Enantioselective oxidation of di-tert-butyl disulfide with a vanadium catalyst: Progress toward mechanism elucidation

Article abstract of DOI:10.1021/jo0205560

The mechanism of the oxidation of di-tert-butyl disulfide (1) to the chiral thiosulfinate (2) by H₂O₂ catalyzed by bis(acetylacetonato)oxovanadium and a chiral Schiff-base ligand (3) has been investigated. Techniques included ⁵¹V NMR spectroscopy, solvent effects on reaction enantioselectivity, and the isolation and full characterization of a 2:1 ligand-to-vanadium catalyst precursor. A model for the dramatic solvent effect on the enantioselectivity of this reaction was developed, based on the identification of a competing nonselective oxidation pathway. From this model, strategies for limiting this competing pathway were developed.

Full text of DOI:10.1021/jo0205560

En a n tioselective Oxid a tion of Di-ter t-Bu tyl Disu lfid e w ith a
Va n a d iu m Ca ta lyst: P r ogr ess tow a r d Mech a n ism Elu cid a tion †
Suzanne A. Blum, Robert G. Bergman,* and J onathan A. Ellman*
Department of Chemistry and Center for New Directions in Organic Synthesis, University of California,
Berkeley, California 94720
bergman@cchem.berkeley.edu; ellman@cchem.berkeley.edu
Received August 21, 2002
The mechanism of the oxidation of di-tert-butyl disulfide (1) to the chiral thiosulfinate (2) by H2O2
catalyzed by bis(acetylacetonato)oxovanadium and a chiral Schiff-base ligand (3) has been
investigated. Techniques included 51V NMR spectroscopy, solvent effects on reaction enantioselec-
tivity, and the isolation and full characterization of a 2:1 ligand-to-vanadium catalyst precursor. A
model for the dramatic solvent effect on the enantioselectivity of this reaction was developed, based
on the identification of a competing nonselective oxidation pathway. From this model, strategies
for limiting this competing pathway were developed.
In tr od u ction
di-tert-butylsalicylaldehyde (eq 1).3,15 This reaction was
based on the Bolm protocol for oxidation of thioethers.16,17
Thiosulfinate 2 does not undergo further oxidation under
the reaction conditions. Furthermore, these conditions
can be efficiently used for mole-scale reactions. Thiosul-
finate 2 has proven to be a versatile intermediate for the
synthesis of tert-butanesulfinamides18-21 and tert-butyl
sulfoxides22,23 that have seen widespread use in organic
synthesis.
Chiral sulfinyl groups have been used as auxiliaries
in a variety of highly diastereoselective carbon-carbon
bond-forming reactions, including the syntheses of chiral
R-branched amines,1-4 R- and â-amino acids,5 aziridines,6-8
and aminophosphonic acids.9-11 Many of these transfor-
mations used arenesulfinyl auxiliaries because practical
methods for their preparation had been developed.12-14
In 1997, we reported the biphasic oxidation of di-tert-
butyl disulfide (1) to the corresponding chiral thiosulfi-
nate (2) (89% yield, 91% ee) with cost-effective hydrogen
peroxide (30% aq), employing catalytic amounts of bis-
(acetylacetonato)oxovanadium (VO(acac)2) and a chiral
Schiff-base ligand (3) derived from tert-leucinol and 3,5-
† This paper is dedicated to the memory of Prof. Henry Rapoport,
whose dedication to chemical research and education illuminated our
department for many years.
(1) Yang, T.-K.; Chen, R.-Y.; Lee, D.-S.; J iang, Y.-Z.; Mi, A.-Q.; J ong,
T.-T. J . Org. Chem. 1994, 59, 914-921.
(2) Moreau, P.; Essiz, M.; Me´rour, J .-Y.; Bouzard, D. Tetrahedron:
Asymmetry 1997, 8, 591-98.
(3) Liu, G.; Cogan, D.; Ellman, J . A. J . Am. Chem. Soc. 1997, 119,
9913-9914.
(4) Annunziata, R.; Cinquini, M.; Cozzi, F. J . Chem. Soc., Perkin
Trans. 1 1982, 341-343.
The mechanism of this oxidation is interesting from
both practical and fundamental standpoints. Mechanistic
(5) Davis, F. A.; Zhou, P.; Chen, B.-C. Chem. Soc. Rev. 1998, 27,
13-18.
(6) Davis, F. A.; Zhou, P.; Reddy, G. V. J . Org. Chem. 1994, 59,
3243-3245.
(15) Cogan, D.; Liu, G.; Kim, K.; Backes, B. J .; Ellman, J . A. J . Am.
Chem. Soc. 1998, 120, 8011-8019.
(7) Ruano, J . L. G.; Ferna´ndez, I.; Catalina, M. D.; Cruz, A. A.
Tetrahedron: Asymmetry 1996, 7, 3407-3414.
(8) Davis, F. A.; Reddy, G. V.; Liang, C.-H. Tetrahedron Lett. 1997,
38, 5139-5142.
(16) Bolm, C.; Bienewald, F. Angew. Chem., Int. Ed. Engl. 1995,
34, 2640-2642.
