Transition metal-catalyzed dissociation of phosphine-gallane adducts: Isolation of mechanistic model complexes and heterogeneous catalyst poisoning studies

Article abstract of DOI:10.1021/ic700573j

Attempts to induce the catalytic dehydrocoupling of the phosphine-gallane adduct Cy₂PH·GaH₃ (Cy = cyclohexyl) (1) by treatment with ca. 5 mol % of either the Rh(I) complex [{Rh(μ-Cl)(1,5-cod)} ₂] (cod = cyclooctadiene) or the Rh(0) species Rh/Al ₂O₃ and [Oct₄N]Cl-stabilized colloidal Rh led to catalytic P-Ga bond cleavage to generate the phosphine, H₂, and Ga metal. Interestingly, subsequent treatment of the reaction mixtures with Me₂NH·BH₃ failed to lead to the formation of [Me₂N-BH₂]₂ via Rh-catalyzed dehydrocoupling, which suggested that catalyst deactivation was taking place. Poisoning studies involving the treatment of the active Rh(0) catalyst with Cy₂PH, PMe₃, or GaH₃·OEt₂ showed that deactivation indeed occurred as the dehydrocoupling of Me₂NH· BH₃ either dramatically decreased in rate or did not take place at all. The X-ray photoelectron spectroscopy analysis of colloidal Rh(0) that had been treated with Cy₂PH and PMe₃ confirmed the presence of phosphorus on the catalyst surface in each case, consistent with catalyst poisoning via phosphine ligation. A mechanism for the Rh-catalyzed P-Ga bond cleavage reaction of 1 and Me₃P·GaH₃ (2) is proposed and involves the initial reaction of Ga-H bonds with the Rh colloid surface, which weakens and ultimately breaks the P-Ga bond. The reasonable nature of this mechanism is supported by a model reaction between the zerovalent group 9 complex Co₂(CO)₈ and 2 which afforded Me ₃P·Ga[Co(CO)₄]₃ (3). Consistent with the elongated and thus weakened P-Ga bond in 3, solutions of this species in Et₂O subsequently form the known complex [(Me₃P)Co(CO) ₃]₂ (4) and Ga metal after 4 h at 25°C.

Full text of DOI:10.1021/ic700573j

Inorg. Chem. 2007, 46, 7394−7402
Transition Metal-Catalyzed Dissociation of Phosphine
−Gallane Adducts:
Isolation of Mechanistic Model Complexes and Heterogeneous Catalyst
Poisoning Studies
Timothy J. Clark,† Cory A. Jaska,† Ayse Turak,‡ Alan J. Lough,† Zheng-Hong Lu,‡ and Ian Manners*,†,§
Department of Chemistry, UniVersity of Toronto, 80 St. George Street, Toronto, Ontario M5S 3H6,
Canada, Department of Materials Science and Engineering, UniVersity of Toronto, 184 College
Street, Toronto, Ontario M5S 3E4, Canada, and School of Chemistry, UniVersity of Bristol,
Bristol BS8 1TS, England
Received March 26, 2007
Attempts to induce the catalytic dehydrocoupling of the phosphine
(1) by treatment with ca. 5 mol % of either the Rh(I) complex [ Rh(
Rh(0) species Rh/Al2O3 and [Oct4N]Cl-stabilized colloidal Rh led to catalytic P
phosphine, H2, and Ga metal. Interestingly, subsequent treatment of the reaction mixtures with Me2NH
to lead to the formation of [Me2N BH2]2 via Rh-catalyzed dehydrocoupling, which suggested that catalyst deactivation
was taking place. Poisoning studies involving the treatment of the active Rh(0) catalyst with Cy2PH, PMe3, or
GaH3 OEt2 showed that deactivation indeed occurred as the dehydrocoupling of Me2NH BH3 either dramatically
−
µ
gallane adduct Cy2PH
-Cl)(1,5-cod) 2] (cod
Ga bond cleavage to generate the
BH3 failed
‚
GaH3 (Cy
) cyclohexyl)
{
}
) cyclooctadiene) or the
−
‚
−
‚
‚
decreased in rate or did not take place at all. The X-ray photoelectron spectroscopy analysis of colloidal Rh(0) that
had been treated with Cy2PH and PMe3 confirmed the presence of phosphorus on the catalyst surface in each
case, consistent with catalyst poisoning via phosphine ligation. A mechanism for the Rh-catalyzed P
cleavage reaction of 1 and Me3P GaH3 (2) is proposed and involves the initial reaction of Ga H bonds with the
Rh colloid surface, which weakens and ultimately breaks the P Ga bond. The reasonable nature of this mechanism
is supported by a model reaction between the zerovalent group 9 complex Co2(CO)8 and 2 which afforded Me3P
Ga[Co(CO)4]3 (3). Consistent with the elongated and thus weakened P Ga bond in 3, solutions of this species in
Et2O subsequently form the known complex [(Me3P)Co(CO)3]2 (4) and Ga metal after 4 h at 25 C.
−Ga bond
‚
−
−
‚
−
°
Introduction
ization catalyst for boron hydrides and carboranes by
Sneddon and co-workers1 and of Ti-catalyzed dehydropo-
lymerization of primary silanes by Harrod and co-workers2
launched the development of catalytic routes to rings, chains,
and polymers based on p-block elements.3
Although metal-catalyzed bond formation and cleavage
reactions have played a critical and widespread role in
modern synthetic organic chemistry, the use of transition
metal catalysts to form or cleave hetero- and homonuclear
bonds between main-group elements is still in its relative
infancy. Traditionally, synthetic methods in main-group
chemistry often utilize metathesis reactions such as salt
elimination to generate new bonds. However, the discoveries
in the early 1980s that PtBr2 acts as a general dehydrodimer-
Our research group has reported late transition metal-
catalyzed dehydrocoupling reactions as an alternative route
(1) Corcoran, E. W.; Sneddon, L. G. J. Am. Chem. Soc. 1984, 106, 7793.
(2) Aitken, C.; Harrod, J. F.; Samuel, E. J. Organomet. Chem. 1985, 279,
C11.
(3) For a recent review on metal-catalyzed dehydrocoupling, see: Clark,
T. J.; Lee, K.; Manners, I. Chem.sEur. J. 2006, 12, 8634.
(4) (a) Dorn, H.; Singh, R. A.; Massey, J. A.; Lough, A. J.; Manners, I.
Angew. Chem., Int. Ed. 1999, 38, 3321. (b) Dorn, H.; Singh, R. A.;
Massey, J. A.; Nelson, J. M.; Jaska, C. A.; Lough, A. J.; Manners, I.
J. Am. Chem. Soc. 2000, 122, 6669. (c) Dorn, H.; Rodezno, J. M.;
Brunnho¨fer, B.; Rivard, E.; Massey, J. A.; Manners, I. Macromolecules
2003, 36, 291.
* To whom correspondence should be addressed. E-mail: Ian.Manners@
bristol.ac.uk.
† Department of Chemistry, University of Toronto.
‡ Department of Materials Science and Engineering, University of
Toronto.
§ School of Chemistry, University of Bristol.
