Metallaborane reactivity. A stoichiometric mechanism for the insertion of two alkynes into an iridaborane framework via a disposable molybdenum chaperone

Article abstract of DOI:10.1021/ja068999z

Building on earlier work that showed the formation of [1-Cp*-2,2,2- (CO)₃-2-THF-nido-1,2-IrMoB₄H₈], 2, from the reaction of [1-Cp*-arachno-1-IrB₄H₁₀], 1, with (arene)Mo(CO)₃, the stoichiometric mechanism for the generation of [1-Cp*-5,6,7,8-(R)₄-nido-1,5,6,7,8-IrC₄B ₃H₃], 8, from the reaction of 2 with RC≡CR, R = Me, Ph, has been identified. For R = Me, the major product in solution is [1-Cp*-5,6,7,8-(CH₃)₄-closo-1,5,6,7,8-IrC ₄B₃H₃Mo(CO)₃], 7, which is in equilibrium with 8. The equilibrium 8 + Mo(THF)₃-(CO)₃ ? 7 + 3THF is characterized by ΔH = 8 kcal/mol and ΔS = 34 cal/mol K. Density functional theory calculations carried out on 7 indicate that the Mo(CO)₃ moiety is weakly bound to the cluster mainly through Mo-C rather than Mo-B interactions. Under alkyne deficient conditions, the product [1-Cp*-2,2,2-(CO)₃-(μ-CO)-3,4-(CH₃) ₂-closo-1,2,3,4-IrMoC₂B₃H₃], 6, can be isolated. Solid-state structures of 1 and 2 have been reported previously, and those of 6, 7, and 8, R = Me, Ph, are reported here. The evolution of products with time was monitored by ¹H and ¹¹B NMR and showed the formation and decay of two additional species which have been identified as the structural isomers [1-Cp*-7,7,7-(CO)₃-7-THF- 2,3-(CH₃)₂-nido-1,7,2,3-IrMoC₂B ₃H₅], 4, and [5-Cp*-7,7,7-(CO)₃-7-THF-2, 3-(CH₃)₂-nido-5,7,2,3-IrMoC₂B₃H ₅], 5, with the metals nonadjacent in 4 and adjacent in 5. Circumstantial evidence suggests that 4 is the precursor to 5 and 5 is the precursor to both 6 and 7. Cluster 2 also is a catalyst or catalyst precursor for the isomerization of olefins, namely, hex-1-ene to cis-hex-2-ene and tetramethyl allene to 2,4-dimethylpenta-1,3-diene. These novel results also establish that [1-Cp*-2,2,2-(CO)₃-2-(alkyne)-nido-1,2- IrMoB₄H₈], 3, forms from 2 and constitutes a logical precursor to 4. The entire process, 1 + 2alkyne = 8 + BH₃ + 2H ₂, which is promoted by (arene)Mo(CO)₃, constitutes an explicit example of a transition-metal-facilitated process analogous to metal-facilitated organic transformations observed in organometallic chemistry.

Full text of DOI:10.1021/ja068999z

Published on Web 02/24/2007
Metallaborane Reactivity. A Stoichiometric Mechanism for the
Insertion of Two Alkynes into an Iridaborane Framework via a
Disposable Molybdenum Chaperone
Fre´de´ric de Montigny,† Ramo´n Macias,† Bruce C. Noll,† Thomas P. Fehlner,*,†
Karine Costuas,‡ Jean-Yves Saillard,‡ and Jean-Franc¸ois Halet‡
Contribution from the Department of Chemistry and Biochemistry, UniVersity of Notre Dame,
Notre Dame, Indiana 46556, and Sciences Chimiques de Rennes, UMR 6226 CNRS-UniVersite´
de Rennes 1, AVenue du Ge´ne´ral Leclerc, F-35042 Rennes cedex, France
Received December 15, 2006; E-mail: fehlner.1@nd.edu
Abstract: Building on earlier work that showed the formation of [1-Cp*-2,2,2-(CO)3-2-THF-nido-1,2-
IrMoB4H8], 2, from the reaction of [1-Cp*-arachno-1-IrB4H10], 1, with (arene)Mo(CO)3, the stoichiometric
mechanism for the generation of [1-Cp*-5,6,7,8-(R)4-nido-1,5,6,7,8-IrC4B3H3], 8, from the reaction of 2 with
RCtCR, R ) Me, Ph, has been identified. For R ) Me, the major product in solution is [1-Cp*-5,6,7,8-
(CH3)4-closo-1,5,6,7,8-IrC4B3H3Mo(CO)3], 7, which is in equilibrium with 8. The equilibrium 8 + Mo(THF)3-
(CO)3 a 7 + 3THF is characterized by ∆H ) 8 kcal/mol and ∆S ) 34 cal/mol K. Density functional theory
calculations carried out on 7 indicate that the Mo(CO)3 moiety is weakly bound to the cluster mainly through
Mo-C rather than Mo-B interactions. Under alkyne deficient conditions, the product [1-Cp*-2,2,2-(CO)3-
(µ-CO)-3,4-(CH3)2-closo-1,2,3,4-IrMoC2B3H3], 6, can be isolated. Solid-state structures of 1 and 2 have
been reported previously, and those of 6, 7, and 8, R ) Me, Ph, are reported here. The evolution of products
with time was monitored by 1H and 11B NMR and showed the formation and decay of two additional species
which have been identified as the structural isomers [1-Cp*-7,7,7-(CO)3-7-THF-2,3-(CH3)2-nido-1,7,2,3-
IrMoC2B3H5], 4, and [5-Cp*-7,7,7-(CO)3-7-THF-2,3-(CH3)2-nido-5,7,2,3-IrMoC2B3H5], 5, with the metals
nonadjacent in 4 and adjacent in 5. Circumstantial evidence suggests that 4 is the precursor to 5 and 5 is
the precursor to both 6 and 7. Cluster 2 also is a catalyst or catalyst precursor for the isomerization of
olefins, namely, hex-1-ene to cis-hex-2-ene and tetramethyl allene to 2,4-dimethylpenta-1,3-diene. These
novel results also establish that [1-Cp*-2,2,2-(CO)3-2-(alkyne)-nido-1,2-IrMoB4H8], 3, forms from 2 and
constitutes a logical precursor to 4. The entire process, 1 + 2alkyne ) 8 + BH3 + 2H2, which is promoted
by (arene)Mo(CO)3, constitutes an explicit example of a transition-metal-facilitated process analogous to
metal-facilitated organic transformations observed in organometallic chemistry.
Introduction
has permitted a focus on reactivity with an eye to comparisons
with organometallic chemistry. Perhaps the epitome of the latter
The classic development of a subarea of chemistry begins
with the discovery of a new compound type and proceeds
through synthetic improvements and structural development to
systematic examination of reactivity. The vagaries of funding
and the “herd instincts” of chemists determine where the process
peters out for any given area. But it is unfortunate when
reactivity is totally neglected as it is in the delineation of
mechanistic pathways that true understanding and some measure
of control of electronic and geometric properties are achieved.
All the bombast of application, application, application obscures
the fact that this understanding is fundamental to the develop-
ment of rational solutions to the tough practical problems
modern life presents.
is the utilization of metals to control the derivatization of an
organic moiety in stoichiometric reactions or catalytic cycles.9
Adding metals to carbon chemistry permits the modest barriers
and small thermodynamic changes required for such chemistry
to be achievedsdeep energy wells brings chemistry to a halt
and high intrinsic barriers increase the probability of competitive
side reactions. Adding metals to boron chemistry should do the
same; and the stepwise binding, modification, and release of a
borane substrate is an attractive goal with potential practical
(1) Barton, L.; Srivastava, D. K. In ComprehensiVe Organometallic Chem. II;
Abel, E., Stone, F. G. A., Wilkinson, G., Eds.; Pergamon: New York,
1995; Vol. 1.
(2) Braunschweig, H. Angew. Chem., Int. Ed. 1998, 37, 1787.
(3) Fehlner, T. P. Organometallics 2000, 19, 2643.
(4) Kawano, Y.; Yasue, T.; Shimoi, M. J. Am. Chem. Soc. 1999, 121, 11744.
(5) Kennedy, J. D. Prog. Inorg. Chem. 1984, 32, 519.
(6) Kennedy, J. D. Prog. Inorg. Chem. 1986, 34, 211.
(7) Grimes, R. N. In Metal Interactions with Boron Clusters; Grimes, R. N.,
Ed.; Plenum: New York, 1982; p 269.
For the past couple of decades we, and others, have worked
on transition metal-borane cluster compounds,1-8 but it is only
in the past decade that development of the synthetic chemistry
(8) Coffy, T. J.; Medford, G.; Plotkin, J.; Long, G. J.; Huffman, J. C.; Shore,
S. G. Organometallics 1989, 8, 2404.
(9) Elschenbroich, C.; Salzer, A. Organometallics; VCH: New York, 1989.
