C-H activations at iridium(I) square-planar complexes promoted by a fifth ligand

Article abstract of DOI:10.1021/ja0557233

In the presence of ligands such as acetonitrile, ethylene, or propylene, the Ir(I) complex [Ir(1,2,5,6-η-C₈H₁₂)(NCMe)(PMe ₃)]BF₄ (1) transforms into the Ir(III) derivatives [Ir(1-κ-4,5,6-η-C₈H₁₂)(NCMe)(L)(PMe ₃)]BF₄ (L = NCMe, 2; η²-C₂H ₄, 3; η²-C₃H₆, 4), respectively, through a sequence of C-H oxidative addition and insertion elementary steps. The rate of this transformation depends on the nature of L and, in the case of NCMe, the pseudo-first-order rate constants display a dependence upon ligand concentration suggesting the formation of five-coordinate reaction intermediates. A similar reaction between 1 and vinyl acetate affords the Ir(III) complex [Ir(1-κ-4,5,6-η-C₈H₁₂){κ- O-η²-OC(Me)OC₂H₃}(PMe₃)]BF ₄ (7) via the isolable five-coordinate Ir(I) compound [Ir(1,2,5,6-η-C₈H₁₂){κ-O-η²- OC(Me)OC₂H₃}(PMe₃)]BF₄ (6). DFT (B3LYP) calculations in model complexes show that reactions initiated by acetonitrile or ethylene five-coordinate adducts involve C-H oxidative addition transition states of lower energy than that found in the absence of these ligands. Key species in these ligand-assisted transformations are the distorted (nonsquare-planar) intermediates preceding the intramolecular C-H oxidative addition step, which are generated after release of one cyclooctadiene double bond from the five-coordinate species. The feasibility of this mechanism is also investigated for complexes [IrCl(L)(PiPr₃)₂] (L = η²-C₂H₄, 27; η²-C ₃H₆, 28). In the presence of NCMe, these complexes afford the C-H activation products [IrClH(CH=CHR)(NCMe)(PiPr₃)₂] (R = H, 29; Me, 30) via the common cyclometalated intermediate [IrClH{κ-P,C-P(iPr)₂CH(CH₃)CH₂}(NCMe) (PiPr₃)] (31). The most effective C-H oxidative addition mechanism seems to involve three-coordinate intermediates generated by photochemical release of the alkene ligand. However, in the absence of light, the reaction rates display dependences upon NCMe concentration again indicating the intermediacy of five-coordinate acetonitrile adducts.

Full text of DOI:10.1021/ja0557233

Published on Web 12/03/2005
C-H Activations at Iridium(I) Square-Planar Complexes
Promoted by a Fifth Ligand
Marta Mart´ın,* Olga Torres, Enrique On˜ate, Eduardo Sola,* and Luis A. Oro
Contribution from the Departamento de Compuestos de Coordinacio´n y Cata´lisis Homoge´nea,
Instituto de Ciencia de Materiales de Arago´n, UniVersidad de Zaragoza-CSIC and Instituto
UniVersitario de Cata´lisis Homoge´nea, UniVersidad de Zaragoza, 50009 Zaragoza, Spain
Received September 2, 2005; E-mail: sola@unizar.es.
Abstract: In the presence of ligands such as acetonitrile, ethylene, or propylene, the Ir(I) complex [Ir(1,2,5,6-
η-C8H12)(NCMe)(PMe3)]BF4 (1) transforms into the Ir(III) derivatives [Ir(1-κ-4,5,6-η-C8H12)(NCMe)(L)(PMe3)]-
BF4 (L ) NCMe, 2; η2-C2H4, 3; η2-C3H6, 4), respectively, through a sequence of C-H oxidative addition
and insertion elementary steps. The rate of this transformation depends on the nature of L and, in the case
of NCMe, the pseudo-first-order rate constants display a dependence upon ligand concentration suggesting
the formation of five-coordinate reaction intermediates. A similar reaction between 1 and vinyl acetate affords
the Ir(III) complex [Ir(1-κ-4,5,6-η-C8H12){κ-O-η2-OC(Me)OC2H3}(PMe3)]BF4 (7) via the isolable five-coordinate
Ir(I) compound [Ir(1,2,5,6-η-C8H12){κ-O-η2-OC(Me)OC2H3}(PMe3)]BF4 (6). DFT (B3LYP) calculations in model
complexes show that reactions initiated by acetonitrile or ethylene five-coordinate adducts involve C-H
oxidative addition transition states of lower energy than that found in the absence of these ligands. Key
species in these ligand-assisted transformations are the distorted (nonsquare-planar) intermediates
preceding the intramolecular C-H oxidative addition step, which are generated after release of one
cyclooctadiene double bond from the five-coordinate species. The feasibility of this mechanism is also
investigated for complexes [IrCl(L)(PiPr3)2] (L ) η2-C2H4, 27; η2-C3H6, 28). In the presence of NCMe, these
complexes afford the C-H activation products [IrClH(CHdCHR)(NCMe)(PiPr3)2] (R ) H, 29; Me, 30) via
the common cyclometalated intermediate [IrClH{κ-P,C-P(iPr)2CH(CH3)CH2}(NCMe)(PiPr3)] (31). The most
effective C-H oxidative addition mechanism seems to involve three-coordinate intermediates generated
by photochemical release of the alkene ligand. However, in the absence of light, the reaction rates display
dependences upon NCMe concentration again indicating the intermediacy of five-coordinate acetonitrile
adducts.
saturated or unsaturated d8 C-H activating systems seems to
Introduction
rely on the generation of very reactive three-coordinate inter-
The need for sustainable industrial chemical processes has
boosted the research on functionalization methods involving
C-H activations at metal complexes, since they may lead to
atom-economic catalytic transformations of readily available
hydrocarbon substrates. This intense research has already
afforded various well-established functionalization strategies,1
together with a wealth of well-characterized C-H activating
metal systems.2 A substantial part of such findings is based on
the solution chemistry of rhodium and iridium complexes,
although a noticeably minor role is attributed within this context
to the very common unsaturated d8 square-planar compounds.
In fact, the most effective Rh(I) and Ir(I) precursors for C-H
oxidative addition hitherto reported are five-coordinate com-
plexes bearing Cp-type ligands,3 while the success of other
mediates via ligand dissociations.4 Even though thermodynamics
are obviously crucial to determine whether the corresponding
d6 C-H activation products can be observed, the relative
inadequacy of d8 square-planar species in these activations might
have a kinetic origin.
The currently accepted mechanistic rationalization of C-H
oxidative additions to metal complexes indicates the availability
of empty metal orbitals allowing side-on C-H coordination to
be a requisite.5 As indicated by the theoretical work of Saillard
and Hoffmann,6 the generation of such suitable empty orbitals
from unsaturated Rh and Ir d8 complexes involves a significant
distortion of their square-planar ground-state geometry, which
(4) (a) Goldman, A. S.; Goldberg, K. I. In ActiVation and Functionalization
of C-H Bonds; Goldberg, K. I.; Goldman, A. S., Eds.; ACS Symposium
Series 885; American Chemical Society: Washington, DC, 2004. See
also: (b) Krogh-Jespersen, K.; Czerw, M.; Goldman, A. S. J. Mol. Catal.
A: Chem. 2002, 189, 95-110. (c) Bronberg, S. E.; Yang, H.; Asplund, M.
C.; Lian, T.; McNamara, B. K.; Kotz, K. T.; Yeston, J. S.; Wilkens, M.;
Frei, H.; Bergman, R. G.; Harris, C. B. Science 1997, 278, 260-263. (d)
Cundari, T. R. J. Am. Chem. Soc. 1994, 116, 340-347.
(1) Labinger, J. A.; Bercaw, J. E. Nature 2002, 417, 507-514.
(2) Shilov, A. E.; Shulpin, G. B. Chem. ReV. 1997, 97, 2879-2932.
(3) (a) Jones, W. D. In ActiVation and Functionalization of Alkanes; Hill, C.
L., Ed.; John Wiley and Sons: New York, 1989; p 111. (b) Jones, W. D.;
Feher, F. J. Acc. Chem. Res. 1989, 22, 91-100. (c) Arndtsen, B. A.;
Bergman, R. G.; Mobley, T. A.; Peterson, T. H. Acc. Chem. Res. 1995,
28, 154-162. See also: (d) Asbury, J. B.; Hang, K.; Yeston, J. S.; Cordaro,
J. G.; Bergman, R. G.; Lian, T. J. Am. Chem. Soc. 2000, 122, 12870-
12871. (e) Smith, K. M.; Poli, R.; Harvey, J. N. Chem. Eur. J. 2001, 7,
1679-1690.
(5) (a) Crabtree, R. H. Chem. ReV. 1985, 85, 245-269. (b) Hall, C.; Perutz,
R. Chem. ReV. 1996, 96, 3125-3146.
(6) Saillard, J.; Hoffmann, R. J. Am. Chem. Soc. 1984, 106, 2006-2026.
