New reactions of N-tert-butylimines; Formation of N-heterocycles by methyl radical elimination on flash vacuum thermolysis of N-benzylidene- and N-(2-pyridylmethylidene)-tert-butylamines

Article abstract of DOI:10.1002/chem.201301663

Thermal reactions of N-benzylidene- and N-(2-pyridylmethylidene)-tert- butylamines (5 and 13) under FVT conditions have been investigated. Unexpectedly, at 800 °C, compound 5 yields 1,2-dimethylindole and 3-methylisoquinoline. In the reaction of 13 at 800 °C, 3-methylimidazo[1,5- a]pyridine was obtained as the major product. Mechanisms of these reactions have been proposed on the basis of DFT calculations. Furthermore, UV-photoelectron spectroscopy combined with FVT has been applied for direct monitoring and characterization of the thermolysis products in situ. Copyright

Full text of DOI:10.1002/chem.201301663

FULL PAPER
DOI: 10.1002/chem.201301663
New Reactions of N-tert-Butylimines; Formation of N-Heterocycles by
Methyl Radical Elimination on Flash Vacuum Thermolysis of N-Benzylidene-
and N-(2-Pyridylmethylidene)-tert-butylamines
Thien Y Vu,[a, b] Anna Chrostowska,*[a] Thi Kieu Xuan Huynh,[b] Saꢀd Khayar,[a]
Alain Dargelos,[a] Katarzyna Justyna,[c] Beata Pasternak,[c] Stanisław Les´niak,*[c] and
Curt Wentrup*[d]
Abstract: Thermal reactions of N-benzylidene- and N-(2-pyridylmethylidene)-tert-
butylamines (5 and 13) under FVT conditions have been investigated. Unexpect-
edly, at 8008C, compound 5 yields 1,2-dimethylindole and 3-methylisoquinoline. In
the reaction of 13 at 8008C, 3-methylimidazo[1,5-a]pyridine was obtained as the
major product. Mechanisms of these reactions have been proposed on the basis of
DFT calculations. Furthermore, UV-photoelectron spectroscopy combined with
FVT has been applied for direct monitoring and characterization of the thermoly-
sis products in situ.
Keywords: density functional calcu-
lations · flash pyrolysis · heterocy-
cles · photoelectron spectroscopy ·
radicals
AHCTUNGTRENNUNG
Introduction
However, the details of these reactions are not well under-
stood. It is proposed that isobutene elimination takes place
via a four-membered cyclic transition state, analogously to
the previously described reactions of O- and S-tert-butyl de-
rivatives.[4] An alternative pathway involves the formation of
Flash vacuum thermolysis (FVT) is a special tool in organic
synthesis. In general, high temperature and low pressure
promote fragmentations, symmetry-allowed cycloreversions,
and rearrangement reactions. FVT has been successfully ap-
plied, for example, in the synthesis of numerous carbo- and
heterocyclic systems. Some typical reactive intermediates
generated under FVT conditions are ketenes, radicals, car-
benes, and nitrenes.[1,2] For compounds containing a tert-
butyl group attached to an sp2 or sp3 nitrogen atom, a typical
and facile reaction is the intramolecular elimination of iso-
butene and formation of the N-unsubstituted imine.[3]
À
a tert-butyl radical as a result of homolysis of a tert-butyl N
bond, which is known to take place in tert-butyl azo-com-
pounds.[5] The formation of isobutene from the tert-butyl
radical is readily understood,[4b] but formation of NH-imines
would seem to require another process. In a previous study,
McNab et al.[6] described transformations of iminyl radicals
into aza-heterocycles under the conditions of FVT. In our
recent work,[7] we have demonstrated that elimination of
isobutene from aliphatic N-tert-butylimines 1 may take place
(Scheme 1), but in some cases the course of the reaction is
entirely different, leading instead to the formation of N-het-
erocycles 2 or 4.
