FT-IR kinetic and product study of the OH radical and Cl-atom - Initiated oxidation of dibasic esters

Article abstract of DOI:10.1002/kin.1039

Using a relative kinetic technique, rate coefficients have been measured, at 296 ± 2 K and 740 Torr total pressure of synthetic air, for the gas-phase reaction of OH radicals with the dibasic esters dimethyl succinate [CH₃OC(O)CH₂CH₂C(O)OCH₃], dimethyl glutarate [CH₃OC(O)CH₂CH₂C(O)OCH₃], and dimethyl adipate [CH₃OC(O)CH₂CH₂CH₂CH₂ C(O)OCH₃]. The rate coefficients obtained were (in units of cm³ molecule^-1 s^-1): dimethyl succinate (1.89 ± 0.26) × 10^-12; dimethyl glutarate (2.13 ± 0.28) × 10^-12; and dimethyl adipate (3.64 ± 0.66) × 10^-12. Rate coefficients have been also measured for the reaction of chlorine atoms with the three dibasic esters; the rate coefficients obtained were (in units of cm³ molecule^-1 s^-1): dimethyl succinate (6.79 ± 0.93) × 10^-12; dimethyl glutarate (1.90 ± 0.33) × 10^-11 and dimethyl adipate (6.08 ± 0.86) × 10^-11. Dibasic esters are industrial solvents, and their increased use will lead to their possible release into the atmosphere, where they may contribute to the formation of photochemical air pollution in urban and regional areas. Consequently, the products formed from the oxidation of dimethyl succinate have been investigated in a 405-L Pyrex glass reactor using Cl-atom-initiated oxidation as a surrogate for the OH radical. The products observed using in situ Fourier transform infrared (FT-IR) absorption spectroscopy and their fractional molar yields were: succinic formic anhydride (0.341 ± 0.068), monomethyl succinate (0.447 ± 0.111), carbon monoxide (0.307 ± 0.061), dimethyl oxaloacetate (0.176 ± 0.044), and methoxy formylperoxynitrate (0.032-0.084). These products account for 82.4 ± 16.4% C of the total reaction products. Although there are large uncertainties in the quantification of monomethyl succinate and dimethyl oxaloacetate, the product study allows the elucidation of an oxidation mechanism for dimethyl succinate.

Full text of DOI:10.1002/kin.1039

FT-IR Kinetic and Product
Study of the OH Radical
and Cl-Atom–Initiated
Oxidation of Dibasic Esters
FABRIZIA CAVALLI, IAN BARNES, KARL HEINZ BECKER
Bergische Universita¨t–GH Wuppertal, Physikalische Chemie/FB 9, Gauss Strasse 20, D-42097 Wuppertal, Germany
Received 27 July 2000; accepted 28 March 2001
ABSTRACT: Using a relative kinetic technique, rate coefficients have been measured, at 296 Ϯ
2 K and 740 Torr total pressure of synthetic air, for the gas-phase reaction of OH radicals with
the dibasic esters dimethyl succinate [CH3OC(O)CH2CH2C(O)OCH3], dimethyl glutarate
[CH3OC(O)CH2CH2CH2C(O)OCH3], and dimethyl adipate [CH3OC(O)CH2CH2CH2CH2C(O)OCH3].
1
The rate coefficients obtained were (in units of cm3 moleculeϪ sϪ1): dimethyl succinate
(1.89 Ϯ 0.26) ϫ 10Ϫ12; dimethyl glutarate (2.13 Ϯ 0.28) ϫ 10Ϫ12; and dimethyl adipate (3.64 Ϯ
0.66) ϫ 10Ϫ12. Rate coefficients have been also measured for the reaction of chlorine atoms
1
with the three dibasic esters; the rate coefficients obtained were (in units of cm3 moleculeϪ
sϪ1): dimethyl succinate (6.79 Ϯ 0.93) ϫ 10Ϫ12; dimethyl glutarate (1.90 Ϯ 0.33) ϫ 10Ϫ11; and
dimethyl adipate (6.08 Ϯ 0.86) ϫ 10Ϫ11. Dibasic esters are industrial solvents, and their in-
creased use will lead to their possible release into the atmosphere, where they may contribute
to the formation of photochemical air pollution in urban and regional areas. Consequently,
the products formed from the oxidation of dimethyl succinate have been investigated in a
405-L Pyrex glass reactor using Cl-atom–initiated oxidation as a surrogate for the OH radical.
The products observed using in situ Fourier transform infrared (FT-IR) absorption spectroscopy
and their fractional molar yields were: succinic formic anhydride (0.341 Ϯ 0.068), monomethyl
succinate (0.447 Ϯ 0.111), carbon monoxide (0.307 Ϯ 0.061), dimethyl oxaloacetate (0.176 Ϯ
0.044), and methoxy formylperoxynitrate (0.032–0.084). These products account for 82.4 Ϯ
16.4% C of the total reaction products. Although there are large uncertainties in the quanti-
fication of monomethyl succinate and dimethyl oxaloacetate, the product study allows the
elucidation of an oxidation mechanism for dimethyl succinate. ᭧ 2001 John Wiley & Sons, Inc. Int
J Chem Kinet 33: 431–439, 2001
INTRODUCTION
vironmentally compatible solvents; in particular, com-
pounds that will help to reduce the level of the oxidant
formation in the troposphere. Dibasic esters are al-
ready employed in a wide variety of applications
(coating solvents, paint stripper, plasticizers, polymer
intermediates, and resin cleanup), and there could be
a significant increase in their usage in the near future.
