Unusual in situ ligand modification to generate a catalyst for room temperature aromatic C-O bond formation [8]

Full text of DOI:10.1021/ja002543e

10718
J. Am. Chem. Soc. 2000, 122, 10718-10719
Unusual in Situ Ligand Modification to Generate a
Catalyst for Room Temperature Aromatic C-O
Bond Formation
Quinetta Shelby, Noriyasu Kataoka, Grace Mann, and
John Hartwig*
Department of Chemistry
Yale UniVersity, P.O. Box 208107
New HaVen, Connecticut 06520-8107
Figure 1. Decay of ArBr and appearance of ArOAr′ vs time for reaction
of 2-bromotoluene (0.08 M) with NaOC6H4-4-OMe catalyzed by 5% Pd-
(dba)2 and 7.5% PFc(t-Bu)2 at 70 °C (left) and catalyzed by 5% Pd(dba)2
and Ph5FeP(t-Bu)3 at 40 °C (right).
ReceiVed July 14, 2000
The lifetime of a catalyst is generally controlled by its
decomposition pathways, such as ligand degradation. Generally,
one seeks to identify these decomposition pathways and to then
prevent them. We report an unusual example of the opposite
scanario: a surprising in situ structural change that transforms a
phosphine-ligated, transition-metal complex displaying low cata-
lytic activity into another system exhibiting high activity.
Identification of the modified catalyst and independent synthesis
of it has led to a series of room-temperature couplings of aryl
bromides with phenoxides, alkoxides, and siloxides, including
cyclizations to form oxygenated heterocycles. Our results em-
phasize that many factors underlie apparent catalyst structure-
reactivity relationships, including the potential to form unexpected
complexes displaying high activity.
Mild, aromatic carbon-oxygen bond formation is a difficult
transformation. For reactions of unactivated aryl halides, direct,
uncatalyzed substitutions and copper-mediated couplings typically
require temperatures of 100 °C or greater.1-3 Diazotization and
displacement with oxygen nucleophiles is generally limited to
phenol synthesis and uses stoichiometric amounts of copper in
its mildest form.4 We recently reported the first palladium catalysts
for the formation of diaryl and alkyl aryl ethers from unactivated
aryl halides.5 Yet, our system, as well as that reported recently
by Buchwald,6 required temperatures similar to those for copper-
mediated processes.1-3,7,8
Scheme 1
crossover experiment in eq 1. Thermolysis of the two arylpalla-
dium phenoxides would produce two ether products if dimer
cleavage prior to reductive elimination was irreversible, but it
would produce four diaryl ethers if dimer cleavage was rapid and
reversible. This experiment produced nearly equal amounts of
the four ethers, suggesting that dimer cleavage was reversible,
and that another explanation for the discrepancy was required.
We felt that a deep-seated understanding of how catalyst
structure affects reactivity could improve this process. Two results
suggested that the system we recently studied was more complex
than expected and that it would serve as an illustrative example
of how in situ changes in structure can affect catalyst activity.
First, reactions of aryl halides with sodium phenoxides catalyzed
by complexes generated from Pd(dba)2 and PFc(t-Bu)2 occurred
in high yields, but stoichiometric reductive elimination from the
presumed arylpalladium phenoxide intermediate occurred in much
lower yields (Scheme 1).5 Second, careful monitoring of the
catalytic reactions (Figure 1) showed that aryl halide was
consumed initially without formation of diaryl ether and that the
time dependence for formation of this product was sigmoidal.
These results suggested the presence of an induction period.
Several mechanistic hypotheses were tested to explain one or
both of these observations. First, we considered whether cleavage
of the isolated dimer was slower than direct decomposition to
form products other than diaryl ether, a hypothesis that would
explain the discrepancy in yields. Thus, we conducted the
Second, we evaluated whether the active catalyst was a species
other than that generated by simple ligation of PFc(P-t-Bu)2 to
the metal center. We evaluated possible ligand modifications by
heating palladium acetate with 20 equiv of ligand relative to
palladium and sodium tert-butoxide in chlorobenzene at 110 °C.
