Mechanism of the benzidine disproportionation of (arylhydrazo)pyridines. The reactions of 4-(4-chlorophenylazo)pyridine and -hydrazo)pyridine in acid media

Article abstract of DOI:10.1139/v98-082

A kinetic and product analysis study of the reactions of 4-(4- chlorophenylhydrazo)pyridine (1) and 4-(4-chlorophenylazo)pyridine (2) in acid media is reported. The disproportionation of two moles of 1 in aqueous sulfuric acid gives one mole of the oxidized product 2, and one mole each of the reduced products 4-chloroaniline and 4-aminopyridine. The azo compound 2, a product of the reaction of 1, undergoes a slower hydroxylation reaction in acid media, and this process was also investigated. The first-formed product in the reaction of 2 is probably 4-(4-hydroxyphenylazo)pyridine, the 4- chlorine being displaced. At the low substrate concentrations used for the kinetic measurements both reactions are straightforward, but at the much higher concentrations used for product isolation studies a complex product mixture results, one of the products being a dimer. The complexity is increased because 1 had to be used as its hydrochloride salt for reasons of stability, and the chloride ion can also react with the diprotonated substrates; chloride ion also accumulates in the system as it is displaced from 2 during the hydroxylation. Thus chloride ion competes with bisulfate ion (present in large excess) for the diprotonated substrates. Nevertheless, complete mechanistic schemes accounting for all of the observed products, including the dimer, could be derived and are presented. Values of pK(BH₂²⁺) for 2 and two of the azo products, needed for the kinetic analysis, were measured using the excess acidity equilibrium method. The nucleophile reacting with protonated 2 in the rate-determining step of the hydroxylation was positively identified as being bisulfate ion by an excess acidity analysis. A comparison of the reaction of 1 with the equivalent reactions of two previously studied 4-(arylhydrazo)pyridines, 9 and 10, reveals that the benzidine disproportionation of these molecules is an A1 process with the second protonation being a pre-equilibrium; pK(BH₂²⁺) values for the hydrazo compounds could not be measured as they react too quickly, but they could be determined from the kinetic data, using the excess acidity method. The rate-determining step is a thermally allowed 10-electron electrocyclic process of the diprotonated substrate, giving rise to an intermediate that undergoes fast reaction with a molecule of protonated substrate in a thermally allowed 14-electron electrocyclic process to give the observed products.

Full text of DOI:10.1139/v98-082

896
Mechanism of the benzidine disproportionation
of (arylhydrazo)pyridines. The reactions of
4-(4-chlorophenylazo)pyridine and
-hydrazo)pyridine in acid media
Robin A. Cox, Kap-Soo Cheon, Sam-Rok Keum, and Erwin Buncel
Abstract: A kinetic and product analysis study of the reactions of 4-(4-chlorophenylhydrazo)pyridine (1) and
4-(4-chlorophenylazo)pyridine (2) in acid media is reported. The disproportionation of two moles of 1 in aqueous sulfuric acid
gives one mole of the oxidized product 2, and one mole each of the reduced products 4-chloroaniline and 4-aminopyridine.
The azo compound 2, a product of the reaction of 1, undergoes a slower hydroxylation reaction in acid media, and this process
was also investigated. The first-formed product in the reaction of 2 is probably 4-(4-hydroxyphenylazo)pyridine, the
4-chlorine being displaced. At the low substrate concentrations used for the kinetic measurements both reactions are
straightforward, but at the much higher concentrations used for product isolation studies a complex product mixture results,
one of the products being a dimer. The complexity is increased because 1 had to be used as its hydrochloride salt for reasons
of stability, and the chloride ion can also react with the diprotonated substrates; chloride ion also accumulates in the system as
it is displaced from 2 during the hydroxylation. Thus chloride ion competes with bisulfate ion (present in large excess) for the
diprotonated substrates. Nevertheless, complete mechanistic schemes accounting for all of the observed products, including
the dimer, could be derived and are presented. Values of pKBH 2+ for 2 and two of the azo products, needed for the kinetic
2
analysis, were measured using the excess acidity equilibrium method. The nucleophile reacting with protonated 2 in the
rate-determining step of the hydroxylation was positively identified as being bisulfate ion by an excess acidity analysis. A
comparison of the reaction of 1 with the equivalent reactions of two previously studied 4-(arylhydrazo)pyridines, 9 and 10,
reveals that the benzidine disproportionation of these molecules is an A1 process with the second protonation being a
pre-equilibrium; pKBH 2+ values for the hydrazo compounds could not be measured as they react too quickly, but they could be
2
determined from the kinetic data, using the excess acidity method. The rate-determining step is a thermally allowed
10-electron electrocyclic process of the diprotonated substrate, giving rise to an intermediate that undergoes fast reaction with
a molecule of protonated substrate in a thermally allowed 14-electron electrocyclic process to give the observed products.
Key words: azopyridines, benzidine, disproportionation, excess acidity, hydrazopyridines, hydroxylation, mechanism.
Résumé : On a réalisé une étude cinétique ainsi qu’une analyse des produits des réactions des
4-(4-chlorophénylhydrazo)pyridine (1) et 4-(4-chlorophénylazo)pyridine (2) en milieu acide. La dismutation de deux
molécules de 1 dans de l’acide sulfurique conduit à une mole du produit oxydé 2 et à une mole chacun des produits réduits,
4-chloroaniline et 4-aminopyridine. En milieu acide, le composé azo 2, un produit de la réaction de 1, subit une réaction
d’hydroxylation plus lente; on a aussi examiné ce processus. Le premier produit formé dans la réaction de 2 est probablement
la 4-(4-hydroxyphénylhydrazo)pyridine (1) provenant d’un remplacement du chlore en position 4. Aux faibles concentrations
utilisées dans les mesures cinétiques, les deux réactions sont simples; toutefois, aux concentrations plus élevées utilisées pour
les études en vue d’isoler les produits, il y a formation d’un mélange complexe de produits comprenant un dimère. La
complexité est accrue à cause du fait qu’il a été nécessaire, pour des raisons de stabilité, d’utiliser le chlorhydrate du composé
1 et que l’ion chlorure peut aussi réagir avec les substrats diprotonés; de plus, l’ion chlorure s’accumule aussi dans le système
lorsqu’il est déplacé du composé 2 par hydroxylation. L’ion chlorure est donc en compétition avec l’ion bisulfate (présent en
très grand excès) pour les substrats diprotonés. Malgré tout, on peut présenter des schémas mécanistiques complets permettant
d’expliquer tous les produits observés, y compris le dimère. Utilisant la méthode d’équilibre d’acidité en excès, on a mesuré
les valeurs de pKBH 2+ du composé 2 et de deux produits azo nécessaires pour l’analyse cinétique. Faisant appel à l’analyse de
2
Received January 14, 1998.
This paper is dedicated to Professor Erwin Buncel in recognition of his contributions to Canadian chemistry.
R.A. Cox.1 Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, ON M5S 3H6, Canada.
K.-S. Cheon and E. Buncel.2 Department of Chemistry, Queen’s University, Kingston, ON K7L 3N6, Canada.
