Singlet oxygen production in the reaction of superoxide with organic peroxides

Article abstract of DOI:10.1021/jo051736n

A selective chemiluminescent probe for singlet oxygen has been employed to detect and quantify singlet oxygen in the reactions of superoxide with organic peroxides. The production of singlet oxygen has been quantified in the reaction of superoxide with benzoyl peroxide (BP). No singlet oxygen was detected in the reactions of superoxide with cumyl peroxide, tert-butyl peroxide, or tert-butyl hydroperoxide. On the basis of these results and on the temperature dependence of the reaction, we proposed a mechanism for singlet oxygen formation in the reaction of superoxide with BP.

Full text of DOI:10.1021/jo051736n

SCHEME 1
Singlet Oxygen Production in the Reaction of
Superoxide with Organic Peroxides
Laura A. MacManus-Spencer, Betsy L. Edhlund, and
Kristopher McNeill*
acids.13-17 Danen and Arudi, using chemical 1O2 traps, reported
high yields of transformation products (ca. 100% combined
yield) in the reactions of benzoyl and lauroyl peroxides with
O2•-, which were taken as evidence of efficient 1O2 production,
although they acknowledged the potential for false positives as
a result of other oxidation pathways.1 They also proposed a
UniVersity of Minnesota, Department of Chemistry,
207 Pleasant Street SE, Minneapolis, Minnesota 55455
mcneill@chem.umn.edu
ReceiVed August 17, 2005
•-
stoichiometry for the reaction, in which 2 equiv of O2 react
with an acyl peroxide, leading to 2 equiv of dioxygen (a fraction
1
as O2) and two carboxylate anions (Scheme 1).1
Khan obtained strong qualitative evidence of the production
of 1O2 in the reaction of benzoyl peroxide (BP) with O2•- from
the direct detection of its characteristic 1270-nm phosphorescent
•-
signal.2 We recently confirmed the production of 1O2 in the O2
/
BP system using a chemiluminescence-based detection method.3
In this contribution, we quantify the yield of 1O2 in this reaction
and expand the scope of the study to include alkyl peroxides
(tert-butyl and cumyl peroxides) and a hydroperoxide (tert-butyl
hydroperoxide).
A selective chemiluminescent probe for singlet oxygen has
been employed to detect and quantify singlet oxygen in the
reactions of superoxide with organic peroxides. The produc-
tion of singlet oxygen has been quantified in the reaction of
superoxide with benzoyl peroxide (BP). No singlet oxygen
was detected in the reactions of superoxide with cumyl
peroxide, tert-butyl peroxide, or tert-butyl hydroperoxide.
On the basis of these results and on the temperature
dependence of the reaction, we proposed a mechanism for
singlet oxygen formation in the reaction of superoxide with
BP.
1
To quantify the production of O2 in these reactions, we
applied a recently developed method in which a stable dioxetane
1
precursor is used to selectively detect O2 in the presence of
other reactive oxygen species (ROS). This detection method is
based on thermally stable spiroadamantylidene-substituted di-
oxetanes, initially developed by Schaap et al.18-20 and subse-
quently studied by others.21-24 In this detection scheme, 1O2 is
trapped by a vinyl ether probe (1) in the form of a stable
dioxetane (2) (Scheme 2). The trapped 1O2 is quantified by the
addition of a chemical trigger, tetra-n-butylammonium fluoride
(TBAF), to induce a chemically initiated electron exchange
luminescence (CIEEL) process by which dioxetane 2 decom-
poses to give a chemiluminescent (CL) signal.
The reaction of acyl peroxides with superoxide (O2•-) has
1
been shown to generate O2.1-3 Interest in this reaction lies in
its potential to be a source of 1O2 in biological systems, where
•- 4-7
both O2
and peroxides8 are formed. Singlet oxygen is a
We have reported the selectivity of this detection method for
known cytotoxic species, oxidizing a variety of biological
1O2 over O2•-,3 a problematic interferent in the chemilumines-
substrates, such as proteins,9 amino acids,10-12 and nucleic
1
cence-based detection of O2, and have also used this method
* Corresponding author. Phone: 612-625-0781. Fax: 612-626-4751.
(1) Danen, W. C.; Arudi, R. L. J. Am. Chem. Soc. 1978, 100, 3944-
3945.
