Computational and experimental test of steric influence on agostic interactions: A homologous series for Ir(III)

Article abstract of DOI:10.1021/ja981727e

Chloride abstraction (using NaBAr'₄, Ar' = 3,5-(CF₃)₂C₆H₃) from Ir(H)₂Cl(P(t)Bu₂Ph)₂ gives cis, trans-Ir(H)₂(P(t)Bu₂Ph)₂⁺, which has two agostic interactions with methyl C-H groups on different ¹Bu groups. The molecule exists as diastereomers, due to stereochemistry at P. Chloride can be similarly abstracted from orthometalated IrH(η²- C₆H₄P(t)Bu₂)Cl(P(t)Bu₂Ph) to give square-pyramidal IrH(η²- C₆H₄P(t)Bu₂)(P(t)Bu₂Ph)⁺, which has only one agostic interaction, involving a (t)BuC - H bond; steric constraints on each phosphine leave no more C-H bonds available to donate to the remaining empty Ir(III) orbital. The smaller ligand PCy₂Ph yields only the tris-phosphine complex Ir(H)₂(PCy₂Ph)₃⁺, and this is shown to have a square-pyramidal structure with one agostic cyclohexyl group and large Pax-Ir-P angles (104-106°). The analogous Ir(H)₂(P(i)Pr₂Ph)₃⁺ has similar inter-phosphorus angles, but no agostic interaction. Geometrical optimization of IrH2L3⁺ (PCy₂Ph, P(i)Pr₂Ph) with the hybrid quantum mechanics/molecular mechanics (QM/MM) method (IMOMM) at the IMOMM (B3LYP:MM3) and IMOMM (MP2:MM3) levels permits a more detailed understanding of the influence of steric factors on the occurrence of an agostic bond. The MP2/MM3 method gives the results in closer agreement with experiment. Steric factors place the agostic bond in the vicinity of the metal center but at a distance that is too long to be considered as bonding. The electron-donating ability of the C-H bond and the electron accepting capacity of the metal center, which are introduced only at the QM level, bring the two partners in a bonding situation.

Full text of DOI:10.1021/ja981727e

J. Am. Chem. Soc. 1999, 121, 97-106
97
Computational and Experimental Test of Steric Influence on Agostic
Interactions: A Homologous Series for Ir(III)
Alan C. Cooper,† Eric Clot,‡ John C. Huffman,† William E. Streib,† Feliu Maseras,*,‡,§
Odile Eisenstein,*,‡ and Kenneth G. Caulton*,†
Contribution from the Department of Chemistry, Indiana UniVersity, Bloomington, Indiana 47405-4001,
and LSDSMS (UMR 5636) Case Courrier 14, UniVersite´ de Montpellier 2,
34095 Montpellier Cedex 5, France
ReceiVed May 18, 1998
Abstract: Chloride abstraction (using NaBAr′4, Ar′ ) 3,5-(CF3)2C6H3) from Ir(H)2Cl(PtBu2Ph)2 gives cis,trans-
Ir(H)2(PtBu2Ph)2+, which has two agostic interactions with methyl C-H groups on different tBu groups. The
molecule exists as diastereomers, due to stereochemistry at P. Chloride can be similarly abstracted from ortho-
metalated IrH(η2-C6H4PtBu2)Cl(PtBu2Ph) to give square-pyramidal IrH(η2-C6H4PtBu2)(PtBu2Ph)+, which has
t
only one agostic interaction, involving a BuC-H bond; steric constraints on each phosphine leave no more
C-H bonds available to donate to the remaining empty Ir(III) orbital. The smaller ligand PCy2Ph yields only
the tris-phosphine complex Ir(H)2(PCy2Ph)3+, and this is shown to have a square-pyramidal structure with one
agostic cyclohexyl group and large Pax-Ir-P angles (104-106°). The analogous Ir(H)2(PiPr2Ph)3+ has similar
inter-phosphorus angles, but no agostic interaction. Geometrical optimization of IrH2L3 (PCy2Ph, PiPr2Ph)
+
with the hybrid quantum mechanics/molecular mechanics (QM/MM) method (IMOMM) at the IMOMM
(B3LYP:MM3) and IMOMM (MP2:MM3) levels permits a more detailed understanding of the influence of
steric factors on the occurrence of an agostic bond. The MP2/MM3 method gives the results in closer agreement
with experiment. Steric factors place the agostic bond in the vicinity of the metal center but at a distance that
is too long to be considered as bonding. The electron-donating ability of the C-H bond and the electron
accepting capacity of the metal center, which are introduced only at the QM level, bring the two partners in
a bonding situation.
Introduction
However, notable exceptions exist where intramolecular C-H
activation (i.e., oxidative addition) of alkyl groups on a
phosphine occurs to give metalated-phosphine complexes.4
Indeed, the metalation of phosphine C-H bonds has been
demonstrated to facilitate reductive elimination from Ir(III)
complexes.5 A more subtle influence of bulky phosphine ligands
on coordinatively unsaturated metal complexes is found in the
interactions of phosphine C-H bonds with the metal center that
do not result in a formal oxidative addition of the C-H bond.
These interactions, known as agostic bonding,6a involve 3-center,
2-electron bonding similar to that observed in main group Lewis
acids such as diborane. Agostic bonding has been observed in
transition metal complexes for over 30 years but has received
more attention in recent years as the number of documented
examples have increased greatly.6b The bonding of a C-H
fragment to a empty metal coordination site is analogous to the
binding of dihydrogen. Indeed, a number of dihydrogen
The quest for the synthesis of new coordinatively unsaturated
transition metal complexes has been driven by the role that these
complexes play in many catalytic processes and the useful
reactivity that can be exploited for the functionalization of
simple organic molecules.1 It has been proposed that highly
unsaturated intermediates in homogeneous catalysis may be
stabilized by agostic interactions from ligands on the metal.2
Shaw et al. first observed that the coordination of bulky tertiary
phosphine ligands with the formula PtBu2R may stabilize
unsaturation in Ir(III) complexes by prohibiting dimerization
and solvent coordination.3 In most studies of unsaturated
complexes which include bulky phosphine ligands, the ligands
have exhibited a “benign” influence in terms of providing steric
protection of empty coordination sites without any direct
interaction with the metal beyond metal-phosphorus bonding.
† Indiana University.
(4) (a) Cooper, A. C.; Huffman, J. C.; Folting, K.; Caulton, K. G.
Organometallics 1997, 16, 505. (b) For an example of phosphite metalation,
see: Bedford, R. B.; Castillo´n, S.; Chaloner, P. A.; Claver, C.; Fernandez,
E.; Hitchcock, P. B.; Ruiz, A. Organometallics 1996, 15, 3990.
(5) (a) Albe´niz, A. C.; Schulte, G.; Crabtree, R. H. Organometallics 1992,
11, 242. (b) Aizenberg, M.; Milstein, D. Organometallics 1996, 15, 3317.
(c) Cooper, A. C.; Huffman, J. C.; Caulton, K. G. Organometallics 1997,
16, 1974.
(6) (a) Brookhart, M.; Green, M. L. H. J. Organomet. Chem. 1983, 250,
395. Brookhart, M.; Green, M. L. H.; Wong, L.-L. Prog. Inorg. Chem.
1988, 36, 1. (b) Crabtree, R. H. Angew. Chem., Int. Ed. Engl. 1993, 32,
789.
‡ Universite´ de Montpellier 2.
§ Present address: Unitat de Qu´ımica F´ısica, Edifici C.n., Universitat
Auto`noma de Barcelona, 08193 Bellatera, Catalonia, Spain.
* Corresponding
odile.eisenstein@lsd.univ-montp2.fr, caulton@indiana.edu.
(1) (a) Collman, J. P.; Hegedus, L. S.; Norton, J. R.; Finke, R. G. In
Principles and Applications of Organotransition Metal Chemistry; Univer-
sity Science: Mill Valley, CA; 1987. (b) Herrmann, W. A.; Cornils, B.
Angew. Chem., Int. Ed. Engl. 1997, 36, 1048.
(2) (a) Margl, P.; Lohrenz, J. C. W.; Ziegler, T.; Blo¨chl, P. E. J. Am.
Chem. Soc. 1996, 118, 4434. (b) Grubbs, R. H.; Coates, G. W. Acc. Chem.
Res. 1996, 29, 85.
authors.
E-mail:
feliu@klingon.uab.es,
(7) (a) Kubas, G. J.; Ryan, R. R.; Swanson, B. I.; Vergamini, P. J.;
Wasserman, H. J. J. Am. Chem. Soc. 1984, 106, 451. (b) Kubas, G. J. Acc.
Chem. Res. 1988, 21, 120.
(3) Empsall, H. D.; Hyde, E. M.; Shaw, B. L.; Uttley, M. F. J. Chem.
Soc., Dalton Trans. 1976, 2069.