S.-R. Keum. Department of Chemistry, Graduate Studies of Korea University, Seoul, Korea 136-701.
1
Mechanistic studies in strong acids, part 24. Part 23 is ref. 26. Telephone: (416) 978-4581. Fax: (416) 978-8775.
E-mail: rcox@alchemy.chem.utoronto.ca
Studies of azo and azoxy dyestuffs, part 25; part 24 is ref. 24. Telephone: (613) 545-2653. Fax: (613) 545-6669.
E-mail: buncele@chem.queensu.ca
2
Can. J. Chem. 76: 896–906 (1998)
© 1998 NRC Canada

897
Cox et al.
l’acidité en excès, il a été possible d’identifier sans ambiguïté que l’ion bisulfate est le nucléophile qui réagit avec la forme
protonée de 2 dans l’étape cinétiquement limitante de l’hydroxylation. Une comparaison de la réaction de 1 avec les réactions
équivalentes de deux 4-(arylhydrazo)pyridines (9 et 10) étudiées antérieurement révèle que la dismutation de la benzidine de
ces molécules est un processus A1 dans lequel la deuxième protonation est un prééquilibre; il n’a pas été possible de mesuré
les valeurs des pKBH 2+ de ces composés hydrazo à cause de leur trop grande vitesse de réaction, mais il a été possible de les
2
déterminer à partir de données cinétiques faisant appel à la méthode d’acidité en excès. L’étape cinétiquement limitante est un
processus électrocyclique thermique permis, à 10 électrons, du substrat diprotoné qui conduit à un intermédiaire qui réagit
rapidement avec une molécule de substrat protoné par le biais d’un processus électrocyclique thermique permis, à 14
électrons, conduisant aux produits observés.
Mots clés : azopyridines, benzidine, dismutation, acidité en excès, hydrazopyridines, hydroxylation, mécanisme.
[Traduit par la rédaction]
Thus, in our quest to resolve the mechanistic ambiguity in the
Introduction
benzidine-type disproportionations of the 4-(arylhydrazo)pyri-
dines (24), we have extended these studies to include 4-(4-
chlorophenylhydrazo)pyridine 1. We also examined the azo
analog 2, the results for which can be compared with those
previously obtained for other (arylazo)pyridines (20). These
substrates form the subject of the present paper; it will be
shown that while the results obtained using 1 and 2 are more
complex in terms of reaction products, the kinetic data ob-
tained do, in fact, allow the definitive resolution of the mecha-
nism of this benzidine rearrangement-like process.
Studies of azo and azoxy compounds have revealed a remark-
able diversity of structure–reactivity relationships, and some
challenging problems in mechanism. Our work in this area was
originally concerned with the mechanism of the acid-catalyzed
Wallach rearrangement of azoxyarenes (1–7). This has devel-
oped in recent years to encompass a variety of photoactive
molecules and processes, including the synthesis (8) and elec-
tronic spectral characteristics (9) of novel azo merocyanine
dyes, and their solvatochromism (9, 10), which led to the es-
tablishment of the π*azo solvent polarity scale (10). The photo-
chemical and thermal interconversion of merocyanines and
their spiro precursors (11–15) has led most recently to studies
of the chiroptical properties of spiro molecules, with a view to
developing a ferroelectric liquid crystal optical switch based
on the principle of photoresolution (16, 17).
The present series of studies is an extension of the Wallach
rearrangement studies of phenyl and naphthyl substrates (2–6)
to the heteroaromatic series, with the observed >104 reactivity
difference between the isomeric α- and β-phenylazoxypyridi-
nes involving di-and tri-protonation processes in the 95–100
wt.% H2SO4 acidity region (7). This led to the discovery of the
remarkably facile hydrolysis of the aryl azo ethers in acid me-
dia, with protonation on nitrogen, oxygen or carbon centers
occurring as fast equilibria or as the rate-determining step (18,
19). In the course of these studies it was discovered that pheny-
lazopyridine is itself unstable in strong acid media, undergoing
hydroxylation and disproportionation (20). Since this was
reminiscent of the classical benzidine rearrangement (21–23),
kinetic studies were extended to the surmised intermediate,
4-(4-hydroxyphenylhydrazo)pyridine (24). However, an ex-
cess acidity treatment of the results (25) led to uncertainties in
the postulated reaction mechanism. This was unsatisfactory,
since the excess acidity method (25) has proven to be very
useful in the determination of the mechanisms of reactions that
take place in strong acid media. For instance, lactam (26),
alkylnitramine (27), N-nitro amide (28), and acylimidazole
(29) hydrolyses, as well as imidazoline cyclizations (30),
alkene hydrations (31), aromatic hydrogen exchange processes
(32), and ether (19, 33) and thioacetal (34) hydrolyses have
been studied using this approach.
Experimental
The hydrochloride salt of 4-(4-chlorophenylhydrazo)pyridine
1 was obtained in 33.8% yield when 4-phenylazopyridine was
simultaneously chlorinated and reduced (35), mp 241–242°C
(lit. (35) mp 243°C). This compound was used as its hydro-
chloride salt rather than in the free base form, since the former
proved to be much more air-stable. 4-(4-Chloropheny-
lazo)pyridine 2 resulted in 22.3% yield when 1 was oxidized
(35), mp 97–98°C (lit. (35) mp 99–100°C). The synthesized
compounds were characterized as before (20).
Substrates and isolated products were purified by several tech-
niques, recrystallization, and column and thin-layer chroma-
tography using silica gel 60 F254 plates (BDH) in various
eluting solvent systems. The products of the reactions of 1 and
2 in 65 wt.% aqueous H2SO4 were identified by UV–vis,
NMR, and GC–MS analysis following direct isolation as pre-
viously described (20); some of the isolated products, 3–6, are
shown above. The MS results were particularly useful for
© 1998 NRC Canada

898
Can. J. Chem. Vol. 76, 1998
Table 1. Isolated products and yields in the reactions of 1 and 2 in 65 wt.% H2SO4.
Substrate
Product
Yield, %
Total yield, % a
48
1
(5 h)
4-(4-Chlorophenylazo)pyridine 2
4-Aminopyridine
22
23
21
4
4-Chloroaniline
4-(2,4-Dichlorophenylhydrazo)pyridine 5
4-(2,4-Dichlorophenylazo)pyridine 4
4-Aminopyridine
1
(24 h)
21
25
20
5
50
57
2,4-Dichloroaniline
4-(4-Chlorophenylazo)pyridine 2
4-Chloroaniline
3
Dimeric product 6
Trace
15
13
14
26
10
2
2
4-(2-Chloro-4-hydroxyphenylazo)pyridine 3
4-(2,4-Dichlorophenylazo)pyridine 4
2-Chloro-4-hydroxyaniline
4-Aminopyridine
2,4-Dichloroaniline
4-(4-Hydroxyphenylazo)pyridine 7
Dimeric product 6
2
a Since two aniline products result from one substrate molecule, their yields have to be averaged in calculating the overall yield,
which is why the numbers in this column are not the simple sum of those in the previous column.
