Flash vacuum thermolysis of 1-azabicyclo[1.1.0]butanes. Photoelectron spectrum of 3-phenyl-2-azabuta-1,3-diene

Article abstract of DOI:10.1002/ejoc.200300060

The reaction behaviour of 1-azabicyclo[1.1.0]butanes under flash vacuum thermolysis (FVT) conditions was studied. It has been found that the thermal rearrangement of title compounds produces unstable 2-aza-1,3-dienes. The formation of these products was established by reduction, by Diels-Alder reactions and by UV photoelectron spectroscopy. Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim, Germany, 2003.

Full text of DOI:10.1002/ejoc.200300060

FULL PAPER
Flash Vacuum Thermolysis of 1-Azabicyclo[1.1.0]butanes.
Photoelectron Spectrum of 3-Phenyl-2-azabuta-1,3-diene[‡]
[a]
[b]
[b]
[c]
`
Romuald Bartnik, Patrick Baylere, Anna Chrostowska, Alberto Galindo,
[a]
[b]
´
`
Stanislaw Lesniak,* and Genevieve Pfister-Guillouzo*
Keywords: Flash pyrolysis / Ionisation potentials / Photoelectron spectroscopy / Quantum mechanical calculations /
Small ring systems
The reaction behaviour of 1-azabicyclo[1.1.0]butanes under
flash vacuum thermolysis (FVT) conditions was studied. It
these products was established by reduction, by Diels−Alder
reactions and by UV photoelectron spectroscopy.
has been found that the thermal rearrangement of title com- ( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
pounds produces unstable 2-aza-1,3-dienes. The formation of
Germany, 2003)
Introduction
Results and Discussion
Thermolysis of 3-Phenyl-1-azabicyclo[1.1.0]butane
Substituted 1-azabicyclo[1.1.0]butanes were first synthe-
sized in the late 1960s.[2Ϫ4] Until recently, however, rela-
tively little interest in synthetic applications of this highly
strained and unusual ring system had been shown. Particu-
larly important synthetic applications of 1-azabicyclo-
[1.1.0]butanes include additions of reagents of the XϪY
type across the highly strained NϪC(3) σ-bond, a variety
of such reagents having thus been treated with 3- and 2,3-
disubstituted 1-azabicyclo[1.1.0]butanes to produce corre-
sponding substituted azetidines.[5,6] The aim of our study
was to determine the structures of products formed from 1-
azabicylo[1.1.0]butane under FVT conditions, by chemical
trapping and by photoelectron spectroscopy. In analogy
with the thermal rearrangement of bicyclo[1.1.0]butane to
buta-1,3-diene,[7,8] the corresponding 2-aza-1,3-dienes
would be expected. The formation of similar 2-aza-1,3-
dienes was previously observed[9] during thermolysis of
1-azetines (structural isomers of 1-azabicyclo[1.1.0]butane).
Study of the thermal rearrangement of 1-azabicyclo-
[1.1.0]butanes was at first performed for 3-phenyl-1-aza-
bicyclo[1.1.0]butane (1a). Compound 1a was thermolysed
in the gaseous phase at 1.5·10Ϫ3 Torr at 400 °C. (The
chosen temperature was the lowest at which the starting
material was found not to be present in the reaction mixture
after thermolysis.) The product of the thermolysis of 1a was
collected in the cooling trap at Ϫ78 °C, then rinsed off with
cold CCl4 and examined by IR and 1H NMR spectroscopy.
1
This procedure enabled us to record the IR and H NMR
spectra of the product, which was identified as 3-phenyl-2-
azabuta-1,3-diene (2). Warming of the pyrolysate to room
temperature each time provoked the polymerisation of
products. It is well known that 2-aza-1,3-dienes are too un-
stable to be isolated and that oligomerisation takes place in
solution around 0 °C.[9,10] In addition to the IR and H
1
NMR spectroscopic data, the formation of 2-azadiene 2
from 1-azabicyclo[1.1.0]butane was also verified chemically.
