Solid-acid catalysed rearrangement of cyclic α,β-epoxy ketones

Article abstract of DOI:10.1002/(sici)1099-0690(200005)2000:10<1905::aid-ejoc1905>3.0.co;2-o

Various cyclic α,β-epoxy ketones rearranged to α-formyl ketones and/or vicdiones in the presence of catalytic amounts of zeolites and montmorillonite K10. This provides an excellent alternative to conventional homogeneous systems with respect to yields and workup. Differences in product distribution and type of products in the rearrangement of pulegone oxide could be reasonably explained by invoking different pathways for homogeneous and heterogeneous catalysts.

Full text of DOI:10.1002/(sici)1099-0690(200005)2000:10<1905::aid-ejoc1905>3.0.co;2-o

FULL PAPER
Solid-Acid Catalysed Rearrangement of Cyclic α,β-Epoxy Ketones
Jacob A. Elings,[a][°] Hans E. B. Lempers,[a] and Roger A. Sheldon*[a]
Keywords: α,β-Epoxy ketones / Rearrangements / Zeolites / Clays
Various cyclic α,β-epoxy ketones rearranged to α-formyl ke-
tones and/or vic-diones in the presence of catalytic amounts
respect to yields and workup. Differences in product distribu-
tion and type of products in the rearrangement of pulegone
of zeolites and montmorillonite K10. This provides an excel- oxide could be reasonably explained by invoking different
lent alternative to conventional homogeneous systems with
pathways for homogeneous and heterogeneous catalysts.
Introduction
Several products resulting from rearrangement of cyclic
α,β-epoxy ketones are of significant synthetic interest. Vari-
ous 1,2-diones, for example, are widely used as food fla-
vours,[9] and BF3 Et2O-mediated rearrangement of optically
active cyclic α,β-epoxy ketones was a key step in the asym-
metric synthesis of (Ϫ)-frontalin (a pheromone) and (ϩ)-
malyngolide (an antibiotic).[10]
Although they are attractive from the viewpoint of ease
of recovery and recycling, heterogeneous catalysts have
scarcely been used in the rearrangement of α,β-epoxy ke-
tones. Ho and Liu used an aqueous suspension of an ion-
exchange resin (Hϩ form) in the rearrangement of 2,3-
epoxy-3-methylcyclopentanone.[9] Rao and Rao reported
the rearrangement of highly active chalcone oxides to 1,3-
diphenylpropane-1,2-diones at room temperature when sil-
ica gel was used as a catalyst.[11] Zeolites have been used in
the industrially important rearrangement of phenyl glycidic
acid esters, which are structurally closely related to acyclic
α,β-epoxy ketones, to phenyl α-keto carboxylic acid es-
ters.[12,13]
α,β-Epoxy ketones are interesting intermediates in or-
ganic synthesis. They not only undergo the usual reactions
of epoxides,[1] but are also susceptible to several useful reac-
tions owing to the presence of a carbonyl group.[2] Cyclic
α,β-epoxy ketones, which are readily available from the ap-
propriate α,β-unsaturated ketones by epoxidation with an
alkaline solution of a hydroperoxide, are easily converted
into aldehydes and diones upon treatment with an acid. In
the presence of a Lewis acid they cleave at the β-carbon
atom as depicted in Scheme 1.[3]
We have investigated the rearrangement of various cyclic
α,β-epoxy ketones with the use of solid (Brønsted) acids
extensively, and in this paper we will show that, because of
the excellent yields and the facile workup, it provides an
excellent alternative to the classical homogeneous system
with BF3 Et2O.
Scheme 1. Lewis acid catalysed rearrangement of a cyclic α,β-
epoxy ketone
Cleavage at the α-carbon atom is energetically unfavour-
able since this generates partial positive charges on adjacent
atoms.[4] Cleavage is followed by hydrogen migration
(route A) or acyl migration (route B). Acyl migration re-
sults in ring contraction and is favoured over hydrogen mi-
gration when BF3 Et2O is used as a catalyst, particularly if
an additional substituent is present at the β-position.[3,5,6]
Conversely, 1,2-diones were found to be the major products
upon treatment of cyclic α,β-epoxy ketones with aqueous
solutions of hydrochloric acid[7] or sulfuric acid.[8] Probably,
hydrolysis of the epoxide to its corresponding diol precedes
the actual rearrangement under these conditions.
Results and Discussion
Rearrangement of Isophorone Oxide
We chose isophorone oxide as a model compound, which
is easily made from the cheap isophorone. Rearrangement
of isophorone oxide (1) with BF3 Et2O or the heterogen-
eous catalysts gave, as major products, 3,5,5-trimethyl-1,2-
cyclohexanedione (2) and 1,4,4-trimethyl-2-oxocyclopen-
tane-1-carbaldehyde (3) (Scheme 2). The results of the reac-
tion with the various catalysts are shown in Table 1.
[a]
Laboratory of Organic Chemistry and Catalysis, Delft
University of Technology,
Julianalaan 136, 2628 BL Delft, The Netherlands
Fax: (internat.) ϩ 31-15/2781415
E-mail: R.A.Sheldon@stm.tudelft.nl
[°]
Present address: Quest International, Ashford, Kent TN24
0LT, England
Eur. J. Org. Chem. 2000, 1905Ϫ1911
WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
1434Ϫ193X/00/0510Ϫ1905 $ 17.50ϩ.50/0
1905

FULL PAPER
70 to 94% and are therefore comparable to or even better
than those found for the homogeneously catalysed reaction
with BF3 Et2O (Table 1).
H-mordenite-, H-beta- and montmorillonite-catalysed
rapid rearrangement of 1 in refluxing benzene, led, for ex-
ample, to complete conversion in 15 min with H-beta. Sub-
stantial deformylation was observed, however. Since 3 is the
precursor of 4, the total (initial) selectivity to 2 and 3, de-
fined as the sum of the selectivities to 2, 3 and 4, was the
highest for H-beta (96%). Replacement of benzene by the
less polar cyclohexane in the reaction with H-beta partially
suppressed the deformylation. In contrast, montmorillonite
Scheme 2. Rearrangement of isophorone oxide (1) to 3,5,5-trime-
thyl-1,2-cyclohexanedione (2) and 1,4,4-trimethyl-2-oxocyclopen-
tane-1-carbaldehyde (3)
Table 1. Conversion of isophorone oxide (1) and selectivities to
3,5,5-trimethyl-1,2-cyclohexanedione (2), 1,4,4-trimethyl-2-oxocy-
clopentane-1-carbaldehyde (3) and 2,4,4-trimethylcyclopentanone
(4)
Catalyst
Solvent[a] Temp. Time Conv. Selectivity
Ratio
(°C) (min) (%) (%)
(3ϩ4/2) gave little deformylation in chlorobenzene, despite the fact
that this catalyst had high activity (conversion 100% after
15 min) and was used at high temperature. Similarly, no de-
formylation was observed with montmorillonite in refluxing
benzene, despite the high activity (nearly 100% conversion
after 15 min). The observed lack of reaction of 1 in benzene
at room temperature in the presence of this clay further
showed that heating is necessary to establish effective re-
arrangement of 1.