(2) Khan, A. U.; Kasha, M. Proc. Natl. Acad. Sci. U.S.A. 1994, 91,
12365-12367.
(3) MacManus-Spencer, L. A.; Latch, D. E.; Kroncke, K. M.; McNeill,
K. Anal. Chem. 2005, 77, 1200-1205.
(13) Canva, J. J.; Balny, C. Int. J. Radiat. Phys. Chem. 1971, 3, 451-
455.
(14) Clagett, D. C.; Galen, T. J. Arch. Biochem. Biophys. 1971, 146,
196-201.
(15) Hallett, F. R.; Hallett, B. P.; Snipes, W. Biophys. J. 1970, 10, 305-
315.
(4) McCord, J. M.; Fridovich, I. J. Biol. Chem. 1968, 243, 5753-5760.
(5) Massey, V.; Strickland, S.; Mayhew, S. G.; Howell, L. G.; Engel, P.
C.; Matthews, R. G.; Schuman, M.; Sullivan, P. A. Biochem. Biophys. Res.
Commun. 1969, 36, 891-897.
(16) Rosenthal, I.; Pitts, J. N., Jr. Biophys. J. 1971, 11, 963-966.
(17) Sysak, P. K.; Foote, C. S.; Ching, T.-Y. Photochem. Photobiol. 1977,
26, 19-27.
(18) Schaap, A. P.; Chen, T. S.; Handley, R. S.; DeSilva, R.; Giri, P. P.
Tetrahedron Lett. 1987, 28, 1155-1158.
(6) Knowles, P. F.; Gibson, J. F.; Pick, F. M.; Bray, R. C. Biochem. J.
1969, 111, 53-58.
(19) Schaap, A. P.; Handley, R. S.; Giri, B. P. Tetrahedron Lett. 1987,
28, 935-938.
(7) Bernacchia, A.; Biondi, A.; Genova, M. L.; Lenaz, G.; Falasca, A.
Toxicol. Mech. Methods 2004, 14, 25-30.
(20) Schaap, A. P.; Sandison, M. D.; Handley, R. S. Tetrahedron Lett.
1987, 28, 1159-1162.
(8) Gutteridge, J. M. C.; Halliwell, B. Trends Biochem. Sci 1990, 15,
129-135.
(21) Adam, W.; Bronstein, I.; Edwards, B.; Engel, T.; Reinhardt, D.;
Schneider, F. W.; Trofimov, A. V.; Vasil’ev, R. F. J. Am. Chem. Soc. 1996,
118, 10400-10407.
(9) Davies, M. J. Biochem. Biophys. Res. Commun. 2003, 305, 761-
770.
(10) Matheson, I. B. C.; Lee, J. Photochem. Photobiol. 1979, 29, 879-
881.
(22) Adam, W.; Fell, R.; Schulz, M. H. Tetrahedron 1993, 49, 2227-
2238.
(11) Michaeli, A.; Feitelson, J. Photochem. Photobiol. 1994, 59, 284-
(23) Matsumoto, M.; Watanabe, N.; Shiono, T.; Suganuma, H.; Mat-
subara, J. Tetrahedron Lett. 1997, 38, 5825-5828.
(24) Watanabe, N.; Suganuma, H.; Kobayashi, H.; Mutoh, H.; Katao,
Y.; Matsumoto, M. Tetrahedron 1999, 55, 4287-4298.
289.
(12) Nilsson, R.; Merkel, P. B.; Kearns, D. R. Photochem. Photobiol.
1972, 16, 117-124.
10.1021/jo051736n CCC: $33.50 © 2006 American Chemical Society
Published on Web 12/23/2005
796
J. Org. Chem. 2006, 71, 796-799

peroxides. No evidence for the formation of 1O2 in the reactions
•-
of O2 with either CP or TBP was found (Figure 1), though
the reactions were carried out under conditions (O2•- concentra-
tion, solvent, and temperature) similar to those in the BP
reactions. Shown in Figure 1 are data from reactions in which
either CP or TBP was present at 1, 5, and 10 mM. Similarly,
no evidence of 1O2 formation was found in the reaction of O2
•-
with tert-butyl hydroperoxide (TBHP) (Figure 1).
Benzoyl Peroxide: Kinetics and 1O2 Yield. The formation
•-
of dioxetane 2 in the reaction of BP with O2 in the presence
of probe 1 was fit well by a second-order kinetic model
incorporating the stoichiometry shown in Scheme 1 (eq 1).
