Aqueous Photochemistry of Glyoxylic Acid

Article abstract of DOI:10.1021/acs.jpca.6b00225

Aerosols affect climate change, the energy balance of the atmosphere, and public health due to their variable chemical composition, size, and shape. While the formation of secondary organic aerosols (SOA) from gas phase precursors is relatively well understood, studying aqueous chemical reactions contributing to the total SOA budget is the current focus of major attention. Field measurements have revealed that mono-, di-, and oxo-carboxylic acids are abundant species present in SOA and atmospheric waters. This work explores the fate of one of these 2-oxocarboxylic acids, glyoxylic acid, which can photogenerate reactive species under solar irradiation. Additionally, the dark thermal aging of photoproducts is studied by UV-visible and fluorescence spectroscopies to reveal that the optical properties are altered by the glyoxal produced. The optical properties display periodicity in the time domain of the UV-visible spectrum of chromophores with absorption enhancement (thermochromism) or loss (photobleaching) during nighttime and daytime cycles, respectively. During irradiation, excited state glyoxylic acid can undergo α-cleavage or participate in hydrogen abstractions. The use of ¹³C nuclear magnetic resonance spectroscopy (NMR) analysis shows that glyoxal is an important intermediate produced during direct photolysis. Glyoxal quickly reaches a quasi-steady state as confirmed by UHPLC-MS analysis of its corresponding (E) and (Z) 2,4-dinitrophenylhydrazones. The homolytic cleavage of glyoxylic acid is proposed as a fundamental step for the production of glyoxal. Both carbon oxides, CO₂(g) and CO(g) evolving to the gas-phase, are quantified by FTIR spectroscopy. Finally, formic acid, oxalic acid, and tartaric acid photoproducts are identified by ion chromatography (IC) with conductivity and electrospray (ESI) mass spectrometry (MS) detection and ¹H NMR spectroscopy. A reaction mechanism is proposed based on all experimental observations.

Full text of DOI:10.1021/acs.jpca.6b00225

This is an open access article published under an ACS AuthorChoice License, which permits
copying and redistribution of the article or any adaptations for non-commercial purposes.
Article
pubs.acs.org/JPCA
Aqueous Photochemistry of Glyoxylic Acid
Alexis J. Eugene, Sha-Sha Xia, and Marcelo I. Guzman*
Department of Chemistry, University of Kentucky, Lexington, Kentucky 40506, United States
*
S Supporting Information
ABSTRACT: Aerosols affect climate change, the energy balance of the atmosphere, and
public health due to their variable chemical composition, size, and shape. While the
formation of secondary organic aerosols (SOA) from gas phase precursors is relatively well
understood, studying aqueous chemical reactions contributing to the total SOA budget is the
current focus of major attention. Field measurements have revealed that mono-, di-, and
oxo-carboxylic acids are abundant species present in SOA and atmospheric waters. This
work explores the fate of one of these 2-oxocarboxylic acids, glyoxylic acid, which can
photogenerate reactive species under solar irradiation. Additionally, the dark thermal aging
of photoproducts is studied by UV−visible and fluorescence spectroscopies to reveal that the
optical properties are altered by the glyoxal produced. The optical properties display
periodicity in the time domain of the UV−visible spectrum of chromophores with
absorption enhancement (thermochromism) or loss (photobleaching) during nighttime and
daytime cycles, respectively. During irradiation, excited state glyoxylic acid can undergo α-
13
cleavage or participate in hydrogen abstractions. The use of C nuclear magnetic resonance spectroscopy (NMR) analysis shows
that glyoxal is an important intermediate produced during direct photolysis. Glyoxal quickly reaches a quasi-steady state as
confirmed by UHPLC-MS analysis of its corresponding (E) and (Z) 2,4-dinitrophenylhydrazones. The homolytic cleavage of
glyoxylic acid is proposed as a fundamental step for the production of glyoxal. Both carbon oxides, CO (g) and CO(g) evolving
2
to the gas-phase, are quantified by FTIR spectroscopy. Finally, formic acid, oxalic acid, and tartaric acid photoproducts are
1
identified by ion chromatography (IC) with conductivity and electrospray (ESI) mass spectrometry (MS) detection and H
NMR spectroscopy. A reaction mechanism is proposed based on all experimental observations.
INTRODUCTION
In this research, the identification of missing mechanisms of
SOA production is investigated by monitoring the sunlight
photolysis of aqueous glyoxylic acid, which is followed by
thermal reactions during dark periods. The work (1) explores
how atmospheric chemical reactions of interest proceed in
atmospheric water mimics, affecting the optical properties of
model aqSOAs; (2) identifies the reaction products by ion
chromatography with dual conductivity and electrospray (ESI)
mass spectrometric detection (IC-MS), ultrahigh pressure
liquid chromatography mass spectrometry (UHPLC-MS) of
derivatized samples with 2,4-dinitrophenylhydrazine (DNPH),
■
Recently, aqueous-phase chemistry was suggested to contribute
to the total SOA budget,
oxocarboxylic acid, such as pyruvic acid, can initiate the
formation of oligomers in the aqueous phase and even in ice.
Simultaneously, inorganic electrolytes, such as ammonium and
sulfate, are now known to play a role during dark processes.
Among all the species found in aqueous SOAs (aqSOA), 2-
oxocarboxylic acids, such as glyoxylic acid (GA, pK = 3.13),
are widely present in different environments. Photo-oxidative
processing of isoprene
1
,2
for example, a simple 2-
3
−5
6
,7
8
a
9
−11
12−16
13
and aromatics
produce first
and C nuclear magnetic resonance spectroscopy (NMR); and
generation products such as methylglyoxal, glyoxal, phenol,
catechol, and benzaldehyde. In more detail, glyoxylic acid can
(3) reports the time series of products involved in the reaction
mechanism. The research contrasts the importance of
dicarbonyl intermediates affecting the optical properties of
model aqSOAs during photochemical and thermal processes.
On the basis of the information gathered, a mechanism that
agrees with all experimental observations for the photoreaction
of glyoxylic acid is presented.
•
be directly produced from HO attack on glyoxal, glycolic acid
17
or hydroperoxy acetic acid, and during the oxidation of
14,15
oxygenated aromatic species.
Glyoxylic acid is the most
abundant 2-oxocarboxylic acid present in organic aerosols in
1
8−20
diverse environments.
The annual net production of
10
glyoxylic acid in atmospheric waters totals 2.95 (±0.89) × 10
mol year−1 ≡ 0.71 (±0.21) TgC year , with up to 85%
−1
21
EXPERIMENTAL SECTION
■
22
concentrated in the particle phase. The small contribution of
the aldehyde form of glyoxylic acid to its total speciation in
water provides a tail for absorbing actinic radiation, allowing
photoreactions to occur. Interestingly, the direct photoreaction
of glyoxylic acid in water has remained previously unexplored
and is investigated herein.
Preparation of Experiments and Controls. In a typical
experiment, solutions of glyoxylic acid (Sigma-Aldrich, 52.1 wt
Received: January 9, 2016
Revised: May 10, 2016
©
XXXX American Chemical Society
A
DOI: 10.1021/acs.jpca.6b00225
J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
in H O) were prepared daily in water (Elga Purelab flex, 18.2
Article
%
Scheme 1. Representation of the Processing of Aqueous
a
2
−1
MΩ cm ) at a concentration of ca. 250 mM. Additionally,
experiments studied the effect of glyoxylic acid concentration
covering the range 5−250 mM (Supporting Information). The
solution of glyoxylic acid was doped with the most abundant
inorganic electrolytes found in seawater as a precursor of sea
Glyoxylic Acid Solutions
2
3
+
−
2−
spray: [Na ] = 468 mM, [Cl ] = 545 mM, and [SO ] =
4
2
8.2 mM. Electrolytes were provided similarly as described
2
4
before from NaCl (Fisher, 99%), Na SO (Fisher, 99%), and
HCl (EMD, 38%). The absence of NH4 in the mixture allows
2
4
+
for focusing the work on the effects caused by photoirradiation
2
5
instead of reiterating the reported catalysis by ammonium
that can lead to decarboxylation.
