Cycloaddition reactions of 7-benzylidenecycloocta-1,3,5-triene with ethenetetracarbonitrile and 4-phenyl-3H-1,2,4-triazole-3,5(4H)-dione

Article abstract of DOI:10.1002/hlca.200590153

An (E)/(Z) mixture (3:2) of 7-benzylidenecycloocta-1,3,5-triene (5) is obtained when 1-benzylcycloocta-1,3,5,7-tetraene (7), prepared by an improved procedure, is treated with t-BuOK in THF. Alternatively, a ca. 9:1 mixture (E)/(Z)-5 can be prepared in a Wittig reaction involving benzaldehyde and cycloocta-2,4,6-trien-1-ylidenetriphenylphoshorane (9). Treatment of (E)/(Z)-5 88:12 with ethenetetracarbonitrile (TCNE) gave a complex mixture of products, from which seven mono-adducts and two bis-adducts were isolated (Sect. 2.2.1). Of the mono-adducts, four are π4 + π2 adducts: two ((E)- and (Z)-isomers) are derived from valence tautomers of the two isomers of (E)/(Z)-5, while it is tentatively suggested that the other two (again (E)- and (Z)-isomers) are formed from the intermediacy of a pentadienyl zwitterion (Sect. 2.3). The remaining three mono-adducts, two of which are epimers, are π8 + π2 adducts. It is suggested that they are derived from the intermediacy of homotropylium zwitterions (Sect. 2.3). For the two bis-adducts, it is postulated that they are derived from an initial π2 + π2 cycloaddition involving the homotropylium zwitterions followed by π4 + π2 cycloaddition to the valence tautomer of each of the π2 + π2 cycloadducts. With 4-phenyl-3H-1,2,4-triazole-3,5(4H)- dione (6), (E)/(Z)-5 91:9 yielded two π4 + π2 cycloadducts ((E)- and (Z)-isomers) as well as two epimeric π8 + π2 cycloadducts (Sect. 2.2.2). The intermediacy of pentadienyl (tentative suggestion) and homotropylium zwitterions accounts for the formation of the products (Sect. 2.3).

Full text of DOI:10.1002/hlca.200590153

Helvetica Chimica Acta Vol. 88 (2005)
2003
Cycloaddition Reactions of 7-Benzylidenecycloocta-1,3,5-triene with
Ethenetetracarbonitrile and 4-Phenyl-3H-1,2,4-triazole-3,5(4H)-dione
by Philip Clements, George E. Gream*, Paul K. Kirkbride1), and Simon M. Pyke
School of Chemistry and Physics, University of Adelaide, Adelaide, S. A. 5005, Australia
(phone: ‡ 61883035652; fax: ‡ 61883034358; e-mail:george.gream@adelaide.edu.au)
Dedicated to Professor Rolf Huisgen on the occasion of his 85th birthday
An (E)/(Z) mixture (3 :2) of 7-benzylidenecycloocta-1,3,5-triene ( 5) is obtained when 1-benzylcycloocta-
1,3,5,7-tetraene (7), prepared by an improved procedure, is treated with t-BuOK in THF. Alternatively, a ca. 9 : 1
mixture (E)/(Z)-5 can be prepared in a Wittig reaction involving benzaldehyde and cycloocta-2,4,6-trien-1-
ylidenetriphenylphoshorane (9). Treatment of (E)/(Z)-5 88 :12 with ethenetetracarbonitrile (TCNE) gave a
complex mixture of products, from which seven mono-adducts and two bis-adducts were isolated (Sect. 2.2.1).
Of the mono-adducts, four are p4 ‡ p2 adducts:two (( E)- and (Z)-isomers) are derived from valence tautomers
of the two isomers of (E)/(Z)-5, while it is tentatively suggested that the other two (again (E)- and (Z)-isomers)
are formed from the intermediacy of a pentadienyl zwitterion (Sect. 2.3). The remaining three mono-adducts,
two of which are epimers, are p8 ‡ p2 adducts. It is suggested that they are derived from the intermediacy of
homotropylium zwitterions (Sect. 2.3). For the two bis-adducts, it is postulated that they are derived from an
initial p2 ‡ p2 cycloaddition involving the homotropylium zwitterions followed by p4 ‡ p2 cycloaddition to the
valence tautomer of each of the p2 ‡ p2 cycloadducts. With 4-phenyl-3H-1,2,4-triazole-3,5(4H)-dione (6), (E)/
(Z)-5 91:9 yielded two p4 ‡ p2 cycloadducts ((E)- and (Z)-isomers) as well as two epimeric p8 ‡ p2
cycloadducts (Sect. 2.2.2). The intermediacy of pentadienyl (tentative suggestion) and homotropylium
zwitterions accounts for the formation of the products (Sect. 2.3).
1. Introduction. In 1990, it was reported [1] that a number of 7-alkylidenecy-
cloocta-1,3,5-trienes undergo interesting p8 ‡ p2 cycloaddition reactions with a variety
of p2-addends. As a representative example, 7-methylenecycloocta-1,3,5-triene (1)
reacts with ethenetetracarbonitrile (ˆtetracyanoethylene ˆ TCNE) to give the adducts
2 and 3. It was postulated [1] that the intermediate homotropylium zwitterion 4 was
involved with charge annihilation occurring by bond formation at C(2) and C(7) to give
2 and 3, respectively.
As an extension of the above studies, the synthesis of 7-benzylidenecyclocta-1,3,5-
triene (5) (as an (E)/(Z) mixture) and its addition reactions with TCNE and 4-phenyl-
3H-1,2,4-triazole-3,5(4H)-dione (ˆ PTAD; 6) are the subject of this paper.
2. Results and Discussion. 2.1. Synthesis of 7-Benzylidenecycloocta-1,3,5-triene
(5). The fact that alkylcyclooctatetraenes can be converted into equilibrium mixtures
with 7-alkylidenecycloocta-1,3,5-trienes by treatment with t-BuOK [1] led to the first of
two successful routes to 7-benzylidenecycloocta-1,3,5-triene (5). Treatment of 1-
benzylcycloocta-1,3,5,7-tetraene (7) with t-BuOK in tetrahydrofuran (THF) gave an
equilibrium mixture (95 :5) of 5 and starting material 7 which was freed of the latter by
1
)
Present address: Forensic Science South Australia, 21 Divett Place, Adelaide, S.A. 5000, Australia.
¹ 2005 Verlag Helvetica Chimica Acta AG, Z¸rich

2004
Helvetica Chimica Acta Vol. 88 (2005)
extraction with a 20% aqueous AgNO3 solution [2] to give 5 (84%) as a 3 :2 ( E)/(Z)
mixture. Benzylcyclooctatetraene 7, which had been prepared previously in low yields
by two procedures [2], was now obtained in high yield (88%), but mixed with bibenzyl
(4%), from the reaction of 1-bromocycloocta-1,3,5,7-tetraene with an excess of
benzylmagnesium chloride catalyzed by FeCl3 [3]. Purification, though relatively
inefficient, was effected by extraction with 20% aqueous AgNO3 solution.
Since 7-methylenecycloocta-1,3,5-triene (1) can be prepared by a Wittig reaction
between cycloocta-2,4,6-trienone (8) [4] and methylenetriphenylphosphorane [1], it
was envisaged that the benzylidenecyclooctatriene 5 might be synthesized from the
same ketone and benzylidenetriphenylphosphorane. The reaction between the
phosphorane, generated from benzyltriphenylphosphonium chloride [5] under differ-
ent conditions, and the ketone gave, however, only a trace, if any, of 5. The compound
was, however, prepared as a (E)/(Z) mixture in ratios varying from 84 :16 to 91:9 by a
Wittig reaction between benzaldehyde and cycloocta-2,4,6-trien-1-ylidenetriphenyl-
phosphorane (9). The precursor of 9, cycloocta-2,4,6-trien-1-yltriphenylphosphonium
bromide (10), was prepared from a ca. 7:3 mixture of 7-bromocycloocta-1,3,5-triene
(11) and its valence tautomer 12 [6]. It was a most unstable and difficult-to-handle
compound and could not be recrystallized. The selective reaction of 11, and not 12, with
triphenylphosphine is the result of the much greater reactivity of an allylic bromide

Helvetica Chimica Acta Vol. 88 (2005)
2005
compared with a cyclobutyl bromide [7] for such a reaction. A similar selectivity was
observed in the reaction of 11/12 with sodium azide to give 7-azidocycloocta-1,3,5-
triene [6]. The two isomers (E)- and (Z)-5 could not be separated by gas-liquid (GLC)
or thin-layer chromatography (TLC). Recognition of their presence and assignment of
their configurations was based on NMR spectroscopy.
In the 1H-NMR spectrum of 5, a pair of d at d 3.30 and 3.40 was present for the two H-atoms at C(8) of each
isomer. Although the relative areas of each of the d enabled the proportions of the two isomers to be
determined, 13C-NMR spectroscopy was used initially to assign the configurations. Resonances for C(8) of each
isomer ((E)/(Z)-5 9 :1) were observed as a t at d 29.39 and a much less intense one at d 37.76 in the off-resonance
spectrum. Because the two H-atoms at C(8) are shielded by the Ph group in the (E)-isomer, the higher-field and
1
more intense t was assigned to (E)-5 [8]. For the H-NMR spectrum, it was therefore concluded that the d of
greater intensity and at slightly lower field (d 3.40) corresponded to (E)-5. This assignment was confirmed by
examination of the ROESY plot where interaction between the olefinic HÀCˆC(7) at d 6.23 with HÀC(8) at d
3.30 clearly established the (Z) configuration at C(7) of the minor isomer. It should be noted that the 1H-NMR
spectrum of the mixture (E)/(Z)-5 did not contain any resonances that could be attributed to the presence of the
bicyclic valence tautomers (E)- and (Z)-13.
Essentially pure (E)-5 can be obtained as recovered starting material (ca. 20%)
when the initially obtained (E)/(Z) mixture (ca. 9 :1) is treated with 0.8 equiv. of
PTAD (6). Other 7-arylidenecycloocta-1,3,5-trienes have been prepared successfully
from phosphorane 9 [9].
2.2. Products of Cycloadditions. 2.2.1. With Ethenetetracarbonitrile (TCNE).
Treatment of a 88 :12 mixture of ( E)/(Z)-5 with an excess of TCNE in boiling AcOEt
gave a complex mixture from which nine products, (E)-14 (8%), (Z)-14 (0.6%), 15
(8%), 16 (11%), endo-17 (8%), exo-17 (20%), 18 (5%), (E)-19 (15%), and (Z)-19
(0.3%) were isolated. After fractional crystallization of the crude mixture had given
some (E)-14 and 15, chromatography (silica gel) of the remaining mixture gave four
fractions. From these, the nine products were isolated, in general after exhaustive
fractional crystallization. Structures of the products were elucidated by the extensive
use of NMR spectroscopy and, in three cases, confirmed by X-ray analysis of single
crystals.
Compounds (E)- and (Z)-14 were shown to be a pair of isomeric mono-adducts by
elemental analysis, the observation of four signals attributable to the CN groups in their
13C-NMR spectra, and mass spectrometry. The structural framework of (E)-14, the 3-
benzylidenetricyclo[4.2.2.02,5]dec-7-ene skeleton, was readily elucidated on the basis of
the 1H-and 13C-NMR and COSY data. Assignment of (3E)- configuration was derived
from the ROESY plot. In particular, interactions between HÀC(1) and HÀC(2) with
the olefinic HÀCˆC(3) clearly indicated the orientation of the benzylidene group (see
Fig. 1). The structure of (E)-14 was confirmed by X-ray crystallographic analysis [10].
Although only a small quantity of (Z)-14 was isolated, its structure was readily deduced
1
by NMR spectroscopy. As would be expected, its H- and 13C-NMR spectra had
features similar to those of the geometric isomer (E)-14. The resonance for HÀC(2) in
(Z)-14 was shifted downfield by ca. 0.3 ppm compared with that for HÀC(2) in (E)-14
due to deshielding by the Ph group. Confirmation of the (3Z)-configuration was
derived from the ROESY plot. In particular, the interaction between the olefinic
HÀCˆC(3) and HaÀC(4)/HbÀC(4) together with the interaction of the Ho of Ph with
HÀC(1)/HÀC(2) clearly indicated the opposite geometry of the benzylidene residue

