Speciation of water-soluble organic carbon compoundsin boundary layer aerosols during theLBA/CLAIRE/SMOCC-2002 campaign

Magda Claeys1, Vlada Pashynska1, Reinhilde Vermeylen1, Gyorgy Vas1, Jan Cafmeyer2, Willy Maenhaut2, and Paulo Artaxo3

1 Department of Pharmaceutical Sciences, University of Antwerp, Antwerp, Belgium (magda.claeys@ua.ac.be)2 Institute for Nuclear Sciences, Ghent University, Gent, Belgium3 Institute for Physics, University of São Paulo, São Paulo, Brazil

Organic aerosol is the major aerosol type in biomass burning (pyrogenic) and biogenic aerosols A major fraction of the organic aerosol is water-soluble and may affect the cloud-nucleating properties of

aerosol particles Characterisation of the water-soluble organic carbon (WSOC) is a complex task, but is of importance for

assessing the physico-chemical properties of the organic aerosol and for estimating the relative contributions from pyrogenic and biogenic sources

1. Introduction

Tropospheric aerosol research team from Ghent University (UGent) participated in the field work of the LBA/CLAIRE/SMOCC experiment in Amazonia, September-November 2002 [Maenhaut et al., 2003] several UGent aerosol collection devices set up at Fazenda Nossa Senhora (FNS), Rondônia

UGent high-volume filter sampler for organic analysis R1HiVo: Hi-Vol sampler, which provides separate fine (<2.5 µm aerodynamic diameter (AD)) and coarse

(>2.5 µm AD) size fractions– double Pallflex quartz fibre filters (of 102 mm diameter) used for each of the two size fractions– the filters had been pre-heated at 550C to remove organic contaminants

Aerosol samplings: from 9 September to 14 November 2002 (66 days)

Figure 3 shows the time trends for fine PM (PM2), and for organic matter(OM, estimated as 1.6*OC), EC, and levoglucosan in the fine filters of R1HiVo all 4 parameters show substantial variation throughout the campaign and are

fairly well correlated with each other several episodes with high levels, because of the impact from biomass

burning

Figure 4 shows levoglucosan carbon as a percentage of the OC data for daytime samples are in green, for night time samples in red interesting diel differences are observed, which may be related to the more

shallow boundary layer during night clear that levoglucosan contributes more to the OC during night (on average

3.31.0% during night versus 2.00.6% during the day)

Figure 5 shows the time trends for levoglucosan, malic acid, and2-methylerythritol malic acid and 2-methylerythritol exibit much less (and different) variability

than levoglucosan (an excellent indicator for biomass burning)– indicates that they originate from other sources

Median concentrations (and concentration ranges) for OC, EC, and the variouscompounds measured by GC/ion trap MS are given in Table 1 levoglucosan is the dominant organic compound

– its concentration (and that of OC) resembles that observed at the same site during the 1999 “burning” season [Graham et al., 2002; Zdráhal et al., 2002]

malic acid and 2-methylerythritol (indicators of photo-oxidation of natural volatile organic compounds) are characteristic of the natural background and are also important organic compounds

– the concentration of malic acid is similar to that reported for the 1999 “burning” season at the same site [Graham et al., 2002]

– the levels of malic acid and of 2-methylerythritol are of the same order as those found during the CLAIRE-2001 “wet” season campaign at Balbina (north of Manaus) [Claeys et al., 2003]

as could be seen from Figure 3, PM2, OC, EC, and levoglucosan were fairly well correlated with each other

– this was also apparent from the R2 values with OC, which were 0.97 for PM2, 0.92 for EC, and 0.86 for levoglucosan (with N = 74)

several organic compounds were well correlated with levoglucosan, i.e., mannosan (0.98), glucofuranose (0.91), and galactosan (0.93) [data in parentheses are R2 values], indicating that they are also associated with pyrogenic emissions

– also glucose somewhat correlated with levoglucosan (R2 of 0.61) in contrast, other compounds were not at all correlated with levoglucosan,

e.g., malic acid and 2-methylerythritol with R2 values of 0.08 and 0.07, resp. the enantiomers 2-methylthreitol and 2-methylerythritol were well correlated

with each other (R2 of 0.87)

Acknowledgements

This research was funded by the European Commission (contract no. EVK2‑CT2001‑00110), the Belgian Federal Office for Scientific, Technical and Cultural Affairs, and the Special Research Fund of Antwerp University. We are grateful to Xuguang Chi for assistance with the analyses for OC and EC, and to Sheila Dunphy for technical assistance.