Scheme 1.
identifying the chlorine-containing products; for instance, the
mass spectrum of 4 with its two chlorine atoms showed a char-
acteristic intensity pattern of 100:65.3:10.6, in which each
peak is separated by two mass units due to the natural abun-
dance of the two chlorine isotopes, 35Cl:37Cl being 100:32.5.
The structural identification of the products was achieved us-
1
ing H NMR spectroscopy and MS. For 3: δH (DMSO-d6):
12.82 (OH), 8.79 (H3), 7.83 (H2), 7.68 (H6′), 7.17 (H3′), 6.98
(H5′); m/z: 233.0 (M). For 4: δH (DMSO-d6): 8.86 (H3), 7.96
(H3′), 7.77 (H2), 7.73 (H6′), 7.60 (H5′); m/z: 251.0 (M). For 5: δH
(DMSO-d6): 8.90 (NHα), 8.14 (H3), 8.08 (NHβ), 7.48 (H3′),
7.20 (H5′), 6.74 (H6′), 6.67 (H2); m/z: 235.1 (M). For the dimer
6: δH (DMSO-d6): 9.98 (NH), 8.72 (H1), 8.14 (H5), 7.86 (H3),
7.65 (H2), 7.42 (H7,8), 7.16 (H4), 6.60 (H6); m/z: 401.2 (M+1).
Scheme 2.
The pKBH 2+ values of the azo substrate 2 and the azo prod-
2
ucts 3 and 4 were determined from the variation of the UV–vis
spectra with sulfuric acid concentration as previously de-
scribed (20); the acid concentration and extinction data are
given here as supplementary information, Tables S1–S3.3 The
kinetic techniques used were similar to those used before (20).
Results
Products
The products formed when 4-(4-chlorophenylhydrazo)pyri-
dine 1 is allowed to react in 65 wt.% H2SO4 for 5 h at a sub-
strate concentration of 0.05 M are given in Table 1, and
illustrated in Scheme 1. Under these conditions, and at the
much lower substrate concentrations used for the kinetic runs
(as determined from the UV–vis spectra after the kinetic runs),
the reaction gives the three products shown, 4-(4-chloropheny-
lazo)pyridine 2, 4-chloroaniline, and 4-aminopyridine. Trace
amounts of 4-(2,4-dichlorophenylhydrazo)pyridine 5 were
also present in the mixture analyzed.
If the reaction of 1 is allowed to continue for longer periods
the number of products increases, as shown in Table 1 and
illustrated in Scheme 2. The products still include 2, 4-chlo-
roaniline, and 4-aminopyridine, but 4-(2,4-dichloropheny-
lazo)pyridine 4 and 2,4-dichloroaniline are also present, with
trace amounts of the dimeric product 6. There is no evidence
for the presence of 5, leading to the inference that at least some
of the new products may have been formed from it. The pres-
ence of extra Cl– (1 had to be used as its hydrochloride salt)
3
Tables S1–S3, supplementary material, may be purchased from:
The Depository of Unpublished Data, Document Delivery,
CISTI, National Research Council of Canada, Ottawa, Canada,
K1A 0S2.
© 1998 NRC Canada

899
Cox et al.
Scheme 3.
Table 2. Azo-compound pKBH 2+ values obtained using the excess
2
acidity method.a
Compound
Substituents
pKBH
m*
2+
2
2
3
4
7b
8b
4-Cl
4-OH, 2-Cl
2,4-Di-Cl
4-OH
–6.47 ± 0.37
–4.15 ± 0.26
–6.61 ± 0.32
–3.07 ± 0.21
–4.97 ± 0.22
0.991 ± 0.070
0.760 ± 0.063
0.875 ± 0.053
0.994 ± 0.089
0.807 ± 0.045
H
a Errors quoted are standard deviations.
b From ref. 20.
Scheme 4.
behaviour, for instance), and spectra of both forms contribute
to the resulting pKBH 2+ values. The data were treated according
to the excess acidity2method; for equilibria the relevant equa-
+
+
tion is log I – log CH = pKBH 2+ + m*X (37), where CH is the
proton concentration in the sulfuric acid solution and X is the
2
excess acidity, both obtained from published sources (37). The
ionization ratio I is given in terms of δε by I =
+
+
(δεBH − δε)/(δε − δεBH 2+) (38), where δεBH and δεBH 2+ apply
2
2
to the pure monoprotonated form and the pure diprotonated
form, respectively. This expression was substituted in the ex-
cess acidity equation and the result recast in terms of δε and
curve-fitted directly, as before (38); previous experience (39)
shows that this method gives reliable protonation constants.
The slope parameter m* is characteristic of the substrate, hav-
ing values of 1.0 for primary nitroanilines, 0.6 for amides and
also complicates the situation. The product recovery in Table 1
is less than 100%, probably due to losses incurred during the
TLC technique used for product separation (20).
The first-formed product 2 also reacts in aqueous H2SO4,
albeit more slowly than does 1, and the products of the reaction
of 2 at 0.05 M substrate concentration were also investigated,
see Table 1. A complex mixture of 4, 4-(2-chloro-4-hy-
droxyphenylazo)pyridine 3, 4-(4-hydroxyphenylazo)pyridine
7, 2-chloro-4-hydroxyaniline, 2,4-dichloroaniline, 4-ami-
nopyridine, and the dimeric product 6 was found, as shown in
Scheme 3.
+
so on (37). The parameters given by the curve fit are δεH ,
δεBH 2+, pKBH 2+, and m*; the latter two are listed for 2–4 in
2
Table 2. Also2 listed in Table 2 are the pKBH 2+ and m* values
for 4-(4-hydroxyphenylazo)pyridine 7 and 2the parent com-
pound 8, already available (20), for comparison. The experi-
However, at the substrate concentrations typically used for
the kinetic runs, ~8 × 10–5 M, the situation for 2 is much sim-
pler, as shown in Scheme 4. The primary product is 3, which
could be identified by its UV–vis spectrum; the concentration
of this product was shown not to depend on the medium acid-
ity, averaging (46.8 ± 4.0)% over the 65–81 wt.% H2SO4 con-
centration range, as determined from the known extinction
coefficients and the observed absorbances after 10 half-lives.
The two aniline products also formed could not be directly
identified, but undoubtedly form in equivalent amounts (20).
Evidently the reaction is uncomplicated; according to Table 1
two moles of 2 give one mole of oxidized products (total 30%
azo compounds) and one mole of reduced products with the
N—N bond broken (total 24% anilines, 26% 4-aminopyrid-
ine).
+
mental data obtained and the values of X and log CH used for
2–4 are given as supplementary material in Tables S1–S3,3
+
together with the calculated δεBH and δεBH 2+ values.
2
Kinetics
The kinetics of the reaction of 4-(4-chlorophenylazo)pyridine
2 (see Scheme 4) were monitored spectrophotometrically by
following the disappearance of the reactant, in either its mono-
protonated SH+ form at 354 nm (up to 72 wt.% H2SO4), or its
diprotonated SH22+ form at 444 nm (at the higher acidities).
The observed pseudo-first-order rate constants as a function of
acidity, kψ, and other parameters relevant to the kinetic analy-
sis, are given in Table 3. All the kinetic plots showed excellent
first-order linearity over the three half-lives for which the re-
action was typically followed.
