Experimental and theoretical characterization of the valence isomerization of Bi-2H-azirin-2-yls to diazabenzenes

Article abstract of DOI:10.1002/chem.200600318

3,4-Diazidocyclobutenes 16 were prepared from the corresponding dihalides. Some of these diazides, such as parent compound 16 d and phenyl-substituted derivatives 16c,f, underwent spontaneous stereoselective electrocyclic ring opening below room temperature, whereas the tetraalkyl derivatives of 16 had to be heated to force the same reaction. In most cases, the resulting 1,4-diazidobuta-1,3-dienes 8 were isolated to study their photochemical transformation into bi-2Hazirin-2-yls 9 via intermediate monoazirines 17. Except for starting materials with a low number of substituents such as 9d and 9f, title compounds 9 underwent a thermal valence isomerization which led exclusively to pyridazines 18 at surprisingly low temperatures. Based on quantum-chemical calculations for the parent bi-2H-azirinyl 2-yl 9d at the UB3LYP/6-31+G(d) and MR-MP2/TZV(2df,2p) levels, the valence isomerization process is best explained by simultaneous homolytic cleavage of both C-N single bonds of 9 to generate energetically favorable N,N′ diradicals 26, which cyclize to 18. The theoretical studies indicate also that one stereoisomer of 9, namely, the rac compound, should undergo valence isomerization more easily than the other, which is in conformity with different rates of these rearrangement reactions found experimentally. For the tetramethyl-bi-2H-azirin-2-yls 9g, which are better models for the experimentally studied compounds, simultaneous homolytic cleavage of both C-N single bonds is also predicted by the calculations, although the intermediate diradicals 26 g are significantly higher in energy than those of the parent system 9d.

Full text of DOI:10.1002/chem.200600318

FULL PAPER
DOI: 10.1002/chem.200600318
Experimental and Theoretical Characterization of the Valence Isomerization
of Bi-2H-azirin-2-yls to Diazabenzenes**
Klaus Banert,*[a] Stefan Grimme,*[b] Rainer Herges,*[c] Kirsten Heß,[c] Frank Kçhler,[a]
Christian Mück-Lichtenfeld,[b] and Ernst-Ulrich Würthwein*[b]
Abstract: 3,4-Diazidocyclobutenes 16
were prepared from the corresponding
dihalides. Some of these diazides, such
as parent compound 16d and phenyl-
substituted derivatives 16c,f, under-
went spontaneous stereoselective elec-
trocyclic ring opening below room tem-
perature, whereas the tetraalkyl deriva-
tives of 16 had to be heated to force
the same reaction. In most cases, the
underwent a thermal valence isomeri-
zation which led exclusively to pyrida-
zines 18 at surprisingly low tempera-
tures. Based on quantum-chemical cal-
culations for the parent bi-2H-azirinyl-
2-yl 9d at the UB3LYP/6-31+G(d)
N,N’ diradicals 26, which cyclize to 18.
The theoretical studies indicate also
that one stereoisomer of 9, namely, the
rac compound, should undergo valence
isomerization more easily than the
other, which is in conformity with dif-
ferent rates of these rearrangement re-
actions found experimentally. For the
and MR-MP2/TZV(2df,2p) levels, the
A
valence isomerization process is best
explained by simultaneous homolytic
tetramethyl-bi-2H-azirin-2-yls
9g,
ꢀ
cleavage of both C N single bonds of 9
which are better models for the experi-
mentally studied compounds, simulta-
resulting 1,4-diazidobuta-1,3-dienes
8
to generate energetically favorable
ꢀ
were isolated to study their photo-
chemical transformation into bi-2H-
azirin-2-yls 9 via intermediate mono-
azirines 17. Except for starting materi-
als with a low number of substituents
such as 9d and 9 f, title compounds 9
neous homolytic cleavage of both C N
single bonds is also predicted by the
calculations, although the intermediate
diradicals 26g are significantly higher
in energy than those of the parent
system 9d.
Keywords: azides
·
density
functional calculations · nitrogen
heterocycles · reactive intermedi-
ates · small ring systems
Introduction
The extremely strained bicycloprop-2-enyl 5a was isolated
first by Billups et al.[1] in 1989 as the last remaining valence
isomer of benzene (1). This compound, which was calculated
to be the highest in energy of the (CH)6 species 1–4 and
5a,[2,3] polymerizes above ꢀ108C.[1] Thirty years before, Bre-
slow et al.[4] discovered that the highly substituted derivative
5b undergoes valence isomerization to give hexaphenylben-
zene on heating. Less substituted starting materials, for ex-
[a] Prof. Dr. K. Banert, Dr. F. Kçhler
Institut für Chemie der Technischen Universität Chemnitz
09107 Chemnitz (Germany)
Fax : (+49)371-531-1839
E-mail: klaus.banert@chemie.tu-chemnitz.de
[b] Prof. Dr. S. Grimme, Dr. C. Mück-Lichtenfeld,
Prof. Dr. E.-U. Würthwein
Institut für Organische Chemie der Universität Münster
Corrensstrasse 40, 48149 Münster (Germany)
Fax : (+49)251-83-39772
E-mail: grimmes@uni-muenster.de
Wurthwe@uni-muenster.de
[c] Prof. Dr. R. Herges, K. Heß
Institut für Organische Chemie der Universität Kiel
Otto-Hahn-Platz 4, 24098 Kiel (Germany)
Fax : (+49)431-880-1558
E-mail: rherges@oc.uni-kiel.de
[**] Reactions of Unsaturated Azides, part 20. For part 19, see K. Banert,
B. Meier, Angew. Chem. 2006, 118, 4120–4123; Angew. Chem. Int Ed.
2006, 45, 4015–4019.
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
Chem. Eur. J. 2006, 12, 7467 – 7481
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7467

ample, meso-5c, rac-5c, and 5d, rearrange at 160–2008C in
the gas phase each to give mixtures of o-, m-, and p-xy-
lenes.[5] Several mechanisms were proposed to explain these
aromatization reactions,[5b,6] which are accompanied by
Cope rearrangements such as the isomerization 5d!5c.[5c]
Up to now, all attempts to prepare the diaza analogues of
5, namely, the heterocycles 9, failed (Scheme 1).[7] Neither
the double Neber reaction of hydrazonium salts 6a,b or sim-
Thus, we planned to generate these compounds by conrota-
tory ring opening of cyclobutenes of type 16 (Scheme 2). To
prepare such 3,4-diazidocyclobutenes, we used the known
Scheme 1.
ilar starting materials nor the coupling reaction of 2-chloro-
2H-azirine 7b with various reagents generated the bi-2H-
azirin-2-yls 9.[8a] Later, Storr and Gallagher[9] reported the
treatment of 7c with lithium in THF, which was said to lead
to pyrimidine 10c (10% yield) and pyrazine 11c (10%) but
not to tetraphenylpyridazine (18c). The authors postulated
the intermediate 9c, which could not be observed. Attempts
to generate 9a from diazide 8a failed immediately because
the starting material 8a could not be prepared.[5] To the best
of our knowledge, 1,4-diazidobuta-1,3-dienes are un-
known.[7]
Nevertheless, we consider access via azides to be the most
promising approach to investigate highly strained heterocy-
cles of type 9 and their valence isomerization to yield diaza-
benzenes. If compounds 9 are unstable at room temperature,
the reaction 8!9 can be performed not only by thermolysis
but also by photolysis at low temperature. We describe here
the synthesis of diazides 8c–g,j,k and their transformation
into biazirinyls 9c,d,f,g,j,k, as well as their aromatization
and the theoretical characterization of the last-named proc-
ess.
Scheme 2.
dihalides 15c[12] (X=Br), cis-15d (X=Cl), trans-15d[13] (X=
Cl, Br, I), a mixture[14] of the dibromides 15e, 15e’, and
15e’’, or 15 f[15] (X=Br) as well as 15g,[16] 15i,[17] 15j,[18] and
15k[18] (all X=Cl). Alternatively to the procedure of Brune
et al.[18,19] via bis(methylene)cyclobutenes, some of these
known and several new compounds were accessible more
conveniently from alkynes 12 via cyclobutenediones 13[20]
and diols 14,[21] which were transformed into dihalides 15
(X=Cl, Br, I) on treatment with SOCl2, PBr3, or PI3. We
prepared 15 f (X=I), 15h (X=Br), and 15i (X=Cl, Br) as
pure substances, whereas 15j (X=Cl, Br) and 15k (X=Cl,
Br) resulted each as mixtures with dihalides 15j’ or 15k’ and
15k’’, respectively. The equilibration of 3,4-dihalocyclobut-1-
enes substituted with alkyl groups is notorious as a rapid
process.[18,19] The synthesis of 15c (X=Br) was performed
by a different literature method.[12]
Results and Discussion
When the commercially available dichloride cis-15d[22]
was treated with NaN3, LiN3, or tributylhexadecylphospho-
nium azide (QN3)[23] under different conditions, we got only
mixtures of cis-16d, cis-19d, (E,E)-20d, and 18d (Scheme 3,
Table 3). After isolation of the diazide cis-16d, we were not
Results in the literature[8,10] but also of our own efforts indi-
cated quickly that classical methods[11] to prepare vinyl
azides are unsuited to the synthesis of diazides of type 8.
7468
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 7467 – 7481

Valence Isomerization of Bi-2H-azirin-2-yls to Diazabenzenes
FULL PAPER
toring in an amount of up to 60%. The best yield (70%)
and highest purity of (E,E)-8 f were achieved when diiodide
15 f had been treated with QN3.
On photolysis in dichloromethane or chloroform at ꢀ50
to ꢀ858C, diazides (E,E)-8d, (E,E)-8e, and (E,E)-8 f led to
mono-2H-azirines (E)-17d (up to 60%), (E)-17e (up to
37%), and (E)-17 f (up to 18%), respectively (Scheme 2).
At first, we achieved the formation of bi-2H-azirin-2-yls 9
only in the case of irradiation of (E,E)-8 f. Thus, both dia-
stereomers of 9 f were generated in 20% yield in a ratio of
1.2:1. Small amounts of the very unstable compound 9d
(8% yield, 3:2 mixture of diastereomers) could only be de-
1
tected by H and 13C NMR spectroscopy when (E,E)-8d was
irradiated in the presence of the sensitizing agent 9,10-dicya-
noanthracene in chloroform at ꢀ508C. On warming this
mixture of products from photolysis, 9d degraded to furnish
at best traces of 18d. Different stabilities of the diastereom-
ers of 9 f were observed by NMR spectroscopy at room tem-
perature. The minor isomer of 9 f decomposed completely
after a few minutes, whereas the major isomer could be de-
tected even after several hours. Just like hydrocarbon 5a, 9d
and also 9 f did not tend to aromatization.
Scheme 3.
able to perform electrocyclic ring opening by heating a solu-
tion of this compound. However, the reaction of the dibro-
mide or the analogous diiodide trans-15d with a concentrat-
ed solution of QN3 in chloroform gave, in addition to small
amounts of the diazide cis-16d (7 and 6%, respectively),
only the diazide (E,E)-8d (89 and 62%, respectively). Clear-
ly, ring opening of trans-16d was very rapid, and therefore
this intermediate was not detected by NMR spectroscopic
monitoring of the substitution reaction. Nevertheless, the
products of monosubstitution, 3-azido-4-bromocyclobut-1-
ene and the analogous iodo compound, could be observed
as intermediates. Evidently, the azido groups accelerate elec-
trocyclic ring opening as effectively as other donor substitu-
ents such as alkoxyl groups. As shown by the work of
Kirmse, Houk, and Rondan,[24] the maximum effect could be
expected if the conrotatory process allows both donors to
rotate “outward”. The parent diazide (E,E)-15d could not
be prepared with the help of alternative reagents, for exam-
ple, sodium azide in DMSO, or by starting with less reactive
dihalides. On treatment with QN3 at 208C, dichloride trans-
15d gave a mixture of (E,E)-20d (62% yield) and 18d
(10%). Obviously, the second nucleophilic substitution
could not compete with the rapid conrotatory ring opening
generating monoazide (E,E)-20d.
However, different reaction conditions, such as sodium
azide or lithium azide in DMSO or QN3 in chloroform,
proved to be appropriate to transform a mixture of dibro-
mide 15e and allyl isomers 15e’ and 15e’’ into the open-
chain product (E,E)-8e (34–37%) in addition to cis-16e
(yield ca. 16%). Similar procedures were not optimal in the
case of starting material 15 f (X=Br) because the synthesis
of (E,E)-8 f is accompanied by formation of monoazide 22 f
and its succeeding product 23 f. The intermediate 21 f, which
is generated by SN2’ reaction from dibromide 15 f and is a
precursor of 22 f, was detected by NMR spectroscopic moni-
To synthesize more stable diazides of type 8 and bi-2H-
azirin-2-yls 9, we tried to prepare tetrasubstituted 3,4-diazi-
docyclobut-1-enes 16. Unfortunately, even when different
reagents and conditions were used, dichloride 15i as well as
dibromides 15h and 15i could not be transformed into dia-
zides, because elimination instead of substitution occurred.
For example, the reaction of 15i (X=Br) with QN3 pro-
duced monoazide 24i (18% yield) and the known hydrocar-
bon 25i[25] (48%). However, treatment of trans-15g (X=Cl)
with a concentrated solution of QN3 in chloroform at 208C
gave a mixture of diazides trans-16g (55% yield) and cis-
16g (9%) as well as the monoazides 19g (5%, 3:2 mixture
of diastereomers) and 24g (17%). After separation by chro-
matography, trans-16g could be isolated, since the conrota-
tory electrocyclic ring opening to afford (E,E)-8g is strongly
decelerated by inward rotation of the methyl groups. The
target compound (E,E)-8g was generated only on warming
in benzene (808C), whereby the secondary products (E)-17g
and known 18g[26] are also formed. Prolonged heating led
exclusively to 18g (88% yield). On incomplete thermolysis
of trans-16g, (E,E)-8g could be isolated in 14% yield based
on the amount of converted starting material. We were not
able to prepare 8g by ring opening of cis-16g.
The E configuration of (E)-17e,g was determined from
1
two-dimensional H NMR nuclear Overhauser spectroscopy
(2D-NOESY) experiments and nuclear Overhauser en-
hancement (NOE) difference spectra. Thus the configura-
tion of (E,E)-8e,g is also confirmed. The trans structure of
trans-16g was established by the signals of the 1H NMR
spectra, which exhibited very small low-field shifts in the
presence of [Eu(fod)3] (fod=1,1,1,2,2,3,3-heptafluoro-7,7-di-
G
methyloctane-4,6-dionato). The analogous spectra of the
compounds cis-16d,e,g showed changes of the chemical
shifts at least 300 times greater, since these diazides are able
to interact with the shift reagent in a chelating[27] manner.
Chem. Eur. J. 2006, 12, 7467 – 7481
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7469

