Communications to the Editor
J. Am. Chem. Soc., Vol. 119, No. 47, 1997 11539
Scheme 3
and 13CH4, which is much slower. The reaction of 1 with 13CH4
results in the incorporation of label into the iridium methyl
position,2 but the half-life for this reaction at 45 °C in CD2Cl2
is approximately 6 h, substantially longer than that for which 1
is converted to Cp*(PMe3)Ir(Ph)(OTf) (t1/2 ) 6 h at 9.6 °C).17
This establishes that methane extrusion must occur irreVersibly
in the rate determining step of the benzene C-H activation
reaction. (3) If intramolecular C-H activation occurred either
simultaneously with or immediately following this irreversible
methane-elimination step, there could be no dependence of the
rate of the reaction on the concentration or nature of the external
organic reactant. Exactly the opposite is true: in the reaction
of 1 with benzene the rate depends on the concentration of the
organic reactant, and in general the C-H activation rate shows
wide qualitative variation with different reactants (e.g., benzal-
dehyde19 reacts instantaneously with 1 at -60 °C). (4) A similar
conclusion can be derived from the observed deuterium isotope
effect. If irreversible methane elimination/cyclometalation were
the first step in the reaction, the rate of disappearance of 1 should
be the same for C6H6 and C6D6. The contrary is true: there is
a large primary isotope effect on the rate of activation of C6H6
vs C6D6 (kH/kD ) 4.0).17
We conclude that cyclometalated complexes such as 6 and
12 undergo C-H activation more rapidly than their unstrained
analogues. However, the intermolecular C-H activation reac-
tions of 1 in solution cannot be proceeding by initial cyclo-
metalation. Since the highly electrophilic species [Cp*(PMe3)-
Ir(CH3)]+ is known to coordinate extremely poor donor ligands
(i.e., CH2Cl2 as shown in the solid-state structure of 2),1 we do
not find it surprising that in the absence of donor stabilization
and at low (or zero, such as in Chen’s gas-phase experiments)
concentration of substrate, the inherently slower cyclometalation
pathway dominates.20
Finally, even in cases where cyclometalation does occur,21
we feel it is misleading to characterize this transformation as a
new mechanism for the C-H activation process. Cyclometal-
ation can in principle proceed by σ-bond metathesis or oxidative
addition/reductive elimination, just as in the intermolecular
reaction. Focusing on the intra- rather than the intermolecular
reaction does not somehow absolve mechanistic chemists from
the need to achieve the resolution of this dichotomy. Theory
seems to be strongly favoring the oxidative addition pathway.
We plan to continue our work on this problem to confirm this
prediction experimentally, and also hope to eventually under-
stand why the cyclometalated complexes, if they are given an
opportunity to form, react so readily with external C-H bonds.
Grignard reagent (Me2PCH2)MgCl (10) was synthesized by
reaction of LiCH2PMe2 with MgCl2 in THF and used as a
stock solution. Addition of 1 equiv of this reagent to
Cp*(DMSO)IrCl2 at -78 °C results in the formation of
cyclometalated chloride 11 (Scheme 3) in 30% yield. The H
NMR spectrum of 11 in C6D6 displays the expected character-
istics. The phosphine-bound methyl groups are diastereotopic
and resonate at 1.45 and 0.98 ppm. The methylene protons are
also diastereotopic and resonate at 1.32 and 0.78 ppm, and the
Cp* methyl resonance is coupled to phosphorus (JP-H ) 1.6
Hz). The most compelling spectroscopic characteristic is the
31P resonance at -64.8 ppm. The extremely high field chemical
shift is characteristic of phosphametalacyclopropane com-
plexes.14,15 A molecular ion is observed in the mass spectrum
at m/z 438 with the predicted isotopic pattern and no higher
mass peaks are observed.
