Poisoning of heterogeneous, late transition metal dehydrocoupling catalysts by boranes and other group 13 hydrides

Article abstract of DOI:10.1021/ja0447412

Borane reagents are widely used as reductants for the generation of colloidal metals. When treated with a variety of heterogeneous catalysts such as colloidal Rh, Rh/Al₂O₃, and Rh(0) black, BH ₃· THF (THF = tetrahydrofuran) was found to generate H ₂ gas with the concomitant formation of a passivating boron layer on the surface of the Rh metal, thereby acting as a poison and rendering the catalyst inactive toward the dehydrocoupling of Me₂NH-BH₃. Analogous poisoning effects were also detected for (i) colloidal Rh treated with other species containing B-H bonds such as [HB-NH]₃, or Ga-H bonds such as those present in GaH₃-OEt₂, (ii) colloidal Rh that was generated from Rh(I) and Rh(III) salts using borane or borohydrides as reductants, and (iii) for other metals such as Ru and Pd. In contrast, analogous poisoning effects were not detected for the catalytic hydrogenation of cyclohexene using Rh/Al₂O₃ or the Pd-catalyzed Suzuki cross-coupling of PhB(OH)₂ and Phl. These results suggest that although this poisoning behavior is not a universal phenomenon, the observation that such boron layers are formed and surface passivation may exist needs to be carefully considered when borane reagents are used for the generation of metal colloids for catalytic or materials science applications.

Full text of DOI:10.1021/ja0447412

Published on Web 03/17/2005
Poisoning of Heterogeneous, Late Transition Metal
Dehydrocoupling Catalysts by Boranes and Other Group 13
Hydrides
Cory A. Jaska,† Timothy J. Clark,† Scott B. Clendenning,† Dan Grozea,‡
Ayse Turak,‡ Zheng-Hong Lu,‡ and Ian Manners*,†
Contribution from the Department of Chemistry, UniVersity of Toronto, 80 St. George Street,
Toronto, Ontario, Canada M5S 3H6, and Department of Materials Science and Engineering,
UniVersity of Toronto, Ontario, Canada M5S 3E4
Received August 31, 2004; E-mail: imanners@chem.utoronto.ca
Abstract: Borane reagents are widely used as reductants for the generation of colloidal metals. When
treated with a variety of heterogeneous catalysts such as colloidal Rh, Rh/Al2O3, and Rh(0) black, BH3‚
THF (THF ) tetrahydrofuran) was found to generate H2 gas with the concomitant formation of a passivating
boron layer on the surface of the Rh metal, thereby acting as a poison and rendering the catalyst inactive
toward the dehydrocoupling of Me2NH‚BH3. Analogous poisoning effects were also detected for (i) colloidal
Rh treated with other species containing B-H bonds such as [HB-NH]3, or Ga-H bonds such as those
present in GaH3‚OEt2, (ii) colloidal Rh that was generated from Rh(I) and Rh(III) salts using borane or
borohydrides as reductants, and (iii) for other metals such as Ru and Pd. In contrast, analogous poisoning
effects were not detected for the catalytic hydrogenation of cyclohexene using Rh/Al2O3 or the Pd-catalyzed
Suzuki cross-coupling of PhB(OH)2 and PhI. These results suggest that although this poisoning behavior
is not a universal phenomenon, the observation that such boron layers are formed and surface passivation
may exist needs to be carefully considered when borane reagents are used for the generation of metal
colloids for catalytic or materials science applications.
Introduction
in synthetic transformations, including alkene and arene hydro-
genation,7 Heck8,9 and Suzuki9,10 couplings, Ullmann-type
In recent years, studies of the synthesis and catalytic activity
of stable transition metal colloids and nanoclusters have attracted
intense interest.1 A multitude of synthetic strategies can now
be employed to afford these types of heterogeneous species that
find applications primarily as catalysts for a number of different
bond-forming reactions.2 For example, nanoclusters and colloids
can be synthesized by electrochemical3 or chemical reduction4
methods employing either electrostatic (e.g. polyoxoanions),4a
steric (e.g. polymers, dendrimers, or microgels)5,6 or electrosteric
(e.g. R4N+ where R is a long chain alkyl group)4b stabilization
in order to prevent aggregation. The chemical reduction of
transition metal salts can be performed by a variety of reducing
agents such as alcohols, dihydrogen, borohydrides, or boranes
and is probably the most common synthetic route to colloids.2
Nanoclusters and colloids have found extensive use as catalysts
aminations,11 ring-opening polymerization,12 silaesterification,13
(5) For representative examples, see the following. Polymer stabilized: (a)
Chan, Y. N. C.; Schrock, R. R.; Cohen, R. E. J. Am. Chem. Soc. 1992,
114, 7295. (b) Golden, J. H.; Deng, H.; DiSalvo, F. J.; Fre´chet, J. M. J.;
Thompson, P. M. Science 1995, 268, 1463. (c) Watkins, J. J.; McCarthy,
T. J. Chem. Mater. 1995, 7, 1991. (d) Teranishi, T.; Hosoe, M.; Miyake,
M. AdV. Mater. 1997, 9, 65. Dendrimer stabilized: (e) Zhao, M.; Crooks,
R. M. Angew. Chem., Int. Ed. 1999, 38, 364. (f) Zhao, M.; Crooks, R. M.
AdV. Mater. 1999, 11, 217. Microgel stabilized: (g) Antonietti, M.; Gro¨hn,
F.; Hartmann, J.; Bronstein, L. Angew. Chem., Int. Ed. Engl. 1997, 36,
2080. (h) Whilton, N. T.; Berton, B.; Bronstein, L.; Hentze, H. P.;
Antonietti, M. AdV. Mater. 1999, 11, 1014. (i) Biffis, A.; Orlandi, N.;
Corain, B. AdV. Mater. 2003, 15, 1551.
(6) For an example of using proteins to stabilize nanoclusters, see: Ueno, T.;
Suzuki, M.; Goto, T.; Matsumoto, T.; Nagayama, K.; Watanabe, Y. Angew.
Chem., Int. Ed. 2004, 43, 2527.
(7) (a) Bo¨nnemann, H.; Braun, G. A. Angew. Chem., Int. Ed. Engl. 1996, 35,
1992. (b) Weddle, K. S.; Aiken, J. D., III; Finke, R. G. J. Am. Chem. Soc.
1998, 120, 5653. (c) Schulz, J.; Roucoux, A.; Patin, H. Chem. Eur. J. 2000,
6, 618. (d) Dyson, P. J. Dalton Trans. 2003, 2964. (e) Doyle, A. M.;
Shaikhutdinov, S. K.; Jackson, S. D.; Freund, H.-J. Angew. Chem., Int.
Ed. 2003, 42, 5240.
† Department of Chemistry.
(8) (a) Reetz, M. T.; Lohmer, G. Chem. Commun. 1996, 1921. (b) Yeung, L.
K.; Crooks, R. M. Nano Lett. 2001, 1, 14. (c) Rahim, E. H.; Kamounah, F.
S.; Frederiksen, J.; Christensen, J. B. Nano Lett. 2001, 1, 499. (d) Calo´,
V.; Nacci, A.; Monopoli, A.; Detomaso, A.; Ilaide, P. Organometallics
2003, 22, 4193. (e) Gopidas, K. R.; Whitesell, J. K.; Fox, M. A. Nano
Lett. 2003, 3, 1757. (f) Na, Y.; Park, S.; Han, S. B.; Han, H.; Ko, S.; Chang,
S. J. Am. Chem. Soc. 2004, 126, 250.
‡ Department of Materials Science and Engineering.
(1) (a) Metal Nanoparticles: Synthesis, Characterization and Applications;
Feldheim, D. L., Ed.; Dekker: New York, 2002. (b) Nanoparticles: From
Theory to Application; Schmid, G., Ed.; Wiley-VCH: Weinheim, Germany,
2004. (c) Nanoparticles: Building Blocks for Nanotechnology; Rotello, V.
M., Ed.; Kluwer Academic/Plenum: New York, 2004.
(2) Roucoux, A.; Schulz, J.; Patin, H. Chem. ReV. 2002, 102, 3757.
(3) (a) Reetz, M. T.; Helbig, W. J. Am. Chem. Soc. 1994, 116, 7401. (b) Reetz,
M. T.; Quaiser, S. A. Angew. Chem., Int. Ed. Engl. 1995, 34, 2240.
(4) See for example: (a) Lin, Y.; Finke, R. G. J. Am. Chem. Soc. 1994, 116,
8335. (b) Bo¨nnemann, H.; Brijoux, W.; Brinkmann, R.; Dinjus, E.; Joussen,
T.; Korall, B. Angew. Chem., Int. Ed. Engl. 1991, 30, 1312. (c) Raveendran,
P.; Fu, J.; Wallen, S. L. J. Am. Chem. Soc. 2003, 125, 13940.
(9) Moreno-Man˜as, M.; Pleixats, R. Acc. Chem. Res. 2003, 36, 638.
(10) (a) Reetz, M. T.; Breinbaur, R.; Wanninger, K. Tetrahedron Lett. 1996,
37, 4499. (b) Li, Y.; El-Sayed, M. A. J. Phys. Chem. B 2001, 105, 8938.