FIGURE 1. Formation of dioxetane 2 over time in the reaction of 2
mM superoxide with 0.71 mM benzoyl peroxide in toluene at 25 °C
(circles). Error bars represent one standard deviation from the average
of four measurements. Shown for comparison are the results of
experiments in which benzoyl peroxide was replaced with 1, 5, and 10
mM cumyl peroxide (squares), 1, 5, and 10 mM tert-butyl peroxide
(triangles), or 10 mM tert-butyl hydroperoxide (diamonds). Note that
data at all concentrations are shown and are overlapping.
∆0[BP]0
[2]t ) R × f × [BP]0 -
(1)
-
[O2 ]0e∆ kt - 2[BP]0
•
0
(
)
Here [2]t is the concentration of dioxetane 2 at time t, ∆0 )
[O2•-]0 - 2[BP]0, R is the yield of 1O2, and f is the fraction of
the O2 that is trapped by probe 1 (eq 2):
1
krxn,1[1]
SCHEME 2
f )
(2)
ksolv + kDTPA[DTPA] + krxn,1[1]
where krxn,1 is the rate constant for reaction of 1 with 1O2, while
1
ksolv and kDTPA are the rate constants for quenching of O2 by
the solvent and DTPA, respectively.
The second-order kinetics observed in the case of the reaction
•-
•-
of O2 with BP enabled the estimation of the initial O2
concentration, [O2•-]0, from kinetic data obtained in this study.
Using global fitting software, we simultaneously fit several sets
of data obtained under identical conditions to eq 1, with the
value for [O2•-]0 shared among the data sets. Using this method,
we obtained a value of 2.0 (( 0.2) mM for [O2•-]0. Because
stock solutions of KO2 in toluene were always prepared
identically, this value for [O2•-]0 was used throughout the studies
described here.
1
•-
to quantify the production of O2 in the reaction of O2 with
hydrogen peroxide.25 Here we report that a significant yield of
1O2 is only observed in the reaction of O2 with BP and not
with the alkyl peroxides or hydroperoxide tested. We also report
the temperature-dependent reaction kinetics for the case of O2
•-
1
•-
The yield of O2 from the reaction of 0.71 mM BP with 2
mM O2•- was calculated from the kinetic data fit and the known
1O2 trapping fraction, f. Given a rate constant for the reaction
of probe 1 with 1O2 (4.5 × 105 M-1 s-1, measured in methylene
chloride),3 the rate constant for deactivation of 1O2 by the solvent
(ksolv,toluene ) 3.4 × 104 s-1),27 the rate constant for quenching
of 1O2 by DTPA (1.0 × 107 M-1 s-1),3 and with 100 µM 1 and
4 mM DTPA, the quantity of 1O2 produced that is trapped under
these conditions is 0.06%. Despite the small trapping yield, the
sensitivity of CL detection enables the use of this detection
scheme. On the basis of the kinetic data fit and the known
and BP.
Reaction of Organic Peroxides with Superoxide: General
Observations. The reaction of O2 (KO2, 2 mM, solubilized
•-
with 18-crown-6 ether) with BP was carried out in toluene in
the presence of 1 (100 µM) and diethylenetriaminepentaacetic
acid (DTPA, 4 mM, added to bind trace metal impurities).
Kinetic time points were obtained by sampling aliquots of the
reaction mixture and quenching the reaction by dilution in
acetonitrile. The formation of dioxetane 2 over time was
monitored by measuring its CL signal, triggered by TBAF-
induced desilylation (Figure 1).
1
1O2 trapping fraction, the average yield of O2 at 25 °C is
The identity of 1O2 as the ROS responsible for the formation
of dioxetane 2 in this reaction was supported by the fact that
neither O2 nor BP alone elicited a signal from 1. The O2
yield was predictably enhanced in 85% deuterated toluene, in
which the lifetime of O2 is 10 times longer than in toluene,
and completely depressed in the presence of 520 µM 1,4-
diazabicyclo[2.2.2]octane (DABCO),26 a known 1O2 quencher.3
Two similar peroxides, cumyl peroxide (CP) and tert-butyl
peroxide (TBP), were also investigated to explore the possibility
5 (( 1)%. The observed second-order rate constant for the
reaction is 0.3 (( 0.2) M-1 s-1
.