Therefore, because of their possible release into the
atmosphere, an evaluation of the atmospheric fate of
Dibasic esters are currently under consideration as po-
tentially attractive alternatives to established solvents
in attempts to introduce low health risk and more en-
Correspondence to: I. Barnes (barnes@physchem.uni-
wuppertal.de)
᭧ 2001 John Wiley & Sons, Inc.

432
CAVALLI ET AL.
these substances is necessary. To this end, detailed ki-
netic data and atmospheric reaction mechanistic infor-
mation are essential in order to be in a position to
assess their contribution to the formation of ozone and
other photooxidants. To improve our knowledge of the
atmospheric chemistry of esters, we report here the
results of a kinetic study of the OH radical and chlo-
rine-atom–initiated oxidation of dimethyl succinate
[CH3OC(O)CH2CH2C(O)OCH3 DBE-4], dimethyl
glutarate [CH3OC(O)CH2CH2CH2C(O)OCH3 BDE-
5], and dimethyl adipate [CH3OC(O)CH2CH2CH2CH2
C(O)OCH3 DBE-6]. Furthermore, we have investi-
gated the products of the reaction of chlorine atoms
with DBE-4 in the presence of NO.
The major atmospheric sink for the dibasic esters
is expected to be reaction with OH radicals, which is,
however, relatively slow. This makes photoreactor
chamber studies of the products difficult because of
the long-time coefficients of the experiments required
to convert sufficient reactant amounts to the concen-
trations of products necessary for an accurate analysis.
Thus, reaction with chlorine atoms was employed as
a surrogate to mimic the OH-radical–induced atmo-
spheric oxidation of dimethyl succinate. Although re-
action of VOCs with Cl atoms is somewhat less selec-
tive as with OH radicals, the subsequent chemistry is
the same and thus it is a convenient method to emulate
OH-radical–initiated chemistry.
the evacuated chamber. Quantitative reference spectra
of formaldehyde, methoxy formylperoxynitrate, and
carbon monoxide were taken from a calibrated infra-
red spectra data bank archived by this laboratory.
Chemicals
Succinic formic anhydride was synthesized following
a procedure similar to that described in the literature
for acetic formic anhydride [2]. Briefly, methyl suc-
cinyl chloride (Fluka, 98%) was added to finely
ground sodium formate (Aldrich, 99%) while the tem-
perature was maintained at 23–27ЊC with a cooling
bath. The mixture was then filtered under suction and
the solid residue rinsed with diethylether. After re-
moval of the solvent, succinic formic anhydride, a
slightly yellow and viscous liquid, was obtained. The
compound had a purity of 90% as determined by
NMR.
Dimethyl oxaloacetate (2-oxo-succinic acid di-
methyl ester, oxalic acid dimethyl ester) was synthe-
sized according to the following procedure: to a stirred
solution of oxalacetic acid (Aldrich, 98%) dissolved
in dry methanol, concentrated sulfuric acid was added
dropwise. The mixture was heated for a few hours un-
der reflux. The solution was allowed to cool to room
temperature, evaporated under vacuum, and redissol-
ved in ethyl acetate. The organic layer was extracted
twice with water, using a saturated aqueous solution
of NaCl, and dried with anhydrous Na2SO4. The re-
moval of the solvent afforded the product, dimethyl
oxaloacetate, pure to such an extent that further puri-
fication was not necessary.
EXPERIMENTAL
All the experiments were carried out in a 405-L Pyrex
cylindrical glass reactor with Teflon-coated metal-end
flanges, which has been previously described in detail
[1]. A White mirror system mounted inside the reactor
and coupled with an external mirror system to a Fou-
rier Transform-Spectrometer (Nicolet Magna 550) en-
ables the in situ monitoring of both reactants and prod-
ucts by long-path infrared absorption spectroscopy
(total path length 50.4 m). The reactor is equipped with
18 fluorescent lamps (Philips TLA 40W/05, 300 Յ ␭
Dimethyl succinate DBE-4 (98% GC), dimethyl
glutarate DBE-5 (98% GC), dimethyl adipate DBE-6
(99ϩ% GC), monomethyl succinate (95%), n-butane
2.5 (Ն99.5%), ethanol (Ն99.8%), 1-propanol
(99.5ϩ%), and cyclohexane (99.9ϩ%) were pur-
chased from Aldrich; methyl nitrite was prepared and
stored as described by Taylor et al. [3]; Cl2 (Ͼ99.8%),
NO (Յ99.95%), and synthetic air were supplied by
Messer Griesheim.
Յ 450 nm; ␭ ϭ 365 nm) arranged concentrically
max
around the outside of the chamber and with 3 low-
pressure mercury vapor lamps (Philips TUV 40W;
RESULTS AND DISCUSSION
␭
max ϭ 254 nm) contained inside an inner quartz glass
Rate Coefficients for the Gas-Phase
Reactions of Dibasic Esters with OH
Radicals and Cl Atoms
tube mounted between the centers of the end flanges.