2
3
4
None
B
B
B
80
120
10
0
0
2
0
5
0
0
0
0
0
0
0
0
1
1
1
1
0
0
-
BF3 Et2O
Silica gel
rT
80
100
3
76
0
36
-
120
120
H-ZSM-5 CB
132
80
6
71
16
4.5
10
6.2
9.3
6.4
22
18
27
8.2
Na-Mor
H-Mor
H-Mor
H-Mor
H-Mor
H-Mor[b]
H-Mor[c]
H-Mor[d]
H-Mor[d]
H-beta
B
1200 26
14 63
B
80
120 100
8
81
B
80
1320 100 13 81
CB
CB
B
80
120 98
9
80
132
80
120 100 13 81
120 30
120 100
120 59
1500 99
4
5
3
80
85
78
Replacement of benzene by the more polar chloroben-
zene in the H-mordenite-catalysed rearrangement afforded
a slower reaction rate at the same temperature. The use of
refluxing chlorobenzene (b.p. 132 °C) instead of refluxing
benzene afforded, with H-mordenite, a different product
distribution after the same reaction time but after pro-
longed refluxing in benzene the final product distribution
was similar.
B
80
B
80
B
80
10 78
B
80
120 100 15 56 25 5.6
120 100 16 59 15 4.6
H-beta
CH
B
81
H-beta[b]
NaHY
80
120 100 11 71
9
1
1
7
0
0
2
7.2
3.7
3.3
2.2
6.3
5.5
4.9
B
80
120 63
1380 97
18 67
21 69
NaHY
NaHY
B
80
CB
132
RT
80
120 100 29 58
Mont.K10 B
Mont.K10 B
Mont.K10 CB
1350 40
10 63
120 100 14 78
120 100 16 78
Turnover numbers (defined as the molar ratio between
rearranged epoxide and bulk aluminium) of up to 20 were
132
[a]
found, but since they are a result of the choice of the reac-
tion conditions they are probably amenable to further im-
provement.
B ϭ Benzene, CB ϭ chlorobenzene and CH ϭ cyclohexane. Ϫ
[b]
[c]
[d]
Poisoned with triphenylphosphane. Ϫ Dealuminated. Ϫ Si-
lylated.
Remarkably, all heterogeneous catalysts showed a lower
ratio (3ϩ4)/2 than that observed with BF3 Et2O. The differ-
ences in the ratios (3ϩ4)/2 of the reactions performed with
H-mordenite, H-beta and montmorillonite in refluxing ben-
zene, on the other hand, were small. The results for H-
mordenite in refluxing benzene show that this ratio can
change when the reaction is prolonged beyond 100% con-
version. This is probably due to a slower desorption rate of
2 from the catalyst than with 3 and 4 since prolongation of
the reaction increased the selectivity to 2, while the selectiv-
ities to 3 and 4 remained unchanged. In general, compar-
ison of selectivities of various catalysts should be based on
the same conversions. Moreover, as mentioned above, the
GC analyses were based only on the liquid phase of the
reaction mixture, and, owing to the relatively large amount
of catalyst used with respect to substrate (30 wt%), signific-
ant quantities of product can be adsorbed on the catalyst.
For these reasons, comparison of the ratios (3ϩ4)/2 is
meaningful only when there is a large difference.
No reaction was observed in the absence of catalyst. The
low conversion observed with silica gel is presumably due
to adsorption of a small amount of 1 on the catalyst, since
the GC analyses were based only on the liquid phase of the
reaction mixture. In most cases, 2,4,4-trimethylcyclopen-
tanone (4) was also observed as a product. The formation
of 4 can be explained by deformylation of 3 as depicted in
Scheme 3. Deformylation of 3 can also be effected by treat-
ment with an aqueous sodium hydroxide solution.[14] Prod-
uct 2 can be separated simultaneously, because this α-dike-
tone is alkali-soluble.
Scheme 3. Deformylation of 1,4,4-trimethyl-2-oxocyclopentane-1-
carbaldehyde (3) to 2,4,4-trimethylcyclopentanone (4)
When zeolite NaHY was used as the catalyst, a relatively
Table 1 shows that the acidic form of the zeolites, with large amount of 2 was formed, but 3 was still the major
the exception of H-ZSM-5, and montmorillonite K10 were product. This relatively high selectivity towards the dione
able to rearrange 1 to 2 and 3 with high conversions and was also observed for other cyclic α,β-epoxy ketones,
selectivities. The total selectivities to 2 and 3 ranged from thereby affording a convenient synthesis of various 1,2-di-
1906
Eur. J. Org. Chem. 2000, 1905Ϫ1911

FULL PAPER
ones, albeit in moderate yields. The reason for this special ture see Scheme 5) was formed as the other major product.
property of zeolite NaHY is not yet clear. The larger pore Table 2 shows that the formation of both products (6 and 8)
size might play a role. Moreover, zeolite NaHY is less acidic proceeded with excellent conversions and total selectivities.
than H-mordenite, H-beta and montmorillonite, resulting
in a weaker interaction with the substrate and/or products
and a more facile desorption of 2.
