Photochemically generated cyclopentane-1,3-diyl triplet diradicals as model systems for the assessment of spin delocalization in heteroaryl-substituted benzyl-type monoradicals through the EPR spectral D parameter

Article abstract of DOI:10.1039/a607513d

The zero-field D parameters of a comprehensive set of cyclopentane-1,3-diyl triplet diradicals 2 were determined at 77 K in a 2-MTHF glass matrix. The D values were found to be dependent on the heteroaryl substituents in the decreasing order 3-furyl ? 3-t

Full text of DOI:10.1039/a607513d

View Article Online / Journal Homepage / Table of Contents for this issue
Photochemically generated cyclopentane-1,3-diyl triplet diradicals as
model systems for the assessment of spin delocalization in heteroaryl-
substituted benzyl-type monoradicals through the EPR spectral D
parameter
Waldemar Adam,* Oliver Emmert † and Heinrich M. Harrer
Institute of Organic Chemistry, University of Würzburg, Am Hubland, D-97074 Würzburg,
Germany
The zero-field D parameters of a comprehensive set of cyclopentane-1,3-diyl triplet diradicals 2 were
determined at 77 K in a 2-M TH F glass matrix. The D values were found to be dependent on the heteroaryl
substituents in the decreasing order 3-furyl ӷ 3-thienyl > 4-pyridyl ≈ 3-N-oxypyridyl > 2-pyridyl ≈ 3-
pyridylium > 3-pyridyl ≈ phenyl ӷ 2-pyridylium ӷ 4-pyridylium ӷ 2-furyl ӷ 2-thienyl ӷ 4-N-oxypyridyl.
G ood linear correlations were obtained with the reported aá coupling constants (r2 = 0.953) for the benzyl
monoradicals 4 and with the semiempirically calculated (PM 3) á spin densities (r2 = 0.928) for the cumyl
monoradicals 3. To rationalize the observed electronic effects, for convenience the ÄDAr scale was defined
as a measure of the spin-delocalizing ability of the different heteroaryl substituents relative to the phenyl
group as reference. The hitherto unknown electronic effects of N-oxidation and protonation for the
different pyridyl regioisomers as well as the regioisomeric effects of the furyl and thienyl substituents,
are experimentally reflected accurately by the changes in the D parameter of the triplet diradicals 2 and
explained theoretically with the help of M O calculations for the corresponding monoradicals 3.
has been difficult because insufficiently high concentrations of
Introduction
these radicals could be obtained in solution, while the add-
itional structural complexity has encumbered the reliable
assignment of coupling constants even in computer-aided
analyses.4 Hence, the effects on the spin density distribution in
such species has remained doubtful. For example, the assign-
ment of the aα hyperfine coupling constants for the pyridyl-
methyl radicals 4 in solution,4 which are a direct measure of the
α spin density (McConnell equation),1b showed considerable
disagreement with results obtained for these radicals in ada-
mantane matrices.5 However, all pyridylmethyl radicals showed
larger aα hfc constants (cf. Table 1) than the parent benzyl
case, i.e. 4- > 2- > 3-pyridylmethyl > benzyl, which was also in
line with β muon couplings of muonated aryl-substituted ethyl
radicals.6 Semiempirical calculations have failed so far to
account for these experimental results.5,7
Of relevance in this context are also the thienyl- and furyl-
methyl radicals 4b–c and l–m, since they constitute five-
membered ring heteroaromatic analogues of the parent benzyl
radical. Of these, only the radicals 4b–c and l have been stud-
ied by EPR spectroscopy in solution and shown to exhibit also
complex spectra. The aα hfc constants8 differ strongly from
those of the benzyl radical (cf. Table 1), which is especially the
case for the ortho isomers. Moreover, these electronic substitu-
ent effects have not been rationalized in terms of spin delocal-
ization by these heteroaryl groups.
The photochemical deazetation of diazabicyclo[2.2.1]heptane
(DBH) derivatives generates conveniently cyclopentane-1,3-diyl
triplet diradicals, which are persistent in low-temperature
matrices.1 These intermediates may be detected by EPR spec-
troscopy and characterized by the zero-field splitting (zfs)
parameters D and E.1 For such localized triplet 1,3-diradicals,
the D parameter derives from a two-centre dipolar interaction
between the radical termini, whose magnitude depends on the
interspin distance dAB and the spin densities ρA and ρB at the
respective radical sites A and B [eqn. (1)]. The spin density2a,b as
3µ0g2µB2 ρAρB
(1)
D =
ͩ ͪ
16π
dA3 B
well as the geometry dependence2c have recently been con-
firmed experimentally and theoretically. Therefore, such local-
ized 1,3-disubstituted cyclopentane-1,3-diyl triplet diradicals
can be considered as a composite of two geometrically fixed
benzyl-type monoradicals. Hence, at constant dAB, the D par-
ameter is a sensitive probe for electronic substituent effects
through the α spin density dependence and provides a measure
of radical stabilization through spin delocalization.
In view of the aforementioned inconsistencies, it was of inter-
est to examine the electronic influence of the various regioiso-
meric pyridyl-, furyl- and thienyl-methyl substituents on the
D parameter in the triplet 1,3-diradicals 2. Specifically, it was
to be assessed whether the α spin density at the radical site
serves as a measure of the spin-delocalizing propensity of these
heteroaryl units. Since the D parameter derives solely from
the dipolar spin–spin interactions,2a this quantity is not
encumbered by hyperfine splittings at the radical site in the
EPR spectra, which should facilitate the diagnosis of substitu-
ent effects. Herein, we report the results of our investigation for
the unsymmetrical mono-substituted diradicals 2, also for the
An interesting electronic modification of the well-studied
benzyl (and cumyl) radicals 4f3 are the pyridylmethyl deriv-
atives 4g,i,k.‡ However, their accurate EPR spectral examination
† Undergraduate research participant, Spring semester 1996.
‡ For the various substituents of Ar (a–m) of compounds 1–5 see
Table 1.
J. Chem. Soc., Perkin Trans. 2, 1997
687

View Article Online
Scheme 1 Preparation of the pyridinium derivatives 1d,e and h and the pyridine-N-oxide derivatives 1a and j
protonated and N-oxidized pyridylmethyl derivatives, the latter
examined EPR-spectrally for the first time. We demonstrate
unequivocally that the D parameter is a reliable measure of the
spin-delocalizing properties of heteroaryl π systems.
Results
The synthesis of the heteroaryl-substituted azoalkanes 1b–c,g,i
and k–m was carried out according to literature procedures,9a,b
whereas the derivative 1f has been reported earlier.9c The
pyridyl-substituted azoalkanes 1g,i and k were further func-
tionalized by nitrogen protonation and oxidation (Scheme 1).