Calibration of dimethyl succinate [CH3OC
(O)CH2CH2C(O)OCH3 DBE-4], succinic formic an-
hydride [CH3OC(O)CH2CH2C(O)OC(O)H], mono-
methyl succinate [CH3OC(O)CH2CH2C(O)OH], and
dimethyl oxaloacetate [CH3OC(O)C(O)CH2C(O)
OCH3] was achieved by expanding known amounts of
these compounds, dissolved in CH2Cl2 or CHCl3, into
Control experiments on each of the three dibasic esters
were performed (1) in the dark and (2) at maximum
light intensity for up to 40 min to check for losses of
reactants via wall adsorption and photolysis, respec-
tively. There were no differences between the loss

THE OH RADICAL AND Cl-ATOM–INITIATED OXIDATION OF DIBASIC ESTERS
433
Reaction (1) was measured relative to reactions (4)
and (5), reaction (2) was measured relative to reactions
(5) and (6), and reaction (3) was measured relative to
reactions (6) and (7).
OH ϩ DBE-4 : products
OH ϩ DBE-5 : products
OH ϩ DBE-6 : products
OH ϩ n-C4H10 : products
(1)
(2)
(3)
(4)
OH ϩ C2H5OH : products
OH ϩ n-C3H7OH : products
OH ϩ C6H12 : products
(5)
(6)
(7)
Figure 1 Kinetic data plotted according to eq. (I) from
experiments performed in 760 Torr synthetic air at 296 K
on the reaction of OH with dimethyl succinate relative to
ethanol (circles) and n-butane (triangles). The data for the
measurement relative to n-butane have been displaced ver-
tically by 0.04 units for clarity.
The initial concentrations (in molecule cmϪ3 units) of
the reference compounds were (4–6) ϫ 1014 n-C4H10,
(5–9) ϫ 1014 C2H5OH, (5–9) ϫ 1014 n-C3H7OH, and
(1–2) ϫ 1014 C6H12.
rates observed in the dark and under irradiation con-
ditions, supporting, therefore, that photolysis of the
dibasic esters was negligible. Wall loss rates of
OH radicals were generated by the photolysis of
5
1
5
1
5
5.05 ϫ 10Ϫ sϪ , 5.28 ϫ 10Ϫ sϪ , and 7.03 ϫ 10Ϫ
3
between (6–12) ϫ 1014 molecule cmϪ of methyl ni-
1
sϪ were measured for DBE-4, DBE-5, and DBE-6,
respectively. Rate coefficients for the OH-radical re-
action were determined using a relative rate method
[4], in which the relative disappearance rates of the
dibasic ester and a reference compound, whose OH-
radical reaction-rate coefficient is reliably known,
were monitored parallel in the presence of OH radi-
cals. Taking into consideration the losses of dibasic
esters to the wall of the reactor, the kinetic experi-
mental data can be treated using the following equa-
tion:
trite (CH3ONO) in synthetic air using the fluorescent
lamps. Approximately 6 ϫ 1014 molecule cmϪ3 of NO
was added to the reactant mixtures to suppress the for-
mation of O3 and NO3 radicals. In typical kinetic ex-
periments, DBE–methyl nitrite–NO–air mixtures
were irradiated for 20 min and 15 spectra were re-
corded. Four experimental runs were conducted on
each dibasic ester–reference pair. The observed dis-
appearance of dimethyl succinate, dimethyl glutarate,
[DBE]t
0
ln
Ϫ kwall ϫ t
ͭ ͮ
[DBE]t
kDBE
Kref
[reference]t
0
ϭ
ln
(I)
ͭ ͮ
[reference]t
where [DBE]t and [reference]t are the concentrations
0
0
of the dibasic ester and the reference compound, re-
spectively, at time t0; [DBE]t and [reference]t are the
corresponding concentrations at time t; kDBE and kref
are the rate coefficients for the reaction with OH rad-
icals of the dibasic ester and the reference compound,
respectively, and kwall is the measured wall loss
rate coefficient for the dibasic ester. Typical initial
reactant concentrations (in molecule cmϪ3 units) were
(2–2.5) ϫ 1014 DBE-4; (3–3.5) ϫ 1014 DBE-5; and
(4–4.5) ϫ 1014 DBE-6.
Figure 2 Kinetic data plotted according to eq. (I) from
experiments performed in 760 Torr synthetic air at 296 K
on the reaction of OH with dimethyl glutarate relative to
ethanol (circles) and 1-propanol (squares).

434
CAVALLI ET AL.
choose to cite final values of k1, k2, and k3 as averages
of the individual determinations together with error
limits that encompass the extremes of the individual
12
determinations. Hence, k1 ϭ (1.89 Ϯ 0.26) ϫ 10Ϫ
cm3 moleculeϪ1 sϪ , k2 ϭ (2.13 Ϯ 0.28) ϫ 10Ϫ12 cm3
1
moleculeϪ sϪ , and k3 ϭ (3.64 Ϯ 0.66) ϫ 10Ϫ cm3
1
1
12
moleculeϪ1 sϪ . The rate coefficients show the reactiv-
1
ity order, DBE-4 Ͻ DBE-5 Ͻ DBE-6, that is, an in-
crease in the rate coefficient with an increase in the
number of CH2 groups in the central carbon moiety.
