Radical Nature of Pathways to Alkene and Ester from Thermal Decomposition of Primary Alkyl Diacyl Peroxide

Article abstract of DOI:10.1021/jo9512766

Thermal decomposition of a primary alkyl diacyl peroxide 2 is investigated.Dependence of product yields on temperature, viscosity, and solvent polarity is examined in a variety of media.The excess of the alkene disproportionation product 4 and the presence of ester 3 and acid 5 is argued to demonstrate the existence of a discrete acyloxy-alkyl geminate radical pair.Stereoselective deuterium labeling of 2 and subsequent 1H-NMR analysis of the resulting isotopomers of 4 confirm the radical nature of detected decomposition products.

Full text of DOI:10.1021/jo9512766

J . Org. Chem. 1996, 61, 2801-2808
2801
Ra d ica l Na tu r e of P a th w a ys to Alk en e a n d Ester fr om Th er m a l
Decom p osition of P r im a r y Alk yl Dia cyl P er oxid e
Lev R. Ryzhkov
Department of Chemistry, Towson State University, Baltimore, Maryland 21204
Received J uly 17, 1995 (Revised Manuscript Received J anuary 30, 1996X)
Thermal decomposition of a primary alkyl diacyl peroxide 2 is investigated. Dependence of product
yields on temperature, viscosity, and solvent polarity is examined in a variety of media. The excess
of the alkene disproportionation product 4 and the presence of ester 3 and acid 5 is argued to
demonstrate the existence of a discrete acyloxy-alkyl geminate radical pair. Stereoselective
1
deuterium labeling of 2 and subsequent H-NMR analysis of the resulting isotopomers of 4 confirm
the radical nature of detected decomposition products.
In tr od u ction
and by comparison with nuclear hyperfine frequencies
1
0
via measurement of CIDNP enhancement factors. These
studies indicate decarboxylation rate constants at 80 °C
Acyloxy radicals, 1, are reactive intermediates in
1
organic reactions ranging from the Barton and related
7
-1
of the order of 10 s , and an activation energy of 8-9
2
reactions, Hundsdiecker reactions to Kolbe electrolysis
kcal/mol.
and thermal or photodecomposition of diacyl peroxides
Much less certain is the decarboxylation rate of RCO
when R is primary, secondary, or tertiary alkyl or benzyl.
The acetoxy radical (R ) CH ) decarboxylation rate has
been estimated from product yields and CIDNP en-
•,
2
3
and other acyloxy derivatives. They also serve as
intermediates in generation and calibration of many of
the “radical clocks”.4
3
1
1
hancement factors12 to be 109
s
-1
at 80 °C with an
•
•
RCO2
R + CO2
activation energy of approximately 7 kcal/mol. Estimates
•
1
for the decarboxylation rate of RCO
CH , (CH CH, (CH C, PhCH , and PhCH
recently obtained by comparison with electron-transfer
2
with R ) CH
3
, CH
3
-
2
3
)
2
3
)
3
2
2
CH
2
were
The most obvious chemical property of these radicals
is the relative ease with which they undergo decarboxy-
lation to produce the corresponding R radical. Knowl-
edge of the rate of this reaction is not only of intrinsic
interest because of the simplicity of the process, but is
also essential if one is to understand fully the distribution
of radical-derived products formed from acyloxy radical
precursors. Determination of the electronic structures
of these species has also proven to present interesting
1
3
rates within the naphthylmethyl-acyloxy radical pairs,
.
assumed independent of the nature of R, and calibrated
by the known decarboxylation rate of 9-methylfluorenoyl-
oxy radical.14 However, the former investigation suggests
that the decarboxylation rate for the propanoyloxy radical
9
-1
is 2 × 10 s . This implies that it should be possible to
observe net CIDNP polarization in photolysis or ther-
molysis of dipropanoyl or other primary peroxides which
5
12
theoretical and experimental challenges.
is contrary to the results reported previously.
Decarboxylation rates of aroyloxy radicals (R ) sub-
stituted phenyl) have been measured, or estimated, by a
variety of methods including indirect comparison with
Without exception primary, secondary, or tertiary alkyl
or benzyl acyloxy radicals decarboxylate too rapidly to
maintain concentrations sufficient for direct detection
using currently available techniques. For example, only
alkyl radical pairs and radicals formed by H abstraction
trapping rates6 direct detection optically, or by EPR
,7
8
9
X
15
Abstract published in Advance ACS Abstracts, March 15, 1996.
1) Barton, D. H. R. et al. J . Am. Chem. Soc. 1991, 113, 6937 and
from the alkyl chains or reactant molecules are seen
(
by EPR during photolysis of single crystals of bis-
decanoyl peroxide at 10-20 K. Furthermore, in many
cases products containing the acyloxy group have not
been detected during the thermal decomposition of diacyl
peroxides, as would be expected for recombination of
intermediate acyloxy-alkyl radical pair (Scheme 1). This
has given rise to an assumption that acyloxy radicals
references therein. See also Newcomb, M.; Dhanabalasingham, B.
Tetrahedron Lett. 1994, 35, 5193; Ziegler, F. E.; Harran, P. G.
Tetrahedron Lett. 1993, 34, 4505. Ziegler, F. E.; Harran, P. G. J . Org.
Chem. 1993, 58, 2768.
(
2) For reviews see Wilson, J . Org. React. 1957, 9, 332. See also
Skell, P. S.; May, D. D. J . Am. Chem. Soc. 1981, 103, 967.
3) (a) Hiatt, R. In Organic Peroxides; Swern, D., Ed.; Wiley-
(
Interscience: New York; 1971; Vol. II, p 799. (b) Koenig, T. In Free
Radicals; Kochi, J . K., Ed.; Wiley-Interscience: New York; 1973; Vol.
I, Chapter 3.
•
•
which produce R more stable then CH
3
do not exist as
(
4) (a) Newcomb, M. Tetrahedron 1993, 49(6), 1151-1176. (b)
discrete chemical intermediates but undergo C-C bond
Newcomb, M.; Manek, M. B. J . Am. Chem. Soc. 1990, 114, 9662. (c)
Choi, S. Y.; Eaton, P. E.; Newcomb, M.; Yip, Y. C. Ibid. 1990, 114,
6
1
326. (d) Newcomb, M.; Varick, T.; Ha, C.; Manek, M. B.; Xu Yue. Ibid.
990, 114, 8158.
(10) (a) Schwerzel, R.; Lawler, R.; Evans, G. Chem. Phys. Lett. 1974,
29, 106. (b) Den Hollander, J . Chem. Phys. 1975, 10, 167.