9
18074
J. AM. CHEM. SOC. 2005, 127, 18074-18084
10.1021/ja0557233 CCC: $30.25 © 2005 American Chemical Society

Ligand-Promoted C−H Activations at Ir(I) Complexes
A R T I C L E S
is likely associated with significant activation energies and short
lifetimes of the reactive intermediates. While these latter features
may still allow facile oxidative additions of reactive substrates
such as dihydrogen,7 they seemingly render this C-H activation
pathway unfavorable against other routes based on ligand
dissociation.4
Alternatively, the postulated electronic and geometric requi-
sites for C-H oxidative addition can be easily fulfilled by
unsaturated intermediates generated after ligand dissociation
from five-coordinate complexes, although such intermediates
are expected to readily evolve inactive square-planar states in
the absence of restraints.6,8 With regard to the experimental
evidence, the rigid, fac-coordinating and strong Cp-type ligands
effectively restrict such undesirable evolution, rendering the
lifetime of the distorted intermediates long enough to allow for
intermolecular activations of a variety of substrates, alkanes
included.3 Unfortunately, such advantageous features of the Cp
ligands become disadvantages for the subsequent functional-
ization of the activated substrate, since any further generation
of vacancies at the activation product is hindered by the strong
Cp ligand coordination. In this context, the following pages will
illustrate that such strong and rigid polydentate ligands are not
strictly necessary for the success of C-H oxidative additions
initiated by Ir(I) five-coordinate compounds, by showing that
distorted four-coordinate intermediates exclusively bearing
monodentate ligands can live long enough to accomplish
intramolecular C-H activations. Furthermore, the five-coordi-
nate precursors of such activations are not required to be
thermodynamically stable species, whenever significant amounts
of them could be generated in situ from square-planar com-
pounds in the presence of weak ligands or coordinating solvents.
This leads to intramolecular C-H activation reactions at square-
planar complexes promoted by weakly coordinating ligands
which afford labile Ir(III) compounds capable of undergoing
further reaction.
Figure 1. Pseudo-first-order rate constants obtained by 31P NMR for the
transformation of 1 into 2 at various acetonitrile concentrations. Solvent:
1,2-dichloroethane; [1]0 ) 6.3 × 10-2 M; T ) 343 K.
indirect experimental evidence for the proposed hydride inter-
mediate was obtained in the form of alkenyl complexes resulting
from its intermolecular reaction with added alkynes.
The need for acetonitrile in this reaction was rationalized in
thermodynamic terms, just as a requisite to turn the final product
into a stable octahedral species, thus shifting the equilibria in
eq 1 toward complex 2. However, in a further investigation of
this transformation, we have found the reaction rate to be directly
dependent upon acetonitrile concentration. This new observation
indicates that, in addition to the proposed thermodynamic
contribution, the acetonitrile plays a kinetic role during the
transformation. The plot of the pseudo-first-order reaction rates
(kobs), obtained by 31P{1H} NMR in 1,2-dichloroethane as
solvent, versus acetonitrile concentration resembles those com-
mon for saturation kinetics (Figure 1), suggesting an involve-
ment of acetonitrile in the reaction mechanism beyond possible
solvent effects.
Acetonitrile is not the only ligand accelerating this reaction.
In fact, upon treatment with an excess of simple alkenes such
as ethylene or propylene, complex 1 has been found to afford
analogues of 2 containing an η2-alkene ligand (eq 2). The rates
Results and Discussion
Studies in Complex [Ir(1,2,5,6-η-C8H12)(NCMe)(PMe3)]-
BF4. In a recent paper on the reactivity of cationic cycloocta-
diene iridium complexes,9 we described the transformation of
the Ir(I) complex [Ir(1,2,5,6-η-C8H12)(NCMe)(PMe3)]BF4 (1)
into the Ir(III) derivative [Ir(1-κ-4,5,6-η-C8H12)(NCMe)2(PMe3)]-
BF4 (2). The reaction was found to slowly take place in
chlorinated solvents in the presence of acetonitrile excess. Our
rationalization of this process assumed a reversible sequence
of C-H activation and insertion elementary steps (eq 1), since
of these latter transformations, in 1,2-dichloroethane solutions
of 1 saturated in the corresponding alkenes, are about 2 orders
of magnitude faster than those obtained with acetonitrile as
additive. This kinetic dependence upon the nature of the additive
further suggests the involvement of the added ligands in the
reaction mechanism.
The spectroscopic data collected for the complexes [Ir(1-κ-
4,5,6-η-C8H12)(NCMe)(η2-C2H4)(PMe3)]BF4 (3) and [Ir(1-κ-
4,5,6-η-C8H12)(NCMe)(η2-C3H6)(PMe3)]BF4 (4) agree with the
(7) See, for example: Johnson, C. E.; Eisenberg, R. J. Am. Chem. Soc. 1985,
107, 3148-3160.
(8) Nu¨ckel, S.; Burger, P. Angew. Chem., Int. Ed. 2003, 42, 1632-1636.
(9) Mart´ın, M.; Sola, E.; Torres, O.; Plou, P.; Oro, L. A. Organometallics 2003,
22, 5406-5417.
9
J. AM. CHEM. SOC. VOL. 127, NO. 51, 2005 18075

A R T I C L E S
Mart´ın et al.
structures proposed in eq 2. They also confirm that, despite the
use of additives susceptible to C-H activation, this process
1
remains restricted to the cyclooctadiene ligand. The H NMR
signals corresponding to the η2-ethylene of 3 are two multiplets
at δ 2.83 and 3.59 which indicate the fast rotation of this ligand
in the NMR time scale. The coordination position of the alkene
ligands, cis to the phosphine, the σ-carbon, and one side of the
1
allyl moiety, has been deduced from the H NOESY NMR
spectra.
The potentially chelating vinyl acetate has provided further
insight into the possible mechanistic role of the additional ligand
in this metal-mediated isomerization of the cyclooctadiene
ligand. The CD2Cl2 solutions obtained by treatment of 1 with 1
equiv of vinyl acetate have been observed by NMR to contain
the five-coordinate compound [Ir(1,2,5,6-η-C8H12){κ-O-η2-
OC(Me)OC2H3}(PMe3)]BF4 (6) together with the precursor
complex in fast equilibrium with two other species. The low-
temperature NMR spectra (see the Experimental Section and
the Supporting Information) indicate these two new species to
coordinate both an acetonitrile and an η2-bonded molecule of
the additive, thus allowing their tentative formulation as two
isomers of the five-coordinate complex [Ir(1,2,5,6-η-C8H12)-
(NCMe)(η2-C2H3OC(O)CH3}(PMe3)]BF4 (5 in eq 3, assuming
trigonal bipyramidal structures). These isomers have been
Figure 2. Molecular structure of the cations of complexes 6-BPh4 (above)
and 7 (below).
Table 1. Selected Bond Distances (angstrom) and Angles (deg)
for Complexes 6-BPh4 and 7
6-BPh4
7
Ir-P
2.3813(16)
2.155(7)
2.132(7)
2.140(7)
2.3215(14)
2.043(5)
2.196(5)
2.168(6)
2.289(6)
Ir-C(1)
Ir-C(4)
Ir-C(5)
Ir-C(6)
Ir-C(8)
2.156(7)
2.102(8)
2.063(6)
2.081(4)
1.457(7)
1.311(7)
1.267(7)
85.55(19)
93.74(18)
129.39(19)
Ir-C(9)
2.156(6)
2.128(5)
2.238(3)
1.443(6)
1.320(6)
1.236(6)
91.49(16)
91.01(17)
124.55(18)
156.95(16)
observed to be the major components of the solutions when
using 1 equiv of vinyl acetate, although they disappear under a
ca. 10-fold excess of added ligand allowing quantitative
formation of 6. Crystals suitable for an X-ray diffraction
experiment have been obtained for the analogue of 6 with the
BPh4- anion (6-BPh4). The cation of this complex is shown in
Figure 2, and its relevant distances and angles are collected in
Table 1.
Ir-C(10)
Ir-O(1)
C(10)-O(2)
C(11)-O(2)
C(11)-O(1)
P-Ir-C(1)
P-Ir-C(4)
P-Ir-C(5)
P-Ir-C(6)
P-Ir-C(8)
P-Ir-C(9)
P-Ir-C(10)
P-Ir-O(1)
O(1)-Ir-C(1)
122.7(2)
90.0(2)
125.48(18)
85.77(11)
169.8(2)
The preparation of 6 has been found to require either low
temperature or fast reaction manipulation, since the complex
has been observed to transform into its Ir(III) isomer [Ir(1-κ-
4,5,6-η-C8H12){κ-O-η2-OC(Me)OC2H3}(PMe3)]BF4 (7) (eq 3).
The X-ray structure of this analogue of complexes 2-4 is
depicted in Figure 2, and its main distances and angles are listed
in Table 1.
84.56(16)
121.37(16)
91.77(9)
175.41(18)
similar in magnitude to those of the aforementioned reactions
with ethylene or propene excess (kobs ) 4.5 × 10-3 s-1 at 323
K in 1,2-dichloroethane). Interestingly, the same rate constant
has been obtained either under larger vinyl acetate excess or
The pseudo-first-order rates for the transformation of 1 into
7 in the presence of 10 equiv of vinyl acetate have been found
9
18076 J. AM. CHEM. SOC. VOL. 127, NO. 51, 2005

Ligand-Promoted C−H Activations at Ir(I) Complexes
A R T I C L E S
by using 6 as precursor of 7 without ligand addition. These latter
observations strongly suggest that acceleration of the 1,2,5,6-
η-C8H12 to 1-κ-4,5,6-η-C8H12 isomerization in the presence of
added ligands is related to the formation of five-coordinate
species from the square-planar precursor 1. This is expected to
result into reaction rates independent upon the concentration
of those additives affording thermodynamically stable five-
coordinate adducts but, in general, should lead to rate depend-
ences reflecting adduct formation preequilibria. Under this
interpretation, the analysis of the rates determined at 343 K with
acetonitrile as the additive (Figure 1), using the equation for
saturation kinetics (kobs ) k2[NCMe]/{[NCMe] + (k-1 + k2)/
k1}), leads to an estimated equilibrium constant for five-
coordinate acetonitrile adduct formation (k1/k-1) of 0.2 ( 0.1
mol-1 L (assuming k-1 . k2) and a first-order rate constant for
the 1,2,5,6-η-C8H12 to 1-κ-4,5,6-η-C8H12 isomerization at this
adduct (k2) of 1.0 × 10-4 ( 2 × 10-5 s-1
.