Because the outer surface of zeolites may contribute to
their reactivity,[15] we investigated the effect of some outer
surface treatments on the product distribution in the re-
Scheme 5. Acid-catalysed conversion of 2-hydroxy-2-isopropenyl-5-
arrangement of 1. It was previously shown that treatment
methylcyclohexanone (8) to 2,2,5-trimethylcycloheptane-1,3-dione
of H-mordenite[15] and H-beta[16] with triphenylphosphane (6)
and H-mordenite with tetraethyl orthosilicate[16] resulted in
total deactivation of the outer surface activity. Dealu-
Table 2. Solid-acid catalysed rearrangement of pulegone oxide (5)
mination of mordenite by treatment with nitric acid re-
duced the outer surface activity to approximately 7% of the
original activity, without loss of the activity of the inner-
pore system.[15]
Catalyst
Time Conv.[a] Selectivity
Ratio
6/8
(min) (%)
(%)
6
8
Triphenylphosphane-poisoned H-beta gave the same con-
version after 120 min as that observed for untreated H-beta;
this is consistent with the notion that the reaction takes
place in the micropores. A significant decrease in the forma-
tion of 4 by deformylation was observed, suggesting that
deformylation takes place primarily on the outer surface.
Dealuminated mordenite showed the same conversion
after 120 min as that found with untreated H-mordenite. In
this case, it should be noted that dealumination of morden-
ite has a significant, but minor effect on the ratio (3ϩ4)/2,
because the total selectivities are nearly the same. Silylated
mordenite showed a much lower reaction rate than un-
treated H-mordenite. Initially, the ratio (3ϩ4)/2 was higher,
but it decreased significantly when the reflux time was pro-
longed. This experiment with silylated mordenite illustrates
clearly the slow desorption of product 2 as mentioned
above.
NaHY
HM
1200
240
15
97
67 29 2.3
70 22 3.2
70 15 4.8
98
H-beta
H-beta
100
100
100
120
78
6
12
Montmorillonite K10 120
62 22 2.8
[a]
Performed in refluxing benzene.
In striking contrast, homogeneous rearrangement of 5
with BF3 Et2O in refluxing benzene gave 6 in an isolated
yield of 92%.[5] In this case, it was shown[5] that the reaction
proceeds via fluorohydrin intermediates (see Scheme 6).
Since the flexible fluorohydrins could readily achieve the
required conformation for concerted 1,2-acyl rearrange-
ment, both diastereomers of pulegone oxide resulted in the
formation of 6 as the sole product.
Rearrangement of Pulegone Oxide
Analogously to the rearrangement of isophorone oxide,
epoxide ring opening of pulegone oxide (5) is expected to
preferentially occur at the β-carbon atom; this is further
facilitated by the presence of two methyl groups (tertiary
carbon atom). Subsequent acyl or alkyl migration will re-
sult in the formation of 2,2,5-trimethylcycloheptane-1,3-di-
one (6) or 3,3,6-trimethylcycloheptane-1,2-dione (7), re-
spectively (see Scheme 4).
Scheme 6. Fluorohydrins in the BF3 Et2O-catalysed rearrangement
of 5[5]
Scheme 4. Expected rearrangement pathway of pulegone oxide (5)
to 2,2,5-trimethylcycloheptane-1,3-dione (6) and 3,3,6-trimethylcy-
cloheptane-1,2-dione (7)
Since involvement of fluorohydrin-like intermediates is
precluded with the heterogeneous catalysts, rearrangement
of pulegone oxide induced by solid acids will, therefore,
Indeed, when a diastereomeric mixture of pulegone oxide probably occur via an unfavourable transition state. In such
(cis/trans ϭ 1.7) was refluxed in benzene in the presence of cases, the rate of substituent migration is probably much
various solid acids, 1,3-dione 6 was obtained as the major faster than the rate of rotation to an optimal conformation,
product. However, instead of the expected 1,2-dione 7, 2- and the result is therefore likely to be a divergent package
hydroxy-2-isopropenyl-5-methylcyclohexanone (8, for struc- of products.
Eur. J. Org. Chem. 2000, 1905Ϫ1911
1907

FULL PAPER
Table 2 shows that when the reaction with H-beta (after
2,3-Epoxy-3-methylcyclohexanone (11) bears a close re-
100% conversion) was prolonged, the selectivity to 8 slowly semblance to isophorone oxide. Indeed, rearrangement of
decreased in favour of 6, a tendency that was also found, the former compound afforded similar products to those
to a lesser degree, for the other solids. A possible explana- found for isophorone oxide (see Scheme 8). Table 4 shows,
tion for this phenomenon is depicted in Scheme 5. Pro- however, that lower selectivities and, more strikingly, a
tonation of allylic alcohol 8 by the solid acid may result in much lower reaction rate were obtained with 11 than with
a similar reaction intermediate to that formed in the re- isophorone oxide, upon treatment with H-mordenite in re-
arrangement of pulegone oxide (((5) to 1,3-dione 6, enab- fluxing benzene (conversion of 1: 100% after 60 min). The
ling the latter product to be formed also in an indirect way. lower reactivity of 11 is remarkable because this epoxide is
Such a formation of 6 from 8 implies that the 1,3-dione expected to fit better in a porous structure than 1, due to
is the thermodynamically more stable product. The slower its somewhat smaller size. Apparently, the presence of two
conversion of 8 compared with 5 can be explained by the methyl groups at the 5-position of isophorone oxide has a
fact that it is more difficult to protonate a double bond beneficial effect on the achievement of an optimum con-
than an epoxide oxygen. This may also clarify why H-beta formation for rearrangement. A recent study on the stereo-
with its higher acidity gives a faster conversion of 8 and a chemical aspects of the BF3 Et2O-catalysed rearrangement
higher selectivity to 6 than the other solid acids.
of 5-substituted 2,3-epoxycyclohexanones revealed that
substituents at the 5-position indeed play an important role
in directing the reaction of these conformationally con-
strained molecules.[17]
Rearrangement of Other α,β-Epoxy Ketones
Rearrangement of piperitone oxide (9) in the presence of
various solid acids afforded two stereoisomeric forms of 3-
isopropyl-1-methyl-2-oxocyclopentane-1-carbaldehyde (10)
as major products (see Scheme 7). Like piperitone oxide
itself, the two rearrangement products were very noticeable
by their very pleasant odour.
Scheme 8. Rearrangement of 2,3-epoxy-3-methylcyclohexan-1-one
(11) to 3-methyl-1,2-cyclohexanedione (12) and 1-methyl-2-oxocy-
clopentane-1-carbaldehyde (13)
Table 4. Solid-acid catalysed rearrangement of 2,3-epoxy-3-methyl-
cyclohexan-1-one (11)
Catalyst Solvent
Time
(min)
Conv.
