Spin trapping chemistry of iminyl free radicals

Article abstract of DOI:10.1002/(sici)1097-458x(199702)35:2<131::aid-omr23>3.0.co;2-3

The iminyl radicals formed from hydrogen atom abstraction between tert-butoxyl radicals and benzylidene-N-alkyl-or N-arylamines were trapped by 2-methyl-2-nitrosopropane and investigated by EPR spectroscopy. The compounds investigated were benzylidene N-methyl, ethyl, 1-propyl, 1-butyl, 2-methylpropyl, 1-methylethyl, 1-methylpropyl, 1-ethylpropyl, 1-methylbutyl and cyclohexyl derivative and also benzylidene N-phenyl, 4-tolyl, 4-fluorophenyl, 4-methoxyphenyl, 4-chlorophenyl, 4-nitrophenyl and 4-trifluoromethylphenyl derivatives. In every case the iminyl nitroxide (aminoxyl) was produced in benzene at room temperature. The nitrogen hyperfine splitting constants were in the ranges 3.39-3.56 and 9.68-9.77 G for the iminyl and nitroxyl nitrogens, respectively, for the benzylidene-N-alkylamines and 3.60-3.77 and 8.45-9.15 G for the iminyl and nitroxyl nitrogens, respectively, for the benzylidene-N-arylamines. Very little evidence was found for hydrogen atom abstraction from the alkyl groups attached to the imine function. The absolute rate constant for hydrogen atom abstraction of the iminyl hydrogen was estimated to be 1.2 × 10⁴ M^-1 s^-1 based on competitive experiments with addition of tert-butoxyl radicals to 2-methyl-2-nitrosopropane (1.5 × 10⁶ M^-1 s^-1). This value is considerably slower than that for benzaldehyde (2.4 × 10⁷ M^-1 s^-1).

Full text of DOI:10.1002/(sici)1097-458x(199702)35:2<131::aid-omr23>3.0.co;2-3

MAGNETIC RESONANCE IN CHEMISTRY, VOL. 35, 131È140 (1997)
Spin Trapping Chemistry of Iminyl Free Radicals
Edward G. Janzen* and Dale E. Nutter¤
Departments of Clinical Studies and Biomedical Sciences, The MRI Facility, Ontario Veterinary College, University of
Guelph, Guelph, Ontario N1G 2W1, Canada
The iminyl radicals formed from hydrogen atom abstraction between tert-butoxyl radicals and benzylidene-N-
alkyl- or N-arylamines were trapped by 2-methyl-2-nitrosopropane and investigated by EPR spectroscopy. The
compounds investigated were benzylidene N-methyl, ethyl, 1-propyl, 1-butyl, 2-methylpropyl, 1-methylethyl, 1-
methylpropyl, 1-ethylpropyl, 1-methylbutyl and cyclohexyl derivative and also benzylidene N-phenyl, 4-tolyl, 4-
Ñuorophenyl, 4-methoxyphenyl, 4-chlorophenyl, 4-nitrophenyl and 4-triÑuoromethylphenyl derivatives. In every
case the iminyl nitroxide (aminoxyl) was produced in benzene at room temperature. The nitrogen hyperÐne split-
ting constants were in the ranges 3.39–3.56 and 9.68–9.77 G for the iminyl and nitroxyl nitrogens, respectively, for
the benzylidene-N-alkylamines and 3.60–3.77 and 8.45–9.15 G for the iminyl and nitroxyl nitrogens, respectively,
for the benzylidene-N-arylamines. Very little evidence was found for hydrogen atom abstraction from the alkyl
groups attached to the imine function. The absolute rate constant for hydrogen atom abstraction of the iminyl
hydrogen was estimated to be 1.2 Â 104 M—1 s—1 based on competitive experiments with addition of tert-butoxyl
radicals to 2-methyl-2-nitrosopropane (1.5 Â 106 M—1 s—1). This value is considerably slower than that for benz-
aldehyde (2.4 Â 107 M—1 s—1). ( 1997 by John Wiley & Sons, Ltd.
Magn. Reson. Chem. 35, 131È140 (1997) No. of Figures: 3 No. of Tables: 3 No. of References: 24
Keywords: ESR; EPR; spin trapping; 2-methyl-2-nitrosopropane; free radicals; iminyl radicals; acyl radicals
Received 26 March 1996; accepted 10 August 1996
This is because the acyl group does not allow transfer of
INTRODUCTION
spin from the nitroxyl function past the carbonyl
carbon (a few exceptions exist, however4). The iminyl
radical spin adduct spectrum of a nitroso spin trap
would be expected to show more lines because the spin
is delocalized on to the second iminyl nitrogen atom:
It is well known that free radical acyl hydrogen
abstraction is readily accomplished by various radicals
to produce acyl radicals:1
Æ
RwCHO ] XÆ ] RwCxO ] HX
(1)
OÆ
=
Æ
The question of interest in this paper is whether imines
derived from aldehydes undergo the same reaction to
form iminyl radicals:2
R@wCxNwRA ] RÓwNxO ] R@wCwNwRÓ
(4)
>
RAwN
Æ
O~
O~
R@wCHxNwRA ] XÆ ] R@wCxNwRA ] HX (2)
=
=
R@wCw~NwRÓ % R@wCxNwRÓ
We sought to answer this question by spin trapping. A
nitroso spin trap was selected to intercept the iminyl
radical so as to produce an EPR distinctive spin adduct
in each case.
When acyl are trapped the acyl nitroxide (aminoxyl)
spin adduct spectrum usually displays only three lines
with about 8 G spacing:3
>
]
=
]
RAwN
RAwNÆ
These iminyl nitroxides (or aminoxyls) have been
detected before by EPR spectroscopy.5 For example,
the following amidine N-oxide was prepared syntheti-
cally and oxidized with a mild oxidizing agent:
O~
OÆ
=
=
Æ
K3Fe(CN)6
C H wCxNwC H ÈÈÈÈ
RwCxO ] RÓwNxO ] RwCwNwRÓ
(3)
6 5
4 9
benzene
=
`
>
C H wNH
O
6 5
O~
=
C H wCw~NwC H (5)
* Correspondence to: E. G. Janzen.
