Determination of the isomerization rate constant HOCH2CH2CH2CH(OO·)CH3→HOC·HCH2 CH2CH(OOH)CH3. Importance of intramolecular hydroperoxy isomerization in tropospheric chemistry

Article abstract of DOI:10.1002/(SICI)1097-4601(1998)30:12<875::AID-KIN2>3.0.CO;2-8

The rate constant of the title reaction is determined during thermal decomposition of di-n-pentyl peroxide C₅H₁₁O-OC₅H₁₁ in oxygen over the temperature range 463-523 K. The pyrolysis of di-n-pentyl peroxide in O₂/N₂ mixtures is studied at atmospheric pressure in passivated quartz vessels. The reaction products are sampled through a micro-probe, collected on a liquid-nitrogen trap and solubilized in liquid acetonitrile. Analysis of the main compound, peroxide C₅H₁₀O₃, was carried out by GC/MS, GC/MS/MS [electron impact El and NH₃ chemical ionization Cl conditions]. After micro-preparative GC separation of this peroxide, the structure of two cyclic isomers (3S*,6S*)3α-hydroxy-6-methyl-1,2-dioxane and (3R*,6S*)3α-hydroxy-6-methyl-1,2-dioxane was determined from ¹H NMR spectra. The hydroperoxy-pentanal OHC-(CH₂)₂-CH(OOH)-CH₃ is formed in the gas phase and is in equilibrium with these two cyclic epimers, which are predominant in the liquid phase at room temperature. This peroxide is produced by successive reactions of the n-pentoxy radical: a first one generates the CH₃C^·H(CH₂)₃OH radical which reacts with O₂ to form CH₃CH(OO^·)(CH₂)₃OH; this hydroxy-peroxy radical isomerizes and forms the hydroperoxy HOC^·H(CH₂)₂CH(OOH)CH₃ radical. This last species leads to the pentanal-hydroperoxide (also called oxo-hydroperoxide, or carbonyl-hydroperoxide, or hydroperoxypentanal). By adding small amounts of NO (0-1.6×10¹⁵ molecules cm_-3) to the di-n-pentyl peroxide/O₂/N₂ mixtures, the pentanal-hydroperoxide concentration was decreased, due to the consumption of RO₂ radicals by reaction (7). The pentanal-hyroperoxide concentration was measured vs. NO concentration at ten temperatures (463-523 K). The isomerization rate constant involving the H atoms of the CH₂-OH group was deduced. The comparison of this rate constant to thermokinetics estimations leads to the conclusion that the strain energy barrier of a seven-member ring transition state is low and near that of a six-member ring. Intramolecular hydroperoxy isomerization reactions produce carbonyl-hydroperoxides which (through atmospheric decomposition) increase concentration of radicals and consequently increase atmospheric pollution, especially tropospheric ozone, during summer anticyclonic periods. Therefore hydrocarbons used in summer should contain only short chains (4) hydrocarbons or totally branched hydrocarbons, for which isomerization reactions unlikely.

Full text of DOI:10.1002/(SICI)1097-4601(1998)30:12<875::AID-KIN2>3.0.CO;2-8

Determination of the
Isomerization Rate
Constant
и
HOCH CH CH CH(OO )CH :
2
2 2
3
и
HOC HCH CH CH(OOH)CH .
2
2
3
Importance of
Intramolecular
Hydroperoxy Isomerization
in Tropospheric Chemistry
1
1
1
2
OLIVIER PERRIN, ADOLPHE HEISS, KRIKOR SAHETCHIAN, LUCIEN KERHOAS,
JACQUES EINHORN2
1Laboratoire de M e´ canique Physique, CNRS UPRESA 7068, Universit e´ P. & M. Curie (Paris 6) 2, Place de la Gare de
Ceinture, 78210 Saint-Cyr l’Ecole, France
2
Unit e´ de Phytopharmacie et M e´ diateurs Chimiques, INRA, Route de St-Cyr, 78026 Versailles Cedex, France
Received 9 June 1997; accepted 21 May 1998
ABSTRACT: The rate constant of the title reaction is determined during thermal decomposition
of di-n-pentyl peroxide C H O9OC H in oxygen over the temperature range 463–523 K. The
5
11
5
11
pyrolysis of di-n-pentyl peroxide in O /N mixtures is studied at atmospheric pressure in pas-
2
2
sivated quartz vessels. The reaction products are sampled through a micro-probe, collected
on a liquid-nitrogen trap and solubilized in liquid acetonitrile. Analysis of the main compound,
peroxide C H O , was carried out by GC/MS, GC/MS/MS [electron impact EI and NH chemical
5
10
3
3
ionization CI conditions]. After micro-preparative GC separation of this peroxide, the structure
of two cyclic isomers (3S*,6S*)3␣-hydroxy-6-methyl-1,2-dioxane and (3R*,6S*)3␣-hydroxy-6-
1
methyl-1,2-dioxane was determined from H NMR spectra. The hydroperoxy-pentanal OHC
9
(CH ) 9CH(OOH)9CH is formed in the gas phase and is in equilibrium with these two
2 2 3
cyclic epimers, which are predominant in the liquid phase at room temperature. This peroxide
is produced by successive reactions of the n-pentoxy radical: a first one generates the
CH3C
и
H(CH ) OH radical which reacts with O to form CH CH(OO )(CH ) OH; this hydroxy-
и
2
3
2
3
2 3
peroxy radical isomerizes and forms the hydroperoxy HOC
и
H(CH ) CH(OOH)CH radical. This
2 2 3
Correspondence to: A. Heiss
Contract grant Sponsor: EUROTRAC-LACTOZ
1998 John Wiley & Sons, Inc. CCC 0538-8066/98/120875-13
᭧

