Mechanisms of reaction of aminoxyl (nitroxide), iminoxyl, and imidoxyl radicals with alkenes and evidence that in the presence of lead tetraacetate, N-hydroxyphthalimide reacts with alkenes by both radical and nonradical mechanisms

Article abstract of DOI:10.1021/jo0504060

1,2-Dideuterio-cyclohexene, 1,2-dideuterio-cyclooctene, and trans-3,4-dideuterio-hex-3-ene were reacted with three >NO^. radicals: 4-hydroxyTempo, di-tert-butyliminoxyl, both used as the actual radicals, and phthalimide-N-oxyl (PINO) generated from N-hydroxyphthalimide (NHPI) by its reaction with tert-alkoxyl radicals (t-RO^.) and with lead tetraacetate. In all cases, except the NHPI/Pb(OAc)₄ system, only mono >NO^.-substituted alkenes were produced. The ²H NMR spectra imply that 88-92% of monoadducts were formed by the initial abstraction of an allylic H-atom, followed by capture of the allylic radical by a second >NO^., while the remaining 12-8% appear to be formed by an initial addition of >NO^. to the double bond followed by H-atom abstraction by a second >NO^.. A substantial and sometimes the major product formed with the NHPI/Pb(OAc)₄ system has two PINO moieties added across the double bond. Since such diadducts are not formed with the NHPI/t-RO^. system, a heterolytic mechanism is proposed, analogous to that known for the Pb(OAc)₄-induced acetoxylation of alkenes. A detailed analysis of the NHPI/Pb(OAc) ₄/alkene products indicates that monosubstitution occurs by both homolytic and heterolytic processes.

Full text of DOI:10.1021/jo0504060

Mechanisms of Reaction of Aminoxyl (Nitroxide), Iminoxyl, and
Imidoxyl Radicals with Alkenes and Evidence that in the Presence
of Lead Tetraacetate, N-Hydroxyphthalimide Reacts with Alkenes
by Both Radical and Nonradical Mechanisms
,†
‡
Sergiu Coseri,* G. David Mendenhall, and K. U. Ingold
National Research Council, Ottawa, ON, Canada K1A 0R6
coseris@icmpp.ro
Received March 3, 2005
1
,2-Dideuterio-cyclohexene, 1,2-dideuterio-cyclooctene, and trans-3,4-dideuterio-hex-3-ene were
•
reacted with three >NO radicals: 4-hydroxyTempo, di-tert-butyliminoxyl, both used as the actual
radicals, and phthalimide-N-oxyl (PINO) generated from N-hydroxyphthalimide (NHPI) by its
reaction with tert-alkoxyl radicals (t-RO ) and with lead tetraacetate. In all cases, except the NHPI/
Pb(OAc)
•
2
4
system, only mono >NO-substituted alkenes were produced. The H NMR spectra imply
that 88-92% of monoadducts were formed by the initial abstraction of an allylic H-atom, followed
•
by capture of the allylic radical by a second >NO , while the remaining 12-8% appear to be formed
•
•
by an initial addition of >NO to the double bond followed by H-atom abstraction by a second >NO .
A substantial and sometimes the major product formed with the NHPI/Pb(OAc)
4
system has two
PINO moieties added across the double bond. Since such diadducts are not formed with the NHPI/
•
t-RO system, a heterolytic mechanism is proposed, analogous to that known for the Pb(OAc)
induced acetoxylation of alkenes. A detailed analysis of the NHPI/Pb(OAc) /alkene products
indicates that monosubstitution occurs by both homolytic and heterolytic processes.
4
-
4
Introduction
particularly di-tert-butyliminoxyl (2)3,4 and imidoxyls
2,5-7
such as phthalimide-N-oxyl, PINO (3),
and related
radicals.8
In recent years, there has been an upsurge in research
directed toward improving the rates of oxidation of
organic compounds and improving product selectivities.
One promising approach uses radicals bearing the >NO
moiety. Persistent >NO radicals have often been em-
•
•
ployed as neat, stoichiometric oxidants. Both persistent
•
and transient >NO have been even more frequently
employed in catalytic quantities in radical chain reactions
in the presence of oxygen and, frequently, also in the
presence of a transition metal ion as a cocatalyst. The
•
These radicals have been used to oxidize hydrocarbons,
alcohols, and aldehydes at higher rates and with better
>
NO radicals used include aminoxyls (nitroxides) such
as Tempo (1a) and related species, iminoxyl radicals,
1,2
(
1) Reviews: (a) Sheldon, R. A.; Arends, I. W. C. E.; Brink, G.-J. T.;
†
Permanent address: P. Poni Institute of Macromolecular Chem-
Dijksman, A. Acc. Chem. Res. 2002, 35, 774-781. (b) Minisci, F.;
Recupero, F.; Pedulli, G. F.; Lucarini, M. J. Mol. Catal. A: Chem. 2003,
204-205, 63-90. (c) Sheldon, R. A.; Arends, I. W. C. E. Adv. Synth.
Catal. 2004, 346, 1051-1071.
istry, Gr. Ghica Voda Alley, 41A, 6600, lasi, Romania.
‡
Current address: Eastern Sources, Inc., 8 Westchester Plaza,
Elmsford, NY 10523.
1
0.1021/jo0504060 CCC: $30.25 © 2005 American Chemical Society
Published on Web 05/10/2005
J. Org. Chem. 2005, 70, 4629-4636
4629

Coseri et al.
SCHEME 1
selectivities than can be achieved using oxygen alone.