The thermolysis was repeated and followed by chemical
trapping of the product at low temperature, the product of
[‡]
Application of Photoelectron Spectroscopy to Molecular the thermolysis of 1a being rinsed from cold finger with
Properties, 65. Part 64: Ref.[1]
cold THF into a suspension of LiAlH4 in THF at Ϫ20 °C.
[a]
[b]
[c]
Department of Organic and Applied Chemistry, University of
After workup the acetophenone N-methylimine (3) was ob-
tained as the only product. Similarly, trapping of 2 through
a DielsϪAlder reaction with the electron-rich N-(1-ethoxy-
vinyl)-N,N-dimethylamine produced the [4ϩ2] cycloadduct
in the reaction mixture. This was aromatised to the
4-(dimethylamino)-2-phenylpyridine (4) with chloranil
(Scheme 1).
´ ´
Łodz,
´ ´
Narutowicza 68, 90Ϫ136 Łodz, Poland
E-mail: slesniak@chemul.uni.lodz.pl
´
Laboratoire de Physico-Chimie Moleculaire, CNRS UMR
´
5624, Universite de Pau et des Pays de l’Adour,
´
Avenue de l’Universite, 64000 Pau, France
E-mail: genevieve.pfister@univ-pau.fr
Centro de Quimica, Instituto de Ciencias, BUAP,
Puebla 72570, Mexico
Eur. J. Org. Chem. 2003, 2475Ϫ2479
DOI: 10.1002/ejoc.200300060
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2475

´
S. Lesniak, G. Pfister-Guillouzo et al.
FULL PAPER
rise to the spectrum with two new peaks at 8.2Ϫ8.4 and
9.2Ϫ9.5 eV (Figure 1, b). We observe the disappearance of
the intense band belonging to the starting compound at
9.9 eV. The product spectrum has a set of three broad bands
around 8.2, 9.2, and 10.5 eV, with marked shoulders at 8.4,
8.6, and 9.5 eV.
The photoelectron spectrum of 2-azabuta-1,3-diene, pre-
viously described by Bock,[11] has the first main band at
9.61 eV with a well resolved vibrational structure (1600
cmϪ1 νCϭC stretch), corresponding to the ejection of an
electron from the πCϭC bond and the second one at
10.30 eV associated with the ejection of an electron from
the nitrogen lone pair.
Scheme 1
In our case, the first broad band could be attributed to
the vibrational structure, but for the analysis of the other
bands, the co-existence of different conformers has to be
taken into account. We consequently decided to perform
studies on conformational and electronic structures of 3-
phenyl-1-azabicyclo[1.1.0]butane (1a) and 3-phenyl-2-aza-
buta-1,3-diene (2).
The above reactions showed that the thermolysis of 3-
phenyl-1-azabicyclo[1.1.0]butane under FVT conditions re-
sulted in the formation of 3-phenyl-2-azabuta-1,3-diene (2)
as a primary product. Thus, the cleavage of the azabicyclo-
butane ring proceeds with breaking of two opposite bonds.
The 2-azadiene formed is too unstable to be isolated and
fully analysed.
Independently of the chemical trapping reactions, the
thermolysis of 3-phenyl-1-azabicyclo[1.1.0]butane (1a) was
performed in a FVT oven directly coupled to the ultra-
violet photoelectron spectrometer (UV-PES).[1] We first per-
formed studies on conformational favoured structure of 3-
phenyl-2-azabuta-1,3-diene (2), and then compared the cal-
culated values obtained by Density Functional Theory with
the experimentally determined PES ionization potential
data.
Theoretical Results and Interpretation of UV Photoelectron
Spectra
3-Phenyl-1-azabicyclo[1.1.0]butane (1a)
Two minima for the precursor 1 were confirmed by fre-
quency calculations. The favoured conformer A (total en-
ergy ET ϭ Ϫ403.1389 Hartrees) corresponds to the struc-
ture with the phenyl ring in the bisector plane of the bicycle.
