Rates and Products of Reactions of N-Benzoyloxy-2-nitrobenzenamine with Metal Alkoxides

Article abstract of DOI:10.1071/CH01110

A major pathway when N-benzoyloxy-2-nitrobenzenamine (1b) reacts with sodium methoxide is attack on carbonyl and production of methyl benzoate. In a competing pathway, proton abstraction is followed by loss of benzoate and formation of 2,1,3-benzoxadiazole N-oxide (4). With the more bulky t-butoxide base, only the latter pathway is followed, though the heterocyclic product undergoes subsequent destruction. Both the detection in the visible spectrum of a red transient intermediate, and a lack of an electron spin resonance spectrum, indicate it to be the anion (2b). The rates of the reactions of potassium t-butoxide with (1b) and N-benzoyloxy-4-nitrobenzenamine (7) are very similar, which rules out the possibility that the ortho-nitro group in (1b) provides neighbouring group assistance for loss of benzoate ion from the anion (2b). Rate measurements show that the loss of benzoate from (2b) is 39 times slower than chloride loss from the analogous anion (2a).

Full text of DOI:10.1071/CH01110

P u b l i s h i n g
AUSTRALIAN JOURNAL OF
CHEMISTRY
A N I N T E R N A T I O N A L J O U R N A L F O R C H E M I C A L S C I E N C E
publishing research papers from all fields of chemical
science, including synthesis, structure, new materials,
macromolecules, supramolecular chemistry, biological
chemistry, nanotechnology, surface chemistry, and
analytical techniques.
Volume 54, 2001
© CSIRO 2001
All enquiries and manuscripts should be directed to:
Dr Alison Green
Australian Journal of Chemistry–
an International Journal for Chemical Science
CSIRO
PUBLISHING
PO Box 1139 (150 Oxford St)
Collingwood, Vic. 3066, Australia
Telephone: +61 3 9662 7630
Fax: +61 3 9662 7611
E-mail: publishing.ajc@csiro.au
CSIRO
Published by
PUBLISHING
for CSIRO and the Australian Academy of Science
w w w. p u b l i s h . c s i r o . a u / j o u r n a l s / a j c

Aust. J. Chem. 2001, 54, 763–768
Rates and Products of Reactions of N-Benzoyloxy-2-nitrobenzenamine
with Metal Alkoxides
Leonard K. Dyall
School of Biological and Chemical Sciences, The University of Newcastle, Callaghan, NSW 2308,
Australia (e-mail: lendyall@yahoo.com.au).
A major pathway when N-benzoyloxy-2-nitrobenzenamine (1b) reacts with sodium methoxide is attack on carbonyl
and production of methyl benzoate. In a competing pathway, proton abstraction is followed by loss of benzoate and
formation of 2,1,3-benzoxadiazole N-oxide (4). With the more bulky t-butoxide base, only the latter pathway is
followed, though the heterocyclic product undergoes subsequent destruction. Both the detection in the visible
spectrum of a red transient intermediate, and a lack of an electron spin resonance spectrum, indicate it to be the
anion (2b). The rates of the reactions of potassium t-butoxide with (1b) and N-benzoyloxy-4-nitrobenzenamine (7)
are very similar, which rules out the possibility that the ortho-nitro group in (1b) provides neighbouring group
assistance for loss of benzoate ion from the anion (2b).
Rate measurements show that the loss of benzoate from (2b) is 39 times slower than chloride loss from the
analogous anion (2a).
Manuscript received: 30 August 2001.
Final version: 14 November 2001.
C
H
01110
L
-.B
K
e
z
o
p
l
o
x
-
2
-
t
r
b
e
n
z
n
a
m
i
e
r
e
ac
t
ions
w
i
h
M
t
a
A
l
o
x
i
e
s
Introduction
act as nucleophiles on amino nitrogen,[2,3] or acyl carbon,[3,4]
in (1b). The present paper reports experiments in which the
reaction described by Scheme 1 was achieved despite the
competing nucleophilic attack on acyl carbon with no
evidence seen for the attack on amino nitrogen. It has proved
possible to compare the relative abilities of chloride and
benzoate anions to act as leaving groups from the anion (2).
We have previously reported[1] the base-promoted stepwise
α-elimination of hydrogen chloride from N-chloro-2-ni-
trobenzenamine (1a) to form 2,1,3-benzoxadiazole* N-oxide
(4) (henceforth called benzofuroxan) (Scheme 1).
Results and Discussion
Reaction of N-Benzoyloxy-2-nitrobenzenamine (1b) with
Sodium Methoxide
When equal volumes of (1b) (7.9 × 10–4 M in methanol) and
sodium methoxide (0.10 M) were mixed, a red colour was
obtained instantaneously. In the first time-resolved
ultraviolet–visible (UV–vis) measurement (2 s after mixing),
the spectrum of (1b) (λmax 362 nm) was already replaced by
onewithmaximaat308and515nm,andashoulderat541 nm.
