Higher selectivity at higher temperatures! Effect of precursor stereochemistry on diastereoselectivity in radical allylations. Insight into the role of the Lewis acid

Article abstract of DOI:10.1021/ja001456j

A detailed investigation of the effects of Lewis acid, temperature, and trapping efficiency of functionalized allylstannanes on diastereoselective radical allylation reactions of α-bromooxazolidinone imides 1 and 2 was conducted. Results indicate that des

Full text of DOI:10.1021/ja001456j

J. Am. Chem. Soc. 2000, 122, 8873-8879
8873
Higher Selectivity at Higher Temperatures! Effect of Precursor
Stereochemistry on Diastereoselectivity in Radical Allylations. Insight
into the Role of the Lewis Acid
Mukund P. Sibi* and Tara R. Rheault1
Contribution from the Department of Chemistry, North Dakota State UniVersity,
Fargo, North Dakota 58105-5516
ReceiVed April 26, 2000
Abstract: A detailed investigation of the effects of Lewis acid, temperature, and trapping efficiency of
functionalized allylstannanes on diastereoselective radical allylation reactions of R-bromooxazolidinone imides
1 and 2 was conducted. Results indicate that despite the addition of Lewis acids, a bidentate chelated radical
intermediate 21 may be accessible from only one diastereomer of starting material due to steric interactions in
20 that are not present in chelated intermediate 17. It is shown that application of the appropriate Lewis acid,
increasing temperatures, and slower allylstannane traps all facilitate formation of 21. Thus highly stereoselective
radical allylations (>50:1, Tables 2 and 3) can be performed at room temperature as well as low temperatures.
Construction of carbon-carbon bonds with high selectivity
in acyclic systems by radical processes has been the subject of
investigation in many laboratories.2,3 A major driving force in
the recent success of such processes was the realization that
Lewis acids can be effectively used to control the conformations
of either the radical intermediate or the radical trap and also
enhance the reactivity of certain substrates for radical addition.4
The dogma in free radical chemistry is that the precursor
configuration has little or no impact on the levels of stereoin-
duction in diastereoselective transformations and this is often
cited as an advantage over ionic processes. The reason for this
conclusion is that a prochiral radical intermediate generated is
generally planar or slightly tetrahedral with a very low barrier
for interconversion, and thus it has no memory of its origin. As
part of an ongoing program toward understanding the basis for
diastereoselectivity in Lewis acid-mediated radical reactions,
we became interested in exploring the issue of precursor
stereochemistry on the levels of selectivity in radical allylations.
The basic postulate we wished to examine was whether
diastereomeric bromides 1 and 2 will lead to the same
conformationally locked radical intermediate 3 or whether non-
Lewis acid (or singly bound) controlled rotamers 5 or 6 play a
role in the diastereoselective outcome of these allylations.5 In
addition, we wished to probe whether variations in temperature
and/or tuning of allylstannane reactivity by appropriate func-
tionalization would have an impact on which reactive rotamer
is finally trapped. The evaluation of this hypothesis is the focus
of this paper.
(1) This material is based upon work supported under a National Science
Foundation Graduate Fellowship.
(2) For monographs and leading review articles, see: (a) Curran, D. P.;
Porter, N. A.; Giese, B. Stereochemistry of Radical Reactions; VCH:
Weinheim, 1995. (b) Sibi, M. P.; Porter, N. A. Acc. Chem. Res. 1999, 32,
163. (c) Porter, N. A.; Giese, B.; Curran, D. P. Acc. Chem. Res. 1991, 24,
296. (d) Smadja, W. Synlett 1994, 1.
(3) For selected recent reports on stereoselective allylation, see: (a)
Renaud, P.; Moufid, N.; Kuo, L. H.; Curran, D. P. J. Org. Chem. 1994, 59,
3547. (b) Yamamoto, Y.; Onuki, S.; Yumoto, M.; Asao, N. J. Am. Chem.
Soc. 1994, 116, 421. (c) Wu, J. H.; Radinov, R.; Porter, N. A. J. Am. Chem.
Soc. 1995, 117, 11029. (d) Porter, N. A.; Wu, J. H.; Zhang, G.; Reed, A.
D. J. Org. Chem. 1997, 62, 6702. (e) Murakata, M.; Jono, T.; Mizuno, Y.;
Hoshino, O. J. Am. Chem. Soc. 1997, 119, 11713. (f) Fhal, A.-R.; Renaud,
P. Tetrahedron Lett. 1997, 38, 2661. (g) Rosenstein, I. J.; Tynan, T. A.
Tetrahedron Lett. 1998, 39, 8429.
(4) For a recent review, see: Renaud, P.; Gerster, M. Angew. Chem.,
Int. Ed. Engl. 1998, 37, 2562. For Lewis acid-mediated enantioselective
conjugate additions, see: Sibi, M. P.; Ji, J.; Wu, J. H.; Gu¨rtler, S.; Porter,
N. A. J. Am. Chem. Soc. 1996, 118, 9200. Sibi, M. P.; Ji, J. J. Org. Chem.
1997, 62, 3800. For a discussion on Lewis acid-assisted atom-transfer
reactions of R-halo amides, see: Mero, C. L.; Porter, N. A. J. Am. Chem.
Soc. 1999, 121, 5155.