Thus, with 70% perchloric acid they gave the pyridinium per-
chlorates 1d–e and h in excellent yields, while the N-oxides
1a and 1j were obtained by dimethyldioxirane oxidation.10
Unfortunately, the oxidation with (a) dimethyldioxirane or (b)
m-CPBA of the ortho regioisomer 1i did not result in the
desired N-oxide, instead the two azoxy derivatives 5i and iЈ
were formed in high yields (cf. Scheme 1).
Fig. 1 Representative EPR spectrum of the triplet diradical 2i. The
important Z signals for the determination of the D parameter are also
indicated.
Functionalization (nitrogen protonation and oxidation) of
The diradicals 2 were generated in a 2-methyltetrahydrofuran
(MTHF) matrix at 77 K by means of irradiation with the 364
nm line of an argon ion laser and a representative EPR spec-
trum is given in Fig. 1. The results of the EPR measurements
are summarized in Table 1. A marked dependence of the
experimental D values on the heteroaryl substituent is evident,
which is in general larger than the previously observed elec-
tronic effects of para and meta substituents on the phenyl
derivative 2f.2 Hence, the smallest D value is displayed by the
diradical 2a (0.0430 cmϪ1), whereas diradical 2m possesses the
largest (0.0539 cmϪ1) and both flank the parent phenyl case 2f
(0.0506 cmϪ1), which is taken as a reference system for these
heteroaryl substituents. While the meta-pyridyl regioisomer 2g
(0.0507 cmϪ1) has the same D value within the experimental
error as the phenyl reference system 2f (0.0506 cmϪ1), both the
ortho-2i (0.0510 cmϪ1) and para-2k (0.0512 cmϪ1) regioisomers
exhibit definitely larger values.
the pyridine lone pair affects significantly the D parameter of
the corresponding diradicals 2. This is best exemplified for the
para-pyridyl regioisomer 2k (0.0512 cmϪ1), for which the D
value decreases substantially on protonation to afford the
diradical 2d (0.0489 cmϪ1). A similar trend, but less pro-
nounced, applies to the ortho regioisomers, i.e. 2i (0.0510 cmϪ1
)
versus 2e (0.0496 cmϪ1). For the corresponding meta derivatives
2g (0.0507 cmϪ1) versus 2h (0.0509 cmϪ1), the effect of proton-
ation is nominal, if at all significant. Contrary to the ortho and
para regioisomers, the D parameter is displaced to slightly
larger values.
N-Oxidation is substantially more effective than protonation
in changing the D parameters and, hence, the electronic proper-
ties of the pyridine ring. Thus, the para-N-oxide 2a (0.0430
cmϪ1) stands out of all heteroaryl substituents investigated here
with the smallest D value and implies some dramatic electronic
changes. In contrast, for the meta regioisomer 2j (0.0511 cmϪ1
)
More dramatic effects are witnessed for the thienyl (2b,l) and
furyl (2c,m) derivatives. Thus, the ortho isomers 2b (0.0445
cmϪ1) and 2c (0.0457 cmϪ1) of these five-membered heteroaryl
substituents have substantially smaller D values than the refer-
ence system 2f (0.0506 cmϪ1). However, the meta isomers 2l
(0.0518 cmϪ1) and 2m (0.0539 cmϪ1) possess larger D values.
a small but definite displacement to a higher D value is
observed on N-oxidation of the pyridyl derivative 2g (0.0507
cmϪ1). These electronic effects of the heteroaryl substituents
shall now be interpreted in terms of their spin-delocalizing pro-
pensity with respect to the parent phenyl derivative 2f as refer-
ence system.
688
J. Chem. Soc., Perkin Trans. 2, 1997

View Article Online
Table 1 Experimental D values of the triplet diradicals 2, calculated α
spin densities of the cumyl-type monoradicals 3, ∆DAr values for the
heteroaryl substituents and a hyperfine coupling constants of the cor-
responding benzyl-type monoradicals 4
∆DAr/102
aα of
4/G
Ar
|D/hc|/cmϪ1 a
ρα of 3b
cmϪ1 c
2a
0.0430
0.423
ϩ0.76
—
2b
2c
2d
0.0445
0.0457
0.0489
0.452
0.475
0.469
ϩ0.61
ϩ0.49
ϩ0.17
13.84e
13.85e
—
Fig. 2 D values of the triplet diradicals 2 versus the experimental
a hyperfine splitting constants of the benzyl-type monoradicals 4
taken from refs. 3, 4 and 8; (᭿) marks the phenyl reference system 2f/4f
2e
2f
0.0496
0.498
0.522
ϩ0.10
—
0.0506d
0.00
16.35f
2g
2h
0.0507
0.0509
0.538
0.524
Ϫ0.01
Ϫ0.03
16.51g
—
2i
2j
0.0510
0.0511
0.544
0.532
Ϫ0.04
Ϫ0.05
16.80g
Fig. 3 D values of the triplet diradicals 2 versus the calculated α spin
densities (ρα) of the cumyl-type radicals 3; (᭿) marks the phenyl refer-
ence system 2f/3f
—
Discussion
The radical centres in the localized triplet 1,3-diradicals act
independently of each other, except for dipolar spin–spin inter-
actions, as confirmed through the additivity of substituent
effects in 1,3-diaryl-substituted derivatives.2 Since in all
unsymmetrical derivatives 2 one radical side is kept constant
(phenyl substitution), the experimentally assessed changes in
the D parameter must derive from the different heteroaryl sub-
stituents a–m. This is experimentally verified by the good linear
correlation between the reported a hyperfine couplings of the
monoradicals 4 and the D parameters of the triplet diradicals 2
in Fig. 2. Hence, the two EPR spectroscopic |D/hc| and a hfc
quantities manifest that the spin-delocalizing propensity of the
heteroaryl substituents decrease in the order 2-thienyl > 2-
furyl > phenyl > 3-pyridyl > 2-pyridyl > 4-pyridyl > 3-thienyl >
3-furyl.
2k
2l
0.0512
0.551
Ϫ0.06
17.20g
0.0518
0.0539
0.519
0.552
Ϫ0.12
Ϫ0.33
16.73e
2m
—
a
Measured in a 2-MeTHF matrix at 77 K, error ±0.0001 cmϪ1
,
|E/hc| < 0.002 cmϪ1 b Calculated α spin densities, cf. text. c Calculated
.
according to eqn. (2). d Ref. 2. e Ref. 8, stated error ±0.03 G. f Ref. 3,
stated error ±0.02 G. g Ref. 4, stated error ±0.01 G.
