The synthesis of alkyl aryl nitriles from N-(1-arylalkylidene)-cyanomethylamines. Part 2. Mechanism

Article abstract of DOI:10.1039/b106259j

The mechanism of the rearrangement of N-(1-arylalkylidene)cyanomethylamines ArC(=NCH₂CN)R 1 to the corresponding nitriles ArCH(CN)R2 (in DMF, at 150°C, with K₂CO₃) is described. Reaction 1 → 2 was investigated for different types of imine 1, and it was found that with a leaving group other than CN^- the reaction does not proceed to yield the nitrile, whereas imines such as PhC(=NCH₂CN)H, prepared starting from aldehydes rather than ketones, yield the expected phenylacetonitrile even at temperatures as low as 120°C. Evidence for the mechanism comes from a study of the reactivity of the postulated intermediates: 2-cyano-3-phenylaziridine 4c, and 2,2-diphenyl-2H-azirine 5b. The route involving aziridine 4c is ruled out, since this compound does not react at all under the investigated conditions. The 2H-azirine 5b instead, yields the corresponding diphenylacetonitrile in DMF with K₂CO₃, at 150°C. The transformation seems to involve an initial deprotonation, followed by an intramolecular ring closure-CN elimination step, which yields the 2H-azirine. The azirine then isomerizes to the nitrile. Additional evidence for the intermediacy of the 2H-azirine, based on ¹H NMR monitoring of the reaction 1 → 2, is described. Finally, the results of a simple isotope exchange experiment provide a rationale for the previously observed scrambling of labels, and further confirm the proposed mechanism.

Full text of DOI:10.1039/b106259j

View Article Online / Journal Homepage / Table of Contents for this issue
The synthesis of alkyl aryl nitriles from N-(1-arylalkylidene)-
cyanomethylamines. Part 2. Mechanism
Alvise Perosa, Maurizio Selva* and Pietro Tundo
2
Dipartimento di Scienze Ambientali, Università Ca’ Foscari, Dorsoduro 2137-30123, Venezia,
Italy. E-mail: selva@unive.it; Fax: ϩ39-041-234 8620; Tel: ϩ39-041-234 8687
Received (in Cambridge, UK) 13th July 2001, Accepted 1st March 2002
First published as an Advance Article on the web 2nd April 2002
The mechanism of the rearrangement of N-(1-arylalkylidene)cyanomethylamines ArC(᎐NCH CN)R 1 to the
᎐
2
corresponding nitriles ArCH(CN)R 2 (in DMF, at 150 ЊC, with K2CO3) is described. Reaction 1
2 was investigated
for different types of imine 1, and it was found that with a leaving group other than CNϪ the reaction does not
proceed to yield the nitrile, whereas imines such as PhC(᎐NCH CN)H, prepared starting from aldehydes rather than
᎐
2
ketones, yield the expected phenylacetonitrile even at temperatures as low as 120 ЊC. Evidence for the mechanism
comes from a study of the reactivity of the postulated intermediates: 2-cyano-3-phenylaziridine 4c, and 2,2-diphenyl-
2H-azirine 5b. The route involving aziridine 4c is ruled out, since this compound does not react at all under the
investigated conditions. The 2H-azirine 5b instead, yields the corresponding diphenylacetonitrile in DMF with
K2CO3, at 150 ЊC. The transformation seems to involve an initial deprotonation, followed by an intramolecular ring
closure–CN elimination step, which yields the 2H-azirine. The azirine then isomerizes to the nitrile. Additional
evidence for the intermediacy of the 2H-azirine, based on 1H NMR monitoring of the reaction 1
2, is described.
Finally, the results of a simple isotope exchange experiment provide a rationale for the previously observed
scrambling of labels, and further confirm the proposed mechanism.
