вход по аккаунту


Benzene Oxide-Oxepin Valence Tautomerism.

код для вставкиСкачать
MAY 1 9 6 7
P A G E S 385-476
Benzene Oxide-Oxepin Valence Tautomerism
Dedicated to Prof. R. Criegee on the occasion of his 65th birthday
Benzene oxide and the potential 8x-electron system oxepin exist in valence-tautomeric
equilibrium with each other, to which both components contribute to approximately the
same extent. N M R spectroscopic measurements show that the equilibrium is rapidly
established (activation energies of the forward and reverse reactions 9.1 and 7.2 kcal
mole-1, respectively). The present knowledge of the properties of oxepinjustifies its classification as a “heterotropilidene”. Benzene oxide-oxepin represents a system having
fluctuating bonds, the equilibrium of’ which can be displaced from one extreme fo the
other by means of suitable substituents. The oxide component determines the reactions
of the system with most agents.
With 1,6-oxido[ IO]annulene, which is formally a 2,7-bridged oxepin, the oxepin character
is completely suppressed by the formation of a delocalized Ion-electron system extending over the Clo perimeter. The existence and aromatic character of l,doxido[l0]annulene give rise to the conception of a homologous series of oxygen bridged annulenes
(I,6; 8,13-bisoxido[I4]annulene, 1,6;4 1 7 ; 10,15-trisoxido[l8]annulene etc.), which,
like the parent acenes, possess a (4 n + 2)x-electron system. Molecular models
demonstrate that a considerable flattening of the C4,+2 perimeter is achievable in the
case of a syn or all-syn arrangement of the oxygen bridges, and thaf the requirement for
aromaticity is thus satisfied. This is confirmed in a striking manner by the synthesis
and properties of syn-l,6; 8,13-bisoxido[l4]annuiene.
I. Introduction
The discovery of the tropolones and the tropylium
ion 111 resulted in the rapid development of the chemistry of carbocyclic aromatic seven-membered ring
systems and an increased interest in the as yet unknown heterotropilidenes oxepin ( I ) , azepine (2), and
thiepin (3), which are formally homologues of furan,
pyrrole, and thiophene. ( I ) , (2), and (3) are potential
8 x-electron systems, isoelectronic with the unstable
cycloheptatrienyl anion 121 and cyclooctatetraene; thus
they do not satisfy the Hiickel rule, and therefore
should not exhibit an electronic stability comparable
with that of the tropylium compounds. Consequently,
heterotropilidenes may provide an interesting test of
the modern theories regarding the nature of aromaticity 131. A knowledge of the heterotropilidenes above all
promises valuable information on similarities and
[*I Prof. Dr. E.Vogel and Dr. H. Giinther
Institut fur Organische Chemie der Universitat
Ziilpicher Str. 47
5 Koln (Germany)
[l] T. Nozoe in D . Ginsburg: Non-Benzenoid Aromatic Compounds. Interscience, New York 1959, p. 339; W. v. E. Doering:
Theoretical Organic Chemistry. Butterworths, London 1959. p.
35; D. M. G. Lloyd: Carbocyclic Non-Benzenoid Aromatic
Compounds. Elsevier, Amsterdam 1966.
Angew. Chem. internat. Edit. / VoI. 6 (1967) No.5
121 H . .
j r . and M . R . Rifi, J. Amer. chem. SOC.85,3041
(1963); W.V.E. Doering and P. P. Gaspar, ibid. 85, 3043 (1963);
R. Breslow and H. W . Chang, ibid. 84, 1484 (1962).
[3] E. Hiickel, Z. Physik 70, 204 (1931); Grundziige derTheorie
ungesattigter und aromatischer Verbindungen. Verlag Chemie,
Berlin 1938; A. Streitwieser: Molecular Orbital Theory for
Organic Chemists. J. Wiley, New York 1961, p. 256.
differencesin the properties of 4 n- and (4n
tron systems.
+ 2)-~elec-
Despite their simple structure, all attempts to synthesize oxepin and thiepin have so far been unsuccessful 141;however, N-alkoxycarbonylazepines[~l
and N-cyanoazepines [61 have become readily accessible by thermolysis of azidoformates and cyanazide in benzene and
its derivatives. No detailed information is yet available
concerning the parent compound 1H-azepine (2),
which, it is claimed, results from the alkaline hydrolysis of N-ethoxycarbonylazepine.It is evidently a very
reactive substance, which rearranges rapidly into the
tautomeric 3H-azepine [7,81. N-Methylazepine, which
is prepared from N-ethoxycarbonylazepine[7,91, and
whose spectral properties should be very close to those
of (2), is also rather unstable; on standing for a short
time it changes into a dimer [91.
The successful preparation of the N-methylated
azepine suggested that, of the heterotropilidenes ( I )
to (3), at least oxepin should be isolable. Like its
analogues, this compound would be able to isomerize
into the benzenoid system (formation of phenol) ;
however, there is no possibility of tautomerization
(such as exists with (2)) or of spontaneous elimination
of the hetero atom (as is conceivable for (3), via
benzene sulfide).
Most of the earlier attempts to synthesize oxepin were
patterned o n Willstutter’s cycloheptatriene synthesis, i.e.
attempts were made t o introduce the missing double bonds
into tetrahydro- or dihydrooxepins[lo]. Critical examination
of the results of these efforts suggests that n o oxepin was
formed in most cases. However, if the desired product had
been formed in one of these reactions, it could have had little
chance of survival because of the drastic conditions used.
In view of the difficulty of obtaining oxepin via
partially hydrogenated derivatives, an attractive alternative method of preparation would seem to be the
thermal rearrangement of strained valence isomers of
oxepin 1111. The obvious choice of isomer is the hypothetical benzene oxide (4), which was assumed, more
intuitively than from thermodynamic considerations,
[4] 3-Benzoxepin has become known through investigations by
K. Dimroth [K. Dimroth, G . Pohl, and H . Follmann, Chem. Ber.
99, 634 (1966); K . Dimroth and G. Pohl, Angew. Chem. 73, 436
(1961)l; earlier papers deal with derivatives of 3-benzoxepin and
its N and S analogues.
[ 5 ] K . Hafner and C. Konig, Angew. Chem. 75,89 (1963); K. Ha&
ner, D. Zinser, and K.-L. Moritz, Tetrahedron Letters 26, 1733
(1964); R. J. Cotter and W. F. Beach, J. org. Chemistry 29, 751
(1964); W. Lwowski, T. J. Maricich, and T. W . Mattingly jr., J.
Amer. chem. SOC.85, 1200 (1963).
[6] F. D . Marsh and H . E. Simmons, J. Amer. chem. SOC.87,
3529 (1965).
171 K. Hafner, Angew. Chem. 75, 1041 (1963); Angew. Chem.
internat. Edit. 3, 165 (1964).
[8] A tricarbonyliron complex of 1 H-azepine was recently
obtained by E. 0. Fischer and H . Riihle, Z . anorg. allg. Chem.
341, 137 (1965).
[9] K. Hafner and J . Mondt, Angew. Chem. 78,822 (1966); Angew. Chem. internat. Edit. 5, 839 (1966).
[lo] S. Olsen and R. Bredoch, Chem. Ber. 91, 1589 (1958); E. E.
Schweizer and W. E. Parham, J. Amer. chem. SOC.82,4085(1960);
J . Meinwald, D . W . Dicker, and N. Danieli, ibid. 82, 4087 (1960);
W . Kinzebach, Diploma Thesis, Universitat Marburg, 1962.
[ l l ] Review on valence isomerinations in strained ring systems:
E. Vogel, Angew. Chem. 74, 829 (1962); see also S. J . Rhoads in
P. deMayo: Molecular Rearrangements. Wiley. New York 1963,
p. 655.
to be less stable than oxepin. On the basis of this
estimated stability relationship a ready benzene
oxide + oxepin isomerization could be envisaged,
particularly since the disrotatory course of the reaction
and the geometry of the valence-isomeric
system are matched to each other. In the case of the
four-membered ring valence isomer 2-oxabicyclo[3.2.0]hepta-3,6-diene (5), on the other hand, the
result of thermolysis is more difficult to predict. The
cleavage of the cyclobutene ring in ( 5 ) does not
conform stereochemically (conrotatory process) with
the rearrangement to the oxepin [12,131, so that the isomerization temperatures required would almost certainly prove too high for the product. In the final case
to be considered, i.e. the oxide of Dewar-benzene (6),
the conversion into oxepin would again appear
promising, since the extremely high ring strain should
allow homolytic cleavage of the bond common to the
two four-membered rings, even at low temperatures.
The sterically unfavorable synchronous isomerization
of the cyclobutene ring would thus be bypassed.
II. Thermolysis of Epoxide C -C Bonds
The dehydrohalogenation of 1,2-epoxy-4,5-dibromocyclohexane (17) [I41 provides a simple scheme for the
synthesis of benzene oxide. However, according to
investigations by Meinwald and Nozaki 1151, only
phenol can be isolated when (17) is heated with ycollidine or sodium glycolate in an excess of ethylene
glycol. Dehydrohalogenation experiments on (17)
carried out simultaneously by our own research
group[161 with a series of nitrogen bases were also
The fact that dehydrohalogenation of (17) to benzene
oxide under very mild conditions, as well as the
benzene oxide-oxepin isomerization (which involved
cleavage of an epoxide C-C bond), had to be regarded
as unusual reactions at that time, may explain the
reluctance to pursue further the above route to oxepin.
However, following recent reports of the cleavage of the
C-C bond in epoxides, it became almost certain
that at least the second reaction step was feasible.
Braun’s study [I71 of the LiC1-catalysed pyrolysis of
1,2-divinylethylene carbonate (7) (mixture of cis and
1121 R. B. Woodward and R. Hoffmann, J. Amer. chem. SOC.87,
395 (1965).
[13] For the stereochemistry of the valence isomerization of
cyclobutenes, see R. Criegee, D . Seebach, R. E. Winter;
B. Borretzen, and H.-A. Brune, Chem. Ber. 98, 2339 (1965).
[14] E. E.vanTamelen, J. Amer. chem. SOC.77, 1704 (1955).
[15] J. Meinwald and H. Nozaki, J. Amer. chem. SOC.80, 3132
[16] J. Ferry, Diploma Thesis, Technische Hochschule Karlsruhe, 1959.
[17] R. A. Braun, J. org. Chemistry 28, 1383 (1963).
