The Henry's law constants of glyoxal, glycolic acid and glyoxylic acid in pure water were determined over the range of 278 and 308 K for the first time by a bubble column technique. These compounds were chosen because of their perceived involvement in the formation of secondary organic aerosol through in-cloud processing pathways. The experimentally determined Henry's law constants are: glyoxal, KH = 4.19 × 105 × exp[(62.2 × 103/R) × (1/T − 1/298)]; glycolic acid, KH = 2.83 × 104 × exp[(33.5 × 103/R) × (1/T − 1/298)]; and glyoxylic acid, KH = 1.09 × 104 × exp[(40.0 × 103/R) × (1/T − 1/298)]. The Henry's law constants of glyoxal in the presence of sodium chloride and sodium sulfate were also determined at 298 K. While the glyoxal KH is enhanced by less than three times in the presence of chloride in the range of 0.05–4.0 M ionic strength, the presence of sulfate at 0.03 M ionic strength increases the glyoxal KH by 50 times.
 The Henry's law constant, KH, is a key parameter in estimating the magnitude, rate, and direction of the flux of solutes between gas and aqueous phases [Betterton, 1992]. Recent research has revealed that a number of C2 bifunctional compounds (i.e., glycolaldehyde, glyoxal, glyoxylic acid and glycolic acid) are precursors in the aqueous-phase formation of oxalic acid [Blando and Turpin, 2000; Warneck, 2003]. Oxalic acid is the single most abundant water-soluble organic compound that has been identified in ambient aerosols and in-cloud processing is its dominant formation pathway [Sorooshian et al., 2006; Yu et al., 2005]. The C2 bifunctional compounds, especially glyoxal, have numerous volatile organic compound (VOC) precursors and are relatively abundant in the atmosphere. Typical ambient gas-phase concentrations of glyoxal range from 20 pptv (∼40 ng m−3) in rural environments to 2.0 ppbv (∼4.0 μg m−3) in urban environments [e.g., Munger et al., 1995; Ho and Yu, 2002; Volkamer et al., 2005]. The ambient abundance of the other three C2 bifunctional compounds was less frequently measured. The cloud-processing of glycolaldehyde and glyoxal, which are formed mainly in the gas phase, is a potentially important pathway leading to formation of secondary organic aerosols (SOA). The evaluation of this pathway requires the knowledge of the Henry's law constants of these compounds and their intermediate aqueous-phase oxidation products (i.e., glycolic acid and glyoxylic acid).
 Of the four C2 bifunctional compounds, glycolaldehyde was measured to have a KH value of 4.14 × 104 M atm−1 at 298 K and 1.56 × 104 M atm−1 at 318 K [Betterton and Hoffmann, 1988]. Two studies in the literature reported KH measurements of glyoxal; one study provided a lower limit estimate of 3 × 105 M atm−1 [Betterton and Hoffmann, 1988] and the second study reported a value of 3.6 × 105 M atm−1 in 100% seawater, based on only two measurements [Zhou and Mopper, 1990]. The KH values of glycolic acid and glyoxylic acid have not been measured experimentally.
 Both glyoxal and glyoxylic acid undergo hydration in water (see reactions (R1) and (R2) below). Glyoxylic acid and glycolic acid also undergo acid dissociation (Reaction (R2)). The experimentally determined values in this work are the effective Henry's law constants, which include terms for the hydration constants and the acid dissociation constants in the case of carboxylic acids [Betterton and Hoffmann, 1988]. That is, for glyoxal, KH = KH′(1 + K1hyd + K1hydK2hyd); for glyoxylic acid, KH = KH′(1 + Khyd + KhydKa/[H+]); and for glycolic acid, KH = KH′(1 + Ka/[H+]), where KH′ is the intrinsic Henry's law constant. The intrinsic Henry's law constant can be calculated from the effective Henry's law constant for compounds with known hydration and acid dissociation constants.
2. Experimental Section
 Glyoxal trimer dihydrate (Sigma), glycolic acid (Acros), glyoxylic acid (Acros), 2, 4-dinitrophenylhydrazine (DNPH) (Sigma) were used as received. A glyoxal solution of ∼5 mM was prepared by dissolving glyoxal trimer dihydrate in water and the solution was allowed to sit in the dark overnight to ensure complete hydrolysis. The glyoxal solution at such a concentration level contained only monomers [Whipple, 1970].
