Photochemical reactions of nitroso oxides at low temperatures: The first experimental evidence for dioxaziridines

Article abstract of DOI:10.1021/ja9803660

Several singlet nitroso oxides (3a - e) were generated by the thermal reaction of triplet nitrenes (2a - e) with triplet oxygen at 95 K in 2-methyltetrahydrofuran. After photolysis of the nitroso oxides at 77 K using strong light intensities, the formation of intermediates was observed for the first time. From spectroscopic, kinetic, and chemical arguments, we postulate the formation of the following dioxaziridines: 4-(dioxaziridinyl)stilbene (4a), 4-(dioxaziridin-yl)-4′-nitrostilbene (4b), 4′-(dioxaziridin-yl)-4-(dimethylamino)stilbene (4c), 4′-(dioxaziridin-yl)-4-aminobiphenyl (4d), and 4-(dioxaziridin-yl)-4′-(nitrene-substituted)stilbene (4f). All dioxaziridines observed are highly reactive species. At 77 K, they react thermally to form the corresponding nitro compounds (5). The velocity of the ring opening reaction of the dioxaziridines (4 → 5) is not significantly influenced by substituents; the rate constants at 77 K are all equal to 0.0030 ± 0.0005 s^-1. The transients were characterized by stationary UV/vis and/or ESR spectroscopy. Ab initio calculations of the thermal reaction of the nonsubstituted dioxaziridine (6) and N-phenyldioxaziridine (9) were performed. From this, it follows that dioxaziridines are experimentally observable species which are separated from the corresponding nitro products by an orbital symmetry-forbidden barrier.

Full text of DOI:10.1021/ja9803660

6580
J. Am. Chem. Soc. 1999, 121, 6580-6588
Photochemical Reactions of Nitroso Oxides at Low Temperatures:
The First Experimental Evidence for Dioxaziridines
T. Harder, P. Wessig, J. Bendig,* and R. Sto1sser
Contribution from the Institute of Chemistry, Humboldt UniVersity Berlin, Hessische Strasse 1-2,
10115 Berlin, Germany
ReceiVed February 2, 1998. ReVised Manuscript ReceiVed March 25, 1999
Abstract: Several singlet nitroso oxides (3a-e) were generated by the thermal reaction of triplet nitrenes
(2a-e) with triplet oxygen at 95 K in 2-methyltetrahydrofuran. After photolysis of the nitroso oxides at 77 K
using strong light intensities, the formation of intermediates was observed for the first time. From spectroscopic,
kinetic, and chemical arguments, we postulate the formation of the following dioxaziridines: 4-(dioxaziridin-
yl)stilbene (4a), 4-(dioxaziridin-yl)-4′-nitrostilbene (4b), 4′-(dioxaziridin-yl)-4-(dimethylamino)stilbene (4c),
4′-(dioxaziridin-yl)-4-aminobiphenyl (4d), and 4-(dioxaziridin-yl)-4′-(nitrene-substituted)stilbene (4f). All
dioxaziridines observed are highly reactive species. At 77 K, they react thermally to form the corresponding
nitro compounds (5). The velocity of the ring opening reaction of the dioxaziridines (4 f 5) is not significantly
influenced by substituents; the rate constants at 77 K are all equal to 0.0030 ( 0.0005 s-1. The transients were
characterized by stationary UV/vis and/or ESR spectroscopy. Ab initio calculations of the thermal reaction of
the nonsubstituted dioxaziridine (6) and N-phenyldioxaziridine (9) were performed. From this, it follows that
dioxaziridines are experimentally observable species which are separated from the corresponding nitro products
by an orbital symmetry-forbidden barrier.
Introduction
amounts to about 45 kcal/mol.6 It is therefore assumed that a
possible intramolecular conversion of nitroso oxides into nitro
compounds via dioxaziridines has to be initialized photochemi-
cally.
Organic chemists have been interested in the properties of
ROO species such as carbonyl oxides, dioxiranes, nitroso oxides,
and dioxaziridines for years.1-3 While a large amount of experi-
mental and theoretical information about the carbon species is
available, little is known about the analogous nitrogen inter-
mediates. Trapping experiments indicate that nitroso oxides are
formed by the reaction of p-nitrophenylnitrene and oxygen at
room temperature.2 Furthermore nitroso oxides were detected
at low temperatures (77-120 K). They are spectroscopically
characterized by broad long-wavelength UV/vis absorption
bands.4,5
Only little information is available about the reactions of
nitroso oxides. After excitation of nitroso oxides at 77 K, the
exclusive formation of the corresponding nitro compounds was
found by product analysis and by spectroscopic results.4,5 In
contrast to the photochemical reaction at 77 K, the thermal
reaction of nitroso oxides (3) at room temperature is reported
to result in the formation of both the corresponding nitro (5)
and the nitroso compound. It is assumed that the nitro compound
is formed via a dioxaziridine intermediate.2 But the existence
of dioxaziridines could not be confirmed so far by direct
spectroscopic observations.
In this paper, we report on the generation of nitroso oxides
at low temperatures, on their UV/vis and ESR spectroscopic
properties, and on their photochemical reactions. Analyzing the
photochemical reactions of nitroso oxides in detail, we found
the first experimental evidence for the existence of dioxaziri-
dines. On the basis of their UV/vis spectra, we present the
kinetics of the thermal reactions of the dioxaziridines 4a-d and
4f. Furthermore, we present our results of ab initio calculations
(obtained using model compounds) of the thermal reactions of
dioxaziridines. They include molecular structures and geo-
metrical parameters of the involved states, their thermodynamic
properties, and the stereoelectronic course of the ring opening
of dioxaziridines.
