The reaction of NH-indazoles with 1-fluoro-2,4-dinitrobenzene: The unusual formation of benzotriazole-N-oxides

Article abstract of DOI:10.1039/c3nj00186e

When N-unsubstituted indazoles, like indazole itself, reacted with 1-fluoro-2,4-dinitrobenzene or 1-chloro-2,4,6-trinitrobenzene, three products were obtained whose structures were determined by X-ray diffraction. Besides the two N-substituted nitroaryl derivatives, a third compound was obtained with the same molecular formula (C₁₃H₈N₄O₄) to which was assigned the structure of a derivative of benzotriazole N-oxide. With the combined use of crystallography, NMR and DFT calculations this reaction was studied with special stress on the mechanism of formation of the benzotriazole-N-oxide.

Full text of DOI:10.1039/c3nj00186e

NJC
View Article Online
View Journal
PAPER
The reaction of NH-indazoles with 1-fluoro-2,4-
dinitrobenzene: the unusual formation of
Cite this: DOI: 10.1039/c3nj00186e
benzotriazole-N-oxides†‡
a
b
a
a
´
Ibon Alkorta,* Fernando P. Coss´ıo, Jose Elguero, Nieves Fresno,
a
c
c
´
Laura Hernandez-Folgado, Santiago Garc´ıa-Granda, Laura Menendez-Taboada,
Ruth Perez-Fernandez, Felipe Reviriego and Lucia Vazquez-Vin˜uela
a
a
c
´
´
´
When N-unsubstituted indazoles, like indazole itself, reacted with 1-fluoro-2,4-dinitrobenzene or
1-chloro-2,4,6-trinitrobenzene, three products were obtained whose structures were determined by
X-ray diffraction. Besides the two N-substituted nitroaryl derivatives, a third compound was obtained
with the same molecular formula (C13H8N4O4) to which was assigned the structure of a derivative of
benzotriazole N-oxide. With the combined use of crystallography, NMR and DFT calculations this
reaction was studied with special stress on the mechanism of formation of the benzotriazole-N-oxide.
Received (in Montpellier, France)
18th February 2013,
Accepted 30th April 2013
DOI: 10.1039/c3nj00186e
www.rsc.org/njc
formyl group (185 Hz).2 The results obtained with 1-chloro-
2,4,6-trinitrobenzene were similar.
Introduction
Many years ago, some of us reported an unprecedented reaction
between NH-indazoles and 1-halo-2-nitrobenzenes, such as
1-fluoro-2,4-dinitrobenzene (2) and 1-chloro-2,4,6-trinitrobenzene
(picryl chloride).1,2 In Fig. 1 are reported the results obtained in
the first case, using three conditions, reflux in ethanol (10 min),
reflux in xylene (60 h) and fusion. The percentages vary from one
experiment to the other, here are the values obtained in the first
case: 29% of 3, 62% of 4 and 9% of 5.
NH-Indazoles such as indazole itself 1 and 7-methylindazole
exist as a mixture of two tautomers, a and b, but the 1H-tautomer a
is largely predominant.3–5 When reacting 1 with 1-fluoro-2,4-dini-
trobenzene (2) three products were isolated: the 1-dinitrophenyl
derivative 3, which is not formed, for obvious steric reasons, when
there is a methyl group at the 7-position; the 2-dinitrophenyl
derivative 4, and a third isomer 5.1 The compounds were tentatively
identified by 1H NMR1 and, in the case of 5, by the 1J(1H–13C) of the
To explain the formation of the benzotriazole N-oxide 5 we
proposed the following mechanism (Fig. 2).2 It was based on
general knowledge but without any experimental proof.
It involves the minor 2H-tautomer 1b but according to the
Gustafsson paradox,3,6–10 the less stable tautomer 1b should be
the most reactive. A 1,3-dipolar cycloaddition of the o-nitro
group on the 1-azadiene moiety of 2H-indazole 1b affords 6 that
followed by a nucleophilic substitution of the fluorine atom
yields 7. A [2+2+2] retrocycloaddition would transform 7 into
the azo–nitroso derivative 8, an intermediate in the synthesis of
benzotriazole N-oxides such as 5 (usually not isolated in the
reduction of an azo–nitro derivative).11,12
We decided to study again the reaction of indazole 1 with
1-fluoro-2,4-dinitrobenzene (2) yielding 3, 4 and 5 with the
following purposes:
(i) To verify the structure of these compounds in the solid
state (X-ray crystallography).
a
´
´
Instituto de Quımica Medica, C.S.I.C., Juan de la Cierva, 3, E-28006 Madrid,
(ii) To carry out GIAO/DFT calculations to determine, by compar-
ison with NMR chemical shifts, some aspects of their structure.
(iii) To search for a mechanism for the reaction 1 + 2 - 5 to
be more consistent with present day knowledge.
Spain. E-mail: ibon@iqm.csic.es
b
c
´
´
Departamento de Quımica Organica I, Kimika Fakultatea,
´
Universidad del Paıs Vasco – Euskal Herriko Unibertsitatea,
´
Manuel de Lardizabal 3, 20018 San Sebastian-Donostia, Spain
´
´
´
´
Departamento de Quımica Fısica y Analıtica, Facultad de Quımica,
Universidad de Oviedo-CINN, E-33006 Oviedo, Spain
Results and discussion
† Dedicated to Professor Alan R. Katritzky on the occasion of his 85th birthday.
‡ Electronic supplementary information (ESI) available: General experimental
methods; 1H and 13C NMR spectra; electronic spectra; computational data. CCDC
806879, 806880, 806881 and 923952. For ESI and crystallographic data in CIF or
other electronic format see DOI: 10.1039/c3nj00186e
Since we have conserved the reaction crude of 1967, we carried
out the separation of the three isomers (C13H8N4O4) using
column chromatography (dichloromethane/hexane/methanol).
c
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013
New J. Chem.

View Article Online
Paper
NJC
Fig. 1 The 1967–1968 results (under the three conditions the relative abundances were similar).
Fig. 2 The 1968 proposed mechanism of the formation of benzotriazole N-oxide 5.
A fourth product was isolated: the hemiacetal of 5, a rather
unstable compound that is irrelevant for the problem under
study (in 1967 we used fractional crystallization). We were
fortunate to obtain suitable single crystals of 3, 4 and 5.
Part I: structure and the mechanism of formation of the
1-(2,4-dinitrophenyl)-1H-indazole (3) and 2-(2,4-dinitrophenyl)-
2H-indazole (4)
Crystallography. The structures of 3 and 4 display no noticeable
modifications in crystal packing at different temperatures. Both
compounds form parallel chains to the bc plane, the molecules
forming the chains are linked by H-bonds. In both cases, there
are weak intramolecular interactions. For compound 3, the
structure at room temperature does not present changes with
Fig. 3 A view of 3 with displacement ellipsoids drawn at the 50% probability
that derived from low temperature data, except for a higher
structure expansion due to higher atomic displacements.
level. The numbering does not follow the systematic one.
Crystal structure of 3. The structure is represented in Fig. 3.
The 3D structure is formed by chains parallel to the bc plane N(4) of one molecule, O(3) and O(4), and the phenyl carbons of
(Fig. 4). In these chains the molecules are linked by hydrogen the other molecule, C(12) and C(13). The O(4) forms a hydrogen
bonds C–Hꢀ ꢀ ꢀO between the oxygen atoms of the nitro group bond with C(6) of the indazole benzene ring.
c
New J. Chem.
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013