The proposed acetonitrile adducts of 1 could not be detected
by NMR, although the dynamic behavior of the complex in
solutions containing added acetonitrile has provided indirect
evidence for the formation of five-coordinate intermediates. In
fact, 1 has been observed to undergo a rearrangement in solution
that eventually averages, inter alia, the two 1H NMR resonances
due to nonequivalent COD olefinic protons. Pseudo-first-order
rate constants for this dynamic process have been obtained by
line-shape analysis of the CD2Cl2 1H NMR spectra of 1 at
various temperatures and different acetonitrile concentrations
(see the Supporting Information). This study has concluded a
first-order dependence of the observed rate constants upon
acetonitrile concentration, in agreement with a process driven
by acetonitrile coordination. Moreover, the activation parameters
Figure 3. DFT (B3LYP)-computed reaction coordinate for the 1,2,5,6-η-
C8H12 to 1-κ-4,5,6-η-C8H12 isomerization of the cyclooctadiene ligand at
the model cation 1′. Values in parentheses are relative free energies (kcal
mol-1) in standard conditions.
obtained for the dynamic process, ∆Hq ) 4.3 ( 0.5 kcal mol-1
,
∆Sq ) -13 ( 1 cal K-1 mol-1, are consistent with an
associative transition state, also indicating the energetic barrier
for the proposed acetonitrile coordination to be very low.
nitrile, in agreement with the experimental observations. The
geometric optimization of the various possible four-coordinate
species generated after release of one COD alkene moiety from
compounds 12′ has been found to afford the square-planar
compounds 13′-15′ but also a high-energy isomer adopting a
nonplanar structure, 16′. This latter species results from the most
stable adduct 12b′, after decoordination of the CdC trans to
phosphine. Such a process is likely to be kinetically preferred
over the release of the equatorial CdC, which with regard to
the shorter Ir-C distances (2.11 vs 2.29 and 2.32 Å) and the
longer CdC bond length (1.46 vs 1.39 Å) seems to coordinate
stronger.
Theoretical Calculations. The consequences to the C-H
activation mechanism of the postulated five-coordination have
been examined in the model cation [Ir(1,2,5,6-η-C8H12)(NCMe)-
(PH3)]+ (1′) by use of DFT (B3LYP) calculations. Initially, for
comparison purposes, the best possible reaction sequence for
the transformation of 1′ into an unsaturated analogue of 2, the
model cation [Ir(1-κ-4,5,6-η-C8H12)(NCMe)(PH3)]+ (11′), has
been investigated. The calculations indicate this transformation
to be thermodynamically disfavored in the absence of added
ligands. The calculated potential energy surface (Figure 3)
features two elementary steps of allylic C-H oxidative addition
and insertion, respectively, mediated by an unsaturated hydride
species, 9′, coordinating a 1-κ-5,6-η-cyclooctadienyl ligand. The
highest point of the computed reaction coordinate corresponds
to the C-H oxidative addition transition state, TS8, the free
energy of which under standard conditions being as large as
The species 16′ nearly retains the ligand arrangement of its
tbp precursor 12b′ without the participation of agostic interac-
tions. Nevertheless, an agostic interaction can be readily
developed within such a distorted structure, to eventually afford
a product of allylic C-H oxidative addition, [IrH(1,2,3-η-
C8H11)(NCMe)2(PH3)]+ (19′) (Figure 5). The transition state for
this C-H oxidative addition, TS18, lies 11.7 kcal mol-1 above
the ground state of 16′, and only 6 kcal mol-1 above TS17,
that required to restore the five-coordinate precursor 12b′. This
is in contrast to the behavior of the square-planar isomers 13′-
15′, for which our attempts to find a low-energy C-H activation
pathway have been unsuccessful.
45.6 kcal mol-1
.
In the presence of acetonitrile, the model cation 1′ has been
found to form two five-coordinate adducts of formula [Ir(1,2,5,6-
η-C8H12)(NCMe)2(PH3)]+, 12a′ and 12b′ (Figure 4). These
adducts display trigonal bipyramidal structures, with one or two
acetonitrile ligands in equatorial positions, respectively. Forma-
tion of the adducts is exothermic by ca. 5 kcal mol-1 although,
due to the unfavorable entropy, both isomers of 12′ are
thermodynamically disfavored against separate 1′ and aceto-
The calculations on the hydride-η3-allyl C-H oxidative
addition product 19′ have found a more stable coordination
isomer, 20′, which displays a cyclooctadienyl ligand coordinated
9
J. AM. CHEM. SOC. VOL. 127, NO. 51, 2005 18077

A R T I C L E S
Mart´ın et al.
Figure 4. DFT (B3LYP)-calculated five- and four-coordinate complexes formed from the model cation 1′ and acetonitrile. Values in parentheses are relative
free energies (kcal mol-1) in standard conditions.
in the 1-κ-5,6-η- fashion. This latter isomer can undergo an
insertion elementary step to afford the model final product [Ir(1-
κ-4,5,6-η-C8H12)(NCMe)2(PH3)]+ (2′), thus completing the
experimentally observed transformation. The relative free ener-
gies, intermediates, and transition states involved in this
calculated transformation are depicted in Figure 5. Assuming
the isomerization of 19′ into 20′ to be kinetically facile, the
highest point of the reaction coordinate (37.4 kcal mol-1) is
significantly lower than that calculated for cation 1′, thus
providing a possible rationalization for the accelerating effect
exhibited by acetonitrile. An analogous reaction coordinate can
be started from the five-coordinate ethylene adduct [Ir(1,2,5,6-
η-C8H12)(NCMe)(η2-C2H4)(PH3)]+ (22b′). This model com-
pound has been found to behave similarly to 12b′, leading to a
C-H oxidative addition activation energy 4 kcal mol-1 below
that of the acetonitrile-promoted reaction. This calculated result
is also in qualitative agreement with the experimental observa-
tion of faster reactions in the presence of ethylene. Further
energetic details for all these calculated transformations and
detailed geometric parameters of their intermediates and transi-
tion states are included as Supporting Information.
The calculations in this section indicate that, in the presence
of potential ligands or coordinating solvents, square-planar
iridium(I) compounds can undergo sequences of ligand
coordination/dissociation affording distorted intermediates ca-
pable of activating intramolecular C-H bonds. As in the
calculated examples, this ligand-assisted distortion of the
ground-state square-planar geometry can result in C-H
oxidative additions which are faster than those involving
simpler nonassisted distortions, thus rationalizing the ex-
perimentally observed reaction acceleration upon addition of
ligands.
Studies in Complexes [IrCl(η2-C2H3R)(PiPr3)2]. Due to
their chelating and relatively rigid cyclooctadiene ligands, the
complexes studied in the previous sections could be regarded
as peculiar, and the deduced C-H activation mechanism might
be considered just a particular consequence of the restraints
introduced by the COD ligand. To check the feasibility of this
mechanism in a more general ligand environment, we have
investigated the behavior of the neutral square-planar Ir(I)
complex [IrCl(η2-C2H4)(PiPr3)2] (27). This compound was
previously reported to activate the ethylene ligand under UV
irradiation, to afford an unsaturated hydride-vinyl Ir(III)
derivative.10 The addition of ligands such as pyridine or CO
was found necessary to obtain thermodynamically stable octa-
hedral complexes from the activation product, although possible
kinetic effects of these added ligands were not considered.
In agreement with the previously reported results, we have
observed that addition of acetonitrile excess to solutions of 27
or its η2-propylene analogue [IrCl(η2-C3H6)(PiPr3)2] (28) affords
hydride compounds. Nevertheless, the composition of the
reaction outcome has been found to depend on the reaction
conditions. The expected ethylene C-H activation product,
[IrClH(CHdCH2)(NCMe)(PiPr3)2] (29), has been obtained by
treatment of 27 with 2 equiv of acetonitrile at room temperature.
Its solution 1H, 13C{1H}, and 31P{1H} NMR spectra at temper-
atures above 303 K are consistent with a single reaction product,
but the signals in the spectra broaden upon decreasing the
temperature, eventually splitting into two sets of signals of
approximately equal intensity. Even though the low-temperature
signals (183 K) do not resolve enough to permit the identifica-
tion of each component in the mixture, the observed behavior
(10) Schulz, M.; Werner, H. Organometallics 1992, 11, 2790-2795.
9
18078 J. AM. CHEM. SOC. VOL. 127, NO. 51, 2005

Ligand-Promoted C−H Activations at Ir(I) Complexes
A R T I C L E S
Figure 5. DFT (B3LYP)-computed reaction coordinate for the 1,2,5,6-η-C8H12 to 1-κ-4,5,6-η-C8H12 isomerization of the cyclooctadiene ligand at the
model five-coordinate acetonitrile adduct 12b′. Values in parentheses are relative free energies (kcal mol-1) in standard conditions.
is consistent with a fast exchange in the NMR time scale
between two isomers of 29 (eq 4), as previously observed for
the analogous reaction with pyridine.10 The crystals obtained
from solutions of 29 have allowed the X-ray characterization
of one of the isomers, 29b, the structure of which is shown in
Figure 6. The most noticeable feature in this structure is the
very long Ir-N bond distance (2.102(4) Å, Table 2), which
should correspond to a readily dissociable acetonitrile ligand.11,12
Since the other expected isomer, 29a, is proposed to coordinate
acetonitrile trans to the also trans-labilizing hydride ligand,13
the observed fast exchange between isomers of 29 is likely to
result from facile acetonitrile dissociations in both isomers.
