EPR and ENDOR spectroscopic study of the reactions of aromatic azides with gallium trichloride

Article abstract of DOI:10.1039/c0ob00084a

The reactions of gallium trichloride with phenyl and deuterio-phenyl azides, as well as with 4-methoxyphenyl azide and deuterium isotopomers, were examined by product analysis, CW EPR spectroscopy and pulsed ENDOR spectroscopy. The products included the corresponding anilines together with 4-aminodiphenylamine type dimers, and polyanilines. Complex CW EPR spectra of the radical cations of the dimers [ArNHC₆H₄NH ₂]⁺ and trimers [ArNHC₆H₄NHC ₆H₄NH₂]⁺ were obtained. These EPR spectra were analysed with the help of data from the deuterium-substituted analogues as well as the pulse Davies ENDOR spectra. DFT computations of the radical cations provided corroborating evidence and suggested the unpaired electrons were accommodated in extensive π-delocalised orbitals. A mechanism to account for the reductive conversion of aromatic azides to the corresponding anilines and thence to the dimers and trimers is proposed.

Full text of DOI:10.1039/c0ob00084a

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
EPR and ENDOR spectroscopic study of the reactions of aromatic azides
with gallium trichloride†
Giorgio Bencivenni,a Riccardo Cesari,a Daniele Nanni,a Hassane El Mkamib and John C. Walton*c
Received 5th May 2010, Accepted 20th July 2010
DOI: 10.1039/c0ob00084a
The reactions of gallium trichloride with phenyl and deuterio-phenyl azides, as well as with
4-methoxyphenyl azide and deuterium isotopomers, were examined by product analysis, CW EPR
spectroscopy and pulsed ENDOR spectroscopy. The products included the corresponding anilines
together with 4-aminodiphenylamine type dimers, and polyanilines. Complex CW EPR spectra of the
∑
∑
radical cations of the dimers [ArNHC6H4NH2]+ and trimers [ArNHC6H4NHC6H4NH2]+ were
obtained. These EPR spectra were analysed with the help of data from the deuterium-substituted
analogues as well as the pulse Davies ENDOR spectra. DFT computations of the radical cations
provided corroborating evidence and suggested the unpaired electrons were accommodated in extensive
p-delocalised orbitals. A mechanism to account for the reductive conversion of aromatic azides to the
corresponding anilines and thence to the dimers and trimers is proposed.
Allylated nitrogen heterocycles have been obtained by reaction
Introduction
of d-azido esters and chlorides with allylindium dichloride.7b In
the course of the latter studies, only indirect evidence of the
intermediacy of nitrogen-centred radicals was obtained, including
the EPR spectrum of the allyl radical during photolysis of
allylindium dichloride.7b Any attempt to detect aminyl radicals
by treatment of azides with indium (or other Group 13 metal)
reagents failed, independently of the technique of generation of
the initial metal-centred radicals. However, during one of those
attempts, we realised that Group 13 metal halides, when treated
with organic azides, gave rise to strong, persistent EPR signals that
were particularly consistent in the case of gallium trichloride.8
Here we report the results obtained from the reactions of some
aromatic azides with gallium trichloride.
Many novel applications of organic azides in synthetic organic
chemistry have been appearing in the literature in recent years.1
One reason for this is the facility with which nitrogen-centred
radicals can be generated from the azido group.2 Alkyl, vinyl,
aryl, and acyl radicals, for example, can ring close onto aliphatic
and aromatic azides yielding five and/or six-membered cyclic
aminyl radicals, which in turn are key intermediates for the
construction of N-heterocycles.3 Recently, azides have been shown
to be useful intermediates for generation of iminyl radicals as
well.4 The realm of radical reactions in general, and of radical
generation from azides in particular, has long been dominated
by toxic metal reagents and, especially, organotin compounds.
In the search for efficient purification procedures and sound tin
substitutes the last years have witnessed thorough investigations
of novel ways of carrying out radical reactions in the absence
of tin reagents.5 Concerning this, the use of indium derivatives
(dichloroindium hydride in particular), as well as other Group
13 metal compounds,6 including organogallium compounds,6c,h,l
has come to the forefront of radical organic synthesis as a more
benign alternative to other metal-based catalysts. This class of
reagents has recently found applications in the field of azide
radical chemistry as well, resulting in reduction of a variety of
organic azides to the corresponding amines and cyclisation of g-
azidonitriles to pyrrolidin-2-imines with dichloroindium hydride.7a
Results and discussion
Products from reactions of aryl azides with gallium trichloride
Gallium trichloride (0.25 mmol) in pentane (0.55 mL) was added
to phenyl azide 1a (0.25 mmol) in dichloromethane (4.0 mL)
at room temperature. A vigorous reaction with gas evolution
(probably nitrogen) was observed and the colour changed from
pale yellow to dark green and eventually to deep violet. The
solution was examined by 9 GHz EPR spectroscopy and found to
exhibit strong and complex spectra that persisted for several weeks
in the absence of oxygen. The resultant mixture was analysed by
GC-MS, and the major products were isolated.
