Synthesis of polyalkylphenyl prop-2-ynoates and their flash vacuum pyrolysis to polyalkylcyclohepta[b]furan-2(2H)-ones

Article abstract of DOI:10.1002/(SICI)1522-2675(20000510)83:5<1022::AID-HLCA1022>3.0.CO;2-Y

A new method for the smooth and highly efficient preparation of polyalkylated aryl propiolates has been developed. It is based on the formation of the corresponding aryl carbonochloridates (cf Scheme 1 and Table 1) that react with sodium (or lithium) propiolate in THF at 25-65°, with intermediate generation of the mixed anhydrides of the arylcarbonic acids and prop-2-ynoic acid, which then decompose almost quantitatively into CO₂ and the aryl propiolates (cf. Scheme 11). This procedure is superior to the transformation of propynoic acid into its difficult-to-handle acid chloride, which is then reacted with sodium (or lithium) arenolates. A number of the polyalkylated aryl propiolates were subjected to flash vacuum pyrolysis (FVP) at 600-650°and 10^-2 Torr which led to the formation of the corresponding cyclohepta[b]furan-2(2H)-ones in average yields of 25-45% (cf. Scheme 14). It has further been found in pilot experiments that the polyalkylated cyclohepta[b]furan-2(2H)-ones react with 1-(pyrrolidin-1-yl)cyclohexene in toluene at 120-130°to yield the corresponding 1,2,3,4- tetrahydrobenz[a]azulenes, which become, with the growing number of Me groups at the seven-membered ring, more and more sensitive to oxidative destruction by air (cf. Scheme 15).

Full text of DOI:10.1002/(SICI)1522-2675(20000510)83:5<1022::AID-HLCA1022>3.0.CO;2-Y

1022
Helvetica Chimica Acta ± Vol. 83 (2000)
Synthesis of Polyalkylphenyl Prop-2-ynoates and Their Flash Vacuum
Pyrolysis to Polyalkylcyclohepta[b]furan-2(2H)-ones
by Matthias Nagel1) and Hans-Jürgen Hansen*
Organisch-chemisches Institut der Universität, Winterthurerstrasse 190, CH-8057Zürich
Dedicated to Manfred Hesse on the occasion of his 65th birthday
A new method for the smooth and highly efficient preparation of polyalkylated aryl propiolates has been
developed. It is based on the formation of the corresponding aryl carbonochloridates (cf. Scheme 1 and Table 1)
that react with sodium (or lithium) propiolate in THF at 25 ± 658, with intermediate generation of the mixed
anhydrides of the arylcarbonic acids and prop-2-ynoic acid, which then decompose almost quantitatively into
CO2 and the aryl propiolates (cf. Scheme 11). This procedure is superior to the transformation of propynoic acid
into its difficult-to-handle acid chloride, which is then reacted with sodium (or lithium) arenolates. A number of
the polyalkylated aryl propiolates were subjected to flash vacuum pyrolysis (FVP) at 600 ± 6508 and 10À2 Torr
which led to the formation of the corresponding cyclohepta[b]furan-2(2H)-ones in average yields of 25 ± 45%
(cf. Scheme 14). It has further been found in pilot experiments that the polyalkylated cyclohepta[b]furan-2(2H)-
ones react with 1-(pyrrolidin-1-yl)cyclohexene in toluene at 120 ± 1308 to yield the corresponding 1,2,3,4-
tetrahydrobenz[a]azulenes, which become, with the growing number of Me groups at the seven-membered ring,
more and more sensitive to oxidative destruction by air (cf. Scheme 15).
1. Introduction. ± In connection with our interest in thermally or photochemically
induced double-bond shifts (DBS) in heptalenes as elements of potential molecular
switches and data storage devices [1][2], we were interested in the synthesis of
polyalkylated heptalene-1,2- and heptalene-4,5-dicarboxylates as starting materials for
heptalenes substituted with chromogenic groups [2]. General observations allow the
conclusion that the stability of heptalenes as [4n]annulenes will increase with the
degree of twisting of the heptalene core that, in turn, is dependent on the number of
substituents at the core, especially at the peri-positions (cf., e.g., [3]). Moreover, it can
be expected that a buttressing effect of vicinally positioned alkyl groups at the
heptalene skeleton will increase the thermal barriers that separate the DBS isomers2),
as well as the two antipodes of the inherently chiral heptalenes [4][5].
Since heptalene-dicarboxylates are best prepared by thermal reaction of azulenes
with dialkyl acetylenedicarboxylates, by a procedure originally developed by Hafner
et al. [7] (see also [4][8]), we were interested in a practicable and variable synthesis of
polyalkylated azulenes as precursors for the corresponding heptalene-dicarboxylates [9].
1
)
)
Part of the diploma thesis of M. N., University of Zurich, 1998.
2
For example, the establishment of the thermal equilibrium of dimethyl 5,6,8,10-tetramethylheptalene-
1,2-dicarboxylate and dimethyl 1,6,8,10-tetramethylheptalene-4,5-dicarboxylate is characterized by
t1/2 (k1 ‡ kÀ1) ˆ 8.5 min at 1008 ((D4)diglyme) [5]; with dimethyl 5,6,7,9,10-pentamethylheptalene-1,2-
dicarboxylate and dimethyl 1,6,7,9,10-pentamethylheptalene-4,5-dicarboxylate with two pairs of vicinal Me
groups, however, the same peri-substitution pattern shows t1/2 (k1 ‡ kÀ1) ˆ 75 min at 1008 (toluene) [6].

Helvetica Chimica Acta ± Vol. 83 (2000)
1023
Until now, practical syntheses of highly alkylated azulenes have not been published.
In a broad and systematic study of the photoelectron (PE) spectra of alkylated
azulenes, Heilbronner, Hafner et al. [10] have also investigated 1,4,5,6,7,8- and
2,4,5,6,7,8-hexamethylazulene. However, no details of their syntheses were given.
One can speculate that a mixture of both azulenes was obtained by the established
Ziegler-Hafner method (cf. [11]) from a pentamethylpyrylium salt and sodium
methylcyclopentadienide. However, our own experience with the reaction between
4-ethyl-2,3,5,6-tetramethylpyrylium perchlorate [12] and sodium cyclopentadienide
taught us that, in such cases, the cyclopentadienide attacks mainly C(4) of the pyrylium
salt [13], perhaps due to maximum steric relief in the addition step, so that the
formation of polyalkylated azulenes by the Ziegler-Hafner method is extremely
inefficient3). Alder and Whittaker have introduced a versatile modification of the
Ziegler-Hafner method, whereby 1-butylpyridinium salts are reacted with sodium
cyclopentadienide in boiling DMF [18]. This variant is perfectly well-suited for the
synthesis of azulenes without substituents at C(4) and C(8) (cf. also [19]), which are not
available by the original pyrylium salt route (cf. [20]). However, the reaction of
hexamethylpyridinium trifluoromethanesulfonate, which is accessible from penta-
methylpyridine [16] and methyl trifluoromethanesulfonate, with sodium cyclopenta-
dienide in boiling DMF gave yields of 4,5,6,7,8-pentamethylazulene in the range of 2%
[17]4). Therefore, the pyridinium-salt modification of the Ziegler-Hafner method is
also not applicable to the synthesis of polyalkylated azulenes.
Fortunately, numerous further syntheses of azulenes, among them a number where
cycloaddition reactions were utilized, have been exploited in recent years (cf. [21]). An
interesting and variable synthesis of azulenes, based on the cycloaddition of cyclo-
hepta[b]furan-2(2H)-ones and enamines or enol ethers, has been developed by Nozoe
et al. (see, e.g., [22]). The principle is shown in Scheme 1. Mainly, five-ring-substituted
azulenes are conveniently prepared by this procedure, since it depends on the
accessibility of the cyclohepta[b]furan-2(2H)-ones, for which Nozoe et al. (loc. cit.)
have found a versatile synthesis in the base-catalyzed condensation reaction of
tropolone derivatives with CÀH acidic compounds such as alkyl carboxylates,
cyanoacetates, malonates, or malononitrile. An example, starting with g-thujaplicin
as the tropolone component, is illustrated in Scheme 2.
However, a disadvantage is that the condensation reaction does not work well with
alkylated tropolones such as 4,6,8-trimethyltropolone [23], as we have found [24]. In
addition, the synthesis of variably substituted tropolones as precursors for the
correspondingly substituted azulenes may become very tedious (cf. [25]). There is,
however, an inspiring other way to synthesize substituted cyclohepta[b]furan-2(2H)-
ones. More than 20 years ago, Trahanovsky et al. have investigated the flash vacuum
pyrolysis (FVP; 6508/10À4 Torr) of phenyl propiolates 2 [26] and found that these
3
)
This is a general experience, as experiments by Kurokawa and Anderson [14], Muth et al. [15a], and
Porshnev et al. [15b] reveal. Moreover, a systematic study of the reaction of pentamethylpyrylium
methanesulfonate [16] with sodium cyclopentadienide showed that the maximum yields of 4,5,6,7,8-
pentamethylazulene are well below 5% [17].
The pentamethylazulene is extremely sensitive to exposure to air [9]. Therefore, it cannot be excluded that
some of the azulene, which is formed by no means in very low yield, is destroyed in the course of the tedious
workup procedure.
4
)

1024
Helvetica Chimica Acta ± Vol. 83 (2000)
Scheme 1
Scheme 2
compounds rearrange cleanly and in good yields into cyclohepta[b]furan-2(2H)-ones 3
(Scheme 3)5). By this way, the authors also obtained 4,6,8-trimethylcyclohepta[b]fur-
an-2(2H)-one (3c) in yields up to 50%. The pyrolysis of o-cresyl propiolate led to the
formation of a ca. 3 :1 mixture of 4- and 8-methylcyclohepta[b]furan-2(2H)-one (4-
Me-3d, R1 ˆ Me, R2 ˆ R3 ˆ H, and 8-Me-3d, R1 ˆ R2 ˆ H, R3 ˆ Me, resp.), an
observation that indicates that the addition of the ethenylidene component of 2d is
slightly favored at the unsubstituted side of the o-cresyl part. Since highly alkylated
phenols are commercially available or can easily be synthesized, Trahanovskyꢀs
5
)
We will not go into the details of this remarkable high-temperature rearrangement. But it may be
mentioned that it is crucially connected to the temperature dependence of the alkyne-ethenylidene
equilibrium (ÀCꢀCÀH ÀCHˆC:), which lies at temperatures < 6008 completely on the side of the
!
alkyne structure (see, e.g., [27]), but is shifted towards the ethenylidene structure at temperatures > 6008
(cf. [28]). At 6508, the ethenylidene component of the phenyl propiolates 2 adds intramolecularly to the
adjacent CˆC bonds of the phenolic part, followed by ring enlargement of the norcaradiene ! tropilidene
type to the observed furan-2(2H)-ones 3.

