Experimental investigation into the mechanism of the epoxidation of aldehydes with sulfur ylides

Article abstract of DOI:10.1021/ol702300d

A mechanistic study of the epoxidation of aldehydes with sulfur ylides has been carried out. The ΔG^? of the reaction was determined to be 22.2 kcal/mol at 298 K. A ¹³C kinetic isotope effect was determined to be 1.026 for the carbonyl carbon of benzaldehyde. A secondary deuterium isotope effect was determined to be 0.93 for the aldehydic hydrogen atom of benzaldehyde. Substituent effects on reaction rate were studied, and a Hammett p of +2.50 was found.

Full text of DOI:10.1021/ol702300d

ORGANIC
LETTERS
2
007
Vol. 9, No. 26
481-5484
Experimental Investigation into the
Mechanism of the Epoxidation of
Aldehydes with Sulfur Ylides
5
David R. Edwards,* Pedro Montoya-Peleaz, and Cathleen M. Crudden*
Department of Chemistry, Queen’s UniVersity, 90 Bader Lane, Kingston, ON, Canada
K7M 3N6
cruddenc@chem.queensu.ca
Received October 4, 2007
ABSTRACT
A mechanistic study of the epoxidation of aldehydes with sulfur ylides has been carried out. The
2.2 kcal/mol at 298 K. A 13C kinetic isotope effect was determined to be 1.026 for the carbonyl carbon of benzaldehyde. A secondary deuterium
isotope effect was determined to be 0.93 for the aldehydic hydrogen atom of benzaldehyde. Substituent effects on reaction rate were studied,
and a Hammett of 2.50 was found.
∆
Gq of the reaction was determined to be
2
G
+
1
The reaction between an aldehyde and an ylide yielding an
epoxide has proven to be a versatile and valuable method
cidic amides, and acids. From a variety of mechanistic
3
probes, in addition to theoretical calculations, a general
consensus on the mechanism of the reaction has been
reached, which is illustrated in Scheme 1.
1
for the production of epoxides. Since the initial discovery
2
of this reaction, catalytic enantioselective variants have been
developed that permit the synthesis of a broad range of
products including epoxides, cyclopropanes, aziridines, gly-
(
1) (a) Aggarwal, V. K.; Winn, C. L. Acc. Chem. Res. 2004, 37, 611.
Scheme 1. Mechanism of Epoxidation
(
b) Aggarwal, V. K.; Alonso, E.; Fang, G.; Ferrara, M.; Hynd, G.; Porcelloni,
M. Angew. Chem., Int. Ed. 2001, 40, 1433. (c) Aggarwal, V. K.; Charmant,
J. P. H.; Fuentes, D.; Harvey, J. N.; Hynd, G.; Ohara, D.; Picoul, W.;
Robiette, R. I.; Smith, C.; Vasse, J.-L.; Winn, C. L. J. Am. Chem. Soc.
2
006, 128, 2105. (d) Aggarwal, V. K; Hebach, C. Org. Biomol. Chem. 2005,
3
, 1419. (e) Aggarwal, V. K.; Richardson, J. Chem. Commun. 2003, 2644.
(2) (a) Johnson, A. W.; LaCount, R. B. Chem. Ind. 1958, 1440. (b)
Johnson, A. W.; LaCount, R. B. J. Am. Chem. Soc. 1961, 83, 417. (c) The
+
Corey-Chaykovsky version of this reaction employs Me3S species as the
ylide precursor rather than the substituted ylides employed herein: Corey,
Cross-over experiments indicate that the addition of
E. J.; Chaykovsky, M. J. Am. Chem. Soc. 1962, 84, 3782.
dimethylsulfonium benzylide to benzaldehyde (k
nonreversible and rate-limiting. The activation energy (E
) is both
2
(3) (a) Aggarwal, V. K.; Harvey, J. N.; Richardson, J. J. Am. Chem.
4
a
)
Soc. 2002, 124, 5747. (b) Volatron, F.; Eisentein, O. J. Am. Chem. Soc.
