OH radical-initiated reaction of 3-hexene-2,5-dione: Formation of methylglyoxal

Article abstract of DOI:10.1021/es034091m

3-Hexene-2,5-dione [CH₃C(O)CH=CHC(O)CH₃] and other unsaturated 1,4-dicarbonyls are formed from the atmospheric photooxidations of aromatic hydrocarbons. We have reinvestigated the formation of methylglyoxal from the gas-phase reaction of OH radicals with 3-hexene-2,5-dione in the presence of NO at room temperature and atmospheric pressure of air using in situ Fourier transform infrared spectroscopy. No evidence for the formation of methylglyoxal was obtained, with the IR spectra showing that methylglyoxal is, at most, a minor reaction product with a molar formation yield of <10% (and more likely <1%). This confirms our earlier study (Tuazon et al. Environ. Sci. Technol. 1985, 19, 265) and suggests that the CH₃C(O)CH(OH)CHO and CH₃C(O)CH(OH)CH(ONO₂)C(O)CH₃ observed by Bethel et al. (Environ. Sci. Technol. 2001, 35, 4477) are the major first-generation reaction products.

Full text of DOI:10.1021/es034091m

Environ. Sci. Technol. 2003, 37, 3339-3341
with the alkoxy radical CH3C(O)CH(OH)CH(O•)C(O)CH3
decomposing by the pathway
OH Radical-Initiated Reaction of
3-Hexene-2,5-dione: Formation of
Methylglyoxal
CH3C(O)CH(OH)CH(O•)C(O)CH3 f
CH3C(O)CH(OH)CHO + CH3C•O (4a)
E R N E S T O C . T U A Z O N * , † A N D
R O G E R A T K I N S O N * , † , ‡
rather than by the alternative pathway which leads to
methylglyoxal formation.
Air Pollution Research Center and Departments of
Environmental Sciences and Chemistry, University of
California, Riverside, California 92521
CH3C(O)CH(OH)CH(O•)C(O)CH3 f
CH3C(O)C•HOH + CH3C(O)CHO (4b)
VO2
CH3C(O)CHO + HO2
3-Hexene-2,5-dione [CH C(O)CHdCHC(O)CH ] and other
3
3
In previous studies of the reaction of OH radicals with
3-hexene-2,5-dione in the presence of NO using in situ Fourier
transform infrared spectroscopy, Becker and Klein (6) and
Bierbach et al. (3) observed the formation of methylglyoxal
in 30-32% yield (as carbon) from the trans-isomer [corre-
sponding to a molar yield of methylglyoxal of 60-64% if two
molecules of methylglyoxal are formed per molecule of
3-hexene-2,5-dione reacted; see reaction 4b], while Tuazon
et al. (2) reported no evidence for methylglyoxal formation
from either the cis- (<10% molar yield) or trans- (<15% molar
yield) isomers. In the Tuazon et al. (2) study, product data
were obtained from kinetic experiments in which propene
was also present in the reactant mixture. To resolve this
discrepancy (2, 3, 6), we have reinvestigated this reaction
using in situ FT-IR spectroscopy to determine the formation
of methylglyoxal from an irradiated mixture of 3-hexene-
2,5-dione and OH radical precursors only, instead of from
a mixture containing an additional reference compound.
unsaturated 1,4-dicarbonyls are formed from the atmospheric
photooxidations of aromatic hydrocarbons. We have
reinvestigated the formation of methylglyoxal from the gas-
phase reaction of OH radicals with 3-hexene-2,5-dione in
the presence of NO at room temperature and atmospheric
pressure of air using in situ Fourier transform infrared
spectroscopy. No evidence for the formation of methylglyoxal
was obtained, with the IR spectra showing that methylglyoxal
is, at most, a minor reaction product with a molar
formation yield of <10% (and more likely <1%). This
confirms our earlier study (Tuazon et al. Environ. Sci.
Technol. 1985, 19, 265) and suggests that the CH C(O)CH-
3
(OH)CHO and CH C(O)CH(OH)CH(ONO )C(O)CH observed
3
2
3
by Bethel et al. (Environ. Sci. Technol. 2001, 35, 4477) are
the major first-generation reaction products.