In this case, there is no interaction between the π-system
and the nitrogen atom lone pair. The full interaction of
these two systems is present in the second conformer B
(total energy ET ϭ Ϫ403.1380 Hartrees), which is less stable
than A by 0.57 kcal·molϪ1. The first ionic state of A, calcu-
lated at 9.1 eV, corresponds to the b1 and a2 ionizations of
the phenyl ring. The following ionizations are associated
with the σC-C orbitals of 1-azabicyclo[1.1.0]butane and the
nitrogen lone pair. Taking the perpendicular phenyl ring
orientation (without any interaction) into account, these ex-
perimental bands should be clearly separated. Considering
Photoelectron Spectra
The photoelectron spectrum of 3-phenyl-1-azabicyclo-
[1.1.0]butane (1a, Figure 1, a) displays four broad bands: at
9.0, 9.4, 9.9, and 10.65 eV. The precursor begins to decom-
pose at 500 °C and disappears completely at 630 °C, giving
Table 1. Experimentally determined and estimated values of
ionization potentials, KohnϪShan (K-S) energies (in eV) and the
molecular orbital natures for two conformers A and B of 3-phenyl-
1-azabicyclo[1.1.0]butane (1a)
A
MO’s
B
IP
∆SCF ϭ 9.10 MO’s
∆SCF ϭ 8.35 exp.
nature K-S Estimated nature K-S Estimated
b1-nN
Ϫ6.90 9.0Ϫ9.2 a2
Ϫ7.07
Ϫ6.49 8.6
Ϫ7.11 9.2
8.7
9.0
b1
a2
nN-σC-C Ϫ7.32 9.4
nN-σC-H Ϫ7.83 9.9
nN-σC-C Ϫ7.44 9.65
9.4
9.9
b1ϩnN Ϫ8.37 10.6
10.65
11.4Ϫ11.7
b1, σC-C Ϫ9.40 11.5
b2
Ϫ9.23 11.4
b2
Ϫ9.57 11.7
Figure 1. Photoelectron spectra of 3-phenyl-1-azabicyclo[1.1.0]-
butane, a) at room temp. b) at 630 °C
2476
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2003, 2475Ϫ2479

Flash Vacuum Thermolysis of 1-Azabicyclo[1.1.0]butanes
FULL PAPER
KohnϪShan energies corrected by about 2.1 eV (Table 1) culated total energy differences are particularly small, so
(see Exp. Sect. for calculation of the x shift), the first esti- under our experimental conditions (T ϭ 630 °C) we should
mated ionization potential for A should be at 9.0Ϫ9.2 eV, observe an equilibrium between these two forms and not
the second one at 9.4 eV and the third one at about 9.9 eV. only the one twisted conformer. This hypothesis may there-
The more energetic ionizations have to be above 11.5 eV.
fore explain the widening of the first three bands. These
In the case of the B conformer, the positions of the first arguments, though, are based only on the total energy
and the fourth bands are modified by the nitrogen lone pair analysis. In order to corroborate these arguments, the first
and phenyl ring interactions, so the first ionic state is calcu- ionic states were calculated for the four conformers. They
lated to be at 8.35 eV and the fourth ionization is estimated are presented in Table 2, with the KohnϪShan energies and
at about 10.6 eV.
the estimated ionization potentials (correcting factor x of
The interpretation of the photoelectron spectrum of 1 is about 2 eV).
based on free rotation of the phenyl ring because of the
presence of the ionization potentials expected for A and
also for B (8.7 eV shoulder and the large band at 10.65 eV).
For the T and TSC we expect the first band, correspond-
ing to the πCϭC orbital interacting with the phenyl ring (b1
orbital), at 8.2Ϫ8.6 eV. The second one, at 9Ϫ9.2 eV, is as-
sociated with two ionizations: that of the nN nitrogen lone
pair in interaction with phenyl group and that of the a2
orbital. The third band, at 9.5 eV, is due to the ejection of
an electron from the πCϭN Ϫ πCϭC orbital in interaction
with the phenyl ring (b1). The σ-ionizations would appear
beyond 11.3 eV.
