The autoxidation of aliphatic esters. Part 3. The reactions of alkoxyl and methyl radicals, from the thermolysis and photolysis of peroxides, with neopentyl esters

Article abstract of DOI:10.1039/b103555j

This paper reports a study of the dimerisation of ester radicals arising from the thermolysis and photolysis of di-tert-butyl peroxide (DTBPO) and dicumyl peroxide [DCPO, bis(α,α-dimethylbenzyl) peroxide] in neopentyl butanoate and a selection of structurally related neopentyl esters, in the temperature range 298 to 438 K. The acyl moieties of these esters were chosen to incorporate a variety of structural types to provide mechanistic information about the reactions. At 438 K, the thermolyses of DTBPO and DCPO in neopentyl butanoate give six ester radical dimers (three pairs of diastereoisomers). The two major diastereoisomers threo- and meso-dineopentyl 2,3-diethylbutane-dioate have been prepared and the crystal structure of the meso compound determined. Interestingly, the dimer product distribution is independent of the peroxide used. By contrast, at 298 K more than twice as many dimers are observed and the product distributions from the two peroxides are no longer the same. Similar results are also observed for the other neopentyl esters. Evidence is presented to show that the ester radicals arise from hydrogen atom abstraction from the esters by alkoxyl and methyl radicals; the latter being formed by the fragmentation of the alkoxyls. At 438 K the dimer product distributions are determined predominantly by a thermodynamically controlled equilibrium of ester radicals prior to dimerisation. Lowering the temperature leads to the increased importance of kinetic effects in determining the product distribution.

Full text of DOI:10.1039/b103555j

View Article Online / Journal Homepage / Table of Contents for this issue
The autoxidation of aliphatic esters. Part 3.1 The reactions of
alkoxyl and methyl radicals, from the thermolysis and photolysis
of peroxides, with neopentyl esters
2
John R. Lindsay Smith,* Eiji Nagatomi,† Angela Stead and David J. Waddington*
Department of Chemistry, University of York, Heslington, York, UK YO10 5DD
Received (in Cambridge, UK) 20th April 2001, Accepted 25th June 2001
First published as an Advance Article on the web 8th August 2001
This paper reports a study of the dimerisation of ester radicals arising from the thermolysis and photolysis of di-tert-
butyl peroxide (DTBPO) and dicumyl peroxide [DCPO, bis(α,α-dimethylbenzyl) peroxide] in neopentyl butanoate
and a selection of structurally related neopentyl esters, in the temperature range 298 to 438 K. The acyl moieties of
these esters were chosen to incorporate a variety of structural types to provide mechanistic information about the
reactions. At 438 K, the thermolyses of DTBPO and DCPO in neopentyl butanoate give six ester radical dimers
(three pairs of diastereoisomers). The two major diastereoisomers threo- and meso-dineopentyl 2,3-diethylbutane-
dioate have been prepared and the crystal structure of the meso compound determined. Interestingly, the dimer
product distribution is independent of the peroxide used. By contrast, at 298 K more than twice as many dimers
are observed and the product distributions from the two peroxides are no longer the same. Similar results are also
observed for the other neopentyl esters. Evidence is presented to show that the ester radicals arise from hydrogen
atom abstraction from the esters by alkoxyl and methyl radicals; the latter being formed by the fragmentation of the
alkoxyls. At 438 K the dimer product distributions are determined predominantly by a thermodynamically controlled
equilibrium of ester radicals prior to dimerisation. Lowering the temperature leads to the increased importance of
kinetic effects in determining the product distribution.
Introduction
Results
Modern automotive engines subject the lubricant to a high
degree of stress. Emission control legislation and new engine
designs, to meet the demands for improved economy and per-
formance, require the development of new and improved lubri-
cant formulations. Thus, much current research is focused on
prolonging the working lifetime of the lubricant. This involves
detailed studies on the thermo-oxidative degradation of the
base fluid and the use of additives designed to minimise this
degradation and thus to prevent deposit formation.2
The products from the thermolysis and photolysis of di-tert-butyl
peroxide (DTBPO) and dicumyl‡ peroxide (DCPO) in neopentyl
esters 1–4 under nitrogen
The thermolyses and photolyses of DTBPO and DCPO in
neopentyl esters 1–4, under nitrogen, have been studied in the
It is generally agreed that lubricants are degraded by radical
autoxidation processes and that the primary products react
further to give undesirable acidic and polymeric materials.3 The
details of the initial processes are well defined for mineral oil
base fluids and, although less thoroughly studied, the same
general mechanisms are assumed to occur with synthetic ester
lubricants.1,4
At York we have embarked on a research programme aimed
at understanding the radical oxidation mechanisms of models
for pentaerythritol, trimethylolpropane and neopentylglycol
ester lubricants. In previous studies we have investigated the
selectivity of hydrogen atom abstraction from esters by alkoxyl
radicals and the mechanisms of ester autoxidation.1,5 In this
paper we examine the products from the homolysis of peroxides
in esters arising from alkoxyl and methyl radicals.
temperature range 298–438 K. Analysis by GC shows that
the gaseous products are a mixture of methane and ethane
and in the liquid phase the peroxides form an alcohol (tert-
butyl alcohol or 2-phenylpropan-2-ol) and a ketone (acetone
or acetophenone) and the esters give some long retention
time products.
The notation employed to identify the different positions on
the ester substrates is the same as that used previously and is
illustrated in Scheme 1.
