Surprising formation of highly substituted azulenes on thermolysis of 4,5,6,7,8-pentamethyl-2h-eyelohepta[b]furan-2-one and heptalene formation with the new azulenes

Article abstract of DOI:10.1002/1522-2675(20010613)84:6<1712::AID-HLCA1712>3.0.CO;2-E

Heating of 4,5,6,7,8-pentamethyl-2H-cyclohepta[b]furan-2-one (1a) in decalin at temperatures > 170° leads to the development of a blue color, typical for azulenes. It belongs, indeed, to two formed azulenes, namely 4,5,6,7,8-pentamethyl-2-(2,3,4,5,6-pentamethylphenyl)azulene (4a) and 4,5,6,7,8-pentamethylazulene (5a) (cf. Scheme 2 and Table 1). As a third product, 4,5,6,7-tetramethyl-2-(2,3,4,5,6-pentamethylphenyl)-1H-indene (6a) is also found in the reaction mixture. Neither 4,6,8-trimethyl-2H-cyclohepta[b]furan-2-one (1b) nor 2Hcyclohepta[b]furan-2-one (1c) exhibit, on heating, such reactivity. However, heating of mixtures 1a/1b or 1a/1c results in the formation of crossed azulenes, namely 4,6,8-trimethyl-2-(2,3,4,5,6-pentamethylphenyl)azulene (4ba) and 2-(2,3,4,5,6-pentamethylphenyl)azulene (4ca), respectively (cf. Scheme 3). The formation of small amounts of 4,6,8-trimethylazulene (5ba) and azulene (5ca), respectively, besides 1H-indene 6a is also observed. The observed product types speak for an [8 + 2]-cycloaddition reaction between two molecules of la or between 1b and 1c, respectively, with 1a, whereby la plays in the latter two cases the part of the two-atom component (cf. Figs. 5-7 and Schemes 4-6). Strain release, due to the five adjacent Me groups in 1a, in the [8+2]cycloaddition step seems to be the driving force for these transformations (cf. Table3), which are further promoted by the consecutive loss of two molecules of CO₂ and concomitant formation of the 10π-electron system of the azulenes. The new azulenes react with dimethyl acetylenedicarboxylate (ADM) to form the corresponding dimethyl heptalene-4,5-dicarboxylates 20, 22, and 24 (cf. Scheme 7), which give thermally or photochemically the corresponding double-bond-shifted (DBS) isomers 20′, 22′, and 24′, respectively. The five adjacent Me groups in 20/20′ and 24/24′ exert a certain buttressing effect, whereby their thermal DBS process is distinctly retarded in comparison to 22/22′, which carry 'isolated' Me groups at C(6), C(8), and C(10). This view is supported by X-ray crystal-structure analyses of 22 and 24 (cf Fig. 8 and Table 5).

Full text of DOI:10.1002/1522-2675(20010613)84:6<1712::AID-HLCA1712>3.0.CO;2-E

1712
Helvetica Chimica Acta ± Vol. 84 (2001)
Surprising Formation of Highly Substituted Azulenes on Thermolysis of
4,5,6,7,8-Pentamethyl-2H-cyclohepta[b]furan-2-one and Heptalene
Formation with the New Azulenes
by Vit Lellek and Hans-Jürgen Hansen*
Organisch-chemisches Institut der Universität Zürich, Winterthurerstr. 190, CH-8057 Zürich
Dedicated to Edgar Heilbronner on the occasion of his 80th birthday
Der Mensch kommt nicht um die Tatsache herum, dass die Sinne zwar als höchste Herren seine
Erkenntnisfähigkeit bestimmen, ihm aber keine Sicherheit bieten, da sie in jeder Lage sich leicht täuschen lassen.
Michel de Montaigne, Essais (Übers. Hans Stilett)
Heating of 4,5,6,7,8-pentamethyl-2H-cyclohepta[b]furan-2-one (1a) in decalin at temperatures > 1708
leads to the development of a blue color, typical for azulenes. It belongs, indeed, to two formed azulenes, namely
4,5,6,7,8-pentamethyl-2-(2,3,4,5,6-pentamethylphenyl)azulene (4a) and 4,5,6,7,8-pentamethylazulene (5a) (cf.
Scheme 2 and Table 1). As a third product, 4,5,6,7-tetramethyl-2-(2,3,4,5,6-pentamethylphenyl)-1H-indene (6a)
is also found in the reaction mixture. Neither 4,6,8-trimethyl-2H-cyclohepta[b]furan-2-one (1b) nor 2H-
cyclohepta[b]furan-2-one (1c) exhibit, on heating, such reactivity. However, heating of mixtures 1a/1b or 1a/1c
results in the formation of crossed azulenes, namely 4,6,8-trimethyl-2-(2,3,4,5,6-pentamethylphenyl)azulene
(4ba) and 2-(2,3,4,5,6-pentamethylphenyl)azulene (4ca), respectively (cf. Scheme 3). The formation of small
amounts of 4,6,8-trimethylazulene (5ba) and azulene (5ca), respectively, besides 1H-indene 6a is also observed.
The observed product types speak for an [8 ‡ 2]-cycloaddition reaction between two molecules of 1a or between
1b and 1c, respectively, with 1a, whereby 1a plays in the latter two cases the part of the two-atom component (cf.
Figs. 5 ± 7 and Schemes 4 ± 6). Strain release, due to the five adjacent Me groups in 1a, in the [8 ‡ 2]-
cycloaddition step seems to be the driving force for these transformations (cf. Table 3), which are further
promoted by the consecutive loss of two molecules of CO2 and concomitant formation of the 10p-electron
system of the azulenes. The new azulenes react with dimethyl acetylenedicarboxylate (ADM) to form the
corresponding dimethyl heptalene-4,5-dicarboxylates 20, 22, and 24 (cf. Scheme 7), which give thermally or
photochemically the corresponding double-bond-shifted (DBS) isomers 20', 22', and 24', respectively. The five
adjacent Me groups in 20/20' and 24/24' exert a certain buttressing effect, whereby their thermal DBS process is
distinctly retarded in comparison to 22/22', which carry ꢀisolatedꢁ Me groups at C(6), C(8), and C(10). This view
is supported by X-ray crystal-structure analyses of 22 and 24 (cf. Fig. 8 and Table 5).
1. Introduction. ± Recently, we reported on the synthesis of highly alkylated 2H-
cyclohepta[b]furan-2-ones by vacuum flash pyrolysis (VFP) of polyalkylphenyl prop-2-
ynoates [1], following a procedure originally described by Trahanovsky et al. [2]. We
also showed, inter alia, that 4,5,6,7,8-pentamethyl-2H-cyclohepta[b]furan-2-one (1a)
reacts thermally with 1-(cyclohex-1-en-yl)pyrrolidine in toluene to give the corre-
sponding 1,2,3,4-tetrahydrobenz[a]azulene 2 (Scheme 1) according to a general method
of azulene synthesis established by Nozoe et al. (see ref. cit. in [1]). Since we were
especially interested in the synthesis of 4,5,6,7,8-pentamethyl-(3a) as well as 1,4,5,6,7,8-
hexamethylazulene (3b), we tested a whole number of electron-donor substituted
ethenes and propenes in the thermal addition reaction with 1a ± but with only moderate

Helvetica Chimica Acta ± Vol. 84 (2001)
1713
Scheme 1
120 – 130°
toluene
N
+
O
O
– CO2,
N
H
1a
2
200°
D
R
R
D = donor
3a R = H
3b R = Me
success [3]. In particular, we did not succeed in reacting 1a thermally with easily
available enol ethers as the most convenient source for the necessary two C-atoms of
the azulene core that is formed in the [8 ‡ 2]-cycloaddition step, followed by the loss of
CO2 in a cycloreversion reaction and then by elimination of the corresponding alcohol
component of the used enol ether. However, we always observed the development of a
blue color, typical for the presence of azulenes in reaction mixtures, so that we thought,
at first, that the expected reaction between 1a and the enol ethers had taken place to a
certain extent. But, we then found that the blue color also developed when 1a was
heated in decalin in the absence of any enol ether. Isolation of the blue-color-carrying
.
‡
component left no doubt that it indeed represented an azulene with a M peak at m/z
344 and at least six sharp s signals in the 1H-NMR spectrum, attributable to Me groups
linked to aromatic C-atoms.
In Sect. 2, we describe the details of our investigations on the formation of azulenes
upon thermolysis of 4,5,6,7,8-pentamethyl-2H-cyclohepta[b]furan-2-one (1a).
2. Azulene Formation. ± Heating a 0.1m solution of 1a in decalin (ˆdecahydro-
naphthalene) at 1908 during 48 h or at 2308 during 3.5 h gave the highest yield of the
new azulene 4a (see Table 1). However, thorough chromatography (silica gel, hexane)
disclosed the presence of a second azulene 5a as well as of a colorless compound 6a
(Scheme 2). The structure of the second azulene could easily be solved on the basis of
its UV/VIS (see Fig. 1), 1H- and 13C-NMR, and mass spectra. The UV/VIS spectrum of
the main azulene 4a (see Fig. 2) was similar to that of 5a, however, in contrast to 5a,
with a vibrationally nonstructured, broad azulene absorption band centered at 570 nm.
The third colorless compound 6a exhibited a UV spectrum (see Fig. 3) with a strong
absorption band at 280 nm that indicated the presence of a styrene-type chromophor.
All spectral data of 4a and 6a were in full agreement with the proposed structures.

1714
Helvetica Chimica Acta ± Vol. 84 (2001)
Fig. 1. UV/VIS Spectrum (hexane) of 4,5,6,7,8-pentamethylazulene (5a)
Fig. 2. UV/VIS Spectrum (hexane) of 4,5,6,7,8-pentamethyl-2-(2,3,4,5,6-pentamethylphenyl)azulene (4a)
The 1H-NMR spectrum (C6D6) of 5a exhibited two signals in a ratio of 1:2 in the aromatic region, namely a
t at 7.81 and a d at 7.28 ppm with J ˆ 4.0 Hz, characteristic for the nonsubstituted five-membered ring of an
azulene. In agreement with this assignment, three sharp s signals in a ratio of 2 :1:2 at 2.62, 2.11, and 2.09 ppm,
attributable to Me C(4)/Me C(8), Me C(6), and Me C(5)/Me C(7), respectively, were found in the Me-
.
‡
group region of the spectrum. The MS (70 eV) was in accordance with the assigned structure, since the M peak
‡
was observed at m/z 198 (98%) and accompanied by a strong [M Me] peak at m/z 183 (100%).
The 1H-NMR spectrum (C6D6) of 4a resembled that of 5a, however, with the exception of there being only
one sharp s in the aromatic region at 7.30 ppm, counting for two H-atoms, which thus had to be linked
symmetrically to the five-membered ring of a 4,5,6,7,8-pentamethylated azulene core. This view was fully
supported by the observed s signals in the region of aromatically bound Me groups. The already mentioned 6 s
appeared in a ratio of 2 :2 :1 :2 :1:2 at 2.61, 2.27, 2.25, 2.23, 2.17, and 2.14 ppm, respectively. This means that the
outer s signals at 2.61, 2.17, and 2.14 ppm could be attributed to 5 Me groups at the seven-membered ring of an

Helvetica Chimica Acta ± Vol. 84 (2001)
1715
Fig. 3. UV/VIS Spectrum (hexane) of 4,5,6,7-tetramethyl-2-(2,3,4,5,6-pentamethylphenyl)-1H-indene (6a)
Table 1. Product Composition of the Thermal Transformation of 4,5,6,7,8-Pentamethyl-2H-cyclohepta[b]furan-
2-one (1a)
Solvent
Conc. [mol/l]
Time [h]
Temp. [8]
Yield [%]
1a
4aa)
5aa)
6aa)
Decalin
Decalin
Decalin
0.60
0.09
0.10
0.15
0.15
3.5
48
24
5
230
190
170
230
210
9
23
50
34
29
53
3.4
30
±
11
13
< 0.5
6
7
n.d.b)
n.d.
n.d.
±
Diethylene glycol diethyl ether
DMSO
6
0c)
±
a) Yield of purified compound with respect to reacted 1a. b) n.d. ˆ compound observed by TLC and/or HPLC,
but yield not determined. c) Complete decomposition of 1a.
Scheme 2a)
190 – 230°
+
O
O
5a
4a
1a
+
6a
a) See Table 1 for details.

