The regioselectivity of addition of organolithium reagents to enones and enals: The role of HMPA

Article abstract of DOI:10.1021/ja010053w

The role of polar solvents (particularly HMPA) in controlling the ratio of 1.2 to 1,4 addition of sulfur-substituted organolithium reagents to cyclohexenones and hexenal was studied. Low-temperature, multinuclear NMR studies provided quantitative informat

Full text of DOI:10.1021/ja010053w

J. Am. Chem. Soc. 2001, 123, 6527-6535
6527
The Regioselectivity of Addition of Organolithium Reagents to
Enones and Enals: The Role of HMPA
William H. Sikorski and Hans J. Reich*
Contribution from the Department of Chemistry, UniVersity of Wisconsin, Madison, Wisconsin 53706
ReceiVed January 4, 2001
Abstract: The role of polar solvents (particularly HMPA) in controlling the ratio of 1,2 to 1,4 addition of
sulfur-substituted organolithium reagents to cyclohexenones and hexenal was studied. Low-temperature,
multinuclear NMR studies provided quantitative information about the ratio of contact (CIP) and solvent-
separated (SIP) ion pairs in solutions of dithianyllithiums and phenylthiobenzyllithiums in THF-HMPA
solutions. The ratio of contact and separated ion pairs was manipulated by changes in the strength of solvation
(generally through the addition of HMPA). Although the results are consistent with the CIP/SIP distribution
being an important factor in determining the regioselectivity of these additions (Curtin-Hammett limitations
prevent a direct correlation), it cannot be the only one. Changes in diastereomeric product ratios upon addition
of HMPA suggest that complexation of HMPA to lithium has two effects. First, it causes ion pair separation,
which enhances 1,4 addition. Second, it lowers the Lewis acidity and catalytic effectiveness of the lithium
cation, which also favors 1,4 addition. For most sulfur-stabilized lithium reagents, 2 equiv of HMPA suffice
to achieve >95% 1,4 addition, whereas 4 equiv of DMPU are required to achieve identical regiochemical and
stereochemical results.
Introduction
Hexamethylphosphoramide (HMPA) is a highly polar, aprotic
solvent that has found extensive use as a solvent additive in
organolithium chemistry. Its usefulness stems from its ability
to coordinate to lithium very stronglysby one measurement
approximately 300 times more strongly than tetrahydrofuran
(THF).1a HMPA is frequently used to accelerate organolithium
reactions, but far more intriguing are the instances where it has
been used to alter the course of a chemical reaction. For
example, it has been used to change the stereochemistry of
enolate formation,2a alkylation,2b and protonation,2c as well as
the regiochemistry of imine metalation,3a allyl anion alkylations,3b
and additions to R,â-unsaturated carbonyl compounds (Figure
1).1b,4,5 This last effect will be the topic of this paper.
Figure 1. A demonstration of the effect of HMPA on the kinetic
selectivity of enone addition.5
A mechanistic understanding of such results has been
hampered by a lack of information about the solution structures
of organolithium reagent/HMPA complexes, particularly the ion
pair structure. Two distinctions will be made: contact ion pairs
(CIP), in which there is a C-Li bond, and separated ion pairs
(SIP), in which the carbanion and lithium counterion are
separated by at least one layer of solvent molecules (yet remain
intimately associated based on chemical shift1c and reactivity6
effects). We have developed a low-temperature, multinuclear
NMR technique for determining the nature of ion-pairing,
aggregation, HMPA-solvation, and intramolecular chelation of
organolithium reagents in solution.1c
(1) (a) Reich, H. J.; Kulicke, K. J. J. Am. Chem. Soc. 1996, 118, 273-
274. (b) Preliminary communication: Reich, H. J.; Sikorski, W. H. J. Org.
Chem. 1999, 64, 14-15. (c) Reich, H. J.; Borst, J. P.; Dykstra, R. R.; Green,
D. P. J. Am. Chem. Soc. 1993, 115, 8728-8741. (d) Reich, H. J.; Borst, J.
P.; Dykstra, R. R. Tetrahedron 1994, 50, 5869-5880. (e) Reich, H. J.;
Green, D. P.; Phillips, N. H. J. Am. Chem. Soc. 1989, 111, 3444-3445. (f)
Reich, H. J.; Green, D. P. J. Am. Chem. Soc. 1989, 111, 8729-8731. (g)
Reich, H. J.; Borst, J. P. J. Am. Chem. Soc. 1991, 113, 1835-1837. (h)
Reich, H. J.; Sikorski, W. H.; Gudmundsson, B. O¨ .; Dykstra, R. R. J. Am.
Chem. Soc. 1998, 120, 4035-4036. (i) Sikorski, W. H.; Sanders, A. W.;
Reich, H. J. Magn. Reson. Chem. 1998, 36, S118-S124. (j) Reich, H. J.;
Green, D. P.; Medina, M. A.; Goldenberg, W. S.; Gudmundsson, B. O¨ .;
Dykstra, R. R.; Phillips, N. H. J. Am. Chem. Soc. 1998, 120, 7201-7210.
(k) Reich, H. J.; Gudmundsson, B. O¨ . J. Am. Chem. Soc. 1996, 118, 6074-
6075. (l) Reich, H. J.; Dykstra, R. R. Organometallics 1994, 13, 4578.
(2) (a) Ireland, R. E.; Mueller, R. H.; Willard, A. K. J. Am. Chem. Soc.
1976, 98, 2868-2877. (b) Ando, K.; Takemasa, Y.; Tomioka, K.; Koga,
K. Tetrahedron 1993, 49, 1579-1588. (c) Panek, E. J.; Rodgers, T. J. J.
Am. Chem. Soc. 1974, 96, 6921-6928.
Background
The addition chemistry to R,â-unsaturated carbonyl com-
pounds can only be of practical synthetic utility if one of the
two regioisomers is generated selectively. The synthetic im-
portance of these reactions is reflected by the extensive efforts
directed toward elucidating the effects that changes in solvent,
(3) (a) Hosomi, A.; Araki, Y.; Sakurai, H. J. Am. Chem. Soc. 1982, 104,
2081-2083. (b) Wolf, G.; Wu¨rthwein, E.-U. Tetrahedron Lett. 1988, 29,
3647-3650.
(4) Hunt, D. A. Org. Prep. Proced. Int. 1989, 21, 705-749.
(5) Brown, C. A.; Yamaichi, A. J. Chem. Soc., Chem. Commun. 1979,
100-101.
(6) (a) Hogen-Esch, T. E.; Smid, J. J. Am. Chem. Soc. 1966, 88, 307-
318. (b) Smid, J. Angew. Chem., Int. Ed. Engl. 1972, 11, 112-127.(c) Smid,
J. In Ions and Ion Pairs in Organic Reactions; Szwarc, J., Ed.; Wiley: New
York, 1972; Chapter 3.
10.1021/ja010053w CCC: $20.00 © 2001 American Chemical Society
Published on Web 06/13/2001

6528 J. Am. Chem. Soc., Vol. 123, No. 27, 2001
Sikorski and Reich
temperature, and steric bulk as well as main group and transition
metal additives have on the regioselectivity of addition of
organolithium reagents to enones and enals.2,7-15
O-Li bond is necessary to compensate for the rupture of the
C-Li bond (the 1,4:1,2 product ratio was shown to be
concentration independent, ruling out termolecular reactions
which could provide a mechanism for the CIP to add 1,4). This
hypothesis could not be directly tested since no information
about the solution structures of organolithium reagent/HMPA
complexes was available.
One plausible explanation for the propensity of SIP’s to add
1,4 is that the higher HOMO of the SIP vs the CIP promotes a
single electron transfer (SET) reaction, leading to conjugate
addition. However, a mechanistic study using radical probes
failed to demonstrate that trapable radicals were involved in
the reactions of lithiodithianes.13
Other explanations have been proposed to account for the
effect of HMPA in promoting 1,4 addition. Ab initio calculations
indicate that the carbonyl coefficient of the enone LUMO is
increased relative to that of the remote carbon when lithium
coordinates to the carbonyl, and hence coordination should
promote 1,2 addition. It is suggested, therefore, that HMPA
coordination renders the lithium less electrophilic, making
carbonyl activation insignificant and leading to more 1,4
addition.12
In the research that follows, we used our NMR technique to
test Cohen’s proposal.11 We were looking for a correlation
between the amount of CIP and SIP present in solution (which
we are in a position to determine quantitatively) and the amount
of 1,2 and 1,4 addition produced. We limited this study to
stabilized organolithium reagents which are known or expected
to be monomeric in THF, to simplify the analysis.