Under similar reaction conditions, the propylene compound
28 has been found to afford the hydride-alkenyl complex
[IrClH(E-CHdCH2CH3)(NCMe)(PiPr3)2] (30) (eq 4). With
regard to the variable temperature NMR data, the solutions of
this compound also consist of a mixture of the two expected
labile isomers in fast exchange. In this case, the crystals obtained
from the mixture correspond to the isomer with acetonitrile trans
to hydride, 30a, the structure of which is shown in Figure 6. In
agreement with the reasons given to rationalize the fast exchange
between isomers, the Ir-N bond distance in this structure is
indeed very long (2.122(3) Å, Table 2). The selectivity of this
(11) Sola, E.; Navarro, J.; Lo´pez, J. A.; Lahoz, F. J.; Oro, L. A.; Werner, H.
Organometallics 1999, 18, 3534-3546.
(12) Navarro, J.; Sola, E.; Mart´ın, M.; Dobrinovitch, I. T.; Lahoz, F. J.; Oro, L.
A. Organometallics 2004, 23, 1908-1917.
(13) Coe, B. J.; Glenwright, S. J. Coord. Chem. ReV. 2000, 203, 5-80.
9
J. AM. CHEM. SOC. VOL. 127, NO. 51, 2005 18079

A R T I C L E S
Mart´ın et al.
Table 2. Selected Bond Distances (angstrom) and Angles (deg)
for Complexes 29b and 30a
29b
30a
Ir-Cl
2.5239(13)
2.3424(14)
2.3383(15)
2.102(4)
2.020(5)
1.310(7)
2.4988(8)
2.3405(8)
2.3357(8)
2.122(3)
2.047(3)
1.334(5)
1.504(4)
164.37(3)
91.51(12)
96.6(13)
178.92(10)
89.55(7)
127.7(3)
125.7(3)
Ir-P(1)
Ir-P(2)
Ir-N
Ir-C(1)
C(1)-C(2)
C(2)-C(3)
P(1)-Ir-P(2)
N-Ir-C(1)
Cl-Ir-H
Cl-Ir-C(1)
Cl-Ir-N
Ir-C(1)-C(2)
C(1)-C(2)-C(3)
166.88(5)
176.5(2)
173.5(15)
88.21(16)
88.28(12)
134.1(5)
(NCMe)(PiPr3)] (31). This mixture of isomers has been prepared
in analytically pure form either by using a large acetonitrile
excess (ca. 80 equiv) at 353 K or, more conveniently, by reacting
the cyclooctene dimer [Ir(µ-Cl)(coe)2]2 with 4 equiv of phos-
phine and acetonitrile. Contrarily to the thermodynamic mixtures
of fast exchanging isomers observed for 29 and 30, the mixture
31 seems to consist of two kinetic isomers in ca. 7:3 molar
ratio. The isomers do not convert each other in the NMR time
scale despite the lability of their acetonitrile ligands, which show
characteristic broad signals in the RT NMR spectra due to their
fast dissociation. This observation suggests the isomers differ
from each other in the syn or anti orientation of the activated
isopropyl group relative to the hydride, rather than in the
acetonitrile coordination position (eq 4). Unfortunately, further
structural assignments have been found impossible due to the
absence of measurable NOE effects among the 1H NMR signals
of hydrides and those due to relevant protons of the activated
isopropyl group. The structure of 31 depicted in eq 4 also
assumes the trans relative position between hydride and chloride
ligands previously proposed for the analogue of this complex
with ammonia.15
The mixture of isomers 31 has been observed to react with
ethylene or propylene affording compound 29 or 30, respectively
(eq 4). These reactions are complete after few seconds at room
temperature, seemly being much faster than those forming the
same compounds from the starting complexes 27 and 28. This
observation strongly suggests that isomers 31 constitute inter-
mediates in the formation of the alkene activation compounds,
being the final reaction products only under experimental
conditions diminishing alkene concentration in solution and
disfavoring alkene coordination (high temperature and large
acetonitrile excess, respectively). This suggested mechanism
would imply that alkene activation takes place at unsaturated
Ir(III)16 centers generated after acetonitrile dissociation from
the kinetic C-H activation products 31, rather than at any Ir(I)
species.
Figure 6. Molecular structures of complexes 29b (above) and 30a (below).
C-H oxidative addition is noticeable, since isomeric compounds
resulting from the activation of the methyl group or other alkene
protons have not been observed. This is in contrast with the
cases examined in the previous sections, in which allylic
activations are those kinetically favored, and against the
thermodynamic selectivity found in related propylene activa-
tions, which also favors hydride-η3-allyl compounds.11,14
Higher reaction temperatures and/or the use of larger aceto-
nitrile excess during the synthesis of 29 and 30 has been
observed to result into the partial formation of two new hydride
products. With regard to their NMR data and the previously
reported result of a similar reaction with ammonia as additive,15
these new complexes can be formulated as two isomers of the
cyclometalated complex [IrClH{κ-P,C-P(iPr)2CH(CH3)CH2}-
The rates of transformation of 27 and 28 into any of their
C-H activation products have been observed to strongly depend
on the presence or absence of light. In agreement with the
previously reported photochemical character of these activa-
tions,10 the reactions exposed to daylight have been found to
(16) For recent work on C-H activation by Ir(III) complexes with leading
references, see: (a) Klei, S. R.; Golden, J. T.; Burger, P.; Bergman, R. G.
J. Mol. Catal. A: Chem. 2002, 189, 79-94. (b) Wong-Foy, A. G.; Bhalla,
G.; Lui, X. Y.; Periana, R. A. J. Am. Chem. Soc. 2003, 125, 14292-14293.
(14) Esteruelas, M. A.; Lahoz, F. J.; On˜ate, E.; Oro, L. A.; Sola, E. J. Am.
Chem. Soc. 1996, 118, 89-99.
(15) Schulz, M.; Milstein, D. J. Chem. Soc., Chem. Commun. 1993, 318-319.
9
18080 J. AM. CHEM. SOC. VOL. 127, NO. 51, 2005

Ligand-Promoted C−H Activations at Ir(I) Complexes
A R T I C L E S
by such a sequence have been found to be faster than those
involving simpler, nonassisted, distortions of the square-planar
geometry. Even though the distorted intermediates generated
in this way seem only capable of activating intramolecular C-H
bonds, the weakly coordinating nature of the ligands promoting
this mechanism permit subsequent reactions on the activation
products, such as intermolecular C-H oxidative additions of
alkenes. This characteristic suggests the possible exploitation
of this fifth ligand effect in new C-H activation/functionaliza-
tion strategies using seemly nonreactive d8 square-planar
complexes.
Experimental Section
Equipment. Infrared spectra were recorded in KBr in a Bruker
Equinox 55 (4000-400 cm-1) spectrometer. C, H, and N elemental
analyses were carried out in a Perkin-Elmer 2400 CHNS/O analyzer.
NMR spectra were recorded on Bruker Avance 400 or 300 MHz
Figure 7. Pseudo-first-order rate constants obtained by 31P NMR for the
transformation of 28 into mixtures of the C-H activation products 30 and
31, at various acetonitrile concentrations. Solvent: toluene-d8; [28]0 ) 4.2
× 10-2 M; T ) 303 K.
1
spectrometers. H (400 or 300 MHz) and 13C (100.6 or 75.5 MHz)
NMR chemical shifts were measured relative to partially deuterated
solvent peaks but are reported in ppm relative to tetramethylsilane. 31
P
occur at relatively fast rates. Nevertheless, the C-H oxidative
additions have also been observed to take place under strict
absence of light, although at slower rates. Under the latter
conditions, the rates estimated by 31P NMR depend on the nature
of the alkene ligand in the starting compound (even when the
final product is the common mixture of isomers 31), being faster
for the propylene derivative 28. Furthermore, the pseudo-first-
order rate constants measured in toluene-d8 solutions of 28
display a direct dependence upon acetonitrile concentration
(Figure 7), with a diagnostic saturation profile similar to that
obtained for the cationic cyclooctadiene complex 1. The analysis
of these kinetic data affords an estimated value of 0.7 ( 0.1
mol-1 L for the equilibrium constant of five-coordinate aceto-
nitrile adduct formation from 28 and a first-order rate constant
for C-H oxidative addition at this adduct of 3.7 × 10-4 ( 2 ×
(162.0 or 121.5 MHz) NMR chemical shifts were measured relative to
H3PO4 (85%). Coupling constants (J and N) are given in hertz. Spectral
assignments were achieved through 1H-COSY, 1H-NOESY, 13C-DEPT,
and 1H,13C-HSQC experiments. The sample temperature was calibrated
using the temperature dependence of the chemical shifts of MeOH.
MS data were recorded on a VG Autospec double-focusing mass
spectrometer operating in the positive mode; ions were produced with
the Cs+ gun at ca. 30 kV, and 3-nitrobenzyl alcohol (NBA) was used
as the matrix.
