Stereoselective coupling of prochiral radicals with a chiral C2-symmetric nitroxide

Article abstract of DOI:10.1021/ja000520u

The coupling reaction beween a chiral C₂-symmetric nitroxide, trans-2,5-dimethyl-2,5-diphenylpyrrolidin-1-oxyl (DPPO; 1), and a series of stabilized secondary prochiral radicals was studied to determine the factors that affect stereoselectivity. Both steric and electronic perturbations on the selectivity by the substituents of the prochiral radical were observed. The effects of temperature, solvent polarity, and solvent viscosity were examined. High selectivity for reactions carried out in solvents of low viscosity provides evidence for the formation of an encounter complex on the reaction path. Ab initio calculations on simplified model systems predict the C-O-N angle of attack to be greater than 110°at a carbon-oxygen bond-forming distance of approximately 2.2 A, although no transition state was found.

Full text of DOI:10.1021/ja000520u

J. Am. Chem. Soc. 2000, 122, 8421-8434
8421
Stereoselective Coupling of Prochiral Radicals with a Chiral
C2-Symmetric Nitroxide
Rebecca Braslau,*,† Neeta Naik,† and Hendrik Zipse‡
Contribution from the Department of Chemistry and Biochemistry, UniVersity of California,
Santa Cruz, California 95064, and the Department of Chemistry, Ludwig-Maximilians-UniVersita¨t,
81377 Mu¨nchen, Germany
ReceiVed February 11, 2000
Abstract: The coupling reaction between a chiral C2-symmetric nitroxide, trans-2,5-dimethyl-2,5-diphenylpyr-
rolidin-1-oxyl (DPPO; 1), and a series of stabilized secondary prochiral radicals was studied to determine the
factors that affect stereoselectivity. Both steric and electronic perturbations on the selectivity by the substituents
of the prochiral radical were observed. The effects of temperature, solvent polarity, and solvent viscosity were
examined. High selectivity for reactions carried out in solvents of low viscosity provides evidence for the
formation of an encounter complex on the reaction path. Ab initio calculations on simplified model systems
predict the C-O-N angle of attack to be greater than 110° at a carbon-oxygen bond-forming distance of
approximately 2.2 Å, although no transition state was found.
Introduction
optically active Lewis acids6). The third intermolecular reaction
possibility, coupling reactions between prochiral radicals and
chiral radical partners, has not been addressed.7 Nitroxides are
persistent species8 and are highly effective traps for carbon
radicals,9 with coupling rate constants on the order of 107 to
109 M-1 s-1 at room temperature.10 We have previously
published stereochemical studies on the coupling of chiral,
optically active steroid or terpene-derived nitroxides with
prochiral carbon radicals.11,12 Although moderate to good
stereoselectivity was observed, in the best cases, analysis of
the products was challenging due to dynamic behavior in the
NMR spectra of the N-alkoxyamine products. Herein we
describe stereoselectivity studies in the reaction of a chiral C2-
symmetric nitroxide with a variety of prochiral carbon radicals.
The ratios of the product N-alkoxyamines were readily deter-
mined by integration of the proton NMR spectra. This made
possible a series of experiments to probe the stereoselectivity
in terms of both steric and electronic effects by varying
substituents on the prochiral radical and examining the tem-
perature and solvent effects. As these reactions probe the
stereoselectivity of two radicals approaching each other without
Free radicals are highly reactive species which can be used
to effect organic transformations under mild conditions. Ad-
vantages of radical over ionic reactions include chemoselectivity
in the tolerance of a wide range of functionalities without the
need for protecting groups, the general absence of solvent
effects, and the predictive power afforded by numerous mea-
sured rate constants for radical reactions. Over the past two
decades, extensive use has been made of radical intermediates
in connective transformations to assemble target skeletons.1
More recently, advances in the control of stereochemistry in
radical reactions have been demonstrated.2,3 Not surprisingly,
radical intermediates are subject to the same types of confor-
mational analysis and predictive transition states as their ionic
counterparts. We have been engaged in a research program to
study the ability of nonbonded, noncoordinated chiral reagents
to discriminate between the two faces of a prochiral carbon
radical. Only a few stereochemical studies of intermolecular
reactions involving prochiral radicals have been reported, which
include hydrogen abstraction reactions from optically active
reagents4 and the addition of prochiral radicals to optically active
unsaturated substrates5 (or to activated alkenes coordinated to
(6) Mikami, K.; Yamaoka, M. Tetrahedron Lett. 1998, 39, 4501-4504.
(7) (a) For a discussion of diastereoselective radical-radical self-
dimerizations and stereoselectivity in geminate radical pair recombinations,
see: ref 2, pp 242-250. (b) Enantiomeric recoginition in radical recombina-
tion has been studied in the context of the 1,2-Wittig rearrangement:
Tomooka, K.; Igarashi, T.; Nakai, T. Tetrahedron Lett. 1993, 34, 8139-
8142.
† Department of Chemistry and Biochemistry, University of California,
Santa Cruz.
‡ Department of Chemistry, Ludwig-Maximilians-Universita¨t.
(1) (a) Curran, D. P. Synthesis 1988, 417-439. (b) Curran, D. P. Synthesis
1988, 489-513. (c) Curran, D. P. Chem. ReV. 1991, 91, 1237-1286.
(2) Curran, D. P.; Porter, N. A.; Giese, B. Stereochemistry of Radical
Reactions; VCH: New York, 1996.
(8) For a discussion of persistent radicals reacting with transient radicals,
see: Fischer, H. J. Am. Chem. Soc. 1986, 108, 3925-3927.
(9) Griller, D.; Ingold, K. U. Acc. Chem. Res. 1976, 8, 13-19.
(10) (a) Bowry, V. W.; Ingold, K. U. J. Am. Chem. Soc. 1992, 114,
4992-4996. (b) Chateauneuf, J.; Lusztyk, J.; Ingold, K. U. J. Org. Chem.
1988, 53, 1629-1632. (c) Arends, I. W. C. E.; Mulder, P.; Clark, K. B.;
Wayner, D. D. M. J. Phys. Chem. 1995, 99, 8182-8189.
(11) (a) Braslau, R.; Burrill, L. C.; Mahal, L. K.; Wedeking, T. Ang.
Chem., Int. Ed. Eng. 1997, 36, 237-238. (b) Braslau, R.; Burrill, L. C.;
Chaplinski, V.; Howden, R.; Papa, P. W. Tetrahedron: Asymmetry 1997,
8, 3209-3212.
(3) Sibi, M. P.; Porter, N. A. Acc. Chem. Res. 1999, 32, 163-171.
(4) (a) Nanni, D.; Curran, D. P. Tetrahedron: Asymmetry 1996, 7, 2417-
2422. (b) Blumenstein, M.; Schwarzkopf, K.; Metzger, J. O. Angew. Chem.,
Int. Ed. Engl. 1997, 36, 235-236. (c) Schwarzkopf, K.; Blumenstein, M.;
Hayen, A.; Metzger, J. O. Eur. J. Org. Chem. 1998, 177, 7-181. (d) Haque,
M. B.; Roberts, B. P. Tetrahedron Lett. 1996, 37, 9123-9126. (e) Haque,
M. B.; Roberts, B. P.; Tocher, D. A. J. Chem. Soc., Perkin Trans. 1 1998,
2881-2889. (f) Dakternieks, D.; Dunn, K.; Perchyonok, V. T.; Schiesser,
C. H. J. Chem. Soc., Chem. Commun. 1999, 1665-1666. (g) A tin hydride
bearing (L)-borneol prepared in situ: Schumann, H.; Pachaly, B.; Schuetze,
B. C. J. Organomet. Chem. 1984, 265, 145-152.
(12) Stereoselective coupling of acyclic chiral nitroxides with prochiral
radicals will be reported separately. Braslau, R.; Chaplinski, V. Manuscript
in preparation.
(5) (a) Fukuzawa, S.; Seki, K.; Tatsuzawa, M.; Mutoh, K. J. Am. Chem.
Soc. 1997, 119, 1482-1483. (b) Sibi, M. Personal communication.
10.1021/ja000520u CCC: $19.00 © 2000 American Chemical Society
Published on Web 08/12/2000

8422 J. Am. Chem. Soc., Vol. 122, No. 35, 2000
Braslau et al.
Scheme 1
Scheme 2
Table 1. Stereochemical Results from Coupling of Prochiral
Radicals Generated by CuCl2 Oxidation of Stabilized Carbanions
with the Nitroxide DPPO
the presence of chelating agents, such as metals or Lewis acids,
these studies provide an opportunity to study the fundamental
behavior of free radical reactivity in solution.
Results and Discussion
Choice of Chiral Nitroxide. The nitroxide trans-2,5-di-
methyl-2,5-diphenyl- pyrrolidin-1-oxyl (DPPO;13 1) was chosen
for this study. It was prepared and used in racemic form
following a slight modification of the synthetic scheme devel-
oped by Puts and Sogah13a,14 (see Scheme 18 in the Supporting
Information).
The C2-symmetry of the nitroxide ensures that the N-O π
bond has enantiotopic faces, thus simplifying the number of
stereochemical approaches in the key coupling reaction. Since
both faces of the N-O π bond are the same, there are only two
possibilities: reaction of the nitroxide from one or the other
enantiotopic face of the prochiral carbon radical (Scheme 1).
In reactions of a non-C2-symmetric chiral nitroxide, the nitroxide
three electron N-O π bond has two diastereotopic faces, each
of which will have different reactivity to the prochiral radical.
Because of the cyclic structure of DPPO, the nitroxide is
conformationally restrained and cannot exist as a collection of
rotamer populations. This simplifies the reaction pathway to
only two possible trajectories.
Generation of Prochiral Radicals. Prochiral radicals were
generated using stoichiometric methods under oxidizing condi-
tions which were compatible with the nitroxide functionality.
Three different methods were utilized: the oxidation of
stabilized carbanions with CuCl2,15 the Mn(salen) method
developed by Hawker,16 and the oxidation of benzyl hydrazines
with lead oxide.15
Metal-mediated oxidation of carbanions to radicals using
copper salts17 has been used synthetically, such as in the
oxidative coupling of enolate anions. Iron salts18 have been used
as well; for example, Jahn has utilized ferrocenium hexafluo-
rophosphonate as a mild single-electron oxidant to form carbon
radicals.19 In the case of the copper-mediated oxidative coupling
of two enolates, mechanistic studies by Porter20 indicate that
radical addition to a metal enolate is probably more realistic
than true radical-radical coupling. In previous studies from our
labs, 2,2,6,6-tetramethylpiperidinyloxy (TEMPO) has been
coupled with carbon radicals generated from lithium ester
enolates.15 Generation of prochiral radical 2 from ester or nitrile
substrates followed by coupling with DPPO gave the expected
N-alkoxyamine products (Scheme 2). In all experiments, the
chiral nitroxide was the limiting reagent. The lithium anions
were generated at -78 °C, using LDA in THF, and added by
cannula to a solution of DPPO and anhydrous CuCl2. In the
first experiment, using tert-butyl propionate, quenching of the
reaction mixture at -78 °C gave the same results as allowing
the reaction mixture to warm to room temperature overnight.
All subsequent reactions were allowed to warm overnight as
the standard protocol.
The selectivities observed in this series of coupling reactions
were modest; however, the diastereoselectivity ratios (dr) were
reproducible. In comparing the first two entries of Table 1, the
bulky tert-butyl ester substituted prochiral radical was much
more selective than the carbomethoxy analogue: dr ) 1.9:1
compared to 1.1:1. Comparing entries 2 and 3, substitution of
the methyl group by the larger cyclohexyl group again gave
increased selectivity: dr ) 2.0:1 (cyclohexyl) as compared to
(13) (a) Puts, R. S.; Sogah D. Y. Macromolecules 1996, 29, 3323-3325.
(b) Benfaremo, N.; Steenbock, M.; Klapper, M.; Mu¨llen, K.; Enkelmann,
V.; Cabrera, K. Liebigs Ann. Chem. 1996, 1413-1415. (c) Einhorn, J.;
Einhorn, C.; Ratajczak, F.; Gautier-Luneau, I.; Pierre, J.-L. J. Org. Chem.
1997, 62, 9385-9388.
(14) Kornblum, N.; Wade, P. A. J. Org. Chem. 1973, 38, 1418-1420.
(15) Braslau, R.; Burrill, L. C.; Siano, M.; Naik, N.; Howden, R. K.;
Mahal, L. K. Macromolecules 1997, 30, 6445-6450.
(16) Dao, J.; Benoit, D.; Hawker, C. J. J. Polym. Sci., Part A: Polym.
Chem. 1998, 36, 2161-2167.
(17) (a) Falck, J. R.; Mekonnen, B.; Yu, J. R.; Lai, J. Y. J. Am. Chem.
Soc. 1996, 118, 6096-6097. (b) You, S.; Neuenschwander, M. Chimia 1992,
46, 377-380.
(18) (a) Poupart, M.-A.; Lassalle, G.; Paquette, L. A. Org. Synth. 1990,
69, 173-179. (b) Ramig, K.; Kuzemko, M. A.; McNamara, K.; Cohen, T.
J. Org. Chem. 1992, 57, 1968-1969. (c) Schwartz, E. B.; Knobler, C. B.;
Cram, D. J. J. Am. Chem. Soc. 1992, 114, 10775-10784.
(19) Jahn, U. J. Org. Chem. 1998, 63, 7130-7131.
(20) (a) Porter, N. A.; Rosenstein, I. J. Tetrahedron Lett. 1993, 34, 7865-
7868. (b) Porter, N. A.; Su, Q.; Harp, J. J.; Rosenstein, I. J.; McPhail, A.
T. Tetrahedron Lett. 1993, 34, 4457-4460.

