Metal-metal cooperativity effects in promoting C-H bond cleavage of a methyl group by an adjacent metal center

Article abstract of DOI:10.1021/ja983525o

The reaction of [Ir₂(CO)₃(dppm)₂] (dppm = Ph₂PCH₂PPh₂) with methyl triflate yields the methylene-bridged hydride [Ir₂H(CO)₃(μ- CH₂)(dppm)₂][CF₃SO₃] (2), in which the hydride and methylene hydrogens are rapidly scrambling at ambient temperature. Under CO this species yields the methyl and acyl products [Ir₂-(R)(CO)₄(dppm)₂][CF₃SO₃] (R = CH₃, C(O)CH₃). Removal of one carbonyl from 2 yields the fluxional methyl complex [Ir₂(CH₃)(CO)₂(dppm)₂] [CF₃SO₃] (3) in which the methyl group readily migrates from metal to metal. Addition of CO, PR₃, CN(t)Bu or SO₂ to 3 results in C-H bond Cleavage of the methyl group yielding the methylene- bridged, hydride species, [Ir₂H(CO)₂L(μ-CH₂)(dppm)₂][CF₃SO₃] (L = CO, PR₃, CN(t)Bu) or [Ir₂H-(CO)₂(μ-CH₂)(μ-SO₂)(dppm)₂][CF₃SO₃] (11). Both the carbonyl and SO₂ adducts are fluxional, having the hydrogens of the hydrido ligand and the methylene group exchanging rapidly at ambient temperature. The activation parameters for this reversible C-H bond-making and -breaking step have been determined (ΔH(+) = 10.3 kcal/mol, ΔS(+) = - 11.2 cal/mol K (CO); ΔH(+) = 6.1 kcal/mol, ΔS(+) = -6.2 cal/mol K (SO₂)). X-ray structure determinations of compound 3, [Ir₂H(CO)₂(PMe₃)(μ- CH₂)(dppm)₂][CF₃SO₃] (6), and compound 11 have been determined to confirm the proposed structures. Density functional theory calculations have been carried out on cations related to 3 and 11 by substitution of the phenyl substituents on the dppm ligands by hydrogens, and on key isomers of these, to gain an understanding of the factors promoting C-H bond cleavage in this system. A proposal is presented rationalizing the facile C-H bond cleavage in 3 upon addition of the substrate molecules, in which the roles of the adjacent metals are described.

Full text of DOI:10.1021/ja983525o

3666
J. Am. Chem. Soc. 1999, 121, 3666-3683
Metal-Metal Cooperativity Effects in Promoting C-H Bond
Cleavage of a Methyl Group by an Adjacent Metal Center
Jeffrey R. Torkelson,‡ Frederick H. Antwi-Nsiah,‡ Robert McDonald,†,‡
Martin Cowie,*,‡ Justin G. Pruis,§ Karl J. Jalkanen,§ and Roger L. DeKock§
Contribution from the Department of Chemistry, UniVersity of Alberta, Edmonton, AB, Canada T6G 2G2,
and Department of Chemistry and Biochemistry, CalVin College, Grand Rapids, Michigan 49546-4388
ReceiVed October 5, 1998
Abstract: The reaction of [Ir2(CO)3(dppm)2] (dppm ) Ph2PCH2PPh2) with methyl triflate yields the methylene-
bridged hydride [Ir2H(CO)3(µ-CH2)(dppm)2][CF3SO3] (2), in which the hydride and methylene hydrogens are
rapidly scrambling at ambient temperature. Under CO this species yields the methyl and acyl products [Ir2-
(R)(CO)4(dppm)2][CF3SO3] (R ) CH3, C(O)CH3). Removal of one carbonyl from 2 yields the fluxional methyl
complex [Ir2(CH3)(CO)2(dppm)2][CF3SO3] (3) in which the methyl group readily migrates from metal to metal.
Addition of CO, PR3, CNtBu or SO2 to 3 results in C-H bond cleavage of the methyl group yielding the
methylene-bridged, hydride species, [Ir2H(CO)2L(µ-CH2)(dppm)2][CF3SO3] (L ) CO, PR3, CNtBu) or [Ir2H-
(CO)2(µ-CH2)(µ-SO2)(dppm)2][CF3SO3] (11). Both the carbonyl and SO2 adducts are fluxional, having the
hydrogens of the hydrido ligand and the methylene group exchanging rapidly at ambient temperature. The
activation parameters for this reversible C-H bond-making and -breaking step have been determined (∆H‡ )
10.3 kcal/mol, ∆S‡ ) -11.2 cal/mol K (CO); ∆H‡ ) 6.1 kcal/mol, ∆S‡ ) -6.2 cal/mol K (SO2)). X-ray
structure determinations of compound 3, [Ir2H(CO)2(PMe3)(µ-CH2)(dppm)2][CF3SO3] (6), and compound 11
have been determined to confirm the proposed structures. Density functional theory calculations have been
carried out on cations related to 3 and 11 by substitution of the phenyl substituents on the dppm ligands by
hydrogens, and on key isomers of these, to gain an understanding of the factors promoting C-H bond cleavage
in this system. A proposal is presented rationalizing the facile C-H bond cleavage in 3 upon addition of the
substrate molecules, in which the roles of the adjacent metals are described.
Introduction
analogous clusters of two or more metals often give rise to
R-hydrogen elimination.4,5 This difference is readily rationalized
upon consideration of the geometries of both systems. In a
mononuclear complex, represented by structure A, the â-hy-
drogens are favorably disposed for formation of an agostic
interaction6 with the metal, often leading to C-H bond cleavage.
As pointed out by Earl Muetterties more than 15 years ago,1
“Surface chemistry is coordination chemistry. If the molecule
or molecules in question are organic then the surface chemistry
is organometallic chemistry.” We therefore expect that organic
transformations occurring on metal surfaces will be fundamen-
tally no different than those occurring with homogeneous metal
complexes. With these parallels in mind, it is therefore not
surprising that homogeneous, organometallic complexes can
serve as useful models for heterogeneous catalysts in organic
transformations.2 Although heterogeneous systems are notori-
ously difficult to study, homogeneous systems are more
amenable, with the aid of a battery of spectroscopic techniques.
The simplest model systems are based on single-metal (mono-
nuclear) complexes; however, one aspect that distinguishes metal
surfaces from mononuclear systems is the presence of adjacent
metals in the former, which may have a significant effect on
the chemistry of interest. The influence of an adjacent metal is
clearly indicated in a comparison of alkyl complexes of
mononuclear systems and those of clusters. In mononuclear late-
metal systems â-hydrogen elimination is common,3 whereas
In clusters, in which the alkyl group is bound to one metal while
adjacent to at least one additional metal, as represented by B,
the R-hydrogens are oriented so that favorable interactions with
the adjacent metal can occur,7 leading to activation of this C-H
bond.
Our current interest in binuclear complexes is aimed at
determining how adjacent metals can influence the reactivity
(3) Collman, J. P.; Hegedus, L. S.; Norton, J. R.; Finke, R. G. Principles
and Applications of Organotransition Metal Chemistry; University Science
Books: Mill Valley, CA, 1987; p 383.
* To whom correspondence should be addressed.
‡ University of Alberta.
(4) Crabtree, R. H. The Organometallic Chemistry of the Transition
Metals; John Wiley & Sons: New York, 1988; p 327.
(5) Calvert, R. B.; Shapley, J. R. J. Am. Chem. Soc. 1977, 99, 5225.
(6) (a) Brookhart, M.; Green, M. L. H. J. Organomet. Chem. 1983, 250,
395. (b) Brookhart, M.; Green, M. L. H.; Wong, L.-L. Prog. Inorg. Chem.
1988, 36, 1.
† Faculty Service Officer, Structure Determination Laboratory.
§ Calvin College.
(1) Muetterties, E. L. Pure Appl. Chem. 1982, 54, 83.
(2) See for example: (a) Marks, T. J. Acc. Chem. Res. 1992, 25, 57. (b)
Zaera, F. Chem. ReV. 1995, 95, 2651. (c) Atwood, J. D. Coord. Chem.
ReV. 1988, 83, 93.
(7) See for example: Jeffrey, J. C.; Orpen, A. G.; Stone, F. G. A.; Went,
M. J. J. Chem. Soc., Dalton Trans. 1986, 173.
10.1021/ja983525o CCC: $18.00 © 1999 American Chemical Society
Published on Web 04/06/1999

Metal-Metal CooperatiVity and C-H Bond CleaVage
J. Am. Chem. Soc., Vol. 121, No. 15, 1999 3667
or as solutions in KCl cells with 0.5-mm-window path lengths. Carbonyl
stretches reported are for isotopically nonenriched samples. The
elemental analyses were performed by the microanalytical service within
the department. Spectroscopic data for all compounds are given in Table
1.
Preparation of Compounds. (a) [Ir2(H)(CO)3(µ-CH2)(dppm)2]-
[CF3SO3] (2). The compound [Ir2(CO)3(dppm)2] (1) (100 mg, 0.081
mmol) was dissolved in 10 mL of benzene, to which was added 1 equiv
of methyl triflate (9.1 µL, 0.081 mmol). The light yellow solution
immediately darkened to orange and turned cloudy after about 10 min.
The mixture was stirred for 3 h by which time a yellow precipitate
had formed. Removal of the solvent under vacuum gave a yellow
product which was recrystallized from CH2Cl2/Et2O. This solid was
dried under a stream of argon and put under vacuum for about 18 h,
during which time the color changed from yellow to light brown (86%
yield). Anal. Calcd for Ir2SP4F3O6C55H47: C, 47.20; H, 3.37. Found:
C, 47.32; H, 3.40.
(b) [Ir2(CH3)(CO)(µ-CO)(dppm)2][CF3SO3] (3). Compound 2 (80
mg, 0.057 mmol) was dissolved in 5 mL of CH2Cl2. One equivalent of
anhydrous Me3NO (4.3 mg, 0.057 mmol), dissolved in 5 mL of CH2-
Cl2, was added dropwise from an addition funnel to the solution of 2
during 15 min, while maintaining a slow stream of argon over the
solution. The mixture was stirred for 45 min, resulting in a color change
to dark red, and the solvent was removed under vacuum. The solid
was then recrystallized from CH2Cl2/Et2O, dried under argon, and then
in vacuo (78% yield). Anal. Calcd for Ir2SP4F3O5C54H47: C, 47.23; H,
3.45. Found: C, 46.81; H, 3.24.
(c) [Ir2(H)(CO)2(PMe3)(µ-CH2)(dppm)2][CF3SO3] (6). Method A.
A solution of PMe3 [71.4 µL of a 1 M solution in tetrahydrofuran (THF)
(0.071 mmol)] in 10 mL of THF was added to a 20-mL THF solution
of compound 2 (100 mg, 0.071 mmol), resulting in an immediate color
change from orange to bright yellow. The solution was stirred for 2 h
during which time a yellow precipitate formed. Removal of the solvent
under vacuum gave a light yellow microcrystalline solid, which was
then recrystallized from THF/Et2O (70% yield).
Method B. To a solution of compound 3 (30 mg, 0.022 mol) in 0.6
mL of CD2Cl2, in an NMR tube capped with a rubber septum, was
added 1 equiv (2.3 µL, 0.022 mmol) of PMe3 leading to an immediate
color change to bright yellow. The NMR spectra showed that the
product was compound 6. Anal. Calcd for Ir2SP5F3O5C57H56: C, 47.23;
H, 3.89. Found: C, 47.12; H, 4.12.
(d) [Ir2(H)(CO)2(PMe2Ph)(µ-CH2)(dppm)2][CF3SO3] (7). Method
A. The procedure was the same as that described for compound 6 except
that PMe2Ph (99%) was used instead of a 1 M THF solution of PMe3
(67% yield).
Method B. To a solution of compound 3 (30 mg, 0.022 mmol) in
0.6 mL of CD2Cl2, in an NMR tube capped with a rubber septum, was
added 1 equiv (3.1 µL, 0.022 mmol) of PMe2Ph leading to an immediate
color change to yellow. The NMR spectra showed that the product
was compound 8. Anal. Calcd for Ir2SP5F3O5C62H58: C, 49.27; H, 3.87.
Found: C, 49.08; H, 3.71.
of catalytically relevant molecules, with a view toward modeling
substrate activation by adjacent metals in heterogeneous systems.
We begin by considering one of the simplest multinuclear
organometallic systems, the binuclear methyl complex. In such
complexes the methyl group is capable of displaying several
coordination modes.8 For late-metal systems, three arrangements
are possible as diagrammed below. In C the methyl group is
terminally bound to one metal and to a first approximation
resembles a mononuclear alkyl species. When methyl groups
bridge two late metals, two binding modes have been identified.
In structure D the methyl group bridges the metals sym-
metrically, giving rise to a three-center metal-carbon-metal
interaction,9 whereas in structure E, the methyl group bridges
in an unsymmetrical manner, η1 bound to one metal via carbon
and η2 bound to the adjacent metal via a C-H bond.10 This
latter agostic interaction has important implications relating to
the activation of the C-H bond, as noted earlier. In this article
we attempt to determine the factors influencing the coordination
modes of methyl groups in a binuclear Ir complex, with
emphasis on factors leading to C-H bond activation.
Experimental Section
General Comments. All solvents were dried (using appropriate
drying agents), distilled before use, and stored under nitrogen. Deu-
terated solvents used for NMR experiments were degassed and stored
under argon over molecular sieves. Reactions were carried out at room
temperature (unless otherwise stated) using standard Schlenk proce-
dures, and compounds, which were isolated as solids, were purified
by recrystallization. A flow rate of ca. 0.2 mL s-1 was used for all
reactions that involved purging a solution with a gas. Prepurified argon
and hydrogen were purchased from Linde, carbon monoxide and sulfur
dioxide were purchased from Matheson, and carbon-13-enriched CO
(99%) was supplied by Isotec Inc. All gases were used as received.
Ammonium hexachloroiridate(IV) was purchased from Vancouver
Island Precious Metals, and the chemicals, methyl triflate, d3-methyl
triflate, 13C-methyl triflate, tert-butyl isocyanide, 2,3,6,6-tetramethylpi-
peridine, all phosphines, and triphenyl phosphite were purchased from
Aldrich. The compound [Ir2(CO)3(dppm)2] (1) was prepared as previ-
ously reported.11
Standard NMR spectra were recorded on a Bruker AM-400
1
spectrometer operating at 400.1 MHz for H, 161.9 MHz for 31P, and
(e) [Ir2(H)(CO)2(PPh3)(µ-CH2)(dppm)2][CF3SO3] (8). Method A.
Compound 2 (100 mg, 0.071 mmol) was dissolved in 20 mL of THF,
to which was added 1 equiv of PPh3 (19 mg, 0.071 mmol) dissolved in
10 mL of THF. The solution color changed from orange to yellow.
After stirring for 2 h, the solvent was removed under vacuum, and the
yellow solid recrystallized from CH2Cl2/Et2O.
Method B. Compound 3 (20 mg, 0.015 mmol) and PPh3 (4.3 mg,
0.016 mmol) were placed in an NMR tube to which 0.6 mL of CD2Cl2
was added resulting in an immediate color change to orange. The NMR
spectra showed that the product was compound 7.
100.6 MHz for 13C spectra. The solid-state 13C{1H} NMR spectrum
was recorded on a Bruker AMR-300 spectrometer [with a magic-angle
spinning (MAS) accessory] operating at 75.5 MHz. The 13C{1H}{31P}
NMR spectra were obtained on a Bruker WH-200 spectrometer
operating at 50.3 MHz. All 13C NMR spectra were obtained using 13CO-
or 13CH3-enriched samples (the latter obtained from 13C-methyl triflate).
Infrared spectra were obtained on a Nicolet 7199 Fourier transform
interferometer, either as solids in Nujol, as CH2Cl2 casts on KBr plates,
(8) See for example: Schwartz, D. J.; Ball, G. E.; Andersen, R. A. J.
Am. Chem. Soc. 1995, 117, 6027.
(9) See for example: (a) Schmidt, G. F.; Muetterties, E. L.; Beno, M.
A.; Williams, J. M. Proc. Natl. Acad. Sci. U.S.A. 1981, 78, 1318. (b) Kulzick,
M. A.; Price, R. T.; Andersen, R. A.; Muetterties, E. L. J. Organomet.
Chem. 1987, 333, 105. (c) Reinking, M. K.; Fanwick, P. E.; Kubiak, C. P.
Angew. Chem., Int. Ed. Engl. 1989, 28, 1377.
(10) See for example: (a) Dawkins, G. M.; Green, M.; Orpen, A. G.;
Stone, F. G. A. J. Chem. Soc. Chem. Commun. 1982, 41. (b) Calvert, R.
B.; Shapley, J. R. J. Am. Chem. Soc. 1978, 100, 7726. (c) Brookhart, M.;
Green, M. L. H.; Wong, L.-L. Prog. Inorg. Chem. 1988, 36, 1; and
references therein.
(f) [Ir2(H)(CO)2(P(OPh)3)(µ-CH2)(dppm)2][CF3SO3] (9a and 9b).
The procedure was the same as that described for 6 except that P(OPh)3
(97%) was used instead of a 1 M THF solution of PMe3, and the mixture
was stirred for 3 h instead of 2 h. Two isomers were formed (see Results
and Characterization) (60% total yield). Anal. Calcd for Ir2SP5F3O8-
C72H62: C, 51.36; H, 3.71. Found: C, 51.40; H, 3.82.
(g) [Ir2(H)(CO)2(tBuNC)(µ-CH2)(dppm)2][CF3SO3] (10). Method
A. To a solution of compound 2 (100 mg, 0.071 mmol) in 5 mL of
t
CH2Cl2, was added 1 equiv of BuNC (8.1 µL, 0.071 mmol), causing
(11) Sutherland, B. R.; Cowie, M. Organometallics 1985, 4, 1637.
an immediate color change from orange to yellow. The mixture was

