Alkene hydrogenation on an 11-vertex rhodathiaborane with full cluster participation

Article abstract of DOI:10.1021/ja802993m

The facile synthesis of the metallaheteroborane [8,8-(PPh₃) ₂-nido-8,7-RhSB₉H₁₀] (1) makes possible the systematic study of its reactivity. Addition of pyridine to 1 gives in high yield the 11-vertex nido-hydridorhodathiaborane [8,8,8-(PPh₃) ₂H-9-(NC₅H₅)-nido-8,7-RhSB₉H ₉] (2). 2 reacts with C₂H₄ or CO to form [1,1-(PPh₃)(L)-3-(NC₅H₅)-closo-RhSB ₉H₈] [L = C₂H₄ (3), CO (4)]. In CH₂Cl₂ at reflux temperature 2 undergoes a nido to closo transformation to afford [1,1-(PPh₃)₂-3-(NC ₅H₅)-closo-1,2-RhSB₉H₈] (5). Reaction of 2 with alkenes leads to hydrogenation and isomerization of the olefins. NMR spectroscopy indicates the presence of a labile phosphine ligand in 2, and DFT calculations have been used to determine which of the two phosphine groups is labile. Rationalization of the hydrogenation mechanism and the part played by the 2 → 3 nido to closo cluster change during the reaction cycle is suggested. In the proposed mechanism the classical hydrogen transfer from hydride metal complexes to olefins occurs twice: first upon coordination of the alkene to the rhodium centre in 2, and second concomitant with formation of a closo-hydridorhodathiaborane intermediate by migration of a BHB-bridging hydrogen atom to the metal. Reaction of H₂ with 3 or 5 regenerates 2, closing a reaction cycle that under catalytic conditions is capable of hydrogenating alkenes. Single-site versus cluster-bifunctional mechanisms are discussed as possible routes for H₂ activation.

Full text of DOI:10.1021/ja802993m

Published on Web 08/02/2008
Alkene Hydrogenation on an 11-Vertex Rhodathiaborane with
Full Cluster Participation
†
,†
Alvaro Alvarez, Ramo´n Mac´ıas,* Jonathan Bould,§, Mar´ıa Jose´ Fabra,‡
´
´
Fernando J. Lahoz,‡ and Luis A. Oro†,‡
Departamento de Qu´ımica Inorga´nica, Instituto UniVersitario de Cata´lisis Homoge´nea,
Instituto de Ciencia de Materiales de Arago´n, UniVersidad de Zaragoza-Consejo Superior de
InVestigaciones Cient´ıficas, 50009-Zaragoza, Spain, Institute of Inorganic Chemistry, Academy
ˇ
of Sciences of the Czech Republic 250 68 Husinec-Rezˇ 1001, Czech Republic, and School of
Chemistry, UniVersity of Leeds, LS2 9JT, United Kingdom
Received April 23, 2008; E-mail: rmacias@unizar.es
Abstract: The facile synthesis of the metallaheteroborane [8,8-(PPh3)2-nido-8,7-RhSB9H10] (1) makes
possible the systematic study of its reactivity. Addition of pyridine to 1 gives in high yield the 11-vertex
nido-hydridorhodathiaborane [8,8,8-(PPh3)2H-9-(NC5H5)-nido-8,7-RhSB9H9] (2). 2 reacts with C2H4 or CO
to form [1,1-(PPh3)(L)-3-(NC5H5)-closo-RhSB9H8] [L ) C2H4 (3), CO (4)]. In CH2Cl2 at reflux temperature 2
undergoes a nido to closo transformation to afford [1,1-(PPh3)2-3-(NC5H5)-closo-1,2-RhSB9H8] (5). Reaction
of 2 with alkenes leads to hydrogenation and isomerization of the olefins. NMR spectroscopy indicates the
presence of a labile phosphine ligand in 2, and DFT calculations have been used to determine which of
the two phosphine groups is labile. Rationalization of the hydrogenation mechanism and the part played
by the 2 f 3 nido to closo cluster change during the reaction cycle is suggested. In the proposed mechanism
the classical hydrogen transfer from hydride metal complexes to olefins occurs twice: first upon coordination
of the alkene to the rhodium centre in 2, and second concomitant with formation of a closo-hydridorho-
dathiaborane intermediate by migration of a BHB-bridging hydrogen atom to the metal. Reaction of H2 with
3 or 5 regenerates 2, closing a reaction cycle that under catalytic conditions is capable of hydrogenating
alkenes. Single-site versus cluster-bifunctional mechanisms are discussed as possible routes for H2
activation.
Introduction
formulae [8,8,8-(L)2(L′)-nido-8,7-RhSB9H10] (L, L′ ) PMe2Ph;
L ) PPh3, L′ ) CO; L ) PPh3, L′ ) CH3CN; L ) η2-dppm,
The high-yield synthesis of the metallaheteroborane [8,8-
(PPh3)2-nido-8,7-RhSB9H10] (1)1 from [RhCl(PPh3)3] and Cs-
[SB9H12] makes the systematic reaction chemistry of this
compound accessible. For example, it has been observed that
monodentate or bidentate phosphine ligands can replace the
metal-bound PPh3 groups in 1 and also fill the vacant coordina-
tion site on the formally 16-electron rhodium center. Alterna-
tively, some phosphine ligands have been found to substitute
the exo-terminal hydrogen atom on the boron vertex at the 9
position of 1.2–5 These reactions have resulted in the preparation
of a series of 11-vertex nido-rhodathiaboranes of general
L′ ) η1-dppm) and [8,8,8-H(η2-L2)-nido-8,7-RhSB9H9-9-(η1-
L2)] (L2 ) dppe, dppp). Further observation made while studying
these systems was a nido to closo transformation in the clusters
with formal release of H2 leading to 11-vertex closo-rhodathi-
aboranes with B-L bonds.2,3 Some of these metallathiaboranes
have interesting structures and chemistry. For example, [8,8,8-
(η2-dppm)(η1-dppm)-nido-8,7-RhSB9H10] reacts with BH3 ·THF
to give an 11-vertex rhodathiaborane in which a BH3 adduct
coordinates to the rhodium atom in a bidentate fashion via its
hydrogen atoms.6 Recently, we extended this family of 11-vertex
metallathiaboranes with the development of an effective route
to N-heterocyclic-ligated clusters from reactions of [Rh(η4-
diene)(L2)] [diene ) cod or nbd; L2 ) bpy, Me2bpy, phen, (py)2]
with Cs[SB9H12].7 This synthetic procedure allows the prepara-
tion of hybrid salts [Rh(η4diene)(L2)][SB9H12] that afford
unsaturated and saturated clusters of general formulation [8,8-
(L2)-nido-RhSB9H10] [L2 ) bpy, Me2bpy, phen, (py)2] and
[8,8,8-(L2)(L′)-nido-RhSB9H10] (L2 ) bpy, Me2bpy, phen; L′
) PPh3 or CH3CN), respectively. This group of 11-vertex
rhodathiaboranes is an emerging example of the potential for
† Instituto Universitario de Cata´lisis Homoge´nea.
‡ Instituto de Ciencia de Materiales de Arago´n.
§ Academy of Sciences of the Czech Republic.
University of Leeds.
(1) Ferguson, G.; Jennings, M. C.; Lough, A. J.; Coughlan, S.; Spalding,
T. R.; Kennedy, J. D.; Fontaine, X. L. R.; Stibr, B. J. Chem. Soc.,
Chem. Commun. 1990, 891–894.
(2) Mac´ıas, R.; Rath, N. P.; Barton, L. Organometallics 1999, 18, 3637–
3648.
(3) Coughlan, S.; Spalding, T. R.; Ferguson, G.; Gallagher, J. F.; Lough,
A. J.; Fontaine, X. L. R.; Kennedy, J. D.; Stibr, B. J. Chem. Soc.,
Dalton Trans. 1992, 2865–2871.
(4) Adams, K. J.; McGrath, T. D.; Rosair, G. M.; Weller, A. S.; Welch,
A. J. J. Organomet. Chem. 1998, 550, 315–329.
(5) Ferguson, G.; Lough, A. L.; Coughlan, S.; Spalding, T. R. Acta
Crystallogr., Sect. C: Cryst. Struct. Commun. 1992, C48, 440–443.
(6) Mac´ıas, R.; Rath, N. P.; Barton, L. Angew. Chem., Int. Ed. 1999, 38,
162–164.
(7) Alvarez, A.; Mac´ıas, R.; Fabra, M. J.; Martin, M. L.; Lahoz, F. J.;
Oro, L. A. Inorg. Chem. 2007, 46, 6811–6826.
9
10.1021/ja802993m CCC: $40.75
2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 11455–11466 11455