The experiments were carried out under conditions which
allow separation of all photochemical and thermal processes
from each other. This was realized by photolyzing exclusively
azides, which form at 77 K only the corresponding triplet
nitrenes (without subsequent thermal and photochemical reac-
tions). This is the most important precondition for the success
of our experiments, but this condition is not met in the case of
phenylnitrene derivatives. On the other hand, the large molecules
used in our experiments cannot be treated by quantum chemical
calculations at a high level of theory. Qualitative results were
therefore obtained by theoretical investigations of C6 derivatives.
From theoretical investigations, it follows that the activation
barrier of the conversion of nitroso oxides into dioxaziridines
(1) Sawaki, Y.; Ishikawa, S.; Iwamura, H. J. Am. Chem. Soc. 1987, 109,
584.
(2) Ishikawa, S.; Nojima, T.; Sawaki, Y. J. Chem. Soc., Perkin Trans. 2
1996, 127.
(3) Casal, H. L.; Sugamori, S. E.; Scaiano, J. C. J. Am. Chem. Soc. 1984,
106, 7623. (b) Scaiano, J. C.; McGimpsey, W. G.; Casal, H. L. J. Org.
Chem. 1989, 54, 1612. (c) Sander, W. J. Org. Chem. 1989, 54, 333. (d)
Chapman, O. L.; Hess, T. C. J. Am. Chem. Soc. 1984, 106, 1842. (e) Liang,
T.-Y.; Schuster, G. B. J. Am. Chem. Soc. 1987, 109, 7803.
(4) Brinen, J. S.; Singh, B. J. Am. Chem. Soc. 1971, 93, 6623.
(5) Harder, T.; Bendig, J.; Scholz, G.; Sto¨sser, R. J. Inf. Rec. Mater.
1996, 23, 147.
Synthesis
The azides 1a-e were synthesized as shown in Figure 1. At
first, the corresponding nitro compounds were synthesized by
the reaction of 4-(nitrophenyl)acetic acid with the corresponding
(6) Fueno, T.; Yokoyama, K.; Takanc, S. Theor. Chim. Acta 1992, 82,
299.
10.1021/ja9803660 CCC: $18.00 © 1999 American Chemical Society
Published on Web 06/30/1999

Photochemical Reactions of Nitroso Oxides
J. Am. Chem. Soc., Vol. 121, No. 28, 1999 6581
Figure 1. Synthesis of the azides and their precursors: (a) piperidine as a catalyst, 140 °C, 2 h; (b) 80% aqueous ethanol, FeSO4/NH3, 100 °C, 2
h; (c) aqueous HCl/NaNO2, 0 °C, NaN3; and (d) 85% aqueous ethanol, NaHS, 70 °C.
Scheme 1
4-substituted benzaldehyde using piperidine as a catalyst (T ≈
140 °C). Alternatively, 4,4′-dinitrostilbene can also be synthe-
sized by the treatment of 4-nitrobenzyl chloride with KOH. The
synthesis of 4-aminostilbene and 4-amino-4′-(dimethylamino)-
stilbene was realized by a reduction of the corresponding nitro
compounds using FeSO4/NH3. The 4-amino-4′-nitrostilbene was
synthesized by a partial reduction of 4,4′-dinitrostilbene using
sodium sulfide which was saturated with hydrogen sulfide
before. Azides 1a-c, diazide 1e, and 4,4′-diazidobiphenyl were
synthesized by the diazotation of the corresponding amines or
diamines (4,4′-diaminostilbene and 4,4′-diaminobiphenyl were
obtained from Aldrich Chemical Co.) and by further treatment
with sodium azide at 0 °C. Azide 1d was obtained by a partial
reduction of 4,4′-diazidobiphenyl using sodium sulfide which
Scheme 2
was saturated with hydrogen sulfide before. All syntheses were
carried out under yellow light; the azides were stored in a
refrigerator at 5 °C in aluminum foil-wrapped vials.
Results and Discussion
Photochemistry and Spectroscopy. The investigated reaction
cascade involving all photochemical and thermal steps and
including the pathway for generating 4a-d is illustrated in
Scheme 2. The specific procedure for synthesizing species 4f
is shown in Scheme 3.
formed.8 As an example, these reaction steps are illustrated in
Figure 2 using the UV/vis spectroscopic investigations of the
The triplet nitrenes 2a-d were formed after the photochemi-
reactions of 4-azidostilbene (1a). These are as follows. The
cal excitation of the corresponding azides 1a-d at 77 K in
excitation of 1a (λmax ) 319, 333, and 350 nm) at 77 K in
MTHF.5,7 The samples containing the nitrenes were then
MTHF leads to the quantitative formation of the corresponding
4-nitrene-substituted stilbene (2a, λmax ) 369, 378, 614, and
immediately warmed to 95 K by a temperature jump (5 K/min).
In this way, only nitroso oxides 3a-d but no amines were
687 nm, Figure 2a; ESR: |D/hc| ) 0.8 cm-1). The linear ED
curves indicate that the nitrene 2a is formed in one observable
reaction step and furthermore that 2a is photochemically and
(7) Harder, T.; Bendig, J.; Scholz, G.; Sto¨sser, R. J. Am. Chem. Soc.
1996, 118, 2497.

6582 J. Am. Chem. Soc., Vol. 121, No. 28, 1999
Harder et al.
Figure 3. UV/vis spectra of the “low light intensity photolysis” (λexc
) 313 nm) of 4,4′-diazidostilbene (1e, λmax ) 332, 348, and 366 nm)
at 115 K in MTHF and formation of the nitroso oxide 3e (λmax ) 537
nm).