During this experiment with large amounts of phosphine, we could
analyze for changes in the ligand structure. We observed the
remarkable catalytic perphenylation of the unsubstituted cyclo-
pentadienyl group (eq 2).9 This material was formed in essentially
quantitative yields, and was isolated in 60-70% yields. This result
suggested that FcP(t-Bu)2 underwent arylation in the catalytic
system to provide the ligand of the active catalyst.
Several additional results indicated that arylated ferrocenyl
ligands were components of the active catalyst. For example, the
catalytic reaction of excess ligand with PhCl using phenoxide as
base provided several arylated ligands, including Ph5FcP(t-Bu)2.
(1) Bacon, R. G. R.; Rennison, S. C. J. Chem. Soc. (C) 1969, 312-315.
(2) Marcoux, J. F.; Doye, S.; Buchwald, S. L. J. Am. Chem. Soc. 1997,
119, 10539-10540.
(3) Kalinin, A. V.; Bower, J. F.; Riebel, P.; Snieckus, V. J. Org. Chem.
1999, 64, 2986-2987.
1
In addition, H and 31P NMR spectra of reactions of PhBr with
(4) March, J. AdVanced Organic Chemistry; John Wiley and Sons: New
York, 1985; p 601.
(7) Boger, D. L.; Yohannes, D. J. Org. Chem. 1991, 56, 1763.
(8) Fagan, P. J.; Hauptman, E.; Shapiro, R.; Casalnuovo, A. J. Am. Chem.
Soc. 2000, 122, 5043-5051.
(9) Miura, M.; Pivsaart, S.; Dyker, G.; Heiermann, J.; Satoh, T.; Nomura,
M. Chem. Commun. 1998, 1889-1890.
(5) Mann, G.; Incarvito, C.; Rheingold, A. L.; Hartwig, J. F. J. Am. Chem.
Soc. 1999, 121, 3224-3225.
(6) Aranyos, A.; Old, D. W.; Kiyomori, A.; Wolfe, J. P.; Sadighi, J. P.;
Buchwald, S. L. J. Am. Chem. Soc. 1999, 121, 4369-4378.
10.1021/ja002543e CCC: $19.00 © 2000 American Chemical Society
Published on Web 10/18/2000

Communications to the Editor
J. Am. Chem. Soc., Vol. 122, No. 43, 2000 10719
Table 1. Aromatic C-O Bond Formation Catalyzed by 5 mol %
a
Pd(dba)2/Ph5FcP(t-Bu)2
NaO-t-Bu using a more conventional 1.1:1 ratio of FcP(t-Bu)2 to
Pd(dba)2 as catalyst showed the decay of FcP(t-Bu)2 and growth
of Ph5FcP(t-Bu)2 and its Pd(0) complex {Pd[Ph5FcP(t-Bu)2]2. This
complex was identified by the formation of [Pd(PCy3)2] and
additional free Ph5FcP(t-Bu)2 after treatment of the reaction
solution with PCy3 and by independently generating the same
species by addition of ligand to Pd(dba)2 at room temperature or
CpPd(allyl) at 50 °C. Finally, reactions catalyzed by a combination
of Pd(dba)2 and Ph5FcP(t-Bu)2 as ligand occurred more rapidly
than those employing PFc(t-Bu)2, occurred in higher yield, and
occurred without a large induction period (Figure 1). Reaction
of 2-MeC6H4Br with NaOC6H4-p-OMe at 80 °C gave complete
reaction after 1 h, and reaction at room temperature gave
quantitative yields after 70 h.
If complexes containing Ar5FcP(t-Bu)2 as ligand are the true
catalysts in the formation of diaryl ethers, then reaction yields
for reductive elimination of diaryl ether should be higher when
this ligand is bound to the arylpalladium phenoxide fragment.
Due to the steric properties of this ligand, we have not been able
to isolate arylpalladium halide or phenoxide complexes containing
this ligand. However, addition of a 20-fold excess of Ph5FcP(t-
Bu)2 to FcP(t-Bu)2-ligated phenoxide 1 led to the reductive
elimination of diaryl ether with rates that were faster and with
yields that were higher than in the absence of this ligand. Reaction
at 70 °C for 1 h occurred in 66% yield (eq 3). Reactions conducted
with a smaller excess of P(Ph5Fc)(t-Bu)2 gave lower yields and
slower rates. Endothermic ligand exchange to generate the Ph5-
FcP(t-Bu)2 analogue of 1 and rapid reductive elimination from
this complex accounts for this observation.
a Reactions conducted in toluene solvent with 0.5 mmol aryl halide
substrate and 1.2 equiv of alkoxide or siloxide in 2 mL of toluene.