In this way it was shown that the mixture contained unreacted
1a together with aniline (2a), 4- and 2-chloroanilines (3a and 4a),
4-aminodiphenylamine (5a), plus traces of 2-aminodiphenylamine
(6a) and 5,10-dihydrophenazine (7a) (Scheme 1).
aDipartimento di Chimica Organica “A. Mangini”, Universita` di Bologna,
Viale del Risorgimento 4, I-40136 Bologna, Italy. E-mail: nanni@
ms.fci.unibo.it
bSchool of Physics and Astronomy, University of St. Andrews, St. Andrews,
Fife, UK, KY16 9SS. E-mail: hem2@st-andrews.ac.uk
cSchool of Chemistry, University of St. Andrews, EaStChem, St. Andrews,
Fife, UK, KY16 9ST. E-mail: jcw@st-and.ac.uk; Fax: +44 (0)1334 463808;
Tel: +44 (0)1334 463863
The reaction was repeated with several sets of starting condi-
tions and the product yields are set out in Table 1.
† Electronic supplementary information (ESI) available: Details of prepa-
rations of isotopically substituted anilines. EPR spectra of deuterium
substituted species from azides 8b–8e. ENDOR simulations for species
from azides 1a and 1d. Cartesian coordinates for DFT computed
structures. See DOI: 10.1039/c0ob00084a
When the reaction was carried out with a 3-fold excess of
GaCl3 (entry 2) the starting azide was completely consumed and
4-chloroaniline (3a) was the almost exclusive product with only
traces of 4a and 5a. This suggested GaCl3 was involved in the
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 5097–5104 | 5097
©

View Article Online
a
Table 1 Product yields from reaction of 1a with GaCl3
Entry
1a : GaCl3
Solvent
1ab
2a
3a & 4a
5a
1
2
1 : 1
1 : 3
3 : 1
1 : 1
CH2Cl2 pentane
CH2Cl2 pentane
CH2Cl2 pentane
CH2Br2 pentane
45
0
nd
nd
5
0
t
30
15 & 22
90 & t
t & t
10
t
50
20
3
4c
25 & 15
a Yields in mol%. b Unreacted 1a; t = trace, nd = not determined. c Traces
of chlorinated 4-aminodiphenylamine were also observed.
Scheme 2 Products from reaction of azide 8a with gallium trichloride.
analysis of the total reaction mixture in positive ion mode showed
a compound with mass 213 amu. This corresponds to protonated
quinodiimine 13a. It seems probable that a sequence of oxidations
is responsible for the series 9a, 13a, 11a, 12a, possibly during work-
up.13 Analysis in negative ion mode showed a peak corresponding
Scheme
trichloride.
1
Products from reaction of phenyl azide with gallium
-
to GaCl4 , suggesting this was the counter-ion for the organic
radical cations.14
chlorination of aniline, probably by an electrophilic mechanism.
When CH2Br2 was used in place of CH2Cl2 (entry 4) the products
were again aniline and chloroanilines together with “dimer” 5a.
No bromoanilines were detected. This confirmed that the GaCl3,
rather than the CH2Cl2 solvent, was the source of the Cl atoms.
When a 3-fold excess of azide was used (entry 3) only traces of
aniline and chloroanilines were observed and the ‘head to tail’
dimer 5a became the main product. Significant amounts of tars
were also noted.
Literature research has shown that anilines can easily be
oxidised to the corresponding resonance-stabilised radical cations,
which can couple with more aniline to afford very persistent radical
cation dimers.9 It seemed possible, therefore, that the species we
observed by EPR spectroscopy might be related to dimer 5a. The
generation of these radical cations depends critically on reaction
conditions, in particular the degree of protonation, which can
facilitate the electron transfer (ET).10 It has also been reported
that electrochemical oxidation of aromatic amines can generate
the same radical cations which can polymerize giving oligo- and
poly-anilines.11
Preparation of deuteriated and 15N-substituted aromatic azides
As an aid to analyzing the complex EPR spectra, a series of
deuterium-substituted phenyl azides (1b–d) and 4-methoxyphenyl
azides (8b–e) was prepared as well as the 15N-isotopomer 8f
(Scheme 3). Deuteriated azides 1b–d and 8b–e were prepared by
standard diazotisation techniques starting from the previously
reported deuteriated anilines; azide 8f was synthesised from
the known 4-methoxyaniline-15N by diazo-transfer reaction with
azido tris(diethylamino)phosphonium bromide (see Electronic
Supplementary Information for details and references).
4-Methoxyphenyl azide (8a) also reacted vigorously with GaCl3
leading to evolution of nitrogen and formation of a deep blue
colour.12 The EPR spectra exhibited strong signals that persisted
for many weeks. Product analyses demonstrated the reaction
was less sensitive to reaction conditions and always furnished 4-
amino-4¢-methoxydiphenylamine (9a, Variamine blue) as the main
product, together with traces of 4-methoxyaniline (10a), 4-(4-
methoxyphenylamino)phenol (11a), its oxidised quinonic form
12a (Scheme 2), 2,7-dimethoxy-5,10-dihydrophenazine and much
dark coloured polymer.
The observation of 10a suggested that this amine was again
the precursor for the other products. Thus, 9a might result from
substitution on 10a by a radical, or radical-cation related to
10a, accompanied by loss of the methoxy group. Direct ESI-MS
Scheme 3 The set of deuterium- and 15N-substituted aromatic azides.
5098 | Org. Biomol. Chem., 2010, 8, 5097–5104
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
Table 2 EPR parameters for paramagnetic “trimer” species 15 from 1a–
EPR spectroscopic study of intermediates from phenyl and
deuterio-phenyl azides
da,b
Precursor azide N-8 N-15 N-1 2 ¥ H-1 H-8 H-15 H-rings
The aromatic azides were reacted with GaCl3 in CH2Cl2–pentane
solution. Aliquots (ca. 0.1 mL) were placed in quartz capillary
tubes (dia. 1 mm), purged with nitrogen for 15 min, and transferred
to the resonant cavity of a 9 GHz EPR spectrometer. Strong
spectra of persistent species were observed, but the pattern evolved
with time and reaction conditions. A set of spectra obtained from
PhN3 is shown in Fig. 1.
