One product, two pathways: Initially divergent radical reactions reconverge to form a single product in high yield

Article abstract of DOI:10.1021/ja2070347

The paper describes examples of net diastereotopic-group-selective radical processes having the unusual feature that a single product is formed even though the key reaction of the two diastereotopic radical precursors is nonselective. For example, reactio

Full text of DOI:10.1021/ja2070347

ARTICLE
pubs.acs.org/JACS
One Product, Two Pathways: Initially Divergent Radical Reactions
Reconverge To Form a Single Product in High Yield
Achim Bruch,† Roland Fr€ohlich,† Stefan Grimme,† Armido Studer,*,† and Dennis P. Curran*,‡
†Organisch-Chemisches Institut, Westf€alische Wilhelms-Universit€at, Corrensstrasse 40, 48149 M€unster, Germany
‡Department of Chemistry, University of Pittsburgh, Pittsburgh, Pennsylvania 15260, United States
S Supporting Information
b
ABSTRACT: The paper describes examples of net diastereo-
topic-group-selective radical processes having the unusual fea-
ture that a single product is formed even though the key reaction
of the two diastereotopic radical precursors is nonselective. For
example, reaction of (R)-N-(cyclohex-2-en-1-yl)-N-(2,6-diio-
do-4-methylphenyl)acetamide with tributyltin hydride pro-
duces 1-((4aR,9aR)-6-methyl-2,3,4,4a-tetrahydro-1H-carbazol-
9(9aH)-yl)ethanone with high product selectivity and in high
yield. Analysis of the concentration profiles of the closed-shell
intermediates at the halfway point of the reaction shows that
nonselective abstraction of diastereotopic iodides by tin radicals occurs, leading to diastereomeric aryl radicals. These isomeric
intermediates evolve via two nonintersecting reaction pathways, cyclization and bimolecular trapping or vice versa, into the same
final product. Origins of the selectivity are suggested on the basis of conformational analysis of the products using both X-ray
crystallography and density functional theory calculations.
’ INTRODUCTION
called stereoselection at the steady state, an apparent group-
selective reaction can be accomplished in an indirect way by
controlling the pathways by which nonselectively generated
intermediates react.
In many multistep organic transformations, there is generally
only one low-energy path to get from a precursor to a product. If
the reaction wanders onto the wrong path, then it may not be
possible to get back on track. Here we report a stereotopic-group-
selective reaction that has two completely different paths to the
same product. There are no U-turns, nor is it possible to go
around the block; paths intersect nowhere but at their origin and
destination. Despite that, it does not matter which path is taken;
the same product is formed either way.
Stereotopic-group-selective reactions are generally per-
formed in a direct manner by pitting two (or more) diaster-
eomeric transition states (TSs) in competition with each
other.1 The product derived from the lower-energy TS, of
course, wins the competition. This is illustrated in Figure 1 for
the conversion of two diastereotopic groups X and X0 in a
precursor into different substituents A and B in a product.2
The first group-selective reaction, here conversion of either X
or X0 to A, completely controls the overall stereoselectivity.
The subsequent reaction with B occurs after the key selective
step is over.
Consider the divergence of the same precursor with X and X0
to a pair of radicals and assume that this step occurs without
selectivity (Figure 1, bottom). These intermediate radicals are
diastereomers that can react differently. Now imagine that one
diastereomer reacts with A before B while the other reacts with B
before A. At the end, these divergent paths reconverge to the
same product. A net group-selective reaction results even though
there was no selectivity in the initial (and only) opportunity for
direct differentiation of the two groups X and X0.
The direct and indirect processes differ in many ways,3b one of
the most important being the theoretical yield of the major
product. In the case of the classical group-selective process, the
yield of the major product can never exceed the level of
stereoselectivity in the first step. In the steady-state process,
the converging topography can in theory override the lack of
group selectivity, providing a high yield of one product.
Reaction time course analyses of net diastereotopic-group-
selective radical reactions have provided convincing evidence
of the principle of stereoselection at the steady state; however,
the selectivities observed to date have been low.4a For example,
reduction of diiodide 1 with Bu3SnH at 0.7 M provides three
A long-held corollary to this type of reaction is that if the initial
group-selective reaction occurs without selectivity, then the
overall process cannot be selective. However, this corollary is
not correct.