(%)
Selectivity
(%)
Scheme 7. Rearrangement of piperitone oxide (9) to 3-isopropyl-1-
methyl-2-oxocyclopentane-1-carbaldehyde (10)
12
6
10
11
13
54
63
75
HM
HM
HM
benzene
30
240
30
23
68
99
benzene
Table 3. Solid-acid catalysed rearrangement of piperitone oxide (9)
chlorobenzene
Catalyst
Time
Conv.[a] Selectivity 10Ratio
(%)
isomers[b]
(min) (%)
3-Methyl-1,2-cyclopentanedione (15) is a flavour ingredi-
ent of maple syrup and roasted coffee,[18] and has found
widespread use as a food flavour.[9] It should be noted that
1,2-dione 15 largely exists as its enolic tautomer; this pro-
pensity is often observed for 1,2-diones because, in this way,
a higher degree of conjugation is obtained and dipole repul-
sions between the carbonyl oxygens are relieved.[19] Its syn-
thesis has been extensively investigated and one route com-
prises the rearrangement of 2,3-epoxy-3-methylcyclopen-
tanone (14) in the presence of acids (Scheme 9). Because of
the industrial importance of this reaction, we investigated
the potential of H-beta in this reaction.
HM
120
30
6
80
78
70
78
45
1.4
1.6
1.3
1.2
1.0
HM[c]
H-beta
H-beta
97
93
100
14
15
120
Montmorillonite K10 120
[a]
[b]
Performed in refluxing benzene. Ϫ
Determined by gas chro-
[c]
matography. Ϫ Performed in chlorobenzene at 132 °C.
Table 3 shows that the two isomers of 10 were rapidly
obtained in refluxing benzene in good selectivities when H-
beta was used as a catalyst. Interestingly, no detectable
amounts of deformylation products were observed. How-
ever, the use of H-mordenite and montmorillonite K10 in
refluxing benzene was accompanied by low conversions. A
rapid conversion of 9 over H-mordenite was observed only
when the reaction was performed at a higher temperature,
in refluxing chlorobenzene. Obviously, piperitone oxide is
more difficult to rearrange than isophorone oxide; this is
probably due to the relatively bulky isopropyl group at the
6-position, which has an unfavourable effect on the forma-
tion of the transition state for rearrangement.
Scheme 9. Rearrangement of 2,3-epoxy-3-methylcyclopentanone
(14) to 3-methyl-1,2-cyclopentanedione (15)
1908
Eur. J. Org. Chem. 2000, 1905Ϫ1911

FULL PAPER
tents of the solid catalysts were determined by elemental analysis
with inductive coupled plasma atomic emission spectroscopy (ICP-
AES) and atomic absorption spectroscopy (AAS). The samples of
the solid catalysts were prepared by successive addition of 45 mL
water, 2.25 mL 20 wt-% aqueous sulfuric acid and 0.45 mL 40 wt%
aqueous hydrofluoric acid to 30 mg catalyst, which was weighed
into a plastic bottle. The bottle was subsequently shaken overnight
at room temperature. The measurements were recorded on a
PerkinϪElmer Plasma 40 (ICP-AES) or a PerkinϪElmer 1100
(AAS) instrument.
Refluxing of 14 in benzene in the presence of H-beta re-
sulted in rapid conversion of this epoxide (100% conversion
after 15 min), with 15 forming as the major product. Pro-
longation of refluxing resulted in a gradual increase of the
selectivity to 15 (84% after 180 min), again illustrating the
strong adsorption behaviour of 1,2-diones on solid acids, as
mentioned above.
Analogously to the rearrangement of its homologue 11
(see above), acyl migration in 14 might be expected to af-
ford an oxocyclobutanecarbaldehyde derivative. However,
such a product was not observed. Similarly, contraction of
14 to this four-ring derivative did not occur with
BF3 Et2O; 15 was formed in 80% yield.[3]
Origin of the Catalysts: Mordenite (sodium form, Si/Al ϭ 7.3, Al/
Na ϭ 1.0) was obtained from PQ zeolites. The ammonium form of
mordenite (Si/Al ϭ 7, Al/Na ϭ 17) was obtained by ion exchange
of Na-mordenite with 1  aqueous ammonium nitrate at 80 °C for
24 h. Ϫ Silylated mordenite was prepared from 2.5 g H-mordenite,
which was stirred in a solution of 1.5 mmol tetraethyl orthosilicate
in n-hexane (50 mL) at room temperature for 16 h. H-mordenite,
for this purpose, was freshly prepared by the heating of NH4-mord-
enite at 450 °C for 26 h (including heating up at 1 °C/min) and was
then allowed to cool to room temperature by being exposed to the
open air. After centrifugation and being washed with n-hexane, the
zeolite was dried at 120 °C. Ϫ Dealuminated mordenite (Si/Al ϭ
33) was prepared by the reflux of Na-mordenite in 6  aqueous
nitric acid, according to a Dow patent.[20] Ϫ Zeolite beta (Si/Al ϭ
12, Al/Na ϭ 10) was synthesised according to a slightly modified
method of Wadlinger et al.[21] A polypropylene bottle equipped
with a magnetic stirrer bar was charged with sodium aluminate
(8.71 g, Riedel-de Hae¨n, 54% Al2O3, 92.3 mmol Al), 40 wt-% aque-
ous tetraethylammonium hydroxide (82.52 g, 224 mmol) and water
(5 mL). After the bottle was closed, the mixture was stirred in a
boiling water bath for 1 h. The resulting clear solution was trans-
ferred to a teflon insert, which was, prior to its use, successively
cleaned with diluted aqueous hydrogen fluoride (at room temper-
ature), 2  aqueous sodium hydroxide (at 80 °C) and copious water.
Silica sol (218.72 g, Ludox LS, 30 wt%, 1092 mmol) was slowly ad-
ded to this solution, which was stirred mechanically. After the im-
mediately formed thick smooth gel was stirred at room temperature
for one night, the teflon insert was placed into an autoclave and
Conclusions
In conclusion, we have shown that zeolite- and clay-cata-
lysed rearrangement of cyclic α,β-epoxy ketones forms an
attractive synthetic methodology. Total yields and selectivit-
ies are equal to or even better than the corresponding reac-
tions with homogeneous reagents, but sometimes a more
divergent package of products is obtained. Furthermore,
the solid catalysts can be easily recovered and recycled. H-
beta proved to be the most active and versatile catalyst and
outer-surface passivation experiments seem to be consistent
with the reaction taking place in the micropores.
