Concerning the reactivity of dioxiranes. Observations from experiments and theory

Article abstract of DOI:10.1021/ja075068u

The challenging hypothesis of a biphilic (i.e., electrophilic vs nucleophilic) character for dioxirane reactivity, which envisages that electron-poor alkenes are attacked by dioxiranes in a nucleophilic fashion, could not be sustained experimentally. Rate data, which estimate Hammett rho values for the epoxidation of 3- or 4-substituted cinnamonitriles X·Ph-CH=CH-CN, unequivocally allow one to establish that dioxiranes epoxidize electrophilically even alkenes carrying electron-withdrawing-groups. The greater propensity of methyl(trifluoromethyl) dioxirane TFDO (1b) to act as an electrophilic oxidant with respect to dimethyldioxirane DDO (1a) parallels the cathode reduction potentials for the two dioxiranes, as measured by cyclic voltammetry. A simple FMO approach for alkene epoxidation is helpful to conceive a likely rationale for the greater oxidizing power of TFDO as compared to DDO.

Full text of DOI:10.1021/ja075068u

Published on Web 01/05/2008
Concerning the Reactivity of Dioxiranes. Observations from
Experiments and Theory
Cosimo Annese,† Lucia D’Accolti,† Anna Dinoi,† Caterina Fusco,§
Remo Gandolfi,*,‡ and Ruggero Curci*,†
Dipartimento Chimica, UniVersita` di Bari, CNR.-ICCOM, V. Amendola 173,
I-70126 Bari, Italy, and Dipartimento di Chimica Organica,
UniVersita` di PaVia, V.le Taramelli 10, 27100 PaVia, Italy
Received July 9, 2007; E-mail: curci@ba.iccom.cnr.it
Abstract: The challenging hypothesis of a “biphilic” (i.e., electrophilic vs nucleophilic) character for dioxirane
reactivity, which envisages that electron-poor alkenes are attacked by dioxiranes in a nucleophilic fashion,
could not be sustained experimentally. Rate data, which estimate Hammett “rho” values for the epoxidation
of 3- or 4-substituted cinnamonitriles X‚PhsCHdCHsCN, unequivocally allow one to establish that
dioxiranes epoxidize electrophilically even alkenes carrying electron-withdrawing groups. The greater
propensity of methyl(trifluoromethyl)dioxirane TFDO (1b) to act as an electrophilic oxidant with respect to
dimethyldioxirane DDO (1a) parallels the cathode reduction potentials for the two dioxiranes, as measured
by cyclic voltammetry. A simple FMO approach for alkene epoxidation is helpful to conceive a likely rationale
for the greater oxidizing power of TFDO as compared to DDO.
at present as the highlight of dioxirane chemistry.1 For this
Introduction
remarkable transformation, the high stereoselectivity as well as
kinetic evidence all point to an oxenoid (electrophilic) mech-
anism of insertion.1
In the past several years dioxiranes have rapidly grown into
one of the more useful oxidation processes in the arsenal of
synthetic chemists.1 Their unique reactivity ranges from the
transfer of an oxygen atom to “unactivated” C-H bonds of
saturated hydrocarbons, to epoxidations, to the oxidation of
heteroatoms containing a lone pair of electrons such as amines
and sulfides. In fact, applications in synthesis of the currently
popular dimethyldioxirane (1a) (DDO)2a and of methyl(trifluo-
romethyl)dioxirane (1b) (TFDO)3 in isolated form have prolifer-
ated the access to key products that are useful in organic
synthesis.1-5
Given the value of epoxides as synthons, among the myriad
of useful dioxirane oxidations, the most common reaction
undoubtedly remains the oxidation of carbon-carbon double
bonds of alkenes.1-5 In fact, the use of dioxiranes today rivals
that of peracids in epoxidation reactions because of the enhanced
rates of reaction and of their performance under neutral
conditions, providing access to even highly sensitive epoxides.
Alkenes of various kinds, either electron-rich or electron-poor,
have been successfully epoxidized employing dioxiranes.5
The efficient oxyfunctionalization of unactivated C-H bonds
of alkanes under extremely mild conditions undoubtedly ranks
†
University of Bari.
University of Pavia.
C.N.R.- Istituto di Chimica dei Composti Organometallici (ICCOM),
‡
§
Bari Section, Italy.
(1) For a recent review, see: Curci, R.; D’Accolti, L.; Fusco, C. Acc. Chem.
Res. 2006, 39, 1.
(2) (a) Murray, R. W.; Jeyaraman, R. J. Org. Chem. 1985, 50, 2847. (b)
Cassidei, L.; Fiorentino, M.; Mello, R.; Sciacovelli, O.; Curci, R. J. Org.
Chem. 1987, 52, 699. (c) Murray, R. W.; Singh, M. Org. Synth. 1997, 74,
91.
(3) (a) Mello, R.; Fiorentino, M.; Sciacovelli, O.; Curci, R. J. Org. Chem. 1988,
53, 3890. (b) Mello, R.; Fiorentino, M.; Fusco, C.; Curci, R. J. Am. Chem.
Soc. 1989, 111, 6749. (c) D’Accolti, L.; Fusco, C.; Rella, M. R.; Curci, R.
Synth. Commun. 2003, 33, 3009. (d) Adam. W.; Curci, R.; Gonzale`s-Nun˜ez,
M. E.; Mello, R. J. Am. Chem. Soc. 1991, 113, 7654.
(4) (a) Curci, R.; Fiorentino, M.; Troisi, L.; Edwards, J. O.; Pater, R. H. J.
Org. Chem. 1980, 45, 4758. (b) Baumstark, A. L.; Vasquez, P. C. J. Org.
Chem. 1988, 53, 3437. (c) Murray, R. W.; Gu, D. J. Chem. Soc., Perkin
Trans. 2 1993, 2203. (d) Curci, R.; Dinoi, A.; Rubino, M. F. Pure Appl.
Chem. 1995, 67, 811. (e) Liu, J.; Houk, K. N.; Dinoi, A.; Fusco, C.; Curci,
R. J. Org. Chem. 1998, 63, 8565. (f) Curci, R. IASOC - Ischia AdVanced
School of Organic Chemistry, VIII Session, Ischia (Italy); September 26-
October 1, 1998; Abstracts, L9.
Much of the early4a and subsequent experimental evidence,
including the syn-stereospecific course of dioxirane epoxidations
and the greater reactivity of cis alkenes with respect to their
trans isomers,4 kinetics, and H/D isotope effects all have pointed
to a substantially concerted mechanism,4 akin to that of
epoxidation by peracids.6,7 A spiro transition state (TS) for the
(6) For instance, see: (a) Roof, A. A. M.; Winter, W. J.; Bartlett, P. D. J.
Org. Chem. 1985, 50, 4093. (b) Hoveyda, A. H.; Evans, D. A.; Fu, G. C.