Synthesis. All experiments were carried out under argon atmosphere
by Schlenk techniques. Solvents were dried by known procedures and
distilled under argon before use. The complexes [Ir(µ-Cl)(coe)2]2,18
[Ir(1,2,5,6-η-C8H12)(NCMe)(PMe3)]BF4 (1), [Ir(1-κ-4,5,6-η-C8H12)-
(NCMe)2(PMe3)]BF4 (2),9 and [IrCl(η2-C2H4)(PiPr3)2] (27),10 were
prepared as previously reported. The analogue of 1 with tetraphenyl-
borate as anion, [Ir(1,2,5,6-η-C8H12)(NCMe)(PMe3)]BPh4, was prepared
10-5 s-1
.
19
as described for 1 starting from [Ir(µ-OMe)(cod)]2 (150.0 mg, 0.22
The above qualitative and quantitative kinetic observations
suggest the feasibility of two possible C-H activation mech-
anisms in the square-planar complexes 27 and 28. The most
effective one seems to be that initiated by a photochemical
release of the alkene ligand, whereas in the absence of light,
alkene dissociation requires the previous coordination of ac-
etonitrile. With regard to reported results and the conclusions
of the previous sections, such activation steps would afford
three-coordinate17 or distorted four-coordinate intermediates,
respectively, both capable of undergoing intramolecular C-H
activation of the phosphine ligand and, depending on the reaction
conditions, subsequent intermolecular reaction with the previ-
ously dissociated alkene.
mmol) and [HPMe3]BPh4 (68.0 mg, 0.44 mmol, prepared by anion
exchange from [HPMe3]BF4 and NaBPh4 in methanol): yield 231.0
mg (71%). Anal. Calcd for C37H44NBIrP: C, 60.32; H, 6.02; N, 1.90.
Found: C, 60.32; H, 5.62; N, 1.83. 1H NMR (CD2Cl2, 293 K): δ 1.36
(d, JHP ) 9.6, 9H, PCH3), 1.85 (s, 3H, NCCH3), 1.99, 2.25 (both m,
4H each, CH2), 4.30 (br, 4H, CH), 6.92 (m, 4H, CH), 7.05 (m, 8H,
CH), 7.36 (br, 8H, CH). 31P{1H} NMR (CD2Cl2, 293 K): δ -19.82
(s). 13C{1H} NMR (CD2Cl2, 183 K): δ 2.01 (s, NCCH3), 12.13 (d, JCP
) 35.4, PCH3), 28.56, 32.25 (both br, CH2), 64.73, 91.93, 121.49,
125.48 (all br, CH), 126.08 (s, NCCH3), 134.86 (br, CH), 163.13 (m,
JCB ) 49.0, C). All other reagents were obtained from commercial
sources and used as received. The new compounds described below
are air-sensitive in solution.
Preparation of [Ir(1-K-4,5,6-η-C8H12)(NCMe)(η2-C2H4)(PMe3)]-
BF4 (3). Method A: Ethylene was bubbled during 1 min through a
solution of 1 (20 mg, 0.04 mmol) in 1,2-dichloroethane (0.5 mL)
contained in an NMR tube. The tube was introduced in the NMR
spectrometer at 333 K, and the reaction was monitored by 31P{1H}
NMR. Quantitative transformation into 3 was complete after 15 min.
Method B: A solution of 2 (100 mg, 0.18 mmol) in CH2Cl2 (8 mL)
was stirred under ethylene atmosphere (1 bar) for 1 h at room
temperature. The resulting solution was concentrated to 0.5 mL, cooled
at 233 K, and treated with diethyl ether to give a pale yellow solid.
The solid was separated by decantation, washed with diethyl ether,
and dried in vacuo: yield 53.5 mg (55%). Anal. Calcd for C15H28-
Conclusion
The counterintuitive observation that ligands such as aceto-
nitrile, ethylene, or propylene accelerate the isomerization of
cyclooctadiene ligands at iridium centers has led to the
recognition of a mechanism for the activation of Ir(I) square-
planar compounds toward C-H oxidative additions. This
mechanism affords distorted (nonsquare-planar) four-coordinate
intermediates through a sequence of ligand coordination/
dissociation steps mediated by five-coordinate species. In the
examples examined in this work, the C-H activations preceded
(17) Werner, H.; Ho¨hn, A.; Dziallas, M. Angew. Chem., Int. Ed. Engl. 1986,
25, 1090-1092.
(18) van der Ent, A.; Onderdelinden, A. L. Inorg. Synth. 1990, 28, 90-92.
(19) Uso´n, R.; Oro, L. A.; Cabeza, J. Inorg. Synth. 1985, 23, 126-130.
9
J. AM. CHEM. SOC. VOL. 127, NO. 51, 2005 18081

A R T I C L E S
Mart´ın et al.
NBF4IrP: C, 33.84; H, 5.30; N, 2.63. Found: C, 33.44; H, 5.30; N,
2.63. MS (FAB+, m/z (%)): 405 (96) [M+-NCMe]. 1H NMR (CDCl3,
293 K): δ 0.77 (m, 1H, IrCH), 0.89, 1.16 (both m, 1H each, CH2),
1.54 (m, 4H, CH2), 1.83 (d, JHP ) 10.1, 9H, PCH3), 2.07 (m, 2H, CH2),
2.19 (br, 3H, NCCH3), 2.83, 3.59 (both m, 2H each, η2-CH2dCH2),
4.72, 5.69, 6.31 (all m, 1H each, CH). 31P{1H} NMR (CDCl3, 293 K):
δ -41.33 (s). 13C{1H} NMR (CDCl3, 293 K): δ 2.82 (s, NCCH3),
15.75 (d, JCP ) 38.1, PCH3), 21.84 (br, IrCH), 23.44 (d, JCP ) 2.2,
CH2), 25.06 (s, CH2), 48.45 (s, η2-CH2dCH2), 48.84 (d, JCP ) 7.6,
CH2), 50.16 (d, JCP ) 6.3, CH2), 77.33 (s, CH), 88.08 (d, JCP ) 19.6,
CH), 98.36 (s, CH), 117.89 (br, NCCH3).
(s, CH), 72.17, (d, JCP ) 10.2, CH), 73.83 (d, JCP ) 8.2, CH), 83.24
(s, η2-CH2dCHOC(O)CH3), 118.50 (s, NCCH3), 171.07 (s, η2-CH2d
CHOC(O)CH3).
Preparation of [Ir(1,2,5,6-η-C8H12){K-O-η2-OC(Me)OC2H3}-
(PMe3)]BF4 (6). A solution of 1 (200 mg, 0.41 mmol) in CH2Cl2 (8
mL) was reacted with vinyl acetate (2.3 mL, 24.6 mmol) and rapidly
concentrated to ca. 0.5 mL under reduced pressure. The addition of
diethyl ether produced the precipitation of a pale yellow solid, which
was separated by decantation, washed with diethyl ether, and dried in
vacuo: yield 180.9 mg (83%). Anal. Calcd for C15H27BF4IrO2P: C,
32.80; H, 4.95. Found: C, 32.38; H, 4.82. MS (FAB+, m/z (%)) 463
1
(57) [M+]. IR (cm-1): 1760, 1647, 1596 ν(CO). H NMR (CD2Cl2,
Preparation of [Ir(1-K-4,5,6-η-C8H12)(NCMe)(η2-C3H6)(PMe3)]-
BF4 (4). The compound was prepared through any of the two methods
described for 3 by using propylene instead of ethylene. Method B gave
4 as a pale yellow solid in 73% yield. Anal. Calcd for C16H30NBF4IrP:
C, 35.17; H, 5.53; N, 2.56. Found: C, 34.78; H, 5.23; N, 2.35. MS
253 K): δ 1.58 (m, 1H, CH2), 1.61 (m, 1H, κ-O-η2-CH2dCHOC(O)-
CH3), 1.76 (d, JHP ) 10.1, 9H, PCH3), 1.82 (m, 1H, κ-O-η2-CH2d
CHOC(O)CH3), 1.86, 2.09 (both m, 1H each, CH2), 2.24 (s, 3H, κ-O-
η2-CH2dCHOC(O)CH3), 2.32, 2.37, 2.40 (all m, 1H each, CH2), 2.43
(m, 1H, CH), 2.59, 2.96 (both m, 1H each, CH2), 3.02, 3.65, 3.82 (all
m, 1H each, CH), 7.26 (m, 1H, κ-O-η2-CH2dCHOC(O)CH3). 31P{1H}
NMR (CD2Cl2, 253 K): δ -39.63 (s). 13C{1H} NMR (CD2Cl2, 253
K): δ 15.01 (d, JCP ) 31.8, PCH3), 18.42 (s, κ-O-η2-CH2dCHOC-
(O)CH3), 19.79 (d, JCP ) 6.4, κ-O-η2-CH2dCHOC(O)CH3), 26.02 (d,
JCP ) 4.1, CH2), 31.57, 33.87 (both s, CH2), 35.79 (d, JCP ) 2.4, CH2),
62.38 (d, JCP ) 3.2, CH), 68.37 (d, JCP ) 6.0, CH), 73.63 (s, CH),
75.76 (d, JCP ) 9.1, CH), 90.91 (d, JCP ) 22.9, κ-O-η2-CH2dCHOC-
(O)CH3), 187.74 (d, JCP ) 10.2, κ-O-η2-CH2dCHOC(O)CH3).