Generally, increasing reaction temperatures has a dramatic
effect on stereoselectivity in that higher temperatures lead to
decreased levels of stereocontrol. However, there have been
sporadic reports in the literature illustrating an inverse relation-
10.1021/ja001456j CCC: $19.00 © 2000 American Chemical Society
Published on Web 08/30/2000

8874 J. Am. Chem. Soc., Vol. 122, No. 37, 2000
Table 1. Effect of Lewis Acid on F-Diastereoselectivitya
Sibi and Rheault
Figure 1.
yield,b
%
ratio
entry lewis acid
reaction conditions
(RS:RR)c
the same as that of enolate allylation and the two reactions
provide comparable diastereoselectivity.10
Commencing our study of diastereoselective radical allyla-
tions, three readily accessible allylstannanes were chosen: 9,
10, and 11. The relative reactivity of the three different
1
2
3
4
none
-78 °C, CH2Cl2, 3 h
93
85
94
1:1.8
1:1.4
39:1
5:1
BF3‚OEt2 -78 °C, CH2Cl2, 2.5 h
MgBr2
-78 °C, CH2Cl2, 2 h
d
Yb(OTf)3 -78 °C, CH2Cl2/THF (1:1), 2 h 64 (14)e
a Two equivs of Lewis acid was used in all reactions unless otherwise
noted. b Yield of isolated product. c Diastereomer ratios were determined
by 1H NMR (400 MHz). d 1 equiv of the Lewis acid was used. e Yield
of the reduced product.
ship between temperature and stereoselectivity.6 In this work
we show a consistent pattern of higher temperatures leading to
significantly increased levels of diastereoselectivity.
allylstannanes toward the electrophilic radical intermediate
(generated from either 1 or 2) was assessed qualitatively through
competitive experiments. Control experiments using equal molar
amounts of two allylstannanes in trapping the intermediate
radical established the reactivity in the following order 11 >
10 > 9, with a relative reactivity ratio of 25:5:1.11
To probe the dependence of selectivity on precursor stereo-
chemistry, we undertook a detailed study of reactions of 1 and
2 with each allylstannane (9-11), various Lewis acids, and
various temperatures. Diastereomerically pure starting materials
were prepared by acylation of the chiral auxiliary with racemic
2-bromopropionyl bromide under standard conditions (n-BuLi,
THF, -78 °C) followed by column chromatographic separation
of the diastereomers.12 With the pure compounds in hand, two
Lewis acids, MgBr2‚Et2O and Yb(OTf)3, were chosen for the
study. Initial reactions conditions employed were -78 °C, Et3B/
O2, and 2.5 equivs of the allylating agent. Results from these
experiments are presented in Tables 2 and 3.
Allylation of either 1 or 2 with allyltributyl tin (9) using 2
equivs of MgBr2 at -78 °C gave the allylated product 8 in good
yield and ∼40:1 selectivity (Table 2, entries 1 and 2).13 The
stereochemistry of the product was the same regardless of the
starting geometry, i.e., starting from either 1 or 2 gave 8 of the
same configuration. The observed high selectivity suggests the
following: (1) both diastereomers produce the same radical
intermediate, (2) the intermediate radical is chelated to the Lewis
acid as an s-cis rotamer and approach of the reagent occurs from
the face opposite to the bulky chiral auxiliary, (3) assuming
that bromine atom abstraction may or may not occur from a
chelated form of the substrate 1 and 2, rotation across the
radical-carbonyl (COCHCH3) bond (structures 4f3) as well as
Results
We have previously reported on the allylation of oxazolidi-
nones of the type 7 under radical conditions using allyltributyltin
and triethylborane-oxygen as an initiator at -78 °C. Salient
results from this study are tabulated in Table 1.7 General
conclusions from this study are (1) the allylations are nonselec-
tive in the absence of Lewis acid additives (entry 1), (2) single
point binding Lewis acids such as BF3‚Et2O gave the allylated
product with low selectivity, (3) Lewis acids capable of
coordinating to both the carbonyl oxygens gave high selectivity
(entry 3, data with other Lewis acids not shown), (4) the use of
ytterbium triflate, which was the best Lewis acid for conjugate
additions, gave only moderate selectivity (entry 4), and (5) the
chiral oxazolidinone derived from diphenylalaninol8 was found
to be optimal for obtaining high selectivity (compared to
oxazolidinone auxiliaries derived from phenylglycinol or phen-
ylalaninol). The absolute configuration of the newly formed
stereocenter was established as S by hydrolysis and comparison
of the sign of rotation with that of a known compound.9 This
result agrees with the model shown in Figure 1where the si
face of the molecule is blocked by the bulky portion of the chiral
auxiliary, leaving the re face exposed to trap the allyltin. The
sense of stereoinduction in the radical mediated allylation was
(5) For limited examples of selectivity dependence on precursor stereo-
chemistry, see: (a) Guindon, Y.; Rancourt, J. J. Org. Chem. 1998, 63, 6554.
(b) Gerster, M.; Audergon, L.; Moufid, N.; Renaud, P. Tetrahedron Lett.
1996, 35, 6335. (c) Hanessian, S.; Yang, H.; Schaum, R. J. Am. Chem.
Soc. 1996, 118, 2507.
(6) For examples describing an inverse relationship between temperature
and stereoselectivity, see the following references. Asymmetric decarboxy-
lation: Musart, J.; He´nin, F.; Aboulhoda, S. J. Tetrahedron: Asymmetry
1997, 8, 381. Hydrogenation: Landis, C. R.; Halpern, J. J. Am. Chem. Soc.
1987, 109, 1746. Oxazaborolidine reductions: Zhang, Y.-W.; Shen, Z.-X.;
Liu, C.-L.; Chen, W.-Y. Synth. Commun. 1995, 25, 3407. Stone, G.
Tetrahedron: Asymmetry 1994, 5, 465. Asymmetric protonation using
SmI2: Takeuchi, S.; Miyoshi, N.; Hirata, K.; Hayashida, H.; Ohgo, Y. Bull.
Chem. Soc. Jpn. 1992, 65, 2001. Asymmetric addition of Grignards to
aldehydes: Marko´, I. E.; Chesney, A.; Hollinshead, D. M. Tetrahedron:
Asymmetry 1994, 5, 569.