1a species (i)
5.0 4.9
3.0 6.5
4.9
2.1
2.1(3H)
1.0(3H)
2.0(4H)
—
1b
5.2 4.8
5.4 5.0
5.2 4.9
5.3 3.4
2.5 6.5
2.9 6.4
2.5 6.5
2.1 -3.0
Dc
Dc
Dc
Dc
Dc
Dc
1c
1d
—
DFTd for 15a
-7.7 -4.8 -1.6(4H)
-1.1(3H)
-0.6(2H)
0.6(2H)
-0.2(2H)
a Hfs in G; note that the experimental EPR spectra can only give the
magnitude (not the sign) of hfs. b The atom numbers in line 1 are defined
in Scheme 4. c Hfs from D-atoms not resolved. d UB3LYP/6-31G(d).
Fig. 1 EPR spectra from PhN3 and GaCl3. (a) 1st derivative spectrum
on first mixing. (b) 1st derivative spectrum after several hours. (c) 2nd
derivative of b [species (i)]. (d) 2nd derivative of another species (ii) in
CD2Cl2 solution.
On first mixing the reactants, broad strong spectra (Fig. 1a) were
usually obtained but over a matter of hours, or after irradiation
with UV light, fine structure usually developed (Fig. 1b). A
broad central component was always present and we attribute
this to persistent polymeric and oligomeric paramagnetic species.
By using 2nd derivative presentation with a low modulation
amplitude (0.2 Gpp) discrimination against the broad component
was achieved and a well resolved spectrum resulted (Fig. 1c).
When CD2Cl2–pentane was used as solvent in place of CH2Cl2–
pentane another spectrum, well resolved in 2nd derivative mode,
was observed (Fig. 1d). The same spectrum as in (d) was also
observed on occasion in CH2Cl2–pentane, during the evolution of
the spectral pattern with time. The spectra shown in (c) and (d)
clearly correspond to two different species (i) and (ii).
To assist in analyzing the spectra and identifying the para-
magnetic species, the deuteriated azides 1b–d were individually
treated with GaCl3 under the same conditions. The best resolved
spectra, together with computer simulations, are shown in Fig. 2.
As expected, the deuterium substitution caused some broadening
of the spectra, but resolution was still satisfactory.
Fig. 2 2nd Derivative EPR spectra from 1b–1d and GaCl3 at 300 K in
DCM–pentane. Top: (a) Spectrum from 1b: upper, experimental; lower,
simulation. Centre: (b) Spectrum from 1c: upper, experimental; lower,
simulation. Bottom: (c) Spectrum from 1d: upper, experimental; lower,
simulation.
H-atoms (Table 2 and Fig. 2). Although the simulations were not
unique, and satisfactory computer simulations could be obtained
for individual spectra with other sets of parameters, the data of
Table 2 and Fig. 2 was the only global set that gave a consistent
picture for all three isotopically substituted species.
By utilising the three N-atom and two equivalent H-atom hfs
derived from the deuterium-substituted species we were able to
obtain a satisfactory simulation of species (i) derived from the
parent azide 1a (Fig. 3a). The good quality of the agreement is
shown in Fig. 3 and the corresponding hfs are noted in Table 2.
A different set of parameters was required for the analysis
of the spectrum of species (ii). In this case good fits were only
obtained with two equivalent N-atoms, three sizeable H-atom hfs
and several additional ring H-atom hfs (Fig. 3 and Table 3).
The hfs we observed for species (ii) are quite similar to
those reported in the literature for the radical cation of 4-
aminodiphenylamine (14a)9b,15 (Table 3). The small differences
The spectrum from the fully deuteriated azide 1d was well simu-
lated by using three non-equivalent N-atom hyperfine splittings
(hfs) and two equivalent H-atom hfs (Table 2). Similarly, the
spectrum from the tetra-deuterio azide 1c was well simulated with
a quite similar set of hfs (Table 2). A good simulation of the
spectrum from the tri-deuterio azide 1b was obtained with again
nearly the same three non-equivalent N-atom hfs, the same two
equivalent H-atom hfs and additional hfs from four equivalent
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 5097–5104 | 5099
©

View Article Online
Table 3 EPR parameters for dimer species (ii) from 1aa
N-8, N-1
H-8
2 ¥ H-1
H-14, 5
H-6, 10
H-11, 13,3
H-4, 12
Exptl.
4.9,4.9
5.4,4.9
6.1,3.8
2.6,3.1
6.8
6.6
-8.7
-4.8
5.6
6.0
-5.7
-5.7
3.1
1.7
-2.6, -2.2
-2.6, -2.1
2.0
1.7
-2.1, -2.0
-2.3, -2.1
1.0
1.1
0.6
0.6
-1.1, 1.0
-2.3, -0.9
Lit. for 14ab
DFTc
-1.9,1.2, -1.1
-0.8, -0.1, -1.8
DFTd
a hfs in G; note that the experimental EPR spectra can only provide the magnitude (not the sign) of hfs. The atom numbers in line 1 are defined in Scheme
4. b See ref. 9, 15. c UB3LYP/6-31G(d). d UB3LYP/epr-iii (see ref. 19).