(c) Pittelkow, M.; Moth-Poulsen, K.; Boas, U.; Christensen, J. B. Langmuir
2003, 19, 7682. (d) Biffis, A.; Sperotto, E. Langmuir 2003, 19, 9548. (e)
Narayanan, R.; El-Sayed, M. A. J. Am. Chem Soc. 2003, 125, 8340.
(11) Son, S. U.; Park, I. K.; Park, J.; Hyeon, T. Chem. Commun. 2004, 778.
9
5116
J. AM. CHEM. SOC. 2005, 127, 5116-5124
10.1021/ja0447412 CCC: $30.25 © 2005 American Chemical Society

Poisoning of Late Transition Metal Catalysts
A R T I C L E S
and allylic alkylation14 reactions. However, one disadvantage
of colloids and nanoclusters (and heterogeneous catalysts in
general) is their ability to be poisoned by substoichiometric
quantities of reagent, which correspondingly reduces or inhibits
their catalytic activity. For example, mercury is a well-known
poison of heterogeneous catalysts through the formation of an
amalgam or adsorption onto the catalyst surface.15 Strongly
coordinating ligands such as PPh3 16 or CS2 17 can also be used
in fractional poisoning experiments.18 Interestingly, these types
of ligands can also function as colloid and nanocluster
stabilizers,19a,20 however this is generally accompanied by a
corresponding reduction of their catalytic activity.21 In some
cases, this poisoning effect can be used as a reliable and
reproducible test to differentiate between homogeneous and
heterogeneous catalysis.16
(eq 5).26 In all cases involving amine-borane adducts, evidence
has been found that the catalytic dehydrocoupling reactions
involve some form of heterogeneous catalyst species, whether
it be oxide-supported metals (e.g. Rh/Al2O3), transition metal
colloids (e.g. [Oct4N]Cl stabilized Rh colloids hereafter referred
to as Rhcolloid/[Oct4N]Cl; Oct ) n-octyl) or the in situ formation
of elemental metal (e.g. Rh(0)) from homogeneous precur-
sors.27,28 Interestingly, the analogous catalytic dehydrocoupling
of phosphine-borane adducts was found to proceed via a
homogeneous mechanism only. It was reasoned that one of the
contributing factors for this disparate behavior may involve
adduct dissociation and subsequent poisoning of heterogeneous
catalysts by free phosphine.28 However, another potential
interfering species is the other dissociation product “BH3”. Thus,
in this paper we report on our further investigations of adduct
dissociation and discuss, in particular, our findings on the
poisoning behavior of BH3‚THF (THF ) tetrahydrofuran) and
other related group 13 hydride species with heterogeneous
transition metal catalysts used for the dehydrocoupling of Me2-
NH‚BH3.
Our group is interested in the development of catalytic routes
for the formation of new bonds between inorganic elements.
Recent work has focused on the catalytic dehydrocoupling of
primary and secondary phosphine-borane R′RPH‚BH3 and
amine-borane adducts R′RNH‚BH3 using primarily Rh (pre)-
catalysts.22 For example, the catalytic dehydrocoupling of
phosphine-borane adducts has been shown to provide access
to oligomeric chains R2PH-BH2-PR2-BH3 (R ) Ph, tBu; eq
1), six- and eight-membered rings [R2P-BH2]x [R ) Ph, x )
3, 4; eq 1), and high molecular weight polymers [RPH-BH2]n
(R ) Ph, iBu, p-BuC6H4; eq 2).23,24 For amine-borane adducts,
catalytic dehydrocoupling was found to provide a mild and
convenient route to cycloaminoboranes [R2N-BH2]2 (R ) Me,
1,4-C4H8; eq 3) and borazines [RN-BH]3 (R ) H, Me, Ph; eq
4).25 In addition, a tandem catalytic dehydrocoupling-hydro-
genation reaction involving a variety of Rh (pre)catalysts and
Me2NH‚BH3 as a stoichiometric hydrogen source for the
hydrogenation of alkenes at 25 °C has been recently developed
(12) (a) Wu, X.; Neckers, D. C. Macromolecules 1999, 32, 6003. (b) Temple,
K.; Ja¨kle, F.; Sheridan, J. B.; Manners, I. J. Am. Chem. Soc. 2001, 123,
1355.
(13) Chauhan, B. P. S.; Rathore, J. S.; Chauhan, M.; Krawicz, A. J. Am. Chem.
Soc. 2003, 125, 2876.
(14) Jansat, S.; Go´mez, M.; Philippot, K.; Muller, G.; Guiu, E.; Claver, C.;
Castillo´n, S.; Chaudret, B. J. Am. Chem. Soc. 2004, 126, 1592.
(15) (a) Anton, D. R.; Crabtree, R. H. Organometallics 1983, 2, 855. (b)
Whitesides, G. M.; Hackett, M.; Brainard, R. L.; Lavalleye, J. P. P. M.;
Sowinski, A. F.; Izumi, A. N.; Moore, S. S.; Brown, D. W.; Staudt, E. M.
Organometallics 1985, 4, 1819.
(16) Widegren, J. A.; Finke, R. G. J. Mol. Catal. A: Chem. 2003, 198, 317.
(17) Hornstein, B. J.; Aiken, J. D., III; Finke, R. G. Inorg. Chem. 2002, 41,
1625.
Results and Discussion
Evidence for Metal-Assisted Dissociation of Me3E‚XH3 (E
) N, P; X ) B, Al, Ga) and the Effect on Catalytic
Dehydrocoupling Activity. During our earlier studies, evidence
was provided supporting the metal-assisted dissociation of
phosphine-borane adducts such as Ph2PR‚BH3 (R ) H, Ph) in
the presence of Rhcolloid/[Oct4N]Cl.28 These dissociation reactions
would generate free phosphine, a well-known poison for
heterogeneous colloidal metal catalysts.16 To establish the
generality of this adduct dissociation chemistry, reactions
involving Rhcolloid/[Oct4N]Cl with the series of phosphine
adducts Me3P‚BH3 and Me3P‚GaH3 and amine adducts Me3N‚
BH3, Me3N‚AlH3, and Me3N‚GaH3 were performed. We also
probed for catalyst poisoning effects by evaluating the subse-
quent catalytic activity of the Rh colloids toward the dehydro-
coupling of Me2NH‚BH3 (eq 3, R ) R′ ) Me). As a control,
(18) A heterogeneous catalyst can be completely poisoned by <1 equiv of ligand
(per metal atom) because only a fraction of the metal atoms are on the
surface, whereas a homogeneous catalyst would require >1 equiv of ligand
to completely poison the active site.
(19) (a) Amiens, C.; de Caro, D.; Chaudret, B.; Bradley, J. S.; Mazel, R.; Roucau,
C. J. Am. Chem. Soc. 1993, 115, 11638. (b) de Caro, D.; Wally, H.; Amiens,
C.; Chaudret, B. J. Chem. Soc., Chem. Commun. 1994, 1891.
(20) (a) Duteil, A.; Schmid, G.; Meyer-Zaika, W. J. Chem. Soc., Chem. Commun.
1995, 31. (b) Jiang, X.; Xie, Y.; Lu, J.; Zhu, L.; He, W.; Qian, Y. Langmuir
2001, 17, 3795.
(21) The stability and catalytic activity of colloidal nanoparticles is anticorre-
lated: the most stable is the least catalytically active. See: Li, Y.; El-
Sayed, M. A. J. Phys. Chem. B 2001, 105, 8938.
(22) Jaska, C. A.; Bartole-Scott, A.; Manners, I. Dalton Trans. 2003, 4015.
(23) (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.; Vejzovic, E.; Lough, A. J.; Manners, I.
Inorg. Chem. 2001, 40, 4327. (d) Dorn, H.; Rodezno, J. M.; Brunnho¨fer,
B.; Rivard, E.; Massey, J. A.; Manners, I. Macromolecules 2003, 36, 291.
(24) For a recent report of 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.
(25) (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.
(26) Jaska, C. A.; Manners, I. J. Am. Chem. Soc. 2004, 126, 2698.
(27) Jaska, C. A.; Manners, I. J. Am. Chem. Soc. 2004, 126, 1334.
(28) Jaska, C. A.; Manners, I. J. Am. Chem. Soc. 2004, 126, 9776.
9
J. AM. CHEM. SOC. VOL. 127, NO. 14, 2005 5117

A R T I C L E S
Jaska et al.
the stability of the adducts in C6D6 solution was first verified.
Indeed, no dissociation of any of the species was observed by
NMR spectroscopy after 24 h at 25 °C.
