Reaction products and mechanism of the regioselective oxidation of N-phenylmorpholine by ozone

Article abstract of DOI:10.1016/j.tet.2012.07.055

The regioselective oxidation of N-phenylmorpholine by ozone in dichloromethane or acetonitrile produced a lactam and a diformylderivative. These products derive from the selective attack of ozone at the heterocyclic ring in one of the two non-equivalent reactive carbons. The reaction mechanism has been investigated by DFT calculations, which show that the reaction occurs through the insertion of ozone at the carbon-hydrogen bond of a methylene group of the morpholine ring. The regioselectivity is due to the significantly lower energy barrier calculated for the attack of ozone α to nitrogen than α to oxygen. In addition, the energy barrier decreases with increasing the polarity of the solvent, explaining the higher conversions observed for the reaction carried out in acetonitrile than in dichloromethane.

Full text of DOI:10.1016/j.tet.2012.07.055

Tetrahedron 68 (2012) 8267e8275
Contents lists available at SciVerse ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
Reaction products and mechanism of the regioselective oxidation
of N-phenylmorpholine by ozone
*
Ricardo Suarez-Bertoa, Francesco Saliu, Maurizio Bruschi , Bruno Rindone
Department of Environmental Sciences, University of Milano-Bicocca, Piazza della Scienza 1, 20126 Milano, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
The regioselective oxidation of N-phenylmorpholine by ozone in dichloromethane or acetonitrile pro-
duced a lactam and a diformylderivative. These products derive from the selective attack of ozone at the
heterocyclic ring in one of the two non-equivalent reactive carbons. The reaction mechanism has been
investigated by DFT calculations, which show that the reaction occurs through the insertion of ozone at
the carbonehydrogen bond of a methylene group of the morpholine ring. The regioselectivity is due to
Received 12 February 2012
Received in revised form 27 June 2012
Accepted 17 July 2012
Available online 25 July 2012
the significantly lower energy barrier calculated for the attack of ozone
a to nitrogen than a to oxygen. In
Keywords:
Ozone
Oxidation
N-Phenylmorpholine
Lactam
addition, the energy barrier decreases with increasing the polarity of the solvent, explaining the higher
conversions observed for the reaction carried out in acetonitrile than in dichloromethane.
Ó 2012 Elsevier Ltd. All rights reserved.
DFT calculations
Reaction mechanism
1. Introduction
In this work we present a combination of experimental and
quantum chemical studies for elucidating the reaction products and
The fundamental problem in the functionalization of saturated
hydrocarbons is that their components, carbon and hydrogen, do
not have electron pairs, and do not have an easily accessible orbital
of proper energy. Thus, very reactive reagents and/or extreme re-
action conditions are usually required. Moreover, reaction products
are usually more reactive than the starting materials, and thus
the mechanism of the regioselective oxidation with ozone of the
aromatic amine N-phenylmorpholine 1. This molecule has been
selected as a suitable model for this study as non-equivalent CH
groups to nitrogen and to oxygen are present.
2
a
a
Ozonation of aliphatic tertiary amines has been extensively in-
vestigated inwater as well as in organic solvents showing that the main
product in most of the reaction conditions is the corresponding amine-
1
undesired side reactions may occur. Ozone is a metal-free alter-
12e15
native to these methods, and the reactivity of this compound with
organic molecules has been the focus of organic synthesis research
for decades.2
oxide.
However, strong solvent and temperature effects were ob-
served in the products distribution: high yields of amine-oxide were
reported in solvents such as chloroform, methanol, and water, while
only a small amount of amine-oxide is obtained in butyl-chloride and
pentane. In the proposed mechanism the amine-oxide is formed
through addition of O to nitrogen of the amine to give the R NeO
3 3 3
adduct followed by elimination of singlet dioxygen (Eq.1 in Scheme 1).
The other reaction products occur from the oxidation of the amine side
e4
Ozone has been used by some of us for the oxidation of cyclo-
5
6
alkanes to cycloalkanol and cycloalkanones, ethers to esters,
amines to amides.7 These studies suggest that ozone is a suit-
able oxidant, but also that the process requires improvement of
both conversion and selectivity. In particular, with this oxidant it is
difficult to control the oxidation position within the carbon chain.
In fact, to our knowledge, the only studies where ozone was used in
regioselective synthesis are the oxidation of functionalized imide
e9
chains. In this case the mechanism should proceed through the disso-
þ
ciation of the NeO
3
adduct to give the amine radical cation R
3
N and
ꢀ
the ozonide radical O , or from an intramolecular hydrogen abstraction
3
10
proposed by Poschenrieder et al., and the stereoselective oxida-
3 3
of the R NeO adduct and the subsequent rearrangement of the in-
tion of (þ)-aromadendrene proposed by van Lier et al. where ozone
termediate (Eqs. 2 and 3 in Scheme 1). An alternative mechanism
considered in Ref. 14 is the 1,3 dipolar insertion reaction at one CeH
bond of the side chains to give the hydrotrioxide intermediate (Eq. 4 in
Scheme 1). Indeed, the latter mechanism has been proposed to account
for the reaction products upon ozonation of aromatic aldehydes and
aliphatic ethers. This mechanism may proceed by a hydrogen
11
reacts with a more reactive tertiary carbon.
*
Corresponding author. Tel.: þ39 (0)2 64482816; fax: þ39 (0)2 64482890;
e-mail address: Maurizio.Bruschi@unimib.it (M. Bruschi).
0
040-4020/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.tet.2012.07.055

8268
R. Suarez-Bertoa et al. / Tetrahedron 68 (2012) 8267e8275
+
3
-
R N
+ O3
R N-O
O2
Eq. 1
Eq. 2
3
R N-O-O-O
3
+
-
-
O-O-O
O-O-OH
OH
R N
+ O
+ O
3
3
2
NR CH R'
R N CHR'
2
2
R N CHR'
2
2
+
+
-
O-O-O
O-O-OH
OH
R N
+ O3
Eq. 3
Eq. 4
3
+
O2
NR CH R'
NR CHR'
R N CHR'
2
2
2
2
+
R N CHR'
2
R N CHR'
OH
R N CHR'
2
H
O
+ O2
O-O-O-H
2
O O
-
+
-
/.
