Source and reaction pathways of dicarboxylic acids, ketoacids and dicarbonyls in arctic aerosols: One year of observations

Pergamon

Atmospheric Encironment Vol. 30, Nos 10/11, pp. 1709-1722, 1996 Copyrighyt © 1996 Elsevier Science Ltd Printed in Great Britain. All rights reserved 1352-2310/96 $15.00 + 0.00

1352-2310(95) 00395-9

SOURCE AND REACTION PATHWAYS OF DICARBOXYLIC ACIDS, KETOACIDS AND DICARBONYLS IN ARCTIC AEROSOLS: ONE YEAR OF OBSERVATIONS* KIMITAKA KAWAMURA and HIDEKI KASUKABE Department of Chemistry, Faculty of Science, Tokyo Metropolitan University, 1-1 Minami-Ohsawa, Hachioji, Tokyo 192-03,Japan and L E O N A R D A. B A R R I E Atmospheric Environment Service, Environment Canada, Downsview,Ontario, Canada (First received 1 February 1995 and in final form 1 September 1995)

Abstract--Normal saturated (C2-C~t) and unsaturated (C4-C5, Ca) dicarboxylic acids were measured in arctic aerosol samples collected weekly at Alert, Canada in 1987-1988. In all seasons, oxalic (C2) acid was usually the dominant diacid species (1.8-70 ng m-3, av. 14 --!-_12 ng m-3) followed by malonic (C3; 0.0519 ngm -3, av. 2.5 + 3.3 ngm -3) and succinic (C4; 0.51-18 ngm -3, av. 3.8 + 3.5 ngm -3) acids. The total concentrations of dicarboxylicacids showed a seasonal variation (4.3-97 ng m- 3, av. 25 + 20 ng m- a), with two maxima in September to October and in March to April. The autumn peak is characterized by high concentrations of oxalic acid and azelaic(C9)acids, which were probably caused by enhanced contributions from anthropogenic and biogenic sources, respectively, followed by photochemical reactions. This is consistent with higher concentrations of n-alkanes from terrestrial plant waxes and of soil-derived aluminum in the autumn aerosol samples. On the other hand, during "Arctic Sunrise" in March to April, oxalic, malonic and succinicacids as well as some other (C5-C6) diacids were 5 to 20 times more abundant than in the preceding dark winter months, suggestingthat diacids are produced in situ by secondary photochemical oxidation of organic pollutants carried to the Arctic. ¢o-Oxocarboxylicacids (C2-C5, C9), pyruvic acid and ~-dicarbonyls (methylglyoxaland glyoxal) were also detected in the arctic aerosols. Their concentration also showed spring maxima; however, they were observed a few weeks earlier than the spring peak of diacids. The og-oxoacids are likely intermediates to the production of ~,¢o-dicarboxylicacids at the polar sunrise. Key word imtex: Water soluble organic compounds, diacids, ketoacids, dicarbonyls, arctic aerosols, oxalic acid, malonic acid, succinic acid.

INTRODUCTION Studies of low molecular weight dicarboxylic acids in urban and suburban aerosols demonstrated that the smallest diacid, oxalic acid, is the most abundant dicarboxylic acid in the range of C2-Clo (Kawamura and Ikushima, 1993; Sempere and Kawamura, 1994). The amounts of these diacids were found to account for up to 1% of total aerosol mass and 2% of the total carbon content of urban aerosols. Concentrations of aerosol dicarboxylic acids in the Tokyo atmosphere showed a summer maximum and a positive correlation with oxidants:. These results suggested that, for

*This paper was presented at the Fuji-Yoshida CACGP/IGAC meel:ing,Japan, September, 1994.

the most part, the small dicarboxylic acids are produced by photochemicaily induced reactions of anthropogenic organic pollutants in the urban atmosphere (Kawamura and Ikushima, 1993). Laboratory experiments have shown that some diacids such as adipic (C6) acid are produced by photochemical oxidation of cyclic alkenes (Grosjean et al., 1978; Grosjean and Friedlander, 1980; Hatakeyama et al., 1985, 1987), which are present in the atmosphere as a result of atmospheric emissions from internal combustion engines (Grosjean and Fung, 1984). Long-term observations of arctic aerosols (more than five years) have demonstrated the long-range atmospheric transport of pollutants such as sulfate from mid-latitudes to the Arctic maximized in winter to spring seasons (Barrie and Hoff, 1985; Barrie, 1986; Barrie and Barrie, 1990). Ozone destruction and photochemical reactions have been proposed to occur at

1709

K. KAWAMURA et al.

1710

polar sunrise in the lower arctic troposphere (Barrie et al., 1988). However, arctic aerosol chemical observations have been mostly focused on inorganic species while organic compounds were rarely studied. K a w a m u r a et al. (1995) briefly examined one year of particulate observations taken at Alert, Canada in 1987/88 for dicarboxylic acids and found a correlation with spring time particulate Br associated with tropospheric ozone depletion at polar sunrise (Barrie et al., 1988). In this paper, that Alert data set of dicarboxylic acids is discussed in more detail together with cooxocarboxylic acids and ~-dicarbonyls detected in the arctic aerosols with an emphasis on reaction pathway. Based on the seasonal variation of their molecular distribution, we explore the photochemical production of dicarboxylic acids and the long-range atmospheric transport of anthropogenic and biogenic organic compounds which are their precursors.

EXPERIMENTAL

The aerosol samples were collected on Whatmann 41 filters (20 x 25 cm) at Alert (82.5°N, 62.3°W) from July 1987 to June 1988 using a high-volume air sampler. The sampling period was generally 7 days (Barrie and Hoff, 1985). Over the year, weekly mean ambient temperature ranged from - 38 to 5°C. The sample filters collected during summer showed a white or very faint dark color whereas those collected in the rest of the year were light grey to black. The filters were blackest from December to February, suggesting an enhanced long-range transport of fossil fuel combustion products such as black carbon to the Arctic atmosphere from industrialized areas including Europe (Barrie and Bottenheim, 1991). Some autumn samples showed a yellowish dark color, suggesting a possible contribution from soil particles. This is consistent with a peak in aerosol A1 concentrations in September/October at Alert (Barrie and Barrie, 1990). Filter samples were analyzed by the method of Kawarnura and Ikushima (1993). Briefly, aliquots of filters (c. 12 or 25 cm 2) were extracted with pure water, which was prepared in an all glass apparatus by distillation of Milli-Q water after oxidizing organic impurities with potassium permanganate. The water extracts containing dicarboxylic acids were concentrated to c. 1 ml by rotary evaporation under vacuum, and then passed through a Pasteur pipet column packed with quartz wool to remove particles. The water soluble fraction was concentrated to nearly dryness and esterified with 14% BF3/n-butanol (0.25 ml) at 100°C for 30 min. The dicarboxylic acid dibutyl esters were extracted with n-hexane after adding pure water and a small amount (0.2 ml) of acetonitrile, and then dissolved in 50 pl n-hexane. The aldehyde group in ~o-oxocarboxylic acids and dicarbonyls were derivatized to dibutoxy acetal. The derivatives were separated in the same fraction of diacid esters by extraction with n-hexane (Kawamura, 1993). Dibutyl esters and other derivatives were determined with a Hewlett-Packard 5890 gas chromatograph (GC) equipped with a split/splitless injector, fused silica capillary column (HP-5, 0.3 mm ID x 25 m long × 0.52/tm film thickness) and an FID detector. The GC peak areas were calculated with Shimadzu C-R7A integrator. Identification of the compounds were performed by comparing GC retention times with those of authentic standards. The compounds were further analyzed by a Finnigan-MAT ITS 40 GC-mass spectrometer for mass spectral identification using authentic standards. Mass spectra (m/z 45-500) were obtained every 1 s

