Effects of superoxide anion generated from aromatic radical anions produced in nucleophilic aromatic photosubstitution reactions

Article abstract of DOI:10.1139/v98-044

Superoxide anion is generated from aromatic radical anions produced in nucleophilic aromatic photosubstitutions when the reactions are carried out in non-deoxygenated solutions of polar aprotic solvents. Superoxide anion thus generated displaces cyanide anion from acetonitrile and benzyl cyanide, ethoxide anion from ethyl acetate, and methanesulfenate anion from dimethyl sulfoxide. Hence, non-deoxygenated polar aprotic solvents should be avoided in nucleophilic aromatic photosubstitution reactions.

Full text of DOI:10.1139/v98-044

966
COMMUNICATION
Effects of superoxide anion generated from
aromatic radical anions produced in
nucleophilic aromatic photosubstitution
reactions
Maria Cervera and Jordi Marquet
Abstract: Superoxide anion is generated from aromatic radical anions produced in nucleophilic aromatic photosubstitutions
when the reactions are carried out in non-deoxygenated solutions of polar aprotic solvents. Superoxide anion thus generated
displaces cyanide anion from acetonitrile and benzyl cyanide, ethoxide anion from ethyl acetate, and methanesulfenate anion
from dimethyl sulfoxide. Hence, non-deoxygenated polar aprotic solvents should be avoided in nucleophilic aromatic
photosubstitution reactions.
Key words: superoxide, fluoride, nucleophilic aromatic photosubstitution.
Résumé : Lorsqu’on effectue des photosubstitutions aromatiques nucléophiliques dans des solutions non désoxygénées de
solvants aromatiques polaires, on génère l’anion superoxyde à partir des anions radicaux aromatiques alors produits. L’anion
superoxyde ainsi produit déplace l’anion cyanure de l’acétonitrile et du cyanure de benzyle, l’anion éthylate de l’acétate
d’éthyle et l’anion méthanesulfonate du diméthylsulfoxyde. Lorsqu’on effectue des réactions de photosubstitution
aromatiques nucléophiles, on doit donc éviter les solvants aprotiques polaires non désoxygénés.
Mots clés : superoxyde, fluorure, photosubstitution aromatique nucléophilique.
[Traduit par la rédaction]
class of compounds because of their wide applications, espe-
cially in medicinal chemistry (7). The most common way of
introducing a fluorine atom into an aromatic nucleus is based
on the use of nucleophilic fluoride anion through nucleophilic
aromatic substitution reactions (known as the “Halex
method”) (8). The efficiency of this method has been much
improved by the use of tetralkylammonium fluorides (9). Fluo-
ride anion is strongly solvated in protic solvents, its reactivity
being very reduced in these media. Therefore, reactions with
fluoride anion are normally carried out in anhydrous polar
aprotic solvents where fluoride behaves as a very strong base
and nucleophile (10).
Nucleophilic aromatic photosubstitutions show in many
cases a different regioselectivity when compared with the cor-
responding ground state reaction. Moreover, inefficient leav-
ing groups in thermal reactions, such as alkoxy groups, are
operative in the photochemical version. Therefore, we decided
to explore the nucleophilic aromatic photosubstitution of 4-ni-
troveratrole (NVT), a typical nitrophenyl ether for which the
regioselectivity of the photoreaction can be established, with
fluoride anion in non-deoxygenated polar aprotic solvents
such as acetonitrile (ACN). However, no fluorinated products
were obtained. In Scheme 1 the results of the photoreactions
carried out in acetonitrile are described. There is no precedent
in the literature for the reaction reported in the Scheme 1.
Nucleophilic aromatic photosubstitutions have been the object
of intense research since their discovery in 1956 (1). A mecha-
nistic borderline between polar SN2Ar* (2) and SN2Ar*-SET
(reactions that take place through single electron transfer from
the nucleophile to the excited nitroaromatic substrate) has been
well established (3, 4). These photoreactions are normally car-
ried out in protic solvents (water is the more common) and
without any precautions with regard to the presence of oxygen
since its effect on the photoreaction is generally negligible.
This has been attributed to the very short lifetime of the ni-
troaromatic excited states (5) and to the fact that the SN2Ar*-
SET reactions involve a collapse of radicals in the solvent cage
(6).
Fluorinated aromatics nowadays constitute an important
Received December 15, 1997.
This paper is dedicated to Professor Erwin Buncel in
recognition of his contributions to Canadian chemistry.
M. Cervera and J. Marquet.1 Departament de Química,
Universitat Autònoma de Barcelona. 08193 Bellaterra,
Barcelona. Spain.
1
Author to whom correspondence may be addressed.
Telephone: (3) 581-1029 / (3) 581-1265. Fax:(3) 581-1265 /
(3) 581-2477. e-mail: iqor3@cc.uab.es
Can. J. Chem. 76: 966–969 (1998)
© 1998 NRC Canada