3
,4,7
After adjusting the pH of
the solution of glyoxylic acid to 1.0 with [HCl] = 2.0 M, 180
mL of the solution were transferred to a fused silica
photoreactor (220 mL capacity) with a jacket for temperature
control at 298 K (Thermo Scientific A25 circulating bath),
magnetic stirring, and a septum for sampling gases. The sealed
reactor underwent continuous sparging (unless indicated
a
Stages I and III are both 8 h photolysis periods. Stages II and IV
sequentially refer to (dark) thermal aging during 15 h at 298 K
followed by 9 h processing at 323 K.
−1
otherwise) with 100 mL min air (Scott-Gross, UHP) starting
0 min before photolysis. Experiments were performed under 1
3
atm air and the conditions of ionic strength, temperatures, and
photon flux are chosen to simulate those encountered by
nascent sea spray aerosols mixing with pollution plumes at
coastal regions.
Analysis of Products. Samples were analyzed by (1) UV−
visible spectroscopy (Evolution 220 spectrophotometer,
Thermo Scientific) and fluorescence spectroscopy (Fluoro-
Max-4 fluorimeter, Horiba Scientific); (2) IC-MS with a
Dionex ICS-2000 instrument (IonPack AS11-HC, 2 mm
analytical column) equipped with a conductivity detector and
a mass spectrometer (MS) interfaced with an ESI probe
Control experiments (Table S1) were designed to study the
effect of inorganic electrolytes and [O (aq)] on the photo-
2
reaction of glyoxylic acid. Table S1 indicates which gas was
used and if electrolytes were not present in the controls:
27
(
1
Thermo Scientific, MSQ Plus); (3) UHPLC-MS (Accela
250 with MSQ Plus detector, Thermo Fisher Scientific) using
Control A under 1 atm N (g) (Scott-Gross, UHP), control B
2
under 1 atm O (g) (Scott-Gross, UHP), control C without
2
a C18-selectivity column (Hypersil GOLD, 1.9 μm, 50 × 2.1
mm, Thermo Scientific) for separation of hydrazones from
carbonyls derivatized with freshly prepared 2,4-dinitrophenyl-
electrolytes, control D without electrolytes under 1 atm N (g),
2
and control E without electrolytes under 1 atm O (g). From
2
−3
−1
26
Henry’s law (K = 1.28 × 10 M atm for O at 298 K),
H
2
30
1
hydrazine (DNPH, Sigma-Aldrich, HPLC grade); (4) H
[
O (aq)]
= 0.26 mM and [O (aq)]
= 1.24 mM.
2
1atmair
2
1atmO
2
13
(
(
(
(
using a WET suppression method) and C NMR experiments
Varian 600 MHz) recorded at 298 K using 5 mm NMR tubes
Wilmad); and 5) Fourier transform infrared spectroscopy
FTIR) for evolving gases in a 2.4 m path length infrared gas
Photochemical Experiments and Thermal Treatment
of Photolyzed Samples. The photochemical setup (New-
2
7,28
port) employed was previously described.
Briefly, it
includes a 1 kW high-pressure Xe−Hg lamp, a water filter to
remove infrared radiation, and a cutoff filter for wavelength λ ≥
cell with ZnSe windows (PIKE) mounted in an iZ10 FTIR
module connected to an infrared microscope (Thermo
Scientific Nicolet iN10). Data from all these methods is
reported as the average of duplicate experiments with error bars
corresponding to one standard deviation. Detailed information
on all these experimental methods, instrumental parameters,
samples and standards preparation, and quantification methods
is given in the accompanying Supporting Information.
3
05 nm to provide actinic radiation in the solar window. The
−5
−1
incident light intensity, I = 3.892 (±0.001) × 10 Einstein L
0
−
1
s , was quantified by chemical actinometry using the concerted
29
photodecarboxylation of phenylglyoxylic acid at pH 1.0.
A
calibrated polarographic probe (081010MD) connected to a
meter (both Thermo Scientific) was used to monitor dissolved
oxygen ([O (aq)]) levels during irradiation. During the first
2
stage of processing (stage I in Scheme 1), samples of the
experiment and controls (Table S1) were irradiated for 8 h, and
RESULTS AND DISCUSSION
■
Optical Properties. UV−visible spectroscopy is used to
study the optical property changes of glyoxylic acid photolyzed
under the conditions listed in Table S1. Figure 1 displays the
UV−visible absorption spectra of 252.1 mM glyoxylic acid
before and after photolysis in the presence of electrolytes under
1 atm air (experiment in Table S1) for the four stages listed in
Scheme 1. The reversible hydration of glyoxylic acid into
dihydroxyacetic acid is shifted toward the hydrate form, as
5
4
mL aliquots were withdrawn from the reactor at 0, 0.5, 1, 2, 3,
, 5, 6, 7, and 8 h. A sample volume of 3.5 mL was transferred
to a Suprasil cuvette (Starna Cells, 10 mm optical path length)
to monitor its aging (stage II in Scheme 1) in the dark by UV−
visible spectroscopy for 15 h at 298 K, followed by 9 h at 323 K.
The remaining aliquot was frozen at 253 K and stored in the
dark for later analyses by IC-MS, UHPLC-MS, and NMR
spectroscopy. The remaining 130 mL of photolyzed from the
original solution was stored in the dark for 24 h (stage II in
Scheme 1) and then rephotolyzed (stage III in Scheme 1).
Rephotolyzed samples underwent a second aging process in the
dark, monitored by UV−visible spectroscopy (stage IV in
Scheme 1).
indicated by the large equilibrium constant KHyd
=
[(OH) CH−COOH]/[OCH−COOH] = 300 (at 298
2
3
1
K). Because the hydration equilibrium is shifted toward 2,2-
dihydroxyacetic acid, this species partitions to aerosols and
32
clouds, and it is unlikely to return to the gas-phase. Although
KH is large for glyoxylic acid, the carbonyl group is still
B
DOI: 10.1021/acs.jpca.6b00225
J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
possible direct reaction with carbonyls resulting in heterocyclic
35
products3, as well as the possible formation of alkenes from
6
alcohols, cannot be recalled to provide an explanation for the
data in Figure 1. Beer’s law was obeyed between 280 and 500
nm within the experimental accuracy. An example of the loss of
linearity between 280 and 450 nm due to typical instrumental
limitations is given in Figure S1.
Figure 2A shows the normalized areasto initial values
integrated under the UV−visible absorption spectra in the
Figure 1. UV−visible absorption spectra during stages (A) I−II and
(
B) III−IV in Scheme 1. Spectrum (red line) before irradiation, after 8
h photolysis (green line) for the end of stages I or III), followed by
thermal aging during 15 h of at 298 K (blue line) and 9 h at 323 K
(
purple line), for the end of stages II or IV). Experimental conditions:
+
2−
Solution of [glyoxylic acid] = 252.1 mM, [Na ] = 545.0 mM, [SO
=
]
4
−
28.2 mM, and [Cl ] = 468.0 mM, at pH = 1.0 under 1 atm air.
observed in the spectrum before irradiation (red line in Figure
1
A). Contrarily, other 2-oxocarboxylic acids such as pyruvic
33
acid display a large fraction of carbonyl-form in equilibrium.
The carbonyl absorption of glyoxylic acid at λ = 276 nm in
Figure 1A comprises an absorbing tail that extends above λ =
3
05 nm.
After 8 h of irradiation, the green line in Figure 1A shows
that for stage I photobleaching occurs. This change reflects the
loss of the carbonyl functionality absorption band of the
reagent and its associated tail in the UV. For the following 24 h
of thermal aging in stage II, the blue and purple lines with
increased absorbance resemble the effects of thermochromism
in the sample. Figure 1B presents the results from a second
irradiation period, such as a second daytime cycle (stage III)
following the nighttime period for the last aged sample in
Figure 1A. The mixture of products at time zero before the
second photolysis is shown in red color in Figure 1B, which
undergoes photobleaching (green line) producing a similar
spectrum to that observed at the end of stage I in Figure 1A.
Similarly, moderate thermochromism occurs in stage IV, during
a second dark cycle, recovering a slightly higher absorption than
after stage II.