10.1021/ja981727e CCC: $18.00 © 1999 American Chemical Society
Published on Web 12/16/1998

98 J. Am. Chem. Soc., Vol. 121, No. 1, 1999
Cooper et al.
complexes have been shown to form agostic C-H bonds in
method enables a deeper understanding of the influence of
electronic and steric factors on the occurrence of agostic
interactions.
the absence of the dihydrogen ligand.7
The importance of steric influences in promoting the meta-
lation of phosphine C-H bonds has been revealed in a number
of elegant studies by Shaw et al.8 Since the earliest examples
of phosphine metalation,9 it has been found that the identity of
both the metalated alkyl group and pendant alkyl groups on
the phosphine ligand play a crucial role in determining whether
metalation will occur. Phosphine metalation often results in the
formation of highly strained metallacycles due to the small
number of members in the ring. A number of examples have
been characterized where four-membered rings10 and even
highly strained three-membered rings11 have been formed via
phosphine C-H activation by unsaturated iridium. This steric
enhancement of closure of small rings is why (Ph2P)2CMe2
favors chelation to a single metal, while (Ph2P)2CH2 is more
often found bridging two metals.12 The effect of alkyl substit-
uents on the formation of strained organic ring systems is
described by Ingold13 and has come to be known as the Thorpe-
Ingold or “gem-dimethyl” effect.14 A similar effect has been
proposed for the pendant alkyl groups of a metalated phosphine
ligand. Due to the larger atomic radius of phosphorus vs carbon,
it was found that alkyl groups with a large steric impact (i.e.,
more bulky than methyl) were necessary to promote the
formation of phosphine metallacycles. Shaw has appropriately
suggested the terminology “gem-tert-butyl” effect upon the basis
of his extensive studies with PtBu2R phosphine ligands.8c
Experimental Section
General Procedures. All manipulations were carried out using
standard Schlenk and glovebox techniques under argon. Toluene,
pentane, THF, and benzene were dried and deoxygenated over sodium
or potassium benzophenone and distilled under argon. Fluorobenzene
was distilled from P2O5 under argon and stored over activated molecular
sieves. C6D6, d8-THF, and d8-toluene were dried over sodium metal
and vacuum distilled before use in a glovebox. CD2Cl2 and CDCl3 were
1
dried over CaH2 and vacuum distilled before use in a glovebox. H
(referenced to residual solvent impurity), 13C, 31P (referenced to external
85% H3PO4), and 19F (referenced to external CFCl3) NMR spectra were
collected on Varian Gemini-300 and Inova-400 spectrometers. IR
spectra were collected on a Nicolet 510T FT-IR spectrometer. H2 (Air
Products, zero grade), PCy2Ph (Aldrich), and PiPr2Ph (Organometallics,
Inc.) were used as purchased. Na[BAr′4] was prepared according to a
literature procedure19 and dried under dynamic vacuum (1 × 10-3 Torr)
1
at 150 °C until H NMR assay confirmed the complete removal of
water from the bulk sample. [Ir(COE)2Cl]2,20 Ir(H)2Cl(PtBu2Ph)2,21 and
IrH(η2-C6H4PtBu2)(Cl)(PtBu2Ph)21 were prepared using literature meth-
ods or modification of the literature method.
[Ir(H)2(PtBu2Ph)2][BAr′4]. Sodium tetrakis[3,5-bis(trifluoromethyl)-
phenyl]borate (1.0 g, 1.13 mmol) and Ir(H)2Cl(PtBu2Ph)2 (762 mg, 1.13
mmol) were dissolved in fluorobenzene (50 mL) with stirring. This
homogeneous orange solution was stirred for 2 h at room temperature,
precipitating a fine white solid during this time. The solution was filtered
and concentrated to 3 mL in vacuo. After layering with ca. 5 mL of
pentane, the solution was placed in a -20 °C freezer for 1 week. Yellow
crystals were separated from the mother liquor and washed with pentane
While the role of bulky alkyl groups in phosphine metalation
is well documented, their possible role in promoting agostic
interactions has not been explored extensively. The characteriza-
tion of agostic interactions can be very difficult due to the weak
nature of the interactions. However, solid-state methods (neutron
and X-ray crystallography) and solution methods (low-temper-
ature NMR spectroscopy and IR spectroscopy) can characterize
the presence of agostic bonds in coordinatively unsaturated metal
complexes. These experimental methods, combined with hybrid
quantum mechanics/molecular mechanics calculations (QM/
MM), have revealed that changes in the steric profile and
geometry of phosphine ligands can determine whether agostic
interactions will be formed and, in complexes with two empty
coordination sites, whether one or two agostic interactions can
be formed.15 The examples of unsaturated Ir(III) complexes,
containing agostic interactions, described herein provide an ideal
situation to combine experimental results with several “com-
putational experiments”, using a hybrid (QM/MM) methodology
(IMOMM).16,17,18 This method has proven to be successful in
the quantification of electronic and steric effects in a number
of transition metal systems.17 In the present study, the IMOMM
1
(3 × 10 mL). Yield: (1.5 g, 87%). H NMR (CD2Cl2, 25 °C): 7.83
(br s), 7.73 (m), 7.63 (m), 7.57 (m), 1.36 (vt, JPH ) 7.6 Hz), -36.51
1
(br s). H NMR (C7D8, 25 °C): 8.30 (s), 8.15 (s), 7.67 (s), 7.55 (s),
7.24 (m), 7.20 (br s), 7.07 (br s), 0.88 (vt, JPH ) 7.2 Hz), 0.77 (vt, JPH
) 7.6 Hz), 0.72 (vt, JPH ) 6.0 Hz), -36.93 (br apparent t), -37.06 (br
apparent t). 13C{1H} NMR (CD2Cl2, 25 °C): 162.51 (m), 136.38 (vt,
JPC ) 6.0 Hz), 135.47 (s), 132.48 (s), 129.62 (m), 129.34 (s), 129.31
(vt, JPC ) 5.0 Hz), 128.99 (vt, JPC ) 21.5 Hz), 126.63 (s), 123.92 (s),
121.22 (s), 118.16 (m), 39.52 (vt, JPC ) 12.4 Hz), 29.69 (vt, JPC ) 9.9
Hz). 31P{1H} NMR (CD2Cl2, 25 °C): 60.2 (s). 31P{1H} NMR (C7D8,
25 °C): 61.7 (s), 61.9 (s). 19F NMR (CD2Cl2, 25 °C): -62.3 (s). IR
(C6D6): 2625, 2593, 2552 cm-1
.
[IrH(η2-C6H4PtBu2)(PtBu2Ph)][BAr′4]. A solution of IrH(η2-C6H4P-
tBu2)(Cl)(PtBu2Ph) (0.4 g, 0.60 mmol) dissolved in 5 mL of CH2Cl2
was added to a suspension of sodium tetrakis[3,5-bis(trifluoromethyl)-
phenyl]borate (0.68 g, 0.60 mmol) in CH2Cl2 (10 mL). This red
suspension was stirred at room temperature for 30 min, changing color
to orange over this time, and filtered. The resulting orange solution
was concentrated to dryness in vacuo. The resulting solid was dissolved
in fluorobenzene (10 mL) and again concentrated to dryness. The
resulting orange solid was suspended in C6H6 (2 mL) and heated to 40
°C, causing the solid to form a dense orange oil at the bottom of the
flask. This oil (under C6H6) was allowed to cool to room temperature
and stand overnight, forming a mass of orange crystals. The crystals
were separated from the oil and washed with C6H6 (3 × 2 mL). Yield:
(8) (a) Cheney, A. J.; Mann, B. E.; Shaw, B. L.; Slade, R. M. J. Chem.
Soc., Chem. Commun. 1970, 1176. (b) Shaw, B. L. J. Am. Chem. Soc. 1975,
97, 3856. (c) Shaw, B. L. J. Organomet. Chem. 1980, 200, 307.
(9) Parshall, G. W. Acc. Chem. Res. 1970, 3, 139 and references therein.
(10) (a) Perego, G.; del Piero, G.; Cesari, M.; Clerici, M. G.; Perrotti,
E. J. Organomet. Chem. 1973, 54, C51. (b) Empsall, H. D.; Heys, P. N.;
McDonald, W. S.; Norton, M. C.; Shaw, B. L. J. Chem. Soc., Dalton Trans.
1978, 1119.
(11) (a) Fryzuk, M. D.; Joshi, K.; Chadha, R. K.; Rettig, S. J. J. Am.
Chem. Soc. 1991, 113, 8724. (b) Al-Jibori, S.; Crocker, C.; McDonald, W.
S.; Shaw, B. L. J. Chem. Soc., Dalton Trans. 1981, 1572.
(12) Barkley, J.; Ellis, M.; Higgins, S. J.; McCart, M. K. Organometallics
1998, 17, 1725.
(13) Ingold, C. K. J. Chem. Soc. 1921, 305, 951.
(14) Eliel, E. L.; Wilen, S. H.; Mander, L. N. Stereochemistry of Organic
Compounds; Wiley: New York, 1994; p 682.
(15) Ujaque, G.; Cooper, A. C.; Maseras, F.; Eisenstein, O.; Caulton,
K. G. J. Am. Chem. Soc. 1998, 120, 361.
(16) Maseras, F.; Morokuma, K. J. Comput. Chem. 1995, 16, 1170.
1
(0.65 g, 71%). H NMR (CDCl3, 25 °C): 7.69 (m), 7.61 (m), 7.51
(17) (a) Wakatsuki, Y.; Koga, N.; Werner, H.; Morokuma, K. J. Am.
Chem. Soc. 1997, 119, 360. (b) Ogasawara, M.; Maseras, F.; Gallego-Planas,
N.; Kawamura, K.; Ito, K.; Toyota, K.; Streib, W. E.; Komiya, S.; Eisenstein,
O.; Caulton, K. G. Organometallics 1997, 16, 1979. (c) Ujaque, G.; Maseras,
F.; Eisenstein, O. Theor. Chem. Acc. 1997, 96, 146. (d) Maseras, F.;
Eisenstein, O. New J. Chem. 1998, 22, 5.
(18) (a) Svensson, M.; Humbel, S.; Morokuma, K. J. Chem. Phys. 1996,
105, 3654. (b) Matsubara, T.; Sieber, S.; Morokuma, K. Int. J. Quantum
Chem. 1996, 60, 1101.
(19) Brookhart, M.; Grant, R. G.; Volpe, A. F., Jr. Organometallics 1992,
11, 3920.
(20) van der Ent, A.; Onderlinden, A. L. Inorg. Synth. 1990, 28, 90.
(21) Cooper, A. C.; Caulton, K. G. Inorg. Chim. Acta 1996, 251, 41.