For the adducts Me3P‚BH3 and Me3P‚GaH3, slight (ca. 7%)
and complete (100%) dissociation was observed, respectively,
to generate free PMe3 upon treatment with Rhcolloid/[Oct4N]Cl
(24 h, 25 °C, C6D6). Subsequent treatment of these mixtures
with Me2NH‚BH3 resulted in no catalytic dehydrocoupling
activity to form [Me2N-BH2]2. Based on the detection of adduct
dissociation, the poisoning effect observed presumably arises
from ligation of the PMe3 groups to the surface of the Rh(0)
colloids, which blocks the active catalytic sites. We have
previously shown that free phosphine is a potent, substoichio-
metric poison for colloidal Rh-catalyzed dehydrocoupling
reactions.28 In addition, in the case of Me3P‚BH3, the 31P{1H}
NMR spectrum of the reaction mixture showed the presence of
a cluster of signals centered at δ -10 ppm that can be assigned
to the surface ligated PMe3 groups (e.g. {Rhcolloid(PMe3)/[Oct4N]-
Cl}). Indeed, treatment of Rhcolloid/[Oct4N]Cl with PMe3 showed
a similar cluster of signals in the 31P{1H} NMR spectrum (δ
ca. -10 ppm), which were shifted downfield relative to free
PMe3 (δ ca. -61 ppm). Previous work by Chaudret and co-
workers support this 31P NMR assignment of {Rhcolloid(PMe3)/
[Oct4N]Cl}, as they have also observed a general downfield
shift for PPh3 on the surface of Pt, Pd, and Cu particles
compared to the free phosphine.19,29
Previous work on the dissociation of Me2NR‚BH3 (R ) H,
Me) in the presence of colloidal Rh(0) metal has suggested that
cleavage of the adduct N-B bonds either does not occur or
occurs to a very small extent.28 Thus, treatment of the series of
adducts Me3N‚XH3 (X ) B, Al, Ga) with Rhcolloid/[Oct4N]Cl
was found to result in partial (ca. 6 and 31%) dissociation for
the borane and alane, respectively, and complete (100%)
dissociation for the gallane as indicated by the 1H NMR spectra
(24 h, 25 °C, C6D6). Subsequent treatment of these mixtures
with Me2NH‚BH3 was found to result in catalytic dehydro-
coupling to yield [Me2N-BH2]2 for the borane case only, while
the analogous alane and gallane treated Rh colloids did not show
any significant catalytic activity. One possible explanation to
consider is that in the latter cases NMe3 might be poisoning
the catalyst in a manner similar to that of PMe3. Significantly,
treatment of Rhcolloid/[Oct4N]Cl with a variety of amines (e.g.
NEt3, Me2NH, and pyridine) did not show any suppression of
catalytic activity toward the dehydrocoupling of Me2NH‚BH3.
This suggested that the species responsible for the poisoning
of the catalyst might be the group 13 hydride species “XH3”
instead. Further detailed experiments were therefore performed
to test this hypothesis (vide infra). However, first we conclude
this section with some further considerations concerning the
metal-assisted adduct bond cleavage process.
Figure 1. Possible coordination modes for Me3E‚XH3 adducts (E ) N, P;
X ) B, Al, Ga) with transition metal (M) centers.
BH3 (145.6 kJ/mol) < Me3P‚BH3 (159.4 kJ/mol).30-33 The
extent of dissociation of these adducts was indeed observed to
broadly follow the same general trend, with weaker adducts
undergoing a higher degree of dissociation in C6D6 at 25 °C
over 24 h in the presence of Rhcolloid/[Oct4N]Cl: Me3P‚GaH3
(100%) ) Me3N‚GaH3 (100%) > Me3N‚AlH3 (31%) > Me3N‚
BH3 (6%) ≈ Me3P‚BH3 (7%). This dissociation process may
proceed via coordination of the XH3 (X ) B, Al, Ga) moiety
to the surface Rh atoms by means of the hydridic hydrogen
atoms (e.g. Hδ-). This coordination might involve the formation
of a transient η1, η2, or η3 complex or alternatively by the
formation of a σ complex (Figure 1). These types of coordination
modes are all well-established in the literature, including, for
example, complexes that contain amine- and phosphine-borane
adducts.34 Coordination in this manner may then weaken the
adduct dative bond (e.g. EfX), thereby leading to displacement
of the Lewis base and adduct dissociation.35
Treatment of Rhcolloid/[Oct4N]Cl with BH3‚THF and
Analysis of the Subsequent Catalytic Dehydrocoupling
Activity. To further explore the potential poisoning effects by
“BH3” generated in metal-assisted adduct dissociation reactions,
a series of experiments involving colloidal Rh and BH3‚THF
were performed. Treatment of Rhcolloid/[Oct4N]Cl36 in THF with
excess BH3‚THF (ca. 5 equivalents per Rh atom) was found to
result in an immediate reaction as indicated by the release of a
gas, which was determined to be H2 (Hδ 4.46; lit. Hδ 4.46).37
(30) 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.
(31) Loschen, C.; Voigt, K.; Frunzke, J.; Diefenbach, A.; Diedenhofen, M.;
Frenking, G. Z. Anorg. Allg. Chem. 2002, 628, 1294.
(32) Haaland, A. Angew. Chem., Int. Ed. Engl. 1989, 28, 992.
(33) The adduct dissociation energy is also dependent on the type of substituents
present on N and P. For example, the unsubstituted adducts NH3‚BH3 and
PH3‚BH3 have dissociation energies of 116 and 91.2 kJ mol-1, respectively.
See: Rablen, P. R. J. Am. Chem. Soc. 1997, 119, 8350.
(34) For η1 complexes, see: (a) Shimoi, M.; Nagai, S. I.; Ichikawa, M.; Kawano,
Y.; Katoh, K.; Uruichi, M.; Ogino, H. J. Am. Chem. Soc. 1999, 121, 11704.
(b) Kakizawa, T.; Kawano, Y.; Shimoi, M. Organometallics 2001, 20, 3211.
(c) Bajko, Z.; Daniels, J.; Gudat, D.; Ha¨p, S.; Nieger, M. Organometallics
2002, 21, 5182. For η2 complexes, see: (d) Khan, K.; Raston, C. L.;
McGrady, J. E.; Skelton, B. W.; White, A. H. Organometallics 1997, 16,
3252. (e) Ingleson, M.; Patmore, N. J.; Ruggiero, G. D.; Frost, C. G.;
Mahon, M. F.; Willis, M. C.; Weller, A. S. Organometallics 2001, 20,
4434. (f) Volkov, O.; Mac´ıas, R.; Rath, N. P.; Barton, L. Inorg. Chem.
2002, 41, 5837. For η3 complexes, see: (g) White, J. P., III; Deng, H.;
Shore, S. G. Inorg. Chem. 1991, 30, 2337. (h) Gozum, J. E.; Wilson, S.
R.; Girolami, G. S. J. Am. Chem. Soc. 1992, 114, 9483. For σ complexes,
see: (i) Hartwig, J. F.; Muhoro, C. N.; He, X.; Eisenstein, O.; Bosque, R.;
Maseras, F. J. Am. Chem. Soc. 1996, 118, 10936. (j) Muhoro, C. N.; He,
X.; Hartwig, J. F. J. Am. Chem. Soc. 1999, 121, 5033. (k) Schlecht, S.;
Hartwig, J. F. J. Am. Chem. Soc. 2000, 122, 9435.
One might expect that the degree of adduct dissociation in
the presence of colloidal metal would be significantly affected
by the dissociation energy of the individual adducts, with weaker
dative bonds being more prone to dissociation in the presence
of elemental metal. For example, the calculated dissociation
energies follow the trend: Me3P‚GaH3 (82.7 kJ/mol) < Me3N‚
GaH3 (102.1 kJ/mol) < Me3N‚AlH3 (129.2 kJ/mol) < Me3N‚
(35) (a) Cowley, A. H.; Mills, J. L. J. Am. Chem. Soc. 1969, 91, 2911. (b)
Toyota, S.; Futawaka, T.; Asakura, M.; Ikeda, H.; Oki, M. Organometallics
1998, 17, 4155. (c) In the case of Me3P‚BH3, we have recently found
evidence for this process in the model reaction with the Co(0) complex
Co2(CO)8 which affords first Me3P‚Ga[Co(CO)4]3 and subsequently Co2-
(CO)6(PMe3)2. Unpublished results.
(36) It should be noted that the preparation of Rhcolloid/[Oct4N]Cl (and the
analogous Ru colloids) involves the use of [Oct4N][BEt3H] as the reducing
agent. Thus, BEt3 is formed as a byproduct, which is fully removed upon
workup as indicated by the lack of a B(1s) peak in the XPS spectrum of
the colloidal product. See Figure 2.
(29) In our case, the parts per million difference between free and complexed
phosphine amounts to a ∆δ of ca. 48 ppm for PMe3 on Rh colloids. For
PPh3 on Pt and Cu colloids, ∆δ values of ca. 51-56 and 34 ppm have
been observed, respectively. See ref 19.
(37) Bo¨hm, V. P. W.; Brookhart, M. Angew. Chem., Int. Ed. 2001, 40, 4694.
9
5118 J. AM. CHEM. SOC. VOL. 127, NO. 14, 2005

Poisoning of Late Transition Metal Catalysts
A R T I C L E S
characteristic of metallic rhodium (cf. Rh(0) ca. 307.3 eV)38
(Figure S1, Supporting Information). It is likely that Rh and B
both form oxide layers when the treated colloids were exposed
to the atmosphere during preparation of the sample for XPS.
However, sputtering with Ar+ ions would remove the oxide layer
from Rh, resulting in a different binding energy. No change
was observed for the B peak, likely due to the fact that an
extremely thin film of boron may be formed, which could be
entirely converted to boric oxide upon exposure to air.
As to the nature of the passivating boron layer on the surface
of the Rh colloid, the formation of rhodium boride may be a
distinct possibility. For example, the structure of rhodium boride
(RhB) has been determined39 and metal borides are well-known
materials.40 This boride layer may be formed from a catalyzed
decomposition of “BH3”, similar to a reaction that has been
previously reported involving Co particles and B2H6 in which
elemental boron and H2 are formed.41 Another possibility might
involve the formation of a metal boryl layer (e.g. M-BH2)42
or a bridging structure with a residual terminal hydrogen (e.g.
