Mechanisms for the reactions of OH with two unsaturated aldehydes: Crotonaldehyde and acrolein

Article abstract of DOI:10.1021/jp021530f

The mechanisms for the reaction of OH with the unsaturated aldehydes, acrolein, and crotonaldehyde, were determined at 1 atm total pressure in the presence of NO_x utilizing an environmental chamber/FTIR spectrometer system. Products in the OH-initiated oxidation of acrolein were CO, CO₂, CH₂O, HOCH₂CHO, and HCOOH. The major products in the OH-initiated oxidation of crotonaldehyde were CO, CO₂, CH₃CHO, and HC(O)CHO. Also observed were two PAN-type species, CH₂=CH-C(O)O₂NO₂ (APAN) from acrolein oxidation and CH₃-CH=CH-C(O)O₂NO₂ (CPAN) from crotonaldehyde. The near-complete mass balance obtained in these experiments allowed for a quantitative assessment of the branching ratios for abstraction and addition in these reactions. About 68% of the OH reaction with acrolein (crotonaldehyde) proceeded via abstraction of the aldehydic H, with the remainder occurring via addition to the double bond. The data allow for a more accurate assessment of the atmospheric source strength of APAN. Trends in the reactivity of acrolein and its methylated derivatives, methacrolein and crotonaldehyde, were elucidated. Data were consistent with structure-reactivity considerations.

Full text of DOI:10.1021/jp021530f

12252
J. Phys. Chem. A 2002, 106, 12252-12259
Mechanisms for the Reactions of OH with Two Unsaturated Aldehydes: Crotonaldehyde
and Acrolein
John J. Orlando* and Geoffrey S. Tyndall
Atmospheric Chemistry DiVision, National Center for Atmospheric Research, P.O. Box 3000,
Boulder, Colorado 80307-3000
ReceiVed: June 28, 2002; In Final Form: October 11, 2002
The mechanisms for the reaction of OH with the unsaturated aldehydes, acrolein and crotonaldehyde, have
been determined at 1 atm total pressure in the presence of NOx using an environmental chamber/FTIR
spectrometer system. Products observed in the OH-initiated oxidation of acrolein were CO, CO2, CH2O,
HOCH2CHO (glycolaldehyde), and HCOOH, while the major products identified in the OH-initiated oxida-
tion of crotonaldehyde were CO, CO2, CH3CHO, and HC(O)CHO (glyoxal). Also observed were two PAN-
type species, identified as CH2dCH-C(O)O2NO2 (APAN) from acrolein oxidation and CH3-CHdCH-
C(O)O2NO2 (CPAN) from crotonaldehyde. The near-complete mass balance obtained in these experiments
allows for a quantitative assessment of the branching ratios for abstraction and addition in these reactions. It
is shown that about 68% (50%) of the OH reaction with acrolein (crotonaldehyde) proceeds via abstraction
of the aldehydic H, with the remainder occurring via addition to the double bond. The data allow for a more
accurate assessment of the atmospheric source strength of APAN, a species which has now been identified
in ambient air. Trends in the reactivity of acrolein and its methylated derivatives, methacrolein and
crotonaldehyde, are also discussed; data are shown to be consistent with structure-reactivity considerations.
Introduction
CH2dC(CH3)C(O)O2 + NO2 + M f
CH2dC(CH3)C(O)O2NO2 + M (R2)
Unsaturated aldehydes (acrolein, CH2dCH-CHO, and its
methylated derivatives methacrolein, CH2dC(CH3)-CHO, and
crotonaldehyde, CH3-CHdCH-CHO) are present in the
atmosphere as the result of direct anthropogenic emissions
(e.g., from combustion sources), or as the result of the oxi-
dation of dienes. For example, it is well know that methacrolein
is a major first-generation product of isoprene degradation,1
while acrolein is generated as a byproduct of butadiene
oxidation.2,3
Due to its central importance in the oxidation of isoprene, a
great deal is known about the atmospheric fate of methacrolein.4-9
The dominant atmospheric loss for this species is via reaction
with OH, with photolysis and reaction with O3 and NO3 playing
at most a minor role. With a value for k1 of 2.9 × 10-11 cm3
molecule-1 s-1 at 298 K,6 the atmospheric lifetime during the
day for this species is ≈2-3 h. It has been shown that the OH
reaction occurs 50-55% via addition of OH to the CdC double
bond (R1a and R1b), and 45-50% via abstraction of the
aldehydic hydrogen (R1c), leading to the formation of three
peroxy radicals in the atmosphere:4,5
In fact, simultaneous measurements of MPAN (believed to be
derived from almost exclusively biogenic sources), PPN
(CH3CH2C(O)O2NO2, of anthropogenic origin), PAN (of mixed
biogenic/ anthropogenic origin), and O3 have been used to
quantify the relative contribution of biogenic and anthropogenic
hydrocarbons to ozone production.13,16 However, the recent
finding of a rapid loss of MPAN via reaction with OH20 calls
into question some of the assumptions made in these analyses.
At present, less is known about the mechanism of the reaction
of OH with either acrolein or crotonaldehyde:
OH + CH2dCH-CHO f products
(R3)
(R4)
OH + CH3-CHdCH-CHO f products
Rate coefficients for these reactions are well established,21-25
k3 ) 2.0 × 10-11 cm3 molecule-1 s-1 and k4 ) 3.4 × 10-11
cm3 molecule-1 s-1, and some products of the oxidation have
been identified and quantified.3,25,26 In particular, during the
course of our investigations, Magneron et al.25 published data
which suggested that ≈20-25% of (R3) occurs via addition
pathways, and that both addition and abstraction pathways are
operative in (R4). Nonetheless, a complete accounting of the
product yields, and hence an accurate assessment of the
branching ratio for OH addition and abstraction channels in (R3)
and particularly (R4), has yet to be achieved.