Table 1. Spectroscopic Data for the Compounds
compounda,b
δ(31P{1H})c
δ(1H)d,e
δ(13C{1H})d
IR, cm-1 f,g
[Ir2(H)(CO)3(µ-CH2)(dppm)2][CF3SO3] (2)h
-0.9(m), -12.5(m) 5.19 (m, 2H), 4.60 (m, 2H),
177.4(t, 1C), 177.0 (t, 1C), 162.0 (t, 1C),
49.4 (b, 1C)
2018 (vs), 2000 (vs), 1970 (vs)
3.59 (m, 1JC-H ) 140.6 Hz, 2H),c
-11.35 (t, 1H)
[Ir2(CH3)(CO)2(dppm)2][CF3SO3] (3)
22.1(s)
4.45 (m, 2H), 3.79 (m, 2H),
0.58 (q, 3JP-H ) 4.1 Hz, 3H,
1JC-H ) 127 Hz)j
171.4 (b, 2C), 17.5 (q, 2JC-P ) 3.9 Hz, 1C) 1964 (s),i 1886 (w)
187.8 (m, 1C),k 167.0 (m, 1C)k
194.2 (t, 2C), 182.9 (t, 2C)
225.1 (t, 1C), 195.6 (b, 2C), 179.9 (t, 2C)
1974 (vs), 1832 (s)
[Ir2(CH3)(CO)4(dppm)2][CF3SO3] (4)
[Ir2(COCH3)(CO)4(dppm)2][CF3SO3] (5)
-8.9(m), -16.0(m) 4.59 (m, 4H), 0.38 (t, 3H)
-9.6(m), -25.5(m) 4.61 (b, 4H), 0.63 (s, 3H)
2038 (m),i 1977 (b, vs), 1956 (vs)
2019 (m),i 1988 (sh),
1959 (s), 1622 (w)
[Ir2(H)(CO)2(PMe3)(µ-CH2)(dppm)2][CF3SO3] (6)
-7.7(m), -11.3(m), 4.05 (m, 2H), 3.45 (m, 2H),
183.0 (t, 1C), 180.1 (b, 1C)
1961 (vs), 1950 (vs)
-79.2(b)
3.05 (m, 2H), 0.75 (d, 9H),
-12.35 (dm, 3JP-H ) 21.1, 4.8,
2JP-H ) 13.5 Hz, 1H)
[Ir2(H)(CO)2(PMe2Ph)(µ-CH2)(dppm)2][CF3SO3 (7)
[Ir2(H)(CO)2(PPh3)(µ-CH2)(dppm)2][CF3SO3] (8)
5.7(m), -12.1(m),
-71.6(b)
3.82 (m, 2H), 3.35 (m, 2H),
3.30 (m, 2H), 1.12 (d, 6H),
-12.2 (dm, 3JP-H ) 21.6 Hz, 1H)
182.6 (t, 1C), 180.6 (b, 1C)
183.4 (t, 1C), 180.5 (b, 1C)
1963 (vs), 1952 (s)
1975 (vs), 1948 (vs)
-4.1(m), -13.7(m), 3.84 (m, 2H), 3.71 (m, 2H),
-33.9(b)
3.13 (M, 2H), -13.25
(dm, 3JP-H ) 22.3 Hz, 1H)
5.2-3.7 (m), -12.69
[Ir2(H)(CO)2(P(OPh)3)(µ-CH2)(dppm)2][CF3SO3] (9a) 43.9(m), -4.1(m),
180.4 (b, 1C), 179.7 (t, 1C), 167.1 (dt, 1C) 1985 (vs), 1975 (vs),
1951 (s)
-12.6(m)
[Ir2(H)(CO)2(P(OPh)3)(µ-CH2)(dppm)2][CF3SO3] (9b) 82.7(t), -5.6(m),
-8.1(m)
(dm, 3JP-H ) 37.3 Hz, 1H)
5.2-3.7 (m), -11.71
dt, 2JP-H ) 17.7 Hz, 1H)
3.92 (m, 2H), 3.58 (m, 2H),
3.57 (m, 2H), 0.60 (s, 9H),
-11.46 (t, 1H)
[Ir2(H)(CO)2(CNtBu)(µ-CH2)(dppm)2][CF3SO3] (10)
-6.6(m), -8.5(m)
179.2 (b, 1C), 179.1 (t, 1C)
2163 (s),l 1975 (s), 1970 (s)
[Ir2(H)(CO)2(µ-CH2)(µ-SO2)(dppm)2][CF3SO3] (11)
-1.0(s)
4.50 (m, 2H), 3.94 (m, 2H),
170.8 (t, 2C), 36.1 (q, 2JC-P ) 4.5 Hz, 1C) 2044 (vs), 1997 (vs)
177.9 (t, 1C), 177.8 (t, 1C), 160.2 (t, 1C)
2042 (s), 2010 (vs), 1987 (s), 1070 (m)m
-2.15 (b, 1JC-H ) 98 Hz, 3H)j
[Ir2(CH3)(CO)3(µ-SO2)(dppm)2][CF3SO3] (12)
[Ir2(CH3CO)(CO)3(µ-SO2)(dppm)2][CF3SO3] (13)
-0.4(m), -14.7(m) 4.9 (m, 2H), 3.40 (m, 2H), 0.01 (t, 3H)
-5.0(m), -13.8(m) 5.03 (m, 2H), 3.18 (m, 2H), 1.28 (s, 3H)
2013 (s), 1990 (vs), 1646 (m)
a NMR abbreviations: s ) singlet, d ) doublet, t ) triplet, m ) multiplet, q ) quintet, dm ) doublet of multiplets, b ) broad. b NMR data at 25 °C in CD2Cl2 unless otherwise stated. c 31P{1H} chemical
shifts are referenced vs external 85% H3PO4. d 1H and 13C vs external TMS. e Chemical shifts for the phenyl hydrogens are not given in the 1H NMR data. f IR abbreviations [ν(CO) unless otherwise stated]:
j 1
vs ) very strong, s ) strong, m ) medium, w ) weak, sh ) shoulder, b ) broad. g Nujol mull or CH2Cl2 cast unless otherwise stated. h Data at -40 °C. i CH2Cl2 solution.
13C-labeled compounds. k Solid-state NMR. l ν(CN). m ν(SO).
J
coupling values for
C-H

Metal-Metal CooperatiVity and C-H Bond CleaVage
Table 2. Crystallographic Data
J. Am. Chem. Soc., Vol. 121, No. 15, 1999 3669
compd
[Ir2(CH3)(CO)(µ-CO)(dppm)2][CF3SO3] (3) [Ir2(H)(PMe3)(CO)2(µ-CH2)(dppm)2]- [Ir2(H)(CO)2(µ-CH2)(µ-SO2)(dppm)2]-
[CF3SO3]‚2CH2Cl2 (6)
C59H60Cl4F3Ir2O5P5S
1619.18
[CF3SO3]‚CH2Cl2 (11)a
C54.875H48.625Cl2.125F3Ir2O7P4S2
1524.79
a
formula
formula wt
color
C54H47F3Ir2O5P4S
1373.26
dark red
yellow
dark red
crystal system
space group
unit cell parameters
a (Å)
b (Å)
c (Å)
monoclinic
triclinic
monoclinic
P21/c (No. 14)
P21/n (an alternate setting of P21/c [No. 14]) P1h (No. 2)
13.8372 (14)
15.569 (2)
23.629 (2)
90.0
96.413 (9)
90.0
5058.5 (10)
4
0.0405
11.910 (2)
12.8934 (15)
20.532 (3)
91.256 (10)
95.308 (13)
102.601 (12)
3060.9 (8)
2
13.974 (2)
18.422 (3)
22.264 (3)
90.0
105.25 (1)
90.0
5532.6 (12)
4
0.0505
R (deg)
â (deg)
γ (deg)
V (Å3)
Z
Rb
0.0598
[Fo2 > 2σ(Fo )]
2
b
wR2
0.0815
0.1798
0.1298
[Fo2 > -3σ(Fo )]
2
GOF(S)c
1.003[Fo2 g -3σ(Fo )]
1.082[Fo2 g -3σ(Fo )]
0.993[Fo2 g -3σ(Fo )]
2
2
2
a Compound 11 cocrystallized with 12.5% [Ir2(CO)2(µ-Cl)(µ-SO2)(dppm)2][CF3SO3]‚CH2Cl2. b R1 ) ∑||Fo| - |Fc||/∑|Fo|; wR2 ) [∑w(Fo
-
2
Fc2)2/∑w(Fo )]1/2
.
c S ) [∑w(Fo2 - Fc2)2/(n - p)]1/2 (n ) number of data; p ) number of parameters varied; w ) [σ2(Fo ) + (a0P)2 + a1P]-1, where
4
2
2
P ) [Max(Fo , 0) + 2Fc2]/3), and a0 and a1 are suggested by the least-squares program for each structure (for 3, a0 ) 0.0216, a1 ) 2.2944; for 6,
a0 ) 0.1278, a1 )10.1840; for 11, a0 ) 0.0554, a1 ) 0).
stirred for 2 h by which time the color was green. The solvent volume
was reduced to about 2 mL, and the green product was precipitated
and washed with Et2O. After drying under an argon stream, the solid
was recrystallized from CH2Cl2/Et2O (70% yield).
Method B. Compound 3 was substituted for compound 2 in the
above-mentioned procedure yielding compound 10 essentially quanti-
tatively (by NMR).
1H NMR Magnetization Transfer Experiments on Compound
2. Magnetization transfer between the hydride and the methylene-ligand
protons was examined by selective inversion. The hydride signal was
selectively inverted by means of a DANTE12 pulse sequence. Typically,
after a delay of 1.5 s, a series of eighteen 5-µs pulses, each separated
by a 550-µs delay was used. The relaxation of this signal after the
inversion was monitored using a 90° nonselective observation pulse,
applied at variable delays of between 0.001 and 2.5 s, after the selective
pulse.
Elemental analyses for compound 10 always gave low values for
%C and %H because of decomposition in air.
(h) [Ir2(H)(CO)2(µ-CH2)(µ-SO2)(dppm)2][CF3SO3] (11). Sulfur
dioxide gas was passed through a solution of compound 3 (60 mg,
0.044 mmol) in 5 mL of CH2Cl2 for 1 min, leading to an immediate
color change of the solution from red to dark red. After stirring for 10
min, the solution was concentrated to about 2 mL under an argon stream
and the dark red-brown solid was precipitated and washed twice with
Et2O. The solid was dried under an argon stream and then under vacuum
(72% yield). Anal. Calcd for Ir2S2P4F3O7C54H47: C, 45.12; H, 3.30; S,
4.46. Found: C, 45.40; H, 3.16; S, 4.19.
(i) [Ir2(CH3)(CO)3(µ-SO2)(dppm)2][CF3SO3] (12). Sulfur dioxide
gas was passed through a solution of 3 (30 mg, 0.022 mmol) in 5 mL
of CH2Cl2 for about 2 min, resulting in a color change from red to
dark red indicating the formation of 11. CO was then passed through
the solution for 5 min and the color changed from dark red to light
yellow, then to green. The solution was stirred under a static atmosphere
of the gas for 15 min, after which time the green solid was precipitated
with Et2O. The solid was collected, washed with Et2O and dried under
an argon stream. It was then recrystallized from CH2Cl2/Et2O (73%
yield). Anal. Calcd. for Ir2S2P4F3O8C55H47: C, 45.19; H, 3.23; S, 4.37.
Found: C, 44.54; H, 3.21; S, 4.77.
(j) Reaction of Compound 3 with CO. To a solution of compound
3 (30 mg, 0.022 mmol) in 0.6 mL of CD2Cl2, in an NMR tube capped
with a rubber septum, was added 2 mL of CO, by means of a gas-tight
syringe, leading to an immediate color change from red to yellow. The
NMR spectra showed that the reaction mixture contained [Ir2(CH3)-
(CO)4(dppm)2][CF3SO3 (4) and [Ir2(COCH3)(CO)4(dppm)2][CF3SO3] (5)
in a typical ratio of about 1.0:0.2. However, the products were not
isolated as solids, because they are unstable in the absence of a CO
atmosphere.
(k) Reaction of Compound 12 with CO. The procedure was the
same as the reaction of compound 3 with CO. The NMR spectrum of
the reaction mixture showed that it contained unreacted compound 12
and [Ir2(CH3CO)(CO)3(µ-SO2)(dppm)2][CF3SO3] (13), in a typical ratio
of 3:1. However the acyl product was not isolated as a solid, because
it reverted to 12 in the absence of a CO atmosphere.
Nonselective inversion recovery experiments were performed initially
to measure the relaxation time (T1) for the hydride as well as that for
the methylene ligand protons. The nonselective 90° pulse width for
1H, with an attenuator set at 10 dB, was 44.75 µs. This pulse width
was calibrated at the beginning of the experiment. NMR probe
temperatures were measured with an independently calibrated thermo-
couple immersed in the appropriate volume of toluene in an NMR tube
before and after the experiments.
X-ray Data Collection. For each of compounds 3, 6, and 11, crystals
suitable for X-ray diffraction were grown via slow diffusion of diethyl
ether into a concentrated CH2Cl2 solution of the compound. Crystals
were mounted and flame-sealed in glass capillaries under solvent vapor
to minimize decomposition or deterioration caused by solvent loss. Data
were collected at -50 °C on an Enraf-Nonius CAD4 diffractometer
using graphite-monochromated Mo KR radiation. Unit cell parameters
and space group assignments were obtained as described below. For
each compound, three reflections were chosen as intensity standards
and were remeasured every 120 min of X-ray exposure time; in no
case was decay evident. Absorption corrections were applied to the
data as described below. Crystal parameters and final agreement factors
are summarized in Table 2.
Unit cell parameters for compounds 3, 6, and 11 were obtained from
a least-squares refinement of 24 reflections in the approximate range
20° < 2θ < 24°. For compounds 3 and 11 the cell parameters and the
systematic absences defined the space groups as P21/n and P21/c,
respectively, whereas for 6 the cell parameters, the lack of absences,
and the diffraction symmetry suggested the space group P1 or P1h, the
latter of which was established by successful refinement of the structure.
Absorption corrections to 3 and 6 were applied by the method of Walker
and Stuart,13 whereas for 11 the crystal faces were indexed and
measured, and a Gaussian integration carried out.
(12) Bodenhausen, G.; Freeman, R.; Morris, G. A. J. Magn. Reson. 1976,
23, 171.
(13) Walker, N.; Stuart, D. Acta Crystallogr., Sect. A 1983, 39, 158.