Fig. 3 Experimental and simulated EPR spectra from PhN3 and GaCl3
at 300 K. Top panel (a): 2nd derivative experimental and simulated spectra
of species (i). Bottom panel (b): 2nd derivative experimental and simulated
spectra of species (ii).
are probably due to the different solvent and counter ions. In
view of our observation of the amine 5a as a major product, we
can confidently identify species (ii) as radical cation 14a. On the
other hand, species (i) from 1a, and all the paramagnetic species
from the deuteriated azides 1b–d, showed the presence of 3 non-
equivalent N-atoms. The most probable explanation is that these
signals correspond to the “trimer” 15a [N1-(4-aminophenyl)-N4-
phenylbenzene-1,4-diamine] and deuterium isotopomers thereof
Scheme 4 Radical cations derived from phenyl azide.
15x–z (Scheme 4).
In the case of 15a, derived from the parent azide, the hfs from
all three N-atoms were resolved, as well as the hfs from the amino
H-atoms. Hfs from six of the total thirteen ring H-atoms were
also resolved. The DFT computation (Table 2) was in reasonable
EPR spectra of intermediates from 4-methoxyphenyl azide,
deuteriated analogues and 15N-substituted azide 8f
agreement and suggested that at least six of the ring H-atoms
would be too small to be resolved. In agreement with our analysis,
the trimer cations 15y and 15z showed no resolved hfs from their
almost fully deuteriated rings. The spectral data established that
for both 15y and 15z the terminal NH2 groups contained H-atoms
rather than D-atoms. Interestingly, however, in both cases, the
chain amino groups (N-8, N-15) gave no resolved hfs, and therefore
they were present as ND groups. In the case of 15x, derived from
the 2,4,6-trideuterio-azide 1b, a similar picture emerged, except
that hfs from four of the ring positions were now resolved. Thus,
the EPR spectra suggested an evolution in the GaCl3 promoted
reaction of phenyl azide from the long-lived radical cation of 4-
aminodiphenylamine (the dimer), to a long-lived trimer radical
cation, and thence to long-lived oligomeric and polymeric species
with broad EPR spectra.
The 4-methoxyphenyl azides 8a–f were reacted individually at
room temperature with GaCl3 in CH2Cl2–pentane solution in
the same way as the phenyl azides. The reactions followed a
similar course in that nitrogen was evolved and deep blue colours
(sometimes appearing green) developed along with much dark
polymer. The EPR spectra showed strong and very persistent
signals which were initially broad, somewhat like the top spectrum
(a) in Fig. 1, but which developed fine structure on standing.
All the spectra contained broad central components which were
discriminated against by use of low modulation intensities. Final
traces of the broad components were digitally removed. The
spectrum from 8a and its 15N-isotopomer 8f are shown in Fig.
4. Spectra from deuteriated azides 8b–e are in the ESI.
Computer simulations were non-trivial because of the multiplic-
ity of lines and because small differences in line width, as well as
5100 | Org. Biomol. Chem., 2010, 8, 5097–5104
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
a
Table 4 EPR data from 4-amino-4¢-methoxydiphenylamine radical cations 14 derived from 4-MeOC6HnD4-nxNN2 and GaCl3
Precursor azide [H or D]
N-8
N-1
H-8
2H-1
2H/2D-5,6
2H/2D-10,11
2H-3,4
2H-12,13
8a [H4]
5.2
4.4
7.3
-8.7
7.4
7.3
7.0
1.3 (D)
7.1
3.8
-4.6
3.8
3.9
3.6
4.0
3.6
3.1
-2.8
3.0
3.1
0.5 (2D)
3.0
2.2
-2.1
2.2
2.2
0.5 (2D)
2.3
0.8
-0.7
0.8
0.8
0.8
—
0.4
0.4
0.4
—
—
—
DFT for 14cb
8f [H4,15N]
8b [OCD3]
8c [D2]
6.1
3.1
7.4 (15N)
5.2
6.1 (15N)
4.3
5.3
5.3
5.2
4.1
3.8
4.4
8d [D2]
8e [D4]
0.66 (2D)
0.55 (2D)
—
—
a Hfs in G for CH2Cl2–pentane soln. at 300 K. Atom numbers in column headings refer to the positions defined in structure 15a. b DFT computation for
radical cation 14c using UB3LYP/6-31G(d).
agreement (Table 4, line 3). No “trimer” radical cations, analogous
to 15, were observed in any of the experiments. Previously reported
EPR spectra of the Variamine blue radical cation were not well
resolved and hence the derived hfs [a(2 N) = 5.6, a(3H) = 7.6,
a(4H) = 2.0 G at rt in water] were approximate estimates.16 The
reasonable correspondence between these values and those of
Table 4 supports our identification.
One difference between the spectra obtained from the
deuterium-substituted versions of the “trimer” radical cations,
that is 15b–d, and the deuterium isotopomers of 14 was that the
latter were better resolved with smaller line widths [compare Fig.
4 (and the examples in the ESI) with Fig. 2]. The probable reason
was that in the deuterium isotopomers of 14 all the amino groups
remained as NH or NH2 (Table 4) whereas in 15y–z the chain
amino groups appeared as ND. The presence of 2 non-equivalent,
exchangeable, ND atoms, with comparatively large unresolved,
a(D) values, probably accounts for the somewhat broader lines
of the 15y–z spectra in Fig. 2. One exception to this was the
spectrum of 14d from 8d (see ESI) which did have broader lines.
The simulations suggested that the central NH group had become
ND in this case, although slightly better fits were obtained by
adding in ca. 30% of the species retaining NH. The assignments
of the hfs to particular atoms in Table 4 are, of course, tentative:
the DFT computed values were used as a guide.