1716
Helvetica Chimica Acta ± Vol. 84 (2001)
azulene, substituted at C(2) with a residue, which had to account for the other 5, symmetrically arranged Me
groups. The only structure that fulfilled these requirements was that of the (pentamethylphenyl)-substituted
azulene 4a. Indeed, the three s at 2.27, 2.25, and 2.23 ppm (2 :1:2) could be assigned to Me C(2')/Me C(4') and
Me C(4') via a 1H-NOESY spectrum of 4a, and to Me C(3')/Me C(5'), respectively, of the phenyl
.
‡
substituent at C(2). The MS (70 eV) of 4a with the M peak at m/z 344 (100%), followed by mass peaks at m/z
329 (49%), 314 (84%), and 299 (55%), indicating the consecutive loss of Me groups, was in full agreement with
the structure of 4a.
The 1H-NMR spectrum (500 MHz, CDCl3) of 6a displayed in the region for olefinic H-atoms an s at
6.72 ppm, counting for one H-atom, and a second s at 3.45 ppm, standing for two H-atoms, which could be
assigned to a CH2 group in allylic and/or benzylic surrounding. The region for aromatically positioned Me
groups showed three sharp s in a ratio of 1:2 :2 at 2.40, 2.26, and 2.14 ppm as well as a broad s at 2.30 ppm with a
slightly resolved shoulder at the lower-field flank of the signal, counting in a total for 4 Me groups. The presence
of indeed 9, aromatically bound Me groups could much better be seen in the 13C-NMR spectrum (CDCl3) of 6a,
which displayed, in the range of 18.2 ± 15.9 ppm, 7 Me signals in a ratio of 2 :1 :1:2 :1:1 :1. The fact that the
13C-NMR spectrum indicated two sets of (in pairs equivalent) Me groups was only compatible with the existence
of a (pentamethylphenyl) substituent. Indeed, specific 1H-NOE experiments and a 1H-NOESY plot of 6a
admitted the assignment of the two sharp s signals at 2.26 and 2.14 ppm to Me C(3')/Me C(5') and Me C(2')/
Me C(6') of the (pentamethylphenyl) unit. The signal of Me C(4') was buried in the broad s at 2.30 ppm. The
residual signals in the NMR spectra were compatible only with the presence of a 2-substituted 4,5,6,7-
tetramethyl-1H-indene unit. Moreover, the 1H-NOE experiments demonstrated that the s at 2.40 ppm belonged
to Me C(4), whereas Me C(7) was responsible for the shoulder of the s at 2.30 ppm. The two other Me groups
at C(5) and C(6), and Me C(4') caused the broad s at 2.30 ppm. These assignments were confirmed by the fully
‡
analyzed 13C-NMR spectrum of 6a. The mass spectrum (Cl) of 6a exhibited the expected [M ‡ 1] peak at m/z
.
‡
319 (100%), followed by a weak [(M ‡ 1) Me] peak at m/z 305 (6%).
The formation of 4a, 5a, and 6a was also observed in diethylene glycol diethyl ether
as solvent (cf. Table 1). However, in dimethyl sulfoxide (DMSO), complete decom-
position of 1a without product formation was recognized. On the other side, the
observed reaction seems to be restricted to 1a, since neither 4,6,8-trimethyl- (1b) nor
4,5,6,7,8-pentafluoro-2H-cyclohepta[b]furan-2-one (1d) gave, on heating in decalin as
the most suitable solvent (cf. Table 1), any products comparable to those from 1a
(Scheme 3). Cross experiments were more successful. When 1a was heated in 0.1m
solution in decalin in the presence of 3 mol-equiv. of 1b1) or equimolar amounts of 2H-
cyclohepta[b]furan-2-one (1c) itself, we found the cross products 4ba/5ba and 4ca/5ca,
respectively, as well as 6a, derived from 1a, in the corresponding reaction mixtures
(Scheme 3).
We can only speculate on the mechanisms of product formation. Nevertheless, some
facts are obvious. The appearance of 4a, 5a, and 6a on thermolysis of 1a must be the
result of bimolecular reactivities of 1a. The formation of 4a seems in this context most
transparent according to the mass equation: 2 Â [1a CO2] ! 4a. The fundamental
questions are, therefore: At which state does CO2 extrusion occur, and what
differentiates 1a from the other 2H-cyclohepta[b]furan-2-ones? X-Ray crystal-
structure analyses of 1a and, on grounds of comparison, nonsubstituted 2H-cyclo-
hepta[b]furan-2-one (1c) (both this work) as well as AM1 calculations of the structure
of 2H-cyclohepta[b]furan-2-ones, including 1b and 1d (see Table 2), demonstrate that
all of these compounds are perfectly planar, with the only exception of 1a. It is forced,
due to the crowding effect of the Me substituents, in a flat boat-like conformation of the
seven-membered ring, comparable to that of heptafulvenes, whereby the calculated
1
)
In the presence of equimolar amounts of 1b, mainly the products of 1a were formed.

Helvetica Chimica Acta ± Vol. 84 (2001)
1717
Scheme 3
R1
R2
R2
1b: 190°, 18 h; decalin
R1
no products
O
1d: 230°, 22 h; decalin
O
R1
1b R1 = Me, R2 = H
1d R1 = R2 = F
R
R
R
R
+
R
O
O
R
190°, 18 h
decalin
4ba (12.0%)
4ca (18.3%)
1a (1 mol-equiv.)
5ba (2.6%)
5ca (1.2%)
R
R
O
O
+
R
1b R = Me (3 mol-equiv.)
1c R = H (1 mol-equiv.)
6a (5%)
6a (6%)
Table 2. Comparison of Some Structural Data of 2H-Cyclohepta[b]furan-2-ones 1 from an X-Ray Crystal
a
Structure and AM1 Calculations
)
Parameter
1ab)
1bb)
1cb)
1db)
Interatomic distances d [pm]:
C(3) C(3a)
C(4) C(5)
C(6) C(7)
C(8) C(8a)
137.5(3)/137.3
137.6(3)/136.3
138.7(3)/136.5
136.1(3)/135.7
±/137.4
±/135.6
±/135.8
±/135.7
136.1(4)/137.2
134.4(4)/135.1
134.9(4)/135.3
134.9(4)/135.1
±/137.0
±/137.9
±/137.9
±/136.2
Bond angles W [8]:
O(1) C(2) C(3)
C(2) C(3) C(3a)
C(5) C(6) C(7)
Torsion angles V [8]:
C(3) C(3a) C(4) C(5)
106.9(2)/108.0
109.0(2)/108.0
129.6(2)/127.8
±/108.0
±/108.0
±/128.6
106.7(2)/108.0
109.4(2)/108.0
129.2(3)/129.8
±/107.8
±/107.9
±/129.8
178.6(2)/168.5
0.3(3)/14.8
179.1(2)/172.2
7.4(4)/10.7
16.3(4)/37.7
10.9/16.9
±/180.0
±/0.0
±/180.0
±/0.0
±/0.0
±/0.0
177.8(3)/180.0
±/0.0
178.3(4)/180.0
0.6(6)/0.0
0.8(6)/0.0
1.3(7)/0.0
±/180.0
±/0.0
±/180.0
±/0.0
±/0.0
±/0.0
C(3) C(3a) C(4)
R
C(3) C(3a) C(8a) C(8)
C(3a) C(4) C(5) C(6)
C(4) C(5) C(6) C(7)
C(5) C(6) C(7) C(8)
Least-squares deviations from plane [pm]:
5-membered ring
0.1/2.9
3.1/18.1
planar
planar
planar/planar
planar/planar
planar
planar
5- and 7-membered ring
a) See Schemes 2 and 3 for substitution patterns. b) First value, X-ray data; second value, AM1 data; for an older
X-ray crystal-structure analysis of 1c, see [5].

1718
Helvetica Chimica Acta ± Vol. 84 (2001)
structure of 1a is somewhat more twisted than indicated by its X-ray crystal structure. If
o
one compares the calculated DHof value for 1a ( 22.16 kcal ´ mol ) with DHf (1a) ˆ
1
34.69 kcal ´ mol that is obtained from DHof (1c) ˆ 0.56 kcal ´ mol and the
1
1
calculated Me-group increments of the corresponding 5 monomethyl-substituted 2H-cyclo-
1
hepta[b]furan-2-ones (S ˆ 34.13 kcal ´ mol ), the observed difference of 12.5 kcal ´
mol 1 can be taken as the strain energy of deformation of 1a, due to the Me-crowding
effect in 1a. Indeed, the calculated DHfo value for 1b ( 20.18 kcal ´ mol ) deviates only
1
1
slightly (by 0.26 kcal ´ mol ) from the value estimated by the described incremental
procedure ( 20.44 kcal ´ mol ), indicating that 1b is free of group-crowding strain.
1
On the other hand, the HOMO, LUMO, and NLUMO coefficients of 1a ± 1c are as
expected quite similar (cf. Fig. 4) [4] (cf. also [1]). They should admit at least two
different types of orbital-symmetry controlled, thermal reactions, namely a first-order
transformation, steered by the HOMO of 1, which will result in disrotatory ring closure
by bond formation between C(3a) and C(8) as well as a second-order [8 ‡ 2]
cycloaddition reaction between two molecules of 1, governed by HOMO-LUMO
interactions. Despite the fact that 1a is formed by vacuum flash pyrolysis of (2,3,4,5,6-
pentamethyl)phenyl prop-2-ynoate at ca. 6508 [2], it may undergo in solution, due to its
Me-crowding strain, intramolecular ring closure to 7a, which can relieve strain by
homolysis of the O C bond. The thus formed diradical 8a may then undergo a
O
1
1
1
1
O
HOMO (-8.799)
LUMO (-1.249)
NLUMO
(-0.540)
1a
O
1
1
1
1
O
HOMO (-8.604)
LUMO (-1.225)
NLUMO
(-1.011)
1b
O
1
1
1
1
O
HOMO (-8.660)
LUMO (-1.177)
NLUMO
(-0.999)
1c
0.30 - 0.399
Coefficients:
> 0.40
0.20 - 0.299
0.10 - 0.199
0 - 0.099
Fig. 4. HOMO, LUMO, and NLUMO coefficients of the 2H-cyclohepta[b]furan-2-ones 1a ± c (taken from [4];
geometry optimization with ꢀGaussian-94ꢁ; semi-empirical calculations with MNDO)

Helvetica Chimica Acta ± Vol. 84 (2001)
1719
fragmentation reaction under extrusion of CO2 and formation of 1-ethynyl-2,3,4,5,6-
pentamethylbenzene (9a) (cf. Scheme 4). The latter compound could add in a [8 ‡ 2]
fashion to 1a (cf. [4]). Loss of CO2 from the cycloadduct 10a would finally yield 4a. We
studied this reaction sequence by thermal reaction of independently synthesized 9a
with 1a in decalin at 1908. However, the only product that was isolated in 89% yield,
after 15 h heating, was the [4 ‡ 2] cycloadduct 11a. This result demonstrates that the
discussed reaction sequence (cf. Scheme 4) cannot be responsible for the formation of
the main product 4a.
The [8 ‡ 2] cycloaddition reaction between two molecules of 1a follow, in principle,
the normal reactivity of 2H-cyclohepta[b]furan-2-ones in the presence of electron-
donor substituted ethenes, with the sole difference that a second molecule of 1a takes
the part of the electron-donor substituted ethene. Four different transition-state
arrangements of the two molecules of 1a are thinkable, resulting in four diastereoiso-
meric dimers 12a, which may be named according to the relative orientation of the two
CˆO groups syn or anti and exo or endo with respect to the relative position of the
anellated furanone ring at the formed central 2-oxanorbornan-3-one substructure. The
AM1-calculated structures of the four possible diastereoisomers of 12a are displayed
together with their calculated DHof values in Fig. 5. The enthalpy of formation of two
1
1
mol of 1a amounts to 2 ´ ( 22.16 kcal ´ mol ) ˆ 44.32 kcal ´ mol , which demon-
Scheme 4
190°
O
O
O
•
–CO2
O
O
•
O
1a
7a
8a
4a
O
–CO2
O
[8 + 2]
[4 + 2]
+
1a
10a
9a
O
O
11a (89%)

1720
Helvetica Chimica Acta ± Vol. 84 (2001)
Fig. 5. Stereoscopic view of the AM1-calculated structures of the four possible [8 ‡ 2] cycloaddition products of
1a, C ˆ magenta, H ˆ blue, O ˆ yellow.