The first demonstration that the 1,4:1,2 ratio could in some
cases be controlled by proper choice of temperature or solvent
appeared in 1977,7 when it was found that the kinetic preference
for 1,2 addition by metalated 2-phenyl-1,3-dithiane could be
overcome by performing the reaction at higher temperatures and
in more polar solvent, leading to the thermodynamically favored
1,4 product. A year later, for a reaction determined to be entirely
under kinetic control, the opposite temperature dependence was
observed: lower temperatures favored the 1,4 product and higher
temperatures favored the 1,2 product.8 No explanation was
offered.
In 1977, Dolak and Bryson observed that isopropyl(phen-
ylthio)methyllithium underwent 1,4 addition to cyclohexenone.9
They attributed this unexpected regioselectivity to the presence
of HMPA, which was needed to perform the difficult metalation
of isobutyl phenyl sulfide. In a footnote they theorized that the
HMPA “may be forming solvent separated ion pairs that allow
the lithium ion to complex with oxygen while the sulfur bearing
carbon adds to the terminus of the conjugated system.”9 Seyden-
Penne and Krief subsequently showed that in the absence of
HMPA this lithium reagent (generated by Li-Se exchange) gave
exclusively 1,2 addition in THF, but confirmed that when
80:20 THF/HMPA was used, only 1,4 addition was observed.10a
More importantly, they demonstrated that this was an HMPA-
promoted kinetic 1,4 addition: when the HMPA was added
subsequently to the addition reaction, no equilibration occurred.
As little as 1 or 2 equiv of HMPA can have quite a dramatic
effect on the kinetic regioselectivity, as exemplified in Figure
1 in which an essentially complete reversal of regioselectivity
was observed by Brown and Yamaichi for the addition of
2-lithio-1,3-dithiane (1a-Li) and several substituted dithianes
(including 1b-Li) to cyclohexenone (Figure 1).5
Results and Discussion
Solution Structure of Lithiodithianes. We began by inves-
tigating the synthetically useful lithiodithianes,16 starting with
the 2-methyl derivative 1b-Li. By titrating small amounts of
HMPA to a solution of the lithium reagent and studying the
solution by NMR spectroscopy at -135 °C, we can detect the
individual HMPA-solvated lithium species, determine the
number of HMPA molecules coordinated to the lithium, and
detect when the transition from a contact to a separated ion
occurs. The method for making these assignments has been
described elsewhere.1d-g
We have not experimentally addressed the aggregation states
of the lithiodithianes, but previous work has shown that both
1a-Li and 1b-Li17d as well as the 2-tert-butyl analog1d in THF
show C-Li coupling to only a single lithium in the 13C NMR
spectra, so the species either are monomeric or are dimerized
by S-Li bridging, as found for the crystals of the 1b-Li‚
TMEDA complex grown from hydrocarbon solvent.17e Cryo-
scopic measurements in THF confirm that 1a-Li is monomeric
in solution,17a and we presume that 1b-Li is also monomeric,
since higher levels of alkyl substitution at a carbanion center
invariably lead to lower levels of aggregation. The HMPA
titration technique1c can also detect aggregation, since dimers
and higher aggregates usually show a more complex sequence
of HMPA solvates in the 6/7Li and 31P NMR spectra than do
In 1987, Cohen extended the separated ion pair hypothesis
of Dolak and Bryson to explain temperature and solvent effects
for the addition of bis(phenylthio)methyllithium to 2-cyclohex-
enone.11 Lower temperatures and the addition of HMPA favored
1,4 addition, conditions also known to favor separated ion pairs.6
Cohen rationalized that separated ion pairs react to give
predominantly 1,4 addition, while contact ion pairs, which are
favored at higher temperatures, undergo 1,2 addition via a four-
centered transition state, since simultaneous formation of the
(7) Ostrowski, P. C.; Kane, V. V. Tetrahedron Lett. 1977, 3549-3552.
(8) Ogura, K.; Yamashita, M.; Tsuchihashi, G.-i. Tetrahedron Lett. 1978,
1303-1306.
(9) Dolak, T. M.; Bryson, T. A. Tetrahedron Lett. 1977, 1961-1964.
(10) (a) Wartski, L.; El Bouz, M.; Seyden-Penne, J.; Dumont, W.; Krief,
A. Tetrahedron Lett. 1979, 1543-1546. (b) Lucchetti, J.; Krief, A. J.
Organomet. Chem. 1980, 194, C49-C52. (c) Krief, A. Tetrahedron 1980,
36, 2531-2640. (d) Dumont, W.; Lucchetti, J.; Krief, A. J. Chem. Soc.,
Chem. Commun. 1983, 66-68.
(11) Cohen, T.; Abraham, W. D.; Myers, M. J. Am. Chem. Soc. 1987,
109, 7923-7924.
(12) Lefour, J.-M.; Loupy, A. Tetrahedron 1978, 34, 2597-2605.
(13) Chung, S. K.; Dunn, L. B., Jr. J. Org. Chem. 1984, 49, 935-939.
Juaristi, E.; Gordillo, B.; Aparicio, D. M. Tetrahedron Lett. 1985, 26, 1927-
1930.
(14) (a) Ager, D. J.; East, M. B. J. Org. Chem. 1986, 51, 3983-3992.
(b) Binns, M. R.; Haynes, R. K.; Katsifis, A. G.; Schober, P. A.; Vonwiller,
S. C. J. Org. Chem. 1989, 54, 1960-1968. (c) Hirama, M. Tetrahedron.
Lett. 1981, 22, 1905-1908. (d) Wilson, S. R.; Price, M. F. Synth. Commun.
1982, 12, 657-663. (e) Still, W. C.; Mitra, A. Tetrahedron Lett. 1978,
2659-2662. (f) Loupy, A.; Lefour, J.-M.; Deschamps, B.; Seyden-Penne,
J. NouV. J. Chim. 1980, 4, 121-126. (g) Seebach, D.; Locher, R. Angew.
Chem., Int. Ed. Engl. 1979, 18, 957-958.
(16) Corey, E. J.; Seebach, D. Angew. Chem., Int. Ed. Engl. 1965, 4,
1075-1077.
(17) (a) Bauer, W.; Seebach, D. HelV. Chim. Acta 1984, 67, 1972-1988.
(b) Ahlbrecht, H.; Harbach, J.; Hoffman, R. W.; Ruhland, T. Liebigs Ann.
1995, 211-216. (c) Ruhland, T.; Hoffman, R. W.; Shade, S.; Boche, G.
Chem. Ber. 1995, 128, 551-556. (d) Seebach, D.; Gabriel, J.; Ha¨ssig, R.
HelV. Chim. Acta 1984, 67, 1083-1099. (e) Amstutz, R.; Seebach, D.;
Seiler, P.; Schweizer, W. B.; Dunitz, J. D. Angew. Chem., Int. Ed. Engl.
1980, 19, 53-54. (f) Zarges, W.; Marsch, M.; Harms, K.; Koch, W.;
Frenking, G.; Boche, G. Chem. Ber. 1991, 124, 543-549. (g) West, P.;
Waack, R. J. Am. Chem. Soc. 1967, 89, 4395-4399.
(15) (a) Maruoka, K.; Shimada, I.; Imoto, H.; Yamamoto, H. Synlett 1994,
519-520. (b) Maruoka, K.; Ito, M.; Yamamoto, H. J. Am. Chem. Soc. 1995,
117, 9091-9092.

Addition of Organolithium Reagents to Enones and Enals
J. Am. Chem. Soc., Vol. 123, No. 27, 2001 6529
Figure 3. Graph of the percentage of each HMPA-containing ion pair
species for 1b-Li (dotted lines) overlaid with the percentage of 1,4
addition of 2-methyl-1,3-dithianyllithium (1b-Li) toward 2-cyclohex-
enone (2) and trans-2-hexenal (5) as functions of HMPA for 0.16 M
1b-Li in 3:2 THF/Et2O at -125 °C (solid lines).