7394 Inorganic Chemistry, Vol. 46, No. 18, 2007
10.1021/ic700573j CCC: $37.00
© 2007 American Chemical Society
Published on Web 07/31/2007

Catalyzed Dissociation of Phosphine-Gallane Adducts
Scheme 1
to generating P-B4 bonds from phosphine-borane adducts,
which has allowed routes to novel rings, chains, and polymers
[RR′P-BH2]n (eqs 1 and 2 in Scheme 1).5 Such reactions
proceed at particularly mild temperatures for fluoroaryl
phosphine-borane derivatives, and preliminary investigations
have shown that the resultant polyphosphinoboranes function
as a negative-tone resist in electron-beam lithography.6 In
addition, we have reported that the dehydrocoupling of
primary and secondary amine-borane adducts and H3N‚BH3
is promoted by Rh (pre)catalysts affording oligomers and/
or insoluble, presumably cross-linked materials [HxN-BHy]n
(eq 3 in Scheme 1).7 Transition metal-catalyzed dehydro-
coupling also has potential utility in hydrogen storage
applications, which are desirable, as hydrogen is a clean
energy source that can potentially replace the petroleum-
derived products that currently fuel vehicles.8 For example,
there is considerable current interest in amine-borane
adducts such as H3N‚BH3, which possesses one of the highest
densities of hydrogen available (19.6 wt % H2), and there
have been numerous recent reports related to the thermal or
catalytic dehydrogenation of H3N‚BH3 or other high H2
content amine-borane adducts.9,10 In addition, the catalytic
dehydrocoupling of Me2NH‚BH3 to afford the cyclic dimer
[Me2N-BH2]2 has been employed as a means of transfer
hydrogenation when carried out in a closed system in the
presence of unsaturated organic species.11
[{Rh(1,5-cod)(µ-Cl)}2] (cod ) cyclooctadiene) were found
to operate under different mechanisms where the former is
a homogeneous process and the latter is a heterogeneous
process involving colloidal Rh metal.12,13 We have also
detailed that heterogeneous Rh(0) catalysts, when treated with
a variety of group 13 hydrides, generate H2 gas and a
passivating layer incorporating the group 13 element on the
surface of the metal, thereby deactivating the catalyst toward
the dehydrocoupling of amine-borane adducts.14 Recently,
we have demonstrated the use of the early transition-metal
complex [Cp2Ti] for amine-borane adduct dehydrocoupling
and found that it exhibits increased catalytic activity relative
to Rh catalysts.15 Notably, this catalyst appears to operate
in a homogeneous fashion in contrast to the heterogeneous
Rh-based system. Goldberg, Heinekey, and co-workers have
described a highly active, homogeneous Ir catalyst that
dehydrocouples H3N‚BH3 at room temperature to afford the
(9) (a) Gutowska, A.; Li, L.; Shin, Y.; Wang, C. M.; Li, X. S.; Linehan,
J. C.; Smith, R. S.; Kay, B. D.; Schmid, B.; Shaw, W.; Gutowski, M.;
Autrey, T. Angew. Chem., Int. Ed. 2005, 44, 3578. (b) Chandra, M.;
Xu, Q. J. Power Sources 2006, 156, 190. (c) Bluhm, M. E.; Bradley,
M. G.; Butterick, R., III; Kusari, U.; Sneddon, L. G. J. Am. Chem.
Soc. 2006, 128, 7748. (d) Denney, M. C.; Pons, V.; Hebden, T. J.;
Heinekey, D. M.; Goldberg, K. I. J. Am. Chem. Soc. 2006, 128, 12048.
(e) Yoon, C. W.; Sneddon, L. G. J. Am. Chem. Soc. 2006, 128, 13992.
(f) Stephens, F. H.; Baker, R. T.; Matus, M. H.; Grant, D. J.; Dixon,
D. A. Angew. Chem., Int. Ed. 2007, 46, 746. (g) Nguyen, M. T.;
Nguyen, V. S.; Matus, M. H.; Gopakumar, G.; Dixon, D. A. J. Phys.
Chem. A 2007, 111, 679. (h) Cheng, F.; Ma, H.; Li, Y.; Chen, J. Inorg.
Chem. 2007, 46, 788. (i) Keaton, R. J.; Blacquiere, J. M.; Baker, R.
T. J. Am. Chem. Soc. 2007, 129, 1844.
(10) It has proven challenging to regenerate H3N‚BH3 from the aminobo-
rane-based products, rendering the system reversible. In this context,
Stephan and co-workers have recently reported a remarkable phos-
phonium-borate species that demonstrates reversible H2 activation
in the absence of a metal, see: Welch, G. C.; San Juan, R. R.; Masuda,
J. D.; Stephan, D. W. Science, 2006, 314, 1124.
Interestingly, catalytic dehydrocoupling reactions of phos-
phine- and amine-borane adducts using the (pre)catalyst
(5) For a report on the B(C6F5)3-catalyzed formation of P-B bonds via
dehydrocoupling, see: Denis, J.-M.; Forintos, H.; Szelke, H.; Toupet,
L.; Pham, T.-N.; Madec, P.-J.; Gaumont, A.-C. Chem. Commun. 2003,
54.
(6) Clark, T. J.; Rodezno, J. M.; Clendenning, S. B.; Aouba, S.; Brodersen,
P. M; Lough, A. J.; Ruda, H. E.; Manners, I. Chem.sEur. J. 2005,
11, 4526.
(7) (a) Jaska, C. A.; Temple, K.; Lough, A. J.; Manners, I. Chem. Commun.
2001, 962. (b) Jaska, C. A.; Temple, K.; Lough, A. J.; Manners, I. J.
Am. Chem. Soc. 2003, 125, 9424.
(8) (a) Basic Research Needs for the Hydrogen Economy; Dresselhaus,
M., Crabtree, G., Buchanan M., Eds.; Basic Energy Sciences, Office
of Science, U.S. Department of Energy: Washington, DC, 2003. (b)
The Hydrogen Economy: Opportunities, Costs, Barriers, and R&D
Needs; National Academies Press: Washington, DC, 2004.
(11) Jaska, C. A.; Manners, I. J. Am. Chem. Soc. 2004, 126, 2698.
(12) Jaska, C. A.; Manners, I. J. Am. Chem. Soc. 2004, 126, 9776.
(13) Autrey and co-workers have studied the dehydrocoupling of Me2NH‚
BH3 using Rh catalysts and reaction conditions that differ from those
used in our group. On the basis of their observations, they propose
that the process may be homogeneous: Chen, Y.; Fulton, J. L.;
Linehan, J. C.; Autrey, T. J. Am. Chem. Soc. 2005, 127, 3254.
(14) Jaska, C. A.; Clark, T. J.; Clendenning, S. B.; Grozea, D.; Turak, A.;
Lu, Z.-H.; Manners, I. J. Am. Chem. Soc. 2005, 127, 5116.
(15) Clark, T. J.; Russell, C. A.; Manners, I. J. Am. Chem. Soc. 2006, 128,
9582.
Inorganic Chemistry, Vol. 46, No. 18, 2007 7395

Clark et al.
Table 1. Crystallographic Data and Summary of Data Collection and
Refinement for Structure 3
cyclic pentamer [H2N-BH2]5, as characterized by powder
X-ray diffraction and IR spectroscopy.9d Similarly, Baker and
co-workers have also shown that a Ni-based catalyst
promotes the dehydrocoupling of H3N‚BH3, yielding a
soluble, cross-linked borazine structure in 4 h at 60 °C.9i
As part of our continuing research program to develop
novel extended structures based on main-group elements, we
have investigated the reactivity of other Lewis acid-base
adducts toward the dehydrocoupling process. In this paper,
we report on our attempts to extend this new synthetic
method to some analogous adducts containing heavier group
13 elements, namely, phosphine-gallanes.16
emp formula
fw
T (K)
λ (Å)
cryst syst
space group
C15H9Co3GaO12P
658.70
150(2)
0.71073
triclinic
P1h
Z
D
2
C (g/cm3)
1.928
3.456
644
2.71-27.58
-11 e h e 11
-13 e k e 13
-17 e l e 17
14 601
5181
0.0886
1.040
0.0458
µ (mm-1
F(000)
)
q range (deg)
index ranges
cryst size (mm) 0.60 × 0.10 × 0.04
a (Å)
8.7826(4)
10.0257(4)
13.4787(5)
84.370(3)
85.783(3)
74.101(2)
1134.57(8)
b (Å)
c (Å)
R (deg)
â (deg)
γ (deg)
V (Å3)
reflns collected
ind reflns
Rint
GOF on F2
R1a (I > 2σ(I))
wR2b (all data)
0.1266
peak/hole (e‚Å3) 0.673/ -1.247
Experimental Section
2
2
2
a R1 ) S||Fo| - |Fc||/S|Fo|. b wR2 ) {S[w(Fo - Fc )2]/S[w(Fo )2]}1/2
.
