Are heterogeneous catalysts precursors to homogeneous catalysts? [21]

Full text of DOI:10.1021/ja016877v

J. Am. Chem. Soc. 2001, 123, 10139-10140
10139
Supporting Information).9 The product was isolated by crystal-
lization from the reaction stream after removal of the carbon by
filtration and typically contained 200 ppm of Pd as determined
by ICP-AES analysis.
Analysis of the reaction mixture after filtration of the carbon
demonstrated >125 ppm of Pd in the process stream, i.e. 25% of
the initial catalyst charge. The addition of mercury limited the
reaction to two turnovers providing circumstantial evidence
regarding the phase of the active catalyst.10,11 To unambiguously
assign the role of a homogeneous catalyst we envisaged applica-
tion of a three-phase test.12
Are Heterogeneous Catalysts Precursors to
Homogeneous Catalysts?
Ian W. Davies,* Louis Matty, David L. Hughes, and
Paul J. Reider
Department of Process Research, Merck & Co., Inc.
P.O. Box 2000, Rahway, New Jersey 07065-0900
ReceiVed August 17, 2001
ReVised Manuscript ReceiVed August 29, 2001
Advances in the design of solid supports and the accompanying
developments in analysis have been driven by the combinatorial
revolution and a wide range of supports are now available. Since
the carbonylation reaction is performed under basic conditions
we chose to use a Novasyn TGR resin allowing the attachment
of the substrate with an amide bond. The design features involving
the use of an aryl iodide in addition to the presence of a para-
electron-withdrawing group used as attachment ensure that the
oxidative addition with a homogeneous catalyst would be
extremely facile.
Homogeneous and heterogeneous catalysts offer their own
distinct advantages.1 Heterogeneous catalysts have an advantage
that at the end of reaction the catalyst can be removed by simple
filtration. In principle the product is un-contaminated with a
transition metal or ligand and allows the catalyst to be recycled
into the next reaction.2 While the distinction between homoge-
neous and heterogeneous catalysis seems well-defined, in many
cases there may be leaching of the transition metal into solution.
In these instances the question that always remains is whether
the catalytic activity resides with the leached metal. In other
situations, it is unclear whether a “release and capture” of the
transition metal catalyst has occurred. We report a simple
unambiguous test to determine the presence of a homogeneous
catalyst and demonstrate its application in three prototypical
reactions. Not only does this method clarify the phase of the
catalytically active species it also allows additional mechanistic
information to be obtained regarding the system.
Our initial studies were prompted by the carbonylation of 1 to
give the carbomethoxycyclopentenone 2 (eq 1). Although this
Control experiments with resin 3 under our standard conditions
(1 mol % Pd/C, 60 °C, 80 psig CO, 1 M DMA, 2 equiv of Bu3N,
5 equiv of MeOH, 10 h) resulted in quantitative recovery of 5
following cleavage with TFA, i.e. there was no catalytically active
species present in solution. However, addition of resin 3 to the
carbonylation reaction in the presence of 1 led to a 95% yield of
2 and a quantitative formation of ester 6 following TFA cleavage
of the resin-bound product. Iodobenzene was equally effective
at generating a homogeneous catalyst and was itself converted
to methylbenzoate in 97% yield.
transformation worked quite well with the conventional homo-
geneous catalyst Pd(PPh3)2Cl2,3 a heterogeneous catalyst was
sought for this reaction to simplify product isolation. In many of
our programs, the fate of the catalyst is especially important during
processing since the acceptable level of transition metal residuals
is highly regulated in pharmaceutical products. Pd/C has been
used for a number of palladium-catalyzed reactions.4,5 After
parallel catalyst screening, the optimal conditions were identified
at small scale.6 Using these conditions at 1 mol scale (1 mol %
Pd/C, 60 °C, 80 psig CO, 1 M DMA, 2 equiv Bu3N, 5 equiv
MeOH, 10 h) led to the formation of 2 in 98-99% yield.7 There
was no trace of reduction, dimerization, or amide formation8 and
mass balance was accounted for by two low-level impurities (see
The requirement for the organohalide in the generation of an
active catalyst suggests that upon oxidative addition, desorption
of the Pd(II) species, e.g. Ph-Pd-I, occurs which enters a
conventional solution phase catalytic cycle generating a soluble
(1) Gates, B. C. In Catalytic Chemistry; Wiley: New York, 1992.
(2) For the sake of simplicity, in this article we specifically address transition
metal-catalyzed processes and we use palladium in particular to illustrate the
approach. The principles and observations discussed should be broadly
applicable.
(9) All new compounds were fully characterized. The identity of known
compounds were confirmed by comparison (LCMS) with authentic samples.
