вход по аккаунту


Aromatic Fluorine as a Chemical Label for Detecting Reaction Mechanism.

код для вставкиСкачать
method and its ability to cope with higher temperatures and pressures [201.
In the condensed phase, thermodynamic data have
been obtained mainly for stoichiometric compounds.
The present techniques readily permit one to study
phase diagrams and thermodynamic properties as
functions of composition of non-stoichiometric compounds [75,145,1461. Likewise, other aspects of heterogeneous reactions and of the mechanism and kinetics
of vaporization, e.g. in electrochemical cells, can be
Problems of electronic and molecular structure of gaseous molecules can not be solved directly by mass
spectrometric techniques. The use of this technique
for detecting and identifying species in magnetic and
electric molecular beam experiments has, however,
already given interesting results 11471. The simultaneous study of gaseous phases by mass spectrometry and
by optical spectroscopy, including microwave and
[1451 K . A. Gingerich, J. physic. Chem. 68, 2514 (1964); J. Amer.
chem. SOC.87, 1660 (1965).
11461 A. R. Miller and A. W . Searcy, J. physic. Chem. 69, 3826
(1 965).
11471 L . Wharton, R . A. Berg, and W . Klemperer, J. chem.
Physics 39,2023 (1963); A. Biichler, J. L . Stauffer, W . Klemperer,
and L. Wharton, ibid. 39, 2299 (1963); M . Kaufman, L. Wharton,
and W . Klemperer, ibid. 43, 943 (1965); R . A . Berg, L . Wharion,
and W . Klemperer, ibid. 43,2416 (1965); A. Biichler, J. L . Staufer,
and W . Klemperer, J. Amer. chem. SOC. 86, 4544 (1964).
matrix techniques 11481, are also promising developments, as is the molecular structure determination by
electron diffraction [149J.
Many problems and systems mentioned here are regularly covered and discussed in review articles 11501.
We thank our co-workers for most valuable discussions
and contributions.
Acknowledgment is also due to the European Office of
Aerospace Research, the Institut pour I'Encouragernent
de la Recherche Scientifique dans I'lndustrie et I'Agriculture, the Fonds National de la Recherche Scientifique,
and the Fonds National de la Recherche Fondamentale
Collective who made the necessary funds available for
rhe experimental work which forms the basis of this
Received: Januaty 5th. 1966;
revised: February 13th, 1967
[ A 574 IE]
German version: Angew. Chem. 79, 589 1967
11481 W . Condensation and Evaporation of Solids.
Gordon and Breach, New York, p. 243; W . Weltner j r . and D.
McLeodjr., J. Molecular Spectroscopy 17, 276 (1965); J. physic.
Chem. 69, 3488 (1965); D. Wite, K . S . Seshadri, D . F. Mann, and
M . J . Linevsky, J . chem. Physics 39,2463; R. L . Barger and H . P.
Broida, ibid. 43, 2364, 2371 (1965); P. H . Kasai and W . Weltner
.jr., ibid. 43, 2553 (1965).
11491 A. N . Khodchenkov, V . P. Spiridonov, and P . A. Akishin, 2.
strukturnoij Chirn. 6, 634, 764, 765 (1965).
[150] Annual Review of Physical Chemistry, Ann. Rev. Inc.,
Palo Alto, Calif. (U.S.A.); Progress in Inorganic Chemistry,
Interscience, New York; Advances in High Temperature Chemistry. Academic Press, New York.
Aromatic Fluorine as a Chemical Label for Detecting Reaction Mechanism'**1
This article explains how nucleophilically activated fluorine can be used for the detection
of ion-pair formation during the decomposition of various aryl radical generators. Tlius a
concurrent dual mechanism for the brenlcdown of certain aryl radical sources can be experitiientallj: demonstrated by fluorine labeling, The bearing of this informcition on the
mechanism of diaio coupling is also discussed.
In addition, labeling with aromatic fluorine is applied to the thermolysis of aryl azides, the
aniinolysis of 3-aryl-4-bromosydnones,and the hydrolysis of 3-arylsydnone imiries. Finally
some mechanistic problems are posed which may possibly be solved by this labeling
I. Principle of the Method
Nucleophilic replacement of suitably activated fluorine
in an aromatic structure usually occurs more readily
than that of any other group or atom[lJ. Foremost in
[ * ] Dr. H. Suschitzky
Department of Chemistry and Applied Chemistry,
University of Salford, Salford 5,
Lancashire (England)
[**I Extended version of a lecture presented at the 3rd International Fluorine Symposium, Munich, September 1965.
[ l ] J. F. Bunnett, Quart. Rev. 12, 1 (1958).
this activating tendency is undoubtedly the diazonium
ion -N2@, which can render aromatic fluorine (and to
a lesser extent other halogens or a nitro group) so
labile that the latter can be lost inadvertently during a
synthesis by being displaced in the presence of even a
weak nucleophilic reagent. Bunnett and Zahler [21 have
called this annoying property of a diazonium group
the "nuisance effect" and the aptness of the metaphor
will be appreciated by anyone who has experienced
this phenomenon. An instance of such an unwanted
121 J. F. Bunnett and R. E. Zahler, Chem. Revs. 49, 273 (1951).
Angew. Chem. internat. Edit.
Vol. 6 (1967) / N o . 7
fluorine displacement is the loss of the fluorine atom
from 4-fluoro-3-nitroaniline ( I ) which occurs during
diazotization in hydrochloric acid [31 to give unexpectedly the chloro compound (3) via the intermediate (2).
Another example is the formation of the chlorodiazonium salt (5) when the dry 2,4-difluorobenzenediazonium fluoroborate ( 4 ) is heated to its melting point
(which lies below its decomposition temperature) in the
presence of sodium chloride 141. Even a hydrochloric
acid solution of diazotized 0- or p-fluoroaniline will
eventually give a positive test for fluoride ions indicating fluorine-chlorine exchange 131. Examples of the
effect of a diazonium ion on the ease of replacement of
other groups, are also well known 151.