Angew. Chem. internat. Edit. 1 Vol. 6 (1967) 1 No. 5
trans isomers) is particularly relevant to the benzene
oxide-oxepin problem. This reaction, which requires
a temperature of 200 "C, gives not only trans-l,2-divinylethylene oxide (8), but also 4,5-dihydrooxepin
(9), the formation of which is probably due to a Cope
rearrangement of initially formed cis-l,2-divinylethylene oxide ( I I ) , facilitated by the strain in the
epoxide ring. In order to avoid the pyrolysis of the
carbonate, which is prohibitive to the isolation of ( I I ) ,
Stogryn, Gianni, and Passannante
studied the action
of alkali metal hydroxide on 3-chloro-4-acetoxy-l,5hexadiene (10) (presumably a mixture of the two
diastereoisomers). In this case, epoxide formation
occurred at 40 to 50 OC, and it was possible to isolate
cis-l ,Zdivinylethylene oxide. As expected, the cis-l,2divinylethylene oxide undergoes the Cope rearrangement (11) -+ (12) below 100°C, with quantitative
formation of 4,5-dihydrooxepin (9).
We investigated the thermal behavior of the relatively
stable trans-l,2-divinylethyleneoxide 1191 and found
that, at 17O-20O0C, it also rearranged completely
with cleavage of the epoxide C-C bond. As in the
case of the cis-isomer, 4,5-dihydrooxepin was obtained;
however, this was accompanied by a second product,
which had escaped the earlier investigators, and which
was identified as 2-vinyl-2,3-dihydrofuran(14). Between 170 and 200°C, the ratio of the quantities of
seven-membered to five-membered ring ethers (30: 70)
is practically independent of the temperature and of
the pyrolysis time. It is deduced from the relatively
high isomerization temperatures and from the observed branching of the reaction that the rearrangement of trans-l,2-divinylethyleneoxide (8) proceeds
via the biradical intermediate (13).
1181 E. L . Stogryn, M . H . Gianni, and A . J . Passannante, J. org.
Chemistry 29, 1275 (1964).
1191 E. Vogel, R. Sundermann, and R. Schubart, unpublished.
Angew. Chem. internat. Edit. 1 Vol. 6 (1967) 1 No. 5
The mechanistic interpretation of the pyrolysis of cisand trans-l,2-divinylethyleneoxide, which is based
largely on analogy with the thermal behavior of cis,transisomeric 1,2-divinylcycloalkanes containing strained
rings [11,20-241, was supported by a kinetic study. The
isomerizations of the two compounds in the gas phase
strictly obeyed the first-order rate law. The reaction of
the cis oxide was not appreciably accelerated by an
increase in surface area, whereas slight wall catalysis
was observed in the case of the trans oxide. The
activation parameters found for the rearrangement of
the cis compound were AH* = 24.6 kcal mole-1 and
AS* = -11.3 cal deg-1 mole-1, while those for the
trans compound were AH* = 36.0 kcal mole-1 and
AS+ = -0.4 cal deg-1 mole-' [193.
Thus cis-1,2-divinylethylene oxide has a low activation
energy and ahighly negativeactivation entropy, whereas
the trans compound has a relatively high activation
energy and only a slightly negative activation entropy. If
in addition one considers the nature of the rearrangement products, the essential criteria for a synchronous
mechanism in the isomerization of the cis oxide and
for a free-radical mechanism in the case of the trans
oxide are satisfied. The difference of 11 kcal mole-' in
the activation energies for the concerted rearrangement
of ( 1 1 ) and the free-radical cleavage of (8) is of the
order of magnitude found for analogous cases [23-261.
Linn, Webster, and Benson 1271 observed another example of
the cleavage of a n epoxide a t the C-C bond, namely in
tetracyanoethylene oxide (15). This compound can undergo
cycloadditions with olefins and acetylenes to form tetracyanotetrahydrofurans (16) and tetracyanodihydrofurans.
Tetracyanoethylene oxide can add to aromatic systems with
formation of tetracyanotetrahydroisobenzofurans.
A kinetic investigation of the reaction of tetracyanoethylene oxide with styrene and trans-stilbenes [281
showed that the addition step is preceded by a thermal
1201 E. VogeI, K.-H. Ott, and K . Gaiek, Liebigs Ann. Chem. 644,
172 (1961).
[21] W.V. E. Doering and W . R. Roth, Angew. Chem. 75, 27
(1963); Angew. Chem. internat. Edit. 2, 115 (1963).
[22] W . V . E. Doering and W . R. Roth, Tetrahedron 19, 715
[23] G. S. Hammond and C. D . DeBoer, J. Amer. chem. SOC.86,
899 (1964); E. Schmidt, Diploma Thesis, Universitat Koln, 1964.
I241 W . R . Roth, personal communication.
1251 According to W . v. E. Doering and J. C. Gilbert, Tetrahedron
Suppl. No. 7, 397 (1966), footnotes 36 and 31, the synchronous
process in the case of 1,5-hexadiene leads to a lowering of the
activation energy by 11 kcal mole-1 [cleavage of 1,5-hexadiene
into free ally1 radicals: AH+ = 46.5 kcal mole-1; Cope rearrangement of l,l-dideuterio-1,5-hexadiene: AH* = 35.5 kcal mole-11.
1261 R. Sundermann found the following activation parameters
for the isomerization of trans-l,2-divinylcyclopropaneinto 1.4cycloheptadiene; A H + = 32 kcal mole-1 and AS* = -3.6
cal deg-1 mole-1 (cf. [Zl]).
1271 W . J. Linn, 0. W . Webster, and R. E. Benson, J. Amer. chem.
SOC.85,2032 (1963); W . J. Linn and R. E. Benson, ibid. 87, 3657
[281 W . J . Linn, J. Amer. chem. SOC.87, 3665 (1965).
equilibrium between tetracyanoethylene oxide and an
activated form thereof. When sufficiently reactive
olefins, such as styrene, are used, the formation of
excited tetracyanoethylene oxide becomes rate-determining. The addition taking place in the second step
has the characteristics of a multi-center process.
III. Benzene Oxide-Oxepin
The course of the thermolysis of cis- and trans-1,2divinylethylene oxide gave new support to the concept
of synthesizing oxepin via benzene oxide. The real
problem was to find a suitable route to benzene
oxide. Owing to the ready availability of the starting
material, a reinvestigation of the dehydrohalogenation
of 1,2-epoxy-4,5-dibromocyclohexane
( I 7) [15,161 suggested itself. We found that the action of sodium
methoxide on ( 1 7 ) in boiling ether gave an orange
product C6H60, (yield > 80 %) which was isomeric
with phenol [29,301. The question was whether this was
benzene oxide, oxepin, or a valence-tautomeric mixture of benzene oxide and oxepin. As in the case of
tropilidene [Ill, it turned out to be problematical to
decide between these three possibilities by chemical
The substance C6H60, which is stable at room temperature, readily isomerizes under the influence of
Bronsted (proton) and Lewis-acids to form phenol;
aromatization takes place, though only slowly, even
in neutral aqueous solution. Distillation at normal
pressure also leads to more or less complete rearrangement to phenol. Hydrogenation with palladium-charcoal produces mainly oxepane (18) (about 70%)
together with cyclohexanol and unidentified products.
Reaction with maleic anhydride, even at 20 "C, leads
within a few minutes to an adduct (19), which must,
according to its NMR spectrum and the absence of a
UV-maximum above 220 mp, be derived from benzene
oxide. Dimethyl acetylenedicarboxylate gives an
adduct (20) of the same type after a few days (20 "C).
Attempts to fragment (20) by the Alder-Rickert
method into acetylene oxide (21) and dimethyl
phthalate have not so far yielded conclusive results.
The dehydrohalogenation product C6H60 can be
reduced with lithium aluminum hydride to form the
vacuum-distillable cyclohexa-l,3-dien-5-01 (22) ("hydrated benzene"), which undergoes wall-catalysed
decomposition into benzene and water on standing.
Methyl-lithium gives a mixture of the relatively stable
cis- and trans-l-methylcyclohexa-2,4-dien6-01s in
which the cis:trans ratio is higher than 90:lO.
While the result of hydrogenation is equally compatible
with either the benzene oxide or the oxepin formula,
the relatively easy diene synthesis, the isomerization
into phenol, and the reduction with lithium aluminum
hydride can best be explained by the benzene oxide
formula. On the other hand, the recently observed,
almost quantitative photocheniical conversion of the
product C6H60 into 2-oxabicyclo[3.2.0]hepta-3,6diene ( 5 ) is indicative of the presence of oxepin [311. It
seems unlikely that oxepin is formedfrombenzeneoxide
on irradiation, since a photochemical benzene oxideoxepin isomerization should be forbidden according
to the Woodward-Hoffmann rule [121. The photoisomer
(S), the structure of which is proved mainly by the
formation of 2-oxabicyclo[3.2.0]heptane (23) on
hydrogenation [32J, exhibits the anticipated thermal
stability. At the temperature required to isomerize ( 5 )
back into oxepin (150-160 "C) the latter is no longer
stable, and thus only phenol is obtained as the pyrolysis
Though the above chemical reactions point very
strongly to the existence of a valence-tautomeric
mixture of benzene oxide and oxepin, they do not
constitute conclusive proof, since a change in structure
caused by the agent in question cannot be ruled out.
Consequently, spectroscopic methods must be used
in an attempt to clarify the benzene oxide-oxepin
The UV spectrum of the product (Fig. 1) shows an
unusually strong dependence on the solvent. The
absorption curves show an isosbestic point, a sure
indication of the presence of two compounds, benzene
oxide and oxepin, in equilibrium with each other. In
order to assign the maxima and to obtain some
information regarding the ratio of the components,
[29] E . Vogel, R . Schubarr and W . A . BOII, Angew. Chem. 76,535
(1964); Angew. Chem. internat. Edit. 3,510(1964); Chem. Engng.
News 42, No. 40, 40 (1964).
1301 E. Vogel, W . A . Boll, and H. Giinther, Tetrahedron Letters
1965, 609.
[31] A similar photoisomerization was observed for 2,7-dimethyloxepin by L. A . Paquette and J . H . Barrett, J. Amer. chem. SOC.
88, 1718 (1966).
I321 L. A. Paquette, J . H . Barrett, R . P. Spitz, and R . Pitcher,
3. Amer. chem. SOC. 87, 3417 (1965).
Angew. Chem. internat. Edit./ Vol. 6 (1967) 1No. 5
we tried to derive the UV spectra of benzene oxide
and oxepin with the aid of suitable reference subStances.
h I m p 1 -=-
Fig. 1. (a) UV-Spectrum of benzene oxide-oxepin (4)
I. in isooctane; 2. in methanol; 3. in water-methanol (85:15).
(b) UV-spectra in isooctane
I. benzene oxide-oxepin (4)
3. 2,7-dimethyloxepin (25).