Indeed, theoretical spin densities, calculated by the semiem-
pirical PM3 method for the heteroaryl-substituted monoradi-
cals 3, confirm the above trend (Fig. 3).5,7 Consequently, the
J. Chem. Soc., Perkin Trans. 2, 1997
689

View Article Online
Fig. 4 Calculated (PM3-AUHF) spin densities (ρ) for the cumyl-type monoradicals 3 and experimental D parameters of the triplet diradicals 2
(values given in parentheses)
good correlation of the D parameters for the triplet diradicals 2
with the experimental hyperfine coupling constants (Fig. 2) and
the calculated α spin densities (Fig. 3) of the corresponding
monoradicals 3 demonstrate that electronic substituent effects
for heteroaryl groups are reliably reproduced by the D values in
terms of spin delocalization. The better the heteroaryl substitu-
ent delocalizes spin, the more spin diffuses into the aromatic
moiety (mainly at the ortho and para positions) and less spin
resides at the radical centre and a lower D value results, as
demanded by eqn. (1).
The D values in Table 1 display some interesting and remark-
able trends in the spin-delocalizing ability of the heteroaryl sub-
stituents. These experimental data, which have been for the first
time acquired and provide a quantitative measure of heteroaryl
conjugation, shall now be compared with the phenyl group as
reference point. For this purpose it is convenient to define the
∆DAr parameter [eqn. (2)] as the difference between the D
group, but in the ortho and para isomers 2i and k delocaliz-
ation is definitely less effective. This trend is also evident in the
calculated spin densities (ρα) at the radical site of the corre-
sponding heteroaryl-type cumyl radicals 3g (0.538), 3i (0.544)
and 3k (0.551) versus the parent phenyl reference system 3f
(0.522). Again, relative to phenyl, the pyridyl derivatives act as
spin donors by enhancing the spin density at the radical site,
although the effects are relatively small. Furthermore, inspec-
tion of the spin distribution within the aromatic ring of the
monoradicals 3 (Fig. 4) reveals that for the more effective spin
donors ortho- and para-pyridyl, the respective spin densities at
the nitrogen sites, namely 3i (0.097) and 3k (0.098), are substan-
tially reduced compared to those of the parent cumyl radical 3f
(0.120 at ortho and 0.128 at para). This indicates that aminyl-
type radical structures such as in the resonance hybrid A,
exemplified for the para-pyridyl case, are discouraged due to
unfavourable spin accumulation at the nitrogen atom. This
aminyl-type radical destabilization, a well documented fact,11a
expresses itself in the reluctance of the para- and ortho-pyridyl
substituents to delocalize spin into the aromatic ring in the
order para (3k) > ortho (3i) > meta (3g) ≈ phenyl (3f). Note in
Fig. 4 that for the meta-pyridyl derivative 3g the ortho and para
ring spin densities are essentially the same as those for the par-
ent cumyl case 3f. Thus, in this regioisomer the pyridyl sub-
stituent is as effective in delocalizing spin as the phenyl group
because no unfavourable aminyl-type radical structures apply.
Protonation of the nitrogen atom in the pyridine moiety
changes substantially its spin-delocalizing propensity. This is
most prominently seen for the former weakly spin-donating
para-pyridyl substituent in 2k (∆DAr = Ϫ0.06), which becomes
now a relatively good spin-accepting one in 2d (∆DAr = ϩ0.17).
∆DAr = [DPh Ϫ D(2)] × 100
(2)
values of the phenyl (DPh) and the heteroaryl group [D(2)] in
the triplet diradicals 2. Thus, positive ∆DAr values indicate that
the heteroaryl substituent delocalizes spin better than the
phenyl group and may be classified as spin acceptor. Heteroaryl
substituents with negative ∆DAr values delocalize spin worse
than the phenyl group and may be classified as spin donors.
Let us first consider the three regioisomeric pyridyl-
substituted triplet diradicals 2g (Ϫ0.01), 2i (Ϫ0.05) and 2k
(Ϫ0.06), for which the ∆DAr values are given in parentheses.
Clearly, within the experimental error, the pyridyl substituent in
the meta isomer 2g conjugates about as well as the phenyl
690
J. Chem. Soc., Perkin Trans. 2, 1997

View Article Online
This may be rationalized in terms of the stabilizing nitrogen-
centred radical–cation resonance structure B for the monoradi-
cal 3d, a stabilization that also applies for the ortho-pyridyl
substituent in 3e. Still more dramatic changes of the experi-
mental D values result from N-oxidation, as exemplified for the
4-N-oxypridyl substituent in the triplet diradical 2a. The ∆DAr
value of ϩ0.76, the largest in the set of heteroaryl substituents
examined herein (Table 1), indicates massive spin delocalization
into the aromatic ring. The semiempirical MO calculations for
the monoradical 3a make evident that the nitroxide resonance
hybrid C is significantly populated, as expressed by the large
spin density at the nitroxide functionality (Fig. 4). This nicely
accounts for the observed strong spin-delocalizing ability of
this heteroaryl substituent. A similar explanation was given
recently for the N-oxypyridyl-2-thio radical 5a, which adds con-
siderably slower to dienes compared to the non-oxidized
pyridyl-2-thio radical 5b.12 The corresponding nitroxide reson-
phenyl group in the parent triplet diradical 2f. This is particu-
larly well illustrated by the ring spin densities of the corre-
sponding monoradicals 3 (Fig. 4). Almost 40% resides at the 3
and 5 positions in the ortho regioisomers 3b and c, ca. 25% at
the 2 positions of the meta derivatives 3l and m. Thus, versus
the phenyl group, the ortho regioisomers 3b and c qualify as
strong spin acceptors, while the meta ones operate as moderate
spin donors. Why? As the theoretical ring spin densities (Fig. 4)
substantiate, for the ortho isomers 3b and c an extended
pentadienyl-type spin delocalization applies in terms of the res-
onance structures D and F. Thus, the synergistic effect of rad-
ical and ring conjugation provides for extensive delocalization.
The reason why this is much more effective than for the phenyl
group is apparently, as already stated, the less pronounced aro-
matic character of the furyl and thienyl moieties.13 In con-
trast, for the meta derivatives 3l and m, cross-conjugation
between the aromatic ring (Hückel-type) and the radical site
(allyl-type) compete and spin delocalization is significantly less
effective for these heteroaryl substituents relative to the phenyl
group. This must not be construed that there is no spin delocal-
ization in the meta regioisomers 3l and m at all. Indeed, sub-
stantial (ca. 25%) spin density diffuses into the heteroaryl ring
through allylic conjugation, as the calculated ring spin densities
at the 2-positions, namely 3l (0.262) and 3m (0.250), portray
unmistakenly; however, we reiterate that compared to the
phenyl group, the 3-furyl and 3-thienyl substituents delocal-
ize spin density in cumyl-type radicals 3 (triplet diradicals 2)
less efficiently. In fact, a better comparison of these heteroaryl
substituents would be with the cyclopentadienyl anion as refer-
ence system, but all synthetic efforts on the corresponding trip-
let diradical have failed so far. Nevertheless, we anticipate that
the delocalized negative charge in the cyclopentadienyl anion
moiety would resist spin delocalization from the radical site
into the aromatic ring. Consequently, the uncharged 2-furyl
and 2-thienyl groups should act as spin acceptors relative to
the cyclopentadienyl anion as reference group.