(10) Enolate allylation of 3-(1-oxopropyl)-4-(diphenylmethyl)-2-oxazo-
lidinone gave 8 in 63% yield with >99% de. Sibi, M. P.; Deshpande, P.
K.; Ji, J. Tetrahedron Lett. 1995, 36, 8965.
(11) Allyltributyl tin is commercially available. Methylallylstannane and
carbomethoxyallylstannane were prepared by known procedures, see: (a)
Baldwin, J. E.; Adlington, R. M.; Mitchell, M. B.; Robertson, J. Tetrahedron
1991, 47, 5901. (b) Schwartz, W. T., Jr.; Post, H. W. J. Organomet. Chem.
1964, 2, 357. For similar competitive experiments, see: Landais, Y.;
Planchenault, D. Tetrahedron 1995, 44, 12 097. Renaud, P.; Gerster, M.;
Ribezzo, M. Chimia 1994, 48, 366.
(7) Sibi, M. P.; Ji, J. Angew. Chem., Int. Ed. Engl. 1996, 35, 190.
(8) (a) For synthesis of the auxiliary, see: Sibi, M. P.; Deshpande, P.
K.; La Loggia, A. J.; Christensen, J. W. Tetrahedron Lett. 1995, 36, 8961.
(b) Sibi, M. P. Aldrichim. Acta 1999, 32, 93. This chiral auxiliary is now
available commercially from Aldrich Chemical Co.
(12) Absolute configuration of R-bromoimides 1 and 2 were determined
by hydrolysis with LiOH/H2O2 and comparison of the optical rotation with
commercially available (S)-2-bromopropionic acid.
(9) On hydrolysis with LiOH/H2O2 8 gave (S)-2-methyl-4-butenoic acid
[R]26D ) 10.5° (c 1.15, CHCl3); lit. [R]26D ) 10.5° (CHCl3). Riley, R. G.;
Silverstein, R. M. Tetrahedron 1974, 30, 1171.
(13) Two equivalents of MgBr2 were used in order to ensure that at least
one equivalent was indeed the Lewis acid in case of possible contamina-
tion by oxidation products.

Higher SelectiVity at Higher Temperatures
J. Am. Chem. Soc., Vol. 122, No. 37, 2000 8875
Table 2. Effect of Substrate Stereochemistry and Reaction
Temperature on Product Diastereoselectivity
Table 3. Effect of Substrate Stereochemistry and Reaction
Temperature on Product Diastereoselectivity
substrate
entry stereochem
substrate
entry stereochem
temp
R
% yielda,b
ratioc
temp
R
% yielda,b
ratioc
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
(R)
(S)
(R)
(S)
(R)
(S)
(R)
(S)
(R)
(S)
(R)
(S)
(R)
(S)
(R)
(S)
(R)
(S)
-78 °C
-78 °C
0 °C
0 °C
RT
H
H
H
H
H
H
70
91
90
92
82
88
78
92
78
95
75
95
25(50)
90(10)
72
70
80
39:1 (S)
39:1 (S)
>50:1 (S)
>50:1 (S)
>50:1 (S)
>50:1 (S)
24:1 (S)
12:1 (S)
>50:1 (S)
20:1 (S)
>50:1 (S)
30:1 (S)
3:1 (S)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
(R)
(S)
(R)
(S)
(R)
(S)
(R)
(S)
(R)
(S)
(R)
(S)
(R)
(S)
(R)
(S)
(R)
(S)
-78 °C
-78 °C
0 °C
0 °C
RT
H
H
H
H
H
H
46 (49)
90
95
93
92
95
81
93
88
86
71
95
41 (45)
70 (23)
81
95
70
2:1 (S)
5:1 (S)
50:1 (S)
50:1 (S)
>50:1 (S)
>50:1 (S)
1:1
RT
RT
-78 °C Me
-78 °C Me
0 °C
0 °C
RT
-78 °C Me
-78 °C Me
0 °C
0 °C
RT
1:1
Me
Me
Me
Me
Me
Me
Me
Me
>50:1 (S)
22:1 (S)
>50:1 (S)
20:1 (S)
1:2 (R)
1:2 (R)
4:1 (S)
RT
RT
-78 °C CO2Me
-78 °C CO2Me
0 °C
0 °C
RT
-78 °C CO2Me
-78 °C CO2Me
0 °C
0 °C
RT
1:1
CO2Me
CO2Me
CO2Me
CO2Me
43:1 (S)
23:1 (S)
40:1 (S)
20:1 (S)
CO2Me
CO2Me
CO2Me
CO2Me
4:1 (S)
5:1 (S)
4:1 (S)
RT
68(27)
RT
76
a Yield of isolated product. b Yield of recovered starting material in
parentheses. c Diastereomer ratios were determined by 1H NMR
spectroscopy (400 MHz).
a Yield of isolated product. b Yield of recovered starting material in
parentheses. c Diastereomer ratios were determined by 1H NMR
spectroscopy (400 MHz).
rotation across the N-carbonyl (NCO) bond (structures 5 or 6f3)
in the intermediate is facile in comparison to its trapping with
allyltin. Surprisingly, increasing the reaction temperature (entries
3-6) led to substantial improvements in selectivity as compared
to -78 °C experiments irrespective of precursor stereochemistry.