[a] T. Y. Vu, Prof. A. Chrostowska, Prof. S. Khayar, Prof. A. Dargelos
Universitꢁ de Pau et des Pays de lꢂAdour
Institut des Sciences Analytiques
et de Physico-Chimie pour lꢂEnvironnement
et les Matꢁriaux, UMR CNRS 5254, 64000 Pau (France)
E-mail: anna.chrostowska@univ-pau.fr
[b] T. Y. Vu, Prof. T. K. X. Huynh
Universitꢁ Nationale du Vietnam, ꢃ Ho Chi Minh Ville
227 Nguen Van Cu, 5ꢄ arr Ho Chi Minh Ville (Vietnam)
´
[c] K. Justyna, Dr. B. Pasternak, Prof. S. Lesniak
Department of Organic and Applied Chemistry
´
´
University of Łꢅdz, Tamka 12, 91-403 Łꢅdz (Poland)
Scheme 1. FVT of N-tert-butyl- and N-isopropylimines and diimines.
Fax : (+48)42-635-57-44
E-mail: slesniak@chemia.uni.lodz.pl
[d] Prof. C. Wentrup
School of Chemistry and Molecular Biosciences
The University of Queensland, Brisbane (Australia)
Fax : (+61)7-3365-4299
Results and Discussion
FVT of N-benzylidene-tert-butylamine (5): Initial FVT of 5
was performed at 8008C and 3ꢀ10À4 Torr. The products
E-mail: wentrup@uq.edu.au
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201301663.
1
were examined by H NMR spectroscopy immediately fol-
Chem. Eur. J. 2013, 19, 14983 – 14988
ꢆ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
14983

lowing the thermolysis. Surprisingly, FVT of 5 afforded two
heterocyclic products, 1,2-dimethylindole (6; 55%) and 3-
methylisoquinoline (7; 24%), as well as benzonitrile (8;
17%). In addition, a small amount (5%) of benzaldehyde
(10) was detected; this is ascribed to hydrolysis of benzyli-
denamine (benzaldimine; 9), which was not isolated, but
identified by photoelectron spectroscopy as described below.
The identities of all other products were established by com-
parison with authentic materials. The ratio of products was
1
determined by integration of the H NMR spectra, and their
full characterization was confirmed after separation by prep-
arative thin-layer chromatography.
FVT-PES of N-benzylidene-tert-butylamine (5): The reac-
tion shown in Scheme 2 was also monitored by direct, on-
line UV photoelectron spectroscopy (PES). The PES of N-
Scheme 2. Products of FVT of N-benzylidene-tert-butylamine (5).
benzylidene-tert-butylamine (5) and the spectra resulting
from FVT at 620 and 6908C are presented in Figure 1 and
compared to the spectrum of an authentic sample of 1,2-di-
methylindole (6). DFT [DSCF/TD-DFT (CAM-B3LYP)]
and ab initio (OVGF, P3 and SAC-CI) calculations of ioni-
Figure 1. UV-photoelectron spectrum of a) N-benzylidene-tert-butylamine
(5); b) FVT of 5 at 6208C giving benzylidenamine 9 (8.8 eV); c) FVT of 5
at 6908C giving 6; d) 1,2-dimethylindole 6 (grey curve and numbers).
zation energies using the 6-311G
G
We propose possible reaction mechanisms in the light of
out on the optimized geometries of each compound. The
data for 5 are presented in Table S1 (see the Supporting In-
formation).
DFT calculations at the CAM-B3LYP/6-311GACTHNGUTER(NNUG d,p) level.
The most puzzling result, which requires an explanation, is
the formation of indole 6 as a major product. Firstly, two
methyl groups are attached to the same carbon atom in radi-
cal a1 (Scheme 3), whereas in the final product 6 they are
located on two different atoms (carbon and nitrogen). Sec-
ondly, in the imine moiety of 5, the sequence of atoms is C-
N-C, whereas in the product 6 the sequence is C-C-N. Third-
ly, the major cyclic product is 1,2-dimethylindole (6), a com-
pound that cannot be formed by any of the pathways previ-
ously described for imines.[6,7]
Considering that 6 and 7 are the major products, it would
seem that elimination of a methyl radical from the tert-butyl
group in 5 is the initial reaction. This has a calculated activa-
tion barrier of 77 kcalmolÀ1 (Figure 2 and Figure 3). Impor-
tantly, this is the highest activation barrier on all the energy
profiles examined. Thus, as soon as the initial radical a1 is
formed, the subsequent steps should be relatively facile. The
2-aza-allyl radical a1, so formed, is postulated to undergo
a 1,4-H shift to generate an isomeric aza-allyl radical a2,
from which cyclization to a3 can take place (Scheme 3).