Figure 2. Graph of the percentage of each ion pair species as a function
of HMPA for 0.16 M 2-methyl-1,3-dithianyllithium (1b-Li) in 3:2 THF/
Et2O at -135 °C.
monomers. Thus, mono and bis HMPA complexes of the dimers
of PhLi,1j 2-(dimethylaminomethyl)phenyllithium,1k MeSLi,1c
LiCl,1c LiPPh2,1l LiNiPr2,18a and lithium bis(trimethylsilyl)-
amide18b have been observed, in addition to one or more
monomer HMPA complexes. Studies of the effect of HMPA
on the parent 2-lithio-1,3-dithiane 1a-Li, as well as on the
2-silyl-, 2-phenyl-, and 2-tert-butyl-substituted lithiodithianes,
revealed an effect of HMPA on ion pairing,1d but gave no
indication of the presence of dimers.
overlaid with the growth of each HMPA-containing lithium
species. The graph demonstrates that there is no direct correla-
tion between the amount of separated ion S4 and the amount of
1,4 addition produced in the reactions with either 2 or 5.
Therefore, SIPs can only be responsible for the 1,4 product
if two conditions are met: the SIPs are exceedingly more
reactive than the CIPs, and the SIPs and CIPs are in rapid
equilibrium. This is a manifestation of the Curtin-Hammett
principle,20 which states that if two reactive species interconvert
more readily than they react, then the product ratio will not
reflect the ground-state distribution (which is what is measured
by the NMR experiment), but, rather, it will be a function of
the energy difference between the product-forming transition
states for the two reactive species. In other words, although the
CIPs are the ground-state structures (the NMR results prove
this), the product distribution could be dominated by more
reactive SIPs in rapid equilibration with the CIPs, even though
their equilibrium concentrations might be negligible.
The results from the NMR study are shown graphically in
Figure 2, where the relative amounts of each of the species based
7
on curve fitting of the Li spectra19 are plotted. In a notation
that will be used throughout this paper, the various organo-
lithium species will be designated as Cn or Sn, where C and S
denote contact ion pair and solvent-separated ion pair, respec-
tively, and the superscript indicates the number of HMPA
molecules coordinated to the lithium. The addition of HMPA
converts the C0 species first to C1, then to C2, followed by a
jump in coordination to the separated ion, S4 (this is, in effect,
a disproportionation of C3 to a C2 and an S4). Only the first
molecule of HMPA coordinates quantitatively; subsequent
coordinations occur with increasing difficulty (i.e., C2 maximizes
at 2.75 equiv of HMPA), probably due to both steric and
electronic effects. Free HMPA is detectable in the solution (by
31P NMR) as early as 1.0 equiv of HMPA, which is character-
istic for strongly contact ions (separated ions show no free
HMPA until >3 equiv of HMPA). A very important result for
the purposes of this study is that the separated ion, S4, is not
detectable until after 1.5 equiv of HMPA, and it does not become
the dominant species until after 4.25 equiv of HMPA.
Regioselectivity of Lithiodithiane Additions. With this
detailed knowledge of which species are present in solution,
we looked for a correlation with the regiochemistry of additions
of 1b-Li to 2-cyclohexenone (2) and trans-2-hexenal (5).10b The
reactions were performed in the same solvent and at comparable
temperatures as the NMR studies. Refer to the Experimental
Section for a description of how we addressed the technical
problem of adding the electrophile to the lithium reagent at these
very low temperatures while maintaining good temperature
control and efficient mixing.
Even before the fully HMPA-solvated S4 species becomes
detectable, SIPs with lower HMPA coordination numbers will
exist in equilibrium with their CIP counterparts (Scheme 1).
As the HMPA coordination number of the CIP increases, the
equilibrium constant favoring the SIP increases (i.e., K0
<
eq
K1eq < K2eq) and the rate at which the CIP equilibrates with its
SIP probably also increases (i.e., kS0 < kS1 < kS2). Thus the
apparent correlation of the 1,4-addition product of cyclohex-
enone with the fraction of C1 in Figure 2 could actually be a
correlation with the undetected species S1 in equilibrium with
it, if K1eq‚kS1 . kC112. Under this analysis, C0 and S1 are the
14
principal 1,2- and 1,4-forming species, respectively, and the
The results of the reactivity studies of 1b-Li with 2 (to give
3b and 4b) and 5 (to give 6 and 7) are shown in Figure 3,
product distribution reflects the C0/C1 equilibrium.
Alternatively, the rate of formation of the SIP (kS0 and/or
kS1) could be rate determining for the 1,4 addition, since the
ion pair separation is substantially endothermic when little or
no HMPA is present. With no HMPA present, kC012 > kS0 and
the 1,2 product predominates, whereas when HMPA is present,
(18) (a) Romesberg, F. E.; Gilchrist, J. H.; Harrison, A. T.; Fuller, D. J.;
Collum, D. B. J. Am. Chem. Soc. 1991, 113, 5751-5757. (b) Romesberg,
F. E.; Bernstein, M. P.; Gilchrist, J. H.; Harrison, A. T.; Fuller, D. J.; Collum,
D. B. J. Am. Chem. Soc. 1993, 115, 3475-3483. (c) Galiano-Roth, A. S.;
Collum, D. B. J. Am. Chem. Soc. 1988, 110, 3546-3553.
(19) Reich, H. J. WinDNMR: Dynamic NMR Spectra for Windows. In
J. Chem. Educ.: Software, Ser. D 1996, 3D, No. 2.
(20) Curtin, D. Y. Rec. Chem. Prog. 1954, 15, 111-128.

6530 J. Am. Chem. Soc., Vol. 123, No. 27, 2001
Sikorski and Reich
Scheme 1. Expected Rates and Equilibria for 1b-Li in
THF/HMPA Solution
Figure 4. Graph of 1,4 addition and 1,2 diastereomer distribution as
functions of HMPA for the reaction of 1,3-dithianyllithium (1a-Li) with
5-trimethylsilyl-2-cyclohexenone (8) at -120 °C in 3:2 THF/Et2O and
in 3:2 Me2THF/Et2O.
titration of 1a-Li has been previously reported,1d and the profile
is very comparable to that of 1b-Li (Figure 2) and will not be
reproduced here. To test the proposal that small amounts of
SIPs in equilibrium with the ground-state CIPs were producing
the 1,4 product, we also performed a series of experiments in
which we substituted for THF the less polar analogue 2,5-
dimethyltetrahydrofuran, in the hopes of disfavoring the SIPs.
The results of the experiments in both solvent systems are shown
in Figure 4.
kC1 < kS1 and the 1,4 product predominates. In this case one
12
can make the prediction that by switching to a sufficiently
reactive substrate, one could tip the balance in favor of kC1
,
12
since kS1 is independent of substrate and kC1 would now be
12
larger. We have tested this hypothesis with the enal 5, expected
to be 3 orders of magnitude more reactive than the enone.21 In
Figure 3, the growth of the 1,4 addition curve for the enal no
longer follows C1, but rather parallels the growth of the C2
species remarkably well. This suggests that when C2 is present,
the enal is reacting with the S2 that is in equilibrium with it
(and whose concentration is always directly proportional to it).
With less HMPA present, in the absence of C2, the enal reacts
with C0 or C1, i.e., the rates of conversion of these CIPs to
their SIPs (kS0 and kS1) are too slow to permit significant reaction
between the enal and S0 or S1. Note that in this system, the SIP
produces a 60:40 ratio of 1,4:1,2 addition rather than pure 1,4
addition.
The rates of all of the reactions discussed in this paper are
too fast to measure by the techniques and analytical methods
used (>90% complete in <1 min at -120 °C), so a more direct
test of the proposed explanation is not possible. We therefore
turned to stereochemical probes to provide further insight into
the reactive intermediates involved.