Products arising from the peroxide. The yields of methane,
ethane, alcohol and ketone from the reactions of both per-
oxides in each ester at 438 K are reported in Tables 1 and 2. The
data show that for all the esters methane is the major gaseous
product and, for a given substrate, the yield of methane is
significantly higher for reactions using DCPO. The account-
abilities of all the reactions were very high with (yields of
ketone ϩ alcohol)/2(peroxide consumed) being close to unity
and the [yield of methane ϩ 2(yield of ethane)]/(yield of
ketone) being >0.80 for almost all the substrates.
Scheme 1 Notation used to identify C–H bonds in esters.
† Present address: Lubricants Business Group, OGDL/1, Shell Global
Solutions (UK), a division of Shell Research Ltd., Cheshire Innovation
Park, PO Box 1, Chester, UK CH1 3SH.
‡ The IUPAC name for cumyl is α,α-dimethylbenzyl.
DOI: 10.1039/b103555j
J. Chem. Soc., Perkin Trans. 2, 2001, 1527–1533
This journal is © The Royal Society of Chemistry 2001
1527

View Article Online
Table 1 Product yields (10Ϫ4 mol) from the thermolysis of DTBPO in neopentyl butanoates 1–4 under nitrogen at 438 K
Ester
Acetone
tert-Butyl alcohol
Methane
Ethane
Methane/acetone
1
2
3
4
6.2
6.6
6.6
6.0
4.2
3.2
3.4
4.6
6.0
6.2
6.3
5.7
0.16
0.21
0.16
0.12
0.96
0.94
0.95
0.95
Table 2 Product yields (10Ϫ4 mol) from the thermolysis of DCPO in neopentyl butanoates 1–4 under nitrogen at 438 K
Ester
Acetophenone
2-Phenylpropan-2-ol
Methane
Ethane
Methane/acetophenone
1
2
3
4
9.0
6.9
9.5
7.6
1.6
0.8
1.3
1.4
7.6
5.2
7.6
5.9
0.12
0.10
0.15
0.05
0.84
0.86
0.80
0.78
Fig.
2
ORTEP representation of meso-dineopentyl 2,3-diethyl-
butanedioate (50% probability ellipsoids).
Fig. 1 GC chromatogram of the radical dimers formed by the
thermolysis of DTBPO in neopentyl butanoate (1) at 438 K.
Products arising from ester 1 at 438 K. The reaction of
DTBPO in ester 1 at 438 K gives six longer retention time prod-
ucts (labelled A1, A2, B1, B2, C1 and C2 in the chromatogram
in Fig. 1, Table 3) each of which by GC-MS has a molecular ion
with m/z = 314 (see Experimental for mass spectral details). The
similarity of the MS fragmentation patterns of A1 and A2, B1
and B2, and C1 and C2 indicates that these products are three
pairs of diastereoisomers arising from the dimerisation of sec-
ondary ester radicals. Column chromatography of the reaction
mixture led to the isolation and purification of A1 which was
Scheme 2 The product diastereoisomers from the self-dimerisation of
α-acyl neopentyl butanoate radicals.
dimers as were formed from the reaction using DTBPO.
Furthermore, the dimer distributions for both peroxides were
essentially identical (Fig. 3).
Products arising from the esters 2–4 at 438 K. As noted above
for ester 1, the other neopentyl esters also gave radical dimer
products from their reactions with DTBPO or DCPO and, for a
given ester, the dimer product distribution was independent of
the peroxide used (Table 3). Ester 2 gave four radical dimers
(labelled D1, D2, E and F in Fig. 4). GC-MS analysis showed
that D1 and D2 are a pair of diastereoisomers and E and F are
structural isomers. Thermolysis of the peroxides in esters 3 and
4 each gave three diastereoisomeric pairs of radical dimers
(Table 3).
1
identified by H and 13C NMR spectroscopy as a dineopentyl
2,3-diethylbutanedioate from the self-dimerisation of two α-
acyl radicals of ester 1 (Scheme 2).
The identities of A1 and A2 were confirmed by their syn-
thesis using the iodine promoted self-coupling of the α-anion of
ester 1.6 One of the diastereoisomers was obtained as crystals
which single crystal X-ray crystallography showed to be the
meso compound (Fig. 2). By elimination the other, which
remained as a liquid, is the racemic mixture of threo
enantiomers.
Effect of temperature on the radical dimer distributions from the
thermolyses and photolyses of DTBPO and DCPO in esters 1–4
GC analysis of the products from the thermolysis of DCPO
in ester 1 under nitrogen at 438 K showed the same six radical
A solution of DTBPO in ester 1 was photolysed at 298 K using
1528
J. Chem. Soc., Perkin Trans. 2, 2001, 1527–1533

View Article Online
Table 3 The relative peak areas of the ester radical dimers from the thermolysis of DTBPO and DCPO in esters 1–4, under nitrogen at 438 K
Ester
Peroxide
Peak 1
Peak 2
Peak 3
Peak 4
Peak 5
Peak 6
1
1
2
2
3
3
4
4
DTBPO
DCPO
DTBPO
DCPO
DTBPO
DCPO
DTBPO
DCPO
100
100
80
80
40
50
50
45
100
100
90
100
100
100
100
100
20
20
85
85
35
50
55
50
20
20
100
95
95
100
90
20
20
20
20
20
20
35
50
10
20
45
60
90
Fig. 5 The distribution of radical dimers from the photolysis of
DTBPO and DCPO in neopentyl butanoate (1) at 298 K.
Fig. 3 The distribution of radical dimers from the thermolysis of
DTBPO and DCPO in neopentyl butanoate (1) at 438 K.
products (esters 1, 3 and 4 gave 14, 9 and 15 GC peaks with
DTBPO and 17, 12 and 17 respectively with DCPO) and a
different product distribution to DTBPO (see for example
Fig. 5).
