Hydrocarbons in Particulate Fallout of Alexandria, Egypt: Sources and Implications TAREK A. T. ABOUL-KASSIM AND BERND R. T. SIMONEIT* Petroleum and Environmental Geochemistry Group, College of Oceanic and Atmospheric Sciences, Oregon State University, Corvallis, Oregon 9 7331
Particulate fallout samples (PFS) were collected in Alexandria, and their aliphatic and aromatic hydrocarbon compositions were determined both quantitatively and qualitatively t o characterize the homologous and biomarker compounds in terms of their original sources. The results show that all samples contain aliphatic hydrocarbons, including n-alkanes, UCM, isoprenoids, tri- and tetracyclic terpanes, hopanes, and steranes/diasteranes. The main source of these compounds is from petrochemical contamination with trace input of terrestrial higher plant wax. In addition, polycyclic aromatic hydrocarbons, which are considered to be combustion products from fossil fuels such as petroleum, are also widely distributed in all samples. Multivariate statistical analysis, including extended Q-mode factor analysis and linear programming technique, was performed in order to reduce the hydrocarbon data set into a meaningful number of end members (sources). This analysis indicates that there are two significant end members explaining 90% of the total variation among the samples and confirming petrochemical (79.6%), and thermogenic/pyrolytic (10.4%) sources in the PFS model.
Introduction Hydrocarbons (aliphatic and aromatic) are widespread organic compounds that are part of the carbon cycle in the contemporary environment. Natural hydrocarbons produced by terrestrial and marine plants are generally encountered at trace levels, with their characteristic components or distributions reflecting their origin and transport. Anthropogenic hydrocarbons are widely distributed and originate from different sources, such as petroleum, coal, and wood via their combustion products. Accordingly, the anthropogenic hydrocarbon sources are * Author for correspondence; Fax: 503-737-2064.
0013-936W95/0929-2473$09.00/0
0 1995 American Chemical Society
found mostly in heavily populated cities associated with industrial and urban activities. The transport of pollutants by atmospheric particulate matter can expose large populations to toxic substances (carcinogenic or mutagenic), representing the environmental process ofhighest potential hazard to human health, especially when the proximity between humans and pollutant sources is closest. Thus, a direct cause-effect relationship between the atmospheric release of allergenic products and urban asthma outbreaks has been documented (e.g.,ref 1). In addition, many organic compounds, i.e., hydrocarbons, found in urban air or in fallout deposits have been shown to be mutagens or carcinogens (e.g.,refs 2-5). The importance of long-range atmospheric transport to the global distribution of pollutants and natural products is well documented (€4. Major environmental problems in the southeastern Mediterranean region are the result of anthropogenic activity and, over remote areas, of dust storms prevailing between September and June, mostly during the spring season (7-10). These dust storms are called Saharan dust (Haboob in Sudan and Khamasin in Egypt),which are carried in the atmosphere by high velocity winds with large vertical motions that are associated with cold or warm low pressure systems (9, 11-12). Thus, the dust is transported by westerly or southwesterly winds to Egypt (3,13)andthentoIsrael(14-16). GanorandMamane (17) reported that dust storms in the eastern Mediterranean usually contain high concentrations of total particulate solids. Alexandria (Egypt) in the southeastern Mediterranean region is one location for assessing the hydrocarbon composition in the atmosphere, because it received particulate mixtures from local and Khamasin dust sources during the study period. Thus, the characterization of the hydrocarbon constituents in particulate fallout in terms of their natural or anthropogenic sources serves to define and elucidate the regional transport pattern. This is the first comprehensive characterization of the hydrocarbon components in particulate fallout in the southeastern Mediterranean. Previous research was conducted in the northeastern (18-21) and western Mediterranean (4, 22-27). The objectives in this paper are to (a) characterize both aliphatic and aromatic hydrocarbons in the solventextractable fraction of particulate fallout collected from different areas in the Alexandria region (Figure 1);(b) use biomarkers to characterize the different sources; and (c) carry out statistical data analyses using both univariate and multivariate statisticalprocedures to examine the variations in the data, determine the regions of hydrocarbon concentration, and find statisticallysignificant associations (end members) in the data set that will help to assess and identify hydrocarbon composition sources in the atmospheric dry fallout over the city.
VOL. 29, NO. 10, 1995 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 12473
Mediterranean Sea
. . ..
,." .:1: '.'..
0
0
City ,Center
I
I
I
29*d
30. E
31. IO'
J
FIGURE 1. Study area around Alexandria (sampling stations: solid squares, beach area (A); solid triangles, city center (B); open squares, eastern area (C); solid circles, western area (D)).
Experimental Methods Sampling. Since this study is a preliminary investigation about the hydrocarbon composition in the southeastern Mediterranean, atmospheric particulate fallout samples (PFS) were collected from locations in Alexandria representing 4 main zones, namely, beach area (A), eastern (C) and western city (D), and central city (B) (Figure 1). The characteristics of each zone are described in Table 1. Samples were collected by exposing plastic trays covered with solvent-precleaned aluminum foil for typically 45-67 days. These trays were installed on the top of high buildings in each zone (ranging from 18-28 floors high Le., 50-80 m height) to avoid any significant interferences from the ground. Sample Extraction and Separation. To minimize contamination, all glassware was cleaned with soap and water, rinsed with distilled water, heated in an oven at 550 "C for 8 h to combust any traces of surficial organic matter, and finally rinsed twice with ultrapure methanol and methylene chloride. The KOH used for saponification was extracted three times with n-hexane and once with methylene chloride in a separatory funnel to remove organic interferences. An extractionprotocol modified from Aceves and Grimalt (27)was designed for the qualitative and quantitative analyses of different organic compounds found in the lipids of the PFS and explained in detail in Aboul-Kassim and
Simoneit (28). In brief, samples (typically 1.5 gl were extracted in a Soxhlet apparatus with methylene chloridelmethan01 (2:l). This extract is a measure of the amount of extractable organic matter (EOM)in a sample. All the extracts (EOM)were concentrated to 2 mL and hydrolyzed overnight with 35 mL of 6% KOH/methanol. The corresponding neutral and acidic fractions were successively recovered with n-hexane (4 x 30 mL). The neutral fraction was fractionated by column chromatography, using a column (50 x 1.2cm) filled with 8 g each of alumina (top) and silica (bottom), both deactivated with 5% water. The following fractions were eluted and collected: (a) 45 mL of n-hexane (aliphatic hydrocarbons, Fl), (b) 25 mL of 10%methylene chloride in n-hexane (monoaromatic hydrocarbons, F2), (c)40 mLof 20%methylene chloride in n-hexane (polycyclic aromatic hydrocarbons PAHs, F3), followed by three additional fractions representing ketones, aldehydes,esters, and alcohols. Here we describe the results for fraction 1 representing the aliphatic hydrocarbons and for fractions 2 and 3 representing the aromatic hydrocarbons. Organic CarbonAnalysis. Organic carbon analyseswere carried out for all the particulate fallout samples using a Carlo Erba NA-1500 CNS analyzer. The concentrations of the hydrocarbon fractions were calculated relative to the total organic carbon (TOC) content of the PFS. Instrumental Analyses. High-resolution gas chromatography (HRGC) of the hydrocarbon fraction of the PFS was conducted on a Hewlett Packard (HP) 5890A gas chromatograph (GC), equipped with a splitlsplitless capillary injection system and a flame ionization detector (FID). The samples were analyzed in the splitless mode using a fused silica capillary column (30 m x 0.25 mm id., DB-5, J&W Scientific) with a 0.25 pm film thickness and using helium as the carrier gas. The analog signal was monitored and/or integrated with an HP 3393A integrator. The GC conditions were as follows: FID 300 "C, injector 300 "C, initial oven temperature 65 "C, programmed to 290 "C at 4 W m i n , isothermal at 290 "C (60 min). The gas chromatography/mass spectrometry (GUMS) analyses of the samples were performed using a Finnigan 9610 GC (identical column with initial temperature 50 "C, isothermal 6 min, programmed at 4 "C/min to 310 "C, isothermal 60 min) interfaced directly with a Finnigan 4021 quadrupole mass spectrometer (electronimpact, emission current -0.45 mA, electron energy 70eV, scanned from 50 to 650 Da). Data were acquired and processed with a Finnigan-INCOS Model 2300 data system. Compound Identificationand Quantification. Compound identification was based on comparison with the retention times and mass fragmentation patterns of stan-
TABLE 1
locations, Characteristic Features, and TOC Contents of Particulate Fallout Samples from Alexandria, Egypt zone
zone location
A
beach area
6 C D
city center eastern side western side
characteristic features residential area, in full contact with marine aerosols residential area with heavy traffic industrial region, with traffic. industrial region, heavy traffic, oil refineries, factories, garbage incineration
sampling period in 1992
no. of samples
mean TOC concn (mg of C/g)
mean concn of EOMa(mg/g)
April 27-July 4
4
89.6 f 12.2
17.1 f 5.3
M a y 12-June 16 M a y 5-June 27 M a y 6-July 7
4 3 3
average a
Extractable organic matter.