On the basis of the subsequent photochemical and thermal
reactions of 3a (including product analysis)9 and considering
the spectroscopic results of Brinen et al.,4 it is concluded that
3a is the nitroso oxide. Our parallel ESR spectroscopic
investigations showed that no signal exists for 3a at 77 K. This
indicates that the observed subsequent reactions should start
from the singlet state of 3a.10
As mentioned above, the triplet nitrenes 2b-d were generated
in the same way and the nitroso oxides 3b-d were formed under
conditions similar to those of 3a (T ) 95 K, in MTHF). The
nitroso oxide 3e was formed in a different way (Scheme 3).
First, 4,4′-diazidostilbene (1e) was photolyzed using low light
intensities11 (λexc ) 313 nm) at 115 K in MTHF. Under these
conditions, the primarily formed triplet azidonitrene 2e reacts
immediately with dissolved triplet oxygen to form species 3e
(λmax ) 537 nm, Figure 3).
Figure 2. UV/vis spectroscopic investigations of 1a f 3a in MTHF:
(a) low-light intensity photolysis (λexc ) 313 nm) of 4-azidostilbene
(1a, λmax ) 319, 333, and 350 nm) and an ED diagram and (b) thermal
reaction of 4-nitrene-substituted stilbene (2a, λmax ) 369, 378, 614,
and 687 nm) in forming the nitroso oxide 3a (λmax ) 514 nm) and an
ED diagram.
Scheme 3
The UV/vis spectra of all nitroso oxides (3a-e) are shown
in Figures 4 and 5 (broken lines), the spectroscopic properties
of all involved species are summarized in Table 1.
After photolysis of the nitroso oxides 3a-e, different
intermediates are observed using stationary UV/vis spectroscopy.
The kind of observed intermediates strongly depends on light
intensity and temperature. If the nitroso oxides 3a-e are excited
monochromatically (λexc ) 550 nm for 3a and 3e, λexc ) 480
nm for 3b, λexc ) 675 nm for 3c, and λexc ) 625 nm for 3d) at
95 K using a low light intensity,11 then only the nitro products
5a-d and 5f are observed in the UV/vis spectrum. As an
example, this is shown in Figure 4a for the stepwise low-
intensity irradiation of 3b (λmax ) 483 nm for 3b, and λmax
)
357, 374, and 394 nm for 5b).13 While the ED curves of the
stepwise low-intensity photolyses of 3a-e are linear at 95 K,
the ED curves of the same reactions at 77 K are generally
nonlinear. If additionally strong light intensities11 are used (at
thermally stable at 77 K. This was also observed in the case of
2b-d.
The UV/vis spectra obtained after the thermal reaction of the
triplet nitrene 2a in air-saturated MTHF at 95 K are shown in
Figure 2b. From this, it follows that 2a reacts quantitatively to
form species 3a (λmax ) 514 nm). The 4-aminostilbene (λmax
) 351 nm) is not formed here at 95 K.8 The linear ED curves
in the inset of Figure 2b indicate that 3a is formed from 2a in
only one observable reaction step.
(9) Species 3a-d are not formed in anaerobic MTHF. After excitation
of 3a-d between 77 and 100 K, the corresponding nitro compounds are
the final products. The thermal reaction of 3a-d (T > 120 K) leads finally
to the formation of the nitroso compounds. The UV/vis spectra of nitroso
compounds are always red-shifted compared with those of the corresponding
nitro compounds [e.g., for 4-nitrostilbene, λmax ) 354 nm; for 4-nitrosos-
tilbene, λmax ) 386 nm (in methanol at 25 °C)].
(10) The species 3a-e have similar UV/vis spectroscopic properties.
Furthermore, they show the same consecutive photochemical and thermal
reactions. It is therefore assumed that at 77 K 3b-e also exist in their singlet
states. Their electronic states below 77 K were not investigated in this work.
(11) The experimental conditions for the “low light intensity” and for
the “strong light intensity” photolyses are described in the Experimental
Section.
(8) The triplet nitrenes that were used are photochemically stable. Since
MTHF is totally rigid at 77 K, significant bimolecular thermal reactions
can also be excluded. Between 77 and 100 K, two bimolecular thermal
reactions occur. These are the hydrogen abstraction from the matrix solvent
and the reaction of the triplet nitrene with molecular oxygen. The yields of
these reactions are significantly influenced by the matrix viscosity and
therefore by the temperature. Below 86 K, the triplet nitrenes prefer to react
under hydrogen abstraction to form the corresponding primary amines.
Above 92 K, the MTHF matrix is soft, which only leads to the formation
of the corresponding nitroso oxides 3a-d.5,7,16
(12) The zfs for 2f is similar to the value observed by: Minato, M.;
Lathi, P. M. J. Am. Chem. Soc. 1997, 119, 2187.

