Free radical-mediated intermolecular conjugate additions. Effect of the lewis acid, chiral auxiliary, and additives on diastereoselectivity

Article abstract of DOI:10.1021/ja991205e

We have developed a highly diastereoselective method for the conjugate addition of carbon radicals to chiral α,β-unsaturated N-enoyloxazolidinones using Bu₃SnH as chain carrier and Et₃B/O₂ as radical initiator. Lewis acids

Full text of DOI:10.1021/ja991205e

J. Am. Chem. Soc. 1999, 121, 7517-7526
7517
Free Radical-Mediated Intermolecular Conjugate Additions. Effect of
the Lewis Acid, Chiral Auxiliary, and Additives on
Diastereoselectivity
Mukund P. Sibi,* Jianguo Ji, Justin B. Sausker, and Craig P. Jasperse
Contribution from the Department of Chemistry, North Dakota State UniVersity,
Fargo, North Dakota 58105-5516
ReceiVed April 15, 1999
Abstract: We have developed a highly diastereoselective method for the conjugate addition of carbon radicals
to chiral R,â-unsaturated N-enoyloxazolidinones using Bu3SnH as chain carrier and Et3B/O2 as radical initiator.
Lewis acids have been screened, and Yb(OTf)3 proved to give optimized results for both chemical yield (88%
for 1a and 94% for 1b) and diastereoselectivity (25:1 for 1a and 46:1 for 1b). The selectivity is solvent-
dependent, CH2Cl2-THF being an ideal combination. Scrupulously dry solvents or reaction conditions were
not required. Substoichiometric amounts of Yb(OTf)3 provided efficient reaction with minimal sacrifice in
diastereoselectivity. Carbon radicals with reasonable nucleophilicity were generally successful, including
functionalized radicals such as acetyl or methoxymethyl. Electrophilic radicals were not successful. A model
which accounts for most of our observations is presented.
Formation of a carbon-carbon bond by addition to an R,â-
unsaturated system is one of the premiere reactions in synthetic
organic chemistry.1 In the majority of these reactions, the carbon
nucleophile is an ionic species and most often an organocopper
reagent. A large number of chiral auxiliaries and chiral ligands
have been described which provide good to excellent diaste-
reoselectivity and enantioselectivity, respectively, in anionic
conjugate additions to acyclic R,â-unsaturated systems.2,3
Intermolecular conjugate additions of free radicals to enones
and enoates have been reported in the literature.4 However,
stereoselective addition to nonterminal alkenes has met with
limited success. Only in the last several years have successful
examples of diastereoselective radical conjugate additions been
reported. In one notable example, Curran used a complex
auxiliary derived from Kemp’s triacid to obtain excellent levels
of diastereoselectivity.5 Other examples include methyl radical
addition to an enoate with a sugar-derived template,6 alkyl
radicals to R-ketosulfoxides,7 and ketyl radicals addition to
enoates with high selectivity have also been reported.8 However,
a general solution to the problem of acyclic diastereoselection9
in â-radical additions remained elusive until 1995 when we
demonstrated highly diastereoselective conjugate reactions using
simple and readily available oxazolidinones as chiral auxilia-
ries.10 After this initial study, several other examples of highly
diastereoselective conjugate additions11 as well as their enan-
tioselective variants have been reported.12
When we initiated this study, oxazolidinone auxiliaries had
found limited application in radical reactions. Crich had used
an ephedrine-derived oxazolidinone under non-Lewis acid
conditions with limited selectivity in an alkylation/trapping
reaction.13 We felt that the low selectivity was due to a lack of
appropriate rotamer control with N-acyloxazolidinones. Several
rotamers are available for free N-acyloxazolidinones (Figure 1).
High stereoselectivity requires a dominant reactive rotamer in
which one face is effectively blocked, but in ground state
conformer, C the R group is too far away for effective face
shielding. We hypothesized that a chelating Lewis acid additive
could enforce predominant reaction via rotamer A so that with
an appropriate R group in the auxiliary facial shielding in the
â-addition of radicals could be possible. The increasing ap-
plication of Lewis acids in radical reactions14 and the excellent
diastereofacial control in Lewis acid-mediated Diels-Alder
(9) For a discussion of acyclic diastereoselection in radical reactions,
see: (a) Curran, D. P.; Porter, N. A.; Giese, B. Stereochemistry of Radical
Reactions; VCH: Weinheim, Germany, 1995. (b) Porter, N. A.; Giese, B.;
Curran, D. P. Acc. Chem. Res. 1991, 24, 296. (c) Smadja, W. Synlett 1994,
1.
(10) For a preliminary communication, see: Sibi, M. P.; Jasperse, C.
P.; Ji, J. J. Am. Chem. Soc. 1995, 117, 10779.
(1) For an excellent monograph, see: Perlmutter, P. Conjugate Addition
Reactions in Organic Synthesis; Pergamon: Oxford, U.K., 1992. For a recent
review, see: Leonard, J.; Diez-Barra, E.; Merino, S. Eur. J. Org. Chem.
1998, 2051.
(2) For a recent review on conjugate additions using copper reagents,
see: Krause, N.; Gerold, A. Angew. Chem., Int. Ed. Engl. 1997, 36, 184.
(3) Rossiter, B. E.; Swingle, N. M. Chem. ReV. 1992, 92, 771 and
references therein.
(4) Nozaki, K.; Oshima, K.; Utimoto, K. Bull. Chem. Soc. Jpn. 1991,
64, 403.
(5) Stack, J. G.; Curran, D. P.; Geib, S. V.; Rebek, J., Jr.; Ballester, P.
J. Am. Chem. Soc. 1992, 114, 7007.
(11) (a) Sibi, M. P.; Ji, J. J. Org. Chem. 1996, 61, 6090. (b) Sibi, M. P.;
Ji, J. Angew. Chem., Int. Ed. Engl. 1997, 35, 2274. (c) Toru, T.; Watanabe,
M.; Mase, N.; Tsusaka, Y.; Hayakawa, T.; Ueno, Y. Pure Appl. Chem.
1996, 68, 711. (d) Mase, N.; Watanabe, Y.; Ueno, Y.; Toru, T. J. Org.
Chem. 1997, 62, 7794. (e) Nishida, M.; Ueyama, E.; Hayashi, H.; Ohtake,
Y.; Yamaura, Y.; Yanaginuma, E.; Yonemitsu, O.; Nishida, A.; Kawahara,
N. J. Am. Chem. Soc. 1994, 116, 6455. (f) Nishida, M.; Hayashi, H.;
Yamaura, Y.; Yanaginuma, E.; Yonemitsu, O.; Nishida, A.; Kawahara, N.
Tetrahedron Lett. 1995, 36, 269. (g) Badone, D.; Bernassau, J. M.;
Cardamone, R.; Guzzi, U. Angew. Chem., Int. Ed. Engl. 1996, 35, 535. (h)
Piber, M.; Leahy, J. W. Tetrahedron Lett. 1998, 39, 2043. (i) Merlic, C.
A.; Walsh, J. C. Tetrahedron Lett. 1998, 39, 2083.
(6) Ru¨ck, K.; Kunz, H. Synthesis 1993, 1018.
(7) Toru, T.; Watanabe, Y.; Tsusaka, M.; Ueno, Y. J. Am. Chem. Soc.
1993, 115, 10464.
(8) For seminal contributions, see: Molander G.; Harris, C. R. Chem.
ReV. 1996, 96, 307. Kawatsura, M.; Deckura, F.; Shirahama, H.; Matsuda,
F. Synlett 1996, 373. Mikami, K.; Yamaoka, M. Tetrahedron Lett. 1998,
39, 4501.
(12) For a recent account, see: Sibi, M. P.; Porter, N. A. Acc. Chem.
Res. 1999, 32, 163. For examples of enantioselective free radical conjugate
additions, see: (a) Wu, J. H.; Zhang, G.; Porter, N. A. Tetrahedron Lett.
1997, 38, 2067. (b) Urabe, H.; Yamashita, K.; Suzuki, K.; Kobayashi, K.;
Sato, F. J. Org. Chem. 1995, 60, 3576.
(13) Crich, D.; Davies, J. W. Tetrahedron Lett. 1987, 28, 4205.
10.1021/ja991205e CCC: $18.00 © 1999 American Chemical Society
Published on Web 08/05/1999

7518 J. Am. Chem. Soc., Vol. 121, No. 33, 1999
Sibi et al.
Table 1. Effect of Lewis Acids and Solvents on Diastereoselective
Conjugate â-Radical Additiona
entry substrate Lewis acid
solvent
product yield (%)b ratioc
1
2
3
4
5
6
7
8
9
(R)-1a none
CH2Cl2
2a
30d,e
93f
94f
88f
90f
82f
93f
90
1.3:1
25:1
45:1
24:1
18:1
9:1
7:1
7:1
6:1
6:1
(R)-1a Yb(OTf)3
(R)-1b Yb(OTf)3
(R)-1a Y(OTf)3
(R)-1a Sm(OTf)3
(R)-1a Sc(OTf)3
(R)-1a La(OTf)3
(R)-1a ZrCl4
CH2Cl2/THF (4:1) 2a
CH2Cl2/THF (4:1) 2b
CH2Cl2/THF (4:1) 2a
CH2Cl2/THF (4:1) 2a
CH2Cl2/THF (4:1) 2a
CH2Cl2/THF (4:1) 2a
CH2Cl2/THF (4:1) 2b
(R)-1a MgBr2‚OEt2 CH2Cl2/THF (4:1) 2a
90
90
80
10 (R)-1a MgBr2‚OEt2 CH2Cl2/ether
11 (R)-1a MgI2
12 (R)-1a ZnCl2 (2)
13 (R)-1a ZnCl2 (2)
14 (R)-1a ZnCl2 (2)
2a
CH2Cl2/Et2O (4:1) 2a
CH2Cl2/Et2O (4:1) 2a
CH2Cl2
6:1
6:1
1.3:1
1.3:1
90
2a
30d,g
20d
Figure 1.