M-B(H)-M). However, we were unable to confirm the
presence of B-H bonds on the colloids by IR spectroscopy.
Poisoning of Other Heterogeneous Transition Metal
Catalysts by Treatment with BH3‚THF. To determine the
scope of this poisoning behavior toward the catalytic dehydro-
coupling of Me2NH‚BH3, a variety of heterogeneous transition
metal catalysts were pretreated with BH3‚THF and the resulting
catalytic dehydrocoupling activities were then explored. For all
the metal catalysts tested, their activity toward the dehydro-
coupling of Me2NH‚BH3 was first verified by performing control
Figure 2. XPS spectra (B(1s) region) of Rhcolloid/[Oct4N]Cl before treatment
(dashed line) and after treatment with BH3‚THF (solid line).
The 11B NMR spectrum of the reaction mixture indicated the
presence of unreacted BH3‚THF and the formation of the
monochlorinated adduct BH2Cl‚THF (Bδ 4.3 ppm), which may
arise from reaction with the colloid stabilizing agent [Oct4N]-
Cl as it is the only source of chlorine. Upon subsequent treatment
of the solution with Me2NH‚BH3, no evidence for any catalytic
dehydrocoupling activity was observed by 11B NMR. Thus, it
was apparent that the BH3‚THF was reacting with the Rh
colloids to effectively act as a catalyst poison. This poisoning
cannot be attributed to a solvent effect in which the THF
molecules may block access to the active sites, as removal of
the volatiles after the BH3‚THF addition was still found to result
in an inactive catalyst when the dehydrocoupling reaction with
Me2NH‚BH3 was attempted in toluene. The colloidal catalyst
was examined by X-ray photoelectron spectroscopy (XPS)
before and after treatment with BH3‚THF. The XPS spectrum
of the colloids before treatment showed essentially no boron
content as indicated by the lack of a B(1s) peak (Figure 2,
dashed line). However, the colloids after treatment with BH3‚
THF showed a B(1s) peak at 193.4 eV (Figure 2, solid line) in
addition to the expected Rh(3d5/2) peak at 308.0 eV. Both of
these peaks are characteristic of oxide species as they fall into
the region typical for B2O3 (ca. 192-193.5 eV) and Rh2O3 (ca.
308.5-309 eV).38 Neglecting the other elements present (e.g.
C, Cl, O, and N), the relative atomic concentrations of Rh and
B were determined to be approximately in a ratio of 1 to 2.5,
respectively. After sputtering the surface with Ar+, it was found
that the relative atomic concentrations of Rh increased while
that of B decreased, giving a ratio of approximately 1 to 1. This
is indicative of a higher concentration of boron on the surface
of the colloids and suggests the formation of a passivated surface
layer. For example, if the Rh colloids simply catalyzed the
decomposition of BH3‚THF into bulk boron (e.g. B12), segre-
gated domains of Rh and B would result. Sputtering with Ar+
might then be expected to remove equal quantities of both Rh
and B from the surface of these domains, and thus the relative
atomic concentrations would not be affected to a great extent.
However, this was not observed as there was a substantial
decrease in the concentration of B from the surface. Upon
sputtering, the B(1s) peak was unchanged but the Rh(3d5/2) peak
was observed at slightly lower energy (307.6 eV), which is
experiments. Significantly, the stabilized colloids Rucolloid
/
[Oct4N]Cl, the supported species Rh/Al2O3 and Pd/C and Rh(0)
black all displayed a reaction with BH3‚THF (ca. 5 equiv) with
the elimination of H2 gas. Subsequent addition of Me2NH‚BH3
was found to result in the absence of catalytic dehydrocoupling
activity for all of the samples, as the expected cyclic dimer
[Me2N-BH2]2 was not observed by 11B NMR. In addition,
Rh(0) metal prepared using borane or borohydride reducing
agents also showed similar poisoning effects. For example, two
separate samples of Rh(0) metal were prepared from the reaction
of [Rh(1,5-cod)(µ-Cl)]2 with BH3‚THF and the reduction of
RhCl3 with Na[BH4]. The subsequent addition of Me2NH‚BH3
to each of these potential catalysts was found to result in no
observable dehydrocoupling activity (24 h, 25 °C), suggestive
of poisoning by the reduction byproduct species BH2Cl‚THF
or B2H6. It is likely that poisoning of these catalysts is similar
to that observed for Rhcolloid/[Oct4N]Cl, in which a passivated
boron surface layer was formed, thereby preventing subsequent
dehydrocoupling.
A series of fractional poisoning experiments involving Rh/
Al2O3 and BH3‚THF were also performed in order to determine
how much borane was required to completely passivate the
catalyst surface toward catalytic dehydrocoupling over 7 h at
(39) Aronsson, B.; Aselius, J.; Stenberg, E. Nature 1959, 183, 1318.
(40) Hoard, J. L.; Hughes, R. E. In The Chemistry of Boron and Its Compounds;
Muetterties, E. L., Ed.; Wiley: New York, 1967; p 25.
(41) Glavee, G. N.; Klabunde, K. J.; Sorensen, C. M.; Hadjipanayis, G. C. Inorg.
Chem. 1993, 32, 474.
(42) For examples involving reactions of B-H bonds at Rh centers, see: (a)
Baker, R. T.; Ovenall, D. W.; Harlow, R. L.; Westcott, S. A.; Taylor, N.
J.; Marder, T. B. Organometallics 1990, 9, 3028. (b) Burgess, K.; van der
Donk, W. A.; Westcott, S. A.; Marder, T. B.; Baker, R. T.; Calabrese, J.
C. J. Am. Chem. Soc. 1992, 114, 9350. For a review on metal-boryl
complexes see: (c) Braunschweig, H. Angew. Chem. Int. Ed. Engl. 1998,
37, 1786.
(38) Moulder, J. F.; Stickle, W. F.; Sobol, P. E.; Bomben, K. D. Handbook of
X-ray Photoelectron Spectroscopy; Perkin-Elmer: Eden Prairie, MN, 1992.
9
J. AM. CHEM. SOC. VOL. 127, NO. 14, 2005 5119

A R T I C L E S
Jaska et al.
catalysts.25 Again, reaction of the B-H bonds at the catalyst
surface may effectively poison the active sites and inhibit or
suppress the catalytic activity.43 As partial dehydrocoupling
activity was observed for the intermediate [H2B-MeNH]3, this
implies that the borazine [HB-NMe]3 is more effective as a
catalyst poison.
We also considered the possibility that the above boranes
might poison the active catalyst by adsorption to the surface,
rather than reacting at the surface to form a passivated layer.
This was investigated by treating Rhcolloid/[Oct4N]Cl with
i
(µ-NMe2)B2H5, [HB-NH]3, 9-BBN, [HB-NMe]3, or Pr2Nd
BH2 followed by removal of these and all other volatile
components of the reaction mixture. The results obtained
suggested that the poisoning mechanism involved reaction at
the surface, as the Rh colloids treated with (µ-NMe2)B2H5,
[HB-NH]3, or 9-BBN displayed no catalytic activity. The
dehydrocoupling results involving Me2NH‚BH3 are identical to
those obtained when the excess borane was left in solution (vide
Figure 3. Graph of percent conversion vs time for the catalytic dehydro-
coupling of Me2NH‚BH3 using Rh/Al2O3 without treatment ([) or after
treatment with 0.1 (9), 1 (2), 2 ()), 3 (0), 4 (4) or 5 (b) equiv of BH3‚
THF/(Rh atom).
25 °C. Thus, Rh/Al2O3 was treated with varying quantities of
BH3‚THF (0, 0.1, 1, 2, 3, 4, or 5 equiv/(Rh atom)), which was
followed by the addition of Me2NH‚BH3 and monitoring of the
subsequent dehydrocoupling reaction by 11B NMR. Figure 3
shows the resulting percent conversion versus time curves. These
studies demonstrate that for increasing amounts of BH3‚THF
added to the catalyst, there is a corresponding decrease in the
dehydrocoupling rate. Thus, it is evident that a significant excess
of borane (5 equiv) is required to completely poison the catalyst
surface toward dehydrocoupling activity under these conditions.
This implies that if a small amount of amine-borane adduct
dissociation occurred during dehydrocoupling, the generated free
BH3 might not affect the catalytic activity to a significant extent.
This result can then explain the previous observation that the
dehydrocoupling of Me2NH‚BH3 proceeds efficiently using
Me3N‚BH3 treated Rhcolloid/[Oct4N]Cl, despite there being ca.
6% free BH3 present in solution.
i
supra). For the case of [HB-NMe]3 and Pr2NdBH2, partial
dehydrocoupling (66 and 33%, respectively) was observed,
which might imply that the surface reaction is slow or that
adsorption to the surface is more significant for these species.
The formation of the borane byproduct BEt3 may also be
anticipated to affect the catalytic activity of metal colloids that
are prepared using “superhydride” reducing agents such as
M[BEt3H] (M ) Li, [R4N]). Indeed, treatment of Rhcolloid
/
[Oct4N]Cl with BEt3 followed by addition of Me2NH‚BH3 was
found to result in no observable dehydrocoupling. However,
when the volatiles were removed prior to adduct addition, partial
dehydrocoupling (35%) was observed, suggesting that this may
be a surface adsorption phenomenon in this particular case.