Fig. 4 EPR spectra from 4-methoxyphenyl azides. Top panel (a):
Experimental EPR spectrum from 8a in CH2Cl2–pentane at 300 K with
the computer simulation below. Bottom panel (b): Experimental EPR
spectrum from 15N- isotopomer 8f with the simulation below.
in the hfs, had a large effect on the appearance of the spectra. The
spectra of the deuterium-substituted species were well resolved
and this, together with the good spectrum from 8f, enabled a
complete analysis to be carried out. The derived hfs are listed in
Table 4.
The hfs form a very consistent set and the magnitudes of both
N-hfs and H-hfs are conserved from one species to another as
would be expected for D- and 15N-substitution. The ratio of the
nuclear magnetic moments 14N/15N is 1.4 so the a(14N) values
calculated from the a(15N) values (Table 4, line 4) are 5.3 and 4.4
G; in close agreement with the observed a(14N) values from 8a
(Table 4, line 2).
Similarly the ratio [mD/ID]]/[IH/mH] is 6.5 giving calculated
a(2D)-5,6 and a(2D)-10,11 values of 0.55 and 0.3 G respectively.
These are in satisfactory agreement with the data in Table 4 (lines
5, 6) taking into account the fact that the line widths of 0.1 to 0.2 G
introduced error limits of this order. Interaction of the unpaired
electron with the 4-CH3O group was too small for CH3 hfs to
be resolved. The spectrum of the species from 8a was, therefore,
almost identical to that from the CD3O-analogue 8b (compare
Table 4, line 2 with line 5). The overall analysis of Table 4 shows
that a single paramagnetic species containing 2 non-equivalent N-
atoms, 3 (N)-H atoms and 8 H/D-atoms was responsible for all the
spectra. There can be little doubt that this is the radical cation of
4-amino-4¢-methoxydiphenylamine (9a, Variamine blue). A DFT
computation on 14a [UB3LYP/6-31G(d)] gave hfs in satisfactory
Pulse ENDOR spectra of intermediates from 1a and
tetra-deuteriophenyl azide 1c
In seeking support for our identifications of the intermediates,
pulsed ENDOR experiments, based on the ESE effect, were carried
out on the frozen solutions from azides 1a and 1c. The echo signal
was created by the microwave pulse sequence and an rf pulse was
applied during the “mixing period” which corresponded to the
time T in the Davies ENDOR sequence. The rf pulse drove the
nuclear spin transitions which led to a change in the ESE intensity.
The ENDOR signal was therefore measured by monitoring the
ESE intensity while the rf frequency was varied. In the case of
1
2
an S = system coupled with a nucleus with nuclear spin I =
12 , the Davies ENDOR spectrum consists of two lines at the
nuclear resonance frequencies na and nb which correspond to the
transitions associated to the electron spin manifold Ms = +1 and
2
1
Ms = -2 respectively. If the Larmor frequency (nn) of the nucleus in
question is larger than the hyperfine interaction then the resonance
frequencies are given by: nab = |nn 12 aiso|; if nn is less than aiso/2
the frequencies are then given by: nab = |1 aiso nn|. The presence
2
of anisotropy considerably alters this picture, and aiso is replaced
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 5097–5104 | 5101
©

View Article Online
by Ai (i.e., one of the principal components of the hyperfine tensor)
to take account all the orientations of the molecule with respect
to the applied magnetic field.
difficult. Therefore, to interpret the ENDOR data, our ENDOR
simulations were based mainly on the CW EPR results. The
simulated spectra are displayed in Fig. 5 & 6.
In Fig. 5, we report the Davies ENDOR spectrum from
the species derived from the 1a sample at 50 K. The inset
shows the ESE-EPR spectrum, with the arrow indicating the
magnetic field position at which the ENDOR experiment was
performed. The ENDOR spectra show powder pattern lineshapes,
as expected for frozen solutions, due to the anisotropic hyperfine
interactions.
A decomposition of each simulation into the contributions
from individual magnetic nuclei is given in the ESI. The ENDOR
simulation tensors are also given in the ESI. Each 1H contributes
three sets of peaks to the spectrum times the number of 1H’s
present. This represents an enormous number of lines in one
spectrum. Obviously they cannot all be assigned from this broad
unresolved powder pattern. Almost axial tensors were assumed
(see ESI). However, it should be noted that it may well be
possible to simulate these spectra with other parameter sets. The
experimental Davies ENDOR data supports the CW EPR data
in confirming the magnitudes of the hyperfine couplings and the
nitrogen interactions.
There were some differences in the case of the deuteriated trimer
15y. Only one D coupling of ca. 0.3 G was resolved. This is close to
the value expected for (N)D-15 (see Table 2). A second D hyperfine
from (N)D-8 (ca. 0.8 G based on the H-8 value) was expected.
However, these amino D/H atoms are readily exchangeable and
it is probable that (N)D-8 exchanged for (N)H-8 in the sample
studied by ENDOR. This sample was several days older than
the CW EPR sample. The region of Fig. 6 centred around the
1H Larmor frequency also shows more couplings than expected
Fig. 5 Davies ENDOR spectrum of species (14a) from 1a at 50 K.
1
for 15y which has only 2 types of H. However the additional
couplings can also be attributed to H/D exchange.
