Gas-phase cyclisation reactions of 1-(2-arylaminophenyl)alkaniminyl radicals

Article abstract of DOI:10.1039/a800886h

Flash vacuum pyrolysis (FVP) of the oxime ethers 9-11 at 650°C (10^-2-10^-3 Torr) gives products such as the nitrile 17, carbazoles 19 and 20 and acridines 18 and 21 derived from the corresponding iminyl radicals 13-15. The mechanism proposed for the formation of the acridines involves a key hydrogen abstraction by the iminyl of the adjacent N-H atom. When this route is blocked by an N-methyl group, as in 12, alternative cyclisations ensue, yielding the dihydroquinazoline 26 (via another hydrogen abstraction process) and the benzimidazole 25 (via an iminyl-imidoyl interconversion).

Full text of DOI:10.1039/a800886h

View Article Online / Journal Homepage / Table of Contents for this issue
Gas-phase cyclisation reactions of 1-(2-arylaminophenyl)alkaniminyl
radicals
Rino Leardini,a Hamish McNab,*,b Daniele Nanni,a Simon Parsons,b
David Reedb and Anton G. Tenana
a Dipartimento di Chimica Organica ‘A Mangini’, Università di Bologna, Viale Risorgimento 4,
I-40136, Bologna, Italy
b Department of Chemistry, The University of Edinburgh, West Mains Road, Edinburgh,
UK EH9 3JJ
Flash vacuum pyrolysis (FVP) of the oxime ethers 9–11 at 650 ЊC (10Ϫ2–10Ϫ3 Torr) gives products such
as the nitrile 17, carbazoles 19 and 20 and acridines 18 and 21 derived from the corresponding iminyl
radicals 13–15. The mechanism proposed for the formation of the acridines involves a key hydrogen
abstraction by the iminyl of the adjacent N᎐H atom. When this route is blocked by an N-methyl group,
as in 12, alternative cyclisations ensue, yielding the dihydroquinazoline 26 (via another hydrogen
abstraction process) and the benzimidazole 25 (via an iminyl–imidoyl interconversion).
In an accompanying paper,1 we have reported the reactions of
R2
N
R2
N
R2
N
1-(2-aryloxyphenyl)- and 1-(2-arylaminophenyl)-alkaniminyl
radicals generated in solution. Our corresponding work on
1-(2-aryloxyphenyl)alkaniminyls 1 generated in the gas phase
by flash vacuum pyrolysis (FVP) has led to the observation
of interconversion processes relating these iminyls and
2-(aryliminomethyl)phenoxyl radicals 2 (Scheme 1, X = O).2
Ph
O
Ph
Ph
N•
NOMe
R1
R1
R1
5 R1 = R2 = H
9 R1 = R2 = H
13 R1 = R2 = H
6 R1 = Me, R2 = H
7 R1 = Ph, R2 = H
8 R1 = H, R2 = Me
10 R1 = Me, R2 = H
11 R1 = Ph, R2 = H
12 R1 = H, R2 = Me
14 R1 = Me, R2 = H
15 R1 = Ph, R2 = H
16 R1 = H, R2 = Me
X
X•
N •
N
oxime ethers 9–12 also show common peaks at m/z 167, which
may be due to a carbazole-like species, after cleavage of the
entire oxime unit and cyclisation. Some of these processes are
reflected in the thermal behaviour of the compounds (see
below).
R
R
1 (X = O)
3 (X = NR)
2 (X = O)
4 (X = NR)
Scheme 1
Under our standard conditions for oxime ether pyrolysis,2,3
[650 ЊC (0.001–0.01 Torr)] the substrates 9–12 were transformed
completely into products, presumably via the iminyls 13–16
respectively. Mixtures of products were generally obtained, and
these were separated by chromatography on silica.
Here we discuss the properties of the 1-(2-arylaminophenyl)-
alkaniminyls 3 in the gas phase which surprisingly show quite
different cyclisation behaviour; in particular we have no
evidence for the interconversion of these species and 2-(arylimino-
methyl)aminyl radicals 4 (Scheme 1, X = NR).
The acetophenone derivative 10 showed the simplest pyrolys-
ate, being transformed exclusively into 2-(N-phenylamino)-
benzonitrile 17 (92%) (Scheme 2). FVP of the oxime ether
As before,2,3 the iminyls were generated by FVP of oxime
O-methyl ethers and we chose to study three series of radicals,
bearing respectively a hydrogen atom, a methyl group and a
phenyl group on the iminyl carbon atom.1,2 The precursor alde-
hydes 54 and 8 were obtained in four-step McFadyen–Stevens
sequences starting from N-phenylanthranilic acid and N-methyl-
N-phenylanthranilic acid respectively, whereas the ketones 6
and 7 were made by standard methods involving copper-
catalysed N-arylation of 2-aminoacetophenone5 and 2-amino-
benzophenone6 respectively with iodobenzene. The oxime
ethers 9–12 were made in 70–92% yield by condensation of the
carbonyl compounds 5–8 with O-methylhydroxylamine hydro-
chloride in the presence of pyridine.