Introduction
We have recently reported a new reaction of N-(1-arylalkyl-
idene)cyanomethylamines ArC(᎐NCH CN)R 1:1 when 1 is
᎐
2
heated in DMF in the presence of potassium carbonate, it
undergoes a clean rearrangement to alkyl aryl nitriles [ArCH-
(CN)R, 2] (Scheme 1). The reaction proceeds only at relatively
Scheme 2
In compound 7, the aryl ring is not conjugated to the reactive
site, a fact not compatible with the large positive ρ value
obtained by the Hammett treatment. In addition, when imine 1
(with R = Ar = Ph) was deprotonated with n-BuLi in THF at
Ϫ90 ЊC, and then heated at the reflux temperature of THF
(67 ЊC), no reaction was observed. However, when THF was
replaced with DMF, and the solution refluxed at 150 ЊC, the
usual clean rearrangement to nitrile Ph2CH(CN) 2 was
observed.
Scheme 1 Rearrangement of N-(1-arylalkylidene)cyanomethylamines
to alkyl aryl nitriles.
high temperatures (T ≥ 150 ЊC), through the formal loss of
HCN, and both alkyl aryl-, substituted alkyl aryl-, and diaryl-
imines 1 can be used as reagents.2
(2) The cyano group of ArC(᎐NCH CN)R acted as the
᎐
2
leaving group. In fact cyanide was titrated at the end of
the reaction, and when it was replaced with other substituents
In a previous paper on the mechanism of the 1
2
rearrangement, some unequivocal mechanistic evidence was
collected through the investigation of the reaction kinetics and
through the study of the reactivity of N-alkylformimidoyl
(e.g. phenyl), reaction 1
2 did not take place.
(3) The process was mainly intramolecular and the cyano
functionality of nitriles 2 derived primarily from the methylene-
imino (᎐NCH –) fragment of 1. This was inferred by running a
cyanides [ArCH(N᎐CHCN)R],3 and of labeled imines [ArC-
᎐
᎐
2
(᎐NCH 13CN)R (*1) and ArC(᎐14NCH CN)R (1*)].4 In
series of reactions using isotopically labeled PhC(᎐NCH -
᎐
᎐
᎐
2
2
2
particular, it was concluded that:
13CN)CH , and PhC(᎐15NCH CN)CH , under the conditions
᎐
3
2
3
(1) the first, and rate-determining step of the mechanism was
deprotonation at the methylene of imines 1 to yield 1Ϫ. This was
derived from the Hammett free-energy relationship obtained
from the reactions of different p-substituted N-(1-arylethyl-
of Scheme 1. The isotope distribution in the two products
[PhCH(CH3)13CN] and [PhCH(CH3)C15N] showed enrichments
of 32 and 57%, respectively.
Accordingly, two possible reaction pathways were proposed
(Scheme 3). They involved the common initial formation of an
aziridine ring system, either followed by expulsion of CNϪ to
give a 2H-azirine, and isomerization to the nitrile (path a); or
by loss of cyanide through a symmetrical intermediate, and
rearrangement to the nitrile (path b).4 Although path a
appeared more elegant and would somewhat account for the
predominant source of CNϪ, the experimental evidence was
not conclusive in its favor. In the present work the proposed
Ϫ
Ϫ
idene)cyanomethylamines 1 (Ar = p-BrC6H4 , p-ClC6H4 ,
Ϫ
Ϫ
p-CH3C6H4 and p-CH3OC6H4 ; R = CH3). The reaction was
first order and fitted the Hammett equation (log k/ko = σρ),
giving a ρ value of 1.86, consistent with the fact that an electron
withdrawing group (EWG) on the aryl ring stabilizes the
developing negative charge in 1Ϫ. This observation rules out any
mechanisms initiated by an SN1Ј or SN2Ј isomerization, such as
the one depicted in Scheme 2.
DOI: 10.1039/b106259j
J. Chem. Soc., Perkin Trans. 2, 2002, 1033–1037
This journal is © The Royal Society of Chemistry 2002
1033

View Article Online
as well as an aldehyde derivative, PhC(᎐NCH CN)H, were
᎐
2
investigated.