¤ Work performed in the Department of Chemistry, University of
Georgia, Athens, GA, USA.
6 5
4 9
>
`
C H wN
6 5
In this case three 1 : 1 : 1 triplets could be resolved
which were assigned to two di†erent nitrogen hyperÐne
Contract grant sponsor: Natural Sciences and Engineering Research
Council of Canada.
( 1997 by John Wiley & Sons, Ltd.
CCC 0749È1581/97/020131È10 $17.50

132
E. G. JANZEN AND D. E. NUTTER
splitting constants (N-HFSCs), plus a complex pattern
of multiplets believed to come from the N-aryl protons.
The larger N-HFSC was assigned to the nitroxyl nitro-
equivalent b-H and b-N HFSCs,12 whereas route (8)
should produce nitrogen-centered adduct with
perhaps a very small splitting from the c-hydrogen13
(see Experimental section for deÐnition of b- and c-
positions):
a
gen (aN \ 8.92 G) and the smaller to the iminyl nitro-
NO
gen (aN \ 3.87 G). The remaining HFSCs were
NC
attributed to aryl hydrogens assumed to be all approx-
RA
OÆ
imately magnetically equivalent (aH
o/p/m
however, see Results section and Table 3).
\ 0.45 G;
=
=
C H OwNwCHwNwC H
4 9
4 9
Cyclic iminyl aminoxyls also give similar N-HFSCs,
e.g.
=
R@
aN B 15 G; aN \ aH B 3 G
NO
(estimated, not detected)
b
b
R@
OÆ
=
=
C H OwCHwNwNwC H
4 9
4 9
=
aN \ 9.10; aN \ 4.37 G6
NO NC
in benzene
aN \ 10.0 G7 plus other HFS
RA
NO
in water
aN B 17 G; aN \ 2 G
NO
(estimated, not detected)
a
We chose to use tert-butoxyl radicals as the hydrogen
atom-abstracting radical using di-tert-butyl per-
oxyoxalate (DBPO) as the room-temperature thermal
radical source:8
In cases where R@ or RA are aliphatic groups, addi-
tional sites for hydrogen atom abstraction might be
present and spin adducts from these avenues of radical
chemistry might be expected. In favourable cases it
might be possible to estimate relative reactivities of dif-
ferent hydrogens in the same molecule. This kind of
information might be very helpful to biological chemists
struggling with identiÐcation of possible products
resulting from reactions of bioradicals produced in dis-
eased tissue.
OO
> >
C H OOCCOOC H ] 2C H O Æ ] 2CO (6)
4 9
4 9
4 9
2
Iminyl radicals have been produced by this method
before except that photolysis of tert-butyl peroxide at
[100 ¡C in cyclopropane was the radical-generating
system.9
The nitroso spin trap of choice in hydrocarbon sol-
vents is 2-methyl-2-nitrosopropane (MNP). In this
study one could expect to generate iminyl spin adducts
of MNP with characteristic EPR spectra consisting of
1 : 1 : 1 nitroxyl triplets split into 1 : 1 : 1 iminyl triplets
with substantially di†erent magnitudes for the HFSCs
(as discussed above). In addition, tert-butoxyl radicals
would add to MNP to give tert-butoxyl-tert-butyl-
aminoxyl.10 This spin adduct has a very large N-HFSC
EXPERIMENTAL
Chemicals
The imines were prepared in general by using benzaldehyde (or a sub-
stituted benzaldehyde) and the primary amine or aniline (or substi-
tuted aniline). The imine was vacuum distilled if
(aN B 27 G) and the outer lines would not overlap with
a liquid or
NO
recrystallized from diethyl ether if a solid. The imines were stored in a
vacuum desiccator over drying agent at 0.5 mmHg and kept in a
refrigerator over nitrogen gas. A general method of preparation is
given below for the synthesis of C-phenyl-tert-butylimine or N-
benzylidene-tert-butylamine. The boiling points and/or melting points
for all the imines synthesized are collected in Table 1.
For the preparation of N-benzylidnen-tert-butylamine, freshly dis-
tilled benzaldehyde (44.5 g) was added to 30.0 g of redistilled tert-
butylamine at 2È3 ¡C with stirring. The stirring was continued for 30
min and the reaction mixture allowed to stand for 2 days over few
grams of NaOH. The aqueous layer was removed and extracted with
diethyl ether. The ether extract was added to the water-insoluble
material and dried over KOH. After removal of the ether, the product
was distilled to give a 60 g yield (90%), b.p. 92 ¡C/8 mmHg.
the expected spectra. Also present perhaps as impurities
would be di-tert-butylaminoxyl (DTBA) and acyl-tert-
butylaminoxyls:11
OÆ
OÆ
OÆ
=
=
=
C H ONC H C H NC H RCNC H
4 9
4 9
4 9
4 9
4 9
>
O
aN \ 27 G
aN \ 15.5 G aN \ 8.0 G
NO
NO
NO
An additional question of interest in this study was
the possibility of radical addition to the imine function:
MNP was prepared from the oxidation of tert-butylamine to make
2-methyl-2-nitropropane followed by reduction to the hydroxylamine
and oxidation to the nitroso dimer.15,16
A 325 g amount of KMnO was stirred into 1.5 l of water and 71.4
4
g of freshly distilled tert-butylamine were added over the course of 15
min. The reaction Ñask was cooled to 11 ¡C during the addition. Note
that without adequate cooling the reaction mixture can “explodeÏ with
messy consequences. The contents were stirred and allowed to come
to room temperature for 8 h, then heated to 55 ¡C and stirred for 18 h.
The mixture was subsequently steam distilled directly from the reac-
tion Ñask (b.p. 89 ¡C) and the distillate extracted with four 30 ml por-
tions of diethyl ether. The ether extract was washed with 30 ml of
The spin adduct of these addition products would give
di†erent EPR spectra with MNP. Reaction (7) should
produce a carbon-centered adduct with approximately
10% HCl and 30 ml of water and dried over anhydrous NaSO . The
4
yield of 2-methyl-2-nitrosopropane was 55.3 g (78%) after removal of
the ether.