•-
•-
1
The second-order kinetics in the reaction of BP with O2
were confirmed by following the disappearance of BP in the
•-
1
presence of O2 in toluene at 22 °C over 75 min (Figure 2).
Over this time period, there was a 14% decrease in the BP
concentration. Under these conditions, a bimolecular rate
constant of 0.8 (( 0.2) M-1 s-1 was observed.
Benzoyl Peroxide: Temperature Dependence. To probe
the mechanism of 1O2 production, the temperature dependence
•-
of a general mechanism for the reaction of O2 with organic
•-
of the reaction of O2 with BP was examined. The reaction
(25) MacManus-Spencer, L. A.; McNeill, K. J. Am. Chem. Soc. 2005,
127, 8954-8955.
(26) Ouannes, C.; Wilson, T. J. Am. Chem. Soc. 1968, 90, 6527-6528.
(27) Ogilby, P. R.; Kristiansen, M.; Clough, R. L. Macromolecules 1990,
23, 2698-2704.
J. Org. Chem, Vol. 71, No. 2, 2006 797

SCHEME 4
SCHEME 5
FIGURE 2. Disappearance of benzoyl peroxide over time in the
reaction of 0.2 mM superoxide with 0.71 mM BP in toluene at 22 °C.
The data are fit to a second-order exponential decay (R2 ) 0.93).
first step is assumed to be the rate-determining step because
the data are fit well by a mixed second-order kinetic model
incorporating first-order dependence on each of the individual
reactants. If the first step were not the rate-determining step,
kinetics indicative of a formal termolecular reaction should be
observed. Thus, it is proposed that the calculated activation
FIGURE 3. Temperature dependence of the observed second-order
rate constant, kobs, for the reaction of superoxide with benzoyl peroxide.
The reaction rate constant is plotted as ln(kobs/T) according to the Eyring
equation. The equation corresponding to the best-fit line is y )
(-4.3 ( 0.5)x + (8 ( 2). Errors in the slope and intercept represent a
95% confidence level. The R2 value is 0.9537. Error bars represent
one standard deviation from the average of three measurements.
•-
parameters correspond to the initial step of the reaction of O2
with BP. The large negative ∆Sq value suggests a highly ordered
transition state with respect to the ground state. A plausible
mechanism may be found in the known cleavage of esters by
nucleophilic attack of O2 at the carbonyl carbon.28-30 The
•-
•-
reduction of an ester by O2 is proposed to proceed through
nucleophilic addition and elimination.30 This initial step may
SCHEME 3
•-
then be followed by reduction of the peroxy radical by O2
(Scheme 4). Thus, two superoxide ions are consumed in this
mechanism, one as a nucleophile and one as a reductant.
If O2•- attacks one of the carbonyl carbons in BP in a similar
manner, a peroxy intermediate (I) would be formed, which could
decompose as shown in Scheme 5 to yield 1O2, benzoate anion,
and a benzoyloxy radical. Reduction of the benzoyloxy radical
by O2 would result in the proposed reaction products.1
•-
A key feature of the proposed mechanism for 1O2 formation
•-
is attack by O2 at the carbonyl carbon of BP. The fact that
CP, TBP, and TBHP lack carbonyl functionalities and that 1O2
is not observed upon their exposure to O2 supports this
was carried out over a temperature range from 7 to 60 °C (280
•-
to 333 K), and the observed second-order rate constant, kobs
,
mechanistic hypothesis.
was determined at each temperature. This temperature depen-
dence is illustrated in the form of an Eyring plot in Figure 3.
As illustrated, the reaction of O2•- with BP exhibits a normal
temperature dependence with an increasing reaction rate at
elevated temperatures. The following activation parameters were
obtained from the temperature dependence data: ∆Hq, 8.6 kcal
mol-1 (36 kJ mol-1); ∆Sq, -31 cal mol-1 K-1 (-131 J mol-1
K-1); activation energy, Ea, 9.1 kcal mol-1 (38 kJ mol-1). At
25 °C (298 K), ∆Gq is 18 kcal mol-1 (75 kJ mol-1).
A remaining possibility for the reaction of an organic peroxide
•-
with O2 is addition to one of the peroxide oxygen atoms,
yielding Haber-Weiss-type products, RO-, RO•, and O2
(Scheme 6); however, our results suggest that this pathway, if
1
it occurs, is not the source of O2 in this reaction.