´
Alvarez et al.
A R T I C L E S
novel chemistry that exists in metallaheteroboranes other than
metallacarboranes.
development of new rhodacarboranes and metal complexes
partnered with monocarborane ligands, which exhibit catalytic
activity in the hydrogenation of olefins.27–33 These studies,
however, have not given rise to fundamental new mechanisms
for either activation of H2 or hydrogenation and isomerization
of olefins. Here we report on the reactivity of 11-vertex nido-
and closo-rhodathiaboranes with alkenes and H2, which reveals
new mechanisms for hydrogenation of olefins and activation
of H2 where the flexible structure of the metallaheteroborane
cluster fully participates in the reactions.
The reactivity of metallaboranes and metallacarboranes with
unsaturated organic molecules is well documented,8–14 and the
outcomes are interesting and diverse. In some cases, the
metallaborane clusters react with the unsaturated molecule,
leading to insertion of heteroatoms,9,14 reduction of the organic
substrate, or both.10,12 Some of the reported reactions are
remarkable. For example, reaction of the iridaborane [6,6-
(PPh3)H-µ-6P,5C-(Ph2P-o-C6H4)-nido-6-IrB9H12] with acetylene
affords [1,1,1-(C4H4)-µ-1P,2C-(Ph2P-o-C6H4)-isocloso-1-IrB9H7-
5-(PPh3)] and [10-(PPh3)-2,2,2-(PH3)2(Ph2P-o-C6H4)-closo-2-
IrB9H7-1].15 The former compound is the result of dimerization
of HCCH at the metal center, whereas the latter is the result of
the reductive stripping of PPh3 ligands and a nido to closo cluster
transformation. Other reductions of alkynes by metallaboranes
are better understood and lead to highly functionalized metall-
aboranes with organic substituents at the boron vertices.12
In polyhedral boron chemistry it has long been a goal to
develop catalytic cycles by combining the oxidative and coor-
dinative flexibility of transition-metal elements with the capabil-
ity of boron clusters to exhibit oxidative/reductive flexibility in
their classical closo-nido-arachno transformations. Some
metallaboranes do exhibit catalytic activity in, for example,
oligomerization of alkynes,16 but although significant examples
exist,17–19 polyhedral boron chemistry shows a marked lack of
reaction cycles either catalytic or stoichiometric. Of these, the
catalytic hydrogenation and isomerization of olefins by 12-vertex
rhodadicarboranes are probably the best-known catalytic pro-
cesses in boron chemistry.20–26 These pioneering studies led to
new mechanisms that later served as inspiration for the
Results and Discussion
Scheme 1 can be used as a guide to the sections that follow.
The synthesis and selected reactivity of 2 and 3 have been
described earlier.34 Herein we discuss the structures and
reactivities of these 11-vertex clusters in more detail and present
the energy-optimized structures of these and some new rho-
dathiaboranes as well as gauge-independent atomic orbital
(GIAO) NMR nuclear shielding predictions. Additionally, the
behavior of 2 in its reaction with ethene is analyzed using DFT
calculations, leading to new proposed mechanisms for the
activation of dihydrogen and the catalytic isomerization and
hydrogenation of olefins by these rhodathiaboranes.
[8,8,8-(PPh3)2H-9-(NC5H5)-nido-8,7-RhSB9H9] (2). Reaction
of [8,8-(PPh3)2-nido-8,7-RhSB9H10] (1) with a 4-fold excess of
pyridine at room temperature affords the nido-hydridorhodathi-
aborane 2 in 75 % yield. The solid-state structure of 2 comprises
an 11-vertex nido-RhSB9 skeleton of the same type as the parent
compound 1 (Figure 1). Electronically, however, 2 has an
additional skeletal electron pair (13 sep), conforming to the
electron counting formalism,35,36 and it is thus saturated with
respect its precursor 1 (12 sep). This does not, however, lead
to substantial differences of intramolecular distances and angles
between 1 and 2 (Table 1). It is noteworthy, nevertheless, that
1 exhibits a larger disparity in the Rh-P distances than its
pyridine-ligated derivative 2. The longest Rh-P distance
between the two compounds corresponds to the PPh3 trans to
the B(3)-B(4) edge in 1 (see Figure 1 for cluster numbering).
Interestingly, the disposition of the two PPh3 ligands with respect
to the {RhSB9} fragment changes upon reaction with the
N-heterocyclic ligand, yielding a configuration with one PPh3
group trans to the B(3)-B(4) edge and the other trans to the
boron in the 9 position. The metal hydride occupies the third
pseudo-octahedral metal position trans to the sulfur atom, which,
in the unsaturated rhodathiaborane 1, is occupied by a PPh3
(8) Kalb, W. C.; Demidowicz, Z.; Speckman, D. M.; Knobler, C.; Teller,
R. G.; Hawthorne, M. F. Inorg. Chem. 1982, 21, 4027–4036.
(9) de Montigny, F.; Mac´ıas, R.; Noll, B. C.; Fehlner, T. P.; Costuas, K.;
Saillard, J.-Y.; Halet, J.-F. J. Am. Chem. Soc. 2007, 129, 3392–3401.
(10) Yan, H.; Noll, B. C.; Fehlner, T. P. J. Organomet. Chem. 2006, 691,
5060–5064.
(11) Fehlner, T. P. Pure Appl. Chem. 2006, 78, 1323–1331.
(12) Yan, H.; Noll, B. C.; Fehlner, T. P. J. Am. Chem. Soc. 2005, 127,
4831–4844.
(13) Bould, J.; Bown, M.; Coldicott, R. J.; Ditzel, E. J.; Greenwood, N. N.;
Macpherson, I.; MacKinnon, P.; Thornton-Pett, M.; Kennedy, J. D.
J. Organomet. Chem. 2005, 690, 2701–2720.
(14) Bould, J.; Rath, N. P.; Barton, L. Organometallics 1996, 15, 4916–
4929.
(15) Bould, J.; Brint, P.; Fontaine, X. L. R.; Kennedy, J. D.; Thornton-
Pett, M. J. Chem. Soc., Chem. Commun. 1989, 1763–1765.
(16) Yan, H.; Beatty, A. M.; Fehlner, T. P. Angew. Chem., Int. Ed. 2001,
40, 4498–4501.
(17) Kadlecek, D. E.; Carroll, P. J.; Sneddon, L. G. J. Am. Chem. Soc.
2000, 122, 10868–10877.
(26) Belmont, J. A.; Soto, J.; King, R. E.; Donaldson, A. J.; Hewes, J. D.;
Hawthorne, M. F. J. Am. Chem. Soc. 1989, 111, 7475–7486.
(27) Teixidor, F.; Nun˜ez, R.; Flores, M. A.; Demonceau, A.; Vin˜as, C. J.
Organomet. Chem. 2000, 614-615, 48–56.
(18) Littger, R.; Englich, U.; Ruhlandt-Senge, K.; Spencer, J. T. Angew.
Chem. Int. Ed 2000, 39, 1472–1474.
(19) Mac´ıas, R.; Fehlner, T. P.; Beatty, A. M. Angew. Chem. Int. Ed. 2002,
41, 3860–3862.
(28) Teixidor, F.; Flores, M. A.; Vinas, C.; Kivekas, R.; Sillanpaa, R.
Angew. Chem., Int. Ed. 1996, 35, 2251–2253.
(20) Long, J. A.; Marder, T. B.; Behnken, P. E.; Hawthorne, M. F. J. Am.
Chem. Soc. 1984, 106, 2979–2989.
(29) Vinas, C.; Flores, M. A.; Nunez, R.; Teixidor, F.; Kivekas, R.;
Sillanpaa, R. Organometallics 1998, 17, 2278–2289.
(30) Douglas, T. M.; Molinos, E.; Brayshaw, S. K.; Weller, A. S.
Organometallics 2007, 26, 463–465.
(21) Behnken, P. E.; Busby, D. C.; Delaney, M. S.; King, R. E.;
Kreimendahl, C. W.; Marder, T. B.; Wilczynski, J. J.; Hawthorne,
M. F. J. Am. Chem. Soc. 1984, 106, 7444–7450.
(22) Baker, R. T.; Delaney, M. S.; King, R. E.; Knobler, C. B.; Long, J. A.;
Marder, T. B.; Paxson, T. E.; Teller, R. G.; Hawthorne, M. F. J. Am.
Chem. Soc. 1984, 106, 2965–2978.
(31) Moxham, G. L.; Douglas, T. M.; Brayshaw, S. K.; Kociok-Kohn, G.;
Lowe, J. P.; Weller, A. S. Dalton Trans. 2006, 5492–5505.
(32) Molinos, E.; Kociok-Kohn, G.; Weller, A. S. Chem. Commun. 2005,
3609–3611.
(23) Knobler, C. B.; Marder, T. B.; Mizusawa, E. A.; Teller, R. G.; Long,
J. A.; Behnken, P. E.; Hawthorne, M. F. J. Am. Chem. Soc. 1984,
106, 2990–3004.
(33) Rifat, A.; Patmore, N. J.; Mahon, M. F.; Weller, A. S. Organometallics
2002, 21, 2856–2865.
(24) Long, J. A.; Marder, T. B.; Hawthorne, M. F. J. Am. Chem. Soc. 1984,
106, 3004–3010.
(34) Alvarez, A.; Mac´ıas, R.; Fabra, M. J.; Lahoz, F. J.; Oro, L. A. J. Am.
Chem. Soc. 2008, 130, 2148–2149.
(25) Behnken, P. E.; Belmont, J. A.; Busby, D. C.; Delaney, M. S.; King,
R. E.; Kreimendahl, C. W.; Marder, T. B.; Wilczynski, J. J.;
Hawthorne, M. F. J. Am. Chem. Soc. 1984, 106, 3011–3025.
(35) Wade, K. AdV. Inorg. Chem. Radiochem. 1976, 18, 1–66.
(36) Mingos, D. M. P.; Wales, D. J. Introduction to Cluster Chemistry;
Prentice Hall: London, 1990.
9
11456 J. AM. CHEM. SOC. VOL. 130, NO. 34, 2008