The red colour faded rapidly and was swamped by the
development of an intense blue colour (λmax 552 nm, with
shoulders at 512 and 585 nm). Weaker maxima at 353 and
371 nm corresponded to those of benzofuroxan, which
formed in yields of 16, 12, and 8% for sodium methoxide
concentrations (pre-mix) of 0.10, 0.20, and 1.0 M,
respectively. The 308 nm peak also faded as the reaction
progressed. The spectra underwent no further change after
60 s.
We found that mixing the base (methoxide or t-butoxide)
with the N-chloro compound produced a red colour that
faded rapidly (t 2 s at 0°C). Arguments based on the visible
½
spectrum of the red species, rate measurements on its fading,
the absence of an electron spin resonance (ESR) spectrum
during the reaction, and the steric effects of a 6-methyl
group, were used to formulate the above mechanism. The red
species was identified as the anion (2a), and we concluded
that the loss of the chloride ion to form the singlet nitrene (3)
is rate-determining.
It should be possible to achieve the chemistry outlined in
Scheme
1
with other starting materials such as
The λmax values of the blue species are close to those
reported by Kuhn and Weygand[5] for the conjugate base of
2-nitrophenylhydroxylamine (500, 546, and 587 nm, in
aqueous solution).
N-benzoyloxy-2-nitrobenzenamine (1b). This compound has
the advantage over (1a) of being stable. On the other hand, the
literature on N-acyloxyarenamines indicates that bases might
* Common name is benzofurazan.
© CSIRO 2001
10.1071/CH01110
0004-9425/01/120763

764
L. K. Dyall
A product run ((1b) 0.05 M, sodium methoxide 0.4M)
Similar results were obtained when the experiment was
repeated with 0.02 M alkoxide, except that the benzofuroxan
was consumed less rapidly.
yielded methyl benzoate (59%), benzofuroxan (53%), and
benzoic acid (50%). No attempt was made to isolate
2-nitrophenylhydroxylamine, which is unstable. Instead, it
was allowed to stand in alkaline solution and undergo aerial
oxidation to the corresponding nitroarene.[3] In the course of
five days, the yield of 1,2-dinitrobenzene increased from 4 to
40%. It is interesting that no 2,2´-dinitroazoxybenzene was
formed during the oxidation. Possibly its formation, by
nucleophilic addition of the arylhydroxylamine to the
intermediate nitroso compound, is prevented by severe steric
hindrance from the ortho-nitro group in each reactant.
Formation of methyl benzoate and the anion (5) indicated
that basic methanolysis (Scheme 2) is a serious competitor to
the reaction of α-elimination (Scheme 1) that we were
attempting to achieve. This methanolysis is very fast, which
indicates that the anion (5) is an extremely good leaving
group. Scheme 2 is undoubtedly a simplified version: it is
probable that (1b) reacts as its conjugate base and proton
transfers must occur.
The destruction of the benzofuroxan in a secondary
reaction complicated the study of reaction products. At the
high-substrate concentration of a product run, only 2%
benzofuroxan survived, and the only useful results were the
isolation of benzoic acid in 95% yield and the absence of
t-butyl benzoate. These results, and the absence of the blue
anion (5) in the time-resolved spectra, indicate that the
pathway shown in Scheme 2 is not followed with
t-butoxide base.
The best yields of benzofuroxan, measured
spectrophotometrically after 15 elapsed half-lives at 25°C,
were 87 and 77%. (Initial [KOBut] 0.01 and 0.05 M
respectively; initial [(1b)] 5.1 × 10–4 M.) At higher base
concentrations, yields were as low as 60%. It is probable that
conversion of (1b) to benzofuroxan in the primary reaction is
quantitative. The products of the secondary reaction did not
interfere with the spectroscopic measurement of the rate of
decay of the intermediate red species and, with the
competing reaction of Scheme 2 absent, it was practicable to
carry out the kinetic measurements described later.
The reason for methoxide following the path of
nucleophilic addition to carbonyl to a considerable extent,
and t-butoxide not doing so at all, cannot simply be
attributed to the latter base shifting the equilibrium
between (1b) and (2b) further to the right (see Scheme 1).
The time-resolved spectra show that all, or nearly all, of
the substrate (1b) is converted to the transient red species
immediately upon adding the large excess of base,
irrespective of whether it is methoxide or t-butoxide. These
observations are best accommodated by concluding that
both bases achieve essentially complete conversion of (1b)
to (2b). If we assume that the pKa of
N-benzoyloxy-2-nitrobenzenamine is not too different from
the value of 13.2 reported by Novak, Martin, and
The formation of benzofuroxan and benzoic acid in
essentially equal proportions is consistent with the
α-elimination mechanism shown in Scheme 1. The internal
trapping of the nitrene centre in (3) must be very efficient.