The next set of experiments used (2-methylallyl)tributyltin
(10) as the trapping agent. Based on our model, we assigned
the stereochemistry of products 12 and 13 by analogy, assuming
the chelated radical intermediate will trap allyltin compounds
predominantly from the re face. In comparison to the parent
system, this trap gave lower levels of selectivity (compare entry
1 with 7). More interestingly, the two diastereomeric starting
materials furnished 12 with a large difference in level of
selectivity: 24:1 for 1 and 12:1 for 2 (entry 7 and 8). However,
the product configuration was the same in both cases. Of the
two diastereomeric starting materials, the R,S isomer 2 was
found to be more reactive and less selective.14 A similar trend
as in the case of the parent allyltin (9) was also observed for
the 2-methylallyltin system 10: increasing reaction temperature
led to higher selectivity (entries 9-12).
As before, the acrylate tin reagent 11 was examined as the
allylating agent. Unlike the previous two allyl tins (9 and 10),
reactions with compound 11 at -78 °C were nonselective with
either diastereomer 1 or 2 (compare entry 1 or 7 with 13 and
14). However, increasing the reaction temperature led to
dramatic changes in levels of selectivity with a maximum of
43:1 for the R,R isomer 1 at room temperature (entry 15). The
configuration of product 13 was the same starting from 1 or 2.
A similar trend in precursor geometry-selectivity was observed
for both 11 and 10 in that the R,R isomer 1 was more selective
than the R,S isomer 2 at all temperatures (for example, compare
entry 7 with 8; 9 with 10; 15 with 16; 17 with 18).
Lanthanide triflates have proven to be a versatile class of
Lewis acids.15 We have successfully used these Lewis acids in
conjugate radical reactions.16 Ytterbium triflate was chosen as
a representative Lewis acid and results from its use in allylations
are tabulated in Table 3. A parallel series of experiments with
magnesium bromide was undertaken (eq 3). Because of limited
solubility in CH2Cl2, reactions with Yb(OTf)3 as a Lewis acid
used a mixed solvent system (CH2Cl2/THF 2:1).
Reaction of 1 or 2 with allyltin 9 and 1 eq of Yb(OTf)3 as a
Lewis acid at -78 °C was only moderately selective (entries 1
and 2). This is in contrast to the high selectivity observed with
magnesium bromide as a Lewis acid. Temperature had a
significant effect on the selectivity (entries 3-6). Reactions at
room temperature gave >50:1 starting with either 1 or 2.
Reactions with allyltin 10 showed a similar dependency in that
higher temperatures gave higher selectivity. Additionally, the
R,R diastereomer 1 gave higher selectivity (>50:1, entry 9)
compared to the R,S diastereomer 2 (22:1, entry 10). Reactions
with the allyltin reagent 11 were very different and were
essentially nonselective at low temperature (entries 13 and 14);
increase in temperature led to only marginal improvement in
selectivity (entries 15-18).
The low selectivity with Yb(OTf)3 at all temperatures for the
acrylate tin 11 implied that the R,â-unsaturated carbonyl
compound sequestered the Lewis acid from the less Lewis basic
(15) For a recent review, see: Kobayashi, S. Eur. J. Org. Chem. 1999,
1. Also see: Kobayashi, S Synlett 1994, 689.
(16) (a) Sibi, M. P.; Jasperse, C. P.; Ji, J. J. Am. Chem. Soc. 1995, 117,
10779. (b) Sibi, M. P.; Ji, J. J. Org. Chem. 1996, 61, 6090. (c) Sibi, M. P.;
Ji, J. Angew. Chem., Int. Ed. Engl. 1997, 36, 274. (d) Sibi, M. P.; Ji, J.;
Sausker, J. B.; Jasperse, C. P. J. Am. Chem. Soc. 1999, 121, 7717.
(14) This is based on recovered starting material when a 3:1 diastereo-
meric mixture was used as the reactant and reactions were stopped after
partial conversion.

8876 J. Am. Chem. Soc., Vol. 122, No. 37, 2000
Table 4. Effect of Lewis Acid on Substrate Epimerizationa
Sibi and Rheault
pure (RR)1
pure (RS)2
pure (RS)2
entry temp MgBr2 (RR:RS) MgBr2 (RR:RS) Yb(OTf)3 (RR:RS)
1
2
3
-78 °C
0 °C
RT
13:1
8:1
3:1
1:12
1:5
3:1
0:100
0:100
0:100
a Diastereomerically pure starting materials were stirred with the
Lewis acid for 2 h and the amount of epimerization was determined
by H NMR spectroscopy (400 MHz).
example shows that the intermediate radical, if generated from
a precomplexed precursor, can be trapped with high selectivity
even at low temperatures (see Scheme 1, path D).
1
R-bromo substrates (1 or 2).17 In such a case, no Lewis acid
would be left over to chelate the substrate (1 or 2) and hence
no facial selectivity would be provided for diastereoselective
trapping by the allyltin. To ensure that uncomplexed Lewis acid
would remain to coordinate to the substrate, 6 equivs of Yb-
(OTf)3 were used. Employing these reaction conditions at room
temperature led to product 13 with diastereoselective ratios of
45:1 and 49:1 (R,S:R,R) for reactions using substrates 1 and 2,
respectively.
To rule out any competition from ionic reactions, control
experiments were performed. The allylation was not possible
in the absence of the radical initiator triethylborane/oxygen using
either MgBr2 or Yb(OTf)3 as a Lewis acid. Conducting the
allylation in the presence of a radical inhibitor, galvinoxyl, led
to no reaction and thus provides additional support for a radical
pathway.
The formation of product with S configuration starting from
either diastereomeric starting material and the variation in level
of selectivity with precursor stereochemistry led us to explore
if there was epimerization during the reaction.18 Diastereomeri-
cally pure starting materials were stirred in the presence of the
Lewis acids (MgBr2 or Yb(OTf)3) at different temperatures for
2 h and the amount of epimerization was determined by NMR.