General Methods and Materials. All reactions and product
manipulations were performed under an atmosphere of dry nitrogen
using standard Schlenk techniques or in a Mbraun glovebox filled
with dry nitrogen. Tetrahydrofuran (THF) and diethyl ether were
dried over Na/benzophenone and distilled prior to use. Toluene and
hexanes were purified using the Grubbs method.17 Ga metal
(99.9995%), PMe3, Rh/Al2O3 (5 wt. % Rh), and HCl (2.0 M solution
in Et2O) were purchased from Aldrich and used as received. Cy2-
PH was acquired through a generous donation from Cytec. Me2-
NH‚BH3 and [Co2(CO)8] were purchased from Strem. Me2NH‚BH3
was further purified by sublimation at 25 °C whereas [Co2(CO)8]
was sublimed immediately before use. [{Rh(µ-Cl)(1,5-cod)}2],18
Rhcolloid/[Oct4N]Cl,19 Me3P‚GaH3 (2),20 and Li[GaH4]21 were syn-
thesized by literature procedures.
Equipment. NMR spectra were recorded on either Varian
Gemini 300 MHz or Varian VRX-S (Unity) 400 MHz spectrom-
eters. Chemical shifts are reported relative to residual solvent peaks
(1H, 13C) or external H3PO4 (31P) or BF3‚OEt2 (11B) references.
The NMR spectra were obtained at 300 MHz (1H), 100 MHz (13C),
121 MHz (31P), and 96 MHz (11B). The mass spectra were obtained
with a VG 70-250S mass spectrometer operating in electron impact
(EI) mode. Elemental analyses were obtained on a Perkin-Elmer
Series 2400 CHNS analyzer maintained by the Analest facility at
the University of Toronto. Infrared spectra were obtained on a
Perkin-Elmer Spectrum One FT-IR spectrometer using KBr win-
dows. The silicon (100) substrates were purchased from Wafer
World, Inc., and cleaned by successive sonication treatments in
CH2Cl2, isopropanol, piranha solution (H2O2/H2SO4, 1:3 vol %;
Caution: piranha solution is extremely corrosiVe!), and deionized
water followed by drying in air. X-ray photoelectron spectroscopy
(XPS) samples were solution deposited onto a clean silicon substrate
and analyzed on a PHI 500 ESCA system. A monochromated Al
KR source with a photon energy of 1486.6 eV was used and
recorded at a photoelectron takeoff angle of 45°. The peak positions
were aligned and corrected for charge effects based on an
adventitious C(1s) peak position at 284.8 eV. For the depth-profiling
analysis, an Ar+ ion beam of 3 keV was used for 15 min and the
spectra were reacquired.
X-Ray Structural Determination. Diffraction data were col-
lected on a Nonius Kappa-CCD with graphite-monochromated Mo
KR radiation (λ ) 0.71073 Å). The data were integrated and scaled
with the Denzo-SMN package.22 The structures were solved and
refined with the SHELXTL-PC version 5.1 software package.23
Refinement was by full-matrix least squares on F2 of all data
(negative intensities included). All molecular structures are pre-
sented with thermal ellipsoids at a 30% probability level. In all
structures, hydrogens bonded to carbon atoms were included in
calculated positions and treated as riding atoms. Crystallographic
data and the summary of data collection and refinement for structure
3 are presented in Table 1. CCDC-249316 (3) contains the
supplementary crystallographic data for this paper. This data can
be obtained free of charge from The Cambridge Crystallographic
Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
Synthesis of Cy2PH‚GaH3 (1). [Cy2PH2]Cl was prepared by
adding HCl (4.03 mL of a 2.0 M solution in diethyl ether, 8.06
mmol) to a solution of Cy2PH (1.63 mL, 8.06 mmol) in 5 mL Et2O
at 25 °C causing a white solid to precipitate immediately. A solution
of Li[GaH4] in Et2O (0.65 g, 8.05 mmol) was added to [Cy2PH2]-
Cl at -78 °C and stirred for 3 h at 20 °C. The solution was filtered
to separate LiCl, and the solvent was removed in vacuo affording
1 as a white solid in 79% yield (1.72 g, 6.35 mmol). In some cases,
a white viscous residue was obtained upon solvent removal.
Subsequent recrystallization of this oil from hexane at -35 °C also
1
afforded 1 as a white solid. H NMR (C6D6): δ 4.36 (s, GaH3),
3.37 (d, PH, 1JP-H ) 125.8 Hz), 1.70 (m, i-CH), 1.47 (m, â-CH2),
1.21 (m, γ-CH2), 0.95 (m, δ-CH2). 13C{1H} NMR (C6D6): δ 31.2
(s), 30.1 (d, JCP ) 6.8 Hz), 29.1 (d, JCP ) 22 Hz), 27.1 (d, JCP
)
)
1
27.3 Hz), 26.3 (s). 31P NMR (C6D6): δ -15.7 (d, PH, JP-H
308.5 Hz). EI-MS (70 eV): m/z 198 (M+ - GaH3). Elemental Anal.
Calcd for C12H26PGa (271.04): C, 53.18; H, 9.67. Found: C, 53.40;
H, 9.45.
Reaction of 1 with [{Rh(µ-Cl)(1,5-cod)}2]. A catalytic amount
of [{Rh(µ-Cl)(1,5-cod)}2] (ca. 5 mol % Rh) was added to a solution
of 1 (ca. 0.125 g) in toluene (ca. 2 mL), causing the reaction mixture
to immediately turn black with vigorous gas evolution. After 16 h
of stirring at 25 °C, an aliquot of the reaction mixture was removed
and the 31P NMR spectrum showed a single resonance at δ -28.1
ppm corresponding to free Cy2PH.
Reaction of 1 with Rh/Al2O3 or Rhcolloid/[Oct4N]Cl. In a typical
experiment, a catalytic amount of Rh/Al2O3 (ca. 5 mol % Rh) was
added to a solution of 1 (ca. 0.125 g) in toluene (ca. 2 mL), resulting
(16) The thermal dehydrocoupling of P-H and Al/Ga-H bonds has been
used to prepare donor-stabilized phosphanyl-alane and phosphanyl-
gallane species: Vogel, U.; Timoshkin, A. Y.; Scheer, M. Angew.
Chem., Int. Ed. 2001, 40, 4409.
(17) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.;
Timmers, F. J. Organometallics 1996, 15, 1518.
(18) Giordano, G.; Crabtree, R. H. Inorg. Synth. 1979, 19, 218.
(19) Bo¨nnemann, H.; Brijoux, W.; Brinkmann, R.; Dinjus, E.; Jouben, T.;
Korall, B.; Dinjus, E. Angew. Chem., Int. Ed. Engl. 1991, 30, 1312.
(20) Greenwood, N. N.; Ross, E. J. F.; Storr, A. J. Chem. Soc. 1965, 1400.
(21) Shirk, A. E.; Shriver, D. F. Inorg. Synth. 1977, 17, 45.
(22) Otwinowski, Z.; Minor, W. Methods Enzymol. 1997, 276, 307.
(23) Sheldrick, G. M. SHELXTL-PC, version 5.1; Bruker Analytical X-ray
Systems, Inc.: Madison, WI, 1997.
7396 Inorganic Chemistry, Vol. 46, No. 18, 2007

Catalyzed Dissociation of Phosphine-Gallane Adducts
in vigorous gas evolution. After 16 h of stirring at 25 °C, an aliquot
of the reaction mixture was removed and the 31P NMR spectrum
showed a single resonance at δ -28.1 corresponding to free Cy2-
PH. Similar results were obtained with Rhcolloid/[Oct4N]Cl.
Treatment of 1 with Ga Metal. To a solution of 1 (0.150 g,
0.55 mmol) in toluene (ca. 1 mL) a catalytic amount of Ga
metal (ca. 5 mol % Ga) was added, giving a slightly gray solution,
which was stirred at 25 °C. After 9 h, an aliquot of the reaction
mixture was removed and the 31P NMR spectrum showed only
unreacted 1.