Scheme 1. The preparation of 5-fluoroindazole.
In the above reactions ( I ) + (3) and ( 4 ) -+ (S), aromatic fluorine has in a sense acted as a chemical label
since it records the presence of the diazonium ion by
appearing as easily Petectable ionic fluorine (HF) in
the reaction mixture. This information is obviously
superflous in the cases mentioned. However, a situation could be envisaged in which the intermediacy of a
diazonium ion, or any other group that causes aromatic fluorine to be anionically labile, is only postulated
or even unsuspected. Under these circumstances the
conversion of covalent (aromatic) into ionic fluorine
would clearly be of value in diagnosing the mechanism
involved. Some of the results, which we obtained from
experiments in which aromatic fluorine was used as a
“chemical marker”, are discussed in the subsequent
sections. It is hoped that other workers will be stimulated to apply this simple and novel method to suitable
It. Mechanism of Indazole Formation
The idea of probing into the mechanism of a reaction
by fluorine labeling suggested itself to us in the course
of making 5-fluoroindazole by Jacobson’s method [6,71,
which entails the spontaneous decomposition of the
N-nitrosobenz-o-toluide (6) in benzene [see Scheme11.
The expected 5-Auoroindazole (8) was obtained together with an approximately equal quantity of a
fluorine-free substance which proved to be 5-benzoyloxyindazole (12). Fluoride ions were detected[*] in
the reaction mixture 181. Halogen replacement did not
occur with the meta isomer (F meta to N(NO)COC6H5
in (6)) and the behavior of fluorine in the ortho position was not tested as the required nitroso compound
could not be prepared.
Huisgen and Nakaten 191 have shown that the Jacobson
indazole synthesis involves a rate-determining rearrangement of the nitroso compound to a diazoester
(e.g. (6) + (7)) before cyclization. In this context the
ready replacement of the fluorine in the para position
is rationally explained if the covalent diazoester (7)
dissociates into the ion pair (9). Fluorine replacement
by the anion CsH5C02Q is then strongly facilitated
because the charged diazonium group -N2 activates
the fluorine towards nucleophilic attack. Thereby a
new ion pair (10) is formed in which the previously
covalent fluorine has taken on the role of anionic
partner. Aromatic, covalent fluorine has been changed
into a fluoride ion and thus has functioned as a sensitive chemical label since it has revealed the unsuspected
and transient formation of a diazonium ion pair during
indazole cyclization in benzene. Moreover, fluorine is
found to be specific in providing this mechanistic information since the chloro or bromo analogue of (6),
F = C1 or Br, yields exclusively the corresponding
halogenoindazoles (8), F = C1 or Br.
The existence of ion pairs raises the obvious question
whether it is the covalent trans-diazoester, e.g. (7), as
previously postulated 191, or the diazonium ion pair,
e.g. (9), which is involved in the ring closure to give
the indazole. De Tar [lo] has investigated the cyclization
131 H . Suschitzky, unreported result.
[4] G. C. Finger and R. E. Oesterling, J. Amer. chem. SOC. 78,
2593 (1956).
[5] H . Suschitzky in: Advances of Fluorine Chemistry. Butterworths, London 1965, Vol. IV, p. 12.
161 I. K. Barben and H. Suschitzky, J. chem. SOC.(London) 1960,
171 P. Jacobsen and L. Huber, Ber. dtsch. chem. Ges. 41, 660
Angew. Chem. internat. Edit.
1 YoI. 6 (1967) / No. 7
[*I Fluoride ions can be conveniently detected by means of the
zirconium alizarin test; F. Feigl: Spot Tests. Elsevier, New
York 1954, Vol. 11, p. 67.
[8] I. K. Barben and H. Suschitzky, J. chem. SOC.(London) 1960,
[9] R. Huisgen and ff. Nakaten, Liebigs Ann. Chem. 586, 84
[lo] D. F. De Tar and Y. W. Chu, J. Amer. chem. SOC.76,1686
of cis-o-diazostilbene (13) and shown that indazole
(14) is formed under conditions favoring a heterolytic
process (aqueous acid) and that phenanthrene (15) is
the main product from homolysis (copper powder).
To this extent it is feasible that the cyclization step in
um compounds exist to a considerable extent in their
covalent form, which must be in equilibrium not only
with “free” ions but also with tightly held ion pairs.
III. Decomposition of Fluorine Substituted
the Jacobson synthesis is also ionic in nature [see
Scheme 1: (9)
(S)]. The ionic route accounts satisfactorily for the ease of internal coupling between
diazo and methyl group since the diazonium ion can
activate its coupling partner by a powerful +I effect.
Moreover, the closeness of the carboxylate ion would
promote cyclization by providing a suitable acceptor
for the methyl proton which is released during ring
The elegant, kinetic studies which led Huisgen I91 to
postulate a four-center concerted mechanism within
the covalent trans-diazoester (16) -+ (18) do not
It was a logical step to extend the application of
fluorine labeling to acylarylnitrosamines without an
ortho methyl substituent [81 since these compounds are
known to break down homolytically in non-polar
solvents to provide aryl radicals. A benzene solution
of N-(4-fluorophenyl)-N-nitrosobenzamide (19) kept
at room temperature gave the expected 4-fluorobiphenyl (21) by homolytic aromatic substitution of the
p-FC6H4 radical in the solvent (c6H6) and an equal
amount of 4-benzoyloxybiphenyl (25). Corresponding
Scheme 2. Decomposition of N-@-Fluorophenyi)-N-nitrosobenzamide.
necessarily invalidate our ionic reaction scheme for
indazole formation (17) + (18). It must be realized
that association of two ions to form either the covalent trans-diazoester (16) or the intimate ion pair
(17) is a gradual and continuous process and the two
species (16) and (17) would be indistinguishable if it
were not for fluorine labeling.
