Products of the OH radical-initiated reactions of furan, 2- and 3-methylfuran, and 2,3- and 2,5-dimethylfuran in the presence of NO

Article abstract of DOI:10.1021/jp410345k

Products of the gas-phase reactions of OH radicals with furan, furan-d ₄, 2- and 3-methylfuran, and 2,3- and 2,5-dimethylfuran have been investigated in the presence of NO using direct air sampling atmospheric pressure ionization tandem mass spectrometry (API-MS and API-MS/MS), and gas chromatography with flame ionization and mass spectrometric detectors (GC-FID and GC-MS) to analyze samples collected onto annular denuders coated with XAD solid adsorbent and further coated with O-(2,3,4,5,6-pentafluorobenzyl) hydroxylamine for derivatization of carbonyl-containing compounds to their oximes. The products observed were unsaturated 1,4-dicarbonyls, unsaturated carbonyl-acids and/or hydroxy-furanones, and from 2,5-dimethylfuran, an unsaturated carbonyl-ester. Quantification of the unsaturated 1,4-dicarbonyls was carried out by GC-FID using 2,5-hexanedione as an internal standard, and the measured molar formation yields were: HC(O)CH=CHCHO (dominantly the E-isomer) from OH + furan, 75 ± 5%; CH₃C(O)CH=CHCHO (dominantly the E-isomer) from OH + 2-methylfuran, 31 ± 5%; HC(O)C(CH₃)=CHCHO (a E-/Z-mixture) from OH + 3-methylfuran, 38 ± 2%; and CH ₃C(O)C(CH₃)=CHCHO from OH + 2,3-dimethylfuran, 8 ± 2%. In addition, a formation yield of 3-hexene-2,5-dione from OH + 2,5-dimethylfuran of 27% was obtained from a single experiment, in good agreement with a previous value of 24 ± 3% from GC-FID analyses of samples collected onto Tenax solid adsorbent without derivatization.

Full text of DOI:10.1021/jp410345k

Article
pubs.acs.org/JPCA
Products of the OH Radical-Initiated Reactions of Furan, 2- and
3‑Methylfuran, and 2,3- and 2,5-Dimethylfuran in the Presence of NO
Sara M. Aschmann, Noriko Nishino,† Janet Arey,*,‡ and Roger Atkinson*,‡
Air Pollution Research Center, University of California, Riverside, California 92521, United States
S
* Supporting Information
ABSTRACT: Products of the gas-phase reactions of OH radicals with furan, furan-d4, 2- and
3-methylfuran, and 2,3- and 2,5-dimethylfuran have been investigated in the presence of NO
using direct air sampling atmospheric pressure ionization tandem mass spectrometry (API-MS
and API-MS/MS), and gas chromatography with flame ionization and mass spectrometric
detectors (GC−FID and GC−MS) to analyze samples collected onto annular denuders
coated with XAD solid adsorbent and further coated with O-(2,3,4,5,6-pentafluorobenzyl)-
hydroxylamine for derivatization of carbonyl-containing compounds to their oximes. The
products observed were unsaturated 1,4-dicarbonyls, unsaturated carbonyl-acids and/or
hydroxy-furanones, and from 2,5-dimethylfuran, an unsaturated carbonyl-ester. Quantification
of the unsaturated 1,4-dicarbonyls was carried out by GC−FID using 2,5-hexanedione as an
internal standard, and the measured molar formation yields were: HC(O)CHCHCHO
(dominantly the E-isomer) from OH + furan, 75
5%; CH3C(O)CHCHCHO
(dominantly the E-isomer) from OH + 2-methylfuran, 31 5%; HC(O)C(CH3)CHCHO (a E-/Z-mixture) from OH +
3-methylfuran, 38 2%; and CH3C(O)C(CH3)CHCHO from OH + 2,3-dimethylfuran, 8 2%. In addition, a formation
yield of 3-hexene-2,5-dione from OH + 2,5-dimethylfuran of 27% was obtained from a single experiment, in good agreement with
a previous value of 24 3% from GC−FID analyses of samples collected onto Tenax solid adsorbent without derivatization.
OH radicals were generated by the photolysis of methyl nitrite
(CH3ONO), and NO was also included in the reactant
mixtures to suppress the formation of O3 and hence of NO3
radicals. Unless noted otherwise, all irradiations were carried
out at a light intensity corresponding to an NO2 photolysis rate
of 0.14 min−1.
INTRODUCTION
■
Furan and alkylfurans are emitted directly into the atmosphere
as well as being formed in situ from atmospheric degradation of
certain other volatile organic compounds.1−3 Furan and
alkylfurans are transformed further in the atmosphere by
reactions with OH radicals, NO3 radicals, O3 and Cl atoms,4
and the products of their OH radical-initiated reactions include
unsaturated 1,4-dicarbonyls.5−8 These reactions may therefore
be useful as potentially clean in situ laboratory sources of
unsaturated 1,4-dicarbonyls, which are otherwise not readily
available.6,7 In this work, we have extended our previous
measurements of the rate constants for the reactions of OH
radicals with 2- and 3-methylfuran and 2,3- and 2,5-
dimethylfuran and of the formation yield of 3-hexene-2,5-
dione from OH + 2,5-dimethylfuran7 to investigate the
products formed from OH + furan, 2- and 3-methylfuran and
2,3-dimethylfuran in the presence of NO. Additional product
analyses are presented for OH + 2,5-dimethylfuran for
comparison with the furan, 2- and 3-methylfuran and 2,3-
dimethylfuran reactions. The loss processes of the unsaturated
1,4-dicarbonyls, which are also products of OH + aromatic
hydrocarbon reactions,9−14 were also explored.
Analyses by Gas Chromatography. A series of
CH3ONO−NO−furan−air irradiations were carried out, with
initial concentrations (molecules cm−3) of CH3ONO and NO,
∼1.2 × 1014 each; furan, (1.91−2.54) × 1013; and 2,5-
hexanedione (used as an internal standard; see below), (6.10−
8.50) × 1012. Each experiment involved a single irradiation of
duration 1.5−8 min, resulting in 35−64% consumption of the
initially present furan. Three preliminary experiments with 2,3-
dimethylfuran were carried out with similar initial CH3ONO,
NO and 2,3-dimethylfuran concentrations, with the 2,5-
hexanedione internal standard being present initially in one
experiment (2.27 × 1012 molecules cm−3) and being added
after the irradiation in the other two (2.30 × 1012 and 8.99 ×
1012 molecules cm−3). In these three experiments, the mixtures
were irradiated for 3−4 min, with 49−61% consumption of 2,3-
dimethylfuran.
The concentrations of the furans and 2,5-hexanedione (and
3-hexene-2,5-dione in the OH + 2,5-dimethylfuran reaction)
were measured before and after the irradiation by gas
EXPERIMENTAL METHODS
■
Experiments were carried out at 296
2 K and ∼735 Torr
pressure of dry purified air in ∼7000 L Teflon chambers, each
equipped with two parallel banks of black lamps for irradiation
at >300 nm and with a Teflon-coated fan to ensure rapid
mixing during the introduction of chemicals into the chamber.
Received: October 18, 2013
Revised: December 12, 2013
Published: December 31, 2013
© 2013 American Chemical Society
457
dx.doi.org/10.1021/jp410345k | J. Phys. Chem. A 2014, 118, 457−466

The Journal of Physical Chemistry A
Article
Scheme 1
chromatography with flame ionization detection (GC−FID).
Gas samples of 100 cm3 volume were collected from the
chamber onto Tenax-TA solid adsorbent, with subsequent
thermal desorption at ∼205 °C onto a 30 m DB-1701
megabore column, temperature programmed from −40 °C at 8
°C min−1. For the analysis of furan, gas samples were collected
from the chamber into a 100 cm3 volume all-glass gastight
syringe and transferred via a 1 cm3 gas sampling loop onto a 30
m DB-5 megabore column initially held at −25 °C and then
temperature programmed at 8 °C min−1. Replicate analyses
before and after the irradiations typically agreed to within 3%.
Samples were collected, starting immediately after the lights
were turned off, for 60 min at 15 L min−1 using an XAD-coated
denuder, further coated with O-(2,3,4,5,6-pentafluorobenzyl)-
hydroxylamine (PFBHA) prior to sampling,13,15 and extracted
as described previously.13 The extracts were analyzed by
positive chemical ionization gas chromatography−mass spec-
trometry (PCI GC−MS) and GC−FID, both analyses using 30
m DB-5 columns. The PCI GC−MS analyses used an Agilent
6890N GC interfaced to an Agilent 5975 Inert XL Mass
Selective Detector operated with methane as the reagent gas.