It was found early on that MgBr2 epimerized 1 and 2 to different
extents at different temperatures (Table 4). At low temperatures
the epimerization was very slow (entry 1), but at higher
temperatures it occurred readily (entry 3). In contrast, experi-
ments with Yb(OTf)3 showed no epimerization over the same
time period regardless of the temperature. Initially, it appeared
that the amount of substrate epimerization positively correlated
with the diastereoselectivity of the allylation. However, high
levels of selectivity are achieved with Yb(OTf)3 and the R,S
diastereomer of substrate (2) at 0 °C and room temperature,
even though substrate epimerization is not a factor (Table 3,
entries 3-6 and 9-12). These results suggest that substrate
epimerization and resultant radical generation exclusively from
1 is not responsible for the seemingly contradictory high
selectivity at higher temperatures.
The lack of selectivity with Yb(OTf)3 at low temperatures
with any of the three traps was puzzling. We wanted to
demonstrate that a similar radical intermediate, generated in an
alternate fashion from a substrate precomplexed with the Lewis
acid Yb(OTf)3, could be selectively trapped at low tempera-
tures.19 As acrylate 14 is a better donor than the bromides 1 or
2, it will chelate Yb(OTf)3 and generate the radical in the desired
s-cis conformation. Thus addition of ethyl radical to acrylate
14 in the presence of 1 equiv of Yb(OTf)3 followed by trapping
with methylallyl stannane (10) was carried out (eq 4). This
The main facts which emerge from the data presented in the
results section are as follows: (1) higher temperatures lead to
higher selectivity in allylations, (2) the R,R starting material is
more selective and less reactive, (3) 2-carboxymethylallylstan-
nane is the most reactive trap, followed by 2-methylallylstannane
and last by the simple allylstannane, (4) magnesium bromide
epimerizes the starting material slowly at low temperatures and
the rate increases at higher temperatures, (5) no epimerization
occurs with Yb(OTf)3, (6) reactions with MgBr2 as a Lewis
acid are selective at low temperature in contrast to nonselective
reactions using Yb(OTf)3, and (7) tandem addition/allylation
with a precomplexed substrate is highly selective.
Discussion
Initially, our goal was to use different allylstannanes as
trapping reagents for the allylation of 1 and 2 under the optimal
conditions established for reactions with allyltributyltin. It was
assumed that both diastereomers 1 and 2 should form chelated
intermediates and that the product radical formed from either
diastereomer of starting material should be identical due to the
trigonal planar nature of radical intermediates. To this end our
initial experiments were conducted on a diastereomeric mixture
of the R-bromoimides (epimeric at the R-carbon) using substi-
tuted allyltributyltin 10 and 11 as the trap. Careful analysis of
the reaction mixture including unreacted starting material
suggested that our initial assumption was incorrect: the two
diastereomers reacted at different rates and different selectiVity.
This contrasts with the indiscriminate trapping by the parent
allyltributyl tin 9. This result revived questions regarding the
reactive radical conformation and rotamer issues thought
previously solved by the addition of chelating Lewis acids. In
addition, relative reactivity differences of the functionalized traps
compared to the parent allylstannane needed to be carefully
reconsidered.
The major issue we needed to address involved the relative
orientation, syn vs anti, of the two carbonyl oxygens present in
the substrate. In the absence of chelating Lewis acids, dipole-
dipole interactions favor an anti ground-state orientation with
respect to the two carbonyl oxygens (16 and 19). Recently,
Lewis acids have been employed in such systems in order to
hold the two carbonyl oxygens in a syn fashion, thus providing
rotamer control for pending asymmetric manipulations. Even
though the rotamer issues surrounding oxazolidinone systems
appear resolved by the addition of chelating Lewis acids, it is
possible that subtle differences in the Lewis acids themselves
as well as steric interactions that may arise in the chelated form
of one diastereomer (17 vs 20) may also play a role in the
stereochemical outcome of these reactions (Scheme 1).5
Scheme 1 takes into account all of our observations. Of the
two diastereomers (R,R and R,S), only R,R is capable of forming
a chelate with a Lewis acid prior to radical generation (compare
17 and 20). However, only a portion of the R,R substrate will
exist in a conformation which is predisposed to form a chelate
with minimal energy constraints. This scenario renders the R,R
(17) Wu, J. H.; Zhang, G.; Porter, N. A. Tetrahedron Lett. 1997, 38,
2067.
(18) For examples using epimerization of R-halo esters in dynamic kinetic
resolution processes, see: Ben, R. N.; Durst, T. J. Org. Chem. 1999, 64,
7700. O’Meara, J. A.; Gardee, N.; Jung, M.; Ben, R. N.; Durst, T. J. Org.
Chem. 1998, 63, 3117. Koh, K.; Durst, T. J. Org. Chem. 1994, 59, 4683.
Durst, T.; Koh, K.; Ben, R. N. Tetrahedron Lett. 1993, 34, 4476.
(19) Sibi, M. P.; Ji, J. J. Org. Chem. 1996, 61, 6090.

Higher SelectiVity at Higher Temperatures
J. Am. Chem. Soc., Vol. 122, No. 37, 2000 8877
Scheme 1
substrate only partially chelated at low temperatures. Reaction
occurs through path A, leading to product 22 with high levels
of selectivity, and alternatively reaction can also occur through
path A1, where increasing the trapping time through variation
in the allylstannane leads to higher levels of selectivity. On the
other hand, in order for the Lewis acid to form a chelated
intermediate with the R,S diastereomer 19, it must adopt a
sterically hindered conformation 20, as a result of steric
interactions that arise between the bromide and the bulky
diphenylmethyl substituent of the of the oxazolidinone chiral
auxiliary. Rotamer 20 is not energetically accessible, disfavoring
a chelated intermediate and resulting in loss of facial discrimi-
nation and little or no diastereoselectivity in the products (path
B to product 24).
important factor in determining the diastereoselectivity of the
functionalized allylstannane additions is the bond rotation, which
allows the nonchelated (monocoordinated) radical intermediate
23 to form the chelated intermediate 21. Increasing reaction
temperatures, slower radical traps, and the appropriate choice
of Lewis acid all function to facilitate this bond rotation and
lead to selective trapping through intermediate 21.