Poisoning of Rhcolloid/[Oct4N]Cl by Treatment with GaH3‚
OEt2. With the poisoning behavior of BH3‚THF and heteroge-
neous metal catalysts now established for the dehydrocoupling
of Me2NH‚BH3, it was of interest to examine if similar poisoning
behavior would occur with other types of group 13 hydrides
species. After BH3‚THF, the next two simplest congener group
13 hydrides are AlH3‚THF and GaH3‚THF. However, both alane
and gallane adducts containing oxygen donors such as THF and
Et2O tend to be unstable species.44 Nevertheless, modification
of an existing literature preparation45 allowed us to synthesize
GaH3‚OEt2. Thus, treatment of Rhcolloid/[Oct4N]Cl with GaH3‚
OEt2 (ca. 5 equiv) was observed to result in the immediate
release of H2 gas and the slow formation of Ga metal. Addition
of Me2NH‚BH3 to this treated catalyst was found to result in
no dehydrocoupling activity after 24 h, as only unreacted starting
material was observed in the 11B NMR spectrum. Again, it is
likely that the gallane reacts at the metal surface to form a
protective layer that prevents any catalytic dehydrocoupling
activity. The gallium metal observed in the reaction mixture
Poisoning of Rhcolloid/[Oct4N]Cl by Treatment with Vari-
ous Other Species Containing B-H Bonds. The use of other
borane species containing B-H bonds was also investigated to
determine if they react with heterogeneous metal surfaces to
inhibit the catalytic dehydrocoupling of Me2NH‚BH3. For
example, treatment of Rhcolloid/[Oct4N]Cl with 5 equiv of
(µ-NMe2)B2H5, [HB-NH]3, [HB-NMe]3 or 9-BBN (BBN )
borabicyclo[3.3.1]nonane) followed by addition of Me2NH‚BH3
was found to result in no catalytic dehydrocoupling activity after
20 h. The boranes iPr2NdBH2 or [H2B-MeNH]3 were observed
to only partially poison the catalyst, with 74 and 61% conversion
after the same time period, respectively. The catalyst poisoning
may result from reaction with, or adsorption to, the surface Rh
atoms and the formation of a protective layer. For example,
the boranes (µ-NMe2)B2H5, [HB-NH]3, and [H2B-MeNH]3
were all observed to form H2 when reacted with the Rh colloids,
i
(43) An alternative explanation is that the formation of the dehydrocoupling
product (e.g. [HB-NH]3) may actually block the active catalytic sites by
binding to the surface Rh atoms, in a fashion similar to the well-known
case of (CO)3Cr([RB-NR′]3) in which the borazine ring is bound to the
metal center through the nitrogen atoms. See: Lagowski, J. J. Coord. Chem.
ReV. 1977, 22, 185. However, dissociation of the bound borazine may
require higher reaction temperatures in order to generate an active site where
dehydrocoupling could occur. An equilibrium between bound and unbound
borazine on the Rh surface might help to suppress the catalysis, resulting
in much longer dehydrocoupling times for NH3‚BH3.
(44) Unfortunately, attempts to prepare stable solutions of AlH3‚OEt2 in our
laboratory have been unsuccessful to date, and thus the reactivity of this
species with heterogeneous metal catalysts could not be investigated.
(45) Shriver, D. F.; Parry, R. W.; Greenwood, N. N.; Storr, A.; Wallbridge, M.
G. H. Inorg. Chem. 1963, 2, 867.
whereas [HB-NMe]3, 9-BBN, and Pr2NdBH2 showed no
evidence of H2 formation. This suggests that the former group
may react with the metal surface to poison catalytic activity,
while the latter group may simply adsorb to the surface and
block the substrate from accessing the active sites. The results
involving [HB-NH]3, [HB-NMe]3, and [H2B-MeNH]3 are
significant as they may explain why the catalytic dehydro-
coupling of NH3‚BH3 or MeNH2‚BH3 requires longer reaction
times (ca. 3 days) and higher temperatures (45 °C) compared
to that of Me2NH‚BH3 (ca. 8 h, 25 °C) using identical (pre)-
9
5120 J. AM. CHEM. SOC. VOL. 127, NO. 14, 2005

Poisoning of Late Transition Metal Catalysts
A R T I C L E S
may arise from the thermal (25 °C) decomposition of GaH3‚
OEt2 over time or from a metal-catalyzed decomposition process
as the gallane-etherate is prone to decompose in the presence
of trace impurities.46
Influence of Borane Treatment on the Activity of Het-
erogeneous Catalysts for Other Transformations. (i) Cata-
lytic Hydrogenation with Borane Treated, Heterogeneous
Rh Catalysts. It has been previously demonstrated that colloids
or nanoclusters and oxide-supported metals are very efficient
heterogeneous catalysts for the hydrogenation of olefins.2 It was
of interest to determine if treatment of these catalysts with BH3‚
THF would lead to surface passivating boron layers, which could
poison the catalyst surface and inhibit hydrogenation reactivity.
This type of poisoning behavior might be of profound signifi-
cance, as one of the most common routes to heterogeneous metal
colloids or nanoclusters used in catalysis involves the reduction
of transition metal salts using borohydride reducing agents such
as Na[BH4], Li[BH4], or [R4N][BEt3H] (R ) long-chain alkyl
group).5e,f,7a,8b,c,e,10c,47c,48 If the byproducts from these reduction
reactions (e.g. B2H6 or BEt3) react with the generated metal
catalyst to form a passivated layer, a significant reduction in
the observed catalytic activity might result. Thus, the hydro-
genation of cyclohexene to cyclohexane (eq 6) under H2 was
monitored using untreated and BH3‚THF treated (5 equiv) Rh/
Al2O3 as a catalyst at 25 °C. However, it was found that the
relative rate of hydrogenation was comparable for trials involv-
ing both untreated and treated catalysts after 1 h of reaction
time (Figure 4). Thus, it can be concluded that treatment of
Rh/Al2O3 with BH3‚THF does not poison the surface of the
catalyst toward catalytic hydrogenation, in stark contrast to the
case of catalytic dehydrocoupling of Me2NH‚BH3.49
Figure 4. Graph of percent conversion vs time for the catalytic hydrogena-
tion of cyclohexene under H2 using Rh/Al2O3 without treatment ([, dashed
line) or after treatment with 5 equiv of BH3‚THF/(Rh atom) (b, solid line).
coupling reactivity. This was tested by using Pd(0) metal formed
by three different reducing agents (BH3‚THF, EtOH, and H2)
followed by an evaluation of the catalytic activity in a Suzuki
reaction involving phenylboronic acid and iodobenzene at 100
°C (eq 7). For example, treatment of H2PdCl4 with BH3‚THF
resulted in the formation of a black precipitate of Pd(0) metal.
Use of this precipitate as a catalyst was found to result in the
formation of biphenyl in an isolated yield of 45%. However,
similar yields were observed for Pd(0) catalysts formed via
EtOH and H2 reduction methods (44 and 52%, respectively).
This implies that there is no poisoning effect by BH3‚THF for
this particular catalytic reaction.
Summary
Metal-assisted dissociation reactions of group 13 - group
15 adducts have been shown to generate both well-known
heterogeneous catalyst poisons, such as free phosphines (e.g.
PMe3), and also the group 13 hydride species EH3 (E ) B, Al,
Ga) via the formation of a passivating boron/boride layer.
Analogous poisoning effects were noted in the cases of
heterogeneous Ru and Pd catalysts. These results explain the
need for higher temperatures and longer reaction times for the
synthesis of certain borazines such as [HB-NH]3 from NH3‚
BH3 by catalytic dehydrocoupling as this product also exhibits
a significant poisoning effect. The results also strongly imply
that catalyst poisoning/surface passivation effects need to be
carefully considered in catalytic reactions in which heteroge-
neous transition metal catalysts are prepared using the commonly
applied borane or borohydride reduction procedures. Although
no analogous poisoning effects were observed for the hydro-
genation and Suzuki coupling reactions we studied using
colloidal metal treated with, or prepared using BH3‚THF, the
formation of boron layers and the potential for surface passi-
vation is likely of direct relevance for many other reactions and
in materials science applications.51,52
(ii) Suzuki Coupling with Borane Treated, Heterogeneous
Pd Catalysts. Pd colloids and nanoclusters have been previously
shown to be excellent catalysts in C-C bond forming reactions
such as Suzuki cross-coupling.10e,50 It was of interest to
determine if treatment of these types of catalysts with BH3‚
THF could poison the catalyst surface and inhibit the cross-
(46) Greenwood, N. N. In 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.
(47) (a) Bo¨nnemann, H.; Brijoux, W.; Brinkmann, R.; Dinjus, E.; Joussen, T.;
Korall, B. Angew. Chem., Int. Ed. Eng. 1991, 30, 1312. (b) Bo¨nnemann,
H.; Braun, G.; Brijoux, W.; Brinkmann, R.; Schulze Tilling, A.; Seevogel,
K.; Siepen, K. J. Organomet. Chem. 1996, 520, 143. (c) Shulz, J.; Roucoux,
A.; Patin, H. Chem. Eur. J. 2000, 6, 618.
