Iridium compounds with κ- P,P,Si (biPSi) pincer ligands: Favoring reactive structures in unsaturated complexes

Article abstract of DOI:10.1021/ja102479h

The structure, coordination properties, insertion processes, and dynamic behavior in solution of the five-coordinate complexes [IrXH(biPSi)] (biPSi = κ-P,P,Si-Si(Me){(CH₂)₃PPh₂}₂; X = Cl (1), Br (2), or I (3)) h

Full text of DOI:10.1021/ja102479h

Published on Web 06/11/2010
Iridium Compounds with K-P,P,Si (biPSi) Pincer Ligands:
Favoring Reactive Structures in Unsaturated Complexes
Eduardo Sola,* Alba Garc´ıa-Camprub´ı, Jose´ L. Andre´s, Marta Mart´ın, and
Pablo Plou
Departamento de Qu´ımica de Coordinacio´n y Cata´lisis Homoge´nea, Instituto de Ciencia de
Materiales de Arago´n, CSIC-UniVersidad de Zaragoza, E-50009 Zaragoza, Spain
Received March 24, 2010; E-mail: sola@unizar.es
Abstract: The structure, coordination properties, insertion processes, and dynamic behavior in solution of
the five-coordinate complexes [IrXH(biPSi)] (biPSi ) κ-P,P,Si-Si(Me){(CH2)3PPh2}2; X ) Cl (1), Br (2), or
I (3)) have been investigated. The compounds are formed as mixtures of two isomers, anti and syn, in
slow equilibrium in solution. The equilibrium position depends on the halogen and the solvent. Both isomers
display distorted square-based pyramidal structures in which the vacant position sits trans to silicon. The
equatorial plane of the syn isomer is closer to the T structure due to distortions of steric origin. The small
structural differences between the isomers trigger remarkable differences in reactivity. The syn isomers
form six-coordinate adducts with chlorinated solvents, CO, P(OMe)3, or NCMe, always after ligand
coordination trans to silicon. The anti isomers do not form detectable adducts with chlorinated solvents
and coordinate CO or P(OMe)3 either trans to silicon (kinetic) or trans to hydride (thermodynamic). NCMe
coordinates the anti isomers exclusively at the position trans to hydride. Qualitative and quantitative details
(equilibrium constants, enthalpies, entropies, etc.) on these coordination processes are given and discussed.
As a result of the different coordination properties, insertion reagents such as acetylene, diphenylacetylene,
or the alkylidene resulting from the decomposition of ethyl diazoacetate selectively insert into the Ir-H
bond of 1-syn, not into that of 1-anti. These reactions give five-coordinate syn alkenyl or alkyl compounds
in which the vacancy also sits trans to silicon. Acetylene is polymerized in the coordination sphere of 1.
The nonreactive isomer 1-anti also evolves into the syn insertion products via antiTsyn isomerizations,
the rates of which are notably dependent on the nature of the insertion reactants. H2 renders antiTsyn
isomerization rates of the same order as the NMR time scale. The reactions are second order (kobs
)
k(antiTsyn)[H2]) and do not involve H2/IrH hydrogen atom scrambling. A possible isomerization mechanism,
supported by MP2 calculations and compatible with the various experimental observations, is described.
It involves Ir(V) intermediates and a key σ Ir-(η2-SiH) agostic transition state. A similar transition state
could also explain the antiTsyn isomerizations in the absence of oxidative addition reactants, although at
the expense of high kinetic barriers strongly dependent on the presence of potential ligands and their nature.
Introduction
contributions that may prejudice kinetics, although, especially
in cases of ligand dissociation, the negative effects can go even
Coordinatively unsaturated transition metal complexes are
ubiquitous in catalytic transformations as either catalyst precur-
sors or reaction intermediates, or both.1 Depending on the
catalytic application, these complexes coordinate “reactive”
ligands such as hydride, alkyl, alkenyl, acyl, carbene, etc.sall
of them strong σ-donors and therefore good trans-directors.2
Accordingly, these ligands push coordination vacancies trans
to them,3 thus favoring structures which are generally inap-
propriate for catalytic applications following ordinary inner-
sphere mechanisms. As a consequence, the catalytic cycles must
include reorganization steps through which the catalysts make
available to the incoming substrates vacancies cis to the reactive
ligands. Such reorganizationsstypically distortions, isomeriza-
tions, or further ligand dissociationssrequire additional energetic
further and jeopardize catalyst stability. A good example of this
can be found in Grubbs’ catalyst [RuCl2(dCHPh)(PCy3)2],4 a
d6 square-pyramidal compound with the vacancy trans to the
carbene ligand. Despite its coordinative unsaturation, this
complex starts catalysis only after phosphine dissociation to form
a 14 e- species; this process is responsible for practical
difficulties such as the need for an initiation period, the limited
catalyst life, and the incompatibility of this catalyst with others
in tandem processes.5 Subsequent generations of olefin metath-
esis catalysts derived from this compound have minimized the
extent of such drawbacks,6 although they retain the adverse
structure that is their origin.
(1) Van Leeuwen, P. W. N. M. Homogeneous Catalysis. Understanding
the Art; Kluwer Academic Publishers: Dordrecht, 2004.
(2) Coe, B. J.; Glenwright, S. J. Coord. Chem. ReV. 2000, 203, 5–80.
(3) Jean, Y. Molecular Orbitals of Transition Metal Complexes; Oxford
University Press: New York, 2005.
(4) Grubbs, R. H. Tetrahedron 2004, 60, 7117–7140.
(5) Goldman, A. S.; Roy, A. H.; Huang, Z.; Ahuja, R.; Schinski, W.;
Brookhart, M. Science 2006, 312, 257–261.
(6) Handbook of Metathesis; Grubbs, R. H., Ed.; Wiley-VCH: Weinheim,
2003.
9
10.1021/ja102479h  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 9111–9121 9111

A R T I C L E S
Sola et al.
The adoption of such unfavorable structures can be combated
through ligand design. An obvious strategy toward this end is
to use polydentate rigid ligands such as the fac-coordinating
Cp and Tp,7 much of whose success in catalysis might be
attributed to their ability to confine reactive ligands and
coordination vacancies mutually cis.8 A second alternative, a
priori compatible with different ligand types and coordination
modes, would be based on ancillary ligands more trans-directing
than those referred to above as reactive. Following generally
accepted conclusions on this subject,2,9 this would only be within
the reach of anionic ligands such as carbine, sulfide, oxide,
nitride, bent nitrosyl, stannyl, silyl, or boryl. Among them,
stannyl,10 silyl,11 and boryl12 ligands could indeed engender a
variety of stereoelectronic environments attractive for catalysis,
although, due to their high reactivity, a stabilization rendering
them ancillary would be necessary. This has already been found
feasible for silyl moieties within polydentate ligands of the types
κ-P,P,Si (biPSi)13,14 and κ-N,N,Si15 which, in addition, have
emerged capable of stabilizing coordinatively unsaturated
compounds and catalysts.16
The above considerations suggest that this type of silyl-
containing ligands is likely to confer kinetic advantages to their
complexes and drive genuine reactivities and catalytic properties.
In order to substantiate these potential pros and characterize
possible cons, we have conducted this study on the five-
coordinate d6 complexes [IrXH(biPSi)] (X ) halogenide; biPSi
) κ-P,P,Si-Si(Me){(CH2)3PPh2}2) focusing on basic aspects of
their chemistry that are likely affected by the peculiar charac-
teristics of the biPSi pincer ligand:17 structure, coordination
properties, and dynamic behavior in solution. The study confirms
that the silyl entity can indeed induce distinctive, more reactive,
ground-state structures in five-coordinate hydride, alkenyl, and
alkyl iridium(III) complexes but also concludes other less-
anticipated features. Thus, the severe weakening of the bonds
trans to silicon intensifies the relative impact of steric factors
over coordination, enabling effective discriminations from
otherwise negligible differences. Also worth mentioning is the
accessibility of agostic Ir-(η2-SiH) transition states (TSs), a
genuine reactivity resource that widens and simplifies the
chemistry of hydride derivatives.