Two main features cover the whole spectrum; a powder pattern
centred about the 1H Larmor frequency and a second broad signal
located at lower frequency and spread over 8 MHz width. A well-
marked singlet appears at the centre of the 1H pattern due to very
small, unresolved coupling corresponding to interaction of the
unpaired electron with the matrix and probably to the presence
of polymer radical cations. The ENDOR spectrum of the species
from 1a appears somewhat asymmetric which could be related to
the saturation of the nuclear transition in one of the electron spin
manifolds.17
DFT computations on radical cations
Quantum chemical calculations were carried out with the Gaus-
sian 03 programme package.18,19 Density functional theory with the
UB3LYP functional, was employed. The equilibrium geometries
were fully optimised with respect to all geometric variables, no
symmetry being assumed, with the 6-31G(d) basis set. Computed
<S2> values for the radical cations were 0.7500 after annihilation
of higher multiplicity spin states. Isotropic EPR hfs were derived
from computed Fermi contact integrals evaluated at the H- and
N-nuclei. The hfs were taken directly from the Gaussian output
files.
The ENDOR spectrum of the species from 1c is shown in Fig.
1
6. This contains a powder pattern centred about the H Larmor
frequency, again with a strong central singlet, together with a
broad signal at lower frequency and an additional well resolved
2
doublet with a splitting of 0.94 MHz centred around the H (D)
The optimum structures of the radical cations 14a and 15a, and
their associated SOMOs, are shown in Fig. 7.
Larmor frequency.
Fig. 6 Davies ENDOR spectrum of species (15y) obtained from deuteri-
ated azide 1c at 50 K.
The lack of resolution encountered in these ENDOR spectra
makes a direct assignment of these signals to the different nuclei
Fig. 7 DFT computed structures and SOMOs for dimer and trimer
radical cations.
5102 | Org. Biomol. Chem., 2010, 8, 5097–5104
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
˚
The C–NH2 bond lengths in radical cation 14a (1.35 A) and
˚
15a (1.36 A) indicated significant double bond character, as did
˚
the central C–NH bond lengths which were all £ 1.415 A. The
CNC angle in 14a was 130.3◦ and the CNC angles in 15a were
130.8 and 129.5◦ for 7-8-9 and 14-15-16 respectively, showing
significant widening from trigonal. The aromatic rings in both
structures twisted significantly out of co-planarity. It seems that a
compromise was reached in which the repulsive steric interaction
between ortho-H-atoms of neighbouring rings was balanced
against the stabilising effect from conjugation of the p-systems.
The SOMOs depicted in Fig. 7 show that there was still sufficient
orbital overlap to support lengthy p-systems extending over all the
rings and N-atoms in both structures. This is in good accord with
the EPR spectroscopic data which shows extensive delocalisation
of the unpaired electron. The computed hfs are in Tables 2 and 3
and they show reasonable correspondence with experiment.
Conclusions
Aniline 2a and 4-chloroaniline 3a as well as 4-aminodiphenyl
amine 5a were major products from the reaction of 1a. Similarly,
azide 8a gave the corresponding aniline 10a and dimer 9a. It is
probable that the chloroanilines were formed by an electrophilic
substitution on aniline in which GaCl3 acted as Lewis acid and
supplied the chlorine atoms. The presence of the chloroanilines
and the rapid evolution of N2 suggested that aniline and 4-
methoxyaniline were formed early in the respective reactions.
Literature reports have indicated that anilines can be converted
to 4-aminodiaryl amines (the dimers) under various oxidative
conditions.9b,11,15 Mechanisms suggested for dimer formation from
anilines include:20 (i) initial formation of the radical cation
∑
+
Scheme 5 Mechanism of reductive formation of anilines, dimers and
trimers from aromatic azides.
ArNH2 which then couples with more aniline and forms the 4-
aminodiaryl amine radical cation after loss of HX, (ii) formation
∑
+
of the aniline radical ArNH which couples with ArNH3 , ArNH2,
evidently easier for the PhN3 system than the 4-MeO-analogue.
This might be because elimination of Cl- to afford 22 takes place
more easily than elimination of MeO-. Two such eliminations are
needed for the formation of 15.
or ArNH2 ∑. A possible mechanism for production of the anilines
+
is shown in Scheme 5. Coordination of the GaCl3 to 1a or 8a will
give rise to an adduct 17 that can be reduced by more azide to
give, after nitrogen loss, a gallium-coordinated aminyl radical 18.
The H-atom source is probably mainly solvent (SH), rather than
the aromatic rings, because even the fully deuteriated precursor
1d gave intermediate radical cations containing H-atoms. Thus 18
picks up an H-atom to produce 19 which will protonate to yield
We have shown that two aromatic azides react with gallium
trichloride in hydrocarbon solvents to yield deeply coloured solu-
tions of very persistent radical cations. Well-resolved EPR spectra
of these species were eventually obtained and analysed. The exper-
imental isotropic hfs, coupled with DFT computations, showed
these were radical cations of the dimer 4-aminodiphenylamine
and trimer (N1-(4-aminophenyl)-N4-phenylbenzene-1,4-diamine)
and these species were comprehensively characterised. ENDOR
spectra of the frozen solutions provided supporting information.
It appears that the GaCl3 promoted reactions of azides resemble
those of aromatic amines under oxidative conditions. In CH2Cl2–
pentane solvent the conditions are just right for initial conversion
of the azides to anilines which then couple to give dimers, trimers
and, eventually, polyanilines.
-
the aniline [2a, 3a or 10a] and GaCl4 .
There are several potential routes to the dimer radical cations.