Some time ago, we recognized that a pair of stereotopic-group-
selective reactions could be performed in a fundamentally dif-
ferent way that is based on reaction topography.3,4 By this method,
Received: August 1, 2011
Published: September 01, 2011
r
2011 American Chemical Society
16270
dx.doi.org/10.1021/ja2070347 J. Am. Chem. Soc. 2011, 133, 16270–16276
|

Journal of the American Chemical Society
ARTICLE
Figure 3. Precursor 7 designed on the basis of widely differing out-
comes in radical reactions of atropisomers 4-syn and 4-anti.
Figure 1. Direct (traditional) and indirect (steady-state) pathways
for accomplishing a pair of net group-selective reactions with A and B.
It is assumed that R* has an element of chirality that makes X and X0
diastereotopic.
reactions (cyclization and bimolecular trapping) but changing
the prochiral element of substrates like 1 from a stereocenter
into an axis.
’ RESULTS AND DISCUSSION
The opportunity to design a substrate for a highly selec-
tive process presented itself during seemingly unrelated work on
radical reactions of axially chiral anilides.5 In particular, atropi-
somers 4-syn and 4-anti are separated by high rotational barriers
of 36ꢀ37 kcal molꢀ1 and can be readily obtained in pure form
(Figure 3). Remarkably, atropisomer 4-syn reacts with Bu3SnH
to give principally 5, the product of cyclization followed by
trapping, whereas atropisomer 4-anti gives exclusively 6, the
intercepted product of direct trapping without cyclization. The
direct-trapping product 6 was obtained even at relatively low
concentrations of the tin hydride.
Evidently, there are large differences in the rates of cycliza-
tion of intermediate radicals derived from 4-syn (fast) and 4-
anti (slow). To capitalize on these differences in a steady-state
setting, we “fused” diastereomers 4-syn and 4-anti into a single
precursor 7 bearing two iodine atoms at the 2- and 6-positions
of the N-aryl ring. This compound shares a key feature of
previously studied (but poorly selective) precursors in that it
contains two diastereomeric radical precursors.4b However,
the overall process is different because the prochiral element
of the precursor is a NꢀAr axis (not a stereocenter) and
because the net group selectivity is expressed as a ratio of
constitutional isomers (reduced and cyclized) rather than
stereoisomers.
Figure 2. Typical example of product selection at the steady state
involving a substrate 1 bearing two radical precursors and a prochiral
center.
products, 2-exo, 2-endo, and 3, in yields of 60, 26, and 14%,
respectively (Figure 2). The abstraction of an iodine atom by tin
radical to start this reaction is nonselective, yet the 2-exo/2-endo
ratio exceeds 1/1 and the yield of 2-exo exceeds 50%. This occurs
because one diastereomeric radical goes to 2-exo by cyclization
followed by reductive deiodination while the other goes by
reductive deiodination followed by cyclization.
The object of this work was to discover examples of pro-
duct selection at the steady state that occur with high selectivities.
This was accomplished by using the same types of competing
Equation 1 shows the synthesis of the precursor. Diiodoani-
lide 7 was prepared by acylation of 2,6-diiodo-4-methylaniline
16271
dx.doi.org/10.1021/ja2070347 |J. Am. Chem. Soc. 2011, 133, 16270–16276

Journal of the American Chemical Society
ARTICLE
Scheme 1. Structures of the Intermediate (8 and 9) and Final
(10 and 11) Products Formed during the Tin Hydride
Reduction of 7
Figure 4. X-ray crystal structure of diiodide 7. Selected torsion
angles: O1ꢀC1ꢀN1ꢀC21, ꢀ177.2°; C1ꢀN1ꢀC21ꢀC26, ꢀ90.6°;
C1ꢀN1ꢀC11ꢀH11, ꢀ50.5°.
followed by allylation of the resulting product with cyclohexenyl
bromide (C6H11Br) and base.
