Peroxyesters as precursors to peroxyl radical clocks

Article abstract of DOI:10.1021/jo201862s

The reactions of peroxyl radicals are at the center of the oxidative degradation of essentially all petroleum-derived hydrocarbons and biological lipids and consequently, the inhibition of these processes by radical-trapping antioxidants. Recently describ

Full text of DOI:10.1021/jo201862s

Article
pubs.acs.org/joc
Peroxyesters As Precursors to Peroxyl Radical Clocks
Jason J. Hanthorn† and Derek A. Pratt*,†,‡
†Department of Chemistry, Queen’s University, 90 Bader Lane, Kingston, K7L 3N6 Canada
‡Department of Chemistry, University of Ottawa, 10 Marie Curie, Ottawa, K1N 6N5 Canada
S
* Supporting Information
ABSTRACT: The reactions of peroxyl radicals are at the
center of the oxidative degradation of essentially all petroleum-
derived hydrocarbons and biological lipids and consequently,
the inhibition of these processes by radical-trapping anti-
oxidants. Recently described peroxyl radical clocks offer a
simple, convenient, and inexpensive method of determining
rate constants for H-atom transfer reactions to peroxyl radicals,
greatly enabling the kinetic and mechanistic characterization of
compounds with antioxidant properties. We follow up our
preliminary communication on the development of a method-
ology utilizing tert-butyl styrylperacetate as a precursor to a
versatile peroxyl radical clock with the present paper, wherein we describe a novel naphthyl analogue, which provides for much
improved product resolution for analysis, and provide the complete details associated with its development and application.
Using this new precursor, and with consideration of the expanded set of reaction products, inhibition rate constants were
measured for a variety of representative phenolic and diarylamine radical-trapping antioxidants. We also provide details for the
use of this methodology for the determination of mechanistic information, such as kinetic solvent effects, Arrhenius parameters,
and kinetic isotope effects.
A−H. This methodology requires no specialized equipment other
than a GC or HPLC for product analysis. Although the approach
is effective for measuring rates of H-atom transfer from many
phenols to peroxyl radicals in benzene, a major limitation of this
method is that it requires the antioxidant-derived radical (A·) to
abstract a hydrogen atom from allylbenzene to propagate the
chain reaction. This becomes problematic if the antioxidant gives
rise to either persistent (e.g., BHT) or highly stabilized (e.g., 6-
amino-3-pyridinols)17 radicals that cannot efficiently propagate
the chain. To offset low propagation rate constants, even for the
most ideal A−H, a high concentration of oxidizable substrate
(e.g., 2.3 M allylbenzene) must be used to ensure enough
oxidation products are formed for reproducible analysis. A
consequence of using such high concentrations of oxidizable
substrate is that it precludes accurate kinetic or mechanistic
studies in any solvent other than benzene.
To address the limitations associated with this approach, we
recently described the use of the homoconjugated peroxyester
29 as a precursor to the same delocalized allylbenzene-derived
radical essential for the kinetic competition experiment. In our pre-
liminary report, we showed that compound 2 decomposes under
aerobic conditions either “thermally” (37 °C) or photolytically
(300 nm) to generate nonconjugated peroxyl radical 1, permitting
clock experiments using as little as 10 mM of compound 2.
Furthermore, the rate constant for β-fragmentation (kβ) of 1
was measured in a variety of organic solvents, which subsequently
INTRODUCTION
■
Peroxyl radicals are the key intermediates in the radical-
mediated oxidative degradation (autoxidation) of all organic
materials, including primary petroleum products, polymers,
foodstuffs, and biological molecules.1−3 To slow the rate of
autoxidation, radical-trapping antioxidants (e.g., phenols and
aromatic amines) are employed industrially (e.g., BHT and
4,4′-dialkyldiphenylamines) and in nature (e.g., α-tocopherol,
the most potent form of vitamin E) to “trap” peroxyl radicals as
hydroperoxides to prevent radical-chain propagation.4−6 These
compounds donate a hydrogen atom to a chain-carrying
peroxyl radical, forming a hydroperoxide and a more stable/
persistent (i.e., less reactive) antioxidant-derived radical at a
rate dependent on antioxidant structure and reaction medium.7
Understanding the kinetics of H-atom transfer from antiox-
idants to peroxyl radicals provides important mechanistic
information about structure−activity relationships relevant to
radical-trapping antioxidants. Several methods have been
developed for measuring inhibition rate constants, including
inhibited autoxidation, laser flash photolysis and pulse
radiolysis. However, these methods all suffer the drawback of
requiring specialized equipment and expertise to obtain reliable
kinetic data, and they are often limited in the time domain and
solvents in which they can be carried out.
Recently, Porter and co-workers8 introduced a peroxyl radi-
cal clock approach based on the kinetic competition between
the unimolecular β-fragmentation of a nonconjugated peroxyl
radical (e.g., 1 derived from allylbenzene as in Scheme 1)
and its trapping in a bimolecular reaction with an antioxidant
Received: September 12, 2011
Published: November 3, 2011
© 2011 American Chemical Society
276
dx.doi.org/10.1021/jo201862s|J. Org. Chem. 2012, 77, 276−284

The Journal of Organic Chemistry
Article
Scheme 1
allowed for the measurement of kinetic solvent effects on the rates
of H-atom transfer from phenolic antioxidants.
material of higher purity that gave rise to a long shelf life at
ambient temperatures.