NHPh
(i)
10
CN
17 (92%)
Scheme 2 Reagents and conditions: (i) 650 ЊC, 0.005 Torr
9 gave three products which were readily identified as 9-amino-
acridine 18 (40%), 2-(N-phenylamino)benzonitrile 17 (40%)
and a trace of 2-(N-phenylamino)benzaldehyde 5 (Scheme 3).
The mass spectra of all the oxime ethers 9–12 show loss of
m/z 31 (MeO) from their molecular ions. The ketone derivatives
10 and 11 both show significant peaks at m/z 194, correspond-
ing to subsequent loss of the substituent attached to the oxime
carbon atom, whereas the aldehyde derivatives 9 and 12 display
a loss of m/z 15 as the next breakdown. Whereas this is prob-
ably due to cleavage of the N-methyl group in 12 (giving rise to
the base peak at m/z 194) this is not possible for 9 and the peak
is most likely due to loss of an NH unit (m/z 180, 30%). The
NH2
(i)
9
17 (40%) + 5 (trace) +
N
18 (40%)
Scheme 3 Reagents and conditions: (i) 650 ЊC, 0.003 Torr
J. Chem. Soc., Perkin Trans. 1, 1998
1833

View Article Online
deduced on the basis of previous work with iminyls. Thus the
nitrile 17, which is common to all three precursors 9, 10 and
11, is formed by β-cleavage of a hydrogen atom, a methyl
group and a phenyl group respectively from the iminyls 13–15
(Scheme 5, route a). Since such β-cleavage of a methyl group is
NHPh
C
N
17
a
– R •
NHPh
N•
b
– RCN
(R = Ph)
R
H
N
H
N
13 R = H
14 R = Me
15 R = Ph
•
22
19
Scheme 5
Fig. 1 A view of the acridine 21 showing the crystallographic number-
known to be facile relative to that of hydrogen atoms or aryl
groups,2,8 it is not surprising that this mechanism dominates the
behaviour of the acetophenone derivative 10. The alternative
cleavage of an RCN group (Scheme 5, route b), which leads
ultimately to carbazole 19 via cyclisation of the resulting aryl
radical 22,9 is found only from the phenyl substituted iminyl
radical 15 where β-cleavage by the two routes a and b leads to
different aryl radicals. This route does not compete to any sig-
nificant extent in the pyrolyses of 9 or 10, which is in general
agreement with the results from correspondingly substituted
1-(2-aryloxyphenyl)alkaniminyls 1.2
ing scheme: displacement ellipsoids enclose 50% probability surfaces
Five products were obtained from the pyrolysis of the benzo-
phenone derivative 11, viz the nitrile 17 (45%), the benzo-
phenone 7 (6%), carbazole 19 (12%) and its 9-methanol deriv-
ative 20 (10%), and 9-phenylaminoacridine 21 (16%) (Scheme
4). The initial identification of the nitrile 17, the acridine 18 and
CH2O
The source(s) of the acridines 18 and 21 are of particular
interest, since acridines (though not the 9-amino derivatives)
are significant products from the generation of the iminyls 13–
15 in solution1 and no corresponding products are obtained
from the related 1-(o-phenoxyphenyl)alkaniminyls.1,2 It is clear
that the N᎐H atom is transferred at some stage of the process,
and we initially believed that 18 might be formed by a non-
radical mechanism involving a 1,5-hydrogen shift, electrocyclis-
ation and elimination (Scheme 6), similar in the early stages
i
+
11
17 (45%) + 7 (6%) +
N
H
N
CH2OH
19 (12%)
20 (10%)
+
NHPh
Ph
N
N
N
R
N
N
21 (16%)
H
N
Scheme 4 Reagents and conditions: (i) 650 ЊC, 0.005 Torr
OMe
NHOMe
H
R
R
NHOMe
the carbazoles 19 and 20 was made by mass spectrometry fol-
lowed by comparison of their NMR spectra and other physical
properties with those quoted in the literature (see Experimental
section). The carbazol-9-ylmethanol 20 was unexpected, and
may be formed by secondary reaction during work-up of the
carbazole 19 with formaldehyde (Scheme 4).7 Although we have
not previously observed products derived from formaldehyde in
oxime ether pyrolyses, this may be formed from the methoxyl
radicals co-produced with the iminyls by homolysis of the N᎐O
bond of the oxime ethers. The NMR spectra of the fraction
which ultimately proved to be 9-phenylaminoacridine 21
showed temperature dependent broad signals owing to unidenti-
fied exchange processes which meant that it could not be identi-
fied spectroscopically. This product was therefore characterised
unambiguously by X-ray crystallography (Fig. 1) (see below).