2006
Helvetica Chimica Acta Vol. 88 (2005)
‡
(see Fig. 1). The mass spectrum of (Z)-14 showed a major peak at m/z 322 for M and
another one at m/z 194 as a result of the expected retro-Diels Alder fragmentation.
Fig. 1. ROESY Interactions for (E) and (Z)-14
Compounds 15 and 18 were shown to be a pair of bis-adducts by consideration of
1
microanalyses, and H- and 13C-NMR data. For each of 15 and 18, the key skeletal
fragment, tricyclo[4.2.2.02,5]dec-7-ene, was again readily elucidated from their 1H-NMR
and COSY data. The attachment of the additional spiro-linked ring was then derived by
interpretation of the HMQC and HMBC data. For both isomers, the orientation of the
spiro-linked ring and the configuration at C(2')2) was determined by consideration of
their ROESY plots. The key ROESY interactions that support these structures are
shown in Fig. 2. The proposed structures for 15 and 18 are also supported by a number
of chemical shift differences observed for these structures. In particular, those for
2
)
Arbitrary numbering; systematic names are given in parentheses in the Exper. Part

Helvetica Chimica Acta Vol. 88 (2005)
2007
HÀC(7) and HÀC(8) of 15 are shifted 0.80 and 1.01 ppm downfield in comparison to
the same H-atoms of 18 (consistent with substantial shielding of these olefinic H-atoms
by the proximal CN groups in 15). A similar observation is made for the chemical shift
of HaÀC(4); its proximity to the C(7)ˆC(8) p-bond results in an anisotropic shielding
in 18 (an upfield shift of 0.46 ppm).
Fig. 2. ROESY Interactions for 15 and 18. An asterisk is used to indicate the fragment C(CN)2 which is omitted
for clarity.
Importantly, 15 and 18 are not derived simply by the addition of TCNE to any of the
mono-adducts (E)-14, 16, endo-17, exo-17, or (E)-19. This was established by the
finding that each of these compounds was recovered unchanged when treated with the
p2-addend under the conditions of the initial reaction. Attempts to produce crystals of
15 and 18 suitable for X-ray analysis were unsuccessful.
The structure of 16 was deduced from 1H- and 13C-NMR data which, when account
is taken of the Ph group at C(11), were similar to those of the adduct 2 formed from 7-
methylenecycloocta-1,3,5-triene (1) [1]. Assignment of the configuration was deduced
by reference to the ROESY plot where HÀC(3) showed a cross-peak with one of the
cyclopropyl H-atoms at C(2), and the Ho of Ph and HÀC(11) also showed cross-peaks
with one of the cyclopropyl H-atoms at C(2). The proposed configuration was
confirmed by X-ray crystallography [11].
The spectral properties of endo-17 bore similarities to those of the 1:1 adduct 3
formed in the reaction between 1 and TCNE [1]. The replacement of one of the H-
atoms at C(10) in the adduct 3 by the Ph group caused the expected [12] downfield shift
1
(from d 3.26 to 4.25) in the chemical shift of HÀC(10). Analysis of the H-NMR and
COSY data revealed the bicyclo[5.3.1]deca-1,3,5-triene framework. Dreiding models
show that the six-membered ring in endo-17 would exist in a chair conformation with
the Ph group in an equatorial position and HÀC(10) axial. Evidence to support this
deduction was obtained from NOE difference experiments (irradiation at d 2.46 (d,
HbÀC(11)) ! NOEs at d 4.25 (13.7%, HÀC(10)), 3.34 (35.0%, HaÀC(11)), 3.88
(8.0%, HÀC(7)). These results demonstrate that HÀC(7) and HÀC(10) are cis to each
other as shown.
The 1H-NMR spectral features of exo-17, the epimer of endo-17, allowed assignment
of this structure to be made without difficulty. For exo-17, steric interactions between
the Ph group and one of the CN groups at C(8) is likely to cause the ring to adopt a
conformation other than the chair form. The Ph group causes considerable deshielding
of HbÀC(11) and this, combined with an unexpected upfield shift in the resonance of

2008
Helvetica Chimica Acta Vol. 88 (2005)
HaÀC(11), results in the chemical-shift difference between the two H-atoms at C(11) in
exo-17 being only 0.14 ppm compared with 0.88 ppm for the corresponding H-atoms in
endo-17. The inverted configuration at C(10) is supported by the lack of NOE between
HÀC(10) and either of the H-atoms at C(11).
1
Assignment of the structure to (E)-19 rested initially on H-NMR and micro-
analytical data. Dreiding models showed that the preferred conformation is as shown
for (E)-19 with C(4) pointing away from C(9) and C(10)2) which bear the CN groups.
Shielding of HaÀC(4) by the p-bond C(7)ˆC(8) and deshielding of HbÀC(4) by the Ph
group account for the chemical shifts of this pair of H-atoms. The magnitudes of the
coupling constants are consistent with the dihedral angles in the proposed conforma-
tion (E)-19, particularly J(3,4b) ˆ 10.0 Hz (consistent with dihedral angle of ca. 08) and
J(3,4a) ˆ 4.5 Hz (higher than normal for a dihedral angle of ca. 1108). It should be
noted that the small difference between the chemical shifts of HÀC(1) and HÀC(6) is
consistent with the compound being the (E)-isomer (E)-19 and not the (Z)-isomer
(Z)-19 (see below). The characteristics of the UV spectrum of (E)-19 resembled
closely those of b,b-dimethylstyrene (ˆ(2-methylprop-1-enyl)benzene) [13] and
benzylidenecyclohexane [14] as would be expected. Finally, the structure was
confirmed by X-ray crystallography [15].
That (Z)-19 was an isomer of (E)-19 was clear from both the 13C-NMR spectrum
‡
which showed the presence of four CN groups and the MS which showed an M at m/z
322 and an intense peak for the fragment ion at m/z 194, which corresponds to a loss of
1
TCNE. As would be expected, the H-NMR spectrum of (Z)-19 showed features
allowing ready recognition that the compound was the (Z)-isomer of (E)-19.
Assignment of the configuration was made by consideration of the ROESY data
where interactions between the Ho of Ph and HÀC(6) (d 4.41) and HÀC(7) (d 6.22)
confirmed this geometry.
2.2.2. With 4-Phenyl-3H-1,2,4-triazole-3,5(4H)-dione (6). A (91:9) mixture ( E)/
(Z)-5 in AcOEt reacted readily with 6 at room temperature to give a mixture of four
mono-adducts: exo-20 (46%), endo-20 (28%), (Z)-21 (4%), and (E)-21 (16%). The
first two and the last compounds were obtained in crystalline forms after chromatog-
raphy (silica gel). That exo- and endo-20 were related to 22 [1], and were epimeric at
C(2), was clearly indicated by the 1H- and 13C-NMR data. Replacement of one of the
H-atoms at C(2) in 22 by the Ph group caused the expected downfield shift [12] of the
remaining H-atom at C(2) (d 2.06 in exo-20 and 1.26 in endo-20). As well, the
introduction of the Ph group at C(2) caused the expected downfield shift of that C-atom
in the 13C-NMR spectrum:the downfield shifts were 11.3 and 15.4 ppm, respectively,
compared to 22. Assignment of the relative configurations at C(2) and C(8) in exo-20
rested on i) the relatively small chemical-shift difference (Dd ꢀ 0.3 ppm) for HÀC(13)
between exo-20 and 22 (models show that the pseudoaxially disposed Ph group in exo-
20 should not have a significant effect on HÀC(13)), ii) absence of allylic coupling
between HÀC(13) and HÀC(2) (this is consistent with the dihedral angle between the
two H-atoms being close to 08 as shown by Dreiding models), and iii) the absence of
an NOE at HÀC(2) (d 6.05) after irradiation at HbÀC(14) (d 2.23). Expected NOEs
were observed for the signals of HaÀC(14) (d 2.86, 36.8%) and HÀC(8) (d 5.26,
5.2%). Confirmation of the structure of exo-20 was provided by X-ray crystallography
[16].