View Article Online
NJC
Paper
The molecules of the dimer are linked by C9B–H9Bꢀ ꢀ ꢀO1A
hydrogen bonds and by p–p stacking interactions, with centroid–
centroid distances of 3.804(1) and 3.709(1) Å (Fig. 7). The crystal
packing is formed by chains parallel to the bc plane. In these
chains the molecules are linked by weak hydrogen bonds
C–Hꢀ ꢀ ꢀO between the oxygen atoms of both nitro groups and
the benzene CHs of the other molecule along the b axis
(C2B–H2Bꢀ ꢀ ꢀO4B, C7B–H7Bꢀ ꢀ ꢀO2B, C9B–H9Bꢀ ꢀ ꢀO2B) (Fig. 8).
The hydrogen bonds of structures 3 and 4 are listed in Table 7.
Optimized vs. experimental geometries. Calculations were
carried out at the B3LYP/6-311++G(d,p) level. For structures 3 and 4
we have started with geometries where the nitro group is on the
same side as the ‘‘pyridine-like’’ nitrogen atom of the indazole ring
(the so-called ‘‘orthogonal interaction’’).13 Then, we found that the
calculated chemical shifts (next section) were consistent with the
experimental values in the case of 3 but not in the case of 4.
Therefore, we calculated the opposite conformation (40) finding that
it was slightly more stable than 4 (0.7 kJ molꢁ1) and subsequently
we used this conformation for the NMR calculations. Remember
that the X-ray structure of the 2-dinitrophenyl derivative, above
discussed, has the 4 or Z conformation.
Fig. 4 View of the layers along the bc plane of 3.
The bond lengths and bond angles are similar for the
experimental and calculated geometries of 3 and 40. Table 1
reports the torsion angles (in 1) that show only small differences.
NMR spectroscopy and GIAO/DFT calculations. The
chemical shifts of both dinitrophenyl indazoles (see Experi-
mental section) are reported in Table 2 together with the
chemical shifts calculated at the GIAO//B3LYP/6-311++G(d,p)
level (see Computational details).
Fig. 5 Packing diagram of 3.
There are intramolecular interactions between N(2) and N(3)
which fix the orientation of the N(3) nitro group. The distance
between N(2) and N(3) is 2.765(2) while N(2) and O(1) is
2.834(2) Å. The chains are supported by weak C–Hꢀ ꢀ ꢀO hydro-
gen bonds between the oxygen atoms of both nitro groups and
the phenyl carbons of the other molecule, without noticeable
strong interactions located (Fig. 5).
Crystal structure of 4. The asymmetric unit of 2-(2,4-dinitro-
phenyl)-2H-indazole is a dimer (Fig. 6). There are intra-
molecular interactions between N(1) and N(3) stabilizing the
orientation of the N(3) nitro group in both molecules of the
dimer, with distances 2.739(3) and 2.737(3) Å respectively. The
We have transformed the absolute shieldings (s, ppm) into
distances between N(1) and O(1) are 2.881(3) Å for molecule A, chemical shifts (d, ppm) using the two empirical equations we
and 2.828(2) Å for molecule B. have established previously:
Fig. 6 A view of 4 at 120 K with displacement ellipsoids drawn at the 50% probability level. The numbering does not follow the systematic one.
c
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013
New J. Chem.

View Article Online
Paper
NJC
Fig. 7 View of the layers along the bc plane of 4 at 120 K.
Table 2 Experimental chemical shifts (d, ppm, in DMSO-d6) and calculated
chemical shifts (d, ppm) of compounds 3 and 40
0
0
0
0
H6
Comp.
H3
H4
H5
H6
H7
H3
H4 H5
CHO
1H NMR
3
Exp. 8.52 7.95 7.38 7.61 7.82 8.92
Calcd 8.00 7.75 7.31 7.51 7.56 8.71
—
—
—
—
8.65 8.29
8.50 7.67
8.69 8.29
8.37 8.34
—
—
—
—
40 Exp. 9.08 7.81 7.15 7.35 7.64 8.93
Calcd 7.74 7.56 7.08 7.26 7.70 8.72
Comp.
C3
C3a
C4
C5
C6
C7
C7a
CHO
13C NMR
3
Exp.
Calcd 138.7 127.9 122.0 123.0 127.9 108.4 139.2
40 Exp.
128.5 123.2 121.8 122.0 128.8 118.2 150.8
Calcd 122.9 126.0 120.8 124.2 128.9 119.6 152.2
139.0 125.3 122.1 123.2 128.5 110.2 136.1
—
—
—
—
Fig. 8 Packing diagram of 4 at 120 K.
0
0
0
0
0
0
C6
C1
C2
C3
C4
C5
3
Exp.
Calcd
Exp.
143.5
140.3
137.3
140.0
138.6
147.2
144.6
144.4
121.7
123.7
122.0
122.5
145.2
145.0
147.2
145.5
128.7
126.7
128.5
127.4
126.2
123.1
128.7
128.3
d1H = 31.0 ꢁ 0.97s1H (ref. 14)
d13C = 175.7 ꢁ 0.963s13C (ref. 15)
(1)
(2)
40
Calcd
reflux in ethanol, reflux in xylene and fusion without solvent.1,2
The computed values corresponding to part I of this manu-
script are gathered in Table 3 and represented in Fig. 9.
Starting from the most stable tautomer, 1H-indazole (1a), to
reach 4ꢀHF the barrier has a height of 92.9 kJ molꢁ1 while to
reach 3ꢀHF the barrier is lower, 88.3 kJ molꢁ1. This is in
We have checked that the experimental chemical shifts of
1-(2,4-dinitrophenyl)-2H-indazole correspond to rotamer 40 and
not to 4.
Energies. Concerning the formation of 3 vs. 4 (actually 40) we
have reported in Fig. 1 that the isomer 4 is the most abundant,
this being true for the three experimental conditions used:
Table 1 Torsion angles (1, absolute values) calculated at the B3LYP/6-311++G(d,p) level
Molecule
Calculated
Experimental
3
N2–N1–C10–C20 = 41.0
C10–C20–N–O = 46.4
C30–C40–N–O = 0.6
N2–N1–C10–C20 = 35.5
C10–C20–N–O = 53.1
C30–C40–N–O = 4.5
4a
N1–N2–C10–C20 = 42.2
C10–C20–N–O = 48.6
C30–C40–N–O = 0.0
N1–N2–C10–C20 = 29.1
C10–C20–N–O = 63.1
C30–C40–N–O = 9.0
N1–N2–C10–C20 = 28.0
C10–C20–N–O = 66.0
C30–C40–N–O = 6.7
40
N1–N2–C10–C20 = 149.8
C10–C20–N–O = 47.1
C30–C40–N–O = 0.1
a
Two independent molecules in the unit cell.
c
New J. Chem.
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013

View Article Online
NJC
Paper
Table 3 Computed energy values for the formation of 1-(2,4-dinitrophenyl)-1H-
indazole (3) and 2-(2,4-dinitrophenyl)-2H-indazole (4) corresponding to
PCM(EtOH)/B3LYP/6-31G(d)+ZPE calculations (absolute values in hartrees, relative
the nitro group N(4) of one molecule, and the hydrogen (H4) of
the phenyl carbon (C4) of other molecule along the ab plane,
and is connected to an antiparallel chain by C–Hꢀ ꢀ ꢀO hydrogen
bonds between the oxygen atom O(1) of the formyl group of one
chain and the hydrogen (H11) of the phenyl carbon (C11) of the
other chain, forming a double chain. This interaction fixes the
unexpected disposition of the formyl group (Fig. 11–13).
Optimized geometries vs. experimental geometries. Neglecting
the conformation of the nitro group, compound 5 is characterized
by two torsion angles, the first one (y1) about the N2–C10 bond
(interring, N1–N2–C10–C60) and the second one (y2) about the
C20–C(O) bond (formyl group, C30–C20–C–O). Thus, four minima
are possible if all structures corresponding to two different angles
are stable. Actually, one of them is not a minimum (Fig. 14).
The three other structures have very similar energies with
differences below 2 kJ molꢁ1, the most stable being 5b
(ꢁ1020.16370 hartrees).
values in kJ molꢁ1
)
Erel (with regard
to 1aꢀ ꢀ ꢀ2)
Erel (with regard
to 1bꢀ ꢀ ꢀ2)
Entry
Total energy
1a + 2
1aꢀ ꢀ ꢀ2
TS1
INT1
TS2
4 + HF
4
1b + 2
1bꢀ ꢀ ꢀ2
TS3
INT2
TS4
3 + HF
3
ꢁ1120.11382
ꢁ1120.11834
ꢁ1120.08842
ꢁ1120.09189
ꢁ1120.07854
ꢁ1120.11488
ꢁ1019.68197
ꢁ1120.10692
ꢁ1120.11182
ꢁ1120.08263
ꢁ1120.08644
ꢁ1120.08025
ꢁ1120.12156
ꢁ1019.69031
11.7
0.0
78.7
69.5
104.6
9.2
30.1
17.2
93.7
83.7
100.0
ꢁ8.4
12.9
0.0
76.5
66.5
82.8
ꢁ25.6
contradiction with the data of Fig. 1 (the most abundant being
isomer 4) but the difference only amounts to 4.6 kJ molꢁ1
.
The reaction between 2 and saturated amines has activation
barriers in the 65–80 kJ molꢁ1 range,16 in reasonable agreement
with those we have calculated taking into account the differences
between indazole and aliphatic amines.
Part II: structure and the mechanism of formation of the
2-(1N-oxide-6-nitro-2H-benzotriazol-2-yl)benzaldehyde (5)
Crystallography. Crystal packing is formed by double chains
parallel to the ab plane (Fig. 10). One chain is formed by
hydrogen bonds C–Hꢀ ꢀ ꢀO between the oxygen atom O(4) of
Fig. 10 A view of 5 with displacement ellipsoids drawn at the 50% probability
level. The numbering does not follow the systematic one.
Fig. 9 Energy profile (kJ molꢁ1) of the formation of 3 and 4 from 1 and 2b (PCM: ethanol).
c
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013
New J. Chem.