For two twisted conformers we would expect the first ion-
ization potential at the lower energies (ca. 8Ϫ8.2 eV) corre-
sponding to the ejection of an electron from the πCϭC or-
bital interacting with the nitrogen lone pair. The second and
third ones, similarly, should be closer (between 9.0 and
9.2 eV) to one another and would be associated with the
π-aromatic ring (a2) and the nitrogen lone pair. The main
difference should be seen in the fourth ionization potential,
due to the πCϭN Ϫ πCϭC ϩ nN orbital, at about 10.3 eV.
The observed photoelectron spectrum corresponds well
to the twisted synclinal and antiperiplanar conformer equi-
librium of 3-phenyl-2-azabuta-1,3-diene.
3-Phenyl-2-azabuta-1,3-diene (2)
Figure 2 displays the stationary points of the potential
energy surfaces obtained by varying the CϭNϪCϭC and
phenylϪCϭC dihedral angles in 3-phenyl-2-azabuta-1,3-di-
ene 2. The C twisted synclinal (ϩsc) conformer, in which
the dihedral angle between CϭC and CϭN is 55.6° and
the phenyl ring is rotated by Ϫ30.45°, is favoured. The T
antiperiplanar (ap) conformer (the dihedral angle between
CϭC and CϭN of Ϫ168.0°) with the phenyl ring rotated
by 61.3° is less stable than C by only 2.19 kcal·molϪ1
.
The first intense band (at 8.2Ϫ8.4 eV) and that at
10.4Ϫ10.6 eV originate from favoured twisted C and TST
conformers. The existence of planar T and TSC conformers
is confirmed through 8.6 and 9.5 eV ionization fingerprints.
The calculated energy values for these two conformers,
and the inter-conversion barriers and the good correlation
between the observed and calculated ionization potentials,
allowed us to infer, without ambiguity, the formation of 3-
phenyl-2-azabuta-1,3-diene from 3-phenyl-1-azabicyclo-
[1.1.0]butane.
Figure 2. Calculated stationary points of the potential energy sur-
faces (ET Ϫ total energies in Hartrees) for 3-phenyl-2-azabuta-1,3-
diene (2). * values in kcal·molϪ1
The transition state TSC (synperiplanar sp) lies at 4.49
kcal·molϪ1 above C and has the planar structure (phenyl
ring twisted 19.5°). TST is the highest transition state (5.97
kcal·molϪ1 above the C), with anticlinal conformation
(ϩac) and coplanar to the CϭC bond phenyl ring. The cal-
Table 2. Experimentally determined and estimated values of ionization potentials, KohnϪShan (K-S) energies (in eV) and the molecular
orbital natures for four stationary points of 3-phenyl-2-azabuta-1,3-diene (2) and corresponding total energies (ET in Hartrees).
TWISTED CONFORMERS
PLANAR CONFORMERS
IP
exp.
MO’s nature
C
TST
MO’s nature
T
TSC
ET ϭ Ϫ403.1853
∆SCF ϭ 8.06
ET ϭ Ϫ403.1758
∆SCF ϭ 7.88
ET ϭ Ϫ403.1818
∆SCF ϭ 8.47
ET ϭ Ϫ403.1782
∆SCF ϭ 8.09
K-S
Estim.
K-S
Estim.
K-S
Estim.
K-S
Estim.
