EPR Kinetic Evidence for Radical Intermediacy in the Oxidation of Secondary Amines to Nitrones by [Wo(O2)2OCOC5H4N]- [Bu4N+]

Article abstract of DOI:10.1021/jp9940598

Oxidation reactions of N,N-benzylalkylamines by [WO(O₂)₂OCOC₅H₄N] ^-[Bu₄N⁺] to nitrones were kinetically studied by UV and EPR techniques. The reactions follow a second-order rate law and the rate-determining step is a simple bimolecular attack of amine onto the peroxide oxygen of the peroxometal complex, which leads to the formation of the corresponding hydroxylamine. However, EPR measurements and iterative procedures point out that the reaction occurs through the intermediacy of aminoxyl radicals, formed by oxidation of hydroxylamine generated in situ in the rate-determining step, and subsequent oxidations of these radicals to nitrones by the starting peroxo complex. It is suggested that the oxidation of hydroxylamine to aminoxyl radical as well as the oxidation of aminoxyl radical to nitrone occurs through an electron transfer step associated with a proton transfer.

Full text of DOI:10.1021/jp9940598

2710
J. Phys. Chem. A 2000, 104, 2710-2715
EPR Kinetic Evidence for Radical Intermediacy in the Oxidation of Secondary Amines to
-
+ †
Nitrones by [Wo(O2)2OCOC5H4N] [Bu4N ]
‡
,§
,⊥
,‡
‡
F. P. Ballistreri, R. Bianchini,* C. Pinzino,* G. A. Tomaselli,* and R. M. Toscano
Dipartimento di Scienze Chimiche, UniVersit a` di Catania, V. le A. Doria 6, I-95125 Catania, Italy;
Dipartimento di Chimica Organica “U. Schiff”, UniVersit a` di Firenze, Via G. Capponi 9, 50121 Florence,
Italy; and Istituto di Chimica Quantistica ed Energetica Molecolare, CNR, Via V. Alfieri 1, 56010 Ghezzano,
S. Giuliano Terme, Pisa, Italy
ReceiVed: NoVember 16, 1999
-
+
2 2 5 4 4
Oxidation reactions of N,N-benzylalkylamines by [WO(O ) OCOC H N] [Bu N ] to nitrones were kinetically
studied by UV and EPR techniques. The reactions follow a second-order rate law and the rate-determining
step is a simple bimolecular attack of amine onto the peroxide oxygen of the peroxometal complex, which
leads to the formation of the corresponding hydroxylamine. However, EPR measurements and iterative
procedures point out that the reaction occurs through the intermediacy of aminoxyl radicals, formed by oxidation
of hydroxylamine generated in situ in the rate-determining step, and subsequent oxidations of these radicals
to nitrones by the starting peroxo complex. It is suggested that the oxidation of hydroxylamine to aminoxyl
radical as well as the oxidation of aminoxyl radical to nitrone occurs through an electron transfer step associated
with a proton transfer.
Introduction
sponding hydroxylamine, which then is oxidized to nitrone in
a fast step. Nevertheless, the question concerning the nature of
The oxidations of amines to nitrones is a very appealing
reaction because of both the relevance of the metabolic fate of
these compounds in vivo and of the synthetic interest of the
reaction products. Important enzymes are involved in the
metabolic oxidation of amines.1a-c For example, flavin mono-
oxygenase and related compounds such as 5-ethyl-4a-hydro-
peroxyflavin oxidize secondary amines to nitrones via hydroxyl-
the fast steps following the nucleophilic attack of the amine on
the oxidant and the possible involvement of radical intermediacy
along the reaction coordinate has not been definitely addressed.
Here we report on the detection and kinetic fate of aminoxyl
free radical intermediates in the oxidation reactions of N,N-
benzylalkylamines with the anionic oxodiperoxo complex
-
+
[
WO(O2)2OCOC5H4N] [Bu4N ] (PIC-W). These nitroxide radi-
2
amines. The simulation of the activity of these enzymes using
cals are shown to be actual reaction intermediates, responsible
for the production of the final nitrone derivatives.
metal complexes is of large interest, potentially providing
mimetic methods for catalytic oxidation of the amines.