Preparation of [Ir(1,2,5,6-η-C8H12){K-O-η2-OC(Me)OC2H3}-
(PMe3)]BPh4 (6-BPh4). The compound was prepared as described for
6 but starting from [Ir(1,2,5,6-η-C8H12)(NCMe)(PMe3)]BPh4 (200 mg,
0.27 mmol) and vinyl acetate (1.50 mL, 16.29 mmol): yield 186.8 mg
(88%). Anal. Calcd for C39H47BIrO2P: C, 59.92; H, 6.06. Found: C,
59.47; H, 5.69. The crystals used in the X-ray diffraction experiment
were obtained by slow diffusion of hexane into a saturated solution of
6-BPh4 in CH2Cl2 at 253 K.
1
(FAB+, m/z (%)): 459 (12) [M+], 418 (16) [M+-NCMe]. H NMR
(CD2Cl2, 233 K): δ 0.69 (m, 1H, IrCH), 0.87, 1.14, 1.51, 1.56, 1.58
(all m, 1H each, CH2), 1.68 (d, JHH ) 3.3, 3H, η2-CH2dCHCH3), 1.72
(m, 1H, CH2), 1.77 (d, JHP ) 9.9, 9H, PCH3), 2.07, 2.12 (both m, 1H
each, CH2), 2.16 (s, 3H, NCCH3), 2.85 (dd, JHH ) 9.0, 7.8, 1H, η2-
CH2dCHCH3), 3.02 (dd, JHH ) 11.7, 7.8, 1H, η2-CH2dCHCH3), 4.64
(m, 1H, CH), 4.98 (m, 1H, η2-CH2dCHCH3), 5.32, 5.75 (both m, 1H
each, CH). 31P{1H} NMR (CD2Cl2, 233 K): δ -41.23 (s). 13C{1H}
NMR (CD2Cl2, 233 K): δ 3.28 (s, NCCH3), 15.44 (d, JCP ) 37.9,
PCH3), 21.17 (s, η2-CH2dCHCH3), 22.72 (d, JCP ) 4.0, IrCH), 24.14
(d, JCP ) 2.6, CH2), 25.02 (s, CH2), 48.47 (d, JCP ) 7.4, CH2), 50.01
(d, JCP ) 6.6, CH2), 53.44 (d, JCP ) 5.1, η2-CH2dCHCH3), 64.56 (s,
η2-CH2dCHCH3), 74.29 (s, CH), 94.33 (d, JCP ) 21.7, CH), 98.71 (s,
CH), 117.72 (s, NCCH3).
Reaction of 1 with Vinyl Acetate. A solution of 1 (60.82 mg, 0.12
mmol) in CD2Cl2 (0.45 mL) contained in an NMR tube was treated
with vinyl acetate (11 µL, 0.12 mmol) and introduced in the NMR
spectrometer at 293 K. The 31P{1H} NMR spectrum displayed a singlet
attributable to complex [Ir(1,2,5,6-η-C8H12){κ-O-η2-OC(Me)OC2H3}-
(PMe3)]BF4 (6) together with a broad signal at ca. δ -19. Upon
decreasing the temperature this latter signal was found to progressively
shift to higher field and simultaneously broaden. After decoalescence,
the signal gave rise in the 193 K NMR spectrum to two major
compounds in 2:1 molar ratio, identified as the isomers [Ir(1,2,5,6-η-
C8H12)(NCMe)(η2-C2H3OC(O)CH3}(PMe3)]BF4 (5). A more detailed
evolution of the 31P{1H} NMR spectrum with the temperature can be
found as Supporting Information. Data for the major isomer of 5 is as
Preparation of [Ir(1-K-4,5,6-η-C8H12){K-O-η2-OC(Me)OC2H3}-
(PMe3)]BF4 (7). A solution of 1 (100 mg, 0.20 mmol) in CH2Cl2 (8
mL) was treated with vinyl acetate (56 µL, 0.60 mmol) and allowed to
react at the reflux temperature for 8 h. The resulting solution was filtered
through Celite, concentrated to ca. 0.5 mL, and cooled at 233 K.
Addition of diethyl ether produced the precipitation of a pale yellow
solid, which was separated by decantation, washed with diethyl ether,
and dried in vacuo: yield 81.3 mg (74%). Anal. Calcd for C15H27BF4-
IrO2P: C, 32.79; H, 4.69. Found: C, 32.59; H, 4.69. MS (FAB+, m/z
(%)): 463 (100) [M+]. IR (cm-1): 1740, 1654, 1597 ν(CO). 1H NMR
(acetone-d6, 293 K): δ 0.88 (m, 2H, CH2), 1.33 (m, 2H, CH2 + CH),
1.49 (m, 1H, CH2), 1.69 (m, 2H, CH2), 1.95 (d, JHP ) 10.8, 9H, PCH3),
2.04 (s, 3H, κ-O-η2-CH2dCHOC(O)CH3), 2.20 (m, 2H, CH2), 2.92
(ddd, JHH ) 5.5, 4.2, JHP ) 4.6, 1H, κ-O-η2-CH2dCHOC(O)CH3), 3.28
(ddd, JHP ) 10.9, JHH ) 7.2, 4.2, 1H, κ-O-η2-CH2dCHOC(O)CH3),
4.58, 5.93, 6.67 (all m, 1H each, CH), 7.95 (ddd, JHH ) 7.2, 5.5, JHP
) 4.3, 1H, κ-O-η2-CH2dCHOC(O)CH3). 31P{1H} NMR (acetone-d6,
293 K): δ -34.41 (s). 13C{1H} NMR (acetone-d6, 293 K): δ 14.64
(d, JCP ) 39.0, PCH3), 18.61 (s, κ-O-η2-CH2dCHOC(O)CH3), 22.89
(d, JCP ) 2.7, CH2), 24.48 (s, CH2), 24.91 (d, JCP ) 3.6, CH), 37.36
(d, JCP ) 6.7, κ-O-η2-CH2dCHOC(O)CH3), 48.14 (d, JCP ) 6.4, CH2),
49.78 (d, JCP ) 7.2, CH2), 79.21 (s, CH), 85.58 (d, JCP ) 6.1, κ-O-
1
follows. H NMR (CD2Cl2, 193 K): δ 1.19 (m, 1H, η2-CH2dCHOC-
(O)CH3), 1.55 (m, 1H, CH2), 1.72 (d, JHP ) 9.6, 9H, PCH3), 1.82 (m,
1H, CH2), 2.09 (s, 3H, η2-CH2dCHOC(O)CH3), 2.14 (m, 1H, CH),
2.19 (m, 1H, CH2), 2.20 (m, 1H, η2-CH2dCHOC(O)CH3), 2.27, 2.51,
2.55, 2.62, 3.13 (all m, 1H each, CH2), 3.13, 3.45, 3.75 (all m, 1H
each, CH), 6.87 (m, 1H, η2-CH2dCHOC(O)CH3). 31P{1H} NMR
(CD2Cl2, 193 K): δ -47.94 (s). 13C{1H} NMR (CD2Cl2, 193 K): δ
4.08 (s, NCCH3), 15.16 (d, JCP ) 30.4, PCH3), 20.83 (d, JCP ) 6.7,
η2-CH2dCHOC(O)CH3), 21.37 (s, η2-CH2dCHOC(O)CH3), 27.24,
30.69, 36.00, 36.46 (all s, CH2), 63.85 (d, JCP ) 5.3, CH), 68.96 (d,
JCP ) 11.5, CH), 71.78 (s, CH), 75.36 (d, JCP ) 20.3, CH), 84.81 (s,
η2-CH2dCHOC(O)CH3), 117.25 (s, NCCH3), 169.94 (s, η2-CH2d
CHOC(O)CH3). Data for the minor isomer of 5 is as follows. 1H NMR
η2-CH2dCHOC(O)CH3), 94.40 (d, JCP ) 17.7, CH), 102.41 (d, JCP
)
0.8, CH), 183.24 (d, JCP ) 3.4, κ-O-η2-CH2dCHOC(O)CH3). The
crystals used in the X-ray determination were obtained by slow diffusion
of diethyl ether into a saturated solution of 7 in CH2Cl2 at 253 K.