Helvetica Chimica Acta Vol. 88 (2005)
2009
For endo-20, the relative configurations at C(2) and C(8) were determined from the
1H-NMR spectrum by i) the marked chemical-shift difference (1.05 ppm) for HÀC(13)
in exo-20 compared to endo-20 (models show that the pseudoequatorially disposed Ph
group in endo-20 should shield HÀC(13) compared with exo-20), ii) the allylic
coupling, as revealed by the 1H,1H-COSY plot, between HÀC(13) and HÀC(2)
(models show that the dihedral angle between the two H-atoms is close to 908 thus
leading to a large allylic coupling (J(2,13) ˆ 3.0 Hz)), and iii) an NOE for the s of
HÀC(2) (d 5.79, 7.8%) on irradiation of the signal of HbÀC(14) (d 2.64). The
observation of the NOE between HÀC(2) and HbÀC(14) is compatible with the 1,3-di-
pseudoaxial orientation as is clearly shown in the Dreiding model. Unfortunately,
satisfactory crystals of endo-20 for an X-ray crystallographic analysis could not be
obtained. The correctness of the deduced structure was supported, however, by an X-
ray analysis [17] of the corresponding 4-methoxy-substituted derivative (formed from
1
7-(4-methoxybenzylidene)cycloocta-1,3,5-triene and 6 [9]) whose H- and 13C-NMR
spectral characteristics for the appropriate resonances were very similar to those of
endo-20.
Although it could be obtained only in a noncrystalline form and not entirely pure,
1
(Z)-21 possessed H- and 13C-NMR and COSY characteristics which bore marked
1
similarities to those of (E)-21. In the H-NMR spectrum of (Z)-21, the signal for
HÀC(1) was observed at d 5.85 compared with d 5.57 for HÀC(1) in (E)-21 (as would
be expected for deshielding by the Ph group in (Z)-21). Confirmation of the deduced
configuration was provided by the ROESY plot of (Z)-21 where cross-peaks were
observed between the signals for HÀCˆC(11) and both HaÀC(10) and HbÀC(10)
together with a cross-peak between the signals for HÀC(1) and the Ho of Ph. Initially,
the structure of (E)-21 was deduced from 1H- and 13C-NMR data which, not
surprisingly, were similar to those of (E)-19 when account is taken of the anticipated
downfield shifts of the resonances of HÀC(1) and HÀC(7) caused by the two N-atoms
in (E)-21. Models of (E)-21 do not show that there should be a preference for one of

2010
Helvetica Chimica Acta Vol. 88 (2005)
the two conformations with C(10) pointing towards, or away from, the p-bond
C(12)ˆC(13) (cf. (E)-19). Indeed, it would seem that interconversion between the two
conformations should occur readily. This is consistent with the coupling constants
observed between HÀC(9) and the two H-atoms at C(10) where the values appear to
be averaged (5.4 and 6.6 Hz). That the compound in the crystalline state exists in the
conformation with C(10) pointing towards the p-bond C(12)ˆC(13) was determined
by X-ray crystallography [18].
2.3. Mechanisms of the Cycloadditions. Let us now consider mechanisms which can
be postulated to account for the cycloadditions involving 5 reported here. As was the
case for 7-alkylidenecycloocta-1,3,5-trienes [1], cycloaddition reactions involving
concerted processes, and thus being thermally allowed by the Woodward Hoffmann
rules [19], for the tetraene 5 might be considered to be unlikely because its eight-
membered ring is tub-shaped and the extent of p-orbital overlap is limited. This is in
harmony with its UV spectrum in which the absorption maxima are at considerably
shorter wavelengths and are much less intense than would be expected for a completely
conjugated system. Indeed, the UV spectrum of 5 was similar to that of b,b-
dimethylstyrene [13] and benzylidenecyclohexane [14] establishing that conjugation
with the endocyclic CˆC bonds was absent. It cannot be assumed, however, that 5 is
unable to undergo cycloaddition reactions via concerted processes. It has been
postulated by Isaken and Snyder [20] that cyclooctatetraene, which is tub-shaped and,
therefore, in general unable to undergo concerted cycloaddition reactions, does so with
the very reactive PTAD (6) via the planar arrangement of the parent compound
involved in its ring-inversion process. It was suggested that the free-energy difference
(estimated to be 4.4 kcal/mol at 208) between cyclooctatetraene and its planar form is
sufficiently low to allow the concerted process to take place. At present, the free-energy
difference between the tub and planar forms of 7-methylenecycloocta-1,3,5-triene (1),
or any of its analogues, is not known. It should be noted that models indicate that ring
inversion for this class of molecules can occur with planarity of the ring p-system only
and with the alkylidene or arylidene portion almost perpendicular to the plane. As a
result, the ring-inversion process might be similar to that of cycloocta-1,3,5-triene (23)
where a free energy of 6.7 kcal/mol has been determined for the barrier to ring
inversion involving −near coplanarity× of the three CˆC bonds in the transition state
[21] (see also [22]). At this stage, the possibility that some of the p4 ‡ p2 cycloadditions
that take place with 5 might involve its planar form during ring inversion cannot be
discounted. On the other hand, the p8 ‡ p2 cycloadditions which occur with 5 cannot be
concerted not only because of the difficulty of attaining planarity of all four p-bonds
but more importantly because the distance between the two C-atoms that are involved
is too great for a p2-addend to be able to react in this way.

Helvetica Chimica Acta Vol. 88 (2005)
2011
Compound (E)-19 is a p4 ‡ p2 cycloadduct, and three possibilities can be
considered for its origin. In the first, (E)-5 in its planar form during ring inversion
undergoes an allowed concerted p4 ‡ p2 cycloaddition with TCNE; this of course can
occur only when the free energy for the formation of the planar form is sufficiently low.
The second possibility is outlined in Scheme 1. It involves the intermediacy of the
pentadienyl zwitterion 24 formed by initial attack of the reagent at C(6) followed by
charge annihilation with bond-formation at C(3) to give (E)-19. Under super-acid
conditions, cyloocta-1,3,5-triene (23) gives the pentadienyl cation 25 [23], while its
p4 ‡ p2 cycloaddition reactions with chlorosulfonyl isocyanate are believed to proceed
via the zwitterion 26 [24]. The remaining possibility (also shown in Scheme 1) for the
formation of (E)-19 involves initial reaction at C(3) to give the allylic zwitterion 27a or
the pentadienyl zwitterion 27b stabilized by the Ph group followed by bond-formation
as a result of charge annihilation at C(6). There would, however, be very considerable
strain associated with the planar portion of 27b. As well, conjugation of the dienyl
portion with the Ph group would be hindered because of a steric interaction with one of
the H-atoms at the methylene C-atom if formed from (E)-5 or at C(6) (as in (Z)-5) if
formed from (Z)-5. At this stage, it is tentatively suggested that (E)-19 is formed from
the pentadienyl zwitterion 24. It follows that the corresponding zwitterion is tentatively
suggested for the origin of (Z)-19.
Scheme 1
In the case of the p8 ‡ p2 adducts 2 and 3 formed from 1 [1] (see Introduction), it
was postulated that they arose via the intermediacy of the homotropylium zwitterion 4.
Similarly, it is suggested that the three corresponding adducts 16, endo-17 and exo-17

2012
Helvetica Chimica Acta Vol. 88 (2005)
from 5 and TCNE arise from the intermediacy of homotropylium zwitterions. Before
their nature is discussed, it is most relevant to take account of the classical studies of
Huisgen and co-workers [25] on the chlorination of cyclooctatetraene at low
temperatures. The reaction involves initial endo-attack (within the −tub× conformation
of the molecule as a result of the formation of a p-complex 28) by Cl2 to give the
homotropylium salt 29 followed by a second endo-attack by a ClÀ ion to give the cis-
dichloro compound cis-30 (Scheme 2). A rise in temperature gives the trans-dichloro
derivative trans-30 as a result of regeneration of the homotropylium salt 29 and ring
inversion of the homotropylium ion to give 31 followed by endo-attack by a ClÀ ion
(Scheme 2).
Scheme 2
Evidence that the parent homotropylium ion undergoes ring inversion is available
[26] and that it might be occurring with the zwitterions formed from (E)-5 and (Z)-5 is
presented below. Before considering whether the products 16, endo-17, and exo-
17 might be formed in endo,endo processes or involve one or two exo processes
(reactions occurring outside of the tub or ring) involving homotropylium zwitterions as
intermediates, it is pertinent to mention the following: 1) when (E)-5 which was
essentially free of its (Z)-isomer was treated with TCNE, 16 was formed in an
estimated yield of 14% compared with 11% when the (E)/(Z) ratio was 3 :2. It would
seem that the formation of 16 is dependent on the (E)/(Z) composition of 5 used in the
reactions. A method to prepare pure (Z)-5, or mixtures of (E)- and (Z)-5 with the
latter dominating, is needed, however, in order to confirm this preliminary finding; 2)
the ratio of the two diastereoisomers endo- and exo-17 is definitely dependent on the
ratio of (E)- and (Z)-5 used in the reactions. When the (E)(Z) ratios were 60 :40,
72 :28, and 88 :12, the ratios endo/exo-17 were 1.00 :0.72, 1.00 :1.16 and 1.00 :2.45,
respectively. Most importantly, however, a sample of (E)-5 in which (Z)-5 was barely
detectable (by 1H-NMR) gave with TCNE a ca. 1:7 mixture of endo- and exo-17. In this
case, the yield of endo-17 was estimated to be at least 3% which clearly shows that the
compound could have been derived only from (E)-5; 3) when 16 was heated at 1008 in
MeCN for 24 h, it was converted into a 1.00 :8.2 mixture of endo- and exo-17. The ratio
is not too different from that mentioned in 2) above.

Helvetica Chimica Acta Vol. 88 (2005)
2013
It is now appropriate to discuss the origin of 16, endo-17, and exo-17. That initial
formation of p-(charge-transfer) complexes with 5 and TCNE occurred was shown by
the strongly colored solutions formed when the TCNE was added to 5. In one scenario,
the zwitterions 32 and 33 can be formed as follows:if ( E)-5 were to react initially with
the complexed TCNE within the tub (endo-mode) of the ring as in 34, the zwitterion 32,
which is the same as that formed from (Z)-5 by exo-attack, would be generated
(Scheme 3). Ring inversion of 32 would yield 33 which is the same species as that
formed from (Z)-5 by endo-attack by TCNE (Note:the apparent change in
configuration at the benzylic C-atom in the side-chain in 33 is the result of a
conformational change caused by rotation of the bond between C(1) and C(1') in 32).
On the other hand, if the initial reaction of (E)-5 with TCNE were to occur as in 35 via
the exo-mode, the zwitterion 36, which is the same as that formed from (Z)-5 by endo-
attack, would be produced. Ring inversion of 36 would give 37 which is the same
species as that formed by exo-attack with TCNE on (Z)-5.
Scheme 3
After consideration of the various endo- and exo-modes of attack by the internal
nucleophile in each of the zwitterions 32, 33, 36, and 37 (Scheme 3), it can be concluded
that the latter two (as a result of initial exo-attack) with a slow equilibrium between
them accommodate the three findings 1) 3) mentioned above. Most important is the
slow equilibrium between 36 and 37; it accounts for the conversion of 16 into endo-17
and exo-17 on heating and for the formation of endo-17 from (E)-5 essentially free of