(10) For suppression of a heterogeneous Pt-catalyzed reaction by Hg(0)
see: 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. Interpetation of the results from the Hg(0)
test should be viewed with caution: Lin, Y.; Finke, R. G. Inorg. Chem. 1994,
33, 4891. The presence of Hg(0) may lead to suppression of catalysis via
amalgamation, which inhibits the proposed oxidative addition/desorption
mechanism.
(11) For application of dibenzo[a,e]cyclooctetraene as a potent poison for
homogeneous catalysis see: Anton, D. R.; Crabtree, R. H. Organometallics
1983, 2, 855. For application of polymer substrates as a test in hydrogenation
reactions see: Collman, J. P.; Kosydar, K. M.; Bressan, M.; Lamanna, W.;
Garrett, T. J. Am. Chem. Soc. 1984, 106, 2569.
(3) Schoenberg, A.; Heck, R. F. J. Org. Chem. 1974, 39, 3327.
(4) (a) Suzuki-Miyaura reaction: LeBlond, C. R.; Andrews, A. T.; Sun,
Y.; Sowa, J. R. Org. Lett. 2001, 3, 1555. (b) Amidocarbonylation of imines:
Beller, M.; Moradi, W. A.; Eckert, M.; Neumann, H. Tetrahedron Lett. 1999,
40, 4523. (c) Heck reaction: Augustine, R. A.; O’Leary, S. T. J. Mol. Catal.
1992, 72, 229. (d) Stille reaction: Liebeskind, L. S.; Penn-Cabrera, E. Org.
Synth. 1999, 77, 135.
(5) For related Ni/C catalyzed reactions see: Lipshutz, B. H.; Ueda, H.
Angew. Chem., Int. Ed. 2000, 39, 4492.
(6) Parallel reaction screening using the Argonaut Endeavor.
(7) The iodide reacted with similar efficiency where as the chloride was
unreactive even at higher temperatures.
(12) Rebek J.; Gavina, F. J. Am. Chem. Soc. 1974, 96, 7112. Rebek, J.;
Brown, D.; Zimmerman, S. J. Am. Chem. Soc. 1975, 97, 454. Rebek, J.
Tetrahedron 1979, 35, 723.
(8) Schnyder A.; Beller, M.; Mehltretter, G.; Nsenda, T.; Studer, M.;
Indolese, A. F. J. Org. Chem. 2001, 66, 4311.
10.1021/ja016877v CCC: $20.00 © 2001 American Chemical Society
Published on Web 09/25/2001

10140 J. Am. Chem. Soc., Vol. 123, No. 41, 2001
Communications to the Editor
Pd(0) catalyst.13 Carbon monoxide is not required for the
generation of the active homogeneous catalyst. The homogeneous
catalyst can be prepared by aging the reaction mixture under
nitrogen followed by filtration of the carbon prior to the
introduction of CO. The homogeneous catalyst prepared in this
manner has a minimum turnover number (ton) of 850 and turnover
frequency (tof) of 250/h for the conversion of 1 to 2.14
A further illustration is provided by an investigation of the Heck
reaction. Recently, indirect methods have implicated the involve-
ment of a homogeneous catalyst involving Pd/alumina, Pd/C, or
polystyrene-supported palladium.15 In the case of Pd/C it has been
suggested that a homogeneous mechanism may be in operation
but is redeposited on the support.16 Heck reaction of bromoace-
tophenone 7 and n-butylacrylate (NBA) promoted by Pd/alumina
(2 mol %) proceeded to give cinnamate 8 in 98% yield (eq 2).
Iodobenzene and the iodoamide 5 coupled with similar efficiency.
the presence of resin 3 (0.5 mol %) and 10 (3 mol %) led to the
clean formation of biphenyl. The biphenylamide 11 was obtained
in 85% yield following TFA cleavage of the resin. The reaction
stream contained 2.5 ppm Pd by ICP-AES analysis. Perhaps the
most unsatisfactory shortcoming of the test is that the method
offers unambiguous evidence for the existence of an intermediate
but only inferences can be drawn concerning its structure. In this
particular example further studies will be required to elucidate if
all of the reactivity is associated with the homogeneous catalyst.
However, the resin may serve as a convenient catalyst precursor
for “release” and acts as a scavenger for palladium at the end of
reactions“capture”.21
In conclusion, we have unequivocally demonstrated in two
reactions involving palladium on heterogeneous supports that the
reaction proceeds via a homogeneous catalyst. In the case of the
Suzuki-Miyaura coupling with a polymeric phosphine we have
demonstrated the existence of a competent homogeneous catalyst.
These examples clearly challenge the assumption that all hetero-
geneous catalysts operate in a well-defined regimen. The use of
polymer supports that allow for designed, sequential, and multistep
organic synthesis is a rapidly expanding arena.22 Our observations
emphasize that caution needs to be exercised in planning multistep
synthesis involving transition metal resins since this endeavor
relies upon the effective isolation of the reactive centers. The
three-phase test should find renewed application in a wide range
of reactions involving supported catalysts.23
Performing the Heck reaction in the presence of resin 3 led to
quantitative formation of cinnamate 9 following TFA cleavage.