967
Communication
Scheme 1.
Scheme 4.
a
Medium-pressure Hg lamp (Pyrex filter, λ > 290 nm). Room
temperature. The products were characterized by comparison
with authentic samples. Absolute preparative yields based on
nonrecovered starting material. Blank irradiations in the absence
of fluoride led in all cases to recovery of the starting material.
Scheme 2.
complete recovery of the starting material. These results rule
out the hypothesis of a reductive cleavage in the mechanism
for the photoreactions described in Scheme 1.
To investigate the evolution of the alkyl fragment, the pho-
toreaction was carried out in benzyl cyanide (instead of ACN)
and with air bubbling. The reaction crude was directly ana-
lyzed by GC–MS and the products detected are described in
the Scheme 2.
Scheme 3.
From the results reported in Scheme 2 it was clear that the
photofragmentation of alkylnitriles is a general process under
our conditions, and that the agent responsible for it is an oxi-
dizing species with nucleophilic properties. All products de-
rived from the benzyl moiety are easily justified by
considering the radical PhCH2OO• as a key intermediate (13).
These results led us to the hypothesis that the “oxidizing spe-
cies with nucleophilic properties” was the superoxide anion
(O2–•) (14). The excellent nucleophilic properties of superox-
ide anion in polar aprotic solvents are well documented (it
reacts with alkyl halides (15) or carbonylic compounds (16));
however, as far as we know no displacement of cyanide from
nitriles has been reported (Sawyer and co-workers (18) de-
scribe the action of superoxide as a base with acetonitrile, ob-
taining condensation products). On the contrary, acetonitrile
has been used as a solvent in fundamental studies on the reac-
tivity of superoxide anion and its derivatives (18); however, it
has been noted that the reactivity of O2–• in acetonitrile is
different from that in other solvents (for example, ref. 19), the
lifetime of O2–• being typically 10 times shorter in the nitrile.
Sawaki and Ogaka (20) reported the reaction between su-
peroxide and acetonitrile in the presence of an easily oxi-
dizable species such as dimethyl sulfoxide as leading to
acetamide and dimethylsulfone through the attack of superox-
ide on the nitrile. However, when the reaction was carried out
For the sake of comparison, we carried out the photoreac-
tion of NVT with sodium cyanide in ACN–H2O (1 h), obtaining
the same photosubstitution product, 2-cyano-4-nitroanisole
(46% yield), as the only cyanide substitution product, indicat-
ing complete meta regioselectivity. Therefore, it seems that the
photoreaction described in Scheme 1 involves the fragmenta-
tion of the solvent. However, the only related cases described
in the literature refer to reductive cleavage of tertiary or sec-
ondary alkylnitriles (11) promoted either electrochemically or
by solvated electrons, and to photoinduced reductive cleavage
of benzylnitriles (12). In these cases the resulting free radical
would be stabilized by conjugative or hyperconjugative ef-
fects. No such effects are present in the methyl radical and, in
addition, our photoreactions proved to be very sensitive to the
presence of oxygen. Thus, the photoreaction in non-deoxygen-
ated ACN led to 17% yield of photosubstitution product, but
this yield was increased to 67% by bubbling air into the solu-
tion. The photoreaction in the absence of oxygen led to
© 1998 NRC Canada