Generally, there is no mechanistic ambiguity: the >NO -
challenged by Babiarz et al.10 These workers suggested
that the reaction of 4-hydroxyTempo, 1b, with cyclohex-
ene involved an initial addition of the radical to the
•
abstracts a hydrogen atom from the substrate, RH, and
the derived carbon-centered radical may react with a
double bond followed by H-atom abstraction from the CH
2
•
second >NO , or with dioxygen, depending on the condi-
group adjacent to the new radical center. That is, Babiarz
tions:
et al.10 proposed that the monosubstitution product from
1
b and cyclohexene followed an addition-abstraction
mechanism rather than an abstraction-addition mech-
anism as had normally been assumed. As we have noted
previously,11 the addition-abstraction mechanism was
proposed “because the bond dissociation energy (enthal-
pies) for allylic and benzylic C-H bonds is expected to
be nearly the same.”10 This mechanism was also claimed
to be supported by semiempirical UHF/AM1 calcula-
tions.10 Unfortunately, even if the enthalpic (thermo-
12
chemical) arguments were correct, entropic effects
favoring H-abstraction from cyclohexene had been
However, for some compounds, particularly the cy-
cloalkenes, there is a mechanistic ambiguity. The com-
mon assumption is that the >NO will abstract an allylic
H-atom, thus forming an allylic radical that undergoes
the further reactions. This assumption was recently
13
ignored.
9
•
In a preliminary communication,11 we reported a
simple and convenient method for distinguishing between
these two mechanisms and mixtures thereof. We em-
ployed the same reaction as Babiarz et al. except that
cyclohexene was replaced by its 1,2-dideuterio isoto-
pomer, 4. This deuterium label was chosen because, for
both mechanisms, there would be no primary deuterium
kinetic isotope effect and hence little isotopic “biasing”
(
2) Minisci, F.; Recupero, F.; Cecchetto, A.; Gambarotti, C.; Punta,
C.; Faletti, R.; Paganelli, R.; Pedulli, G. F. Eur. J. Org. Chem. 2004,
1
2
1
09-119.
(
3) Mendenhall, G. D.; Ingold, K. U. J. Am. Chem. Soc. 1973, 95,
963-2971.
(4) (a) Lornejo, J. J.; Larson, K. D.; Mendenhall, G. D. J. Org. Chem.
2
985, 50, 5382-5383. (b) Ngo, M.; Larson, K. R.; Mendenhall, G. D. J.
of the results. Furthermore, H NMR of the products
Org. Chem. 1986, 51, 5390-5393. (c) Eisenhauer, B. M.; Wang, M.;
Brown, R. E.; Labaziewicz, H.; Ngo, M.; Kettinger, K. W.; Mendenhall,
G. D. J. Phys. Org. Chem. 1997, 10, 737-746. (d) Eisenhauer, B. M.;
Wang, M.; Labaziewicz, H.; Ngo, M.; Mendenhall, G. D. J. Org. Chem.
would provide a clear marker of the reaction mechanism
since one product would be unique to the abstraction-
addition mechanism (see Scheme 1).
1
997, 62, 2050-2053.
5) Reviews: (a) Ref 1b. (b) Ref 1c. (c) Ishi, Y.; Sakaguchi, S.;
Iwahama, T. Adv. Synth. Catal. 2001, 343, 393-427.
6) No transition metal ion cocatalyst; see e.g.: (a) Ueda, C.;
Noyama, M.; Ohmori, H.; Masui, M. Chem. Pharm. Bull. 1987, 35,
372-1377. (b) Amorati, R.; Lucarini, M.; Mugnaini, V.; Pedulli, G.
To our surprise, both mechanisms are operational in
the 1b + 4 reaction, the more important being abstrac-
tion-addition (see Scheme 1). We have carried this work
forward with eight additional reactions, involving 4 and
(
(
1
F.; Minisci, F.; Recupero, F.; Fontana, F.; Astolfi, P.; Greci, L. J. Org.
Chem. 2003, 68, 1747-1754. (c) Aoki, Y.; Sakaguchi, S.; Ishii, Y. Adv.
Synth. Catal. 2004, 346, 199-202. (d) Koshino, N.; Saha, B.; Espenson,
J. H. J. Org. Chem. 2003, 68, 9364-9370. (e) Koshino, N.; Cai, Y.;
Espenson, J. H. J. Phys. Chem. A 2003, 107, 4262-4267. (f) Baciocchi,
E.; Gerini, M. F.; Lanzalunga, O. J. Org. Chem. 2004, 69, 8963-8966.
two new alkenes, 1,2-dideuterio-cyclooctene, 5, and trans-
•
3
2
,4-dideuterio-hex-3-ene, 6, using the >NO radicals 1b,
, and 3.
(7) With a transition metal ion cocatalyst; see e.g.: (a) Ishii, Y.;
Iwahama, T.; Sakaguchi, S.; Nakayama, K.; Nishiyama, Y. J. Org.
Chem. 1996, 61, 4520-4526. (b) Ishii, Y.; Sakaguchi, S. Catal. Surv.
Jpn. 1999, 3, 27-35. (c) Minisci, F.; Punta, C.; Recupero, F.; Fontana,
F.; Pedulli, G. F. Chem. Commun. 2002, 688-689. (d) Figiel, P. J.;
Sobczak, J. M.; Ziolkowski, J. J. Chem. Commun. 2004, 244-245. (e)
Minisci, F.; Porta, O.; Recupero, F.; Gambarotti, C.; Paganelli, R.;
Pedulli, G. F.; Fontana, F. Tetrahedron Lett. 2004, 45, 1607-1609.
Both mechanisms are operative for all of the eight new
reactions and, astonishingly, for both the nitroxide, 1b,
and the iminoxyl, 2; the partitioning between the two
(
8) (a) Annunziatini, C.; Gerini, M. F.; Lanzalunga, O.; Lucarini,
M. J. Org. Chem. 2004, 69, 3431-3438. (b) Baucherel, X.; Gonsalvi,
L.; Arends, I. W. C. E.; Ellwood, S.; Sheldon, R. A. Adv. Synth. Catal.
2
004, 346, 286-296. (c) Cai, Y.; Koshino, N.; Saha, B.; Espenson, J.
H. J. Org. Chem. 2005, 70, 238-243.
9) See e.g.: Bottle, S. E.; Busfield, W. K.; Jenkins, I. D. J. Chem.
Soc., Perkins Trans. 2 1992, 2145-2150.
(
(10) Babiarz, J. E.; Cunkle, G. T.; DeBellis, A. D.; Eveland, D.;
Pastor, S. D.; Shum, S. P. J. Org. Chem. 2002, 67, 6831-6834.
