7358 J . Org. Chem., Vol. 61, No. 21, 1996
Bagnell et al.
suggested either that expulsion of water may not have
occurred from 6 primarily via a Saytzeff-type process by
which the more substituted, fully conjugated olefin 5
would have been expected19 or that 2 may have predomi-
nated as a result of allylic strain implicating the aryl
hydroxyl group.20 The unusual selectivity did not extend
to the behavior of 1-phenyl-2-propanol, however. In
water at 250 °C for 1 h, this phenyl-2-propanol under-
went only 5% conversion, but gave a mixture of cis and
trans-phenylpropenes that accorded with the Saytzeff
rule. Heating in dilute H2SO4 also afforded the 1-phen-
ylprop-2-enes primarily, but in higher conversion. These
results thus rendered the second suggestion more likely.
The accumulation of dihydrobenzofuran 7 at higher
temperature, its formation from 1, 2, 5, or 6, and its
relative lack of reactivity at 290 °C all indicated that of
the C9 compounds, 7 was the most thermodynamically
stable product. Even in dilute H2SO4, ring opening of 7
occurred to only a minor extent (entry 13, Table 1).
Nonetheless, this highlighted the complexity of the
equilibria involved, and confirmed that slight adjust-
ments to the reaction conditions could lead to signifi-
cantly different (but reproducible) outcomes.
dihydrobenzofuran 7 (64%), 3 (5%), 4 (6%), and 5 (20%),
a product distribution comparable with that obtained
from 1 in water at 240-250 °C (Figure 1). These results
were consistent with recent findings that addition of
traces of acid can substantially enhance the rates of
reactions in high-temperature water.8
Com p a r ison of Au tocla ve a n d MBR Rea ction s of
1. Many publications on microwave-assisted organic
chemistry have referred to accelerated reactions, and
workers have speculated about the role of microwaves.
However, suggestions of specific activation at a controlled
temperature in homogeneous media now have been
rejected.21 Complex reactions may be more difficult to
explore in this regard though, and enhanced rates of
catalytic transfer hydrogenation were recently attributed
to microwave-assistance in transport processes at the
catalyst and liquid interfaces in a multiphase system.22
In the present work, the main differences between the
sets of results from the MBR and autoclave concerned
the relative rates of diminution of 1; temperatures at
which maximal conversion to specific products occurred;
the extent of conversion to some of these products; and
the relative rates of formation of 7.
The pathway proposed in Scheme 1 appears to satisfy
data in Figure 1, Table 1, and the above discussion.
However, owing to the interrelationship among phenols
2, 4, 5, and to some extent 6, it was difficult to establish
which of these alcohols could give dihydrofuran 7 directly.
The influence of water on the course of reaction was
next assessed. In a control experiment, neat allyl phenyl
ether (1) was heated in the autoclave for 1 h at 290 °C
(entry 3, Table 1). Allylphenol 2 was formed in high
conversion, with only small amounts of 3, 4, 5, and 7
detected. This result indicated that in the absence of
water the phenolic product 2 could not readily promote
autocatalytic reactions such as ring closure to 7 or
isomerization to 5. The product distribution contrasted
markedly with that obtained after 1 was heated at lower
temperatures in water for the same time to afford
products in common with those reported in TFA at 60
°C (see Figure 1, Table 1, and discussion above). This
showed that water behaved primarily as a medium for
Claisen rearrangement at about 200 °C but indicated that
at higher temperature it also played catalytic and
participatory roles.
In the MBR and autoclave reactions of 1 in water, the
(unbuffered) aqueous phase had a pH ranging between
3.5 and 4.5 after completion of heating, cooling, and
extraction of the products. Initially, this acidity was
attributed to the presence of small amounts of unex-
tracted phenolic products. Significantly though, after the
phenolics 2, 5, and 6 were heated individually (see Table
1) and the products were extracted, the pH of the aqueous
phase was higher, in the range 4.8-5.4. The conse-
quences of these differences in pH were investigated by
heating 2 in dilute sulfuric acid at 290 °C (entry 7, Table
1). After extraction, the pH of the aqueous phase was
3.8. In sharp contrast with reactions of 2 at 290 °C in
unacidified water (entry 6 Table 1), only 5% of the
starting material was recovered. Major products were
To elaborate, 95% conversion of ether 1 was achieved
at 230 °C in the MBR, but nearly 20 °C higher in the
autoclave. The greatest concentration of allylphenol 2
was obtained at 200 °C in the MBR, but 15 °C higher in
the autoclave; hydroxyphenol 6 at 225 °C (MBR) and 245
°C (autoclave); and ether 7 at 250 °C (MBR) and 275 °C
(autoclave). The highest conversions observed for 2 were
56% (MBR) and 67% (autoclave); hydroxyphenol 6, 37%
(both methods); ether 7, 72% (MBR) and 66% (autoclave);
isopropenylphenols 4 and 5, 12% (MBR) and 23% (auto-
clave); and phenol (3), 9% (MBR) and 5% (autoclave).
From 230 to 250 °C conversion to 7 increased at the rate
of 2%/°C with the MBR, but only at 1.3%/°C in the
autoclave.
These results indicate that in aspects other than the
greatest extent of conversion of 1 to 2-allylphenol (2), and
to the mixture of propenylphenols 4 and 5, microwave
heating afforded higher conversions. However, re-
arrangement of 1 in water was already known to proceed
in 84% conversion at 240 °C within only 10 min.1 So the
present work has shown that if aqueous conditions are
to be employed for preparative conversion of 1 to products
such as 2, the optimal heating time may not be 1 h.
With conductive heating by conventional autoclave,
thermal gradients develop, and despite stirring, only part
of the sample is at the temperature of the applied heat.
Accordingly, in this work the measured temperature was
dependent upon the positioning of thermal sensors. In
calibrations of the vessel, differences of (7 °C were
observed within the sample.
On the other hand, with the MBR the microwave
energy was primarily absorbed by the sample directly,
and since the vessel was fabricated from polytetra-
fluoroethylene (PTFE), conductive heat losses were mini-
mal. Further, owing to bulk heating by microwaves, the
whole of the sample could be irradiated simultaneously.
These factors, combined with efficient stirring, virtually
eliminate temperature gradients. The differences be-
tween the nature and properties of conventional and
(19) Saunders, W. H., J r. The Chemistry of Alkenes; Patai S., Ed.;
Interscience: New York, 1964; Chapter 2, p 149. Cram, D. J . J . Am.
Chem. Soc. 1952, 74, 2137. Manassen, J .; Klein, F. S. J . Chem. Soc.
1960, 4203.
(20) For reviews on allylic strain, see: J ohnson, F. Chem. Rev. 1968,
68, 375. Hoffmann, R. W. Chem. Rev. 1989, 89, 1841. We thank Dr. S.
Marcuccio for this suggestion.
(21) Rate studies and investigations into “microwave effects” have
been recently summarized and discussed.3
(22) Leskovsek, S.; Smidovnik, A.; Koloini, T. J . Org. Chem. 1994,
59, 7433.