8
76
PERRIN ET AL.
last species leads to the pentanal-hydroperoxide (also called oxo-hydroperoxide, or carbonyl-
hydroperoxide, or hydroperoxypentanal), by the reaction HOC
и
H(CH ) CH(OOH)CH ϩ
2
2
3
O : O"CH(CH ) CH(OOH)CH ϩ HO .
2
2
2
3
2
The
isomerization
rate
constant
HOCH CH CH CH(OO
и
)CH :
2
2
2
3
HOC HCH CH CH(OOH)CH (k ) has been determined by comparison to the competing well-
и
2 2 3 3
known reaction RO ϩ NO : RO ϩ NO (k ). By adding small amounts of NO (0–1.6 ϫ
0 molecules cm ) to the di-n-pentyl peroxide/O /N mixtures, the pentanal-hydroperoxide
2 2
2
Ϫ3
2
7
1
5
1
concentration was decreased, due to the consumption of RO radicals by reaction (7). The
2
pentanal-hydroperoxide concentration was measured vs. NO concentration at ten tempera-
tures (463–523 K). The isomerization rate constant involving the H atoms of the CH 9OH
2
group was deduced:
10
Ϫ1
Ϫ1
k ϭ (6.4 Ϯ 0.6) ϫ 10 exp{ Ϫ (16,900 Ϯ 700)cal mol /RT}s
3
or per H atom:
k3(H) ϭ (3.2 Ϯ 0.3) ϫ 10 exp{ Ϫ (16,900 Ϯ 700)cal mol /RT}s
1
0
Ϫ1
Ϫ1
The comparison of this rate constant to thermokinetics estimations leads to the conclusion
that the strain energy barrier of a seven-member ring transition state is low and near that of
a six-member ring. Intramolecular hydroperoxy isomerization reactions produce carbonyl-hy-
droperoxides which (through atmospheric decomposition) increase concentration of radicals
and consequently increase atmospheric pollution, especially tropospheric ozone, during sum-
mer anticyclonic periods. Therefore, hydrocarbons used in summer should contain only short
chains (ϽC ) hydrocarbons or totally branched hydrocarbons, for which isomerization reac-
4
tions are unlikely. ᭧ 1998 John Wiley & Sons, Inc. Int J Chem Kinet 30: 875–887, 1998
INTRODUCTION
oxide [5]. This compound of formula O"
CHCH CH CH O9OH (MW 104, C H O ) is an al-
2
2
2
4
8
3
The role of reactions involving alkoxy RO and peroxy
dehyde-hydroperoxide in the gas phase, which is in
equilibrium with its acyclic peroxide form 3␣-hydroxy
1,2-dioxane in the liquid phase. The formation of this
complex hydroperoxide occurs through two consecu-
tive isomerization reactions of alkoxy radicals; it can
be expected that such hydroperoxides could also be
generated after abstraction of an H atom from oxy-
genated hydrocarbons.
RO radicals in combustion and in atmospheric chem-
2
istry needs significant elucidation. Up to now, decades
of investigations show that for higher alkanes (ՆC4 )
RH complexities arise, due to the diversity of the re-
actions of radicals as well as to the formation of many
species. In the field of combustion, isomerization re-
actions of RO radicals through the formation of ke-
2
tohydroperoxides [1] play a main role in the slow ox-
It is well established that the formation of pollutants
from hydrocarbons occurs through complex processes,
but their chemical mechanisms are not all well char-
acterized. For decades, many suggestions have been
made, and a large number of theoretical investigations
idation of (ՆC ) alkanes, and introduction of those
5
isomerization reactions enabled prediction of the onset
of autoignition for (ՆC ) alkanes in motored-engines.
5
In modeling atmospheric pollution, this isomerization
reaction of ordinary alkyl peroxy radicals plays no role
about isomerization reactions of RO and RO radicals
2
[2]. We have previously shown experimentally that the
have been performed [4,6–10]. Only two experimen-
isomerization [3] reaction of the OOCH CH -
tal determinations of RO radical isomerization kinet-
2
2
2
CH CH OH radical becomes the main process, in the
temperature range 343–503 K. This is due to the
weakening, by ca. 4 kcal/mol, of the C9H bonds in
ics can be found in the literature, one by Walker and
co-workers [11] and the other by Pilling and co-work-
ers [12], ten years later. Arrhenius parameters were
2
2
the CH OH group, because of the presence of the
determined
for
the
isomerization reaction
2
C9O bond as shown in thermokinetic studies on hy-
droperoxy-propyl radicals [4]. The n-OOC H OH rad-
ical is generated by an initial isomerization reaction of
neo-C H O : C H OOH. Neopentane was chosen
5
11
2
5
10
as reference fuel by these two groups for two reasons:
(i) all H atoms are equivalent and (ii) it cannot react
with oxygen to generate an alkene. Small amounts
(1%) of neopentane [11] were added to slowly reacting
4
8
the n-C H O radical followed by a subsequent addition
4
9
of O which leads to a previously unreported per-
2

ISOMERIZATION RATE CONSTANT
877
mixtures of H ϩ O over the temperature range
The addition of small quantities (0–22.5 ppm) of
2
2
6
53–793 K and the following initial products were
nitric oxide NO to this system n-C H O/O /N , con-
4
9
2
2
observed: 3,3-dimethyloxetan (DMO), acetone, i-bu-
tene, methane, and formaldehyde. From initial values
sumed the peroxy radicals formed and, consequently,
lowered the formation rate of this hydroperoxide; on
the contrary, butanal yield remained unchanged with
NO concentration. The competing reactions are either
isomerization (HOROO : HORϪHOOH) or reaction
of the ratios R ϭ ([DMO] ϩ [acetone])/[i-butene]
1
and R ϭ [acetone]/[DMO] at each temperature
2
used, Arrhenius parameters were determined for
several elementary steps, among which was the iso-
merization reaction of neopentylperoxy radicals
with NO (HOROO ϩ NO : HORO ϩ NO ). Using
2
the known rate constant for the NO reaction, an esti-
mation of the isomerization reaction rate constant
HOCH CH CH CH OO : HOCHCH CH CH OOH
(
CH ) CCH O : (CH ) C(CH )CH OOH (3). The
3 3 2 2 3 2 2 2
1
3
recommended value of k3 is 1.20 ϫ 10 exp
2
2
2
2
2
2
2
Ϫ1
Ϫ1
(Ϫ120 kJ mol /RT) s . The same rate constant k
was made. Flow experiments were performed in a pas-
sivated quartz vessel (id 10mm) at temperature
487 K, and linear decrease of the concentration ratio
[O"CHCH CH CH O9OH]/[butanal] vs. [NO] in
3
has been determined directly [12] over the temperature
range 660–750 K. The neopentylperoxy radical is
generated by flash photolysis of neo-pentyliodide in
2
2
2
3
Ϫ1
the presence of O and He. The reactions of the hy-
ppm was observed; a value of 1.5 ϫ 10 s was
2
droperoxy radical (CH ) C(CH )CH OOH, formed by
isomerization, lead to the generation of OH which is
detected at high sensitivity by laser induced fluores-
found for the isomerization rate constant, at this tem-
3
2
2
2
1
1
perature. Assuming a value of 10 for the preexpo-
nential factor, this led to an activation energy of
17.6 kcal/mol, and to an estimate of this previously
cence (LIF). An Arrhenius analysis gives k ϭ
3
12.2
Ϯ
0.77
Ϫ1
Ϫ1
1
0
exp Ϫ(122.7 Ϯ 10.3 kJ mol /RT) s . Pil-
undetermined isomerization rate constant: k
ϭ
isom
1
1
ling and co-workers [12] noticed that the values ob-
10 exp(Ϫ17,600/RT). Recent theoretical investiga-
tions about atmospheric oxidation mechanism of n-
butane and the fate of alkoxy radicals [13] explain our
results, and reach conclusions which are in good
agreement with ours, especially about atmospheric im-
plication.
tained for k are over a factor of 10 lower than the
3
values of Walker and co-workers [11], which likely
means that the rate constant of the reverse reac-
tion (CH ) CCH O : (CH ) CCH ϩ O , would be
3
3
2
2
3
3
2
2
much higher than expected.
During recent investigations on n-butoxy [3] radi-
cals in the temperature range 343–503 K in mixtures
of O /N , the formation of an aldehyde-hydroperox-
After this first attempt on a C species (the n-butoxy
radical), our intention was to carry out investigations
4
on the next homologue (C ) isomerization, in this un-
2
2
5
ide, as a major compound, was shown, formed by two
consecutive isomerizations of these radicals, accord-
ing to the following scheme (only H atoms playing a
significant role are written):
usual series of peroxy radicals. The behavior
of these HOCH CH CH CH OO
и
and HOCH2
2
2
2
2
CH CH CH(OO
и
)CH3 radicals is indeed different
2
2
from that of RO radicals containing only two O at-
2
oms, because of the ␣ C9H bond which is weakened
by the presence of the C9O bond. This weakness
favours the H atom migration from the ␣ C to the OO
group (linked to the ␦ C), leading to the isomerized
iso
O
2
и
OCCCC 9
9
: HOCCCC
и
99 : HOCCCCOO
и
iso
O
2
99
: HOC CCCOOH 99 : O"CCCCOOH ϩ HO2
и
HOC HCH CH CH(OOH)CH radical and, finally, to
и
2 2 3
the identified aldehyde-hydroperoxide. This situation
contrasts with that of neo-pentoxy radicals where all
C9H bonds are equivalent. The interest of the present
work is twofold: (i) from a theoretical point of view,
to determine the isomerization rate constant of an
The ratio butanal/(butanal ϩ isomerization products)
was systematically investigated in flow experiments at
six temperatures ranging from 448 to 496 K, as a func-
tion of the oxygen percentage range (from 20 to 80%
of O ). This ratio was found equal to 0.290 Ϯ 0.035
␴) and independent of oxygen concentration. Butanal
is mainly produced consecutively to the first isomeri-
zation [3], and its formation was proposed to occur
according to the following sequence:
HOCCCC : HOC CCC, followed by the O addi-
tion reaction HOC
HO ; the direct oxidation route
2
HORO radical, and (ii) in practice, this type of isom-
2
(
erization should play a main role in the field of com-
bustion and in atmospheric chemistry.
The isomerization studied in this paper,
и
OCCCC :
HORO : HORϪHOOH is in competition with the re-
2
и
и
action with nitric oxide HORO ϩ NO : HORO ϩ
2
2
и
CCC ϩ O : O"CCCC ϩ
NO , as previously [3] noted. Explanation of the isom-
2
2
и
OCCCC ϩ O :
erization mechanism occurring inside the HORO rad-
2
2
2
HO"CCCC ϩ HO is of minor importance, from
room temperature to ca. 500 K.
ical, i.e., from which C an H atom is internally shifted
to the OO group, needs accurate knowledge of the
2