References

Birch, M.E., R.A. Cary, Aerosol Sci. Technol. 25 : 221, 1996.Claeys, M., et al., submitted manuscript, 2003.Graham, B., O.L. Mayol-Bracero, P. Guyon, G.C. Roberts, S. Decesari, M.C.

Facchini, P. Artaxo, W. Maenhaut, P. Köll, M.O. Andreae, J. Geophys. Res. 107 (D20), 8047, doi:10.1029/2001JD000336, 2002.

Maenhaut, W., A. Onjia, X. Chi, J. Cafmeyer, V. Pashynska, R. Vermeylen, G. Vas, M. Claeys, Joint International Symposium on Atmospheric Chemistry, Program and Abstracts, Heraklion, Crete, Greece, p. 161, 2002.

Maenhaut, W., J. Cafmeyer, X. Chi, S. Dunphy, N. Raes, P. Artaxo, see poster board P0823 at this Assembly, 2003.

Pashynska, V., R. Vermeylen, G. Vas, W. Maenhaut, M. Claeys, J. Mass Spectrom. 37 : 1249, 2002.

Zdráhal, Z., J. Oliveira, R. Vermeylen, M. Claeys, W. Maenhaut, Environ. Sci. Technol. 36 : 747, 2002.

2. Aerosol collections

4. Time trends and median concentrations

5. Correlations between selected species

3. Chemical analyses

6. Mass fraction of OC explained by the organic compounds

Table 2 presents the mean percent carbon (and associated standard deviation) ofthe OC that is attributed to the various organic compounds measured in this study levoglucosan provides the major contribution (2.5%) and all compounds

together account for 4.1% these two percentages are quite similar to those found for “total” (PM10) filter

samples from SAFARI 2000 (where they were 2.5% and 4.9%, resp.) [Maenhaut et al., 2002]

in contrast, for arabitol, mannitol, glucose, and fructose, 5 to 10 times lower percentages are found than for the SAFARI 2000 “total” filter samples

– these compounds are presumably mainly associated with the coarse particles

Table 2Mean percentages (and associated standard deviations) of the fine (<2.5 µm AD) OC attributable to the carbon in the organic compounds (data derived from R1HiVo; N = 74).

Species Mean % ± std. dev.

Levoglucosan 2.5 ± 1.0 Mannosan 0.18 ± 0.07 Galactosan 0.07 ± 0.03 Glucofuranose 0.12 ± 0.05 Arabitol 0.04 ± 0.03 Mannitol 0.07 ± 0.06 Erythritol 0.02 ± 0.01 Glucose 0.12 ± 0.10 Fructose 0.03 ± 0.02 2-Methylthreitol 0.10 ± 0.07 2-Methylerythritol 0.32 ± 0.24 Malic acid 0.45 ± 0.32

Sum 4.1 ± 0.9

Table 1Median concentrations and concentration ranges in the fine (<2.5 µm AD) size fraction during the LBA/CLAIRE/SMOCC-2002 campaign (derived from R1HiVo; N = 74).Data for OC and EC are in µg.m–3; for all other species in ng.m–3.