The 298 K rate coefficients were also calculated
with the structure–reactivity estimation method of
Kwok and Atkinson [7]. The resulting rate coefficients
(in cm3 moleculeϪ sϪ units) are 1.15 ϫ 10Ϫ12, 2.56
1
1
12
ϫ 10Ϫ12, and 3.97 ϫ 10Ϫ for DBE-4, DBE-5, and
Figure 3 Kinetic data plotted according to eq. (I) from
experiments performed in 760 Torr synthetic air at 296 K
on the reaction of OH with dimethyl adipate relative to 1-
propanol (squares) and cyclohexane (crosses).
DBE-6, respectively, which are in reasonable agree-
ment with the measured values.
Aschmann and Atkinson [8] have recently mea-
sured rate coefficients for the reaction of OH radicals
with DBE-4, DBE-5, and DBE-6. The reported values
(in cm3 moleculeϪ sϪ units) were: (1.4 Ϯ 0.6) ϫ
1
1
and dimethyl adipate, against those of the reference
compounds, in the presence of OH radicals, are plotted
in Figures 1–3 according to eq. (I). The contribution
of the correction for the wall losses, that is, (kwall ϫ t)/
10Ϫ12, (3.3 Ϯ 1.1) ϫ 10Ϫ12, and (8.4 Ϯ 2.5) ϫ 10Ϫ
,
12
respectively. Although the rate coefficients measured
by Aschmann and Atkinson for DBE-4 and DBE-5 are
in reasonable agreement with our measured values,
there is a significant discrepancy in the rate coefficient
measured for DBE-6 compared to this study. The rea-
son for this large difference is not presently clear. In
our experiments, it was generally possible to find IR
structural features characteristic only of the DBE re-
actants and, therefore, to separate them quite accu-
rately from those of the degradation products. How-
ever, in the case of the DBE-6, the smooth IR features
made a very accurate analysis of the reaction spectra
more difficult. The study of Aschmann and Atkinson
has, however, also potential problems associated with
the analytical procedure used. In particular, the ad-
sorption and thermal desorption employed for the
analysis of the dibasic esters could lead to losses of
the sticky substrate compounds, especially of those
with high boiling point, resulting in larger rate
{[DBE]t /[DBE]t, accounts for 25–35% of the mea-
0
sured disappearance of the DBE during the 20-min
duration of the experiments. Good straight-line plots
were obtained, and the rate coefficient ratios obtained
from least-squares analyses of these data are given in
Table I. These rate-coefficient ratios have been placed
on an absolute basis by use of the following rate co-
1
1
efficients (in cm3 moleculeϪ sϪ units): k4 ϭ 2.44 ϫ
10Ϫ12 [5], k5 ϭ 3.3 ϫ 10Ϫ12 [6], k6 ϭ 5.5 ϫ 10Ϫ12 [6],
12
and k7 ϭ 7.21 ϫ 10Ϫ [5]. The resulting rate coeffi-
cients k1, k2, and k3 are given in Table I. The errors
indicated for the rate coefficients are 2␴ plus an ad-
ditional uncertainty of 10% due to the potential sys-
tematic errors associated with uncertainties in the ref-
erence rate coefficient. As seen from Table I,
indistinguishable results were obtained from experi-
ments using different reference compounds. We
Table I Kinetic Data for the Reactions of OH Radicals with Dimethyl Succinate, Dimethyl Glutarate, and Dimethyl
Adipate Measured at 296 Ϯ 2 K
CH3O(O)CH2CH2C(O)OCH3
CH3O(O)CH2CH2CH2C(O)OCH3
CH3O(O)CH2CH2CH2CH2C(O)OCH3
a
b,c
a
b,c
a
b,c
Reference
k1/kreference
k1
k2/kreference
k2
k3/kreference
k3
12
12
n-C4H10
0.78 Ϯ 0.04
0.59 Ϯ 0.05
(1.90 Ϯ 0.21) ϫ 10Ϫ
(1.88 Ϯ 0.25) ϫ 10Ϫ
12
12
C2H5OH
n-CH3H7OH
C6H12
0.68 Ϯ 0.03
0.38 Ϯ 0.02
(2.17 Ϯ 0.24) ϫ 10Ϫ
(2.12 Ϯ 0.24) ϫ 10Ϫ
12
12
0.71 Ϯ 0.02
0.47 Ϯ 0.02
(3.90 Ϯ 0.40) ϫ 10Ϫ
(3.38 Ϯ 0.37) ϫ 10Ϫ
a Quoted errors are 2␴ from a least-squares analysis.
b Quoted errors are 2␴ plus an additional 10% uncertainty for uncertainties in the reference rate coefficients (see text).
1
c The rate coefficients are given in units of cm3 moleculeϪ1 sϪ
.

THE OH RADICAL AND Cl-ATOM–INITIATED OXIDATION OF DIBASIC ESTERS
435
coefficients for the investigated hydroxyl radical re-
actions.
seen from Table II, indistinguishable results were ob-
tained from experiments using different reference
compounds. We choose to cite final values for k(Cl ϩ
DBE-4), k(Cl ϩ DBE-5), and k(Cl ϩ DBE-6), which
are averages of individual determinations with error
limits encompassing the extremes of the individual de-
The rate coefficients measured here can be com-
bined with a diurnally, seasonally, and annually av-
eraged global tropospheric OH radical concentration
of 1.0 ϫ 106 molecule cmϪ3 (24-h average) [9] to ob-
tain estimates of the tropospheric lifetimes of the di-
basic esters due to gas phase reaction with OH radi-
cals. Using our room-temperature rate constants, the
calculated lifetimes range from 6 days for dimethyl
succinate to 3 days for dimethyl adipate.