Steric Influence on Agostic Interactions
J. Am. Chem. Soc., Vol. 121, No. 1, 1999 99
(m), 7.50 (s), 7.21-7.00 (overlapping m), 1.35 (d, JPH ) 14.4 Hz),
1.34 (d, JPH ) 14.8 Hz), 1.32 (d, JPH ) 16.0 Hz), 1.03 (d, JPH ) 14.4
Hz), -41.6 (dd, JPH ) 12.0 Hz, JP′H ) 9.6 Hz). 31P{1H} NMR (CDCl3,
25 °C): 39.7 (d, JP′P ) 278 Hz), 6.2 (d, JP′P ) 278 Hz). 19F NMR
(CDCl3, 25 °C): -62.5 (s).
Table 1. Crystallographic Data for [Ir(H)2(PtBu2Ph)2][BAr′4]
a ) 20.023(4) Å
b ) 28.492(6) Å
c ) 19.579(4) Å
R ) 109.47(1)
â ) 89.89(1)
γ ) 110.40(1)
V ) 9788.58 Å3
Z ) 2
fwa ) 4602.33
space group: P1h
T ) -167 °C
λ ) 0.71069 Åb
Fcalcd ) 1.562 g cm-1
µ ) 22.082 cm-1
R(Fo)c ) 0.0660
Rw(Fo)d ) 0.0595
[Ir(H)2(PCy2Ph)3][BAr′4]. To a slurry of [Ir(COE)2Cl]2 (400 mg,
0.50 mmol) in 10 mL of C6H6 was added a solution of PCy2Ph (825
mg, 3.0 mmol) in 5 mL of C6H6. Upon addition, the color changed
from yellow to orange and the solution became homogeneous. H2 was
bubbled through the solution for 20 min at room temperature. The
volatiles were removed from the resulting orange solution in vacuo to
yield an orange solid. This solid was dissolved in fluorobenzene (10
mL) and a solution of Na[BAr′4] (885 mg, 1.0 mmol) in 5 mL of
fluorobenzene added with stirring. This solution was stirred at room
temperature for 1 h and concentrated in vacuo to 5 mL before filtering.
The remaining fluorobenzene was removed in vacuo to yield a orange
solid. After transferring the solid to a 5-mL flask, toluene (1.5 mL)
was added to form a suspension. Upon heating this suspension to 45
°C the solid dissolved to form a dense red oil. The solution was allowed
to cool to room temperature and stand overnight, forming a mass of
crystals in the dense oil. The solution was decanted, and the crystals
were washed with toluene (3 × 2 mL) to yield a dark yellow crystalline
a Formula C180H180B3F72P6Ir3; C6H5F. b Graphite monochromator. c R
2
) ∑|Fo| - |Fc|/∑|Fo|. d Rw ) [∑w(|Fo| - |Fc|)2/∑w|Fo| ]1/2, where w
) 1/σ2(|Fo|).
Table 2. Crystallographic Data for
[IrH(η2-C6H4PtBu2)(PtBu2Ph)][BAr′4]
a ) 12.798(1) Å
b ) 23.148(2) Å
c ) 12.220(1) Å
R ) 92.91(1)
â ) 92.99(1)
γ ) 74.50(1)
V ) 3481.46 Å3
Z ) 2
fwa ) 1617.23
space group: P1h
T ) -170 °C
λ ) 0.71069 Åb
Fcalcd ) 1.543 g cm-1
µ ) 20.566 cm-1
R(Fo)c ) 0.068
Rw(Fo)d ) 0.055
1
solid (1.22 g, 64%). H NMR (d8-THF, 25 °C): 7.76 (m), 7.55 (s),
a Formula C69H67BF24IrP2. b Graphite monochromator. c R ) ∑|Fo|
1/σ2(|Fo|).
7.40-7.07 (m), 2.41 (br s), 1.69 (br s), 1.40-0.85 (m), -26.05 (br s).
1H NMR (d8-THF, -110 °C, hydride ligands only): -5.4 (d, JPH
)
2
- |Fc|/∑|Fo|. d Rw ) [∑w(|Fo| - |Fc|)2/∑w|Fo| ]1/2, where w )
107 Hz), -44.8 (br s). 13C{1H} NMR (d8-THF, 25 °C): 162.99 (m),
135.77 (s), 133.56 (s), 131.70 (s), 130.66 (m), 130.35 (m), 130.03 (m),
129.75 (s), 129.55 (m), 127.04 (s), 124.34 (s), 121.63 (s), 118.36 (m),
40.05 (br s), 30.58 (s), 30.17 (br s), 27.77 (br s), 27.53 (s), 27.50 (s),
26.96 (s). 31P{1H} NMR (d8-THF, 25 °C): 28.3 (br s). 31P{1H} NMR
(d8-THF, -30 °C): 28.9 (br s, 2P), 28.2 (br s, 1P).19F NMR (d8-THF,
25 °C): -63.4 (s).
in fixed calculated positions with thermal parameters fixed at one plus
the isotropic thermal parameter of the parent carbon atoms. The third
anion was badly disordered and was modeled as 98 atoms, many at
partial occupancy. The occupancies were refined, scaled to give a total
occupancy of 57 atoms that should have been present, and then fixed.
No attempt was made to include hydrogens on the disordered anion.
In the final cycles of refinement, isotropic thermal parameters were
varied for one atom in the second anion, C(180) whose anisotropic
thermal parameters did not refine properly, and for all atoms in the
disordered anion. The largest peak in the final difference map was 1.5
e/Å3 near the disordered anion, and the deepest hole was -1.6 e/Å3.
Two anticipated hydride ligands on each of the iridium atoms were
not observed and were not included in the refinement.
[Ir(H)2(PiPr2Ph)3][BAr′4]. To a slurry of [Ir(COE)2Cl]2 (450 mg,
0.56 mmol) in 15 mL of C6H6 was added a solution of PiPr2Ph (1.16
mL, 3.36 mmol) in 5 mL of C6H6. Upon addition, the color changed
from yellow to red and the solution became homogeneous. H2 was
bubbled through the solution for 20 min at room temperature. The
volatiles were removed from the resulting orange solution in vacuo to
yield an orange solid. This solid was dissolved in fluorobenzene (10
mL) and a solution of Na[BAr′4] (990 mg, 1.12 mmol) in 5 mL of
fluorobenzene added with stirring. This solution was stirred at room
temperature for 2 h and concentrated in vacuo to 5 mL before filtering.
The remaining fluorobenzene was removed in vacuo to yield a dark
orange solid. After transferring the solid to a 5-mL flask, toluene (2
mL) was added to form a suspension. Upon heating this suspension to
60 °C the solid dissolved to form a dense red oil. The solution was
allowed to cool to room temperature and stand overnight, forming a
mass of crystals in the dense oil. The solution was decanted and the
crystals washed with toluene (3 × 2 mL) to yield an orange crystalline
Crystallographic Details for [IrH(η2-C6H4PtBu2)(PtBu2Ph)]-
[BAr′4]. An orange crystal was cleaved to form a nearly equidimen-
sional trigonal prism, affixed to the end of a glass fiber using silicon
grease, and cooled to -170 °C for characterization and data collection
(Table 2). A systematic search of a limited hemisphere of reciprocal
space located 86 reflections which were used to determine that the
crystal possessed no symmetry or systematic absences, indicating a
triclinic space group. Subsequent solution and refinement confirmed
the centrosymmetric choice, P1h. Data were corrected for Lorentz and
polarization effects, equivalent reflections, and averaged after an
analytical absorption correction (transmission factors 0.41-0.60). The
structure was solved using direct methods (MULTAN78) and Fourier
techniques. Hydrogen atoms were generally visible in a difference
Fourier phased on the non-hydrogen atoms but were placed in idealized
fixed positions for the final cycles of refinement. The data were not of
sufficient quality to locate the hydride ligand. A final difference Fourier
was featureless. There was one peak of 2.45 e/Å3 at the metal site, and
several peaks of up to 2.0 e/Å3 in the vicinity of the fluorine atoms of
the anion.
Crystallographic Details for [Ir(H)2(PCy2Ph)3][BAr′4]. A crystal
of suitable size was mounted in a nitrogen atmosphere glovebag using
silicone grease. It was then transferred to a goniostat where it was cooled
to -168 °C for characterization and data collection (Table 3). A
preliminary search for peaks and then analysis using programs DIRAX
and TRACER revealed a triclinic cell. Following intensity data
collection and an analytical correction for absorption (transmission
factors 0.65-0.86), the structure was solved using a combination of
direct methods (MULTAN78) and Fourier techniques. The positions
of the Ir and 22 other non-hydrogen atoms were obtained from an initial
E-map. The positions of the other non-hydrogen atoms were obtained
from iterations of a least-squares refinement followed by a difference
1
solid (1.42 g, 76%). H NMR (d8-THF, 25 °C): 7.80 (m), 7.58 (s),
7.48-7.27 (m), 2.69 (br s), 1.11 (m), 0.92 (m), -25.38 (br s). 1H NMR
(d8-THF, -110 °C, hydride ligands only): -5.8 (d, JPH ) 104 Hz),
-43.8 (br s). 13C{1H} NMR (d8-THF, 25 °C): 163.01 (m), 135.77 (s),
133.29 (s), 131.68 (m), 130.37 (m), 130.06 (m), 129.75 (s), 129.40
(m), 127.04 (s), 124.33 (s), 121.63 (s), 118.36 (m), 28.92 (br s), 20.09
(s), 19.18 (s). 31P{1H} NMR (d8-THF, 25 °C): 35.2 (br s). 31P{1H}
NMR (d8-THF, -30 °C): 35.2 (br s, 2P), 34.2 (br s, 1P).19F NMR
(d8-THF, 25 °C): -63.9 (s).