K. Banert, S. Grimme, R. Herges, E.-U. Würthwein et al.
The synthesis of more unsymmetrical tetraalkyl-substitut-
ed (E,E)-1,4-diazidobuta-1,3-dienes 8 was performed like
that of (E,E)-8g. Thus, mixtures of the chlorides 15j and
15j’ or, alternatively, mixtures of the corresponding bro-
mides were treated with QN3 to afford 1:1 mixtures of the
azides trans-16j and trans-16j’ in 73–75% yield. Both the
bromides 15k, 15k’, and 15k’’ as well as the analogous mix-
ture of allylic chlorides led similarly to a 1:2:1 mixture of
the azides trans-16k, trans-16k’, and trans-16k’’ with 51 and
62% yield, respectively. Partial separation of the constitu-
tional isomers of these trans-diazidocyclobutenes was possi-
ble. On heating, however, thermal electrocyclic ring opening
competed with scrambling of the allylic isomers by [3,3] sig-
matropic rearrangement[28] of the azido group. Therefore,
mixtures of (E,E)-8j and (E,E)-8j’ as well as mixtures of
(E,E)-8k, (E,E)-8k’, and (E,E)-8k’’, which unfortunately
both could not be separated, were produced.
and 18g were photochemically stable also under these con-
ditions. Similar prolonged photolysis of the diazides cis-16g
and trans-16g in CD3CN proceeded via 9g and resulted in
the formation of both aromatics 10g (15% in both cases)
and 18g (33 and 41%, respectively) as well as traces of but-
2-yne and acetonitrile.
To investigate whether the aromatization 9!18 is accom-
panied by the Cope rearrangement of 9, which would ex-
change the “inner” (R1,R2) and the “outer” substituents
(R3,R4), we included starting materials with different alkyl
substituents such as (E,E)-8j and (E,E)-8k. Because we
were only able to use (E,E)-8j as an inseparable mixture
with (E,E)-8j’ as well as (E,E)-8k as a mixture with (E,E)-
8k’ and (E,E)-8k’’, we could not prove strictly that the
transformation of one constitutional isomer of 8 led exclu-
sively to one isomer of 18 without any scrambling of the
substituents. However, the ratio (E,E)-8j:(E,E)-8j’ agreed
G
As expected, (E,E)-8g was transformed via (E)-17g into
18g with high yield (97%) on thermolysis. However, the
photochemical reaction of (E,E)-8g in CDCl3 at ꢀ608C,
which proceeded also via (E)-17g with a proportion of up to
61%, led quantitatively to a 1:1 mixture of meso-9g and
rac-9g. On warming the irradiated solution, both stereo-
isomers were converted to 18g in quantitative first-order re-
actions. One of these valence isomerizations occurred al-
ready with k=3.65ꢁ10ꢀ4 sꢀ1 at ꢀ258C, and the other with
k=1.73ꢁ10ꢀ4 sꢀ1 at +108C. In the presence of silver(i) tet-
rafluoroborate at ꢀ258C, the more stable diastereomer of
9g was slowly (7 d) but completely transformed into
known[29] 10g, whereas the other stereoisomer gave 18g
again. A mechanism including the intermediates A, B, and
C may be responsible for formation of 10g (Scheme 4). Sim-
with the ratio 9j:9j’ after quantitative photochemical reac-
tion, and these ratios correspond also to the ratio 18j:18j’
after quantitative thermal aromatization. Concerning the
analogous quantitative transformations, the same held true
for the ratios (E,E)-8k:
N
G
1
and 18k:18k’:18k’’, easily measured by H NMR spectrosco-
py. Therefore, scrambling of the substituents during the iso-
merization 9!18 seems to be unlikely.
We monitored the reaction of 15c (X=Br) with QN3 in
CDCl3 by NMR spectroscopy at low temperature. Even in
the range of ꢀ50 to ꢀ258C, nucleophilic substitution and
electrocyclic ring opening proceeded so rapidly that only
some of the signals of 8c could be assigned with certainty.
At +58C, slow evolution of nitrogen started, in the course
of which one geometrical isomer of 17c with a maximum
amount of 87% and after that both stereoisomers of 9c,
generated in a ratio of 2:1, could be detected unequivocally.
This thermal formation of bi-2H-azirin-2-yls was not surpris-
ing, because there are several reports in the literature on
thermal transformation of other 1-azido-1,2-diphenylethenes
into 2H-azirines, which occurred already at low tempera-
tures.[30] After prolonged reaction times or after slight warm-
ing of the reaction mixture including 9c, only the known ar-
omatic compound 18c[31] was found with 87% NMR yield
based on dibromide 15c. Although we prepared the known
heterocycles 10c[32] and 11c[33] for comparison, they could
not be detected as thermal secondary products of 9c. This
should exclude the postulated[9] bi-2H-azirin-2-yl 9c being
intermediate during the transformation 7c!10c+11c as-
suming that the presence of the generated lithium chloride
did not change the thermal successive reactions of 9c dra-
matically. When a reaction mixture with a large amount of
8c resulting from 15c (X=Br) and QN3 was irradiated at
low temperature, pyrimidine 10c was formed in 44% yield,
whereas 11c and 18c could not be detected.
Scheme 4.
ilar mechanisms were discussed for the reaction of bicyclo-
prop-2-enyls to benzene derivatives. In the case of the latter
transformation, Dewar benzenes were proved as intermedi-
ates or even isolated.[1a,6b] Photolysis of (E,E)-8g in the
more UV transparent CD3CN using quartz equipment at
ꢀ408C furnished (E)-17g and 9g first and then, after pro-
longed irradiation, a mixture of 10g (40%) and 18g (22%)
as well as the fragmentation products but-2-yne (3–4%) and
acetonitrile (3–4%). As shown in control experiments, 10g
The thermal valence isomerization 9!18 showed some
remarkable singularities, which deserve special attention
with regard to its reaction mechanism. Compared to thermal
reactions of simple 2H-azirines,[34] the low temperature at
which the aromatization 9!18 could be performed was sur-
7470
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 7467 – 7481

Valence Isomerization of Bi-2H-azirin-2-yls to Diazabenzenes
FULL PAPER
prising, especially when compared to the temperatures
which were necessary for analogous valence isomerization
of bicycloprop-2-enyls.[5,6] The very different reactivities of
meso- and rac-9 or unlike- and like-9, respectively, are also
noteworthy. Furthermore, the exclusive thermal transforma-
tion of 9 to yield pyridazines 18 should be explained, since
bicycloprop-2-enyls such as 5c or 5d led to mixtures of o-,
m-, and p-xylenes.[5] The formation of pyridazines from two
molecules of simple 2H-azirines was observed only in rare
cases and with very low yield.[35] On the other hand, similar
dimerizations to yield pyrazine derivatives were reported
several times.[34]
biazirinyls 9g. All optimizations were performed at the
UB3LYP/6-31+G* level using the Gaussian98 program,[36]
and the stationary points were characterized as transition
states or minima by harmonic frequency analyses. All re-
ported relative energies are corrected for zero-point vibra-
tional energies (ZPVE) as obtained from the DFT calcula-
tions (unscaled).
In the DFT calculations, the guess=mix keyword was em-
ployed to enforce convergence of the SCF procedure to a
spin- and spatial-symmetry broken solution (BS-UDFT).
For the biradicals or transition states with significant biradi-
cal character, hS2i values between 0.4 and 1.1 were found
(see Table 1).[37,38] In view of the thus-introduced uncertainty
and the complexity of the electronic structure problem, we
decided to check the most relevant species by single-point
calculations at the MR-MP2[39] and UCCSD(T) levels. Note
that the Hartree–Fock solutions (on which CCSD(T) is
based) suffer even more from spin contamination than DFT,
as is evident from hS2i values between 1.22 and 1.98. We
also applied the spin-correction procedure proposed by Ya-
Quantum-Chemical Calculations
To elucidate the mechanisms of the thermal biazirinyl ring-
opening processes and to understand the experimental re-
sults, we investigated parts of the energy hypersurface of the
parent system 9d and that of the tetramethyl-substituted
Table 1. Relative energies Erel(0 K) [kcalmolꢀ1] with respect to rac-syn-9d from UB3LYP/6-31+G*, spin-corrected (SC) UB3LYP/6-31+G*,[40] MR-MP2/
TZV
(2df,2p), and UCCSD(T)/TZVP calculations, and hS2i values from UB3LYP and UHF calculations (which are the basis for the UCCSD(T) treat-
ment) for the species 9d, 18d, 26d, 27d, 28d, 29d, and the related transition states.
ACHTREUNG
Erel(0 K)
(UB3LYP/6-
31+G*)
hS2i
Erel(0 K)
Erel(0 K)
[MR-MP2/TZV
(2df,2p)]
Erel(0 K)
[UCCSD(T)/TZVP-
(d,p)]
hS2i
(UB3LYP) [UB3LYP/6-31+G*
G
A
(SC)]
rac-syn-9d
rac-anti-9d
0.0
0.1
2.1
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
1.0429
1.0437
2.0718
0.4431
0.4271
0.5256
1.0244
1.1014
1.0148
0.6658
0.7337
0.6705
0.6433
1.0200
1.0201
0.9938
0.0000
1.0253
1.0507
1.0562
1.0229
1.0261
0.9587
0.9391
0.9914
1.0006
0.0000
0.5047
0.0
0.0
0.6
0.0
0.0000
TS rac-syn-9d!rac-anti-9d
meso-gauche-9d
1.2
meso-anti-9d
ꢀ0.1
TS meso-gauche-9d!meso-anti-9d
1.2
18d
ꢀ70.8
ꢀ9.2
ꢀ11.7
29.0
17.4
18.6
19.9
ꢀ2.1
16.1
15.1
19.7
19.5
20.0
21.4
15.8
14.7
21.9
ꢀ62.6
ꢀ4.9
ꢀ9.1
ꢀ7.7
ꢀ5.7
ꢀ3.4
20.7
23.2
26.2
28.4
ꢀ13.2
22.9
ꢀ63.3
ꢀ6.4
ꢀ6.4
cis-26d
trans-26d
ꢀ9.7
0.3
1.9880
TS cis-26d!trans-26d
TS rac-syn-9d!cis-26d
TS rac-anti-9d!trans-26d
TS meso-anti-9d!trans-26d
TS cis-26d!18d
11.6
13.1
13.4
14.8
24.4
26.0
1.5461
1.5476
ꢀ2.2
22.4
20.8
3.6
27.0
3.2
29.7
22.6
1.8060
1.1790
1.2615
syn-27d
anti-27d
TS rac-anti-9d!anti-27d
TS rac-syn-9d!syn-27d
TS meso-anti-9d!anti-27d
TS meso-gauche-9d!syn-27d
TS syn-27d!cis-26d
TS anti-27d!trans-26d
TS syn-27d!anti-27d
28d
16.3
20.6
25.3
26.4
25.8
23.2
1.2278
1.4195
TS cis-26d!28d
ꢀ4.9
1.1
ꢀ3.4
1.5
1.8250
trans-26d’
trans-26d’’
TS trans-26d!trans-26d’
TS trans-26d’!trans-26d’’
s-trans-29d+HCN
s-cis-29d+HCN
TS trans-26d’!s-trans-29d+HCN
TS trans-26d’’!s-cis-29d+HCN
HCCH+2HCN
20.6
23.5
TS s-trans-29d+HCN!
18.2
25.9
HCCH+2HCN
TS trans-26d!HCCH+2HCN
35.4
0.0000
Chem. Eur. J. 2006, 12, 7467 – 7481
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7471