3670 J. Am. Chem. Soc., Vol. 121, No. 15, 1999
Torkelson et al.
Structure Solution and Refinement. For each structure, the
positions of the iridium and phosphorus atoms were found using the
direct-methods program SHELXS-86,14 and the remaining atoms were
found using a succession of least-squares and difference Fourier maps.
Refinement of each structure proceeded using the program SHELXL-
93.15 Hydrogen atom positions (except for hydride ligands as noted
below) were calculated by assuming idealized sp2 or sp3 geometries
about their attached carbon atoms (as appropriate), and were given
thermal parameters 120% of the equivalent isotropic displacement
parameters of their attached carbons. Further details of structure
refinement (other than described below) and final residual indices may
be found in Table 2.
For [Ir2(CH3)(CO)(µ-CO)(dppm)2][CF3SO3] (3) electron density
maps in the equatorial plane, essentially perpendicular to the Ir-P
vectors, indicated similar coordination environments for both iridium
atoms. Although the bridging carbonyl group was well behaved, each
Ir apparently had a terminal CO in which the peaks corresponding to
the oxygens were of low intensity. This suggested a disorder involving
superposition of the terminal carbonyl and methyl ligands, resulting
from the two orientations of the complex as diagrammed below (dppm
groups, above and below the plane of the paper, omitted for clarity).
Figure 1. A view of the equatorial plane of compound 11 (87.5%
occupancy), showing the disorder involving the superimposed 12.5%
occupancy [Ir2(CO)2(µ-Cl)(µ-SO2)(dppm)2]+. The major species is
shown with solid bonds, whereas the minor species having Cl(1)
replacing C(3)H2 and H(1), and C(1′)O(1′) replacing C(1)O(1) is shown
with these changes designated by open bonds.
Disorder of the carbonyl group attached to Ir(1) was suggested by
the elongation of the thermal ellipsoids of C(1) and O(1) in a direction
perpendicular to the Ir(1)-C(1) bond axis. In addition, the Ir(2)-Ir-
(1)-C(1) angle observed for 11 (ca. 144°) was substantially different
from that previously determined for the chloro-bridged complex (ca.
167°) indicating that these carbonyl groups for the two disordered
molecules would not be superimposed. This carbonyl group was
therefore disordered over two sites having occupancy factors of 0.875
and 0.125, corresponding to those of the bridging methylene and chloro
groups, respectively. The placement of C(1′)O(1′) was initially set at
the coordinates previously found for the pure chloro-bridged species,
and was refined as a rigid group having Ir(1)-C(1′), C(1′)-O(1′), and
Ir(1)-O(1′) distances approximating those within the Ir(2)-C(2)-O(2)
unit. Atom C(1′) was given a fixed isotropic thermal parameter roughly
equal to Ueq for C(2) whereas O(1′) was refined isotropically. Carbonyl
C(1)O(1) having the major occupancy was refined anisotropically
without restriction. A drawing of the equatorial plane of the disordered
structure (dppm groups omitted) is shown in Figure 1. Atoms Ir(1),
Ir(2), C(2), O(2), S(1), O(3), and O(4) are common to both disordered
molecules and C(1′), O(1′), and Cl(1) corresponding to the chloro-
bridged species and C(3)H2, C(1), O(1), and H(1) belonging to the
methylene-bridged complex (11). Superposition of the atom positions
of complex 11 with the pure chloro-bridged species shows that, apart
from the atoms involved in the disorder, all atoms superimpose almost
exactly.
Subsequent difference Fourier maps showed that the disordered
carbonyl and methyl carbons were not exactly superimposed but were
slightly offset from each other, and these two half-occupancy positions
on each Ir refined well. The atoms corresponding to the disordered
CO and CH3 groups were refined isotropically with an occupancy factor
of 0.5.
Location of all atoms in [Ir2(H)(CO)2(PMe3)(µ-CH2)(dppm)2][CF3-
SO3]‚2CH2Cl2 (6) proceeded smoothly, and even the hydride ligand
H(2) was located and refined with a fixed Ir(2)-H(2) distance of 1.70
Å.
Although all the nonhydrogen atoms of [Ir2(H)(CO)2(µ-CH2)(µ-
SO2)(dppm)2][CF3SO3]‚CH2Cl2 (11) were located, the carbon atom of
the bridging methylene group [C(3)] did not behave well in least-squares
refinements. A difference Fourier map at this stage located another
peak near C(3), but at approximately 2.5 Å from each metal. A 31P-
{1H} NMR spectrum obtained on a few of the single crystals showed
about 10-15% of the previously characterized [Ir2(CO)2(µ-Cl)(µ-
SO2)(dppm)2][CF3SO3],16 which had apparently resulted from reaction
of 11 with the chlorinated solvent. All subsequent attempts to obtain
suitable crystals of 11 without the chloro-bridged impurity failed;
attempts to grow crystals in nonchlorinated solvents gave poor quality
crystals. Inclusion of a chlorine atom [Cl(1)] in the least-squares
refinement with an occupancy factor of 0.125 [and concomitant
reduction of the occupancy factor for C(3) to 0.875] yielded the most
satisfactory results. The hydride ligand H(1) was not located from a
difference Fourier map, but its position was idealized at 1.70 Å from
Ir(1) in a position that minimized contacts with the surrounding ortho-
hydrogen atoms of the dppm groups. Unfortunately, owing to the
disorder in which the fractional Cl atom was essentially superimposed
on the CH2 group, the methylene hydrogens could not be located.
Theoretical Methods and Software. We have used the Gaussian
9417 and Mulliken 2.0.118 software suites running on the SGI Indigo II
(G94, Rev. D.1) and IBM RS/6000 (G94, Rev. D.2) workstations at
Calvin College. Selected Gaussian 94 calculations were performed on
the SGI Power Challenge Array at the National Center for Supercom-
puting Applications (NCSA). Details regarding methodology and basis
sets may be found in the Gaussian User’s Reference.19 In Gaussian 94
we used density functional theory (DFT) with the B3LYP methodol-
(17) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T. A.; Petersson,
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.;
Nanyakkara, 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, Revisions D.1, D.3,
and D.4; Gaussian, Inc.: Pittsburgh, PA, 1995.
(18) Lengsfield, B. H., III; Horn, H.; Swoe, W. C.; Carter, J. T.; McLean,
A. D.; Carter, J. T.; Replogle, E. S.; Rice, J. E.; Barnes, L. A.; Maluendes,
S. A.; Lie, G. C.; Gutowski, M.; Rudge, W. E.; Sauer, S. P. A.; Lindh, R.;
Andersson, K.; Chevalier, T. S.; Widmark, P.-O.; Bouzida, D.; Nesbet, R.;
Singh, K.; Gillan, C. J.; Carnevali, P.; Liu, B.; Pacansky, J. Mulliken, 2.0
ed.; IBM Research Division, Scientific & Technical Application Software,
Almaden Research Center: San Jose, CA, 1996.
(14) Sheldrick, G. M. Acta Crystallogr., Sect. A 1990, 46, 467-473.
(15) Sheldrick, G. M. SHELXL-93, program for crystal structure deter-
mination; University of Go¨ttingen: Germany, 1993. Refinement on Fo2 for
2
all reflections except for those having Fo2 < -3σ(Fo ). Weighted R-factors
2
wR2 and all goodnesses-of-fit S are based on Fo ; conventional R-factors
2
R1 are based on Fo, with Fo set to zero for negative Fo . The observed
2
criterion of Fo2 > 2σ(Fo ) is used only for calculating R1, and is not relevant
2
to the choice of reflections for refinement. R-factors based on Fo are
statistically about twice as large as those based on Fo, and R-factors based
on ALL data will be even larger.
(16) (a) Torkelson, J. R.; Cowie, M. Unpublished results, 1994. (b)
Crystal data for [Ir2(CO)2)(µ-Cl)(µ-SO2)(dppm)2][CF3SO3]‚CH2Cl2: mono-
clinic; P21/c; a ) 14.053(3) Å, b ) 18.485(4) Å, c ) 22.544(6) Å, â )
105.89(1)°; V ) 5633(2) Å3; Z ) 4; Dcalc ) 1.819 g cm-3; µ(Mo KR) )
2
2
51.11 cm-1; T ) 22 °C; 6489 independent observations Fo g 2σ (Fo );
(19) Frisch, M. J.; Frisch, A.; Foresman, J. B. Gaussian 94 User’s
Reference; Gaussian, Inc.: Pittsburgh, PA, 1996.
676 variables; R1(F) ) 0.0532 (observed data), wR2(F2) ) 0.1273 (all data).

Metal-Metal CooperatiVity and C-H Bond CleaVage
J. Am. Chem. Soc., Vol. 121, No. 15, 1999 3671
ogy.20 The basis set made use of effective core potentials and is
designated LANL2DZ in the Gaussian library.21-24 This basis set is of
double-ú quality but does not include polarization functions. Because
of the long Ir-Ir bond length that we obtained with this basis set (see
results), and the long bond lengths surrounding the sulfur atom, on the
calculations for [Ir2(H)(CO)2(µ-CH2)(µ-SO2)(dhpm)2]+ (dhpm ) H2-
PCH2PH2), we completed additional studies using a modified set of
valence p orbitals on the iridium atom as specified by Hall and co-
workers.25 We also added a set of polarization functions to the sulfur
atom with an exponent obtained from the database of Feller and co-
workers.26 Minima on the potential energy surface were located using
standard methodologies available in Gaussian 94.
Scheme 1
The Mulliken software was used on our IBM RS/6000 workstation.
We used DFT and made use of the effective core potential basis library
called ECP•QD95*. This basis set is of double-ú quality and includes
polarization functions on the nontransition metal atoms. The effective
core is of type ECP•CEP-DZ.27-29 The B3LYP results within Mulliken
use the local correlation functional of Perdew and Wang30 instead of
the Vosko et al.31 which is used in Gaussian 94.19
Within both Gaussian 94 and Mulliken, full-geometry optimizations
were completed on each of the cations for which we report data. No
theoretical vibrational frequencies were computed for any of these
species. The bis(diphenylphosphino)methane (dppm) and bis(dimeth-
ylphosphino)methane (dmpm) ligands were modeled by replacing the
organic substituents on the phosphorus atom with hydrogen atoms to
form the bis(dihydrogenphosphino)methane (dhpm) ligand. In all our
work we began with the Ir2(dhpm)2 portion of each molecule in a
boatlike configuration, and it retained that configuration throughout
the geometry optimization procedure.
In the discussion of the theoretical results (vide infra) we refer to
energy differences between different isomers. These are simply
differences in internal energy at 0 K. We have not computed vibrational
frequencies, so we cannot include any zero-point energy. However,
because we are comparing only differences in isomeric energies, the
differences in zero-point energy will be miniscule. Furthermore, we
will simply compare our differences in internal energy to experimental
enthalpy changes.
obtained for dirhodium34 and a product containing a methyl
group terminally bound to Ir was obtained for RhIr.35 The
reaction with the diiridium precursor also does not proceed
strictly according to expectation, although, as will be explained,
a methyl-bridged species is no doubt involved. Instead of the
targeted methyl-bridged product, the methylene-bridged hydrido
species [Ir2(H)(CO)3(µ-CH2)(dppm)2][CF3SO3] (2) results from
C-H bond cleavage of the methyl group, as shown in Scheme
1.36 This result parallels an earlier study in which methyl triflate
addition to [Cp*Ir(µ-CO)]2 yielded the methylene-bridged
hydride [Cp2*Ir2(CO)2(µ-CH2)(µ-H)][CF3SO3].37
Results and Characterization of Compounds
Earlier work in this research group and by Eisenberg and
co-workers had shown that protonation of the complexes [Ir2-
(CO)3(dppm)2] (1), [Rh2(CO)3(dppm)2], and their heterobi-
nuclear analogue, [RhIr(CO)3(dppm)2], yielded the respective
hydrido-bridged complexes, [Ir2(CO)2(µ-H)(µ-CO)(dppm)2]-
[BF4],11 [Rh2(CO)2(µ-H)(µ-CO)(dppm)2][BF4],32 and [RhIr-
(CO)3(µ-H)(dppm)2][BF4],33 by apparent electrophilic attack at
the metal-metal bond. An obvious extension of these proto-
Compound 2 was characterized by extensive spectroscopic
investigations. 1H NMR experiments with selective and broad-
band 31P-decoupling, carried out at -40 °C, clearly differentiate
the methylene group bridging the metals (δ 4.60) from the two
dppm methylene resonances (see Table 1). The Ir2-bridging
methylene appears as an apparent quintet in the 31P-coupled
1H NMR spectrum and simplifies to a singlet upon broad-band
decoupling and to a triplet upon selective decoupling of each
+
nations was to substitute H+ by CH3 [in the form of methyl
triflate (MeOTf), for example] in attempts to obtain methyl-
bridged analogues. This expectation was not realized for the
Rh-containing complexes, for which an acyl-bridged species was
1
31P resonance in turn, whereas the two H resonances for the
dppm methylenes simplify to the expected AB quartet upon
broad-band 31P-decoupling. The hydride resonance, which
appears as a triplet in the 1H NMR spectrum at -40 °C,
collapses into a singlet upon selectively decoupling the lower
field 31P resonance but remains unchanged on decoupling the
higher field 31P resonance, showing that it is terminally bound
to one of the metals, and is therefore coupling to only the pair
of 31P nuclei on this metal.
The 13C{1H} NMR spectrum of 2 shows triplets at δ 177.4,
177.0, and 162.0, corresponding to three terminally bound
carbonyl groups; and selective 31P-decoupling experiments
(20) Becke, A. D. J. Chem. Phys. 1993, 93, 5648-5652.
(21) Dunning, T. H., Jr.; Hay, P. J. In Modern Theoretical Chemistry;
Schaefer, H. F., III, Ed.; Plenum: New York, 1976; pp 1-28.
(22) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270.
(23) Wadt, W. R.; Hay, P. J. J. Chem. Phys. 1985, 82, 284.
(24) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299.
(25) Couty, M.; Hall, M. B. J. Comput. Chem. 1996, 17, 1359-1370.
(26) Feller, D.; Schuchardt, K.; Jones, D. Extensible Computational
Chemistry EnVironment Basis Set Database, 1.0 ed.; Feller, D., Schuchardt,
K., Jones, D., Eds.; Molecular Sciences Computing Facility, Environmental
and Molecular Sciences Laboratory, 1998.
(27) Stevens, W.; Basch, H.; Krauss, M. J. Chem. Phys. 1984, 81, 6026.
(28) Stevens, W.; Jasien, P. G.; Krauss, M.; Basch, H. Can. J. Chem.
1992, 70, 612.
(29) Cundari, T. R.; Stevens, W. J. J. Chem. Phys. 1993, 98, 5555.
(30) Perdew, J. P.; Wang, Y. Phys. ReV. B. 1992, 45, 1324.
(31) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1980, 58, 1200.
(32) (a) Kubiak, C. P.; Woodcock, C.; Eisenberg, R. Inorg. Chem. 1982,
21, 2119. (b) Kubiak, C. P.; Eisenberg, R. J. Am. Chem. Soc. 1980, 102,
3637.
(34) Shafiq, F.; Kramarz, K. W.; Eisenberg, R. Inorg. Chim. Acta 1993,
213, 111.
(35) Antwi-Nsiah, F. H.; Oke, O.; Cowie, M. Organometallics 1996,
15, 1042.
(36) A preliminary report of this has appeared: Antwi-Nsiah, F.; Cowie,
M. Organometallics 1992, 11, 3157.
(37) Heinekey, D. M.; Michel, S. T.; Schulte, G. K. Organometallics
1989, 8, 1241.
(33) McDonald, R.; Cowie, M. Inorg. Chem. 1990, 29, 1564.