View Article Online
Paper
NJC
in Table 4. The experimental chemical shifts are the geometrical
average of two or three conformations if one assumes the
simplification of considering only the minima. The test of
validity of the multiple regressions is not only the correlation
coefficient but, mainly, the fact that the sum of individual
populations should be close to 1 (or the percentages close to
100). Imposing the intercept to be 0, the following results are
obtained:
1
From H NMR chemical shifts: 23.7% 5a, 23.6% 5b, 54.0%
5c (sum: 101.3%). From 13C NMR chemical shifts: 33.2% 5a,
23.1% 5b, 43.8% 5c (sum: 100.1%).
Fig. 11 View of the chains along the ab plane of 5.
The results based on 13C chemical shifts should be more
reliable (greater range) but, in any case, there are about 25% of
each 5a and 5b (sE isomers) and 50% of 5c (sZ isomer).
If all the values of Tables 2 and 4 are considered together,
the 1H chemical shifts (24 values) are well correlated (R2 > 0.90)
and the 13C chemical shifts (39 values) still better (R2 > 0.98).
The mechanism of formation of 5. All our attempts to
characterize a reaction mechanism compatible with that
proposed in 1967–1968 met with no success. Formation of 5
was found to be symmetry forbidden and other possible (3+2)
reaction paths led to dead ends. We therefore investigated the
reaction mechanism outlined in Fig. 16 by computing the
reaction profile. The possible stationary points were located
and characterized at the B3LYP/6-31G* level of theory in
ethanol solution (vide infra). The relative energies of all the
stationary points and the shape of the corresponding transition
structures are reported in Fig. 17 (all values in kJ molꢁ1).
Formation of both 40 (a rotamer of 4) and 5 starts with
the stepwise aromatic nucleophilic substitution reaction of
1H-indazole 1 with 1-fluoro-2-nitrobenzene 2. Despite the partial
stabilization of saddle point TS1 by a weak N–Oꢀ ꢀ ꢀH hydrogen
bond (see Fig. S1 in the ESI‡), the computed activation energy for
the addition step is the largest one along the reaction profile
leading to 4, as it is expected for this kind of reactions. The
substitution product 4 is obtained by syn-elimination of HF from
the unstable addition intermediate INT1, recovery of aromaticity
of the nitrophenyl moiety being the driving force for this step. It
is noteworthy that 4 lies 9.2 kJ molꢁ1 above the complex 1aꢀ ꢀ ꢀ2
(Fig. 16).
Fig. 12 View of 5 along the c axis. Chains parallel to the ab plane.
The sequence of stationary points leading to 5 involves the
addition of oxygen Od of the nitro-nitronate group on the Ce
atom of 1 via transition structure TS5. As it can be seen
by inspection of the geometric features of TS5, at this step
the Na–Cb bond is already formed, and the whole structure is
Fig. 13 Packing diagram of 5.
The experimental X-ray geometry corresponds to 5c. The stabili- also partially stabilized by an intramolecular hydrogen bond
zation of these structures is of different origin. For 5a and 5b, they (see Fig. S1 in the ESI‡). The activation energy associated with
are due to hydrogen bonds (HBs) but for 5c there is a single HB, the this step is also noticeable and corresponds to a stepwise (4+2)
other stabilizing force is an orthogonal interaction between the O cycloaddition between the CeQNa moiety of 1 and 2. It is
of the carbonyl group and the N2 nitrogen of the benzotriazole interesting to note that (4+2) cycloaddition (hetero Diels–Alder)
N-oxide,13 which bears a partial positive charge (note that the chemistry involving nitroalkenes in the presence of Lewis acids
corresponding resonance form is benzenoid while the more com- has been described by Denmark et al.17 and, more recently and
mon representation is quinonoid). This orthogonal interaction is within a different context, by Danishefsky and Houk et al.18
clearly apparent in the AIM analysis (Fig. 15).
The (4+2) hetero Diels–Alder cycloadduct INT3 can eliminate
NMR spectroscopy and GIAO/DFT calculations. The chemical one equivalent of HF to yield the aromatized intermediate INT4 via
shifts, experimental and calculated, for compound 5 are reported TS6 in a similar way to that computed for INT1 (Fig. 16 and 17).
c
New J. Chem.
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013

View Article Online
NJC
Paper
Fig. 14 The calculated and experimental conformations of benzotriazole-N-oxide 5 (in parentheses, relative energies).
Table 4 Experimental chemical shifts [d, ppm, 1H (400 MHz) and 13C (100 MHz)
in CDCl3] of compound 5 and calculated chemical shifts (d, ppm) for three
minimum conformations of compound 5
0
0
0
0
Comp.
H4
H5
H7
H3
H4
H5
H6
CHO
1H NMR
5
Exp.
7.94 8.31 8.85 8.14 7.90 7.90 7.74
9.93
9.90
9.65
5a Calcd 7.68 8.36 8.86 8.44 7.65 7.66 7.91
5b Calcd 7.70 8.27 8.84 8.44 7.72 7.69 7.42
Fig. 15 AIM molecular graphs of compound 5c optimized at the B3LYP/6-
311++G(d,p) computational level. The atoms are labeled O, N, C and H. The bond
critical points are pale grey and the ring critical points are dark grey. The
following colors will appear in the electronic version: red: oxygen; blue: nitrogen;
gray: carbon; white: hydrogen. Bond critical points: green; ring critical points: red.
5c
Calcd 7.64 8.25 8.78 7.81 7.83 7.73 7.45 10.05
Comp.
C3a
C4
C5
C6
C7
C7a
CHO
13C NMR
5
Exp.
142.6 120.9 128.2 145.7 113.1 124.4 187.5
5a Calcd 142.6 119.7 124.7 145.6 114.6 125.5 184.1
5b Calcd 142.8 120.0 124.6 145.6 114.5 124.8 182.7
Unstable zwitterionic species INT4 yields (E)-2-((2-nitroso-phenyl)-
diazenyl) benzaldehyde conformers 8 and 80 via concerted cleavage
of Na–Ce and Od–Cf bonds. Surprisingly, the activation energy
associated with this retro-(4+2) cycloaddition is quite low, the
reaction energy leading to benzaldehyde derivative 8 being quite
exothermic (Fig. 16). From this latter local minimum, cyclization to
5 takes place smoothly via a kinetically favored [5-exo-trig] process19
whose associated transition structure is TS9 (Fig. 15 and 16).
5c
Calcd 142.7 119.8 123.8 145.2 114.9 125.6 185.3
0
0
0
0
0
0
C6
C1
C2
C3
C4
C5
5
Exp.
132.6
137.4
136.8
133.6
131.9
131.1
131.8
133.4
131.6
129.9
128.8
135.0
132.5
131.0
131.5
131.3
134.5
132.8
133.5
133.6
123.7
127.3
127.3
129.4
5a
5b
5c
Calcd
Calcd
Calcd
Interestingly, the reaction energy (DErxn) associated with the computational results, we dare to predict that microwave
1aꢀ ꢀ ꢀ2 - 5 reaction is computed to be ꢁ30.0 kJ molꢁ1, whereas irradiation20 should increase the yield of 5 and its analogues,
for the 1aꢀ ꢀ ꢀ2 - 4 transformation it is found that DErxn
=
the calculated barrier to reach 5 from 1aꢀ ꢀ ꢀ2 is 231.4 kJ molꢁ1
,
+9.2 kJ molꢁ1. Therefore, thermodynamic control favors for- too high to be involved in the formation of the benzotriazole-
mation of 5 as highly polar zwitterionic intermediates are N-oxide and the mechanism depicted in Fig. 17 must be
involved in its formation. Although on the basis of our rejected.
c
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013
New J. Chem.