Although the peroxyester approach was successful in obtaining
both inhibition rate constants (kinh in Scheme 1) and the effect of
solvent upon them for several phenolic antioxidants, one of the
aspects we sought to improve prior to any further development
or application of the approach was the resolution of the product
alcohols (the hydroperoxides are reduced with PPh3 for analysis
by gas chromatography with flame ionization detection), which
often had poorly resolved peaks and/or peaks with similar
retention times to common phenolic antioxidants and their
derivatives. Herein, we describe a second generation peroxyl
radical clock, derived from the naphthyl-based peroxyester 3,
which yields greatly improved chromatographic resolution and
revealed previously unidentified products in the region of the
conjugated and nonconjugated alcohols. The origin of these
products and their contribution to the accurate measurement of
inhibition rate constants using this approach was explored, as
well as the role played by the antioxidants in their formation. To
demonstrate the utility of the peroxyester approach beyond
simple phenolic compounds, we use compound 3 to measure
inhibition rate constants for some highly reactive pyrimidinols,
pyridinols, and diarylamines. Furthermore, we use 3 to measure
deuterium kinetic isotope effects (DKIEs) for a variety of
phenolic antioxidants and are able to provide Arrhenius
parameters for the β-fragmentation of the nonconjugated peroxyl
radical derived from 3, such that it may be applied in the future
to study the temperature dependence of inhibition reactions.
The corresponding nonconjugated (4) and conjugated (6)
alcohol products used as GC standards for kinetic analysis were
also synthesized from 2-naphthaldehyde. Compound 4 was
prepared via a Grignard reaction with vinylmagnesium bromide,
and compound 6 via a Horner−Wadsworth−Emmons reaction
followed by DIBAL reduction of the resulting ester.
Although the conjugated and nonconjugated alcohols were
the only products observed in the use of peroxyester 2,9 we
observed additional peaks in the same region as the
corresponding alcohols 4 and 6 in our calibration experiments
with peroxyester 3 with α-tocopherol (vide infra). GC−MS
studies indicated they were the corresponding higher oxidized
products, ketone 5 and aldehyde 7, seemingly derived from
dehydration of the nonconjugated and conjugated hydro-
peroxides, respectively. Therefore, we prepared authentic
standards of these compounds as well. The nonconjugated
ketone 5 was prepared by CrO3/tert-butyl hydroperoxide
oxidation of alcohol 4, and the conjugated aldehyde 7 via PCC
oxidation of alcohol 6 (Scheme 2).
II. Calibration of Peroxyl Radical Clock Derived From
3. In order to “calibrate” the radical clock derived from 3
(determination of kβ), we decomposed 3 in the presence of
various concentrations of α-tocopherol (α-TOH, 8) at 37 °C.
The resulting hydroperoxides were reduced to their corre-
sponding alcohols with PPh3, and the products were analyzed
by GC-FID. Examples of gas chromatograms obtained at high
and low concentrations of α-TOH are shown in Figure 1. The
resulting ratios of nonconjugated/conjugated products were
plotted vs [α-TOH] and the non-linear relationship fit to the
equation shown in Figure 2 to determine the rate constant for
β-fragmentation of the nonconjugated peroxyl radical in a given
solvent (Table 1). Values for kinh used to determine kβ were
obtained by laser flash photolysis for the reaction of α-TOH
with cumylperoxyl radicals in several different solvents.10
III. Representative Clocking Experiments Using Peroxy-
ester 3. To confirm the accuracy of the peroxyl radical clock
methodology utilizing 3, kinh values were determined in
chlorobenzene for a representative set of antioxidants having
RESULTS
■
I. Synthesis of Peroxyester 3 and its Derived
Products. Synthesis of the acid precursor to (E)-tert-butyl 4-
(naphthalen-2-yl)but-3-eneperoxoate 3 was achieved via a
Wittig reaction between commercially available 2-naphthalde-
hyde and the phosphonium salt derived from 3-bromopro-
pionic acid. Subsequent formation of the acid chloride and
substitution with tert-butyl hydroperoxide afforded the product
in 60% yield over the three steps. Upon its preparation,
isolation and purification, the first advantage of peroxyester 3
over the analogous phenyl compound 2 was evident: its higher
crystallinity enabled easy recrystallization and resulted in
277
dx.doi.org/10.1021/jo201862s|J. Org. Chem. 2012, 77, 276−284

The Journal of Organic Chemistry
Article
Scheme 2
Figure 1. Representative gas chromatograms (GC-FID) in the retention time range of product elution following the incubation of peroxyester 3 with
1.0 M α-TOH (A) and 0.03 M α-TOH (B) for 12 h at 37 °C in chlorobenzene. The peak labeled * is observed in most chromatograms and could
not be identified; however, the peak area does not change with antioxidant concentration and therefore appears not to be relevant in our kinetic
analysis.
Table 1. β-Fragmentation Rate Constants and Oxygen
Partition Coefficients of the Nonconjugated Peroxyl
a
Radical Derived from 3 as a Function of Solvent
b
c
d
solvent
hexanes
kβ (s−1
)
α
kinh (M−1 s−1
2.1 × 107
)
1.6 (±0.2) × 106
5.7 (±0.3) × 105
3.0 (±0.9) × 105
1.9 (±0.5) × 105
1.4 (±0.1) × 105
9.5 (±0.3) × 104
5.1 (±0.2) × 104
2.8 (±0.1) × 104
0.65
0.79
0.74
0.69
0.78
0.76
0.57
0.65
e
chlorobenzene
benzene
7.1 × 106
3.9 × 106
1.8 × 106
4.7 × 105
2.9 × 105
1.0 × 106
4.7 × 105
anisole
f
f
benzonitrile
ethyl acetate
acetic acid
propionitrile
a
Inhibition rate constants used in the determination of kβ and α
b
c
are given for reference. From decomposition of 3 at 37 °C. Values
d
are dimensionless and have errors of ± 0.01. Determined directly by
laser flash photolysis in the given solvent at 25 °C10 and corrected to
e
37 °C using log A = 8 and Ea = 1.6 kcal/mol, see text. See ref 11.