2474
ENVIRONMENTAL SCIENCE &TECHNOLOGY / VOL. 29. NO. 10, 1995
96.8 f 9.5 92.8 f 6.2 127.0 f 8.4
10.0 f 3.1 19.5 f 4.9 11.2 f 2.1
101.6 3~ 14.9
14.5 f 3.9
dard reference materials and with the help of the National Institute of Standards and Technology (NIST) standard library (incorporated in the INCOS data system). This was achieved using the following standard mixtures injected in both GC and GUMS: (a) a series of normal alkanes ranging from n-Clo to n-C36; (b) regular isoprenoids as pristane and phytane: and (c) a suite of about 30 polycyclic aromatic hydrocarbons (PAHs),including naphthalene, methyl- and dimethylnaphthalene, fluorene, 9-methylfluorene, dibenzothiophene, phenanthrene, 3-, 2-, 9-, and l-methylphenanthrene, anthracene, fluoranthene, pyrene, 2,3-benzofluorene, 1,l’-binaphthalene, benz[a]anthracene, chrysene, benzo [b+k]fluoranthene,benzo [elpyrene,benzo [alpyrene, perylene, 9,IO-diphenylanthracene,dibenz[ah]anthracene, benzo [ghQperylene, anthanthrene, coronene, and dibenzo[aelpyrene. Quantification was based on the application of perdeuterated compounds (e.g., nG2Dssand pyrenedlo)as internal standards for the aliphatic and aromatic fractions, respectively (after fractionation of the total extracts). In order to correct for detector response, sets of relative response factors were determined for everyfraction from multiple injections. Normal and isoprenoid alkanes were quantified on the GC, while the terpane and sterane biomarkers and PAH were determined by GCIMS. Statistical Analyses. Organic geochemical data on the hydrocarbon compositions and biomarkers were examined statistically in order to determine any significant environmental variation. All statistical analyses were performed using the SPSS/PC+statistical package (29)as well as the statistical package provided on the SUN system of the College of Oceanic and Atmospheric Sciences at Oregon State University. These analysesinclude standard deviation (measure of dispersion), correlation between variables, extended Q-mode factor analysis, and linear programming techniques. The particulate fallout deposits (PFD) model of Alexandria consists mainly of both factor analysis and linear programming techniques. Factor analysis (FA) is the statistical technique that effectively reduces the raw data into a smaller number of hypothetical variables called factors. There are two types of FA, known as R- and Q-mode techniques. The first is concerned with the relationships between variables while the second is Concerned with the interrelationships between samples. In the present work, Q-mode FA will be employed. The quantitative compositions of the hydrocarbon fractions from each sample were submitted to extended Q-mode FAand linear programming technique (LPT). LPT was used to identify end member compositions (28,30). Q-mode FA provides a description of the multivariate data set in terms of a few orthogonal end members (factors), which account for the variance within the data set. Because transformations of the original data variables during the analysis result in negative factor scores for some variables and negative concentrations for others in the end member, we used the “new rotation” technique proposed by Leinen and Pisias (30) which, in addition, does not require the assumption of having sampled pure end members. The criteria for choosing the number of end members used to model the data were as follows: (a) at least 90% of data set variance was explained by the sums of squares of the end members, and (b) all end member factors that explained less than 2% of the total variance were rejected. Once the number and composition of the end members are determined, the next step is to obtain a quantitative
estimate of the relative amount of each end member in each sample. Because the fallout samples are regarded as simple mixtures, the bulk composition of each fallout sample is assumed to consist of some linear combination of end member compositions, each sample can be represented mathematically as a system of n equations ( nis the number of individual organic geochemical variables used to identify the compositional end members) in m unknowns (mis the number of major compositional end members that are present) of the form:
where Spl, Sp2, ... Spnare the measured concentrations of variables p1,2 ... pn in the sample; Elpl ... Empn are the concentrations of variables pl ... pn in the compositional end members El ... Em determined by factor analysis; Kl, K2, ... K,, are unknowns whose magnitude for each fallout sample reflects the relative contributions of each compositional end member in that sample; and Rpl, Rp2, ... Rpn are residual terms reflecting the fact that each equation is inexact due to sampling andlor analytical error. These systems of equations are usually overdetermined ( n > m) in organic geochemical applications, thus optimum solutions can be obtained using linear programming methods (31, 32). The major advantage of obtaining a linear programming solution is that certain physical constraints can be incorporated into the mathematical calculations (33, 34). For example, the linear programming solution specifies that no compositional end member can have a negative contribution to the total composition of a fallout sample. The residual terms associated with each system of equations represent the difference between the linear programming estimate and the actual concentration of each organic chemical parameter in the sample. The optimum solution for each system of equations is that for which the residual terms are minimized. Since a perfect modeling solution would account for 100% of the measured concentration for each organic chemical compound, we have evaluated the validity of our model by calculating a mean residual percent of each compound (i.e.,the mean residual for each compound divided by the mean compound concentration). Thus, the use of LPT partitioning helped to correct the initial end member compositions and their abundance to better fit the observed multivariate data set as well as to specify and select the compositions of the end members. Organic Geochemical Ratios. The following calculations based on biomarker indices (35)were made to show relationships among samples (Table 2 ) : (1)Carbon Preference Index (CPI),ameasure of the ratios of odd-to-even carbon numbered n-alkanes, is calculated according to the following equations (36): CPI, (whole range, petroleum) =