1
Because O2 is not observed in the cases of CP or TBP, if
such a reaction occurs, either it generates ground-state O2 or
Mechanistic Implications. The kinetic data obtained for the
(28) Magno, F.; Bontempelli, G. J. Electroanal. Chem. Interfacial
Electrochem. 1976, 68, 337-344.
•-
reaction of O2 with BP are consistent with a two-step
(29) San Filippo, J., Jr.; Romano, L. J.; Chern, C.-I.; Valentine, J. S. J.
Org. Chem. 1976, 41, 586-588.
•-
mechanism, in which 1 equiv of O2 is involved in each step
(Scheme 3). The rate of the final step of the reaction, trapping
(30) Gibian, M. J.; Sawyer, D. T.; Ungermann, T.; Tangpoonpholvivat,
R.; Morrison, M. M. J. Am. Chem. Soc. 1979, 101, 640-644.
1
of O2 by probe 1, is known (45 s-1 for [1] ) 100 µM).3 The
798 J. Org. Chem., Vol. 71, No. 2, 2006

SCHEME 6
1 (430 µM) was prepared in a 50% toluene, 50% THF mixture.
For each trial, the following were combined in an amber 2-mL
autosampler vial containing a micro stirbar: 363 µL of KO2
solution, 117 µL of stock solution of 1 (100 µM final concentration),
and an aliquot of stock BP solution (0.71 mM final concentration).
The total reaction volume was 500 µL. For all trials, the BP was
added last and marked time zero for the reaction. Reaction vials
were thermostated at 25 °C for the duration of the kinetic assay.
Time points were obtained by sampling aliquots of the reaction
mixture and quenching the reaction by 1:1000 dilution in CH3CN.
The CL signal was then measured for each time point as described
above.
In the deuterated toluene experiment, the final composition of
the reaction was 85% toluene-d8. In the DABCO quenching
experiment, the reaction contents were changed as follows: 363
µL of KO2 solution (toluene), 117 µL of stock solution of 1 (100
µM final concentration), 10 µL of stock BP solution (0.71 mM
final concentration), and 10 µL of stock DABCO solution (520
µM final concentration). In the trial without DABCO, the 10-µL
aliquot of stock DABCO solution was replaced with 10 µL of
toluene. The reaction of BP (0.71 mM) with KO2 in toluene was
also carried out at various temperatures, ranging from 7 to 60 °C,
using a thermostated water bath. All other conditions were the same
as the reactions at 25 °C.
In a separate experiment, the reaction of BP with KO2 was carried
out in a manner similar to that described above, and the disappear-
ance of BP was monitored by high-performance liquid chroma-
tography with UV absorbance detection (HPLC-UV). For this
kinetics experiment, the following were combined in a 2-mL
autosampler vial: 726 µL of KO2 solution, 117 µL of THF, 137
µL of toluene, and 20 µL of stock BP solution (0.71 mM final
concentration). The total reaction volume was 1000 µL. The BP
was added last, and analysis by HPLC-UV commenced im-
mediately after this addition. The HPLC analyses were performed
on an 1100 series Hewlett-Packard liquid chromatograph equipped
with a UV-absorbance detector. The chromatographic column used
was a Supelco C18, 150 × 4.6 mm, 5-µm particle size column.
Benzoyl peroxide was analyzed by an isocratic method with a
mobile phase of 75:25 methanol/water, a flow rate of 1.0 mL/min,
an injection volume of 10 µL, and a detection wavelength of 238
nm. Time points were obtained every 7.5 min. The reaction was
carried out at 22 °C. Chromatographic peak areas were converted
to BP concentration using an external calibration curve.
1
the yield of O2 is too low to be trapped and observed by this
detection method (<0.1%). Gibian and Ungermann also noted
•- 31
that TBP is apparently inert toward O2
.
Peters and Foote
confirmed this result.32 Thus, it appears that 1O2 formation may
•-
be limited to the reaction of O2 with acyl peroxides.