Alkene Hydrogenation on an 11-Vertex Rhodathiaborane
A R T I C L E S
Scheme 1
ligand. The length of the N-B distance falls within the reported
range of 1.230-1.978 Å found for this type of interaction (CSD
search, version 5.29).
compound 2a from the energy-optimized Rh-P distance of
2.4-4.0 Å leads to an increase in energy of 56 kJ/mol. This
energy difference is very similar to the experimentally deter-
mined ∆G‡ of 55 kJ/mol.
278
Although the two Rh-P distances in 2 are similar, the
variable-temperature 31P{1H} NMR spectra demonstrate that one
PPh3 ligand is labile, undergoing a fast dissociation at ambient
temperature. In order to ascertain the position of the labile
phosphine, we performed a relaxed scan of the potential-energy
surface on the model [8,8,8-(PH3)2H-9-(NC5H5)-nido-8,7-
RhSB9H9] (2a) by varying the Rh-P distances of the phos-
phines. The results of these studies are summarized in Figure
2, which illustrates that elongation of the PPh3 trans to the
B(3)-B(4) edge [P(2) in Figure 2] leads to a higher destabiliza-
tion of the cluster, and we therefore conclude that the PPh3
ligand trans to the pyridine-ligated boron atom B(9) is the labile
phosphine [P(1) in Figure 2]. Elongation of P(1) in the model
Table 2 gathers the measured NMR data of [8,8-(PPh3)2-nido-
8,7-RhSB9H10] (1) together with those for [8,8,8-(PPh3)2H-9-
(NC5H5)-nido-8,7-RhSB9H9] (2). A comparison of GIAO NMR
shift calculations with 11B NMR measured data was used to
assign individual boron resonances to structural positions as
defined by X-ray crystallography. The calculated values in Table
2 are presented in brackets. The agreement between the
computed 11B chemical shifts for [8,8-(PH3)2-nido-8,7-RhSB9-
H10] (1a) and [8,8,8-(PH3)2H-9-(NC5H5)-nido-8,7-RhSB9H9]
(2a) and the measured values for compounds 1 and 2 are
satisfactory, indicating that the DFT-calculated structures are
good models of the crystallographically determined structures.
It is interesting to note that the 1H NMR resonance of the BHB
bridging hydrogen atom in the hydridorhodathiaborane exhibits
a shift to high field of 0.07 ppm with respect to that of 1, a
small change considering that the BHB resonance is very
sensitive to substitution at either the rhodium center or the boron
atom at the 9 position.7 Overall, the 11B NMR spectrum of 2
exhibits a slight shielding with respect to that of 1. It is
surprising however that the 11B chemical shift of the pyridine-
substituted boron atom located at the 9 position in 2 does not
significantly change with respect to the parent unsubstituted
rhodathiaborane 1. These observations suggest that the higher
electron density in 2 is delocalized over the whole framework,
and we therefore attribute the labile nature of one of the PPh3
ligands to the global effect of the N-substituted {RhSB9} cluster
rather than to localized trans effects.
Figure 1. Molecular structure of [8,8,8-(PPh3)2H-9-(NC5H5)-nido-8,7-
RhSB9H9] (2).
9
J. AM. CHEM. SOC. VOL. 130, NO. 34, 2008 11457

´
Alvarez et al.
A R T I C L E S
Table 1. Selected Interatomic Distances (Å) for 2 and 3 with Estimated Standard Uncertainties (s.u.) in Parenthesesa
2
3
Rh(8)-S(7)
2.431(2)[2.431]
2.354(2)[2.387]
2.341(2)[2.347]
2.201(10)[2.303]
2.217(11)[2.243]
2.220(9)[2.201]
1.980(9)[1.999]
2.059(11)[2.073]
1.953(9)[1.943]
1.547(11)[1.565]
1.906(14) [1.902]
1.727(14) [1.737]
1.7845(14)
100.27(7) [98.71]
96.38(7)[97.91]
102.05(8) [101.36]
120.6(5) [117.4]
162.4(3) [159.3]
95.8(3) [99.82]
87.1(3) [87.5]
Rh(1)-S(2)
2.3721(8) [2.453]
2.2980(8) [2.311]
2.171(4) [2.205]
2.166(4) [2.200]
2.080(4) [2.094]
2.503(4) [2.516]
2.427(4) [2.590]
2.365(4) [2.396]
2.376(4) [2.404]
1.383(5) [1.405]
1.539(5)[1.539]
1.903(6) [1.904]
1.697(5) [1.702]
1.791(5)
112.86(3) [104.25]
109.12(11) [116.25]
93.39(10) [93.34]
89.33(10) [94.11]
71.60(14) [71.21]
127.8(2) [126.9]
Rh(8)-P(1)
Rh(1)-P(1)
Rh(8)-P(2)
Rh(1)-C(1)
Rh(8)-B(3)
Rh(1)-C(2)
Rh(8)-B(4)
Rh(1)-B(3)
Rh(8)-B(9)
Rh(1)-B(4)
S(7)-B(2)
Rh(1)-B(5)
S(7)-B(3)
Rh(1)-B(6)
S(7)-B(11)
Rh(1)-B(7)
N(1)-B(9)
C(1)-C(2)
B(2)-B(3) (longest)
B(6)-B(11) (shortest)
B-B (average)
P(1)-Rh(8)-P(2)
P(1)-Rh(8)-S(7)
P(2)-Rh(8)-S(7)
Rh(8)-B(9)-N(1)
P(1)-Rh(8)-B(9)
P(2)-Rh(8)-B(9)
S(7)-Rh(8)-B(9)
N(1)-B(3)
B(4)-B(8) (longest)
B(3)-B(6) (shortest)
B-B (average)
S(2)-Rh(1)-P(1)
B(3)-Rh(1)-P(1)
C(1)-Rh(1)-P(1)
C(2)-Rh(1)-P(1)
C(1)-Rh(1)-C(2)
N(1)-B(3)-Rh(1)
a Corresponding calculated distances for the model compounds 2a and 3a are enclosed in brackets.
Figure 2. DFT-calculated energies (kJ/mol) computed at the B3LYP/6-31G*/LANL2DZ level for elongation of the PH3 ligands in the model [8,8,8-
(PH3)2H-9-(NC5H5)-nido-8,7-RhSB9H9] (2a).
As previously reported,2 the formally unsaturated nido-
rhodathiaborane 1 contains two main centers of reactivity: the
rhodium atom and the boron atom at the 9 position. The rhodium
center undergoes ligand substitution reactions and, in some
cases, addition of another ligand at the third available coordina-
tion site (Scheme 1, structures I-III). It is also known that
some bidentate phosphines such as dppe and dppp can attack
the B(9) vertex, engendering a concomitant shift of the B-
H(terminal) hydrogen atom to the metal, which leads to
hydridorhodathiaborane analogues of 2 (Scheme 1, structure I).2
However, these phosphine-ligated derivatives exhibit labile P-B
bonds, which in dilute solutions lead to mixtures of the
corresponding free phosphine and the unsaturated rhodathiabo-
rane chelates (Scheme 1). Thus, it is somewhat surprising that
the pyridine attacks selectively the boron vertex in the 9 position
to afford a high yield of the new hydridorhodathiaborane 2.
[1,1-(PPh3)(η2-C2H4)-3-(NC5H5)-closo-1,2-RhSB9H8] (3). In an
atmosphere of CH2dCH2 the hydridorhodathiaborane 2 affords
CH3-CH3 and [1,1-(PPh3)(η2-C2H4)-3-(NC5H5)-closo-1,2-
RhSB9H8] (3). The solid-state structure of 3 is shown in Figure
3, where it can be seen that this rhodathiaborane is an 11-vertex
cluster based on the canonical structure of an octadecahedron
that may be rationalized by simple application of the Wade/
Mingos approach.35,36 The rhodium atom occupies the apical
cluster position connected to six atoms in the {SB9} fragment
with two exo-polyhedral metal coordination sites occupied by
the PPh3 and C2H4 ligands. Compound 3 is isoelectronic with
previously reported 11-vertex rhodathiaboranes such as [1,1-
(PPh3)(CO)-3-(PPh3)-closo-1,2-RhSB9H8],[1,1-(η2-Ph2PCH2CH2PPh2)-
3-(η1-Ph2PCH2CH2PPh2O)-closo-1,2-RhSB9H8] and others,2,3
which bear a phosphine substituent on B(3). The metrics do
not differ significantly among these rhodathiaborane adducts,
9
11458 J. AM. CHEM. SOC. VOL. 130, NO. 34, 2008