f
Value determined in acetonitrile, which has the same βH2 value
(0.45).12
Figure 2. Ratio of nonconjugated ([4] + [5]) to conjugated ([6] + [7])
oxidation products formed in the decomposition of 3 as a function of
[α-TOH, 8] following incubation for 12 h at 37 °C in benzene.
of varying amounts of the antioxidant and plotting the ratio of
nonconjugated to conjugated products as a function of
antioxidant concentration as in Figure 3. The inhibition rate
constants determined in this way are shown in Table 2 and are
in good-to-excellent agreement with literature values obtained
by the inhibited autoxidation of styrene.
varied structure and reactivity: 2,4,6-trimethylphenol (9), 2,6-
di-tert-butyl-4-methoxyphenol (10), diphenylamine (11), phe-
nothiazine (12), a pyrimidinol (13), and a pyridinol (14). The
values were obtained simply by decomposing 3 in the presence
278
dx.doi.org/10.1021/jo201862s|J. Org. Chem. 2012, 77, 276−284

The Journal of Organic Chemistry
Article
15 and 16 were incubated at 37 °C for 14 h in chlorobenzene
or acetonitrile. The mixture was then reduced with PPh3 and
analyzed by GC. The product analysis showed only alcohols 4
and 6 in the same 4:1 ratio observed when the hydroperoxide
mixture was reduced with PPh3 and analyzed immediately.
Next, the hydroperoxides were incubated at 37 °C for 14 h in
chlorobenzene or acetonitrile containing either 10, 11, 2,2,5,7,
8-pentamethyl-6-chromanol (a truncated version of α-tocopherol,
17), or N1,N1-dimethyl-N4-phenylbenzene-1,4-diamine (18)
over 14 h at different concentrations. Representative chromato-
grams are shown in Figure 4.
Importantly, although there was a difference in the product
composition (alcohols and carbonyls) as a function of antioxidant
Figure 3. Representative double-reciprocal plot used to obtain
inhibition rate constants. Shown is data obtained for 2,4,6-
trimethylphenol (9) in chlorobenzene at 37 °C.
IV. Origin of Carbonyl-Containing Products. Authentic
conjugated and nonconjugated hydroperoxides were prepared
in order to probe the conditions under which they could serve
as a source for the carbonyl products observed when
phenothiazine 12, pyrimidinol 13, or pyridinol 14 were clocked
or when the clock was calibrated with α-TOH in polar media.
Bloodworth et al. have reported that cinnamyl-derived
hydroperoxides can be prepared from the corresponding
cinnamyl chloride by reaction with excess hydrogen peroxide
and NaOH, albeit in low yield (11%) as an inseparable mixture
of isomers (conjugated/nonconjugated).18 When the analogous
(E)-2-(3-chloroprop-1-enyl)naphthalene was subjected to the
same conditions in a number of different solvents, the desired
hydroperoxides were not isolated, likely because of the
solubility of the starting material under these conditions.
Therefore, we decided to study the reactivity of the cinnamyl-
derived products, given that their reactivity should be
essentially the same as their naphthyl analogues. Following
the literature procedure, we were able to isolate the cinnamyl-
derived hydroperoxides as a 4:1 mixture of conjugated (15) and
nonconjugated (16) hydroperoxides.
structure and solvent, there was no significant change in the 4:1
ratio of conjugated/nonconjugated alcohol + carbonyl products.
While it would appear that the hydroperoxide products persist
in the presence of the phenols 10 and 17 and the diarylamine
11, they dehydrate to carbonyls in the presence of more basic
diarylamines, such as 18.19 Furthermore, the yield of carbonyl
products increases upon changing the solvent from chloro-
benzene to acetonitrile (see the Supporting Information). In
fact, in acetonitrile, carbonyl products are now also observed for
phenol 17.
V. Temperature Dependence of β-Fragmentation. To
estimate Arrhenius parameters for the β-fragmentation of the
nonconjugated peroxyl radical, 3 was decomposed in the
presence of varying concentrations (0.02−1.0 M) of α-
tocopherol (α-TOH, 8) at six additional temperatures (45,
50, 60, 70, 80, and 95 °C) in chlorobenzene. Values for kinh
used to determine kβ at these temperatures were estimated on
the basis of kinh = 6.4 × 106 M−1 s−1 in chlorobenzene at 25 °C
and log A = 8 (an estimated value, vide infra). The resulting
Arrhenius plot is shown in Figure 5, whose slope and intercept
yield Ea = 9.6 (±0.9) kcal/mol and log A = 12.8 (±0.6).
To test the stability of the hydroperoxides under the
conditions of the clocking experiment, the mixture of isomers
a
Table 2. Rate Constants for Reactions of 9−14 with Peroxyl Radicals Generated from Peroxyester 3 at 37 °C in Chlorobenzene
compound
clock kinh (M−1 s−1
)
lit kinh (M−1 s−1
)
T, ref
9
2.3 (±0.1) × 105
5.6 (±1.7) × 105
4.6 (±0.5) × 104
8.5 (±0.6) × 106
1.7 (±2.5) × 107
3.2 (±1.1) × 107
8.5 × 104
1.1 × 105
4.4 × 104
8.8 × 106
30 °C, 13
30 °C, 13
65 °C, 14
50 °C, 15
50 °C, 16
30 °C, 17
10
11
12
13
14
b
8.6 × 106
1.6 × 107
a
b
Literature values obtained by inhibited autoxidation of styrene at the given temperatures are shown for comparison. In benzene.