+ ( I ) ; 2. 8,9-indan oxide
(24) ;
The model used for benzene oxide was 8,9-indan oxide
(24), in which the rearrangement into the oxepin
isomer is blocked by the bracket effect of the trimethylene bridge (cf. Section IV. 5). The same principle had previously been used to fix the norcaradiene
skeleton [33,341. 8,9-Indan oxide (24) is colorless, and
in isooctane possesses a UV spectrum of the cyclohexa-l,3-diene type with Amax = 258 mp (E = 4900).
A suitable reference substance for oxepin is its 2,7dimethyl derivative (25), in which the concentration
of the oxide tautomer (if present) is below the detection
limit of spectroscopic methods. The electronic spectrum of 2,7-dimethyloxepin contains a broad absorption band at Amax (in isooctane) = 297 mp (E = 1800),
which extends to about 430 mp and gives the compound a yellow-orange color. Significantly, the U V
spectra of both (24) and (25) are practically
independent of the solvent.
These observations show that in the UV spectrum of
the dehydrohalogenation product, the maximum at
271 mp (E = 1430) in isooctane is due to the benzene
oxide component, whereas the shoulder at 305 mp
(E = 900) corresponds to the oxepin maximum which
must be located in this region. The position of the
equilibrium cannot be accurately determined without
knowledge of the extinction coefficients of the two
valence tautomers, but it is estimated from comparison
with the spectra of (24) and (25) that the benzene
oxide concentration in isooctane is about 30 %, while
that in water-methanol (85:lS) is about 90%. Concerning the solvent-dependence of the benzene oxideoxepin equilibrium, which is presumably due to
various factors, it is noteworthy that the proportion
of benzene oxide increases with increasing dielectric
constant of the solvent [3OJ.
The existence of a benzene oxide-oxepin valence
tautomerism with comDarable concentrations of the
components is clearly shown by NMR-spectroscopic
investigations, which may also be used to deduce
the thermodynamic and kinetic parameters of the
equilibri um.
At room temperature, the IH-NMR spectrum of the
product (Fig. 2) contains three groups of signals in the
olefinic proton absorption region, the chemical shifts
LOO of which differ sufficiently to allow an assignment to
the three pairs of protons of benzene oxide or oxepin
purely on the basis of multiplicity. Thus the group of
signals at T = 4.8, which may be regarded to a first
approximation as a doublet, must be assigned to the
a-protons, while the multiplets at T = 3.9 and 7 = 4.4
can be assigned to the y-and P-protons, respectively.
3 75
L 50
5 00
Fig. 2. IH-NMR spectrum of benzene oxide-oxepin at room
temperature (in CSz; 60 MHz; internal standard: tetramethylsilane).
The greatest difference in the spectra of benzene oxide
and oxepin should be found for the position of the aproton resonances. Taking cyclic enol ethers as a
guide for oxepin, the a-protons of oxepin should
absorb close to T = 4.0. The resonance signals of the
benzene oxide a-protons, on the other hand, would
be expected to occur in the region of T = 6.5 to 7.0,
on the basis of experience with vinyl-substituted
epoxides (Table 1).
Table 1. T values of the a-protons of cyclic enol ethers and vinylsubstituted epoxides.
3.77 1351
3.78 1351
3.84 1361
3.9 1371
3.69 [381
3.9 [371
frans-Divinylethylene oxide
cis-Divinylethylene oxide
Cyclooctatetraene oxide
6.9 1371
6.6 [371
6.7 I371
6.5 [391
I331 E. VogeI, W . Wiedemann, H . Kiefer, and W. F. Harrison,
Tetrahedron Letters 1963, 673.
[341 J . Schreiber, W. Leimgruber, M. Pesaro, P. Schudel, T. Threl/aN, and A . Eschenmoser, Helv. chim. Acta 44, 5 4 0 (1961); R .
Darms, T.ThreIfalI, M . Pesaro, and A . Eschenmoser, ibid. 46,
[35] L. M . Jackman: Applications of NMR-Spectroscopy in
Organic Chemistry. Pergamon Press, London 1959, p. 62.
[36] S. Masamune and N. T. Castellucci, J. Amer. chem. SOC.84,
2452 (1962).
[37] E. Vogel, R . Schubart and H. Gunther, unpublished.
[38] E. E. Schweizer and W . E. Parham, J. Amer. chem. SOC.82,
4085 (1960).
[39] P. Rnd/ick and S . Winstein, J . Amer. chem. SOC.86, 1866
2893 (1963).
Angew. Chem. internat. Edit.
1 VoI. 6 (1967) / No. 5
The observation that the cc-proton resonance va of the
product lies between the expected vaIues for benzene
oxide and oxepin suggests a rapid valence tautomerism
with comparable concentrations of the two forms,
since for a rapid exchange ( k > lo3 sec-1) of corresponding benzene oxide and oxepin protons,
according to eq. (a), va may be regarded as the average
of the chemical shifts vY4) and vyl) ( P ( ~ )p(rl
= mole
fractions) 1401; the same is true of v@and vy.
The assumption of an equilibrium of this type is
confirmed by the temperature-dependence of the NMR
spectrum 1411 (Fig. 3).
When the temperature is lowered below -80 "C, the exchange
rate ultimately becomes so slow that the "average spectrum"
is replaced by the individual spectra of the two equilibrium
components (Fig. 3f). This process is shown particularly
clearly by the a-proton signal, which reaches its maximum
halfwidth at about -113'C and then splits into two new
absorptions which, as expected, are strongly separated, and
which must be due to the a-protons in benzene oxide and
In order to deduce the kinetics of the equilibrium (4)
from the changes in the line shape of the spectrum, the
assignment of the low temperature spectrum is essential.
This can be made o n the basis of the ratio of the signal areas
and of the following considerations:
1. The resonance of the oc-protons in benzene oxide must
occur at the highest field strength [421.
2. The shift difference between the p- and y-protons in
benzene oxide must be small 1431.
3. The shift difference between the p- and y-protons in oxepin
must be large[431.
The equilibrium constant K of the valence tautomerism
( I ) was determined below the coalescence temperature
(-113 " C ) from the ratio of the signal areas, and above this
temperature from the resonance position of the a-proton
signal. Its temperature-dependence gave a value of AH0
(=Ho(1)--HO(q)) = 1.7
0.4 kcal mole-1 for the enthalpy
of the isomerization (4) + ( I ) .
The principal changes observed down to -80 "C are shifts of
the resonance frequencies. The a-proton resonance undergoes
a steady upfield shift, while the @-proton resonance is
displaced in the opposite direction; the y-proton resonance
remains approximately constant. The same phenomena are
observed in the room temperature spectra o n changing the
solvent from carbon tetrachloride to a polar medium such as
methanol or dimethyl sulfoxide-water (cf. the solventdependence of the UV spectra, Fig. 1). The resonance shifts
are therefore due to a displacement of the equilibrium. As can
be deduced from the increasing shielding of the a-protons,
the benzene oxide component is favored by a decrease in
temperature, as well as by a change to a polar solvent.
The lifetimes ~ ( 4 )and ~ ( 1 of
) the valence tautomers, which
must be known in order to find the activation parameters,
are obtained from the halfwidths A' of the a-proton signal
in the regions of slow and fast exchange (below -1 18 OC and
above -110OC respectively). A' was corrected for field
inhomogeneity and coupling with the @-protons,whereas the
long-range coupling with the y-protons was not taken into
account. For the limiting cases mentioned, the following
relationships exist between the lifetime T and the exchange
broadening A, of a resonance signal [40.441:
= 7cAa
slow exchange: 1 / (4)
fast exchange: l/z( 4 )
= kl = 4np ( 4 ) p2
The rate constants kl ( = 1 / ~
( 4 ) ) and k-1 ( = 1 / ~
(1)) as found
from equations (b) and (c) for ten different temperatures
satisfy the Arrhenius equations:
* exp [-(9100 f 800)/RT](sec-1)
1012.1* 1.4 exp
f 1000)/RT](sec-1)
The ratio of the frequency factors gives the entropy difference
AS0 (=SO (1l-So ( 4 ) ) = 10.5 f 8.3 cal deg-1 mole-1. Using
the equilibrium constant K for the calculation of AS0 one
obtains 11.0 & 5.0 cal deg-1 mole-1.
2 * 2
TFig. 3. 'H-NMR spectra of benzene oxide-oxepin at various
temperatures in CFJBr/pentane (2: 1) (60 MHz; internal standard:
[401 J . A . Pople, W. G. Schneider, and H . J. Bernstein: High
Resolution Nuclear Magnetic Resonance. McGraw-Hill Book
Co., Inc., New York 1959, p. 218.
[41] ff. Giinther, Tetrahedron Letters 1965, 4085.
The favorable position of the equilibrium ( K < 20)
permits the observation and evaluation of the benzene
oxide-oxepin valence tautomerism by NMR spectroscopy. According to these investigations the energy
of benzene oxide is ca. 1.7 kcal mole-1 lower than
that of oxepin. However, at room temperature the
entropy gain associated with this rearrangement causes
a considerable displacement of the equilibrium in
favor of oxepin (AG m -1.3 kcal mole-'). Owing to
the low activation energies of the forward and reverse
reactions (9.1 and 7.2 kcal mole-1, respectively), which
[42] This follows from the data o f Table 1.
[43] This condition follows from the changes in the resonance
frequencies in the average spectrum with temperature.
[44] L. H . Piette and W. A . Anderson, J. chem. Physics 30, 899
Angew. Chem. internat. Edit. / VoI. 6 (1967) No. 5
permit the iscimers to be isolated at best by crystallization, benzene oxide-oxepin may be regarded as a system
with fluctuating bonds (at 0 "C the valence isomerization of the benzene oxide-oxepin system proceeds approximately lo4 times faster than that of bullvalene) 1451.
How are the stability relationships of the benzene
oxide-oxepin system to be interpreted? An enthalpy of
formation of 29 kcal mole-1 has been calculated for
oxepin (three cis-CH=CH groups and one 0 group)
by the method of Franklinc461. A corresponding
calculation for benzene oxide (two cis-CH=CH, two
CH, and one 0 group) gives values of 21 and 35 kcal
mole-1, respectively, depending upon whether the
value of 13 1471 or 27 kcal mole-1 1481 is taken for the
epoxide ring strain. The difference between the enthalpies of formation of oxepin and benzene oxide is
therefore calculated as 8 or -6 kcal mole-1, whereas
the experimentally determined value is 1.7 kcal mole-1.
The conclusion to be drawn 1491 from these considerations is that the stabilization energy of oxepin (with
an epoxide ring strain of 13 kcal mole-1) is at most
of the order of magnitude of that of tropilidene (7 kcal
mole-') [501.
extreme to the other. Substituents in the u-position
would be expected, a priori, to favor the oxepin form
electronically, and possibly sterically ; the effects of
substituents in the p- and y-positions are more difficult
to predict. A suitable model for the expected substituent-dependence of the equilibrium is the valencetautomeric system bicyclo [4.2.0]octa-2,4-diene + cycloocta-1,3,5-triene 1511, in which the components are
also present in comparable quantities.