† University of Notre Dame.
‡ University of Rennes.
9
3392
J. AM. CHEM. SOC. 2007, 129, 3392-3401
10.1021/ja068999z CCC: $37.00 © 2007 American Chemical Society

Stoichiometric Metallaborane Reaction Mechanism
A R T I C L E S
payoff. We describe one such system in which the modification
of a metallaborane is mediated by coordination to a molybdenum
carbonyl fragment.
rane, [1-Cp*-2,2,2-(CO)3-2-THF-nido-1,2-IrMoB4H8], 2, pos-
sessed a weakly coordinated THF molecule.25 Normally such a
metallaborane would lose the THF ligand and form a closo
cluster; however, an octahedral MM′B4 skeleton is not well
suited to accommodate the four bridging hydrogen atoms
required to meet the seven skeletal electron pair (sep) cluster
count. As a consequence 2 can be isolated and readily undergoes
displacement of THF with a variety of Lewis bases.25 Thus, 2
is a potential reaction partner for the incorporation of alkynes,
i.e., it possesses both an accessible alkyne binding site as well
as “extra” skeletal H atoms. It does react, and the essentials of
the overall reaction were communicated earlier.26 We now
present the full story including that of intermediates as well as
experiments that reveal some of the early mechanistic details.
Important earlier studies suggest that the present findings are
part of a growing body of reaction chemistry supporting the
thesis outlined above. Thus, we have the elegant series of papers
in which a rhodacarborane serves as a homogeneous hydrogena-
tion catalyst for organic substrates.10 Closely related is the
utilization of transition metal catalysts to effect derivatization
of boranes and carboranessa process that involves metal-
borane interactions.11 Properties of compounds containing
metal-boron bonds are critical in the creation of B-C bonds
to effect the functionalization of hydrocarbons in stoichiometric
and catalytic processes.12-14 Equally fascinating is the hydrobo-
ration of an unsaturated triosmium cluster to a metal cluster
substituted boroxime ring versus an osmaborane.15,16
Results and Discussion
Scheme 1 can be used as a guide to the following sections.
The synthesis as well as selected reactivity of 1 has been
described earlier.27,28 The first reaction step, 1 f 2, as well as
the structure of 2 and displacement of THF by a variety of Lewis
bases is also the principal topic of an earlier paper.25 We begin
here with the isolated products, 6, 7, and 8 and then continue
with the less precisely characterized intermediates 4 and 5.
Finally, reactions of 2 with olefins, which are unlikely to insert,
are used to refine understanding of the initial alkyne adduct 3.
Isolated Products. The reaction of 2 with excess RCtCR,
R ) Me, Ph, followed by product extraction and chromatog-
raphy yielded a single pure product. Solid-state X-ray structure
determinations revealed that the new compounds are the eight-
vertex nido-iridacarboranes, [1-Cp*-5,6,7,8-(R)4-nido-1,5,6,7,8-
IrC4B3H3], 8, with structures shown in Figure 1 (R ) Me) and
Figure 1 in the Supporting Information (R ) Ph). The latter
suffers from the presence of unresolved solvent(s); however, it
is sufficient to establish the essential similarity of the Me and
Ph derivative structures. The structures in the solid state are
consistent with the spectroscopic data and similar to those of
known examples of the compound type, e.g., the cobaltacar-
borane, [Cp*CoC4Ph4B3H3], as well as main group analogs.29-31
Simple application of the Wade/Mingos approach predicts a
cluster structure based on a tricapped trigonal prism with either
a five-connect or four-connect vertex unoccupied.32-35 However,
8 is another example of an eight-vertex, 10 skeletal electron
pairs (sep) cluster that adopts a more open structure. Williams
has provided one rationalization (maximized charge smooth-
ing),36 but it now appears clear that the number of bridging H
atoms is also an important factor in determining the form of
the 10 sep/8 fragment cluster shape.37,38
In exploring the reactivity of a class of metallaboranes3
containing Cp*M (Cp* ) η5-C5Me5, M ) Cr, Mo, W, Re, Ru,
Co, Rh, Ir) fragments with alkynes, we were surprised to observe
facile reactivity involving the borane fragment only in the case
of nido-(Cp*RuH)2B3H7. For this compound, examination of
the dependence of intermediates and final products on alkyne
substituents as well as related experiments revealed the sto-
ichiometric mechanism and the source of the variety of
metallacarborane products.17-22 The key structural feature of
nido-(Cp*RuH)2B3H7 relative to isostructural and isoelectronic
nido-(Cp*Rh)2B3H7, for which cyclotrimerization alone was
observed,23 is the “extra” two hydrogen atoms present in the
ruthenaborane. These foster facile hydrometallation and/or ready
incorporation of the alkyne carbon atoms into the cluster
framework. Although the mechanistic evidence suggests both
rhoda- and ruthenaboranes reversibly add alkyne in a first step,
only the ruthenaborane with its “extra” hydrides provides a
pathway to alkyne incorporation into the cluster framework.
However, “extra” hydrides are not a sufficient condition for
alkyne incorporation. For example, the iridaborane, [1-Cp*-
arachno-1-IrB4H10], 1,24 with six H atoms on the framework
does not react with alkynes under mild conditions. Although
other reasons are possible, given the high reactivity of alkynes,
a low iridaborane/alkyne association constant is a likely cause.
Subsequent study of this iridaborane suggested a test. That is,
we found that a group 6 metal fragment could be efficiently
incorporated into 1 and that the resulting nido-dimetallahexabo-
(10) Belmont, J. A.; Soto, J.; King, R. E., III; Donaldson, A. J.; Hewes, J. D.;
Hawthorne, M. F. J. Am. Chem. Soc. 1989, 111, 7475.
(11) Kadlecek, D. E.; Carroll, P. J.; Sneddon, L. G. J. Am. Chem. Soc. 2000,
122, 10686.
(12) Chen, H.; Hartwig, J. F. Angew. Chem., Int. Ed. 1999, 38, 3391.
(13) Chen, H.; Schlecht, S.; Semple, T. C.; Hartwig, J. F. Science 2000, 287,
1995.
(14) Duan, Z.; Hampden-Smith, M. J.; Sylwester, A. P. Chem. Mater. 1992, 4,
1146.
(15) Shore, S. G.; Jan, D.-Y.; Hsu, W.-L.; Hsu, L.-Y.; Kennedy, S.; Huffman,
J. C.; Wang, T.-C. L.; Marshall, A. G. J. Chem. Soc., Chem. Commun.
1984, 392.
(16) Chung, J.-H.; Knoeppel, D.; McCarthy, D.; Columbie, A.; Shore, S. G.
Inorg. Chem. 1993, 32, 3391.
(17) Yan, H.; Beatty, A. M.; Fehlner, T. P. J. Am. Chem. Soc. 2002, 124, 10280.
(18) Yan, H.; Beatty, A. M.; Fehlner, T. P. Angew. Chem., Int. Ed. 2001, 40,
4498.
(19) Yan, H.; Beatty, A. M.; Fehlner, T. P. Angew. Chem., Int. Ed. 2002, 41,
2578.
(20) Yan, H.; Beatty, A. M.; Fehlner, T. P. J. Organomet. Chem. 2003, 680,
66.
(21) Yan, H.; Beatty, A. M.; Fehlner, T. P. J. Am. Chem. Soc. 2003, 125, 16367.
(22) Yan, H.; Noll, B. C.; Fehlner, T. P. J. Am. Chem. Soc. 2005, 127, 4831.
(23) Yan, H.; Beatty, A. M.; Fehlner, T. P. Organometallics 2002, 21, 5029.
(24) Lei, X.; Shang, M.; Fehlner, T. P. Chem. Eur. J. 2000, 6, 2653.
(25) Mac´ıas, R.; Fehlner, T. P.; Beatty, A. M.; Noll, B. Organometallics 2004,
23, 5994.
(26) de Montigny, F.; Macias, R.; Noll, B.; Fehlner, T. P. Angew. Chem., Int.
Ed. 2006, 45, 2119.
(27) Lei, X.; Bandyopadhyay, A. K.; Shang, M.; Fehlner, T. P. Organometallics
1999, 18, 2294.
(28) Mac´ıas, R.; Fehlner, T. P.; Beatty, A. M. Angew. Chem., Int. Ed. 2002, 41,
3860.
(29) Zimmerman, G. J.; Sneddon, L. G. Inorg. Chem. 1980, 19, 3650.
(30) Micciche, R. P.; Briguglio, J. J.; Sneddon, L. G. Organometallics 1984, 3,
1396.
(31) Wrackmeyer, B.; Schanz, H.-J.; Hofmann, M.; Schleyer, P.; Boese, R. Eur.