Our data on the D parameters (Table 1) for the heteroaryl-
substituted triplet diradicals 2 and the theoretical spin densities
for the corresponding heteroaryl derivatives of the cumyl-type
monoradicals 3 (Fig. 4) have allowed us for the first time to
probe experimentally and theoretically the electronic effects of
pyridyl, furyl and thienyl groups on cumyl-related radical
centres through spin delocalization effects. Some unprecedented
electronic trends have become evident through nitrogen proto-
nation and oxidation in the pyridine series 2g,i,k and regio-
isomeric effects in the furan and thiophene derivatives 2c,m
and 2b,l. While in principle such information may be acquired
through EPR hyperfine couplings on the monoradicals,1b,4,8 the
synthetic ease of preparing the required azoalkane pre-
cursors2,9 for the photochemical generation of the matrix-
ance structure for the N-oxide radical reduces the α spin density
at the sulfur radical site and, hence, lowers the reactivity.
Additional experimental evidence for the nitroxide structure
in the resonance hybrid C provides the g values of the diradicals
2. Normally these fall in the range of 2.0020 to 2.0025, but for
the triplet diradical 2a a substantial increase to 2.0045 is found,
which definitively expresses significant nitroxide radical char-
acter, since the latter have g values in the range from 2.0055 to
2.0065.11b In this context it is relevant to note that for the triplet
diradical 2j with the meta-N-oxypyridyl substituent a normal g
value of 2.0027 is observed. This is expected because in the meta-
N-oxypyridyl group the N-oxide functionality is not in direct
conjugation with the radical centre.
The comparison of the ∆DAr parameters of the unfunctional-
ized, pyridyl-substituted triplet diradicals 2g (Ϫ0.01), 2i
(Ϫ0.04) and 2k (Ϫ0.06) with the furyl derivatives 2c (ϩ0.49)
and 2m (Ϫ0.33) and the thienyl ones 2b (ϩ0.61) and 2l (Ϫ0.12)
makes clearly evident that the five-membered ring heteroaryl
substituents interact much more strongly with the radical
centre. In view of the lower aromatic character of the five- ver-
sus the six-membered aryl substituents, the 6π electron system
of the former is more easily perturbed by spin delocalizing
effects.13 Also the calculated spin densities (ρα) at the radical
sites for the furyl and thienyl derivatives (Table 1) bring out
this trend in that the changes in ρα are more pronounced in the
five- versus six-membered heteroaryl substituents. Notable is the
fact that the ring spin densities at the oxygen and sulfur atoms
are quite low, which implies relatively little interaction of the
heteroatom with the spin centre.
Interestingly, the ∆DAr parameters of the furyl- and
thienyl-substituted triplet diradicals 2 vary strongly with the
substitution pattern. The heteroaryl substituents in the ortho
isomers 2b (ϩ0.67) and 2c (ϩ0.49) are much better spin delocal-
izers, whereas in the meta isomers 2l (Ϫ0.12) and 2m (Ϫ0.33)
these heteroaryl substituents delocalize spin worse than the
J. Chem. Soc., Perkin Trans. 2, 1997
691

View Article Online
isolated triplet diradicals and the convenience of measuring the
EPR spectra for the latter,1a,2 offers definite advantages for the
herein presented methodology to assess electronic substituent
effects in radical species.
(1á,4á,4aá,7aá)-4,4a,5,6,7,7a-Hexahydro-8,8-dimethyl-1-
(2Ј-furyl)-4-phenyl-1,4-methano-1H-cyclopenta[d]pyridazine
(1c). Colourless needles (137 mg, 89%), mp 67–68 ЊC (decomp.)
(Found: C, 78.15; H, 7.41; N, 8.89. C20H22N2O requires C,
78.40; H, 7.24; N, 9.14%); λmax(C6H6)/nm 359 (log ε 2.01);
νmax(KBr)/cmϪ1 3060, 2970, 2940, 1680, 1660, 1460; δH (CDCl3)
0.35 (3 H, s, endo-CH3), 1.02 (3 H, s, exo-CH3), 1.56 (6 H, m,
Experimental
3
3
Computations
CH2), 3.41 (2 H, mc, 4a-H and 7a-H), 6.51 (1 H, dd, J 3.2, J
3 4
Full geometry optimization of the monoradicals 3 was carried
out on the highest molecular symmetry with a planar arrange-
ment of the aryl groups at the radical site by using the PM3
method 14a and the A(nnihilated) UHF wave function,14b which
are provided in the VAMP5.0 program 15 and run on an IRIS
INDIGO Silicon Graphics Workstation. The α spin densities
were determined with a single-point CI calculation, which
results in excellent spin expectation 〈S2〉 values between 0.76
and 0.78 for these radicals.
1.8, 4Ј-H), 6.73 (1 H, dd, J 3.2, J 0.8, 3Ј-H), 7.40 (3 H, m,
3 4 3 4
Ph), 7.53 (1 H, dd, J 1.8, J 0.8, 5Ј-H), 7.73 (2 H, dt, J 6.3, J
1.7, ortho-Hphenyl); δC(CDCl3) 17.2 (q), 18.0 (q), 25.3 (t), 25.8 (t),
28.4 (t), 48.5 (d), 48.6 (d), 65.7 (s), 95.4 (s), 97.3 (s), 108.8 (d),
110.3 (d), 127.4 (2 × d), 127.7 (d), 128.3 (2 × d), 135.8 (s), 142.7
(d), 151.4 (s).
(1á,4á,4aá,7aá)-4,4a,5,6,7,7a-Hexahydro-8,8-dimethyl-1-(3Ј-
pyridyl)-4-phenyl-1,4-methano-1H-cyclopenta[d]pyridazine (1g).
Colourless needles (159 mg, 99%), mp 88–89 ЊC (decomp.)