1987, 109, 1. (c) Lindvall, M. K.; Koskinen, A. M. P. J. Org. Chem. 1999,
64, 4595. (d) Lindvall, M. K.; Koskinen, A. M. P. Tetrahedron 2001, 57,
4629.
for this step was predicted to be 4.5 kcal/mol by DFT
3a
calculations. No experimentally determined activation
parameters have been reported for the epoxidation of
benzaldehyde with dimethylsulfonium benzylide. In fact, to
the best of our knowledge, no kinetic parameters have ever
been reported for this reaction or the Corey-Chaykovsky2c
(
4) (a) Aggarwal, V. K.; Calamai, S.; Ford, J. G. J. Chem. Soc., Perkin
Trans. 1 1997, 593. (b) Aggarwal, V. K Charmant, J. P. H; Ciampi, C.;
Hornby, J. M.; O’Brien, C. J.; Hynd, G.; Parsons, R. J. Chem. Soc., Perkin
Trans. 1 2001, 3159. (c) Yoshimine, M.; Hatch, M. J. J. Am. Chem. Soc.
1
967, 89, 5831.
1
0.1021/ol702300d CCC: $37.00
© 2007 American Chemical Society
Published on Web 11/28/2007

epoxidation of any substrate. Owing to this fact, and the
overall importance of this reaction in organic synthesis, we
undertook to experimentally determine the rate expression
and activation parameters for the sulfur ylide epoxidation
of benzaldehyde. Last, it was deemed that the transition state
characteristics of the reaction could be obtained by a series
of mechanistic probes such as isotope and substituent effects.
Doing so, the sulfur ylide epoxidation of benzaldehyde could
be placed into context with other similar reactions involving
additions to arylaldehydes such as the related Wittig olefi-
nation, which have been studied in considerable detail.
Using our standard conditions, epoxidations were carried
out in CD
3
CN under pseudo-first-order conditions of excess
Figure 2. Rate versus [PhCHO].
benzaldehyde and limiting sulfonium salt (eq 1). The progress
of the reaction could be effectively monitored by high-field
H NMR (600 MHz) for loss of sulfonium salt. Under these
1
5
derived. From eq 4, then, the apparent second-order rate
constant kapp was derived. The apparent second-order rate
constant determined in this manner is a composite of the
individual rate constants of eq 2 and may simply be expressed
conditions, nearly complete selectivity for the trans-epoxide
isomer is observed (ca. 98:2).
2 eq
as kapp ) k K .
An Eyring plot was generated by plotting log(kapp/T) versus
q
1
/T from which the enthalpy of activation (∆H ) 11.8 kcal/
q
mol) and the entropy of activation (∆S ) -35 cal/mol/K)
were determined. This corresponds to an overall free energy
q
of activation (∆G ) for the reaction of 22.2 kcal/mol at 298
K. The large negative ∆S is consistent with a bimolecular
addition resulting in the conversion of two molecules into a
single activated complex. The analogous Arrhenius treatment
q
Monitoring the reaction with increasing concentrations of
DBU (0.029-0.16 M) and plotting kobs versus [DBU] yields
a straight line (Figure 1), indicating that the reaction is first-
of the temperature-dependent rate data results in an E
2.3 kcal.
a
of
1
-
d[SM]
dt
[DBU]
)
K k
[PhCHO][sulfonium salt]
(2)
(3)
(4)
eq
2
[DBUH]
[
DBU]
where kobs ) Keqk2
[ PhCHO]
[
DBUH]
kobs[DBUH]
and kapp
)
[PhCHO][DBU]
The irreversibility of sulfur ylide addition to benzaldehyde
2
(k ) has previously been demonstrated in the solvents DMSO
4a
Figure 1. Rate versus [DBU].
2 2
and CH Cl by cross-over experiments. The results of these
previous studies have been confirmed under the current
reaction conditions, by treating betaine precursor anti-2 with
base in the presence of a more reactive aldehyde (eq 5).