Results and Discussion
Isomers of [IrXH(biPSi)] and Structures. The complex
[IrClH(biPSi)] (1) was previously reported to result from the
reaction between dimer [Ir(µ-Cl)(cod)]2 and the biPSi ligand
precursor HSi(Me){(CH2)3PPh2}2.13 On the basis of the NMR
data obtained from freshly prepared solutions in CDCl3, 1 was
described as a single isomer displaying an anti orientation of
the hydride ligand relative to the methyl group at silicon (1-
anti, eq 1). Nevertheless, we have now observed that, after
several days at room temperature, these solutions evolve
equilibrium mixtures of the two isomers, 1-anti and 1-syn. The
position of the 1-anti:1-syn equilibrium is 93:7 in C6D6 or
toluene-d8, irrespective of the temperature, but slightly dimin-
ishes to ca. 83:17 in solvents such as CDCl3 or CD2Cl2. The 1H
NOESY NMR spectra of these mixtures display clear cross-
peaks between the hydride and methyl signals of the minor
isomer, thus confirming the relative syn orientation of these
groups (eq 1).
(7) Trofimenko, S. Scorpionates. The Coordination Chemistry of Poly-
pyrazolylborate Ligands; Imperial College Press: London, 1999.
(8) Best examples are in olefin polymerizations catalyzed by group 4
metalocenes. See: Resconi, L.; Chadwick, J. C.; Cavallo, L. In
ComprehensiVe Organometallic Chemistry, 3rd ed.; Mingos, D. M. P.,
Crabtree, R. H., Eds.; Elsevier: Amsterdam, 2006; Vol. 4, pp 1006-
1166.
Analogues of 1 with bromide and iodide ligands instead of
chloride have been obtained by reaction of CH2Cl2 solutions of
1 with the corresponding sodium halogenide in excess. The
compounds [IrBrH(biPSi)] (2) and [IrIH(biPSi)] (3) have also
been isolated as kinetic mixtures of anti and syn isomers, the
compositions of which depend on the actual reaction conditions
and isolation procedures. As for 1, these kinetic mixtures have
been observed to slowly evolve solvent-dependent anti:syn
equilibria. The bromide complex 2 behaves similarly to its
chloride precursor, affording anti:syn equilibria of molar
compositions 92:8 and 82:18 in C6D6 and CDCl3, respectively.
The equilibrium proportion of the syn isomer is larger for the
iodide compound 3, which gives rise to nearly equimolar
mixtures in C6D6, 45:55, but enriched in the syn isomer in
CD2Cl2 or CDCl3, ca. 22:78.
The X-ray structures of the two isomeric bromide complexes,
2-anti and 2-syn, are shown in Figure 1. Relevant bond distances
and angles of these two structures, together with those of 3-syn,
can be found as Supporting Information. Both isomers 2 display
distorted square-based pyramidal structures in which the vacant
position sits trans to silicon. This is in contrast with the
structures adopted by isoelectronic analogues of 2 coordinating
other pincer diphosphines with sp2 carbon or nitrogen donor
atoms, in which the vacant site is trans to hydride.18 Then the
structures confirm that, as expected, the trans influence of the
sp3 silicon dominates over that of hydride, rendering isoelec-
tronic five-coordinate d6 complexes non-isostructural.
(9) Zhu, J.; Lin, Z.; Marder, T. B. Inorg. Chem. 2005, 44, 9384–9390.
(10) Smith, N. D.; Mancuso, J.; Lautens, M. Chem. ReV. 2000, 100, 3257–
3282.
(11) Corey, J. Y.; Braddock-Wilking, J. Chem. ReV. 1999, 99, 175–292.
(12) Irvine, G. J.; Lesley, M. J. G.; Marder, T. B.; Norman, N. C.; Rice,
C. R.; Robins, E. G.; Roper, W. R.; Whittell, G. R.; Wright, L. J.
Chem. ReV. 1998, 98, 2685–2722.
(13) Stobart, S. R.; Brost, R. D.; Bruce, G. C.; Joslin, F. L. Organometallics
1997, 16, 5669–5680.
(14) (a) Bushnell, G. W.; Casado, M. A.; Stobart, S. R. Organometallics
2001, 20, 601–603. (b) Auburn, M. J.; Holmes-Smith, R. D.; Stobart,
S. R.; Bakshi, P. K.; Cameron, T. S. Organometallics 1996, 15, 3032–
3036. (c) Grundy, S. L.; Holmes-Smith, R. D.; Stobart, S. R.; Williams,
M. A. Inorg. Chem. 1991, 30, 3333–3337. (d) MacInnis, M. C.;
MacLean, D. F.; Lundgren, R. J.; McDonald, R.; Turculet, L.
Organometallics 2007, 26, 6522–6525. (e) MacLean, D. F.; McDonald,
R.; Ferguson, M. J.; Caddell, A. J.; Turculet, L. Chem. Commun. 2008,
5146–5148. (f) Morgan, E.; MacLean, D. F.; McDonald, R.; Turculet,
L. J. Am. Chem. Soc. 2009, 131, 14234–14236.
(15) (a) Sangtrirutnugul, P.; Tilley, T. D. Organometallics 2007, 26, 5557–
5568. (b) Sangtrirutnugul, P.; Tilley, T. D. Organometallics 2008, 27,
2223–2230. (c) Sangtrirutnugul, P.; Stradiotto, M.; Tilley, T. D.
Organometallics 2006, 25, 1607–1617. (d) Stradiotto, M.; Fujdala,
K. L.; Tilley, T. D. Chem. Commun. 2001, 1200–1201.
(16) This capability has also been recognized for boryl moieties in pincer
phosphines: (a) Segawa, Y.; Yamashita, M.; Nozaki, K. J. Am. Chem.
Soc. 2009, 131, 9201–9203. (b) Segawa, Y.; Yamashita, M.; Nozaki,
K. Organometallics 2009, 28, 6234–6242.
(17) (a) The Chemistry of Pincer Compounds; Morales-Morales, D., Jensen,
C., Eds.; Elsevier: Amsterdam, 2007. (b) van der Boom, M. E.;
Milstein, D. Chem. ReV. 2003, 103, 1759–1792. (c) Albrecht, M.; van
Koten, G. Angew. Chem., Int. Ed. 2001, 41, 3750–3781.
The small structural differences between isomers 2 deserve
careful consideration since, as will be described below, they
9
9112 J. AM. CHEM. SOC. VOL. 132, NO. 26, 2010

biPSi Complexes of Iridium
A R T I C L E S
Figure 2. (Left) Variation with temperature of the 31P NMR chemical shifts
of isomers 1 in CD2Cl2. (Right) Sigmoidal plots on the 31P NMR chemical
shift dependence upon 1/T for 1-syn.
(UFF methods in ONIOM approach) to afford an optimized
structure close to the experimental one.
Thus, contrary to what the geometric regularity of isomer
2-syn seems to suggest, the structure of 2-anti is the electroni-
cally preferred one, whereas that of 2-syn results from distortions
provoked by the sterics. Examination of space-filling models
indicates that the bromide ligand of 2-anti occupies a position
very conflicting in steric terms, wedged among four phenyl rings
regardless of its exact position in the equatorial plane. Since
there is no way to avoid or minimize such steric conflict, there
is no reason to distort. In contrast, the halide of the syn isomer
can elude any conflict with the phenyls by approaching the silyl
moiety. Hence, there is a privileged halide position from the
steric point of view and, consequently, a good reason to distort.
According to the available structural data, the characteristic
structure of each isomer is retained by changing the halogenide.
In contrast, the relative energies of the isomers are likely to
vary with halogen size, progressively disfavoring the more
encumbered anti isomers. This qualitative observation explains
the relative destabilization of the iodide complex 3-anti observed
experimentally.
Figure 1. (Top) X-ray structures of isomers 2 and angles of their equatorial
planes. (Bottom) Equatorial planes of MP2-optimized models of isomers 2
without and with phenyl rings.
trigger remarkable differences in reactivity. The geometry of
the syn isomer displays little deviation from an octahedron with
a vacant position, showing bond angles around the metal close
to 90 and 180° (T structure).3 On the other hand, the P-Ir-P
angle of the anti isomer decreases to 163°, while the distribution
of ligands in the equatorial plane approaches that of a distorted
trigonal bipyramid (Y structure). Since these differences are
likely to result from steric and conformational adjustments
within a common electronically preferred structure,19 we have
optimized the geometries of various models for isomers 2 to
understand the structural impact of each separate contribution.
The MP2-based optimization of model complexes without
phenyl rings (H instead of Ph) reproduces the experimental
disposition of equatorial ligands only for the most stable isomer
2-anti, suggesting that repulsions coming from the axial
substituents do not contribute to this structure (Figure 1, bottom).