Photochemical Reactions of Nitroso Oxides
J. Am. Chem. Soc., Vol. 121, No. 28, 1999 6583
thus forming species 4f (λmax ) 355 and 371 nm, Figure 5a,
Scheme 3).15
It follows from Figures 4b and 5a-d that the intermediates
4a-d and 4f are highly reactive at 77 K in the dark (τ1/2 ≈ 4
min). As mentioned, their thermal reactions lead in all the cases
that were investigated to the formation of the corresponding
nitro compounds 5a-d and 5f. The ED curves of the thermal
reactions of 4a-d and 4f in forming 5a-d and 5f at 77 K are
strongly linear. This demonstrates that only one thermal process
is observed. This process is very fast compared with thermal
bimolecular processes in MTHF at 77 K.7,16 It is therefore
concluded that the reactions of 4a-d and 4f in forming 5a-d
and 5f are intramolecular processes. Considering the exclusive
formation of the nitro derivatives 5a-d and 5f and the inability
of oxygen and other species to diffuse at this temperature, it is
concluded from mechanistic arguments that the intermediates
4a-d and 4f are the corresponding dioxaziridines (Schemes 2
and 3). The proposed reactions of the nitroso oxides 3a-d and
3f are supported by the fact that similar 1,3-dipoles show similar
intramolecular reactions. The photochemical disrotatory ring
closure in forming three-membered heterocycles is allowed by
orbital symmetry rules and well-known for other 1,3-dipoles
such as carbonyl oxides or azomethine ylides. The conrotatory
ring opening reaction proceeds thermally. The product is another
isomeric 1,3-dipole.17
Figure 4. UV/vis spectroscopic investigations of 3b f 5b in MTHF:
(a) “low light intensity photolysis” (λexc ) 480 nm, irradiation time is
given in the picture) of the nitroso oxide 3b (λmax ) 483 nm) at 95 K
and formation of 4,4′-dinitrostilbene (5b, λmax ) 357, 374, and 394
nm) without observable intermediates; and (b) “strong light intensity
photolysis” (λexc > 440 nm) of the nitroso oxide 3b (λmax ) 486 nm)
at 77 K, formation of the dioxaziridine 4b (λmax ) 367 nm), and its
thermal reaction (the time is given in the picture) in forming
4,4′-dinitrostilbene (5b, λmax ) 357, 374, and 394 nm).
On the basis of the analogy of the reactions of 1,3-dipoles
such as carbonyl oxides, another possible candidate for 4 might
be the nitroso compound formed after excitation of the nitroso
oxide. The nitroso compound would then react ultimately with
oxygen to form the corresponding nitro compound. But this
reaction way can be excluded because of our experimental data
since it is well-known that UV/vis absorption bands of aromatic
nitroso compounds are always strongly red shifted compared
to those of the corresponding nitro compounds.9 On the basis
of the kinetic analyses of the reactions 3 f 4 f 5 (described
above) and on the basis of mechanistic arguments, the inter-
mediate 4 can also excluded to be ArNHOO• or a dehy-
droazepine. Furthermore, 4 cannot be a product with the solvent
(ArNH•) since its formation and the formation of the final nitro
product depend on the presence of oxygen.9
77 K), then the absorption bands of unstable intermediates
(4a-d and 4f) are observed. At 77 K, these intermediates (4a-d
and 4f) react thermally to form the corresponding nitro
compounds. As an example, this is demonstrated in Figure 4b
for the photochemically induced reaction of 3b (λmax ) 486
nm), which forms 4b (λmax ) 367 nm), and for the subsequent
thermal reaction of 4b, which forms 5b (λmax ) 374 and 394
nm). This means that a high stationary concentration of 4b is
obtained using strong light intensities at 77 K (Figure 4b) but
not if low light intensities are used at 95 K (Figure 4a). If 3b
is excited using strong light intensities at 95 K, then only a
small concentration of 4b is obtained. This indicates that the
primary photochemical reaction after excitation of 3b is the same
at both temperatures, but a sufficient stabilization of 4b is only
possible at 77 K.
The irradiation of 3a, 3c, 3d, and 3e leads qualitatively to
the same reactions; all important experimental results are shown
in Figure 5a-d (for the assignment of the intermediates,
compare to Table 1). The photolysis of 3e at 77 K (λexc ) 550
nm) leads additionally to the loss of nitrogen and to the
formation of 3f (λmax ) 554 nm).14 Species 3f cannot be
stabilized since it is also photolyzed under the given conditions,
When all these arguments are considered, it is almost certain
that the intermediates 4a-d and 4f are the dioxaziridines.
The fast thermal reactions of 4a-d and 4f at 77 K indicate
that the dioxaziridines are highly reactive species. The first-
order rate constants (k) of the ring opening reactions of 4a-d
and 4f were obtained by monitoring the absorbance (E) as a
function of time. The k values (at 77 K) were calculated from
the slopes of these plots. They are, within experimental error,
all equal to 0.0030 ((0.0005) s-1. This means that the reaction
is not influenced by mesomerically (4a) and inductively (4b)
electron-withdrawing or mesomerically electron-donating (4d)
substituents at the 4′-position of the stilbene or biphenyl.
Ab Initio Calculations. It was the aim of our ab initio
calculations to find a confirmation of our experimental findings
described above. We have chosen the nonsubstituted dioxaziri-
(13) The nitro products 5a-d were identified by their UV/vis spectra at
77 K by comparing them with spectra of authentic samples (at 77 K) as
well as by product analysis (HPLC using reference substances) at 300 K.
Species 5f (UV/vis: λmax ) 379, 396, 454 nm; ESR: Hx,y ) 620.7 mT,
|D/hc| ) 0.8 cm-1) was identified by comparing its UV/vis and ESR
spectra with those of 4-nitrene-substituted 4′-nitrostilbene which was
generated independently by the photolysis of 4-azido-4′-nitrostilbene (at
77 K).7
(15) The absorption bands at 541 and 573 nm are caused by the di-
(nitroso oxide) which is formed in a side reaction during the photolysis of
1e at 115 K in MTHF.5 This intermediate is photochemically stable and
can be formed either by a strong light intensity photolysis of 1e at 115 K
in MTHF or by the thermal reaction of 2f (generated before at 77 K) at T
> 95 K.
(16) Harder, T.; Sto¨sser, R.; Wessig, P.; Bendig, J. J. Photochem.
Photobiol., A 1997, 103, 105.
(14) Since the quantum yield of the nitrogen extrusion (3e f 3f) is much
higher than that of the electrocyclization (3e f 4e), the observation of the
azidodioxaziridine 4e is impossible.
(17) Scaiano, J. C. In Handbook of Organic Photochemistry; CRC
Press: Boca Raton, FL, 1989.

6584 J. Am. Chem. Soc., Vol. 121, No. 28, 1999
Harder et al.