StereoselectiVe Coupling of Prochiral Radicals
J. Am. Chem. Soc., Vol. 122, No. 35, 2000 8423
Scheme 3
Scheme 4
Scheme 5
1.1:1 (methyl). Although the phenyl group has a larger “A
value”21 than the cyclohexyl group, it is flat: in the prochiral
radical, the π electrons will be coplanar with the SOMO in order
to delocalize the unpaired electron density. Thus, the effective
size of the phenyl substituent is less than the standard “A value”
for phenyl, which assumes a freely rotating aromatic ring.
Comparison of entries 4 and 5 examines the perturbation by
introducing an electron-donating substituent on the aromatic
ring. Substitution in the para position was chosen to minimize
steric effects upon introduction of a methoxy group. A small
increase of dr ) 1.6:1as compared to dr ) 1.4:1 was seen upon
introduction of this electron-donating group. In another example,
the very small nitrile group was used in place of an ester
substituent. Comparing entries 4 and 6 shows that the selectivity
decreases slightly in going from carbomethoxy, “A value” of
1.2, to the much smaller nitrile, “A value” of 0.2: dr ) 1.4:1
as compared to dr ) 1.3:1. The data from Table 1 imply that
increasing the steric bulk22 of one of the substituents on the
prochiral radical tends to lead to higher diastereoselectivity.
To probe the electronic contribution to the stereoselectivity,
prochiral radicals bearing para-substituted aromatic rings were
examined using the Mn(salen) method developed by Hawker
to prepare N-alkoxyamines as initiators for nitroxide mediated
“living” free radical polymerization.16 This method utilizes the
catalyst systems developed by Jacobsen and Katsuki for
asymmetric epoxidations.23 In the Hawker protocol, a manganese
oxo species reacts with styrene to generate a carbon radical 10,
which is trapped by a nitroxide before collapse to epoxide 11
can take place (Scheme 3). Sodium borohydride reduces the
coupled product 12 in situ to give the phenethyl N-alkoxyamine
and to free the manganese complex to return to the catalytic
cycle. In Hawker’s work, the commercially available optically
active Jacobsen manganese salen complex has been used.
However, in this study, the use of a chiral manganese complex
would make the two faces of the carbon radical 10 diastereotopic
rather than enantiotopic and could bias the stereoselectivity as
well as lead to kinetic resolution of the racemic DPPO. To make
the prochiral radical 10 stereochemically neutral, an achiral Mn-
(salen) catalyst 13 was prepared in one step24 from manga-
nese diacetate tetrahydrate and N,N′-bis(salicylidiene)ethylene-
diamine (see Scheme 19 in the Supporting Information). With
this achiral catalyst in hand, DPPO was coupled to an electron-
poor styrene, an electron-rich styrene, and the parent unsubsti-
tuted styrene (Scheme 4). These reactions were carried out at
room temperature.
As can be seen from the results, both the electron-with-
drawing nitro group (σF value ) -0.57)25 and the electron-
donating amine group (σF value ) +0.81) lead to a diminution
of the selectivity compared to that of the unsubstituted styrene
(RdH, σF value ) 0). To rule out the possibility that the aniline
nitrogen was coordinated to a boron species derived from the
sodium borohydride, the reaction was repeated with a stoichio-
metric amount of the Mn(salen) reagent, followed by a
subsequent reduction step with sodium borohydride (Scheme
5). The stereoselectivity was not significantly changed (dr )
3.0:1 as compared to dr ) 2.9:1).
Recent calculations by Svensson et al.26 on the Mn(salen)
epoxidation pathway indicate that reaction between the man-
ganese oxo species and alkene begins on the triplet surface,
followed by spin change to the quintet surface, which then leads
to the quintet epoxide complexed manganese product without
a barrier. This implies that intermediate 10 is not a typical benzyl
radical but rather has a significantly different reactivity profile.
Since no epoxide product was found in the presence of DPPO,
we assume that the nitroxide intercepts the triplet species before
spin-state crossing can occur. They also found that alkene
substituents influence the timing of the spin-crossing, with
electron-withdrawing substituents causing spin-crossing to occur
later in the reaction path. In epoxidation reactions, this allows
more carbon-carbon bond rotation to occur in intermediate 10,
(21) Eliel, E. L.; Wilen, S. H.; Mander, L. N. Stereochemistry of Organic
Compounds; Wiley: New York, 1994; p 695.
(22) For a study on the effect of steric bulk on the rate of nitroxides
coupling with carbon radicals, see: Iwao, K.; Sakakibara, K.; Hirota, M. J.
Comput. Chem. 1998, 19, 215-221.
(23) (a) Finney, N. S.; Pospisil, P. J.; Chang, S.; Palucki, M.; Konsler,
R. G.; Hansen, K. B.; Jacobsen, E. N. Angew. Chem., Int. Ed. Engl. 1997,
36, 1720-1723 and references therein. (b) Katsuki, T. Coord. Chem. ReV.
1995, 140, 189-214.
(24) Palucki, M.; Finney, N. S.; Pospisil, P. J.; Guler, M. L.; Ishida, T.;
Jacobsen, E. N. J. Am. Chem. Soc. 1998, 120, 948-954.
(25) March, J. AdVanced Organic Chemistry: Reactions, Mechanisms,
and Structure; Wiley: New York, 1992; p 280.
(26) Linde, C.; Åermark, B.; Norrby, P.-O.; Svensson, M. J. Am. Chem.
Soc. 1999, 121, 5083-5084.

8424 J. Am. Chem. Soc., Vol. 122, No. 35, 2000
Braslau et al.
Scheme 6
Scheme 8
Scheme 7
Table 2. Stereochemical Results from Coupling of Prochiral
Radicals Generated by Oxidation of Benzyl Hydrazines Derived
from Benzyl Bromides with the Nitroxide DPPO
leading to a mixture of cis and trans epoxides from a cis olefin
substrate. In our reactions, both electron-donating and electron-
withdrawing substituents on the aromatic ring are expected to
help stabilize a benzylic radical intermediate and will lower the
rotational barrier. Thus, the lower selectivities for the two
substituted styrene substrates may derive from nitroxide trapping
a less favorable mixture of rotational conformers compared to
the unsubstituted styrene substrate.
The third method used to generate stoichiometric radicals
under mild conditions is the lead dioxide oxidation of alkyl
hydrazines 17, utilized by Corey to add carbon radicals to nitroso
compounds.27 A diazene intermediate 18 forms, which decom-
poses to the carbon radical with the extrusion of dinitrogen
(Scheme 6). This is consistent with the work of Meyers, in which
a diazene formed by an alternate route, anionic desilylation,
decomposes to a free radical which has been trapped by
TEMPO.28
We have demonstrated that benzyl hydrazines are easily
formed from benzyl bromides by SN2 displacement using
sonication to facilitate mixing of the two reactants.15 Benzyl
bromides 19b-e, other than commercially available phenethyl
bromide 19a, were prepared via bromination of the correspond-
ing benzyl alcohols using the bromine 1,2-bis(diphenylphos-
phino)ethane (DIPHOS) procedure developed by Schmidt and
Brooks.29 Use of DIPHOS is advantageous over the more
commonly used triphenylphosphine, as the phosphine oxide
byproduct is easily removed by a simple precipitation-filtration
procedure, employing the prescribed solvent mixture. Thus, the
sensitive benzyl bromides can be obtained cleanly in high yield
without resorting to silica gel chromatography. However, in two
cases, the standard procedure gave 1,2-dibromides as the major
products. These presumably form by elimination of HBr from
the desired benzyl bromide followed by addition of bromine to
the styrene intermediate (Scheme 7). This undesired dibromi-
nation could be avoided by carrying out the reaction between
-15 and 15 °C instead of allowing the reactions to warm to
room temperature (see Table 4 in the Supporting Information).
The resulting benzyl bromides were then reacted with
hydrazine, followed by treatment with lead dioxide in the
presence of DPPO at -78 °C, and allowed to warm to room
temperature to give the N-alkoxyamine products (Scheme 8,
Table 2). The selectivities observed in the hydrazine/lead dioxide
method were much better than those observed with the LDA/
CuCl2 method, presumably due to the slower coupling rate of
benzyl versus R-ester-substituted radicals with nitroxides.30 The
parent 1-phenethyl radical gave a dr of 2.5:1. In entry 2, the
sterically very hindered di-ortho-methyl-substituted radical gave
a dr of only 1.5:1. This decrease in selectivity is a reflection of
slight pyramidalization due to severe steric interactions between
the two ortho-methyl substituents on the benzene ring with the
methyl and hydrogen at the prochiral radical center. The
pyramidalization causes the substituents on the prochiral radical
to be pointed away from the incoming nitroxide, in contrast to
the planar orientation of substituents on the 1-phenethyl radical.
The electron-poor p-trifluoromethyl-substituted benzyl radical
in entry 3 gave a slightly improved diastereoselectivity of 3.5:
1. The 1-tetralinyl and 1-indanyl radicals gave the best diaste-
reoselectivity of the series: dr ) 5:1. Structurally, the 1-tetralinyl
radical is very similar to the 1-phenethyl radical; the only
difference is that the “methyl group” is tied back into a rigid
ring. This enhanced selectivity of a cyclic radical is very similar
to that observed by Guindon in diastereoselective hydrogen
transfer reactions.31 He reported a significant increase in
diastereoselectivity when an acyclic radical was transformed
into an endocyclic radical intermediate by chelation with a Lewis
acid. Similarly, in this work, the endocyclic tetralinyl and
indanyl radicals were more selective than the acyclic 1-phenethyl
radical. As these cyclic radicals gave good diastereoselectivity,
the 1-tetralinyl case was chosen for further studies.
(27) (a) Corey, E. J.; Gross, A. W. Tetrahedron Lett. 1984, 25, 491-
494. (b) Corey, E. J.; Gross, A. W. J. Org. Chem. 1985, 50, 5391-5393.
(28) Myers, A. G.; Movassaghi, M. J. Am. Chem. Soc. 1998, 120, 8891-
8892.
(29) Schmidt, S. P.; Brooks, D. W. Tetrahedron Lett. 1987, 28, 767-
768.
We have recently prepared another interesting C2-symmetric
chiral nitroxide in our laboratory, trans-1,3-dimethyl-1,3-
(30) Keoshkerian, B.; Georges, M.; Quinlan, M.; Veregin, R.; Goodbrand,
B. Macromolecules 1998, 31, 7559-7561.
(31) Guindon, Y.; Liu, Z. P.; Jung, G. J. Am. Chem. Soc. 1997, 119,
9289-9290.