Each carbonyl group derivatized to an oxime added 195 mass
units to the compound’s molecular weight (mw), and methane-
PCI gave characteristic protonated molecules ([M+H]+) and
smaller adduct ions at [M+29]+ and [M+41]+.13
An analogous CH3ONO−NO−3-methyl-2-butenal−air irra-
diation was carried out, with 2,5-hexanedione present as an
internal standard, to measure the glyoxal formation yield from
OH + 3-methyl-2-butenal and compare it with the yield
previously measured using in situ Fourier transform infrared
(FTIR) spectroscopy.16 The initial concentrations (molecules
cm−3) were CH3ONO and NO, ∼1.2 × 1014 each; 3-methyl-2-
butenal, 2.11 × 1013; and 2,5-hexanedione, 4.06 × 1012. The
mixture was irradiated for 3 min, resulting in 50% reaction of
the initially present 3-methyl-2-butenal. The concentrations of
3-methyl-2-butenal and 2,5-hexanedione were measured by
GC−FID analyses of samples collected onto Tenax solid
adsorbent, and those of the dioximes of 2,5-hexanedione and
glyoxal were measured by GC−FID analyses of an extract of a
PFBHA-coated denuder sample, as described above.
Analyses by API-MS/MS. CH3ONO−NO−furan−air
irradiations were carried out during which the chamber
contents were sampled through a 25 mm diameter × 75 cm
length Pyrex tube at ∼20 L min−1 directly into the API-MS
source. The operation of the API-MS in the MS (scanning) and
MS/MS [with collision activated dissociation (CAD)] modes
has been described previously.7,17 Both positive and negative
ion modes were used in this work. In positive ion mode,
protonated water hydrates (H3O+(H2O)n) generated by the
corona discharge in the chamber diluent air were responsible
458
dx.doi.org/10.1021/jp410345k | J. Phys. Chem. A 2014, 118, 457−466

The Journal of Physical Chemistry A
Article
Table 1. Products Formed from the Furans Studied from API-MS and API-MS/MS Analyses
molecular weight of potential products (Scheme 1)
furan
pathway A
84
pathway B
100
pathway C
100
pathway D
100
observed products from API-MS analyses and ion peak attribution
a,b
furan
84, HC(O)CHCHCHO
a,c
d
100, HC(O)CHCHC(O)OH and/or isomer
a
furan-d4
88
98
103
103
104
114
88, DC(O)CDCDCDO
a,
c
d
103, DC(O)CDCDC(O)OH and/or isomer
a,b
2-methylfuran
100, 114
100, 114
98, CH3C(O)CHCHCHO
a,c
d
100, HC(O)CHCHC(O)OH and/or isomer
a,
c
d
114, CH3C(O)CHCHC(O)OH and/or isomer
a,e
3-methylfuran
98
114
114
114
128
98, HC(O)C(CH3)CHCHO
a,c
d
114 HC(O)C(CH3)CHC(O)OH and/or isomer
a,e
2,3-dimethylfuran
112
114, 128
114, 128
112, CH3C(O)C(CH3)CHCHO
a,c
d
114, HC(O)CC(CH3)C(O)OH and/or isomer
a,
c
d
128, CH3C(O)C(CH3)CHC(O)OH and/or isomer
a,e
2,5-dimethylfuran
112
114
114
128
112 CH3C(O)CHCHC(O)CH3
a,c
d
114, CH3C(O)CHCHC(O)OH and/or isomer
a
128, CH3C(O)OCHCHC(O)CH3
a
b
−
d
Observed in positive ion mode. Small NO2 adduct peak observed in negative ion mode (see Figure 2 and its caption for the case of 2-
methylfuran). Observed in negative ion mode. See Scheme 1, pathways B and C. Small ion peak present in negative ion mode which could have
been the NO2 adduct ion.
c
e
−
for the formation of protonated molecules ([M+H]+), water
adduct ions [M+H+H2O]+, and protonated homo- and
and D in Scheme 1. Table 1 lists the molecular weights of the
potential products formed from these four pathways, noting
that for OH radical addition at the 2- or 5-positions, different
molecular weight products are formed from 2-methylfuran and
2,3-dimethylfuran through pathways B and C depending on
whether the initial OH radical addition is at the 2- or 5-
position. OH radical addition at the 3- or 4-positions would
lead to the same carbonyl-ester products as formed after
addition at the 2- or 5-positions, respectively, via pathway D
(see Supporting Information to ref 7).
Analysis by API-MS. In positive ion mode prior to reaction,
ion peaks due to the furans were observed at M+, [M+H]+ and
[M+NO]+. Additional ion peaks were observed after reaction,
and API-MS spectra from the reactions of furan, 2- and 3-
methylfuran and 2,3-dimethylfuran during the initial stages of
the reactions are shown in Figure 1. An analogous API-MS
spectrum from the 2,5-dimethylfuran reaction was presented in
Aschmann et al.7 The assignments of these ion peaks, based on
API-MS/MS spectra, are noted in the caption to Figure 1 and
the molecular weights of the products observed, and whether
they were observed in negative ion mode as well as in positive
ion mode, are listed in Table 1. An example of the negative ion
mode API-MS spectra obtained is shown in Figure 2 for the
OH + 2-methylfuran reaction, and the ion peak assignments are
given in the caption to Figure 2. The API-MS negative ion
mode spectra for furan, furan-d4, 3-methylfuran, and 2,3- and
2,5-dimethylfuran were analogous, with the ion peaks being
assigned to either one (furan, furan-d4, 3-methylfuran, and 2,5-
dimethylfuran) or two (2-methylfuran and 2,3-dimethylfuran)
product masses (Table 1).
heterodimers,7,17 while in negative ion mode O2 , NO2 and
−
−
NO3 ions were responsible for formation of adduct ions.17
−
The initial concentrations and experimental conditions were
similar to those of the product studies, except that 2,5-
hexanedione was not present.
Chemicals. The chemicals used, and their stated purities,
were 2,3-dimethylfuran (99%), 2,5-dimethylfuran (99%), furan
(99+%), 2,5-hexanedione (98+%), 3-methyl-2-butenal (97%),
2-methylfuran (99%), O-(2,3,4,5,6-pentafluorobenzyl)-
hydroxylamine hydrochloride (>98%), 1,2,3-trimethylbenzene
(90%), o-xylene (97%), and p-xylene (99+%), Aldrich; E-3-
hexene-2,5-dione (GC−MS analyses of its dioximes in samples
collected from the chamber onto a PFBHA-coated denuder
showed no detectable Z-isomer), Hestia Laboratories Inc.;
furan-d4, (98 atom % D), Isotec; 3-methylfuran (98%), Arcos
Organic; 5-hydroxy-2-pentanone (≥96%), TCI America; and
NO (≥99.0%), Matheson Gas Products. Methyl nitrite was
prepared as described by Taylor et al.18 and stored at 77 K
under vacuum.
RESULTS AND DISCUSSION
■
Our recent product study of the OH + 2,5-dimethylfuran
reaction7 showed the formation of the unsaturated 1,4-
dicarbonyl 3-hexene-2,5-dione (mw 112) plus additional
products of mw 114 and mw 128 in both the presence and
absence of NOx, with the 3-hexene-2,5-dione yield being higher
in the absence of NOx (34 3%) than in the presence of NOx
(24 3%, independent of initial NO concentration over the
range (2.4−24) × 1013 molecules cm−3).7 The formation of 1,4-
butenedial from OH + furan,5,6 4-oxo-2-pentenal from OH + 2-
methylfuran,5,6 2-methyl-1,4-butenedial from OH + 3-methyl-
furan,6 and 3-hexene-2,5-dione from OH + 2,5-dimethylfuran7
implies that these unsaturated 1,4-dicarbonyls are formed after
initial OH addition at the 2- or 5-positions (pathway A in
Scheme 1).5−7 However, the formation of additional products
from OH + 2,5-dimethylfuran and the effect of NOx on the 3-
hexene-2,5-dione yield showed that additional reaction path-
ways are operative,7 and these are denoted as pathways B, C
All of the furans studied here show the presence of product
masses consistent with those expected from pathway A and
pathways B and/or C in Scheme 1, noting that pathways B and
C lead to the same product masses. Pathway D can only be
discriminated from pathways B and C for the furan-d4 and 2,5-
dimethylfuran reactions (Table 1). The API-MS spectra
showed evidence for the occurrence of pathway D in the OH
+ 2,5-dimethylfuran reaction (ref 7 and Table 1), but not in the
OH + furan-d4 reaction. The products attributed to formation
via pathways B and/or C were observed in both positive and
459
dx.doi.org/10.1021/jp410345k | J. Phys. Chem. A 2014, 118, 457−466

The Journal of Physical Chemistry A
Article
Figure 2. Negative mode API-MS spectrum of an irradiated
CH3ONO−NO−2-methylfuran−air mixture. The ion peaks are
attributed to: m/z 132, [100+O2]−; m/z 144, [98+NO2]−; m/z 146,
[100+NO2]− and [114+O2]−; m/z 160, [114+NO2]−; m/z 162,
[100+NO3]−; m/z 174, [114+60]−; m/z 176, [114+NO3]−; m/z 199,
[100+100−H]−; m/z 213, [100+114−H]−; m/z 227, [114+114−H]−;
m/z 246, [100+114+O2]− and [100+100+NO2]−; and m/z 260,
[114+114+O2]− and [100+114+NO2]−.