Alkene Hydrogenation on an 11-Vertex Rhodathiaborane
A R T I C L E S
Table 2. 11B (96 MHz), 1H (300 MHz), and 31P (162 MHz) NMR
Data for 1 and 2 in CD2Cl2 Compared to the Corresponding
DFT/GIAO-Calculated 11B-Nuclear Shielding Values [in brackets]
for the Related P-Hydryl models 1a and 2a
A plausible reaction pathway for formation of the ethene-
ligated closo-rhodathiborane 3 from the nido-hydridorhoda-
thiaborane 2 might start with substitution of the labile phosphine
[P(1) in Figure 2] and subsequent coordination of ethene to the
rhodium center. In order to identify the potential course of the
reductive elimination, we carried out preliminary DFT calcula-
tions using the model [8,8-(PH3)(η2-C2H4)H-9-(NC5H5)-nido-
8,7-RhSB9H9] (3a) as a starting point in the search for
intermediates that may play a part in this reaction. The cal-
culations suggest that reduction of the coordinated ethene group
by the metal hydride produces an ethyl-ligated intermediate Ia
stabilized by 17 kJ/mol. Ia can be formed as the result of an
intramolecular ethene migratory insertion into the Rh-H bond
followed by rearrangement of the exo-polyhedral ligands.
Reductive elimination of the ethyl group would then be mediated
by an intermediate IIa, which lies 8 kJ/mol higher than the ethyl-
ligated intermediate Ia. Formation of IIa implies migration of
the BHB hydrogen atom from the B(9)-B(10) edge to the
rhodium atom and a nido to closo cluster oxidation. This
transformation resembles the recently calculated transition state
along the reaction coordinate for the observed fluxional process
that interchanges the two enantiomeric 11-vertex nido cages of
the parent compound 1.38 The final reductive elimination of
ethane results in formation of a system IIIa that is stabilized
by an agostic interaction between ethane and rhodium as
illustrated in Figure 4 (Rh · · ·H ) 2.13 Å).39 In this intermediate,
ethane could be easily displaced by ethene, which is in excess
in the reaction mixture.
Cluster data at 300 K
1
2
1
assignmenta
δ(11B)
δ(1H)
δ(11B)
δ(1H) 1J(11B- H)
9
3
6
11
4
5
10
1
+12.6 [+18.7] +3.59 +11.8 [+17.7] NC5H5
+16.3 [+23.1] +4.14 +7.9 [+9.4] +3.51 127 Hz
+4.5 [+7.1] +1.56 +3.2 [+7.4] +2.61b
+6.0 [+10.1] +2.32 +0.3 [+2.3] +4.12
+2.5 [+2.1] +2.38 -3.7 [+0.0] +2.84
-9.3 [-9.2] +1.78 -10.0 [-10.7] +1.84 133 Hz
-21.1 [-19.5] +1.24 -18.1 [-23.0] +0.91
-18.6 [-17.7] +3.26 -25.6 [-24.9] +1.49 142 Hz
-27.5 [-28.6] +1.28 -28.5 [-28.3] +1.12
-1.32 -1.39
c
c
c
c
2
c
µ (9,10)
PPh3 data for 1 at 223 K and 2 at 198 K
1
2
assignment δ(31P) 1J(135Rh-31P) 2J(31P1-31P2) δ(31P) 1J(135Rh-31P) 2J(31P1-31P2)
P(1) (br) +43.0 160 Hz
P(2) +20.3 128 Hz
+37.0d 106 Hz
36 Hz
16 Hz
+29.7 128 Hz
a Assignments based on 11B selective experiments and DFT calcu-
lations. b Doublet, 25 Hz. c Coupling constants not measured due to
broadness of peaks relative to 1J. d At higher temperatures P(1) broadens
losing the 1J(135Rh-31P) and 2J(31P1-31P2) couplings; coalescence
temperature 278 K.
We have not carried out a comprehensive mapping of the
potential-energy surface, leading to characterization of the
intermediate transition states between the potential-energy wells
illustrated in Figure 4. However, these energy-optimized minima
provide a useful guide for consideration of the reaction
mechanism and a useful base for future experimental and
calculational work. The results demonstrate that elimination of
ethane from the ethene-ligated nido-hydridorhodathiaborane
model 3a is theoretically exothermic by 57 kJ/mol (Figure 4),
which conforms to the energy required for dissociation of the
PPh3 ligand in 2 (Figure 2).
Other possible reaction pathways for formation of 3 from 2
are (i) direct loss of dihydrogen upon displacement of PPh3 by
ethene and (ii) spontaneous dehydrogenation of 2 to give a bis-
PPh3-ligated closo derivative that could react with C2H4. In this
regard, we found that bubbling CO(g) through a solution of 2
in dichloromethane at room temperature affords [1,1-(PPh3)(CO)-
3-(NC5H5)-closo-1,2-RhSB9H8] (4) in about 20 min. In addition,
the nido-hydridorhodathiaborane 2 loses dihydrogen after reflux
in CH2Cl2 for 48 h, leading to formation of [1,1-(PPh3)2-3-
(NC5H5)-closo-1,2-RhSB9H8] (5). This suggests that both
pathways are in principle possible for formation of the ethene-
ligated closo-rhodathiaborane 3. The second route may, how-
ever, be discounted because the nido compound 2 is stable in
CH2Cl2 at room temperature for at least 5 days, whereas in an
atmosphere of ethene it affords the closo derivative 3 in 2 h
(see Figures S1 and S2 in the Supporting Information). However,
we cannot discount that, similarly to the CO ligand, ethene could
promote dihydrogen elimination upon displacement of the labile
Figure 3. Molecular structure of [1,1-(PPh3)(η2-C2H4)-3-(NC5H5)-closo-
1,2-RhSB9H8] (3).
which, as a common pattern, exhibit a Rh(1)-B(3) length
shorter than the other four Rh-B distances;2,3 this is general to
11-vertex closo-metallaborane and binary borane clusters due
to the geometry of the canonical octadecahedron. The CdC
bond distance in 3 [1.383(5) Å] is similar to those in reported
ethene-ligated organometallic compounds [CSD search (version
5.29) for the M(η2-C2H4) fragment; M ) transition element].
The 1H NMR spectrum at room temperature exhibits two
multiplets at +2.27 and +2.05 ppm, which spilt into three
signals of 1:1:2 relative intensity at lower temperatures and
coalesce over the range 240-260 K (see Figure S5 in the
Supporting Information). This temperature-dependent behavior
demonstrates that the protons of the ethene ligand undergo
exchange. The chemical nonrigidity of the Rh-C2H4 linkage
has been previously reported, and the proposed mechanism
consists of rotation of the ethene about the rhodium-carbon
bonds.37
(37) Cramer, R.; Kline, J. B.; Roberts, J. D. J. Am. Chem. Soc. 1969, 91,
2519–2524.
ˇ
(38) Mac´ıas, R.; Bould, J.; Holub, J.; Kennedy, J. D.; Stibr, B.; Thornton-
Pett, M. Dalton Trans. 2007, 2885–2897.
(39) Brookhart, M.; Green, M. L. H.; Parkin, G. Proc. Natl. Acad. Sci.
U.S.A. 2007, 104, 6908–6914.
9
J. AM. CHEM. SOC. VOL. 130, NO. 34, 2008 11459

´
Alvarez et al.
A R T I C L E S
Figure 4. Potential-energy profile for the reductive elimination of CH2dCH2 from 3a to give CH3-CH3 and a nido to closo cluster transformation.
Table 3. 11B (96 MHz) and 1H (300 MHz) NMR Data for 3-5 in CD2Cl2 Compared to the Corresponding DFT/GIAO-Calculated 11B-Nuclear
Shielding Values [in Brackets] For the Related P-Hydryl Models 3b, 4a, and 5a
3
4
5
assignmenta
δ(11B)
δ(1H)b
δ(11B)
δ(1H)
δ(11B)
δ(1H)
3
9
5
4
8
7
6
10
11
+55.7
+26.4c
+0.8
[+58.5]
[+28.5]
[+9.0]
NC5H5
+4.36
+1.80
+2.06
+2.46
+0.54
-0.17
+0.42
-0.01
+55.4
+28.1e
+1.1
[+58.5]
[+30.4]
[+8.9]
NC5H5
+4.34
+2.39
+1.08
+2.37
+0.66
+0.08
+0.07
+0.00
+54.6
+27.3g
-0.5
[+58.4]
[+30.1]
[+6.7]
NC5H5
+4.09
-0.1
[+4.2]
-0.7
[+4.2]
(2B)
[+5.4]
+1.27
+2.16
-14.5
-22.2
-24.6
-30.3d
(2B)
[-13.3]
[-16.8]
[-22.0]
[-27.4]
[-27.0]
-14.4
-26.1
-24.4
-31.9f
(2B)
[-13.4]
[-21.8]
[-17.1]
[-26.7]
[-29.2]
-15.2
-24.2h
(2B)
[-12.2]
[-18.1]
[-21.8]
[-29.4]
[-29.6]
-0.22
-0.27
-30.4i
(2B)
a Assignments based on 1H{11B} selective experiments and DFT calculations. b Boron-bound proton resonances. c 1J(11B-1H) ) 124 Hz.
d Accidentally coincident 11B resonances; 1J(11B-1H) ) 97 Hz. e 1J(11B-1H) ) 118 Hz. f Accidentally coincident 11B resonances; 1J(11B-1H) ) 137
Hz. g 1J(11B-1H) ) 144 Hz. h 1J(11B-1H) ) 129 Hz. i 1J(11B-1H) ) 139 Hz.
PPh3 and subsequent coordination to the rhodium atom: dihy-
drogen loss and ethene reductive elimination may be competing
reactions in formation of 3.
the absence of free PPh3, the closo-rhodathiaborane decomposes
within several hours. By contrast, reaction of 3 with H2(g) in
acetone leads to formation of the nido-hydridorhodathiaborane
2 together with decomposition of the closo-rhodathiaborane 3
[Figure 5c]. In the case of the bis-PPh3-ligated closo-rhodathi-
aborane 5, reaction with dihydrogen proceeds at the same initial
rate as for 3 in the presence of PPh3 [Figure 5b].
Hydrogenation of ethene upon reaction of [(PPh3)(η2-C2-
H4)-closo-RhS9BH8(NC5H5)] (3) with H2(g) takes place most
probably after formation of an ethene-ligated derivative of
formulation [(PPh3)(η2-C2H4)H-nido-RhS9BH9(NC5H5)] for which,
according to calculation, elimination of ethane should be
energetically favored. Hydrogenation of ethene would lead to
a coordinatively unsaturated system (IIIa in Figure 4) that, in
dichloromethane and the absence of free PPh3, decomposes.
However, in more coordinating solvents such as acetone, the
unsaturated species appears to be stabilized by formation of
solvent-ligated intermediates such as, for example, [(PPh3)(Me2-
CO)RhS9BH8(NC5H5)], which then reacts further with other
molecules of 3, sequestrating the PPh3 ligand and leading to
As shown in Table 3, the 11B-calculated nuclear shieldings
at the B3LYP/6-31G*/LANL2DZ level for the [1,1-(PH3)(L)-
3-(NC5H5)-closo-1,2-RhSB9H8] [L ) C2H4 (3b), CO (4a), PH3
(5a)] models are in good agreement with the measured 11B
chemical shifts for 3-5. The lowest field 11B resonance
corresponds to the pyridine-substituted boron atom at the 3
position, which appears at around +55 ppm. The second 11B
resonance at low field corresponds to the boron atom in the 9
position adjacent to the B(3) atom, after which there is a
significant gap (ca. 26 ppm) between these low-field resonances
and the rest.
Reactions with H2(g). The 11-vertex closo-rhodathiaboranes
3 and 5 react with dihydrogen (Figure 5), whereas the CO-
ligated analogue 4 does not. When the metal is ethene ligated,
as in compound 3, the outcome depends on the conditions, and
thus, in CH2Cl2 solvent with 1 equiv of free PPh3 the reaction
regenerates the nido-hydridorhodathiaborane 2 with formation
of ethane [Figure 5a]. However, if the reaction is carried out in
9
11460 J. AM. CHEM. SOC. VOL. 130, NO. 34, 2008