Alternative X groups
In view of the proposed mechanism, whereby the CN moiety
of the reagent ArC(᎐NCH CN)R 1 is expelled during the
᎐
2
reaction, other groups able to stabilize the negative charge on
the methylene were considered in place of CN. Compounds
PhC(᎐NCH X)Me (X = CO Me,1 3a; CF , 3b) were prepared by
᎐
2
2
3
condensation of acetophenone with hydrochloride salts of
glycine methyl ester and of 2,2,2-trifluoroethylamine,† respec-
tively. The synthesis of the mesyl (X = Ms), tosyl (X = Ts), and
oxazole derivatives was attempted, with no success as of yet.‡
Imines 3a,b were made to react under the conditions of
Scheme 1, and would have been expected to yield the corre-
sponding nitrile PhCH(CN)CH3 2a. Compound 3a proved to
be completely unreactive: it was recovered unchanged without
any other reaction having taken place: a result likely due to the
poor leaving group ability of the –COOCH3 group.5 Compound
3b, possessing an electron withdrawing CF3 group able to
promote deprotonation and perhaps to act as a leaving group,6
did not yield the expected nitrile 2a either; however, it did
undergo base-promoted HF elimination to the corresponding
conjugated product [eqn. (1)].
(1)
Under the same reaction conditions, a similar product was
observed for PhCH(Me)N᎐CHCCl (3f ).§ In this case, com-
᎐
3
pound 3f underwent a 1,4-elimination of HCl to yield PhC-
(᎐NCH᎐CCl )Me (3g) as the product; behavior analogous to
᎐
᎐
2
that of base-promoted reactions of 7α-amino-1-oxacephems.7
On the whole, since no other groups but cyano were suitable
for this reaction, it appeared that the presence of electron with-
drawing groups capable of negative charge stabilization was not
enough to promote the examined reaction: the X group must
also have leaving group ability. This dual behavior of CN as (a)
negative charge stabilizer and (b) leaving group, has an analogy
in the benzoin condensation, where cyanide first promotes
α-deprotonation, and then acts as a leaving group.8
Imines derived from aldehydes
In order to investigate the possible effect of the alkyl group (R)
on the iminic carbon, analogs of imines 1 were considered.
Specifically, PhC(᎐NCH CN)H 1c (R = H) was prepared start-
᎐
2
ing from benzaldehyde following the usual procedure.1 When
reacted under the conditions of Scheme 1, compound 1c proved
much more reactive than methyl aryl imines 1, and it was con-
verted very quickly (10 min) into phenylacetonitrile 2c (66%).
Scheme 3 Two possible pathways for the 1
2 rearrangement.
intermediacy of an azirine in the mechanism of the reaction
2, was tested and the results are reported. The behavior
of representative examples of the proposed intermediates,
synthesized independently by other routes, is described. The
outcome of the reaction on substrates with different leaving
groups (other than cyano), as well as extension of the process
1
The double bond isomer of the reagent PhCH N᎐CHCN 1c
᎐
2
i
(24%) appeared as a by-product, implying that the isomeriz-
ation 1c 1ci—forbidden for ketone derivatives—took place
in this case. However, the rearrangement of 1c to 2c proved
feasible at a lower temperature (80 ЊC instead of 150 ЊC; con-
version: 84%; 2c: 78%, after 180 min), thereby drastically
decreasing 1ci formation (6%).
to aldehyde derivatives [PhC(᎐NCH CN)H] will be discussed.
᎐
2
Finally, a new experiment is described to explain the observed
isotope scrambling.
† Compounds 3a and 3b were prepared by the same general procedure
used for 1.1
Results and discussion
Before exploring the reactivity of aziridines and azirines—the
proposed intermediates—attention was focused on determining
how general the examined reaction could be. Thus, func-
‡ The syntheses of PhC(᎐NCH2X)Me (X = LG) were carried out analo-
᎐
gously to the above imines, starting from a ketone and the appropriate
amine. However, attempts to synthesize H2NCH2XؒHCl (X = Ms, Ts,
oxazole) proved fruitless.
tionalized imines of general formula PhC(᎐NCH X)R were
᎐
2
§ PhCH(Me)N᎐CHCCl3 was prepared, rather than its isomer PhC-
᎐
considered. To understand whether nitriles would still be
observable as products and consequently, whether azirines were
possible intermediates for such compounds, different X groups,
(᎐NCH2CCl3)Me, since its synthesis is easier and since it is known
᎐
that under the conditions of Scheme 1, R2CHN᎐CHCN undergoes
᎐
isomerization to yield R2C᎐NCH2CN.3
᎐
1034
J. Chem. Soc., Perkin Trans. 2, 2002, 1033–1037

View Article Online
Aziridines 4 and azirines 5
Direct and indirect methods were used to probe the inter-
mediacy of the three-membered heterocycles postulated for
both pathways a and b in the mechanism of Scheme 3. The
following section describes how the search for evidence of
the mechanism was approached through the study of these
compounds. Indirect tests were carried out by investigating the
reactivity of compounds 4 and 5 under the conditions of
Scheme 1. Direct evidence of an intermediate was sought
Fig. 1 Comparison between Scheme 4 and the benzoin condensation.