Chem. ReV. 1993, 93, 1307. (c) Rebek, J., Jr Heterocycles 1981, 15, 517.
(7) Bach, R. D. In The Chemistry of Peroxides; Patai, S., Ed.; Wiley: New
York, 2006; Vol. 1, Chapter 1. See references therein.
(5) (a) Adam, W.; Zhao, C.-G.; Saha-Mo¨ller, C. R. Org. React. 2002, 61, 219.
(b) Adam, W.; Zhao, C.-G. In The Chemistry of Peroxides; Patai, S., Ed.;
Wiley: New York, 2006; Vol. 2, Chapter 14. See references therein.
9
10.1021/ja075068u CCC: $40.75 © 2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 1197-1204
1197

A R T I C L E S
Chart 1a
Annese et al.
Table 1. Relative Rates of Epoxidation of Cinammonitriles with
Dioxiranes 1a,b in Acetone Solvent
rel. rate
(kX/kH)a
103 k2 1
(M-1 s-
)
DDO
(1a)b
TFDO
(1b)c
DDO
TFDO
entry
substrate
substituent
(1a)
(1b)
1
2
3
4
5
6
7
8
(E)-4a
(E)-4b
(Z)-4b′
(E)-4c
(Z)-4c′
(E)-4d
(E)-4e
(E)-4f
p-OCH3
p-CH3
p-CH3
H
H
p-Cl
7.02
2.21
2.19
1.00
1.10
0.84
0.42
0.28
4.90
1.00
1.95
0.53
0.24
0.045
0.7d
37e
m-Cl
m-NO2
a No products derived from rearrangement were detected (ref 11).
a Relative rate coefficients estimated as averages ((2-8%) from log([A]/
[A]o)/log([B]/[B]o); the concentrations of two given competing substrates
(A and B) were determined (HPLC) at various times (usually over the first
15-50% of reaction). bAt 25.0 °C, with [DDO]o ) 0.05-0.08 M and
dioxirane approach to the double bond was suggested to be
consistent with all these data.4b,d Following the pioneering
theoretical work by Bach,7 numerous computations also sub-
stantiated the claim of a largely concerted mechanism. The most
recent calculations at a high-level of theory on this topic confirm
the previous notion envisaging a concerted nucleophilic attack
by the alkene π bond at the O-O bond of the dioxirane
electrophile along an SN2-type coordinate, much in the fashion
of the mechanism well established for nucleophilic attack of
common peroxides.8 Calculations have shown that uneven
substitution of the CdC can cause the transition state to become
unsymmetrical, but no intermediates were located on the
potential surface.7
Because of their relevance in synthesis, mechanistic studies
of key transformations involving dioxiranes have continued
unabated during the past decade.1 In this regard, the involvement
of radical pathways9 could be discounted also for dioxirane
epoxidations on the basis of considerable evidence derived from
careful experiments and computations.4e,10
Furthermore, the application of the ultrafast (2,2-diphenyl)-
cyclopropylcarbinyl radical (3) probe11a to the transformation
of 1,1-diphenyl-2-vinylcyclopropane (2a) into the corresponding
epoxide (3a), yielding no rearrangement products (Chart 1),
allowed us to rule out pathways involving discreet radical
intermediates.11b
Concerning the electronic character of the dioxirane oxygen
transfer from dioxiranes to alkenes, a more appealing mecha-
nistic question has been advanced recently by Deubel and co-
workers.12 In fact, using high-level computational methods, these
authors employed Charge Decomposition Analysis (CDA) to
estimate the donor-acceptor interactions in the transition
c
[Substrate]o ) (0.1-0.5) × 10-2 M. At 5.0 °C, with [TFDO]o ) 0.13-
0.17 M and [Substrate]o ) (0.2-0.6) × 10-2 M. dAt 25.0 °C, absolute rate
constant k2, estimated ((4-8%) as (k1/[E-4c]o) from runs under pseudo
first-order conditions, with [DDO]o ) (0.05-0.08) × 10-2 M and [E-4c]o
e
) (4.0-6.5) × 10-2 M. At 5.0 °C, absolute rate constant k2 ((5-10%)
from kinetics run under second-order conditions, with [TFDO]o and [E-4c]o
initial concentrations both ranging (1.5-5.0) × 10-2 M.
structure for the dioxirane epoxidation of alkenes substituted
by electron-donor (ED) or electron-withdrawing (EW) substit-
uents. Their conclusion was that the electronic character
(electrophilic vs nucleophilic) of the dioxirane depends upon
the substituents on the CdC bond of the alkene. For the model
alkenes examined, EW substituents on the CdC bond (sCN
and sCHdO) designate the dioxirane to behave as a nucleo-
philic oxygen-transfer agent. Deubel cautioned that this peculiar
feature of dioxirane reactivity does not derive from the inherent
electronic properties of the dioxirane itself, but it becomes
manifest from the computed parameters of the transition
structures for the dioxirane/alkene pair.12
In our view, the interesting hypothesis of a possible electro-
philic/nucleophilic dichotomy for dioxirane reactivity, a novel
feature not recognized previously, demanded rigorous scrutiny.
Therefore, we set out to revisit the transformation above, starting
with determining the rates of dioxirane oxidation of a few
substituted cinnamonitriles aimed at establishing a Hammett
LFER. Obviously, this simple approach would allow one to
assess experimentally the electronic character of the oxidant in
the transformation at hand. Along with other pertinent observa-
tions concerning dioxirane reactivity, our results are reported
below.
(8) (a) Dankleff, M. A. P.; Curci, R.; Edwards, J. O.; Pyun, H.-Y. J. Am. Chem.
Soc. 1968, 90, 3209. (b) Curci, R.; Edwards, J. O. In Catalytic Oxidations
with H2O2 as Oxidant; Strukul, G., Ed.; Kluwer: Dodrecht, The Nether-
lands, 1992; Chapter 3, pp 45-95. See references therein.
(9) Bravo, A.; Fontana, F.; Fronza, G.; Minisci, F.; Zhao, L. J. Org. Chem.
1998, 63, 254.
(10) (a) Curci, R.; Dinoi, A.; Fusco, C.; Lillo, M. A. Tetrahedron Lett. 1996,
37, 249. (b) Curci, R.; D’Accolti, L.; Fusco, C. Tetrahedron Lett. 2001,
42, 7087.
(11) (a) Newcomb, M. Tetrahedron 1993, 49, 1151. (b) Adam, W.; Curci, R.;
D’Accolti, L.; Dinoi, A.; Fusco, C.; Gasparrini, F.; Kluge, R.; Paredes, R.;
Schulz, M.; Smerz, A. K.; Veloza, L. A.; Weinko¨etz, S.; Winde, R. Chem.
Eur. J. 1997, 3, 105.