This is despite the fact that such steric contributions are usually
important in the presence of pincer-type diphophines, for which
the ligand skeleton imposes an eclipsed conformation of the
two PR2 moieties that maximizes their steric impact on the eq-
uatorial plane.20 In agreement with this latter observation, the
model of isomer 2-syn indeed requires the inclusion of phenyls
Coordination Properties. As a result of their subtle structural
differences, the syn and anti isomers of [IrXH(biPSi)] complexes
constitute slightly different starting points for the formation of
six-coordinate adducts. Coordination trans to silicon should be
relatively favored in the distorted syn isomers, which are
preorganized to form octahedral compounds. This advantage
can be readily recognized by the potential ligands, especially
by those weakly coordinating. Figure 2 illustrates the temper-
ature dependence of the 31P NMR chemical shifts corresponding
to the anti and syn isomers of 1 in CD2Cl2, which can be
interpreted as the result of a solvent coordination equilibrium
exclusively affecting to the syn isomer (eq 2).21 The phenom-
1
enon has also been observed, by H and 31P NMR, for the syn
isomers of the analogous complexes 2 and 3 in both CDCl3
and CD2Cl2. The dependences of 31P NMR chemical shifts on
the inverse of temperature have been adjusted to symmetric
sigmoidal functions (Figure 2, right) to directly estimate the
∆H° and ∆S° thermodynamic parameters of the proposed
solvent coordination. Values of ∆H° in the range -5.5((0.5)
kcal mol-1 and negative enthalpy increments ∆S° of about
-26((2) cal K-1 mol-1 have been obtained, irrespective of the
compound and the solvent (see Experimental Section and
Supporting Information for details). This small additional
(18) Representative structures: (a) Go¨ttker-Schnetmann, I.; White, P.;
Brookhart, M. J. Am. Chem. Soc. 2004, 126, 1804–1811. (b) Grimm,
J. C.; Nachtigal, C.; Mack, H.-G.; Kaska, W. C.; Mayer, H. A. Inorg.
Chem. Commun. 2000, 3, 511–514. (c) Ben-Ari, E.; Gandelman, M.;
Rozenberg, H.; Shimon, L. J. W.; Milstein, D. J. Am. Chem. Soc.
2003, 125, 4714–4715.
(19) Other possible causes of distortion such as the formation of agostic
or hypervalent interactions can be discarded in the view of the distances
around the Si atoms. See: Nikonov, G. I. AdV. Organomet. Chem.
2005, 53, 217–309.
(20) (a) Gusev, D. G.; Lough, A. J. Organometallics 2002, 21, 5091–5099.
(b) Ujaque, G.; Maseras, F.; Eisenstein, O.; Liable-Sands, L.; Rhei-
ngold, A. L.; Yao, W.; Crabtree, R. H. New J. Chem. 1998, 1493–
1498.
(21) Precedents for CH2Cl2 coordination to Ir(III): (a) Klei, S. R.; Golden,
J. T.; Burger, P.; Bergman, R. G. J. Mol. Catal. A 2002, 189, 79–94.
(b) Arndtsen, B. A.; Bergman, R. G. Science 1995, 270, 1970–1973.
9
J. AM. CHEM. SOC. VOL. 132, NO. 26, 2010 9113

A R T I C L E S
Sola et al.
The anti*:syn equilibrium distributions found for the CO and
P(OMe)3 adducts 4 and 5 (98:2 and 85:15, respectively) are
rather similar despite the different ligand size. In contrast,
changing the halogenide ligand has a very large effect on the
composition of the anti*:syn equilibrium, which shifts to 48:52
for the bromide adduct [IrBrH(biPSi)(CO)] (6) and to 0:100
for the iodide analogue 7. Thus, looking exclusively at the
thermodynamic distributions of isomers (those observable after
several hours under CO), the simple replacement of chloride
by iodide inverts the stereochemistry of the reaction product
and switches the CO coordination position from trans to hydride
(98%) to trans to silicon (100%). Even though this result can
be rationalized on the basis of conventional steric arguments
already pointed out in the previous section, its forcefulness
exceeds the commonly modest nature of the halide effects.23
In the same line of extreme behaviors, a ligand such as
acetonitrile has been found to bind each isomer of complexes
1-3 at a different site. At room temperature and in the presence
of a moderate acetonitrile excess, the 1H NMR spectra of
solutions of 1-3 in CDCl3 or toluene-d8 have been observed to
display narrow signals corresponding to the syn adducts
[IrXH(biPSi)(NCMe)] {X ) Cl (8), Br (9), and I (10)} but broad
ones attributable to anti isomers (Figure 4). Decoalescence of
the latter signals at low temperature has evidenced slow
equilibria between the five-coordinate anti precursors and their
acetonitrile adducts. Integration of the corresponding 1H hydride
and acetonitrile signals in the low-temperature NMR spectra
has afforded the equilibrium constants at each temperature. In
addition, these spectra have allowed identification of the
coordination position for acetonitrile, which is that trans to
hydride (anti*). This follows from the 13C signals corresponding
to the quaternary carbon of the acetonitrile ligand (NCMe) with
1
Figure 3. (Top) Hydride H NMR signals of 1 (upper spectrum) and its
kinetic and thermodynamic products of reaction with CO, respectively (C6D6,
300 K). (Bottom) Hydride 1H and 29Si{1H} NMR signals of isomers 5-anti*
(left) and 5-syn (right) in CDCl3.
stabilization of the syn isomers accounts for the aforementioned
dependence of the anti:syn equilibrium position upon the solvent.
A more complete picture on the coordination capabilities of
the [IrXH(biPSi)] complexes has been obtained from their
reactions with the small and strongly coordinating CO. Figure
3 illustrates how both isomers of 1 initially coordinate CO at
the position trans to silicon to form the [IrClH(biPSi)(CO)]
adducts, 4-anti and 4-syn. The kinetic product 4-anti then further
evolves into a new compound which, attending to its NMR
parameters, is another anti isomer (NOE) and coordinates CO
at the position trans to hydride. The latter is inferred from the
1
and without decoupling of the hydride H resonance (Figure
1
marked displacement of the H NMR signal corresponding to
3
4), which evidence a JCH coupling constant of 9.8 Hz only
the hydride, from δ -19.97 to -7.05, consistent with the
coordination of a π-acceptor ligand trans to it.22 The structure
of this new coordination isomer, referred to as anti*, can be
unambiguously assigned with the help of adduct [IrClH-
(biPSi){P(OMe)3}] (5), the P(OMe)3 analogue of 4. This
compound has also been observed to initially form a kinetic
mixture of isomers 5-anti and 5-syn, subsequently evolving into
an equilibrium of 5-anti* and 5-syn. Figure 3 shows the
characteristic JHP and JSiP NMR coupling constants that confirm
the structures proposed in eq 3.
compatible with a trans relative position of hydride and NCMe.
On the other hand, the NMR signals corresponding to the syn
(22) For examples, see: (a) Sola, E.; Navarro, J.; Lo´pez, J. A.; Lahoz, F. J.;
Oro, L. A.; Werner, H. Organometallics 1999, 18, 3534–3546. (b)
Navarro, J.; Sola, E.; Mart´ın, M.; Dobrinovitch, I. T.; Lahoz, F. J.;
Oro, L. A. Organometallics 2004, 23, 1908–1917. (c) Sola, E.;
Bakhmutov, V. I.; Torres, F.; Elduque, A.; Lo´pez, J. A.; Lahoz, F. J.;
Werner, H.; Oro, L. A. Organometallics 1998, 17, 683–696.
Figure 4. (Left) Room-temperature hydride 1H NMR signals of 3 in CDCl3,
before and after addition of 10 equiv of acetonitrile. (Right) 13C NMR signal
corresponding to the NCMe ligand of 8-anti* at 223 K: (top) 13C{1H} and
(bottom) allowing hydride coupling.
9
9114 J. AM. CHEM. SOC. VOL. 132, NO. 26, 2010

biPSi Complexes of Iridium
A R T I C L E S
Figure 6. Experimental enthalpy profile for the coordination of acetonitrile
to 1 in toluene. The 1.5 kcal mol-1 difference between isomers 1 follows
from the anti:syn equilibrium position in toluene (93:7), considering ∆S°
) 0.
Figure 5. Van’t Hoff representation of the equilibrium constants for
acetonitrile coordination to 1-3 to form 8-10, experimentally determined
by NMR in toluene-d8.
Table 1. Enthalpies and Entropies of Formation of the Acetonitrile
Adducts 8-10
librium constants of the iodide analogue 10-anti* are about 1
order of magnitude higher and lead to a higher formation
enthalpy. Rather than a greater stability of this iodide adduct,
the improved thermodynamic balance likely reflects, once again,
the lower stability of the precursor 3-anti. Besides steric fac-
tors, the smaller π-donating capability of iodide compared to
that of chloride and bromide could be added to the list of
possible causes for the different thermodynamics,23 since such
a bond contribution must disappear in the six-coordinated
adducts.
