Reactions of SmI2 with olefins: Mechanism and complexation effect on chemoselectivity

Full text of DOI:10.1021/ja950937d

J. Am. Chem. Soc. 1996, 118, 261-262
261
Reactions of SmI2 with Olefins: Mechanism and
Complexation Effect on Chemoselectivity
Avihai Yacovan and Shmaryahu Hoz*
Department of Chemistry, Bar-Ilan UniVersity
Ramat Gan, Israel 52900
Itzhak Bilkis
Faculty of Agriculture, Department of Biochemistry
The Hebrew UniVersity, RehoVot, Israel
ReceiVed March 21, 1995
ReVised Manuscript ReceiVed July 17, 1995
Radical anions of activated olefins may serve as good models
for the transition state of Michael addition reactions.1,2 In our
search for a suitable reducing agent to effect the generation of
these radical anions3 our attention was drawn to SmI2, which is
well-known to have exceptional qualities as a single electron
transfer reductant.4,5 Our initial studies revealed, however, that
SmI2 displays an extremely interesting and multifaceted mecha-
nistic chemistry of which very little is well understood.6 We
believe therefore that an exploration of the mechanistic chem-
istry of SmI2 is justified on its own merits. Equation 1 outlines
the reactions studied.
Figure 1. Ratio of the products derived from MA and DP as a function
of the identity and concentration of the proton donor.
value. For MeOH and trifluoroacetic acid (TFA), at high
concentration, the product ratio approaches unity (Figure 1).
These results are consistent with a mechanism where the
protonation competes with the equilibration of the radical anions.
At the lower proton donor concentration range, the lifetime of
the radical anions is long enough to permit equilibration resulting
in a relatively high selectivity. As the proton donor concentra-
tion increases, the equilibration is suppressed, thus leading to a
lower selectivity. Consistent with the suggested mechanism,
the plateau for the much stronger acid TFA is achieved at a
concentration (0.25 M) much lower than that of MeOH (2.5
M).
Surprisingly, for PrOH and MeOD, proton donors having
kinetic acidity lower than MeOH, the plateau level is achieved
at a product ratio below unity (0.7 and 0.8, respectively). The
absence of selectivity at the plateau region for TFA and MeOH
indicates a fast and unselective reaction of SmI2 with the two
substrates.10 The lack of dependence of the selectivity on the
“external” concentration of the various proton donors at the
plateau regions, combined with the fact that for MeOD and
PrOH plateau is achieved below unity, suggests that protonation
occurs internally, probably within a triple complex such as
ROH‚SmI2‚MA•- (or ROH‚SmI2‚DP•-).11 The triple complexes
can either undergo equilibration (and gravitate to their thermo-
dynamic distribution ratio) or undergo an internal protonation
which will “lock” the product distribution. The height of the
plateau is determined by a competition between two processes:
internal protonation within the triple complex and equilibration
of the radical anions of the two substrates. In the case of MeOH
this protonation is fast enough to prohibit any appreciable
equilibration. Slowing down the protonation rate by using acids
with lower kinetic acidity (MeOD and PrOH) permits a certain
extent of equilibration. The level of the plateau is, hence,
determined by the relative rates of the two processes, internal
protonation and equilibration. The lower the kinetic acidity,
the lower will be the plateau level.
In a competition experiment, a mixture of MA and DP in
THF was reacted with SmI2 in the presence of a proton donor.7,8
The product ratio was found to depend on the concentration of
the latter. At low proton donor concentration, the selectivity is
relatively high and approaches the thermodynamic stability ratio
of the two radical anions.9 As the proton donor concentration
increases, the selectivity decreases, finally reaching a constant
(1) Hoz, S. Acc. Chem. Res. 1993, 26, 69.
(2) Gross, Z.; Hoz, S. J. Am. Chem. Soc. 1994, 116, 7489.
(3) The scope included Na in liquid ammonia, Na/naphthalenide,
Na2S2O4, electrochemistry (see also: Avaca, L. A.; Utely, J. H. P. J. Chem.
Soc., Perkin Trans. 1 1975, 971; J. Chem. Soc., Perkin Trans. 2 1975, 161);
all gave reduction of the double bond but were less amenable for detailed
studies.
(4) Namy, J. L.; Girard, P.; Kagan, H. B. New J. Chem. 1977, 1, 5.
(5) (a) Molander, G. A. Chem. ReV. 1992, 92, 26. (b) Molander, G. A.
In ComprehensiVe Organic Synthesis; Trost, B. M., Fleming, I., Eds.;
Pergamon: Oxford, 1991; Vol. 4, pp 251. (c) Soderquist, J. A. Aldrichimica
Acta 1991, 24, 15. (d) Kagan, H. B. New J. Chem. 1990, 14, 453. (e)
Molander, G. A. In Chemistry of the Carbon-Metal Bond; Hartley, F. R.,
Patai, S., Eds.; Wiley: New York, 1989; Vol. 5, p 319. (f) Kagan, H. B.;
Sasaki, M.; Collin, J. Pure Appl. Chem. 1988, 60, 1725. (g) Kagan, H. B.;
Namy, J. L. Tetrahedron 1986, 42, 6573. (h) Hasegawa, E.; Curran, D. P.
J. Org. Chem. 1993, 58, 5008 and references cited therein.
(6) In spite of the wealth of literature in the rapidly developing field of
SmI2, it has been studied primarily from a synthetic perspective. To the
best of our knowledge, the leading reference for mechanisms in SmI2
chemistry is ref 14 below.
As shown in Scheme 1, the triple complex may be formed
in a reaction between the alcohol and the SmI2‚substrate ion
(7) Unless otherwise indicated, the total concentration of the substrate-
(s) was in the range of 0.012 M (depending on the concentration of the
SmI2 as determined by titration) and that of SmI2 was 0.006 M. Two
equivalents of SmI2 is needed to fully reduce 1 equiv of the olefin.
(8) The reactions are very colorful. Immediately upon mixing of the
reactants, the blue color of SmI2 vanishes and a red color appears. This
color gives way to a yellow one (probably Sm3+) at a rate which depends
on the concentration and acidity of the proton donor. In the experiments
with TFA and DP the intermediate red color is not seen at all; in the presence
of 2.5 M MeOH it persists for less than 1 s. After the completion of the
reaction, the solutions were taken out of the glovebox, treated with CH2-
Cl2-water, and analyzed (reactants and products) by NMR and/or by HPLC.
The agreement between the two methods was better than 5%. Products
were stable with time.
(10) The E1/2 of SmI2 in acetonitrile is -1.62 V Vs SCE (Kolthoff, I.
M.; Coetzee, J. F. J. Am. Chem. Soc. 1957, 79, 1852). According to this
value there is a thermodynamic driving force of ca. 13 kcal/mol for the
electron transfer reaction. This value is sufficient for a diffusion-controlled
reaction. However, the intrinsic barrier due to the internal reorganization
of the ligation sphere may be too high to enable such a fast process (Eberson,
L. Electron Transfer in Organic Chemistry; Springer-Verlag: Berlin, 1987;
Chapter 4).
(11) (a) Throughout this paper we use the term radical anion although
other intermediates such as charge transfer complex or Grignard samarium
may be the actual chemical species involved. (b) On the basis of other
results it seems to us that, unlike the alcohols, TFA does not form complexes
with samarium.
(9) E1/2 -1.08 and -1.16 V Vs SCE for DP and MA respectively,
unpublished results.
0002-7863/96/1518-0261$12.00/0 © 1996 American Chemical Society