Stereochemistry of Additions. We used the chiral enone
5-trimethylsilyl-2-cyclohexen-1-one (8)22 with 1a and the chiral
lithium reagent phenylthio(3-methyl)benzyllithium (9-Li) in the
hope that the CIPs and SIPs would exhibit unique signatures
or fingerprints in the form of diastereomeric product ratios which
would allow their participation in addition reactions to be
tracked. We switched to the parent dithianyllithium 1a-Li from
the methyl homologue 1b-Li used above because the bis(sulfur)-
substituted carbon has an easily observable proton for NMR
analysis of the diastereomeric product mixtures. The HMPA
The diastereoselectivities for this system to form the adducts
10 and 11 were constant; however, they were too large23 (and
therefore had too little dynamic range available) to use them as
evidence that a CIP was solely responsible for the 1,2 addition
and a SIP was solely responsible for the 1,4 addition. For
simplicity, only the 1,2 diastereomer distributions are shown
(the selectivity was even greater for the 1,4 products 11, more
than 95:5 throughout in favor of the trans product of axial
attack). The most significant aspect of this graph is the phase
shift of the total 1,4 addition curves between the two solvent
systems. The change in solvent polarity and donor strength is
not likely to appreciably change the relative ratios of C0, C1,
and C2 over the range of 0 to 1.5 equiv of HMPA for this
strongly contact ion, because they are formed essentially
quantitatively at these HMPA levels even in THF. What the
change in solvent is likely to do is affect the stabilization of
any small amounts of separated ions that are in equilibrium with
these contact ions, through primary and secondary solvent shells.
Hence, the sharp rise in the total 1,4 addition occurs at a later
point in the less polar solvent system.
(21) Pentanal is 2700 times as reactive as cyclohexanone toward
pinacolone lithium enolate in THF (∆∆Gq )3.1 kcal/mol): Das, G.;
Thornton, E. R. J. Am. Chem. Soc. 1993, 115, 1302-1312. Benzaldehyde
is >3000 times more reactive than benzophenone toward methylmagnesium
bromide: Holm, T. J. Org. Chem. 2000, 65, 1188-1192.
(22) Asaoka, M.; Shima, K.; Takei, H. Tetrahedron Lett. 1987, 28, 5669-
5672.
We obtained our most dramatic stereochemical results using
the chiral organolithium reagent phenylthio(3-methyl)benzyl-

Addition of Organolithium Reagents to Enones and Enals
J. Am. Chem. Soc., Vol. 123, No. 27, 2001 6531
Figure 6. Graphs of total 1,4 addition and 1,4 diastereomer distribution
as functions of THF in ether (left) and HMPA in 3:2 THF/ether (right)
for the reaction of 9-Li with 8 at -120 °C.
Scheme 2. Mechanistic Proposal
Figure 5. 7Li and 31P NMR spectra from an HMPA titration of 0.16
M phenylthio(3-methyl)benzyllithium (9-Li) in 3:2 THF/Et2O at -129
°C.
lithium (9-Li). Benzyllithium itself is monomeric in THF,17g
and R-phenylthiobenzyllithium crystallizes as a monomeric tris-
solvate from THF.17f Other benzyllithium reagents such as
2-phenyl-1,3-dithiane, R-methylsulfonylbenzyllithium, and R-di-
methylaminobenzyllithium are also monomeric in THF solution.17b
We therefore assume that 9-Li is monomeric.
Compound 9-Li is unambiguously a contact ion in THF, but
a very weak one. This is shown by the HMPA titration (Figure
5), in which the addition of HMPA converts C0 directly to S1.
Consistent with this, a substantial fraction of 1,4 addition occurs
even without the addition of HMPA. It was necessary to reduce
solvent polarity to pure ether before 1,4 addition was fully
suppressed.
Reactivity studies were performed in Et2O with incremental
amounts of THF added, and also in 3:2 THF/Et2O (which
corresponds to 90 equiv of THF) with incremental amounts of
HMPA added. Reaction of 9-Li with the chiral enone 8 produced
four diastereomers of both the 1,2 and 1,4 products 12 and 13.
The diastereomer distribution of the 1,4 products will be
considered first (Figure 6).
The most significant result from Figure 6 is that the 1,4
diastereomer ratio was essentially invariant as THF was added
to an Et2O solution, despite a change in the total amount of 1,4
addition from 0 to 65%. The constant diastereomer ratio is
consistent with the 1,4 products being produced by the same
species throughout. In the absence of THF, only 1,2 addition
was observed, and we propose that a CIP (Mechanism A in
Scheme 2) is the only reactive species. The addition of THF to
the Et2O solution increases the solvent strength enough to
stabilize small amounts of SIP (S0), which are solely responsible
for the 1,4 product under THF/Et2O conditions (Mechanism B).
We have checked the effect of concentration of 9-Li on the ratio
of 1,2 to 1,4 addition. 1,4-Addition to 2-cyclohexenone is
constant within experimental error over a 12-fold change in
concentration (61% 1,4 product at 0.02 M, 60% at 0.08 M, 64%
at 0.25 M), and the ratio of addition to 8 is constant over a
4-fold range of concentrations (41% 1,4 product at 0.02 M, 39%
at 0.08 M), as also found previously for bis(phenylthio)-
methyllithium.11 This rules out a difference in molecularity (i.e.,
from bimolecular to unimolecular in RLi) as the origin of the
changes in the ratio of 1,2 and 1,4 additions which we have
attributed to a switch from Mechanism A to Mechanism B.
The addition of HMPA causes a dramatic change in the 1,4
diastereomer distribution as well as the total 1,4 addition. We
propose that the change in diastereomer ratios signals the onset
of an uncatalyzed pathway (Mechanism C, Scheme 2), which
strongly favors 1,4 addition (this is consistent with calculations
which predict that Lewis acid complexation of the carbonyl
enhances 1,2 addition12). The diastereoselectivity reaches a
limiting value at 1.0 equiv of HMPA, suggesting that with this
system an HMPA-coordinated lithium is a catalytically inactive
species. HMPA certainly does create more SIPs (this is clear
from the NMR experiment), but the increase in diastereoselec-
tivity from 30:28:26:16 to 59:30:5:6 is the result of a concomi-
tant destruction of lithium catalysis, and it is the combination
of these effects which accounts for the potent ability of HMPA
to promote 1,4 addition.
An alternative explanation might be that the dramatic effect
of HMPA is steric in nature, as for Yamamoto’s bulky aluminum
catalysts.15 We disfavor this rationalization, since Yamamoto’s
catalysts are strongly electrophilic, whereas an HMPA-solvated
lithium would be expected to be much less so. Further, the
diastereoselectivity change is very sensitive to only the first
equivalent of HMPA. If the effect were purely steric in nature,
continued changes might be expected as additional HMPA was
added. Yet, from 1 to 4 equiv of HMPA, the diastereoselectivity
remains essentially constant.24
More evidence against the steric explanation comes from a
study of the effect of alternative lithium-complexing agents such
(23) The high stereoselectivity in this system is surprising, since the
addition of 1a-Li to 4-tert-butylcyclohexanone gives a 60/40 ratio of axial
to equatorial attack: Juaristi, E.; Cruz-Sa´nchez, J. S.; Ramos-Morales, F.
R. J. Org. Chem. 1984, 49, 4912-4917.
(24) With large excesses of HMPA, the results become less reproducible
and there is a greater scattering of data. Apparently, in this highly polar
medium the reaction becomes sensitive to other, unknown variables which
are not being properly controlled in these experiments.

6532 J. Am. Chem. Soc., Vol. 123, No. 27, 2001
Sikorski and Reich
Figure 8. Graph of total 1,2 addition and 1,2 diastereomer distribution
as functions of THF for the reaction of phenylthio(3-methyl)benzyl-
lithium (9-Li) with 5-trimethylsilyl-2-cyclohexen-1-one (8) in Et2O at
-120 °C.
Figure 7. Graph of 1,4 addition and 1,4 diastereomer distribution as
functions of HMPA and DMPU for the reaction of 9-Li with 8 at -120
°C. Twice as much DMPU relative to HMPA is necessary for the curves
to overlap.
with C0, which is also capable of some 1,2 addition, but through
Mechanism B (Scheme 2), hence producing a different diaste-
reomer ratio. In this interpretation the 1,2 diastereomer distribu-
tion at 5 equiv of THF represents the diastereoselectivity of
the THF-solvated C0 (operating through Mechanism A). That
at 90 equiv of THF represents a significant influence of S0 in
producing the 1,2 product (through Mechanism B).
as N,N′-dimethylpropylene urea (DMPU), which has been
proposed as an alternative to the carcinogenic HMPA,25 and
[2.1.1]cryptand.26 When the series of experiments was repeated
using the sterically different DMPU, the same limiting diaste-
reomer ratio was reached (Figure 7). Note, however, that twice
as much DMPU relative to HMPA is needed27 to reach both
the same 1,4:1,2 ratio and the same diastereomer distribution.
This suggests an electronic rather than a steric effect.