K. Banert, S. Grimme, R. Herges, E.-U. Würthwein et al.
maguchi, Jensen, Dorigo, and Houk,[40] using the UB3LYP
wavefunction. For diradicals (hS2iꢁ1), we found corrections
ranging from +0.1 to +6 kcalmolꢀ1, and for systems with
less spin contamination (hS2iꢁ0.5) corrections of ꢀ3.2 to
ꢀ5.5 kcalmolꢀ1 that disagree with the results from the MR-
MP2 method. Table 1 compares the relative energies from
the three mentioned methods for the selected set of station-
ary points. The qualitative picture of the energetics is the
same for all methods, although significant differences are
noted, for example, between DFT or MR-MP2 and
CCSD(T) for the transition state TS rac-anti-9d!trans-26d
as well as between DFT and MR-MP2 or CCSD(T) for syn-
27d and TS syn-27d!cis-26d. However, for the most im-
portant question whether the reaction proceeds via inter-
mediate 27d or directly from rac-9d to cis-26d, all methods
agree qualitatively, that is, the barrier rac-syn-9d!syn-27d
is higher than that between rac-syn-9d and cis-26d. If we
consider the MR-MP2 results as most accurate (because it is
not spin-symmetry broken and accounts for static and dy-
namic electron correlation effects), the difference between
these two barriers is more than 10 kcalmolꢀ1 in favor of the
concerted mechanism.
Parent bi-2H-azirin-2-yls 9d: As aforementioned, the biazir-
inyls exist in two diastereomeric forms (rac and meso), each
in two rotational conformations. The structures of the poten-
tial energy minima of the parent biazirinyls, C2-symmetric
rac-syn-9d and rac-anti-9d, as well as meso-gauche-9d (C1)
and meso-anti-9d (Ci), are plotted in Figure 1. These struc-
tures have similar energies (within 1.3 kcalmolꢀ1) and there
ꢀ
is only a small barrier for rotation around the central C C
single bond (about 2.1 kcalmolꢀ1, see Table 1).
Figure 1. UB3LYP/6-31+G* optimized structures for bi-2H-azirin-2-yls
9d and N,N’ diradicals 26d (bond lengths in ꢂ).
There are two obvious reaction paths to trigger thermal
ꢀ
isomerization of rac- and meso-9d: breaking either the C C
ꢀ
or the C N single bonds in the azirine rings. For the parent
ꢀ
ꢀ
azirine, it is experimentally well known that fission of a C
N single bond is much easier than C C bond breaking.
detect a reaction pathway with simultaneous C C breaking
of both azirine rings (see below).
[34]
ꢀ
ꢀ
Our calculations are well in accord with these observations.
To break the two C N single bonds in both azirine rings
ꢀ
For the thermal C C breakage of azirine to yield the nitrile
of 9d homolytically, there are in principle two reaction path-
ylide, an activation energy of 46.7 kcalmolꢀ1 (UB3LYP/6-
ways: breaking of the C N single bonds one after the other
ꢀ
31+G*, resulting in the closed-shell transition state) and
or simultaneous breaking of both bonds (two-stage or con-
certed). Both reactions lead to N,N’ diradicals. From rac-
syn-9d and meso-gauche-9d, the Z diradical cis-26d is
formed, while rac-anti-9d and meso-anti-9d give rise to the
E diradical trans-26d (Figure 1). With DH° in the range of
19.5–21.4 kcalmolꢀ1, the activation barriers for the first step
of the two-stage reaction pathway (9d!27d!26d) are
52.2 kcalmolꢀ1 [MR-MP2/TZV
(2df,2p)] was calculated.
H
ꢀ
However, for the C N breakage of azirine an open-shell
transition state (24.8 kcalmolꢀ1 at UB3LYP, 27.3 kcalmolꢀ1
at MR-MP2) was localized to give diradicaloid vinylnitrene
(19.9 kcalmolꢀ1 at UB3LYP, open-shell singlet with some
mixing in of the triplet state, and 31.4 kcalmolꢀ1 at MR-
MP2, open-shell singlet).
ꢀ
much lower than those of C C bond cleavage (Figure 2).
ꢀ
Heterolytic C C bond breaking of bi-2H-azirin-2-yls 9d,
Upon including the zero-point energy, the energy of the
transition state (TS syn-27d!cis-26d) drops below those of
the intermediate C,N diradicals syn-27d and anti-27d.
Therefore, this reaction pathway is concerted (unsymmetri-
cal) rather than stepwise. On the other hand, there is also a
with calculated activation energies of 40–47 kcalmolꢀ1 (all
UB3LYP calculations give closed-shell species), is compara-
bly unfavorable.[41] Here, in the first step, C C bond fission
ꢀ
of one of the two azirine rings would lead to a nitrile ylide.
This could either attack the second azirine ring at the nitro-
gen atom to produce pyrimidine or, after thermal opening
of the second ring, could give a bis(nitrile ylide), which
would aromatize immediately to pyrazine. We could not
ꢀ
synchronous pathway with symmetric cleavage of both C N
bonds (rac-syn-9d!cis-26d, rac-anti-9d!trans-26d) with
slightly lower activation energy of 17.4 or 18.6 kcalmolꢀ1. At
the more reliable MR-MP2/TZV(2df,2p) level of theory, the
A
7472
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 7467 – 7481

Valence Isomerization of Bi-2H-azirin-2-yls to Diazabenzenes
FULL PAPER
Led by the observation that
in case of other substituted bi-
azirinyl systems thermal frag-
mentation to alkynes and ni-
triles was observed as the domi-
nant process,[42] we also investi-
gated this pathway for the
parent system. There is no such
pathway for Z diradical cis-26d,
but all rotational isomers of E
diradical trans-26d (Figure 3)
can undergo fragmentation.
Ethyne and two molecules of
hydrocyanic acid are either
eliminated simultaneously from
trans-26d or in two stages from
trans-26d’ and trans-26d’’. Both
pathways have relatively high
barriers
(trans-26d!C2H2 +
2HCN 47.1, trans-26d’!s-trans-
29d 35.3, trans-26d’’!s-cis-29d
36.1 kcalmolꢀ1).
Elimination of hydrogen cya-
nide leads to unstable C,N di-
A
ethyne and a second hydrogen
cyanide molecule with activa-
tion energies of about 2 kcal
molꢀ1. Again, comparison of
the energies at the UB3LYP/6-
31G*
and
MR-MP2/TZV-
ꢀ
Figure 2. Reactions of the parent biazirinyls 9d that are triggered by C N bond breaking: Energies at
UB3LYP/6-31+G* (ZPVE corrected) relative to rac-syn-9d are given in kcalmolꢀ1. Relative energies of tran-
sition states are given at the corresponding arrows; relative energies (ZPVE-corrected), calculated at the MR-
AHCTREUNG
that the DFT method is appro-
priate to describe these struc-
MP2/TZV(2df,2p) level, are given in parentheses.
G
tures. Both methods agree that
the activation barriers for the
following barriers for the symmetrical cleavage were ob-
tained: 13.4 and 14.8 kcalmolꢀ1. Interestingly, the diradicals
cis-26d and trans-26d are more stable than the closed-shell
(but highly strained) biazirinyls [ꢀ9.2 and ꢀ11.7 kcalmolꢀ1
at UB3LYP/6-31+G* and ꢀ6.4 kcalmolꢀ1 for both at MR-
fragmentation are too high to be overcome at experimental
ambient conditions in the parent system.
Tetramethylbi-2H-azirin-2-yls 9g: We extended our calcula-
tions further to the tetramethylbiazirinyls 9g, which serve as
models for the experimentally studied alkyl-substituted de-
rivatives. We were especially interested to learn about the
substituent effect of the methyl groups on the course of the
ring-opening process. Due to the presence of the methyl
groups, we were not able to localize all the transition states
which interconnect the various species involved in the rear-
rangement. Tetramethylbiazirinyl 9g exhibits similar config-
urations and conformations to the parent system (Figure 4).
MP2/TZV(2df,2p)]. The stability of the diradicals probably
A
arises from the planar geometry and hence effective conju-
gation of the p system. Summing up, it can be stated, that
there is a low energy pathway, either concerted or even syn-
chronous (operative at room temperature), that converts the
parent biazirinyls rac- and meso-9d to the diradicals cis-26d
and trans-26d, respectively.
To obtain information on the fate of the diradicals in the
experimental setup, we performed additional calculations on
the C4H4N2 energy hypersurface starting from cis-26d and
ꢀ
We found that simultaneous homolytic C N bond breaking
of both azirine rings is preferred in comparison to sequential
homolytic opening of the three-membered rings. Compared
to the parent system, the activation barriers for the simulta-
C
trans-26d. Rotation of one or both C=N groups around the
ꢀ
C C single bonds of cis-N,N’ diradical cis-26d leads to ni-
trile 28d via a 1,5-H shift or to pyridazine 18d, respectively,
ꢀ
neous cleavage of the C N bonds are here somewhat lower
via aromatization. Both reactions have low activation ener-
for the rac isomers but higher for the meso isomers
gies
(cis-26d!28d
4.3 kcalmolꢀ1,
cis-26d!18d
(Table 2).
7.1 kcalmolꢀ1).
Chem. Eur. J. 2006, 12, 7467 – 7481
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7473

K. Banert, S. Grimme, R. Herges, E.-U. Würthwein et al.
suggested.[9] Pyridazines 18 are produced instead. On the
basis of quantum-chemical calculations, their formation is
explained by short-lived but energetically favorable N,N’ di-
radicals 26, which are generated by simultaneous homolytic
ꢀ
cleavage of both C N bonds of the azirine rings. Assuming
these intermediates, the low barrier for the aromatization
process 9!18, which excludes any competitive diaza Cope
rearrangement, can easily be rationalized. The more rapid
reaction of the rac stereoisomers of 9 compared to the cor-
responding meso compounds is also explained by the theo-
retical studies. Furthermore, the missing formation of aro-
matic compounds 18 in the case of heterocycles 9 with a low
number of substituents is interpreted by competitive reac-
tions of the involved intermediates 26.
Presently, we are trying to prepare more stable bi-2H-
azirin-2-yls which will allow experimental assignment of
their stereochemistry and direct correlation of their rates of
valence isomerization.
Experimental Section
Figure 3. Fragmentation pathways for the E diradical trans-26d: Energies
at UB3LYP/6-31+G* level (ZPVE corrected) in kcalmolꢀ1 are relative
to biazirinyl rac-syn-9d. Energies of transition states are given at the cor-
responding reaction arrows; relative energies (ZPVE-corrected), calcu-
Caution: Care must be taken in handling azides, which are explosive. Es-
pecially neat (E,E)-8d can lead to heavy explosions on friction.
lated at the MR-MP2/TZV(2df,2p) level, are given in parentheses.
A
Instrumentation and measurement: Melting points were determined with
a Pentakon Dresden Boetius apparatus and are uncorrected. FTIR spec-
tra were recorded on a BRUKER IFS 28 FTIR spectrophotometer. IR
measurements were made on solutions in KBr cuvettes. UV/Vis spectra
were acquired by using a Carl Zeiss Jena SPECORD UV/Vis spectro-
photometer in cuvettes of 1 cm thickness. 1H and 13C NMR spectra were
recorded on a Varian Gemini 2000 spectrometer operating at 300 and
75 MHz, respectively. NMR signals were referenced to TMS (d=0) or
solvent signals and recalculated relative to TMS. The multiplicities of
13C NMR signals were determined with the aid of gated spectra and/or
DEPT135 experiments. GC-MS spectra were acquired with a Hewlett-
Packard model 5989 quadrupole mass spectrometer. Ionization was per-
formed by EI (70 eV). For the previous separation, a Hewlett-Packard
model 5890 series II gas chromatograph with thermal conductivity detec-
tor was used (column: HP-MS 5 (60 m), carrier gas: helium). HRMS
(ESI) spectra were recorded on an Applied Biosystems Mariner 5229
mass spectrometer. The elemental analyses were performed with a Vario
EL elemental analyzer from Elementar Analysensysteme GmbH Hanau.
Elemental analyses of explosive azides and highly unstable azirines could
not be performed. The gas chromatograms were recorded with a Hew-
lett-Packard model 5890 series II gas chromatograph with flame ioniza-
tion detector (column: HP-MS 5 capillary column (5% phenylmethylsili-
cone gum), carrier gas: nitrogen). Preparative GC separations were car-
ried out with a Shimadzu model GC 8A gas chromatograph. Separation
conditions and columns used are given at the separated compounds. TLC
was performed with Macherey-Nagel POLYGRAM SIL G/UV254 poly-
However, the geometries and energies of the tetramethyl
N,N’ diradicals 26g formed by the initial bond-breaking
process are markedly different. They are nonplanar because
of steric hindrance of the methyl groups. In comparison to
the parent system, where diradicals 26d are more stable
than bisazirinyls 9d, here tetramethyl N,N’ diradicals 26g
are 7–10 kcalmolꢀ1 less stable than tetramethylbiazirinyls
ꢀ
9g. The C C single bonds of the tetramethyl diradicals are
longer (1.51 ꢂ) than in the parent system (1.47 ꢂ), and this
indicates reduced conjugation and thus destabilization,
while the central C=C bond is similar in length in both sys-
tems. Being nonplanar, the tetramethyl N,N’ diradicals are
more easily converted to tetramethylpyridazine 18g, which
in turn also suffers from steric interaction of the methyl
groups. Thus, we explain the different chemical behavior of
the methyl-substituted biazirinyls 9g mainly as a result of
steric repulsion of the methyl groups in diradicals 26g and
cyclic compound 9g in comparison to the parent biazirinyls
9d.
ester sheets. Flash column chromatography was performed with 32–
63 mm silica gel. HPLC separations were carried out under isocratic flow
with distilled and degassed solvents. A KNAUER model 64 HPLC pump
equipped with MERCK LiChrospher Si 60, 5 mm (1 20 mm, l=20 cm)
column and UV detector (l=254 nm) was used.
Conclusion
Electrocyclic ring opening of 3,4-diazidocyclobutenes 16 is a
successful strategy to prepare 1,4-diazidobuta-1,3-dienes 8,
which were unknown before. By irradiation of the latter,
long-sought[8,9] bi-2H-azirin-2-yls 9 are accessible for the
first time. The postulated[8] valence isomerization of these
heterocycles can indeed be realized, but the thermal trans-
formation does not lead to pyrimidines and pyrazines as was
General instructions for photolysis: Irradiation was conducted at ꢀ40 to
ꢀ608C by using a high-pressure mercury lamp (TQ150, Heraeus GmbH)
supplied with glass or quartz equipment and an ethanol cryostat. Most of
these photolyses were monitored by NMR spectroscopy. Thus, the solu-
tion of the appropriate starting material was irradiated in an NMR tube.
To exclude oxygen, this solution was flushed with argon in an ultrasonic
bath prior to irradiation. Dioxane was used as standard to determine
1
yields based on H NMR data. When a photosensitizer was employed, sa-
7474
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 7467 – 7481