(17) Bolm, C.; Schlingloff, G.; Bienewald, F. J . Mol. Catal. A 1997,
117, 347-350.
(9) Mikolajczyk, M.; Lyzwa, P.; Drabowicz, J . Tetrahedron: Asym-
metry 1997, 8, 3991-3994.
(18) Owens, T. D.; Tang, T. P.; Ellman, J . A. Acc. Chem. Res. 2002,
35, 984-995.
(10) Lefebvre, I. M.; Evans, S. A., J r. J . Org. Chem. 1997, 62, 7532-
7533.
(19) Prakash, S. G. K.; Mandal, M.; Olah, G. A. Angew. Chem., Int.
Ed. 2001, 40, 589-590.
(11) Mikolajczyk, M.; Lyzwa, P.; Drabowicz, J .; Wieczorek, M. W.;
Blaszczyk, J . J . Chem. Soc., Chem Commun. 1996, 1503-1504.
(12) Andersen, K. K. In The Chemistry of Sulfones and Sulfoxides;
Patai, S., Rappoport, Z., Stirling, C. H., Eds.; J on Wiley & Sons: New
York, 1988; pp 56-94.
(13) Kagan, H. B.; Rebiere, F. Synlett 1990, 643-649.
(14) Solladie´, G. Synthesis 1981, 185-196.
(20) Cogan, D.; Ellman, J . A. J . Am. Chem. Soc. 1999, 121, 268-
269.
(21) Tang, T. P.; Ellman, J . A. J . Org. Chem. 1999, 64, 12-13.
(22) Adrio, J .; Carretero, J . C. J . Am. Chem. Soc. 1999, 121, 7411-
7412.
(23) Priego, J .; Carratero, J . C. Synlett 1999, 1603-1605.
10.1021/jo0205560 CCC: $25.00 © 2003 American Chemical Society
Published on Web 12/12/2002
150
J . Org. Chem. 2003, 68, 150-155

Enantioselective Oxidation of Di-tert-Butyl Disulfide
insight may aid in the design of an efficient industrial
scale-up of this reaction to the ton scale, since the current
biphasic conditions become inefficient on a multimole
scale. Additionally, insight into this system will contrib-
ute to a broader understanding of vanadium-catalyzed
oxidations employing O,N-donor ligands and stoichio-
metric peroxide oxidants. Other such processes include
epoxidation of allylic alcohols,24-29 development of bio-
mimetic analogues of vanadium bromoperoxidase,30-33
functionalization of alkanes,34,35 and asymmetric oxida-
tion of thioethers.36-44
Our progress toward elucidating the mechanism of the
vanadium-catalyzed enantioselective oxidation of 1 using
H2O2 and ligand 3 is discussed herein. In particular, we
discuss isolation and X-ray crystallographic characteriza-
tion of two catalyst precursors, and describe a significant
solvent effect in which biphasic conditions are necessary
for high enantioselectivity. Analysis of this effect led to
the discovery that slow addition of H2O2 under fully
soluble conditions results in dramatic increases in reac-
tion enantioselectivity.
F IGURE 1. Organic cosolvent solubility (g/L) with H2O vs
ee. All reactions were run with 0.025 equiv of VO(acac)2 and
3, 1 mmol 1 scale.
TABLE 1. In cr ea se in Rea ction En a n tioselectivities
u p on Slow H2O2 Ad d ition
fast addition
H2O2 ee (%)
slow addition
H2O2 ee (%)
solvent
CH3CNa
CF3CH2OHa
i-PrOHa
t-BuOHb
THFb
0
0
0
11
9
87
72
74
54
64
46
87
Resu lts a n d Discu ssion
E ffect of Solven t a n d Slow H 2O2 Ad d it ion on
En a n tioselectivity. The biphasic nature of the opti-
mized di-tert-butyl disulfide oxidation system complicates
mechanistic studies and industrial scale-up. For these
reasons, we investigated potential homogeneous reaction
conditions. We observed that all reactions in which the
organic cosolvent was fully miscible with the aqueous
H2O2 yielded racemic product. Furthermore, a screen of
b
CHCl3
a
b
Reaction is homogeneous. Reaction is biphasic.
organic cosolvents revealed an inverse relationship be-
tween the enantiopurity of 2 and the miscibility of that
cosolvent with water (Figure 1). In the absence of
vanadium catalyst, oxidation of 1 was slow under both
biphasic and homogeneous conditions, yielding less than
5% of 2 after 7 h. Enantioselectivities of the biphasic
systems were found to vary slightly ((10%) with interface
area (i.e., flask shape and reaction scale).
(24) Michaelson, R. C.; Palermo, R. E.; Sharpless, K. B. J . Am. Chem.
Soc. 1977, 99, 1990-1992.
(25) Hoshino, Y.; Yamamoto, H. J . Am. Chem. Soc. 2000, 122,
10452-10453.
(26) Murase, N.; Hoshino, Y.; Oishi, M.; Yamamoto, H. J . Am. Chem.
Soc. 1942, 64, 338-339.
(27) Hoshino, Y.; Murase, N.; Oishi, M.; Yamamoto, H. Bull. Chem.