CH2Cl2/THF (4:1) 2a
15 (R)-1a Zn(OTf)2 (2) CH2Cl2/Et2O (4:1) 2a
50 (20) 3:1
20 (80) 5:1
70 (20) 4:1
30 (60) 3:1
reactions of N-enoyloxazolidinones15 (involving a type A
rotamer) supported our approach. We also anticipated that Lewis
acids should greatly enhance the electrophilicity of the â-car-
bons, obviating the problems normally associated with the slow
addition of radicals to nonterminal alkenes.16 We also recognized
that, if the reaction proceeds exclusively via rotamer A, the
distance between the reactive â-center and the chiral carbon is
still significant; thus we assumed that the R groups familiar
from Evans chemistry (isopropyl, benzyl, phenyl) might prove
too small and that a larger substituent might be necessary for
high selectivity.
In this paper we describe a comprehensive investigation on
diastereoselective free-radical additions to enoyl oxazolidinones.
The use of a novel oxazolidinone auxiliary was found to be
superior to the traditional Evans auxiliaries. High selectivity
required the use of a two-point binding Lewis acid. We detail
the effect of Lewis acid, solvent, and radical precursors on the
facility of the reaction as well as the levels of diastereoselec-
tivity. A model which accounts for most of our observations is
presented.
16 (R)-1a MeAlCl2
17 (R)-1a Et2AlCl
18 (R)-1a SnCl4 (2)
19 (R)-1a Yb(OTf)3
20 (R)-1a Bu2SnCl2
21 (R)-1a BF3.Et2O
22 (R)-1a TiCl4
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
2a
2a
2a
2a
2a
2a
2a
2a
2b
50d,g
90d
1.3:1
1.3:1
1.3:1
-
80 (5)
-(90)
-(70)
90
80
70
23 (R)-1a Bu2BOTf
-
24 (R)-1b MgBr2‚OEt2 CH2Cl2/ether
25 (R)-1b La(OTf)3
26 (R)-1b ZnCl2 (2)
20:1
12:1
9:1
CH2Cl2/THF (4:1) 2b
CH2Cl2/Et2O (4:1) 2b
a Two equivalents of the Lewis acid, 10 equiv of Pr-I, 5 equiv of
Bu3SnH, and 10 equiv of Et3B were used at -78 °C. b Yields were
determined by NMR integration, except when purified yields are
indicated. Yields in parentheses are for the alkene reduction product.
c Diastereomer ratios were determined by 1H NMR (400 MHz).
d Starting material accounted for most of the remaining mass balance.
e Sixty percent of the starting material was recovered. f Purified yield.
g The Lewis acid was insoluble.
i
reaction, and the screening reactions were worked up after 3 h.
The solvent varied depending on Lewis acid. Dichloromethane
was used for soluble Lewis acids, but THF or ether was often
included to solubilize Lewis acids that were otherwise insoluble.
During the evaluation of the Lewis acids, we used large excesses
of isopropyl iodide, tributyl tin hydride, and triethylborane in
order to give maximum opportunity for reaction. These quanti-
ties in eq 1 do not reflect those necessary under optimized
conditions.
Results
Effect of Lewis Acids and Solvent on Diastereoselectivity.
We began our study by examining radical additions to the
crotonate and cinnamate derived from 4-(diphenylmethyl)-2-
oxazolidinone, a new chiral auxiliary introduced by our group.17
The required starting materials 1a and 1b were prepared by
standard acylation protocols. We originally screened a variety
of different Lewis acids for their ability to mediate diastereo-
selective and high-yielding addition of isopropyl radical (eq 1,
Table 1). Reactions were conducted at -78 °C, using Bu3SnH
as the radical chain carrier and triethylborane/oxygen as the
radical initiator.18 In the standard Lewis acid screening experi-
ment, two equivalents of Lewis acid was used relative to the
substrate. As discussed later, substoichiometric quantities of
Lewis acid also proved quite satisfactory under optimized
conditions. The solvent, substrate, and Lewis acid were always
premixed at room temperature; the solution was then cooled,
the isopropyl iodide and Bu3SnH were added at -78 °C,19 and
the Et3B was then rapidly added last to initiate the reaction.
Oxygen was added at 30 min intervals to reinitiate radical
(14) For an excellent recent review, see: Renaud, P.; Gerster, M. Angew.
Chem., Int. Ed. Engl. 1998, 37, 2562.
(15) Evans, D. A.; Chapman, K. T.; Bisaha, J. J. Am. Chem. Soc. 1988,
110, 1238.
(16) Giese, B. Angew. Chem., Int. Ed. Engl. 1983, 22, 753.
(17) This auxiliary is now commercially available from Aldrich chemical
company. For its synthesis, see: Sibi, M. P.; Deshpande, P. K.; LaLoggia,
A. J.; Christensen, J. W. Tetrahedron Lett. 1995, 36, 8961.
(18) Miura, K.; Ichinose, Y.; Nozaki, K.; Fugami, K.; Oshima, K.;
Utimoto, K. Bull. Soc. Chem. Jpn. 1989, 62, 143.
In the control reaction in which no Lewis acid additive was
present, a nonselective 1.3:1 diastereomeric mixture of 2a
formed in low yield (entry 1).20 The reaction was extremely
sluggish, as is typical for radical additions to nonterminal
alkenes.16 Most of the remaining mass was recovered starting
material. The yield without added Lewis acid could be boosted
to 60%, but only by using large excesses of isopropyl iodide

Free Radical-Mediated Intermolecular Additions
J. Am. Chem. Soc., Vol. 121, No. 33, 1999 7519
Table 2. Effect of Rare Earth Lewis Acid on Diastereoselective
Conjugate â-Radical Addition
and tributyltin hydride, using multiple additions of initiator, and
conducting the reaction for 12 h. It is worth noting that, even
in the absence of added Lewis acid, Bu3SnI forms as a byproduct
and may function as a weak but noncoordinating Lewis acid.21
(R)-1a f 2a
yield (%)b drc
(R)-1b f 2b
yield (%)b drc
entry
Lewis acida
The best Lewis acid proved to be Yb(OTf)3, which gave
essentially quantitative yields and diastereoselectivities of 25:1
(crotonate 1a, entry 2) and 45:1 (cinnamate 1b, entry 3).22 We
emphasize that these diastereoselectiVities are comparable to
or better than those aVailable using ionic reaction conditions!23
Other Lewis acids generally gave good conversion, confirming
the importance of alkene activation. Starting material was
recovered only when the Lewis acid was insoluble and thus
ineffective (entries 13, 19). Ether or THF was often added to
the dichloromethane solvent mixture in order to solubilize the
Lewis acid. There are many reactions in which Lewis acid
solubility is not a necessity, but since radicals are relatively
short-lived reactants, it is probable that Lewis acid solubility is
a necessity in radical reactions.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Sc(OTf)3
Y(OTf)3
82
88
93
82
90
85
93
88
93
96
90
96
93
90
88
90
8.6:1
24:1
3.5:1
6.4:1
7:1
90
88
87
85
88
90
93
85
96
95
95
90
87
87
94
90
8.0:1
37:1
5:1
La(OTf)3
Ce(OTf)3
Pr(OTf)3
Nd(OTf)3
Sm(OTf)3
Eu(OTf)3
Gd(OTf)3
Tb(OTf)3
Dy(OTf)3
Ho(OTf)3
Er(OTf)3
Tm(OTf)3
Yb(OTf)3
Lu(OTf)3
7.5:1
11:1
18:1
32:1
36:1
32:1
46:1
17:1
29:1
36:1
41:1
46:1
50:1
7:1
10:1
20:1
18:1
23:1
23:1
23:1
25:1
24:1
25:1
20:1
i
a One equivalent of the Lewis acid, 5 equiv of Pr-I, 2 equiv of
Bu3SnH, and 2 equiv of Et3B were used at -78 °C. b Yields are for
isolated and column-purified materials. c Diastereomer ratios were
When very strong Lewis acids were used (entries 15-18,
21-23), alkene reduction (H-H addition instead of R-H
addition) was observed. In a control reaction using EtAlCl2,
alkene reduction was also observed even when radical initiator
(Et3B/O2) was omitted. We surmise that, when the Lewis acid
is relatively strong, the crotonate is sufficiently activated so that
tin hydride may serve as an ionic hydride source.24 Optimal
results thus require that the Lewis acid be soluble and strong
enough to activate radical addition but not strong enough to
activate direct reaction with tin hydride.25
1
determined by H NMR (400 MHz).
In terms of diastereoselectivity, Table 1 shows that in general
lanthanide and pre-lanthanide triflates gave the best results
(entries 2-7). Lewis acids incapable of simultaneously binding
both substrate carbonyls gave low selectivity. For example, BF3‚
Et2O gave high chemical reactivity but essentially no higher
selectivity than was observed in the Lewis acid-free control
reaction (entry 21). Chelating Lewis acids capable of binding
both substrate carbonyls consistently showed improved diaste-
reoselectivity relative to the control but surprisingly gave a wide
range of selectivities; with the crotonate 1a, for example, Yb-
(OTf)3 gave 25:1 selectivity whereas MgBr2 and ZnCl2 gave
only 6:1 ratios (entries 2, 9-12).