 A bubble-column technique was used to measure KH. The apparatus, shown schematically in Figure 1, was modified from the design by Mackay et al.  and Betterton and Hoffmann . It consisted of a conditioning column, a stripped column and an absorbing column of the same dimensions (70 cm long and 4 cm i.d.), all placed in a temperature-controlled water bath. The columns were connected using silicone rubber tubing. A heating tape was used to maintain the tubing connection at least 5°C above the operating temperature of the water bath to prevent condensation. Glass-wool plugs were inserted to eliminate mist carry-over. The gas exiting the stripped column was collected by an absorbing solution. In a typical experiment, ultra pure nitrogen gas passed through the first column containing 1 L of water (at a liquid depth of 64 cm) where it was conditioned to the required temperature and humidity. The conditioned gas was then bubbled through the stripped column containing 1 L of 5 mM solution of the target compound at a flow rate of 250 mL min−1. All solutions were prepared within three days prior to each experiment. The system was equilibrated at a specific temperature for 12 h before sample collection for KH measurements at this temperature. The experiment at each temperature was repeated three times or more.
Mackay et al.  have shown that the mass transfer rate of a solute from the aqueous phase to the gas phase in a bubble column is described by the following equation:
where G is the gas flow rate (L min−1), V is the volume of the liquid (L) in the stripped column, R is the gas constant (0.082 L atm K−1 mol−1), T is the system temperature (K), Co is the initial solute concentration and Ct is the solute concentration after time t (min) in the stripped column. A plot of ln(Ct/Co) against time yields a linear correlation with a slope of (−G/(KHVRT)), from which KH can be calculated. Ct in the stripped column was computed to be Co − Ca, where Ca is the concentration in the absorbing column. Ca was monitored by removing aliquots of the absorbing solution at different time intervals for chemical analysis. Five or more samples were taken in each experiment at a given temperature. The sampling time interval between two consecutive samples was 30 min for glyoxylic acid and glycolic acid and varied from 1 to 6 h for glyoxal. An aliquot of 1 mL was withdrawn for glyoxylic acid and glycolic acid while an aliquot of 10–20 mL was removed for glyoxal. The higher KH values for glyoxal required longer sampling intervals and larger aliquots of solution for analysis. Ca was corrected to the volume of the stripped column. Glycolic acid and glyoxylic acid were analyzed using ion chromatography (IC) with 5 mM NaOH as eluent. Glyoxal was quantified in its DNPH derivative using High-Performance Liquid Chromatography with detection of light absorption at 420 nm.
3. Results and Discussion
3.1. Validation of the Bubble Column Technique
 The set-up was first optimized and validated by measuring the KH of acetic acid and formaldehyde. Their experimental KH values are known from previous studies. In the bubble column technique to obtain KH, the solute in the exit vapor must be in equilibrium with the liquid. The lower KH a compound has, the longer it takes to attain the gas-liquid equilibrium [Mackay et al., 1979]. Formaldehyde and acetic acid, having a lower KH than the three target compounds, require more stringent conditions (e.g., lower flow rate, higher liquid depth) to attain the gas-liquid equilibrium. A few studies measured the effective KH of formaldehyde at 298 K, reporting values of 3.0 × 103 M atm−1 [Betterton and Hoffmann, 1988], 3.1 × 103 M atm−1 in a seawater matrix [Zhou and Mopper, 1990], and 6.0 × 103 M atm−1 [Gaffney and Senum, 1984]. We measured formaldehyde KH to be 4.8 × 103 M atm−1, falling within the range of these reported values. Three studies reported the KH value of acetic acid at 298 K. They were 4.1 × 103 M atm−1 by Johnson et al. , 5.5 × 103 M atm−1 by Khan et al. , and 9.3 × 103 M atm−1 by Servant et al. . Our measurement of acetic acid KH was 5.0 × 103 M atm−1, also in the range of the published measurements. The measurement results for the KH of formaldehyde and acetic acid have demonstrated that the flow rate in the range of 200–300 mL min−1 and the liquid depth (64 cm) used in this work was adequate to ensure liquid-gas equilibrium.
3.2. Henry's Law Constant of Glyoxal
 Measurements of the glyoxal KH were made at four temperatures and the results are listed in Table 1. A plot of the KH measurements as a function of 1/T yields a solution enthalpy of −62.2 kJ mol−1. The measured glyoxal KH at 298 K was 4.2 × 105 M atm−1, in good agreement with the value of 3.6 × 105 M atm−1 in the seawater matrix measured by Zhou and Mopper  and consistent with the lower limit value of 3.0 × 105 M atm−1 reported by Betterton and Hoffmann  based on their analytical detection limit (∼5 μM) for aqueous-phase glyoxal.