Starting with the R,S substrate 19, this follows with Scheme
1 in that at low temperatures where epimerization is not an issue
(see Table 4, entry 1), the substrate is unable to form a chelate
with either Lewis acid before radical generation. As a result,
uncomplexed or singly bound substrate will react with tributyltin
radical to form intermediate radical 23. If the trap is too fast,
as it is in the case with R ) CO2Me, then radical 23 is trapped
indiscriminately to form equal amounts of diastereomeric
products (Table 2, entries 13 and 14). With a slightly slower
trap, 10, the bond rotation and formation of chelated intermediate
21 is not complete and both 21 and 23 are trapped, leading to
products enriched in the R,S diastereomer resulting in a ratio
of 12:1 (Table 2; entry 8). Finally, using the slowest trap, simple
allylstannane, enough time is available for formation of the
bidentate complex 21 and trapping results in highly diastereo-
Contrary to convention, increasing reaction temperatures led
to higher diastereoselectivity with maximum selectivity being
achieved not at -78 °C but at 0 °C or room temperature. Higher
temperatures allow enough energy for the nonchelated (mono-
coordinated) radical intermediate 23 to undergo bond rotation
followed by chelation to form the rigid intermediate 21 (path
C). It is this chelated intermediate 21 which can then be
selectively trapped by the functionalized allylstannane. The most

8878 J. Am. Chem. Soc., Vol. 122, No. 37, 2000
Sibi and Rheault
addition, the solution was stirred for 10 min at -78 °C. A solution of
2-bromopropionyl bromide (1.296 g, 6 mmol) in 5 mL of THF was
then added dropwise at -78 °C over 10 min and the mixture was stirred
at -78 °C for 0.5 h. The reaction was quenched with 1 mL of saturated
NH4Cl solution and extracted with ethyl acetate (3 × 30 mL). The
organic extracts were combined, washed with brine (2 × 5 mL), dried
(MgSO4), and concentrated under reduced pressure. The product N-(R-
bromopropionyl)-2-oxazolidinone (a diastereomeric mixture of 1 and
2) was purified by chromatography on silica gel using hexane/ethyl
acetate (4:1) as the eluent, yield 1.59 g (82%).
selective products irrespective of the starting diastereomer used,
16 or 19. Conversely, selectivities starting with the R,R substrate
16 are almost always higher since at least a portion of the radical
generation occurs through a substrate precomplexed with the
Lewis acid.
An explanation for the observed differences in selectivity with
changes in Lewis acid also involves differing abilities to access
the chelated complex 21. At -78 °C ytterbium triflate provides
minimal selectivity (Table 3, entries 1, 2, 7, and 8), indicating
that much of the radical is trapped from a nonchelated form.
While the selectivity for the chelate-derived S-product increases
as the trap speed decreases, it never exceeds 5:1 even with the
slowest allylstannane trap. The failure of ytterbium triflate to
form the chelate easily at low temperatures results from the
presence of THF solvent,20 which is required for solubility
reasons. The starting bromide (16 or 19) is a poorer Lewis base
than THF (the electronegativity of the bromine weakens the
carbonyl basicity).21 As a result, only a small population of
bromide substrate is initially coordinated to ytterbium triflate,
and most of the radical is formed without coordinated Lewis
acid. At low-temperature dynamic equilibration is too slow to
enable the ligand exchange and complex reorganization required
to form a chelate 21. In contrast, magnesium bromide in the
nondonor solvent CH2Cl2 has no choice but to complex the
bromide; even if some of the radical is formed in a nonchelated
form, the rate of equilibration from a monodentate complex to
the bidentate chelate is able to compete with all but the most
reactive traps. In the case of ytterbium triflate, as the temperature
increases, the rate of equilibration from the free radical 23 to
the chelated complex 21 increases rapidly (more so than does
the rate of trapping), and trapping of the chelated radical
intermediate provides high selectivity.
The reason that ytterbium triflate gives high selectivity in
the addition-trapping experiment of acrylate 25 even at low
temperature is because the radical is generated in the chelated
form 21. Addition predominantly occurs to the preactivated
acrylate-ytterbium triflate chelate. Radicals that are monoco-
ordinated to ytterbium triflate, or not coordinated at all, are
unlikely, and so no complex equilibration is required.
In conclusion, the application of Lewis acids in radical
reactions to control rotamer populations is a more complex
situation than previously hypothesized. Several factors need to
be carefully examined including the interactions of the Lewis
acid with the substrate (or solvent), the reactive conformation
required to obtain high selectivity and the variations in how to
access this conformation, the time scale of the reaction, and
finally the temperature dependence of all the above. Realization
of these intricacies has provided access to functionally allylated
products with excellent diastereoselectivity (>50:1) and yield
(>90%) even at room temperature.
1: Rf ) 0.3 (80:20 hexane/ethyl acetate); mp 120-123 °C; 1H NMR
(400 MHz, CDCl3) δ 7.36-7.09 (m, 10 H), 5.61 (q, J ) 6.7 Hz, 1 H),
5.32 (q, J ) 5.4 Hz, 1 H), 4.75 (d, J ) 4.8 Hz, 1 H), 4.47 (d, J ) 5.6
Hz, 2 H), 1.80 (d, J ) 6.7 Hz, 3 H); 13C NMR (400 MHz, CDCl3) δ
169.4, 152.5, 139.5, 137.8, 129.6, 129.0, 128.9, 128.4, 128.0, 127.3,
64.9, 56.5, 50.4, 38.2, 20.8; [R]26 ) -152.2° (c 1, CH2Cl2). Anal.