(Found: C, 79.87; H, 7.64; N, 13.33. C21H23N3 requires C, 79.46;
H, 7.30; N, 13.24%); λmax(C6H6)/nm 363 (log ε 2.00); νmax(KBr)/
cmϪ1 2940, 2820, 1480, 1460, 1400; δH (CDCl3) 0.18 (3 H, s,
endo-CH3), 1.00 (3 H, s, exo-CH3), 1.55 (6 H, mc, CH2), 3.54 (2
H, mc, 4a-H and 7a-H), 7.40–7.54 (4 H, m, arom. H), 7.72–7.77
EPR spectroscopy
A sample (ca. 5 × 10Ϫ4 mmol) of the azoalkanes 1 was dissolved
in 0.3 cm3 of MTHF, placed into an EPR sample tube (inner
diameter ca. 2 mm) and thoroughly degassed by purging with
argon gas. The sample was sealed and the matrix was prepared
at 77 K by freezing the samples in liquid nitrogen. The triplet
diradicals 2 were generated by irradiation with the 364 nm line
of an INNOVA-100 CW argon-ion laser (widened beam, 1.5 W,
2 min) at 77 K. Their EPR spectra were recorded with a Bruker
ESP-300 spectrometer (9.43 GHz, spectra accumulation with
the Bruker 1620 data system, n у 5). The D values were deter-
mined by a manual analysis of the Z signals.2
3
4
4
(2 H, m, ortho-Hphenyl), 8.15 (1 H, ddd, J 8.0, J 2.3, J 1.7, 4Ј-
H), 8.67 (1 H, dd, 3J 4.8, 4J 1.4, 6Ј-H), 8.96 (1 H, d, 4J 1.5, 2Ј-H);
δC(CDCl3) 16.9 (q), 17.7 (q), 25.39 (t), 25.41 (t), 28.5 (t), 48.8
(2 × d), 66.4 (s), 96.6 (s), 98.7 (s), 123.4 (d), 127.4 (2 × d), 127.9
(d), 128.4 (2 × d), 132.0 (s), 135.4 (d), 135.7 (s), 148.5 (d), 149.2
(d).
(1á,4á,4aá,7aá)-4,4a,5,6,7,7a-Hexahydro-8,8-dimethyl-1-
(2Ј-pyridyl)-4-phenyl-1,4-methano-1H-cyclopenta[d]pyridazine
(1i). Colourless needles (159 mg, 99%), mp 100–101 ЊC
(decomp.) (Found: C, 79.91; H, 7.72; N, 13.38. C21H23N3
requires C, 79.46; H, 7.30; N, 13.24%); λmax(C6H6)/nm 362 (log ε
2.01); νmax(KBr)/cmϪ1 2930, 2840, 1460, 1420, 1360; δH (CDCl3)
0.19 (3 H, s, endo-CH3), 1.11 (3 H, s, exo-CH3), 1.56 (6 H, m,
CH2), 3.57 (1 H, mc, 4a-H), 3.84 (1 H, mc, 7a-H), 7.26–7.50 (4
H, m, arom. H), 7.74–7.90 (3 H, m, Ph), 8.14 (1 H, dt, 3J 7.9, 4J
1.1, 3Ј-H), 8.70 (1 H, ddd, 3J 4.9, 4J 1.9, 5J 0.9, 6Ј-H); δC(CDCl3)
16.9 (q), 18.2 (q), 25.5 (2 × t), 28.5 (t), 48.9 (d), 49.0 (d), 66.7 (s),
98.8 (s), 99.5 (s), 122.5 (d), 123.6 (d), 127.6 (2 × d), 127.7 (d),
128.3 (2 × d), 136.0 (s), 136.2 (d), 149.3 (d), 156.8 (s).
(1á,4á,4aá,7aá)-4,4a,5,6,7,7a-Hexahydro-8,8-dimethyl-1-
(4Ј-pyridyl)-4-phenyl-1,4-methano-1H-cyclopenta[d]pyridazine
(1k). Colourless powder (159 mg, 99%), mp 142–143 ЊC
(decomp.) (Found: C, 79.55; H, 7.28; N, 13.55. C21H23N3
requires C, 79.46; H, 7.30; N, 13.24%); λmax(C6H6)/nm 365 (log ε
1.99); νmax(KBr)/cmϪ1 2920, 2880, 1580, 1450, 1430; δH (CDCl3)
0.17 (3 H, s, endo-CH3), 1.02 (3 H, s, exo-CH3), 1.52 (6 H, mc,
CH2), 3.52 (2 H, mc, 4a-H and 7a-H), 7.43–7.54 (3 H, m, Ph),
7.68–7.76 (4 H, m, arom. H), 8.73 (2 H, d, 3J 5.4, 2Ј-H);
δC(CDCl3) 16.8 (q), 17.8 (q), 25.4 (2 × t), 28.5 (t), 48.8 (d), 49.0
(d), 66.5 (s), 97.0 (s), 99.0 (s), 122.4 (2 × d), 127.5 (2 × d), 127.9
(d), 128.4 (2 × d), 135.4 (s), 145.2 (s), 150.0 (2 × d).
General aspects
1H and 13C NMR spectra were measured on a Bruker AC 200
(1H: 200 MHz, 13C: 50 MHz) spectrometer with deuteriochloro-
form or deuterioacetonitrile as internal standard. J values are
given in Hz. IR spectra were recorded on a Perkin-Elmer 1420
ratio recording IR spectrophotometer. UV spectra were taken
on an Hitachi U 3200 spectrophotometer. Elemental analyses
were carried out by the Microanalytical Division of the Insti-
tute of Inorganic Chemistry, University of Würzburg. Melting
points were taken on a Büchi apparatus B-545. TLC analyses
were conducted on precoated silica-gel foils Polygram SIL G/
UV254 (40 × 80 mm) from Macherey & Nagel. Spots were iden-
tified under a UV lamp or by exposure to iodine vapour. Silica
gel (63–200 µm; Woelm) was used for column chromatography,
the adsorbant :substrate ratio was ca. 100:1.
Preparation of the azoalkanes 1b–c,g,i and k–m
The corresponding unsaturated azoalkane9b (0.500 mmol) was
dissolved in ethyl acetate (40 cm3) and ca. 10 mg of palladium
on charcoal catalyst were added. The suspension was deareated
and saturated with hydrogen gas. The hydrogenation was car-
ried out at ca. 20 ЊC for 24 h by using a slight pressure of
hydrogen gas. The catalyst was removed from the reaction mix-
ture by filtration and the solvent evaporated (ca. 40 ЊC/15 Torr)
to afford the azoalkane 1.
(1á,4á,4aá,7aá)-4,4a,5,6,7,7a-Hexahydro-8,8-dimethyl-1-
(3Ј-thienyl)-4-phenyl-1,4-methano-1H-cyclopenta[d]pyridazine
(1l). Colourless needles (143 mg, 88%), mp 116–117 ЊC
(decomp.) (Found: C, 74.44; H, 7.15; N, 8.44; S, 9.73. C20H22N2S
requires C, 74.49; H, 6.87; N, 8.69; S, 9.94%); λmax(C6H6)/nm
362 (log ε 2.08); νmax(KBr)/cmϪ1 3100, 2960, 2925, 1480, 1460,
1405; δH (CDCl3) 0.23 (3 H, s, endo-CH3), 1.00 (3 H, s, exo-CH3),
(1á,4á,4aá,7aá)-4,4a,5,6,7a-Hexahydro-8,8-dimethyl-1-(2Ј-
thienyl)-4-phenyl-1,4-methano-1H-cyclopenta[d]pyridazine
(1b). Colourless needles (157 mg, 95%), mp 109–110 ЊC
(decomp.) (Found: C, 74.42; H, 7.18; N, 8.44; S, 9.82.