(48) (a) Mayer, A. B. R.; Mark, J. E. J. Polym. Sci., Part A: Polym. Chem.
1997, 35, 3151. (b) Chechik, V.; Crooks, R. M. J. Am. Chem. Soc. 2000,
122, 1243. (c) Ooe, M.; Murata, M.; Mizugaki, T.; Ebitani, K.; Kaneda,
K. Nano Lett. 2002, 2, 999. (d) Kidambi, S.; Dai, J.; Li, J.; Bruening, M.
L. J. Am. Chem. Soc. 2004, 126, 2658.
(49) It is possible that the generated boron-rich passivating layer is subsequently
removed during the hydrogenation reaction. However, no change in the
boron content was observed by XPS for samples of the BH3‚THF treated
Rhcolloid/[Oct4N]Cl catalyst both before and after the hydrogenation reaction.
In addition, treatment of the BH3‚THF poisoned catalyst with a hydrogen
atmosphere did not alter the reactivity of the catalyst; the catalyst was found
to remain inactive toward dehydrocoupling. These two results indicate that
substantial leaching of boron from the catalyst surface did not occur and
that the surface coverage is incomplete, despite the poisoning of the sites
responsible for dehydrocoupling. We thank a reviewer for suggesting these
experiments.
Experimental Section
General Procedures 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
(50) (a) Li, Y.; Hong, X. M.; Collard, D. M.; El-Sayed, M. A. Org. Lett. 2000,
2, 2385. (b) Kogan, V.; Aizenshtat, Z.; Popovitz-Biro, R.; Neumann, R.
Org. Lett. 2002, 4, 3529.
9
J. AM. CHEM. SOC. VOL. 127, NO. 14, 2005 5121

A R T I C L E S
Jaska et al.
1
dry nitrogen unless otherwise specified. THF and Et2O were dried over
Na/benzophenone and distilled prior to use. Toluene was dried via the
Grubb’s method.53 ACS grade EtOH, pentane, and CH3CN were used
without further purification. PhB(OH)2, PhI, sodium acetate, NEt3,
NMe3, Me2NH (2.0 M in THF) BH3‚THF (1.0 M in THF), PMe3 (1.0
M in toluene), Rh/Al2O3 (5 wt % Rh), Pd/C (10 wt % Pd), 9-BBN (0.5
M in THF), BEt3 (1.0 M in THF), GaCl3, Me3N‚BH3 (Aldrich), Rh(0)
black (Strem), and PdCl2 (Pressure) were purchased and used as
received. Pyridine (Aldrich) was dried over Na and distilled prior to
use. Me2NH‚BH3 (Strem) was purified by sublimation at 25 °C. Li-
[AlH4] (Aldrich) was purified by dissolving a sample in Et2O, followed
by filtration and solvent removal to afford a white powder. Me3P‚
For Me3N‚GaH3. H NMR (C6D6): δ ) 5.00 (Br, GaH3), 1.84 (s,
NMe3). 13C{1H} NMR (C6D6): δ ) 49.8. (i) 0% dissociation. (ii) 100%
dissociation. (iii) 89% unreacted Me2NH‚BH3, 9% Me3N‚BH3 (via
amine exchange) and 2% [Me2N-BH2]2.
1
For Me3P‚BH3. H NMR (C6D6): δ ) 1.32 (qd, JHB ) 95 Hz, JHP
) 15 Hz), 0.64 (d, JHP ) 10.5 Hz). 11B{1H} NMR (C6D6): δ ) -36.9
(d, JBP ) 62 Hz). 13C{1H} NMR (C6D6): δ ) 12.7 (d, JCP ) 37 Hz).
31P{1H} NMR (C6D6): δ ) -2.15 (q, JPB ) 59 Hz). (i) 0% dissociation.
(ii) 7% dissociation into PMe3. (iii) 92% unreacted Me2NH‚BH3, 8%
[Me2N-BH2]2 and Me3P‚BH3.
1
For Me3P‚GaH3. H NMR (C6D6): δ ) 4.30 (br, GaH3), 0.59 (d,
JHP ) 7.5 Hz). 13C{1H} NMR (C6D6): δ ) 11.7. 31P{1H} NMR
(C6D6): δ ) -40.2 (s). (i) 0% dissociation. (ii) 100% dissociation into
PMe3. (iii) 100% unreacted Me2NH‚BH3.
GaH3,54 Me3N‚AlH3,55 Me3P‚BH3,56 Me3N‚GaH3,57
Mcolloid/[Oct4N]Cl
(M ) Rh, Ru),4b (µ-NMe2)B2H5,58 [HB-NH]3,59 iPr2NdBH2,25b [MeN-
BH]3,25b [MeNH-BH2]3,60 and [Rh(1,5-cod)(µ-Cl)]261 were synthesized
by literature procedures.
For NMe3. H NMR (C6D6): δ ) 2.06 (s). 13C{1H} NMR (C6D6):
1
δ ) 48.1.
Equipment. NMR spectra were recorded on a Varian Gemini 300
MHz spectrometer. Chemical shifts are reported relative to residual
protonated solvent peaks (1H, 13C) or external BF3‚Et2O (11B) or H3-
PO4 (31P) references. NMR spectra were obtained at 300 MHz (1H),
96 MHz (11B), 75 MHz (13C) or 121 MHz (31P). Silicon(100) substrates
were purchased from Wafer World Inc. and cleaned by successive
sonication treatments in CH2Cl2, 2-propanol, piranha solution (H2O2/
H2SO4, 1:3 vol. %; caution: extremely corrosiVe!), and deionized water
followed by drying in air. XPS samples were solution deposited onto
a clean silicon substrate and analyzed on a PHI 5500 ESCA system. A
Mg KR source with a photon energy of 1253.6 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 depth profiling analysis, an Ar+
ion beam of 3 keV was used for 15 min and the spectra were reacquired.
Treatment of Rhcolloid/[Oct4N]Cl with Me3E‚XH3 (E ) N, P; X
) B, Al, Ga). A typical reaction was performed as follows: A solution
of Me3E‚XH3 (ca. 0.03 g) in C6D6 (0.75 mL) was stirred at 25 °C for
For PMe3. H NMR (C6D6): δ ) 0.78 (s). 13C{1H} NMR (C6D6):
1
δ ) 15.3. 31P{1H} NMR (C6D6): δ ) -61 (s).
PMe3 Ligation to Rhcolloid/[Oct4N]Cl; Preparation of {Rhcolloid
-
(PMe3)/[Oct4N]Cl}). (a) To a solution of Rhcolloid/[Oct4N]Cl (0.012 g,
ca. 0.058 mmol Rh) in THF (1 mL), a solution of PMe3 in toluene
(0.12 mL, 0.12 mmol, ca. 2 equiv/(Rh atom)) was added, and the
mixture was stirred for 7 h at 25 °C. 31P{1H} NMR (THF): δ ) -10
(m, {Rhcolloid(PMe3)/[Oct4N]Cl}), -61 (s, PMe3). Treatment of this
solution with Me2NH‚BH3 (ca. 0.1 g) was found to result in no
dehydrocoupling activity (24 h, 25 °C), indicative of PMe3 poisoning.
(b) To a solution of Rhcolloid/[Oct4N]Cl (0.002 g, ca. 0.01 mmol of
Rh) in THF (1 mL), a solution of Me3P‚BH3 (0.016 g, 0.18 mmol) in
THF (1 mL) was added, and the mixture was stirred for 18 h at 25 °C.
31P{1H} NMR (THF): δ ) -0.7 (q, Me3P‚BH3), -10 (m, {Rhcolloid
(PMe3)/[Oct4N]Cl}), -61 (s, PMe3).
Attempted Poisoning of Rhcolloid/[Oct4N]Cl by the Nitrogen
Donors Me2NH, NEt3, and Pyridine. A typical reaction was performed
as follows: To a solution of Rhcolloid/[Oct4N]Cl (0.005 g, ca. 0.02 mmol
of Rh) in THF (1 mL), a solution of Me2NH in THF (1.5 mL, 3.0
mmol, ca. 150 equiv/(Rh atom)) was added at 25 °C. The mixture was
stirred for 6 h, and then Me2NH‚BH3 (0.109 g, 1.85 mmol) was added
as a solid. After 18 h, only unreacted Me2NH‚BH3 was observed in
the 11B NMR spectrum. Similar results were obtained using NEt3 and
pyridine.
-
1
18 h. The presence or absence of free EMe3 was determined by H
NMR (i). Solid Rhcolloid/[Oct4N]Cl (ca. 0.002 g) was then added, and
the mixture was stirred for 24 h. Again, the presence or absence of
1
free EMe3 was determined by H NMR (ii). Solid Me2NH‚BH3 (ca.
0.05 g) was then added, and the mixture was stirred for 24 h. Integration
B
of the product ([Me2N-BH2]2, δ 5 ppm) and reactant (Me2NH‚BH3,
Bδ -13 ppm) resonances in the 11B NMR spectrum was used to
determine the extent of the catalytic dehydrocoupling reaction (iii).
For Me3N‚BH3. 1H NMR (C6D6): δ ) 2.4 (br q, JHB ) 98 Hz, BH3),
2.01 (s, NMe3). 11B{1H} NMR (C6D6): δ ) -7.7 (s). 13C{1H} NMR
(C6D6): δ ) 53.9. (i) 0% dissociation. (ii) 6% dissociation into NMe3.