279
dx.doi.org/10.1021/jo201862s|J. Org. Chem. 2012, 77, 276−284

The Journal of Organic Chemistry
Article
Figure 4. Stacked plot of gas chromatograms (GC-FID) showing product distribution (4, 5, 6, and 7 with R = Ph) arising from hydroperoxide
decomposition (5 mM) in the presence of various additives (50 mM) incubated at 37 °C in chlorobenzene. Initial ratio of conjugated/
nonconjugated products (15/16) was 4:1.
Figure 5. Temperature dependence of the β-fragmentation of the
Figure 6. Kinetic isotope effect on the product distribution arising from
nonconjugated peroxyl radical derived from 3, which yields Ea = 9.6
the decomposition of 3 in the presence of 2,2,5,7,8-pentamethyl-6-
(±0.9) kcal/mol and log A = 12.8 (±0.6). Corresponding oxygen
chromanol (17) and either 1% D2O (■) or 1% H2O (●), which yield
partition coefficients (α) for each temperature are given in parentheses
kH = 5.0 × 106 M−1 s−1 and kD = 2.4 × 106 M−1 s−1 for kH/kD = 2.1.
next to each data point.
DISCUSSION
■
Radical clocks based on the carbon-skeleton rearrangements of
VI. Kinetic Isotope Effect Measurements. Deuterium
alkyl radicals have become an indispensable tool in kinetic and
mechanistic investigations of reactions where alkyl radicals have
been proposed as potential intermediates. Despite the ease with
which these types of experiments can be carried out,
heteroatom-centered radical clock approaches are few. Among
the limited examples are the particularly useful radical clocks
based upon competition between the β-fragmentation of a
nonconjugated peroxyl radical and its reduction to the
corresponding hydroperoxide by an H-atom donor, which
greatly enables the determination of rate constants for peroxyl-
molecule reactions.8
kinetic isotope effects (DKIEs) were measured for a
representative set of phenolic antioxidants using the peroxyl
radical clock methodology. Prior to each experiment, the
antioxidant was stirred with an excess (typically 1−5%) of D2O
(and an equivalent amount of H2O in a parallel experiment as a
control).20 Furthermore, the experiment was carried out in the
presence of 1% D2O to ensure adventitious water in the solvent
did not shift the equilibrium. A representative plot is shown in
Figure 6 for 2,2,5,7,8-pentamethyl-6-chromanol (17), and the
DKIEs for it and the other four phenols we examined are
summarized in Table 3. For comparison, DKIE values were also
measured using the methyl linoleate clock of Porter and co-
workers8 and are reported alongside.
The first peroxyl radical clocks relied on the chain transfer
reaction of the antioxidant-derived radical A· with the oxidizable
280
dx.doi.org/10.1021/jo201862s|J. Org. Chem. 2012, 77, 276−284

The Journal of Organic Chemistry
Article
phenyl π-system. Likewise, the oxygen partition coefficient
(α) is highly similar: 0.74 ± 0.01 for the naphthyl analogue
versus 0.76 ± 0.01 for the phenyl analogue.9 This suggested
that the clocks could be used essentially interchangeably.
The solvent effects on the β-fragmentation rate constants
obtained for the nonconjugated peroxyl radical derived from 3
reflect the same trends as those observed for that derived from
2,9 although the absolute values vary slightly from solvent to
solvent. The overall trend is generally one wherein solvents
with greater polarity generally slow the rate of β-fragmentation,
consistent with the idea that the peroxyl radical is better
solvated than the transition state for C−OO· bond dissociation
as the polarity of the medium increases.9 The oxygen partition
coefficients (α) show much higher variance between the two
clocks, presumably because of slightly differing orbital
interactions of the dioxygen SOMO with the SOMOs of the
naphthyl- or phenyl-substituted allyl radicals.3,25 Since oxygen
partitioning across delocalized carbon-centered radicals is
influenced by both sterics and electronics, rationalization of
solvent effects on these trends beyond this is difficult.
Nevertheless, with kβ and α defined for a given solvent, the
clock can be applied with confidence. We were able to
accurately clock reactions of phenols, diarylamines, pyridinols,
and pyrimidinols with inhibition rate constants ranging over
roughly three orders of magnitude (104−107 M−1 s−1): the
range of greatest interest for the assessment of natural products
with perceived radical-trapping antioxidant activity and the
development of novel synthetic compounds. The rate constants
were generally within a factor of 2 or better than literature
values obtained using the conventional inhibited autoxidation
of styrene methodology.
In carrying out these experiments, we noticed that the
amount of products observed in the chromatograms varied with
antioxidant. This suggested that the antioxidant was playing a
role in the decomposition of the peroxyester. The decom-
position of peroxyesters under both photolytic and thermal
conditions is well documented (reaction A in Scheme 4).
Homolytic cleavage of the weak O−O bond in conjunction
with (what is believed to be concerted)26 rupture of the C−C
bond results in extrusion of CO2 and formation of a stabilized
allylic radical and a tert-butoxyl radical (in the case of 2 or 3).