Attempted Catalytic Dehydrocoupling of Me2NH‚BH3 in the
Presence of 1. A solution of 1 (0.217 g, 0.801 mmol) and Me2-
NH‚BH3 (0.047 g, 0.798 mmol) in toluene (ca. 1 mL) was stirred
with no color change or gas evolution observed for 3 h at 25 °C.
An aliquot of the reaction mixture was removed, and both 31P and
11B NMR spectra indicated no dehydrocoupling activity, as only
the starting materials were observed. To the above solution a
catalytic amount of [{Rh(µ-Cl)(1,5-cod)}2] (ca. 5 mol %) was
added, causing the reaction mixture to immediately turn black with
vigorous gas evolution. After 5 h of stirring at 25 °C, an aliquot of
the reaction mixture was removed and the 31P NMR spectrum
showed a single resonance at δ -28.1 corresponding to free Cy2-
PH whereas the 11B NMR spectrum showed a single resonance at
δ -13.9 corresponding to unreacted Me2NH‚BH3.
consumption of 1 and new resonances at Pδ 19 and Bδ -45
corresponding to Cy2PH‚BH3.24
Testing for Lewis Acid Exchange between 2 and BH3‚THF.
To a solution of 2 (0.080 g, 0.538 mmol) in toluene (ca. 2 mL)
BH3•THF (1.10 mL, 1.10 mmol) was added and stirred at 25 °C.
After 75 min, an aliquot of the reaction mixture was removed and
the 11B and 31P NMR spectra were obtained, showing complete
consumption of 2 and new resonances at Pδ -1 and Bδ -37
corresponding to Me3P‚BH3.25
Treatment of 2 with [{Rh(µ-Cl)(1,5-cod)}2]. To a solution of
2 (0.088 g, 0.591 mmol) in toluene (ca. 2 mL) a catalytic amount
of [{Rh(µ-Cl)(1,5-cod)}2] (ca. 2 mol %) was added, giving a red
solution which was stirred at 25 °C. After 16 h, the 31P NMR
spectrum showed a single resonance at δ -39.1 corresponding to
unreacted 2.
Treatment of 2 with a Stoichiometric Amount of [{Rh(µ-Cl)-
(1,5-cod)}2]. To a solution of 2 (0.100 g, 0.672 mmol) in toluene
(2 mL) 1 equiv of [{Rh(µ-Cl)(1,5-cod)}2] (0.166 g, 0.337 mmol)
was added, resulting in vigorous gas evolution and the formation
of an amber solution and black precipitate. After 16 h, the 31P and
1H NMR spectra were obtained, showing complete consumption
of 2 and resonances corresponding to [(Me3P)RhCl(1,5-cod)].26
Treatment of 2 with Substoichiometric Amounts of [{Rh(µ-
Cl)(1,5-cod)}2] and Subsequent Testing of Catalytic Dehydro-
coupling Activity Toward Me2NH•BH3. To a solution of 2 (0.100
g, 0.672 mmol) and Me2NH‚BH3 (0.040 g, 0.679 mmol) in toluene
(ca. 2 mL) a substoichiometric amount of [{Rh(µ-Cl)(1,5-cod)}2]
(ca. 50 mol %) was added, resulting in vigorous gas evolution and
the formation of a dark brown solution and black precipitate. After
16 h, an aliquot of the reaction mixture was removed and the 11B
and 31P NMR spectra were obtained. The % conversion was
calculated from the integration of the product (Bδ 5 ppm) and
reactant (Bδ -13 ppm) resonances whereas the 31P NMR spectrum
revealed several resonances at δ -8 ([(Me3P)RhCl(1,5-cod)]),26 -18
(d; J ) 28 Hz), -22 (d; J ) 97 Hz), and -29 (d; J ) 176 Hz)
(unidentified).
Treatment of [{Rh(µ-Cl)(1,5-cod)}2] with 1 equiv of Cy2PH
and Testing for Catalytic Activity. To a solution of Me2NH‚BH3
(0.240 g, 4.07 mmol) in toluene (ca. 1 mL) a catalytic amount of
[{Rh(µ-Cl)(1,5-cod)}2] (ca. 2 mol %) and Cy2PH (ca. 2 mol %)
was added and stirred at 25 °C, giving a red-brown solution over
time. After 16 h, an aliquot of the reaction mixture was removed
and the 11B NMR spectra were obtained. The % conversion was
calculated from the integration of the product (Bδ 5 ppm) and
reactant (Bδ -13 ppm) resonances.
Treatment of [{Rh(µ-Cl)(1,5-cod)}2] with an Excess of Cy2PH
and Testing for Catalytic Activity. To a solution of Me2NH‚BH3
(0.100 g, 1.69 mmol) in toluene (ca. 1 mL) a catalytic amount of
[{Rh(µ-Cl)(1,5-cod)}2] (ca. 3 mol %) and Cy2PH (ca. 10 mol %)
was added and stirred at 25 °C, giving an orange solution over
time. After 16 h, an aliquot of the reaction mixture was removed
and the 11B and 31P NMR spectra obtained. The % conversion was
calculated from the integration of the product (Bδ 5 ppm) and
reactant (Bδ -13 ppm) resonances whereas the 31P NMR spectrum
revealed a resonance at δ -28.1 corresponding to free Cy2PH in
Similar results were obtained for the reaction employing 10 mol
% [{Rh(µ-Cl)(1,5-cod)}2].
Treatment of 2 with [{Rh(µ-Cl)(1,5-cod)}2] and Subsequent
Testing of Catalytic Dehydrocoupling Activity. To a solution of
2 (0.088 g, 0.591 mmol) in toluene (ca. 1 mL) a catalytic amount
of [{Rh(µ-Cl)(1,5-cod)}2] (ca. 2 mol %) was added, giving a red
solution which was stirred at 25 °C. After 5 h, the 31P NMR
spectrum showed a single resonance at δ -39.1 corresponding to
unreacted 2. To this solution, Me2NH‚BH3 (0.035 g, 0.594 mmol)
was added and stirred at 25 °C. After 16 h, an aliquot of the reaction
mixture was removed and the 31P NMR spectrum showed unreacted
2 whereas the 11B NMR spectrum showed a single resonance at δ
-13.9 corresponding to unreacted Me2NH‚BH3.
addition to two multiplets at Pδ 33 (dd; 1JP-Rh ) 131 Hz, 2JP-P
40 Hz) and 50 (dt; JP-Rh ) 174 Hz, JP-P ) 40 Hz).
)
1
2
Treatment of Rhcolloid/[Oct4N]Cl with Cy2PH and Subsequent
Testing for Catalytic Activity. Cy2PH (0.013 g, 0.066 mmol) was
added to a solution of Rhcolloid/[Oct4N]Cl (0.042 g, 0.135 mmol) in
THF (ca. 1 mL) and stirred at 25 °C. After 16 h, all volatiles were
removed and the crude reaction mixture was washed with 2-3 mL
portions of hexanes. The treated Rh(0) colloids were then dried
overnight. To a solution of Me2NH‚BH3 (0.200 g, 3.39 mmol) in
toluene (ca. 1 mL) a catalytic amount of the treated Rhcolloid/[Oct4N]-
Cl (0.006 g, 5 mol % Rh) was added at 25 °C. After 16 h of stirring,
an aliquot of the reaction mixture was removed and the 11B NMR
spectrum was obtained. The % conversion was calculated from the
integration of the product (Bδ 5 ppm) and reactant (Bδ -13 ppm)
resonances.