8
78
PERRIN ET AL.
structure of the complex hydroperoxide resulting from
this isomerization, and this has been done in the
present work by using a combination of modern ana-
lytical techniques.
GC/MS CHROMATOGRAM
The GC/MS chromatogram, obtained under Electron
Impact EI conditions and displayed (total ion current
TIC vs. retention time in s) in Figure 1, shows good
separation between the species formed during a flow
oxidation experiment performed at 463 K in a
EXPERIMENTAL
1
6 mm id passivated quartz vessel, O /N 7:3, resi-
2 2
The experimental set up and most operating conditions
were essentially the same as those previously de-
scribed [3], a brief review is presented here. Reactions
of alkoxy radicals are carried out in a tubular quartz
flow reactor located in a thermoregulated (Ϯ0.5 K)
oven. The n-pentoxy radical is generated through py-
rolysis of small amounts (tenths of ppm) of di-n-pen-
tylperoxide in O /N mixtures at atmospheric pressure.
dence time ca. 35 s. Separation was carried out on
capillary column HP-1, 25 m ϫ 0.20 mm ϫ
a
0.50 ␮m, the carrier gas being He 99.9999%, and the
following GC oven temperature programme: 313 to
363 K at a rate of 10 K/min, 363 to 373 K at a rate of
2 K/min, 373 to 423 K at a rate of 10 K/min. It is
obvious that the peroxide to be identified is the major
compound. In fact, it elutes over a long time (ca.
100 s) and appears mainly under the form of two peaks
of respective retention times at 459 (major) and
536 s, both (7 and 8) having the same mass spectrum.
The other species include (in order of elution) 2,3 di-
2
2
The di-n-pentylperoxide was synthesized according to
the method of Welch et al. [14], purified by vacuum
distillation and checked by GC/MS. Gaseous flows of
O , of N and of N carrying vapours of di-n-pentyl-
2
2
2
peroxide, electronically regulated (by mass flow con-
trollers RDM 280 from l’Air Liquide-Alphagaz, flow
accuracy Ϯ1% of full scale) are mixed before entering
the quartz reactor at atmospheric pressure. Gas flow,
at the reactor outlet, is continuously sampled through
a microprobe and the condensable species are trapped
hydro-5-methyl-furan,
tetrahydro-2-methyl-furan,
pentanal (a main compound), 1-pentanol, 2-(3H)-di-
hydro-5-methyl-furanone (␥-valerolactone). These
species were first compared by computer analysis of
the observed mass spectra and that of the NIST 92
library. The proposed products, obtained from either
Aldrich or Fluka, and injected under identical chro-
matographic conditions gave the same retention times
and the same EI mass spectra, thus, providing positive
identification. The unreacted di-n pentyl peroxide
elutes at 849 s and is not present on the chromatogram.
on the finger of a Dewar, cooled by liquid N . At the
2
end of an experiment, the Dewar is dismantled, the
condensed end-products solubilized in liquid acetoni-
trile (ca. 50 ␮l), and samples of ca. 1 ␮l injected into
the GC column of the GC/MS where they are ana-
lysed. Determination of the isomerization rate constant
under study has been made in the temperature range
4
63–523 K, by adding small amounts of NO (from 0
IDENTIFICATION OF THE ALDEHYDE-
HYDROPEROXIDE
1
5
Ϫ3
to 1.6 ϫ 10 molecules cm at the maximum) to the
system O /N /ROOR. Different sizes of quartz reac-
2
2
tors have been used (inner diameter, id, of 10, 16, 20,
and 30 mm) whose walls have been passivated by a
boric acid coating followed by treatment via the slow
hydrogen/oxygen reaction at ca. 773 K.
Analysis was carried out by GC/MS, GC/MS/MS [EI
and NH 9CI (chemical ionization) conditions]. After
3
micro-preparative GC separation of the complex per-
1
oxide had been achieved, H NMR spectra were also
Precise identification of the peroxide MW 118
could only be made after micropreparative separation
of this species from the other products had been per-
formed. Experimental conditions in the quartz reactor
were chosen in order to achieve a good consumption
of the di-n-pentylperoxide ROOR and to obtain a high
yield of the peroxide MW 118: initial concentration of
used in order to identify unambiguously this species.
Conventional GC/MS and GC/MS/MS spectra were
obtained with a Nermag R-30-10 (Quad Service,
Poissy, France) triple quadrupole instrument. Source
conditions were set as follows: temperature, 463 K (EI
mode), and 413 K (CI mode); filament current,
0.2 mA (EI), and 0.1 mA (CI); electron energy,
70 eV (EI), or 95 eV (CI). NH (or ND ) was used as
1
4
Ϫ3
ROOR [(4–6) ϫ 10 molecules cm ], temperature
3
3
Ϫ1
Ϫ4
of 483 K with an overall flow of 100 ml min , pas-
sivated quartz vessel id 10 mm, residence time ca.
reagent gas for CI at 10 torr measured in the source
housing. For MS-MS experiments, CAD was achieved
Ϫ2
1
0 s. Five experiments were carried out, each consist-
with Ar as collision gas at 3.5 10 torr and 5, 10 and
25 eV collision energy. Samples were introduced via
GC by split/splitless injection using a 25 m ϫ
0.32 mm DB5 MS (J & W Scientific, Fisons) silica
ing of trapping products for about 2 h. Sufficient
amounts were, thus, collected to carry out further sep-
aration.