Recent work from our laboratory pointed out that Mo(VI)
and W(VI) oxodiperoxo complexes are able to oxidize secondary
amines to nitrones through the intermediacy of the corresponding
Results and Discussion
(i) UV Measurements. Model N,N-benzylalkylamines have
been reacted with PIC-W in chloroform, and reactions monitored
both by UV and by EPR techniques. The choice of PIC-W as
oxidant was determined by the observation that amine is not
able to enter the coordination sphere of this peroxo complex
3
hydroxylamines.
A kinetic study, concerning oxidation of N,N-benzylalkyl-
amines to the corresponding nitrones by both anionic and neutral
Mo(VI) and W(VI) oxodiperoxo complexes, indicated that the
oxidation reactions involve a rate-determining nucleophilic
attack of the amine onto the peroxide oxygen of the oxidant,
with a transition state in which N-O bond formation and O-O
bond cleavage occur in a concerted way through a Bartlett-
3
and therefore the reactivity scheme is more simplified. In fact,
in this case no coordination preequilibria steps involving the
amine and the oxidant should be taken into consideration and
the rate-limiting step is a simple bimolecular step in which the
amine attacks the peroxide oxygen and yields the hydroxylamine
3,4
type transition state (electrophilic oxygen transfer mechanism).
4
On the other hand, in the case of neutral oxidants a rapid
exchange of ligands occurs in a preequilibrium step. That is,
the nucleophile amine coordinates the metal center replacing
the original ligand HMPA and yields a new peroxo complex
bearing now amine molecules in the coordination sphere. This
complex formed in situ then undergoes a nucleophilic attack
on the peroxide oxygen by external amine molecules. In both
cases the nucleophilic attack is believed to yield the corre-
through a Bartlett-like transition state. N,N-Benzylisopropyl-
i
t
amine (Pr ), N,N-benzyltertbutylamine (Bu ), and N,N-dibenzyl-
amine (DBA) are converted quantitatively into nitrones by
PIC-W according to eq 1:
C H CH NHR + WO(O ) L f
6
5
2
2 2
+
-
C H CHdN (O )R + H O + WO + L (1)
6
5
2
3
*
Corresponding authors.
Dedicated to the memory of Prof. Fulvio Di Furia (1944-1997).
Universit a` di Catania.
Universit a` di Firenze. E-mail: bianchini@cesit1.unifi.it.
Istituto di Chimica Quantistica ed Energetica Molecolare, CNR.
where R ) -CH(CH3)2, -C(CH3)3, or -CH2C6H5. The nitrones
adsorb remarkably in the UV region, so that the kinetics of their
formation can be easily followed (see Experimental Section).
The kinetics were carried out under pseudo-first-order condi-
†
‡
§
⊥
1
0.1021/jp9940598 CCC: $19.00 © 2000 American Chemical Society
Published on Web 02/29/2000

Oxidation of Secondary Amines to Nitrones
J. Phys. Chem. A, Vol. 104, No. 12, 2000 2711
Figure 1. Typical kinetic plot of ln(A
at 40 °C; R ) 0.9995).
∞
- A
t
) vs time (oxidation of Pri
2
TABLE 1: Second-Order Rate Constants for the Reaction
of N,N-Benzylalkylamines with PIC-W in Chloroform
2
q
q
T
(°C)
10 k
2
∆H
∆S
-
1
(kcal mol-1)
9.93
(cal mol deg-1
)
-1
amine
(M s-1
)
Figure 2. Experimental (a, b) and computed (a′ , b′) EPR spectra
obtained from the reaction of PIC-W with Bu and Pr , respectively.
t
Bu
22
40
60
10
20
35
55
10
35
55
0.23
0.66
1.6
-23.6
t
i
t
Bu
t
Bu
i
Pr
0.51
0.94
1.5
8.59
10.5
-40.5
-34.6
i
Pr
i
Pr
i
Pr
4.5
DBA
DBA
DBA
0.35
1.3
4.7
tions, using higher than 10-fold excesses of the amine and
monitoring the reaction for at least two half-lives. Good straight
lines were always obtained upon plotting ln(A∞ - At)/ (A∞ -
A0) versus time (Figure 1).