Helvetica Chimica Acta ± Vol. 83 (2000)
1025
Scheme 3
1, 2, 3
R1
R2
R3
Yield 2 ! 3 [%]
a
b
c
H
H
H
Me
H
H
H
30 ± 45
45 ± 50
45 ± 50
50 ± 60 b)
Me
Me
Me
H
Me
Me
H
d
Me
a) According to Trahanovsky et al. [26].
b) A mixture of 4-Me-3d and 8-Me-3d is formed (see text).
procedure may open a variable and practical pathway to polyalkylated cyclohepta[b]-
furan-2(2H)-ones 3 as attractive precursors for corresponding azulenes6).
In the following sections, we report a new esterification procedure for polyalkylated
phenols with propiolic acid and on the general applicability of Trahanovskyꢀs method
for the synthesis of highly substituted cyclohepta[b]furan-2(2H)-ones 3.
2. Preparation of Aryl Propiolates. ± The established procedure for the synthesis of
these esters is the acylation of the corresponding phenolates with propynoyl chloride in
a two-phase system in the presence of aqueous NaOH (Schotten-Baumann reaction;
see, e.g., [29]). Jung and Buszek described the formation of (2,6-dimethylphenyl)
propiolate (2b) in 94% yield by deprotonation of 1b with NaH in Et2O, followed by the
reaction with propynoyl chloride at À788 [30]. The disadvantage of these procedures is
that propynoyl chloride is not commercially available due to its instability7). It can be
prepared from propiolic acid and PCl5 at 48 in yields of 45 ± 50% [31] (see also [32]). To
remove POCl3 as the second product, it has to be purified by trap-to-trap condensation.
However, the thus-treated propynoyl chloride still contains some POCl3 and also traces
6
)
It is worthwhile to mention that, in the very first of our FVP experiments, M. N. sublimed a small quantity of
crystalline 2c through the red-hot zone, generated with the flame of a Bunsen burner, of a quartz tube (inner
diameter ca. 5 mm) under vacuum (ca. 5 ´ 10À2 Torr). The formation of 3c was at once indicated by the
development of an intensely yellow ring in the cold zone of the quartz tube. It was contaminated only with
some starting material. This pilot experiment convinced us that, in principle, FVP of phenyl propiolates is
an easy-to-handle procedure that should also be applicable to higher alkylated phenyl propiolates.
The acid chloride is a lachrymose, low-boiling compound (58 ± 608/760 Torr) that readily decomposes in air
and daylight. It can be stored for some time in the dark at dry-ice temperature. It decarbonylates already at
the boiling point under normal pressure to chloroacetylene, which, in turn, is a difficult-to-handle
compound.
7
)

1026
Helvetica Chimica Acta ± Vol. 83 (2000)
of HCl, which form with propynoyl chloride 3-chloroacryloyl chloride as the third
component built in the chloride-formation reaction with PCl5.
We tested the ester formation with mesitol (1c) in a number of runs. Following Jung
and Buszekꢀs protocol, we obtained the phenyl ester 2c in average yields of 70 ± 80%
(Scheme 4). Nevertheless, the raw propiolate had to be purified by chromatography on
silica gel, because it was contaminated with the b-chloroacrylate 4 and diphenyl
chlorophosphate 5 as a result of the presence of POCl3 and HCl in the propynoyl
chloride. The comparably expensive propiolic acid, the moderate yields of acid-
chloride formation, and the difficulties obtaining really pure material led us to search
for other ester-forming processes of propiolic acid and alkylated phenols.
Scheme 4
a) 1. 1.1 Mol-equiv. NaH, Et2O, 208; 2. Propynoyl chlorideb), À808 ! r.t.
a) R ˆ 2,4,6-Trimethylphenyl. b) From the reaction of propiolic acid with PCl5 (see text).
Phenyl acrylates have been successfully prepared by reaction of phenols with the
mixed anhydride of acrylic acid and CF3COOH [33]. The latter compound is formed in
an exothermic reaction from acrylic acid and (CF3CO)2O (TFAA) [33]. Mixing of
propiolic acid and TFAA showed no exothermicity (cf. [34]) and the reaction resulted
mainly in the formation of 2,4,6-trimethylphenyl trifluoroacetate (6; Scheme 5). The
addition of TFAA to lithium propiolate followed by addition of 1c gave again mainly
the trifluoroacetate 6, accompanied by small amounts of 2c and some enol ester 7.
Brown and Eastwood prepared phenyl 2-butynoate from but-2-ynoic acid (tetrolic
acid) and phenol in CH2Cl2 in the presence of dicyclohexylcarbodiimide (DCC) and
catalytic amounts of 4-(dimethylamino)pyridine (DMAP) [35]. The same conditions
applied to propiolic acid and 1c led to an average turnover of 50 ± 70% of 1c
(Scheme 6). The expected ester 2c was formed in moderate yields, but, accompanied by
appreciable amounts of (E)-8 and (Z)-8, the Michael adducts of 1c and 2c.
Phenyl carboxylates have also been prepared from carboxylic acids and phenols in
the presence of Ph3P/CCl4 and a tertiary base [36]. However, all reactions performed
with propiolic acid and 1c led within minutes to a blackening of the reaction mixture
and resin formation. N,N'-Carbonylbis[imidazol] has been similarly applied in phenyl-
ester formation [37]. Indeed, the addition of propiolic acid to the coupling reagent,
dissolved in THF, led to the expected CO2 evolution, but, the reaction mixture turned
yellow, and an insoluble gum-like material started to separate before any phenol 1c
could be added.

Helvetica Chimica Acta ± Vol. 83 (2000)
1027
Scheme 5
a) 1. TFAA, 208; 1c in toluene, 208, 36 h. b) 1. LiH, THF, 58, 3 h; 2. TFAA, 208, 1 h. c) 1c, 208, 12 h.
Scheme 6
a) DCC ‡ 1 mol-% DMAP, CH2Cl2, r.t.
a) 30 ± 50% 1c are recovered. b) R ˆ 2,4,6-Trimethylphenyl.
In 1993, Coppola and Damon reported an elegant and efficient synthesis of N-
benzylpropynamide (10), which otherwise is difficult to prepare (Scheme 7) [38]. The
authors formed the mixed anhydride 9 of ethyl carbonic acid and propiolic acid in THF
in situ, which was then reacted with PhCH2NH2 to give amide 10 in 77% yield.
We tried to apply this procedure also to the formation of 2c from propiolic acid and
1c. The reaction of the in situ generated mixed anhydride 9 with 1c led to a weak
Scheme 7
a) 1. LiH, THF, r.t., 18 h; 2. Ethyl carbonochloridate, À108 ! r.t. b) PhCH2NH2, 58 ! r.t., 4 h.

1028
Helvetica Chimica Acta ± Vol. 83 (2000)
Scheme 8
a) See a) in Scheme 7. b) 1c, THF, 20 ± 458, 5 ± 24 h. c) 2,4,6-Me3C6H2OM (M ˆ Li, Na), THF, 20 ± 458, 2 ± 12 h.
evolution of CO2, which could be enhanced by warming of the reaction mixture to
40 ± 458. However, the results of our first experiments were disappointing (Scheme 8).
The expected ester 2c could indeed be observed and obtained in yields of ca. 20%, but
up to 50% of 1c was still present. Interestingly, beside that we also found in the reaction
mixture ca. 30% of a 1:1 mixture of the (E)- and (Z)-isomer of the b-chloroacrylate 4.
Moreover, appreciable amounts of ethyl propiolate (11) were finally identified8). The
situation did not improve when we worked with the sodium salt of 1c in order to utilize
the enhanced nucleophilicity of the phenolate of 1c. In addition to the discussed
products, we found the (E)- and (Z)-isomer of the b-aryloxyacrylate 8, which we had
already identified in the DCC-mediated reaction of 1c and propiolic acid (Scheme 6).
At the end, the mixed carbonate 12 appeared also in the reaction mixture.
The adaptation of the Coppola-Damon procedure to the preparation of aryl
propiolates was initially not very promising. Nonetheless, it was evident from a
thorough review of the origin of all observed side products that most were secondary
products from Michael-type addition of the initially formed 2c. Only ethyl propiolate
(11) and the mixed carbonate 12 were exceptions. The appearance of 11 in the reaction
mixture finally proved to be the key step in the aryl-ester formation. At the moment
when the mixed anhydride 9 is attacked by the phenolate of 1c at the desired position
under formation of 2c, EtOÀ ions enter the scene and become, during the course of the
reaction, in an autocatalytic manner, competitors of the phenolate ions (Scheme 9). In
other words and in view of earlier work on mixed anhydrides (see later), the reaction
would be likely to run much cleaner if the EtO residue in the mixed anhydride were to
be replaced by the corresponding aryloxy moiety of the phenol to be esterified with
propiolic acid (Scheme 10). The only by-products that could also be formed in analogy
8
)
In our first experiments with the Coppola-Damon system, we had overlooked the formation of 11 due to its
volatility (b.p. 117± 120 8/760 Torr). However, the mass balance of the identified products accounted only
for ca. 60%. Thereafter, we observed that the THF, distilled off from the reaction mixture, exhibited a
pugnant smell, which could finally be attributed to the ester 11 (for details of the analytical procedures see
Exper. Part).

Helvetica Chimica Acta ± Vol. 83 (2000)
1029
Scheme 9
Scheme 10
to 12 would be diaryl carbonates. The corresponding aryl carbonochloridates 14,
necessary for the mixed-anhydride formation, could be prepared in excellent yields
from COCl2 in toluene in the presence of N,N-dimethylaniline (Scheme 11 and Table 1)
(cf. [39][40]). When these carbonochloridates were added to a suspension of lithium
propiolate, generated from propiolic acid and LiH in THF 9), and the mixture was
heated at 25 to 658, a gentle CO2 evolution took place, and the pure aryl propiolates 2
could be isolated in superb yields (Scheme 11 and Table 2). In only one case did our
procedure fail: [2,4,6-tri(tert-butyl)phenoxy]carbonyl chloride (14k) could be obtained
in 75% yield (cf. [41]), but no reaction was observed with sodium propiolate9) in THF.
The yield of the aryl propiolates dropped below 80% in three other cases (2g, 2h, and
2l) due to the accompanying formation of diaryl carbonate. These observations clearly
indicate that propiolate formation from the mixed anhydrides 13 generated in situ is
governed mainly by steric factors. However, the leaving group quality of the aryloxy
9
)
We found in later experiments that it is much easier to work with sodium propiolate, which can be prepared
by neutralization of aqueous solutions of propiolic acid with concentrated aqueous NaOH. The removal of
H2O and drying in high-vacuum leads to a cream-colored salt whose color changes slowly to pale violet in
the air and on exposure to daylight. However, an ivory color persists when the salt is protected against light
and stored in the dark at À208. The deprotonation of propiolic acid in THF with LiH has the disadvantage
that the lithium propiolate forms as a suspension in THF, and 12 ± 18 h are needed for the completion of the
deprotonation reaction with the suspended LiH.

1030
Helvetica Chimica Acta ± Vol. 83 (2000)
Scheme 11
a) 1. 1.3 Mol-equiv. COCl2 (20% in toluene), r.t.; 2. 1 mol-equiv. N,N-dimethylaniline, 0 ! 58, 5 ± 24 h. b) 1 Mol-
equiv. HCꢀCÀCOOM (M ˆ Li, Na), THF, 25 ± 658, 2 ± 12 h.
Table 1. Preparation of Aryl Carbonochloridates 14 a)
1, 14
R1
R2
R3
R4
R5
Reaction
Purification c)
Yield [%]
Reference
yield [%]
time [h] b)
c
e
f
g
h
i
Me
Me
Me
H
H
Me
Me
H
Me
H
H
Me
Me
Me
H
18
18
18
15
8
24
2.5
12
D
C
93
95
93
90
77
ca. 90
7
85
95.5 [39]
±
±
±
Me
Me
Me
Et
H
Me
Me
Me
Et
H
D ‡ C
D
H
H
D
±
i-Pr
t-Bu
H
H
i-Pr
t-Bu
H
D ‡ C d)
C d)
D ‡ C
98 [39]
ca5. 50 f) [41a]
52 [39]
e
k
l
H
t-Bu
H
H
)
g
ÀC4H4À )
a) According to Scheme 11. b) After addition of N,N-dimethylaniline. c) D ˆ bulb-to-bulb distillation, C ˆ
crystallization at 48. d) Crystallization at À208. e) Deprotonation with NaH in THF (cf. [41a]). Some not
reacted 1k could not be separated from 14k. f) As intermediate; via the Li salt a quantitative yield of 14k has
been reported [41b]. g) Naphthalen-2-yl compound.
moiety in the mixed anhydrides, as expressed by the pkA values of the corresponding
phenols, seems to play a role, too. Otherwise, the enhanced carbonate formation in the
case of naphthalen-2-ol (1l; pkA(H2O, 258) ˆ 9.51) in comparison with 3,4,5-trimethyl-
or 3,5-diethylphenol (1g and 1h, resp.; pkA(H2O, 258) of 1g ˆ 10.25), which all have
non-substituted o-positions, is difficult to understand. All our observations are in
agreement with a controlled autocatalytic decomposition of the mixed anhydrides 13
initiated by any nucleophile that attacks their carbonic part and liberates the
corresponding aryloxides as chain reactants (Scheme 10). The nucleophilic chain
reaction is interrupted by diaryl-carbonate formation. Nevertheless, propiolate ions are
generated by this process that can ignite again the nucleophilic chain process.
Moreover, according to our reaction protocol, the mixed anhydrides 13 are formed in