Experimental Section
General: All chemicals used were analytical grade products pur-
chased from commercial suppliers and were used without further
purification. Ϫ All glassware was dried at 140 °C for at least 1 h
prior to use in the catalytic reactions. Ϫ Gas chromatography (GC)
analyses were performed on Varian 3400 and 3600 gas chromato-
graphs with flame-ionisation detectors and Chrompack CP Sil 5 was heated at 150 °C for 166 h. A crystalline material was obtained,
CB wide-bore columns (50 m ϫ 0.53 mm) with nitrogen as carrier
gas. Ϫ Gas chromatography/mass spectrometry (GC/MS) analyses
were performed on a VG 70-SE mass spectrometer equipped with
a CP Sil 5 CB column. The mass spectra were obtained in the
electron impact (EI) mode, with 70 eV as ionisation energy. Ϫ Op-
tical rotations were measured on a PerkinϪElmer P-141 polari-
which was centrifuged and thoroughly washed with water until the
supernatant liquid was pH-neutral. It was then dried over potas-
sium hydroxide under vacuum. This resulted in 89 g beta in the
form of a fine white powder. For complete removal of the template,
the above-synthesized zeolite was subjected to three subsequent cal-
cination steps. First, it was calcined in air to 450 °C for 23 h (heat-
ing rate 1 °C/min). After it was cooled to room temperature, a
1
meter with ethanol as solvent. Ϫ H NMR spectra of solutions in
CDCl3 were recorded on a Varian T-60 spectrometer (at 60 MHz), second calcination in 1% ozone in oxygen at 120 °C for 2 h took
an INOVA-300 spectrometer (at 300 MHz) or a Varian VXR-400S
spectrometer (at 400 MHz). Ϫ 13C NMR spectra of solutions in
CDCl3 were recorded on an INOVA-300 spectrometer (at 75 MHz)
place. Finally, the zeolite was calcined in oxygen at 450 °C for 2 h
(heating rate 1 °C/min). Afterwards, the zeolite was converted into
its ammonium form (Si/Al ϭ 16, Al/Na Ͼ150) by three-fold ion
or a Varian VXR-400S spectrometer (at 100 MHz). The chemical exchange with 1  aqueous ammonium nitrate at 60 °C for 1 h. Ϫ
shifts, δ, are reported relative to tetramethylsilane (TMS).
13C NMR signals were assigned with the use of the attached proton
test method (APT) and correlation tables. Ϫ Column chromato-
graphy was performed with silica gel 60 (particle size
0.063Ϫ0.200 mm) from Merck. The fractions were investigated by
TLC and GC. Ϫ Thin layer chromatography (TLC) was performed
on precoated silica gel plates (Merck silica gel 60F254). The plates
were first observed by UV (254 nm) and then by staining with a
2.5 wt-% aqueous KMnO4/K2CO3 solution, followed by warming
with a blow-drier. Ϫ The solid catalysts were characterised by X-
ray powder diffraction (XRD) on a Philips PW 1877 diffractometer
with Cu-Kα radiation. Ϫ The silicon, aluminium and sodium con-
NH4-ZSM-5 (Si/Al ϭ 45) was kindly supplied by Dr. P. J. Kunkeler,
Delft University of Technology. Ϫ Zeolite NaY (Si/Al ϭ 2.7, Al/
Na ϭ 1.0) was kindly donated by Akzo. The zeolite was exchanged
with 1  aqueous ammonium nitrate at 80 °C for 24 h resulting in
NaNH4Y (Si/Al ϭ 2.6, Al/Na ϭ 2.9). Ϫ Montmorillonite K10 (Si/
Al ϭ 5.1, Al/Na ϭ 45) was purchased from Fluka. The clay was
washed with 1  aqueous ammonium chloride at room temperature
for 23 h to eliminate possible metal impurities (Si/Al ϭ 4.9, Al/
Na ϭ 41).[22] Ϫ All solid catalysts were preactivated at 450 °C in a
calcination oven for 17Ϫ25 h (including heating up at 1 °C/min)
prior to use in the catalytic reactions. By this thermal treatment,
the ammonium forms of the solids were converted into their protic
Eur. J. Org. Chem. 2000, 1905Ϫ1911
1909

FULL PAPER
forms. Silica gel was used without further activating or drying, as
previously described.[11]
δ ϭ 16.9 (CH2), 18.2 (CH3), 18.5 (CH3), 20.0 (CH3), 20.9 (CH3),
21.8 (CH3), 23.0 (CH3), 23.8 (CH2), 26.2 (CH), 27.8 (CH2), 28.5
(CH2), 28.9 (CH), 48.2 (CH), 52.1 (CH), 61.6 (CϪO), 62.5 (CHϪO,
2 ϫ ), 65.7 (CϪO), 208.5 (CϭO), 209.4 (CϭO).
Isophorone Oxide (1): Isophorone oxide was synthesised by the
epoxidation of isophorone with alkaline hydrogen peroxide accord-
ing to a method of House and Wasson.[3,23] Because the crude iso-
phorone oxide may contain a potentially hazardous amount of per-
oxide, which was not reported by House and Wasson, it was treated
with aqueous sodium sulfite, to destroy peroxide impurities. During
this treatment, the presence of sodium hydrogen carbonate was re-
quired to avoid the formation of bisulfite adducts. In a 2 L, three-
necked flask, equipped with a dropping funnel, a stirrer, and a ther-
mometer, was placed a solution of isophorone (111 g, 0.803 mol)
and 30 wt-% aqueous hydrogen peroxide (240 mL, 2.35 mol) in
800 mL of methanol. After the contents of the flask had been co-
oled to 15 °C by means of an ice bath, 66 mL of 6  aqueous
sodium hydroxide was added, dropwise and with stirring, over a
period of 1 h. During this addition, the temperature of the reaction
mixture was maintained at 15Ϫ20 °C. After the addition was com-
plete, the resulting mixture was stirred at the same temperature for
an additional 1 h, and was then allowed to warm to room temper-
ature and stirred overnight. The reaction mixture was then poured
into 1 L of water and was extracted with diethyl ether
(3 ϫ 600 mL). The combined extracts were washed with water
(3 ϫ 800 mL), followed by addition of saturated aqueous NaHCO3
(200 mL) and 20 wt% aqueous Na2SO3 (200 mL). The resulting
mixture was vigorously stirred for one night. After separation of
the water layer, the organic layer was successively washed with
water (800 mL) and saturated aqueous NaCl (400 mL), and dried
over anhydrous Na2SO4. After evaporation of the major part of
the solvent, the residual liquid was distilled. The epoxide was ob-
tained in 65% yield as a colourless oil, b.p. 75 °C/8 Torr. Ϫ
1H NMR (300 MHz): δ ϭ 0.92 (s, 3 H), 1.02 (s, 3 H), 1.42 (s, 3 H),
1.70 (dd, J ϭ 2 Hz, J ϭ 15 Hz, 1 H), 1.81 (2d overlapping, J ϭ
14 Hz, 1 H), 2.08 (d, J ϭ 15 Hz, 1 H), 2.62 (d, J ϭ 13 Hz, 1 H),
3.06 (s, 1 H). Ϫ 13C NMR (75 MHz): δ ϭ 24.1 (CH3), 27.9 (CH3),
30.9 (CH3), 36.1, 42.8 (CH2), 48.0 (CH2), 61.5 (CHϪO), 64.3
(CϪO), 207.9 (CϭO). Ϫ MS; m/z (%): 154 (19) [Mϩ], 139 (26), 126
(15), 111 (12), 97 (29), 83 (100), 69 (54), 56 (20), 55 (37), 43 (29).