As can be seen in Figure 1, FVT of 5 at 6208C results in
consumption of the starting material and appearance of new
ionization bands in the UV spectra at 9.45, 11.8, and 12.9 eV
due to isobutene[8] as well as bands at 8.8, 9.5, and 9.8 eV,
which are ascribed to the first two p orbitals and the nitro-
gen lone pair of N-benzylideneamine 9, respectively (see
Table S2 in the Supporting Information for details).
An increase of the FVT temperature to 6908C led to the
formation of 1,2-dimethylindole (6) as a major product. This
is confirmed by the presence of the relevant ionization
bands at 7.5, 7.9, 9.5 and 10.6 eV. Isoquinoline 7 is also un-
doubtedly present (8.4 eV),[9] but overlapping bands make
its observation by PES difficult. It should be noted that dif-
ferent dimensions and pressures used in the preparative and
analytical FVT apparatus means that temperatures and
yields are not directly comparable. Further theoretical eval-
uation of ionization energies is presented in the Supporting
Information.
14984
www.chemeurj.org
ꢆ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 14983 – 14988

Reactions of N-tert-Butylimines
FULL PAPER
of CH3 to afford the 2H-indole
intermediate 12 and the final
product 6, respectively, are
straightforward (Scheme 3 and
Figure 3a).
Alternatively,
1-aziridinyl
radical a4 may rearrange to en-
aminyl radical a8 by means of
a methyl group shift and ring
opening. Cyclization and elimi-
nation of a hydrogen atom then
leads to the final product 6
(Scheme 3). The calculated
energy profile for this mecha-
nism is shown in Figure 3b. The
aziridinyl radical a4 is formed
relatively easily. The highest ac-
tivation barrier of 70 kcalmolÀ1
is for the methyl group shift,
which is still below the value of
77 kcalmolÀ1 calculated for the
initial step. A concerted path
from a4 to a8 was sought but
not found, which is not surpris-
ing in view of the nontrivial
energy profile for ring opening
of the cyclopropyl radical.[11]
Once the rearranged aziridinyl
radical a7 has formed, the re-
maining steps have readily ac-
cessible activation barriers.
These calculated energy profiles
serve to illustrate that the ob-
served reactions are readily ac-
counted for by means of rea-
sonable reaction mechanisms.
However, further refinement of
the mechanisms by substituent
and isotope labeling is desira-
ble. The system 13, which will
be discussed next, was con-
ceived to achieve a more clear-
cut outcome.
Scheme 3. Proposed mechanisms for rearrangements of N-benzylidene-tert-butylamine (5).
Figure 2. Calculated energy profile (CAM-B3LYP/6-311GACTHNUTRGNEUGN(d,p) calculations) for the rearrangement of benzyli-
dene-tert-butylamine (5) to 3-methylisoquinoline (7). Relative energies in kcalmolÀ1. The activation barriers
for each individual step are shown.
There is precedence for the cyclization of aza-allyl radi-
cals.[10] Elimination of a hydrogen atom from a3 followed by
aromatization affords isoquinoline 7 (Scheme 3). The calcu-
lated energy profile is shown in Figure 2. Apart from the
first step (77 kcalmolÀ1), the highest barrier is 58 kcalmolÀ1.
Such reactions are feasible under high-temperature FVT
conditions. Several alternative pathways were examined
computationally, but the reaction shown in Scheme 3 and
Figure 2 has the lowest overall activation energies.
The initial aza-allyl radical a1 may also cyclize to an aziri-
dinyl radical a4. Attack on the benzene ring affords a new
and strained radical a5, which will undergo ring expansion
to the more stabilized dihydro-N-indolyl radical a6
(Scheme 3). From here, the elimination of H and a 1,5-shift
The formation of benzonitrile 8 from 5 is readily ex-
plained by direct homolysis of the N-tert-butyl bond, which
results in the formation of an iminyl radical from which 8
can form by loss of a hydrogen atom. Benzylidenamine 9
will result from elimination of isobutene from 5 via a four-
membered cyclic transition state.[4] The hydrolysis of 9 yields
benzaldehyde 10 in the preparative experiments.