( 1997 by John Wiley & Sons, Ltd.
MAGNETIC RESONANCE IN CHEMISTRY, VOL. 35, 131È140 (1997)

SPIN TRAPPING CHEMISTRY OF IMINYL FREE RADICALS
133
Table 1. Boiling and melting points of C-phenyl-N-alkyl and N-phenylimines
RºCH = NR/a
R¾
RÂ
Boiling point (¡C/mmHg)
Melting point (¡C)
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Methyl
Ethyl
60/6
61–62/4
80/5
100/6.5
85.5/5.9
88/8
95–96/5
97–98/8
96–97/8
137/10
1-Propyl
1-Butyl
2-Methylpropyl
2-Propyl
2-Butyl
2-pentyl
3-Pentyl
Cyclohexyl
tert-Butyl
92/8
tert-Octyl
99–100/1.7
130–133/1.7
136–138/1.5
153–155/10
Benzyl
a-Methylbenzyl
Phenyl
52–53
58–59
52–54
71–72.5
30–32
51–54
94–97
60.5–62
41–45
42–45
4-Fluorophenyl
4-Chlorophenyl
4-Methoxyphenyl
4-Tolyl
165–166/9
4-Trifluoromethylphenyl
4-Nitrophenyl
Phenyl
4-Methoxyphenyl
4-Tolyl
Phenyl
4-Nitrophenyl
Phenyl
a Taken from D. E. Nutter, Jr, Two applications of ESR spin trapping investigation of reac-
tion of t-butoxy radicals with benzaldimines and photolysis of alcohol/nitrone solutions,
PhD Dissertation, University of Georgia, Athens, GA (1974).
A 25 g amount of 2-methyl-2-nitrosopropane was added to 16.05 g
pentane, dried by suction and stored immediately at [20 ¡C. The
of NH Cl dissolved in 300 ml of water in three-necked 500 ml Ñask
yield was 2.08 g (23.4%).
4
Ðtted with a mechanical stirrer, then 31.7 g of zinc dust were added in
small portions over 15 min. Ice was added to keep the temperature
below 65 ¡C. When the temperature had fallen to below 45 ¡C the pre-
cipitated zinc oxide was Ðltered and the Ðltrate placed in a two-
necked 500 ml Ñask.
Equipment
EPR spectra were obtained using a Varian V4502 EPR
spectrometer. The calibration of the magnetic Ðeld in
gauss was e†ected by use of a Magnion Model G-502
gaussmeter with a manual Ðeld marker.
The Ñask was heated to 70 ¡C and a solution of 23 g of K CrO in
2
4
50 ml of 1 M H SO were added slowly from an addition funnel. The
2
4
water-bath temperature may be increased to 95 ¡C to drive over the
nitroso product, which is a blue gas over the reaction mixture. The
monomer dimerizes in the condenser and it may be necessary to warm
the condenser to move the dimer to a receiving Ñask (which should be
kept cool with an iceÈmethanol mixture). The white solid dimer of
MNP was dried over P O in a desiccator and recrystallized from a
minimum amount of n-pentane. The yield was 9.7 g of MNP dimer.
Di-tert-butyl peroxyoxalate (DBPO)8 was prepared as follows.
Note: DBPO is highly explosive. This material should never be
scratched or jarred. DBPO should be stored in very small quantities in
a deep-freeze. Do not allow it to melt and refreeze. Wear a protective
face shield and gloves. Further note: no explosion has ever been expe-
rienced during the preparation of DBPO or use of fresh DBPO.
However, repeated use by removal of small samples from a stored
container perhaps with slight warming followed by refreezing, or
exposure to humid air and/or room light, has resulted in two violent
spontaneous explosions of solid DBPO in the crystalline or glassy
state.
Spin trapping method
2
5
The correct amount (by weight) of the imine was placed
in one arm of a U-cell17 and dissolved in benzene. A
weighed amount of MNP was added to the benzene
solution of the imine. An appropriate amount of DBPO
was weighed and placed as dry crystals in the second
arm of the U-cell. Dry nitrogen was bubbled through
the benzene solution and over the DBPO crystals for 15
min or longer. A round EPR cell was degassed with
nitrogen and Ðtted to the U-cell for the EPR determi-
nations. Vacuum cell preparations did not improve the
linewidth of EPR spectra, and were used only for veriÐ-
cation purposes. The additional time needed to prepare
samples was not warranted for routine samples.
Oxalyl chloride (3.3 ml) in 40 ml of pentane was added to a solution
of 6.25 ml of pyridine and 5.6 ml of tert-butyl hydroperoxide in 60 ml
of pentane at 0 ¡C. The mixture can be allowed to warm to room
temperature with continuous stirring. The precipitated pyridine
hydrochloride was Ðlterd and washed with pentane. The pentane solu-
tions were combined and washed with 10% NaHCO solution, Ðltered
In this type of EPR spectroscopy, a-, b-, c- or d-
positions are designated for an atom or a group one,
two, three or four bonds away from the indicated spin
center in the structure of the molecule.
3
and cooled for 5 h at [20 ¡C. This cold solution was placed in a
dry-iceÈacetone bath for 20 min and the DBPO crystals were Ðltered.
The solid was dissolved in pentane, dried over MgSO and cooled at
4
[20 ¡C for 14 h. The crystals were Ðltered again, washed with cold
( 1997 by John Wiley & Sons, Ltd.
MAGNETIC RESONANCE IN CHEMISTRY, VOL. 35, 131È140 (1997)

134
E. G. JANZEN AND D. E. NUTTER
RESULTS
N-Benzylidene-N-alkylamines
When N-benzylidene-N-alkylamines reacted with tert-
butoxyl radicals from room temperature thermal
decomposition of DBPO in benzene in the presence of
MNP, EPR spectra could be obtained in situ for hours.
The intensity of the spectra increased to a plateau con-
centration after about 2È3 h, and thereafter remained at
that level for weeks.