Additionally, peroxyester decomposition can be induced by
electron-rich compounds (e.g., dimethylaniline) through
donation of an electron into the σ* orbital of the O−O
bond, affording either an acyloxyl radical and an alkoxide (B) or
a carboxylate anion and an alkoxyl radical (C) as shown in
Scheme 3.27 Indeed, when the total area of the product peaks
Table 3. Deuterium Kinetic Isotope Effects (kH/kD) on the
Reactions of a Representative Group of Phenols with
Peroxyl Radicals Determined by the Peroxyl Radical Clock
a
Methodology with Peroxyester 3 at 37°C
kH/kD by methodology
peroxyester clock
methyl linoleate
clock
inhibited autoxidation
(ref)
(3)
b
b
b
c
19
10
17
13
14
2.5
2.2
2.1
2.1
1.6
2.5
2.3
2.4
1.5
1.5
n/a
n/a
5.1 (21)
3.1 (16)
n/a
c
a
Values obtained by inhibited autoxidation of styrene and the methyl
linoleate-based peroxyl radical clock are presented alongside for
b
comparison. Obtained in chlorobenzene containing 1% D2O (or 1%
c
H2O) as described below. Obtained in benzene containing 1% D2O
(or 1% H2O) as described below.22
substrate R−H in order to set up the clock (Scheme 1). Since this
chain-transfer reaction is quite slow (e.g., 0.1 M−1 s−1 for A· = α-
tocopheroxyl and R−H = methyl linoleate,23 for which ΔH ∼
2 kcal/mol) and will be slowed as A· becomes more stable/
persistent and R−H less oxidizable (e.g., allylbenzene, vide
supra), this approach is limited. This prompted us to consider
other precursors to delocalized radicals that could set up the
clock reaction. We first considered diacylperoxides, such as
21, but quickly found that they were simply too labile to be
practical.
We next turned to peroxyester 2, which worked nicely as
reported in our preliminary communication,9 but whose
oxidation products often appear in the same part of the (gas)
chromatogram as the antioxidant or antioxidant-derived
products. We surmised that by replacing the phenyl ring in 2
with a naphthyl ring as in 3, we could push the retention times
of the products much later in the chromatogram, improving the
resolution and reliability of the analysis. This worked well, with
the retention times shifting from ca. 5−10 min on our
optimized column/temperature conditions to ca. 12−22 min
under the same conditions and, importantly, resolving the
expected alcohols and unexpected corresponding carbonyl
compounds that were not apparent when 2 was used as
precursor to set up the clock (vide infra).
Scheme 3
The rate constant for β-fragmentation (kβ) of the non-
conjugated peroxyl radical derived from 3 was determined in
the same way as we did for 2 and was expectedly similar: kβ =
3.0 (±0.9) × 105 s−1 for the naphthyl analogue compared to
kβ = 1.7 (±0.1) × 105 s−1 for the phenyl analogue9 at 37 °C
using kinh = 3.9 × 106 M−1 s−1 at 37 °C for the standard reaction
of α-tocopherol with peroxyl radicals in benzene.24 The slight
increase in rate constant may reflect the greater delocalization of
the allylic radical by the naphthyl π-system as opposed to the
obtained from clocking experiments at the same concentration
of antioxidant ([ArOH] = 0.06 M) was plotted as a function of
the standard potential of the phenolic antioxidant under study
281
dx.doi.org/10.1021/jo201862s|J. Org. Chem. 2012, 77, 276−284

The Journal of Organic Chemistry
Article
(E° ranging from 0.12 to 0.9 V28), a linear correlation was
obtained (see Figure S2, Supporting Information), supporting
induced decomposition of the peroxyester by the antioxidant.
Since the clocking experiments are carried out under pseudo-
first-order conditions where the antioxidant is in excess, this
reaction does not confound our kinetic analysis and serves to
improve the yield of products for more reactive (i.e., electron
rich) compounds.
antioxidants. There is little of this data in the literature, owing
to how difficult it has been to carry out the measurements, and
the trends in the limited data available are difficult to
understand. For example, while phenol and aniline were
found to have log A = 7.2 (±0.5), Ea = 5.2 (±0.5) kcal/mol34
and log A = 6.6 (±0.5), Ea = 5.0 (±0.5) kcal/mol,34
respectively, butylated hydroxytoluene (BHT) and N-phenyl-
α-napthylamine were found to have log A = 4.2 (±0.3), Ea = 1.4
(±0.3) kcal/mol35 and log A = 5.1 (±0.5), Ea = 1.0 (±0.4)
kcal/mol,35 respectively. The latter result is particularly
troublesome given that inhibition rate constants of up to
107 M−1 s−1 have been measured for diarylamines.15
Deuterium kinetic isotope effects (DKIEs) have long provided
direct mechanistic support for H-atom transfer (HAT) or
proton-coupled electron transfer (PCET)36 reactions between
phenolic and aromatic amine antioxidants and peroxyl radicals.
Hammond and co-workers originally found very small KIEs for
these reactions, leading them to suggest that the mechanism of
their chain-breaking activity was due to rate-determining electron
transfer followed by a rapid proton transfer.37 Shortly
thereafter, Ingold and Howard demonstrated that Hammond’s
experimental approach had been flawed (protium from product
hydroperoxides in inhibited autoxidations could rapidly
exchange with the deuterium in the deuterated phenol, pre-
cluding the observation of the DKIE) and demonstrated that
there was indeed a sizable DKIE in these reactions.20 Since
then, DKIEs for the reactions of several phenols with peroxyl
radicals have been measured by inhibited autoxidation of
styrene in the presence of excess D2O (usually 1% by volume)
and generally fall into a range of ∼3−5 in chlorobenzene at
ambient temperatures.16,21,38
In the clocking experiments for compounds 9−14 shown in
Table 2, carbonyl products were observed in chromatograms
from reactions containing 12−14 but not in those containing
9−11. Since 12−14 each contain basic nitrogen atoms, we
surmised that a base-catalyzed dehydration mechanism may be
responsible for the formation of these carbonyls.29 Indeed,
when mixtures of the conjugated and nonconjugated hydro-
peroxides 15 and 16 were incubated with varying concen-
trations of different antioxidants under the same experimental
conditions as used in the clocking experiments, the
corresponding carbonyl-containing products were observed
for those antioxidants containing basic nitrogen atoms.