Interestingly, no isoindoles were detected, although they
could have been expected on the basis of the sequence of
atoms in the starting material 5. Furthermore, control ex-
periments established that indole 6 was not converted into
isoquinoline 7 on FVT up to 10008C; no traces of either 7
or methylquinolines were detectable by high-field NMR
spectroscopy. These control experiments were important be-
Chem. Eur. J. 2013, 19, 14983 – 14988
ꢆ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14985

´
A. Chrostowska, S. Lesniak, C. Wentrup et al.
FVT of N-(2-pyridylmethyli-
dene)-tert-butylamine
Preparative thermolysis
(13):
of
imine 13 was performed at
8008C and 3ꢀ10À4 Torr. This
temperature was optimal for
the formation of the major
product 14, although a part of
the starting material 13 re-
mained unchanged. The FVT
products were examined by
high-resolution 1H NMR spec-
troscopic analysis performed
immediately after the reaction.
The three products 14, 15, and
17 were isolated and separated
(Scheme 4), and ca. 4% of
other, unidentified products
were formed in yields too low
for separation and characteriza-
tion (the yield of 4% is based
on integration for four un-
known compounds possessing
protons in the 2-position of pyr-
idine). The major product of
the reaction, formed in 80%
yield, was identified as 3-
methylimidazoACTHNUTRGNEUNG
[1,5-a]pyridine[15]
(14). In addition, pyridine 2-car-
bonitrile (15) and pyridine 2-
carboxaldehyde
(17)
were
formed in low yield.
FVT-PES of N-(2-pyridylme-
thylidene)-tert-butylamine (13):
The UV-photoelectron spec-
trum of 13 and the spectrum of
the products of FVT at 6508C
are shown in Figure 4. The in-
terpretation of these spectra is
corroborated by the theoretical
evaluation of the ionization en-
ergies (Tables S4 and S5, Sup-
porting Information). Thus, for
13 the first PE band at 9.0 eV
corresponds to the ionization of
the HOMO of the imine nitro-
Figure 3. Two alternative mechanisms for the rearrangement of benzylidene-tert-butylamine (5) to 1,2-di-
methylindole (6). Relative energies in kcalmolÀ1. The activation barriers for each individual step are shown
(CAM-B3LYP/6-311G(d,p) calculations).
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
gen lone pair, and the second
band at 9.15 eV to an MO aris-
cause there are known cases of ring expansion of methyl-
pyrroles and -indoles to pyridines and quinolines under
high-temperature flow pyrolysis conditions.[12] Neither dihy-
droisoquinoline 11 nor 2H-indole 12 were detectable either,
and they are expected to aromatize readily under the reac-
tion conditions. There are very few studies on nitrogen-cen-
tered aziridinyl radicals in the literature,[13] and their use in
synthesis has been little explored.[14]
ing from an antibonding interaction between pyridine and
imine C=N bonds. The third band at 9.65 eV is assigned to
the pyridine nitrogen lone pair.
Upon FVT at 6508C the bands due to the precursor 13
disappear, and new bands observed are in good agreement
with calculations for 14 (Table S5). Thus, the first band at
very low IE (7.6 eV) corresponds to the HOMO, and the
bands at 9.4 and 10.6 eV are ascribed to orbitals arising from
p–p interactions. The band at 11.7 eV reflects an antibond-
14986
www.chemeurj.org
ꢆ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 14983 – 14988

Reactions of N-tert-Butylimines
FULL PAPER
action between the pyridine
lone pair and the free radical
electron located in a p-orbital
stabilizes radical b1, thereby
lowering the activation energy
for its formation by about
7 kcalmolÀ1 compared with the
formation of a1 (cf. Figure 3).
The formation of a bond be-
tween the two centers in b1
generates the similarly reso-
Scheme 4. Products of FVT of N-(2-pyridylmethylidene)-tert-butylamine (13).
nance stabilized radical b2. This
is formally a 5-endo-trig cycliza-
tion,[16] but the usually favored 4-exo-trig reaction would not
lead to a stable species in this case.
Scheme 5. Proposed mechanism for formation of imidazopyridine 14.
Figure 4. UV-photoelectron spectrum of a) N-(2-pyridinylmethylidene)-
tert-butylamine (13); b) Products of FVT of 13 at 6508C.