(12) (a) Deubel, D. V. J. Org. Chem. 2001, 66, 3790. (b) Deubel, D. V.; Frenking
G.; Senn, H. M.; Sundermeyer, J. Chem. Commun. 2000, 2469. (c) Deubel,
D. V.; Sundermeyer, J.; Frenking, G. J. Am. Chem. Soc. 2000, 122, 10101.
Results and Discussion
Using both DDO and TFDO, we have examined the oxidation
of cinnamonitrile and of a few substituted cinnamonitriles (3-,
or 4-)X‚PhsCHdCHsCN carrying either EW or ED substit-
uents X on the phenyl moiety (Table 1). Prior to determining
the reaction rates, it was verified that at the given conditions
the compounds considered are cleanly transformed into their
corresponding epoxides in a stereospecific manner, with high
yield (>95%) and practically complete conversion, with the
9
1198 J. AM. CHEM. SOC. VOL. 130, NO. 4, 2008

Concerning the Reactivity of Dioxiranes
A R T I C L E S
notable exception of the TFDO oxidation of E or Z cinnamoni-
triles carrying the sOCH3 substituted phenyl moiety. In fact,
limited to this case, the oxidative scission of the electron-rich
aryl moiety by the powerful TFDO was found to prevail. This
finding is in agreement with our early reports concerning the
TFDO oxidation of catechols and of benzene itself to muconic
acids,13a as well as with more recent literature data.13b In most
cases, it was found that substrate epoxidation is quite fast,
especially when the powerful TFDO is the oxidant. Therefore,
we found it appropriate to determine the relative rates by kinetics
runs wherein the concentration decay of two substrates compet-
ing for excess dioxirane was determined at various times using
HPLC. In consideration of the fast rates and of the volatility of
TFDO, kinetic runs using this oxidant were conveniently carried
out at subambient temperature (5 °C).
Relative rates data are presented in Table 1. For unsubstituted
cinnamonitrile E-4c, the absolute epoxidation rate constants with
DDO and TFDO were also determined (footnotes d and e, Table
1). Data in Table 1 show that, with both DDO and TFDO, EW
groups in the substrate decrease the rate of epoxidation, whereas
the opposite is true with ED substituents. For the DDO runs, a
Hammett plot of log(kX/kH) vs σ+ for the E-alkenes (entries 1,
2, 4, 6, 7, and 8) provides an F+ value of -0.98 (r 0.991). A
similar plot of relative rates for the TFDO reactions yields F+
) -1.9 (r 0.991). Although just a few substituted cinnamoni-
triles were examined, these markedly negative F+ values point
to a TS wherein a partial positive charge is developing on the
alkene, leaving no doubt that the dioxiranes act as the electro-
phile in these epoxidations.
Kinetic studies by Baumstark and Vasquez allowed these
authors to estimate a Hammett F+ of -0.90 for the dioxirane
epoxidation of substituted styrenes; the greater reactivity of
Z-alkenes in the DDO epoxidation of E/Z pairs of alkenes was
ascribed to steric interactions developing in the TS.14a Also,
Murray reported the relative epoxidation rates of a series of
4-substituted (E)-ethylcinnamates by DDO, yielding a Hammett
F ) -1.53.14b
Data in Table 1 suggest that steric effects are relatively
unimportant here, since it is found that the cis alkenes are only
just a trifle more reactive than their trans isomers (cf., entries
2-5). Worthy of note is the finding that the measured absolute
rates (Table 1, footnotes d and e) show that TFDO is over 50
times more reactive than DDO in carrying out the transformation
at hand. And this, despite the fact that epoxidations with TFDO
were conveniently run at a 5 °C, i.e., well below those using
DDO (25 °C). Clearly, TFDO should be the reagent of choice
to perform the epoxidation of electron-poor alkenes efficiently.
Thus, the reactivity trend TFDO > DDO as electrophilic oxidant
is confirmed in the case at hand also. Actually, we found that
in the case of substrates more reluctant to undergo dioxirane
O-transfer, such as either simple or structurally complex
hydrocarbons, TFDO is more reactive than DDO by a factor of
almost 103.1,4d
Figure 1. Cyclic voltammetry of DDO 1a and of TFDO 1b (red trace),
both ca. 2 × 10-3 M in MeCN, at 20 °C. TBAPF6 (0.1 mol L-1) supporting
electrolyte, glassy carbon working electrode, platinum wire counter
electrode, SCE reference electrode, 200 mV s-1 scan rate. Wave peaks a
cat
and x are taken to correspond to the one-electron reduction potentials (Ep
)
of TFDO and of DDO, respectively.
mmetry (CV). Among the readily available electrochemical
methods, cyclic voltammetry (CV) techniques are the simplest
and the most convenient to use in organic solvents, offering
direct access to reduction potentials.15,16 Typically, the cyclic
voltammograms of most peroxide compounds exhibit irrevers-
ible behavior at sweep rates up to 200 mV s-1 and more. This
results in the absence of the anodic component on the return
potential sweep, largely owing to competition from fast follow-
up reactions of the metastable peroxide anion radicals.16
We ran the CVs of DDO and of TFDO in acetonitrile solvent
with0.1Mtetrabutylammoniumhexafluorophosphate(Bu4N+PF6-,
TBAP) as supporting electrolyte. In Figure 1 typical CV traces
are shown, which were recorded at the conditions given in the
footnote.
Carefully purged (Ar gas) DDO solutions gave two irrevers-
ible waves with peaks x and y located at -0.94 and at -2.1 V,
respectively (Figure 1). Control experiments allowed us to rule
•-
out that any of these waves are due to the O2/O2 couple,
residual acetone (also the reduction product of DDO), or
impurities. As found for other peroxides,16 it is likely that the
initial electron transfer to give rise to the radical-anion [(CH3)2-
CO2] •- is accompanied by the scission of the weak O-O bond,
forming a ring-opened radical-anion species such as •O-
C(CH3)2-O-;16 from the latter, the rapid formation of a second
reducible species should account for the wave with peak y. The
CV of TFDO presented a more complex case. Ketone-free
TFDO was employed in these experiments.3d However, oxygen
could not be completely removed by bubbling inert gas through
the solutions, since the procedure also depletes substantially this
highly volatile dioxirane species.3 Thus, interference by the O2/
•-
O2 couple cannot be avoided. As a result, the electrode
reduction of TFDO does not result in a “clean” process. It is
apparent that, ensuing the initial reduction to the radical anion
•
(peak a, + 0.17 V), hence to O-C(CH3)(CF3)-O-,17 a rapid
consecutive chemical reaction yields yet another metastable
Yet another experimental hint of the more powerful oxidizing
power of TFDO with respect to DDO came from the measure
of their electrode potential measurement Ep by cyclic volta-
(15) Bard, A. J.; Faulkner, L. R. Electrochemical Methods. Fundamentals and
Applications; Wiley: New York, 1980.
cat
(16) (a) Baron, R.; Darchen, A.; Hauchard, D. Electrochim. Acta 2006, 51, 1336.