The X-ray crystal structure of 7 is shown in Figure 4. As
expected from related molecules such as 4, the planes of the
amide and the aryl ring of 7 are virtually orthogonal. The barrier
for rotation around the NꢀAr bond cannot be directly measured
because of the symmetry of the aryl ring substituents, but by
analogy to 4 it must be well over 30 kcal/mol. Also, as in most
anilides, the amide C2(O1)ꢀN1 bond of 7 adopts the E
geometry in both solution and the solid state.6 The amide
nitrogen substituent adopts a pseudoequatorial position on the
cyclohexenyl ring, and the amide plane is roughly staggered
between the C11 substituents H11 and C12. This geometry is
typical for amides with secondary-alkyl N-substituents and
results from minimization of A-strain.7
To study the time-dependent concentration profiles of pro-
ducts and intermediates during a typical radical reaction, diiodide
7 was treated with Bu3SnH (4.0 equiv) in the presence of the
initiator 1,10-azobis(cyclohexane-1-carbonitrile) (V-40, 0.2 equiv)
in benzene at 80 °C, and the reaction course was followed by GC
analysis (Scheme 1 and Figure 5).
Authentic samples of the intermediate products 8 and 9 (one
iodide reacted) and final products 10 and 119 (both iodides
reacted) were synthesized as standards (see the Supporting
Information). Isomers 8-syn and 8-anti exhibited a single peak
upon GC analysis as a result of their rapid interconversion under
the high-temperature conditions of the analysis (see below), so
only a total amount of 8 was determined.
The two monodehalogenated intermediates 8 and 9 were
identified after 16 min in 8 and 12% yield, respectively. The
targeted final product 11 was already present in 11% yield at that
time. The yields of 8 and 9 soon rose to just over 15%, remained
nearly constant for a time, and then diminished. Meanwhile, the
starting diiodide 7 was consumed, and the final product 11 was
Figure 5. Concentration profiles of major products in the Bu3SnH-
mediated reaction of diiodoanilide 7.
eventually formed in high yield. The reaction was complete after
4 h, and the major product 11 was formed in 86% GC yield along
with a small amount of doubly reduced anilide 10 (5%).8 The
outcome is far superior to prior examples of product selection
at the steady state in terms of both yield (86%) and selectivity
(94/6) of the major product, 11.
The observed time-course profiles in Figure 5 conforms well
with that expected on the basis of the kinetic theory that follows
16272
dx.doi.org/10.1021/ja2070347 |J. Am. Chem. Soc. 2011, 133, 16270–16276

Journal of the American Chemical Society
ARTICLE
Figure 6. Simplified mechanistic analysis of the tin hydride reduction of 7 showing the “major convergence” to 11 and a proposed leakage to the doubly
reduced product 10.
Figure 7. Assessment of the importance of residual axial chirality in the cyclizations of radicals 12 along with four possible conformations of product 9.
from the simplified mechanistic framework shown in Figure 6.
Product selection at the steady state is complex because all of the
intermediates and products can arise from two partially intercon-
nected pathways, a major convergence and a minor convergence.3a
Conveniently, we can neglect the minor convergence for the
analysis in Figure 6 because the process is so selective (Figure S1 in
the Supporting Information shows all of the pathways.)
diastereomeric radicals 12a and 12b. That this diastereotopic-
group-selective process occurs without selectivity is shown by the
observation of equal amounts of the immediately downstream
products 8 and 9 derived from these radicals throughout the
reaction. Radical 12b resembles the radical derived from 4-syn in
Figure 3, so it cyclizes very rapidly to provide 13, which in turn
delivers 9 by hydrogen atom abstraction. In contrast, radical 12a
resembles the radical derived from 4-anti in Figure 3, so it does
not cyclize and instead is reduced to 8-syn. This part of the
Paths diverge immediately with abstraction of either the α- or
β-oriented iodine atom from 7 by Bu3Sn to provide the two
3
16273
dx.doi.org/10.1021/ja2070347 |J. Am. Chem. Soc. 2011, 133, 16270–16276

Journal of the American Chemical Society
ARTICLE
result of the two cyclizations is the formation not of the same radical
13 but of the diastereomeric radicals M*-13 and P*-13.
The ultimate closed-shell products 9 from reduction of these
radicals can presumably interconvert thermally because the
newly formed ring lowers the barrier to NꢀAr rotation (now
associated with a conformational change of the five-membered
ring). Also, the amide CꢀN bond can rotate between E and
Z.12,13 This results in the four conformers of 9 shown in Figure 7.
However, these rotation processes occur after all of the action
with the radicals is over.
Efforts to study 9 by NMR spectroscopic analysis were
complicated by the presence of overlapping broad peaks that
could not be unambiguously assigned to the candidate rotamers.
However, we were able to solve the X-ray crystal structure of 9,
and this is shown in Figure 8.