StereoselectiVe Coupling of Prochiral Radicals
J. Am. Chem. Soc., Vol. 122, No. 35, 2000 8425
Scheme 9
Scheme 12
Scheme 13
Scheme 10
ature was raised in increments: the disappearance of the
nitroxide signal is indicative of carbon radical formation.
The reaction was started at -78 °C; ESR spectra were
acquired every 30 min. After 2 h, no significant change in the
TEMPO signal had been detected. The temperature was raised
to -38 °C and the TEMPO signal monitored every 15 min for
2.5 h. Again, no change was detected in the nitroxide ESR
signal. The temperature was then raised to 2 °C, and the TEMPO
signal monitored every 10 min. At this temperature, the ESR
indicated an ∼50% decrease in nitroxide concentration every
30 min period. The temperature was then raised to 22 °C, and
ESR spectra were measured at 5 min intervals, revealing an
even more rapid decrease in the nitroxide signal. Thus, the
benzyl hydrazine reacts to generate benzyl radical at a temper-
ature somewhere between -38 and 2 °C. This result explains
the poorer diastereoselectivities observed for experiments started
at -78 °C and warmed to room temperature compared to
experiments held at 0 °C. In the -78 °C to room-temperature
experiments, the reaction does not proceed at temperatures much
lower than 0 °C; however, when the prochiral radical is finally
generated, the reaction vessel does not linger at 0 °C but quickly
approaches room temperature.
A similar ESR study was carried out for radicals generated
by the Cu(II) oxidation of stabilized lithium anions. In this case,
the coupling of the lithium enolate of methyl propionate with
TEMPO was examined (Scheme 13). This reaction was carried
out in a 0.2 mm o.d. ESR tube. At -78 °C, no change in the
nitroxide signal34 was observed after 2 h. At -38 °C, the
TEMPO signal was no longer a clean triplet but had broadened
out, possibly because of interaction with copper ion. Integration
of the nitroxide signal showed a 60% decrease, indicating that
the radical formed in the range between -78 and -38 °C.
Solvent Effects. Kinetic studies by Beckwith, Bowry, and
Ingold35 using both laser flash photolysis and the radical “clock”
method have shown that solvent polarity affects the rate of the
coupling reaction between nitroxides and carbon radicals. The
rates were found to be strongly correlated with nitroxide
solvation, as determined by the solvent’s effect on the nitrogen
hyperfine splitting in the ESR. Nitroxides have a significant
dipole moment, resulting from a contribution from the dipolar
canonical resonance structure which localizes the unpaired
electron on the nitrogen (Scheme 14). Polar solvents will
stabilize the dipolar form, enhancing the distribution of unpaired
spin on nitrogen. In order for coupling of a nitroxide with a
Scheme 11
diphenylisoindoline-1-oxyl (21), and showed that it reacted with
1-phenethyl radical at 0 °C to give 81% of the N-alkoxyamine
product with a dr ) 2.2:1.32 When it was coupled with
1-tetralinyl radical at 0 °C, the N-alkoxyamine product 22 was
obtained in a 3.2:1 diastereomeric ratio (Scheme 9).
Temperature Effects. The coupling of the 1-tetralinyl radical
with DPPO was studied to determine the effect of temperature
on the stereoselectivity. The reaction was carried out at four
different temperatures: -78 °C to room temperature, 0 °C, room
temperature, and at reflux in toluene. The results obtained are
shown in Scheme 10. The reaction run at a constant 0 °C was
more selective than the reaction started at -78 °C and allowed
to warm to room temperature. This same phenomenon was
observed in using the 1-(4-trifluoromethylphenyl)ethyl radical
(Scheme 11). This led us to question at what temperature carbon
radicals are formed under our reaction conditions. Thus, ESR
experiments were undertaken to monitor the disappearance of
the nitroxide signal.
ESR Studies with TEMPO. As a model reaction, 1-phen-
ethyl radical was generated from the corresponding benzyl
hydrazine in the presence of TEMPO at low temperature in an
NMR tube and monitored by ESR33 (Scheme 12). The temper-
(32) Braslau, R.; Chaplinski, V.; Goodson, P. J. Org. Chem. 1998, 63,
9857-9864.
(33) The variable temperature ESR spectra can be obtained from the
Supporting Information.
(34) The ESR spectra at -78 °C and -38 °C can be obtained from the
Supporting Information.
(35) Beckwith, A. L. J.; Bowry, V. W.; Ingold, K. U. J. Am. Chem. Soc.
1992, 114, 4983-4992.

8426 J. Am. Chem. Soc., Vol. 122, No. 35, 2000
Braslau et al.
Scheme 14
Scheme 16
Scheme 15
the other extreme, the very viscous ethylene glycol solvent
resulted in a far less selective reaction: dr ) 2.1:1.
To understand these findings, insight into the details of the
carbon radical-nitroxide coupling process is necessary. It is
well-known that the rates of stabilized carbon radical coupling
reactions with nitroxides are significantly slower than radical-
radical self-recombination processes, which are generally con-
sidered close to diffusion controlled. Even the rates of radical-
radical self-reactions are affected by solvent viscosity. Typically,
reaction rates decrease with more viscous solvents, providing
evidence for rate determination by formation of diffusive
encounter pairs.40 Fischer has probed radical-radical self-
reactions by studying the product distribution between dimer-
ization and disproportionation as a function of solvent viscosity.
The ratio between dimers and products of disproportionation is
controlled by the ability of the radical pairs to reorient
themselves between repeated collisions. For example, with tert-
butyl radicals, the geometry required for dimerization, coinci-
dence of the principal axis in a parallel orientation, is much
more restrictive than that required for disproportionation. As
solvent viscosity increases, reorientational motions are inhibited.
Thus, the ratio of disproportionation is much higher in viscous
solvents compared to dimerization.
In carbon radical coupling reactions with nitroxides, slower
than diffusion controlled reactions have been attributed to an
entropic barrier lying between the caged singlet radical pair and
the N-alkoxyamine product.35 Beckwith, Bowry, and Ingold
speculated that this entropic barrier arises from the need to
localize the unpaired electron on the nitroxide oxygen before
coupling can occur. With delocalized benzyl radicals, a com-
ponent of the barrier is also due to the need to localize spin at
the benzylic carbon before coupling can proceed. Earlier work
along these same lines using laser flash photolysis41 compared
nitroxide trapping of carbon radicals with similar steric demands
but differing resonance stabilization. They found that more
stabilized radicals react significantly more slowly. For example,
the rate of TEMPO trapping at 20 °C by benzyl radical was 4.9
× 108 M-1 s-1 and the rate by 1-naphthylmethyl radical was
8.2 × 107 M-1 s-1. In this case, the steric demands of the two
radicals are approximately the same, whereas the 1-naphthyl-
methyl radical has approximately 4.5 kcal/mol greater resonance
carbon radical to occur, polar solvent molecules must be
displaced, leading to slower rate constants in comparison to
reaction in a nonpolar solvent. In addition to polar solvent
effects, coupling reactions were found to be unexpectedly fast
in hydroxylic solvents.36 This was termed the “anomalous”
nitroxide radical trapping effect. It is postulated that a directional
hydrogen bond between the nitroxide and the solvent is
responsible for this rate enhancement, as stabilization of the
dipolar canonical resonance structure by hydroxylic solvents
can occur with an orientation that does not impede the carbon
radical coupling. Solvent polarity also has an influence on the
bond dissociation enthalpies of N-alkoxyamines, as determined
by photoaccoustic calorimetry.37 This is again due to the
difference in solvation between the N-alkoxyamine and the
nitroxide in polar solvents. Since both the rate of N-alkoxyamine
formation and N-O bond strength are affected by solvent
polarity, we examined the diastereoselectivity of the coupling
reaction in both hydroxylic and nonhydroxylic solvents of
varying polarity.
The 1-tetralinyl radical displayed good selectivity in coupling
with DPPO; thus, it was utilized for probing the effects of
solvent polarity (Scheme 15). Four solvents with similar
viscosity but increasing polarity38 were examined: toluene,
tetrahydrofuran, ethanol, and methanol. Surprisingly, no sig-
nificant trend was observed in the diastereoselectivity: the
average dr for all four solvents is 4.8:1, which is not far from
the value of the least polar of the group, toluene, or the most
polar of the group, methanol.
Solvent viscosity is also known to influence the rate of
nitroxide coupling with carbon radicals, although this effect is
generally smaller than the influence of solvent polarity.33 We
examined the diastereoselectivity in a series of solvents of
increasing viscosity39 (Scheme 16). Here the results were dra-
matic. In the least viscous solvent, diethyl ether, the selectivity
at room temperature was the best of the series: dr ) 5.2:1. At
(39) Viscosity values for (a) Et2O and toluene: Handbook of Chemistry
and Physics, 56th ed.; CRC Press: Boca Raton, FL, 1975; pp F-52-F-55.
(b) t-BuOH: Viswanth, D. S.; Natarajan, G. Data Book on the Viscosity of
Liquids; Hemisphere Pub. Corp.: New York, 1989. (c) THF: Carvajal, C.;
Tolle, K. J.; Smid, J.; Szwarc, M. J. Am. Chem. Soc. 1965, 87, 5548-
5553.
(40) Fischer, H.; Paul, H. Acc. Chem. Res. 1987, 20, 200-206.
(41) Chateauneuf, J.; Lusztyk, J.; Ingold, K. U. J. Org. Chem. 1988, 53,
1629-1632.
(36) Valgimigli, L.; Banks, J. T.; Ingold, K. U.; Lusztyk, J. J. Am. Chem.
Soc. 1995, 117, 9966-9971.
(37) Ciriano, M. V.; Korth, H.-G.; van Scheppingen, W. B.; Mulder, P.
J. Am. Chem. Soc. 1999, 121, 6375-6381.
(38) All values were obtained from the Handbook of Chemistry and
Physics, 56th ed.; CRC Press: Boca Raton, FL, 1975; pp E-56-E-57, except
for THF: Carvajal, C.; Tolle, K. J.; Smid, J.; Szwarc, M. J. Am. Chem.
Soc. 1965, 87, 5548-5553.

StereoselectiVe Coupling of Prochiral Radicals
J. Am. Chem. Soc., Vol. 122, No. 35, 2000 8427
stabilization. Even the least stabilized carbon radicals (primary
alkyl and cyclopropyl) react about 10 times slower than dif-
fusion-controlled reactions. By studying the temperature de-
pendence of these reactions, it was concluded that this is due
to an unfavorable entropy of activation. Mulder and Wayner10c
have also studied the rate constants for benzylic radicals
coupling with TEMPO. They too postulate the formation of a
loose-encounter complex followed by a solvent-dependent
rearrangement to form the N-alkoxyamine product. They noted
that the rate constants decrease as the N-O bond in the coupled
product becomes weaker, and concluded that both steric factors
and thermodynamic considerations influence the dynamics in
this complex recombination system. In recent work from the
Mulder labs investigating the thermal stability of N-alkoxy-
amines,37 the coupling reaction of benzyl radical with TEMPO
is again described as occurring through a reversible encounter
of the two radical species to from a short-lived intermediate.
Subsequent rearrangement through an activated process gives
the N-alkoxyamine. Fischer42 has recently studied the rate con-
stants of cumyl radical coupling with TEMPO under condi-
tions of kinetic isolation and determined an Ea ) 0.9 kcal/
mol with a reaction entropy of ∆S ) 38 kcal mol-1 K-1
.
Figure 1. Variable temperature 1H NMR spectra of the N-alkoxyamine
20b in CDCl3 showing sharpening of signals of the pyrrolidine ring
methyl groups at lower temperature.
Similarly, Scaiano43 has determined an Ea ) 2.3 kcal/mol for
1-phenethyl radical coupling with TEMPO using Ingold’s kinetic
data.10a
With this model of an intermediary encounter complex along
the pathway to the formation of the N-alkoxyamine, the higher
diastereoselectivity observed in the less viscous solvents for our
coupling reactions of the chiral DPPO nitroxide can be under-
stood. With less viscous solvents, the prochiral radical has
more opportunity to reorient itself in a looser-encounter com-
plex to achieve a more favorable coupling product, as com-
pared to that in a more viscous, tighter-encounter complex.
Clearly the rate of the coupling, which is influenced by solvent
polarity, is less important in the diastereoselectivity than the
looseness or tightness of the solvent cage around the encounter
complex.
Dynamic NMR of N-Alkoxyamine Products. Dynamic
NMR characteristics were observed in the room temperature
1H and 13C spectra of all N-alkoxyamines derived from DPPO.
The signals for the methyl and phenyl groups on the pyrrolidine
ring and the methylene of the pyrrolidine ring were shorter and
broader than normal signals and were seen as nonequivalent
pairs of peaks.44 This type of dynamic NMR behavior has been
previously reported for a number of N-alkoxyamines, based on
the piperidine45 and isoindoline46 ring systems. The temperature
dependence of these NMR spectra is attributed to a conforma-
tional process involving ring inversion, nitrogen inversion, and
rotation about the nitrogen-oxygen bond. The five-membered
rings of the isoindoline and the pyrrolidine-based N-alkoxy-
amines have a less pronounced conformational preference
compared to that of the six-membered piperidine rings. Thus,
Figure 2. Variable temperature 1H NMR spectra of the N-alkoxyamine
20b in DMSO-d6 showing coalescence of the pyrrolidine ring methyl
groups at high temperature.
the dynamic behavior is attributed to hindered nitrogen inversion
and nitrogen-oxygen bond rotation in the pyrrolidinine systems.
1
Multitemperature H NMR studies were carried out on one
of the DPPO diastereomeric coupling products, N-alkoxyamine
20b. The methyl groups on the pyrrolidine ring should be homo-
topic because of the C2-symmetry and, thus, would be expected
to show only one signal for each diastereomer. However, due
to hindered nitrogen inversion and hindered rotation around the
nitrogen-oxygen bond, four methyl signals were seen for the
mixture of two diastereomers. At room temperature in CDCl3,
the methyl groups of the minor diastereomer are seen as a set
of short, broad signals separated by 0.7 ppm, whereas the methyl
groups of the major diastereomer are seen together as a larger,
broad peak (Figure 1). When cooled to -25 °C, all of the methyl
signals sharpen, and the two peaks from the major diastereomer
can be discerned. In a separate series of experiments in DMSO-
d6, the temperature was raised over a series of steps (Figure 2).
At 115 °C, the methyl signals corresponding to the minor
(42) (a) Kothe, T.; Marque, S.; Martschke, R.; Popov, M.; Fischer, H. J.
Chem. Soc., Perkin Trans. 2 1998, 1553-1559. (b) Fischer, H. J. Polym.
Sci., Part A: Polym. Chem. 1999, 37, 1885-1901.
(43) Skene, W. G.; Belt, S. T.; Connolly, J. J.; Hahn, P.; Scaiano, J. C.
Macromolecules 1998, 31, 9103-9105.
1
(44) For a typical example of H and 13C NMR spectra illustrating this
dynamic behavior, see the Supporting Information.
(45) (a) Anderson, J. E.; Corrie, J. E. T. J. Chem. Soc., Perkin Trans. 2
1992, 1027-1031. (b) Anderson, J. E.; Casarini, D.; Corrie, J. E. T.;
Lunazzi, L. J. Chem. Soc., Perkin Trans. 2 1993, 1299-1304. (c) Brown,
T. M.; Cooksey, C. J.; Crich, D.; Dronsfield, A. T.; Ellis, R. J. Chem. Soc.,
Perkin Trans. 1 1993, 2131-2136.
(46) (a) Busfield, W. K.; Jenkins, I. A.; Thang, S. H.; Moad, G.; Rizzardo,
E.; Solomon, D. H. J. Chem. Soc., Chem. Commun. 1985, 1249-1250. (b)
Beckwith, A. L. J.; Bowry, V. W.; Moad, G. J. Org. Chem. 1988, 53, 1632-
1641.