Figure 5. UV/vis spectroscopic investigations of the “strong light intensity photolyses” of the nitroso oxides and of the thermal reactions of
dioxaziridines at 77 K (the time of the thermal reaction 4 f 5 is given in each figure): (a) photolysis (λexc ) 550 nm) of the nitroso oxide 3e (λmax
) 541 nm), formation of the dioxaziridine 4f (λmax ) 355 and 371 nm), and its thermal reaction in forming the 4-nitrene-substituted 4′-nitrostilbene
(5f, λmax ) 379, 396, 425, and 454 nm);15 (b) photolysis (λexc ) 550 nm) of the nitroso oxide 3a (λmax ) 514 nm, compare to Figure 2b), formation
of the dioxaziridine 4a (λmax ) 333 and 348 nm), and its thermal reaction in forming 4-nitrostilbene (5a, λmax ) 373 nm); (c) photolysis (λexc > 590
nm) of the nitroso oxide 3c (λmax ) 703 nm), formation of the dioxaziridine 4c (λmax ) 414 nm), and its thermal reaction in forming 4-(dimethylamino)-
4′-nitrostilbene (5c, λmax ) 477 nm); and (d) photolysis (λexc > 510 nm) of the nitroso oxide 3d (λmax ) 612 nm), formation of the dioxaziridine
4d (λmax ) 362 nm), and its thermal reaction in forming 4-amino-4′-nitrobiphenyl (5d, λmax ) 432 nm).
Table 1. UV/Vis and ESR Spectroscopic Properties of the Compounds Investigated at 77 K in MTHF
dine 6 and N-phenyldioxaziridine 9 as model systems for the
calculations. All calculations were performed with the program
GAUSSIAN 94.18
First, we have investigated the structures of the dioxaziridines
6 and 9 using the well-established hybrid DFT method
Becke3Lyp19 and the 6-31G* basis set.20 Figure 6 shows the

Photochemical Reactions of Nitroso Oxides
J. Am. Chem. Soc., Vol. 121, No. 28, 1999 6585
Figure 7. Energy of HNO2 upon variation of bond angle (R) O-N-O
(HF/6-31G*): (a) electronic state of 6 kept along the reaction path,
(b) electronic state of 8 kept along the reaction path, and (c) new guess
at every point of the reaction path.
Figure 6. Calculated structures of the dioxaziridines 6 and 9, the nitro
derivatives 8 and 11 formed by ring opening, and the corresponding
transition states 7 and 10.
6 f 8 should also demand a spin-unrestricted treatment. Figure
7 shows the results of reaction path calculations upon variation
of the bond angle O-N-O (R). It was surprising that starting
with the geometry of 6 increasing values of R provided a
continuous increase of energy (Figure 7, curve a). The same
result was obtained in the reverse direction starting with the
geometry of 8 (curve b). Obviously, a far-reaching change of
the MO scheme takes place along the reaction path. The
steady energy (curve c) was first obtained by a new guess of
the UHF wave function at every point and second by mixing
the HOMO and LUMO in the wave function to destroy R-â
and spatial symmetries (this is often necessary if a stable UHF
wave function of a singlet state with diradical character is
produced).
Since the electron correlation is neglected in this method and
due to the well-known tendency of UHF wave functions to be
spin-contaminated, the Hartree-Fock values shown in Figure
7 should not be overestimated. Nevertheless, these results
encouraged us to investigate the stereoelectronic course of the
reaction by means of Hartree-Fock theory. Figure 8 shows the
orbital correlation diagram obtained on the HF/6-31G* level.
The five highest occupied molecular orbitals and the lowest
unoccupied molecular orbital of both 6 (left) and 8 (right) are
shown. It is easy to recognize that the LUMO of 6, an
antibonding Walsh orbital, corresponds with HOMO-2 of 8, a
nonbonding n orbital. On the other hand, the HOMO of 6, an
n orbital with a week O-O π interaction, corresponds with the
LUMO of 8, an antibonding π* orbital.
optimized structures and the most important geometrical pa-
rameters of the calculated species. One can see that the structures
of both dioxaziridines are very similar concerning the three-
membered ring. The nitrogen atom is strongly pyramidalized
even though the reason is not a sp3 hybridization but a
cyclopropane-like electronic structure of the ring. This is
confirmed by the structure of N-phenyldioxaziridine 9. The
phenyl substituent prefers a bisected conformation, which
points to an overlap of π orbitals of the aromatic ring and a
Walsh orbital of the nitrogen atom, a phenomenon well-known
from substituted cyclopropanes.21 It should be noted that the
perpendicular conformation of 9 was found to be a transition
state. The rotation barrier amounts to 3.5 kcal/mol (B3Lyp/6-
31G*).
Second, we were interested in the transition state of the ring
opening in forming isonitrous acid 8 and nitrobenzene 11. Kahn
et al.22 recently concluded that the isonitrous acid 8 has a
remarkably diradical character. This was concluded by the fact
that a spin-unrestricted Hartree-Fock (UHF) wave function
provides a lower energy than a spin-restricted Hartree-Fock
(RHF) treatment. Therefore, an investigation of the reaction path
(18) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M.; Johnson,
B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.; Petersson, G. A.;
Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski, V.
G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.;
Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.;
Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.;
Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. P.; Head-
Gordon, M.; Gonzalez, C.; Pople, J. A. Gaussian 94; Gaussian, Inc.:
Pittsburgh, PA, 1995.
(19) Parr, R. G.; Yang, W. Density-Functional Theory of Atom and
Molecules; Oxford University Press: New York, 1989.
(20) Hariharan, P. C.; Pople, J. A. Chem. Phys. Lett. 1972, 66, 217.