X
isomer
∆H° (kcal mol-1
)
∆S° (cal K-1 mol-1
)
Cl
Cl
Br
Br
I
8-anti*
8-syn
9-anti*
9-syn
10-anti*
10-syn
-7.7((0.4)
-13.4((0.6)
-8.7((0.5)
-15.1((0.9)
-11.2((0.7)
-13.9((1.3)
-26((2)
-31((2)
-30((2)
-36((2)
-34((3)
-33((3)
I
isomers of 8-10 reveal the coordination equilibria of eq 4 only
above 300 K, when the equilibrium concentrations of the five-
coordinate precursors 1-3-syn become significant. Under such
conditions, the equilibria are fast on the NMR time scale and,
as detailed above for the coordination of chlorinated solvents,
generate averaged signals whose temperature-dependent chemi-
cal shifts permit quantitative determinations of the equilibrium
positions.
First-order rate constants for acetonitrile dissociation (k-1
)
have been measured for complex 8-anti* in toluene-d8 between
1
273 and 236 K. Within this temperature range, the hydride H
NMR signals of both complexes in equilibrium were simulta-
neously observable, while the rate constants remained of a
magnitude accessible by spin saturation transfer methods. These
rate constants have allowed an estimation of the activation
parameters for acetonitrile dissociation from 8-anti* as ∆Hq )
19.1((0.7) kcal mol-1 and ∆Sq ) 20((2) cal K-1 mol-1
.
Unfortunately, a similar quantitative determination has not been
1
possible for the 8-syn isomer, although its narrow H NMR
averaged acetonitrile resonance (free + coordinated to 8-syn)
at 236 K ensures that the equilibrium between syn species
remains fast on the NMR time scale. Assuming similar
activation entropies for ligand dissociation from both isomers
8, this acetonitrile resonance line width would imply a dis-
sociation enthalpy from 8-syn significantly below that estimated
for 8-anti*. This latter qualitative consideration, together with
the quantitative data collected for 8, are used to assemble the
energy profile of Figure 6, which summarizes the experimental
observations on acetonitrile coordination to complex 1. The
figure highlights the different coordination position of aceto-
nitrile in each isomer and the fact that the syn adduct is
simultaneously more stable and more labile than its anti* isomer.
Figure 7 shows the X-ray structures of two syn adducts
already described in this section, 6-syn and 8-syn. At this point
of the discussion, their perfectly octahedral structures deserve
no comment except for their long Ir-C and Ir-N distances trans
to silicon, 1.958(3) and 2.148(5) Å, respectively (see Supporting
Information for more structural details). The latter is similar to
the bond lengths of other labile acetonitrile ligands trans to
hydrides or alkenyls in related Ir(III) compounds.22,24 Although
we have not been able to grow suitable crystals for any of the
anti or anti* isomers in this section, Figure 7 includes the
Before discussing the quantitative aspects of acetonitrile
coordination to 1-3, it should be noted that, irrespective of the
compound and the isomer, the equilibria in eq 4 are established
much faster (seconds) than those converting anti species into
syn and vice versa (days). Hence, each five-coordinate isomer
can be assumed to be in equilibrium with its corresponding
acetonitrile adduct even if the anti:syn composition, which
evolves very slowly, is outside equilibrium. In fact, the equilibria
of eq 4 have been quantified by NMR despite the isomers 8-10-
anti* eventually disappearing in favor of their more stable syn
isomers. The Van’t Hoff representations (ln K vs 1/T) of the
experimentally determined equilibrium constants are shown in
Figure 5, and the adduct formation enthalpies and entropies
estimated from these representations are listed in Table 1.
According to these experimental determinations, the three syn
acetonitrile adducts display indistinguishable thermodynamic
properties, including a common formation enthalpy of about
14 kcal mol-1 and negative entropy increments around -33 cal
K-1 mol-1. The anti* adducts are less stable than their syn
isomers, in agreement with their eventual disappearance from
the product mixtures. The chloride and bromide complexes 8-
and 9-anti* show equal thermodynamic values, but the equi-
(24) (a) Mart´ın, M.; Torres, O.; On˜ate, E.; Sola, E.; Oro, L. A. J. Am.
Chem. Soc. 2005, 127, 18074–18084. (b) Canepa, G.; Sola, E.; Mart´ın,
M.; Lahoz, F. J.; Oro, L. A.; Werner, H. Organometallics 2003, 22,
2151–2160.
(23) Fagnou, K.; Lautens, M. Angew. Chem., Int. Ed. 2002, 41, 26–47.
9
J. AM. CHEM. SOC. VOL. 132, NO. 26, 2010 9115

A R T I C L E S
Sola et al.
Many features of the insertion reaction in eq 5 change
substantially by changing the alkyne from diphenylacetylene
(bulky and internal) to acetylene (small and terminal). Actually,
the only similarity between the two reactions has been the
isolation of analogous syn insertion productssthe vinyl complex
syn-[IrCl(CHdCH2)(biPSi)] (13) in the case of acetylene. In
connection with the main issue this work addresses, the T
structure of 13 shown in Figure 8 confirms that the sp3 silicon
donor atom is able to place the coordination vacancy in the
trans position also in the presence of an alkenyl, another trans-
directing ligand assiduous in many catalytic transformations.
Forcomparison,thestructureofcomplex[IrCl(CHdCHPh)(PCP)],
an analogue of 13 with a conventional pincer ligand, is also an
almost perfect T, but the vacancy lies trans to the alkenyl ligand
instead of cis.26
Unlike the reaction with diphenylacetylene, the disappearance
of 1-anti in the presence of acetylene has been observed to be
fast. Furthermore, under excess acetylene, the vinyl complex
13 has been found to be a transient reaction product rather than
the final one. Due to its high solubility in hexane, 13 has been
isolated pure after treating with this solvent the dry residues of
reactions carried out in CH2Cl2. However, the yield of 13, always
poor, has been found to decrease with time due to the increasing
formation of a relatively insoluble dark-colored material. This
material remained soluble enough in THF for GPC analysis,
which suggests that it is (or contains) polyacetylene. Two
different samples, obtained after 14 and 24 h of reaction between
1 and excess acetylene (ca. 1 bar) at room temperature, have
produced GPC peaks of number-average molecular weights (Mn)
2600 and 4900, respectively, with a common polydispersity ratio
(Mw/Mn) of 1.19. These features suggest a living polymeriza-
tion,27 as might be expected from a structure such as that of
13, suitable to undergo consecutive insertions (Figure 8). Despite
its modest activity, complex 13 constitutes a rare example of
iridium alkyne polymerization catalyst,28 as well as a limpid
illustration of the potential benefits behind the catalytic applica-
tion of biPSi pincer ligands. Unfortunately, so far, similar poly-
or oligomerization reactions have not been observed for other
1-alkynes.
Figure 7. X-ray structures of 6-syn, 8-syn, and 11-anti and detail of the
11-anti equatorial plane.
illustrative structure of the complex anti-[Ir(O2CMe)H(biPSi)]
(11-anti). This compound has been prepared from 1 and sodium
acetate in an analogous manner to 2 and 3, and its anti:syn
equilibrium contains 95% of the anti isomer. From the perspec-
tive of the acetate ligand, this compound is six-coordinate since,
as inferred from the two equal C-O distances, 1.253(6) and
1.255(6) Å, the coordination mode of acetate is chelate (κ2-O).
However, attending to the iridium coordination sphere, the
complex is best described as five-coordinate, since the Ir-O
distance trans to silicon, 2.405(4) Å, is too long for a bond.25
The example illustrates how, even under almost optimal
conditions, the anti fragments hardly show the Lewis acidity
that can be expected from its coordinative unsaturation.
Insertions. Not surprisingly, given the coordination behavior
described in the previous section, an insertion reagent such as
diphenylacetylene has been found to selectively react with the
syn isomer of complex 1, leaving unreacted 1-anti. The new
reaction product is the five-coordinate alkenyl complex syn-
[IrCl{Z-C(Ph)dCHPh}(biPSi)] (12) (eq 5), as indicated by the
clear 1H NOESY NMR cross peaks relating the vinylic hydrogen
signal with that corresponding to the methyl substituent at
silicon. Monitoring the reaction by NMR at room temperature
for weeks (or days at high temperature in toluene-d8) has
revealed the eventual quantitative formation of 12 as a conse-
quence of the slow equilibration between the anti and syn
isomers of precursor 1. Actually, this evolution has been used
to estimate the rate constant for the antiTsyn isomerization
process, assumed to be rate-limiting in the reaction sequence
leading to 12. The kinetic information will be discussed in the
next section.