+
−
[1]
log kψ − log (CSH 2+ / (CSH + CSH 2+)) − log aNu
2
2
pKBH 2+ values
= log k0 + (m‡ − 1)m X
2
The azo compounds 2–4 are all protonated on the pyridine
nitrogen in the pH range (8), so the protonation of interest in
this work is the second protonation at the azo group. The
Previously it was found that eq. [1] describes the kinetic
behaviour of 4-(phenylazo)pyridine (8) (and that of 2-(pheny-
lazo)pyridine) at these acidities (20), and the behaviour of 2
was found to be very similar. Protonation of the substrate is a
pKBH values were measured as described before (20). Ex-
2+
2
+
2+
tinction differences (δε = εBH − εBH ) were obtained from
UV–vis spectra measured in different2sulfuric acid solutions at
wavelengths appropriate for the peak maxima of the pro-
tonated and diprotonated forms; this is the Davis–Geissman
method (36), preferred because errors due to small differences
between the individual samples tend to cancel out when the
difference is taken (these may cause irregular isosbestic point
pre-equilibrium process (20), and so the pKBH 2+ and m* values
2
given in Table 2 were used with the equilibrium excess acidity
equation to calculate protonation correction terms (PCT) on
the
basis
of
full
substrate
protonation,
log
+
(CSH 2+ / (CSH + CSH 2+)), or log (I/(1 + I)) after division by
2
2
+
+
CSH (since the ionization ratio I = CBH 2+ / CBH ), which correct
2
© 1998 NRC Canada

900
Can. J. Chem. Vol. 76, 1998
Table 3. Kinetic data for the hydroxylation of
4-(4-chlorophenylazo)pyridine 2.
Fig. 1. Excess acidity plots for the reactions of 2 (open circles) and
8 (closed circles) (20) in aqueous sulfuric acid, assuming reaction
with one bisulfate ion.
a
b
+
H2SO4,
wt.%
kψ
X b
log CH
log CHSO
PCT c log kψ
–
4
– PCT
64.99
2.01 3.793
2.03
1.078
1.100
1.122
1.144
1.154
1.157
1.118
0.932
0.975
1.020
1.071
1.096
1.118
1.102
–1.640 –3.057
–3.053
68.22
71.67
75.86
78.28
81.36
85.98
2.90 4.208
2.91
–1.224 –3.314
–3.312
7.32 4.708
7.35
–0.762 –3.373
–3.372
9.12 5.397
9.16
–0.289 –3.751
–3.749
7.99 5.830
8.01
–0.128 –3.969
–3.968
3.91 6.407
3.92
–0.038 –4.370
–4.369
0.81 7.274
0.81
–0.006 –5.086
–5.086
a Observed pseudo-first-order rate constants, ×105 s–1. Both duplicate
determinations given.
b From ref. 37.
c Protonation Correction Term; log (CSH 2+ / (CSH + CSH 2+)). Calculated
Table 4. Kinetic data for the disproportionation of
4-(4-chlorophenylhydrazo)pyridine 1.
+
2
2
using pKBH 2+= –6.47 and m* = 0.991 from Table 2.
2
H2SO4,
wt.%
kψ
X b
log CH
PCT c
log kψ
a
b
+
for less than full substrate protonation at the acidity under
investigation. The nucleophile involved at these acidities is
bisulfate ion, and log bisulfate ion concentrations were used
–PCT
59.96
0.0384
0.0386
0.162
0.163
0.362
0.363
0.839
0.843
2.02
3.234
3.793
4.208
4.708
5.397
5.830
6.407
7.274
1.039
1.078
1.100
1.122
1.144
1.154
1.157
1.118
–0.644
–0.270
–0.119
–0.039
–0.008
–0.003
–0.001
0
–3.772
–3.769
–3.520
–3.518
–3.322
–3.321
–3.037
–3.035
–2.687
–2.687
–2.518
–2.513
–2.263
–2.260
–1.777
–1.772
−
for log aNu as before (20). Log k0 refers to the rate constant
64.99
68.22
71.67
75.86
78.28
81.36
85.98
for the reaction in the hypothetical ideal 1 M acid solution
which is the standard state (37), and m‡ refers to the transition
state, having values of >1 for an A1 reaction, <1 for an A-SE2
reaction, and probably ~1 for an A2 reaction with water (25).
Plots according to eq. [1] are given in Fig. 1 for 2, with that
for the parent compound 8 for comparison (20). For the reac-
tion of 8 with bisulfate, log k0 = 0.36 ± 0.11 and m*(m‡ – 1)
= –0.645 ± 0.018, i.e. m‡ = 0.201 ± 0.050 (20). In this work we
find, for the reaction of 2 with bisulfate, log k0 = –1.60 ± 0.13
and m*(m‡ – 1) = –0.612 ± 0.024, i.e. m‡ = 0.383 ± 0.031.
2.02
3.01
3.05
5.44
5.48
16.7
16.9
a Observed pseudo-first-order rate constants, ×103 s–1. Both duplicate
determinations given.
b From ref. 37.
c Protonation Correction Term; log (CBH 2+ / (CBH + CBH 2+)). Calculated
The kinetics of the disproportionation of 1 could be con-
veniently followed spectrophotometrically over the 60–86
wt.% H2SO4 range, by monitoring the formation of the azo
product, 4-(4-chlorophenylazo)pyridine 2 (see Scheme 1) at
either 354 nm (up to 72 wt.% H2SO4, monoprotonated form)
or at 444 nm (at the higher acidities, diprotonated form). Re-
actions were generally followed over three half-lives, and all
showed excellent first-order behaviour over this range. The
observed pseudo-first-order rate constants, kψ, are given in Ta-
ble 4, together with other pertinent information.
+
2
2
using the pKBH 2+ value of –4.806 (m* = 1) given in Table 5.
2
These rate data were also treated according to the excess
+
acidity method (25). To this end plots of log kψ – log CH vs.
X for 1, together with the previously published equivalent plots
for the parent 4-(phenylhydrazo)pyridine 9 and 4-(4-hy-
droxyphenylhydrazo)pyridine 10 for comparison (24), are
shown here as Fig. 2. The observed downward curvature in
Fig. 2 indicates (a) that the reaction is not an A-SE2 process,
as otherwise these plots would be straight, and (b) that the data
go through a pKa value (substrate titration) in an A1 reaction,
which means that a protonation correction term has to be ap-
plied (25). Reaction with water in an A2 process would also
© 1998 NRC Canada

901
Cox et al.
+
+
Fig. 2. Excess acidity plots of log kψ – log CH vs. X for the
reactions of 1 (open triangles), 9 (open circles) (24), and 10 (closed
circles) (24) in aqueous sulfuric acid, illustrating the downward
curvature due to substrate titration. The points are experimental and
the curves theoretical, see text.
Fig. 3. Excess acidity plots of log kψ − log (CBH 2+ / (CBH + CBH 2+))
2
2
(PCT) vs. X for the reactions of 1 (open triangles), 9 (open circles),
and 10 (closed circles) in aqueous sulfuric acid.