Reaction of 2 with Rh/Al2O3 or Rhcolloid/[Oct4N]Cl. To a
solution of 2 (0.088 g, 0.591 mmol) in toluene (ca. 2 mL) a catalytic
amount of Rh/Al2O3 (ca. 2 mol %) was added, resulting in vigorous
gas evolution and giving a black solution, which was stirred at
(24) The 31P NMR spectroscopic characterization of Cy2PH‚BH3 in CDCl3
(Pδ 19 ppm) has been previously reported, see: Stankevic, M.;
Pietrusiewicz, K. M.; Synthesis 2005, 8, 1279. We independently
carried out the synthesis of the compound by treatment of Cy2PH with
BH3‚THF in THF and observed, in situ, a multiplet centered at δ -45
Testing for Lewis Acid Exchange between 1 and BH3‚THF.
To a solution of 1 (0.100 g, 0.369 mmol) in toluene (ca. 2 mL)
BH3‚THF (0.74 mL, 0.740 mmol) was added and stirred at 25 °C.
After 75 min, an aliquot of the reaction mixture was removed and
the 11B and 31P NMR spectra were obtained, showing complete
1
1
in the 11B NMR spectrum (d of q; JB-P ) 47 Hz, JB-H ) 97 Hz).
(25) Schmidbaur, H.; Weiss, E.; Mu¨ller, G. Synth. React. Inorg. Met-Org.
Chem. 1985, 15, 401.
(26) Kulzick, M. A.; Price, R. T.; Andersen, R. A.; Muetterties, E. L. J.
Organomet. Chem. 1987, 333, 105.
Inorganic Chemistry, Vol. 46, No. 18, 2007 7397

Clark et al.
25 °C. After 5 h, the 31P NMR spectrum showed a single resonance
at δ -60.9 corresponding to free PMe3. Similar results were
obtained with Rhcolloid/[Oct4N]Cl.
change from orange to black, and gas evolution was also
observed. The volatile product was assumed to be H2 based
on the detection of this dissolved gas in the reaction solution
1
by H NMR (Hδ 4.46; lit. Hδ 4.46).29 The color change
Treatment of 2 with Rh/Al2O3 and Subsequent Testing of
Catalytic Dehydrocoupling Activity. To a solution of 2 (0.088 g,
0.591 mmol) and Me2NH‚BH3 (0.035 g, 0.594 mmol) in toluene
(ca. 2 mL) a catalytic amount of Rh/Al2O3 (ca. 2 mol %) was added,
resulting in vigorous bubbling and giving a black solution, which
was stirred at 25 °C. After 16 h, the 31P NMR spectrum showed a
single resonance at δ -60.9 corresponding to free PMe3 whereas
the 11B NMR spectrum showed a single resonance at δ -13.9
corresponding to unreacted Me2NH‚BH3.
observed was similar to that detected during the catalytic
dehydrocoupling of amine-borane adducts, in which the
reduction of the Rh(I) (pre)catalyst occurs over a period of
45-200 min to give catalytically active Rh(0) colloids.7,12
In addition, complete adduct dissociation rather than dehy-
drocoupling was observed based on the quantitative forma-
tion of free Cy2PH detected by 31P NMR (Pδ -28.1) and
the apparent generation of Ga metal as a precipitate
(eq 4).
Synthesis of Me3P‚Ga[Co(CO)4]3 (3). A solution of 2 in Et2O
(0.100 g, 0.672 mmol) was added to a solution of [Co2(CO)8] in
Et2O (0.115 g, 0.336 mmol) at 25 °C. This led to vigorous bubbling
1
ca. 5 mol % [Rh]
due to the release of a gas that was identified as H2 by H NMR
Cy2PH‚GaH3
8 Cy2PH + Ga
(4)
when the reaction was performed in C6D6 (δ 4.46). After 1 min,
the solvent was removed in vacuo and the residue dissolved in
toluene, giving a maroon solution that was recrystallized at -30
°C affording orange-red crystals of 3 in 78% yield (0.115 g, 0.175
toluene, 25 °C
-H2
1
It has been previously reported that Ga metal autocata-
lytically promotes the decomposition of Li[GaH4].30 To test
for similar chemistry, we treated 1 with a catalytic amount
of Ga metal but did not observe any reaction, which suggests
that Rh(0) is the sole contributor to the metal-assisted
dissociation process. Further evidence for the reduction of
the Rh(I) (pre)catalyst to catalytically active Rh(0) metal
particles was provided by the observation that treatment of
1 with 5 mol % of Rh/Al2O3 or [Oct4N]Cl-stabilized Rh
colloids (hereafter referred to as Rhcolloid/[Oct4N]Cl) similarly
resulted in H2 evolution and the quantitative formation of
Cy2PH. Thus, the gallane moiety in 1 presumably behaves
as a “borohydride-type” reducing agent to afford Rh(0) metal.
1
2
mmol). H NMR (C6D6): δ 0.92 (d, PMe3, JPH ) 10.5 Hz). 13C
1
NMR (C6D6): δ 198.9 (s, CO), 10.5 (d, PMe3, JCP ) 31.1 Hz).
31P{1H} NMR (C6D6): δ -32 (s, PMe3). IR (toluene solution):
νCO ) 1966, 1982, 2000, 2025, 2072, 2088 cm-1. EI-MS charac-
terization was attempted numerous times but did not yield useful
data presumably due to the exceptional reactivity and sensitivity
of the compound.
Synthesis of [(Me3P)Co(CO)3]2 (4).27 A solution of 2 in Et2O
(0.100 g, 0.672 mmol) was added to a solution of [Co2(CO)8] in
Et2O (0.115 g, 0.336 mmol) at 25 °C. This led to vigorous bubbling
1
due to the release of a gas that was identified as H2 by H NMR
when the reaction was performed in C6D6 (δ 4.46). After 16 h, the
solvent was removed in vacuo and the residue dissolved in hexane,
giving a maroon solution that was recrystallized at -30 °C affording
dark red-brown crystals of 4 in 78% yield (0.038 g, 0.087 mmol).
Previously, our group has reported that Me2NH‚BH3
undergoes Rh-catalyzed dehydrocoupling at 25 °C to afford
the cyclic dimer [Me2N_BH2]2 (eq 3).7 A mixture of 1 and
Me2NH‚BH3 was stirred for several hours with no change
detected in the 31P and 11B NMR spectra. However, upon
the addition of a catalytic amount of [{Rh(µ-Cl)(1,5-cod)}2]
to this mixture, the solution again immediately turned black
and H2 evolution was observed. The 31P and 11B NMR
spectra were recorded and showed the presence of free Cy2-
PH and, surprisingly, unreacted Me2NH‚BH3, respectively.
The absence of dehydrocoupling of the latter was intriguing
and implied that deactivation of the Rh(0) catalyst was taking
place. There is well-established precedent for the poisoning
of transition metal catalysts by strongly coordinating ligands
such as PPh331 or CS232, which block access of the substrate
to the active site. Indeed, the stoichiometry of the donor
ligand per metal atom required for effective catalyst poison-
ing acts as an aid in determining whether a process proceeds
by a heterogeneous or homogeneous mechanism. Thus, Finke
et al. have demonstrated that if substantially less than 1 equiv
of the added ligand (per metal atom) poisons the catalyst
completely, then it is highly suggestive of a heterogeneous
2
1H NMR (C6D6): δ 0.65 (d, PMe3, JPH ) 10.8 Hz). 13C NMR
(C6D6): δ 203.1 (s, CO), 18.8 (d, PMe3, 1JCP ) 31.9 Hz). 31P{1H}
NMR (C6D6): δ 26 (s, PMe3). EI-MS (70 eV) m/z: 409 (M+
-
CO). IR (hexadecane solution): νCO ) 1950, 1972 cm-1 27a
.
