Diethyl ketone triperoxide: Thermal decomposition reaction in chlorobenzene solution and its application as initiator of polymerization

Article abstract of DOI:10.1002/poc.716

The kinetics of the thermal decomposition reaction of diethyl ketone triperoxide (3,3,6,6,9,9-hexaethyl-1,2,4,5,7,8-hexaoxacyclononane, DEKT) in chlorobenzene solution were studied in the temperature range 99.0-148.0°C and at initial concentrations of (1.65-4.97) × 10^-2 M. The thermolysis of DEKT follows a first-order kinetic law up to at least ca 80% triperoxide conversion. The activation parameter values for the initial O - O bond rupture in chlorobenzene (ΔH? = 134.6 ± 1.7 kJ mol^-1; ΔS? = 4.2 ± 3.8 J mol^-1 K^-1) and the reaction products observed support a stepwise reaction mechanism which includes as a first step the unimolecular homolytic cleavage of one peroxidic bond of the DEKT molecule giving rise to a biradical as intermediate. Additionally, the results obtained were compared with those obtained in toluene, toluene-styrene (50%, v/v) and chlorobenzene-styrene (50%, v/v) solution, showing that the decomposition reaction is strongly solvent dependant. Further, the biradical intermediate obtained in all cases was used to initiate styrene polymerization. It was demonstrated that DEKT can effectively act as initiator in styrene polymerization and its performance is similar to that presented by a multifunctional initiator giving rise to high molecular weight polystyrene at a high reaction rate. Copyright

Full text of DOI:10.1002/poc.716

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 2004; 17: 215–220
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/poc.716
Diethyl ketone triperoxide: thermal decomposition
reaction in chlorobenzene solution and its application
as initiator of polymerization
Adriana I. Can˜ izo,1* Gladys N. Eyler,1 Graciela Morales2 and Jorge R. Cerna2
1
´
´
Laboratorio de Quımica, Facultad de Ingenierıa, Universidad Nacional del Centro de la Provincia de Buenos Aires, Avda. del Valle 5737,
´
B7400JWI, Olavarrıa, Argentina
2
´
´
´
Centro de Investigacion en Quımica Aplicada, Blvd. Enrique Reyna 140, Apdo. Postal 379, 25100 Saltillo, Coahuila, Mexico
Received 31 March 2003; revised 28 August 2003; accepted 16 September 2003
ABSTRACT: The kinetics of the thermal decomposition reaction of diethyl ketone triperoxide (3,3,6,6,9,9-hexaethyl-
1,2,4,5,7,8-hexaoxacyclononane, DEKT) in chlorobenzene solution were studied in the temperature range 99.0–
148.0 ꢀC and at initial concentrations of (1.65–4.97) ꢁ 10ꢂ2 M. The thermolysis of DEKT follows a first-order kinetic
law up to at least ca 80% triperoxide conversion. The activation parameter values for the initial O—O bond rupture in
chlorobenzene (ÁHz ¼ 134.6 ꢃ 1.7 kJ molꢂ1; ÁSz ¼ 4.2 ꢃ 3.8 J molꢂ1 Kꢂ1) and the reaction products observed
support a stepwise reaction mechanism which includes as a first step the unimolecular homolytic cleavage of one
peroxidic bond of the DEKT molecule giving rise to a biradical as intermediate. Additionally, the results obtained
were compared with those obtained in toluene, toluene–styrene (50%, v/v) and chlorobenzene–styrene (50%, v/v)
solution, showing that the decomposition reaction is strongly solvent dependant. Further, the biradical intermediate
obtained in all cases was used to initiate styrene polymerization. It was demonstrated that DEKT can effectively act as
initiator in styrene polymerization and its performance is similar to that presented by a multifunctional initiator giving
rise to high molecular weight polystyrene at a high reaction rate. Copyright # 2004 John Wiley & Sons, Ltd.
KEYWORDS: cyclic organic peroxide; diethyl ketone triperoxide; hexaoxacyclononane; solution kinetics; multifunctional
initiator; cyclic peroxide initiator; styrene polymerization; polystyrene
INTRODUCTION
in particular di- and triperoxides, which can be consid-
ered as multifunctional initiators as a consequence of the
two or three active sites of the peroxide molecule.