4630 J. Org. Chem., Vol. 70, No. 12, 2005

Aminoxyl (Nitroxide), Iminoxyl, and Imidoxyl Radicals
2
mechanisms for all three alkenes is roughly the same as
that found for the 1b + 4 reaction.11 The same partition-
CHCl
3
. The T
1
relaxation times of the H signals were ∼100
ms, and the delay time between pulses was therefore chosen
to be 2s (.5 × T
1
) to ensure accurate integration of these
ing was also found for the imidoxyl, PINO (3), when this
radical was generated by H-atom abstraction from N-
hydroxyphthalimide, NHPI (3H), using tert-alkoxyl radi-
signals.
(
i) 1b. Each of the three neat alkenes was reacted with ca.
0 mol % of the neat nitroxide under nitrogen at the temper-
1
•
cals, t-RO . However, the same partitioning was not found
atures and for the times indicated: 4, 70 °C, 72 h; 5, 88 °C, 48
h; 6, ∼64 °C (reflux), 72 h. The workup of the 1b (mono)-
when the alkenes were allowed to react with NHPI and
lead tetracetate in benzene. This is a chemical system
that was shown to form PINO radicals as early as 1964
by Lemaire and Rassat15 using EPR spectroscopy (reac-
2
substituted dideuterio-alkene products for H NMR spectro-
scopic analysis was the same as that described previously for
11
the 1b/4 reaction. For both the abstraction-addition and
addition-abstraction mechanisms, half of 1b abstracts a
hydrogen atom to form 1bH. Thus, the maximum yield of the
adducts is 50% on the basis of 1b. The actual yields were 15.5%
4
tion 5). Furthermore, for the NHPI/Pb(OAc) system, a
significant, even major, product was a diadduct in which
two PINO moieties had added across the double bond of
the alkene.
(
4), 18% (5), and 20% (6). These yields would presumably have
been doubled if the reactions had been carried out in the
presence of an oxidizing agent (such as silver oxide), which
would convert the 1bH product back to 1b. Even higher yields
would have been obtained if the reactions had been run for
2
longer times or at higher temperatures, since the H NMR
showed no sign of other products (see Supporting Information).
Higher yields were not, however, required for our purposes.
(
ii) 2. The three neat alkenes were mixed under nitrogen
with the neat iminoxyl radical (molar ratio ca. 2.2:1.0) and
allowed to stand (still under nitrogen) at room temperature.
The blue color of the radical slowly faded, and crystals of di-
tert-butyl ketoxime, 2H, were formed. After 90 min, a roughly
equal volume (0.5 mL) of pentane was added, and the mixture
was cooled to -60 °C for 1 h. The supernatant was removed
and passed through a short column of basic alumina to remove
the remaining 2H. The eluate and pentane washings were
concentrated to give the O-(dideuterio-alkenyl) oxime products.
Again, the maximum yield of the adducts is 50% on the basis
of 2. The actual yields were 28% (4, the isolation and
characterization of the allylic substitution product of 2 +
nondeuterated 4 have been described elsewhere ), 30% (5),
and 31% (6). These yields could also, presumably, have been
doubled if the reaction had been carried out in the presence
of an oxidizing agent to regenerate 2 from 2H and could have
been further improved by a longer reaction time and/or a
higher reaction temperature.
Experimental Section
Materials. (i) >NO• Radicals 1, 2, and 3. The stable
nitroxide radical, 1b, was purchased (Aldrich). Syntheses of
the persistent iminoxyl radical, 2, and its precursor, di-tert-
butyl ketoxime, 2H, have been described. N-Hydroxyphthal-
imide, NHPI (3H, Aldrich), was used as the precursor for the
nonpersistent PINO (3) radical.
3
(
ii) Dideuterio-alkenes 4, 5, and 6. The synthesis of 4 has
11
been described. Alkenes 5 and 6 were synthesized by lithium
4
d
16
17
aluminum deuteride reduction of cyclooctyne and 3-hexyne
Aldrich), respectively. For details, see Supporting Information.
iii) Other. Di-tert-butyl hyponitrite and dicumyl hyponitrite
(
(
1
8
were synthesized by standard methods.
•
>
NO + Dideuterio-alkene Reactions. The alkene was
•
always used in large excess over the >NO radical to minimize
any further reactions of the initial products. In all cases, the
radical/alkene adducts were placed under a high vacuum
overnight to remove any traces of the alkenes. If present, these
(
iiia) 3. In our first investigation of the mechanism of
reaction of this imidoxyl radical with our three alkenes, we
oxidized NHPI (3H) with lead tetraacetate in the presence of
dideuterio-alkenes at room temperature in deoxygenated ac-
etonitrile, a solvent chosen to provide both reasonable solubili-
2
alkenes could, of course, confound appropriate analyses by H
NMR of the products and, hence, the determination of the
2
ties for NHPI and Pb(OAc)
alkene, NHPI, and Pb(OAc)
4
and easy product workup. The
were always reacted together at
molar ratios of 10:2:1. The alkene (49 mM 4, 76 mM 5, and 83
mM 6) and NHPI in 5 mL of acetonitrile were added to 5 mL
reaction mechanism. All H NMR spectra were obtained in
4
(
11) Coseri, S.; Ingold, K. U. Org. Lett. 2004, 6, 1641-1643.
(
12) Allylic C-H bonds may actually by slightly weaker than
4
of acetonitrile solutions of Pb(OAc) . Directly after mixing, the
comparable benzylic bonds; e.g., for primary C-H, the bond dissocia-
tion enthalpies (kcal/mol) are 88.8 ( 0.4 for propylene and 89.7 ( 0.6
for toluene. See Table 2 in: Blanksby, S. J.; Ellison, G. B. Acc. Chem.
