Investigations into the mechanism of action of nitrobenzene as a mild dehydrogenating agent under acid-catalysed conditions

Article abstract of DOI:10.1039/b210887a

Protonated nitrobenzene can be used to dehydrogenate a range of hydrocarbons, which already possess at least one double bond. Kinetic and spectroscopic results, together with known electrode potentials, yield approximate limits within which protonated nitrobenzenes can be expected to effect dehydrogenation of hydroaromatic compounds. A high yielding synthesis of benzo[j]fluoranthene is described.

Full text of DOI:10.1039/b210887a

View Article Online / Journal Homepage / Table of Contents for this issue
Investigations into the mechanism of action of nitrobenzene as a
mild dehydrogenating agent under acid-catalysed conditions
M. Lurdes S. Cristiano,a David J. P. Gago,a Antonio M. d’A. Rocha Gonsalves,b
Robert A. W. Johnstone,*c Moya McCarronc and Jorge M. T. B. Varejãoc,d
a FCT, Universidade do Algarve, Campus de Gambelas, Faro 8000, Portugal
b Departamento de Quimica, Universidade de Coimbra, Coimbra 3000, Portugal
c Department of Chemistry, University of Liverpool, Liverpool, UK L69 3BX
d Escola Superior Agraria, IPC, Bencanta, Coimbra 3040, Portugal
Received 5th November 2002, Accepted 12th December 2002
First published as an Advance Article on the web 20th January 2003
Protonated nitrobenzene can be used to dehydrogenate a range of hydrocarbons, which already possess at least one
double bond. Kinetic and spectroscopic results, together with known electrode potentials, yield approximate limits
within which protonated nitrobenzenes can be expected to effect dehydrogenation of hydroaromatic compounds. A
high yielding synthesis of benzo[j]fluoranthene is described.
amine, aminophenol or amine depends very much on pH and
Introduction
temperature, as well as electrode potential. The range of
products observed following electron transfer reduction of
nitrobenzene is very characteristic of the pH used.
Nitrobenzene and occasionally other nitroarenes have been
used for a long time as mild oxidants, which effect dehydrogen-
ation under both acidic and basic conditions.1a–d Although
various reduction products of nitrobenzene have been des-
cribed as side-products of these dehydrogenations, and there is
a general impression that reduction of nitrobenzene with con-
comitant oxidation of a hydroaromatic to a fully aromatic
compound constitutes the “driving force”, there is little inform-
ation on the thermochemistry or mechanism of the reaction
and there are no kinetic results. There is no guidance concern-
ing when a particular dehydrogenation might or might not be
effected by nitrobenzene and, although there are examples of
hydroaromatic systems that did respond, there are reports
of some that did not.2 Nitrobenzene used under acidic condi-
tions is generally a moderately high temperature process for
dehydrogenating hydroaromatic compounds. Dehydrogenation
of hydroaromatics under neutral conditions requires very high
temperatures and catalysts.3 Under neutral or alkaline condi-
tions, nitrobenzene can be used at ambient temperatures
for dehydrogenation of thiols,4 alkoxides,5 carbanions,6 and
radicals;7 it also plays a prominent role in SET reactions.8
Under acidic or aqueous conditions or with protic solvents, it
is unlikely that nitrobenzene would be reduced concomitantly
by hydride transfer from a hydroaromatic compound because
no evolution of hydrogen has ever been observed. Also, the
range of reduction products from nitrobenzene are not those
expected for hydride reduction.9 Hydride transfer mechanisms
have been suggested for some low yielding conversions of
chloroxylenes into triarylmethane derivatives but these reac-
tions can be accounted for also by Friedel–Crafts/electron
transfer mechanisms.10 The role of nitrobenzene in Friedel–
Crafts reactions has been cogently discussed with the conclu-
sion that electron transfer mechanisms were most likely in many
cases.11 Hydride type transfer on the surface of a heterogeneous
catalyst has been considered as a possibility for reduction of
nitro compounds by hydrogen-donors.12 In contrast to the
hydride type mechanism, reduction of nitrobenzene is well
known to take place easily by electron transfer.13 Electro-
chemical and EPR investigations into the reduction of nitro-
benzene clearly reveal initial formation of the corresponding
anion-radical as the rate determining step,4,14,15 This anion-
radical is then protonated in a second step, which becomes rate
determining at high pH.15 Whether a nitro compound is
reduced to nitroso, azo or azoxy compounds or to hydroxyl-
In the present work, results of dehydrogenation of hydro-
aromatic compounds with nitroarenes are described, together
with other experimental data. There is strong evidence for
initial complex formation between protonated nitroarene and
the hydroaromatic compound before an electron transfer step
and the release of protons that finally leads to dehydrogenation.
Analysis of thermochemical data is entirely compatible with
such a mechanism.
Results and discussion
(a) Kinetic experiments
In the presence of nitrobenzenes, dehydrogenation of 9,10-di-
hydroanthracene (DHA) under acidic conditions has been
found to proceed in very high yield and selectivity to form
anthracene;16 if acid is omitted, no significant dehydrogenation
is observed even at high temperatures or for extended reaction
times. Because of the excellent yield and specificity in producing
anthracene in this reaction, dehydrogenation of DHA with
nitrobenzene was selected as the basis of kinetic experiments in
the present work. Trifluoromethanesulfonic acid (TFMSA) was
chosen as the acid catalyst because it is strong, non-nucleo-
philic, non-oxidising and dissolves readily in nitrobenzene.
Initially, it was found that, in neat nitrobenzene as solvent,
DHA was oxidised to anthracene at 100 ЊC in times varying
between 2–3 minutes and an hour, depending on the hydrogen
ion concentration. Accordingly, in a first series of kinetic
experiments, DHA was dehydrogenated by nitrobenzene to
form anthracene, in the presence of various known proportions
of TFMSA. The resulting graphs for fractional conversion of
DHA into anthracene with time (Fig. 1) show a “burst” phase in
that some anthracene appears within a few minutes but this is
followed by a steady state phase when anthracene continues to
be produced at an almost constant rate. If the molar concen-
tration of acid is increased to be similar to or greater than that
of DHA, the burst phase of anthracene formation affords an
almost 100% yield within 1–2 minutes (Fig. 1) and no steady
state phase was discernible.
In general terms, the rate of dehydrogenation may be con-
sidered to result from an equilibrium constant (K) and the con-
centrations of reactants. Thus, for protonation of nitrobenzene,
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 5 6 5 – 5 7 4
565

View Article Online
Table 1 Kinetic data for conversion of 9,10-dihydroanthracene into anthracene in protonated nitrobenzene
[Hϩ]/mol dmϪ3 T /K
T med/K
kЈ × 105 /sϪ1 1/T /KϪ1 Corr.a Ea/kJ molϪ1 Corr.b ∆H*/kJ molϪ1 ∆S*/J molϪ1 KϪ1 ∆G*/kJ molϪ1
0.0565
0.1130
0.1695
0.2260
0.2825
353.15
373.15 373.15
393.15
353.15
373.15 373.155
393.15
353.15
373.15 373.15
393.15
353.15
373.15 373.15
393.15
0.5
1.8
11.9
1.5
5.9
30.5
3.9
14.1
44.8
6.5
19.5
59.5
11.7
26.8
84.2
0.00283
0.00268
0.00254
0.00283
0.00268
0.00254
0.00283
0.00268
0.00254
0.00283
0.00268
0.00254
0.00283
0.00268
0.00254
0.954
0.952 93.4
0.987
0.973
0.981 87.7
0.964
0.995
0.971 70.3
0.985
0.992
0.970 63.8
0.991
0.991 90.3
0.997 84.6
1.000 67.2
0.999 60.7
0.992 53.6
Ϫ96.0
Ϫ101.1
Ϫ140.5
Ϫ155.2
Ϫ171.5
126.1
122.3
119.6
118.6
117.6
353.15
373.15 373.15
393.15
0.995
0.968 56.7
0.989
a Correlation coefficients for conversion of DHA to anthracene with time; b Correlation coefficients for log(kЈ) and 1/T .
Eqn. (3) represents a first order process, in which kЈ =
kؒK [Hϩ]n[PhNO2] and [DHA]0 is the concentration of DHA at
time zero. For any one hydrogen ion concentration, a plot of
ln([DHA]/[DHA]0) versus t should be linear with a slope equal
to kЈ. If the hydrogen ion concentration is varied then kЈ should
vary in proportion to [Hϩ]n.