When Novak and co-workers[4] reacted N-pivaloyloxy-3(or
4-)-chlorobenzenamine with basic amines in methanolic
solution, they obtained the corresponding azo and azoxy
compounds, and arenamine. They attributed these products
to the reactions of an arylnitrene. Nitrene-derived products
(azepines, arenamines, and azo compounds) have also been
reported[6] in the reactions of N-sulfonyloxyarenamines with
amines. There are no such products from the reaction of our
substrate (1b) with sodium methoxide.
Heinrich[7]
for
N-pivaloyloxy-4-nitrobenzenamine
(methanol solution), then the argument is reasonable.
The failure of t-butoxide to displace the blue anion (5)
from N-benzoyloxy-2-nitrobenzenamine is most likely due
to severe steric hindrance to attack on the carbonyl group by
this bulky nucleophile.
Reaction of (1b) with Potassium t-Butoxide in t-Butanol
Mixing equal volumes of the alcoholic solutions of (1b) (9.76
× 10–4 M) and potassium t-butoxide (0.196 M) at 25°C
immediately produced a red colour. When the first spectrum
was recorded (t = 2 s), the broad band (λmax 363 nm) of (1b)
was no longer visible. New bands had appeared at 355 and
379 nm (benzofuroxan), and 516 and 542(sh) nm.
Extrapolation of the decay curve for the 516 nm peak to the
time of mixing gave ε ca.5000 cm2 mol–1. After 30 s, the 516
and 542 nm maxima were replaced by a single maximum at
460 nm, and at 120 s this band had resolved into a broad
maximum at 440 nm with a shoulder at 476 nm. At t = 5 min,
most of the benzofuroxan was gone, and none of it remained
after 10 min. At this time, the broad maxima of the product(s)
were at 449 and 478 nm, with a shoulder at 425 nm. This last
set of maxima corresponded reasonably well with those
observed in a separate experiment when benzofuroxan was
allowed to react with potassium t-butoxide.
The Nature of the Red Species Formed From (1b) and
Potassium t-Butoxide
The data indicate that the red species is the conjugate base
(2b) of N-benzoyloxy-2-nitrobenzenamine. We have argued
elsewhere,[1] for the N-chloro analogue (1a), that a λmax value
above 500 nm is too high for a singlet arylnitrene, and that
the bases used in this work are not powerful enough to
produce a Meisenheimer adduct. Another possibility, that the
red colour is due to a radical anion, is discounted by our ESR
experiments. The ESR signal observed when
N-benzoyloxy-2-nitrobenzenamine was treated with
potassium t-butoxide was much too feeble to belong to the
red intermediate. Moreover, after the red colour had all
faded, the observed ESR signal had actually become more

N-Benzpyloxy-2-nitrobenzenamine reactions with Metal Alkoxides
765
intense. It probably arose from the secondary reaction
between the benzofuroxan product and the t-butoxide ion.
If the mechanism portrayed in Scheme 1 is correct, then
the red species is the same irrespective of whether the base is
methoxide or t-butoxide. In the UV–vis spectra, the λmax
values of the red species (515 and 516 nm in methanol and
t-butanol respectively) match, as do the shoulders (541 and
542 nm respectively) on these peak maxima.
suggests its kinetics would be second-order in t-butoxide.
Therefore, it is not surprising that, in our experiments with
(1b), little benzofuroxan survived at high t-butoxide
concentrations. We noted some variation in the extent of this
benzofuroxan destruction, probably because it was
impractical to deoxygenate our reactants.
In our earlier kinetic work[1] with N-chloro-
2-nitrobenzenamine (1a) and t-butoxide base, benzofuroxan
(4) was produced in quantitative yield, and its subsequent
destruction by the base was not noted. However, in those
experiments the yield was measured spectrophotometrically
after the elapse of ten half-lives, which was only 25 s. No
significant destruction of (4) is to be expected in that brief
time,atalowertemperature(0insteadof25°C),andwithquite
low concentrations of base (≤ 0.074 M).
Sodium methoxide did not react with benzofuroxan. The
UV–vis spectrum of a solution 1 × 10–4 M in benzofuroxan
and 0.2 M in sodium methoxide did not change over a period
of 5 h.
It follows that the α-elimination of benzoic acid from (1b)
is a stepwise process.