and processed with the ITS 40 systems with INCOS mass spectral data library. Recoveries of authentic standards spiked to a pre-combusted quartz fiber filter were 71% for oxalic acid and better than 90% for malonic, succinic and adipic acids (Kawamura and Ikushima, 1993). Duplicate analyses of some sample filters showed an error of less than 15% for major diacids. The gas chromatograms of blank filters showed peaks of phthalic, oxalic, succinic, and other acids. However, their concentration levels were generally less than 10% of those for sample filters, except for phthalic acid (up to c. 50 %). The concentrations reported in this paper are corrected for the blanks but not for recoveries. Actual recoveries on samples are expected to be higher than for spiked filters because of metal ions in aerosols.

RESULTS AND DISCUSSION

Molecular distributions and origin of dicarboxylic acids, ketocarboxylie acids and ~-dicarbonyls Figure 1 shows a typical gas chromatogram of dicarboxylic acid dibutyl esters and related polar compounds isolated from the arctic aerosol samples (4-11 April 1988). Three types of organic compounds were detected in the water soluble fraction of the aerosols: (1) an homologous series of dicarboxylic acids including aliphatic ~,co-dicarboxylic acids (C2-Cll) and aromatic (phthalic) acid, (2) ketocarboxylic acids including co-oxocarboxylic acids (C2-C9) and pyruvic acid, (3) c~-dicarbonyls (glyoxal and methylglyoxal) (see Fig. 2 for chemical structures). Table 1 summarizes the polar organic compounds measured, their mean concentrations, concentration ranges and the month when their maximum concentration was observed.

Homologous series of ~, co-dicarboxylic acids (C2-Cll). Throughout all the samples, oxalic acid (C2) was found as the most abundant diacid species. The second most abundant diacid was malonic (C3) or succinic (C4) acid. The C2-C4 diacids accounted for more than 80% of total diacids (4.3-97 ng m - 3 ) in the arctic aerosols studied. Longer chain normal diacids were less abundant and their concentrations generally decreased with increasing carbon number. The exception was C9 diacid which was more abundant than Cs species. Branched chain saturated dicarboxylic acids were also detected in the Alert aerosols, including methylmalonic (iC4), methylsuccinic (iCs) and 2methylglutaric (iC6) acids. They are less abundant than corresponding straight-chain dicarboxylic acids. In addition to saturated ~, co-dicarboxylic acids, unsaturated diacids were detected in the arctic aerosols, including maleic, fumaric and methylmaleic acids. The abundance of the unsaturated diacids are lower than those of corresponding saturated species, although maleic acid was relatively abundant in winter samples. Aromatic acid (phthalic) was detected as the fourth most abundant dicarboxylic acid. ~,co-Dicarboxylic acids with an additional functional group were also detected in the arctic aerosols. They include ketomalonic acid (kC3), 4-oxopimelic

Dicarboxylic acids in arctic aerosols

C~ C~

C4 eC~

C6

l

_

1711

Ph

C5 ul

31y

I a rr

.L

II 1'°'

M°°'Y/I

l

i I"iii

J

cel ! \.

!

C9

I Cr t0% ;1 i

lO

C~0 1

20

:30

RETENTION TIME (rain.)

Fig. 1. Capillary gas chromatogram of dicarboxylic acid dibutyl esters and related polar compounds isolated from the arctic aerosol sample ( 4-11 April 1988).

HOOC -- COO H (a) Oxalic acid (C2)

HOOC ~ C O O H

HOOC ~ COOH (b) Malonicacid (C3)

.CH3 HOOC . , . ~ ¢ / C O O H

(e) Succinicacid (C4)

HOOC~COOH

(d) Methylsuccinicacid (iCS)

H O o c ~ C O O H

(e) Glutaric acicl (C5)

(f) Adipic acid (C6)

COOH

HOOC ~

i•

C O O H (g) Azelaic acid (C9)

y

COOH

HOOC" (j) Fumarie acid (F)

I ~ COOH (h) Maleicacid (M)

N~COOH (i) Methylmaleic acid (mM) O CII

[~COOH v,~--,. COOH

HOOC /

(k) Phthalicacid (Ph)

OII HOOC ~ C

H3C,,,~ COOH

~COOH

(I) Ketornalonieacid (kC3) OH

~

COOH

HOOC ~ ' ~

(m) 4-Ketopimeticacid (kC7)

HOC ~ C O O H (p) 4-Oxobutanoic acid (wC4)

O II H3C- C--COOH (q) Pyruvic acid (Pyr)

COOH

(n) Malieacid (he4)

HOC--COH (r) Glyoxal (Gly)

HOC -COOH (o) Glyoxylic acid (we2) O II H3C--C--COH (s) Methylglyoxal (MeGly)

Fig. 2, Chemical structures of dicarboxylic acids (a-n), ketoacids (o-q) and ct-dicarbonyls (r-s) detected in the Alert aerosols. acid (kCT) and hydroxysuccinic (malic) acid (hC4). However, these tri-.functional compounds were found as relatively minor components compared to bifunctional C2--C4 diacids.

Low molecular weight dicarboxylic acids with oxalic acid being the predominant, have been reported in aerosols collected in urban Los Angeles (Kawamura and Kaplan, 1987) and in Tokyo

1712

K. KAWAMURA et al. Table 1. Dicarboxylic acids (ng m-3) in the Arctic aerosol samples collected from Alert in July 1987 June 1988

Compounds Dicarboxylic acids Normal saturated diacids Oxalic Malonic Succinic Glutaric Adipic Pimelic Suberic Azelaic Sebacic Undecanedioic Branched saturated diacids Methylmalonic Methylsuccinic 2-Methylglutaric Unsaturated diacids Maleic Fumaric Methylmaleic Phthalic

C,, Abbr.