Figure 1. Positive mode API-MS spectra of irradiated CH3ONO−
NO−furan−air mixtures. The ion peaks are attributed to: OH + furan,
m/z 85, [84+H]+; m/z 101, [100+H]+; m/z 119, [100+H+H2O]+; m/
z 185, [84+100+H]+; and m/z 201, [100+100+H]+; OH + 2-
methylfuran, m/z 99, [98+H]+; m/z 101, [100+H]+; m/z 115,
[114+H]+; m/z 117, [98+H+H2O]+; m/z 119, [100+H+H2O]+; m/z
133, [114+H+H2O]+; m/z 199, [98+100+H]+; m/z 201,
[100+100+H]+; m/z 213, [98+114+H]+; m/z 215, [100+114+H]+;
and m/z 229, [114+114+H]+; OH + 3-methylfuran, m/z 99, [98+H]+;
m/z 115, [114+H]+; m/z 133, [114+H+H2O]+; m/z 213,
[98+114+H]+; and m/z 229, [114+114+H]+; OH + 2,3-dimethylfur-
an, m/z 97, [2,3-dimethylfuran+H]+; m/z 113, [112+H]+; m/z 115,
[114+H]+; m/z 129, [128+H]+; m/z 131, [112+H+H2O]+; m/z 133,
[114+H+H2O]+; m/z 147, [128+H+H2O]+; m/z 225, [112+112+H]+;
m/z 227, [112+114+H]+; m/z 229, [114+114+H]+; m/z 241,
[112+128+H]+; m/z 243, [114+128+H]+; and m/z 257,
[128+128+H]+.
negative ion modes, whereas those from pathways A and D
were observed mainly in positive ion mode (see footnotes to
Table 1).
Analyses by GC−MS. GC−MS analyses of extracts of
samples collected using PFBHA-coated denuders showed the
presence of the expected unsaturated 1,4-dicarbonyls (pathway
A in Scheme 1, and Table 1). A PCI GC−MS total ion
chromatogram (TIC) is shown in Figure 3 for the furan
reaction, and Figures S1−S4 in the Supporting Information
show the PCI GC−MS TICs for 2- and 3-methylfuran and 2,3-
and 2,5-dimethylfuran. Most experiments were carried out with
the internal standard 2,5-hexanedione present in the initial
reactant mixtures (i.e., before reaction). The rate constant for
the reaction of OH radicals with 2,5-hexanedione has been
measured to be (7.13 0.34) × 10−12 cm3 molecule−1 s−1 at
298 K,19 and hence 2,5-hexanedione is less reactive toward OH
radicals than the furans studied here, by factors of 6−18.4,7
Based on the measured concentrations of 2,5-hexanedione
before and after reaction, the percentages of 2,5-hexanedione,
which were consumed by reaction were 11−17% in the furan
reactions and ≤6% in the 2- and 3-methylfuran and 2,3- and
Figure 3. Positive chemical ionization GC−MS total ion chromato-
gram (TIC) of an extract of a PFBHA-coated denuder sample
collected after 4 min irradiation of a CH3ONO−NO−furan−2,5-
hexanedione (internal standard, IS)−air mixture, with 38% con-
sumption of the initially present furan. IS = dioximes of 2,5-
hexanedione, the internal standard; peak #1, mono-oxime of molecular
weight (mw) 92 carbonyl; peak #2, mono-oxime of CH3C(O)-
CH2CHO (mw 86) formed from OH + 2,5-hexanedione, the internal
standard; peaks #3 and #4, from derivatizing agent; peaks #5, dioximes
of glyoxal; peaks #6, dioximes of CH3C(O)CH2CHO (mw 86)
formed from OH + 2,5-hexanedione; and mw 84 dioximes of
HC(O)CHCHCHO.
2,5-dimethylfuran reactions, consistent with expectations based
on the relative reactivities of 2,5-hexanedione and the furans
and the fractions of the furans reacted. The unsaturated 1,4-
dicarbonyls and glyoxal, methylglyoxal, 2,3-butanedione and 3-
460
dx.doi.org/10.1021/jp410345k | J. Phys. Chem. A 2014, 118, 457−466

The Journal of Physical Chemistry A
Article
oxobutanal [CH3C(O)CH2CHO] were identified by matching
GC retention times and mass spectra of their oximes with those
generated in situ from OH radical-initiated reactions of o-
xylene, p-xylene, 1,2,3-trimethylbenzene, and 5-hydroxy-2-
pentanone, which form these compounds.13,20,21 A major
expected product of OH + 2,5-hexanedione is 3-oxobutanal,
and this mw 86 dicarbonyl was observed in all PCI GC−MS
analyses when 2,5-hexanedione was present prior to reaction
(and in much lesser amounts when 2,5-hexanedione was not
present prior to reaction, presumably due to carry-over from
previous experiments).
The products observed from the OH + furan reactions by
GC−MS were then as follows: from furan: 1,4-butenedial
[HC(O)CHCHCHO], glyoxal, and a mw 92 carbonyl; from
2-methylfuran, 4-oxo-2-pentenal [CH3C(O)CHCHCHO],
glyoxal, methylglyoxal, a mw 100 product, and a mw 114
product(s); from 3-methylfuran, 2-methyl-1,4-butenedial [HC-
(O)C(CH3)CHCHO], glyoxal, methylglyoxal, and mw 114
product(s); from 2,3-dimethylfuran, 3-methyl-4-oxo-2-pentenal
[CH3C(O)C(CH3)CHCHO], glyoxal, methylglyoxal, biace-
tyl, mw 60 carbonyl (most likely a hydroxycarbonyl and hence
glycolaldehyde, HOCH2CHO), mw 114 carbonyl, plus other
unidentified peaks; and from 2,5-dimethylfuran, 3-hexene-2,5-
dione [CH3C(O)CHCHC(O)CH3], glyoxal, methylglyoxal,
and mw 114 carbonyl. Based on the PCI GC−MS TICs,
glyoxal, methylglyoxal and 2,3-butanedione formation was
minor in all cases, and could have arisen as second-generation
products.5 Bierbach et al.5 also observed the formation of
glyoxal from OH + furan (with a formation yield of 8.2% in the
initial stages of reaction, decreasing to 4.4% at higher extents of
reaction), and of glyoxal and methylglyoxal from OH + 2-
methylfuran (with formation yields of 1.5−4.8% and 1.6−2.3%,
respectively, both increasing with the extent of reaction), both
reactions being studied in the absence of NOx.
Figure 4. Positive chemical ionization GC−MS analyses of extracts of
PFBHA-coated denuder samples collected after CH3ONO−NO−air
irradiations. 1A: m/z = 475 [M+H]+ of dioximes of HC(O)CH
CHCHO from a furan reaction after 6 min irradiation; 1B: m/z = 475
[M+H]+ of dioximes of HC(O)CHCHCHO from an o-xylene
reaction after 4 min irradiation; 2A: m/z = 489 [M+H]+ of dioximes of
CH3C(O)CHCHCHO from a 2-methylfuran reaction after 2 min
irradiation; and 2B: m/z = 489 [M+H]+ of dioximes of CH3C(O)-
CHCHCHO from an o-xylene reaction after 4 min irradiation.
