Organic Letters
Letter
the 1,2-dioxene product 3 under catalytic conditions. However,
due to conjugation with an additional double bond, a
peroxycarbenium ion may react with the nucleophile at the
less hindered C-5 site, thus rendering formation of the target
3,3,6,6-tetrasubstituted 1,2-dioxenes bearing two consecutive
quaternary carbons more difficult. Herein, we wish to report
our recent progress on such a catalytic reaction by identifying
scandium(III) triflate as an efficient catalyst to promote the
C−C bond-forming reactions of 3-(2-methoxyethoxy)-endo-
peroxy ketals with various nucleophiles such as silyl ketene
acetals, silyl enol ethers, allyltrimethylsilane, and trimethylsilyl
cyanide at room temperature. To our knowledge, this approach
establishes the first general route to construct the medicinally
important 3,3,6,6-tetrasubstituted 1,2-dioxenes bearing 3-
acetyl/allyl/cyano functional groups and adjacent quaternary
carbon centers.
Scheme 1. Research Background and Our Approach
Towards 1,2-Dioxenes
We initiated our study by synthesizing a series of 3-phenyl-3-
alkoxy endoperoxy ketals 1a and 6a−d as the model substrates
via two simple steps of photooxygenation and ketalization
reactions from enone 4a9a (Table 1). Since the simple
hemiketal 5a might be also possible for the C−C bond-
forming reaction, we first tested its reaction with trimethylsilyl
ketene acetal 2 in the presence of 10 mol % of Sc(OTf)3 at
room temperature. As expected, the conversion of this catalytic
reaction was very low, and a large amount of starting material
was left even after prolonged reaction time. However, we were
pleased to find that the target 3,3,6,6-tetrasubstituted 1,2-
dioxene product 3a could be isolated in 14% yield (Table 1,
entry 1), indicating that generation of the peroxycarbenium ion
and regioselective formation of the C−C bond at the C-3
position were feasible via manipulation of the catalytic
conditions. Subsequently, other 3-alkoxy endoperoxy ketals
6a−c with increased leaving ability were screened. Although
faster reactions were observed with these substrates, the
desired product 3a was obtained with 29−40% yield (Table 1,
entries 2−4). Next, we turned our attention to endoperoxy
ketals 1a and 6d bearing 3-(2-methoxyethoxy) and 3-(3-
methoxypropoxy) groups. We envisioned that the pendant
bidentate ligands in these substrates might selectively chelate
with Sc(OTf)3 and consequently facilitated the formation of
the peroxycarbenium intermediate. Indeed, the reaction of 1a
with 2 proceeded to full conversion within 1 h and improved
the reaction yield to 57%. For comparison, substrate 6d with a
three-carbon linkage turned out to be less efficient, only
affording product 3a in 36% yield (Table 1, entries 5 and 6).
At this stage, other commonly used Lewis acids were further
screened, but they all proved to be inferior to Sc(OTf)3. For
example, using 10 mol % of La(OTf)3 and In(OTf)3 only
resulted in a trace amount of product 3a (Table 1, entries 7−
11). Later, the impact of catalyst loading was evaluated.
Increasing the loading of Sc(OTf)3 to 20 mol % resulted in a
lower yield of 3a, while decreasing its loading to 2.5 mol %
completely shut down the reaction. The best choice was 5 mol
% catalyst, which improved the reaction yield to 71% (Table 1,
entries 12−14). Finally, after further screening of other
reaction parameters such as reaction temperature and solvents,
we established the optimal conditions by conducting the
reaction of endoperoxy ketal 1a and silyl ketene acetal 2 with 5
mol % of Sc(OTf)3 in DCM at room temperature.
such as silyl ketene acetals in the presence of proper Lewis acid
catalysts.
Previously, C−C bond-forming reactions via reactive
oxycarbenium ions were well developed,10 but the related
reactions of peroxycarbenium ions only received less
attention,11 in spite of their similar capability of stabilizing
carbon cations by the vicinal peroxy group. In 1982, Nojima
and co-workers reported the BF3-mediated cyclization of
ozonides with olefins via the peroxycarbenium intermediate-
s.11a,b In 1993, Dussault and co-workers described the
synthesis of functionalized peroxides based upon the TiCl4-
or SnCl4-promoted C−C bond-forming reactions of peroxy
acetals or ketals with allylsilanes and silyl enol ethers at low
temperature.11c,g Thereafter, these authors11d−f as well as
11k,l
Woerpel,11h,i Wu,11j and Ferrie
extended this C−C bond-
́
forming method to the synthesis of five-membered 1,2-
dioxolanes under similar strong Lewis acid conditions.
Nevertheless, employing such an approach to construct the
six-membered endoperoxides was rare.12 One example was the
allylation of a 3-methoxy endoperoxy acetal substrate, which
was also reported by Dussault to build 3,3,6-trisubstituted 1,2-
dioxene with a stoichiometric amount of SnCl4 or TiCl4
(Scheme 1B).11c,g Probably, the strong and excessive Lewis
acid conditions might lead to degradation of the starting
materials. Thus, a mild catalytic condition might be essential
for the C−C bond-forming reactions of six-membered
endoperoxy ketals.
We envision that selective binding of a Lewis acid catalyst to
an exocyclic OR2 group in endoperoxy ketal 1 to form the
peroxycarbenium ion may be possible since generation of the
undesired oxycarbenium or carbenium ion via opening of the
dioxene ring is disfavored (Scheme 1C). Subsequently, the
reactive peroxycarbenium intermediate may undertake the C−
C bond-forming reaction with the nucleophile 2 and produce
After establishing the optimal conditions, the scope of
endoperoxy ketals was probed. As summarized in Scheme 2,
this new catalytic reaction was quite general. A range of 3-aryl
endoperoxy ketals were well tolerated, delivering highly
1633
Org. Lett. 2021, 23, 1632−1637
Feng, Haowei