The carbonyl compounds 5 and 7 may be formed either by
adventitious hydrolysis of the oxime ethers in the FVP inlet
system during sublimation, or via (intermolecular) hydrogen
capture by the iminyl and hydrolysis of the resulting imine dur-
ing chromatography (cf. ref. 1). Otherwise the mechanism of
formation of most of the products of these pyrolyses can be
R = H – MeOH
N
H
NH2
NH
18
Scheme 6
to the formation of acridones from N-phenylanthranilic acid
esters.10 However, this route cannot readily account for the
production of 21 and we now believe that the mechanism of
Scheme 7 is most likely. In this case, the iminyl is involved in the
hydrogen transfer and the aminyl 23 thus obtained initiates the
cyclisation. Further evidence for intramolecular hydrogen
abstraction by iminyls is presented below, and the correspond-
ing intermolecular process has been observed under solution
conditions.1 The primary aminyl 24 so obtained can achieve
greater stability via a 1,2- (or possibly a 1,3-) shift of a hydrogen
1834
J. Chem. Soc., Perkin Trans. 1, 1998

View Article Online
7.16
Ph
N
•
N
H
7.40
N
H
H
N •
R
7.24
3%
H
R
• NH
6.79 H
H
R
NH
4.5%
H
H
N
13 R = H
15 R = Ph
23
24
7.21
H
N
R = H
5.26
H 6.84
R = Ph
7.30
H
H 8.18
N
N
•
2%
4.5%
H
H
1.5%
NH
NH2
•
Fig. 2 1H NMR data for the 1,2-dihydroquinazoline 26
– H •
suggests that they are attached to non-adjacent carbon atoms.
The characteristic triplet (2H), doublet (2H), triplet (1H)
pattern due to the meta, ortho and para protons respectively of
a phenyl ring could also be discerned at δH 7.40, 7.24 and 7.16.
The remaining signals were an equally characteristic doublet,
triplet, triplet, doublet pattern at δH 7.30, 7.21, 6.84 and 6.79
due to an ortho-disubstituted (fused) benzene ring. NOE
experiments (Fig. 2) related the imine proton to the doublet at
δH 7.30, which allowed the assignments of the other protons in
the fused ring as shown. Another key NOE experiment related
the methylene protons to the ortho protons of the phenyl group
(Fig. 2). The structure 26 is consistent with these observations
and in particular the tautomer 26Ј can be unambiguously
excluded.
N
•
18
H
NHPh
– H •
21
Scheme 7
atom (which may be intermolecular) giving the immediate pre-
cursor to 18, or a 1,2-shift of the phenyl group (via a neophyl-
type rearrangement) giving the immediate precursor to 21.
It is clear that N-methylation would block the formation of
acridines by these mechanisms and so the precursor 12 was
synthesised in the hope that the pyrolysate mixture might be
simplified. In the event, although no acridines were obtained,
the availability of hydrogen atoms in the methyl group for
hydrogen transfer caused further complications and the benz-
imidazole 25 (m/z 194) (18%) and the dihydroquinazoline 26
(m/z 208) (24%) were obtained along with the expected nitrile
27b (20%) (Scheme 8). The component with m/z 194 was
The formation of the dihydroquinazoline 26 clearly requires
functionalisation of the N-methyl group of the substrate 12,
and this is most likely achieved by hydrogen atom abstraction
by the iminyl (Scheme 9). A 6-endo-trig cyclisation onto the
Ph
H
N
– H •
H
26
N
•
H
Ph
N
Ph
N
H
CH2
CH2
•
27a
Ph
N
Me
N
H
N•
N
H
N
N
H
(i)
Ph
H
+
+
12
H
H
N
Ph
N
CN
16
Ph
25 (18%)
26 (24%)
27b (20%)
12
NH•
Scheme 8 Reagents and conditions: (i) 650 ЊC, 0.005 Torr
H
29
Ph
Ph
N
Ph
Ph
N
N
N
N
N
– Me•
Me
N•
Me
Me
N
•
25
N
CH•
N
Ph
H
H
H
28
26′
16
initially thought to be the indazole 28, but this structure was
readily excluded on the basis of its NMR spectra.11 In particu-
lar the 13C NMR spectrum of 1-phenylindazole shows no
signals at δC >138 whereas the observed peaks at δC 143.70
(quaternary) and 141.96 are in excellent agreement with those
known for 25 (δC 143.7 and 142.3 respectively11). Product 27b
was also easily identified (see Experimental section), but the
structure of the unusual quinazoline 26 was deduced from the
following NMR experiments. The 1H NMR spectrum (360
MHz) showed a slightly broadened imine triplet at δH 8.18
which was shown by decoupling to interact with a methylene
doublet at δH 5.26 (J 1.9 Hz). The size of the coupling constant
supports the idea that these protons are in the same ring, but
Scheme 9
aldimine function generates the quinazoline ring system which
leads to 26 by loss of a hydrogen atom. The driving force for the
6-endo-trig mode of cyclisation is presumably the formation of
the stabilised benzyl radical 27a rather than the primary aminyl
29 obtained by the alternative 5-exo-trig ring closure.