Finally, essentially the same limiting diastereomer ratio was
observed with 1 equiv of [2.1.1]cryptand (53/11/5/31) as was
found for 1 equiv of HMPA (59/7/5/29) or 2 equiv of DMPU
(53/8/10/29). The barrier to intermolecular exchange of lithium
cations complexed to [2.1.1]cryptand has been measured in
several solvents, ∆Gq >19 kcal/mol.28 The reaction of 9-Li with
cyclohexenone has ∆Gq < 9.7 kcal/mol, so in the presence of
excess cryptand participation by free lithium is not possible
because lithium cannot escape at a fast enough rate (i.e., the
Curtin-Hammett principle does not apply here). However, a
cryptand-encapsulated lithium could conceivably still act as a
catalyst.18c,29
Summary
Curtin-Hammett limitations precluded us from proving a
correlation between the ratio of CIP and SIP and the amount of
1,2 and 1,4 addition. Unless a system can be found where the
lithium reagent reacts more readily than its CIPs and SIPs
interconvert, such a correlation will not be possible. However,
the large amount of circumstantial evidence that we have
collected indicates that the type of ion pair is an important factor,
but it cannot be the only one. In addition to inducing CIP to
SIP transitions, we believe that HMPA also affects the 1,4:1,2
ratio by decreasing or preventing lithium catalysis. We propose
that the absence of Li catalysis is an important factor in
achieving clean 1,4 addition. For well-stabilized anions in the
absence of HMPA, when lithium catalysis is possible and SIPs
are energetically accessible reactive intermediates, mixtures of
1,2 and 1,4 addition are observed.
Figure 8 shows the results for the 1,2 diastereomers. Only
the Et2O/THF experiments are shown because 1,2 addition
became negligible as soon as the first 0.25 equiv of HMPA
was added.
Many stabilized lithium reagents can be induced to cleanly
add either 1,4 or 1,2 to enones. The addition of HMPA or
DMPU favors 1,4 addition (there is rarely an advantage to the
use of more than 2 equiv of HMPA or 4 equiv of DMPU).
Conversely, the use of Et2O as a solvent leads in most cases to
clean 1,2 addition (although with strongly contact ions such as
dithianyllithium, exclusive 1,2 addition is achieved even in
THF).
There was a noticeable change with two of the diastereomers
upon adding 5 equiv of THF to the Et2O solution. This jump is
reproducible and very likely represents a change in the nature
of the C0 species that is generating the 1,2 product (from one
that is Et2O solvated to one that is THF solvated). With further
addition of THF, there was a mild drift in the diastereomer ratio.
This can be interpreted as a direct effect of the change in bulk
solvent polarity, or as an indirect effect resulting from the
stabilization of a small amount of an S0 species in equilibrium
Experimental Section
(25) Mukhopadhyay, T.; Seebach, D. HelV. Chim. Acta 1982, 65, 385-
391.
General. Tetrahydrofuran (THF) and Et2O were freshly distilled from
sodium benzophenone ketyl before use. Hexamethylphosphoramide
(HMPA) was distilled from CaH2 at reduced pressure and stored under
N2 over molecular sieves. Glassware was placed overnight in a 110 °C
oven or flame-dried before purging with N2 to remove moisture.
Common lithium reagents were titrated using n-propanol in THF with
1,10-phenanthroline as an indicator.30 Temperatures of -78 °C were
achieved with a dry ice/acetone bath and -120 °C with an N2(l)/pentane
bath. Melting and boiling points are not corrected.
(26) (a) Dietrich, B.; Lehn, J. M.; Sauvage, J. P. Tetrahedron Lett. 1969,
34, 2885-2888. (b) Dietrich, B.; Lehn, J. M.; Sauvage, J. P. Tetrahedron
Lett. 1969, 34, 2889-2892. (c) Lehn, J. M. Pure Appl. Chem. 1980, 52,
2303-2319.
(27) This is consistent with other studies on DMPU,25 and with several
NMR studies that we have performed which show that twice as much DMPU
is needed to reach the same SIP/CIP ratio in solution.
(28) Cahen, Y. M.; Dye, J. L.; Popov, A. I. J. Phys. Chem. 1975, 79,
1292-1295. Shamsipur, M.; Popov, A. I. J. Phys. Chem. 1986, 90, 5997-
5999.
(29) For the X-ray crystal structure of a [2.1.1]cryptand complex of
EtMg+ see: Squiller, E. P.; Whittle, R. R.; Richey, H. G., Jr. J. Am. Chem.
Soc. 1985, 107, 432-435.
(30) Watson, S. C.; Eastham, J. F. J. Organomet. Chem. 1967, 9, 165-
168.

Addition of Organolithium Reagents to Enones and Enals
J. Am. Chem. Soc., Vol. 123, No. 27, 2001 6533
Commercially available starting materials and reagents (Aldrich
Chemical Co.) include the following: R-chloro-m-xylene, 2-cyclohex-
enone (2), 1,3-dithiane (1a-H), hexamethylphosphoramide (HMPA),
2-methyl-1,3-dithiane (1b-H), thiophenol, trans-2-hexenal (5), and
triethylamine.
A syringe was used to transfer 4.0 mL (0.25 mmol) of the 1b-Li
stock solution to the flask containing the cannula, the necessary amount
of HMPA (1 equiv ) 43.5 µL) was added, and the flask was cooled to
-78 °C under positive N2 pressure.
A dried, N2-purged, 5 mL conical flask fitted with a septum was
charged with 0.6 mL of THF, 0.4 mL of Et2O, and 12.0 µL (11.9 mg,
0.124 mmol) of 2, and both flasks were cooled under positive N2
pressure to -120 °C using an N2(l)/pentane slurry. A N2 inlet needle
was introduced into the conical flask containing 2 and turned on
momentarily to build up pressure. The septum plug was removed from
the exposed end of the cannula, and the cannula was immediately
inserted into the conical flask, with the N2 pressure inside being
dissipated through the cannula, purging it of any liquids and preventing
a back flow. The N2 gas through the inlet needle was again turned on,
and the end of the cannula was submerged in the liquid of the conical
flask, transferring the solution to the lithium reagent over the course
of approximately 5 s (occasionally an experiment had to be abandoned
due to a plugged cannula).
Syntheses: 5-Trimethylsilyl-2-cyclohexen-1-one (8). This com-
pound was synthesized according to a literature procedure22 and purified
by distillation (60-63 °C at 1 mmHg [lit.22 65.5-67 °C at 2.0 mmHg]).
1
Density: 0.9020 g/mL. H NMR (300 MHz, CDCl3): δ -0.22 (CH3,
3
3
3
s, 9H), 1.18 (H5, dddd, J5-6a ) 14.9 Hz, J5-4a ) 11.6 Hz, J5-4e
)
)
3
2
3
4.8 Hz, J5-6e ) 3.9 Hz), 1.94 (H4a, ddt, J4a-4e ) 18.9 Hz, J4a-5
3
4
2
11.4 Hz, J4a-3 ) J4a-2 ) 2.6 Hz), 1.96 (H6a, dd, J6a-6e ) 16.6 Hz,
3J6a-5 ) 14.5 Hz), 2.08 (H4e, dtt, 2J4e-4a ) 19.1 Hz, 3J4e-3e ) 3J4e-5
)
4
4
2
5.2 Hz, J4e-2 ) J4e-6e ) 1.3 Hz), 2.17 (H6e, ddt, J6e-6a ) 16.4 Hz,
3J6e-5 ) 3.9 Hz, 4J6e-2 ) 4J6e-4e ) 1.1 Hz), 5.74 (H2, ddt, 3J2-3 ) 9.8
Hz, 4J2-4a ) 2.7 Hz, 4J2-4e ) 4J2-6e ) 1.1 Hz), 6.79 (H3, 3J3-2 ) 10.1
Hz, J3-4e ) 5.5 Hz, J3-4a ) 2.6 Hz). 13C NMR (75.4 MHz, CDCl3):
δ -3.91 (CH3), 22.98 (CH), 26.61 (CH2), 38.45 (CH2), 129.23 (CH),
151.20 (CH), 199.81 (C).