Conc. range

1.8-70 0.048-19 0.51-18 0.16-4.3 < 0.003-9.5 0016-1.8 0.015-1.8 0.036-2.0 < 0.003-1.7 < 0.003-1.1

13.6 + 12.1 2.46 ± 3.3 3.73 + 3.55 0.90 + 0.91 0.82 + 1.43 0.13 + 0.26 0.15 _+ 0.26 0.26 ± 0.32 Not calc. 0.064 _+ 0.15

9 4 4 4 10 10 10 10 10 10

iC4 iC5 iC6

< 0.003-0.64 < 0.003-1.8 < 0.003-0.21

0.13 _+ 0.12 0.36 + 0.37 0.016 ___0.038

4 4 4

C4, M C4, F C5, mM Ca, Ph

0.016-0.61 0.017-1.1 < 0.003-0.22 < 0.005-5.8

0.19 _+ 0.14 0.14 ± 0.19 0.034 __+0.056 1.5 _+ 1.5

12 8 10 3

0.31 ___0.38 0.20 ___0.17

4

0.026 ___0.038

10

25 + 20

4

1.7 ___2.4 0.07 ___0.05 0.35 ± 0.24 0.02 __+0.02 0.01 ___0.02

3 4 5 4 10

kC3 kC 7

0.026-2.1 0.18-0.62

Hydroxylated diacid Malic

hC4

< 0.003-0.14

Total

4.3-97 rnC2 ~oC3 coC4 ~oC5 ~oC9

Total ct-Ketocarboxylic acid Pyruvic Dicarbonyls Glyoxal Methylglyoxal Total

Month (max conc.)

C2 C3 C4 C5 C6 C7 C8 C9 C10 C11

Ketodiacids Ketomalonic 4-Ketopimelic

~o-Oxocarboxylic acids Glyoxylic 3-Oxopropanoic 4-Oxobutanoic 5-Oxopentanoic 9-Oxononanoic

Mean conc. _+ st. dev.

< 0.003-10.1 0.012-0.20 < 0.003-0.94 < 0.003-0.10 < 0.003-0.11 0.18-10.8

2.2 ___2.5

C3, Pyr

< 0.003-0.52

0.13 ± 0.13

C2, Gly Ca, MeGly

< 0.003-2.29 < 0.003-0.59 < 0.003-2.66

0.54 _+ 0.56 0.16+0.14 0.71 ___0.67

(Kawamura and Ikushima, 1993; Sempere and Kawamura, 1994). The predominance of oxalic acid was also recognized in the molecular distribution of diacids in remote marine aerosols from the North Pacific (Kawamura and Usukura, 1993). Oxalic acid has been detected in Alaska aerosol samples by Li and Winchester (1993), using ion chromatography. Other diacids were not measured. Diacids with oxalic acid being relatively less abundant were reported in aerosols from Takasaki, central Japan using methyl ester derivatization followed by G C determination (Satsumabayashi et al., 1990), which results in a serious underestimation of oxalic and malonic acids (Kawamura and Ikushima, 1993).

4

The atmospheric chemical sources and sinks of dicarboxylic acids are not quantitatively well known. However, there is a growing body of knowledge that allows a qualitative picture to be painted, as illustrated in Fig. 3. Laboratory experiments have shown that they are produced in the ozone and cycloalkene reaction (Hatakeyama et al., 1987). In addition, evidence suggests that they are produced in the atmosphere as a result of secondary photochemical reaction of anthropogenic aromatic hydrocarbons such as benzene and toluene and their oxidation intermediates such as glyoxal and methylglyoxal (Norton et al., 1983; K a w a m u r a and Ikushima, 1993). The abundant presence of cis configuration (maleic acid

Dicarboxylic acids in arctic aerosols

"m--:' i I/

t

i J_ / l#/

t____,__.,

*

---

kkc3}

I

Lm_.aleic_I /

7~IT~

r"

i

I---c-I/

1713

7----'1 t I v-. /'7 / . i k,,.,or,o.rbo,,y.,,:; /

_

--

v .-"'

~Mid-chainketo~

"t& hydroxyacidsJ "

/

/

l

'

[

J / • V

C6, iC6 I Sea salt spray iSeawate r surface I

Fig. 3. Possible photochemical production of oxalic (C2), malonic (C3), succinic (C4) and azelaic (C4) acids in the atmosphere. For abbreviations and chemical structures, see Table 1 and Fig. 2.

and methylmaleic acid) in the urban atmosphere supports an oxidation of these aromatic hydrocarbons as a precursor of oxalic acid. In automobile exhausts, oxalic acid was predominant followed by succinic acid and, to a lesser extent, malonic acid (Kawamura and Kaplan, 1987). In contrast, malonic acid which is hardly produced by the oxidation of aromatic structures having conjugated double bonds, is likely produced by the photochemical oxidation of succinic acid with malic acid as an intermediate (Kawamura and Ikushima, 1993). In turn succinic acid and longer chain diacids are likely oxidation products of gaseous aliphatic monocarboxylic acids, which are in part produced by photo-induced oxidations of biogenic unsaturated fatty acids (Kawamura and Gagosian, 1987) and other precursors such as n-alkanes, aldehydes and midchain ketocarboxy]~ic acids (see Fig. 3). A one-year observational study of low molecular weight dicarboxylic acids in urban aerosols of Tokyo suggested that photochemica7 reactions which peak in summer result in production of LMW diacids and in summer contribute more to the particulate diacids than direct emissions from anthropogenic and natural sources (Kawamura and Ikushima, 1993). og-Oxocarboxylic acids and ketoacid. The homologous series of aldehydic acids were detected in the range of C2-Clo in the arctic aerosols. Although the presence of Cr--Ca and C1 o o9-oxoacids was confirmed by mass chromatography techniques using characteristic mass fragment ions (e.g. m/z M-73, Kawamura, 1993), their concentrations were not measured due to the small and, sometimes, non-detectable peaks on the GC-FID chromatograms, thus not reported in Table 1. Interestingly, g3tyoxylic acid (eJC2), the smallest

~o-oxoacid, was generally the most abundant species in winter to spring seasons. In contrast, 09oxobutanoic acid (o9C4) overwhelmed the oJC2 in the summer aerosols, although the concentrations were lower than in other seasons. Total concentrations of ~-oxoacids (09C2 to ogCs and ~C9) ranged from 0.18 to 10.8 ngm -3 and largely fluctuated with seasons. Their concentrations (av. 2.2 ng m - 3) were lower than those of ct,~o-dicarboxylic acids (av. 2 5 n g m -3) throughout four seasons. The ketoacid, pyruvic acid, was also detected in the aerosols. However, concentrations (av. 0.13ngm -3) were much lower than glyoxylic acid (~oC2, av. 1.7 ng m-3). The predominance of glyoxylic acid in the oJ-oxocarboxylic acids has been reported in the urban atmosphere (Kawamura, 1993). Glyoxylic acid may be produced by atmospheric oxidation of glyoxal (Kawamura, 1993). Pyruvic acid was also reported in urban aerosols and rainwaters (Kawamura, 1993; Sempere and Kawamura, 1994), remote continental rain and aerosols from Amazon (Talbot et al., 1990) and remote marine atmosphere (Kawamura and Usukura, 1993; Zhou and Mopper, 1990). These ketoacids may be intermediate in the photochemical oxidation of anthropogenic and natural organic compounds to the production of oxalic acid. ct-Dicarbonyls. Glyoxal and methylglyoxal were detected in Alert aerosols. Glyoxal atmospheric concentrations were highest in the winter-spring. The\ dicarbonyls were detected at much lower concentrations < 0.003 to 2.66 ngm -3 than the dicarboxylic acids (4.3-97 ng m - 3) and ~-oxocarboxylic acids (0.18-10.8 ng m-3). This is likely because dicarbonyls are largely present in the gas phase.

