Organic Letters
Letter
arene (Ar2) originally linked to the triple bond (Scheme 1b).
Thus, the catalytic method developed provides a comple-
mentary synthetic strategy to the most common one depicted in
Scheme 1a that differs in the substitution pattern at the internal
cycle of the phenanthrene skeleton constructed. Moreover,
upon careful selection of the appropriate reaction conditions,
mainly the solvent, unsymmetrically (9,10)-disubstitued
dihydrophenanthrenes can also be prepared from the same
starting 1,3,5-trien-7-ynes (Scheme 1b).
Based on our experience in the electrophilic cycloisomeriza-
tion of o-alkynylstyrenes,12 we selected 2-(2-methylprop-1-en-
1-yl)-2′-(phenylethynyl)-1,1′-biphenyl (1a) as the model for
assessing the reactivity in the presence of gold catalysts. This
substrate possesses two nucleophilic entitiesan arene and the
olefinin suitable locations to react with the alkyne that would
render different cycloadducts. However, we envisioned that the
high substitution of the alkene would facilitate its selective 6-exo
nucleophilic addition to the metal-activated acetylene, thus
furnishing the desired phenanthrene skeleton. At the outset,
several gold-derived complexes as catalysts, and different
solvents, were studied, and the most significant results are
summarized in Table 1.
occurred when THF was employed as solvent and the reaction
was conducted for an extended reaction time (72 h). Thus,
under these conditions, dihydrophenanthrene 3a could be
selectively obtained in good yield (entry 3). Analogous
experiments in THF varying the gold catalyst gave varying
mixtures of 2a, 3a, and other unidentified products, whereas no
evolution was observed in other solvents such as acetonitrile or
DMF.13
The minimal influence of the nature of the cationic gold(I)
complex was determined from the outcome of the reactions
conducted in DCM. Thus, phenanthrene 2a was obtained
exclusively from model substrate 1a in quantitative yields with
most of the catalysts tested, including less active gold(III) salts
such as AuCl3 (entries 4−9). Of these gold complexes,
JohnPhosAu(MeCN)SbF6 exhibited a slightly improved
selectivity and no traces of the regioisomeric cycloadduct 3a
could be detected. Moreover, different temperatures or
concentrations resulted in a higher percentage of byproducts,
whereas a limited screening with silver salts showed no
beneficial effect or improvement in this process.13 Finally,
lowering the catalyst loading to 1 mol% or scaling the reaction
to 1 mmol had no impact on either the yield or the reaction
time (entry 10). In summary, an appropriate choice of catalyst
and solvent allows the selective formation of phenanthrene 2a
(entry 10) and dihydrophenanthrene 3a (entry 3).
a
Table 1. Optimization of Reaction Conditions
With these results in hand, and based on previous studies in
the cycloisomerization of 1,n-enynes,10,11c,12,14 we propose the
following mechanism that accounts for the formation of both
phenanthrene derivatives 2a and 3a (Scheme 2).
b
Scheme 2. Proposed Mechanism
entry
[Au]+
PPh3AuNTf2
PPh3AuNTf2
PPh3AuNTf2
ArO3PAuCl/AgNTf2
tBu3PAuNTf2
XPhosAuNTf2
JohnPhosAu(MeCN)SbF6
IPrAuCl/AgNTf2
AuCl3
solvent
product
conv (%)
1
2
3
CH2Cl2
toluene
THF
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
2a
2a
3a
2a
2a
2a
2a
2a
2a
2a
100 (94)
c
70
d
100 (79)
e
c
4
5
6
7
8
9
100
80
100 (98)
100 (99)
100 (97)
100 (98)
f
g
10
JohnPhosAu(MeCN)SbF6
100 (99)
a
Reactions conducted using 0.05 mmol of 1a in 1 mL of solvent at 25
b
1
°C for 2 h. Conversion estimated by H NMR spectroscopy (300
Hz); isolated yield in parentheses for an experiment conducted with
0.2 mmol of 1a. < 5% of 3a was detected. Significant amounts of 3a
c
d
and other unidentified products were observed. Conducted for 72 h.
e
f
g
Ar = 2,4-(tBu)2C6H3. Conducted with 1.0 mol% of catalyst. Same
isolated yield was obtained in an experiment conducted with 1.0
mmol of 1a.
The reaction of 1a in the presence of 5 mol% PPh3AuNTf2 as
catalyst in dichloromethane at room temperature selectively
afforded the 9-substituted phenanthrene 2a in less than 2 h
(entry 1). Remarkably, the competitive pathway resulting from
addition of the arene to the alkyne, which would produce the
corresponding regioisomeric phenanthrene, described in
Scheme 1a, was not detected. Encouraged by this initial result,
the influence of the solvent on the cycloisomerization of 1a was
explored. Thus, the reaction conducted in toluene also gave 2a
as the major product, albeit with lower conversion and
selectivity (entry 2). From that crude mixture, the formation
of a disubstituted 9,10-dihydrophenanthrene 3a was deter-
mined. Interestingly, a complete switch in the regioselectivity
The reaction is initiated upon activation of the acetylene of
the starting enyne 1a upon coordination to the gold complex,
followed by an intramolecular 6-exo-dig nucleophilic addition of
the alkene moiety to give cationic intermediate I. This
intermediate can be described as the resonance hybrid of two
structures, namely the cyclopropylgold(I) carbene Ia and the
gold-stabilized homoallylic carbocation intermediate Ib, which
delocalizes the positive charge over the molecule. Cyclopropyl
ring expansion of Ia then furnishes the (η2-cyclobutene)gold(I)
complex II,15 which, after ring opening of the cyclobutene and
subsequent demetalation, would lead to phenanthrene 2a and
release the gold catalyst for a new cycle (path a). An alternative
pathway (a′) involving the direct transformation of inter-
B
Org. Lett. XXXX, XXX, XXX−XXX