The dioxime profiles from the GC−MS analyses can be
compared to those for the reactions of OH radicals with
aromatic hydrocarbons,7,13,20 where the unsaturated 1,4-
dicarbonyls are assumed to be formed in the Z-configuration
based on the observation that 3-hexene-2,5-dione is formed
from OH + p-xylene and 1,2,4-trimethylbenzene in the Z-
configuration,7,11 recognizing that Z-/E-photoisomerization
occurs and that short irradiation times are hence necessary to
maintain the initial Z-/E-distribution. For dicarbonyls not
containing a CC bond, there are up to four dioximes,
depending on symmetry, these being syn,syn; syn,anti; anti,syn
and anti,anti. For the unsaturated 1,4-dicarbonyls that exist as
Z- and E-isomers, there are up to eight dioximes, with six for
those with a center of symmetry such as 1,4-butenedial and 3-
hexene-2,5-dione, noting that because of coelution and/or
dominance of certain configurations, not all possible dioximes
may be observed.13
Methyl-4-oxo-2-pentenal from OH + 2,3-dimethylfuran also
appears to be formed in a Z-/E-mixture (Figure 5, traces 2A
and 2B), although this interpretation is less straightforward
than for the furan and 2- and 3-methylfuran reactions. We
previously found,7 from GC−FID measurements of under-
ivatized 3-hexene-2,5-dione, that in the presence of NO the OH
+ 2,5-dimethylfuran reaction forms 3-hexene-2,5-dione in the
ratio 69 4% Z/31 4% E. In the presence of NO, there
appears to be a gradual change from dominantly E-isomer
formation of 1,4-butenedial from OH + furan to mainly Z-
isomer formation of 3-hexene-2,5-dione from OH + 2,5-
dimethylfuran. Our observation of mainly E unsaturated 1,4-
dicarbonyl formation from OH + furan and 2-methylfuran is in
agreement with the previous in situ Fourier transform infrared
(FT-IR) spectral data of Bierbach et al.5 for OH + furan and 2-
methylfuran and the solid phase microextraction (SPME) fiber
analyses of Gom
́
ez Alvarez et al.6 for OH + furan.
Figures 4 and 5 compare the dioxime profiles from the furan
reactions with those from OH + aromatic experiments
conducted at low extents of reaction. Assuming that the
unsaturated 1,4-dicarbonyls formed from the OH + aromatic
reactions are totally or almost totally in the Z-form (recognizing
that some Z-/E-photoisomerization will occur, even for short
irradiation times), it is obvious that 1,4-butenedial from the OH
+ furan reaction is formed dominantly as the E-isomer (Figure
4, traces 1A and 1B), 4-oxo-2-pentenal from OH + 2-
methylfuran is mainly in the E-form (Figure 4, traces 2A and
2B), and 2-methyl-1,4-butenedial from OH + 3-methylfuran is
formed in a Z-/E-mixture (Figure 5, traces 1A and 1B). 3-
Quantification of the Unsaturated 1,4-Dicarbonyls.
Concentrations Measured from Denuder Samples with
Internal Standards. Since unsaturated 1,4-dicarbonyls, other
than diketones like 3-hexene-2,5-dione7,11 and 3-methyl-3-
hexene-2,5-dione,14 do not appear to elute from GC columns
without prior derivatization, the unsaturated 1,4-dicarbonyls
were quantified as their dioximes. For the 3-methylfuran and
2,3-dimethylfuran reactions, minor amounts of mono-oximes
were also observed and quantified. 2,5-Hexanedione was used
as an internal standard, with its concentration in the chamber
being measured by GC−FID analyses of samples collected onto
Tenax solid adsorbent (without derivatization) and with it
461
dx.doi.org/10.1021/jp410345k | J. Phys. Chem. A 2014, 118, 457−466

The Journal of Physical Chemistry A
Article
and derivatization efficiencies of 2,5-hexanedione and glyoxal
were essentially identical and that the use of the calculated
ECNs was appropriate.
The before and after reaction concentrations of the furans
and the after reaction concentrations of the unsaturated 1,4-
dicarbonyls are listed in Table S1 (Supporting Information) (in
parts-per-billion (ppbV) mixing ratio; 1 ppbV = 2.40 × 1010
molecules cm−3 for the temperature and total pressure
conditions of this study). As evident from Table S1, up to
five replicate postreaction GC−FID analyses of the extract of
the PFBHA-coated denuder sample were conducted, with
excellent reproducibility relative to the internal standard.
Corrections for Losses of Unsaturated 1,4-Dicarbonyls.
The unsaturated 1,4-dicarbonyls also react with OH radicals
and undergo photolysis, and these unsaturated dicarbonyl loss
processes need to be taken into account in order to obtain their
formation yields. For a given furan, the OH radical
concentrations averaged over the irradiation period were
independent of the irradiation time, with ≤13% change in the
OH radical concentration for a factor of 2 change in the
irradiation time (see the denuder samples in Table S2
(Supporting Information), for which the irradiation times
were varied for identical initial CH3ONO and NO concen-
trations). Hence the OH radical concentrations could be taken
to be constant during an individual experiment, and corrections
for secondary reactions were carried out as described
previously,24 with k(unsaturated 1,4-dicarbonyl)/k(furan) =
k2/k1 = {k(OH + unsaturated 1,4-dicarbonyl)[OH] +
k(photolysis of unsaturated 1,4-dicarbonyl) + k(wall loss of
unsaturated 1,4-dicarbonyl)}/k(OH + furan)[OH], and with
the correction factors depending on the ratio k2/k1 and the
extent of reaction.24
Rate constants for the OH radical reactions of the Z and E
unsaturated 1,4-dicarbonyls and their rates of photolysis by
blacklamps are, however, not well-known. Rate constants have
been measured for the reactions of OH radicals with 1,4-
butenedial, 4-oxopentenal and 3-hexene-2,5-dione, with room
temperature rate constants of (in units of 10−11 cm3 molecule−1
s−1): Z-1,4-butenedial, 5.29 0.11;25,26 E-1,4-butenedial, 3.45
0.34;27 Z-/E-mixture of 4-oxo-2-pentenal, 5.67 0.22;25,26
Z-3-hexene-2,5-dione, 5.90 0.57;7 and E-3-hexene-2,5-dione,
4.14 0.02.7
We obtained additional information concerning the rate
ratios k2/k1 pertinent to our experiments by analyzing the
time−concentration behavior of the unsaturated 1,4-dicarbon-
yls during OH radical-initiated reactions of the furans. This
included additional experiments comprising continuous irradi-
ations of CH3ONO−NO−furan−air mixtures with analysis of
the unsaturated 1,4-dicarbonyls every ∼3 min by API-MS.
These data are presented and discussed in the Supporting
Information, with Table S2 including data from the experi-
ments with denuder sampling as well as from the additional
experiments with API-MS analyses. On the basis of the available
literature data concerning photolysis and reaction with OH
radicals, and consistent with the values of k2/k1 obtained for
unsaturated 1,4-dicarbonyl formation from our OH + furan
reactions (see Supporting Information), we used rate constants
for the OH radical reactions of (in units of 10−11 cm3
molecule−1 s−1) of: E-1,4-butenedial, 3.45;27 4-oxo-2-pentenal,
5.67;25,26 2-methyl-1,4-butenedial and 3-methyl-4-oxo-2-pente-
nal, 6.0 each (estimated); and 3-hexene-2,5-dione, 5.35
(applicable for a 69% Z-/31% E-mixture7). A photolysis rate
of J(unsaturated 1,4-dicarbonyl)/J(NO2) = 0.14 was used for
Figure 5. Positive chemical ionization GC−MS analyses of extracts of
PFBHA-coated denuder samples collected after CH3ONO−NO−air
irradiations. 1A: m/z = 489 [M+H]+ of dioximes of HC(O)C(CH3)
CHCHO from a 3-methylfuran reaction after 4 min irradiation; 1B:
m/z = 489 [M+H]+ of dioximes of HC(O)C(CH3)CHCHO from a
p-xylene reaction after 4 min irradiation; 2A: m/z = 503 [M+H]+ of
dioximes of CH3C(O)C(CH3)CHCHO from a 2,3-dimethylfuran
reaction after 3 min irradiation; and 2B: m/z = 503 [M+H]+ of
dioximes of CH3C(O)C(CH3)CHCHO from a 1,2,3-trimethylben-
zene (1,2,3-TMB) reaction after 2 min irradiation. The peaks labeled
with asterisks in 2B are not due to CH3C(O)C(CH3)CHCHO (the
first asterisked peak being a m/z = 503 [M+H−H2O]+ of dioximes of a
molecular weight 130 hydroxydicarbonyl, possibly CH3C(O)C(O)C-
(CH3)(OH)C(O)CH3 or isomer formed as a second-generation
product from OH + unsaturated 1,4-dicarbonyl). An additional ion
peak must not be due to CH3C(O)C(CH3)CHCHO, since a
maximum of eight are expected.
being observed in the PFBHA-coated denuder samples as its
dioximes (no mono-oximes of 2,5-hexanedione were observed).