1. 2-Methyloxepin
2-Methyloxepin (27), an orange liquid (b.p. 33.5 to
34"C/11 mm Hg), has been obtained both from the
dibromide (26a) and from the allyl bromide (26b) by
dehydrobromination (with sodium methoxide and with
potassium t-butoxide, respectively). It reacts with
maleic anhydride to give a 1 :1 adduct (m.p. 197 to
198"C), the NMR spectrum of which indicates that
it is derived from 1-methylbenzene oxide (26).
A clearer indication of the expected non-aromatic
character of oxepin is given by spectroscopic investigation of 2,7-dimethyl- and 2,7-pentamethyleneoxepin (cf. Section IV.2). These investigations indicate
that oxepin possesses a nonplanar ring skeleton, which
should correspond closely to that of tropilidene. It is
thus justified to speak of oxepin as a heterotropilidene.
IV. Substituted Oxepins
The synthesis of oxepin described above is versatile,
and has already been used for the preparation of a
variety of substituted oxepins. The limiting factor of the
method appears to be that the 1,4-cyclohexadienesused
as starting materials are sometimes difficult to convert
into 1,2-epoxy-4,5-dibromocyclohexanes,the imrnediate precursors of the arene oxide-oxepin systems. If
the 1,Ccyclohexadienes are first converted into 1,2epoxycyclohex-4-enes (instead of into dibromocyclohexenes), it can be advantageous to introduce the
missing double bond by monohalogenation in the
allyl position (with N-bromosuccinimide or t-butyl
hypochlorite) followed by dehydrohalogenation, and
not by bromination-dehydrobromination.
According to NMR- and UV-spectroscopic studies,
2-methyloxepin, unlike 2,7-dimethyloxepin, still exists
in "rapid" equilibrium with a considerable quantity
of the corresponding arene oxide, i.e. 1-methylbenzene
oxide [493.
The NMR spectrum, like that of oxepin, is temperaturedependent (Fig. 4). A decrease in temperature leads first to
broadening and then to splitting of the signals (-113 " C ) ; on
further cooling (-119OC), the individual spectra of (26) and
(27) appear. The differences with respect to the spectrum at
Owing to the sensitivity of the benzene oxide-oxepin
equilibrium, it takes only small substituent effects to
displace the position of the equilibrium from one
[45] G . Schroder, J. F. M . Oth, and R. Merdnyi, Angew. Chem.
77, 774 (1965); Angew. Chem. internat. Edit. 4, 752 (1965).
[46] J . L. Franklin, Ind. Engng. Chem. 41,1070 (1949).
1471 R . A . Nelson and R. S. Jessup, J. Res. nat. Bur. Standards
48, 2307 (1952).
[481 A . S. PeN and G. Pitcher, Trans. Faraday SOC.61,71 (1965).
[49] H . Gunther, R . Schubart, and E. Vogel, Z . Naturforsch. 22b,
25 (1967).
[SO] R . B. Turner, W . R . Meador, and R. E. Winkler, J. Amer.
chem. SOC.79, 4116 (1957); R . B.Turner: Theoretical Organic
Chemistry. Butterworths, London 1959, p. 67.
Angew. Chem. internat. Edit. 1 Vol. 6 (1967)
1 No. 5
A 5 1 2 L]
Fig. 4. 'H-NMR spectrum of 2-methyloxepin (60 MHz; internal
standard: tetramethylsilane) a) in C S 2 at 37 "C, b) in CF3Br at -1 19 "C.
[Sll A . C. Cope, A . C . Haven jr.. F. L. Ramp, and E. R. Trumbull,
J. Amer. chem. SOC.74, 4867 (1952); R. Huisgen: Organic Reaction Mechanisms, Special Publ., Chem. SOC.(London) No. 19,
3 (1965).
3 7 O C are particularly noticeable in the case of the methyl
proton absorption. In the low-temperature spectrum the
doublet (J = 0.8 Hz) at T = 8.20 initially observed for these
protons is split into two signals (T = 8.09 and 8.45), the
positions of which correspond to those of the methyl protons
in (27) and (26). The intensities show that, under these conditions, the two compounds are present in a molar ratio
of 7:3. The new signal at T = 6.26 must be assigned to the
cc-proton in (26), wh,ereas the doublet at T = 4.22 must be due
to the corresponding proton in (27). This is compatible with
the coupling of about 5 Hz found for the latter proton. The
intensity ratio of these protons again shows that the oxepin
form predominates.
The equilibrium constant K of the valence tautomerism
(26) $ (27) below the coalescence temperature (-113OC)
was derived from the intensities of the methyl proton signals,
and above this temperature from the chemical shift of the
methyl protons. The reaction enthalpy A H 0 of the l-methylbenzene oxide + 2-methyloxepin isomerization as calculated
from the temperature-dependence of K is 0.4 -I 0.2 kcal
mole-1, i.e. 1.3 kcal mole-1 lower than that of the corresponding reaction in the parent system. A S 0 is calculated to
be 5.0 cal deg-1 mole-1.
The activation parameters for the equilibrium were deduced
in this case from the changes in the methyl proton resonances
with temperature. The evaluation was carried out with the
aid of the complete equation of the absorption curve[40~521,
the coupling of the methyl protons with the adjacent olefinic
protons in the oxepin form also being taken into account.
The theoretical spectra were calculated by means of a
Fortran I1 program. Fitting calculated and experimental
spectra by variation of the exchange rate permitted the
determination of the lifetimes T of the isomers. In the regions
of slow and fast exchange, the approximations (b) and (c)
could also be used for the determination of the T values
(lifetimes) .
tion of the oxide form in the equilibrium is below the
spectroscopic detection limit ( t 5 %). However, 2,7dimethyloxepin readily forms a nialeic anhydride
adduct derived from the oxide form. The strong
displacement of the equilibrium to the oxepin side on
or,or'-dimethyl substitution is undoubtedly not only the
result of electronic interaction of the methyl groups
with the double bond system in (25), but is probably
also due to the less favorable conformational situation
in the oxide form (eclipsed cis-methyl groups).
The uniformity of 2,7-dimethyloxepin is shown in
particular by the NMR spectrum (Fig. 5a). The positions of the methyl proton (T = 8.19) and ring proton
resonances (H455: T = 4.19, H336: T = 4.75) are in
agreement with the triene structure (25), as are the
coupling parameters obtained by analysis of the
AA'XX-system of the ring protons. Practically the
same couplings are found for the corresponding
protons in 1,6-dirnethylcycloheptatriene[54,551. Comparison of the spectra of (25) and 1,6-dimethylcycloheptatriene (Fig. 5b) clearIy shows the relationship
between the two compounds. The NMR spectrum of
(25) shows no sign of line broadening, even at -110 "C.
The Arrhenius equations obtained by these methods for the
(27) are:
valence tautomerism (26)
kl =
0.9 exp
[-(9200 i 700)/RT] (sec-1)
exp 1-(8700 & 700)/RT](sec-1).
In a comparison of the thermodynamic and kinetic
quantities for the equilibrium (26) F' (27) with those
for the parent system, the decrease in the AH0 value
of the arene oxide-oxepin isomerization is particularly
striking. Thus the introduction of the methyl group
lowers the energy of the oxepin form in relation to that
of the oxide form, either by hyperconjugation or by
o-bond stabilization (531. The activation energy for the
conversion of (26) into (27), on the other hand, is
practically the same as for the benzene oxide-oxepin
2. 2,7-Dimethyloxepin
In 2,7-dimethyloxepin (25) [291, whichis obtainedin high
yield on dehydrobromination of 4,5-dibromo-l,2-epoxy-l,2-dirnethylcyclohexane(28) with sodium methoxide, and which contains two or-methyl groups, the
oxepin form is so strongly stabilized that the propor[52] H. S . Gutowskv, D. W . McCalI, and C. P. S k h t e r , J. chem.
Physics 21, 279 (1953); H. S. Gutowsky and C. H . Holm, ibid. 25,
1228 (1956).
1531 A . Streitwieser: Molecular Orbital Theory for Organic
Chemists. J. Wiley, New York 1961, p. 247.
Fig. 5. 1H-NMR spectra of a) 2,7-dimethyloxepin (in CS2), b) 1,6-dimethylcycloheptatriene (in CC14) (60 MHz; internal standard: tetramethylsilane).
At the same time, the strong solvent-dependence of the
UV spectrum observed for ( I ) and (27) is absent for
(2.5). Thus 2,7-dimethyloxepin appears to be best
suited for physical studies on the oxepin system 1561.
[54] H . D . Roth, Dissertation, Universitat Koln, 1965.
[55] H. Giinther and H.-H. Hinrichs, Tetrahedron Letters 1966,
[56] The I R spectrum of 2,7-dimethyloxepin contains an enol
ether band at 1155 cm-1 and intense C=C stretching bands at
1655 and 1584 cm-1.
Angew. Chem. internat. Edit.
Vol. 6 (1967) I No. 5
Oxepin should have a conformation (30) similar to that of
cycloheptatriene, i.e. the molecule should be in a boat form,
presumably in equilibrium with its mirror-image conformer.
Evidence for such a nonplanar shape is provide by the
coupling parameters of the olefinic protons of (25), in
conjunction with the UV spectrum of the 2,7-pentamethylenebridged oxepin (58). The coupling constant J34 (= J56) should
show a similar dependence on the dihedral angle to that
observed for the analogous coupling in cycloheptatriene 1571.
Thus it should be possible to deduce qualitatively the geometry of the triene part of (25) from the magnitude of J34.
The coupling constant J34 (5.5 Hz) is practically equal to the
corresponding constants for cycloheptatriene and its 1,6dimethyl derivative (5.5 and 5.4 Hz, respectively). It may
therefore be taken as certain that the c 4 - C ~ double bond
does not lie in the plane of C-atoms 2,3,6, and 7. Owing to
the good agreement between the UV spectra of (25) and of
2,7-pentamethyleneoxepin (58) (cf. Section IV.5), in which a
deflection of the oxygen in relation to the triene part is
forced by steric factors, it is likely that the oxygen atom also
lies out of this plane.
4. 4,5-Bis(methoxycarbonyl)oxepin
The y,y'-disubstituted oxepin (36) can be obtained
from the oxide (34) of dimethyl 3,6-dihydrophthalate
by allylic bromination with N-bromosuccinimide
(NBS) and dehalogenation of the crude bromination
product (35) with sodium iodide in acetone[601. The
same substance was very recently synthesized in an
original manner by Prinzbach, Arguelles, and Druckrey 1611 from the Diels-Alder adduct (37) of furan and
dimethyl acetylenedicarboxylate. Irradiation of the
adduct yields the tetracyclic isomer (38), which gives
(36) on pyrolysis with cleavage of the indicated bonds
in the three-membered rings.