The source of the benzimidazole 25 is also of interest; in this
case a skeletal rearrangement and loss of a C1 unit has taken
place. A number of mechanisms can be drawn for this trans-
formation, of which the most likely is an iminyl to imidoyl
radical interconversion followed by cyclisation and ejection of
the N-methyl group (Scheme 9). It is perhaps surprising that
J. Chem. Soc., Perkin Trans. 1, 1998
1835

View Article Online
such neophyl-type rearrangements of iminyls to imidoyls has
not been previously observed, and experiments are in progress
to probe the feasibility of this step in more detail. At this stage,
we cannot completely exclude the possibility that the primary
product of the reaction is the indazole 28 which rearranges
under the reaction conditions to the benzimidazole 25; con-
version of certain indazoles to benzimidazole derivatives is
known under photochemical conditions.12 However, FVP of
1-adamantylindazole under comparable conditions to ours
rearranges only to its 2-isomer and to 2-(N-adamantylamino)-
benzonitrile, with no benzimidazoles being formed.13 In
our case, the corresponding product [2-(N-phenylamino)benzo-
nitrile 17] was not found in the pyrolysate from 12 and so we
believe that the indazole 28 is probably not involved in our
transformations. A third possible mechanism involving initial
loss of the N-methyl group is also unlikely because such homo-
lysis is only partially complete even at 900 ЊC in diarylamine
derivatives using our apparatus, compared with the furnace
temperature of 650 ЊC used in the present study.9 In any event,
these three possible mechanisms differ only in the timing of the
loss of the N-methyl group.
organic fractions were dried (MgSO4) and the solvent was
removed to give the oxime (1.04 g, 92%), mp 75–77 ЊC (from
isopropyl alcohol) (Found: C, 74.25; H, 6.2; N, 12.45.
C14H14N2O requires C, 74.35; H, 6.2; N, 12.4%); δH 9.02 (1H, br
s), 8.22 (1H, s), 7.40–7.03 (8H, m), 6.79 (1H, m) and 3.98 (3H,
s); δC 151.83, 143.71 (q), 141.12 (q), 132.73, 130.14, 129.24,
122.93, 121.60, 117.74, 115.61 (q), 113.20 and 62.05; m/z 226
(Mϩ, 82%), 195 (100), 180 (30), 168 (18) and 167 (46).
2Ј-(N-Phenylamino)acetophenone O-methyloxime 10
A solution of 2Ј-(N-phenylamino)acetophenone 65 (1.06 g, 5
mmol) and O-methylhydroxylamine hydrochloride (0.83 g, 10
mmol) was heated under reflux in ethanol (25 cm3) in the pres-
ence of pyridine (0.79 g, 10 mmol) for 6 h. A similar work-up to
that described for 9 gave the oxime 10 (0.98 g, 82%), mp 85–
86 ЊC (from methanol) (Found: C, 74.8; H, 6.9; N, 11.5.
C15H16N2O requires C, 75.0; H, 6.65; N, 11.65%); δH 9.48 (1H,
br s), 7.46–6.84 (9H, m), 4.02 (3H, s) and 2.31 (3H, s); δC 156.61
(q), 142.35 (q), 142.10 (q), 129.05, 129.18, 121.78, 121.18,
120.01, 120.67 (q), 118.38, 115.48, 61.92 and 13.53; m/z 240
(Mϩ, 100%), 209 (87), 194 (16), 182 (8) and 167 (46).
Finally, the crystal structure of the 9-(phenylamino)acridine
21 will be briefly discussed (Fig. 1). Previous crystallographic
studies of 9-(arylamino)acridines have focused on derivatives
with substituted N-aryl groups which have anticancer proper-
ties14,15 and our compound is therefore the parent of this
important class of materials. The acridine moiety of 21 is
approximately planar, with the maximum deviation [at C(7)]
being 0.145 Å. However, a minor ‘butterfly’ distortion about the
C(9)᎐N(10) axis is observed, similar to that which has been
previously noted for other 9-(arylamino)acridines.15 The plane
of the phenyl group forms an angle of 67.4Њ with the best plane
of the acridine. The acridine ring is highly symmetric about a
plane through C(9)᎐N(10), but the bonds linking C(1)᎐C(2)
and C(3)᎐C(4) [and C(5)᎐C(6) and C(7)᎐C(8)] are, at a mean of
1.359(2) Å, significantly shorter than the other bonds in the
fused benzene rings [mean 1.426(2) Å]. This feature is also
common to the other acridines.14,15 Bond angles in 21 are all
close to 120Њ, with the largest deviations (up to ϩ3.5Њ and
Ϫ2.5Њ) being associated with angles in the central ring. The
intermolecular packing of 21 involves the formation of chains
2-(N-Phenylamino)benzophenone O-methyloxime 11
A solution of 2-(N-phenylamino)benzophenone 76 (1.6 g, 5.8
mmol) and O-methylhydroxylamine hydrochloride (1.08 g, 5.8
mmol) was heated under reflux in a mixture of ethanol (30 cm3)
and pyridine (15 cm3) for 2 days. A similar work-up to those
described above gave the oxime 11 (1.61 g, 92%), as a yellow oil
which was a mixture of Z and E isomers (Found: Mϩ, 302.1434.
C20H18N2O requires M, 302.1419); δH 9.80 (1H, br s), 7.82–6.74
(28H, m), 6.08 (1H, br s), 4.20 (3H, s) and 4.09 (3H, s) (13C
NMR spectrum not quoted because of its complexity); m/z 302
(Mϩ, 93%), 271 (52), 269 (100), 195 (48), 194 (40), 167 (54) and
77 (40).