Reaction of 2H-azirine
2,2-Diphenyl-2H-azirine 5b—the proposed intermediate for
through NMR analysis of the reaction 1
2.
the rearrangement of PhC(᎐NCH CN)Ph 1b to diphenylaceto-
᎐
2
nitrile 2b—was synthesized,6,11 as shown in Scheme 6, by
Reaction of aziridines
For the synthesis of an intermediate in the mechanism shown in
Scheme 3, two general considerations led to the choice of
2-cyano-3-phenylaziridine 4c (Scheme 4), as a model com-
Scheme 4 Synthesis of 2-cyano-3-phenylaziridine 4c.
pound: (i) as described above, the rearrangement 1
2 occurs
more readily on the benzaldehyde-derived imine PhC(᎐NCH -
᎐
2
CN)H 1c with respect to ketone-derived imines; (ii) both
reaction pathways (a and b) of compounds 1 in Scheme 3,
require an aziridine as a common precursor, and 4c was one
of the more readily accessible synthetically.9 Thus 2-cyano-3-
phenylaziridine was obtained in the cis configuration by a
literature method,9 by reacting α,β-dibromocinnamonitrile with
dry ammonia. Once prepared, the aziridine 4c was made to
react under the same conditions as for 1c. No reaction what-
soever occurred [eqn. (2)] and the reagent was recovered
unchanged. This observation allowed us to rule out path b of
Scheme 3.
Scheme 6 Synthesis of 2,2-diphenyl-2H-azirine 5b.
irradiation of diphenylvinyl azide 6b. Since 2H-azirines are
quite air sensitive, once a 37% yield was reached,¶ the solvent
(CDCl3) was immediately removed with a stream of nitrogen,
and the mixture containing 5b was set to react at 150 ЊC with
K2CO3 in DMF. Rewardingly, the reaction yielded predomin-
antly diphenylacetonitrile 2b [85%, eqn. (3)] along with 15%
phenylindole, known to derive from the thermal treatment of
6b.6 This result supported the mechanism delineated in path a
of Scheme 3 for the transformation 1
2, although some
doubt still persisted as to whether the vinyl azide 6b, present in
the reaction mixture, might itself have generated diphenyl-
acetonitrile 2b under the same conditions. This hypothesis was
ruled out by treating compound 6b at 150 ЊC with K2CO3 in
DMF: benzophenone (74%) and phenylindole (23%) were the
only products.
(2)
Path a of Scheme 3 implies the initial deprotonation of the
reactant imine 1, followed by ring closure and CNϪ expulsion.
The proposed ring closure is thermal, 4-electron, conrotatory,
and allowed by the Woodward–Hoffmann rules.10 Scheme 5
(3)
Having established that nitriles 2 can originate from azirines
5, direct evidence for the intermediacy of azirines 5 in the over-
all rearrangement of Scheme 1 was sought. This was done by
1H NMR monitoring of the reaction of N-(diphenylmethylene)-
cyanomethylamine [PhC(᎐NCH CN)Ph, 1b] at 150 ЊC, in
᎐
2
DMF-d7, and with K2CO3. The reaction proceeded as expected,
with the disappearance of the 2H methylene singlet of 1b
(4.5 ppm) and the appearance of the 1H singlet of diphenyl-
acetonitrile 2b (6.0 ppm). The signal at 10.7 ppm—ascribable to
the 2H-azirine 5b—was detected, and remained at a constant
Scheme 5 Electrocyclic ring closure of 1Ϫ.
shows the allylic anion type HOMO of 1Ϫ, which controls the
course of the reaction, and its conrotatory closure to the inter-
mediate cyclized anion. In this case, the subsequent step of
cyanide elimination (path a of Scheme 3), to give the 2H-
azirine, must necessarily be favored over protonation at the
nitrogen under the investigated conditions, to avoid the
bottleneck posed by the aziridine of path b. The mechanistic
analogy is again with the benzoin condensation,8 where the
intermediate, with a negative charge on a heteroatom in the α
position with respect to the CN, prefers to eliminate cyanide,
rather than to protonate at the oxygen (Fig. 1). To test this
hypothesis the synthesis of an appropriate 2H-azirine was
undertaken.
intensity with respect to the starting material PhC(᎐NCH CN)-
᎐
2
Ph, 1b until the reaction was over (Fig. 2).