Conclusion
ing interaction between the two rings. The shoulder at
9.8 eV is likely due to 2-pyridylmethanimine 16, but this is
tentative because of overlapping bands. Further analysis of
ionization energies is presented in the Supporting Informa-
tion. It is clear from both preparative and analytical FVT
that imidazopyridine 14 is the major product. Thus, the reac-
tion of 13 is relatively clean, yielding 14 regioselectively and
with retention of the atom sequence C-N-C. The formation
of pyridine-2-carbonitrile 15 and pyridine-2-carboxaldehyde
17 is readily understood in the same manner as described
for 5 above.
A new reaction of aromatic tert-butylimines under FVT con-
ditions is described. At high temperature, the elimination of
a methyl radical from the N-tert-butyl group is the first step
of a cascade reaction, accounting for at least 80% of the
products. The resulting, resonance-stabilized aza-allyl radical
undergoes cyclization to isoquinoline 7, indole 6, or imida-
A
reaction mechanism, similar to that depicted in
Scheme 3, starts with the elimination of a methyl radical
from the tert-butyl group to give resonance-stabilized aza-
allyl radical b1. Due to the presence of the pyridine-N,
a facile cyclization to the intermediate aminyl b2 can occur
(Scheme 5). In contrast to the reaction presented in
Scheme 2, the transformation of 13 into 14 requires sequen-
tial elimination of two methyl radicals. This mechanism is
supported by the calculated energy profile (Figure 5). Here,
the highest activation barrier (70 kcalmolÀ1) is again for the
À
initial C C bond cleavage in the tert-butyl group. All the
Figure 5. Calculated energy profile for the rearrangement of N-(2-pyri-
dylmethylidene)-tert-butylamine (13) to 3-methylimidazoACTHNUTRGNE[UNG 1,5-a]pyridine
subsequent steps have relatively low barriers. The ease and
cleanliness of the reaction can be ascribed to the pyridine
lone pair, which facilitates cyclization of b1 to b2. An inter-
(14). The activation barriers for each individual step are shown (relative
energies in kcalmolÀ1 calculated at the CAM-B3LYP/6-311G
ACTHNUGRTENUNG(d,p) level).
Chem. Eur. J. 2013, 19, 14983 – 14988
ꢆ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14987

´
A. Chrostowska, S. Lesniak, C. Wentrup et al.
Phase Reaction in Organic Synthesis Gordon and Breach, 1997;
e) H. McNab, Aldrichimica Acta 2004, 37, 19–26; f) G. I. Yranzo, J.
Elguero, R. Flammang, C. Wentrup, Eur. J. Org. Chem. 2001, 2209–
2220.
zopyridine 14. In the case of 6, a rearrangement is required.
This work opens up avenues to unexpected chemistry of
imines and aza-allyl radicals. We will investigate the synthet-
ic scope and further mechanistic details of these reactions.
[2] For recent examples, see: a) S. Lesniak, B. Pasternak, Tetrahedron
Lett. 2005, 46, 3093–3095; b) S. Lesniak, R. B. Nazarski, B. Paster-
nak, Tetrahedron 2009, 65, 6364–6369; c) C. Wentrup, Acc. Chem.
Res. 2011, 44, 393–404; d) R. Leardini, H. McNab, D. Nanni, A. G.
Tenan, A. Thomson, Org. Biomol. Chem. 2012, 10, 623–630.
[3] a) J. D. Perez, G. I. Yranzo, L. M. Phagouape, Bull. Soc. Chim Fr.
1986, 1, 129–132b) V. Metail, S. Joanteguy, A. Chrostowska-Senio,
G. Pfister-Guillouzo, A. Systermans, J. L. Ripoll, Inorg. Chem. 1997,
36, 1482–1487; c) V. Lefꢄvre, J. L. Ripoll, Y. Dat, S. Joanteguy, V.
Metail, A. Chrostowska-Senio, G. Pfister-Guillouzo, Organometallics
1997, 16, 1635–1640; d) J. Levillain, G. Pfister-Guillouzo, J. L.
Ripoll, Eur. J. Org. Chem. 2000, 3253–3259.
Experimental Section
All reagents were purchased from commercial suppliers and used without
further purification. NMR spectra were recorded with a Bruker instru-
ment at 600 MHz with CDCl3 as solvent and TMS as internal standard.