Helvetica Chimica Acta ± Vol. 84 (2001)
1721
strates that all four cycloaddition reactions are possible, whereby the formation of
1
anti,exo-12a is most favorable with a gain in enthalpy of 11.60 kcal ´ mol . The next
step, to be expected according to the known cycloaddition/cycloreversion chemistry of
2H-cyclohepta[b]furan-2-ones, should be the extrusion of CO2 from the 2-oxanor-
bornan-3-one part of 12a. The steps leading finally to 4a are displayed in Scheme 5, and
the corresponding AM1-calculated structures are shown in Fig. 6. At a certain stage of
the transformations, ring contraction of one of the seven-membered rings to the
(pentamethylphenyl)-substituent has to take place. This normally happens in heptalene
chemistry via the cycloheptatriene > norcaradiene equilibrium, followed by cleavage
of one of the ꢀoldꢁ cyclopropane bonds (cf. [6][7] as well as Sect. 3). We believe that in
the present case, intermediate 13a undergoes ring contraction with its cycloheptatriene
ring B, which possesses, in contrast to ring A, the shorter C,C distance for bond
formation (ring B, 247 pm, and ring A, 255 pm) due to the larger torsion angles along its
conjugated triene system. The DDHof value of 13a and its ring-closed form 14a amounts
1
to 18.37 kcal ´ mol (cf. Fig. 6), i.e., does not forbid the ring-closure reaction.
Homolysis of the cyclopropane bond that includes C(2) of the 1,2-dihydroazulene
substructure creates a strain-relaxed highly delocalized diradical 15a, which, on loss of
CO2, concerted or non-concerted stepwise, gives 4a in an overall strongly exothermic
1
reaction (cf. Fig. 6). The dimer anti,endo-12a (cf. Fig. 5), which lies by 4.79 kcal ´ mol
higher in DHof than anti,exo-12a will also give 13a on loss of CO2, so that it also may
take part to a certain extent in the formation of 4a. However, the discussed reaction
sequence, applied to the other two dimers of 12a, with syn orientation of the two CˆO
groups and the two seven-ring systems, therefore higher lying in DHof than the two anti-
forms of 12a (cf. Fig. 5), would lead finally to 4,5,6,7,8-pentamethyl-1-[(2,3,4,5,6-
pentamethyl)phenylazulene, an isomer of 4a, which was not observed.
The failure of the other 2H-cyclohepta[b]furan-2-ones to undergo the reactions
observed with 1a as well as the successful cross experiments with 1b and 1c (cf.
Scheme 3) are in full agreement with this sketched reactivity of 1a. The formation of
the azulenes 4ba and 4ca, respectively, in the cross reactions clearly indicate, on the
basis of the postulated Scheme 5 and Figs. 5 and 6, that 1a entered the cycloaddition
Scheme 5
B
anti,exo-12a
A
– CO2
O
O
13a
•
•
4a
– CO2
O
H
O
H
O
O
14a
15a

1722
Helvetica Chimica Acta ± Vol. 84 (2001)
Fig. 6. Stereoscopic view of the calculated structures of the intermediates 13a, 14a (AM1), and the p2-diradical
15a (MM2; cf. Scheme 5) on the postulated reaction path of anti,exo-12a to 4a. For colors, see Fig. 5

Helvetica Chimica Acta ± Vol. 84 (2001)
1723
reaction as 2-atom component, whereas 1b or 1c acted as its 8-atom reaction partner.
The other two possible azulenes 4ab and 4ac, respectively, with exchanged azulene and
phenyl part were not observed in the corresponding reaction mixtures. Table 3 contains
the AM1-calculated DHof values of all anti-intermediates 12 of the cycloaddition
reactions of 1, including the cross reactions of 1a with 1b or 1c. The relevant structures
of anti,exo-12ba and anti,exo-12ca that will lead to the observed ꢀcrossꢁ azulenes 4ba
and 4ca, respectively, are displayed in Fig. 7.
Table 3. Comparison of the AM1-Calculated DHof Values of the Possible anti-Addition Products 12 of the 2H-
Cyclohepta[b]furan-2-ones 1a)
Addition products 12
Calculated DHof
Sum of DHof of starting 1
Gain in DHof of 12 (DDHof )b)
anti,exo-12ac)
anti,exo-12b
anti,exo-12c
55.92
45.03
6.36
48.00
52.65
28.72
33.77
51.13
43.00
4.90
44.32
40.36
1.12
42.34
42.34
22.72
22.72
44.32
40.36
1.12
11.60
4.67
5.24
5.66
10.31
6.00
11.05
6.81
2.64
3.78
0.76
8.71
1.43
9.28
anti,exo-12abd)
anti,exo-12bad)
anti,exo-12acd)
anti,exo-12cad)
anti,endo-12ac)
anti,endo-12b
anti,endo-12c
anti,endo-12abd)
anti,endo-12bad)
anti,endo-12acd)
anti,endo-12cad)
43.10
51.05
24.15
32.00
42.34
42.34
22.72
22.72
a) DHof and DDHof in kcal ´ mol
.
b) Underlined DDHof values belong to the reaction sequence of the observed
1
azulenes 4. c) See Fig. 5. d) The first letter indicates the component 1 that will become in the course of the
reaction the azulene part; the second letter stands for the component 1 that will become later the phenyl
substituent at C(2) of the formed azulene ring.
It is striking that all postulated anti-intermediates 12 that should lead, according to
Scheme 5, to the observed azulenes 4a, 4ba, and 4ca exhibit indeed the largest gain in
enthalpy in comparison to the sum of DHof of the starting 2H-cyclohepta[b]furan-2-
ones 1, whereby the anti,exo-diastereoisomers of 12 possess in all cases the larger
difference than the corresponding anti,endo-diastereoisomers. The DDHof values for
anti,exo-12b and -12c in comparison to that of anti,exo-12a also explain why pure 1b and
1c, respectively, do not undergo the dimerization reaction under the applied conditions.
o
1
The gain in enthalpy (DDHf ) is well below 10.3 kcal ´ mol , which is the lowest for the
cycloaddition reaction that finally gives 4ba. Moreover, this value is by 0.74 kcal ´ mol
1
lower than that for the formation of anti,exo-12ca. The results are in full accord with
these findings, since 4ca is already formed as the major azulene when equimolar
amounts of 1a and 1c are reacted, whereas the formation of 4ba as major azulene
requires a 3-fold molar amount of 1b in the presence of 1 mol-equiv. of 1a. The
conclusion is that in crossed cycloaddition reactions of 1a with other 2H-cyclo-
hepta[b]furan-2-ones, 1a is an equivalent for the introduction of the (2,3,4,5,6-
pentamethyl)phenyl substituent at C(2) of azulenes, with a substitution pattern at the
seven-membered ring that is determined by the other component 1. The reason for this

1724
Helvetica Chimica Acta ± Vol. 84 (2001)
Fig. 7. Stereoscopic view of the AM1-calculated structures of the energetically most favorable [8 ‡ 2] anti,exo-
cycloadducts 12ba and 12ca, leading to the ꢀmixedꢁ azulenes 4bc and 4ca, respectively (for colors, see Fig. 5)
chemoselectivity is apparent. The Me-crowding energy of 1a (vide supra) is best
reduced in the transition state of the cycloaddition reaction when 1a takes part as 2-
atom component, which leads to a cycloadduct 12 where the seven-membered ring of
1a becomes the strongly puckered cycloheptatriene ring, which is spiro-linked to the
central 2-oxanorban-3-one unit. The other cycloheptatriene ring, which is anellated to
this central unit along the C(1) C(7) bond, is forced in a planar conformation due to
the bicyclo[2.2.1] nature of the central unit.

Helvetica Chimica Acta ± Vol. 84 (2001)
1725
We believe that the central diradical 15a is also responsible for the formation of the
pentamethylazulene 5a as well as for the indene 6a in the thermal reaction of 1a.
Intramolecular disproportionation of 15a via 1,4-H transfer will lead to the
intermediate 16a (Scheme 6). The AM1-calculated DHof value of 16a (4.29 kcal ´
1
1
mol ) is close to that of the precursor 14a, which amounts to 1.00 kcal ´ mol .
Concerted ± in the sense of a linear cheletropic fragmentation ± or non-concerted, via
stepwise homolysis, loss of CO2 will give azulene 5a and the benzyl-phenyl-centered
diradical 17a, which may also be formulated as a 6-methylidenecyclohexa-2,4-dien-1-
ylidene. It seems that 17a is trapped by 1a, but neither by 1b nor 1c in the cross
experiments, in a [2 ‡ 3] 6e-cycloaddition reaction resulting in the formation of 18a2).
The concertedness of [2 ‡ 3] cycloaddition reactions with singulet 1,3-diradicals and
alkenes is well-established (see, e.g., [8]). As an intermediate with a cycloheptatriene
substructure, 18a may undergo intramolecular disrotatory ring closure to 19a. This
intermediate possesses again the possibility of a concerted or non-concerted, stepwise
extrusion of CO2 under cleavage of one of the ꢀoldꢁ cyclopropane bonds and formation
of the 1H-indene derivative 6a.
3. Heptalene Formation. ± The thermal reaction of the 2-(pentamethylphenyl)-
substituted azulenes 4a and 4ba and the pentamethyl-substituted heptalene 5a with an
excess of dimethyl acetylenedicarboxylate (ADM) in toluene gave in all three cases the
expected heptalenedicarboxylates as a mixture of their double-bond-shifted (DBS)
isomers (Scheme 7). Small amounts of the corresponding azulene-1,2-dicarboxylates 21
and 23 were also formed, as it is generally observed for thermal reactions of azulenes
Scheme 6a)
H
1,4-H
+
15a
•
transfer
– CO2
O
H
•
O
16a (4.29)
5a
17a
[2 + 3]
+
1a
17a
O
6a
– CO2
O
H
O
(–79.86)
O
18a (–45.55)
19a (–29.00)
(15.14)
a) In parentheses, AM1-calculated DHof values in kcal ´ mol
.
1
2
)
The product resulting from the reverse regioselectivity may also be formed. However, its DHof value is,
according to AM1 calculations, by more than 10 kcal ´ mol higher than that of 18a, which makes its
1
formation unlikely. Nevertheless, its transformation, following Scheme 6, will also lead to 6a and 17a, due to
the pseudo-symmetry plane of 6a.