J. Inorg. Chem. 1999, 533.
(32) Wade, K. Inorg. Nucl. Chem. Lett. 1972, 8, 559.
(33) Wade, K. AdV. Inorg. Chem. Radiochem. 1976, 18, 1.
(34) Mingos, D. M. P. Nature (London), Phys. Sci. 1972, 236, 99.
(35) Mingos, D. M. P.; Wales, D. J. Introduction to Cluster Chemistry; Prentice
Hall: New York, 1990.
(36) Williams, R. E. In The Borane, Carborane, Carbocation Continuum;
Casanova, J., Ed.; Wiley-Interscience: New York, 1997; p 3.
9
J. AM. CHEM. SOC. VOL. 129, NO. 11, 2007 3393

A R T I C L E S
Scheme 1
de Montigny et al.
As 8 reversibly coordinates to a Mo center (see below), the
C-C distances of 8 are of interest. That for the C2 unit not
bonded to Ir in 8 is similar to that in its cobaltacarborane
analogue (1.379(5) Å) as well as those in the isolobal main
group analogue, C4Et4B4H4 (1.380(3) and 1.385(3) Å). That of
the C2 unit bound to Ir is slightly longer. As the angles around
the carbon atoms in these clusters approach 120°, the skeletal
carbons have sp2 character similar to that of the C-C interaction
in nido-2,3-C2B4H8 and the B-B interaction in B6H10. Both of
the latter compounds exhibit Lewis basicity, and cluster 8 is
expected to do so as well. This conclusion is useful in explaining
the reversible coordination of 8 to Mo described below.
With limited alkyne the same reaction generated a small yield
of a different compound 6 in addition to 8. Addition of alkyne
to pure 6 under similar reaction conditions did not generate 8;
hence, 6 is a “dead-end” intermediate. Nonetheless, the NMR
data for 6 are closely related to those of 8, R ) Me. There are
two exceptions: the Me resonance integrates to 6 protons rather
than 12, and the infrared spectrum shows metal CO bands
including one absorption frequency consistent with a bridging
carbonyl ligand. The spectrometric data suggest a composition
Cp*IrMo(CO)4B3H3C2R2 with 8 sep. A closo seven-atom cluster
containing Mo and Ir is predicted, and the solid-state structure
in Figure 2 shows this to be the case. The observed geometric
isomer, [1-Cp*-2,2,2-(CO)3(µ-CO)-3,4-(CH3)2-closo-1,2,3,4-
IrMoC2B3H3], 6, possesses a plane of symmetry containing B(2),
Mo, and Ir. The carbon atoms are adjacent in the equatorial
ring and Mo and B atoms occupy the apical positions. One of
the four CO ligands bridges the Mo-Ir edge between the Mo-
(CO)3 and Cp*Ir fragments. For comparison with 8, note that
the C-C distance in 6 is considerably longer at 1.508 Å.
Penultimate Product. When the reaction mixture for R )
Me was analyzed by NMR before workup, we were surprised
to find that the signals due to 8 constituted a small fraction of
the total intensity observed even though reactant 2 had been
consumed. The major product was another species that had a
(37) King, R. B. Inorg. Chem. 2001, 40, 6369.
(38) Mavunkal, I. J.; Noll, B. C.; Meijboom, R.; Muller, A. Fehlner, T. P.
Organometallics 2006, 25, 2906.
9
3394 J. AM. CHEM. SOC. VOL. 129, NO. 11, 2007

Stoichiometric Metallaborane Reaction Mechanism
A R T I C L E S
Figure 3. Molecular structure of [1-Cp*-5,6,7,8-(CH3)4-closo-1,5,6,7,8-
IrC4B3H3Mo(CO)3], 7. Selected bond lengths [angstrom] and interatomic
angles [deg]: Ir(1)-B(1) 2.053(2), Ir(1)-C(5) 2.1433(13), Ir(1)-B(2)
2.3077(16), Mo(1)-C(3) 2.2355(13), Mo(1)-C(5) 2.3884(13), Mo(1)-B(2)
2.5716(16), Mo(1)-C(1) 1.999(2), Mo(1)-C(2) 1.9525(15), O(1)-C(1)
1.149(3), O(2)-C(2) 1.1603(18), C(5)-C(5) 1.448(3), C(3)-C(3) 1.451-
(3), B(1)-C(3) 1.722(2), B(1)-B(2) 1.974(2), B(2)-C(5) 1.613(2), B(2)-
C(3) 1.641(2), B(1)-B(2) 1.974(2), B(1)-Ir(1)-C(5) 86.23(7), C(5)-Ir(1)-
C(5) 39.47(7), B(1)-Ir(1)-C(9) 99.62(7), B(1)-Ir(1)-B(2) 53.45(6), C(3)-
B(1)-C(3) 49.83(11), C(3)-B(1)-B(2) 89.02(11), B(2)-B(1)-B(2)
90.05(13), C(3)-C(3)-C(4) 122.18(8), C(3)-C(3)-B(2) 114.12(7), C(5)-
C(5)-C(6) 123.04(8), C(5)-C(5)-B(2) 114.64(7), C(6)-C(5)-B(2) 122.11-
(12).
Figure 1. Molecular structure of [1-Cp*-5,6,7,8-(R)4-nido-1,5,6,7,8-
IrC4B3H3], 8, R ) Me. Selected bond lengths [angstrom] and interatomic
angles [deg]: Ir1-B2 2.089(2), Ir1-B3 2.305(2), Ir1-B4 2.289(2), Ir1-
C5 2.1619(19), Ir1-C6 2.1619(19), B2-B3 1.953(3), B2-B4 1.943(3),
C5-B3 1.572(3), C8-B3 1.603(3), C7-B4 1.600(3), C6-B4 1.577(3),
C5-C6 1.414(3), C7-C8 1.384(3), C5-B3-C8 113.13(17), C7-B4-C6
113.04(18), B3-C5-C51 121.51(18), B3-C5-C6 115.64(17), C51-C5-
C6 122.84(18), C51-C5-Ir1 123.64(14), B3-C8-C81 120.95(17), B3-
C8-C7 115.30(17), C81-C8-C7 123.62(19), C81-C8-B2 129.14(19).
Scheme 2
The solid-state structure of 7 is shown in Figure 3 and Scheme
1, where it can be seen that [1-Cp*-5,6,7,8-(CH3)4-closo-
1,5,6,7,8-IrC4B3H3Mo(CO)3], 7, is indeed closely related to 8s
a Mo(CO)3 fragment has simply been added to the open face
of 8. As expected from the sep count, this is a closo cluster but
is only so provided the Mo-B distance of 2.572(2) Å is
considered bonding (more about this point below). If considered
fully closed, the cluster geometry exhibited is not the canonical
tricapped trigonal prismatic shape of the [B9H9]2- borane
paradigm but rather a shape derived from it by two diamond-
square-diamond rearrangements (Scheme 2). The six-connect
vertex thereby generated is occupied by the Mo atom.
Figure 2. Molecular structure of [1-Cp*-2,2,2-(CO)3(µ-CO)-3,4-(CH3)2-
closo-1,2,3,4-IrMoC2B3H3], 6. Selected bond lengths [angstrom]: Ir(1)-
C(7) 2.138(3), Ir(1)-B(1) 2.146(2), Ir(1)-B(2) 2.214(3), Ir(1)-Mo(1)
2.8751(6), Mo(1)-C(7) 2.152(3), Mo(1)-C(10) 2.3185(17), Mo(1)-B(1)
2.461(2), O(1)-C(7) 1.183(3).
Considering the easily reversed formation of 7 from 8, the
iridatetracarbon-metallacarborane is usefully viewed as a
complex ligand that can displace the six-electron arene in
(arene)Mo(CO)3 or, as the arene is probably largely lost in THF,
the three THF ligands of Mo(THF)3(CO)3. The residual unsat-
uration of the C-C interactions in 8 mentioned above now
becomes relevant. The C-C distances in 8 (1.38 and 1.41 Å)
increase to 1.45 Å on binding the Mo(CO)3 fragment to give 7.
Compare this to the C-C distance of 1.51 Å in 6 where the
alkyne is strongly retained in the cluster network. There is a
small increase of shielding of the alkyne carbons of 7 (δ 133.5,
133.4 ppm) on coordination to the Mo center (δ 129.6, 129.7
ppm). If the two “enes” are exclusive donors to the Mo atom
then we have a formal 16-electron Mo complex. Hence, the
Mo-B interaction is important in determining how to view
coordination of 8 to Mo. The 11B NMR shift change on
spectroscopic signature similar to that of 8 but one that retained
the molybdenum carbonyl fragment of 2. The spectroscopic
characterization of mixtures of the new product with 8 was
consistent with a composition Cp*IrB3H3C4R4Mo(CO)3, 7. This
cluster possesses 10 sep if the Mo(CO)3 fragment is counted as
a zero-electron fragment. Hence, with nine cluster fragments
the counting rules suggest a closo geometry.