ISOMERIZATION RATE CONSTANT
879
Figure 1 GC/MS chromatogram of the species formed during flow experiment at 463 K in a 16 mm id passivated quartz
vessel, residence time ca. 35 s, carrier gas He 99.9999%, capillary column HP-1 25 m ϫ 0.20 mm ϫ 0.50 ␮m: 1 2-methyl-
furan, 2 2,3 dihydro-5-methyl-furan, 3 tetrahydro-2-methyl-furan, 4 pentanal, 5 1-pentanol, 6 ␥-valerolactone, 7 complex
peroxide, C H O , 8 complex peroxide, C H O .
5
10
3
5
10
3
ϩ
fused capillary column at 313 K (4 min) and up to
at m/z 118 (of very low abundance, M Њ), 101, 100,
1
4
23 K (5 K/min) and He as carrier gas. H NMR spec-
85, 83, 72, 67, 57, 43 (base peak), and 29 (Fig. 2).
Ions 101, 100, and 85 could be interpreted as
[M9ЊOH] , [M9H O] Њ and [M9ЊOOH] frag-
tra were obtained with a Varian Gemini 300 MHz
equipment with C D as solvent.
ϩ
ϩ
ϩ
6
6
2
The EI spectrum obtained for 7 (or 8) shows ions
ment-ions, respectively. On the other hand, ions m/z
Figure 2 EI mass spectrum of peroxide 7 (or 8).

8
80
PERRIN ET AL.
Figure 3 NH 9CI mass spectrum of peroxide 7 (or 8).
3
ϩ
8
3, 72, and 67 could result from consecutive losses
[M ϩ ND 9HDO9ND ] , respectively, (with
D 4 3
M ϭ M ϩ 1).
D
ϩ
of neutrals giving rise [M9ЊOH9H O] ,
2
ϩ
ϩ
[M9H O928 u] Њ and [M9ЊOOH9H O] . All
MS9MS experiments were also performed under
EI conditions. The CAD daughter ion spectra obtained
from ions m/z 85 and 83 showed (among others) frag-
2
2
these ions strongly resemble the EIMS behavior of
the hydroperoxybutyraldehyde (or its cyclic form)
previously identified [5]. The [M9H O] Њ and
ϩ
ment-ions at m/z 67 (loss of H O) and at m/z 53 (cor-
2
2
ϩ
[M9ЊOOH] ions might herein also correspond spe-
responding to an elimination of HCHO), respectively.
These results and above assignments of the examined
parent ions supported the presence of three oxygens
in the original molecule thus suggesting a C H O
cifically to the hydroperoxyaldehyde form likely to be
present in the conditions of our gas-phase experiments.
To confirm the molecular weight and to provide
5
10
3
further structure evidence, the NH (CI) spectrum was
recorded (Fig. 3). This spectrum showed ions at m/z
molecular formula.
3
1
The H NMR spectrum and decoupling experi-
ments led to the following interpretation (␦ in ppm, in
C D ): 4.94 (H-3, m), 4.01–3.95 (H-6 m), 2.53, and
1
1
53 (of low abundance), 136, 118 (base peak), 102,
01, and 85. The two former could correspond to ad-
6
6
ϩ
ϩ
duct-ions [MNH , NH ] and [M ϩ NH ] whereas
2.37 (OH 2 d, J ϭ 5.5 Hz), 1.6–1.3 (4H, H-4, HЈ-4,
4
3
4
ion m/z 118 might be interpreted as [M ϩ
H-5, HЈ-5 m), 0.92, and 0.90 (H C 2 d, J ϭ 6.5 Hz).
3
ϩ
NH 9H O] due to the presence of an hydroxyl
Since two cyclic epimers, (3S*,6S*)3␣-hydroxy-6-
methyl-1,2-dioxane and (3R*,6S*)3␣-hydroxy-6-
methyl-1,2-dioxane, are found by this method, it may
be assumed that these correspond to the two forms also
observed in the gas phase by GC/MS (although in a
different ratio) and that epimerization is a slow pro-
cess.
4
2
group. Other ions of lower m/z values would then arise
from further decompositions most probably as fol-
ϩ
lows: [M ϩ NH 9H O ] for ion m/z 102 and
4
2
2
ϩ
[
M ϩ NH 9H O9NH ] for ion m/z 101. The
4 2 3
ND (CI) spectrum exhibiting ions at m/z 161
3
(
(
ϩ8 u), 141 (ϩ5), 122 (ϩ4 u), 105 (ϩ3 u), and 102
ϩ1 u) was in agreement with such assignments. The
All these data led us to conclude that the aldehyde-
hydroperoxide was transformed into its corresponding
cyclic form in the liquid sample (two epimers accord-
ing to the NMR, Fig. 4). In our system, in presence
of NO, the formation of the two cyclic epimers from
the HOROи2 radical can only occur through the in-
termediate of the pentanal-hydroperoxide O"
CH(CH ) CH(OOH)CH . Moreover, cyclization of
observed shifts not only confirmed the above interpre-
tation for the adduct-ions and thus the molecular
weight to be 118, but also proved the occurrence of
one mobile hydrogen most likely from an hydroxyl
group (cf. supplementary ϩ 1 shift due to H9D ex-
change). The adduct-ion fragments at m/z 141, 122,
ϩ
1
[
05, and 102 should be explained as [M ϩ ND ] ,
D
4
2
2
3
ϩ
ϩ
M ϩ ND 9HDO] , [M ϩ ND 9D O ] , and
hemi-ketal-like compounds is a well known and fast
D
4
D
4
2
2