2-(N-Methyl-N-phenylamino)benzaldehyde 8
This compound was made by McFadyen–Stevens degradation
of N-methyl-N-phenylanthranilic acid16 without full character-
isation of the intermediates involved. Thus, N-methylation of
N-phenylanthranilic acid was achieved by heating overnight
with an excess of iodomethane in dilute sodium hydroxide solu-
tion.16 The resulting mixture of N-methyl-N-phenylanthranilic
acid (ca. 65%) and starting material (both partially esterified)
was hydrolysed to the corresponding acids under strongly basic
conditions, and these precipitated on acidification. Without
purification, this mixture (18.69 g, ca. 0.088 mol) was esterified
overnight using iodomethane (14.2 g, 0.1 mol) in dimethyl-
formamide (250 cm3) in the presence of potassium carbonate
(12.2 g, 0.088 mol). After the usual work-up,2 the resulting mix-
ture of esters was separated using column chromatography on
silica [using hexane and ethyl acetate (2.5%) as eluents] to give
methyl N-methyl-N-phenylanthranilate (8.5 g, 40%) as a yellow
oil, δH 7.85 (1H, m), 7.55 (1H, m), 7.34–7.18 (4H, m), 6.82–
6.68 (3H, m), 3.62 (3H, s) and 3.33 (3H, s); δC 167.04 (q), 148.84
(q), 147.73 (q), 132.91, 131.03, 128.84 (q), 128.64, 128.52,
124.84, 117.63, 113.87, 51.62 and 40.00; m/z 241 (Mϩ, 100%),
210 (27), 208 (30), 195 (10), 182 (23), 180 (30), 167 (23) and 106
(13).
This ester (7.5 g, 0.031 mol) was then heated under reflux
with hydrazine hydrate (2.35 g, 0.047 mol) for 10 h, cooled and
washed twice with water. The thick oily residue soon solidified
and was recrystallised from ethanol to give the hydrazide (6.90
g, 92%), mp 88–90 ЊC, δH 9.35 (1H, br s), 8.09 (1H, m), 7.47–
7.04 (5H, m), 6.86–6.57 (3H, m), 4.18 (2H, br s) and 3.11 (3H,
s); δC 166.46 (q), 148.66 (q), 147.80 (q), 132.48, 130.72 (q),
128.69, 127.35, 126.11, 119.74, 115.83, 113.78 and 40.44; m/z
241 (Mϩ, 49%), 210 (100), 195 (10) and 180 (10).
1
1
–
–
via hydrogen bonding between N(10) and H(1Ј) (2 ϩ x, Ϫ2 Ϫ y,
z) [N(10)᎐N(1Ј) 2.975(2) Å].
In conclusion, this work has shown that in the gas phase,
the cyclisation reactions of 1-(2-arylaminophenyl)alkaniminyl
radicals are controlled by hydrogen transfer processes from the
NH (if present) to give acridines, or from an N-alkyl group (if
present) to give benzimidazole or quinazoline ring systems.
Although the results have raised interesting questions—such as
the possibility of iminyl–imidoyl interconversions—none of the
cyclisation routes are particularly efficient and no evidence
for the anticipated 1-(2-arylaminophenyl)alkaniminyl–2-(aryl-
iminomethyl)aminyl radical interconversion was obtained
though this is known to take place in solution.1 Good evidence
for the analogous gas-phase rearrangement in the ‘oxygen’
series is presented in the accompanying paper.2
Experimental
1H and 13C NMR species are recorded at 250 (or 200) and 63 (or
50) MHz respectively for solutions in [2H]chloroform unless
otherwise stated. Coupling constants are quoted in Hz.
2-(N-Phenylamino)benzaldehyde O-methyloxime 9
A solution of 2-(N-phenylamino)benzaldehyde 54 (0.99 g, 5
mmol) and O-methylhydroxylamine hydrochloride (0.83 g, 10
mmol) was heated under reflux in ethanol (25 cm3) in the pres-
ence of pyridine (0.79 g, 10 mmol) for 1 h. The solvent was
removed under reduced pressure and the residue was par-
titioned between ether and dilute hydrochloric acid. The
The hydrazide (5.68 g, 0.024 mol) was tosylated by slow add-
ition of toluene-p-sulfonyl chloride (4.49 g, 0.024 mol) to a
solution in dry pyridine (10 cm3). The reaction temperature
1836
J. Chem. Soc., Perkin Trans. 1, 1998

View Article Online
was kept below 60 ЊC. The solution was left to stand at room
temperature for 2 h, and then poured into a mixture of ice and
dilute hydrochloric acid to give the toluene-p-sulfonohydrazide
as a thick oil which soon solidified. It was filtered, washed with
water and recrystallised from ethanol to give yellow crystals
(7.56 g, 80%), mp 148–150 ЊC, δH 10.56 (1H, br s), 7.95 (1H,
dd, 3J 7.8, 4J 1.7), 7.70–7.60 (2H, d, 3J 8.3), 7.49 (1H, td, 3J 7.5,
4J 1.7), 7.35–7.05 (7H, m), 6.97 (1H, m), 6.80–6.70 (2H, d,
3J 8.3), 3.19 (3H, s) and 2.36 (3H, s); δC 163.76 (q), 148.92 (q),
148.86 (q), 144.42 (q), 133.84, 133.39 (q), 131.29, 129.36,
129.08, 128.24, 127.69, 127.32 (q), 126.67, 121.05, 117.01, 41.28
and 21.51; m/z 395 (Mϩ, 12%), 210 (100), 195 (17), 180 (9), 167
(10), 91 (15) and 77 (10).