The major spectrum obtained was consistent with the
conjugated a-iminylaminoxyl structure expected when
iminyl radicals are trapped by the nitroso function:
O
OÆ
>
=
C H wCÆ
6 5
NwC H ] C H wCwNwC H (9)
4 9
6 5
RAwN
4 9
>
>
RAwN
A variety of N-alkyl groups were investigated and the
EPR spectra interpreted in terms of a delocalized n-spin
system with two nitrogen hyperÐne splitting constants
(see Table 2); in general, the values found were as
follows: nitroxyl N-HFSC, aN B 9.7 G; iminyl
NO
N-HFSC, aN B 3.4 G.
NC
O~
O~
=
=
C H wCw~NwC H % C H wCxNwC H
6 5
RAwN
4 9
6 5
RAwNÆ
4 9
>
`
=
`
In Fig. 1 is shown the EPR spectrum from N-
benzylidene-3-pentylamine as an example of an adduct
with one hydrogen in the RA group in the b-position to
the iminyl nitrogen atom:
O~
=
C H wCxNwC H
6 5
4 9
=
`
(CH CH ) CHwNÆ
2 2
3
Figure 1. EPR spectrum from 0.005
benzylidene-3-pentylamine and 0.01
computer simulation using: main component aN ¼9.7 G; aN ¼
M
DBPO,
1
M N-
Inspection of the main pattern of lines reveals nine
groups of multiplets arranged as three 1 : 1 : 1 triplets.
The larger triplet splitting is assigned to the nitroxyl
M
MNP in benzene with
NO NC
3.4 G; aH ¼0.6 G; aH ¼0.35 G (2H); second component a
15.35 G; third component a ¼14.75 G; aH ¼1.3 G; relative area:
¼
R(b)
R(c)
N
nitrogen (aN B 9.7 G) and the smaller triplet splitting is
N
0.53, 0.42, 0.05.
NC
assigned to the iminyl nitrogen (aN B 3.4 G). These
NC
lines are again split into a doublet of 1 : 2 : 1 triplets.
This further splitting is assigned to one b-hydrogen
The latter N-HFSC is 14.75 G, which is slightly smaller
than the N-HFSC for DTBA. The di†erence is consis-
tent with the presence of an electron-withdrawing
iminyl group in the b-position. However, additional
small HFSs would be expected and a 1.3 G doublet pro-
vides a good Ðt in the computer-simulated spectrum.
Perhaps the dihedral angle is not favourable for N-HFS
in this case. It should be noted that this assignment is
speculative and no proof of structure is available at this
time.
The other examples of N-benzylidene-N-alkylamines
with one b-hydrogen in the R group give similar EPR
spectra (i.e. RA \ isopropyl, sec-butyl, sec-pentyl and
cyclohexyl). The appropriate EPR parameters are given
in Table 2.
[aH \ 0.60 G] and two c-hydrogens [a \ 0.35 G].
R(b)
R(c)
Although there are in total four c-hydrogens, appar-
ently only two give resolvable hyperÐne splitting. This
situation could be complicated because the c-methylene
hydrogens are also enantiotopic.
The EPR spectrum in Fig. 1 also shows at least two
more triplets, one undoubtedly from di-tert-butyl-
aminoxyl (DTBA), aN \ 15.35 G, and the other perhaps
from trapping the carbon-centered radical derived from
abstraction of the b-hydrogen in the RA group:
C H wCHxN tvBuOÕ
MNP
ÈÈÈ ÈÈÈ
6 5
(CH CH ) CH
=
?
3
2 2
C H wCHxN OÆ
6 5
=
=
In Fig. 2 is shown the EPR spectrum from N-
benzylideneisobutylamine as an example of an adduct
(CH CH ) CwNwC H (10)
2 2 4 9
3
( 1997 by John Wiley & Sons, Ltd.
MAGNETIC RESONANCE IN CHEMISTRY, VOL. 35, 131È140 (1997)

SPIN TRAPPING CHEMISTRY OF IMINYL FREE RADICALS
135
Table 2. HyperÐne splitting constants from MNP spin adducts of N-alkyl-
C-phenyliminyl radicalsa
R
aN
NO
aN
NC
aH
R(b)
aH
R(a)
Fig.
CH
9.70
9.68
9.73
9.72
9.84
9.78
9.74
9.70
9.75
9.75
3.53
3.45
3.47
3.47
3.36
3.56
3.39
3.40
3.44
3.42
5.34 (3H)
4.62 (2H)
5.10 (2H)
5.03 (2H)
5.66 (2H)
1.08 (1H)
0.78 (1H)
0.60 (1H)
0.81 (1H)
0.94 (1H)
—
—
3
3
CH CH
—
3
CH CH CH
2
3
2
CH CH CH CH
2
—
—
—
2
3
2
2
2
(CH ) CHCH
3 2
(CH ) CH
3 2
CH CH C(CH )H
2
—b
—
—
1
0.23 (4H)
0.35 (2H)
0.21 (4H)
—b
3
2
3
(CH CH ) CH
2 2
CH CH CH C(CH )H
3
—
—
3
2
2
3
Cyclohexyl
a EPR spectra obtained in benzene at room temperature from
1
M
N-
benzylidene-N-alkylamine, 0.005 M DBPO and 0.01 M MNP; hyperfine split-
tings constants measured in gauss.
b Additional splittings observed but not resolved.
with two hydrogens in the RA group present in the b-
position to the iminyl nitrogen atom:
O~
=
C H wCxNwC H
6 5
4 9
=
`
(CH ) CHCH NÆ
3 2
2
The spectrum looks complicated but if the large DTBA
triplet is ignored, the remaining 24 lines are reasonably
well resolved into intensity “1Ï and intensity “2Ï lines. In
fact, three groups of 1 : 2 : 1 triplets of 1 :1 :1 triplets can
be found consistent with two magnetically equivalent b-
methylene hydrogens, aH \ 5.66 G (2G), and one
R(b)
iminyl nitrogen, aN \ 3.36 G (27 lines total expected;
NC
three lines lost under the large DTBA triplet). Other
lines near the DTBA triplet hint at additional adducts
produced by MNP [aN \ 12.60; aH \ 2.50(1H) G], but
these products were not investigated further. A com-
puter simulation under the spectrum in Fig. 2 accounts
for all the components described.
