.
Angewandte
Communications
which is stored as a 50% toluene solution, afforded the
desired product 3a in good yield (entry 1), a quantitative yield
was observed when 2a was used in pure form (entry 2). The
reaction time could be shortened to 5 hours without a reduc-
tion in yield (entry 3). Using 2.0 equivalents of 2a is crucial
because the yield dropped significantly when only 1.0 equiv-
alent was employed (entry 4). The use of the preformed
cationic rhodium(III) precursor [Cp*Rh(CH3CN)3(SbF6)2]
slightly reduced the yield (entry 5). A lower catalyst loading
proved to be effective. At a 2.5 mol% catalyst loading,
complete conversion into product was maintained (entry 6),
while at a 1.25 mol% loading only a slight drop in yield was
observed (entry 7). Conversion into product was not observed
when either [{Cp*RhCl2}2] or AgSbF6 alone were used
(entries 8 and 9), which suggests that a cationic rhodium(III)
substituent, R3, is not essential for the success of the reaction,
thereby enabling the preparation of trisubstituted furans
(3h,i). We also explored variation at R1 and established that
different alkyl groups work well in this transformation (3e–g).
Varying the R2 substituent demonstrated that alkyl-, aryl-
(3h), and benzyl-substituted (3i) oximes are effective sub-
strates. However, a substituent at this position is crucial
because no product was obtained when R2 was H. To enhance
the practicality of the method, we performed the reaction on
the benchtop and found that the yield was comparable to that
observed when the reaction was set up in a glovebox under
inert atmosphere conditions (see 3a).
Having demonstrated a broad scope for coupling oximes,
bearing different substitution patterns, with 2a, we next
explored the possibility of extending this transformation to
less activated aldehydes. To apply the transformation to
aromatic aldehydes, or even to aliphatic aldehydes, the
À
species is required for this C H bond functionalization
process. Although reaction of 1a with 2a proceeds to
complete conversion within 5 hours at 508C (entry 6),
a longer reaction time and a higher temperature (758C) did
not result in any reduction in yield (entry 10). These more
forcing conditions were found to be necessary for less reactive
oxime substrates.
À
reversible nature of metal-catalyzed C H bond addition to
aldehydes must be considered (Scheme 1). In contrast to the
reaction of the electron-deficient and destabilized 2a, which
Having defined an effective catalyst system and reaction
conditions for the synthesis of furan 3a in high yield, we next
explored the substrate scope with a range of a,b-unsaturated
oximes (1) having different substitution patterns (Table 2).
Annulation of cyclic and acyclic trisubstituted oximes resulted
in fully substituted furan products in good yields (3a–g), with
cyclohexyl (3a,e,g) and cyclopentyl (3b) derivatives provid-
ing access to different bicyclic frameworks. The terminal
Table 2: Oxime substrate scope with ethyl glyoxylate.[a]
Scheme 1. Proposed reaction pathways.
likely favors formation of alcohol intermediates such as those
illustrated in Scheme 1, literature reports suggest that starting
materials are generally more stable than the addition
products when less-electron-deficient aldehydes are used.[12]
For unactivated aldehydes we therefore envisioned that
capture of the initial alcohol product by cyclization and
aromatization might be necessary to drive the reaction to
completion.[13]
We initially focused on the identification of suitable
reaction conditions for the efficient coupling of 4-trifluoro-
methylbenzaldehyde (5a) and the oxime 1b (Table 3). The
use of DCE as a solvent, which we have previously found to
À
be one of the most effective for rhodium(III)-catalyzed C H
À
À
additions to polarized C N and C O multiple bonds,
provided the desired furan 7a in poor yield (entry 1). A
range of solvents with different polarities and hydrogen-
bonding capabilities were therefore evaluated because differ-
ent solvation properties might be expected to affect the
equilibrium between the starting materials and the alcohol
intermediate, and could also impact the rate of cyclization and
aromatization. A variety of alcohols with different steric
properties and acidities were first explored, with CF3CH2OH
increasing the yield to 50% (entries 2–4). Evaluation of
aprotic solvents resulted in an interesting trend. Polar and
[a] Reaction conditions: 1 (0.20 mmol), purified ethyl glyoxylate (2a;
0.40 mmol) in 0.67 mL of DCE. The yield of isolated, purified material is
reported for each product. [b] Conducted in a Schlenk flask under N2 on
the benchtop.
630
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 629 –633