3672 J. Am. Chem. Soc., Vol. 121, No. 15, 1999
Torkelson et al.
establish that the high- and low-field carbonyls are bound to
one iridium center whereas the third is bound to the other
iridium, to which the hydride is also bound. The 13C{1H}
resonance for the bridging methylene group, obtained on a
sample prepared using 13C-methyl triflate, appears as a broad
singlet at δ 49.9; the carbon-hydrogen coupling constant (1JC-H
and the 13C{1H} NMR spectrum shows one singlet carbonyl
resonance at δ 171.4 and a sharp quintet at δ 17.5 (2JP-C ) 4.5
Hz) for the methyl group. However, at low temperatures the
spectral parameters are more in keeping with an unsymmetrical
species having a terminally bound methyl group. Thus at -80
°C the 31P{1H} NMR spectrum shows two resonances for two
chemically distinct phosphorus environments (by virtue of the
chemically inequivalent metal centers), and the 13C{1H} NMR
spectrum displays two carbonyl resonances at δ 182.5 and 162.1.
In the solid-state 13C NMR spectrum the carbonyl resonances
(δ 187.8 and 167.0) are close to those observed in solution at
-80 °C, whereas the IR spectrum (Nujol mull) shows carbonyl
bands at 1974 and 1832 cm-1, suggesting one terminal and one
bridging carbonyl group. Compound 3 is therefore formulated
as [Ir2(CH3)(CO)(µ-CO)(dppm)2][CF3SO3] in the solid state,
having a structure much like that of Eisenberg’s Rh2 analogue.34
Upon dissolving compound 3, the strong IR band at 1832
cm-1 disappears and is replaced by a Very weak band at ca.
1886 cm-1, and the stretch caused by the terminal carbonyl
moves from 1974 to 1964 cm-1. It would appear that in solution,
the dominant species is not the same as that found in the solid
state, and instead has the structure shown below for 3a. The
1
) 140.6 Hz) was obtained from the H NMR spectrum of the
13C-labeled sample.
Although the high-field 1H resonance could result from either
a hydride ligand or an agostic methyl interaction, it can be
unambiguously ascribed to a hydride ligand because no 13C-
1H coupling is observed. In addition, agostic methyl groups are
highly fluxional; for mononuclear complexes we know of no
case in which such groups are static on the NMR time scale,
even at low temperature,38 whereas for polynuclear species we
know of only two examples in which a static structure (at
between -80° and -90 °C) was observed in solution for a
bridging agostic methyl group.39 The observation of separate
hydride and methylene signals at the relatively high temperature
of -40 °C is more consistent with the formulation shown for 2
than for a bridged agostic methyl interaction.
At temperatures greater than -40 °C the 1H NMR resonances
for the methylene and the hydride ligands of 2 broaden,
suggesting an exchange between the two, which has been
confirmed by a 1H NMR, spin-saturation-transfer experiment.
Presumably the exchange between the hydride and the meth-
ylene hydrogens occurs through an intermediate methyl com-
plex, not unlike the targeted methyl-bridged species. The rates
of exchange between the bridging methylene protons and the
hydride ligand were determined at various temperatures between
1
-29.7 °C and -56.4 °C, by selective inversion recovery H
two terminal carbonyls in this compound absorb at ap-
proximately the same frequency, giving rise to the broad band
at 1964 cm-1. Although this structure type has not, to our
knowledge, been observed in dppm-bridged complexes of Rh
or Ir, related structures have been observed or proposed for
closely related PNP-bridged complexes of rhodium (PNP )
(RO)2PN(Et)P(OR)2)42 and dppm complexes of platinum and
palladium.43 In addition, as will be addressed later, DFT
calculations have shown that structure 3a is the ground-state
structure for the isolated cation. In any case, in all subsequent
schemes, compound 3 will be shown having the structure
established in the solid state (vide infra).
The fluxionality of 3a, which gives rise to time-averaged
equivalence of both metals, results from transfer of the methyl
group from one metal to the other (presumably via a methyl-
bridged intermediate) accompanied by slight movement of the
carbonyls from opposite the Ir-Ir bond to opposite the methyl
group, and vice versa. A full merry-go-round motion of these
ligands around the Ir2P4 core is ruled out because this would
give rise to a time-averaged symmetry on either side of the Ir2P4
plane, leading to only one resonance for the dppm methylene
hydrogens; at ambient temperature two resonances are observed
indicating different environments on either side of the Ir2P4
plane.
experiments, and the activation parameters for this exchange
process (∆Hq ) 10.3 ( 0.4 kcal/mol and ∆Sq ) -11.2 ( 1.5
cal/mol K) have been obtained.40 Compound 2 is therefore the
result of an unusually facile and reversible C-H activation of
the methyl group.
Compound 2 bears an interesting relationship to two previ-
ously reported dicarbonyl methyl compounds. In [Rh2(CH3)-
(CO)(µ-CO)(dppm)2][CF3SO3]34 the methyl group is terminally
bound to one rhodium center, whereas in [Ir2(CO)2(µ-CH3)-
(dmpm)2][CF3SO3]41 (dmpm ) Me2PCH2PMe2) the methyl
group bridges the metals in a symmetrical fashion. The existence
of these dicarbonyl, methyl complexes suggested that CO
removal from 2 may reVerse the C-H bond cleavage step,
regenerating a related methyl product. If this were the case, the
further question arose as to the subsequent binding mode of
the methyl group. The answer to the first suggestion is
affirmative; one carbonyl is readily removed from 2 by reaction
with Me3NO, generating the dicarbonyl methyl complex [Ir2-
(CH3)(CO)2(dppm)2][CF3SO3] (3). This species is highly flux-
ional, so at ambient temperature all spectral parameters suggest
a symmetrical species having a bridging methyl group. The 31P-
{1H} NMR spectrum at this temperature shows a singlet at δ
1
22.1, the H NMR spectrum displays the methyl signal as a
quintet at δ 0.58, with equal coupling to all four 31P nuclei,
Confirmation of the proposed structure for 3 in the solid state
comes from the X-ray structure determination, as shown in
Figure 2, with selected distances and angles given in Table 3.
(38) Green, M. L. H.; Hughes, A. K.; Popham, N. A.; Stephens, A. H.
H.; Wong, L.-L. J. Chem. Soc., Dalton Trans. 1992, 307.
(39) (a) Park, J. W.; Mackenzie, P. B.; Schaefer, W. P.; Grubbs, R. H.
J. Am. Chem. Soc. 1986, 108, 6402. (b) Ozawa, F.; Park, J. W.; Mackenzie,
P. B.; Schaefer, W. P.; Henling, L. M.; Grubbs, R. H. J. Am. Chem. Soc.
1989, 111, 1319.
(42) (a) Haines, R. J.; Meintjies, E.; Laing, M. Inorg. Chim. Acta 1979,
36, L403. (b) Haines, R. J.; Laing, M.; Meintjies, E.; Sommerville, P. J.
Organomet. Chem. 1981, 215, C17.
(40) Data analysis was carried out according to the method of McClung
and co-workers: Muhandiram, D. R.; McClung, R. E. D. J. Magn. Reson.
1987, 71, 187.
(41) Reinking, M. K.; Fanwick, P. E.; Kubiak, C. P. Angew. Chem., Int.
Ed. Engl. 1989, 28, 1377.
(43) (a) Brown, M. P.; Cooper, S. J.; Frew, A. A.; Manojlovic-Muir, L.;
Muir, K. W.; Puddephatt, R. J.; Seddon, K. R.; Thomson, M. A. Inorg.
Chem. 1981, 20, 1500. (b) Hill R. H.; Puddephatt, R. J. J. Am. Chem. Soc.
1983, 105, 5797. (c) Xu, C.; Anderson, G. K. Organometallics 1996, 15,
1760.

Metal-Metal CooperatiVity and C-H Bond CleaVage
J. Am. Chem. Soc., Vol. 121, No. 15, 1999 3673
resolved satisfactorily, the almost superposition of the terminal
methyl and carbonyl groups does result in an uncertainty in their
exact positions precluding an in-depth discussion of the
structural parameters.
The conversion of a hydrido methylene species (2) into a
methyl complex (3) upon removal of a carbonyl group, and the
reverse process, C-H bond activation of a methyl group upon
addition of a carbonyl ligand, is a fascinating transformation.
Usually C-H bond activation in low-valent, late transition-metal
complexes is induced by ligand loss to generate the required
coordinative unsaturation.44 We know of one other example of
ligand addition resulting in C-H bond activation, in which phos-
phine addition to a diruthenium complex resulted in cleavage
of an agostic C-H bond.45 It was therefore of interest to deter-
mine the scope of the present C-H bond cleavage by reacting
compound 3 with several ligands having a range of electronic
properties. Although the reactions of 3 with alkynes, allenes, and
olefins were also investigated as part of this study, the diversity
observed in this series of reactions requires separate reports.46
Reaction of 3 with 1 equiv of carbon monoxide regenerates
the tricarbonyl complex 2, accompanied by C-H bond cleavage
of the methyl group and formation of the bridging methylene
and terminal hydrido groups. This is, of course, essentially the
reverse of the conversion of 2 to 3 upon reaction with Me3NO.
Under excess CO compound 2 transforms into a mixture of two
tetracarbonyl species, the methyl complex [Ir2(CH3)(CO)4-
(dppm)2][CF3SO3] (4) and the related acyl product [Ir2(C(O)-
CH3)(CO)4(dppm)2][CF3SO3] (5), in varying ratios depending
on the CO pressure (see Scheme 1). In compound 4 the methyl
resonance appears at δ 0.38 as a triplet in the 1H NMR spectrum,
displaying coupling to the two adjacent phosphorus nuclei,
whereas in 5 the 1H resonance for the methyl group appears as
a singlet at δ 0.63. The 13C{1H} spectra for both compounds
are comparable apart from the low-field resonance for the acyl
carbonyl at δ 225.1; this group also displays a characteristic IR
stretch at 1622 cm-1. It is an interesting dichotomy that
conversion of the methylene and hydrido moieties of 2 into a
methyl group occurs both on addition and remoVal of CO
(yielding 4 and 3, respectively). The migratory insertion,
transforming 4 to 5, is readily reversible, so that in the absence
of excess CO the mixture of 5 quickly regenerates 4, and under
vacuum 4 transforms quantitatively to 2.
Figure 2. A perspective view of the [Ir2(CH3)(CO)(µ-CO)(dppm)2]+
cation of 3. Thermal ellipsoids are drawn at the 20% level except for
methylene and methyl hydrogens which are shown artificially small.
Phenyl groups are numbered starting at the ipso carbon, and hydrogens
on these groups are omitted.
Table 3. Selected Interatomic Distances and Angles for Compound
3; X-Ray Data
(a) Distances (Å)
atom 1
atom 2
distance
atom 1
atom 2
distance
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(2)
Ir(2)
Ir(2)
Ir(2)
P(1)
P(3))
C(1)
C(2)
C(3′)a
P(2)
2.7775(5)
2.307(2)
2.308(2)
1.77(2)
2.110(8)
2.15(2)
2.315(2)
2.313(2)
1.76(2)
Ir(2)
Ir(2)
P(1)
P(2)
P(3)
P(4)
O(1)
O(1′)
O(2)
C(2)
C(3)
C(4)
C(4)
C(5)
C(5)
C(1)
C(1′)
C(2)
2.089(9)
2.16(2)
1.816(8)
1.828(8)
1.837(8)
1.825(8)
1.19(2)
P(4)
C(1′)
1.15(2)
1.095(9)
(b) Angles (deg)
atom 1 atom 2 atom 3 angle atom 1 atom 2 atom 3 angle
Ir(2) Ir(1) C(1) 176.4(6) P(2) Ir(2) P(4) 173.46(8)
Ir(2) Ir(1) C(2) 48.3(2) C(1′) Ir(2) C(2) 134.5(7)
Ir(2) Ir(1) C(3′) 166.0(5) C(2) Ir(2) C(3) 148.5(6)
P(1) Ir(1) P(3) 170.96(8) Ir(1) C(1) O(1) 177.6(16)
C(1) Ir(1) C(2) 128.2(7) Ir(2) C(1′) O(1′) 175.8(20)
C(2) Ir(1) C(3′) 145.7(6) Ir(1) C(2) Ir(2) 82.8(3)
Ir(1) Ir(2) C(1′) 176.6(6) Ir(1) C(2) O(2) 137.9(8)
Ir(1) Ir(2) C(2) 48.9(2) Ir(2) C(2) O(2) 138.8(7)
Reaction of 3 with various phosphines and triphenyl phosphite
also results in methyl C-H bond cleavage, yielding the
methylene-bridged, hydride products [Ir2(H)(CO)2(PR3)(µ-
CH2)(dppm)2][CF3SO3] [PR3 ) PMe3 (6), PMe2Ph (7), PPh3
(8), and P(OPh)3 (9)] as shown in Scheme 2. Compounds 6-8
all have very similar spectral parameters, apart from the 31P-
{1H} resonances for the monodentate phosphine groups, and
also spectrally resemble compound 2 in which a PR3 group is
replaced by CO; therefore all are assumed to have similar
structures. This has been confirmed by the X-ray structure
determination of 6 (vide infra). However, unlike the tricarbonyl
(2), the phosphine adducts do not undergo exchange of the
hydride and methylene hydrogens at ambient temperature.
Ir(1) Ir(2) C(3) 162.4(5)
a Primed atoms are those of the carbonyl [C(1′)O(1′)] and methyl
[C(3′)] groups and are 50:50 disordered with C(1)O(1) and C(3). See
the experimental section.
Compound 3 has an unsymmetrical A-frame structure in which
one carbonyl group is bridging, but the other carbonyl and the
methyl group occupy a terminal position on each metal. These
terminal ligands are not opposite the bridging carbonyl, but are
in a “flattened A-frame” arrangement being almost trans to the
Ir-Ir bond [C(1)-Ir(1)-Ir(2) ) 176.4(6)°, Ir(1)-Ir(2)-C(3) )
162.4(5)°]. This arrangement results in somewhat different
angles between the bridging carbonyl and the different terminal
groups [C(1)-Ir(1)-C(2) ) 128.2(7)° and C(2)-Ir(2)-C(3) )
148.5(6)°]. The Ir-Ir separation of 2.7775(5) Å is consistent
with a single bond and is shorter than the Rh-Rh separation of
2.811(2) Å in the dirhodium analogue34; otherwise the two
structures are closely comparable. Although the disorder in the
equatorial ligand positions (see Experimental Section) was
(44) (a) Collman, J. P.; Hegedus, L. S.; Norton, J. R.; Finke, R. G.
Principles and Applications of Organotransition Metal Chemistry; Univer-
sity Science Books: Mill Valley, CA, 1987; pp 298-304, and references
therein. (b) Arndtsen, B. A.; Bergman, R. G.; Mobley, T. A.; Peterson, T.
H. Acc. Chem. Res. 1995, 28, 154. (c) All publications in J. Organomet.
Chem 1995, 504, 1-155.
(45) Kuhlman, R.; Folting, K.; Caulton, K. G. Organometallics 1995,
14, 3188.
(46) (a) A preliminary report of some of this chemistry has appeared:
Torkelson, J. R.; McDonald, R.; Cowie, M. J. Am. Chem. Soc. 1998, 120,
4047. (b) Torkelson, J. R.; McDonald, R.; Cowie, M. Manuscript in
preparation.