(b) Workentin, M. S.; Maran, F.; Wayner, D. D. M. J. Am. Chem. Soc.
1995, 117, 2120. (c) Vasudevan, D. Bull. Electrochem. 2000, 16, 277. (d)
Bonchio, M; Conte, V.; DiFuria, F.; Modena, G.; Moro, S.; Carofiglio, T.;
Magno, F.; Pastore, P. Inorg. Chem. 1993, 32, 5797. See also references
therein.
(13) (a) Altamura, A.; Fusco, C.; D’Accolti, L.; Mello, R.; Prencipe, T.; Curci,
R. Tetrahedron Lett. 1991, 32, 5445. (b) Paquette, L. A.; Kreileinm, M.
M.; Bedore, M. W.; Friedrich, D. Org. Lett. 2005, 7, 4665.
(14) (a) Baumstark, A. L.; Vasquez, P. C. J. Org. Chem. 1988, 53, 3437. (b)
Murray, R. W.; Shiang, D. L. J. Chem. Soc., Perkin Trans. 2 1990, 349.
(17) Adam, W.; Asensio, G.; Curci, R.; Gonzale`s-Nun˜ez, M. E.; Mello, R. J.
Am. Chem. Soc. 1992, 114, 8345.
9
J. AM. CHEM. SOC. VOL. 130, NO. 4, 2008 1199

A R T I C L E S
Annese et al.
corresponding radical anions, i.e., R(CH3) CO2 (1a,b) + e- f
[R(CH3) CO2]•-. In fact, our computations for this process at
the B3LYP/6-311++G(2d,p) level, including a suitable solvent
model (see below), gave ∆E ) -28.9 and -30.4 kcal mol-1 in
acetone and acetonitrile, respectively, for DDO. In contrast, for
TFDO the computed values were ∆E ) -42.6 kcal mol-1 in
acetone and -43.9 kcal mol-1 in acetonitrile. Therefore, it is
reasonable to assume that the relatively large change monitored
in reduction energetics is reflected, at least in part, by the greater
propensity for electrophilic O-transfer to electron-donor sub-
strates manifested by TFDO with respect to DDO and other
classical peroxides.
Table 2. Cathodic Reduction Peak Potentials of Some Organic
Peroxides from Cyclic Voltammetrya
entry
peroxide
Epcat (V/SCE)
reference
1
2
3
4
5
6
7
TFDO (1b)
DDO (1a)
+0.17
-0.94
-1.26
-1.30
-1.44
-1.49
-2.70
this work
this work
16d
16d
16d
VO(O2)(Pic)(H2O)2
WO(O2)2(HMPT)
MoO(O2)2(HMPT)
CH3C(d)OsOBu-t
t-BuOsOBu-t
16a
16a
a
In acetonitrile solvent with TBAPF6 (0.1 M) supporting electrolyte
(entry 1 and 2), in DMF with 0.1 M TBABF4 supporting electrolyte for the
other peroxide species (entry 3 to 7); see also text and caption of Figure 1.
Theoretical studies support the enhanced reactivity of TFDO
with respect to DDO. Although the driving force of dioxirane
reactivity has customarily been ascribed to a relief of ring
strain,1-5 the recent reassessment of the strain energy of DDO
(SE ) 11 kcal mol-1) by Bach et al.20 suggests this should not
play a major role. Rather, dioxirane inclination to peroxide
O-transfer might just be due to the favorable enthalpy change
attending the formation of a strong CdO π-bond in its reduction
product, i.e., the parent ketone. Thus, as in the classic case
comparing epoxidations with trifluoroperacetic vs peracetic
acid,21 the increased reactivity of TFDO with respect to DDO
has been ascribed largely to the inductive effect exercised by
the CF3 group.7
electroactive species (peak b, ca. -0.28 V), which is in turn
rapidly consumed. The third shallow wave with its maximum
at ca. -0.9 V (peak c) appears in a region that is characteristic
•-
of the O2/O2 couple in acetonitrile or DMF;15,18 the absence
•-
of a reversible wave suggests that O2 (similarly to other
radicals)19 contributes to the overall chemical process of
dioxirane decomposition and it is quickly consumed.17 In fact,
we previously reported evidence concerning a rapid electron-
transfer chain reaction mediated by the superoxide ion for TFDO
decomposition.17 Finally, peak d (at ca. -2.0 V) might be
attributed to the reduction of one ending electroactive species
derived from dioxirane decomposition (most likely, 1,1,1-
trifluoroacetone).17 In fact, its depth was observed to increase
continuously with time at the expense of the others through the
repeated scans.
High-level theoretical work on dioxirane epoxidations of
representative alkenes has supported the higher reactivity of
TFDO. For instance, at the B3LYP/6-31+G(d,p) level of theory
the computed activation barriers (∆E*) for the epoxidation of
Obviously, elucidation of the detailed reaction mechanism
of dioxirane electrode reduction must await further experiments.
ethene with DDO and with TFDO are 17.7 and 11.1 kcal mol-1
,
respectively; at the QCISD(T)//QCISD/6-31+G(d,p) level the
computed ∆E values were 15.5 kcal mol-1 for DDO and 11.1
kcal mol-1 for TFDO.7,21a However, it is well recognized that
in the condensed phase these barriers could be significantly
lower.7,23
cat
Nonetheless, it is clear that the cathodic Ep data support the
electrophilic nature of these oxidants.1
Baron and co-workers investigated the voltammetry of a
variety of classic organic peroxides in DMF under comparable
conditions to those adopted for the dioxiranes at hand. For
simple dialkylperoxides and peresters, all yield irreversible broad
CV waves, and peaks are positioned over a wide potential range,
from-2.70to-0.59V/SCEfornonarylsubstitutedcompounds.16a
Besides, Bonchio et al. have reported the reduction potentials
of several “side-on” peroxo complexes of d° transition metals
in DMF, again determined by CV under conditions very similar
In Figure 2 we report a summary of our own computations
regarding the optimized transition structures of dioxirane
epoxidation of ethene and of acrylonitrile both in the gas phase
and in acetone solvent. Inspection of Figure 2 shows that
replacing one hydrogen atom with the cyano group in ethene
makes the transition structures nonsymmetrical, as expected.7,22
These data suggest that upon passing from a gas to condensed
phase (acetone solvent) the epoxidations rates should benefit
cat
to those mentioned above.16d A few representative cathodic Ep
values of the latter and those of dialkyl peroxides of choice are
listed in Table 2, in comparison to the analogous values
determined for dioxiranes.
by a decrease of activation barriers of some 1.5-2.0 kcal mol-1
,
except the case of TFDO reaction with acrylonitrile (TS d) where
there is practically no change in the computed ∆E*.