Since lanthanide triflates appeared to give the best results
we hoped to gain a better appreciation of what factors influence
the magnitude of selectivity. We have conducted a systematic
study of the dependence of selectivity on lanthanide and
prelanthanide triflates as Lewis acid (Table 2). Yields were
consistently high. The general pattern observed is that the late
lanthanide triflates gave higher selectivity than the earlier
lanthanides. Within the Sc-Y-La period, yttrium triflate gave
much higher selectivity than either the scandium or lanthanum
analogue and came close to matching the selectivity obtained
with ytterbium. Moving from the early to late lanthanides both
reduces the ionic radii (“the lanthanide contraction”) and also
increases the Lewis acidity of the ions. The ionic radius is
probably the more critical influence on selectivity in the present
reaction. If selectivity simply increased with increasing Lewis
acidity, then Sc(OTf)3 should have given higher selectivity than
Y(OTf)3, contrary to observation. Ytterbium and yttrium ions
have similar ionic radii (∼0.9 Å),26 so their similar selectivity
is in agreement with the dependence of stereoselectivity on ionic
radii. Lewis acids such as MgBr2, ZnCl2, ZrCl4, and Sc(OTf)3
are probably too small; early lanthanides such as La(OTf)3 and
Ce(OTf)3 are probably too big.
A trace side product formed in variable yields was ethyl
addition compound 2c. The use of Et3B/O2 as radical initiator
generates ethyl radical, which can add to the substrate. The
amount of ethyl addition was 0.5-2.0% when Yb(OTf)3 was
used as Lewis acid. Although the quantities were too small to
quantify carefully by NMR, we have qualitatively observed that
ethyl addition is minimized when reactivity is maximum. When
weaker Lewis acids (ZnCl2), less reactive substrates, or less
reactive radicals were used (vide infra) the amount of ethyl
transfer product was sometimes as high as 5-10%. This
observation is consistent with the expectation that short radical
chains require more initiation events and provide more op-
portunities for ethyl addition. In the case of Yb(OTf)3-mediated
addition of isopropyl radical to the crotonate 1a, it appears that
the radical chain has an average chain length of between 50
and 200.
(19) Bu3SnH often reacted at room temperature with substrate-Lewis
acid complexes.
(20) No regioisomeric product, that is, the product arising from addition
to the R-carbon, was detected in the reaction. Belokon et al. report the
formation of R-addition products in their work on electrophilic radical
addition to oxazolidinone cinnamates in the absence of Lewis acids:
Tararov, V. I.; Kuzentanov, N. Yu.; Bakhmutov, V. I.; Ikonnikov, N. S.;
Bubnov, Y. N.; Khrustalev, V. N.; Saveleva, T. F.; Belokon, Y. N. J. Chem.
Soc., Perkin Trans 1 1997, 3101.
(21) The Bu3SnI generated in situ in the reaction may conceivably serve
as a Lewis acid. See: Sibi, M. P.; Ji, J. J. Am. Chem. Soc. 1996, 118,
3063; Porter, N. A.; Wu, J. H.; Zhang, G.; Reed, A. D. J. Org. Chem.
1997, 62, 6702.
(22) Yb(OTf)3 was purchased from Aldrich Chemical company and
contained ∼0.5-1% water by mass. The use of anhydrous Yb(OTf)3 did
not lead to improvement in selectivity.
(23) For conjugate addition to chiral N-enoyl oxazolidinones under ionic
conditions using copper nucleophiles, see: Nicola´s, E.; Russell, K. C.;
Hruby, V. J. J. Org. Chem. 1993, 58, 766. For a comparative study of chiral
auxiliaries, see: Andersson, P. G.; Schink, H. E.; O¨ sterlund, K. J. Org.
Chem. 1998, 63, 8067.
The dependence of the reaction on solvent is shown in Table
3. As discussed previously, dichloromethane was a poor solvent
because of its inability to dissolve Yb(OTf)3 (entry 1). The
results and selectivity in dichloromethane were fairly similar
to reaction run in the absence of any Lewis acid additive. Use
(24) Nozaki, K.; Oshima, K.; Utimoto, K. Bull. Chem. Soc. Jpn. 1991,
64, 2585.
(25) We have observed that adding THF as a cosolvent minimized the
problem of direct reduction with several Lewis acids.
(26) Cotton, F. A.; Wilkinson, G. AdVanced Inorganic Chemistry, 5th
5d.; Wiley: New York, 1988; p 955.

7520 J. Am. Chem. Soc., Vol. 121, No. 33, 1999
Sibi et al.
Table 3. Effect of Solvent and Yb(OTf)3 Stoichiometry on
Table 4. Effect of Additives on Diastereoselectivity in Isopropyl
Radical Addition to 1b Using 0.1 Equivalent Yb(OTf)3
Selectivitya
a
Lewis acid
(equiv)
entry
additive (equiv)
yield (%)b
drc
entry
solventb
CH2Cl2
CH2Cl2/Et2O (4:1)
toluene/THF (4:1)
THF
yield (%)c
50 (40)e
90
90
90
drd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
none
82
90
92
89
82
81
76
61
92
92
91
92
89
90
88
90
75
95
86
60
10:1
8:1
10:1
8:1
3:1
10:1
4:1
1.4:1
22:1
23:1
21:1
17:1
14:1
10:1
4:1
1
2
3
4
5
Yb(OTf)3 (2)
Yb(OTf)3 (2)
Yb(OTf)3 (2)
Yb(OTf)3 (2)
Yb(OTf)3 (2)
1.3:1
9:1
11:1
15:1
20:1
HMPA (0.1)
DMSO (0.1)
H3COCH2CH2OCH3 (0.1)
12-Crown-4 (0.1)
15-Crown-5 (0.1)
18-Crown-6 (0.1)
(H3C)2NCH2CH2N(CH3)2 (0.1)
Methanol (0.2)
HOCH2CH2OH (0.1)
2,3-Butanediol (0.1)
Pinacol (0.1)
CH2Cl2/THF (4:1)
H2O (6 equiv.)
90
6
7
Yb(OTf)3 (2)
Yb(OTf)3 (2)
CH2Cl2/THFa (4:1)
CH2Cl2/THF (4:1)
H2O (30 equiv.)
CH2Cl2/THF (4:1)
CH2Cl2/THF (4:1)
93
60
25:1
1.7:1
8
9
10
11
Yb(OTf)3 (2)
Yb(OTf)3 (1)
Yb(OTf)3 (0.3) CH2Cl2/THF (4:1)
Yb(OTf)3 (0.1) CH2Cl2/THF (4:1)
93
90
90
88
25:1
25:1
20:1
16:1
Catechol (0.1)
HO(CH2CH2O)2OH (0.1)
HO(CH2CH2O)3OH (0.1)
HO(CH2CH2O)4OH (0.1)
HO(CH2CH2O)5OH (0.1)
HOCH2CH2OH (0.2)
HOCH2CH2OH (0.3)
HOCH2CH2OH (0.4)
a Five equivalents of iPr-I, 2 equiv of Bu3SnH, and 2 equiv of Et3B
were used at -78 °C. b Commercial THF and CH2Cl2 were used without
any attempt to predry or predistill them. c Yields are for isolated and
column-purified materials. d Diastereomer ratios were determined by
1H NMR (400 MHz). e The yield in parentheses is for the reduction
product.
3:1
6:1
18:1
6:1
7:1
a For typical reaction conditions (5 equiv of Pr-I, 2 equiv of
Bu3SnH, 4 equiv of Et3B, 4:1 CH2Cl2/THF, -78 °C); see experimental
section. A total of 2.5 mL was used for 0.1 mmol scale reaction. b Yields
are for isolated and column-purified materials. c Diastereomer ratios
i
of ether instead of THF (entry 2), toluene instead of CH2Cl2
(entry 3), or THF only (entry 4) all gave good yields but reduced
selectivity. Moderate quantities of water did little harm (entry
5). The ability to use Yb(OTf)3 under slightly wet conditions is
of tremendous practical impact. Many other Lewis acids are
acutely moisture-sensitive and must be handled in scrupulously
dry solvents using drybox or syringe techniques. By contrast,
we were able to routinely weigh Yb(OTf)3 in the air, and our
Yb(OTf)3 bottles showed no sign of deterioration over months
of usage despite not using glovebox procedures. We also found
that it was unnecessary to predry our solvents (entry 6); when
CH2Cl2 and THF were used “straight from the bottle” the
reaction results were not compromised!27 A large excess of
water28 did appear to deactivate the Yb(OTf)3, however (entry
7), even though Yb(OTf)3 has been found to be active in aqueous
solvent in other applications.29
We have also evaluated the use of lesser excesses of
tributyltin hydride and isopropyl iodide. With 2 equiv of
tributyltin hydride, 2 equiv of isopropyl iodide, and 1 equiv of
Yb(OTf)3, the yield of 2a was 84%. With 1.2 equiv of tributyltin
hydride, 1.2 equiv of isopropyl iodide, and 1 equiv of Yb(OTf)3,
the yield dropped slightly to 75%. The diastereoselectivities were
not affected. When tributyltin hydride was replaced by (Me3-
Si)3SiH or Ph2SiH2, very low yields of 2a resulted, combined
with complex mixtures of unreacted starting material 1a, ethyl
addition product 2c, silane addition products, and unidentified
side products.