The reaction of DTBPO in ester 1 was also examined at the
following temperatures: between 298 and 438 K; 353, 373 and
393 K (photolysis) and 393 and 408 K (thermolysis). All the
radical dimers were detectable up to 408 K, although their
distributions changed with temperature (Table 4). One of the
most significant differences is in the proportions of the α-acyl
dimers A1 and A2. The reaction at 393 K was duplicated, using
thermolysis as the source of radicals in one experiment and
photolysis (with a small proportion of thermolysis) in the other.
The distribution of radical dimers from the two experiments
is almost identical (Table 4).
The thermolysis of neopentyl 2-bromobutanoate in neopentyl
butanoate (1) at 438 K
Solutions of neopentyl 2-bromobutanoate in ester 1 were
heated under nitrogen at 438 K. GC analysis showed that
the mixtures, which had changed from colourless to light
brown, contained the radical dimers of 1 (Fig. 6) and that the
distribution of dimer products was almost identical to those
from the thermolyses of DTBPO and DCPO in ester 1 at
438 K.
Fig. 4 GC chromatogram of the radical dimers formed by the
thermolysis of DTBPO in neopentyl 2,2-dimethylpropanoate (2) at
438 K.
a xenon arc light source (λ > 300 nm). The peaks from at least
fourteen radical dimers were observed by GC and GC-MS,
including the six observed in the reaction at 438 K (Fig. 5); GC-
MS analysis showed that most of these are present as pairs of
diastereoisomers. Significantly more of A2 than A1 appears to
be formed at 298 K. For esters 3 and 4 the number of radical
dimers also increased, from 6 to 9 and from 6 to 15 respectively,
when the reactions were carried out at 298 K. In contrast, for
ester 2 the number of radical dimers formed at 298 K was the
same as that obtained at 438 K.
Similar results were generally obtained with DCPO to those
from DTBPO. In both systems, the number of radical dimers
formed from ester 2 was independent of reaction temperature
whereas esters 1, 3 and 4 yielded many more radical dimers at
298 K compared to their reactions at 438 K. However, there are
clear distinctions between the reactions of the two peroxides
with the latter group of esters, at 298 K. DCPO gives more
Discussion
The peroxides used in this study have been thermolysed and
photolysed to give alkoxyl radicals [reaction (1)] and these in
turn react further in two competing processes: fragmentation
[reaction (2)] and hydrogen atom abstraction [reaction (3)].
Reaction (2) is more significant with DCPO than DTBPO7 and
for both peroxides it becomes increasingly important as the
temperature is increased.1
The reactions of the methyl radicals
The methyl radicals abstract hydrogen atoms from the ester to
give methane [reaction (4)] and are involved in a number of
other radical–radical reactions [reactions (5)–(7)]. From the
J. Chem. Soc., Perkin Trans. 2, 2001, 1527–1533
1529

View Article Online
Table 4 The temperature dependence of the relative yields of the ester
radical dimers from the thermolysis and the photolysis of DTBPO
under nitrogen, in neopentyl butanoate 1
GC-MS, it was not possible to separate them from the ester
substrates, making it impossible to quantify them in this
study.
Temperature/K
Ester radical dimerisation at 438 K
Dimera
298b
353b
373b
393b
393c
408c
438c
The ester radicals are consumed in disproportionation and
dimerisation processes [reactions (8) and (9)].
1 (A1)
2 (X)
3 (A2)d
4
5
6
7 (B1)
8 (B2)
9
100
83
167
10
10
50
25
20
18
22
23
23
10
10
100
53
161
19
23
31
50
53
29
16
31
29
19
29
100
48
167
20
17
22
38
43
28
20
35
30
12
12
100
27
137
11
5
11
38
34
12
7
100
24
125
9
6
9
24
29
9
100
10
112
4
100
100
ؒ
ؒ
R ϩ R → RH ϩ R(ϪH)
(8)
(9)
4
11
22
25
8
1
12
9
ؒ
ؒ
R ϩ R → R–R
20
20
Product analyses of the thermolyses of the peroxides in ester
1 at 438 K reveal three key results: (i) only 6 of a possible 21
ester radical dimers (excluding enantiomers) are detected; (ii)
the six dimers that are observed are formed as three pairs of
diastereoisomers; (iii) the reactions with two peroxides give
essentially identical dimer distributions. These points are
discussed in detail below.
Diastereoisomers can only be formed from ester 1 by
combinations of the three secondary radicals (α- and β-acyl
and α-alkyl). The preferential hydrogen abstraction by the
alkoxyl and methyl radicals from secondary C–H bonds can
be accounted for in terms of the relative bond dissociation
energies of primary and secondary C–H bonds.9 However, the
observed selectivity of the radical dimerisation was unexpected
since only half of the possible 6 diastereoisomer pairs was
detected. The major products, A1 and A2 (∼80%) have been
identified as α-acyl–α-acyl dimers, meso- and ( )-threo-
dineopentyl 2,3-diethylbutanedioate (Scheme 2), indicating
that the α-acyl radical is the major ester radical in the reaction
system. It follows that B1 and B2, and C1 and C2 are formed by
cross-dimerisations of the α-acyl radical with the β-acyl and the
α-alkyl radicals.
10
7
11 (C1)
12 (C2)
13
18
11
5
19
11
8
19
18
2
3
14
4
3
a Dimer peaks numbered in order of elution from GC column.
b Peroxide photolysed. c Peroxide thermolysed. d This GC peak is
believed to contain the diastereoisomer of dimer X.