OH + CH2dC(CH3)-CHO (+O2)
f HOCH2-C(OO)(CH3)-CHO (R1a)
f OOCH2-C(CH3)(OH)-CHO (R1b)
f CH2dC(CH3)-C(O)O2
(R1c)
Interest in the chemistry of acrolein has been stimulated
by the recent observation of acryloylperoxy nitrate (CH2d
CHCOO2NO2, APAN, referred to as vinyl PAN in refs 25 and
26) in ambient air.27,28 Tanimoto and Akimoto27 observed this
species at both an urban site (suburban Tokyo) and a more
Among the suite of products generated from the ensuing
chemistry is methacryloylperoxynitrate (MPAN, an unsaturated
PAN analogue), which has been identified in ambient air in
many continental regions:10-19
10.1021/jp021530f CCC: $22.00 © 2002 American Chemical Society
Published on Web 12/03/2002

Reactions of OH with Unsaturated Aldehydes
J. Phys. Chem. A, Vol. 106, No. 51, 2002 12253
remote site (Rishiri Island, northern Japan). Roberts et al.28 have
identified APAN in ambient air near Houston, TX as part of
the recently completed TEXAS 2000 Air Quality Study.
Elevated APAN mixing ratios in this region were attributed in
part to large emissions of 1,3-butadiene and/or acrolein from
the petrochemical industry.
Interpretation of these APAN data will require knowledge
of both its sources (acrolein oxidation) and its sinks (expected
to be reaction with OH and thermal decomposition, by analogy
to MPAN20,29). Regarding the source strength, it is clear that
an understanding of the relative contribution of addition and
abstraction in the reaction of OH with acrolein is required, since
only the abstraction channel (R3c) can lead to APAN produc-
tion:
OH + CH2dCH-CHO (+O2)
f HOCH2-CH(OO)-CHO
f OOCH2-CH(OH)-CHO
f CH2dCH-C(O)O2
(R3a)
(R3b)
(R3c)
Figure 1. Product concentrations observed from the OH-initiated
oxidation of acrolein in 700 Torr of synthetic air, plotted as a function
of acrolein consumption. Details of the experimental conditions are
given in the text. Open squares, CO2; filled circles, CO; filled triangles,
CH2O; open triangles, glycolaldehyde; filled diamonds, APAN; open
circles, HCOOH.
In this study, experiments are conducted in an environmental
chamber/FTIR absorption system to determine the end products
of the OH-initiated oxidation of both acrolein and crotonalde-
hyde. These data are interpreted in terms of the branching ratios
for reactions (R3a-c), and for the analogous reaction channels
occurring in the reaction of OH with crotonaldehyde:
photolyzed for four to five periods (each of duration 2-3 min),
with an IR spectrum recorded following each irradiation.
Some experiments, using Cl-atoms to initiate oxidation, were
carried out specifically to determine IR spectra for the unsatur-
ated PAN analogues of MPAN, i.e., APAN from acrolein and
CPAN from crotonaldehyde (see results section for details).
These experiments were conducted in 1 atm of synthetic air,
OH + CH3-CHdCH-CHO (+O2)
f CH3-CH(OH)-CH(OO)-CHO (R4a)
and involved the photolysis of Cl2 (≈5 × 1015 molecule cm-3
)
f CH3-CH(OO)-CH(OH)-CHO (R4b)
in the presence of either crotonaldehyde or acrolein (≈7 × 1014
molecule cm-3) and NO2 (1.5 × 1014 molecule cm-3).
Quantification of reactants and products was done by
comparison with reference spectra recorded with our spectrom-
eter system,5,31-33 with the exception of the unsaturated PAN
species (CH2dCH-C(O)O2NO2 and CH3-CHdCH-C(O)O2-
NO2) whose identification and quantification is described in the
results section. Quantification was done primarily using the fol-
lowing absorption bands: CO, 2080-2200 cm-1; CO2, 2300-
2380 cm-1; CH2O, near 1745 cm-1; glycolaldehyde, near 1115
cm-1; glyoxal, near 1730 cm-1 and/or 2835 cm-1; CH3CHO,
1320-1490 cm-1 and/or 1745 cm-1. Our absorption cross
sections for glycolaldehyde are consistent with previous stud-
ies.34,35 Minor corrections (less than 10%) to product yield data
were made to account for product loss via reaction with OH.
Chemicals used in this study and their sources and purities
were as follows: methyl nitrite and ethyl nitrite, synthesized
according to the procedure of Taylor et al.;36 O2, UHP, U. S.
Welding; N2, boil-off from liquid N2, U. S. Welding; acrolein,
90%, Aldrich; crotonaldehyde, 99%, Aldrich; NO, UHP, Linde;
Cl2, HP, Matheson. Gases were used as received. Acrolein and
crotonaldehyde were subjected to several freeze-pump-thaw
cycles prior to use.
f CH3-CHdCH-C(O)O2
(R4c)
Trends in the reactivity of acrolein, methacrolein, and crotonal-
dehyde are discussed, as are the implications of the results for
the atmospheric source of APAN.
Experimental Section
All experiments described here were conducted in a 47 L
stainless steel environmental chamber,20,30 which is interfaced
to a Fourier transform infrared spectrometer via a set of Hanst-
type multipass optics (observational path length 32.6 m). Spectra
were obtained at a resolution of 1 cm-1 from the coaddition of
200 scans (acquisition time 3-4 min), and covered the range
800-3900 cm-1. Photolyses were carried out along the length
of the chamber, using the output of a Xe-arc lamp filtered to
provide radiation in the range 240-400 nm.