Treatment of 11 with AgOTf in C6H6 results in the formation
of Cp*(PMe3)Ir(Ph)(OTf)16 (13) in 90% yield (Scheme 3). This
presumably involves initial metathesis of the iridium-bound Cl-
with OTf- to form iridium triflate 12, which immediately reacts
with C6H6 (Scheme 3). The rate of disappearance of 11 under
these conditions is extremely rapid (t1/2 ) 5 min at 25 °C), and
there is no spectroscopic evidence that 12 builds up during the
reaction. This indicates that the rate of reaction of 12 with C6H6
must be considerably faster than the ligand metathesis reaction,
and remarkably faster than the analogous reaction of 1 in neat
C6H6 (t1/2 ) 24 h at 25 °C).17 In contrast to its behavior in
benzene, however, treatment of 11 with AgOTf in CH2Cl2sthe
solvent normally used for our C-H activation experimentsseven
in the presence of 20 equiv of C6H6 (conditions under which 1
cleanly reacts to give 13), results in decomposition to unidentifi-
able products. The fact that no detectable Cp*(PMe3)Ir(Ph)-
(OTf) (13) is observed by NMR provides suggestive evidence
that 12 is not a viable intermediate in the reaction of 1.
The following considerations confirm the implication of our
observations on the chemistry of 11. (1) In the reaction of C6D6
with 1 no deuterium is incorporated into the phosphine methyl
groups, which would be required if the cyclometalation pathway
were operative.18 (2) Additionally, careful monitoring of the
reactions of benzene or other organic compounds with 1 or 2
shows that no detectable concentrations of any intermediates
build up during the reaction. Therefore, if initial cyclometalation
were to occur, it would have to be either the rate-limiting step
in the overall reaction or exist in a pre-equilibrium with the
starting complex. In the case of benzene as the organic reactant,
the rapid pre-equilibrium can be ruled out on the basis of the
relative rate of this reaction and that of the reaction between 1
13
1
Acknowledgment. We are grateful for support of this work by the
Director, Office of Energy Research, Office of Basic Energy Sciences,
Chemical Sciences Division, U.S. Department of Energy, under Contract
No. DE-AC03-76SF00098.
Supporting Information Available: Text giving full experimental
details and characterization data for compounds 6-8 and 11 and
spectroscopic data for 6, 8, and 11 (7 pages). See any current masthead
page for ordering and Internet access instructions.
JA972340Z
(19) Arndtsen, B. A.; Bergman, R. G.; Mobley, T. A.; Peterson, T. H.
Acc. Chem. Res. 1995, 28, 154.
(20) Prof. P. Chen has informed us that he, too, has carried out studies
on the solution C-H activation reactions of PMe3 complexes such as the
ones described here, and finds results that are in agreement with ours
indicating that the intermolecular C-H activation reactions do not proceed
by cyclometalation under these conditions. In the manuscript detailing this
work (ref 6b), Chen also finds that cyclometalated complexes react more
rapidly than their analogous noncyclometalated analogues, and that donor
stabilization likely inhibits the cyclometalation reaction in solution. We are
grateful to Prof. Chen for providing a draft copy of his full paper describing
these results prior to publication.
(13) Karsch, H. H. Z. Naturforsch. B 1982, 37B, 284.
(14) Wenzel, T. T.; Bergman, R. G. J. Am. Chem. Soc. 1986, 108, 4856.
(15) Bergman, R. G.; Seidler, P. F.; Wenzel, T. T. J. Am. Chem. Soc.
1985, 107, 4358.
(16) Woerpel, K. A.; Bergman, R. G. J. Am. Chem. Soc. 1993, 115, 7888.
(17) Arndtsen, B. A.; Bergman, R. G. Unpublished results.
(18) Similarly, the reaction of 1 with CH3CDO results in exclusive
formation of CH3D. Alaimo, P. J.; Arndtsen, B. A.; Bergman, R. G. J. Am.
Chem. Soc. 1997, 119, 5269.
(21) For a recent example of γ-cyclometalation in Cp*(P(OiPr)3)Ir(Me)-
(OTf), see: Simpson, R. D. Organometallics 1997, 16, 1797.