An obvious objection to the ionic mechanism of cyclization is the lack of indazole formation from omethyl substituted diazonium compounds in aqueous
mineral acid. It is, however, feasible that an activated
methyl-group will undergo internal coupling only if
closely associated with a proton accepting base, a situation realized in an ion pair. The fact that various
nitro- and polynitro-indazoles can be made from the
appropriate nitrobenzene-diazonium salts in aqueous
mineral acid
supports the postulated mechanism :
it is well known that in aqueous solutions such diazoni[ll] E. Nolting, Ber. dtsch. chem. Ges. 37, 2556 (1904); H. D.
Porter and W.D . Peterson, Org. Syntheses, Coll. Vol. 111, p. 660.
results were observed with fluorine in the ortho-position and also with other acyl groups, and fluoride ions
were present in each reaction mixture. Fluorine meta
to the N-nitroso group, or any other halogen in ortho
or para position, was not expelled. In the naphthalenes
(26), X = F, C1, or Br, however, chlorine as well
as bromine suffered replacement 1121 in conformance with the higher reactivity of activated halogen in
this ring system 1131.
These results are again rationally accommodated
within the accepted reaction scheme by interpreting
[12] P . Miles and H. Suschitzky, Tetrahedron 18, 1369 (1962).
1131 H. J. Van Opstall, Recueil Trav. Chim. Pays-Bas 52, 901
Angew. Chem. internat. Edit.
Vol. 6 (1967) / No. 7
with a number of independent observations: Both the
the appearance of fluoride ions as evidence for the
nitroso compound ArN(N0)-COR (19) and
heterolysis of the covalent diazoester (20) into an ion
diazoester ArN=NOCOR (20) can dissopair (22). At this stage the fluorine is rendered labile
towards nucleophilic attack by the anionic partner
[C6H5C0ze in (22)], and the steps leading to a mixture of biphenyls (21) and (25) are readily discernible
in Scheme 2. The stages (19) + (21) represent the
well-known rearrangement and homolytic substitution
as revealed by the elegant work of Hey, Huisgen, and
their collaborators [14,153. This reaction scheme must,
however, be augmented by the additional steps (22)
(25) to include a concomitant heterolysis, which is in0
ferred from fluorine-labeling experiments. By contrast,
heterolysis of acylarylnitrosamines in a polar medium
such as acetic acid or methanol has been postulated
beforer15.161 but was precluded for benzene solutions 1151. It is of interest that acylalkylnitrosamines
have recently been shown to decompose by a polar
mechanism even in non-polar solvents as a result of
1 8 0 tracer studies carried out by White and Aufdermarsh [171. For instance, N-(n-hexyl)-N-nitroso-2naphthamide (27) disintegrates via formation of the
diazonium salt (29) to give the ester (30) with completely equilibrated oxygens. This technique is obviously not applicable to the aromatic nitroso compounds
since the acyloxy moiety does not recombine with the
aryl group.
Scheme 3.
Formation of benzoylfluoride in the decom-
When the products from the above fluorine-labeling
experiments were separated by fractional distillation
of the reaction mixture under anhydrous conditions,
and not by extraction with aqueous solvents, the acid
fluoride RCOF derived from the acyl moiety of the
starting nitroso compound F-ArN(N0)COR was obtained in high yield. For instance, the benzoylarylnitrosamine (19) gave benzoyl fluoride, and analogous
acetyl or formyl compounds in the benzene or naphthalene series yielded acetyl or formyl fluoride respectively [*21. Blank experiments excluded the possibility
that these products were formed purely incidentally to
the main reaction, perhaps just by combination of the
organic acid RCOzH and hydrogen fluoride, both of
which are present during the reaction (see Scheme 2).
We interpret the origin of the acyl fluoride RCOF in
the following way because this explanation correlates
[14] W. S. M . Grieve and D. H. Hey, J. chem. SOC.(London)
1934, 1797; E. C. Butterworth and D . H . Hey, ibid. 1938, 116;
Q. H. Hey, A . Nechvatal, and T. S. Robinson, ibid. 1951, 2892.
1151 R . Huisgen and G. Horeld, Liebigs Ann. Chem. 562, 137
[16] D . F. De Tar, J. Amer. chem. SOC.73, 1446 (1951).
[I71 E. H. White and C. A . Aufdermarsh, J. Amer. chem. SOC. 83,
1174, 1179 (1961).
Angew. Chem. internat. Edit. f VoI. 6 (1967) No. 7
ciate into their respective ion pairs. These relatively
stable pairs can interact and exchange partners to
form the new ion pairs (32) and (33). In the case of the
decomposition of the 4-fluorophenylnitroso compound
(19) an interchange of the cationic moieties between
the intimate ion pairs (31) and (23) [for the formation
of the ion pair (23) see Scheme 21 leads to the observed
benzoyl fluoride (35) as shown in Scheme 3.
The simultaneously formed diazoanhydride (38) is a
highly unstable free-radical source 1181 whose breakdown products are indistinguishable from those expected from the starting nitroso compound (19).
Intermediacy of diazoanhydrides has indeed been
postulated again recently in a modified reaction
scheme for the thermolysis of acylarylnitrosamines
(see Scheme 4) 1191.
[18] E. Bamberger, Ber. dtsch. chem. Ges. 29, 446 (1896); 53,
2314 (1920); Th. Kauffmann, H . 0.Friestad, and H . Henkler,
Liebigs Ann. Chem. 634, 64 (1960).
[19] C. Riichardt and B. Freudenberg, Tetrahedron Letters 1964,
3623; C. Riichardt, B. Freudenberg, and E. Merz, Abstracts of
Papers of the International Symposium on Organic Reaction
Mechanism, Cork, Ireland 1964, p. 28; C. Riichardt, Angew.