The other examples of N-benzylidene-N-alkylamines
with two b-hydrogens in the RA group gave similar EPR
spectra (i.e. R \ ethyl, n-propyl and n-butyl). The
appropriate EPR parameters are given in Table 2.
In Fig. 3, the EPR spectrum from N-
benzylidenemethylamine is shown. Here there are three
b-hydrogens in the RA group next to the iminyl nitrogen
atom:
O~
=
C H wCxNwC H
6 5
4 9
=
`
CH NÆ
3
In this case 33 of the possible 36 lines are visible. The
resolution is adequate for recognition of 1 : 3 : 3 : 1 multi-
plets. In addition to the large triplet due to DTBA,
Figure 2. EPR spectrum from 0.005 M N-benzylideneisobuty-
lamine and 0.01 M MNP in benzene with computer simulation
using: main component aN ¼9.84 G; aN ¼3.36 G;
another triplet with a smaller N-HFSC, namely a B
N
NO
NC
8.0 G due to benzoyl-tert-butylaminoxyl is detected.
aH ¼5.66G(2H);secondcomponenta ¼15.36G;thirdcomponent
R(b)
¼12.60 G; aH ¼2.50 G (1H); relative area: 0.50, 0.39, 0.11.
N
a
This nitroxide comes from benzaldehyde impurity in the
N
( 1997 by John Wiley & Sons, Ltd.
MAGNETIC RESONANCE IN CHEMISTRY, VOL. 35, 131È140 (1997)

136
E. G. JANZEN AND D. E. NUTTER
Instead, a very strong EPR spectrum due to DTBA was
detected, indicating that b-cleavage of the intermediate
iminyl radical occurs [eqn (13)] to produce tert-butyl
radicals and benzonitrile (cyanobenzene) before spin
trapping with MNP can be accomplished (spectrum not
shown). Also, when N-benzylidene-N-tert-octylamine
was subjected to the same conditions, a triplet pattern
consistent with tert-octyl-tert-butylaminoxyl was pro-
duced, with little else. This kind of b-cleavage of iminyl
radicals has been proposed by Ohta and Tokumaru18
to account for their observations.
When N-benzylidene-N-a-methylbenzylamine was
investigated, a strong EPR spectrum composed of a
triplet of doublets was detected with DBPO and MNP.
By allowing ethylbenzene in a di†erent experiment to
react with tert-butoxyl radicals from DBPO in the pres-
ence of MNP, the same spectrum was obtained. There-
fore,
the
assignment
of
a-methylbenzyl-tert-
butylaminoxyl to the Ðrst EPR spectrum could be veri-
Ðed (spectra not shown):
OÆ
C H wCÆ
=
6 5
MNP
> ÈÈÈ C H CH(CH )NwC H
6 5
3
4 9
C H CH(CH )wN
6 5
3
aN \ 14.95; aH \ 3.59 G
b
DBPO
C H CH CH ÈÈÈ ÈÈÈ
6 5
MNP
(14)
2
3
aN \ 14.87; aH \ 3.58 G
This spin adduct is also the product of spin trapping
b
Figure 3. EPR spectrum from 0.005
M
DBPO,
1
M N-
benzylidenemethylamine and 0.01 M MNP in benzene with com-
methyl
radical
by
C-phenyl-N-tert-butylnitrone
puter simulation using: main component aN ¼9.70 G; aN ¼3.53
G; aH ¼5.34 G (3H); second component a ¼15.36 G; third
component a ¼8.01 G; relative area: 0.75, 0.14, 0.1.
NO
NC
(PBN):19 aN \ 14.81; aH \ 3.48 G.
b
R(b)
N
When N-benzylidene-N-benzylamine was subjected
N
to tert-butoxyl radicals from DBPO in benzene in the
presence of MNP, a relatively weak EPR spectrum was
observed. The strongest components after 10 min were
DTBA, tert-butoxyl-tert-butylaminoxyl (aN \ 27.1 G),
an unknown triplet with aN B 10.38 G (see later) and a
small amount of benzyl-tert-butylaminoxyl (aN \ 14.8;
imine reacting with the tert-butoxyl radicals followed by
addition to MNP:
O
O
O
OÆ
>
>
>
=
C H wCÆ ] NwC H ] C H wCwNwC H (11)
6 5
4 9
4 5
4 9
aH \ 7.4 G). After 8 h the benzyl-tert-butylaminoxyl
b
The computer simulation under the spectrum in Fig.
3 accounts for the N-methylimine adduct using the EPR
parameters shown in Table 2, plus DTBA and benzoyl-
tert-butylaminoxyl in small amounts.
pattern was still discernible as a weak set of lines
dwarfed by a large DTBA triplet. During the interim
period (e.g. after 1 h), spectra were obtained which con-
tained more patterns of lines but the signal intensity
and resolution were insufficient for good spectral
analysis to be accomplished (spectra not shown).
In comparing the last three compounds, it is clear
that the release of the RA group as an MNP detectable
free radical produced from b-cleavage of the N-
benzylidene-N-alkyliminyl radical follows the following
order: tert-butyl [ a-methylbenzyl [ benzyl.
The compounds where RA is N-tert-butyl, N-benzyl
and N-a-methylbenzyl constitute a special case. In the
Ðrst example when N-benzylidene-N-tert-butylamine
was reacted with tert-butoxyl radical in the presence of
MNP, no spectrum consistent with the “normalÏ spin
adduct was detected [Eqn (12)]:
The question of hydrogen atom abstraction from a
site remote from the iminyl position is also of interest.
Inspection of the EPR spectra obtained from the N-
alkylimines indicated that typically another smaller
triplet of moderate intensity was present which could be
due to the MNP adduct of an alkyl adduct with very
small b-H HFSs. The following data were gleaned from
the spectra but estimates are very approximate:
aN (G)
Notes
NO
methyl
ethyl
È
only a small amount of DTBA
very little evidence; perhaps only
DTBA
È
( 1997 by John Wiley & Sons, Ltd.