All initial attempts to isolate pure 7 by chromatography for
a structure determination resulted solely in the isolation of 8.
Reactivity of 7 observed by NMR provided the key to solving
the problem. Addition of PPh3, toluene, or CH3CN resulted in
the loss of 7 and the formation of 8. On the other hand, addition
of excess (arene)Mo(CO)3 resulted in the nearly quantitative
formation of 7. It was from the latter reaction mixture that single
crystals of 7 were finally obtained.
9
J. AM. CHEM. SOC. VOL. 129, NO. 11, 2007 3395

A R T I C L E S
de Montigny et al.
coordination of 8 is -10.4 to -10.6 for the two B atoms bonded
to Mo, whereas it changes from 11.6 to 33.1 for the one B atom
that is not. This is not helpful.
Quantum Chemical Analysis. Neither the structural metrics
nor the spectroscopic shifts on coordination are definitive
concerning the nature of the Mo-B interaction in 7. If bonding,
7 is properly described as a closo cluster and can be viewed as
an unusual cluster coordination compound with an 18-electron
Mo center. If nonbonding, it would be described as a hypo-
electronic open cluster and a cluster coordination compound
with a 16-electron Mo center.39 The nature of the Mo-B
bonding must be addressed if one is to understand why the
iridacarborane 8 is so weakly bound to the Mo fragment in 7.
Hence, 7, as well as 6 and 8 for comparison, were investigated
by quantitative quantum chemical methods at the density
functional theory (DFT) level (see Computational Details).
Models [1-Cp-2,2,2-(CO)3(µ-CO)-3,4-(CH3)2-closo-1,2,3,4-
IrMoC2B3H3] (6′), [1-Cp-5,6,7,8-(CH3)4-closo-1,5,6,7,8-IrC4B3H3-
Mo(CO)3] (7′), and [1-Cp*-5,6,7,8-(R)4-nido-1,5,6,7,8-IrC4B3H3]
(8′) in which the Cp* ligand was replaced by Cp were used in
order to reduce computational effort. These models were first
optimized and compared to X-ray structures previously described
(Cartesian coordinates of the optimized geometries are given
in the Supporting Information).
Model 8′ within Cs symmetry, i.e., with the Cp and B3C2
(B(2)B(3)B(4)C(5)C(6)) rings eclipsed, is computed to be nearly
isoenergetic (0.3 kJ/mol more stable) to the structure where the
Cp and the B3C2 rings are slightly staggered with a torsion angle
of 14°, close to that observed experimentally in 8 (12°). This
slight staggering is probably due to weak steric hindrance
between the Me groups of Cp* and the R groups attached to
the carbon atom of the B3C2 ring. The metrical data computed
for 8′ are in good agreement with those experimentally measured
for 8. The largest bond distance and bond angle deviations are
0.02 Å and 2°, respectively. A good agreement between theory
and experiment is also observed between 6 and 6′. The main
discrepancy is for the Mo-C(5) bond (see Figure 2 for atom
labeling), which is computed to be 0.04 Å longer than that
measured experimentally. Model 7′ also quite satisfactorily
mimics the arrangement of the crystallographically characterized
cluster 7 with a slight overestimation of the Mo-C and Mo-B
distances (0.03-0.05 Å for Mo-C, 0.03 Å for Mo-B).
The long Mo-B separations, which we seek to interpret, are
well reproduced in 7′ (2.602 vs 2.572 Å experimental) and the
normal separations are also reproduced in 6′ (2.459 vs 2.461 Å
experimental). Do the calculations also show a difference in
the Mo(CO)3-cluster binding energies? Investigation of these
interactions utilizes an energy-partitioning analysis.40,41 This
approach allows one to estimate the interaction energy (∆Eint)
between the two fragments as the sum of the energy contribu-
tions of the stabilizing orbital interactions (∆Eorb) and the steric
interaction (∆Este), which is the sum of the attractive electrostatic
contributions (∆Eelst) and the Pauli repulsions (∆EPauli). ∆Eint
between the fragments Mo(CO)3 and [1-Cp-5,6,7,8-(Me)4-nido-
1,5,6,7,8-IrC4B3H3] is -3.16 eV (305 kJ/mol), mainly due to
Figure 4. Total electron density plots in the metal-B-B plane for (a) 7′
and (b) [1-Cp-5,6,7,8-(CH3)4-closo-1,5,6,7,8-IrC4B3H3RuCp]+. Contour plots
range from 0.0 (red) to 0.13 e/bohr3 (blue).
(CO)IrC2Me4B3H3, gives an interaction energy of -5.81 eV,
which results from ∆Eorb ) -10.25 eV and ∆Este ) 4.45 eV.
Consistent with the experimental observations, these results
show that the Mo-cluster binding is considerably stronger in
the case of 6′ than 7′.
Is the difference in binding energies caused by lack of Mo-B
interactions in 7′, i.e., is 7 best described as a 16-electron Mo
cluster complex or, alternatively, a hypoelectronic metallacar-
borane? This is a much more difficult question to answer but
can be addressed by partitioning the electronic structure solution
with methods previously found useful. Partitioning of the
binding energies of 6′ and 7′ is problematical as the two Mo-
cage interactions are significantly different in the two cases.
That is, 6′ has a Mo-Ir bridging carbonyl, as well as a Mo-Ir
bond, and its larger binding energy does not necessarily mean
the Mo-B binding is weak in 7′.
For this reason the nature of the Mo-B bonds in 7′ was
examined by analyzing the electron density distribution. A cut
corresponding to the Mo-B-B plane is shown in Figure 4.
The density in this region is small and supports a weak Mo-B
interaction at best. A frontier molecular orbital (FMO) fragment
analysis of the MO diagram of 7′ is consistent with this
conclusion. The FMOs of the Mo(CO)3 moiety (a typical ML3
fragment with three hybrids FMOs of σ and π type above three
occupied “t2g” FMOs) mainly interact with FMOs of the cluster
which have a carbon character. Indeed, only two MOs of 7′
with a tiny Mo-B bonding character can be identified at low
energy (HOMO-6 and HOMO-9), leading to a very weak Mo-B
interaction.
The “synthesis” of a late metal derivative was carried out
for comparison, i.e., the Mo(CO)3 fragment was replaced with
an isolobal [RuCp]+ fragment as this particular fragment is
experimentally accessible via [Cp*Ru(NCCH3)3][PF6].42 This
hypothetical molecule has an optimized geometry similar to that
of 7′ and a computed binding energy 2 eV higher (5.12 eV vs
3.16 for Mo)! The HOMO-LUMO gap also increases signifi-
cantly in going from 7′ to the [RuCp]+ analog (1.3-1.8 eV).
These results led to several attempts to synthesize the ruthenium
derivative by the reaction of 8 with a source of the [RuCp*]+
fragment. Unfortunately, none were successful, but the molecule
should be able to be isolated if an appropriate route can be
discovered.
attractive orbital interactions (∆Eorb ) -5.88 eV and ∆Este
)
A possible explanation for the larger binding energy between
the ruthenium entity and the cluster might be that the [RuCp]+
fragment interacts with the four carbon atoms and the two boron
2.72 eV). A similar fragmentation for 6′, i.e., Mo(CO)3 and Cp-
(39) Ghosh, S.; Beatty, A. M.; Fehlner, T. P. Angew. Chem., Int. Ed. 2003, 42,
4678.
(40) Morokuma, K. J. Chem. Phys. 1971, 55, 1236-1244.
(41) Ziegler, T.; Rauk, A. Theor. Chim. Acta 1977, 46, 1-10.
(42) Francis, M. D.; Holtel, C.; Jones, C.; Rose, R. P. Organometallics 2005,
24, 4216.
9
3396 J. AM. CHEM. SOC. VOL. 129, NO. 11, 2007

Stoichiometric Metallaborane Reaction Mechanism
A R T I C L E S
Figure 5. Plot of ln K′ vs 1/T derived from 11B NMR intensities signals
from 7 and 8 in equilibrium 1.
atoms, whereas the Mo(CO)3 moiety interacts only with four
carbon atoms in 7′. To check this hypothesis, the total electron
density between B and the metal for 7′ was compared with that
in the [RuCp]+ analog (Figure 4). The magnitude of the density
in the latter is also consistent with a weak Ru-B interaction
only slightly more important than in 7′ at best. In essence, the
binding in the [RuCp]+ derivative is similar to that in 7′, i.e.,
Mo-cluster interactions are mainly via M-C contacts, which
are somewhat stronger because of better orbital interactions
between the FMOs of the [RuCp]+ and the iridacarborane
cluster.