The observed evolution of carbon species formed after stage
I is related to the presence of coexistent aldehydes and gem-
diols, alcohols, and carboxylic acids that are characterized below
by several spectroscopic techniques. In the dark under mild
atmospheric conditions, it is anticipated that glyoxal possesses
the key aldehyde functional group that displays a continuum of
absorption toward longer wavelengths. Interestingly, organic
products absorbing in the 300−400 nm range were also
observed during the thermal processing of solutions with
Figure 2. (A) Normalized area (295 ≤ λ ≤ 500 nm) under the UV−
visible absorption spectra for (red ○) an experiment, (blue ▲) control
A in N , (pink ▼) control B in O , (violet □) control C without
2
2
electrolytes, (orange ) control D without electrolytes in N , and
⧫
2
(black ×) control E without electrolytes in O . Experimental
2
conditions defined in Figure 1. Dashed lines are to guide the eye
and do not represent a model fit. (B) Fluorescence emission spectra
under variable excitation wavelengths (λ ) for the experiment in
exc
panel A at the end of Stage II. (C) Position of emission maxima
Max
(
λem) for spectra in (black ▲ and blue solid line) panel B vs λexc
compared to (red □ and dashed black line) a 24 h aged solution of
glyoxal] = 11.8 mM with the same electrolytes.
[
wavelength interval 295 ≤ λ ≤ 500 nm during the four stages in
Scheme 1 for the experiment and controls (Table S1). Control
F in the absence of light remained stable over the four stages
and is not included in Figure 2A. There is a large contrast
between photobleaching stages I and III vs thermochromism
[
(
glyoxal] = 220 μM−2.21 M (pH 4) in the presence of
34
6
NH ) SO or NH NO . Because our experiment was
stages II and IV. Each half cycle, simulating daytime vs
4
2
4
4
3
purposely performed in the absence of ammonium ion, its
nighttime periods, shows a dependence on the presence of
25,34
catalytic effect in the self-condensation of aldehydes,
its
inorganic electrolytes and [O (aq)].
2
C
DOI: 10.1021/acs.jpca.6b00225
J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
Although some dispersion is observed in the data set at the
end of stage I in Figure 2A, thermochromism during stage II
In summary, after irradiation and initial loss of absorption, at
least one new chromophore appears but the presence of O2
diminishes its production (Supporting Information). Electro-
lytes play a significantly different role enhancing thermochrom-
ism in stage II, implying interactions of products with
electrolytes. It is important to contrast present results with
provides a better picture of how electrolytes and O (aq) act.
2
Once stage II is complete, a comparison of the effect of
[
O (aq)] (experiment, control A, and control B in Table S1) in
2
the presence of electrolytes, reveals that when O is present
2
6
thermochromism is lower than in an inert atmosphere. The
magnitude for the areas at the end of stage II follows the order
O (g) ≲ air < N (g): 0.38 ≲ 0.40 < 0.62. A similar qualitative
previous experiments for the reaction of pyruvic acid. While
the previous study was focused on the optical properties of
complex organic matter originating from pyruvic acid in 2 M
ammonium bisulfate, this work investigates the behavior of
2
2
conclusion is valid for the subset of experiments in the absence
of electrolytes (controls C, D, and E) for the areas in air ≲
O (g) < N (g): 0.58 ≲ 0.59 < 1.23. The direct effect of
+
glyoxylic acid with a mixture of electrolytes ([Na ] = 545 mM,
−
2−
[Cl ] = 468 mM, and [SO ] = 28.2 mM) that does not
2
2
4
+
electrolytes during thermochromism for stage II can be
contrasted for the pair of subsets in air (experiment vs control
C), in N (g) (controls A vs D), and in O (g) (controls B vs E).
include NH . The effect of electrolytes in present experiments
agrees well with previous observations that electrolytes
4
6
enhance thermochromism of postphotolyzed pyruvic acid
solutions. Similarly, thermochromism increases with higher
ionic strength (from inert electrolytes) and temperature, which
2
2
The results show that the lack of electrolytes in the sample
similarly enhances thermochromism in air and O (g), which
2
6
increases further in N (g), suggesting that the concentration of
may promote dehydration reactions.
2
at least one species contributes to enhance the optical
properties of the sample even in the absence of inorganic
electrolytes in this simple system. After the second photo-
bleaching and thermochromism periods in stages III and IV, the
area of all samples start to merge but control D without
The proposed red shift in the emission spectra (Figure 2B−
C) resulting from the carbonyl moiety of glyoxal provides the
6
,7
new absorption in this kind of model system, and should
contribute to the optical properties of water-extracted ambient
39
particulate matter. Other atmospheric chromophoric species
have been proposed to arise from the oxidation of
polyhydroxylated aromatic rings into polyhydroxylated qui-
electrolytes in N (g) is the only one showing a different
2
behavior. In general, the normalized areas follow a periodic
function in the time domain with slightly decreasing amplitude
for the different stages, an important observation to model the
evolution of absorption properties of atmospheric brown
14
nones absorbing in the 400−600 nm range. In addition, the
direct oxidation of biomass burning and combustion emissions
provided a direct link to explain the production of glyoxylic acid
under humid conditions such as those found in clouds, fogs, or
3
7
carbon.
1
4,15
A recent chromatographic analysis of a nitrogen-containing
brown carbon mixture concluded the red tail observed in the
absorption spectrum could be explained as a linear combination
aqueous secondary aerosols.
The rapid photobleaching of
40
aged solutions of glyoxal by solar illumination was proposed
to result from the loss of imines generated in reactions between
carbonyls and nitrogen-containing nucleophiles. Our findings
38
40
of absorbing molecules. To assess if this was the case for
present experiments, fluorescence spectra for the aged
photoproducts are presented in Figure 2B. The composition
of the mixture of photoproducts at the end of stage II (Scheme
cannot be explained using the same concept due to the lack of
imine formation, suggesting that glyoxal itself is a relevant
atmospheric chromophore. Thus, photobleaching is interpreted
as the disruption of chromophores’ key function of interacting
1) was characterized (see Analysis of Products section to
6
,7
include 11.8 (±0.1) mM glyoxal, 2.00 (±0.03) mM formic acid,
and 493 (±25) μM oxalic acid. For comparison, the individual
emission spectra of glyoxylic acid as well as the identified
products at the target concentrations in a matrix with
electrolytes was also registered. It was confirmed that glyoxylic,
formic, and oxalic acids did not contribute to the emission
weakly among each other and with electrolytes.
Identification of Products. Among the several methods
available to study the photoreaction products of glyoxylic acid,
useful information is provided by analyzing IC-ESI-MS
chromatograms of photolysis products. The information from
IC-ESI-MS reveals the mass-to-charge ratio (m/z) for anionic
products present in the sample after chromatographic
separation. Figure 3 shows IC chromatograms with con-
ductivity and ESI-MS detection in the negative ion mode for
the same experiment in Figure 1 before (A) and after (B) 8 h
photolysis. Before irradiation (Figure 3A), a shoulder and two
peaks elute at 11.86, 12.79, and 19.55 min in the conductivity
detector, which are assigned with matching standards and by
spectra of the aged photolyzed sample in the interval 350 ≤ λ
exc
≤
400 nm. However, Figure 2C shows similar linear
correlations between the maximum wavelength of emission
Max
(
λ ) versus the excitation wavelength (350 ≤ λ ≤ 400
em
exc
nm) exist for the mixture of photoproducts and for aged 11.8
mM glyoxal (24 h in darkness) with electrolytes. We conclude
the red shifting of λem for longer λexc is due to the presence
Max
−
of glyoxal as an individual fluorophore. Therefore, the
absorption tail of aged photoproducts (Figure 1) is thought
to be related to the behavior of glyoxal in the emission spectra
for λexc ≥ 350 nm (Figure 2B−C). Emission measurements also
confirmed no other fluorophore species (or corresponding
chromophore) is present in the starting experimental mixture
their m/z values to (1) glyoxylate (C H O , m/z 72.99), (2)
2
1
3
−
−
chloride (Cl ), and (3) bisulfate (HSO , m/z 96.99),
respectively. Although Cl could not be detected by MS, its
4
−
conductivity peak was shown to grow after spike addition with a
standard. Clearly, some conductivity peaks correspond to
overlapping species based on the distinctive m/z values
extracted with the help of the MS detector. Upon 8 h of
photolysis (Figure 3B), new peaks elute at 18.60 and 20.33 min,
corresponding to two products from the photoreaction of
Max
before photolysis, as revealed by the unchanged value of λem
=
441.3 ± 2.1 nm for the excitation range 350 ≤ λexc ≤ 400 nm.