The identification of the conjugate base (2b) as the red
intermediate has support from studies by Novak and
coworkers[4] with N-pivaloyloxy-4-nitrobenzenamine. This
compound reacted with methanolic dimethylamine to give a
red-orange solution, λmax 474 nm. This colour did not
develop when N-methyl-N-pivaloyloxy-4-nitrobenzenamine
was treated with the same base.
Another reaction pathway that must be considered is
attack on the arenamino nitrogen by the alkoxide nucleophile
(Scheme 3). Such reactions of N-acyloxyarenamines have
been reported[2,3] when the nucleophile is an amine. Related
substrates, O-diphenylphosphinoyl arylhydroxylamines
(ArNHOPOPh2), are reported[8] to undergo competing
α-elimination and attack on amino nitrogen when treated
with amine nucleophiles.
Kinetic Results for Reactions of (1b) with Potassium
t-Butoxide
The decompositions of the ion (2b) in t-butanol solvent were
smoothly first-order, even as far as five half-lives. There was
an effect of changing the concentration of potassium
t-butoxide (see Table 1), but it was not one to suggest a
kinetic dependence. At low values of [KOBut] (0.01–0.02 M)
the observed rate constants were appreciably lower than
those obtained in the range 0.05–0.25 M. At the lowest end
of the [KOBut] range, the ratio of base to substrate (1b) was
only 21:1, and may in reality have been much smaller since
no precautions were taken against uptake of carbon dioxide.
The absorbance of the red species in the kinetic runs with
very low base concentration was lower than usual, and the
corresponding low rate constants are attributed to an
incomplete conversion of (1b) to the anion (2b). The rate
constants attained a broad plateau with [KOBut] in the range
0.05–0.25 M, and then slowly decreased at still higher base
concentrations. This overall pattern of rate-constant
variation is strikingly similar to that we reported for the
N-chloro compound (1a). With the reactions between (1a)
and lithium methoxide, we demonstrated that lithium salts
The transient red intermediate would then be (6), and our
spectroscopic evidence that the same intermediate is formed
by both methoxide and t-butoxide bases cannot be
accommodated. Moreover, it is difficult to believe that the
first step in Scheme 3 would occur instantaneously on
mixing the reagents.
Reaction Between Benzofuroxan (4) and Potassium
t-Butoxide or Sodium Methoxide
When equal volumes of benzofuroxan (2.08 × 10–3 M in
t-butanol) and potassium t-butoxide (0.5 M) were mixed at
25°C, the peaks in the UV–vis spectrum due to substrate
steadily diminished in intensity. After 5 min, little
benzofuroxan remained, and none could be detected after 30
min. New peaks (λmax 432, with a shoulder at 480 nm) were
visible after 1 min, and became more intense until t = 30 min,
with the maximum shifting to 442 nm (480 nm shoulder). In
repeat experiments with pre-mix base concentrations of
0.0095 and 0.978 M, the amounts of benzofuroxan surviving
after 5 min were 100 and 0% respectively.
This reaction is presumed to be the one that consumes the
benzofuroxan produced when the N-benzoyloxy compound
(1b) is treated with potassium t-butoxide. ESR studies on
benzofuroxan by Russell, Janzen, and Strom[9] have shown
that electron transfer from the t-butoxide ion to
benzofuroxan is an efficient process, and work[10] on the
related electron transfer to aromatic nitro compounds
Table 1. Effect of varied [KOBut] on measured rate constants
for N-benzoyloxy-2-nitrobenzenamine (1b)
Initial [(1b)] 4.5 × 10–4 M. t-Butanol solvent
Temperature
K
Initial [KOBut]
M
102k1
s
102(deviation)A
298.2
0.00960
0.0239
0.0477
0.0960
0.162
0.245
0.343
0.489
03.94
06.28
08.38
08.74
08.67
08.52
07.73
06.49
0.11
0.09
0.04
0.05
0.10
0.17
0.10
0.05
413.2
0.096
0.162
27.30
27.80
1.00
0.60
A These are the deviations from the mean of duplicate runs.

766
L. K. Dyall
Table 4. Effect of varied [KOBut] on measured rate constants
added to the methanolic reaction mixture depressed the rate
constants. It is probable that similar salt effects operate at
high [KOBut] for the reactions of (1b), but other potassium
salts are not sufficiently soluble in t-butanol to check this
point experimentally.
for N-benzoyloxy-4-nitrobenzenamine (7) at 298.2 K
Initial [(7)] ca. 3 × 10–4 M. t-Butanol solvent
Initial [KOBut]
k1 (s)
Subsequent rate data for (1b) were obtained with [KOBut]
values in the range 0.05–0.25 M to avoid either incomplete
deprotonation or sizeable salt effects.