Preparation of [IrCl(η2-C3H6)(PiPr3)2] (28). A suspension of [Ir(µ-
Cl)(coe)2]2 (679.4 mg, 0.75 mmol) in pentane (15 mL) was treated with
PiPr3 (637 µL, 3.34 mmol) at 273 K. After 10 min of reaction, the
resulting suspension was protected from the light, treated with propylene
at atmospheric pressure, and then cooled at 233 K. The resulting solution
was concentrated to ca. 0.5 mL to give an orange solid, which was
separated by decantation, washed with pentane at 233 K, and dried in
(CD2Cl2, 193 K): δ 1.36, 1.60 (both m, 1H each, CH2), 1.72 (d, JHP
)
9.6, 9H, PCH3), 1.85 (s, 3H, η2-CH2dCHOC(O)CH3), 2.22, 2.38, 2.62
(all m, 1H each, CH2), 2.67 (m, 1H, η2-CH2dCHOC(O)CH3), 2.90 (m,
1H, CH2), 2.99 (m, 1H, CH), 3.00, 3.13 (both m, 1H each, CH2), 3.33
(m, 1H, CH), 3.34 (m, 1H, η2-CH2dCHOC(O)CH3dCH2), 3.40, 3.93
(both m, 1H each, CH), 5.63 (m, 1H, η2-CH2dCHOC(O)CH3). 31P{1H}
NMR (CD2Cl2, 193 K): δ -47.73 (s). 13C{1H} NMR (CD2Cl2, 193
K): δ 4.60 (s, NCCH3), 14.66 (d, JCP ) 31.4, PCH3), 21.13 (s, η2-
CH2dCHOC(O)CH3), 24.65 (d, JCP ) 16.4, η2-CH2dCHOC(O)CH3),
25.57, 28.59, 36.00, 37.90 (all s, CH2), 64.16 (d, JCP ) 5.0, CH), 71.60
9
18082 J. AM. CHEM. SOC. VOL. 127, NO. 51, 2005

Ligand-Promoted C−H Activations at Ir(I) Complexes
A R T I C L E S
vacuo: yield 239 mg (27%). Anal. Calcd for C21H48ClIrP2: C, 42.73;
H, 8.20. Found: C, 42.55; H, 8.38. 1H NMR (CD2Cl2, 263 K): δ 1.21
(d, JHH ) 6.21, 3H, η2-CH2dCHCH3), 1.26-1.39 (m, 36H, PCHCH3),
1.55, 1.75 (both m, 1H each, η2-CH2dCHCH3), 2.47, 2.66 (both m,
3H each, PCHCH3), 2.71 (m, 1H, η2-CH2dCHCH3). 31P{1H} NMR
(CD2Cl2, 263 K): δ 13.37, 17.55 (both d, JPP ) 360.7). 13C{1H} NMR
(CD2Cl2, 263 K): δ 18.83 (br, η2-CH2dCHCH3), 19.60, 19.70, 20.08,
(m, 1H, IrCH2), 1.17 (m, 3H, PCHCH3), 1.26-1.67 (m, 15H, PCHCH3),
1.80 (m, 1H, IrCH2), 1.92 (m, 1H, PCHCH3), 2.22 (m, 3H, PCHCH3),
3.14, 3.70 (both m, 1H each, PCHCH3). 31P{1H} NMR (toluene-d8,
253 K): δ -25.75, 15.13 (both d, JPP ) 361.8). 13C{1H} NMR (toluene-
d8, 253 K): δ -22.95 (dd, JCP ) 22.9, 3.2, IrCH2), 1.23 (s, NCCH3),
19.07, 20.39, 20.43, 21.97, 22.04 (all s, PCHCH3), 22.35 (d, JCP
)
21.1, PCHCH3), 23.82 (d, JCP ) 21.5, PCHCH3), 24.26 (d, JCP ) 22.4,
PCHCH3), 44.25 (d, JCP ) 31.2, PCHCH3), 116.48 (s, NCCH3). Data
20.36 (all s, PCHCH3), 22.58 (br, η2-CH2dCHCH3), 22.18 (d, JCP
)
17.0, PCHCH3), 23.24 (d, JCP ) 17.7, PCHCH3), 33.27 (s, η2-CH2d
CHCH3).
for the minor isomer is as follows. H NMR (toluene-d8, 253 K): δ
1
-21.63 (dd, JHP ) 14.0, 14.4, 1H, IrH), 0.96 (s, 3H, NCCH3), 1.17
(m, 3H, PCHCH3), 1.26-1.40 (m, 15H, PCHCH3), 1.46, 1.86 (both
m, 1H each, IrCH2), 1.92 (m, 1H, PCHCH3), 2.22 (m, 3H, PCHCH3),
3.25, 3.34 (both m, 1H each, PCHCH3). 31P{1H} NMR (toluene-d8,
253 K): δ -32.04, 13.43 (both d, JPP ) 362.5). 13C{1H} NMR (toluene-
d8, 253 K): δ -24.17 (d, JCP ) 24.5, IrCH2), 1.13 (s, NCCH3), 18.85
(s, PCHCH3), 19.05 (d, JCP ) 19.3, PCHCH3), 20.25, (s, PCHCH3),
Preparation of [IrClH(CHdCH2)(NCMe)(PiPr3)2] (29). A solution
of 27 (143.6 mg, 0.25 mmol) in benzene (3 mL) was treated with
acetonitrile (2.6 µL, 0.50 mmol) and stirred for 1 h while exposed to
the daylight. The resulting solution was concentrated to ca. 0.5 mL
and treated with hexane to give a white solid. The solid was separated
by decantation, washed with hexane, and dried in vacuo: yield 98.6
mg (64%). Anal. Calcd for C22H49NClIrP2: C, 42.81; H, 8.00; N, 2.27.
Found: C, 43.02; H, 7.96; N, 2.05. IR (cm-1): 2285, 2256 ν(IrH). 1H
NMR (toluene-d8, 308 K): δ -21.05 (t, JHP ) 15.9, 1H, IrH), 0.79 (s,
3H, NCCH3), 1.22 (dvt, JHH ) 7.2, N ) 12.8, 18H, PCHCH3), 1.23
(dvt, JHH ) 7.2, N ) 12.0, 18H, PCHCH3), 2.82 (m, 6H, PCHCH3),
5.26 (ddt, JHH ) 18.0, 3.6, JHP ) 1.6, 1H, IrCHdCH2), 5.98 (ddt, JHH
) 10.4, 3.6, JHP ) 1.6, 1H, IrCHdCH2), 7.89 (dd, JHH ) 18.0, 10.4,
1H, IrCHdCH2). 31P{1H} NMR (toluene-d8, 308 K): δ 11.41 (br).
13C{1H} NMR (toluene-d8, 308 K): δ 0.15 (s, NCCH3), 18.86 (s,
PCHCH3), 22.97 (vt, N ) 26.5, PCHCH3), 115.68 (s, NCCH3), 116.76
(s, IrCHdCH2), 127.87 (s, IrCHdCH2). The crystals used in the X-ray
structural determination were obtained by cooling at 253 K a saturated
solution of 29 in toluene.
21.13 (d, JCP ) 19.0, PCHCH3), 21.84 (s, PCHCH3), 24.26 (d, JCP
)
22.4, PCHCH3), 44.48 (d, JCP ) 30.5, PCHCH3), 116.67 (s, NCCH3).
Kinetics. The transformations of 1 and 28 into any of their products
were followed by registering the intensity decrease of their 31P{1H}
NMR signals as a function of time, in 1,2-dichloroethane or toluene-
d8 solutions, respectively. The lock and field homogenization of the
NMR samples in 1,2-dichloroethane were achieved by introducing a
capillary tube containing toluene-d8. For the studies under acetonitrile
excess, solutions of known concentration were prepared by mixing exact
volumes of previously prepared concentrated solutions of the starting
complex and acetonitrile and subsequent adjustment of the sample total
volume to 0.5 mL using the corresponding solvent. These manipulations
were done avoiding exposition to daylight in the case of 28. Pseudo-
first-order rate constants were obtained either directly, from the
adjustment of the entire reaction to an exponential decay function, or
indirectly from the initial zero-order rates. The CDCl3 solutions of 1
used in the variable temperature studies under acetonitrile excess were
Preparation of [IrClH(E-CHdCH2CH3)(NCMe)(PiPr3)2] (30). A
solution of 28 (100.0 mg, 0.17 mmol) in benzene (5 mL) was treated
with acetonitrile (9 µL, 0.18 mmol) and stirred for 15 min while exposed
to the daylight. The resulting solution was concentrated to 0.5 mL and
treated with hexane to give a white solid. The solid was separated by
decantation, washed with hexane, and dried in vacuo: yield 77.3 mg
(72%). Anal. Calcd for C23H51NClIrP2: C, 43.76; H, 8.08; N, 2.22.
Found: C, 43.80; H, 8.30; N, 2.21. 1H NMR (C6D6, 293 K): δ -20.89
(br, 1H, IrH), 1.00 (s, 3H, NCCH3), 1.44 (dvt, JHH ) 6.2, N ) 12.8,
18H, PCHCH3), 1.45 (dvt, JHH ) 6.3, N ) 13.0, 18H, PCHCH3), 2.04
(dt, JHH ) 6.0, JHP ) 1.7, 3H, IrCHdCHCH3), 3.00 (m, 6H, PCHCH3),
5.54 (dq, JHH ) 15.6, 6.0, 1H, IrCHdCHCH3), 7.31 (brd, JHH ) 15.6,
1H, IrCHdCHCH3). 31P{1H} NMR (C6D6, 293 K): δ 12.28 (s). 13C{1H}
NMR (C6D6, 293 K): δ 1.66 (s, NCCH3), 19.17 (s, PCHCH3), 23.61
(vt, N ) 26.3, PCHCH3), 24.83 (s, IrCHdCHCH3), 115.02 (br, IrCHd
CHCH3), 118.20 (s, NCCH3), 124.77 (br, IrCHdCHCH3). The crystals
used in the X-ray structural determination were obtained by cooling at
253 K a saturated solution of 30 in toluene.
1
also prepared as described above. The line-shape analysis of the H
NMR spectra was performed with the g-NMR program. Errors in the
activation parameters obtained from the Eyring regression were
estimated through conventional error propagation formulas,20 assuming
1K error in the temperature and a 10% error in the rate constant.
X-ray Structure Analysis of 6-BPh4, 7, 29b, and 30a. X-ray data
were collected for all complexes on a Bruker SMART APEX CCD
diffractometer with graphite monochromated Mo KR radiation (λ )
0.71073 Å) using ω scans (0.3°). Data were collected over the complete
sphere by a combination of four sets and corrected for absorption using
a multiscan method applied with the SADABS program.21 The structures
were solved by the Patterson method. Refinement, by full-matrix least-
squares on F2 using SHELXL97,22 was similar for all complexes,
including isotropic and subsequently anisotropic displacement param-
eters for all non-hydrogen nondisordered atoms. Hydrogen atoms were
included in calculated positions and refined riding on carbon atoms with
the thermal parameter related to bonded atoms or in observed posi-
tions and refined freely. Particular details concerning the presence of
solvent, static disorder, and hydride ligand refinement are listed below.