1714

K. KAWAMURA et al.

The 7-dicarbonyls have been reported in the urban (Kawamura, 1993) and marine atmosphere (Zhou and Mopper, 1990). Laboratory studies show that glyoxal and methylglyoxal are produced as photochemical oxidation products of aromatic hydrocarbons such as benzene and toluene (Bandow and Washida, 1985; Bandow et al., 1985).

IOO

Ill

(a) Total Diacids 8o6040. 20-

Seasonal variations of polar compounds: long-range transport and photochemical production One year of observations of arctic aerosols showed strong variations in the concentrations. Figure 4 presents overall seasonal variations in the total concentrations of dicarboxylic acids, ~o-oxocarboxylic acids, and ~-dicarbonyls in the arctic aerosol samples collected in 1987-1988. Total dicarboxylic acids showed two major peaks in autumn and spring. In contrast, concentrations of eJ-oxoacids and ~-dicarbonyls showed a maximum in the spring and no significant peaks in autumn. However, individual species fluctuated in a different manner, as follows. Saturated diacids (C2-C9). Figure 5 gives seasonal variations of individual ~,og-dicarboxylic acids in the Alert aerosols. The most abundant diacid (oxalic acid) showed two major peaks in September to October and March to April. The highest concentration (ca. 70 ngm-3) was obtained in the aerosol sample collected on 14-21 September 1987. In late October to January, the concentrations stayed at rather low levels with a relatively small fluctuation (2-20 ng m-3). Its concentration started to increase significantly in March and peaked in early April (Fig. 5a). The C2 concentration at the April peak was ca. 3 times more abundant than that of the preceding dark winter months. The April peak of oxalic acid appeared at the time of arctic sunrise, suggesting that oxalic acid is significantly produced in the arctic atmosphere by photo-induced reactions. During arctic sunrise, much chemical activity involving CI and Br atoms reacting with hydrocarbons takes place (Barrie et al., 1994b; Jobson et al., 1994), resulting in oxidants such as ozone, hydrogen peroxide and hydroxyl radicals. The oxidants may react with various organic compounds which are transported and accumulated in the Arctic during dark winter to produce oxalic acid. After the spring peak, oxalic acid concentration decreased in the summer to a level similar to that observed in the winter months. Lower concentrations in the summer are due to scavenging of the particulate diacids by more frequently occurring precipitation, coupled with a weakened transport of polluted air parcels from mid-latitudes (Barrie, 1986). Similar seasonal variations were observed for C3-C~ c~,og-dicarboxylic acids (Figs 5b-d), except that autumn peaks are relatively weaker and amplitudes of the spring peaks are larger (10 to 20 times) than that of oxalic acid (3 times). Their concentrations started to increase at the beginning of March, peaked in April and returned to the background levels in late June. These results again suggest in situ photochemical

i i i I i I I I I I I I I n n

o

.

12

.

.

.

I,,

,I

. . . . . .

i,

I,,

i i . l , , t l n . a l J J J l t J l l

•

,,

(b) Total w - O x o a c i d s i

i

i

,

8.

~

6

0

3

ii

I l i ] l l l l l l l l l l l l l l

L)

ii

( 2

1

0 6

8

10

12

2

4

6

8

Month (1987-1988, Alert) Fig. 4. Seasonal variations of total concentrations of (a) dicarboxylic acids (C2-C10, (b) ~o-dicarboxylicacids (C2-C5, C9), and (c) c~-dicarbonyls (Cz-C~) in the Alert aerosols, 1987-1988. production of these diacids in the arctic atmosphere. However, the peak shapes are slightly different from each other. Thus C3 and C5 diacids showed maxima in the 4-11 April sample whereas C4 diacid peaked three weeks later. Such a variation may be associated with different source inputs and/or different production mechanisms. The C6 diacid showed a seasonal distribution similar to that of C3-C5 diacids, however, the amplitude of the April peak was much greater than that of C3-C5 acids (Fig. 5e). On the other hand, longer chain diacids (C7-Cll) showed seasonal variations different from those of shorter chain diacids (C2-C6), i.e. they showed a large peak in early October, and no significant peak in early spring. Except for the autumn peak, their concentrations stayed low in dark winter months and become only slightly higher in summer, although higher concentrations were observed in early April (Fig. 5). It is important to note that, in the early October samples, C7, C8, C9, Cio and Cll diacids were found as

Dicarboxylic acids in arctic aerosols I t al

70

nl

,,~1

,1,

nil

la

le[,,,l,,,

i,

Sq

60 i (a) C2

1715 ,,

, I , , ,

(e) C6

, , ,

I , , ,

=1=

l l l l

!

50: 4O~ 3O~

=

i

ZO~

10: Oi 20-

_L_Illl

I l i r l

a*,

n

~a,

III

!

,,

l , l l l l

,=

P

to.

e 5.

o.6-~

i

0.4-1

i

0.2-t

= l l

tl

iJ

,

1.8i

o.8-1 (f) C7

(b) C3 i

15.

II

~

W~

@

o.

o!. ,,;-.,

20-

~

i

i

I

i

i

i

I

I

I

i l l

I

1

I

,i

I

l~,,,t,,,

*l

(c) C4

O.8q

15. = @

, ,

(g) c s

o.6-~

i

0.44

i

, 1 , , , I

i,

I , = , t = t ,

1,si

r

i02

0.2-1 o 5

01..,i.

• " I ' ' ....1.111111

i

i

n

I

I

,i,

l

~

l * , l J l

(d) C5

4 3

~ !