VOL. 29, NO. 10, 1995 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
2475
TABLE 2
Mean Aliphatic Hylrecerben Colllperitiens Relative to Weight and Total Organic Carbon Content, as Well as Biomarker Ratios of Particulate Fallout Samples from Different Zones of Alexandria concentrations' compound class
composition MW
zona A
zone B
zone C
zone D
IDb
Aliphatic Hydrocarbons
Emalkanes (cl4-c38) (pg/g)' Zterrestrial wax n-alkanes (c21-c37) (pg/g)c)d UCM (,ug/gP isoprenoids (pg/g)c 2,6,10,14-tetramethylpentadecane (pristane) 2,6,10,14-tetramethylhexadecane (phytane) Zisoprenoids tricyclic terpanes (ng/g) Clg-tricyclic Czo-tricyclic Czl-tricyclic C23-tri~y~li~ C24-tricyclic C 2 5 - t r i ~ ic y~I C26-tricyclic (S) C26-tricyclic ( R ) C28-tricyclic C29-tricyclic Ztricyclics tetracyclic terpanes (ng/g) Cz4-tetracyclic(17,21 -secc-hopane) Cz8-tetracyclic(18,14-seco-hopane) CZg-tetracyclic(18,14-seco-hopane) Ztetracyclics pentacyclic triterpanes (ng/g) 18a(H)-22,29,30-trisnorneohopane (T,) 17a(H)-22,29,30-trisnorhopane (T,) 17a(H),21P(H)-norhopane 17p(H),Pla(H)-norhopane 17a(H),2 IP(M-hopane li'P(H),Pla(H)-hopane 17a(H),21/3(H)-homohopane (22s) 17a(H),2 I/3(H)-homohopane (22R) gammacerane 17a(H),21~(H)-bishomohopane(22s) 17a(H),21P(H)-bishomohopane(22R) 17a(H),21P(H)-trishomohopane(22s) 17a(H),21P(H)-trishomohopane(22R) 17a(H),21/3(H)-tetrakishomohopane(22s) 17a(H),21~(H)-tetrakishomohopane(22R) 17a(H),21P(H)-pentakishomohopane(22s) 17a(H),21/3(H)-pentakishomohopane(22R) Zpentacyclics diasteranes (ng/g) 13a(H),17P(H)-diacholestane (20s) 13a(H),17P(H)-diacholestane (20R) Zdiasteranes steranes (ngig) 5a(H),14a(H),17a(H)-cholestane (20s) 5a(H),14P(H),17P(H)-cholestane (20R) 5a(~,14p(H),17p(H)-cholestane (205) 5a(H),14a(H),17a(H)-cholestane (20R) 5a(~),14a(H),17a(H)-ergostane (20s) 5a(H),14p(H),17P(H)-ergostane (20R)
5a(H),14p(H),17/3(H)-ergostane (20s) 5a(H),14a(H),17a(H)-ergostane (20R) 5a(H),14a( H), 17a(H)-sitostane (20.9 5a(~),14p(H),178(H)-sitostane (20R) 5a(H),1 4/3(H),17P( H)-sitostane (20s) 5a(~),14a(H),17a(H)-sitostane (20R) Zsteranes
727 (7509) 325 (3666) 72.5 (764.5) 2.7 (29.4) 2562 (28611) 1717 (19174)
937 (11515) a 45.1 (354.4) 5903 (72545)
254 0.2 (2) 268 1.4 (15) 1.6 (17)
1.2 (13) 1.7 (17) 2.9 (30)
0.6 (7) 1.6 (17) 2.2 (24)
BD 2.1 (16) 2.1 (16)
a a
262 276 290 318 332 346 360 360 388 402
BDe 0.01 (0.12) BD 0.03 (0.33) 0.02 (0.23) 0.02 (0.23) 0.01 (0.11) 0.01 (0.1 1) 0.01 (0.11) 0.01 (0.11) 0.14 (1.41)
0.01 (0.10) 0.02 (0.21) 0.01 (0.11) 0.06 (0.62) 0.04 (0.41) 0.04 (0.39) 0.02 (0.21) 0.01 (0.10) 0.02 (0.21) 0.04 (0.40) 0.27 (2.76)
0.01 0.04 0.01 0.11 0.09 0.10 0.05 0.04 0.06 0.08 0.59
(0.11) (0.43) (0.11) (1.19) (0.97) (1.08) (0.54) (0.43) (0.65) (0.86) (6.39)
0.02 (0.16) 0.05 (0.39) 0.01 (0.08) 0.14 (1.10) 0.09 (0.71) 0.11 (0.87) 0.02 (0.16) 0.03 (0.24) 0.05 (0.39) 0.11 (0.87) 0.63 (4.97)
b b b b b b b b
330 0.01 (0.12) 386 0.01 (0.11) 400 ED 0.02 (0.23)
0.02 (0.21) 0.03 (0.31) 0.04(0.41) 0.09 (0.93)
0.07 (0.75) 0.06 (0.65) 0.11 (1.19) 0.24 (2.59)
0.03 (0.24) 0.08 (0.63) 0.11 (0.87) 0.22 (1.74)
b b b
370 370 398 398 412 412 426 426 412 440 440 454 454 468 468 482 482
0.10 (1.03) 0.13 (1.34) 0.94 (9.71) 0.20 (2.07) 1.05 (10.85) BD 0.55 (5.68) 0.43 (4.44) 0.07 (0.72) 0.36 (3.72) 0.26 (2.69) 0.26 (2.68) 0.19 (1.96) 0.19 (1.97) 0.16 (1.65) 0.21 (2.17) 0.10 (1.03) 5.20 (53.701
0.33 (3.56) 0.68 (7.33) 2.23 (24.03) 0.16 (1.72) 2.50 (26.94) 0.25 (2.69) 1.79 (19.29) 1.24 (13.36) 0.18 (1.94) 0.92 (9.91) 1.04 (11.21) 0.91 (9.81) 0.50 (5.39) 0.54 (5.82) 0.42 (4.53) 0.75 (8.08) 0.63 (6.79) 15.10 (162.40)
0.44 (3.46) 0.48 (3.78) 2.01 (15.82) 0.20 (1.57) 2.26 (17.79) BD 1.49 (11.73) 1.06 (8.34) 0.10 (0.79) 0.96 (7.56) 0.77 (6.06) 0.73 (5.75) 0.49 (3.86) BD BD 0.49 (3.86) 0.49 (3.87) 12.00 (94.20)
a b b b b b b b
372 0.01 (0.11) 372 0.01 (0.11) 0.02 (0.22)
0.02 (0.21) 0.02 (0.22) 0.04 (0.43)
0.05 (0.54) 0.09 (0.97) 0.14 (1.51)
0.06 (0.47) 0.05 (0.39) 0.11 (0.86)
b, d b, d
372 372 372 372 386 386 386 386 400 400 400 400
0.04 0.02 0.02 0.07 0.09 0.03 0.05 0.04 0.06 0.04 0.05 0.09 0.60
0.21 (2.26) 0.31 (3.34) 0.18 (1.94) 0.21 (2.26) 0.06 (0.66) 0.30 (3.23) 0.25 (2.69) 0.21 (2.26) 0.24 (2.59) 0.32 (3.45) 0.31 (3.34) 0.24 (2.59) 2.84 (30.61)
0.40 0.33 0.20 0.41 0.19 0.32 0.27 0.32 0.38 0.39 0.31 0.32 3.84
b b b b b
0.50 1.20 0.30 5.30 776
1.10 1.20 0.90 6.30 236
60 (673) 2.0 (23.7) 186 (2086)
0.03 (0.33) 0.10 (1.12) 0.48 (5.36) 0.05 (0.56) 0.35 (3.91) BD 0.27 (3.11) 0.18 (2.11) 0.05 (0.56) 0.15 (1.67) 0.12 (1.34) 0.12 (1.34) 0.10 (1.12) 0.11 (1.23) 0.07 (0.78) 0.10 (1.12) 0.09 (1.00) 2.40 (26.70)
0.02 0.03 0.03 0.02 0.03 0.04 0.03 0.03 0.03 0.05 0.03 0.03 0.37
(0.22) (0.33) (0.34) (0.21) (0.33) (0.46) (0.33) (0.32) (0.31) (0.56) (0.32) (0.31) (4.05)
Biomarker Ratiosd 1.10 1.30 0.80 3.60 115
2476 1 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 29, NO. 10, 1995
1.20 1.40 1.10 6.52 46
(0.41) (0.21) (0.21) (0.72) (0.93) (0.27) (0.52) (0.41) (0.62) (0.41) (0.43) (0.92) (6.02)
(3.15) (2.60) (1.57) (3.23) (1.50) (2.52) (2.13) (2.52) (2.99) (2.98) (2.44) (2.52) (30.15)
C C
C
b b b b b b b b
b b b b b b
b. d
Table 2 Continuedl concentrations'
compound class
composition
Mw
zone B
zone C
zone D
0.60 0.56 0.55 0.61 0.53
0.56 0.58 0.58 0.54 0.68
0.59 0.46 0.65 0.56 0.54
0.58 0.55 0.60
0.50 0.38
0.40 0.56
0.50 0.49
0.54 0.44
zone A
IDb
Bionarker Ratioid
c35
0.54
sterane epirnerization
(ana)
(@PI
*Expressed as weighvg of particulate fallout; v s l u ~ sin parentheses are concentrations relative to g of TOC. IO = compound identification 18, positive; b, probable; C. possible: d, tentative), for more details see Rogge etaf. (ref 411. =Values relative to organic carbon were rounded off. See t e n (Organic Geochemical Ratios) for more details. *BO. below detection.