3. 2-Acetyloxepin
An cr-substituent that can enter into conjugation, e.g.
an acetyl group, should be more effective than a n
K:-methyl group in stabilizing the oxepin form. The
red-brown 2-acetyloxepin (33) [b.p. 51 "C/0.4 mm) [591
obtained by dehydrobromination of l-acetyl-l,2epoxy-4,5-dibromocyclohexane (31) with 1,5-diazabicyclo[4.3.0]non - 5 - ene [581 (other bases always gave polymeric products) is uniform. This is
shown by the N M R spectrum (complex multiplet for
the olefinic protons between 7 = 3.3 and 4.5, in which
a doublet of triplets due to the proton H7 is observed
at T = 4.06) and by the fact that the spectrum shows
practically n o change with temperature. The same
conclusion can be drawn from the UV spectrum [Amax
(cyclohexane) = 222 m p (E = 10900), 315 m p (2500),
and 360 m p (1200)l in conjunction with its weak
Though the oxide component (32) cannot be detected,
it appears to determine the chemical reactivity, as
in the case of 2,7-dimethyloxepin. 2-Acetyloxepin
gives the adduct derived from the oxide form with
maleic anhydride, and is aromatized by catalytic
quantities of mineral acid, with formation of 2-hydroxyacetophenone.
[571 J. B. Lambert, L. J . Durham, P. Lepoutere, and J. D . Roberts,
J . Amer. chem. SOC.87, 3896 (1965); H. Giinther and R. Wenzl,
2. Naturforsch. 226, 389 (1967).
1581 H. Oediger, H.-J. Kabbe, F. Mofler, and K . Eiter, Chem.
Ber. 99, 2012 (1966). We are grateful to Dr. K . Eiter, Farbenfabriken Bayer, for supplying this base.
1591 F.-G. Klurner, Diploma Thesis, Universitat Koln, 1965.
Angew. Chem. internat. Edit.
VoI. 6 (1967)
i No. 5
The N M R spectrum of (36) consists o f a n A A X X ' system
of the olefinic protons and a singlet due to the ester methyl
protons. It is strongly temperature-dependent, so that a
considerable proportion of the oxide tautomer (39) may be
assumed to be present in the equilibrium. The changes in the
T- and N-values of the AA'XX' system when the temperature
is lowered are shown in Table 2. The increasing shielding of
Table 2. ' H - N M R
spectroscopic data for the system (39) ;.' (36).
the a-protons (Hx and Hx,) and the simuftaneous deshielding of the @-protons(HA and HK) (owing to the conjugation
position of these protons in relation t o the ester group in the
arene oxide (39), their resonance frequency should occur
at lower field strengths than in (36)) show that the decrease
in temperature results in a displacement of the equilibrium
in favor of the oxide.
The same conclusion can be drawn from the changes in the
parameter N, which represents the sum of the coupling
constants JAX and JAX. J A is~ negligibly small in both
isomers, so that in the present case N depends essentially o n
the coupling constant J A X . Since (39) undoubtedly has the
lower coupling constant, a decrease in N must correspond to
a n increase in the oxide concentration.
The UV spectrum of (36) is strongly solvent-dependent,
again indicating a valence tautomerism with comparable
[601 R. Srhuhart, Dissertation, Universitat Koln, 1967.
[Sl] H . Prinzbach, M . Arguelles, and E . Druckrey, Angew. Chem
78, 1057 (1966); Angew. Chem. internat. Edit. 5, 1039 (1966).
concentrations of the components. As in the parent system,
a solvent change from cyclohexane to methanol causes a
displacement of the equilibrium in favor of the arene oxide
An N M R spectrum similar to that of (36) (even in its
temperature dependence) is found for 4,5-dimethyloxepin (42), which can be prepared from 1,f-epoxy4,5-dimethyl-4,5-dibromocyclohexane(40), but which
has so far been obtained only in solution [601.
The room-temperature spectrum of the compound (42)
shows an AAXX' system for the oIefinic protons with TA =
4.2 and TX = 5.68 and a methyl proton signal at T = 8.16,
The difference TX - TA increases noticeably with decreasing
temperature, while the position of the methyl proton signal,
as expected, changes only slightly. The chemical shifts at
-65 OC are TA = 4.0 and TX = 6.2, showing that under these
conditions the system consists almost entirely of the oxide
In both 4,5-disubstituted oxepins therefore, as in the
parent system, the arene oxide is the low energy component.
5. Fixed Arene Oxides
Taking into account the different equilibrium positions in the
parent systems norcaradiene-cycloheptatriene and benzene
oxide-oxepin, it is to be expected from the above findings that
the bridged arene oxide-oxepin systems will still contain
measurable concentrations of the tricyclic isomers in the
equilibrium at n = 4, and possibly even at n = 5.
The bridged arene oxide-oxepin systems (52) + (53)
were synthesized by dehydrohalogenation of the dibromides (51) with potassium t-butoxide in ether 1631.
(51) with n = 3 gives the uniform, colorless arene
oxide (24) (8,g-indan oxide). The structure of the
n = 3, 4, 5
= 3, 4, 5
n = 3, 4, 5
compound is shown by the similarity of its N M R
spectrum to that of (45). The spectrum shows a n
AA'BB' system of the olefinic protons at T = 3.4 to 4.1
(JAM= 6.0, JAB= 9.3, JAB'= 0.7 and JBB' = 1.2 Hz) [641
and a multiplet due to the methylene protons at
T = 7.5 to 8.8 (intensity ratio 4:6). The values of the
coupling constants are compatible with the assumed
diene structure, but not with the triene structure (concerning the UV spectra, cf. Section 111).
By bridging the oxepin in the 2,7-position with a
methylene chain having a sufficiently small number
of chain members n, it is possible, because of steric
reasons, to achieve complete displacement of the
arene oxide-oxepin equilibrium in favor of the arene
The influence of the number n of chain-CHz groups on the
position of the equilibrium can be qualitatively estimated
from the data for the corresponding bridged norcaradienecycloheptatriene systems (43) + (44) [33,54,621. The strain
imposed on the cycloheptatriene form (46) by a bridge
having n = 3 is so great that the valence-tautomeric norcaradiene form (45) is frozen in. Surprisingly, however, a
The product obtained from (51) with n = 4 was also
found to be uniform, but the NMR spectrum in this
case does not permit a definite assignment, since the
olefinic proton absorption consists simply of a sharp
signal at T = 3.85. A decision in favor of the arene
oxide structure (55) (9,lO-tetralin oxide) is possible
from the UV spectrum, since the maximum at 259 m p
(E = 3900) corresponds exactly in its position (though
the extinction is lower) to that of 8,g-indan oxide (24).
complete reversal of the equilibrium is observed when n is as
low as 4. Though the cycloheptatriene form (48) is still
appreciably strained, as is clearly shown by its spectral data
and by its reactivity toward maleic anhydride, the norcaradiene isomer (47) can no longer be spectroscopically detected.
A chain of five methylene groups gives a cycloheptatriene
form (50) that is almost strain-free, though its conformation
isIf%ed by the bridge. (50) is in fact uniform, and is very
similar in i ts~spectral~properties-to-l,6-dimethylcycloheptaThe reaction product obtained from (51) with n
triene 1541.
= 5
differs from (24) and (55) even in its external appearance, in that it has the characteristic orange color
of oxepins. Its spectral properties indicate the presence
of an arene oxide-oxepin mixture with comparable
concentrations of the components. Both the UV
spectrum and the N M R spectrum are strongly solventdependent.
In cyclohexane, two UV maxima are observed at 276 and
300 my, both with E = 1400; corresponding bands are
observed in methanol at 270 and 300 mp, but in this case
1631 M . Wiesel, Dissertation, Universitat Koln, 1966.
1641 H . Giinther
E. Vogel and J. Eimer, unpublished.
and H.-H. Hinrichs, Liebigs Ann. Chem., in
Angew. Chem. internat. Edit. I Vol. 6 (1967)
1 No. 5
2100 and 1100. The oxide component, as in the parent
system, is evidently favored in methanol. The changes in the
NMR spectrum with the solvent are less pronounced, owing
to the absence of a-protons, but point in the same direction.
The olefinic protons absorb as an AA'XX system, the N
value (JAX + JAX,) of which is 7.2 Hz in cyclohexane and
8.1 Hz in methanol. Since a considerable amount of experimental evidence indicates that cyclic trienes have lower N
values than the corresponding diene valence isomers 1551, the
increase in this quantity suggests an increase in the proportion of the oxide. The NMR spectrum shows no appreciable
change when the temperature is lowered to -70 OC. This is
not surprising, since the most pronounced changes are again
associated with the presence of a-protons.
E =
V. Chemistry of Arene Oxide-Oxepin Systems
Little investigation has been carried out on arene
oxide-oxepin systems beyond the reactions mentioned
in connection with the parent system. All the systems
known are sensitive to Bronsted and Lewis acids, the
action of which normally leads to aromatization. Unlike cycloheptatriene, these systems readily undergo
Diels-Alder reactions, which yield the adduct derived
from the arene oxide component. Another apparently
general reaction is that with triphenylphosphine (at
120 to 150°C), which leads to the hydrocarbon corresponding to the arene oxide. Removal of oxygen has
also been observed under the action of carbonylchromium complexes 1651, but in these cases the hydrocarbons were partly obtained as metal complexes. It
is not yet known to what extent the addition of
nucleophilic reagents can be achieved. Such additions
are of interest, since as was shown by the formation
of cyclohexa-l,3-dien-5-01 (22), they would offer a
possible route to 1,2-dihydrobenzene derivatives that
are otherwise difficult to prepare.
In the study of the chemistry of arene oxide-oxepin
systems, the important question is: which reactions
can be ascribed to arene oxides and which to oxepins?
Now that the benzene oxide-oxepin equilibrium has
been established, the Diels-Alder reaction can be
definitely ascribed to the oxide component. If, as is
likely, arene oxides are also the reacting species in
acid-induced isomerizations, nucleophilic additions,
and other reactions, the chemical behavior of arene
oxide-oxepin equilibrium systems should be similar to
that of fixed arene oxides. It therefore seemed advisable
that the properties of arene oxides such as 9,10-tetralin
oxide (55) and 8,9-indan oxide (24) should be investigated in greater detail 1661.
(60) ; formation o f 6-tetra101 was not detected. Since
the spiro ketone (59) undergoes the dienone-phenol
rearrangement 1681 into 5-tetra101 under comparable
conditions, it may be assumed to occur as an intermediate in the aromatization of (55).