Alkene Hydrogenation on an 11-Vertex Rhodathiaborane
A R T I C L E S
1
Figure 5. Plots of reaction time vs conversion based on H NMR signal intensities for reaction of 3 and 5 with H2(g): (a) in CD2Cl2 with a 1:1 ratio of 3
to free PPh3, (b) 5 in CD2Cl2, and (c) 3 in (CD3)2CO (closo compounds 3 and 5, dark blue diamonds; nido-rhodathiaborane 2, magenta squares).
formation of the nido derivative 2 and sacrificial decomposition
of the parent ethene-ligated rhodathiaborane 3.
(Scheme 2). In this regard, it has been reported that the closo-
rhodadicarboranes [3-(η3-C3H5)-closo-3,1,2-RhC2B9H11] and
[3-(η3-C8H13)-1-SR-2-R′-closo-3,1,2-RhC2B9H9] (R′ ) Ph, Me)
activate H2.43,44 As a tentative mechanism, reaction with H2
was proposed to occur through a η3 to η1 rearrangement of the
η3-allyl ligand followed by oxidative addition of H2 to produce
an unstable Rh(V) alkyl dihydride species or alternatively by a
concerted four-center transition state involving the η1-allyl-Rh
moiety and H2.
The closo-rhodathiaboranes 3-5 may be related to the cy-
clopentadienyl-containing complexes [(η5-C5H5)Rh(η2-C2H4)2]
and [(η5-C5H5)Rh(PPh3)(η2-C2H4)]40,41 by regarding the {SB9-
H8(NC5H5)} fragment in compounds 3-5 as a polyhedral
analogue of the C5H5 ligand. On the basis of this comparison,
these 11-vertex closo-rhodathiaboranes may be thought of as
18-electron complexes of {Rh(PPh3)(L)}+ [L ) η2-C2H4 (3),
CO (4), and PPh3 (5)] metal moieties with the {SB9H8(NC5H5)}
fragment as a hexahapto monoanionic ligand. It has been
reported that [(η5-C5H5)Rh(η2-C2H4)2] undergoes photochemi-
cally initiated oxidative addition of H2(g) producing [(η5-
C5H5)Rh(η2-C2H4)H2], which eliminates ethane in a slow
thermal process.40 Thus, the η5-C5H5 organometallic analogue
activates dihydrogen after formation of a coordinatively unsatur-
ated 16-electron complex. At first sight, this single-site H2
activation appears to be unlikely in the closo-rhodathiaboranes
3 and 5 since the rhodium centers are coordinatively saturated,
and we have not observed exo-polyhedral ligand dissociation
processes that could facilitate dihydrogen coordination at the
metal center. On the other hand, it has been reported that
heterometallic complexes of formula [Pt3Re2(CO)6(PtBu3)3] and
[Pt2Re2(CO)7(PtBu3)3(µ-H)2] activate H2 by addition to a single
metal site without generation of a vacant site.42 These mixed-
metal clusters are highly unsaturated, and therefore, they
facilitate coordination of H2 to a metal vertex and its consequent
cleavage. Compounds 3 and 5 are saturated 11-vertex closo
clusters with 12 sep, and a priori, H2 addition to the rhodium
vertex does not appear to be favorable. However, given the
stereochemical flexibility of polyhedral boron-containing com-
pounds, a slight rearrangement of the 11-vertex closo-rhodathi-
aboranes could facilitate addition of H2, leading to its activation
Single-site H2 activation could be possible in the closo-
rhodathiaboranes 3 and 5 through slippage of the {(L)2Rh}
fragment above the boat-shaped {S(2)B(3)B(4)B(5)B(6)B(7)}
face, leading to a vacant metal site where H2 could bind.
However, given the bifunctional character of the ethene
hydrogenation mechanism illustrated in Figure 4, it is tempting
to speculate that H2 activation takes place on the rhodathiaborane
cluster in a concerted manner. A possible mechanism, depicted
in Scheme 2, could start with a closo to isonido rearrangement,
leading to a structure that features a quadrilateral [Rh(1)B(3)B(7)-
B(4)] open face.38 This isonido structure would accommodate
the H2 molecule in a bridging position between the Rh(1)-B(3)
edge and either of the B(3)-B(7) or B(4)-B(7) edges (closo
numbering). Cleavage of the H-H bond would then lead directly
to formation of the nido-hydridorhodathiaborane 2, or it could
afford an intermediate containing a Rh(1)-H-B(3) bridging
hydrogen atom as in Ib, Scheme 2, which would then rearrange
to 2. Calculations demonstrate that an intermediate such as Ib
lies 15 kJ/mol below the [8,8,8-(PH3)(L)H-9-(NC5H5)-nido-8,7-
RhSB9H9] models [L ) η2-C2H4 (3a) and CO (4b)] and 9 kJ/
mol above [8,8,8-(PH3)2H-9-(NC5H5)-nido-8,7-RhSB9H9] (2a)
(Figure 6). On the basis of these energy differences, we suggest
that a Rh(8)HB(9) hydrido intermediate with the structure
illustrated by Ib in Scheme 2 may play a role in the dehydro-
genative nido to closo transformation of 2 and in the reverse
H2 activation reaction by 5. The fact that the calculated energy
(40) Duckett, S. B.; Haddleton, D. M.; Jackson, S. A.; Perutz, R. N.;
Poliakoff, M.; Upmacis, R. K. Organometallics 1988, 7, 1526–1532.
(41) Campian, M. V.; Harris, J. L.; Jasim, N.; Perutz, R. N.; Marder, T. B.;
Whitwood, A. C. Organometallics 2006, 25, 5093–5104.
(42) Adams, R. D.; Captain, B. Angew. Chem., Int. Ed. 2008, 47, 252–
257.
(43) Walker, J. A.; Zheng, L.; Knobler, C. B.; Soto, J.; Hawthorne, M. F.
Inorg. Chem. 1987, 26, 1608–1614.
(44) Teixidor, F.; Flores, M. A.; Vinas, C.; Sillanpaa, R.; Kivekas, R. J. Am.
Chem. Soc. 2000, 122, 1963–1973.
9
J. AM. CHEM. SOC. VOL. 130, NO. 34, 2008 11461