1
were determined by H NMR (400 MHz).
potential of the reaction was obviously appealing because it
suggested to us that chiral Lewis acids could mediate enanti-
oselective conjugate addition reactions on achiral substrates; we
have already reduced this possibility to practice.31
It is well-known that the exact complexation environment of
lanthanide ions is extremely complex.32 Additives can influence
ion pairing to counterions, Lewis acidity, steric volume, and
the redox chemistry of lanthanide reagents. As background to
other investigations of enantioselective reactions involving chiral
lanthanides, we have tested the effect of several additives on
Yb(OTf)3-catalyzed radical addition to cinnamate 1b (Table 4).
Good Lewis bases such as HMPA, DMSO, and DME had little
effect relative to the reference reaction (entries 1-4), crown
ethers reduced the selectivity somewhat (entries 5-7), and
TMEDA was very harmful for both reactivity and selectivity
(entry 8). Methanol and simple diol additives, however, had a
generally favorable impact on both yield and selectivity (entries
9-13). Poly(ethylene glycol) had a negative impact on selectiv-
ity (entries 14-17), as did increasing amounts of ethylene glycol
(entries 18-20). It was qualitatively observed that there was a
correlation between selectivity and reaction speed. We believe
that those additives that reduced selectivity did so by sequester-
ing the Yb(OTf)3, such that the slower reactions and increased
production of minor isomer resulted from nonselective reaction
by “free” substrate competing with reaction of Yb(OTf)3-bound
substrate. The simple alcohols and diols may have enhanced
the selectivity by modestly enhancing the reactivity of the
substrate-Lewis acid complex for reasons we do not yet
The dependence of the selectivity on the Yb(OTf)3 stoichi-
ometry is also shown in Table 3 (entries 8-11). The use of
10% Yb(OTf)3 resulted in only a modest reduction in selectivity
and minimal increase in the time required for completion (6 h).
Successful catalysis shows that Yb(OTf)3 transfers readily from
adduct 2a to reactant substrate 1a. This is in keeping with the
typically facile ligand exchange kinetics of lanthanide Lewis
acids, and with the observations that R,â-unsaturated carbonyls
are normally more basic than saturated analogues.30 The catalytic
(30) Hunt, I. R.; Rogers, C.; Woo, S.; Rauk, A.; Keay, B. J. Am. Chem.
Soc. 1995, 117, 1049.
(31) (a) Sibi, M. P.; Ji, J.; Wu, J. H.; Gurtler, S.; Porter, N. A. J. Am.
Chem. Soc. 1996, 118, 9200. (b) Sibi, M. P.; Ji, J. J. Org. Chem. 1997, 62,
3800. (c) Sibi, M. P.; Shay, J. J.; Ji, J. Tetrahedron Lett. 1997, 38, 5955.
(32) For modification of lanthanide reactivity and or structure by addition
of ligands, see: Aspinall, H. C.; Dwyer, J. L. M.; Greeves, N.; McIver, E.
G.; Wooley, J. C. Organometallics 1998, 17, 1884. Aspinall, H. C.; Greeves,
N.; McIver, E. G. Tetrahedron Lett. 1998, 39, 9283. Greeves, N.; Aspinall,
H. C.; Browning, A. F.; Ravenscroft, P. Tetrahedron Lett. 1994, 35, 4639.
Lacote, E.; Renaud, P. Angew. Chem., Int. Ed. Engl. 1998, 37, 2259.
Fukuzaka, S.-I.; Seki, K.; Tatsuzawa, M.; Mutoh, K. J. Am. Chem. Soc.
1997, 119, 1482.
(27) 99.6% CH2Cl2 from Aldrich chemical company, listed as <0.02%
water; 99.5% THF, listed as <0.02% water.
(28) Triethylborane is stable to water and is capable of initiating radical
reactions under aqueous conditions. See: Yorimitsu, H.; Nakamura, T.;
Shinokubo, H.; Oshima, K. J. Org. Chem. 1998, 63, 8604.
(29) (a) Kobayashi, S. Synlett 1994, 689 and references therein. (b)
Kobayashi, S. Eur. J. Org. Chem. 1999, 15. (c) Keller, E.; Feringa, B. L.
Tetrahedron Lett. 1996, 37, 1879.

Free Radical-Mediated Intermolecular Additions
J. Am. Chem. Soc., Vol. 121, No. 33, 1999 7521
Table 5. Effect of Temperature on Diastereoselectivity
much bulkier auxiliary (R1 ) CHPh2, entries 1, 5). In the
absence of Lewis acid, substrates 1c-e reacted sluggishly and
nonselectively.
We have shown elsewhere that our serine-derived auxiliary
also induces outstanding stereoselectivity in aldol reactions,
alkylation reactions, Diels-Alder reactions, and radical R-al-
lylation reactions.36 In all of these reactions, the diastereose-
lectivity observed using diphenylmethyl-substituted oxazolidi-
nones was equal to or superior to that of benzyl, isopropyl, or
phenyl-substituted oxazolidinones. The superiority of the diphe-
nylmethyl-substituted oxazolidinone auxiliary is particularly
striking in the present conjugate addition context, however. The
majority of chiral oxazolidinone applications have involved
chemistry alpha to the carbonyl, where even the isopropyl and
benzyl groups are close enough to provide efficient face-
shielding (so long as the conformation of the acyloxazolidinone
is controlled). Conjugate addition is a much more demanding
reaction, however, because the reactive beta carbon is one bond
further removed from the oxazolidinone stereocenter; thus
effective face shielding requires a significantly larger blocking
group. While the isopropyl, benzyl, and phenyl groups are too
small, the larger steric volume of the diphenylmethyl group
evidently has sufficient extension to block one face of the beta
carbon. As a general note, it may be that conjugate additions
may serve as a nice test reaction to discriminate between chiral
auxiliaries. It may also be that the diphenylmethyl-substituted
chiral auxiliary may be of common value for “tough cases”, in
applications where simpler oxazolidinone auxiliaries provide
inadequate selectivity.
The Effect of Radical Precursors. A wide range of
organohalides serve as radical precursors and undergo addition
cleanly and with good selectivity under our optimized conditions
(Table 7, eq 5). We generally used iodides, but entry 2 shows
that isopropyl bromide gave the same yield and selectivity as
the iodide, demonstrating that the halogen used as the proradical
species does not impact the selectivity. Primary and secondary
alkyl groups added effectively and easily (entries 1-4, 8-10).37
Addition of tert-butyl radical was sluggish at -78 °C but
proceeded efficiently at -40 °C, although with accordingly
reduced selectivity (entries 5, 11). The lesser reactivity of the
tert-butyl radical was surprising, since tertiary alkyl radicals
are considered to be relatively nucleophilic and normally show
good reactivity toward addition.38 This may reflect a sensitivity
to steric size.
entry
temp (°C)
yield (%)a
drb
1
2
3
4
-78
-40
0
93
90
25:1
15:1
5:1
80 (15%)c
30 (65%)c
25
4:1
a Yields are for isolated and column-purified materials. Five
i
equivalents of Pr-I, 2 equiv of Bu3SnH, and 2 equiv of Et3B were
used. b Diastereomer ratios were determined by H NMR (400 MHz).
1
c
The yield in parentheses is the yield of the reduction product 5.
understand. The success of the reaction in the presence of these
miscellaneous additives also suggests good compatibility with
a wide variety of functionality.
The Yb(OTf)3-mediated alkylation of crotonate 1a was found
to be significantly compromised at temperatures higher than -78
°C (Table 5, eq 2). The stereoselectivity diminished appreciably,
and at room temperature, direct reduction of the alkene became
the predominant reaction pathway, just as in -78 °C reactions
with very strong Lewis acids (Table 1).
The newly created stereocenter in the major diastereomer of
2a and 2b had the S configuration shown (eq 3). The absolute
stereochemistry of the products was established by hydrolysis
(LiOH, H2O2) to give known optically active carboxylic acids
3a and 3b.33 The chiral auxiliary can be easily recovered in
essentially quantitative yield.
The â-substituent of the enoyl group had a significant impact
on the diastereoselectiVity of the radical addition. Under
identical reaction conditions, isopropyl radical addition to 1b
gave 45:1 selectivity as compared to 25:1 for 1a (entries 1 vs
9). Ethyl and cyclohexyl additions to the cinnamate 1b were
also more selective than their additions to the crotonate 1a
(entries 3, 4 vs 8, 10).39
Of special note is the efficient addition of alkoxyalkyl radicals
and especially acyl radicals (entries 6, 7, 12, 13). The ability to
introduce functionalized organic moieties such as acyl groups
under radical conditions highlights one of the advantages of
Relationship between Chiral Auxiliary and Diastereose-
lectivity. Optically active oxazolidinone auxiliaries are readily
derived from amino acids and have been widely applied to
control stereoselectivity in a host of synthetic methods, such as
R-alkylation reactions, aldol reactions, and Diels-Alder reac-
tions.34 The most commonly used optically active oxazolidinones
are the “Evans auxiliaries”, in which the R1 group is phenyl,
benzyl,35 or isopropyl. We have screened the ability of these
auxiliaries to control the stereoselectivity of radical conjugate
additions under our Yb(OTf)3-activated conditions, and the
results are shown in Table 6 (eq 4). Chemical yields were good
in all cases. The results show that traditional oxazolidinone
auxiliaries (R1 ) Ph, benzyl, i-Pr) give modest but not useful
levels of stereocontrol () 3:1, entries 2-4, 6-8), unlike our
(36) Aldol: Sibi, M. P.; Lu, J.; Talbacka, C. L. J. Org. Chem. 1996, 61,
7848. Alkylation: Sibi, M. P.; Deshpande, P. K.; LaLoggia, A. J. Synlett
1996, 343. Diels-Alder: Sibi, M. P.; Deshpande, P. K.; Ji, J. Tetrahedron
Lett. 1995, 36, 8965. R-allylation: Sibi, M. P.; Ji, J. Angew. Chem., Int.