3
3
After 1 min at this temperature, 3 mL (0.8 mmol) of 0.27 M
propionic acid in Et2O (1:49 v/v) were added via syringe to quench
the unreacted lithium reagent. The solution was allowed to warm to
room temperature, the contents were transferred to a separatory funnel,
and 20 mL of 1:1 Et2O/hexane were added. The solution was washed
with 3 × 20 mL of H2O, dried over anhydrous MgSO4, and concentrated
by rotary evaporation.
The ratios of products were determined using the following 1H NMR
signals with C6D6 as the solvent (300 MHz spectrometer): starting
material, δ 3.81 (CH, q, J ) 7.0 Hz, 1H); 1,2 product (3b), δ 6.32
(dCH, dq, J ) 10.3, 1.8 Hz, 1H); and 1,4 product (4b), δ 2.96 (CH,
dq, J ) 12.9, 2.5 Hz, 1H). Although, in general, a relaxation delay
was not used during the NMR acquisitions, occasionally the reliability
of the integrations was checked by introducing delays of up to 20 s,
which, without exception, had absolutely no effect on either the ratio
of the two products or their ratios to the returned starting material.
3-Methylbenzyl Phenyl Sulfide (9-H). To a 250 mL round-bottom
flask with a stir bar were added 125 mL of MeCN, 5.32 g (37.8 mmol)
of R-chloro-m-xylene, and 4.18 g (38.0 mmol) of PhSH. With vigorous
stirring, 3.85 g (38 mmol) of Et3N was added dropwise. The solid Et3N‚
HCl was filtered using a fritted glass funnel, and the solid was rinsed
with an additional 50 mL of MeCN. The solution was concentrated by
rotary evaporation, taken up in 100 mL of 1:1 Et2O/hexane, and washed
with 2 × 50 mL of 10% NaOH, 50 mL of H2O, and 50 mL of brine.
The solution was dried over anhydrous MgSO4, the solvents were
removed by rotary evaporation, and the product was distilled (109-
110 °C at 0.02 mmHg), yielding 5.19 g (24.2 mmol, 64.1%) of a pale
yellow liquid. 1H NMR (300 MHz, CDCl3): δ 2.29 (CH3, s), 4.06 (CH2,
s), 7.00-7.33 (Ar, m, 9H). 13C NMR (75.4 MHz, CDCl3): δ 21.29
(CH3), 38.88 (CH2), 125.80 (CH), 126.16 (CH), 127.90 (CH), 128.31
(CH), 128.75 (CH), 129.52 (CH), 129.56 (CH), 136.57 (C), 137.18
(C), 138.07 (C). Mass Spec: M+ 214.0819 (calcd for C14H14S,
214.0816).
Trapping Studies. All of the enone trapping studies were performed
using the same general procedure, which is written out in detail for
the reaction between 2-methyl-1,3-dithianyllithium (1b-Li) and 2-cy-
clohexenone (2). Because of the potential problem that the product
alkoxides and enolates, which are good electron pair donors, could
displace HMPA from lithium, thereby changing the ratio of HMPA to
unreacted lithium reagent at latter stages in the reaction, generally 0.5
equiv of the electrophile was added and the remaining lithium reagent
was quenched with a propionic acid/Et2O solution, which does not
freeze under the reaction conditions.
The samples from experiments which strongly favored one or the
other isomer were pooled together, and the major component was
isolated by preparative TLC for characterization.
1-[2-(2-Methyl-1,3-dithianyl)]-2-cyclohexen-1-ol (3b: 1,2 product
of 1b-Li and 2). Rf (20% EtOAc/hexane): 0.32. 1H NMR (300 MHz,
CDCl3): δ 1.70-2.10 (m, 8H), 1.77 (CH3, s, 3H), 2.28 (OH, S, 1H),
2.88-2.98 (m, 4H), 5.97-6.05 (dCH, m, 1H), 6.10-6.17 (dCH, m,
1H). 13C NMR (75.4 MHz, CDCl3): δ 18.83 (CH2), 24.31 (CH3), 25.00
(CH2), 25.13 (CH2), 26.69 (CH2), 26.82 (CH2), 31.01 (CH2), 59.44
(S2C), 75.62 (COH), 128.35 (dCH), 132.84 (dCH). IR (neat): 3466
cm-1 (OH). MS: M+ 230.0793 (calcd for C11H18OS2, 230.07991).
2-Methyl-1,3-dithianyllithium (1b-Li) and 2-cyclohexenone (2):
HMPA dependence in 3:2 THF/Et2O at -120 °C. A stock solution of
1b-Li was prepared by dissolving 0.24 mL (0.27 g, 2.0 mmol) of 1b-H
in 18.6 mL of THF and 12.4 mL of Et2O, cooling to -78 °C, adding
0.95 mL (2.02 mmol) of 2.13 M n-BuLi in pentane, and storing
overnight at -20 °C (the resulting solution, 0.0626 M in 1b-Li, is good
for at least 1 week at -20 °C).
3-[2-(2-Methyl-1,3-dithianyl)]cyclohexanone (4b: 1,4 product of
1b-Li and 2). Rf (20% EtOAc/hexane): 0.41. H NMR (300 MHz,
1
CDCl3): δ 1.50-1.70 (m, 2H), 1.62 (OH, s, 1H), 1.90-2.03 (m, 2H),
2.10-2.45 (m, 6H), 2.77-2.90 (m, 5H). 13C NMR (75.4 MHz,
CDCl3): δ 24.40 (CH3), 24.84 (CH2), 25.06 (CH2), 26.07 (CH2), 26.21
(CH2), 26.29 (CH2), 41.09 (CH2), 43.08 (CH2), 46.27 (CH), 52.50 (S2C),
211.34 (CdO). IR (neat): 1708 cm-1 (CdO). MS: M+ 230.0804 (calcd
for C11H18OS2, 230.07991).
A 10 mL round-bottom flask was prepared in advance with one end
of a cannula coiled inside. The flask (with cannula and stir bar) was
2-Methyl-1,3-dithianyllithium (1b-Li) and trans-2-hexenal (5):
HMPA dependence in 3:2 THF/Et2O at -120 °C. The general procedure
was used with the following changes. A stock solution of 1b-Li was
prepared by dissolving 0.24 mL (0.27 g, 2.0 mmol) of 1b-H in 23.25
mL of THF and 15.50 mL of Et2O, cooling to -78 °C, adding 1.0 mL
(2.0 mmol) of 2.0 M n-BuLi in pentane, and storing overnight at -20
°C (the resulting solution, 0.052 M in 1b-Li, is stable at least 1 week
at -20 °C). A syringe was used to transfer 4.0 mL (0.21 mmol) of
1b-Li to the flask containing the cannula. A stock solution of 5 was
also prepared, with 0.23 mL (0.20 g, 2.0 mmol) of 5 in 5.86 mL of
THF and 3.91 mL of Et2O (the resulting solution is 0.20 M in 5). A
syringe was used to transfer 1 mL (0.2 mmol) of 5 to the conical flask.
For these experiments, 1 equiv of HMPA ) 34.8 µL. Product ratios
were determined using the following 1H NMR signals in CDCl3:
starting material, δ 4.13 (S-CH-S, q, J ) 7.0 Hz, 1H); 1,2 product
(6b), δ 4.47 (O-CH, dm, J ) 6.2 Hz, 1H); and 1,4 product (7b), δ
9.86 (CHO, t, J ) 2.6 Hz, 1H).
dried overnight in an oven. The external end of the cannula was poked
through the bottom of a septum, and the septum was passed along it
into position. An N2 inlet needle was introduced into the flask, purging
the flask and the cannula. When purging was complete, the free end of
the cannula was plugged by imbedding it into a thick piece of septum.

6534 J. Am. Chem. Soc., Vol. 123, No. 27, 2001
Sikorski and Reich
1-[2-(2-Methyl-1,3-dithianyl)]-2-hexen-1-ol (6b: 1,2 product of
1b-Li and 5). Rf (20% EtOAc/hexane): 0.41. H NMR (300 MHz,
lithium reagent was made fresh for each experiment, directly in the
flask containing the cannula at -78 °C, because the metalation is
instantaneous, using 1.8 mL of THF, 1.2 mL of Et2O, 49.0 µL (52.7
mg, 0.25 mmol) of 9-H, and 0.125 mL (0.25 mmol) of 2.0 M n-BuLi
in pentane. For these experiments, 1 equiv of HMPA ) 43.5 µL. The
lithium reagent was trapped with 46.6 µL (42.0 mg, 0.25 mmol) of 8.