Soc. J pn. 2000, 73, 1653-1658.
These results are consistent with the operation of a
nonstereoselective catalytic pathway at high H2O2 con-
centrations, perhaps involving displacement of all or part
of ligand 3 from the vanadium center by excess H2O2 to
form a nonselective oxidant. As this model predicts,
adding H2O2 slowly via syringe pump over 12 h under
fully miscible conditions resulted in dramatic increases
in reaction enantioselectivity to as high as 72% ee,
compared to 0% ee if the oxidant is added in one portion
(Table 1). No additional increase in enantioselectivity was
observed with longer addition times. Slow addition of
H2O2 into highly soluble but still biphasic cosolvents such
as THF and tert-butyl alcohol also resulted in large
increases in reaction enantioselectivities, whereas slow
addition into the sparingly soluble CHCl3 system did not
affect the reaction enantioselectivity. Modest increases
in enantioselectivity in the oxidation of thioethers under
biphasic conditions upon slow H2O2 addition have been
reported recently by Karpyshev and co-workers.36 Specif-
ically, portionwise addition of H2O2 into the biphasic CH2-
Cl2 system increased enantioselectivities from at most
29% (one portion) to 44% ee (slow addition). No homo-
geneous conditions, however, were reported. Homoge-
neous conditions that limit the nonselective pathway are
particularly attractive since they may provide an efficient
method for ton-scale synthesis. Because slow addition of
(28) Bolm, C.; Ku¨hn, T. Synlett 2000, 899-901.
(29) Traber, B.; J ung, Y.-G.; Park, T. K.; Hong, J .-I. Bull. Korean
Chem. Soc 2001, 22, 547-548.
(30) Colpas, G. J .; Hamstra, B. J .; Kampf, J . W.; Pecoraro, V. L. J .
Am. Chem. Soc. 1996, 118, 3469-3478.
(31) Clague, M. J .; Butler, A. J . Am. Chem. Soc. 1995, 117, 3475-
3484.
(32) Clague, M. J .; Keder, N. L.; Butler, A. Inorg. Chem. 1993, 32,
4754-4761.
(33) Bortolini, O.; Carraro, M.; Conte, V.; Moro, S. Eur. J . Inorg.
Chem. 1999, 1489-1495.
(34) Su¨ss-Fink, G.; Stanislas, S.; Shul’pin, G. B.; Nizova, G. V.;
Stoeckli-Evans, H.; Neels, A.; Bobillier, C.; Claude, S. J . Chem. Soc.,
Dalton Trans. 1999, 3169-3175.
(35) Su¨ss-Fink, G.; Stanislas, S.; Shul’pin, G. B.; Nizova, G. V. Appl.
Organomet. Chem 2000, 14, 623-628.
(36) Karpyshev, N. N.; Yakovleva, O. D.; Talsi, E. P.; Bryliakov, K.
P.; Tolstikova, O. V.; Tolstikov, A. G. J . Mol. Catal. A 2000, 157, 91-
95.
(37) Skarzewski, J .; Ostrycharz, E.; Siedlecka, R. Tetrahedron:
Asymmetry 1999, 10, 3457-3461.
(38) Bolm, C.; Bienewald, F. Synlett 1998, 1327-1328.
(39) Schmidt, H.; Bashirpoor, M.; Rehder, D. J . Chem. Soc., Dalton
Trans. 1996, 3865-3870.
(40) Nakajima, K.; Kojima, K.; Kojima, M.; Fujita, J . Bull. Chem.
Soc. J pn. 1990, 63, 2620-2630.
(41) Nakajima, K.; Kojima, M.; Toriumi, K.; Saito, K.; Fujita, J . Bull.
Chem. Soc. J pn. 1989, 62, 760-767.
(42) Ohta, C.; Shimizu, H.; Kondo, A.; Katsuki, T. Synlett 2002, 161-
163.
(43) Vetter, A. H.; Berkessel, A. Tetrahedron Lett. 1998, 39, 1741-
1744.
(44) Skarzewski, J .; Ostrycharz, E.; Siedlecka, R.; Zielinska-Blajet,
B.; Pisarski, J . J . Chem. Res. S 2001, 263-264.
J . Org. Chem, Vol. 68, No. 1, 2003 151

Blum et al.
biphasic and homogeneous conditions is a result of the
increase in H2O2 concentration and not an effect of
changing solvent.
We initially considered that some of the nine peaks
observed under low H2O2 concentrations were oxo-
peroxo complexes (4) with ligand 3 bound to the metal
center (Scheme 1). Vanadium oxo-peroxo complexes with
N,O-donor ligands derived from amino acids and alcohols
have been isolated previously and characterized by X-ray
diffraction.30 Such complexes have also been suggested
to play a role in oxidation of thioethers, based on 51V
NMR spectroscopy.36,45 Typical 51V NMR resonances for
such complexes appear between δ -500 and -600 ppm,
in agreement with our observation. Unfortunately, at-
tempts to isolate and characterize these presumed peroxo
complexes were unsuccessful.