0.0477
0.096
0.162
0.245
0.215, 0.191
0.334, 0.329
0.513, 0.494
0.746
In Table 2, it is demonstrated that changing the
concentration of the substrate (1b) does not affect the
observed first-order rate constants. Thus, the fading of the
red-coloured species is a unimolecular process. Since there
was no evidence for neighbouring group participation by the
ortho-nitro group (see next section of this discussion) the
mechanism described in Scheme 1 adequately describes the
overall reaction, with stage (2b)→(3) assumed to be
rate-determining. The insolubility of potassium benzoate in
t-butanol precluded direct testing of whether this step is
reversible. However, the observation of strict first-order
kinetics even as far as five half-lives suggests that benzoate
ion has no effect on the reaction rate.
observed value (–77 J K–1 mol–1) is a little larger than that
(–51) we observed[1] for N-chloro-2-nitrobenzenamine.
Possibility of Neighbouring Group Participation
In principle, the ortho-nitro group in the anion (2b) could
assist the loss of benzoate ion, and then the free nitrene (3)
would not be an intermediate. Scheme 4 represents one
possible formulation of this neighbouring group
participation.
Table 2. Effect of varied initial
[N-benzoyloxy–2-nitrobenzenamine (1b)] on rate constants
t-Butanol solvent.Temperature 298.2 K
Initial 104[(1b)]
M
Initial [KOBut]
M
102k1
s
05.1
08.8
04.5
10.3
0.0477
0.0477
0.2450
0.2450
8.41
8.34
8.52
8.53
Whether such assistance to benzoate ion loss actually
occurs can be determined by comparing rates for the
ortho-isomer (1b) and the para-isomer, N-benz-
oyloxy-4-nitrobenzenamine (7).
Slow unimolecular solvolyses have been reported[11,12]
for certain N-acyloxyarenamines. However, when
N-benzoyloxy-2-nitrobenzenamine was dissolved in either
methanol or t-butanol, the UV–vis spectrum did not change
in 48 h. The very fast reactions with t-butoxide described in
the present paper clearly involve a different sort of reaction
requiring a strong base, such as the postulated process
(2b)→(3).
The effects of temperature on rate (Table 3) reveal that the
entropy of activation is large and negative, which is
consistent with a requirement of substantial solvent
reorganization to solvate the departing benzoate ion. The
While the very low solubility of N-benzoyloxy-4-nitro-
benzenamine in t-butanol severely limited the scope of
kinetic experiments, sufficient data were obtained for our
present purpose. On mixing with potassium t-butoxide, the
peak maximum at 332 nm was immediately replaced by a
new one at 480 nm. This new spectrum can confidently be
assigned totheconjugatebaseof(7), beingcloseinλmax value
to the figure of 474 nm reported by Novak and his
coworkers[4] for the conjugate base of N-pivaloyloxy
-4-nitrobenzenamine (methanol solution).
This 480 nm peak underwent rapid first-order decay, and
a new peak developed at 390 nm. This 390 nm peak, which
corresponds in λmax to the anion of 4-nitro-phenyl-hydroxyl-
amine, also decayed, but much more slowly than the peak at
480 nm had done.
The observed pseudo-first order rate constants for decay
of absorbance at 480 nm (see Table 4) showed a precise
linear dependence on [KOBut], and extrapolation to zero
concentration of the base yielded a non-zero rate constant
(7.00 × 10–2 s–1). (See Equation 1.)
Table 3. Arrhenius data for reaction of
N-benzoyloxy-2-nitrobenzenamine (1b) with KOBut in ButOH
solution
Initial [(1b)] 4.8 × 10–4 M. Initial [KOBut] 0.162 M
Temperature
K
102k1
s
Arrhenius parametersA
298.2
303.2
308.2
313.2
8.77, 8.56 Eact 58.4 2.4 kJ mol–1
13.8, 14.2 ∆Sact –77.4 8.1 J K–1 mol–1
18.8, 17.8 loge A 21.163 6.100
27.2, 28.4
k1 = 0.0700 + 2.72[KOBut]
(1)
A The errors in Eact, ∆Sact, and loge A are quoted as 90% confidence
limits.
(correlation coefficient r = 0.9983)

N-Benzpyloxy-2-nitrobenzenamine reactions with Metal Alkoxides
767
This behaviour fits a scenario in which a bimolecular
process (presumed from the absorbance spectrum to be
displacement of the anion of 4-nitrophenylhydroxylamine)
runs parallel to the unimolecular loss of benzoate from the
conjugate base of (7).
identified as (2b), the conjugate base of (1b). The α-elimi-
nation therefore occurs stepwise, and a comparison of reac-
tion rates with those of N-benzoyloxy-4-nitrobenzenamine
show that there is no neighbouring-group participation in the
loss of benzoate ion from the species (2b).