(21) Hehre, W. J.; Rado, L.; Schleyer, P. v.; Pople, J. A. Ab initio
Molecular Orbital Theory; John-Wiley & Sons: New York, 1986.
(22) Kahn, S. D.; Hehre, W. J.; Pople, J. A. J. Am. Chem. Soc. 1987,
109, 1871.
One can therefore deduce from the orbital correlation diagram
which reactant and product configurations correlate with each
other. The diagram shown in Figure 8 suggests that correlation
lines of equal symmetry cross in a configuration diagram. This
is a violation of the noncrossing rule, and therefore, 6 f 8 must
be considered as a symmetrical ground state-forbidden concerted
reaction path.
Nevertheless, the reaction was found to be strongly exother-
mic, which means that a little activation barrier is expected. To
obtain more accurate activation barriers, the ring openings of 6

6586 J. Am. Chem. Soc., Vol. 121, No. 28, 1999
Harder et al.
Figure 8. Orbital correlation diagram for the reaction path 6 f 8 (HF/6-31G*).
Table 2. Calculated Energies of 6-11
saturated MTHF at 95 K. At temperatures below 100 K,
aromatic nitroso oxides are thermally stable. The investigated
nitroso oxides 3a-e exist at 77 K in their singlet states. Hitherto
assumed but not found, further intermediates were directly
observed by UV/vis spectroscopy after optical excitation of
nitroso oxides at 77 K. On the basis of their UV/vis spectra
and chemical, mechanistic, and kinetic arguments, these inter-
mediates are assumed to be the corresponding dioxaziridines,
but more analytical results (IR spectra) are necessary to prove
this statement. The assumed dioxaziridines 4a-d and 4f are
highly reactive species. They react thermally (k77K ) 0.003
s-1) to form the corresponding nitro compounds (5a-d and
5f). The velocity of the ring opening reaction is mainly
determined by the strained dioxaziridine structure but not by
substituents.
(UB3Lyp/6-31G*//UB3Lyp/6-31G*)
entry
E (au)
ZPVE (kcal/mol)
∆Ea (kcal/mol)
6
7
8
9
10
11
-205.5670
-205.5411
-205.6869
-436.6261
-436.5867
-436.7506
12.6
10.5
13.8
63.2
61.6
65.0
0
14.2
-74.0
0
18.8
-80.7
a Relative energies including zero-point energy, ∆E ) E - E(6)
and ∆E ) E - E(9), respectively.
and 9 were reinvestigated using a wave function which includes
electron correlation (B3Lyp/6-31G*). The resulting structures
of the transition states 7 and 10 are pictured in Figure 6. We
have found energy differences of 14.2 kcal/mol (6 f 7) and
18.8 kcal/mol (9 f 10). All thermodynamic parameters of the
model compounds 6-11 are shown in Table 2.
It follows from ab initio calculations that dioxaziridines are
experimentally observable species. They are separated from the
nitro products by an orbital symmetry-forbidden barrier.
Unfortunately, it is not possible to compare our calculated
energy barriers with the experimental data because different
molecules were investigated experimentally and theoretically.
As mentioned in the Introduction, this was inevitable because
of two reasons. On one hand, large molecules are the precondi-
tion for an experimental success (photochemically stable triplet
nitrenes must be generated and different species must be able
to be distinguished by UV/vis spectroscopy), but on the other
hand, only small molecules can be treated quantum chemically
at a high level of theory. Furthermore, it is generally impossible
to estimate a reliable energy barrier on the basis of only one
rate constant determined experimentally at one temperature
(especially if the reaction was carried out in a low temperature
matrix). Energy barriers can therefore not be compared quan-
titatively. But it can be concluded from the theoretical results
that dioxaziridines are separated from their appropriate nitro
products by an orbital symmetry-forbidden barrier and that they
are experimentally observable species.
Experimental Section
General. The starting reagents were used as purchased from Aldrich
Chemical Co. MTHF (2-methyltetrahydrofuran) was distilled from
calcium hydride. Optical absorption measurements were carried out
with a Hitachi U-3410 double-beam spectrophotometer using a 1 mm
quartz cell. The temperature could be adjusted and kept constant
between 77 and 300 K with a precision of 0.1 K using a cryostat
equipped with a temperature controller. The ESR spectra were recorded
on a ERS 300 spectrometer (ZWG Berlin, X-band microwave unit,
100 kHz field modulation, f ) 9.18 GHz). The samples were cooled
within the cavity using a liquid nitrogen Dewar insert. The product
analysis was realized by HPLC investigations (Kontron, DAD 440)
comparing both the retention time and the UV/vis spectra of the
products with those of reference substances.
Irradiation. For both the UV/vis and ESR experiments, the light
sensitive compounds were irradiated through a water cell (10 cm), metal
interference filters, and/or glass filters using a high-pressure mercury
lamp (500 W) at a distance of 20 cm. The samples were irradiated
within the sample chambers of the spectrometers.
Conclusions
Aromatic nitroso oxides can be synthesized quantitatively by
a thermal reaction of the corresponding triplet nitrenes in air-
The so-called “low light intensity photolysis” was realized by the
combination of both a metal interference filter (for λexc see the text)

Photochemical Reactions of Nitroso Oxides
J. Am. Chem. Soc., Vol. 121, No. 28, 1999 6587
and glass filters which decreased the light intensity additionally. The
light intensity was adjusted according to the following criteria. In the
starting period of the photolysis, 5 s of irradiation must decrease E by
about 0.05 [E0 (λmax) ≈ 1]. It was irradiated in intervals of about 5 s,
and after every irradiation, the sample was in the dark for at least 3
min. The spectra were recorded during this time. This kind of irradiation
was applied to the photolysis of all nitroso oxides and azides [λexc(1a,
1d, 1e) ) 313 nm, λexc(1b, 1c) ) 365 nm]. UV/vis spectra obtained
by this procedure are shown in Figure 2a (photolysis of 1a, formation
of 2a), in Figure 3 (photolysis of 1e, formation of 3e), and in Figure
4a (photolysis of 3b, formation of 5b).