Valence Isomerization of Bi-2H-azirin-2-yls to Diazabenzenes
FULL PAPER
(2.76 g, 20.0 mmol) in anhydrous Et2O
was added dropwise at a temperature
not above ꢀ558C. The reaction mix-
ture was stirred at ꢀ608C for 1 h and
then added slowly at a temperature
not above ꢀ558C (caution: strongly
exothermic reaction) to a previously
prepared mixture of glacial acetic acid
(6.00 g, 100 mmol) and anhydrous
Et2O (200 mL), which was cooled to
ꢀ608C. After the addition was finish-
ed, stirring was continued for 30 min.
Then the reaction mixture was washed
(1ꢁH2O, 2ꢁaq NaHCO3, 2ꢁH2O) and
dried (MgSO4). After removing the
solvent under reduced pressure, 14h
(3.64 g, 18.4 mmol, 92%) was obtained
as a yellow oil which was recondensed
in vacuum and crystallized (hexane,
ꢀ258C) to give
a white solid, m.p.
698C (hexane). IR (CDCl3): n˜ =
3479 cmꢀ1 (OH); 1H NMR (CDCl3):
3
d=1.03 (t, J=7.4 Hz, 6H; CH3CH2 at
C-3/C-4), 1.06 (t, 3J=7.8 Hz, 6H;
CH3CH2 at C-1/C-2), 1.71 (m, 4H;
3
CH2 at C-3/C-4), 2.08 (brq, J=7.8 Hz,
4H; CH2 at C-1/C-2), 4.40 ppm (brs,
2H; OH); 13C NMR (CDCl3): d=8.53
(q), 12.99 (q), 18.00 (t, CH2 at C-1/C-
2), 27.74 (t, CH2 at C-3/C-4), 84.79 (s,
C-1/C-2), 147.51 ppm (s, C-3/C-4); MS
(ESI): m/z (%): 181.1601 (100)
[MꢀOHꢀ; calcd 181.1587]; elemental
analysis calcd (%) for C12H22O2:
72.68, 11.18; found: 72.40,
C
H
H
C
10.85. If hydrolysis of the reaction
mixture obtained after treatment of
the dione with ethyllithium was per-
formed at higher temperature (08C)
with ice/H2O, open-chain diketones
were found as main products (see Sup-
porting Information for details).
Figure 4. Reactions of tetramethyl-bi-2H-azirin-2-yls 9g: Energies at UB3LYP/6-31+G* level (ZPVE-correct-
ed) relative to tetramethylbiazirinyl rac-syn-9g are given in kcalmolꢀ1. Energies of transition states are given
above the corresponding arrows.
3-Ethyl-1,2,4-trimethylcyclobut-3-ene-
1,2-diol (14j) and 3,4-diethyl-1,2-dime-
thylcyclobut-3-ene-1,2-diol
(14k):
Treatment of 3-ethyl-4-methylcyclo-
turated solutions of the sensitizer were used as solvent. Exact conditions
are given at the compounds irradiated.
but-3-ene-1,2-dione (13j=13k’) or 3,4-diethylcyclobut-3-ene-1,2-dione
(13h=13k) with methylmagnesium bromide was carried out analogously
to a known procedure[21b] to get 14j and 14k, respectively.
3-Ethyl-4-methylcyclobut-3-ene-1,2-dione (13j): Using a known proce-
dure,[20] pent-2-yne (12j) was treated with activated zinc powder and tri-
chloroacetyl chloride to give regioisomeric 4,4-dichloroethylmethylcyclo-
but-2-enones, which were hydrolyzed in the presence of sulfuric acid to
yield 13j as a pale yellow liquid (see Supporting Information for details).
IR (CDCl3): n˜ =1771 cmꢀ1 (C=O); 1H NMR (CDCl3): d=1.29 (t, 3J=
14j: Yield 86%, white solid, m.p. 408C (hexane); IR (CDCl3): n˜ =
3396 cmꢀ1 (OH); 1H NMR (CDCl3): d=1.02 (t, 3J=7.7 Hz, 3H; 2’-H),
1.20 (s, 3H), 1.23 (s, 3H), 1.54 (t, 5J=1.1 Hz, 3H; Me-4), 2.00 (qq, 3J=
7.7 Hz, 5J=1.1 Hz, 2H; 1’-H), 3.10 ppm (brs, 2H; OH); 13C NMR
(CDCl3): d=7.93 (q), 12.17 (q), 17.58 (t, C-1’), 18.35 (q), 19.11 (q), 79.71
(s), 80.04 (s), 143.18 (s), 149.04 ppm (s); MS (ESI): m/z (%): 139.1144
(18) [MꢀOHꢀ; calcd 139.1117], 179.1051 (100) [M+Na+; calcd
179.1043], 335.2185 (44) [2M+Na+; calcd 335.2193]; elemental analysis
calcd (%) for C9H16O2: C 69.19, H 10.32; found: C 68.96, H 10.24.
5
3
5
7.7 Hz, 3H; 2’-H), 2.34 (t, J=1.0 Hz, 3H; Me), 2.75 (qq, J=7.7 Hz, J=
1.0 Hz, 2H; 1’-H); 13C NMR (CDCl3): d=10.13 (q), 11.14 (q), 19.88 (t, C-
1’), 198.58 (s), 199.12 (s), 199.63 (s), 204.10 (s); MS (ESI): m/z (%):
125.0535 (0.3) [M+H+; calcd 125.0597], 249.1092 (100) [2M+H+; calcd
249.1121]; elemental analysis calcd (%) for C7H8O2: C 67.73, H 6.50;
found: C 67.65, H 6.57.
14k: Yield 82%, white solid, m.p. 85–868C (hexane); IR (CCl4): n˜ =
3386 cmꢀ1 (OH); 1H NMR (CDCl3): d=1.08 (t, 3J=7.7 Hz, 6H; 2’-H),
1.30 (s, 6H; 1-Me/2-Me), 2.07 (q, 3J=7.6 Hz, 2H; 1’-H), 2.08 (q, 3J=
7.7 Hz, 2H; 1’-H), 2.45 ppm (brs, 2H; OH); 13C NMR (CDCl3): d=12.60
(q, C-2’), 17.66 (t, C-1’), 19.19 (q, Me-1/Me-2), 79.85 (s, C-1/C-2),
148.37 ppm (s, C-3/C-4); MS (ESI): m/z (%): 153.1267 (80) [MꢀOHꢀ;
calcd 153.1274], 193.1213 (100) [M+Na+; calcd 193.1199], 363.2471 (57)
[2M+Na+; calcd 363.2506]; elemental analysis calcd (%) for C10H18O2: C
70.55, H 10.66; found: C 70.72, H 10.90.
cis-3,4-Diphenylcyclobut-3-ene-1,2-diol (cis-14 f): Cyclobutenedione 13 f
was treated with LiAlH4 to prepare cis-14 f in 23% yield.[21a] We found
that the yield could be increased to 71% by using NaBH4/CeCl3 instead
of LiAlH4, analogously to a known procedure.[43]
1,2,3,4-Tetraethylcyclobut-3-ene-1,2-diol (14h): A solution of ethyllithi-
um[44] (0.75m, 100 mL, 75 mmol) in Et2O was cooled to ꢀ608C. There-
after
a
solution of 3,4-diethylcyclobut-3-ene-1,2-dione (13h=13k)[20]
Chem. Eur. J. 2006, 12, 7467 – 7481
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7475

K. Banert, S. Grimme, R. Herges, E.-U. Würthwein et al.
Table 2. Relative energies Erel(0 K) [kcalmolꢀ1] from UB3LYP/6-31+G*
calculations for the species 9g, 18g, and 26d and the related transition
states.
cis-3-Azido-4-chlorocyclobutene (cis-19d): Colorless liquid; IR (CDCl3):
1
n˜ =2108 cmꢀ1 (N3); H NMR (CDCl3): d=4.44 (brd, J=2.8 Hz, 1H), 5.15
(m, 1H), 6.33 (m, 1H), 6.42 ppm (m, 1H); 13C NMR (CDCl3): d=62.57
(d), 64.25 (d), 138.16 (d), 142.83 ppm (d).
Erel(0 K)
cis-3,4-Diazidocyclobutene (cis-16d): Colorless liquid; IR (CDCl3): n˜ =
2109 cmꢀ1 (N3), 1261 (N3); 1H NMR (CDCl3): d=4.53 (brs, 2H; 3-H/4-
H), 6.40 ppm (t, J=1.1 Hz, 2H; 1-H/2-H); 13C NMR (CDCl3): d=65.26
(d, C-3/C-4), 140.24 ppm (d, C-1/C-2).
rac-syn-9g
rac-anti-9g
meso-gauche-9g
meso-anti-9g
TS rac-syn-9g!(Z)-rac-26g
TS rac-anti-9g!(E)-rac-26g
TS meso-anti-9g!(E)-meso-26g
TS meso-gauche-9g!(Z)-meso-26g
(Z)-rac-26g
0.0
1.5
0.5
0.7
A
19.1
21.3
22.2
27.4
6.9
(CHCl3): n˜ =2111 cmꢀ1 (N3), 1261 (N3); 1H NMR (CDCl3): d=5.90 (ddd,
3Jtrans =13.3 Hz, 3J=11.2 Hz, 4J=0.6 Hz, 1H; 2-H), 6.14 (ddd, Jtrans
13.1 Hz, J=0.8 Hz, J=0.6 Hz, 1H; 4-H), 6.20 (brd, 3Jtrans =13.3 Hz, 1H;
1-H), 6.39 ppm (ddd, 3Jtrans =13.1 Hz, 3J=11.2 Hz, 4J=0.6 Hz, 1H; 3-H);
13C NMR (CDCl3): d=116.1 (d), 119.6 (d), 129.6 (d), 129.8 ppm (d). The
(E)-rac-26g
(E)-meso-26g
7.7
7.4
assignment of H NMR signals was possible by means of the greater peak
width at half-height of the signal of 1-H.
(Z)-meso-26g
18g
10.1
ꢀ60.0
(E,E)-1,4-Diazidobuta-1,3-diene ((E,E)-8d): QN3 (2.0 g, 4.3 mmol) was
melted with CHCl3 (1 mL) to a homogenous solution. After cooling to
room temperature, trans-15d (X=Br)[13] or trans-15d (X=I)[13]
(1.5 mmol) was added. After 20 h [trans-15d (X=Br)] or 80 min [trans-
15d (X=I)] at room temperature, the reaction mixture was separated by
chromatography (Et2O/hexane, silica gel). The first fraction contained
butadiene (E,E)-8d [89% from trans-15d (X=Br) and 62% from trans-
15d (X=I)]. As the second fraction cis-16d [7% from trans-15d (X=
Br) and 6% from trans-15d (X=I)] was eluted. (E,E)-8d: Explosive
yellow needles (hexane), m.p. (decomp). IR (CDCl3): n˜ =2099 cmꢀ1 (N3),
1257 (N3); UV/Vis (CH3CN): lmax (lge)=296 nm (4.52); 1H NMR
Table 3. Product distribution after treatment of cis-15d (X=Cl) with
azide-transfer reagents.
Reagent Solvent
t/T
[h/8C] (X=Cl) 19d
[%] [%] [%]
cis-15d cis- cis-16d
C
NaN3
LiN3
QN3
QN3
[D6]DMSO 72/50 15
17
6
31
12
20
–-
13
18
12
–-
5
34
traces
12
(CDCl3): AA’XX’ system, dA =5.87 (m, 2H; 2-H/3-H), dX =6.12 ppm (m,
CD3OD
C6D6
72/70 27
75/50 33
23/50 21
4
2H; 1-H/4-H); 3JAX =3JA’X’ =13.3 Hz, 3JAA’ =11.5 Hz, 5JXX’ =1.0 Hz, JAX’
=
4JA’X =ꢀ0.5 Hz (coupling constants obtained by spectrum simulation);
molten
12
4
13C NMR (CDCl3): d=116.39 (d), 127.97 ppm (d).
Reaction of trans-3,4-dichlorocyclobutene [trans-15d (X=Cl)] with
QN3: QN3 (1.0 g, 2.13 mmol) was melted with CHCl3 (1 mL) to a homo-
genous solution. After cooling to room temperature, trans-15d (X=
Cl)[13] (88.0 mg, 0.72 mmol) was added. After 24 h at room temperature,
the reaction mixture was recondensed in vacuum and the solvent re-
moved from the condensate under reduced pressure to yield a mixture of
(E,E)-20d (62%) and 18d (10%) as a light yellow oil.
3,4-Diiodo-1,2-diphenylcyclobutene [15 f (X=I)]: A solution of diol cis-
14 f (2.14 g, 8.99 mmol) in dry CHCl3 (100 mL) was added to a solution
of PI3 (3.80 g, 9.22 mmol) in dry CHCl3 (200 mL) at ꢀ508C. The reaction
mixture was kept for 4 days at this temperature, then warmed to room
temperature slowly, and stirred for 2 days. After washing (H2O, aq
NaHCO3, aq Na2S2O3, H2O), the organic layer was separated and dried
(MgSO4). Evaporating the solvent under reduced pressure led to a resi-
due which was recrystallized (Et2O/hexane) to yield 15 f (X=I) (3.18 g,
6.94 mmol, 77%) as a pale beige crystalline solid, m.p. 1278C (Et2O/
Reaction of dibromides 15e, 15e’, and 15e’’ with azide-transfer reagents:
a) A mixture of isomeric dibromides 15e, 15e’, and 15e’’[14] (1.00 g,
4.17 mmol) in DMSO (5 mL) was added dropwise to a solution of NaN3
(1.08 g, 16.6 mmol) in DMSO (40 mL) at room temperature. The reaction
mixture was stirred for 45 min at 508C. After cooling, it was poured onto
ice/H2O and extracted with Et2O. The organic layer was washed (H2O),
dried (MgSO4), and the solvent was removed in vacuum. A mixture of
(E,E)-8e and cis-16e (650 mg, ratio (E,E)-8e:cis-16e=1.8:1) and a small
amount of unconverted bromides was obtained. Crystallization (CH2Cl2
at ꢀ258C) yielded (E,E)-8e (230 mg, 34%) as orange-yellow crystals.
From the remaining solution, some more (E,E)-8e and cis-16e (108 mg,
16%, yellow oil) could be obtained by HPLC (Et2O/hexane 1/40). b) A
mixture of isomeric dibromides 15e, 15e’, and 15e’’ (1.00 g, 4.17 mmol)
in DMSO (5 mL) was added dropwise to a solution of LiN3 (820 mg,
16.7 mmol) in DMSO (40 mL) at room temperature. The reaction mix-
ture was stirred for 2 h at 408C. Workup followed the procedure descri-
bed under a) to yield (E,E)-8e (250 mg, 37%) as orange-yellow crystals.
c) A mixture of isomeric dibromides 15e, 15e’, and 15e’’ (650 mg,
2.71 mmol) in CHCl3 (2 mL) was added dropwise to a solution of QN3
(3.80 g, 8.09 mmol) in CHCl3 (3 mL) at room temperature. The reaction
mixture was stirred for 16 h at room temperature. The solvent was re-
moved under reduced pressure, and the residue was separated by chro-
matography (silica gel, Et2O). Crystallization (hexane) yielded (E,E)-8e
(152 mg, 34%) as orange-yellow crystals.
1
hexane). H NMR (CDCl3): d=5.59 (s, 2H; 3-H/4-H), 7.41–7.46 (m, 6H),
7.53–7.58 ppm (m, 4H); 13C NMR (CDCl3): d=27.51 (d, C-3/C-4), 127.53
(d), 128.65 (d), 129.66 (d, p-Ph), 131.31 (s, i-Ph), 139.19 ppm (s, C-1/C-2);
elemental analysis calcd (%) for C16H12I2: C 41.95, H 2.64; found: C
41.93, H 2.81.
Reaction of cis-3,4-dichlorocyclobutene [cis-15d (X=Cl)] with azide-
transfer reagents: Reaction with NaN3: Finely powdered NaN3 (101.3 mg,
1.56 mmol) was dissolved in [D6]DMSO in an NMR tube with warming.
Then the solution was allowed to cool to room temperature, and cis-15d
(X=Cl) (52.2 mg, 0.42 mmol) was added. After 72 h at 508C the reaction
mixture showed the composition given in Table 3. Separation was carried
out by preparative GC on a nonpolar column (1 m, OV-101) at a column
temperature of 808C. Reaction with LiN3: Finely powdered LiN3
(101.3 mg, 1.56 mmol) was dissolved in CD3OD in an NMR tube. Then
cis-15d (X=Cl) (52.2 mg, 0.42 mmol) was added. The progress of the re-
action was monitored by 1H NMR spectroscopy. After 70 h at 708C the
reaction was stopped. The composition of the mixture is given in Table 3.
Reaction with QN3: a) Hexadecyltributylphosphonium azide (QN3,
165.3 mg, 0.352 mmol) was solved in C6D6 in an NMR tube and cis-15d
(X=Cl) (14.0 mg, 0.114 mmol) was added. The progress of the reaction
1
was monitored by H NMR spectroscopy. After 75 h at 508C the reaction
was stopped. The composition of the mixture is given in Table 3. b) QN3
(580 mg, 1.23 mmol) was melted and cis-15d (X=Cl) (50.0 mg,
0.41 mmol) was added at 508C. After 23 h at 508C the reaction mixture
was recondensed in oil-pump vacuum to yield a colorless oil (34.6 mg)
composed as given in Table 3.
A
(hexane, decomp); IR (CDCl3): n˜ =2100 cmꢀ1 (N3), 1258 (N3); UV/Vis
(cyclohexane): lmax (lge)=297 nm (4.46); 1H NMR (CDCl3): d=1.75 (d,
4J=1.1 Hz, 6H; Me), 6.26 ppm (q, 4J=1.1 Hz, 2H; 1-H/4-H); 13C NMR
(CDCl3): d=11.97 (q, Me), 123.14 (d, C-1/C-4), 125.13 ppm (s, C-2/C-3).
7476
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 7467 – 7481