1726
Helvetica Chimica Acta ± Vol. 84 (2001)
Scheme 7a)
R2
E
R2
R2
R2
R3
R3
R3
R1
R3
E
E
a)
R1
R2
E
R2
R2
R2
+
+
R3
R1
R3
E
R3
R3
R2
R2
E
R2
R2
5a R1 = H, R2 = R3 = Me
20
20'
22'
(<1%)
(9%)
(16%)
(34%)
21
23
4ba R1 = PMP, R2 = Me, R3 = H
4a R1 = PMP, R2 = R3 = Me
22
24
(<1%)
(39%)
R1
21
24'
R2
R3
E
E = COOMe
PMP =
+
E
R2
R2
(n.o.)
25
26
(n.o.)
(3%)
27
a) 4 ± 8 mol-equiv. of ADM; toluene, 1308, 10 ± 1800 min.
a) For details, see Exper. Part.
with ADM (cf. [9 ± 11]). The total yield of the DBS mixture of the pentamethyl-
heptalenedicarboxylates 20/20' was distinctly lower than those of the heptalene-
dicarboxylates 22/22' and 24/24' with the (pentamethylphenyl) group at C(2)/C(4) due
to the fact that azulene 5a is thermally much less stable (cf. [1]) than the azulenes 4a and
4ba with the (pentamethylphenyl) residue at C(2).
The thermal reaction of 4a and ADM gave, as further product, the acenaphthylene-
3,4-dicarboxylate 27 (Scheme 7). It must have been formed by a similar mechanism as
other products of this type (cf. [5][6]). However, in contrast to our former findings,
where dehydrogenation or leaving-group elimination has to take place in the final step
of aromatization, we have to assume that in the present case, methane is split off
(Scheme 8). The formation of analog products (25 or 26) were not observed in the two
Scheme 8
– CH4
H
PMP
PMP
+
4
ADM
27
H
E
E
E
E
28
29

Helvetica Chimica Acta ± Vol. 84 (2001)
1727
other thermal reactions. The structure of 27 was solved unambiguously by an X-ray
crystal-structure determination (see Exper. Part).
All new heptalenedicarboxylates could easily be equilibrated thermally (cf.
Scheme 9). The DBS process could also be induced photochemically (see Exper.
Part). The thermal equilibrium ratios of the heptalenedicarboxylates do not differ very
much. It seems, however, that the Me5 substitution pattern as in 20/20' or 24/24' favors
the 4,5-dicarboxylate structures slightly more than the 6,8,10-Me3 pattern as present in
22/22' or 30/30', in agreement with corresponding AM1 calculations of DHof , whereby
the (pentamethylphenyl) or a phenyl substituent at C(2)/C(4) has no significant
influence on the equilibrium ratios. The two Me substitution patterns have a much
stronger impact on the rate (k1 ‡ k ) of the equilibration process. One observes an
1
expressed buttressing effect of the 5 adjacent Me groups in 20/20' and 24/24' in
comparison with the 3 non-adjacent Me groups in 22/22' and 30/30'. The two DBS
isomers 24 and 24' are thermally stable enough at room temperature that both isomers
can be separated by chromatography. Similarly, the ratio of the photo-equilibrium
mixture 20/20' (54 :46; see Exper. Part) is thermally not altered at room temperature.
First kinetic experiments in toluene gave for 24/24' a total half-life (t1/2(24) ‡ t1/2(24'))
of ca. 45 min and for 20/20' of ca. 22 min at 1008. In contrast to these findings, pure
crystalline 22 equilibrates already on dissolution in toluene at room temperature, so
that we were not able to measure the kinetics of equilibration under our standard
conditions. The same is true for 30/30' (cf. [12]).
Scheme 9a)
R2
E
R2
R3
R3
E
R1
1
2
100°
R2
R2
toluene
R3
R1
R3
E
R2
R2
E
20
90 : 10
74 : 26
84 : 16
20'
22'
24'
30'
22
24
30
77 : 23b)
a) R1 to R3, see Scheme 7 and Table 5 for 30; equilibrium ratios in %. b) Hexane, 208 [12].
The structure of the 2-(pentamethylphenyl)-substituted heptalene-4,5-dicarboxy-
lates 22 and 24 were ascertained by X-ray crystal-structure analyses (see Exper. Part
and Tables 4 and 5). The crystals of 24 exhibited unfortunately twinned regions3). The
structure refinement was, therefore, distinctly worse than that of 22, which gave
standard R values. Nevertheless, the structure of 24 and its relative configuration was
clearly solved (cf. Fig. 8). A comparison of the skeletal torsion angles V of 22 and 24
3
)
A number of further attempts to crystallize 24 did not improve the quality of the crystals due to the fact that
24 is highly soluble in all common organic solvents, so that only hexane was left for crystallization
experiments.

1728
Helvetica Chimica Acta ± Vol. 84 (2001)
Fig. 8. Stereoscopic view of the X-ray crystal structure of the sterically most crowded 4,5,6,7,8-pentamethyl-2-
(2,3,4,5,6-pentamethylphenyl)azulene (24)
Table 4. Comparison of Some X-Ray Crystal-Structure Data of the Heptalene-4,5-dicarboxylates 22, 24, and 30a)
22
24
30
Bond angles W [8]:
C(1) C(10a) C(5a)
C(5) C(5a) C(10a)
C(5a) C(10a) C(10)
C(6) C(5a) C(10a)
C(7) C(8) C(9)
113.5(2)
121.5(2)
123.7(2)
113.0(2)
124.3(3)
116.4(10)
121.2(9)
124.1(10)
113.3(10)
125.2(9)
115.1(1)
121.8(1)
122.7(1)
114.4(1)
123.0(1)
Torsion angles V [8]:
C(4) C(5) C(5a) C(6)
C(5) C(5a) C(6) C(7)
C(5) C(5a) C(10a) C(10)
C(1) C(10a) C(5a) C(6)
C(1) C(10a) C(10) C(9)
C(10) C(10a) C(1) C(2)
C(3) C(4) C(1'')ˆO
178.5(2)
123.5(3)
122.7(3)
116.9(3)
173.6(3)
121.7(3)
178.1(3)
44.4(4)
172.3(9)
127.5(11)
125.4(11)
115.2(10)
179.7(10)
127.7(12)
21.5(17)
31.6(16)
177.8(10)
176.8(9)
85.6(12)
40.5(14)
172.8(1)
120.2(1)
123.6(1)
114.4(1)
175.3(1)
124.3(1)
25.1(2)
154.8(1)
176.2(1)
75.6(1)
C(5a) C(5) C(1''')ˆO
C(2) C(3) C(4) (1'')
C(1''') C(5) C(5a) C(10a)
C(1) C(2) C(1') C(2')
C(1'') C(4) C(5) C(1''')
177.7(3)
176.6(2)
86.9(3)
31.2(2)
41.6(2)
37.5(3)
a) The quality of the crystals of 24 could not be improved by recrystallization. For heptalenedicarboxylate 30,
see Table 5.
and of heptalene-4,5-dicarboxylate 30 as the phenyl analog of 22 (see Tables 4 and 5)
shows that all three heptalene-4,5-dicarboxylates possess in the crystals similar Vat the
MeOCO-substituted ring as well as at the central s-bond, common for both rings. It
seems that an average V value of 60 ± 618 at the central s-bond represents the utmost
size that is structurally possible by the anellation of two seven-membered, fully
unsaturated rings. The central V values of 24 as the sterically most encumbered
heptalene-4,5-dicarboxylate fall exactly in this range. The steric congestion of the 5
adjacent Me groups of 24 is compensated by the peripheral torsion angles of the seven-

Helvetica Chimica Acta ± Vol. 84 (2001)
1729
Table 5. Torsion Angles V from X-Ray Crystal Structure Analyses of the 2-(Pentamethylphenyl)-Substituted
Heptalene-4,5-dicarboxylates 22 and 24 in Comparison with Those of the 2-Phenyl Analog 30 of 22
Heptalene-4,5-dicarboxylatea)
22
30b)
24c)
s-cis-Buta-1,3-diene subunits torsion angles V [8]:
C(1) C(2) C(3) C(4)
C(3) C(4) C(5) C(5a)
C(6) C(7) C(8) C(9)
C(8) C(9) C(10) C(10a)
38.3(5)
39.5(4)
34.0(5)
29.8(5)
31.4(2)
37.2(15)
36.9(2)
34.1(2)
34.4(2)
36.3(16)
48.3(19)
37.4(18)
cis-Ethene subunits torsion angles V [8]:
C(2) C(3) C(4) C(5)
C(10a) C(1) C(2) C(3)
C(5a) C(6) C(7) C(8)
0.4(5)
1.9(5)
2.9(5)
2.2(5)
4.5(2)
1.1(2)
5.6(2)
1.5(2)
1.2(18)
5.5(16)
7.8(16)
6(2)
C(7) C(8) C(9) C(10)
Central s-bond cisoid torsion angles V [8]:
C(10a) C(1) C(2) C(3)
C(10a) C(1) C(2) C(3)
59.3(4)
61.2(4)
60.7(2)
61.4(2)
58.0(13)
61.4(13)
a) Torsion angles V are given for the (P)-configuration. b) Data for 30 are taken from [12];
1
2
1'
3
2'
30
5a
4
1''
5
COOMe
COOMe
1'''
c) The large e.s.d. values are due to the bad quality of the crystals of 24, which could not be improved despite
several recrystallizations.
membered ring that carries the Me groups. Its torsion angles at the s-cis-buta-1,3-dienyl
subunits are in average 108 larger than those at the 6,8,10-Me3 substituted rings in 22
and 30. The transition state of the DBS process in heptalenes must be still twisted since
it occurs with retention of the configuration of the inherently chiral heptalene skeleton
[13] (see also [14]). Therefore, we have to expect that adjacent Me groups will come
closer in the transition state since all peripheral torsion angles will be reduced in the
transition state to allow the cyclic p-bond jump from one DBS isomer to the other4).
Finally, it is worthwhile to mention that the 1H- and 13C-NMR spectra of 22 and 24
and their DBS isomers 22' and 24' indicate a restricted rotational freedom of the 2-
4
)
Similar observations have been made by Paquette et al. for the cyclic p-bond shift in highly methylated
cyclooctatetraenes (see [15] and ref. cit. therein). DFT Calculations that we have performed on heptalenes
[16], partly in cooperation with Prof. W. Thiel and Dr. E.-U. Wallenborn, show that the lowest excited state
of heptalenes is still twisted, however, with equalized bond lengths. Therefore, we believe that the DBS
process in heptalene is characterized by touching of the hypersurfaces of the ground and excited singlet
state, in agreement with all our experiments, especially with respect to the observed retention of
configuration in thermally or photochemically induced DBS in optically active heptalenes [13]. We will
come back to this point in a planned publication [16].