For the interval 280 ≤ λ ≤ 345 nm, the emission spectra of
the aged photolyzed experiment with glyoxylic acid only
displayed a practically constant value for λem = 444.1 (±1.8)
nm. The unchanged
exc
−
glyoxylic acid, (4) tartrate (C H O , m/z 149.01), and (5)
4
5
6
Max
−
oxalate (C HO , m/z 88.99), respectively.
2 4
Max
λem for the wavelength interval 280−
The similar product composition found in the experiment
and control C indicates that the presence of inorganic
345 nm is due to the fluorescence of unreacted glyoxylic acid.
D
DOI: 10.1021/acs.jpca.6b00225
J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
Figure 4. UHPLC-MS for the extracted ion at m/z 417.05 for (E) and
(Z) DNPH-glyoxal hydrazones in the experiment of Figure 1. The red
trace corresponds to a standard.
Figure 3. Ion chromatogram with conductivity and ESI-MS(−)
detection for the experiment in Figure 1 (A) before and (B) after 8 h
of irradiation. Key for peaks: (1) glyoxylic acid (m/z 72.99), (2)
chloride, (3) sulfate (m/z 96.99), (4) tartaric acid (m/z 149.01), and
(5) oxalic acid (m/z 88.99).
electrolytes does not determine the formation of the low
molecular weight products. Therefore, the major new photo-
products identified by IC-MS are oxalic (C H O ) and tartaric
2
2
4
(
C H O ) acids with molecular weights of 90.03 and 150.09
4 6 6
amu, respectively. The formation of these two species requires
the presence of glyoxylic acid to form the C - and C -carboxylic
2
4
acids. Although formic acid (HCOOH) escaped MS detection
at m/z 45, its presence was confirmed by IC separation with a
Figure 5. (Blue) Infrared spectrum of gas-phase products after
photolysis (λ > 305 nm) of glyoxylic acid for 8 h. All other
experimental conditions are indicated in Figure 1. The red trace shows
a 1:18 dilution of the spectrum.
1
.0 mM KOH isocratic gradient and conductivity detection.
1
Instead, the production of formic acid was quantified by H
NMR.
Figure 4 shows the production of glyoxal during irradiation,
as registered in the extracted ion chromatogram at m/z 417.05
for its (E) and (Z) hydrazones, corresponding to the reaction 2
DNPH + 1 glyoxal. While the hydrazone with m/z 417.05 is
absent before irradiation (blue trace in Figure 4), its presence is
confirmed for the experiment and matched to a standard (red
trace) at 7.02 and 7.65 min. After irradiation starts, from 30 min
to 8 h, the concentration of glyoxal remains relatively stable
collected over the headspace of the reactor containing carbon
dioxide (CO ) and a lower level of carbon monoxide (CO).
2
The spectrum shows the asymmetric stretching vibration of
−
1
CO at 2349 cm , and the well-resolved CO bands centered at
2
−1
2143 cm . The presence of formaldehyde was also discarded in
the gaseous products because its very strong absorption at 1771
−1
[
glyoxal] = 2.02 (±0.10) mM. Therefore, during the experi-
cm was missing from the spectrum.
ment glyoxal is an intermediate that remains in steady state.
The origin of glyoxal is discussed below, where its presence is
The use of 13C NMR spectroscopy provides additional
information from the chemical shifts (δ) of reactants and
products in the photolysis mixture. 3-(Trimethylsilyl)-1-
propanesulfonic acid sodium salt (DSS) was added to the
NMR tube as a reference (δ = 0.00 ppm) in the aqueous
1
3
also confirmed by C NMR spectroscopy. The formation of
formaldehyde (H CO) as a primary photoproduct was
2
discarded after performing UHPLC-MS analysis of derivatized
3
0
13
samples. The hydrazone from derivatized formaldehyde
would have eluted at 2.39 min in the extracted ion
chromatogram at m/z 209.03, as confirmed after spike addition
of the samples with a 5 μM standard.
solutions for all samples. Figure 6 presents the C NMR
spectrum for an experiment after 8 h of irradiation. In addition
to the reference peak at δ = 0.00 ppm, DSS signals also appear
13
at 17.60, 21.71, and 56.97 ppm in the C NMR spectrum. The
The experimental results point to the different photolytic
behavior of glyoxylic acid in water and the gas-phase, where
abundant gem-diol (−C(OH) ) group of glyoxylic acid, with
2
3
sp carbon, is observed at δ 88.89 ppm, while its carboxylic acid
41
CO was accompanied by formaldehyde generation. Figure 5
group (−COOH) is at δ 175.90. The −COOH of formic acid
product is observed δ 179.41 ppm in Figure 6, while oxalic acid
2
shows the major products in the infrared spectrum of the gas
E
DOI: 10.1021/acs.jpca.6b00225
J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
Figure 7. (Red ◊) Normalized loss of glyoxylic acid and production of
(
brown ∇) CO , (green ●) CO, (gray ×) glyoxal, (black Δ) formic
2
acid, (pink ○) oxalic acid, (blue ■) tartaric acid, and (teal hexagon)
dissolved O during 8 h photolysis under the conditions given in
2
Figure 6. 13C NMR spectrum after 8 h photolysis of glyoxylic acid
Figure 1.
under the experimental conditions given in Figure 1.
acid, CO (g), oxalic acid, CO(g), and tartaric acid follows zero-
2
−1
−1
2
order kinetics with rate constants k (mM h ): 2.39 × 10 (R
= 0.998), 2.28 × 10− (R = 0.985, excluding the outlier at the
2 2
1
2
(
δ = 168.51 ppm) was only detectable after spike addition of a
−
standard (Sigma-Aldrich, 99.7%) to the sample. The origin of
95% confidence level for t = 8 h), 6.34 × 10 (R = 0.980),
2 2
•
−2
−3
formic acid can be ascribed to the formyloxyl radical COOH
4.92 × 10 (R = 0.961), and 6.77 × 10 (R = 0.984),
respectively. Therefore, the uniform changes observed in Figure
7 indicate the persistent photoproducts formed undergo
reversible thermochromism and photobleaching in water.
This versatile behavior is mainly provided by the aldehyde
functional group of glyoxal and/or its derived material in the
mixture. Finally, it is important to indicate that the behavior of
dissolved oxygen in this experimental system is quite different
created during the homolytic cleavage of excited state glyoxylic
•
acid. This COOH radical is converted to formic acid after
abstracting a hydrogen atom, for example, from dihydroxyacetic
acid. However, a comparison of product yields suggests that the
pathway including H-abstraction may be as favorable as the
channel resulting in CO evolution.
2
Because tartaric acid is about 10-times less abundant than
oxalic acid, as quantified by IC-MS, indeed this species remains
below the limit of detection in Figure 6. Spike addition of L-
42,43
to that reported in studies with methyl vinyl ketone.
Under
irradiation, [O (aq)] in Figure 7 (external right-hand side
2
(
+)-tartaric acid (Sigma-Aldrich, 99.8%) standard to the
vertical axis) only decreases 70% (from 0.26 mM to 0.08 mM)
within the first 2 h, and then recovers close to 80% of the
expected equilibrium concentration. In other words, the
product mixture indicates its peaks for −COOH and
−
CH(OH)− should be observed at δ 177.41 and δ 74.72
ppm, respectively. Remarkably, the new peak appearing for
production of species that consume O (aq) is never large
2
glyoxal at δ 93.22 ppm matches that of a spiked standard
enough to completely deplete this important gas in solution
during the whole experiment.
(
Sigma-Aldrich, 38.5%) in the products’ spectrum, indicating
•
the homolytic reaction HOOC-O(H)C + hν → HOOC +
Proposed Reaction Mechanism. From the experiments
described in the discussion above, a mechanism for the
photoreaction of glyoxylic acid via radical chemistry is proposed
in Scheme 2. The proposed mechanism, which advances
previous knowledge in this field, serves as a first representation
that could be further improved in the future. Dissolved organic
•
C(H)O proceeds upon λ ≥ 305 nm irradiation.