Radical 18 might add to the aniline with production of delocalised
radical 21. Elimination of GaCl3X- would then yield radical
22 which on protonation would afford the observed long-lived
dimer radical cations 14. Of course, proton transfer could occur
earlier in the reaction, such that coupling takes place with the
anilinium cation instead. The presence of much 4-chloroaniline
in the reaction with 1a means that coupling with 18 could take
place with 3a as well as 2a so there is potential in this system
for formation of both the 4-aminodiphenyl amine radical cation
14a and N1-(4-chlorophenyl)benzene-1,4-diamine 14b and/or the
corresponding trimers 15a,b. The EPR spectrum of species (ii)
from 1a showed unequivocally that 14a with no Cl substituent
was formed (see Table 3). However, in the case of the trimer
radical cations 15a,x–z, hfs were not observable from all sites so the
presence of a Cl-substituent cannot be ruled out. Trimerisation was
Experimental
EPR and ENDOR spectroscopy
EPR spectra were obtained with a Bruker EMX 10/12 spectrom-
eter fitted with a rectangular ER4122 SP resonant cavity and
operating at 9.5 GHz with 100 kHz modulation. An aliquot (ca.
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 5097–5104 | 5103
©

View Article Online
0.1 mL) of the reaction mixture from each aromatic azide and
GaCl3 in CH2Cl2–pentane solution was placed in a 1 mm o.d.
quartz capillary tube, de-aerated by bubbling nitrogen for 20 min
and transferred to the resonant cavity. Spectra were examined
at several temperatures but generally best resolution and signal
intensity were obtained at around 300 K. Most of the EPR spectra
were recorded with 2.0 mW power, 1.0 to 0.2 Gpp modulation
intensity and a gain of ca. 106. In all cases where spectra were
obtained, hfs were assigned with the aid of computer simulations
using the Bruker SimFonia and NIEHS Winsim2002 software
packages.
Pulsed EPR and ENDOR were performed using a pulsed
EPR spectrometer (Bruker Elexsys E580) equipped with a Dice-
ENDOR accessory, a radio frequency (rf) amplifier and a
dielectric-ring ENDOR resonator (Bruker EN4118X-MD-4-W1).
Samples were maintained at 50 K using liquid helium in an
Oxford CF-935 cryostat. Field-swept electron spin echos (ESE)
were recorded using a 2-pulse ESE sequence while ESE-ENDOR
experiments were carried out using Davies three-pulse sequence
p-T-p/2-t-p-echo with a selective rf pulse of variable frequency
applied during time T. The pulse length used were 128 and 256 ns
for p/2 and p respectively, and 10 ms for p-rf pulse. ENDOR data
was processed and simulated using the EasySpin package (free
download from http://www.easyspin.org/ web page).
and K. Oshima, Org. Lett., 2002, 4, 2993; (f) S.-I. Usugi, T. Tsuritani,
H. Yorimitsu, H. Shinokubo and K. Oshima, Bull. Chem. Soc. Jpn.,
2002, 75, 841; (g) K. Takami, S. Mikami, H. Yorimitsu, H. Shinokubo
and K. Oshima, Tetrahedron, 2003, 59, 6627; (h) K. Takami, S. Mikami,
H. Yorimitsu, H. Shinokubo and K. Oshima, J. Org. Chem., 2003, 68,
6627; (i) N. Hayashi, I. Shibata and A. Baba, Org. Lett., 2004, 6, 4981;
(j) K. Takami, H. Yorimitsu and K. Oshima, Org. Lett., 2004, 6, 4555;
(k) N. Hayashi, I. Shibata and A. Baba, Org. Lett., 2005, 7, 3093; (l) K.
Takami, S. Usugi, H. Yorimitsu and K. Oshima, Synthesis, 2005, 5,
824; (m) N. Hayashi, H. Honda, M. Yasuda, I. Shibata and A. Baba,
Org. Lett., 2006, 8, 4553; For recent general reviews on use of Group 13
organometallic compounds in organic synthesis, see: (n) V. Nair, S. Ros,
C. N. Jayan and B. S. Pillai, Tetrahedron, 2004, 60, 1959; (o) S. Araki and
T. Hirashita, in Comprehensive Organometallic Chemistry III, 2007, vol.
9, pp. 649–751; (p) M. Yamaguchi and Y. Nishimura, Chem. Commun.,
2008, 35; (q) Main Group Metals in Organic Synthesis, H. Yamamoto
and K. Oshima, ed., Wiley-VCH, Weinheim, 2004.
7 (a) L. Benati, G. Bencivenni, R. Leardini, D. Nanni, M. Minozzi, P.
Spagnolo, R. Scialpi and G. Zanardi, Org. Lett., 2006, 8, 2499; (b) G.
Bencivenni, T. Lanza, M. Minozzi, D. Nanni, P. Spagnolo and G.
Zanardi, Org. Biomol. Chem., 2010, 8, 3444.
8 As far as we know, the only example of reaction between an azide and
a Group 13 metal trihalide was reported in 1987 by Takeuchi et al. (H.
Takeuchi, M. Maeda, M. Mitani and K. Koyama, J. Chem. Soc., Perkin
Trans. 1, 1987, 57), who noted that aromatic azides, when treated in
aromatic solvents in the presence of AlCl3, give rise to a decomposition
reaction entailing formation of extremely reactive N-containing species
able to give aromatic substitution reactions. He suggested that those
species were AlCl3-complexed arylnitrenes ensuing from aryl azide–
AlCl3 complexes. Takeuchi also marginally observed that by mixing
the aryl azide and AlCl3 in CH2Cl2 at 243 K a strong blue colour
developed; this colour faded completely, with concomitant nitrogen
gas evolution, when the solution was warmed to 273 K. However, no
product analysis of the resulting mixture was carried out and no other
mechanistic or synthetic data were subsequently obtained about Lewis
acid-catalysed decompositions of azides.