2014
Helvetica Chimica Acta Vol. 88 (2005)
(Z)-5. The origin of 16, endo-17, and exo-17 can be accommodated as follows: exo-
attack by the nucleophilic part of 36 at C(2) gives 16 and at C(7) gives exo-17, while the
corresponding reaction at C(7) in 37 yields endo-17.
The formation of (E)- and (Z)-14 can be readily accounted for:they are the result
of p4 ‡ p2 cycloadditions involving the valence tautomers (E)- and (Z)-13 of (E)- and
(Z)-5, respectively. In the valence tautomers, the diene portions are planar and the
cycloadditions are thermally allowed by the Woodward Hoffmann rules [19]. That
cycloocta-1,3,5-trienes can exist in equilibrium with bicyclic tautomers and that these
are the reactants with dienophiles is well established [27]. As already mentioned, the
bicyclic tautomers of (E)- and (Z)-5 could not be detected in the 1H-NMR spectrum of
the mixture of the two compounds. Their presence in the mixture must, therefore, be
very small but once formed they react readily with the p2-addend.
It is postulated that the two bis-adducts 15 and 18 are formed in three steps. Initially,
the mono-adducts 38 and 39 are formed as a result of a p2 ‡ p2 cycloaddition resulting
from charge annihilation by bond-formation at C(1) in one or more of the
homotropylium zwitterions 32, 33, 36, and 37 (see Scheme 3). These compounds 38
and 39, which have not been detected in the mixture formed from (E)- and (Z)-5 with
TCNE, exist in a valence-tautomeric equilibrium with the bicyclic forms 40 and 41
which undergo allowed p4 ‡ p2 cycloadditions with the p2-addend to give 15 and 18. It
should be noted that p2 ‡ p2 cycloadditions in homotropylium zwitterions involving
TCNE have been reported [28]. The formation of 38 can be ascribed to the attack by
the internal nucleophile in 36 at C(1) by the exo-mode. Formation of 39 is, however, at
variance with the exo-attack by the internal nucleophile postulated for the other
products discussed above involving the homotropylium zwitterions 36 and 37. It can be
formed by endo-attack in 37, the species formed by the exo-mode initially from (Z)-5. It
must be borne in mind, however, that the ratio 15/18 of ca. 2 :1 did not vary with
changes in the (E)/(Z) ratio in 5.
In the case of the reaction between (E)/(Z)-5 91:9 with 6, the number of adducts
isolated was four compared with nine when TCNE was the p2-addend. In discussing the
origin of exo- and endo-20 which are p8 ‡ p2 adducts, it is most important to note that
both were formed in significantly greater amounts (46 and 28%, resp.) than the amount
(9%) of (Z)-5 in the starting material; the two compounds must be derived mostly from
(E)-5. Their origin can be rationalized in the same way that was suggested for the
corresponding two adducts endo- and exo-17 from TCNE, namely via the equilibrating
zwitterions 42 and 43 (for the case of initial exo-attack by the p2-addend on (E)- and
(Z)-5; Scheme 4). Charge annihilation by bond-formation at C(7) in these species by
the exo-mode gives exo-20 and endo-20. In a preliminary study involving the reaction of
a (E)/(Z)-5 65 :35 with 6, the yield of exo-20 was found to be 45% (the same as for

Helvetica Chimica Acta Vol. 88 (2005)
2015
(E)/(Z)-5 91:9) implying that its formation is independent of the nature of the
composition of 5 and that it involves equilibrating homotropylium zwitterions. On the
other hand, the yield of endo-20 was reduced (from 28 to 17%) indicating that its
formation is dependent on the composition of 5. Clearly more work is required to
define this aspect more precisely.
Scheme 4
As in the case of (E)- and (Z)-19, (Z)- and (E)-21 are p4 ‡ p2 adducts, and their
origin cannot be explained with confidence. Tentatively, it is suggested that they are
derived from pentadienyl zwitterions corresponding to the species 24 suggested for the
corresponding product from (E)-5 and TCNE. The absence of adducts corresponding
to (E)- and (Z)-14 in the reaction with 6 can be attributed to the rapidity of the reaction
compared with that of TCNE; reaction of (E)- and (Z)-5 with 6 occurs before
meaningful quantities of the bicyclic valence tautomers (E)- and (Z)-13 to be trapped
by the dienophile can be formed. The absence of a cyclopropyl derivative correspond-
ing to 2 and 16 formed by bond-formation with charge annihilation at C(2) in 42 is
interesting. Perhaps it is formed but is rapidly isomerized into compounds exo- and
endo-20 under the conditions of the reaction. Bis-adducts corresponding to 15 and 18,
or of their precursor mono-adducts corresponding to 38 and 39 (or their tricyclic forms
40 and 41) were not detected in the mixture of products formed in the reaction of (E)/
(Z)-5 with 6.
To conclude, it is clear that access to pure (Z)-5, or at least mixtures with (E)-5
highly enriched in (Z)-5, is highly desirable in order that more-definitive conclusions
can be drawn about its cycloaddition reactions with typical p2-addends.
The authors wish to thank Prof. J. H. Bowie for helpful comments about mass spectra, Mr. T. Blumenthal for
recording mass spectra, and Dr. E. R. T. Tiekink for collaboration with X-ray crystallography. ACommonwealth
Postgraduate Award, a University of Adelaide Research Scholarship (to P. H. K.), and support of the project by
the former Australian Research Grants Committee are gratefully acknowledged.

2016
Helvetica Chimica Acta Vol. 88 (2005)
Experimental Part
1. General. All solvents were distilled before use. Petroleum ether refers to the fraction of b.p. 64 688. All
org. extracts were dried over anh. MgSO4. Anal. gas-liquid chromatography (GLC): Perkin-Elmer Sigma-3B
instrument; flame-ionization detector. Flash chromatography (FC): Merck silica gel 60; CC ˆ column
chromatography. M.p.: Kofler hot-stage apparatus, uncorrected. B.p.:uncorrected. UV Spectra:EtOH solns.,
unless stated otherwise; Unicam SP8-100 spectrophotometer; lmax (e) in nm. IR Spectra: Jasco IRA-1 grating
spectrometer for IR and ATI-Mattson Genesis-Series-FTIR spectrometer for FT-IR; films for liquids and nujol
mulls for solids; in cmÀ1. 1H- and 13C-NMR Spectra:CDCl 3 solns., unless stated otherwise, with Me4Si as internal
standard; Bruker WP-80 (20.1 MHz, 13C), Varian Gemini (200 MHz, 1H; 50.32 MHz, 13C), Varian Gemini-2000
(300 MHz, 1H; 75.47 MHz, 13C), Bruker ACP-300 (300 MHz, 1H; 75.47 MHz, 13C), and Varian INOVA
(600 MHz, 1H; 150.84 MHz, 13C) instruments; routine 1H-NMR at 200 and 300 MHz, spectra of adducts at
600 MHz, unless stated otherwise; multiplicities for 13C-NMR were determined at 20.1 MHz, if quoted. Mass
spectra: VG-ZAB-2 HF mass spectrometer; at 70 eV; in m/z (rel. %); in general, recording of m/z of abundance
greater than 25%. Elemental analyses were carried out by the Australian Microanalytical Service (Melbourne)
and the Chemistry Department, University of Otago (New Zealand). The accurate mass measurement was
carried out in the Chemistry Department, Monash University.
2. 1-Benzylcycloocta-1,3,5,7-tetraene (7). A soln. of benzyl chloride (4.92 g, 38.8 mmol) in dry THF (40 ml)
was added dropwise over 80 min to a stirred suspension of Mg turnings (2.10 g, 86.4 mmol) in THF (40 ml) (to
which a small crystal of I2 had been added) at r.t. under N2. After the mixture had been stirred at r.t. for a further
1.5 h, it was heated at 65 708 for another 1.5 h and kept overnight at r.t. The soln. was separated from
unchanged Mg by syringe and transferred to a 250-ml flask and cooled to À 158. Bromocyclooctatetraene [29]
(4.00 g, 21.9 mmol) in THF (35 ml) was added rapidly to the stirred soln. under N2, followed immediately by a
soln. of anh. FeCl3 (23 mg) in THF (1.0 ml) ! immediately deep-red soln. After the mixture had been stirred at
À 108 for 1.5 h, it was stirred at r.t. for a further 3.5 h, and then cooled (ice-water). Sat. NH4Cl soln. (50 ml) was
slowly added to the mixture, followed by petroleum ether (120 ml). The org. phase was washed with H2O (3 Â
100 ml), 20% aq. AgNO3 soln. (5 Â 30 ml), and H2O (3 Â 100 ml), dried, and evaporated. The residue was
distilled to give a mixture (2.68 g; b.p. 83 878/0.1 Torr), consisting of 7 (94%) and bibenzyl (6%) (1H-NMR).
The AgNO3 extract was treated with excess of 25% NH4OH soln., and the mixture was extracted with
petroleum ether (3 Â 75 ml). The combined org. extracts were washed with H2O (4 Â 75 ml), dried, and
evaporated to give pure (1H-NMR) 7 (1.22 g). Bulb-to-bulb distillation at 90 958 (block)/0.08 Torr gave 7
(1.16 g) whose 1H-NMR and IR data were identical to reported data [2]. Overall, the yields (including the
product extracted with AgNO3) of 7 and bibenzyl were 88% and 4%, resp.
Note:The use of excess of the Grignard reagent in the reaction avoided the formation of bi(cyclooctate-
traenyl) which caused problems during purification of the desired product when an equiv. of the reagent was
used.
3. Cycloocta-2,4,6-trien-1-yltriphenylphosphonium Bromide (10). A ca. 7:3 mixture (by 1H-NMR; see [30]
where a ratio of 65 :35 has been reported) of 7-bromocycloocta-1,3,5-triene (11) and its valence tautomer 12
(2.82 g, 15.2 mmol), prepared by the method of Krˆner [6], in benzene (5 ml) was added dropwise to a stirred
soln. of PPh3 (3.72 g, 14.2 mmol) in benzene (15 ml) cooled to 38. After the mixture had been stirred at 3 58 for
3.5 h, it was filtered under N2, and the solid material was washed with benzene (2 Â 10 ml) and dry Et2O (6 Â
20 ml) and −dried× at r.t./0.01 Torr for 40 min: 10 (4.39 g, ca. 91%, based on estimated 11 in the starting material)
as a highly electrostatic powder. On exposure to air, 10 was rapidly converted into a sticky gum. Even at À 188
under N2, the lifetime of 10 was only a few days. 1H-NMR (200 MHz):2.67 ( q, J ˆ 7.0, 2 HÀC(8)); 5.45 5.57 (m,
HÀC(1)); 5.68 5.84 (m, 2 olef. H):5.85 6.46 ( m, 3 olef. H); 6.30 6.46 (m, 1 olef. H); 7.60 7.85 (m, 10 arom.
H); 7.85 8.06 (m, 5 arom. H).
4. (7E)/(7Z)-7-Benzylidenecycloocta-1,3,5-triene ((E)/(Z)-5). 4.1. In preliminary experiments, it was
shown by GLC (15% SE-30 on Varaport 30) that an equilibrium (95 :5) between 5 and starting material was
established readily when 7 was treated with t-BuOK in THF. A stirred soln. of 7 (220 mg, 1.13 mmol) in dry THF
(8.0 ml) at 08 under N2 was treated with finely powdered t-BuOK (64 mg, 0.57 mmol). After 5 min, the cooling
bath was removed and the mixture was stirred for a further 5.5 h. The deep-brown mixture was diluted with
petroleum ether (25 ml), washed with brine (1 Â 15 ml), H2O (3 Â 20 ml), 20% aq. AgNO3 soln. (2 Â 10 ml),
and H2O (3 Â 20 ml), dried, and evaporated:( E)/(Z)-5 (185 mg, 84%). Yellow liquid. B.p. 1058 (block)/
0.03 Torr. UV:209 (1.80 ¥ 10 4), 252 (1.57 ¥ 104), 306 (1.20 ¥ 104). IR:3099 s, 1602m, 753m, 711s. 1H-NMR
(300 MHz):3.30 ( d, J(1,8) ˆ 8.1, HÀC(8) (Z)); 3.40 (d, J (1,8) ˆ 8.1, HÀC(8) (E)); 5.75 6.18 (m, 5 olef. H);
6.23 (s, HÀCˆC(7) (Z)); 6.48 (d, J(6,5) ˆ 12.9, HÀC(6) (E)); 6.57 (s, HÀCˆC(7) (E)); 6.85 (d, J (6,5) ˆ 12.9,