Table 1. Summary of Effective Henry's Law Constants for Glyoxal, Glyoxylic and Glycolic Acid
KH (M atm−1)
KH, intrinsic (M atm−1)
The KH values as a function of temperature in the form of KH = KH,298K × exp[ × ( − )] derived from measurements at different temperatures in this work are: Glyoxal: KH = 4.19 × 105 × exp[ × ( − )]; glycolic acid: KH = 2.83 × 104 × exp[ × ( − )]; and glyoxylic acid: KH = 1.09 × 104 × exp[ × ( − )].
The intrinsic KH at 298 K for glyoxal was calculated from its effective KH using a hydration constant of 2.2 × 105 M measured by Wasa and Musha .
The intrinsic KH at 298 K for glyoxylic acid was calculated from its effective KH using a hydration constant of 3.0 × 102 M [Sørensen et al., 1974] and an acid dissociation constant of 3.47 × 10−4 M for the hydrated form of glyoxylic acid [Smith and Martell., 1977].
The intrinsic KH at 298 K for glycolic acid was calculated from its effective KH using an acid dissociation constant of 1.58 × 10−4 M [vel Leitner and Dore, 1997].
 Recently, Kroll et al.  measured the effective glyoxal KH to be 2.6 × 107 M atm−1 in a chamber study. This value was obtained from partition of glyoxal between gas phase and ammonium sulfate seed particles or mixed ammonium sulfate/sulfuric acid seed particles. The partition of glyoxal onto the two types of seed particles was found to yield the same effective Henry's law constant. On the basis of this observation, Kroll et al.  suggested that the high ionic strength (19.5–22.5 M) in the aerosol aqueous phase was responsible for the enhanced KH in comparison with the KH measured in the seawater matrix by Zhou and Mopper .
 We measured KH of glyoxal in NaCl and Na2SO4 solutions to investigate the effect of ionic strength. A series of experiments were conducted in NaCl solutions of different ionic strength ranging from 0.05 to 4.0 M. The measured KH was 1.90 × 106 M atm−1 at an ionic strength of 0.05 M and dropped by ∼50% to 8.50 × 105 M atm−1 when the ionic strength increased to 4.0 M. The KH value measured in the highest ionic strength NaCl solution was twice that measured in pure water. The enhancement of KH was possibly facilitated by the weak hydrogen bonding formed between Cl− ion and the −OH groups in the hydrated form of glyoxal [Wan and Yu, 2007]. Further increase in the NaCl concentration lowered the KH. It was likely a result of shift in the hydration equilibrium in favor of the aldehyde form since Cl− and Na+ ions compete for water molecules to form hydration shells [Schwarzenbach et al., 2003], leading to more prominent salt-out effect with increasing NaCl concentration.
 In the experiments carried out in Na2SO4 solutions, the ionic strength varied from 0.0003 to 0.03 M while the concentration of glyoxal was fixed at 10 mM. The corresponding sulfate:glyoxal molar ratio ranged from 0.01:1 to 1:1. KH was found to increase significantly with increasing SO42− concentration and reached a value of 2.40 × 107 M atm−1 in the solution having the highest ionic strength (0.03 M). This value was similar to the KH value derived by Kroll et al.  in their chamber study, but it was ∼50 times higher than the KH measured in pure water and 12 times higher than the KH measured in the NaCl solution at an ionic strength of 0.05 M. More experiments were conducted, in which the SO42− concentration was further increased to 0.15 M (ionic strength 0.225 M, and sulfate:glyoxal = 15:1) to make the sulfate:glyoxal ratio closer to those observed in ambient clouds and aerosols. It was found that the resulting effective KH was too high (>109 M atm−1) for measurement by our method, since a 6-h collection time was not long enough to obtain sufficient glyoxal in the absorbing column for analysis. Our measurement results pointed to that sulfate was a more important factor than the ionic strength in affecting the glyoxal KH. We have yet to understand the reactive interaction of sulfate ions with the hydrated glyoxal forms.