The crystal exhibited only conformer P*-9Z, in which the
cyclohexyl ring exists in a chair conformation with the nitro-
gen substituent occupying a pseudoequatorial position and the
aryl substituent a pseudoaxial position. This (Z)-amide rotamer
cannot be the primary product of the radical cyclization. The
residual axial chirality of the NꢀAr-bond is clearly visible in the
X-ray structure (bridgehead H atoms and the I atom anti to each
other). At about 50°, the NꢀAr twist angle is reduced relative
to that precursor 7 but is still substantial. The amide CꢀN bond
now nearly eclipses the former allylic CꢀH bond (H12), thus
avoiding A-strain between the methyl group of the acyl sub-
stituent (C15) and the methylene group (C11) of the cyclo-
hexyl ring.
Loosely related amides exhibit both E and Z rotamers in
solution, with the E rotamer often being favored.12,13 We were
not confident that the crystal structure provided the favored
ground-state amide rotamer. Although it seems less likely (based
on A-strain), we also considered that the NꢀAr twist direction in
the crystal structure might also not be favored in the ground state.
To address these issues, the conformational space of 9 with
respect to the axial chirality (torsion around the CArꢀN bond
vector) and rotation of the acetyl groups was explored by state-
of-the-art dispersion-corrected density functional theory (DFT)
at the DSD-BLYP-D3/def2-QZVP//TPSS-D3/def2-TZVP level.14
Details of the calculations and absolute energies (in hartrees) as
well as coordinates of geometry-optimized structures are pro-
vided in the Supporting Information.
In the gas phase, the relative free energies ΔG(298 K) are
0.0 kcal/mol for P*-9E, 0.96 kcal/mol for P*-9Z, 8.66 kcal/mol
for M*-9E, and 7.26 kcal/mol for M*-9Z (for the structures,
see Figure 7). In all of the structures, the cyclohexane ring is in
the same chair-type conformation.
When corrections for the solvent (chloroform) were included,
the ΔG(298 K) values for P*-9E (0.0 kcal/mol) and P*-9Z
(ꢀ0.19 kcal/mol) were reversed, but the NꢀAr twist isomers
M*-9E (8.86 kcal/mol) and M*-9Z (6.65 kcal/mol) remained
much higher in energy. The reversal to favor P*-9Z slightly in
solution results from the larger dipole moment (by about 1.3 D)
of P*-9Z relative to P*-9E. Thus, it seems likely that substantial
amounts of both P*-9E and P*-9Z are present in solution and that
the Z rotamer selectively crystallizes.
Figure 8. X-ray crystal structure of P*-9Z. Selected torsion angles:
O1ꢀC14ꢀN1ꢀC1, 6.5°; C14ꢀN1ꢀC1ꢀC2, ꢀ50.8°; C14ꢀN1ꢀ
C12ꢀH12, ꢀ4.0°.
process is an example in which stereoisomeric precursors (12a
and 12b) subjected to identical reaction conditions give con-
stitutionally different products (8 and 9) rather than stereoiso-
meric products.10
From there, the roles are reversed; abstraction of iodine from 9
now gives a radical that cannot again cyclize and thus is reduced,
while radical 14a derived from 8 is very similar to 12b and rapidly
cyclizes prior to reduction. The net result is that the pathways
reconverge to give the same product 11 in high yield. In short, if
one diastereotopic iodine of 7 is abstracted, then cyclization
precedes direct reduction. If the other is abstracted, then direct
reduction followed by cyclization occurs.
The high selectivity for formation of 11 may seem surprising
since aryl radicals 12a and 12b must react with tin hydride with
virtually the same rate constant.11 However, because the process
occurs at the steady state and the cyclization of 12b is much faster
than 12a, the concentration of 12a must always greatly exceed
that of 12b. This concentration gradient ensures that tin hydride
reacts with 12a and not with 12b. Though the end result is
selectivity of two constitutional isomers (10 and 11), there is a
hidden yet crucial feature of stereoselectivity. In principle, radical
12a can cyclize. But if it were to do so, it would not give the same
radical 13 resulting from cyclization of 12b because the NꢀAr
axis cannot rotate in competition with cyclization.