The signal at δ -680 ppm observed at high H2O2
concentrations likely results from a complex formed by
full displacement of ligand 3 by H2O2. The resulting
achiral oxo-diperoxo complex, VO(O2)O2H (5) (Scheme
1),46 could nonselectively oxidize 1 to give racemic 2.
Formation of complex 5 has been previously linked to a
loss of enantioselectivity in the oxidation of thioethers.36
Additional evidence for formation of 5 is that addition of
H2O2 to VO(acac)2 in the absence of ligand 3 results in
formation of the species that gives rise to the peak at δ
-680 ppm. Furthermore, nonenantioselective oxidation
of 1 occurs in the absence of ligand 3 in CH3CN and in
iPrOH (homogeneous).
Isola tion of a Ca ta lyst P r ecu r sor . Aerobic oxidation
of VO(acac)2 in the presence of 3 in CH3CN produces red
crystals of a 2:1 ligand-to-vanadium complex, VOL*2H
(6) (eq 2). An X-ray diffraction study revealed a square
pyramidal vanadium complex of the endo isomer, where
the oxo ligand and the tert-butyl group of the tridentate
ligand are located on opposite sides of the plane of the
square pyramid, presumably for steric reasons.
F IGURE 2. 51V NMR spectra of VO(acac)2 and ligand 3 (1:1)
(a) with 100 equiv of H2O2 in CD2Cl2, (b) with 1 equiv of H2O2
in THF-d8, and (c) with 100 equiv of H2O2 in THF-d8.
H2O2 maintains a low peroxide concentration even under
homogeneous conditions, these results are consistent with
a stereoselective catalytic pathway operating at low
peroxide concentrations and a competing nonstereose-
lective catalytic pathway operating at high peroxide
concentrations. This provides an explanation for the
strong dependence of enantioselectivity on the solvent
system.
We considered that using stoichiometric amounts of
ligand 3 may prevent formation of the nonstereoselective
catalyst that forms under homogeneous conditions. Use
of a stoichiometric amount of ligand relative to H2O2
(0.05:1:1:1, VO(acac)2:1:3:H2O2) in CH3CN, however, still
produces racemic 2. Addition of 10 equiv of 3 relative to
H2O2 at 40% VO(acac)2 catalyst loading in CH3CN
inhibits oxidation and results in the formation of only
trace amounts of 2.
Dep en d en ce of Va n a d iu m Sp ecies on H2O2 Con -
cen tr a tion . Analysis of reaction mixtures by 51V NMR
spectroscopy provided further insight into the dramatic
solvent dependence on enantioselectivity. 51V NMR spec-
tra revealed that different species in observable diamag-
netic oxidation states (+5) are present under biphasic
conditions than under homogeneous conditions. A 51V
NMR spectrum of the organic layer of a solution of VO-
(acac)2 and ligand 3 in CD2Cl2 (biphasic) with 100 equiv
H2O2 layered on top showed nine resonances from δ -460
to -580 ppm (Figure 2a). In contrast, a 51V NMR
spectrum of the same materials in THF-d8 (homogeneous)
shows only one peak, at δ -680 ppm (Figure 2c).
Importantly, a 51V NMR spectrum of VO(acac)2 and
ligand 3 with only one equiv H2O2 in THF-d8 (homoge-
neous) is similar to the spectrum obtained under biphasic
conditions in CD2Cl2, and shows multiple peaks from δ
-500 to -580 ppm, and no upfield peak (Figure 2b). This
implies that the difference in spectra obtained under
Addition of 2.5% catalyst 6 to 1 and H2O2 in CHCl3
under biphasic conditions results in the formation of 2
with high conversion and ee (92%, 89% ee). This yield
and enantioselectivity are similar to those obtained using
VO(acac)2 and 3.
Sp ectr oscop ic Stu d ies of Ca ta lyst P r ecu r sor . Dis-
solution of crystalline 6 in CD2Cl2 resulted in the
observation of multiple species by 51V NMR (Figure 3)
1
and H NMR spectroscopy. Due to the complexity of the
1H NMR spectra, 51V NMR spectroscopy was used as the
primary tool to monitor the behavior of 6 in solution.
Importantly, the six resonances observed by 51V NMR
spectroscopy match six of the nine signals observed when
(45) Bryliakov, K. P.; Karpyshev, N. N.; Fominsky, S. S.; Tolstikov,
A. G.; Talsi, E. P. J . Mol. Catal. A 2001, 171, 73-80.
(46) Conte, V.; DiFuria, F.; Moro, S. J . Mol. Catal. A 1997, 117, 139-
149.