I i

,.

t i I i i i

i

I i i 1 1

i1

l l , l

_ (h) C 9

].5-1

i

]

1-1

2

o.5-1

Ol

o 6

8

10

12

2

6

8

6

8

10

12

2

4

6

8

Month (1987-1988, Alert) Fig. 5. Seasonal distribution of individual normal saturated ~,~o-dicarboxylicacid in the arctic aerosol

samples collected from Alert, 1987-1988. For abbreviations, see Table 1 and Fig. 3.

abundant as C3 to C6 diacids (Fig. 5). Furthermore, longer chain dicarboxylic acids (C 12 to C 16) were also detected in the ear],y October sample (5-11 October 1987; for the gas chromatogram, see Fig. 6). The characteristic molecular distribution in the late September to October samples is very different from the rest of the samples, suggesting a different source• Because long chain normal alkanes with odd carbon n u m b e r

predominance

(C27

,

C29

,

C31

,

C33

)

were

abundantly detected in the 5-11 October sample

(Kawamura et al., unpublished results), the aerosols are heavily influenced by terrestrial organic matter associated with higher plant waxes and soil dust particles• Aluminum concentration in the October aerosols was significantly higher (1500 ng m-3) compared to the background level of c. 100 ngm -3 observed in winter (Barrie et al., 1994a). This is likely a sample affected by local wind blown dust, The seasonal variation of branched chain dicarboxylic acids (iso C4-C6) seems to be similar to those

1716

K. KAWAMURA et al. r~

C3 C4

Ph

I

G~,

C6

I C5

I

C8 (7

0C4

I MeGI~~c,

C~0

i ''c''cls i3

M~Fh U~'~

lO

C9

"

20

3O

RETENTITIOMNE(rain.) Fig. 6. Capillary gas cbromatogram of dicarboxylic acid dibutyl esters and related polar compounds isolated from the arctic aerosol sample (5-12 October 1987).

of straight chain C3-C5 diacids (Figs 7a--c). Their concentrations were low in the dark winter, started to increase in February and peaked in April at concentrations 5-10 times greater than those of the preceding dark winter months. The amplitude in the April peak is a factor of c. 5-10 higher than background, being similar to that of the corresponding straight chain diacids. However, peak shapes are different. The iso C4 and C5 diacids showed a maximum in late April similar to that of succinic acid. In contrast, iso C6 diacid showed a maximum three weeks earlier than the iso C4 and C5 peaks. The iso C6 diacid may be in situ produced by atmospheric oxidation of 2methyl cyclohexene which is of anthropogenic origin whereas other branched diacids may be derived from different precursors. The difference in the peak timing suggests that the molecular distribution of precursor cyclic alkenes has shifted in the arctic atmosphere probably due to changes in the origin of air mass. Unsaturated diacids. Maleic (cis configuration) acid showed a unique seasonal distribution pattern with a maximum in winter (Fig. 7e), being different from those of saturated diacids. Its concentration began to increase between summer and autumn and peaked in December. During dark winter, it remained high and then began to steadily decrease in February to a summer minimum. Such a distribution suggests that maleic acid is not produced in situ in the arctic atmosphere, but rather, is transported from mid-latitudes where it is produced by the atmospheric photochemical oxidation of aromatic hydrocarbons. High concentration of maleic acid has been measured in the urban atmosphere of Tokyo (Kawamura and Ikushima, 1993). In contrast, fumaric acid, the trans positional isomer of maleic acid, showed a distribution pattern different from the cis configuration (Fig. 7f). It was highest in summer and lowest in the winter

season. In addition a secondary spring maximum was apparent. This acid may be produced in situ by photochemically induced isomerization of maleic acid in the Arctic during arctic sunrise and afterwards. The summer peak of fumaric acid suggests the presence of a non-anthropogenic precursor. Phenolic compounds may be a likely candidate, which have been reported to originate from macroalgae and enriched in the sea surface slicks (Carlson, 1982). Keto dicarboxylic acids. Oxomalonic acid, which was detected as the predominant ketodicarboxylic acid in the Alert aerosol samples, showed a concentration maximum in the early spring (Fig. 7g). Its seasonal variation is almost the same as that of malonic acid (Fig. 5b), but its concentration is much lower. Their carbon skeletons are the same (see Figs 2b and 1). This parallel seasonal distribution and chemical consideration suggest that ketomalonic acid is likely produced by the atmospheric oxidation of malonic acid. 4-Oxopimelic acid also gave the April maximum; however, its concentrations were also high during the summer season (Fig. 7h). A parallel distribution was not recognized between the seasonal variations of 4-oxopimelic acid (kC7) and pimelic acid

(c~). Oxocarboxylic acids. ¢0C2 which is the smallest but most abundant ketoacid species in the arctic aerosols, showed a gradual increase in late winter season and maximum concentration in late March (Fig. 8a). After the March maximum, its concentration decreased steadily through summer to an autumn/winter minimum. The March peak is more than 50 times higher than the background concentrations in the summer and early winter months. This seasonal pattern was similar to that of C3 and C4 diacids (Figs 5b, c); however, the concentration of ¢oC2 started to increase two months earlier (December) than those of C3 and

Dicarboxylic acids in arctic aerosols i|l..i

0.7

=,,

-

0.6,

,=l

lltlJll

llll

!

0.7

!

(a) iC4

0.5

!

0.4 -:

i

i

7

i

1717 ll[

i,,,1,,,

IIIII

(e) M

0.6.

i

0.3.

0.2

0.2"

,

~

~

i

0

~ = ..~ ......

2-

I

(b) iC5

1.5

I

I

'

= '

1. t

I

l

i i

0

•

,

r ,

I

t

I

~

,

,

I

I

u

.

.

.

.

i

i

i

1

i

1

.

u'''

.

I

I

I

¢

l i t

I

I

I

g

0

I

I

I

I

I

' •

25

I

I

I

i . . . . . . I

I

'

'

'

i

t

I

!

I

I

I

I

i .......

t

I

I

•

•

•

i

•

,

a

i

l

i.

I

I

•

•

i

* = ,

I

I

I

.

•

,,i

,,

I

I

I

i

.

.

I

,

i

I

I

I

(g) kC3

2,

i

i

i

1 i

0.1

i

=

:

x

i

,

l. 0~"

0.05 0

5

I

i

O.lS

6-

I

0.2.

i l

(c) IC6

I

! 1

i ...... I

I

i...~..

0.4,

.m

~t

I

. . . . .

i[

0.6-

0.5

@

Z

(f) F

1.

i

' I

0.8-

0.2

II

OJ. tt~ i

02

@

*

i

0.4.

i

0.3 -:

,I

I

0.5.

0,],

e~

,I

,p-. i

i

I

n

I

n

I

J

i

i

I

(d) Ph

:

; . . . . . n

r

r

I

,

I

I

n

;_L_

0

,,,I

,

0.7

i

,

,

n

n

i

•

•

•

m

n

,

i

,

,

.

•

I

I

i

(h) kC7

0.6.

i

i

0.5.

4

7

0.4.

3

03.

2

0.2'

I

OJ.

0

0 8

10

12

2

4

6

8

i .

6

8

.

10

.

12

.

2

4

6

Month (1987-1988, Alert) Fig. 7. Seasonal distribution of individual branched-chain saturated and unsaturated ct,oJ-dicarboxylicacid as well as ct,o>dicarboxylicacid with keto functional group detected in the arctic aerosol samples collected from Alert, 1987-1988. For abbreviations, see Table 1 and Fig. 3.