Results and Discussion
Catalytic Dissociation and Poisoning Behavior of Cy2PH‚
GaH3 (1). Lewis acid-base adducts of group 13/15 com-
pounds such as phosphine-gallanes have been the subject
of considerable interest as single-source metal-organic
chemical vapor deposition (MOCVD) precursors in thin-film
deposition for microelectronic device fabrication.28 Generally,
compounds of this type are synthesized by salt elimination
reactions. The secondary phosphine-gallane Cy2PH‚GaH3
(1) (Cy ) cyclohexyl) was prepared in 79% yield from the
dehydrogenative metathesis reaction of Li[GaH4] and [Cy2-
PH2]Cl. Compound 1 is stable in solution at room temper-
ature for several hours before decomposition to Cy2PH and
Ga metal is detected. However, treatment of 1 with 5 mol
% of [{Rh(µ-Cl)(1,5-cod)}2] in toluene led to a rapid color
(27) (a) Samir Arabi, M.; Maisonnat, A.; Attali, S.; Poilblanc, R. J.
Organomet. Chem. 1974, 67, 109. (b) Jones, R. A.; Seeberger, M. H.;
Stuart, A. L.; Whittlesey, B. R.; Wright, T. C. Acta Crystallogr. 1986,
C42, 399.
(29) Bo¨hm, V. P. W.; Brookhart, M. Angew. Chem., Int. Ed. 2001, 40,
4694.
(30) Greenwood, N. N. New Pathways in Inorganic Chemistry; Ebsworth,
E. A. V., Maddock, A. G., Sharpe, A. G., Eds.; Cambridge University
Press: Cambridge, U.K., 1968; p 37.
(28) Cowley, A. H.; Jones, R. A. Angew. Chem., Int. Ed. 1989, 28, 1208.
7398 Inorganic Chemistry, Vol. 46, No. 18, 2007

Catalyzed Dissociation of Phosphine-Gallane Adducts
catalyst.31 The presence of Cy2PH along with unreacted Me2-
NH‚BH3 suggested that the decomposition of Cy2PH‚GaH3
occurred with subsequent binding of the phosphine to the
active Rh catalyst, poisoning it toward the dehydrocoupling
of Me2NH‚BH3.
with the Ar+ ion beam resulting in decomposition to
elemental phosphorus. Nevertheless, the key result is that
the studies of colloidal Rh treated with Cy2PH employing
XPS verified the presence of phosphorus on the surface of
the catalyst particles.
We were also interested in investigating the fate of the
GaH3 moiety in the catalytic P-Ga bond cleavage reactions
and the possibility that this may also act as a potential source
of catalyst deactivation. As noted above, we have previously
reported the poisoning of heterogeneous Rh(0) dehydrocou-
pling catalysts by group 13 hydride species including Ga-H
bonds such as those present in GaH3‚OEt2.14 In light of these
observations, the potential role of the GaH3 moiety in 1 as
an additional contributor to the deactivation of the Rh(0)
catalyst cannot be ruled out. Unfortunately, accurate XPS
studies of Rhcolloid/[Oct4N]Cl samples treated with GaH3‚OEt2
or 1 could not be carried out, as Ga metal, a byproduct of
both reactions, and colloidal Rh(0) could not be separated.
To investigate the poisoning of the Rh catalyst toward the
dehydrocoupling of Me2NH‚BH3 in the presence of substo-
ichiometric amounts of Cy2PH, Rhcolloid/[Oct4N]Cl was
treated with 0.5 equiv of Cy2PH and allowed to react for 1
day. Following the evacuation of volatiles, washing with
hexanes, and drying overnight to remove excess phosphine,
a portion of the Rhcolloid/[Oct4N]Cl was removed and used
in an attempt to catalyze the dehydrocoupling of Me2NH‚
BH3.33 It was found that there was greatly reduced catalytic
activity, as only 2.5% conversion to [Me2N-BH2]2 was
observed by 11B NMR spectroscopy after 16 h. This
suggested that Cy2PH does indeed bind to most of the active
sites on the Rh(0) catalyst surface. Furthermore, the strength
of binding is significant, as the phosphine ligands were not
readily eliminated under reduced pressure or solvent wash-
ing.34 The colloidal Rh catalyst was also examined by XPS
before and after treatment with Cy2PH. As expected, the XPS
spectrum of the colloids before treatment showed essentially
no phosphorus content, as indicated by the lack of a P(2p)
peak. However, the colloids after treatment with Cy2PH
showed a P(2p) peak at 132.6 eV, which falls into the typical
range for phosphine oxides (ca. 132.0-134.0 eV).35 In
addition, the expected Rh(3d5/2) peak at 308.0 eV was
observed, which is characteristic of an oxide species, as it
falls into the typical range for Rh2O3 (ca. 308.2-308.8 eV).35
In the case of both Rh and P, it is likely that an oxide species
forms when the colloids are exposed to the atmosphere
during preparation of the sample for XPS analysis. Neglect-
ing the other elements present (e.g., C, O, Si), the relative
atomic concentrations of Rh and P in the colloids exposed
to Cy2PH were found to be approximately 1:2. After
sputtering the surface with Ar+ ions, it was found that the
relative atomic concentration of Rh increased while that of
P decreased (3:1 ratio). This is indicative of a higher
concentration of phosphorus on the surface of the colloids.
Upon sputtering, the Rh(3d5/2) peak remained unchanged but
the P(2p) peak was observed at a lower binding energy of
130.2 eV which is characteristic of elemental phosphorus
(ca. 129.7-130.3 eV), although the presence of free Cy2PH
cannot be ruled out, as the binding energy values for
phosphines span the range of 130.9-132.5 eV.35 It is possible
that the phosphine may undergo degradation upon treatment
Catalytic Dissociation and Poisoning Behavior of Me3P-
GaH3 (2). Next, we turned our attention toward the tertiary
phosphine-gallane Me3‚GaH3 2.36 Compound 2 was pre-
pared, as previously reported, from the dehydrogenative
metathesis reaction of Li[GaH4] and [Me3PH]Cl as a white
crystalline solid which is stable in solution under N2 at room
temperature for over 24 h.20
In contrast to that of 1, treatment of 2 with a catalytic
quantity (ca. 2 mol %) of [{Rh(µ-Cl)(1,5-cod)}2] did not
result in gas evolution or a color change to black, charac-
teristic of substantial Rh(I) reduction to Rh(0) colloidal metal.
Furthermore, a 31P NMR spectrum of the mixture showed
no evidence of reaction after 16 h. This suggested that, in
contrast to 1, adduct 2 is not sufficiently hydridic to reduce
the Rh(I) (pre)catalyst to Rh(0) metal, although a reaction
that prevents Rh(I) reduction may occur below the spectro-
scopic detection limit and therefore could not be ruled out.
With this in mind, we explored the reaction of 2 with a
stoichiometric amount of [{Rh(µ-Cl)(1,5-cod)}2] (e.g., P/Rh
) 1:1). Significantly, this solely afforded the known
compound [(Me3P)RhCl(1,5-cod)] as determined by 31P and
1H NMR spectroscopy.26 The reactivity of 2 with 50 and 10
mol % [{Rh(µ-Cl)(1,5-cod)}2] was also investigated, and in
both cases, 31P NMR spectra revealed a complex mixture of
products, including [(Me3P)RhCl(1,5-cod)], with several
unassigned doublets observed consistent with the presence
of P-Rh coupling. As a result of these observations, it
appears that although the reduction of the Rh(I) species by
2 does not take place, an alternative reaction indeed occurs.
Consistent with this behavior, the addition of 2 mol % [{Rh-
(µ-Cl)(1,5-cod)}2] to an equimolar mixture of 2 and Me2-
NH‚BH3 resulted in no reaction after 16 h, as detected by
31P and 11B NMR spectroscopy.37 This and the previous
(31) Widegren, J. A.; Finke, R. G. J. Mol. Catal. A: Chem. 2003, 198,
317.
(32) Hornstein, B. J.; Aiken, J. D., III.; Finke, R. G. Inorg. Chem. 2002,
41, 1625.
(33) In stoichiometric reactions no evidence for ligand exchange between
Cy2PH and Me2NH‚BH3, to yield Cy2PH‚BH3, was observed by either
31P or 11B NMR spectroscopy. See ref 24 for spectroscopic details
for Cy2PH‚BH3.