152 J . Org. Chem., Vol. 68, No. 1, 2003

Enantioselective Oxidation of Di-tert-Butyl Disulfide
SCHEME 1. Mod el for th e Obser ved Loss of En a n tioselectivity a t High H2O2 Con cen tr a tion s
Sp ectr oscop ic Stu d ies of Hom ogen eou s Oxid a -
tion of Disu lfid e. Addition of H2O2 (0.5 equiv) to
solutions of VO(acac)2 (0.1 equiv), 3 (0.11 equiv), and
disulfide 1 (1.0 equiv) in CH3CN (homogeneous) at 22 °C
results in the formation of multiple short-lived vanadium-
(V) species, as monitored by 51V NMR spectroscopy (δ
-620 to -690 ppm, broad). These species decompose to
give the characteristic multiple resonances formed on
dissolution of crystalline 6 after three minutes, at which
time oxidation of 1 is complete. Analysis of the oxidized
product showed it to have 0% ee. Addition of another
portion of H2O2 (0.5 equiv) re-forms the species that give
rise to the short-lived signals, and again these revert to
species present prior to addition of the second portion of
H2O2 within three minutes, at which time oxidation of 1
is complete (0% ee). This shows that all the short-lived
species are capable of accessing the nonstereoselective
oxidation pathway.
Mod el for Solven t a n d Slow H2O2 Ad d ition Effect.
Information from 51V NMR spectroscopy, the considerable
dependence of enantioselectivity on the solvent system,
and the dramatic increase in enantioselectivity upon slow
H2O2 addition into homogeneous systems strongly sup-
port a model in which a nonselective oxidant is the
dominant reactive species at high H2O2 concentrations
(Scheme 1). The relative ratio of enantioselective oxida-
tion to nonenantioselective oxidation depends both on the
relative concentrations of 4 and 5 and on their relative
oxidation rates. This model also rationalizes the similar
solvent dependence of enantioselectivity observed by
Bolm when this catalyst system is used in the oxidation
of thioethers.47
Effect of H218O. To determine whether the original
oxo group exchanged with water during the oxidation
reaction, H218O was used. Addition of H218O to a solution
of 6 in THF did not result in incorporation of 18O into
the oxovanadium complex, in the absence or presence of
(unlabeled) H2O2. Addition of H218O to standard oxida-
tions of 1 did not result in 18O incorporation into product
2.
F IGURE 3. 51V NMR spectrum of 6 in CDCl3 at 298 K.
H2O2 (100 equiv) is added to VO(acac)2 and ligand 3 in
CD2Cl2 (vide supra). Thus, 6 dissolves to give most of the
major species present under oxidation conditions, in the
absence of other potential ligands (peroxide or acetylac-
etonate). Solutions of 6 exhibit temperature- and con-
centration-dependent spectra. Reversible coalescence of
some peaks at higher temperatures suggest these vanan-
dium(V) species are in rapid equilibrium on the NMR
time scale. Rapid change of spectra upon dilution in CD2-
Cl2 suggests the presence of aggregates. Signals may also
be from both endo and exo isomers, despite the fact that
only the endo isomer is observed by X-ray crystal-
lography.
The three remaining unaccounted for signals observed
when H2O2 (100 equiv) is added to VO(acac)2 in CD2Cl2
(δ -474, -504, -516 ppm) are assigned to vanadium oxo-
peroxo complexes with ligand 3 bound, based on analogy
to the literature.30,36,45 To confirm that the new peaks
formed upon addition of H2O2 (30% aq) were from species
with incorporated peroxide and not water, additional H2O
(100 equiv) was added to a solution of H2O2 (1 equiv),
VO(acac)2, and 3 in CH3CN. The peaks at δ -474, -504,
and -516 ppm did not increase in intensity, nor were
any new peaks observed. 51V NMR spectra of VO(acac)2
and 3 in CH3CN with H2O, in the absence of peroxide,
show no resonances since VO(acac)2 is a paramagnetic
vanadium(IV) complex. Addition of H2O to 6 in THF does
not result in formation of new species. This further
suggests that the changes in the 51V NMR spectra arise
from reaction with peroxide and not H2O.
Isola tion of a Secon d Ca ta lyst P r ecu r sor . Due to
the complexity of the mixture of species formed upon
dissolution of catalyst precursor 6, we desired to synthe-
size a catalyst precursor that dissolved to give a simpler
mixture, ideally a single species. Additionally, we wanted
to examine the ease of exchange of the monodentate
ligand in 6, since it is likely that this ligand exchanges
(47) Bienewald, F. In Entwicklung neuer Katalysatorsysteme fur
asymmetrische Oxidationen. Ph.D. Thesis, Philipps-Universitat, Mar-
burg, 1996.
J . Org. Chem, Vol. 68, No. 1, 2003 153

Blum et al.
(95-98%) was purchased from Cambridge Isotope Laboratories
and used as received. All manipulations were performed in
air unless otherwise mentioned.
with peroxide in order to initiate the catalysis. Accord-
ingly, one equivalent of HCl was added to 6. 51V NMR
spectroscopy shows conversion to a single new vanadium-
(V) species in CD2Cl2 solution. A X-ray structure of
crystals grown from this solution shows the chloride
complex, 7 (eq 3). The crystals redissolve to give a
solution that exhibits a 51V NMR spectrum identical to
the spectrum of the solution from which they initially
crystallized. No reaction was observed upon addition of
HBr or HI to 6.
Caution: Vanadium-catalyzed H2O2 decomposition can re-
sult in O2 release. Therefore, care should be taken to use only
vented reaction vessels.