C4 diacids did (February). Furthermore, ~oC2 peaked one or two weeks earlier (Fig. 8a) than the Ca-C5 diacids (Figs 5b-d) and Ca ketodiacid did (Fig. 7g). This suggests that glyoxylic acid is produced in winter under weak solar radiation probably during the southto-north transport of polluted air mass which originate in mid-latitudes and then transported and accumulated in the dark Arctic. As described earlier, glyoxylic acid may be extensively produced at polar

sunrise by in situ photo-oxidation of accumulated organic pollutants and then serve as a precursor of oxalic acid, which peaked one or two weeks later. On the other hand, concentrations of c0C3 which were much lower than those of coC2 and peaked in early April, although the concentrations largely scattered (Fig. 8b). The time lag between peaks of these two acids suggests that the ¢0C3 is produced from the precursor differently than eJC2. The seasonal

1718

K. KAWAMURA et al. 12

l l l l l , l

I l l

III

i i i I i i i

<

10

II

' .

+

-

6-

,

,11

i l l

Ii

ii

Jl

0.08

0.04 -

(.._,

i

2-

i

0 i

0.2

=

i I | l l

+

ii

l l l l

(b) wC3

0.15

iI

L : : R I . I

i

i

0.1. 0.05

l

0

:

. . . . . . i

1

l l l l l =

*l

i l l

I I I

0.02 -

*i

0

'

0.6 ~

t

(f:

''

ol,.

I E I I

,

2.5~

II

i i i 1 ,

Pyr

=i

'

'

*

|

I

I

I

i

i

Ug

i I

ii

-~

0.2 -] 1 [

}

i , I i i i

0.~t0.3-] 0.4

,

~, . . . . . . .

0

!...

............ ,~,

!

I

!

i

(c) we4

0.8•

@

, 1 + ,

0.06 -

4-

i m

i i 1 , 1

O.t-

8-'

r~ @

i

0.12

0.6

0A 0.5 ]

0.2 0 0.12

•

,

,

i

,

,-

i

i

a

I

i

i

0.1

I

l

,

+,

I

i

i

i (d) wC5 j

i i

,

,

,

I

i

i

0

i i ,

i

,

~

I

0.6 ! , , , I ,

I I

'1

.

.

.

.

.

,,I,,

o.s (h) MeGly

0.08

.

.

.

.

.

.

.

.

.

.

i

.....

,I

,,,

i

0.4

i

0.06

0.3

i

i

0.04

0.2

0.02

0.1

0

I

6

,

•

,"I

8

. . . . .

" + '

10

I

'

12

•

'

.

2

.

.

.

4

.

0,

I

6

8

6

L

,,.i,,,i..,i 8

I0

12

2

4

6

8

Month (1987-1988, Alert) Fig. 8. Seasonal distribution of individual oJ-oxocarboxylic acids, pyruvic acid and c<-dicarbonylsin the arctic aerosol samples collected from Alert, 1987-1988• For abbreviations, see Table 1 and Fig. 3.

distribution of coC+ showed lower concentrations in winter and higher values in spring and summer months (Fig. 8c). o9C5 was also detected in the aerosol samples; however, its concentration levels were roughly one-tenth of the concentration of o~C+. The seasonal distribution of the ~oCs species showed two peaks in October and April (Fig. 8d). coC9 has been proposed as the photochemically-induced oxidation product of biogenic unsaturated fatty acids, which contain a double bond predominantly at C-9 position

(Kawamura and Gagosian, 1987). This compound was detected in the samples as a relatively minor compound in the co-oxoacid series• The seasonal variation of pyruvic acid concentration peaked in February to April, although higher concentrations were sporadically observed in October, November, January, May (see Fig. 8f). ~-Dicarbonyls. Glyoxal and methylglyoxal showed a seasonal variation similar to that of coC2 (Figs 8g and h), although concentration levels of these

Diearboxylic acids in arctic aerosols .

.

.

.

.

.

.

.

.

.

.

.

.

.

6 / |

i i i t l

1719

i

t

t

i

|~1

l i t

!

i i i

II

!

4

S-~ (d)C9

!

i

f

2-1 ."

1

o-t- .......

10

II

i

i

i...i...i...i..

2o~'15(b)'' 'Ca ..... i ......

'i

t

i i i

i i I i t l

~M

8-

i

l i t

ii

il

6:

<

t l l I

!

4~

i

i

2-"

*llll

m

. ~ l l

I

. I

.

I

I

. I

I

. 1

. I

I

. I

(c) C4

i

I

I

I

.

.

I

I

.

.

I

. I

.

.

I

~

0

.

I

I

,

•

30

i

I

5J

I

4q

i

1

l l l * l l

ii

i

kC7

(f)

it

i

t l l l l

t i t

i

tl

3-4 20'

24 10. O'

lq -

6

. .

.

8

.

.

I

10

.

.

.

.

12

.

.

.

2

.

.

.

.

4

.

.

.

O I .

.

6

8

6

•

i...~...~......,

8

lO

12

i...;...

2

4

6

8

Month (1987-1988) Fig, 9. Seasonal changes in the relative abundance (%) of selected dicarboxylic acid in the total diacid concentrations. For abbreviations, see Table 1 and Fig. 3.

dicarbonyls were c. 5 times lower that of t~C2. Their concentrations started to increase in December and peaked in late March or early April. The amplitudes of the peaks are roughly 5-10 times greater than the background concenlration levels in the summer. Relative abundance of dicarboxylic acids and og-oxoacids: preferential production

Spring peaks of the major dicarboxylic acids and related polar compounds suggested photochemically induced reactions produce small diacids, which are end products and in'cermediates of the chain reactions occurring on the organic molecules transported to the Arctic. An increase i:a the diacid at the polar sunrise is consistent with the spring maximum of sulfate, which is in significant part produced in the arctic atmosphere by the photochemical oxidation of sulfur dioxide (Barrie and Banie, 1990; Barrie et al., 1994a). An in situ production :is supported by the presence of a strong linear correlation between shorter chain

diacids (C2-C5) and particulate Br in the arctic tropospheric aerosols at Alert (February-June, 1988) (Kawamura et al., 1995). Particulate Br is produced during the depletion of ozone in the Arctic boundary layer (Barrie et al., 1988, 1994). Bromine may be involved with the initiation of oxidative chain reactions on the organic molecules under strong solar radiation. In contrast, enhanced transport of small diacids in spring is not likely because of an absence of a strong peak of primary anthropogenic particulate matter such as vanadium (Barrie et al., 1994a) or lead (Barrie and Barrie, 1990). During in situ production of the diacids in the arctic troposphere, molecular distributions are likely shifted in time probably by preferential production of certain diacids. Figure 9 presents seasonal changes in the relative abundance of individual diacids (C2-C4, C9, M, kCT) in the total diacid mixture in arctic aerosols. C2 showed higher relative abundance in dark winter (Fig. 9a). This is consistent with the higher