2,3-Epoxy-3-methylcyclohexanone (11): Compound 11 was pre-
pared from 3-methyl-2-cyclohexen-1-one according to the proced-
ure described above. Crude yield 37%. The epoxide was used with-
1
out further purification. Ϫ H NMR (60 MHz): δ ϭ 1.45 (s, 3 H),
1.51Ϫ2.77 (m, 6 H), 3.02 (s, 1 H).
2,3-Epoxy-3-methylcyclopentanone (14): Compound 14 was pre-
pared from 3-methyl-2-cyclopenten-1-one according to a similar
procedure to that described above. The epoxide was purified by
distillation. Yield 22%, b.p. 150 °C (bath temperature)/22 Torr.
Catalytic Reactions with Solid Acids: The reactions were carried
out in a 250 mL three-necked, round-bottom flask equipped with
a magnetic stirrer bar and a condenser fitted with a calcium chlor-
ide tube. The solid acid (0.5 g), pre-activated (450 °C, overnight),
was suspended in a solution of the epoxide (10 mmol) and 1,3,5-
tri-tert-butylbenzene (1 mmol, bulky internal standard) in benzene
(50 mL). The reaction mixture was heated to the desired temper-
ature with stirring at 1000 rpm. Aliquots were taken during the
reaction and were analysed by GC. Products were isolated by col-
umn chromatography and were identified by NMR and GC/MS.
Poisoning of the catalyst was performed by stirring the catalyst in
a solution of triphenylphosphane in benzene (ratio triphenylphos-
phane:bulk aluminium ϭ 1Ϫ2) at room temperature for 45 min. In
this case, the reaction was started by the addition of a solution
of the appropriate amounts of isophorone oxide and 1,3,5-tri-tert-
butylbenzene in benzene.
Homogeneous Rearrangement of Isophorone Oxide: The homogen-
eous rearrangement of isophorone oxide was based on a method
of House and Wasson.[3,14] To a solution of isophorone oxide
(32 mmol) in 100 mL of benzene was added BF3 Et2O (32 mmol).
The resulting solution was mixed by swirling and was allowed to
stand for 10 min. After addition of 100 mL benzene, the solution
was washed with water (2 ϫ 100 mL) and saturated aqueous so-
dium chloride solution (50 mL); it was then dried with sodium sulf-
ate. The solution was analysed by GC using an internal standard.
Pulegone Oxide (5, cis/trans ؍ 1.7): Pulegone oxide was prepared
from (R)-(ϩ)-pulegone according to the procedure described above,
affording a cis/trans mixture in a ratio of 1.7 (determined by GC).
Yield 47%. The epoxide was used without further purification. The
assignments cis and trans were based on measurement of the op-
tical rotation of this mixture and known optical rotation values for
Homogeneous Rearrangement of Pulegone Oxide. Synthesis of 2,2,5-
Trimethylcycloheptane-1,3-dione (6): The homogeneous rearrange-
ment of pulegone oxide was based on a method of Bach and Klix.[5]
BF3 Et2O (32 mmol) was added to a solution of pulegone oxide
(16 mmol) in 50 mL of benzene. The resulting solution was refluxed
for 1 min. After addition of 50 mL benzene, the solution was
washed with water (2 ϫ 50 mL) and saturated aqueous sodium
chloride solution (25 mL), and was then dried with sodium sulfate.
Removal of the solvent, followed by chromatography over a short
column afforded 2,2,5-trimethylcycloheptane-1,3-dione (6) as a
1
the separate isomers from ref.[5] Ϫ H NMR (300 MHz): δ ϭ 1.06
and 1.09 (dϩd, J ϭ 6 Hz, J ϭ 5 Hz, 3 H), 1.22 (s, 1.1 H), 1.23 (s,
1.9 H), 1.44 (s, 3 H), 1.70Ϫ2.10 (m, 4 H), 2.14Ϫ2.66 (m, 3 H). Ϫ
13C NMR (75 MHz): δ ϭ 19.0, 19.4, 19.7, 19.8, 20.0, 22.1, 26.4
(CH2), 30.0 (CH2), 30.3 (CH2), 30.8 (CH3), 33.1 (CH2), 34.0 (CH3),
49.5 (CH2), 51.4 (CH2), 63.2 (CϪO), 63.4 (CϪO), 70.1 (CϪO),
70.3 (CϪO), 206.4 (CϭO), 207.5 (CϭO). Ϫ MS; m/z (%, cis): 168
(43) [Mϩ], 153 (100), 125 (27), 111 (34), 98 (15), 86 (41), 83 (34),
70 (44), 69 (26), 67 (26), 55 (27), 43 (47). Ϫ MS; m/z (%, trans):
168 (41) [Mϩ], 153 (100), 125 (20), 111 (35), 98 (11), 86 (44), 83
(30), 70 (35), 69 (23), 67 (23), 55 (25), 43 (45).