πCϭC-nσ N-b1
a2
Ϫ6.23
Ϫ6.98
Ϫ7.18
8.2
Ϫ6.05
Ϫ7.13
8.0
πCϭC-b1
πCϭC-πCϭN- b1
nN-(ϕ)
Ϫ6.27
8.2
8.2
8.6
9.0Ϫ9.2
Ϫ6.61
Ϫ6.97
Ϫ7.19
8.6
9.0
9.2
9.55
9.0
9.1
Ϫ7.20
Ϫ6.89
Ϫ7.72
9.1
8.8
9.65
a2
nNϩb1
πCϭC-πCϭNϩnN Ϫ8.24
9.2
10.3
Ϫ7.25
Ϫ8.26
9.3
10.3
πCϭC-πCϭNϩ b1 Ϫ7.54
9.5
10.5
11.3
σb2 Ϫ9.58
11.59
Ϫ9.35
11.3
Eur. J. Org. Chem. 2003, 2475Ϫ2479
www.eurjoc.org
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2477

´
S. Lesniak, G. Pfister-Guillouzo et al.
FULL PAPER
Thermolysis of 1b؊d Precursors with Chemical Trapping
Experiments
is different and more complicated, because of the presence
of the nitrogen atom. It is possible that the intermediate
carbene (or nitrene) is not involved in the reaction. In our
opinion, thermal opening of the azbicyclobutane ring pro-
ceeds as a biradical stepwise process. It seems possible that
the cleavage of the NϪC(2) bond produces the more stable
biradical which, in a further step, is rearranged to 2-aza-
1,3-diene through cleavage of the C(3)ϪC(4) bond. The in-
verse order of bond cleavage is less probable. The cleavage
of the C(2)ϪC(3) bond affords a diradical, the stability of
which changes with variation of substituents at C(2). If we
presume that the C(2)ϪC(3) bond reacted at the first step
we can explain the changes of the regiochemistry caused by
a substituent at C(2).
Let us now consider 2-substituted precursors. We showed
above that, when the ring opening of 1-aza-bicyclo[1.1.0]-
butane 1a to 1,3-azadiene 2 is carried out thermally, the
central bond remains intact while two opposite lateral CϪC
and CϪN bonds break with formation of the product. It is
not possible to consider the regiochemistry of this reaction
in the case of the symmetrical 3-phenyl-1-aza-bicyclo[1.1.0]-
butane molecule. However, compounds 1bϪd, substituted
in the 2-position, should allow us to determine the regio-
chemistry of this reaction. In an unsymmetrically substi-
tuted molecule the cleavage of a pair of bonds [NϪC(2) and
C(3)ϪC(4) (path a) or C(2)ϪC(3) and C(4)ϪN (path b)]
should produce two isomeric substituted 2-aza-1,3-dienes.
These unstable products were trapped by acid-promoted hy-
drolysis at low temperature, giving the corresponding ke-
tones (Scheme 2). The ratios of ketones formed were deter-
Conclusion
1
Thermal degradation of unsubstituted or 2-substituted 3-
phenyl-1-azabicyclo[1.1.0]butanes under FVT conditions
favours the formation of the corresponding 3-phenyl-2-aza-
buta-1,3-dienes. The experimentally ascertained UV-PES
studies and the Density Functional Theory calculations
also highlight the great thermodynamic instability of the 2-
azabuta-1,3-diene system and the predominance of the
twisted synclinal conformer. The ring opening of azabicy-
clobutanes to 2-azadienes proceeds as biradical process
with cleavage of the two opposite bonds, the cleavage of the
CϪN bond in the first step being strongly favoured.
mined (see Exp. Sect.) from the H NMR spectra (Table 3).
We presumed that the relative ratios of ketones formed
correspond to the ratio of primary 2-azadienes.
Experimental Section
General Remarks: Melting points were determined in a Boe¨tius ap-
paratus and were uncorrected. The 1H NMR spectra were recorded
on a VARIAN GEMINI 200 spectrometer with TMS as an internal
standard. IR spectra were recorded on a Specord 75 IR instrument.