Res. 2003, 36, 255-263.
acetonitrile and excess alkene were removed under reduced
pressure. The mono- and di-PINO alkene adducts were
separated and purified by preparative TLC using hexane/ethyl
acetate (2:1, v/v) as an eluent. Product yields based on NHPI
(
13) H-abstraction from ethylbenzene requires the free rotation
•
about the Ph-C bond to be “frozen out” in the transition state for PhC -
HCH radical formation. There are no such restrictions on cyclohexene.
were 4-PINO 17%, 4-(PINO)
1 ) 19%); 5-PINO 9%, 5-(PINO)
and 6-PINO 9%, 6-(PINO) 10% (NHPI reacted ) 29%). In
the case of cyclooctene, the normal H NMR analyses were
2
1% (NHPI reacted ) 17 + 2 ×
3
Moreover, two of the allylic C-H bonds in cyclohexene are perpen-
dicular to the CCHdCHC plane and are therefore well positioned for
C-H bond rupture with full allylic stabilization. In contrast, the two
benzylic C-H bonds in ethylbenzene lie at an angle of 30° to the
aromatic plane and are not well positioned for C-H rupture. These
factors are manifest in the rate constants for H-abstraction by the
2
16% (NHPI reacted ) 41%);
2
2
1
13
supplemented by H and C NMR and X-ray crystallography
on the products formed from nondeuterated cyclooctene (see
Supporting Information).
14
-1
corresponding peroxyl radicals at 30 °C, viz., cyclohexene 6.0 M
-
1
-1
-1
s
(
, ethylbenzene 1.3 M
s
, and 1,2,3,4-tetrahydronaphthalene
In view of the unusual pattern of products (see Results) and
-1
-1
where addition-abstraction is not possible) 6.4 M
s
.
19
4
the well-known fact that Pb(OAc) can, itself, oxidize alkenes,
(
(
(
14) Howard, J. A.; Ingold, K. U. Can. J. Chem. 1967, 45, 793-802.
15) Lemaire, E.; Rassat, A. Tetrahedron. Lett. 1964, 2245-2248.
16) Kroll, J. H.; Donahue, N. M.; Cee, V. J.; Demerjian, K. L.;
2
0
including, e.g., cyclohexene, we turned next to a system in
which free PINO radicals would be the unquestioned product
of a one-electron oxidation of NHPI.
Anderson, J. G. J. Am. Chem. Soc. 2002, 124, 8518-8519.
(
(
17) Brandsma, L.; Verkruijsse, H. D. Synthesis 1978, 290.
18) (a) Kiefer, H.; Traylor, T. G. Tetrahedron Lett. 1966, 6163-
6
168. (b) Dulog, L.; Klein, P. Chem. Ber. 1971, 104, 895-901. (c) Ogle,
(19) Review: Moriarty, R. M. In Selective Organic Transformations;
Thyagarajan, B. S., Ed.; Interscience: New York, 1972; pp 183-237.
(20) Anderson, C. B.; Winstein, S. J. Org. Chem. 1963, 28, 605-
606. Wiberg, K. B.; Nielsen, S. D. J. Org. Chem. 1964, 29, 3353-3361.
C. A.; Martin, S. W.; Dziobak, M. P.; Urban, M. W.; Mendenhall, G. D.
J. Org. Chem. 1983, 48, 3728-3733. (d) Quinga, E. M. Y.; Bieker, T.;
Dziobak, M. P.; Mendenhall, G. D. J. Org. Chem. 1989, 54, 2769-2771.
J. Org. Chem, Vol. 70, No. 12, 2005 4631

Coseri et al.
TABLE 1. Rate Constants for the Reactions of t-RO•
(
iiib) 3. NHPI was oxidized with tert-alkoxyl radicals
generated by the thermal decomposition of di-tert-alkyl hy-
Radicals with NHPI at 30 °C and with Three
6
b
Nondeuterated Alkenes at Ambient Temperatures
ponitrites (reactions 6 and 7). Pedulli and co-workers have
shown that reaction 7
0-6 k (M-1 s-1
1
)
substrate
NHPI
cyclohexene
cyclooctene
trans-3-hexene
this work
literature
ref
∆
•
t-RONdNOR-t
9
8 2t-RO + N
2
1050
4.6
5.7
2.6
21
22
t-R ) Me C, PhMe C
(6)
3
2
2.3
2.7
occurs because in the presence of NHPI, photochemically
2
6
absorption band at ∼485 nm with at least five different
concentrations of the three (nondeuterated) alkenes. Pseudo-
-
1
first-order rate constants (kexptl (s )) were determined using
digitally averaged decay traces from 5-10 laser flashes.
-
1
-1
Second-order rate constants, k
8
(M
s ) (see Results), were
calculated from the slopes of the straight lines obtained by
least-squares fitting of kexptl vs [alkene]; see Supporting
Information.
The LFP method could not be employed for reaction 7
because NHPI absorbed too strongly at the laser wavelength.
This rate constant was therefore obtained by thermally
decomposing dicumyl hyponitrite at 30 °C in benzene in the
presence of five different concentrations of NHPI for 14 days
•
generated t-RO radicals yield the characteristic PINO EPR
spectrum. The hyponitrite/NHPI/alkene experiment must, of
course, be designed to ensure that essentially the only radical
to react with the alkene is PINO. This can only be done with
a full knowledge of the rate constant for reaction 7 and the
rate constants for reaction of t-RO radicals with the three
alkenes (reaction 8), which includes both H-atom abstraction
and addition to the double bond.
•
(
to ensure complete decomposition of the hyponitrite, vide
•
supra). Under these conditions, reaction 7 (RO ) cumyloxyl)
and reaction 9 are in competition. The ratio of the yields
k8
•
t-RO + alkene
98 products
(8)
•
•
PhCMe O f PhCOMe + Me
(9)
2
There has been no measurement of k
7
, while values of k
8
are available only for the reaction of tert-butoxyl with non-
2
1
22
of the products of these two reactions, viz., [cumyl alcohol]/
7 8
deuterated 4 and 5. We therefore measured k and k for
all three nondeuterated alkenes (see later).