For very short times in the burst phase, [DHA]/[DHA]0 ∼
1 Ϫ kЈؒt or, [A]/[DHA]0 = kЈؒt = kؒKؒ[Hϩ]n[PhNO2]t, in which
[A] is the concentration of anthracene produced in the burst
(= [DHA]0 Ϫ [DHA]). For the initial burst phase, a plot of
ln([A]/[DHA]0) versus ln[Hϩ] should then be a straight line of
slope n. When this was done from the graphs in Fig. 1, the value
of n was found to be close to 2. No attempt was made to obtain
a more accurate value because accurate ones for n were later
extracted from the full rate measurements described below.
However, the value does serve to indicate that TFMSA appears
to doubly protonate nitrobenzene. The result is in keeping
with UV/visible measurements on complex formation between
protonated nitrobenzene and DHA (discussed below).
In the main second series of kinetic experiments, the initial
rates of the burst phase for oxidation of DHA were found to
obey a simple first order plot, as expected from the above analy-
sis. In all the following experiments, the initial concentration of
DHA was the same. For each of three temperatures (80, 100,
120 ЊC), the rate of formation of anthracene was monitored
with time and, by application of eqn. (3), a value for kЈ was
obtained at each hydrogen ion concentration (Table 1). For
each value of [Hϩ], a plot of ln(kЈ) versus 1/T yielded the
Arrhenius activation energy, and the enthalpy, entropy and
the free energy of activation (Table 1).
Fig. 1 Conversion of 9,10-dihydroanthracene (1.1 mmol) to anthra-
cene at 100 ЊC in nitrobenzene (2 mL) containing the amounts of
trifluoromethanesulfonic acid shown in the figure. For 0.84 mol LϪ1 of
acid, the reaction is complete as a burst within about 4 minutes. For
0.057 mol LϪ1 of acid the burst is almost non-existent and a gradual
steady formation of anthracene occurs over a period of hours. Other
acid concentrations produce mixtures of burst and steady state phases
ranging between these extremes.
eqn. (1) affords an expression for K, where PhNO2/Hnϩ is a
proton–nitrobenzene complex with n undefined.
The enthalpy of activation falls by nearly 40 kJ.molϪ1 as the
acidity of the medium increases five-fold (Table 1). At the same
time, the entropy of activation becomes more negative by
almost 80 J KϪ1 molϪ1. The combined enthalpies and entropies
offset each other such that the overall free energy change varies
by only 9 kJ molϪ1 over a 40Њ range in temperature and the five-
fold increase in acidity. The changes imply an activation energy
barrier that becomes thermally easier to cross with increase in
acidity but a transition state that becomes more difficult to
form.
PhNO2 ϩ nHϩ ↔ PhNO2/Hnϩ
;
K = [PhNO2/Hnϩ]/([PhNO2][Hϩ]n) (1)
A rate of dehydrogenation can be written as in eqn. (2).
Rate of disappearance of DHA =
Ϫd[DHA]/dt = k[PhNO2/Hnϩ][DHA] (2)
To examine the effect of acidity on kЈ, ln(kЈ) was plotted
against ln[Hϩ] for each temperature. These graph plots were
straight lines with slopes (n) equal to 2.0 (80 ЊC), 1.72 (100 ЊC)
and 1.18 (120 ЊC). New rate constants (kЉ) were calculated from
the expression, kЉ = kЈ/[Hϩ]n (Table 2). At any one temperature,
the rate constants (kЉ) are almost equal within experimental
error. Since [1]n = 1, the new kЉ values are essentially just kЈ for a
uniform hydrogen ion concentration of 1 mol.LϪ1. The kЉ
values were used in a second series of Arrhenius plots of ln(kЉ)
against the reciprocal of absolute temperature to give new
By insertion from eqn. (1) and rearranging, eqn. (3) is obtained,
in which [PhNO2] is considered to be constant because it is used
in large excess as solvent.
Ϫd[DHA]/[DHA] =kЈdt
or, on integration,
ln[DHA]/[DHA]0 = ϪkЈ.t
(3)
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 5 6 5 – 5 7 4
566

View Article Online
Table 2 Kinetic data for conversion of 9,10-dihydroanthracene into anthracene changed to a uniform acid concentration, [Hϩ] = 1 mol LϪ1
kЉ × 103/
EA/
∆H*/
∆S*/
∆G*/
[Hϩ]/mol dmϪ3
0.057
T /K
T med/K
sϪ1 b
ln(kЉ)
1/T /KϪ1
Corr.a
kJ molϪ1
kJ molϪ1
kJ molϪ1 KϪ1 kJ molϪ1
353.15
373.15
393.15
353.15
373.15
393.15
353.15
373.15
393.15
353.15
373.15
393.15
353.15
373.15
393.15
1.47
2.48
3.52
1.15
2.51
3.63
1.38
2.99
3.63
1.29
2.52
3.43
1.48
2.36
3.74
Ϫ6.5208
Ϫ6.0011
Ϫ5.6497
Ϫ6.7675
Ϫ5.9864
Ϫ5.6188
Ϫ6.5863
Ϫ5.8114
Ϫ5.6188
Ϫ6.6562
Ϫ5.9824
Ϫ5.6741
Ϫ6.5175
Ϫ6.0485
Ϫ5.5899
0.00283
0.00268
0.00254
0.00283
0.00268
0.00254
0.00283
0.00268
0.00254
0.00283
0.00268
0.00254
0.00283
0.00268
0.00254
373.15
373.15
373.15
373.15
373.15
0.997
0.985
0.954
0.984
1.000
25.2
33.3
28.2
28.5
26.8
22.1
30.2
25.1
25.4
Ϫ237.5
Ϫ215.6
Ϫ227.9
Ϫ228.5
110.7
110.7
110.1
110.7
0.113
0.170
0.226
0.283
23.6
25.2
Ϫ233.7
Ϫ228.6
110.9
110.6
Mean values = 28.4
a Correlation coefficient for ln(kЈ) and 1/T . b kЉ = kЈ/[Hϩ]n, where n = 2.0 at 353.15, 1.72 at 373.15 and 1.18 at 393.15 K.
values for Ea, ∆H*, ∆S* and ∆G* for reaction at [Hϩ] = 1 mol
Ϫ1. Once the acidity term is made uniform, the enthalpy and
entropy terms also become uniform (Table 2).
reduction of nitrobenzene to aniline and not to the “aromatiza-
tion” process. By postulates relating exo- and endothermicities
of reactions to their activation energies,21 it is clear that exo-
thermic dehydrogenation implies a structure of the transition
state resembling more the starting materials than it does the
products of reaction, viz., the transition state for this dehydro-
genation resembles the initial complex, DHA–PhNO2–Hϩ.
Once reaction is complete, return to the starting materials is
unlikely because of the high energy demand required to offset
the heat of formation of water.
L
The new values of Ea, ∆H*, ∆S* and ∆G* for [Hϩ] = 1 allow
some deductions to be made. First, the free energy of activation
is constant. Second, the activation energy for the burst phase is
quite low and is commensurate with the sorts of values found
for electron transfer in simple redox reactions.17 Third, the
entropic factor is quite large and negative. In terms of random-
ness in the transition state, this result means that it becomes
increasingly difficult for the reaction to proceed as temperature
rises. The power of the hydrogen ion concentration decreases
with increasing temperature as would be expected for a complex
of nitrobenzene and Hϩ that increasingly dissociates as the
temperature rises. The enthalpy and entropy are in accord with
the concept of an initial complex formed from DHA, nitro-
benzene and protons, which decomposes easily by a mechanism
that has very low enthalpy of activation, similar to the require-
ments of simple redox reactions. The complex becomes increas-
ingly difficult to form as temperature increases because of
increasing translational and rotational motion of the com-
ponents. The rapid burst phase of release of anthracene early in
the reaction is characterised by an activation energy that is
small at high acidities, but which increases with either or both
decreasing acidity and increasing temperature.
(c) The complex of DHA–PhNO2–H؉
When DHA is added to a solution of TFMSA in PhNO2, a
yellow colour develops at room temperature. If the acidity is
increased, the colour deepens and becomes orange–brown at
the highest acid concentrations used in these experiments. On
examination of the UV/visible spectra of DHA–PhNO2–Hϩ
mixtures, it is found that, as acid strength is increased, a new
charge transfer band22 begins to appear at about 470 nm. In
a standard process for examining complexes in solution,23a
log(absorption at 520 nm) was plotted against log[Hϩ]. A
straight line was obtained of slope equal to 1.81. The result
indicates that, over the hydrogen ion concentration considered,
a complex is formed between protonated nitrobenzene and
DHA, in which the proton concentration is effective to a power
value close to 2, which accords with the results of the kinetic
experiments described above. Other methods have been used to
examine room temperature complexes of nitrobenzene with
acids and have produced relationships in which the nitro-
benzene/proton number varies from 2 : 1 to 1 : 2,24 and a value
of 1 : 1 : 1 has been reported for a complex of a Lewis acid
(AlCl3) with PhNO2 and anthracene.25
These kinetic results from DHA are next compared with
information gleaned from other experiments described below or
from thermochemical constants and other data found in the
literature.