8428 J. Am. Chem. Soc., Vol. 122, No. 35, 2000
Braslau et al.
Table 3. Reaction Energy for the Recombination of Methyl
Scheme 17
Radical 23 with Nitroxide Radical 24 To Give Methoxyamine 25a
level of theory
UB3LYP/SVP
UB3LYP/SVP + ∆ZPE
UCCSD(T)/cc-pVDZ//B3LYP/SVP
UCCSD(T)/aug-cc-pVDZ//B3LYP/SVP
UCCSD(T)/cc-pVTZ//B3LYP/SVP
UCCSD(T)/aug-cc-pVTZ//B3LYP/SVP
UB3LYP/6-31G(d)
UB3LYP/6-31G(d) + ∆ZPE
UB3LYP/DZVP
UB3LYP/DZVP + ∆ZPE
UBLYP/DZVP
UBLYP/DZVP + ∆ZPE
∆Erxn
-56.9
-49.3
-60.0
-59.4
-61.9
-61.9
-58.6
-51.2
-56.7
-48.2
-55.4
-48.2
diastereomer coalesced. At 174 °C, each diastereomer displays
a single, sharp methyl signal.
a All energies are in kcal/mol.
It is noteworthy that the two diastereomers are expected to
interconvert at temperatures in the 85-100 °C range, by way
of bond homolysis,47 to form nitroxide and secondary benzyl
radical, followed by recombination. This is the underlying basis
for nitroxide-mediated “living” free radical polymerization,
which has been studied by Puts and Sogah13a using the nitroxide
DPPO. In their work, polymerization of styrene with DPPO was
investigated both experimentally and by semiempirical PM3
calculations on the coupling reaction with 1-phenethyl radical.
Experimentally, they found that polymerization proceeded (albeit
slowly) at 90 °C but was completely suppressed at 75 °C. Their
calculations gave a rough estimate of ∆(∆Hq) ) 1.08 kcal/mol
for the difference in enthalpies between the formation of the
(R,R)-S and (R,R)-R diastereomers of 20a, with the later being
favored. Making a crude assumption that ∆(∆Sq) ) 0 for the
formation of the two diastereomers, this value predicts the
formation of diastereomers in a 7.4:1 ratio at 0 °C, a 6.2:1 ratio
at room temperature, and a 4.1:1 ratio in refluxing toluene. Our
results gave a ratio of 2.5:1 for the reaction started at -78 °C
and allowed to warm to room temperature, in which the coupling
probably occurred over an actual range of 0 °C to room
temperature. The results of 1-tetralinyl radical coupling with
DPPO in reactions held at constant temperatures (a 6.2:1 ratio
at 0 °C, a 4.7:1 ratio at room temperature, and a 1.5:1 ratio in
refluxing toluene) indicate that the value derived by Puts and
Sogah is not unreasonable.
Theoretical Studies on the Geometry of Approach. As a
carbon radical and a nitroxide come together to form an
N-alkoxyamine, the carbon radical approaches the three-electron
π-bond of the nitroxide in a pathway reminiscent of the Bu¨rgi-
Dunitz trajectory for nucleophilic addition to a carbonyl
π-bond.48 To probe the energetics and geometry of approach
of carbon radicals reacting with nitroxides, a series of calcula-
tions were carried out.
To evaluate which theoretical methods give a reliable
description of the recombination reaction between nitroxides
and carbon-centered radicals, we first took a detailed look at a
small model system. Thus, the recombination of methyl radical
23 and nitroxide radical 24 to give methoxy amine 25 was
examined (Scheme 17).
combination with the polarized VDZ basis set by Ahlrichs et
al.50,51 In keeping with the terminology used in Gaussian 98,
this level of theory will be termed “B3LYP/SVP”. A similar
level of theory has recently been used by Tordo et al.52 in studies
of the H2NO•
+
•H recombination reaction. The reaction
energies of 23 coupling with 24 calculated at this level of theory
are -49.3 and -56.9 kcal/mol with and without accounting for
differences in zero-point vibrational energy (Table 3). Using
these geometries, a more reliable estimate of the reaction energy
was obtained by single-point calculations at the UCCSD(T) level
of theory in combination with various correlation consistent basis
sets by Dunning et al.53 Quite surprisingly, the results obtained
at the coupled cluster level are quite similar to those obtained
with B3LYP/SVP. Moreover, the CCSD(T) results appear to
depend only marginally on the choice of the basis set. The best
estimate of the reaction energy at 0 K is obtained at the UCCSD-
(T)/aug-cc-pVTZ//B3LYP/SVP level of theory in combination
with the unscaled B3LYP/SVP zero-point vibrational energy
and amounts to ∆Erxn(0 K) ) -54.3 kcal/mol.
The basis set dependence of the B3LYP result was also tested
by choosing two smaller basis sets: the economical 6-31G(d)
basis set, as well as the even smaller DZVP basis set specifically
optimized for density functional methods by Godbout et al.54
The reaction energies calculated with these smaller basis sets
are still very close to the B3LYP/SVP result; thus, we conclude
that, at least the reaction energies can be calculated with good
accuracy with even a medium-sized basis set. An even more
(49) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652. (b) Lee,
C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785-789. (c) Hertwig,
R. H.; Koch, W. J. Comput. Chem. 1995, 16, 576-585.
(50) Schaefer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97, 2571-
2577.
(51) All calculations have been performed with Gaussian 98, Revision
A.7, Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.;
Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A.
D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi,
M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.;
Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick,
D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.;
Ortiz, J. V.; Baboul, A. G.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz,
P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-
Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe,
M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.;
Gonzalez, C.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian,
Inc.: Pittsburgh, PA, 1998.
(52) Marsal, P.; Roche, M.; Tordo, P.; de Sainte Claire, P. D. J. Phys.
Chem. A 1999, 103, 2899-2905.
(53) (a) Dunning, T. H. Jr. J. Chem. Phys. 1989, 90, 1007-1023. (b)
Kendall, R. A.; Dunning, T. H. Jr.; Harrison, R. J. J. Chem. Phys. 1992,
96, 6769-6806. (c) Davidson, E. R. Chem. Phys. Lett. 1996, 220, 514-
518.
Of particular interest were those methods that not only yielded
reliable reaction energetics but also allowed for geometry opti-
mization of larger model systems. A first test of the performance
of various methods was the correct prediction of the reaction
energy for the recombination of 23 and 24. As no experimental
value is currently available for this system, all reactants were
optimized using the hybrid density functional Becke3LYP49 in
(47) Moad, G.; Rizzardo, E. Macromolecules 1995, 28, 8722-8728.
(48) Eliel, E. L.; Wilen S. H.; Mander, L. N. Stereochemistry of Organic
Compounds; Wiley-Interscience: New York, 1994; p 877.
(54) Godbout, N.; Salahub, D. R.; Andzelm, J.; Wimmer, E. Can. J.
Chem. 1992, 70, 560-571.

StereoselectiVe Coupling of Prochiral Radicals
J. Am. Chem. Soc., Vol. 122, No. 35, 2000 8429
performed using the UBLYP geometries. The UCCSD(T) level
of theory has performed exceedingly well in earlier studies of
bond dissociation processes, even in combination with medium-
sized basis sets.55 The UCCSD(T) dissociation curve is slightly
steeper as compared to that of the UBLYP level of theory, but
also flattens out after reaching the energy of the separate
products (shown as a dashed line in Figure 3). The energy of
the system at 2.4 Å exceeds that of the separate reactants by
1.1 kcal/mol and, thus, indicates a slight recombination barrier.
However, the accuracy of the UCCSD(T) level of theory is of
similar magnitude to this 1.1 kcal/mol barrier. Thus, we conclude
that all theoretical methods examined in this study indicate that
recombination of radicals 23 and 24 proceeds with a negligible
enthalpic barrier. One of the important structural characteristics
of the C-O bond formation process concerns the angle of attack
(lower part of Figure 3). Both the RBLYP and the UBLYP
methods agree, in that the C-O-N angle is larger than 105°
throughout the addition process. The angle of attack varies
around 117° at large C-O distances, and the two methods only
differ in how quickly this initial value approaches smaller values
upon shortening the C-O bond.
How relevant are these findings for recombination reactions
involving sterically demanding chiral nitroxides? To answer this
question, to at least some extent, we have also studied the
reaction between methyl radical 23 and nitroxide 26 (Scheme
17). Nitroxide 26 was chosen as a simple model for the chiral
steroidal doxyl nitroxide used in our earlier experiments.11a
These radicals recombine to give product 27. The main
difference between this and the first reaction is the enhanced
steric bulk present by virtue of the four methyl substituents in
26. As a consequence of these steric effects, the reaction
exothermicity is reduced to -42.4 kcal/mol at the BLYP/DZVP
level of theory. Accounting for differences in zero-point
vibrational energy, this value is reduced to -36.3 kcal/mol.
Using the latter value, this represents a reduction of around 12
kcal/mol as compared to the recombination reaction involving
nitroxide 24.
To determine whether the steric effects would be sufficiently
large to cause formation of an enthalpic barrier along the
recombination pathway, the recombination of 23 and 26 was
followed from a C-O distance of 3.0 Å, using the same
theoretical approach as before (Figure 4). The RBLYP/UBLYP
instability occurs at the same C-O distance as seen in the
previous case. In this instance we did not follow the RBLYP
curve beyond 2.5 Å, but instead concentrated on the UBLYP
results. For C-O distances larger than 2.5 Å, the UBLYP
energies are almost identical to the sum of the energies of the
two reactants (shown as a dotted line in Figure 4). Even with
the enhanced steric bulk present in 26, no transition state could
be found for the recombination process on the potential energy
surface. Variations in the C-O-N angle along the recombina-
tion pathway appear to be similar to that seen with 24 (Figure
3) and 26 (Figure 4). It is only at the longer C-O bond distances
around 3.0 Å that the angle of attack is somewhat larger for
26, with values around 130°. If one assumes that an enthalpic
barrier, if it exists, would be located at C-O distances around
2.2 Å,56 it is safe to conclude that the angle of attack would be
larger than 110°. A three-dimensional representation of 27 at
this bond distance is shown in Figure 5. The C-O-N bond
Figure 3. Relative energies as well as C-O-N angle deformation
along the pathway for dissociation of 25 into methyl radical 23 and
nitroxide radical 24.
economical procedure involves the combination of a gradient-
corrected pure density functional without admixture of Hartree-
Fock exchange, such as the BLYP functional with the DZVP
basis set. The reaction energy calculated at this level of theory
is only slightly smaller than with the B3LYP functional, and
we have therefore decided to perform all calculations for the
larger systems at this level of theory.
The next important consideration concerns the recombination
process. It is widely known that a correct description of the
recombination process requires a multiconfigurational treatment.
However, such an approach is not feasible for geometry
optimizations of even moderately large systems. We have
therefore used the same density functional methods employed
before to follow the reaction pathway for recombination of 23
and 24.
Because we did not expect to find a transition state for the
recombination process, we started the investigation of the
potential energy surface by incrementally stretching the C-O
bond in product 25 (0.1 Å per step) while all other structural
parameters were allowed to vary. The relaxed potential energy
scan obtained in this way at the BLYP/DZVP level of theory is
shown in Figure 3. Using restricted Kohn-Sham orbitals as in
RBLYP, the energy of the system increases steadily, surpassing
the energy of the separate reactants 23 and 24 (shown as a dotted
line in Figure 1) beyond a C-O distance of approximately 2.5
Å. It is in this region that the use of unrestricted Kohn-Sham
orbitals starts to lead to an energetically more favorable solution.
Moreover, with increasing C-O bond distance, the UBLYP
curve smoothly approaches the energy of the separate reactants
(shown as a dotted line in Figure 3). On the downside, the
UBLYP solution ceases to describe a singlet state beyond 3.0
Å, as the <S2> expectation values are larger than 1.0 at these
distances. A very similar picture emerges upon using the hybrid
B3LYP functional (not shown), and it therefore appears that
the use of a hybrid density functional does not yield a decisive
advantage in this case. To obtain more reliable reference values,
a series of UCCSD(T)/cc-pVDZ single-point calculations were
(55) Chen, W.; Schlegel, H. B. J. Chem. Phys. 1994, 101, 5957-5968.
(56) This choice is motivated by the CCSD(T) results which indicate
that the onset of C-O bond formation is located at around 2.3 Å. The
onset for bond formation in larger model systems occurs at slightly shorter
bond distances as compared to the minimal system at the BLYP level of
theory.