1730
Helvetica Chimica Acta ± Vol. 84 (2001)
(pentamethylphenyl) substituents, in a way that all four forms exhibit 5 sharp s for the 5
aromatically bound Me groups due to the inherent chirality of the heptalene skeleton.
We thank Dr. H. W. Schmalle, Anorganisch-chemisches Institut der Universität Zürich, and Dr. A. Linden
for the X-ray crystal-structure determinations, Dipl.-Chem. M. Nagel for the arrangement of Fig. 4, our NMR
laboratory, in particular Nadja Walch and Dr. Gudrun Hoppe, for specific NMR measurements, Prof. M. Hesse
and his group for mass-spectrometric measurements, and our Analytical Laboratory for elemental analyses. The
financial support of this work by the Swiss National Science Foundation is gratefully acknowledged.
Experimental Part
General. HPLC: Bischoff HPLC pump model 2200 with an anal. column Spherisorb Nitrile (3 mm, 10 Â
3 mm); photodiode array detector (Waters 991). Prep. HPLC: Du Pont Instruments-830 liquid chromatograph;
prep. column (250 Â 20 mm) with the stationary phase Spherisorb S5 CN. UV Spectra: Perkin-Elmer Lambda-19
1
spectrometer, lmax (log e) in nm. IR Spectra: Perkin-Elmer Spectrum-One spectrometer; in cm
.
1H- and
13C-NMR Spectra: Bruker instruments ARX-300 (300 and 75 MHz, resp.), Avance DRX-500 (500 and 125 MHz,
resp.), and Avance DRX-600 (600 and 150 MHz, resp.); chemical shifts d in ppm rel. to CD(H)Cl3 (7.270 and
77.00 ppm, resp.); complete assignments of 1H signals by COSY, TOCSY, NOESY, ROESY 2D or 1D NMR
methods; complete assignments of 13C-signals by 2D HMBC and HSQC methods. Electron-impact (EI) MS:
Finnigan SSQ-700 mass spectrometer. Chemical-ionization (CI) MS: Varian MAT-90 mass spectrometer,
ionization via methane.
2H-Cyclohepta[b]furan-2-ones 1a ± c were synthesized according to [1]. The 4,5,6,7,8-pentafluoro-2H-
cyclohepta[b]furan-2-one (1d) was prepared according to a published procedure [17]. All reactions were
performed under Ar.
1. Azulenes by Thermolysis of 2H-Cyclohepta[b]furan-2-ones 1. 1.1. Thermolysis of 4,5,6,7,8-Pentamethyl-
2H-cyclohepta[b]furan-2-one (1a). A soln. of 1a (1.3 g, 6.01 mmol) in decalin (10 ml) was flushed with Ar in a
Schlenk vessel. The vessel was closed and heated at 230 ± 2408 during 3.5 h. The solvent was evaporated. The
residual black solid was then immediately chromatographed on a short column (Al2O3, gradient hexane !
CH2Cl2/acetone 90 :10) to separate the blue azulene fraction (390 mg) from starting material (0.113 g, 9%) and
polymeric stuff. The subsequent CC (silica gel, hexane) of the azulene fraction gave 4a (0.269 g, 26%), 5a
(0.06 g, 10%), and 6a (0.06 g, 6%).
Data of 4,5,6,7,8-Pentamethyl-2-(2,3,4,5,6-pentamethylphenyl)azulene (4a): Light blue crystals. M.p. 272.3 ±
272.78 (hexane). UV/VIS (hexane; cf. Fig. 2): 566 (2.76), 370 (sh, 3.48), 360 (3.74), 351 (sh, 3.70), 304 (4.64),
1
255 (4.29), 247 (sh, 4.23), 200 (4.77); min. 425 (2.08), 339 (3.65), 267 (4.02), 232 (4.20). H-NMR (500 MHz,
C6D6): 7.30 (s, H C(1,3)); 2.62 (s, Me C(4,8)); 2.28 (s, Me C(2',6')); 2.25 (s Me C(4')); 2.23 (s, Me C(3',5'));
2.18 (s, Me C(6)); 2.15 (s, Me C(7,5)). 13C-NMR (125 MHz, CDCl3): 150.12 (s, C(2)); 144.13 (s, C(6)); 143.36
(s, C(4,8)); 139.29 (s, C(3a,8a)); 138.56 (s, C(1')); 133.64 (s, C(4')); 132.33 (s, C(3',5')); 132.26 (s, C(2',6')); 130.98
(s, C(5,7)); 116.28 (d, C(1,3)); 24.72 (q, Me C(6)); 22.99 (q, Me C(5,7)); 22.86 (q, Me C(4,8)); 19.63 (q,
.
‡
‡
Me C(2',6')); 17.19 (q, Me C(4')); 17.13 (q, Me C(3',5')). EI-MS: 344.2 (100, M ), 329.2 (49, [M Me] ),
‡
‡
314.1 (84, [M 2 Me] ), 299.1 (55, [M 3 Me] ). Anal. calc. for C26H32 (344.53): C 90.64, H 9.36; found: C
90.34, H 9.44.
Data of 4,5,6,7,8-Pentamethylazulene (5a): Blue crystals. M.p. 86.3 ± 93.18 (hexane). UV/VIS (hexane; cf.
Fig. 1): 574 (2.78), 366 (sh, 3.46), 356 (3.46), 344 (sh, 3.60), 294 (4.57), 252 (4.38), 200 (4.21); min 420 (2.10), 330
(2.52), 266 (4.00), 224 (4.04). 1H-NMR (300 MHz, C6D6): 7.64 (t, 3J(2,1) ˆ 3J(2,3) ˆ 4.0, H C(2)); 7.27 (d,
3J(1,2) ˆ 3J(3,2) ˆ 4.0, H C(1,3)); 2.45 (s, Me C(4,8)); 1.94 (s, Me C(6)); 1.92 (s, Me C(5,7)). 13C-NMR
(150 MHz, CDCl3): 145.30 (s, C(6)); 144.34 (s, C(4,8)); 137.53 (s, C(3a,8a)); 132.29 (d, C(2)); 130.44 (s, C(5,7));
.
‡
114.10 (d, C(1,3)); 24.72 (q, Me C(6)); 22.96 (q, Me C(5,7)); 22.77 (q, Me C(4,8)). EI-MS: 198.0 (95, M
C15H18 ), 183.0 (100, [M Me] ).
,
.
‡
‡
Data of 4,5,6,7-Tetramethyl-2-(2,3,4,5,6-pentamethylphenyl)-1H-indene (6a): Colorless crystals. M.p.
280.0 ± 282.88 (hexane). UV (hexane): 272 (4.11), 220 (sh, 4.44), 202 (4.67); min. 249 (3.83). 1H-NMR
(500 MHz, CDCl3): 6.72 (s, H C(3)); 3.46 (s, 2 H C(1)); 2.41 (s, Me C(4)); 2.31 (shoulder of s, Me C(7));
2.30 (s, Me C(5), Me C(6), Me C(4')); 2.67 (s, Me C(3'), Me C(5')); 2.15 (s, Me C(2'), Me C(6')).
13C-NMR (125 MHz, CDCl3): 147.63 (s, C(2)); 141.60 (s, C(3a)); 139.47 (s, C(7a)); 135.86 (s, C(1')); 133.97 (s,
C(4')); 133.43 (s, C(5)); 132.36 (s, C(3', C(5')); 131.83 (s, C(2'), C(6')); 131.20 (s, C(7)); 129.03 (d, C(3)); 128.81
(s, C(6)); 126.04 (s, C(4)); 43.13 (t, C(1)); 18.16 (q, Me C(2'), Me C(6')); 16.71 (q, Me C(4') or Me C(5) or

Helvetica Chimica Acta ± Vol. 84 (2001)
1731
Me C(6)); 16.63 (q, Me C(7)); 16.53 (q, Me C(3'), Me C(5')); 16.32 (q, Me C(4)); 16.10 (q, Me C(4') or
.
.
‡
‡
Me C(5) or Me (6)); 15.97 (q, Me C(4') or Me C(5) or Me C(6)). EI-MS: 318.1 (100, M , C24H30 ),
.
.
‡
‡
‡
‡
303.1 (80, [M Me] ), 288.1 (94, [M 2 Me] ), 147.0 (52, [Me5C6] ). CI-MS: 319.3 (100, [M ‡ 1] ), 305.3 (6,
.
‡
[M ‡ 1 Me] ).
1.2. Cross Experiment with 1a and 4,6,8-Trimethyl-2H-cyclohepta[b]furan-2-one (1b). A mixture of 1a
(1.26 g, 5.8 mmol) and 1b (3.28 g, 17.4 mmol) in decaline (40 ml) was stirred under Ar at 1808. After 18 h, the
mixture was worked up according to Exper. 1.1. Repetitive CC yielded 4ba (0.218 g, 12%), 5ba (0.026 g, 2.6%),
colorless 6a (0.045 g, 4.9%), and recovered 1a (0.06 g, 5%) and 1b (2.42 g, 74%).
Data of 4,6,8-Trimethyl-2-(2,3,4,5,6-pentamethylphenyl)azulene (4ba): Violet-blue crystals. M.p. 188.5 ±
192.48 (hexane). UV/VIS (hexane): 560 (sh, 2.64), 537 (266), 372 (sh, 3.40), 355 (3.79), 340 (3.63), 314 (4.26),
292 (4.69), 242 (4.37), 199 (4.67); min. 415 (1.87), 345 (3.62), 332 (3.60), 261 (3.78), 232 (4.30). 1H-NMR
(600 MHz, C6D6): 7.34 (s, H CH(1,3)); 6.85 (s, H C(5,7)); 2.65 (s, Me C(4,8)); 2.32 (s, Me C(6)); 2.24 (s,
Me C(4')); 2.21 (s, Me C(3',5')); 2.20 (s, Me C(2',6')). 13C-NMR (150 MHz, C6D6): 150.62 (s, C(2)); 144.89 (s,
C(6)); 144.65 (s, C(4,8)); 138.07 (s, C(1')); 137.68 (s, C(3a,8a)); 133.86 (s, C(4')); 132.40 (s, C(3',5')); 132.19 (s,
C(2',6')); 127.89 (d, C(5,7)); 118.29 (d, C(1,3)); 28.77 (q, Me C(6)); 25.28 (q, Me C(4,8)); 19.58 (q,
.
‡
‡
Me C(2',6')); 17.18 (q, Me C(4')); 17.09 (q, Me C(3',5')). EI-MS: 316.1 (96, M ), 301.1 (72, [M Me] ),
‡
‡
286.1 (100, [M 2 Me] ), 271.0 (60, [M 3 Me] ). Anal. calc. for C24H28 (316.48): C 91.08, H 8.92; found: C
90.91, H 9.05.
1.3. Cross Experiment with 1a and Cyclohepta[b]furan-2-one (1c). According to Exper.1.2, a mixture of 1a
(0.013 g, 0.062 mmol) and 1c (0.009 g, 0.062 mmol) in decalin (0.5 ml) was stirred in a Schlenk vessel during 2.5 h
at 2308. HPLC/UVand GC/MS (Spherisorb CN, 3 mm; 125 Â 40 mm; hexane) of a raw reaction mixture, before
chromatography, confirmed the presence of 4ca, 4a, 5ca, 5a, and 6a; HPLC/UV: 4ca/4a/5ca/5a/6a
17.4 :5.8 :6.2 :1:6.6. Workup according to Exper. 1.1 gave 4ca (3.1 mg, 18.3%), 5ca (0.1 mg, 1.2%), 1a
(0.0016 g, 12%), 1c (4.4 mg, 49%), and 4a/5a/6a (2.6 mg).
Data of 2-(2,3,4,5,6-Pentamethylphenyl)azulene (4ca): Blue crystals. M.p. 180.5 ± 186.28 (hexane). UV/VIS
(hexane): 672 (2.10), 613 (2.41), 569 (2.44), 545 (sh, 2.32), 365 (sh, 3.32), 347 (3.62), 330 (sh, 3.63), 317 (3.93),
302 (4.15), 282 (4.72), 275 (4.71), 232 (sh, 4.20), 200 (4.66); min. 439 (1.67), 363 (3.32), 340 (3.56), 315 (3.92),
300 (4.15), 278 (4.70), 251 (4.05). 1H-NMR (600 MHz, CDCl3): 8.31 (d, 3J(4,5) ˆ 3J(8,7) ˆ 9.2, H C(4,8)); 7.57
(t, 3J(6,5) ˆ 3J(6,7) ˆ 9.9, H C(4,8)); 7.22 (s, H C(1,3)); 7.20 (t, 3J(5,4) ˆ 3J(5,6) ˆ 3J(7,6) ˆ 3J(7,8) ˆ 9.8,
H
C(5,7)); 2.34 (s, Me C(4')); 2.29 (s, Me C(3',5')); 1.97 (s, Me C(2',6')). 13C-NMR (150 MHz, CDCl3):
153.72 (s, C(2)); 140.32 (s, C(3a,8a)); 136.36 (s, C(1')); 136.10 (d, C(6)); 135.40 (d, C(4,8)); 134.22 (s, C(4'));
132.28 (s, C(3',5')); 131.70 (s, C(2',6')); 123.09 (d, C(5,7)); 119.21 (d, C(1,3)); 18.69 (q, Me C(2',6')); 16.80 (q,
.
.
‡
‡
‡
Me C(4')); 16.56 (q, Me C(3',5')). EI-MS: 274.1 (97, M , C21H22 ), 259.1 (100, [M Me] ), 244.0 (92,
[M 2 Me] ), 229.0 (90, [M 3 Me] ).
‡
‡
2. Reaction of 1a with 1-Ethynyl-2,3,4,5,6-pentamethylbenzene (9a). 2.1. 1-Ethynyl-2,3,4,5,6-pentamethyl-
benzene (9a). In analogy to [18]: Phosphorous trichloride (11.5 ml, 131 mmol) was mixed with (2,3,4,5,6-
pentamethylphenyl)ethanone (6.0 g, 28.9 mmol; Avocado Comp.) in toluene (30 ml). Pentachlorophosphorane
(12.6 g, 60 mmol) was added and the mixture heated at 558 for two weeks. Nonreacted PCl5 and PCl3 were
removed in vacuo. The residue was washed with H2O and dried. CC (silica gel, hexane/Et2O 9 :1) gave colorless
crystals of 1-(1-chloroethenyl)-2,3,4,5,6-pentamethylbenzene (3.45 g, 63%). Colorless crystals. M.p. 73.5 ± 75.58
(hexane/Et2O). 1H-NMR (300 MHz, CDCl3): 5.94 (br. s, 1 H, CH2); 5.48 (br. s, 1 H, CH2); 2.57 (s, 2 Me); 2.49 (s,
Me C(4)); 2.47 (s, Me C(2,6)). 13C-NMR (75 MHz, CDCl3): 140.19 (s); 136.56 (s); 135.69 (s); 132.67 (s);
131.26 (s); 116.71 (d, CH2); 17.46 (q, 2 Me); 16.90 (q, Me C(4)); 16.49 (q, 2 Me). EI-MS: 207.9, 210.0 (100, 70,
.
.
‡
‡
‡
‡
‡
M
, C13H17Cl ), 193.1, 195.1 (44, 14 [M Me] ), 173.0 (98, [M Cl] ), 158.1 (81, [M Cl Me] ).
The pure (chloroethenyl)benzene (2.42 g, 11.6 mmol) was dissolved in dry THF (20 ml) and added
dropwise within 0.5 h to a stirred suspension of lithium diisopropylamide at 788, prepared at 788 from 2.5m
BuLi in hexane (12 ml, 30 mmol) and diisopropylamine (5 ml, 36 mmol) in THF (12 ml). The mixture was
allowed to warm to r.t., and stirring was continued for 6 h. The reaction was quenched by the addition of cold
H2O (10 ml), the org. layer formed was washed with dil. HCl soln. (10 ml; conc. HCl soln./H2O 1:10) and H2O,
and evaporated, and the crude product was purified by sublimation and recrystallized from hexane: pure 9a
(1.19 g, 60%). Colorless crystals. M.p. 90.2 ± 92.78 (hexane). IR (CHCl3): 3307s (H CꢀC); 2096w (CꢀC).
1H-NMR (300 MHz, CDCl3): 3.47 (s, H CꢀC); 2.49 (s, 2 Me); 2.27 (s, Me C(4)); 2.24 (s, Me C(2',6')).
13C-NMR (75 MHz, CDCl3): 136.02 (s); 135.65 (s); 132.19 (s); 119.75 (s); 83.75 (d, HCꢀC); 83.04 (s, HCꢀC);
.
.
‡
‡
‡
18.59 (q, 2 Me); 16.75 (q, Me C(4)); 16.42 (q, 2 Me). EI-MS: 172.1 (100, M , C13H16 ), 157.1 (94, [M Me] ),
142.1 (88, [M 2 Me] ), 229.0 (90, [M 3 Me] ).
‡
‡