The consequences of the prostereogenic axis in 7 for the
radical cyclization are explored in Figure 7. Radical 12b (lower
line of structures) is well-positioned to cyclize with simple
twisting of the NꢀAr bond and the allylic CꢀN bond to bring
the radical and the alkene together. However, this does not
provide a “flat” five-membered ring in product 13 (as suggested
by the simplified structures in Figure 6) but instead one with
residual axial chirality, as shown in P*-13 (Figure 7).12,13
In contrast, radical 12a (upper line of structures in Figure 7) is
not well-positioned to cyclize until the allylic CꢀN bond rotates
substantially, with an attendant increase in A-strain. Cyclization
through a conformer like 12a0 now gives a product M*-13 with
axial chirality opposite that of P*-13. In other words, the immediate
The high relative energy of M*-9E with respect to P*-9E (7ꢀ9
kcal/mol) is consistent with the notion that M*-9E must deal
with significant A-strain issues because the NꢀAr twist direction
and the stereocenter configurations are mismatched. The same
mismatching is present in the cyclized radical M*-13 in compar-
ison with P*-13. Because of the developing strain, the activation
16274
dx.doi.org/10.1021/ja2070347 |J. Am. Chem. Soc. 2011, 133, 16270–16276

Journal of the American Chemical Society
ARTICLE
Scheme 2. Radical Cyclization and Trapping Mediated by
Me3SnPPh2
If this notion is correct, then the leaks at both the closed-shell
(8) and open-shell (14) intermediates can be plugged by using a
radical-trapping agent that is larger than a hydrogen atom. Aryl
radicals react with Me3SnPPh2 at rates similar to those with
Bu3SnH,5b,15 so transformation of 7 to a major double-phospha-
nylated product should be feasible.
To this end, we treated 7 with Me3SnPPh2 (6.0 equiv) and
initiator V-40 in benzene (Scheme 2). To simplify the analysis
and product isolation, the reaction mixture was treated with S8 to
afford bis(phosphine sulfide) 15, which was isolated in 63% yield
after chromatography. The transfer of a phosphine group rather
than a hydrogen atom provides a third stereocenter in the
cyclohexyl ring of the product 15. This traditional face-selective
reaction of a radical also occurs with complete stereoselectivity
by reaction from the less hindered exo face of the adjacent
bicyclic ring fusion.16
The structure of 15 was determined by X-ray analysis (see
Figure S5 in the Supporting Information), which confirmed the
relative configuration of the new stereocenter. Otherwise, the
structure of 15 was similar to that of 9 in three key aspects:
(1) the (Z)-amide rotamer was present; (2) the NꢀAr twist
angle was substantial (45°); and (3) the amide CꢀN bond
roughly eclipsed the adjacent cyclohexyl CꢀH bond. These fea-
tures reinforce the stereochemical analysis in Figure 7.
The double phosphanylation is not as clean and high-yielding
as the tin hydride-mediated reaction, and minor side products
resulting from monophosphanylation were detected (see the
Supporting Information). These side products are common in
aryl radical phosphanylation,5b,15 and their amounts can be
reduced to <10% by using a relatively high concentration of
Me3SnPPh2 (0.1 M). Importantly, we looked by both GC
analysis and 31P NMR spectroscopy but were not able to identify
the double-trapped side product 18 that is analogous to 10.
The absence of 18 can be understood by the partial mechan-
istic analysis shown in the lower part of Scheme 2. Abstraction of
the β-iodine atom from 7 is not shown because it follows a path
directly analogous to that in Figure 6; cyclization is followed by
phosphanylation to give 15. Abstraction of the α-iodine atom
gives the aryl radical that does not cyclize and instead is phos-
phanylated with memory of axial chirality (MOC)5b to give 16-
syn. Now this closed-shell intermediate has two large ortho sub-
stituents, so the NꢀAr bond cannot rotate to give 16-anti in
competition with onward reaction, even at 80 °C. Abstraction of
the remaining iodine atom from 16-syn provides radical 17a,
whose cyclization is faster than rotation to 17b again because of
the presence of the larger Ph2P substituent. In short, phospha-
nylation plugs both potential leaks.
energy for the cyclization of 12a0 leading to M*-13 must be
far higher than the activation energy for cyclization of 12b to
P*-13. Neither cyclization nor NꢀAr rotation of aryl radical 12a
or 12a0 can compete with bimolecular tin hydride reduction, so
12a is funneled selectively into 8-syn, as observed in the experi-
ment (see the concentration profiles of the intermediates in
Figure 5).
