Determination of Gas/Particle Partitioning of Glyoxal and Other Bifunctional Species using the Annular Denuder-Filter Sampling Technique

Transcription

1 Determination of Gas/Particle Partitioning of Glyoxal and Other Bifunctional Species using the Annular Denuder-Filter Sampling Technique Simon Ip, Hilda Huang, Jian Zhen Yu Hong Kong University of Science and Technology May, 2010

2 Introduction OUTLINE Sources and abundance of glyoxal Secondary organic aerosols from glyoxal Literature of Gas/Particle measurements of glyoxal Experimental Sulfite-coated method H 2 O 2 vs. Barium Chloride Recovery and Collection Efficiency Results G/P measurements of glyoxal in Hong Kong Gas/particle partition coefficients from field study Estimated Effective Henry s law constant of glyoxal from field study Conclusions 2

3 Introduction

4 Sources of glyoxal VOC precursors Isoprene Terpenes Benzene Toluene Xylene Ethylene O glyoxal O 18 mm 20 C 4

5 Why glyoxal is special? No growth is observed for most carbonyls studied, even at high conc. (500 ppb to 5 ppm) At RH = 12%, no growth was observed at conc. as high as ~1 ppm of glyoxal Increase in particle volume was observed only when RH was higher than 40% No change in particle volume INCREASE in particle volume Liquid Water Content is Important! Kroll et al., JGR, ppb 2,4-hexadienal at t = 50 min 200 ppb glyoxal at t = 53 min 5

6 Knowledge Gap: SOA Model vs. Measurement Volkamer et al., GRL, 2006 Large Gap between modeled and measured SOA formation Current model parameterizations based on G/P thermodynamic partitioning of semi-volatile species of VOC oxidation underestimate OC aerosol concentrations as a function of photochemical age. 6

7 Available Literature for gas/particle glyoxal Matsunaga et al., Ortiz et al., Ortiz et al., Denuder XAD-7 XAD-4 coating + PFBHA + method BHA PFBHA Particle type PM4 PM2.5 TSP P, ng m G, ng m P/(P+G) 46% 80% 58% Flow rate 10 L/min 16.7 L/min 16.7 L/min Sampling Duration 3.5 hr 5.5 hr min 7

8 Experimental Methods

9 Annular Denuder-Filter Setup Quartz Filter Air Flow Denuder coated with sorbent KI-coated Denuder Cyclone XAD-7 Types of sorbents 1. XAD-4 2. XAD-4 + PFBHA 3. XAD-7 + OBE 4. PFBHA only XAD-4 9

10 The Use of Sulfite Coating Carbonyls + S(IV) Carbonyls S(IV) adduct Volatile Non-volatile Advantages Unlimited sulfite can be used for coating The reaction between sulfite and aldehydes is generally fast Carbonyl S(IV) adducts are non-volatile and therefore the effective sampling time is longer Disadvantages Removed by precipitation Not applicable for every carbonyl species The reaction between sulfite and ketone is generally much slower than aldehyde S(IV)-adduct formation 10

11 Extraction Procedures for Sulfite-Coated Denuder The denuder were extracted with DI water x 3 BaCl 2 is added into the sample extract The mixture is then centrifuged at RT for 30min Acidified PFBHA is added The extract was extracted x 3 with DCM Purged to near dryness under N 2 Reconstitution with Hexane and DCM Then 100 µl of BSTFA is added Baked at 60 C for an hour GC-MS analysis 11

12 Sulfite coated vs. XAD coated denuder This study XAD coated Sampling time Integrity of GC mass spectrum Coating and extraction Procedure Loading derivatizing reagent, mole 12 hours Clean Simple 0.24 Sulfite 5.5 hours Collection efficiency a ~85% ~ 60% Deteriorated Tedious PFBHA a Average collection efficiencies for glyoxal, methylglyoxal, glycoaldehyde, glyoxylic acid and pyruvic acid 12

13 Gas Chromatogram Glycoaldehyde Relative Abundance Hydroxyacetone Glyoxylic acid Pyruvic acid Glyoxal Methylglyoxal Time (min) Relative Abundance Glycoaldehyde Hydroxyacetone Glyoxylic acid Pyruvic acid Glyoxal Methylglyoxal Time (min) Figure 2. Total ion chromatogram (Top) and m/z = 181 ion chromatogram (bottom) of daytime denuder sample on 15 th January,

14 Sampling at Air Pollution Monitoring Site Important gas-phase species (CO, NO x, O 3, SO 2, etc.) and PM 10 - EPD PM 2.5 OC, EC Ionic species (SO 4 2-, NO 3-, Cl -, oxalate, Na +, NH 4+, K +, Ca 2+, Mg 2+, and other dicarboxylic acid) Sampling at Tsuen Wan Gas- and particle-phase carbonyl species 14