Products of the OH-initiated oxidation of both acrolein and
crotonaldehyde were measured at 700-720 Torr total pressure
(O2 partial pressure 150 Torr, balance N2) from the photolysis
of mixtures of an OH precursor (methyl or ethyl nitrite), the
unsaturated aldehyde, and NO. Ethyl nitrite was used as the
OH source for the acrolein experiments, since methyl nitrite
photolysis generates formaldehyde, a likely acrolein oxidation
product. For studies of crotonaldehyde, most experiments were
carried out using methyl nitrite, to avoid acetaldehyde production
from ethyl nitrite. A few experiments were conducted using ethyl
nitrite, however, to assess the possibility of CH2O production
from crotonaldehyde oxidation. Typical initial concentrations
were as follows: methyl or ethyl nitrite, (1.5-2.8) × 1015
molecule cm-3; acrolein, (3.5-8) × 1014 molecule cm-3 or
crotonaldehyde, (2.4-14) × 1014 molecule cm-3; NO, ≈3.5 ×
1014 molecule cm-3. To obtain product yield data, mixtures were
Results and Discussion
Products of Reaction of OH with Acrolein. Products
observed following the OH-initiated oxidation of acrolein were
CO2, CO, CH2O, glycolaldehyde (HOCH2CHO), and formic
acid; concentrations of these species, corrected for the occur-
rence of secondary reactions,37 as a function of acrolein
consumption, are shown in Figure 1. These product concentra-
tions were observed to be linear with acrolein consumption,
indicating that all are likely primary products (with the possible

12254 J. Phys. Chem. A, Vol. 106, No. 51, 2002
Orlando and Tyndall
TABLE 1: Peak Absorption Cross Sections (10-18 cm2
molecule-1) for PAN,38 MPAN,4 APAN, and CPAN (This
Work, Assumed by Analogy, See Text for Details)
PAN
1.14
MPAN
1.2
CPAN
APAN
794
901
950
0.22
0.65
0.65
0.3
1070
1125
1163
1193
1204
1300
1452
1740
1810
1840
1.6
1.8
1.47
0.18
0.34
1.3
1.1
2.9
1
1.28
1.1
0.21
3.2
3.04
0.8
2.8
0.8
0.6
O2 and NO2, APAN is expected to be the sole product obtained
from abstraction, while Cl addition under these conditions would
generate chlorinated peroxynitrate species, products which ob-
viously will not be formed following OH attack:
Figure 2. Infrared absorption spectra observed following the photolysis
of a mixture of Cl2 (5 × 1015 molecule cm-3), acrolein (upper trace, 7
× 1014 molecule cm-3) or crotonaldehyde (lower trace, 7 × 1014
molecule cm-3), NO2 (1.5 × 1014 molecule cm-3), O2 (150 Torr) and
N2 (550 Torr). Absorption features due to the parent aldehyde have
been subtracted for clarity. Labeled peaks are assigned to APAN,
CH2dCH-C(O)OONO2, or CPAN (CH2dCH-C(O)OONO2), see text
for details.
Cl + CH2dCH-CHO (+O2,NO2)
f ClCH2-CH(OONO2)-CHO (R5a)
f O2NOOCH2-CHCl-CHO
f CH2dCH-C(O)O2NO2
(R5b)
(R5c)
exception of formic acid, see below). As will be detailed later,
concentrations of CO2, CO, and CH2O are expected to level
off somewhat toward the end of an experiment (due to APAN
formation), but this effect is not immediately evident in the data.
Product yields, obtained from the slopes of the data shown in
Figure 1, are as follows: CO2, 70 ( 5%; CO, 62 ( 5%; CH2O,
49 ( 4%; and glycolaldehyde, 32 ( 4%, and formic acid, 1.5
( 0.7%. The origin of the minor product formic acid is not
known, but may be heterogeneous in nature. As its formation
at such low levels has no bearing on the data interpretation, it
will not be discussed further.
Thus, it seems reasonable to assume that the absorption features
that are common to the Cl- and OH-initiated experiments (1075,
1204, 1298, 1745, and 1820 cm-1) are indeed due to APAN.
Furthermore, the observed “dark” behavior is as expected:
APAN is stable over the course of an hour or more, while loss
of HO2NO2 occurs to a significant extent over this time period.
An absolute calibration of the APAN infrared spectrum cannot
be done without prior knowledge of the branching ratios for
(R5). Instead, an approximate calibration is obtained from an
assumption of similar intensities for APAN bands compared to
the analogous absorption features in MPAN4 and PAN.38 To
accomplish this, an average absorption cross section for the 1300
and 1740 cm-1 bands was established using the MPAN and
PAN data, and the APAN spectrum was adjusted to achieve
best agreement with these average absorption cross sections.
Use of these absorption cross sections (which are given in Table
1) leads to the conclusion that about 20% of the reaction of Cl
with acrolein occurs via abstraction, (R5c). This branching ratio
is in quantitative agreement with conclusions drawn from a re-
cently published assessment of Cl-atom reactivity with unsatur-
ated oxygenates,40 thus lending support to our APAN calibration.
The assumed cross sections can also be used to quantify the
APAN yields following reaction of OH with acrolein, as shown
in Figure 1. As discussed earlier, these yields vary with the
extent of acrolein consumption, reaching a maximum level of
22% at the highest acrolein conversion. Once the contribution
of APAN is included, the observed products in the OH-initiated
oxidation of acrolein represent on average 97 ( 7% of the
reacted carbon, indicating that no major product(s) remain
unaccounted for. It now remains to assess these product yields
in terms of the branching ratios for (R3a)-(R3c). As pointed
out above, reaction of OH with acrolein will result in the
generation of three peroxy radicals, HOCH2-CH(OO)-CHO,
OOCH2-CH(OH)-CHO, and CH2dCH-C(O)O2. Under the
conditions of our experiments, the main fate of these species
will be reaction with NO to generate the corresponding oxy
radicals:
Also observed in the product spectra were absorption features
at 1075, 1204, 1298, 1745, and 1820 cm-1. The growth of these
features was found to be nonlinear with respect to acrolein
consumption, accelerating toward the end of an experiment when
NO2/NO concentration ratios were at a maximum. Since the
intensities of these absorption features grew in unison (i.e., their
relative intensities remained constant), it is likely that they are
all due to a single species (or at least that they are all obtained
from the same reaction). On the basis of the position of the
absorption bands (note that PAN and MPAN both have
absorption features near 1300, 1740, and 1820 cm-1 4,38
and
)
on the basis of the observed temporal profile, we assign these
absorption features to the PAN analogue, acryloylperoxynitrate,
CH2dCH-C(O)O2NO2 (APAN).