View Article Online
Paper
NJC
Fig. 16 Mechanism connecting 1H-indazole (1a) and 1-fluoro-2,4-dinitrobenzene (2) to the final products 4 and 5.
Fig. 17 Reaction profile for the transformation of 1a and 2 into either 4 or 5. All the relative energies have been computed in ethanol at the B3LYP(PCM)/6-31G*+
DZPVE level and are given in kJ molꢁ1. The part concerning the formation of 40 is common with Fig. 9.
We think that 5 could be obtained by thermal isomerization
The 1967 reaction1 was carried out using ordinary glass (not
of 4 but all our attempts (see Experimental section) failed; then, Pyrex) and compound 5 was isolated pure by fractional crystal-
since one mole of HF was formed during the reaction (see Fig. 9 lization in ethanol. We repeated the reaction (10 min reflux in
and 16), we carried out experiments under conditions as ethanol) using actinic glassware: we obtained 30% of 3 and
energetic as MW irradiation of a trifluoroacetic solution of 4 70% of 4 (by HPLC) without traces of 5. The proportions are
containing some drops of H2SO4 100%. In all cases, compound similar to our previous ones taking into account that the
4 was recovered unchanged. The only remaining possibility21 amount of 5 should be added to that of 4. At 300 K, the 30/70
was that 5 was formed by a photochemical process.
ratio corresponds to a difference of 2.1 kJ molꢁ1
.
c
New J. Chem.
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013

View Article Online
NJC
Paper
Finally, we irradiated a sample of 4 in ethanol and obtained reaction occurs via the singlet.23 The formation of oxygen–
5 in quantitative yield; the reaction was the same in a degassed carbon bonds upon irradiation has been reported previously
solution (Ar bubbled through it) and in solution where oxygen but on very different compounds: 1-(2,4-dinitrophenyl) and
(air) as a quencher22 had been bubbled, thus proving that the 1-(2,4,6-trinitrophenyl)aziridines,24 and 1-aryl-2-nitroalkenes
(photochromism).25
In order to investigate the photochemical conversion of 4 to
5 on the S0 and S1 potential energy hypersurfaces associated
with the singlet ground state and the first excited singlet state,
respectively, we decided to explore the behavior of a computa-
tionally simpler system. The critical step on the S0 ground state
reaction coordinate was the formation of INT4, whose com-
bined activation energies from 4 were calculated to be prohibi-
tively high. Since in the photochemical experiments the neutral
species 4 was used as starting material, we explored this step
using the 9 - 10 transformation as a model system (Fig. 18).
A standard DFT study (Fig. 19A) on the thermal 9 - 10
reaction at the B3LYP/6-31G* level of theory led to the char-
acterization of transition structure 12TS(S0) on the ground state
S0 hypersurface, with an activation energy of 192 kJ molꢁ1, in
line with the high barriers previously computed for this kind
of reaction. Time-dependent DFT (TDDFT) calculations per-
formed with the same hybrid functional and the basis set led to
a vertical transition of type np*. This electronic transition
involves a double excitation from the HOMO ꢁ 2 Kohn–Sham
Fig. 18 Model system used to explore the photochemical formation of six-
membered intermediate INT4. The species included in the experimental studies
are indicated in grey.
Fig. 19 Computational profiles associated with the 9 - 10 reaction. Bond distances are given in Å. Relative energies are reported in kJ molꢁ1. Grey arrows connect
fully optimized structures. Dashed arrows indicate vertical transitions. (A) B3LYP/6-61G* and time-dependent DFT (TDDFT) B3LYP/6-31G* profiles. KSi - KSj indicates
the Kohn–Sham (KS) orbitals involved in the corresponding transition. 12TS(S0) is the transition structure (TS) that connects local minima 9 and 10 in the S0 ground
state. (B) CASSCF(4,4)-MP2/6-31G*//CASSCF(4,4)/6-31G* profile for the same process. 12CI(S1–S0) is the conical intersection (CI) that connects 9 on the S1 hypersurface
with 10 in the S0 ground state.
c
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013
New J. Chem.

View Article Online
Paper
NJC
(KS) orbital, mainly formed by lone pairs of the nitro and near the N atom (3) while 2-(2,4-dinitrophenyl)-2H-indazole
pyrazole groups, to the LUMO and LUMO + 1, both of them has the nitro group near the CH (conformation 40).
involving delocalized p-systems. Despite its low absorptive
– The mechanism of formation of both N-substituted
capacity, at this point the S1 state has a vertical excitation indazoles is a classical one and the calculated transition states
energy of 341 nm (equivalent to 350.6 kJ molꢁ1), a transition have reasonable values.
therefore quite close to the visible region (380–740 nm), within
– The only found thermal mechanism of formation of
the accuracy limits of TDDFT.26 It is expected that in more 2-(1N-oxide-6-nitro-2H-benzotriazol-2-yl)benzaldehyde differs consid-
conjugated systems such as 4 the wavelength associated with erably from that proposed in 1967–1968. It involves five TSs and five
this kind of transition lies in the visible region. Interestingly, intermediates that have been theoretically characterized. However, it
the S0–S1 vertical transition of 12TS was calculated to be of only is too high in energy (231 kJ molꢁ1) to be operative.
ca. 121 kJ molꢁ1 and is associated with a symmetry allowed pp*
– The 4 to 5 transformation is a photoisomerization going
HOMO–LUMO transition (Fig. 19A). As a consequence, 12TS in through the singlet state. This is the first reported example of
the S1 state lies ca. 33 kJ molꢁ1 below 9 and ca. 84 kJ molꢁ1 such a reaction.
below 10. This situation suggested that there should be a
– A mechanism has been theoretically calculated for the
conical intersection (CI) connecting both S0 and S1 potential singlet state photoisomerization.
energy hypersurfaces and therefore stationary points 9 and 10.
We carried out CASSCF(4,4)-MP2/6-31G*//CASSCF(4,4)/
Experimental section
Chemistry
6-31G* calculation27 (Fig. 19B) in order to locate this hypothe-
tical conical intersection. These multiconfigurational SCF (MC-
SCF) calculations were performed by checking the Kohn–Sham
orbitals involved in the S0–S1 transitions and the corresponding
canonical MOs. According to the TDDFT results, the starting
guess was modified in order to include MO34 (which is similar
to KS34) in the active space of 9. In the case of 12 and 10,
the simpler HOMO–LUMO transition found in the TDDFT
calculations resulted in rapidly convergent CASSCF solutions.
Actually, for these latter stationary points the CASSCF(2,2) and
CASSCF(4,4) results were very similar. Under these conditions
we readily located conical intersection 12CI at the CASSCF(4,4)/
6-31G* level by starting the optimization from the B3LYP/
6-31G* geometry and wave function of 12TS (Fig. 19A). As
expected, the chief geometric features of 12CI were found to
be very similar to those computed for 12TS, the critical C–O
bond distance being a little bit shorter, thus indicating a more
advanced structure towards intermediate 10.
General methods. All reagents were purchased from com-
mercial suppliers and used without further purification. Flash
chromatography was performed on silica gel (230–400 mesh).
Melting points are uncorrected. Microwave reactions were
carried out using a Biotage Initiatort 2,0. Most 1H and 13C
NMR spectra were obtained at 300 MHz (1H; 13C 75 MHz) at
298 K and one at 400 MHz for the 1H NMR spectrum of 5.
Abbreviations for multiplicity are as follows: d indicates doub-
let, t indicates triplet, m indicates multiplet, bs indicates broad
singlet, bd indicates broad doublet, dd indicates double doublet, dt
indicates double triplet. Chemical shifts are reported in ppm
referenced to TMS for CDCl3 and to the internal solvent residual
of DMSO-d6 at 2.50 ppm for 1H NMR and at 39.5 ppm for 13C NMR.
Assignment of the 1H and 13C NMR signals was based on the 1H–1H
spin–spin coupling constants as well as on COSY, HMBC and HSQC
methods. LC-MS analyses were performed using an Alliance 2695
(Waters) with a diode array UV/Vis detector Waters 2996 and
interfaced to a Micromass ZQ mass spectrometer; analyses were
performed using reversed phase HPLC silica based columns: col-
umn bridge C18 3.5 mm. (2.1 ꢂ 100 mm). Using an injection volume
of 3 mL, a flow rate of 0.25 mL minꢁ1 and gradient elution (5 to 95%
over 5 min) of acetonitrile in water. Acetonitrile contains 0.08% v/v
and water contains 0.1% v/v formic acid. Analyses were monitored
at 254 nm wavelength.
From these TDDFT and MC-SCF computational results, we
conclude that the reactions indicated in Fig. 17 can be readily
performed under UV-visible irradiation. The photochemical
reaction path starts with a relatively low energy S0–S1 excitation.
From this np* S1 excited state the system evolves to a pp* state
and relaxes to the S0 state via a conical intersection very similar
to the transition structure that connects reactants and products
under purely thermal conditions. Beyond this latter stationary
point, the rest of the elementary steps leading to 5 can occur on
the S0 potential energy hypersurface.
The photoreactions were performed in quartz cells. The
solutions were irradiated through a dichroic mirror (Dichroic
Beam Turning Mirror, Oriel) to cut off UV light (transmission
range 260–320 nm) by using a mercury (Xe) lamp (Oriel Research
Arc Lamp Source, 400 W). A focusing lens (Fiber Bundle Focus-
ing Assembly, Fused Silica Aspheric, F/2.2, 4 mm Typical Spot
Size) was also used. The irradiation was controlled using an
electronic shutter (Electronic Shutter Driver, Oriel), a light
intensity controller and a digital timer (Oriel, RS-232).
Conclusions
The most significant conclusions obtained from the results
reported in this paper are:
– The X-ray structures of 3 and 4 have been determined
both having the nitro group near the N atom of the indazole
ring. The structure of 5 presents an interesting example of
orthogonal interaction.
Purification of the three compounds previously obtained1. Using a
–
1H and specially 13C NMR chemical shifts show that in flash chromatography column (eluent 5 : 1 dichloromethane–hexane
DMSO 1-(2,4-dinitrophenyl)-2H-indazole has the nitro group - 9 : 1 dichloromethane–methanol), four fractions were isolated: 3
c
New J. Chem.
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013