´
Alvarez et al.
A R T I C L E S
Scheme 2
of the intermediate Ib is higher than that of 2a is consistent
with the stability of the nido-hydridorhodathiaborane 2 at room
temperature and its thermally promoted nido to closo transfor-
mation. Alternatively, a CO-ligated intermediate with the stru-
cture of Ib, which exhibits lower energy than a putative [8,8,8-
(PPh3)(CO)H-nido-8,7-RhSB9H9(NC5H5)] (analogue of 2), agrees
with the facile H2 elimination and nido to closo change upon
reaction with CO to give 4.
similar to the ethyl-nido-intermediate in the reductive elimination
of ethene (Ia in Figure 4). In line with the proposed H2 activation
mechanism (Scheme 2), the presumed intermediate [8,8,8-
(PPh3)(η2-C2H4)H-9-(NC5H5)-nido-8,7-RhSB9H9] could be an-
other intermediate that, in the reductive elimination of ethane
from the ethene-ligated closo-compound 3, lies above its
Rh-H-B isomer (Ib) and the ethyl-ligated intermediate (Figure
6).
In the case of the ethene-ligated system, the PPh3-metal-
bound model with the structure of Ib has a calculated energy
Reduction of the 11-vertex closo-rhodathiaboranes 3 and 5
with H2 is an interesting reaction since closo to nido transforma-
Figure 6. Energy comparisons for DFT-optimized isomeric structures of the model compounds (a) [8,8-(PH3)(η2-C2H4)-9-(NC5H5)-nido-8,7-RhSB9H8]
(3a), (b) [8,8,8-(PH3)2H-9-(NC5H5)-nido-8,7-RhSB9H8] (2a), and (c) [8,8,8-(PH3)(CO)H-9-(NC5H5)-nido-8,7-RhSB9H8] (4b).
9
11462 J. AM. CHEM. SOC. VOL. 130, NO. 34, 2008

Alkene Hydrogenation on an 11-Vertex Rhodathiaborane
A R T I C L E S
Scheme 3
tions in polyhedral boron-containing compounds usually require
strong reducing conditions.45–49 An indication of the difficulty
of this reaction is the fact that the CO-ligated closo-rhodathi-
aborane 4 does not react with H2, and this suggests that the CO
ligand stabilizes the 18-electron {RhPPh3(CO)}+ center by
means of its higher π-back-bonding capabilities, decreasing
consequently the half-cell reduction potential of the cluster. In
contrast, activation of H2 by the bis-PPh3-ligated closo-rho-
dathiaborane 5 occurs reversibly (Scheme 2) and is thus an
example of a cluster-assisted H2 activation and dehydrogenation.
Reaction with D2(g). To gain further insight into the mech-
anism of activation of H2 by the closo clusters 3 and 5, we
carried out the reaction with D2. A sample of 3 and 1 equiv of
PPh3 were placed in a screw-capped NMR tube, which was
opened to an atmosphere of D2. After 4 days, the 11B and
1H{11B} NMR spectra showed a 1:1 mixture of the nido-
hydridorhodathiaborane 2 and the bis-PPh3-ligated closo com-
pound 5. All the BH terminal resonances corresponding to the
door to the catalytic hydrogenation of olefins. To begin the study
of the catalytic activity of the nido/closo rhodathiaborane system,
we first investigated reactions of the nido cluster 2 with
cyclohexene and 1-hexene. A 20-fold excess of cyclohexene in
CH2Cl2 solvent at room temperature does not react with the
nido-hydridorhodathiaborane after 3 days. However, under the
same conditions 1-hexene afforded 17% of 2-hexene, 7% of
3-hexene, and 1% of hexane. This indicates that the system
undergoes phosphine dissociation followed by alkene coordina-
tion and Rh-alkyl formation. Reductive elimination of the
alkane is hindered by olefins higher than ethane: thus, with
1-hexene we observed formation of small amounts of hexane,
but with cyclohexene no cyclohexane was formed. Steric effects
may increase the energy barrier in the nido to closo transforma-
tion that, as calculated for ethene and depicted in Figure 4, may
mediate on the final reductive elimination of the Rh-ligated alkyl
group.
Hydrogenation experiments using 2 as the catalyst precursor
and cyclohexene as substrate afforded cyclohexane. In refluxing
CH2Cl2 the rhodathiaborane system reduces 19% of the olefin
in 8 h (GC yield, unoptimized). Similarly, in catalytic amounts,
the rhodathiaborane system promotes hydrogenation of 1-hexene
and its isomerization to the 2- and 3-isomers. For this terminal
olefin, after 5 h in CH2Cl2 at reflux, there is 4% formation of
hexane together with 36% of the isomers 2- and 3-hexene.
1
closo and the nido species were present in the H{11B} NMR
spectrum, but the BHB and RhH proton resonances of the nido
compound were absent. This indicates that there is no H/D
exchange at a reasonable rate between the directly boron-bound
hydrogen atoms and D2 and that the deuterium gas is activated
by the closo clusters, leading selectively to formation of the
deuterated derivative [8,8,8-(PPh3)2(D)-nido-8,7-RhSB9H8D-
(NC5H5)] containing a Rh(8)-D group and a B(9)-D-B(10)
bridging deuterium atom. These results are consistent with the
H2 activation mechanism proposed above.
On the basis of the computational and experimental studies
described above we propose a mechanistic model depicted in
Scheme 3 for isomerization and hydrogenation of alkenes by
the 2/3/5 system. The initial step in the catalytic cycle is
dissociation of the PPh3 ligand trans to the N-B-substituted
boron atom at the 9 position in 2 (Figure 2). This first step is
performed by both lower and higher olefins, leading to formation
of a Rh-alkyl group. At this point the system may evolve
toward the final reductive elimination or ꢀ-elimination of the
alkyl group. The latter reaction would be therefore responsible
for the observed isomerization of terminal alkenes such as the
studied 1-hexene.
Catalytic Hydrogenation and Isomerization of Olefins. The
reaction between ethene and the nido-hydridorhodathiaborane
2, combined with the activation of H2 by 3 and 5, opens a new
(45) Evans, W. J.; Hawthorne, M. F. Inorg. Chem. 1974, 13, 869–874.
(46) Evans, W. J.; Hawthorne, M. F. J. Chem. Soc., Chem. Commun. 1974,
38–9.
(47) Lopez, M. E.; Edie, M. J.; Ellis, D.; Horneber, A.; Macgregor, S. A.;
Rosair, G. M.; Welch, A. J. Chem. Commun. 2007, 2243–2245.
(48) Xie, Z. Acc. Chem. Res. 2003, 36, 1–9.
(49) Deng, L.; Xie, Z. Organometallics 2007, 26, 1832–1845.
9
J. AM. CHEM. SOC. VOL. 130, NO. 34, 2008 11463