Recently, we have published5 some results dealing with
styrene polymerization in the presence of diethyl ketone
triperoxide (3,3,6,6,9,9-hexaethyl-1,2,4,5,7,8-hexaoxa-
cyclononane) (DEKT) in the bulk and in toluene solution.
In this paper, we report a study of the thermal decom-
position reaction of DEKT and its application as a
trifunctional initiator of styrene polymerization. The
kinetic experiments were carried out in chloroben-
zene, chlorobenzene–styrene (50%, v/v), toluene–styrene
(50%, v/v) and styrene. The selection of chlorobenzene in
particular was made taking into account its low transfer
Most commercial vinyl polymers such as polystyrene,
polyethylene and poly(vinyl chloride) are produced in-
dustrially by means of free radical polymerization, where
organic peroxides are usually used as free radical initia-
tors. Cyclic organic di- and triperoxides can decompose
thermally by homolytic cleavage of the O—O bond to
give an intermediate biradical,1–4 which in turn initiates
the polymerization of vinyl monomers. The radical thus
produced is incorporated into the polymer structure
yielding polymeric species with embedded peroxide
groups that decompose later during the polymerization
process. Since the nature of the biradical formed is
dependent on peroxide structures and has a profound
influence on the final polymer properties, it is important
to clarify the initiation mechanism of organic peroxides,
constant in a polymerization process (Cs ¼ 0.56 ꢁ 10ꢂ4
)
and the value of half-life of the DEKT in that solvent
(t1/2 ¼ 5.1 h at 130 ꢀC).
´
*Correspondence to: A. I. Can˜izo, Laboratorio de Quımica, Facultad
de Ingenierıa, Universidad Nacional del Centro de la Provincia de
´
Buenos Aires, Avda. del Valle 5737, B7400JWI, Olavarrıa, Argentina.
E-mail: acanizo@fio.unicen.edu.ar
Contract/grant sponsors: Universidad Nacional del Centro de la
´
RESULTS AND DISCUSSION
Thermal decomposition of DEKT in solution
´
Provincia de Buenos Aires (UNC); Comision de Investigaciones
´
Cientıficas de la Provincia de Buenos Aires.
We have shown previously1 that there is a significant
solvent effect on the kinetics and activation parameters
´
Contract/grant sponsor: Consejo Nacional de Ciencia y Tecnologıa
(CONACYT); Contract/grant number: 31249-U.
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 215–220

˜
A. I. CANIZO ET AL.
216
corresponding to the thermal decomposition reaction of
DEKT in solution in benzene, toluene and benzene–2-
propanol (1 : 1, v/v). The unimolecular thermal decom-
position reaction of this molecule can be represented by a
stepwise homolysis where the rate of ring cyclization
within the solvent cage after the initial O—O bond
cleavage would be a relatively fast step6,7 [Eqn (1)].
Thus, the rate-determining step is biradical (dialcoxy
radical) formation that further decomposes by either
C—C, C—O or O—O bond rupture, originating differ-
ent final products.
conditions employed in this study, there are no contribu-
tions from second-order processes inducing the decom-
position of the triperoxide, either at higher conversions
(ca three half-lives) or at higher initial DEKT concentra-
tion. Therefore, the observed rate constant values
(Table 1) can be ascribed to the unimolecular initial
step of DEKT thermolysis [Eqn (1)].
The reaction products in chlorobenzene detected by
gas chromatography were diethyl ketone, ethane, butane,
4,40-dichloro-1,10-biphenyl,
1-chloro-3-ethylbenzene,
1-chloro-4-ethylbenzene, 1-chloro-2-ethylbenzene and
methyl propionate. These reaction products can be inter-
preted in terms of the initial O—O bond homolytic
rupture to give the corresponding biradical [Eqn (1)].