Isolated yields are an average of at least two runs.
room temperature with these ligands. Although we have not
explored in detail the activity of catalysts generated from Ph4-
FcP(t-Bu)2, preliminary data have suggested that these complexes
are much less reactive than those generated from Ph5FcP(t-Bu)2.
The most common means to conduct palladium-catalyzed
coupling reactions is to mix a catalyst precursor with a phosphine
ligand. Generally, coordination of the selected or designed ligand
occurs to generate the active catalyst, while P-C bond
cleavage,11-14 ligand oxidation, phosphorus quaternization, or
P-O bond hydrolysis deactivates the catalyst. In our case, it
appears that intramolecular ligand metalation and coupling with
aryl halide substrate leads to a pentaarylferrocenyl ligand.
Identification and independent synthesis of the pentaphenyl
version of this ligand has revealed a ligand structural class that
creates catalysts for mild, aromatic carbon-oxygen bond forma-
tion.
The use of isolated Ph5FcP(t-Bu)2 in combination with a
palladium catalyst precursor may then provide faster rates and
higher yields than previous catalysts for this process. Indeed, we
found that a combination of Ph5FcP(t-Bu)2 and Pd(dba)2 catalyzes
the formation of several types of aromatic carbon-oxygen bonds
at room temperature. Table 1 summarizes this chemistry. As stated
above, reaction of 2-MeC6H4Br with NaOC6H4-p-OMe occurred
at room temperature in 70 h. Reaction of NaO-t-Bu with a variety
of electron-neutral or electron-poor aryl bromides also occurred
at room temperature. Deactivated substrates such as 4-bromoani-
sole required heating but gave the protected phenol as product in
good yield. Similar chemistry occurred with the siloxide NaOSi-
(t-Bu)Me2 to give TBS-protected phenol products. Again, reac-
tions with electron-neutral or electron-poor substrates occurred
in good yield. The most deactivated substrates gave low yields.
Intramolecular formation of aryl alkyl ethers also occurred under
these conditions. Entries 15-19 show examples of cyclizations
to form oxygen heterocycles rapidly at room temperature.
Importantly, this ligand system allows for the reaction of substrates
that bear â-hydrogens. The reactions in entries 17 and 19 occur
in much higher yield than observed previously.10 In all cases, we
compared qualitatively the rates of these reactions to those
catalyzed by Pd(dba)2 and FcP(t-Bu)2 or by 2-methyl, 2-(di-tert-
butylphosphino)-1,1-biphenyl;6 little or no reaction occurred at
Modification of this ligand structure by altering the steric and
electronic properties of the Cp-bound aryl group will be inves-
tigated in the future to improve catalyst activity further. Moreover,
the stability of the current catalysts and the mild conditions for
the process should allow for future kinetic studies on this process
and for future synthetic studies on related couplings.
Acknowledgment. We thank the NIH (R29-GM55382) for support
of this work. N.K. thanks Ajinomoto Co., Inc. for support. We also thank
Felix Goodson for his initial observation of Cp arylation in a different
system that led us to consider this process with PFc(t-Bu)2.
Supporting Information Available: Ligand preparation procedures
and representative catalytic reaction procedures (PDF). This material is
available free of charge via the Internet at http://pubs.acs.org.
JA002543E
(11) Garrou, P. E. Chem. ReV. 1985, 85, 171-85.
(12) Kong, K.-C.; Cheng, C.-H. J. Am. Chem. Soc. 1991, 113, 6313-5.
(13) Morita, D. K.; Stille, J. K.; Norton, J. R. J. Am. Chem. Soc. 1995,
117, 8576-81.
(14) Goodson, F. E.; Wallow, T. I.; Novak, B. M. J. Am. Chem. Soc. 1997,
119, 12441-12453.
(10) Palucki, M.; Wolfe, J. P.; Buchwald, S. L. J. Am. Chem. Soc. 1996,
118, 10333-10334.