Helvetica Chimica Acta ± Vol. 83 (2000)
Table 2. Preparation of Aryl Propiolates 2 a)
1031
14, 2
R1
R2
R3
R4
R5
Reaction
Reaction
temp. [8]
Yield [%]
time [h] b)
a
c
e
f
g
h
i
H
H
H
H
H
H
H
2
18
24
15
2.5
2
25 ! 40
25 ! 45
25 ! 65
25 ! 65
25 ! 40
30 ! 55
25 ! 65
no reaction
25 ! 65
70 c)
90
91
Me
Me
Me
H
Me
Me
H
Me
H
Me
Me
Me
H
Me
Me
Me
Et
H
Me
Me
Me
Et
H
95
75 c)
72 c)
85
H
H
i-Pr
t-Bu
H
H
t-Bu
i-Pr
t-Bu
H
18
k
l
H
H
H
d
ÀC6H4À )
4
36 e)
a) According to Scheme 11. b) After crystallization and/or distillation; not optimized. c) 20 ± 30% of the
corresponding diaryl carbonate was also isolated. d) Naphthalen-2-yl compound. e) 40% of di(naphthalen-2-yl)
carbonate was also isolated.
situ from the corresponding (aryloxy)carbonyl chlorides and propiolate ions, and,
hence, propiolate and chloride ions as nucleophiles are present during the whole
reaction in molar proportion, whereby progressively propiolate ions are exchanged by
chloride ions.
The mechanistic picture of the autocatalytic decomposition of the carboxylic
carbonic anhydrides 13 that we have sketched here is in agreement with earlier
investigations and observations (cf. [42])10), in particular with those of Windholz [44]
and Green [45], who found that mixed anhydrides of type 9, composed of carboxylic
acids with electron-withdrawing groups, such as in cyano-, (p-nitrophenyl)- or
trifluoroacetic acid, and alkylcarbonic acids give, on gentle heating, a nearly
quantitative evolution of CO2 with formation of the corresponding alkyl carboxylates
in 80 ± 95% yield. On the other hand, steric crowding in the a-position of the carboxylic
part, e.g., diphenylacetic acid, favors formation of the symmetric carboxylic anhydrides,
and the corresponding esters are the minor part of the reaction mixtures [44]. Heating
of pure mixed anhydrides of type 9 in the absence of external nucleophiles also reduces
the ester formation in favor of the generation of the symmetric carboxylic anhydrides
[44][46]. Added tertiary amines act as catalysts and lower the decomposition
temperatures without markedly affecting the product composition [46b]. Moreover,
Tarbell and Longosz [46b] have demonstrated that the purely thermal reaction of the
mixed anhydride of benzoic and (R)-2-octylcarbonic acid takes place with complete
retention of configuration with respect to the formed (oct-2-yl) benzoate as well as
di(oct-2-yl) carbonate that was built as minor product. The observed retention of
10
)
The chemistry of mixed carboxylic carbonic anhydrides can be traced back to the time of Alfred Einhorn
[43].

1032
Helvetica Chimica Acta ± Vol. 83 (2000)
configuration excludes any intramolecular cyclic transition state for ester formation as
has been proposed [44].
3. Preparation of Highly Alkylated Phenols. ± A number of highly methylated
phenols such as 1c, 1f, or 1g are or have been commercially available (see Exper. Part).
Others, such as pentamethylphenol (1e) are available in only small quantities at
Scheme 12
a) CH2O (30% in H2O)/HCl, AcOH, 558 [47] or MeOCH2Cl/SnCl4, CH2Cl2, r.t. [48]. b) H2 (Pd/C), MeOH [47]
or LiAlH4/THF [49].
comparably high prices. The established procedure for the preparation of 1e is the
bis(chloromethylation) of mesitol (1c), followed by hydrogenolysis of the two ClÀC
bonds (Scheme 12).
Wegler and Regel performed the chloromethylation reaction in glacial AcOH with
HCHO in excess (30% aqueous solution) and gaseous HCl at 558 [47]. The authors
reported yields of 16 of ca. 45% and mentioned no further detailed difficulties in the
purification of 16. Some years later, the chloromethylation reaction of 1c has been
described with chloromethyl methyl ether in CH2Cl2 and SnCl4 as catalyst [48]. The
bis(chloromethylated) phenol 16 was obtained in a yield of 65%.
The hydrogenolysis of 16 in MeOH in the presence of Pd/C as catalyst led to
pentamethylphenol (1e) in a yield of 58%. However, the reductive cleavage of the
ClÀC bonds in 16 with LiAlH4 in THF seems to be more effective since a yield of 1e of
95% has been reported by H. Schmid and co-workers [49].
We tried to substitute the gaseous HCl in the protocol of Wegler and Regel with 39%
aqueous HCl. However, GC/MS analyses of the reaction mixture showed that, with
increasing reaction time, growing amounts of several highly substituted benzyl phenyl
ethers were also formed and could not be removed chromatographically in a
satisfactory manner. Hydrogenolysis of the whole reaction mixture as such led to
only small amounts of 1e. Also the stepwise preparation of 1e via the isolated
3-chloromethylated phenol 15 did not markedly improve the yields of 16. Therefore, we
searched for other procedures for the preparation of 1e from 1c.
Recently, the bromomethylation reaction of benzene derivatives has been redis-
covered (cf. [50][51])11). The reaction is performed in glacial AcOH, containing HBr,
11
)
In a footnote A. W. and R. H. Van der Made [50] remarked that the bromomethylation reaction for benzene
derivatives is not well-known very probably due to a mistranslation. Darzens, who was one of the first to
investigate the bromomethylation of benzene derivatives already in 1936 [52], reported in French language
that the yields are much better than in chloromethylations. Fuson and McKeever translated this statement
incorrectly in their review entitled ꢁChloromethylation of Aromatic Compoundsꢀ [53], and from this time
on, the wrong statement was perpetuated in the later literature.

Helvetica Chimica Acta ± Vol. 83 (2000)
1033
with paraformaldehyde as the C1 source. These conditions could also be applied for the
bisbromomethylation of mesitol (Scheme 13). With dependence on the reaction time
and temperature, the mono- or disubstituted phenol, 17 and 18, respectively, is formed.
We could not suppress the formation of up to 20% of the acetate of 1c, which made the
purification of the phenol 18 more difficult (see Exper. Part).
Scheme 13
a) (CH2O)n, HBr in AcOH, 50 ! 858, 7 h. b) H2 (Pd/C), AcOH/AcONa (or N,N-dimethylaniline).
The hydrogenolysis of 18 could be performed in the presence of Pd/C in EtOH,
AcOEt or AcOH. Without any further addition, the uptake of H2 slowed distinctly after
a third of the calculated volume of H2 had been consumed. However, it could be started
again by addition of AcONa12). Under these conditions, 1e was obtained in 70 ± 85%
yield.
We also prepared 1e in 65% yield by Friedel-Crafts alkylation of 1c in boiling MeI in
the presence of AlCl3 (see Exper. Part). These conditions, however, could not be
applied to the synthesis of 3,5-diethyl-2,4,6-trimethylphenol. The reaction of 1c in EtBr
in the presence of AlCl3 at ambient temperature led after several hours to a reaction
mixture from which a crystalline phenol with the expected mass was obtained.
1
Nevertheless, the H- as well as the 13C-NMR spectrum of this compound exhibited
three s signals for three Me groups, as well as signals for two non-equivalent Et groups,
which, for symmetry reasons, excluded the structure of a 3,5-diethyl-2,4,6-trimethyl-
phenol13).
4. Formation of Cyclohepta[b]furan-2(2H)-ones. ± We performed the FVP
reactions of the aryl propiolates as pilot experiments in an improvised apparatus
consisting of an horizontally mounted, 40-cm-long quartz tube with an outer diameter
of 10 mm and filled in its middle part to a length of 8 ± 12 cm with quartz splinters, kept
in place by loose plugs of quartz wadding at each end. One end of the quartz tube was
linked to a small evaporation bulb and the other end to a cooling trap for the
pyrolysate. The middle part of the quartz tube was heated with an oven for combustion
analyses. The temperature in the hot zone of the filled quartz tube was controlled with
an electronic digital thermometer and kept at 650 ± 6708 during the pyrolysis
experiments. The pressure in the apparatus of ca. 5 ´ 10À2 Torr was maintained with
12
)
)
In later experiments, we used N,N-dimethylaniline as base for the neutralization of the HBr formed.
If we allow a complete permutation of the Me and Et substituents as the result of Jacobsen-type
rearrangements four unsymmetrically substituted diethyltrimethylphenols are possible.
13

1034
Helvetica Chimica Acta ± Vol. 83 (2000)
an oil pump. The dimensions of the apparatus allowed the pyrolysis of 0.5 ± 2 g of the
aryl propiolates.
The six cyclohepta[b]furan-2(2H)-ones that we prepared with this improvised FVP
apparatus are shown in Table 3. All FVP reactions were performed several times, and
the range of observed yields is also indicated in Table 3. The formation of furanones 3a
Table 3. Formation of Highly Alkylated Cyclohepta[b]furan-2(2H)-ones 3 via FVP (at 650 ± 6708/0.01 Torr)
Rn in 2
No.
Rn in 3
No.
Observed range of
yields [%] a)
None
2a
2c
2e
2f
2g
2h
none
3a
3c
3e
3f
3g
3h
35 ± 50
30 ± 50
15 ± 40
25 ± 45
30 ± 45
35 ± 45
2,4,6-Me3
2,3,4,5,6-Me5
2,3,4,6-Me4
3,4,5-Me3
3,5-Et2
4,6,8-Me3
4,5,6,7,8-Me5
4,5,7,8-Me4
5,6,7-Me3
5,7-Et2
a) After chromatography on silica gel and crystallization.
and 3c has already been studied by Trahanovsky et al., and yields of 30 ± 50% have
been reported (Scheme 3) [26]. We obtained similar yields with the simplified
apparatus. The range of yields that we observed in the formation of the new
cyclohepta[b]furan-2(2H)-ones can, therefore, be regarded as representative.
The general spectral features of a large number of simply substituted cyclo-
hepta[b]furan-2(2H)-ones have already been reported by various authors (see, e.g.,
[22][26][54][55]), but, to the best of our knowledge, not in a systematic and
comparative way. Since we had new furanones 3 with an increasing number of alkyl
substituents at the seven-membered ring in our hands, we studied their spectroscopic
properties in greater detail.
4.1. Structural and Spectroscopic Properties. All furanones possess an almost planar
structure, with the exception of the pentamethyl derivative 3e. Its X-ray crystal-
structure analysis revealed a slight deviation from planarity with ring-torsion angles of
7.4(4)8, À16.3(4)8, and 10.9(4)8 at the fragment C(4)ˆC(5)ÀC(6)ˆC(7) [9]. More-
over, the bonds C(5)ÀC(6) and C(6)ˆC(7) represent with 146.2(3) and 138.7(3) pm,
respectively, the longest s and p bonds in the molecule, and also exhibit the largest
bond-length deviations from the almost perfectly planar parent structure 3a
(C(5)ÀC(6): 142.4(3) pm and C(6)ˆC(7): 134.9(3) pm [9]14).
4.1.1. UV/VIS Spectra. The spectra of the parent compound 3a and its pentamethyl
derivative 3e in cyclohexane and in MeCN are displayed in Figs. 1 and 2. The
absorbance data of all measured furanones 3 are collected in Table 4. The basal
structure 3a clearly exhibits in cyclohexane three spectral regions defined by a broad,
14
)
We have redetermined the X-ray crystal structure of 3a, which was first published by Saseda in 1959 [56].