[acetophenone], when plotted against [NHPI] gave a straight
-
1
Two sets of hyponitrite/NHPI/alkene experiments were
carried out under oxygen-free conditions. In one, di-tert-butyl
hyponitrite was employed in refluxing benzene (to reflect the
temperatures of the 1b/alkene reactions). Under these condi-
line, the slope of which is k
7
/k
9
) 2760 M (see Supporting
5
Information). In benzene at this temperature, k
9
) 3.8 × 10
-
1 27
s
,
7
and hence k is readily obtained.
The LFP and competition kinetic data are presented in
1
8a
tions the half-life of the hyponitrite will be about 5 min. In
the other, dicumyl hyponitrite was employed in benzene at
room temperature. The reaction was run in the dark for 14
days, which is sufficient for >99% of the hyponitrite to
decompose and reflects the temperatures used in the NHPI/
4
Pb(OAc) experiments and those used in the iminoxyl, 2,
Table 1. For cyclohexene and cyclooctene, agreement with the
literature is satisfactory. Even with the most reactive alkene,
cyclohexene, the magnitudes of these rate constants ensure
that with equimolar NHPI and alkene (vide supra), at least
23
9
9.5% of the alkoxyls generated will react with NHPI and form
PINO (reaction 7) and only 0.5% will react with the alkene.
experiments.
In refluxing benzene, the reactants were NHPI, 1.35 mmol;
dideuterio-alkene, 1.35 mmol; and hyponitrite, 0.34 mmol
Results
•
(
which will yield slightly less than 0.68 mmol of t-RO because
2
4
of some in-cage coupling of the geminate radical pair). The
products of interest were again separated by preparative TLC
2
The H NMR spectra for 1,2-dideuterio-cyclohexene, 4,
and the products of its reaction with nitroxide 1b have
been presented previously.11 Figure 1 shows the H NMR
spectra for 1,2-dideuterio-cyclooctene, 5, and the products
4
as in the Pb(OAc) experiments.
2
In room-temperature benzene, the reactants were NHPI,
0
.4 mmol; dideuterio-alkene, 0.4 mmol; and hyponitrite, 0.125
2
mmol. The products were separated in the usual way.
Measurement of the Rate Constants for Reactions 7
and 8. The rate constants for the three reactions 8 were
measured in benzene at room temperature by 308 nm laser
flash photolysis (LFP) of dicumyl peroxide (0.13 M; at this
concentration, the OD is ∼0.3 at 308 nm). The LFP equipment
of its reaction with 1b and with the iminoxyl, 2. The H
NMR spectra for the products obtained when NHPI is
•
oxidized by t-RO radicals and by Pb(OAc)
4
in the pres-
2
ence of 5 are shown in Figure 2. Additional H NMR
spectra are available in Supporting Information. The
relevant structures and 2H NMR chemical shifts are
summarized in Table 2.
21,25
and general procedure have been adequately described.
The
decay of the cumyloxyl radical was monitored using its visible
Even a cursory examination of Table 2 reveals that the
two deuterium atoms labeled D and D for each pair of
a d
monoadducts have, as would be expected, identical
(
21) Paul, H.; Small, R. D., Jr.; Scaiano, J. C. J. Am. Chem. Soc.
978, 100, 4520-4527.
22) Encinas, M. V.; Scaiano, J. C. J. Am. Chem. Soc. 1981, 103,
393-6397.
23) Mendenhall, G. D.; Chen, H.-T. E. J. Phys. Chem. 1985, 89,
849-2851. See also: Snelgrove, D. W.; Lusztyk, J.; Banks, J. T.;
1
6
2
(
(
(26) Avila, D. V.; Lusztyk, J.; Ingold, K. U. J. Am. Chem. Soc. 1992,
114, 6576-6577. Avila, D. V.; Ingold, K. U.; Di Nardo, A. A.; Zerbetto,
F.; Zgierski, M. Z.; Lusztyk, J. J. Am. Chem. Soc. 1995, 117, 2711-
2718.
(27) Avila, D. V.; Brown, C. E.; Ingold, K. U.; Lusztyk, J. J. Am.
Chem. Soc. 1993, 115, 466-470.
Mulder, P.; Ingold, K. U. J. Am. Chem. Soc. 2001, 123, 469-477.
(
24) Kiefer, H.; Traylor, T. G. J. Am. Chem. Soc. 1967, 89, 6667-
6
671.
(
25) Kazanis, S.; Azarani, A.; Johnston, L. J. J. Phys. Chem. 1991,
9
5, 4430-4435.
4632 J. Org. Chem., Vol. 70, No. 12, 2005

Aminoxyl (Nitroxide), Iminoxyl, and Imidoxyl Radicals
2
FIGURE 2. H NMR spectra for the products of 1,2-dideuter-
•
iocyclooctene reaction with NHPI oxidized with t-RO in
benzene under reflux (top) and oxidized with lead tetraacetate
(
bottom). The peaks at 7.406 and 7.383 are due to natural
abundance CDCl
3
.
has a unique chemical shift. The product containing D
b
can be formed both by abstraction-addition and by
addition-abstraction (see Scheme 1). Fortunately, the
relative importance of the two mechanisms of formation
of monoadducts can be determined by integration of the
two signals, one arising from D
a
+ D
c
d
+ D and another
from D . Since D ) D (Scheme 1) we have:
b
a
b
(
D + D + D ) - D ) [D + D ]
a c d b c d
•
Because the >NO radical will add with equal prob-
ability to both ends of the allyl radicals, two of which
are completely symmetric, the total amount of the
abstraction-addition products is: 2[D
c
d
+ D ]. The per-
FIGURE 1. 2H NMR spectra for 1,2-dideuteriocyclooctene
centage of the monoadduct products from abstraction-
(
neat, top) and for the products of its reaction with 1b (middle)
addition is therefore given by equation I:
and 2 (bottom) in CHCl
3
. The peaks at 7.382, 7.310, and 7.265
are due to natural abundance CDCl
3
.
2
[D + D ]
c d
×
100 )
chemical shifts. The chemical shift for D
c
also overlaps
(D + D + D ) + D
a
c
d
b
1
those for D
a
and D
d
except for the 1/4 products.1
2
{(D + D + D ) - D }
3
a
c
d
b
However, the deuterium D
b
, which is bonded to the sp -
×
100 (I)
(
D + D + D ) + D
hybridized carbon bearing the radical moiety, n, always
a c d b
J. Org. Chem, Vol. 70, No. 12, 2005 4633

Coseri et al.