Valence Isomerization of Bi-2H-azirin-2-yls to Diazabenzenes
cis-3,4-Diazido-1,2-dimethylcyclobutene (cis-16e): IR (CDCl3): n˜ =
FULL PAPER
(s, 6H; 1-Me/2-Me); 13C NMR (CDCl3): d=8.80 (q, Me-1/Me-2), 17.55
4
2101 cmꢀ1 (N3), 1241 (N3); 1H NMR (CDCl3): d=1.74 (d, J=0.7 Hz, 6H;
(q, Me-3/Me-4), 73.41 (s, C-3/C-4), 141.94 ppm (s, C-1/C-2).
Me), 4.24 ppm (brs, 2H; 3-H/4-H); 13C NMR (CDCl3): d=11.48 (q, Me),
64.75 (d, C-3/C-4), 141.49 ppm (s, C-1/C-2).
3-Azido-1,2,3-trimethyl-4-methylencyclobutene (24g): Yellow oil; IR
(CDCl3): n˜ =2096 cmꢀ1 (N3), 1247 (N3); 1H NMR (CDCl3): d=1.36 (s,
3H; 3-Me), 1.70 (q, 5J=1.2 Hz, 3H; 2-Me), 1.77 (m, 3H; 1-Me), 4.50
(brs, 1H), 4.54 ppm (brs, 1H); 13C NMR (CDCl3): d=8.95 (q), 9.10 (q),
19.35 (q, 3-Me), 71.30 (s, C-3), 90.85 (t, =CH2), 142.76 (s), 150.04 (s),
153.09 ppm (s). The assignment of the 1H NMR signals of the methyl
groups was performed by double resonance experiments.
Reaction of 3,4-diiodo-1,2-diphenylcyclobutene [15 f (X=I)] with QN3:
3,4-Diiodo-1,2-diphenylcyclobutene 15 f (X=I) (1.15 g, 2.51 mmol) was
dissolved in a solution of QN3 (4.20 g, 8.94 mmol) in CHCl3 (5 mL). The
reaction mixture was stirred at room temperature for 12 h and the color
changed to bright yellow. The phosphonium salts were separated by chro-
matography (Et2O/hexane), and the organic solvents were removed from
the eluate in vacuum. Recrystallizing (Et2O/hexane) the residue yielded
1,4-diazido-2,3-diphenylbuta-1,3-diene [(E,E)-8 f] (508 mg, 1.76 mmol,
70%), yellow crystalline solid, m.p. 808C (Et2O/hexane, decomp). IR
(CDCl3): n˜ =2096 cmꢀ1 (N3), 1269 (N3); 1H NMR (CDCl3): d=6.01 (s,
2H), 7.25–7.48 ppm (m, 10H); 13C NMR (CDCl3): d=126.61 (d), 127.91
(d), 128.41 (d, 2C), 129.74 (d, 2C), 131.55 (s), 135.24 ppm (s).
3-Azido-4-chloro-1,2,3,4-tetramethylcyclobutene [19g (major isomer)]:
Colorless oil; IR (CDCl3): n˜ =2099 cmꢀ1 (N3), 1255 (N3); 1H NMR
5
(CDCl3): d=1.26 (s, 3H), 1.36 (s, 3H), 1.58 (q, J=1.2 Hz, 3H), 1.61 ppm
(q, 5J=1.2 Hz, 3H); 13C NMR (CDCl3): d=7.59 (q), 8.57 (q), 16.10 (q),
18.94 (q), 73.78 (s, C-3), 80.48 (s, C-4), 140.36 (s), 145.72 ppm (s).
3-Azido-4-chloro-1,2,3,4-tetramethylcyclobutene [19g (minor isomer)]:
Colorless oil; IR (CDCl3): n˜ =2089 cmꢀ1 (N3), 1252 (N3); 1H NMR
5
(CDCl3): d=1.34 (s, 3H), 1.37 (s, 3H), 1.58 (q, J=1.2 Hz, 3H), 1.61 ppm
Reaction of 3,4-dibromo-3,4-dimethyl-1,2-diphenylcyclobutene [15i (X=
Br)] with QN3: Treatment of dibromide 15i (X=Br) with QN3 was car-
ried out as described for the analogous transformation of 15j (X=Br);
see Table 4. However, the reaction did not yield the wanted diazidocyclo-
butene but led to 24i (18%) and the known[25] bis(methylene)cyclobu-
tene 25i (48%).
(q, 5J=1.2 Hz, 3H); 13C NMR (CDCl3): d=7.52 (q), 8.66 (q), 17.63 (q),
20.28 (q), 73.03 (s, C-3), 81.26 (s, C-4), 140.19 (s), 143.55 ppm (s).
The assignment of the signals of C-3 and C-4 of both isomers of 19g was
possible by comparison with NMR data of 15g (X=Cl), trans-16g, and
cis-16g.
Reactions of tetraalkyl-3,4-dichlorocy-
clobutenes and tetraalkyl-3,4-dibromo-
cyclobutenes 15j and 15j’ as well as
15k, 15k’, and 15k’’ with QN3: The ap-
Table 4. Synthesis of trans-3,4-diazidocyclobutenes from dihalides.
Dihalides
QN3
trans-3,4-Diazidocyclobutenes
propriate dihalides were dissolved
(CHCl3) and added with stirring to
molten QN3 at room temperature. Stir-
ring was continued for 12 h at room
temperature. After adding Et2O
(100 mL), the reaction mixture was fil-
tered over silica gel by eluting with
Et2O. The solvent was evaporated
under reduced pressure, and the crude
product was purified by chromatogra-
phy (hexane/Et2O 2/1, silica gel) to
yield mixtures of the isomeric trans-
15j (X=Cl) and 15j’ (X=Cl) (1.13 g, 5.85 mmol) in 11.0 g
trans-16j:trans-16j’ꢁ1:1 (0.90 g,
trans-16j:trans-16j’ꢁ1:1 (0.78 g,
trans-16k:trans-16k’:trans-16k’’ꢁ1:2:1
CHCl3 (5 mL)
(23.4 mmol) 4.39 mmol, 75%)
15j (X=Br) and 15j’ (X=Br) (1.46 g, 5.18 mmol) in 10.8 g
CHCl3 (5 mL)
(23.0 mmol) 3.78 mmol, 73%)
15k (X=Cl), 15k’ (X=Cl), and 15k’’ (X=Cl)
(1.48 g, 7.14 mmol) in CHCl3 (2 mL)
15k (X=Br), 15k’ (X=Br), and 15k’’ (X=Br)
(1.84 g, 6.22 mmol) in CHCl3 (5 mL)
10.0 g
(21.3 mmol) (914 mg, 4.43 mmol, 62%)
10.2 g
(21.7 mmol) (696 mg, 3.16 mmol, 51%)
trans-16k:trans-16k’:trans-16k’’ꢁ1:2:1
3-Azido-3-methyl-4-methylene-1,2-diphenylcyclobutene (24i): Colorless
oil; IR (CDCl3): n˜ =2101 cmꢀ1 (N3), 1246 (N3); 1H NMR (CDCl3): d=
3,4-diazidocyclobutenes as yellow liquids (Table 4).
trans-3,4-Diazido-1-ethyl-2,3,4-trimethylcyclobutene (trans-16j): Color-
2
2
less liquid; IR (CDCl3): n˜ =2094 cmꢀ1 (N3), 1247 (N3); H NMR (CDCl3):
1
1.62 (s, 3H; Me), 4.88 (d, J=1.4 Hz, 1H), 4.98 (d, J=1.4 Hz, 1H), 7.30–
7.64 ppm (m, 10H); 13C NMR (CDCl3): d=20.73 (q, Me), 70.89 (s, C-3),
95.53 (t, =CH2), 126.93 (d), 127.66 (d), 128.59 (d), 128.63 (d), 128.75 (d,
p-Ph), 129.04 (d, p-Ph), 131.48 (s), 131.92 (s), 149.31 (s), 150.86 ppm (s),
one signal (s) was hidden by another.
d=1.10 (t, 3J=7.6 Hz, 3H; 2’-H), 1.41 (s, 3H), 1.45 (s, 3H), 1.67 (t, 5J=
1.2 Hz, 3H; 2-Me), 2.09 ppm (qq, 3J=7.6 Hz, 5J=1.2 Hz, 2H; 1’-H);
13C NMR (CDCl3): d=9.15 (q), 11.71 (q), 17.08 (q), 18.39 (t, C-1’), 18.69
(q), 72.52 (s), 72.67 (s), 140.51 (s, C-2), 146.28 ppm (s, C-1).
trans-3,4-Diazido-3-ethyl-1,2,4-trimethylcyclobutene (trans-16j’): Color-
Reaction of 3,4-dichloro-1,2,3,4-tetramethylcyclobutene [15g (X=Cl)]
with QN3: 15g (X=Cl)[16] (1.79 g, 10 mmol) was added to a solution of
QN3 (11.0 g, 23.4 mmol) in CHCl3 (50 mL) under cooling with water. The
reaction mixture was evaporated in vacuum after 20 h of stirring. The res-
idue was separated by chromatography (Et2O, silica gel) to remove the
phosphonium salts. After removal of the solvent under reduced pressure,
a pale yellow oil (2.04 g) was obtained. NMR analyses indicated that
trans-16g (55%, 5.5 mmol), cis-16g (9%, 0.9 mmol), 24g (17%,
1.7 mmol), 19g (major isomer, 3%, 0.3 mmol), and 19g (minor isomer,
2%, 0.2 mmol) were formed in an overall yield of 86%. The products
could be separated by flash chromatography (hexane/Et2O, silica gel)
with the following order of elution: trans-16g, 24g, cis-16g, 19g (minor
isomer), 19g (major isomer).
1
less liquid; IR (CDCl3): n˜ =2096 cmꢀ1 (N3), 1254 (N3); H NMR (CDCl3):
3
5
d=1.04 (t, J=7.4 Hz, 3H; 2’-H), 1.42 (s, 3H; 4-Me), 1.67 (q, J=1.2 Hz,
3H), 1.70 (q, 5J=1.2 Hz, 3H), 1.65–1.85 ppm (m, 2H; 1’-H); 13C NMR
(CDCl3): d=8.75 (q), 8.98 (q), 10.23 (q), 18.28 (q, Me-4), 26.52 (t, C-1’),
72.69 (s, C-4), 76.61 (s, C-3), 141.16 (s), 142.43 ppm (s).
Mixture of trans-16k, trans-16k’, and trans-16k’’ (1:2:1): IR (CDCl3): n˜ =
2096 (N3), 1252 cmꢀ1 (N3).
trans-3,4-Diazido-1,2-diethyl-3,4-dimethylcyclobutene (trans-16k): Color-
less liquid; 1H NMR (CDCl3): d=1.10 (t, 3J=7.5 Hz, 6H; 2’-H), 1.46 (s,
3
6H; 3-Me/4-Me), 2.12 ppm (q, J=7.5 Hz, 4H; 1’-H); 13C NMR (CDCl3):
d=12.18 (q, C-2’), 18.48 (t, C-1’), 18.75 (q, Me-3/Me-4), 72.66 (s, C-3/C-
4), 145.87 ppm (s, C-1/C-2).
trans-3,4-Diazido-1,4-diethyl-2,3-dimethylcyclobutene (trans-16k’): Col-
orless liquid; 1H NMR (CDCl3): d=1.03 (t, 3J=7.4 Hz, 3H; CH3CH2 at
C-4), 1.13 (t, 3J=7.7 Hz, 3H; CH3CH2 at C-1), 1.42 (s, 3H; Me-3), 1.71
trans-3,4-Diazido-1,2,3,4-tetramethylcyclobutene (trans-16g): Colorless
oil; IR (CDCl3): n˜ =2093 cmꢀ1 (N3), 1251 (N3); UV/Vis (CH3CN): lmax
(lge)=208 nm (3.74); 1H NMR (CDCl3): d=1.42 (s, 6H; 3-Me/4-Me),
1.64 ppm (s, 6H; 1-Me/2-Me); 13C NMR (CDCl3): d=8.75 (q, Me-1/Me-
2), 17.91 (q, Me-3/Me-4), 72.67 (s, C-3/C-4), 141.24 ppm (s, C-1/C-2).
5
(t, J=1.4 Hz, 3H; Me-2), 1.65–1.85 (m, 2H; CH2 at C-4), 2.12 ppm (brq,
3J=7.7 Hz, 2H; CH2 at C-1); 13C NMR (CDCl3): d=8.85 (q), 9.25 (q),
11.78 (q), 18.17 (q, Me-3), 19.24 (t, CH2 at C-1), 26.79 (t, CH2 at C-4),
72.49 (s, C-3), 76.76 (s, C-4), 141.56 (s, C-2), 146.29 ppm (s, C-1).
cis-3,4-Diazido-1,2,3,4-tetramethylcyclobutene (cis-16g): Colorless oil;
IR (CDCl3): n˜ =2102 cmꢀ1 (N3), 1248 (N3); UV/Vis (Ch3CN): lmax (lge)=
205 nm (3.75); 1H NMR (CDCl3): d=1.29 (s, 6H; 3-Me/4-Me), 1.63 ppm
trans-3,4-Diazido-3,4-diethyl-1,2-dimethylcyclobutene (trans-16k’’): Col-
orless liquid; 1H NMR (CDCl3): d=1.04 (t, 3J=7.5 Hz, 6H; 2’-H), 1.60–
Chem. Eur. J. 2006, 12, 7467 – 7481
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7477