8430 J. Am. Chem. Soc., Vol. 122, No. 35, 2000
Braslau et al.
secondary prochiral radicals, sterics make a significant contribu-
tion to the selectivity; however, there is a small but real
electronic component. The effect of temperature was exam-
ined: higher stereoselectivity was obtained when the reaction
was carried out at 0 °C than when it was started at -78 °C and
allowed to warm to room temperature. ESR studies indicate that
the generation of carbon radicals using the hydrazine method
takes place at noticeable rates at 0 °C and higher.
The effects of solvent polarity and viscosity were examined.
Although nitroxide trapping rates are affected by polar and protic
solvents, this rate differential did not manifest itself into a change
in stereoselectivity. The diastereoselectivity for the coupling of
DPPO with tetralinyl radical was relatively constant at 4.8 (
0.3 for toluene, THF, EtOH, and MeOH solvents. However,
viscosity does play a role in stereoselectivity; selectivity was
highest in nonviscous solvents (diethyl ether dr ) 5.2:1) and
was much poorer in highly viscous ethylene glycol (dr ) 2.1:
1). This viscosity dependence is interpreted as supporting
evidence for the formation of an encounter complex in these
non-diffusion-controlled reactions. In a less viscous solvent, the
prochiral radical has more opportunity to reorient itself in a
looser encounter complex to produce the more favorable
coupling product as compared to a tighter encounter complex
formed in a more viscous solvent.
Ab initio modeling studies predict that the C-O-N angle
of attack of the carbon radical adding to the nitroxide is
somewhat greater than 110° at a critical carbon-oxygen bond-
forming distance of approximately 2.2 Å, although no transition
state could be found. The angle of approach of the carbon radical
adding to the N-O π bond of the nitroxide is very similar to
that of a nucleophile adding to a carbonyl π bond.
Figure 4. Relative energies as well as C-O-N angle deformation
along the pathway for dissociation of 27 into methyl radical 23 and
nitroxide radical 26.
Experimental Section
General. All reactions were run under nitrogen. Solvents were dried
under nitrogen as follows: THF, diethyl ether, and toluene were distilled
from sodium benzophenone. Dichloromethane was distilled from
calcium hydride. Methanol was distilled from magnesium metal. Fuming
hydrazine (anhydrous) was purchased from FisherBiotech (sequencing
grade). Copper(II) chloride was dried over P2O5 by heating at 150 °C
under house vacuum overnight. Reactions requiring sonication were
performed by placing the reaction flask into a Fisher Scientific solid
state/ultrasonic FS-14 bath cleaner; the height of the flask was adjusted
so that the solution received maximum agitation. Flash chromatography
was performed using Universal Scientific Inc. Silica Gel 63-200. IR
spectra were recorded in deuterochloroform solution. Mass spectra were
obtained using fast atom bombardment (FAB) on a VG ZAB-SE reverse
geometry spectrometer with a VG 11/250 data system at the University
of Illinois, and at UC Santa Cruz using a triple-quadrupole ESI
spectrometer. Semipreparative HPLC was carried out using a Waters
instrument with UV detection at 254 nm on a 150 × 4.6 mm Hasil 5
µm silica column. Elemental analysis was carried out by M-H-W
Laboratories in Phoenix, AZ. Melting points are uncorrected.
CuCl2 Oxidation of Lithium Enolates to Generate Prochiral
Carbon Radicals: Coupling with DPPO. The following procedure
is a typical example:
tert-Butyl 2-(trans-2,5-Dimethyl-2,5-diphenylpyrrolidine-1-oxy)
Propionate (3). A solution of tert-butyl propionate (61 µL, 0.400 mmol)
in 1.4 mL of THF was cooled to -78 °C, and 2 M LDA in heptane/
THF/ethylbenzene (266 µL, 0.532 mmol) was added. The mixture was
stirred at -78 °C for 4 h. The C2-symmetric nitroxide trans-2,5-
dimethyl-2,5-diphenylpyrrolidine-1-oxyl (70.8 mg, 0.266 mmol) and
CuCl2 (anhyd, 57.3 mg, 0.426 mmol) were suspended in 1.7 mL of
THF and cooled to -78 °C, and the enolate solution was added
dropwise by cannula. The residues of enolate were washed with two
Figure 5. Structure of N-alkoxyamine 27 at a C-O bond distance of
2.2 Å.
angle at this stage is 117.2°. In comparison, Puts and Sogah,
using PM3 calculations, found a transition state at a C-O
distance of 2.07 Å and a C-O-N bond angle of 123.0° for
DPPO, adding to 1-phenethyl radical.
It is interesting to compare our results to the classic work of
Bu¨rgi, Dunitz, Lehn, and Wipf, in which they used early ab
initio calculations to look at the reaction pathway for the addition
of hydride to the carbonyl group of formaldehyde.57 They found,
that at H-C distances of 2.5 Å, the H-C-O angle of attack
was 126°, which decreased to 119° at 2.0 Å, to 118° at 1.5 Å,
and finally to 109.5° at 1.12 Å. As in the carbon radical +
nitroxide case, the hydride + formaldehyde calculations found
no enthalpic barrier. Similarly, calculations on the addition of
ammonia to formaldehyde at 2.0 Å gave an N-C-O angle of
105°, which compared favorably with bond angles of 107 ( 5°
at 1.99 Å extracted from a series of crystal structures of amino
carbonyl compounds.58
Conclusions
In the stereoselective coupling reaction between the sterically
demanding chiral DPPO nitroxide and a series of stabilized
(57) (a) Dunitz, J. D.; Lehn, J. M.; Wipf, G. J. Am. Chem. Soc. 1974,
96, 1956-1957. (b) Bu¨rgi, H. B.; Dunitz, J. D.; Lehn, J. M.; Wipf, G.
Tetrahedron 1974, 30, 1563-1572.
(58) (a) Bu¨rgi, H. B.; Dunitz, J. D.; Shefter, E. J. Am. Chem. Soc. 1973,
95, 5065-5067. (b) Bu¨rgi, H. B.; Dunitz, J. D. Acc. Chem. Res. 1983, 16,
153-161.