Control experiments indicated that in the absence of the aryl halide
low conversions of 3 were observed and the rate of reaction
increased with the sodium acetate charge (2-5 equiv). Conversion
and rate of reaction of 3 were greatly enhanced by the presence
of both aryl halide 7 and sodium acetate, i.e. they act coopera-
tively. Surface oxidative addition occurs and desorption of the
palladium(II) species is presumably accelerated by the formation
17
of ArPdBr(OAc)n or ArPd(OAc)n,18 which are free to enter a
conventional solution-phase reaction.19 Our results firmly establish
the involvement of a homogeneous catalyst.
Finally, a demonstration of the utility of this approach in a
system that has direct implications for library generation involved
the use of phosphine complexes. Polymer-supported phosphines
have now been used in a growing number of transformations and
a large number are now commercially available. The palladium
complex 10, prepared from triphenylphosphine-NovaGel and
palladium π-allyl chloride dimer, has successfully been applied
to Suzuki-Miyaura cross-coupling and arylation of allylic
acetates.20
Acknowledgment. We would like to thank Professor Barry Trost for
his insightful suggestions regarding application of a three-phase test and
his continued interest in our programs.
Supporting Information Available: Experimental procedures and
characterization data for all new compounds, conversion vs CO uptake
data as a function of pressure and temperature, a comparison of Pd/C vs
Pd(PPh3)2Cl2 as a function of temperature, and data to support the
calculation of ton and tof (PDF). This material is available free of charge
via the Internet at http://pubs.acs.org.
Performing the Suzuki-Miyaura coupling reaction of iodo-
benzene and phenylboronic acid (1.5 M KOH, 20 h, 25 °C) in
(13) PhCl was unreactive even at higher temperatures. Similarly the chloro
analogue of 1 was unreactive. These observations provide additional circum-
stantial evidence that the desorption mechanism is driven by surface oxidative
or electron transfer.
JA016877V
(21) In the generally accepted mechanism, an 18-electron palladium(0)
complex PdLn must undergo a ligand loss to PdLn-2 prior to generation of
a competent catalyst. In the case of a resin bound monodentate phosphine
(X-P) the complex Pd(X-P)Ln-1 may lose X-P based on steric or electronic
arguments (e.g. trans-effect) releasing a free PdLn-1 into the solution phase
as a precursor to a conventional catalytic cycle. See ref 17 for an alternative
mechanism involving anionic palladium complexes.
(22) Ley, S. V.; Baxendale, I. R.; Bream, R. N.; Jackson, P. S.; Leach, A.
G.; Longbottom, D. A.; Nesi, M.; Scott. J. S.; Storer, I. R.; Taylor, S. J. J.
Chem. Soc., Perkin Trans 1 2000, 3815. For the first example of a sequential
reaction see: Pittman, C. U.; Wuu, S. K.; Jacobsen, J. J. Catal. 1976, 44, 87.
(23) For example, Raney-Ni-tartaric acid methylacetoacetate reduction in
the presence of bromide ion, see: Harada, T.; Yamamoto, M.; Onaka, S.;
Imaida, M.; Tai, A.; Izumi, Y. Bull. Chem. Soc. Jpn. 1981, 54, 2323. Pt-
chinchona ketoester reduction, see: LeBlond, C.; Wang, J.; Liu, A.; Andrews,
A. T.; Sun, Y.-K. J. Am. Chem. Soc. 1999, 121, 4920.
(14) The catalyst was prepared by using 10 equiv of bromide 1 vs 5 wt %
Pd/C. See Supporting Information for details.
(15) Biffis, A.; Zecca, M.; Basato, M. Eur. J. Inorg. Chem. 2001, 1131.
(16) Zhao, F.; Bhanage, B. M.; Shirai, M.; Arai, M. Chem. Eur. 2000, 6,
843. Zhao, F.; Shirai, M.; Arai, M. J. Mol. Catal. 2000, 154, 39. Jayasaree,
J.; Seayad, A.; Chaudhari, R. V. Chem. Commun. 1999, 1067.
(17) For a discussion of the formation of anionic palladium(II) complexes
in the Heck reaction see: Amatore, C.; Jutand, A. Acc. Chem. Res. 2000, 33,
314.
(18) Nilsson, P.; Larhed, M.; Hallberg, A. J. Am. Chem. Soc. 2001, 123,
8217.
(19) For an intriguing mechanistic proposal for the role of acetate ion and
anionic palladium complexes in the Heck reaction see: Shaw, B. L. New J.
Chem. 1998, 77.
(20) Uozumi, Y.; Danjo, H.; Hayashi, T. J. Org. Chem. 1999, 64, 3384.