(36) We probed the existence of exchange reactions between BH3‚THF
and both 1 and 2. In both cases, complete P-Ga bond cleavage was
observed with concomitant formation of Cy2PH‚BH3 and Me3P‚BH3,
respectively. This indicates that for both adducts, the P-Ga bond is
thermodynamically labile. See the Experimental Section for details.
(37) Similar to the case of 1, no exchange reaction between 2 and Me2-
NH‚BH3 was detected by 31P and 11B NMR spectra.
(34) Analysis of these Rh(0) colloids treated with Cy2PH by 31P NMR
spectroscopy has been unsuccessful despite numerous attempts.
(35) Moulder, J. F.; Stickle, W. F.; Sobol, P. E.; Bomben, K. D. Handbook
of X-ray Photoelectron Spectroscopy; Chastain, J., Ed.; Perkin-
Elmer: Eden Prairie, MN, 1992.
Inorganic Chemistry, Vol. 46, No. 18, 2007 7399

Clark et al.
Scheme 2. Postulated Mechanism of Rh(0)-catalyzed P-Ga Bond Dissociation in 2
observations strongly suggest that the treatment of 2 with a
catalytic quantity of [{Rh(µ-Cl)(1,5-cod)}2] leads to PMe3
binding to the Rh(I) center (which is below the spectroscopic
detection limit). This prevents reduction to catalytically active
Rh(0) metal and consequently hinders the dehydrocoupling
of Me2NH‚BH3. Indeed, even when 10 or 50 mol % [{Rh-
(µ-Cl)(1,5-cod)}2] was added, only 2% and 6% conversion
to [Me2N-BH2]2 was observed, respectively, by 11B NMR
spectroscopy.
is more phosphorus on the surface of the colloidal Rh with
the secondary phosphine than the tertiary phosphine. Similar
to the case for Cy2PH, a higher concentration of phosphorus
on the surface of the colloids was indicated as the atomic
concentration of Rh to P increased to 3.5:1 after sputtering
with Ar+ ions. Upon sputtering, the Rh(3d5/2) peak remained
unchanged but the P(2p) peak was again observed at a lower
binding energy of 130.0 eV, which is characteristic of
elemental phosphorus, possibly due to ion beam-induced
degradation effects (129.7-130.3 eV).35 In summary, studies
of colloidal Rh treated with PMe3 employing XPS again
verified the presence of phosphorus on the surface of the
catalyst particles. As was the case with 1, despite previous
evidence for GaH3‚OEt2 acting as a dehydrocoupling catalyst
poison,14 accurate XPS studies of Rhcolloid/[Oct4N]Cl samples
treated with GaH3‚OEt2 or 2 could not be carried out.
Synthesis, Characterization, and Reactivity of Model
Compounds Me3P‚Ga(Co(CO)4)3 (3) and [(Me3P)Co-
(CO)3]2 (4). The Rh(0)-catalyzed P-Ga bond dissociation
to afford the free phosphine and Ga metal that is observed
when 1 and 2 are treated with Rh(0) suggests that cleavage
of the P-Ga bond occurs through an interaction with the
metal surface. A possible initial step involves the interaction
of the hydridic hydrogen substituents in the GaH3 group of
1 and 2 with the Rh(0) surface. This might weaken the P-Ga
bond in the adduct, resulting in dissociation. Subsequently,
this would be accompanied by decomposition of the GaH3
moiety, as Ga metal and H2 gas were both observed (Scheme
2).
In contrast to the results with Rh(I) precatalysts, treatment
of 2, which is stable in solution at room temperature for up
to 24 h, with 2 mol % Rh/Al2O3 in toluene or Rhcolloid/[Oct4N]-
1
Cl in THF, resulted in immediate gas evolution (H2 by H
NMR) and complete adduct dissociation to afford free PMe3,
which was detected by 31P NMR spectroscopy (Pδ -60.9)
as the only product after 16 h (eq 5). Similar to the reaction
with [{Rh(µ-Cl)(1,5-cod)}2] noted above, a mixture of 2 and
Me2NH‚BH3 was also treated with a catalytic amount of the
Rh(0) source Rh/Al2O3. After 16 h, 31P and 11B NMR
characterization of the black suspension showed free PMe3
and unreacted Me2NH‚BH3, respectively. This observation,
similar to the case with 1, supports the hypothesis that PMe3
binds to the Rh(0) surface (below the spectroscopic detection
limit), poisoning the active sites on the heterogeneous catalyst
and consequently preventing the dehydrocoupling of Me2-
NH‚BH3.
[Rh0]
Me3P‚GaH3
8 PMe3 + Ga
(5)
-H2
2
To gain further insight into this interesting reactivity, we
attempted to model the process by studying the reaction of
2 with a convenient, zerovalent group 9 metal complex. A
reaction between 2 and [Co2(CO)8] in Et2O at 25 °C caused
immediate gas evolution (H2 by 1H NMR) after only 1 min
along with a small change in the 31P NMR chemical shift to
-32 ppm (cf. Pδ -39 for Me3P‚GaH3), affording the cobalt
carbonyl-substituted phosphine-gallane complex 3 in 78%
yield (eq 7).
In an analogous manner to the case of 1, the colloidal
catalyst was examined by XPS before and after treatment
with PMe3. The XPS spectrum of the colloids before
treatment showed essentially no phosphorus present, as
indicated by the lack of a P(2p) peak. However, similar to
1, the colloids after treatment showed a P(2p) peak at 131.9
eV, falling close to the typical range for phosphine oxides
(ca. 132.0-134.0 eV).35 Here, neglecting the other elements
present (e.g., C, O, Si), the relative atomic concentrations
of Rh and P were found to be approximately 2:1. This can
be compared to the 1:2 ratio found in the case of Rh colloids
that had been treated with Cy2PH, which indicated that there
Et2O, 25 °C
1 min
Me3P‚GaH
Me P‚Ga[Co(CO) ]
(7)
4 3
3 + Co2(CO)8
8
3
-H2
2
3
7400 Inorganic Chemistry, Vol. 46, No. 18, 2007

Catalyzed Dissociation of Phosphine-Gallane Adducts
various chlorogallium species.42 Cowley and co-workers have
described the unique cobalt-gallium complex [{2,6-Me2-
NCH2)2C6H3}Ga{Co(CO)4}2], which features intramolecular
base stabilization by one of the dimethylamine groups.43
More recently, the first gallane-coordinated metal complex,
(OC)5W(η1-GaH3‚(quinuclidine)), was synthesized from
W(CO)6 and H3Ga‚(quinuclidine) via a photolysis method.44
Structural characterization of the product revealed that the
gallane is bound to the tungsten fragment via a W-H-Ga
three-center two-electron bond. Shimoi and co-workers have
reported the reaction of [Co2(CO)8] with the diadduct B2H4‚
2PMe3 that yielded [{Co(CO)3}2(µ-CO)(µ-BH-PMe3)], which
contains a bridging borylene-Lewis base unit.45 In the case
of 3, the binding of three Co(CO)4 fragments to a single Ga
center is a unique entry in bonding for these types of systems.
The IR spectrum of 3 in toluene was recorded and showed
six bands corresponding to CO stretches ranging from 1966
to 2088 cm-1, values which fall in the accepted approximate
values of 1850-2125 cm-1 for a terminal carbonyl moiety.
Interestingly, the IR spectrum for the previously mentioned
cobalt-gallium complex reported by Cowley and co-workers
also showed six bands attributed to the CO ligands on the
Co(CO)4 fragments in the region between 1878 and 2080
cm-1.43
Consistent with the weakened P-Ga bond, solutions of 3
in Et2O form the known complex 4 over several hours at
25 °C (eq 8). To the best of our knowledge, 4 has only been
characterized by IR spectroscopy27a and X-ray crystallo-
graphy.27b We confirmed the identity of 4 by these methods
as well as multinuclear NMR spectroscopy and mass
spectrometry. For instance, the 31P NMR spectrum revealed
a single resonance at 26 ppm whereas two resonances were
observed in the 13C NMR spectrum, a singlet at 203 ppm
and a doublet at 18 ppm (1JCP ) 31.9 Hz), corresponding to
CO and Me carbon atoms, respectively. The IR spectrum of
4 contains two bands corresponding to CO stretches at 1950
and 1972 cm-1, which, on average, appear at lower wave-
number values than those observed for 3. This can be
attributed to the presence of the strong σ-donor PMe3
molecules capping both ends of the Co2(CO)6 chain in 4,
placing more electron density on the metal centers and, thus,
increased back donation into the CO π* orbitals. In contrast,
the Lewis acidic gallium center in 3 does not donate
significant electron density onto the Co centers, as reflected
in the relatively high CO stretching frequencies.