TABLE 2. Assignmentsa (and Relative Intensities for
b
2
Monoadducts, in Parentheses) of H NMR Signals in
CHCl3 of Radical (n)-Alkene Adducts, δ (ppm)c
FIGURE 3. X-ray crystallographic structures for cyclooctene
mono- (left) and diadduct (right) with the PINO moiety.
percentage of the total of all substitution products, i.e.,
equation II:
diadduct % yield ) 100 D /(D + D + D + D + D )
e
a
b
c
d
e
(
II)
That the diadducts are saturated30 organic compounds
2
is, of course, obvious from their H NMR spectra.
Nevertheless, the structure of one of them was further
confirmed by reacting NHPI, Pb(OAc) , and nondeuter-
4
ated cyclooctene under the usual conditions, but on a
larger scale. The mono- and diadducts were separated
1
(
TLC) in sufficient quantity to measure their H NMR
13
and C NMR spectra (see Supporting Information) and
to obtain X-ray crystallographic structures (Figure 3 and
Supporting Information).
Discussion
Formation of Monoadducts. The nitroxide, 1b, the
iminoxyl, 2, and the imidoxyl, PINO, when generated
•
from NHPI and t-RO radicals, react with the three
a
Assignments are based on ACD Labs NMR Prediction Soft-
b
2
alkenes mainly by an initial abstraction of an allylic
H-atom. The allyl radical so formed then couples with a
second >NO radical (see Scheme 1). In these nine
reactions, 88-92% of the monoadducts are formed by this
abstraction-addition mechanism (see Table 3), and the
remaining 12-8% are formed by the addition-abstrac-
tion mechanism. Such a high (and constant) percentage
of abstraction-addition raises the disturbing possibility
that abstraction-addition is the only operative reaction
mechanism. That is, the 88-92% value might be due to
some artifact such as a systematic H NMR peak inte-
gration error or incorporation of the deuterium into more
than just the two vinylic positions. The latter possibility
can be discarded on the basis of the clean H NMR
spectra of the starting alkenes (see, e.g., Figure 1 and
ware. For nonresolved H NMR peaks, the total relative intensity
c
2
is given only once. All H NMR peak positions are given in parts
2
•
per million relative to the H NMR internal standard arising from
the solvent’s natural abundance CDCl3, which has been set at δ
7
.34 ppm. d These data were obtained in refluxing benzene. The
results obtained at room temperature were essentially identical,
see Supporting Information. e Data from ref 11. f Not detected.
g
Yield of the disubstituted alkenes as a percentage of the total
product, i.e., 100 De/(Da + Db + Dc + Dd + De).
TABLE 3. Percentages of Monoadducts Formed by the
Abstraction-Addition Mechanism and Percentage Yields
of Diadducts as a Percentage of Total Adducts
2
radical
alkene % abstraction-additiona % diadduct
2
1
2
3
3
1
2
3
3
1
2
3
3
b
4
4
4
4
5
5
5
5
6
6
6
6
91
91
88
73
92
90
89
65
91
92
90
50
0
0
(t-RO•)
(Pb(OAc)4)
b
b
b
b
0
2
Supporting Information). A systematic error in the H
3
signal integrations (which were done both by machine
0
and by hand)29 appeared unlikely because the NMR pulse
0
(t-RO•)
(Pb(OAc)4)
b
0
2
time (2 s) was much more than 5 times the H T
1
relaxation time31 (∼100 ms). However, to be “on the safe
61
0
0
0
27
(
a c d b
29) Repeated machine integration of the D + D + D and D peaks
(t-RO•)
(Pb(OAc)4)
using different ranges of δ showed that random errors for the
abstraction-addition percentages listed in Table 3 are less than 2.5%.
In addition, integration by hand gave abstraction-addition percentages
fully consistent with those listed in Table 3.
a
Random error < (2.5%; see ref 29. b See footnote d to Table
2
.
(
30) Unsaturated diadducts containing the following structural
element were not detected in any system:
The results of these calculations are given in Table 3.28
Table 3 also includes the yields of diadducts as a
(
28) Regretfully, this percentage was incorrectly calculated for the
(31) Mantsch, H. H.; Saito, H.; Smith, I. C. P. Progress in NMR
Spectroscopy; 1977, Vol. II, Part 4, pp 211-271.
11
1
b/4 reaction in our original communication.
4634 J. Org. Chem., Vol. 70, No. 12, 2005

Aminoxyl (Nitroxide), Iminoxyl, and Imidoxyl Radicals
side”, the PINO radical was reacted with trans-3,4-
dideuterio-hex-3-ene on a sufficient scale to isolate the
diadduct. This compound was separately mixed with an
SCHEME 2
3
equimolecular quantity of the dideuterioalkene in CHCl .
2
The H NMR integrations were fully consistent with the
experimental 1:1 molarities of the adducts and alkene
(
see Supporting Information). We therefore conclude that
•
in all the >NO /alkene reactions under consideration,
there is good evidence that ca. 8-12% of the (monoad-
duct) product is formed by the addition-abstraction
mechanism, although it must be admitted that the
absence of diadducts, particularly for the reactions
involving 1b and 2 at the relatively high radical concen-
trations employed, is a worry and may indicate that we
are overlooking some aspect of the chemistry or of the
2
H NMR analytical procedure. However, in this connec-
(
OAc)
4
were mixed together and allowed to stand for 1,
tion, we note that other oxygen-centered radicals have
been demonstrated to add to alkenes as well as to
abstract allylic H-atoms. For example, there is 3.8%
3
, and 10 min prior to the addition of the cyclooctene,
with product workup taking place 2 min after the
addition of the alkene. In the second set of three experi-
•
32
addition in the t-RO + cyclohexene reaction, 4.4%
•
33
4
ments, the cyclooctene and Pb(OAc) were mixed and
addition in the ROO + cyclohexene reaction, and an
•
allowed to stand for 1, 3, and 10 min prior to the addition
of the NHPI, with product workup taking place 2 min
later. Analyses of the products from these six reactions
by CI-MS showed always the same monoadduct/diadduct
ratio, and furthermore, this ratio was the same as that
obtained with the original experiment protocol (see
Supporting Information). In none of the experiments was
there a CI-MS M + 1 peak corresponding to 2-acetoxy-
cyclooctene or a peak for 1,2-diacetoxycyclooctane.
amazing 71% addition in the ROO + cyclooctene reac-
3
3,34
tion.