1732
Helvetica Chimica Acta ± Vol. 84 (2001)
2.2. Thermal Reaction of 1a and 9a: Formation of 4,5,6,7,8-Pentamethyl-9-(2,3,4,5,6-pentamethylphenyl)-
6,8a-etheno-8aH-cyclohepta[b]furan-2(6 H)-one (11a): A Schlenk vessel was charged with 1a (0.030 g,
0.139 mmol) and 9a (0.096 g, 0.555 mmol) in decalin (10 ml). The vessel was flushed with Ar, closed, and
heated at 1908 during 15 h. The solvent was then evaporated. The residual red solid was subjected to sublimation
(1208/0.04 mbar) to remove nonreacted 9a (0.048 g, 50%). CC (hexane/E2O 9 :1) of residue gave, according to
1H-NMR, pure 11a (0.048 g, 89%). Rose crystals. M.p. 139.2 ± 143.48 (hexane/Et2O 9 :1). UV (hexane): 369 (sh,
2.82), 309 (sh, 3.55), 266 (3.99), 240 (sh, 4.00), 201 (4.71); min. 249 (3.92). IR (CHCl3): 1742vs (CˆO).
1H-NMR (600 MHz, CDCl3): 5.69 (s, H C(10)); 5.57 (s, H C(3)); 2.20 (s, Me C(4')); 2.19 (s, Me C(5')); 2.13
(s, Me C(3')); 2.09 (br. s, Me C(5)); 2.05 (s, Me C(6')); 1.88 (d-like, 5J(4,5) ˆ 1,2, Me C(4)); 1.87 (s,
Me C(2')); 1.84 (d-like, 5J(7,8) ˆ 1.0, Me C(7)); 1.77 (d-like, 5J(8,7) ˆ 1.2, Me C(8)); 1.56 (s, ME C(6)).
13C-NMR (150 MHz, CDCl3): 173.9 (s, C(2)); 162.62 (s, C(3a)); 149.45 (s, C(5)); 144.02 (s, C(9)); 137.29 (s,
C(7)); 135.65 (d, C(10)); 134.23 (s, C(4')); 133.35 (s, C(6')); 132.13 (s, C(1')); 132.03 (s, C(5')); 131.98 (s, C(3'));
131.62 (s, C(8)); 131.52 (s, C(2')); 119.90 (s, C(4)); 108.09 (d, C(3)); 90.96 (s, C(8a)); 49.46 (s, C(6)); 23.25 (q,
Me C(6)); 18.71 (q, Me C(6')); 18.10 (q, Me C(5)); 17.66 (q, Me C(2')); 16.70 (q, Me C(4)); 16.66 (q,
Me C(4')); 16.55 (q, Me C(3')); 16.41 (q, Me C(5')); 14.62 (q, Me C(7)); 11.37 (q, Me C(8)). EI-MS: 388.2
.
‡
‡
‡
(100, M ), 373.2 (83, [M Me] ), 202 (75, C13H14O2 ). Anal. calc. for C27H32O2 ´ 0.33 H2O (394.54): C 82.19, H
8.35; found: C 82.08, H 8.19.
3. Thermal Reaction of the New Azulenes with ADM. 3.1. Formation of Dimethyl 6,7,8,9,10-Penta-
methylheptalene-4,5-dicarboxylate (20). Azulene 5a (0.050 g, 0.252 mmol) was dissolved in toluene (2 ml) under
Ar in a flame-dired Schlenk vessel, and ADM (0.093 ml, 0.757 mmol) was added. The vessel was flushed again
with Ar and then closed. The dark blue soln. was heated at 120 ± 1308. It turned immediately to a deep orange
color. After 25 min of stirring at 120 ± 1308, the solvent and unreacted ADM were evaporated at 508. The
residual orange solid was subjected to flash CC (silica gel, hexane/Et2O 1:1). From the main yellow fraction,
pure 20 (0.014 g, 16%) was obtained. For NMR and UV/VIS characterization of its DBS isomer 20', a CDCl3
soln. of 20 was irradiated in an NMR tube with a high-pressure Hg lamp. After 1 h of irradiation, a photo-
stationary mixture 20/20' 54 :46 was obtained.
Data of 20: Orange crystals. M.p. 102.3 ± 105.98 (hexane). UV/VIS (hexane): tailing up to > 400, 289 (sh,
3.98), 254 (4.21), 234 (sh, 4.26), 219 (4.37); min. 247 (4.20). Ir (CHCl3): 1716vs (CˆO), 1266s (C O). 1H-NMR
(600 MHz, CDCl3): 7.51 (d, 3J(3,2) ˆ 6.2, H C(3)); 6.26 (dd, 3J(2,3) ˆ 6.3, 3J(2,1) ˆ 10.2, H C(2)); 6.20 (d,
3J(1,2) ˆ 10.2, H C(1)); 3.73 (s, MeOCO C(4)); 3.67 (s, MeOCO C(5)); 1.96 (s, Me C(8)); 1.85 (s,
Me C(9)); 1.80 (s, Me C(6), Me C(7)); 1.76 (s, Me C(10)). 13C-NMR (150 MHz, CDCl3): 168.14 (s,
OˆC C(5)); 167.53 (s, OˆC C(4)); 150.32 (s, C(5a)); 139.18 (s, C(8)); 138.56 (d, C(3)); 134.58 (s, C(7));
133.45 (s, C(10)); 133.39 (s, C(4)); 133.21 (d, C(1)); 132.16 (s, C(9)); 128.61 (s, C(6)); 128.29 (s, C(10a)); 126.85
(d, C(2)); 122.54 (s, C(5)); 52.05 (q, MeOCO C(4)); 51.86 (q, MeOCO C(5)); 20.17 (q, Me C(6)); 18.75 (q,
.
.
‡
‡
Me C(8)); 17.59 (q, Me (7)); 17.28 (q, Me C(9)); 15.66 (q, Me C(10)). EI-MS: 340.1 (61, M , C21H24O4 ),
‡
‡
‡
325.0 (30, [M Me] ), 308.0 (18, [M MeOH] ), 293.0 (57, [M MeOH Me] ), 286.0 (24, [M
‡
‡
‡
MeCꢀCMe] ), 281.1 (100, [M COOMe] ), 266.0 (83, [M COOMe Me] ), 254.0 (59, [M
‡
‡
MeCꢀCMe MeOH] ), 249.0 (83, [M COOMe MeOH] ), 239.0 (78, [M MeOH Me
‡
MeCꢀCMe] ).
Data of Dimethyl 6,7,8,9,10-Pentamethylheptalene-1,2-dicarboxylate (20'): UV/VIS (hexane; taken from the
photo-stationary mixture with the HPLC/UV detector; rel. intensities with respect to the strongest band
intensity ˆ 100%): 367 (4), 271 (69), 227 (sh, 83), 208 (100); min. 359 (3), 252 (63). 1H-NMR (600 MHz, CDCl3;
taken from the photo-stationary mixture): 6.69 (dd, 3J(4,5) ˆ 6.0, 3J(4,3) ˆ 11.5, H C(4)); 6.52 (d, 3J(3,4) ˆ 11.5,
H
C(3)); 5.79 (d, 3J(5,4) ˆ 6.0, H C(5)); 3.83 (s, MeOCO C(2)); 3.73 (s, MeOCO C(1)); 1.95 (s,
Me C(6)); 1.92 (s, Me C(8)); 1.86 (s, Me C(7)); 1.83 (s, Me C(9)); 1.70 (s, Me C(10)). 13C-NMR
(150 MHz, CDCl3; taken from the photo-stationary mixture): 168.60 (s, OˆC C(2)); 167.21 (s, OˆC C(1));
147.09 (s, C(5a)); 138.66 (s, C(8)); 137.02 (s, C(2)); 135.22 (d, C(4)); 134.76 (s, C(7)); 134.37 (s, C(10)); 132.70 (s,
C(9)); 132.05 (s, C(6)); 126.86 (s, C(10a)); 125.30 (d, C(3)); 124.90 (s, C(1)); 121.43 (d, C(5)); 52.56 (q,
MeOCO C(2)); 52.35 (q, MeOCO C(1)); 20.95 (q, Me C(6)); 19.08 (q, Me C(8)); 18.08 (q, Me C(7));
17.64 (q, Me C(9)); 15.26 (q, Me C(10)).
3.2. Formation of Dimethyl 6,8,10-Trimethyl-2-(2,3,4,5,6-pentamethylphenyl)heptalene-4,5-dicarboxylate
(22). Azulene 4ba (0.070 g, 0.221 mmol) was dissolved in toluene (1.5 ml) under Ar in a flame-dried Schlenk
vessel, and ADM (0.082 ml, 0.664 mmol) was added. The vessel was flushed again with Ar and then closed. The
dark blue soln. was heated at 120 ± 1308. After 20 h, a second portion of ADM (0.082 ml, 0.664 mmol) was added
and heating continued for additional 10 h. The solvent and unreacted ADM were evaporated at 508. The
residual orange solid was subjected to CC (silica gel, hexane/Et2O 1:1). The yellow main fraction (0.068 g) gave,