(iii) 95% [Me2N-BH2]2, 5% unreacted Me2NH‚BH3, and Me3N‚BH3.
Reaction of Rhcolloid/[Oct4N]Cl with BH3‚THF. To a solution of
Rhcolloid/[Oct4N]Cl (0.045 g, ca. 0.22 mmol of Rh) in THF (1 mL), a
solution of BH3‚THF (1.0 mL, 1.0 mmol, ca. 5 equiv/(Rh atom)) was
added at 25 °C. An immediate reaction was observed with the formation
of bubbles and the release of a gas. 11B NMR (THF): δ ) 4.3 (t, JBH
) 139 Hz; lit. 4.6, JBH ) 135 Hz;62 BH2Cl‚THF, ca. 11%), 0.6 (q,
BH3‚THF, ca. 85%), -30.0 (s, unidentified product, ca. 4%). The gas
was identified as being H2 by 1H NMR (δ 4.46). After 2 h, the reaction
mixture was evaporated to dryness, and the black residue was washed
with toluene (2 × 10 mL). The toluene washings were found to contain
[Oct4N]Cl by 1H NMR (Hδ ) 3.95 (m), 1.59 (m), 1.34 (m), 0.93 (m)).
The residue was dried in Vacuo to yield a black powder (0.048 g) that
was only slightly soluble in THF. XPS: Rh (308.0 eV, relative atomic
concentration ) 28.3%), B (193.4 eV, relative atomic concentration
) 71.7%). After Ar+ sputter: Rh (307.6 eV, relative atomic concentra-
tion ) 42.7%), B (193.4, relative atomic concentration ) 57.3%).
Attempted Catalytic Dehydrocoupling of Me2NH‚BH3 Using
BH3‚THF Treated Rhcolloid/[Oct4N]Cl. To a solution of Rhcolloid/[Oct4N]-
Cl (0.008 g, ca. 0.04 mmol Rh) in THF (1 mL), a solution of BH3‚
THF (0.18 mL, 0.18 mmol) in THF was added at 25 °C. After 2 h,
solid Me2NH‚BH3 (0.087 g) was added and the mixture was left to stir
at 25 °C. After 18 h, only unreacted Me2NH‚BH3 was observed in the
11B NMR spectrum. In a repeat trial, the volatiles were removed after
the addition of BH3‚THF. The addition of solid Me2NH‚BH3 was found
to result in no dehydrocoupling activity.
1
For Me3N‚AlH3. H NMR (C6D6): δ ) 4.01 (br, AlH3), 1.84 (s,
NMe3). 13C{1H} NMR (C6D6): δ ) 47.7. (i) 0% dissociation. (ii) 31%
dissociation into NMe3. (iii) 100% unreacted Me2NH‚BH3.
(51) One example may be metal-catalyzed hydroboration reactions, which also
involve the use of reactants with B-H bonds. See ref 52. Interestingly,
the metal-catalyzed hydroboration of 1-octene has been recently reported
using Rh(0) metal formed in situ by the reaction of RhCl3 with BH3‚THF.
Morrill, T. C.; D’Souza, C. A. Organometallics 2003, 22, 1626. However,
to date, the exact nature of the active catalyst was not determined, and our
results suggest that a homogeneous catalyst may need to be considered as
the actual species responsible for the catalysis.
(52) (a) Burgess, K.; Ohlmeyer, M. J. Chem. ReV. 1991, 91, 1179. (b) Beletskaya,
I.; Pelter, A. Tetrahedron 1997, 53, 4957. (c) Crudden, C. M.; Edwards,
D. Eur. J. Org. Chem. 2003, 4695.
(53) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.; Timmers,
F. J. Organometallics 1996, 15, 1518.
(54) Greenwood, N. N.; Ross, E. J. F.; Storr, A. J. Chem. Soc. 1965, 1400.
(55) Ruff, J. K. Inorg. Synth. 1967, 9, 30.
(56) Schmidbaur, H.; Weiss, E.; Mu¨ller, G. Synth. React. Inorg. Met.-Org. Chem.
1985, 15, 401.
(57) Shriver, D. F.; Shirk, A. E. Inorg. Synth. 1977, 17, 42.
(58) Keller, P. C. Synth. Inorg. Met.-Org. Chem. 1973, 3, 307.
(59) Wideman, T.; Sneddon, L. G. Inorg. Chem. 1995, 34, 1002.
(60) Narula, C. K.; Janik, J. F.; Duesler, E. N.; Paine, R. T.; Schaeffer, R. Inorg.
Chem. 1986, 25, 3346.
(61) Giordano, G.; Crabtree, R. H. Inorg. Synth. 1979, 19, 218.
(62) Dodds, A. R.; Kodama, G. Inorg. Chem. 1979, 18, 1465.
9
5122 J. AM. CHEM. SOC. VOL. 127, NO. 14, 2005

Poisoning of Late Transition Metal Catalysts
A R T I C L E S
Treatment of Heterogeneous Transition Metal Catalysts with
BH3‚THF and Attempted Catalytic Dehydrocoupling of Me2NH‚
BH3. (a) A typical reaction was performed as follows: To a solution
of Rucolloid/[Oct4N]Cl (0.008 g, ca. 0.06 mmol of Ru) in THF (1 mL),
a solution of BH3‚THF (0.30 mL, 0.30 mmol, ca. 5 equiv/(Ru atom))
was added at 25 °C. An immediate reaction was observed with the
formation of bubbles and the release of a gas (H2, Hδ 4.46). After 4 h,
Me2NH‚BH3 (0.050 g, 0.85 mmol) was added and the mixture was
stirred at 25 °C. After ca. 20 h, only unreacted Me2NH‚BH3 was
observed in the 11B NMR spectrum. Other heterogeneous catalysts tested
were Rh/Al2O3, Pd/C and Rh(0) black which all showed similar
poisoning behavior.
(b) The above experiments were repeated by treating Rhcolloid/[Oct4N]-
Cl with the appropriate borane, followed by removal of volatiles. The
residue was redissolved in THF, Me2NH‚BH3 was added and the
mixture stirred for 24 h. For (µ-NMe2)B2H5, [HB-NH]3, and 9-BBN,
no dehydrocoupling was observed. For [HB-NMe]3, iPr2NdBH2, and
BEt3, only 66, 33, and 35% dehydrocoupling were observed, respec-
tively.
Preparation of GaH3‚OEt2. GaH3‚OEt2 was prepared by a modified
literature procedure.45 A solution of GaCl3 (0.167 g, 0.948 mmol) in
Et2O (5 mL) was added dropwise to a solution of Li[AlH4] (0.109 g,
2.87 mmol, 3 equiv) in Et2O (10 mL) at 0 °C. The formation of a
white precipitate was observed. The mixture was stirred for 2 h at 0
°C and then stored overnight at -20 °C in order to completely
decompose the proposed intermediate mixed hydride [Ga(AlH4)3].46 The
solution was quickly filtered cold to afford a colorless solution of GaH3‚
OEt2 (ca. 0.059 M based on 100% yield) that was stored at -55 °C to
prevent decomposition. An NMR sample was prepared by removing
(b) Blank reactions of Rucolloid/[Oct4N]Cl, Rh/Al2O3, Pd/C, and Rh(0)
black were performed to verify their catalytic activity toward Me2NH‚
BH3 dehydrocoupling. In all cases, complete dehydrocoupling was
observed after 24 h at 25 °C.
Preparation of Rh(0) Metal by Reaction of [Rh(1,5-cod)(µ-
Cl)]2with BH3‚THF and Attempted Catalytic Dehydrocoupling of
Me2NH‚BH3. To a solution of [Rh(1,5-cod)(µ-Cl)]2 (0.020 g, 0.041
mmol) in THF (1 mL), a solution of BH3‚THF in THF (0.41 mL, 0.041
mmol) was added at 25 °C. The solution immediately turned black,
and the evolution of a gas and the formation of a black precipitate
were both observed. 11B NMR (THF): δ ) 7.2 (d, JBH ) 164 Hz; lit.
7.0, JBH ) 162 Hz; BHCl2‚THF, ca. 15%),63 4.3 (BH2Cl‚THF, ca.
85%).62 Solid Me2NH‚BH3 (0.054 g, 0.92 mmol) was added, and the
mixture was stirred at 25 °C overnight. After 24 h, the 11B NMR
spectrum of the reaction mixture indicated only the presence of
unreacted Me2NH‚BH3.
Preparation of Rh(0) Metal by Reaction of RhCl3 with Na[BH4]
and Attempted Catalytic Dehydrocoupling of Me2NH‚BH3. To a
solution of RhCl3 (0.050 g, 0.24 mmol) in THF (2 mL), a solution of
Na[BH4] (0.027 g, 0.73 mmol) in THF (2 mL) was added at 25 °C.
The solution turned black over ca. 5 min, and the evolution of a gas
and the formation of a black precipitate were both observed. After 24
h, Me2NH‚BH3 (ca. 0.1 g) was added and stirred. The 11B NMR
spectrum indicated no dehydrocoupling activity after 24 h at 25 °C.