(11) Braun, W.; Rajerbach, L.; Eirich, F. J . Phys. Chem. 1962, 66,
1591.
(12) (a) Kaptein, R.; Brokken-Zijp, J .; de Kanter, F. J . Am. Chem.
Soc. 1972, 94, 6280. (b) Glazov, Y. et al. Russ. J . Appl. Spectr. 1978,
28, 607. (c) Turetskaya, E. et al. Dokl. Acad. Nauk BSSR 1980, 24,
57.
(5) (a) Feller, D.; Huyser, E. S.; Borden, W. T.; Davidson, E. R. J .
Am. Chem. Soc. 1983, 105, 1459. (b) Feller, D.; Borden, W. T.;
Davidson, E. R. J . Am. Chem. Soc. 1983, 105, 3347. (c) Rauk, A.; Yu,
D.; Armstrong, D. A. J . Am. Chem. Soc. 1994, 116, 8222.
(
(
6) Bevington, J . C.; Lewis, T. D. Trans. Faraday Soc. 1958, 94, 8236.
7) J anzen, E. G.; Evans, C. A.; Nishi, T. J . Am. Chem. Soc. 1972,
1
04, 4515.
8) (a) Chateauneuf, J .; Lusztyk, J .; Ingold, K. U. Ibid. 1988, 110,
877. (b) Ingold, K.; Lusztyc, J .; Chateanneuf, J . Ibid. 1988, 110, 2886.
9) Yamauchi, S.; Hirota, N.; Takahara, S.; Sakuragi, H.; Tokamaru,
K. Ibid. 1985, 107, 5021.
(13) Hilborn, J . W.; Pincock, J . A. J . Am. Chem. Soc. 1991, 113, 2683.
(14) Falvey, D. E.; Schuster, G. B. J . Am. Chem. Soc. 1986, 108,
7419.
(15) Karch, N. J .; Koh, E. T.; Whitsel, B. L.; McBride, J . M. J . Am.
Chem. Soc. 1975, 97, 6729.
(
2
(
0
022-3263/96/1961-2801$12.00/0 © 1996 American Chemical Society

2
802 J . Org. Chem., Vol. 61, No. 8, 1996
Ryzhkov
Sch em e 1
O
O
R′
RCH2CH2
O
O
carboxy inversion product, 12
kpolar
O
RCH2CH2
O
O
R′
O
2: R = C9H19, R′ = RCH2CH2
kdecomp
λIC
O
•
•
RCH2CH2
O
O
R′
O
1
cage I
coupling
2k
disproportionation
IIC
O
IID
O
O
λIIC
λIID
RCH2CH2
O
R′
RCH CH + HO
R′
RCH CH • CO •O
2
2
2
R′
2
3
4
5
cage II
k
IIIC
IIID
λIIIC
λIIID
RCH2CH2 R′
RCH2CH2• CO2 O2C •R′
RCH CH2 + R′(+H)
7
cage III
4
6
scission in synchrony with reaction of the precursor
molecule.16 Early evidence for such a conclusion came
from observation of declining activation energies for
deuterium labeling to further explore the origin of the
products listed above.
1
5a
15b
thermolysis of diacyl peroxides
and peresters
with
Resu lts
.
increasing R stability. Perhaps the strongest evidence
against a discrete acyloxy intermediate is failure to detect
CIDNP in products of acyloxy-alkyl cage II coupling and
disproportionation.17 In those cases where products
containing the acyloxy groups, such as esters or carboxy-
lic acids, have been isolated, it has been assumed that
they arise from ionic side reactions related to the forma-
Decom p osition Ra tes. First-order rate constants for
thermolysis of 2 in four solvents representing the range
of polarity and radical reactivity used in this study were
1
determined from semilog plots of the H-NMR signal of
2 vs time.20 In all cases, decomposition of 2 at 89 °C
showed clean first-order behavior over at least 4 half-
lives. First-order rate constants, kdecomp, were (3.1 ( 0.5)
1
8
tion of the carboxy inversion product or else has been
-
4
-1
-4
reported without comment.19
× 10
s
in C D CD (toluene-d ) and (1.9 ( 0.2) × 10
6
5
3
8
-
1
s
in PFMCH (perfluoromethylcyclohexane). In both
We report here the results of a detailed study of the
products formed when lauroyl peroxide (2) is decomposed
thermally in a variety of solvents at temperatures rang-
ing from 80-140 °C. This investigation was prompted
both by the isolation of appreciable quantities of the
corresponding ester 3 and carboxylic acid 5, and by the
discovery of an excess of 1-undecene (4), above what is
expected from the yield of n-undecane (6), by simple
disproportionation of two undecyl radicals in cage III.
Both observations could be explained by the presence of
the acyloxy radical 1 as a discrete intermediate in cage
II recombination. We have also used stereospecific
cases the initial concentration of 2 was 0.01 M. At higher
initial concentration of 2, 0.1M, kdecomp were (2.0 ( 0.2)
-
4
-1
×
10
0
s
in HCA (hexachloroacetone) and (2.6 ( 0.1) ×
s-1 in C
(benzene-d ). The values of these rate
constants are consistent with the reported value of 3.4
-
4
1
6
D
6
6
2
1
-
4
-1
×
10
s
for the rate constant for decomposition of 10%
w/v 2 in Primol C mineral oil at 89 °C. The lack of
dependence of kdecomp on solvent polarity supports the
predominantly nonionic nature of the reaction. The
insensitivity of kdecomp to changes in peroxide concentra-
tion rules out a substantial contribution from induced
peroxide decomposition.18 Measurements of kdecomp, how-
ever, cannot detect minor reaction pathways of either
type.
(
16) (a) Bartlett, P. D.; Hiatt, R. R. J . Am. Chem. Soc. 1958, 80,
398. (b) Szwarc, M. In: Peroxide Reaction Mechanisms: Edwards, J .
O., Ed.; Wiley-Interscience: New York, 1962; pp 153-174.
17) (a) Cooper, R. A.; Lawler, R. G.; Ward, H. R. J . Am. Chem. Soc.
972, 94, 545. (b) Kaptein, R. Adv. Free-Radical Chem. 1975, 4, 319.
18) (a) Walling, C.; Waits, H.; Milanovic, J .; Pappiaonnou, C. J .
1
P r od u ct Yield s. The yields of hydrocarbon products
of decomposition of 2, 1-undecene (4), n-undecane (6), and
(
1
(
Am. Chem. Soc. 1970, 92, 4928. (b) Liu, Y.-C.; Guo, Q.-X. Acta Chim.
Sin. 1985, 43, 1171.
(20) Two different samples were thermolyzed and at least 10-18
points were used for a semilog first-order plot for each sample. The
quoted error of the slope (kdecomp) represents two standard deviations.