3674 J. Am. Chem. Soc., Vol. 121, No. 15, 1999
Torkelson et al.
Scheme 2
The reaction of 3 with P(OPh)3 actually yields two isomers
(9a and 9b). Isomer 9a has spectral parameters closely
resembling those of compounds 6-8 so it is assumed to have
an analogous structure, whereas the structure of 9b could not
be established conclusively by NMR spectroscopy. The two
structures shown below (as F and G) for isomer 9b are
suggested on the basis of the two carbonyl resonances, one of
which is in the region of those in 9a, suggesting that it also is
opposite the µ-CH2 group, and one of which is clearly in a
different environment. We favor structure F for 9b, because
both dppm 31P resonances are at similar chemical shifts,
suggesting similar environments with one carbonyl on each
metal. However we could not carry out a selective 31P-
decoupling experiment to conclusively rule out the alternate
structure (G).
Figure 3. A perspective view of the [Ir2H(CO)2(PMe3)(µ-CH2)-
(dppm)2]+ cation of 6. Thermal parameters and numbering as in Figure
2.
Table 4. Selected Interatomic Distances and Angles for Compound
6; X-Ray Data
(a) Distances (Å)
atom 1
atom 2
distance
atom 1
atom 2
distance
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(2)
Ir(2)
Ir(2)
Ir(2)
P(1)
P(3)
P(5)
C(1)
C(2)
P(2)
P(4)
C(2)
2.8308(6)
2.332(2)
2.330(2)
2.359(3)
1.887(9)
2.172(9)
2.304(2)
2.305(2)
2.098(9)
Ir(2)
Ir(2)
P(1)
P(2)
P(3)
P(4)
O(1)
O(3)
C(3)
H(2)
C(4)
C(4)
C(5)
C(5)
C(1)
C(3)
1.909(9)
1.71
1.819(9)
1.830(9)
1.818(9)
1.850(9)
1.141(11)
1.117(12)
(b) Angles (deg)
atom 1 atom 2 atom 3 angle atom 1 atom 2 atom 3 angle
Ir(2)
Ir(2)
Ir(2)
P(1)
P(5)
P(5)
C(1)
Ir(1)
Ir(1)
P(2)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(2)
Ir(2)
Ir(2)
P(5) 131.72(7) C(2)
C(1) 123.3(3) Ir(1)
Ir(2) C(3) 157.9(4)
P(5) C(91) 116.7(4)
P(5) C(92) 116.9(5)
P(5) C(93) 115.5(4)
C(2)
47.4(2) Ir(1)
P(3) 163.16(9) Ir(1)
C(1) 104.9(3) C(91) P(5) C(92) 99.5(6)
C(2) 84.4(2) C(91) P(5) C(93) 102.6(7)
C(2) 170.7(4) C(92) P(5) C(93) 103.2(8)
C(3) 108.3(3) Ir(1) C(1) O(1) 178.5(9)
C(2) 49.6(2) Ir(1) C(2) Ir(2) 83.0(3)
P(4) 154.79(8) Ir(2) C(3) O(3) 177.4(9)
The structure of 6, as determined by X-ray diffraction is
shown in Figure 3. A compilation of important bond lengths
and angles is given in Table 4. This structure confirms the
methylene-bridged, hydride formulation and also displays the
arrangement of carbonyl, hydride, and PMe3 groups as proposed
from the spectral results. The overall geometry is that of an
A-frame structure having the carbonyls essentially opposite the
bridging methylene group, with the PMe3 and hydride ligands
on the outside of the A-frame, almost bisecting the angles
between the µ-CH2 and terminal carbonyl groups. Both ends
of the diphosphine ligands are bent significantly from a trans
alignment owing to repulsions involving the PMe3 group, (P(1)-
Ir(1)-P(3) ) 163.16(9)° and P(2)-Ir(2)-P(4) ) 154.79(8)°).
These repulsions may also be responsible for larger Ir-P (dppm)
distances at the more crowded Ir(1) center (2.330(2), 2.332(2)
Å] than at Ir(2) (2.304(2), 2.305(2) Å], as well as the larger
Ir(1)-C(2) distance [2.172(9) Å] compared with Ir(2)-C(2)
[2.098(9) Å] involving the methylene group. The angle at the
methylene carbon [83.0(3)°] is typical of such a group when
bridging a metal-metal bond,47 as suggested by the short Ir-
Ir separation of 2.8308(6) Å.
The reaction of 3 with tBuNC also occurs readily, producing
the C-H-activated product [Ir2(H)(CO)2(CNtBu)(µ-CH2)(dppm)2]-
[CF3SO3] (10) analogous to the phosphine adducts (6-9a). As
with the phosphine adducts, no exchange between the hydride
and methylene protons is observed at ambient temperature. All
spectroscopic parameters support the structure shown, and in
particular, the characteristic CtN stretch in the IR spectrum
for the tBuNC appears at 2163 cm-1, consistent with this group
being terminally bound.48 The increase in frequency of the Ct
N stretch compared with that in free BuNC (2125 cm-1) is
consistent with this group functioning primarily as a σ donor.
As for the CO (2) and PR3 (6-9) adducts, no Ir-H stretch could
be detected in the IR spectrum of compound 10.
t
(48) (a) Treichel, P. M. AdV. Organomet. Chem. 1973, 11, 21. (b)
Lukehart, C. Fundamental Transition Metal Organometallic Chemistry;
Brooks/Cole Publishing Company: Belmont, CA, 1985; p 77.
(47) Herrmann, W. A. AdV. Organomet. Chem. 1982, 20, 159.

Metal-Metal CooperatiVity and C-H Bond CleaVage
The reaction of compound 3 with SO2 appears to be somewhat
J. Am. Chem. Soc., Vol. 121, No. 15, 1999 3675
t
analogous to the reactions with CO, phosphines, and BuNC,
producing a methylene-bridged, hydrido complex [Ir2(H)(CO)2-
(µ-CH2)(µ-SO2)(dppm)2][CF3SO3] (11). Like the CO adduct (2),
compound 11 is fluxional, undergoing facile exchange of the
hydride ligand and the methylene hydrogens. The unusually
facile exchange process in 11 (vide infra) caused some initial
uncertainty about its formulation, because it was necessary to
rule out a species, [Ir2(CO)2(µ-η1,η2-CH3)(µ-SO2)(dppm)2][CF3-
SO3], containing an asymmetrically bound methyl group of the
type shown in structure E. In this case hydrogen exchange would
occur by rotation about the Ir-CH3 bond, interchanging the
agostic interaction between the three C-H bonds. At ambient
temperature the 31P{1H} NMR spectrum appears as a singlet at
δ -1.0, consistent with the chemical equivalence of all four
phosphorus nuclei. In the 1H NMR spectrum the three hydrogens
of the methyl group appear as a broad singlet at δ -2.15, with
coupling of 98 Hz observed to carbon when 13CH3+ is used. At
50 °C the “CH3 signal” appears as a well-resolved quintet,
coupling equivalently to all four phosphorus nuclei. The high-
field shift of this resonance and the small average C-H coupling
constant are consistent with a fluxional process involving either
the methylene-hydride or the agostic methyl formulations,
Figure 4. 1H NMR spectrum in the “methyl” region for the CH3,
CH2D, and CH2D isotopomers of compound 11.
1
although the average value of JC-H is slightly lower than has
Table 5. Variation of Chemical Shift with Temperature for the
Isotopomers of Compound 11
previously been suggested for asymmetrically bridged methyl
groups.49 At ambient temperature the 13C NMR spectra add
additional support for the symmetric time-averaged structure
with only one signal (at δ 170.8) for both carbonyls and a sharp
quartet at δ 36.8 for the methyl carbon, showing equivalent
coupling to the three hydrogens.
temp (°C)
δ(CH3)
δ(CH2D)
δ(CD2H)
∆
∆
2
1
25
10
0
-2.07
-2.05
-2.05
-2.02
-3.48
-3.52
-3.56
-3.62
-5.72
-5.90
-6.02
-6.40
1.41
1.47
1.51
1.60
2.24
2.38
2.46
2.78
-20
Lowering the temperature causes the methyl resonance in the
1H NMR spectrum to broaden until at -60 °C the signal has
fully collapsed into the baseline. At -110 °C a very broad signal
at ca. δ 4.2 (2H) and an equally broad signal at ca. δ -14.6
(1H) start to appear, but are only barely discernible above the
baseline. Support for the exchange process was shown by
irradiating the low-field signal at -110 °C and observing that
the high-field signal disappeared (although the barely discernible
signal at this temperature makes such a conclusion equivocal).
The activation parameters for this exchange process (∆Hq )
6.1 ( 0.1 kcal/mol and ∆Sq ) -6.2 ( 0.5 cal/mol K) were
obtained from a line-shape analysis on variable-temperature 1H
NMR spectra.50 At the outset, these activation parameters
appeared to us to be too small to correspond to a C-H bond-
cleavage process, being about half the value obtained in the
tricarbonyl compound, 2, and much less than values near 20
kcal/mol previously observed in binuclear systems.51 On the
other hand, these values are in line with the activation energy
(∆Gq ) 9.8 kcal/mol) for exchange of the methyl hydrogens in
a bridging agostic methyl interaction, by rotation about the
metal-carbon bond.39
a For a definition of ∆1 and ∆2 see Figure 4.
of resonance (IPR) in compound 11. Although first described
by Saunders et al.52 in studies of carbocations, this effect was
later used (first by Calvert and Shapley53 and subsequently by
others10a,54) to help establish the presence of agostic interactions
involving methyl groups. Owing to zero-point energy differences
an agostic interaction of a C-H bond is favored over that of a
1
C-D bond. As a result, the H chemical shift of an agostic
CHD2 group appears at a higher field than that of a CH2D group,
which is in turn at a higher field than that of a CH3 group.
Furthermore, the chemical shifts of the CHD2 and CH2D groups
are also temperature dependent, shifting upfield at lower
temperatures; this effect is greater for CHD2 than for CH2D.
The 1H NMR spectrum in the methyl region for compound 11,
which was prepared from d3-methyl triflate, is shown in Figure
4, and the temperature variation of the chemical shifts is
presented in Table 5. Although these results clearly demonstrate
the IPR phenomenon, consistent with hydrogen exchange
between the agostic and anagostic hydrogens, they do not by
themselves establish the existence of an agostic interaction,
because it has been pointed out that this effect will also be
observed in a system in which facile exchange occurs between
a hydrido ligand and methylene hydrogens.49 From the IPR data,
the chemical shifts for the two different hydrogen environments
are calculated49b,53 at δ 4.13 and -14.47. It is encouraging that
these values are in very close agreement with the positions of
the very broad resonances that are beginning to appear in the
Attempts to prepare compound 11 with d3-methyl triflate
resulted in partial hydrogen incorporation (presumably from
solvent) such that the desired product was contaminated with
small, but significant amounts of the CHD2, CH2D, and CH3
isotopomers. This inadvertent result presented us with the
opportunity to investigate the effects of isotopic perturbation
(49) (a) Brookhart, M.; Green, M. L. H.; Wong, L.-L. Prog. Inorg. Chem.
1988, 36, 1. (b) Tolman, C. A.; Faller, J. W. In Homogeneous Catalysis
with Metal Phosphine Complexes; Pignolet, L. H., Ed.; Plenum Press: New
York, 1983; pp 30-34.
(50) (a) The program used was NMR Exchange for Windows 95, written
by R. E. D. McClung, 1997. (b) McConnell, H. M. J. Chem. Phys. 1958,
28, 430.
(52) Saunders: M.; Jaffe, M. H.; Vogel, P. J. Am. Chem. Soc. 1971, 93,
2558.
(53) Calvert, R. B.; Shapley, J. R. J. Am. Chem. Soc. 1978, 100, 7726.
(54) See for example: (a) Casey, C. P.; Fagan, P. J.; Miles, W. H. J.
Am. Chem. Soc. 1982, 104, 1134. (b) Green, M. L. H.; Hughes, A. K.;
Popham, N. A.; Stephens, A. H. H.; Wong, L.-L. J. Chem. Soc., Dalton
Trans. 1992, 3077.
(51) Jacobsen, E. N.; Goldberg, K. I.; Bergman, R. G. J. Am. Chem.
Soc. 1988, 110, 3706.