CP,I
(split range,
I.
CPI, (split range, higher plant wax) = ~~~C21-C3,~/~~C20-C3dl (2) Terrestrial higher plant n-alkane signature is calculated by subtraction of the average of the next higher and lower even carbon numbered homolog (37) as follows:
8
a
where negative values of wax C, were taken as zero. (3) Ratio of unresolved to resolved hydrocarbons (U/R) i s calculated (36) as U / R = concentration of unresolved complex mixture (UCM)/concentration o f resolved hydrocarbon peaks (mostly n-alkanes) in this case. the P/W (petroleum to resolved biogenic wax hydrocarbons) = concentration of IUCM T (R - W)I/W. (4) Homohopane index (HHl) i s the ratio between the epimer at C-22S and R for the 17a(H)-homohopane series [C31-C35):
Time-
FIGURE 2 GC traces of n-alkanes [dots over peaks. IS = internal standard)of particulatefallout samplesfromthe la) western and lbl city center areas of Alexandria. Egypt.
HHI = [22S/(22S 7 22R)I (51 Sterane epimerization parameterat C-20i s calculated
for Csi as 5a(H).14a(H).17a(H)-C29-sterane=
[(2OS)/(2OS7 20R)I
111~1 f l I I il II lb
I
Results a d Discussion I h e sampling locations in Alexandria with their specific charactenstics.dates. andT0Ccontentsofthe PFSaregiven in Table I . The total solvent-extractable organic matter (EOMI was a maximum of 19.5 mg/g P I 3 for zone C and a low of 10.0 mg/g for zone B (Table 1). with an average of 14.5 mg/g. The EOM yield relative to TOC reached a maximum of 19% in zone A. a low of 11% in zone R. and averaged 14%. The average total hydrocarbon (HC) content of the EOM was 37% and was comprised of 68% aliphatic and 32% aromatic hydrocarbons. AUphatic Hydrocarbon Composition. Normal and isoprenoid alkanes and an envelope (hump) of an UCM of branchedand cyclic hydrocarbons were present and ranged incarhonchainlengrhfromC14toCla(Figure2). The total
FIGURES. Aliphatic hydrocarbon concenhations reletiveto organic carbon content of the PFS from Alexandria.
concentrations for the aliphatic hydrocarbons relative to both weight and TOC are given in Table 2. The average n-alkane concentrations in the study area relative to TOC are shown as an overview in Figure 3. All zones exhibit a minor predominance of the odd carbon n-alkanes from C2?to C,l, which indicates a trace input of terrestrial plant VOL 29. NO 10. 1995 ENVIROhMENTAL SCIENCE 6 TEChhOLOGY ,2477
21
23
25
27
29
31
33
Carbon number
I E -wiu
i Iwc-w
E4 umM.(cl
m
yrmm”BI
I
FIGURE 4. Wax malkane concentrationsrelative to organic carbon contem of the PFS from Alexandria.
wax. The total n-alkane concentrations were about 0.7, 7.5, 3.7, and 11.5 mg/g OC for zones A-D, respectively. Figure 2 indicates that the hydrocarbon distributions have differentcontributions from both petroleum and biogenic sources. The identification of the homologous n-alkanes in the hydrocarbon fractions allowed the determination of the CPI, C,,, UCM, and wax n-alkanes for each sample set. Each provides supportive evidence for the relative incorporation of both biogenic and anthropogenic components. The CPl (363,which is a measure of biologically synthesized n-alkanes, was determined for these samples. CPI values of automobile exhaust particulate matter, diesel soot, and motor oil are near unity (e.g., refs 18,36, and 38). On the other hand, major alkanes present in waxes ofplants and algae, pollen, and fungi (e.g., ref391 are generally oddcarbon numbered, with CPI values higher than unity. CPI values can therefore indicate the relative contributions of n-alkanes fromnaturalcompared to artificialsources. Thus, a low CPI (-1) indicates significanthydrocarbon pollution and/or degradation effects. In the PFS from Alexandria (Table2). the CPI, (full range) values had an average of 1.0 with small differencesbetween the sampling sites, indicating significant petroleum pollution. A more sensitive method to calculate CPI is to split the carbon number range into low (CL3-CI9) and high (CZaC37) ends (36). This split demonstrates the relative input of contemporary biogenic versus anthropogenic materials with greater clarity, i.e., the low end (C13-CL9)is produced bymicrobiotaoris present involatile fossilfuels (e.g., diesel) (e.g., refs 39 and 40). On the other hand, the high weight range (CZO-c37) occurs in higher plant waxes and in fossil fuel detritus ofurban areas (e.g., refs 39-42). The CPIbfor the PFS was relatively high for zone B (1.41, while the CPI, was a low of 0.3 for zone C (Table 2). Because the n-alkanes of petroleum and terrestrial biogenicoriginsaremixedtogetherinthePFS,asubtraction of the corresponding n-alkane concentration with CPI, = 1 was carried out (37) to determine the distribution signatures of the residual plant wax alkanes. There is a significant error associated with this approach when the petroleumcomponent islarge (37).TableZshowsthemean terrestrial wax n-alkane series relative to the weight and TOC content of the PFS, while Figure 4 shows the distribution patterns of the remaining odd-carhon numbered alkanes. It is obvious that each sample has a different n-alkane distribution derived from epicuticular waxes and 2478 m ENVIRONMENTAL SCIENCE &TECHNOLOGY I VOL. 29.