The so-called “strong light intensity photolysis” was realized by using
either only metal interference filters or only glass filters. The light
intensity was adjusted here according to the following criteria. The
light sensitive intermediate must be completely decomposed after a
single irradiation process which must not exceed 30 s. After this
single photolysis, the UV/vis spectra were quickly recorded using the
highest speed of the spectrophotometer (1200 nm/min). Species 3a and
3e (λexc ) 550 nm) were irradiated using only a metal interference
filter, and species 3b (λexc > 440 nm), 3c (λexc > 590 nm), and 3d
(λexc > 510 nm) were irradiated using only a glass filter. Reference
experiments showed that qualitatively the same reactions occur when
the intermediates 3b-d are irradiated using metal interference filters.
UV/vis spectra obtained by this procedure are shown in Figure 4b
(photolysis of 3b, formation of 4b) and Figure 5 (photolysis of 3a, 3c,
3d, and 3e).
trans-4-Nitrostilbene. Benzaldehyde (9.2 g), 1.5 mL of freshly
distilled piperidine, and 16 g of (nitrophenyl)acetic acid were combined
and refluxed for 2 h in an oil bath at 140 °C. After cooling to room
temperature, the yellow residue was treated with ethanol. It was filtered
and recrystallized from ethanol (85% yield). Anal. Calcd: C, 74.65;
H, 4.9; N, 6.2. Found: C, 74.4; H, 4.7; N, 6.0. UV/vis: λmax ) 354
nm (methanol). Mp: 155 °C.
trans-4-Aminostilbene. trans-4′-Nitrostilbene (3.5 g) was suspended
in 100 mL of 80% ethanol. Then a solution of 35 g of FeSO4‚7H2O in
175 mL of water treated with 175 mL of concentrated NH3 was added.
Subsequently, the reaction mixture was heated on a steam bath for 2 h
and cooled to room temperature. After 12 h, the mixture was filtered
and the black residue was extracted several times using diethyl ether.
After evaporation of the solvent, the product was recrystallized
from ethanol to provide a yield of 90%. Anal. Calcd: C, 86.1; H, 6.7;
N, 7.1. Found: C, 85.7; H, 6.6; N, 6.7. UV/vis: λmax ) 324 nm
(methanol).
solution had turned dark red, it was evaporated to dryness. The residue
was extracted with acetone and recrystallized from acetone. The crystals
were red (52% yield). Anal. Calcd: C, 70.0; H, 5.0; N, 11.7; O, 13.3.
Found: C, 69.3; H, 5.0; N, 11.5. UV/Vis: λmax ) 287, 403 nm
(methanol). Mp: 201 °C.
trans-4-Azido-4′-nitrostilbene (1b). Concentrated HCl (20 mL) was
added to 1.56 g of trans-4-amino-4′-nitrostilbene. The mixture was
cooled to -10 °C and treated dropwise (1.5 h) with a solution of 0.46
g of NaNO2 in 5 mL of water and then stirred for another 6 h at -10
°C. Then the reaction mixture was filtered at 0 °C. A solution of 0.46
g of NaN3 in 10 mL of water was added, and the solution was stirred
for another 2 h at 0 °C. The solid was filtered, rinsed with water and
then with diethyl ether, and dried (70% yield). Anal. Calcd: C, 63.2;
H, 3.8; N, 21.0. Found: C, 62.2; H, 3.8; N, 20.6. UV/Vis: λmax
)
283, 364 nm (methanol). IR (KBr): 2114.4 cm-1 (str NdNdN).
trans-4-(Dimethylamino)-4′-nitrostilbene. 4-(Dimethylamino)ben-
zaldehyde (13.2 g), 1.5 mL of freshly distilled piperidine, and 16 g of
(nitrophenyl)acetic acid were combined and refluxed for 2 h in an oil
bath at 140 °C. After cooling to room temperature, the red solid was
treated with ethanol, filtered, and recrystallized from ethanol (85%
yield). Anal. Calcd: C, 71.6; H, 6.0; N, 10.4; O, 11.9. Found: C, 71.7;
H, 6.1; N, 10.4. UV/vis: λmax ) 297, 425 nm (methanol). Mp: 257
°C.
trans-4-Amino-4′-(dimethylamino)stilbene. trans-4-(Dimethylamino)-
4′-nitrostilbene (4.2 g) was suspended in 100 mL of 80% ethanol. Then
a solution of 35 g of FeSO4‚7H2O in 175 mL of water treated with
175 mL of concentrated NH3 was added. Subsequently, the reaction
mixture was heated on a steam bath for 2 h and then cooled to room
temperature. After 12 h, the mixture was filtered and the black residue
was extracted several times using diethyl ether to isolate the product.
After evaporation of the solvent, the product was recrystallized from
ethanol to provide a yield of 68%. Anal. Calcd: C, 80.6; H, 7.6; N,
11.8. Found: C, 76.9; H, 8.0; N, 11.2. UV/vis: λmax ) 353 nm
(methanol). Mp: 179 °C.
trans-4-Azido-4′-(dimethylamino)stilbene (1c). Concentrated HCl
(20 mL) was added to 1.5 g of trans-4-amino-4′-(dimethylamino)-
stilbene. The mixture was cooled in the dark to -10 °C, treated
dropwise (1.5 h) with a solution of 0.48 g of NaNO2 in 1 mL of water,
and then stirred for another 5 h at -10 °C. Then the reaction mixture
was filtered in the dark at 0 °C. A solution of 0.45 g of NaN3 in 10
mL of water was added, and the solution was stirred for another 2 h at
0 °C (dark). The solid was filtered, rinsed with water, and dried (10%
yield). Anal. Calcd: C, 72.7; H, 6.1; N, 21.2. Found: C, 71.9; H, 5.9;
N, 20.8. UV/vis: λmax ) 358 nm (methanol). IR (KBr): 2113.7 cm-1
(str NdNdN).