The reactions of the other neopentyl esters, 2–4, with
DTBPO and DCPO at 438 K show similar trends to those of
ester 1. Thus, the dimerisations are relatively selective and for a
given ester both peroxides give the same dimer product distribu-
tion. Ester 2 has only one secondary carbon and, as expected,
gives one pair of diastereoisomers (D1 and D2) from the self-
reaction of the α-alkyl radical. The other two dimers detected
(E and F) are not diastereoisomers and must arise from the
cross-dimerisation of this radical with the two primary ester
radicals (Scheme 3). Esters 3 and 4, each with two secondary
carbon atoms, give the expected three diastereoisomeric pairs
of radical dimers.
If tert-butoxyl and cumyloxyl were the only radicals involved
in abstracting hydrogen atoms from the esters then, even
though the alkoxyl radicals are different, it would perhaps not
be surprising that the radical dimer distributions from the
reactions are so similar. However, it is clear that methyl radicals
play a significant role in these reactions and, from the methane
yields, that virtually all of them react by hydrogen atom
abstraction from the ester. Indeed, there are several reports in
the literature of methyl radicals abstracting the α-acyl hydrogen
atoms from methyl esters to give the corresponding α-acyl–α-
acyl dimers.10 With larger alkyl groups, hydrogen abstraction
from the alkyl portion is also observed.11 Furthermore, methyl
and alkoxyl radicals, on electronic grounds, would not be
expected to have the same selectivity for hydrogen atom abstrac-
tion: alkoxyls are electrophilic whereas methyl is nucleophilic.12
Thus, since the two peroxides produce distinctly different ratios
of alkoxyl and methyl radicals, with DCPO giving a much
higher proportion of methyl radicals, the reactions of the
two peroxides would be expected to give different dimer
distributions.
Fig. 6 The distribution of radical dimers from the thermolysis of
neopentyl 2-bromobutanoate in neopentyl butanoate (1) at 438 K.
ؒ
RЈ(CH3)2CO–OC(CH3)2RЈ → 2RЈ(CH3)2CO
(1)
(2)
(3)
(4)
(5)
(6)
(7)
ؒ
ؒ
RЈ(CH3)2CO → RЈCOCH3 ϩ CH3
ؒ
ؒ
RЈ(CH3)2CO ϩ RH → RЈ(CH3)2COH ϩ R
ؒ
ؒ
CH3 ϩ RH → CH4 ϩ R
ؒ
ؒ
CH3 ϩ R → CH4 ϩ R(ϪH)
ؒ
ؒ
CH3 ϩ R → CH3R
ؒ
ؒ
CH3 ϩ CH3 → CH3CH3
ؒ
ؒ
[RH = ester and R = (ester Ϫ H) ]
methane : ketone ratios in Tables 1 and 2, it is clear that 80–90%
of the methyl radicals are consumed in reactions (4) and (5)
rather than in reactions (6) and (7). Since the rate constants for
the radical–radical reactions are very large compared with the
expected value for hydrogen abstraction by methyl,8 this indi-
cates that the concentration of the radicals in the reactions
ؒ
must be very low and k4[ester] ӷ k5[(ester Ϫ H) ], k6[(ester Ϫ
ؒ
ؒ
H) ] and k7[Me ]. The high methane to ethane ratios confirm
this conclusion. It is likely that the methyl radicals unaccounted
for as either methane or ethane are involved in reactions (5)
and (6) to give methylated and unsaturated neopentyl esters.
However, although some of these products could be detected by
The above conclusions assume that when the abstracting
radicals remove a hydrogen atom, the resulting ester radicals
1530
J. Chem. Soc., Perkin Trans. 2, 2001, 1527–1533

View Article Online
react either by dimerisation or by disproportionation and that
the position of the initial hydrogen abstraction is locked in the
dimer products. Thus, the carbon atoms involved in the new
C–C bond correspond to the sites of hydrogen abstraction.
However, we argue that once an ester radical is produced it
should also be able to abstract a hydrogen atom from the ester,
which is present in a very large excess as both substrate and
solvent. Assuming this happens several times before two rad-
icals dimerise, the products would not reveal the identities of
the initial sites of hydrogen abstraction but rather the thermo-
dynamic equilibrium distribution of ester radicals (Scheme 4).
Consequently, it is not surprising that for a given ester both
peroxides give essentially the same dimer product distribution.
To explore this hypothesis in more detail, the temperature of
the reaction was lowered from 438 to 298 K.
than increase the number of products obtained. We believe that
at 298 K kinetic effects are important and that the ratio of
intermolecular radical scrambling to dimerisation and dis-
proportionation has decreased [reactions (8) and (9)]. Thus,
the product analysis of reactions at 298 K should give more
information about the initial sites of hydrogen abstraction,
although it is still likely that ester radical hydrogen abstraction
is occurring.
Neopentyl esters 3 and 4 behave like ester 1 in their reaction
with DTBPO giving a greater number of products at 298 than
at 438 K. In contrast, however, the number of radical dimers
from ester 2 is independent of temperature. This, we believe, is
associated with their structures: esters 1, 3 and 4 have two or
more secondary carbon atoms whereas ester 2 has only one
(α-alkyl). It is likely that for the latter substrate, the ester radical
that predominates is the same whether the reaction is favoured
by kinetic or by thermodynamic factors.
Ester radical dimerisation at 298 K
The results above support those from our previous studies on
the autoxidation of neopentyl esters 1–4 where all the primary
and secondary C–H bonds were found to be susceptible to
oxidation.1,5 However, on closer examination, the dominance of
dimer products from the α-acyl radical of ester 1 at all temper-
atures in this study seems contrary to the selectivity observed in
the autoxidation of this ester where the β-acyl and α-alkyl C–H
bonds were found to be more reactive. This suggests that, under
autoxidative conditions, the ester radicals are rapidly trapped
by oxygen before they have an opportunity to undergo the
intermolecular radical scrambling.