The relative rate constants for the reactions of different
alkenes with a common radical are determined by the
interplay of entropic and enthalpic effects. The former
arise because in the reactions of alkenes there is either
a partial “freeing” or a partial “freezing” of a bond
rotation in the transition state. For this reason, entropic
effects favor acyclic over cyclic alkenes for addition
reactions but favor cyclic over acyclic alkenes for H-atom
abstractions. Enthalpic effects on the rates of addition
to, and H-atom abstraction from, cyclic alkenes are
determined, in part, by changes in ring-strain in the
transition states. Convoluted with “normal” ring-strain
in these H-atom abstractions will be the extent to which
the cyclic allyl radicals are, perforce, twisted away from
their preferred planar geometry. Since only two cyclic and
one acyclic alkene were examined and since the percent-
age yields of monoadducts formed by the abstraction-
addition mechanism are essentially identical (88-92%)
for reaction of the three alkenes with 1b, 2, and PINO
Our results prove that free PINO radicals cannot be
involved in the formation of the diadducts in the alkene/
NHPI/Pb(OAc) system. We therefore suggest a mecha-
4
nism based on that proposed for the lead tetraacetate
oxidation of cyclohexene to cis- and trans-1,2-diacetoxy-
cyclohexanes and 3-acetoxycyclohexene via a symmetric
intermediate (see Scheme 2).1
9,20,35,36
This mechanism accounts both for the formation of
PINO diadducts and for the reduced percentage of the
abstraction-addition mechanism for the three alkene/
4
NHPI/Pb(OAc) reactions relative to all the other alk-
ene-radical reactions (see Table 3). That is, the PINO
•
generated from NHPI with t-RO radicals, it would seem
monoadduct that is formed according to Scheme 2 gives
2
that the entropic and enthalpic effects on the rates of the
rise to the D
a
and D
b
H NMR signals. This is the product
•
two processes are roughly in balance for these nine >NO
that is formed by both the abstraction-addition and the
addition-abstraction mechanisms (see Scheme 1). The
product that is unique to the addition-abstraction mech-
+
alkene reactions.
Formation of Diadducts. The alkene/NHPI/Pb-
2
(OAc)
4
system yields diadducts, which are not formed in
anism gives rise to the D
(Scheme 1). Formation of the D
Scheme 2 will cause the calculated percentage of abstrac-
tion-addition to decrease (see eq I). Indeed, if all D were
formed according to Scheme 2, the calculated percentage
of abstraction-addition would be zero (as it should be).
That this percentage is not zero, implies that some of the
monoadduct products must be formed by PINO radicals.
c
and D
d
H NMR signals
•
the alkene/NHPI/t-RO system (see Tables 2 and 3). This
raises the question, is diadduct formation specific to the
experimental procedure? This involved mixing the alkene
and NHPI in deoxygenated acetonitrile followed by the
b
product according to
b
4
addition of Pb(OAc) in deoxygenated acetonitrile and
product workup (see Experimental Section). Using cy-
clooctene (because this gave the highest yields of diad-
duct) and deoxygenated acetonitrile as the solvent, we
carried out reactions with different experimental proto-
cols. In one set of three experiments, the NHPI and Pb-
(
35) Separate experiments showed that NHPI did not replace the
acetate group in acetoxycyclohexane.
36) We cannot rule out the possibility that the first steps in Scheme
(
2
involve the substitution of one or two acetate groups in the Pb(OAc)
4
(
32) Walling, C.; Clark, R. T. J. Am. Chem. Soc. 1974, 96, 4530-
by PINO moieties, reaction 10 and 11, followed by chemistry similar
4
534.
to that shown in Scheme 2.
(
33) Van Sickle, D. E.; Mayo, F. R.; Arluck, R. M. J. Am. Chem. Soc.
1
965, 87, 4824-4832.
NHPI + Pb(OAc)
4
a PINOPb(OAc)
3
+ AcOH (10)
•
(
34) For a review of ROO + alkene addition and abstraction
reactions, see: Mayo, F. R. Acc. Chem. Res. 1968, 1, 193-201.
NHPI + PINOPb(OAc)
3
(PINO)
2
Pb(OAc) + AcOH (11)
2
J. Org. Chem, Vol. 70, No. 12, 2005 4635

Coseri et al.
This is hardly surprising in view of the known15 produc-
tion of PINO radicals from NHPI and Pb(OAc)
The relative importance of the radical (Scheme 1) and
nonradical (Scheme 2) routes to PINO monoadducts can
be readily estimated. For example, trans-hex-3-ene has
the lowest calculated percentage of the abstraction-
addition mechanism, viz., 50% (Table 3). Using eq I and
acetonitrile oxidized benzyl alcohol quantitatively to
benzaldehyde without the formation of significant amounts
of benzoic acid. However, Espenson and co-workers
4
.
6
e
using the NHPI/Pb(OAc) system later reported that the
4
rate constant for reaction of PINO with benzyl alcohol
-
1
-1
(5.65 M s ) was lower than the rate constant for
-1
-1
PINO’s reaction with benzaldehyde 10.6 M s , imply-
ing that a clean conversion of benzyl alcohol to benzal-
dehyde would not be possible. More recently, Pedulli and
taking D
1.667, and since D
.667. This compound is probably formed solely via PINO
radicals, and it will be formed in equal amounts to the
+ D ) monoadduct arising from the PINO radical
reaction (Scheme 1). The PINO abstraction-addition
reaction therefore forms 0.667/2 ) 0.334 D . The remain-
der of the D
b
) 1.00, the calculated value of (D
a
+ D
c
+ D
d
)
)
0
a
) D , the (D + D ) monoadduct )
b
c
d
6b
co-workers reported a much higher rate constant for
-1 -1
the reaction of PINO with benzyl alcohol (28.3 M s ),
(D
a
b
7c,43
which would be consistent with Minisci’s observations.