The observed pseudo-first-order rate constants were reckoned
from the slopes of these plots by using a linear least-squares
computer program and were reproducible within (5%. In all
examined cases an overall second-order rate law is obeyed (eq
2
).
d[nitrone]/dt ) k [amine][PIC-W]
(2)
2
The pertinent observed second-order kinetic constants, which
are the average of at least three independent measurements, are
t
reported in Table 1. The rate values indicate that Bu is the
slowest amine. The highest kinetic constant value observed for
Pr is probably ascribable to the less relevant steric hindrance
Figure 3. Experimental (a) and computed (b) spectra for the reaction
of PIC-W and DBA, and experimental spectrum (c) for the reaction of
PIC-W with N,N-dibenzylhydroxylamine.
i
of the isopropyl substituent, during the nucleophilic attack of
the nitrogen onto the peroxide oxygen, as compared with that
of tert-butyl and benzyl substituents, respectively. Activation
enthalpies and entropies are also reported and are in agreement
with a simple bimolecular rate-determining step.
TABLE 2: Hyperfine Splitting Constants (G) of Aminoxyl
Radicals Obtained in the Oxidation Reactions of
N,N-Benzylalkylamines and of N,N-Dibenzylhydroxylamine
(
DBH) by PIC-W in Chloroforma
(ii) EPR Measurements. To check the eventual formation
H
benzyl
H
of intermediate free radical species, we monitored the oxidation
process by the EPR technique. Amine and oxidant solutions
were prepared in accurately degassed chloroform, under argon
atmosphere, in order to avoid any contamination from oxygen.
Then appropriate volumes of both solutions (kept at 210 K)
were mixed under dry argon atmosphere in an EPR tube. The
tube was thereafter positioned in the EPR cavity, still keeping
the temperature at 210 K. Then the temperature was allowed to
rise until the signals reported in Figures 2 and 3 appeared
substrate
g
AN
A
A
i-propyl
t
Bu
Pr
2.0057
2.0050
2.0062
2.0062
13.87
16.78
16.81
16.81
7.35 (2)
10.20 (2)
9.95 (4)
9.95
i
4.90 (1)
DBA
DBH
a
Measured on a Varian EPR instrument at 210°K.
spectra registered during the oxidation process (Figure 2, a,a′,
and b, b′; Figure 3, a and b).
(
normally at nearly 260 K). The observed signals were ascribed
In fact the computed hyperfine coupling constants, reported
in Table 2, are in agreement with the spectral parameters
reported in the literature for the aminoxyl radicals having the
to the presence of the aminoxyl radicals generated by the parent
amines, by matching the simulated spectra with the experimental

2712 J. Phys. Chem. A, Vol. 104, No. 12, 2000
Ballistreri et al.
expected structure.5,6 Therefore, we can reasonably conclude
that in all the investigated oxidation reactions of the starting
amines aminoxyl radical species are detected as intermediates.
In the case of DBA oxidation, the presence of the corre-
This scheme involves first the oxidation of the amine by PIC-W
leading to the formation of aminoxyl radicals through the
intermediacy of N-hydroxylamine derivatives. The intermediacy
of N-hydroxylamine derivatives along the reaction pathway has
been previously addressed, but here it receives further support,
due to the EPR observation that the same aminoxyl radical is
generated as an intermediate in the oxidation reactions by PIC-W
of both N,N-dibenzylamine and N,N-hydroxydibenzylamine,
respectively.
•
sponding aminoxyl radical (C6H5CH2)2NO was further con-
firmed by comparing the EPR spectrum registered during the
oxidation reaction of the amine by PIC-W with the spectra
registered, respectively, during the oxidation process of the
corresponding N,N-dibenzylhydroxylamine by PIC-W (Figure
3
c) and during the irradiation of N,N-dibenzylhydroxylamine
The formation of the N-hydroxy derivative due to the
nucleophilic attack of amine on the peroxide oxygen is the
slowest step (step 5), as suggested by the observed second-order
t
itself in the presence of Bu OOH.
In the latter case, the N,N-dibenzylaminoxyl radical is
generated as reported in the following reaction.