Scheme 2
Starting Materials: All starting 1-azabicyclo[1.1.0]butanes were
prepared by described procedures: 3-phenyl-1-azabicyclo[1.1.0]but-
ane,[4] 2-methyl-3-phenyl-1-azabicyclo[1.1.0]butane,[4] 2,2-dimethyl-
3-phenyl-1-azabicyclo[1.1.0]butane,[4] and cis-2,3-diphenyl-1-azab-
icyclo[1.1.0]butane.[6]
Table 3. The ratios of ketones obtained after hydrolyses of products
formed in FVT of 1aϪd
Azabicyclobutane
Path a
Path b
1a R1 ϭ R2 ϭ H
PhCOCH3 100%
FVT: The flash vacuum thermolyses were carried out in a 30 ϫ
2 cm horizontal quartz tube packed with quartz rings, electrically
heated to 400 °C at 1.5·10Ϫ3 Torr. Compounds 1aϪd were slowly
distilled into the thermolysis tube. The products were collected in
a CO2/acetone trap.
1b R1 ϭ Me, R2 ϭ H
1c R1 ϭ R2 ϭ Me
1d R1 ϭ Ph, R2 ϭ H
PhCOCH2CH3 93%
PhCOCH(CH3)2 76%
PhCOCH2Ph 87%
PhCOCH3 7%
PhCOCH3 24%
PhCOCH3 13%
Table 3 shows that the regiochemistry of the reaction de-
pends greatly on the presence of the substituents in the 2-
position. The thermal fragmentation of 1bϪd occurred
mainly through the cleavage of the NϪC(2) and C(3)ϪC(4)
bonds (pathway a).
Thermolysis of 1a
3-Phenyl-2-azabuta-1,3-diene (2): IR (CCl4, 25 °C): νCϭN 1640
cmϪ1 1H NMR (200 MHz, CCl4, 25 °C): δ ϭ 4.55 (br. s, 1 H),
.
4.85 (br. s, 1 H), 7.30Ϫ7.80 (m, 7 H) ppm.
For the ring opening of bicyclobutanes to butadienes,
Shevlin and McKee[8] predicted two-step mechanisms with
intermediate 3-butenylidene (especially for photolytic reac-
tion) or cyclopropylcarbinyl biradicals for this substituent-
containing system. The biradical mechanism was proposed
Reduction of 2 with LiAlH4: The product of the thermolysis of 1a
(655 mg, 5 mmol) was rinsed from the cold finger with cold THF
(5 mL) at 0 °C into a suspension of LiAlH4 (114 mg, 3 mmol) in
THF (10 mL). The mixture was heated to room temperature and
stirred for 1 hour, and was then hydrolysed with 20% NaOH. The
earlier by Dewar.[12] However, the azabicyclobutane system organic phase was separated and the water layer was extracted with
2478  2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.eurjoc.org
Eur. J. Org. Chem. 2003, 2475Ϫ2479

Flash Vacuum Thermolysis of 1-Azabicyclo[1.1.0]butanes
FULL PAPER
Et2O. The combined organic phases were dried (MgSO4) and con-
centrated to give a pale yellow liquid methylimine of acetophenone
(3, 598 mg, 90%). The IR and 1H NMR spectra of 3 were identical
with the spectroscopic data of an authentic sample of aceto-
phenone methylimine.[13]
Acknowledgments
We thank IDRIS (CNRS, Orsay) for calculation facilities.
[1]
Application of Photoelectron Spectroscopy to Molecular
Diels Alder Reaction between 3-Phenyl-2-azabuta-1,3-diene (2) and
N-(1-Ethoxyvinyl)-N,N-dimethylamine: The product of the ther-
molysis of 1a (655 mg, 5 mmol) was rinsed from the cold finger
with cold Et2O (5 mL) into a solution of N-(1-ethoxyvinyl)-N,N-
dimethylamine (585 mg, 5 mmol) in Et2O (10 mL) at Ϫ10 °C. The
mixture was stirred at Ϫ8 °C for 24 hours and concentrated. A
solution of chloranil (1.148 g, 4.8 mmol) in benzene (50 mL) was
added to the residue, and the reaction mixture was stirred for
30 min. The solvent was removed under reduced pressure to give a
crystalline solid. Recrystallization from a pentane/ethyl ether mix-
ture gave pure 4-dimethylamino-2-phenylpyridine (4, 230 mg, 23%),
identical with an authentic sample.[14]
Properties, Part 64. For part 63, see: K. Miqueu, G. Pfister-
Guillouzo, J. M. Sotiropoulos, V. L. Rudzevich, V. D. Rom-
anenko, G. Bertrand, Organometallics 2002, submitted.