Helvetica Chimica Acta ± Vol. 83 (2000)
1035
Fig. 1. UV/VIS Spectrum of cyclohepta[b]furan-2(2H)-one (3a): a) in cyclohexane; b) in MeCN
vibrationally structured absorption band in the range of 512 ± 414 nm. This is followed
by a strong and broad absorption band at 350 ± 390 nm with a spectral half-width of ca.
60 nm, fully separated from a third strong absorption band at 252 nm. The spectral
habitus of 3a is markedly changed in MeCN. The broad and vibrationally structured

1036
Helvetica Chimica Acta ± Vol. 83 (2000)
Fig. 2. UV/VIS Spectrum of 4,5,6,7,8-pentamethylcyclohepta[b]furan-2(2H)-one (3e): a) in cyclohexane; b) in
MeCN
absorption band is hypsochromically shifted by ca. 23 nm, whereas the strong
absorption band centered at ca. 370 nm has still a half-width of ca. 60 nm, but with a
slight bathochromic shift of ca. 7nm of the gravity of the band. The third and strongest
absorption band appears more or less at the same place as in cyclohexane with almost
no alteration in its habitus. The UV/VIS spectra of the pentamethyl form 3e are much

Helvetica Chimica Acta ± Vol. 83 (2000)
1037
simpler than those of 3a. The vibrationally structured band at the longest wavelength
region is no longer recognizable, neither in cyclohexane nor in MeCN. The second
strong absorption band appears symmetrically centered and with a spectral half-width
of 66 and 61 nm at 395 and 405 nm in cyclohexane and MeCN, respectively.
Table 4. UV/VIS Data of the Cyclohepta[b]furan-2(2H)-ones 3 a)
Substituents in 3
No.
Transitions
n ! p*
p ! p* (1)
p ! p* (2)
Minima
None
3a
489 (sh, 150)
457 (sh, 610)
430 (sh, 1500)
387(12550)
373 (12970)
369 (sh, 12810)
351 (sh, 8660)
387(14700)
256 (sh, 19160)
251 (20520)
225 (12720)
381 (11740)
289 (520)
229 (12350)
3a b)
3a c)
488 (sh, 190)
453 (sh, 780)
427(sh, 1980)
258 (sh, 20960)
251 (23150)
223 (14350)
380 (13650)
289 (440)
230 (13300)
373 (14660)
370 (sh, 14340)
351 (sh, 9220)
389 (10680)
367 (13030)
362 (sh, 12030)
352 (sh, 10220)
512 (220)
474 (710)
443 (1350)
414 (2240)
259 (sh, 20130)
252 (21850)
248 (sh, 20960)
238 (sh, 17570)
503 (100)
467 (430)
436 (1050)
411 (2180)
384 (10010)
290 (1100)
211 (10280)
311 (1220)
239 (16170)
4,6,8-Me3
3c
± d)
392 (16470)
368 (sh, 12350)
266 (sh, 21320)
256 (24360)
231 (18310)
5,6,7-Me3
3g
ca. 490 (sh, ca. 300)
396 (17400)
385 (sh, 15780)
271 (sh, 15200)
260 (22680)
300 (480)
221 (11440)
458 (sh, 1340)
± d)
236 (sh, 16300)
272 (sh, 20600)
265 (24220)
232 (17120)
277 (sh, 12000)
265 (16900)
4,5,7,8-Me4
4,5,6,7,8-Me5
3f
397(15680)
384 (sh, 14000)
302 (410)
243 (13570)
215 (10540)
313 (590)
254 (15910)
221 (8930)
312 (990)
3e
430 (sh, 7700)
±
406 (14330)
245 (16900)
3e c)
395 (13080)
368 (8900)
277 (sh, 11330)
261 (sh, 17830)
248 (19410)
221 (10090)
a) In MeCN if not otherwise stated; l [nm]; in parentheses, e. b) In MeOH. c) In cyclohexane. d) Weak tailing
recognizable.
All these spectral features, which are also reflected in the UV/VIS spectra of the
other furanones in MeCN (cf. Table 4), characterize the longest-wavelength absorption
as an n ! p* transition at the CˆO group. The vibrational splitting of ca. 1570 cmÀ1
corresponds quite well with the ground-state vibration of the CˆO group of 3a at
1747 cmÀ1 (cf. Table 5) and indicates the CˆO bond weakening in the excited state. In
agreement with this assignment is the negative solvatochromism of the absorption band
and the dependence of its complete disappearance on the increasing degree of alkyl
substitution at the seven-membered ring of 3 (cf. Table 4). The second strong
absorption band has to be regarded as a p ! p* transition (1) with strong charge-

1038
Helvetica Chimica Acta ± Vol. 83 (2000)
transfer character, i.e., electron density from the CˆO region, especially from C(3)
(see later), is moved to the seven-membered ring in the excited state, in agreement with
SCF calculations of 3a. The expected positive solvatochromism for such an electronic
transition is indeed observed. Moreover, the successive addition of Me groups at the
seven-membered ring has no great influence on the position of this absorption band,
which is in agreement with the observation that the hyperconjugative, electron-
donating properties of the Me groups reduces the electron-accepting character of the
seven-membered ring in the excited state. However, a bathochromic shift of ca. 10 nm
is observed when the unsubstituted position at C(6) in 3f is also occupied by a Me
group, due to the already discussed non-planarity of the seven-membered ring in 3e
that raises the ground-state energy of 3e in comparison to 3f and the other furanones,
so that 3e profits more than the other furanones from the steric relief in the excited
state caused by bond elongation. The magnitude of the absorption coefficients e can
serve as a second indicator for the change in the ground-state energy of the furanones in
going from 3a via 3c to 3e. It increases, especially for the third absorption band at ca.
250 ± 260 nm, up to 24000 for 3c and 3f, but decreases to ca. 2/3 for the pentamethyl-
furanone 3e, an effect that can again be ascribed to the non-planarity of this compound
affecting its optimal conjugation.
Table 5. IR Data of the Cyclohepta[b]furan-2(2H)-ones 3 a)
Substituents in 3
No.
nÄ [cmÀ1
]
None
3a
1780m b)
1747s
1604s
1540s
1513s
1262s
1236s
1702m b)
1732s
4,6,8-Me3
3c
3g
3f
3e
1596m
1599s
1540m
1518m
1490m
1502w
1506s
1473s
1481m
1454s
1702s b)
1732s
5,6,7-Me3
1277m
1223m
1230m
1690m b)
1738s
4,5,7,8-Me4
4,5,6,7,8-Me5
1571m
1556m
1715s
1732s b)
1716s
1240m
a) KBr pellets; only the most important bands in the region of 1800 ± 1000 cmÀ1 are reported; s ˆ strong, m ˆ
medium; w ˆ weak. b) Shoulder.
Finally, the third absorption band can be regarded as being due to a more or less
pure p ! p* transition (2). The negligible solvent dependence of this band speaks for
this assumption.
4.1.2. IR Spectra. We recorded the IR spectra of the furanones in KBr pellets. The
most intense bands for the spectral region of 1800 ± 1000 cmÀ1 are collected in Table 5.
The picture of this region is more or less the same, with the most intense band for
nÄ (CˆO) at 1780 ± 1716 cmÀ1. The intensities of the other skeletal vibrations between
1600 and 1200 cmÀ1 may vary substantially with dependence on the pattern and degree
of substitution at the seven-membered ring. The general trend is that all bands are
shifted to lower wave numbers with the increase of the number of alkyl substituents at
the seven-membered ring, whereby the unsubstituted 3a and its pentamethyl derivative

Helvetica Chimica Acta ± Vol. 83 (2000)
1039
3e can again be regarded as the two borderline cases. Also substitution at C(3) does not
alter this vibrational pattern as is evident from the IR spectrum of 3-phenyl-
cyclohepta[b]furan-2(2H)-one [24][57].
As a rule, all furanones show at least two vibrational bands in the nÄ (CˆO) region
that may appear as two distinct bands as in the case of 3f or the previously mentioned
3-phenylfuranone derivative. Already, Nozoe et al. pointed out [57] that this band-
splitting resembles those for nÄ (CˆO) of a,b-unsaturated five-ring lactones [58].
However, this splitting has been ascribed to a Fermi resonance coupling with a
vibration involving the a-CÀH bond, because a-substituted a,b-unsaturated g-lactones
do not exhibit this splitting. Since the 3-phenylfuranone derivative displays also two
separated bands at 1736 and 1716 cmÀ1, the splitting in the furanones should have
another cause. Nevertheless, the position of the main band is dependent on the degree
of substitution and is shifted from 1747 cmÀ1 for 3a to 1716 cmÀ1 for its pentamethyl
form 3e. Similar shifts are observed for the other reported bands, which should be due
to skeletal vibrations of the cyclohepta[b]furanone-ring system.
Table 6. 1H-NMR Data of the Cyclohepta[b]furan-2(2 H)-ones 3
Substituents in 3
No.
Chemical shift [ppm] a)
HÀC(3)
H/RÀC(4)
H/RÀC(5)
H/RÀC(6)
H/RÀC(7)
H/RÀC(8)
None
3a
3c
3g
3f
3e
3h
5.76 b)
5.572.32
5.49 c)
5.56 d)
5.53
7.29
6.93
7.23
2.22
2.22
7.10
7.02
6.81
6.80
2.19
6.88
2.25
6.62
6.99
6.94
4,6,8-Me3
5,6,7-Me3
4,5,7,8-Me4
4,5,6,7,8-Me5
5,7-Et2
2.33
2.42
2.39
2.35
2.343
2.40
6.90 d)
2.40
2.350
2.272.23
2.57/1.25
5.61 f)
2.56/1.25
6.82 g)
b
a) In CDCl3 at 300 and 600 MHz (3a), respectively. ) As d, J(3,8) ˆ 1.2 Hz; other couplings [Hz]: 3J(4,5) ˆ
11.4, 4J(4,6) ˆ 1.2; 3J(6,7) ˆ 10.2, 4J(6,8) ˆ 1.2; 3J(7,8) ˆ 9.0. c) As d, J ˆ 0.6 Hz. d) As d, J ˆ 1.3 Hz. e) As d, J ˆ
0.6 Hz. f) As d, J ˆ 1.4 Hz. g) As t, J ˆ 1.4 Hz.
5
Table 7. 13C-NMR Data of Cyclohepta[b]furan-2(2H)-ones 3
Substituents in 3 No. Chemical shift [ppm] a)
C(2)
C(3) C(3a) C(4)/
C(5)/
C(6)/
C(7)/
C(8)/
C(8a)
RÀC(4) RÀC(5) RÀC(6) RÀC(7) RÀC(8)
None
3a
3c
3g
3f
3e
169.47 98.77 153.19 127.85
135.38
±
137.68
±
147.30
28.78
141.91
26.42
130.45
±
132.52
±
134.10
±
139.98
27.23
141.60
26.42
113.77
±
125.28
19.85
118.09
±
158.41
151.62
155.36
153.49
±
4,6,8-Me3
5,6,7-Me3
4,5,7,8-Me4
4,5,6,7,8-Me5
5,7-Et2
169.05 95.80 151.62 137.23
139.01
27.49
139.20
21.32
134.85
±
24.40
170.24 94.43 151.15 125.80
±
169.04 94.83 153.13 133.64
20.21
169.3795.18 151.91 132.58
20.77
126.53
16.85
144.37 138.93
138.22
126.48
151.91
23.75
151.95
34.77
15.23
23.94
130.92
±
21.04
148.57
34.29
15.11
18.08
115.32
±
3h 169.70 96.45 152.31 123.53
157.07
±
a) In CDCl3 at 75 MHz; for data of 3a see also [55].

1040
Helvetica Chimica Acta ± Vol. 83 (2000)
1
4.1.3. NMR Spectra. The data of the H- and 13C-NMR spectra are assembled in
Tables 6 and 7. All signal assignments were made by 1H,13C correlation spectra for each
compound with the exception of 5,7-diethylfuranone 3h, signal assignments of which
were made by comparison with the chemical shifts of the other furanones. The high-
field shift of the signal of HÀC(3) is in agreement with the high electron density at
C(3), which is also reflected in d(C(3)) in the 13C-NMR spectra. The increasing
number of alkyl substituents at the seven-membered ring displaces the signals, as
expected, to higher fields. The fact that all cyclohepta[b]furan-2(2H)-ones 3 that carry
5
an H-atom at C(8) display J(HÀC(3),HÀ(8)) ˆ 0.6 ± 1.2 Hz is also of diagnostic
value. The smallest shift differences are exhibited by the H-atoms or alkyl substituents
at C(5) and C(7), and most obviously are demonstrated by the NMR spectra of 3f. It
can also be recognized that the resonance position of the CˆO C-atom is not very
sensitive to the alkyl substitution pattern at the seven-membered ring of the furanones.
4.2.4. Mass Spectra. We measured the mass spectra of the furanones 3 by GC/MS
.
‡
coupling. The most prominent signal is that of the M ion in all spectra. Further
characteristic signals and their relative intensities are listed in Table 8. As expected, the
spectra are dominated by the successive loss of two CO molecules. The structure and
.
‡
further fate of the generated [M À 2CO] ions can only be speculated upon. However,
.
‡
it is remarkable that the most intense signal, after that of the M peak, is found at the
‡
m/z value corresponding to [M À 2CO À R] , whereby R represents Me, Et, or H,
according to the given substituents of 3. The intensity of this signal for all furanones
speaks for the generation of stabilized cations, which may be represented by
cyclopropenylium ions (see Scheme 14). The further fragmentation becomes more
and more clouded. However, the loss of acetylene seems to play a role.
.
The loss of the first molecule of CO is in competition with the loss of [HÀCˆO] ,
and it seems that this fragmentation plays a role when a Me group at C(4) can act as an
H-atom donor, since this fragmentation pattern is not observed for 3a and is also of no
importance for 3g and 3h (cf. Table 8).
5. Formation of 1,2,3,4-Tetrahydrobenz[a]azulenes. ± We reacted the furanones 3c,
3e, and 3f with 1-(cyclohex-1-enyl)pyrrolidine (19) in toluene in order to test the
azulene-forming propensity of these highly methylated cyclohepta[b]furan-2(2H)-ones
(Scheme 15)15). Indeed, in all three cases the generation of the blue azulene color was
recognized. The chromatographic isolation and purification of azulene 20c created no
specific problems. It was obtained in > 50% yield. However, azulenes 20e and 20f
turned out to be very sensitive to exposure to air, so that their chromatographic
purification on Al2O3 was difficult. The blue color of purified samples of 20f and
especially of 20e faded within hours when exposed to air.
1
Nevertheless, the new azulenes could be characterized by their H- and 13C-NMR
spectra, as well as by their mass spectra (see Table 9 and Exper. Part). Quite typical are
the positions of the signals for HÀC(10) which appear in the 1H-NMR spectra as ss in
the range of 7.07 ± 6.86 ppm, and in the13C-NMR spectra as ds in the range of 112 ±
115 ppm. The fragmentation behavior of 20c and 20e at 70 eV is distinctly different.
15
)
The new furanones 3 could also be transformed into azulenes with enol ethers at temperatures of 130 ± 2008.
We will report these results soon [9].