2
rate law. If we assume that the N-hydroxylamine derivative is
able to bind the metal by its hydroxyl group, its oxidation might
occur within the coordination sphere of the peroxo complex
and also the generation of aminoxyl radicals (step 6) might occur
there. A similar behavior is exhibited by alcohols in the
oxidation process to carbonyl compounds by PIC-W through
hν
t
•
(
C H CH ) NOH + Bu OOH
9
8 (C H CH ) NO (3)
6
5
2 2
6
5
2 2
7
8
This procedure was employed by Coppinger and Ingold and
the pertinent reactions involving the N,N-dibenzylaminoxyl
radical are represented by the following equations of Scheme
9
the probable intermediacy of ketyl radicals. In such a case the
1
:
aminoxyl radical might be itself oxidized to nitrone within the
coordination sphere of the oxidant (step 7) or it might be able
to escape to yield the corresponding nitrone through the steps
SCHEME 1
8
and 9 without the further involvement of the oxidant.
hν
t
t
•
Disproportionation of MO4 to MO5 and MO3 has already been
suggested in the oxidation reactions employing MoO5 and WO5
Bu -OOH
9
8 Bu O
(4a)
k4b
98
t
•
as oxidant agents to rationalize the kinetic observations that only
Bu O + (C H CH ) NOH
6
5
2 2
10,11
one oxidant species carries on the oxidation process.
Steps
t
•
Bu OH + (C H CH ) NO (4b)
8 and 9 are the same steps 4c and 4d reported in Scheme 1.
They represent, respectively, the equilibrium of formation of a
presumably dimeric species and the disproportionation step of
this dimer to nitrone and to N-hydroxy derivative.
6
5
2 2
k4c
•
•
2
(C H CH ) NO y z [(C H CH ) NO ]
(4c)
6
5
2 2
6
5
2 2
2
k-4c
k4d
The kinetic equations of Scheme 2 can be written as follows
(A ) amine, O ) oxidant, H ) N-hydroxylamine, R ) radical,
D ) dimer, and N ) nitrone): (which is the rate-limiting step)
•
[
(C H CH ) NO ] 98
6 5 2 2 2
+
-
C H CHdN (O )CH C H + (C H CH ) NOH (4d)
6
5
2
6
5
6
5
2 2
k5
98 H
The pertinent EPR spectra, shown in Figure 3a,c, appear
overlapping and match perfectly with the simulated spectrum
A + O
H + O
(5)
(6)
5,6
(
Figure 3b) (spectral parameters are reported in the literature).
k6
9
8 R
(iii) Kinetic Approach. We followed both the nitrone
formation by UV technique (Table 1) and the aminoxyl radicals
build up and decay by EPR spectroscopy. In order to check
whether the observed aminoxyl radicals are intermediates in the
oxidation process of amines to nitrones, we employed a fitting
iterative procedure, based on Scheme 2:
k7
98 N
R + O
(7)
k8
R + R
9
8 D
8 2R
8 N + H
(8a)
SCHEME 2
k
-8
D
(8b)
(9)
k5
WO(O ) + ArCH NHR
9
8 ArCH N(OH)R + WO (O )
2
2
2
2
2
+5
2
k9
+6
D
9
(5)
k6
98
Reaction 5, which leads to the formation of N,N-hydroxylamine,
is transparent with respect to the EPR technique. Therefore, with
reference to the process monitored by EPR, we can rewrite the
two steps 5 and 6 as a unique step:
WO(O ) + ArCH N(OH)R
2
2
2
+5
•
+
ArCH N(O )R + WO(O ) + H (6)
2
2 2
k7
98
k5
•
+
WO(O ) + ArCH N(O )R + H
2
2
2
A + O 98 R
+
-
ArCHdN (O )R + WO (O ) + H O (7)
2
2
2
The relative differential equations, describing the evolution of
concentrations with time, are reported as follows:
2
WO (O ) f WO(O ) + WO
2
2
2 2
3
k8
•
•
d[A]
dt
2
ArCH N(O )R y z [ArCH N(O )R]
(8)
ArCH N(O )R] 98 ArCHdN (O )R + ArCH N(OH)R
2 2
-
) k [A][O]
(10)
(11)
2
2
2
5
k-8
k9
•
+
-
d[O]
[
2
-
) k [A][O] + k [R][O]
5
7
dt
(9)

Oxidation of Secondary Amines to Nitrones
J. Phys. Chem. A, Vol. 104, No. 12, 2000 2713
d[R]
dt
2
)
k [A][O] - k [R][O] - 2k [R] + k [D]
(12)
(13)
(14)
5
7
8
-8
d[N]
dt
)
k [R][O] + k [D]
7
9
d[D]
dt
2
)
2k [R] - k [D] - k [D]
8 -8 9
In the fitting procedure the program took into account both
the EPR and the UV spectra, registered at the same temperatures
with the same starting reagent concentrations, and generated
by experimental measurements performed at identical time
intervals and for a coincident time duration (Figure 4) (see
Experimental Section). In each fitting we tried to simulate the
two different spectral curves. In order to check the involvement
of the aminoxyl free radical along the reaction coordinate, we
looked for a correlation between the nitroxide radical build up
and decay process and the nitrone formation. Therefore, the
starting parameters were used to fit the EPR curve, then the
fitted values found were employed to fit the UV absorption/
time curve, and then again back to the EPR, in an iterative
process whose final convergence consisted of an acceptably
good reproduction of both curves.12 The final values of the fitted
kinetic constants are shown in Table 3.