8,9-Indan oxide (24) behaves differently toward Bronsted and Lewis acids. Rearrangement into a spiro
ketone, i.e. (63), is associated in this case with a n
energetically unfavorable contraction of the fivemembered ring to give a four-membered ring, so that
side reactions that are not encountered with (55) can
prevail. Thus the aromatization product obtained in
aqueous media is found to be pure 5-indanol (62), and
not the 4-indanol (64) that would be expected if the
reaction proceeded via the spiro ketone (63). The
possibility that (64) is in fact formed via (63), but
undergoes a phenol-phenol rearrangement 1691 into
(62), can be ruled out on the basis of control experiments. The OH group of (62) appears instead to
be derived from the solvent by the route (24)
since a n attempted isomerization in methanol gave the
5-methoxyindan corresponding to (62), again in a
uniform reaction. It should be possible to determine
definitely the source of the OH group in (62) by
labeling with 1 8 0 . In the absence of prototropic solvents, the reaction sequence carbonium ion rearrangement - dienone-phenol rearrangement [(24) +
(63) + (64)] can also be observed with 8,9-indan
oxide (24). The dienone (63) can be isolated as a
dimer when the oxide is treated with catalytic quantities of a proton acid in ether. The tendency of (63)
to dimerize is evidently even stronger than that of
(59), a property that protects it against the dienonephenol rearrangement into (64) and allows its detection.
At room temperature in the presence of catalytic
quantities of a proton acid (aqueous-ethereal medium),
9,10-tetralin oxide (55) undergoes a n almost quantitative carbonium ion rearrangement into spiro[5.4]deca-7,9-dien-6-one (59) 1671. At higher acid concentrations and when Lewis acids are used, complete
aromatization takes place with formation of 5-tetra101
On the other hand, when Lewis acids such as zinc
chloride are used instead of proton acids, the catalyst
becomes fully effective, so that both the carbonium
ion rearrangement and the dienone-phenol rearrangement are observed as in the corresponding experiment
with (55).
I651 W. Grimme and B. Haas, unpublished.
[66] Cf. the 9,lO-phenanthrene oxide recently described by M.S.
Newman and S . Blum, J. Amer. chem. SOC.86, 5598 (1964).
I671 K . Alder, F. H . Flock, and H . Lessenich, Chem. Ber. 90,
1709 (1957).
I681 E . N . Marvel1 and E. Magoon, J. Amer. chem. SOC.77,2542
(1955); L. Fieser and M . Fieser: Steroide. Verlag Chemie, Weinheim 1961, p. 360.
[69] W. H. Hopff and A. S. Dreiding, Angew. Chem. 77, 717
(1965); Angew. Chem. internat. Edit. 4, 690 (1965).
Angew. Chem. internat. Edit.
1 Vol. 6 (1967) No. 5
2,7-Dimethyloxepin (25), which may be used for
comparison purposes owing to its similarity of substitution, behaves toward dilute acids in much the
same way as the two arene oxides (55) and (24). The
isomerization products obtained from (25) are 6,6dimethylcyclohexadienone (65) and 2,3-, 3,4-, and 2,6dimethylphenol [(66), (67), and (68)l.
6,6-Dimethylcyclohexadienone1681 is much more resistant to acids than the spiro ketones (59) and (63).
This explains why 2,3-dimethylphenol, which is
related to this compound via the dienone-phenol
rearrangement, is present only to a very small extent.
The formation of 3,4-dimethylphenol corresponds to
that of (62) from (24), and as in that case, is best
rationalized by assuming that the OH group is derived from the solvent. Finally, 2,6-dimethylphenol,
which [like 6,6-dimethylcyclohexadienone (65)] is
probably formed from the mesomeric carbonium ion
(69) derived from (29) by cleavage o f the epoxide
ring, can have no analogues in the case of (55) and
(24), since this would involve meta-bridging of the
aromatic nucleus by a chain of only 4 or 3 methyIene
The similarity in the chemical properties of 2,7-dimethyloxepin (25) and of the above arene oxides also
extends to the reactions with triphenylphosphine and
Cr(C0)3(NH3)3 [65,701, in which both types of compounds lose oxygen and undergo aromatization.
Finally, (25) resembles the arene oxides in that the
oxygen can also be removed by catalytic hydrogenation (with platinum in ether at 0 "C) 1711; like the triphenylphosphine reaction, this hydrogenation leads to
the corresponding aromatic hydrocarbons.
[70]For the preparation of Cr(C0)3(NH&, see W. Hieber,
W.Abeck, and H.K.Platzer, Z . anorg. allg. Chem. 280,252 (1955).
1711 With palladium-charcoal as the catalyst, on the other hand,
(25) affords 2,7-dimethyloxepane.
Though the comparative investigations of arene oxideoxepin systems and fixed arene oxides have by no
means been exhaustive, the results reported clearly
indicate that the arene oxides are the more reactive
components of the valence-tautomeric systems toward
most reagents. The only chemical indication of the
presence of the oxepins is the formation of 2-oxabicyclo[3.2.0]hepta-3,6-dienes[31,6oJ on photolysis.
VI. l,dOxido[ lO]annuIene
The synthesis and the aromatic natdre of the cyclic
conjugated 10 x-electron system 1,6-methano[10]annulene (70) 1721 directed interest toward the analogous 1,6-oxido[lO]annulene (71), which can also be
regarded as a 2,7-bridged oxepin. That (71) might be
unstable in comparison with the valence-isomeric 9,lOnaphthalene oxide (72) [cf. the position of the equilibrium (55) + (56)] seemed unlikely in view of the
properties of (70), which indicate that the 10 x-electron
system possesses considerable resonance energy. (71)
was synthesized simultaneously by Sondheimer 1731 and
by Vogel et a l . [ 7 4 1 by dehydrobromination of
the tetrabromo epoxide (76) with sodium methoxide
in ether, (76) being obtained from 1,4,5,8-tetrahydro4a,8a-oxidonaphthalene (75) 1751 which in turn is
readily available from (73) via (74). Though a total
of four moles of hydrogen bromide had to be
eliminated in this step, (71) was formed in yields of
up to 90%.
1721 E. Vogel and H . D. Roth, Angew. Chem. 76, 145 (1964);
Angew. Chem. internat. Edit. 3, 228 (1964); E. Vogel and
W . A. Boll, Angew. Chem. 76, 784 (1964); Angew. Chem.
internat. Edit. 3, 642 (1964); E.Voge1, Special Publ., Chem.
SOC.(London), No. 21, in press.
[73] F.Sondheimerand A.Shani, J.Amer.chern.SOC.86,3168 (1964).
[74] E. Vogel, M . Biskup, W. Prefzer, and W . A . Boll, Angew.
Chem. 76, 785 (1964); Angew. Chem. internat. Edit. 3, 642
(1964); M . Biskup, Dissertation, Universitat Koln, 1966.
[75] W. Huckel and H . Schlee, Chem. Ber. 88, 346 (1955).
Angew. Chem. interitat. Edit. / Vol. 6 (1967) 1 No. 5
1,6-Oxido[lO]annulene is a pale yellow crystalline
substance (with an odor of naphthalene), which shows
no tendency toward polymerization in air, despite
the formal presence of five conjugated double bonds.
The properties of the 10 x-electron system are evident
from the spectra.
The 1H-NMR spectrum (see Fig. 6) consists of an
AA'BB' system with TA= 2.54 and 'tg = 2.14 and the
coupling constants JAB = 8.8, JBB, = 9.3, JAB, = 0.3,
and JAA' = 1.1 Hz [76J. The absorption of the protons
at relatively low field and the type of spectrum,
according to which Hz to H5 and H7 to HI0 have
identical NMR parameters, appear to be incompatible
with a 1,6-0xido[lO]annulene having localized double
bonds. The NMR findings rather indicate that a ring
current is present in the peripheral 10x-electron
system, and hence that the compound has an aromatic
structure k771. Though the symmetry of the spectrum
could likewise be explained by a molecule with rapidly
fluctuating x-bonds, such a structure is not consistent
with the position of the proton resonances.
Fig. 6. 1H-NMR spectrum of 1,6-0xido[lOlannulene (in CC14;60 MHz,
internal standard : tetramethylsilane).
The coupling constants found for the AA'BB system also fit
into the above picture. This is particularly true of the vicinal
couplings, which are almost identical, as is to be expected
for a delocalized x-electron system in (71). On the other
hand, the vicinal coupling constants of olefinic protons in
conjugated cycloolefins generally show considerable differencesf57.64,781,which must be primarily due to the alternance
of the bonds, if the double bond system is sufficiently planarC791. It is interesting in this connection to note that the
a-protons in (71) absorb a t lower field strengths than the
@-protons1801, i.e. the resonances occur in the opposite order
[76] H . Giinther, Z . Naturforsch. 20b, 948 (1965).
[77] J. A. Elvidge and L. M . Jackman, 3. chem. SOC.(London)
1961, 859; L. M . Jackman, F. Sondheimer, Y. Amiel, D. A . BenEfraim, Y . Gaoni, R . Wolovsky, and A . A . Bothner-By, J. Amer.
chem. SOC. 84, 4307 (1962).
1781 C . Ganter and J. D. Roberts, J. Amer. chem. SOC. 88, 741
[79] According to theoretical [ M . Karplus, J. Amer. chem. SOC.
85, 2870 (1963)l and experimental [D. R . Eafon, A . D . Josey,
W . D . Phillips, and R.E. Benson, J. chem. Physics 39,3513 (1963)]
findings, vicinal coupling constants depend on the bond length
in the CH-CH segment in question.
[SO] This assignment, which cannot be made purely on the basis
of analysis, is established by specific deuteration in the a-position
[F. Gerson, E. Heilbronner, W . A . Boii, and E. Vogel, Helv. chim.
Acta 48, 1494 (1965)].
Angew. Chem. internat. Edit. J Vol. 6 (1967) J No. 5
to that found for the corresponding protons in oxepin and
cycloheptatriene. In fact, the a-protons should be more
strongly deshielded in (71) according to a n estimate of
the resonance positions using Pople's point-dipole modeWll.
The electronic spectrum of 1,6-0xido[lO]annulene contains
three bands with maxima a t 257 (E = 74500), 302(7100), and
about 390 rnp (300), the last of these showing pronounced
vibrational fine structure. 1,6-Methano[lO]annulene gives a
very similar spectrum. It may be concluded from this that
there is no appreciable interaction between the free electron
pairs o n the bridge oxygen atom and the peripheral 10 xelectron system in the oxido compound. The habit of the
spectra of (71) and (70) corresponds largely to that of the
spectrum of benzene, and so points to a close relationship
between the electronic structures of the x-systems in the
1,6-bridged [lO]annulenes and in benzene [8*1.