The rate constant for the unimolecular reaction (k1 = 7.0
× 10–2 s–1) is so close to the value of 8.67 × 10–2 s–1 recorded
for N-benzoyloxy-2-nitrobenzenamine that one must
conclude there is no significant neighbouring group
participation when the conjugate base of the latter compound
loses benzoate ion.
Kinetic studies found that benzoate expulsion from this
intermediate (2b) occurs 39 times slower than chloride loss
from the corresponding chloro anion (2a).
Experimental
Materials
We have reported the synthesis and characterization of
N-benzoyloxy-2-nitrobenzenamine previously.[13] A freshly-prepared
sample was recrystallized from ethyl acetate/light petroleum prior to
use, and purity was checked by thin-layer chromatography (TLC) (silica
gel G, CHCl3 elution).
The t-butanol (2-methylpropan-2-ol) was dried on Linde type 5A
molecular sieves and was then distilled from sodium to remove residual
water. Methanol was analytical reagent grade (Univar) containing less
than 0.1 mmol of water per litre. Potassium t-butoxide was prepared by
the reaction of t-butanol with potassium,[14] and sodium methoxide by
the reaction of sodium with methanol under a nitrogen atmosphere.
These metal alkoxide solutions were stored under nitrogen, and their
molarities were established by acid titration.
In extrapolating the observed rate constants for (7) to zero
[KOBut] to obtain the rate constant for the unimolecular
reaction, it is assumed that (7) has been completely
converted to its conjugate base at all the concentrations of
potassium t-butoxide used in the kinetic experiments. The
lack of curvature in the plot of k versus [KOBut] supports the
view that this complete conversion to conjugate base has
occurred.
Leaving Group Abilities
Table 5 lists data obtained with substrate (1b) to enable a
comparison of its rates with those for the N-chloro
compound (1a) obtained in the same solvent mixture (3:1
v/v CCl4/t-BuOH). There is no evidence for an effect on rate
from changing [KOBut] over a small range.
Synthesis of N-Benzoyloxy-4-nitrobenzenamine (7)
A suspension of 4-nitrophenylhydroxylamine (175 mg, 1.14 mmol) in
benzene (20 mL) was degassed with a nitrogen stream and then stirred
at room temperature while redistilled benzoyl chloride (175 mg,
1.25 mmol) in dry pyridine (3 mL) was added during 1 min. The stirring
was continued for 30 min, and the reaction mixture was then washed
with water (100 mL), sulfuric acid (1 M, 10 mL), and water again
(2 × 10 mL). Evaporation of the dried benzene solution yielded a bright
yellow solid, which was crystallized from ethyl acetate/light petroleum
to obtain pale yellow micro-needles (91 mg, 31%). Melting point (m.p.)
139–142°C. TLC (silica gel G, CHCl3 elution) gave a major and a
minor spot. The purity was not improved by recrystallization and the
material was not sufficiently soluble for chromatographic purification.
The infrared (IR) spectrum detected no 4-nitrophenylhydroxylamine:
Table 5. Rate data for the reaction between
N-benzoyloxy-2-nitrobenzenamine (1b) and KOBut in 3:1 (v/v)
CCl4/ButOH at 298.2K
Initial [(1b)] 4.42 × 10–4
M
A
Initial [KOBut]
102k1
M
s
0.081
0.171
0.171
6.91
7.08
6.83
ν
max(CHCl3) 3255 (NH); 1740 (C=O); 1522 and 1347 (NO2) cm–1. On
a direct insertion mass spectrometer probe the substance decomposed
thermally. Mass spectrum (direct CI, NH3) m/z 288 (M+• for a
dinitroazoxybenzene, 5%), 276 (MNH4+ for (7), 38), 259 (MH+ for (7),
99), and 105 (100).
A Average 102k1 = 6.94 0.09 s–1.
The rate constant for (1b) at 298.2K is 6.94 × 10–2 s–1, and
extrapolation of the published data[1] for (1a) yields k1 =
2.73 s–1 at this temperature. Thus chloride ion is 39.3 times
better than benzoate as a leaving group in this situation.
Apparatus
Routine UV–vis spectra were recorded with a Varian DMS-90
instrument, and IR spectra with a Perkin Elmer Paragon 1000 Fourier
transform spectrometer. ESR spectra were measured with a Bruker ER
200D-SRC instrument by Dr S. Brumby.
Conclusions
Kinetic Measurements and Time-Resolved Spectra
N-Benzoyloxy-2-nitrobenzenamine (1b) reacts with sodium
methoxide in two fast, competing reactions. Nucleophilic
addition to acyl carbon yields methyl benzoate and
2-nitrophenylhydroxylamine, while α-elimination leads to
benzofuroxan (2,1,3-benzoxadiazole N-oxide) and benzoic
acid. The bulkier t-butoxide base reacts only by the
α-elimination route, but destroys some of the benzofuroxan
in a secondary reaction. The nitrene centre formed in the
α-elimination is efficiently captured internally by the
adjacent nitro group and does not yield products in
intermolecular reactions.