In short, the overall transformation changes from a stereo-
selective one to a product-selective one (11/10 ratio) because
the difference ratio of cyclization of isomeric radicals 12a and 12b
is so high that effectively no minor stereoisomer results from 12a.
What is the origin of the doubly reduced product 10? In
principle, this could arise from a leakage into the minor con-
vergence at one of several branch points. For example, if radical
12b leaks by reduction with tin hydride rather than cyclization,
then the resulting product will probably also be reduced, giving
10 (see Figure S1 in the Supporting Information). However, this
seems unlikely on the basis of the cyclization results with 4 and
related o-iodoanilides.
’ CONCLUSIONS
Prior measurements of the rotational barrier and cyclization
experiments with 7 and 8 suggest a different explanation for the
origin of 10. Reduction of 12a by tin hydride should occur with
memory of axial chirality to produce only 8-syn,5b but the
attendant replacement of the iodine atom by hydrogen reduces
the rotational barrier for interconversion of 8-syn and 8-anti to
23ꢀ24 kcal molꢀ1.5a At 80 °C, rotation of the NꢀAr bond of 8 is
probably competitive to some extent with onward radical gen-
eration. In addition, the previous results suggest that radicals 14a
and 14b might also partially equilibrate.5a However, radical 14b is
of the slow-cyclizing variety, so the return back to the major
convergence is the more likely of the two rotations here.
We have developed a high-yielding and highly selective
process for the formation of cyclized products such as 11 and
15 from substrate 7 containing diastereotopic radical precursors.
This transformation is founded on the concept of product
selection at the steady state. In classical group-selective pro-
cesses, the maximum yield of a major product is fixed in the initial
group-selective step. If there is no selectivity in this step, then a
product yield cannot exceed 50%. In essence, there is one path
from the precursor to the product, so a wrong turn at the group-
selection step cannot be tolerated.
At the steady state, the initial group selectivity plays only a
supporting role. The lead role is played by the reaction topology
16275
dx.doi.org/10.1021/ja2070347 |J. Am. Chem. Soc. 2011, 133, 16270–16276

Journal of the American Chemical Society
ARTICLE
because there are two main paths to the product. If one
diastereotopic iodine atom is abstracted in the first step, then
the system follows one path to the product. Abstraction of the
other iodine atom might initially appear to be a wrong turn, but
instead the product is simply reached by the other path. In
essence, the system is self-correcting provided that the kinetics
are in order to prevent any of several other possible wrong turns
along the way.
This unusual case of reaction selectivity has been demon-
strated on aryl radical cascades comprising cyclization and
bimolecular trapping. The net result is product selection between
two constitutional isomers (reduced and cyclized), and high
yields of cyclized products are obtained. However, underlying
this outcome is a key aspect of stereoselectivity. The initial
rotationally isomeric radicals 12 cyclize not to the same product
at different rates but instead to different products (NꢀAr twist
isomers 13) at different rates. One of these stereoisomeric pro-
ducts is so high in energy that the radical cyclization producing it
cannot compete at all with bimolecular reduction.
1997, 8, 3955–3975. (c) Itai, A.; Toriumi, Y.; Saito, S.; Kagechika, H.;
Shudo, K. J. Am. Chem. Soc. 1992, 114, 10649–10650.
(7) Hoffmann, R. W. Chem. Rev. 1989, 89, 1841–1860.
(8) Traces of two unidentified byproducts (7% and 2%) accounted
for the balance of the starting material.
(9) According to NMR analysis, final product 11 was a 78:22 mixture
of amide CꢀN rotamers. The proton in the position ortho to the
nitrogen atom of the major rotamer showed a downfield shift of about
1.0 ppm relative to the other signals of the aromatic protons. This may
have been caused by the proximity of the carbonyl group to this proton
in the Z isomer, and this isomer can be tentatively assigned as the
predominant one.
(10) (a) Kagan, H. B. Croat. Chem. Acta 1996, 69, 669–680.
(b) Kumar, R. R.; Kagan, H. B. Adv. Synth. Catal. 2010, 352, 231–242.
(11) Chatgilialoglu, C.; Newcomb, M. Adv. Organomet. Chem. 1999,
44, 67–112.
(12) (a) Tabata, H.; Nakagomi, J.; Morizono, D.; Oshitari, T.;
Takahashi, H.; Natsugari, H. Angew. Chem., Int. Ed. 2011, 50, 3075–3079.