Ed. Engl. 1996, 35, 190.
(37) The configuration of 2c-l is assumed by analogy to 2a and 2b to
be S.
(38) Fossey, J.; Lefort, D.; Sorba, J. Free Radicals in Organic Synthesis;
Wiley: New York, 1995; pp 49-71.
(39) We have not carried out careful kinetic studies, but in general, the
cinnamate was qualitatively observed to be less reactive than the crotonate.
(33) The absolute stereochemistry of the products was established by
hydrolysis: Enders, D.; Rendenbach, B. E. M. Tetrahedron 1986, 42, 2235.
Lardicci, L.; Salvadori, P.; Caporusso, A. M.; Menicagli, R.; Belgodere, E.
Gazz. Chim. Ital. 1972, 102, 64.
(34) For a recent monograph, see: Ager, D. A.; East, M. B. Asymmetric
Synthetic Methodology, CRC: Boca Raton, FL, 1996.
(35) Gage J. R.; Evans, D. A. Org. Synth. 1989, 68, 77.

7522 J. Am. Chem. Soc., Vol. 121, No. 33, 1999
Sibi et al.
Table 6. Relationship between Chiral Auxiliary and Diastereoselectivity
entry
substrate
R1
R2
product
yield (%)a
drb
entry
substrate
R1
R2
product
yield (%)a
drb
1
2
3
4
(R)-1a
(S)-1c
(S)-1d
(S)-1e
CHPh2
Ph
CH2Ph
i-PrI
Me
Me
Me
Me
2a
2r
2s
2t
90
82
81
82
25:1
3:1
2:1
3:1
5
6
7
8
(R)-1b
(R)-1f
(S)-1g
(S)-1h
CHPh2
Ph
CH2Ph
i-Pr
Ph
Ph
Ph
Ph
2b
2u
2v
2w
94
92
85
90
45:1
2.2:1
2.2:1
3.4:1
b
1
a Yields are for isolated and column-purified materials. Diastereomer ratios were determined by H NMR (400 MHz).
Table 7. Effect of Radical Precursors on Conjugate Additions
Lewis acid
(equiv)
yield
Lewis acid
(equiv)
yield
product (%)a
entry substrate
R3X
i-PrI
i-PrBr
EtI
product (%)a
drb entry substrate
R3X
drb
1
2
(R)-1a
(R)-1a
(R)-1a
(R)-1a
(R)-1a
(R)-1a
(R)-1a
(R)-1b
(R)-1b
(R)-1b
Yb(OTf)3 (1)
Yb(OTf)3 (1)
Yb(OTf)3 (1)
Yb(OTf)3 (1)
2a
2a
2c
2e
2f
2g
2h
2d
2b
2i
90
90
84
92
25:1
25:1
12:1
16:1
11
12
13
14
15
16
17
18
19
(R)-1b
(R)-1b
(R)-1b
(R)-1a
(R)-1a
(R)-1a
(R)-1a
(R)-1a
(R)-1a
tert-ButylI
MeOCH2Br
MeCOBr
PhI
Yb(OTf)3 (1)
Yb(OTf)3 (1)
Yb(OTf)3 (1)
Yb(OTf)3 (1)
Yb(OTf)3 (1)
Yb(OTf)3 (1)
Yb(OTf)3 (1)
Yb(OTf)3 (1)
2j
2k
2l
90
81
84
9:1c
10:1
8:1
3
4
5
6
c-C6H11I
2m
2n
2o
2p
2q
2r
<5%d
<5%d
<5%d
<5%d
<5%d
<5%d
tert-ButylI Yb(OTf)3 (1)
MeOCH2Br Yb(OTf)3 (1)
82 14:1c
MeI
84
85
80
94
88
14:1
7:1
allyl-I
PhCH2I
AcOCH2Br
7
8
9
10
MeCOBr
EtI
i-PrI
Yb(OTf)3 (1)
Yb(OTf)3 (1)
Yb(OTf)3 (1)
Yb(OTf)3 (1)
20:1
45:1
26:1
BrCH2CO2Bn Yb(OTf)3 (1)
c-C6H11I
a Yields are for isolated and column-purified materials. b Diastereomer ratios were determined by 1H NMR (400 MHz). c Reactions were conducted
1
at -40 °C. d The product mixtures from entries 14-19 were analyzed by H NMR only.
radical procedures relative to ionic methods.40 We were also
delighted to find that simple acyl bromides functioned as
convenient acyl radical precursors under our reaction condi-
tions.41 Methoxymethyl and acetyl radical additions were
qualitatively observed to react faster than simple alkyl radicals,42
in keeping with their nucleophilic character.43
Several organohalides did not undergo addition. Phenyl,
methyl, allyl, and benzyl iodide, bromomethyl acetate, and
benzyl bromoacetate all failed to react with crotonate 1a under
the standard conditions (entries 14-18).44 Aryl radicals are
known to react fairly rapidly with THF and some THF radical
addition was observed when PhI was used, so we conclude that,
under our reaction conditions, the Ph• radical is hydrogenated
by Bu3SnH/THF too quickly to allow clean addition. That allyl
and benzyl radicals have low reactivity is well-known, so their
failure to add is not at all surprising.45 When benzyl iodide was
used, some dibenzyl was observed in addition to mostly toluene.
The use of (Me3Si)3SiH or syringe pump addition of Bu3SnH
did not make reaction of benzyl iodide successful. We believe
•
•
that the radicals AcOCH2 and CH2CO2Bn are “mismatched”
(42) While the results in the table reflect standard reaction times, TLC
analysis showing reactions involving methoxymethyl bromide or acetyl
bromide were complete within 15 min. NMR analysis of crude reaction
mixtures also showed no indication of ethyl addition, indicating long,
efficient radical chains; 1-3% ethyl addition was often observed when
simple alkyl radicals were used.
(43) Alkoxyalkyl radicals are “nucleophilic”. See: Giese, B. Radicals
in Organic Synthesis: Formation of Carbon-Carbon Bonds; Pergamon:
Oxford, U.K., 1986, Chapter 2. Giese, B.; Dupuis, J.; Haâkerl, T.; Meixner,
J. Tetrahedron Lett. 1983, 24, 703. Acyl radicals are also nucleophilic; see
ref 41.
(40) We have also found that haloalkyl radicals derived from reagents
such as CH2ICl and CH3CHBr2 add very efficiently to provide halogenated
products (unpublished results).
(41) Acyl radicals have usually been prepared from acyl selenides. See:
Boger, D. L.; Mathvink, R. J. J. Org. Chem. 1992, 57, 1429.
(44) The reactions were analyzed by 1H NMR only. No further attempt
was made to characterize the products.

Free Radical-Mediated Intermolecular Additions
J. Am. Chem. Soc., Vol. 121, No. 33, 1999 7523
Scheme 1
with the electrophilic Lewis acid-complexed enamides such that
addition is apparently too slow to compete with radical
hydrogenation. It appears that, while the addition is efficient
for radicals of reasonable reactivity and nucleophilicity, reactiv-
ity problems occur under our conditions when electrophilic
radicals, highly stabilized radicals, or too highly reactive radicals
(aryl radicals) are used.
hydrogen donor will prevent hydrogenation, our initial attempts
to use (Me3Si)3SiH as a chain carrier have failed.
Following radical addition to the Lewis acid activated alkene,
the adduct radical then reacts with tin hydride. It is significant
that we did not observe any polymerization when soluble Lewis
acids were used. The kinetic partitioning is balanced such that
a nucleophilic radical selectively adds to the Lewis acid-
activated substrate, faster than it abstracts hydrogen from Bu3-
SnH; however the adduct radical reacts with Bu3SnH much
faster than it adds to another substrate. This reversal in
chemoselectivity reflects that the substrate is electrophilic and
reacts faster with a nucleophilic isopropyl radical than with
electrophilic radicals. As discussed earlier, when the radical R•
is electrophilic, then addition to the activated alkene is too slow
and the electrophilic radical instead abstracts hydrogen atom
from Bu3SnH. It is also notable that we saw no evidence for
the addition of the Bu3Sn• radical, halogen transfer apparently
remaining much faster. This was true even when only 1.2 equiv
of isopropyl iodide was used or when isopropyl iodide was
replaced by isopropyl bromide.
That a radical mechanism was operative under our standard
procedure is supported by the observation that, when Et3B was
omitted, no reaction proceeded and starting materials could be
recovered. When Bu3SnH was omitted, no product 2a formed,
although some side reactions (including ethyl transfer)46 were
observed.
The nature of the radical chain initiation involves the reaction
of triethylborane with molecular oxygen to produce ethyl radical.
The ethyl radical can initiate the chain in any of three ways.
Since use of iodoethane gave alkylation in high yield, it is
evident that ethyl radical adds effectively to the activated alkene
under our conditions, regardless of whether the ethyl radical is
Mechanism and Model. The probable mechanism for the
reaction is shown in Scheme 1. Initiation indirectly generates
an alkyl radical R• that adds to the Lewis acid-activated alkene.