Crystallographic Details for [Ir(H)2(PtBu2Ph)2][BAr′4]. A crystal
of suitable size was obtained by cleaving a large piece of the sample
in a nitrogen atmosphere glovebag. The crystal was mounted using
silicone grease, and it was then transferred to a goniostat where it was
cooled to -167 °C for characterization and data collection (Table 1).
A preliminary search for peaks followed by analysis using programs
DIRAX and TRACER revealed a triclinic cell. Following intensity data
collection a correction for drift in the data was made, based on four
standards (0 0 6, -7 0 0, 0-5 0, -2-10 2) measured every 400 data.
An analytical correction was also made for absorption. The asymmetric
unit contains three cations, three anions, and one molecule of fluo-
robenzene. The cations, two of the anions, and the solvent molecule
were reasonably well ordered, and hydrogens were included for them

100 J. Am. Chem. Soc., Vol. 121, No. 1, 1999
Cooper et al.
Table 3. Crystallographic Data for [Ir(H)2(PCy2Ph)3][BAr′4]
polarization shells on carbon and hydrogen atoms27 involved in the
agostic interaction was discarded because preliminary calculations
showed that it brought only very minor differences.
a ) 17.627(3) Å
b ) 18.678(3) Å
c ) 13.772(2) Å
R ) 91.13(1)
â ) 105.65(1)
γ ) 99.59(1)
V ) 4295.09 Å3
Z ) 2
fw ) 1919.68
space group: P1h
T ) -168 °C
Hybrid IMOMM (integrated molecular orbital/molecular mechanics)
calculations were performed with a program built from modified
versions of two standard programs: Gaussian 92/DFT28 for the quantum
mechanics (QM) part and MM3(92)29 for the molecular mechanics
(MM) part. A number of different partitions of the molecules in QM
and MM regions were used, and they will be detailed as they are
presented. The computational level for the QM part was always that
described in the previous paragraph. For the MM part, the MM3(92)
force field was used.30Van der Waals parameters for the iridium atom
are taken from the UFF force field,31 and torsional contributions
involving dihedral angles with the metal atom in terminal position are
set to zero. All geometrical parameters are optimized without symmetry
restrictions except the bond distances between the QM and MM regions
of the molecules. The frozen values are 1.420 Å (P-H) and 1.112 Å
(C-H) in the QM part and 1.843 Å (P-Csp3), 1.828 Å (P-Csp2),
1.5247 Å (C-C) in the MM part. The starting point of all geometry
optimizations was the X-ray structure coordinates of the cationic iridium
complex.
λ ) 0.71069 Åa
Fcalcd ) 1.483 g cm-1
µ ) 17.112 cm-1
R(Fo)b ) 0.042
Rw(Fo)c ) 0.041
a Graphite monochromator. b R ) ∑|Fo| - |Fc|/∑|Fo|. c Rw ) [∑w(|Fo|
2
- |Fc|)2/∑w|Fo| ]1/2, where w ) 1/σ2(|Fo|).
Table 4. Crystallographic Data for [Ir(H)2(PiPr2Ph)3][BAr′4]
formula ) C68H71BF24IrP3
a ) 17.174(4) Å
b ) 24.980(6) Å
c ) 16.466(3) Å
V ) 7064.02 Å3
Z ) 4
space group ) P212121
T ) -170 °C
λ ) 0.71069 Åa
Fcalcd ) 1.542 g cm-3
µ ) 20.495 cm-1
R(Fo)b ) 0.0886
Rw(Fo)c ) 0.0592
fw ) 1640.22
a Graphite monochromator. b R ) ∑|Fo| - |Fc|/∑|Fo|. c Rw ) [∑w(|Fo|
2
- |Fc|)2/∑w|Fo| ]1/2, where w ) 1/σ2(|Fo|).
Results and Discussion
Synthesis and Characterization of [Ir(H)2(PtBu2Ph)2]-
[BAr′4]. Abstraction of the X ligand from coordinatively
unsaturated Ir(H)2(X)(PtBu2Ph)2 (X ) Cl, F, OSO2CF3) by Na-
[BAr′]4 (Ar′ ) 3,5-bis(trifluoromethyl)phenyl) in fluorobenzene
yields the solvent-ligand-free cationic iridium(III) complex [Ir-
(H)2(PtBu2Ph)2][BAr′4] in nearly quantitative yield. An important
conclusion from this result is the potent Lewis acidity of sodium
in Na[BAr′4], for its ability to abstract the X ligand even when
there is multiple metal-ligand bond character,32 and when a
14 e- species is produced. Crystallization occurs from slow
diffusion of pentane into a concentrated fluorobenzene solution
at -20 °C. This yellow air-sensitive complex is highly soluble
in THF, CH2Cl2, and acetone and demonstrates moderate
solubility in nonpolar arene solvents. The 31P{1H} NMR in d8-
toluene shows two singlets, separated by 0.2 ppm, in the
temperature range +75 to -40 °C. Agostic bonding from C-H
bonds of both phosphines in noncoordinating toluene makes the
phosphorus atoms chiral and, therefore, makes possible the
observation of diastereomers in the NMR spectra. However, the
observation of singlets for both diastereomers indicates that the
agostic bonding is fluxional, with a rapid (on the NMR time
scale) exchange of agostic bonding to both coordination sites
Fourier calculation. The asymmetric unit contains one-half molecule
of benzene solvent, the center of the benzene molecule being at a
crystallographic center of symmetry. Hydrogens bonded to carbons were
included in fixed calculated positions with thermal parameters fixed at
one plus the isotropic thermal parameter of the parent carbon atom.
The hydride ligands could not be identified from the most significant
peaks in the final difference Fourier, and none were included in any of
the calculations. In the final cycles of refinement, all of the non-
hydrogen atoms were varied with anisotropic thermal parameters. The
largest peak in the final difference map was 1.7 e/A3 at the Ir position,
and the deepest hole was -0.9 e/A3.
Crystallographic Details for [Ir(H)2(PiPr2Ph)3][BAr′4]. A suitable
fragment of a larger crystal was mounted on a glass fiber in a nitrogen
atmosphere glovebag using silicone grease. It was then transferred to
a goniostat where it was cooled to -170 °C for characterization and
data collection (Table 4). A systematic search of a limited hemisphere
of reciprocal space was used to determine that the crystal possessed
orthorhombic symmetry and systematic absences indicating the unique
space group P212121. The structure was solved with absorption corrected
data (transmission factors 0.70 to 0.73) using a combination of direct
methods (SHELXTL-PC) and Fourier techniques. Hydrogen atoms were
visible in a difference Fourier phased on the non-hydrogen atoms and
were included as isotropic contributors in the final cycles of refinement.
Some difficulty was encountered in modeling disorder in the CF3 groups
of the anion. Several attempts were made to locate the hydride ligands,
with no success. A final difference Fourier was featureless, the largest
peak being 1.23 e/A3, located adjacent to the Ir position.
1
by each phosphine (Figure 1). The H NMR spectrum (25 °C)
in d8-toluene displays multiple signals for the tBu protons (with
virtual triplet splitting and shifted upfield by 0.5-0.7 ppm in
comparison to the tBu resonance in CD2Cl2), indicating that the
Computational Details. Pure quantum mechanics calculations on
the model systems Ir(H)2[P(Et)H2]2+, Ir(H)2[P(Et)H(CHCH2)]2+, Ir(H)2-
t
phosphines have inequivalent Bu groups. This is attributed to
(PH3)3+, and Ir(H2)(P(Et)H2)3 are carried out with Gaussian 94.22
+
agostic bonding, which causes the inequivalence of agostic tBu
vs pendant tBu groups. In addition, there are two resolved signals
Quasirelativistic effective core potentials replace the 60-electron core
of the Ir atom23 and the 10-electron core of the P atoms.24 The basis
set was valence double-ú for all atoms,24,25 with the addition of a
polarization d shell on phosphorus atoms.26 A larger basis set including
1
in the upfield region of the H NMR spectrum for the hydride
resonances of two diastereomers. The variable-temperature 1H
NMR spectra in d8-toluene reveal two dynamic processes
(22) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.; Peterson, G. A.;
Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski, V.
G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.;
Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.;
Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.;
Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. P.; Head-
Gordon, M.; Gonzalez, C.; Pople, J. A. Gaussian 94: Revision, D. I.;
Gaussian, Inc.: Pittsburgh, PA, 1995.
(23) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299.
(24) Wadt, W. R. Hay, P. J.; J. Chem. Phys. 1985, 82, 284.
(25) Hehre, W. J.; Ditchfield, R.; Pople, J. A. J. Chem. Phys. 1972, 56,
2257.
(27) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28, 213.
(28) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
Johnson, B. G.; Wong, M. W.; Foresman, J. B.; Robb, M. A.; Head-Gordon,
M.; Replogle, E. S.; Gomperts, R.; Andres, J. L.; Raghavachari, K.; Binkley,
J. S.; Gonzalez, C.; Martin, R. L.; Fox, D. J.; Defrees, D. J.; Baker, J.;
Stewart, J. P.; Pople, J. A. Gaussian 92/DFT; Gaussian, Inc.: Pittsburgh,
PA, 1993.
(29) Allinger, N. L. MM3(92); QCPE: Indiana University, 1992.
(30) (a) Allinger, N. L.; Yuh, Y. H.; Lii, J. R. J. Am. Chem. Soc. 1989,
111, 8551. (b) Lii, J. H.; Allinger, N. L. J. Am. Chem. Soc. 1989, 111,
8566. (c) Lii, J. H.; Allinger, N. L. J. Am. Chem. Soc. 1989, 111, 8576.
(31) Rappe´, A. K.; Casewit, C. J.; Colwell, K. S.; Goddard, W. A., III;
Skiff, W. M. J. Am. Chem. Soc. 1992, 114, 10024.
(26) 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.
(32) Caulton, K. G. New. J. Chem. 1994, 18, 25 and references therein.