(b) Heat of reaction
Having obtained the activation energy, it becomes of interest to
obtain the heat of reaction (∆HR) for the conversion of DHA
and nitrobenzene into anthracene and aniline, the end products.
It seems to be generally believed that the “driving force” for
these sorts of dehydrogenations results from the energy gained
by converting nitrobenzene to aniline and by the “aromatiz-
ation” process of DHA being converted into anthracene (A;
reaction 4).1a,b,16
(d) Electron transfer from DHA to protonated nitrobenzene
The basic electron transfer reaction (eqn. (5)) can be viewed as
two half reactions, for which electrochemical reduction and
oxidation potentials are known.
(5)
3DHA ϩ PhNO2
3A ϩ PhNH2 ϩ 2H2O
(4)
The heats of formation of all these components are known.18
From these values and eqn. (4), it is easily calculated that, if
water is neglected, the reaction would be expected to be endo-
thermic but inclusion of water makes the reaction exothermic
by 87 kJ.molϪ1(of DHA). If a thermal “driving force” is
considered, it is clearly due to the formation of water in the
The values for E1/2(1) and E1/2(2) are 1.87 and Ϫ0.56 V
respectively,26 indicating an overall energy requirement of
Ϫ2.43 V (233 kJ molϪ1) at pH 7.0. This requirement is endo-
thermic and far greater than the observed kinetic activation
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 5 6 5 – 5 7 4
567

View Article Online
energy of, for example 70.3 kJ molϪ1 at [Hϩ] = 0.17 mol LϪ1
(Table 1). There are other factors that must be taken into
account in estimating ∆HR of eqn. (5). The first of these is the
change in E1/2(2) as PhNO2 is acidified. Two separate reports
yield information on the magnitude of E1/2(2) with change in
pH.27 If the two sets of data are plotted on the same graph, a
linear correlation is observed, from which the reduction poten-
tial for nitrobenzene can be represented by the equation, E1/2(2)
= Ϫ0.14–0.06[pH] volts (correlation coefficient = 0.984).
The slope is close to the expected value of 0.059 expected for
single electron transfer.28 For example, at a hydrogen ion
concentration of 0.85 mol LϪ1, as used in some of the experi-
ments described here, the pH of 0.07 gives a value for E1/2(2) of
Ϫ0.14 V. This new value decreases the overall energy require-
ment from 2.43 V to 2.01 V (193 kJ molϪ1). Thus, the effect of
acidifying a mixture of DHA and nitrobenzene to pH 0.07 is to
reduce the energy requirement for electron transfer by 40 kJ
molϪ1 (Fig. 2). However, at 193 kJ molϪ1, the process still
appears to be very endothermic and substantially greater than
the thermal component of the activation energy (about 60–
100 kJ molϪ1 depending on acidity; Table 1).
methods. The resulting crystal structure showed that each ben-
zene ring of each DHA molecule independently formed a com-
plex with an acceptor molecule (1,3,5-trinitrobenzene) thereby
giving a 1 : 2 complex for the DHA. The interaction between
each DHA and the two acceptor molecules was sufficiently
strong as to cause a shape change in the DHA.38 If there are
effectively two separate complexation sites in the same mole-
cule, the total resonance energy is approximately doubled.
Good acceptors such as tetracyanoethylene having a similar
electron affinity to that of protonated nitrobenzene form 1 : 1
complexes with even simple aromatic π-donors such as hexa-
methylbenzene and these complexes have stabilities of about
30 kJ mol .
Ϫ1 39 These figures indicate that a resonance energy of
78 kJ molϪ1 is reasonable for a strong double complex of DHA
with protonated nitrobenzene.
Together, the resonance and Coulombic energy terms reduce
the overall requirement for transfer of an electron from DHA
to protonated PhNO2 from 193 to 86 kJ molϪ1 (Fig. 2). Since
the reaction is carried out at 100 ЊC, there is an additional heat
component of about 15 kJ molϪ1 40
bringing the total energy
,
requirement for electron transfer to 63 kJ molϪ1. This estimated
value for the heat energy needed to allow electron transfer from
DHA to protonated nitrobenzene is similar to the enthalpies
of activation found in the kinetic experiments (Fig. 2). This
close similarity between the enthalpy for an electron-transfer
process and the observed enthalpy of activation increases the
probability that the mechanism starts with formation of
a strong complex between DHA and protonated benzene (a
charge-transfer complex). Heating to 100 ЊC effects complete
transfer of the electron. As shown in electrochemical work
on PhNO2, once initial electron–proton transfer has occurred
at the reduction potential, a further cascade of electrons and
protons completes the reduction to aniline at the same
potential.15
(f ) The importance of the ionization energy of the substance to
be dehydrogenated
If the “driving force” for the reaction is due to the ease of
transfer of an electron from the substance to be dehydrogenated
(DHA in this case) and the electron affinity of protonated
nitrobenzene then it is clear that the ionization energy (I ) of the
hydrogen donor must play an important role. For all dehydro-
genations involving protonated nitrobenzene, the electron affin-
ity of the protonated nitrobenzene (EA) remains constant for
any one hydrogen ion concentration.32 The Coulombic energy is
small and may be neglected in this discussion. Therefore, the
energy requirement for electron transfer can be approximated
to the difference in the energy (I ) needed to ionize DHA and
the energy (EA) gained, viz., I Ϫ EA. It is now possible to
consider what effect a change in ionization energy would have
on the rate of dehydrogenation. The ionization energy (I ) of
the substrate to be dehydrogenated is important in determining
whether or not an initial charge transfer complex will be formed
with sufficient gain in energy as to offset the notional energy
requirement for complete transfer of an electron in the absence
of complex formation.
For the hydrobenzene compounds, cyclohexane, cyclohexene
and cyclohexadiene, the ionization energies are 9.86, 8.95,
and 8.40 eV respectively.41 In the present work, attempts were
made to dehydrogenate these compounds under similar
condition to those used for the kinetic experiments with DHA.
There were three quite different results. At 50 ЊC, cyclohexane
was not dehydrogenated, even on extended heating. Cyclo-
hexene underwent slow conversion to benzene at 50 ЊC (40% in
one hour). In marked contrast, cyclohexa-1,3-diene itself and
three other cyclohexadienes (α-terpinene, 4(R)- and 4(S)-iso-
propenylmethylcyclohex-1-ene) underwent rapid dehydrogen-
ation to benzene or substituted benzenes at 22 ЊC within about
1 minute. These last materials are all excellent hydrogen donors
Fig. 2 An outline of the various energies discussed in the text. The
thermal gap represents the difference in energy between the starting
materials (0 kJ molϪ1) and the same materials after mixing and allowing
for acidity effects, complex formation, Coulombic energy and added
heat energy in raising the temperature of the mixture to 100 ЊC. The
thermal gap is of a similar magnitude to the observed enthalpy of
activation, shown here bracketed over a range of acidities.
(e) Energetics of complex formation between protonated
PhNO2 and DHA
The appearance of a new charge-transfer band at 470 nm on
acidifying a solution of DHA in nitrobenzene with TFMSA
indicates the formation of a stable entity, in which an electron is
already partially transferred from DHA to PhNO2. This sort of
charge transfer complex formation has been well researched in
the past.23b,29 The stability of the complex can be estimated
from three major factors: the ionization energy of DHA, the
electron affinity of protonated PhNO2 and a Coulombic term
resulting from the charges arising when an electron is trans-
ferred.30 From these values, a resonance energy can be calcu-
lated for the complex. As shown,30 the resonance energy is
calculated to be 0.81 eV (78 kJ molϪ1) and the Coulombic
energy 0.3 eV (29 kJ molϪ1).33 The calculated resonance energy
for the DHA complex may at first appear to be high but it
should be noted that a 1 : 2 complex of DHA with 1,3,5-trini-
trobenzene has been prepared and is sufficiently stable for its
structure to have been determined by X-ray diffraction
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 5 6 5 – 5 7 4
568

View Article Online
in catalysed heterogeneous transfer hydrogenation.12 In chang-
ing from an ionization energy of 8.4 eV to 8.95 eV and then to
9.86 eV, dehydrogenation changes from being very easy (cyclo-
hexadiene) to being unobservable (cyclohexane) under the
reaction conditions used here. Cyclohexene is intermediate in
behaviour, undergoing slow dehydrogenation. The behaviour is
in keeping with the strengths of complexes formed between
such compounds and electron acceptors.22 In practical terms,
an ionization energy of about 8.9 to 9.0 eV (or I Ϫ E = 6.3–6.4
eV) can be regarded as an upper limit for deciding whether or
not dehydrogenation is likely to occur with protonated nitro-
benzene at modest temperatures.