The insertion reactions in complex 1 and their products can
be further illustrated with another type of stable compounds:
alkyl derivatives resulting from reactions against diazoalkanes.
The reaction between 1 and the commercially available ethyl
diazoacetate (EDA) has been found to afford the alkyl derivative
syn-[IrCl{CH2C(O)OEt}(biPSi)] (14). According to the gener-
ally accepted mechanism for reactions between diazoalkanes
and transition metal complexes, the transformation in eq 6 likely
(26) Ghosh, R.; Zhang, X.; Achord, P.; Emge, T. J.; Krogh-Jespersen, K.;
Goldman, A. S. J. Am. Chem. Soc. 2007, 129, 853–866.
(27) See, for example: Misumi, Y.; Masuda, T. Macromolecules 1998, 31,
7572–7573.
(28) (a) Lui, J.; Lam, J. W.; Tang, B. Z. Chem. ReV. 2009, 109, 5799–
5867. (b) Joo, K.-S.; Kim, S. Y.; Chin, C. S. Bull. Kor. Chem. Soc.
1997, 18, 1296–1301. (c) Marigo, M.; Marsich, N.; Farnetti, E. J. Mol.
Catal. A: Chem. 2003, 206, 319–329. (d) Zhang, Y.; Wang, D.; Wurst,
K.; Buchmeiser, M. R. J. Organomet. Chem. 2005, 690, 5728–5735.
(25) Examples of this long distance are restricted to water and hydroxide
ligands formally acting as asymmetric bridges between metals. See:
Allen, F. H. Acta Crystallogr. 2002, B58, 380–388.
9
9116 J. AM. CHEM. SOC. VOL. 132, NO. 26, 2010

biPSi Complexes of Iridium
A R T I C L E S
Figure 9. Evolution with temperature of the T1 relaxation times corre-
1
sponding to the H NMR hydride signals of isomers 1 in toluene-d8 (300
MHz): (left) in the absence and (right) in the presence of dissolved H2.
10-6 s-1. This value has been found independent of the
diphenylacetylene concentration and therefore corresponds to
a true first-order constant. This small rate constant seems
compatible with the very slow isomerizations observed for the
five-coordinate complexes 1-3 or their acetonitrile adducts
8-10. In contrast, the carbonyl compounds 4, 6, and 7 isomerize
much faster, needing just several hours to reach equilibrium
distributions.
Figure 8. X-ray structures of complexes 13 and 14 and details of their
equatorial planes.
An extreme case of fast antiTsyn exchange has been
observed in the presence of dihydrogen, a reagent that does not
form any detectable new product with 1-3 but renders
isomerization rates of the same order than as NMR time scale.
The fast antiTsyn exchange can be visualized with the help of
Figure 9, which represents the evolution with temperature of
proceeds via a hydride-carbene intermediate,29 which would
undergo R-insertion to form 14. Again, the insertion reaction
seems to be exclusive of syn isomers since, together with traces
of diethyl fumarate and maleate, compound 14 is the only
detectable reaction product. The reaction has been observed to
complete in about 20 min at room temperature.
1
the T1 relaxation times corresponding to the H NMR hydride
signals of isomers 1. Even though the exchange does not become
fast enough to average or broaden the hydride NMR signals, it
provokes an average of their relaxation times above 260 K.
Interestingly, the T1 values corresponding to the 1H NMR signal
of dissolved H2 do not average with those of the hydrides,
suggesting the lack of Ir-H/H2 atom exchange during the
antiTsyn isomerization. The same conclusion can be derived
from the use of D2 instead of H2 since, despite the fast exchange,
incorporation of deuterium into the hydride position of isomers
1 is significant only after at least 1 h of reaction. The progress
of such slow H/D substitution can be conveniently followed by
31P{1H} NMR, since it causes observable isotopic shifts. The
values of ∆δ ) +0.12 and +0.16 ppm found for the anti and
syn isomers of 1, respectively, are of the same sign and
magnitude as those previously observed in related Ir(III)
complexes.22
The structure of 14 determined by X-ray diffraction is shown
in Figure 8. Although the orientation of the ester substituent
may suggest that an oxygen atom occupies the sixth coordination
position, the observed Ir-O distance is again too long for a
bond. Once again, the structure reveals the capability of the
biPSi ligand to control the position of the coordination vacancy,
in this example in the presence of an alkyl ligand.
Isomerization antiTsyn. All reactions in the previous section
illustrate kinetic resolutions of equilibrium mixtures of isomers
1 due to the preferred (or exclusive) insertion into 1-syn. All
insertions seem to be relatively fast, but the time required for
quantitative disappearance of the nonreactive 1-anti strongly
depends on the nature of the insertion reagent. Thus, it takes
just several minutes for reaction with acetylene or EDA at room
temperature but weeks in the presence of diphenylacetylene.
This seems to indicate that the antiTsyn equilibration rate
depends to a large extent on the presence of additives and their
nature. As reference value, the pseudo-first-order rate constant
for the disappearance of 1-anti in the presence of dipheny-
lacetylene has been estimated by 31P NMR at 363 K as 4.32 ×
A quantitative determination of antiTsyn isomerization rate
constants has been carried out on solutions of the bromide
complex 2 in toluene-d8 at different dihydrogen pressures and
temperatures, using spin saturation transfer methods on the 1H
NMR hydride signals. All kinetic information is summarized
in Figure 10, which concludes that the isomerization rate
depends directly on the H2 pressure and estimates the activation
parameters as ∆Hq ) 7.4((1) kcal mol-1 and ∆Sq ) -22((2)
cal K-1 mol-1 (see Experimental Section and Supporting
Information for details). The large negative enthalpy increment
agrees with a second-order reaction and suggests that, despite
the absence of observable intermediates, this antiTsyn isomer-
ization begins with the coordination of dihydrogen to the five-
coordinate complex. Accordingly, the fast isomerization could
be related to the formation of polyhydride intermediates, the
(29) (a) Cohen, R.; Rybtchinski, B.; Gandelman, M.; Rozenberg, H.; Mart´ın,
J. M. L.; Milstein, D. J. Am. Chem. Soc. 2003, 125, 6532–6546. (b)
Doyle, M. P.; McKervey, M. A.; Ye, T. Modern Catalytic Methods
for Organic Synthesis with Diazo Compounds; Wiley: New York,
1998.
9
J. AM. CHEM. SOC. VOL. 132, NO. 26, 2010 9117

A R T I C L E S
Sola et al.
simple model of complex 2 and are not intended to accurately
reproduce experimental energies, the low overall isomerization
barrier seems compatible with the experimental activation
enthalpy (10.1 vs 7.4 ( 1 kcal mol-1).
The calculated mechanism does not include a TS for
dihydrogen coordination to 2-syn because it could not be located.
As discussed in previous sections, on the basis of experimental
data, given the structure of the complex, the activation energy
for ligand coordination trans to silicon is expected to be very
low, especially for a small incoming ligand such as dihydrogen
in a model complex without phenyl rings. A similar circum-
stance has been experimentally observed for other d6 complexes
of T structure, whose reaction kinetics with dihydrogen depend
only on the rates of reactant diffusion in solution.34 By contrast,
the model of 2-anti needs to invest 5.8 kcal mol-1 to reorganize
its equatorial plane for dihydrogen coordination trans to hydride.
Interestingly, the experiments described previously have shown
that a similar reorganization of 1-anti to coordinate acetonitrile
requires an activation enthalpy of 11.4((1.1) kcal mol-1, higher
than that of the entire isomerization process in the presence of
H2. This suggests that the mere coordination of dihydrogen to2-
anti should not be discarded as potential rate-determining step
of the whole antiTsyn isomerization process. In fact, this
possibility would perfectly fit with the negligible kinetic isotope
effects (KIEs) (very close to 1) that have been observed for
isomerizations using either the deuteride isotopomer of 2 or D2
instead of H2. With regard to the literature data, the elementary
steps of H2 coordination to unsaturated complexes typically
display KIE values (kH/kD) very close to 134a,35 but equilibrium
isotope effects (EIE ) KH/KD) significantly lower than this
value.36 Due to the latter, any rate-limiting step beyond the
reversible H2 coordination should lead to a significant inverse
KIE, reflecting such an EIE.