Furthermore, when the solvent was changed from chloroben-
zene to acetonitrile, the carbonyl-containing compounds were
observed in solutions containing the same additive but in
greater yield. When the antioxidant was replaced with
triethylamine, the same carbonyl compounds were formed in
quantities that correlated with triethylamine concentration. The
analogous carbonyl products derived from 2 were not identified
in our preliminary report,9 presumably because no compounds
having sufficient basicity were examined. In fact, when 2 is used
to clock these “basic” antioxidants (i.e., 12−14 or 18),
cinnamaldehyde (conjugated aldehyde) and 1-phenylprop-2-
en-1-one (nonconjugated ketone) are indeed observed as major
products in approximately the same ratios as 5 and 7 when
peroxyester 3 was used as the clock.
Expectedly, decomposition of 3 at higher temperatures gave
rise to higher yields of products. The product ratios were again
dependent on α-TOH concentration and could be fit to the
kinetics in Scheme 1. The solution of this expression for kβ and
α at the higher temperatures required kinh for α-tocopherol at
these temperatures. To the best of our knowledge, Arrhenius
parameters for the reaction of α-TOH with peroxyl radicals are
not available, but given the typical log A ∼ 8 for H-atom
transfer reactions as suggested by Benson30 and the inhibition
rate constant for α-TOH in chlorobenzene of 6.4 × 106 M−1 s−1
at 25 °C,10 we calculated Ea = 1.6 kcal/mol and were able to
estimate the kinh at the required temperatures (37, 45, 50, 60,
70, 80, and 95 °C). The resulting Arrhenius plot for the β-
fragmentation of the nonconjugated peroxyl radical yields Ea =
9.6 ± 0.9 kcal/mol and log A = 12.8 ± 0.6.31 Although these are
only estimates, they seem reasonable. The calculated bond
dissociation enthalpy (BDE) of the C−OO· bond in the
nonconjugated peroxyl radical calculated by ROB3P86/6-
311G**//UB3P86/6-311G** is 9.8 kcal/mol.32 This method
has been shown to give C−OO· BDEs in excellent agreement
with experimental values; e.g., for benzylperoxyl, the calculated
C−OO· BDE value was 22.2 kcal/mol, and it was determined
to be 21.8 ± 0.9 kcal/mol experimentally.33 Furthermore,
log A ∼ 13 for a bond dissociation where a minimum of two
bond rotations must be restricted also seems reasonable from
first principles.
Utilizing peroxyester 3 (and excess D2O/H2O), we obtained
highly reproducible primary DKIEs for a representative series of
compounds, whose magnitude were roughly half of what were
obtained by inhibited autoxidation of styrene where compar-
ison was possible. As an independent check, we measured
DKIEs for the same compounds using the first reported peroxyl
radical clock methodology with methyl linoleate as the
oxidizable substrate8 and found essentially the same isotope
effects as those obtained using the peroxyester approach. The
reason for the discrepancy between our measured DKIEs and
those obtained using the inhibited autoxidation of styrene
method is unclear and under further investigation.
CONCLUSION
■
β,γ-Unsaturated peroxyesters are convenient and versatile
precursors to delocalized radicals for use in peroxyl radical
clock experiments. In particular, peroxyester 3 yields products
that are most conveniently separated by gas chromatography,
and resolution of the corresponding alcohol and carbonyl
products has allowed us to expand the application of the clock
methodology to compounds containing basic moieties with
confidence. Additionally, we have demonstrated the utility of
peroxyester 3 in measuring deuterium kinetic isotope effects
(DKIEs) for a variety of phenolic antioxidants and also
discussed its potential in the determination of Arrhenius
parameters for peroxyl radical-trapping reactions.
EXPERIMENTAL SECTION
■
Synthesis of (E)-tert-butyl 4-(naphthalen-2-yl)but-3-eneper-
oxoate (3). To a suspension of (E)-4-(naphthalen-2-yl)but-3-enoic
acid (3.00 g, 14.2 mmol) in dry benzene (45 mL) was added SOCl2
(3.38 g, 28.4 mmol) in one portion. The reaction mixture was heated
We anticipate that the use of 3 to measure inhibition rate
constants at higher temperatures will be useful in estimating
Arrhenius parameters for various phenolic and aromatic amine
282
dx.doi.org/10.1021/jo201862s|J. Org. Chem. 2012, 77, 276−284

The Journal of Organic Chemistry
Article
at 50 °C until complete as indicated by TLC (ca. 2 h). The solvent was
evaporated in vacuo, and the yellow solid obtained was dissolved in
dry CH2Cl2 (75 mL) and cooled to 0 °C. tert-Butyl hydroperoxide (6.7
mL of 5.5 M solution in decane, 36.9 mmol) was then added dropwise,
followed by pyridine (freshly distilled over CaH2, 4.04 g, 51.1 mmol)
as a solution in 20 mL of CH2Cl2 (dropwise). The reaction was stirred
at 0 °C until complete as indicated by TLC (ca. 40 min). The reaction
mixture was poured into 50 mL of ice water and extracted with
CH2Cl2 (2 × 30 mL). The organics were washed with 10% HCl (2 ×
20 mL), saturated NaHCO3 (2 × 20 mL), and brine (20 mL) and
dried over MgSO4. Column chromatography (1:9 Et2O/petroleum
ether) afforded pure product as a white solid. Crystalline product was
obtained by recrystallization from ether/hexanes at −20 °C. Yield:
Engineering Research Council of Canada, the Ontario Ministry
of Research and Innovation, Queen’s University, and the
University of Ottawa. D.A.P. also acknowledges the support of
the Canada Research Chairs program.