Helvetica Chimica Acta ± Vol. 83 (2000)
1041
Scheme 14
a) Tentative fragmentation pattern, exemplified for 3c.
.
‡
The spectrum of 20c with M at m/z 224 (90%) is dominated by the loss of ethene as
the result of a retro-Diels-Alder fragmentation leading to the signal at m/z 196 (100%).
The loss of a Me group is indicated by the signal at m/z 209 (50%). The loss of a second
Me group does not play a role according to the weak signal at m/z 194 (<6%). The
.
‡
fragmentation behavior of 20e is quite different. The M signal at m/z 252 (100%) is
Table 8. Mass Spectra of the Cyclohepta[b]furan-2(2H)-ones 3 a)
Substituents No.
Ions b)
in 3
.
‡
‡
M
[M À
[M À
[M À CO [M À
[M À CO [M À 2 CO [M À X]
.
‡
‡
‡
‡
‡
‡
CO]
CHO]
À Me]
2 CO]
À CHO] À R]
None
3a 146 (100) 118 (25)
3c
±
±
90 (85) ? c)
89 (49)
117 (36)
63 (20)
d
e
4,6,8-Me3
5,6,7-Me3
4,5,7,8-Me4
188 (100) 160 (6) 159 (3) 145 (7)
132 (18) 131 (8)
)
)
91 (13)
91 (17)
±
3g 188 (100) 160 (12) 159 (4) 145 (12) 132 (18) 131 (13) 117(47)
3f
202 (100) 174 (7) 173 (14) 159 (15) 146 (25) 145 (7)
131 (40) f)
145 (27)
131 (22)
165 (103)
103 (28)
4,5,6,7,8-Me5 3e 216 (100) 188 (5) 187 (17) 173 (20) 160 (22) 159 (5)
±
±
5,7-Et2
3-Ph g)
3-Me h)
3h 202 (100) 174 (8) 173 (2) 159 (15) 146 (2) 145 (9)
3m 222 (100) 194 (9) 193 (7)
3n 160 (100) 132 (21) 131 (92)
±
±
166 (61) ? c)
104 (27) ? c)
139 (77)
77 (12)
a) If not otherwise stated, spectra were recorded by GC/MS; carrier gas He, EI at 70 eV. b) Ions in m/z (rel. %);
R corresponds to substituents at the seven-membered ring in 3; X ˆ (2 CO ‡ R ‡ HÀCꢀCÀH). c) m/z Values
are identical for [M À CO À CHO] and [M À 2 CO À R] (R ˆ H). d) Further signal at m/z 115 (19).
e) Further signal at m/z 115 (25). f) Further signals at m/z 129 (7) and 128 (7). g) Taken from [24] and
redetermined with GC/MS. h) Data taken from [54].
‡
‡

1042
Helvetica Chimica Acta ± Vol. 83 (2000)
Scheme 15
a) Toluene, 120 ± 1308, 9 ± 12 h.
Table 9. NMR Data of the 1,2,3,4-Tetrahydrobenz[a]azulenes 20
Substituents No. 1H-NMR a)
in
20 (solvent)
13C-NMR b)
C(10) C(1 ± 4) C(6,7,8) c) Other C
HÀC(10) HÀC(1 ± 4) HÀC(6 ± 8) Me
Me
5,7,9-Me3
(CDCl3)
20c 7.07
3.37 d)
6.83 (s)
6.82 (s)
2.96
2.76
2.51
114.728.8
27.1
128.2
126.2
146.2, 145.0, 28.3
143.9, 142.8, 27.7
135.4, 133.6, 25.4
126.1
3.03 e)
1.93 ± 1.80 f)
24.8
22.8
5,7,9-Me3
(C6D6)
20c 7.18
3.24 ± 3.22 d) 6.70 (br.s) 2.75
115.5 28.9
27.3
127.9
125.8
146.2, 144.2, 27.9
143.0, 142.2, 27.4
136.0, 134.4, 25.2
126.2
3.03 e)
2.64
2.29
1.80 ± 1.71 f)
24.9
23.0
g
5,6,8,9-Me4 20f 7.07
(CDCl3)
3.34 d)
7.40 (s)
2.83
2.71
2.51
2.50
112.1 29.9
27.0
139.1
)
27.2
26.2
23.2
21.1
3.05 e)
1.80 ± 1.92 f)
25.1
22.9
5,6,7,8,9-Me5 20e 6.86
(CDCl3)
3.19 d)
±
±
2.84
2.62
2.93 e)
1.78 ± 1.90 f)
2.36
2.33 h)
5,6,7,8,9-Me5 20e 7.04
(C6D6)
3.12 d)
2.57
3.01 e)
2.54
1.71 ± 1.89 f)
2.09 h)
2.08
a) In the indicated solvent at 300 MHz. b) In the indicated solvent at 75 MHz; d in ppm. c) d of HÀC(X).
f
d) Centered m of CH2(1). e) Centered m of CH2(4). ) m of CH2(2,3). g) Signals of the quaternary C-atoms not
recorded. h) Broad s of two Me groups.
.
‡
here the reference signal, followed by the [M À Me] signal at m/z 237 (77%) and the
.
‡
[M À 2 Me] signal at m/z 222 (21%). The retro-Diels-Alder fragmentation is of no
importance according to the intensity of the signal at m/z 224 (7%).
We thank our NMR laboratory, in particular Nadja Walch, Martin Binder, and Dr. Gudrun Hoppe, for
specific NMR measurements, Dr. Philipp Ott for UV/VIS measurements and presentations, and Dr. Christophe
Weymuth for his helpful discussions. H.-J. H. thanks especially Manfred Hesse for a discussion on the
fragmentation behavior of the furanones. The financial support of this work by the Swiss National Science
Foundation is gratefully acknowledged.

Helvetica Chimica Acta ± Vol. 83 (2000)
1043
Experimental Part
General. TLC: glass plates pre-coated with silica gel 60 F254 from Merck or Polygram Alox N/UV254
(Macherey-Nagel); visualization by UV light and/or I2 vapor, respectively. Column chromatography (CC) and
flash chromatography (FC): on silica gel 60 (40 ± 63 mm), Chemie Uetikon AG, or aluminium oxide, Fluka, type
5016 A basic, Brockmann grade IV, or Alox ICN Alumina N neutral, Brockmann grade III. UV/VIS Spectra:
Perkin-Elmer spectrophotometer (model Lambda 9), l in nm (log e). IR Spectra: Perkin-Elmer spectropho-
tometer (model FT-IR 1600), nÄmax in cmÀ1. 1H- and 13C-NMR spectra: Bruker instruments (AC 300, ARX 300,
AMX 600); d(1H and 13C) in ppm rel. to CHCl3 and CDCl3 as internal standard 7.26 and 77.00 ppm, resp.;
J(H,H) in Hz; assignments by 1H,13C-correlation spectra (HSQC and HMBC technique). MS: GC/MS Hewlett-
Packard HP5890 Series II (GC), HP-5971 MSD (mass-selective detector (EI, 70 eV)), WCOT cap. column HP-
5, 25 m  0.2 mm; carrier gas: He; GC conditions: injector temp. 2808, starting temp. 608 over 2 min, rate 208/
min, term. temp. 2508 over 5 min.
1. Syntheses. ± 1.1. Reagents and Solvents. If not otherwise stated, reagents and solvents are from Fluka.
2,4,6-Trimethylphenol (mesitol; 1c), 2,3,5,6-tetramethylphenol (1f), Aldrich, purum; 3,4,5-trimethylphenol (1g),
EGA-Chemie; 3,5-diethylphenol (1h), EGA-Chemie; 2,6-diisopropylphenol (1i), EGA-Chemie, techn. (90%);
2,4,6-tri(tert-butyl)phenol (1k), purum; 2-naphthol (1l), purum; phenylcarbonochloridate (14a), purum; AlCl3,
H2O-free, pract.; HBr, 5.7m soln. in AcOH, pract.; cyclohexanone, purum; dicyclohexylcarbodiimide (DCC),
purum; 4-(dimethylamino)pyridine (DMAP), purum; N,N-dimethylaniline (DMA), purum; formaldehyde
soln. (30% in H2O), techn.; HCl (33 wt.-%), Merck, p. a. ; KOH, Siegfried, puriss.; MeI, purum; morpholine,
puriss.; Pd/C (10%), puriss.; paraformaldehyde, pract.; COCl2 in toluene (20%), purum; PCl5, purum, p. a. ;
H3PO4 (85% in H2O), Merck, p. a.; propiolic acid (ˆ propynoic acid), pract.; trifluoroacetic anhydride (TFAA),
purum; Ph3P, purum. Solvents. Solvents of the grade ꢁpuriss.ꢀ were used without further purification. Solvents of
the grade ꢁpurumꢀ were distilled and, when necessary, dried (Et2O and hexane over NaH; CH2Cl2 and AcOEt
over P2O5) before distillation. THF, distilled from potassium ketyl; toluene and o-xylene, distilled from Na.
Further solvents: tert-butyl methyl ether (TBME; purum), acetic acid (AcOH; Merck, p. a.), heptane fraction,
purum, 1-methyl-2-pyrrolidone (NMP; purum).
1.2. Aryl Propynoates 2. ± 1.2.1. 2,3,4,5,6-Pentamethylphenol (1e). 1.2.1.1. 3,5-Bis(bromomethyl)-2,4,6-
trimethylphenol (18). Mesitol (1c, 68.5 g, 0.5 mol) and paraformaldehyde (31 g, 1.0 mol) were suspended in
250 ml of AcOH, whereby the temp. of the mixture decreased. HBr/AcOH (200 ml, 1.14 mol HBr) was added
rapidly with stirring and the resulting pink-colored mixture warmed to 408. After ca. 10 min, a solid precipated
from the reaction mixture (GC/MS indicated the predominant presence of 3-(bromomethyl)mesitol (17; m/z
.
‡
228/230 (M )). After heating to > 608, the solid dissolved again, and the soln. turned reddish brown. The
reaction of 1c ! 17, 18 was completed by further addition of paraformaldehyde (3.1 g, 0.1 mol) and stirring at
70 ± 908 during 8 h (GC/MS control). The hot mixture was cautiously poured on to 1.5 l of ice and H2O. The
resulting greyish precipitate was vacuum-filtered and dissolved in Et2O, washed 3 times with H2O, and filtered
through a silica-gel pad. After addition of hexane, a fine powder precipitated at À208 (158 g, > 90%). It
contained mainly 18, but also some acetate of 18, which could be removed only by several recrystallizations.
However, the presence of 18-acetate did not disturb the subsequent hydrogenolysis reaction.
Data of Purified 18: colorless needles (felt). M.p. 162 ± 1648 (heptane). Rf(hexane/Et2O 4 :1; I2) 0.4.
1H-NMR (CDCl3 ‡ (D6)acetone): 6.5 (br. s, OH); 4.59 (s, 2 CH2Br); 2.38 (s, MeÀC(4)); 2.32 (s, MeÀC(2,6)).
13C-NMR (CDCl3 ‡ (D6)acetone): 151.1 (C(1)); 132.9 (C(3,5)); 128.8 (C(2,6)); 125.3 (C(4)); 30.52 (t, 2 CH2);
.
‡
14.5 (q, MeÀC(4)); 12.2 (q, MeÀC(2,6)). EI-MS (GC/MS): 324/322/320 (8/16/8, M ), 243/241 (100, [M À
.
.
‡
‡
HBr] ), 162 (75, [M À 2 HBr] ).
1.2.1.2. 2,3,4,6-Tetramethylphenol (1o) via 17. The reaction of 1c with 1 mol-equiv. of paraformaldehyde and
HBr/AcOH at < 658 led after 2 h to almost pure 17 contaminated with ca. 10% of 17-acetate (GC/MS). The
crude compound was dissolved in a mixture of AcOEt/EtOH/AcOH (ca. 2 :2 :1) and hydrogenated in the
presence of Pd/C (10%) (0.50 g of catalyst/100 g of 17). For details see 1.2.1.3. The phenol 1o was obtained in
yields of 75 ± 85% (with respect to 1c) as colorless needles.
1
Data of 1o: M.p. 79 ± 818: H-NMR (CDCl3): 6.77 (s, HÀC(5)); 4.45 (s, OH); 2.19, 2.18, 2.15 (3s, 4 Me).
13C-NMR (CDCl3): 150.0 (s, C(1)); 133.7( s, arom. C); 129.1 (d, C(5)); 128.0, 121.9, 119.5 (3s, arom. C); 20.0 (q,
.
‡
‡
Me); 15.7( q, 2 Me); 12.1 (q, Me). EI-MS (GC/MS): 150 (70, M ), 135 (100, [M À Me] ).
1.2.1.3. Hydrogenolysis of 18. Crude 18 was dissolved in AcOEt/EtOH/AcOH (ca. 2 :2 :1), and Pd/C (10%)
was added (0.50 g/100 g 18). Hydrogenation was performed with H2 under normal pressure. At the beginning,
the temp. increased to 30 ± 358. After 1ꢂ3 of H2 had been consumed, an equimolar amount of conc. aq. AcONa was
added. The two-phase mixture was stirred at r.t. until the total amount of H2 had been taken up (after 12 ± 24 h).