Further confirmation of the identity of the unidentified
absorber as APAN was obtained from experiments involving
photolysis of Cl2/acrolein/NO2/air mixtures. Following a brief
photolysis period (30 s), absorption features at 1075, 1204, 1298,
1745, and 1820 cm-1 again appeared along with other absorption
features at 1304 cm-1 and 1725 cm-1 due to HO2NO2,39 see
Figure 2 (upper trace). Upon sitting in the dark, the features at
1075, 1204, 1298, 1745, and 1820 cm-1 remained unchanged,
while the HO2NO2 features decreased (presumably due to
thermal decomposition). As is the case with OH, Cl-atom
reaction with acrolein can occur via addition to the double bond,
or via abstraction of the aldehydic H atom. In the presence of

Reactions of OH with Unsaturated Aldehydes
J. Phys. Chem. A, Vol. 106, No. 51, 2002 12255
branching ratio to OH addition of (k3a+k3b)/k3 of e 44%, with
HOCH2-CH(OO)-CHO + NO f
k
3a g 4 k3b. The larger rate coefficient for external OH addition
HOCH2-CH(O)-CHO + NO2 (R6)
to acrolein is consistent with findings in the OH/methacrolein
reaction, k1a g 5 k1b.4,5
OOCH2-CH(OH)-CHO + NO f
It now remains to consider the end-products of OH abstrac-
tion, (R3c). As shown above, the CH2dCH-C(O)O2 radical
generated following (R3c) will either react with NO2 to form
APAN (R11), or with NO to form CH2dCHC(O)O, (R8). The
CH2dCHC(O)O species is then expected to undergo rapid
decomposition to generate CO2 and the vinyl radical:
OCH2-CH(OH)-CHO + NO2 (R7)
CH2dCH-C(O)O2 + NO f
CH2dCH-C(O)O + NO2 (R8)
Organic nitrates are also possible products of these RO2/NO
reactions, though yields are expected41 to be small, probably
<5%. Reaction of the peroxy radicals with NO2 will also occur,
CH2dCH-C(O)O f CH2dCH + CO2
(R17)
Thus, abstraction (R3c) is expected to yield either APAN or
(at least one) CO2, and the sum of the yields of these two species
represents an upper limit to the abstraction channel branching
ratio, k3c/k3 e 0.9. As will be shown below, further CO2 appears
to be generated from the vinyl radical, and hence k3c/k3 is indeed
less than 0.9.
HOCH2-CH(OO)-CHO + NO2 f
HOCH2-CH(OONO2)-CHO (R9)
OOCH2-CH(OH)-CHO + NO2 f
O2NOOCH2-CH(OH)-CHO (R10)
The reaction of vinyl radical with O2 is known to be rapid,
k18 ) 1 × 10-11 cm3 molecule-1 s-1 43-47
,
and thus represents
the sole fate of this species under our experimental conditions.
Products of this reaction have not been quantitatively determined
as yet, however. Gutman and co-workers44,45 have shown that
formation of HCO and CH2O is important at low pressures, a
result recently confirmed by Wang et al.48 Furthermore, Wang
et al.48 demonstrated the existence of a measurable yield of CH3
and CO2, (R18b), at low pressures:
CH2dCH-C(O)O2 + NO2 f
CH2dCH-C(O)O2NO2 (APAN) (R11)
though the peroxynitrates formed in (R9) and (R10) are likely
to be thermally unstable at 298 K. On the other hand, APAN
formed in (R11) is stable and, as shown above, is a detectable
end-product of OH abstraction.
The chemistry of the oxy species generated in (R6) to (R8)
can be inferred by analogy to the chemistry of other oxy radi-
cals,41,42 and more specifically by analogy to chemistry occurring
in methacrolein oxidation.4,5 We will first consider the two oxy
radicals formed via OH-addition pathways, HOCH2-CH(O)-
CHO and OCH2CH(OH)-CHO. The HOCH2-CH(O)-CHO
species will decompose, with the most likely end products being
glycolaldehyde and CO, with a possible minor channel to form
CH2O and glyoxal:
CH2dCH + O2 f CH2O + HCO
f CH3 + CO2
(R18a)
(R18b)
In our system, HCO will be rapidly converted to CO, while
CH3 radicals will be converted to CH2O:
CH3 + O2 f CH3O2
(R19)
(R20)
(R21)
CH3O2 + NO f CH3O + NO2
CH3O + O2 f CH2O + HO2
HOCH2-CH(O)-CHO f HOCH2-CHO + HCO (R12a)
f CH2OH + HC(O)CHO (R12b)
At atmospheric pressure, the formation of a stable vinyl peroxy
radical may also occur:49
HCO + O2 f HO2 + CO
(R13)
(R14)
CH2dCH + O2 f CH2dCHOO
(R18c)
CH2OH + O2 f CH2O + HO2
The fate of this species is open to speculation, but may include
formation of ketene, glyoxal, or CO and CH2O.50,51
The OCH2-CH(OH)-CHO radical will also be subject to rapid
decomposition, yielding formaldehyde and glyoxal:
Upon consideration of the data in Figure 1, and taking into
account the 32% yield of glycolaldehyde and CO from OH
addition channels, it is apparent that the following products are
obtained via abstraction: CO2 (yield ≈70%), CH2O (≈49%),
CO (≈30%), and APAN (≈5-22%, depending on the extent
of acrolein conversion). Note that, since the APAN yield
increases with extent of conversion, some downward curvature
in the yields of CO, CO2, and CH2O is likely. Given our
observation of essentially 100% carbon balance, it follows that
these products represent the sole end-products of abstraction,
which must contribute ≈68% to the total reaction. From the
discussion given above, it seems likely that three sets of products
can be obtained from the abstraction pathway: APAN, via
(R11); CO2 + CH2O + CO, via (R8), (R17), and (R18a); and
2CO2 + CH2O via (R8), (R17), (R18b), and (R19)-(R21). The
yield data are consistent with this supposition. First, a plot of
the sum of the concentrations of CO (in excess to that formed
OCH2-CH(OH)-CHO f CH2O + HOCH-CHO (R15)
HOCH-CHO + O2 f HC(O)CHO + HO2 (R16)
To assess the importance of the two addition pathways, we
first note that glyoxal was not detected as a reaction product
(yield of glyoxal and hence branching ratio to (R3b) e 8%).
Furthermore, the lack of significant glyoxal formation indicates
that the majority of the HOCH2-CH(O)-CHO radicals (formed
via (R3a), followed by (R6)), must lead to glycolaldehyde. Since
glycolaldehyde is not likely to be a product of OH abstraction
(see below), and assuming minimal nitrate production in (R6),
a branching ratio to (R3a) of ≈32% can be derived. Allowing
for a possible 8% glyoxal yield gives an upper limit for the

12256 J. Phys. Chem. A, Vol. 106, No. 51, 2002
Orlando and Tyndall
After accounting for the products listed above, a number of
additional unassigned absorption features (901, 950, 1070, 1193,
1300, 1452, 1740, and 1820 cm-1) were apparent in the product
spectra. As was the case for the absorption bands attributed to
APAN, the appearance profile of these features was nonlinear
with crotonaldehyde consumption, accelerating toward the end
of an experiment. Furthermore, these same absorption features
were evident (with the same relative intensities) following the
photolysis of mixtures of Cl2, crotonaldehyde, and NO2 in air,
see Figure 2 (lower trace). Using the same logic applied above
for APAN, it seems reasonable to assign these absorption
features to a crotonaldehyde-derived unsaturated PAN species,
CH3-CHdCH-C(O)O2NO2, hereafter referred to as CPAN.