NO. 10.1995
representing about 0.3% maximum of the total resolved aliphatic hydrocarhons. The wax alkane concentrations range between 23.7 and 764.5 pglg of OC, with an average of 293 pg/g of OC. Thedeterminationofthe C,,canalsogiveanindication of the relative source input (36). In general, a C, 2 CZ5 for n-alkanes reflects the incorporation of higher plant wax and C, at lower carbon numbers (Figures 2 and 3) may indicate a major input from microbial or petroliferous sources. The dominant C, determined for the n-alkanes of the PFS are CZ9(zone B), CZ5(zones A and C), and CZ3 (zone D).These C, support the interpretation of a minor terrestrial contribution to the fallout. while different C, values may result from regional differences in source strengths of terrestrial plants dominant in each zone. Another diagnostic parameter is the UCM of branched and cyclic hydrocarbons. This UCM is interpreted here to be derived from the utilization of petroleum products (e.g., refs40-43). The highvalue, i.e., 72.6mg/gofOC, occurred in zone B, while a low d u e of 2.1 mg/g of OC occurred in zone A (Table 2), with a regional average of 30.6 mg/g of OC. The values of the U/R ratio (36)and the P/W ratio are also used as a criterion to assess anthropogenic input. The U / R values for the hydrocarbons from the different zones (Table 2) ranged from 3.5 to 6.3 (mean 4.3). and the P/W ratios ranged from 46 to 776 (mean 294). Plant wax hydrocarbons have UIR < 0.1. i.e., they have no UCM (36, 43). The inverse significant correlation (r = -0.7821) between wax n-alkanes and P/W ratios also indicates the usefulness of the P/W ratio as a measure of the anthropogenicinputinsample mixtures. AU/Rvalue >2 reflects significantcontamination by petroleum products (e.g., refs 36 3and 7). Aceves and Grimalt (27)reported that the U/R ratio for the paniculate aerosol fraction samples collected over Barcelona (Spain, Western Mediterranean) ranged between 6.9 and 15.0,with an average of 8.0. On the other hand, Stephanou and Stratigakis (20)gave a value of 4.06.6 for aerosol particles over the coastal area of Heraklion, Island of Crete (Greece, Northeastern Mediterranean). Gogou et al. (21) compared the differences in lipid composition of aerosols from Eastern and Western Mediterranean coastal cities, reporting U/R values ranging between 4.0 and 6.6 and between 17.7and 22.0 for Heraklion and Barcelona, respectively. For atmospheric fallout in the Tokyo area, Matsumoto and Hanya (44)reported a U/R ratio of 2.6-9.2 (average Of 5.61, reflecting varying degrees of urban pollution. The strong inverse correlation between U/R and CPI of these samples (r = -0.9684, p 5 0.001) can be explained by the admixture of petroleum contaminants, which reduces the CPI to values of ahout 1. Biomarkers. The second main group of organic compounds studied in the aliphatic fraction of the EOM from the PFS are the molecular biomarkers. They are specific organic indicator compounds (found in geological and environmental samples) that can be utilized for genetic source correlations (e.g., refs 38, 40, 43, and 45). Such molecules are characterized bytheir restricted occurrence, source specificity, molecular stability, and suitable concentration for analytical detection (43). Biomarkers have been utilized as confirmation indicators for petroleum residues, higher plant waxes, and pyrogenic components in the solvent-earactable organic matter of aerosols and PFS. The biomarkers examined in this study are the isoprenoids, tri- and tetracyclic terpanes, 17a(H),21p(Ifhopanes, steranes, and diasteranes. The following will
)'
1
'"i
24T
29lT 29T
FIGURE 5. Massfragmentogramsrepresenting(a)tri- and tetracyclic terpane (nJz 191), (b) hopane (summed d z 149, 177, 191,205, 219, 233,247, and 261), and (c) starane/diesterana (summed dz217.219, and 259) series representing the anthropogenic biomarkers in the PFS from Alexandria [T = tricyclana, TT = tetracyclane, T, = 18a(H)-22R9,30-trisnorneohopane, Tm= 17a(H)-2229,30-trisnorhopane, I@= 17a(H),21/9(H)-hopanes (R&S at C-22), G = gammacerane, @/9 = 5a(H),14/9(H),17/9(H)-steranes (R&S at C-20), aaa = 5a(H),14a(H),l7a(H)-steranes (R&S at C-201, I@D = l~(M,17a(H)-diasteranes (R&S at C-20)l.
21-seco-hopanes were proposed to be derived from either thermocatalytic degradation of hopane precursors during geological maturation, microbial ring opening of hopanoids during early diagenesis, or cyclization of squalene to only ring-D (48).In the case of the 8,14-seco-hopanes,a ring-C opened hopane structure was suggested. The PFS contained a C~~-(17,21-seco-hopane) and C28 and C29-(8,14seco-hopanes) (Figure 5a), which were quantified for the different zones (Table 2). The total tetracyclic terpanes in the study area ranged from 0.23 to 2.59 nglg of OC, with an average of 1.37 nglg of OC. PentacyclicMterpanes. The predominant triterpenoid hydrocarbons from petroleum found in atmospheric particulate matter are the 17a(H),21@(H)-hopanes(43).The identification of these compounds is based primarily on their mass spectra and GC retention time. Their occurrence is determined by GClMS using the mlz 191 ion intensity, which is the base peak of most triterpanes (47).An example of the rnlz 191 data is shown in Figure 5b, where the predominant homolog is 17a(H),21@(H)-hopane,with subordinate amounts of 18a(H)-22,29,30-trisnorneohopane;
17a(H)-22,29,30-trisnorhopane; 17a(H),21B(H)-norhopane;
the 17@(H),21a(H)-hopaneseries; and the extended CS1C35 17a(H),21@(H)-hopanes.The hopane series (c27-c35) was present in the PFS of all zones from Alexandria and showed a predominance of 17a(H)-hopane in zones B, C, D, while zone A had a predominance of 17a(H)-norhopane. The distributions of the hopane series are similar for auto confirmingvehicular emissions and diesel exhausts (40,411, as the major source of petroleum residues in PFS. Gasoline examine these biomarkers in more detail. and diesel fuels do not contain these triterpanes, but they Isoprenoid Hydrocarbons. Pristane (2,6,10,14-tetraare present in lubricating oil (40). This indicates that methylpentadecane), phytane (2,6,10,14-tetramethylhexa- lubricants adsorbed as vapor microdroplets on particles decane), and lower molecular weight homologs are geologic carry the biomarker signature of fossil fuels to vehicular alteration products of phytol and are not primary conemissions (40,43). The concentrations of all hopanes in stituents of most terrestrial biota (46).The presence of PFS are listed in Table 2, and the total ranged from 26.7 these isoprenoid biomarkers in the PFS hydrocarbon nglg of OC in zone A to 162.4 nglg of OC in zone C. In typical petroleums, the extended a-hopane homologs >C31 fractions coupled with UCM confirms an origin from petroleum mainly from vehicular exhausts (40,43). The have the epimers at C-22 at an equilibrium ratio {SI (S+R)} of 0.6 [HHI, Seifert and Moldowan (49)].The HHI (Table concentration of pristane in the PFS ranged between 2 and 13pglg of OC, with an average of 3.1 pglg of OC (Table 2). 