The source of glyoxal could be linked to formyl radical
•
•
recombination through the reaction: O(H)C + C(H)O
→
O(H)C−C(H)O. However, more likely the addition of
•
O(H)C to the carbonyl group of glyoxylic acid generates an
alkoxy radical that, after homolysis, produces a molecule of
matter, including triplet-state carbonyls, can play a significant
•
44,45
glyoxal and a second HOOC . More details of these reactions
role as a photosensitizer in the environment.
For example,
are provided in the mechanism below.
Figure 7 shows the first-order decay of [glyoxylic acid] at
the oligomerization of methyl vinyl ketone proceeds after
42,43
reaction with OH or excited state pyruvic acid in water.
In
t
each time (t) point relative to its initial value [glyoxylic acid]0
during 8 h irradiation of stage I, which is described by the
equation [glyoxylic acid] /[glyoxylic acid] . = 0.998 × exp-
present experiments, as for other carbonyls and due to its
absorbing tail for λ ≥ 300 nm, the lowest energy n → π*
transition is activated by solar light to produce a very reactive
singlet excited state of glyoxylic acid. The singlet excited state
can be considered to react directly, or undergo intersystem
crossing producing a triplet, which can then react, as it has been
t
0
−3
−1
2
(−8.85 × 10
h
× t) with R = 0.980. The right-hand side
vertical axis within Figure 7 displays the scale for the
production of glyoxal, CO , CO, formic acid, oxalic acid, and
2
41
tartaric acid for the same time series. Photoproducts can be
consumed during reactions, e.g. with carbonyls or reactive
radical species generated during photolysis. For example,
glyoxal quickly reaches a quasi-steady state in Figure 7 with a
characteristic time τ = 7.5 min as described by the exponential
rise to maximum fitting [glyoxal] (mM) = 2.04 × (1 −
suggested for the gas-phase photolysis of glyoxylic acid.
Therefore, Scheme 2 depicts a generic excited state glyoxylic
acid GA* to be produced upon photon absorption by reaction
R1.
The excited state GA* can react directly undergoing
homolytic cleavage by reaction R2 into formyl-oxyl radical
exp(−5.31 h− × t)) with R = 0.981. The production of formic
1
2
HOOC and formyl radical C(H)O, the simplest existing
•
•
F
DOI: 10.1021/acs.jpca.6b00225
J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
Scheme 2. Proposed Mechanism for the Photooxidation of Glyoxylic Acid (GA) in Water
49
Alternatively, HOOC• could abstract a hydrogen from
dihydroxyacetic acid to produce formic acid and G .
acyl radical. The alternative fate of GA*, as a triplet, is to
abstract a hydrogen atom from the very abundant hydrate of
glyoxylic acid via reaction R3 to generate the radicals of glycolic
•
The production of oxalic acid observed can be explained by
•
•
•
acid (K ) and dihydroxyacetic acid (G ). Hydrated glyoxylic
the direct oxidation of G after combining with O and in situ
2
53
acid cannot accept energy from GA* but hydrogen abstraction
elimination of hydroperoxyl radical via reaction R9. Once
oxalic acid is formed, it can undergo in situ reduction to form
G again by any of the α-hydroxyalkyl radicals available in the
46
is a characteristic reaction of triplet 2-oxocarboxylic acids. The
doublet displayed in electron paramagnetic resonance measure-
•
4
7
•
47
•
ments in the liquid phase shows that K can also be generated
system. The combination of two K radicals via reaction R10,
•
−1 −1
54
8
when H-abstraction from glycolic acid occurs. Similarly, G
with rate constant k
•
•
= 5.5 × 10 M s , results in the
K +K
corresponds to a triplet previously registered during the
formation of tartaric acid, a less likely process for the fate of this
radical, as represented by the low level of product quantified. As
observed for the production of dimethyltartaric acid from the
recombination of lactic acid radicals in the photolysis of pyruvic
48
reduction of oxalic acid at low pH.
Reaction R4 (Scheme 2) generates carbon monoxide from
•
•
49
50
C(H)O and also provides a pathway to form K from GA.
The acid dissociation constant of HOOC• (pK = 2.3)
3,4
a
acid, the pathway for the production of tartaric acid in this
indicates that 95% of the speciation of this radical at the
experimental condition of pH 1.0 is in the undissociated form.
work is not fully quenched in the presence of [O (aq)]
=
2
1atmair
0
.26 mM. The recombination of α-hydroxyperoxyl radicals
•−
Interestingly, the fate of the conjugated base CO2 for the pH
should become more important at higher radical concentration
55
range 3−9, is to form oxalate through bimolecular recombina-
−9
than 5 × 10 M, and also as proposed here for decreasing
9
−1 −1 50
•−
•−
2
•
tion with a rate constant kCO2
+CO
= 1.4 × 10 M s .
[O ] found at higher altitudes. The reaction of K with O
2
2
•
However, at lower pH (e.g., pH 1.0) the yield of oxalate from
proceeds by reaction R11, with a rate constant k
= 1.76 ×
K +O
2
the previous reaction becomes negligible while the yield of CO
9
−1 −1 56
2
10 M s , to regenerate the carbonyl-form of glyoxylic acid
after elimination of hydroperoxyl radical with a rate constant
5
0
increases. Therefore, discarding the recombination reaction
•
•
HOOC + C(H)O → HOOCC(H)O as viable under
−1 55
•
kKO2 = 122 s .
•
continuous irradiation, the fate of HOOC at low pH is (1) to
An alkoxy radical is generated following the addition of
•
•
self-disproportionate with rate constant k
= 1.7 ×
HOOC +HOOC
•
9
−1 −1
51
C(H)O to the carbonyl group of glyoxylic acid (reaction
10 M
s
(reported at pH 0) into CO and HCOOH by
reaction R5 (Scheme 2); (2) to react with another radical,
2
50
49
R12, Scheme 2). It has been previously proposed that at high
pH the semidione radical would be an important intermedi-
that is, to produce CO and regenerate the hydrate form of
2
47
•
ate. The production of glyoxal in reaction R13 is proposed to
result from the cleavage of the alkoxy radical formed in reaction
glyoxylic acid by reaction R6; or (3) to react with C(H)O
49
to yield formic acid and CO. Carbon monoxide can also be
generated from the formyl radical in the presence of dissolved
O2 by reaction R7 that eliminates hydroperoxyl radical
R12. Reaction R13 simultaneously generates a second source of
•
COOH that contributes CO , the major reaction product. It is
2
•
52
•−
(
HO ). As indicated above, CO2 formed by reaction R8a
important to highlight that the presence of carbonyls (e.g.,
2
can undergo efficient acid catalyzed reaction with glyoxylic acid
glyoxylic acid or in situ produced glyoxal) acting as sensitizers
•
47
to release CO (g) and K by reaction R8b. At very acidic pH,
can result in a positive feedback for the production of CO .
2
2
•
•
the sequence R8a + R8b proceeds directly from HOOC + GA.
Because the possible recombination of two C(H)O radicals
G
DOI: 10.1021/acs.jpca.6b00225
J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
•
seems an unlikely process (i.e., the recombination of two
COOH yields little oxalic acid), a reaction such as R13 is
reaction R7 decreases the availability of C(H)O for the
glyoxal production channel R12 + R13. Indeed, the growing
trends for ΦCO2 and ΦCO with increasing [glyoxylic acid]0
•
thought to dominate the production of glyoxal. The carbonyl
group of the alkoxy radical formed in reaction R12 likely
undergoes fast hydration and actually generates the mono-
hydrate of glyoxal, which incorporates a second molecule of
provide mechanistic support for direct reactions R4 and R8a +
R8b, and indirectly by enhancing the production of G in
reaction R3, which is available to facilitate the decarboxylation
R6. A related discussion of the effect of [O (aq)] and [glyoxylic
•
5
57
water (Khyd = 2.2 × 10 ) to produce the unique tetrol signal
2
13
observed in the C NMR spectrum (equilibrium R14, Scheme
acid] in the production of formic, oxalic, and tartaric acids is
0
5
7
2).
presented in the Supporting Information.
Reaction R15 shows that hydrogen abstraction by any of the
radicals present in the system, denoted by the generic radical
CONCLUSIONS AND ATMOSPHERIC
IMPLICATIONS
•
■
L , yields the hydrated glyoxal radical. In reaction R16, the
•
hydrated glyoxal radical M combines with O forming the
2
•
Addressing the significance of aqueous photochemistry in the
processing of atmospheric particles is an important issue.