As was the case for APAN, an absolute calibration of the CPAN
infrared spectrum is not possible without prior knowledge of
the mechanism for reaction of Cl with crotonaldehyde. However,
approximate absorption cross sections can be obtained from an
assumption of similar band strengths for absorption features
common to PAN, MPAN, APAN, and CPAN. Absorption cross
sections assigned to the various CPAN absorption features
(obtained using the same logic as was use for APAN spectral
calibration discussed earlier) are given in Table 1, and CPAN
concentrations observed in the OH-initiated crotonaldehyde
oxidation experiments are shown in Figure 3. The identified
products (CO, CO2, CH3CHO, glyoxal, and CPAN) account for,
on average, 77 ( 12% of the reacted crotonaldehyde carbon,
indicating the likelihood of the existence of an additional
undetected product.
Figure 3. Product concentrations observed from the OH-initiated
oxidation of crotonaldehyde in 700 Torr of synthetic air, plotted as a
function of crotonaldehyde consumption. Details of the experimental
conditions are given in the text. Filled triangles, acetaldehyde; open
squares, CO2; filled circles, CO; open triangles, glyoxal; filled diamonds,
CPAN.
in the addition channel), CH2O, CO2 and APAN versus acrolein
consumption (on a per carbon basis) yielded a slope of 68 (
8%. Furthermore, a plot of the sum of the concentrations of
APAN and CH2O versus acrolein consumption, on a molar basis,
also yielded a slope of 68 ( 8%. Further consideration of the
CO and CO2 data indicate that toward the end of an experiment,
the CO2 + HCHO + CO channel was occurring with about a
30% yield, the 2CO2 + CH2O channel with about a 20% yield,
and the APAN channel with about a 20% yield. Thus, it appears
likely that pathways (R18a) and (R18b) are both occurring in
the vinyl + O2 reaction, consistent with the findings of Wang
et al.48 However, yields of CO2 are higher relative to CO than
might be expected if the CH3 + CO2 channel (R18b) is minor,
and may indicate that other sources of CO2 are present in our
experiments which skew our determination of the relative
importance of (R18a) and (R18b).
Previous studies of the reaction of OH with acrolein have
been carried out by Grosjean et al.,26 Liu et al.3 and, very
recently, by Magneron et al.25 Grosjean et al. identified APAN
(referred to as vinyl PAN in their paper), CH2O, and glycol-
aldehyde and/or glyoxal as products, though no yield data were
reported. Similarly, Liu et al. reported the formation of CH2O,
glyoxal, glycolaldehyde, glycidaldehyde, malonaldehyde, and
3-hydroxy-propanal following reaction of OH with acrolein;
again, no yield data were reported. Magneron et al.25 identi-
fied glyoxal (yield 5-10%), glycolaldehyde (12-25%), ketene
(2-3%), as well as APAN, CH2O, and CO (no yield data given)
as products of (R3). Where comparisons are possible, the yields
reported by Magneron et al.25 are in agreement with our data,
within the combined uncertainties. On the basis of their
combined glycolaldehyde and glyoxal yields,25 these workers
suggested a ≈25% occurrence of addition pathways in (R3),
similar to our conclusion.
The chemistry pursuant to reaction of OH with crotonalde-
hyde, (R4a)-(R4c),
OH + CH3-CHdCH-CHO (+O2)
f CH3-CH(OH)-CH(OO)-CHO (R4a)
f CH3-CH(OO)-CH(OH)-CHO (R4b)
f CH3-CHdCH-C(O)O2
(R4c)
is expected to be similar to that observed in the case of acrolein.
Peroxy radicals generated by OH addition, (R4a) and (R4b),
will be converted predominantly to the corresponding oxy
radicals, whose subsequent fate will be decomposition to form
either glyoxal and acetaldehyde or CO and 2-hydroxypropanal,
CH3CH(OH)CHO:
CH3-CH(OH)-CH(OO)-CHO + NO f
CH3-CH(OH)-CH(O)-CHO + NO2 (R22)
CH3-CH(OO)-CH(OH)-CHO + NO f
CH3-CH(O)-CH(OH)-CHO + NO2 (R23)
CH3-CH(OH)-CH(O)-CHO
f HCO + CH3-CH(OH)CHO (R24a)
f CH3CHOH + HCOCHO
HCO + O2 f CO + HO2
(R24b)
(R13)
(R25)
Products of Reaction of OH with Crotonaldehyde. Prod-
ucts observed following the OH-initiated oxidation of crotonal-
dehyde were CO (42 ( 4% yield), CO2 (41 ( 4%), glyoxal
(21 ( 5%), and acetaldehyde (51 ( 9%), see Figure 3. Some
formaldehyde may also be formed, yield <12%. Uncertainties
in the crotonaldehyde study are larger than in the case of
acrolein, owing to the more complex nature of the IR spectra
involved.
CH3CHOH + O2 f CH3CHO + HO2
CH3-CH(O)-CH(OH)-CHO f
CH3CHO + CH(OH)-CHO (R26)
CH(OH)-CHO + O2 f HCOCHO + HO2 (R27)

Reactions of OH with Unsaturated Aldehydes
J. Phys. Chem. A, Vol. 106, No. 51, 2002 12257
By analogy to acrolein (see above) and methacrolein,4,5 products
likely to arise from OH abstraction, (R4c), include CPAN:
TABLE 2: Total Rate Coefficients and Rate Coefficients for
Addition and Abstraction in the Reactions of OH with
Unsaturated Aldehydes: All Rate Coefficient Data in Units
of 10-12 cm3 molecule-1 s-1
CH3-CHdCH-C(O)O2 + NO2 f
compound
total k
% addition
kabs
kaddition
CH3-CHdCH-C(O)O2NO2 (CPAN) (R28)
acrolein
methacrolein
crotonaldehyde
20
28
34
32 ( 5
55 ( 5
50 ( 15
14 ( 2
13 ( 2
17 ( 5
6 ( 1
15 ( 2
17 ( 5
or CO2 and (via 2-methyl vinyl radical formation) either
methyglyoxal, acetaldehyde, methyl ketene, CO, or additional
CO2:
(note that quantification of acetaldehyde in our spectra is
extremely difficult). Other products reported, but for which yield
data are not given, include CO, CPAN (referred to as methyl
vinyl PAN), formaldehyde, and either methylglyoxal or 2-hy-
droxypropanal. Though they presented evidence for the occur-
rence of both abstraction and addition, quantification of the
branching ratios to the competing pathways was not possible.