1044
Helvetica Chimica Acta ± Vol. 83 (2000)
The mixture was filtered over sea sand/Celite/silica gel, and Et2O was added to the filtrate. The org. layer was
washed with H2O and twice with sat. NaCl soln. The residue was recrystallized twice from hot heptane and
provided pure 1e in 70 ± 85% yield with respect to 1c in colorless crystals. Two times recrystallization from hot
heptane delivered pure 1e as colorless needles (purity according to GC: > 98%).
Data of 1e: M.p. 128 ± 1298 ([47]: 1288; [49]: 128 ± 1298). Rf(hexane/Et2O 4 :1, I2) 0.65. 1H-NMR (CDCl3):
4.50 (br. s, OH); 2.20, 2.18, 2.17(3 s, 5 Me). 13C-NMR (CDCl3): 149.5 (C(1)); 132.9, 126.9, 119.0 (3s, 5 arom. C);
.
‡
‡
16.4, 16.3 (2q, MeÀC(3,4,5)); 12.5 (q, MeÀC(2,6)). EI-MS (GC/MS): 164 (65, M ), 149 (100, [M À Me] ).
1.2.1.4. 2,3,4,5,6-Pentamethylphenol (1e) by Friedel-Crafts Reaction. Mesitol (1c) (1.36 g, 10 mmol) and
AlCl3 (2.7g, 20 mmol) were added to MeI (30 ml) and the suspension heated under reflux for 2 h. GC/MS
Control revealed the presence of greater amounts of 1o. After additional heating for 4 h, further MeI (5 ml) was
added, and heating was continued for 18 h. The mixture was hydrolyzed with ice-water, and the formed 1e was
recrystallized twice from hot heptane: 1.07g (65%) of 1e.
1.2.2. 2,4,6-Trimethylphenyl Prop-2-ynoate (2c) with Prop-2-ynoyl Chloride. 1.2.2.1. Prop-2-ynoyl Chloride.
Prop-2-ynoic acid (5.0 g, 0.07mol) was added dropwise to a small excess of PCl 5 (16.0 g, 0.0075 mol) at 5 ± 108.
Vivid evolution of HCl was recognized. The prop-2-ynoyl chloride formed was separated from the mixture, kept
at liquid N2 temp., and then slowly thawed by evaporation and condensation at 12 Torr. In a first trap, kept at
À788, condensed mainly POCl3. In a second trap, kept at ca. À1308, the chloride was collected. At temp. below
À1358, the chloride was co-condensed with HCl, which partially reacted on thawing with the chloride to give 3-
chloroprop-2-enoyl chloride. The procedure led to yields of the chloride in the range of 30 ± 35%. 1H-NMR
.
.
‡
‡
‡
(CDCl3): 3.7( s, HÀC(3)). EI-MS (GC/MS): 90/88 ( ± /2, M ), 62/60 (3/10, [M À CO] ), 53 (100, [M À Cl] ).
The thus prepared prop-2-ynoyl chloride contained always some 3-chloroprop-2-enoyl chloride (ꢁ10%), as
indicated by an AB system at 7.68 and 6.52 ppm with J(AB) ˆ 13.6 Hz (cf. [31]), as well as 10 ± 20% POCl3.
1.2.2.2. Formation of 2c. Mesitol (1c) (4.22 g, 0.31 mmol) was added to a suspension of oil-free NaH (0.83 g,
35 mmol) in Et2O (300 ml). The mixture was stirred until the evolution of H2 had ceased, and then, after
additional stirring for 0.5 h, cooled to À788. Freshly prepared prop-2-ynoyl chloride (see above; 3.5 g, ca.
34 mmol) was added dropwise within 15 min. The cooling bath was removed after additional stirring for 30 min,
and the mixture was allowed to warm up to r.t. H2O (100 ml) was added, and 2c formed was isolated by filtration
over a short silica-gel column with hexane/Et2O 4 :1. The solvent mixture was removed from the almost colorless
filtrate and the residue redissolved in hexane/Et2O 9 :1. Ester 2c crystallized at À208. Average yields of 2c were
75 ± 80%, which could be increased to 85 ± 90% by workup of the mother liquors.
Data of 2c: Colorless, transparent crystals. M.p. 58 ± 608 ([26]: 57± 58 8). Rf(hexane/Et2O 4 :1, I2) 0.65. IR
(KBr): 3240s (ꢀCÀH), 2130s (ÀCꢀCÀ), 1725vs (CˆO). 1H-NMR (CDCl3): 6.88 (s, HÀ(3',5')); 3.03 (s,
HÀC(3)); 2.27( s, MeÀC(4')); 2.14 (s, MeÀC(2',6')). 13C-NMR (CDCl3): 150.7(C ˆO); 145.1 (C(1')); 136.2
(C(4')); 129.5 (C(2',6')); 129.4 (C(3',5')); 76.4 (C(3)); 74.2 (C(2)); 20.8 (MeÀC(4')); 16.1 (MeÀC(2',6')). EI-MS
.
‡
‡
(GC/MS): 188 (85, M ), 173 (60, [M À Me] ), 145 (65), 136 (100), 91 (90).
The by-products 4 and 5 in crude 2c were identified by their GC/MS.
Also other products, i.e., 6 and (E)-7 and (Z)-7 (cf. Scheme 5), 8 (cf. Scheme 6), 11 and 12 (cf. Scheme 8)
that had been observed as by-products, were identified mainly by GC/MS. Moreover, ethyl prop-2-ynoate
(11) was unequivocally identified by its 1H-NMR (CDCl3) with 2.97( s, HÀC(3)), and 4.28/1.32 (q and t, resp.,
MeCH2O) as well as by comparison with an authentic sample.
1.2.3. Aryl Prop-2-ynoates 2 via Mixed Anhydrides. 1.2.3.1. Aryl Carbonochloridates 14. General Procedure:
COCl2 in toluene (20%; 1.3 ± 1.35 mol-equiv.) was added to the phenol (1 mol-equiv.), whereby the temp.
decreased slightly without reaction. The mixture was cooled to 08, and N,N-dimethylaniline (1 mol-equiv.) was
added dropwise under stirring in a way that the temp. not surpassed 58. The resulting creamy suspension was
stirred for 2 ± 22 h, whereby the temp. slowly reached r.t. The formation of the aryl carbonochloridates 14 was
followed by GC/MS. Hydrolysis of the excess of COCl2 was performed by dropwise addition of H2O (20 ± 30 ml)
at 08. After the precipitate of the anilinium hydrochloride had been dissolved, the org. phase was washed with
0.1n aq. HCl, and then with diluted NaOH and H2O. The org. layer was dried (MgSO4), and the residue was
filtered over a short column of silica gel. The residues of the now colorless filtrates were dissolved in hexane, and
14 crystallized at 48 or À208, or purified by bulb-to-bulb distillation in vacuo. (cf. Table 1).
Data of 2,4,6-Trimethylphenyl Carbonochloridate (14c). Colorless Oil. 1H-NMR (CDCl3): 6.85 (s,
HÀC(3',5')); 2.20 (s, MeÀC(4')); 2.08 (s, MeÀC(2',6')). 13C-NMR (CDCl3): 149.3 (CˆO); 145.8 (s, C(1'));
135.3 (s, C(4')); 130.4 (s, C(2',6')); 128.8 (d, C(3',5')); 20.0 (q, MeÀC(4')); 16.5 (q, MeÀC(2',6')). EI-MS (GC/
.
‡
‡
MS): 200/198 (17/50, M ), 135 (100, [M À COCl] ), 119 (45), 91 (35).
Data of 2,3,4,5,6-Pentamethylphenyl Carbonochloridate (14e). Colorless crystals. M.p. 60.5 ± 628. 1H-NMR
(CDCl3): 2.20 (s, MeÀC(2',4',6')); 2.15 (s, MeÀC(3',5')). 13C-NMR (CDCl3): 149.5 (CˆO); 145.7( s, C(1'));