Typ ica l P r oced u r e for Syn th esis of ter t-Bu ta n e ter t-
Bu tylth iosu lfin a te (2). Ligand 3, VO(acac)2, and di-tert-butyl
disulfide 1 were dissolved in the desired solvent, giving a pale
green solution. The specified amount of H2O2 (30% aq) was
added and the solution turned deep red or brown immediately.
Gentle bubbling was often observed, presumably from O2
evolution. Conversion was monitored by 1H NMR spectroscopy.
1H NMR of 2 (400 MHz, CDCl3): δ 1.38 (s, 9H), 1.56 (s, 9H).
Identity was established by comparison with literature 1H
NMR data.15 The enantiomeric excess was determined by
chiral HPLC analysis (Diacel Chiralpak AS column, 97:3
hexanes:2-propanol; 1 mL/min, 258 nm; (R)-2, tR ) 6.8 min;
(S)-2, tR ) 8.4 min).
A blank reaction in CH3CN (homogeneous) with no added
VO(acac)2 showed less than 5% conversion to 2 by 1H NMR
spectroscopy after 7 h. A reaction carried out in CH3CN
(homogeneous) with no ligand 3 showed 92% conversion to 2
Use of 7 as the precatalyst in the oxidation of disulfide
1 resulted in conversions and enantioselectivities similar
to those produced with the catalyst that is formed in situ
upon addition of VO(acac)2 and ligand 3 separately.
Monitoring the oxidation reaction by 51V NMR spectros-
copy shows that 7 is the major vanadium-containing
species throughout the reaction. Two weak new low-field
signals appear upon addition of H2O2 to 7. The multiple
resonances from 6 gradually appear, leaving open the
possibility that catalysis proceeds through one of the
species responsible for these absorptions.
Su m m a r y a n d Con clu sion . The mechanism of the
vanadium-catalyzed enantioselective oxidation of disul-
fide 1 to thiosulfinate 2 has been studied. A 1:2 complex
of a vanadium(V) fragment and chiral ligand 3 has been
isolated and shown to be a catalyst precursor. Dramatic
increases in reaction enantioselectivity under homoge-
neous conditions result when the H2O2 oxidant is added
slowly. Additionally, a rationale for the observed depen-
dence of enantioselectivities of 2 on solvent has been
offered and experimentally supported. Different species
observed by 51V NMR spectroscopy at low and high H2O2
concentrations, an inverse relationship between organic
cosolvent miscibility with water and enantioselectivity,
and the dramatic increase in enantioselectivity upon slow
H2O2 addition support the competitive displacement of
all or part of 3 by H2O2 to form nonstereoselective
catalysts.
1
by H NMR spectroscopy after 6 h.
P r oced u r e for F or m a tion of 2 via Slow H2O2 Ad d ition .
Ligand 3 (10 mg, 0.030 mmol), VO(acac)2 (5 mg, 0.02 mmol),
and 1 (98 µL, 0.51 mmol) were dissolved in CH3CN to give a
pale green solution. The H2O2 (30% aq) (58 µL, 0.65 mmol)
was added dropwise via syringe pump over the specified time
(3-15 h) with rapid stirring. Upon addition of the first drop
of H2O2 the solution turned deep brown. Aliquots were
removed at the specified times and analyzed by HPLC. t ) 3
h, 68% ee; t ) 7 h, 76% ee; t ) 12 h, 66% ee, t ) 15 h, 70% ee.
Effect of Excess Liga n d 3. Ligand 3 (183 mg, 0.55 mmol),
VO(acac)2 (3 mg, 0.01 mmol), and 1 (30 µL, 0.17 mmol) were
dissolved in CH3CN (0.5 mL) to give a yellow oil. The H2O2
(30% aq) (5.5 µL, 0.055 mmol) was added in one portion to
produce a brown solution. While stirring over 21 h the solution
gradually turned orange and a pale precipitate formed. Four
drops of CHCl3 were added, dissolving the precipitate. Analysis
of an aliquot showed starting material 1 and ligand 3, but no
2. To be certain the ligand signal was not obscuring the signal
from 2, the remaining solution was frozen at -196 °C and the
volatile materials were removed under vacuum. The solution
was then heated to 37 °C and distilled into a round-bottom
flask submerged in liquid nitrogen, over 2 h. The distillate was
concentrated in vacuo to a clear residue and by HPLC showed
starting material 1, but no 2.
P r oced u r e for Mon itor in g Ad d ition of H 2O2 to VO-
(a ca c)2 a n d 3 (Hom ogen eou s Con d it ion s) by 51V NMR
Sp ectr oscop y. Ligand 3 (17 mg, 0.050 mmol) and VO(acac)2
(12 mg, 0.045 mmol) were dissolved in THF-d8 (∼0.5 mL) to
give a pale green solution. The solution was transferred to a
NMR tube, at which time the specified amount of H2O2 (30%
aq) was added. The solution immediately turned brown. For
H2O2 (1.0 equiv): 51V NMR (295 K, THF-d8) δ -505, -515,
-514, -525, -537, -544, -556, -562, -576 ppm. For H2O2
(10 equiv): 51V NMR (295 K, THF-d8) δ -437, -559, -619,
-626, -692, -712, -735 ppm. For H2O2 (100 equiv): 51V NMR
(295 K, THF-d8) δ -680 ppm.