Under these conditions, clean conversion to trans-stilbene
oxide was observed. No cross-over products were observed,
as would be expected if the betaine reverted to benzaldehyde
and ylide. Accordingly, the addition of sulfur ylide to
benzaldehyde is interpreted as being an nonreversible
process.
order with respect to base. The reaction was also monitored
under increasing concentrations of benzaldehyde (0.017-
0.26 M), and in this instance, plotting kobs versus [benzal-
dehyde] revealed saturation behavior (Figure 2).
The reaction was monitored over a 30 °C temperature
range using a concentration of benzaldehyde in which the
plot of rate versus [PhCHO] is linear (0.017 M). This
permitted us to determine kobs in the usual manner. Working
under the assumption that, under these conditions, the
addition of 1 to benzaldehyde (k ) is rate-limiting and is
2
preceded by a pre-equilibrium proton transfer (Keq) to
generate the ylide from sulfonium salt, eq 2 can then be
(5) Treating the proton transfer (Keq) as a pre-equilibrium is justified
since the ratio of k-1[DBUH]/k2[PhCHO] is 22.7 under the conditions
employed ([DBUH] ) 0.021 M, [PhCHO] ) 0.017 M). The ratio of k-1/k2
was determined from the slope of the double reciprocal plot 1/kobs vs
1/[benzaldehyde]; see Supporting Information. Treating k2 as rate-limiting
is justified by the combined results of cross-over experiments, isotope
effects, as well as substituent effects determined for the reaction.
5482
Org. Lett., Vol. 9, No. 26, 2007

in the transition state and is consistent with rate-limiting
9
addition of the sulfur ylide to benzaldehyde.
In order to gain further insight into the nature of the rate-
limiting step in the sulfur ylide addition to benzaldehyde, a
1
3
C kinetic isotope effect study was carried out using the
A Hammett study was then conducted in order to assess
the nature of charge development in the transition state of
the reaction. The relative reactivity of substituted benzalde-
hydes was determined by a series of competition experiments
in which at least a 10-fold excess of benzaldehyde and p-X-
benzaldehyde was allowed to compete for a limiting amount
6
Singleton protocol. Using this technique, unlabeled, natural
1
3
abundance C-containing benzaldehyde is subjected to a
large-scale epoxidation and taken to high conversion. Fol-
lowing reaction, the unreacted starting material is recovered
and the enrichment of the slower reacting isotope is
determined via quantitative C NMR. The para-carbon of
benzaldehyde is assumed to undergo a negligible KIE during
the reaction and, as such, serves as a reference for the
13
1
0
of sulfonium salt. Under these conditions, krel can be
11
calculated from eq 8. The resulting Hammett plot is shown
in Figure 3.
1
3
enrichment of C at the carbonyl center. From these data, a
13
C KIE can be calculated in accordance with standard theory
X
H
utilizing eq 67
X
H
[Product ]/[Product ]
k /k )
(8)
X
H
[
Aldehyde ]/[Aldehyde ]
ln(1 - F)
ln[(1 - F)(R/R )]
KIE )
(6)
The Hammett slope, F, is determined to be +2.50 for the
reaction of dimethylsulfonium benzylide with substituted
o
where F ) fractional conversion and R/R
o
) ratio of 13C in
benzaldehyde prior to and after reaction as determined by
quantitative 13C NMR. The C KIE of the carbonyl carbon
of benzaldehyde was determined to be 1.026 ( 0.003
13
(average of three experiments, standard deviations calculated
by propagation of errors). Full details are provided in the
Supporting Information. This relatively large 13C KIE is
consistent with a rate-determining addition to the carbonyl.
Secondary deuterium kinetic isotope effects (SDKIEs) are
also useful probes for the analysis of rehybridization at the
attached carbon undergoing direct bond change during the
reaction. Rate-limiting addition of a nucleophilic species to
an aldehyde is generally associated with an inverse SDKIE
at the aldehydic proton. The inverse nature of the KIE, where
Figure 3. Hammett plot for reaction of ArCHO.