Chemical shifts d are reported in parts per million, and coupling con-
stants J are given in Hertz. Data are reported as s=singlet, d=doublet,
t=triplet m=multiplet. Melting points were measured with a MELTEMP
apparatus and are uncorrected. Analytical TLC was performed on Merck
5554 aluminum-backed SiO2 plates. Preparative TLC was performed by
using Merck silica gel 60 (PF254) and a mixture of hexane and ethyl ace-
tate as eluent. Products were visualized by UV light (264 nm). IR spectra
were recorded with a Thermo-Nicolet Nexus spectrometer as neat sam-
ples (liquid films on NaCl plates or in KBr discs).
[4] a) T. Zhang, L. Zhang, J. Wang, T. Yuan, X. Hong, F. Qi, J. Phys.
Chem. A 2008, 112, 10495–10501; b) B. Mondal, D. Mandal, A. K.
Das, J. Phys. Chem. A 2011, 115, 3068–3078.
[5] a) J. Pacansky, J. S. Chang, D. Brown, Tetrahedron 1982, 38, 257–
260; b) P. S. Engel, P. Li, K. H. Whitmire, J. Guzman-Himenez,
M. R. Willcot, W. B. Smith, J. Org. Chem. 2000, 65, 1016–1021.
[6] a) H. McNab, J. Chem. Soc. Perkin Trans. 1 1980, 2200; b) C. L.
Hickson, H. McNab, Synthesis 1981, 464; c) H. McNab, G. S. Smith,
J. Chem. Soc. Chem. Commun. 1982, 996–997; d) C. L. Hickson, H.
McNab, J. Chem. Soc. Perkin Trans. 1 1984, 1569; e) H. McNab, Al-
drichimica Acta 2004, 37, 19–26.
[7] S. Lesniak, B. Pasternak, K. Justyna, T. Y. Vu, T. K. X. Huynh, S.
Khayar, A. Dargelos, A. Chrostowska, Tetrahedron 2013, 69, 722–
729.
[8] K. Kimura, S. Katsumata, Y. Achiba, T. Yamazaki, S. Iwata, Hand-
book of He I Photoelectron Spectra of Fundamental Organic Mole-
cules, Japan Scientific, Tokyo, 1981 and references cited therein.
[9] W. Schꢇfer, A. Schweig, H. Vermeer, F. Bickelhaupt, H. de Graaf, J.
Electron Spectrosc. Relat. Phenom. 1975, 6, 91–98.
Details of the FVT-photoelectron spectroscopy procedure are given in
the Supporting Information.
Preparative FVT; General procedure: The FVT reactions were carried
out in a 30ꢀ2.5 cm (lengthꢀinternal diameter) electrically heated hori-
zontal quartz tube packed with quartz rings (1 cm diameter). Imines 5
and 13 (2 mmol) were slowly vaporized from a flask held at RT into
a thermolysis tube preheated to 8008C at 3ꢀ10À4 hPa. The vacuum was
achieved with the use of silicone oil diffusion pump, and its value was
measured continuously during the pyrolysis. The products were collected
in a solid CO2/acetone trap. After thermolysis, the system was brought to
atmospheric pressure and allowed to warm slowly to RT. A small portion
1
of the pyrolyzate was dissolved in CDCl3 for the H NMR measurement,
and the remainder was dissolved in Et2O for chromatographic separation
and purification.
[10] For cyclization under FVT conditions, see: S. Lesniak, B. Pasternak,
Synth. Commun. 2002, 32, 875–880.
Thermolysis of 5: The mixture of products 6, 7, benzonitrile (8), and
benzaldehyde (10) was separated by preparative TLC (SiO2; hexane/
EtOAc, 6:4) and identified by comparison with commercially available
compounds.
[11] a) S. Olivella, A. Solꢁ, J. M. Bofill, J. Am. Chem. Soc. 1990, 112,
2160–2167; b) P. A. Arnold, B. K. Carpenter, Chem. Phys. Lett.
2000, 328, 90–96; c) D. J. Mann, W. L. Hase, J. Am. Chem. Soc.
2002, 124, 3208–3209; d) D. J. Mann, W. L. Hase, Phys. Chem.
Chem. Phys. 2002, 4, 5066–5071; e) W. Quapp, J. M. Bofill, J. Math.
Chem. 2012, 50, 2061–2085.