Using the relative rate method and following a sim-
ilar experimental procedure as described previously
for OH, rate coefficients have also been measured for
the gas-phase reaction of chlorine atoms with the three
dibasic esters, reactions (8) to (10). The reaction of
each dibasic ester with Cl was measured relative to
chloroethane [CH3CH2Cl; reaction (11)] and 2-chlo-
ropropane [CH3CHClCH3; reaction (12)].
terminations. Hence, k(Cl ϩ DBE-4) ϭ (6.79 Ϯ 0.93)
11
ϫ 10Ϫ12, k(Cl ϩ DBE-5) ϭ (1.90 Ϯ 0.33) ϫ 10Ϫ
,
k(Cl ϩ DBE-6) ϭ (6.08 Ϯ 0.86) ϫ 10Ϫ11. Again, the
rate coefficients show the expected reactivity order,
DBE-4 Ͻ DBE-5 Ͻ DBE-6, that is, an increase in the
rate coefficient with an increase in the number of CH2
groups in the central carbon moiety.
The chlorine-atom reaction-rate coefficients of the
dibasic esters were also calculated using the structure–
reactivity relationship method developed by Notario
1
et al. [12]; the calculated values (in cm3 moleculeϪ
sϪ units) k(Cl ϩ DBE-4) ϭ 6.39 ϫ 10Ϫ12, k(Cl ϩ
1
DBE-5) ϭ 2.62 ϫ 10Ϫ11, and k(Cl ϩ DBE-6) ϭ 6.92
ϫ 10Ϫ11 are in reasonably good agreement with those
measured.
Cl ϩ DBE-4 : products
Cl ϩ DBE-5 : products
(8)
(9)
It is interesting at this stage to compare the relative
reactivities of OH and Cl with respect to reaction with
the dibasic esters investigated in this work. For the OH
reactions, the relative reactivities are DBE-4, 1.0;
DBE-5, 1.1; and DBE-6, 1.9; but for the Cl reactions,
they are DBE-4, 1.0; DBE-5, 2.8; and DBE-6, 8.9. The
apparent higher selectivity of Cl atoms toward di-
methyl esters than OH radicals can be mainly ratio-
nalized in terms of the addition of further CH2 groups
to the central carbon chain of the dibasic esters. In the
case of reaction with OH, the addition of further CH2
groups to the central carbon moiety in DBE-4 only
results in modest enhancements to the overall rate co-
efficient because of the low reactivity of CH2 groups
in this position toward OH. In the case of the Cl re-
actions, however, the addition of further CH2 groups
results in a large contribution to the overall reactivity
because of the very high reactivity of the CH2 groups
in the central position of the molecule toward Cl at-
oms. Factors such as complex formation, particularly
in the case of the OH reactions, may also play a role.
Cl ϩ DBE-6 : products
(10)
(11)
OH ϩ CH3CHClCH3 : products
Cl ϩ CH3CH2Cl : products
(12)
The rate constant ratios, in Table II, obtained from
least-squares analysis of plots of eq. (I), have been
placed on an absolute basis using the following rate
constants (in cm3 moleculeϪ1 sϪ1 units): k11(Cl ϩ iso-
11
C3H7Cl) ϭ 2 ϫ 10Ϫ [10] and k12(Cl ϩ C2H5Cl) ϭ
12
8.04 ϫ 10Ϫ [11]. The resulting rate coefficients for
the Cl-atom–initiated reaction of DBE-4, DBE-5, and
DBE-6 are listed in Table II. Uncertainties quoted for
k(Cl ϩ DBE-4), k(Cl ϩ DBE-5), and k(Cl ϩ DBE-6)
are 2␴ plus an additional 10% uncertainty to account
for uncertainties in the reference rate coefficients. As
Table II Kinetic Data for the Reactions of Cl Atoms with Dimethyl Succinate, Dimethyl Glutarate, and Dimethyl
Adipate Measured at 296 Ϯ 2 K
CH3O(O)CH2CH2C(O)OCH3
CH3O(O)CH2CH2CH2C(O)OCH3
CH3O(O)CH2CH2CH2CH2C(O)OCH3
a
b,c
a
b,c
a
b,c
Reference
k8/kreference
k8
k9/kreference
k9
k10/kreference
k10
12
12
11
11
11
11
C2H5Cl
Iso-C3H7Cl
0.87 Ϯ 0.02
0.33 Ϯ 0.01
(7.00 Ϯ 0.72) ϫ 10Ϫ
(6.58 Ϯ 0.68) ϫ 10Ϫ
2.19 Ϯ 0.08
1.02 Ϯ 0.02
(1.77 Ϯ 0.19) ϫ 10Ϫ
(2.03 Ϯ 0.20) ϫ 10Ϫ
7.80 Ϯ 0.30
2.94 Ϯ 0.10
(6.27 Ϯ 0.67) ϫ 10Ϫ
(5.88 Ϯ 0.62) ϫ 10Ϫ
a Quoted errors are 2␴ from a least-squares analysis.
b Quoted errors are 2␴ plus an additional 10% uncertainty for uncertainties in the reference rate coefficients (see text).