-
O-O-OH
O-O-O
O-O-OH
NR2
.
OH
CHR'
+/
R N CH R'
2
R N CHR'
+
R
N
CHR'
R N
+
O3
+ O
2
Eq. 5
Scheme 1.
abstraction followed by the formation of the radical pair R/OOOH,
In the section ‘Results and discussion’ we first report on the
product distribution obtained from the ozonation of N-phenyl-
morpholineatseveraltemperaturesand indifferentsolvents, thenwe
discuss the reaction mechanism on the basis of the results obtained
from the theoretical investigation. In particular, it will be shown that
a low energy pathway can account for the product distribution.
16
which then recombine to form ROOOH, or by a hydride abstraction
þ ꢀ
with the formation of the ion pair R OOOH that again recombine to
17
form ROOOH, as also reviewed for the reactions of acetals with
1
8e22
ozone.
mechanism for aliphatic tertiary amines.
Ozonation of tertiary aromatic amines has been less studied than
However, experimental results seem to exclude this
14
2
3,24
the corresponding reaction with aliphatic amines.
In particular, it
was demonstrated that amine-oxide, which is the major product for
aliphatic amines, is not formed in the ozonation of aromatic amines.
Meth-Cohn and Kerralsoshowed that the absence of the amine-oxide
2. Results and discussion
2.1. Experimental results
24
is not due to further reactions of this species with ozone. On this
basis they proposed a mechanism in which O adds to the nitrogen
atom to give the NeO adduct, followed by an intermolecular hy-
3
N-Phenylmorpholine 1 was reacted in dichloromethane or
acetonitrile with excess ozone at different temperatures and at
different reaction times in order to study the effect of the reaction
conditions over the distribution of the products. Table 1 summa-
rizes the obtained results.
3
drogen abstraction and a decomposition of the corresponding adduct
to give radical and/or cationic intermediates, which further recom-
bine to give the final products (Eq. 5 in Scheme 1).
Table 1
Conversions and isolated yields of the products 2
N
and 3 in the reaction of N-phenylmorpholine 1 with ozone (DCM¼dichloromethane; ACN¼acetonitrile)
ꢁ
Conversiona
Lactam 2 b
%
Entry
T ( C)
t (h)
Solvent
Ozone
N
Diformylderivative
3 (%)
Other (%)
b
(%)
1
2
3
4
5
6
7
8
9
25
0
1.5
1.5
0.5
24
DCM
DCM
DCM
DCM
DCM
ACN
ACN
ACN
ACN
ACN
Excess
Excess
Excess
10:1
80
84
100
38
6
99
100
98
100
30
20
24
23
4
6
7
4
3
3
17
14
6
5:tr
4:tr; 5:tr
6:tr
ꢀ20
25
25
25
0
ꢀ20
25
0
24
5:1
1
6:tr
0.5
0.5
0.5
24
Excess
Excess
Excess
10:1
24
34
27
46
16
8
3
1
0
1
1:1
a
Conversion is calculated on the basis of the unreacted N-phenylmorpholine recovered.
Isolated yield after silica gel chromatography (see the Experimental part).
b

R. Suarez-Bertoa et al. / Tetrahedron 68 (2012) 8267e8275
8269
ꢁ
In dichloromethane at 25 C, after 1.5 h, 80% of the starting
material was converted into products. Two reaction products could
be separated by silica gel chromatography and were identified as
In summary, the collected data show that:
1. A major conversion is always observed in acetonitrile compared
to dichloromethane. In fact, in the reactions with excess ozone
the conversion of N-phenylmorpholine 1 in acetonitrile is
complete at all of the temperatures investigated. In dichloro-
N
the lactam 2 and the diformylderivative 3 (Scheme 2) by com-
parison of their IR, MS, H NMR, and C NMR spectra with those of
1
13
authentic samples.2
5,26
ꢁ
methane the conversion is complete only at ꢀ20 C, while par-
O
N
O
N
O
N
O
N
ꢁ
tial conversion is observed at 0 and 25 C. Moreover, with a 10:1
CHO
CHO
excess of ozone the starting material is completely converted in
acetonitrile, not in dichloromethane. This may be explained by
a solvent polarity effect on the reaction mechanism (vide infra).
O
2
3
. In dichloromethane the conversion seems to be higher at lower
temperatures than at higher temperatures.
. Mass balance is generally poor, probably due to the further oxi-
dation of the primary reaction products. The higher recovery was
obtained in the reactions performed in acetonitrile with a 10:1
N
excess of ozone, but the yields in compounds 2 and 3 were very
1
2N
3
4
similar to those of the reactions performed with excess ozone.
This suggests that further oxidation of the primary reaction
products is faster than the oxidation of the starting material.
. Notwithstanding the poor recovery, all the reactions showed
a very interesting regioselectivity and further developments
are underway in our laboratory in order to develop a valuable
synthetic procedure. In particular, the main isolated reaction
4
O
N
O
N
O
*
O
N
*
*
N
products, N-phenylmorpholinone 2 and the diformylder-
N
OH
ivative of the N-phenylmorpholine 3, both derive from the
regioselective attack of ozone at the heterocyclic ring, involving
at some step of the reaction pathway the activation of the
Cl
methylene group
from the attack
a
a
to nitrogen. No traces of products deriving
to oxygen were isolated. Moreover, the
diformylderivative could be one of the degradation products
deriving from further oxidation of the intermediate alcohol 7
see Scheme 2) that seems to be preferred.
N
5
6
7N
(
Scheme 2. Reaction products from the ozonation of N-phenylmorpholine 1.
These results are in line with those obtained earlier.24 In fact, the
formation of lactam 2 from N-phenylmorpholine 1 corresponds to
the formation of an N-methylformanilide derivative from N,N-
dimethylaniline and its 4-substituted derivatives. However, in the
case of N,N-dimethylaniline the path, which leads to the dehy-
drointermediate 4, and consequently to the diformylderivative 3, is
not possible since the substituent at nitrogen is a one-carbon unit.
N
ꢁ
The reaction was repeated in dichloromethane at 0 C for 1 h
ꢁ
and at ꢀ20 C for 0.5 h. The conversion was higher compared to the
reaction carried out at room temperature and products 2 and 3
were again isolated.