In other experiments described below, 1,2,3,4-tetrahydro-
naphthalene (tetralin; an “analogue” of cyclohexane) gave very
small amounts of naphthalene and resisted attempts to
dehydrogenate it even at 150 ЊC and high acidity. However, GC/
MS revealed some formation of two isomers with Mr = 262
resulting from acid-catalysed condensation of two molecules
of tetralin: these were not investigated further as they were
not products of dehydrogenation. In contrast, 1,2-dihydro-
naphthalene (dialin; an analogue of cyclohexene) was readily
dehydrogenated at 150 ЊC to give some naphthalene but there
were added complications from acid-induced reactions; the
major product was benzo[j]fluoranthene.
Dehydrogenation of 9,10-dihydrophenanthrene was slow
even at 100 ЊC. Toluene and 1,2-dimethylbenzene did not
dehydrogenate at 100 ЊC. Perhydroaromatic compounds seem
to resist dehydrogenation by nitrobenzene, as would be
expected from a mechanism involving electron transfer in an
initially formed complex of substrate and nitrobenzene because
the ionization energy for an alkane is much greater than that
for an alkene. Examples, from the literature support this
proposal: 1,2,3,4,5,7,8,8a,9,10,10a-dodecahydrophenanthrene
dicarboxylic acid anhydride could not be dehydrogenated by
nitrobenzene2 and dibenzyl with nitrobenzene could only be
converted into stilbene in modest yield even at 245 ЊC, under
pressure and over a period of 16 hours in the presence of a
palladium catalyst.3a
(h) The effect of a change in reduction potential through the use
of other nitroarenes
The above discussion has shown that the effect of protonation
of nitrobenzene is to lead to an increase in its electron affinity
(a decrease in its reduction potential), which in turn leads to a
concomitant ease of use of nitrobenzene as an oxidant. How-
ever, the electron affinity of nitroarenes can also be changed
considerably by varying the substituents in the aromatic ring.
Thus, for a series of 4-substituted nitrobenzenes, electron
affinity varies from 2.0 eV for 1,4-dinitrobenzene to 0.91 eV for
4-methoxynitrobenzene. From the equation, EA = 1.49(E1/2) ϩ
2.79,32 this difference in EAs translates into a change in E1/2 of
0.73 V, viz., it becomes much more difficult to reduce the nitro
group in 4-methoxynitrobenzene than it does in 1,4-dinitro
benzene. These changes should make 1,4-dinitrobenzene a
better oxidant (I Ϫ EA is smaller) than nitrobenzene and
4-methoxynitrobenzene a worse oxidant (I Ϫ EA is larger).
Electron affinities correlate linearly with reduction potentials
(E1/2).32 Also, reduction potentials of substituted nitrobenzenes
in DMSO have been found to correlate with σpϪ values.42 When
electron affinities for p-substituted nitrobenzenes were plotted
against σpϪ constants,43 a linear correlation was found with EA
= 0.69(σpϪ) ϩ 1.09 eV. Therefore, logarithms of rate constants
for dehydrogenation of substrates using p-substituted nitro-
Ϫ
benzenes should vary linearly with σp constants for the sub-
stituents (I Ϫ EA = I Ϫ 0.69 σpϪ ϩ 1.09). This effect of change
of substituent was examined by dehydrogenation of a por-
phyrinogen with different nitrobenzenes. Nitrobenzene in acetic
acid is a good oxidant for converting various porphyrinogens
(made from substituted benzaldehydes and pyrrole) into the
porphyrin itself.44 The reaction has been optimised44 but it is
sometimes accompanied by significant quantities of the corre-
sponding chlorins, in which two of the hydrogens in the por-
phyrin have not been removed. Nitrobenzene in acetic acid is
not a sufficiently powerful oxidant as to convert at least some
chlorins (3; Scheme 1) into porphyrins.
In the present work, solutions of 5,10,15,20-tetrakis-
(4-methoxyphenyl)porphyrinogen (2; Scheme 1) were prepared
(g) Ionization energy and the dehydrogenation of anions
There are many instances of the ease with which anions and
radicals may be dehydrogenated with neutral or alkaline nitro-
benzene. In these cases, the concentrations of hydrogen ion are
very low and the energy gain of 40 kJ molϪ1 from protonation
of nitrobenzene cannot apply. However, for anions, their ionis-
ation energies are simply the negative values of the electron
affinities of the corresponding radicals. Electron affinities for
radicals are generally of the order of 1–2 eV and therefore, for
an anion, the ionization term (I ) in complex formation30 is only
about 1–2 eV, which is about 6–7 eV below the ionization
energy even of an easily dehydrogenated cyclohexadiene. The
electron affinity of non-protonated nitrobenzene is 1.01 eV.35
Thus, I Ϫ EA for dehydrogenation of anions is only about
0–2 eV compared with the 6.3–6.4 eV needed to dehydrogenate
alkenes. Complex formation between nitrobenzene and an
anion should then be expected to be very easy and provide more
than enough energy to offset the loss of electron affinity
entailed in changing to unprotonated nitrobenzene.38 Oxidation
of anions by nitrobenzene can be expected to be easy. This
ease of dehydrogenation has been found for SET reactions, and
in the easy oxidations of thiolates, ascorbates and carbanions
of various kinds.4–9 Dehydrogenation of all such entities has
been examined by EPR and been shown to proceed through
initial complex formation with nitrobenzene, followed by
electron transfer to form the anion-radical of nitrobenzene.
Similar reactions have been investigated in cellular systems, in
which various drugs containing nitro groups have been shown
by EPR to react by initial transfer of an electron to the
nitro group.6b
Scheme 1
at room temperature by reducing solutions of 5,10,15,20-
tetrakis(4-methoxyphenyl)porphyrin (1; Scheme 1) in acetic acid
with zinc dust. None of the chlorin (3; Scheme 1) was observed
in these experiments. As expected, solutions of the porphyrino-
gen were colourless and showed no UV/visible absorption at the
usual porphyrin Soret band wavelength near 420 nm. The por-
phyrinogen was dehydrogenated by adding a solution of it to a
solution of a p-substituted nitrobenzene in acetic acid at 90 ЊC.
The appearance of the Soret band representing the reformation
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 5 6 5 – 5 7 4
569

View Article Online
of porphyrin was monitored by UV/visible measurements.
Initial rate constants were obtained for the first few minutes of
reaction by recording the visible absorption at the Soret wave-
length with time. When the logarithms of these first order rate
constants are plotted against σpϪ constants43 for several substi-
tuted nitrobenzenes, an approximately linear correlation is
colour and fluorescence.45a In the present work, only the [j] iso-
mer was formed with nitrobenzene as oxidant and at a much
lower temperature than in the earlier work. This suggests that
the acid-catalysed condensation followed by dehydrogenation
did not proceed through the same mechanisms as those in
the earlier work.45a The formation of only the [j] isomer pro-
vides a credible mechanistic route (Scheme 2), whereby acid
catalysed condensation of 1,2-dihydronaphthalene leads first to
a specific dimer. The dimer can ring close in only one way to
form a hydrobenzo[j]fluoranthene, which can be dehydrogen-
ated to benzo[j]fluoranthene itself (Scheme 2). Confirmation of
this pathway was obtained by isolation of a second benzo-
[j]fluoranthene, which was not completely identified but, from
its UV and mass spectra, it is almost certainly a tetrahydro-
naphthyl-substituted benzo[j]fluroranthene (Scheme 2), formed
by dehydrogenation of a trimer of 1,2-dihydronaphthalene.
This tetrahydronaphthyl-susbstituted aromatic compound has
an almost identical long wavelength UV absorption to that of
benzo[j]fluoranthene but its molecular mass is 382, rather than
the 252 of a simple benzofluoranthene. The results suggest that
the likelihood of any Scholl reactions being observed during
nitrobenzene dehydrogenation requires first a sufficiently strong
Lewis acid to induce condensation and then a nitrobenzene to
effect dehydrogenation of this intermediate.