Figure 4. EPR and UV (335 nm) kinetics for the reaction of PIC-W
-
3
t
-2
(
1.2 × 10 M) with Bu (2.0 × 10 M) in chloroform at 60 °C.
2
Absorption data have been multiplied by a factor of 10 to be in a
comparable scale with EPR curve. Full lines are experimental curves,
crosses are computed EPR, and triangles computed absorptions.
TABLE 3: Kinetic Rate Constants Obtained by the Fitting
Procedure for the Oxidation Reactions of
N,N-Benzylalkylamines with PIC-W in Chloroform at 40 °C
2
2
2
PIC-W
(M)
10 k
5
10 k
7
10 k
8
amine (M)
(M s-1
-
1
)
(M s-1
-1
)
(M s-1
)
-1
The rate constant values reported in Table 3 indicate that the
agreement between the k5 value obtained on the basis of EPR
data and the k2 value (shown in Table 1), obtained by UV
measurements, is quite good, supporting therefore the idea that
the observed second-order rate constant k2, describing the attack
of the amine onto the peroxide oxygen to yield the correspond-
ing hydroxylamine, is indeed the rate-determining step. In fact,
by stopped flow technique we were able to measure the N,N-
dibenzylhydroxylamine disappearance rate constant k6 (eq 6)
in the presence of PIC-W in chloroform at 40 °C. The observed
i
(0.018) 3.0 ( 1.2a
Pr (0.012)
9.1 ( 2.6 3.3 ( 1.1
20 ( 6.0 3.7 ( 1.1
5.0 ( 1.6 1.7 ( 0.5
2.5 ( 0.5 0.76 ( 0.16
11 ( 2.0 4.1 ( 1.0
DBA (0.020) (0.001) 2.5 ( 0.6
Bu (0.021)
Bu (0.021)
Bu (0.020)
t
(0.001) 0.66 ( 0.21
(0.001) 0.28 ( 0.12
(0.001) 1.6 ( 0.5
t
b
c
t
a
Standard deviation. b T ) 22 °C. c T ) 60 °C.
-
1
-1
value k6 ) 12 ( 2 M
.6) × 10
s
is much larger than k5 ) (2.5 (
-
2
-1 -1
0
M
s
and then we can conclude that the rate of
formation of N,N-dibenzylaminoxyl radicals is controlled by the
k5 value. Therefore, on this basis, it would be expected, and
indeed it is found, that the formation rate k5 of aminoxyl radicals
measured by EPR should correspond, at least within the
experimental errors, to the k2 value determined by UV measure-
ments. The agreement among k5 values obtained by EPR
technique and k2 values measured by UV is good also for the
i
t
other two amines, Pr and Bu . Finally, the above approach
allowed a good stability of the fitting procedure, since even
considerably different starting guesses (3 times larger, or
smaller) of the fitting parameters with respect to the final fitting
results converged to the same final results. Only when these
guesses were orders of magnitude far from reasonable values
the fitting diverged. But in any case no different final fitting
values were obtained for the fitting parameters from those
reported in the table, the only alternative being just a total
divergence of the fitting procedure.
Figure 5. Experimental (line) and computed (symbols) kinetics for
the reaction of N,N-dibenzylhydroxylamine (0.1 M) with tert-butyl
hydroperoxide in chloroform under irradiation.
Figure 5 shows the buildup and decay, monitored by EPR,
of N,N-dibenzylaminoxyl radicals obtained by UV irradiation
of the corresponding N,N-dibenzylhydroxylamine (eq 3, Scheme
1
).