2) for the PFS varies from 0.46 to 0.68. On the other hand, phytane had a higher concentration Steranes and Diasteranes. Steranes and diasteranes with a maximum of 17 pglg of OC (zones B and C) and a are biomarkers present in fossil fuels and are also useful low of 15 pglg of OC (zone A) (Table 2). The presence of indicators for emissionsfrom vehicular traffic in urban areas isoprenoids in each zone (Table 2) points to a significant (40,43).Both steranes and diasteranes were detected and common petrochemical source in the Alexandria fallout. quantified in the PFS by their mlz 217, 218, and 259 Tricyclic Terpanes. Tricyclic terpanes are important fragmentograms (Figure 5c), and their yields are given in geochemical tracers occurring in almost all crude oils Table 2. The steranes have mainly the 5a(H),14,8(H),17/3(exceptthose derived from terrestrial source materials) and (H)-configuration and a minor amount of the 5a(H),14arange from C19H34 to C45H~6, possibly higher (e.g., ref 43). (H), 17a(H)-configuration. The epimerization ratio at C-20 The tricyclic terpane series present in the PFS ranges from of the C Z sterane ~ is high (0.40-0.54, Table 2), which indicates that the PFS contain mature petroleum residues C19H34 to C29H54, with no CZZand Cz7, and a C23 predominance (Figure 5a, mlz 191 key ion for the series). The C23 (43)* tricyclane was 0.1-1.2 nglg of OC, with an average To summarize, coupling CPI, C ,, UCM, UIR, and wax characteristic for Alexandria of 0.76 nglg of OC (Table 2). n-alkane values with the quantitation of biomarker data Higher concentrations of the tricyclic series occurred in allows the definitionof the different sources for the aliphatic zones C and D. The total tricyclic terpane series concenhydrocarbon fraction of the PFS (terrestrial vs anthropogenic). Experimental analysis allowed only the determitration is 3.87 nglg of OC. The occurrence and variation in the relative distribution of the homologs of this series, nation of different sources in the PFS and not source as determined by GClMS, in these samples makes them strengths (cf. the statistical part). useful tracers for petroleum source identification. Polycyclic Aromatic Hydrocarbons. The second main class of compounds in the PFS consists of PAHs. They are Tetracyclic Terpanes. Another group of biomarkers found in the PFS comprises the tetracyclic terpanes, which considered to be a class of chemical carcinogens and are derivatives of the hopanes (47). Both 17,21- and mutagenic pollutants derived from anthropogenic sources, 8,14-seco-hopanes are found in fossil fuels, and the 17,such as vehicular exhaust (41,50),emissions from coke VOL. 29, NO. 10, 1995 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2479
TABLE 3
Mean Polycyclic Aromatic H y d n s a h Concentrations Relative to Weight and Total Organic Cadon Content of Particulate Fallout Samples from Different Zones of Alexandria concentrationsa compound class
composition
CDolvcvclic aromatic hvdrocarbons (PAHs, ng/g) ~.
alkyl1substituted PAHs ing/g) 3-methylphenanthrene (3MP) 2-methylphenanthrene (2MP) 9-methylphenanthrene (9MP) 1-methylphenanthrene (1MP) dimethylphenanthrenes trimethylphenanthrenes tetramethylphenanthrenes pentamethylphenanthrenes Cphenanthrene series methylpyrenes/fluoranthenes dimethylpyrenes/fluoranthenes trimethylpyrenes/fluoranthenes tetramethylpyrenes/fluoranthenes Xalkyl-202 seriesC methyl-228 dimethyl-228 trimethyl-228 tetramethyl-228 Calkyl-228 seriescrd methyl-252 dimethyl-252 trimethyl-252 tetramethyl-252 Zalkyl-252 seriesCje Ctotal PAHs alkyl PAHs (pglg)
+
zone A
zone B
zone C
zone D
IDb
6362 (71001)
4277 (44186)
82997 (881614)
55426 (43731)
a
7 (82) 7 (75) 10 (106) 11 (118) 114 (1270) 97 (1090) 42 (468) 64 (716) 351.6 (3925) BD 121 (1350) BD BD 121 (1350) 157 (1750) 140 (1560) 107 (1200) BD 404 (4510) 208 (2320) 238 (2660) 173 (1930) 162 (1810) 782 (8720) 8.0 (89.5)
BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD 4.3 (44.2)
478 (5140) 352 (3800) 166 (1790) 253 (2730) 4510 (48590) 5750 (61970) 4900 (31530) 2930 (52760) 19340 (208310) 1530 (16500) 4780 (51470) 5890 (63470) 2430 (26180) 14630 (157600) 392 (4220) 1027 (11070) 617 (6650) 554 (5980) 2590 (27920) 2400 (25840) 5270 (56790) 4170 (44950) 3930 (42320) 15770 (169900) 135.3 (1445.3)
61 (481) 111 (873) 128 (1010) 105 (825) 1420 (11190) 1650 (12960) BD BD 3475 (27339) 578 (4550) 1510 (11850) 2740 (21560) 409 (3220) 5236.9 (41180) 2840 (22350) 1770 (13910) 617 (4860) 465 (3660) 5692 (44780) 1120 (8780) 1350 (10590) 358 (2820) 452 (3560) 3281 (25750) 73.1 (576.4)
a
MW
192 192 192 192 206 220 234 2 48 216 230 244 258 242 256 270 284 266 280 294 308
a
a a b b b d
b b d b b b b b b b b b
a Expressed as weight/g of particulate fallout; values in parentheses are concentrations relative t o g TOC; BD, below detection limit (concentrations are ,-Cn-steranes IS+R, aoo @p), respectively; letters A-R indicate the aromatic hydrocarbons: A = Xphenanthrenes. E =anthracene; C = fluoranthene; 0 = Zpyrenes; E = terphenyl; F = 2,3-benzofluorene; G = Zbenz[alanthracene; H = chqsenettriphenylene: I = henzo[b+k]lluoranthene: J = benzo[e]pyrene; K = benzo[a]pyrene; 1 = Zperylenes; M = indeno[lZ3-cdpyrene: N = dibendahlanthracene: 0 = benzo[ghdperylene; P = anthanthrene; (1 = coronene: and R = dihenzo[aelpyrenel.
+
vehicular exhaust. The second end member (Figure 7) is represented by 10.4% of the total variation in the data set ofthe hydrocarbon fractions. It is dominated mainlybyhigh molecularweight PAH,consisting of benzlnlanthracene (4.5%), benzofluoranthene (10.6%), benzo[al- and benzolelpyrene (8.8%). perylene (7.6%).anthanthrene (6.8%),coronene (8.9%). and dihenzo[aelpyrene(7%). Since combustion processes occur rapidly [consideringa radical formation mechanism (57)1, PAHformationandgrowth bythe addition ofhydrocarbon radicals also occurs quickly, leading to condensation of
heavy PAH on particles. Thus, the presence of high molecularweight PAH in the second end member confirms its origin from thermogenidpyrolytic sources in the PFS model.