View Article Online
NJC
Paper
was obtained pure in the first fraction, 4 was recrystallized from The solvent was evaporated at reduced pressure yielding a light
ethanol from the second fraction, and 5 was recrystallized from yellow solid. LC-MS retention time 4.88 min [M + H]+ = 285. Acid
acetonitrile from the third fraction. Fourth fraction was a mixture of conditions: to a solution of 2-(2,4-dinitrophenyl)-2H-indazole (6 mg,
5 and an unknown compound, which was identified as the hemi- 0.02 mmol) in trifluoroacetic acid (10 mL) two drops of H2SO4 100%
acetal of 5 formed by reaction with traces of ethanol (see ESI‡).
1-(2,4-Dinitrophenyl)-1H-indazole (3): yellow crystals. Mp
were added. The reaction mixture was refluxed for 22 h. The solvent
was evaporated at reduced pressure yielding light yellow solid.
=
172–173 1C. Lit. m.p. 172 1C.1 1H-NMR (300 MHz, DMSO-d6) d LC-MS retention time 4.86 min [M + H]+ = 285; 2-(2,4-dinitrophe-
0
0
ppm: 8.92 (H3 Ar, d, J = 2.5 Hz, 1H), 8.65 (H5 Ar, dd, J = 8.9 Hz, nyl)-2H-indazole (6 mg, 0.02 mmol), H2SO4 100% (two drops) in TFA
0
J = 2.5 Hz, 1H), 8.52 (H3 indazole, bd, J = 0.9 Hz, 1H), 8.29 (H6 (10 mL) were heated at 100 1C for 7 h in a microwave reactor. The
Ar, d, J = 8.9 Hz, 1H), 7.95 (H4 indazole, dd, J = 8.1 Hz, J = 0.9 Hz, mixture was evaporated at reduced pressure yielding a light yellow
1H), 7.82 (H7 indazole, d, J = 8.1 Hz, 1H), 7.61 (H6 indazole, dt, solid. LC-MS retention time 4.84 min [M + H]+ = 285. Compound 5
J = 8.1 Hz, J = 0.9 Hz, 1H), 7.38 (H5 indazole, dt, J = 8.1 Hz, J = has a LC-MS retention time of 4.12 min [M + H]+ = 285.
0.9 Hz, 1H). 13C-NMR (75 MHz, DMSO-d6) d ppm: 145.2 C4 , 143.5
0
0
0
0
0
C1 , 139.0 C3, 138.6 C2 , 136.1 C7a, 128.7 C5 , 128.5 C6, 126.2 C6 ,
Preparation of compounds 3 and 4. A solution of indazole
0
125.3 C3a, 123.2 C5, 122.1 C4, 121.7 C3 , 110.2 C7. LC/MS: (500 mg, 4.2 mmol, 1 equiv.) and 1-fluoro-2,4-dinitrobenzene
positive ESI-MS: m/z = 285.20 [M + H]+ (100%); HPLC: tR
=
(530 mL, 4.2 mmol, 1 equiv.) in EtOH (15 mL) was heated at
5.06 min (99%); eluent: gradient from 10% MeCN + 0.08% reflux for 19 h in an actinic amber round-bottom flask pro-
formic acid and 90% H2O + 0.1% formic acid to 100% MeCN + tected from light. Afterwards, the yellow solid formed was
0.08% formic acid; flow = 1 mL minꢁ1; l = 256 nm. Proper filtered and washed with cold EtOH several times. Solvent
crystals of 3 for X-ray diffraction were obtained by recrystalliza- was removed from the liquid layer, obtaining an oily brown
tion from EtOH with some drops of hexane.
residue. The solid and the oily residue were dried under
vacuum. Solid (832 mg, 69% yield) was identified by TLC and
2-(2,4-Dinitrophenyl)-2H-indazole (4): yellow solid. Mp
=
152–154 1C. Lit. m.p. 158 1C.1 1H-NMR (400 MHz, DMSO-d6) d LC/MS as a mixture (68 : 32) of 2-(2,4-dinitrophenyl)-2H-inda-
0
ppm: 9.08 (H3 indazole, bs, 1H), 8.93 (H3 Ar, d, J = 2.4 Hz, 1H), zole and 1-(2,4-dinitrophenyl)-1H-indazole, respectively. In the
0
0
8.69 (H5 Ar, dd, J = 8.7 Hz, J = 2.4 Hz, 1H), 8.29 (H6 Ar, d, J = oily residue were found unreacted indazole and two subpro-
8.7 Hz, 1H), 7.81 (H4 indazole, d, J = 8.6 Hz, 1H), 7.64 ducts 2,4-dinitro-1-ethoxybenzene and an unknown one. No
(H7 indazole, d, J = 8.6 Hz, 1H), 7.35 (H6 indazole, t, J = traces of 5 were found (1H NMR and LC/MS).
7.2 Hz, 1H), 7.15 (H5 indazole, t, J = 7.2 Hz, 1H). 13C-NMR
0
0
(75 MHz, DMSO-d6) d ppm: 150.8 C7a, 147.2 C4 , 144.6 C2 , 137.3
Photoisomerization of 2-(2,4-dinitrophenyl)-2H-indazole (4). A
C1 , 128.8. C6, 128.7 C6 , 128.5 C3, C5 , 126.2 C6 , 123.7 C3 , 123.2 solution of 2-(2,4-dinitrophenyl)-2H-indazole (10ꢁ4 M) in EtOH
C3a, 122.0 C5, 121.8 C4, 118.2 C7. LC/MS: positive ESI-MS: m/z = was placed in a quartz cell and bubbled with Ar for 40 min
285.20 [M + H]+ (100%); HPLC: tR = 4.90 min (90%); eluent: before irradiation. Then, the solution was irradiated from 260
gradient from 10% MeCN + 0.08% formic acid and 90% to 320 nm for 2 h using a mercury (Xe) lamp (400 W). An
H2O + 0.1% formic acid to 100% MeCN + 0.08% formic acid; analogous experiment was performed by the irradiation of the
flow = 1 mL minꢁ1; l = 256 nm. Proper crystals of 4 for X-ray solution, this time, in the presence of air. The reactions were
diffraction were obtained by recrystallization from dichloro- monitored by HPLC-MS.
0
0
0
0
0
methane with some drops of hexane.
Crystallographic study. The structures display no noticeable
2-(1N-Oxide-6-nitro-2H-benzotriazol-2-yl)-benzaldehyde (5): modifications in crystal packing at different temperatures.
light yellow powder. Mp = 212–214 1C. Lit. m.p. 206 1C.1 1H- Compounds 3 and 4 are formed by chains parallel to the
NMR (400 MHz, CDCl3) d ppm: 9.93 (CHO, s, 1H); 8.85 (H7, d, bc plane while compound 5 is formed by chains parallel to
0
J = 2.0, 1H), 8.31 (H5, q, J = 9.4, J = 2.0 Hz, 1H), 8.14 (H3 , q, J = the ab plane. Molecules forming the layers are linked by
0
0
6.6, J = 2.3, 1H), 7.94 (H4, d, J = 9.4 Hz, 1H), 7.9 (H4 and H5 , H-bonds. There are weak intramolecular interactions. For com-
septuplet), 7.74 (H6 , q, J = 7.0, J = 2.0 Hz, 1H). 13C-NMR pound 4, the structure at room temperature does not present
0
(75 MHz, DMSO-d6) d ppm: 190.2 CHO, 145.2 C6, 142.0 C3a, changes with that derived from low temperature data, except
0
0
0
0
0
135.0 C5 , 132.7 C4 , 132.4 C1 , 131.2 C2 , 123.4 C6 , 128.7 C5, for the higher structure expansion due to the higher atomic
0
0
124.1 C7a, 123.4 C6 , 131.0 C3 , 121.1 C4, 112.6 C7. LC/MS: positive displacements.
ESI-MS: m/z = 284.94 [M + H]+ (100%); HPLC: tR = 4.28 min
Details of crystal structure determinations including crystal
(99%); eluent: gradient from 10% MeCN + 0.08% formic acid data, data collection, structure determination and refinement
and 90% H2O + 0.1% formic acid to 100% MeCN + 0.08% parameters are shown in Table 5. X-ray diffraction measure-
formic acid; flow = 1 mL minꢁ1; l = 256 nm. Proper crystals of ments were performed at room temperature for 4 (298.0 K) and
5 for X-ray diffraction were obtained by recrystallization from 5 (293 K) and at low temperature (120 K) for compounds 3 and
EtOH–water 5 : 1 with some drops of hexane.
4, with mirror-monochromated CuKa radiation (l = 1.5418 Å).
Images were collected at 55 mm fixed crystal-detector distance,
Thermal isomerization experiments of 2-(2,4-dinitrophenyl)-2H-inda- using the oscillation method with variable exposure time per
zole (4). Neutral conditions: a solution of 2-(2,4-dinitrophenyl)-2H- image. The structures were solved using direct methods and
indazole (6 mg, 0.02 mmol) in xylene (10 mL) was refluxed for 22 h. subsequent Fourier differences using in all cases the WingX
c
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013
New J. Chem.