1720

K. KAWAMURA et al.

percentages of maleic acid (M) (Fig. 9e), which is a photo-oxidation product of aromatic hydrocarbons and transported from mid-latitudes to the Arctic• The relative abundance of oxalic acid began to decrease towards the spring season as a result of preferential production and enrichment of other diacids. Especially, production of malonic acid (C3) is remarkable, that is, its relative abundance in the total diacids jumped from less than 10% to c. 20% at the polar sunrise (Fig. 9b). Selective enrichment of C3 diacid has been reported in the mid-latitudinal urban atmosphere during the seasons of summer to autumn as a result of enhanced photochemical oxidation of longer chain compounds such as succinic acid (Kawamura and Ikushima, 1993)• After polar sunrise, the relative abundance of malonic acid significantly decreased towards the summer season• In contrast to this, succinic acid (C4) peaked in late May (Fig. 9c), probably due to enhanced atmospheric transport of soil particles from China and other arid areas (Barrie and Barrie, 1990; Welch et al., 1991; Franzen et al., 1994). Except for a few summer samples, the relative abundance of succinic acid seems to be higher in summer season than winter. This may suggest that the precursor of succinic acid supplied in summer season originate from biogenic sources rather than anthropogenic sources. Succinic acid could be produced by photochemical oxidation of > C4 carboxylic acids, including o~C4 (Fig. 3) and carboxylic acids containing an additional keto group at C-4 position, which were not measured in this study. The former acid is detected in the arctic aerosol samples and its concentration increased from winter to summer seasons (Fig. 8c). Further, its relative abundance in the total oJ-oxoacids also showed an increase in summer season (Fig. 10c), being in contrast to C2 o~-oxoacid (Fig. 10a). On the other hand, 4-oxocarboxylic acids have been detected among the mid-chain aliphatic ketocarboxylic acids (C7-C13) in the remote marine aerosols (Kawamura and Gagosian, 1988) and in the present arctic aerosols (Kawamura et al., unpublished results). These ketoacids could serve as precursors of succinic acid. Figure 11 presents behaviors of M, o~C2 and C2 in winter to summer seasons, whose concentrations are normalized by the highest one and plotted by 5 point running means. M showed a maximum in December and decreased toward spring and summer. In contrast, ogC2 and C2 showed maximum in late March, however, the former acid seemed to peak slightly earlier than the latter. Such a shift of the concentration peaks from M (intermediate) in winter to C2 (end product) in spring suggests that chain reactions occur in the anthropogenic organic compounds transported to the Arctic predominantly in winter. Interestingly, the relative abundance of azelaic acid (C9 diacid) largely increased after May and showed higher values in summer to autumn (Fig. 9d). Because azelaic acid is a photo-oxidation product of biogenic unsaturated fatty acids emitted to the atmosphere on

10o

wc2

i

J

i

(,,)

8o:

i

~o! 40 20 0

~"

'

"

. . . . . .

5o

,,

'

'

'

i

. . . . . .

. l , , l l , . ,

I

,

,,

"

"

a,

(b)wC31 i 20 :

i

,

lO:

o

[

i

100

,,

I i . i

,1

i

(C)wC4

80

i i 1 1

| 1 1

l , , l l l l

i {

i

{

60

40 20 1

•

6

.

.

ii

8

.

.

10

12

2

4

6

8

M o n t h (1987-1988) Fig. 10. Seasonal changes in the relative abundance (%) of selected og-oxocarboxylicacid in the total oJ-oxoacid concentrations. For abbreviations, see Table 1 and Fig. 3.

1 q_____a_~__ I I .,%~-~ . q

L~ ~' ~

*

C2dl I

I - - -o---M 1

I

0.6 ..4

0.~

t.~

It

12

2

Month

4

5

6

7

Fig. l 1. Winter to summer changes in the concentrations of maleic acid, glyoxylic acid and oxalic acid relative to the highest concentration obtained for the arctic aerosols• Points are plotted by 5 point running means.

Dicarboxylic acids in arctic aerosols sea-salt particles (Kawamura and Gagosian, 1987), this result probably indicates that sea-to-air emission of phytoplankton derived organic matter is enhanced in the Arctic. However, the C9 diacid concentrations did not show a positive correlation either with methanesulphonic acid, l, Br, Na or A1 in the aerosols collected from Apri!: to June. The C9 concentration may be primarily controlled by atmospheric oxidation of precursor unsaturated fatty acids, whose sea-to-air emission should be enhanced in summer and autumn during open water and high biological activity conditions. A similar variation was observed for C~, C8 and C11 diacids as well as 4-oxopimelic acid (Fig. 9f). Enhanced emission of marine derived org~tnic matter to the Arctic may be caused as a result of warmer climate in spring to summer, which may cause a retreat of sea ice and activate sea/air exclhange of seawater lipid components, which are enriched in a microlayer of seawater surface (Marty et al. 1979). Enhanced relative abundance of natural organic matter during summer is also coupled with weakened transport of anthropogenic organic matter from mid-latitudes at that time of year (Barrie, 1986).

SUMMAI~:YAND CONCLUSIONS To understand the molecular distributions and seasonal changes of water soluble organic compounds including dicarboxylic acids, ketocarboxylic acids and dicarbonyls, a capillary gas chromatographic and mass spectroscopic analysis of 47 arctic aerosoll samples collected weekly from Alert, Canada was conducted. Major results and conclusions are as follows. 1. Oxalic (C2) acid was the most abundant diacid species followed by malonic (C3) and succinic (C4) acids. These small diacids comprised more than 80% of the total diacid concentrations (4-97 ng m 3). 2. During polar sunrise (late March to early April), concentrations of major diacids (C2-C5) became 3-20 times higher than those in the preceding winter months. The large increase in diacid concentrations was explained as a result of enhanced photochemical oxidation of organi= pollutants such as unsaturated hydrocarbons, which were transported and accumulated in the Arctic atmosphere during dark winter. ~o-Oxocarboxylic ac,ids and c~-dicarbonyls which are a likely intermediate to the production of dicarboxylic acids also showed a concentration maximum about two weeks earlier than the peak of diacids. 3. Selective production of malonic acid was observed during polar sunrise: its relative abundance in the total diacids jumped from c. 7-8% in the winter season to 18% just after the polar sunrise. Malonic acid was suggested to be produced by the photochemical oxidation of succinic acid through intermediates such as hydroxysuccinic (malic) acid (Fig. 3).

AE 30:IO/II-H

1721

Malonic acid may be further oxidized to ketomalonic acid and oxalic acid (Fig. 3). Such a chain reaction may also control the molecular distribution of dicarboxylic acids. 4. After the polar sunrise, concentrations of low molecular weight diacids decreased toward summer. However, azelaic acid (C9), a unique photochemical oxidation product of biogenic unsaturated fatty acids containing a double bond predominantly at the C-9 position, showed a peak in summer likely due to enhanced sea-to-air emission of marine derived organic matter to the Arctic due to the warming and retreat of sea ice in the polar ocean. 5. In September to October, concentrations of azelaic acid in the Arctic aerosols largely increased, probably due to local soil peaking in this time (Barrie and Barrie, 1990). Finally, this study demonstrates that the Arctic is an excellent natural laboratory for the study of photochemical transformation of anthropogenic and natural organic matter in the troposphere, and is useful for understanding the complex chemistry of dicarboxylic acids and related organic compounds. More laboratory studies of the reaction mechanisms of these compounds are needed. Acknowledgements--This work was in part supported by the

Ministry of Education, Scienceand Culture through Grantin-Aid No. 06453009 and by the Air Quality Research Branch of the Atmospheric Environment Serviceof Canada.