H
D
k /k < 1, is interpreted as arising from an increase in the
aldehydic C-H(D) force constant for out-of-plane bending
benzaldehydes in CH CN. The positive sign of F reflects
3
2
on going from the ground state (sp ) to the transition state
the fact that electron-withdrawing substituents on the aro-
matic ring accelerate the reaction, which is consistent with
a nucleophilic attack of the sulfur ylide on the aldehyde. The
large absolute value of F indicates that there is a substantial
buildup of negative charge in the activated complex as it
proceeds through to the transition state.
The results of the current kinetic study reveal that the
epoxidation reaction is first-order with respect to [DBU] and
[sulfonium salt], while saturation behavior is observed in
[benzaldehyde]. These results are consistent with the mecha-
nistic proposal depicted in Scheme 1 in which addition of
3
8
(
sp -like).
The SDKIE has been determined for the epoxidation of
benzaldehyde with dimethylbenzylsulfonium tetrafluorobo-
rate. A 55:45 mixture of benzaldehyde and d -benzaldehyde
was subjected to epoxidation under typical reaction condi-
tions in CD CN. The relative conversion of each isotopomer
was determined by H NMR following reaction. From this,
1
3
1
H
D
a krel ) k /k was calculated using eq 6. An inverse SDKIE
of 0.93 was determined for the addition of dimethylsulfonium
benzylide to benzaldehyde. The inverse nature of the SDKIE
2
3
is typical of processes involving sp to sp rehybridization
(9) (a) Gajewski, J. J.; Wojciech, B.; Harris, N. J.; Olson, L. P.; Gajewski,
J. P. J. Am. Chem. Soc. 1999, 121, 326. (b) Gajewski, J. J.; Wojciech, B.;
Brichford, N.; Henderson, J. L. J. Org. Chem. 2002, 67, 4236.
(10) An excess of the slower reacting aldehyde was used in order to
obtain a significant proportion of the product derived from this species and
allow for accurate integrations to be made in the proton NMR spectrum.
(11) Lansbury, P. T.; MacLeay, R. E. J. Am. Chem. Soc. 1965, 87, 831.
(
(
6) Singleton, D. A.; Thomas, A. A. J. Am. Chem. Soc. 1995, 117, 9357.
7) Melander, L.; Saunders, W. H. Reaction Rates of Isotopic Molecules;
Wiley-Interscience: New York, 1980; Chapter 4, pp 95-102.
(
8) Buncel, E.; Lee, C. C. Secondary and SolVent Isotope Effects (Isotopes
in Organic Chemistry); Elsevier Science Ltd.: Amsterdam, 1997
Org. Lett., Vol. 9, No. 26, 2007
5483

i
sulfur ylide to benzaldehyde is rate-limiting. This result is
expected based on previously reported results of cross-over
experiments, which were verified under the current set of
KIE ) EIE
(9)
The experimentally determined EIE for the hydrocyanation
of p-methoxybenzaldehyde is reported to be 0.78. Hill has
used a similar value to estimate the EIE for borohydride
reductions of benzaldehyde. An EIE of 0.78 has also been
used in a study of organolithium and Grignard additions to
benzaldehyde. Employing an EIE of 0.78 along with eq 8
leads to the calculation of i ) 0.29 for the sulfur ylide
epoxidation of benzaldehyde. This would indicate moderate
product character in the activated complex. However, it is
not completely clear whether or not the EIE of 0.78 is an
appropriate model for the current reaction. In fact, Gajewski
has demonstrated that the identity of the countercation in
the reaction strongly affects the magnitude of the predicted
4
15
reaction conditions. The reaction is shown to proceed with
q
a ∆G of 22.2 kcal/mol at 298 K. The analogous Arrhenius
16
treatment of the temperature-dependent rate data results in
the calculation of E
This value is larger than the E
Aggarwal using DFT calculations for the title reaction in
a
) 12.3 kcal/mol.
9a
a
of 4.5 kcal calculated by
3
a
3
CH CN. The discrepancy between the two values is likely
a result of the means by which each was determined. In the
DFT method, the energies of the various intermediates and
transition states for the reaction between preformed dimeth-
ylsulfonium benzylide and benzaldehyde were calculated. On
the other hand, the results presented here were obtained from
a kinetic analysis of the reaction and necessarily include the
acid-base reaction between DBU and the sulfonium salt to
generate the ylide species.