Lowering the reaction temperature to 298 K resulted in a sig-
nificant increase in the number of radical dimers from the reac-
tion of ester 1 with DTBPO. This supports the suggestion that
the reaction at 438 K is thermodynamically rather than kinetic-
ally controlled, since lowering the temperature of a reaction
that is kinetically controlled would be expected to reduce rather
The reaction of ester 1 with DCPO at 298 K gives more
radical dimer products and in a different distribution to that
obtained with DTBPO. The main differences in the two product
distributions can be attributed to the methyl radical. At 298 K,
fragmentation of tert-butoxyl [reaction (2)] is unimportant and
the alkoxyl radical is the only abstracting radical with the latter
peroxide, whereas with DCPO both cumyloxyl and methyl are
involved.
The effect of temperature on the reaction of ester 1 with
DTBPO was explored further between 298 and 438 K. This
reveals that all the products observed at 298 K were detectable
up to 408 K suggesting that thermodynamic control of the
products may only be complete at 438 K. This study also
confirmed that the method of generating the alkoxyl radicals
(thermolysis versus photolysis) has no effect on the product
Scheme 3 Cross-dimerisation of α-alkyl and primary radicals of
neopentyl 2,2-dimethylpropanoate.
Scheme 4 Intermolecular rearrangement of ester radicals from neopentyl butanoate.
J. Chem. Soc., Perkin Trans. 2, 2001, 1527–1533
1531

View Article Online
ratios. Thus, at 393 K the dimer distribution from ester 1
is similar whether the reaction is initiated thermally or
photochemically.
(inside volume 4.87 cm3) which has been described in previous
papers.1,5 The autoclave, containing a 0.5 mol dmϪ3 solution of
the peroxide in the ester (1 cm3), was evacuated and flushed with
nitrogen several times before finally being repressurised to 1 bar
with nitrogen and placed in an aluminium block which was
heated to the reaction temperature. After the required reaction
time, the autoclave was cooled with ice, connected to a vacuum
line and the gas from the reaction was shared into a sample
loop for GC analysis. The autoclave was then opened and the
liquid products also analysed by GC.
For the isolation of ester dimer A1, a larger scale reaction
was carried out. A solution of DTBPO (0.1 mol dmϪ3) in neo-
pentyl butanoate (10 cm3) was thoroughly purged with argon
and then heated at 438 K under argon. After 2 h the solution
was cooled and the volatile products were removed under
vacuum. The residue which contained the radical dimers A1,
A2, B1, B2, C1 and C2 was purified by flash chromatography
(silica gel with dichloromethane) to give A1; δH (CDCl3) 0.89
(6H, t), 0.96 (18H, s), 1.60 (4H, m), 2.62 (2H, m), 3.78 (4H, s);
δC (CDCl3) 11.7 (–CH2CH3), 24.3 (–CH2CH3), 26.6 [(CH3)3C],
31.2 [(CH3)3C], 50.0 (–O2CH2), 74.0 (–CH2O2C), 174.5
(ϪO2CCH2); MS(CI) m/z 315 (MHϩ).
The photolyses were carried out in a cylindrical stainless steel
cell (id 10 mm, length 20 mm) fitted at each end with Pyrex
optical windows.16 Three stainless steel side-arms were con-
nected to the cell, one was sealed with a Viton septum and the
others were connected via glass-to-metal joints to greaseless
‘Rotaflo’ taps. The temperature of the cell was controlled by
placing it in a thermostatted aluminium jacket, heated by a
Jencons (230 V, 320 W) heating band. The heating was moni-
tored by a thermocouple connected to an electroserve tem-
perature control unit. The light source used was a 300 W, high
intensity xenon arc lamp fitted with a light guide (Laser Lines
Ltd.). The reactions used 0.5 mol dmϪ3 solutions of the per-
oxide in the ester (5 cm3) and were thoroughly deoxygenated
with nitrogen before photolysis.
The three pairs of diastereoisomers at 438 K, detected from
ester 1, are major products at all the temperatures investigated.
However, whereas the ratios of B1 : B2 and C1 : C2 are
approximately 1.0 at all the temperatures, the A1 : A2 ratio
decreases from 1.0 at 438 K to 0.6 at 298 K. This could arise
from a change in the diastereoselectivity in the dimerisation
process, however, we believe it is due to the co-elution of A2
with the diastereoisomer of the radical dimer which elutes
between A1 and A2 (X in Fig. 5). In agreement with this conclu-
sion the yield of A2 at all temperatures studied is approximately
equal to the sum of the yields of A1 and X.
Thermolysis of neopentyl 2-bromobutanoate in neopentyl
butanoate
To confirm the existence of the intermolecular rearrangement
of the ester radicals we examined the fate of a specific ester
radical under the reaction conditions. For this purpose we
chose the thermolysis of neopentyl 2-bromobutanoate in neo-
pentyl butanoate at 438 K. This gave not only the two α-acyl–α-
acyl dimers, A1 and A2, expected from the self-combination
of the α-acyl radicals but also dimers B1, B2, C1 and C2.
Furthermore, in agreement with the thermodynamic control
of the intermolecular rearrangement at 438 K, the distribution
of dimers is the effectively same as that obtained from the
reactions with DTBPO and DCPO at this temperature.