ISOMERIZATION RATE CONSTANT
881
Figure 4 Equilibrium between the open form and the two cyclic epimers.
process in the liquid phase, as it appears in a review
article about peroxy natural products [15].
The isomerization reaction (1) of the n-pentoxy
CCCCCO radical generates a CC CCCOH radical
via a six-member transition state (this isomerization is
и
и
faster than that of the n-butoxy radical). This
FORMATION MECHANISM FOR THE
ALDEHYDE-HYDROPEROXIDE
CC CCCOH radical further reacts by two parallel
и
pathways: one by O addition (2) which is stabilized,
2
and leads to the following hydroxy-peroxy radical
Previous work on development of a mechanism [3,5]
for formation of the butanal-hydroperoxide and of bu-
tanal serves as a guideline to develop the mechanism
accounting for reactions occurring in the (n-pentoxy/
O ) system. In the two systems, (n-butoxy/O ) and (n-
CH CH(O9O )CH CH CH OH and the other
и
3
2 2 2
through reaction (5) leading to the CCCCC OH rad-
и
ical, which after O addition (6) produces pentanal and
2
HO2 .
2
2
This formation of pentanal is supported by the fact
that the pentanal yield is independent of oxygen for a
very wide range of O concentration (5–100% O ).
pentoxy/O ), observed products are similar, i.e., the
2
same reactions form the following sets of correspond-
ing products: (butanal/pentanal, furans/methyl-furans,
butyrolactone/␥-valerolactone, butanal-hydroperox-
ide/pentanal-hydroperoxide, etc.). The reaction paths
to explain the formation of pentanal-hydroperoxide
and pentanal are in the following mechanism I (only
reactions of H atoms playing a significant role are writ-
ten):
2
2
Therefore, production of pentanal through the direct
reaction RO ϩ O can be practically ruled out.
2
Potential further reaction of CH CH(OO )
и
3
CH CH CH OH radical can be an internal H-shift
2
2
2
isomerization, either via a seven-member transition
state, reaction (3), this isomerization being favored by
the C9O bond [3,4] (of the CH OH group), or via a
2
six-member transition state according to reaction (X):
a
b
1
2
3
CCCCCO
и
CC
иCCCOH
CCCCCOH
O2
X
XЈ
CCCCCOH
O9O
CCCC COH
и
OO
и
O2
5
Ϫ5
g
C
и
O9OH
4
CCCCC
и
OH
CCCCC"O ϩ HO2
O2
XЉ
CCCCCOH
CCC9C9C ϩ HO2
NO
7
OOH
OOH
O
O9O
O9OH
и
6
CCCCdC
и
OH
CCCCC"O ϩ HO2
O9OH
O2
pentanal
hydroperoxypentanal
A
Reactions (X), (XЈ), and (XЉ) result in an epoxide-
hydroperoxide, a species we searched for, but that we
were unable to observe. Only the pentanal-hydrope-
roxide (formed by reaction 4) was identified. The sec-
CCCCCOH ϩ NO2
products
Oи

8
82
PERRIN ET AL.
Table I Ratios (C/A) and C/A) ϫ [NO] vs. Temperature T (K) and Concentration of NO
Temperature T (K)
residence time (s)
463
41.2
7.25
2.521
18.276
463
41.2
10.0
1.916
19.164
463
463
41.2
12.0
1.613
19.355
463
463
41.2
13.8
1.418
19.574
463
41.2
16.0
1.235
19.753
1
4
Ϫ3
[
(
(
NO] 10 cm
C/A)
C/A) ϫ [NO]
Temperature T (K)
463
residence time (s)
41.2
1.7
3.861
6.564
473
41.5
1.60
4.000
6.400
41.2
3.5
2.274
7.960
473
41.5
2.30
3.101
7.132
41.2
3.5
2.440
8.541
473
41.5
2.30
3.404
7.828
1
4
Ϫ3
[
(
(
NO] 10 cm
C/A)
C/A) ϫ [NO]
Temperature T (K)
residence time (s)
473
41.5
3.0
473
41.5
6.1
473
41.5
9.2
1
4
Ϫ3
[
(
(
NO] 10 cm
C/A)
2.941
8.824
473
1.818
11.091
473
1.111
10.222
C/A) ϫ [NO]
Temperature T (K)
473
473
473
residence time (s)
6.5
0
3.373
0
6.5
0
3.521
0
6.5
0
3.186
0
6.5
6.5
1
4
Ϫ3
[
(
(
NO] 10 cm
C/A)
C/A) ϫ [NO]
1.90
1.685
3.202
2.90
1.299
3.767
Temperature T (K)
483
483
483
483
residence time (s)
5.1
0
4.348
0
5.1
5.1
5.1
3.00
1.770
5.310
1
4
Ϫ3
[
(
(
NO] 10 cm
C/A)
C/A) ϫ [NO]
0.75
3.125
2.344
1.50
2.468
3.702
Temperature T (K)
493
493
493
493
493
residence time (s)
5.0
2.95
3.448
10.172
498
5.0
4.45
2.564
11.410
498
5.0
5.9
2.342
13.817
498
5.0
7.05
2.062
14.536
498
5.0
10.2
1.587
16.190
498
1
4
Ϫ3
[
(
(
NO] 10 cm
C/A)
C/A) ϫ [NO]
Temperature T (K)
498
residence time (s)
4.9
0
2.439
0
4.9
1.0
1.961
1.961
4.9
2.9
1.429
4.143
4.9
5.8
0.972
5.639
4.9
5.8
0.933
5.413
4.9
7.0
0.855
5.983
1
4
Ϫ3
[
(
(
NO] 10 cm
C/A)
C/A) ϫ [NO]
Temperature T (K)
residence time (s)
503
3.25
0
2.833
0
503
3.25
2.0
1.923
3.847
503
3.25
5.0
1.282
6.410
503
3.25
10.0
0.870
8.696
503
3.25
16.0
0.612
9.798
1
4
Ϫ3
[
(
(
NO] 10 cm
C/A)
C/A) ϫ [NO]
Temperature T (K)
518
518
518
518
518
residence time (s)
2.0
2.0
3
1.176
3.529
523
2.0
6.1
0.971
5.922
523
2.0
9.2
0.833
7.667
523
2.0
12.5
0.669
8.361
523
1
4
Ϫ3
[
(
(
NO] 10 cm
C/A)
C/A) ϫ [NO]
1.38
1.504
2.075
Temperature T (K)
523
523
residence time (s)
1.95
0.59
2.358
1.392
1.95
0.80
2.212
1.770
1.95
3.05
1.754
5.351
1.95
4.25
1.681
7.143
1.95
6.05
1.538
9.308
1.95
7.05
1.418
10.00
1
4
Ϫ3
[
(
(
NO] 10 cm
C/A)
C/A) ϫ [NO]
C/A) is the ratio (oxo-hydroperoxide/pentanal); [NO] concentration in 1014 molecules cm 3
.
Ϫ
(

