Synthesis and properties of a stable, cationic, rhodium Lewis-acid catalyst for hydrosilation, Mukaiyama aldol and cyclopropanation reactions

Article abstract of DOI:10.1039/b007815h

The remarkably stable cationic, three-coordinate, 14-electron rhodium complex 1 has been synthesized, isolated and used as a catalyst for hydrosilation, Mukaiyama aldol and cyclopropanation reactions.

Full text of DOI:10.1039/b007815h

Synthesis and properties of a stable, cationic, rhodium Lewis-acid catalyst for
hydrosilation, Mukaiyama aldol and cyclopropanation reactions†
Eric L. Dias, Maurice Brookhart* and Peter S. White
Department of Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-3290, USA.
E-mail: mbrookhart@unc.edu
Received (in Irvine, CA, USA) 22nd September 2000, Accepted 9th January 2001
First published as an Advance Article on the web 14th February 2001
The remarkably stable cationic, three-coordinate, 14-elec-
tron rhodium complex 1 has been synthesized, isolated and
used as a catalyst for hydrosilation, Mukaiyama aldol and
cyclopropanation reactions.
solution of 1, generated in situ from 2 and NaB(ArA)4, results in
quantitative formation of the rhodium(III) methyl iodide com-
plex 3 after 15–20 min [eqn. (2)]. While we had previously
Of the several different catalyst types available for organic
reactions, the most diverse are probably the Lewis acids.
Ranging from a simple proton to boranes or main group and
transition metal complexes, Lewis-acid catalysts now allow for
a variety of transformations to be accomplished efficiently and,
in many cases, selectively.1 The requirements for a Lewis-acid
catalyst are straightforward: (1) the complex should be
electrophilic, and (2) there should be a vacant coordination site.
Here, we report the synthesis of the surprisingly stable, cationic,
three-coordinate, 14-electron rhodium compound 1.2 Com-
plexes of 1 with aldehydes and ketones are also described, in
addition to preliminary studies employing 1 in catalytic
reactions.
(2)
synthesized 3 from the ethylene-bound adduct of 1,5 this has
proven to be a much more simple and convenient route.
The use of chlorinated solvents, whch are very weakly
coordinating at best, allows the Lewis-acid characteristics of 1
to be fully exploited. For example, addition of 1 equiv. of either
p-tolualdehyde or acetophenone to a dichloromethane solution
of 1 produces the corresponding adducts 4a and 4b
[eqn. (3)], which were isolated as stable solids. From the crystal
(3)
Addition of NaB(ArA)4 [ArA = 3,5-bis(trifluoromethyl)phe-
nyl] to a solution of the rhodium( ) chloride compound 2 in
I
dichloromethane results in quantitative formation of the
cationic complex 1 after 1–2 h at room temp., as evidenced by
the change in solution color from dark green to orange [eqn. (1)].
structure of the aldehyde complex 4a (Fig. 1),6 it can be seen
that the aldehyde coordinates in an h fashion. It is also apparent
1
that the imine moieties of the ligand are pulled close to the metal
center, such that the attached aryl groups create a ‘wedge’ in
which the aldehyde binds. In fact, when hydrogens are included
in the structure, the intramolecular distance of 2.70 Å between
the aldehydic hydrogen and the centroid of one of the aryl rings
(1)
After removing the NaCl by cannula filtration, evaporation of
the solvent yields 1 as a stable, dark orange solid which usually
contains ca. 0.25 equiv. of CH2Cl2, as determined by 1H NMR
spectroscopy in d8-THF. Although 1 can be stored under an
inert atmosphere for an indefinite period of time, it is extremely
susceptible to hydration, such that it is often more convenient to
generate 1 in situ immediately prior to use in any subsequent
experiments.
The stability of 1 towards the oxidative addition of dichloro-
methane, a reaction common to rhodium( ) complexes of this
I
type,3 is of particular significance.4 Complex 1 was found to be
stable in dichloromethane solution indefinitely at room tem-
perature, and even at 40 °C. In addition, 1 can be heated to 70 °C
in chlorobenzene without any decomposition. On the other
hand, addition of 1 equiv. of methyl iodide to a dichloromethane
Fig. 1 X-Ray crystal structure of 4a. Thermal ellipsoids are drawn at the
50% probability level. Selected bond lengths (Å) and angles (°): Rh(1)–O(7)
2.0728(18), Rh(1)–N(3) 2.0124(23), Rh(1)–N(11) 1.8854(20), O(7)–C(8)
1.227(3); Rh(1)–O(7)–C(8) 129.44(19), O(7)–Rh(1)–N(11) 174.81(9),
N(6)–C(4)–C(12) 113.80(25), C(1)–N(3)–C(22) 2104.8(6), N(3)–Rh(1)–
O(7)–C(8) 41.5(3).
† Electronic supplementary information (ESI) available: experimental
information, van’t Hoff plot, temperature and equilibrium data, 1H NMR
spectrum of 1 and crystallographic data. See http://www.rsc.org/suppdata/
cc/b0/b007815h/
DOI: 10.1039/b007815h
Chem. Commun., 2001, 423–424
This journal is © The Royal Society of Chemistry 2001
423