Ϫ
observed. The nitroarene with the largest σp value (1,4-di-
nitrobenzene) was found to increase the rate of dehydrogen-
ation 10-fold compared with the rate for the nitro compounds
Ϫ
with the smallest σp (4-methoxynitrobenzene; Fig. 3). The
(j) Effects of other acids
Other acids were examined for their ability to accelerate
dehydrogenation of DHA by nitrobenzene. All the experiments
were carried out at 100 ЊC for 24 hours. For HCl (conc.), no
conversion was found after 24 hours. However, when the reac-
tion was left for over a week, a maximum 30% conversion was
observed after 6 days. With acetic, trichloroacetic and tri-
fluoroacetic acid, no reaction was observed. For concentrated
nitric, sulfuric and chlorosulfonic acid, reaction was complete.
The last acids are very strong and nitric acid is an oxidant in its
own right. Compared with these acids, acetic, trichloroacetic
and trifluoroacetic acids are quite weak and they seem not to
be able to form an adequate complex with nitrobenzene. The
reason for the failure of hydrochloric acid is not obvious except
that, at 100 ЊC, hydrochloric acid would evaporate from the
reaction medium quite quickly. Also, it is notable that, whereas
concentrated sulfuric acid has been reported to form a yellow
solution with nitrobenzene, no colour developed in liquid
HCl.24b
Fig. 3 Logarithms of the initial rate of formation of 5,10,15,20-
tetrakis(4-methoxyphenyl)porphyrin (2; Scheme 1) from the
corresponding porphyrinogen (1; Scheme 1) in the presence of a
nitrobenzene. Seven p-substituted nitrobenzenes, RC6H4NO2, were used
with σp values as follows: NO2 (1.27), H (0), NH3 (0.56), N(CH3)3
(0.77), CH3O (Ϫ0.26), CO2H (0.77), all from reference 43.
Ϫ
ϩ
ϩ
slope of the graph in Fig. 3 (0.52) is very similar to the slope of
Ϫ
the correlation between EA and σp (0.69). Therefore, the rate
of dehydrogenation of porphyrinogen to porphyrin does vary
with the reduction potential of the substituted nitrobenzene
used as oxidant.
(i) Dehydrogenation of 1,2-dihydronaphthalene by nitrobenzene
(k) Reduction products of nitrobenzene
Cyclohexane, cyclohexene, cyclohexadienes and tetralin have
already been commented on but the dehydrogenation of 1,2-
dihydronaphthalene needs further comment. Under the usual
conditions with protonated nitrobenzene at 100 ЊC, the start-
ing material disappeared quickly but little dehydrogenation to
naphthalene was observed; a similar small quantity of tetralin
was formed, the naphthalene and tetralin presumably being
formed by disproportionation. However, two major higher boil-
ing components were found. By GC/MS, these had molecular
masses of 262 and 258 and appear to be formed from an acid-
catalysed dimerization to mass 260, followed by simple dispro-
portionation to 262 and 258. With continued heating at 150 ЊC,
the yield of the dimers went down and a third major compon-
ent appeared. This was isolated from the reaction mixture as a
crystalline solid and found to be benzo[j]fluoranthene. Further
acid-catalysed ring closure of the initial dimers followed by
dehydrogenation is the most likely mechanism for this reaction
(Scheme 2). Both benzo[j]fluoranthenes and benzo[k]fluor-
anthenes have been prepared as a mixture in a Scholl type
synthesis by oxidising 1,2Ј-binaphthylene over CrO3–Al2O3 at
495–500 ЊC.45a Similar reactions have led also to ill-defined mix-
tures of acid-catalysed condensation products of bisnaphthal-
enes, believed to contain benzofluoranthenes.45b The [j] and [k]
isomers are easily distinguished by melting point, UV spectra,
When nitrobenzene is reduced electrochemically, two distinct
patterns of final reduction products are observed. Under acidic
conditions, nitrobenzene gives aniline and 4-hydroxyaniline (by
acid-catalysed rearrangement of the intermediate reduction
product, phenylhydroxylamine). At low pH, the addition of a
first electron is rate determining, followed by addition of a pro-
ton and then the rapid addition of three more electrons and
protons to reach the phenylhydroxylamine stage, which then
rearranges. Reduction to aniline requires another two electrons
and two protons at a slightly higher reduction potential. With
dehydrogenation in the presence of benzenesulfonic acid, the
benzenesulfonate ester of 4-hydroxyaniline is observed.16 The
dehydrogenations described here always resulted in the form-
ation of aniline and the trifluoromethanesulfonate ester of
4-hydroxyaniline, This sulfonate has been synthesised earlier.10
Under alkaline conditions, proton addition becomes the rate
controlling step and condensation of nitrosobenzene with
aniline or phenylhydroxylamine to give azo and azoxy com-
pounds is common. Thus, the reduction products of nitro-
benzene (aniline and 4-aminophenyl trifluoromethylsulfonate)
in the dehydrogenations described here are of the same pattern
as those observed during electrochemical reduction under
strongly acidic conditions.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 5 6 5 – 5 7 4
570

View Article Online
Scheme 2
(l) Effect of aniline on the reaction rate for dehydrogenation of
DHA
In burst kinetics, after an initial burst phase, the subsequent
steady state is controlled by one of the products of reaction.46
In the dehydrogenation of DHA by protonated nitrobenzene, it
was found in the present work that intentional addition of
anthracene at the beginning of reaction did not affect the burst
or steady state phases. The other major reaction products are
aniline and/or 4-aminophenol, which react with protons to
form ammonium salts. The production of ammonium salts
means that one product of reaction remains “complexed” to the
catalyst (the acid), viz., the requirements for burst kinetics are
met.46 As revealed by Fig. 1, with sufficient acid (more than
equimolar to DHA), the burst in formation of anthracene is
virtually quantitative within 2–3 minutes at 100 ЊC (Fig. 1)
and there was no steady state; in this case, the formation of
ammonium salts is not sufficient to slow down anthracene
formation. For amounts of acid less than equimolar to DHA,
as acidity decreases the burst formation of anthracene becomes
progressively less and longer steady state processes are
observed. The data show that, after the burst phase, the result-
ing ammonium salt must control the subsequent steady state
phase through control of hydrogen ion concentration; the con-
centration of hydrogen ions is controlled by dissociation of the
Scheme 3
anilinium salt. The process is illustrated in Scheme 3, in which
the overall stoichiometric process is shown first and then the
burst and steady state phases are outlined.
Further support for the process shown in Scheme 3 comes
from two observations. First, in a system of nitrobenzene–acid–
DHA in which dehydrogenation is proceeding quite rapidly to
give anthracene plus aniline, addition of a small molar excess of
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 5 6 5 – 5 7 4
571

View Article Online
aniline to a reacting DHA–nitrobenzene–acid solution stops
the reaction immediately because the additional aniline
removes most of the protons. Second, when nitrobenzene
is replaced as oxidant by 4-trifluoromethylnitrobenzene, rate
changes are observed. For this last oxidant, formation of
an ammonium salt from any 4-trifluoromethylaniline produced
on reduction is much less easy than for aniline because the
strongly electron-withdrawing CF3 group makes this aniline a
very weak base. This means that one of the products of reaction
does not remove so many protons and therefore the steady
state rate is improved. Additionally, 4-trifluoromethylnitro-
benzene has a lower reduction potential than does nitro-
benzene. Therefore, during dehydrogenation of DHA, the elec-
tron-withdrawing CF3 substituent eases electron transfer and
therefore speeds the burst phase. Both these effects (an
increased burst and an increased steady state rate) were
observed when DHA was dehydrogenated with 4-trifluoro-
methylnitrobenzene and trifluoromethanesulfonic acid at
100 ЊC. The reduction products of 4-trifluoronitrobenzene were
4-trifluoromethylaniline and 2-amino-5-trifluoromethylphenyl
trifluoromethanesulfonate.
Experimental
All compounds were used as received (Aldrich). GC analyses
were carried out on either a Shimadzu GC14B instrument and
an EC1 column or on a Chrompack CP-9001 with an Optima-5
column (FID). GC/MS analyses were carried out with an EC1
column on a Fisons Trio 1000 (EI at 70 V or CI with ammonia
reagent gas).
Kinetic experiments on the dehydrogenation of 9,10-dihydro-
anthracene.
(a) Initial experiments showing the burst phase (Fig. 1). In a
typical experiment, 9,10-dihydroanthracene (0.2 g; 0.0011 mol)
was dissolved in nitrobenzene (2 mL) and trifluoromethane-
sulfonic acid (x µL) was added. The mixture was placed in an
oil bath at 373 K and small aliquots were removed at approxi-
mately 15 min intervals over a period of about 120 min. These
aliquots were examined by GC and the peak areas for anthra-
cene (A) and dihydroanthracene (DHA) were measured. The
conversion (c = A/(AϩDHA)) into anthracene was plotted
against time to give the graphs shown in Fig. 1. The experi-
ments were repeated for x = 10, 20, 30, 40, 50, 150.