C20H22N2S requires C, 74.49; H, 6.87; N, 8.69; S, 9.94%);
λmax(C6H6)/nm 362 (log ε 2.09); νmax(KBr)/cmϪ1 3100, 2990,
2940, 1700, 1490, 1460; δH (CDCl3) 0.27 (3 H, s, endo-CH3),
1.03 (3 H, s, exo-CH3), 1.58 (6 H, m, CH2), 3.32 (1 H, dt, 3J 8.4,
3J 8.0, 3J 5.0, 7a-H), 3.48 (1 H, mc, 4a-H), 7.16 (1 H, dd, 3J 5.0,
3
3
3
1.56 (6 H, m, CH2), 3.35 (1 H, dt, J 8.7, J 7.6, J 4.6, 7a-H),
3.50 (1 H, mc, 4a-H), 7.35 (1 H, dd, 3J 5.0, 4J 1.3, 4Ј-H), 7.45 (4
H, m, 4 H, 5Ј-H and Ph), 7.66 (1 H, dd, 4J 2.9, 4J 1.3, 2Ј-H), 7.75
3
4
(2 H, dt, J 6.7, J 1.3, ortho-Hphenyl); δC(CDCl3) 17.1 (q), 17.9
(q), 25.5 (t), 25.6 (t), 28.5 (t), 48.7 (d), 50.3 (d), 65.8 (s), 96.9 (s),
97.6 (s), 122.8 (d), 125.5 (d), 126.8 (d), 127.5 (2 × d), 127.7 (d),
128.3 (2 × d), 136.1 (s), 137.6 (s).
3
4
3J 3.6, 4Ј-H), 7.34 (1 H, dd, J 3.6, J 1.1, 3Ј-H), 7.40 (4 H, m,
5Ј-H and Ph), 7.78 (2 H, dt, 3J 6.7, 4J 1.7, ortho-Hphenyl);
δC(CDCl3) 17.0 (q), 17.8 (q), 25.5 (t), 25.7 (t), 28.4 (t), 48.7 (d),
51.7 (d), 66.1 (s), 96.8 (s), 98.1 (s), 124.8 (d), 125.3 (d), 127.1 (d),
127.4 (2 × d), 127.8 (d), 128.3 (2 × d), 129.5 (s), 135.9 (s).
(1á,4á,4aá,7aá)-4,4a,5,6,7,7a-Hexahydro-8,8-dimethyl-1-(3-
furyl)-4-phenyl-1,4-methano-1H-cyclopenta[d]pyridazine (1m).
Colourless needles (125 mg, 82%), mp 98–99 ЊC (decomp.)
692
J. Chem. Soc., Perkin Trans. 2, 1997

View Article Online
(Found: C, 78.23; H, 7.30; N, 8.85. C20H22N2O requires C,
78.40; H, 7.24; N, 9.14%); λmax(C6H6)/nm 361 (log ε 2.52);
νmax(KBr)/cmϪ1 3060, 3020, 2970, 1570, 1470, 1440; δH (CDCl3)
0.26 (3 H, s, endo-CH3), 0.97 (3 H, s, exo-CH3), 1.44 (6 H, m,
CH2), 3.15 (1 H, m, 7a-H), 3.45 (1 H, m, 4a-H), 6.62 (1 H, m, 4Ј-
H), 7.45–7.54 (4 H, m, 5Ј-H and Ph), 7.74 (2 H, m, ortho-
Hphenyl), 7.81 (1 H, m, 2Ј-H); δC(CDCl3) 17.0 (q), 17.8 (q), 25.5
(t), 25.6 (t), 28.7 (t), 48.6 (d), 49.8 (d), 65.2 (s), 94.0 (s), 97.3 (s),
109.6 (d), 121.1 (s), 127.4 (2 × d), 127.7 (d), 128.3 (2 × d), 136.1
(s), 140.6 (d), 143.2 (d).
59 ЊC (decomp.) (Found: C, 75.47; H, 7.26; N, 12.79.
C21H23N3O requires C, 75.65; H, 6.95; N, 12.60%); λmax(C6H6)/
nm 362 (log ε 1.89); νmax(BKr)/cmϪ1 3000, 2920, 1460, 1460,
1280, 1230 (N᎐O); δH (CDCl3) 0.18 (3 H, s, endo-CH3), 1.01 (3 H,
s, exo-CH3), 1.46 (6 H, mc, CH2), 3.40 (1 H, m, 7a-H), 3.48 (1 H,
m, 4a-H), 7.27–7.54 (3 H, m, Ph), 7.67–7.74 (4 H, m, 3Ј-H,
ortho-Hphenyl), 8.33 (2 H, dd, 3J 5.2, 4J 2.0, 2Ј-H); δC(CDCl3) 16.8
(q), 17.8 (q), 25.3 (2 × t), 28.5 (t), 48.9 (d), 49.4 (d), 66.7 (s), 96.3
(s), 99.1 (s), 124.9 (2 × d), 127.4 (2 × d), 128.1 (s), 128.5 (2 × d),
135.4 (s), 136.4 (s), 139.2 (2 × d).
(1á,4á,4aá,7aá)-4,4a,5,6,7,7a-Hexahydro-8,8-dimethyl-1-(N-
oxidopyridinium-3-yl)-4-phenyl-1,4-methano-1H-cyclopenta[d]-
pyridazine (1j). Colourless powder (32.8 mg, 97%), mp 132–
133 ЊC (decomp.) (Found: C, 75.31; H, 7.27; N, 12.87.
C21H23N3O requires C, 75.65; H, 6.95; N, 12.60%); λmax(C6H6)/
nm 361 (log ε 1.38); νmax(KBr)/cmϪ1 3020, 2920, 1570, 1460,
1280 (N᎐O), 1200; δH (CDCl3) 0.19 (3 H, s, endo-CH3), 1.03 (3 H,
s, exo-CH3), 1.54 (6 H, mc, CH2), 3.34 (1 H, dt, 3J 8.7, 3J 6.2, 7a-
Preparation of the azoalkanes 1d–e and h
To a solution of the azoalkanes 1g,i or k (15.6 mg, 50.0 µmol) in
dry diethyl ether (20 cm3) were added 7.17 mg (50.0 µmol) of
70% perchloric acid and the mixture stirred for 15 min at ca.
20 ЊC. The precipitate was then collected by filtration and
washed twice with cold diethyl ether (10 cm3).