Chem. 77, 974 (1965); Angew. Chem. internat. Edit. 4, 964
Moreover, the participation of an acylium ion in the
isomerization of nitroso compound to diazoester
[C6H5-C08 in Scheme 3, (19) + (2O)] is in some
agreement with the kinetic results obtained by Hey [201
and Huisgelz [211 although these workers prefer a nonionic interpretation of these data. Yet it appears that
the rate of this rearrangement depends on the structure of the acyl group in the nitroso compound in a
way expected for an S N process,
i.e. by a mechanism
involving an acylium ion. Thus the isomerization
velocity of ArN(N0)CO-R increases in the order
R = H < CH3 < C2H5 < ~ s o - C ~but
H ~ is little affected by changes in the aryl moiety; this is feasible
since the aromatic ring neither participates in the
migration nor has it any conjugative influence on the
moving acyl or receiving nitroso group.
Since our experiments were carried out, Riichardt [I91
has suggested a modified decomposition mechanism
of acylarylnitrosamines, which embodies several features inferred above from fluorine-labeling experiments. This new reaction scheme overcomes an
intrinsic weakness of the older one[z21 which assumed
intermediacy not only of aryl but also of acyloxy
radicals (RC02.) yet could not account for the failure
of the latter to liberate carbon dioxide. This anomaly
was explained away by a “cage mechanism” [15,231
which, however, became incompatible with later
results “241. While we fully agree with Ruchardt’s reaction scheme we consider it unnecessary to postulate
that the initial ’diazotate anion (37) arises from the
interaction of an acetate ion and the nitrosamine.
+ Ac@
The spontaneous dissociation of the nitroso compound (36), as evidenced by acyl fluoride formation in
our labeling experiments (see Scheme 3), seems to us a
more feasible source for this key intermediate. Consequently the appropriate acid anhydride, e.g. (38),
which has been observed in several reactions [19,251 is
not formed as originally suggested from the nitroso
compound (36), but arises later in the reaction by a
process analogous to that of acyl fluoride formation in
Scheme 3. This process is shown in Scheme 4, which
represents essentially the reaction scheme proposed by
Riichardt [I91 except for the modifications mentioned
above with some consequential alterations. Further
support of this scheme comes from recent ESR measurements [25al which confirmed the presence of aryldiazotate radicals (ArNzO-).
N 2 - 0 AC
[ A c - 0 Ac]
m ]
[20]D . H. Hey, J. Stuart- Webb, and G . H. Williams, J. chem. SOC.
(London) 1952,4657.
[21]R . Huisgen and L. Krause, Liebigs Ann. Chem. 574, 157
[22]H. Zollinger : Azo and Diazo Chemistry. Interscience, New
York 1961,pp. 153-159.
[23]E. L. Eliel, M . Eberhardt, 0. Simamura, and S . Meyerson,
Tetrahedron Letters 1962,749.
1241 E. L. EIiel and J . G . Saha, J. Amer. chem. SOC.86, 3581
(1964); D. B. Denney, N . E. Gershman, and A. Appelbaum, ibid.
86,3180 (1964).
Scheme 4.
Decomposition mechanism of acetylarylnitrosamine.
Square brackets denote ion pairs and isolated reaction products are
[25]The reaction mixture from the decomposition of
C~H5-N(NO)-CO-C6H5 in benzene showed two sharp bands
at 1790 and 1725 cm-1 characteristic of benzoic anhydride in the
infrared spectrum. H . Suschitzky, unpublished result.
[25a]G. Binsch, E. Merz, and C. Riichardt, Chem. Ber. 100, 247
Angew. Chem. internat. Edit. 1 Vol. 6 (1967) No. 7
So far it has been tacitly assumed that the decomposition mechanism deduced for fluorine-substituted
acylarylnitrosamines is applicable to the whole class
of these compounds regardless of aromatic substituent.
Fluorine is probably unique in providing mechanistic
evidence in this reaction since the numerous observations of decomposition of non-fluorinated nitrosamines
reported over the last twenty-five years d o not contain
a single observation of anionic displacement affecting
another substituent. The specificity of fluorine was
confirmed when chlorine replacement could not be
detected by even a most careful product analysis of the
reaction mixture obtained from N-nitroso-p-chlorophenylacetamide. Yet it is unlikely that only fluorine
substituted nitrosamines have the prerogative of undergoing heterolysis. I n order to find out whether other
nitrosamines form ion pairs besides free radicals,
regardless of the nature of their ring-substituent, we
carried out several mixed decompositions [121. Thus
various acetylarylnitrosamines with substituents of
different electronic properties [(39), R = C1, CH3,
NO2, H in the para position as well as CH3 in the meta
position] were made to decompose (each separately)
in dry benzene containing some N-nitroso-p-fluorophenylbenzamide (40). If the two nitrosoamines (39)
and (40) disintegrated at a similar rate and both
heterolysed to give diazonium ions, two sets of ion
clearly indicative of the fact that the breakdown of the
acetylnitrosamine (39) under test is not purely homolytic but has also included the ionic intermediate (41).
This was confirmed by product analysis using gasliquid and column chromatography since the diagnostic
4-acetoxybiphenyl (47) was present (up to 17%) in
each of the mixed decompositions in addition to the
expected three biphenyls (45), (46), and (48). The
possibility that the crossed product (47) is due to
substitution of the diazonium compound (42) by
acetic acid rather than by the anion from the ion pair
(41) was discounted by control experiments. It is thus
shown conclusively that acylarylnitrosamines of all
types decompose via a route involving homolytic and
heterolytic steps.
IV. Coupling Mechanism of Acylarylnitrosamines
in Non-Aqueous Media
The information gathered from fluorine labeling on
the behavior of acylarylnitrosamines allows us to make
some relevant observations on the mechanism of azo
coupling in non-aqueous solvents. Huisgen [15,21 261
was able t o determine the isomerization rate of acylarylnitrosamines (49) into diazoesters (50) simply by
measuring colorimetrically their azo-dye production
Scheme 5.