968
Can. J. Chem. Vol. 76, 1998
Scheme 5.
Scheme 6.
in the absence of the oxidizable species, these authors de-
scribed the disappearance of the superoxide anion. No expla-
nation was advanced for this behavior.
in anhydrous acetonitrile) instead of fluoride. In addition, the
radical anion of NVT (23) was detected by UV–vis spectros-
copy when NVT was irradiated in the presence of TBAF in
acetonitrile, under inert atmosphere, but this radical anion
could not be detected when bromide was used. These results
strongly support our hypothesis. In Scheme 4 a simplified
mechanistic proposal is described.
As expected, this behavior was not restricted to nitriles. The
photoreaction of NVT in the presence of TBAF 3H2O in ethyl
acetate gave rise to 2-ethoxy-4-nitroanisole (Scheme 5A),
whereas in dimethyl sulfoxide (DMSO), 1,2-dimethoxy-3-
methyl-4-nitrobenzene, 1,2-dimethoxy-4-methyl-5-nitroben-
zene, and 2-methoxy-5-nitrophenol were obtained
(Scheme 5B). These results support our previous conclusions
and can also be justified by considering the superoxide anion
as responsible. Thus, and for the case of ethyl acetate, it is well
known that superoxide anion reacts with esters producing alk-
oxy anions (17). Here, the ethoxy anion would act as a nucleo-
We have confirmed the role played by the superoxide anion
in our reactions by carrying out the photoreaction of NVT with
potassium superoxide in acetonitrile and in inert atmosphere,
in the presence of the 18-crown-6 (Scheme 3). Direct analysis
of the reaction crude by GC showed only the presence of start-
ing NVT and 2-cyano-4-nitroanisole. The displacement of
cyanide from acetonitrile in the presence of superoxide anion
seems to be a ground state reaction, and this was shown by the
detection of cyanide anion (Prussian blue test) in a solution of
potassium superoxide and 18-crown-6 in acetonitrile.
After the identification of the agent responsible for our pho-
toreactions an immediate question arises: how is the superox-
ide anion produced? A still unsolved paradox in ground state
aromatic chemistry is the fact that when an electronically poor
aromatic (e.g., polynitrobenzenes) substrate is placed in the
presence of not easily oxidizable anions such as alkoxy or
cyanide, not only the Meisenheimer complexes, but also the
corresponding radical anions, can be detected (21). It is known
that the concentration of radical anions depends on the con-
centration and nature of the nucleophile, and on the concen-
tration of Meisenheimer complexes present in the solution
(normally very low in ground state chemistry) (22). Therefore,
the simplest hypothesis is that production of significant
amounts of superoxide anion in our photochemical reactions
originates by interaction of dioxygen with the radical anion of
the aromatic substrate (NVT), and that this radical anion is
present in a significant concentration under our photochemical
conditions due to the enhanced reactivity of the nucleophile
(fluoride) versus the excited state of the aromatic substrate
(NVT is not a very electronically poor substrate), affording a
sufficiently large concentration of Meisenheimer complexes.
The dependence of the photoreaction upon the nature of the
nucleophile and on the concentration of Meisenheimer com-
plexes is demonstrated by the fact that the reactions of
Scheme 1 failwhen attempted with bromide(aweakernucleophile
phile towards NVT in
a
nucleophilic aromatic
photosubstitution reaction (last step of Scheme 4, RO– instead
of CN–). In the case of the photoreaction carried out in DMSO,
the main photoreaction corresponds to the methylation of the
aromatic ring. This result suggests the operation of methyl
radicals as reaction intermediates. It is known (24) that hy-
droxyl radicals (•OH) add to DMSO producing methyl radicals
through decomposition of intermediate species such as
Me2S(O•)OH. The presence of hydroxyl radicals can be justi-
fied by assuming that superoxide anion behaves towards
DMSO in a similar way as towards ACN (Scheme 6), through
nucleophilic displacement of, in this case, methanesulfenate
anion. The methanesulfenate anion thus produced would be
responsible for the observed 2-methoxy-5-nitrophenol through
a nucleophilic attack upon the excited state of NVT and sub-
sequent hydrolysis, probably at work-up, since no phenol was
observed in the photoreactions carried out in ACN, either in
ethyl acetate or in a blank irradiation in the absence of fluoride
anion.
© 1998 NRC Canada