Species Median conc. Conc. range

OC (µg.m–3) 15.0 1.59 – 61EC 1.23 0.181 – 3.8

Levoglucosan 770 86 – 6100Mannosan 54 6.0 – 440Galactosan 18.2 1.96 – 210Glucofuranose 34 3.3 – 360Arabitol 10.4 DL – 34Mannitol 16.1 3.2 – 43Erythritol 6.9 0.94 – 30Glucose 35 12.6 – 107Fructose 9.8 DL – 362-Methylthreitol 34 2.5 – 812-Methylerythritol 107 11.0 – 250Malic acid 153 23 – 570

Fig. 1

All filters from R1HiVo analysed for organic carbon (OC) and elemental carbon (EC) by a thermal-optical transmission (TOT) technique [Birch and Cary, 1996]

Fine front filters for 74 samples: 1/16 or 1/32 fraction of the front filter extracted with

methanol:dichloromethane mixture analysis for levoglucosan, related saccharidic compounds, and malic acid with a gas

chromatography/ion trap mass spectrometry method (GC/ion trap MS) adapted from Pashynska et al. [2002]

the analytical procedure is presented in Figure 1 total ion chromatogram obtained for sample 15D (collected on 23 Sept., daytime) is

shown in Figure 2– levoglucosan is the major organic compound detected, but it can be noted that

malic acid and 2-methylerythritol [Claeys et al., 2003] also give rise to prominent peaks in the chromatogram

Part of quartz filter(in glass flask)

Extraction with 3 x 20 mL CH2Cl2 :CH3OH(80:20; v/v) under ultrasonic agitation

Concentration to 1 mL by rotary evaporation at 213 hPa and 35ºC

Filtration on Teflonfilter (0.45 μm)

Silylation with MSTFA + 1% TMCS: pyridine (2:1; v/v) at 70ºC for 60 min

Addition of recovery standards(methyl-β-D-xylopyranoside)

(D3-malic acid)

GC/MS: full scan mode

Mass chromatography:

malic acid - m/z 307

D3-malic acid - m/z 310

Concentrations of PM2 and of Fine OC, EC and levoglucosan (R1HiVo fine filter)

0

20

40

60

80

100

120

140

160

1N (

9 S

ep. ) 2

3D 3N 4D 4N 5D 5N 6D 6N 7D 7N 8D 8N 9D 9N 10D

10N

11D

11N

12D

12N

13D

13N

14D

14N

15D

15N

16D

16N

17D

17N

18D

18N

19D

19N

20D

20N 21 22

23 (

1 O

ct.)

24D

24N

25D

25N

26D

26N

27D

27N

28D

28N

29D

29N 30 31 32 33 34

35D

35N 36 37 38 39 40 41

42D

43N

44D

45N

46D

47N

48D

49N

50D

(1

Nov

.)51

N52

D53

N 54 55

Con

c. (

µg/

m3)

PM2OC * 1.6EC (x 10)Levoglucosan (x 10)

R1HiVo fine filter (<2.5 µm AD): Levoglucosan C as percent of OC

0

1

2

3

4

5

6

1N2

3D 3N 4D 4N 5D 5N 6D 6N 7D 7N 8D 8N 9D 9N 10D

10N

11D

11N

12D

12N

13D

13N

14D

14N

15D

15N

16D

16N

17D

17N

18D

18N

19D

19N

20D

20N 21 22 23

24D

24N

25D

25N

26D

26N

27D

27N

28D

28N

29D

29N 30 31 32 33 34

35D

35N 36 37 38 39 40 41

42D

43N

44D

45N

46D

47N

48D

49N

50D

Per

cent

Fig. 2

Fig. 3

Fig. 4

R1HiVo Fine size fraction (<2.5 µm AD)

0

1

2

3

4

5

6

7

1N (

9 S

ep.) 2

3D 3N 4D 4N 5D 5N 6D 6N 7D 7N 8D 8N 9D 9N 10D

10N

11D

11N

12D

12N

13D

13N

14D

14N

15D

15N

16D

16N

17D

17N

18D

18N

19D

19N

20D

20N 21 22

23 (

1 O

ct.)

24D

24N

25D

25N

26D

26N

27D

27N

28D

28N

29D

29N 30 31 32 33 34

35D

35N 36 37 38 39 40 41

42D

43N

44D

45N

46D

47N

48D

49N

50D

(1

Nov

. )51

N52

D53

N 54 55

Con

c. (

µg/

m3)

LevoglucosanMalic acid2-Methylerythritol

Fig. 5