ISOMERIZATION RATE CONSTANT
883
ond isomerization reaction occurs via the described
seven-member transition state (reaction 3) and absence
of XЉ suggests the six-member transition state (reac-
tion X) can be ruled out. Experimental identification
of the peroxide chemical structure enables positive de-
termination of the rate constant of isomerization.
a set of experiments was performed at a given tem-
perature with ratio O /N 1 : 4 and fixed flows of O
2
2
2
and N (giving a fixed residence time and a constant
2
extent of reaction) over several days. Passivation of
the chosen quartz reactor had been performed by slow
reaction H /O before each experimental set was be-
2
2
gun. Small amounts of NO, ranging from 0.59 ϫ
1
4
15
Ϫ3
1
0 to 1.6 ϫ 10 molecules cm , were added to the
DETERMINATION OF THE
di-n-pentyl peroxide/O /N mixtures ([NO] was the
2
2
ISOMERIZATION RATE CONSTANT OF
unique variable). From the GC/MS chromatogram cor-
responding to given NO concentration, areas of the
pentanal and oxo-hydroperoxide (pentanal-hydrope-
roxide) peaks were measured, and the ratio oxo-hy-
droperoxide/pentanal calculated. Pentanal is formed
by reactions (5) and (6) and is observed to be inde-
THE CH CH(O9Oи)CH CH CH OH
3
2
2
2
RADICAL
Determination of the isomerization rate constant of the
CH CH(O9O )CH CH CH OH radical has been
и
3
2
2
2
performed by the relative rate technique: by compar-
pendent of oxygen level (5–100% O ); this observa-
2
ison to the competing well-known reaction RO ϩ
tion is similar to the published independence between
butanal [3] and oxygen. We observed also that the pen-
tanal yield is independent of NO, a result which is in
agreement with the above mechanism I, whereas, the
oxo-hydroperoxide yield decreases with NO concen-
tration.
2
NO : RO ϩ NO (7). This peroxy radical is either
2
consumed by isomerization reaction (3) leading, after
reaction (4), to the pentanal-hydroperoxide plus for-
mation of HO , or by reaction (7) with NO, generating
2
an other alkoxy radical which does not lead to the
formation of pentanal-hydroperoxide.
We assign the following symbols to the interme-
diates and end products (mechanism I):
[
CC
и
CCCOH] ϭ ␣,
OH] ϭ ␥,
[CC(OO
и
)CCCOH] ϭ ␤,
OH] ϭ ␦;
Experiments
[CC(OOH)CCC
и
[CCCCC
и
Determination of the isomerization rate constant was
made by using two quartz reactors, one of id. 20 mm
and volume 120 ml, the other of id. 10 mm and vol-
ume 15 ml (the residence time for each experiment is
specified in Table I). The experimental protocol was:
[CC(OOH)CCC"O] ϭ C] (the oxo-hydroperoxide)
and [CCCCC"O] ϭ A (aldehyde, i.e., pentanal). The
formation rates of pentanal, the oxo-hydroperoxide,
and the intermediates ␣, ␤, ␥, and ␦ follow the rela-
tions:
dA
dt
dC
dt
d␦
ϭ k O ␦
ϭ k O ␥
ϭ k ␣ Ϫ k ␦ Ϫ k O ␦
5 5 6 2
6
2
4
2
Ϫ
dt
d␥
d␤
ϭ k ␤ Ϫ k O ␥
ϭ k O ␣ Ϫ k ␤ Ϫ k [NO]␤
3
4
2
2
2
3
7
dt
dt
Applying the quasi-stationary state to intermediates ␤, ␥, and ␦, we find:
k O ␣
k k O ␣
k5␣
kϪ5 ϩ k O
2
2
2
2
3
2
␤
ϭ
;
␥ ϭ
;
␦ ϭ
;
k ϩ k NO
k O (k ϩ k NO)
3
7
4
2
3
7
6
dA
dt
␣k k O
2
dC
dt
6
5
ϭ k O ␦ ϭ
;
and
ϭ k O ␥ ϭ k ␤
4 2 3
6
2
kϪ5 ϩ k O
6
2
To express the formation rates of the peroxide and the
aldehyde from the same radical intermediate ␣, we
perform the integration of dC and dA from time 0 to
time t, leading to the ratio (oxo-hydroperoxide/pen-
tanal):
t
t
t
␣
k k O
k k O
2 3 2
2
3
2
͵
dC
dA
͵
dt
dt
͵
␣dt
␣dt
0
0
k ϩ k NO
k ϩ k NO
0
C
A
k k (k ϩ k O ) oxo Ϫ hydroperoxide
2 3 5 6 2
3
7
3
7
Ϫ
ϭ
ϭ
;
ϭ
ϭ
t
t
t
␣k k O
k k O
k k (k ϩ k NO)
pentanal
5
6
2
5
6
2
5
6
3
7
͵
͵
͵
0
0
kϪ5 ϩ k O
kϪ5 ϩ k O
2
0
6
2
6

8
84
PERRIN ET AL.
If we express the ratios (C /A ) corresponding to a first
1
1
concentration of NO, [NO] , and (C /A ) correspond-
1
2
2
ing to an other concentration of NO, [NO] , the fol-
2
lowing equalities can be established, leading to rela-
tion (I):
C1
A1
k k (k ϩ k O )
2
3
Ϫ5
6
2
(
k ϩ k [NO] ) ϭ
3 7 1
k k
5
6
C2
A2
ϭ
(k ϩ k [NO] )
3
7
2
C
C1
2
[
NO] Ϫ
[NO]1
ͩA
2
ͪ
2
A1
k3 ϭ Ϫk7
(I)
C2 C1
Ϫ
ͩ
ͪ
Figure
5 Plot of (C/A) ϫ [NO] vs. C/A, at
A2
A1
14
Ϫ3
5
03 K; [NO] unit ϭ 10 molecules cm .
According to eq. (I), the rate constant k is calculated
3
at each temperature T knowing k , and the slope of the
7
2
correlation factor r ϭ 0.988 and errors ϭ 1␴, of Fig-
ure 6 was obtained from the data of Table II. The best
fit to our data is a straight line of equation y ϭ
line (C/A)[NO] vs. (C/A). An example is given on
Figure 5, at 503 K, showing how this calculation is
performed: five ratios (C/A) were measured for reac-
tion without NO and four NO concentrations ranging
8
519.5 x ϩ 24.883; therefore activation energy E
and A factor are: Ea(3) ϭ (16.9 Ϯ 0.7) kcal mol and
A3 ϭ (6.4 Ϯ 0.6) ϫ 1010 s .
The isomerization rate constant of reaction
HOCH CH CH CH(OO
CH(OOH)CH (4), involving the H atoms of the
CH 9OH group is:
a
Ϫ1
1
4
14
Ϫ3
from 2.0 ϫ 10 to 16.0 ϫ 10 molecules cm (the
Ϫ1
same reactor and the same O and N flow conditions
2
2
were used three days, passivation conditions of the
reactor remained unchanged). The measured ratios (C/
A) for each [NO] (see Table I) gave the plot of
и
)CH3 : HOC HCH CH2
и
2
2
2
2
3
2
(
C/A) ϫ [NO] vs. C/A, with slope of the straight line
1
4
Ϫ3
of 4.4107 ϫ 10 molecules cm , at 503 K. Using
k3 ϭ (6.4 Ϯ 0.6) ϫ 1010 exp{Ϫ(16,900
for k the recommended value of Atkinson et al.
7
Ϫ1
Ϫ1
Ϯ 700) cal mol /RT}s
[
4
16], the same value for all RO radicals, k ϭ
2 7
Ϫ12
3
Ϫ1
Ϫ1
.2 ϫ 10
exp (180/T) cm molecule
s
The isomerization rate constant for 1 H atom is there-
fore:
Ϫ12
3
Ϫ1
Ϫ1
(k7(503K) ϭ 6.007 ϫ 10 cm molecule s ), k at
3
Ϫ1
5
03 K is found k3(503 K) ϭ 2649.5 s . This procedure
was used at eight temperatures 463, 473, 483, 493,
98, 503, 518, and 523 K (two series of measure-
ments, each series differing by an interval of about one
month, were made at 463 and 473 K), and gave ten
k3(H) ϭ (3.2 Ϯ 0.3) ϫ 1010 exp{Ϫ(16,900
4
Ϫ1
Ϫ1
Ϯ 700) cal mol /RT}s .
coupled values (k , T) which are summarized in Table
II.
DISCUSSION
3
This isomerization rate constant k (which is pro-
A parallel can be established between reaction
OOCCCOOH : HOOCCC OOH, from estimations
for the reaction of hydroperoxy-propyl [4] radicals
3
portional to k ) would change if the literature recom-
и
и
7
mended value of k were modified. The Arrhenius plot,
7
Table II Isomerization Rate Constant k at Temperature T (K)
3
T (K)
k (s )
3
463
700.0
463
610.3
473
948.1
473
1127.3
483
1248.4
Ϫ
1
1
ln k3
T (K)
/T
0.002160
6.55106
493
2029.3
0.002028
7.61543
0.002160
6.41391
498
2227.2
0.002008
7.70852
0.002114
6.85447
503
2649.5
0.001988
7.88214
0.002114
7.02758
518
4824.0
0.001931
8.48135
0.00207
7.12959
523
5682.2
0.001912
8.64510
Ϫ1
k (s )
3
1
/T
ln k3