REFERENCES
■
(1) Ingold, K. U. Acc. Chem. Res. 1969, 2, 1.
(2) Porter, N. A. Acc. Chem. Res. 1986, 19, 262.
(3) Pratt, D. A.; Porter, N. A.; Tallman, K. A. Acc. Chem. Res. 2011,
44, 458.
(4) Ingold, K. U. Chem. Rev. 1961, 61, 563.
(5) Mahoney, L. R. Angew. Chem., Int. Ed. 1969, 8, 547.
(6) Burton, G. W.; Ingold, K. U. Acc. Chem. Res. 1986, 19, 194.
(7) Valgimigli, L.; Pratt, D. A. In Encyclopedia of Radicals in
Chemistry, Biology and Materials; Chatgillialogu, C., Studer, A., Eds.;
Wiley: Hoboken, NJ, 2011.
(8) Roschek, B.; Tallman, K. A.; Rector, C. L.; Gillmore, J. G.; Pratt,
D. A.; Punta, C.; Porter, N. A. J. Org. Chem. 2006, 71, 3527.
(9) Jha, M.; Pratt, D. A. Chem. Commun. 2008, 1252.
(10) Valgimigli, L.; Banks, J. T.; Lusztyk, J.; Ingold, K. U. J. Org.
Chem. 1999, 64, 3381.
(11) When the log kinh values for α-TOH obtained by Valgimigli et
al. using laser flash photolysis are plotted vs solvent βH2, an excellent
linear correlation is obtained on which all points reside with the
exception of chlorobenzene (see Figure S1, Supporting Information).
Since the rate constant obtained in chlorobenzene is almost certainly
in error (it is lower than that obtained in benzene, which has a higher
βH2 value), we used the kinh value derived from the log kinh−βH2
correlation, which is 6.4 × 106 M−1 s−1 at 25 °C.
(12) Abraham, M. H.; Grellier, P. L.; Prior, D. V.; Duce, P. P.;
Morris, J. J.; Taylor, P. J. J. Chem. Soc., Perkin Trans. 1989, 2, 699.
(13) Burton, G. W.; Doba, T.; Gabe, E. J.; Hughes, L.; Lee, F. L.;
Prasad, L.; Ingold, K. U. J. Am. Chem. Soc. 1985, 107, 7053.
(14) Brownlie, I. T.; Ingold, K. Can. J. Chem. 1966, 44, 861.
(15) Lucarini, M.; Pedrielli, P.; Pedulli, G. F.; Valgimigli, L.; Gigmes,
D.; Tordo, P. J. Am. Chem. Soc. 1999, 121, 11546.
1
2.67 g (66%); H NMR (CDCl3, 400 MHz) δ 7.81−7.77 (m, 3H),
7.72 (s, 1H), 7.58 (dd, J = 2.0, 4.8 Hz, 1H), 7.49−7.42 (m, 2H), 6.72
(d, J = 15.8 Hz, 1H), 6.39 (td, J = 7.1, 15.8 Hz, 1H), 3.33 (dd, J = 1.4,
7.1 Hz, 2H), 1.35 (s, 9H); 13C NMR (CDCl3, 100 MHz) δ 168.7,
134.3, 133.9, 133.4, 133.0, 128.2, 127.9, 127.6, 126.2, 126.2, 125.9,
123.3, 120.5, 83.6, 35.6, 26.0. HRMS (EI) Calculated: 284.1412.
Actual: 284.1451. The starting acid, nonconjugated alcohol 4,39
nonconjugated ketone 5,39 conjugated alcohol 6,40 and conjugated
aldehyde 740 were prepared according to literature procedures.
Calibration Experiments. To a screw-capped GC vial was added
peroxyester 3 (0.01 M final concentration), α-TOH (8) (0.02−1.0 M
final concentration), and the desired solvent to a total volume of
100 μL. The samples were incubated for 2−14 h, quenched with
100 μL of 1 M PPh3, and diluted to 1 mL with acetonitrile for analysis.
GC analysis was carried out using an Agilent DB-5 column (30 m ×
0.32 μm × 0.25 μm) with the following temperature profile: 130 °C
hold 5 min, 2 °C/min to 162 °C, 30 °C/min to 280 °C, hold
5 min. Response factors for the nonconjugated alcohol 4, non-
conjugated ketone 5, conjugated alcohol 6, and conjugated aldehyde 7
are 1.85, 1.25, 1.21, and 1.83, respectively, relative to benzyl alcohol. The
resulting plot of ([4] + [5])/([6] + [7]) vs [α-TOH] was fit using
nonlinear regression to obtain kβ and α.
Clocking Experiments. To a screw-capped GC vial was added
peroxyester 3 (0.01 M final concentration), H-atom donor (0.02−
1.0 M final concentration depending on kinh), and the desired solvent
to a total volume of 100 μL. The samples were incubated for 2−14 h,
quenched with 100 μL of 1 M PPh3, and diluted to 1 mL with
acetonitrile for analysis. GC analysis was carried out using an Agilent
DB-5 column (30 m × 0.32 μm × 0.25 μm) with the following
temperature profile: 130 °C hold 5 min, 2 °C/min to 162 °C, 30 °C/
min to 280 °C, hold 5 min. A plot of ([6] + [7])/([4] + [5]) vs 1/[H-
atom donor] was fit linearly to obtain kinh. Deuterium kinetic isotope
effects were obtained by carrying out measurements in the same
manner as above, with the addition of 1% D2O (or 1% H2O as a
control) to the solvent (after distillation over CaH2).