Figure 1. Molecular structure of 3. Selected bond lengths (Å) and angles
(deg): Ga(1)-P(1) 2.5108(12), Ga(1)-Co(1) 2.5557(7), Ga(1)-Co(2)
2.5552(7), Ga(1)-Co(3) 2.5686(7), Co-C 1.790(5) av.; P(1)-Ga(1)-Co-
(1) 103.91(3), P(1)-Ga(1)-Co(2) 109.10(3), P(1)-Ga(1)-Co(3) 102.23-
(3), Co-Ga(1)-Co 113.49(3) av., C-P(1)-C 104.0(2) av., C-P(1)-Ga(1)
114.50(16) av.
Orange-red crystals were obtained from toluene at
-30 °C, and analysis by X-ray crystallography confirmed
the structure of 3 (Figure 1). The P-Ga bond length in 3
was found to be 2.5108(12) Å, significantly longer than that
found in 2 (2.3857(6) Å).38 Notably, the P-Ga bond length
in the bulky adduct Cy3P‚GaH3 is also considerably longer
than that in 2 (2.460(2) Å).39 Therefore, the lengthened P-Ga
bond length in 3 may be a consequence of the increased steric
bulk of the three Co(CO)4 ligands around the Ga center
relative to the smaller hydrogen atoms. However, an
electronic effect probably also plays a role as the electron-
rich Co(CO)4 substituents will tend to relieve the electron
deficiency at Ga, thereby decreasing the strength of the P-Ga
bond. Both the P and Ga centers adopt a distorted tetrahedral
geometry with angles ranging from 102° to 115° whereas
the Me and Co(CO)4 substituents around those centers are
staggered relative to each other. The average Co-Ga bond
length in 3 is 2.5598(7) Å, which falls into the range of
2.38-2.58 Å for typical Co-Ga single bonds.40
Transition metal complexes containing bonds with group
13 elements are an intriguing class of compounds that have
received attention due to their use as potential precursors
for MOCVD.41 Fischer and co-workers have reported the
preparation of transition metal substituted gallanes by salt
elimination reactions of transition carbonyl metallates with
The conversion of 3 to 4 intriguingly mimics the proposed
mechanism in Scheme 2 for the reaction involving colloidal
Rh with 1 and 2. Weakening of the P-Ga bond is clearly
observed in 3 and subsequently results in P-Ga bond
cleavage; in this case, free PMe3 is not generated, as this
(38) Tang, C. Y.; Coxall, R. A.; Downs, A. J.; Greene, T. M.; Kettle, L.;
Parsons, S.; Rankin, D. W. H.; Robertson, H. E.; Turner, A. R. Dalton
Trans. 2003, 3526.
(39) Atwood, J. L.; Robinson, K. D.; Bennett, F. R.; Elms, F. M.;
Koutsantonis, G. A.; Raston, C. L.; Young, D. J. Inorg. Chem. 1992,
31, 2673.
(42) (a) Fischer, R. A.; Miehr, A.; Priermeier, T. Chem. Ber. 1995, 128,
831. (b) Fischer, R. A.; Miehr, A.; Hoffmann, H.; Rogge, W.; Boehme,
C.; Frenking, G.; Herdtweck, E. Z. Anorg. Allg. Chem. 1999, 625,
1466.
(40) Compton, N. A.; Errington, R. J.; Norman, N. C. AdV. Organomet.
Chem. 1990, 31, 91.
(41) For reviews on transition metal complexes of boron and other group
13 elements, see: (a) Irvine, G. J.; Lesley, M. J. G.; Marder, T. B.;
Norman, N. C.; Rice, C. R.; Robins, E. G.; Roper, W. R.; Whittell,
G. R.; Wright, L. J. Chem. ReV. 1998, 98, 2685. (b) Fischer, R. A.;
Weib, J. Angew. Chem., Int. Ed. 1999, 38, 2830. (c) Braunschweig,
H.; Kollann, C.; Rais, D. Angew. Chem., Int. Ed. 2006, 45, 5254.
(43) Olaza´bal C. A.; Gabba¨ı, F. P.; Cowley, A. H.; Carrano, C. J.; Mokry,
L. M.; Bond, M. R. Organometallics 1994, 13, 421.
(44) Ueno, K.; Yamaguchi, T.; Uchiyama, K.; Ogino, H. Organometallics
2002, 21, 2347.
(45) Shimoi, M.; Ikubo, S.; Kawano, Y.; Katuh, K.; Ogino, H. J. Am. Chem.
Soc. 1998, 120, 4222.
Inorganic Chemistry, Vol. 46, No. 18, 2007 7401

Clark et al.
coordinates to reactive Co carbonyl fragments to form 4.46
involving the treatment of the Rh(0) colloidal catalyst with
Cy2PH, PMe3, and GaH3‚OEt2 revealed that both phosphines
and the gallane poison the catalyst to varying degrees. XPS
studies verified the presence of phosphorus on the surface
of the Rh(0) colloids. A model reaction with a zerovalent
group 9 metal complex was performed in an effort to probe
the mechanism behind the metal-assisted dissociation pro-
cess. Specifically, the reaction between [Co2(CO)8] and 2
first afforded 3, which possesses a P-Ga bond, and
subsequently afforded 4, with concomitant Ga metal pre-
cipitation. This suggests that a reasonable mechanism
involves the reaction of the Ga-H bonds on the Rh colloid
surface and the weakening of the P-Ga bond leading to
adduct dissociation. Future work is directed at studying the
catalytic P-Ga bond cleavage process in more detail and
extending our studies to adducts containing other heavier
group 13 and 15 elements.
Et2O, 25 °C
4 h
Me3P‚Ga[Co(CO)4]3
[(Me P)Co(CO) ]
3 3 2
8
(8)
-H2, -Ga
3
4
Conclusions
The successful formation of P-B and N-B bonds by the
transition metal-catalyzed dehydrocoupling of phosphine-
and amine-borane adducts, respectively, prompted our
investigations into the potential extension of this synthetic
approach to P-Ga systems. Consequently, the secondary
phosphine-gallane 1 was treated with Rh(I) or Rh(0)
catalysts in toluene at 25 °C. However, instead of dehydro-
coupling, this resulted in the formation of Ga metal, release
of H2 gas, and catalytic P-Ga bond cleavage generating the
phosphine quantitatively. Similarly, the tertiary phosphine-
gallane 2 also dissociated when treated with a catalytic
amount of Rh(0). Interestingly, when Me2NH‚BH3 was
subsequently added to the reaction mixture, no dehydrocou-
pling to afford the cyclic dimer [Me2N-BH2]2 was detected,
indicating that catalyst deactivation is taking place. Studies
Acknowledgment. T.J.C. is grateful to the Ontario
Government for an OGS fellowship as well as a CIBA
Specialty Chemicals Award, and C.A.J. thanks the NSERC
of Canada for a scholarship. I.M. would like to thank the
Canadian Government for a Canada Research Chair at
Toronto as well as the EU for a Marie Curie Chair and the
Royal Society for a Wolfson Research Merit Award at
Bristol.
(46) There is also a possibility that the reactions with [Co2(CO)8] may
proceed via routes involving radicals, and this cannot be ruled out at
this time.
IC700573J
7402 Inorganic Chemistry, Vol. 46, No. 18, 2007