This biradical can reform the DEKT molecule or undergo
either a C—O, C—C or O—O bond cleavage. Quanti-
tative analysis of 3-pentanone, generated by C—O bond
ruptures, shows a molar yield of 0.27 mol per mole of
DEKT decomposed. Ethyl radicals derived from C—C
bond ruptures can abstract a hydrogen atom from the
solvent to give ethane and other products by combination
of free radicals (1-chloro-3-ethylbenzene, 1-chloro-4-
ethylbenzene, 1-chloro-2-ethylbenzene, butane, 4,40-
dichloro-1,10-biphenyl). Product analysis confirms that
the C—C bond rupture to give ethyl radicals makes an
important contribution to the decomposition of the bir-
adical. The presence of methyl propionate can be as-
sumed to be a result of C—C and O—O bond ruptures
coming from the former biradical [Eqn (1)] to generate a
propionate radical. Oxygen and carbon dioxide probably
are also formed as reaction products but the instrumental
methods used made it impossible to detect them (see
Experimental). At present, quantitative analysis of the
reaction products is insufficient for the complete elucida-
tion of the thermolysis mechanism of DEKT in chloro-
benzene solution, and further work is in progress.
In a previous study (A. I. Can˜izo, unpublished work,
2001), it was also demonstrated that the biradical decom-
poses thermally to form free ethyl radicals that react with
the solvent, benzene or toluene, to generate different
products such as ethylbenzene, biphenyl or n-propylben-
zene and bibenzyl, respectively. In this way, this cyclic
peroxide can be considered as an important source of
ethyl radicals.
The thermal decomposition reaction of DEKT in
chlorobenzene solution, at temperature and initial con-
centration ranges of 99.0–148.0 ꢀC and (1.65–
1.76) ꢁ 10ꢂ2 M, respectively, follows a first-order kinetic
law up to at least 80% triperoxide conversion. In order to
discard a radical-induced decomposition reaction as a
competing mechanism, the kinetics of the DEKT thermal
decomposition reaction in chlorobenzene solution were
studied at 135.0 ꢀC with a higher initial peroxide con-
centration (4.97 ꢁ 10ꢂ2 M) (Table 1). The results of the
experiments show that the rate constant values are prac-
tically independent of the initial DEKT concentration.
Hence it can be assumed that, under the experimental
Nevertheless, the temperature effect on the rate con-
stant values (kexp) for the unimolecular reaction [Eqn (1)]
can be represented by the Arrhenius equation. The
Table 1. First-order rate constants for DEKT decomposition in chlorobenzene solution
Temperature ( ꢀC)
[DEKT]a (10ꢂ2M)
kexp (10ꢂ5 sꢂ1
)
Experimental pointse
5
DEKT conversion (%)
26
rf
99.0
118.0
1.65
1.65
0.167
ꢂ0.990
0.925d
7.44d
7.84
9
9
5
9
80
ꢂ0.994
ꢂ0.998
ꢂ0.988
ꢂ0.990
1.76
135.0
148.0
1.65b
1.76c
4.97b
1.65
52
80
50
28.0c
78
1.76
a
Initial DEKT concentration.
Naphthalene 0.7% as internal standard.
Decalin 0.12% as internal standard.
b
c
d
e
f
Average kexp values with 1.65 ꢁ 10ꢂ2 and 1.76 ꢁ 10ꢂ2 M as initial DEKT concentration.
Number of experimental point obtained for the kinetic curve.
r ¼ correlation coefficient from Arrhenius equation.
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 215–220

DIETHYL KETONE TRIPEROXIDE: DECOMPOSITION AND APPLICATION INITIATOR
217
power of the solvent used. In this case we postulate the
existence of a rather dipolar activation complex in com-
parison with the initial state. As a consequence of
differential solvation, in more polar solvents the reaction
is accelerated (kexp in chlorobenzene is three times higher
than that in toluene) because of higher solvation of the
activation complex. A more highly ordered activated
complex results because of the presence of some kind
of association with the solvent molecule resulting in a
more rigid species. Therefore, the observed activation
entropy values are lower in polar solvents, reflecting the
decrease in the degrees of freedom of the molecules.
On the other hand, when the thermal decomposition
reaction of DEKT was carried out in styrene–
chlorobenzene or styrene–toluene mixtures (Table 2), it
could be observed that the reactivity of DEKT in the
presence of styrene is higher when the organic solvent is
toluene, in contract to the behavior observed in neat
solvents. Again, it is evident that solvent effects are
closely related to the nature and extent of solute–solvent
interactions developed locally in the immediate vicinity
of the solutes. When the same decomposition reaction is
evaluated in pure styrene, the rate constant (kexp) reaches
the highest value, probably owing to the fast reaction that
takes place between the biradical initially formed [Eqn
(1)] and styrene molecules, which are excellent radical
scavengers. The multifunctional character of the ini-
tiator makes the polymerization reaction rate increase
as a consequence of successive initiation–propagation–
termination reactions during the course of the polymer-
ization process. In this scenario, polymerization of
styrene in the presence of DEKT is analyzed on the basis
of the initial step previously proposed that includes the
presence of a biradical initiating species (Scheme 1).
