may be a possible intermediate for the present reaction. We
found, however, that BF3·OEt2 also catalyzed the rearrange-
ment to give lactone 4g, though only in 13% yield, when
aldehyde 1g was treated with ketene (method B) in Et2O (200
cm3) at room temperature in the presence of 1.5 equiv. of the
acid.10 This observation, along with the result that lactone 3h
did not isomerize to lactone 4h, may suggest another possibility,
that coordination of the carbonyl oxygen of 4-vinyl lactones 3g,
i–k to Lewis acid 2d promoted the heterolytic cleavage of the
C(4)–O bond of the lactones to form a zwitterion, recombina-
tion of which at the other allylic terminus afforded lactones 4.
Further treatment of lactone 4g with EtOH in the presence of
HCl followed by saponification of the resulting ethyl sorbate,
gave hexa-2,4-dienoic acid (sorbic acid) 5 in 90% yield
(Scheme 2). Therefore, the present method provides an easy
access to the acid 5. It should be noted that sorbic acid 5 is
important as a mould and yeast inhibitor, the first step of an
industrial synthesis of which relies on the cycloaddition of
ketene with crotonaldehyde 1g catalyzed by a zinc carbox-
ylate.11 However, the product obtained from the reaction is not
the b-lactone 3g but its ring-opening polymer, poly(3-hydroxy-
hex-4-enoic acid), the viscosity of which causes a great deal of
trouble during the subsequent destructive distillation of the
polyester to the acid 5. Further studies on the scope and
limitations of the allylic rearrangement, as well as the [2 + 2]
cycloaddition, are in progress.
Scheme 2 Reagents and conditions: i, 2d, CH2Cl2; ii, EtOH, conc. HCl,
reflux; iii, KOH, aqueous EtOH, reflux.
Table 2 Tandem cycloaddition–allylic rearrangement of ketene with a,b-
unsaturated aldehydes 1g–k catalyzed by palladium complex 2d
Entry
1
Methoda CH2Cl2/cm3
4
Yield (%)b
1
2
3
4
5
6
7
8
9
1g
1g
1g
1g
1g
1gd
1h
1i
A
A
B
B
B
B
B
B
B
B
20
200
200
200
500
200
200
200
200
200
4g
4g
4g
4g
4g
4g
4h
4i
13
55
70
50c
81
65
0e
This work was supported in part by grants from the Center of
Interdisciplinary Research (Tohoku University), the Takasago
International Corporation and the Chisso Corporation.
77c
58c
66c
1j
1k
4j
4k
10
a See text. b Determined by GC analysis on Quadrex MPS-10 column (0.32
mm i.d. 3 25 m) by the internal standard method. c Isolated yield after
column chromatography on silica gel with hexane–EtOAc (1+1) as the
eluent. d 2.00 mmol. e Lactone 3h was obtained in 96% yield.
Notes and references
† Method B: To a solution of complex 2d (50.0 mmol) in CH2Cl2 (200 or
500 cm3) was added aldehyde 1 (200 mmol). Ketene (ca. 250 mmol) was
bubbled into the mixture over a period of 1 min and the mixture was stirred
for 5 min. This series of operations was repeated until added aldehyde 1
reached the total amount of 1.00 mmol. To the mixture was added an
additional amount of ketene (ca. 1.0 mmol) and the resulting mixture was
stirred for 1 h before work-up.
ering the reaction temperature and changing the molar ratios of
ketene and catalyst 2d to aldehyde 1g did not improve the
product yield, while dilution of the reaction solution was found
to be highly effective (entry 2). Eventually, lactone 4g could be
obtained in good yields by adding aldehyde 1g and ketene
portionwise to a dilute solution of catalyst 2d (Method B†)
(entries 3 and 5). Under these conditions, 2.5 mol% of the
catalyst 2d was sufficient to complete the reaction (compare
entry 6 with entry 3). Similar d-lactones 4i–k were also
obtained in the reaction of a,b-unsaturated aldehydes 1i–k,
while acrolein 1h afforded b-lactone 3h under the same
conditions (entries 7–10).
The formation of lactone 4 can be rationalized by the initial
[2 + 2] cycloaddition of aldehyde 1 with ketene to give the allyl
ester 3, followed by its allylic rearrangement to form lactone 4.
It is known that this type of 1,3-rearrangement of allylic esters
is promoted by Pd0 and PdII complexes.8 It should be noted,
however, that the palladium(II)-catalyzed reaction is reportedly
a [3,3]-sigmatropic rearrangement of allyl esters, which is
impossible for the said lactones 3g–k due to steric reasons. On
the other hand, the palladium(0)-catalyzed rearrangement is
believed to involve a p-allylpalladium(II) intermediate. It is also
reported that palladium(II) salts promote the ring opening of
4-vinyl- (3h), and 4-isopropenyl-oxetan-2-one to afford the
corresponding penta-2,4-dienoic acids.9 A metallacyclic s-
allylpalladium intermediate generated by oxidative addition of
the C(4)–O bond of the oxetan-2-ones to a palladium(0) species
is proposed for the reaction. Thus, an allylpalladium species
1 Review: A. Pommier and J.-M. Pons, Synthesis, 1993, 441.
2 Y. Tamai, H. Yoshiwara, M. Someya, J. Fukumoto and S. Miyano,
J. Chem. Soc., Chem. Commun., 1994, 2281.
3 Review: B. Bosnich, Aldrichim. Acta, 1998, 31, 76.
4 W. Odenkirk, A. L. Rheingold and B. Bosnich, J. Am. Chem. Soc., 1992,
114, 6392; S. Kanemasa, Y. Oderaotoshi, S. Sakaguchi, H. Yamamoto,
J. Tanaka, E. Wada and D. P. Curran, J. Am. Chem. Soc., 1998, 120,
3074.
5 J. A. Davies, F. R. Hartley and S. G. Murray, J. Chem. Soc., Dalton
Trans., 1980, 2246; S. Oi, K. Kashiwagi, E. Terada, K. Ohuchi and Y.
Inoue, Tetrahedron Lett., 1996, 37, 6351.
6 R. Stevenson and J. V. Weber, J. Nat. Prod., 1988, 51, 1215.
7 Only one report was found in literature, which described a formation of
similar d-lactones in the cycloaddition of ketene with ketones: F. G.
Young, J. Am. Chem. Soc., 1949, 71, 1346.
8 Reviews: R. P. Lutz, Chem. Rev., 1984, 84, 205; L. E. Overman, Angew.
Chem., Int. Ed. Engl., 1984, 23, 579.
9 A. F. Noels, J. J. Herman and P. Teyssie´, J. Org. Chem., 1976, 41,
2527.
10 Hagemeyer reported that the boron trifluoride-catalyzed reaction of
ketene with crotonaldehyde 1g gave lactone 3g: H. J. Hagemeyer, Jr.,
US 2,478,388/1949 (Chem. Abstr., 1950, 44, 1132). See also: J. H.
McCain and E. Marcus, J. Org. Chem., 1970, 35, 2414.
11 H. J. Hagemeyer Jr., Ind. Eng. Chem., 1949, 41, 765.
Communication a908709e
74
Chem. Commun., 2000, 73–74