Hence by ratioing the GC−FID peaks areas of the oximes of
the unsaturated 1,4-dicarbonyls to those of 2,5-hexanedione
and taking into account the differing FID responses of the
oximes of the unsaturated 1,4-dicarbonyls and of 2,5-
hexanedione (using the Effective Carbon Numbers (ECNs)
of Scanlon and Willis22 and Nishino et al.23), the concen-
trations of the unsaturated dicarbonyls in the chamber could be
derived.
As a check on this approach, the glyoxal formation yield from
OH + 3-methyl-2-butenal was measured using this procedure.
After correcting for the GC−FID responses of the dioximes of
2,5-hexanedione and glyoxal using their calculated ECNs,22,23
and for reaction of glyoxal with OH radicals,16 the glyoxal
formation yield was determined to be 40
5%, where the
indicated error is two standard deviations of the four replicate
analyses of the extract combined with estimated uncertainties in
the GC−FID response factors for 3-methyl-2-butenal and 2,5-
hexanedione of 5% each. This formation yield is in excellent
agreement with that of 40 3% measured by Tuazon et al.16
using in situ FT-IR spectroscopy, indicating that the collection
462
dx.doi.org/10.1021/jp410345k | J. Phys. Chem. A 2014, 118, 457−466

The Journal of Physical Chemistry A
Article
Table 2. Measured Formation Yields of Unsaturated 1,4-Dicarbonyls from the OH Radical-Initiated Reactions of Furans,
Together with Literature Data
a
furan
unsaturated 1,4-dicarbonyl
molar yield (%)
ref
b
furan
HC(O)CHCHCHO
75
5
this work
Bierbach et al.5
c
>70
d
109 41
Gom
Gom
́
ez Alvarez et al.6
ez Alvarez et al.6
d
90 36
́
b
2-methylfuran
3-methylfuran
CH3C(O)CHCHCHO
HC(O)C(CH3)CHCHO
31
5
this work
c
∼70
Bierbach et al.5
60 24
Gom
this work
Gom
́
ez Alvarez et al.6
b
38
2
83 33
́
ez Alvarez et al.6
1.4 0.3
Tapia et al.8
this work
b
2,3-dimethylfuran
2,5-dimethylfuran
CH3C(O)C(CH3)CHCHO
CH3C(O)CHCHC(O)CH3
8
2
e
27
24
this work
Aschmann et al.7
3
a
Experiments carried out in the presence of NOx unless noted otherwise. In this work, analyses of the unsaturated 1,4-dicarbonyls was as their mono-
b
(minor or not observed) and dioximes from GC−FID analyses of extracts of PFBHA-coated denuder samples (see text). Indicated errors are two
c
d
standard deviations and indicate precision only. Experiments carried out in the absence of NOx. The two values are the HC(O)CHCHCHO
e
yields from each of two experiments conducted. From a single experiment. Concurrent GC−FID analyses of samples collected onto Tenax solid
adsorbent with subsequent thermal desorption resulted in a 3-hexene-2,5-dione yield of 22%.
1,4-butenedial and of 0.20 for 4-oxo-2-pentenal and 2-methyl-
1,4-butenedial, based on the experimental data of Bierbach et
al.25 for Z- and E-1,4-butenedial and a Z-/E-mixture of 4-oxo-2-
butenedial formation from OH + furan, 1.25−1.40 for 4-oxo-2-
pentenal formation from OH + 2-methylfuran, 1.27−1.43 for 2-
methyl-1,4-butenedial formation from OH + 3-methylfuran,
1.23−1.32 for 3-methyl-4-oxo-2-pentenal formation from OH +
2,3-dimethylfuran, and 1.10 for 3-hexene-2,5-dione formation
from the single OH + 2,5-dimethylfuran reaction. While there
are significant uncertainties in the k2/k1 ratios, especially for the
formation of 4-oxo-2-pentenal from OH + 2-methylfuran, 2-
methyl-1,4-butenedial from OH + 3-methylfuran, and 3-
methyl-4-oxo-2-pentenal from OH + 2,3-dimethylfuran, only
for formation of 1,4-butenedial from OH + furan would
significant uncertainties in k2/k1 translate into significant
uncertainties in the multiplicative correction factors (a 25%
uncertainty in k2/k1 would result in uncertainties in the
multiplicative correction factors of <10% for all but 1,4-
butenedial from OH + furan). During the experiments, the
denuder samples were collected over a period of 1.5 h, and the
consistency of the k2/k1 ratios that fit the experimental data
with measured25−27 or estimated rate constants for the
reactions of OH radicals with unsaturated 1,4-dicarbonyls
indicates that their loss rates during our experiments were
controlled mainly by reaction with OH radicals, and that losses
to the chamber walls were minor and that photolysis was
consistent with the literature data and of relatively minor
importance. That wall losses were minor was confirmed by the
preliminary experiments with 2,3-dimethylfuran, with yields
obtained in experiments where PFBHA-coated denuder
sampling was conducted 45 and 156 min after the irradiation
period, being in good agreement with those where sampling
was initiated immediately after the irradiation period (Table S1,
Supporting Information), and by the additional experiments
with in situ analyses of the unsaturated 1,4-dicarbonyls by API-
MS (see, for example, Figures S9 and S10, Supporting
Information).
pentenal and of Thuner et al.28 for E-1,4-butenedial and Z-4-
̈
oxo-2-pentenal. For our experimental conditions, photolysis
was a minor fraction (14−22%) of the OH radical reaction rate
for 1,4-butenedial, 4-oxo-2-pentenal, and 2-methyl-1,4-butene-
dial formed in the furan and 2- and 3-methylfuran reactions.
Photolysis (other than photoisomerization) of 3-hexene-2,5-
dione by black lamps is not important7,11 and this was also
assumed to be the case for 3-methyl-4-oxo-2-pentenal.
The use of these photolysis rates and OH radical reaction
rate constants for the unsaturated 1,4-dicarbonyls, combined
with OH radical reaction rate constants for the furans,7 resulted
in ratios k2/k1 of 1.02−1.04 for formation of 1,4-butenedial
(predominantly the E-isomer) from OH + furan, 0.899−0.914
for formation of 4-oxo-2-pentenal from OH + 2-methylfuran,
0.783−0.792 for formation of 2-methyl-1,4-butenedial from
OH + 3-methylfuran, 0.476 for formation of 3-methyl-4-oxo-2-
pentenal from OH + 2,3-dimethylfuran, and 0.428 for
formation of 3-hexene-2,5-dione from OH + 2,5-dimethylfuran,
where the ranges for the furan and 2- and 3-methylfuran
reactions are due to the varying importance of photolysis versus
OH radical reaction of the unsaturated 1,4-dicarbonyl formed
in the individual denuder experiments (Table S2, Supporting
Information). The use of these k2/k1 ratios to correct for
secondary reactions resulted in unsaturated 1,4-dicarbonyl
formation yields which showed no obvious trend with extent of
reaction, and therefore indicating that these values of k2/k1
were reasonable (see also Figures S5−S8 and the associated
text in the Supporting Information). The resulting unsaturated
1,4-dicarbonyl formation yields are listed in Table 2. Note that
in the single OH + 2,5-dimethylfuran experiment, 3-hexene-2,5-
dione was also quantified from GC−FID analyses of samples
collected onto Tenax solid adsorbent, with agreement to within
20% between analyses of 3-hexene-2,5-dione analyses from
Tenax samples (without derivatization) and as its dioximes
from extracts of PFBHA-coated denuder samples (Table 2).