In all these experiments recovery of material after silica gel
chromatography was poor: approximately one half of the starting
material was converted into highly polar products produced by the
decomposition of the parent compound.
2.2. Theoretical investigation of the reaction mechanism
Moreover, traces of other reaction products were observed in the
GCeMS chromatograms of the crude reaction mixture. On the basis of
the detected mass spectra and tentative structural elucidation they
The mechanism and the regiochemistry of the reaction have
been investigated by quantum mechanics calculations through the
identification of the relevant stationary points (minima and saddle
points) along the reaction coordinate of the pathways A and B
reported in Scheme 3. The concurrent ozonation of the aromatic
ring by the mechanism shown in Scheme 4 has been studied, as
well. Indeed, decomposition of the aromatic ring may account for
the partial recovery of the reaction products. The calculations were
performed in vacuum, as well as in acetonitrile and dichloro-
methane to evaluate the solvent effect on the reaction mechanism.
In the following, for the sake of clarity, the subscripts N and O
will be added to labels of chemical species, to indicate regioisomers
þ
were recognized as the dehydroderivative 4 (m/z¼161 (M ),132,104,
þ
7
7) and stereoisomers of the dimeric structure 5 (m/z¼322 (M )).
Oxidation of N-phenylmorpholine 1 inacetonitrileassolvent, with
ꢁ
excess of ozone at 25, 0 and ꢀ20 C for 0.5 h, gave a complete con-
N
version, and again lactam 2 and diformylderivative 3 were isolated.
Also in these cases, poor recovery was observed (less than 50%).
Since shorter reaction times did not give higher mass balance, ex-
periments were performed using a controlled amount of ozone in the
attempt to minimize the further oxidation of primary reaction
products.
differing by substituents or reacting centers
ygen, respectively.
a to nitrogen and ox-
Using a 10:1 excess of ozone and a 24 h reaction time at room
temperature, in dichloromethane a better recovery was obtained:
The first step in the proposed mechanism could be the forma-
tion of intermediates A and A , or B and B , through pathways A,
3
8% of the starting material and minute amounts of compounds 2
N
N
O
N
O
and 3 were present. Also isomers deriving from the chlorination of
the starting material 6 were detected in the GCeMS of the crude
reaction mixture. By replacing dichloromethane with acetonitrile
and using the same conditions, complete conversion was instead
or B, respectively (Scheme 3).
Pathway A, as proposed for the oxidation of N,N-dimethylaniline
in Ref. 24, occurs with the addition of O to nitrogen leading to Int
3
A
[reaction (1a)], followed by an intramolecular hydrogen transfer
from a methylene group to nitrogen or to oxygen to the terminal
oxygen atom to give the intermediate AN or A , respectively
observed. Compounds 2
N
and 3 were present in 31 and 15% yields,
a
a
respectively.
O

8270
R. Suarez-Bertoa et al. / Tetrahedron 68 (2012) 8267e8275
O O O O N N N N
Scheme 3. The regiochemistry of the ozonation of N-phenylmorpholine. A , B , Rad and Cat , not reported in the scheme, are derived from A , B , Rad and Cat , respectively, by
placing the reacting center or the functional group to the oxygen atom.
a
ꢀ
ꢂ
[
A
reaction (2a)]. Finally, elimination of O
2
and OH /OH from A
N
or
A
discussed below. Int may also loose singlet dioxygen producing
ꢀ
O
, and the subsequent recombination of OH /OH with the radical/
cationic substrate leads to the N-phenylmorpholinone 2
may dissociate into
the ozonide radical O3 and the N-phenylmorpholyl radical cation
the amine-oxide 12 [reaction (6a)]. The last reaction, which is the
major route for aliphatic amines, should not occur for tertiary ar-
omatic amines, as the amine-oxide has never been isolated from
N
or 2
O
[reaction (3a)]. Alternatively, the adduct Int
A
ꢀ
24
the ozonation of this class of compounds.
D
D
1
[reaction (4a)]. The ozonide may further react with 1 by
In the pathway B the first step is a hydride or hydrogen transfer from
abstracting a hydrogen atom or a proton from a methylene group of
a methylene carbon atom to the terminal oxygen atom of O
dride abstraction from a methylene group to nitrogen or to oxygen
of the morpholine ring leads to the formation of the N-phenyl-
3
. The hy-
ꢂ
the morpholine ring to form HO
3
and Rad
(hydrogen abstraction) [re-
action (5a)]. This pathway then proceeds with the recombination of
the two species to form the hydrotrioxide B or B of the pathway B
N
or Rad
O
(proton ab-
a
a
e
3 N O
straction), or HO and Cat or Cat
morpholyl cation Cat
N
O
or Cat , respectively, and the hydrotrioxide
ꢀ
N
O
anion ðHO Þ [reaction (1b
3
2
)]. Alternatively, the hydrogen transfer leads
Scheme 4. Mechanisms for the ozonation of the aromatic ring.

R. Suarez-Bertoa et al. / Tetrahedron 68 (2012) 8267e8275
8271
to the formation of the N-phenylmorpholyl radicals Rad
N
or Rad
O
and
regioselectivity and the product distribution observed experimen-
tally. Energies and relevant energy differences for the species
considered in this work are reported in Supplementary data.
ꢂ
the hydrotrioxide radical HO
the terminal oxygen atom of HO3 or HO
atom then gives the hydrotrioxide B or B
3
[reaction (1b
1
)]. The recombination of
with the methylene carbon
[reaction (2b or 2b )]. The
ꢀ
ꢂ
3
N
O
1
2
The full electron transfer from 1 to O
3
to give the pair of separate
radical ions 1 and O3 [reaction (7a)] is a process in which the
products are less stable than 1 and O by about 112, 11, and
CN, respectively. It is worth
noting that in a strongly polar solvent such as water the pair of
D
ꢀ
process may also occur through a concerted mechanism where hy-
drogen transfer and CeO bond formation occurs simultaneously.