The data also show that the powerful TFDO is a more
effective O-transfer agent than DDO in its reaction with both
ethene and with acrylonitrile either in the gas phase or in
Inspection of Table 2 indicates that the dioxiranes rank among
cat
the most easily reducible peroxides. The Ep value recorded
for DDO (entry 2) is markedly less negative even than that for
the molybdenum side-on peroxometal complex (entry 5),
deemed to be the ultimate electrophilic epoxidation reagent.12c
It is worthy of note that the reduction potential estimated for
TFDO is astoundingly more positive even with respect to DDO
by over 1 V. The considerable ease with which TFDO undergoes
one-electron reduction is also pointed out by calculations
involving the conversion of the dioxiranes at hand into the
(20) (a) Bach, R. D.; Dmitrenko, O. J. Org. Chem. 2002, 67, 2588. (b) Bach,
R. D.; Dmitrenko, O. J. Org. Chem. 2002, 67, 3884.
(21) (a) Bach, R. D.; Dmitrenko, O.; Adam, W.; Schambony, S. J. Am. Chem.
Soc. 2003, 125, 924. (b) Estevez, C. M.; Dmitrenko, O.; Winter, J. E.;
Bach, R. D. J. Org. Chem. 2000, 65, 8629. See references therein. (c) Bach,
R. D. In DFG Research Report on Peroxide Chemistry: Mechanistic and
PreparatiVe Aspects of Oxygen Transfer; Adam, W., Saha-Mo¨ller, C. R.,
Eds.; Wiley-VCH: Weinheim, 2000; p 569. (d) Bach, R. D.; Glukhovtsev,
M. N.; Gonzales, C.; Marquez, M.; Este´vez, C. M.; Baboul, A. G.; Schlegel,
H. B. J. Phys. Chem. A 1997, 101, 6092. (e) Houk, K. N.; Liu, J.; DeMello,
N. C.; Condroski, K. R. J. Am. Chem. Soc. 1997, 119, 10147.
(22) Selc¸uki, C.; Aviyente, V. THEOCHEM 1999, 492, 165.
(23) Freccero, M.; Gandolfi, R.; Sarzi-Amade`, M.; Rastelli, A. (a) J. Org. Chem.
2002, 67, 8519; (b) J. Org. Chem. 2003, 68, 811; (c) J. Org. Chem. 2000,
65, 8948; (d) Tetrahedron 1998, 54, 12323; (e) J. Org. Chem. 2005, 70,
9573.
(18) For instance, see: (a) Guo, Z.; Lin, X. J. Electroanal. Chem. 2005, 576,
95. (b) Ortiz, M. E.; Nunez-Vergara, L. J.; Squella, J. A. J. Electroanal.
Chem. 2003, 549, 157. (c) Araki, T.; Kitaoka, H. Chem. Pharm. Bull. 2001,
49, 541. See also references therein.
(19) Dinoi, A.; Curci, R.; Carloni, P.; Damiani, E.; Stipa, P.; Greci, L. Eur. J.
Org. Chem. 1998, 871.
9
1200 J. AM. CHEM. SOC. VOL. 130, NO. 4, 2008

Concerning the Reactivity of Dioxiranes
A R T I C L E S
Figure 2. Transition structures for the epoxidation of ethene with DDO (a) and by TFDO (b), as compared to their analogue TSs for the epoxidation of
acrilonitrile, (c) and (d) respectively. All structures were optimized at the B3LYP/6-31+G(2d,p) level of theory, both in the gas phase and in solvent acetone.
Bond lengths for structures in solution are shown in parentheses. The fraction of charge transfer (CT) from alkene to dioxirane in acetone is given in electron
units.
acetone; in fact, a ∆∆E* of ca. -6 kcal mol-1 applies (cf., TS
a and b), whereas the computed activation barriers for the TFDO
epoxidation of acrylonitrile are lower than those for DDO by
about 4.2 and 2.8 kcal mol-1 in the gas phase and acetone,
respectively (cf., TS c and d). Notice that a difference of 2.8
kcal mol-1 in Ea would make TFDO epoxidations of acryloni-
trile in solvent acetone to occur faster than those with DDO by
a factor of over 102 at 25 °C. More important, the charge-transfer
(CT) figures computed for all transition structures leave no doubt
that the electron donation is from the alkene to the dioxirane.
This is in complete agreement with the vast majority of
theoretical and experimental literature reports.7
work,7,21d,e there is a certain amount of diradicaloid character
in the TS for DDO alkene epoxidation.
We believe that, in addition to precise computations of
transition state energies, an analysis of FMO interactions also
corroborates the higher reactivity found for TFDO. In Table 3
we report the energy levels for the significant MO of DDO and
of TFDO calculated at various levels of theory, along with the
relevant LUMO and HOMO for archetypical electron donors
such as ethene and dimethylsulfide.
It is seen that the three computation methods agree concerning
the relative ordering of energy levels of the given MOs.24 The
dioxirane LUMO lies in the molecular plane, and it has the
While long-lived radical intermediates could be ruled out on
the basis of both theoretical and experimental data, it should
be mentioned that, consistent with previously reported theoretical
(24) For an excellent discussion about the qualitative use of Kohn-Sham orbitals
to rationalize chemical phenomena, see: Stowasser, R.; Hoffmann, R. J.
Am. Chem. Soc. 1999, 121, 3414.
9
J. AM. CHEM. SOC. VOL. 130, NO. 4, 2008 1201

A R T I C L E S
Annese et al.
Table 3. Energy Levels (eV) Computed for the Significant
Molecular Orbitals (MO) of Dimethyldioxirane (1a) and of
Methyl(trifluoromethyl)dioxirane (1b)a as Compared with Those of
Performic Acid, Ethene, Acrylonitrile, and Dimethylsulfide
Chart 2. Relevant HOMO-LUMO Interaction for the Electrophilic
Oxygen Transfer from Dioxiranes 1a,b to Ethenea
a
Energies (eV) estimated by B3LYP/6-31G(d) calculations.
appears to be in line with experimental findings.25 Further
inspection of data in Table 3 (entries 1-3) suggests that, based
on LUMO energy levels, the TFDO oxidizing power is well
above that of both performic acid and DDO by ca. 23 kcal mol-1
(1.03 eV). This again agrees with the experimental findings for
reactions in the condensed phase, which strongly suggests that
TFDO is best tailored for carrying out the electrophilic epoxi-
dation of electron-poor substrates.4d,5
In conclusion, the claim of a “biphilic” character of dioxirane
reactivity12 could not be substantiated by experiments. In fact,
rate data herein unequivocally show that dioxiranes epoxidize
electrophilically even with alkenes carrying electron-withdraw-
ing groups. For epoxidation of electron deficient olefins, peracids
can be as effective as dioxiranes; in analogy to the well-known
Weitz-Scheffer epoxidation using alkaline hydrogen peroxide,26a
a two-step nucleophilic epoxidation avenue is open for the
reaction of peracids with electron-poor olefins.26 Thus, while
the electrophilic/nucleophilic dichotomy for peracids as epoxi-
dation reagents is amply supported experimentally, for dioxiranes
this does not apply.