(21) Guillet, J .; Gilmer, J . Can. J . Chem. 1969, 47, 4405.
(19) Smid, J .; Rembaum, A.; Szwarc, M. J . Am. Chem. Soc. 1956,
7
8, 3315.

Thermal Decomposition of Primary Alkyl Diacyl Peroxide
J . Org. Chem., Vol. 61, No. 8, 1996 2803
Ta ble 1. Rela tive Excess of 1-Un d ecen e (4) Obser ved in
Va r iou s Med ia
Ta ble 2. Yield s of 3, 5, a n d Excess 4 fr om Th er m olysis of
1 in Va r iou s Solven tsa
relative
excess at excess at
85 °C 105 °C
1.4 ( 0.1 1.3 ( 0.4d
relative
yield
(%) of 4
solvent and
conditions
solvent
HCAb
4
6
(excess)
5
3b
ETa
85 °C
105 °C
5.6 ( 0.4 2.8 ( 0.1d 2.8 ( 0.4 2.6 ( 0.3 11.3 ( 1.2
2 2
Cl(CH ) Cl 6.2 ( 0.1 3.2 ( 0.1 3.0 ( 0.2 4.2 ( 0.2 10.3 ( 1.1
b
c
n-C8H18
30.9 6.8 ( 0.2
.9 ( 0.1
2
C
C
6
D
D
5
CD
3
6.5 ( 0.3c 2.9 ( 0.1 3.6 ( 0.3 3.4 ( 0.2 11.9 ( 1.0
6.1 ( 0.2 2.8 ( 0.2 3.3 ( 0.3 2.9 ( 0.2 10.6 ( 1.0
6.8 ( 0.2 2.9 ( 0.1 3.9 ( 0.2 3.3 ( 0.3 12.3 ( 1.3
C6H5CH3
33.9 6.2 ( 0.2 5.7 ( 0.2 1.1 ( 0.1 1.2 ( 0.1
6
6
2
.9 ( 0.1 2.5 ( 0.1
6.1 ( 0.2
.8 ( 0.1
8 18
n-C H
e
f
1.1 ( 0.1
a
T ) 83 °C. [2]o is 10-4 M unless noted otherwise. b 0.1 M 2, T
) 83 °C, reaction time 5 min. 0.1 M 2, plus 0.2 M cyclohexene.
2
c
5.6 ( 0.2
.8 ( 0.1
1.0 ( 0.1
d
2
0.1/M 2, plus 0.1 M BrCCl3.
ClCH2CH2Cl 41.9 6.2 ( 0.1
0.9 ( 0.1
1.0 ( 0.1
3
.2 ( 0.1
46.9 4.4 ( 0.1
.2 ( 0.1
scale.22
T
solvent polarity as measured by the Dimroth E
CH3CN
As will be discussed further below, this is evidence
2
against excess 4 arising from the rearrangement product,
a
Dimroth solvent polarity parameter, reference 22. b % yield
1
2, which is known in other cases to be formed by a polar
of 4 (an average of at least three separate samples, the yield for
each sample is an average of at least four separate injections). %
yield of 6 (an average of at least three separate samples, the yield
1
7
c
3
mechanism. Indeed, in CH CN, the most polar solvent
used, the reduction in yields of both 4 and 6 (row 6, Table
for each sample is an average of at least four separate injections).
1) would be expected, with a corresponding increase in
the yield of “polar” product.
In order to confirm that 4 and 6 arise from geminate
radical coupling, experiments were carried out using
several scavengers and solvents in addition to those
shown in Table 1. The results are summarized in Table
d
Determined from molar ratio, [4]/[6]. e C6H5CH3 with 0.01 M
18
BrCCl3. f C6H5CH3 mixed with crushed glass ampule.
n-docosane (7), were determined from thermolysis of
-
3
-4
dilute (10 to 10 M) solutions and analyzed using
capillary gas chromatography. In the dilute solutions O
2
2
. Decomposition of 0.1 M 2 in HCA (hexachloroacetone)
serves as the radical scavenger, whereas at higher
peroxide concentrations the solvent itself, or an added
at 140 °C was examined because this is the concentration,
solvent, and temperature often chosen for CIDNP stud-
scavenger such as BrCCl
lauroyl lauroate (3), undecyl chloride or bromide (13a or
3b, products of undecyl radical scavenging in chlorine
3
, plays that role. The yields of
ies.1
0,12
Furthermore, HCA is usually considered to be
capable of efficient scavenging of radicals escaping
recombination, thereby assuring that alkane 6 arises
solely from geminate pairs. It was determined, however,
1
or bromine donating solvents), lauric acid (5), and
1
carbonic anhydride (12) were obtained by H-NMR on
that HCA is not as efficient a scavenger as BrCCl
average yield of 6 drops from 3.6 ( 0.3% to 2.8 ( 0.1%
when BrCCl is added to the HCA solution, suggesting
3
. The
partially reacted 0.1 M solutions. Under the conditions
employed for product analysis radical-radical products
derived from 2 arise exclusively from geminate recom-
bination, i.e., terminal encounters of escaped radicals
have been eliminated. Overall, 7 accounted for 24-28%
of alkyl radicals produced by each mole of 1, cage escape
products (either 13a or 6 in chlorine- or hydrogen-
donating solvents, respectively) for 25-37%, 4 for 6-7%,
3
that in the pure solvent some of the alkane product arises
from random encounters of nonscavenged radicals or,
more likely, by abstraction of hydrogen from the relative
high concentration of unreacted 2 or reaction products.
Although the yields of 4 and 6 in 0.1 M HCA solution at
-4
1
40 °C are very similar to those obtained from 10
M
6
for 3%, carbonic anhydride (12) for 3-12%, 5 for 3%,
solutions of 2 in toluene at 105 °C when efficient oxygen
scavenging is employed (see Table 1, rows 2,4), it must
be cautioned that in the presence of peroxide, HCA
removes some of the alkene, presumably by radical chain
addition. This was confirmed by observing that the
recovery of alkene from solutions of 4 in HCA heated at
and 3 for 10-13%. Identified products yielded material
balance of 74-92% in the solvents studied. The existence
of the acyloxy-alkyl radical pair (cage II) for a sufficient
time to permit formation of recombination products is
consistent with the following observations: (1) products
3
and 5 contain intact acyloxy groups, (2) yields of 4 are
1
40 °C for 12 min was lessened when 2 was present
consistently higher than those of 6, and (3) the yields of
whereas 4 itself does not react under these conditions.