Introduction
The 1,4-unsaturated dicarbonyl 3-hexene-2,5-dione [CH3C-
(O)CHdCHC(O)CH3] is formed in significant yields from the
atmospheric OH radical-initiated reactions of the aromatic
hydrocarbons p-xylene and 1,2,4-trimethylbenzene (1). In
the atmosphere, 3-hexene-2,5-dione is predicted to undergo
mainly photolysis and reaction with OH radicals (2-4), with
photolysis proceeding to a large extent through cis-/ trans-
isomerization (2, 3). In a recent qualitative product study of
the OH radical-initiated reaction of 3-hexene-2,5-dione using
in situ atmospheric pressure ionization tandem mass spec-
trometry (API-MS), ion peaks attributed to CH3C(O)CH(OH)-
CHO and CH3C(O)CH(OH)CH(ONO2)C(O)CH3 were observed
(5), indicating that the reactions involved are
Experimental Methods
A mixture of trans-3-hexene-2,5-dione (1.76 × 1014 molecule
cm-3), CH3ONO (2.46 × 1014 molecule cm-3), and NO (2.46
× 1014 molecule cm-3) in air was irradiated at 298 K and 740
Torr total pressure in a 5800 L evacuable, Teflon-walled
chamber. The irradiation source was a high-pressure xenon
arc lamp with its collimated beam being passed through a
6 mm thick Pyrex pane to filter out wavelengths <300 nm,
with the resulting radiation having an NO2 photolysis rate of
J(NO2) ) 0.12 min-1. The concentrations of reactants and
products were monitored by in situ FT-IR spectroscopy with
0.7 cm-1 (full width at half-maximum) resolution and a path
length of 62.9 m. Spectra were recorded before and during
photolysis at 3-5 min intervals with an averaging time of 72
s (32 scans) per spectrum. The FT-IR instrumentation
consisted of a White-type folded-path optics interfaced to a
modified Mattson Galaxy 5020 spectrometer equipped with
a liquid N2 cooled HgCdTe detector.
OH + CH3C(O)CHdCHC(O)CH3 f
CH3C(O)CH(OH)C•HC(O)CH3 (1)
CH3C(O)CH(OH)C•HC(O)CH3 + O2 f
trans-3-Hexene-2,5-dione was prepared by oxidation of
2,5-dimethylfuran by bromine in methanol under acidic
condition as described by Levisalles (7). The dark, syrupy
extract obtained did not immediately yield crystals of the
product as implied by the procedure (7), but the concentrate
formed a good yield of crystals after standing for several days
in a refrigerator. The faint yellow compound was purified by
repeated recrystallization in cyclohexane. The IR spectrum
of the purified trans-sample and its calibration was identical
to those obtained previously (2) using the synthesis described
by Armstrong and Robinson (8). A calibrated IR spectrum of
cis-3-hexene-2,5-dione was derived by subtraction of the
trans-spectrum from a spectrum of the cis-/trans-mixture
obtained by irradiation of the trans-isomer in N2 diluent (2).
CH3C(O)CH(OH)CH(OO•)C(O)CH3 (2)
CH3C(O)CH(OH)CH(OO•)C(O)CH3 + NO f
CH3C(O)CH(OH)CH(ONO2)C(O)CH3 (3a)
CH3C(O)CH(OH)CH(OO•)C(O)CH3 + NO f
CH3C(O)CH(OH)CH(O•)C(O)CH3 + NO2 (3b)
* Corresponding author phone: (909)787-5140; fax: (909)787-
5004: e-mail: ectuazon@ucrac1.ucr.edu (E.C.T.); phone: (909)787-
4191; fax: (909)787-5004; e-mail: ratkins@mail.ucr.edu (R.A.).
† Air Pollution Research Center.
‡ Departments of Environmental Sciences and Chemistry.
9
10.1021/es034091m CCC: $25.00
Published on Web 06/21/2003
2003 Am erican Chem ical Society
VOL. 37, NO. 15, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 3 3 9

due to methylglyoxal and was verified to not be a subtraction
remnant of the HCHO peak 1 (Figure 1A).
It is essential to have a spectrum of pure methylglyoxal,
devoid of any HCHO impurity, which can be overlaid and
expanded on a screen simultaneously with the product
spectra (such as traces B, C, and E of Figure 1) in order to
determine whether methylglyoxal is formed as a reaction
product. Bierbach et al. (3) did not present such a spectrum
of methylglyoxal but marked two peaks in the expanded inset
of their Figure 9 as proof of its presence as a product of their
experiments. Their spectrum was presented with unsub-
tracted HCHO peaks, and indeed those two peaks which
were marked “methylglyoxal” correspond in positions to the
HCHO peaks labeled 1 and 2 in Figure 1A of the present
work. Bierbach et al. (3) employed a spectral resolution lower
than that used in this work, but comparison of relative widths
among the HCHO peaks (e.g., 1 and 3 in Figure 1) for this
and Bierbach et al.’s work (3) and with the widths of the P
and R branches of methylglyoxal (Figure 1D, this work) reveals
that the peak denoted by Bierbach et al. (3) as the R branch
of methylglyoxal is significantly narrower than that of the
actual compound. In fact, the peak denoted by Bierbach et
al. (3) as the R branch of methylglyoxal clearly has its largest
component from HCHO itself (see peak 1 of Figure 1A) and,
possibly, with some contribution from the compound
responsible for peak 5 in Figure 1B,E.