ISOMERIZATION RATE CONSTANT
885
Figure 6 Arrhenius plot of the isomerization rate constant ln (k ) vs. 1/T (K). The experimental data are shown by the
3
triangles, and the straight line is the best fit to these data.
with O (A and E were estimated to be 1.14 ϫ
tochemical decomposition of these aldehyde-hydro-
peroxides (more generally oxo-hydroperoxides or car-
bonyl-hydroperoxides).
Thermochemical estimations of Bozzelli et al. [4]
might be extended to the HOROO radical resulting
from the s-pentoxy radical. They report that isomeri-
zation reaction HOROO : HROϪHOOH is enhanced
relative to ROOH by the presence of the C9O bond
2
Ϫ1
a
10
1
0
s
and 16.80 kcal/mole, respectively), and the
isomerization
HOCCCC(OO
reaction
)C : HOC
we
investigated
и
Ϫ1
и
CCC(OOH)C (A(H)
ϭ
3
7
.2 ϫ 1010 s , and E ϭ 16.9 kcal/mole). ⌬H ϭ
a
.5 kcal/mole and Eabstraction ϭ 9.2 kcal/mole are in-
deed the same since in both cases the C9H primary
bonds are weakened because of the presence of the
C9O bond. Bozzelli et al. [4] used 0.1 kcal/mole for
the ring strain of a six-member ring, whereas we find
a low strain energy barrier of ca. 0.2 kcal/mole for a
seven-member ring, practically equivalent to that of a
six-member ring. This value is consistent with previ-
ous investigations. Twenty years ago, formation of
significant yields of O-heterocycles from the oxidation
(of the CH OH group) which weakens the RC(OH)
2
9H bonds from their usual primary C9H bond en-
ergy value. Abstraction of an H atom from a carbon
bearing a C9O bond, by an isomerization reaction
occurring through a five-membered ring transition
state,
for
example
the
isomerization
CC(OOH)COO
comparable to the second isomerization reaction of the
secondary pentoxy radical in this study:
и
: CC (OOH)COOH would be
и
of (ϾC ) alkanes had been explained by the predom-
5
inant chain cycle involving initial attack at a secondary
H, followed by addition of O , 1 : 6-H transfer from
2
another secondary H (this is a seven-member ring
CC(OH)CCCOO
и
!: CC (OH)CCCOOH (i)
и
[17]) and decomposition of the ␥-hydroperoxyalkyl
radical by simple cyclization and loss of OH. Recently,
observation was made that isomerization reactions
involving 7-membered ring transition state or
There is a major difference, however: the activation
energy of this isomerization reaction (i) would be
14.3 kcal/mole instead of 20.1 kcal/mole, because the
ring strain of 6 kcal/mole (five-member ring) has to
be replaced by 0.2 (seven-member ring in our case).
With an activation energy Ea(i) Ϸ 14.3 kcal/mole, and
6
-membered ring have low or zero strain energy bar-
riers [18].
It is worthwhile noticing that the isomerization rate
constant of reaction n-OOC H OH : HOOC H OH,
4
8
4
7
1
0
that we had previously [3] measured at the temperature
assuming the same A factor A(H) ϭ 3.2 ϫ 10 in
agreement with our determination, the rate constant of
reaction (i) at room temperature would be ca. 70 times
3
Ϫ1
of 487 K, k(487 K) ϭ 1.5 ϫ 10 s , is in agreement
with the present work. It can, therefore, be inferred
from these results that conclusions which were drawn
about the role of these isomerization reactions on tro-
pospheric ozone [3,5] remain valid. At low NO con-
centrations, as prevail in rural areas, tropospheric
ozone concentration was observed to be very high un-
der summer anticyclonic conditions: this can be due
to multiplication of radicals formed by thermal or pho-
1
0
higher than the value [A(H) ϭ 3.2 ϫ 10 , E 16.9
a
kcal/mole] determined here. As a consequence, this
isomerization reaction (of HORO radicals (ՆC )
2
5
resulting from s-alkoxy radicals) would become
important in atmospheric chemistry. It can be ex-
pected that the, by analogy assumed reaction with
O (iЈ):
2

8
86
PERRIN ET AL.
CC
и
CCCOOH
ϩ O !: CC(O)CCCOOH ϩ HO (iЈ)
H
(IV)
2
2
H H9C9H OH
H
would form, as main product, a ketone-hydroperoxide
and not an aldehyde-hydroperoxide as it occurs with
n-RO radicals. This carbonyl-hydroperoxide, 2-pen-
tanone-hydroperoxide, should act as a much stronger
source of radicals; it will increase atmospheric pollu-
tion, and result in increased ozone relative to primary
species.
We can extend thermochemical estimates of Boz-
zelli et al. to our s-pentoxy system, show good agree-
ment and make important predictions for reactions of
larger alkanes. Taking into account the above discus-
sion of isomerization reactions involving a secondary
H9C
H H9C9H
O9OH
C
C
C9H
и
H
(ii) a branched hexane where a secondary H is ab-
stracted by OH, reacts with O , gives an RO radical
2
2
which is transformed into an RO radical through re-
action with NO of the atmosphere.
H
carbon bearing a C9O bond, the use of (ՆC ) al-
4
kanes, not totally branched, should be strictly limited
during summer time. We consider the following have
strong effects in atmospheric chemistry: (i) branched
alcohols; (ii) branched hydrocarbons (alkanes); and
H H9C9H H
H9C
H
OH
C
C
C9H
H H9C9H H
H
(iii) higher aldehydes.
(i) a branched hexanol, where an H atom can be
H
abstracted from a secondary site weakened by the pres-
ence of the C9O bond from the alcohol, (where a
primary H atom has been abstracted by an OH radical)
for which isomerization occurs through a six-member
ring.
H
H H9C9H H
H9C
H
C
C
и
C9H
H H9C9H
H
H
H
(I)
O2
H H9C9H OH
H
OH
H9C
C
C
C9H
H
H H9C9H H
H
H H9C9H H
H
NO
H
H9C
C
C
C9H
H
(II)
H H9C9H O
H
H H9C9H OH
H9C
H
H
Oи
H
C
C
C9H
H H9C9H H
H9C
H
H H9C9H H
H
и
C
C
C9H
O2
H H9C9H O
и
H
H
(III)
H
H H9C9H OH
H9C
H H9C9H H
O9O
H
isomerization
This RO (hexoxy) radical has a configuration equiv-
alent to that of previous case (i II). It can subsequently
isomerize, according to configurations (i III) and (i
IV), and leads to a complex hydroperoxide.
C
C
C9H
H
и
(iii) in our last example, (ՆC ) aldehydes are in-
3