This result is not unexpected, since alkenes tend to
undergo efficient radical-initiated chlorination by halo-
genated solvents.23 Indeed, < 0.1% yield of 4 is detected
5
and excess 4 are nearly the same.
Con sta n cy of Excess Of Alk en e. The yields of 4 and
6
were determined under a variety of conditions in order
to explore possible sources of excess 4. Table 1 sum-
marizes the relative excess of 4 obtained in various
media. It is apparent that the relative excess of 4 is
nearly independent of temperature and insensitive to
solvent polarity. The excess cannot be attributed to the
reaction of 2 at a glass surface, since the yields are
unchanged in samples decomposed in the presence of
finely powdered Pyrex glass of the type used in ampoules
employed for decomposition (row 4 in Table 1). Samples
when 2 is thermalized in the presence of BrCCl . It is
3
therefore likely that the actual yield of 4 in HCA is
somewhat higher than 5.6%.
When 0.1 M 2 is decomposed in C
6 5 3
D CD , the yield of
4 is 8.4 ( 0.5%, higher than observed when oxygen
-
4
scavenging is employed with 10 M 2 in this solvent
(Table 1, entries 2, 3). This suggests a contribution from
random radical encounters. Addition of 0.2 M cyclohex-
ene as a radical scavenger reduced the yield of 4 to 6.5
( 0.3%, but as expected the yield of 6 remained high at
decomposed with addition of 0.01 M BrCCl
3
as an extra
scavenger (row 3 in Table 1) do not differ from those
samples which rely on oxygen scavenging only, thus
precluding the formation of excess 4 from reaction of the
radicals with oxygen. An interesting feature of Table 1
is the invariance of relative excess of 4 with change in
26.3 ( 1.7%. With 0.1 M added BrCCl , however, the
3
yield of 6 from 0.1 M solution of 2 in C D CD was
6
5
3
(
22) Dimroth, K.; Reichardt, C. Ann. Chem. 1963, 1, 661.
(23) De Tar, D.; Wells, D. J . Am. Chem. Soc. 1960, 82, 5839.

2
804 J . Org. Chem., Vol. 61, No. 8, 1996
Ryzhkov
reduced to 2.9 ( 0.1%, in good agreement with the yields
obtained with efficient oxygen scavenging of the dilute
solution (Table 1, entries 2, 3). Since it was important
Sch em e 2
(a)
O
H
O
O
R′
2
C
for H-labeling experiments using NMR (see below) that
O
O
kconcerted
H
D
D
all of 4 come from geminate radical sources, 0.2 M
2
D
cyclohexene was employed as a scavenger when H-
+ HO
R′
H
RCH2CH2
6 5 3
labeled 2 was thermalized in C D CD .
D
Table 2 also lists the yields of ester 3 and acid 5 in the
solvents discussed above, along with the corresponding
yields of excess 4. It is apparent that the observed yield
of 5 correlates well with the excess of 4. Also notable is
the insensitivity of the yield of ester 3 to the type of
solvent or the temperature employed. This supports the
assumption that 3, 5, and excess 4 are produced from
reactions within the geminate alkyl acyloxy radical pair
9
5
CH2CH2R
O
(b)
RCHDCHD
O
O
R′
O
2
: R′ = RCHDCHD
kpolar
O
(cage II). It should be noted, however, that variable
amounts of undecyl alcohol were detected in some
samples. One possible source of alcohol is the hydrolysis
of carboxy inversion product 12.18 Since the other
product of the hydrolysis would be acid 5, it is possible
that unless care is taken to exclude moisture the yield
of 5 might be higher than expected from the radical
pathway alone.
C+
ion pair
O
O
D
H
–O
R′
RCH2CH2
D
"
SN2"
H
The yields of the various products from thermolysis of
reported here compare favorably with previously
observed yields of similar products from thermolysis of
"E2"
"
SN2"
"
E2"
HO
2
2
1
19
O
D
H
2
and the related symmetric dipropanoyl and dihep-
D
H
D
tanoyl2 peroxides containing primary alkyl groups. In
these reports the yields of alkane disproportionation
product, when determined under efficient scavenging
conditions, range from 2.8% to 3.8%. Significantly, a two-
to-three-fold excess of alkene disproportionation products
is also observed.23 Yields of ester, ranging from 11% to
3
O
R′
+
R′
RCH2CH2
D
RCH2CH2
O
H
8
5
erythro-3
signals of 12, 3, and 4 was monitored over a period of 18
h. It was found that 12 is formed in 6% yield at the rate
of peroxide decomposition and then decomposes slowly
over a period of 12 h to produce an equivalent amount of
1
3%, are similar to those seen here. Yields of heptanoic
acid, estimated from the titration of scaled up thermolysis
samples of heptanoyl peroxide, ranged from 1% to 16%
but were generally not considered reliable indicators of
-4
ester 3. The first-order rate constants (2.6 ( 0.1) × 10
2
3
a free-radical process.
A reliable acid yield of 3%,
-1
-5 -1
s
for decomposition of 2 and (1.0 ( 0.1) × 10
s
for
however, reported by Pryor et al.24 from 0.6 M 2 in C
D
6
6
18b
1
2 are in good agreement with previous reports,
at 80 °C, is similar to our findings.
indicating that a negligible amount of 3 over the time
required for decomposition of 2 is the result of thermoly-
sis of 12. It is also found that the yield of 4 remains
constant after all of 2 is decomposed. Concentration of
in another sample was followed by GC over a total of
2 half-lives of peroxide decomposition, and its yield
remained constant at 6.5 ( 0.2%.
Ster eoch em istr y of F or m a tion of Excess Alk en e.
Scheme 2 illustrates two other possible sources of excess
in thermolysis of 2. Process a is a concerted pathway,
previously suggested by Szwarc, not unlike the well
documented pyrolysis of esters.
proceeds through a six-membered transition state, follows
Th er m a l Sta bility of P r od u cts. It has been sug-
gested that carboxy inversion products analogous to 12
may decompose by either concerted or ionic mechanisms
to form ester25 and alkene. Previous investigations of
the thermal stability of carboxy inversion products from
primary alkyl-aroyl peroxides, however, suggest that the
rate of thermal decomposition of 12 should be ca. 20-fold
slower than that of the precursor peroxide.27 Measure-
ment of the stability of 12 confirmed this prediction. Liu
and co-workers1 determined that 12 decomposes with
26
4
1
4
8b
19
-
6
-1
first-order rate constant of 7 × 10
compared to 2-3 × 10 for decomposition of 2 at the
same temperature. We conducted a more detailed analy-
sis of the kinetics of decomposition of 2 in C in order
s
at 80 °C in C
6
H
6
,
28
The latter reaction
-
4
-1
s
the Ei mechanism and gives the syn-elimination alkene
D
6
29
6
product exclusively.