Helvetica Chimica Acta ± Vol. 83 (2000)
1045
134.1 (s, C(4')); 134.0 (s, C(3',5')); 124.6 (s, C(2',6')); 16.7( q, MeÀC(4')); 16.4 (q, MeÀC(2',6')); 13.2 (q,
.
‡
‡
MeÀC(3',5')). EI-MS (GC/MS): 228/226 (27/80, M ), 163 (100, [M À COCl] ), 147(45).
1
Data of 2,3,5,6-Tetramethylphenyl Carbonochloridate (14f). Colorless needles. M.p. 66.5 ± 67.58. H-NMR
(CDCl3): 6.92 (s, HÀC(4')); 2.23 (s, MeÀC(3',5')); 2.10 (s, MeÀC(2',6')). 13C-NMR (CDCl3): 149.6 (CˆO);
149.1 (s, C(1')), 135.24 (s, C(3',5')); 130.1 (d, C(4')); 125.1 (s, C(2',6')); 19.6 (q, MeÀC(3',5')); 12.4 (q,
.
‡
‡
MeÀC(2',6')). EI-MS (GC/MS): 214/212 (20/60, M ), 149 (70, [M À COCl] ), 133 (100).
Data of 3,4,5-Trimethylphenyl Carbonochloridate (14g). Colorless oil, which solidified after several days
standing at r.t. H-NMR (CDCl3): 6.83 (s, HÀC(2',6')); 2.27( s, MeÀC(3',5')); 2.13 (s, MeÀC(4')). 13C-NMR
1
(CDCl3): 149.8 (s, CˆO); 149.0 (s, C(1')); 138.2 (s, C(3',5')); 134.2 (s, C(4')); 118.9 (d, C(2',6')); 20.7( q,
.
‡
‡
MeÀC(3',5')); 15.0 (q, MeÀC(4')). EI-MS (GC/MS): 200/198 (35/100, M ), 135 (70, [M À COCl] ), 119 (80).
1
Data of 3,5-Diethylphenyl Carbonochloridate (14h). Colorless oil. H-NMR (CDCl3): 6.96 (s, HÀC(4'));
6.85 (s, HÀC(2'6')); 2.62 (q, 2 MeCH2); 1.22 (t, 2 MeCH2). 13C-NMR (CDCl3): 151.9 (CˆO); 149.5 (s, C(1'));
146.3 (s, C(3',5')); 126.4 (d, C(4')); 116.9 (d, C(2',6')); 28.7( t, 2 MeCH2); 15.2 (q, 2 MeCH2).
1
Data of 2,6-Diisopropylphenyl Carbonochloridate (14i). Colorless oil, which solidified at À208. H-NMR
(CDCl3): 7.26 ± 7.15 (m, HÀC(3',4',5')); 3.00 (sept., 2 Me2CH); 1.23 (d, 2 Me2CH). 13C-NMR (CDCl3): 149.7( s,
CˆO); 147.1 (s, C(1')); 139.8 (s, C(2',6')); 127.8 (d, C(4')); 124.5 (d, C(3',5')); 27.7 (d, 2 Me2CH); 23.2 (q,
2 Me2CH).
Data of 2,4,6-Tri(tert-butyl)phenyl Carbonochloridate (14k). Yellowish solid. M.p. 80 ± 848. 1H-NMR
(CDCl3): 7.38 (s, HÀC(3',5')); 1.39 (s, Me3CÀC(2',6')); 1.29 (s, Me3CÀC(4')). 13C-NMR (CDCl3): 151.0 (CˆO);
148.8 (s, C(1')); 146.7( s, C(4')); 140.4 (s, C(2',6')); 123.8 (d, C(3',5')); 35.1 (s, Me3CÀC(2',6')); 34.5 (s,
.
‡
‡
Me3CÀC(4')); 31.2 (q, 3 Me3C). EI-MS (GC/MS): 326/324 (10/30, M ), 311/309 (30/100, [M À Me] ), 265 (100,
‡
[M À COCl] ).
Data of Naphthalen-2-yl Carbonochloridate (14l). Colorless crystals. M.p. 62 ± 648 ([40]: 65 ± 65.58).
1H-NMR (CDCl3): 7.76 ± 7.68 (m, 3 H); 7.57 (d-like, 1 H); 7.44 ± 7.36 (m, 2 H); 7.22 ± 7.18 (m, 1 H). 13C-NMR
(CDCl3): 149.6 (s, CˆO); 149.1 (s, C(2')); 133.4, 131.9 (2s, C(4'a,8'a)); 130.0, 127.8 (enhanced intensity), 127.1,
.
‡
126.5, 119.2, 117.8 (7d, C(1',3',4',5',6',7',8')). EI-MS (GC/MS): 208/206 (30/90, M ), 162 (40), 143 (40, [M À
‡
COCl] ), 127(45), 115 (100).
1.2.3.2. Sodium Prop-2-ynoate. Distilled prop-2-ynoic acid (35 g, 0.5 mol) was neutralized under cooling by
stirring with conc. aq. NaOH (20.02 g, 0.501 mol). The temp. was kept between 10 ± 208. The slight excess of
NaOH was neutralized with dropwise addition of prop-2-ynoic acid to adjust pH 6.8 ± 7.0. H O was then
2
removed on a rotatory evaporator, and the residue was dried under high vacuum until the weight was constant
(12 ± 14 h). The solid salt was powdered in a mortar and dried again at 408. Sodium prop-2-ynoate was thus
obtained as an ivory-colored powder that was protected from light and stored at À208.
1.2.3.3. Formation of Prop-2-ynoates 2. General Procedure: The aryl carbonochloridate (0.1 mol) was
dissolved in THF (100 ml), and sodium prop-2-ynoate (9.3 g, 0.1 mol) was added under stirring at r.t. In cases
where the evolution of CO2 did not start at r.t., the mixture was gently heated at 308, 408, or, for 2,6-disubstituted
aryl compounds, at 40 ± 658. The progress of the reaction was followed by GC. Within 2 ± 12 h, all reactions were
complete. The product mixture was filtered over sea sand and silica gel. The solvent was distilled off, and the
residue was dissolved in hexane or hexane/Et2O, and filtered again over silica gel. The filtrate was cooled to 4 or
À208 to induce crystallization, or the residue was purified by bulb-to-bulb distillation, when no crystallization
occurred. For yields, see Table 2.
1
Data of Phenyl Prop-2-ynoate (2a): Colorless oil. H-NMR (CDCl3): 7.41 ± 7.34 (t, with f.s., HÀC(3',5'));
7.27 ± 7.22 (t, with f.s., HÀC(4')); 7.15 ± 7.10 (d, with f.s., HÀC(2',6')); 3.05 (s, HÀC(3)). 13C-NMR (CDCl3): 151.0
(s, CˆO); 149.8 (s, C(1')); 129.6 (d, C(3',5')); 126.6 (d, C(4')); 121.2 (d, C(2',6')); 76.7 (d, C(3)); 74.2 (C(2)); EI-
.
‡
‡
MS (GC/MS): 146 (94, M ), 145 (78), 118 (100, [M À CO] ), 94 (78), 93 (39), 90 (100), 89 (87).
As by-product, diphenyl carbonate was obtained and spectroscopically identified.
Data of 2c: See 1.2.2.2.
Data of 2,3,4,5,6-Pentamethylphenyl Prop-2-ynoate (2e): Colorless crystals. M.p. 114 ± 1158. 1H-NMR
(CDCl3): 3.03 (s, HÀC(3)); 2.20 (s, MeÀC(3',4',5')); 2.10 (s, MeÀC(2',6')). 13C-NMR (CDCl3): 151.1 (s, CˆO);
145.0 (s, C(1')); 133.7( s, C(3',5')); 133.4 (s, C(4')); 125.0 (s, C(2',6')); 76.3 (d, C(3)); 74.3 (C(2)); 16.6/16.4 (2 q,
MeÀC(3',4',5)); 13.4 (q, MeÀC(2',6')). EI-MS (GC/MS): 216 (65, M ), 201 (100, [M À Me] ), 164 (60).
Data of 2,3,5,6-Tetramethylphenyl Prop-2-ynoate (2f). Colorless crystals. M.p. 86 ± 898. 1H-NMR (CDCl3):
6.88 (s, C(4')); 3.03 (s, HÀC(3)); 2.22 (s, MeÀC(3',5')); 2.04 (s, MeÀC(2',6')). 13C-NMR (CDCl3): 151.8 (s,
CˆO); 148.1 (s, C(1')); 135.92 (s, C(3',5')); 130.6 (d, C(4')); 126.5 (s, C(2',6')); 77.37 (d, C(3)); 75.2 (s, C(2)); 20.6
(q, MeÀC(3',5')); 13.5 (q, MeÀC(2',6')).
.
‡
‡

1046
Helvetica Chimica Acta ± Vol. 83 (2000)
1
Data of 3,4,5-Trimethylphenyl Prop-2-ynoate (2g). White solid. M.p. 54 ± 568. H-NMR (CDCl3): 6.77 (s,
HÀC(2',6')); 3.02 (s, HÀC(3)); 2.27( s, MeÀC(3',5')); 2.13 (s, MeÀC(4')). 13C-NMR (CDCl3): 151.4 (s, CˆO);
147.0 (s, C(1')); 137.9 (s, C(3',5')); 133.5 (s, C(4')); 119.8 (d, HÀC(2',6')); 76.4 (d, C(3)); 74.5 (s, C(2)); 20.6 (q,
MeÀC(3',5')); 15.0 (q, MeÀC(4')).
As by-product, bis(3,4,5-trimethylphenyl) carbonate was obtained and spectroscopically identified.
Data of 3,5-Diethylphenyl Prop-2-ynoate (2h): Colorless oil. 1H-NMR (CDCl3): 6.96 (br. s, HÀC(4')); 6.79
(s, HÀC(2',6')); 3.03 (s, HÀC(3)); 2.62 (q, 2 MeCH2); 1.22 (t, 2 MeCH2). 13C-NMR (CDCl3): 150.2 (CˆO);
149.0 (C(1')); 145.0 (s, C(3',5')); 124.9 (d, C(4')); 116.8 (d, C(2',6')); 75.5 (d, C(3)); 73.5 (s, C(2)); 27.7 (t,
2 MeCH2); 14.3 (q, 2 MeCH2).
As by-product, bis(3,5-diethylphenyl) carbonate was obtained and identified spectroscopically.
Data of 2,6-Diisopropylphenyl Prop-2-ynoate (2i). Colorless oil. 1H-NMR (CDCl3): 7.23 (m, Jo ˆ 7.5,
HÀC(4')); 7.17 (m, Jo ˆ 7.5, HÀC(3',5')); 3.05 (s, HÀC(3)); 2.95 (sept., 2 Me2CH); 1.21 (d, 2 Me2CH). 13C-NMR
(CDCl3): 151.4 (s, CˆO); 144.7( s, C(1')); 140.3 (s, C(2',6')); 127.2 (d, C(4')); 124.2 (d, C(3',5')); 76.6 (d, C(3));
74.2 (s, C(2)); 27.6 (d, 2 Me2CH); 23.2 (q, 2 Me2CH). EI-MS (GC/MS): 230 (73), 215 (17), 187 (95), 177 (100),
163 (82).
Data of Naphthalen-2-yl Prop-2-ynoate (2l): Colorless crystals. M.p. 73.5 ± 758 ([29]: 74 ± 758). 1H-NMR
(CDCl3): 8.00 ± 7.90 (m, 3 H); 7 .7 5d(, J ˆ 1.9, HÀC(1')); 7.65 ± 7.55 (m, 2 H); 7 .39 d(d, J ˆ 8.9, 2.2, HÀC(8'));
3.21 (s, HÀC(3)). 13C-NMR (CDCl3): 151.0 (s, CˆO); 147.4 (s, C(2')); 133.6, 131.7(2 s, C(4'a,8'a)); 129.6, 127.7
(enhanced intensity), 126.8, 126.1, 120.3, 118.4 (7d, C(1',3',4',5',6',7',8')).
As by-product, 40% of di(naphthalen-2-yl) carbonate was obtained and spectroscopically identified.
2. Cyclohepta[b]furan-2(2H)-ones 3 via Pyrolysis of Aryl Prop-2-ynoates 2. 2.1. General Remarks. The
essential semi-micro apparatus for FVP for 650 ± 6708 had been already described in the theoretical part. A more
detailed description is given in the work cited in Footnote 1. Some FVP experiments have been performed with a
prep. FVP apparatus at Givaudan-Roure (Dübendorf/Zurich)16). The heated quartz tube (length 40 cm,
diameter 2.5 cm; filled with quartz tubes 8 Â 8 mm) was vertically amounted in this apparatus, which allowed the
introduction of 2 in xylene soln. by dropping the soln. into the heated, vertical quartz tube. This apparatus
allowed the pyrolysis of 10 ± 20 g of 2e. The yields of furanone 3e were at the first attempts better, i.e., at the
upper end of the given yields in Table 3, than with the improvised semi-micro apparatus. All spectral data of the
furanones 3 are compiled in Tables 4 ± 8.
The following melting points were recorded (in parentheses: color of the crystals): 3a (lemon), 73 ± 748; 3c
(yellow), 186 ± 1888; 3e (lemon); 169 ± 1728; 3f (greenish yellow) 180 ± 1838; 3g (yellow-orange) 144 ± 1468.
3. Formation of Azulenes. General Remarks. 1-(Cyclohex-1-enyl)pyrrolidine (19) was prepared in the usual
way from cyclohexanone and pyrrolidine. It was obtained after distillation as a colorless oil and in a purity of >
99% (GC analysis). Furanones 3 and a 3 ± 5-fold amount of 19 were dissolved in toluene (ca. 5% soln. with
respect to 3), and the mixture was heated at 120 ± 1308 in a sealed Pyrex tube17). After heating, TLC analysis
(hexane) showed in all cases an intense blue spot with Rf of ca. 0.8 ± 0.9.
The solvent was removed by distillation, and the residue was dissolved again in hexane. H2O and aq. H3PO4
(dropwise) were added to adjust the pH to 4 ± 5. The greenish-blue hexane phase was washed with H2O and
dried (Na2SO4). The azulene was further purified by chromatography on Alox (basic, act. IV), and the solvent
was distilled off. Residual hexane was removed in high vacuum.
3.1. 1,2,3,4-Tetrahydro-5,7,9-trimethylbenz[a]azulene (20c). Reaction time 12 ± 14 h, average yields 55 ±
.
‡
‡
75%. 1H- and 13C-NMR: see Table 9. EI-MS (GC/MS): 224 (90, M ), 209 (50, [M À Me] ), 196 (100, [M À
.
‡
CH2ˆCH2] ), 194 (<6), 165 (27).
3.2. 1,2,3,4-Tetrahydro-5,6,7,8,9-pentamethylbenz[a]azulene (20e). Reaction time 8 ± 12 h. The sensitivity of
this compound to oxidative degradation by air did not allow complete purification by chromatography (Alox;
basic, act. IV) under standard conditions. Hexane fractions of 20e lost their intense blue color within hours on
.
standing in air. 1H- and 13C-NMR: see Table 9. EI-MS (GC/MS): 252 (100, M ), 237(77, [ M À Me] ), 224 (7,
‡
‡
.
‡
‡
‡
[M À CH2ˆCH2] ), 222 (21, [M À 2 Me] ), 209 (14, [M À (Me ‡ CH2ˆCH2)] ), 165 (14).
3.3. 1,2,3,4-Tetrahydro-5,6,8,9-tetramethylbenz[a]azulene (20f). Reaction time 9 ± 11. The azulene could not
be purified completely by chromatography under standard conditions. Hexane fractions of 20f lost their intense
blue color on standing, but not as rapidly as those of 20e. H- and 13C-NMR: see Table 9.
1
16
Â
)
)
We thank PD Dr. G. Frater for allowing us to perform some FVP experiments with the Givaudan apparatus.
17
The azulene formation could also be performed by heating the reaction partners in boiling EtOH.