Decomposition of an acetyl and benzoylnitrosamine in
pairs, namely (41) and (42), would be present in the
reaction mixture. As a consequence, two species of
anions, namely benzoate and acetate, would be competing for displacement of aromatic fluorine in (42)
and four different biphenyls (45)-(48) would be
produced by subsequent events as shown in Scheme 5.
Formation of the 4-acetoxybiphenyl (47) would be
Angew. Chem. internat. Edit. 1 Yal. 6 (1967) f No. 7
with added P-naphthol since the latter process is a
much faster reaction than the former. Dye formation
proceeded more rapidly in benzene than in acetic acid
and as the non-polar solvent was presumed to contain
only the diazoester (SO) coupling was ascribed to this
species rather than to its "dissociated" form, the diazonium ion (51). This postulate is, however, at
[26] R . Huisgen, Liebigs Ann. Chem. 573, 163 (1951); 574, 184
60 1
a) Phenylazotriphenylmethane
R (49)
The decomposition of phenylazotriphenylmethane in
hot benzene was first studied in 1934 by Hey[2Yal.
Subsequently a number of workers, Hey, Wieland,
Huisgen, and their respective collaborators [2Yb-2YgJ,
have shown that the reaction involves free-radical
intermediates while other workers have put forward
a cage process in which free radicals are not involved [23J. The accepted mechanism for the homolytic
decomposition of phenylazotriphenylmethane in an
aromatic solvent ArH is outlined in Scheme 7.
Scheme 6. Coupling of acylarylnitrosamine with @-naphtholin benzene.
variance with Bradley and Thompson’s observation [271
on diazonium salts such as p-(n-decy1oxy)benzenediazonium-p-toluenesulfonate (52a) or the p-(n-hexadecy1oxy)benzene diazonium chloride (52b) which in
Scheme 7.
Homolytic decomposition of phenylazotriphenylmethane.
An unsuccessful attempt has been made to induce
heterolysis by introducing electronically different substituents into the azo compound, e.g. (53), and by the
use of various solvents [3OJ.
spite of their moderate solubility in benzene nevertheless couple readily with P-naphthol in this medium to
give azo dyes. These discrepancies can, however, be
resolved on the basis that diazonium ions (as borne
out by fluorine-labeling experiments) exist in the nonpolar media and, like carbonium ions, are prone to
form intimate ion pairs as the polarity of the solvent
decreases. In a solvent such as benzene it is most probably the diazo ion pair rather than the covalent ester
that undergoes azo coupling (see Scheme 6). The
increased azo-coupling rate in benzene is possibly due
to the favorable geometrical position as well as the
polarity of the anionic partner, factors which ensure
a rapid proton loss from the transition state (52).
V. Fluorine Labeling of Various Aryl Radical
It was thought that “fluorine tagging” of other aromatic radical sources, particularly those which operate
by homolytic cleavage of an aryl-nitrogen bond, would
be a simple way to discover any inherent but hidden
tendency to heterolytic fission. The results of this work
in relation to currently accepted mechanisms are
briefly summarized in the following sections 1281.
[27] W . Bradley and J. D . Thompson,Nature (London) 178, 1069
1281 P. Miles and H. Suschitrky, Tetrahedron 19, 385 (1963).
Only products indicative of a free-radical breakdown
were obtained.
We studied the same problem by causing p-fluorophenylazotriphenylmethane (54) to decompose in hot
benzene. Heterolysis of the former to give the diazonium ion pair (55) could possibly lead to fluorine replacement by the triphenylmethyl anion of (55). However,
n o ionic fluorine was detected; this could be attributed
to the unreactivity of this bulky, resonance-stabilized
triphenylmethyl anion. In order to provide a more
nucleophilic agent the reaction was carried out in the
presence of N-nitrosoacetanilide (56),which we know
[29a] D . H. Hey, J . chem. SOC. (London) 1934, 1966. [b] D. H.
Hey, C . J. M . Stirling, and G. H. Williams, J. chem. SOC. (London) 1955, 3963. [c] ,G. L . Davies, D. H . Hey, and G . H. Williams,
ibid. 1956, 4397. [d] D. H. Hey, M . J . Perkins, and G . H. Williams, Tetrahedron Letters 1963,445. [el D . H. Hey, M. J. Perkins,
and G. H. Williams, I. chem. SOC. (London) 1965, 110. [f] H.
Wieland, K . Heyman, T. Tsatsas, D . Juchum, G . Varvoglis, G .
Labriola, G. Dobbelstein, and H. S . Boyd-Barreft, Liebigs Ann.
Chem. 514, 145 (1935). [g] R. Huisgen and H. Nakatan, ibid. 586,
70 (1954). [h] R. Huisgen and R. Grashey, ibid. 607, 46 (1957).
[30] M . D. Cohen, J . E. Leffler, and L . M . Barbato, J. Amer.
chern. SOC. 76, 4149 (1954).
Angew. Chem. internat. Edit. / Val. 6 (1967) No. 7
from previous work (see Scheme 4) to release the
reactive acetate ion under these conditions. Fluoride
ions and the diagnostic 4-acetoxybiphenyl (58) were
formation of free radicals, nitrogen, a biphenyl, and
an amine (see Scheme 9: Route a). When the fluorinelabeled triazene (59),Y = F, was made to decompose
in a boiling mixture of benzene (10 parts) and acetic
acid (1 part), 4-acetoxybiphenyl (62), X = AcO, and
fluoride ions in addition to the products expected
from the homolysis (60), X = AcO, were formed.
An analogous result was observed when hydrochloric
acid was used instead of acetic acid; the diagnostic
product 4-chlorobiphenyl (62), X = C1, was formed.
(55 J
+ Nz + H F
Scheme 8.
Mixed decomposition of p-fluorophenylazatriphenylmethane and N-nitrosoacetanilide.
in fact detected, a result which is best accounted for by
interaction of the two ion pairs (55) and (57). As is
shown in Scheme 8, the ion pair (55) is a heterolytic
product of the azo compound (54).