Conclusions
(1) Ester radicals, generated from the thermolysis or photolysis
of peroxides in esters, rearrange by hydrogen atom abstraction
from the ester. (2) As a consequence of this radical scrambling,
the ester radical dimer product distributions do not provide
reliable data to calculate the selectivity of hydrogen abstraction
by alkoxyl and methyl radicals. (3) The high temperature stud-
ies with ester 1, where the radical equilibration is complete,
show that the most stable radical is α to the carbonyl group.
Materials
All materials were commercially available and used as
purchased unless otherwise stated. Tetrahydrofuran (THF)
was dried over sodium under a nitrogen atmosphere using
benzophenone as an indicator.
Experimental
Instrumentation
The neopentyl ester substrates were prepared as reported
previously.5 Dineopentyl 2,3-diethylbutanedioate was prepared
by adding neopentyl butanoate (0.01 mol) in dry THF (2 cm3)
to a stirred solution of lithium diisopropylamide (0.01 mol) in
dry THF (10 cm3) under nitrogen at 195 K. After 10 min, iodine
(0.005 mol) in dry THF was added and the mixture was allowed
to warm to room temperature. After a further 20 min stirring
the reaction was quenched with an aqueous solution of ammo-
nium chloride and extracted with light petroleum (40–60 ЊC).
The organic solution was washed with an aqueous solution of
sodium chloride, dried, concentrated under vacuum and puri-
fied using column chromatography (silica gel with dichloro-
methane as eluant) to give a 19.5% yield of a mixture of meso-
and ( )-threo-dineopentyl 2,3-diethylbutanedioate; δH (CDCl3)
0.89 (6H, t), 0.96 (18H, s), 1.60 (4H, m), 2.62 (2H, m), 3.82
(4H, s); δC (CDCl3) 11.7 (–CH2CH3), 24.3 (–CH2CH3), 26.4
[(CH3)3C], 31.2 [(CH3)3C], 49.9 (–O2CCH2), 73.9 (–CH2O2C),
174.4 (–O2CCH2); MS(EI) m/z 299 (25%), 227 (80), 157 (88),
GC analyses were carried out on a Pye Unicam GCD gas
chromatograph fitted with a flame ionisation detector. The gas
samples were separated with a glass column packed with silica
gel (80–120 mesh) and for the liquid samples a Carbowax 20M
capillary column (30 m length, 0.25 mm id, 0.25 µm film
thickness) was employed. Data collection used a Trio Trivector
integrator. GC mass spectra were obtained with a VG Autospec
S Series A027 mass spectrometer linked to a Hewlett Packard
5890 Series 2 gas chromatograph. The spectra were analysed
using a VAX3100 Workstation.
1H and 13C NMR spectra were recorded using JEOL 270 and
Bruker MSL 300 spectrometers with tetramethylsilane as the
internal standard.
The X-ray crystal structure was obtained using an automated
Rigaku AFC6S diffractometer, ω–2θ scan mode, with graphite-
monochromated Mo-Kα radiation, µ(Mo-Kα) = 0.07 mmϪ1,
ω scan width = 1.21 ϩ tan θ, ω scan speed 1Њ minϪ1. The
unit cell was indexed by least mean squares refinement on
diffractometer angles for 20 automatically centred reflections,
λ = 0.71069 Å. The structure was solved by direct methods
with SHELXS8613 and expanded using Fourier techniques with
DIRDIF.14 Full matrix least squares refinements on F 2
with SHELXL9715 with all non-hydrogen atoms anisotropic
and hydrogens refined using a rigid model.
ϩ
71 (100), 55 (27), 43 (57); MS(CI) m/z 332 (MNH4 ), 315
(MHϩ).
When the purified mixture of diastereoisomers was left to
stand, crystals of the meso compound were obtained. Crystal
data§ for meso-dineopentyl 2,3-diethylbutanedioate C18H34O4,
§ CCDC reference number 162741. See http://www.rsc.org/suppdata/
p2/b1/b103555j/ for crystallographic files in .cif or other electronic
format.
Apparatus and methods
The thermolyses were carried out in a stainless steel autoclave
1532
J. Chem. Soc., Perkin Trans. 2, 2001, 1527–1533

View Article Online
M = 314.46, monoclinic, a = 8.359(5), b = 13.505(4), c =
9.538(4) Å, β = 111.74(3)Њ, V = 1000.1(8) Å3, T = 150 K, space
group P21/a, Z = 2, µ(Mo-Kα) = 0.072 mmϪ1, 1841 reflections
measured, 1750 unique (Rint = 0.012). Final R1, wR2 on all data
(1750 reflections) 0.0725, 0.1209. R1, wR2 on [Io > 2σ(Io)] (1216
reflections) 0.0400, 0.1047.
55 (5), 43 (25), 41 (9.0); dimer 6: 355 (Mϩ Ϫ Me, 0.3%), 215 (5),
99 (45), 72 (7), 71 (100), 57 (11), 55 (7.0), 43 (33), 41 (12).
Acknowledgements
We thank Showa Shell Sekiyu KK, Shell Global Solutions and
the University of York for their support, Dr S. J. Archibald
for solving the crystal structure of meso-dineopentyl 2,3-
diethylbutanedioate and Dr H. Gillespie, Dr S. Bévière and
Dr S. Lawrence for very helpful discussions.
2-Bromobutanoic acid was prepared by heating a mixture of
butanoic acid (20 mmol), bromine (20 mmol), chlorosulfonic
acid (0.5 cm3) in dichloroethane (50 cm3) at 85 ЊC. After 2 h the
unreacted bromine and dichloroethane were removed under
vacuum and the residue was esterified with neopentanol and
a small amount of concentrated sulfuric acid in toluene.