262 J. Am. Chem. Soc., Vol. 118, No. 1, 1996
Communications to the Editor
this complexation was demonstrated by using a variation of the
experimental procedure. In the “normal” addition mode, the
THF solution of SmI2 was mixed with a THF solution containing
the substrate and the proton donor. In the “inverse” addition
mode, the substrate solution was added to a mixture of SmI2
and the proton donor in THF.13 Since formation of the triple
complex MeOH‚SmI2‚MA•- (or DP•-) retards the equilibration
at the radical anion level, in an inverse addition experiment,
the plateau is expected to be reached at a lower concentration
of the alcohol. This was indeed observed, and the plateau was
achieved at ca. 0.8 M rather than at 2.5 M MeOH.
Thus, the overall reaction mechanism involves two paths: one
in which the free SmI2 reacts with the substrate prior to the
reaction with the alcohol; and a second in which the complex
SmI2‚ROH reacts with the substrate. The relatiVe contribution
of each of these reaction paths is goVerned by the addition mode
as well as by the reactants’ concentration.
Much to our surprise it was found that, in the inverse addition
mode, not only is the alcohol molecule bound firmly to SmI2,
but the isotopic identity of the labile proton is also largely
retained. This was demonstrated in the following experiments.
When SmI2 was added to a substrate solution containing equal
amounts of MeOH and MeOD (1.25 M each), the H/D
incorporation ratio into the benzylic position of DP was 1.2.
Under the same conditions but having the MeOH in the SmI2
solution and MeOD in the substrate solution, the incorporation
ratio increased to 7.15.
Figure 2. Variations of the UV spectrum of SmI2 as a function of the
concentration of the added MeOH. λmax: for SmI2, 616 and 552 nm;
for the MeOH complex, 584 nm.
Scheme 1
In his 1981 review Kagan stated14 that Sm in THF is bound
to several molecules of THF. It would seem that the bond of
Sm (especially the trivalent samarium) to alcohols (and their
labile proton) is stronger than one would expect.15 As we have
shown, this feature of Sm can be used to affect chemoselectivity
in reduction as well as in isotopic labeling reactions.
Acknowledgment. This study was supported by the Research
Authority of Bar-Ilan University.
JA950937D
(12) Calculations were based on a 1:1 complex composition. Correlation
coefficients are higher than 0.994.
(13) The alcohol was preincubated with the SmI2 usually for about a
minute, the time necessary to complete the operations in the glovebox.
(14) Kagan, H. B. Tetrahedron 1981, 37, Suppl. 1, 175.
(15) A referee had sugggested that the samarium may insert into the
O-H bond to give a hydride which does not easily exchange the proton.
We have tried to computationally explore this avenue. However, although
pseudopotentials are available for Sm, we found that, with programs such
as Gaussian and Gamess, polyatomic gradient optimization is not possible
with f-functions.
pair (e.g., SmI2‚MA•-) as well as in the reaction of a SmI2‚
ROH complex with the substrate. The ability of SmI2 to form
a complex with ROH was spectroscopically demonstrated.
Figure 2 shows the effect of added MeOH on the UV absorption
spectra of SmI2. Regression analysis of the data for MeOH
and PrOH gave the complex formation equilibrium constants12
of 0.56 and 0.22 M-1, respectively. The kinetic implication of