Trends in Reactivity of OH with the Unsaturated Alde-
hydes. As mentioned in the Introduction, the rate coefficients
for reaction of OH with acrolein, methacrolein, and crotonal-
dehyde are well-known. With these total rate coefficient data,
the branching ratio data for acrolein and crotonaldehyde
presented above, and previously published data on methacrolein
branching ratios,4,5 it becomes possible to calculate site-specific
rate coefficients for abstraction and addition, as is done in Table
2. Inspection of the data shows that the rate coefficients for
abstraction from acrolein, methacrolein, and crotonaldehyde are
very similar to one another, and are quite similar to the rate
coefficient for reaction of OH with acetaldehyde,52 k ) 1.6 ×
10-11 cm3 molecule-1 s-1. On the other hand, the rate coefficient
for OH addition to acrolein (6 × 10-12 cm3 molecule-1 s-1), is
considerably smaller than the rate coefficients for the addition
of OH to the methyl-substituted species (about 1.6 × 10-11 cm3
molecule-1 s-1). Thus, as expected from structure-reactivity
considerations,53,54 the presence of a methyl group leads to a
large enhancement of the reactivity of the double bond (by a
factor of 2.5 or so).
As has been presented previously,53,54 the presence of the
adjacent carbonyl group provides a deactivating effect on the
reactivity of the double bond toward OH. Note that addition of
OH to acrolein (methacrolein or crotonaldehyde) is slower than
addition of OH to the corresponding “parent” alkene, propene
(2-methyl-propene or 2-butene). In the structure-reactivity
framework of Kwok and Atkinson,54 the -CHO group is
assigned a substituent factor of 0.34 relative to a methyl group
(that is, the rate coefficient for addition of OH to acrolein is
expected to be one-third that of OH addition to propene).
Consideration of the data in Table 2 suggests a modest decrease
in this parameter, to 0.27.
Finally, it is interesting to speculate upon the actual site of
OH addition (i.e., adjacent (R) to the carbonyl group or on the
double-bonded carbon furthest from the carbonyl group (â)) for
the three unsaturated aldehydes. For methacrolein, the large yield
of hydroxyacetone (about 45%)4,5 out of a total of 55% of the
reaction occurring via addition, indicates that at least 80% of
the addition occurs at the â-carbon. Similarly, the large yield
of glycolaldehyde (32%) relative to that of glyoxal (<8%)
following addition of OH to acrolein again indicates at least
80% OH addition to the (terminal) â-carbon. The situation is
less clear in the case of crotonaldehyde. However, the upper
limit to the yield of 2-hydroxypropanal (35%, see above)
coupled with a 20% yield of glyoxal may be an indication of
measurable (about 40%) addition occurring at the R-carbon (i.e.,
adjacent to the carbonyl group), possibly a consequence of the
stabilization of the resulting hydroxyalkyl radical by the methyl
group in crotonaldehyde. Alternatively, the glyoxal (and co-
CH3-CHdCH-C(O)O2 + NO f
CH3-CHdCH-C(OO) + NO2 (R29)
CH3-CHdCH-C(OO) f CH3-CHdCH + CO2 (R30)
CH3-CHdCH + O2 f CH3CHO + HCO (R31a)
f CH3CH2 + CO2 ? (R31b)
f other products ? (R31c)
CH3CH2 + O2 f CH3CH2O2
CH3CH2O2 + NO f CH3CH2O + NO2
CH3CH2O + O2 f CH3CHO + HO2
(R32)
(R33)
(R34)
As mentioned above, the identified products only account
for about 77% of the oxidized crotonaldehyde (on a per carbon
basis). From a consideration of the chemistry just presented,
the most likely candidate for the missing product is 2-hydroxy-
propanal, CH3CH(OH)CHO. Given the lack of a full carbon
balance, determination of the branching ratios for addition and
abstraction is not as clear-cut as was the case for acrolein.
Nonetheless, a reasonable estimate is attainable. Glyoxal appears
to be a product unique to addition, thus providing a lower limit
of 21% to addition channels. The sum of the CO2 and CPAN
yields, provides an upper limit (e64%) to abstraction, see
(R28)-(R30). We also note that the acetaldehyde yield (51%)
exceeds that of glyoxal (21%). Thus, a significant amount of
acetaldehyde must be derived from OH abstraction. In fact, if
(R31a) and (R31b) represent the major channels in the reaction
of 2-methylvinyl with O2, as was the case for reaction of vinyl
with O2, the sum of the CPAN yield and “excess” acetaldehyde
yield (i.e., the yield of acetaldehyde in excess of its addition
coproduct, glyoxal) would provide a measure of the abstraction
branching ratio; this term results in a best estimate for the
branching to abstraction, k4c/k4, of ≈50%. Furthermore, from
mass balance arguments, the yield of 2-hydroxypropanal cannot
exceed a value of 35%. This assumption, coupled with the
measured yield of glyoxal (21%), provides an upper limit of
about 56% for addition channels. Finally, if it is assumed that
the majority of the missing carbon in our product distribution
is indeed attributable to 2-hydroxypropanal, the following
product channel yields are most consistent with our data toward
the end of a run: acetaldehyde and glyoxal via addition, 21%;
2-hydroxypropanal and CO via addition, 35%; CPAN via
abstraction, about 20%; CO + CO2 + acetaldehyde via addition,
7%; 2 CO2 + acetaldehyde via addition, 19%. While the
conclusions drawn here clearly involve some speculation, it
seems unlikely that the branching ratios for abstraction and
addition are significantly different from each other.