´
Alvarez et al.
A R T I C L E S
According to the DFT calculations on the model 3a final
reductive elimination of the Rh-alkyl intermediate could
proceed through a closo-hydridorhodathiaborane intermediate
(Figure 4 IIa) formed by migration of the BHB bridging
hydrogen atom to the metal center and subsequent nido to closo
transformation. For alkenes higher than ethene, such as cyclo-
hexene and 1-hexene, reductive elimination from the Rh-alkyl
nido intermediate is hindered, suggesting that the nido to closo
conversion is the rate-determining step in the hydrogenation
process. After reductive elimination of the alkane, the resulting
unsaturated rhodathiaborane can be stabilized by either free PPh3
or the excess of olefin, forming the closo derivative 5 or alkene-
ligated analogues of 3, respectively, which, in an atmosphere
of H2, would regenerate the nido-hydridorhodathiaborane sys-
tem, thus closing the catalytic cycle (Scheme 3).
Catalytic hydrogenation and isomerization of olefins by 12-
vertex rhodacarboranes was first reported by Hawthorne and
co-workers.20–26 The catalytic activity of these compounds was
attributed to formation of B-Rh(III)-H species formed by
oxidative addition of terminal B-H bonds to Rh(I) centers in
exo-nido-rhodacarboranes. These exo-nido precursors are best
described as 16-electron Rh(I) square-planar complexes com-
prising an organometallic {RhL2} fragment bound to an 11-
vertex nido-carborane cluster, which is acting as a bidentate
chelating ligand via two B-H bonds. In the proposed
mechanism20–26 the B-H-Rh(I) species is in equilibrium with
the catalytically active B-Rh(III)-H intermediates in which
the nido-carborane is now bound in a monodentate fashion to
the metal center through a boron atom. Hydrogenation of olefins
by these metallacarboranes may therefore be regarded as a
single-site catalytic process similar to the mechanism usually
adopted by monohydride metal complexes.50
PPh3-ligated nido-hydridorhodathiaborane 2 can be thermally
or chemically promoted. Not surprisingly, the nature of the exo-
polyhedral metal ligands change the reactivity of the clusters
with dehydrogenation/H2-activation being reversible for the bis-
PPh3-ligated 2/5 system and irreversible for the CO-ligated closo
compound 4. The reactivity of the clusters described here
combines fundamental principles of organometallic chemistry
and metallaborane chemistry, leading to new mechanisms for
hydrogenation of olefins and activation of H2. These results are
an unusual overlap between polyhedral boron chemistry and
organometallic chemistry, demonstrating that, similarly to
transition-metal complexes, metallaheteroboranes may become
useful reagents for functionalization of organic substrates.
Although more mechanistic studies are needed in order to fully
determine the kinetics of the alkene hydrogenation and activation
of dihydrogen, the observed reactivity appears to be driven by
the redox and structural flexibility of the 11-vertex rhoda-
thiaboranes. These polyhedral compounds can be, in principle,
tailored for chemical hydrogen storage and hydrogenation of
olefins, suggesting a rich chemistry based on 11-vertex closo/
nido metallaheteroboranes.
Experimental Section
General Procedures. All reactions were carried out under an
argon atmosphere using standard Schlenk-line techniques.51 Sol-
vents were obtained from an Innovative Technology Solvent
Purification System. All commercial reagents were used as received
without further purification. The 10-vertex arachno-thiaborane salt,
CsSB9H12, was purchased from Katchem. The starting rhodathi-
aborane, [8,8-(PPh3)2-8,7-RhSB9H10] (1), was prepared by published
methods.1 C, H, N, and S analyses were carried out in a Perkin-
Elmer 2400 CHNS/O analyzer. Infrared spectra were recorded in
KBr with a Perkin-Elmer Spectrum One spectrometer. NMR spectra
were recorded on Bruker Avance 300 MHz and AV 400 MHz
spectrometers using 11B, 11B{1H}, 1H, 1H{11B}, 1H{11B(sele-
ctive)}, 13C{1H}, and [1H-1H]-COSY techniques. 1H and 13C
NMR chemical shifts were measured relative to partially deuterated
solvent peaks but are reported in ppm relative to tetramethylsilane.
11B chemical shifts were measured relative to [BF3(OEt)2)]. 31P
(121.48 MHz) chemical shifts were measured relative to H3PO4
(85%). The temperature of the probe was calibrated using the
temperature dependence of the chemical shifts of methanol. Mass
spectrometry data were recorded on a VG Autospec double-focusing
mass spectrometer, a microflex MALDI-TOF, and a ESQUIRE
3000+ API-TRAP, operating in either positive or negative modes.
GC analyses were performed on an Agilent 6890N Network System
equipped with a flame ionization detector and a 30 m (0.25 mm
i.d., 0.25 µm film thickness) HP-Ultra-1 column. X-ray crystal-
lographic files for 2 and 3 were deposited in the Supporting
Information of the preliminary communication (ref 34).
According to the mechanism proposed here, reduction of the
alkene takes place at the rhodium center in the 2/3/5 rhodathi-
aborane system, but supported by calculation, the cluster then
promotes the final transfer of the second hydrogen atom to the
Rh-alkyl group by means of a nido to closo transformation,
conferring a rhodium/thioborane bifunctional character to the
process. Nido to closo transformations are fairly common in
polyhedral boron chemistry.8–14 This is, however, the first time
that effective oxidation/reduction has been coupled catalytically
with hydrogenation of olefins.
Conclusions
The labile nature of the PPh3 ligand trans to the 9 boron in
[8,8,8-(PPh3)2H-9-(NC5H5)-nido-RhSB9H9] (2) promotes isomer-
ization and hydrogenation of olefins. DFT calculations and
experimental results support a rhodium thiaborane bifunctional
mechanism that, for the transfer of two hydrogen atoms to the
CdC bond, combines the first part of the classical single-site
monohydride route with a nido to closo cluster conversion. The
resulting 11-vertex closo-rhodathiaboranes 3 and 5 activate
dihydrogen, regenerating the nido parent compound 2 and
initiating a new catalytic cycle. The mechanism for activation
of H2 by the closo species is uncertain, but according to DFT
calculations and our experimental observations, a concerted
process with some bifunctional character involving interaction
of H2 with the metal center and thiaborane fragment is
reasonable. Dehydrogenative nido to closo oxidation of the bis-
Reaction of [8,8-(PPh3)2-nido-RhSB9H10] (1) with Pyridine.
An 8 mg amount of 1 (0.0104 mmol) was dissolved in 1 mL of
CD2Cl2 in a 5 mm NMR tube, and then 3.37 mL of pyridine was
added to the bright-red solution of the rhodathiaborane. The
resulting solution was examined by NMR spectroscopy at room
temperature. The data revealed the quantitative formation of [8,8,8-
(PPh3)2H-9-(NC5H5)-nido-RhSB9H9] (2). Anal. Calcd for C41H45B9-
NP2RhS·(CH2Cl2): C, 54.19; H, 5.09; N, 1.50; S, 3.44. Found: C,
54.24; H, 5.03; N, 1.60; S, 3.69. IR (KBr, disk): ν 2533 vs (BH),
1
2042 m (RhH). H NMR (300 MHz, CD2Cl2, 300 K): δ 8.56 (m,
2H; NC5H5), 7.86 (m, 1H; NC5H5), 7.74 (t, 7.6 Hz, 1H; NC5H5),
7.67 (m, 1H; NC5H5), 7.37-7.02 (30H; 2PPh3), -12.46 (apparent
q, 2J(H,P) + 1J(H,Rh) )18.3 Hz, 1H, Rh(PPh3)2H). 31P{1H}NMR
1
(162 MHz, CDCl3, 300 K): δ 33.2 [d, J(Rh,P) ) 128 Hz; PPh3].
(50) Oro, L. A.; Carmona, D. In Handbook of Homogeneous Hydrogena-
tion; De Vries, J. G., Elsevier, C. J., Eds.; Wiley-VCH: Weinheim,
Germany, 2007; Vol. 1
(51) Shriver, D. F.; Drezdzon, M. A. The Manipulation of Air-SensitiVe
Compounds, 2nd ed.; Wiley-Interscience: New York, 1986.
9
11464 J. AM. CHEM. SOC. VOL. 130, NO. 34, 2008