Figure 1. Eyring plot corresponding to the thermal decom-
position reaction of DEKT in chlorobenzene and toluene
solution
corresponding activation parameters can be deduced
from the Eyring plot (Fig. 1) and the kinetic data for
the thermal decomposition reaction of DEKT in toluene
solution obtained in previous studies (Ref. 1 and A. I.
Can˜izo, unpublished work, 2001) were included. The
representation in Fig. 1 is linear (r ¼ 0.996) over a
relatively wide temperature range (ca 49 ꢀC), which
suggests that the activation parameters (Table 2) belong
to a single process, which corresponds to the unimole-
cular homolysis of the DEKT molecule [Eqn (1)]. The
errors shown in Table 2 are standard deviations from a
least mean squares treatment8,9 of the kinetic data.
The marked solvent effect observed1 on the unimole-
cular decomposition reaction of the DEKT molecule is
confirmed once again with the results obtained in other
polar solvents such as chlorobenzene and solvent–styrene
mixtures. In neat solvent both activation parameters
(enthalpy and entropy of activation) decrease when the
solvent polarity increases (Table 2); this effect is the
‘compensation effect’. The initial state of the DEKT
molecule, and also the corresponding activated complex,
is solvated to different extents depending on the solvating
Polymerization of styrene in the presence
of DEKT
The effect of solvent on the monomer conversion and
polymer molecular weight was studied using DEKT as
initiator and the behavior of this cyclic peroxide in
styrene solution polymerization was also evaluated. The
temperature employed was 130 ꢀC because at this tem-
perature and in toluene solution, it was previously
Table 2. Kinetic and activation parameters for DEKT thermal decomposition in different solvents
130 ꢀC
kexp
Solvent
t1/2 (h)
(10ꢂ5 sꢂ1
)
Rp (10ꢂ3 mol lꢂ1 sꢂ1
)
ÁHz (kJ molꢂ1 a,b
)
ÁSz (J molꢂ1 Kꢂ1 a,b
)
Styrene
Styrene–toluene (50%, v/v)
0.13
0.34
14.8
0.41
153
56.5
1.3
7.5
7.2
n.a.
3.8
n.a.
n.a.
n.a.
n.a.
Toluene1
203.6 ꢃ 6.3 (48.7)
169.3 ꢃ 14.2 (40.5)
Styrene–chlorobenzene
(50%, v/v)
Chlorobenzene
46.8
n.a.
n.a.
5.1
3.8
n.a.
134.6 ꢃ 1.7 (32.2)
4.2 ꢃ 3.8 (1.0)
a
Standard deviation from a least mean squares data treatment.
In parentheses, activation parameters expressed in kcal molꢂ1 or cal molꢂ1 Kꢂ1, respectively (1 kcal ¼ 4.184 kJ).
b
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 215–220

˜
A. I. CANIZO ET AL.
218
Scheme 1. Different routes proposed for the decomposition of DEKT and characteristics of the polymers obtained
demonstrated that this initiator can act as a multifunc-
Moreover, the polymerization rates in bulk or solution are
higher than that obtained for the spontaneous process.
Nevertheless, when the conversion values and, in con-
sequence, the polymerization rates in chlorobenzene and
toluene are compared, it can be observed that conversion
is lower in chlorobenzene than toluene (Fig. 2). This
behavior is in accord with the observed half-lives of
DEKT in each solvent (t1/2 ¼ 14.8 h in toluene and 5.1 h
in chlorobenzene) and is reflected in an increase in the
polymerization rate (Rp) in toluene–styrene mixtures
(Table 2).