Steric Influence on Agostic Interactions
J. Am. Chem. Soc., Vol. 121, No. 1, 1999 101
Figure 1. Two diastereomers of doubly agostic Ir(H)2(PtBu2Ph)2
.
+
Figure 3. ORTEP drawing of cation 3 of Ir(H)2(PtBu2Ph)2+, showing
selected atom labeling. Only the hydrogens of the agostic methyl groups
(placed in idealized positions) are shown.
interactions involving other 5d metals.6a,38 Due to the presence
of two tBu groups on each phosphine, it is possible to have an
“internal standard” to accurately gauge the magnitude of bond
deformation inherent in the agostic interactions. For example,
in the cation shown in Figure 3, the Ir(63)-P(64)-C(tBu) angle
t
of the agostic Bu group is 20.5° less than the noninteracting
tBu group of the same phosphine. The six agostic (96.7-98.4°)/
nonagostic (114.3-117.3°) tBu groups in the three cations give
an average Ir-P-C angle decrease of 18.6°.
Figure 2. ORTEP drawing of cation 1 of Ir(H)2(PtBu2Ph)2+, showing
selected atom labeling. Only the hydrogens of the agostic methyl groups
(placed in idealized positions) are shown.
Since the two agostic interactions are mutually cis, the
hydrides must be cis also. In an effort to understand whether
the cis orientation of the hydride ligands is a result of the agostic
interactions from phosphine C-H bonds, the structure of Ir-
involving tBu protons. Coalescence of pendant and agostic tBu
groups is only evident at +75 °C. There is also an observed
+
t
(H)2(PH3)2 was optimized at the B3LYP level with no
broadening of the agostic Bu resonances below -20 °C,
symmetry constraint.39a Despite the lack of steric hindrance or
Ir-agostic interactions from PH3 ligands, the optimized geom-
etry shows a preference for trans phosphines and a cis
arrangement for the hydrides (to minimize mutual influence of
trans hydrides).39b The calculated H-Ir-H angle of 88.2° in
the absence of agostic interactions shows that the cis hydride
orientation in the experimental structure is consistent with a
minimum-energy conformation in the absence of, and not caused
by, the cis agostic interactions. The bent “saw-horse” ML4
structure (essentially an octahedron with two cis ligands missing)
of d6 Ir(III) is more stable than square-planar or tetrahedral since
it is the only geometry associated simultaneously with three
nonbonding occupied d orbitals and two strong M-H interac-
tions.The LUMO is the out-of-phase combination of a metal d
orbital (xy) and the hydride orbitals. Because of the strong
overlap and of the electropositive character of hydride, the
LUMO is at high energy, leading to a large HOMO-LUMO
gap (larger than in a tetrahedron or a square-pyramid structure).
Synthesis and Characterization of [IrH(η2-C6H4PtBu2)-
(PtBu2Ph)][BAr′4]. Ortho-metalation of a phenyl group on a
phosphine ligand is a common reaction for low-valent unsatur-
ated metal complexes, and is promoted by the presence of large
attributed to decoalescence of the methyl groups within an
t
1
agostic Bu. Due to the complexity of the H NMR spectra
resulting from these several dynamic processes, a technique with
a faster time scale was necessary for spectroscopic characteriza-
tion of agostic interactions in solution. The IR spectrum of [Ir-
(H)2(PtBu2Ph)2][BAr′4] in C6D6 displays three bands of medium
intensity at 2625, 2593, and 2552 cm-1. These are assigned to
the agostic C-H stretches of the two diastereomers (Figure 1).33
The X-ray structure of [Ir(H)2(PtBu2Ph)2][BAr′4] shows that
there are three independent [Ir(H)2(PtBu2Ph)2]+, three nonco-
ordinating34 [BAr′4]-, and one (noninteracting) fluorobenzene
per asymmetric unit. The fortuitous crystallization of three
independent molecules in the unit cell allows for six separate
observations of the agostic interactions in three different
positions within the crystal lattice, allowing for the averaging
of crystal packing effects on bond lengths and angles.35 Two
of the cations (Figure 2) are very similar, with only minor
conformational differences within the phosphine ligands, while
the third cation (Figure 3) shows opposite chirality at one
phosphine relative to the first two cations. In all three cations,
there are no close Ir/C(phenyl) contacts (<3.48 Å), but there
are two separate agostic interactions from tert-butyl C-H bonds
to the unsaturated iridium center.36
t
(i.e., Bu) pendant groups on the phosphine,8 in contrast to the
rarity of ortho-metalation involving PMe2Ph.40 The metalation
of a phosphine can have a profound influence on the reactivity
of the complex and radically change the steric profile of the
phosphine ligand. The effective cone angle of the metalated
phosphine increases and can promote the dissociation of other
ligands via increased steric repulsions (i.e., labilizing effect).21
The addition of 1 equiv of Na[BAr′4] to IrH(η2-C6H4P-
tBu2)(Cl)(PtBu2Ph) in CH2Cl2 leads to a rapid abstraction of
the chloride ligand and formation of [IrH(η2-C6H4PtBu2)(PtBu2-
Ph)][BAr′4]. This complex can be isolated as an orange
The metal hydrides and the agostic H were not located by
X-ray diffraction. The six independently measured agostic
interactions are characterized by Ir-C distances of 2.81-2.94
Å. These Ir-C distances are longer than those reported for other
agostic interactions among crystallographically characterized Ir-
(III) complexes,5a,37 but within the range of reported agostic
(33) For comparison, see: Wasserman, H. J.; Kubas, G. J.; Ryan, R. R.
J. Am. Chem. Soc. 1986, 108, 2294.
(34) Strauss, S. H. Chem. ReV. 1993, 93, 927.
(35) For a discussion of this concept in the structural analysis of inorganic
complexes, see: Martin, A.; Orpen, A. G. J. Am. Chem. Soc. 1996, 118,
1464.
(38) Crabtree, R. H. Chem. ReV. 1985, 85, 245.
(36) For an example of a bis-agostic interaction, involving two phenyls,
to a single coordination site, see: King, W. A.; Luo, X.-L.; Scott, B. L.;
Kubas, G. J.; Zilm, K. W. J. Am. Chem. Soc. 1996, 118, 6782.
(37) (a) Crabtree, R. H.; Holt, E. M.; Lavin, M.; Morehouse, S. M. Inorg.
Chem. 1985, 24, 1986. (b) Robertson, G. B.; Wickramasinghe, W. A. Acta
Crystallogr., Sect. C: Cryst. Struct. Commun. 1988, 44, 1383.
(39) (a) Cooper, A. C.; Streib, W. E.; Eisenstein, O.; Caulton, K. C. J.
Am. Chem. Soc. 1997, 119, 9069. (b) For relevant theoretical precedent,
see: Burdett, J. K. J. Chem. Soc., Faraday Trans. 2 1974, 70, 1599. Elian,
M.; Hoffmann, R. Inorg. Chem. 1975, 14, 1058.
(40) Green, M. A.; Huffman, J. C.; Caulton, K. G.; Rybak, W. K.;
Ziolkowski, J. J. Organomet. Chem. 1981, 218, C39.

102 J. Am. Chem. Soc., Vol. 121, No. 1, 1999
Cooper et al.
Table 5. Selected Bond Distances (Å) and Angles (deg) for
[IrH(η2-C6H4PtBu2)(PtBu2Ph)][BAr′4]
Ir(1)-C(29)
Ir(1)-C(8)
2.745(9)
2.042(9)
Ir(1)-P(2)
Ir(1)-P(17)
2.319(3)
2.333(3)
P(2)-Ir(1)-P(17)
P(2)-Ir(1)-C(8)
P(2)-C(3)-C(8)
Ir(1)-P(2)-C(3)
177.58(13) Ir(1)-P(17)-C(28)
94.2(4)
68.7(3)
101.5(7)
83.9(3)
Ir(1)-P(17)-C(24) 116.9(4)
Ir(1)-P(2)-C(9)
Ir(1)-P(2)-C(13)
P(17)-C(28)-C(29) 104.0(7)
115.82(10)
117.17(10)
P(17)-Ir(1)-C(8) 111.2(3)
Ir(1)-P(17)-C(18) 118.9(4)
allacycle occur at the Ir-P-C(ipso) and P-C(ipso)-C(ortho)
angles of the ring. The Ir-P-C(ipso) and P-C(ipso)-C(ortho)
angles (83.9° and 101.5°) are 35.0° and 19.5° smaller than the
corresponding angles of the nonmetalated phosphine. The
agostic interaction (trans to metalated phenyl) is characterized
by a short Ir-C distance (2.745 Å, 0.06-0.19 Å shorter than
agostic Ir-C in [Ir(H)2(PtBu2Ph)2][BAr′4]) and a contraction
Figure 4. ORTEP drawing of IrH(η2-C6H4PtBu2)(PtBu2Ph)+ showing
selected atom labeling. Only the hydrogens of the agostic methyl groups
(placed in idealized positions) are shown.
t
of the Ir-P-C angle involving the agostic Bu group (94.2°,
2.5-4.2° smaller than the agostic Ir-P-C in [Ir(H)2(PtBu2Ph)2]-
[BAr′4]). By these criteria, this agostic interaction is stronger
than those observed for [Ir(H)2(PtBu2Ph)2][BAr′4]. The hydride
ligand and phosphine hydrogens of [IrH(η2-C6H4PtBu2)(PtBu2-
Ph)][BAr′4] were not located by the X-ray diffraction. Placing
the hydrogens of the agostic methyl group in idealized positions
and setting the C-H bond distance at 1.05 Å results in a short
t
(2.032 Å) Ir-H distance. These restrict the ability of the Bu
groups of the metalated phosphine to reach the empty coordina-
tion site, which might have furnished another agostic interaction.
t
To form an agostic interaction from a Bu group of the
metalated phosphine, two of the Ir-P-C angles would have to
be less than 100°. The resulting geometry about the phosphorus
atom would be significantly distorted from tetrahedral and
t
should be very high in energy. The failure of the second Bu
group of the already agostic phosphine ligand (i.e., P(17)) to
form another agostic interaction, giving two agostic interactions
from one phosphine, can be attributed to the same effect. Thus,
the geometrical constraints imposed by ortho-metalation and
the agostic interaction trans to phenyl prevent agostic bonding
to the empty coordination site despite the presence of three
Figure 5. Edge-on view of the Ir-C-P-P plane of IrH(η2-C6H4P-
tBu2)(PtBu2Ph)+.
t
pendant Bu groups.