Alkene Hydrogenation on an 11-Vertex Rhodathiaborane
A R T I C L E S
Table 4. Crystallographic Data and Structure Refinement Information for [8,8,8-(PPh3)2H-9-(NC5H5)-nido-8,7-RhSB9H9] (2) and
[1-(PPh3)-1-(η2-C2H4)-3-(NC5H5)-closo-1,2-RhSB9H8] (3)
2
3
empirical formula
mol wt
C41H45B9NP2RhS·2CHCl3
1084.71
C25H32B9NPRh S ·CDCl3
730.12
cryst dimens (mm)
cryst syst
space group
a (Å)
0.07× 0.10× 0.16
monoclinic
P21
0.13×0.19×0.19
orthorhombic
Pbca
12.922(3)
13.0047(8)
b (Å)
c (Å)
15.436(3)
12.930(3)
19.9544(12)
24.4736(15)
ꢀ (deg)
109.481(3)
2431.7(8)
V (Å3)
6350.9(7)
Z
2
8
Dcalcd (g/cm3)
1.480
0.823
1.7-27.0
0.8795, 0.9469
15295
7878 (0.058)
6575
7878/25/587
1.04
1.527
0.928
1.7-28.0
0.8449, 0.8921
40397
7537
6324
7537/0/483
µ (mm-1
θ range (deg)
min and max trans
no. of reflns collected
no. of unique reflns (Rint
no. of reflns obsd (I > 2σ(I))
no. of data/restraints/params
GOF
)
)
1.21
R indices (I > 2σ(I))
R1 ) 0.0644 wR2 ) 0.1306
R1 ) 0.0814 wR2 ) 0.1387
0.95, -1.73
R1 ) 0.0512 wR2 ) 0.1081
R1 ) 0.0644 wR2 ) 0.1306
0.95, -0.85
R indices (all data)
largest diff peak and hole (e/ Å3)
31P{1H} NMR (162 MHz, CDCl3, 223 K): δ 36.4 [br dd, 1J(Rh,P)
) 128 Hz, 2J(P,P) ≈ 14 Hz not well resolved due to the broadness
of the peak], 30.6 [dd, 1J(Rh,P) ) 128 Hz, 2J(P,P) ) 19 Hz]. LRMS
(MALDI+): m/z 582 [M - (PPh3) - 2H]+ (isotope envelope).
Crystallographic data and structure refinement information are
gathered in Table 4.
means of a balloon attached to the Schlenk tube, then afforded
compound 3 in 4 h. Purification of the product was accomplished
by crystallization from CH2Cl2/hexane in 70 % yield.
Reaction of 2 with 1-Hexene. In a Schlenk tube, 2 mg (0.0025
mmol) of 2 was dissolved in 0.6 mL of CH2Cl2 together with a
20-fold excess of 1-hexene. The resulting solution was stirred at
room temperature under an argon atmosphere. The reaction was
monitored by GC, and after 3 days, the mixture contained 17 % of
2-hexene, 7 % of 3-hexene, and 1 % of hexane. We did not observe
formation of hexane or the associated nido to closo transformation
of 2.
Reaction of 2 with Cyclohexane. A 2 mg (0.0025 mmol)
amount of 2 was dissolved in 0.6 mL of CD2Cl2 together with a
20-fold excess of cyclohexane. No products were observed after
seven days.
Compound 2 was also prepared on a larger scale from reaction
of 1 (80 mg, 0.104 mmol) with pyridine (33 µL, 0.4173 mmol) in
dichloromethane. The reaction mixture was stirred at room tem-
perature for 2 h, and the rhodathiaborane 2 was crystallized from
CH2Cl2/hexane to give 61 mg of product (0.07 mmol, 75 %).
Reaction of [8,8,8-(PPh3)2H-9-(NC5H5)-nido-RhSB9H9] (2) with
C2H4. In a 5 mm NMR tube, 8 mg of 1, dissolved in 0.7 mL of
CD2Cl2, was treated with a four-fold excess of pyridine at room
temperature to give the hydride-containing species 2. After bubbling
ethene through the solution, the reaction mixture was studied by
NMR spectroscopy, which showed, after several hours, the quan-
titative formation of the 11-vertex closo rhodathiaborane, [1,1-
(PPh3)(η2-C2H4)-3-(NC5H5)-closo-1,2-RhSB9H8] (3). The nido to
closo oxidation is concomitant with reduction of ethene to yield
ethane (see Figures S1 and S2 in the Supporting Information). Anal.
Calcd for C25H32B9NPRhS·(CH2Cl2): C, 41.59; H, 4.65; N, 1.80;
S, 4.11. Found: C, 41.33; H, 4.35; N, 1.88; S, 4.48. IR (KBr, disk,
4000-450 cm-1): ν 2514 vs (BH), 2468 vs (BH), 2033 w, 1619 m,
1479 s, 1458s s, 1433 s, 1263 m, 1156 m, 1090 s, 1004 m, 933 m,
744 m, 693 s, 527 s, 493 m. IR (neat, 480-200 cm-1): ν 455 m
(Rh-C2H4), ν 424 m (Rh-C2H4), 329 (RhP). 1H NMR (400 MHz,
CD2Cl2, 300 K): δ 9.28 (m, 1H, NC5H5), 8.57 (m, 2H, NC5H5),
7.74 (t, 7.6 Hz, 1H, NC5H5), 7.67 (m, 1H, NC5H5), 7.79-7.02 (15H,
1P(C6H5)3), 2.27 (m, 2H, C2H4), 2.05 (m, 2H, C2H4). 31P{1H} NMR
[1,1-(PPh3)(CO)-3-(NC5H5)-closo-1,2-RhSB9H8] (4). Method
1. An orange-red solution of 3 (100 mg, 0.130 mmol) in 5 mL of
CH2Cl2 was stirred under an atmosphere of CO(g) at room
temperature for 2 days during which time the initial bright-red
solution became yellow. The final mixture was filtered through
Celite, the solvent reduced in volume, and hexane added leading
to precipitation of a yellow product, which was washed repeatedly
with hexane to give 63 mg of 4 (0.104 mmol, 80 %). IR (KBr,
disk, 4000-450 cm-1): ν 2524 s (BH), 2474 vs (BH), 1980 vs
(CO), 1619 m, 1480 m, 1456 m, 1433 m, 1167 m, 1095 m, 1005 s,
935 m, 745 m, 692 s, 526 s. 1H NMR (400 MHz, CD2Cl2, 300 K):
δ 9.12 (d, 6.0 Hz, 2H; NC5H5), 8.20 (t, 7.7 Hz, 1H; NC5H5), 7.64
(t, 7.7 Hz,2H; NC5H5), 7.40-7.26 (15H; P(C6H5)3). 31P{1H} NMR
1
(121 MHz, CD2Cl2, 300 K): 37.9 [d, J(Rh,P) ) 132 Hz; 1PPh3].
Method 2. A 25 mg (0.029 mmol) amount of 2 was dissolved
in 8 mL of CH2Cl2; then, after three freeze-pump-thaw cycles,
the frozen-evacuated solution was exposed to an atmosphere of
CO(g) and stirred at room temperature for 18 h. The final orange
solution was reduced in volume, and hexane was added, precipitat-
ing an orange-yellow product, which was washed repeatedly with
hexane to give 15 mg (0.025 mmol; 85 %) of 4.
1
(121 MHz, CD2Cl2, 300 K): 39.7 [d, J(Rh,P) ) 139 Hz, 1PPh3].
13C{1H} NMR (75 MHz, CDCl3, 300 K): δ 48.9, br s, 2C,
H2CdCH2 (correlates with the two broad multiples at δΗ 2.27 and
2.05). 13C{1H} NMR (100 MHz, CD2Cl2, 193 K): δ 49.3, br d,
1J(Rh,C) ) 55 Hz; 46.2, br d, 1J(Rh,C) ) 40 Hz. LRMS
(MALDI+): m/z 582 [M - C2H4]+ (isotope envelope). Crystal-
lographic data and structure refinement information are gathered
in Table 4.
[1,1-(PPh3)2-3-(NC5H5)-closo-1,2-RhSB9H8] (5). A solution of
2 (30 mg, 0.035 mmol) in CH2Cl2 was heated at reflux in an argon
atmosphere for 48 h. The final red solution was evaporated under
vacuum to give a red residue, which was subsequently crystallized
from dichloromethane/hexane to give 22 mg of product (0.026mmol;
74%). IR (KBr, disk, 4000-450 cm-1): ν 2506 vs (BH), 1261 s
(BH), 1093 s, 1022 s, 799 s, 694 m. 1H NMR (400 MHz, CD2Cl2,
Compound 3 may be prepared on a larger scale. For example,
45 mg of the ethene-ligated closo-rhodathiaborane was prepared
by stirring of 80 mg (0.104 mmol) of 1 with four-fold excess of
pyridine in 5 mL of CH2Cl2 for 2 h to form the hydridorhodathi-
aborane 2. Exposure of the solution to an ethene atmosphere, by
9
J. AM. CHEM. SOC. VOL. 130, NO. 34, 2008 11465

´
Alvarez et al.
A R T I C L E S
300 K): δ 8.90 (d, 3J(H,H) ) 5.4 Hz, 2H; C5H5N), 8.26 (t, 3J(H,H)
confirm the true minima, together with GIAO NMR nuclear-
shielding predictions, were performed using B3LYP methodology
with the 6-31G* and LANL2DZ basis sets. GIAO NMR nuclear
shielding predictions were performed on the final optimized
geometries, and computed 11B shielding values were related to
chemical shifts by comparison with the calculated value for B2H6,
which was taken to be δ(11B) +16.6 ppm relative to the F3B·OEt2
) 0.0 ppm standard. In each case the compounds were modeled
using hydrogen atoms rather than phenyl groups on the phosphine
ligands in order to reduce computation time.
3
) 6.6 Hz, 1H; C5H5N), 7.66 (t, J(H,H) ) 6.6 Hz, 2H, C5H5N),
7.32-6.99 (30H, 2P(C6H5)3). 31P{1H} NMR (121 MHz, CD2Cl2,
300 K): 39.6 [d, 1J(103Rh-31P) ) 147 Hz, 2PPh3]. LRMS
(MALDI+): m/z 571 [M - BH - PPh3]+ (isotope envelope).
Reaction of 3-5 with H2(g). In a typical reaction, around 10
mg (0.016 mmol) of the closo-rhodathiaborane was placed in a
screw-capped NMR tube and dissolved in 0.6 mL of CD2Cl2. The
NMR tube was cooled in liquid nitrogen, evacuated, and then filled
with dihydrogen. The reaction was monitored at different times by
1H NMR spectroscopy, and conversion was followed by integration
of particular peaks in the 1H{11B} spectrum. The integrals for each
experiment are listed in the Supporting Information. In addition,
reaction of the C2H4-ligated closo-rhodathiaborane 3 with dihy-
drogen was studied under the following conditions: (i) In CD2Cl2
solvent, (iii) in CD2Cl2 solvent with 1 equiv of free PPh3, and (iv)
in (CD3)2CO.
Reaction of 3 with D2(g). As in the reaction with H2(g) above.
Catalytic Hydrogenation. To test the catalytic activity of 2 for
the hydrogenation of olefins, two reactions between nido-hydri-
dorhodathiaborane 2 and cyclohexene or 1-hexene were conducted
in screw-capped NMR tubes: (i) 5 µL (0.050 mmol) of cyclohexane
and 6 µL mL of iso-octane were added to 2 (2 mg, 0.003 mmol)
dissolved in 0.8 mL of CH2Cl2; (ii) 5 µL (0.040 mmol) of 1-hexene
and 6 µL of iso-octane were added to a solution of 2 (1.7 mg,
0.002mmol)in0.8mLofCHCl3.Eachsamplewasfreeze-pump-thawed
and then exposed, while still immersed in the liquid nitrogen, to
an atmosphere of dihydrogen (rubber-gas balloon) for 1 min. The
NMR tubes were then immersed in an oil bath at 60 °C, and the
reactions were monitored by GC using the iso-octane peak as an
internal reference.
Acknowledgment. This work was supported by the Ministerio
de Educacio´n y Ciencia (MEC, Spain) (Factor´ıa de Cristalizacio´n,
CONSOLIDER INGENIO-2010 and grant CTQ2006-01629). R.M.
thanks the MEC-Universidad de Zaragoza and the European Social
Fund for his Research Contract in the framework of the “Ramo´n y
Cajal” Program. A.A. thanks the IUCH for a fellowship. J.B. thanks
professors J.D. Kennedy and M.J. Pilling for the use of facilities.
J.B. was supported in part by the Czech Academy of Sciences Grant
No. IAA400320601 and the Grant Agency of the Czech Republic
Project No. 203 3982.
Supporting Information Available: Complete ref 52; figures
for 1H and 31P spectra of 2 and 3; attached proton test spectrum
of 2; conversions of 2, 3, and 5 in their reaction with H2; GC
data; computational details, coordinates of optimized geometries,
and calculated energies for all themodel compounds. This
material is available free of charge via the Internet at http://
pubs.acs.org.
Calculations. All calculations were performed using the Gaussian
98 and Gaussian 03 packages.52 Structures were initially optimized
using standard ab initio methods with the STO-3G* basis sets for
B, C, P, S, N, and H with the LANL2DZ basis set for the metal
atoms. The final optimizations, including frequency analyses to
JA802993M
(52) Pople, J. A., et al. Gaussian 03, revision B.05; Gaussian Inc.:
Wallingford, CT, 2004.
9
11466 J. AM. CHEM. SOC. VOL. 130, NO. 34, 2008