969
Communication
13. J.R. Kanofsky, M. Matsumoto, H. Sugimoto, and D.T. Sawyer.
J. Am. Chem. Soc. 110, 5193 (1988).
14. (a) D.T. Sawyer and J.S. Valentine. Acc. Chem. Res. 14, 393
(1981); (b) M. Chanon, M. Julliard, J. Santamaria, and F.
Chanon. New J. Chem. 16, 171 (1992).
Acknowledgements
Financial support from the Ministerio de Educación y Culture
(M.E.C.) of Spain through project DGES PB96–1145 and from
the Generalitat de Catalunya through project 1995SGR00469
is gratefully acknowledged.
15. (a) J.L. Roberts, Jr., T.S. Calderwood, and D.T. Sawyer. J. Am.
Chem. Soc. 105, 3337 (1983); (b) T.S. Calderwood and D.T.
Sawyer. J. Am. Chem. Soc. 106, 7185 (1984); (c) H. Sugimoto,
S. Matsumoto, D.T. Sawyer, J.R. Kanofsky, A.K. Chowdhury,
and C.L. Wilkins. J. Am. Chem. Soc. 110, 5193 (1988).
16. (a) J.L. Roberts, Jr. and D.T. Sawyer. J. Am. Chem. Soc. 106,
4667 (1984); (b) F. Magno and G. Bontempelli. J. Electroanal.
Chem. 68, 337 (1976); (c) D.T. Sawyer, J.J. Stamp, and K.A.
Menton. J. Org. Chem. 48, 3733 (1983).
17. M.J. Gibian, D.T. Sawyer, T. Ungermann, R. Tangpoonpholvi-
vat, and M.M. Morrison. J. Am. Chem. Soc. 101, 640 (1979).
18. (a) J.W. Peter and C.S. Foote. J. Am. Chem. Soc. 98, 875 (1976);
(b) H. Sugimoto, S. Matsumoto, and D.T. Sawyer. J. Am. Chem.
Soc. 109, 8081 (1987).
19. (a) D.B. Denney and J.D. Rosen. Tetrahedron, 20, 271 (1964);
(b) M.J Gibian and T. Ungermann. J. Am. Chem. Soc. 101, 1291
(1979); (c) M. Tezuka, H. Hamada, Y. Ohkatsu, and T. Osa.
Denki Kagaku, 44, 17 (1976).
20. Y. Sawaki and Y. Ogata. Bull. Chem. Soc. Jpn. 54, 793 (1981).
21. S.M. Shein, L.V. Bryukhovetskaya, F.V. Pishchugin, V.F.
Starichenko, V.N. Panfilov, and V.V. Voevodskii. J. Struct.
Chem. 11, 228 (1970).
References
1. J. Cornelisse and E. Havinga. Chem. Rev. 75, 353 (1975), and
references therein.
2. H.C.H.A. Van Riel, G. Lodder, and E. Havinga. J. Am. Chem.
Soc. 103, 7257 (1981).
3. (a) K. Mutai, K. Kobayashi, and K. Yokoyama. Tetrahedron, 40,
1755 (1984); (b) N.J. Bunce, S.R. Cater, J.C. Scaiano, and L.J.
Johnston. J. Org. Chem. 52, 4214 (1987).
4. (a) A. Cantos, J. Marquet, M. Moreno-Mañas, and A. Castelló.
Tetrahedron, 44, 2607 (1988); (b) A. Cantos, J. Marquet, M.
Moreno-Mañas, A. González-Lafont, J.M. Lluch, and J. Bertrán.
J. Org. Chem. 55, 3303 (1990).
5. (a) J. Cornelisse, G.P. De Gunst, and E. Havinga. Adv. Phys.
Org. Chem. 11, 225 (1975); (b) P.H.M. Van Zeijl, A.M.J. van
Eijk, and C.A.G.O. Varma. J. Photochem. 29, 415 (1985).
6. A.M.J. Van Eijk, H. Huizer, C.A.G.O. Varma, and J. Marquet. J.
Am. Chem. Soc. 111, 88 (1989).
7. J. Mann. Chem. Soc. Rev. 16, 381 (1987), and references therein.
8. S. Brown and M. Bowden. Chem. Ind. (London), 143 (1993),
and references therein.
9. J.H. Clark and D.K. Smith. Tetrahedron Lett. 26, 2233 (1985).
10. J.H. Clark. Chem. Rev. 80, 429 (1980).
11. (a) P.B. Arapakos. J. Am. Chem. Soc. 89, 6917 (1967); (b) P.G.
Arapakos and M.K. Scott. Tetrahedron Lett. 1975 (1968);
(c) P.G. Arapakos, M.K. Scott, and F.E. Huber, Jr. J. Am. Chem.
Soc. 91, 6425 (1969).
22. (a) C. Mariani, G. Modena, G.P. Pizzo, and G. Scorrano. J.
Chem. Soc. Perkin Trans. 2, 1187 (1979); (b) A. Sauer, F. Was-
gestian, and B. Barbasch. J. Chem. Soc. Perkin Trans. 2, 1317
(1990).
23. K. Mutai, K. Yokoyama, S. Kanno, and K. Kobayashi. Bull.
Chem. Soc. Jpn. 55, 1112 (1982).
24. B.C. Gilbert, R.O.C. Norman, and R.C. Sealy. J. Chem. Soc.
12. P. Maslak and J. Kula. Tetrahedron Lett. 31, 4969 (1990).
Perkin Trans. 2, 303 (1975).
© 1998 NRC Canada