1
crystalline material by recrystallization from benzene. The H
Synthesis and Characterization of [Ir(H)2(PCy2Ph)3]-
[BAr′4]. An analogue of [Ir(H)2(PtBu2Ph)2][BAr′4], incorporat-
ing smaller alkyl groups on the phosphines, was desired to probe
for the effect of these groups on the formation (or lack) of
agostic interactions. Attempts at the synthesis of [Ir(H)2(PCy2-
Ph)2][BAr′4] were unsuccessful. Abstraction of the chloride
ligand from Ir(H)2Cl(PCy2Ph)2 by Na[BAr′4] gave a mixture
of products including [Ir(H)2(PCy2Ph)3][BAr′4] and monophos-
phine hydride complexes (characterized by doublet splitting of
the hydrides in the 1H NMR spectrum). [Ir(H)2(PCy2Ph)3][BAr′4]
could be isolated in low yield from these mixtures by crystal-
lization from hot benzene. The following rational synthesis of
[Ir(H)2(PCy2Ph)3][BAr′4] was therefore devised: the addition
of 1 equiv of Na[BAr′4] to “Ir(H)2Cl(PCy2Ph)3” (generated in
situ from [Ir(COE)2Cl]2, 6 equiv of PCy2Ph, and excess H2) in
CH2Cl2 or fluorobenzene results in the formation of [Ir(H)2-
(PCy2Ph)3][BAr′4] in high yield. This formally 16-electron
complex can be isolated as a crystalline solid, free of solvent
ligands, by recrystallization from hot benzene.
and 31P{1H} NMR spectra of the product confirms that the
ortho-metalation present in the starting material is retained after
chloride abstraction. 1H NMR shows four doublets for inequiva-
lent tBu groups and a doublet of doublets in the upfield region
for the hydride ligand, due to coupling of the hydride to two
inequivalent phosphorus nuclei. 31P{1H} NMR shows an AX
pattern for the two phosphines, with a large coupling constant
of 278 Hz, characteristic of a mutually trans configuration.
Variable-temperature NMR studies in CD2Cl2 are not able to
show any features consistent with agostic bonding by either tBu
or phenyl groups. However, the crystal structure of [IrH(η2-
C6H4PtBu2)(PtBu2Ph)][BAr′4] (Figures4 and 5) shows a strong
agostic interaction from a tBu group of the terminal phosphine
ligand.
Despite the presence of two empty coordination sites, only
one agostic interaction is observed. The anion is noncoordinat-
ing, and no solvent ligands are present. Examination of the
structural data (Table 5) shows the extreme deformations of
the metalated phosphine that are necessary to form the covalent
Ir-C bond. The P-Ir-P angle (177.58°) is very close to 180°,
showing that the cyclometalation does not significantly affect
the trans coordination of the phosphines. The bond angle
contractions needed to form the strained four-membered met-
1
The H NMR spectrum of [Ir(H)2(PCy2Ph)3][BAr′4] in d8-
THF shows resonances in the phenyl region for the BAr′4 anion
and phenyl groups of the phosphines. A number of broad,
overlapping resonances from 2.41 to 0.85 ppm are assigned to
cyclohexyl protons. The hydride ligands are observed as a single

Steric Influence on Agostic Interactions
J. Am. Chem. Soc., Vol. 121, No. 1, 1999 103
Table 6. Selected Bond Distances (Å) and Angles (deg) for
[Ir(H)2(PCy2Ph)3][BAr′4]
Ir(1)-P(2)
Ir(1)-P(21)
2.4005(16)
2.3163(17)
Ir(1)-P(40)
Ir(1)-C(33)
2.3587(16)
2.923(10)
P(2)-Ir(1)-P(21)
P(2)-Ir(1)-P(40)
103.95(6) C(33)-Ir(1)-P(2)
106.09(6) Ir(1)-P(21)-C(28)
108.33(4)
100.94(21)
113.86(21)
P(21)-Ir(1)-P(40) 149.95(6) Ir(1)-P(21)-C(22)
C(33)-Ir(1)-P(40) 108.29(4) P(21)-C(28)-C(33) 107.1(4)
C(33)-Ir(1)-P(21) 61.44(6) P(21)-C(22)-C(27) 115.8(4)
106.09°) much greater than 90° and a trans P-Ir-P angle of
the axial phosphines (149.95°), greatly reduced from 180°. The
most interesting feature of the crystal structure of [Ir(H)2(PCy2-
Ph)3][BAr′4] is an agostic interaction from a C-H bond of one
of the cyclohexyl groups on an axial phosphine ligand. As
expected, the BAr′4 anion is noncoordinating, and crystallization
from C6H6 resulted in no solvent coordination to the empty site.
While the hydride ligands and hydrogens of the phosphines were
not located in the X-ray study, the agostic interaction can be
characterized by a number of structural parameters presented
in Table 6.
The Ir-C distance (2.923 Å) to the agostic C-H is within
the range of distances observed for the agostic Ir-C distances
in [Ir(H)2(PtBu2Ph)2][BAr′4] (2.81-2.94 Å), implying that the
strength of the interactions may be similar. The Ir-P-C(R) and
P-C(R)-C(â) angles (100.94° and 107.1°) involving the agostic
cyclohexyl group are significantly decreased from the corre-
sponding angles of the pendant cyclohexyl of the same
phosphine (113.86° and 115.8°).
Figure 6. ORTEP drawing of Ir(H)2(PCy2Ph)3+, showing selected atom
labeling. Only the hydrogens of the agostic methylene groups (placed
in idealized positions) are shown.
broad resonance (singlet, -26.1 ppm) at room temperature,
suggestive of a rapid site exchange between the hydride
ligands.41 Lowering the temperature results in a broadening of
the hydride resonance and eventual decoalescence of two
1
hydride signals at ca. -80 °C. At -110 °C, the H NMR
spectrum shows signals at -5.4 ppm (br doublet, JPH ) 107
Hz) and -44.8 ppm (br s, w1/2 ) 70 Hz). These signals suggest
a static structure (I) at this temperature, in which one hydride
Five-membered rings may be formed by agostic interaction
with either a C-H bond of a cyclohexyl group or an o-phenyl
C-H bond, but a preference to form the agostic interaction with
the aliphatic group is demonstrated by the results presented here,
as it also is in [Ir(H)2(PtBu2Ph)2][BAr′4] and [IrH(η2-C6H4P-
tBu2)(PtBu2Ph)][BAr′4]. While the smaller cone angle of PCy2-
Ph (estimated at 162°),42 compared to PtBu2Ph, is apparent by
the coordination of three phosphines in [Ir(H)2(PCy2Ph)3]-
[BAr′4], the steric bulk of the cyclohexyl and phenyl groups of
the phosphine is still sufficient to promote an agostic interaction
in [Ir(H)2(PCy2Ph)3][BAr′4]. Inter-phosphine repulsions are
clearly apparent in the structure by the large deformations from
a rigorous mer geometry (i.e., P-Ir-P bond angles of 90 and
180°). Therefore, both substituent effects from the pendant
groups of the agostic phosphine and inter-phosphine steric
repulsions must be considered in this case. This makes the
consideration of the influence of steric factors on the agostic
interaction more complicated than in the case of [Ir(H)2(PtBu2-
Ph)2][BAr′4], where the steric repulsions between mutually trans
phosphines (P-Ir-P angle close to 180°) could be discounted.
Synthesis and Characterization of [Ir(H)2(PiPr2Ph)3]-
[BAr′4]. The synthesis and characterization of [Ir(H)2(PiPr2Ph)3]-
[BAr′4] were accomplished in an effort to study the effect of
changing the bulky substituents of the phosphine ligands from
those in [Ir(H)2(PCy2Ph)3][BAr′4] while minimizing the differ-
ence in phosphine donating ability. PCy2Ph and PiPr2Ph should
have very similar donating abilities, both containing P-C bonds
to two secondary alkyls and one phenyl substituent.
is trans to a phosphine, the other hydride is trans to the empty
coordination site, and phosphines adopt a mer arrangement in
a square pyramid. There are no significant changes in the phenyl
or cyclohexyl resonances in the temperature range +25 to -110
°C.
Another fluxional process is observed in the variable-
temperature 31P{1H} NMR spectra. At 25 °C, the three phos-
phines are observed as a single resonance at 28.3 ppm. Cooling
the solution causes a broadening of this signal and decoalescence
is observed at ca. -20 °C. At temperatures below -30 °C, two
resonances are present at 28.9 and 28.2 ppm in a 2:1 integrated
ratio. These data are consistent with a decoalescence of two
phosphine environments in either a mer or fac IrH2P3+ structure.
Upon the basis of the strong trans influence of the hydrides,
the mer configuration (I) should be favored. These signals
remain broad down to -100 °C, and no phosphorus-
phosphorus coupling is observed.
Unlike [Ir(H)2(PtBu2Ph)2][BAr′4], [Ir(H)2(PCy2Ph)3][BAr′4]
has very poor solubility in arene solvents. This unfortunate
solubility problem has curtailed the study of a possible agostic
interaction by NMR or IR spectroscopy in noncoordinating
solvents. Therefore, the determination of the structure of [Ir-
(H)2(PCy2Ph)3][BAr′4] by X-ray diffraction was performed
(Figure 6).