Helvetica Chimica Acta ± Vol. 84 (2001)
1733
after crystallization from hexane at 188, 22 (0.035 g, 34%). According to 1H-NMR (CDCl3), 22 reacted rapidly
at 208 to a thermal equilibrium mixture of 87% of 22 and 13% of its DBS isomer 22'. From a second blue
fraction of CC, the known dimethyl 4,6,8-trimethylazulene-1,2-dicarboxylate (23, 6.0 mg, 9%) was isolated.
The thermal reaction 4ba (0.070 g, 0.221 mmol) and ADM (0.082 ml, 0.664 mmol) in tetralin (ˆ1,2,3,4-
tetrahydronaphthalene; 1,5 ml) at 2058/6 h gave 22 (0.036 g, 36%) and 23 (4.0 mg, 6%), whereby 30% of non-
reacted 4ba (0.021 g) were recovered.
Data of 22: Orange crystals. M.p. 178.5 ± 181.78 (hexane). UV/VIS (hexane soln. of the thermal equilibrium
mixture): tailing from > 400, 334 (sh, 3.57), 267 (4.27), 222 (sh, 4.51); min. 251 (4.24). UV/VIS (hexane; taken
from the thermal equilibrium mixture by HPLC/UV; see above): tailing from > 400. 325 (sh, 12), 262 (41), 212
(100); min. 250 (38). IR (CHCl3): 1716vs (CˆO), 1270vs (C O), 1252vs (C O). 1H-NMR (600 MHz, CDCl3;
taken from the thermal equilibrium mixture)5): 7.44 (s, H C(3)); 6.20 (s, H C(9)); 6.00 (s, H C(7)); 5.79 (s,
H
C(1)); 3.73 (s, MeOCO C(5)); 3.70 (s, MeOCO C(4)); 2.31 (s, Me C(2')); 2.28 (s, Me C(4')); 2.27 (s,
4
Me C(3')); 2.22 (s, Me C(5')); 2.08 (s, Me C(6')); 2.06 (br. s, Me C(8)); 2.04 (d, J(6,7) ˆ 1.2, Me C(6));
1.83 (s, Me C(10)). 13C-NMR (75 MHz, CDCl3; taken from the thermal equilibrium mixture): 167.93 (s,
OˆC C(5)); 167.26 (s, OˆC C(4)); 148.18 (s, C(5a)); 143.40 (s, C(2)); 142.83 (d, C(3)); 139.93 (s, C(8));
136.88 (s, C(1')); 134.42 (s, C(4')); 133.06 (d, C(1); C(2')); 130.13 (superimposed d and s, C(9), C(6)); 129.32 (d,
C(7)); 123.08 (s, C(5)); 122.34 (s, C(10a)); signal of C(4) (s) not recognizable; 52.00 (q, MeOCO C(4)); 51.94
(q, MeOCO C(5)); 24.95 (q, Me C(8)); 23.21 (q, Me C(6)); 18.14 (q, Me C(2')); 17.90 (q, Me C(10)); 17.16
.
‡
(q, Me C(6')); 16.75 (q, Me C(4')); 16.58 (q,Me C(3')); 16.37 (q, Me C(5')). EI-MS: 458.1 (100, M
C30H34O4 ), 399.1 (36, [M COOMe] ).
,
.
‡
‡
The structure of 22 was confirmed by an X-ray crystal-structure analysis (cf. below).
Data of Dimethyl 6,8,10-Trimethyl-4-(2,3,4,5,6-pentamethylphenyl)-heptalene-1,2-dicarboxylate (22'): UV/
VIS (hexane; taken from the thermal equilibrium mixture by HPLC/UV; see above): tailing from > 400, 317
(sh, 15), 272 (48), 218 (sh, 80), 204 (100); min. 252 (38). 1H-NMR (600 MHz, CDCl3; taken from the thermal
equilibrium mixture)5): 6.28 (d, 4J(3,5) ˆ 0.8, H C(3)); 6.12 (s, H C(9)); 5.89 (s, H C(7)); 5.61 (d, 4J(5,3) ˆ
0.8, H C(5)); 3.83 (s, MeOCO C(2)); 3.72 (s, MeOCO C(1)); 2.23 (s, Me C(4')); 2.21 (s, Me C(3')); 2.16
(s, Me C(2'), Me C(5')); 2.10 (s, Me C(6)); 1.94 (s, Me C(8)); 1.90 (s, Me C(6')); 1.74 (s, Me C(10)).
3.3. Formation of Dimethyl 6,7,8,9,10-Pentamethyl-2-(2,3,4,5,6-pentamethylphenyl)heptalene-4,5-dicarbox-
ylate (24). As described in Exper. 3.1, 4a (0.070 g, 0.203 mmol) and ADM (0.125 ml, 1.02 mmol) in toluene
(2.0 ml) gave, after 50 min of heating at 1208, followed by CC and crystallization from hexane, 24 (0.038 g, 39%).
The 1H-NMR (500 MHz; CDCl3) of the crude 24 indicated the presence of a second product 27 (total amount
ca. 3%), which was also found in crystalline 24 as yellow crystals that could be separated mechanically from the
crystal mixture. For NMR and UV/VIS characterization of the DBS isomer 24' of 24, a dioxane soln. (1 ml) of
24 (0.020 g) was irradiated in a UV cuvette with a high-pressure Hg lamp. After 4 h of irradiation, a photo-
stationary mixture 24/24' 63 :37 was obtained. The DBS isomer 24' (5.0 mg, 25%) was separated by prep. TLC
(hexane/t-BuOMe 9 :1).
Data of 24: Orange crystals. M.p. 181.2 ± 183.58 (hexane). UV/VIS (hexane): tailing from > 400, 291 (sh,
4.07), 257 (4.33), 226 (4.47); min. 245 (4.31). IR (CHCl3): 1714vs (CˆO), 1268vs (C O), 1252vs (C O).
1H-NMR (500MHz, CDCl3)5): 7.39 (s, H C(3)); 5.92 (s, H C(1)); 3.70 (s, MeOCO C(4), MeOCO ((5));
2.31 (s, Me C(2')); 2.28 (s, Me C(3'), Me C(4')); 2.21 (s, Me C(5')); 2.03 (s, Me C(6')); 1.98 (s, Me C(8));
1.88 (s, Me C(9)); 1.87 (s, Me C(6)); 1.83 (s, Me C(7)); 1.80 (s, Me C(10)). 13C-NMR (125 MHz, CDCl3):
168.09 (s, OˆC C(5)); 167.43 (s, OˆC C(4)); 151.23 (s, C(5a)); 141.25 (d, C(3)); 141.16 (s, C(2) or C(4));
138.75 (s, C(8)); 137.27 (s, C(1')); 134.31 (s, C(4')); 134.18 (s, C(7)); 133.22 (s, C(10)); 132.69 (s, C(2') or C(3'));
132.50 (s, C(2) or C(4)); 132.48 (s, C(5')); 132.45 (s, C(6')); 132.27 (s, C(9)); 131.68 (d, C(1)); 130.53 (s, C(2') or
C(3')); 128.36 (s, C(6)); 127.89 (s, C(10a)); 121.95 (s, C(5)); 51.98, 51.86 (q, MeOCO C(4), MeOCO C(5));
20.17 (q, Me C(6)); 18.84 (q, Me C(8)); 18.18 (q, Me C(2')); 17.62 (q, Me C(7)); 17.34 (q, Me C(9)); 17.24
(q, Me C(6')); 16.76, 16.64 (q, Me C(3'), Me C(4')); 16.39 (q, Me C(5')); 15.86 (q, Me C(10)). EI-MS:
.
‡
‡
‡
486.3 (54, M ), 432.2 (35, [M MeCꢀCMe] ), 427.2 (100, [M COOMe] ), 395.2 (25, [M COOMe
‡
‡
‡
MeOH] ), 385.2 (14, [M MeOH Me MeCꢀCMe] ), 299.1 (22, [M Me5C6 Me] ). Anal. calc. for
C32H38O4 (486.64): C 78.98, H 7.87; found: C 79.04, H 7.78.
The structure of 24 was confirmed by an X-ray crystal-structure analysis (cf. below).
5
)
For the 2-(2,3,4,5,6-pentamethylphenyl)heptalene derivatives, the (P)-chirality was chosen arbitrarily for
1
the assignment of the H- and 13C-signals of the diastereoisomeric C-atoms and Me groups, whereby the
(pro-P) atoms or groups were given priority.