K. Banert, S. Grimme, R. Herges, E.-U. Würthwein et al.
1.90 (m, 4H; 1’-H) 1.73 ppm (s, 6H; 1-Me/2-Me); 13C NMR (CDCl3): d=
8.89 (q), 9.98 (q), 26.38 (t, C-1’), 76.64 (s, C-3/C-4), 142.14 ppm (s, C-1/C-
2).
((E,E)-8k’)+2-H/7-H ((E,E)-8k’’)]; 13C NMR (CDCl3, (E,E)-8k): d=
12.38 (q), 14.76 (q), 22.72 (t), 126.78 (s), 127.52 ppm (s); 13C NMR
(CDCl3, (E,E)-8k’): d=12.48 (q), 12.74 (q), 15.05 (q), 17.18 (q), 22.62 (t),
23.52 (t), 121.55 (s), 127.10 (s), 128.80 (s), 131.87 ppm (s); 13C NMR
(CDCl3, (E,E)-8k’’): d=12.86 (q), 16.86 (q), 22.71 (t), 122.76 (s),
132.30 ppm (s).
Thermal ring opening of trans-3,4-diazidotetraalkylcyclobutenes: Ther-
molyses of the 3,4-diazidotetraalkylcyclobutenes trans-16g or the mixture
of trans-16j and trans-16j’ as well as mixtures of trans-16k, trans-16k’,
and trans-16k’’ were performed in benzene at 70–808C and monitored by
NMR spectroscopy. The reactions were stopped when the amount of
(E,E)-8g, (E,E)-8j, and (E,E)-8j’, as well as (E,E)-8k, (E,E)-8k’, and
(E,E)-8k’’ reached a maximum. The resulting mixtures were separated
by flash chromatography (Et2O/hexane, silica gel) which allowed also the
isolation of unconverted starting materials. In a typical experiment,
(E,E)-8g was obtained in 6% yield based on applied starting material
trans-16g, which correspond to 14% yield related to the amount of con-
verted trans-16g. Monitoring the thermolysis of (E,E)-8g by NMR spec-
troscopy showed a maximum amount of (E)-17g (42%) after 50 min at
808C in CDCl3. In the case of the azides trans-16j and trans-16j’ as well
as trans-16k, trans-16k’, and trans-16k’’, separation of isomeric 1,4-diazi-
dobuta-1,3-dienes was not possible (10% yield of (E,E)-8j and (E,E)-8j’
based on applied azides trans-16j and trans-16j’ as well as 10% yield of
(E,E)-8k, (E,E)-8k’, and (E,E)-8k’’ based on applied azides trans-16k,
trans-16k’, and trans-16k’’). However, changing the ratio of the mixture
of the 3,4-diazidocyclobutenes trans-16k, trans-16k’, and trans-16k’’ led
to product mixtures with different distribution of (E,E)-8k, (E,E)-8k’,
and (E,E)-8k’’, although electrocyclic ring opening competed with scram-
bling of the allylic isomers of the starting material (see Supporting Infor-
mation). These different product mixtures were used for assignment of
NMR signals and also for irradiation to give different distributions of 9k,
9k’, and 9k’’.
Photolyses of (E,E)-1,4-diazidobuta-1,3-diene ((E,E)-8d): (E,E)-8d
(18.6 mg, 0.023 mmol) and dioxane (3 mL, as standard) were dissolved in
a saturated and degassed solution of 9,10-dicyanoanthracene (DCA) in
CDCl3. The NMR tube was irradiated at ꢀ508C, and the progress of the
reaction was monitored by NMR spectroscopy at ꢀ508C. The maximum
yield of 9d (8%, determined by NMR spectroscopy) was reached after
95 min of irradiation. Formation of (E)-17d (up to 24%) as an intermedi-
ate could also be observed. Without DCA, irradiation of (E,E)-8d led to
(E)-17d in up to 60% yield.
2-[(E)-2-Azidoethenyl]-2H-azirine ((E)-17d): Unstable, not isolable;
3
3
1H NMR (CDCl3, ꢀ508C): d=2.40 (dd, J=8.5 Hz, J=2.1 Hz, 1H; 2-H),
4.89 (ddd, 3Jtrans =13.7 Hz, 3J=8.5 Hz, 4J=0.6 Hz, 1H; 1’-H), 6.25 (d,
3Jtrans =13.7 Hz, 1H; 2’-H), 10.07 ppm (dd, 3J=2.1 Hz, 4J=0.6 Hz, 1H; 3-
H); 13C NMR (CDCl3, ꢀ508C): d=26.70 (d, C-2), 119.81 (d), 128.65 (d),
164.95 ppm (d, C-3).
meso-/rac-Bi-2H-azirin-2-yl (meso-9d/rac-9d): Unstable, not isolable;
1H NMR (CDCl3, ꢀ508C): d=1.68 (s, 2H; 2-H, major isomer), 1.98 (s,
2H; 2-H, minor isomer), 9.91 (s, 2H; 3-H, minor isomer), 10.09 ppm (s,
2H; 3-H, major isomer); 13C NMR (CDCl3, ꢀ508C): d=28.32 (d, C-2),
28.70 (d, C-2), 165.75 (d, C-3), 166.89 ppm (d, C-3). The ratio of diaster-
eomers was 3:2. Assignment of meso and rac isomers could not be per-
formed.
Photolyses of (E,E)-1,4-diazido-2,3-dimethyl-1,3-butadiene ((E,E)-8e): A
solution of (E,E)-8e in CDCl3 was irradiated in an NMR tube at ꢀ508C.
The progress of the reaction was monitored by NMR spectroscopy at
ꢀ508C. The maximum yield of (E)-17e (37%, determined by NMR spec-
troscopy) was reached after 20 min of irradiation. When DCA was used
as sensitizer, no new products could be found, but the maximum yield of
(E)-17e decreased to 26%.
A
IR (CHCl3): n˜ =2108 cmꢀ1 (N3), 1287 (N3); UV/Vis (Ch3CN): lmax (lge)=
255 nm (4.08); 1H NMR (CDCl3): d=1.66 (q, 5J=1.5 Hz, 6H; 3-Me/4-
Me), 1.86 ppm (q, 5J=1.5 Hz, 6H; 1-H/6-H); 13C NMR (CDCl3): d=
14.31 (q), 15.78 (q), 123.00 (s), 126.33 ppm (s). The assignment of the
1H NMR signals was possible by comparison of the 1H NMR signals of
(E,E)-8g, (E,E)-8j, (E,E)-8j’, (E,E)-8k, (E,E)-8k’, and (E,E)-8k’’.
Methyl groups in positions 2 and 3 of the 1,4-diazidobuta-1,3-dienes
showed chemical shifts in the range of 1.65–1.70 ppm, in contrast to
methyl groups in positions 1 and 4 with values in the range of 1.85–
1.90 ppm.
2-[(E)-2-Azido-1-methylethenyl]-2-methyl-2H-azirine ((E)-17e): Unsta-
ble, not isolable; 1H NMR (CDCl3, ꢀ508C): d=1.14 (d, 4J=1.3 Hz, 3H;
1’-Me), 1.31 (d, 4J=1.5 Hz, 3H; 2-Me), 6.20 (qd, 4J=1.3 Hz, 5J=0.8 Hz,
1H; 2’-H), 10.26 ppm (qd, 4J=1.5 Hz, 5J=0.8 Hz, 1H; 3-H); 13C NMR
(CDCl3, ꢀ508C): d=13.65 (q, Me-1’), 20.73 (q, Me-2), 33.65 (s, C-2),
2-[(E)-2-Azido-1-methylprop-1-enyl]-2,3-dimethyl-2H-azirine ((E)-17g):
Unstable, not isolable; H NMR (CDCl3): d=1.28 (s, 3H; 2-Me), 1.63 (q,
1
122.48 (d, C-2’), 128.31 (s, C-1’), 172.53 ppm (d, C-3). The H NMR signal
assignment of the methyl groups of (E)-17e was performed by double
resonance experiments. The configuration of (E)-17e was proved by
NOE difference spectroscopy. When the signal of 2’-H was irradiated,
only one NOE of 5.3% could be observed at the signal of 2-Me.
5J=1.5 Hz, 3H; 1’-Me), 2.12 (q, 5J=1.5 Hz, 3H; 3’-H), 2.44 ppm (s, 3H;
3-Me); 13C NMR (CDCl3): d=12.58 (q), 13.40 (q), 14.39 (q), 22.39 (q,
Me-2), 36.13 (s, C-2), 124.75 (s), 127.61 (s), 174.49 ppm (s, C-3). The as-
signment of the 1H NMR signals of 1’-Me and 3’-H was proved by a
NOESY NMR experiment.
Photolysis of 1,4-diazido-2,3-diphenylbuta-1,3-diene ((E,E)-8 f): In an
NMR tube, (E,E)-8 f (44.6 mg, 0.155 mmol) was dissolved in CDCl3
(700 mL) and irradiated at ꢀ508C. The reaction was monitored by NMR
spectroscopy at this temperature. The maximum percentage of (E)-17 f
(18%) was determined after 25 min of irradiation. Isomers 9 f (20%,
ratio of more stable isomer:less stable isomer=1.2:1) were detected after
40 min of photolysis. The less stable diastereomer was no longer detecta-
ble after 20 min at room temperature, whereas complete decomposition
of the more stable isomer could be observed only after a few hours at
room temperature.
A
diazido-3,4-dimethylhepta-2,4-diene ((E,E)-8j’): Mixture (E,E)-8j:(E,E)-
8j’=1:2.3, yellow liquid; IR (CDCl3): n˜ =2104 (N3), 1285 cmꢀ1 (N3);
1H NMR (CDCl3, (E,E)-8j): d=0.86 (t, 3J=7.6 Hz, 3H; 2’-H), 1.647 (q,
5J=1.5 Hz, 3H; 4-Me), 1.86 (t, 5J=0.9 Hz, 3H; 1-H), 1.87 (q, J=1.5 Hz,
3H; 6-H), 2.1–2.4 ppm (m, 2H; 1’-h); 1H NMR (CDCl3, (E,E)-8j’): d=
1.08 (t, J=7.6 Hz, 3H; 7-H), 1.65 (t, J=0.9 Hz, 3H; 4-Me), 1.66 (q, J=
1.5 Hz, 3H; 3-Me), 1.88 (q, J=1.5 Hz, 3H; 1-H), 2.1–2.4 ppm (m, 2H; 6-
H); 13C NMR (CDCl3, (E,E)-8j): d=12.45 (q), 14.43 (q), 14.67 (q), 16.26
(q), 22.87 (t), 121.62 (s), 125.84 (s), 127.33 (s), 129.15 ppm (s); 13C NMR
(CDCl3, (E,E)-8j’): d=12.67 (q), 14.55 (q), 16.19 (q), 16.46 (q), 22.55 (t),
122.90 (s), 123.02 (s), 126.32 (s), 132.35 ppm (s).
2-[(E)-2-Azido-1-phenylethenyl)]-2-phenyl-2H-azirine ((E)-17 f): Not
isolable; 1H NMR (ꢀ508C, CDCl3): d=6.70 (s, 1H), 7.20–7.70 (m, 10H),
10.08 ppm (s, 1H); 13C NMR (ꢀ508C, CDCl3): d=39.06 (s, C-2), 126.78
(d), 126.84 (d, 2C), 127.60 (s), 127.64 (d), 128.01 (d, 2C), 128.12 (d, 2C),
128.33 (d, 3C), 133.97 (s), 141.01 (s, i-Ph at C-2), 162.93 ppm (d, C-3).
ACHTREUNG
meso-/rac-2,2’-Diphenylbi-2H-azirin-2-yl (meso-9 f/rac-9 f): Unstable, not
methylocta-3,5-diene ((E,E)-8k’’): Mixture (E,E)-8k:
C
G
1
isolable; more stable isomer: H NMR (ꢀ508C, CDCl3): d=7.20–7.60 (m,
10H), 9.85 (s, 2H); 13C NMR (ꢀ508C, CDCl3): d=39.09 (s, C-2), 126.42
(d), 127.12 (d, p-Ph), 128.24 (d), 139.35 (s, i-Ph), 158.16 (d, C-3); less
5
5
1
((E,E)-8k’’)], 1.66 (q, J=1.5 Hz, 3-Me, (E,E)-8k’), 1.67 (t, J=0.8 Hz, 4-
Me/5-Me, (E,E)-8k’’), 1.87 (s, 1-H/6-H, (E,E)-8k), 1.89 (q, 5J=1.5 Hz, 1-
H, (E,E)-8k’), 1.9–2.4 ppm [m, 1’-H ((E,E)-8k)+6-H ((E,E)-8k’)+1’-H
stable isomer: H NMR (ꢀ508C, CDCl3): d=7.20–7.60 (m, 10H), 9.80 (s,
2H); 13C NMR (ꢀ508C, CDCl3): d=38.59 (s, C-2), 126.15 (d), 127.08 (d,
p-Ph), 128.19 (d), 139.75 (s, i-Ph), 157.20 (d, C-3).
7478
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 7467 – 7481