StereoselectiVe Coupling of Prochiral Radicals
J. Am. Chem. Soc., Vol. 122, No. 35, 2000 8431
0.5 mL portions of THF. The mixture was stirred at -78 °C overnight.
It was then quenched with 2 mL of saturated NaHSO4 at -78 °C,
diluted with 15 mL of ether, and washed with 5 mL of saturated
NaHSO4 and 8 mL of saturated NaCl. The ether layer was dried over
magnesium sulfate and filtered, and the volatiles were removed in vacuo
to give a yellow oil which was purified by flash column chromatography
(2 cm column, 95:5 hexane/ethyl acetate gradient) to yield 84 mg of
colorless oil (80% yield). The major and minor diastereomers were
obtained as a mixture in a 1.9:1 ratio, as elucidated by integration of
s), 144 (minor, s), 143.23 (major, s), 128.89 (d), 128.69 (d), 127.96
(d), 127.77 (d), 127.62 (d), 126.73 (d), 126.36 (d), 126.09 (d), 125.87
(d), 88.65 (minor, d), 86.77 (major, d), 71.98 (minor, s), 70.18 (minor,
s), 68.39 (major, s), 67.41 (major, s), 51.15 (major, q), 50.89 (minor,
q), 41.14 (major, t),40.93 (minor, t), 40.58 (minor, d), 40.47 (major,
d), 34.49 (minor, t), 33.17 (major, t), 30.30 (major, q), 29.66 (minor,
q), 29.30 (major, t), 29.19 (minor, t), 28.76 (major, t), 28.52 (minor,
t), 26.47 (minor, t), 26.41 (major, t), 26.28 (t), 26.22 (t), 26.06 (t),
26.06 (t), 25.36 (minor, q), 22.49 (major, q). Elemental analysis: Anal.
Calcd for C27H35NO3: C, 76.92; H, 8.37; N, 3.32. Found: C, 76.88;
H, 8.12; N, 3.25.
1
the 500 MHz H NMR spectrum in CDCl3. Further purification was
done by HPLC to obtain analytically pure samples of each diastereomer.
TLC: 95:5 hexane/ethyl acetate, UV, p-anisaldehyde stain, Rf ) 0.4.
IR (CDCl3): major diastereomer 2980, 2934, 1735, 1601, 1492, 1445,
1360, 1275, 1224, 1146, 1087, 1030 cm-1; minor diastereomer 2981,
2936, 1737, 1601, 1492, 1445, 1370, 1265, 1222, 1162, 1134, 1084,
1030 cm-1. 1H NMR (500 MHz, CDCl3): major diastereomer δ 7.22-
7.83 (m, 10H), 3.86 (q, 1H, J ) 7), 2.53-2.57 (m, 1H), 2.18-2.22
(m, 1H), 1.93-2.07 (m, 2H), 1.68 (s, 3H), 1.51 (s, 9H), 1.12 (s, 3H),
0.90 (d, 3H, J ) 7); minor diastereomer δ 7.17-7.65 (m, 10H), 4.02
(q, 1H, J ) 7), 2.54-2.60 (m, 1H), 2.16-2.24 (m, 1H), 2.02-2.08
(m, 1H), 1.88-1.96 (m, 1H), 1.51 (s, 3H), 1.41 (s, 3H), 1.23 (s, 9H),
1.07 (d, 3H, J ) 7). 13C NMR (125 MHz, APT, CDCl3): major
diastereomer δ 173.3 (s), 151.88 (s), 143.2 (s), 128.73 (d), 127.96 (d),
127.57 (d), 126.6 (d), 126.17 (d), 126.0 (d), 80.75 (s), 80.1 (d), 68.17
(s), 67.08 (s), 40.48 (t), 32.78 (t), 30.29 (q), 28.18 (q), 22.5 (q), 17.79
(q); minor diastereomer δ 172.53 (s), 151.99 (s), 143.54 (s), 128.42
(d), 127.99 (d), 127.66 (d), 126.55 (d), 126.0 (d), 125.86 (d), 80.66
(d), 80.55 (s), 70.37 (s), 69.19 (s), 40.43 (t), 33.23 (t), 29.56 (q), 27.9
(q), 25.1 (q), 17.6 (q). MS: major diastereomer m/z 395 [M+], 266,
250, 236, 208, 131, 118, 103, 91, 84, 77, 57, 47; minor diastereomer
m/z 395 [M+], 266, 250, 236, 208, 131, 118, 103, 91, 84, 77, 57, 47.
HRMS Calcd for C25H33NO3: 395.2460 [M+]. Found for major
diastereomer: 395.2458. Found for minor diastereomer: 395.2458.
Methyl 2-(trans-2,5-Dimethyl-2,5-diphenylpyrrolidine-1-oxy)-
phenyl Acetate (6). The same procedure was followed to obtain 45
mg of colorless oil (72% yield). The major and minor diastereomers
were obtained as an inseparable mixture in a 1.4:1 ratio, as elucidated
1
by integration of the 500 MHz H NMR spectrum in CDCl3. Further
purification was done by HPLC. TLC: 95:5 hexane/ethyl acetate, UV,
p-anisaldehyde stain, Rf ) 0.24. IR(CDCl3): 2966, 1743, 1596, 1490,
1443, 1367, 1267, 1208, 1167, 1049, 1026 cm-1. 1H NMR (500 MHz,
CDCl3): major and minor diastereomers δ 7.02-7.88 (m, 30H), 5.09
(minor, s, 1H), 4.8 (major, s, 1H), 3.68 (major, s, 3H), 3.14 (minor, s,
3H), 2.56-2.68 (m, 2H), 2.21-2.29 (m, 2H), 1.96-2.12 (m, 4H), 1.71
(major, s, 3H), 1.51 (minor, s, 3H), 1.47 (minor, s, 3H), 1.02 (major,
s, 3H). 13C NMR (125 MHz, CDCl3): major and minor diastereomers:
δ 172.21 (major, s), 171.74 (minor, s), 151.74 (minor, s), 151.32 (major,
s), 143.04 (major, s), 142.65 (minor, s), 136.34 (minor, s), 136.28
(major, s), 128.72 (d), 128.59 (d), 128.56 (d), 128.42 (d), 128.2 (d),
128.03 (d), 127.96 (d), 127.72 (d), 127.58 (d), 127.33 (d), 126.79 (d),
126.76 (d), 126.26 (d), 126 (d), 125.86 (d), 86.86 (minor, d), 85.2
(major, d), 69.64 (minor, s), 69.08 (major, s), 68.51 (minor, s), 67.84
(major, s), 51.88 (major, q), 51.51 (minor, q), 40.62 (minor, t), 40.32
(major, t), 33.37 (major, t), 33.06 (minor, t), 30.09 (major, q), 29.62
(minor, q), 23.71 (minor, q), 23.01 (major, q). MS: m/z 416 [M +
H]+, 266, 250, 236, 131, 118. HRMS Calcd for C27H29NO3: 416.2226
(M + H). Found: 416.2232.
Methyl 2-(trans-2,5-Dimethyl-2,5-diphenylpyrrolidine-1-oxy) Pro-
pionate (4). The same procedure was followed to obtain 58 mg of
pale yellow oil (55% yield). The major and minor diastereomers were
obtained as a mixture in a 1.1:1 ratio, as elucidated by integration of
Methyl (2-(trans-2,5-Dimethyl-2,5-diphenylpyrrolidine-1-oxy))-
4-methoxyphenyl Acetate (7). The same procedure was followed to
obtain 28 mg of pale yellow oil (42% yield). The major and minor
diastereomers were obtained as a mixture in a 1.65:1 ratio, as elucidated
1
the 500 MHz H NMR spectrum in CDCl3. TLC: 93:7 hexane/ethyl
acetate, UV, p-anisaldehyde stain, Rf ) 0.44. IR (CDCl3): major
diastereomer 2978, 1743, 1490, 1449, 1373, 1279, 1208, 1091 cm-1
.
1
by integration of the 500 MHz H NMR spectrum in CDCl3. Further
1H NMR (500 MHz, CDCl3): major and minor diastereomers δ 7.18-
7.77 (m, 20H), 4.22 (minor, q, 1H, J ) 7.5), 3.97 (major, q, 1H, J )
7), 3.73 (major, s, 3H), 3.15 (minor, s, 3H), 2.52-2.62 (m, 2H), 2.14-
2.30 (m, 2H), 1.90-2.06 (m, 4H), 1.60 (s, 3H), 1.59 (s, 3H), 1.37 (s,
3H), 1.15 (s, 3H), 1.14 (minor, d, 3H, J ) 7.5), 0.97 (d, 3H, J ) 7).
13C NMR (125 MHz, APT, CDCl3): major and minor diastereomers δ
174.00 (major, s), 173.76 (minor, s), 151.60 (s), 151.52 (s), 142.93
(major, s), 142.93 (minor, s), 128.47 (d), 128.25 (d), 128.05 (d), 127.83
(d), 127.60 (d), 127.52 (d), 126.68 (d), 126.54 (d), 126.15 (d), 125.95
(d), 125.81 (d), 125.75 (d), 125.56 (d), 125.40 (d), 80.56 (minor, d),
79.25 (major, d), 69.95 (minor, s), 68.49 (minor, s), 68.32 (minor, s),
67.26 (major, s), 51.44 (q), 51.05 (minor, q), 40.35 (minor, t), 40.21
(major, t), 33.17 (minor, t), 32.83 (major, t), 29.90 (major, q), 29.51
(minor, q), 23.99 (minor, q), 22.58 (major, q), 17.52 (major, q), 17.38
(minor, q). MS: m/z 354 [M + H]+, 338, 266, 250, 236, 222, 157,
131, 118. HRMS Calcd for C22H27NO3 353.1990 [M+]. Found:
353.1990.
purification was done by HPLC. TLC: 95:5 hexane/ethyl acetate, UV,
p-anisaldehyde stain, Rf ) 0.13. IR (CDCl3): major diastereomer 2955,
1737, 1602, 1508, 1443, 1373, 1255, 1167, 1091, 1026 cm-1; minor
diastereomer 2966, 1743, 1602, 1508, 1461, 1443, 1373, 1255, 1167,
1091, 1026 cm-1. 1H NMR (500 MHz, CDCl3): major diastereomer δ
6.67-7.84 (m, 14H), 4.71 (s, 1H), 3.75 (s, 3H), 3.65 (s, 3H), 2.61-
2.65 (m, 1H), 2.18-2.22 (m, 1H), 1.99-2.09 (m, 2H), 1.67 (s, 3H),
0.99 (s, 3H); minor diastereomer δ 6.83-7.62 (m, 14H), 4.98 (s, 1H),
3.81 (s, 3H), 3.13 (s, 3H), 2.52-2.57 (m, 1H), 2.19-2.26 (m, 1H),
1.93-2.06 (m, 2H), 1.48 (s, 3H), 1.43 (s, 3H). 13C NMR (125 MHz,
CDCl3): major diastereomer δ 172.5 (s), 159.62 (s), 151.5 (s), 143.12
(s), 128.87 (d), 128.76 (d), 128.54 (d), 128.01 (d), 127.75 (d), 126.77
(d), 126.3 (d), 126.08 (d), 113.66 (d), 84.71 (d), 69.16 (s), 67.93 (s),
55.34 (q), 51.89 (q), 40.38 (t), 33.43 (t), 30.15 (q), 23.19 (q); minor
diastereomer δ 172.06 (s), 159.9 (s), 151.86 (s), 142.742 (s), 129.13
(d), 128.64 (d), 128.59 (d), 128.06 (d), 127.64 (d), 126.85 (d), 126.02
(d), 125.91 (d), 113.86 (d), 86.31 (d), 69.69 (s), 68.55 (s), 55.42 (q),
51.53 (q), 40.67 (t), 33.05 (t), 29.66 (q), 23.86 (q). MS: major
diastereomer m/z 446 [M + H]+, 307, 266, 236, 179, 154, 118; minor
diastereomer m/z 446 [M + H]+, 266, 236, 179, 118. HRMS Calcd
for C28H31NO4: 446.2331 [M + H]. Found for major diastereomer:
416.2328. Found for minor diastereomer: 416.2334.
Methyl 2-(trans-2,5-Dimethyl-2,5-diphenylpyrrolidine-1-oxy)cy-
clohexyl Acetate (5). The same procedure was followed to obtain 48
mg of colorless oil (76% yield). The major and minor diastereomers
were obtained as a mixture in a 2:1 ratio, as elucidated by integration
1
of the 500 MHz H NMR spectrum in CDCl3, which were purified
further by HPLC. TLC: 95:5 hexane/ethyl acetate, UV, p-anisaldehyde
stain, Rf ) 0.2. IR (CDCl3): 2928, 1743, 1492, 1446, 1373, 1275, 1202,
1172, 1059, 1029 cm-1. 1H NMR (500 MHz, CDCl3): major and minor
diastereomers δ 7.18-7.75 (m, 20H), 3.89 (minor, d, 1H, J ) 6), 3.71
(major, s, 3H), 3.52 (major, d, 1H, J ) 7.5), 3.15 (minor, s, 3H), 2.51-
2.55 (m, 2H), 1.86-2.23 (m, 6H), 1.62 (minor, s, 3H), 1.57 (major, s,
3H), 1.55 (major, s, 3H), 1.15 (minor, s, 3H), 0.68-1.68 (m, 20H).
13C NMR (125 MHz, APT, CDCl3): major and minor diastereomers δ
173.44 (major, s), 172.88 (minor, s), 151.94 (minor, s), 151.66 (major,
2-(trans-2,5-Dimethyl-2,5-diphenylpyrrolidine-1-oxy) Benzyl Ni-
trile (8). The same procedure was followed to obtain 34 mg of colorless
oil (50% yield). The major and minor diastereomers were obtained as
a mixture in a 1.4:1 ratio, as elucidated by integration of the 500 MHz
1H NMR spectrum in CDCl3. Further purification was done by HPLC.
TLC: 95:5 hexane/ethyl acetate, UV, p-anisaldehyde stain, Rf ) 0.15.
1
IR (CDCl3): 2966.2, 1460.7, 1378.4, 1260.9, 1096.4 cm-1. H NMR
(500 MHz, CDCl3): major and minor diastereomers δ 6.70-7.86 (m,
30H), 4.96 (major, s, 1H), 4.82 (minor, s, 1H), 2.62-2.80 (m, 4H),