1734
Helvetica Chimica Acta ± Vol. 84 (2001)
Data of Dimethyl 6,7,8,9,10-Pentamethyl-4-(2,3,4,5,6-pentamethylphenyl)heptalene-1,2-dicarboxylate (24'):
Orange crystals. M.p. 183.0 ± 187.58 (hexane). UV/VIS (hexane): tailing up to > 400, 357 (sh, 3.14), 270 (sh,
4.17), 219 (sh, 4.51), 199 (4.80); min. not present down to 200. IR (CHCl3): 1721vs (CˆO), 1268s (C O), 1236s
(C O). 1H-NMR (600 MHz, CDCl3)5): 6.31 (d, 4J(3,5) ˆ 0.5, H C(3)); 5.45 (s, 4J(5,3) ˆ 0.5, H C(5)); 3.83 (s,
MeOCO C(2)); 3.77 (s, MeOCO C(1)); 2.23 (s, Me C(4')); 2.22 (s, Me C(3')); 2.18 (s, Me C(2')); 2.16
(MeOCO C(1)); 2.23 (s, Me C(4')); 2.22 (s, Me C(3')); 2.18 (s, Me C(2')); 2.16 (s, Me C(5')); 2.96 (s,
Me C(6)); 1.86 (s, Me C(8)); 1.82 (s, Me C(7)); 1.81 (s, Me C(9)), 1.78 (s, Me C(6')); 1.77 (s, Me C(10)).
13C-NMR (150 MHz, CDCl3): 168.57 (s, OˆC C(2)); 167.04 (s, OˆC C(1)); 150.03 (s, C(4)); 147.81 (s, C(5a));
139.80 (s, C(1')); 138.25 (s, C(8)); 137.06 (s, C(2)); 134.22 (s, C(10)); 133.99 (s, C(4')); 133.95 (s, C(7)); 132.41 (s,
C(9), C(3')); 132.28 (s, C(5')); 132.05 (s, C(6)); 131.46 (s, C(6')); 129.48 (s, C(2')); 127.84 (s, C(10a)); 125.51 (s,
C(1)); 124.89 (d, C(3)); 123.78 (s, C(5)); 52.44 (q, MeOCO C(2)); 52.29 (q, MeOCO C(2)); 20.48 (q,
Me C(6)); 18.77 (q, Me C(8)); 18.06 (q, Me C(2')); 17.69 (q, Me C(7)); 17.26 (q, Me C(9)); 16.69 (q,
Me C(4')); 16.65 (q, Me C(6')); 16.50 (q, Me C(3')); 16.36 (q, Me C(5')); 15.37 (q, Me C(10)).
Data of Dimethyl 5,6,7,8-Tetramethyl-2-(2,3,4,5,6-pentamethylphenyl)acenaphthalene-3,4-dicarboxylate (27):
Yellow crystals.M.p. 241.3±243.38 (hexane). UV/VIS (hexane): 359 (sh, 3.79), 338 (3.98), 327 (sh, 3.87), 255 (4.55),
223 (sh, 4.58), 202 (4.85); min. 298 (3.66), 241 (4.46). IR (CHCl3): 1726vs (CˆO), 1250s (C O). 1H-NMR
(600 MHz, CDCl3): 6.82 (s, H C(1)); 3.62 (s, MeOCO C(4)); 2.91 (s, MeOCO C(3)); 2.70 (s, Me C(5));
2.31 (s, Me C(6)); 2.25 (s, Me C(2'), Me C(6')); 2.18 (s, Me C(8)); 2.15 (s, Me C(4')); 2.14 (s, Me C(3'),
Me C(5')); 2.00 (s, Me C(7)). 13C-NMR (150 MHz, CDCl3): 170.26 (s, OˆC C(4)); 168.66 (s, OˆC C(4));
143.00 (s, C(2)); 137.81 (s, C(7)); 136.95 (s, C(8a)); 136.85 (s, C(8b)); 136.52 (s, C(8)); 136.48 (s, C(5) or C(4));
135.09 (s, C(6)); 134.79 (s, C(1')); 133.81 (s, C(4')); 133.29 (s, C(2'), C(6')); 133.09 (s, C(5) or C(4)); 132.34 (s,
C(3'), C(5')); 129.63 (s, C(5a)); 129.16 (s, C(2a)); 128.68 (s, C(3)); 128.36 (d, C(1)); 52.30 (q, MeOCO C(4));
51.51 (q, MeOCO C(3)); 22.41 (q, Me C(5)); 20.63 (q, Me C(6)); 19.04 (q, Me C(2'), Me C(6')); 17.92 (q,
.
‡
Me C(8)); 17.04 (q, Me C(7), Me C(4')); 16.88 (q, Me C(3'), Me C(5')). EI-MS: 470.2 (100, M
,
.
‡
‡
‡
‡
C31H34O4 ), 438.1 (25, [M MeO] ), 423.0 (84, [M Me MeO] ), 379.1 (88, [M COOMe MeOH] ).
The structure of 27 was confirmed by X-ray crystal-structure analysis (cf. below).
4. Thermal Equilibration Experiments of the New Heptalenes. 4.1. Thermal Equilibration of 20 and 20'. A
soln. of 20 (ca. 2 mg) in dioxane (0.2 ml) was heated 1 h at 1008. After heating, the probe was immediately
subjected to HPLC (Spherisorb CN; hexane/iPrOH 99.5 :0.5; flow rate 1 ml/min): tR 2.7 (20') and 4.7 min (20);
UV detection at the isosbestic point of the isomers (230 nm).
The photochemically prepared mixture 20/20' 54 :46 (ca. 1 mg) was dissolved in toluene (0.1 ml) and heated
during 1.5 h at 1008. HPLC Analyses (time [min], 20/20' [%]): 0, 54 :46; 30, 87:13; 60, 89 :11; 90, 90 :10; no
further changes in composition. t1/2(1008) ꢁ 22 min.
4.2. Thermal Equilibration of 22 and 22'. Heptalene 22 (ca. 5 mg) was dissolved in (D8)toluene (0.6 ml).
Equilibration occurred already at r.t. 1H-NMR Analyses at 1008: equilibrium ratio 22/22' of 74 :26.
4.3. Thermal Equilibration of 24 and 24'. A soln. of 24 (ca. 2 mg) in dioxane (0.2 ml) was heated 2 h at 1008.
After heating, the probe was immediately subjected to HPLC (Spherisorb CN; hexane/iPrOH 99.5 :0.5; flow
rate 1 ml/min): tR 2.5 (24') and 4.4 min (24). UV Detection at the isosbestic point of the isomers (286 nm) gave
an equilibrium ratio 24/24' of 84 :16. The photochemically prepared 24' (ca. 1 mg) was dissolved in toluene
(0.1 ml) and heated at 1008. At regular time intervals, samples were taken and immediately subjected to HPLC
analysis. Results (time [min], 24/24' [%]): 30, 44 :56; 60, 65 :35; 90, 76 :24; 150, 83 :17; 210, 84 :16; no further
changes in composition. t1/2 (1008) ca. 45 min.
5. X-Ray Crystal-Structure Determination of 22, 24, and 27. 5.1. Heptalene 22 and Acenaphthylene 276). All
measurements were made on a Rigaku AFC5R diffractometer with graphite-monochromated MoKa radiation
(cf. Table 6) and a 12 kW rotating anode generator. The structures of 22 (C30H34O4) and 27 (C31H34O4) were
solved and refined successfully with no unusual features. The three fused rings of 27 form a planar system, and
the plane of the pentamethylphenyl ring lies almost perpendicular to this plane with the interplanar angle being
83.90(9)8. Similarly, the pentamethylphenyl ring of 22 is almost perpendicularly oriented with respect to the p-
plane of the C(1)ˆC(2) bond (Tables 4 and 5).
6
)
Crystallographic date (excluding structure factors) for the structures reported here have been deposited
with the Cambridge Crystallographic Data Center as supplementary publication nos. CCDC-161722 and
161723 for 22 and 27, respectively. Copies of the data can be obtained, free of charge, on application to the
CCDC, 12 Union Road, Cambridge CB21EZ, UK (fax: ‡44-(0)1223-336033; e-mail: deposit@ccdc.ca-
m.ac.uk).

Helvetica Chimica Acta ± Vol. 84 (2001)
1735
Table 6. Crystallographic Data of Heptalene-4,5-dicarboxylates 22 and 24 and Acenaphthylene-3,4-dicarboxylate 27
22
27
24
Crystallized from
Empirical formula
Mr
hexane
C30H34O4
458.59
hexane/Et2O
C31H34O4
470.61
hexane
C32H38O4
486.62
Crystal color, habit
Crystal dimensions [mm]
Temperature [K]
yellow, prism
0.23 Â 0.38 Â 0.45
173(1)
yellow, prism
0.38 Â 0.43 Â 0.45
173(1)
yellow, prism
0.08 Â 0.14 Â 0.52
183(1)
Crystal system
Space group
Z
monoclinic
P21/n (#14)
4
triclinic
P1 (#2)
2
monoclinic
P21/n (#14)
4
Å
Reflections for cell determination
25
25
8000
2qRange for cell determination [8]
25 ± 38
38 ± 40
5 ± 56
Unit-cell parameters
a [Š]
b [Š]
c [Š]
a [8]
b [8]
g [8]
V [Š3]
F(000)
Dx [gcm
m(MoKa) [mm
Scan type
2q(max) [8]
13.969(4)
8.744(3)
20.803(2)
90
95.57(1)
90
2529(1)
984
1.204
0.0784
w/2q
8.905(3)
16.739(7)
8.908(3)
97.91(4)
106.61(3)
86.26(4)
1259.8(8)
504
9.007(6)
16.4382(15)
17.4171(12)
90
99.909(8)
90
2735.9(4)
1048
1.181
3
]
1.241
0.0805
w/2q
1
]
0.0760
f-oscill. scan
56
55
55
Total reflections measured
Symmetry-independent reflections
Rint
Reflections used (l > 2s(l))
Parameters refined
Reflection/parameter ratio
Final R
6439
5806
0.049
3561
308
6183
5810
0.024
4124
317
13.0
0.0523
0.0496
0.005
2.843
6(1) ´ 10
0.0002
0.28; 0.24
0.003
26021
6122
0.136
2442
338
7.2
0.1895
0.4967
0.1000
2.188
0.046(8)
0.007
11.6
0.0577
0.0541
0.005
2.126
2.2(4) ´ 10
0.0002
wR
Weights: p in w ˆ [s2(Fo) ‡ (pFo)2]
1
Goodness-of-fit
7
7
Secondary extinction coefficient
Final Dmax/s
D1 (max; min) [e Š
s(d(C C)) [Š]
3
]
0.28; 0.28
0.003 ± 0.004
0.74; 0.64
0.013 ± 0.015
5.2. Heptalene 247). Intensity data were collected with an imaging plate detector system (Stoe IPDS) [19]
with graphite-monochromated MoKa radiation (cf. Table 6). A total of 167 images were exposed at a constant
time of 5 min/image. Direct methods with SHELXS-97 [20] were used to solve the structure. All non-H-atoms
were refined anisotropically. H-Atom positions were calculated after each refinement cycle. The bad R values
for 24 are due to a crystal of minor quality with many twinned domains with different orientations. Crystals of
better quality could not be obtained by further crystallization experiments. Despite this fact, the twisted
structure of 24 could be unambiguously determined (cf. Fig. 8 and Tables 4 and 5).
7
)
Further information on crystallographic data of the structure of 24 are available from Dr. H. W. Schmalle,
Anorganisch-chemisches Institut der Universität, CH-8057 Zürich, Winterthurerstrasse 190.

1736
Helvetica Chimica Acta ± Vol. 84 (2001)
REFERENCES
[1] M. Nagel, H.-J. Hansen, Helv. Chim. Acta 2000, 83, 1022.
[2] W. S. Trahanovsky, S. L. Emeis, A. S. Lee, J. Org. Chem. 1976, 41, 4044.
[3] V. Lellek, M. Nagel, H.-J. Hansen, Helv. Chim. Acta, in preparation.
[4] M. Nagel, Diploma Thesis, University of Zürich, 1997.
[5] Y. Saseda, Bull. Chem. Soc. Jpn. 1959, 32, 165.
[6] A. Linden, M. Meyer, P. Mohler, A. J. Rippert, H.-J. Hansen, Helv. Chim. Acta 1999, 82, 2274.
[7] G. Melone, Ph. D. Thesis, University of Zürich, 1999.
[8] J. A. Berson, Acc. Chem. Res. 1978, 11, 446.
[9] W. Bernhard, H.-R. Zumbrunnen, H.-J. Hansen, Chimia 1979, 33, 324.
[10] Y. Chen, R. W. Kunz, P. Uebelhart, R. H. Weber, Helv. Chim. Acta 1992, 75, 2447.
[11] K. Hafner, G. L. Knaup, H. J. Lindner, Bull. Chem. Soc. Jpn. 1988, 61, 155.
[12] A. A. S. Briquet, P. Uebelhart, H.-J. Hansen, Helv. Chim. Acta 1996, 79, 2282.
[13] W. Bernhard, P. Brügger, J. J. Daly, P. Schönholzer, R. H. Weber, H.-J. Hansen, Helv. Chim. Acta 1985, 68,
415.
[14] K. Hafner, G. L. Knaup, H. J. Lindner, H.-C. Flöter, Angew. Chem. 1985, 97, 209; Angew. Chem., Int. Ed.
1985, 24, 212.
[15] L. A. Paquette, Pure Appl. Chem. 1982, 54, 987; L. A. Paquette, T.-Z. J. Luo, Ch. E. Cottrell, A. E. Clough,
L. B. Anderson, J. Am. Chem. Soc. 1990, 112, 239; L. A. Paquette, M. P. Trova, J. Luo, A. E. Clough, L. B.
Anderson, J. Am. Chem. Soc. 1990, 112, 228.
[16] D. Laikov, H.-J. Hansen, unpublished results.
[17] G. M. Brooke, R. Matthews, S. Raymond, M. E. Harman, M. B. Hursthouse, J. Fluorine Chem. 1991, 53,
339.
[18] K. L. Evans, P. Prince, E. T. Huang, K. R. Boss, R. D. Gandour, Tetrahedron Lett. 1990, 31, 6753.
[19] Stoe & Cie., Version 2.92, Darmstadt, Germany.
[20] G. M. Sheldrick, Acta Crystallogr. Sect. A 1990, 46, 467; G. M. Sheldrick, ꢀProgram for the Refinement of
Crystal Structuresꢁ, University of Göttingen, 1997.
Received April 7, 2001