3
ꢀ
1
The key hydrotrioxide intermediates B
N
and B
O
may further
2 2 3
2 kcal mol in vacuum, CH Cl and CH
evolve following the pathways:
D
ꢀ
radical ions 1 and O3 are likely to be significantly more stable
(
(
(
1) Dissociation to a carbonyl compound (2
N
or 2
O
) and hydrogen
than separate reactants, due to the large solvation energy of ionic
species. This is consistent with the large yield of OH formed from
ꢂ
peroxide [reaction (3b)].
2) Dismutation to an alcohol (7
N
or 7
O
) and dioxygen [reaction
the ozonation of aniline in aqueous solution, for which the pre-
27
ꢀ
28
(
4b)]. Singlet dioxygen is reported to be formed in this step.
3) Rearrangement to an aldehyde hydroperoxide (8 or 8 ) with
further oxidation to carboxylic acids [reaction (5b)].
cursor is O . These results also suggest that the reaction could
3
N
O
occur with different mechanisms depending on the polarity of the
solvent.
Notably, any attempt to find an energy minimum for Int
A
in
or to the evolution of
with the formation of the amine-oxide 12 [reaction (6a)],
depending on the initial geometry for the optimization. Therefore,
the key intermediate Int , with an oxygen atom of O covalently
bounded to nitrogen, in pathway A is not likely to be formed. In
addition, A and A are not stable species as the geometry opti-
mization of these compounds leads to the dissociation of the O
Previous studies on cycloalkanes5 and ethers6 reported that
decomposition of the intermediate hydrotrioxide to form carbonyl
compounds predominates over their fragmentation to the corre-
sponding alcohols or to open chain compounds.
pathway A leads to the dissociation of O
singlet O
29
3
,
2
A
3
Concerning the regiochemistry of the reaction, fragmentation of
hydrotrioxide B
is the major product observed experimentally. The alternative for-
mation of the isomeric hydrotrioxide B seems not to occur be-
cause the lactone 2 has never been found.
Scheme 5 shows the decomposition of the intermediate alcohols
and 7 . Dehydration of alcohol 7 should give the dehydroder-
N
gives the lactam 2
N
[reaction (3b)], which, in fact
N
O
3
H
O
group. These results clearly indicate that the mechanism proposed
in Ref. 24, corresponding to pathway A in Scheme 3 does not occur
for N-phenylmorpholine. In this respect, we note that the formation
O
7
N
O
N
A
of a positive nitrogen in the rate-determining step, as for Int ,
ivative 4, which was found in trace amounts because of its fast
oxidation to the diformylderivative 3. Alternatively, the de-
should imply a much higher reactivity of 4-methoxy-N,N-dime-
thylaniline over N,N-dimethylaniline due to an electronic effect.
This effect was not observed in Ref. 24 suggesting that also for N,N-
dimethylaniline and its 4-substituted derivatives the reaction may
occur with a different mechanism than that illustrated in path A.
The results discussed above suggest that the amine-oxide 12
hydration of 7
material 5, in accord with the dimerization of N-phenylpyrrolidine
upon ozonation.24 Again, alcohol 7
seems not to be formed, since
N
to the immonium ion 9 is the origin of the dimeric
O
the final formylderivative 11 was never found.
[reaction (6a)] may form directly from the reagents. However, the
system composed by the amine-oxide 12 and singlet dioxygen is
less stable than the separate reactants by about 7, 3 and
ꢀ
1
2
kcal mol , in vacuum, dichloromethane, and acetonitrile. Nota-
bly, in the case of the tertiary aliphatic amine (CH N, the system
composed by the corresponding amine-oxide and singlet dioxygen
is more stable than the separate reactants (CH N and O by 4 and
in dichloromethane and acetonitrile, respectively,
3 3
)
3
)
3
3
ꢀ
kcal mol
1
7
ꢀ
1
while it remains less stable by about 3 kcal mol in vacuum (data
not shown). The different stability between the amine-oxide pro-
3 3
duced from the aliphatic tertiary amine (CH ) N and the aromatic
tertiary amine N-phenylmorpholine would explain why the amine-
oxide is the main product in the ozonation of an aliphatic amine,
while it is not observed in the ozonation of an aromatic amine.
In the pathway B, the reaction, starting from the reagents, can
proceed along two different paths depending on the secondary
carbon atom attacked by ozone, namely either
oxygen.
a to nitrogen or a to
In both cases the formation of the hydrotrioxide intermediates
and B proceeds through a hydrogen abstraction mechanism
B
N
O
followed by recombination of the ionic or radical pairs (see below).
In spite of the many attempts made, we were not able to find
a transition state for the concerted mechanism.
The transition state for the hydrogen abstraction
is shown in Fig. 2. In vacuum, the corresponding activation
energy (E ), calculated with respect to the separate reactants, is
equal to 7.6 kcal mol . The hydrogen abstraction
to the transition state TS (see Fig. 2), which is 13.4 kcal mol
higher in energy than the separate reactants; a value about
a to nitrogen
TS
N
a
ꢀ
1
a
to oxygen leads
ꢀ1
O
Scheme 5. The fate of the intermediate alcohols.
ꢀ
1
6
kcal mol higher than that calculated for TS
N
.
It should be noted that in both cases the transition state involves
the abstraction of an axial hydrogen. Any attempts to find a tran-
sition state for the abstraction of an equatorial hydrogen leads to
The computational investigation of the reaction mechanism,
summarized by the energy profile shown in Fig. 1, explains the

8272
R. Suarez-Bertoa et al. / Tetrahedron 68 (2012) 8267e8275
Fig. 1. (a) Energy profile along the reaction coordinate for the reaction of 1 with ozone calculated in vacuum. (b) Energy profile along the reaction coordinate for the first step of the
reaction of 1 with ozone calculated in vacuum (
), dichloromethane (
), and acetonitrile (
). Solid and dashed lines indicate the pathway with H-abstraction a to
nitrogen and oxygen, respectively.
a rearrangement of the heterocycle and a convergence of the ge-
ometry optimization to TS or TS
The activation energies for both TS
cantly smaller when the reaction is carried out in dichloromethane
or acetonitrile solutions. The reduction of the activation energies is
also a function of the polarity of the solvent. In fact, the activation
smaller than that calculated for the H-abstraction of an aliphatic
carbon atom. For example, the activation energy calculated in
vacuum for the for the H-abstraction of cyclohexane is about
N
O
.