Conclusions
(b) More accurate measurements over both burst and steady
state phases. In a typical experiment, 9,10-dihydroanthracene
(0.2 g; 1.11 mmol) was dissolved in nitrobenzene (2 mL). Tri-
fluoromethanesulfonic acid (x µL) was added and the final mix-
ture was stirred in an oil bath stabilised at 373.15 1 K. The
temperature was noted throughout the kinetic runs to monitor
its constancy. Small aliquots (50 µL) of the reaction mixture
were removed at regular intervals and each sample was added to
dichloromethane (1 mL) ready for GC, which was used to
monitor the formation of anthracene and the disappearance of
9,10-dihydroanthracene. For each sample, peak areas for both
anthracene (A) and 9,10-dihydroanthracene (DHA) were
measured. A dimensionless conversion factor (c) was calculated
from, c = A/(AϩDHA). Rate constants were then determined
as described in the Discussion section. This kinetic experi-
ment was repeated at two more temperatures, 353.15 1 and
393.15 1 K.
For each of the three temperatures, rate constants were
determined at acidities corresponding to the addition of x mL
of trifluoromethanesulfonic acid, where x = 10, 20, 30, 40, and
50.
Each rate constant was measured in duplicate from 6 to 8
data points, giving a coefficient of variation between sets of
2.4%. The rate constants are listed in Table 1.
Rate measurements and other experiments have shown that
9,10-dihydroanthracene is dehydrogenated to form anthracene
by protonated nitrobenzenes in a process of burst kinetics. The
results can be interpreted as initial complex formation between
protonated nitrobenzene and dihydroanthracene, followed by
electron and proton transfers from the DHA to the nitro-
benzene. The action of trifluoromethanesulfonic acid is to
increase the electron affinity (reduce the reduction potential) of
the nitrobenzene through its protonation. The kinetics of
dehydrogenation are very dependent on acid concentration, in
keeping with the increased ease of electrochemical reduction of
nitrobenzene in acidic media. A similar decrease in reduction
potential can be effected through use of p-substituted nitro-
benzenes, in which the substituent is electron-withdrawing.
The results provide a guide to the ease with which
hydroarenes can be expected to be dehydrogenated to arenes
with protonated nitrobenzenes. The major influence lies in the
ionization energy (I ) of the substance to be dehydrogenated
and the electron affinity (EA) of the nitrobenzene used as
oxidant, viz., on the difference I Ϫ EA. Since electron affinity
varies over only a small range (0–3 eV) but ionization energies
for hydrocarbons vary over a much larger range (0–10 eV),
changing the ionization energy of the substrate to be dehydro-
genated has a more substantial impact on the ease of
dehydrogenation. The lower the ionization energy of the
substrate, the easier it is to effect its dehydrogenation and
vice versa. Experiments with p-substituted nitrobenzenes reveal
that electron-withdrawing substituents accelerate the rate of
dehydrogenation of porphyrinogens to porphyrins. The pres-
ence of the strong acid can lead to Friedel–Crafts type reactions
and these in turn furnish compounds other than those expected
of simple dehydrogenation. For example, 1,2-dihydronaphthal-
ene forms little naphthalene on attempted dehydrogenation but
does form a relatively high yield of benzo[j]fluoranthene by
condensation and dehydrogenation.
Kinetic experiments on the dehydrogenation of porphyrinogens.
(a) Preparation of 5,10,15,20-tetrakis(4-methoxyphenyl)-
porphyrinogen.
5,10,15,20-Tetrakis(4-methoxyphenyl)-
porphyrin (0.5 g; 0.68 mmol) was dissolved in a mixture of
glacial acetic acid (50 mL) and trifluoroacetic acid (2.5 mL) and
the resulting solution was degassed by refluxing it gently under
a slow stream of nitrogen. After cooling the solution to 20 ЊC,
zinc dust (2.5 g; 0.38 mol) was added and the solution was
stirred for 1 h.47 To ensure complete reduction, visible spectra
of small aliquots were obtained until no bands could be
observed in the 400–600 nm region. This solution of por-
phyrinogen (S) was filtered and used in the following kinetic
work.
The same criterion of the magnitude of I Ϫ EA provides an
explanation for the ease with which anions and radicals can be
dehydrogenated by nitrobenzenes under neutral conditions. For
anions and radicals in neutral or alkaline solution, their ioniz-
ation energies are very much less than those of hydroarenes and
the electron affinity of nitrobenzene decreases by only about 2
eV. Therefore, I Ϫ EA is very much less than it is in the case of
protonated nitrobenzene and a hydroarene. Dehydrogenation
of anions and radicals by nitrobenzene is expected to proceed
easily, as is observed for many reactions, including SET
processes.
(b) Dehydrogenation back to porphyrin. In a typical experi-
ment, two vessels were used, which could be heated in the same
oil bath and the contents of which could be stirred. Nitrogen
was used to ensure the atmospheres within the vessels contained
no air. Into one vessel was placed an aliquot (10 mL) of the
porphyrinogen solution (S), which served as a “blank” control.
Into the other vessel was placed an aliquot of solution (S) plus
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 5 6 5 – 5 7 4
572

View Article Online
nitrobenzene (1.5 mmol). Both vessels were heated to 363 1 K
and stirring was initiated at zero time. Aliquots (25 µL) of each
solution were removed at intervals and each was dissolved in
CHCl3 (3 mL, containing 13 mL of triethylamine per litre of
CHCl3). Visible spectra were recorded and absorbances due
to the appearance of 5,10,15, 20-tetrakis(4-methoxyphenyl)-
porphyrin were measured at 420 nm (the Soret band). The
background absorbance of the “blank” at 420 nm was sub-
tracted from the measured absorbances from the reaction
vessel. Concentrations of porphyrin were calculated from the
net absorbance and the molar absorption coefficient for the
Soret band of the porphyrin.48 The initial rate of formation of
porphyrin was determined from its yield over the first few
(f ) (4-R)(؉) and (S)(؊)-isopropenyl-1-methylcyclohexene
(limonenes). Complete conversion to 1-methy-4-isopropenyl-
benzene was observed within 12 min.
(g) Cyclohexane. No change was observed at 323 K after
extended periods (24 h).
Formation of reduction products from nitrobenzene
In all experiments using nitrobenzene as dehydrogenating
agent, its reduction product, aniline, was observed and con-
firmed by GC/MS. Additionally, a second reduction product
was found (4-aminophenoxytrifluoromethyl sulfonate; MS, m/z
241 (Mϩ, 8%), 108 (M Ϫ CF3SO2; 100%), 80 (M Ϫ CF3SO2 Ϫ
CO; 45%). The identity of this product of acid catalysed
rearrangement of initially formed phenylhydroxylamine was
confirmed by independent synthesis.10 When 4-trifluoromethyl-
nitrobenzene was used as dehydrogenating agent, the analogous
reduction product (2-amino-5-trifluoromethylphenoxytrifluoro-
methyl sulfonate) was observed; MS, (Mϩ, m/z 309; 10%), 176
(M Ϫ CF3SO2; 100%), 148 (M Ϫ CF3SO2 Ϫ CO; 85%).
minutes of reaction. The experiment was repeated for other
ϩ
substituted nitrobenzenes (4-RC6H4NO2; R = NO2, NH3
,
N(CH3)3ϩ, CH3O, CO2H) in place of nitrobenzene itself (R =
H). Logarithms of the initial rates of formation of porphyrin
for the various nitrobenzenes are shown in Fig. 3.
Dehydrogenation of 9,10-dihydroanthracene by other acids
Each experiment was carried out using a solution of 9,10-di-
hydroanthracene (0.2 g; 1.11 mmol) in nitrobenzene (2 mL), to
which had been added an acid (50 µL). Each solution was
heated at 373.15 1 K for 24 h and the amount of conversion
to anthracene was determined by gas chromatography, as
described above. The acids used were either concentrated or
neat: HCl(conc.), HNO3(conc.), H2SO4(conc.), HSO3Cl(conc.),
CH3CO2H, CF3CO2H, CCl3CO2H. The results are described in
the main text.
Complex formation between protonated nitrobenzene and
9,10-dihydroanthracene
A solution (S1) of 9,10-dihydroanthracene in nitrobenzene
(0.29 mol LϪ1) and a solution (S2) of trifluoromethanesulfonic
acid in nitrobenzene (1.28 mol LϪ1) were prepared. To aliquots
(5.2 mL) of S1 were added aliquots of S2 (200 µL) to a limit of
1000 µL. UV/visible spectra for each mix were recorded and
compared with a blank solution of S1 containing no acid. A
plot of net absorbance at 520 nm against amount of added acid
was linear, showing that the Beer–Lambert laws were obeyed.