8432 J. Am. Chem. Soc., Vol. 122, No. 35, 2000
Braslau et al.
2.09-2.30 (m, 4H), 1.87 (minor, s, 3H), 1.69 (major, s, 3H), 1.16
(minor, s, 3H), 1.06 (major, s, 3H). 13C NMR (125 MHz, APT,
CDCl3): major and minor diastereomers δ 151.32 (s), 150.89 (s), 143.57
(minor, s), 142.63 (s), 132.92 (s), 132.44 (s), 129.79 (d), 128.91 (d),
128.66 (d), 128.53 (d), 128.46 (d), 128.22 (d), 128.22 (d), 127.93 (d),
127.82 (d), 127.32 (d), 127.04 (d), 126.89(d), 126.23 (d), 126.14 (d),
125. 94 (d), 119.15 (s), 119.11 (s), 75.11 (s), 74.81 (s), 74.75 (d), 74.55
(d), 69.27 (s), 68.03 (s), 40.36 (t), 38.49 (t), 33.33 (t), 33.22 (t), 31.21
(q), 30.19 (q), 28.54 (q), 22.92 (q). MS: m/z 383 [M + H]+, 266, 236,
190, 131, 118. HRMS Calcd for C26H26N2O: 383.2123 [M + H].
Found: 372.2121.
123.24 (major, d), 123.10 (minor, d), 81.17 (minor, d), 80.02 (major,
d), 69.47 (major, s), 69.38 (minor, s), 68.60 (major, s), 68.01 (minor,
s), 40.57 (minor, t), 40.48 (major, t), 34.21 (minor, t), 34.21 (major, t),
30.21 (minor, q), 29.65 (major, q), 23.73 (major, q), 22.98 (minor, q),
22.25 (minor, q), 21.69 (major, q). MS: m/z 417 [M + H]+, 339, 309,
282, 266, 190, 152, 135, 119. HRMS Calcd for C26H28N2O3: 417.2178
[M + H]. Found: 417.2180.
1-(trans-2,5-Dimethyl-2,5-diphenylpyrrolidine-1-oxy)-1-(4-ami-
nophenyl)ethane (16). The same procedure was followed to obtain
27 mg of pale yellow oil (39% yield). The major and minor
diastereomers were obtained as a mixture in a 2.9:1 ratio, as elucidated
1
by integration of the 500 MHz H NMR spectrum in CDCl3. TLC:
1-(trans-2,5-Dimethyl-2,5-diphenylpyrrolidine-1-oxy)isochro-
manone (9). The same procedure was followed to obtain 48 mg of
pale yellow oil (45% yield). The major and minor diastereomers were
obtained as a mixture in a 1.3:1 ratio, as elucidated by integration of
95:5 hexane/ethyl acetate, UV, p-anisaldehyde stain, Rf ) 0.13. IR-
(CDCl3): 3389, 3060, 2971, 2931, 2494, 1951, 1704, 1621, 1519, 1417,
1318, 1274, 1180, 1133, 1061, 1029 cm-1 1H NMR (500 MHz,
.
1
CDCl3): major and minor diastereomers δ 7.25-7.69 (m, 20H), 6.94
(major, d, 2H, J ) 7.5), 6.84 (minor, d, 2H, J ) 8), 6.63 (major, d,
2H, J ) 8), 6.49 (minor, d, 2H, J ) 7.5), 4.16 (minor, q, 1H, J ) 6.5),
4.14 (major, q, J ) 6.5), 3.61 (br S, 2H), 2.47-2.64 (m, 2H), 1.87-
2.21 (m, 6H), 1.58 (minor, s, 3H), 1.28 (major, s, 3H), 1.22 (minor, d,
3H, J ) 6.5), 1.14 (major, s, 3H), 1.11 (minor, s, 3H), 1.06 (major, d,
3H, J ) 6.0). 13C NMR (125 MHz, APT, CDCl3): major and minor
diastereomers δ 152.56 (minor, s), 152.56 (major, s), 145.69 (major,
s), 145.42 (minor, s), 144.1 (minor, s), 143.90 (major, s), 134.74 (major,
s), 134.02 (minor, s), 128.67 (d), 128.05 (d), 127.91 (d), 127.78 (d),
127.46 (d), 126.59 (d), 126.45 (d), 126.15 (d), 126.06 (d), 125.96 (d),
125.75 (d), 114.75 (major, d), 114.75 (minor, d), 80.76 (major, d), 79.99
(minor, d), 68.84 (major, s), 68.84 (minor, s), 67.93 (major, s), 67.93
(minor, s), 40.63 (major, t), 40.42 (minor, t), 34.09 (minor, t), 33.82
(major, t), 30.00 (minor, q), 30.00 (major, q), 24.89 (minor, q), 23.48
(major, q), 21.66 (major, q), 21.55 (minor, q). MS: m/z 387 [M +
H]+, 282, 268, 252, 239, 224, 207. 190, 174, 157, 140, 120. HRMS
Calcd for C26H30N2O: 387.2435 [M + H]. Found: 387.2436.
1-(trans-2,5-Dimethyl-2,5-diphenylpyrrolidine-1-oxy)-1-phenyl-
ethane (14). The same procedure was followed to obtain 25 mg of a
pale yellow oil (56% yield). The major and minor diastereomers were
obtained as an inseparable mixture in a 3.4:1 ratio, as elucidated by
integration of the 500 MHz 1H NMR spectrum in CD3CN. See
compound 20a for characterization.
the 500 MHz H NMR spectrum in CDCl3. TLC: 4:1 hexane/ethyl
acetate, UV, p-anisaldehyde stain, Rf ) 0.39. IR(CDCl3): 2966, 1755,
1708, 1604, 1443, 1255, 1085, 1026 cm-1 1H NMR (500 MHz,
.
CDCl3): major and minor diastereomers δ 7.07-7.78 (m, 26H), 6.71
(d, 1H, J ) 7.5), 6.38 (d, 1H, J ) 7.5), 5.19 (d, 1H, J ) 14.0), 4.85
(d, 1H, J ) 13.5), 4.84 (d, 1H, J ) 14.0), 4.82 (s, 1H), 4.76 (d, 1H, J
) 13.5), 4.73 (s, 1H), 1.86-2.68 (m, 8H), 1.64 (major, s, 3H), 1.57
(minor, s, 3H), 1.29 (major, s, 3H), 1.16 (minor, s, 3H). 13C NMR
(125 MHz, APT, CDCl3): major and minor diastereomers δ 168.85
(major, s), 168.43 (minor, s), 150.84 (major, s), 150.73 (minor, s),
143.52 (minor, s), 143.10 (major, s), 134.16 (major, s), 134.16 (minor,
s), 131.53 (major, s), 131.42 (minor, s), 129.74 (d), 129.46 (d), 129.34
(d), 128.87 (d), 128.44 (d), 128.25 (d), 128.16 (d), 128.11 (d), 127.97
(d), 126.82 (d), 126.54 (d), 126.43 (d), 124.77 (d), 124.67 (d), 79.94
(major, d), 78.74 (minor, d), 73.05 (minor, s), 72.80 (minor, s), 70.72
(major, s), 70.23 (major, s), 69.71 (minor, t), 69.47 (major, t), 39.97
(major, t), 39.37 (minor, t), 35.04 (minor, t), 34.57 (major, t), 29.75
(major, q), 28.91 (minor, q), 26.62 (minor, q), 24.38 (major, q). MS:
m/z 414 [M + H]+, 266, 221, 190, 147, 131. HRMS Calcd for C27H27-
NO3: 414.2069 [M + H]. Found: 414.2075.
Manganese Salen Method. For the synthesis of [N, N′-bis-
(salicylidene)-1,2-ethanediaminato]manganese(III) chloride (13), see the
Supporting Information.
Coupling of 1-(4-Nitrophenyl)ethyl Radical Generated by the
Mn(salen) Method with DPPO To Form 1-(trans-2,5-Dimethyl-2,5-
diphenylpyrrolidine-1-oxy)-1-(4-nitrophenyl)ethane (15). To a mix-
ture of 4-nitrostyrene (22.4 mg, 0.150 mmol) and the C2-symmetric
nitroxide DPPO, trans-2,5-dimethyl-2,5-diphenylpyrrolidine-1-oxyl (40
mg, 0.150 mmol), in 0.74 mL of ethanol/ toluene (1:1) was added [N,N′-
bis(salicylidene)-1,2-ethanediaminato]manganese(III) chloride (8 mg,
0.023 mmol), followed by di-tert-butyl peroxide (27.5 µL, 0.150 mmol)
and sodium borohydride (11.3 mg, 0.300 mmol). The mixture was
stirred at room temperature under nitrogen for 22 h. It was then
evaporated to dryness and partitioned between 5 mL of dichloromethane
and 8 mL of water. The aqueous layer was washed with 3 × 10 mL of
dichloromethane. The combined dichloromethane layer was dried over
magnesium sulfate, the solution was filtered, and the volatiles were
removed in vacuo to give 85 mg of a dark brown, thick oil which was
purified by flash column chromatography (1 cm column, 9:1 hexane/
dichloromethane gradient) to give 47 mg of colorless oil (75% yield).
The major and minor diastereomers were obtained as an inseparable
mixture in a 2.4:1 ratio, as elucidated by integration of the 500 MHz
1H NMR spectrum in CDCl3. TLC: 95:5 hexane/ethyl acetate, UV,
p-anisaldehyde stain, Rf ) 0.37. IR (CDCl3): 2966, 1796, 1731, 1602,
1519, 1467, 1373, 1343, 1261, 1091, 1009 cm-1. 1H NMR (500 MHz,
CDCl3): major and minor diastereomers δ 8.13 (major, d, 2H, J )
8.5), 7.85 (minor, d, 2H, J ) 9.0), 7.18 (major, d, 2H, J ) 9.0), 7.06-
7.72 (m, 20H), 7.03 (minor, d, 2H, J ) 8.5), 4.50 (minor, q, 1H, J )
6.5), 4.33 (major, q, J ) 7), 2.53-2.64 (m, 2H), 1.91-2.21 (m, 6H),
1.73 (minor, s, 3H), 1.36 (major, s, 3H), 1.28 (minor, d, 3H, J ) 6.5),
1.14 (minor, s, 3H), 1.12 (major, s, 3H), 1.11 (major, d, 3H, J ) 6.5).
13C NMR (125 MHz, APT, CDCl3): major and minor diastereomers δ
151.88 (minor, s), 151.67 (major, s), 151.63 (major, s), 150.98 (minor,
s), 147.24 (major, s), 146.82 (minor, s), 143.18 (major, s), 143.14
(minor, s), 128.59 (d), 128.44 (d), 128.02 (d), 127.69 (d), 127.57 (d),
127.13 (d), 126.88 (d), 126.68 (d), 126.38 (d), 126.20 (d), 125.81 (d),
Benzyl Hydrazine, Lead Dioxide Method for Generating Prochiral
Carbon Radicals: Coupling to DPPO. For bromination of benzyl
alcohols with Br2/DIPHOS to give benzyl bromides, see the Supporting
Information. The bromides were then reacted with hydrazine and then
oxidized to form the prochiral carbon radicals. The following is a typical
example.
1-(trans-2,5-Dimethyl-2,5-diphenylpyrrolidine-1-oxy)-1-phenyl-
ethane (20a). A mixture of (1-bromoethyl)benzene (205 µL, 1.5 mmol)
and fuming hydrazine (940 µL, 30 mmol) was sonicated under nitrogen
for 30 min. The cloudy solution was diluted with 30 mL of diethyl
ether, and the hydrazine layer was washed with 10 mL of ether. The
combined ether layer was washed with 20 mL of 10% potassium
hydroxide solution followed by 10 mL of saturated sodium chloride
solution. After drying over magnesium sulfate, the solution was filtered,
and the volatiles were removed in vacuo to give 138 mg of a slightly
yellow viscous oil. This oil was diluted with 0.6 mL of toluene. The
C2-symmetric nitroxide DPPO, trans-2,5-dimethyl-2,5-diphenylpyrro-
lidine-1-oxyl (67.4 mg, 0.253 mmol), and lead dioxide (121 mg, 0.506
mmol) were suspended in 0.6 mL of toluene, sonicated for 3 min under
nitrogen, and cooled to -78 °C, and the benzyl hydrazine solution was
added dropwise by cannula. The residues of benzyl hydrazine were
washed in with two 0.3 mL portions of toluene. The mixture was
allowed slowly to warm to room temperature, followed by filtration
through Celite with dichloromethane rinses. The volatiles were removed
in vacuo to give a thick, yellow oil which was purified by flash column
chromatography (2 cm column, 95:5 hexane/ethyl acetate gradient)
followed by a second column (1 cm, 95:5 hexane/ethyl acetate) to yield
58 mg of colorless oil (62% yield). The major and minor diastereomers
were obtained as a mixture in a 2.5:1 ratio, as elucidated by integration
1
of the 500 MHz H NMR spectrum in CD3CN. Further purification
was done by HPLC to obtain an analytically pure sample of the mixture
of diastereomers. TLC: 95:5 hexane/ethyl acetate, UV, p-anisaldehyde