P r oced u r e for Mon itor in g Ad d ition of H 2O2 to VO-
(a ca c)2 a n d 3 (Heter ogen eou s Con d ition s) by 51V NMR
Sp ectr oscop y. Ligand 3 (6.3 mg, 0.019 mmol) and VO(acac)2
(5.0 mg, 0.019 mmol) were dissolved in CD2Cl2 (0.5 mL) to give
a pale green solution. The solution was transferred to a NMR
tube, at which time the specified amount of H2O2 (30% aq)
was added as a second layer. The solution was mixed, and the
organic layer immediately turned brown. 51V NMR spectra on
the organic layer were recorded while the aqueous layer
remained on top. For H2O2 (1.0 and 4.0 equiv): 51V NMR (295
K, CD2Cl2) δ -476, -504, -520, -536, -545, -560 ppm. For
Exp er im en ta l Section
Gen er a l Meth od s. All 1H NMR and 13C{1H} NMR spectra
1
were recorded on 400 and 500 MHz spectrometers. H NMR
chemical shifts (δ) are reported in parts per million (ppm)
downfield of tetramethylsilane, and referenced to the residual
protiated solvent peak. 51V NMR spectra were recorded at
131.53 MHz and referenced against an external VOCl3 stan-
dard (δ 0 ppm). X-ray structural analysis was performed by
Dr. Fred Hollander and Dr. Allen Oliver in the University of
California, Berkeley CHEXRAY facility. X-ray diffraction
measurements were made on a SMART CCD area detector
with graphite monochromated Mo-KR radiation. Schiff-base
ligand 3 was prepared according to previously published
syntheses.3 Di-tert-butyl disulfide was distilled prior to use.
Bis(acetylacetonato)oxovanadium (VO(acac)2) was purchased
from Strem Chemical Company and used as received. H218O
154 J . Org. Chem., Vol. 68, No. 1, 2003

Enantioselective Oxidation of Di-tert-Butyl Disulfide
H2O2 (100 equiv): 51V NMR (295 K, CD2Cl2) δ -474, -504,
-516, -518, -527, -533, -540, -557, -573 ppm.
VOL*Cl (7). To a deep red solution of 6 (100 mg, 0.140
mmol) in THF (6 mL) was added HCl (1.0 M in ether) (140
VOL*2H (6). Ligand 3 (1.02 g, 3.05 mmol) and VO(acac)2
(403 mg, 1.52 mmol) were dissolved in CH3CN (40 mL) to give
a green solution, and the flask was loosely capped. Over the
course of 5 days the solution turned brown with accompanying
formation of red crystals. The brown mother liquor was
removed by pipet and the remaining X-ray quality crystals
were washed with CH3CN and air-dried to afford 812 mg (73%)
of VOL*2H (6). A second crop of crystals was collected by slow
evaporation of the mother liquor over 4 days to half the original
solvent volume, affording 68 mg of crystals for a total yield of
880 mg (79%). 1H NMR, 13C{1H} NMR, and 51V NMR (CD2-
Cl2) analysis showed a complex mixture of species. 1H NMR
(500 MHz, 295 K, 0.11 M, CD2Cl2): δ 14.02 (s), 13.99 (s), 13.74
(s), 8.54 (s), 8.49 (s), 8.39 (s), 8.26 (s), 8.23 (s), 7.66 (d, J ) 2
Hz), 7.64 (d, J ) 2 Hz), 7.42 (d, J ) 2 Hz), 7.39 (d, J ) 2 Hz),
7.36 (d, J ) 2 Hz), 7.34 (d, J ) 2 Hz), 7.20-7.18 (m), 6.91 (d,
J ) 2), 4.01 (d, J ) 6 Hz), 3.94 (d, J ) 11 Hz), 3.73 (t, J ) 11
Hz), 3.35 (dd, J ) 3, 9 Hz), 3.14 (dd, J ) 2, 9 Hz), 2.95 (dd, J
) 3, 9 Hz), 1.48 (s), 1.47 (s), 1.46 (s), 1.37 (s), 1.36 (s), 1.34 (s),
1.33 (s), 1.23 (s), 1.10 (s), 1.09 (s), 1.05 (s), 1.01 (s), 0.89 (s)
ppm. 51V NMR (295 K, 0.03 M, CD2Cl2): δ -518, -533, -538,
-551, -556, -573 ppm. Mp: 231-232 °C. Anal. Calcd for
µL, 0.14 mmol). The solution immediately turned black. A 51
V
NMR spectrum of an aliquot taken after 30 min showed a
single vanadium(V) species. The reaction mixture was con-
centrated in vacuo to a black gum. Hexane (4 mL) was added.