3676 J. Am. Chem. Soc., Vol. 121, No. 15, 1999
Torkelson et al.
Table 6. Selected Interatomic Distances and Angles for Compound
11; X-Ray Data
(a) Distances (Å)
atom 1
atom 2
distance
atom 1
atom 2
distance
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(2)
Ir(2)
Ir(1)
Ir(1)
Ir(2)
Cl(1)
S(1)
C(1)
C(1′)a
C(3)
Cl(1)
S(1)
2.8534(7)
2.66(2)
Ir(2)
Ir(2)
S(1)
S(1)
O(1)
O(1′)
O(2)
Ir(2)
Ir(2)
C(2)
C(3)
O(3)
O(4)
C(1)
C(1′)
C(2)
P(2)
P(4)
1.85(2)
2.123(14)
1.459(8)
1.450(8)
1.16(2)
1.15a
1.14(2)
2.337(3)
2.330(3)
2.449(3)
1.88(2)
1.85b
2.190(12)
2.68(2)
2.273(3)
2.353(3)
2.364(3)
P(1)
P(3)
(b) Angles (deg)
angle atom 1 atom 2 atom 3
atom 1 atom 2 atom 3
Ir(2) Ir(1) S(1)
angle
50.06(7) Ir(1) Ir(2) C(3) 49.6(3)
Cl(1) Ir(2) C(2) 141.3(7)
Ir(2) Ir(1) C(1′) 170.3(32) S(1) Ir(2) C(2) 105.6(5)
Ir(2) Ir(1) C(3) 47.6(4) S(1) Ir(2) C(3) 105.3(3)
Cl(1) Ir(1) C(1′) 130.7(32) P(2) Ir(2) P(4) 161.90(11)
Ir(2) Ir(1) C(1) 141.6(5)
S(1)
S(1)
S(1)
P(1)
Ir(1) C(1) 91.6(5)
Ir(1) C(1′) 120.9(33) Ir(1) Cl(1) Ir(2) 64.6(5)
Ir(1) C(3) 97.6(4) Ir(1) S(1) Ir(2) 74.24(8)
Ir(1) P(3) 170.60(10) Ir(1) C(1) O(1) 172.3(17)
C(2) Ir(2) C(3) 149.1(6)
Figure 5. A perspective view of the [Ir2H(CO)2(µ-CH2)(µ-SO2)-
(dppm)2]+ cation of 11. Thermal parameters and numbering as in Figure
2. The hydride ligand [H(1)] and the hydrogens on C(3) were not located
in the X-ray study, but are placed in idealized positions.
C(1) Ir(1) C(3) 170.7(6)
Ir(1) Ir(2) Cl(1) 57.4(5)
Ir(1) Ir(2) S(1)
Ir(1) C(1′) O(1′) 175.7(77)
Ir(2) C(2) O(2) 177.4(16)
55.70(8) Ir(1) C(3) Ir(2) 82.8(5)
1H NMR spectrum at -110 °C. Furthermore, the chemical shift
of the low-field signal is more consistent with a methylene group
than a methyl group. This conclusion is supported by the low
average value of 1JC-H (98 Hz), which would then result from
Ir(1) Ir(2) C(2) 161.3(5)
a Primed atoms belong to [Ir2(CO)2(µ-Cl)(µ-SO2)(dppm)2][CF3SO3]
which is superimposed on compound 11 in a 12.5:87.5 disorder.
b Distance fixed (within 0.01 Å of this value) during refinement.
1
averaging of the two single-bond coupling constants of JC-H
≈ 147 Hz with a two-bond coupling (between the methylene
asymmetric manner with the Ir(1)-S(1) bond length [2.449(3)
Å] longer than the Ir(2)-S(1) bond length [2.273(3) Å]; this
difference is consistent with the high trans influence of the
hydride ligand, which is assumed to be in the vacant coordina-
tion site on Ir(1) opposite S(1).
2
carbon and the hydride ligand) of JC-H ≈ 0 Hz. The C-H
coupling constant proposed for the methylene group is in good
agreement with previous determinations.55
In hopes of establishing conclusively (at least in the solid
state) whether compound 11 contained an agostic methyl group
or methylene and hydride groups, the X-ray structure determi-
nation of 11 was undertaken. Unfortunately, the conclusions
were equivocal here also. Our only successful attempts to obtain
suitable crystals were from CH2Cl2 solvent, from which the
desired crystals were always contaminated by varying amounts
of [Ir2(CO)2(µ-Cl)(µ-SO2)(dppm)2][CF3SO3],16 which resulted
from replacement of the methyl group by a chlorine atom
(crystals of compound 11 and this chloro-bridged species are
isomorphous). The resulting placement of the fractional oc-
cupancy chloro ligand on top of the CH3 or CH2 group prevents
location of the attached hydrogens; however, the metrical
parameters of the complex (vide infra) are consistent with the
methylene-hydride formulation suggested by the NMR study.
An ORTEP diagram for compound 11 is shown in Figure 5
with important bond lengths and angles given in Table 6.
Compound 11 has the SO2 group in a bridging arrangement
with the carbonyl ligands bound terminally, one on each metal.
Although the hydrogen atoms originating on the methyl group
of the precursor (3) could not be located, C(3) is assumed to be
associated with a methylene group which bridges the metals
on the face opposite the SO2 group; the hydride ligand and the
methylene hydrogens have been idealized in the positions
shown. Excluding the metal-metal bond the geometry around
the metals is best described as octahedral at Ir(1) and trigonal
bipyramidal at Ir(2), with both diphosphine groups in the
expected trans arrangements [P(1)-Ir(1)-P(3) ) 170.6(1)° and
P(2)-Ir(2)-P(4) ) 161.9(1)°]. The SO2 group bridges in an
Although the crystallographic data do not unambiguously
differentiate between an asymmetrically bridged methyl group
and a methylene-bridged hydride structure, the structural
parameters are more supportive of the latter assignment. It is
clear that carbonyl group C(1)O(1) is bent back toward the
µ-SO2 group compared with C(2)O(2) [Ir(2)-Ir(1)-C(1) )
141.6(5)°, Ir(1)-Ir(2)-C(2) ) 161.3(5)°] giving rise to a larger
C(1)-Ir(1)-C(3) angle [170.7(6)°] than the C(2)-Ir(2)-C(3)
angle of 149.1(6)°, and suggesting the location of the hydride
ligand in the resulting vacant site, as shown in Figure 5. All
spectroscopic and crystallographic data together support the
methylene-bridged, hydride structure for compound 11, as do
the DFT calculations reported below.
Reaction of compound 11 with CO parallels that of the
methylene-bridged tricarbonyl (2), yielding the SO2-bridged,
tricarbonyl methyl species, [Ir2(CH3)(CO)3(µ-SO2)(dppm)2][CF3-
SO3] (12), in which the methyl group is now terminally bound
to one of the metals. The three carbonyls are also shown to be
terminally bound, on the basis of 13C NMR and IR spectroscopy.
Although the relative positions of the methyl and µ-SO2 groups
are not known, the geometry shown for 12 in Scheme 3 is based
on the assumption that hydride transfer from Ir to the methylene
group, yielding a methyl group σ-bound to the adjacent metal,
would leave coordinative unsaturation at the first metal, which
would be alleviated by CO coordination. In a further parallel
of the reaction of 2 with CO, compound 12 reacts further with
CO to produce the acyl compound [Ir2(CH3CO)(CO)3(µ-
SO2)(dppm)2][CF3SO3] (13), as shown by the loss of coupling
for the methyl group which now appears as a singlet in the 1H
(55) See for example, compound 2 and refs 49a and 54.

Metal-Metal CooperatiVity and C-H Bond CleaVage
J. Am. Chem. Soc., Vol. 121, No. 15, 1999 3677
Scheme 3
corresponding to the X-ray-derived structure of 3. Surprisingly,
the calculations did not converge to a structure close to that
input, but instead the bridging carbonyl migrated away from
the bridging site, with concomitant movement of the methyl
group, yielding the geometry shown in Figure 6a. In the
computed structure IIIt, the square plane about Ir(2), defined
by P(2), P(4), C(2), and C(3), is not perpendicular to the Ir-
(1)-Ir(2) vector, but is tilted significantly such that the carbonyl
C(2)O(2) is tipped toward Ir(1), with the methyl group tipped
away [Ir(1)-Ir(2)-C(2) ) 78.6°; Ir(1)-Ir(2)-C(3) ) 111.6°].
This tilt may be favored by the retention of a small component
of π back-bonding from Ir(1) to C(2)O(2). It is significant that
the proposed structure of 3 in solution (3a) corresponds to that
calculated for the isolated cation, having a geometry rather
different from that in the solid state. The transformation from
the solid-state structure of 3 to that calculated for IIIt involves
very little change at Ir(1), because in both structures carbonyl
C(1)O(1) is essentially opposite the metal-metal bond. The
change primarily involves twisting of the O(2)C(2)-Ir(2)-C(3)
unit so that C(2)O(2) moves away from the bridging position.
We next carried out calculations on [Ir2(CO)2(µ-CH3)-
(dhpm)2]+ (isomer IIIb), in which the methyl group in IIIt has
moved from a terminal position to a bridging site. Mulliken
computational studies on two isomers of IIIb have been carried
out; in the first, the methyl group is unsymmetrically bridged,
involving an agostic interaction, as shown earlier in structure
E, whereas the second isomer has the methyl group bridging
symmetrically (structure D). Both results are shown in Figure
6b and 6c, respectively. The unsymmetrically bridged isomer
is more stable by only 0.8 kcal/mol, and significantly, even the
more stable methyl-bridged structure is approximately 6 kcal/
mol less stable than structure IIIt, having the terminal methyl
group. Table 7 contains a comparison of the pertinent geometric
features of these three isomers [IIIt, IIIb (agostic), and IIIb
(symmetric)]. In general, the parameters for IIIb (symmetric)
agree reasonably well with those of the X-ray determination of
the dmpm analogue,41 and in particular the calculated Ir-Ir bond
length is comparable with that in the X-ray structure.
NMR spectrum at δ 1.28, and by the IR spectrum which shows
an acyl band at 1646 cm-1. Again the geometry shown for 13
is not known but the structure shown is based on the assumed
migration of the methyl group to the adjacent carbonyl on the
same metal.
Theoretical Results
We have carried out DFT calculations on two different
cationic species related to compounds 3 and 11 and on isomers
of each, in efforts to learn more about the bonding involving
the methyl groups and about the C-H bond-activation process
of methyl groups in the vicinity of adjacent metals. The first
compound investigated was [Ir2(CH3)(CO)2(dhpm)2]+ (IIIt); this
is the analogue of 3 in which the phenyl substituents on the
bridging diphosphine ligands were input as hydrogens to
facilitate the calculations (i.e., dhpm ) H2PCH2PH2). Calcula-
tions on the related A-frame, [Ir2(CO)2(µ-CH3)(dhpm)2]+, in
which the methyl group now bridges the metals, were also
carried out in attempts to establish which (if any) of the above-
mentioned structures (terminal or bridging CH3) was favored
for the isolated cation, and how the energies of these isomers
compared. In our calculations, the structure having a terminal
methyl group is labeled (IIIt), whereas the dhpm analogue of
Kubiak’s methyl-bridged compound41 is labeled IIIb.
Although conversion of the methyl compound 3 to the
respective methylene hydride products requires the addition of
t
small molecules such as CO, PR3, BuNC, and SO2, it is not
Using the X-ray coordinates of 3 as a starting point for the
calculations on IIIt, we looked for an energy minimum
clear whether ligand addition induces C-H bond cleavage or
whether prior C-H bond activation in 3 occurs to yield
Figure 6. The calculated structures of three isomers of the [Ir2(CH3)(CO)2(dhpm)2]+ cation: (a) structure IIIt; (b) structure IIIb (agostic); (c)
structure IIIb (symmetric).

