Angewandte
Communications
Chemie
reaction to a good extent (see Supporting Information).
Among these bases, K2CO3 was selected for further studies on
the grounds of cost and practicality. The reaction could
proceed without base (Table 1, entry 1), but its presence led to
Table 1: Most influential factors on the HFIP-promoted cyclization of a-
carbonyl sulfoxonium ylide 2 into compound 3.
Entry
Variation of conditions[a]
Yield[b]
1
2
3
4
5
No base
63%
0%[c]
20%
0%[c]
0%[c]
K2CO3 (1 equiv), either TFE, i-PrOH, or 1,2-DCE
No base, HFIP (5 equiv), 1,2-DCE
No base, 2,6-Me2-C6H3OH (5 equiv), 1,2-DCE
No base, TMP·HCl (5 equiv), i-PrOH
[a] From those depicted in Equation (1). [b] Yield of isolated product.
[c] The starting material was recovered in at least 90% yield. 1,2-DCE:
1,2-dichloroethane; TFE: trifluoroethanol; TMP: 2,2,6,6-tetramethylpi-
peridine.
Figure 2. Cyclization of a-carbonyl sulfoxonium ylides promoted by
HFIP. All yields given are for isolated products from reactions
conducted with 0.66 mmol of 4 (0.2m). [a] At 908C. [b] Microwave
heating at 808C for 1 hour. [c] 48 hours. [d] Ratio of regioisomers; the
position of the alternative carbon-carbon bond formation is denoted
by an asterisk. [e] 0.1 mmol of 4. [f] 0.26 mmol of 4.
a higher yield of 3. Significantly, the presence of HFIP was
essential, whereas the conversion remained null in other
solvents (entry 2). Using HFIP as additive in another solvent
partially restored the reactivity (entry 3). Replacing HFIP
with additives of similar acidity left the starting material 2
intact (entries 4 and 5), which suggests that the protic nature
of HFIP is not sufficient to explain the cyclization of 2 into
indanone 3.
With optimised conditions in hands, we examined the
scope with substrates 4a–o (Figure 2). Indan-2-ones 5a–f
were obtained in good yields, whereas gem-dimethylated 4g
failed to afford 5g. Nevertheless, we were pleased to observe
that halogen-substituted 4d and 4e underwent the cyclization
to give 5d and 5e in excellent yield, although more forcing
conditions were necessary, as was the case for the cyclization
of electron-poor substrate 4 f into 5 f. In the case of substrate
4c, the expected cyclization product 5c was obtained in 65%
yield alongside solvolysis product 6. We then studied the
regioselectivity of this cyclization with 4h–n, and observed
excellent selectivity in the reactions that gave 5h–m in very
good yields. Thus, no six-membered ring was formed in the
cyclization of 4k into 5k. However, the two possible
regioisomers of 5o were obtained in an almost equimolar
ratio.
Figure 3. Plausible mechanism of the HFIP-promoted cyclization of a-
carbonyl sulfoxonium ylides.
promote the formation of oxy-allyl cations in the case of other
leaving groups.[15] Secondly, when keeping R1 as Me and
varying R2 (H, p-iBu, p-OMe, p-Cl), a good correlation (R2 =
0.97) of the relative rates with Hammett sp parameters[16] was
obtained and gave a reaction constant 1 of À0.6, in good
agreement with those found for other examples of antarafa-
cial five-centres 4p-electrocyclization of cationic intermedi-
ates.[17] A similar mechanism could therefore be plausible for
the rearrangement of V into IV. Thirdly, in the absence of
substituent R1, and when R2 was p-OMe, the equilibrium
between V and VI led to solvolysis product 7 in 54% yield as
sole product of the reaction. This side reaction was only partly
prevented when R1 was a methyl group, that is, in the case of
4c, and a mixture of 5c and 6 was obtained (Figure 2).
Moreover, in the absence of substituent R1, and when R2 was
p-CF3, a Favorskii rearrangement via cyclopropanone VII led
We propose the following mechanism to account for these
results (Figure 3). In view of infrared spectra of substrates 4a–
o (n(C O): 1557–1572 cmÀ1), it is reasonable to consider
=
intermediate I as starting point, and it would be in equilibrium
with II and III under the conditions. Cyclization of III into IV
by pathway (a) would lead to the observed products after
rearomatization. However, several observations point to oxy-
allyl cation V as a plausible intermediate from III to IV and
suggest pathway (b) as a possibly more likely alternative.
Firstly, HFIP is a very strong H-bond donor[14] and it could
À
promote the cleavage of the C S bond in III. Furthermore,
the combination of HFIP and a base has been reported to
2
ꢀ 2019 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2019, 58, 1 – 6
These are not the final page numbers!