However, uncertainties in this field remain because photolysis
quantum yields for many compounds are unknown. The data
•
glyoxyl peroxyl radical MO with a rate constant k
= 1.1 ×
2
M +O
2
1
,2,59
09 M−1 s . After elimination of HO , also included in
−1
55
•
1
2
•
−1 55
reaction R16 (k
= 122 s ), the gem-diol form of glyoxylic
2
59
MO
31
acid is generated, which is in equilibrium R17 (KHyd = 300)
in this work allow us to report for the first time an overall
quantum yield for the photolytic loss of aqueous glyoxylic acid
(Φ−GA) in air under solar irradiation. The ratio of the initial
reaction rate of aqueous glyoxylic acid in the experiment of
with the carbonyl form of glyoxylic acid. Alternatively,
•
52
disproportionation during the recombination of MO radicals
2
produces glyoxal and glyoxylic acid. Different carbonyls have
been shown to favor both α-cleavage and hydrogen abstraction
occurring simultaneously during primary photochemical
−
7
−1 −1
Figure 7, (−d[GA]/dt) = 6.75 × 10 mol L s , to I
0
0
provides Φ−GA, = 0.017 for λ = 325 (±20) nm. This value can
be used to estimate the photolysis rate in the aqueous phase
5
8
reactions and glyoxylic acid is not the exception. For this
•
•
•
•
•
•
•
•
mechanism, reactions G + G , G + K , G + M , and K + M
(
j ) from the integral of the variable actinic flux F and the
aq A
were considered and disregarded because no corresponding
products with m/z 165 181, 183, and 167 were found in the IC-
MS chromatogram. Remarkably, despite the low levels of
products generated in the condensed phase, this mechanism
clearly highlights the importance of understanding further
mechanistic steps occurring in atmospheric waters. The
mechanism provides details of several reactions (i.e., R7, R9,
R11, and R16) capable of generating reactive oxygen species in
situ from a photoinitiated process previously unknown.
cross section of glyoxylic acid σ : j = ∫ F (λ) Φ(λ) σ(λ) dλ.
λ
aq
A
Available data for surface altitude, a solar zenith angle of 40°,
60
and a surface albedo of 80% is chosen to represent FA, while
σλ is estimated from the known concentration that simulates
atmospheric conditions (Supporting Information) and the
corresponding absorbance values in Figure 1. Thus, under
atmospheric conditions a value of jaq = 9.9 × 10−
7
s
−1
is
obtained, the reciprocal of which can be used to estimate a
rough photolysis lifetime in the aqueous phase of τ ≈ 11 d. For
comparison, the loss of glyoxylic acids by indirect photolysis in
the aqueous phase proceeds only two times faster (e.g,
It is interesting to contrast the effect of dissolved O in the
2
formation of both CO (g) and CO(g) in experiments
2
61
supersaturated in O (g) ([O (aq)] = 1.24 mM) vs 1 atm
τaq,GA+OH ≈ 5 d). While the rate constant for the gas phase
2
2
0
•
N (g). For this purpose, the evolution of CO (g) and CO(g)
reaction glyoxylic acid + OH is unknown, structure activity
2
2
relationships give an estimated k
= 1.28 × 10− cm
11
3
6
during irradiation was monitored for [glyoxylic acid] = 5, 25,
00, and 250 mM in the electrolyte matrix during irradiation.
GA+OH
−1
−1 59
1
molecule s . Using an average daytime [OH(g)] 1.6 × 10
3
molecules cm− leads to a lifetime τgas,GA+OH ≈ 0.56 day. The
The ratio of the initial rates of CO (g) and CO(g) formation to
2
large gas phase photolysis rate of glyoxylic acid j = 8.7 × 10−
5
I0 were employed to calculate the corresponding apparent
quantum yields ΦCO2 and ΦCO displayed in Figure S4. The fact
gas
−1
17
s
provides a shorter lifetime τgas,hν = 3.2 h. However, as the
hydration equilibrium of glyoxylic acid is shifted toward 2,2-
dihydroxyacetic acid, this species preferentially and quickly
that the ratio of quantum yields (Figure S4) in the presence of
dissolved oxygen is always larger than the corresponding ratio
under 1 atm N (g), (ΦCO2/Φ ) 2 > (ΦCO2/Φ ) , clearly
32
partitions to the particle phase. Because of the improbable
presence of glyoxylic acid in the gas-phase, its fate should not
2
CO O
CO N
2
indicates that O (aq) favors the formation of CO (g), the fully
•
2
2
be dominated by reactivity with OH(g) radical.
oxidized product. The values of ΦCO2 are enhanced for larger
glyoxylic acid], suggesting that some bimolecular processes
This work studies for the first time the photoreaction of
aqueous glyoxylic acid under simulated solar irradiation and
thermal processing of the photolyzed in the dark. The
mechanism proposed via radicals is initiated by a photoexcited
α-dicarbonyl species in water and results in the production of
species with large O:C ratio. The results provide an explanation
to the possible role of glyoxylic acid in the generation of
complex organic matter in aqueous aerosols. Although glyoxylic
acid is not a major absorber of solar radiation, the tail of its
electronic spectrum is extended well into the region of solar
actinic radiation available in the troposphere. After light
absorption, glyoxylic acid can induce secondary reactivity of
other molecules as demonstrated here. This information
contributes new experimental evidence to explain the
mechanisms of aqSOA production of oxalic acid and establishes
[
and/or the participation of reaction intermediates could be
important to facilitate CO production during irradiation. In
2
contrast, Φ appears to follow a similar growing trend for
CO
increasing [glyoxylic acid] with dissolved O (g) or N (g),
2
2
suggesting the production of CO(g) has a much smaller
dependence on [O (aq)] than CO (g) does. For the two
2
2
higher [glyoxylic acid] (Figure S4), the significant difference in
under supersaturated dissolved oxygen versus N (g) might
Φ
CO
2
be attributed to reaction R7 (Scheme 2) with direct
participation of O (aq). A related conclusion can be drawn
2
from the larger quasi-steady state glyoxal production obtained
by extrapolating the [glyoxal]inf as time → ∞ (Figure S5)
without dissolved O . Figure S5 suggests that O (aq) in
2
2
H
DOI: 10.1021/acs.jpca.6b00225
J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
a link to glyoxal chemistry. The low production of tartaric acid,
a dimer of a primary radical, is also detected. However, it is
clear that glyoxylic acid is also converted to carbon oxides.
Experiments and controls have also explored the effects
(3) Guzman, M. I.; Colussi, A. J.; Hoffmann, M. R. Photoinduced
oligomerization of aqueous pyruvic acid. J. Phys. Chem. A 2006, 110,
3
(
619−3626.
4) Guzman, M.; Hoffmann, M.; Colussi, A. Photolysis of pyruvic
́
acid in ice: Possible relevance to CO and CO2 ice core record
caused by inorganic electrolytes and [O (aq)]. Chemical
2
anomalies. J. Geophys. Res. 2007, 112, D10123.
identification is obtained from a combination of spectroscopies
that supports the direct participation of radicals in the
mechanism. In addition, this work has contrasted the optical
properties showing the periodic behavior of samples during
daytime and nighttime cycles of photobleaching and
thermochromism, respectively. Overall, the optical properties
of atmospheric brown carbon species change periodically in the
(
5) Guzman, M. I.; Colussi, A. J.; Hoffmann, M. R. Photogeneration
of distant radical pairs in aqueous pyruvic acid glasses. J. Phys. Chem. A
006, 110, 931−935.
6) Rincon, A. G.; Guzman, M. I.; Hoffmann, M. R.; Colussi, A. J.
Thermochromism of model organic aerosol matter. J. Phys. Chem. Lett.
2
(
2
010, 1, 368−373.
(7) Rincon, A. G.; Guzman, M. I.; Hoffmann, M. R.; Colussi, A. J.