Crystal Data for 6-BPh4. C39H47BIrO2P‚CH2Cl2, M ) 866.67;
colorless plate prism, 0.18 × 0.08 × 0.04 mm3; monoclinic, P21/c; a
) 12.976(3), b ) 12.175(3), c ) 23.421(6) Å, â ) 91.686(5); Z ) 4;
V ) 3698.4(16) Å3; Dc ) 1.557 g/cm3; µ ) 3.832 mm-1, minimum
and maximum transmission factors 0.600 and 0.756; 2θmax ) 56.9°;
temperature 100.0(2) K; 29 752 reflections collected, 8688 unique
[R(int) ) 0.076]; number of data/restrains/parameters 8688/0/431; final
GOF 0.848, R1 ) 0.0468 [5554 reflections I > 2σ(I)], wR2 ) 0.0713
Preparation of [IrClH{K-P,C-P(iPr)2CH(CH3)CH2}(NCMe)-
(PiPr3)] (31). Method A: A solution of 27 (96.5 mg, 0.17 mmol) in
benzene (2 mL) was treated with acetonitrile (704 µL, 13.39 mmol)
and stirred for 2 h at 353 K. The resulting solution was concentrated
to ca. 0.5 mL, and hexane was slowly added to give a white solid. The
solid was separated by decantation, washed with hexane, and dried in
vacuo: yield 42.4 mg (43%). Method B: A suspension of [Ir(µ-Cl)-
(coe)2]2 (226.8 mg, 0.25 mmol) in pentane (15 mL) was treated with
PiPr3 (212 µL, 1.11 mmol). After 15 min at room temperature, 4 equiv
of acetonitrile (53 µL, 1.01 mmol) was added. The resulting yellow
solution was stirred during 10 min, concentrated to ca. 0.5 mL, and
treated with hexane to give a white solid. The solid was separated by
decantation, washed with hexane, and dried in vacuo: yield 198.8 mg
(67%). Anal. Calcd for C20H45NClIrP2: C, 40.77; H, 7.70; N, 2.38.
Found: C, 40.60; H, 7.40; N, 2.26. MS (FAB+, m/z (%)): 548 (32)
[M+-NCCH3]. IR (cm-1): 2271, 2207 ν(IrH). The compound was
observed by NMR to consist of two isomers in molar ratio of 7:3. Data
(20) Morse, P. M.; Spencer, M. O.; Wilson, S. R.; Girolami, G. S. Organo-
metallics 1994, 13, 1646-1655.
(21) (a) Blessing, R. H. Acta Crystallogr. 1995, A51, 33-38. (b) SADABS: Area-
detector absorption correction; Bruker-AXS: Madison, WI, 1996.
(22) (a) SHELXTL Package, v. 6.10; Bruker-AXS: Madison, WI, 2000. (b)
Sheldrick, G. M. SHELXS-86 and SHELXL-97; University of Go¨ttingen:
Go¨ttingen, Germany, 1997.
1
for the major isomer is as follows. H NMR (toluene-d8, 253 K): δ
-21.86 (dd, JHP ) 16.2, 11.7, 1H, IrH), 0.96 (s, 3H, NCCH3), 1.15
9
J. AM. CHEM. SOC. VOL. 127, NO. 51, 2005 18083

A R T I C L E S
Mart´ın et al.
for all data; largest difference peak 1.87 e‚Å-3. A dichloromethane was
observed disordered in two positions. Both dichloromethane molecules
share the same carbon atom with a turn angle of 49.7° and were refined
with occupancy factors 0.54 and 0.46, respectively. The highest
electronic residuals were observed close to the solvent.
e‚Å-3. This compound crystallizes with two molecules of benzene, one
of them located over a symmetry center and shared by two asymmetric
units. The hydride ligand was refined with fixed Ir-H bond length
and thermal parameter. The highest electronic residual was observed
close to the Ir atom and has no chemical sense.
Crystal Data for 7. C15H27BF4IrO2P, M ) 549.35; irregular block
pale yellow, 0.10 × 0.010 × 0.06 mm3; monoclinic, P21/c; a )
12.4262(10), b ) 10.0804(8), c ) 15.6876(12) Å, â ) 113.1280(10);
Computation. All calculations were carried out with the Gaussian
98 package of programs23 using the density functional theory (DFT)24
with the B3LYP functional.25 An effective core potential operator was
used to replace the innermost electrons of Ir and P.26 The basis set for
the metal was that associated with the pseudopotential,26 with a standard
valence double-ê LANL2DZ contraction.23 A 6-31G(d) basis was used
for all the ethylene and diene C atoms.27a The nitrogen N atoms were
also described with a 6-31G(d) basis. The rest of the atoms were
described with a 6-31G basis set.27 The transition states found have
been confirmed by frequency calculations, and the connection between
the starting and final reactants has been checked by slightly perturbing
the TS geometry toward the minima geometries and reoptimizing.
Z ) 4; V ) 1807.1(2) Å3; Dc ) 2.019 g/cm3; µ ) 7.521 mm-1
,
minimum and maximum transmission factors 0.463 and 0.625; 2θmax
) 57.0°; temperature 100.0(2) K; 21 415 reflections collected, 4331
unique [R(int) ) 0.050]; number of data/restrains/parameters 4331/1/
242; final GOF 0.900, R1 ) 0.0307 [3248 reflections I > 2σ(I)], wR2
) 0.0568 for all data; largest difference peak 2.08 e‚Å-3. The highest
electronic residual was observed close to the Ir atom and has no
chemical sense.
Crystal Data for 29b. C22H49ClIrNP2‚0.50(C14H16), M ) 709.35;
colorless prism, 0.10 × 0.08 × 0.08 mm3; triclinic, P-1; a )
8.4587(19), b ) 11.443(3), c ) 17.379(4) Å, R ) 78.694(6), â )
78.537(7), γ ) 87.473(6); Z ) 2; V ) 1616.6(6) Å3; Dc ) 1.457 g/cm3;
µ ) 4.328 mm-1, minimum and maximum transmission factors 0.714
and 0.759; 2θmax ) 58.0°; temperature 100.0(2) K; 20 500 reflections
collected, 7777 unique [R(int) ) 0.041]; number of data/restrains/
parameters 7777/48/298; final GOF 0.919, R1 ) 0.0418 [6435
reflections I > 2σ(I)], wR2 ) 0.0692 for all data; largest difference
peak 1.80 e‚Å-3. In the last cycles of the refinement a phosphine ligand
was observed disordered in two positions, due to a turn on the P-Ir
bond of about 21°. This ligand was refined with restrained geometry
and thermal parameters. Two toluene molecules were also observed
disordered over symmetry centers. The hydride ligand was refined with
fixed Ir-H bond length. The highest electronic residual was observed
close to the Ir atom and has no chemical sense.
Acknowledgment. This research was supported by Plan
Nacional de Investigacio´n MEC/FEDER (Project BQU2003-
05412).
Supporting Information Available: Complete ref 23; VT
NMR experiments and simulations; electronic energies and
Cartesian coordinates for the computed species; X-ray crystal-
lographic file (CIF). This material is available free of charge
via the Internet at http://pubs.acs.org.
JA0557233
(23) Frisch, M. J.; et al. Gaussian 98, revision A.6; Gaussian, Inc.: Pittsburgh,
PA, 1998.
(24) (a) Parr, R. G.; Yang, W. Density Functional Theory of Atoms and
Molecules; Oxford University Press: Oxford, U.K., 1989. (b) Ziegler, T.
Chem. ReV. 1991, 91, 651-667.
Crystal Data for 30a. C23H51ClIrP2‚1.5C6H6, M ) 748.40; colorless
irregular block, 0.24 × 0.14 × 0.08 mm3; triclinic, P-1; a ) 8.9291(4),
b ) 11.4598(5), c ) 18.4112(8) Å, R ) 84.5070(10), â ) 82.5280(10),
γ ) 70.6030(10); Z ) 2; V ) 1759.14(13) Å3; Dc ) 1.413 g/cm3; µ )
3.982 mm-1, minimum and maximum transmission factors 0.522 and
0.643; 2θmax ) 57.8°; temperature 100.0(2) K; 22 169 reflections
collected, 8437 unique [R(int) ) 0.031]; number of data/restrains/
parameters 8437/1/357; final GOF 0.901, R1 ) 0.0266 [7447 reflections
I > 2σ(I)], wR2 ) 0.0528 for all data; largest difference peak 1.69
(25) (a) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785-789. (b)
Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652. (c) Stephens, P. J.;
Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. J. Phys. Chem. 1994, 98,
11623-11627.
(26) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299-310.
(27) (a) Hehre, W. J.; Ditchfield, R.; Pople, J. A. J. Chem. Phys. 1972, 56,
2257-2261. (b) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973,
28, 213-222. (c) Francl, M. M.; Pietro, W. J.; Hehre, W. J.; Binkley, J.
S.; Gordon, M. S.; DeFrees, D. J.; Pople, J. A. J. Chem. Phys. 1982, 77,
3654-3665.
9
18084 J. AM. CHEM. SOC. VOL. 127, NO. 51, 2005