I
In addition, a minor insigniscant end member, descrihing ~ 2 . 0 %of the sample variations, was characterized and dominated by odd-carbon numbered n-alkanes madmizing between C19andCu(calculatedasinref37). This represents the trace terrestrialplantwaxcontribution to theAlexandria PFS model VOL. 29. NO. io. 1995 I E N V I R O N M E ~ T A SCIENCE L &TECHNOLOGY
24a1
TABLE 4
Comparison between Characteristic Ratios Dduced from PAH Compositions of Several Potential Source Inputs and Those from Particulate Fallout Samples from Various locations in the Mediterranean ratios'
source
1, PAP+ An)
3, BeP/(BeP+BaP)
refs
4, Inpy/(lnpy+Bper)
this study
Alexandria zone A
zone B zone C zone D crude oil used motor oil car emission, gasoline coal soot coke oven emission Northeastern Mediterranean Heraklion, Greece
Western Mediterranean Barcelona, Spain a
2. Ban/(Ban+Chr+Tri)
0.58 0.96 0.50 0.48 0.98 0.78 0.77
0.34 0.21 0.34 0.19 0.16 0.50
0.67
0.43
0.40-0.46
0.32-0.35
Abbreviations stand forthe following ratios: 1, phenanthrene/(anthracene
0.79 0.40 0.71 0.83 0.87 0.64 0.53 0.30 0.23
0.16 0.00 0.76 0.08
0.80-0.90
0.32-0.45 0.30-0.50
18 21
0.47-0.69 0.50-0.80
0.44-0.52 0.30-0.40
27 21
0.25 0.18 0.56
64 3 65 1
+ phenanthrene);2, benz[alanthracene/(benz[alanthracene+chrysene
+ triphenylene); 3, benzo[elpyrene/(benzo[elpyrene + benzo[alpyrene); 4, indenof 123-cdlpyrene/(indeno[ 123-cdlpyrene+ benzo[ghilperylene). In order to assess the different sources and origins of PAHs present in the PFS, a comparison was made between characteristic ratios calculated from well-characterized emission sources and the same ratios for these samples. Nevertheless, the use of such ratios as definitive identification criteria assumes that only minor modification has occurred from emission ofthe PAH to their deposition (e.g., refs 25 and 58). It has been shown that the degradation rate of PAHs on atmospheric particles depends on the nature of the particles (e.g., refs 59, and 601, where PAHs appear to be better preserved from alteration when associated with soots. Nielsen (61) and Masclet et al. (62)reported the rapid decay of benzo [alpyrene relative to benzo[e]pyrene increased with distance from their source. In the present study, this ratio was 0.79, 0.40, 0.71, and 0.83 for zones A-D, respectively, which may indicate that zone B (the city center) is the source for the benzopyrenes and the ratio increases as the particles are transported to the other zones. Based on the comparisons in Table 4, it can be concluded that more than one source is contributing PAHs to the PFS. Values of the ratios such as [P/(P+ An)]and [BeP/(BeP BaP)] are close to those for crude oil, its products, and vehicular exhaust as well as coke oven emissions. Furthermore, the [Inpy/(Inpy Bper)] values are similar to vehicular emissions in zones A and D, and the [Ban/(Ban Chr Trill are intermediates between crude oil and/or coke oven emissions. In short, a dominant pyrolytic origin can be assigned to the PAH composition of the PFS. Vehicular emissions and higher temperature combustion processes (Le.,industrial emissions) appear to be the major sources for these PAHs. Ratios from both the western and northeastern Mediterranean regions are represented in Table 4. Since carcinogenicity depends mainly on the type and quantity of the aromatic hydrocarbons, human exposure in Alexandria (especially in the east and west, i.e., the industrial areas, Table 1)to atmospheric particulate matter could become a public health concern, especiallywith the high concentrations, which is considered as the most important cause of lung cancer and asthma (e.g., ref 63). Because the present study is preliminqfor that area, future research for health risk assessment should consider some
+
+
+
2482
+
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 29, NO. 10,1995
or all of the following points: (a) point source assessment-the determination of all possible input sources in the area with complete compositional analyses for these sources; (b) hazard identification-the determination of specificchemicalswith detrimental health effects; (c)doseresponse assessment-the determination of the magnitude of exposure and the probability of the occurrence of a health effect; and (d) exposure assessment-the determination of the extent of human exposure before or after application of regulatory controls. Thus, this exposure-response relationship would be based on consistent evidence provided by chemical analyses and information on the occurrence of human lung disease.
Conclusion Particulate fallout samples from a coastal area in the Southeastern Mediterranean (Alexandria, Egypt) contain both anthropogenic and traces of terrestrial organic compounds, which are specific to their emission sources. The anthropogenic components comprise mainly petroleum residues, confirmed by UCM, UIR, pristane, phytane, CPI, C, tricyclic (C19-CZ9) and tetracyclic (&, Cz8 and CZ9 seco-hopanes)terpanes, triterpanes (a@-hopanes),steranes (a#lp configuration with a minor amount of the aaa), diasteranes, and therrnogeniclpyrogenicPAH, ranging from fluorene to coronene with minor alkyl-PAH series. The trace terrestrial source of n-alkanes from plant waxes, determined by subtracting the average alkane distribution (CPI = 11, was observed in the PFS mainly by the slight predominance of Czl, Cz5, CZ7,C29, and C31. The multivariate statistical analyses, including both extended Q-mode FA and LPT, reduced the hydrocarbon data set into two significant end m,embers (sources), explaining 90% of the variation among the PFS. These multivariate techniques represented a useful method for end member source confirmation, representing petrochemical (79.6%) and thermogenic/pyrolytic (10.4%) sources.
Acknowledgments The authors wish to thank Hala Aboul-Kassim and Niven Hamed for sample collection, Fred Prahl for access to TOC
analyses, and Nicklas Pisias for extended Q-mode factor analysis and linear programming technique software. We thankfouranonymous reviewers for their comments,which greatly improved this paper.