The mechanism of Figure 11 involves some Ir(V) intermedi-
ates. This oxidation state is emerging as relatively common in
stable complexes with ligands such as hydride, silyl, alkyl, and
boryl and is increasingly considered a likely alternative in the
context of mechanistic discussion.37 Assuming the key TS of
the mechanism to be an agostic Ir(III) species, the acceleration
of the antiTsyn isomerization under H2 can be attributed to
the ability of this reactant to provoke the oxidation state Ir(V),
from which the formal reductive elimination leading to the
agostic TS should be favored. Accordingly, acetylene, another
potential oxidative addition reagent, has also been observed to
catalyze this isomerization prior to insertion. A calculated
Figure 10. (Left) Dependence upon H2 partial pressure of the pseudo-
first-order rate constants (kobs) for exchange between 2-anti and 2-syn
isomers. (Right) Eyring representation of the isomerization first-order rate
constants (k(antifsyn) ) kobs/[H2]).
behavior of which is often fluxional.30 Nevertheless, this
proposal should be qualified to account for the absence of
exchange between hydrides and dihydrogen.
A possible mechanism for this antiTsyn isomerization,
initiated by coordination of dihydrogen to 2 and compatible with
all experimental observations, is detailed in Figure 11. Its
feasibility is supported by MP2 calculations on the model of
complex 2 without phenyl rings. Under such a mechanism, the
isomerization takes place when the hydride ligand passes from
one face of the complex to the other through the plane defined
by the mer-coordinated biPSi ligand. The TS associated with
this step resembles a σ-M-(η2-SiH)-coordinated silane (agostic)31
or a hypervalent Si · · ·HsM interaction.19 This type of TSs and
intermediates has previously been observed or proposed for the
isomerization of other hydride silyl complexes.32 A similar
mechanism has also previously been analyzed for compounds
analogous to 1-3 with PCP pincer ligands, although it was
found ineffective in that case.33
The mechanistic proposal of Figure 11 involves a syn adduct
coordinating dihydrogen trans to silicon and its anti* isomer
(H2 trans to hydride), those experimentally observed for other
weakly coordinating ligands such as acetonitrile. Also according
to the experimental evidence, the isomerization mechanism does
not necessarily imply an exchange of H atoms between hydride
and dihydrogen ligands. Nevertheless, such an exchange can
occur in the trihydride intermediate 2-syn(H)2 through the high-
energy TS shown in Figure 11 above the isomerization
coordinate. Although the calculations have been limited to a
(30) (a) Loza, M.; Faller, J. W.; Crabtree, R. H. Inorg. Chem. 1995, 34,
2937–294. See also: (b) Kubas, G. J. Metal Dihydrogen and σ-Bond
Complexes; Kluwer Academic/Plenum, New York, 2001. (c) Kubas,
G. J. ComprehensiVe Organometallic Chemistry, 3rd ed.; Mingos,
D. M. P., Crabtree, R. H., Eds.; Elsevier: Amsterdam, 2007; Vol. 1,
pp 671-698. (d) Morris, J. H.; Jessop, P. G. Coord. Chem. ReV. 1992,
121, 155–289. (e) Heinekey, D. M.; Oldham, W. J., Jr. Chem. ReV.
1993, 93, 913–926.
(34) (a) Bakhmutov, V. I.; Bertra´n, J.; Esteruelas, M. A.; Lledo´s, A.;
Maseras, F.; Modrego, J.; Oro, L. A.; Sola, E. Chem. Eur. J. 1996, 2,
815–825. See also: (b) Zhang, K.; Gonza´lez, A. A.; Hoff, C. D. J. Am.
Chem. Soc. 1989, 111, 3627–3632. (c) Hall, C.; Jones, W. D.; Mawby,
R. J.; Osman, R.; Perutz, R. N.; Whittlesey, M. K. J. Am. Chem. Soc.
1992, 114, 7425–7435.
(35) Janak, K. E. In ComprehensiVe Organometallic Chemistry, 3rd ed.;
Mingos, D. M. P., Crabtree, R. H., Eds.; Elsevier: Amsterdam, 2007;
Vol. 1, p 561.
(31) (a) Delpech, F.; Sabo-Etienne, S.; Daran, J.-C.; Chaudret, B.; Hussein,
K.; Marsden, C. J.; Barthelat, J.-C. J. Am. Chem. Soc. 1999, 121, 6668–
6682. (b) Perutz, R. N.; Sabo-Etienne, S. Angew. Chem., Int. Ed. 2007,
46, 2578–2592. Stable M(η2-Si-H) agostic species have been isolated
for Ru(II), Ni(II), and Pd(0): (c) Montiel-Palma, V.; Mun˜oz-Herna´ndez,
M. A.; Ayed, T.; Barthelat, J.-C.; Grellier, M.; Vendier, L.; Sabo-
Etienne, S. Chem. Commun. 2007, 3963–3965. (d) Chen, W.; Shimada,
S.; Tanaka, M.; Kobayashi, Y.; Saigo, K. J. Am. Chem. Soc. 2004,
126, 8072–8073. (e) Takaya, J.; Nobuharu, I. Organometallics 2009,
28, 6636–6638.
(36) (a) Gusev, D. G.; Bakhmutov, V. I.; Grushin, V. V.; Vol’pin, M. E.
Inorg. Chim. Acta 1990, 177, 115–120. (b) Hauger, B. E.; Gusev,
D. G.; Caulton, K. G. J. Am. Chem. Soc. 1994, 116, 208–214. (c)
Kubas, G. J. Chem. ReV. 2007, 107, 4152–4205.
(37) See for example: (a) Mann, B. E.; Masters, C.; Shaw, B. L. J. Inorg.
Nucl. Chem. 1971, 33, 2195–2204. (b) Fernandez, M. J.; Maitlis, P. M.
Organometallics 1983, 2, 164–165. (c) Gutie´rrez-Puebla, E.; Monge,
A.; Nicasio, M. C.; Pe´rez, P. J.; Poveda, M. L.; Carmona, E. Chem.
Eur. J. 1998, 4, 2225–2236. (d) Alaimo, P. J.; Bergman, R. G.
Organometallics 1999, 18, 2707–2717. (e) Klei, S. R.; Tilley, T. D.;
Bergman, R. G. J. Am. Chem. Soc. 2000, 122, 1816–1817. (f)
Kawamura, K.; Hartwig, J. F. J. Am. Chem. Soc. 2001, 123, 8422–
8423.
(32) (a) Ampt, K. A. M.; Duckett, S. B.; Perutz, R. N. Dalton Trans. 2004,
3331–3337. See also: (b) Caˆmpian, M. V.; Clot, E.; Eisenstein, O.;
Helmstedt, U.; Jasim, N.; Perutz, R. N.; Whitwood, A. C.; Williamson,
D. J. Am. Chem. Soc. 2008, 130, 4375–4385.
(33) Li, S.; Hall, M. B. Organometallics 1999, 18, 5682–5687.
9
9118 J. AM. CHEM. SOC. VOL. 132, NO. 26, 2010

biPSi Complexes of Iridium
A R T I C L E S
Figure 11. Mechanism for the antiTsyn isomerization under H2 based on MP2-calculated structures and energies (kcal mol-1, without corrections). Most
transition states have been omitted for clarity.
mol-1 in the six-coordinate carbonyl compound 6. These barriers
are in good qualitative agreement with the expected electronic
effect of each ligand on the energy of an Ir(I) TS and even
seem to correlate with structural parameters such as the Ir-Si
bond distances (Supporting Information). Even though the
quantitative aspects of this mechanism remain susceptible to
further elaboration, so far it is entirely consistent with the
experimental observations on the system. Moreover, the mech-
anism provides a rationale for the previously reported counter-
intuitive observation of five-coordinate Ru(II) hydride complexes
that are rigid in solution but form rapidly isomerizing octahedral
CO adducts.38
Conclusions
The studies described herein offer a detailed description of
the behavior of the biPSi ligand (κ-P,P,Si-Si(Me){(CH2)3PPh2}2)
in the coordination sphere of the five-coordinate iridium(III)
complexes [IrXH(biPSi)], together with representative examples
Figure 12. Calculated energies (kcal mol-1, without corrections) and
of their coordination adducts and insertion products.
optimized transition states for the antiTsyn isomerization of complex 2
This type of diphosphine can be considered a particular class
(center) and its carbonyl (right) and acetonitrile (left) adducts 6 and 9,
of pincer ligand due to its capability of forcing cis relative
respectively.
positions between the vacant sites of unsaturated complexes and
reactive ligands such as hydride, alkyl, and alkenyl. Neverthe-
less, an effective use of such vacant sites for coordination and
reaction coordinate showing the intermediates and TSs of the
antiTsyn isomerization in the presence of acetylene and its
subsequent reactions requires the connivance of steric factors.
subsequent insertion has been included as Supporting Informa-
Both the success of this strategy toward more reactive unsatur-
tion. An agostic TS analogous to that in Figure 11 has been
located just 4.3 kcal mol-1 above 2-anti.