3678 J. Am. Chem. Soc., Vol. 121, No. 15, 1999
Torkelson et al.
Table 7. Geometric Features Calculated for Structures IIIt, Both Agostic and Symmetric
Methyl-Bridged Structures of IIIb, and Structures IIIh and III-TS
structure IIIb
(agostic)
structure IIIb
(symmetric)
atom 1
atom 2
atom 3
structure IIIt
structure IIIh
structure III-TS
Distances (Å)
2.99
Ir(1)
Ir(1)
Ir(2)
Ir
Ir(1)
Ir(2)
Ir(1)
C(3)
Ir(2)
C(3)
C(3)
P (av)
C(1)
C(2)
H(1)
H(1)
2.84
4.14
2.14
2.34
1.84
1.93
2.99
2.34
2.34
2.34
1.86
1.86
2.92
1.09
2.98
2.09
2.19
2.33
1.96
1.91
1.59
3.07
2.16
2.17
2.33
1.92
1.90
1.62
1.80
2.37
2.30
2.34
1.85
1.87
2.09
1.10
1.13
Angles (deg)
79.6
135.8
51.2
49.2
177.1
173.1
173.9
145.5
28.5
Ir(1)
Ir(1)
Ir(1)
Ir(2)
P
C(2)
C(1)
C(1)
C(3)
C(3)
Ir(2)
Ir(2)
Ir(1)
Ir
Ir(2)
Ir(1)
Ir(1)
Ir(1)
Ir(2)
C(2)
C(3)
C(3)
P
C(3)
C(3)
H(1)
H(1)
79.6
138.8
50.2
88.3
129.6
44.4
90.3
132.7
44.7
78.6
111.6
50.2
47.3
44.9
174.5
169.8
177.8
170.6
170.6
170.8
174.0
167.7
90.4
175.2
177.4
175.6
121.1
54.5
101.8
Table 7. Structure III-TS very much resembles IIIh except that
the hydrido ligand is substantially closer to the methylene group
(1.80 vs 2.87 Å) but still retains hydride character, having a
normal Ir-H distance of 1.62 Å.
The initial ambiguities about the fate of the methyl group
(i.e., CH3 or µ-CH2/H) in the SO2 adduct 11 led us to carry out
DFT calculations, using Gaussian 94 software, on this species.
Again, the X-ray-derived positions were used as a starting point
for the [Ir2H(CO)2(µ-CH2)(µ-SO2)(dhpm)2]+ cation (XIh; “h”
signifies hydride). It is encouraging that geometry optimization
yields a structure that agrees with the methylene hydride
formulation deduced from the NMR and X-ray results. The
calculated geometry is diagrammed in Figure 8a. The most
notable discrepancy between the calculated and X-ray structures
is the Ir-Ir distance, which was 0.47 Å too long (3.32 vs 2.85
Å) in the initial calculations. This discrepancy in Ir-Ir bond
length is surprising because all other structures calculated in
this article have the expected metal-metal separations; and in
particular, the structure of IIIb, involving the symmetrically
bridged methyl group, is close to the crystallographic value for
the dmpm analogue. The Couty/Hall modification to the valence
p orbitals of the Ir atom in the Gaussian 94 calculations did not
provide much change in any of the parameters. However,
addition of a polarization orbital to the S atom in the Gaussian
94 calculations resulted in significant improvement to the Ir-S
and S-O distances. The effect of the combined change in the
Ir p orbital and the polarization orbital on the S atom was to
decrease the Ir-Ir distance from 3.32 to 3.21 Åsan improve-
ment, but still a long way from the experimental (X-ray) result
of 2.85 Å. Scanning the remainder of the bond lengths in Table
8, we see that Gaussian 94 (with modifications to the Ir p orbital
and added polarization function on the S atom) and Mulliken
results give overall best agreement with experiment.
Figure 7. Potential energy diagram for the different isomers of the
[Ir2(CH3)(CO)2(dhpm)2]+ cation III. Energies in kcal/mol. Molecular
drawings are shown with the dhpm ligands, above and below the plane
of the paper, omitted for clarity.
undetectable amounts of a methylene hydride [Ir2H(CO)2(µ-
CH2)(dppm)2]+ which subsequently reacts with the added
ligands to yield the final products. We have therefore carried
out DFT calculations on the dhpm analogue of this methylene
hydride cation (called IIIh; “h” signifies hydride) and a drawing
of the equatorial plane of this species (phosphine ligands
omitted) appears in Figure 7 (which shows the relative energies
of all isomers of III), with calculated bond lengths and angles
in Table 7. This structure is only 3.4 kcal/mol less stable than
structure IIIb (agostic). A transition state between IIIb (agostic)
and IIIh has been located at an energy of about 15.0 kcal/mol
above that of the agostic structure.56 A drawing of the equatorial
plane of this transition-state structure (labeled III-TS) is shown
in Figure 7, and important structural parameters are given in
Despite the discrepancy between calculations and experiment
(Table 8), there is a remarkable correlation in the asymmetry
of binding in the SO2 ligand in all models. The experimental
asymmetry in SO2 binding [(Ir(1)-S(1))-(Ir(2)-S(1))] is 0.18
Å, whereas the four theoretical results give asymmetries of 0.16,
0.15, 0.15, and 0.13 Å. We previously attributed this asymmetry
to the trans influence of the hydride ligand. Among the angles
in Table 8, there are several involving C(3) which deviate from
experiment by as much as 10-15°; most of these can be
attributed to the long Ir-Ir distance noted above. One exception
(56) We were unable to locate the transition state between IIIb (agostic)
and IIIh by using automated procedures in the Mulliken software. Therefore,
we completed a linear synchronous transit calculation connecting IIIb
(agostic) with IIIh. Then we picked selected points near the top of the
energy barrier, and carried out several additional calculations in which we
optimized all of the parameters except the Ir-Ir-H angle. We believe that
we have located the transition state by this means to within 0.5 kcal/mol.
Additional studies are in progress to further examine the nature of the
transition state for this system.

Metal-Metal CooperatiVity and C-H Bond CleaVage
J. Am. Chem. Soc., Vol. 121, No. 15, 1999 3679
Figure 8. The calculated structures of two isomers of the [Ir2(CH3)(CO)2(µ-SO2)(dhpm)2]+ cation; (a) [Ir2H(CO)2(µ-CH2)(µ-SO2)(dhpm)2]+ (XIh)
and (b) [Ir2(CO)2(SO2)(µ-CH3)(dhpm)2]+ (XIa), and (c) the transition state (XI-TS) between them.
Table 8. Comparison of Geometries of the Different Calculated Models for Structure XIh and that of the X-Ray Structure of Compound 11
(a) Bond Lengths (Å)
G94b
B3LYP
G94 (with
mod Ir p orb)
G94 (with mod Ir p
orb and S d orb)
Mulliken
B3LYP
atom 1
atom 2
expta
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(2)
Ir(2)
Ir(2)
Ir(2)
Ir(2)
S(1)
S(1)
O(1)
O(2)
Ir(2)
S(1)
P(1)
P(3)
C(1)
C(3)
H(1)
S(1)
P(2)
P(4)
C(2)
C(3)
O(3)
O(4)
C(1)
C(2)
2.85
2.45
2.35
2.36
1.88
2.19
1.70c
2.27
2.34
2.33
1.85
2.12
1.46
1.45
1.16
1.14
3.32
2.64
2.41
2.41
1.94
2.18
1.58
2.48
2.41
2.41
1.93
2.13
1.64
1.64
1.17
1.17
3.29
2.61
2.40
2.41
1.94
2.18
1.58
2.46
2.41
2.41
1.93
2.13
1.64
1.64
1.17
1.17
3.21
2.52
2.40
2.40
1.93
2.20
1.60
2.37
2.41
2.41
1.93
2.12
1.50
1.50
1.17
1.18
3.24
2.54
2.35
2.35
1.95
2.19
1.61
2.41
2.36
2.36
1.94
2.14
1.48
1.48
1.15
1.15
(b) Bond Angles (deg)
G94b
B3LYP
G94 (with
mod Ir p orb)
G94 (with mod Ir p
orb and S d orb)
Mulliken
B3LYP
atom 1
atom 2
atom 3
expta
Ir(2)
Ir(2)
Ir(2)
Ir(2)
S(1)
S(1)
P(1)
C(1)
C(1)
C(3)
Ir(1)
Ir(1)
Ir(1)
S(1)
S(1)
P(2)
C(2)
Ir(1)
Ir(1)
Ir(1)
Ir(2)
Ir(2)
O(3)
Ir(1)
Ir(2)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(2)
Ir(2)
Ir(2)
Ir(2)
Ir(2)
Ir(2)
Ir(2)
S(1)
S(1)
S(1)
S(1)
S(1)
S(1)
C(1)
C(2)
C(3)
S(1)
C(1)
C(3)
H(1)
C(1)
H(1)
P(3)
C(3)
H(1)
H(1)
S(1)
C(2)
C(3)
C(2)
C(3)
P(4)
C(3)
Ir(2)
O(3)
O(4)
O(3)
O(4)
O(4)
O(1)
O(2)
Ir(2)
50.1
143.5
47.6
47.6
143.4
39.1
123.6
95.8
171.2
169.1
177.5
93.0
84.5
51.8
145.5
40.1
93.7
91.9
170.6
174.4
80.6
120.0
119.9
106.2
106.2
114.9
175.3
174.4
100.7
56.4
143.2
39.5
123.8
86.9
177.7
169.1
177.3
93.0
84.3
52.8
145.3
40.7
92.5
93.5
170.4
174.0
70.8
123.6
119.0
109.2
105.3
115.1
175.7
174.5
99.8
50.9
142.1
41.1
124.6
95.1
171.6
168.8
176.7
93.3
83.4
50.9
145.1
42.9
94.2
93.9
167.9
172.0
82.0
116.7
116.8
109.0
109.0
117.0
175.8
173.3
95.9
47.4
142.5
40.9
125.6
95.2
173.0
165.3
176.6
91.8
84.7
50.8
144.7
42.2
93.9
93.1
165.9
173.1
81.8
117.3
117.3
108.6
108.6
116.6
175.1
173.6
96.9
125c
93.5
175c
170.6
168.8
91c
78c
55.7
161.3
49.5
105.7
105.2
161.9
149.1
74.3
115.5
115.6
117.0
115.8
113.3
175.3
177.8
82.9
a X-ray structure of 11. b G94 ) Gaussian 94. c The H(1) position was placed in an idealized position in the X-ray structure.

3680 J. Am. Chem. Soc., Vol. 121, No. 15, 1999
Torkelson et al.
Table 9. Comparison of Geometries of the Different Calculated
Models for Structure XIa
Table 10. Selected Bond Lengths and Angles for the Calculated
(Mulliken, B3LYP) Structure XI-TS
(a) Bond Lengths (Å)
(a) Bond Lengths (Å)
G94a
B3LYP
G94
Mulliken
B3LYP
atom 1
atom 2
distance
atom 1
atom 2
distance
atom 1
atom 2
(with mod Ir p orb)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(2)
Ir(2)
S(1)
P(1)
P(3)
C(1)
C(3)
H(1)
P(2)
3.07
2.57
2.36
2.36
1.91
2.31
1.61
2.36
Ir(2)
Ir(2)
Ir(2)
S(1)
S(1)
O(1)
O(2)
C(3)
P(4)
C(2)
C(3)
O(3)
O(4)
C(1)
C(2)
H(1)
2.36
1.92
2.09
1.48
1.48
1.15
1.16
1.94
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(2)
Ir(2)
Ir(2)
Ir(2)
Ir(2)
S(1)
S(1)
O(1)
O(2)
C(3)
Ir(2)
S(1)
P(1)
P(3)
C(1)
C(3)
H(1)
S(1)
P(2)
P(4)
C(2)
C(3)
O(3)
O(4)
C(1)
C(2)
H(1)
3.06
3.29
2.40
2.40
1.85
2.30
2.07
2.62
2.42
2.42
1.88
2.31
1.63
1.63
1.18
1.17
1.13
3.01
3.30
2.40
2.40
1.86
2.30
2.05
2.59
2.42
2.42
1.88
2.31
1.64
1.64
1.18
1.17
1.14
3.00
3.43
2.34
2.34
1.87
2.28
2.11
2.64
2.36
2.36
1.88
2.38
1.48
1.48
1.16
1.16
1.13
(b) Bond Angles (deg)
atom 1 atom 2 atom 3 angle atom 1 atom 2 atom 3 angle
Ir(2)
Ir(2)
Ir(2)
Ir(2)
S(1)
S(1)
S(1)
C(1)
C(1)
C(3)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(2)
Ir(2)
Ir(2)
S(1)
50.8 S(1)
C(1) 150.2 S(1)
Ir(2)
Ir(2)
Ir(2)
S(1)
S(1)
S(1)
S(1)
S(1)
S(1)
C(1)
C(2)
C(3)
C(2)
94.3
C(3) 162.8
C(3) 162.3
C(3)
H(1)
C(1)
C(3)
42.9 C(2)
98.9 Ir(1)
99.4 Ir(1)
93.7 Ir(1)
Ir(2)
75.1
O(3) 117.1
O(4) 117.1
O(3) 111.7
O(4) 111.7
O(4) 116.8
O(1) 175.4
O(2) 172.8
H(1) 149.7 Ir(2)
C(3) 166.9 Ir(2)
H(1) 110.9 O(3)
(b) Bond Angles
G94a
G94
Mulliken
H(1)
S(1)
56.0 Ir(1)
54.0 Ir(2)
atom 1 atom 2 atom 3 B3LYP (with mod Ir p orb) B3LYP
C(2) 148.4 Ir(1)
C(3) 58.8
Ir(2)
88.3
Ir(2)
Ir(2)
Ir(2)
Ir(2)
S(1)
S(1)
S(1)
P(1)
C(1)
C(1)
C(3)
Ir(1)
Ir(1)
Ir(1)
S(1)
S(1)
P(2)
C(2)
Ir(1)
Ir(1)
Ir(1)
Ir(2)
Ir(2)
O(3)
Ir(1)
Ir(2)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(1)
Ir(2)
Ir(2)
Ir(2)
Ir(2)
Ir(2)
Ir(2)
Ir(2)
S(1)
S(1)
S(1)
S(1)
S(1)
S(1)
C(1)
C(2)
C(3)
S(1)
C(1)
C(3)
H(1)
C(1)
C(3)
H(1)
P(3)
C(3)
H(1)
H(1)
S(1)
C(2)
C(3)
C(2)
C(3)
P(4)
C(3)
Ir(2)
O(3)
O(4)
O(3)
O(4)
O(4)
O(1)
O(2)
Ir(2)
48.6
138.1
48.5
48.1
137.6
49.3
48.0
134.7
51.4
77.9
79.0
80.9
89.5
89.5
86.8
In an attempt to gain information about the putative inter-
mediate, immediately preceding the C-H bond activation
process that yielded XIh, we moved the hydride ligand [within
the Ir(1)Ir(2)C(3) plane] to within bonding distance (1.13 Å)
of C(3), as found in other complexes containing an asymmetri-
cally bridged methyl group.57,58 For this structure (XIa) we did
not complete a Gaussian 94 calculation with S polarization
function included in the basis set. Otherwise, the complete range
of results on XIa is the same as that on XIh. Upon convergence
of XIa, which is only 4.8 kcal/mol higher in energy than XIh,
a dramatic shortening of the calculated Ir(1)-Ir(2) distance by
0.24 Å occurs, resulting in a distance that is now in line with
that observed in the structure of 11. However, the most dramatic
change is in the bonding of the SO2 ligand, which has moved
from an almost symmetrically bridged position in XIh to a
terminal position on Ir(1), leaving it bound in an η1-pyramidal
geometry [Ir(1)-S(1) ) 3.43 Å; Ir(2)-S(1) ) 2.64 Å]. The
resulting geometry at Ir(2), shown in Figure 7b and presented
in Table 9, is reminiscent of the SO2 adduct of Vaska’s
compound in which the Ir(2)-S(1) distance and the tilt of the
SO2-plane normal to the Ir-S vector (27.2°) are both compa-
rable with the X-ray-derived values (2.49(1) Å and 32(2)°,
respectively)59; we will comment later about the significance
of this calculation.
We have successfully located a transition state (labeled XI-
TS) relating the two structures XIh and XIa, at about 4.5 kcal/
mol higher in energy than that of XIa. The methodology we
used to locate this transition state is the same as was described
for III-TS (see ref 56). The structure of XI-TS basically
resembles the hydride species XIh, as shown in Figure 8c. The
major difference in the two can be seen in the hydride position,
which is much closer to the methylene group (1.94 vs 2.59 Å)
in the transition state. Other important bond lengths and angles
are given in Table 10.
97.1
97.4
99.3
126.5
177.5
173.4
144.0
29.4
127.1
177.5
173.1
143.4
29.7
128.8
177.3
173.9
144.4
29.5
70.2
71.7
74.7
157.2
48.4
159.7
48.9
163.3
48.5
87.0
87.9
88.6
118.6
179.0
154.4
61.1
122.2
122.2
100.0
100.0
114.3
179.6
176.2
83.1
120.7
177.6
151.4
60.2
123.2
177.7
148.2
57.4
122.5
122.1
100.6
100.6
114.3
179.6
176.1
81.8
121.3
121.3
103.8
103.8
117.1
178.8
176.7
80.1
a G94 ) Gaussian 94.
is the C(2)-Ir(2)-C(3) angle, which is computed to be near
175° but is observed at about 150°. One troubling aspect of not
including the phenyl groups on the diphosphines in the
calculations is the possibility that they may significantly
influence the positions of the equatorial ligands through
nonbonded contacts. For example, an examination of Figure 5
shows that the phenyl groups are thrust between the equatorial
ligands and thus are capable of a significant influence in the
positions of these ligands. A further discrepancy between the
theoretical models and the X-ray structure lies in the tilt of the
SO2 group with respect to Ir(2). In the X-ray structure the SO2
group is tilted slightly toward Ir(2) such that the angles between
the Ir(1)-S(1) and Ir(2)-S(1) vectors and the SO2-plane normal
are 54.2° and 51.6°. However in the Mulliken calculation, for
example, this asymmetry is exaggerated, giving vector-normal
angles of 60.8° and 37.4°. We do not currently understand the
reasons for the discrepancies between the observed and calcu-
lated structural parameters noted above.
(57) Jime´nez-Catan˜o, R.; Hall, M. B. Organometallics 1996, 15, 1889-
1897.
(58) Su, M.-D.; Chu, S.-Y. J. Am. Chem. Soc. 1997, 119, 5373-
5383.
(59) La Placa, S. J.; Ibers, J. A. Inorg. Chem. 1966, 5, 405.