As more acid was added, the absorbance of nitrobenzene near
470 nm increased until, after addition of 600 µL of acid solu-
tion S2, a clear maximum of the charge-transfer band could
be discerned. The tail of this band at 520 nm was used for
the absorbance measurements. Finally, the logarithm of the
absorbance at 520 nm was plotted against the logarithm of the
hydrogen ion concentration and gave a straight line of slope
1.81 (correlation coefficient, 0.9917).
Dehydrogenation of other hydrocarbons
(a) 1,2-Dihydronaphthalene (dialin). The hydrocarbon (0.14 g;
1.1 mmol) was dissolved in nitrobenzene (2 mL) and tri-
fluoromethanesulfonic acid (30 µL) was added. The mixture
was kept at 295 1 K and was sampled at intervals for GC and
GC/MS analysis. Naphthalene was formed quickly and starting
material disappeared within a few minutes but the conversion to
naphthalene was very low. The experiment was repeated at
373 K and 423 K. Results are described in the main text. After
the experiment at 423 K, some of the product was separated on
silica gel TLC plates to give mainly one yellow band, along with
two other smaller bands. The bands were scraped from the plate
and extracted with DCM. The isolated products were examined
by GC/MS. For the middle-running main yellow band, the
product crystallised on evaporation of the solvent and was
identified as benzo[j]fluoranthene (mp 164 ЊC, lit. 166 ЊC;45a
UV/visible λmax 332, 346, 364, 374, 382 nm; MS (EI), m/z 252
(Mϩ; 100%), 250(M Ϫ H2; 36%), 126(M Ϫ C10H6; 30%),
125(M Ϫ C10H7; 30%). A bottom paler yellow band was tenta-
tively identified as a naphthylbenzo[j]fluoranthene (UV/visible
λmax 332, 338, 346, 366, 374, 386 nm; MS (CI; NH4ϩ), m/z 383
(M ϩ Hϩ; 100%), 131 (M Ϫ C10H11; 22%). The top weakly
fluorescent band produced the same mass spectrum as that
found for 1,2-dihydronaphthalene “dimers”, observed in the
initial GC/MS analysis. Results are discussed in greater detail in
the main text. In the following, experimental conditions were
similar to those in (a) at 295 K.
Acknowledgements
The authors thank FCT, Portugal (JMTBV, DJPG, MLSC)
and the Eschenmoser Trust, UK (MLSC) for generous financial
assistance.
References
1 (a) P. A. Claret, in Comprehensive Organic Chemistry, eds. D. H. R.
Barton and W. D. Ollis, vol. 4, Pergamon Press, Oxford, 1979,
pp. 157–160; (b) L. Golser, in Methoden der Organischen Chemie
(Houben-Weyl), eds. E. Müller and O. Bayer, Oxidation Part 4/1b;
Part II, Oxidation, Georg Thieme, Stuttgart, 1975, pp. 963; (c) F. G.
Holliman, B. A. Jeffrey and D. J. H. Brock, Tetrahedron, 1963, 19,
1841; Y. Lepage, Bull.Soc.Chim. Fr., 1963, 1141; (d ) H. T. Clarke and
A. W. Davis, in Organic Syntheses, collected vol. 1, Wiley, London,
1948, p. 478; S. Filipov, O. Petrov and N. Mollov, Arch. Pharm.,
1982, 315, 498.
(b) 1,2,3,4-tetrahydronaphthalene (tetralin). This showed no
conversion to dehydrogenated products when heated at 423 K
for 60 min.
2 M. Okawara and K. Yamamoto, Kegyo Kagaku Zasschi, 1957, 60,
78 (Chem. Abs., 1959, 53, 6179h).
3 (a) K. G. Allum, I. V. Howell and R. C. Pitkethly, Brit. Pat.,
1,378,151, 1974, (Chem. Abstr., 1975, 82, 125043r); (b) N. N.
Rozhdestvenskaya, G. V. Isagulyanto, L. A. Kupcha and N. S.
Kozlov, Izv. Akad. Nauk SSSR, Ser. Khim., 1991, 1751 (Chem. Abs.,
1992, 116, 20722d); J. Haber, M. Wojciechowska and W. Gut,
Bull. Pol. Acad. Sci. Chem., 1996, 44, 55 (Chem. Abs., 1996, 125,
301640a); N. S. Kozlov, K. Zhavnerlo and M. I. Kozlyak,
Dokl. Akad. Nauk Beloruss SSR, 1970, 14, 914 (Chem. Abs., 1971,
74, 76099n); H. S. Kim, D. I. Kim, C-S. Kim and Y. J. Joo,
Kongop Hwahak, 1994, 5, 871 (Chem. Abs., 1995, 123, 143142s).
(c) Cyclohexene. At 323 K, conversion to benzene was 40%
complete after 1 h.
(d) Cyclohexadiene. Complete conversion to benzene was
observed within 1 minute (GC and GC/MS).
(e) 1-Methyl-4-isopropylcyclohexa-1,3-diene (ꢀ-terpinene).
Complete conversion to p-cymene (1-methyl-4-isopropenyl-
benzene) was observed in 5 min.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 5 6 5 – 5 7 4
573

View Article Online
4 T. J. Wallace, J. M. Miller, H. Probner and A. Schriesheim,
Proc. Chem. Soc., 1962, 384; T. J. Wallace, A. Schriesheim and
W. Bartok, J. Org. Chem., 1963, 28, 1311.
5 G. A. Russel and E. G. Janzen, J. Am. Chem. Soc., 1962, 84, 4513;
G. A. Russel, R. G. Janzen, H.-D. Becker and F. Smentowski,
J. Am. Chem. Soc., 1962, 84, 2652.
6 (a) R. D. Guthrie and N. S. Cho, J. Am. Chem. Soc., 1975, 97, 2280;
(b) D. N. R. Rao and R. F. Mason, Basic Life Sci., 1988, 49, (Oxygen
Radicals in Biology and Medicine, 787); M. Rahman and S. K. Naher,
J. Organomet. Chem., 1987, 329, 133.
resonance integral.23b,29 Using the value for hν of 2.63 eV for the
470 nm charge transfer, I and EA from the literature, and C by
calculation, a value of 1.81 eV for β can be derived, which lies within
the reported small range of values for this integral.34 The resonance
energy gained on forming the tight complex is given by β2/(I Ϫ EA ϩ
C),23b,29 which in the present case for DHA–PhNO2–Hϩ is 78 kJ
molϪ1
.
31 In the absence of an ionization energy for DHA itself, the value for
1,2-dimethylbenzene was used. K. Watanabe, J. Chem. Phys., 1957,
26, 542.
7 M. K. Eberhardt, Tetrahedron Lett., 1984, 25, 3663; J. Eibenberger,
D. Schulte-Frohlinde and S. Steenben, J. Phys. Chem., 1980, 84, 704.
8 N. Kornblum and M. J. Fifolt, Tetrahedron, 1989, 45, 1311;
L. Weisler and R. W. Helmkamp, J. Am. Chem. Soc., 1945, 67, 1167;
R. Bacaloglu, A. Blasko, C. Bunton, E. Dorwin, F. Ortega and
C. Zucco, J. Am. Chem. Soc., 1991, 113, 238.
9 M. Hudlicky, Reductions in Organic Synthesis, ACS Monograph
188, American Chemical Society, Washington DC, 1996, pp. 94–95.
10 R. P. Austin and J. H. Ridd, J. Chem. Soc., Chem. Commun., 1992,
1599; R. P. Austin and J. H. Ridd, J. Chem. Soc., Perkin Trans. 2,
1994, 1411; R. P. Austin and J. H. Ridd, J. Chem. Soc., Perkin Trans.
2, 1994, 1205.
32 The electron affinity for protonated nitrobenzene is unknown but it
can be estimated quite accurately. The largest electron affinities are
about 3 eV and nitrobenzene itself has a value of 1.01 eV.35 It might
therefore be expected that protonated nitrobenzene should have an
electron affinity in the range 1.01 < EA < 3.0 eV. It has been pointed
out that reduction potentials (E1/2) should correlate with the orbital
energy of the first unoccupied molecular orbital in a molecule.36
Electron affinity should also correlate with this same orbital energy
and, therefore, there should be a linear correlation between electron
affinity (EA) and reduction potential (E1/2). By using values for
electron affinities for a series of compounds35 and known E1/2
reduction potentials for the same compounds,37 a graph can be
11 G. A. Clowes, J. Chem. Soc. (C), 1968, 2519.
12 I. D. Entwistle, R. A. W. Johnstone and A. H. Wilby, Chemical
Reviews, 1985, 85, 129.