A solution of the toluene-p-sulfonohydrazide (6.45 g, 0.016
mol) in ethylene glycol (30 cm3) was degraded to the aldehyde
by addition of anhydrous sodium carbonate (4.02 g, 0.038 mol)
at 160 ЊC as fast as foaming allowed, according to the general
method of McFadyen and Stevens (cf. ref. 4). The reaction mix-
ture was maintained at that temperature for 4 min, cooled and
diluted with water to give a brown solid. Careful recrystallis-
ation from methanol gave 2-(N-methyl-N-phenylamino)benz-
aldehyde 8 as a soft yellow solid (2.36 g, 70%), mp 48–50 ЊC, δH
10.15 (1H, s), 7.95 (1H, m), 7.64 (1H, m), 7.37–7.16 (4H, m),
6.86–6.72 (3H, m) and 3.37 (3H, s); δC 191.19, 151.80 (q),
149.84 (q), 135.67, 132.46 (q), 129.07, 128.70, 127.50, 125.80,
118.94, 114.96 and 41.33; m/z 211 (Mϩ, 100%), 195 (11), 194
(45), 182 (68), 180 (20), 168 (93), 167 (83), 106 (24), 91 (30) and
77 (42). This aldehyde was characterised as its oxime ether (see
below).
114.04 and 98.33 (q); m/z 194 (Mϩ, 100%), 193 (28), 168 (10),
167 (15), 166 (11) and 77 (18).
Pyrolysis of 2Ј-(N-phenylamino)acetophenone O-methyloxime
10. (0.42 g, 1.8 mmol) 90 ЊC, 650 ЊC, 5 × 10Ϫ3 Torr, 2 h: the
sole significant product isolated after chromatography was
2-(N-phenylamino)benzonitrile 17 (0.32 g, 92%) (spectra as
above).
Pyrolysis of 2-(N-phenylamino)benzophenone O-methyloxime
11. (0.64 g, 2.1 mmol) 90–120 ЊC, 650 ЊC, 5 × 10Ϫ3 Torr, 3 h: a
chloroform-insoluble fraction was identified as carbazole 19
(0.05 g, 12%), mp 236–238 ЊC (lit.,20 245 ЊC), spectra identical
with those in the literature.21 The chloroform-soluble fraction
was separated by chromatography (eluted with hexane–ethyl
acetate) to give 2-(N-phenylamino)benzonitrile 17 (0.18 g,
45%), (spectra as above): carbazol-9-ylmethanol 20 (0.04 g,
10%), mp 130–132 ЊC (lit.,7 128–139 ЊC), δH 8.11–8.05 (2H, m),
7.55–7.20 (6H, m), 5.80 (2H, s) and 2.70 (1H, br s); δC 139.55
(q), 125.91, 120.47 (q), 120.34, 119.97, 108.73 and 66.40; m/z
197 (Mϩ, 32%), 180 (6), 168 (14), 167 (100), 166 (23) and 139
(13); 9-phenylaminoacridine 21 (0.09 g, 16%), crude mp 198–
200 ЊC (lit.,22 224 ЊC), m/z 270 (Mϩ, 100%), 269 (81), 268 (42)
and 135 (16). This compound showed broad peaks in its NMR
spectra due to exchange effects at room temperature, and so
it was characterised by X-ray crystallography (see below and
Discussion section). A small quantity of 2-(N-phenylamino)-
benzophenone 7 (6%) was also recovered.
Pyrolysis of 2-(N-methyl-N-phenylamino)benzaldehyde O-
methyloxime 12. (0.60 g, 2.5 mmol) 90–120 ЊC, 650 ЊC, 5 × 10Ϫ3
Torr, 2.5 h: 2-(N-methyl-N-phenylamino)benzonitrile 27b (0.10
g, 20%) (Found: Mϩ, 208.0999. C14H12N2 requires M, 208.1000);
δH 7.56 (1H, m), 7.43–7.11 (4H, m), 6.98–6.80 (3H, m),
6.67 (1H, m) and 3.39 (3H, s); δC(CH and CH3 signals only)
134.41, 133.73, 129.08, 125.51, 123.77, 120.90, 117.86 and
40.67; m/z 208 (Mϩ, 100%), 207 (51), 192 (13), 180 (10), 167
(8), 131 (25), 129 (14), 104 (12), 91 (19) and 77 (25); 1-
phenylbenzimidazole 25 (0.09 g, 18%); δH 8.06 (1H, s), 7.79–
7.90 (2H, m), 7.55–7.35 (5H, m) and 7.32–7.20 (2H, m); δC
143.70 (q), 141.96, 135.98 (q), 133.33 (q), 129.72, 127.69,
123.67, 123.38, 122.47, 120.23 and 110.16 (spectrum consist-
ent with literature data11); m/z 194 (Mϩ, 100%), 193 (10), 77
(11) and 51 (13); 1-phenyl-1,2-dihydroquinazoline 26 (0.12 g,
24%) (Found: Mϩ, 208.1008. C14H12N2 requires M, 208.1000);
δH 8.18 (1H, m), 7.46–7.32 (3H, m), 7.26–7.11 (4H, m),
6.85–6.78 (2H, m) and 5.26 (2H, d) (see also Fig. 1); δC
158.63, 144.60 (q), 143.62, 132.43, 129.16, 127.74, 124.23,
123.04, 119.70 (q), 119.23, 114.49 (q) and 61.81; m/z 208 (Mϩ,
55%), 207 (100), 195 (14), 180 (27), 168 (14), 167 (12) and 77
(26).