Lewis acid activation is crucial for this addition step; in the
absence of Lewis acid addition is slow, so that radical
hydrogenation to give R-H occurs preferentially. The competi-
tion between radical addition and hydrogenation explains why
the Lewis acid must be soluble, so that there is a substantial
concentration of activated alkene. The competition between
addition and hydrogenation also explains why electrophilic or
benzylic radicals fail to add; in these cases addition evidently
becomes slower than hydrogenation (even though the rate of
hydrogenation is also reduced when radicals are stabilized by
conjugation). This explanation is supported by the observation
that no tributyltin hydride remained (based on GC or I2 titration)
following reaction with benzyl iodide, methyl iodide, or benzyl
2-bromoacetate. Additionally, in reactions with benzyl iodide
and benzyl 2-bromoacetate, the reduced products toluene and
benzyl acetate were found in the crude reaction mixtures. We
consider the competition between radical addition and hydro-
genation to be the most significant limitation to our procedure.
If structural modifications in either the radical or the alkene
result in significant reductions in the rate of addition, it may
become increasingly difficult for addition to win over hydro-
genation. While it is reasonable that the use of a less reactive
(45) We have recently found that benzyl radicals add to the more reactive
chiral fumarates in good yields.
(46) Pelter, A.; Smith, K.; Brown, H. C. Borane Reagents; Academic:
London, U.K., 1988.

7524 J. Am. Chem. Soc., Vol. 121, No. 33, 1999
Sibi et al.
was still good using only 10 mol % (Table 3, entries 8-11).
Free substrate is also inherently less reactive than Lewis acid-
complexed substrate (see Table 1, entry 1) and thus unlikely to
provide a major competing pathway unless present in substantial
quantities.
Figure 2.
Summary
produced from iodoethane or from triethylborane. Iodine atom
transfer4 may also be important but is not necessary, on the
basis of the observation that replacement of isopropyl iodide
with isopropyl bromide did not inhibit the reaction or complicate
the product mixture. Additionally, hydrogen atom abstraction
from tinhydride could also participate. If initiation involves ethyl
addition, the formation of some ethyl addition product 2c is a
necessary consequence of using triethylborane as initiator. The
absence of significant quantities of ethyl addition product 2c,
however, implies that radical chains are long under our standard
conditions.
The high diastereoselectivity observed in the â-radical
addition can be explained by a chelation model (Figure 2). Upon
Lewis acid chelation, the orientations of the two carbonyls are
fixed. Radical addition to the chelated substrate then takes place
from the face opposite of the bulky diphenylmethyl substituent.
Because the reactive beta carbon is fairly distant from the chiral
carbon in the chiral auxiliary, the oxazolidinone substituent must
be very large to afford adequate face shielding. Thus replace-
ment of the diphenylmethyl group by the smaller phenyl, benzyl,
or isopropyl groups gives unacceptable diastereoselectivity. The
dependence of stereoselectivity on the lanthanide or pre-
lanthanide triflate suggests that ytterbium’s ionic radius of about
0.9 Å is nearly optimal for radical additions. A shorter ionic
radius may “pull” the enoyl carbonyl away from the diphenyl-
methyl group and thus expose the back face of the reactive beta
carbon toward radical addition. We had anticipated that a longer
ionic radius would enhance selectivity by “pushing” the enoyl
portion further over the blocking group, but the results
contradicted this notion. For Lewis acids whose ionic radii are
too long, distortions from planarity may occur that either block
the top face or somehow expose the bottom face.
The significantly higher selectivity observed for cinnamate
1b as compared to crotonate 1a (Table 7) is puzzling. This
observation, combined with the decreased selectivity in the
presence of toluene (Table 3, entries 3 vs 7), raises the possibility
that π-stacking may play a role in organizing the orientation of
the diphenylmethyl group relative to the enoyl portion.47 If so,
perhaps the π-stacking is more effective for cinnamate than
crotonate because the former is more highly conjugated, and
toluene reduces the stereoselectivity because it disrupts the
auxiliary-enoyl π-stack.
With good chelating Lewis acids under stoichiometric condi-
tions, we believe that the reaction proceeds primarily from the
Lewis acid-substrate complex; the minor isomer probably
results from radical addition to the “wrong” face of the complex
rather than from competing nonselective addition to substrate
in which only one or neither carbonyl is bound to Lewis acid.
NMR studies in CD3CN show very strong binding between
acyloxazolidinone substrates and chelating Lewis acids.48 In the
case of MgBr2, replacement of ether with the more Lewis-basic
THF did not reduce the selectivity (Table 1, entries 9 and 10),
as would have been expected if the solvent was competing with
the substrate for Lewis acid binding sites. Only minimal
reduction in selectivity was observed when the amount of Yb-
(OTf)3 was reduced from 2 to 1 to 0.3 equiv, and selectivity
We have developed a highly diastereoselective method for
the conjugate addition of carbon radicals to chiral R,â-
unsaturated N-enoyloxazolidinones using Bu3SnH as chain
carrier and Et3B/O2 as radical initiator. Lewis acids have been
screened, and Yb(OTf)3 proved to give optimized results for
both chemical yield and diastereoselectivity. The virtue of the
Yb(OTf)3 lies in its ability to chelate both substrate carbonyls
and thus control the conformation of the reactive substrate; in
its ionic radius, which is neither too long nor too short; and in
its Lewis acidity, which is strong enough to greatly activate
the substrates but not so strong that hydrogenation of the
substrate interferes at -78 °C. Chemical yields were outstand-
ing, especially given the normally sluggish reactivity of non-
terminal alkenes toward radical addition, and no competing
polymerization was observed. The diastereoselectivity observed
is comparable to or better than that observed in analogous
conjugate additions using ionic methods.
The selectivity is solvent-dependent, CH2Cl2-THF being an
ideal combination. Scrupulously dry solvents or reaction condi-
tions were not required. Substoichiometric amounts of Yb(OTf)3
provided efficient reaction with minimal sacrifice in diastereo-
selectivity. Our diphenylmethyl-substituted oxazolidinone aux-
iliaries gave greatly superior selectivity compared to phenyl-,
benzyl-, or isopropyl-substituted oxazolidinones. Carbon radicals
with reasonable nucleophilicity were generally successful,
including functionalized radicals such as acetyl or methoxy-
methyl. Electrophilic radicals were not successful.
Experimental Procedures
All reagents were used as received from the supplier. Tetrahydro-
furan, ether, and 1,2-dimethoxyethane were distilled from sodium
benzophenone/ketyl prior to use. Chloroform, hexane, and CH2Cl2 were
distilled from calcium hydride. Standard benchtop techniques were
employed for handling air-sensitive reagents, and all reactions were
carried out under nitrogen. Flash column chromatography was per-
formed using Merck 60 silica gel, 230-400 mesh. H and 13C NMR
1
spectra were recorded in CDCl3 at 270 (400) and 65 MHz, respectively.
Chiral HPLC analysis were performed using a Chiralcel OD column
(Chiral Technologies, Inc.) on a ISCO system comprising of a 2360
pump, 2350 gradient programmer, and a variable wavelength UV
detector. Rotations were recorded on a JASCO-DIP-370 instrument.
Elemental analyses were performed in house on a Perkin-Elmer
instrument.
Lewis Acid-Mediated Intermolecular â-Selective Radical Addi-
tion to 1a: Scale-up Procedure for Product Characterization. The
following conditions differ slightly from the typical reaction conditions
stated in the text. The selectivity using this procedure is identical to
that listed in the text under standard conditions. To a flask containing
1a (160 mg, 0.5 mmol) Yb(OTf)3 (210 mg, 0.5 mmol), THF (5 mL),
and CH2Cl2 (5 mL) under N2 was added i-PrI (425 mg, 2.5 mmol),
Bu3SnH (730 mg, 2.5 mmol), and Et3B (1 M in hexane) (1 mL, 1 mmol)
at -78 °C. Five milliliters of O2 was then added via syringe over 2
min. 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 2a was purified by chromatography on silica gel using hexane/
ethyl acetate (9:1) as the eluent, yield 170 mg (93%).
2a. mp 67-70 °C; Rf ) 0.75 (80:20 hexane/EtOAc); 1H NMR (400
MHz, CDCl3) δ 0.81 (d, J ) 6.6 Hz, 3H), 0.86 (d, J ) 7.3 Hz, 3H),
0.89 (d, J ) 6.6 Hz, 3H), 1.50-1.63 (m, 1H), 1.98-2.12 (m, 1H),
(47) Jones, G. J.; Chapman, B. J. Synthesis 1995, 475.
(48) Unpublished results.

Free Radical-Mediated Intermolecular Additions
J. Am. Chem. Soc., Vol. 121, No. 33, 1999 7525
2.73 (dd, J ) 16.1, 9.6 Hz, 1H), 3.04 (dd, J ) 16.1, 4.3 Hz, 1H), 3.36
(d, J ) 9.3 Hz, 1H), 3.80 (dd, J ) 9.3, 2.2 Hz, 1H), 4.66 (d, J ) 5.9
Hz, 1H), 4.86-4.95 (m, 1H), 6.80-7.10 (m, 10H); 13C NMR (270 MHz,
CDCl3) δ 15.8, 18.3, 20.5, 32.3, 35.2, 40.3, 51.8, 56.8, 65.6, 127.5,
128.3, 128.9, 129.1, 129.4, 129.6, 138.6, 140.1, 173.3; [R]D +107.7
(c 0.545, CH2Cl2). Anal. calcd for C23H27NO3: C, 75.59; H, 7.45.