The multiplicative correction factors24 to account for losses
of the unsaturated 1,4-dicarbonyls were 1.28−1.63 for 1,4-
Comparison with Literature Data. Previous literature
unsaturated 1,4-dicarbonyl formation yields are listed in Table
2. Bierbach et al.5 used in situ FT-IR spectroscopy to investigate
the reactions of OH radicals with furan and 2-methylfuran in
the absence of NOx. The major products observed were 1,4-
butenedial from furan and 4-oxo-2-pentenal from 2-methylfur-
463
dx.doi.org/10.1021/jp410345k | J. Phys. Chem. A 2014, 118, 457−466

The Journal of Physical Chemistry A
Article
an. Because of a lack of authentic standards, the absorption
cross-section of the ∼1720 cm−1 CO stretch in 3-hexene-2,5-
dione measured by Tuazon et al.29 was used for quantification
and that of Gom
́
ez Alvarez et al.6 were carried out in the
presence of NO, while that of Bierbach et al.5 was in the
absence of NO. Note that pathway (B) in Scheme 1 as written
would only occur in the presence of NO, and hence a higher
yield of unsaturated 1,4-dicarbonyl would be expected in the
absence of NO then in the presence of NO, as we observed for
3-hexene-2,5-dione formation from OH + 2,5-dimethylfuran,7
with 3-hexene-2,5-dione yields of 34 3% in the absence of
of 1,4-butenedial and 4-oxo-2-pentenal. Gom
́
ez Alvarez et al.6
used SPME with on-fiber derivatization and GC−FID to
quantify the unsaturated 1,4-dicarbonyls as their oximes from
the furan and 2- and 3-methylfuran reactions, with calibration
of E-1,4-butenedial by use of a synthesized standard. The
SPME GC−FID response factor for E-1,4-butenedial was used
for the quantification of Z-1,4-butenedial and the Z- and E-
isomers of 4-oxo-2-pentenal and 2-methyl-1,4-butenedial.
Tapia et al.8 observed the formation of acetic acid, 3-
furaldehyde (3-furancarboxaldehyde), 2-methyl-1,4-butenedial,
3-methyl-2,5-furandione, 3-methyl-2(3H)-furanone and 2-hy-
droxy-3-methyl-5-(2H)-furanone from the OH radical-initiated
reaction of 3-methylfuran in the presence of NO, using
photolysis of CH3ONO at 254 and 360 nm in a 150 L Teflon-
coated chamber with samples being collected onto SPME fibers
without derivatization and with GC−MS analyses. The
reported molar yields, uncorrected for secondary reactions,
were 3.9 1.8% for 3-furaldehyde, 1.4 0.3% for 2-methyl-1,4-
butenedial, 18.1 4.6% for 3-methyl-2,5-furandione, and 1.6
0.2% for 3-methyl-2(3H)-furanone.8 While the much lower
yield of 2-methyl-1,4-butenedial reported by Tapia et al.,8 using
NO and 24
3% in the presence of NO. Hence the
unsaturated 1,4-dicarbonyl yields from the OH + furan, 2- and
3-methylfuran and 2,3-dimethylfuran reactions in the absence
of NO are also likely to be higher than those measured in the
presence of NO.
The unsaturated 1,4-dicarbonyl formation yields measured in
this study could be low because of (a) less than 100% collection
and derivatization efficiency for the C5 and C6 unsaturated 1,4-
dicarbonyls, (b) rapid losses of the unsaturated 1,4-dicarbonyls
́
to the chamber walls or to aerosol (noting that Gomez Alvarez
et al.6 measured aerosol yields from 2- and 3-methylfuran of
<6%), (c) rapid photolysis of the unsaturated 1,4-dicarbonyls,
or (d) partitioning of the unsaturated 1,4-dicarbonyls to the
chamber walls with an equilibrium being attained. Our
discussion above showing no evidence for significant wall
losses of 3-methyl-4-oxo-2-pentenal from experiments in which
sampling was initiated up to 156 min after the irradiation
period compared to sampling immediately after the irradiation
period argues against possibility (b). The fact that the variation
of measured unsaturated 1,4-dicarbonyl concentration with
extent of reaction (Figures S5−S8, Supporting Information)
shows no evidence for unsaturated 1,4-dicarbonyl loss rates
greater than expected from OH radical reaction plus a small
contribution from photolysis argues against possibilities (b) and
(c). For a given carbon number, aldehydes derivatize more
easily than ketones,15,30 and hence 3-hexene-2,5-dione would
be expected to be the slowest to derivatize. However, the 3-
hexene-2,5-dione formation yield from the OH + 2,5-
dimethylfuran reaction as measured using the PFBHA-coated
denuder procedure is in good agreement with the yield we
previously measured7 from samples collected onto Tenax solid
adsorbent with subsequent thermal desorption and GC−FID
analysis (without derivatization). We have also shown that the
collection and derivatization efficiencies of 2,5-hexanedione, 5-
methyl-2-hexanone, 2-octanone, and 2-decanone on PFBHA-
coated denuders are equal to within 10%.21 This argues against
possibility (a), although mono-oximes were observed for 2-
methyl-1,4-butenedial from 3-methylfuran and for 3-methyl-4-
oxo-2-pentenal from 2,3-dimethylfuran, with the dioximes
accounting for ≥94% of the total oximes for 2-methyl-1,4-
butenedial from 3-methylfuran and for ≥75% of the total
oximes for 3-methyl-4-oxo-2-pentenal from 2,3-dimethylfuran
(but only 44−63% of the total oximes in the three preliminary
experiments which, however, gave results in good agreement
with the subsequent four experiments (see Table S1,
Supporting Information)).
́
SPME without derivatization, than measured by Gomez Alvarez
et al.,6 using SPME fibers with on-fiber derivatization, was
ascribed to potentially fast secondary loss processes,8 we have
not been able to quantitatively analyze unsaturated 1,4-
dicarbonyls by gas chromatography without prior derivatiza-
tion, presumably due to adsorption and/or polymerization on
the sampling and/or chromatographic column surfaces. Hence
it is likely that the very low yield of 2-methyl-1,4-butenedial in
the Tapia et al.8 study was due to analytical problems.
Apart from the quantification of E-1,4-butenedial from OH +
furan by Gom
́
ez Alvarez et al.,6 the other literature yield
measurements5,6,8 are subject to analytical uncertainties8 or
uncertainties in the FT-IR absorption cross sections5 or in the
SPME GC−FID response factors,6 which include the on-fiber
collection and derivatization efficiency which can vary
significantly.30 Our 1,4-butenedial yield from OH + furan in
the presence of NO agrees well with those of Bierbach et al.5
(in the absence of NO) and Gom
́
ez Alvarez et al.6 All three
studies show that formation of 1,4-butenedial dominates, and
that 1,4-butenedial is formed primarily in the E-configuration,
with Gom
́
ez Alvarez et al.6 finding a 76% E-/24% Z-isomer
distribution. For OH + 2-methylfuran and OH + 3-
methylfuran, our measured 4-oxo-2-pentenal and 2-methyl-
1,4-butenedial formation yields are lower by a factor of ∼2 than
those reported by Bierbach et al.5 for OH + 2-methylfuran (in
the absence of NO) and Gom
́
ez Alvarez et al.6 for 2- and 3-
methylfuran (in the presence of NO). While the previous
studies did not differentiate the Z-/E-isomer composition of the
unsaturated 1,4-dicarbonyls formed in the 2- and 3-methylfuran
reactions, we find that 4-oxo-2-pentenal formed from 2-
methylfuran is primarily in the E-configuration, while 2-
methyl-1,4-butenedial from 3-methylfuran is a Z-/E-mixture.
It is apparent from Table 2 that our unsaturated 1,4-
dicarbonyl formation yields decrease markedly from 75 5%
Finally, to investigate the possibility of significant equilibrium
partitioning to the chamber walls, two experiments were carried
out for OH + 2,3-dimethylfuran, with 2,5-hexanedione present,
in which PFBHA-coated denuder samples were collected
immediately after a single irradiation period and after the
chamber was emptied and backfilled with dry purified air (a
total period of 2.15−2.55 h elapsing between the two denuder
samples). The chamber contents were diluted by factors of 4.2
0.4 and 14 3 in the two experiments. Hence, if 3-methyl-4-
for the furan reaction to 8
2% for the 2,3-dimethylfuran
reaction and, as noted above and in Table 2, our 4-oxo-2-
pentenal and 2-methyl-1,4-butenedial yields from 2- and 3-
methylfuran, respectively, are a factor of ∼2 lower than those of
́
Bierbach et al.5 and Gomez Alvarez et al.6 The present study
464
dx.doi.org/10.1021/jp410345k | J. Phys. Chem. A 2014, 118, 457−466

The Journal of Physical Chemistry A
Article
oxo-2-pentenal had partitioned mainly to the chamber walls
with an equilibrium being set up between gas- and wall-phase,
then the (3-methyl-4-oxo-2-pentenal dioximes/2,5-hexanedione
dioximes) ratio would increase from the first denuder sample to
the second denuder sample in the two experiments by factors of
up to 4 and 14, respectively, assuming that no such partitioning
of 2,5-hexanedione occurs. The (3-methyl-4-oxo-2-pentenal
dioximes/2,5-hexanedione dioximes) ratio after the dilution
and backfilling compared to that directly after the irradiation
was 1.5 in the first experiment (versus up to 4.2 expected) and
0.6 in the second experiment (versus up to 14 expected). While
the GC−FID signals after the chamber was emptied and
backfilled were low, these two experiments show no clear
evidence for significant off-gassing of 3-methyl-4-oxo-2-
pentenal from the chamber walls, and hence argue against
possibility (d).