1
the solvent at -78 °C. H NMR (C6D6): δ ) 4.01 (br s, GaH3), 3.30
(s, OCH2), 1.05 (s, CH3). 13C NMR (C6D6): δ ) 66.5 (OCH2), 15.6
(CH3). Note: The 1H NMR spectrum did not give acceptable integration
values, as a small amount of decomposition into Ga(s), H2, and Et2O
was consistently observed.
Treatment of Rhcolloid/[Oct4N]Cl with GaH3‚OEt2 and Attempted
Catalytic Dehydrocoupling of Me2NH‚BH3. To a solution of Rhcolloid
/
[Oct4N]Cl (0.005 g, ca. 0.03 mmol Rh) in THF (1 mL), a solution of
GaH3‚OEt2 (2.5 mL, ca. 0.15 mmol, ca. 5 equiv/(Rh atom)) in Et2O
was added at 25 °C. The formation of bubbles and the release of a gas
(H2, Hδ 4.46) was immediately observed. The slow formation of a gray
powder was also observed, which was attributed to gallium metal. The
mixture was stirred for 2 h; then Me2NH‚BH3 (0.098 g, 1.7 mmol)
was added. Upon stirring at 25 °C for 24 h, only unreacted Me2NH‚
BH3 was observed in the 11B NMR spectrum.
Treatment of Rh/Al2O3 with BH3‚THF and Evaluation of the
Catalytic Hydrogenation Activity. To a suspension of Rh/Al2O3 (ca.
0.05 g) in THF (2 mL) in a 10 mL vial, a solution of BH3‚THF (5
equiv/(Rh atom)) was added via syringe. The solution was stirred for
2 h at 25 °C, and the volatiles were removed. A solution of cyclohexene
(ca. 0.15 g) in C6D6 (1 mL) was added to the black residue, and the
vial was capped with a septum. The solution was frozen using liquid
nitrogen, and the headspace of the vial was evacuated. The vial was
refilled with pressurized H2 gas (ca. 0.1 atm) to commence the
hydrogenation reaction. The reaction was monitored by periodically
removing a small aliquot of solution (50 µL) and running the 1H NMR
spectrum in C6D6 via integration of the product and reactant signals.
A control experiment was performed in a similar manner, except that
the catalyst was not treated with BH3‚THF.
Treatment of Rh/Al2O3 with BH3‚THF and Effect on the Rate
of Catalytic Dehydrocoupling of Me2NH‚BH3. To a suspension of
Rh/Al2O3 (ca. 0.05 g) in toluene (2 mL), a solution of BH3‚THF (0,
0.1, 1, 2, 3, 4, or 5 equiv) was added via microsyringe. The solution
was stirred for 18 h at 25 °C, and then Me2NH‚BH3 (ca. 0.15 g) was
added as a solid. The dehydrocoupling reaction was monitored by
removing a small aliquot of solution, diluting with toluene, and
obtaining the 11B NMR spectrum. The percent conversion was
calculated from the integration of the product (Bδ 5 ppm) and reactant
(Bδ -13 ppm) resonances.
Treatment of Rhcolloid/[Oct4N]Cl with Various Boranes and
Attempted Catalytic Dehydrocoupling of Me2NH‚BH3. (a) A typical
reaction was performed as follows: To a solution of Rhcolloid/[Oct4N]-
Cl (0.008 g, ca. 0.06 mmol of Rh) in THF (1 mL), a solution of
(µ-NMe2)B2H5 (0.025 g, 0.35 mmol, ca. 5 equiv/(Rh atom)) in THF (1
mL) was added at 25 °C. When the reaction was repeated in C6D6, the
presence of H2 was detected in solution by 1H NMR. The mixture was
stirred 4 h; then Me2NH‚BH3 (0.053 g, 0.90 mmol) was added as a
solid. After ca. 20 h, only unreacted Me2NH‚BH3 was observed in the
11B NMR spectrum.
1
For cyclohexene. H NMR (C6D6): δ ) 5.69 (s, CHdCH), 1.90
(m, CH2), 1.51 (m, CH2). 13C{1H} NMR (C6D6): δ ) 127.7 (CHd
CH), 25.8 (CH2), 23.3 (CH2).
1
For cyclohexane. H NMR (C6D6): δ ) 1.40 (s). 13C{1H} NMR
(C6D6): δ ) 27.6.
Evaluation of Boron Content by XPS for Poisoned Rhcolloid
/
[Oct4N]Cl Catalyst after Hydrogenation. To a solution of Rhcolloid
/
[Oct4N]Cl (0.082 g, ca. 0.26 mmol Rh) in THF (2 mL), a solution of
BH3‚THF (1.3 mL, 1.3 mmol, ca. 5 equiv/(Rh atom)) was added and
stirred overnight at 25 °C. The solution was split equally into two
fractions. The first fraction was pumped to dryness, washed with toluene
(5 mL), and dried in Vacuo. The black residue was suspended in toluene
and deposited onto a clean Si substrate, and the volatiles were removed
under vacuum for 16 h. The second fraction was stirred under a
hydrogen atmosphere for 3 h, followed by the same workup procedure
for the above blank sample. The XPS spectra for both fractions showed
the presence of boron on the surface of the catalyst.
For [HB-NH]3: H2 observed; no dehydrocoupling of Me2NH‚BH3.
For [MeN-BH]3: no H2 observed; no dehydrocoupling of Me2NH‚
BH3.
For [MeNH-BH2]3: H2 observed; 61% dehydrocoupling of Me2-
NH‚BH3.
For 9-BBN: no H2 observed; no dehydrocoupling of Me2NH‚BH3.
Evaluation of Catalytic Dehydrocoupling Activity for Poisoned
Rhcolloid/[Oct4N]Cl Catalyst after Hydrogenation. To a solution of
i
For Pr2NdBH2: no H2 observed; 74% dehydrocoupling of Me2-
NH‚BH3.
For BEt3: no dehydrocoupling of Me2NH‚BH3.
(63) Brown, H. C.; Ravindran, N. J. Am. Chem. Soc. 1976, 98, 1798.
9
J. AM. CHEM. SOC. VOL. 127, NO. 14, 2005 5123

A R T I C L E S
Jaska et al.
Rhcolloid/[Oct4N]Cl (0.004 g, ca. 0.013 mmol Rh) in THF (2 mL), a
solution of BH3‚THF (0.65 mL, 0.65 mmol, ca. 5 equiv/(Rh atom))
was added. The volatiles were removed after 1 h; the black residue
redissolved in toluene (2 mL) and stirred under a hydrogen atmosphere
for 3 h. Solid Me2NH‚BH3 (0.120 g, 2.04 mmol) was then added, and
the reaction was stirred under N2 at 25 °C. After 16 h, the 11B NMR
spectrum showed that only 4% of the dehydrocoupling product [Me2N-
BH2]2 was present.
Preparation of Pd(0) Metal via BH3‚THF, EtOH, or H2 Reduc-
tion and Evaluation of Suzuki Coupling Activity. A solution of
H2PdCl4 (0.002 M) was prepared by the addition of 0.2 M HCl (6 mL)
to a solution of PdCl2 (0.090 g) in water (100 mL), followed by dilution
to 250 mL in air.9e The solvent was removed from a 15 mL portion of
this to give a red solid, which was redissolved in THF (15 mL). To
this solution under N2, a solution of BH3‚THF in THF (0.45 mL) was
added dropwise via syringe at 25 °C. A black precipitate of Pd(0) was
formed immediately. After 1 h, the solution was decanted and the
precipitate washed with THF (10 mL) and dried under vacuum. A
solution of PhB(OH)2 (6 mmol), PhI (2 mmol), and sodium acetate (7
mmol) in CH3CN/H2O (3:1, 35 mL) was added to the Pd(0) residue.
The mixture was sonicated for 1 min to give a fine dispersion of Pd(0),
and the mixture was refluxed at 100 °C for 16 h under air. The product
was extracted with pentane and dried over MgSO4 and the solvent
removed to give biphenyl as a white solid in 45% yield (crude).
For the reduction of H2PdCl4 with EtOH, EtOH (10 mL) was added
to the THF solution of H2PdCl4 and the mixture was refluxed at 100
°C for 16 h. A similar procedure was followed for the Suzuki coupling,
with an isolated yield of 44%. For the reduction of H2PdCl4 with H2,
the THF solution of H2PdCl4 was stirred under a H2 atmosphere for 1
h at 25 °C. A similar procedure was followed for the Suzuki coupling,
with an isolated yield of 52%.
1
For Ph-Ph. H NMR (CDCl3): δ ) 7.71 (m), 7.54 (m), 7.45 (m).
13C{1H} NMR (CDCl3): δ ) 141.5, 129.0, 127.5, 127.4.
Acknowledgment. C.A.J. is grateful for a Natural Sciences
and Engineering Research Council of Canada (NSERC) scholar-
ship (2002-2004), T.J.C. is grateful for a University of Toronto
Fellowship, S.B.C. is grateful for an NSERC Postdoctoral
Fellowship, and I.M. thanks NSERC for a Discovery Grant and
the Canadian Government for a Canada Research Chair.
Supporting Information Available: XPS spectrum (Rh
region) of Rhcolloid/[Oct4N]Cl after treatment with BH3‚THF
(PDF). This material is available free of charge via the Internet
at http://pubs.acs.org.
JA0447412
9
5124 J. AM. CHEM. SOC. VOL. 127, NO. 14, 2005