All our analyses of the results suggest that the Mo-B
interaction in 7′ is weak and the binding to Mo resides largely
in the Mo-C interactions. We conclude that 7 is appropriately
viewed as a 16-electron cluster Mo coordination compound
where cluster 8 acts as a four-electron donor. Certainly it may
also be viewed as a single hypoelectronic cluster containing
two transition metals; however, this is much less informative.
The solution is pleasing as it accommodates the experimentally
observed weak binding of cluster 8 to the Mo center and its
ready displacement by Lewis bases. Both are of crucial
importance to the ultimate release of the final product 8 by base
displacement.
Equilibrium Constant. The observation that the ratio 7 to 8
in a mixture of the two compounds can be increased by the
addition of (arene)Mo(CO)3 and decreased by the addition of a
Lewis base suggests the existence of equilibrium 1. As equi-
librium properties provide access to important thermodynamic
parameters, the equilibrium constant, Keq, was measured.
Equilibrium mixtures were prepared by adding measured
amounts of (arene)Mo(CO)3 and 8 to a known volume of THF.
The NMR tube containing the mixture was heated at 50 °C to
form 7 and then equilibrated at a chosen temperature for up to
16 h at the lowest temperature. The equilibrium ratio of 7 to 8
was measured by 11B NMR at the temperatures, -20, 0, 20,
and 50 °C. Equilibrium 1 written as a formation reaction for 7
Figure 6. Plot of reaction time vs relative abundances of reactants and
products based on 1H NMR signal intensities (2-butyne, dark blue diamonds;
2, magenta squares; 7, brown double crosses; 5, azure crosses; 4, yellow
triangles).
(R ) Ph). Either the reaction has a substantially higher barrier
or the greater steric bulk of Ph versus Me decreases K′
sufficiently to reduce the concentration of 7 below the level of
NMR detection.
As described above the gas-phase binding energy of a Mo-
(CO)3 fragment to 8 (for Cp rather than Cp*) is exothermic by
3.16 eV. The 8 kcal/mol endothermicity of equilibrium 1 shows
that the binding energy of three THF molecules to Mo(CO)3 is
larger than the binding energy of 8 to the Mo center. The large
positive entropy change associated with equilibrium 1 is a result
of the fact that two entities are converted into four which, in
the gas phase, would constitute an increase of six translational
degrees of freedom. The enthalpic and entropic factors largely
balance each other at room temperature, and the equilibrium
constant is sufficiently close to unity to permit both 7 and 8 to
be observed in the equilibrium mixture at room temperature.
We conclude that, as cluster carbon content increases, the
binding energy to the Mo center decreases mainly due to a loss
of Mo-B bonding. As a consequence, the addition of two alkyne
units leads to sufficiently poor binding that Mo is readily
displaced by basic solvents or silica gel. The presence of the
Mo center in 2 facilitates the addition of alkyne, but it is the
change in Mo-cluster binding with carbon content that leads
to the detachment of the Mo center. As a result, the overall
process in Scheme 1 constitutes a classic example of a metal-
facilitated transformation where the substrate is a metallaborane
rather than an organic molecule.
8 + Mo(THF)3(CO)3 a 7 + 3THF; K′ ) Keq/[THF]3 (1)
Reaction Intermediates. In the first hour of the reaction of
2 with 2-butyne, two species, 4 and 5, are observed by 1H and
11B NMR to form before the fully characterized products 7 and
8 described above. (Product 6 is only abundant under conditions
of alkyne deficiency described above.) A plot of NMR signal
intensities of these species as a function of time relative to those
of the isolated products is shown in Figure 6. The shapes of
the curves are consistent with 4 and 5 arising from 2 and leading
to 7. Although 4 and 5 appear in parallel, 5 is in higher
abundance with a longer life time under the reaction conditions.
Assuming the compositions and structures for 4 and 5 shown
is a slightly endothermic process, i.e., the mole fraction of 7
increases from 0.4 at -20 °C to 0.8 at 50 °C. As shown in
Figure 5 straight line fit to a plot of ln K′ versus 1/T gave ∆H
) 8 kcal/mol and ∆S ) 37 cal/mol K. These are approximate
values because raw NMR intensities were used for concentration
measurements, the presence of arene was ignored, and the
temperature range was small. However, because of the expo-
nential relationship between Keq and free energy even a crude
measure of the former gives useful energy values. Addition of
(arene)Mo(CO)3 to 8 (R ) Ph) did not yield any measurable 7
9
J. AM. CHEM. SOC. VOL. 129, NO. 11, 2007 3397

A R T I C L E S
de Montigny et al.
Table 1. NMR Assignment of IrB3H3C4Me4 and Intermediates
2
4
5
6
7
1H NMR
Cp*
BHt
1.80 (15 H)
2.17 (15 H)
2.06 (15 H)
1.54 (15 H)
1.92 (15 H)
7.0 (1 H)
3.6 (1 H)
3.4 (1 H)
3.1 (1 H)
5.20 (2 H)
2.43 (1 H)
5.16 (2 H)
3.01 (1 H)
5.48 (2 H)
2.77 (1 H)
5.36 (1 H)
1.54 (2 H)
Me
2.0 (6 H)a
2.1 (6 H)
2.24 (6 H)
2.44 (6 H)
2.10 (6 H)
H-B-M
0.3 (1 H)
-3.1 (1 H)
-3.5 (1 H)
-5.2
-5.0 (1H)
H-M
-6.9 (1 H)
-11.51(1 H)
-20.47(1 H)
11B NMR
44.7 (1 B)
9.5 (1 B)
-5.0 (1 B)
32.5 (2 B)
1.1 (1 B)
33.9 (2 B)
-6.1 (1 B)
33.4 (2 B)
1.9 (1 B)
33.1 (1 B)
1.1 (2 B)
a Assignment uncertain.
in Scheme 1 (assignments discussed below), the overall mass
balance is good suggesting there are no more intermediate
products of comparable abundance. An important consequence
is that the mass balance is only good for the Me resonances if
4 and 5 each contain one alkyne moiety. At a reaction time
where the system contains only 5 and 7 (plus a small amount
of 8), addition of excess alkyne converts 5 to 7 quantitatively
by 11B NMR (see Figure B, Supporting Information).
another alkyne must be added to reach 7, we suggest retention
of the Mo(CO)3THF fragment of 2 in 4 and 5. This give a 9
sep count and a nido structure for each.
The formulation of both 4 and 5 as nido-Cp*IrB3H5C2R2-
Mo(CO)3THF requires that they be isomers. It is reasonable to
assume that Cp*IrB3H3 and the R2C2 fragments observed in 6-8
remain intact in 4 and 5. Thus, the difference between 4 and 5
is the relative positions of the Ir and Mo metals. The Ir and Mo
atoms are adjacent in 2 and 6 and should be so placed in either
4 or 5. For reasons discussed next the structures are assigned
as [1-Cp*-7,7,7-(CO)3-7-THF-2,3-(CH3)2-nido-1,7,2,3-IrMo-
C2B3H5], 4, and [5-Cp*-7,7,7-(CO)3-7-THF-2,3-(CH3)2-nido-
5,7,2,3-IrMoC2B3H5], 5, and shown in Scheme 1.
In a mechanistic context, two roles for 4 and 5 are possible:
a consecutive path of 2 to 4 to 5 to 7 or a parallel path 2 to 4,
5 to 7. Unfortunately, as already mentioned, the early time
resolution of the NMR signal intensities does not permit the
two to be distinguished. In the absence of definitive information
we have used a consecutive pathway in Scheme 1 as it is a
more satisfying one. Considering the position of the labile THF
ligand in 2, it is reasonable that alkyne insertion takes place
via the Ir-Mo edge. If so, elimination of BH3 from an
intermediate not shown leads to 4 with nonadjacent metals.
Rearrangement to 5 with adjacent metals permits insertion of
another alkyne or, if alkyne is limiting, unimolecular H2
elimination and formation of 6. Note that the 11B resonance of
the unique boron shifts upfield smoothly in going from 4 (1.1
ppm) to 5 (-6.1 ppm) to 7 (-10.6 ppm) consistent with position
connectivities and nearest neighbor atom identities.
Although the structures of 4 and 5 cannot be viewed as
definitive, the overall mechanistic picture is clearssequential
addition of alkyne competitive with BH3 or H2 fragment
elimination. This stoichiometric process is fully consistent with
earlier cluster mechanistic information in the literature.43
Olefin IsomerizationsMechanistic Probe. What remains
to be defined is the initial step in going from 2 to 4.