9
a
EIEs. Considering the large Hammett F (+2.50) and
1
3
relatively large C KIE (1.026), it appears as though i may
be underestimating the position of the transition state for the
addition of the sulfur ylide to benzaldehyde.
In conclusion, the reaction between dimethylbenzyl sul-
fonium tetrafluoroborate and benzaldehyde has been studied
using a variety of kinetic tools. All of these methods point
13
The observed carbonyl C KIE of 1.026 is most consistent
with a reaction step in which this atom is heavily involved
2
in reaction coordinate motion as in addition (k ) of sulfur
to the addition of the ylide to benzaldehyde as rate-
ylide to benzaldehyde. The large Hammett F of +2.50
indicates that a buildup of negative charge is occurring at
the carbonyl carbon during the course of the reaction. A
nucleophilic addition such as that depicted by k might be
2
expected to involve this magnitude of charge development.
3a,4a
determining.
Despite the importance of the epoxidation
of aldehydes with sulfur ylides, this represents the first
experimental determination of the activation parameters for
this reaction. In addition, transition state characteristics have
been probed by a combination of isotopic and substituent
effects. The results of these studies lend support to the
conclusions drawn from the kinetic analyses we have
performed. Although the results of the current study fall short
of positioning the transition state along the reaction coor-
dinate with confidence, they provide considerable experi-
mental verification for literature and computational postulates
of the mechanism of this important reaction. Most impor-
tantly, these results shed light on charge development in the
activated complex and provide a basis for future study of
transition state characteristics.
For example, the Hammett F for hydrocyanation of benzal-
1
2
dehyde is reported to be +2.33. Similarly, reduction of
1
3
acetophenone with sodium borohydride is +3.06, and
+
-
3
addition of the phosphonium ylide Ph P C HPh to benzal-
dehyde is +2.77.14 The inverse SDKIE of 0.93 for the
2
aldehydic proton of benzaldehyde is indicative of an sp to
3
sp rehybridization occurring during the multistep epoxida-
tion. This interpretation is in line with standard theory, which
predicts this outcome based on changes in out-of-plane
2
3
bending frequencies for sp and sp hybridized CH bonds.
It would therefore appear that treating k of Scheme 1 as
2
Acknowledgment. The Natural Sciences and Engineering
Research Council of Canada (NSERC) is gratefully acknowl-
edged for support of this work in terms of operating and
equipment grants to C.M.C. D.R.E. thanks the government
of Ontario and Queen’s University for graduate scholarships.
rate-determining is fully consistent with the results of heavy
atom and secondary deuterium isotope effects as well as
substituent effects on the reaction.
A useful technique to experimentally probe transition state
structure is to compare observed SDKIEs with the predicted
equilibrium isotope effect for complete bond formation by
Supporting Information Available: Experimental pro-
cedures, details of determination of order in benzaldehyde
and DBU, justification for eq 2, details of temperature
studies, C isotope effects, and H isotope effect determi-
nation. This material is available free of charge via the
Internet at http://pubs.acs.org.
9
means of eq 9. The parameter i is generally believed to be
1
3
2
indicative of progress along the reaction coordinate assuming
that the observed SDKIE varies monotonically from reactant
to product.
OL702300D
(12) Baker, J. W.; Hopkins, H. B. J. Chem. Soc. 1949, 1089.
(13) Bowden, K.; Hardy, M. Tetrahedron 1966, 22, 1169.
(14) Yamataka, H.; Nagareda, K.; Ando, K.; Hanafusa, T. J. Org. Chem.
(15) Cordes, E. H.; Bull, H. G.; Amaral, L. J. Am. Chem. Soc. 1972, 94,
7579.
1
992, 57, 2865.
(16) Hill, E. A.; Milosevich, S. A. Tetrahedron Lett. 1976, 35, 3013.
5484
Org. Lett., Vol. 9, No. 26, 2007