D
Calcd for C19H18BrNO3: C, 58.78; H, 4.67, N, 3.61. Found: C, 58.55;
H, 4.42; N, 3.60.
2: Rf ) 0.6 (80:20 hexane/ethyl acetate); mp: 162-165 °C; 1H NMR
(400 MHz, CDCl3) δ 7.38-7.09 (m, 10 H), 5.55 (q, J ) 6.7 Hz, 1 H),
5.36-5.30 (m, 1 H), 4.65 (d, J ) 6.7 Hz, 1 H), 4.47 (dd, J ) 7.8, 9.6
Hz, 1 H), 4.39 (dd, J ) 2.4, 9.4 Hz, 1 H), 1.64 (d, J ) 6.7 Hz, 3 H);
13C NMR (400 MHz, CDCl3) δ 169.3, 152.7, 139.2, 137.8, 129.2, 129.2,
128.8, 128.5, 128.2, 127.3, 65.8, 57.1, 51.5, 38.8, 21.1; [R]26D ) -88.9°
(c 1, CH2Cl2). Anal. Calcd for C19H18BrNO3: C, 58.78, H, 4.67, N,
3.61. Found: C, 58.60; H, 4.33; N, 3.53.
Lewis Acid-Mediated Radical Allylations of 1. To a flask
containing 1 (78 mg, 0.2 mmol), MgBr2 (103 mg, 0.4 mmol), and CH2-
Cl2 (4 mL) under N2 were added (2-methylallyl)tributyl tin (173 mg,
0.5 mmol) and Et3B (1M in hexane) (0.4 mL, 0.4 mmol) at -78 °C.
Two milliliters of O2 was then added via syringe at once. The reaction
mixture was stirred at -78 °C for 2 h. After completion (TLC), Et2O
(20 mL) was added to the reaction mixture. It was then washed with
brine (3 × 3 mL) and dried with MgSO4. The product 8 was purified
by chromatography on silica gel using hexane/ethyl acetate (9:1) as
the eluent, yield 57 mg (78%)
8: Rf ) 0.6 (80:20 hexane/ethyl acetate); mp 98-100 °C; 1H NMR
(400 MHz, CDCl3) δ 7.08-7.33 (m, 10 H), 5.6 (m, 1 H), 5.29 (m, 1
H), 5.01 (m, 2 H), 4.66 (d, J ) 6.6 Hz, 1 H), 4.35 (d, J ) 5.1 Hz, 2
H), 3.62 (m, 1 H), 2.28 (m, 1 H), 1.89 (m, 1 H), 1.05 (d, J ) 6.6 Hz,
3 H); 13C NMR (270 MHz, CDCl3) δ 176.1, 153.0, 139.5, 138.1, 135.5,
129.2, 128.9, 128.6, 128.4, 128.3, 127.8, 127.0, 117.0, 64.9, 56.5, 51.2,
37.3, 16.0; [R]D26 ) -111.9° (c 0.35, CH2Cl2). Anal. Calcd for C22H23-
NO3: C, 75.62; H, 6.63; N, 4.01. Found: C, 75.34, H, 6.82, N, 4.24.
Hydrolysis of 8. Typical Procedure. To a flask containing 8 (174.5
mg, 0.5 mmol) was added THF (5 mL), H2O (5 mL), and H2O2 (30%)
(0.226 mL, 2 mmol) at 0 °C under N2. LiOH‚H2O (41 mg, 1 mmol)
was added to the reaction mixture; it was stirred at 0 °C for 1 h. After
completion (TLC), most of the THF was evaporated. The aqueous
solution (pH 12) was extracted with CH2Cl2 (3 × 10 mL) (recovery of
chiral auxiliary). Finally, The aqueous solution was acidified with HCl
(3M) to pH ∼1 and reextracted with CH2Cl2 (4 × 15 mL). The organic
solution was dried (MgSO4) and concentrated to yield (S)-2-methyl-
4-butenoic acid (45 mg, 88%). [R]26 ) 10.5° (c 1.15, CHCl3) [lit.
D
[R]26 ) 10.5° (CHCl3). Riley, R. G.; Silverstein, R. M. Tetrahedron
D
1
1974, 30, 1171]; H NMR (400 MHz, CDCl3) δ 5.85-5.72 (m, 1 H),
Experimental Section
5.18-5.05 (m, 2 H), 2.62-2.56 (m, 1 H), 2.48-2.42 (m, 1 H), 2.26-
2.18 (m, 1 H), 1.20 (d, J ) 7.0 Hz, 3 H).
For general experimental see ref 16d.
Products from Allylation of 1 with 10 and 11. 12: Rf ) 0.6 (80:
20 hexane/ethyl acetate); mp 66-68 °C; 1H NMR (400 MHz, CDCl3)
δ 7.34-7.11 (m, 10 H), 5.32 (q, J ) 5.7 Hz, 1 H), 4.79 (s, 1 H), 4.75-
4.68 (m, 2 H), 4.39 (d, J ) 5.4 Hz, 2 H), 3.89-3.80 (m, 1 H), {[2.40
(dd, J ) 7.4, 13.5 Hz)], [2.30 (dd, J ) 7.4, 13.5 Hz)], 1 H}, {[1.99
(dd, J ) 7.0, 14.2 Hz)], [1.81 (dd, J ) 7.0, 14.2 Hz)], 1 H}, {[1.72
(s)], [1.68 (s)], 3 H}, {[1.05 (d, J ) 6.8 Hz)], [0.94 (d, J ) 6.8 Hz)],
3 H}; 13C NMR (400 MHz, CDCl3) δ 176.7, 153.2, 142.9, 139.7, 138.1,
129.4, 128.9, 128.7, 128.5, 127.9, 127.1, 112.9, 68.1, 64.9, 56.6, 51.1,
Preparation of N-(r-bromopropionyl)-2-oxazolidinones 1 and 2.