The curve is composed of two distinct sections, the first one
However, the values of k-8 (step 8b, radical dimerization
process) and k9 (step 9, nitrone formation from the dimer)
obtained by the fitting procedure are not displayed in Table 3,
consisting of a very fast buildup and a rapid decay of the
•
aminoxyl radical (C6H5CH2)2NO down to 10% of its initial
concentration, and the second one representing a signal which
decreases very slowly and then persists for many hours. With
reference to Scheme 1 the following differential equations can
be taken into account:
because they are very low and much smaller than standard
deviation values.12
In order to check whether the very small values obtained for
k-8 and k9 is a pitfall of the iterative procedure or they are
intrinsically so low, we tried to obtain them experimentally by
independent measurements.
2
d[R]/dt ) 2k-4c[D] - 2k [R]
(15)
4c

2714 J. Phys. Chem. A, Vol. 104, No. 12, 2000
Ballistreri et al.
2
SCHEME 4
d[D]/dt ) k [R] - k [D] - k [D]
(16)
(17)
4c
-4c
4d
d[N]/dt ) k [D]
4d
Also in this case the iterative procedure, based on Scheme 1,
provided a very good agreement with the experimental observa-
tions. The convergence of the fitting was reached with a very
low variance of the computed versus experimental curve (Figure
4
1
(
), and the computed kinetic values were k4c ) (3.6 ( 0.03) ×
-2 -1 -1 -5 -1
0
M
s
, k-4c ) (4.6 ( 0.03) × 10 s , and k4d ) (1.3
-
4
-1
0.01) × 10 s . Indeed it is very interesting to observe
that the rate constant value concerning the dimerization process
k4c is, within the experimental error, equal to the rate constant
k8 , reported in Table 3, which represents in Scheme 2 the same
dimerization process. Therefore, since two experiments, per-
formed in different ways, lead to the same rate constant value
for the same reaction step, we can reasonably conclude that the
process reported in Scheme 1 is a good model also for steps
k-8 and k9. In such a case the value of k-8 would correspond to
that of k-4c (4.6 × 10 ). This observed k-8 value is very low
and in agreement with the computed value of the iterative
procedure, which is itself low and smaller than the standard
deviation value. Likewise, the value of k9 would correspond to
In our case, coordination of the N-hydroxylamine derivative to
the metal by the hydroxylic group would lead to the formation
of an association complex, which by two subsequent electron
transfer steps (associated with two proton transfers) might evolve
to nitrone, through the intermediacy of aminoxyl radicals
(Scheme 4). The first ET step might follow an association
complex formation between the oxidant and the hydroxylamine
and therefore it might occur within the coordination sphere of
the metal. Both potential reduction values and literature
examples indicate that Mo(VI) and W(VI) are very often eager
-
5
14-16
to be involved as acceptors in electron transfer steps.
-
4
that of k4d (1.3 × 10 ) which is small too and therefore also
this experimental value is according to the expectation of the
iterative procedure. In conclusion these results suggest that the
iterative procedure provides a reliable outcome. Therefore, the
two reaction steps, (8b) and (9), indeed are so slow that their
significance in the overall kinetic picture of Scheme 2 is
irrelevant and the relative rate constant values, k-8 and k9,
obtained by the iterative procedure are not meaningful param-
eters (they are almost 2 or 3 orders of magnitude lower than
the other rate constants reported in Table 3). The rate constant
values, related to the oxidation process of DBA with PIC-W
and displayed in Table 3, show that processes of type (8b) and
Experimental Section
Materials. Fluka chloroform containing 1% ethanol was used
as such. N,N-Benzylalkylamines (Aldrich) were purified by
distillation over calcium hydride. PIC-W was prepared by the
8
previously reported original procedure.
UV-Vis Spectroscopic Measurements. Chloroform solu-
tions of amine and oxidant were independently prepared. The
chloroform was degassed, and the overall solutions were kept
under argon atmosphere. Then equal volumes of the reagents
were mixed in the cell, and the formation of the nitrone was
followed in a Cary 220 instrument at the selected wavelength,
using a Haake thermostat. At the examined wavelength (320-
(9) can eventually affect the concentration of the final product,
or that of the aminoxyl radical, only after a considerable amount
of time, by far longer than that usually required to accomplish
the oxidation reactions under observation.