W. Funke, Angew. Chem. Int. Ed. Engl. 1969, 8, 70Ϫ71.
A. G. Hortmann, D. A. Robertson, J. Am. Chem. Soc. 1967,
89, 5974Ϫ5975.
A. G. Hortmann, D. A. Robertson, J. Am. Chem. Soc. 1972,
94, 2758Ϫ2765.
For review see: R. Bartnik, A. P. Marchand, Synlett 1997,
[2]
[3]
[4]
[5]
1029Ϫ1039.
[6]
´
G. Alvernhe, A. Laurent, K. Touhami, R. Bartnik, G. Mloston,
J. Fluorine. Chem. 1985, 29, 363Ϫ384.
I. A. Dyakonov, V. V. Razin, M. I. Komendatov, Tetrahedron
Lett. 1966, 1135Ϫ1140.
P. B. Shevlin, M. L. McKee, J. Am. Chem. Soc. 1988, 110,
1666Ϫ1671.
J. C. Guillemin, J. M. Denis, A. Lablache-Combier, J. Am.
Chem. Soc. 1981, 103, 468Ϫ469.
J. C. Guillemin, J. M. Denis, Tetrahedron 1988, 44, 4447Ϫ4456.
H. Bock, R. Dammel, Chem. Ber. 1987, 1971Ϫ1986.
M. J. S. Dewar, J. Am. Chem. Soc. 1984, 106, 209Ϫ219.
A. Cobas, E. Guitian, L. Castedo, J. Org. Chem. 1993, 58,
3113Ϫ3117.
[7]
[8]
[9]
Acid Hydrolysis of 2-Aza-1,3-dienes 2a؊d: The products of the
thermolyses of 1aϪd (5 mmol) were rinsed from the cold finger
with methanol (5 mL) into mixtures of 4% HCl (5 mL) and meth-
anol (10 mL), cooled to Ϫ50 °C. After warming to room tempera-
ture, the mixture was concentrated under reduced pressure and ex-
tracted with Et2O. The extract was dried (MgSO4) and the solvents
were evaporated, yielding mixtures of the corresponding ketones.
[10]
[11]
[12]
[13]
1
The ratios of ketones were determined from the H NMR spectra
by comparison of the integrals of the singlet of the acetophenone
methyl group (δ ϭ 2.59 ppm) with the triplet of the propiophenone
methyl group (δ ϭ 1.19 ppm), the doublets of the two isobutyro-
phenone methyl groups (δ ϭ 1.20 ppm) and the singlet of the
deoxybenzoin methylene group (δ ϭ 4.21 ppm), respectively.
[14]
[15]
H. Neunhoeffer, H. W. Frühauf, Justus Liebigs Ann. Chem.
1972, 758, see pp. 125, 126, 137, 129, 131.
Gaussian 98, Revision A.9, M. J. Frisch, G. W. Trucks, H. B.
Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G.
Zakrzewski, J. A. Montgomery, Jr., R. E. Stratmann, J. C. Bu-
rant, S. Dapprich, J. M. Millam, A. D. Daniels, K. N. Kudin,
M. C. Strain, U. Parkas, J. Tomasi, V. Barone, M. Cossi, R.
Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J.
Uchterski, G. A. Petersson, P. Y. Ayala, Q. Cui, K. Morokuma,
D. K. Malick, A. D. Rabuck, K. Ragavachari, J. B. Foresman,
J. Cioslowski, J. V. Urtiz, A. G. Baboui, B. B. Stefanov, G. Liu,
A. Liashenko, P. Piskorz, L. Komaromi, R. Gomperts, R. L.
Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A.