MAGNETIC RESONANCE IN CHEMISTRY, VOL. 35, 131È140 (1997)

SPIN TRAPPING CHEMISTRY OF IMINYL FREE RADICALS
137
propyl
14.1
D4 ] weaker than iminyl spin
adduct
D9 ] weaker than iminyl adduct
D3 ] weaker than iminyl adduct
D9 ] weaker than iminyl adduct
In this work, the EPR spectrum obtained from the
MNP adduct of the unsubstituted N-benzylideneaniline
was computer simulated and the best Ðt was found
when three aryl protons were assumed to have the same
HFS (namely aH \ 0.52 G, see Table 3) plus two addi-
tional smaller but equal aryl HFSs (almost half, namely
aH \ 0.31 G, see Table 3). The EPR spectrum due to
this aminoxyl has been published before by Aurich et
al.20 but the aryl hydrogen splittings were given as 0.45
G and all positions equal (see Table 3). Since Aurich et
al.Ïs value is the average between the larger and smaller
numbers we report, perhaps with slightly better
resolution obtained in this work, the di†erences
between the HFSs from di†erent positions in the ring
could be distinguished. This makes it unnecessary to
theorize about the possibility that both n- and p-
delocalization of spin at the same time produces equal
HFS in all positions of the N-aryl ring in these amino-
xyl radicals.20
isobutyl
2-butyl
3-pentyl
15.5
14.2
14.1
Although these spectra cannot be assigned with cer-
tainty at this time, it is likely that spin adducts from
alkyl group hydrogen atom abstraction occurred. All
but the n-propyl case are examples where a tertiary
hydrogen exists in this group. The peak heights are
comparable in size but the fact that the iminyl adduct
has many more lines makes this nitroxide stronger in
molar concentration.
N-Benzylideneanilines
After mixing N-benzylideneanilines or substituted N-
benzylideneanilines in benzene with MNP and DBPO,
rich EPR spectra were obtained in situ in the EPR
spectrometer cavity. Although the splitting patterns are
complex, they could be analyzed in terms of one amino-
In the case of N-benzylidene-4-Ñuoroaniline, the EPR
spectrum was obtained was considerably di†erent.
Aurich et al.20 also reported this spectrum and the
assignments of these spectra (theirs and ours) agree well
(see Table 3). In this case, because the hyperÐne splitting
patterns are consistent with one relatively large HFS
(1.1 G), two medium-sized HFSs (but similar to the
larger HFS found in the unsubstituted example, namely
0.52 G) and two relatively small HFSs (0.31 G), it is
suggested that the HFSs from the ortho and para posi-
xyl N-HFS of about 9 G (aN \ 8.45È9.15 G, see Table
NO
3) and another nitrogen triplet with HFS B 3.7 G (aN
NC
3.6È3.77 G, see Table 3). Further splitting is difficult to
resolve completely.
By substitution in the anilino phenyl ring it could be
determined that the detected multiplets originate from
proton hyperÐne splitting in this ring. However, the
assignment to speciÐc positions in the ring is not pos-
sible. It is assumed that limited rotation of this phenyl
ring may make the positions in the ortho and meta posi-
tions unequal. Thus in N-benzylidene aniline the spin
adduct might have Ðve di†erent aryl hydrogen HFSs:
tions are approximately equal (aN B 0.55 G) and the
ohp
HFSs from the meta positions are smaller and also
approximately equal (aH B 0.38 G).
m
Table 3. HyperÐne splitting constants from MNP spin adducts of N-aryl-C-
phenyliminyl radicalsa
X
Y
aN
NO
aN
NC
aH
NhC H hX
Ref.
6
4
H
H
H
H
8.97
8.92
9.1
3.64
3.68
3.66
3.68
3.66
3.68
3.60
3.66
3.77
3.72
3.68
3.70
0.52 (3H); 0.32 (2H)
0.45 (5H)
—
—
20
—
—
—
—
—
—
—
—
—
F
H
0.56 (2H); 0.315 (2H); 1.06 (1F)
0.55 (2H); 0.3 (2H); 1.10 (1F)
0.47 (7H)
F
H
8.84
8.96
8.96
8.95
8.78
8.45
8.70
9.04
9.15
CH
CH
H
3
3
H
0.5 (7H)
CH O
H
0.68 (2H); 0.34 (2H)
0.53 (2H); 0.36 (2H)
0.41 (multiplet spacing)b
0.40 (2H ½3F); 0.35 (2H)
0.41 (multiplet spacing)b
0.40 (multiplet spacing)b
3
Cl
H
NO
2
H
CF
3
H
H
CH
3
CH O
H
3
a EPR spectra obtained in benzene at room temperature from 1 M N-benzylidene-N-
arylamine, 0.005 M DBPO and 0.01 M MNP; hyperfine splittings are measured in
gauss.
b Spacing of peaks in multiplet; resolution inadequate for analysis.
( 1997 by John Wiley & Sons, Ltd.
MAGNETIC RESONANCE IN CHEMISTRY, VOL. 35, 131È140 (1997)

138
E. G. JANZEN AND D. E. NUTTER
Similar conclusions can be reached about the 4-
and oxygen. In this case the oxyl radical proposed is
actually the acylamino radical, which should be readily
trapped by MNP to give the spin adduct indicated.
methyl, 4-methoxy, 4-chloro and 4-nitroaniline deriv-
atives (see Table 3), although sufficient resolution was
not available in every case to distinguish between ortho
and meta positions.