Helvetica Chimica Acta Vol. 88 (2005)
2017
HÀC(6) (Z)); 7.20 7.40 (m, 5 arom. H); (E)/(Z) ratio was 3 :2. 13C-NMR (50.32 MHz):29.39 (C(8) ( E)); 37.76
(C(8) (Z)); 124.71; 126.03; 126.82; 126.84; 127.61; 127.86; 127.91; 128.10; 128.35; 128.52; 128.61; 128.42; 128.92;
.
‡
129.65; 130.81; 130.87; 132.51; 136.56; 137.85; 137.94; 137.98; 138.82; 139.27; 139.62. MS:195 (13, [ M ‡ H] ),
.
‡
194 (100, M ), 179 (56), 178 (61), 165 (56), 152 (25), 128 (26), 117 (26), 116 (43), 115 (81), 105 (58), 91 (40), 77
(25), 57 (35), 51 (38), 41 (70), 39 (69). Anal. calc. for C15H14 (194.26):C 92.74, H 7.26; found:C 92.44, H 7.12.
The mixtures having (E)/(Z) ratios of 73 :27 and 65 :35 mentioned in the General Part were obtained
during preliminary investigations when the amounts of t-BuOK used, and the reaction times, were smaller than
those given above.
4.2. A stirred suspension of 10 (1.802 g, 4.03 mmol) in dry Et2O (45 ml) under N2 at 08 was treated dropwise
with 2.24m BuLi in petroleum ether (1.8 ml, 4.03 mmol) (!deep maroon color). After the mixture had been
stirred at 08 for 45 min and then at r.t. for 15 min, it was cooled to 08 and treated dropwise with a soln. of
benzaldehyde (428 mg, 4.03 mmol) in Et2O (5 ml). The mixture was stirred at r.t. for 3 h and then filtered. The
filtrate was washed with H2O (4 Â 40 ml), dried and evaporated, and the yellow residue subjected to CC
(Sorbsil, (16 g), petroleum ether):( E)/(Z)-5 88 :12 (by 1H-NMR; 314 mg, 40%). Yellow liquid.
In several preparations by this method, the (E)/(Z) ratio ranged from 91 :9 to 84 :16.
5. Reaction of 5 with Ethenetetracarbonitrile (TCNE). 5.1 Isolation of the Products. A stirred soln. of (E)/
(Z)-5 88 :12 (377 mg, 1.94 mmol) and TCNE (448 mg, 3.49 mmol) in dry AcOEt (8 ml) under N2 was heated at
80 838 for 5 h (initially deep blue ! finally golden-yellow). The cooled soln., after the addition of AcOEt
(30 ml), was washed with H2O containing some brine (20 ml), 20% aq. sodium metabisulfite (3 Â 35 ml), and
brine (2 x 20 ml) and dried. TLC (silica gel, AcOEt/petroleum ether 3 :7, UV detection):three distinct and two
overlapping product −spots×. The soln. was evaporated to give a greenish-brown glass (695 mg). A soln. of the
glass in hot AcOEt (3 ml), when diluted with hot petroleum ether (7 ml), gave on cooling (3E)-3-
benzylidenetricyclo[4.2.2.02,5]dec-7-ene-9,9,10,10-tetracarbonitrile (ˆ((3E)-benzylidenetricyclo[4.2.2.02,5]dec-9-
ene-7,7,8,8-tetracarbonitrile; (E)-14; 47 mg). Small, hard, colorless crystals. M.p. 270 2718 (AcOEt/petroleum
ether) with browning. 1H-NMR:2.58 2.64 ( m, HaÀC(4)); 3.25 3.28 (m, HÀC(5)) overlapping 3.26 (ddd,
J(4b,4a) ˆ 16.8, J(4b,5) ˆ 9.6, J(4b,HÀCˆC(3)) ˆ 3.0, HbÀC(4)); 3.71 (ddd, J(6,7) ˆ 6.6, J(6,5) ˆ 3.0, J(6,8) ˆ
1.2, HÀC(6)); 3.78 (ddd, J(1,8) ˆ 6.6, J(1,2) ˆ 4.2, J(1,7) ˆ 1.2, HÀC(1); 3.80 3.84 (m, HÀC(2)); 6.23 (q,
J(4a,HÀCˆC(3)) ˆ J(4b,HÀCˆC(3)) ˆ J(2,HÀCˆC(3)) ˆ 2.4, HÀCˆC(3)); 6.65 (ddt, J(7,8) ˆ 8.4, J(7,6) ˆ
6.6, J(7,5) ˆ J(7,1) ˆ 1.2, HÀC(7)); 6.73 (ddt, J(8,7) ˆ 8.4, J(8,1) ˆ 7.2, J(8,2) ˆ J(8,6) ˆ 1.2, HÀC(8)); 7.10 7.12
(m, 2 Ho); 7.23 (tt, J ˆ 7.2,1.2, Hp); 7.30 7.33 (m, 2 Hm). 13C-NMR (75.47 MHz):27.57; 35.03; 40.86; 41.54; 42.19;
42.37; 42.50; 110.97, 111.05, 111.70 and 111.81 (4 CN); 126.57; 127.68; 127.75; 128.91; 130.31; 133.86; 135.99;
‡¥
136.24. MS:322 (5, M ), 194 (100), 193 (36), 179 (43), 165 (26), 116 (52), 115 (48), 101 (34), 81 (50), 77 (34),
57 (48), 55 (56), 51 (34), 41 (85), 39 (35). Anal. calc. for C21H14N4 (322.35):C 78.24, H 4.38, N 17.38; found:C
77.98, H 4.26, N 17.52.
The residue from the initial mother liquor in hot AcOEt (1.5 ml), on dilution with hot petroleum ether
(3.5 ml), yielded on cooling two crops of spiro[2'-(RS)-phenylcyclobutane-1',3-(1RS,2SR,3RS,5SR,6SR)-
tetracyclo[4.2.2.02,5]dec-7-ene-3',3',4',4',9,9,10,10-octacarbonitrile] (ˆ(1'RS,2'SR,3'RS,4RS,5'SR,6'SR)-4-phenyl-
spiro[cyclobutane-1,3'-tetracyclo[4.2.2.02,5]dec[9]ene]-2,2,3,3,7',7',8',8'-octacarbonitrile; 15, 37 mg). Small, hard,
colorless crystals. M.p. 2508 (dec., AcOEt/petroleum ether). 1H-NMR ((D6)acetone):2.52 ( ddd, J(4a,4b) ˆ 15.0,
J(4a,5) ˆ 5.8, J(4a,2) ˆ 1.3, HaÀC(4)); 2.93 (ddd, J(4b,4a) ˆ 15.0, J(4b,5) ˆ 9.0, J(4b,2) ˆ 1.3, HbÀC(4)); 3.24
(ddddd, J(5,4b) ˆ 9.0, J(5,2) ˆ 8.3, J(5,4a) ˆ 5.8, J(5,6) ˆ 3.6, J(5,7) ˆ 1.2, HÀC(5)); 3.91 (dddt, J(2,5) ˆ 8.3,
J(2,1) ˆ 3.8, J(2,4a) ˆ J(2,4b) ˆ 1.3, J(2,8) ˆ 0.9, HÀC(2)); 4.31 (ddd, J(6,7) ˆ 6.7, J(6,5) ˆ 3.6, J(6,8) ˆ 0.9,
HÀC(6)); 4.65 (ddd, J(1,8) ˆ 6.5, J(1,2) ˆ 3.8, J(1,7) ˆ 1.2, HÀC(1)); 5.45 (s, HÀC(2')); 7.21 (ddt, J(7,8) ˆ 8.1,
J(7,6) ˆ 6.7, J(7,1) ˆ J(7,5) ˆ 1.2, HÀC(7)); 7.32 (ddt, J(8,7) ˆ 8.1, J(8,1) ˆ 6.5, J(8,2) ˆ J(8,6) ˆ 0.8, HÀC(8));
7.59 7.61 (m, Hp); 7.64 7.66 (m, 2 Hm); 7.72 7.74 (m, 2 Ho). 13C-NMR (150 MHz, (D6)acetone):28.48 (C(5));
30.18 (C(4)); 40.04 (C(1)); 41.04 (C(2)); 42.02 (C(6)); 42.43 (C(9)); 42.96 (C(10)); 43.66 (C(3)); 44.19 (C(3'));
54.16 (C(2')); 55.45 C(4')); 110.23, 111.10, 111.59, 111.62, 111.95, 112.16, 112.91, 113.01 (8 CN); 128.11 (arom);
‡¥
130.63 (arom); 130.70 (arom); 131.46 (arom); 133.62 (C(8)); 139.50 (C(7)). MS:322 (2, [ M-TCNE] ), 194
(100), 193 (28), 179 (30), 116 (34), 115 (33), 96 (25), 91 (23), 83 (28), 71 (33), 69 (31), 57 (61), 55 (43), 44 (62),
41 (46), 39 (26). Anal. calc. for C27H14N8 (450.45):C 71.99, H 3.13. N 24.88; found:C 71.70, H 3.16, N 24.85.
The residue from the mother liquor from which 15 was obtained was purified by FC (silica gel (144 g),
AcOEt/petroleum ether 3 :7), 8 to10-ml fractions, TLC monitoring as above. Fractions showing two −spots× were
combined and rechromatographed as above on fresh silica gel (144 g). Fractions showing a single −spot× from the
two chromatographic separations were combined to give four fractions, Fr. A (4 mg), B (140 mg), C (261 mg),
and D (115 mg). Elution of the first column with AcOEt gave Fr. E (27 mg), an amorphous, brown material.
Fr. A: 1H-NMR showed that in part it consisted of recovered (E)/(Z)-5 59 :41.