2-(N-Methyl-N-phenylamino)benzaldehyde O-methyloxime 12
A solution of 2-(N-methyl-N-phenylamino)benzaldehyde 8
(1.88 g, 8.9 mmol) and O-methylhydroxylamine hydrochloride
(1.49 g, 18 mmol) in ethanol (40 cm3) containing pyridine
(3 cm3) was heated under reflux for 1 h. After the standard
work-up, the oxime (1.88 g, 88%), mp 102–103 ЊC (from iso-
propyl alcohol) was obtained (Found: C, 74.7; H, 7.0; N, 11.55.
C15H16N2O requires C, 75.0; H, 6.65; N, 11.65%); δH 8.13 (1H,
s), 7.96 (1H, dd, 3J 7.0 and 4J 1.9), 7.45–7.13 (5H, m), 6.75 (1H,
apparent t, 3J 7.0), 6.59 (2H, m), 3.93 (3H, s) and 3.21 (3H, s);
δC 149.28 (q), 147.51 (q), 145.84, 131.36, 130.40, 128.89, 128.22
(q), 126.73, 126.48, 117.76, 113.70, 61.85 and 40.17; m/z 240
(Mϩ, 27%), 209 (91), 194 (100), 180 (69), 167 (14), 131 (11), 91
(13) and 77 (26).
Flash vacuum pyrolysis experiments
The substrate was sublimed at 0.01–0.001 Torr through a silica
tube which was maintained at the appropriate temperature by
an electrically heated tube furnace. Products were collected in a
U-tube trap cooled by liquid nitrogen located at the exit point
of the furnace and thereafter were usually separated by dry-
flash chromatography on silica. Results are presented in the
form quantity of substrate, inlet temperature, furnace temper-
ature, pressure, pyrolysis time and products.
Pyrolysis of 2-(N-phenylamino)benzaldehyde O-methyloxime
9. (0.37 g, 1.6 mmol) 60–90 ЊC, 650 ЊC, 3 × 10Ϫ3 Torr, 2 h: a
yellow solid was formed at the exit-point of the furnace which
could be dissolved in acetone. Removal of the solvent gave
9-aminoacridine 18 (0.13 g, 40%), mp 233–234 ЊC (lit.,17
233 ЊC), δH([2H6]DMSO) 8.38 (2H, dd), 7.81 (2H, dd), 7.63
(2H, ddd), 7.29 (2H, ddd) and 4.42 (2H, br s); δC([2H6]DMSO)
150.27 (q), 148.79 (q), 129.98, 128.61, 123.41, 121.61 and
112.98 (q), (spectrum consistent with literature data18); m/z 194
(Mϩ, 100%), 193 (15), 167 (6), 166 (9), 97 (7), 84 (7) and 66
(15); chromatography of the remainder of the pyrolysate gave a
trace of 2-(N-phenylamino)benzaldehyde 5 (0.019 g, 6%) and
2-(N-phenylamino)benzonitrile 17 (0.12 g, 40%), mp 55–57 ЊC
(lit.,19 50–52 ЊC), δH 7.50 (1H, m), 7.39–7.20 (3H, m), 7.17–7.09
(4H, m), 6.83 (1H, m) and 6.42 (1H, br s); δC 147.20 (q), 139.80
(q), 133.76, 132.93, 129.46, 124.02, 121.54, 119.09, 117.49 (q),
X-Ray crystallography
Crystal data. C19H14N2, M = 270.32, monoclinic, a =
12.2580(14), b = 8.6809(6), c = 13.0712(10) Å, β = 92.051(10)Њ,
U = 1390 Å3 [from 2θ values of 45 reflections (30 < 2θ < 39Њ)
measured at ω], T = 150 K, space group P21/a, graphite-
monochromated Mo-Kα radiation, λ = 0.710 73 Å, Z = 4,
Dc = 1.292 Mg mϪ3, F(000) = 568, yellow block with dimensions
0.70 × 0.51 × 0.31 mm3. Stoe Stadi-4 diffractometer equipped
with an Oxford Cryosystems low-temperature device, ω–θ
scans, data collection range 5 < 2θ < 50Њ, Ϫ14 < h < 14,
0 < k < 10, 0 < l < 15. Three standard reflections showed no
significant intensity variation; 2457 unique data.