Some products of the OH-initiated oxidation of crotonalde-
hyde have recently been quantified by Magneron et al.25 Their
yield of glyoxal, 16%, is in agreement with ours, while their
yield of acetaldehyde (30 ( 5%) is somewhat lower than ours

12258 J. Phys. Chem. A, Vol. 106, No. 51, 2002
Orlando and Tyndall
(4) Tuazon, E. C.; Atkinson, R. Int. J. Chem. Kinet. 1990, 22, 591.
(5) Orlando, J. J.; Tyndall, G. S.; Paulson, S. E. Geophys. Res. Lett.
1999, 26, 2191.
(6) Gierczak, T.; Burkholder, J. B.; Talukdar, R. K.; Mellouki, A.;
Barone, S. B.; Ravishankara, A. R. J. Photochem. Photobiol. A 1997, 110,
1.
(7) Canosa-Mas, C. E.; Carr, S.; King, M. D.; Shallcross, D. E.;
Thompson, K. C.; Wayne, R. P. Phys. Chem. Chem. Phys. 1999, 1, 4195.
(8) Kwok, E. S. C.; Aschmann, S. M.; Arey, J.; Atkinson, R. Int. J.
Chem. Kinet. 1996, 28, 925.
(9) Carter, W. P. L.; Atkinson, R. Int. J. Chem. Kinet. 1996, 28, 497.
(10) Bertman, S. B.; Roberts, J. M. Geophys. Res. Lett. 1991, 18, 1461.
(11) Grosjean, D.,; Williams, E. L., II; Grosjean, E. EnViron. Sci.
Technol. 1993, 27, 110.
(12) Williams, E. L., II; Grosjean, E.; Grosjean, D. J. Air Waste Manage.
Assoc. 1993, 43, 873.
(13) Williams, J.; Roberts, J. M.; Fehsenfeld, F. C.; Bertman, S. B.;
Buhr, M. P.; Goldan, P. D.; Hu¨bler, G.; Kuster, W. C.; Ryerson, T. B.;
Trainer, M.; Young, V. Geophys. Res. Lett. 1997, 24, 1099.
(14) Williams, J.; Roberts, J. M.; Bertman, S. B.; Stroud, C. A.;
Fehsenfeld, F. C.; Baumann, K.; Buhr, M. P.; Knapp, K.; Murphy, P. C.;
Nowick, M.; Williams, E. J. J. Geophys. Res. 2000, 105, 28943.
(15) Nouaime, G.; Bertman, S. B.; Seaver, C.; Elyea, D.; Huang, H.;
Shepson, P. B.; Starn, T. K.; Riemer, D. D.; Zika, R. G.; Olszyna, K. J.
Geophys. Res. 1998, 103, 22463.
(16) Roberts, J. M.; Williams, J.; Baumann, K.; Buhr, M. P.; Goldan,
P. D.; Holloway, J.; Hu¨bler, G.; Kuster, W. C.; McKeen, S. A.; Ryerson,
T. B.; Trainer, M.; Willliams, E. J.; Fehsenfeld, F. C.; Bertman, S. B.;
Nouaime, G.; Seaver, C.; Grodzinsky, G.; Rodgers, M.; Young, V. L. J.
Geophys. Res. 1998, 103, 22473.
(17) Starn, T. K.; Shepson, P. B.; Bertman, S. B.; White, J. S.; Splawn,
B. G.; Riemer, D. D.; Zika, R. G.; Olszyna, K. J. Geophys. Res. 1998, 103,
22425.
(18) Starn, T. K.; Shepson, P. B.; Bertman, S. B.; Riemer, D. D.; Zika,
R. G.; Olszyna, K. J. Geophys. Res. 1998, 103, 22437.
(19) Hansel, A.; Wisthaler, A. Geophys. Res. Lett. 2000, 27, 895.
(20) Orlando, J. J.; Tyndall, G. S.; Bertman, S. B.; Chen, W.; Burkholder,
J. B. Atmos. EnViron. 2002, 36, 1895.
(21) Maldotti, A.; Chiorboli, C.; Bignozzi, C. A.; Bartocci, C.; Carassiti,
V. Int. J. Chem. Kinet. 1980, 12, 905.
(22) Kerr, J. A.; Sheppard, D. W. EnViron. Sci. Technol. 1981, 15, 960.
(23) Atkinson, R.; Aschmann, S. M.; Pitts, J. N., Jr. Int. J. Chem. Kinet.
1983, 15, 75.
(24) Edney, E. O.; Kleindienst, T. E.; Corse, E. W. Int. J. Chem. Kinet.
1986, 18, 1355.
(25) Magneron, I.; The´venet, R.; Mellouki, A.; Le Bras, G.; Moortgat,
G. K.; Wirtz, K. J. Phys. Chem. A 2002, 106, 2526.
(26) Grosjean, E.; Williams, E. L., II; Grosjean, D. Sci. Total EnViron.
1994, 153, 195.
(27) Tanimoto, H.; Akimoto, H. Geophys. Res. Lett. 2001, 28, 2831.
(28) Roberts, J. M.; Flocke, F.; Weinheimer, A.; Tanimoto, H.; Jobson,
B. T.; Riemer, D.; Apel, E.; Atlas, E.; Donnelly, S.; Stroud, V.; Johnston,
K.; Weaver, R.; Fehsenfeld, F. C. Geophys. Res. Lett. 2001, 28, 4195.
(29) Roberts, J. M.; Bertman, S. B. Int. J. Chem. Kinet. 1992, 24, 297.
(30) Shetter, R. E.; Davidson, J. A.; Cantrell, C. A.; Calvert, J. G. ReV.
Sci. Instrum. 1987, 58, 1427.
product acetaldehyde) may result from (R24b), decomposition
of the CH3-CH(OH)-CH(O)-CHO radical to CH3CHOH and
glyoxal. This reaction may be more favorable than the analogous
processes in the oxidation schemes of acrolein, (R12b), or
methacrolein,
HOCH2-C(CH3)(O)-CHO f CH2OH + CH3C(O)CHO
(R35)
because of the nature of the leaving groups involved (CH3CHOH
in the case of crotonaldehyde, CH2OH in the acrolein and
methacrolein case). Theoretical calculations would be useful in
resolving this issue.