Optical absorptivity versus molecular composition of model organic
aerosol matter. J. Phys. Chem. A 2009, 113, 10512−10520.
aqueous-phase. O (g) in air cannot stop the development of
2
thermochromism, affecting minimally the optical properties of
the model organic material produced, which in the atmosphere
is influenced by aerosols’ variable and complex size, structure,
and composition. These results are directly related to the
ulterior oxidative fate of polyphenols emitted during biomass
burning and combustion processes, which have been proposed
to contribute chromophores and reactive intermediates in route
to the formation of the dicarbonyl, oxo-carboxylic acid, and
(
8) Das, S.; Bhattacharyya, J.; Mukhopadhyay, S. Mechanistic studies
4
+
on the oxidation of glyoxylic and pyruvic acid by a [Mn O ] core in
4
6
aqueous media: Kinetics of oxo-bridge protonation. Helv. Chim. Acta
006, 89, 1947−1958.
9) Carlton, A. G.; Turpin, B. J.; Lim, H. J.; Altieri, K. E.; Seitzinger,
2
(
S. Link between isoprene and secondary organic aerosol (SOA):
Pyruvic acid oxidation yields low volatility organic acids in clouds.
Geophys. Res. Lett. 2006, 33, L06822.
1
4,15
dicarboxylic acid components of brown carbon.
In
(10) Ervens, B.; Carlton, A. G.; Turpin, B. J.; Altieri, K. E.;
particular, oxocarboxylic acids are important species in the
atmosphere because they participate in aqueous-phase reactions
in the presence of electrolytes. The observed behavior of model
organic species should be treated in future climate change
models taking into account thermochromism and photo-
bleaching parametrization factors.
Kreidenweis, S. M.; Feingold, G. Secondary organic aerosol yields from
cloud-processing of isoprene oxidation products. Geophys. Res. Lett.
2008, 35, L02816.
(11) Nguyen, T. B.; Bateman, A. P.; Bones, D. L.; Nizkorodov, S. A.;
Laskin, J.; Laskin, A. High-resolution mass spectrometry analysis of
secondary organic aerosol generated by ozonolysis of isoprene. Atmos.
Environ. 2010, 44, 1032−1042.
(
12) Martín-Reviejo, M.; Wirtz, K. Is benzene a precursor for
secondary organic aerosol? Environ. Sci. Technol. 2005, 39, 1045−1054.
13) Bloss, C.; Wagner, V.; Jenkin, M.; Volkamer, R.; Bloss, W.; Lee,
ASSOCIATED CONTENT
Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.jpca.6b00225.
■
*
S
(
J.; Heard, D.; Wirtz, K.; Martin-Reviejo, M.; Rea, G.; et al.
Development of a detailed chemical mechanism (MCMv3. 1) for
the atmospheric oxidation of aromatic hydrocarbons. Atmos. Chem.
Phys. 2005, 5, 641−664.
Detailed experimental section, additional characterization
of starting material and effects of dissolved oxygen on
products formation, Figures S1−S6, and Table S1 (PDF)
(14) Pillar, E. A.; Camm, R. C.; Guzman, M. I. Catechol oxidation by
ozone and hydroxyl radicals at the air−water interface. Environ. Sci.
Technol. 2014, 48, 14352−14360.
AUTHOR INFORMATION
Corresponding Author
E-mail: marcelo.guzman@uky.edu. Tel: 859-323-2892.
■
*
(15) Pillar, E. A.; Zhou, R.; Guzman, M. I. Heterogeneous oxidation
of catechol. J. Phys. Chem. A 2015, 119, 10349−10359.
(16) Chang, J. L.; Thompson, J. E. Characterization of colored
products formed during irradiation of aqueous solutions containing
H O and phenolic compounds. Atmos. Environ. 2010, 44, 541−551.
Notes
2
2
The authors declare no competing financial interest.
(
17) Warneck, P. Multi-phase chemistry of C2 and C organic
3
compounds in the marine atmosphere. J. Atmos. Chem. 2005, 51, 119−
59.
18) Kawamura, K.; Tachibana, E.; Okuzawa, K.; Aggarwal, S.;
ACKNOWLEDGMENTS
We thank research funding from the National Science
foundation under NSF CAREER award CHE-1255290.
■
1
(
Kanaya, Y.; Wang, Z. High abundances of water-soluble dicarboxylic
acids, ketocarboxylic acids and α-dicarbonyls in the mountaintop
aerosols over the North China Plain during wheat burning season.
Atmos. Chem. Phys. 2013, 13, 8285−8302.
ABBREVIATIONS
aqSOA, aqueous secondary organic aerosol; DNPH, 2,4-
dinitrophenylhydrazine; DSS, 3-(trimethylsilyl)-1-propanesul-
■
(19) Fu, P.; Kawamura, K.; Usukura, K.; Miura, K. Dicarboxylic acids,
•
ketocarboxylic acids and glyoxal in the marine aerosols collected
fonic acid sodium salt; GA, glyoxylic acid; G , dihydroxyacetic
during a round-the-world cruise. Mar. Chem. 2013, 148, 22−32.
•
acid radical; int, integral; K , glycolic acid radical; K , carbonyl
hyd
(
20) Mkoma, S.; Kawamura, K. Molecular composition of
•
hydration equilibrium constant; K , Henry’s law constant; L , a
H
dicarboxylic acids, ketocarboxylic acids, α-dicarbonyls and fatty acids
in atmospheric aerosols from Tanzania, East Africa during wet and dry
seasons. Atmos. Chem. Phys. 2013, 13, 2235−2251.
•
•
generic radical; M , hydrated glyoxal radical; MO , glyoxyl
2
peroxyl radical; δ, chemical shift
(21) Lin, G.; Sillman, S.; Penner, J. E.; Ito, A. Global modeling of
REFERENCES
SOA: the use of different mechanisms for aqueous-phase formation.
Atmos. Chem. Phys. 2014, 14, 5451−5475.
(22) Bao, L.; Matsumoto, M.; Kubota, T.; Sekiguchi, K.; Wang, Q.;
Sakamoto, K. Gas/particle partitioning of low-molecular-weight
dicarboxylic acids at a suburban site in Saitama, Japan. Atmos. Environ.
2012, 47, 546−553.
■
(
1) Ervens, B. Modeling the processing of aerosol and trace gases in
clouds and fogs. Chem. Rev. 2015, 115, 4157−4198.
2) McNeill, V. F. Aqueous organic chemistry in the atmosphere:
Sources and chemical processing of organic aerosols. Environ. Sci.
Technol. 2015, 49, 1237−1244.
(
I
DOI: 10.1021/acs.jpca.6b00225
J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
(
23) Harvey, H. W. The Chemistry and Fertility of Sea Waters, 2nd ed.;
Cambridge University Press: New York, 1957.
24) Guzman, M. I.; Athalye, R. R.; Rodriguez, J. M. Concentration
(45) Fu, H.; Ciuraru, R.; Dupart, Y.; Passananti, M.; Tinel, L.;
Rossignol, S.; Perrier, S.; Donaldson, D. J.; Chen, J.; George, C.
Photosensitized production of atmospherically reactive organic
compounds at the air/aqueous interface. J. Am. Chem. Soc. 2015,
137, 8348−8351.
(
effects and ion properties controlling the fractionation of halides
during aerosol formation. J. Phys. Chem. A 2012, 116, 5428−5435.
(
25) Noziere, B.; Dziedzic, P.; Cor
́
dova, A. Products and kinetics of
(46) Leermakers, P. A.; Vesley, G. F. The photochemistry of α-keto
acids and α-keto esters. I. Photolysis of pyruvic acid and benzoylformic
acid. J. Am. Chem. Soc. 1963, 85, 3776−3779.
the liquid-phase reaction of glyoxal catalyzed by ammonium ions
+
(
NH ). J. Phys. Chem. A 2009, 113, 231−237.
4
(47) Steenken, S.; Sprague, E. D.; Schulte-Frohlinde, D. Photo-
(
26) Manahan, S. E. Environmental Chemistry, 9th ed.; CRC Press:
Boca Raton, FL, 2010.
27) Zhou, R.; Guzman, M. I. CO2 reduction under periodic
illumination of ZnS. J. Phys. Chem. C 2014, 118, 11649−11656.
28) Zhou, R.; Guzman, M. I. Photocatalytic reduction of fumarate to
succinate on ZnS mineral surfaces. J. Phys. Chem. C 2016, 120, 7349−
357.
