previously.
The liquid 1-alkene 1-hexene was simply added to the solution
if and when required.
Acknowledgment. The authors wish to thank BP
Chemicals Ltd. for funding.
Supporting Information Available: 1H NMR spectra
and NOE data for I-VII, coordinates (Molfile) and energies
for computed species 1-16, details of KIE calculations. This
material is available free of charge via the Internet at
http://pubs.acs.org.
Preparation of [CoH{N3(dipp)2}], VIII. A solution of
[CoMe{N3(dipp)2}] (5.6 mg, 0.010 mmol) in C6D6 (0.7 cm3) was
degassed by two freeze-pump-thaw cycles. A stoichiometric
OM049581H