15 Results and Discussion

16 Gas/Particle Distribution in Hong Kong Maximum, minimum, average, and median of individual bifunctional carbonyl concentrations in the gas phase. Summer Gas phase (µg m -3 ) Winter Gas phase (µg m -3 ) Max. Min. Median Mean Max. Min. Median Mean Glycoaldehyde (MW: 60.05) Glyoxylic acid (MW: 74.04) Pyruvic acid (MW: 88.06) Glyoxal (MW: 58.04) Methylglyoxal (MW: 72.06) Maximum, minimum, average, and median of individual bifunctional carbonyl concentrations in the particle phase. Summer Particle phase (ng m -3 ) Winter Particle phase (ng m -3 ) Max. Min. Median Mean Max. Min. Median Mean Glycoaldehyde (MW: 60.05) 39.8 LOD Glyoxylic acid (MW: 74.04) Pyruvic acid (MW: 88.06) Glyoxal (MW: 58.04) Methylglyoxal (MW: 72.06)

17 Literature vs. Findings Matsunaga et al., Ortiz et al., Ortiz et al., This Study This Study Summer Winter Denuder XAD-7 XAD-4 Sulfite Sulfite coating + PFBHA method BHA PFBHA Glycerol Glycerol [Glyoxal] particle, ng m [Glyoxal] gas, ng m P/(P+G), % 46% 80% 58% 2.8% 3.4% Flow rate 10 L min L min L min -1 5 L min -1 5 L min -1 Sampling Duration 3.5 hr 5.5 hr min 12hr 12hr 17

18 Partition coefficient, K p Compounds Summer K p,i (n=17) (theoretical) Summer K p,i (n=17) (experimental) Winter K p,i (n=20) (theoretical) Winter K p,i (n=20) (experimental) Glyoxal (3.95±0.96)x10-10 (1.43±1.33)x10-3 (3.37±1.07)x10-10 (1.18±1.12)x10-3 Methylglyoxal (9.11±2.23)x10-10 (4.62±2.82)x10-4 (7.79±2.48)x10-10 (2.98±5.04)x10-3 Glycoaldehyde (5.33±1.30)x10-8 (6.48 ±5.74)x10-2 (4.56±1.45)x10-8 (2.30±3.47)x10-3 Glyoxylic Acid (5.83±1.42)x10-7 (1.63±1.33)x10-3 (4.98±1.59)x10-7 (1.82±2.19)x10-2 K K Pyruvic Acid (1.43±0.32)x10-7 (1.27±0.89)x10-2 (1.23±0.39)x10-7 (0.68±1.16)x10-2 C p, i (exp erimental = (Odum et al., 1996) C FSP p, i ) p, i ( theoretical) = MW g, i 760RTf om om 6 10 ζ p i L, i (Pankow, 1994a; Pankow, 1994b) 18

19 Experimental H eff of glyoxal Compounds Glyoxal Methylglyoxal Glycoaldehyde Glyoxylic Acid Pyruvic Acid a [Ip et al., 2009] b [Betterton and Hoffmann, 1998] c [Khan et al., 1995] Summer Effective K H in Wet Aerosols, M atm -1 (1.21 ± 1.36)x 10 9 (1.66 ± 1.42)x 10 8 (8.50 ± 7.90)x 10 7 (1.78± 1.30)x 10 8 (1.61 ± 1.30)x 10 9 Winter Effective K H in Wet Aerosols, M atm -1 (1.40 ± 0.85)x 10 9 (3.90 ± 8.86)x 10 8 (2.06 ± 2.89)x 10 8 (1.81 ± 1.72)x 10 9 (0.85 ± 1.21)x 10 9 Effective K H in Pure Water, M atm x 10 5a 3.7 x 10 3b 4.1 x 10 4b 1.1 x 10 4a 3.1 x 10 5c K H = FP ( 1 F) RT ( LWC) F = C p, i ( C p, i + Cg, i ) (Seinfeld and Pandis, 1998) where C g,i and C p,i are the concentrations of one species in gas phase and particle phase, respectively (µg m -3 ), R is the ideal gas constant ( m 3 Pa K -1 mol -1 ), P is the atmospheric pressure (Pa), T is the temperature (K) and LWC is aerosol liquid water content (µg m -3 ) calculated by AIM III. Experimental temperatures were used for calculations. 19

20 Summary CE prove that sulfite-coated denuder is capable of performing 12 hours Recovery studies prove that S(IV) precipitation by BaCl would not cause artifact Large gap between experimental K p and theoretical K P suggested that g/p partition theory cannot fully explain the partition pathways of glyoxal Estimated H eff from field study (1.12 x 10 9 M atm -1 ) are in agreement to H eff measured in acidic solution (>10 9 M atm -1, Ip et al., 2009). More studies are needed to investigate the effect of ionic strength, acidity and other major aerosol components on H eff of methylglyoxal, glyoxylic acid, pyruvic acid and glycoaldehyde, etc. 20

21 Acknowledgement Prof. Jianzhen YU Hilda Huang Dr. Eric Wan and Dr. Steven Ho ENVF Staffs Mr. To, Mr. Tsui, Ah Pong, Waiman, Ah Man Group Mates Elber Sit, Emily Au, Yu Huan, Li yun chun, Hu Di, Tony So, Wu Cheng, Huang xiao feng, Yuan zibing, Eric Xue, Lin Peng Thank you! Q & A 21