Acknowledgements
We thank EaStChem and the EPSRC (UK Basic Technology
Programme grant GR/S85726/01) for financial assistance. We
also acknowledge financial support from MIUR, Italy (2008
PRIN funds for “Properties and reactivity of free radicals in
complex environments and their role in oxidative processes and
in organic synthesis”).
9 (a) A. R. Forrester, J. M. Hay and R. H. Thomson, In Organic Chemistry
of Stable Free Radicals, Academic Press, New York, 1968, pp. 254; (b) R.
Male and R. D. Allendorfer, J. Phys. Chem., 1988, 92, 6237.
10 (a) J. F. Wolf, C. E. Forbes, S. Gould and L. W. Shacklette, J.
Electrochem. Soc., 1989, 136, 2887; (b) M. M. Wienk and R. A. Janssen,
J. Am. Chem. Soc., 1996, 118, 10626.
11 A. Petr and L. Dunsch, J. Phys. Chem., 1996, 100, 4867.
12 Intense colours were also observed by two-electron oxidation of
Variamine blue in aqueous media with oxidising agents such as Ce(IV),
Fe(III), I2, and Tl(III); see, for example: T. Imamura and M. Fujimoto,
Bull. Chem. Soc. Jpn., 1972, 45, 442–444.
References
1 S. Bra¨se, D. Gil, K. Knepper and V. Zimmermann, Angew. Chem., 2005,
117, 5320; S. Bra¨se, D. Gil, K. Knepper and V. Zimmermann, Angew.
Chem., Int. Ed., 2005, 44, 5188.
2 M. Minozzi, D. Nanni and P. Spagnolo, Chem.–Eur. J., 2009, 15, 7830
and references therein.
3 (a) L. Benati, P. C. Montevecchi and P. Spagnolo, Tetrahedron Lett.,
1978, 19, 815; (b) L. Benati and P. C. Montevecchi, J. Org. Chem., 1981,
46, 4570; (c) S. Kim, G. H. Joe and J. Y. Do, J. Am. Chem. Soc., 1994,
116, 5521.
4 (a) G. Bencivenni, T. Lanza, R. Leardini, M. Minozzi, D. Nanni, P.
Spagnolo and G. Zanardi, J. Org. Chem., 2008, 73, 4721; (b) T. Lanza, R.
Leardini, M. Minozzi, D. Nanni, P. Spagnolo and G. Zanardi, Angew.
Chem., Int. Ed., 2008, 47, 9439.
13 ESI analysis also showed another product with mass 243 amu, which
could be related to some unidentified intermediate where loss of the
methoxy group has still to occur.
14 Gallium trichloride also gives rise to a peak for GaCl4-. Our conclusion
-
remained tentative because the GaCl4 ion could have come from
residual reactant.
15 (a) I. Streeter, A. J. Wain, M. Thompson and R. G. Compton, J. Phys.
Chem., 2005, 109, 12636; (b) L. Omelka, S. Ondrasova, L. Dunsch, A.
Petr and A. Stasko, Monatsh. Chem., 2001, 132, 597.
16 L. Erdey, T. Mesiel, B. Mohos and F. Tu¨do¨s, Talanta, 1967, 14, 1477.
17 S. Stoll, B. Epel, S. Vega and D. Goldfard, J. Chem. Phys., 2007, 127,
164511.
5 (a) P. A. Baguley and J. C. Walton, Angew. Chem., Int. Ed., 1998,
37, 3072; (b) A. Studer and S. Amrein, Synthesis, 2002, 835; (c) V.
Darmency and P. Renaud, Top. Curr. Chem., 2006, 263, 71.
6 (a) T. Miyai, K. Inoue, M. Yasuda, I. Shibata and A. Baba, Tetrahedron
Lett., 1998, 39, 1929; (b) K. Inoue, A. Sawada, I. Shibata and A.
Baba, Tetrahedron Lett., 2001, 42, 4661; (c) S. Mikami, K. Fujita, T.
Nakamura, H. Yorimitsu, H. Shinokubo, S. Matsubara and K. Oshima,
Org. Lett., 2001, 3, 1853; (d) K. Inoue, A. Sawada, I. Shibata and A.
Baba, J. Am. Chem. Soc., 2002, 124, 906; (e) K. Takami, H. Yorimitsu
18 Gaussian 03, Revision A.1, M. J. Frisch, et al., Gaussian, Inc., Pittsburgh
PA, 2003 (see ESI for full reference).
19 For the B3LYP/EPR-iii//B3LYP/6-31G(d) basis set, see: V. Barone,
In Recent Advances in Density Functional Theory, D. P. Chong, Ed.;
World Scientific Publishing Co., Singapore, 1996.
20 For further mechanistic studies see: (a) A. Petr and L. Dunsch, J.
Electroanal. Chem., 1996, 419, 55; (b) P. Simon, G. Farsang and C.
Amatore, J. Electroanal. Chem., 1997, 435, 65; (c) M. Goto, K. Otsuka,
X. Chen, Y. Tao and M. Oyama, J. Phys. Chem. A, 2004, 108, 3980.
5104 | Org. Biomol. Chem., 2010, 8, 5097–5104
This journal is
The Royal Society of Chemistry 2010
©