3
40 nm), molar absorption of 1 was around 10, while those of
2
nitrones, the unique reaction product, were around 5 × 10 to
3
-1
-1
2
× 10 M cm . The kinetics were found linear, on plotting
SCHEME 3
ln(A∞ - At) versus time, for at least two half-lives. But generally
they were linear for more than 90% conversion (Figure 1). In
the case of DBH (N,N-dibenzylhydroxylamine) the kinetic
experiments were carried out in a diode-array stopped flow
k5 ) 2.5 × 10-2
k7 ) 20 × 10-2
DBA + PIC-W
8 R + O
8 N
k8 ) 3.7 × 10-
2
k9 ) 1.3 × 10 -4
17
instrument monitoring the UV interval, and using the same
R + R y
5z D
8 N
k-8 ) 4.6 × 10-
mathematical approach to make linear absorption/time data so
obtained.
In fact, considering that the maximum value of concentration
EPR Measurements. EPR spectra and the time evolution of
the amplitude of the highest line in each spectrum (EPR kinetics)
were taken from a Varian E 112 EPR spectrometer. The
temperature of the samples was controlled by an Oxford EPR
900 cryostat. The spectrometer was interfaced to a 486/100 PC
computer through a data acquisition system able to give in real
3
of free radicals R is 1/1000 of the [O]0, a rate7/rate8 = 5 × 10
ratio can be estimated. This value indicates that only when the
oxidant is completely reacted does step 6 supply free radicals
R to the equilibrium 8 (but this situation involves reaction times
out of the present investigation). Thus, it turns out that step 8
yields an amount of nitrone which can be considered insignifi-
cant and that step 7 is the unique pathway which leads to the
nitrone formation. In other words, the observed nitrone is formed
only by oxidation of aminoxyl radicals by the peroxometal
complex.
1
8
time the average signal. The software employed was able to
operate simultaneously the EPR spectrometer, the temperature
1
9
controller, and the data acquisition system. The hyperfine
coupling constants and line widths of the paramagnetic species
were the result of the best-fitting between experimental and
simulated spectra. The spectral optimization software was
We do not have any piece of information on how aminoxyl
radicals are oxidized by PIC-W. However, we can speculate
that such a process might bear closely to the mechanistic scheme
suggested for alcohol oxidations to carbonyl compounds by
1
9
provided by D. Duling, NIEHS, NIC.
Reaction Mixture Preparation. In a quartz EPR tube,
equipped with a Rotaflo stopcock, 25-45 mL of the oxidant
solution was introduced under dry argon; after freezing by
immersion of the EPR tube in a dry ice-acetone refrigerant
bath, an equivalent volume of the amine solution was added,
9
Mo(VI) peroxo complexes, which envisages the association of
the alcohol to the peroxo complex, followed by two electron
transfers to the oxidant from the coordinate alkoxo compound.

Oxidation of Secondary Amines to Nitrones
J. Phys. Chem. A, Vol. 104, No. 12, 2000 2715
taking care that it solidified before reaching the surface of the
oxidant frozen solution. The overall frozen sample was rapidly
transferred into the EPR cavity, where the temperature was
allowed to increase slowly by a heating controller up to the
fusion temperature of the two frozen phases, thus allowing
oxidant and amine to interact each other. The Borning spectrum
was then registered, and the reaction process monitored by a
fast acquisition of EPR data for each programmed temperature
for the time requested by the evolution of the reaction.
In the best fitting procedure of the kinetic data, the fitting
program took into account alternatively the EPR and the UV-
vis spectroscopic curves, both registered with the same starting
reagent concentrations, at the same temperatures, and with
experimental points registered at identical time intervals and
for a coincident time duration. The program fits the two curves,
alternatively, and the computed parameters were the starting
fitting parameters for the computing process of the successive
kinetic curve, until a final, overall convergence was obtained.
Fixed parameters were, of course, the initial reagent concentra-
tions and their molar absorption at the wavelength of the
spectroscopic curve. Fitting parameters were instead the kinetic
constants reported in Table 3 for each run. Even in the case of
EPR kinetics, two half-lives (upon consideration of the product
formation) were considered, at least.