(m, 1H), 2.77-2.83 (dd, J ) 16.1, 9.1 Hz, 1H), 3.06-3.11 (dd, J )
16.1, 9.1 Hz, 1H), 3.35-3.39 (dt, J ) 17.5, 7.7, 1.9 Hz, 1H), 3.79-
3.82 (dd, J ) 9.1, 2.4 Hz, 1H), 4.70-4.71 (d, J ) 5.9 Hz, 1H), 4.90-
4.94 (m, 1H), 6.83-6.85 (m, 2H), 6.93-7.07 (m, 8H); 13C NMR (100
MHz, CDCl3) δ 16.3, 26.6, 26.7, 26.8, 28.7, 30.5, 34.3, 42.6, 51.2,
56.4, 65.2, 127.1, 127.9, 128.5, 128.7, 129.0, 129.3, 138.2, 139.7, 153.5,
172.9, 173.0; [R]D26 -93.82 (c 0.534, CHCl3). Anal. calcd for C26H31-
NO3: C, 77.00; H, 7.70; N, 3.45. Found: C, 76.74; H, 7.36; N, 3.82.
tert-Butyl Radical Addition to 1a. Product 2f. Yield 82%; white
26
Found: C, 75.76, H, 7.45.
2b. mp 173-175 °C; Rf ) 0.8 (70:30 hexane/EtOAc); H NMR
1
(400 MHz, CDCl3) δ 0.76 (d, J ) 6.6 Hz, 3H), 0.98 (d, J ) 6.6 Hz,
3H), 1.84-1.99 (m, 1H), 2.90-3.04 (m, 2H), 3.55-3.66 (m, 1H), 4.16
(t, J ) 8.8 Hz, 1H), 4.28 (dd, J ) 9.5, 2.2 Hz, 1H), 4.64 (d, J ) 5.9
Hz, 1H), 5.05-5.12 (m, 1H), 7.05-7.45 (m, 15H); 13C NMR (270 MHz,
CDCl3) δ 20.8, 21.3, 33.4, 38.9, 48.8, 51.5, 56.9, 65.5, 126.7, 127.5,
128.3, 128.5, 128.8, 128.9, 129.0, 129.3, 129.6, 138.6, 140.0, 143.5,
153.9, 172.6; [R]D26 -142.4 (c 0.460, CH2Cl2). Anal. calcd for C28H29-
NO3: C, 78.66, H, 6.84. Found: C, 78.97, H, 7.05.
Typical Procedure using Ethylene Glycol as an Additive. To a
solution of Yb(OTf)3 (0.01 mmol) in 4:1 CH2Cl2/THF (2.5 mL) was
added ethylene glycol (0.01 mmol), and the mixture was stirred for 5
min. This was followed by addition of substrate 1b (0.1 mmol) and
i-PrI (0.5 mmol), and the reaction was cooled to -78 °C. Bu3SnH (0.2
mmol) and triethylborane (0.4 mmol) were added in sequence, and the
reaction was initiated by oxygen. After completion (TLC), normal
workup gave 2b.
1
solid; mp 75-83°; H NMR (400 MHz, CDCl3) δ 0.72-0.74 (d, J )
6.7 Hz, min), 0.77-0.79 (d, J ) 6.7 Hz, 3H), 0.85 (s, min), 0.86 (s,
9H), 1.71-1.85 (m, 1H), 2.50-2.57 (dd, J ) 16.1, 6.4 Hz, min), 2.58-
2.64 (dd, J ) 15.8, 5.9 Hz, 1H), 2.82-2.87 (dd, J ) 16.5, 2.1 Hz,
1H), 2.92-2.97 (dd, J ) 17.1, 2.4 Hz, min), 4.29-4.42 (m, 2H), 4.66-
4.67 (d, J ) 5.5 Hz, 1H), 4.71-4.72 (d, J ) 5.6 Hz, min), 5.32-5.36
(m, 1H), 7.11-7.36 (m, 10H); 13C NMR (100 MHz, CDCl3) δ 27.3,
32.9, 38.2, 38.7, 51.4, 56.5, 65.3, 127.2, 127.9, 128.5, 128.7, 129.0,
26
129.3, 138.3, 139.7, 153.5, 173.4; [R]D -248.83 (c 0.512, CHCl3).
Anal. calcd for C24H29NO3: C, 75.96; H, 7.70; N, 3.69.
Methoxymethyl Radical Addition to 1a. Product 2g. Yield 84%;
white solid; mp 81-82°; Rf ) 0.18 (70:30 hexanes/EtOAc); 1H NMR
(400 MHz, CDCl3) δ 0.86-0.88 (d, J ) 6.7 Hz, min), 0.90-0.91 (d,
J ) 6.8 Hz, 3H), 2.19-2.31 (m, 1H), 2.53-2.59 (dd, J ) 16.9, 7.4
Hz, min), 2.67-2.73 (dd, J ) 16.9, 7.5 Hz, 1H), 2.86-2.92 (dd, J )
16.9, 5.9 Hz, 1H), 2.98-3.04 (dd, J ) 16.6, 5.9 Hz, min), 3.24-3.14
(m, 2H), 3.28 (s, min), 3.32 (s, 3H), 4.32-4.46 (m, 2H), 4.62-4.64
(d, J ) 6.4 Hz, min), 4.67-4.68 (d, J ) 5.9 Hz, 1H), 5.30-5.34 (m,
1H), 7.10-7.36 (m, 10H); 13C NMR (100 MHz, CDCl3) δ 17.2, 29.7,
39.4, 51.3, 56.4, 58.9, 65.2, 77.3, 127.2, 127.9, 128.5, 128.8, 129.0,
129.3, 138.3, 139.7, 153.6, 172.1; [R]D26 -123.80 (c 1.0, CH2Cl2). Anal.
calcd for C22H25NO4: C, 71.91; H, 6.86; N, 3.81. Found: C, 71.56; H,
6.85; N, 3.80.
Hydrolysis of 2a. Typical Procedure. To a flask containing 2a (256
mg, 0.7 mmol) THF (5 mL) and H2O (5 mL) under N2 was added
H2O2 (30%) (0.317 mL, 2.8 mmol) at 0 °C. LiOH‚H2O (57 mg, 1.4
mmol) was then added, and the reaction mixture was stirred at 0 °C
for 1 h. After completion of the reaction (TLC), most of the THF was
evaporated. The aqueous solution (pH ) 12) was extracted with CH2-
Cl2 (3 × 10 mL) (recovery of chiral auxiliary). The aqueous solution
was acidified with HCl (3 M) until pH ∼1 and extracted again with
CH2Cl2 (4 × 15 mL). The organic solution was dried (MgSO4) and
concentrated to yield (S)-3,4-dimethyl-pentanoic acid (85 mg, 93%).
Acetyl Radical Addition to 1a. Product 2h. Yield 88%; clear solid;
mp 115-119°; Rf ) 0.15 (70:30 hexane/EtOAc); 1H NMR (400 MHz,
CDCl3) δ 1.12-1.14 (d, J ) 7.5 Hz, min), 1.15-1.17 (d, J ) 5.1 Hz,
3H), 2.20 (s, min), 2.25 (s, 3H), 2.67-2.72 (dd, J ) 18.5, 3.8 Hz,
min), 2.79-2.84 (dd, J ) 18.5, 3.8 Hz, 1H), 3.00-3.09 (m, 1H), 3.32-
3.40 (dd, J ) 18.5, 9.8 Hz, min), 3.38-3.45 (dd, J ) 18.5, 9.9 Hz,
1H), 4.34-4.36 (dd, J ) 9.3, 2.4 Hz, min), 4.41-4.49 (m, 2H), 4.64-
4.65 (d, J ) 5.6 Hz, min), 4.69-4.70 (d, J ) 4.6, 1H), 5.28-5.22 (m,
1H), 7.05-7.38 (m, 10H); 13C NMR (100 MHz, CDCl3) δ 16.6, 28.6,
39.0, 41.9, 49.8, 56.0, 64.7, 127.1, 127.9, 128.3, 128.8, 128.9, 129.7,
1
(S)-3,4-Dimethylpentanoic Acid. H NMR (400 MHz, CDCl3) δ
0.88 (d, J ) 6.5 Hz, 3H), 0.92 (d, J ) 7 Hz, 3H), 0.94 (d, J ) 6.4 Hz,
3H), 1.62-1.67 (m, 1H), 1.88-1.94 (m, 1H), 2.14 (dd, J ) 15.1, 9.1
26
Hz, 1H), 2.44 (dd, J ) 15.1, 4.8 Hz, 1H); [R]D -10.7 (c 0.55,
benzene). {lit: [R]D21 -6.9 (c 1.18, benzene); Enders, D.; Rendenbach,
B. E. M. Tetrahedron 1986, 42, 2235}.
(S)-3-Phenyl-4-methylpentanoic Acid. 1H NMR (400 MHz, CDCl3)
δ 0.75 (d, J ) 6.5 Hz, 3H), 0.93 (d, J ) 7.0 Hz, 3H), 1.82-1.90 (m,
1H), 2.62 (dd, J ) 15.6, 9.6 Hz, 1H), 2.80 (dd, J ) 15.1, 5.4 Hz, 1H),
2.82-2.90 (m, 1H), 7.20-7.32 (m, 5H); [R]D26 -33.6 (c 3.77, CHCl3)
or [R]D26 -41.0 (c 0.405, benzene). {lit:[R]D25 -34.4 (c 4.06, CHCl3);
Lardicci, L.; Salvadori, P.; Caporusso, A. M.; Menicagli, R.; Belgodere,
E. Gazz. Chim. Ital. 1972, 102, 64}.
26
137.9, 139.7, 153.4, 171.8, 211.1; [R]D -248.83 (c 0.512, CHCl3).