a
Values for TFDO (1b) are given in boldface. bB3LYP/6-31G(d). cHF/
d
6-31G(d). Peracid σ*(O-O); the lower lying peracid LUMO (-0.63 eV,
DFT) is a π* pertaining to the carbonyl moiety.
proper symmetry to interact with the alkene π bond.4f,7,12 By
analogy to the typical HOMO-LUMO interaction drawn in SN2-
type displacements on peroxide O-O bonds, the dominant
orbital interaction involved in nucleophilic attack on the
dioxirane should be the alkene π molecular orbital with the
empty dioxirane O-O σ* Walsh-type orbital.7
Thus, while the dioxirane oxygen lone pair’s interaction with
the alkene LUMO favor a spiro geometry in the TS, the alkene
HOMO interacts in a stabilizing manner with the dioxirane
LUMO (Chart 2). The same should hold true for powerful σ-
or n-electron donor nucleophiles, such as dimethylsulfide.
The diagram in Chart 2 illustrates the fact that the TFDO
epoxidations benefit by a lower HOMO (alkene)/LUMO (di-
oxirane) energy gap with respect to that of DDO.
Based on the MO energies given in Table 3, the same applies
if the dioxirane epoxidation of the electron-poor acrylonitrile
is considered. The major stabilization occurring on the way to
the activated complex should give rise to a sizable difference
in TS energies, hence to energy barriers of epoxidation
significantly lower for TFDO as compared to DDO. Within this
framework, it is seen that the lower lying HOMO of acrylonitrile
(-7.87 eV) makes it a substrate substantially more reluctant
than ethene to epoxidation by both DDO and TFD. This ordering
Along with high-level calculations concerning TS structure
and energies, we deem that a simple FMO analysis is also useful
(25) It should be mentioned that the energy barriers both from experiments and
from computations should diminish markedly on going from gas to
condensed phase. For instance, an experimental Ea of 8.0 kcal mol-1 at 25
°C (∆H* of 7.4 kcal mol-1) has been measured for the DDO epoxidation
of cyclohexene in an acetone solvent (ref 4c). The Ea values of 11.4 and
10.7 kcal mol-1 have been estimated for the DDO epoxidation of
cyclohexene in CHCl3 and acetone, respectively, using the COSMO solvent
mode [B3LYP/6-311+G(d,p)] (ref 7). The increased reactivity in polar
solvents might be attributed to enhanced polarization of the TS relative to
reactants; theoretical solvation studies are consistent with these observations
(Jenson, C.; Liu, J.; Houk, K. N.; Jorgensen, W. L. J. Am. Chem. Soc.
1997, 119, 12982); a systematic study of the effects of solvation and
hydrogen bonding was also carried out by Gandolfi et al. at the B3LYP/
6-31G(d) level (ref 23)
(26) (a) Weitz, E.; Scheffer, A. Chem. Ber. 1921, 54, 2327. (b) Curci, R.; DiFuria,
F.; Meneghin, M. Gazz. Chim. Ital. 1978, 108, 123. (c) Curci, R.; DiFuria,
F. Tetrahedron Lett. 1974, 4085. (d) Adam,W.; Rao, P.; Degen, H.-G.;
Saha-Mo¨ller, C.; Chantu, R. J. Am. Chem. Soc. 2000, 122, 5654.
9
1202 J. AM. CHEM. SOC. VOL. 130, NO. 4, 2008

Concerning the Reactivity of Dioxiranes
A R T I C L E S
a general procedure reported below and purified by column chroma-
tography (purity g98%, HPLC). The following procedures are repre-
sentative for the epoxidations of the alkenes above using dioxiranes.
Epoxidation of (E)-Cinnamonitrile (4c) with Dimethyldioxirane
(1a). To a stirred solution of (E)-4c (19.4 mg, 0.15 mmol) in acetone
(1 mL) at 25 °C, containing 0.1 M PFMB internal standard, was added
in one portion a standardized cold solution of DDO (1a) in acetone
(33 mL, 0.09 M, 3.00 mmol). The reaction progress was monitored by
HPLC (acetonitrile/methanol/water; flow rate: 1.0-1.3 mL/min; detec-
tor, λ ) 220 nm). Upon completion of the reaction (30 h, conv. >90%),
the solvent was removed under reduced pressure and the residue was
purified by column chromatography. The epoxide (E)-5c (yield >95%,
purity g98% HPLC) gave physical and spectral data identical to those
of an authentic sample.30a,b
Epoxidation of (E)-Cinnamonitrile (4c) with Methyl(trifluorom-
ethyl)dioxirane (1b). To a stirred solution of (E)-4c (12 mg, 0.093
mmol) in acetone (4 mL) at 0 °C, containing 0.05 M PFMB internal
standard, was added in one portion a standardized cold solution of
TFDO (1b) in acetone (0.3 mL, 0.88 M, 0.28 mmol). The reaction
progress was monitored by HPLC. Upon completion of the reaction
(30 min, conv. >98%), the solvent was removed under reduced pressure
and the residue was purified by column chromatography. The epoxide
(E)-5c (yield >95%, purity g98% HPLC) gave physical and spectral
data identical to those of an authentic sample.30a,b
(E)-3-(4-Methoxy-phenyl)-oxirane-2-carbonitrile (5a): colorless oil,
bp 155-158 °C/6 mmHg (lit.30a 105-114 °C/0.28 mmHg). (E)-3-(4-
Methyl-phenyl)-oxirane-2-carbonitrile (5b):30a colorless oil. (Z)-3-(4-
Methyl-phenyl)-oxirane-2-carbonitrile (5b′):30a colorless oil. (E)-3-
Phenyl-oxirane-2-carbonitrile (5c):30a colorless liquid, bp 135-137 °C/
16 mmHg (lit.30b 133-134 °C/16 mmHg). (Z)-3-Phenyl-oxirane-2-
carbonitrile (5c′):30a white solid, mp 55.5-56 °C (lit.30c 55-57 °C).
(E)-3-(4-Chloro-phenyl)-oxirane-2-carbonitrile (5d):30a white solid, mp
67-68 °C (lit.30a 68-69 °C). (E)-3-(3-Chloro-phenyl)-oxirane-2-
carbonitrile (5e):31 white solid, mp 59-61 °C. (E)-3-(3-Nitro-phenyl)-
oxirane-2-carbonitrile (5f):32 white solid, mp 118-120 °C.
Kinetics. RelatiVe rates were determined under pseudo-first-order
conditions, with the dioxirane in large excess (from 20- to 80-fold)
over initial concentrations of the alkenes. The procedure used was based
on the amount of starting material consumed during the kinetic run.
