5732
P. Lemechko et al. / Tetrahedron Letters 48 (2007) 5731–5734
Table 1. Catalyst screening in the hydroalkoxylation of methyl
citronellyl ether 1a in MeOH, 24 h at 60 °C
lol 1b (entry 2), limonene 1c (entry 3), citronellene 1d
entry 4), and a-pinene 1e (entry 5) was achieved in
(
a
0
Entry Catalyst
mol % Conversion Yield 2a/2a
24 h in the presence of 5 mol % catalyst in 15–70%
yields. Hydromethoxylated products were the only
products isolated after the reaction, along with some
unconverted starting material. Conversions were in the
range of 33–100% and yields were sometimes lowered
by a partial evaporation of the volatile products during
solvent removal and possible degradation of the starting
material. The free hydroxyl group of citronellol 1b did
not modify the course of the reaction, and the methoxyl-
ated compound 2b was formed regiospecifically as the
only product in 50% yield. With limonene 1c, the reac-
tion yielded 53% of the mono-hydromethoxylated prod-
uct 2c corresponding to the functionalisation of the
isoprenyl moiety. Additionally, doubly methoxylated
product was formed in 26% yield from the hydrometh-
oxylation of both double bonds of 1c. In the case of citr-
onellene 1d, the terminal double bond was not modified.
During the hydroalkoxylation process, bicyclic a-pinene
rearranged to the terpineol derivative through ring-
(
%)
1
2
3
4
5
6
7
8
9
0
1
None
TfOH
HNTf
Sn(OTf)
Sn(NTf
Al(OTf)
In(OTf)
Sn(OTf)
Pd(CH
Al(OTf)
Sn(OTf)
—
5
5
5
5
5
5
5
5
0
—
—
22
25
70
47
31
39
50
5
13
23
54
45
9
29
28
<5%
39
80/20
95/5
84/16
91/9
89/11
76/24
95/5
—
2
4
2
)
4
3
3
2
3
CN)Æ2OTf
1
1
3
20
20
45
79
85/15
84/16
4
77
a
1
Hydroalkoxylation yield, determined by H NMR.
Heating olefin 1a in anhydrous methanol at 60 °C in the
presence of 5 mol % HOTf or HNTf led only to 22–
5% conversion after 24 h, yielding 13–23% of hydroalk-
2
2
0
oxylated products 2a + 2a (entries 2 and 3). The reac-
tion was more efficient when catalysed by metal-based
Lewis superacids such as Sn(IV) triflate (70% conver-
sion, entry 4) and Sn(IV) triflimidate (47% conversion,
entry 5). Both catalysts were prepared by an electro-
chemical procedure from metallic tin and HOTf or
HNTf2. Al(III) and In(III) triflates led to similar reac-
tion profiles albeit with moderate conversions (entries 6
and 7). Sn(II) triflate qualitatively reproduced the result
obtained with Sn(IV) triflate but in lower yield (entry 8);
the electrophilic activation of the doubly charged tin
cation was probably too weak to efficiently allow the
attack of methanol. The cationic Pd(II) complex
Pd(CH CN)Æ2OTf, formed by treatment of Pd(CH CN)Æ
2
2
opening of the four-membered ring.
The reactivity of the double bond of substrate 1f was
interesting, because of the possibility of an intramolecu-
lar 6-endo-trig cycloisomerisation with the secondary
alcohol function. However, after 24 h at 60 °C in
MeOH, a conversion of 53% was reached and only the
hydromethoxylated product 2f was regioselectively
formed in 32% yield (entry 6). The competition between
the addition of MeOH and the intramolecular addition
of the alcohol was thus in favour of the intermolecular
process.
2
0
3
3
Cl2 with 2 equiv of AgOTf, was not active in this
reaction (entry 9). In general, the regioselectivity was
strongly in favour of the Markovnikov-type product
Examples of functionalisation of exocyclic double bonds
were carried out with substrates 1g and 1h, which were
converted, in methanol, to the corresponding methoxyl-
ated and rearranged derivatives 2e and 2h in 61% and
2
a.
7
0% yields, respectively (entries 7 and 8). The terminal
In each case, additional reaction time did not allow to
increase the conversion. However, a higher catalyst ratio
resulted in the enhancement of both the yield and the
conversion with Al(III) triflate and Sn(IV) triflate
position of the double bond, more available to bind
the bulky Sn(OTf) catalyst, afforded higher conversions
4
as compared to those obtained with 1e.
(
entries 10 and 11), the yields rising from 9% to 39%
Disubstituted internal olefins such as norbornadiene,
norbornene and 2-acetylnorbornene, as well as cyclo-
octene and cyclooctadiene were unreactive, and 1,1-
diphenylethylene yielded only 7% of methoxylated
product (not shown). Replacing methanol by ethanol
and 1-butanol allowed to heat the reaction at higher
temperatures (from 78 °C to 117 °C) but did not afford
better yields, although conversions were slightly
improved (compare entries 1, 5 with EtOH and
n-BuOH). The hydrobutoxylated products 4a and 4e
were formed as the sole products in 40% and 49% yields,
when the reactions were carried out in 1-butanol at 117
and 80 °C, respectively.
and from 54% to 77%, respectively. To check if the
catalyst could suffer from rapid deactivation under the
reaction conditions, we tried to add it portionwise to
the reaction medium, unfortunately without significant
improvement. Having identified Sn(IV) triflate as the
best catalyst for the hydroalkoxylation of 1a, we further
screened a selection of solvents. Running reactions in
methanol remained the best choice, since with 10 equiv
of MeOH in acetonitrile, toluene or dichloroethane,
the rates of conversion remained below 50%. Under
reaction conditions inspired by those of Hartwig and
2
1
coll. (1 mol % catalyst in toluene), HOTf remained
poorly active in the hydromethoxylation reaction, the
conversion remaining below 20%.
The synthesis of the methoxysesquiterpene 8-methoxy-
cedrane 2i was further envisaged as a synthetic applica-
tion of this methodology. This compound is used in
We next extended the substrate scope of this hydroalk-
oxylation reaction catalysed by Sn(IV) triflate (Table
2
Ò
fragrance industry under the tradenames Cedramber
Ò
). In the terpenoid family, the hydroalkoxylation of tri-
and Cedranfix for its cedarwood and amber olfactory
notes. (ꢀ)-a-Cedrene 1i was submitted to the hydro-
2
3
substituted double bonds such as in the case of citronel-