View Article Online
Paper
NJC
Table 5 Crystal data and structure refinement for compounds 3, 4 and 5
Compound
3 (120.0 K)
4 (120.0 K)
4 (298.0 K)
5 (293.0 K)
Formula
Formula weight/g
molꢁ1
Wavelength
Crystal system
Space group
a (Å)
b (Å)
c (Å)
a (1)
C13H8N4O4
284.23
C13H8N4O4
284.23
C13H8N4O4
284.23
C13H8N4O4
284.23
CuKa (1.5418 Å)
Orthorhombic
P212121
7.0459(2)
11.0462(3)
15.4551(4)
90.00
CuKa (1.5418 Å)
Monoclinic
P21/c
7.0730(5)
24.0840(16)
14.3289(10)
90.00
CuKa (1.5418 Å)
Monoclinic
P21/c
7.2107(6)
24.0635(17)
14.4077(11)
90.00
CuKa (1.5418 Å)
Monoclinic
P21/c
13.0325(6)
15.3958(15)
6.1563(3)
90
b (1)
90.00
90.00
1203.1(2)
4
100.633(7)
90.00
2399.0(3)
8
100.334(17)
90.00
2459.4(3)
8
99.558(5)
90
1218.09(14)
4
g (1)
Cell volume/Å3
Z
Calc. density/mg mꢁ3 1.569
1.574
1.535
1.550
Absorption
1.025
1.028
1.003
1.012
coefficient/mmꢁ1
F(000)
584.0
1168.0
0.25 ꢂ 0.05 ꢂ 0.01
1168.0
0.28 ꢂ 0.02 ꢂ 0.01
584.0
0.20 ꢂ 0.10 ꢂ 0.03
Crystal size (mm3)
Index ranges
0.14 ꢂ 0.08 ꢂ 0.06
ꢁ8 r h r 8, ꢁ11 r k r 13, ꢁ7 r h r 8, ꢁ29 r k r 28, ꢁ8 r h r 7, ꢁ23 r k r 29, ꢁ15 r h r 14, ꢁ18 r k r 18,
ꢁ15 r l r 18
ꢁ15 r l r 17
7.28 to 141.341
ꢁ17 r l r 12
ꢁ7 r l r 7
2Y range for data
collection (1)
9.84 to 140.82
7.24 to 141.42
6.88 to 141.62
Reflections collected 4652
16197
4554 [Rint = 0.0795]
8843
13 222
2317 [Rint = 0.0506]
Independent
reflections
2258 [Rint = 0.0311]
4556 [Rint = 0.0663]
Completeness
Absorption
0.998 (y = 68.001)
Analytical
1.000 (y = 68.001)
Analytical
0.986 (y = 68.001)
Analytical
0.989 (y = 70.811)
Empirical
correction
Max. and min.
transmission
1.00000 and 0.98138
0.987 and 0.873
1.0957 and 0.6170
1.00000 and 0.90640
Refinement method Full-matrix least-squares on Full-matrix least-squares on Full-matrix least-squares on Full-matrix least-squares on
F2
F2
F2
F2
Data/restraints/
parameters
2258/0/190
4554/0/379
4556/0/379
2317/0/190
Goodness-of-fit on F2 1.073
0.984
0.998
1.048
Final R indices
R1 = 0.0335, wR2 = 0.0768
R1 = 0.0469, wR2 = 0.0934
R1 = 0.0599, wR2 = 0.1241
R1 = 0.0439, wR2 = 0.1028
[I > 2s(I)]
R indices (all data) R1 = 0.0383, wR2 = 0.0796
Largest diff. peak
and hole
R1 = 0.0822, wR2 = 0.1085
0.21 and ꢁ0.20 e Åꢁ3
R1 = 0.1325, wR2 = 0.1709
0.19 and ꢁ0.29 e Åꢁ3
R1 = 0.0676, wR2 = 0.1171
0.19 and ꢁ0.25 e Åꢁ3
0.14 and ꢁ0.22 e Åꢁ3
Flack parameter
0.3(2)
package and refined by full-matrix least-squares on F2. Empiri- three-parameter functional customarily denoted B3LYP.35 The
cal absorption correction was applied using the program standard 6-31G(d) basis sets36 as implemented in the GAUSSIAN
XABS2. All non-H atoms were anisotropically refined. Hydrogen 0937 suite of programs have been used in all cases. Solvent
atoms attached to carbon atoms were positioned geometrically effects were taken into account by optimizing geometries and
with Uiso values derived from Ueq values of the corresponding computing energies using the PCM method with ethanol para-
carbon atom.
meters.38 All stationary points were characterized by harmonic
Crystallographic calculations were carried out using the analysis.39 All transition structures exhibited one and only one
following programs: CrysAlis CCD and RED28 for data collec- imaginary frequency associated with nuclear motion along the
tion, cell refinement, data reduction and empirical absorption reaction coordinate under study. For calculation of absolute
correction; SIR 200429 for structure solution; SHELXL-9730 for shieldings, the B3LYP/6-31G(d) optimized geometries were
structure refinement and preparation of materials for publica- further optimized at the B3LYP/6-311++G(d,p) level.40 Absolute
tion; PLATON31 for the geometrical calculations; ORTEP32 and shieldings were calculated within the GIAO approximation.41
Mercury33 for molecular graphics. Crystal data and structure
The reactivity of the model systems in the excited state has
refinement details for the three structures are given in Table 5. been carried out using the time-dependent DFT (TDDFT)
Selected bond lengths (Å) and bond angles (1), and hydrogen B3LYP/6-31G(d), CASSCF(4,4)/6-31G(d) and CASSCF(4,4)-MP2/
bonds are listed in Table 6 and 7 for both compounds.
CCDC 806879, 806880, 806881 and 923952, respectively.‡
All the calculations reported in this paper have been per-
6-31G(d)//CASSCF(4,4)/6-31G(d) computational models within
the GAUSSIAN 09 program.
The bonding characteristics were analyzed by means of the
formed within Density Functional Theory,34 using the hybrid atoms in molecules (AIM) theory.42,43 For this purpose we have
c
New J. Chem.
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013