Clearly, the NHPI/Pb(OAc) method for generating PINO
4
b
radicals should be treated with caution until these (and
b
signal, i.e., 1.00-0.33 ) 0.67, must therefore
43
other ) anomalies have been sorted out.
arise mainly from the Pb(IV)-mediated reaction together
with the small amount formed via the PINO radical
Acknowledgment. We thank Don Leek for valuable
NMR advice and Constantin Udachin for the X-ray
analyses. S.C. thanks the Natural Sciences and Engi-
neering Research Council (NSERC) of Canada for
receipt of a Visiting Fellowship.
addition-abstraction mechanism. From the trans-hex-
•
3
-ene/NHPI/t-RO data (Table 3), this small amount is
ca. 10% of the amount of monoadduct formed by abstrac-
tion-addition. Thus, for this alkene, {0.67 - (0.1 × 0.33)}
×
100 ) 64% of the monoadduct is formed according to
the nonradical Scheme 2, and the remaining 36% of the
monoadduct is formed via PINO radicals according to
Scheme 1.
Supporting Information Available: Experimental pro-
cedures for the syntheses of 5 and 6; LFP and competitive
1
13
kinetics results; H NMR and C NMR spectra for cyclooctene
Although iminoxyl radicals have been detected (EPR)
2
mono- and diadducts with the PINO moiety; H NMR spectra
in the reaction of Pb(OAc)
4
with a number of oximes,15
for the 2/4, 3/4, 1b/6, 2/6, and 3/6 reaction products, an
equimolecular mixture of the trans-3,4-dideuteriohex-3-ene
and trans-3,4-dideuteriohex-3-ene-PINO di-adduct, and neat
6; comparison of the CI-MS M + 1 peaks for the cyclooctene
reaction with NHPI in the presence of lead tetraacetate using
the two different protocols described in the text, and X-ray
crystallographic data for the cyclooctene/PINO mono- and
diadducts (CIF). This material is available free of charge via
the Internet at http://pubs.acs.org.
neither mono- nor diadducts of the iminoxyl, 2 were
formed from cyclooctene upon reaction of the ketoxime,
2
H, and Pb(OAc)
4
for 15 min. (The alkene/NHPI/Pb-
(
OAc) reactions yielded products in <1 min.) Similar
4
negative results were obtained with cyclooctene and the
hydroxylamine 1bH. Thus, reaction via the mechanism
in Scheme 2 (or analogous mechanisms) appears to be
confined to NHPI. This we attribute to the much greater
) of NHPI compared with 2H and 1bH.37 The
JO0504060
acidity (pK
absence of radical products in the alkene/2H/Pb(OAc)
and alkene/1bH/Pb(OAc) reactions we attribute to the
a
4
4
(
38) Green, A. L.; Sainsbury, G. L.; Saville, B.; Stansfield, M. J.
Chem. Soc. 1958, 1583-1587.
39) Some papers and databases give the pK
combination of two factors. The first factor is the slow
radical formation in these systems, which is probably due
(
a
for NHPI as 7.0
O/MeOH
without comment. However, this value was measured in H
2
to the difficulty of replacing acetate in Pb(OAc)
4
with the
(1:1, v/v); see: Ames, D. E.; Grey, T. F. J. Chem. Soc. 1955, 3518-
much less acidic 2H and 1bH.37 The second factor is that
the radicals, 2 and 1b, are considerably less reactive than
PINO. Monoadducts of 2 and 1b would presumably have
been formed by radical processes (Scheme 1) if the
reactions had been run for longer times and at higher
temperatures.
3521.
(
957, 79, 796-800.
40) Bissot, T. C.; Parry, R. W.; Campbell, D. H. J. Am. Chem. Soc.
1
(41) The compound that was actually titrated with base was Me -
2
4
0
NOH. HCl and two pKs were found: 5.20 and 8.80. Unfortunately,
the first pK was listed as pK
the pK of Me NOH is often given as 5.20.
(42) King, C. V.; Marion, A. P. J. Am. Chem. Soc. 1944, 66, 977-
80.
43) Pedulli et al. used the NHPI/t-RO system to generate PINO,
a b
and the second as pK . For this reason,
a
2
9
The intervention of some form of Pb(IV) in the NHPI/
2
,6b
•
(
alkene/Pb(OAc)
4
reaction introduces two quite different
•
with t-RO being generated photochemically from the peroxide, t-
ROOR-t. This allowed the measurement by kinetic EPR spectroscopy
of the PINO/benzyl alcohol rate constant. However, it would probably
not have been possible to measure the rate constant for PINO/
benzaldehyde reaction because of benzaldehyde’s strong UV absor-
bance. Interestingly, the rate constants for reaction of PINO with alkyl
channels for product formation, viz., radical and non-
radical reactions. A similar phenomenon may occur with
other substrates and may provide an explanation for a
puzzling kinetic anomaly in the literature. Thus, Minisci
6b
•
6e
7
c
aromatics measured by Pedulli using NHPI/t-RO and by Espenson
using NHPI/Pb(OAc) are in fairly good agreement. However, Espe-
nson and co-workers have very recently reported that the rate
constants for H atom abstraction by PINO from p-xylene and toluene
using NHPI/Co(III) were “slightly different” from those obtained using
NHPI/Pb(IV).
et al. reported that NHPI and a cobalt salt in aerated
4
8
c
40,41
(
37) NHPI (3H), pK
NOH and Et CdNOH, pK
last three compounds must be seen as surrogates for 1bH and 2H.
a
) 6.1;38,39 Me
2
NOH, pK
a
) 8.80;4
2
Me Cd
2
2
a
) 12.4 and 12.6, respectively, where the
4636 J. Org. Chem., Vol. 70, No. 12, 2005