Relative rate constants
Since MNP adducts of the iminyl and tert-butoxyl rad-
icals are produced simultaneously in benzene the rela-
tive rate constants of hydrogen atom abstraction can be
determined if the addition of tert-butoxyl radicals to
MNP is known, i.e. if reactions (16) and (17) are in com-
petition:
Unassigned EPR spectra
In all the experiments described so far an additional
EPR component was present in the spectrum when
oxygen was incompletely removed from the solution or
when oxygen was purposely added by bubbling through
the benzene solution. This spectrum was characterized
by a 10È11 G triplet with a very broad linewidth, some-
times revealing some extra splittings. A total of 23 ben-
zylideneamines including those in Tables 2 and 3 plus
the N-tert-butyl, N-tert-octyl, N-benzyl and N-(a-
methylbenzyl) derivatives showed this kind of triplet
with aN \ 10.2È10.96 G. In some cases evidence of addi-
tional splitting could be resolved which reÑected the
number of magnetically equivalent hydrogens in the
N-alkyl group, e.g. 0.8 G quartets from N-methyl and
0.5 G triplets from N-ethyl and N-propyl groups.
Proposals speculating about a structure for this
aminoxyl must take into account the peculiar size of the
nitrogen triplet hyperÐne splitting and the lack of addi-
tional signiÐcant splitting from another nucleus. The
range of N-HFS as one considers possible aminoxyl
structures might be as follows:
C H wCwH
C H wCÆ
6 5
tvC4H9OÕ 6 5
>
ÈÈÈÈ
Õ
>
(16)
RwN
RwN
OÆ
=
tvC4H9OÕ
C H N \ 0 ÈÈÈÈ C H ONC H
(17)
4 9
4 9
4 9
the following equation applies:
d(SA )/dt
k
[C H NO]
17
17 4 9
\
d(SA )/dt
k [imine]
16
16
where SA \ spin adduct.
Using the equation A \ bhw2, where A \ area of
peak, h \ peak height, w \ linewidth measured hori-
zontally from maximum to minimum of the Ðrst deriv-
ative peak and b is the line multiplicity of the splitting
pattern (e.g. b \ 3 for a 1 :1 :1 triplet), the change in spin
adduct intensity for the imine where RA \ methyl can
be monitored in an experiment performed in the normal
way. For the MNP tert-butoxyl adduct b \ 3,
OÆ
OÆ
OÆ
OÆ
=
=
=
=
wNwCw
wNwCw wNwNw
wNwCw
w2 \ 0.36 and k \ 1.5 ] 106 M~1 s~1;22 therefore, for
>
>
>
>
17
O
Nw
wCw
wCw
the MNP iminyl adduct, b \ 27, w2 \ 0.058 and k
\
16
1.2 [ 104 M~1 s~1 (calculated). This estimate for iminyl
hydrogen atom abstraction is considerably slower than
for the same reaction with benzaldehyde. With benz-
aldehyde, tert-butoxyl radicals abstract the acyl hydro-
gen at a rate of 2 ] 107 M~1 s~1.11
aN 7.8È8.0 G
aN 9.7È9.9 G
10È11 G?
12.5 G21
NO
NO
A possible suitable structure might be produced if a
nitrogen-centered radical which has been trapped by
MNP has a double bond to carbon in its most preva-
lent resonance form. A reaction sequence which could
create this kind of aminoxyl is shown below:
Hydrogen atom abstraction from nitrones
C H wCÆ
C H wCwOOÆ
6 5
6 5
2x
ÈÈÈ
> ] O ÈÈÈ
Õ
>
2
C H wN
C H wN
It was of interest to compare the possibility of hydrogen
atom abstraction from the nitronyl function with the
same hydrogen atom in the iminyl compound of similar
structure:
2 5
2 5
C H wCwOÆ
C H wCxO
6 5
6 5
>
ÈÈÈ
Õ
=
(15)
C H wN
C H wNÆ
2 5
2 5
H
C H
4 9
H
C H
4 9
z
C H wCxO
C H w C wOh
6 5
}
z
z
}
}
z
MNP 6 5
ÈÈÈ
Õ
=
Ô
Õ
>
ÈÈÈ
CxN
ÈÈÈ
CxN
C H wNwNwC H
C H w N wNwC H
2
5
4
9
2
5
4 9
=
]
=
C H
6 5
O
C H
6 5
OÆ
OÆ
If some radical species was capable of abstracting a
nitronyl hydrogen, the resulting radical would give a
nitronyl nitroxide spin adduct with MNP:23
The b-position to the nitroxyl function in acyl nitro-
xides is known to have very little (if any) spin and the
small spin on the a-nitrogen atom in the proposed
structure might also give only a very small hyperÐne
splitting from the methyl or ethyl methylenic hydrogens
in the other b-position (0.8È0.5 G). The iminyl radical
should react with oxygen rapidly to form the peroxyl
radical. Peroxyl radicals are known to couple, and the
tetraoxides produced dissociate to give oxyl radicals
OÆ
C H wCÆ
=
6 5
>
] MNP ÈÈÈ C H CwN wC H
6 5
4 9
N]
}
>
z
N]
}
O
C H
4 9
z
~
O
C H
4 9
(18)
~
( 1997 by John Wiley & Sons, Ltd.
MAGNETIC RESONANCE IN CHEMISTRY, VOL. 35, 131È140 (1997)

SPIN TRAPPING CHEMISTRY OF IMINYL FREE RADICALS
139
Since these nitronyl nitroxides have considerable stabil-
ity, one would expect the presence of even a small
amount of nitronyl hydrogen abstraction to be detect-
able.
Experiments in which DBPO was reacted with C-
phenyl-N-tert-butylnitrone in the presence of MNP
showed no evidence of nitronyl nitroxide formation
even after several days. Only tert-butoxyl addition to
form the spin adduct occurred.22
s~1, respectively.26 The value for iminyl hydrogen
abstraction in this work is 1.2Œ104 M~1 s~1. The esti-
mate from EPR spectral intensity comparison would
place abstraction from remote hydrogens in the N-alkyl
group at about 3È9 times less. However, trapping bulky
tertiary alkyl radicals might be slower than trapping
iminyl radicals. Therefore, these values may be under-
estimated.