(b) Takahashi, H.; Wakamatsu, S.; Tabata, H.; Oshitari, T.; Harada, A.;
Inoue, K.; Natsugari, H. Org. Lett. 2011, 13, 760–763.
(13) Keller, A. I. Ph.D. Thesis, UniversityofPittsburgh, Pittsburgh, PA,
2007; http://etd.library.pitt.edu/ETD/available/etd-04122007-142347/
(14) (a) Kozuch, S.; Gruzman, D.; Martin, J. M. L. J. Phys. Chem. C
2010, 114, 20801–20808. (b) Goerigk, L.; Grimme, S. J. Chem. Theory
Comput. 2011, 7, 291–309. (c) Goerigk, L.; Grimme, S. Phys. Chem.
Chem. Phys. 2011, 13, 6670–6688. (d) Grimme, S.; Antony, J.; Ehrlich,
S.; Krieg, H. J. Chem. Phys. 2010, 132, No. 154104. (e) Grimme, S.;
Ehrlich, S.; Goerigk, L. J. Comput. Chem. 2011, 32, 1456–1465. (f)
Weigend, F.; Furche, F.; Ahlrichs, R. J. Chem. Phys. 2003, 119,
12753–12762. (g) Sch€afer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys.
1994, 100, 5829–5835. (h) Weigend, F.; Ahlrichs, R. Phys. Chem. Chem.
Phys. 2005, 7, 3297–3305.
The ability to map out two separate, nonintersecting pathways
to the same destination is a feature of the steady state, not of
radical reactions per se. Therefore, the principles of product
selection outlined herein are applicable in other areas such as
catalysis where reactive intermediates are involved.
’ ASSOCIATED CONTENT
S
Supporting Information. Experimental procedures and
b
compound characterization data. This material is available free of
(15) Vaillard, S. E.; M€uck-Lichtenfeld, C.; Grimme, S.; Studer, A.
Angew. Chem., Int. Ed. 2007, 46, 6533–6536.
charge via the Internet at http://pubs.acs.org.
(16) Curran, D. P.; Porter, N. A.; Giese, B. Stereochemistry of Radical
Reactions:Concepts, Guidelines, andSyntheticApplications; VCH: Weinheim,
Germany, 1996.
’ AUTHOR INFORMATION
Corresponding Author
studer@uni-muenster.de; curran@pitt.edu
’ ACKNOWLEDGMENT
We thank the DFG and the National Science Foundation
for support of the work in Germany and the U.S., respectively.
D.P.C. thanks the Humboldt Foundation for a Senior Research
Fellowship.
’ REFERENCES
(1) Eliel, E. L.; Wilen, S. Stereochemistry of Organic Compunds; Wiley-
Interscience: New York, 1994.
(2) The analysis assumes that there are no additional reactions
outside of those shown.
(3) (a) DeMello, N. C.; Curran, D. P. J. Am. Chem. Soc. 1998,
120, 329–341. (b) Curran, D. P.; Qi, H. Y.; DeMello, N. C.; Lin, C. H.
J. Am. Chem. Soc. 1994, 116, 8430–8431.
(4) (a) Curran, D. P.; Lin, C. H.; DeMello, N.; Junggebauer, J. J. Am.
Chem. Soc. 1998, 120, 342–351. (b) Andrukiewicz, R.; Cmoch, P.;
Gawez, A.; Staliꢀnski, K. J. Org. Chem. 2004, 69, 1844–1848. (c) Stalinski,
K.; Curran, D. P. J. Org. Chem. 2002, 67, 2982–2988.
(5) (a) Guthrie, D. B.; Geib, S. J.; Curran, D. P. J. Am. Chem. Soc.
2011, 133, 115–122. (b) Bruch, A.; Ambrosius, A.; Fr€ohlich, R.; Studer,
A.; Guthrie, D. B.; Zhang, H.; Curran, D. P. J. Am. Chem. Soc. 2010,
132, 11452–11454.
(6) (a) Stewart, W. E.; Siddall, T. H. Chem. Rev. 1970, 70, 517–551.
(b) Curran, D. P.; Hale, G. R.; Geib, S. J.; Balog, A.; Cass, Q. B.; Degani,
A. L. G.; Hernandes, M. Z.; Freitas, L. C. G. Tetrahedron: Asymmetry
16276
dx.doi.org/10.1021/ja2070347 |J. Am. Chem. Soc. 2011, 133, 16270–16276