From
a mechanistic standpoint, the general scenario
appeared clear: path a of Scheme 3, shown in Scheme 4, was
¶ Yield calculated by acquiring a 1H NMR spectrum of the reaction
mixture, and by integrating the peak of the proton in the 2-position of
the 2H-azirine 5b (10.2 ppm),13 relative to that of the vinyl proton of 6b
(6.7 ppm). This experiment also provided direct confirmation of the
NMR resonance frequency of the azirine proton, which is reported in
the literature, but with some ambiguity on the synthetic procedure.
J. Chem. Soc., Perkin Trans. 2, 2002, 1033–1037
1035

View Article Online
The reaction is very sensitive to the nature of the reacting
imines. Imines derived from aldehydes react much faster than
those prepared from ketones, indicating that the proposed
intramolecular ring closure is influenced by steric factors. How-
ever, even the sterically undemanding methyl group in the α
position of imine 1 disfavors the intramolecular ring closure
to give the 2H-azirine ring, presumably by electron donating
effects.
The data available to date seem to indicate that leaving
groups other than CNϪ inhibit the reaction. Current efforts are
aimed at extending its scope to other substrates, and some
preliminary data indicate that compounds such as PhC(᎐
᎐
NCH2CO2CH3)H, derived from aldehydes, do in fact undergo
the same rearrangement to yield phenylacetonitrile. This obser-
vation may open the way to a more general application of the
reaction, and to its use with reagents possessing poorer leaving
groups. To this end, there are significant challenges in the syn-
theses of imines 1 with such characteristics, which may follow
the same rearrangement mechanism, thereby paving the way to
new synthetic applications.
Fig. 2 1H NMR of reaction PhC(᎐NCH2CN)Ph, 1b
Ph2CH(CN)
᎐
2b, collected at time intervals.
validated experimentally and offered a plausible pattern for the
rearrangement 1 2. However, for a definitive explanation,
one question remained to be answered. The intramolecular
reaction shown in path a implied that the cyano group of the
product nitrile 2 had to originate exclusively from the imine
Experimental
Imine PhC(᎐NCH CF )Me (3b) was obtained through the
᎐
2
3
condensation of acetophenone (2 g, 16.7 mmol) with 2,2,2-
trifluoroethylamine hydrochloride salt (CF3CH2NH2ؒHCl).1,10
Yield of distilled product (bp = 65 ЊC/0.8 mmHg): 1.8 g (54%).
1H NMR (CDCl3) δ 7.87–7.83 and 7.45–7.35 (m, 4 H, Ph), 4.00
(qq, 2H, J = 9.6 Hz, JЈ = 0.98 Hz), 2.28 (t, 3H, CH3, JЈ = 0.98
Hz). Mass spectrum (70 eV) m/z (relative intensity): 201 (Mϩ,
30%), 200 (32), 187 (10), 186 (Mϩ Ϫ CH3, 100), 132 (Mϩ Ϫ CF3,
fragment (᎐N–CH –) of the reactant. By contrast, as deter-
᎐
2
mined by the isotope labeling experiments carried out with
PhC(᎐NCH 13CN)Me (1a*),4 a fraction (32%) of the CN
᎐
2
incorporated in the nitrile 2a resulted from the starting nitrile
functionality of 1a* [eqn. (4)]. This isotope distribution (not
explainable by path a, but partially accounted for by path b),
was rationalized by means of a simple cross-over experiment.
5), 124 (18), 104 (24), 103 [PhC(Me)᎐Nϩ, 10], 91 (26), 83 (10),
᎐
77 (18), 51 (10).