Although the temperature re-
to eliminate the possibility that 3 and excess 4 are
derived from 12. A solution of 2 (0.1 M) was decomposed
quired for ester pyrolysis is relatively high (300-550 °C),
the greater overall reactivity of peroxides might be
expected to lower the activation energy for the analogous
process. If 2 is to produce 4 by the Ei mechanism, it is
likely that the eight-membered transition state shown
in Scheme 2 would be involved and cis-1-undecene-1,2-
d (9) obtained from the stereospecifically deuterated
2
peroxide shown. Although the literature pertaining to
the chemistry of the carbonyl group does not seem to
6 6
at 83 °C in C D
and the time evolution of 1H-NMR
(
(
24) Pryor, W. A.; Davis, W. H. J . Am. Chem. Soc. 1974, 96, 7557.
25) Oae, S.; Kashiwari, T.; Kozuka, S. Chem. Ind. (London) 1965,
1
694.
26) (a) Lamb, R.; Sanderson, J . J . Am. Chem. Soc. 1969, 91, 4034.
b) Lamb, R.; Pacifici, J .; Ayers, P. Ibid 1965, 87, 3928. (c) Hart, H.;
Cipriany, R. Ibid 1962, 84, 3697. (d) Bartlett, P.; Leffler, J . Ibid 1950,
(
(
7
1
2, 3030. (e) Kharasch, M.; Kuderna, J .; Nudenberg, W. J . Org. Chem.
955, 19, 1283. (f) Green, F. J . Am. Chem. Soc. 1955, 77, 4869. (g) de
Tar, D.; Weiss, C. Ibid 1957, 79, 3045. (h) Green, F.; Stein, H.; Chu,
C.; Vane, M. Ibid 1964, 86, 2080.
(28) DePuy, R, King, J . Chem. Rev. 1960, 60, 431.
(29) March, J . Advanced Organic Chemistry; Wiley-Interscience;
New York, 1985, Chap. 17, pp 896-905.
(27) Kline, M., Ph. D. Dissertation, Brown Univ., 1982.

Thermal Decomposition of Primary Alkyl Diacyl Peroxide
J . Org. Chem., Vol. 61, No. 8, 1996 2805
contain an example of a reaction proceeding through an
eight-membered transition state, nevertheless, we felt
that this possibility should be explored.
Also shown in Scheme 2 is a possible mechanism (b)
for formation of 5 and excess 4 via a rearranged ion pair.
This pair is well-known to give rise to carboxy inversion
product 12, which however, we have shown cannot itself
produce a significant amount of 4. If the ion pair were
to give rise to 3 the process would be likely to proceed
with inversion of stereochemistry. Similarily, alkene 4
would be formed by an antiperiplanar stereoelectronic
pathway. These pathways are labeled in Scheme 2b as
“S
N
2” and “E2”, respectively. In an attempt to determine
the stereochemistry of formation of 3 and 4 from 2 we
2
2
prepared H-labeled 2 as the threo-2,3-d stereoisomer
and examined the stereochemistry of the resulting alkene
1
isotopomers by H-NMR. Since the sample of threo-2,3-
d
2
2 also contained monodeuterated 2 and a small
amount of erythro 2 (see Experimental Section), a total
of 6 isotopomers of 4 were observed.
HC
HT
H
D
HX
D
HY
D
D
RCH2CH2
RCH2CH2
RCH2CH2
4
8
9
HA
HB
D
D
HD
H
HE
D
H
RCH2CH2
-d
RCH2CH2
RCH2CH2
4
10
11
As is apparent from Scheme 2, the ratio of 8/9 (or 10/
1) may be used to differentiate the concerted or ion-pair
1
pathways to 4 from the radical route. The latter would
be accompanied by rapid rotation along the C -C bond
and should give equal amounts of C alkene stereoiso-
1
2
1
mers. Pathways a and b, however, would yield unequal
amounts of these isomers. The H-NMR spectrum of the
1
F igu r e 1. Allyl CH -decoupled 1H-NMR spectrum of the
2
alkene region of 0.1 M threo-2,3-d
at 140 °C is shown in Figure 1. The spectra of the same
region both from HCA and 89 °C thermolysis of threo-
2
2 thermalized in HCA
terminal CH
thermalized in C
of isotopomers correspond to the structures shown to the right
of the spectrum.
2
alkene region of 4 from 0.1 M threo-2,3-d
2
-2
6
D
5
CD at 80 °C. The labels of various protons
3
2
,3-d
spectra using the program PANIC are shown in Figure
. The spectra were obtained with simultaneous decou-
pling of the allyl CH protons. The relative amounts of
2 6 5 3
2 in C D CD together with the simulation of these
formation of 3 from threo-2,3-d
butyl group (in the model compound) is much bulkier
than the n-nonyl group in undecyl-1,2-d 3, making the
2
2. However, the tert-
2
2
2
the isotopomers determined from these spectra are listed
in column 5 of Table 3. The observed ratios of deuterated
isotopomers of 4 are compared with ratios predicted for
pure radical and mixed radical/nonradical pathways.
When allowance is made for the fact that isotope effects
on hydrogen removal are not included in the predicted
ratios, the results support a radical route to excess 4.
The attempt to analyze the stereochemistry of the alkyl
group in the ester 3 unfortunately gave ambiguous
equilibrium populations of rotational conformers very
different and a direct comparison impossible. Similarly,
equating the populations of rotational conformers in 13
and 3 is unrealistic because of the differences in steric
bulk between the chlorine and lauroyloxy groups.
In an attempt to determine whether the stereoisomeric
deuterated esters could be distinguished by NMR, a
known mixture of esters with threo and erythro configu-
rations was prepared by thermolysis of a mixture of threo-
2
1
results. Although the H-decoupled H-NMR of the
undecyl-1,2-d group of chloroalkane 13 shows both
and erythro-1,2-d
2
2 (see Experimental Section). The
2 and erythro-2,3-d 2 (in
:1 ratio), as well as 2-d and 3-d 2. The impurity of 2-d
, however, was higher in this mixture than in the
2
mixture contained threo-2,3-d
1
2
2
2
erythro- (J 1,2 of 8.3 Hz) and threo- (J 1,2 of 6.2 Hz) isomers,
only one doublet (J 1,2 of 5.7 Hz) is seen for the alcohol
undecyl-1,2-d
2
groups of ester 3, acid 5, and undecyl
alcohol. Previously reported30 values of J 1,2 in the esters
sample of pure threo peroxide. Figure 3 shows the
superimposed signals of the alcohol R-protons of 3 from
threo- and erythro-1-acetoxy-3,3-dimethylbutane-1,2-d
2
threo-2,3-d
and erythro-2,3-d
of the doublets of undecyl-1,2-d
2 2
2 and from the 1:1 mixture of threo-2,3-d
are 6.1 and 8.8 Hz, respectively. This appears to be
consistent with a threo assignment for the coupling
constant observed from the alcohol group of the undecyl-
2
2
2 obtained H-decoupling. Both peaks
group have nearly
2
identical line widths (ca. 2 Hz), suggesting that regard-
less of whether 3 retains or scrambles the original
1
2
,2-d 3, i.e., with retention of configuration in the
stereochemistry, the difference in J 1,2 between the threo
(
30) (a) Whitesides, G. J . Am. Chem. Soc. 1967, 89, 1135. (b)
1
Whitesides, G. Ibid 1974, 96, 2814.