Nanayakkara, M. Challacombe, P. M. W. Gill, B. G. Jonhson,
W. Chen, M. W. Wong, J. L. Andres, C. Gonzalez, M. Head-
Gordon, E. S. Replogle, J. A. Pople, Gaussian, Inc. Pittsburgh
PA, 1998.
UV-Photoelectron Spectra: The photoelectron spectra were re-
corded on a home-made spectrometerϪthree part assembly: main
body spectrometer (Meca2000), equipped with a spherical analyser
(Omicron), and a photon source (Focus) using HeI 21.21 eV radi-
ation. The spectrometer was monitored by use of a microcomputer
supplemented with a digital analogue converter. The spectra were
calibrated on the basis of the known auto-ionization of xenon at
12.13 and 13.45 eV, and of the nitrogen at 15.59 eV and 16.98 eV.
Thermolyses were carried out in a short path internal oven included
in the apparatus with 2 cm distance between oven exit and ionis-
ation head.
[16]
R. G. Parr, W. Yang, Functional Theory of Atoms and Mol-
ecules, Oxford Univ. Press, New York, 1989.
Computational Methods: Quantum calculations were performed
with Gaussian 98.[15] The optimization of all geometry parameters
and vibrational analysis were carried out by Density Functional
Theory[16] by use of the B3LYP[17Ϫ19] functional with a conjunction
of 6-311G(d,p) basis set. Total energies of molecules (ET in Har-
[17]
[18]
[19]
[20]
A. D. Becke, Phys. Rev. A 1988, 38, 3098Ϫ3100.
A. D. Becke, J. Chem. Phys. 1993, 98, 5648Ϫ5652.
C. Lee, W. Yang, R. G. Parr, Phys. Rev. B. 1988, 37, 785Ϫ789.
A. J. Arduengo, H. Bock, H. Chen, M. Denk, D. A. Dixon, J.
C. Green, W. A. Hermann, N. L. Jones, M. Wagner, R. West,
J. Am. Chem. Soc. 1994, 116, 6641Ϫ6649.
trees) were uncorrected. ∆SCF Ϫ first vertical ionization potential
calcd.
IP1v
was calculated as the difference ET (cation) Ϫ ET (neutral
molecule). Recent works[20Ϫ25] have shown that εiKS could be as-
sociated with experimentally determined vertical ionization poten-
tials (IPv) by uniform shift x ϭ |Ϫεi (HOMO) Ϫ IPvexp|. This ap-
proach gives a remarkable agreement with experimentally ascer-
tained values and is justified by the fact that the first calculated
vertical ionization potential IPvcalcd. [as the difference ET(cation) Ϫ
ET(neutral molecule)] lies very close to the experimentally deter-
mined values. Stowasser and Hoffman[26] have recently shown that
the localizations of KS orbitals are very similar to those obtained
after HF calculations, so it is possible to determine the nature of
the first ionizations and to interpret the photoelectron spectra un-
ambiguously.
[21]
[22]
[23]
[24]
[25]
[26]
H. M. Muchall, N. H. Werstiuk, J. L. Fitters, M. S. Workentin,
Tetrahedron 1999, 55, 3767Ϫ3778.
H. M. Muchall, N. H. Werstiuk, B. Choudhury, J. Ma, J. Wark-
entin, J. Pezacki, Can. J. Chem. 1998, 76, 238Ϫ240.
H. M. Muchall, N. H. Werstiuk, B. Choudhury, Can. J. Chem.
1998, 76, 221Ϫ227.
H. M. Muchall, P. Rademacher, J. Mol. Struct. 1998, 471,
189Ϫ194.
K. Miqueu, J-M. Sotiropoulos, G. Pfister-Guillouzo, H. Ranai-
´
vonjatovo, J. Escudie, J. Mol. Struct. 2001, 545, 139Ϫ146.
R. Stowasser, R. Hoffmann, J. Am. Chem. Soc. 1999, 121,
3414Ϫ3420.
Received February 1, 2003
Eur. J. Org. Chem. 2003, 2475Ϫ2479
www.eurjoc.org
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2479