CONCLUSIONS
DISCUSSION
This study has contributed information about the prob-
ability of abstracting a hydrogen atom from a given site
in a series of benzylidene-N-alkylamines using spin
trapping methodology. On the basis of this work, it is
possible to predict that tert-butoxyl radicals will react
by hydrogen atom abstraction in the following order (1
is the fastest, 5 is the slowest):
The special energetics of radical approach to a nitrone
function have been explored recently in the case of a
nitrone24 and the route towards carbon bond formation
through additions has been clearly delineated theoreti-
cally. When tert-butoxyl radicals react with styrene the
process also seems to be addition.25 However, with
benzaldehyde and benzylidene imines the reaction is
only hydrogen atom abstraction, as illustrated in this
work. No evidence for addition was found:
O
=
radical
C H wCHxCH C H wCHxNw addition only
6 5
2
6 5
hydrogen atom
C H wCHxO
C H wCHxNw abstraction only
6 5
6 5
However, this prediction is speculative for hydrogens at
positions 2, 3 and 4 because the EPR spectra for sec-
ondary adducts were not explicitly analyzed and assign-
ed. No EPR spectra consistent with aryl radical adducts
were found; therefore, position 5 must be the least
likely.
This subtle di†erence would seem to warrant further
study. Perhaps the transition states for the two former
cases are polarized as the tert-butoxyl radical
approaches so that addition is favored over hydrogen
atom abstraction. If polar transition states do not
develop, perhaps the only other route remaining is
hydrogen atom abstraction.
The fact that hydrogen atom abstraction from the
N-alkyl group occurs is also interesting. That the
amount of spin adduct formed is comparable to the
amount from iminyl radicals might not have been
expected. This would indicated that similar absolute
rate constants apply. Thus, the absolute rate constant
for hydrogen atom abstraction from toluene and cyclo-
hexane has been estimated at 2Œ104 and 1Œ105 M~1
Acknowledgements
We thank Dr Hong Sang (National Biomedical Center for Spin Trap-
ping and Free Radicals, Oklahoma Medical Research Foundation,
Oklahoma City, OK, USA) for providing the computer simulations
for the EPR spectra shown. We are also grateful to the Natural Sci-
ences and Engineering Research Council of Canada for research
funding and to Mrs Luci White for excellent assistance in preparing
the manuscript.
REFERENCES
1. W. A. Pryor, Free Radicals. McGraw-Hill, New York (1966).
2. A. R. Forrester, J. M. Hay and R. H. Thomson, Organic Chem-
istry of Stable Free Radicals. Academic Press, London (1968).
3. M. J. Perkins, in Essays on Free Radical Chemistry, edited by
R. O. C. Norman, Chapt. 5. Chemical Society, London (1970).
4. E. G. Janzen, E. R. Davis and D. E. Nutter, Tetrahedron Lett.
3309 (1978).
5. H. G. Aurich, W. Dersch and H. Forster, Chem. Ber. 106, 2854
(1973), and references cited therein.
6. E. F. Ullman, L. Call and J. H. Osiecki, J. Org. Chem. 35, 3623
(1970).
7. E. Boyland and R. Nery, J. Chem. Soc. C 354 (1966), and
references cited therein; A. R. Forrester, J. M. Hay and R. H.
Thomson, Organic Chemistry of Stable Free Radicals, p. 214.
Academic Press, London (1968).
8. P. D. Bartlett, E. P. Benzing and R. E. Pincock, J. Am. Chem.
Soc. 82, 1762 (1960).
9. W. C. Danen and C. T. West, J. Am. Chem. Soc. 95, 6873
(1973).
10. E. G. Janzen and G. A. Coulter, Tetrahedron Lett. 22, 615
(1981).
11. E. G. Janzen, E. R. Davis and C. M. DuBose, Magn. Reson.
Chem. 33, S166 (1995).
12. D. L. Haire and E. G. Janzen, Magn. Reson. Chem. 32, 151
(1994).
13. E. G. Janzen and D. L. Haire, in Two Decades of Spin Trap-
ping, Advances in Free Radical Chemistry, edited by D. D.
Tanner, Chapt. 6, pp. 253–295. JAI Press, Greenwich, CT
(1990).
14. B. L. Emling, R. J. Haworth, A. J. Saraceno, E. F. Ellermeyer,
L. Haile and L. D. Hudac, J. Org. Chem. 24, 657 (1959).
15. W. D. Emmons, J. Am. Chem. Soc. 79, 6522 (1957).
16. N. Kornblum, R. J. Clutter and W. J. Jones, J. Am. Chem.
Soc. 78, 4003 (1956).
( 1997 by John Wiley & Sons, Ltd.
MAGNETIC RESONANCE IN CHEMISTRY, VOL. 35, 131È140 (1997)

140
E. G. JANZEN AND D. E. NUTTER
17. G. A. Russell, E. G. Janzen and E. T. Strom, J. Am. Chem. Soc.
86, 1807 (1964).
22. E. G. Janzen and C. A. Evans, J. Am. Chem. Soc. 95, 8205
(1973).
18. H. Ohta and K. Tokumaru, Chem. Commun. 1601 (1970).
19. E. G. Janzen and B. J. Blackburn, J. Am. Chem. Soc. 91, 4481
(1969).
23. P. W. Kopf, R. Kreilick, D. G. B. Bobcock and E. F. Ullman, J.
Am. Chem. Soc. 92, 4531 (1970), and references cited
therein.
20. H. G. Aurich, H. Forster, A. Lotz and W. Weiss, Chem. Ber.
106, 2832 (1973).
21. C. M. Camaggi, R. Holman, M. J. Perkins, J. Chem. Soc.,
Perkin Trans. 2 501 (1972); vinyl-tert-butylaminoxyl has N-
and H-HFSC (2H) ¼12.5 and 1.1 G, respectively, in benzene;
see also Ref. 2, p. 216.
24. S. L. Boyd and R. J. Boyd, J. Phys. Chem. 98, 11705 (1994).
25. G. R. Chalfont and M. J. Perkins, J. Am. Chem. Soc. 90, 7141
(1968).
26. J. A. Howard, Adv. Free Rad. Chem. 4, 49 (1971).
( 1997 by John Wiley & Sons, Ltd.
MAGNETIC RESONANCE IN CHEMISTRY, VOL. 35, 131È140 (1997)