4,4′-Diazidobiphenyl. Concentrated HCl (20 mL) and 20 mL of
water were added to 1.75 g of 4,4′-diaminobiphenyl. The mixture was
cooled to -10 °C, treated dropwise (0.5 h) with a solution of 0.67 g of
NaNO2 in 5 mL of water, and then stirred for another 1.5 h at -10 °C.
The reaction mixture was filtered at 0 °C. A solution of 0.7 g of NaN3
in 10 mL of water was added (0.5 h). The solution was stirred for
another 1.5 h at 0 °C. The solid was filtered, rinsed with water, and
dried (80% yield). Anal. Calcd: C, 61.0; H, 3.4; N, 36.6. Found: C,
61.2; H, 3.5; N, 36.5. UV/Vis: λmax ) 294 nm (methanol). IR (KBr):
2134.0 cm-1 (str NdNdN). Mp: 126 °C.
trans-4-Azidostilbene (1a). Concentrated HCl (20 mL) and 20 mL
of water were added to 3.7 g of trans-4-aminostilbene. The mixture
was cooled to -10 °C and treated dropwise (0.5 h) with a solution
of 0.67 g of NaNO2 in 5 mL of water, and then stirred for another
1.5 h at -10 °C. The reaction mixture was filtered at 0 °C. A
solution of 0.7 g of NaN3 in 10 mL of water was added (0.5 h). The
solution was stirred for another 1.5 h at 0 °C. The solid was filtered,
rinsed with water, and dried (85% yield). Anal. Calcd: C, 75.99; H,
5.01; N, 18.99. Found: C, 75.03; H, 5.06; N, 18.71. UV/Vis: λmax
)
324 nm (methanol). IR (KBr): 2121.3 cm-1 (str NdNdN). Mp: 104
°C.
trans-4,4′-Dinitrostilbene.23 4-Nitrobenzyl chloride (10 g) was
dissolved in 30 mL of heated ethanol. Then 3.5 g of KOH, dissolved
in a mixture of 3 mL of water and 12 mL of ethanol, was added
dropwise at 25 °C. The temperature of the reaction mixture was
raised to about 70 °C. It was stirred for 2 h at 70 °C, cooled, and
filtered. The residue was washed with ethanol. The crystals were
yellow needles (80% yield). Anal. Calcd: C, 62.2; H, 3.7; N, 10.4.
Found: C, 62.1; H, 3.7; N, 10.1. UV/Vis: λmax ) 353 nm (methanol).
Mp: 282 °C.
trans-4-Amino-4′-nitrostilbene.24 trans-4,4′-Dinitrostilbene (5 g)
was suspended in 200 mL of 95% ethanol. The mixture was stirred
and heated to boiling. Then 13 mL of a 1.25 M solution of sodium
sulfide, treated with hydrogen sulfide for 25 min, was added. After the
4-Amino-4′-azidobiphenyl (1d). 4,4′-Diazidobiphenyl (0.2 g) was
suspended in 20 mL of 85% ethanol. The mixture was stirred and heated
to boiling. Then 0.4 mL of a 1.25 M solution of sodium sulfide, treated
for 25 min with hydrogen sulfide before, was added. After the solution
had turned green, it was stirred at 70 °C for another 30 min. Then it
was evaporated at room temperature. Besides the azidoamine, the
residue contains the diamine, the diazide, and inorganic compounds. It
was extracted and recrystallized from ethanol several times. After these
procedures, the purity of 1d was only about 90%. The final product
was therefore purified by HPLC (11% yield). Anal. Calcd: C, 68.6;
H, 4.8; N, 26.6. Found: C, 67.4; H, 4.1; N, 26.1. UV/Vis: λmax
298, 401 nm (methanol).
)
trans-4,4′-Diazidostilbene (1e). Concentrated HCl (20 mL) and 20
mL of water were added to 2 g of trans-4,4′-diaminostilbene. The
mixture was cooled to -10 °C, treated dropwise (0.5 h) with a solution
of 0.67 g of NaNO2 in 5 mL of water, and then stirred for another 1.5
(23) Walden, P.; Kernbaum, A. Ber. 1890, 23, 1959. Oki, M.; Kunimoto,
H. Spectrochim. Acta 1963, 19, 1463.
(24) Calvin, M.; Buckles J. Am. Chem. Soc. 1940, 62, 3324.

6588 J. Am. Chem. Soc., Vol. 121, No. 28, 1999
Harder et al.
h at -10 °C. The reaction mixture was filtered at 0 °C. A solution of
0.7 g of NaN3 in 10 mL of water was added (0.5 h). The solution was
stirred for another 1.5 h at 0 °C. The solid was filtered, rinsed with
water, and dried (75% yield). Anal. Calcd: C, 64.1; H, 3.8; N, 32.0.
Found: C, 63.7; H, 3.9; N, 31.0. UV/Vis: λmax ) 340 nm (methanol).
IR (KBr): 2120.1 cm-1 (str NdNdN). Mp: 162 °C.
Acknowledgment. We thank the micro resist technology
GmbH (Berlin), the Deutsche Forschungsgemeinschaft (Bonn),
and the Fond der Chemischen Industrie (Frankfurt/Main) for
financial support.
JA9803660