View Article Online
NJC
Paper
Table 6 Selected bond lengths (Å) and bond angles (1) for 3, 4 and 5
Table 7 Hydrogen bonds for compounds 3, 4 and 5
Bond lengths (Å)
Bond angles (1)
D–Hꢀ ꢀ ꢀA (Å)
d(D–H)/Å d(Hꢀ ꢀ ꢀA)/Å d(Dꢀ ꢀ ꢀA)/Å (D–Hꢀ ꢀ ꢀA)/1
3
3
C6–H6ꢀ ꢀ ꢀO4a
C12–H12ꢀ ꢀ ꢀO3a
C13–H13ꢀ ꢀ ꢀO4a
0.95
0.95
0.95
2.53
2.51
2.52
3.237(3)
3.262(3)
3.328(3)
131
137
142
O3–N4
O1–N3
O4–N4
N2–N1
N1–C7
N1–C8
N3–C9
N2–C1
N4–C11
1.224(2)
1.222(2)
1.225(2)
1.381(2)
1.386(2)
1.395(2)
1.472(2)
1.310(3)
1.465(2)
N2–N1–C7
N2–N1–C8
O1–N3–O2
C1–N2–N1
O3–N4–O4
C10–C11–N4
C12–C11–N4
N1–C8–C13
N1–C8–C9
C10–C9–N3
C8–C9–N3
111.2(1)
118.7(2)
124.6(2)
105.5(2)
124.1(2)
118.3(2)
119.3(2)
120.3(2)
121.8(2)
115.8(2)
121.7(2)
4 (120 K)
C2B–H2Bꢀ ꢀ ꢀO4Bb
C7B–H7Bꢀ ꢀ ꢀO2Bc
C9B–H9Bꢀ ꢀ ꢀO1Ad
C9B–H9Bꢀ ꢀ ꢀO2Bc
0.95
0.95
0.95
0.95
2.46
2.48
2.46
2.51
3.2663(5)
3.394(3)
3.213(3)
3.283(3)
143
161
135
139
4 (298 K)
C2B–H2Bꢀ ꢀ ꢀO4Bb
C7B–H7Bꢀ ꢀ ꢀO2Bc
C9B–H9Bꢀ ꢀ ꢀO1Ad
C9B–H9Bꢀ ꢀ ꢀO2Bc
0.93
0.93
0.93
0.93
2.52
2.55
2.50
2.56
3.319(6)
3.440(5)
3.231(6)
3.319(6)
145
161
136
139
4 298 K
O2A–N3A
O2B–N3B
O3A–N4A
O4B–N4B
O3B–N4B
O1A–N3A
O1B–N3B
N1B–C1B
N1B–N2B
N2A–C7A
N2A–N1A
N2A–C8A
O4A–N4A
N1A–C1A
N3B–C13B
N2B–C7B
N2B–C8B
N4B–C11B
N4A–O3A
N4A–C11A
N3A–C13A
1.219(4)
1.221(5)
1.212(5)
1.229(5)
1.212(5)
1.216(4)
1.211(4)
1.341(5)
1.373(4)
1.356(5)
1.367(4)
1.409(5)
1.220(5)
1.351(5)
1.469(5)
1.349(5)
1.410(5)
1.470(6)
1.212(5)
1.459(6)
1.482(5)
C1B–N1B–N2B
C7A–N2A–N1A
N1A–N2A–C8A
C1A–N1A–N2A
O1B–N3B–O2B
C7B–N2B–N1B
O3B–N4B–O4B
O3A–N4A–O4A
O1A–N3A–O2A
103.4(3)
113.5(3)
119.2(3)
103.4(3)
125.2(3)
113.2(3)
123.8(4)
123.7(4)
125.5(4)
5
C(4)–H(4)ꢀ ꢀ ꢀO(4)e
0.93
2.57
2.57
3.251(3)
3.396(3)
131
148
C(11)–H(11)ꢀ ꢀ ꢀO(1)f 0.93
Symmetry codes:a 1 ꢁ x, ꢁ1/2 + y, ꢁ1/2 ꢁ z. x, 1/2 ꢁ y, ꢁ1/2 + z. x, 1/
b
c
d
e
f
2 ꢁ y, 1/2 + z. x, y, z. 1 + x, y, 1 + z. 1 ꢁ x, ꢁy, ꢁz.
located the most relevant bond critical points (BCP), and
evaluated the electron density at each of them, with the facil-
ities of AIMAll programs.44 All the interactions were character-
ized by the presence of a BCP between the atoms involved that
are connected by the corresponding bond paths.
Acknowledgements
4 120 K
This work was supported by the Spanish Ministerio de Econo-
O2A–N3A
O3A–N4A
O4B–N4B
O3B–N4B
O1A–N3A
O1B–N3B
N1B–C1B
N1B–N2B
N2A–C7A
N2A–N1A
N2A–C8A
O4A–N4A
N1A–C1A
N3B–C13B
N2B–C7B
N2B–C8B
N4B–C11B
N4A–C11A
N3A–C13A
1.230(3)
1.222(3)
1.227(3)
1.225(3)
1.214(3)
1.218(3)
1.347(3)
1.370(3)
1.362(3)
1.369(3)
1.409(3)
1.228(3)
1.350(3)
1.477(3)
1.353(3)
1.416(3)
1.461(3)
1.468(3)
1.482(3)
C1B–N1B–N2B
C7A–N2A–N1A
N1A–N2A–C8A
C1A–N1A–N2A
O1B–N3B–O2B
C7B–N2B–N1B
O3B–N4B–O4B
O3A–N4A–O4A
O1A–N3A–O2A
103.0(2)
114.0(2)
119.4(2)
103.2(2)
126.0(2)
114.0(2)
124.0(2)
124.5(2)
125.2(2)
´
mıa y Competitividad (MAT2010-15094, Factoria de Cristaliza-
cion – Consolider Ingenio 2010) and FEDER funding. We also
´
thank the Ministerio de Ciencia e Innovacion (Project No. CTQ
´
2009-13129-C02-02) and the Comunidad Autonoma de Madrid
(Project MADRISOLAR2, ref. S2009/PPQ-1533) for continuing
support. Thanks are given to the CTI (CSIC) for an allocation of
computer time. F. P. C. thanks the Spanish MICINN (Grants
CTQ2010-16959 and Ingenio-Consolider CSD2007-00006) and
the Basque Government (GV-EJ, Grant IT-324-07) for financial
support. Technical and human support provided by SGIker
(UPV/EHU, MICINN, GV/EJ, ESF) is gratefully acknowledged.
For the kinetic data discussion and the photochemical experi-
´
´
ments and advice we warmly thank Drs Jose Ignacio Garcıa
˜
´
(Universidad de Zaragoza), Ulises Acuna (Instituto de Quımica-
´
´ ´
Fısica ‘‘Rocasolano’’, CSIC), Fernando Catalina and Levı Lopez
5
´
´
Vilanova (Instituto de Ciencia y Tecnologıa de Polımeros, CSIC).
N4–O3
N4–O4
N4–C10
N2–N1
N2–C8
N2–O2
N1–C7
N1–N3
C2–C1
O1–C1
1.217(3)
1.227(3)
1.469(3)
1.361(2)
1.360(2)
1.262(2)
1.421(2)
1.339(2)
1.474(3)
1.202(3)
O3–N4–O4
O3–N4–C10
O4–N4–C10
C8–N2–N1
O2–N2–N1
O2–N2–C8
N2–N1–C7
N3–N1–N2
N3–N1–C7
O1–C1–C2
123.7(2)
118.6(2)
117.7(2)
104.9(1)
123.7(1)
131.4(1)
122.4(1)
114.1(1)
123.4(1)
125.8(2)
Notes and references
1 J. Elguero, A. Fruchier and R. Jacquier, Bull. Soc. Chim. Fr.,
1967, 2619.
2 J. Elguero, A. Fruchier and R. Jacquier, Bull. Soc. Chim. Fr.,
1968, 3331.
3 J. Elguero, C. Marzin, A. R. Katritzky and P. Linda, The
Tautomerism of Heterocycles, Advances in Heterocyclic
c
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013
New J. Chem.