Structure solution and refinement. The structure was solved
by direct methods (SIR92)23 and refined anisotropically by full-
matrix least-squares on F 2 (SHELXL-93).24 H-atoms were
located in a difference map and refined freely with isotropic
displacement parameters. The weighting scheme was wϪ1
=
2
σ2(Fo ) ϩ (0.0483P)2 ϩ 0.254P where 3P = (Fo2 ϩ 2Fc2). The
final wR(F2) was 9.34% (based on all 2455 data used for refine-
ment) and the conventional R(F) was 3.65% [based on 1915
data with F > 4σ(F)] (R-factors as defined in ref. 24) for 247
J. Chem. Soc., Perkin Trans. 1, 1998
1837

View Article Online
9 J. I. G. Cadogan, H. S. Hutchison and H. McNab, Tetrahedron,
1992, 48, 7747.
parameters. The final difference map extrema were ϩ0.18 and
Ϫ0.17 e ÅϪ3, and the final shift/esd was 0.001.
10 Y. Mao and V. Boekelheide, J. Org. Chem., 1980, 45, 1547.
11 M. Begtrup, J. Elguero, R. Faure, P. Camps, C. Estopá, D. Ilavský,
A. Fruchier, C. Marzin and J. De Mendoza, Magn. Reson. Chem.,
1988, 26, 134.
12 H. Tiefenthaler, W. Dörscheln, H. Göth and H. Schmidt, Helv.
Chim. Acta, 1967, 50, 2244.
13 J. D. Pérez, G. I. Yranzo, M. A. Ferraris, J. Elguero, R. M.
Claramunt and D. Sanz, Bull. Soc. Chim. Fr., 1991, 592.
14 Z. H. L. Abraham, S. D. Cutbush, R. Kuroda, S. Neidle, R. M.
Acheson and G. N. Taylor, J. Chem. Soc., Perkin Trans. 2, 1985, 461.
15 J. S. Buckleton and G. R. Clark, Acta Crystallogr., Sect. C, 1992, 48,
1085.
Full crystallographic details, excluding structure factor
tables, have been deposited at the Cambridge Crystallographic
Data Centre (CCDC). For details of the deposition scheme, see
‘Instructions for Authors’, J. Chem. Soc., Perkin Trans. 1, avail-
able via the RSC Web page (http://www.rsc.org/authors). Any
request to the CCDC for this material should quote the full
literature citation and the reference number 207/203.
Acknowledgements
16 H. Gilman and S. M. Spatz, J. Org. Chem., 1952, 17, 860.
17 A. Albert, The Acridines, Edward Arnold, London, 1951, p. 151.
18 R. Faure, J.-P. Galy, E.-J. Vincent, J. Elguero, A.-M. Galy and
J. Barbe, Chem. Scripta, 1980, 15, 62.
We are grateful to the EPSRC for the provision of a diffract-
ometer. This investigation was also supported by the University
of Bologna (1997 Funds for selected research topics).
19 M. R. Bryce, T. A. Dransfield, K. A. Kandeel and J. M. Vernon,
J. Chem. Soc., Perkin Trans. 1, 1988, 2141.
References
20 S. H. Tucker, J. Chem. Soc., 1926, 546.
1 G. Calestani, R. Leardini, H. McNab, D. Nanni and G. Zanardi,
J. Chem. Soc., Perkin Trans. 1, 1998, 1813.
2 M. Black, J. I. G. Cadogan, R. Leardini, H. McNab,
G. McDougald, D. Nanni, D. Reed and A. Zompatori, J. Chem.
Soc., Perkin Trans. 1, 1998, 1825.
3 J. I. G. Cadogan, C. L. Hickson and H. McNab, Tetrahedron, 1986,
42, 2135.
4 A. Albert, J. Chem. Soc., 1948, 1225.
5 D. Hellwinkel and P. Ittemann, Chem. Ber., 1986, 119, 3165.
6 cf. J. Itier and A. Casadevall, Bull. Soc. Chim. Fr., 1969, 2342.
7 V. P. Lopatinskii and E. E. Sirotkina, Metody Poluch. Khim. Reakt.
Prep., 1964, 88; (Chem. Abstr., 1969, 64, 15 823h).
8 K. Bird, A. W. K. Chan and W. D. Crow, Aust. J. Chem., 1976, 29,
2281.
21 J. Pouchert and J. Behnke, The Aldrich Library of 13C and 1H
FTNMR spectra, Edition 1, Volume 3, Aldrich Chemical Company
Inc., 1993, 160B.
22 Ref. 17, p. 156.
23 A. Altomare, M. C. Burla, M. Camalli, G. Cascarano, C.
Giacovazzo, A. Guagliardi and G. Polidori, J. Appl. Cryst., 1994, 27,
435.
24 G. M. Sheldrick, SHELXL-97, University of Göttingen, Germany,
1997.
Paper 8/00886H
Received 2nd February 1998
Accepted 1st April 1998
1838
J. Chem. Soc., Perkin Trans. 1, 1998