In agreement with the symmetry shown by t h e N M R spectrum, the IR spectrum of 1,6-oxido[lO]annulene contains
relatively few bands. The C-C double-bond band at the
unusually low frequency of 1538 cm-1, which appears t o be
characteristic of the carbon skeleton, is reminiscent of C-C
stretching vibrations of benzenoid compounds. This parallelism is also strongly suggested by the appearance of welldeveloped combination vibration bands in the range 1940 to
1580 cm-1.
The aromatic nature of 1,6-oxido[lO]annulene (71) is
also apparent from its chemical behavior. The compound is not only relatively stable toward heat,
oxygen, and light, but can also undergo substitution
reactions with electrophilic reagents. Thus it can be
nitrated with copper(I1) nitrate in acetic anhydride,
the nitro group being introduced into the 2- and 3positions [(77) and (78)] [73,747.
However, the possibility of obtaining substituted
products is considerably lower for 1,6-oxido[lO]annulene than for 1,6-methano[lOO]annulene,since
the oxido compound and its derivatives, like the arene
oxide-oxepin systems, are very sensitive toward proton
and Lewis acids. 1,6-Oxido[lO]annulene isomerizes
to u-naphthol (79) under the action of proton acids
in aqueous-ethereal media, while on silica gel in the
presence of nonpolar solvents it rearranges almost uniformly into the intense yellow I-benzoxepin (80) [741.
The formation of u-naphthol cannot involve (80) as an
intermediate, since on treatment with acids (80) does
not give u-naphthol, but products containing carbonyl
groups ; these indicate hydrolytic cleavage of the
oxepin ring. The 1,6-oxido[1O]annulene-l-benzoxepin
rearrangement has also been observed with the 2-nitro
and 2-bromo derivatives of (71). In the resulting
nitro- and bromo-1-benzoxepins, the substituent is
attached to the aromatic nucleus; its position has not
yet been determined.
[Sl] J . A . Pople, J. chem. Physics 24, 1111 (1956).
1821 H.-R. Blattmann, W . A . Boll, E. Heilbronner, G. Hohlneicher,
E. Vogel, and J.-P. Weber, Helv. chim. Acta 49,2017 (1966).
Concerning the mechanism of the acid-catalysed isomerization of 1,6-0xido[lO]annulene into a-naphthol, there is as
yet no evidence (unlike in the oxepin-phenol rearrangement)
that the reacting species is the corresponding arene oxide,
i.e. 9,lO-naphthalene oxide.
An alternative is that the proton attacks thea-carbon atom of
(7I) and so initiates the structural rearrangement (cf. the
formation of an arene oxide on bromination). The same difficulty is encountered in the interpretation of the surfacecatalysed rearrangement of 1,6-oxido[lO]annulene into
I-benzoxepin on silica gel.
I-Benzoxepin is separated from its arene oxide valence
isomer, 1,9-naphthalene oxide, by a relatively high
energy barrier ; consequently the oxepin character
should be much more pronounced (though modified by
the fused aromatic nucleus) than in the oxepin itself.
As far as can be seen from the reactions studied so far,
dibromo adduct which can be isolated. This is identified spectroscopically as the benzene oxide derivative
(86) with the bromine atoms cis to each other.
Whereas symmetry of the N M R spectrum indicates
that the position of the bromine atoms must be
cis-2,5, but gives n o further definite information
(the protons of the double-bond system appear as a
singlet), a concIusive decision between the benzene
oxide and the oxepin structures is possible with the
aid of the UV spectrum. As is to be expected for the
benzene oxide structure (86), the UV spectrum agrees
closely with that of 8,9-indan oxide (24), apart from
- 78OC
the chemical behavior of (80) corresponds to that of
an en01 ether. Like oxepin, (80) can be photochemically isomerized into a cyclobutene derivative (81)1831.
the high end-absorption of the adduct due to the
We have confirmed that the I-benzoxepin obtained from (71)
atoms. An oxepin, on the other hand, should
is free from the isomeric 3-benzoxepin (84), which is also
yellow, and which was first synthesized by Dimroth and
give an absorption extending into the visible region.
Pohlr4J by a Wittig reaction of dimethylether-a,cr'-bis(triIt is not yet certain whether the bromine attacks the
phenylphosphonium) dibromide (83) with phthalaldehyde
from above or below relative to the oxygen bridge.
(82). 3-Benzoxepin has recently also been obtained by
reaction of l-bromo-2,3-epoxy-1,2,3,4-tetrahydronaphtha- On further bromination, the dibromo adduct is
Iene (85) with potassium t-butoxide in ether (841.
converted into a tetrabromo adduct which, according
to the NMR spectrum, has the structure (87) with cis
bromine atoms in both cyclohexene rings ;thus another
cis-l,4 addition has taken place. The pairs of bromine
atoms probably have syn and anti configurations with
respect to the epoxide ring. (71) can be regenerated
quantitatively from both the dibromo and the tetrabromo adduct on treatment with sodium iodide in
1,6-Oxido[lO]annulene also resembles arene oxide
oxepin systems in that the oxygen can be removed by
triphenylphosphine (160 "C) or by Cr(C0)3(NH3)3 (in
boiling hexane) 1651; the reaction product is naphthalene. As a result of the resonance stabilization in 1,6oxido[lO]annulene, however, the reactivity towards
dienophiles is drastically reduced.
A very close chemical relationship between 1,6-oxido[IOIannulene and arene oxides is established by
addition-elimination reactions, a typical example of
which is the bromination of the oxido compound [851.
Under the action of bromine in methylene chloride
at -78 'C, 1,6-0xido[lO]annulene gives a colorless
[83] A derivative of I-henzoxepin was recently described by
H. Hofmann [Angew. Chem. 77, 864 (1965); Angew. Chem.
internat. Edit. 4,872 (1965)J; see also H. Hofmann and H. Westernacher, ibid. 78, 980 (1966); Angew. Chem. internat. Edit. 5 , 958
[84] E. Vogel, M . Biskup, and F.-G. Klarner, unpublished.
(851 E. Vogel, W. A . Boll, and M. Biskup, Tetrahedron Letters
1966, 1569.
Attempted thermal dehydrohalogenation of the dibromo adduct (86) 1861 has so far led only to unidentifiable decomposition products. However, if the dehydrobromination is carried out with bases such as
potassium t-butoxide (in ether), the reaction leads
smoothly to 2-bromo-1,6-oxido[lO]annulene (88), the
NMR and UV spectra of which correspond to those
of 2-bromo-l,6-methano[1O]annulene
[721. Treatment
of (88) with butyl-lithium (ether, -95 "C) yields 2lithio-l,6-oxido[lO]annulene(89), from which it is
possible to obtain 1,6-oxido[l0]annulene-2-carboxylic
acid (90) and other 2-substituted 1,6-oxido[l0]annulenes.
The reaction of the tetrabromo adduct (87) with
potassium t-butoxide led to a dibromo substitution
product of (71), which was found to be pure 2,7-di[86] The dibromo adduct of 1,6-methano[lO]annulene, which
still has the same carbon skeleton as the hydrocarbon, loses
hydrogen bromide even below 0 "C.
Angew. Chem. infernat. Edit. 1 Vol. 6 (1967) No. 5
bromo-l,6-oxido[lO]annulene (91). This structure is
based on a spectral comparison with 2,7-dibromo-l,6methano[lO]annulene 1721, on the dipole moment [871,
and on the formation of 1,5-dibromo-naphthalene
(92) when (91) is heated with triphenylphosphine at
160 “C.
In the compounds in which we are interested, i.e.
1,6;8,13-bisoxido[14]annulenes, the 2p,-orbitals on
C6, C7, and C* (like those on C1, C14, and C13) can
assume an approximately parallel position in the syn
form (95), whereas in the anti form (96) the 2pzorbitals on C6 and C7 as well as those on C7 and C*
exhibit appreciable twisting in relation to each other.
Thus the resonance stabilization should undoubtedly
be highest in the syn form.
VII. 1,6; 8,13-Bisoxido[l4]annulene
The fact that the spectral and, to a large extent, the
chemical properties of 1,6-oxido[lO]annulene (71) are
dominated by the 10 x-electron system of the Clo perimeter and not by the “oxepin structural unit” suggested the idea of a homologous series of bridged
annulenes (71), (93), (94), etc . , which, like thelinearly
fused aromatic hydrocarbons (the acenes) on which
they are based, contain (4n + 2) x-electrons. According
to calculations by Dewar and Gleicher [*81 the Hiickel
rule may be regarded as valid for up to the 22 xelectron system; consequently the question of whether
the aromaticity observed in the prototype (71) also
extends to its nearest homologues should depend mainly
on the geometry of the perimeter.
As soon as the number of bridges is greater than one,
geometrical isomerism occurs, with the result that
(93) can exist in a syn and a n anti form, while (94) can
exist in three forms (syn-syn, anti-anti and syn-anti).
Dreiding and Stuart- Briegleb models of the homologues of (71) show that these molecules, unlike the
flexible annulenes [891, have relatively rigid ring skeletons, and that substantial planarity of the perimeter is
possible only in the syn and all-syn forms. The Stuart
models indicate that there is no mutual steric hindrance
of the syn-oxygen atoms, which would lead to bending
of the carbon skeleton. Since the increase in the
carbon perimeter from 10 to 1 4 or 1 8 C atoms is
accompanied by a n appreciable relief of strain in the
C-0-C angle, the C skeleton can in fact be flatter in
syn-(93) and syn-syn-(94) than in the parent compound (71)[901.
[87] W. Bremser, H . T. Grunder, E. Heilbronner, and E. Vogel,
Helv. chim. Acta, 50, 84 (1967).
[ 8 8 ] M.J. S . Dewar and G. J . Gleicher, J. Amer. chem. SOC.87,
685 (1965).
[89] Y.Gaoni, A. Mtlera, F. Sondheimer, and R. Wolovsky, Proc.
chem. SOC.(London) 1964, 397; I. C. Calder and F. Sondheimer,
Chem. Commun. 1966,904; G. Schroder and J . F. M. Oth, Tetrahedron Letters 1966, 4083.
1901 Cf. the X-ray analysis of 1,6-methano[lO]annulene-Zcarboxylic acid [ M . Dobler and J. D . Dunitz, Helv. chim. Acta
48, 1429 (1965)J.
Angew. Chem. internat. Edit.
1 Vol. 6 (1967)/ No. 5
The qualitative validity of the views deduced from the
Hiickel rule and from molecular models concerning
the homologues of 1,6-0xido[lO]annulene (71) is
demonstrated by attempts to synthesize syn- and anti1,6; 8,13-bisoxido[l4]annulene.