The other reaction products were triphenylmethane,
triphenylmethanol, triphenylmethylperoxide, and 4fluorobiphenyl which are normally expected from the
homolysis of (54). The results of recent experiments [29el in which two different azo compounds were
made to decompose together in benzene for a limited
time and in which the undecomposed azo compounds
were found to have exchanged their trityl groups,
albeit to a small extent, could also be explained on
our ion-pair hypothesis. This evidence is, however,
not fully conclusive as considerable experimental
difficulties in assessing the products were reported.
b) N-(p-Fluoropheny1diazo)piperidine
Thermolysis of triazenes [311, for instance of the piperidene derivative (59), Y = H, in benzene solution
(containing acids), produces, via the intermediate
I311 J. Elks and D . H. Hey, J. chem. SOC.(London) 1934, 441.
Angew. Chem. internat. Edit.
Vol. 6 (I967) /No.7
Scheme 9.
Thermolysis of p-fluorophenyldiazopiperidine.
Again the presence of the two biphenyls (60) and (62),
X = AcO or C1, is indicative of a dual mechanism in
the pyrolysis of the triazene, as set out in Scheme 9.
Fluorine replacement obviously occurs at the ion-pair
stage (6f), X = A c O e or Cle, while the triazene
itself does not dissociate into ions for piperidinobiphenyl is not a reaction product.
c) p,p'-Difluorodiazoaminobenzene
Diazoaminobenzenes have recently been reported by
Hardie and Thornson [321 to be sources of free radicals
when heated in a high-boiling solvent. Nitrogen, and
anilino and aryl radicals are produced ; combination
of the latter with an aromatic solvent leads to biaryl
formation as set out in Scheme 10 (cf. Route a).
Decomposition of the fluorine analogue (63) in boiling
chlorobenzene gave p-fluoroaniline hydrofluoride (67)
which deposited on the walls of the reflux condenser
This substance is formed undoubtedly by interaction
of the homolytically produced p-fluoroaniline (64)
and heterolytically formed hydrogen fluoride (65).
The predicted diphenylamine (66) was not isolated
from the tarry reaction mixture which, however, gave
a strong diphenylamine tess 1331.
1321 R. L. Hardie and R. H. Thomson, J. chem. SOC. (London)
1958, 1286.
I331 Concentrated sulfuric acid and sodium nitrite produced
an immediate blue color on addition of a trace of recation product
cf. F. Feigf,S p o t Tests. Elsevier, New York 1954, Vol. 11, p. 126
Scheme 10.
Thermolysis of p,p’-difluorodiazoaminohenzene in an
aromatic solvent.
d) p-Fluorophenylhydrazine
Oxidation of arylhydrazines in an aromatic solvent
with silver oxide liberates aryl radicals, which attack
the solvent to form biarylsI341. This homolysis probably involves the covalent diazo compound (69),
which could feasibly be in equilibrium with the aryldiazonium hydride (71). If so, replacement of a
fluorine atom at the para-position by the hydride ion
is possible; but this is improbable because the hydride
ion will be too readily removed by the oxidizing agent.
It was, therefore, not unexpected to find that treatment
of p-fluorophenylhydrazine (68) with silver oxide in
benzene gave p-fluorobiphenyl (70), in accordance
with Route (a) in Scheme 11, but no fluoride ions. In
the presence of N-nitrosoacetanilide (72), which
provides a diazonium acetate and with it the much
more ‘ effective acetate anion, the reaction products
were 4-fluorobiphenyl and 4-hydroxybiphenyl[281.
Moreover, an aqueous extract of the silver residues
gave a strongly positive test for ionic fluorine. The
anticipated 4-acetoxybiphenyl(73) could not be isolated as it suffered hydrolysis in situ to its 4-hydroxyderivative (74). By analogy with the previous conclusions a-rational scheme for phenylhydrazine oxidation can be advanced (see Scheme 11) which clearly
illustrates operation of a dual mechanism.
e) p-Fluorobenzoyl peroxide
Thermal decomposition of diaroyl peroxides is judged
to be the best method of providing aryl radicals because it usually gives the “cleanest” reaction products.
Its mechanism has probably received more attention
than that of any other aryl radical source [29a, 351. Unsymmetrical peroxides such as p-methoxy-p’-nitrobenzoylperoxide (75) in which the substituents induce
a unidirectional electron-flow and thereby distort the
molecule electronically can, according to Lefler 1361,
heterolyse in nitrobenzene in presence of acetic acid or
Nz t H F +
Scheme 11.
Oxidation of p-fluorophenylhydrazine with silver oxide in
[34] R . L. Hardie and R. H . Thomson, J. chem. SOC. (London)
1957, 2512.
[35] P. Gelisson and P. H . Hermans, Ber. dtsch. chem. Ges. 58,
285 (1925); D . F. De Tar and R . A . J . Long, J. Amer. chem. Soc.
80, 4742 (1958); E. L. Eliel, S. Meyerson, 2. Welvart, and S. H .
Wilen, ibid. 82, 2936 (1960); M . Eberhardt and E. L. Eliel, 1. org.
Chemistry 27, 2289 (1962); G. H . Wiiliams, Homolytic Aromatic
Substitution. Pergamon Press, Oxford 1960, pp. 34-41.
1361 J . E. Lefler, J. Amer. chem. SOC. 72, 67 (1950).
Angew. Chem. internat. Edit. / Val. 6 (1967)
/ No. 7
thionyl chloride, as inferred from product analysis. If
the fluorine labeled peroxide (76) can dissociate into
ions [Scheme 12) then fluoride ions (77) should be
detectable in the reaction mixture. However, only
compounds expected from a free radical process,
namely p-fluorobenzoic acid (78), 4-fluorobiphenyl
(79), and a little p,p‘-difluoro-p-quaterphenyl (80)
were obtained in benzene [281. Pyrolysis in the presence
of N-nitrosoacetanilide did not lead to fluorine replacement whatever solvent was employed.
that any halogen, even in the meta position, was eliminated mith formation of the corresponding anilinium
halide it became obvious that the loss of halogen was
not due t o a simple nucleophilic aromatic substitution.