Literature Cited (1) Colmsjo, A. L.; Zebuhr, Y. U.; Ostman, C. E.; WLdding, A.; Soderstrom, H. Chemosphere 1986, 15, 169-182. (2) Levin, W.; Wood, W.; Wislocki, P. G.; Chang, R. L.; Kapitulnik,
J.; Mah, H. D.; Yagi, H.; Jerina, D. M.; Conney,A. H. In Polycyclic Hydrocarbons and Cancer. Vol. 1: Environment, Chemistry and Metabolism. Part N: Chemistry, Carcinogenicity and Theory; Gelboin, H. V., TSo, P. 0. P., Eds.; Academic Press: New York, 1978; pp 189-204. (3) Grimmer, G.; Jacob, I.; Naujack, K. W.; Dettbarn, G. FreseniusZ. Anal. Chem. 1981, 309, 13-19. (4) Muel, C.; Saguen, S. Int. J. Environ. Anal. Chem. 1985,19, 111131. (5) Leuenberger, C.; Czuczwa, J.; Heyerdahl, E.; Giger, W. Atmos. Environ. 1988, 22, 695-705. (6) Knap, A. H. The Long-range Atmospheric Transport of Natural and Contaminant Substances; NATO AS1 Series No. 297; Kluwer Academic: Dordrecht, 1990. (7) Chester, R.; Baxter, G. C.; Behairy, A. K. A.; Connor, K.; Cross, D.; Elderfield, H.; Padgham, R. C. Mar. Geol. 1977,24,201-217. (8) Chester, R.; Sharples, E. J.; Sanders, G. S.; Saydam, A. C. Atmos. Environ. 1984, 18, 929-935. (9) Morales, C. Saharan Dust;John Wiley: New York, 1977; 210 pp. (10) Middleton, N. I. J. Arid Environ. 1986, 10, 83-96. (11) Joseph, J. H.; Manes, A.; Ashbel, D. J. Appl. Meteorol. 1973, 12, 792-797. (12) Levin, Z.; Joseph, J. H.; Mekler, Y. J. Atmos. Sci. 1980, 37, 882891. (13) Lunson, E. A. Meteorological Office, Professional Notes 102 MO 5246; HMSO: London, 1950; pp 1-12. (14) Levin, Z.; Lindberg, J. D. J. Geophys. Res. 1979, 84, 6941-6949. (15) Ganor, E. Atmos. Environ. 1991, 25A, 2657-2664. (16) Ganor, E.; Foner, H. A.; Brenner, S.; Neeman, E.; Lavi, N.Atmos. Environ. 1991, 25A, 2665-2670. (17) Ganor, E.; Mamane, Y. Atmos. Environ. 1981, 15, 1-7. (18) Stephanou, E. G. Atmos. Environ. 1992, 26A, 2821-2829. (19) Stephanou, E. G.; Stratigakis, N. Enuiron. Sci. Technol. 1993,27, 1403-1407. (20) Stephanou, E. G.; Stratigakis, N.]. Chromatogr. 1993,644, 141151. (21) Gogou, A.; Stephanou, E. G.; Stratigakis, N.; Grimalt, J. 0.;Simb, R.; Aceves, M.; Albaiges,J.Atmos. Environ. 1994,28, 1301- 1310. (22) Marty, J. C.; Tissier, M. J.; Saliot, A. Atmos. Environ. 1984, 18, 2183-2190. (23) Saliot, A.; Marty, J. C. In Strategies andAdvanced Techniquesfor Marine Pollution Studies: Mediterranean Sea; Gian, C. S., Dou, H. J. M., Eds.; NATO AS1 Series G9; Springer: Berlin, 1986; pp 157- 186. (24) Sicre, M. A.; Marty, J. C.; Saliot, A.; Aparicio, X.; Grimalt, J.; Albaiges, J. Int. J. Environ. Anal. Chem. 1987, 29, 73-94. (25) Grimalt, J. 0.;Albaiges, J.; Sicre, M. A,; Marty, J, C.; Saliot, A. Naturwissenschaften 1988, 75, 39-42. (26) Simb, R.; Colom-Altes, M.; Grimalt, J. 0.;Albaiges, J. Atmos. Environ. 1991, 25, 1463-1471. (27) Aceves, M.; Grimalt, G. 0.Atmos. Environ. 1993,278,251-263. (28) Aboul-Kassim,T. A. T.; Simoneit, B. R. T. Mar. Pollut. Bull. 1995, 30, 63-73. (291 Norusis. M. 1. SPSSPC+Advanced Statistics: SPSS Inc.: Chicago, 1986; 202 pp. (30) Leinen, M.; Pisias, N. Geochim. Cosmochim. Acta. 1984,48,47-
.,
fi2. .~
(31) Gass, S. I. Linear Programming Methods and Applications; McGraw-Hill Book Co.: New York, 1975; 170 pp. (32) Chvatal, V. Linear Programming Freeman Book Co.: San Francisco, 1983; 203 pp.
(33) Dymond, J. Nazca Plate: Crustal Formation and Andean Convergence. Geol. SOC.Am. Mem. 1981, 154, 133-174. (34) Heath, G. R.; Dymond, J. Nazca Plate: Crustal Formation and Andean Convergence. Geol. Soc. Am. Mem. 1981,154,175-189. (35) Tissot, B. P.; Welte, D. H. Petroleum Formation and Occurrence; Springer-Verlag: New York, 1984; p 699. (36) Mazurek, M.; Simoneit, B. R. T. In Identification and Analysis of
Organic Pollutants in Air; Keith, L. H., Ed.; Ann Arbor Science/ Butterworth Publishers: Wobum, MA, 1983; pp 353-370. (37) Simoneit, B. R. T.; Cardoso, J. N.; Robinson, N. Chemosphere 1990, 21, 1285-1301. (38) Simoneit, B. R. T.; Mazurek, M. A.; Cahill,T. A. AirPollut. Control ASSOC.J. 1980, 30, 387-390. (39) Kolattukudy, P. E., Ed. Chemistry and Biochemistry of Natural Waxes; Elsevier: Amsterdam, 1976 418 pp. (40) Simoneit, B. R. T. Int. 1.Environ. Anal. Chem. 1985,22,203-233. (41) Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. Enuiron. Sci. Technol. 1993, 27, 636-651. (42) Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. Environ. Sci. Technol. 1993, 27, 2700-2711. (43) Simoneit, B. R. T. Atinos. Enuiron. 1984, 18, 51-67. (44) Matsumoto, G.; Hanya, T.Atmos. Environ. 1980,14,1409-1419. (45) Simoneit, B. R. T.; Mazurek, M. A. Atmos. Environ. 1982, 16, 2139-2159. (46) Didyk, B. M.; Simoneit, B. R. T.;Brassell,S. C.; Eglinton, G. Nature 1978,272, 216-222. (47) Philp, R. P. Fossil Fuel Biomarkers: Methods in Geochemistry and Geophysics; Elsevier: New York, 1985; Vol. 23, 292 pp. (48) Trendel, J. M.; Restle, A.; Connan, J.; Albrecht, P. J. Chem. SOC. Chem. Commun. 1982, 304-325. (49) Seifert,W. K.; Moldowan, J. M. Geochim. Cosmochim. Acta 1978, 42, 77-95. (50) Tong, H. Y.; Karasek, F. W. Anal. Chem. 1984, 56, 2129-2134. (51) Lao, R. C.; Thomas, R. S.; Monkman, J. L. J. Chromatogr, 1975, 112, 681-700. (52) Wade, T. L. Atmos. Environ. 1983, 17, 2311-2320. (53) Badger, G. M. Natl. Cancer Inst. Monogr. 1962, 9, 1-16. (54) Adams, J. D.; Lavoie, E. J.; Hoffmann, D.J. Chromatogr. Sci. 1982, 20, 274-277. (55) LaFlamme, R. E.; Hites, R. A. Geochim. Cosmochim. Acta 1978, 42, 289-303. (56) Radke, M.; Welte, D. H.; Willsch, H. Geochim. Cosmochim. Acta 1982, 46, 1-10. (57) Haynes, B. S. In Fossil Fuel Combustion: A SourceBook; Bartock, W., Sarafin, A. F., Eds.; Wiley and Sons: New York, 1991; pp 261-326. (58) Gibson, T. L.; Korsog, P. E.; Wolff, G. T. Atmos. Environ. 1986, 20, 1575-1578. (59) Valerio, F.; Lazzarotto, A. Int. 1. Environ. Anal. Chem. 1985,23, 135-151. (60) Behymer, T. D.; Hites, R.A. Environ. Sci. Technol. 1985,19,10041006. (61) Nielsen, T. Atmos. Environ. 1988, 22, 2249-2254. (62) Masclet, P.; Mouvier, G.; Nikolaou, K. Atmos. Environ. 1986,20, 439-446. (63) Interagency Regulatory Liaison Group]. Natl. Cancerlnst. 1979, 63, 241-268. (64) Grimmer, G.; Jacob, J.; Naujack, K. W. Fresenius Z. Anal. Chem. 1985, 314, 29-36. (65) Grimmer, G.; Hildebrandt, A. Zentralbl. Bakteriol. Abt. 1, Orig. B 1975, 8161, 104-124.
Received for review April 11, 1994. Revised manuscript ceived December 6, 1994. AcceptedJune 8, 1995.@
re-
ES940227S @Abstractpublished in Advance ACS Abstracts, August 1, 1995.
VOL. 29, NO. 10, 1995 I ENVIRONMENTAL SCIENCE &TECHNOLOGY rn 2483