+
for equilibria (log I = pKBH 2+ + m*X + log CH ), because these
hydrazo compounds are very similar to the H0 primary aniline
2
indicators, for which m* is known to have the value 1.0 (37).
Using eq. [2] means that we have three unknowns to calculate
Table 5. Second basicities and excess acidity slopes and intercepts
for the phenylhydrazopyridines 1, 9, and 10.a
for each compound, slope (m‡m*), intercept (log (k0/KBH 2+))
and pKBH 2+. A standard curve-fitting computer program was
2
Substrate Substituents
pKBH
log k0
–4.806 ± 0.045 –5.371 ± 0.052 1.4915 ± 0.0092
–0.59 ± 0.45 –4.15 ± 0.52 1.87 ± 0.12
–1.038 ± 0.090 –2.934 ± 0.098 1.872 ± 0.033
m‡
2+
used (40),2and the resulting, reasonably precise, pKBH 2+ values
are given in Table 5. These were then used to calculate proto-
2
2
1
9
4-Cl
H
nation
correction
+
terms
in
the
form
log
10
4-OH
(CBH 2+ / (CBH + CBH 2+)). Subtraction of the latter from log kψ
2
2
a Errors listed are standard deviations.
results in the very good straight lines in Fig. 3. This figure is
drawn on the basis of full substrate protonation, which more
accurately reflects the true situation at these acidities. The
slope is (m‡ –1)m* and the intercept is log k0 (25). These and
the observed m‡ values, with their errors, are also given in
Table 5. The m‡ values, all >1, are only consistent with an A1
mechanism (25). The Table 5 values were used to calculate the
theoretical curves in Fig. 2.
cause downward curvature in these plots (20, 27), but if that
were the case here the curvature would be different for the
different substrates, with much stronger curvature for 2 than
for 10, since the water activity decreases with acidity much
faster at the higher acidities (4). The observed (Fig. 2) very
similar curvature for 2 and 10 is much more likely to be due
to a substrate titration; indeed, the Fig. 2 plots straighten out,
as expected for a substrate protonation, rather than show the
increasing curvature that would be expected for a reaction with
water. This mechanistic distinction can now be made, with the
addition of the data for 2, whereas it could not without it (24).
The protonation involved is again the second one, undoubt-
edly on the hydrazo group, because these compounds are all
already protonated on the pyridine ring, this process occurring
Discussion
Reaction pathways
The products of the reactions of 1 and 2 in acid media, given
in Table 1 and in Schemes 1–4, are quite complex. Although
under kinetic conditions Schemes 1 and 4 prevail, it is never-
theless most convenient to begin with the elements that are
common to several schemes. We have established previously
(24) that 4-(phenylhydrazo)pyridines undergo a dispropor-
tionation in sulfuric acid media, for example, see Scheme 1,
very similar to the processes undergone by other hydrazo com-
pounds that have both 4-positions blocked (21, 41). Two sub-
strate molecules give an azo compound (oxidized product) and
two anilines (reduced products), as shown in Scheme 5. This
disproportionation is much faster than any reaction undergone
by the 4-(phenylazo)pyridine products (24); for the reactions
of 1 and 2 this is readily apparent from a comparison of the rate
constants in Tables 3 and 4. It is also very likely to be irre-
versible, since the reverse would be a termolecular process.
Deferring consideration of the mechanism until later, it is
apparent that as soon as a hydrazo compound is formed it will
react according to Scheme 5. The other products, apart from
the dimer, can then be accounted for as shown in Scheme 6.
in the pH range (8). The pKBH 2+ values for these compounds
are not known, and reaction is too fast to allow them to be
measured, but pKBH 2+ values can be estimated from the curva-
2
2
ture of the kinetic plots in Fig. 2. It has recently been shown
that quite precise values can be obtained in this way (29). The
modified excess acidity equation for an A1 reaction that ap-
plies here, accounting for the changing state of protonation
with acidity, is eq. [2] (25):
+
+
+
[2]
log kψ − log CH − log (CBH / (CBH + C 2+))
BH2
= log (k0/KBH 2+) + m‡m X
2
where k0 is the standard-state rate constant for the reaction. The
log
protonation
+
correction
term
in
eq.
[2],
+
(CBH / (CBH + CBH 2+)), becomes log (1/(1 + I)) upon division
2
+
by CBH , and we can put m* = 1 in the excess acidity equation
© 1998 NRC Canada

902
Can. J. Chem. Vol. 76, 1998
Scheme 5.
Scheme 7.
Scheme 6.
Scheme 8.
(see Fig. 1), but in the identified products it appears as OH,
since these sulfate ester products hydrolyze very rapidly to
phenols in more dilute acid, as we have shown previously (42),
and this will occur during the product work-up.) Chloride ion
is present during the reaction of 2 when it derives from 1, since
1 had to be used as its hydrochloride salt, and it accumulates
as Nu– displaces it from 2; it is the best nucleophile in the
system, and can react at the 2-position of the phenyl group in
2H22+ as shown on the right-hand side of Scheme 8. Tautomeri-
zation leads to the hydrazo compound 5H+, most of which will
react according to Scheme 5, giving the observed product 4
(Schemes 2 and 3). However, in the initial reaction of 2 itself
the chloride concentration is zero, and thus the rate-determin-
ing step involves diprotonated 2 and bisulfate ion, as shown
by the kinetic analysis; this is the left-hand pathway in
Scheme 8. Loss of chloride ion (a good leaving group) as
shown gives 7H22+. With free chloride ion now present, the
right-hand side of Scheme 8 can come into play, and 7H22+ can
also react further, as shown in Scheme 9. Presumably the bet-
ter nucleophile Cl– can react at the more crowded ortho site
(Schemes 8 and 9), whereas the far worse (and much bigger)
nucleophile HSO4– is limited to para attack.
The chlorination of 7H22+, shown in Scheme 9, will be faster
than the chlorination of 2H22+, shown at the right of Scheme 8,
because of the presence of the inductively electron-withdraw-
ing oxygen in the Nu group encouraging nucleophilic attack.
Again, fast tautomerization leads to a hydrazo compound,
11H+, which will immediately react via Scheme 5 to give the
observed product 3 (Schemes 3 and 4); in fact this is the first-
formed major product in the reaction of 2, according to the
kinetics and Scheme 4. There are other possibilities; for
Starting with 1, which has an equimolar amount of chloride
ion present, product formation takes place from top to bottom
on the left-hand side only, with no displacement by hydroxide
occurring; products on the right are not observed if the starting
material is 1. On the other hand, if the starting material is 2
there is no free chloride ion present initially and the first reac-
tion is 2 → 7, with subsequent product formation taking place
down both sides of Scheme 6, chloride ion accumulating as it
is displaced. The hydrazo product in brackets is the only one
we did not observe, presumably because it is too reactive.
The reaction of 2H+ begins with a second protonation (the
reaction is acid catalyzed), undoubtedly on the azo group.
There are two nitrogens that can be protonated, as shown in
Scheme 7. It does not particularly matter which azo nitrogen
bears the proton, since the nitrogen lone pairs do not form part
of the π system, but it seems reasonable to place it on the
nitrogen furthest from the pyridinium positive charge, and the
pKBH 2+ values quoted in Table 2 probably refer to this. How-
ever, 2it is easier to draw the mechanistic schemes for sub-
sequent reactions with the proton on the other nitrogen, 2H22+
in Scheme 7.