Typical Procedure. To a flask containing 4-(diphenylmethyl)-2-
oxazolidinone (1.265 g, 5 mmol) and THF (20 mL) under N2 was added
n-BuLi (2.5 M) (2.0 mL, 5 mmol) at -78 °C dropwise. After complete
(20) For the use of THF (or alkyl ethers) coordinated lanthanides in
synthesis, see: Aspinall, H. C.; Browning, A. F.; Greeves, N.; Ravenscroft,
P. Tetrahedron Lett. 1994, 35, 4639; Aspinall, H. C.; Dwyer, J. L. M.;
Greeves, N.; McIver, E. G.; Woolley, J. C. Organometallics 1998, 17, 1884.
Aspinall, H. C.; Greeves, N.; Lee, W.-M.; McIver, E. G.; Smith, P. M.
Tetrahedron Lett. 1997, 38, 4679.
(21) (a) Wu, J. H.; Zhang, G.; Porter, N. A. Tetrahedron Lett. 1997, 38,
2067. (b) Hunt, I. R.; Rogers, C.; Woo, S.; Rauk, A.; Keay, B. A. J. Am.
Chem. Soc. 1995, 117, 1049.
41.6, 35.6, 25.7, 22.0, 16.2; [R]26 ) -130.9° (c 1, CH2Cl2). Anal.
D
Calcd for C23H25NO3: C, 76.01; H, 6.93; N, 3.85. Found: C, 75.66;
H, 6.64; N, 4.09.
13: Rf ) 0.3 (80:20 hexane/ethyl acetate); oil; 1H NMR (400 MHz,

Higher SelectiVity at Higher Temperatures
J. Am. Chem. Soc., Vol. 122, No. 37, 2000 8879
CDCl3) δ 7.36-7.09 (m, 10 H), {[6.21 (d, J ) 1 Hz)], [6.17 (d, J )
1 Hz)], 1 H}, {[5.56 (d, J ) 1 Hz)], [5.54 (d, J ) 1 Hz)], 1 H}, 5.31
(m, 1 H), {[4.75 (d, J ) 4.8 Hz)], [4.69 (d, J ) 4.8 Hz)], 1 H}, 4.48-
4.36 (m, 2 H), 3.99-3.80 (m, 1 H), {[3.78 (s)], [3.72 (s)], 3 H}, 2.63
(dd, J ) 7.2, 13.8 Hz), {[2.39 (dd, J ) 7.2, 13.8 Hz)], [2.27 (dd, J )
7.2, 13.8 Hz)], 1 H}, {[1.10 (d, J ) 6.7 Hz)], [0.96 (d, J ) 6.7 Hz)],
3 H}; 13C NMR (400 MHz, CDCl3) δ 176.2, 167.3, 153.1, 139.7, 138.0,
137.8, 129.4, 129.0, 128.8, 128.5, 127.9, 127.4, 127.1, 64.9, 56.6, 52.1,
33.8, 22.2, 20.6, 14.2; [R]26 ) -183.3° (c 1, CH2Cl2). Anal. Calcd
for C25H29NO3: C, 76.70, H, 7.47, N, 3.58. Found: C, 76.67; H, 7.52;
N, 3.65.
D
Competition Experiment for Determination of Relative Rates for
Allylstannane Trapping. Typical Procedure. The same procedure was
used as in the preparation of 8 except that in addition to 2 equivs of
allylstannane, 2 equivs of either 10 or 11 were added and allowed to
compete for the 1 equiv of substrate 1 or 2. Each combination of
substrate, allylstannane and Lewis acid was tested and crude product
50.9, 36.8, 35.0, 17.2; [R]26 ) -104.8° (c 1, CH2Cl2). Anal. Calcd
D
for C24H25NO5: C, 70.75, H, 6.18, N, 3.44. Found: C, 70.43; H, 6.27;
N, 3.55.
1
ratios were determined by H NMR (400 MHz).
15: Rf ) 0.7 (80:20 hexane/ethyl acetate); mp 104-105 °C; 1H NMR
(400 MHz, CDCl3) δ 7.35-7.09 (m, 10 H), 5.35-5.30 (m, 1 H), 4.80
(s, 1 H), 4.77-4.72 (m, 3 H), 4.43-4.36 (m, 2 H), 4.01-3.94 (m, 1
H), 2.32 (dd, J ) 7.2, 13.7 Hz, 1 H), {[2.07 (dd, J ) 7.3, 13.8 Hz)],
[1.94 (dd, J ) 7.3, 13.8 Hz)], 1 H}, {[1.74 (s)],[1.70 (s)], 3 H}, 1.61-
1.52 (m, 1 H), 1.44-1.21 (m, 3 H), 0.86 (t, J ) 7.2 Hz, 3 H); 13C
NMR (400 MHz, CDCl3) δ 176.2, 153.2, 143.2, 139.8, 138.1, 129.5,
128.9, 128.8, 128.5, 127.9, 127.1, 112.8, 64.7, 56.7, 50.9, 40.7, 40.4,
Acknowledgment. This work was supported by the National
Institutes of Health (GM-54656). We thank Dr. Jianguo Ji for
initial experimental assistance and Prof. Craig Jasperse for
helpful discussions. Tara Rheault thanks NSF for a graduate
fellowship.
JA001456J