Different NMR solvents had to be used to determine the diastereomer
ratios of the 1,2 and 1,4 products, due to coincident signals. The 1,4
product ratios were determined using 1H NMR signals in CDCl3:
starting material δ 4.04 (s, 2H); 1,4 diastereomers (13), δ 3.85 (d, J )
10.3 Hz, 1H), 3.95 (d, J ) 10.7 Hz, 1H), 4.00 (d, J ) 7.0 Hz, 1H),
4.09 (d, J ) 5.9 Hz, 1H); 1,2 diastereomers (12), δ 4.17 (two coincident
1
CDCl3): δ 0.91 (CH3, t, J ) 7.4 Hz, 3H), 1.39 (H,6 s, 3H), 1.44 (H,5
qt, J ) 7.3, 7.3 Hz, 2H), 1.80-1.96 (m, 1H), 2.02-2.15 (m, 3H), 2.65
(SCHeq, dm, J ) 14.2 Hz, 2H), 2.80 (OH, br s, 1H), 3.03 (SCHax, dddd,
J ) 14.3, 11.2, 6.1, 2.9 Hz, 2H), 4.47 (OCH, br d, J ) 6.1 Hz, 1H),
5.64 (CH2d, ddt, J ) 15.3, 6.1, 1.5 Hz, 1H), 5.85 (dCH3, dtd, J )
15.4, 6.8, 1.1 Hz). 13C NMR (75.4 MHz, CDCl3): δ 13.65 (CH3), 22.17
(CH2), 22.26 (CH3), 24.32 (CH2), 25.75 (CH2), 26.12 (CH2), 34.47
(CH2), 53.29 (S2C), 72.72 (COH), 125.67 (C2Hd), 134.91 (dC3H).
IR (neat): 3466 cm-1 (OH). MS: M+ 232.0963 (calcd for C11H20OS2,
232.0956).
1
signals), 4.25, 4.39. The 1,2 product ratios were determined using H
3-[2-(2-Methyl-1,3-dithianyl)]hexanal (7b: 1,4 product of 1b-Li
1
NMR signals in C6D6: starting material, δ 3.82 (s, 2H); 1,2 diastere-
omers (12), δ 4.21 (s, 1H), 4.29 (s, 1H), 4.38 (s, 1H), 4.44 (s, 1H); 1,4
diastereomers (13), δ 3.87 (d, J ) 10.9 Hz), 3.90 (two coincident
signals, d, J ) 6.6 Hz), 4.10 (d, J ) 10.5 Hz).
and 5). Rf (20% EtOAc/hexane): 0.56. H NMR (300 MHz, CDCl3):
δ 0.94 (CH3, t, J ) 7.2 Hz, 3H), 1.05-1.50 (m, 3H), 1.42 (CH3, s,
3H), 1.80-2.10 (m, 3H), 2.29 (ddd, J ) 17.1, 4.6, 1.3 Hz, 1H), 2.57
(dddd, J ) 14.3, 5.9, 3.3, 0.7 Hz, 1H), 2.68 (dddd, J ) 14.5, 6.1, 3.3,
0.7 Hz, 1H), 2.75-2.85 (m, 2H), 2.90-3.00 (m, 2H), 9.86 (CHO, t, J
) 1.3 Hz). 13C NMR (75.4 MHz, CDCl3): δ 14.12 (CH3), 21.82 (CH2),
23.54 (CH3), 24.55 (CH2), 26.18 (CH2), 26.34 (CH2), 33.31 (CH2), 39.65
(CH), 45.42 (CH2), 52.99 (C), 200.12 (CdO). IR (neat): 1719 cm-1
(CdO).
1,3-Dithianyllithium (1a-Li) and 5-trimethylsilyl-2-cyclohexen-
1-one (8): Effect of HMPA in 3:2 THF/Et2O at -120 °C. The general
procedure was used with the following changes: the lithium reagents
was made fresh for each experiment, directly in the flask containing
the cannula at -78 °C, using 1.8 mL of THF, 1.2 mL of Et2O, 60.2
mg (0.5 mmol) of 1a-H, 0.26 mL (0.52 mmol) of 2.02 M n-BuLi in
pentane, and warming to -20 °C for 2 h. For these experiments, 1
equiv of HMPA ) 87.0 µL. The lithium reagent was trapped with a
solution of 46.6 µL (42.0 mg, 0.25 mmol) of 8. Product ratios were
determined using the following 1H NMR signals in C6D6: starting
material, δ 3.35 (s, 2H); 1,2 diastereomers (10), δ 4.13 (s, 1H), 4.25
(s, 1H); 1,4 diastereomers (11), δ 3.73 (d, J ) 4.5 Hz, 1H), 3.83 (d, J
) 9.0 Hz, 1H).
1-[Phenylthio(3-methyl)benzyl]-5-(trimethylsilyl)-2-cyclohexen-
1-ol (12: 1,2 products of 9-Li and 8, four diastereomers A, B, C,
1
and D). Rf (20% EtOAc/hexane): 0.56. H NMR (300 MHz, C6D6):
δ -0.23 (SiMe3 A), -0.13 (SiMe3 D), -0.11 (SiMe3 C), -0.10 (SiMe3
B), 0.60-1.90 (m, 6H), 2.00-2.35 (m, 4H), 2.69 (B, dm, J ) 13.6
Hz), 4.21 (SCH D), 4.29 (SCH C), 4.38 (SCH B), 4.43 (SCH A), 5.39
(dCH B, dm, J ) 10.3 Hz), 5.56 (dCH B, ddd, J ) 10.1, 4.4, 2.8
Hz), 5.68 (dCH A, ddd, J ) 10.1, 4.4, 2.4 Hz), 5.78-5.94 (dCH, m,
1H), 6.33 (dCH D, dm, J ) 10.1 Hz), 6.46 (dCH A, dm, J ) 10.1
Hz), 6.78-6.94 (m, 4H), 7.00-7.10 (m, 1H), 7.28-7.40 (m, 4H). 13
C
NMR (75.4 MHz, C6D6, selected signals): δ -3.89 (SiMe3 A), -3.71
(SiMe3 D), -3.68 (SiMe3 C), 64.83 (SCH A), 65.51 (SCH B), 65.83
(SCH D), 66.93 (SCH C), 71.19 (COH D), 71.57 (COH C), 73.50
(COH A), 73.54 (COH B). IR (neat): 3456 (OH), 3541 cm-1 (OH).
MS: M+ 382.1781 (calcd for C23H30OSSi, 382.1787).
3-[Phenylthio(3-methyl)benzyl]-5-trimethylsilylcyclohexanone
(13: 1,4 products of 9-Li and 8, four diastereomers A, B, C and
1
D). Rf (20% EtOAc/hexane): 0.44. H NMR (300 MHz, CDCl3): δ
-0.22 (SiMe3 D), -0.07 (SiMe3 A), -0.02 (SiMe3 B or C), 0.01 (SiMe3
C or B), 0.85-1.63 (m, 2H), 1.75-3.06 (m, 9 H), 3.85 (SCH D, d, J
) 10.3 Hz), 3.95 (SCH C, d, J ) 10.7 Hz), 4.00 (SCH B, d, J ) 7.0
Hz), 4.09 (SCH A, d, J ) 5.9 Hz). 13C NMR (75.4 MHz, CDCl3,
selected signals): δ -3.83 (SiCH3 A), -3.77 (SiCH3 D), -3.46 (SiCH3
B or C), 56.93 (SCH B or C), 57.71 (SCH D), 59.31 (SCH A), 59.96
(SCH C or B), 211.45 (CdO A), 211.77 (CdO D). IR (neat): 1709
cm-1 (CdO). MS: M+ 382.1799 (calcd for C23H30OSSi, 382.1787).
Phenylthio(3-methyl)benzyllithium (9-Li) and 5-trimethylsilyl-
2-cyclohexenone (8): Effect of addition of THF to Et2O at -120 °C.
The previous procedure, used for the HMPA experiments with the
system, was adapted in the following way: 3 mL of Et2O were used
as the solvent for 9-H; 1 mL of Et2O was used as the solvent for 8;
and THF was used in place of HMPA (5 equiv ) 0.10 mL). The
metalation was performed at 0 °C for 10 min before cooling to -120
°C.