Two different mechanistic possibilities for 2H22+ are given
in Scheme 8. (In Scheme 8 and subsequently the attacking nu-
cleophile is identified as either Cl– or as Nu–; the latter is
actually HSO4– in these media, as shown by the kinetic analysis
© 1998 NRC Canada

903
Cox et al.
Scheme 9.
Scheme 11.
Scheme 10.
Scheme 12.
instance, it seems reasonable that 4H22+ could undergo a chlo-
rine displacement as in Scheme 8 to give 3H22+ (and hence the
isolated products should contain more 3 than 4, see Table 1),
and any two hydrazo compounds can take part in the dispro-
portionation of Scheme 5. It is also possible (although prob-
ably less likely) that the 4-chloro- and 4-hydroxyanilinium ion
products formed can be chlorinated at the 2-position by the
+
free chloride ion (the presence of the NH3 group increasing
susceptibility to nucleophilic attack), perhaps giving rise to
some of the observed 2,4-dichloro- and 2-chloro-4-hydroxy-
anilinium products.
having an unprotonated pyridine nitrogen present decreases
with increasing acidity, and, being bimolecular, the reaction
will only occur at all at the high substrate concentrations used
for the product analysis.
The only remaining major product to be accounted for is
the dimer 6. It has been mentioned that chloride ion is the best
nucleophile in the system, but if any starting material that is
not protonated on the pyridine nitrogen is present, this will be
quite nucleophilic as well. The simplest dimerization possibil-
ity is a nucleophilic attack by 1 on 2H22+, as shown in
Scheme 10. (It is likely that 1 will be protonated somewhere
on the hydrazo group; for simplicity it is shown as being un-
protonated in Scheme 10.) Loss of chloride ion from the re-
sulting intermediate is straightforward, and simple loss of two
more protons gives the observed dimer product 6.
pKa relationships
The pKBH 2+ values for the various 4-(arylphenylazo)pyri-
2
dines listed in Table 2 seem reasonable, with compounds hav-
ing electron-donating OH groups being easier to protonate on
the azo group, and those with electron-withdrawing Cl being
more difficult. There are insufficient data for a good linear free
energy relationship (some of the substituents are ortho in any
case), but a ρ+ value of –2.9 ± 0.6 can be estimated. The m*
slopes show considerable variation, averaging 0.86 ± 0.09. The
reactions undergone by 2 and 8 are not quite the same, that of
2 being an ipso substitution at a carbon already bearing a sub-
stituent but, as Fig. 1 shows, the kinetic behaviour is very simi-
lar, and reaction of the deactivated chloro compound is slower,
by a factor of ~100 in the standard state according to the inter-
cept difference.
Another possibility results from an attack of 2H+ on 2H22+
,
as shown in Scheme 11. Essentially, chlorine has to be lost as
Cl+ in this scheme, but this has precedence in reactions of this
type; Rhee and Shine have observed a similar Cl+ loss during
the benzidine rearrangement and disproportionation of 4,4′-di-
chlorohydrazobenzene (43). Whatever the dimerization
mechanism may be, presumably dimerization will diminish in
importance as the acidity increases, since the probability of
The m‡ values for both 2 and 8 are <0.5. At the moment
© 1998 NRC Canada

904
Can. J. Chem. Vol. 76, 1998
Scheme 13.
Scheme 14.
there are very few data for comparison, and so it is difficult to
predict m‡ values for reaction with bisulfate. In the case of
protonated benzamide a value of 1.1 was found for nucleo-
philic attack of bisulfate at the carbonyl group (44) (but this
number is subject to revision (26)), and for several substituted
alkylnitramines values of m*m‡ between 0.9 and 1.2 were
found for SN2 displacement by bisulfate (27), the value of m*
unfortunately not being known. Neither case corresponds very
well to the situation in the reaction under discussion here, nu-
cleophilic attack on a benzene ring.
+
[4]
log kψ − log (CBH 2+ / (CBH + CBH 2+))
2
2
= log k0 + (m‡ − 1)m X
Upon replacing activity by concentration and activity coeffi-
cient, taking logs and rearranging, eq. [3] becomes eq. [4],
since the log activity coefficient ratio term here is equal to
(m‡ – 1)m* (25), where the slope parameters m* (= 1 in this
case) and m‡ are defined above. This is the rate equation that
gives the good straight lines in Fig. 3.
Benzidine disproportionation mechanism
Although Scheme 13 explains the products and the kinetics,
there is a problem with it. The proposed chloronium interme-
diate is actually a nitrenium ion:
Proposed mechanisms for the benzidine disproportionation
of 1 (and the other 4-(phenylhydrazo)pyridines involved in this
work) have to account for the observed kinetic behaviour,
which is that the rate-determining step of these disproportiona-
tions is an A1 process involving the diprotonated substrate, as
well as the reaction products. The second protonation is written
as occurring on the hydrazo nitrogen away from the pyrid-
inium ring. There are two reasons for this: apart from simple
electrostatic repulsion of the two positive charges, the lone pair
on the other nitrogen is occupied to some extent in resonance
interaction with the pyridinium centre, as shown, and so is less
available for protonation.
and these are highly reactive, with very short lifetimes (45).
They have carbene-like properties, and would be expected to
react almost instantly with the water present, probably on the
aromatic ring (46). So Scheme 13 can be ruled out on the
grounds that this nitrenium ion would not exist long enough for
the suggested hydride transfer from another 1H+ molecule to
be possible.
Thus the most likely mechanism for the reaction of 1, 9, and
10, and the other hydrazo compounds suggested as intermedi-
ates in this work, is the benzidine-type disproportionation
mechanism shown in Scheme 14 for the general structure BH+,
as proposed previously (24). This would give the same kinetic
behaviour as does Scheme 13, i.e., eq. [4] is followed. The
rate-determining step is a thermally allowed 10-electron elec-
Processes such as the one shown in Scheme 12, although it
gives the correct products, can be ruled out as Scheme 12
would result in second-order kinetics, and all of the observed
kinetics were accurately first order, for 1 and also for 9 and 10
(24).
Scheme 13, involving rate-determining breakup of dipro-
tonated 1H22+, is a more plausible mechanism. Deriving an
excess acidity rate equation based on Scheme 13 is straight-
forward (25). On the basis of full substrate diprotonation, the
observed rate constants kψ can be equated with k0 and the
activities of the species present in the activated complex ac-
cording to eq. [3].
trocyclic reaction involving the diprotonated substrate BH22+
,
giving rise to intermediate 12. This is now firmly established
as being the key step in the rearrangement mechanism for hy-
drazobenzene and many other hydrazo aromatics, primarily by
the application of heavy-atom isotope effect studies to these
processes (21, 23). Also, for hydrazobenzene itself, an inter-
mediate equivalent to 12 has been experimentally observed
under stable-ion conditions (47). Regular hydrazo compounds
can form benzidines by rearomatization of intermediates like
12, but this cannot happen here because the pyridine nitrogen
would have to acquire a permanent positive charge, and
+
[3]
kψ (CBH + CBH 2+) = k0a 2+/f‡
BH2
2
© 1998 NRC Canada

905
Cox et al.
because of the presence of the substituents. So disproportiona-
tion occurs instead, and reaction with another BH+ molecule
is proposed by analogy with a similar proposal made for the
disproportionation of 4,4′-diiodohydrazobenzene (21, 41).