N
O
and TS became signifi-
ꢀ1
ꢀ1
26 kcal mol (data not shown) a value 18 kcal mol higher than
that calculated for TS
O
The reaction, from TS and TS , can evolve through a homolytic
N
.
N
energy for TS
decreases from 7.6 kcal mol
dichloromethane, to a value as small as equal to 1.6 kcal mol in
acetonitrile. In the case of TS the reduction is parallel of that cal-
N
, calculated with respect to the separate reactants,
bond dissociation to form a radical pair, or through a heterolytic
ꢀ
1
ꢀ1
in vacuum to 2.6 kcal mol
in
bond dissociation to form an ionic pair. The homolytic H-abstrac-
ꢀ
1
ꢂ
tion
a
to nitrogen and oxygen leads to O
3
H and the radical species
O
Rad
N
and Rad , respectively [reaction (1b
O
1
)]. In vacuum the system
ꢀ1
ꢂ
culated for TS
N
; E
a
decreases from 13.2 kcal mol in vacuum to
composed by the Rad
than the separate reactants (
the radical pair slightly increases in the two solvents considered;
N
eO
3
H radical pair is slightly lower in energy
ꢀ1
ꢀ1
ꢀ1
9
.7 kcal mol
acetonitrile.
The effect of the aromatic nitrogen on the H-abstraction
atom is remarkable, as the E calculated for TS
in dichloromethane, and to 8.9 kcal mol
in
DE¼ꢀ0.9 kcal mol ). The stability of
ꢀ1
a
to this
D
E¼ꢀ3.0 and ꢀ2.0 kcal mol in dichloromethane and acetonitrile,
a
N
is significantly
respectively (see Fig. 1b).

R. Suarez-Bertoa et al. / Tetrahedron 68 (2012) 8267e8275
8273
The formation of the hydrotrioxide intermediates (B
N
, B
O
) is the
crucial step in order to explain the regioselectivity of the reaction.
The results illustrated above clearly show that the methylene group
a
to nitrogen is more reactive with ozone than the methylene group
a
to oxygen. This is due to the significantly smaller activation en-
ergy for the H-abstraction. In addition, the radical or ionic pair
formed by the hydrogen abstraction to nitrogen is more stable
than that formed by H-abstraction to oxygen. In fact, in
dichloromethane and acetonitrile, the radical or ionic pair given by
the attack of ozone to nitrogen is always more stable than sepa-
rate reactants, while the radical or ionic pair given by the attack of
ozone to oxygen is less stable than separate reactants. The smaller
a
a
a
a
activation energy and the higher stability of the ionic or radical pair
as intermediates in the reaction make significantly more favorable
the functionalization
hydrotrioxide B
In acetonitrile the ionic pair is more stable than the radical pair
a to nitrogen and the formation of the
N
.
ꢀ
1
by about 13 kcal mol
indicating that H-abstraction occurs
through a ionic mechanism, as already observed for the ozonation
19e22
of other substrates.
In dichloromethane the radical and ionic
pairs are almost isoenergetic, suggesting a competition between
the ionic and radical mechanism.
A relevant result illustrated above is that the activation energy
for the H-abstraction process decreases by raising the polarity of
the solvent. This would explain the higher conversions observed
experimentally in acetonitrile than in dichloromethane.
Hydrotrioxide B
following three different pathways [reactions (3b), (4b), (5b)]
Scheme 3). Calculations in vacuum were also performed for these
pathways to dissect the mechanisms, which lead in the formation
N O
or B could evolve to give the final products
(
of the lactam 2
N
, and the hemiaminal 7
N
as precursor of the
diformylderivative 3.
N O
Fig. 2. Geometries of the transition states TS and TS , calculated using the B3LYP/cc-
pVTZ scheme.
Dismutation of hydrotrioxides (B and B ) along the pathway
N O
(
4b) gives hemiaminal 7
N
or hemiacetal 7
O
and singlet O
2
. These
intermediates are energetically less stable than B
N
, and B
O
by about
ꢂ
ꢀ1
O 3
The Rad eO H radical pair is less stable in vacuum than in
6
and 4 kcal mol , respectively. The activation energy of this re-
ꢀ
1
separate reactants by about 6 kcal mol . The formation of the
H radical pair is therefore disfavored with respect to the
formation of the Rad
of the solvents on the stability of the Rad
that discussed for the Rad
acetonitrile the Rad
ꢀ1
action step is somewhat large; 46 and 45 kcal mol for TS-Alc
and TS-Alc (see Fig. 3) relative to the intermediates B and B
However, it should be noted that the hemiaminal 7 could still react
N
ꢂ
Rad
O
eO
3
O
N
O
.
ꢂ
ꢀ1
N
eO
3
H pair by about 7 kcal mol . The effect
N
ꢂ
O
eO
3
H pair is parallel to
H pair; in dichloromethane and
H pair is more stable than in vacuum, but
ꢂ
N
eO
3
ꢂ
O
eO
3
always less stable than the separate reactants (
D
E¼3.6 and
ꢀ1
4
.5 kcal mol in dichloromethane and acetonitrile, respectively).
The abstraction of the hydride ion from the methylene group
to nitrogen and oxygen produces the ionic pair composed by the
a
carbocations Cat
N
and Cat
O
, respectively, and the hydrotrioxide
ꢀ
anion O
3
H
[reaction (1b )]. The formation of the ionic pair is
2
ꢀ
1
clearly disfavored in vacuum by more than 100 kcal mol . As
expected, the presence of a polar solvent strongly stabilizes the
ꢀ
ionic pair. Indeed, the Cat
N
eO
3
H
pair is more stable than the
ꢀ1
separate reactants by 5.7 and 15.5 kcal mol in dichloromethane
and acetonitrile, respectively (see Fig. 1b). The effect of the solvents
ꢀ
is similar for the Cat
O
eO
3
H pair, even though in the latter case the
ionic pair remains less stable than the separate reactants by 15.3
ꢀ1
and 4.8 kcal mol in dichloromethane and acetonitrile. Notably,
these results also indicate that in the two solvents Cat is more
N
ꢀ
1
stable than Cat
After the formation of the ionic or radical pairs, the terminal
O
by more than 20 kcal mol .