A trap-and-trigger detection method was used to quantify the
production of O2 in the reactions of O2 with three organic
peroxides and one hydroperoxide. Only in the case of BP was
a significant yield of 1O2 observed (5%). No 1O2 was observed
1
•-
•-
in the reactions of O2 with alkyl peroxides/hydroperoxides
(cumyl and tert-butyl), indicating the potential importance of
the carbonyl functionality in the reaction. This observation, taken
together with the observed bimolecular reaction kinetics,
indicates that the rate-determining process may be the attack
of O2•- at the BP carbonyl. The formation of 1O2 in the reactions
•-
of O2 with acyl peroxides with yields as high as 5%, as
measured here, represents a potentially significant source of 1O2
via nonphotochemical reactions in both biological and envi-
ronmental systems.
Experimental Section
Materials. Potassium superoxide (KO2), benzoyl peroxide (pu-
rum, moistened (25%) with water), cumyl peroxide (98%), tert-
butyl peroxide (98%), tert-butyl hydroperoxide (anhydrous solution
in decane (5.27 M) over 4 Å molecular sieves), 1,4,7,10,13,16-
hexaoxacyclooctadecane (18-crown-6), 1,4-diazabicyclo-[2.2.2]-
octane (DABCO), diethylenetriaminepentaacetic acid (DTPA),
toluene-d8 (99%), acetonitrile (CH3CN), toluene, and tetrahydro-
furan (THF, without stabilizer) were purchased from commercial
suppliers and used as received. The synthesis and purification of
vinyl ether probe 1 have been previously described.3
Software. OriginPro (v. 7.5) global fitting software was used to
simultaneously fit several sets of data obtained in the reaction of
•-
Reactions of Superoxide with Cumyl- and tert-Butyl Perox-
ides and tert-Butyl Hydroperoxide. All glassware was treated in
the same manner as that for the BP reactions, and the KO2 solutions
were prepared identically. Stock solutions of CP, TBP, and TBHP
were prepared in toluene. Reactions of CP (1, 5, and 10 mM), TBP
(1, 5, and 10 mM), and TBHP (10 mM) with KO2 were each carried
out in a manner similar to that for the BP reactions. Time points
were obtained and analyzed as in the BP reactions. All reactions
were carried out at 25 °C.
O2 with benzoyl peroxide to estimate the concentration of
•-
dissolved O2 in saturated solutions of KO2 in toluene.
Chemiluminescence Measurements. Chemiluminescence (CL)
measurements were made as described previously.3 For all assays,
CL samples were prepared by 1:1000 dilution in CH3CN. The
luminometer sensitivity was set to 60.1%, the default setting. The
CL signal from 2 was triggered by adding 20 µL of 10 mM TBAF,
via syringe through a light-tight manual injection port, to 300 µL
of the reaction solution in CH3CN. The CL signals obtained were
calibrated by comparison to external standards prepared from an
independently synthesized sample of 2.3
Reactions of Superoxide with Benzoyl Peroxide. All glassware
was acid-washed, rinsed thoroughly with Milli-Q water, and dried
in an oven (110 °C) prior to use. Saturated solutions of KO2
containing 18-crown-6 (47 mM) and diethylenetriaminepentaacetic
acid (DTPA, 4 mM) were prepared in toluene and 99% toluene-d8.
The solutions were mixed on a vortex mixer for 1 min, followed
by centrifugation at 3300 rpm for 2 min to allow settling of any
undissolved KO2. Solutions of dissolved KO2 were used im-
mediately upon preparation. Stock solutions of BP (0.035 M) and
DABCO (0.026 M) were prepared in toluene. A stock solution of
•-
Determination of the Dissolved O2 Concentration. The
•-
concentration of dissolved O2 in the saturated solutions of KO2
(solubilized with 18-crown-6) in toluene prepared for this study
•-
was estimated from kinetic data obtained for the reaction of O2
with BP using OriginPro global fitting software. Several sets of
data obtained under identical conditions were simultaneously fit to
eq 1, with the value for the initial O2•- concentration, [O2•-]0, shared
among the data sets. The value for [BP]0 was input, and all other
variables were solved by OriginPro. Using this method, we obtained
a value of 2.0 (( 0.2) mM for [O2•-]0.
Acknowledgment. Financial support was provided by a
Grant-in-Aid from the University of Minnesota.
(31) Gibian, M. J.; Ungermann, T. J. Org. Chem. 1976, 41, 2500-2502.
(32) Peters, J. W.; Foote, C. S. J. Am. Chem. Soc. 1976, 98, 873-875.
JO051736N
J. Org. Chem, Vol. 71, No. 2, 2006 799