Environ. Sci. Technol. 2006, 40, 6318-6323
Oligomer Formation in Evaporating Aqueous Glyoxal and Methyl Glyoxal Solutions KIRSTEN W. LOEFFLER, CHARLES A. KOEHLER, NICHOLE M. PAUL, AND DAVID O. DE HAAN* Chemistry Department, University of San Diego, 5998 Alcala Park, San Diego, California 92110
Glyoxal and methyl glyoxal are common secondary atmospheric pollutants, formed from aromatic and terpene precursors. Both compounds are extremely watersoluble due to dihydrate formation and partition into cloudwater. In this work, FTIR-ATR and mass measurements indicate that both compounds remain primarily in the condensed phase due to oligomer formation when aqueous solution droplets are evaporated, regardless of concentration (g1 mM) or, for glyoxal, droplet evaporation rate. FTIR spectral analyses suggest that oligomer formation is triggered by conversion from dihydrate to monohydrate forms, which are still nonvolatile but contain reactive carbonyl groups. Methyl glyoxal hemiacetal formation is observed by changes in the C-O/CdO stretch peak area ratio. The formation of glyoxal oligomers is detected by a dramatic shift of the C-O stretching peak toward low frequencies. Glyoxal oligomer peaks at 1070 cm-1, 950 cm-1, and 980 cm-1 are assigned to free C-OH stretch, dioxolane-linked C-OC asymmetric stretch, and tentatively to non-dioxolanelinked C-OC stretches, respectively. Acids have little effect on glyoxal oligomer formation; however, base interrupts oligomer formation by catalyzing glyoxal hydration and disproportionation to glycolic acid. Since glyoxal and methyl glyoxal are commonly found in cloudwater and are expected to remain largely in the aerosol phase when cloud droplets evaporate, this process may be a source of secondary organic aerosol by cloud processing.
Introduction
which results in effective Henry’s law constants of K*H ) 2.6 × 107 M/atm for glyoxal in ammonium sulfate aerosol particles at ∼50 %RH (7) and 3.6 × 105 M/atm in bulk seawater (8). Questions about the fate of such compounds have driven recent interest in secondary organic aerosol (SOA) formation by cloud processing (9). In a recent study, glyoxal and methyl glyoxal were identified as key intermediates in a pathway by which isoprene could form SOA (10). In this mechanism, gas-phase oxidation produces both compounds, which are taken up by cloud droplets. There, aqueous phase oxidation by hydroxyl radicals produces nonvolatile species such as glyoxylic acid and oxalic acid (11, 12), which are left in the particle phase when the cloud droplet evaporates. Barring such oxidation, uptake into the aqueous phase is typically assumed to be reversible (12). Parallel interest has focused on glyoxal as a source of recently discovered polymeric material that makes up a substantial fraction of aerosol mass (13-17). Glyoxal has been found to easily transfer to the aerosol phase (7, 18-23) at relative humidity levels above 26% (23). Aerosol mass spectrometer measurements (21) indicate glyoxal forms acetal oligomers in the condensed phase. Oligomer formation involves the nucleophilic attack of an OH group on the reactive carbonyl of a neighboring molecule (reaction 2). If both glyoxal monomers are singly hydrated at the outset, a dioxolane ring can be formed through a second, intramolecular nucleophilic attack without any intervening dehydration steps.
Similarly, if the gem-diol group on the product dimer molecule is dehydrated, a third glyoxal hydrate subunit can be added via a dioxolane ring (reaction 3) without intervening dehydration steps. It is important to note that, while more glyoxal subunits can be added to the trimer, no further dioxolane rings can be formed by this mechanism.
Glyoxal and methyl glyoxal are produced in high yield by the atmospheric oxidation of anthropogenic aromatic hydrocarbons (1-3) and in smaller yields by biogenic terpenes and isoprene (4), making them ubiquitous in the lower terrestrial troposphere. Because of its high solubility in water, glyoxal is the dominant aldehyde found in dew and fogwater collected at urban and suburban sites (5, 6). Of aldehydes, only formaldehyde is present at higher concentrations in rainwater (6). The high solubility of glyoxal is driven by hydration of both aldehyde groups (reaction 1) * Corresponding author phone: (619)260-6882; fax: (619)2602211; e-mail:
[email protected]. 6318
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 20, 2006
Experimental evidence for this mechanism will be presented in this article. The ultimate conversion of glyoxal into these oligomer hydrates may explain why glyoxal uptake can exceed 10.1021/es060810w CCC: $33.50
2006 American Chemical Society Published on Web 09/12/2006
the prediction of Henry’s law equilibrium (7, 24), making a volatile compound a potentially significant source of aerosol mass. Similar oligomer formation mechanisms have been proposed for methyl glyoxal and other small aldehydes (13). In this paper, we report FTIR-attenuated total reflectance (ATR) and mass measurements of drying glyoxal solutions that link the two lines of inquiry, cloud processing, and glyoxal polymerization. We examine the effects of glyoxal concentration and evaporation rate on the ability of glyoxal to escape the aqueous phase. Temperature effects are beyond the scope of this study. Our goal is to answer the following questions: what happens to glyoxal when a glyoxal-containing cloud droplet evaporates and could oligomer formation be significant as a mechanism of SOA formation by cloud processing.