ꢀ
3 3
H or O H species can easily approach the un-
ꢂ
oxygen of the O
saturated carbon atom forming a carboneoxygen bond and giving
the hydrotrioxides intermediates B and B . These intermediates
have similar energies and are significantly more stable than the
N
O
ꢀ
1
separate reactants (in vacuum
D
E¼ꢀ51.0 kcal mol
). In addition, the stability of B
not affected by the solvent.
for B
and B
N
and
is
ꢀ1
ꢀ
51.7 kcal mol for B
O
N
O
N O N O
Fig. 3. Geometries of the transition states TS-Ket , TS-Ket , TS-Alc , and TS-Alc
calculated using the B3LYP/cc-pVTZ scheme.

8274
R. Suarez-Bertoa et al. / Tetrahedron 68 (2012) 8267e8275
until a more stable product is formed, such as the dehydroder-
ivative 4, which then gives the diformylderivative 3.
activation energy and to the larger stability of the ionic pair formed
by the hydrogen abstraction. In addition, the activation energy
calculated for the H-abstraction of N-phenylmorpholine is much
lower than that calculated for an aliphatic hydrocarbon, explaining
The dissociation of hydrotrioxides (B
way (3b) leads to the formation of the lactam 2
and hydrogen peroxide. The final products are significantly more
N
and B
O
) along the path-
N
or the lactone 2
O
,
the faster reaction kinetic. The hydrotrioxide formed by the re-
ꢀ
1
ꢂ
stable than the separate reactants (about 100 kcal mol ) and the
combination of O
3
H with the substrate then evolve toward the
ꢀ
1
barrier that should be overcome is about 38 kcal mol (TS-Ket
N
,
formation of the lactam 2
with an activation energy smaller than that calculated for the for-
mation of the hemiacetal 7
N
, which is a strongly exothermic process
TS-Ket ; see Fig. 3) relative to the intermediates B and B . Al-
O
N
O
though very large, the energy of the transition state is negative if
compared to that of the separate reactants. The excess energy
N
.
The computational investigation also shows that the mecha-
nism suggested in Ref. 24, with the initial addition of ozone to the
nitrogen atom cannot occur in N-phenylmorpholine since the re-
active intermediate is not stable. In addition, for this aromatic
amine, the amine-oxide is less stable than separate reactants, and
therefore is not expected to be formed, in agreement with experi-
mental results. Finally, ozonation of aromatic ring is expected to
compete with ozonation of the morpholine ring since the energy
barrier for the formation of Criegee ozonide is similar with respect
to the energy barrier for the H-abstraction for the morpholine ring.
Oxidation and subsequent fragmentation of the aromatic ring can
account for the partial recovery of the reaction products.
gained from the system to form intermediates B
N
and B
O
may then
be used to overcome this energy barrier.
In summary, the formation of the hemiaminal 7
trioxide B is an endothermic process, while the formation of the
lactam 2 is a strongly exothermic process with an activation en-
ergy smaller than that calculated for the formation of 7 . However,
as shown in Fig. 1 the excess of energy provided by the formation of
is enough to overcome the energy barriers for the formation of
both the lactam 2 and the hemiaminal 7 . These results can ac-
count for the formation of the hemiaminal 7 , observed experi-
mentally, even though it is significantly less stable than the lactam
N
from hydro-
N
N
N
B
N
N
N
N
2
N
, and with a higher energy barrier.
Ozonation of the aromatic ring may compete with the ozonation
4. Experimental and computational methods
4.1. General
of the morpholine ring. The attack of ozone to the aromatic ring can
lead to the formation of the Criegee ozonide 13, or to the formation
of the adduct 14 (see Scheme 4). These reactions have been ex-
3
0e34
tensively studied for benzene and other aromatic compounds.
The Criegee ozonide 13 is more stable than the starting compounds
Ozonations were performed with a corona-discharge-based
ozone generator (Ozono Elettrica Internazionale s.r.l.) using
a 30 L/h oxygen gas steam. The diffusion of the gas in the liquid
phase was obtained through a Pasteur pipette 2 mm wide.
Reactions were monitored by thin-layer chromatography (TLC)
on commercially available precoated plates (silica gel 60 F254) and
the spots of products were visualized under UV light (eluent mix-
ꢀ
1
1
and O
and CH
calculated in vacuum with respect to the separate reactants, is
3
by about ꢀ19, ꢀ21, and ꢀ22 kcal mol in vacuum, CH
2 2
Cl ,
3
CN, respectively. The activation energy of this reaction,
ꢀ
1
equal to about 6 kcal mol , a value in accord with that calculated,
with a similar level of theory, for the ozonation of phenol in Ref. 44.
ꢀ
1
The activation energy decreases up to 1.6 and 0.3 kcal mol in
dichloromethane and acetonitrile, respectively. Notably, these
values are similar to those calculated for the H-abstraction at the
morpholine ring showing that the two mechanisms are kinetically
competitive, and that the partial recovery of the reaction products
can be explained by the decomposition of the aromatic ring.
The adduct 14 in vacuum is higher in energy than the separate
ture dichloromethane/acetate 8:2, 1 R
GCeMS analyses were performed with a Hewlett Packard 5890
f
¼0.74, 2 R
f
¼0.86, 3 R
f
0.40).
instrument equipped with a 5971A Mass Selective Detector. A
Supelco SPB-5 column (30 m, I.D. 0.25 mm, 0.25 mm film thickness)
ꢁ
ꢁ
ꢀ1
ꢁ
was used. Oven temp 50 C for 5 min, then 10 C min until 250 C
ꢁ
was reached. The temperature of 250 C was kept for additional
5 min. Infrared (IR) spectra in solutions were recorded on a Nicolet
Avatar 360 Fourier transform (FT)-IR spectrometer, using calcium
ꢀ
1
reactant by about 9.5 kcal mol . However, the solvent significantly
1
13
affects the stability of the adduct as it becomes more stable than the
fluoride cells previously purged with N
2
. H and C NMR were
. Chemical
d), relative to the internal
ꢀ1
separate reactants by about 5 and 8 kcal mol in dichloromethane
and acetonitrile, respectively. In this case we were not able to locate
the transition state, and we cannot compare the energetics of this
recorded on a Varian Mercury 400 instrument in CDCl
shifts are reported in parts per million (
standard of tetramethylsilane (TMS).