1
white crystalline solid in a yield of 92% (purity 98%). Ϫ H NMR
(400 MHz): δ ϭ 1.03 (d, J ϭ 6.6 Hz, 3 H), 1.23 (sϩs, 6 H),
1.45Ϫ1.56 (m, 1 H), 1.94Ϫ2.10 (m, 2 H), 2.32Ϫ2.44 (m, 3 H), 2.53
(dt, J ϭ 2.9 Hz, J ϭ 12.0 Hz, 1 H). The 1H NMR data matched
with those reported.[24]
Ϫ
13C NMR (100 MHz): δ ϭ 20.5 (2 ϫ ),
22.9, 35.1, 35.8, 39.8, 48.4, 61.5, 211.5, 212.5. Ϫ MS; m/z (%): 168
(35) [Mϩ], 153 (6), 140 (35), 125 (100), 124 (24), 123 (43), 111 (23),
97 (62), 84 (25), 83 (60), 70 (92), 69 (59), 67 (32), 55 (40), 43 (49).
Ϫ [α]2D5 °C ϭ Ϫ65.7 (c ϭ 1.17 EtOH), [α]2D4 5 °C (ref.[5])ϭ Ϫ81.8 (c ϭ
1.5 EtOH).
Piperitone Oxide (9): Piperitone oxide was prepared from
piperitone according to the procedure described above. Yield of
crude product 41%. The product was purified by column chromato-
graphy (EtOAc/hexane ϭ 1:5). NMR revealed the presence of two
stereoisomers. Ϫ 1H NMR (300 MHz): δ ϭ 0.83 (d, J ϭ 7 Hz, 3
H), 0.91 (d, J ϭ 7 Hz, 3 H), 1.42 (s, 0.7 H), 1.45 (s, 2.3 H),
Identification of the Reaction Products: After the completion of the
1.50Ϫ2.50 (m, 6 H), 3.05Ϫ3.15 (m, 1 H). Ϫ 13C NMR (75 MHz): catalytic experiments, the various products were isolated from the
1910
Eur. J. Org. Chem. 2000, 1905Ϫ1911

FULL PAPER
reaction mixtures by removal of the solid catalyst (by filtration or
centrifugation) and solvent (evaporation). In some cases, reactions
on a larger scale were performed, without the presence of internal
standard, to obtain a larger amount of a product. Where necessary,
the products were subjected to further purification.
(12). Ϫ 1H NMR (60 MHz): δ ϭ 1.28 (s, 3 H), 1.43Ϫ2.80 (m, 6 H),
9.42 (s, 1 H, CHO).
3-Methyl-1,2-cyclopentanedione (15): 1H NMR (60 MHz): δ ϭ 1.99
(s, 3 H), 2.39 (s, 4 H), 5.77 (br s, 1 H). The 1H NMR data matched
those reported.[9] Ϫ MS; m/z (%): 112 (100) [Mϩ], 97 (11), 84 (27),
83 (33), 69 (57), 56 (21), 55 (33), 43 (38).
3,5,5-Trimethyl-1,2-cyclohexanedione (2): MS; m/z (%): 154 (91)
[Mϩ], 139 (53), 126 (19), 125 (19), 112 (31), 111 (36), 98 (66), 83
(36), 71 (18), 70 (100), 69 (29), 55 (31), 43 (37). The MS data
matched with those reported.[25]
Acknowledgments
We are grateful to the Dutch Innovation Oriented Research Pro-
gramme on Catalysis (IOP Catalysis), who financially supported
the work described in this paper.
1,4,4-Trimethyl-2-oxocyclopentane-1-carbaldehyde (3): 1H NMR
(60 MHz): δ ϭ 1.03 (s, 3 H), 1.15 (s, 3 H), 1.33 (s, 3 H), 1.40Ϫ2.77
1
(sϩm, 4 H), 9.50 (s, 1 H, CHO). The H NMR data matched with
those reported.[26] Ϫ MS; m/z (%): 154 (15) [Mϩ], 139 (5), 126 (48),
[1]
R. E. Parker, N. S. Isaacs, Chem. Rev. 1959, 59, 737Ϫ799.
111 (19), 97 (22), 83 (100), 69 (9), 56 (15), 55 (21), 43 (21).
[2]
J. Gorzynski Smith, Synthesis 1984, 629Ϫ656.
[3]
H. O. House, R. L. Wasson, J. Am. Chem. Soc. 1957, 79,
1488Ϫ1492.
H. O. House, D. J. Reif, R. L. Wasson, J. Am. Chem. Soc. 1957,
79, 2490Ϫ2495.
2,4,4-Trimethylcyclopentanone (4): Compound 4 was obtained by
shaking of an ethereal solution containing 3 with 5  aqueous so-
dium hydroxide (mol ratio OHϪ/3ϭ4) at room temperature for sev-
eral minutes. The organic layer was separated, washed with satur-
ated aqueous NaCl, dried over Na2SO4, and concentrated. Distilla-
tion afforded the deformylation product (4) as a colourless oil, b.p.
[4]
[5]
R. D. Bach, R. C. Klix, Tetrahedron Lett. 1985, 26, 985Ϫ988.
[6]
R. D. Bach, M. W. Tubergen, R. C. Klix, Tetrahedron Lett.
1986, 27, 3565Ϫ3568.
[7]
G. B. Payne, J. Org. Chem. 1959, 24, 719Ϫ720.
1
42.5 °C/8 Torr. Ϫ H NMR (60 MHz): δ ϭ 1.08 (s, 3 H), 1.10 (d,
[8]
´
L. De Buyck, Y. Zi-Peng, R. Verhe, N. De Kimpe, N. Schamp,
J ϭ 7 Hz, 3 H), 1.17 (s, 3 H), 1.30Ϫ2.70 (sϩm, 5 H). Ϫ MS; m/z
(%): 126 (35) [Mϩ], 111 (11), 84 (16), 83 (96), 70 (11), 69 (73), 67
(8), 57 (21), 56 (100), 55 (61), 53 (14), 43 (10).
Bull. Soc. Chim. Belg. 1985, 94, 75Ϫ80.
[9]
T-L Ho, S-H Liu, Synth. Commun. 1983, 13, 685Ϫ690.
M. Asaoka, S. Hayashibe, S. Sonoda, H. Takei, Tetrahedron
1991, 47, 6967Ϫ6974.