1,2-dicarbonyls in OH + aromatic hydrocarbon reactions), we
have no evidence of problems which would lead to erroneously
low formation yields for these specific unsaturated 1,4-
dicarbonyls in this study.
Theoretical studies of the reactions of OH radicals with
furan,34 2-methylfuran,35,36 and 3-methylfuran37 all conclude
that the dominant pathway at room temperature involves OH
radical addition at the 2- or 5-positions, to form RC(O)CR′
CHC•HOH and/or RC•(OH)CR′CHCHO, where R and
R′ are H and H for furan, CH3 and H for 2-methylfuran, and H
and CH3 for 3-methylfuran (i.e., to form the α-hydroxy radical
product involved in pathways A and B in Scheme 1). Our
experimental data indicate that additional reaction pathways are
involved leading to additional products and suggest that these
are especially significant for 2,3- and 2,5-dimethylfuran. Very
recently, Davis and Sarathy36 have carried out a computational
study of potential reactions occurring subsequent to O2
reaction with the radicals formed by OH addition to 2-
methylfuran in the absence of NO.
Unsaturated 1,4-dicarbonyls, and 1,2-dicarbonyls, are also
formed from the OH radical-initiated reactions of monocyclic
aromatic hydrocarbons.9−14,23,26,31−33 Glyoxal, methylglyoxal
and 2,3-butanedione formation yields from the toluene, o-, m-
and p-xylene, and 1,2,3-, 1,2,4- and 1,3,5-trimethylbenzene
reactions have been obtained at low NO2 concentrations by
Atkinson and Aschmann,33 Bethel et al.11 and Nishino et al.23
The 3-hexene-2,5-dione formation yields of Bethel et al.11 from
the OH + p-xylene and OH + 1,2,4-trimethylbenzene reactions
are identical, within the experimental uncertainties, with the
coproduct glyoxal23,32 and methylglyoxal23 formation yields,
respectively. As discussed and presented in the Supporting
Information, using the same analytical procedures as employed
here, the measured formation yields of 1,4-butenedial13 are
similar to those of its assumed coproducts23,33 from the OH +
toluene and OH + o-xylene reactions (Tables S3 and S4 and
Figure S12, Supporting Information). However, the formation
yields of 4-oxo-2-pentenal, 2-methyl-1,4-butenedial, and 3-
methyl-4-oxo-2-pentenal are significantly lower than those of
their assumed coproducts (Tables S3 and S4 and Figure S12,
Supporting Information).13 These observations suggest that, for
whatever reason (but not including more rapid losses from the
gas-phase than anticipated), the amounts of 4-oxo-2-pentenal,
2-methyl-1,4-butenedial and 3-methyl-4-oxo-2-pentenal pro-
duced in the OH radical-initiated reactions of methyl-
substituted benzenes are not quantitatively observed. The
OH + aromatics data suggest that our measured formation
yields of 4-oxo-2-pentenal, 2-methyl-1,4-butenedial, and 3-
methyl-4-oxo-2-pentenal from 2-methylfuran, 3-methylfuran
and 2,3-dimethylfuran could be low. One possibility not
considered above is that some or most of the unsaturated 1,4-
dialdehydes (other than 1,4-butenedial) and unsaturated 1,4-
keto-aldehydes (which do not appear to elute from GC
columns without prior derivatization) undergo heterogeneous
reactions (possibly oligomer formation) on the denuder and
other surfaces (but obviously not rapidly on Teflon film), or are
in equilibrium with an isomer which was not observed in our
previous study of OH + aromatic reactions nor in the present
study. However, it is quite obvious that other products are also
formed from the OH + furan reactions; indeed, in the presence
of NO the yield of 3-hexene-2,5-dione from OH + 2,5-
dimethylfuran is only 24 3% (Table 2).7 While it is possible
that our measured formation yields of 4-oxo-2-pentenal from
OH + 2-methylfuran, 2-methyl-1,4-butenedial from OH + 3-
methylfuran, and 3-methyl-4-oxo-2-pentenal from OH + 2,3-
dimethylfuran are low (and this would be more consistent with
unsaturated 1,4-dicarbonyls being coproducts to the observed
ASSOCIATED CONTENT
* Supporting Information
■
S
Twenty five pages of Supporting Information are available,
comprising Tables S1−S4, Figures S1−S12, and discussion of
the time−concentration behavior of unsaturated 1,4-dicarbon-
yls during OH + furan reactions and of formation of
unsaturated 1,4-dicarbonyls and 1,2-dicarbonyls from OH +
aromatic hydrocarbon reactions. This information is available
free of charge via the Internet, at http://pubs.acs.org/.
AUTHOR INFORMATION
Corresponding Authors
*J.A.: e-mail, janet.arey@ucr.edu; telephone, 951 452-4502.
*R.A.: e-mail, ratkins@mail.ucr.edu; telephone, 951 452-4503.
■
Present Addresses
†Department of Chemistry, University of California, Irvine, CA
92697.
‡Also Department of Environmental Sciences, University of
California, Riverside, CA 92521.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank the U.S. Environmental Protection Agency
(Grant No. R833752) for supporting this research. While this
work has been supported by this Agency, the results and the
contents of this publication do not necessarily reflect the views
and the opinions of the Agency. J.A. thanks the University of
California Agricultural Experiment Station for partial salary
support.
REFERENCES
■
(1) Graedel, T. E.; Hawkins, D. T.; Claxton, L. D. Atmospheric
Chemical Compounds: Sources, Occurrence and Bioassay; Academic
Press: New York, 1986.
(2) Ciccioli, P.; Brancaleoni, E.; Frattoni, M.; Cecinato, A.; Pinciarelli,
L. Determination of Volatile Organic Compounds (VOC) Emitted
from Biomass Burning of Mediterranean Vegetation Species by GC-
MS. Anal. Lett. 2001, 34, 937−955.
(3) Andreae, M. O.; Merlet, P. Emission of Trace Gases and Aerosols
from Biomass Burning. Global Biogeochem. Cycles 2001, 15, 955−966.
(4) Atkinson, R.; Arey, J. Atmospheric Degradation of Volatile
Organic Compounds. Chem. Rev. 2003, 103, 4605−4638.
465
dx.doi.org/10.1021/jp410345k | J. Phys. Chem. A 2014, 118, 457−466

The Journal of Physical Chemistry A
Article
(5) Bierbach, A.; Barnes, I.; Becker, K. H. Product and Kinetic Study
of the OH-Initiated Gas-Phase Oxidation of Furan, 2-Methylfuran and
Furanaldehydes at ≈300 K. Atmos. Environ. 1995, 29, 2651−2660.
(24) Atkinson, R.; Aschmann, S. M.; Carter, W. P. L.; Winer, A. M.;
Pitts, J. N., Jr. Alkyl Nitrate Formation from the NOx−Air
Photooxidations of C2-C8 n-Alkanes. J. Phys. Chem. 1982, 86, 4563−
4569.
́ ́
(6) Gomez Alvarez, E.; Borras, E.; Viidanoja, J.; Hjorth, J.
(25) Bierbach, A.; Barnes, I.; Becker, K. H.; Wiesen, E. Atmospheric
Chemistry of Unsaturated Carbonyls: Butenedial, 4-Oxo-2-pentenal, 3-
Hexene-2,5-dione, Maleic Anhydride, 3H-Furan-2-one, and 5-Methyl-
3H-furan-2-one. Environ. Sci. Technol. 1994, 28, 715−729.
(26) Calvert, J. G.; Atkinson, R.; Becker, K. H.; Kamens, R. M.;
Seinfeld, J. H.; Wallington, T. J.; Yarwood, G. The Mechanisms of
Atmospheric Oxidation of Aromatic Hydrocarbons, Oxford University
Press, New York, 2002.
Unsaturated Dicarbonyl Products from the OH-Initiated Photo-
oxidation of Furan, 2-Methylfuran and 3-Methylfuran. Atmos. Environ.