(16) Pratt, D. A.; DiLabio, G. A.; Brigati, G.; Pedulli, G. F.;
Valgimigli, L. J. Am. Chem. Soc. 2001, 123, 4625.
(17) Wijtmans, M.; Pratt, D. A.; Valgimigli, L.; DiLabio, G. A.;
Pedulli, G. F.; Porter, N. A. Angew. Chem., Int. Ed. 2003, 42, 4370.
(18) Bloodworth, A. J.; Curtis, R. J.; Spencer, M. D.; Tallant, N. A.
Tetrahedron 1993, 49, 2729.
(19) Kornblum, N.; DeLaMare, H. E. J. Am. Chem. Soc. 1951, 73,
880.
(20) Ingold, K. U.; Howard, J. A. Nature 1962, 195, 280.
(21) Valgimigli, L.; Amorati, R.; Petrucci, S.; Pedulli, G. F.; Hu, D.;
Hanthorn, J. J.; Pratt, D. A. Angew. Chem., Int. Ed. 2009, 48, 8348.
(22) Russell, G. A. J. Am. Chem. Soc. 1957, 79, 3871.
(23) Ingold, K. U.; Bowry, V. W.; Stocker, R.; Walling, C. Proc. Natl.
Acad. Sci. U.S.A. 1993, 90, 45−49.
(24) To exclude any role of self-association of α-TOH on the
determination of kβ at the higher concentrations required to accurately
calibrate the clock, we also calibrated the clock using 2,6-di-tert-butyl-
4-methoxyphenol as the H-atom donor (kinh = 1.1 × 105 M−1 s−1 at
30 °C in chlorobenzene by the inhibited autoxidation of styrene
compared to kinh = 3.2 × 106 M−1 s−1 for α-TOH determined by the
same investigators under the same conditions)13 and found a
difference of 3.1-fold on kβ. In this case, the double-reciprocal plots
were compared for [antioxidant] = 0.02−1.8 M because fitting of the
data obtained for 2,6-di-tert-butyl-4-methoxyphenol was not possible
because of its slower reaction with peroxyl radicals.
ASSOCIATED CONTENT
■
S
* Supporting Information
Figures S1−S34, all raw data for clock calibration, clocking
experiments, temperature-dependence, and isotope effects. This
material is available free of charge via the Internet at http://
pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: dpratt@uottawa.ca.
ACKNOWLEDGMENTS
■
We thank Philip T. Lynett (Queen’s University) for measuring
deuterium kinetic isotope effects using the methyl linoleate
clock methodology and Dr. Luca Valgimigli (University of
Bologna) for discussions regarding the Arrhenius parameters
for the reactions of phenols with peroxyl radicals. We are
grateful for the support of the Natural Sciences and
(25) Hu, D.; Pratt, D. A. Chem. Commun. 2010, 46, 3711.
(26) Bartlett, P. D.; Hiatt, R. R. J. Am. Chem. Soc. 1958, 80, 1398.
(27) Antonello, S.; Formaggio, F.; Moretto, A.; Toniolo, C.; Maran,
F. J. Am. Chem. Soc. 2001, 123, 9577.
(28) Lien, E. J.; Ren, S. J.; Bui, H. Y. H.; Wang, R. B. Free Radical
Biol. Med. 1999, 26, 285.
283
dx.doi.org/10.1021/jo201862s|J. Org. Chem. 2012, 77, 276−284

The Journal of Organic Chemistry
Article
(29) Ellam, R. M.; Padbury, J. M. J. Chem. Soc., Chem. Commun.
1972, 1086.
(30) Benson, S. W. Thermochemical Kinetics, 2nd ed.; John Wiley &
Sons: New York, 1976.
(31) If log A = 7.5 is used for the reaction of α-TOH with peroxyls,
log A and Ea for the β-fragmentation of the nonconjugated peroxyl
radical become 12.3 and 8.9 kcal/mol, respectively. Similarily, if log
A = 8.5 is used for the reaction of α-TOH with peroxyls, log A and Ea
for the β-fragmentation of the nonconjugated peroxyl radical become
13.3 and 10.3 kcal/mol, respectively.
(32) DiLabio, G. A.; Pratt, D. A. J. Phys. Chem. A 2000, 104, 1938.
(33) Fenter, F. F.; Noziere, B.; Carlap, P.; Lesclaux, R. Int. J. Chem.
Kinet. 1994, 26, 171.
(34) Chenier, J. H. B.; Furimsky, E.; Howard, J. A. Can. J. Chem.
1974, 52, 3682.
(35) Howard, J. A.; Furimsky, E. Can. J. Chem. 1973, 51, 3738.
(36) Litwinienko, G.; Ingold, K. U. Acc. Chem. Res. 2007, 40, 222.
(37) Hammond, G. S.; Boozer, C. E.; Hamilton, C. E.; Sen, J. N.
J. Am. Chem. Soc. 1955, 77, 3238.
(38) Burton, G. W.; Ingold, K. U. J. Am. Chem. Soc. 1981, 103, 6472.
(39) Brigham & Women’s Hospital.; Augelli Szafran, C. E.; Wolfe,
M. S.; Han Xun, W.; European Patent Office, 2009, WO2008US11649
20081010.
(40) Avery, T. D.; Caiazza, D.; Culbert, J. A.; Taylor, D. K.; Tiekink,
E. R. T. J. Org. Chem. 2005, 70, 8344.
284
dx.doi.org/10.1021/jo201862s|J. Org. Chem. 2012, 77, 276−284