1
c The rate coefficients are given in units of cm3 moleculeϪ1 sϪ
.

436
CAVALLI ET AL.
methyl succinate by heating and flushing a solid sam-
ple into the evacuated chamber and then filling to at-
mospheric pressure with synthetic air. No change in
the IR features was observed over a time period of
1 h. An attempt was also made to calibrate the solid
sample by dissolving it in a solvent, CH2Cl2, followed
by injection into the chamber. The spectra, in this case,
showed with time a slight decrease in the absorption
bands of what is most likely the dimer and an increase
in the intensities of bands attributed to the monomer.
Neither the use of solvents with higher polarity, the
heating of the injected solution, nor the increase of the
temperature of the reaction chamber resulted in further
conversion of the compound to the monomeric form.
Therefore, it is not presently possible to ascertain if
the calibrated IR spectrum of monomethyl succinate
really represents the pure monomeric form or a mix-
ture of the dimer and monomer. An analysis of the
product spectra from two experiments afforded an av-
erage yield of 0.447 Ϯ 0.111 for methyl succinate. The
estimated overall error (25%) includes uncertainties
associated with the analysis of the features of mono-
methyl succinate.
Dimethyl oxaloacetate ester exhibits, as reported in
the literature [13], complete enolization in the crystal
phase. Being a 1,3-dicarbonyl compound, in solution
it exists as an equilibrium mixture of keto and enol
forms, the proportionation depending on the polarity
of the solvent.
Vapor samples of dimethyl oxaloacetate, intro-
duced into the chamber by heating and flushing sam-
ples of the solid with N2, showed IR features of both
keto and enol forms, with an extremely slow conver-
sion rate between the forms compared to the overall
dark decay loss to the chamber walls. Therefore, in
order to derive a calibrated spectrum for each form,
NMR and FT-IR spectra of dimethyl oxaloacetate dis-
solved in CDCl3 and CHCl3 were recorded over a pe-
Figure 4 Plots of the amounts of succinic formic anhy-
dride (᭝), monomethyl succinate (ٗ), CO (᭺), dimethyl
oxaloacetate (Ϫ), and methoxy formyl peroxynitrate (ϩ)
formed as a function of the amounts of dimethyl succinate
reacted with Cl. The data for formic anhydride (᭝) have
been displaced vertically by 1 ppm for clarity.
Product Studies on the Cl-Atom–Initiated
Oxidation of Dimethyl Succinate
Photolysis experiments were performed on Cl2 –NO–
dimethyl succinate–air mixtures to determine the
products formed in the Cl-atom–initiated oxidation of
dimethyl succinate. Typical initial concentrations (in
3
units of molecule cmϪ ) were Cl2, 6 ϫ 1014; NO, 6 ϫ
1014; and dimethyl succinate, 2.5 ϫ 1014. The mixtures
were irradiated for 10 min, and the dimethyl succinate
consumption was typically ϳ60%. Based on FT-IR
reference spectra, succinic formic anhydride
[CH3OC(O)CH2CH2C(O)OC(O)H], monomethyl suc-
cinate [CH3OC(O)CH2CH2C(O)OH], dimethyl oxalo-
acetate [CH3OC(O)C(O)CH2C(O)OCH3], methoxy
formylperoxynitrate [CH3OC(O)OONO2], and CO
were identified among the degradation products. The
measured concentrations of these products are plotted
as a function of the amounts of dimethyl succinate
reacted in Figure 4. Good straight-line plots were ob-
tained; the lack of curvature in the plots strongly sup-
ports that all the products are primary and that no sec-
ondary reactions generated or removed these reaction
products during the time period of the experiments.
Least-squares analyses of these data led to the product
formation yields given in Table III.
Table III Molar Yields of the Products Observed in
the Cl-Atom–Initiated Oxidation of Dimethyl Succinate
in the Presence of NOx
Product
Molar Yield
Succinic formic anhydride
Monomethyl succinate
Carbon monoxide
Dimethyl oxaloacetate
Methoxy formylperoxynitrate
0.341 Ϯ 0.068a
0.447 Ϯ 0.111b
0.307 Ϯ 0.061a
0.176 Ϯ 0.044b
0.032 Ϫ 0.084c
The calibration of monomethyl succinate and di-
methyl oxaloacetate were not straightforward because
the IR spectra suggested that more than one form of
each compound exists in equilibrium in the liquid state
and that this equilibrium persists in the gas phase. An
attempt was made to obtain an IR spectrum of mono-
a Errors are the total overall estimated uncertainty of 20%.
b Errors are the total overall estimated uncertainty of 25%.
c The measured yield depends on the NO/NO2 ratio in the
system.