29) Go
fragmentation of α-oxocarboxylic acids in aqueous solution. An EPR
study. Formation of semidione radicals by decarboxylative substitution
of α-oxocarboxylic acids by acyl radicals-I: Glyoxylic, pyruvic, α-
oxobutyric, α-oxoglutaric and α-oxoisocaproic acids. Photochem.
Photobiol. 1975, 22, 19−27.
(
(
(
48) Zeldes, H.; Livingston, R. Paramagnetic resonance study of
liquids during photolysis. IX. Oxalic acid and its esters. J. Phys. Chem.
970, 74, 3336−3346.
49) Adams, G. E.; Hart, E. J. Radiolysis and photolysis of aqueous
formic acid. Carbon monoxide formation. J. Am. Chem. Soc. 1962, 84,
994−3999.
50) Flyunt, R.; Schuchmann, M. N.; von Sonntag, C. A common
carbanion intermediate in the recombination and proton-catalysed
7
(
̈
rner, H.; Kuhn, H. J. Photodecarboxylation of phenyl-
1
(
glyoxylic acid: Influence of para-substituents on the triplet state
properties. J. Chem. Soc., Perkin Trans. 2 1999, 2671−2680.
(
30) Eugene, A. J.; Xia, S.-S.; Guzman, M. I. Negative production of
3
(
acetoin in the photochemistry of aqueous pyruvic acid. Proc. Natl.
Acad. Sci. U. S. A. 2013, 110, E4274−E4275.
(
31) Sorensen, P. E.; Bruhn, K.; Lindelov, F.; et al. Kinetics and
.
−
disproportionation of the carboxyl radical anion, CO2 , in aqueous
equilibria for the reversible hydration of the aldehyde group in
glyoxylic acid. Acta Chem. Scand. 1974, 28, 162−168.
(
solution. Chem. - Eur. J. 2001, 7, 796−799.
(51) Buxton, G. V.; Sellers, R. M. Acid dissociation constant of the
32) Plath, K. L.; Axson, J. L.; Nelson, G. C.; Takahashi, K.; Skodje,
carboxyl radical. Pulse radiolysis studies of aqueous solutions of formic
acid and sodium formate. J. Chem. Soc., Faraday Trans. 1 1973, 69,
R. T.; Vaidaa, V. Gas-phase vibrational spectra of glyoxylic acid and its
gem diol monohydrate. Implications for atmospheric chemistry. React.
Kinet. Catal. Lett. 2009, 96, 209−224.
5
(
55−559.
52) Bothe, E.; Schuchmann, M. N.; Schulte-Frohlinde, D.; von
(
33) Guzman, M. I.; Hildebrandt, L.; Colussi, A. J.; Hoffmann, M. R.
Cooperative hydration of pyruvic acid in ice. J. Am. Chem. Soc. 2006,
28, 10621−10624.
34) Shapiro, E. L.; Szprengiel, J.; Sareen, N.; Jen, C. N.; Giordano,
Sonntag, C. Hydroxyl radical-induced oxidation of ethanol in
oxygenated aqueous solutions. A pulse radiolysis and product study.
Z. Naturforsch., B: J. Chem. Sci. 1983, 38, 212−219.
1
(
(
́
53) Vel Leitner, N. K.; Dore, M. Mecanisme d’action des radicaux
M. R.; McNeill, V. F. Light-absorbing secondary organic material
formed by glyoxal in aqueous aerosol mimics. Atmos. Chem. Phys.
OH sur les acides glycolique, glyoxylique, acetique et oxalique en
solution aqueuse: Incidence sur la consammation de peroxyde
d’hydrogene dans les systemes H O /UV et O /H O . Water Res.
2
(
009, 9, 2289−2300.
2
2
3
2
2
35) Updyke, K. M.; Nguyen, T. B.; Nizkorodov, S. A. Formation of
1
(
997, 31, 1383−1397.
brown carbon via reactions of ammonia with secondary organic
aerosols from biogenic and anthropogenic precursors. Atmos. Environ.
54) Simic, M.; Neta, P.; Hayon, E. Pulse radiolysis of aliphatic acids
in aqueous solutions. II. Hydroxy and polycarboxylic acids. J. Phys.
2
012, 63, 22−31.
36) Jones, M.; Fleming, S. A. Organic Chemistry, 4th ed.; W. W.
Norton & Company Inc.: New York, 2010.
37) Laskin, A.; Laskin, J.; Nizkorodov, S. A. Chemistry of
Atmospheric Brown Carbon. Chem. Rev. 2015, 115, 4335−4382.
38) Lin, P.; Laskin, J.; Nizkorodov, S. A.; Laskin, A. Revealing brown
Chem. 1969, 73, 4214−4219.
(
(55) Schaefer, T.; van Pinxteren, D.; Herrmann, H. Multiphase
chemistry of glyoxal: Revised kinetics of the alkyl radical reaction with
−
(
molecular oxygen and the reaction of glyoxal with OH, NO , and SO
3
4
in aqueous solution. Environ. Sci. Technol. 2015, 49, 343−350.
(
(56) Adams, G. E.; Willson, R. L. Pulse radiolysis studies on the
oxidation of organic radicals in aqueous solution. Trans. Faraday Soc.
1969, 65, 2981−2987.
carbon chromophores produced in reactions of methylglyoxal with
ammonium sulfate. Environ. Sci. Technol. 2015, 49, 14257−14266.
(
39) Phillips, S. M.; Smith, G. D. Light absorption by charge transfer
(57) Wasa, T.; Musha, S. The polarographic behavior of glyoxal and
its related compounds. Bull. Univ. Osaka Prefec. Ser. A 1970, 19, 169−
180.
complexes in brown carbon aerosols. Environ. Sci. Technol. Lett. 2014,
, 382−386.
40) Zhao, R.; Lee, A. K. Y.; Huang, L.; Li, X.; Yang, F.; Abbatt, J. P.
1
(
(58) Paul, H.; Fischer, H. Elektronenspinresonanz freier radikale bei
D. Photochemical processing of aqueous atmospheric brown carbon.
Atmos. Chem. Phys. 2015, 15, 6087−6100.
(
photochemischen reaktionen von ketonen in Lo
973, 56, 1575−1594.
59) Epstein, S. A.; Nizkorodov, S. A. A comparison of the chemical
̈
sung. Helv. Chim. Acta
1
(
41) Back, R. A.; Yamamoto, S. The gas-phase photochemistry and
sinks of atmospheric organics in the gas and aqueous phase. Atmos.
thermal decomposition of glyoxylic acid. Can. J. Chem. 1985, 63, 542−
48.
42) Renard, P.; Reed Harris, A. E.; Rapf, R. J.; Ravier, S.; Demelas,
Chem. Phys. 2012, 12, 8205−8222.
5
(
(60) Finlayson-Pitts, B. J.; Pitts, J. Chemistry of the Upper and Lower
Atmosphere; Academic Press: San Diego, CA, 1999.
61) Schone, L.; Herrmann, H. Kinetic measurements of the
C.; Coulomb, B.; Quivet, E.; Vaida, V.; Monod, A. Aqueous phase
oligomerization of methyl vinyl ketone by atmospheric radical
reactions. J. Phys. Chem. C 2014, 118, 29421−29430.
(
̈
reactivity of hydrogen peroxide and ozone towards small atmospheri-
cally relevant aldehydes, ketones and organic acids in aqueous
solutions. Atmos. Chem. Phys. 2014, 14, 4503−4514.
(
43) Renard, P.; Siekmann, F.; Gandolfo, A.; Socorro, P.; Salque, G.;
Ravier, S.; Quivet, E; Clement, J. L.; Traikia, M.; Delort, A. M.; et al.
́
Radical mechanisms of methyl vinyl ketone oligomerization through
aqueous phase OH-oxidation: on the paradoxical role of dissolved
molecular oxygen. Atmos. Chem. Phys. 2013, 13, 6473−6491.
(
44) Ciuraru, R.; Fine, L.; van Pinxteren, M.; D’Anna, B.; Herrmann,
H.; George, C. Photosensitized production of functionalized and
unsaturated organic compounds at the air-sea interface. Sci. Rep. 2015,
5, 12741.
J
DOI: 10.1021/acs.jpca.6b00225
J. Phys. Chem. A XXXX, XXX, XXX−XXX