Structural analysis of the IrP3 unit (Table 6) shows a highly
distorted mer arrangement for the phosphine ligands, due to the
steric interactions inherent in the coordination of three bulky
PCy2Ph to the Ir. The axial phosphines are bent away from the
equatorial phosphine, giving cis Pax-Ir-Peq angles (103.95° and
The addition of 1 equiv of Na[BAr′4] to “Ir(H)2Cl(PiPr2Ph)3”
(generated in situ from [Ir(COE)2Cl]2, 6 equiv of PiPr2Ph, and
excess H2) in CH2Cl2 or fluorobenzene results in the formation
of [Ir(H)2(PiPr2Ph)3][BAr′4] in high yield. This formally 16-
electron complex can be isolated as a crystalline solid, free of
solvent ligands, by recrystallization from benzene.
(41) Gusev, D. G.; Berke, H. Chem. Ber. 1996, 129, 1143.
(42) Tolman, C. A. Chem. ReV. 1977, 77, 313.

104 J. Am. Chem. Soc., Vol. 121, No. 1, 1999
Cooper et al.
Table 7. Selected Bond Distances (Å) and Angles (deg) for
[Ir(H)2(PiPr2Ph)3][BAr′4]
Ir(1)-P(2)
Ir(1)-P(15)
2.323(6)
2.394(8)
Ir(1)-P(28)
Ir(1)-C(37)
2.320(10)
3.463(8)
P(2)-Ir(1)-P(28)
P(2)-Ir(1)-P(15)
146.4(3)
Ir(1)-P(28)-C(29) 126.7(10)
107.79(24) Ir(1)-P(28)-C(35) 108.5(14)
P(15)-Ir(1)-P(28) 105.8(3)
Ir(1)-P(28)-C(38) 110.4(13)
Ir(1)-P(2)-C(12) 108.3(9)
Ir(1)-P(2)-C(3)
Ir(1)-P(2)-C(9)
122.5(10)
114.2(9)
+
Figure 8. ORTEP drawings of Ir(H)2(PCy2Ph)3 (left) and Ir(H)2-
(PiPr2Ph)3 (right) viewed to show the similarity of the phosphine
+
substituent conformations. Only the agostic methylene group hydrogens
are shown at left.
[BAr′4] (Figure 8). The contracted Ir-P-C angle of the agostic
cyclohexyl group in [Ir(H)2(PCy2Ph)3][BAr′4] is clearly not
present in [Ir(H)2(PiPr2Ph)3][BAr′4]. Note that the configurations
of alkyl and phenyl groups in the phosphine ligands of the two
structures are nearly identical, suggesting that the effect of
intermolecular contacts within the crystal on the configuration
of the phosphine substituents is minimal. These conformations
must be determined primarily by intramolecular contacts.
If so many similarities between the two structures exist, how
then can one contain an agostic interaction, while the other does
not? Differences in the electrophilicity of the Ir centers should
be minimal. The donating ability of the two phosphines is
expected to be very similar, and the first coordination spheres
are nearly identical. Thus, it is reasonable to conclude that, in
determining the formation (or lack) of agostic bonds in these
two complexes, steric effects predominate. The origin of the
difference in the structures appears to lie mainly in the identity
of the alkyl groups of the phosphine ligands. Cyclohexyl
apparently has a larger steric impact than isopropyl, as evidenced
by the difference in the cone angles of PCy3 (170°) vs PiPr3
(160°).42 Cyclohexyl ring constraints limit distortions of the
â-methylene groups which might otherwise lower steric repul-
sions. The methyl groups of isopropyl have no such constraint.
For the molecules [Ir(H)2(PR2Ph)3][BAr′4], the average of the
C-P-C angles of the nonagostic phosphines is (slightly) larger
for R ) Cy (103.2°) vs R ) iPr (102.4°). Larger C-P-C angles
are characteristic of increased steric repulsions among the alkyl
groups of the phosphine.
Figure 7. ORTEP drawing of Ir(H)2(PiPr2Ph)3+, showing selected atom
labeling.
1
Variable-temperature H and 31P{1H) NMR spectra of [Ir-
(H)2(PiPr2Ph)3][BAr′4] in d8-THF are very similar to those
observed for [Ir(H)2(PCy2Ph)3][BAr′4]. The room-temperature
1H NMR spectra show the expected resonances for the BAr′4
and PiPr2Ph protons. The hydride ligands are observed at 25
°C as a single broad resonance at -25.4 ppm. Cooling the
solution to ca. -80 °C causes decoalescence of the two hydrides
as separate, broad signals at -5.8 and -43.8 ppm. Large P-H
coupling (104 Hz) can be resolved on the downfield hydride
signal at -100 °C. The 31P{1H} NMR spectrum at room
temperature shows one resonance at 35.2 ppm, which decoa-
lesces into two signals at 35.4 and 34.4 ppm (2:1 integrated
ratio) at ca. -30 °C. Due to the large line width of these signals,
no phosphorus-phosphorus coupling is observed at any tem-
perature down to -100 °C.
Calculational Evidence for the Importance of Substituent
Bulk on Agostic Interactions. (a) Agostic Interaction: Elec-
tronic and/or Steric Origin? The hybrid (QM/MM) IMOMM
method16 has been successful in reproducing various types of
distortion of the coordination sphere caused by bulky ligands.17
In the particular case of unsaturated molecules, the attempt to
decrease intraligand and interligand repulsive interactions can
direct pendant groups toward the space left vacant by the missing
ligand. In doing so, a C-H bond could come sufficiently close
to the metal to be viewed as being involved in agostic
interaction. A fundamental question thus arises: is the short
M‚‚‚H-C distance the result of minimizing steric repulsion with
no additional attractive M-C-H attraction or is there also an
attraction between the weak C-H Lewis base and the Lewis
acid empty metal site? The theoretical studies presented in a
previous communication on Ir(H)2(PtBu2Ph)2+ have shown the
determining influence of steric factors for bringing CH3 within
bonding distance of the metal.15 When the real complex is
modeled by Ir(H)2(PH2C2H5)2+, B3LYP calculations show the
The structure of [Ir(H)2(PiPr2Ph)3][BAr′4] was determined by
X-ray diffraction. While the angles of the IrP3 core (Table 7)
show a distorted mer geometry in which the trans phosphine
ligands are distorted away from the equatorial phosphine ligand
(Figure 7), closely resembling those of [Ir(H)2(PCy2Ph)3][BAr′4],
there is a notable lack of any agostic bond to the empty
coordination site in [Ir(H)2(PiPr2Ph)3][BAr′4]. The shortest Ir-C
distance is 3.338 Å, involving a methine carbon on an axial
phosphine, and the shortest Ir-C(methyl) distance is 3.463 Å,
clearly inconsistent with Ir‚‚‚H-C bonding. There is no
perceptible contraction of any Ir-P-C angles, further proving
the lack of any agostic interaction in [Ir(H)2(PiPr2Ph)3][BAr′4].
i
The Pr and Cy analogues are thus not isostructural! Indeed,
the crystals are not even isomorphous.
+
Comparison of Ir(H)2P3 Structures. The geometrical
distortions necessary to form the five-membered ring of an
agostic bond in [Ir(H)2(PCy2Ph)3][BAr′4] become apparent when
this structure is compared directly to that of [Ir(H)2(PiPr2Ph)3]-

Steric Influence on Agostic Interactions
J. Am. Chem. Soc., Vol. 121, No. 1, 1999 105
Table 8. Experimental and Calculated Structural Parameters for
Ir(H)2(PCy2Ph)3+, Using Various Computational Models (see Text)
exptl B3LYP/I MP2/I B3LYP/II MP2/II
Ir-P(21)
Ir-P(40)
Ir-P(2)
2.316
2.359
2.400
2.381
2.419
2.528
152.6
102.0
105.4
3.178
105.6
111.0
109.6
2.374
2.400
2.508
152.0
102.3
105.6
3.176
105.5
111.4
109.9
2.375
2.424
2.529
152.0
102.9
105.0
3.107
103.8
110.4
109.4
2.350
2.400
2.500
151.0
103.0
105.9
2.881
99.8
P(21)-Ir-P(40) 150.0
P(2)-Ir-P(21)
P(2)-Ir-P(40)
Ir‚‚‚C(33)
104.0
106.0
2.923
Ir-P(21)-C(28) 100.9
Ir-P(21)-C(27) 113.9
111.6
109.9
av angle at P(3)
109.7
absence of agostic interaction since the C atom remains 3.98 Å
from the metal. The calculated Ir-P-CR is 118.8°, compared
to the experimental value of 97°. The energy gain associated
with the weak C-H‚‚‚Ir interaction is thus too small to effect
the necessary bending of the Ir-P-C angle. While IrH2L2+ is
a 14-electron and thus highly electron-deficient complex, the
two hydrides with their strong trans influence limit the energy
gain of bonds trans to the hydrides.
+
Figure 9. Optimized (MP2/MM3) structure of Ir(H)2(PCy2Ph)3
.
tions are given in Table 8, along with the experimental values.
Figure 9 shows the calculated structure resulting from MP2/II
level and is qualitatively similar for all methods of calculations.
Numbering of heavy atoms is identical to that in the experi-
mental structure.
At all levels of calculations, the geometry is that of a strongly
distorted square-based pyramid with one apical hydride. The
overall coordination of the Ir atom is well reproduced, and the
angles P-Ir-P differ from the experimental value by less than
2°. The P(21)-Ir-P(40) angle of 150° and the P(2)-Ir-P(21)
and P(2)-Ir-P(40) angles of 105° clearly reflect the large bulk
of the PCy2Ph groups. The calculated Ir-P(21) and Ir-P(40)
bond distances are longer than the experimental values by 0.065
Å, while the discrepancy is even larger (0.13 Å) for Ir-P(2),
which is trans to a hydride. The best agreement is obtained for
MP2/II. Models I and II give reasonably close results within
the same method of quantum calculations. Using B3LYP in
place of MP2 results in an increase in the calculated Ir-P bond
lengths. It is, however, satisfying that all levels of calculations
reproduce the relative Ir-P bond lengths.