The deduced mechanism of isomerization could be extrapo-
lated to the absence of oxidative addition reagents if the key
ated metal complexes and its limitations have a common origin
in the large trans influence of the sp3 silyl moiety, which
overcomes that of other ligands to dictate the ground-state
geometry around the metal, although at the expense of diminish-
agostic TSs could be generated through formal reductive
ing the binding energies of incoming ligands below contributions
eliminations in Ir(III) species. This possibility has been explored
that are otherwise secondary. As a consequence of the latter,
by calculation, in this case using the complete model of complex
the coordination behavior of these complexes, and hence their
2 (including phenyls by UFF methods in ONIOM approach).
reactivity thereafter, becomes especially sensitive to details of
The TS for a direct isomerization of the five-coordinate species
has been found 42.7 kcal mol-1 above 2-anti (Figure 12).
the ligand environment. As an example, changing the halide
Interestingly, the calculated isomerization barrier further in-
creases in the acetonitrile adduct 9 but decreases to 30.1 kcal
(38) Zhou, X.; Stobart, S. R. Organometallics 2001, 20, 1898–1900.
9
J. AM. CHEM. SOC. VOL. 132, NO. 26, 2010 9119

A R T I C L E S
Sola et al.
magnitudes obtained from Van’t Hoff or Eyring regressions were
estimated through conventional error propagation formulas,41
assuming 1 K error in the temperature and a 10% error in the
constant. All numerical data obtained in these studies are presented
in the Supporting Information.
The coordination of CDCl3 or CD2Cl2 to the syn isomers of 1-3
was studied using the 31P{1H} NMR chemical shifts because their
variation with temperature was greater than that of other chemical
shifts. The equilibrium constants (K) for solvent (sol) coordination
to the five-coordinate complex (5C) are defined in the sense of
formation of the six-coordinate adduct (6C) and can also be
expressed in terms of the molar fraction (x) of the 5C species:
ligand of [IrXH(biPSi)] from chloride to iodide switches the
stable coordination position of CO from cis to hydride to trans.
In the particular case of the [IrXH(biPSi)] complexes, such
a sensitivity to the ligand environment results in more reactive
syn isomers. Actually, all reagents studied in the course of this
work have been found able to discriminate to some extent
between anti and syn isomers, in some cases resolving the
isomeric mixtures by selective reaction with the syn compound.
First, this has minimized the consequences of dealing with
mixtures of isomers instead of pure compoundssanother
inevitable outcome of the sp3 silyl group, especially undesirable
if the objective is a single-site homogeneous catalyst. In addition,
the effectiveness of certain reagents in the kinetic resolution of
isomers has led to the recognition of an antiTsyn isomerization
mechanism via M-(η2-SiH) agostic TSs. In agreement with such
a mechanism, the isomerization is accelerated by potential
ligands that stabilize the oxidation state Ir(I), but it can also be
catalyzed by oxidative addition reagents providing access to
Ir(V) intermediates.
K ) [6C]/[5C][sol] ) (1 - x)/(x[sol])
For a fast equilibrium on the NMR time scale, the chemical shift
of the signal due to an average of the 5C and 6C species is also a
function of the molar fraction and therefore can also be expressed
in terms of the equilibrium constant:
δ ) δ(6C) + {δ(5C) - δ(6C)}x
It turns out from this study that PSiP pincer ligands control
the position of coordination vacancies, produce resolvable
mixtures of isomeric hydrides, and provoke coordination proper-
ties that are especially sensitive to the ligand environment. The
design of complexes, ligands, and reaction conditions that can
turn such exceptional features into advantages and/or new
catalytic applications is a challenge, for which we expect the
experimental observations and conclusions of this work to be a
guide.
δ ) δ(6C) + {δ(5C) - δ(6C)}{1/(1 + K[sol])}
Writing K as a function of ∆H° and ∆S° (ln K ) -∆H°/RT +
∆S°/R) and operating, the dependence of δ upon the inverse of
temperature takes the form typical of a sigmoidal function:
δ ) δ(6C) + {δ(5C) - δ(6C)}/[1 + exp{(1/T) -
(R ln[sol] + ∆S°)/∆H°}/(-R/∆H°)}]
y ) A2 + (A1 - A2)/[1 + exp{(x - x0)/dx}]
Experimental Section
The experimental values of δ vs 1/T can then be adjusted to
sigmoidal functions to give the equilibrium enthalpy and entropy
increments. Since the samples were rather diluted (ca. 10 mg in
0.5 mL), the solvent concentration was directly estimated from its
density: 15.24 M (1.32 g cm-3) for CD2Cl2 and 12.29 M (1.48 g
cm-3) for CDCl3. In this case, the errors in ∆H° and ∆S° were
directly propagated from the assumed errors in temperature and δ:
1 K and 0.1 ppm, respectively.
General Methods. All manipulations were carried out under
argon by standard Schlenk techniques or in an argon-filled MBraun
drybox. Solvents were obtained from a solvent purification system
(Innovative Technologies). Deuterated solvents were dried with
appropriate drying agents and degassed with argon prior to use. C,
H, and N analyses were carried out in a Perkin-Elmer 2400 CHNS/O
analyzer. MALDI-TOF mass spectra were obtained on a Bruker
Microflex mass spectrometer using 1,8,9-trihydroxyanthracene as
matrix. Infrared spectra were recorded in KBr using a Perkin-Elmer
Spectrum One FT-IR spectrometer. NMR spectra were recorded
on Bruker Avance 400 and 300 MHz spectrometers. 1H (400.13 or
300.13 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 TMS. 31P (162.0 or 121.5 MHz)
and 29Si (59.6 or 79.5 MHz) chemical shifts were measured relative
to H3PO4 (85%) and TMS, respectively. In general, NMR spectral
The stability of the acetonitrile adducts was studied in toluene-
d8 samples containing known concentrations of the metal complex
(around 0.02 M) and acetonitrile (in the range 2-4 µL). The
samples used to study anti* isomers were prepared immediately
before use, whereas those dedicated to the syn isomers were allowed
to evolve for several weeks until complete transformation into these
isomers. The equilibrium constants for the formation of 9-11-syn
were obtained from the 31P{1H} NMR spectra of fast equilibria
(above 313 K), using the same equations detailed in the previous
case with [sol] ) [NCMe]. Since, at room temperature, the
equilibria are well displaced toward the 6C species, their chemical
shifts, δ(6C), were taken from these spectra, while δ(5C) were
obtained from samples of the pure 5C precursors. Equilibrium ratios
[6C]/[5C] at various temperatures were then obtained from the
averaged chemical shifts, and the actual acetonitrile concentrations
were calculated from the initial ones and the position of each
equilibrium. The equilibrium constants involving anti* isomers were
obtained from slow equilibria at low temperature. The [6C]/[5C]
ratios were obtained by integration of their corresponding 1H NMR
hydride signals. The acetonitrile concentrations could not be
estimated by integration due to the overlap of various signals. In
turn, [NCMe] values were calculated from the known initial
concentrations, taking into account that, in part, it is coordinated
to both anti* and syn isomers.
assignments were achieved through H COSY, H NOESY, 13C
1
1
1
APT, and H/13C HSQC experiments. The molecular weights (Mn
and Mw) and polydispersity indices (Mw/Mn) of the polyacetylene
samples were measured by GPC with a Waters 2695 autosampler
chromatograph equipped with Phenogel linear mixed columns and
a Waters 2487 dual-lambda UV detector, which was previously
calibrated with polystyrene standards (molecular weight limit up
to 1 × 107). The samples were eluted at 35 °C with THF at a flow
rate of 1 mL min-1. The starting complex [Ir(µ-Cl)(COD)]239 and
the diphosphine bis(3-diphenylphosphinopropyl)methylsilane (bi-
PSiH)40 were prepared by known procedures. All other reagents
were commercially available and were used as received. Experi-
mental details and spectroscopic characterizations of the new
complexes are presented in the Supporting Information.