I n view of the synthesis of (71) from 1,4,5,8-tetrahydronaphthalene, the most suitable starting compound for the synthesis of the two bisoxido[l4]annulenes (101) and (105) appeared to be 1,4,5,8,9,10hexahydroanthracene (97) [911. Reaction of the tetraene (97) with two equivalents of perbenzoic acid leads
almost quantitatively to 1,4,5,8,9,10-hexahydro-4a,9a
8a,10a- bisoxidoanthracene, which, according to the
NMR spectrum, consists of the syn (102) and the anti
isomer (98) (molar ratio approximately 60: 40), since
the central CH2-protons give rise both to a n ABsystem (syn form) and to a singlet (anti form). The
isomer mixture could be separated by fractional
crystallization. Treatment of the syn- and anti-diepoxides (102) and (98) with bromine gives the
corresponding tetrabromo adducts (103) and (99)
(the configurational relationship between the remote
bromine atoms is uncertain). The dehydrobromination
of the two tetrabromides with potassium t-butoxide
in tetrahydrofuran at 0°C leads to a surprising
Whereas (99) lost four moles of hydrogen bromide to
form the colorless bis(arene oxide) (ZOO), the reaction
of (103) led directly to the desired syn-bisoxido[l4]annulene (105), which was obtained as a stable
carmine compound. Thus in the latter case the dehydrobromination was followed by dehydrogenation,
evidently by atmospheric oxygen. The syn configuration
of the bisoxido[l4]annulene is shown, not only by the
[91] J . Runge, Z . Chem. 2, 374 (1962); A. J. Birch, P. Fitton,
D . C . C. Smith, D. E. Steere, and A . R . Stewox, J . chem. SOC.
(London) 1963, 2209; E. Voget, M. Biskup, A . Vogel, U.Haberland, and J. Eimer, Angew. Chem. 78,642 (1966); Angew. Chem.
internat. Edit. 5, 603 (1966).
synthesis [921, but also by the dipole moment of 3.25 D,
which is of the expected order of magnituder931. In
the meantime, it has been possible to detect spectroscopically the dihydro precursor of (105). This is not
the bis(arene oxide) corresponding to (loo), but the
bisoxepin (104).
The compound (100) has so far resisted all attempts
to dehydrogenate it to anti-bisoxido[14]annulene
(101) ; this indicates that (lor), unlike the syn-isomer,
has no appreciable resonance stabilization. A 1,6;8,13bisoxido[l4]annulene of unknown configuration [941,
which was found to be identical with (105), had been
previously obtained from the mixture of syn- and
anti-diepoxides (102) and (98) by reaction with Nbromosuccinimideand treatment of the crude bromination product with potassium t-butoxide in tetrahydrofuran.
A chemical proof of the structure of syn-bisoxido[l4]annulene (105), which, however, does not take into
account the configuration, is based on its reaction
with Cr(C0)3(NH&, which leads to deoxygenation
and formation of anthracene in yields of up to 95 % “551.
This reaction should be useful in the determination of
the structures of substitution products of (105).
The NMR spectrum (Fig. 7) of syn-bisoxido[l4]annulene (105) shows a singlet at z = 2.06 (relative
intensity 2) for the protons H7 and H14 and an AA’BB’
system (TA = 2.25 and TB = 2.40, relative intensity 8),
which corresponds to the protons Hz to H5 and H9 to
H12. The fact that the proton resonances occur at
relatively low field strengths, together with the symmetry of the spectrum, points to a ring current in the
peripheral 14 x-electron system. With 1,6-oxido[lO]annulene (71) as the reference substance, an estimate
of the chemical shift from Pople’s point-dipole
model 1811 gives TC = 1.92, TA = 2.27, and ‘tg = 2.65, in
1921 Oscillation of the bridge oxygen atoms through the peripheral C14 ring seems unlikely on steric grounds [the rotation
of optically active methyl 1,6-methano[lO]annulene-2-carboxylate
remains unchanged even on prolonged heating at 240°C
(experiments by W. Schrock, unpub1ished)l.
[93] According to measurements by W. Bremser, unpublished.
1941 E. Vogel, M . Biskup, A. Vogel, and H. Giinfher, Angew.
Chem. 78, 755(1966); Angew.Chem. internat. Edit. 5 , 734(1966).
good agreement with the experimental values. The
spin-spin interactions found for (105) are in remarkable agreement with those of (71) (cf. Section VI);
= 9.0, J 5 B = 9.2, JAB’= 0 . 3 and JAN = 1.1Hz.
Since the vicinal coupling constants JAB
and JBB, are
practically identical, it is probable (as in the cases of
(70) and (71) 1901) that the lengths of the C-C bonds
in question do not differ appreciably.
2 L
Fig. 7. ‘H-NMR spectrum of syn-1.6; 8,13-bisoxido[l41annulene(in
CDCIJ; 60 MHz; internal standard: tetramethylsilane).
The electronic spectrum of syn-bisoxido[14]annulene (105)
shows a four-band system, as is reported for the aromatic
14 x-electron system trans-l5,16-dimethyldihydropyreneL951.
The maxima, which occur at slightly lower wavelengths than
for the dihydropyrene derivative, are located at 306 ( E =
169000), 345 (14400), 382 (8500), and 555 m p (775), the band
at 555 m y exhibiting vibrational fine structure. Though the
C14 perimeter i n bisoxido[l4]annulene cannot become
completely planar and electronic interaction of the oxygen
atoms with the 14 x-electron system is not ruled out, the
energy transitions calculated by the Pariser-Parr method for
an unperturbed 14 x-electron system are surprisingly well
satisfied. This is particularly true of the transition at the
longest wavelength, whose expected value of 2.17 eV (= 512
my) corresponds to the band between 543 and 565 mp[961.
[951 V. Boekelheide and J . B. Phillips, Proc. nat. Acad. Sci.
USA 51, 550 (1964); F. Cerson, E. Heilbronner, and V. Boekelheide, Helv. chim. Acta 47, 1123 (1964).
1961 H . - P . Blattmann, V . Boekelheide, E. Heilbronner, and J.-P.
Weber, Helv. chim. Acta, 50, 68 (1967).
Angew. Ckem. internat. Edit.
1 Yol. 6 (1967) I No. 5
The I R spectrum, which contains relatively few bands owing
to the symmetry of the compound, contains only one band
at 1536 cm-1 in the C-C stretching region, which practically
coincides with the analogous absorption of 1,6-oxido[lO]annulene (1538 cm-1).
The aromatic character shown by the spectra of (105)
is reflected in the considerable thermal stability of the
compound and in its insensitivity to oxygen. Moreover,
preliminary experiments indicate that (105) gives
substitution products under the action of various
electrophilic reagents.
The substantial similarity between the spectral and
chemical properties of 1,6-oxido[lOlannulene and syn1,6; 8,13-bisoxido[l4]annulene provides a firm experimental basis for the idea of an homologous series
of bridged annulenes having aromatic character
within the range of validity of the Hiickel rule.
VIII. Procedures
Oxepin-Benzene Oxide :
A slurry of pulverized 1,2-epoxy-4,5-dibromocyclohexane
(17) 1141 (12.7 g, 50 mmole) in ether (25 ml) is added over a
period of 10 min to a suspension of freshly prepared sodium
methoxide (8.1 g, 150 mmole) in boiling ether (30 ml). After
the reaction mixture has been refluxed for a further 5 min
i t is cooled in ice, and then 30 ml of water is added. The
ethereal phase is washed with 2x20 ml of water and the
combined washes extracted with 20 ml of ether (to recover
any entrained oxepin-benzene oxide). The ethereal phase and
the ether extract (from the washes) are then combined and
dried over MgS04. The oxepin-benzene oxide is recovered
by distillation using a 15 cm Vigreux column; the ether is
distilled off first and the oxepin-benzene oxide finally recovered by distillhtion under vacuum (water pump). Yield :
3.85 g (80 %) of orange oxepin-benzene oxide (b.p. 27 'C/14
mm, 38OC/30 mm; n: = 1.5163). The product may be
distilled under normal pressure (b.p. 124 "C)in a n apparatus
that has been treated with alkali.
Angew. Chem. internat. Edit. 1 Vol. 6 (1967)1 No. 5
A solution of 4,5-dibromo-1,2-epoxy-l,2-dimethylcyclohexane (28) (20 g) (prepared from 1 ,2-epoxy-l,2-dimethylcyclohex-4-ene 1971 and bromine in methylene chloride
at -78OC, m.p. 85-86°C) in ether (30 ml) is added t o a
stirred suspension of sodium methoxide (14 g, 260 mmole)
in boiling ether (30 ml). The reaction mixture is refluxed for
2 h, cooled o n ice, and water (50 ml) is then added. The product is recovered as in the previous procedure. Yield: 7.3 g
(84 %) yellow-orange 2,7-dimethyloxepin (b.p. 49-50 "C/
11 mm; n: = 1.5045'3.
8,9-Indan Oxide :
A solution of 5,6-dibromo-8,9-epoxy-4,7-dihydroindan
n = 3, (14.8 g) (prepared from 8,9-epoxy-4,7-dihydroindan(981
and bromi ie in methylene chloride a t -78 "C: m.p. 87 t o
88°C) in anhydrous ether (150 ml) is added to a stirred
suspension of potassium t-butoxide (25 g, 220 mmole) in
anhydrous ether (225 ml) at 0 "Cover a period of 1 h.
After the reaction mixture has been stirred for 5 h at room
temperature and then for 15 min at 40 "C,it is cooled o n ice
and 2 N KOH (40 ml) is added. Since the product is unstable
towards water in neutral solution, it is washed twice with
2 N KOH (40 ml); it is subsequently recovered in the normal
way. Yield: 5.7 g (85 %) of 8,9-indan oxide as a colorless
liquid (b.p. 30-31 "C/0.4mm; n: = 1.5255).
Thanks are due to M . Biskiip, R . Schubart, W. A . Boll,
M . Wiesel, F,-G. Klarner, and R. Suridermann for their
enthusiastic cooperation. We aregrateful to H. Friebolin,
Freiburg, for recording the lo w-temperature N M R
spectra. Acknowledgment is also due to the Deutsche
Forsch~ingsgerneinschajtand the Fonds der Chemischen
Industrie for their generous support of our work.
Received: February 9th. 1967
[A 572 IEl
German version: Angew. Chem. 79, 429 (1967)
Translated by Express Translation Service, London
1971 W. Huckel and U.Worffel, Chem. Ber. 88, 338 (1955).
[98] E. Giovannini and H . Wegmiiller, Helv. chim. Acta 41, 933
40 1
Без категории
Размер файла
1 750 Кб
oxide, valence, oxepin, tautomeric, benzenes
Пожаловаться на содержимое документа