Atmospheric Speculation
The atmospheric fate and impact of both acrolein and
crotonaldehyde have been discussed in the recent paper by
Magneron et al.25 Reaction with OH is the dominant tropo-
spheric loss process for both species (lifetime a few hrs. for
typical daytime [OH] ) (2-3) × 106 molecule cm-3), with
photolysis and reaction with either ozone or NO3 expected to
play only a minor role. The suite of oxygenated products
generated in the OH-initiated oxidations of the unsaturated
aldehydes (formaldehyde and glycolaldehyde from acrolein, and
glyoxal, acetaldehyde, and 2-hydroxypropanal from crotonal-
dehyde) are all expected to be reasonably short-lived as
well,25,31,32,34,52 as reaction with OH and/or photolysis present
efficient removal processes for these species. Lifetimes ranging
from a few hours for acetaldehyde to less than 1 h for glyoxal
are likely. Thus, these oxygenates will contribute to ozone and
HOx production close to the acrolein/crotonaldehyde source
region.
The branching ratio data for reaction of OH with acrolein
and crotonaldehyde presented herein provide key information
required to assess the source strength of the unsaturated PAN-
analogues (APAN and CPAN). Our data show that the acyl
peroxy radical, CH2dCH-C(O)O2, will be generated with about
70% yield from the OH/acrolein reaction, while the correspond-
ing CH3-CHdCH-C(O)O2 radical will be generated in about
50% yield from reaction of OH with crotonaldehyde. Formation
of APAN (or CPAN) will also be controlled by the relative
rate coefficients for (R11) and (R8) (or (R28) and (R29)),
expected to be about 0.5 by analogy to the acetylperoxy
radical,52 as well as by the ambient [NO2]/[NO] ratio, typic-
ally 2-8 in the continental boundary layer. By analogy to
MPAN,20,29 OH reaction and thermal decomposition are likely
to be the major loss processes for APAN and CPAN. Since the
-C(O)O2NO2 group does not appear to be strongly deactivat-
ing,20 rate coefficients with OH on the order of (1-2) × 10-11
cm3 molecule-1 s-1 for APAN, and (3-4) × 10-11 cm3
molecule-1 s-1 for CPAN are likely.
(31) Bacher, C.; Tyndall, G. S.; Orlando, J. J. J. Atmos. Chem. 2001,
39, 171.
(32) Orlando, J. J., Tyndall, G. S. Int. J. Chem. Kinet. 2001, 33, 149.
(33) Orlando, J. J.; Tyndall, G. S.; Ceazan, N. J. Phys. Chem. A 2001,
105, 3564.
(34) Niki, H.; Maker, P. D.; Savage, C. M.; Hurley, M. D. J. Phys.
Chem. 1987, 91, 2174.
(35) Tuazon, E. C.; Atkinson, R. Int. J. Chem. Kinet. 1989, 21, 1141.
(36) Taylor, W. D.; Allston, T. D.; Moscato, M. J.; Fazekas, G. B.;
Kozlowski, R.; Takacs, G. A. Int. J. Chem. Kinet. 1980, 12, 231.
(37) Meagher, R. J.; McIntosh, M. E.; Hurley, M. D.; Wallington, T. J.
Int. J. Chem. Kinet. 1997, 29, 619.
(38) Tsalkani, N.; Toupance, G. Atmos. EnViron. 1989, 23, 1849.
(39) Graham, R. A.; Winer, A. M.; Pitts Jr., J. N. Geophys. Res. Lett.
1978, 5, 909.
Acknowledgment. The National Center for Atmospheric
Research is operated by the University Corporation for Atmo-
spheric Research under the sponsorship of the National Science
Foundation. We are indebted to Chris Cantrell and Eric Apel
of NCAR for their insightful comments on the manuscript.
(40) Wang, W.; Ezell, M. J.; Ezell, A. A.; Soskin, G.; Finlayson-Pitts,
B. J. Phys. Chem. Chem. Phys., 2002, 4, 1824.
(41) Atkinson, R. J. Phys. Chem. Ref. Data 1997, 26, 215.
(42) Atkinson, R. Int. J. Chem. Kinet. 1997, 29, 99.
(43) Fahr, A.; Laufer, A. H. J. Phys. Chem. 1988, 92, 7229.
(44) Park, J.-Y.; Heaven, M. C., Gutman, D. Chem. Phys. Lett. 1984,
104, 469.
References and Notes
(1) Tuazon, E. C.; Atkinson, R. Int. J. Chem. Kinet. 1990, 22, 1221.
(2) Tuazon, E. C.; Alvarado, A.; Aschmann, S. M.; Atkinson, R.; Arey,
J. EnViron. Sci. Technol. 1999, 33, 3586.
(3) Liu, X.; Jeffries, H. E.; Sexton, K. G. Atmos. EnViron. 1999, 33,
3005.
(45) Slagle, I. R.; Park, J.-Y.; Heaven, M. C.; Gutman, D. J. Am. Chem.
Soc. 1984, 106, 4356.

Reactions of OH with Unsaturated Aldehydes
J. Phys. Chem. A, Vol. 106, No. 51, 2002 12259
(46) Krueger, H. Weitz, E., J. Chem. Phys. 1988, 88, 1608.
(47) Knyazev, V. D.; Slagle, I. R. J. Phys. Chem. 1995, 99, 2247.
(48) Wang, H., Wang, B.; He, Y.; Kong, F. J. Chem. Phys. 2001, 115,
1742.
(49) Fahr, A.; Laufer, A. H.; Krauss, M.; Osman, R. J. Phys. Chem.
1997, 101, 4879.
(50) Gutman, D.; Nelson, H. H. J. Phys. Chem. 1985, 89, 3902.
(51) Zhu, L.; Johnston, G. J. Phys. Chem. 1995, 99, 15114.
(52) Atkinson, R.; Baulch, D. L.; Cox, R. A.; Hampson, R. F.; Kerr, J.
A.; Rossi, M. J.; Troe, J. J. Phys. Chem. Ref. Data 1999, 28, 191.
(53) Atkinson, R. Int. J. Chem. Kinet. 1987 19, 799.
(54) Kwok, E. S. C.; Atkinson, R. Atmos. EnViron. 1995, 29, 1685.