Organic Letters
Letter
Brønsted acids, no substituted furans or π-bond migration
products were observed.
Scheme 2. Fe-Catalyzed Cycloisomerization
Based on our previous work, we initiated this study by
employing aryl allenyl ketone 2 to the conditions that proved
successful in the cycloisomerization of enyne acetates and
observed a clean cycloisomerization of 2 to the corresponding
indanone 3 in moderate yield (Table 1, entry 1).
a
Table 1. Catalyst Screening
quantitative conversion and isolated yields up to 93% (entries
1−8, Table 2). In the case of meta-substitution, product 20 was
formed as a single isomer (entry 9, Table 2). Substitution in the
ortho position leads to the decreased reactivity of the starting
materials (entry 10, Table 2); however, this effect can be
overcome through substitution in the meta- or para-position as
exemplified for the cycloisomerization of highly decorated
allene 23 to arene 24, which is being formed in a quantitative
fashion (entry 11, Table 2). In the case of nonidentical
substituents at C3 of the allene moiety, a high degree of (Z)/
(E)-selectivity up to 7.5/1 was observed (entries 12−14, Table
2). Substitution in the α-position significantly increased the
reactivity, leading to mostly quantitative isolation of the
cycloisomerization products (entries 15−17, Table 2). While
ortho substitution was of no hindrance now, even a terminal
allene 37 could be transformed to the corresponding product
38 albeit in moderate yields (entry 18, Table 2). In contrast to
oxophilic Lewis acid catalysts, acyl residues in the α-position
could not increase the overall yield of the corresponding
products 42 and 44 (entries 20 and 21, Table 2).8 In no case
was π-bond isomerization to the corresponding endocyclic
indenone products or furan formation observed; even allene 45,
which under thermal conditions undergoes cycloisomerization
and subsequent π-bond migration to the corresponding
indenones, reacted to indanone 46 selectively (entry 22,
Table 2).
b
b
entry
catalyst
3 /4
c
d
1
2
1
67% :0% (5%:0%)
−
−
−
11%
e
3
c
4
HBF4·OEt2
TMSOTf
AuCl3
PtCl2
29% :0%
5
8%:41%
6
−
−
7
8
(o-Ph)C6H4PtBu2)Au(CH3CN)SbF6
(PPh3)AuCl, AgBF4
(PPh3)AuCl
0%:76%
49%:0%
−
9
10
11
AgBF4
47%:0%
a
Reaction conditions: All reactions were performed on a 0.2 mmol
scale using 10 mol % of the catalyst in dry dichloromethane (1 mL) at
b
50 °C under a N2 atmosphere in a sealed tube for 22 h. Yield
obtained through 1H NMR-integration using 1,3,5-trimethoxybenzene
c
d
as internal standard. Isolated yield. The reaction was performed at
room temperature. The reaction was performed at 160 °C in DMSO
e
(0.1 M).
Importantly, the expected furan 4 was not formed under
these conditions. While the blank reaction did not produce any
product (Table 1, entry 2), application of Ren’s conditions
mainly caused decomposition of the starting material (Table 1,
entry 3). Control experiments using HBF4 led to the formation
of indanone 3 (Table 1, entry 4), and TMSOTf led to the
formation of furan 4 along with minor quantities of indanone 3
(Table 1, entry 5). Whereas neither AuCl3 nor PtCl2 showed
any activity (Table 1, entries 6 and 7), a combination of
(Ph3P)AuCl and AgBF4 induced the formation of indanone 3
in moderate yields (Table 1, entry 9). However, as it turned out
AgBF4 was the active catalyst (Table 1, entry 11). (Ph3P)AuCl
did not show any catalytic activity (Table 1, entry 10). In sharp
contrast, Echavarren’s Au(+I)-catalyst was catalytically active,
yet, clean formation of furan 4 was observed (Table 1, entry 8).
Careful optimization of different reaction parameters6
indicated a concentration of 0.1 M in 1,2-dichloroethane
(1,2-DCE) and a temperature of 60 °C to be optimal which
results in the clean formation of product 3 in 83% isolated yield
after 8 h of reaction time using 5 mol % complex 1 (Scheme
2).7
At the outset of this project we envisioned coordination of
the cationic Fe-complex to the C−C π-bond to alter the
reaction pathway;9 hence, steric hindrance close to the carbonyl
group should not have a pronounced effect on the conversion
(Scheme 3).
In order to test this hypothesis, o-methoxy- and o-isopropoxy
aryl allenyl ketones 47 and 49 were prepared and subjected to
the reaction conditions. The corresponding indanones 48 and
50 were isolated in quantitative yields (eqs 1 and 2, Scheme 3).
In a competition experiment both 47 and 49 were reacted using
catalyst 1. The reaction was stopped after 30 min, and the
conversion to 48 and 50, respectively, was analyzed. Sterically
more hindered aryl allenyl ketone 47 reacted even faster, a
result that might reflect the better stabilization of intermediate
carbocations formed within the reaction (eq 3, Scheme 3). In a
final experiment, aldol products 51 and 53 were treated with
catalyst 1 under the established reaction conditions. In both
cases the Fe-catalyzed cycloisomerization took place; however,
the primary formed product underwent fast elimination to the
corresponding 1,3-dienes 52, 54 or a retro-aldol reaction with
formation of indanone 3 in 14% or even 47% yield, respectively
(eq 4, Scheme 3).
With these optimized conditions in hand, we set out to
analyze the scope and limitations (Table 2). The reaction
proved to be broadly applicable, both electron-rich and -poor
ortho-, meta-, and para-substituted aromatic moieties are
reactive. As expected, the substitution pattern at the aromatic
moiety had a profound influence on the conversion rate. para-
Substituents known to stabilize positive charge result in almost
Herein we report an unprecedented metal-catalyzed cyclo-
isomerization of aryl allenyl ketones to 3-aryliden or 3-
alkylidene-indan-1-ones. The cationic Fe(0)-nitrosyl complex
[(Ph3P)2Fe(CO)(NO)]BF4 1 is able to alter the reported
Lewis acid catalyzed pathway in favor of the cycloisomerization
route to give indanones rather than furans. Good functional
B
Org. Lett. XXXX, XXX, XXX−XXX
Teske, Johannes