We are mentioning this experiment as a cautionary
example. Had we been working only with compounds
containing fluorine in the 0-and p-positions, an erroneous conclusion could have easily been drawn. It is in?portan. to realize that mobility of a fluorine atom in
the ortho orparn position is diagnostic of an activating
Scheme 13. Thermal decomposition of halogenophenyl azides
in aniline.
Scheme 12. Thermal decomposition of p-fluorobenzoyl peroxide
in benzene.
VI. Thermolysis of Fluorophenyl h i d e s
The decomposition of aryl azides (81) in hot solvents
leads to the loss of nitrogen and formation of a highly
reactive, unstable intermediate (82) which, because of
its isoelectronic nature with, and chemical resemblance
to, carbene [381, is best called nitrene 1371. As nitrenes
often behave 2s an electrophilic species (Ar-N :), we
thought that the electron deficient nitrogen may render
A r - N 3 . ----+
Nz + A r - N :
in solution
a para- or ortho-positioned fluorine labile towards
nucleophilic attack in a manner analogous to an aromatic diazonium group ArN2 @.
Indeed, thermolysis
of 0- or p-fluorophenyl azide in chlorobenzene [391
containing a little aniline as the nucleophile produced
a sublimate in the reflux condenser which was essentially aniline hydrofluoride. When we found, however,
[371 Naming of this intermediate is unfortunately not uniform
a t present. The terms azene and imidogen are also used in the
Anglo-Saxon literature while in German the intermediate is
referred to as “Imen”.
1381 L. Horner and A . Chnstmann, Angew. Chem. 75,707(1962);
R . A . Abramovitch and B. A . Davies, Chem. Rev. 64, 149 (1964);
M . Apple and R. Huisgen, Chem. Ber. 91, 12 (1958); 92, 2961
(1959); W. von E. Doering and R. A . Odum, Tetrahedron 22,
81 (1966).
[391 R . K . Smalley and H . Suschitzky, J. chern. SOC. (London)
[Suppl. 21 1964, 5922.
Angew. Chem. internat. Edit.f Yo[.6 (1967)1 No. 7
aromatic intermediate only if it has been demonstrated
that this halogen is not replaceable in the meta position.
A full discussion and presentation of the reasons for
this unusual halogen replacement are outside the scope
of this article but can be found elsewhere[391. However, to complete the picture it may be briefly stated
that a phenyl azide with a halogen in the orrho, meta,
or para position (83) decomposes in aniline to give a
small quantity of the 6-, the 3- and the 5-, or the 4halogenoazepine (84), respectively. Halogen in any
position of this heterocycle was found t o be prone to
nucleophilic attack and is thus readily replaceable by
the added aniline or by the corresponding halogenoaniline (85), which is a normal by-product formed by
addition of hydrogen to the halogenonitrene as shown
in Scheme 13.
VII. Aminolysis of 3-Aryl-4-bromosydnones and
Acid Hydrolysis of 3-Arylsydnone Imines
We found [401 that the aminolysis of 3-aryl-4-bromosydnones (86) with a secondary amine, for examole
piperidine, yielded the anticipated amino azid aniide
(87) but primary amines, for instance cyclohexylamine,
gave an unexpected result. Nitrogen was slowly liberated and the main product proved to be the Schiff base
(90) containing two cyclohexyl moieties. Tentatively
we interpreted this unforeseen cleavage of the hetero[40] M. BeNas and H . Suschitzky, Chern. Comrnun. (London) 7,
136 (1965); M. Bellas and H . Suschitzky, J. chern. SOC. (London)
C1966, 189; G. Puranik and H . Suschitzky, J. chem. SOC. (London) C 1967,1006.
cycle as a two-step process the first stage producing the
nitroso-compound (88). A 1,6 interaction between the
nitroso oxygen and the aminohydrogen could feasibly
lead to the observed azomethine (90) and, of necessity,
to the diazohydroxide (89), a reaction scheme for
which we had of course no evidence. However, this
mechanism seemed capable of being tested by fluorine
labeling. If the postulated diazohydroxide contained a
fluorine atom in the para position (89), X F, then
dissociation (cf. Route a) would render this atom
sufficiently labile to promote its replacement in the
nucleophilic environment. Accordingly we tested the
p-fluorophenylsydnone (86), X = p-F, with cyclohexylamine and found not only the expected Schiff
base (90) but also ionic fluorine in the reaction mixture. In the case of m-fluorophenylsydnone (96), X =
m-F, no fluoride ions were produced thus establishing
the nucleophilic nature of the replacement of the
fluorine atom in the para position and thereby
corroborating the proposed mechanism as given in
Scheme 14.
the fluorine would be rendered labile towards nudeophilic attack by the chloride ions. Consequently chlorobenzene, p-chlorophenol, and fluoride ions should
be present in the reaction mixture. On the other hand
m-fluorophenylsydnone imine should not lose its
fluorine atom. All these predictions were experimentally verified and the reaction path in (Scheme 15) for
the acid hydrolysis of aryl sydnone imines is thus
Acid hydrolysis of 3-Arylsydnone irnines
Vm. Potential Scope of Aromatic
Fluorine Labeling
It seems appropriate to conclude this brief review of a
noveI and little explored method by drawing attention
to some mechanistic problems which could well be
amenable to "fluorine tagging". Even a random perusal of the literature yields a crop of reactions of uncertain mechanism where a differential diagnosis between a free radical or a heterolytic pathway could be
made. Certain rearrangements of nitrogen heterocyclic
compounds are a potentially fruitful field in this respect, particularly if intermediacy of a diazonium
structure is suspected.