From the X-ray results, the shortest Ir‚‚‚C(33) distance (2.923
Å) indicates the presence of an agostic C-H bond. Associated
with this short distance is a strong decrease in the Ir-P(21)-
C(28) angle, to 100.9°, while the Ir-P(21)-C(27) (nonagostic
cyclohexyl) angle is 113.9°. For the other phosphine ligands,
the experimental Ir-P-C(cyclohexyl) angles are around 110°.
For B3LYP/I and MP2/I calculations, the angles for the
nonagostic groups are quantitatively reproduced within 0.5°. The
trend in the angles at P(21) is correctly represented since the
smallest value (105.5°) is for Ir-P(21)-C(28) (agostic cyclo-
hexyl), while the Ir-P(21)-C(27) (nonagostic cyclohexyl) is
111.0°. Angles at the carbon centers are all around 109°. Thus,
as for IrH2(PtBu2Ph)2+, the essentials of the Thorpe-Ingold
effect are visible in the distorted Ir-P-C angles. Since Ir-
P(21)-C(28) is slightly too large and since the Ir-P bonds are
also too long, the calculations with model I (B3LYP or MP2)
give an Ir‚‚‚C(33), which is 0.25 Å too long (3.18 Å) and equal
for the two quantum methods.
+
The IMOMM calculations on this same Ir(H)2(PtBu2Ph)2
+
complex were carried out with Ir(H)2(PH2C2H5)2 in the QM
part. The results mimic well the Thorpe-Ingold effect men-
tioned earlier.14 Remarkably, this is achieved primarily by
decreasing the Ir-P-CR angle (102.6°) and leaving the P-CR-
Câ angles essentially undistorted. While the calculations clearly
illustrate the importance of the full ligand in achieving agostic
interactions, they do not permit separation of the electronic and
steric components in the agostic interaction. The carbon chain
carrying the agostic group is part of the QM ensemble, while
the other substituents at the phosphorus are part of the MM
ensemble. We extend here the previous study to the Ir(H)2(PCy2-
Ph)3+ and Ir(H)2(PiPr2Ph)3+ species containing three phosphine
groups, and we propose also a way to have a better understand-
ing of the true factors by doing two sets of calculations with
different partitioning of the substituents of the phosphine ligands
in the QM and MM ensembles.
+
(b) Ir(H)2(PCy2Ph)3+. The complex Ir(H)2(PCy2Ph)3 was
calculated by the IMOMM method. Two sets of partitioning of
atoms between the QM and MM domains have been chosen.
In model I, all atoms that are not directly bonded to the metal
are calculated at the MM level. This means that for the QM
part, the phosphine is represented by PH3 and that the C-H
bonds in the vicinity of the metal carry no electronic density.
In model II, the QM part of the phosphine includes PH2C2H5,
where the terminal C-H bond of the ethyl substituent is that
which could come into proximity to the metal center. The
remaining part of this cyclohexyl as well as the second
cyclohexyl and the phenyl group are part of the MM calcula-
tions. Model II differs from model I in that electron density is
present in the C-H bond which could come into proximity to
the metal center. Two levels of calculations, B3LYP and MP2,
were selected for the QM part. While B3LYP has had large
success for providing very good geometries in transition metal
complexes, it underestimates weak interactions.43 The MP2
method has also given good results for geometry and seems to
give a better representation of weak interactions. Its wide use
has been limited because of its requirement in computational
effort.
Since model I describes the cyclohexyl and phenyl group at
the MM level only, the results of the calculations show that the
bulk of the ligands alone, in the absence of any attractive forces
between the metal and C-H bonds, has positioned one of the
cyclohexyl substituents with one C-H bond in the vicinity of
the empty site of the metal. However, the distance M‚‚‚H is
beyond any bonding interaction. The ability to determine the
consequence of only one type of interaction (here, steric) is a
strength of the hybrid method: the execution of “computational
experiments”.
Ir(H)2(PCy2Ph)3+ was fully optimized at four levels: B3LYP
and MP2 for models I and II (B3LYP or MP2/I or II). The
most relevant geometrical parameters from these four calcula-
(43) Ruiz, E.; Salahub, D. R.; Vela, A. J. Phys. Chem. 1996, 100, 12265.

106 J. Am. Chem. Soc., Vol. 121, No. 1, 1999
Cooper et al.
locate an agostic interaction if it is present. Initiating the
optimization process from the experimental geometry leads to
a structure that is noticeably different from the experimental
one (Figure 10). Several groups have rotated from the observed
orientation in the solid state, and all of the Ir‚‚‚C nonbonding
distances are larger than 3.5 Å (3.4 Å experimentally). These
facts are evidence for a shallow energy surface for the groups
on the phosphine ligands, where orientation is not significantly
influenced by any agostic interaction. The average of the Ir-
P-C(isopropyl) is calculated to be 115.7° (114.2° experimen-
tally). The observed orientation of the ligands can then be
influenced by the crystal packing.
Figure 10. Optimized (MP2/MM3) structure of Ir(H)2(PiPr2Ph)3
.
+
Conclusions
+
With model II, the Ir‚‚‚C(33) distances are shorter than with
model I and the best results are obtained for the MP2
calculations (Ir‚‚‚C(33) ) 2.881 Å, which is very close to the
experimental value of 2.923 Å). For this calculation, there is
excellent agreement for all Ir-P-C angles, including the one
corresponding to the agostic cyclohexyl group. The B3LYP/II
calculations give a significantly greater distance Ir‚‚‚C(33)
(3.107 Å). In fact, the distance Ir‚‚‚C(33) is only slightly shorter
than for the pure molecular mechanics model of cyclohexyl
(method I). This confirms that the B3LYP significantly under-
estimates the energy of the electronic part of the agostic
interaction.
The experimental structural comparison of Ir(H)2(PCy2Ph)3
to Ir(H)2(PiPr2Ph)3 and the computational experiments on Ir-
+
(H)2(PR2Ph)n+, where R ) Bu, Pr, and Cy, and simplified
systems represent the “most pure” tests of the effect of steric
constraints on agostic bonding. Both these experimental and
theoretical studies seek to manipulate steric parameters within
a system while bringing minimal changes to the electronic
(bonding) parameters (i.e., electron density of metal orbitals).
Such careful changes are necessary when studying a phenom-
enon which relies on both steric and electronic factors. Based
upon these studies, the lowering of the energetic barriers of bond
deformation and loss of rotational entropy by bulky substituents
on the agostic ligand, placing the C-H in the proximity of the
empty coordination site, play the critical role in the formation
of agostic bonds to these unsaturated iridium complexes.
The agostic bond is clearly very weak in these systems,
probably because of the presence of the two hydrides. This
permits the steric factors to play an active (“encouraging” or
supporting) role in making the agostic interactions feasible. It
also permits the agostic interaction to be remarkably sensitive
to small changes in the nature of the phosphine. In this work,
t
i
The two factors that are important in this complex for the
occurrence of the agostic interaction are clearly apparent in these
calculations. The steric bulk of the ligands moves one cyclohexyl
group into proximity of the metal empty site as shown by the
pure molecular mechanics calculations (method I). It is remark-
able that the orientation of all the ligands in the calculated
structure is identical to that in the calculated structure. This
shows that the structure adopted by the compound in the solid
state is at least a local minimum on the potential energy surface
for the isolated molecule and also suggests a very small
influence of the crystal packing. When the level of calculations
is changed from method I to method II, the conformations of
the substituents of the phosphine remain identical. The only
consequence of the model I/II upgrade is that the C-H bond
moves toward the metal, showing the presence of a true
attraction between Ir and the C-H bond. As expected, the
agostic C-H bond is longer (1.11 Å) than those that are not in
contact with the metal (1.09 Å) at the (MP2/II) level. The
attraction of C-H toward the iridium center is accomplished
primarily by bending the angle at the phosphine.
i
the subtle change between the two secondary alkyls Pr and
cyclohexyl is enough to cause an agostic interaction to be absent/
present. Of course, steric factors are not mandatory for agostic
interactions, as evidenced by the numerous examples deprived
of any apparent steric strain.44 Finally, the hybrid QM/MM
method is a powerful tool for detecting and analyzing the relative
role of the electronic and steric factors.
Acknowledgment. This work was supported by the U.S.
National Science Foundation, the University of Montpellier, and
the French CNRS. F.M. thanks the French CNRS for a position
of research associate.
+
(c) Ir(H)2(PiPr2Ph)3+. The complex Ir(H)2(PiPr2Ph)3 was
calculated only with MP2/II, which maximizes the chance to
Supporting Information Available: Tables giving full
crystallographic details, positional and thermal parameters, and
bond distances and angles (23 pages, print/PDF). See any current
masthead page for ordering information and Web access
instructions.
(44) For recent references, see inter alia: (a) Han, Y. Z.; Deng, L. Q.;
Ziegler, T. J. Am. Chem. Soc. 1997, 119, 5939. (b) Deng, L. Q.; Margl, P.;
Ziegler, T. J. Am. Chem. Soc. 1997, 119, 1094. (c) Musaev, D. G.; Froese,
R. D. J.; Svensson, M.; Morokuma, K. J. Am. Chem. Soc. 1997, 119, 367.
(d) Musaev, D. G.; Svensson, M.; Morokuma, K.; Stromberg, S.; Zetterberg,
K.; Siegbahn, P. E. H. Organometallics 1997, 16, 1933. (e) Woo, T. K.;
Fan, L.; Ziegler, T. Organometallics 1994, 13, 2252.
JA981727E