The water was removed azeotropically with a Dean–Stark
apparatus. After the reaction, the toluene was washed with an
aqueous solution of sodium hydrogen carbonate, dried
(MgSO4), concentrated under vacuum and purified by flash
chromatography on silica gel, using dichloromethane as the
eluant, to give neopentyl 2-bromobutanoate in 55% yield;
δH(CDCl3) 1.15 (9H, s), 1.26 (3H, t), 2.35 (2H, m), 3.75 (2H,
s), 4.28 (1H, t); δC (CDCl3) 11.83 (–CH2CH3), 26.25
[(CH3)3C–], 28.35 (–CH2CH3), 31.53 [(CH3)3C–], 47.87
(–CHBrCH2CH3), 74.85 (–CH2O2C), 169.69 (–O2CCHBr);
MS(EI), m/z 223 (5%), 221 (5), 183 (85), 181 (85), 151 (73),
149 (73), 123 (75), 121 (75), 62 (100), 57 (98); MS(CI) m/z
References
1 Part 2. J. R. Lindsay Smith, E. Nagatomi and D. J. Waddington,
J. Chem. Soc., Perkin Trans. 2, 2000, 2248.
2 For example, M. Brown, J. D. Fotherington, T. J. Hoyes, R. M.
Mortier, S. T. Orszulik, S. J. Randles and P. M. Stroud, in Chemistry
and Technology of Lubricants, ed. R. M. Mortier and S. T. Orszulik,
Blackie, London, 1997, 2nd edn.
3 T. E. Karis, J. L. Miller, H. E. Hunziker, M. S. de Vries, D. A.
Hopper and H. S. Nagaraj, Tribol. Trans., 1999, 42, 431.
4 R. K. Jensen, S. Korcek, L. R. Mahoney and M. Zinbo, J. Am.
Chem. Soc., 1979, 101, 7574; E. J. Hamilton, S. Korcek, L. R.
Mahoney and M. Zinbo, Int. J. Chem. Kinet., 1980, 12, 577; R. K.
Jensen, S. Korcek and M. Zinbo, J. Synth. Lubr., 1984, 1, 91; V. N.
Bakunin and O. P. Parenago, J. Synth. Lubr., 1992, 9, 127; R. K.
Jensen, S. Korcek and M. Zinbo, J. Am. Chem. Soc., 1992, 114, 7742;
R. K. Jensen, S. Korcek and M. Zinbo, Int. J. Chem. Kinet., 1994,
26, 673; H. S. Aldrich, J. M. Cayce, K. E. Edwards, L. Gschwender,
K. U. Ingold, R. L. Paulson and C. Snyder, presentation at the
Society of Tribologists and Lubrication Engineers Conference,
Las Vegas, 1999; J. R. Lindsay Smith, D. J. Waddington, M. S. Stark,
P. W. Smith and E. D. Pritchard, presentation at Additives 2001,
Oxford, 2001.
5 J. R. Lindsay Smith, E. Nagatomi, A. Stead, D. J. Waddington and
S. D. Bévière, J. Chem. Soc., Perkin Trans. 2, 2000, 1193.
6 T. G. Brocksom, N. Petragnani, R. Rodrigues and H. La Scala
Teixeira, Synthesis, 1975, 396; J. Briggs, A. H. Haines, R. J. K.
Taylor, A. P. Dawson, I. Gibson, J. Hook, A. Lloyd and S. Meiners,
Carbohydr. Res., 1992, 234, 23.
ϩ
256, 254 (MNH4 ).
Mass spectral data for dimers from reactions of esters at 438 K
Ester 1 gave six dimers: MS(EI) m/z (relative intensities) A1: 314
(Mϩ, 0.03%), 227 (6), 157 (81), 129 (12), 88 (21), 87 (20), 71
(100), 57 (17), 55 (28), 43 (57), 41 (22); A2: 314 (Mϩ, 0.03%),
227 (7), 157 (85), 129 (10), 88 (12), 87 (10), 71 (100), 57 (17), 55
(29), 43 (65), 41 (23); B1: 314 (Mϩ, 1.0%), 227 (11), 157 (100),
129 (17), 128 (16), 88 (9), 87 (12), 71 (60), 57 (16), 55 (23), 43
(46), 41 (23); B2: 314 (Mϩ, 1%), 227 (12), 157 (100), 129 (22),
128 (16), 88 (11), 87 (12), 71 (63), 57 (17), 55 (22), 43 (45), 41
(22); C1: 314 (Mϩ, 3.5%), 227 (36), 198 (12), 157 (65), 129 (44),
111 (20), 71 (100), 57 (30), 55 (33), 43 (67), 41 (35); C2: 314 (Mϩ,
3.5%), 227 (36), 198 (12), 157 (65), 129 (45), 111 (20), 71 (100),
57 (30), 55 (33), 43 (68), 41 (35).
7 K. Y. Choo and S. W. Benson, Int. J. Chem. Kinet., 1981, 13, 833;
L. Batt and G. N. Robinson, Int. J. Chem. Kinet., 1982, 14, 1053;
A. Baignée, J. A. Howard, J. C. Scaiano and L. C. Stewart, J. Am.
Chem. Soc., 1983, 105, 6120.
8 Rate constants for abstraction of hydrogen atoms in hydrocarbons
and analogues by alkyl radicals in the liquid phase are scarce, but
sensible estimates can be made from the reliable data that are
available, e.g. K. Munger and H. Fischer, Int. J. Chem. Kinet., 1984,
16, 1213.