Valence Isomerization of Bi-2H-azirin-2-yls to Diazabenzenes
FULL PAPER
Photochemical syntheses of tetraalkylbi-2H-azirin-2-yls 9g, 9j, and 9j’ as
well as 9k, 9k’, and 9k’’: The photochemical transformations of buta-
dienes (E,E)-8g, (E,E)-8j, and (E,E)-8j’ as well as (E,E)-8k, (E,E)-8k’,
and (E,E)-8k’’ were carried out in solution (CDCl3) at temperatures be-
tween ꢀ50 and ꢀ608C. The yields were quantitative. The ratios of 9j:9j’
obtained after photolysis were in accordance with the ratios of the start-
Thermal reactions of tetraalkylbi-2H-azirin-2-yls 9g, 9j, and 9j’ as well
as 9k, 9k’, and 9k’’: After photolyses of the corresponding diazides
(E,E)-8 at ꢀ50 to ꢀ608C, the resulting solutions of 9g, 9j, and 9j’ as well
as 9k, 9k’, and 9k’’ were warmed to temperatures in the range of ꢀ25 to
208C to yield the pyridazines 18 quantitatively. The ratio of the starting
materials 9j and 9j’ were in accordance with the ratio of the thermal
products 18j and 18j’. The same held true for the ratios 9k:9k’:9k’’ and
18k:18k’:18k’’. The heterocyclic aromatic products were isolated by re-
moving the solvent in vacuo. However, if AgBF4 is present, the reaction
mixture must first be washed with aqueous NaBr. Products 18g and 10g
resulting from thermal, silver-ion-induced, or photochemical reactions
were identified by comparing their 1H NMR and 13C NMR data to those
of independently prepared[26,29] samples. These sets of data can be un-
equivocally distinguished from each other and from those of 11g, which
was also synthesized[45] for control experiments including photostability.
ing material (E,E)-8j:
ACHTREUNG
9k:9k’:9k’’ and (E,E)-8k:
A
ACHTREUNG
2,2’,3,3’-Tetramethylbi-2H-azirin-2-yl (9g) (more stable isomer, most
probably meso compound): IR (CDCl3): n˜ =1753 cmꢀ1 (C=N); 1H NMR
(CDCl3, ꢀ508C): d=0.93 (s, 6H; 2-Me/2’-Me), 2.47 ppm (s, 6H; 3-Me/3’-
Me); 13C NMR (CDCl3, ꢀ508C): d=14.53 (q, Me-3/Me-3’), 19.78 (q, Me-
2/Me-2’), 38.95 (s, C-2/C-2’), 176.25 ppm (s, C-3/C-3’).
2,2’,3,3’-Tetramethylbi-2H-azirin-2-yl (9g) (less stable isomer, most prob-
ably rac compound): 1H NMR (CDCl3, ꢀ508C): d=1.16 (s, 6H; 2-Me/2’-
Me), 2.38 ppm (s, 6H; 3-Me/3’-Me); 13C NMR (CDCl3, ꢀ508C): d=14.71
(q, Me-3/Me-3’), 20.25 (q, Me-2/Me-2’), 38.93 (s, C-2/C-2’), 173.75 ppm (s,
C-3/C-3’).
1:2 Mixture of 4-ethyl-3,5,6-trimethylpyridazine (18j) and 3-ethyl-4,5,6-
trimethylpyridazine (18j’): Colorless liquid; MS (ESI): m/z (%): 151.1229
(100) [M+H+; calcd 151.1230], 323.2129 (17) [2M+Na+; calcd 323.2206];
elemental analysis calcd (%) for C9H14N2: C 71.96, H 9.39, N 18.65;
found: C 71.48, H 9.07, N 18.87.
18j: 1H NMR (CDCl3): d=1.11 (t, 3J=7.6 Hz, 3H; 2’-H), 2.21 (s, 3H;
Me-5), 2.57 (s, 3H), 2.61 (s, 3H), 2.63 ppm (q, 3J=7.6 Hz, 2H; 1’-H);
13C NMR (CDCl3): d=12.43 (q), 13.69 (q), 20.01 (q), 20.77 (q), 21.52 (t,
C-1’), 133.56 (s, C-5), 139.37 (s, C-4), 156.85 (s), 157.88 ppm (s).
18j’: 1H NMR (CDCl3): d=1.28 (t, 3J=7.5 Hz, 3H; 2’-H), 2.19 (s, 3H),
2.22 (s, 3H), 2.58 (s, 3H; 6-Me), 2.93 ppm (q, 3J=7.5 Hz, 2H; 1’-H);
13C NMR (CDCl3): d=12.93 (q), 13.78 (q), 14.39 (q), 20.79 (q, Me-6),
27.31 (t, C-1’), 133.55 (s), 134.52 (s), 157.16 (s, C-6), 161.40 ppm (s, C-3).
2-Ethyl-2’,3,3’-trimethylbi-2H-azirin-2-yl (9j) (more stable isomer, most
probably unlike compound): 1H NMR (CDCl3, ꢀ508C): d=0.49 (t, 3J=
7.6 Hz, 3H; CH3CH2), 0.97 (s, 3H; 2’-Me), 1.48–1.60 (m, 2H; CH2), 2.44
(s, 3H), 2.45 ppm (s, 3H); 13C NMR (CDCl3, ꢀ508C): d=10.30 (q,
CH3CH2), 14.32 (q), 15.13 (q), 19.80 (q, Me-2’), 24.00 (t, CH2), 38.54 (s,
C-2’), 43.81 (s, C-2), 174.93 (s), 175.66 ppm (s).
2-Ethyl-2’,3,3’-trimethylbi-2H-azirin-2-yl (9j) (less stable isomer, most
probably like compound): 1H NMR (CDCl3, ꢀ508C): d=0.61 (t, 3J=
7.5 Hz, 3H; CH3CH2), 1.18 (s, 3H; 2’-Me), 1.70–1.84 (m, 2H; CH2), 2.37
(s, 3H), 2.40 ppm (s, 3H); 13C NMR (CDCl3, ꢀ508C): d=10.20 (q,
CH3CH2), 14.60 (q), 15.38 (q), 20.49 (q, Me-2’), 24.77 (t, CH2), 38.01 (s,
C-2’), 43.45 (s, C-2), 172.03 (s), 172.80 ppm (s).
1:2:1 Mixture of 4,5-diethyl-3,6-dimethylpyridazine (18k), 3,4-diethyl-5,6-
dimethylpyridazine (18k’), and 3,6-diethyl-4,5-dimethylpyridazine (18k’’):
Colorless liquid; MS (ESI): m/z (%): 165.1377 (100) [M+H+; calcd
165.1386]; elemental analysis calcd (%) for c10H16N2: C 73.13, H 9.82, N
17.06; found: C 72.75, H 9.79, N 17.24.
3-Ethyl-2,2’,3’-trimethylbi-2H-azirin-2-yl (9j’) (more stable isomer, most
probably unlike compound): 1H NMR (CDCl3, ꢀ508C): d=0.93 (s, 3H),
0.94 (s, 3H), 1.26 (t, 3J=7.4 Hz, 3H; CH3CH2), 2.46 (s, 3H; Me-3’), 2.76
(q, 3J=7.4 Hz, 1H; CH2), 2.78 ppm (q, 3J=7.4 Hz, 1H; CH2); 13C NMR
(CDCl3, ꢀ508C): d=8.59 (q, CH3CH2), 14.53 (q, Me-3’), 19.79 (q), 20.04
(q), 21.96 (t, CH2), 39.01 (s), 39.84 (s), 176.29 (s, C-3’), 179.83 ppm (s, C-
3).
1
3
18k: H NMR (CDCl3): d=1.17 (t, J=7.7 Hz, 6H; 2’-H), 2.65 (s, 6H; 3-
Me/6-Me), 2.65 ppm (q, 3J=7.7 Hz, 4H; 1’-H); 13C NMR (CDCl3): d=
13.50 (q, C-2’), 20.07 (q, Me-3/Me-6), 21.11 (t, C-1’), 138.82 (s, C-4/C-5),
157.44 ppm (s, C-3/C-6).
18k’: 1H NMR (CDCl3): d=1.14 (t, 3J=7.6 Hz, 3H; CH3CH2 at C-4),
3-Ethyl-2,2’,3’-trimethylbi-2H-azirin-2-yl (9j’) (less stable isomer, most
probably like compound): 1H NMR (CDCl3, ꢀ508C): d=1.16 (s, 3H),
1.17 (s, 3H), 1.22 (t, 3J=7.4 Hz, 3H; CH3CH2), 2.34 (s, 3H; Me-3’),
2.66 ppm (q, 3J=7.4 Hz, 2H; CH2); 13C NMR (CDCl3, ꢀ508C): d=8.33
(q, CH3CH2), 14.69 (q, Me-3’), 20.31 (q), 20.57 (q), 22.08 (t, CH2), 38.86
(s), 39.78 (s), 173.53 (s, C-3’), 177.16 ppm (s, C-3).
3
1.34 (t, J=7.5 Hz, 3H; CH3CH2 at C-3), 2.23 (s, 3H; Me-5), 2.60 (s, 3H;
Me-6), 2.67 (q, 3J=7.6 Hz, 2H; CH2 at C-4), 2.94 ppm (q, 3J=7.5 Hz,
2H; CH2 at C-3); 13C NMR (CDCl3): d=13.31 (q), 13.80 (q), 13.85 (q),
20.87 (q, Me-6), 20.94 (t, CH2 at C-4), 26.46 (t, CH2 at C-3), 133.78 (s, C-
5), 138.78 (s, C-4), 157.64 (s, C-6), 161.08 ppm (s, C-3).
1
3
18k’’: H NMR (CDCl3): d=1.34 (t, J=7.5 Hz, 6H; 2’-H), 2.25 (s, 6H; 4-
Me/5-Me), 2.96 ppm (q, 3J=7.5 Hz, 4H; 1’-H); 13C NMR (CDCl3): d=
12.84 (q), 13.91 (q), 27.29 (t, C-1’), 133.86 (s, C-4/C-5), 161.14 ppm (s, C-
3/C-6).
2,2’-Diethyl-3,3’-dimethylbi-2H-azirin-2-yl (9k), 2,3-diethyl-2’,3’-dimethyl-
bi-2H-azirin-2-yl (9k’), 3,3’-diethyl-2,2’-dimethylbi-2H-azirin-2-yl (9k’’)
(more stable isomers, most probably meso or unlike compounds):
1H NMR (CDCl3, ꢀ508C): d=0.52 (t, 3J=7.3 Hz, CH3CH2, 9k), 0.54 (t,
3J=7.4 Hz, CH3CH2 at C-2, 9k’), 0.94 (s, 2-Me/2’-Me, 9k’’), 1.01 (s, 2’-Me,
9k’), 1.29 [t, 3J=7.5 Hz, CH3CH2 at C-3 (9k’)+CH3CH2 (9k’’)], 1.4–1.9
(m, CH2 (9k)+CH2 at C-2 (9k’)), 2.46 (s, 3’-Me, 9k’), 2.47 (s, 3-Me/3’-
Me, 9k), 2.7–2.9 ppm [m, CH2 at C-3 (9k’)+CH2 (9k’’)]; 13C NMR
(CDCl3, ꢀ508C): d=8.56 (q), 8.72 (q), 10.23 (q), 10.32 (q), 14.22 (q),
14.89 (q), 19.71 (q), 19.99 (q), 21.86 (t), 22.25 (t), 24.00 (t), 24.29 (t),
38.47 (s), 39.88 (s), 43.37 (s), 44.85 (s), 174.31 (s), 175.43 (s), 178.50 (s),
179.69 ppm (s).
Reaction of trans-3,4-dibromo-1,2,3,4-tetraphenylcyclobutene [15c (X=
Br)] with QN3: In
a
typical experiment, 15c (X=Br)[12] (21.5 mg,
43.3 mmol) and dioxane (2 mL as standard) were dissolved in CDCl3
(0.7 mL). To this mixture, a solution of QN3 (50.0 mg, 0.106 mmol) in
CDCl3 (0.3 mL) was added at ꢀ508C. The reaction was monitored by
NMR spectroscopy while increasing the temperature stepwise from
ꢀ508C to 608C. Tetraphenylpyridazine 18c (87%) was obtained as the
stable final product and identified by comparing its 1H NMR and
13C NMR data to those of an independently prepared[31] sample. These
data can be unequivocally distinguished from those of 10c[32] and 11c,[33]
which were also synthesized for control experiments including photosta-
bility. The azirine 17c as well as bi-2H-azirin-2-yls meso- and rac-9c
could be observed as intermediates by NMR spectroscopy. The maximum
amount of 17c was 87%.
2,2’-Diethyl-3,3’-dimethylbi-2H-azirin-2-yl (9k), 2,3-diethyl-2’,3’-dimethyl-
bi-2H-azirin-2-yl (9k’), 3,3’-diethyl-2,2’-dimethylbi-2H-azirin-2-yl (9k’’)
(less stable isomers, most probably rac or like compounds): 1H NMR
(CDCl3, ꢀ508C): d=0.677 (t, 3J=7.8 Hz, CH3CH2 at C-2, 9k’), 0.683 (t,
3J=7.5 Hz, CH3CH2, 9k), 1.21 (s, 2-Me/2’-Me, 9k’’), 1.23 (s, 2’-Me, 9k’),
1.24 (t, 3J=7.3 Hz, CH3CH2, 9k’’), 1.27 (t, 3J=7.5 Hz, CH3CH2 at C-3,
9k’), 1.4–1.9 (m, CH2 (9k)+CH2 at C-2 (9k’)), 2.39 (s, 3’-Me, 9k’), 2.42
(s, 3-Me/3’-Me, 9k), 2.7–2.9 ppm (m, CH2 at C-3 (9k’)+CH2 (9k’’));
13C NMR (CDCl3, ꢀ508C): d=8.33 (q), 8.35 (q), 10.09 (q), 10.15 (q),
14.50 (q), 15.13 (q), 20.54 (q), 20.59 (q), 21.98 (t), 22.45 (t), 25.04 (t),
25.10 (t), 37.72 (s), 39.64 (s), 42.27 (s), 44.29 (s), 171.17 (s), 172.38 (s),
175.47 (s), 176.85 ppm (s).
2-(2-Azido-1,2-diphenylethenyl)-2,3-diphenyl-2H-azirine
(17c):
IR
(CDCl3): n˜ =2110 cmꢀ1 (N3), 1738 (C=N), 1255 (N3); 1H NMR (CDCl3):
d=6.88–7.63 ppm (m); 13C NMR (CDCl3): d=44.78 (s, C-2), 123.50 (s),
126.11 (d), 126.37 (d), 127.24 (d), 127.68 (d) 127.75 (d), 128.25 (d), 128.40
(d), 128.53 (d), 128.95 (d), 129.09 (d), 129.75 (d), 131.86 (d), 133.26 (s),
137.63 (s), 139.04 (s), 142.19 (s), 165.51 ppm (s, C-3), one signal (s) was
Chem. Eur. J. 2006, 12, 7467 – 7481
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7479