13 J. March, Advanced Organic Chemistry, 4th edn., Wiley, New York,
1992, pp. 1216–1217.
14 D. H. Geske and A. H. Maki, J. Am. Chem. Soc., 1960, 82, 2671;
B. Kalyanaram, E. Perez-Reyes and R. P. Mason, Mol. Pharm.,
1979, 16, 1059; B. Kalyanaram, R. P. Mason, R. Rowlett and
L. D. Kispert, Biochim. Biophys. Acta, 1981, 660, 102; Y. Ito,
T. Konoika and T. Saeguta, Terahedron Lett., 1974, 1287.
15 A. J. Fry, Synthetic Organic Electrochemistry, Harper and Row, New
York, 1972, pp. 224–234; A.J. Fry, in The Chemistry of Amino,
Nitroso and Nitro Compounds and their Derivatives, suppl. F, part 1,
ch. 8, ed. S. Patai, Wiley, Chichester, 1982, p. 319; B. Kastening,
Electrochim. Acta, 1964, 9, 241.
plotted, which gives a good linear correlation with, EA = 1.49(E1/2)
ϩ 2.79 eV (correlation coefficient = 0.9847). At the acidities used in
the experiments described here, the E1/2 value for nitrobenzene is
Ϫ0.14 V.27 Therefore, from the linear correlation with EA, the
electron affinity of protonated nitrobenzene is estimated to be 2.6
eV.
33 The Coulombic term (C) or potential energy for the complex can be
derived from the expression, C = z1z2NAe2/4πε0r, in which z1,z2 are
the numbers of charges (z1 = z2 = 1 for transfer of one electron), NA
is Avagadro’s number, e is the charge on the electron (1.6 × 10Ϫ19 C),
ε0 is the permittivity of the solution (= permittivity of free space ×
dielectric constant for nitrobenzene = 3.19 × 10Ϫ10 F mϪ1) and r is the
distance apart of the two charged species (closest approach of two
carbon atom radii (1.34 × 10Ϫ10 m). Insertion of these values gives C
= 0.3 eV.
16 C. D. Nenitzescu and A. Balaban, Chem. Ber., 1958, 91, 2109.
17 A. J. Gordon and R. A. Ford, The Chemist’s Companion, Wiley,
New York, 1972, pp. 138–140.
34 R. L. Flurry, Molecular Orbital Theories of Bonding in Organic
Molecules, Edward Arnold, London, 1968, p. 59.
35 S. Chowdhury, T. Heinis, E. Grimsrud and P. Kebarle, J. Phys.
Chem., 1986, 90, 2747.
36 G. J. Hoijtink, Rec. Trav. Chim. Pays-Bas, 1955, 1525; G. Cauquis,
in Organic Electrochemistry, ed. M. M. Baizer, Marcel Dekker,
New York, 1973, pp. 75–79.
18 All heats of formation are known accurately for a temperature
of 25 ЊC: dihydroanthracene (167 kJ molϪ1),19 anthracene (234 kJ
molϪ1),19 nitrobenzene (67 kJ molϪ1),20a aniline (88 kJ molϪ1),20a
20a
water (Ϫ242 kJ molϪ1
)
.
19 A. Magnus, H. Hartman and F. Becker, Z. Physik. Chem., 1951, 197,
37 Reduction potentials were taken from reference 26, (pp. 752, 769,
770, 920, 934–938). Fourteen compounds were found to be common
to references 26,35 and these were used for the correlation of E1/2
with EA. Values for E1/2 from reference 26 were those relating to
solutions in MeCN at pH 7, with a Hg electrode and a standard
calomel electrode as reference.
85.
20 Handbook of Chemistry and Physics, 76th edn., ed. D. R. Lide,
CRC Press, Boca Raton, 1995; (a) pp. 5–36 and 5–16; (b) p. 6–170.
21 J. E. Leffler, Science, 1953, 117, 340; G. S. Hammond, J. Am. Chem.
Soc., 1955, 77, 334.
22 L. J. Andrews, Chem. Rev., 1954, 54, 713; L. J. Andrews and
R. M. Keefer, Molecular Complexes in Organic Chemistry,
Holden-Day, San Francisco, 1964; H. McConnell, J. S. Haw and
R. Platt, J. Phys. Chem., 1953, 21, 66.
23 H. H. Jaffe and M. Orchin, Theory and Applications of Ultraviolet
Spectroscopy, Wiley, New York, 1964; (a) pp. 578–585; (b) pp. 271–
273.
24 (a) O. I. Kachurin and V. M. Belobrov, Ukr. Khim. Zh., 1975, 41,
709; E. M. Arnett, R. P. Quirk and J. J. Burke, J. Am. Chem. Soc.,
1970, 92, 1260; R. C. Paul, K. C. Khanna and K. C. Malhotra,
Res. Bull. Panjab Univ., 1965, 16, 251; R. J. Gillespie and
E. A. Robinson, J. Chem. Soc., 1957, 4233; (b) M. E. Peach and
T. C. Waddington, J. Chem. Soc., 1962, 2680.
25 H. C. Brown and M. Grayson, J. Am. Chem. Soc., 1953, 75, 6285.
26 H. Siegerman, in Technique of Electroorganic Synthesis, vol. V, part
II, ed. N. L. Weinberg, Wiley, New York, 1975, pp. 667.
27 J. E. Page, J. W. Smith and J. G. Weller, J. Phys. Chem., 1949, 53, 545;
S. K. Vijayalakshamma and R. S. Subrahmanya, J. Electroanal.
Chem., 1969, 23, 99.
28 J.O’M. Bockris and A. K. N. Reddy, Modern Electrochemistry,
Plenum Press, New York, p. 906.
29 R. S. Mulliken, J. Am. Chem. Soc., 1952, 74, 811.
30 The three terms are: the ionization energy (I ) for 9,10-
dihydroanthracene (8.56 eV),31 the electron affinity (EA) for
protonated nitrobenzene ( 2.6 eV),32 and the Coulombic energy (C)
of attraction between a positive and a negative ion (0.3 eV).33 The
frequency (ν) for a charge-transfer band at 470 nm is predicted
approximately by the equation, hν = 1 ϩ 2β2/(I Ϫ EA ϩ C); β is a
38 F. H. Herbstein, M. Kapon and G. M. Reisner, Acta Crystallogr.
Sect. B, 1986, 42, 181.
39 L. R. Melby, R. J. Harder, W. R. Hertler, W. Mahler, R. E. Benson
and W. E. Mochel, J. Am. Chem. Soc., 1962, 84, 3374.
20b
40 The heat capacity for nitrobenzene is 186 J molϪ1 KϪ1
.
Heat
content per mole of nitrobenzene heated from 20 to 100 ЊC = 186 ×
80 = 15 kJ molϪ1. This is a minimum value because, in the
experiments described here, the nitrobenzene was always used in
excess as a solvent as well as a reactant.
41 J. L. Franklin, J. G. Dillard, H. M. Rosenstock, J. T. Herron,
K. Draxl and F. H. Field, Ionization Potentials, Appearance
Potentials, and Heats of Formation of Gaseous Positive Ions,
National Bureau of Standards, NSRDS-NBS 26, Washington DC,
1969, cyclohexane 9.86 eV, cyclohexene 8.95 eV, cyclohexadiene
8.4 eV.
42 Y. Zhao and F. G. Bordwell, J. Org. Chem., 1996, 61, 2530.
43 C. Hansch, A. Leo and R. W. Taft, Chem. Rev., 1991, 91, 185.
44 R. A. W. Johnstone, M. L. P. G. Nunes, M. M. Pereira,
A. M. d’A Rocha Gonsalves and A. Serra, Heterocycles, 1996, 42,
1423 and references cited therein.
45 (a) M. Orchin and L. Reggel, J. Am. Chem. Soc., 1947, 69, 505;
(b) C. D. Nenitzescu and M. Auram, J. Am. Chem. Soc., 1950, 72,
3486; A. Dansi and M. Reggiani, Gazz. Chim. Ital., 1948, 78, 801.
46 W. P. Jenks, Catalysis in Chemistry and Enzymology, McGraw-Hill,
New York, 1969; A. Fersht, Enzyme Structure and Mechanism,
W. H. Freeman, London, 1979.
47 D. Dolphin, J. Heterocycl. Chem., 1970, 7, 275.
48 D. W. Thomas and A. E. Martell, J. Am. Chem. Soc., 1956, 78, 1335.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 5 6 5 – 5 7 4
574