Compound 3b (0.2 g, 0.99 mmol) was made to react under
the conditions of Scheme 1, during which elimination of HF
took place. After 310 min in refluxing DMF, the conversion was
(4)
54% (by GC) and a mixture of E/Z isomers [PhC(᎐NCH᎐
᎐
᎐
CF2)Me: 3e (45%) and 3eЈ (3%)] was obtained. Structures were
assigned by GC–MS. Mass spectrum (70 eV) m/z (relative
intensity): 3e, 181 (Mϩ, 47%), 166 (Mϩ Ϫ Me, 100), 154 (15),
104 (25), 103 (12), 77 (38), 51 (13); 3eЈ, 181 (Mϩ, 37%), 104
(Mϩ Ϫ NCH᎐CF , 10), 103 (100), 77 (25), 51 (11).
Reaction of phenylacetonitrile 2c with 13CN؊
A cross-over experiment had been previously reported by us,4
and it had shown that under the conditions of Scheme 1, the
reaction of 1a in the presence 13C-labeled sodium cyanide,
yielded product PhCH(Me)13CN 2a* with 15% 13C enrichment.
Although this result did not exclude an intermolecular attack
of 13CN on the reagent, the existing evidence for the intra-
᎐
2
Imine PhCH(Me)N᎐CHCCl (3f ) was obtained through the
᎐
3
condensation of α-methylbenzylamine (0.5 g, 4.1 mmol) with
2,2,2-trichloroacetaldehyde.12 Yield of distilled product (bp =
62–65 ЊC/0.4 mmHg): 0.73 g (70%). 1H NMR (CDCl3) δ 7.77 (s,
molecular character of the reaction 1
2, strongly suggested
1H, CH᎐), 7.37–7.32 (m, 4 H, Ph), 4.68 (q, 1H, J = 6.7 Hz), 1.59
᎐
that the isotope scrambling occurred on the product nitrile 2,
after the imine had reacted. A more in-depth review of the
literature, supported this hypothesis.12
(d, 3H, CH3, J = 6.7 Hz). Mass spectrum (70 eV) m/z (relative
intensity): 249 (Mϩ, <1%), 132 (Mϩ Ϫ CCl3, 7), 106 (11), 105
(100), 77 (14), 51 (10).
Compound 3f (0.2 g, 0.80 mmol) was made to react under the
conditions of Scheme 1. As already reported for similar condi-
tions,12 elimination of HCl took place. After 90 min in refluxing
DMF, the conversion was 64% (by GC) and a mixture of E/Z
When phenylacetonitrile 2c was allowed to react directly with
13C-labeled sodium cyanide under the conditions of Scheme 1, a
significant isotopic enrichment (25%) of the product nitrile 2c
was observed. Therefore, since the CN scrambling was the final,
independent step, from the preceding ones involving imines 1,
the intramolecular mechanism of Scheme 4 appeared compat-
ible with the results also obtained for the 13C-labeled imine [1a*,
eqn. (4)]. The CNϪ released from the reaction of imines 1
accounted for the observed scrambling.4
isomers [PhC(᎐NCH᎐CCl )Me 3g (41%) and 3gЈ (20%)] was
᎐
᎐
2
obtained. Structures were assigned by GC–MS. Mass spectrum
(70 eV) m/z (relative intensity): 3g, 215 (Mϩ ϩ 2, 33), 213 (Mϩ,
50%), 202 (11), 200 (66), 199 (11), 198 (Mϩ Ϫ CH3, 100), 180
(11), 178 (33), 138 (11), 136 (17), 104 (63), 103 (22), 102 (16), 77
(45), 51 (32); 3gЈ, 213 (Mϩ, <1%), 164 (2), 106 (10), 105 (100),
103 (10), 77 (15).
Conclusions
Imine PhC(᎐NCH CN)H 1c was obtained through the con-
᎐
2
The results presented allow an overall mechanism for the trans-
densation of benzaldehyde (1.0 g, 9.4 mmol) with amino-
acetonitrile hydrochloride.1 Yield of distilled product (bp =
110 ЊC/0.7 mmHg, lit.,13 92–93 ЊC/0.1 mmHg): 0.60 g (44%).