Phenylthio(3-methyl)benzyllithium (9-Li) and 5-trimethylsilyl-
2-cyclohexen-1-one (8): DMPU dependence in 3:2 THF/Et2O at -120
°C. The procedure for the HMPA experiments with this system was
modified by substituting DMPU for HMPA (1 equiv of DMPU ) 30.2
µL).
Phenylthio(3-methyl)benzyllithium (9-Li) and 5-trimethylsilyl-
2-cyclohexen-1-one (8): Effect of 1 equiv of [2.1.1]Cryptand in 3:2
THF/Et2O at -120 °C. The procedure for the HMPA experiments with
this system was modified by substituting [2.1.1]cryptand for HMPA
(1 equiv of [2.1.1]cryptand ) 65.7 µL).
1-[2-(1,3-Dithianyl)]-5-trimethylsilyl-2-cyclohexen-1-ol (10: 1,2
product of 1a-Li and 8, major diastereomer). Rf (20% EtOAc/
hexane): 0.31. 1H NMR (300 MHz, CDCl3): δ -0.03 (SiCH3, s, 9H),
0.99 (SiCH, s, 1H), 1.34 (dd, J ) 14.3, 13.2 Hz, 1H), 1.76-1.93 (m,
2H), 1.94-2.10 (m, 2H), 2.32 (OH, broad s, 1H), 2.37 (dm, J ) 13.2
Hz, 1H), 2.76-2.96 (m, 4H), 4.37 (S2CH, s, 1H), 5.80 (CHd, dm, J
) 9.9 Hz, 1H), 5.86 (CHd, ddd, J ) 10.1, 3.9, 2.2 Hz, 1H). 13C NMR
(75.4 MHz, CDCl3): δ 3.87 (SiCH3), 18.91 (SiCH), 25.55 (CH2), 25.69
(CH2), 30.13 (CH2), 30.33 (CH2), 34.42 (CH2), 57.93 (S2CH), 73.12
(COH), 130.63 (CHd), 131.27 (CHd). IR (neat): 3448 cm-1 (OH).
MS: M+ 288.1031 (calcd for C13H24OS2Si, 288.1038).
trans-3-[2-(1,3-Dithianyl)]-5-trimethylsilylcyclohexanone (11: 1,4
1
product of 1a-Li and 8). Rf (20% EtOAc/hexane): 0.33. H NMR
(500 MHz, CDCl3): δ 0.00 (SiCH3, s, 9H), 1.29 (tt, J ) 11.2, 4.4 Hz,
H5a), 1.77 (ddd, J ) 13.9, 10.9, 4.7 Hz, H4a), 1.85 (dtt, J ) 14, 9.6, 4.9
Hz, H4′), 2.10 (m, H4e, H4′), 2.17 (ddd, J ) 14.4, 11.2, 0.9 Hz, H6a),
2.34 (ddt, J ) 14.4, 4.9, 1.5 Hz, H6e), 2.42 (d p, J ) 8.6, 5.3 Hz, H3e),
2.49 (dd, J ) 14.2, 5.3 Hz, H2a), 2.74 (ddt, J ) 14.2, 5.9, 1.5 Hz, H2e),
2.80-2.88 (m, 4H, SCH2), 3.97 (d, J ) 8.6 Hz, S2CH). The minor cis
diastereomer had a signal at δ 4.09 (S2CH, d, J ) 4.6 Hz). The trans
stereochemistry for the major isomer was assigned on the basis of the
large H6a-H5a trans-diaxial coupling of 11.2 Hz (5-SiMe3 group
equatorial), vs the smaller H2a-H3e coupling of 5.3 Hz (3-dithianyl
group axial). Assignments were confirmed by a COSY spectrum. 13C
NMR (75.4 MHz, CDCl3): δ -3.42 (SiCH3), 21.46 (SiCH), 25.83
(CH2), 27.77 (CH2), 30.42 (CH2), 30.46 (CH2), 41.52 (CH), 41.61 (CH2),
43.83 (CH2), 50.83 (S2CH), 211.62 (CdO). IR (neat): 1709 cm-1
(CdO). MS: M+ 288.1045 (calcd for C13H24OS2Si, 288.1038).
1,3-Dithianyllithium (1a-Li) and 5-trimethylsilyl-2-cyclohexen-
1-one (8): HMPA dependence in 3:2 dimethyl-THF/Et2O at -120 °C.
The preceding experiment was repeated, substituting 2,5-dimethyl-
tetrahydrofuran for THF in both the lithium reagent and the enone
solutions.
NMR Spectroscopy. For the characterization of synthesized com-
1
pounds, H and 13C NMR spectra were acquired on a Bruker AC-300
spectrometer with CDCl3 as the solvent (unless otherwise stated) and
tetramethylsilane as the internal standard. All multinuclear NMR
experiments were performed in 10 mm NMR tubes using a wide-bore
AM-360 spectrometer at 139.962 MHz (7Li) or 145.784 MHz (31P).
7
The digital resolution was 0.51 Hz for Li and 0.61 Hz for 31P. For a
Phenylthio(3-methyl)benzyllithium (9-Li) and 5-trimethylsilyl-
2-cyclohexen-1-one (8): HMPA dependence in 3:2 THF/Et2O at -120
°C. The general procedure was used with the following changes: the
typical 0.15 M solution, excellent signal-to-noise ratios were obtained
after 32 transients for 7Li and 80 for 31P. 7Li and 31P spectra were
generally transformed with Gaussian multiplication, with the GB

Addition of Organolithium Reagents to Enones and Enals
J. Am. Chem. Soc., Vol. 123, No. 27, 2001 6535
parameter set to the fractional duration of the FID and the LB parameter
set to -(digital resolution)/GB.
pentane was added. The NMR tube was warmed to -20 °C for 5 h to
complete the metalation. A series of 7Li and 31P NMR spectra at -135
°C were taken during an HMPA titration (1 equiv of HMPA ) 93.8
µL).
The lithium reagent samples were prepared in 10 mm thin-walled
NMR tubes which were oven-dried, fitted with a septum (9 mm i.d.),
and N2 flushed. The outside top portion of the tube was lightly greased
to make a better seal for the septum, which was held securely in place
with Parafilm. Silicon grease was placed on the septum top to seal
punctures, and the tubes were stored at -78 °C until the experiment
was performed. Since nondeuterated solvents were used, the spectrom-
eter was run unlocked, and shimming was performed on the 13C FID
of carbon 3 of THF. Although the spectrometer was unlocked during
the acquisition, the field was generally very stable and only occasionally
did a spectrum have to be reacquired due to a field shift. When a
substance had to be added (HMPA, for example), the sample was
ejected and placed in a -78 °C bath. To get HMPA to dissolve, the
tube had to be repeatedly shaken, but each time was returned quickly
to the cold bath. Temperatures were measured using a thermocouple
submerged in a second NMR tube containing the same solvent mixture,
or with an internal 13C NMR chemical shift thermometer.1i
R-(Phenylthio)(3-methyl)benzyllithium (9-Li). To a dried, N2-
purged, 10 mm NMR tube fitted with a septum and maintained under
positive N2 pressure were added 103 µL (108 mg, 0.504 mmol) of
3-methylbenzyl phenyl sulfide (9-H), 1.8 mL of THF, and 1.2 mL of
Et2O. The NMR tube was cooled to -78 °C and 0.27 mL (0.513 mmol)
7
of 1.9 M n-BuLi in pentane was added. A series of Li and 31P NMR
spectra at -127 °C were taken during an HMPA titration (1 equiv of
HMPA ) 87.7 µL).
Acknowledgment. We thank the National Science Founda-
tion for financial support of this work and Procter and Gamble
for a fellowship for W.H.S.
Supporting Information Available: 7Li and 31P spectra of
an HMPA titration of 1b-Li, and 1H and 13C spectra of 8, 9-H,
and the products of the trapping experiments (PDF). This
material is available free of charge via the Internet at
http://pubs.acs.org.
2-Methyl-1,3-dithianyllithium (1b-Li). To a dried, N2-purged, 10
mm NMR tube fitted with a septum and maintained under positive N2
pressure were added 72.5 mg (0.540 mmol) of 2-methyl-1,3-dithiane
(1b-H), 2.1 mL of THF, and 1.4 mL of Et2O. The NMR tube was
cooled to -78 °C and 0.25 mL (0.540 mmol) of 2.13 M n-BuLi in
JA010053W