[10]
[11]
[12]
2-Hydroxy-2-isopropyl-5-methylcyclohexanone (8): Compound
8
T. B. Rao, J. M. Rao, Synth. Commun. 1993, 23, 1527Ϫ1533.
W. F. Hoelderich, in Catalytic Science and Technology (Eds.: S.
Yoshida, N. Takezawa, T. Ono), Kodansha-VCH, Tokyo, 1991,
Vol. 1, pp. 31Ϫ46.
was separated from 1,3-dione 6 by column chromatography
(EtOAc/hexane ϭ 1:10): Rf (8) Ͻ Rf (6). GC-analysis: tr (8) Ͼ tr
1
(6). NMR revealed the presence of two stereoisomers. Ϫ H NMR
[13]
W. F Hoelderich, N. Goetz, in Proceedings from the Ninth In-
ternational Zeolite Conference, Montreal 1992 (Eds.: R. von
Ballmoos, J. B. Higgins, M. M. J. Treacy), Butterworth-Heine-
mann, Boston, 1993, pp. 309Ϫ317.
(400 MHz): δ ϭ 0.97 (d, J ϭ 7.1 Hz, 2.3 H, CH3), 1.03 (d, J ϭ
6.4 Hz, 0.7 H, CH3), 1.40Ϫ2.76 (m, 7 H), 1.69 (s, 3 H, CH3ϪCϭ
C), 4.05 (br s, 0.8 H, OH), 4.14 (br s, 0.2 H, OH), 5.08Ϫ5.13 (m,
2 H, CH2ϭC). The 1H NMR data were in agreement with those
[14]
[15]
Org. Synth. Coll. Vol. 1963, 4, 957Ϫ959.
E. J. Creyghton, J. A. Elings, R. S. Downing, R. A. Sheldon,
H. van Bekkum, Microporous Mater. 1996, 5, 299Ϫ307.
J. A. Elings, Solid-Acid Catalysed Reactions with Epoxides and
Allyl Aryl Ethers, Ph. D. Thesis, Delft University Press, Delft,
The Netherlands, 1997.
reported.[27,28]
Ϫ
13C NMR (100 MHz): δ ϭ 18.5, 18.6, 18.9, 22.2,
28.8 (CH2), 31.6 (CH2), 32.5 (CH3), 33.4 (CH2), 36.1 (CH3), 36.9
(CH2), 45.1 (CH2), 46.9 (CH2), 80.1 (CϪOH), 80.4 (CϪOH), 114.1
(CH2ϭC), 114.7 (CH2ϭC), 143.9 (CH2ϭC), 144.1 (CH2ϭC), 212.9
(CϭO), 213.0 (CϭO). Ϫ MS; m/z (%): 168 (20) [Mϩ], 153 (14), 140
(9), 125 (55), 124 (54), 123 (7), 111 (45), 97 (100), 84 (51), 83 (41),
70 (21), 69 (85), 67 (18), 55 (32), 43 (74).
[16]
[17]
[18]
[19]
[20]
[21]
[22]
K. Obuchi, S. Hayashibe, M. Asaoka, H. Takei, Bull. Chem.
Soc. Jpn. 1992, 65, 3206Ϫ3208.
M. A. Gianturco, A. S. Giammarino, R. G. Pitcher, Tetrahed-
ron 1963, 19, 2051Ϫ2059.
R. J. Fessenden, J. S. Fessenden, Organic Chemistry, 5th ed.,
Brooks/Cole, Pacific Grove, 1994.
J. G. Lee, J. M. Garces, G. R. Meima, M. J. M. Van der Aalst,
Eur. Patent No. 0433,932 1990, to Dow.
3-Isopropyl-1-methyl-2-oxocyclopentane-1-carbaldehyde (10): Two
isomers (A and B) of 10 were isolated with column chromatography
(EtOAc/ hexaneϭ1:5): Rf (A) Ͼ Rf (B). GC-analysis: tr (A) Ͻ tr
(B). Ϫ Isomer A: 1H NMR (60 MHz): δ ϭ 0.77 (d, J ϭ 6 Hz, 3 H),
0.95 (d, J ϭ 6 Hz, 3 H), 1.00Ϫ3.00 (m, 6 H), 1.30 (s, 3 H), 9.35 (s,
1 H, CHO). Ϫ MS; m/z (%): 168 (4) [Mϩ], 139 (100), 97 (80), 84
(18), 71 (54), 70 (69), 69 (79), 55 (80), 43 (32). Ϫ Isomer B:
1H NMR (60 MHz): δ ϭ 0.70Ϫ1.18 (m, 6 H), 1.20 (s, 3 H),
1.20Ϫ3.00 (m, 6 H), 9.49 (s, 1 H, CHO). Ϫ MS; m/z (%): 168 (2)
[Mϩ], 139 (100), 97 (82), 84 (17), 71 (55), 70 (71), 69 (87), 55 (93),
43 (35).
R. L. Wadlinger, G. T. Kerr, E. J. Rosinski, US Patent No.
28,341 Reissued 1975, to Mobil Oil Corporation.
B. M. Choudary, M. Ravichandra Sarma, K. Vijaya Kumar, J.
Mol. Catal. 1994, 87, 33Ϫ38.
[23]
[24]
Org. Synth. Coll. Vol. 1963,4, 552Ϫ553.
W. Reusch, D. F. Anderson, C. K. Johnson, J. Am. Chem. Soc.
1968, 90, 4988Ϫ4993.
[25]
´
R. Verhe, N. Schamp, L. De Buyck, R. Van Loocke, Bull. Soc.
Chim. Belg. 1975, 84, 371Ϫ380.
[26]
[27]
T. H. Koch, D. A. Brown, J. Org. Chem. 1971, 36, 1934Ϫ1937.
H. Tokuyama, E. Nakamura, J. Org. Chem. 1994, 59,
1135Ϫ1138.
3-Methyl-1,2-cyclohexanedione (12): 1H NMR (60 MHz): δ ϭ
2.00Ϫ2.70 (m, 6 H), 1.87 (s, 3 H, CH3), 6.16 (br s, 1 H, OH). The
1H NMR data matched those reported.[29]
[28]
[29]
K. H. Schulte-Elte, M. Gadola, B. L. Müller, Helv. Chim. Acta
1971, 54, 1870Ϫ1880.
D. J. Callis, N. F. Thomas, D. P. J. Pearson, B. V. L. Potter, J.
Org. Chem. 1996, 61, 4634Ϫ4640.
1-Methyl-2-oxocyclopentane-1-carbaldehyde (13): Compound 13
was separated from 1,2-dione 12 by column chromatography
(EtOAc/hexaneϭ1:5): Rf (13) Ͻ Rf (12). GC-analysis: tr (13) Ͻ tr
Received June 29, 1999
[O99380]
Eur. J. Org. Chem. 2000, 1905Ϫ1911
1911