Previous studies of the photolysis rates of trans-3-hexene-
2,5-dione (2) and methylglyoxal (10), using the same chamber
and light source as in this work, indicate that for the light
intensity used here the lifetimes of trans-3-hexene-2,5-dione
and methylglyoxal due to photolysis are ∼110 min for trans-
3-hexene-2,5-dione (with g80% of this being trans- to cis-
isomerization) and ∼440 min for methylglyoxal. Hence for
the photolysis times employed in this work (e32 min), ∼7%
or less of the observed loss of 3-hexene-2,5-dione is estimated
to be due to photodecomposition, with the remainder
reacting with the OH radical. If formed, corrections to account
for losses of methylglyoxal due to photolysis and reaction
with OH radicals during our experiments were minor, being
<20%. The experiments of Bierbach et al. (3) employing
photolysis of CH3ONO-NO-trans-3-hexene-2,5-dione-air
mixtures, presumably at 320-480 nm [although this is not
specifically stated (3)], appear to be similar to those conducted
in this work and with photolysis (mainly trans- to cis-
isomerization) accounting for ∼20-30% of the loss of the
trans-3-hexene-2,5-dione (3). In contrast, the experiments
of Liu et al. (4) employed natural sunlight irradiation of a
3-hexene-2,5-dione (2.2 × 1013 molecule cm-3)-NOx (1.5 ×
1013 molecule cm-3)-air mixture and 3-hexene-2,5-dione was
subject to loss by photolysis, reaction with OH radicals, and
reaction with O3 (which attained a concentration of ∼1.7 ×
1013 molecule cm-3 by the time the 3-hexene-2,5-dione was
largely removed). While reaction of 3-hexene-2,5-dione with
O3 in the Liu et al. (4) experiment is estimated to be fairly
minor (∼10-15% of the observed loss), photolysis of both
3-hexene-2,5-dione and first-generation products could well
have been significant over the several hours duration of this
experiment, noting that photolysis was more important
relative to OH radical reaction in the Liu et al. (4) experiment
than in our present and previous (2) studies and that of
Bierbach et al. (3). It is therefore possible that the meth-
ylglyoxal observed (4) was a second-generation product
arising from, for example, photolysis and OH radical-initiated
reaction of CH3C(O)CH(OH)CHO.
FIGURE 1. (A) IR spectrum of a trans-3-hexene-2,5-dione-CH ONO-
3
NO-air mixture after 27 min of irradiation. (B) From (A) after
subtraction of absorptions by HCHO. (C) Spectrum of mixture after
12 min of irradiation, with HCHO absorptions subtracted. (D)
Reference spectrum of methylglyoxal (4.92 × 1013 molecule cm-3).
(E) Segment of the spectrum of the mixture after 32 min of irradiation,
with HCHO absorptions subtracted. Upper traces are offset from
zero for clarity.
Calibrated reference spectra of methylglyoxal were obtained
from a sample prepared as described previously (9).
Results and Discussion
The analysis of reactants and known products by in situ FT-
IR spectroscopy was carried out using mainly the 600-2500
cm-1 region, as described previously (2). However, for the
purpose of measuring the extent of methylglyoxal formation
from 3-hexene-2,5-dione the C-H stretching region was
examined. Figure 1A shows the spectrum of the mixture in
the 2700-3000 cm-1 region after 27 min of irradiation, which
corresponds to a consumption of 8.98 × 1013 molecule cm-3
of 3-hexene-2,5-dione (cis- and trans-) of the initial 1.76 ×
1014 molecule cm-3 of trans-3-hexene-2,5-dione (i.e., 51%
consumption of the initial trans-3-hexene-2,5-dione).
Figure 1B resulted from subtracting only the spectrum of
HCHO from Figure 1A. The subtraction was based on the
1745 cm-1 peak of HCHO, and, as seen in Figure 1B, this
procedure correctly subtracted the HCHO peaks labeled 1,
2, and 3 and the other fine structures comprising the HCHO
absorption in this frequency region. Figure 1C is a similarly
processed spectrum after 12 min of irradiation, which
corresponds to 28% consumption of the initial trans-3-
hexene-2,5-dione. Figure 1D is a reference spectrum of
methylglyoxal (4.92 × 1013 molecule cm3) shown for a direct
comparison of the band contour and intensity with that of
the largely unknown product absorption (labeled 4 in Figure
1B) in the 2800-2860 cm-1 range. Except for HCHO, none
of the starting materials or known methyl nitrite photolysis
products (CH3ONO, CH3ONO2, HC(O)OH, HONO, HNO3, and
NO2) measurably interfere with this product band(s) at 2800-
2860 cm-1. The estimated upper limit (molar) yields of
methylglyoxal corresponding to the spectra shown in Figure
1C (t ) 12 min) and Figure 1B (t ) 27 min) are 1% and 10%,
respectively, with the latter estimate being positively influ-
enced by the growth of a peak labeled 5 in Figure 1B. The
development of peak 5 is further shown in Figure 1E, which
was derived from the last spectrum recorded for this
experiment after 32 min of irradiation and corresponding to
54% consumption of total 3-hexene-2,5-dione. Peak 5 is not
CH3C(O)CH(OH)CHO + hν f
CH3C(O)C•HOH + HC•O (5a)
VO2
CH3C(O)CHO + HO2
9
3 3 4 0 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 15, 2003

band at 1732 cm-1, this is consistent with the possible
presence of CH3C(O)CH(OH)CHO as a reaction product.