View Article Online
1
suggests a van der Waals interaction with the aromatic p-cloud.
Likewise, the distance of 3.06 Å between the ortho-protons on
the p-tolualdehyde and the other ring indicates a potentially
the product distrubution was determined by H NMR to be
85+9+6 respectively.10
Surprisingly, these cyclopropanations do not proceed using
trimethylsilyldiazomethane as a carbene source. Because of the
steric limitations imposed by the ligand, incorporation of the
TMS group appears to make the diazoalkyl fragment too bulky
to coordinate to the rhodium center through the a-carbon, and
thus ultimately produce a reactive carbene. In fact, the N-bound
adduct of TMS–diazomethane is actually stable enough to be
isolated, and we are currently exploring the unique reactivity
that results.11
In summary, we have demonstrated that 1 is an easily
synthesized, isolable compound which shows remarkable
stability for a three-coordinate, 14-electron complex. As
expected, 1 is a fairly potent Lewis acid, and the steric
environment imposed by the ligand creates a binding pocket
that must accommodate potential substrates.12 It is this steric
environment, however, which is responsible for the stability of
1
similar interaction. This geometry is verified by the H NMR
spectrum of 4a. Resonances for the aldehydic and ortho protons
appear at 7.45 and 6.68 ppm, respectively, shifted upfield due to
the shielding provided by the aryl groups.
When the acetophenone complex 4b is redissolved in
dichloromethane, an equilibrium between the bound and free
acetophenone is established. 1H NMR spectroscopy reveals that
at room temp., ca. 25% of the ketone is not bound to the
rhodium center, unlike the aldehyde complex 4a in which all of
the aldehyde remains complexed. At lower temperatures, the
equilibrium lies further towards complexed acetophenone as
expected, and a van’t Hoff plot (see ESI†) provides an estimated
value of DH0
=
216 ± 2 kJ mol21 for acetophenone
binding.
The availability of complexes 4a and 4b led us to screen
transformations involving aldehydes and ketones, utilizing 1 as
a Lewis acid catalyst. Using 2.5 mol% 1, the hydrosilation of
benzaldehyde and acetophenone with triethylsilane could be
effected [eqn. (4)].1a,7 While the reaction of triethylsilane with
1 towards oxidative addition reactions common to Rh( )
I
complexes—in particular, reaction with chlorinated solvents is
suppressed, allowing them to be used as the preferred solvents
for synthetic and catalytic applications. In future papers, we will
describe the integral role that steric effects play in stabilizing
complexes of this type, and report on further studies utilizing 1
as a Lewis acid catalyst.
(4)
Acknowledgment is made to the National Institutes of Health
(GM-29838) for support of this work.
benzaldehyde proceeds to 95% conversion to 5a after 2 h at
70 °C in chlorobenzene, the reaction with acetophenone to
produce 5b required ca. 17 h to reach 90% conversion under
similar conditions. Given the lack of reactivity of triethylsilane
with 1 and the steric environment imposed by the ligand, it is
believed that this reaction in fact proceeds by a Lewis acid-
catalyzed mechanism.
Notes and references
1 (a) Lewis Acid Reagents: A Practical Approach, ed. H. Yamamoto,
Oxford University Press, New York, 1999; (b) Lewis Acids and
Selectivity in Organic Synthesis ed. M. Santelli and J.-M. Pons, CRC
Press, Boca Raton, FL, 1996.
2 A complex employing a similar PNP ligand has been implicated (C.
Hahn, M. Spiegler, E. Herdtweck and R. Tabue, Eur. J. Inorg. Chem.,