2018
Helvetica Chimica Acta Vol. 88 (2005)
Fraction B: Extraction of the solid of Fr. B with warm Et2O (4 Â 5 ml) gave an insoluble portion (43 mg)
which, from acetone/petroleum ether, gave two crops of 15 (25 mg; total yield 62 mg, 8%). The residue from the
mother liquor from which 15 was obtained yielded, from AcOEt/petroleum ether, additional (E)-14 (4 mg; total
yield 51 mg, 8%). The Et2O-soluble portion (92 mg) was crystallized from Et2O, followed by Et2O/petroleum
ether, to give four crops of (1RS,3RS,8SR,11SR)-11-phenyltricyclo[6.3.0.01,3]undeca-4,6-diene-9,9,10,10-tetra-
carbonitrile (16; 67 mg, 11%). M.p. 184 1868, raised to 185 1878 (AcOEt/petroleum ether). UV:257 (4.97 ¥
103), 208 (1.60 ¥ 104). 1H-NMR:0.90 0.96 ( m, 2 HÀC(2)); 1.44 1.48 (m, HÀC(3)); 3.65 3.67 (m, HÀC(8));
4.11 (s, HÀC(11)); 5.92 (dd, J ˆ 11.4, J ˆ 6.0, HÀC(5) or HÀC(6)); 6.27 6.36 (m, 3 olef. H); 7.37 7.40 (m, 2
arom. H); 7.42 7.48 (m, 3 arom. H). 13C-NMR (75.47 MHz):20.73 ( d, C(3)); 22.94 (t, C(2)); 37.03 (s, C(1));
47.97, 52.04 (C(9), C(10)); 55.90 (d, C(8)); 60.82 (d, C(11)); 109.22, 109.55, 111.10, 111.61 (4 CN); 125.89 (d);
.
‡
126.74 (d); 129.65 (d); 130.25; 130.59 (d); 131.40 (d); 131.73; 136.19 (d). MS:322 (30, M ), 194 (100), 193 (33),
179 (38), 178 (26), 165 (25), 116 (46), 115 (43), 103 (31), 91 (21), 78 (34), 77 (37), 76 (28), 51 (55), 50 (34), 39
(62). Anal. calc. for C21H14N4 (322.35):C 78.24, H 4.38, N 17.38; found C 77.85, H 4.61, N 17.67.
Fraction C: Extraction of the solid of Fr. C with warm Et2O (2 Â 5 ml) gave an insoluble, white powder
(Fr. C1, 31 mg; see below). The residue from the soluble portion was recrystallized from Et2O/petroleum ether
1 :1 (10 ml) to give (7RS,10SR)-10-phenylbicyclo[5.3.1]deca-1,3,5-triene-8,8,9,9-tetracarbonitrile (endo-17;
45 mg). M.p. 221.5 222.58 (after sublimation at 1008/0.1 Torr). UV:260 (2.31 ¥ 10 3), 211 (2.59 ¥ 104).
1H-NMR:2.46 ( d with fine splitting, J(11a,11b) ˆ 13.8, HaÀC(11)); 3.34 (dd, J(11b,11a) ˆ 13.8, J(11b,7) ˆ 2.4,
HbÀC(11)); 3.88 (br. s, HÀC(7)); 4.25 (s, HÀC(10)); 6.11 (ddd, J ˆ 11.4, 7.8, 1.2, olef. H); overlapping 6.12 6.14
(m, 4 olef. H); 7.48 7.52 (m, 3 arom. H); 7.58 À 7.60 (m, 2 arom. H). 13C-NMR (75.47 MHz):30.64 ( t, C(11));
46.96; 47.86 (d, C(7)); 56.06 (d, C(10)); 110.24, 110.43, 111.18, 111.28 (4 CN); 126.25 (d); 128.33 (d); 128.68 (d);
.
‡
129.47 (d); 129.69 (d); 130.11 (d); 130.44 (d); 131.14; 131.52 (d); 136.39 (s). MS:322 (4, M ), 194 (100), 179
(33), 116 (15), 115 (15), 77 (8), 51 (11), 39 (13). Anal. calc. for C21H14N4 (322.35):C 78.24, H 4.38; found:C
78.30, H 4.21.
The residue from the mother liquor from which endo-17 had been obtained was crystallized from Et2O/
petroleum ether 2 :1 (15 ml) to give (7RS,10RS)-10-phenylbicyclo[5.3.1]undeca-1,3,5-triene-8,8,9,9-tetracarbo-
nitrile (exo-17; 68 mg). M.p. 155 1578 (AcOEt/petroleum ether). 1H-NMR:2.93 ( dd, J(11b,11a) ˆ 13.8,
J(11b,7) ˆ 7.2, HbÀC(11)); 3.07 (d, J(11a,11b) ˆ 13.8, HaÀC(11)); 3.96 (br. s, HÀC(7)); 4.70 (s, HÀC(10)); 6.09
(dd, J(3,4) ˆ 12.6, J(3,2) ˆ 7.8, HÀC(3)); 6.14 6.21 (m, HÀC(4), and HÀC(6)); 6.38 (br. s, HÀC(2)); 6.42
6.49 (m, HÀC(5)); 7.49 (s, 5 arom. H); assignments by COSY and ROESY. 13C-NMR (50.4 MHz):24.87 ( t,
C(11)); 45.69 (C(8), C(9)); 46.87 (d, C(7)); 55.07 (d, C(10)); 109.50, 110.20, 111.79, 112.34 (4 CN); 125.73;
.
‡
126.70; 127.69; 128.56; 129.75; 129.93; 130.29; 134.12. MS:322 (6, M ), 194 (100), 193 (29), 179 (33), 116 (36),
115 (33), 103 (15), 76 (12), 51 (16), 39 (11). Anal. calc. for C21H14N4 (322.35):C 78.24, H 4.38, N 17.38; found C
78.42, H 4.43, N 17.64.
The residue from the mother liquor from which exo-17 had been obtained was crystallized from Et2O/
petroleum ether 3 :1 (10 ml) to give additional exo-17 (35 mg). Crystallization of the residue from the mother
liquor from AcOEt/petroleum ether 2 :5 (7 ml) gave additional endo-17 (5 mg; total yield 50 mg, 8%). A soln.
of the residue from the mother liquor in AcOEt (ca. 0.5 ml) on dilution with petroleum ether gave two crops of
additional exo-17 (21 mg; total yield 124 mg, 20%). The residue from the mother liquor in acetone/petroleum
ether at À 188 yielded pale-brown crystals of (3Z)-3-benzylidenetricyclo[4.2.2.02,5]dec-7-ene-9,9,10,10-tetracar-
bonitrile ( ˆ (3Z)-3-benzylidenetricyclo[4.2.2.02,5]dec-9-ene-7,7,8,8-tetracarbonitrile; (Z)-14; 4 mg, 0.6%) which
gave colorless crystals (2 mg). M.p. 167 1698 (AcOEt/petroleum ether). 1H-NMR:2.38 ( dddd, J(4a,4b) ˆ 17.4,
J(4a,5) ˆ 3.0, J(4a,HÀCˆC(3)) ˆ 2.4, J(4a,2) ˆ 1.2, HaÀC(4)); 3.10 (ddd, J(4b,4a) ˆ 17.4, J(4b,5) ˆ 9.0, J(4b,
HÀCˆC(3)) ˆ 2.4, HbÀC(4)); 3.20 (dddd, J(5,4b) ˆ 9.0, J(5,6) ˆ 3.6, J(5,4a) ˆ 3.0, J(5,7) ˆ 1.2, HÀC(5)); 3.68
(ddd, J(1,8) ˆ 7.2, J(1,2) ˆ 4.2, J (1,7) ˆ 1.2, HÀC(1)); 3.77 (ddd, J(6,7) ˆ 6.6, J(6,5) ˆ 3.6, J(6,8) ˆ 1.2, HÀC(6));
4.12 (ddt, J(2,1) ˆ 4.2, J(2, HÀCˆC(3)) ˆ 2.4, J(2,4a) ˆ J(2,8) ˆ 1.2, HÀC(2)); 6.15 (q, J(HÀCˆC(3),2) ˆ
J(HÀCˆC(3),4a) ˆ J(HÀCˆC(3),4b) ˆ 2.4, HÀCˆC(3)); 6.47 (ddt, J(7,8) ˆ 8.4, J(7,6) ˆ 6.6, J(7,5) ˆ
J(7,1) ˆ 1.2, HÀC(7)); 6.63 (ddt, J(8,7) ˆ 8.4, J(8,1) ˆ 7.2, J(8,6) ˆ J(8,2) ˆ 1.2, HÀC(8)); 7.18 7.19 (m, 2
Ho); 7.28 7.30 (m, Hp); 7.38 7.41 (m, 2 Hm). 13C-NMR (150.84 MHz):26.97; 31.11; 34.31; 40.06; 40.64; 41.46;
42.19; 111.02, 111.10, 111.62, 111.79 (4 CN); 127.30; 127.55; 127.85; 129.31; 130.08; 133.52; 135.72; 135.92. MS:
.
‡
322 (48, M ), 194 (29), 179 (27), 142 (42), 141 (34), 116 (90), 115 (100), 91 (12), 78 (12), 69 (12), 51 (14), 41
‡
‡
(16), 39 (21). HR-ESI MS:345.1108 ([ M ‡ Na] ); C21H14N4Na ; calc. 345.1116)
1H-NMR ((D6)acetone) of the insoluble Fr. C1 (see above) indicated that it was essentially pure spiro[2'-
(SR)-phenylcyclobutane-1',3-(1RS,2SR,3SR,5SR,6SR)-tetracyclo[4.2.2.02,5]dec-7-ene-3',3',4',4',9,9,10,10-octa-
carbonitrile] ( ˆ (1'RS,2'SR,3'SR,4SR,5'SR,6'SR)-4-phenylspiro[cyclobutane-1,3'-tetracyclo[4.2.2.02,5]dec-
[9]ene]-2,2,3,3,7',7',8',8'-octacarbonitrile; 18; 4%). Recrystallization from AcOEt/petroleum ether gave clusters