Circumstantial evidence implicates dissociation of the labile
THF and coordination of the alkyne to Mo as prerequisite to
the insertion process. If alkynes coordinate to 2, then olefins
should also displace THF from 2. However, in contrast to
The structural assignments for 4 and 5 are based on the known
spectroscopic and structural data of the reactant 2 and products
7 and 8. The time dependence of the signals provides a low-
resolution marker that allows the complex Me and Cp* region
of the 1H NMR to be sorted and assigned. Although the broad
11B NMR signals showed the same trends in composition with
time (Supporting Information, Figure B), it was the well-
1
resolved H data that provided the data for Figure 6.
The NMR assignments of 4 and 5 are given in Table 1, where
it will be seen that the two intermediates are closely related to
each other as well as to 2 and 7. The Me resonance associated
with 4 is tentative, and the mass balance above is a stronger
argument for the presence of one alkyne moiety in this
intermediate. In addition, the individual assignments of the B-H
terminal resonances to 4 versus 5 could be interchanged. On
the other hand, confidence in the assignments of the high-field
hydrides and Cp* resonances is high as they are well resolved
and there is a time period when 5 and 7 are the predominant
species present. The following summarizes the reasoning leading
to the proposed structures of 4 and 5 shown in Scheme 1.
With single alkyne residues 4 and 5 are reasonable precursors
to 7 by addition of alkyne and 6 by loss of H2 (and THF/CO
exchange). Further, like 6 and 7 the two intermediates contain
three B-H fragments; hence, a B-H fragment must be lost in
a relatively rapid step in going from 2 to the intermediates. The
total number of high-field hydride signals associated with 6 and
7 require the loss of 2 H as well; hence, in going from 2 to the
intermediates 1 mol of alkyne is added and the elements of BH3
are lost. The latter has been well documented previously in the
reactions of alkynes with nido-(Cp*RuH)2B3H7.20-22 At this
point we have a composition Cp*IrB3H5C2R2, but a Mo
fragment is almost certainly present. If it is a Mo(CO)3 fragment
the species would have 8 sep and a probable closo structure
like 6. This is unlikely as 6, like closed clusters in general, is
a stable end product. For this reason, as well as the fact that
(43) Shriver, D. F., Kaesz, H. D., Adams, R. D., Eds. The Chemistry of Metal
Cluster Complexes; VCH: New York, 1990.
9
3398 J. AM. CHEM. SOC. VOL. 129, NO. 11, 2007

Stoichiometric Metallaborane Reaction Mechanism
A R T I C L E S
Scheme 3
alkynes, olefins have not been found to yield metal-boron-
containing products under mild conditions for any members of
this class of metallaboranes. Although a disappointment in terms
of new derivatives, this negative observation provides a way
of probing the coordination process. Thus, we now show that
two olefins, tetramethylallene and hexane, provide evidence of
THF displacement and coordination to 2. This supports revers-
ible binding of alkyne to the Mo site of 2 to form 3 as shown
in Scheme 1. As the results on hexene serve as a foundation
for understanding the allene reaction, they are presented first.
The reaction of hex-1-ene with 1 mol % of 2 leads to the
slow generation of cis-hex-2-ene with a turnover rate averaging
6/day (Table A, Supporting Information). If light is excluded,
no trans-hex-2-ene is observed. Unfortunately, 2 degrades to
unknown, intractable products on roughly the same time scale
as the catalytic reaction, but by replacing it daily a total
conversion of 96% is achieved in 10 days (2.3 mmol hex-2-
ene produced with a total of 0.43 mmol 2 added). Continued
addition of catalyst led to a final conversion of 98%. Examina-
tion of the reverse reaction with hex-2-ene containing ap-
proximately 1% hex-1-ene impurity showed no reaction. Note
that thermal isomerization of hex-1-ene requires more vigorous
conditions.44,45
the metal with formation of an allyl complex, H transfer back
to the ligand, and generation of either the original alkene or its
isomer. For the metallaborane we suggest that the olefin
displaces THF from 2, undergoes rearrangement at the Mo site,
and is displaced by THF, i.e., for the alkene, step 2 to 3 in
Scheme 1 takes place but the insertion pathway available to
the alkyne does not. In contrast, the addition of a B-H bond to
an olefin takes place in an anti-Markownikoff sense, and
although hydroboration of alkynes was observed for a ruthen-
aborane,22 the B-H bonds of 2 do not react with either alkene
or alkyne.
When tetramethylallene is added to a solution containing 2
there is an instant color change which we attribute to formation
of a complex, i.e., the binding constant for the allene is larger
than that of the hexene. Once again no reaction is observed
other than the conversion of the cumulene to an isoprene
derivative (2,4-dimethylpenta-1,3-diene) as shown in eq 2. On
1
the basis of the H and 13C NMR the TON after 3 h were 18
and 11, respectively, at 50 °C and 6.5 at room temperature.
Conversions ranged from 7% to 20%. The room-temperature
catalytic rate (2 turnovers/h) is about 1 order of magnitude larger
than that found for hexene (0.25 turnovers/h), which is consistent
with the apparently higher binding constant mentioned above.
The catalytic isomerization of olefins is well studied, and the
general features of the mechanism are understood.9 Hence, our
observations can be used to comment on the nature of 2 provided
decomposition products can be excluded as the active catalyst.
The isomerization of pen-1-ene to a mixture of both cis- and
trans-pen-2-ene by ruthenium carbonyl clusters has been
reported.46 Such clusters are more prone to metal loss than
metallaboranes. This contrasts with the sole conversion to the
cis isomer here and, along with the observations on the allene
below, suggests that the decomposition products of 2 are not
significantly active.
Me2CdCdMe2 f cis-H2CdC(Me)CHdC(Me)2 (2)
In the absence of a catalyst, tetramethylallene dimerizes
quantitatively to tetramethyl-1,2-di-isopropylidenecyclobutane
without isomerization in 72 h at 150 °C in a reaction vessel
pretreated with strong base.44 The isomerization reaction 2 as
well as more complex dimerization occurs on proton acidic
surfaces in 24 h at the same temperature.44 In a cluster, it is the
bridging hydrogen atoms that are most acidic, and in a
metallaborane such as 2, the Mo-H-B hydrogen is expected
to be more acidic than B-H-B hydrogen. The former, being
adjacent to the Mo binding site, is readily available for transfer
to a bound allene. Further, the other Mo-B edge of 2 lacks a
bridging hydrogen atom and is available to accept a proton.
Hence, the mechanism sketched in Scheme 3, where the arrows
show movement of electron density, is a viable one for the
The general mechanism for known mononuclear catalysts
consists of addition of the olefin to a vacant site, H transfer to
(44) Taylor, D. R.; Wright, D. B. Chem. Commun. 1968, 434.
(45) Taylor, D. R.; Wright, D. B. J. Chem. Soc. B 1971, 391.
(46) Gladfelter, W. L.; Roesselet, K. J. In The Chemistry of Metal Cluster
Complexes; Shriver, D. F., Kaesz, H. D., Adams, R. D., Eds.; VCH: New
York, 1990; p 329.
9
J. AM. CHEM. SOC. VOL. 129, NO. 11, 2007 3399

A R T I C L E S
de Montigny et al.
observed isomerization at room temperature. In essence, coor-
dination of one double bond is followed by transfer of a methyl
proton to the Mo-B edge as the Mo-H-B proton is transferred
to the central allenic carbon atom. Release of the less strongly
bound 1,3-diene (see hexene results) completes the isomeriza-
tion.
such a view really is. For the same substrate, metal identity
completely dominates the predominant reaction pathway ob-
served. Metal choice, then, provides a way to control reactivity
and suggests continued study of structurally similar but com-
positionally different metallaboranes will be profitable.
Experimental Section
These observations are consistent with those on hexene and
establish step 2 to 3 in the mechanism of Scheme 1, i.e., the
intermediate is [1-Cp*-2,2,2-(CO)3-2-(alkyne)-nido-1,2-IrMo-
B4H8], 3. Thus we have built up a stoichiometric mechanism
for the complex overall reaction. It does mimic a typical metal-
facilitated reaction in organometallic chemistry. That is, step 1
to 2 constitutes metal incorporation into the iridaborane
substrate. Step 2 to 3 describes the Mo-facilitated assembly of
the second substrate. Step 3 to 4 denotes reaction of the alkyne
and iridaborane substrates which, in this case, is alkyne insertion
into the cluster framework. Steps 4 to 5 and 5 to 7 describe the
incorporation of the second alkyne with concomitant decrease
in the strength of the association of the Mo center with the
original iridaborane framework. Equilibrium 7 to 8 constitutes
product release and reformation of the Mo complex originally
used to generate 2 from 1.