-and erythro-1,2-d
2
3 is too small to be detected by H-

2
806 J . Org. Chem., Vol. 61, No. 8, 1996
Ryzhkov
1
F igu r e 2. Experimental (top) and simulated (bottom) allyl CH
from 0.1 M threo-2,3-d -2. (A) Spectrum obtained after thermolysis of 2 in C
with the relative amounts of alkenes corresponding to 50%:50% polar/radical pathway (see Table 3), (D) 100% radical, and (E)
2
-decoupled H-NMR spectra of the terminal CH
2
alkene region of
4
2
6
D
5
CD
3
at 80 °C, (B) in HCA at 140 °C, (C) simulated
5
0%:50% concerted/radical pathway.
Ta ble 3. Exp ected a n d Obser ved Rela tive Yield of
presence of the appropriate amount of 5 in the product
mixture, as expected from 1:1 stoichiometry of the
disproportionation reaction within the undecyl-lauroy-
loxy radical pair, (2) absence of correlation between the
Isotop om er s of 4 fr om Th er m olysis of th r eo-1,2-d 2-2a
5
0% radical
and 50%
polar
radical
pathway
only
50% radical
and 50%
concerted
alkene
observedb
amount of excess 4 and the polarity of the medium, and
9
1.85
0.73
2.44
1.31
1
1.29
1.29
1.87
1.87
1
0.73
1.85
1.31
2.44
1
1.7 ( 0.1
1.4 ( 0.1
1.4 ( 0.1
1.5 ( 0.1
1
2
(
3) the complete scrambling of the H label in the alkene
8
1
1
4
4
products from stereospecificaly deuterated 2. At an
1
0
-
4
initial peroxide concentration of 10
M, and under
-d
conditions of efficient scavenging by dissolved oxygen, the
excess of 4 was observed to be nearly constant in all
solvents studied, despite significant variations in the
individual yields of 4 and 6. The literature contains two
examples of detection of excess alkene from thermolysis
of symmetrical diacyl peroxides. The alkene excess in
1
2.05
1
1
1
0.48
1.0 ( 0.1
[
11]/[10]
9]/[8]
1.1 ( 0.1
.0 ( 0.1c
1
[
0.34
1
2.98
1.1 ( 0.1
1
.0 ( 0.1c
a
HCA at 140 °C unless indicated otherwise. b Relative yields
are calculated for an initial mixture of 2 consisting of 58% threo-
,3-d2, 4% erythro-2,3-d2, 24% 3-d and 14% 2-d. Possible kinetic
isotope effects have been ignored. All yields are relative to
-undecene-2-d, the alkene expected to show least isotope effect
19
thermolysis of propionyl peroxide was determined to be
2
2
.6-fold (at 65 °C), and a 2-fold excess of hexene over the
yield of hexane was observed in thermolysis of heptanoyl
peroxide.23 It therefore may be confidently stated that
the alkene excess is not an isolated phenomenon, but is
apparently characteristic of the thermal decomposition
of straight chain symmetrical primary diacyl peroxides.
Ester 3, the coupling product of the recombination
reaction within the undecyl-lauroyloxy radical pair, also
does not appear to be a consequence of the polar rear-
rangement. Its yield is insensitive to solvent polarity and
is nearly constant in the temperature range of 60-120
1
on it’s yield. Errors are from replicate integrations of single
c
spectra obtained from HCA or C6D5CD3 solutions. C6D5CD3, 80
°
C, ratios only determined.
NMR. Attempts to convert (in situ) the small yield of 3
to alkyl chloride, for which the signals of erythro and
threo isomers were resolvable, were not successful.
Con clu sion s
°
C, precluding its formation from carboxy inversion
The results presented in this paper suggest that excess
alkene 4, ester 3, and acid 5 seen in thermal decomposi-
tion of lauroyl peroxide are the products of a radical
process and not a result of a random artifact or interven-
tion of the ionic intermediates as was proposed previ-
ously. The evidence for this conclusion includes (1) the
product 12, which is stable at the temperatures em-
ployed. Evaluation of the rate of decomposition of
lauroyloxy radical 1 (R ) C11H23) from the temperature
and viscosity dependence of cage product yields will
appear in future publications.

Thermal Decomposition of Primary Alkyl Diacyl Peroxide
J . Org. Chem., Vol. 61, No. 8, 1996 2807
crystallization from CHCl
3 3
/CH OH (1:1). Purified 2 contained
less than ca. 0.2% lauric acid as evidenced by the 400 MHz
1
3
H-NMR of a 0.6 M sample in CDCl . Lauroyl peroxide was
also prepared by the procedure due to Swern.32 Lauroyl
chloride (3.68 g, 0.017 mol) was added to diethyl ether (30 mL)
in a 50 mL round bottom flask and cooled to 0-2 °C with an
ice-water bath. Hydrogen peroxide (3 mL, 90% in H
mol) was added with vigorous magnetic stirring. Pyridine
1.67 mL, .0204 moles) was added dropwise in 100 µL portions
2
O, 0.013
(
over 1 min time intervals. Following the complete addition
of pyridine, the ice bath was removed, and the reaction mixture
was stirred for an additional 1 h. The pyridinium hydrochlo-
ride precipitate was filtered, and the remaining ether layer
was washed twice with 20 mL portions of 5% HCl, twice with
2
5 mL portions of 5% NaHCO
3
and twice with 25 mL portions
of distilled water. The product was recrystallized from cold
-
1
27
diethyl ether. IR (Nujol mull): 1812, 1780 cm [lit. 1813,
1
(
-
1
1
3
778 cm ]. 400 MHz H-NMR (CDCl ): δ 2.4 (t, 2H), δ 1.7
quintiplet, 2H), δ 1.3 (broad, 16H), δ 0.9 (t,3H).