The electronic spectra of transitory species were recorded
(250–750 nm) with a Hewlett-Packard 8450A UV–vis spectrometer
equipped with a model 7225 plotter and a 89100A temperature
controller ( 0.1°C). The spectra were tape-filed at 5 s intervals.
To conduct
a kinetic run, the solution of N-benzoyloxy-2-
nitrobenzenamine was pipetted into one half of a split sample cell (total
width 1 cm), and a solution of potassium t-butoxide was used to fill the
other half of the cell to the same level. When temperature equilibrium
was reached, the cell was rapidly inverted several times to mix the
contents. A red colour (λmax 516, with shoulder at 542 nm, in t-butanol)
appeared instantaneously on mixing. The fading of this colour was
actually measured at 550 nm to avoid interference from a secondary
product which appeared late in the reaction. The infinity absorbance,
A short-lived red intermediate, shown to be identical in
the reaction of (1b) with either methoxide or t-butoxide, is
A , taken after 10 half-lives, was always close to zero. One reaction
∞

768
L. K. Dyall
product, potassium benzoate, has very low solubility in t-butanol, but in
Products of Reaction of N-Benzoyloxy-2-nitrobenzenamine (1b) with
Potassium t-Butoxide
none of our runs did it precipitate until after A had been recorded.
∞
At the selected wavelength, absorbance readings during the kinetic
runs could be taken at intervals as short as 1 s. Depending on the rate of
the reaction, between 6 and 15 absorbance readings were taken, always
extending across at least two half-lives and sometimes as many as five.
The first-order rate constants were evaluated graphically.
In all the Tables of rate constants reported here, the concentrations
refer to those in the split sample cell after the two reactants had been
mixed.
In order to compare rates for the N-benzoyloxy compound (1b) with
those previously obtained for N-chloro-2-nitrobenzenamine (1a), some
rates were measured with solvent mixtures that ended up three parts
(v/v) CCl4 and one part ButOH after mixing the reactants in the cell.
Occasionally, in this mixed solvent, a fine white suspension (presumed
to be potassium t-butoxide and/or potassium benzoate) appeared before
the run was over and elevated the baseline absorbance. These runs were
discarded.
Stock solutions of (1b) in methanol or t-butanol were used within
2 h of preparation. A check of these solutions by UV–vis spectroscopy
detected no change after 48 h.
For kinetic runs with N-benzoyloxy-4-nitrobenzenamine (7), a
saturated solution of this sparingly soluble material was prepared in
t-butanol. The filtered solution was estimated to be approximately 6 ×
10–4 M. Thereafter, the procedure was as described above for its isomer.
The N-benzoyloxy compound (1b) (216 mg) was dissolved in t-butanol
(20 mL) and treated at room temperature with KOBut (20 mL, 0.1 M).
The immediate red-brown coloration faded to pale brown within
30 min. The mixture was then acidified with HCl (0.2 M), diluted with
water, and extracted with ether. Some material (15 mg) was insoluble.
Extraction of the ether solution with aqueous NaHCO3 recovered
benzoic acid (97 mg, 95%) whose IR spectrum matched that of an
authentic sample. The ether solution yielded a fraction (48 mg) shown
by TLC (silica gel G, CHCl3) to have at least seven components.
Benzofuroxan (2% yield) was identified in the ether distillate by
UV–vis spectroscopy.
ESR Experiments
The compound (1b) (1.0 mg) was dissolved in 1 : 1 v/v
t-butanol/benzene (5 mL) and then mixed at room temperature with
KOBut (0.1 M in 5 mL of the same solvent mixture). Within 2 s of
mixing, some of this solution was transferred into an ESR tube, which
was then capped and transferred into liquid nitrogen. The usual red
colour, obtained on mixing these reagents, persisted in the frozen
sample.
At 160 K, and maximum gain, there was a minor absorption over
and above the quartz tube signal. This signal faded out on warming to
240 K. At 280 K the frozen mixture melted, and after 10 min the red
species was non-existent and a pale yellow solution remained. There
was no ESR spectrum at this temperature, but on cooling to 160 K the
frozen sample had a signal about three times as intense as before. This
signal cannot belong to the red species and must arise in a slower
secondary reaction.