THE OH RADICAL AND Cl-ATOM–INITIATED OXIDATION OF DIBASIC ESTERS
437
riod of time of seven days. Both spectra showed com-
plete enolization for the initial freshly prepared
solution and, with time, a progressive increase in the
fraction of the keto form. Thus, infrared spectra of the
pure enol form and then of the equilibrium mixture of
the keto and enol forms were obtained. Furthermore,
the NMR solution spectra indicated that the position
of the keto–enol equilibrium in CDCl3 was 83.5% for
the keto form and 16.5% for the enol form. Expansions
into the reaction chamber of known amounts of the
solution at equilibrium allowed keto and enol infrared
absorption bands to be calibrated. As evident from the
previous discussion, the FT-IR quantification of mono-
methyl succinate and dimethyl oxaloacetate is subject
to significant uncertainties. Similar difficulties in ob-
taining IR spectra for the products from DBE-5 (mono-
methyl glutarate and dimethyl 1,3-acetonedicarboxy-
late) were observed by Tuazon et al. [14].
tified. However, an estimate of the molar RONO2 con-
centration was made from the integrated intensity of
1
the 1679 cmϪ absorption band and the average inte-
grated absorption coefficient (base 10) of (2.5 Ϯ 0.2)
ϫ 10Ϫ17 cm moleculeϪ1 of the corresponding band of
other organic nitrates [14]. Using this value, the av-
erage RONO2 molar formation yield obtained from
two experiments was 3.4 Ϯ 0.6%. The estimated total
error includes uncertainties associated with the ab-
sorption coefficient used and the generally weak sig-
nals observed. Model simulations of environmental
chamber incremental reactivity experiments [16] in-
volving the reaction of dimethyl succinate with OH
radicals in the presence of NO predict overall nitrate
yields of ϳ10–14%.
Since DBE-4 is symmetrical, the chlorine-atom–
initiated oxidation can proceed by only two path-
ways—H-atom abstraction from the terminal methyl
groups and H-atom abstraction from the central
CH2CH2 entity. In the troposphere, the alkyl radicals
formed after the H-atom abstractions react solely with
O2 to form the corresponding alkyl peroxy radicals
[17]. Further reaction with NO then leads to formation
of the analogous alkoxy radicals.
The possible reaction pathways of both alkoxy rad-
icals are outlined in Schemes I and II. The various
products identified in the analysis of the IR spectra
from the reactions of the two alkoxy radicals are en-
closed in boxes in the respective reaction schemes.
The products highlighted account for 82.4 Ϯ
16.4% C.
In the Scheme I, the alkoxy radical CH3OC(O)
CH2CH2C(O)OCH2O⅐ (I), formed after the H-atom ab-
straction at the terminal CH3 groups of the dibasic es-
ter, undergoes reaction with O2 to form succinic for-
mic anhydride and an ␣ ester rearrangement to form
monomethyl succinate plus the formyl radical, which
In the product spectra, spectral features are ob-
1
served at 1835, 1748, 1308, 1237, and 799 cmϪ .
These characteristic absorptions are assigned to me-
thoxy formylperoxynitrate [CH3OC(O)OONO2]. The
formation of this compound is very dependent on the
NO/NO2 ratio in the experimental system and, con-
sequently, its yield is low at the beginning of the ex-
periment when the NO/NO2 ratio is high and increases
gradually during the course of the experiment as the
NO/NO2 ratio becomes smaller (Fig. 4). An estimation
of the concentration of the peroxynitrate has been
3
1
1
made using the value of 1.09 ϫ 10Ϫ ppmϪ mϪ for
the absorption cross section of acetylperoxynitrate at
1835 cmϪ1 available in the literature [15]; its yield was
found to increase from approximately 3 to 8 mol%
during the progression of an experiment. The residual
product spectra also show the presence of RONO2-
1
type bands at ϳ843, ϳ1298, and ϳ1679 cmϪ . The
specific RONO2 products formed could not be iden-
Scheme I Suggested mechanism for the Cl-atom–initiated oxidation of dimethyl succinate; abstraction from the terminal
9CH3 groups.

438
CAVALLI ET AL.
reacts with O2 to give CO. If there are no additional
sources of CO and monomethyl succinate in the sys-
tem, as supported by the absence of curvature in plots
of their measured concentrations against reacted di-
methyl succinate (Fig. 4), then the molar formation
yield of the two products should be identical. The ob-
served discrepancy of 14% between the molar for-
mation yield measured for CO and that for mono-
methyl succinate is very probably attributable to the
significant uncertainties in the calibration of the acid.
We are, therefore, of the opinion that it is reasonable
to assume that the molar formation yield of CO reflects
the upper limit for the molar yield formation for mono-
methyl succinate.
In the FT-IR spectra of the reaction mixture,
formation of HCHO was not observed, allowing
the decomposition channel of the alkoxy radical
CH3OC(O)CH2CH2C(O)OCH2O⅐ (I) to be excluded.
This conclusion conforms with similar observations in
previous studies on the degradation mechanism of
DBE-5 with OH radicals [14] and of methyl propio-
nate with chlorine atoms [18].
ployed in this study, the major pathway of the
⅐C(O)OCH3 radical is expected to be addition of
O2 and further reaction with NO2 to form mainly
methoxy formylperoxynitrate, rather than de-
composition to CO2 and ⅐CH3 [19]. Because of
the lack of a calibrated spectrum for 3-oxo-
methyl propionate, identification and quantifi-
cation are not possible; however, the identi-
fication of methoxy formylperoxynitrate
provides evidence for the occurrence of this
pathway.