ISOMERIZATION RATE CONSTANT
887
volved and propionaldehyde is chosen. One of the pri-
mary H atoms is abstracted by OH, addition of O2
occurs and finally an H-shift isomerization reaction
through a six-member ring:
OH
O
2
CH 9CH 9CH"O 9
9
:
и
CH 9CH 9CH"O 9
9
: OOCH 9CH 9CH"O
и
3
2
2
2
2 2
isom.
и
OOCH 9CH 9CH"O 9
9
: HOOCH 9CH 9C "O
и
2
2
2
2
These examples illustrate the extension of this type of
isomerization reaction, which must be taken into ac-
count in order to understand low temperature com-
bustion and atmospheric chemistry of hydrocarbons
and, more generally, of volatile organic compounds.
BIBLIOGRAPHY
1. N. Blin-Simiand, R. Rigny, V. Viossat, S. Circan, and
K. Sahetchian, Combust. Sci. and Technol., 88, 329
(1993).
2
. W. P. L. Carter, K. R. Darnall, A. C. Lloyd, A. M.
Winer, and J. M. Winer, and J. N. Pitts, Chem. Phys.
Letters, 42, 22 (1976).
CONCLUSION
3
4
. A. Heiss and K. Sahetchian, Int. J. Chem. Kinet., 28,
5
31 (1996).
. E. R. Ritter and J. W. Bozzelli, Int. J. Chem. Kinet., 23,
67 (1991); J. W. Bozzelli and W. J. Pitz, Twenty-fifth
The rate constant of an isomerization reaction, of im-
portance in atmospheric pollution and in low temper-
ature combustion processes,
7
Symposium (International) on Combustion, The Com-
bustion Institute, Pittsburgh, Pennsylvania, 1994, p.
HOCCCC(OO
и
)C !: HOC
и
CCC(OOH)C
7
83.
5
. F. Jorand, A. Heiss, K. Sahetchian, L. Kerhoas, and J.
Einhorn, J. Chem. Soc., Faraday Trans., 92, 4167
has been determined for the first time, in the temper-
1
0
ature range 463–523 K, [A(H) ϭ 3.2 ϫ 10 , E ϭ
a
(1996).
1
6.9 kcal/mole]. This isomerization reaction leads to
6
7
. S. W. Benson, Thermochemical Kinetics, 2nd ed., Wi-
ley, New York, 1976.
. A. C. Baldwin, J. R. Barker, D. M. Golden, and D. G.
Hendry, J. Phys. Chem., 81, 2483 (1977).
the pentanal-hydroperoxide O"CCCC(OOH)C in
the gas phase, which is in equilibrium with two cyclic
epimers, (3S*,6S*)3␣-hydroxy-6-methyl-1,2-dioxane
and (3R*,6S*)3␣-hydroxy-6-methyl-1,2-dioxane, in
the liquid phase.
The importance of oxo-hydroperoxides (also des-
ignated in the literature as alkyl-carbonyl-hydroperox-
ides) in low-temperature chemistry of hydrocarbon ox-
idation and in atmospheric chemistry is due to their
propensity (capacity) to easily decompose, thermally
or photochemically, and generate radicals. This mul-
tiplication of radicals contributes under some atmo-
spheric conditions to the increase of secondary pol-
lutants, especially tropospheric ozone.
8. L. Batt, Int. Rev. Phys. Chem., 6, 53 (1987).
9. R. Bounaceur, V. Warth, P. A. Claude, F. Battin-Le-
clerc, G. Scacchi, G. M. C oˆ me, T. Faravalli, E. Ranzi,
J. Chim. Phys., 93, 1472 (1996).
1
1
1
0. E. Ranzi, T. Faravelli, P. Gaffuri, and A. Sogaro, Com-
bust. Flame, 102, 179 (1995).
1. R. R. Baldwin, M. W. M. Hisham and R. W. Walker,
J. Chem. Soc., Faraday Trans., 78, 1615 (1982).
2. J. Hughes, P. A. Halford-Maw, P. D. Lightfoot, T. Tur-
anyi and M. J. Pilling, Twenty-fourth Symposium (In-
ternational) on Combustion, The Combustion Institute,
Pittsburgh, Pennsylvania, p. 645, 1992.
There is need for experimental determinations of
13. T. P. W. Jungkamp, J. N. Smith, and J. H. Seinfeld, J.
Phys. Chem. A, 101, 4392 (1997).
the isomerization rate constant of RO radicals for
2
1
4. F. Welch, H. R. Williams and M. S. Mosher, J. Am.
Chem. Soc., 77, 551 (1954).
which the abstraction of an H atom would occur from
a secondary carbon bearing a C9O bond. The anal-
ysis, made in the discussion, points out the absolute
necessity to avoid emissions into the atmosphere of
1
1
5. D. A. Casteel, Nat. Prod. Rep., 9, 289–312 (1992).
6. R. Atkinson, S. M. Aschmann, and A. M. Winer, J.
Atmos. Chem., 91, 5 (1987).
(
ՆC ) hydrocarbons bearing an aldehydic function,
3
1
7. R. T. Pollard, in Comprehensive Chemical Kinetics, (C.
F. H. Tipper and C. J. Bamford, Eds), Elsevier, Am-
sterdam, Vol. 17, p. 249, 1977.
and of (ՆC ) hydrocarbons with carbon chains con-
4
taining secondary H atoms. Short chains (ϽC ) hy-
4
drocarbons and totally branched hydrocarbons are
much less harmful to Environment.
1
8. J. S. Cowart, J. C. Keck, J. B. Heywood, C. K. West-
brook and W. J. Pitz, Twenty-third Symposium (Inter-
national) on Combustion, The Combustion Institute,
Pittsburgh, Pennsylvania, p. 1055, 1990.
Helpful discussions with Dr. E. M. Evleth are gratefully ac-
knowledged. This work was carried out in the frame of the
EUROTRAC-LACTOZ Project.