(1á,4á,4aá,7aá)-4,4a,5,6,7,7a-Hexahydro-8,8-dimethyl-1-
(pyridinium-4-yl)-4-phenyl-1,4-methano-1H-cyclopenta[d]-
pyridazine perchlorate (1d). Pale-yellow powder (20.0 mg, 96%),
mp 110–111 ЊC (decomp.) (Found: C, 60.31; H, 6.16; N,
10.08. C21H24ClN3O4 requires C, 60.36; H, 5.79; N, 10.06%);
λmax(C6H6)/nm 364 (log ε 2.00); νmax(KBr)/cmϪ1 3200, 3100,
2910, 1480, 1100, 1040; δH (CD3CN) 0.12 (3 H, s, endo-CH3),
1.04 (3 H, s, exo-CH3), 1.49 (6 H, mc, CH2), 3.71 (2 H, mc, 4a-H
and 7a-H), 7.45–7.58 (3 H, m, Ph), 7.72–7.78 (2 H, m, ortho-
Hphenyl), 8.43 (2 H, d, 3J 6.9, 3Ј-H), 8.80 (2 H, br s, 2Ј-H), 9.24 (1
H, br s, NH); δC(CD3CN) 17.0 (q), 18.0 (q), 25.8 (2 × t), 29.3 (t),
49.8 (d), 50.6 (d), 69.1 (s), 97.7 (s), 101.2 (s), 127.1 (2 × d), 128.7
(2 × d), 129.3 (2 × d), 129.5 (2 × d), 135.9 (s), 142.5 (s, 2 × d).
(1á,4á,4aá,7aá)-4,4a,5,6,7,7a-Hexahydro-8,8-dimethyl-1-
(pyridinium-2-yl)-4-phenyl-1,4-methano-1H-cyclopenta[d]-
pyridazine perchlorate (1e). Pale-yellow powder (20.0 mg, 96%),
mp 104–105 ЊC (decomp.) (Found: C, 60.33; H, 6.22; N,
9.88. C21H24ClN3O4 requires C, 60.36; H, 5.79; N, 10.06%);
λmax(C6H6)/nm 363 (log ε 1.99); νmax(KBr)/cmϪ1 3200, 3040,
3020, 2920, 2820, 1580, 1500, 1450, 1080; δH (CD3CN) 0.19 (3
H, s, endo-CH3), 1.12 (3 H, s, exo-CH3), 1.52 (6 H, mc, CH2),
3.74 (2 H, mc, 4a-H and 7a-H), 7.46–7.59 (3 H, m, Ph), 7.76 (2
H, dd, 3J 6.6, 4J 1.7, 2Љ-H), 8.03 (1 H, t, 3J 7.9, 3J 5.0, 5Ј-H), 8.23
(1 H, d, 3J 7.9, 3Ј-H), 8.59 (1 H, t, 3J 7.5, 4J 2.0, 4Ј-H), 8.87 (1 H,
3
3
H), 3.55 (1 H, dt, J 8.9, J 6.2, 4a-H), 7.41–7.51 (4 H, m, Ph),
7.70 (2 H, dt, 3J 6.2, 4J 2.0, ortho-Hphenyl), 7.79 (1 H, ddd, 3J 8.0,
4J 1.6, 5J 1.1, 4Ј-H), 8.27 (1 H, ddd, 3J 6.4, 4J 1.7, 5J 1.0, 6Ј-H),
4
5
8.62 (1 H, t, J 1.4, J 1.1, 2Ј-H); δC(CDCl3) 16.8 (q), 17.8 (q),
25.3 (2 × t), 28.5 (t), 48.7 (d), 49.4 (d), 66.6 (s), 95.3 (s), 99.1 (s),
125.6 (d), 125.7 (d), 127.4 (2 × d), 128.1 (d), 128.5 (2 × d), 135.0
(s), 136.6 (s), 138.1 (d), 138.4 (d).
(1á,4á,4aá,7aá)-4,4a,5,6,7,7a-Hexahydro-8,8-dimethyl-1-(4-
pyridyl)-4-phenyl-1,4-methano-1H-cyclopenta[d]pyridazine-1-
oxide (5i) and (1á,4á,4aá,7aá)-4,4a,5,6,7,7a-hexahydro-8,8-
dimethyl-1-(4-pyridyl)-4-phenyl-1,4-methano-1H-cyclopenta-
[d]pyridazine-4-oxide (5iЈ). To a solution of the azoalkane 1i
(31.2 mg, 100 µmol) in dry methylene chloride (20 cm3) were
added 6.83 mg (110 µmol) of dimethyldioxirane as acetone solu-
tion 10 and stirred for 18 h at ca. 20 ЊC. Removal of the solvent at
water aspirator pressure (ca. 20 ЊC/Torr) afforded the azoxy
derivatives 5i/5iЈ as a colourless powder (33.1 mg, 99%), mp
134–135 ЊC (decomp.) (Found: C, 75.51; H, 7.11; N, 12.74.
C21H23N3O requires C, 75.65; H, 6.95; N, 12.60%); λmax(C6H6)/
nm 362 (log ε 1.89); νmax(KBr)/cmϪ1 3040, 2940, 1560, 1550,
1490 (N᎐O), 1460, 1420; δH of 5i (CDCl3) 0.67 (3 H, s, endo-
CH3), 1.05 (3 H, s, exo-CH3), 1.70 (6 H, mc, CH2), 3.72 (1 H, dt,
3
3
3
3J 8.6, J 6.1, 4a-H), 4.28 (1 H, dt, J 8.6, J 4.9, 7a-H), 7.26–
7.86 (8 H, m, 8 H, arom. H), 8.66 (1 H, dt, 3J 5.2, 4J 2.0, 6Ј-H),
δC of 5i (CDCl3) 17.9 (q), 18.2 (q), 25.6 (t), 26.4 (t), 28.0 (t),
47.1 (d), 48.5 (d), 65.8 (s), 83.3 (s), 102.6 (s), 123.4 (d), 125.7
(2 × d), 127.5 (d), 127.9 (d), 128.0 (2 × d), 134.9 (s), 135.6 (d),
148.8 (d), 150.4 (s); δH of 5iЈ (CDCl3) 0.71 (3 H, s, endo-CH3),
1.12 (3 H, s, exo-CH3), 1.69 (6 H, mc, CH2), 3.89 (1 H, mc, 4a-
H), 4.03 (1 H, mc, 7a-H), 7.26–7.86 (8 H, m, arom. H), 8.67 (1
3
4
dd, J 5.0, J 1.0, 6Ј-H), 9.22 (1 H, br s, NH); δC(CD3CN) 17.0
(q), 18.1 (q), 25.8 (t), 25.9 (t), 29.1 (t), 49.7 (d), 51.0 (d), 69.4 (s),
97.0 (s), 101.7 (s), 126.7 (d), 127.4 (d), 128.7 (2 × d), 129.4 (d),
129.6 (2 × d), 135.5 (s), 144.5 (d), 147.5 (d), 152.6 (s).