Toscano, R. M. Tetrahedron 1992, 48, 8677-8684. (c) Murahashi, S. I.;
Mitsui, H.; Tatsuki, S.; Tsudo, T.; Watanabe, S. J. Org. Chem. 1990, 55,
1736-1744.
(2) Ball, S.; Bruice, T. C. J. Am. Chem. Soc. 1980, 102, 6498.
(3) Ballistreri, F. P.; Barbuzzi, E. G.; Tomaselli, G.; Toscano, R. M.
J. Org. Chem. 1996, 61, 6381-6387.
(4) Bartlett, P. D. Record. Chem. Prog. 1950, 11, 47-56.
(5) Coppinger, G. M.; Swalen, J. D. J. Am. Chem. Soc. 1961, 83, 4900-
902.
4
(
(
6) Cowley, D. J.; Waters, W. A. J. Chem. Soc. (B) 1970, 96-101.
7) De La Mare, H. E.; Coppinger, G. M. J. Am. Chem. Soc. 1963, 28,
1
068-1070.
8) Adamic, K.; Bowman, D. F.; Gillan, T.; Ingold, K. J. Am. Chem.
(
Soc. 1971, 93, 902-908.
(9) Campestrini, S.; Di Furia, F. Tetrahedron 1994, 5119-5131.
(10) Bortolini, O.; Di Furia, F.; Modena, G.; Scardellato, C. J. Mol.
Catal. 1981, 11, 107-118.
(11) Amato, G.; Arcoria, A.; Ballistreri, F. P.; Tomaselli, G. A.; Bortolini,
O.; Conte, V.; Di Furia, F.; Modena, G.; Valle, G. J. Mol. Catal. 1986, 37,
165-175.
(12) To evaluate the goodness of the fitting we considered the ratio,
0
expressed in percentage, between the variance of the fitting s and the
0
1/2
experimental squared maximum value (s ) 100(S/n) , where S is the sum
of the squared residuals and n is the number of data points minus the number
of fitting parameters); see: Ambrosetti, R.; Bianchini, R; Bellucci, G. J.
Phys. Chem. 1986, 90, 6261-6266. The minimization of such parameter
was carried out by least-squares methods. We assumed that convergence
would occur for maximum 5% deviation.
2
-1 -1
(
13) The following are reported: for instance, 10 k9 (M
s ) values
Besides the kinetic constants, reported in Table 3, the output
of the fitting program included the standard deviations and the
correlation coefficients between each possible couple of fitting
parameters. These last values never approached too much to
the value of 1. The evaluation of the amount of free spins was
performed using a Varian standard pitch.
i
obtained by the fitting procedure (amine, temp): 0.005 ( 2 (Pr , 40 °C);
0.0067 ( 0.85 (DBA, 40 °C); 0.0045 ( 5 (Bu , 22 °C); 0.01 ( 0.02 (Bu ,
t
t
t
4
0 °C); 0.052 ( 0.62 (Bu , 60 °C).
14) Bonchio, M.; Conte, V.; Di Furia, F.; Modena, G.; Moro, S.;
Carofiglio, T.; Magno, F.; Bonchio, M.; Pastore, P. Inorg. Chem. 1993,
2, 5797-5799.
15) Campestrini, S.; Di Furia, F.; Modena, G.; Bortolini, O. O. J. Org.
(
3
(
Chem. 1990, 55, 3658-3660.
Acknowledgment. Financial support from CNR and from
MURST is gratefully acknowledged.
(16) Campestrini, S.; Di Furia, F.; Novello, F. J. Mol. Catal. 1993, 78,
1
59-168.
17) Ambrosetti, R.; Bellucci, G.; Bianchini, R.; Fontana, E.; Ricci, D.
J. Phys. Chem. 1994, 98, 1620-1625.
18) Ambrosetti, R.; Ricci, D. ReV. Sci. Instrum. 1991, 62, 2281-2283.
(19) Pinzino, C.; Forte, C. ESR-ENDOR, ICQEM-CNR, Pisa, Italy,
1992.
(
References and Notes
(
(
1) (a) Murahashi, S. I.; Angew. Chem., Int. Ed. Engl. 1995, 34, 2443-
2
465. (b) Ballistreri, F. P.; Chiacchio, U.; Rescifina, A.; Tomaselli, G.;