This is likely to be quite a fast reaction, because (a) it is a
highly favourable, thermally allowed 14-electron electrocyclic
reaction, (b) it is energetically downhill because the aromatic
protonated products formed are all much more stable than is
12 in the reaction medium, and (c) formation of three mole-
cules from two counteracts the entropy factor involved in ori-
enting 12 and BH+ for reaction.
The benzidine rearrangement of hydrazobenzenes is a fast
reaction (1). Our reactions are quite fast, but rate comparisons
are difficult because most other kinetic studies have been per-
formed in dilute acid media containing a cosolvent, and the
acid used is usually HCl or HClO4 rather than H2SO4 (1). The
most recent study is that of Bunton and Rubin (48), who stud-
ied hydrazobenzene (i.e., 1,2-diphenylhydrazine) and two sub-
stituted hydrazobenzenes in pure aqueous HCl and HClO4 to
concentrations of about 1 M; their substrates were dipro-
tonated (on both hydrazo nitrogens), and their reaction rates
were much faster than ours, in the stopped-flow range.
Depending on the substrate, the hydrazo group may be
diprotonated, monoprotonated, or even unprotonated during
the rearrangement step, but the process is fastest for dipro-
tonated substrates and slowest for unprotonated ones (1, 21,
48). Our substrates are diprotonated but one of the protons is
on the pyridine nitrogen; because of this, protonation on the
other hydrazo nitrogen (to give a triprotonated species) is ex-
pected to be very difficult, only occurring at acidities much
higher than those used here. Diprotonation in hydrazobenzenes
is assumed to speed reaction by encouraging N—N bond
cleavage and the subsequent charge separation (1). Hydrazo
group protonation presumably does that here too, but since the
two positive charges are further apart in our case the reaction
will not be accelerated nearly so much, which probably ac-
counts for our rates being slower than those of Bunton and
Rubin.
to cleavage of the N—N bond as shown at the top of
Scheme 13, and this was not observed.
The m‡ values in Table 5 are all quite large, 1.5–1.9, values
that are only compatible with an A1 reaction mechanism (25),
as discussed. However, they are not all the same; the value for
1, with a p-Cl substituent, is smaller than the others. Variation
of m‡ with substituent is not unusual; for instance, a gradual
increase with σ* was observed in N-nitro amine decomposi-
tions in aqueous H2SO4 (27), a slow decrease with σ+ was seen
in some phenylpropiolic acid hydrations in these media (50),
and slow increases or decreases with σ or σ+ were found for
some acylal and thioacylal hydrolyses (51), and some thiol-
and thion-benzoate ester hydrolyses, in aqueous H2SO4 (52).
What, if anything, these trends mean is not clear at present,
although as data accumulate it may become possible to use
them for mechanistic diagnosis. It is interesting that Marziano,
Tomasin, and Tortato have commented upon similar trends in
m* value in a variety of aqueous acid systems (53), although
it is not certain with the data available at present that the ad-
duced variations are real.
Hydroxylation of the (arylazo)pyridines is an unusual reac-
tion. Normal azobenzenes are stable in strong acid solution,
indeed, acidity functions based on their protonation behaviour
have been defined (54). Presumably the strongly electron-
withdrawing pyridinium substituent in these molecules
strongly activates the other ring to nucleophilic attack. The
benzidine rearrangement and related processes have always
fascinated chemists, and interest in them continues to be high
(1, 21–24). For instance, an 18-electron thermally allowed
electrocyclic reaction in a benzidine rearrangement has been
described very recently (55), which complements the 10- and
14-electron processes discussed in this work. To our knowl-
edge, the work in this paper is the first definitive delineation
of the mechanism of the acid-catalyzed benzidine dispropor-
tionation reaction as an A1 process.
Acknowledgments
Financial support of this research through a grant by the Natu-
ral Sciences and Engineering Research Council of Canada (to
E.B.) is gratefully acknowledged. We would also like to thank
a referee for helpful comments.
Mechanisms are sometimes suggested for the benzidine re-
arrangement in which the second proton transfer is part of the
rate-determining step (A-SE2 mechanism). For instance, rate-
limiting ring carbon protonation followed by fast N—N bond
cleavage (47) was a popular mechanism at o+ne time (1), as was
+
References
rate-determining formation of the −NH2 − NH2− species with
its subsequent rearrangement being a fast process (48). More
recent heavy-atom isotope effect results (49) have rendered
these mechanisms unlikely, however (21). For the dispropor-
tionations in this paper we have good evidence that the second
protonation is a pre-equilibrium process (A1 mechanism) and
not part of the rate-determining step: the excellent linearity of
the Fig. 3 plots, and the fact that the m‡ values (Table 5) are
all much greater than one (25).
1. R.A. Cox and E. Buncel. In The chemistry of the hydrazo, azo
and azoxy groups. Vol. 1. Edited by S. Patai. Wiley, London.
1975. p. 775.
2. E. Buncel and B.T. Lawton. Can. J. Chem. 43, 862 (1965).
3. E. Buncel and W.M.J. Strachan. Can. J. Chem. 48, 377 (1970).
4. R.A. Cox. J. Am. Chem. Soc. 96, 1059 (1974).
5. R.A. Cox and E. Buncel. J. Am. Chem. Soc. 97, 1871 (1975); E.
Buncel, R.A. Cox, and A.J. Dolenko. Tetrahedron Lett. 215
(1973); R.A. Cox, A.J. Dolenko, and E. Buncel. J. Chem. Soc.
Perkin Trans. 2, 471 (1975).
The log k0 values in Table 5 fall in the expected order,
electron-donor OH fastest, then H, with the electron-with-
drawing Cl slowest. The observed pKBH 2+ values (Table 5) are
indicative of only inductive interacti2on with the positive
charge on the hydrazo group; both OH and Cl are inductively
electron withdrawing, and both 1 and 10 are more basic than
9. Any attempted mesomeric interaction of the lone pairs on
oxygen with the positive charge on the hydrazo nitrogen leads
6. E. Buncel. Acc. Chem. Res. 8, 132 (1975).
7. E. Buncel and S.-R. Keum. J. Chem. Soc. Chem. Commun. 578
(1983).
8. E. Buncel and S.-R. Keum. Tetrahedron, 39, 1091 (1983).
9. S. Rajagopal and E. Buncel. Dyes Pigm. 17, 303 (1991).
10. E. Buncel and S. Rajagopal. J. Org. Chem. 54, 798 (1989); Acc.
Chem. Res. 23, 226 (1990).
© 1998 NRC Canada

906
Can. J. Chem. Vol. 76, 1998
11. S.-R. Keum, S.-S. Hur, P.M. Kazmaier, and E. Buncel. Can. J.
Chem. 69, 1940 (1991).