(1á,4á,4aá,7aá)-4,4a,5,6,7,7a-Hexahydro-8,8-dimethyl-1-
(pyridinium-3-yl)-4-phenyl-1,4-methano-1H-cyclopenta[d]-
pyridazine perchlorate (1h). Colourless powder (19.6 mg, 94%),
mp 98–99 ЊC (decomp.) (Found: C, 60.19; H, 6.09; N, 9.88.
C21H24ClN3O4 requires C, 60.36; H, 5.79; N, 10.06%);
λmax(C6H6)/nm 363 (log ε 1.99); νmax(KBr)/cmϪ1 3200, 3020,
2920, 2880, 1440, 1080, 1010; δH (CD3CN) 0.14 (3 H, s, endo-
CH3), 1.01 (3 H, s, exo-CH3), 1.46 (6 H, mc, CH2), 3.68 (2 H, mc,
4a-H and 7a-H), 7.45–7.58 (3 H, m, Ph), 7.77 (2 H, dt, 3J 6.6, 4J
3
4
H, dt, J 5.0, J 1.8, 6Ј-H); δC of 5iЈ (CDCl3) 17.9 (q), 18.7 (q),
25.5 (t), 26.2 (t), 28.0 (t), 47.0 (d), 48.9 (d), 66.0 (s), 83.5 (s),
102.8 (s), 122.4 (d), 123.3 (d), 128.3 (2 × d), 128.9 (d), 129.3
(2 × d), 129.6 (s), 136.3 (d), 149.1 (d), 155.9 (s).
5
3
4
4
Acknowledgements
1.7, J 1.7, 2Љ-H), 8.10 (1 H, ddd, J 8.1, J 5.8, J 0.7, 5Ј-H),
4
8.81 (2 H, m, arom. H), 9.13 (2 H, d, J 2.0, 2Ј-H and NH);
We thank the Deutsche Forschungsgemeinschaft and the Fonds
der Chemischen Industrie for generous financial support and
Birgit Beck for technical assistance.
δC(CD3CN) 17.0 (q), 17.8 (q), 25.8 (t), 25.9 (t), 29.2 (t), 49.6 (d),
50.0 (d), 68.0 (s), 96.5 (s), 100.6 (s), 128.3 (d), 128.6 (2 × d),
129.2 (d), 129.5 (2 × d), 136.2 (s), 137.8 (s), 142.2 (d), 143.0 (d),
145.7 (d).
References
Preparation of the azoalkanes 1a and j
1 (a) D. A. Dougherty, in Kinetics and Spectroscopy of Carbenes and
Diradicals, ed. M. S. Platz, Plenum, New York, 1990; (b) J. A. Weil,
J. R. Bolton and J. E. Wertz, Electronic Paramagnetic Resonance:
Elementary Theory and Practical Applications, Wiley, New York,
1994.
2 (a) W. Adam, F. Kita, H. M. Harrer, W. M. Nau and R. Zipf, J. Org.
Chem., 1996, 61, 7056; (b) W. Adam, H. M. Harrer, F. Kita and
W. M. Nau, Pure Appl. Chem., 1997, in the press; (c) W. Adam,
H. M. Harrer, T. Heidenfelder, T. Kammel, F. Kita, W. M. Nau and
C. Sahin, J. Chem. Soc., Perkin Trans. 2, 1996, 2085.
To a solution of the azoalkanes 1g or k (31.2 mg, 100 µmol) in
dry methylene chloride (20 cm3) were added 6.83 mg (110 µmol)
of dimethyldioxirane as acetone solution 10 and stirred for 1 h at
ca. 20 ЊC. Removal of the solvent at water aspirator pressure
(ca. 20 ЊC/15 Torr) afforded the pure derivatives 1a and j.
(1á,4á,4aá,7aá)-4,4a,5,6,7,7a-Hexahydro-8,8-dimethyl-1-(N-
oxidopyridinium-4-yl)-4-phenyl-1,4-methano-1H-cyclopenta[d]-
pyridazine (1a). Colourless powder (33.1 mg, 99%), mp 58–
J. Chem. Soc., Perkin Trans. 2, 1997
693

View Article Online
3 (a) D. R. Arnold, in Substituent Effects in Radical Chemistry, NATO
ASI ser. Ser. C, Vol. 189, ed. H. G. Viehe, Z. Janousek and
R. Meréni, Riedel & Dordrecht, Netherlands, 1986, pp. 171–188; (b)
R. A. Jackson and M. Sharifi, J. Chem. Soc., Perkin Trans. 2, 1996,
776.
10 W. Adam, J. Bialas and L. Hadjiarapoglou, Chem. Ber., 1991, 124,
2377.
11 (a) H. G. Aurich and H. Czepluch, Tetrahedron, 1980, 36, 3543; (b)
H. G. Aurich, in Nitrones, Nitronates and Nitroxides, ed. S. Patai
and Z. Rappoport, Wiley, New York, 1989, ch. 4, p. 313.
12 M. M. Alam, A. Watanabe and O. Ito, Photochem. Photobiol., 1996,
63, 53.
4 R. A. Jackson and C. J. Rhodes, J. Chem. Soc., Perkin Trans. 2,
1985, 121.
5 (a) R. V. Lloyd and D. E. Wood, Mol. Phys., 1971, 20, 735; (b)
D. M. Hirst, R. F. Treeweek and C. A. Vilain, Chem. Phys., 1973, 1,
149.
13 (a) M. J. S. Dewar and A. J. Holder, Heterocycles, 1989, 28, 1135;
(b) C. W. Bird, Tetrahedron, 1992, 48, 335; Tetrahedron, 1993, 49,
8441.
6 C. J. Rhodes and E. Roduner, J. Chem. Soc., Chem. Commun., 1988,
1227.
7 R. S. Hutton and H.-D. Roth, J. Am. Chem. Soc., 1982, 104, 7395.
8 (a) W. Lau, J. Bargon and A. C. Ling, J. Magn. Reson., 1980, 38, 93;
(b) R. J. Waltman, A. C. Ling and J. Bargon, J. Magn. Reson., 1982,
48, 314.
14 (a) J. J. P. Stewart, J. Comput. Chem., 1989, 10, 209; (b) T. Kovàr,
Diploma Thesis, University of Erlangen-Nürnberg, 1987.
15 G. Rauhut, J. Chandrasekhar, A. Alex, T. Steinke and T. Clark,
VAMP 5.0, University of Erlangen-Nürnberg, 1993.
9 (a) W. Adam, H. M. Harrer, W. M. Nau and K. Peters, J. Org.
Chem., 1994, 59, 3786; (b) W. Adam, H. M. Harrer and T. Heiden-
felder, unpublished work; (c) K. H. Beck and S. Hünig, Chem. Ber.,
1987, 120, 477.
Paper 6/07513D
Received 4th November 1996
Accepted 4th December 1996
694
J. Chem. Soc., Perkin Trans. 2, 1997