Helvetica Chimica Acta Vol. 88 (2005)
2019
of soft radiating needles (15 mg). M.p. 197 2028 (unchanged on further crystallization). 1H-NMR
((D6)acetone):2.67 ( ddd, J(4a,4b) ˆ 15.1, J(4a,5) ˆ 6.1, J(4a,2) ˆ 1.6, HaÀC(4)); 2.98 (ddd, J(4b,4a) ˆ 15.1,
J(4b,5) ˆ 9.0, J(4b,2) ˆ 2.1, HbÀC(4)); 3.54 (ddddd, J(5,4b) ˆ 9.0, J(5,2) ˆ 8.6, J(5,4a) ˆ 6.1; J(5,6) ˆ 3.9,
J(5,7) ˆ 1.3, HÀC(5)); 3.87 (ddddd, J(2,5) ˆ 8.6, J(2,1) ˆ 3.4, J(2,4b) ˆ 2.1, J(2,4a) ˆ 1.6, J(2,8) ˆ 0.8, HÀC(2));
4.04 (ddd, J(6,7) ˆ 6.8, J(6,5) ˆ 3.9, J(6,8) ˆ 1.3, HÀC(6)); 4.62 (ddd, J(1,8) ˆ 6.5, J(1,2) ˆ 3.4, J(1,7) ˆ 1.3,
HÀC(1)); 4.99 (s, HÀC(2')); 6.31 (dddd, J(8,7) ˆ 8.3, J(8,1) ˆ 6.5, J(8,6) ˆ 1.3, J(8,2) ˆ 0.8, HÀC(8)); 6.41 (ddt,
J(7,8) ˆ 8.3, J(7,6) ˆ 6.8, J(7,5) ˆ J(7,5) ˆ 1.3, HÀC(7)); 7.60 7.63 (m, Ho); 7.64 7.67 (m, 2 Hm, Hp). 13C-NMR
(600 MHz, (D6)acetone):27.59 (C(5)); 30.28 (C(4)); 40.53 (C(1)); 41.25 (C(2)); 42.45 (C(6)); 43.12 (C(9));
43.40 (C(10)); 48.08 (C(3)); 52.80 (C(3')); 52.83 (C(2')); 53.60 C(4')); 110.38, 110.90 (2C), 112.09, 112.13, 112.24,
113.06, 113.09 (8 CN); 130.20 (arom. C); 131.34 (arom. C); 131.36 (arom. C); 131.52 (arom. C); 133.89 (C(8));
‡
135.05 (C(7)). MS:322 (24, [ M À TCNE] ), 194 (73), 193 (23), 189 (29), 154 (35), 128 (65), 116 (38), 115 (42),
103 (23), 78 (32), 77 (27), 76 (58), 71 (60), 69 (42), 68 (53), 57 (100), 41 (69), 39 (33). Anal. calc. for C27H14N8
(450.45):C 71.99, H 3.13, N 24.88; found:C 72.22, H 3.46, N 24.28.
Fraction D: Recystallization of the crude product from AcOEt/ petroleum ether gave (5E)-5-benzyl-
idenebicyclo[4.2.2]deca-2,7-diene-9,9,10,10-tetracarbonitrile (ˆ(5E)-5-benzylidenebicyclo[4.2.2]deca-2,9-diene-
7,7,8,8-tetracarbonitrile; (E)-19; 91 mg, 15%). M.p. 170 1718. 1H-NMR (300 MHz):2.97 ( dquint., J(4a,4b) ˆ
15.3, −J× ˆ 2.4, HaÀC(4)); 3.29 (dd, J(4b,4a) ˆ 15.3, J(4b,3) ˆ 9.9, HbÀC(4)); 3.96 (t, J(1,2) ˆ 7.2, HÀC(1)); 4.22
(d, J(6,7) ˆ 6.4, HÀC(6)); 5.83 (ddd, J(2,3) ˆ 11.6, J(2,1) ˆ 7.2, J(2,4a) ˆ 2.9, HÀC(2)); 6.27 (ddd, J(3,2) ˆ 11.6,
J(3,4b) ˆ 9.9, J(3,4a) ˆ 4.5, HÀC(3)); 6.41 6.56 (overlapping m, HÀC(7), HÀC(8)); 6.92 (d (poorly
resolved), J(HÀCˆC(5),4b) ꢀ 1.6, HÀCˆC(5)); 7.24 7.30 (m, Ho); 7.32 7.43 (m, 2 Hm, Hp); the quint.
within d 2.97 was shown from double irradiation experiments to arise from J(4b,3) ˆ 4.5, J(4b,2) ˆ 2.9,
J(4b,HÀCˆC(5)) ˆ 1.6 Hz. 13C-NMR (50.32 MHz):26.01 ( t, C(4)); 42.45 (d, C(1)); 45.02, 45.63 (C(9), C(10));
51.98 (d, C(6)); 110.81, 110.88, 111.96, 113.41 (4 CN); 120.90 (d); 127.46 (d); 128.19 (d); 128.49; 128.61; 129.32;
.
‡
135.64; 136.94 (d); 139.27 (d). MS:322 (20, M ), 194 (100), 193 (28), 179 (35), 116 (52), 115 (60), 91 (35), 78
(25), 77 (41), 69 (38), 65 (25), 63 (25), 57 (34), 55 (33), 51 (51), 41 (67), 39 (80). Anal. calc. for C21H14N4
(322.35):C 78.24, H 4.38; found:C 78.13, H 4.40.
Fraction E: The complex mixture (1H-NMR) was extracted with warm Et2O (2 Â 3 ml). The Et2O soluble
material (17 mg), a brown powder, was then extracted with CHCl3 (2 Â 3 ml), and the soluble portion (14 mg),
on crystallization from acetone/petroleum ether, gave colorless crystals of (5Z)-5-benzylidenebicyclo[4.2.2]-
deca-2,7-diene-9,9,10,10-tetracarbonitrile ( ˆ (5Z)-5-benzylidene[4.2.2]deca-2,9-diene-7,7,8,8-tetracarbonitrile;
(Z)-19; 2 mg, 0.3%). M.p. 214 2168 (with browning). 1H-NMR:2.96 ( dd, J(4b,4) ˆ 15.6, J (4b,3) ˆ 10.2,
HbÀC(4)); 3.50 (dddd, J(4a,4b) ˆ 15.6, J(4a,3) ˆ 4.2, J(4a,2) ˆ 3.0, J(4a,HÀCˆC(5)) ˆ 2.4, HaÀC(4)); 3.92 (t,
J(1,2) ˆ J(1.8) ˆ 7.2, HÀC(1)); 4.41 (d, J(6,7) ˆ 6.6, HÀC(6)); 5.80 (ddd, J(2,3) ˆ 12.0, J(2,1) ˆ 7.2, J(2,4a) ˆ 3.0,
HÀC(2)); 6.22 (dd, J(7,8) ˆ 9.6, J(7,6) ˆ 6.6, HÀC(7)); 6.36 (dddd, J(3,2) ˆ 12.0, J(3,4b) ˆ 10.2, J(3,4a) ˆ 4.2,
J(3,1) ˆ 0.6, HÀC(3)); 6.48 (dd, J(8,7) ˆ 9.6, J(8,1) 7.2, HÀC(8)); 7.04 (d, J(HÀCˆC(5),4a) ˆ 2.4,
HÀCˆC(5)); 7.26 7.28 (m, 2 Ho); 7.38 7.39 (m, Hp); 7.45 7.48 (m, 2 Hm)). 13C-NMR (75.47):33.33 (C(4));
42.74 (C(1)); 43.91 (C(9) or C(10)); 44.82 (C(6)); 45.67 (C(9) or C(10)); 110.99, 111.53, 111.85, 112.5 (4 CN);
.
‡
120.06; 127.54; 128.42; 128.57; 129.31; 129.68; 135.39; 136.33; 139.07; 139.77. MS:322 (25, M ), 194 (100), 193
(27), 116 (100), 115 (31), 91 (16), 77 (12), 51 (12), 39 (13).
5.2. Composition of the Product Mixtures of Other Reactions of 5 with TCNE. (7E)-7-Benzylidenecy-
cloocta-1,3,5-triene ((E)-5), containing a trace (1H-NMR) of the (Z)-isomer, was recovered (ca. 20%) after CC
(silica gel) when (E)/(Z)-5 ca. 9 :1 was treated with 0.8 equiv. of 6 in AcOEt. Mixtures enriched in (Z)-5 were
obtained from treatment of 7 with t-BuOK in THF. The various mixtures were each treated with TCNE as
1
described in Exper. 5.1 above, and the crude products (after H-NMR ((D6)acetone/CDCl3) determination of
the ratios of some of the products) were separated by FC (silica gel) as above. The composition of each of the
1
appropriate weighed fractions was determined by H-NMR, and the yields were calculated.
5.3. Conversion of 16 into endo-17 and exo-17. A soln. of 16 (3.2 mg) in MeCN (0.3 ml) was heated under N2
in a sealed tube at 1008 for 24 h. The solvent was evaporated, and 1H-NMR showed that the product consisted of
endo- and exo-17 in the ratio of 1.00 :8.2 (mean of two determinations). When endo- and exo-17 were each
subjected to the same conditions, they were recovered unchanged (1H-NMR).
6. Reaction of 5 with PTAD 6. A stirred soln. of (E)/(Z)-5 91:9 (247 mg, 1.27 mmol) in AcOEt (5 ml)
under N2 was treated dropwise at r.t. with a soln. of 6 (223 mg, 1.27 mmol) in AcOEt (4 ml). After the color of
the reagent had been discharged (80 min), the resulting pale yellow soln. was evaporated and the residue
subjected to CC (silica gel (148 g), AcOEt/petroleum ether 1:1, 5- to 10-ml fractions) with TLC (silica gel,
AcOEt/petroleum ether 2 :3, UV detection) monitoring. Fractions giving two −spots× were rechromatographed
to give fractions giving only one; these were combined with the relevant fractions obtained initially. In order of