1999, 435), although it was not observed directly and its stability is not
evident.
3 H. F. Haarman, J. M. Ernsting, M. Kranenburg, H. Kooijman, N.
Veldman, A. L. Spek, P. W. N. M. vanLeeuwen and K. Vrieze,
Organometallics, 1997, 16, 887; K. J. Bradd, B. T. Heaton, C. Jacob,
J. T. Sampanthar and A. Steiner, J. Chem. Soc., Dalton Trans., 1999,
1109; H. Nishiyama, M. Horihata, T. Hirai, S. Wakamatsu and K. Itoh,
Organometallics, 1991, 10, 2706.
The Mukaiyama aldol condensation of a trimethylsilyl enol
ether with benzaldehyde can also be catalyzed by 1 [eqn. (5)].8
(5)
4 In a separate study, we have determined that steric effects are
responsible for the stability of 2 towards the oxidative addition of
dichloromethane, and this stability extends to 1 as well. In the series of
complexes analogous to 2 where only the ortho substituents are varied,
we have found that the oxidative addition of dichloromethane occurs
when R = H or Me.
The reaction of benzaldehyde with 1-trimethylsilyloxy-
1-phenyl ethylene in chlorobenzene proceeds to 88% conver-
sion to 6a after heating at 65 °C for 24 h; however, when
dichloromethane is used as the solvent, the reaction only
proceeds to ca. 50% conversion after 24 h at room tem-
perature.
Finally, since carbenes are also two-electron donors, complex
1 seemed to be ideally suited to catalyze cyclopropanation
reactions.9 The reaction of ethyl diazoacetate with a-olefins
such as hex-1-ene and oct-1-ene proceeds smoothly; however,
carbene insertion into the vinylic C–H bonds was also observed
to some extent [eqn. (6)]. Upon monitoring the reaction of
5 E. L. Dias, M. Brookhart and P. S. White, Organometallics, 2000, 19,
4995.
6 Crystal data. BC73F24H63N3ORh, M = 1567.98, triclinic, space group
¯
P1, a = 13.3839(6), b = 16.3962(7), c = 16.8506(7) Å, a = 80.160(1),
b = 79.835(1), g = 86.890(1)°, U = 3585.1(3) Å3, Z = 2, m(Mo-Ka)
= 0.35 mm21, 45 784 reflections measured, 17 165 unique (Rint
=
0.018) which were used in all calculations. The final wR(F2) was 0.056
(all data). CCDC 182/1892. See http://www.rsc.org/suppdata/cc/b0/
b007815h/ for crystallographic files in .cif format.
7 For examples of Lewis-acid catalyzed hydrosilation, see; D. J. Parks,
J. M. Blackwell and W. M. Piers, J. Org. Chem., 2000, 65, 3090; D. J.
Parks and W. E. Piers, J. Am. Chem. Soc., 1996, 118, 9440; Y. S. Song,
B. R. Yoo, G. H. Lee and I. N. Jung, Organometallics, 1999, 18,
3109.
8 See: T. Mukaiyama, Org. React., 1982, 28, 203 and references therein.
9 For reviews of transition metal-catalyzed cyclopropanations, see: M. P.
Doyle and D. C. Forbes, Chem. Rev., 1998, 98, 911; M. P. Doyle,
Acc.Chem. Res., 1986, 19, 348; G. Maas, Top. Curr. Chem., 1987, 137,
75.
(6)
10 It was observed in NMR experiments that the product distribution does
not change over the course of these reactions, indicating that the
products 8 and 9 are not formed by acid-assisted opening of the
cyclopropanes 7.
11 E. L. Dias, M. Brookhart and P. S. White, J. Am. Chem. Soc., in
press.
12 For a recent review of Lewis-acid catalysts with designed binding
pockets, see: H. Yamamoto and S. Saito, Pure Appl. Chem., 199, 71, 239
and references therein.
1
hex-1-ene and ethyl diazoacetate by H NMR spectroscopy, it
was determined that in addition to forming the cyclopropane 7a,
the C–H insertion product 8a appears initially as well, and
slowly isomerizes to the a,b-unsaturated ester 9a in what is
likely a rhodium-catalyzed process. For oct-1-ene, the com-
bined products 7b, 8b and 9b were isolated in 95% yield, and
424
Chem. Commun., 2001, 423–424