When the evolution of molecular weight with poly-
merization time is considered (Fig. 3), there is clear
evidence that supports the solvent participation in the
initiating step. The molecular weight of the polymers
obtained in chlorobenzene are of the order of
1.7 ꢁ 105 g molꢂ1 with a polydispersity index (Mw/Mn)
of <2, values which are slightly higher than those
obtained in the presence of toluene.
tional initiator.5 One of the potential advantages of using
multifunctional initiators, such as that proposed in this
study, is that a high polymer molecular weight and a high
polymerization rate can be obtained simultaneously by
controlling the initiator decomposition rate through opti-
mal reactor temperature programming, leading to better
control of the radical concentration through sequential
decomposition of labile groups (Scheme 1, route A) of
different thermal stabilities. For free radical polymeriza-
tion with monofunctional initiators, the reaction rate
depends on the initiator concentration and the polymer-
ization is of order 0.5 with respect to initiator concentra-
tion. The presence of additional labile groups in
multifunctional initiators makes the polymerization ki-
netics very complicated, because labile groups such as
peroxide become redistributed repeatedly in the growing
and inactive polymers and they are engaged in further
initiation, propagation, chain transfer and termination
reactions during the course of polymerization.
With respect to the bulk polymerization, the conversion
values reached are greater than those in toluene or
chlorobenzene, as expected in a mass polymerization
process. Otherwise, the molecular weight values present
a singular behavior, as can be observed from Fig. 3 where
molecular weight increases with polymerization time in a
From Fig. 2, it can be seen that whether the polymer-
ization takes place in bulk or in solvent–styrene mixtures,
the initiator efficiently initiates the polymerization of
styrene if curves b, c and d are compared with curve a,
which corresponds to the autoinitiated polymerization.
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 215–220

DIETHYL KETONE TRIPEROXIDE: DECOMPOSITION AND APPLICATION INITIATOR
219
place, allowing the conversion and molecular weights
to increase simultaneously, contrary to a ‘classical’
radical process where an increase in conversion, due to
an increase in radical concentration, leads to a decrease in
molecular weight of the final products. Moreover, the
high molecular weights observed in the first step of the
polymerization are associated with species that result
from the combination of growing primarily radicals
giving rise to species with two O—O bonds within their
structures that subsequently can decompose and initiate
new chains. The intermediate species formed from the
beginning of the reaction has recently been isolated by G.
Morales and co-workers (to be published). This fact
supports the postulated mechanism where the unimole-
cular homolysis of peroxide is the rate-determining step
and the existence of that biradical species is responsible
for the initiation step in polymerization reactions as
described above.
CONCLUSIONS
The thermolysis of DEKT in chlorobenzene solution
follows a first-order kinetic law up to at least ca 80%
triperoxide conversion.
Under the experimental conditions described here, a
radical-induced decomposition reaction as a competing
mechanism with the decomposition of DEKT may be
dismissed, so the activation parameters correspond to the
unimolecular thermal decomposition reaction of the
DEKT molecule.
Figure 2. Evolution of conversion as a function of polymer-
ization time for (a) thermal, (b) bulk, (c) toluene and (d)
chlorobenzene polymerization of styrene using DEKT as
initiator
Analysis of the reaction products is insufficient to
elucidate the real mechanism of the thermolysis of the
diethylketone triperoxide, but studies are in progress to
clarify it.
DEKT can effectively act as an initiator in styrene
polymerization and its performance is similar to that
presented by a multifunctional initiator leading to poly-
mers with high molecular weight at a high reaction rate
compared monofunctional initiated polymerization.
Values of the conversion and molecular weight of
the polymers obtained strongly depend on the solvent
used and the polymerization conditions. It was ob-
served that the DEKT initiator efficiency decreases in
chlorobenzene.
EXPERIMENTAL
Materials. DEKT was prepared by methods described
elsewhere10,11 and its purity was checked by GC, FTIR
Figure 3. Evolution of Mw as a function of time for styrene
polymerization in different solvents: ^, bulk; &, toluene;
~, chlorobenzene
1
(KBr), 13C NMR, H NMR and TLC. Caution: DEKT
must be handled with care because it can be detonated by
shock. Chlorobenzene (Mallinckrodt, AR grade) was
purified by standard techniques12 (b.p. 132–133 ꢀC).