Metal-Metal CooperatiVity and C-H Bond CleaVage
J. Am. Chem. Soc., Vol. 121, No. 15, 1999 3681
Scheme 4
The binding energy of SO2 in the diiridium complex 11 is
calculated as only 4.9 kcal/mol (i.e., the difference between the
energies of IIIt + SO2 and XIh). Although the other energies
resulting from our calculations are consistent with the energy
differences obtained from NMR studies, this binding energy
seems to be underestimated.
bridged intermediate analogous to Kubiak’s dmpm-bridged
compound. Significantly, the predominant structure of 3 in
solution is not that in the solid state; instead the “(µ-CO)-Ir-
Me” moiety appears to have twisted so that now both carbonyls
and the methyl group are terminally bound. This structure is
intermediate between both types of solid-state structures (dppm
and dmpm) in which either the carbonyl or the methyl group
bridges the metals. DFT calculations on the dhpm analogue of
3 converge to a structure (IIIt) which is essentially that proposed
in solution (3a). Two local minima have been obtained for a
methyl-bridged species; the one containing the symmetrically
bridged methyl group, exactly analogous to the Kubiak com-
pound, is 0.8 kcal/mol less stable than the unsymmetrically
bridged methyl species, in which the methyl group has an agostic
interaction with the adjacent metal, and this is in turn ca. 6 kcal/
mol less stable than that in which the methyl group is terminally
bound (structure IIIt). The fluxionality of 3 in solution and the
small energy differences obtained from the DFT calculations
on the different geometries suggest that the differences observed
in the solid state do not reflect electronic stabilization of the
different geometries by the two different phosphines, but more
likely result from subtle nonbonded contacts in the solid. DFT
calculations were also carried out on [Ir2H(CO)2(µ-CH2)-
(dhpm)2]+ (IIIh), the result of cleavage of the agostic C-H
interaction in IIIb (agostic), and a transition state between the
two was located. More will be said later about these structures
when we discuss the C-H bond-activation process.
In the reactions of 3 with the substrates CO, SO2, CNR, and
PR3, C-H bond cleavage of the methyl group occurs to yield
methylene-bridged hydride structures. Two mechanisms appear
possible for this C-H activation process: either C-H activation
is induced by ligand addition, or it precedes ligand addition
and the resulting methylene-bridged hydride is stabilized by
subsequent ligand coordination. Let us begin by considering
the former possibility in which ligand attack occurs on complex
3 (or one of its isomers) in which the methyl group is intact. In
the reactions of 3, two basic structural types result in which
the added ligand occupies a site in the product either adjacent
to (for CO, CNR, or PR3) or opposite (for SO2) the bridging
methylene unit. This difference is reminiscent of the different
sites of attack previously observed in A-frame species, in which
SO2 attack occurred in the A-frame pocket between the metals,
whereas other ligands attacked on the outside of the A-frame
structure, remote from the adjacent metal.60 We therefore suggest
that ligand attack occurs at either of two sites in an A-frame
species such as that shown in Scheme 4 for the agostic, methyl-
Discussion
As noted earlier, the reactions of the complexes [MM′(CO)3-
(dppm)2] (M, M′ ) Rh, Ir) with methyl triflate give rise to three
different classes of products, the natures of which are consistent
with the different properties of Rh and Ir. Possibly the simplest
of these reactions is that involving the mixed Rh/Ir complex,
in which simple electrophilic attack at the metals occurs yielding
a product in which the methyl group is terminally bound to Ir,
presumably reflecting the stronger Ir-C bond. For the dirhodium
species the greater tendency for migratory insertion of the lighter
congener gives rise to an acyl-bridged product. Reaction of the
diiridium complex with methyl triflate is the most novel in this
series, having resulted in facile, reversible C-H bond activation
of the methyl group, to give [Ir2(H)(CO)3(µ-CH2)(dppm)2][CF3-
SO3] (2). Again, the stronger Ir-C and Ir-H bonds, compared
with those of Rh, appear to favor the methylene-hydride species
in this case.
Although the above-mentioned tricarbonyl species 2 repre-
sents our entry point into the chemistry described in this article,
the theme of this article is aimed primarily at an investigation
of ligand addition to the dicarbonyl methyl compound [Ir2(CH3)-
(CO)2(dppm)2][CF3SO3] (3), obtained from 2 through loss of a
carbonyl. We have attempted to determine the factors that
influence the facile C-H bond cleavage of the methyl group
that occurs upon addition of small molecules (CO, PR3, CNR,
and SO2) to 3.
Before considering the reactivity of 3, however, it is instruc-
tive to first consider its structure, particularly with regard to
the bonding of the methyl group. Two analogous complexes
having essentially the same stoichiometry (one having different
substituents on the diphosphine and the other having different
metals) have been characterized. In the solid state the dirhodium
species, [Rh2(CH3)(CO)(µ-CO)(dppm)2]+, has a geometry iden-
tical with 3, having a terminal methyl group, whereas the dmpm
analogue, [Ir2(CO)2(µ-CH3)(dmpm)2]+, has a symmetrically
bridged methyl group. It was clearly of interest to establish
whether the different methyl-binding mode in the dmpm species
had resulted from the electronic influence of the diphosphine
ligands, because the binding mode of the methyl group
presumably influences its reactivity. In any case, it seems clear
that the different geometries do not differ appreciably in energy,
because in solution, compound 3 is fluxional, with the methyl
group transferring from metal to metal, presumably via a methyl-
(60) (a) Cowie, M. Inorg. Chem. 1979, 18, 286. (b) Mague, J. T.; Sanger,
A. R. Inorg. Chem. 1979, 18, 2060. (c) George, D. S. A.; McDonald, R.;
Cowie, M. Organometallics 1998, 17, 2553.

3682 J. Am. Chem. Soc., Vol. 121, No. 15, 1999
Torkelson et al.
bridged structure [structure IIIb (agostic)]. Under such a scheme
ligand attack at site Y or Z could occur, followed (at some stage)
by cleavage of the C-H bond. This proposal is consistent with
the DFT calculations that model the reVerse transformation. In
these calculations movement of the hydride ligand in the
methylene-bridged hydride product (XIh) to give a bridged
agostic methyl group also resulted in movement of the SO2
ligand away from the metal at which C-H activation occurred,
to a terminal position on the adjacent metal. As noted, the
resulting geometry of the SO2 ligand is pyramidal (as in the
SO2 adduct of Vaska’s compound), having the electron lone
pair of the pyramidal SO2 pointed in the direction of the adjacent
metal. Presumably, the reverse movement of the SO2 ligand
toward the adjacent metal results in donation of the SO2 lone
pair to this metal, resulting in an increase in the electron density
at this metal and concomitant C-H bond cleavage of the methyl
group. It is significant that under this scheme, initial ligand
addition to one metal results in C-H activation at the adjacent
metal via transfer of an electron pair from the first metal to the
electrophile (SO2), and subsequent donation of this pair to the
second metal. Again, a transition state between the “agostic”
intermediate XIa and the product of C-H bond cleavage XIh
has been located at only 4.5 kcal/mol above XIa and 9.3 kcal/
mol above XIh. It is encouraging that these energy differences
are in line with the very low barrier of 6.1 kcal/mol obtained
from NMR studies (vide supra).
performed, the activation barrier to C-H cleavage [starting from
the agostic structures IIIb (agostic) and XIb] is substantially
less with SO2 coordinated than without (ca. 5 vs 15 kcal/mol),
indicating that this process is favored by SO2 addition. Second,
the experimental NMR evidence of facile and reversible
exchange of the methylene hydrogens and the hydride ligand
in the CO (2) and SO2 (11) adducts demonstrates that reversible
C-H bond cleavage/formation is occurring in these adducts and
therefore argues strongly for the second mechanism.
Conclusions
Facile C-H bond cleavage of the methyl group in [Ir2(CH3)-
(CO)2(dppm)2]+ occurs upon addition of several small molecules
such as CO, CNtBu, phosphines, triphenyl phosphite, and SO2,
yielding either [Ir2H(CO)2L(µ-CH2)(dppm)2]+ (L ) CO, CNt-
Bu, PR3) or [IrH(CO)2(µ-CH2)(µ-SO2)(dppm)2]+. Although
C-H activation has been observed previously upon addition
of phosphines,45 it has not, to our knowledge, been observed
upon addition of strong π-acid ligands such as CO and SO2. In
all cases reported herein, we propose that C-H bond cleavage
at one metal results from ligand binding at the adjacent metal.
In this regard, SO2 is unique in this study. It functions first as
an electrophile and subsequently as a nucleophile, acting as an
electron conduit, accepting an electron pair from one metal and
passing it on to the adjacent metal. The resulting electron density
increase at the second metal presumably results in C-H bond
cleavage by donation into the antibonding orbital of the C-H
bond that is involved in an agostic interaction with the adjacent
metal. For other ligands the metal at which addition occurs
attains an 18-valence electron count leading to dative metal-
metal bond formation to the adjacent metal. It is this dative
bond that supplies the necessary electron density to cleave the
C-H bond. We are unaware of any other system in which such
metal-metal cooperativity effects result in C-H bond activa-
tion.
When ligand attack is on the outside of the A-frame complex
IIIb (agostic) (site Z in Scheme 4), as with phosphines and
isocyanides, this electronic influence must be transmitted to the
second metal via the metal-metal framework, because the added
ligand is not in a position to interact directly with this adjacent
metal. In this case we propose that addition at site Z gives an
18e center which can result in formation of a dative Ir f Ir
bond, which supplies the additional electron density at the
adjacent metal, leading to C-H bond cleavage.
As in other A-frame complexes, we assume that CO attack
occurs at the outside site (Z), much as for PR3 and CNR groups.
Irrespective of site of attack, however, the net result of attack
at one metal is C-H bond cleavage by the other metal. The
major difference in attack at site Y or Z is whether the electronic
effect of ligand coordination at one metal is transmitted to the
adjacent metal via the attacking ligand, which assumes a
bridging coordination mode, or via a direct metal-metal
interaction. In this scheme the preference of SO2 for site Y is
presumably thermodynamic, resulting from the preference of
the electrophile to access the electron density of both metals
and therefore be in the bridging site, whereas the preference of
other ligands for site Z is assumed to be kinetic, having less
steric repulsions in this site at one end of the complex.
This study has also demonstrated that the activation barrier
to methyl group C-H bond cleavage can be unusually low in
these binuclear systems. The value observed for interconversion
between methyl and methylene groups (∆H‡ ) 6 kcal/mol) in
the SO2 adduct, [Ir2H(CO)2(µ-CH2)(µ-SO2)(dppm)2][CF3SO3],
is in fact very close to values of between 6 and 8 kcal/mol
observed for platinum-surface-bound alkyls.61 This supports
our contention that binuclear complexes can function as useful
models for adjacent-metal involvement in heterogeneous sys-
tems. As stated by Fehlner and co-workers,62 “...reaction paths
depend on small energy differences that can induce large
changes in rates...if one wants to model catalytic reactivity with
cluster chemistry, small barrier processes must be understood.”
In this study we have developed an improved understanding of
the function of the adjacent metals in the facile and reversible
C-H bond cleavage process of a methyl group.
An alternate mechanism for methyl-group activation involves
C-H bond cleavage in the precursor 3, before ligand attack, to
give small, undetectable concentrations of an intermediate such
as IIIh (Figure 7). The final products of ligand attack on this
intermediate would result from coordination at a vacant site on
the adjacent metal. The same factors discussed earlier would
favor SO2 attack at the inside site, leading to a bridging mode,
whereas the other ligands would favor attack at the less sterically
encumbered external site. Certainly the energy of this intermedi-
ate (IIIh) is calculated as only 3.4 kcal/mol less stable than
that of the agostically bound IIIb, therefore it is accessible.
However, there are two aspects of this study that lead us to
favor the first mechanism of C-H bond cleavage over the
second; i.e., C-H bond cleavage induced by ligand addition,
rather than before ligand addition. First, in the calculations
Acknowledgment. We thank the Natural Sciences and
Engineering Research Council of Canada (NSERC) and the
University of Alberta for financial support of the research and
for scholarships (to J.R.T.), and NSERC for funding the P4/
RA diffractometer. The Calvin College contingent wishes to
thank the college for a sabbatical leave to RDK, the donors of
the Petroleum Research Fund, administered by the American
Chemical Society, and the Camille and Henry Dreyfus Founda-
tion for a Scholar/Fellow award to undergraduate institutions.
(61) Zaera, F. Acc. Chem. Res. 1992, 25, 260.
(62) Dutta, T. K.; Vites, J. C.; Jacobsen, G. B.; Fehlner, T. P.
Organometallics 1987, 6, 842.

Metal-Metal CooperatiVity and C-H Bond CleaVage
J. Am. Chem. Soc., Vol. 121, No. 15, 1999 3683
We also thank Dr. R. E. D. McClung and G. Aarts for assistance
with the NMR saturation transfer experiments and the line-shape
analysis on compounds 2 and 11, respectively, and Dr. McClung
for many helpful discussions.
coordinates, bond distances and angles, and thermal parameters
for compounds 3, 6 and 11, Tables of atom coordinates from
DFT calculations for structures IIIt, IIIb (agostic), IIIb
(symmetric), IIIh, III-TS, XIh, XIa, and XI-TS. This material
is available free of charge via the Internet at http://pubs.acs.org.
Supporting Information Available: Listings of crystal data,
data collection, solution and refinement, complete atomic
JA983525O