REFERENCES

Bandow H. and Washida N. (1985) Ring-cleavagereactions of aromatic hydrocarbons studied by FTIR spectroscopy II. Photooxidation of o-, m-, and p-xylenesin the NO:air system. Bull. Chem. Soc. Jpn. 58, 2541-2548. Bandow H., Washida N. and Akimoto H. (1985) Ring-cleavage reactions of aromatic hydrocarbons studied by FT-IR spectroscopy. I. Photooxidation of toluene and benzene in the NOx-air system. Bull. Chem. Soc. Jpn. 58, 2531-2540. Barrie L. A. (1986) Arctic air pollution: an overview of current knowledge. Atmospheric Environment 20, 643-663. Barrie, L. A. and Barrie M. J. (1990)Chemicalcomponents of lower tropospheric aerosols in the high Arctic:Six years of observations. J. atmos. Chem. 11, 211 226. Barrie L. A. and HoffR. M. (1985) Five years of air chemistry observations in the Canadian Arctic. Atmospheric Environment 19, 1995-2010. Barrie L. A., Bottenheim J. W., Schnell R. C., Crutzen P. J. and RasmussenR. A. (1988)Ozone destruction and photochemical reactions at polar sunrise in the lower Arctic atmosphere. Nature 334, 138-141. Barrie L. A., Li S. M., Toom D. L., Landsberger S. and Wu D. (1994a) Lower tropospheric measurements of halogens, nitrates and sulphur oxides during Polar Sunrise Experiment 1992. J. yeophys. Res. 99, 25,453-25,468. Barrie L. A., Bottenheim J. W. and Hart W. R. (1994b) Polar Sunrise Experiment 1992 (PSE92): Preface. J. oeophys. Res. 99D, 25,313-25,314. Carlson D. J. (1982) Surface microlayer phenolic enrichments indicate sea surface slicks. Nature 296, 426-429.

1722

K. KAWAMURA et al.

Franzen L. G., Hjelmroos M., Kallberg P., BrorstromLunden E., Juntto S. and Savolainen A. L. (1994) The 'Yellow Snow' episode of northern Fennoscandia, March 1991--a case study of long-distance transport of soil, pollen and stable organic compounds. Atmospheric Environment 28, 3587-3604. Grosjean D. and Friedlander S. K. (1980) Formation of organic aerosols from cyclic olefins and diolefins. Adv. Envir. Sci. Technol. 9, 435-473. Grosjean D. and Fung K. (1984) Hydrocarbons and carbonyls in Los Angeles air. J. Air. Pollut. Control. Ass 34, 537-543. Grosjean D., Van Cauwenberghe K., Schmid J. P., Kelley P. E. and Pitts J. N. Jr. (1978) Identification of C3-C10 aliphatic dicarboxylic acids in airborne particulate matter. Envir. Sci. Technol. 12, 313-317. Hatakeyama S., Tanonaka T., Weng J., Bandow H., Takagi H. and Akimoto H. (1985) Ozone-cyclohexene reaction in air: quantitative analysis of particulate products and the reaction mechanism. Envir. Sci. Technol. 19, 935-942. Hatakeyama S., Ohno M., Weng J., Takagi H. and Akimoto H. (1987) Mechanism for the formation of gaseous and particulate products from ozone-cycloalkene reactions in air. Envir. Sci. Technol. 21, 52-57. Jobson B. T., Niki H., Yokouchi Y., Bottenheim J., Hopper F. and Leaitch R. (1994) Measurements of C2-C 6 hydrocarbons during Polar Sunrise Experiment 1992: evidence for C1 atom and Br atom chemistry. J. geophys. Res. 99D, 25,355-25,365. Kawamura K. (1993) Identification of C2-C10 og-oxocarboxylic acids, pyruvic acid, and C2-C 3 ~t-dicarbonyls in wet precipitation and aerosol samples by capillary GC and GC/MS. Anal. Chem. 65, 3505-3511. Kawamura K. and Gagosian R. B. (1987) Implication of t~-oxocarboxylic acids in the remote marine atmosphere for photo-oxidation of unsaturated fatty acids. Nature 325, 330-332. Kawamura K. and Gagosian R. B. (1990) Mid-chain ketocarboxylic acids in the remote marine atmosphere: Distribution patterns and possible formation mechanisms. J. atmos. Chem. 11, 107-122.

Kawamura K. and Ikushima K. (1993) Seasonal changes in the distribution of dicarboxylic acids in the urban atmosphere. Envir. Sci. Technol. 27, 2227-2235. Kawamura K. and Kaplan I. R. (1987) Motor exhaust emissions as a primary source for dicarboxylic acids in Los Angeles ambient air. Envir. Sci. Technol. 21, 105-110. Kawamura K. and Usukura K. (1993) Distributions of low molecular weight dicarboxylic acids in the North Pacific aerosol samples. J. Oceanography 49, 271 283. Kawamura K., Kasukabe H., Yasui O. and Barrie L. A. (1995) Production of dicarboxylic acids in the arctic atmosphere at polar sunrise. Geophys. Res. Lett. 22, 1253-1256. Li S. M. and Winchester J. W. (1993) Water soluble organic constituents in arctic aerosols and snow pack. Geophys. Res. Lett. 20, 45-48. Marty J. C., Saliot A., Buat-Menard P., Chesselet R. and Hunter K. A. (1979) Relationship between the lipid composition of marine aerosols, the sea surface microlayers and subsurface water. J. geophys. Res. 84, 5707-5716. Norton R. B., Roberts J. M. and Huebert B. J. (1983) Tropospheric oxalate. Geophys. Res. Lett. 10, 517 520. Satsumabayashi H., Kurita H., Yokouchi Y. and Ueda H. (1990) Photochemical formation of particulate dicarboxylic acids under long-range transport in central Japan. Atmospheric Environment 24A, 1443-1450. Sempere R. and Kawamura K. (1994) Comparative distributions of diearboxylic acids and related polar compounds in snow, rain and aerosols from urban atmosphere. Atmospheric Environment 28, 449-459. Talbot R. W., Andreae M. O., Berresheim H., Jacob D. J. and Beecher K. M. (1990) Source and sink of formic and acetic, and pyruvic acids over central Amazon. J. 9eophys. Res. 95, 16,799-16,811. Welch H. E., Muir D. C. G., Billeck B. N., Lockhart W. L., Brunskill G. J., Kling H. J., Olson M. P. and Lemoine R. M. (1991) Brown snow: a long-range transport event in the Canadian Arctic. Envir. Sci. Technol. 25, 280-286. Zhou X. and Mopper K. (1990) Measurement of sub-partsper-billion levels of carbonyl compounds in marine air by a simple cartridge trapping procedure followed by liquid chromatography. Envir. Sci. Technol. 24, 1482-1485.