7030
Y. Egoshi et al. / Tetrahedron Letters 54 (2013) 7029–7030
Table 1
Table 3
Synthesis of epoxy alcohol 3 from 2 via 5
Optimization of the ring-opening reaction of 4 to 1
Entry
Photosensitizer
(0.2 mol %)
Solvent
Time (h)
Yield (3, %)
Entry
Lewis acid
Equiv
Temperature (°C)
Time (h)
Yield (%)
Ref. 1
1
2
3
4
5
6
7
8
BF3ÁOEt2
BF3ÁOEt2
TiCl4
3
3
0
4
4
6
6
24
24
0.75
19
6
2
19
6
55
8
4
17
Ref. 1
TPP
TPP
RB
RB
RB
CH2Cl2
CH2Cl2
CH2Cl2
Acetone
Ethanol
4
3
2
2
2
95
92
rt
rt
rt
rt
rt
rt
rt
rt
40
rt
rt
rt
1
2
3
4
1.05
1.05
1.5
1.5
1.5
1.5
1.05
1.05
1.5
0.2
3
Trace
Complex mixture
Complex mixture
AlCl3
Pr(OTf)3
Nd(OTf)3
Sc(OTf)3
Sc(OTf)3
Cu(OTf)2
Cu(OTf)2
Cu(OTf)2
Cu(OTf)2
Cu(OTf)2
NR (43a)
NR (58a)
45
20
53
43
55
26
51
9
Table 2
10
11
12
Oxidation of epoxy alcohol 3 to afford epoxy ketone 4
2
a
Yield of recovered starting material.
Cu(OTf)2, however, gave 1 in 53% yield (entry 8). When the reaction
temperature was increased to 40 °C, the yield decreased slightly to
43% (entry 9). An investigation of the effect of the number of molar
equivalents of Cu(OTf)2 used in the reaction on the yield of 1 indi-
cated that the complex works as a non-catalytic reagent (entries
10–12). Consequently, a reproducible procedure using 1.5 equiv
of Cu(OTf)2 in CH2Cl2 at room temperature for 19 h was found to
give the desired (R)-7-hydroxycarvone 1 in 55% yield (entry 10)
as the best condition. Large scale synthesis of 1 under the same
condition gave similar yield with 30–40% recovery of starting
material.8 The relatively mild Lewis acid Cu(OTf)2 may work as a
proper non-catalytic Lewis acid via chelation to the corresponding
oxygen atoms.
Entry
Conditions
Yield (%)
Ref. 1
(COCl)2, DMSO, Et3N, CH2Cl2, À78 °C to rt
(COCl)2, DMSO, Et3N, CH2Cl2, À78 °C to rt, 2 h
Dess–Martin periodinane, CH2Cl2, 45 min
TPAP (5 mol %), NMO, MS4Å, CH2Cl2, 0 °C, 1 h
90
27
85
90
1
2
3
However, only a trace amount of epoxy alcohol 3 was obtained (en-
try 2). In highly polar solvents, such as acetone and ethanol, com-
plex mixtures were observed that included both unreacted starting
material 2 and intermediate 5, as identified by GC–MS (entries 3
and 4). Thus, the reaction using TPP was determined to provide
the most effective conditions.
In conclusion, reproducible reaction conditions for the prepara-
tion of (R)-7-hydroxycarvone 1 from (S)-a-pinene 2 via photooxy-
genation and migration, TPAP oxidation, and Cu(OTf)2-mediated
ring-opening were described. The desired product 1 was thus ob-
tained in 46% yield over three steps.
Oxidation of the obtained epoxy alcohol 3 to generate epoxy ke-
tone 4 was then investigated (Table 2). While McIntosh and co-
workers reported that the Swern oxidation of 3 proceeds in 90%
yield,1 our attempt gave the product 4 in only 27% yield (entry 1).
Dess–Martin periodinane6 and TPAP7 oxidations were attempted
to afford ketone 4 in 85% and 90% yields, respectively (entries 2
and 3). Thus, it was concluded that oxidation using TPAP with N-
methylmorpholine N-oxide (NMO), which is a well-accepted proto-
col for large scale-synthesis, was the best method for this step.
The final step in the synthesis of (R)-7-hydroxycarvone 1 in-
volved ring-opening of the pinene skeleton and the epoxide group
in 4. Previously, it was shown that the use of BF3ÁOEt2 as a Lewis
acid afforded 1 in moderate yield (55%) on a multi-gram scale,
while basic, acidic, and silylative conditions were found to cause
decomposition or no reaction.1 In our hands, however, the reaction
with BF3ÁOEt2 gave the desired product 1 in only 8% yield (Table 3,
entry 1) with poor reproducibility and decomposition of starting
material, particularly on a gram-scale. The ring-opening reaction
under various Lewis acid conditions was thus investigated. When
either TiCl4 or AlCl3 was used as the Lewis acid for 6 h, the product
was obtained in 4% and 17% yields, respectively (entries 2 and 3). In
addition, the use of typical Lewis acids, such as CuCl2, Ti(OiPr)4,
Yb(OTf)3, and Mg(OTf)2 led to no reaction (data not shown). Fur-
thermore, when lanthanide triflates, such as Pr(OTf)3 and Nd(OTf)3,
were applied to this reaction, only recovery of the starting material
was observed in 43% and 58% yields, respectively (entries 4 and 5).
Use of Sc(OTf)3 for 45 min and 19 h gave the desired product 1 in
45% and 20% yields, respectively (entries 6 and 7), along with gen-
eration of byproducts that were difficult to separate. The use of
Acknowledgment
We thank Professor Masahiro Rikukawa (Sophia University) for
his support of equipments. This work was supported in part by a
Grant-in-Aid for Young Scientists (B) from the Ministry of Educa-
tion, Culture, Sports, Science, and Technology (MEXT), Japan.
Supplementary data
Supplementary data (experimental procedures and character-
ization data) associated with this article can be found, in the online
References and notes
8. When the reaction is run on a large scale (i.e., 500 mg or more), it should be run
for less than 2 h due to the generation of byproducts that are difficult to
separate.