2020
Helvetica Chimica Acta Vol. 88 (2005)
elution, the following compounds were obtained:unchanged ( E)-5 (18 mg) free of (Z)-5 (1H-NMR), exo-20
(199 mg, 46%), endo-20 (121 mg, 28%), (Z)-21 (16 mg, 4%; impure), and (E)-21 (68 mg, 16%).
Data of (2RS,8RS)-2,5-Diphenyl-3,5,7-triazatricyclo[6.5.1.03,7]tetradeca-9,11,13-triene-4,6-dione (exo-20);
Colorless crystals. M.p. 169 1708 (AcOEt/petroleum ether). FT-IR:1763 s, 1707vs, 1597w, 1403s, 1284m, 1153w,
764m, 1128w, 764m. 1H-NMR:2.23 ( ddd, J(14b,14a) ˆ 13.0, J(14b,8) ˆ 3.4, J(14b,13) ˆ 1.6, HbÀC(14)); 2.86
(ddd, J(14a,14b) ˆ 13.0, J(14a,8) ˆ 4.2, J(14a,13) ˆ 1.2, HaÀC(14)); 5.26 (br. s, HÀC(8)); 6.05 (s, HÀC(2))
overlapping 5.96 6.10 (m, HÀC(9), HÀC(10), HÀC(11)); 6.25 (dd, J(12,11) ˆ 10.5, J(12,13) ˆ 3.1, HÀC(12));
6.54 (d, J(13,12) ˆ 3.1, HÀC(13)); 7.18 7.63 (m, 10 arom. H); double-irradiation experiments (at d 5.26 and
6.50) and COSY confirmed the values and attributions of the J. 13C-NMR (50.32 MHz):25.81 (C(14)); 57.48
(C(8)); 62.77 (C(2)); 125.68; 126.03; 128.00; 128.19; 128.77; 128.95; 129.15; 129.30; 131.88; 136.43; 137.25;
.
‡
149.39, 151.42 (2 CˆO). MS:369 (63, M ), 256 (28), 225 (30), 206 (25), 197 (31), 157 (32), 115 (30), 101 (40),
91 (38), 81 (32), 69 (81), 42 (100), 40 (64). Anal. calc. for C23H19N3O2 (369.41):C 74.78, H 5.18; found:C 74.42,
H 5.04.
Data of (2RS,8SR)-2,5-Diphenyl-3,5,7-triazatricyclo[6.5.1.03.7]tetradeca-9,11,13-triene-4,6-dione (endo-20).
Colorless crystals. M.p. 205 2078 (AcOEt/petroleum ether). FT-IR:1759 m, 1699vs, 1597w, 1491m, 1415m,
1396w, 1132m, 768m. 1H-NMR:2.65 ( dd, J(14b,14a) ˆ 12.6, J(14b,8 ) ˆ 5.4, HbÀC(14)); 3.17 (dd, J(14a,14b) ˆ
12.6, J(14a,8) ˆ 1.2, HaÀC(14)); 5.25 (ddt, J(8,14b) ˆ 5.4, J(8,9) ˆ 4.1, J(8,14a) ˆ J(8,10) ˆ 1.2, HÀC(8)); 5.50 (t,
J(13,12) ˆ J(13,2) ˆ 3.0, HÀC(13)); 5.79 (t, J(2,9) ˆ J(2,13) ˆ 3.0, HÀC(2)); 5.95 (dd, J(11,12) ˆ 11.4, J(11,10) ˆ
6.0, HÀC(11)); 6.01 6.08 (m, HÀC(10), HÀC(12)); 6.29 (ddd, J(9,10) ˆ 13.2, J(9,8) ˆ 4.1, J(9,11) ˆ 1.2,
HÀC(9)); 7.26 7.56 (m, 10 arom. H); assignments by spin decoupling, NOE, and COSY. 13C-NMR
(75.47 MHz):28.98 (C(14)); 60.78 (C(8)); 66.64 (C(2)); 125.40; 126.44; 127.15; 128.00; 128.74; 128.79; 128.89;
129.05; 129.12; 130.42; 131.63; 136.93; 137.64; 150.12, 152.90 (2 CˆO); assignments of C(2) and C(8) by HMQC.
.
‡
MS:369 (26, M ), 193 (31), 115 (36), 101 (45), 42 (100), 40 (28). Anal. calc. for C13H19N3O2 (369.41):C 74.78,
H 11.38; found:C 74.85, H 5.13.
Data of (11Z)-11-benzylidene-4-phenyl-2,4,6-triazatricyclo[5.4.2.02,6]trideca-8,12-diene-3,5-dione ((Z)-21):
The obtained powder was pure by TLC but 1 H-NMR revealed minor impurities; no crystals could be obtained.
1H-NMR:3.15 ( ddd, J(10b,10a) ˆ 16.2, J(10b,9) ˆ 8.4 , J(10b,1) ˆ 1.2, HbÀC(10); 3.32 (ddt, J(10a,10b) ˆ 16.2,
J(10a,9) ˆ 6.0, J(10a,8) ˆ J(10a,HÀCˆC(11)) ˆ 1.2, HaÀC(10)); 5.23 (t, J(7,8) ˆ J(7,13) ˆ 6.6, HÀC(7)); 5.85
(dd, J(1,12) ˆ 7.2, J(1,10b) ˆ 1.2, HÀC(1)); 5.91 (ddd, J(9,8) ˆ 11.4, J(9,10b) ˆ 8.4, J(9,10a) ˆ 6.0, HÀC(9));
6.03 (ddd, J(8,9) ˆ 11.4, J(8,7) ˆ 6.6, J(8,10b) ˆ 1.2, HÀC(8)); 6.08 (dd, J(12,13) ˆ 9.6, J(12,1) ˆ 7.2, HÀC(12));
6.36 (dd, J(13,12) ˆ 9.6, J(13,7) ˆ 6.6, HÀC(13)); 6.58 (br. s, HÀCˆC(11)); 7.27 7.58 (m, 10 arom. H);
assignments by spin decoupling, COSY, and ROESY. 13C-NMR (50.32 MHz):34.82 (C(4)); 50.27, 51.61 (C(1),
C(6)); 125.79; 126.74; 127.67; 127.72; 128.17; 128.63; 128.85; 128.99; 129.11; 129.89; 130.82; 131.84; 133.64;
136.30; 148.40, 149.01 (2 CˆO).
Data of (11E)-11-benzylidene-4-phenyl-2,4,6-triazatricyclo[5.4.2.02,6]trideca-8,12-diene-3,5-dione ((E)-21).
Colorless crystals. M.p. 170 1728 (AcOEt/petroleum ether). FT-IR:1747 m, 1693s, 1599w, 1502m, 1408m,
1
1377m, 1227w, 762m, 748m. H-NMR:3.11 ( ddd, J(10b,10a) ˆ 15.6, J(10b,9) ˆ 3.6, J(10b,1) ˆ 1.2, HbÀC(10));
3.24 (dd, J(10a,10b) ˆ 15.6, J(10a,9) ˆ 6.6, HaÀC(10)); 5.29 (t, J(7,8) ˆ J(7,13) ˆ 6.6, HÀC(7)); 5.57 (dd,
J(1,12) ˆ 6.6, J(1,10b) ˆ 1.2, HÀC(1)); 5.97 6.01 (m, HÀC(8), HÀC(9)); 6.37 (dd, J(12,13) ˆ 9.6, J(12,1) ˆ 6.6,
HÀC(12)); 6.43 (dd, J(13,12) ˆ 9.6, J (13,7) ˆ 6.6, HÀC(13)); 6.81 (br. s, HÀCˆC(11)); 7.20 7.55 (m, 10 arom.
H). 13C-NMR (50.32 MHz):26.74 (C(4)); 49.60 (C(1)); 57.38 (C(6)); 125.70; 126.34; 127.05; 127.63; 127.66;
128.05; 128.53; 128.95; 129.20; 131.83; 132.10; 132.25; 136.61; 137.29; 148.48, 148.67 (2 CˆO); assignments by
.
‡
spin decoupling, COSY, HMQC, and HMBC. MS:369 (22, M ), 157 (28), 115 (29), 101 (46), 69 (35), 42 (100),
40 (47). Anal. calc. for C23H19N3O2 (369.41):C 74.78, H 5.18; found:C 74.70, H 5.08.
REFERENCES
[1] P. H. Ferber, G. E. Gream, P. K. Kirkbride, E. R. T. Tiekink, Aust. J. Chem. 1990, 43, 463.
[2] S. W. Staley, G. E. Linkowski, A. S. Heyn, Tetrahedron 1975, 31, 1131.
[3] M. Tamura, J. Kochi, Synthesis 1971, 303.
[4] B. F. G. Johnson, J. Lewis, D. Wege, J. Chem. Soc., Dalton Trans. 1976, 1874.
[5] L. D. Bergelson, L. I. Barsukov, M. M. Shemyakin, Tetrahedron 1967, 23, 2709.
[6] M. Krˆner, Chem. Ber. 1967, 100, 3162.
[7] A. Streitwieser, in −Solvolytic Displacement Reactions×, McGraw-Hill, U.S.A., 1962, p. 95.

Helvetica Chimica Acta Vol. 88 (2005)
2021
[8] F. W. Wehrli, T. W. Wirthlin, −Interpretation of Carbon-13 NMR Spectra×, Heydon and Sons Ltd., London,
1978.
[9] G. E. Gream, P. K. Kirkbride, unpublished data.
[10] G. E. Gream, P. K. Kirkbride, E. R. T. Tiekink, Z. Kristallogr., New Cryst. Struct. 2002, 217, 353
[11] G. E. Gream, P. K. Kirkbride, E. R. T. Tiekink, unpublished data.
[12] L. M. Jackman, S. Sternhell, −Applications of Nuclear Magnetic Resonance Spectroscopy in Organic
Chemistry×, 2nd edn., Pergamon Press, Oxford, 1969.
[13] W. H. Urry, M. S. Kharasch, J. Am. Chem. Soc. 1944, 66, 1438.
[14] G. Wittig, W. Haag, Chem. Ber. 1955, 88, 1654.
[15] G. E. Gream, P. K. Kirkbride, E. R. T. Tiekink, Z. Krystallogr., New Cryst. Struct. 2002, 217, 355.
[16] G. E. Gream, P. K. Kirkbride, E. R. T. Tiekink, Z. Krystallogr., New Cryst. Struct. 2002, 217, 351.
[17] G. E. Gream, P. K. Kirkbride, E. R. T. Tiekink, Z. Krystallogr., New Cryst. Struct. 2001, 216, 571
[18] G. E. Gream, P. K. Kirkbride, E. R. T. Tiekink, Z. Krystallogr., New Cryst. Struct. 1994, 209, 545.
[19] R. B. Woodward, R. Hoffmann, −The Conservation of Orbital Symmetry×, Academic Press, New York,
1971.
[20] H. Isaksen, J. P. Snyder, Tetrahedron Lett. 1977, 889.
[21] G. W. Buchanan, Tetrahedron Lett. 1972, 665.
[22] M. St.-Jacques, R. Prud×homme, Tetrahedron Lett. 1970, 4833.
[23] P. Ahlberg, D. L. Harris, M. Roberts, P. Warner, P. Seidel, M. Sakai, D. Cook, A. Diaz, J. P. Dirlam, H.
Hamberger, S. Winstein, J. Am. Chem. Soc. 1972, 94, 7063.
[24] J. R. Malpass, J. Chem. Soc., Chem. Commun. 1972, 1246.
[25] R. Huisgen, G. Boche, W. Hechtl, H. Huber, Angew. Chem., Int. Ed. 1966, 5, 585; R. Huisgen, G. Boche, H.
Huber, J. Am. Chem. Soc. 1967, 89, 3345; R. Huisgen, J. Gasteiger, Angew. Chem., Int. Ed. 1972, 11, 1104.
[26] S. Winstein, C. G. Kreiter, J. I. Bauman, J. Am. Chem. Soc. 1966, 88, 2047.
[27] G. I. Fray, P. R. Saxton, −The Chemistry of Cyclo-Octatetraene and its Derivatives×, Cambridge University
Press, 1978.
[28] J. Gasteiger, R. Huisgen, Angew. Chem., Int. Ed. 1972, 11, 716.
[29] J. Gasteiger, G. E. Gream, R. Huisgen, W. E. Konz, U. Schnegg, Chem. Ber. 1971, 104, 2414.
[30] R. Huisgen, G. Boche, A. Dahmen, W. Hechl, Tetrahedron Lett. 1968, 5212.
Received March 11, 2005