StereoselectiVe Coupling of Prochiral Radicals
J. Am. Chem. Soc., Vol. 122, No. 35, 2000 8433
stain, Rf ) 0.3. IR (CDCl3): 2966, 1461, 1379, 1261, 1091, 1014 cm-1
.
(minor, s), 40.51 (major, t), 40.51 (minor, t), 34.24 (minor, t), 34.02
(major, t), 30.17 (minor, q), 29.67 (major, q), 23.67 (major, q), 23.29
(minor, q), 22.33 (minor, q), 21.77 (major, q). Elemental analysis: Anal.
Calcd for C27H28F3NO: C, 73.78; H, 6.42; F, 12.97; N, 3.19. Found:
C, 73.57; H, 6.21; F, 13.30; N, 3.07.
1H NMR (500 MHz, CD3CN): major and minor diastereomers δ 7.16-
7.81 (m, 30H), 4.37 (major, q, 1H, J ) 7.0), 4.30 (minor, q, 1H, J )
6.5), 2.58-2.79 (m, 2H), 2.05-2.36 (m, 6H), 1.73 (minor, s, 3H), 1.38
(major, d, 3H, J ) 6.5), 1.35 (major, s, 3H), 1.19 (major, s, 3H), 1.14
(minor, s, 3H), 1.13 (major, d, 3H, J ) 7.0). 1H NMR (500 MHz, CD3-
OD): major and minor diastereomers δ 6.96-7.69 (m, 30H), 4.22
(major, q, 1H, J ) 6.5), 4.23 (minor, q, 1H, J ) 6.5), 2.46-2.66 (m,
2H), 2.12-2.22 (m, 2H), 1.86-2.04 (m, 4H), 1.57 (minor, s, 3H), 1.22
(major, s, 3H), 1.20 (minor, d, 3H, J ) 6.5), 1.04 (major, s, 3H), 1.03
(minor, s, 3H), 1.00 (major, d, 3H, J ) 6.5). 13C NMR (125 MHz,
CD3OD): major and minor diastereomers δ 153.61 (s), 153.36 (s),
145.71 (s), 145.71 (s), 145.05 (s), 144.84 (s), 129.78 (d), 129.04 (d),
128.97 (d), 128.81 (d), 128.54 (d), 128.44 (d), 128.11 (d), 127.80 (d),
127.69 (d), 127.48 (d), 127.32 (d), 126.94 (d), 82.73 (major, d), 81.88
(minor, d), 70.80 (minor, s), 69.75 (major, s), 69.75 (minor, s), 68.9
(major, s), 41.63 (major, t), 41.48 (minor, t), 34.95 (minor, t), 34.71
(major, t), 30.51 (major, q), 30.51 (minor, q), 25.02 (minor, q), 23.61
(major, q), 22.43 (minor, q), 22.26 (major, q). MS: m/z 372 [M +
H]+, 356, 267, 252, 190, 118, 105. HRMS Calcd for C26H29NO:
372.2327 [M + H]. Found: 372.2325.
1-(trans-2,5-Dimethyl-2,5-diphenylpyrrolidine-1-oxy)-1,2,3,4-tet-
rahydronaphthalene (20d). The same procedure was followed to
obtain 14 mg of pale yellow oil (41% yield). The major and minor
diastereomers were obtained as a mixture in a 5:1 ratio, as elucidated
1
by integration of the 500 MHz H NMR spectrum in CDCl3. TLC:
93:7 hexane/ethyl acetate, UV, p-anisaldehyde stain, Rf ) 0.45. IR-
1
(CDCl3): 2940, 1462, 1376, 1260, 1093, 1001 cm-1. H NMR (500
MHz, CDCl3): major and minor diastereomers δ 6.75-7.7.74 (m, 28H),
4.34 (major, t, 1H, J ) 3.5), 4.08 (minor, dd, 1H, J ) 5.5, 4), 2.51-
2.82 (m, 6H), 1.67 (minor, s, 3H), 1.41 (major, s, 3H), 1.38 (major, s,
3H), 1.30 (minor, s, 3H), 1.32-2.30 (m, 14H). 13C NMR (62.5 MHz,
APT, CDCl3): major and minor diastereomers δ 152.18 (minor, s),
151.50 (major, s), 143.94 (major, s), 143.72 (minor, s), 137.83 (major,
s), 137.56 (minor, s), 136.43 (minor, s), 135.98 (major, s), 130.83
(major, d), 129.31 (minor, d), 128.55 (d), 128.44 (d), 127.73 (d), 127.64
(d), 127.43 (d), 127.38 (d), 126.95 (d), 126.49 (d), 126.31 (d), 125.99
(d), 125.89 (d), 125.64 (d), 125.22 (d), 124.76 (d), 77.40 (major, d),
76.03 (minor, d), 70.69 (major, s), 69.33 (major, s), 68.13 (minor, s),
67.24 (minor, s), 40.80 (minor, t), 40.27 (major, t), 35.15 (major, t),
33.46 (minor, t), 30.05 (minor, q), 29.95 (major, q), 29.33 (minor, t),
28.90 (major, t), 27.72 (major, t), 27.63 (minor, t), 24.28 (major, q),
23.20 (minor, q), 18.96 (minor, t), 17.96 (major, t). MS: m/z 398 [M
+ H]+, 267, 252, 236, 190, 131, 118. HRMS Calcd for C27H29NO3:
398.2484 [M + H]. Found: 398.2485.
1-(trans-2,5-Dimethyl-2,5-diphenylpyrrolidine-1-oxy)-indane (20e).
The same procedure was followed to obtain 50 mg of a colorless oil
(52% yield). The major and minor diastereomers were obtained as a
mixture in a 5:1 ratio, as elucidated by integration of the 500 MHz 1H
NMR spectrum in CDCl3. Further purification was done by HPLC to
obtain an analytically pure sample of the mixture of diastereomers.
TLC: 98:2 hexane/ethyl acetate, UV, p-anisaldehyde stain, Rf ) 0.25.
IR(CDCl3): 3154, 2978, 1813, 1796, 1602, 1467, 1378, 1261, 1091
cm-1. 1H NMR (500 MHz, CDCl3): major and minor diastereomers δ
7.66-7.74 (m, 8H), 7.02-7.54 (m, 20H), 4.94 (major, t, 1H, J ) 5.0),
4.75 (minor, dd, 1H, J ) 6.5, 5.0), 2.77-2.95 (m, 2H), 2.52-2.67 (m,
4H), 2.02-2.28 (m, 6H), 1.72-1.74 (m, 4H), 1.71 (minor, s, 3H), 1.34
(major, s, 3H), 1.28 (minor, d, 3H, J ) 6.5), 1.13 (minor, s, 3H), 1.11
(minor, s, 3H), 1.09 (major, d, 3H, J ) 6.5). 13C NMR (125 MHz,
APT, CDCl3): major and minor diastereomers δ 152.22 (major, s),
152.17 (minor, s), 144.80 (major, s), 144.15 (minor, s), 143.79 (minor,
s), 142.72 (major, s), 128.61 (d), 128.66 (d), 128.48 (d), 128.25 (d),
128.02 (d), 127.89 (d), 127.71 (d), 126.42 (d), 126.29 (d), 126.18 (d),
126.12 (d), 126.07 (d), 126.00 (d), 125.92 (d), 125.51 (d), 124.63 (d),
124.58 (d), 87.39 (major, d), 86.08 (minor, d), 70.73 (major, s), 69.16
(major, s), 68.63 (major, s), 67.82 (minor, s), 40.68 (minor, t), 40.32
(major, t), 34.12 (major, t), 33.43 (minor, t), 32.98 (minor, t), 32.80
(major, t), 30.40 (minor, q), 30.26 (major, t), 30.00 (minor, t), 29.82
(major, q), 25.00 (major, q), 23.71 (minor, q). Elemental analysis: Anal.
Calcd for C27H29NO: C, 84.56; H, 7.62; N, 3.65. Found: C, 84.34; H,
7.74; N, 3.47.
1-(trans-2,5-Dimethyl-2,5-diphenylpyrrolidine-1-oxy)-1-(2,6-di-
methylphenyl)ethane (20b). The same procedure was followed to
obtain 27 mg of a whitish oil (32% yield). The major and minor
diastereomers were obtained as a mixture in a 1.5:1 ratio, as elucidated
1
by integration of the 500 MHz H NMR spectrum in CDCl3. TLC:
93:7 hexane/ethyl acetate, UV, p-anisaldehyde stain, Rf ) 0.5. IR
1
(CDCl3): 2966, 1496, 1449, 1367, 1261, 1097, 1008 cm-1. H NMR
(500 MHz, CDCl3): major and minor diastereomers δ 6.72-7.73 (m,
26H), 4.63 (major, q, 1H, J ) 7), 4.44 (minor, q, 1H, J ) 7), 2.48-
2.70 (m, 2H), 2.30 (major, s, 3H), 2.22 (minor, s, 3H), 2.07 (major, s,
3H), 1.95-2.25 (m, 6H), 1.82 (minor, s, 3H), 1.70 (minor, s, 3H), 1.28
(major, s, 3H), 1.28 (minor, s, 3H), 1.26 (major, d, 3H, J ) 7), 1.08
(minor, s, 3H), 0.90 (minor, d, 3H, J ) 7). 13C NMR (125 MHz,
CDCl3): major and minor diastereomers δ 152.23 (major, s), 151.97
(minor, s), 143.80 (major, s), 143.74 (minor, s), 140.38 (major, s),
139.59 (minor, s), 136.09 (major, s), 135.90 (minor, s), 134.07 (major,
s), 133.69 (minor, s), 130.74 (d), 130.51 (d), 128.69 (d), 128.59 (d),
127.88 (d), 127.61 (d), 127.40 (d), 126.87 (d), 126.56 (d), 126.48 (d),
126.37 (d), 126.26 (d), 126.20 (d), 126.17 (d), 125.84 (d), 125.75 (d),
79.71 (major, d), 76.78 (minor, d), 68.83 (minor, s), 68.83 (major, s),
67.79 (minor, s), 67.79 (major, s), 40.43 (t), 40.29 (t), 34.10 (t), 33.78
(t), 30.28 (q), 29.92 (q), 23.37 (minor, q), 23.20 (major, q), 22.36 (major,
q), 21.91 (minor, q), 21.15 (major, q), 21,05 (minor, q), 19.20 (major,
q), 18.80 (minor, q). MS: m/z 400 [M + H]+, 309, 267, 252, 190,
155, 133,118. HRMS Calcd for C28H33NO: 400.2640 [M + H].
Found: 400.2642.
1-(trans-2,5-Dimethyl-2,5-diphenylpyrrolidine-1-oxy)-1-(4-trifluo-
romethylphenyl)ethane (20c). The same procedure was followed to
obtain 86 mg of white solid (63% yield), mp 92-96 °C. The major
and minor diastereomers were obtained as a mixture in a 3.5:1 ratio,
1
as elucidated by integration of the 500 MHz H NMR spectrum in
CDCl3. When the reaction was carried out at 0 °C for 4 h, the major
and minor diastereomers were obtained as a mixture in a 4:1 ratio, as
elucidated by integration of the 500 MHz 1H NMR spectrum in CDCl3.
TLC: 7:1 pentane/methylene chloride, UV, p-anisaldehyde stain, Rf )
0.5. IR (CDCl3): 2966, 2931, 1619, 1596, 1490, 1443, 1420, 1373,
1326, 1279, 1167, 1126, 1096, 1061, 1014 cm-1. 1H NMR (500 MHz,
CDCl3): major and minor diastereomers δ 7.01-7.72 (m, 28H), 4.41
(minor, q, 1H, J ) 6.5), 4.29 (major, q, 1H, J ) 6.5), 2.29-2.64 (m,
2H), 1.89-2.25 (m, 6H), 1.71 (minor, s, 3H), 1.34 (major, s, 3H), 1.28
(minor, d, 3H, J ) 6.5), 1.13 (minor, s, 3H), 1.11 (minor, s, 3H), 1.09
(major, d, 3H, J ) 6.5). 13C NMR (125 MHz, APT, CDCl3): major
and minor diastereomers δ 152.11 (major, s), 151.24 (minor, s), 148.22
(major, s), 148.13 (minor, s), 143.38 (major, s), 143.38 (minor, s),
Coupling of 1-Tetralinyl Radical with trans-1,3-Dimethyl-1,3-
diphenylisoindoline-1-oxyl (21) to form 1-(trans-1,3-Dimethyl-1,3-
diphenylisoindoline-2-oxy)-1,2,3,4-tetrahydronaphthalene (22). The
same procedure was followed using the C2-symmetric nitroxide, trans-
1,3-dimethyl-1,3-diphenylisoindoline-1-oxyl (21) (39.0 mg, 0.124
mmol), in place of DPPO to obtain 26 mg of a foamy, white solid
(47% yield). The major and minor diastereomers were obtained as a
mixture in a 3.2:1 ratio, as elucidated by integration of the 500 MHz
1H NMR spectrum in C6D6. Recrystallization from hexane gave a white
solid, mp 105-107 °C (dec). TLC: 5:1 pentane/methylene chloride,
UV, p-anisaldehyde stain, Rf ) 0.33. IR (CDCl3): 2955, 2919, 1595,
2
2
129.471 (major, q, J (CF) ) 31), 128.66 (d), 128.616 (minor, q, J
(CF) ) 33), 128.52 (d), 128.00 (d), 127.69 (d), 127.60 (d), 127.52 (d),
126.80 (d), 126.70 (d), 126.60 (d), 126.28 (d), 125.93 (d), 125.78 (d),
1490, 1443, 1367, 1255, 1067, 1026 cm-1 1H NMR (500 MHz,
.
CDCl3): major and minor diastereomers δ 6.76-7.58 (m, 36H), 4.21
(major, t, 1H, J ) 3.5), 4.14 (minor, t, 1H, J ) 5.0), 2.60-2.91 (m,
4H), 1.83-2.56 (m, 4H), 2.06 (minor, s, 3H), 1.39-1.70 (m, 4H), 1.62
(minor, s, 3H), 1.58 (major, s, 3H), 1.41 (major, s, 3H). [Note: The
1
125.02 (d), 124.91 (d), 124.88 (d), 124.78 (d), 124.42 (major, q, J
(CF) ) 270), 124.39 (minor, q, 1J (CF) )270), 81.03 (minor, d), 80.47
(major, d), 69.33 (minor, s), 69.25 (major, s), 68.35 (major, s), 68.07

8434 J. Am. Chem. Soc., Vol. 122, No. 35, 2000
Braslau et al.
methyl singlets at 1.58 and 1.41 ppm are seen only after cooling the
NMR sample to -10 °C.] 13C NMR (62.5 MHz, CDCl3): major and
minor diastereomers δ 148.67 (major, s), 148.14 (minor, s), 147.89
(s), 143.99 (major, s), 143.88 (s), 143.02 (minor, s), 137.67 (minor, s),
137.61 (major, s), 136.80 (major, s), 136.40 (minor, s), 131.93 (d),
129.79 (d), 129.46 (d), 128.92 (d), 128.65 (d), 128.53 (d), 128.50 (d),
128.17 (d), 127.98 (d), 127.87 (d), 127.73 (d), 127.44 (d), 127.17 (d),
127.10 (d), 127.03 (d), 126.86 (d), 125.50 (d), 124.83 (d), 123.74 (d),
123.61 (d), 77.02 (major, d), 76.66 (minor, d), 72.97 (major, s), 72.97
(major, s), 72.19 (minor, s), 72.07 (minor, s), 29.06 (minor, t), 29.14
(minor, q), 28.95 (major, q), 28.69 (major, t), 28.06 (major, t), 28.02
(minor, t), 20.69 (minor, q), 20.60 (major, q), 19.45 (minor, t), 17.94
(major, t). MS: m/z 446 [M + H]+, 315, 300, 284, 269, 252, 238,
222, 206, 165, 131, 115. HRMS Calcd for C32H31NO: 446.2484 [M +
H]. Found: 446.2483.
Science Foundation (Grant CHE-9527647) for providing finan-
cial support for this project. H.Z. thanks the Fonds der
Chemischen Industrie for generous financial support. We thank
Vladimir Chaplinski for the preparation of nitroxide 21, Rodney
Howden for the preparation of DPPO, and Glenn Millhauser
and D. Joe Anderson for assistance with the ESR measurements.
We thank Hans Fischer for insightful discussion.
Supporting Information Available: Preparation of DPPO
1, preparation of Mn(salen) 13, preparation of benzyl bromides
1
19b-e, H-NMR and 13C-NMR spectra for 20b, experimental
procedures and ESR spectra for the variable temperature ex-
periments monitoring both the hydrazine and enolate oxidation
methods of radical generation (PDF). This material is available
free of charge via the Internet at http://pubs.acs.org.
Acknowledgment. R.B. and N.N. thank the University of
California, Santa Cruz, Faculty Research Funds and the National
JA000520U