Characterization data of 1c are in ref. 14.
formation 1
2, under the conditions of Scheme 1, to be
proposed, such as the one shown in path a of Scheme 3. It
involves a thermal electrocyclic 4-electron reaction followed by
a CN elimination step. While it is observed that the aziridine 3
does not react at all, and therefore cannot be an intermediate,
direct and indirect evidence supports the intermediacy of a
2H-azirine, which does give the expected nitrile under the
conditions investigated.
cis-2-Cyano-3-phenylaziridine (4c)9
4c was obtained by saturating a solution of α,β-dibromocin-
namonitrile (650 mg, 2.14 mmol) solution of α,β-dibromocin-
1036
J. Chem. Soc., Perkin Trans. 2, 2002, 1033–1037

View Article Online
namonitrile in DMSO (20 ml) with dry ammonia gas for one
hour. The mixture was stirred overnight at room temperature,
poured into brine, and extracted with dichloromethane. The
organic layer was washed twice with 50 ml of water, dried over
sodium sulfate and evaporated to dryness. The product solidi-
fied (55% yield) upon standing. H NMR (CDCl3) δ 7.5–7.3
(m, 5H), 3.56 and 3.02 (AB systems, J1 = 5.9 Hz, J2 = 9.7 Hz,
After removal of the CDCl3 with a stream of nitrogen (at
0 ЊC), the residue containing 2,2-diphenyl-2H-azirine 5b was
transferred to a 10 ml flask containing 150 mg (1.1 mmol) of
K2CO3, using 5 ml of dry DMF. The mixture was heated under
nitrogen at 120 ЊC for 2 hours, and then analyzed by GC–MS.
Along with 6b, the mixture showed the presence of diphenyl-
acetonitrile 2b (85%) and a small amount of phenylindole
(15%), all recognized by comparison with authentic samples.
Diphenylvinyl azide 6b was made to react under the same con-
ditions as 5b. After 2 hours at 150 ЊC, at complete conversion,
74% of benzophenone, 23% of phenylindole and <3% of 2b
were observed.
1
cis-aziridine ring protons||), 1.72 (t, 1H, J = 9.7 Hz).
2-Azido-1,1-diphenylethylene (6b)
Compound 6b was prepared following known methods starting
from diphenylethylene oxide, via 1,1-diphenyl-2-azidoethanol.6
Dehydration of the latter was conducted by the following pro-
cedure, adapted from the literature:15 1,1-diphenyl-2-azido-
ethanol (1 g, 4.2 mmol) was dissolved in 7 ml of dry DMF, 3.0
ml of pyridine were added, and the mixture was cooled to 15 ЊC
in an ice bath. In a separate flask, SO2, obtained by the drop-
wise addition of sulfuric acid to Na2SO3, was bubbled through
10 ml of methanesulfonyl chloride (MsCl, Aldrich reagent),
until a 5% wt solution of SO2 in MsCl was obtained. An aliquot
(2 ml, corresponding to 26 mmol of MsCl) of MsCl–SO2 was
added to the 2-azido-1,1-diphenylethylene solution over 15 min,
maintaining the temperature of the mixture below 15 ЊC, during
which the solution turned deep red. It was allowed to stir over-
night at room temperature. Work-up consisted of the dropwise
addition of 3 ml water, followed by a large excess (15 ml), and
extraction of the mixture with methylene chloride (15 ml). The
organic phase was washed successively with saturated brine,
half saturated brine, and water, and dried over Na2SO4. It was
then filtered, and evaporated to dryness at reduced pressure.
Purification was carried out by flash chromatography eluting
with a 1 : 1 mixture of toluene and petroleum ether (Rf = 0.55).
Reaction of phenylacetonitrile 2c with 13CN؊
Phenylacetonitrile (50 mg, 0.43 mmol) was set to react at 120 ЊC
in a nitrogen atmosphere, with an equimolar amount of
Na13CN in the presence of K2CO3 (90 mg, 0.65 mmol) in 5 ml
of dry DMF. After 1 hour the mixture was analyzed by
GC–MS, and the phenylacetonitrile showed a 25% isotope
enrichment of 13C.