K. Banert, S. Grimme, R. Herges, E.-U. Würthwein et al.
hidden by another. The formation of only one stereoisomer could be ob-
served.
[20] M. Sharon, A. Parker, C. J. Rizzo, Synth. Commun. 1995, 25, 2781–
2789.
[21] a) A. T. Blomquist, E. A. LaLancette, J. Org. Chem. 1964, 29, 2331–
2334; b) J. Hambrecht, H. Straub, E. Müller, Chem. Ber. 1974, 107,
3962–3965.
[22] For another example of nucleophilic substitution starting with cis-
15d (X=Cl), see W. Kirmse, F. Scheidt, H.-J. Vater, J. Am. Chem.
Soc. 1978, 100, 3945–3946.
meso-/rac-2,2’,3,3’-Tetraphenylbi-2H-azirin-2-yl
(meso-9c/rac-9c):
1H NMR (CDCl3): d=6.80–7.80 ppm (m); 13C NMR (CDCl3): d=45.49
(s, C-2/C-2’), 45.55 (s, C-2/C-2’), 122.73 (s), 123.14 (s), 126.40 (d), 126.69
(d), 126.80 (d) 127.25 (d), 127.93 (d), 128.10 (d), 128.80 (d), 129.12 (d),
129.61 (d), 129.63 (d), 132.87 (d), 132.97 (d), 139.77 (s), 140.70 (s), 163.72
(s, C-3/C-3’), 163.98 ppm (s, C-3/C-3’).
[23] a) D. Landini, A. Maia, F. Montanari, J. Am. Chem. Soc. 1978, 100,
2796–2801; b) D. Landini, A. Maia, F. Montanari, Nouv. J. Chim.
1979, 3, 575–577; c) D. Landini, A. Maia, F. Montanari, F. Rolla, J.
Org. Chem. 1983, 48, 3774–3777; d) K. Banert, W. Kirmse, J. Am.
Chem. Soc. 1982, 104, 3766–3767; e) K. Banert, Chem. Ber. 1985,
118, 1564–1574.
Acknowledgements
[24] W. Kirmse, N. G. Rondan, K. N. Houk, J. Am. Chem. Soc. 1984, 106,
7989–7991.
[25] A. T. Blomquist, Y. C. Meinwald, J. Am. Chem. Soc. 1959, 81, 667–
672.
Financial support of this work by the Deutsche Forschungsgemeinschaft
[Ba 903/10-1/2 (Chemnitz) and SFB 424 (Münster)] and the Fonds der
Chemischen Industrie is gratefully acknowledged. We thank also Dr. J.
Lehmann for support.
[26] G. Maier, M. Schneider, G. Kreiling, W. Mayer, Chem. Ber. 1981,
114, 3922–3934.
[27] R. S. Hansen, W. S. Trahanovsky, J. Org. Chem. 1974, 39, 570–571.
[28] A. Gagneux, S. Winstein, W. G. Young, J. Am. Chem. Soc. 1960, 82,
5956–5957.
[29] W. Zielinski, Heterocycles 1985, 23, 1639–1644.
[30] a) F. W. Fowler, A. Hassner, L. A. Levy, J. Am. Chem. Soc. 1967, 89,
2077–2082; b) K. Banert, M. Hagedorn, C. Liedtke, A. Melzer, C.
Schçffler, Eur. J. Org. Chem. 2000, 257–267.
[1] a) W. E. Billups, M. M. Haley, Angew. Chem. 1989, 101, 1735–1737;
Angew. Chem. Int. Ed. Engl. 1989, 28, 1711; b) R. Boese, D. Bläser,
R. Gleiter, K.-H. Pfeifer, W. E. Billups, M. M. Haley, J. Am. Chem.
Soc. 1993, 115, 743–746; c) W. E. Billups, M. M. Haley, Tetrahedron
1994, 50, 10693–10700; d) W. E. Billups, D. J. McCord, Angew.
Chem. 1994, 106, 1394–1406; Angew. Chem. Int. Ed. Engl. 1994, 33,
1332.
[2] A. Greenberg, J. F. Liebman, J. Am. Chem. Soc. 1981, 103, 44–48.
[3] J. Spanget-Larsen, C. de Korswagen, M. Eckert-Maksic, R. Gleiter,
Helv. Chim. Acta 1982, 90, 968–982.
[4] a) R. Breslow, P. Gal, J. Am. Chem. Soc. 1959, 81, 4747–4748; b) R.
Breslow, P. Gal, H. W. Chang, L. J. Altman, J. Am. Chem. Soc. 1965,
87, 5139–5144.
[5] a) J. H. Davis, K. J. Shea, R. G. Bergman, Angew. Chem. 1976, 88,
254–255; Angew. Chem. Int. Ed. Engl. 1976, 15, 232; b) J. H. Davis,
K. J. Shea, R. G. Bergman, J. Am. Chem. Soc. 1977, 99, 1499–1507
and references therein; c) W. H. de Wolf, I. J. Landheer, F. Bickel-
haupt, Tetrahedron Lett. 1975, 179–182.
[6] For summaries, see: a) A. Padwa, S. I. Goldstein, R. J. Rosenthal, J.
Org. Chem. 1987, 52, 3278–3285; b) F. Bickelhaupt, W. H. de Wolf,
Recl. Trav. Chim. Pays-Bas 1988, 107, 459–478.
[31] R. A. Carboni, R. V. Lindsey, J. Am. Chem. Soc. 1959, 81, 4342–
4346.
[32] a) P.-L. Compagnon, F. Gasquez, O. Compagnon, T. Kimny, Bull.
Soc. Chim. Belg. 1982, 91, 931–946; b) A. Garcia, E. Lete, M. J.
Villa, E. Dominguez, M. D. Badia, Tetrahedron 1988, 44, 6681–6686.
[33] K. N. Mehrotra, G. Singh, Synthesis 1980, 1001–1004.
[34] For summaries, see a) V. Nair in The Chemistry of Heterocyclic
Compounds, Small-Ring Heterocycles, Vol. 42 (Ed.: A. Hassner),
Wiley, New York, 1983, Part 1, pp. 215–332; b) J. Backes in
Houben-Weyl, Vol. E16c (Ed.: D. Klamann), Thieme, Stuttgart,
1992, pp. 317–369.
[35] J. H. Boyer, W. E. Krueger, R. Modler, Tetrahedron Lett. 1968,
5979–5982.
[36] Gaussian98, RevisionA.11.4, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski,
J. A. Montgomery, R. E. Stratman, J. C. Burant, S. Drappich, J. M.
Millam, A. D. Daniels, M. C. Kudin, M. C. Strain, O. Farkas, J.
Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli,
C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala, Q.
Cui, K. Morokuma, N. Rega, P. Salvador, J. J. Dannenberg, D. K.
Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. Cioslow-
ski, J. V. Ortiz, A. G. Baboul, B. B. Stefanov, G. Liu, A. Liashenko,
P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin, D. J. Fox, T.
Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, C. Gonzalez,
M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong,
J. L. Andres, C. Gonzalez, M. Head-Gordon, E. S. Replogle, J. A.
Pople, Gaussian, Inc., Pittsburgh PA, 2002.
[37] T. Bally, W. T. Borden in Reviews in Computational Chemistry,
Vol. 13 (Eds.: K. B. Lipkowitz, D. B. Boyd), Wiley-VCH, New York,
1999, pp. 1–97.
[38] For a related problem, see D. A. Hrovat, W. T. Borden, J. Am.
Chem. Soc. 2001, 123, 4069–4072.
[39] S. Grimme, M. Waletzke, Phys. Chem. Chem. Phys. 2000, 2, 2075–
2081.
[40] a) K. Yamaguchi, F. Jensen, A. Dorigo, K. N. Houk, Chem. Phys.
Lett. 1988, 149, 537–542; b) S. Yamanaka, T. Kawakami, H. Nagao,
K. Yamaguchi, Chem. Phys. Lett. 1994, 231, 25–33.
[7] Work on compounds 8 and 9 was reported in a preliminary short
communication: K. Banert, F. Kçhler, Angew. Chem. 2001, 113,
173–176; Angew. Chem. Int. Ed. 2001, 40, 174–177.
[8] a) A. Padwa, T. J. Blacklock, P. H. J. Carlsen, M. Pulwer, J. Org.
Chem. 1979, 44, 3281–3287; b) A. Padwa, J. Smolanoff, A. Tremper,
J. Org. Chem. 1976, 41, 543–549.
[9] T. C. Gallagher, R. C. Storr, Tetrahedron Lett. 1981, 22, 2905–2908.
[10] A. Hassner, J. Keogh, Tetrahedron Lett. 1975, 1575–1578.
[11] a) G. L’abbꢃ, A. Hassner, Angew. Chem. 1971, 83, 103–109; Angew.
Chem. Int. Ed. Engl. 1971, 10, 98–104; b) G. Smolinsky, C. A. Pryde
in The Chemistry of the Azido Group (Ed.: S. Patai), Wiley, London,
1971, pp. 555–585; c) A. Hassner in Azides and Nitrenes: Reactivity
and Utility (Ed.: E. F. V. Scriven), Academic Press, Orlando, 1984,
pp. 35–94; d) K. Banert in Houben-Weyl, Vol. E15 (Eds.: H. Kropf,
E. Schaumann), Thieme, Stuttgart, 1993, pp. 818–875, 1344–1347,
2348–2349, 3105–3107.
[12] H. Hoberg, C. Frçhlich, J. Organomet. Chem. 1980, 197, 105–109.
[13] H. Hoberg, C. Frçhlich, Synthesis 1981, 830–831.
[14] H. A. Brune, W. Schwab, Tetrahedron 1969, 25, 4375–4378.
[15] G. Metha, R. Uma, Indian J. Chem. Sect. B 1999, 38, 1154–1158.
[16] R. Criegee, A. Moschel, Chem. Ber. 1959, 92, 2181–2184.
[17] E. Müller, A. Huth, Chem.-Ztg. 1972, 96, 585–586.
[18] H. A. Brune, H. P. Wolff, H. Hüther, Tetrahedron 1971, 27, 3949–
3957.
[41] Compare: a) B. Bigot, A. Sevin, A. Devaquet, J. Am. Chem. Soc.
1978, 100, 6924–6929; b) M. Klessinger, C. Bornemann, J. Phys.
Org. Chem. 2002, 15, 514–518; c) C. Bornemann, M. Klessinger,
Chem. Phys. 2000, 259, 263–271.
[19] a) H. A. Brune, H. P. Wolff, Tetrahedron 1968, 24, 4861–4864;
b) H. A. Brune, H. P. Wolff, H. Hüther, Z. Naturforsch. B 1971, 26,
765–770; c) H. A. Brune, W. Eberius, H. P. Wolff, J. Organomet.
Chem. 1968, 12, 485–490.
7480
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 7467 – 7481

Valence Isomerization of Bi-2H-azirin-2-yls to Diazabenzenes
FULL PAPER
[42] A. Melzer, Ph.D. Dissertation, Chemnitz University of Technology,
2001.
[45] W.-x. Chen, J.-h. Zhang, M.-y. Hu, X.-c. Wang, Synthesis 1990, 701–
702.
[43] Y. Rubin, C. B. Knobler, F. Diederich, J. Am. Chem. Soc. 1990, 112,
1607–1617.
[44] L. Brandsma, H. D. Verkruijsse, Synthesis of Acetylenes, Allenes and
Cumulenes, Elsevier, Amsterdam, 1981, p. 11.
Received: March 7, 2006
Published online: July 3, 2006
Chem. Eur. J. 2006, 12, 7467 – 7481
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7481