Meth yl (tr ip h en ylp h osp h or a n ylid en e)a ceta te. This
compound was prepared by dissolving commercially available
carbomethoxymethyl)triphenylphosphonium bromide (Ald-
(
rich) (12 g, 0.029 moles) in 180 mL distilled water and titrating
with 1.0 M NaOH (phenolphthalein indicator). The resulting
fine white precipitate of the ylide was vacuum filtered and
2 2
redissolved in 100 mL of CH Cl to be used in the step below.
F igu r e 3. Superimposed 1H-NMR signals of the alcohol
Tr a n s-2-Dod ecen oic Acid . The acid was prepared by
R-protons of 3 from 1:1 mixture of threo-2,3-d
2
- and erythro-
-2 (bottom trace) obtained
adding the ylide prepared above to 4.2 mL (0.022 mol) of
2
,3-d
2
- (top trace) and threo-2,3-d
2
decanal in 100 mL of CH
2
Cl
2
in a 500 mL roundbottom flask.
Cl was removed
2
with H-decoupling. Peaks labeled B in the top trace arise from
The solution was refluxed for 5 h, and the CH
2
2
overlap of the signals from threo- and erythro-3-d
2
. Peaks
by distillation. The resulting methyl trans-2-dodecenoate was
saponified by addition of 170 mL of 0.3 M KOH in EtOH/water
(9/8 v/v) and refluxing for 3 h. Ethanol was removed by
distillation, and the aqueous layer was washed with a total of
labeled A correspond to the residual monodeuterated 3, those
marked with * are from unidentified impurities. Details of
spectrum acquisition are given in the Experimental Section.
5
00 mL of diethyl ether and acidified with concentrated HCl.
Exp er im en ta l Section
The acid was extracted into a total of 200 mL of CH Cl , and
2
2
the solvent was removed by rotary evaporation yielding 3.58
Gen er a l Meth od s. Peroxide samples were heated in
sealed glass ampules (4 mm i.d., ca. 5 cm in length) in an oil
bath equipped with proportional temperature controller and
the mercury thermometer readable to (0.05 °C. For the
determination of the yields of 3, 2 was decomposed in 1.7 mL
V-vials equipped with the screw-caps and Teflon septa. Since
oxygen scavenging was desired, the samples were not de-
gassed. The total decomposition time at each temperature was
adjusted to ensure greater than 98% decomposition of the
peroxides. Peroxide samples in HCA were immediately cooled
in an ice-water bath and stored at -20 °C to prevent further
reaction between the alkene product and solvent. For the
kinetic runs the samples were decomposed in 5 mm sealed
NMR tubes for periods of time varying between 5 and 60 min
and were rapidly cooled in an ice-water bath prior to NMR
analysis at room temperature. A single sample was used for
each series of decomposition times, being alternately cooled
and heated.
g (81%) of a mixture of trans- and cis-2-dodecenoic acids: 400
1
MHz H-NMR (CDCl
1
3
): δ 6.92 (dt, 1H, R alkene proton, J )
5.7 Hz (trans acid)], 6.2 (dt, 1H, R alkene proton, J ) 11.6
Hz (cis acid)], 5.78 (dt, 1H, â alkene proton, J ) 15.7 Hz (trans
acid)], 5.73 (dt, 1H, â alkene proton, J ) 11.6 Hz (cis acid)], δ
2
.62 (qd, 2H, γ-CH
2
2
of cis acid), 2.2 (qd, 2H, γ-CH of trans
acid), 1.2-1.5 (broad, 18H), 0.9 (t, 3H). Ratio of trans/ cis
acid: 14/1. The product contained 7% cis acid and was used
without further purification.
th r eo-2,3-Did eu ter iola u r ic Acid . This acid was prepared
3
3
from the trans-2-dodecenoic acid by the procedure of Simon.
trans-2-Dodecenoic acid (0.93 g, 0.0047 mol, containing 7% cis
acid) was added to 20 mL of dry benzene. Wilkinson’s catalyst,
ClRh(PPh ) (0.08 g, 0.000086 mol), was placed into a movable
3 3
side-arm of a 50 mL roundbottom flask attached to the vacuum
line and evacuated. Oxygen was removed from the sample
by three freeze-pump-thaw cycles (pump cycle, 3 min), 99.9%
2
2
H (ca. 150 mL, 0.006 mol) was admitted into the flask, and
Ma ter ia ls. Toluene and benzene were obtained from
Aldrich and distilled from CaH
2
. Hexachloroacetone (HCA)
was obtained from Aldrich and vacuum distilled (80 °C).
n-Octane contained an appreciable amount of n-dodecane as
determined by GC, and was slowly redistilled through a
the side-arm was rotated to deliver the open glass ampule
containing the catalyst into the solution. Vigorous magnetic
stirring with a Teflon-coated stirring bar was commenced at
room temperature. The progress of the reaction was monitored
2
by following the consumption of H
2
via a mercury manometer
Vigreaux fractionating column. ClCH
from Fischer Chemical Co. and purified by repeated washing
with concentrated H SO followed by distillation from molec-
Crown ether (18-crown-6, Aldrich, 95%) was
2 2 2
CH Cl was obtained
2
attached to the vacuum line. After uptake of H
2
ceased (ca.
3
days), the line was opened to air and the solvent was
2
4
3
1
removed by rotary evaporation. The resulting solid residue
was stirred with 0.5 M NaOH. The resulting solid soap was
washed with ether, suspended in distilled water, acidified with
ular sieves.
recrystallized from acetonitrile. Hydrogen peroxide (90%) was
obtained from FMC corporation.
concentrated HCl, and extracted with CH
the solution over anhydrous MgSO , removal of solvent yielded
.6 g (62% yield) of a pure compound. H- and C-NMR
2 2
Cl . After drying
La u r oyl Ch lor id e. A 12 g amount of lauric acid was
refluxed in approximately 50 mL of thionyl chloride. The
product was vacuum distilled (60 °C, 1.0 torr) to give faintly
4
1
13
0
-
1
1
spectra of the product were identical to those of undeuterated
lauric acid, except for small isotope effects on chemical shifts
and line broadening due to interactions with 2H.
yellow liquid. IR (neat) 1800 cm
CDCl ): δ 2.78 (t, 2H), other signals same as for 2 below.
La u r oyl P er oxid e (2). The commercially available com-
pound (Alperox 95% or Aldrich, 98%) was purified by repeated
.
400 MHz H-NMR
(
3
(
32) Silbert, L.; Swern, D. J . Am. Chem. Soc. 1958, 81, 2364.