3. Decomposition of the alkoxy radical CH3OC(O)
CH(O⅐)CH2C(O)OCH3(II) via C–C cleavage of
the central carbon atoms would produce 2 mol-
ecules of methyl glyoxylate. However, this
channel is obviously of negligible importance
because the compound was not observed among
the products. This result confirms results from
a previous study [16], which deduced that
the reaction channel was unimportant because
of its relatively high estimated endothermic-
ity.
In Scheme II, the alkoxy radical CH3OC(O)
CH(O⅐)CH2C(O)OCH3 (II) formed after H-atom ab-
straction at the CH2 groups of DBE-4 can react by four
different pathways:
4. The alkoxy radical CH3OC(O)CH(O⅐)CH2C(O)
OCH3 (II) can also undergo a 1–5 H shift isom-
erization through a 6-membered–ring transition
state; because of a lack of the spectra of the sev-
eral possible polyfunctional products, it is not
possible to gauge the relative importance of this
pathway in the degradation mechanism of di-
methyl succinate.
1. Reaction with O2 will form dimethyl oxalo-
acetate.
2. Decomposition of this alkoxy radical via C–C
cleavage will lead to 3-oxo-methyl propionate
plus the methoxy formyl radical (⅐C(O)OCH3)
radical. Under the experimental conditions em-
The results of the present investigation have helped
to elucidate several aspects of the atmospheric pho-
Scheme II Reaction scheme for the Cl-atom–initiated oxidation of dimethyl succinate; abstraction from the central
9CH2CH2 9 unit.

THE OH RADICAL AND Cl-ATOM–INITIATED OXIDATION OF DIBASIC ESTERS
439
7. Kwok, E. S. C.; Atkinson, R. Atmos Environ 1995, 29,
1685.
8. Aschmann, S. M.; Atkinson, R. Int J Chem Kinet 1998,
30, 471.
9. Prinn, R. G.; Weiss, R. F.; Miller, B. R.; Huang, J.;
Alyea, F. N.; Cunnold, D. M.; Fraser, P. J.; Hartley, D.
E.; Simmonds, P. G. Science 1995, 269, 187.
10. Tyndall, G. S.; Orlando, J. J.; Wallington, T. J.; Dill,
M.; Kaiser, E. W. Int J Chem Kinet 1997, 29, 43.
11. Wine, P. H.; Semmes, D. H. J Phys Chem 1983, 87,
3572.
12. Notario, A.; Le Bras, G.; Mellouki, A. J Phys Chem
1998, 102, 3112.
13. Schiering, D. W.; Katon, J. E. Spectrochim Acta 1986,
42A, 487.
14. Tuazon, E. C.; Aschmann, S. M.; Atkinson, R. Environ
Sci Technol 1999, 33, 2885.
15. Tsalkani, N.; Toupance, G. Atmos Environ 1988, 23,
1849.
tooxidation mechanism of dimethyl succinate. How-
ever, the uncertainties encountered in the characteriza-
tion and quantification of several of the product
species demonstrate the need for the use of comple-
mentary and interdisciplinary analytical methods to
study the complex atmospheric chemistry of larger or-
ganic compounds.
F. Cavalli gratefully acknowledges the European Commis-
sion for funding her research at the Bergische Universita¨t
Wuppertal within the framework of the Environmental and
Climate Research and the Specific RTD Program. Financial
support within the 4th Framework Program, project EU-
ROSOLV, is also gratefully acknowledged.
BIBLIOGRAPHY
1. Bierbach, A. Produktuntersuchungen und Kinetik der
OH-initiierten Gasphasenoxidation aromatischer Koh-
lenwasserstoffe sowie ausgewa¨hlter carbonylischer Fol-
genprodukte; Ph.D. dissertation, BUGH Wuppertal,
Wuppertal, 1994.
16. Carter, W. P. L.; Luo, D.; Malkina, I. L.; Aschmann, S.
M.; Atkinson, R. Investigation of the Atmospheric
Ozone Formation Potential of Selected Dibasic Esters;
Final Report to the Dibasic Esters Group, Synthetic Or-
ganic Chemical Manufacturer’s Association, 1997.
Statewide Air Pollution Research Center, University of
California, Riverside (97-AP-RT71-002-FR). http://
www.cert.ucr.edu/ϳcarter/bycarter.htm.
2. Krimen, L. I. Org Synth 1970, 50, 1.
3. Taylor, W. D.; Allston, T. D.; Moscato, M. J.; Fezekas,
G. B.; Kozlowski, R.; Takacs, G. A. Int J Chem Kinet
1980, 12, 231.
17. Atkinson, R. J. Phys Chem Ref Data, Monograph 2,
4. Bierbach, A.; Barnes, I.; Becker, K. H.; Wiesen, E. En-
viron Sci Technol 1994, 28, 715.
5. Atkinson, R. J. Phys Chem Ref Data 1997, 26, 215.
6. Atkinson, R.; Baulch, D. L.; Cox, R. A.; Hampson, R.
F., Jr.; Kerr, J. A.; Rossi, M. J.; Troe, J. J. Phys Chem
Ref Data 1999, 28, 215, 191.
1994.
18. Cavalli, F.; Barnes, I.; Becker, K. H.; Wallington, T. J.
To be published.
19. Maurer, T. Tropospha¨rische Abbaumechanismen aus-
gesuchter Glykoldiether und Formaldehydacetale;
Ph.D. dissertation, BUGH Wuppertal, Wuppertal 2000.