3
reaction with the attack of O
. Conclusions
This paper reports on the investigation of the reaction mecha-
nism and primary products of the N-phenylmorpholine oxidation
by ozone in different solvents, and at different temperatures. N-
Phenylmorpholine was used in order to study the regiochemistry of
the ozonation reaction with hetero-substituted carbons. The dif-
ferent secondary carbons are adjacent to nitrogen and oxygen
heteroatoms, allowing to explore the regioselectivity in the oxi-
3
at the morpholine ring.
4
.2. Representative experimental procedure for the oxidation
3
of N-phenylmorpholine with excess of ozone
N-Phenylmorpholine 1 of 5 mmol was dissolved in 250 mL of
solvent (dichloromethane or acetonitrile) and ozonized at the se-
lected temperature. After the selected time of reaction, the sus-
pension was filtered, evaporated under a nitrogen stream, and the
residue was separated by flash chromatography over silica gel
(R¼100) eluting with the mixture dichloromethane/ethyl
acetate 9:1.
dation by ozone of the carbon
the carbon to an ether oxygen.
In all of the conditions investigated two major reaction products
have been characterized; the lactam 2 and the diformylderivative
. The lactone 2 was not found, neither was the hemiacetal. The
a to an aromatic amino nitrogen, and
a
4.3. Representative experimental procedure for the oxida-
tions of N-phenylmorpholine with a limited amount of ozone
N
3
O
Ozone was bubbled for 15 min in 1 L of the appropriate solvent
at the selected temperature. The content of ozone was measured by
computational study of the reaction mechanism shows that the
minimum energy pathway leads to the formation of the hydro-
3
5
iodometric titration. N-Phenylmorpholine 1 in the amount for the
appropriate stoichiometry with ozone was then added to the so-
lution. After 24 h reaction, the suspension was filtered, evaporated
at reduced pressure, and the residue was separated by flash
trioxide intermediate B
group to nitrogen by O
carbon atom to nitrogen than
N
through the H-abstraction from the CH
. The preferential attack of O to the
to oxygen is due to the smaller
2
a
3
3
a
a

R. Suarez-Bertoa et al. / Tetrahedron 68 (2012) 8267e8275
8275
chromatography over silica gel (R¼100) eluting with the mixture
dichloromethane/ethyl acetate 9:1.
5. Rindone, B.; Saliu, F.; Suarez Bertoa, R. Ozone-Sci. Eng. 2008, 30, 165e171.
6. Brambilla, A.; Bolzacchini, E.; Orlandi, M.; Polesello, S.; Rindone, B. Water Res.
1997, 31, 1839e1846.
7. Brambilla, A.; Rindone, B.; Polesello, S.; Galassi, S.; Balestrini, R. Sci. Total En-
4
.3.1. N-Phenylmorpholinone (2). White powder: IR (CaF
2
):
viron. 1993, 132, 339e348.
8. Bolzacchini, E.; Brambilla, A.; Orlandi, M.; Polesello, S.; Rindone, B. Water Sci.
Technol. 1994, 30, 129e136.
. Brambilla, A.; Bolzacchini, E.; Orlandi, M.; Polesello, S.; Rindone, B. Life Chem.
Rep. 1995, 13, 79e84.
0. Poschenrieder, H.; Stachel, H.-D.; Eckl, E.; Jax, S.; Polborn, K.; Mayer, P. Helv.
Chim. Acta 2006, 89, 971e981.
ꢀ
1
þ
þ
þ
1
655 cm ; MS: m/z¼177 (M ),148 (M ꢀCHO),119 (M ꢀ2CHO),101
þ
1
(
M ꢀ2CHOꢀCH
.97 (t, J¼5.1 Hz, 2H), 4.28 (s, 2H), 7.10e7.20 (m, 5H); C NMR
100 MHz, CDCl ): 161.0, 129.0, 126.5, 124.0, 60.0, 43.5, 28.5.
3
); H NMR (400 MHz, CDCl
3
):
d
3.07 (t, J¼5.1 Hz, 2H),
9
13
3
(
1
3
d
1
1. van Lier, F. P.; Hesp, T. G. M.; van der Linde, L. M.; van der Weerdt, A. J. A.
Tetrahedron Lett. 1985, 26, 2109e2110.
4
.3.2. Compound 3. Yellow oil: IR (CaF
2
): 1725, 1675 cmꢀ1; MS: m/
þ
þ
þ
þ
12. Henbest, H. B.; Stratford, M. J. W. J. Chem. Soc. 1964, 711e714.
13. Bailey, P. S. Chem. Rev. 1958, 58, 925e1010.
z¼193 (M ), 165 (M ꢀCO), 147 (M ꢀHCOOH), 134 (M ꢀCOꢀCHO);
1
H NMR (400 MHz, CDCl
H), 7.05e7.40 (m, 5H), 7.91 (s, 1H), 8.34 (s, 1H); C NMR (100 MHz,
CDCl ): 161.0, 160.0, 141.0, 130.0, 126.5, 124.7, 60.3, 44.1.
3
):
d
4.03 (t, J¼5.1 Hz, 2H), 4.27 (t, J¼5.1 Hz,
1
4. Bailey, P. S.; Mitchard, D. A.; Khashab, A.-I. J. Org. Chem. 1968, 33, 2675e2680.
15. Munoz, F.; von Sonntag, C. J. Chem. Soc., Perkin Trans. 2 2000, 2029e2033.
6. Giamalva, D. H.; Church, D. F.; Pryor, W. A. J. Am. Chem. Soc. 1981, 108,
678e7681.
7. Nangia, P. S.; Benson, S. W. J. Am. Chem. Soc. 1980, 102, 3105e3115.
13
2
1
3
d
7
1
4
.4. Computational methods
18. Houben, W. Methoden der Organische Chemie; George Thieme: Stuttgart, New
York, NY, 1988; Vol. E13; p 1118.
1
2
9. Plesnicar, B.; Cerkovnik, J.; Tekavec, T.; Koller, J. Chem.dEur. J. 2000, 6, 809e819.
0. Plesnicar, B.; Cerkovnik, J.; Tuttle, T.; Kraka, E.; Cremer, D. J. Am. Chem. Soc.
002, 124, 11260e11261.