2009, 43, 1603−1612.
(7) Aschmann, S. M.; Nishino, N.; Arey, J.; Atkinson, R. Kinetics of
the Reactions of OH Radicals with 2- and 3-Methylfuran, 2,3- and 2,5-
Dimethylfuran, and E- and Z-3-Hexene-2,5-dione, and Products of OH
+ 2,5-Dimethylfuran. Environ. Sci. Technol. 2011, 45, 1859−1865.
(8) Tapia, A.; Villanueva, F.; Salgado, M. S.; Cabanas, B.; Martinez,
̃
(27) Martín, P.; Cabanas, B.; Colmenar, I.; Salgado, M. S.; Villanueva,
̃
E.; Martin, P. Atmospheric Degradation of 3-Methylfuran: Kinetics
and Products Study. Atmos. Chem. Phys. 2011, 11, 3227−3241.
(9) Smith, D. F.; McIver, C. D.; Kleindienst, T. E. Primary Product
Distribution from the Reaction of Hydroxyl Radicals with Toluene at
ppb NOx Mixing Ratios. J. Atmos. Chem. 1998, 30, 209−228.
(10) Smith, D. F.; Kleindienst, T. E.; McIver, C. D. Primary Product
Distributions from the Reaction of OH with m-, p-Xylene, 1,2,4- and
1,3,5-Trimethylbenzene. J. Atmos. Chem. 1999, 34, 339−364.
(11) Bethel, H. L.; Atkinson, R.; Arey, J. Products of the Gas-Phase
Reactions of OH Radicals with p-Xylene and 1,2,3- and 1,2,4-
Trimethylbenzene: Effect of NO2 Concentration. J. Phys. Chem. A
2000, 104, 8922−8929.
F.; Tapia, A. Reactivity of E-Butenedial with the Major Atmospheric
Oxidants. Atmos. Environ. 2013, 70, 351−360.
(28) Thuner, L. P.; Rea, G.; Wenger, J. C. Photolysis of Butenedial and
̈
4-Oxopent-2-enal. The European Photoreactor EUPHORE, 4th Report
2001, Barnes, I. (ed), Institute of Physical Chemistry, Bergische
Universitat Wuppertal, Wuppertal, Germany, November 2003, pp 41−
̈
46.
(29) Tuazon, E. C.; Atkinson, R.; Carter, W. P. L. Atmospheric
Chemistry of cis- and trans-3-Hexene-2,5-dione. Environ. Sci. Technol.
1985, 19, 265−269.
(30) Reisen, F.; Aschmann, S. M.; Atkinson, R.; Arey, J.
Hydroxyaldehyde Products from Hydroxyl Radical Reactions of Z-3-
Hexen-1-ol and 2-Methyl-3-buten-2-ol Quantified by SPME and API-
MS. Environ. Sci. Technol. 2003, 37, 4664−4671.
(31) Volkamer, R.; Platt, U.; Wirtz, K. Primary and Secondary
Glyoxal Formation from Aromatics: Experimental Evidence for the
Bicycloalkyl-Radical Pathway from Benzene, Toluene, and p-Xylene. J.
Phys. Chem. A 2001, 105, 7865−7874.
(32) Volkamer, R.; Spietz, P.; Burrows, J.; Platt, U. High-Resolution
Absorption Cross-Section of Glyoxal in the UV-Vis and IR Spectral
Ranges. J. Photochem. Photobiol. A: Chem. 2005, 172, 35−46.
(33) Atkinson, R.; Aschmann, S. M. Products of the Gas-Phase
Reactions of Aromatic Hydrocarbons: Effect of NO2 Concentration.
Int. J. Chem. Kinet. 1994, 26, 929−944.
(34) Mousavipour, S. H.; Ramazani, S.; Shahkolahi, Z. Multichannel
RRKM-TST and Direct-Dynamics VTST Study of the Reaction of
Hydroxyl Radical with Furan. J. Phys. Chem. A 2009, 113, 2838−2846.
(35) Zhang, W.; Du, B.; Mu, L.; Feng, C. Computational Study on
the Mechanism for the Reaction of OH with 2-Methylfuran.
THEOCHEM 2008, 851, 353−357.
(36) Davis, A. C.; Sarathy, S. M. Computational Study of the
Combustion and Atmospheric Decomposition of 2-Methylfuran. J.
Phys. Chem. A 2013, 117, 7670−7685.
(37) Zhang, W.; Du, B.; Mu, L.; Feng, C. Mechanism for the Gas-
Phase Reaction between OH and 3-Methylfuran: A Theoretical Study.
Int. J. Quantum Chem. 2008, 108, 1232−1238.
́
(12) Gomez Alvarez, E.; Viidanoja, J.; Munoz, A.; Wirtz, K.; Hjorth, J.
̃
Experimental Confirmation of the Dicarbonyl Route in the Photo-
Oxidation of Toluene and Benzene. Environ. Sci. Technol. 2007, 41,
8362−8369.
(13) Arey, J.; Obermeyer, G.; Aschmann, S. M.; Chattopadhyay, S.;
Cusick, R. D.; Atkinson, R. Dicarbonyl Products of the OH Radical-
Initiated Reaction of a Series of Aromatic Hydrocarbons. Environ. Sci.
Technol. 2009, 43, 683−689.
(14) Aschmann, S. M.; Arey, J.; Atkinson, R. Rate Constants for the
Reactions of OH Radicals with 1,2,4,5-Tetramethylbenzene, Pentam-
ethylbenzene, 2,4,5-Trimethylbenzaldehyde, 2,4,5-Trimethylphenol,
and 3-Methyl-3-hexene-2,5-dione and Products of OH + 1,2,4,5-
Tetramethylbenzene. J. Phys. Chem. A 2013, 117, 2556−2568.
(15) Temime, B.; Healy, R. M.; Wenger, J. C. A Denuder-Filter
Sampling Technique for the Detection of Gas and Particle Phase
Carbonyl Compounds. Environ. Sci. Technol. 2007, 41, 6514−6520.
(16) Tuazon, E. C.; Aschmann, S. M.; Nishino, N.; Arey, J.; Atkinson,
R. Kinetics and Products of the OH Radical-Initiated Reaction of 3-
Methyl-2-butenal. Phys. Chem. Chem. Phys. 2005, 7, 2298−2304.
(17) Arey, J.; Aschmann, S. M.; Kwok, E. S. C.; Atkinson, R. Alkyl
Nitrate, Hydroxyalkyl Nitrate, and Hydroxycarbonyl Formation from
the NOx-Air Photooxidations of C5-C8 n-Alkanes. J. Phys. Chem. A
2001, 105, 1020−1027.
(18) Taylor, W. D.; Allston, T. D.; Moscato, M. J.; Fazekas, G. B.;
Kozlowski, R.; Takacs, G. A. Atmospheric Photodissociation Lifetimes
for Nitromethane, Methyl Nitrite, and Methyl Nitrate. Int. J. Chem.
Kinet. 1980, 12, 231−240.
(19) Dagaut, P.; Wallington, T. J.; Liu, R.; Kurylo, M. J. A Kinetic
Investigation of the Gas-Phase Reactions of OH Radicals with Cyclic
Ketones and Diones: Mechanistic Insights. J. Phys. Chem. 1988, 92,
4375−4377.
(20) Obermeyer, G.; Aschmann, S. M.; Atkinson, R.; Arey, J.
Carbonyl Atmospheric Reaction Products of Aromatic Hydrocarbons
in Ambient Air. Atmos. Environ. 2009, 43, 3736−3744.
(21) Aschmann, S. M.; Arey, J.; Atkinson, R. Reaction of OH
Radicals with 5-Hydroxy-2-pentanone: Formation of 4-Oxopentanal
and its OH Radical Reaction Rate Constant. Environ. Chem. 2013, 10,
145−150.
(22) Scanlon, J. T.; Willis, D. E. Calculation of Flame Ionization
Detector Relative Response Factors using the Effective Carbon
Number Concept. J. Chromat. Sci. 1985, 23, 333−340.
(23) Nishino, N.; Arey, J.; Atkinson, R. Formation Yields of Glyoxal
and Methylglyoxal from the Gas-Phase OH Radical-Initiated Reactions
of Toluene, Xylenes, and Trimethylbenzenes as a Function of NO2
Concentration. J. Phys. Chem. A 2010, 114, 10140−10147.
466
dx.doi.org/10.1021/jp410345k | J. Phys. Chem. A 2014, 118, 457−466