[12] a) J. M. Patterson, C. F. Mayer, W. T. Smith, J. Org. Chem. 1975, 40,
1511–1514; b) J. M. Patterson, J. W. de Haan, J. Am. Chem. Soc.
1972, 94, 2487–2494; c) J. M. Patterson, M. Drenchko, J. Org. Chem.
1962, 27, 1650–1652; d) J. M. Patterson, J. Brasch, M. Drenchko, J.
Org. Chem. 1962, 27, 1652–1659.
[13] a) W. C. Danen, T. T. Kensler, Tetrahedron Lett. 1971, 12, 2247;
b) J. C. Brand, M. D. Cook, B. P. Roberts, J. Chem. Soc. Perkin
Trans. 2 1984, 1187–1196; c) I. MacInnes, J. C. Walton, D. C. Nonhe-
bel, J. Chem. Soc. Perkin Trans. 2 1987, 1789–1794; d) X.-Z. Qin, F.
Williams, J. Phys. Chem. 1986, 90, 2292–2296.
[14] a) X. Yang, A. K. Yudin, Synlett 2007, 2912–2918; b) A. Baralle, A.
Baroudi, M. Daniel, L. Fensterbank, J.-P. Goddard, E. Lacꢈte, M.-
H. Larraufie, G. Maestri, M. Malacria, C. Ollivier, Main-Group Ele-
ments in Radical Chemistry, in Encyclopedia of Radicals in Chemis-
try, Biology and Materials, John Wiley & Sons 2012, Vol. 2, Chapter
28, p 767–816.
Thermolysis of 13: The mixture of products 14, pyridine 2-carbonitrile
(15) and pyridine 2-carboxaldehyde (17) was separated by preparative
TLC (SiO2; hexane/EtOAc, 6:4). 3-MethylimidazoACTHNUTRGNE[NUG 1,5-a]pyridine (14)
was recrystallized from cyclohexane; m.p. 49–508C (lit.[15] m.p. 44–488C).
Because the spectral data for 14 described in ref. [15] are incomplete, ad-
ditional data are reported here: 1H NMR (600 MHz, 258C, TMS): d
=7.63 (dd, 3J(HH)=7.2 Hz, 4J
(H,H)=0.6 Hz, 1H), 7.37 (dt, 3J(HH)=
ACHTUNGTRENNUNG
9.6 Hz, 4J
ACHTUNGTRENNUNG(H,H)=0.6 Hz, 1H), 7.32 (s, 1H), 6.64–6.61 (m, 1H), 6.50–6.53
(m, 1H), 2.63 ppm (s, 3H; CH3); 13H NMR (150 MHz, 258C, TMS): d
=134.9 (Cq), 130.4 (Cq), 120.6 (CH), 118.5 (CH), 118.3 (CH), 112.2 (CH),
12.5 ppm (CH3); IR (KBr): n˜ =3019, 2920, 1636, 1520, 1496, 1443, 1396,
1360, 1320, 1171, 1078, 1000, 970, 905, 783 cmÀ1; MS
ACTHNUTRGNEUNG(ESI, +ve): m/z: 133
[M+H]; elemental analysis calcd (%) for C8H8N2 (132.1): C 72.70, H
6.10, N 21.20; found: C 72.68, H 6.13, N 21.16.
Acknowledgements
[15] S. El Khadem, J. Kawai, D. L. Swartz, Heterocycles 1989, 28, 239–
248.
[16] J. E. Baldwin, J. Chem. Soc. Chem. Commun. 1976, 734.
A.C. and T.Y.V. thank the Agence Universitaire de la Francophonie
Asie-Pacifique for financial support towards PhD studies. The authors
thank Mr. Patrick Baylꢄre for efficient technical assistance.
Received: May 1, 2013
Revised: July 23, 2013
[1] For reviews see: a) R. F. C. Brown, Pyrolytic Methods in Organic
Chemistry, Academic Press, New York, 1980; b) U. E. Wiersum, Al-
drichimica Acta 1984, 17, 31–40; c) J. I. G. Cadogan, C. L. Hickson,
H. McNab, Tetrahedron 1986, 42, 2135–2165; d) Y. Vallꢁe, Gas
Published online: September 20, 2013
Minor changes have been made to this manuscript since its publication in
Chemistry–A European Journal Early View. The Editor.
14988
www.chemeurj.org
ꢆ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 14983 – 14988