General Considerations. All reactions were carried out under an
argon atmosphere using standard Schlenk-line techniques.47 Solvents
were distilled immediately before use under a dinitrogen atmosphere:
sodium benzophenone ketyl for hexanes and tetrahydrofuran and
calcium hydride for dichloromethane. All commercial reagents were
used as received without further purification. The iridapentaborane 1
and the (arene)tricarbonylmolybdenum complex were prepared by
following synthetic procedures previously reported.25,27,28 NMR spectra
were recorded on a 400 MHz Bruker instrument. NMR references:
internal (C6D6, δH 7.16 ppm) for 1H; external ([(Me4N)(B3H9)] in
acetone-d6 δB -29.7 ppm) for 11B. IR spectra were recorded on a
Nicolet 205 FT-IR spectrometer. Mass spectra were acquired on a
Finnigan MAT model 8400 mass spectrometer. M-H-W Laboratories,
Phoenix, AZ, performed the elemental analysis.
The heterometallaborane 2 was prepared and used in situ by adding
arachno-[Cp*IrB4H10] (100 mg, 0.26 mmol) to Mo(tol)(CO)3 (100 mg,
0.37 mmol) in freshly distilled THF (5 mL). This mixture was stirred
at 50 °C for 10 min to generate 2, and then a large excess of 2-butyne
(0.1 mL, 1.28 mmol) was added. The reaction mixture was stirred 2 h.
The mixture was purified by TLC (silica, hexane) to afford 8 (R )
Me, 24 mg, 20%) and a trace of 6. Yellow crystals of 8 were obtained
from hexane. By a similar method, a 28% yield of 8, R ) Ph, was
obtained, and yellow crystals were grown from CH2Cl2/hexane.
Improved yields of 6 (4 mg, 2%) were obtained by using less than 1
equiv of 2-butyne (0.02 mL, 0.25 mmol). A mixture of 8 (R ) Me, 10
mg, 0.02 mmol) and Mo(tol)(CO)3 (10 mg, 0.04 mmol) in dry THF-d6
(0.5 mL) was heated at 50 °C for 3 h at which point the 11B NMR
spectrum shows large conversion to 7. Although some 8 was present,
a few brown crystals of 7 were obtained from this solution at -20 °C.
[1-Cp*-5,6,7,8-(CH3)4-nido-1,5,6,7,8-IrC4B3H3], 8, R ) Me. 11B
Conclusions
A comparison of the reactivities of ruthena-, rhoda-, and
iridaboranes with alkynes further resolves the structural factors
that define the productive interaction with alkynes under mild
conditions. The first factor of importance is cluster geometry/
electron count. Earlier work showed that isoelectronic nido-
ruthena- and rhodaborane clusters easily add alkyne giving
arachno cluster intermediates. This leads to exclusively to
cyclotrimerization for the rhodaborane and predominantly to
ruthenacarboranes for the ruthenaborane. Although alkyne
addition is facile for both metallaboranes, only in the case of
ruthenium are there additional cluster hydrogen atoms to
facilitate hydrometallation of the alkyne and initiate the forma-
tion of a variety of novel derivatives containing the alkyne
components.
1
NMR ([D6]benzene, 22 °C, 128 MHz): δ ) 11.6 (d, J(H,B) ) 157
1
1
Hz, 1B; BH), -10.4 (d, J(H,B) ) 128 Hz, 2B; BH). H{11B} NMR
([D6]benzene, 22 °C, 400 MHz): δ ) 5.16 (s, 1H; BHt), 2.30 (s, 2H;
BHt), 2.19 (s, 6H; CH3), 1.88 (s, 6H; CH3), 1.68 (s, 15H; Cp*). 13C-
{1H} NMR ([D6]benzene, 22 °C, 100 MHz): δ ) 133.5 (CCH3), 133.4
(CCH3), 92.9 (C5(CH3)5), 22.7 (CH3), 19.0 (CH3), 9.3 (C5(CH3)5). IR
(KBr): 2956 (s), 2924 (s), 2855 (s), 2453 (s, B-H), 2019 (w), 1950
(w), 1929 (w), 1460 (m), 1379 (m), 1262 (m), 1091 (m), 1029 (m),
951 (w), 798 (s). LR-MS (FAB): m/z isotope envelope; 471 [M]+, 3
B, 1 Ir atoms, calcd for weighted average of isotopomers (C18H30B3Ir)
lying within the instrument resolution 472.2256; obsd, 472.2258.
Elemental analysis (%) calcd for C18H30B3Ir: C, 45.89; H, 6.42.
Found: C, 45.70; H, 6.60.
In the synthesis of the group 9 metallaborane congeners from
monoborane and monocyclopentadienyl metal chlorides, the
cobalt and rhodium derivatives evolve H2 until nido clusters
are isolated, whereas for iridium the process stops at the
electron-rich arachno products. Hence, although the arachno-
iridaborane has an abundance of hydrogen atoms, Lewis base
addition is not favored. The tendency to lose a pair of electrons
is greater than that for addition. As a consequence, the
iridaborane alone does not react with alkynes under mild
conditions. Introduction of a metal center with a weakly
coordinated THF molecule into the iridaborane cluster provides
a nexus for alkyne coordination and leads to metallacarborane
formation, albeit with interesting additional features. Because
the added metal center binding decreases with added carbon
content, it is easily removed, thereby completing the metal-
facilitated process of alkyne addition to the iridaborane cluster.
The introduction of transition metal centers into a borane
framework provides a variety of cluster analogues of boranes
that often are easily understood with the cluster electron counting
rules and isolobal ideas. That is, from a purely geometrical/
structural viewpoint, the Ru, Rh, Ir clusters are routine. The
comparative chemistry, outlined above, shows how deceptive
[1-Cp*-6,7,8,8-Ph5-nido-1,6,7,8,8-IrC5B4H4], 8, R ) Ph. 11B NMR
1
([D6]benzene, 22 °C, 128 MHz): δ ) 9.6 (d, J(H,B) ) 180 Hz, 1B;
BH), -9.1 (d, 1J(H,B) ) 73 Hz, 2B; BH). 1H{11B} NMR ([D6]benzene,
3
22 °C, 400 MHz): δ ) 7.41 (d, J(H,H) ) 8 Hz, 4H, PhH), 7.35 (d,
3J(H,H) ) 8 Hz, 4H, PhH), 7.07 (t, 3J(H,H) ) 7 Hz, 4H, PhH), 7.01-
6.95 (m, 8H, PhH), 5.53 (s, 1H; BHt), 3.07 (s, 2H; BHt), 1.60 (s, 15H;
Cp*). 13C{1H} NMR ([D6]benzene, 22 °C, 100 MHz): δ ) 144.0
(PhCi), 143.6 (PhCi), 132.3 (PhCo), 127.3 (PhCm), 126.7 (PhCp), 94.8
(C5(CH3)5), 9.9 (C5(CH3)5). IR (hexane): 2497 (s, B-H), 2020 (s),
1954 (s), 1940 (w), 1599 (w), 1262 (s), 1099 (s), 1018 (s). LR-MS
(FAB): m/z isotope envelope; 719 with cutoff at 720, [M+], 3 B, 1 Ir
atoms, calcd for weighted average of isotopomers C38H38B3Ir lying
within the instrument resolution 720.2882; obsd, 720.2873.
(47) Shriver, D. F.; Drezdzon, M. A. The Manipulation of Air SensitiVe
Compounds, 2nd ed.; Wiley-Interscience: New York, 1986.
9
3400 J. AM. CHEM. SOC. VOL. 129, NO. 11, 2007

Stoichiometric Metallaborane Reaction Mechanism
A R T I C L E S
[1-Cp*-2,2,2-(CO)3(µ-CO)-3,4-(CH3)2-closo-1,2,3,4-IrMoC2B3-
H3], 6. 11B NMR ([D6]benzene, 22 °C, 128 MHz): δ ) 33.4 (d, 1J(H,B)
Mo(tol)(CO)3 (30 mg, 0.11 mmol) in 1 mL of THF-d8. After 3 h
1
addition of hexane (1 mL) was followed by extraction on Celite. H
) 138 Hz, 2B; BH), 1.9 (d, J(H,B) ) 168 Hz, 1B; BH). H{11B}
NMR ([D6]benzene, 22 °C, 400 MHz): δ ) 5.48 (s, 2H; BHt), 2.77
(s, 1H; BHt), 2.24 (s, 6H; CH3), 1.54 (s, 15H; Cp*). IR (hexane): 2518
(w, B-H), 2137 (w, CO), 2022 (s, CO), 1997 (s, CO), 1925 (s, CO),
1936 (s, CO), 1820 (m, µ-CO), 1271 (m), 1102 (m), 1011 (m). LR-
MS (FAB): m/z calcd for C18H24IrMoB3O4 isotope envelopes; 626,
19% [M+], 598, 18% [M-CO+], 569, 65% [M-(CO)2+], 513, 100%
[M-CH3C2CH3-(CO)2+].
and 13C NMR of this material revealed the presence of the isoprene
derivative (2,4-dimethylpenta-1,3-diene) in addition to the solvents
1
1
1
(THF and hexane) and the starting tetramethylallene. The H NMR
spectra show 20% conversion and a TON ) 18. Because of the
proximity of the proton shifts of the terminal-methyl resonances, the
ratio between the cumulene and σ-tropic product was obtained from
the 13C NMR. This measurement gives 12% conversion and a TON )
11. The σ-tropic product is also observed at room temperature albeit
1
[1-Cp*-5,6,7,8-(CH3)4-closo-1,5,6,7,8-IrC4B3H3Mo(CO)3], 7. 11B
in lower yield (7% conversion based H NMR for a TON ) 6.5 in 3
1
h).