CH3C(O)CH(OH)CHO + hν f
CH3C(O)CH2OH + CO (5b)
Acknowledgments
OH + CH3C(O)CH(OH)CHO f
H2O + CH3C(O)CH(OH)C•O (6)
The authors gratefully thank the U.S. Department of Energy
for support of this research through Contract No. DE-FG03-
01ER63095. While this research has been funded by a Federal
Agency, the results and content of this publication do not
necessarily reflect the views and opinions of the Agency or
the U.S. Government.
VO2, NO, O2
CH3C(O)CHO + CO2
Indeed, photolysis of CH3C(O)CH(OH)CHO by pathway
5b could explain the observed formation of hydroxyacetone
(4), which is otherwise difficult to explain as a product of the
OH radical-initiated reaction of either 3-hexene-2,5-dione
or CH3C(O)CH(OH)CHO.
Literature Cited
(1) Bethel, H. L.; Atkinson, R.; Arey, J. J. Phys. Chem. A 2000, 104,
8922.
(2) Tuazon, E. C.; Atkinson, R.; Carter, W. P. L. Environ. Sci. Technol.
1985, 19, 265.
(3) Bierbach, A.; Barnes, I.; Becker, K. H.; Wiesen, E. Environ. Sci.
Technol. 1994, 28, 715.
(4) Liu, X.; Jeffries, H. E.; Sexton, K. G. Environ. Sci. Technol. 1999,
33, 4212.
(5) Bethel, H. L.; Arey, J.; Atkinson, R. Environ. Sci. Technol. 2001,
35, 4477.
(6) Becker, K. H.; Klein, Th. OH-Initiated Oxidation of p-Xylene
under Atmospheric Conditions, Proceedings of the 4th European
Symposium on the Physico-Chemical Behavior of Atmospheric
Pollutants, D. Reidel Publishing Co.: Dordrecht, The Nether-
lands, 1987; pp 320-326.
Our present reinvestigation of the OH radical-initiated
reaction of trans-3-hexene-2,5-dione in the presence of NO,
using in situ FT-IR spectroscopy shows that methylglyoxal
is, at most, a minor reaction product with a molar formation
yield of <10% (and more likely <1%). This confirms our earlier
study (2) and suggests that the CH3C(O)CH(OH)CHO and
CH3C(O)CH(OH)CH(ONO2)C(O)CH3 observed by Bethel et
al. (5) are the major first-generation reaction products and
that reaction 4a dominates over reaction 4b.
(7) Levisalles, J. Bull. Soc. Chim. Fr. 1957, 997.
(8) Armstrong, K. F.; Robinson, R. J. Chem. Soc. 1934, 1650.
(9) Tuazon, E. C.; Atkinson, R. Int. J. Chem. Kinet. 1989, 21, 1141.
(10) Plum, C. N.; Sanhueza, E.; Atkinson, R.; Carter, W. P. L.; Pitts,
J. N., Jr. Environ. Sci. Technol. 1983, 17, 479.
(11) Smith, B. Infrared Spectral Interpretation; CRC Press: 1999; p
96.
It is likely that previous studies have either misinterpreted
the IR spectra obtained (3) or been subject to secondary
reactions which could have led to methylglyoxal formation
as a second-generation product (4). Although a more positive
identification of the CH3C(O)CH(OH)CHO and CH3C(O)CH-
(OH)CH(ONO2)C(O)CH3 products observed by Bethel et al.
(5) is needed, it may be noted that the 2820 cm-1 position
of peak 4 (Figure 1) is within the frequency region where an
aldehydic C-H stretch occurs (11). Combined with present
and previous (2,3) observations of a residual CdO stretch
Received for review February 2, 2003. Revised manuscript
received May 14, 2003. Accepted May 19, 2003.
ES034091M
9
VOL. 37, NO. 15, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 3 4 1