3.3. Henry's Law Constants of Glyoxylic and Glycolic Acid
 The measured KH values of glyoxylic and glycolic acid at four temperatures are given in Table 1. The solution pH was not controlled in the experiments. Consequently, the concentrations of the acids and their respective acid dissociation constants determined the resulting solution pH (Table 1). Plots of −lnKH versus 1/T yielded ΔH values of −33.5 and −40.0 kJ mol−1 for glycolic acid and glyoxylic acid, respectively. Our measured ΔH for glycolic and glyoxylic acids are in good agreement with ΔH values for the other carboxylic acids. Khan et al.  reported ΔH values of pyruvic, pentanoic, and hexanoic acid as −42.3, −54.7, and −52.4 kJ mol−1, respectively.
3.4. Atmospheric Implications
 The knowledge of KH makes it possible to assess the phase distribution, which in turn is a necessary step in evaluating the fate of glyoxal, glycolic acid and glyoxylic acid and their potential contributions to secondary organic aerosol. The partitioning of atmospheric species into cloud droplets involves multiple steps and the key step is transport across the interface [Schwartz, 1986]. The characteristic time to achieve interfacial equilibrium is strongly dependent on the Henry's law constant and can be calculated with the knowledge of KH and the accommodation coefficient of a species using the following equation [Betterton, 1992; Schwartz, 1986]:
where r is the radius of cloud droplet, Dg is the gas phase diffusion coefficient, α is the mass accommodation coefficient, and υ = (8RgT/πM)1/2 is the mean thermal velocity of molecules in the gas phase (M is the molecular weight). Schweitzer et al.  reported an average accommodation coefficient of glyoxal by water droplets to be 0.023 in the temperature range of 263–283 K. If we assume an accommodation coefficient of 0.01 (a lower value typically employed in cloud physics computation [Schwartz, 1986]), a cloud droplet radius of 5.0 μm and a temperature of 288 K, the characteristic times required to achieve gas-droplet equilibrium are 54 s for glyoxal, 4 s for glycolic acid (pH = 3), and 1.6 s for glyoxylic acid (pH = 3). The calculations indicate that the gas-droplet equilibrium could be readily established under atmospheric conditions.
 Using the KH values measured in pure water, we calculate that with the typical liquid water content of clouds of 0.1–1 g m−3 (i.e., 10−7–10−6 v/v), 78–97% of glyoxal, 17–67% of glycolic acid, and 6–38% of glyoxylic acid partition into the cloud water at equilibrium at 5°C (Figure 2). In other words, glyoxal predominately resides in clouds at equilibrium while a smaller but still significant fraction of glycolic acid and glyoxylic acid is in clouds at equilibrium. The significant partitioning of glyoxal in cloud water predicted by its Henry's law constant is confirmed by field measurements. Munger et al.  observed that up to 50% of glyoxal was partitioned in the aqueous phase during cloudy periods encountered in the Shenandoah National Park, USA during September 1990. Glyoxal in cloud/fog water was also detected in Riverside, California, USA [Munger et al., 1990].
 If we do not consider the enhancement by salt like NaCl or sulfate, the KH values predict that in wet aerosols (typical liquid water content in the range of 1–100 μg m−3), the fractions of glyoxal, glycolic acid, and glyoxylic acid in particle phase are negligible (Figure 2). The effective glyoxal Henry's law constant in pure water is apparently not high enough to explain its observed presence in particle phase. Matsunaga et al.  measured gas-particle partitioning of glyoxal in a forest atmosphere in Japan in August 2002 and reported that the fraction of glyoxal residing in the particle phase ranged from zero to 100% and averaged at 46%.
 In addition to liquid water content, atmospheric factors such as other constituents in cloud water/aerosols, acidity, and ionic strength could significantly affect the partitioning of glyoxal, as demonstrated by the observed higher KH values in the presence of NaCl and Na2SO4 in this work. Volkamer et al.  compared direct measurements of gas-phase glyoxal in Mexico City to experimentally constrained model prediction and concluded that there must be an additional glyoxal sink. They suggested reversible partitioning to aerosol liquid water with a KH as high as 4 × 109 M atm−1 to be a possible sink to account for the discrepancy between the model prediction and the measurements. Certain strong interactions between glyoxal with other aerosol constituents have to be involved to account for the high effective KH suggested by Volkamer et al. , since this suggested KH value is four orders of magnitude higher than the KH in pure water. Such a high KH is possible and in general agreement with our lower-limit estimate for KH in the presence of 0.15 M sulfate and 10 mM glyoxal. It is apparent that further work is needed to quantify the effects of sulfate, acidity, and ionic strength on partitioning of glyoxal between the gas and aqueous phases.
 This work was supported by the Research Grant Council of Hong Kong, China (621806 and 621708).