NMR ([D8]THF, 22 °C, 128 MHz): δ ) 33.1 (d, J(H,B) ) 160 Hz,
1B; BH), -10.6 (d, 1J(H,B) ) 137 Hz, 2B; BH). 1H{11B} NMR ([D8]-
THF, 22 °C, 400 MHz): δ ) 5.36 (s, 1H; BHt), 2.44 (s, 6H; CH3),
2.10 (s, 6H; CH3), 1.92 (s, 15H; Cp*), 1.54 (s, 2H; BHt). 13C{1H} NMR
([D8]THF, 22 °C, 100 MHz): δ ) 202.1 (CO), 129.6 (CMe), 129.5
(CMe), 92.9 (C5(CH3)5), 21.5 (CH3), 18.5 (CH3), 9.3 (C5(CH3)5). IR
([D8]THF): 2495 (s, B-H), 1980 (w, CO), 1941 (w, CO), 1893 (m,
CO), 1854 (m, CO). LR-MS (FAB): m/z isotope envelope; 471 [M]+,
3 B, 1 Ir atoms corresponding to the fragment ion [Cp*IrC4-
(CH3)4B3H3]+.
NMR Equilibrium Experiment. To obtain K′ as a function of T, a
mixture of 8 (31 mg, 0.066 mmol) and Mo(tol)(CO)3 (15 mg, 0.055
mmol) in THF-d8 was heated at 50 °C for 3 h in an NMR tube and
then equilibrated at the temperature of choice. At 50 °C the composition
was unchanged after another 3 h. At 20 and 0 °C measurements were
taken after 5 h, whereas at -20 °C they were taken after 16 h. The
mole fraction of 7 was determined by 11B NMR to be 0.39 (-20 °C),
0.64 (0 °C), 0.68 (20 °C), 0.77 (50 °C) yielding ln K′ ) -3925 (1/T)
+ 18.8 (R2 ) 0.96).
Computational Details. Density functional theory calculations were
carried out on model compounds (non-methylated Cp instead of Cp*)
using the Amsterdam Density Functional (ADF) program49 developed
by Baerends and co-workers.50,51 Electron correlation was treated within
the local density approximation (LDA) in the Vosko-Wilk-Nusair
parametrization.52 The nonlocal corrections of Becke and Perdew were
added to the exchange and correlation energies, respectively.53,54 The
atom electronic configurations were described by a triple-ú Slater-type
orbital (STO) basis set for H 1s, B 2s and 2p, C 2s and 2p, O 2s and
2p, augmented with a 2p single-ú polarization function for H atoms,
with a 3d single-ú polarization function for O, and with a 3d and a 4f
single for B and C. The atom electronic configurations of the Mo, Ru,
and Ir atoms were described by a triple-ú STO for the outer nd and (n
+ 1)s orbitals. A single-ú STO was used for the outer Mo and Ru 5p
orbitals and the outer Ir 6p orbitals. The Mo basis set was augmented
with a 4f polarization function. For all atoms, a frozen core approxima-
tion for the inner shells was used. Geometry optimizations were carried
out using the analytical gradient method implemented by Verluis and
Ziegler.55 Relativistic corrections were added using the zeroth-order
regular approximation scalar Hamiltonian.56-58 Representations of the
total electronic densities were done using MOLEKEL4.1.59
Reaction of 2 with Olefins. The photoisomerization of the alkenes
(cis-hex-2-ene to trans-hex-2-ene or 2,4-dimethylpenta-1,3-diene to 2,4-
dimethylpenta-2,3-diene as examples) was prevented by wrapping the
Schlenk tubes in aluminum foil. Tubes were capped and sealed with
Parafilm to reduce evaporation of products. Celite, not silica gel, was
used for product extraction in order to prevent σ-tropic rearrangements.48
With Hex-1-ene. Cp*IrB4H9 (10 mg, 2.6 10-5 mol) and Mo(tol)-
(CO)3 (10 mg, 3.7 10-5 mol) were first dissolved in 1 mL of THF-d8
and stirred 10 min at 50 °C to generate 2. Then 0.3 mL of hex-1-ene
(2.4 10-3 mol) was added. A slow color change from (light orange) to
dark yellow is observed at room temperature, and the 1H NMR reveals
conversion to cis-hex-2-ene in amounts exceeding initial catalyst
quantities. 1H NMR spectra of the crude were used to follow the reaction
(Supporting Information, Table A). Turnover numbers (TON) were
estimated from the integrated intensities of the alkenic protons lying
between 4.5 and 5.5 ppm: cis-hex-2-ene; 5.34 ppm 2 H, m, and hex-
1-ene; 4.87 ppm 2 H terminal, m. The catalyst was renewed each day
by adding Cp*IrB4H9 and Mo(tol)(CO)3 until complete conversion was
obtained. The data are presented in Table A in the Supporting
Information. For confirmation addition of 1 mL of hexane and extraction
Acknowledgment. This work was supported by the National
Science Foundation (CHE 0304008) and by an International
Grant (INT02-31792) in cooperation with the CNRS, France
(14559-CNRS-NSF). We thank the French CINES (Montpellier)
and IDRIS (Orsay) for computing facilities.
Supporting Information Available: Molecular structure of
8, R ) Ph (Figure A), time dependence of the 11B spectra
(Figure B), catalytic isomerization of hexane (Table A), and
the optimized Cartesian coordinates of 6′, 7′, 8′, and [1-Cp-
5,6,7,8-(CH3)4-closo-1,5,6,7,8-IrC4B3H3RuCp]+. This material
is available free of charge via the Internet at http://pubs.acs.org.
JA068999Z
(49) SCM. ADF, 2005.01 ed.; Theoretical Chemistry, Vrije Universiteit:
Amsterdam, The Netherlands, 2005.
(50) Fonseca Guerra, C.; Snijders, J. G.; te Velde, G.; Baerends, E. J. Theor.
Chem. Acc. 1998, 99, 391-403.
1
by Celite permitted analysis by both H and 13C NMR. They showed
(51) te Velde, G.; Bickelhaupt, F. M.; Fonseca Guerra, C.; van Gisbergen, S. J.
A.; Baerends, E. J.; Snijders, J. G.; Ziegler, T. J. Comput. Chem. 2001,
22, 931-967.
by comparison with standard spectra, hex-1-ene, and the σ-tropic
product, cis-hex-2-ene, as well as THF and hexane. In the absence of
an aluminum foil light shield, a mixture of trans-hex-2-ene and cis-
hex-2-ene was observed. The thermal isomerization of hex-1-ene
requires harsher conditions than those employed here.44,45
With 2,5-Dimethylpenta-2,4-diene. A rapid color change (from light
orange to dark yellow brown) was observed at 50 °C after addition of
0.1 mL of allene (7.2 mmol) to Cp*IrB4H9 (30 mg, 0.078 mmol) and
(52) Vosko, S. D.; Wilk, L.; Nusair, M. Can. J. Chem. 1990, 58, 1200-1211.
(53) Perdew, J. P. Phys. ReV. B 1986, 34, 7406.
(54) Becke, A. D. Phys. ReV. A 1988, 38, 3098-3100.
(55) Verluis, L.; Ziegler, T. J. Chem. Phys. 1988, 88, 322-328.
(56) van Lenthe, E.; Baerends, E. J.; Snijders, J. G. J. Chem. Phys. 1993, 99,
4597-4610.
(57) van Lenthe, E.; Baerends, E. J.; Snijders, J. G. J. Chem. Phys. 1994, 101,
9783-9792.
(58) van Lenthe, E.; van Leeuwen, R.; Baerends, E. J.; Snijders, J. G. Int. J.
Quantum Chem. 1996, 57, 281-293.
(48) Kropp, P. J.; Breton, G. W.; Craig, S. L.; Crawford, S. D.; Durland, W. F.,
(59) Flu¨kiger, P.; Lu¨thi, H. P.; Portmann, S.; Weber, J. MOLEKEL4.1; Swiss
Center for Scientific Computing: Manno, Switzerland, 2002.
Jr.; Jones, J. E., Raleigh, J. S. III. J. Org. Chem. 1995, 60, 4146.
9
J. AM. CHEM. SOC. VOL. 129, NO. 11, 2007 3401