Kinetic Measurements on the Reaction Between Sodium Methoxide and
N-Benzoyloxy-2-nitrobenzenamine (1b)
The overlaps of the broad bands at 515 and 552 nm made it impossible
to measure accurate rate constants for either the fade of the red colour
(
λ
max 515 nm) or the appearance of the blue species (λmax 552 nm). Based
on the first 40% of reaction, the 515 nm band decayed with k1 = 4.3 ×
10–2 s–1 at 25°C. (Initial concentrations, after mixing the two reagents,
were: [(1b)] 3.95 × 10–5 M, [NaOMe] 0.050 M). Under these same
Acknowledgments
I am indebted to Professor Alan Sargeson, of the Research
School of Chemistry, the Australian National University, for
access to his time-resolved UV–vis spectrometer, and to Dr
Stephen Brumby of the same school for measuring the ESR
spectra. Helpful discussions with Drs Michael Novak and
Stephen Glover are gratefully acknowledged. Professor
Geoff Lawrance kindly provided laboratory facilities at the
University of Newcastle.
reaction conditions, the development of the product with max 552 nm
λ
(monitored on its absorption flank at 650 nm) had k1 = 8.7 × 10–2 s–1.
Products of Reaction of N-Benzoyloxy-2-nitrobenzenamine (1b) with
Sodium Methoxide
Methanolic sodium methoxide (0.4 M, 10 mL, 4 mmol) was added to
(1b) (129 mg, 0.5 mmol) at 20°C. There was an instantaneous flash of
red colour, which changed immediately to a dark inky blue. After
30 min, the mixture was diluted with water (50 mL) and then extracted
with ether (2 × 25 mL). The extract was washed with water (10 mL) and
then dried (MgSO4). Careful removal of the solvent (by rotary
evaporation followed by a gentle stream of nitrogen) left a pale yellow
oil (79 mg). The IR spectrum showed this oil to be a mixture of methyl
benzoate, benzofuroxan, and a trace of 1,2-dinitrobenzene.
References
[1] K. J. Chapman, L. K. Dyall, L. K. Frith, Aust. J. Chem. 1984, 37,
341.
[2] G. Boche, F. Bosold, S. Schröder, Angew. Chem., Int. Ed. Engl.
1988, 27, 973.
Five days later, the alkaline-aqueous residue from the ether
extraction had faded in colour from blue to pale orange, and colourless
needles were present. These crystals were collected by filtration and
then dried. Yield, 29 mg. The IR spectrum (CHCl3 solution) was an
exact match to that of 1,2-dinitrobenzene. These needles had m.p. of
119.5–120.5°C and mixed m.p. of 118.5–120°C with an authentic
sample (m.p. 119.5–121°C).
The filtrate from collection of the needles was acidified with dilute
sulfuric acid and then extracted with chloroform to recover a pale
yellow crystalline residue. The IR spectrum (CHCl3 solution) was
identical with that of benzoic acid. Calculations of yield based on key
bands at 1694, 1451, and 1026 cm–1 gave 30.2 1.2 mg (50%).
IR analysis of the neutral oil fraction showed it to consist of 40 0.4
mg (59%) of methyl benzoate (analysed at 1719, 1452, and 1437 cm–1)
and 36 mg (53%) of benzofuroxan (analysed at 1542 cm–1).
1,2-Dinitrobenzene (3 mg) was left behind when the mixture was taken
up in light petroleum.
[3] J. S. Helmick, M. Novak, J. Org. Chem. 1991, 56, 2925.
[4] M. Novak, K. A. Martin, J. L. Heinrich, K. M. Peet, L. K. Mohler,
J. Org. Chem. 1990, 55, 3023.
[5] R. Kuhn, F. Weygand, Chem. Ber. 1936, 69B, 1969.
[6] F. Bosold, G. Boche, W. Kleemiß, Tetrahedron Lett. 1988, 29,
1781.
[7] M. Novak, K. A. Martin, J. L. Heinrich, J. Org. Chem. 1989, 54,
5430.
[8] G. Boche, C. Meier, W. Kleeniß, Tetrahedron. Lett. 1988, 29,
1777.
[9] G. A. Russell, E. G. Janzen, E. T. Strom, J. Am. Chem. Soc. 1964,
86, 1807.
[10] R. D. Guthrie, D. E. Nutter, J. Am. Chem. Soc. 1982, 104, 7478.
[11] M. Novak, R. K. Lagerman, J. Org. Chem. 1988, 53, 4762.
[12] M. Novak, M. J. Kahley, E. Eiger, J. S. Helmick, H. E. Peters,
J. Am. Chem. Soc. 1993, 115, 9453.
[13] I. R. Bryant, L. K. Dyall, Aust. J. Chem. 1989, 42, 2275.
[14] W. S. Johnson, W. P. Schneider, Org. Synth. Coll. Vol. IV 1963,
132.
The various IR spectra gave no indication of 2-nitrobenzenamine,
2,2´-dinitroazobenzene, or the corresponding azoxy compound.