At zero time a thermostatted aliquot (20 to 2 mL) of standard acetone
solution of dioxirane 1a (0.07-0.09 M) or of dioxirane 1b (0.6 - 0.9
M, in TFP) was added to an acetone solution (also thermostatted)
containing the two given alkene substrates (A and B, usually 0.01-
0.04 M for 1a and 0.003-0.008 M for 1b) competing for the oxidant,
as well as the internal standard PFMB. Over the initial 15-50%
reaction, aliquots (0.2-1 mL) were withdrawn periodically and
quenched with 0.3 mL of 0.5 M p-MeC6H4SMe in acetone. The relative
rates were determined from the concentration percentage of alkenes
remaining, [A]/[A]0 and [B]/[B]0. The chromatogram area of the alkenes
was corrected for the detector response, determined separately for each
cinnamonitrile substrate using standard solutions of the alkene contain-
ing the internal standard. Relative rate coefficients were estimated as
kA/kB ) log([A]/[A]0)/log([B]/[B]0).
in explaining the high reactivity of dioxiranes as electrophilic
oxidants, based on the key stabilizing interaction nucleophile
HOMO with dioxirane LUMO. In fact, both the LUMO of both
the DDO (-0.43 eV) and of the TFDO (-1.43 eV) are
energetically much more accessible to the nucleophile electron
pair than the corresponding σ* LUMO of classical peroxides.27
Experimental Section
Materials and Methods. Boiling points and melting points were
not corrected. The HPLC analyses were run using a Supelcosil
ABZ+plus, 5 µm column (15 cm × 4.6 mm id). Column chromatog-
raphy was performed using silicagel (230-400 mesh), eluent n-hexane/
CH2Cl2; GC/MS experiments were run using a ZB-1 column (30 m ×
0.25 µm id), the MS detector in EI mode (70 eV). The 1H NMR spectra
were recorded on a 500 MHz or 400 MHz instrument; resonances are
referenced to residual isotopic impurity (7.26 ppm) of CDCl3 solvent
and/or to TMS. The 13C NMR spectral data (125.76 or 100 MHz) are
referred to the middle peak of CDCl3 solvent (77.0 ppm). FTIR spectra
were run on samples in KBr pellets or films (KBr plates).
Commercial 1,1,1-trifluoro-2-propanone (TFP) (bp 22 °C) was
purified by fractional distillation over granular P2O5, stored over 5 Å
molecular sieves, and routinely redistilled prior to use. Acetone and
other solvents were purified by standard methods. Caroat triple salt
2KHSO5‚KHSO4‚K2SO4 (a gift from Peroxid-Chemie, Degussa, Ger-
many) was our source of potassium peroxomonosulfate employed in
the synthesis of dioxiranes.3 Commercial tetrabutylammonium hexaflu-
orophosphate (Bu4N+PF6-, purity >99%) was used in the cyclic
voltammetry experiments.
Commercial (E)-p-methoxycinnamonitrile (4a), (E)-cinnamonitrile
(4c), and (E)-p-chlorocinnamonitrile (4d) were further purified to >98%
(HPLC) by standard techniques. (E)-p-Methyl-cinnamonitrile (4b),28a
(E)-m-chlorocinnamonitrile (4e),28a and (E)-m-nitrocinnamonitrile (4f)28a
were synthesized by standard procedures upon conversion of com-
mercially available cinnamic acids into corresponding amides, followed
by treatment of the latter with P2O5; column chromatography afforded
the nitriles above at g98% purity (HPLC). The (Z)-p-methyl-cinna-
monitrile (4b′)28b and (Z)-cinnamonitrile (4c′)28b were obtained upon
photoisomerization28a of the corresponding trans-stereoisomer and
purified by column chromatography. All substrates gave physical
constants and spectra (NMR, FTIR, GC-MS) in full agreement with
literature.28
Employed as an inert internal standard in HPLC runs, methyl
2,3,4,5,6-pentafluorobenzoate (C6F5CO2Me, PFMB)29 was obtained as
a colorless liquid from the corresponding acid upon esterification with
MeOH/BF3;29b it was purified by fractional distillation (bp: 44 °C/4
mmHg,29a purity g98%, HPLC). Commercial tetrabutylammonium
hexafluorophosphate (Bu4N+PF6-, TBAP) (purity 98%) was employed
as the supporting electrolyte in cyclic voltammetry runs.
Solutions of 0.8-1.0 M methyl(trifluoromethyl)dioxirane (1b) in
1,1,1-trifluoropropanone (TFP) or in CCl4 (ketone-free)3d and solutions
of 0.08-0.16 M dimethyldioxirane (1a) in acetone were made available
upon adopting procedures, equipment, and precautions already reported
in detail.3 The solutions were assayed for dioxirane content by
iodometry and/or using a GC method after quenching the oxidant with
excess p-MeC6H4SMe, which is rapidly and quantitatively oxidized to
the corresponding sulfoxide.3 Authentic samples of the epoxides 5a-f
were synthesized upon reaction with dioxiranes 1a or 1b according to
In each experiment, at least two values were measured at various
reaction times; data from two or more independent runs were averaged
(estimated error: e10%).
Absolute rates of DDO oxidations were determined at 25.0 ( 0.1
°C under pseudo-first-order conditions, with the alkene substrate in
large excess (from 20- to 50-fold) over the dioxirane initial concentra-
(27) This work, DFT computations; values optimized using B3LYP/6-31G(d)
data.
(28) (a) Shim, S. C.; Yoon, S. K. Bull. Korean Chem. Soc. 1981, 2, 147. (b)
Clovis, P.; de Azevedo Mello, P.; Pavao da Cha˜gas, R. J. Organomet. Chem.
2006, 691, 2335. (c) Hamza, K.; Abu-Reziq, R.; Avinr, D.; Blum, J. Org.
Lett. 2004, 6, 925. (d) Butt, G.; Topsom, R. D. Spectrochim. Acta 1982,
38A, 301. (e) Happer, D. A. R.; Steenson, B. E. J. Chem. Soc., Perkin
Trans. 2 1988, 19.
(30) (a) Svoboda, J.; Kocfeldova´, Z.; Palecˇek, J. Coll. Czech. Chem. Commun.
1988, 53, 822. (b) Stork, W. G.; Worall, W. S.; Pappas, J. J. J. Am. Chem.
Soc. 1960, 82, 4315. (c) Jonczyk, A.; Fedorynoky; M.; Makosza, M.
Tetrahedron Lett. 1972; 23, 2395.
(31) Wang, M.; Lin, S.; Liu, C.; Zheng, Q.; Li, J. J. Org. Chem. 2003, 68,
4570.
(29) (a) Strazzolini, P.; Verardo, G.; Giumanini, A. G. J. Org. Chem. 1988, 53,
3321. (b) AcKmam, R. G. J. Am. Oil Chem. Soc. 1998, 75, 541.