(
31) Perrin, D.; Armargero, W. Purification of Organic Chemicals;
(33) (a) Simon, H.; Berngruber, O. Tetrahedron 1970, 26, 1401. (b)
Simon, H. Z. Naturforsch. 1969, 24b, 1195.
Pergamon Press: New York, 1988.

2
808 J . Org. Chem., Vol. 61, No. 8, 1996
Ryzhkov
Bis(th r eo-2,3-d id eu ter iola u r oyl) P er oxid e. This com-
2.9 mL/min. Linear velocity of the carrier gas (He) was
optimized at 33 cm/s by maximizing the column efficiency as
judged by the ratio of peak retention time to the peak width
at half-height. Since the amount of analyte was very small
(ca. <1 nM for 4 and 6) the splitless injection technique was
used. Injection volumes varied from 0.6 µL (for concentrated
samples) to 1.2 µL for more dilute samples. Retention times
and response factors for all components were determined by
using authentic samples and were checked occasionally and
found not to vary appreciably with time or solvent. When
acetonitrile solutions were analyzed, all of the GC peaks were
small and broad. The cut-and-weigh procedure had to be used
for integration of these peaks since electronic integration did
not yield reproducible data. Time delay for the solvent purge
was optimized by determining the areas of the analyte peaks
while varying the purge vent time. It was found that the
optimal purge time for analysis of 4 and 6 was 0.5 min and
for 7 was 0.3 min. To ensure the precision of analyses,
calibration plots of relative areas of FID signals to the molar
ratios of analytes were prepared using n-dodecane as the
standard for the signals of 4 and 6 and n-eicosane for the 7.
These calibration plots showed that the intercepts were
insignificantly different from 0. Peak areas were determined
using electronic integration and checked by the cut-and-weigh
method.
pound was prepared according to the procedure given for the
1
13
lauroyl peroxide above. H- and C-NMR spectra of the
product were essentially identical to those of 2, except for small
isotope effects on chemical shifts and line broadening due to
2
1
2
interactions with H. 250 MHz H-NMR ( H decoupling)
showed the product to contain 58% of threo-2,3-dideuterio-2
(J H,H 6.49 Hz), 4% of erythro-2,3-dideuterio 2, 24% of 3-d-2,
and 14% 2-d-2. It was not possible to obtain a precise estimate
of the amount of fully protonated 2 in the mixture of deuter-
1
13
ated isomers by either H- or C-NMR. We are confident that
the maximum amount of 2-d present is about the same as that
of erythro-2,3-dideuterio 2. In any event, excess of 2 will only
effect the 4/4-d ratio (see Table 3) which is not used as the
primary indicator for the radical nature of the 3, 5, and excess
4
.
Mixtu r e of tr a n s- a n d cis-2-Dod ecen oic Acid s. The
mixture was prepared from trans-2-dodecenoic acid methyl
3
4
ester by the procedure of Lewis. The methyl ester of trans-
-dodecenoic acid was obtained by stirring 1.5 g. (0.0074 mol)
of the acid in 40 mL of 14% BF /CH OH complex (Aldrich) for
h. The resulting solution was mixed with 50 mL of distilled
water and extracted twice with 20 mL of CH Cl . Removal of
CH Cl yielded 1.38 g (87%) of ester. The resulting ester was
combined with 0.419 g (0.0031 mol) of AlCl
Cl in a Vycor irradiation vessel, degassed by bubbling N
2
3
3
7
2
2
2
2
3
in 40 mL of CH
2
-
2
2
1H-NMR An a lysis of P er oxid e Decom p osition P r od -
through the solution for ca. 3 min and irradiated at 254 nm
for 4 h at room temperature in a Rayonet reactor. Upon
completion of the photolysis, the solution was washed twice
with 40 mL portions of water and the organic phase was dried
u cts. The initial concentration of the peroxide was either 0.01
M or 0.1 M as noted in the text. The chemical shifts of the
signals for the peroxide decomposition products were confirmed
in all solvents using authentic materials. All of the NMR
product yield determinations were performed using the 400
MHz spectra. Spectral size was 64 K in all such samples, and
an acquisition time of 4.6 s was sufficient to eliminate all
relaxation distortions between pulses for all products of
interest. Spectra were recorded using a ca. 15° pulse angle.
To enhance the precision of the determinations the integration
was performed after the FID was exponentially broadened (line
broadening constant varied from 0.3 to not more than 0.5 Hz)
and zero-filled to 128 K. All peaks were baseline corrected
prior to integration. For all of the 0.01 M or 0.1 M samples,
the yield of 4 (determined by GC) was used to convert the
integrals of the other signals to actual yields. For example,
the yields of 3 were obtained by comparing the integral of the
R protons of the ester to the terminal protons of the alkene.
For partially decomposed samples the yield of 4 was deter-
mined by GC of completely reacted samples.
over anhydrous MgSO
of mixed methyl cis- and trans-2-dodecenoates was 0.99 g
63%). The mixture was found to contain 55% trans and 45%
4
. After the removal of solvent, the yield
(
1
cis esters by H-NMR. This mixture was saponified by the
method given for the trans-2-dodecenoate to yield 0.76 g of the
mixture of cis- and trans-2-dodecenoic acids. The resulting
material was converted to a mixture of bis(threo-2,3-dideu-
terio)lauroyl and bis(erythro-2,3-dideuteriolauroyl) peroxides
according to the procedure given for preparation of bis(threo-
2
,3-dideuteriolauroyl) peroxide above.
An a lysis of Decom p osit ion P r od u ct s of P er oxid es.
Ga s Ch r om a togr a p h ic An a lysis. GC analysis was per-
formed using a chromatograph equipped with FID detector and
a wall-coated open tubular column (12 m length × 0.2 mm
i.d.) coated with 0.33 µm thick cross-linked dimethylpoly-
siloxane liquid stationary phase. Injector and detector tem-
peratures were 200 °C for the analysis of 4 and 6 and 340 °C
for the analysis of 7. Column temperature was held constant
at 180 °C for docosane runs and was ramped from 80 °C to
Ack n ow led gm en t. This work was performed while
the author was a graduate student at Brown University,
Providence, RI. The author is greatly indebted to Prof.
Ronald G. Lawler for suggesting this project and provid-
ing lab space, encouragement, and assistance during the
preparation of this manuscript.
1
20 °C at 10 °C/min for undecene and undecane runs. Flow
rates were the following: air (to FID detector) 450 mL/min,
hydrogen (to FID detector) 30 mL/min, nitrogen (make-up gas)
3
0 mL/min. Split vent flow was 8.5 mL/min, septum purge
(34) Lewis, F.; Howard, D.; Barancyk, S.; Oxman, J . J . Am. Chem.
Soc. 1986, 108, 3016.
J O9512766