9 J. F. Seetula and I. R. Slagle, J. Chem. Soc., Faraday Trans., 1997, 93,
1709.
10 M. S. Kharasch, E. V. Jensen and W. H. Urry, J. Org. Chem., 1945,
10, 386; M. S. Kharasch, H. C. McBay and W. H. Urry, J. Org.
Chem., 1945, 10, 394; H. C. McBay, O. Tucker and A. Milligan,
J. Org. Chem., 1954, 19, 1003; S. A. Harrison, L. E. Peterson and
D. H. Wheeler, J. Chem. Soc., 1955, 349; G. A. Razavuov and L. S.
Boguslovskaya, Zh. Obshch. Khim., 1961, 31, 3440; A. Razavuov
and L. S. Boguslovskaya, Bul. Inst. Politeh. Iasi, 1962, 141.
11 M. Lazar, J. Pavlinec and Z. Movasek, Collect. Czech. Chem.
Commun., 1961, 26, 1380.
12 R. L. Huang, S. H. Goh and S. H. Ong, The Chemistry of Free
Radicals, Edward Arnold, London, 1974; J. M. Tedder and J. C.
Walton, Tetrahedron, 1982, 38, 313; J. M. Tedder, Angew. Chem.,
Int. Ed. Engl., 1982, 21, 401; B. Giese, Angew. Chem., Int. Ed. Engl.,
1983, 22, 753; M. J. Perkins, Radical Chemistry, Horwood,
New York, 1994; J. Fossey, D. Lefort and J. Sorba, Free Radicals in
Organic Chemistry, Wiley, Chichester, 1995; T. Zytowski and
H. Fischer, J. Am. Chem. Soc., 1997, 119, 12869.
Ester 2 gave four dimers: MS(EI) m/z (relative intensities) D1:
327 (Mϩ Ϫ Me, 1.0%), 285 (5), 201 (6), 199 (4), 184 (4), 171 (4),
85 (54), 57 (100), 41 (14); D2: 327 (Mϩ Ϫ Me, 1%), 285 (5), 201
(6), 199 (4), 184 (4), 171 (4), 85 (53), 57 (100), 41 (14); E: 327
(Mϩ Ϫ Me, 0.5%), 139 (9), 125 (6), 99 (52), 85 (30), 83 (19), 69
(18), 57 (100), 41 (17); F: 327 (Mϩ Ϫ Me, 2%), 285 (7), 171 (43),
125 (19), 113 (47), 85 (30), 71 (15), 70 (12), 69 (12), 57 (100), 43
(14), 41 (21).
Ester 3 gave six dimers: MS(EI) m/z (relative intensities), in
order of GC elution dimer 1: 355 (Mϩ Ϫ Me, 0.03%), 185 (11),
129 (9), 99 (100), 71 (14), 57 (46), 43 (12); dimer 2: 355
(Mϩ Ϫ Me, 0.03%), 185 (11), 129 (9), 99 (100), 71 (14), 57
(47), 43 (12); dimer 3: 355 (Mϩ Ϫ Me, 0.04%), 167 (18), 129
(10), 99 (100), 71 (17), 57 (50), 43 (12); dimer 4: 355
(Mϩ Ϫ Me, 0.04%), 167 (18), 129 (10), 99 (100), 71 (16), 57
(49), 43 (12); dimer 5: 355 (Mϩ Ϫ Me, 0.2%), 313 (5), 100 (8),
99 (100), 83 (7), 71 (9), 57 (47), 43 (8.0), 41 (6); dimer 6: 355
(Mϩ Ϫ Me, 0.3%), 313 (3), 100 (8), 99 (100), 83 (7), 71 (9), 57
(46), 43 (7), 41 (7).
Ester 4 gave six dimers: MS(EI) m/z (relative intensities), in
order of GC elution dimer 1: 355 (Mϩ Ϫ Me, 1.6%), 185 (33),
139 (14), 127 (19), 99 (34), 83 (22), 71 (100), 69 (30), 57 (27), 43
(48), 41 (23); dimer 2: 355 (Mϩ Ϫ Me, 1.6%), 185 (33), 139 (14),
127 (19), 99 (3), 83 (23), 71 (100), 69 (30), 57 (27), 43 (48), 41
(23); dimer 3: 355 (Mϩ Ϫ Me, 0.7%), 185 (39), 145 (10), 127
(32), 99 (22), 83 (16), 71 (100), 57 (18), 55 (13), 43 (38), 41 (17);
dimer 4: 355 (Mϩ Ϫ Me, 0.9%), 185 (35), 145 (10), 127 (32), 99
(20), 83 (17), 71 (100), 57 (18), 55 (12), 43 (38), 41 (17); dimer 5:
355 (Mϩ Ϫ Me, 0.2%), 215 (7), 99 (30), 72 (6), 71 (100), 57 (8),
13 G. M. Sheldrick, in Crystallographic Computing 3, ed. G. M.
Sheldrick and C. Kruger, Oxford University Press, 1985, 175.
14 P. T. Beurskens, G. Admiral, G. Beurskens, G. Bosman, W. P. Garcia-
Granda, R. O. Gould, J. M. M. Smits and C. Smykalla, The
DIRDIF Programme System, Technical report of the
crystallography laboratory, University of Nijmegen, 1992.
15 G. M. Sheldrick, SHELXL97, Programme for crystal structure
refinement, University of Göttingen, 1997.
16 A. R. Costello, J. R. Lindsay Smith, M. S. Stark and D. J.
Waddington, J. Chem. Soc., Faraday Trans., 1996, 92, 3497.
J. Chem. Soc., Perkin Trans. 2, 2001, 1527–1533
1533