Anal. calcd for C22H23NO4: C, 72.31; H, 6.34; N, 3.83. Found: C,
72.02; H, 6.57; N, 3.69.
Cyclohexyl Radical Addition to 1b. Product 2i. Yield 85%; white
solid; mp. 169-173°; Rf ) 0.40 (70:30 hexane/EtOAc); 1H NMR (400
MHz, CDCl3) δ 0.73-1.85 (m, 11H), 2.90-3.00 (m, 2H), 3.51-3.59
(m, 1H), 4.08-4.13 (app. t, J ) 8.3 Hz, 1H), 4.23-4.26 (dd, J ) 9.1,
2.1 Hz, 1H), 4.37-4.39 (m, min), 4.48-4.49 (d, J ) 4.3 Hz, min),
4.59-4.61 (d, J ) 5.9 Hz, 1H), 5.01-5.05 (m, 1H), 5.15-5.19 (m,
min), 7.03-7.34 (m, 15H); 13C NMR (100 MHz, CDCl3) δ 26.4, 26.5,
30.8, 31.2, 38.2, 42.7, 47.6, 51.0, 56.6, 65.1, 126.3, 127.1, 127.9, 128.1,
128.4, 128.5, 128.7, 129.0, 129.3, 138.2, 139.6, 143.3, 153.9, 172.3;
[R]D26 -130.0 (c 1.00, CH2Cl2). Anal. calcd for C31H33NO4: C, 76.99;
H, 6.86; N, 2.90. Found: C, 77.13; H, 6.78; N, 3.11.
Ethyl Radical Addition to 1a. Product 2c. Yield 83%; white solid;
1
mp 58-60°; Rf ) 0.43 (70:30 hexane/EtOAc); H NMR (400 MHz,
CDCl3) δ 0.84-0.87 (m, 6H), 1.00-1.36 (m, 2H), 1.78-1.90 (m, 1H),
2.58-2.64 (dd, J ) 16.4, 8.1 Hz, 1H), 2.77-2.82 (dd, J ) 16.2, 4.0
Hz, 1H), 4.34-4.41 (m, 2H), 4.67-4.68 (d, J ) 6.2 Hz, 1H), 4.72-
4.74 (d, J ) 5.6 Hz, min), 5.31-5.36 (m, 1H), 7.36-7.11 (m, 10H);
13C NMR (100 MHz, CDCl3) δ 11.4, 19.2, 29.3, 31.0, 42.2, 51.3, 56.4,
26
65.2, 127.1, 127.9, 128.5, 129.0, 138.3, 139.7, 153.5, 172.5; [R]D
-110.09 (c 1.09, CH2Cl2). Anal. calcd for C22H25NO3: C, 75.19; H,
7.17; N, 3.99. Found: C, 75.05; H, 7.20; N, 3.91.
tert-Butyl Radical Addition to 1b. Product 2j. Yield 90%; white
solid; mp 194-195°; Rf ) 0.40 (70:30 hexanes/EtOAc); 1H NMR (400
MHz, CDCl3) δ 0.89 (s, 9H), 2.82-2.88 (dd, J ) 15.5, 3.2 Hz, 1H),
2.99-3.02 (dd, J ) 11.5, 3.0 Hz, 1H), 3.74-3.81 (dd, J ) 17.2, 11.5
Hz, 1H), 4.16-4.18 (t, J ) 5.6 Hz, 1H), 4.24-4.27 (dd, J ) 11.8, 2,4
Hz, 1H), 4.58-4.59 (d, J ) 6.2 Hz, 1H), 4.63-4.64 (d, J ) 5.6 Hz,
min), 5.02-5.04 (m, 1H), 5.10-5.15 (m, min), 6.99-7.35 (m, 15H);
13C NMR (100 MHz, CDCl3) δ 28.2, 33.8, 35.9, 51.1, 55.5, 56.5, 65.2,
126.4, 127.1, 127.8, 127.9, 128.5, 128.7, 129.0, 129.3, 138.2, 139.6,
Ethyl Radical Addition to 1b. Product 2d. Yield 80%; white solid;
mp 140-142°; Rf ) 0.38 (70:30 hexane/EtOAc); 1H NMR (400 MHz,
CDCl3) δ 0.74-0.79 (t, J ) 7.2 Hz, 3H), 1.54-1.65 (m, 2H), 2.91-
3.02 (m, 2H), 3.35-3.42 (dd, J ) 16.5, 8.3 Hz, 1H), 4.18-4.22 (t, J
) 8.1 Hz, 1H), 4.27-4.29 (dd, J ) 9.1, 2.4 Hz, 1H), 4.53-4.54 (d, J
) 4.8 Hz, min), 4.63-4.64 (d, J ) 5.9 Hz, 1H), 5.12-5.16 (m, 1H),
5.22-5.26 (m, min), 7.06-7.36 (m, 15H); 13C NMR (100 MHz, CDCl3)
δ 12.1, 29.3, 41.5, 43.1, 56.5, 65.2, 126.4, 127.1, 127.8, 127.9, 128.4,
26
26
128.7, 129.0, 129.3, 129.4, 138.2, 139.6, 144.1, 153.5, 171.8; [R]D
142.1, 153.7, 172.4; [R]D -117.3 (c 1.10, CH2Cl2). Anal. calcd for
-106.7 (c 0.52, CHCl3). Anal. calcd for C27H27NO3: C, 78.42; H, 6.58;
N, 3.39. Found: C, 78.35; H, 6.54; N, 3.46.
C29H31NO3: C, 78.89; H, 7.08; N, 3.17. Found: C, 78.54; H, 7.28; N,
3.18.
Cyclohexyl Radical Addition to 1a. Product 2e. Yield 87%; white
solid; mp 114-116°; Rf ) 0.40 (70:30 hexanes/EtOAc); 1H NMR (400
MHz, CDCl3) δ 0.84-1.29 (m, 9H), 1.59-1.71 (m, 5H), 2.03-2.11
Methoxymethyl Radical to 2b. Product 2k. Yield 84%; white solid;
mp 153-155°; Rf ) 0.27 (70:30 hexane/EtOAc); 1H NMR (400 MHz,
CDCl3) δ 3.14-3.53 (m, 5H), 3.33 (s, 3H), 4.27-4.42 (m, 2H), 4.55-

7526 J. Am. Chem. Soc., Vol. 121, No. 33, 1999
Sibi et al.
4.54 (d, J ) 5.1 Hz, min), 4.66-4.67 (d, J ) 6.2 Hz, 1H), 5.16-5.20
(m, 1H), 5.25-5.29 (m, min), 7.05-7.12 (m, 4H), 7.18-7.36 (m, 11H);
13C NMR (100 MHz, CDCl3) δ 38.4, 41.2, 50.8, 56.4, 58.9, 65.1, 76.5,
126.9, 127.1, 127.9, 128.0, 128.4, 128.6, 128.7, 129.0, 129.3, 138.1,
137.9, 141.7, 153.5, 171.5; [R]D26 -130.8 (c 0.52, CH2Cl2). Anal. calcd
for C27H27NO4: C, 75.50; H, 6.34; N, 3.26. Found: C, 75.48; H, 6.34;
N, 3.26.
7.22-7.41 (m, 12H); 13C NMR (100 MHz, CDCl3) δ 29.1, 39.1, 49.6,
54.0, 56.0, 64.5, 127.1, 127.8, 127.9, 128.3, 128.5, 128.8, 128.9, 129.3,
129.8, 137.3, 137.9, 139.8, 153.2, 171.6, 206.9; [R]D26 -297.7 (c 0.997,
CH2Cl2). Anal. calcd for C27H25NO4: C, 75.86; H, 5.89; N, 3.27.
Found: C, 75.92; H, 6.24; N, 3.28.
Acknowledgment. This work was supported by the National
Acetyl Radical Addition to 1b. Product 2l. Yield, 84%: mp. 110-
Institutes of Health (GM-54656).
1
113°; Rf ) 0.26 (70:30 hexane/EtOAc); H NMR (400 MHz, CDCl3)
δ 2.04 (s, min), 2.17 (s, 3H), 2.86-2.92 (dd, J ) 18.3, 3.5, 1H), 3.00-
3.05 (dd, J ) 18.8, 3.5 Hz, 1H), 3.81-3.88 (dd, J ) 9.8, 10.6 Hz,
min), 3.88-3.95 (dd, J ) 18.8, 10.7 Hz, 1H), 4.26-4.29 (dd, J )
10.7, 3.5 Hz, 1H), 4.35-4.37 (dd, J ) 9.1, 2.4 Hz, min), 4.42-4.43
(m, 2H), 4.62-4.63 (d, J ) 5.9 Hz, min), 4.75-4.76 (d, J ) 4.0 Hz,
1H), 7.06-7.07 (d, J ) 7.5 Hz, 2H), 7.18-7.20 (d, J ) 7.5 Hz, 1H),
Supporting Information Available: Characterization data
for compounds 1c, 1d, 1e, 1f, 1g, 1h, and 2r, 2s, 2t, 2u, 2v,
and 2w. This material is available free of charge via the Internet
at http://pubs.acs.org.
JA991205E