Helvetica Chimica Acta ± Vol. 83 (2000)
1047
REFERENCES
[1] A. A. S. Briquet, P. Uebelhart, H.-J. Hansen, Helv. Chim. Acta 1996, 79, 2282.
[2] S. El Houar, H.-J. Hansen, Helv. Chim. Acta 1997, 80, 253; S. Maillefer-El Houar, H.-J. Hansen, Helv. Chim.
Acta 2000, 83, in preparation.
[3] Y. Chen, R. W. Kunz, P. Uebelhart, R. H. Weber, H.-J. Hansen, Helv. Chim. Acta 1992, 75, 2447.
[4] a) W. Bernhard, P. Brügger, J. J. Daly, P. Schönholzer, R. H. Weber, H.-J. Hansen, Helv. Chim. Acta 1985,
68, 415; b) W. Bernhard, P. Brügger, P. Schönholzer, R. H. Weber, H.-J. Hansen, Helv. Chim. Acta 1985, 68,
429.
[5] K. Hafner, G. L. Knaup, H. J. Lindner, H.-C. Flöter, Angew. Chem. 1985, 97, 209; Angew. Chem., Int. Ed.
1985, 24, 212.
[6] P. Uebelhart, H.-J. Hansen, unpublished results.
[7] a) K. Hafner, H. Diehl, H.-U. Süss, Angew. Chem. 1976, 88, 121; Angew. Chem., Int. Ed. 1976, 15, 104; b) K.
Hafner, G. L. Knaup, H. J. Lindner, Bull. Chem. Soc. Jpn. 1988, 61, 155.
[8] W. Bernhard, H.-R. Zumbrunnen, H.-J. Hansen, Chimia 1979, 33, 324.
[9] M. Nagel, V. Lellek, H.-J. Hansen, Helv. Chim. Acta 2000, 83, in preparation.
[10] E. Heilbronner, T. Hoshi, J. L. v. Rosenberg, K. Hafner, Nouv. J. Chim. 1977, 1, 105.
[11] K. Hafner, H. Kaiser, Org. Synth., Coll. Vol. 5, 1973, 1088.
[12] F. G. Praill, A. L. Whitear, J. Chem. Soc. 1961, 3573.
[13] R. H. Weber, Ph.D. Thesis, University of Basel, 1988.
[14] S. Kurokawa, A. G. Anderson, Bull. Chem. Soc. Jpn. 1979, 52, 257.
[15] a) C. W. Muth, M. L. DeMatte, A. R. Urbanik, W. G. Isner, J. Org. Chem. 1966, 31, 3013; b) Y. N. Porshnev,
T. N. Ivanova, L. V. Efimova, E. M. Tereshchenko, M. I. Cherkasin, K. M. Dyumaev, Zh. Org. Khim. 1982,
18, 150 and Russ. J. Org. Chem. 1982, 18, 132.
[16] H. G. Rajoharison, C. Roussel, Bull. Soc. Chim. Fr. 1986, 307.
[17] V. Lellek, H.-J. Hansen, unpublished results.
[18] R. W. Alder, G. Whittaker, J. Chem. Soc., Perkin Trans. 2, 1975, 714.
[19] Y. Chen, H.-J. Hansen, Helv. Chim. Acta 1993, 76, 168.
[20] K. Hafner, H. Kaiser, Liebigs Ann. Chem. 1958, 618, 140.
[21] K.-P. Zeller, in ꢁMethoden der organischen Chemieꢀ, 4. Aufl., Georg Thieme Verlag, Stuttgart, 1985, Bd. V/
2c, p. 127.
[22] T. Nozoe, H. Wakabayashi, K. Shindo, S. Ishikawa, C. P. Wu, P. W. Yang, Heterocycles 1991, 32, 213; H.
Wakabayashi, P. W. Yang, C. P. Wu, K. Shindo, S. Ishikawa, T. Nozoe, Heterocycles 1992, 34, 429; M.
Yokota, T. Yanagisawa, K. Kosakai, S. Wakabayashi, T. Tomiyama, M. Yasunami, Chem. Pharm. Bull. 1994,
42, 865.
Â
[23] T. Nozoe, T. Mukai, K. Takase, Science Reports Tokohu Univ., Series 1 1954, 36, 40; cf. Chem. Abstr. 1954,
48, 3946d.
[24] A. A. S. Briquet, H.-J. Hansen, Helv. Chim. Acta 1994, 77, 1577; see also A. A. S. Briquet, Ph.D. Thesis,
University of Zurich, 1993.
[25] a) P. L. Pauson, Chem. Rev. 1955, 55, 9; b) T. Nozoe, Fortschr. Chem. org. Naturst. 1956, 13, 232.
[26] W. S. Trahanovsky, S. L. Emeis, A. S. Lee, J. Org. Chem. 1976, 41, 4044.
[27] J. Wolf, W. Stüer, C. Grünwald, H. Werner, P. Schwab, M. Schulz, Angew. Chem. 1998, 110, 1165; Angew.
Chem., Int. Ed. 1998, 37, 1124.
[28] a) R. F. C. Brown, ꢁPyrolytic Methods in Organic Chemistryꢀ, Academic Press Inc., New York, 1980; b) M.
Karpf, Angew. Chem. 1986, 98, 413; c) R. F. C. Brown, F. W. Eastwood, K. J. Harrington, Aust. J. Chem.
1974, 27, 2393; d) R. F. C. Brown, F. W. Eastwood, G. L. McMullen, J. Chem. Soc., Chem. Commun. 1974,
123; e) R. F. C. Brown, F. W. Eastwood, G. L. McMullen, Aust. J. Chem. 1977, 30, 179; f) M. Karpf, A. S.
Dreiding, Helv. Chim. Acta 1979, 62, 852.
[29] L. A. Miller, US-Patent Nr. 3097230 (1963); Chem. Abstr. 1963, 59, 13891h.
[30] M. E. Jung, K. R. Buszek, J. Am. Chem. Soc. 1988, 110, 3965.
[31] W. J. Balfour, C. C. Greig, S. Visaisouk, J. Org. Chem. 1974, 39, 725.
[32] A. Rieche, H. Gross, Chem. Ber. 1959, 92, 83; see also A. Rieche, H. Gross, Org. Synth. 1967, 47, 47 .
[33] A. H. Ahlbrecht, D. W. Codding, J. Am. Chem. Soc. 1953, 75, 984; see also J. M. Tedder, Chem. Rev. 1955,
55, 787.
[34] H. Pelartzik, B. Irmsch-Pielartzik, T. Eicher, in ꢁHouben-Weyl, Carbonsäuren und Carbonsäure-Derivateꢀ,
Part 1, Ed. J. Falbe, Band E5/1, Georg Thieme Verlag, Stuttgart, New York, 1985, p. 670 ff and p. 692 ff.

1048
Helvetica Chimica Acta ± Vol. 83 (2000)
[35] R. F. C. Brown, F. W. Eastwood, J. Org. Chem. 1981, 46, 4588.
[36] S. Hashimoto, J. Furukowa, Bull. Chem. Soc. Jpn. 1981, 54, 2227.
[37] H. A. Staab, Angew. Chem. 1959, 71, 194; H. A. Staab, Angew. Chem., Int. Ed. 1962, 74, 407.
[38] G. M. Coppola, R. E. Damon, Synth. Commun. 1993, 23, 2003.
[39] G. Heywang, in ꢁHouben-Weyl, Kohlensäure-chlorid-esterꢀ, Ed. H. Hagemann, Band E/4, Georg Thieme
Verlag, Stuttgart, New York, 1983, p. 15 ff.
[40] a) M. J. Zabik, R. D. Schuetz, J. Org. Chem. 1967, 32, 300; b) V. A. Pattison, J. G. Colson, R. L. K. Carr, J.
Org. Chem. 1968, 33, 1084.
[41] a) D. Seebach, T. Hassel, Angew. Chem. 1978, 90, 296; b) M. R. Banks, J. I. G. Cadogan, I. Gosney, P. K. G.
Hodgson, P. R. R. Langridge-Smith, D. W. H. Rankin, J. Chem. Soc., Chem. Commun. 1994, 1365.
[42] In [34], p. 693.
[43] A. Einhorn, Ber. Dtsch. Chem. Ges. 1909, 42, 2772.
[44] T. B. Windholz, J. Org. Chem. 1958, 23, 2044; T. B. Windholz, J. Org. Chem. 1960, 25, 1703.
[45] M. Green, Chem. Ind. (London) 1961, 435.
[46] a) D. S. Tarbell, N. A. Leister, J. Org. Chem. 1958, 23, 1149; b) D. S. Tarbell, E. J. Longosz, J. Org. Chem.
1959, 24, 774.
[47] R. Wegler, E. Regel, Makromol. Chem. 1953, 9, 1.
[48] S. Pappalarde, G. Ferguson, J. F. Gallagher, J. Org. Chem. 1992, 57, 7102.
[49] J. Peter-Katalinic, J. Zsindely, H. Schmid, Helv. Chim. Acta 1973, 56, 2796.
[50] A. W. van der Made, R. H. van der Made, J. Org. Chem. 1993, 58, 1262.
[51] J. Zavada, M. Pankova, P. Holy, M. Tichy, Synthesis 1994, 1132.
Â
Â
Â
[52] G. Darzens, A. Levy, C. R. Hebd. Seances Acad. Sci. 1936, 202, 73; G. Darzens, C. R. Hebd. Seances Acad.
Sci. 1939, 208, 818.
[53] R. C. Fuson, C. H. McKeever, Org. React. 1942, 1, 63; see also G. A. Olah, W. S. Tolgyesi, in ꢁFriedel-Crafts
and Related Reactionsꢀ, Part II, Ed. G. A. Olah, Interscience Publ., New York, 1964, p. 734 ff.
[54] R. F. C. Brown, F. W. Eastwood, J. Org. Chem. 1981, 46, 4588.
[55] M. Nitta, H. Tomioka, A. Akaogi, K. Takahashi, K. Saito, K. Ito, J. Chem. Soc., Perkin Trans. 1 1994, 2625.
[56] Y. Saseda, Bull. Chem. Soc. Jpn. 1959, 32, 165.
[57] T. Nozoe, K. Takase, S. Fukuda, Bull. Chem. Soc. Jpn. 1971, 44, 2210.
[58] R. N. Jones, C. L. Angell, T. Ito, R. J. D. Smith, Can. J. Chem. 1959, 37, 2007; see also: N. B. Colthup, L. H.
Daly, S. E. Wiberly, ꢁIntroduction to Infrared and Raman Spectroscopyꢀ, Academic Press, Inc., New York,
1975, p. 294.
Received January 31, 2000