Sublimed naphthalene (Mallinckrodt, AR grade) or dec-
alin was employed as an internal standard in quantitative
bulk process. This last situation is typical of a polymer-
ization reaction initiated by a multifunctional initiator
where in the course of the polymerization several cycles
of initiation–propagation–termination reactions take
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 215–220

˜
A. I. CANIZO ET AL.
220
determinations of DEKT concentration and the 3-
pentanone concentration was determined by GC. Styrene
(Aldrich) was distilled over sodium under vacuum and
kept in refrigeration until use.13
merization time, the polymer samples were dissolved in
THF and precipitated by adding excess methanol, dried
in vacuo and the monomer conversion was measured
gravimetrically. The molecular weight and molecular
weight distribution of polystyrene were determined by
GPC using THF as a solvent.
Kinetic methods. Pyrex glass ampoules (10 cm
long ꢁ 6 mm o.d.) half filled with DEKT solution and
spiked with naphthalene or decalin as an internal standard
were thoroughly degassed under vacuum atꢂ196 ꢀC and
then sealed with a flame torch. To perform the runs, the
ampoules (5–8 at each temperature) were immersed in a
thermostated silicone oil-bath ( ꢃ 0.1 ꢀC) and withdrawn
after predetermined times. The reaction was stopped by
cooling to 0 ꢀC. The concentrations of DEKT remaining
in the solution and the reaction products were determined
by quantitative GC analysis in a 3% SE-30 column
(silicone gum rubber stationary phase on Chromosorb-
W, 1/8 in o.d. stainless steel) installed in a Konik-2000 C
gas chromatograph with nitrogen as carrier gas and flame
ionization detection. The reaction products were identi-
fied by GC–MS analysis.
Acknowledgements
This study was financially supported by the Universidad
Nacional del Centro de la Provincia de Buenos Aires
´
´
(UNC), the Comision de Investigaciones Cientıficas de la
Provincia de Buenos Aires (Argentina) and the Consejo
´
Nacional de Ciencia y Tecnologıa (CONACYT),
Contract No. 31249-U (Mexico). Jorge R. Cerna thanks
´
CONACYT for a Research Fellowship.
REFERENCES
The corresponding first-order rate constants were cal-
culated from the slope of the line obtained by plotting the
values of ln[DEKT] aganist reaction time. The corre-
sponding activation parameters were deduced from the
Eyring equation and the errors were determined using a
least mean squares data treatment.8,9
1. Eyler GN, Can˜izo AI, Alvarez EE, Cafferata LFR. An. Asoc.
´
Quım. Argent. 1994; 82: 175–181.
2. Cafferata LFR, Eyler GN, Can˜izo AI, Alvarez EE. J. Org. Chem.
1991; 56: 411–414.
3. Eyler GN, Mateo CM, Alvarez EE, Can˜izo AI. J. Org. Chem.
2000; 65: 2319–2321.
4. Sanderson JR, Story PR. J. Org. Chem. 1974; 39: 3463–3469.
5. Cerna JR, Morales GE, Eyler GN, Can˜izo AI. J. Appl. Polym. Sci.
2002; 83: 1–11.
6. Benson SW. Thermochemical Kinetics. Wiley: New York, 1968;
69.
7. Cafferata LFR. Trends Org. Chem. 1993; 4: 773–791.
8. Huyberechts S, Halleux A, Kruys P. Bull. Soc. Chim. Belg. 1955;
64: 203–209.
Polymerization experiments. The syntheses of different
polystyrenes were carried out by dissolving the appro-
priate amount of DEKT in styrene–chlorobenzene (50%,
v/v) and equal amounts of the resulting solution were
placed in glass tubes. The tubes were evacuated, sealed
and kept at 130 ꢀC for 4, 6, 8, 10, 15, 60, 120, 180
and 360 min. The conversion in the pure thermal poly-
merization of styrene was measured under the same
experimental conditions. Experiments in bulk were also
evaluated for comparison with the results obtained in
solution polymerization. After the predetermined poly-
9. Schaleger LL, Long FA. Adv. Phys. Org. Chem. 1963; 1: 1–33.
´
10. Milas NA, Golubovic A. J. Am. Chem. Soc. 1959; 81: 3361–3364.
11. Eyler GN, Can˜izo AI, Alvarez EE, Cafferata LFR. Tetrahedron
Lett. 1993; 34: 1745–1746.
12. Riddick JA, Bunger WB. In Organic Solvents, Weissberger A
(ed). Wiley: New York, 1970; 767.
´
13. Perrin DD, Armarego WLF. Purification of Laboratory Chemicals
(3rd edn). Pergamon Press: Oxford, 1988; 276.
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 215–220