View Article Online
Paper
NJC
Chemistry, Supplement 1, Academic Press, New York, 1976,
and references therein.
and J. A. Sternberg, Helv. Chim. Acta, 1986, 69, 1971;
(d) S. E. Denmark, Y.-C. Moon and C. B. W. Senanyake,
J. Am. Chem. Soc., 1990, 112, 311.
´
4 J. Catalan, J. L. G. de Paz and J. Elguero, J. Chem. Soc., Perkin
Trans. 2, 1996, 57.
5 I. Alkorta and J. Elguero, J. Phys. Org. Chem., 2005, 18, 719.
6 C. H. Gustafsson, Finska Kemistamfundets Medd., 1938, 1, 2.
18 (a) R. S. Paton, J. L. Mackey, W. H. Kim, J. H. Lee,
S. J. Danishefsky and K. N. Houk, J. Am. Chem. Soc., 2010,
¨
132, 9335; (b) N. Çelebi-Olçu¨m, D. H. Ess, V. Aviyente and
¨
7 J. Sandstrom, Acta Chem. Scand., 1964, 18, 871.
K. N. Houk, J. Org. Chem., 2008, 73, 7472.
˜
´
8 J. L. M. Abboud, P. Cabildo, T. Canada, J. Catalan, 19 J. E. Baldwin, J. Chem. Soc., Chem. Commun., 1976, 734.
R. M. Claramunt, J. L. G. de Paz, J. Elguero, H. Homan, 20 (a) C. O. Kappe and A. Stadler, Microwaves in Organic and
R. Notario, C. Toiron and G. I. Yranzo, J. Org. Chem., 1992,
57, 3938.
9 I. Alkorta and J. Elguero, J. Org. Chem., 2002, 67, 1515.
medicinal Chemistry, Wiley-VCH, Weinheim, 2005;
(b) Microwaves in Organic Chemistry, ed. A. Loupy, Wiley-
VCH, Weinheim, 2nd edn, 2006.
´
10 M. A. Garcıa, P. Cabildo, R. M. Claramunt, E. Pinilla, 21 (a) ‘‘How often have I said to you that when you have eliminated
M. R. Torres, I. Alkorta and J. Elguero, Inorg. Chim. Acta,
2010, 363, 1332.
11 W. Baik, T. H. Park, B. H. Kim and Y. M. Jun, J. Org. Chem.,
1995, 60, 5683.
12 B. H. Kim, D. H. Kim, H. J. Park, R. Han, Y. M. Jun and
W. Baik, Bull. Korean Chem. Soc., 2001, 22, 1163.
the impossible, whatever remains, however improbable, must
be the truth?’’ A. Conan Doyle, The Sign of Four; (b) ‘‘It is an old
maxim of mine that when you have excluded the impossible,
whatever remains, however improbable, must be the truth’’.
A. Conan Doyle, The Adventure of the Beryl Coronet.
22 C. Schweitzer and R. Schmidt, Chem. Rev., 2003, 103, 1685.
13 (a) R. Paulini, K. Mu¨ller and F. Diederich, Angew. Chem., Int. 23 Although experimentally eliminated we have calculated the
´
Ed., 2005, 44, 1788; (b) G. P. A. Yap, F. A. Jove,
reaction via triplet [6-31G(d)+EtOH(PCM)]. Compound 4 in
the ground state rotates about the N–C bond to yield 40
which lies 2.6 kJ molꢁ1 higher in energy with a barrier of
19.1 kJ molꢁ1 (this process is preferred to the 4 - 4T-40T).
Excitation of 40 leads to 40T that, without barrier, affords
intermediate INT4 of Fig. 16. From INT4 the reaction
R. M. Claramunt, D. Sanz, I. Alkorta and J. Elguero, Aust.
J. Chem., 2005, 58, 817; (c) E. Pinilla, M. R. Torres,
R. M. Claramunt, D. Sanz, R. Prakash, S. P. Singh,
I. Alkorta and J. Elguero, ARKIVOC, 2006, ii, 136; (d) M. D.
´
´
´
S. Marıa, R. M. Claramunt, M. A. Garcıa and J. Elguero,
Tetrahedron, 2007, 63, 3737; (e) F. R. Fischer, P. A. Wood,
proceeds to 5 with a barrier of 35.6 kJ molꢁ1
.
F. H. Allen and F. Diederich, Proc. Natl. Acad. Sci. U. S. A., 24 H. W. Heine, G. J. Blosick and G. B. Lowrie, Tetrahedron
2008, 195, 17290; ( f ) N. Triballeau, E. Van Name, G. Laslier, Lett., 1968, 9, 4801.
D. Cai, G. Paillard, P. W. Sorensen, R. Hoffmann, 25 J. A. Sousa, J. Weinstein and A. L. Bluhm, J. Org. Chem.,
H.-O. Bertrand, J. Ngai and F. C. Acher, Neuron, 2008, 1969, 34, 3320.
60, 767; (g) M. Solimannejad, I. Alkorta and J. Elguero, 26 R. Bauernschmitt and R. Ahlrichs, Chem. Phys. Lett., 1996,
Chem. Phys. Lett., 2009, 474, 253. 256, 454.
14 A. M. S. Silva, R. M. S. Sousa, M. L. Jimeno, F. Blanco, 27 See for example: (a) T. Vreven, F. Bernardi, M. Garavelli,
I. Alkorta and J. Elguero, Magn. Reson. Chem., 2008, 46, 859.
15 F. Blanco, I. Alkorta and J. Elguero, Magn. Reson. Chem.,
2007, 45, 797.
16 P. M. Mancini, G. G. Fortunato and L. R. Vottero, J. Phys.
Org. Chem., 2005, 18, 336.
M. Olivucci, M. A. Robb and H. B. Schlegel, J. Am. Chem. Soc.,
1997, 119, 12687; (b) N. Yamamoto, T. Vreven, M. A. Robb,
M. J. Frisch and H. B. Schlegel, Chem. Phys. Lett., 1996, 250, 373;
(c) F. Bernardi, A. Bottini, J. J. W. McDougall, M. A. Robb and
H. B. Schlegel, Faraday Symp. Chem. Soc., 1984, 19, 137.
17 (a) S. E. Denmark and A. Thorarensen, Chem. Rev., 1996, 28 Oxford Diffraction, CrysAlis CCD and CrysAlis RED. Versions
96, 137; (b) S. E. Denmark, B. S. Kesler and Y.-C. Moon,
J. Org. Chem., 1992, 57, 4912; (c) S. E. Denmark, C. J. Cramer
(1.171.34.9), Oxford Diffraction Ltd, Abingdon, Oxfordshire,
England, 2009.
c
New J. Chem.
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013

View Article Online
NJC
Paper
29 M. C. Burla, R. Caliandro, M. Camalli, B. Carrozzini,
G. Cascarano, L. De Caro, C. Giacovazzo, G. Polidori and
R. Spagna, J. Appl. Crystallogr., 2005, 38, 381.
F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin,
V. N. Staroverov, R. Kobayashi, J. Normand,
K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar,
J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene,
J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo,
R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin,
R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin,
K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador,
30 G. M. Sheldrick, SHELXS97 and SHELXL97, University of
¨
Gottingen, Germany, 1997.
31 A. L. Spek, J. Appl. Crystallogr., 2003, 36, 7.
32 L. J. Farrugia, J. Appl. Crystallogr., 1997, 30, 565.
33 C. F. Macrae, I. J. Bruno, J. A. Chisholm, P. R. Edgington,
¨
´
P. McCabe, E. Pidcock, L. Rodrıguez-Monge, R. Taylor, J. van
J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas,
de Streek and P. A. Wood, J. Appl. Crystallogr., 2008, 41, 466.
34 (a) C. Lee, W. Yang and R. G. Parr, Phys. Rev. B, 1988,
37, 785; (b) R. G. Parr and W. Yang, Density-Functional
Theory of Atoms and Molecules, Oxford, New York, 1989.
35 (a) A. D. Becke, Phys. Rev. A: At., Mol., Opt. Phys., 1988,
38, 3098; (b) W. Kohn, A. D. Becke and R. G. Parr, J. Phys.
Chem., 1996, 100, 12974; (c) A. D. Becke, J. Chem. Phys., 1993,
98, 5648.
J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox,
GAUSSIAN 09 (Revision A.1), Gaussian, Inc., Wallingford, CT,
2009.
38 (a) S. Miertus, E. Scrocco and J. Tomasi, Chem. Phys., 1981,
55, 117; (b) B. Mennucci and J. Tomasi, J. Chem. Phys., 1997,
106, 5151; (c) R. Cammi, B. Mennucci and J. Tomasi, J. Phys.
Chem. A, 2000, 104, 5631.
39 J. W. McIver Jr. and A. K. Komornicki, J. Am. Chem. Soc.,
1972, 94, 2625.
36 (a) W. J. Hehre, L. Radom, P. v. R. Schleyer and J. A. Pople,
Ab Initio Molecular Orbital Theory, Wiley, New York, 1986, 40 (a) R. Ditchfield, W. J. Hehre and J. A. Pople, J. Chem. Phys.,
p. 76; (b) R. Ditchfield, W. J. Hehre and J. A. Pople, J. Chem.
Phys., 1971, 54, 724; (c) W. J. Hehre, R. Ditchfield and
J. A. Pople, J. Chem. Phys., 1972, 56, 2257.
1971, 54, 724; (b) M. J. Frisch, J. A. Pople and J. S. Binkley,
J. Chem. Phys., 1984, 80, 3265.
41 (a) R. Ditchfield, Mol. Phys., 1974, 27, 789; (b) F. London,
J. Phys. Radium, 1937, 8, 397.
37 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, 42 R. F. W. Bader, Atoms in Molecules: A Quantum Theory,
B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, Clarendon Press, Oxford, 1990.
X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, 43 P. L. A. Popelier, Atoms in Molecules. An introduction,
J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, Prentice Hall, Harlow, England, 2000.
J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, 44 T. A. Keith, 11.10.16 ed., TK Gristmill Software, 2011,
H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta,
http://aim.tkgristmill.com.
c
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013
New J. Chem.