The solution was allowed to sit undisturbed for 2 h, at which
time removal of the red-brown mother liquor revealed clusters
of black needle-shaped crystals coated with a white film
(presumably amine salts) and mixed with a minor amount of
crystalline 6. The crystals were washed with 3 portions of
pentane and air-dried. 51V NMR spectroscopy of the crystals
showed the same signal as that observed on the aliquot taken
earlier, and trace signals from 6. 51V NMR (295 K, CD2Cl2): δ
-432 ppm. 1H NMR (500 MHz, 295 K, CD2Cl2): δ 8.62 (s, 1H),
7.78 (d, J ) 4 Hz, 1H), 7.38 (d, J ) 4 Hz, 1H), 5.61 (m, 1H),
5.21 (m, 1H), 4.22 (m, 1H), 1.51 (s, 9H), 1.35 (s, 9H), 1.17 (s,
9H) ppm. Mp: 193-194 °C. HRMS (EI): m/z calcd (C21H33
VNO3Cl) 433.1589, found 433.1596 [M+].
-
X-r a y Cr ysta l Str u ctu r e of 7. X-ray quality crystals of 7
were grown by addition of ether to the black gum described
above to produce a black solution, followed by complete
evaporation of the ether solvent over 1 h. For X-ray crystal-
lographic analysis, a fragment of a black columnar crystal of
7 having approximate dimensions of 0.03 × 0.10 × 0.19 mm
was mounted on a glass fiber using Paratone N hydrocarbon
oil. The structure was solved by direct methods and expanded
using Fourier techniques. The non-hydrogen atoms were
refined anisotropically. Hydrogen atoms were included but not
refined.
51V NMR Sp ectr oscop y of Ad d ition of H2O2 to 7.
Complex 7 (15 mg, 0.035 mmol) was dissolved in CD2Cl2 to
produce an opaque black solution, H2O2 (30% aq) (32 µL, 0.32
mmol) was added with no color change, and the solution was
transferred to a NMR tube. A 51V NMR spectrum obtained
after 45 min showed two new minor peaks at low field (δ -252,
-264 ppm). In a 51V NMR spectrum obtained after 3 d those
two minor peaks are no longer present, and signals from 6
increased. 51V NMR (295 K): δ -432 (1.00), -540, -554, -556,
-573 (0.55 total) ppm.
C
42H67N2O5V: C, 69.01; H, 9.24; N, 3.83. Found: C, 69.19; H,
9.31; N, 3.84.
X-r a y Cr ysta l Str u ctu r e of 6. X-ray quality crystals were
grown according to the above-described method. For X-ray
crystallographic analysis, a deep red tablet-shaped crystal of
6 having approximate dimensions of 0.22 × 0.18 × 0.08 mm
was mounted on a glass fiber using Paratone N hydrocarbon
oil. The structure was solved by direct methods and expanded
using Fourier techniques. The vanadium, oxygen, nitrogen and
methyl carbon atoms were refined anisotropically, while the
rest of the carbon atoms were refined isotropically. Hydrogen
atoms were included in calculated positions but were not
refined.
51V NMR Stu d ies of Rea ction Mixtu r es. Ligand 3 (24
mg, 0.072 mmol), VO(acac)2 (17 mg, 0.065 mmol), and 1 (124
µL, 0.65 mmol) were dissolved in CD3CN (0.8 mL) to give a
pale green solution. The solution was transferred to a NMR
tube at which time H2O2 (30% aq) (32 µL, 0.28 mmol) was
added, resulting in a brown solution. The solution was mixed
and 51V NMR spectra were obtained. 51V NMR (295 K, CD3-
CN) δ t ) 2 min: -507, -533 (sharp), -600 to -700 (very
broad), -641 (sharp), -680 ppm. t ) 3 min: -480 to -540
(very broad), -510, -520, -533 (sharp), -642, -650 ppm. t )
5 min: -507, -520 (broad), -533 (sharp), -540 ppm.
Test for 18O In cor p or a tion in to 6 fr om H 218O. Complex
6 (4 mg, 0.006 mmol) was dissolved in THF (0.1 mL) to produce
an opaque dark red solution. The H218O was added with no
color change. An aliquot was removed after 6 h and dissolved
in CH3CN. A mass spectrum showed no isotopic enrichment
in 6. MS (ESI): m/z 731 [M + H]+.
Ack n ow led gm en t. This work was supported by the
NSF through Grant No. CHE-019468 to J .A.E., Grant
No. CHE-0094349 to R.G.B., and a graduate fellowship
to S.A.B. We thank Dr. Fred Hollander and Dr. Allen
Oliver of the UC Berkeley CHEXRAY facility for X-ray
structure determination. The Center for New Directions
in Organic Synthesis is supported by Bristol-Meyers
Squibb as a Sponsoring Member and Novartis as a
Supporting Member.
Su p p or tin g In for m a tion Ava ila ble: X-ray crystallo-
graphic data for 6 and 7, variable-temperature 51V NMR
spectra of 6, and a 51V NMR spectrum of 7. This material is
available free of charge via the Internet at http://pubs.acs.org.
The experiment was repeated, with the exception that
(unlabeled) H2O2 (30% aq) (1.1 µL, 0.011 mmol) was added.
An aliquot was removed after 6 h and showed no isotopic
enrichment in 6. MS (ESI): m/z 731 [M + H]+.
J O0205560
J . Org. Chem, Vol. 68, No. 1, 2003 155