All calculations were performed in the framework of the Density
3
6
Functional Theory (DFT) by using the hybrid three parameters
2
3
7e40
B3LYP exchange and correlation functional,
and the valence
21. Wu, A.; Cremer, D.; Plesnicar, B. J. Am. Chem. Soc. 2003, 125, 9395e9402.
22. Tuttle, T.; Cremer, D.; Plesnicar, B.; Cerkovnik, J. J. Am. Chem. Soc. 2004, 126,
41
triple- basissetwith polarizationonall atoms cc-pVTZ. Stationary
z
16093e16104.
points of the energy hypersurface (minima and saddle points) were
located by means of energy gradient techniques using the quantum
2
3. Kolsaker, P.; Meth-Cohn, O. J. Chem. Soc., Chem. Commun. 1965, 423e425.
24. Meth-Cohn, O.; Kerr, G. H. J. Chem. Soc. 1971, 1369e1374.
25. Tulyaganov, S. R.; Khasanov, S. A. Uzb. Khim. Zh. 1966, 10, 32e35.
2
4
2
mechanical Gaussian 03 suite. Geometry optimizations to minima
and transition states were performed in vacuum as well as in the two
solvents employed in the experiments (dichloromethane and ace-
6. Thomas, C.; Berwe, M.; Straub, A. PCT Int. Appl., 2005, 16 pp; CODEN: PIXXD2
WO 2005026135 A1 20050324.
27. Munoz, F.; Mvula, E.; Braslavsky, S. E.; von Sonntag, C. J. Chem. Soc., Perkin Trans.
2001, 1109e1116.
8. Sarasa, J.; Cort eꢀ s, S.; Ormad, P.; Gracia, R.; Ovelleiro, J. L. Water Res. 2002, 36,
035e3044.
29. Preliminary calculations performed at the CASSCF(6,6)/cc-pVDZ level of theory
2
tonitrile). The effect of the solvents (dichloromethane:
acetonitrile:
¼35.688) was evaluated according to the Polarizable
Continuum Model (PCM) approach of Tomasi et al.
3
¼8.93;
2
3
3
4
3e46
as imple-
A
confirm this result; the intermediate Int is unstable with respect to the dis-
sociation of ozone or singlet dioxygen.
0. Hendrickx, M. F. A.; Vinckier, C. J. Phys. Chem. A 2003, 107, 7574e7580.
mented in Gaussian 03. Due to the nature of the reaction in-
vestigated, individual spheres on hydrogen atoms, with UFF radius,
wereusedin the definition of the polarizablecavity. Transition states
were characterized bya full vibrational analysis. Due to the nature of
the process, transition states for the hydrogen abstraction were
calculated using the unrestricted formalism. However, wave-
functions always converged to the restricted solution, suggesting an
ionic rather than a radical character of the TSs. The wavefunctions of
the transition states were also tested for an internal and external
stability47 and in all cases the wavefunctions resulted to be stable.
3
31. Ndassa, I. M.; Silvi, B.; Volatron, F. J. Phys. Chem. A 2010, 114, 12900e12906.
32. Mvula, E.; von Sonntag, C. Org. Biomol. Chem. 2003, 1, 1749e1756.
3
3. Mvula, E.; Naumov, S.; von Sonntag, C. Environ. Sci. Technol. 2009, 43,
275e6282.
4. Naumov, S.; von Sonntag, C. Ozone-Sci. Eng. 2010, 32, 61e65.
6
3
35. Gordon, S. A.; Chughtai, A. R.; Smith, D. M. Am. Lab. 2000, 32, 12e13.
3
6. Parr, R. G.; Yang, W. Density Functional Theory of Atoms and Molecules; Oxford
University Press: New York, NY, 1989.
37. Becke, A. D. Phys. Rev. A 1988, 38, 3098e3104.
38. Becke, A. D. J. Chem. Phys. 1992, 96, 2155e2160.
3
4
9. Becke, A. D. J. Chem. Phys. 1993, 98, 5648e5652.
0. Stevens, P. J.; Devlin, F. J.; Chablowski, C. F.; Frisch, M. J. J. Phys. Chem. 1994, 98,
Acknowledgements
11623e11627.
41. Dunning, T. H., Jr. J. Chem. Phys. 1989, 90, 1007e1023.
4
2. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.;
Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.;
Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.;
Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.;
Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao,
O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken,
V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A.
J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G.
A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.;
Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Fores-
man, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov,
B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.;
Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P.
M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian
03, Revision C.02; Gaussian: Wallingford, CT, 2004.
We thank our students Maria Giusy Cardinale, Francesco Conti,
Maria Giner Vano for their collaboration in the experimental work.
Supplementary data
Supplementary data associated with this article can be found in
the online version, at http://dx.doi.org/10.1016/j.tet.2012.07.055.
References and notes
1. Weissermel, K.; Arpe, H. J. Industrial Organic Chemistry; VCH: Weinheim, 1993.
2
3
. Bailey, P. S. Ozonation in Organic Chemistry; Academic: New York, NY,1981; p 255.
. Kochi, J.; Sheldon, R. Metal Catalysed Oxidations of Organic Compounds; Aca-
demic: New York, NY, 1981.
43. Canc eꢁ s, E.; Mennucci, B.; Tomasi, J. J. Chem. Phys. 1997, 107, 3032e3041.
44. Mennucci, B.; Canc eꢁ s, E.; Tomasi, J. J. Phys. Chem. B 1997, 101, 10506e10517.
45. Cossi, M.; Barone, V.; Cammi, R.; Tomasi, J. Chem. Phys. Lett. 1996, 255, 327e335.
46. Barone, V.; Cossi, M.; Tomasi, J. J. Comput. Chem. 1998, 19, 404e417.
47. Bauernschmitt, R.; Ahlrichs, R. J. Chem. Phys. 1996, 104, 9047e9052.
4
. Arends, I. W. C. E.; Mac Faul, P. A.; Snelgrove, D. W.; Wayer, D. D. M.; Ingold, K. U.
In Free Radicals in Biology and Environment; Minisci, F., Ed.; Kluver Academic:
New York, NY, 1997; pp 79e90.