(32) Taylor, E. C.; Marynoff, C. A.; Skotnicki, J. S. J. Org. Chem. 1980, 45,
2512.
9
J. AM. CHEM. SOC. VOL. 130, NO. 4, 2008 1203

A R T I C L E S
Annese et al.
tion. At zero time, an aliquot (10-20 mL) of a thermostated solution
of dioxirane 1a ((0.55-0.81) × 10-2 M) was added to 180 mL of (4.6-
7.2) × 10-2 M solution of the alkene in acetone (also thermostated).
The decrease of dioxirane concentration with time was measured by
iodometric titrations of aliquots (10 or 20 mL) periodically withdrawn
from the reaction flask. Plots log[a0/(a0 - x)] vs time, linear to over
50% reaction, yielded the pseudo-first-order k1 (s-1) value; from this,
and the known alkene initial concentration, k2 (M-1 s-1) values were
estimated. Data agreeing within (10% were averaged.
Absolute rates of TFDO oxidations were determined at 5.0 ( 0.1
°C under second-order conditions, with initial concentrations of both
reagents ranging 0.015-0.05 M. At zero time, an aliquot (5 mL) of a
dioxirane 1b thermostated solution (0.240-0.280 M) was added to a
0.05-0.06 M solution of the alkene in acetone. The decrease in the
dioxirane concentration was determined by iodometry of aliquots (5
mL or 10 mL) periodically withdrawn from the reaction solution.
Integrated second-order rate plots log[b(a - x)/a(b - x)] vs time were
linear to over 70% reaction, yielding the k2 (M-1 s-1) value. Based on
two or more independent runs, data agreeing within (10% were
averaged.
Cyclic Voltammetry. The CV experiments were run using an
EG&G-PAR 273 potentiostat/galvanostat controlled by a universal
programmer. A plotter was used to record the current voltage output
at 200 mV/s sweep rate. A conventional three-electrode system
consisting of a glassy carbon working electrode (od 3 mm), a platinum
wire counter electrode, and a saturated calomel reference electrode
(SCE) were employed. All electrochemical measurements were carried
out at 20 °C in the same solvent, i.e., dry acetonitrile, thus avoiding
problems related to liquid-junction potentials. Commercial tetrabuty-
lammonium hexafluorophosphate (Bu4N+PF6-, TBAP) ca. 0.1 M served
as the supporting electrolyte. Aliquots of mother solutions that was
either 0.08 M DDO in acetone or 0.2 M TFDO in CCl43d were used to
prepare ca. 2 × 10-3 M dioxirane solutions in dry acetonitrile.
The DDO solutions were carefully deoxygenated by bubbling an
Ar gas stream through them for at least 10 min prior to recording of
the voltammetric data; instead, TFDO solutions could be only partially
freed by atmospheric oxygen by the mentioned procedure since it was
found that the inert gas stream depletes substantially the solution of
this volatile dioxirane species.3 The potential scan range was from -3.0
to +1.0 V vs SCE, allowing us to estimate the electron reduction
potentials (Epcat) of TFDO and of DDO as shown in Figure 1.
Computational Methods. All calculations were carried out using
the Gaussian 2003 suite of programs.33 High-level computations by
Bach et al. (as well as by other authors) have demonstrated that the
B3LYP variant of DFT calculations is a useful method and accurate
enough for studying dioxirane and peroxyacid epoxidations;7 accord-
ingly, we employed the B3LYP method for full geometry optimizations
both in the gas phase and in solution. Solvent effects were evaluated
with the self-consistent reaction field (SCRF) theory using the PCM-
united atom topological model (UAHF, radii of interlocking spheres
obtained through the Barone’s HF parametrization) as implemented in
the C.02 version of Gaussian 2003. The dioxirane reduction to the give
the corresponding radical anion was studied by using the 6-311++G-
(2d,p) basis set of triple ú quality with diffuse and polarization functions,
which is appropriate in describing molecules bearing functionalities
rich in lone pairs, such as the peroxide moiety7,23 as well as anionic
systems. At first, a geometry optimization of DDO and TFDO in acetone
or acetonitrile was performed; then, the dioxirane radical anion energy
was estimated by single-point calculations on the geometries attained
in each of the solvents above. Regarding epoxidations of ethene or
acrylonitrile either by DDO or by TFDO, the somewhat smaller basis
set 6-31+G(2d,p) was employed; in fact, we recently reported that the
latter computational method is quite appropriate to describe peracid
epoxidations, nicely reproducing the results achievable using much
larger basis sets.23e All stationary points were confirmed by vibrational
frequency analysis (i.e., one imaginary frequency was found for all
TSs). The two gas-phase TSs relative to the reaction of ethene with
DDO and TFDO were found stable with the keyword stable)opt.
Concerning the case of the TSs of reactions of DDO and of TFDO
with acrylonitrile in the gas phase, the stability test revealed a small
RHF-UHF instability (0.6 kcal/mol and S2 ) 0.25 for DDO’s TS; 0.34
kcal/mol and S2 ) 0.19 for TFDO’s TS), which might suggest a small
diradical character for these transition structures. For the reaction of
DDO with acrylonitrile, we located two TSs (respectively, exo and
endo), of which just the exo one (carrying the cyano group on the side
opposite to the dioxirane alkyl groups) is given in Figure 2. Actually,
the two TSs above exhibit the same energy in acetone. Similarly, of
the four TSs located in the reaction of TFDO with acrylonitrile, just
the exo,anti structure is presented; again, the energies of the TSs omitted
were found to be quite close to that of the TS represented in Figure 2.
Charge transfer (CT) data were estimated from NPA charges. CHELPG
calculations gave slightly different values, confirming electron donation
from the alkenes to dioxiranes.
Acknowledgment. Partial support of this work by the
National Research Council of Italy (CNR) is gratefully ac-
knowledged. Thanks are due to professor F. Magno and his staff
(Department of Analytical Chemistry, University of Padova,
Italy) for help in running and interpreting the cyclic voltammetry
experiments. We are pleased to dedicate this work to professor
F. Naso (University of Bari, Italy) and to professor W. Adam
(Emeritus, University of Puerto Rico) on the occasion of their
70th birthday.
Supporting Information Available: Experimental details,
supplemental characterization data, and sample HPLC data for
substrates and products. Energy of dioxiranes and dioxirane
radical anions at the B3LYP/6-311++G(2d,p) level (acetone
and acetonitrile solution), energy and Cartesian coordinates of
all TSs at the B3LYP/6-31+G(2d,p) level (gas phase and
solution). Imaginary frequency for all TSs. Complete ref 33.
This material is available free of charge via the Internet at
http://pubs.acs.org.
(33) Frisch, M. J. et al. Gaussian 03, Revision C. 02; Gaussian, Inc.: Pittsburgh,
PA, 2003. For full citation, see Supporting Information.
JA075068U
9
1204 J. AM. CHEM. SOC. VOL. 130, NO. 4, 2008