NMR Equilibrium and Kinetic Studies. The NMR sample
temperature was calibrated using the temperature dependence of
the chemical shifts of MeOH and ethylene glycol. Errors in the
(41) (a) Taylor, J. R. An Introduction to Error Analysis, 2nd ed.; University
Science Books: Mill Valley, CA, 1982. (b) Morse, P. M.; Spencer,
M. O.; Wilson, S. R.; Girolami, G. S. Organometallics 1994, 13, 1646–
1655.
(39) Uso´n, R.; Oro, L. A.; Cabeza, J. A. Inorg. Synth. 1985, 23, 126–130.
(40) Stobart, S. R.; Joslin, F. L. Inorg. Chem. 1993, 32, 2221–2223.
9
9120 J. AM. CHEM. SOC. VOL. 132, NO. 26, 2010

biPSi Complexes of Iridium
A R T I C L E S
First-order rate constants for acetonitrile dissociation from 8-anti*
were determined by spin saturation transfer, following the
Forse´n-Hoffman method.42 The presaturation pulse was applied
on the 1H NMR signal corresponding to the hydride ligand of 1-anti,
and its effect was measured in the intensity of the 8-anti* hydride
signal.
Determination of the 1-antiT1-syn isomerization rate constant
in the presence of diphenylacetylene was carried out by 31P{1H}
NMR monitoring of 0.04 M solutions of 1 at 363 K. Two samples
of alkyne concentrations 0.4 and 1 M, respectively, gave coincident
results.
observed in close proximity of the Ir centers and make no chemical
sense. Crystallographic data are summarized in the Supporting
Information.
Computational Details. Calculations were executed using the
Gaussian suite of programs47 and ONIOM-type approaches. The
low-level calculations were MP2. The inner electrons of Ir, Br, Si,
and P atoms were described using the effective core potential
LANL2DZ.48 The outer electrons were described with a double-ꢀ
basis set,49 plus p polarization functions for the hydrides, d for C,
P, Si, and Br atoms,50 and diffuse d51 and f52 polarization shells
for Ir. The second layer consisted of a QCISD(T) approach for Ir
and the atoms coordinated to it: Br, Si, P, hydrides, and acetylene
C atoms. In calculations including phenyl groups, the third layer
used a molecular mechanics force field (UFF) to estimate their steric
effects.
The characterization of each structure as either minimum or TS used
harmonic vibrational frequency calculations on an early level of
calculation, i.e., at the MP2 level not including polarization functions
for most atoms [ONIOM MP2/LANL2DZ+(f,d,p):HF/LANL1DZ] or
at the DFT level53 by using the B3LYP functional54 [B3LYP/
LANL2DZ+(f,d,p)]. After this characterization, the structures and
energies were reoptimized at the levels MP2/LANL2DZ+(f,d,p) and
ONIOM QCISD(T)/LANL2DZ+(f,d,p):MP2/LANL2DZ+(f,d,p), re-
spectively. For models with phenyl rings, the final levels of calculation
were ONIOM MP2/LANL2DZ+(f,d,p):UFF for the structures and
ONIOM QCISD(T)/LANL2DZ+(f,d,p):MP2/LANL2DZ+(f,d,p):UFF
for the energy.
Determinations of the antiTsyn isomerization rate constants in
presence of H2 were carried out in PTFE-valved sealable 5 mm
NMR tubes (Wilmad). Each tube, previously filled with an aliquot
of a common solution of 2 in toluene-d8 ([2] ca. 0.02 M), was
simultaneously connected to a Schlenk manifold, to allow for
reaction manipulation and degassing, and to a high-pressure H2
source. This latter connection was made through an electronic
pressure meter/controller (El-Press, Bronkhorst HI-TEC) that al-
lowed us to select the operating pressure in the range 0-4 bar
(absolute pressure) from a computer, thereby also allowing us to
measure the resulting pressure once the system reached equilibrium.
The freeze-thawed samples of 2 were loaded with H2 at the desired
pressure, allowed to equilibrate at room temperature for 30 min,
and sealed. The partial pressure of H2 was calculated from the
pressure measured at this point and the vapor pressure of toluene.43
In experiments above room temperature, the pressure that developed
inside the NMR tube was calculated assuming ideal gas behavior.
The values of H2 solubility at different partial pressures and
temperatures were obtained from literature data44 and the application
of Henry’s law. T1 relaxation times were measured by the
inversion-recovery method. The pseudo-first-order rate constants
(kobs) were obtained by spin saturation transfer, applying the
saturating pulse on the hydride signal of the minor isomer 2-syn
and measuring the intensity change of the hydride signal of 2-anti.
X-ray Crystallography. X-ray data were collected at 100.0(2)
K on a Bruker SMART APEX CCD with graphite-monochromated
Mo KR radiation (λ ) 0.71073 Å). Data were collected over the
complete sphere by a combination of four sets. Data were corrected
for absorption by using a multiscan method applied with the
SADABS program.45 The structures were solved by the Patterson
method and refined by full-matrix least-squares on F2 using Bruker
SHELXTL program package,46 including isotropic and subsequently
anisotropic displacement parameters. Weighted R factors (Rw) and
goodness of fit (S) are based on F2, and conventional R factors are
based on F. Hydride ligands were located, but not all of them refined
appropriately. In such cases the Ir-H distance was fixed to 1.59(1)
Å (average value found in the Cambridge Structural Database). The
rest of the hydrogen atoms were calculated using a restricted riding
model on their respective carbon atoms with the thermal parameter
related to the bonded atom. All the highest electronic residuals were
Acknowledgment. We thank Stephen R. Stobart for his valuable
contributions to this work. This research was supported by the
Spanish MICINN (grants CTQ2006-01629, CTQ2009-08023, and
Consolider Ingenio 2010 INTECAT CSD2006-0003).
Supporting Information Available: Synthesis and character-
ization details, tables of bond distances and angles, crystal data,
figures and tables on the thermodynamic and kinetic determina-
tions, complete ref 47, a calculated reaction coordinate for the
antiTsyn isomerization/insertion sequence in presence of
acetylene; information (Cartesian coordinates, absolute energies)
on the calculated structures; and X-ray crystallographic file for
the complexes 2-anti, 2-syn, 3-syn, 6-syn, 8-syn, 11-anti, 13,
and 14 in CIF format. This material is available free of charge
via the Internet at http://pubs.acs.org.
JA102479H
(47) Frisch, M. J.; et al.; Gaussian 03, Revision C.02; Gaussian, Inc.:
Wallingford, CT, 2004.
(48) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299–310.
(49) (a) Huzinaga, S. J. Chem. Phys. 1965, 42, 1293–1302. (b) Dunning,
T. H. J. Chem. Phys. 1970, 53, 2823–2833.
(50) Ho¨llwarth, A.; Bo¨hme, S.; Dapprich, S.; Ehlers, A. W.; Gobbi, A.;
Jonas, V.; Ko¨hler, K. F.; Stegmann, R.; Veldkamp, A.; Frenking, G.
Chem. Phys. Lett. 1993, 208, 237–240.
(51) Couty, M.; Hall, M. B. J. Comput. Chem. 1996, 17, 1359–1370.
(52) Ehlers, A. W.; Bo¨hme, S.; Dapprich, S.; Gobbi, A.; Ho¨llwarth, A.;
Jonas, V.; Ko¨hler, K. F.; Stegmann, R.; Veldkamp, A.; Frenking, G.
Chem. Phys. Lett. 1993, 208, 111–114.
(53) (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.
(42) (a) Faller, J. W. In Determination of Organic Structures by Physical
Methods; Nachod, F. C., Zuckerman, J. J., Eds.; Academic Press: New
York and London, 1973. (b) Sanders, J. K. M.; Hunter, B. K. Modern
NMR Spectroscopy, A Guide for Chemists, 2nd ed.; Oxford University
Press: Oxford, 1993.
(43) TRC Thermodynamic Tables: Hydrocarbons; National Institute of
Standards and Technology: Boulder, CO; Vol. VIII, p Kb3290, 1995.
(44) Cook, M. W.; Hanson, D. N.; Adler, B. J. J. Chem. Phys. 1957, 26,
748–751.
(45) Blessing, R. H. Acta Crystallogr. 1995, A51, 33–38. SADABS: Area-
detector absorption correction; Bruker-AXS: Madison, WI, 1996.
(46) SHELXTL, v. 6.10; Bruker-AXS: Madison, WI, 2000. Sheldrick, G. M.
SHELXS-86 and SHELXL-97; University of Go¨ttingen: Go¨ttingen,
Germany, 1997.
(54) (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.
9
J. AM. CHEM. SOC. VOL. 132, NO. 26, 2010 9121