Virta, and O. Kylläinen. Acta Chem. Scand. 48, 122 (1994), and
earlier papers.
12. S.-R. Keum, K.-B. Lee, P.M. Kazmaier, R.A. Manderville, and
E. Buncel. Magn. Reson. Chem. 30, 1128 (1992).
13. S.-R. Keum, K.-B. Lee, P.M. Kazmaier, and E. Buncel. Tetrahe-
dron Lett. 35, 1015 (1994).
14. S.-R. Keum, S.-S. Lim, B.-H. Min, P.M. Kazmaier, and E.
Buncel. Dyes Pigm. 30, 225 (1996).
15. S.-R. Keum, P.M. Kazmaier, K.-S. Cheon, R.A. Manderville,
and E. Buncel. Bull. Korean Chem. Soc. 17, 391 (1996).
16. S. Swansburg, Y.-K. Choi, S.-R. Keum, E. Buncel, and R.P.
Lemieux. Liq. Cryst. In press (1998).
32. R.A. Cox. J. Phys. Org. Chem. 4, 233 (1991).
33. M. Lajunen, R. Laine, and M. Aaltonen. Acta Chem. Scand. 51,
1155 (1997), and earlier papers.
34. M. Ali and D.P.N. Satchell. J. Chem. Soc. Perkin Trans. 2, 167
(1995); J. Chem. Soc. Perkin Trans. 2, 1825 (1993).
35. E. Koenigs, W. Freigang, G. Lobmayer, and A. Zscharn. Ber.
Dtsch. Chem. Ges. 59B, 321 (1926).
36. C.T. Davis and T.A. Geissman. J. Am. Chem. Soc. 76, 3507
(1954).
37. R.A. Cox and K. Yates. J. Am. Chem. Soc. 100, 3861 (1978).
38. R A. Cox and K. Yates. Can. J. Chem. 59, 1560 (1981).
39. R.A. Cox, L.M. Druet, A.E. Klausner, T.A. Modro, P. Wan, and
K. Yates. Can. J. Chem. 59, 1568 (1981); R.A. Cox and K. Yates.
Can. J. Chem. 59, 2116 (1981); Can. J. Chem. 62, 2155 (1984).
40. P.R. Bevington. Data reduction and error analysis for the physi-
cal sciences. McGraw–Hill, New York. 1969. p. 237.
41. H.J. Shine, J. Habdas, H. Kwart, M. Brechbiel, A.G. Horgan, and
J. San Filippo. J. Am. Chem. Soc. 105, 2823 (1983).
42. E. Buncel and W.M.J. Strachan. Can. J. Chem. 47, 911 (1969).
43. E.-S. Rhee and H.J. Shine. J. Am. Chem. Soc. 108, 1000 (1986).
44. R.A. Cox and K. Yates. Can. J. Chem. 59, 2853 (1981).
45. R.A. McClelland. Tetrahedron, 52, 6823 (1996); J.C. Fishbein
and R.A. McClelland. Can. J. Chem. 74, 1321 (1996).
46. M. Novak, M.J. Kahley, J. Lin, S.A. Kennedy, and L.A.
Swanegan. J. Am. Chem. Soc. 116, 11626 (1994).
47. G.A. Olah, K. Dunne, D.P. Kelly, and Y.K. Mo. J. Am. Chem.
Soc. 94, 7438 (1972).
48. C.A. Bunton and R.J. Rubin. J. Am. Chem. Soc. 98, 4236 (1976).
49. H.J. Shine, G.N. Henderson, A. Cu, and P. Schmid. J. Am. Chem.
Soc. 99, 3719 (1977).
50. R.A. Cox, E. Grant, T. Whitaker, and T.T. Tidwell. Can. J.
Chem. 68, 1876 (1990).
51. R.A. Cox and K. Yates. J. Org. Chem. 51, 3619 (1986).
52. R.A. Cox and K. Yates. Can. J. Chem. 60, 3061 (1982).
53. N.C. Marziano, A. Tomasin, and C. Tortato. Org. React. (Tartu),
30, 39 (1996).
17. S.-R. Keum, M.-J. Lee, S. Swansburg, E. Buncel, and R.P.
Lemieux. Dyes Pigm. In press (1998).
18. E. Buncel and I. Onyido. Can. J. Chem. 64, 2115 (1986).
19. R. A. Cox, I. Onyido, and E. Buncel. J. Am. Chem. Soc. 114,
1358 (1992).
20. K.-S. Cheon, R.A. Cox, S.-R. Keum, and E. Buncel. J. Chem.
Soc. Perkin Trans. 2, 1231 (1998).
21. R.A. Cox and E. Buncel. In The chemistry of the hydrazo, azo
and azoxy groups. Vol. 2. Edited by S. Patai. Wiley, London.
1997. p. 569.
22. D.V.Banthorpe, E.D. Hughes, and C.K. Ingold. J. Chem. Soc.
2864 (1964); D. V. Banthorpe. Top. Carbocycl. Chem. 1, 1
(1969); H.J. Shine. Aromatic rearrangements. Elsevier, Amster-
dam. 1967. p. 126; In Mechanisms of molecular migrations. Vol.
2. Edited by B.S. Thyagarajan. Wiley, New York. 1969. p. 191.
23. H.J. Shine. J. Phys. Org. Chem. 2, 491 (1989); In Isotopes in
organic chemistry. Vol. 8. Edited by E. Buncel and W.H. Saun-
ders. Elsevier, Amsterdam. 1992. p. 1.
24. E. Buncel and K.-S. Cheon. J. Chem. Soc. Perkin Trans. 2, 1241
(1998).
25. R.A. Cox. Acc. Chem. Res. 20, 27 (1987); R.A. Cox and
K.Yates. Can. J. Chem. 57, 2944 (1979).
26. R.A. Cox. Can. J. Chem. 76, 649 (1998).
27. R.A. Cox. Can. J. Chem. 74, 1774 (1996).
28. R.A. Cox. J. Chem. Soc. Perkin Trans. 2, 1743 (1997).
29. R.A. Cox. Can. J. Chem. 75, 1093 (1997).
30. R.A. Cox, D.B. Moore, and R.S. McDonald. Can. J. Chem. 72,
1910 (1994).
54. H.H. Jaffé and R.W. Gardner. J. Am. Chem. Soc. 80, 319 (1958);
S.-J. Yeh and H.H. Jaffé. J. Am. Chem. Soc. 81, 3274 (1959);
A.J. Kresge and H.-J. Chen. J. Am. Chem. Soc. 94, 8192 (1972).
55. K.H. Park and J.S. Kang. J. Org. Chem. 62, 3794 (1997).
31. J.-P. Bégué, F. Benayoud, D. Bonnet-Delpon, A.D. Allen, R.A.
Cox, and T.T. Tidwell. Gazz. Chim. Ital. 125, 399 (1995); Y.
Chiang, A.J. Kresge, P.A. Obraztsov, and J.B. Tobin. Croat.
Chem. Acta, 65, 615 (1992), and earlier papers; M. Lajunen, M.
© 1998 NRC Canada