Acknowledgements
We wish to thank Professor Andrea Maldotti for assistance in
the photochemical synthesis, and Dr Alfonso Zambon for help-
ing with the calculations of the molecular orbital energies. The
Consorzio Interuniversitario “la Chimica per l’Ambiente”
(INCA) is also gratefully acknowledged for funding.
References
1
Yield 300 mg (35%). H NMR (CDCl3) δ 7.4–7.2 (m, 10H),
1 M. Selva, C. A. Marques and P. Tundo, Synth. Commun., 1995, 25,
369.
2 M. Selva, A. Bomben and P. Tundo, Synth. Commun., 1999, 29,
1561.
3 A. Perosa, M. Selva and P. Tundo, Tetrahedron Lett., 1999, 40,
7573.
4 A. Perosa, M. Selva and P. Tundo, J. Chem. Soc., Perkin Trans. 2,
1999, 2485.
6.7 (s, 1H). IR νmax(neat)/cmϪ1 2102 (N ), 1612 (C᎐C). Mass
᎐
3
spectrum (70 eV) m/z (relative intensity): 193 (Mϩ Ϫ N2, 100),
165 (42), 139 (7).
2,2-Diphenyl-2H-azirine (5b) and its reaction under the
conditions of Scheme 1
5 R. G. Pearson, H. Sobel and J. Songstad, J. Am. Chem. Soc., 1968,
90, 319.
6 W. Tumas, R. F. Foster and J. Brauman, J. Am. Chem. Soc., 1988,
110, 2714.
7 T. Aoki, N. Haga, Y. Sendo, T. Konoike, M. Yoshioka and
W. Nagata, Tetrahedron Lett., 1985, 26, 339.
8 M. B. Smith and J. March, in March’s Advanced Organic Chemistry,
5th edn., J. Wiley & Sons, New York, 2001, p. 1243.
9 A. I. Ayi and R. Guedj, J. Chem. Soc., Perkin Trans. 1, 1983, 2045.
10 R. B. Woodward and R. Hoffmann, J. Am. Chem. Soc., 1965, 87,
395; H. C. Longuet-Higgins and E. W. Abrahamson, J. Am. Chem.
Soc., 1965, 87, 2045; R. B. Woodward and R. Hoffmann, J. Am.
Chem. Soc., 1965, 87, 2046.
11 K. Isomura, M. Okada and H. Taniguchi, Tetrahedron Lett., 1969,
46, 4073; K. Isomura, H. Taniguchi, M. Mishima, M. Fujio and
Y. Tsuno, Org. Magn. Reson., 1977, 9, 559.
2-Azido-1,1-diphenylethylene (200 mg, 0.9 mmol) was dis-
solved in 10 ml CDCl3 in a nitrogen atmosphere in a Pyrex
jacketed reactor maintained at Ϫ20 ЊC by circulating ethanol.
The solution was irradiated with a medium pressure Hg lamp
(filter cut-off λ < 350 nm). The tendency of atmospheric humid-
ity to condense on the reactor was controlled by means of a
nitrogen flow over the part exposed to the lamp. After 90 min-
utes, a portion of the solution was analyzed by 1H NMR which
showed the presence of 37% of the desired 2H-azirine 5b, the
remaining being substrate 6b, indicating that the reaction had
proceeded cleanly. 5b: 1H NMR (CDCl3) δ 10.2 (s, 1H), 7.4–7.2
(m, 10H).
|| Aziridine cis-protons are known to have larger coupling constants
and larger chemical shifts with respect to protons in the trans-
configuration; in the latter case the signals of the protons should there-
fore be broad, while in the former two AB systems would be expected.
The observed AB systems (J1 = 5.9 Hz, J2 = 9.7 Hz) are therefore
consistent with an aziridine in the cis-configuration.9
12 M. Jay, W. J. Layton and G. A. Digenis, Tetrahedron Lett., 1980, 21,
2621.
13 A. C. Davis and A. L. Levy, J. Chem. Soc., 1951, 3479.
14 O. Tsuge, K. Ueno, S. Kanemasa and K. Yorozu, Bull. Chem. Soc.
Jpn., 1986, 59, 1809.
15 G. G. Hazen and D. W. Rosenburg, J. Org. Chem., 1964, 29, 1931.
J. Chem. Soc., Perkin Trans. 2, 2002, 1033–1037
1037