Organic Letters
Letter
a
hindrance arising from the presence of two substituents. When
phenols with only one meta substituent were tested (S29−
S32), C−H insertion occurred at both ortho carbons, resulting
in a mixture of regioisomers. Interestingly, for substrates with a
Me (S29), chloro (S30), or bromo (S31) substituent, the
more congested regioisomer was slightly favored over the other
one. In contrast, reactions of 1-naphthol (S33) and 2-naphthol
(S34) were regioselective, affording a single product.
Remarkably, 2-methyl-3-pyridinol (S35) also underwent the
C−H insertion reaction, despite the potential for inhibition of
the catalyst by coordination of the nitrogen atom; the desired
product was obtained in 40% yield.
Scheme 3. Enantioselective Version
We tested four naturally occurring or bioactive small
molecules (S36−S39). Reaction of sesamol (S36) preferen-
tially produced the sterically favored product in 62% yield,
whereas the reactions of formononetin (S37) and 6-
hydroxycoumarin (S38) occurred preferentially at the sterically
hindered site, probably as a result of the strong electronic
influences of their substituents. Note that, with these three
substrates, the other regioisomer was formed but was not
isolated, because of the low yields (<10%). From capsaicin
(S39), the product was isolated in 40% yield with the olefin
and amide functionality intact. Although the yields of the
reactions of these bioactive molecules were relatively low, their
successful transformation demonstrated the potential utility of
the protocol for late-stage C−H functionalization reactions.
We also conducted gram-scale reactions of phenol (S1) and p-
Me-substituted phenol (S23) at a catalyst loading of only 0.5
mol %. These reactions afforded P1 and P23 in 70% and 63%
yields, respectively.
a
Unless otherwise specified, all reactions were performed with 0.1
mmol of substrate and 0.25 mmol of diisopropyl diazomalonate in 1.0
mL of cyclohexane. The percentages were isolated yields and
enantiomeric excesses, respectively. The absolute configuration of
P41 was determined by X-ray crystallography, and those of the other
b
products were assigned by analogy. Reaction temperature: 80 °C.
cEt3BnNCl and toluene were used as the additive and the solvent,
respectively.
We then tried to achieve the enantioselective version of the
reaction by employing chiral diene ligands.11 After an extensive
screening of the ligands and the reaction conditions (see the
use of diene L7, [Rh(ethylene)2Cl]2, and diisopropyl
diazomalonate with nBu3BnNCl as an additive and cyclohexane
as a solvent was able to produce the corresponding products in
good yields and high enantiomeric excess (ee) values (Scheme
3). The reaction is compatible with unsubstituted phenol
(P40) and ortho-isopropyl (P41), benzyl (P42), and
trifluoromethyl-substituted (P43) phenols, as well as phenols
with various ortho-aryl substituents (P44−P48). The hetero-
aromatic substituents (P49−P53) did not influence the
reactivity and selectivity. 6-Bromo-2-naphthol was also
reactive, giving product P54 in 96% yield and 90% ee. With
meta-dimethyl (P55) and para-phenyl-substituted phenols
(P56), the ee values decreased to 83% and 84%, respectively.
We then investigated the reaction mechanism by performing
several control experiments (Scheme 4). First, when we
separately prepared methyl phenyl diazomalonate and then
subjected it to the standard reaction conditions, intramolecular
C−H insertion did not occur; instead, the carbene
dimerization product was obtained in 52% yield (Scheme
4a). This result rules out the possibility that the mechanism
begins with a transesterification reaction between phenol and
1. Moreover, replacement of 1 with the bulkier di-t-butyl
diazomalonate prevented cyclization, and a mixture of
monoalkylated and dialkylated products was obtained (Scheme
4b). By comparison, the absence of the dialkylated compound
in those reactions with 1 demonstrates that a free hydroxyl
group is required for the C−H insertion reaction. Therefore,
we hypothesized that a facile transient formation of an
oxonium ylide intermediate6 from the free hydroxyl group
Scheme 4. Mechanistic Experiments
and the metal carbene might have promoted the ortho-selective
insertion via a favorable cyclic transition state. If so, it would be
possible to transform the O−H insertion product (2) into this
intermediate via reverse proton transfer and attachment of the
rhodium catalyst. To evaluate this possibility, we performed
the reaction between 2 and 1 in the presence of a rhodium
catalyst and Me3BnNOH, and we found that P1 was produced
in 42% yield (Scheme 4c). Furthermore, when a d6-diazo
compound was used, four isotopic isomers were produced,
along with a 16% yield of the deuterated O−H insertion
C
Org. Lett. XXXX, XXX, XXX−XXX