For instance, the conversion of the triazine (91), R =
H, into the cinnoline (93), R = H, by warm hydrochloric acid has tentatively been explained [41J by a
Scheme 14. Aminoly\is of 3-aryl-4-bromosydnones.
It is noteworthy thatp-chlorophenylsydnone (96),X =
p-CI, did not furnish chloride ions when treated with
the amine, which again stresses the distinction of
fluorine as a chemical label. One of the products from
the ammonolysis of chlorosydnone was chlorobenzene,
which is in keeping with the proclivity of the aryl
radical CIC&Lp to abstract hydrogen. Homolysis of
the intermediate diazohydroxide (89), X = p-C1,
(Route b in Scheme 14) is responsible for the presence
of this radical as well as for the observed evolution of
Russian workers [4Oal have recently suggested that 3phenylsydnone imine is hydrolysed on treatment with
hot hydrochloric acid, a diazonium ion being formed
as an intermediate. Since participation of an aryl diazonium cation can be diagnosed by using aromatic
fluorine as a label we prepared the g-fluorophenylsydnone imine (90a), and hydrolysed it with hydrochloric acid. At the postulated diazonium stage (90b),
[40a]L. E. Kholodov and V. G. Yashunskii,
3336 (1963).
z. obSE. Chim. 33,
heterolytic cleavage involving the pyrazolylphenyldiazonium ion (92), R = H, as an intermediate. Its
rotation through 180 about the central bond is followed by cyclization yielding the product (93). Under
similar conditions the related ring system of benzotriazin-Cone (94), R1 = C6H5, yields a small quantity of the phenanthridone (96). On the other hand
thermal decomposition of (94), R1 = C6H5. at 280 "C
[41] M. 5'. Gibson, Chem. and Ind. 1962, 698.
Angew. Chem. internat. Edit. 1 Vol. 6 (1967) 1 No. 7
was reported [421 t o give a mixture of the phenanthridone (96) and acridone (100) and these results were
interpreted in terms of a radical mechanism. However,
the behavior of 3-aminotriazinone (94), R1 = NHZ,
in acid was accounted for by intervention of the diazonium ion (95), R1 = NH21431. Results of recent
studies on the photolysis 1441 of this system, e.g. (94),
R' = C 6 H 5 , have again led to the assumption that it
dissociates first to give a diazonium ion (97). Loss of
nitrogen leads to the formation of the cyclic amide
(98) or its valence tautomer (99) either of which is responsible for all observed products such as acridone
(100). Direct chemical evidence for or against participation of the postulated diazonium ions (92), (95), or
(97) in these reactions could feasibly be obtained by
using the "fluorine-labeled" starting materials (91),
R == F, and (94), R = F.
The scope of fluorine labeling is by no means limited
to reactions with a potential diazo intermediate. For
instance, another problem which would qualify for
this labeling method is the well-known but little understood Meisenheimer rearrangement 1451 of amine
oxides to hydroxylamines e.g., (101) + (104), which
occurs on heating. Recent evidence based on the
stereochemical outcome of this isomerization, namely
substanrial racemization of the optically active [a-D]benzylamine-N-oxide (105) excludes an S,i type of
rearrangement [46J. However, no clear-cut decision between the heterolytic pathway (a) and the homolytic
process (b) can really be made, although the authors
of these experiments are inclined to favor the latter 1461.
Since we have shown that the N-oxide group activates
aromatic fluorine more than any other halogen towards nucleophilic substitution [471 the rearrangement
of a p-fluorophenylamine N-oxide, e.g. (IOl), R1
could throw considerable light on the
mode of this reaction. If the ionic mechanism (a) were
operative the nucleophilically active fluorine in the
starting material (IOI), R1 = p - F - C & ,
or possibly
in the cation (102), R1 = p-F-CgH4, would suffer replacement by attack of the benzyl anion (102). If on
the other hand the Meisenheimer rearrangement involved, as suggested 1461, a homolytic process (b),
ionic fluorine would not be found in the reaction
I am much indebted to my co-workers Dr. I . K. Barben,
Dr P. Miles, Dr. M. Bellas, Dr. R. K . Snialley, Dr. G .
Puraxik, and D. Price for their enthusiastic support and
to the Research Fund of the Chemical Society (London),
the Gergy Co. Ltd., Smith, Kfine and French (Pa.
U.S.A.), and the S. R. C. (London) for generous financial help and to Dr. P. Koch of Koch-Light Laboratories
Lid. j o r gifts of fluorine compounds.
Received: March 24th, 1966,
[A 575 [El
revised: March 14th, 1967
German version: Angew. Chem. 79,636 (1967)
O C H D - r t- R
Ree& and A . R. Todd, Chem. and Ind.
1421 D . H. Hey, c.
1962, 1332.
1431 M . S. Gibson and M . Green, Tetrahedron 21, 2191 (1965).
1441 G. Ege, Angew. Chem. 77, 723 (1965); Angew. Chem. internat. Edit. 4, 699 (1965); E. M . Burgess and G. Milne, Tetra-
hedron Letters 1966, 93.
Angew. Chem. internat. Edit. Vol. 6 (1967) 1 No. 7
1451 J. Meisenheimer, Ber. dtsch. chem. Ges. 52, 1667 (1919);
J. Meisenheimer, H. Greeske, and A . Wilmersdorf, ibid. 55, 513
[46] U.Schollkopf and H . Schafer, Liebigs Ann. Chem. 683, 42
(1965); U. Schollkopf, M . Putsch, and H. Schafer, Tetrahedron
Letters, 1964, 2515.
1471 M. Bellas and H. Suschitzky, J. chem. SOC.(London) 1963,
4007; 1964, 4561; 1965, 2096; Chem. Commun. (London) 1 5 ,
367 (1965); D. Price and H. Suschitzky, unpublished.
Без категории
Размер файла
1 043 Кб
labe, chemical, reaction, mechanism, detecting, fluorine, aromatic
Пожаловаться на содержимое документа