Experimental Section Solutions of glyoxal monomer dihydrate were made by hydrolyzing glyoxal trimer dihydrate (GTD) powder (92%, ICN Biomedicals, Inc.) in deionized water (18 MΩ) overnight, followed by dilution with water or water/acetone mixtures (20 to 99% water by volume). Acetone increased the initial evaporation rate of the solution but was shown to have no effect on subsequent spectra except at the highest glyoxal dilution (1 mM), where nonvolatile impurities in the acetone impeded glyoxal oligomer formation slightly. Glyoxylic acid and methyl glyoxal solutions were diluted from 50% w/w (Sigma Aldrich) and 40% w/w (Alfa Aesar) stock solutions, respectively. Solutions of oxalic acid (98.8%, JT Baker), maleic acid, sodium chloride and malonic acid (99%, Sigma Aldrich), nonanoic acid (97%, MP Biomedicals), and potassium hydrogen phthalate (99.95%, Mallingckrodt) were made from solid standards and used without further purification. Sulfuric and hydrochloric acid solutions were diluted from concentrated stock solutions (certified ACS grade, Fisher). In each experiment involving glyoxal, methyl glyoxal, acid, or base, single 1-µL droplets of solutions in the concentration range 1-1000 mM were autopipetted onto the diamond crystal surface of a 9-reflection ATR attachment (DuraSamplIR, Smiths Detection, Danbury, CT) in an FTIR spectrometer (JASCO Model 480). After a few minutes of solvent evaporation, 12 spectra were recorded (each with 16 accumulations, 2 cm-1 resolution) with short delays (typically ∼15 s) after each spectra were recorded. The spectrum of the clean crystal served as the background. Spectra were observed to be stable by the end of the series of measurements. Standard glyoxal and methyl glyoxal solution spectra were taken using the same conditions but with a water drop as the background and without any evaporation delays. Standard spectra of solid glyoxal trimer dihydrate were recorded by pressing crystals onto the diamond surface of a more rugged single-reflection ATR attachment (Golden Gate model, Specac Inc.) mounted in the same model FTIR spectrometer. Masses were measured to (0.1 mg on an analytical balance (Mettler Toledo AB204S). In mass loss experiments, multiple 25-µL droplets were evaporated to dryness under controlled humidity and then desiccated overnight to remove residual water. Humidity was logged by a NIST-certified hygrometer (Fisher). Humidity was controlled in mass loss and certain ATR experiments by mixing metered humidified and dry flows of either filtered nitrogen (Airgas, Inc.) or purified house air (dried, filtered, and passed through molecular sieve and Purafil) in small chambers made from polyethylene teraphthalate or polypropylene. pH measurements were made by calibrated ISFET sensors (Model IQ150, IQ Scientific Instruments, Inc., San Diego, CA).
Results Glyoxal. During ambient drying of single 1-µL drops of 10 mM glyoxal solution, the observed infrared ATR spectra in
FIGURE 1. 10 mM glyoxal solution drying on 9-reflection ATR diamond crystal at 22 °C and 35% RH. Gray line: 6.7 min. Gray dash: 7.6 min. Black dash: 8.5 min. Black line: 10.2 min. Green line: 16.3 min. Blue line: 0.9 M glyoxal dihydrate standard solution. Red line: solid glyoxal trimer dihydrate pressed onto single-reflection diamond ATR. the C-O stretch region around 1000 cm-1 change dramatically over a 10-min period (Figure 1). In experiments that contained acetone, the acetone evaporated within 1 min and did not affect subsequent spectra. Initial spectra taken as water evaporates show the C-OH stretching band at around 1070 cm-1 growing rapidly as the concentration of glyoxal hydrate increases. This band initially matches the frequency of a 0.9 M standard solution of glyoxal monomer dihydrate, as expected. Interestingly, as the broad liquid water peaks at 1640 and 700 cm-1 disappear at the liquid-to-solid phase change at 7 min, the C-O stretching band of glyoxal hydrate begins a shift toward low frequencies. Within a few minutes, strong bands are visible at 980 cm-1, then 950 cm-1 (Figure 1), the latter of which matches the strongest absorption band of the solid glyoxal trimer. These observations demonstrate directly that when an aqueous solution of glyoxal evaporates, glyoxal itself polymerizes and remains in the condensed phase. This is consistent with earlier particle chamber results where lowering the relative humidity (from 35 to 14%) did not result in size decreases in particles that had earlier been formed from glyoxal and water vapor (23). Simple mass measurements also show that glyoxal remains in the condensed phase when a glyoxal-containing droplet dries. If all glyoxal is converted to the trimer (reaction 3), drying 500 µL of 1.0 M glyoxal will result in 35 mg of residue. Mass-based percent recovery of glyoxal in dried residues was 101 ( 2% for 10 experiments where drying occurred at 22 °C and with humidity levels held at values between 30 and 98 %RH, followed by subsequent desiccation to remove 1-2 mg of residual water. As shown in the Supporting Information, glyoxal recovery showed no dependence on the humidity level during drying: the data include a run where the humidity remained above 90% RH over a 2-day drying process, resulting in a solid-phase glyoxal recovery of 103 ( 2%. A slight excess in residue mass may be due to the presence of some glyoxal dimers, which have a higher water/glyoxal ratio than that of the glyoxal trimers. We attribute the 950 cm-1 band (Figure 1) to an asymmetrical C-O-C stretch of glyoxal oligomers linked by 5-membered dioxolane rings (reaction 2). This band nearly matches the asymmetrical C-OC stretch of 1,3-dioxolane at 940 cm-1 (25). The strong peak around 980 cm-1, though prominent in all IR spectra of dried glyoxal solutions with concentrations
