(9) Metcalfe, L. D.; Schmitz, A. A. Anal. Chem. 1961,33, 363. (10) Morrison, W. R.; Smith, L. M. J . Lipid Res. 1964,5, 600. (11) Farrington, J. W.; Henricks, S. M.; Anderson, R. Geochim. Cosmochim. Acta 1977,41, 289. (12) MacKenzie, M. J.; Hunter, J. V. Enuiron. Sci. Technol. 1979, 13, 179. (13) Zurcher, F.; Thuer, M.; Davis, J. A. “Importance of Particulate Matter on the Load of Hydrocarbons of Motorway Runoff and Secondary Effluents”; in Proceedings of the International Symposium on the AnalysiS of Hydrocarbons and Halogenated Hydrocarbons, Ontario, Canada, May 1978. (14) Van Vleet, E. S.; Quinn, J. G. Enuiron. Sci. Technol. 1977,11, 1086. (15) Hunter, J. V.; Sabatino, T.; Gomperts, R.; Mackenzie, M. J. J . Water Pollut. Control Fed. 1979,51, 2129. (16) Boylan, D. B.; Tripp, B. W. Nature (London)1971,230,44. (17) Zurcher, F.; Thuer, M. Enuiron. Sci. Technol. 1978,12, 838. (18) Winters, K.; Parker, P. L. API Publ. 1977,4284, 579-81. (19) Pierce, R. H.; Olney, C . E.; Felbeck, G. T. Geochim. Cosmochim. Acta 1974,38, 1061. (20) Meyers, P. A.; Quinn, J. G. Nature (London)1973,244, 23. (21) Button, D. K. Geochim. Cosmochim. Acta 1976,40,435. (22) Herbes, S.E. Water Res. 1977,11, 493. (23) Karickhoff, S. W.; Brown, D. S.; and Scott, T. A. Water Res. 1979,13, 241. (24) Thompson, S.; Eglinton, G. Geochim. Cosmochim. Acta 1978, 42, 199. (25) Sheldon, R. W. Limnol. Oceanogr. 1972,17, 494. (26) Stokes, V. K.; Harvey, A.C. Proc. Jt. Conf. Preu. Control Oil Spills, 1973 1973,457-65.
tribute significantly greater quantities of these materials to the ocean on a per capita basis than do nonurban rivers. This probably reflects the higher usage and surface deposition of petroleum in the urban locale. Acknowledgment We gratefully acknowledge Shan-Tan Lu, who provided organic-carbon analyses, and B. R. T. Simoneit for his helpful comments and suggestions. Literature Cited (1) National Academy of Sciences “Petroleum in the Marine Environment”; Washington, D.C., 1975. (2) “The Ecology of the Southern California Bight: Implications for Water Quality Management”; Southern California Coastal Water Research Project; Technical Report 104, March 1973. (3) Reed, W. E.; Kaplan, I. R.; Sandstrom, M.; Mankiewicz, P. API Publ. 1977,4284, 183-8. (4) Peters, K. E.; Sweeney, R. E.; Kaplan, I. R. Limnol. Oceanogr. 1978,23, 598. (5) Eganhouse, R. P.; Simoneit, B. R. T.; Kaplan, I. R. Enuiron. Sci. Technol., following paper in this issue. (6) Eganhouse, R. P.; Lu, S.-T.;Kaplan, I. R. “Elemental and Isotopic ComDosition of Particulates Carried in Los Angeles Stormwaters”, unpGblished. (7) Los Angeles County Flood Control District “Hydrologic Report 1974-75’1 174, October 1, 1976. (8) “Standard Methods for the Examination of Water and Wastewater”, 13th ed; American Public Health Association: New York, 1971; p 537.
Received for review May 19, 1980. Accepted November 6,1980. Financial support was provided by the Department of Energy and the Bureau o f Land Management (Contract No. EY-76-3-03-0034).
Extractable Organic Matter in Urban Stormwater Runoff. 2. Molecular Characterization? Robert P. Eganhouse,” Bernd R. T. Simoneit, and Isaac R. Kaplan Department of Earth and Space Sciences and Institute of Geophysics and Planetary Physics, University of California,
Los Angeles, California 90024
A comprehensive molecular characterization of solventextractable and bound organic constituents of Los Angeles stormwater runoff has been conducted. The four classes of compounds inspected were hydrocarbons, fatty acids, ketones, and polar compounds. On the basis of molecular distributions and abundances, the extractable constituents appear to be largely anthropogenic. This is attributed to the dominance of petroleum residues on street surfaces and to their removal during the storm. The bound fraction comprises mainly microbial and higher plant debris derived from the original structural matrices of the source material or present in association with humic/fulvic acids and mineral phases. Introduction Urban storm drainage is one means by which natural terrigenous and anthropogenic organic matter is transported from land to the ocean. Whereas some efforts have been made to characterize stormwater-borne hydrocarbons ( I - 3 ) , no details have yet been reported for other molecular species which may constitute up to 40% of the solvent-extractable organics ( 4 ) .The identification and quantitative assessment Publication No. 2062, Institute of Geophysics and Planetary Physics, University of California, Los Angeles, CA 90024.
of runoff-transported organic matter in nearshore marine sediments has not been possible owing to a lack of data. Consequently, it is unclear whether a threat is posed to local marine life by these inputs. Although the ability to completely control pollution caused by urban runoff is out of reach, a comparison of its chemical properties with controllable inputs such as municipal/industrial waste disposal and chemical dumping is not only important but attainable. This study, performed in Southern California, attempts to provide new information toward this end. A comprehensive analysis was undertaken on solvent-extractable and bound organic compounds isolated from waters of the Los Angeles River during a storm sequence in the fall of 1978. Our objectives were to characterize the major components, examine specific molecular distributional features, assess the approximate anthropogenic contribution to the overall input, and search for marker compounds to be used in source identification. Experimental Section Samples of Los Angeles River storm runoff were collected at 11intervals during a storm event on November 21,1978 ( 4 ) . Before extraction, the samples were spiked with two recovery standards: triisopropylbenzene (TIB) and n-nonadecanoic acid (NDA). Three basic sample types were prepared: (1)
0013-936X/81/0915-0315$01.25/0 @ 1981 American Chemical Society
Volume 15, Number 3, March 1981 315
unfiltered samples, extracted directly without physical or chemical pretreatment, (2) filtered samples, i.e., runoff samples filtered through Whatman GF/A glass-fiber filters after which the filtrates were extracted, and (3) particulates, isolated by gravitational settling. Details of the sampling, the extraction, and the chromatographic separation of organic compounds into various classes are described in the preceding report ( 4 ) . Briefly, the CHCls-extractable organics from filtered and unfiltered samples were separated into four fractions for molecular analysis: (1)total hydrocarbons (THC), (2) fatty acids (FA), (3) ketones (KET), and (4) polar compounds (PLR). The particulates were extracted exhaustively with CHC13 to remove extractable organics and then saponified (5) to isolate the bound constituents. After methylation of the fatty acids with BF3-MeOH, the bound organics were separated into four fractions as previously described ( 4 ) .Bound fractions are hereafter identified by the prefix "B" before the fraction acronym. For example, BTHC refers to bound total hydrocarbons, whereas ETHC refers to (solvent) extractable total hydrocarbons. All fractions were examined by high-resolution glass capillary gas chromatography using Carlo Erba FTV 2150 and 2350 instruments equipped with split injectors of the Grob design (6)and flame ionization detectors. Capillary columns (15 m) wall-coated with either OV-101 or SE-54 (0.25-mm i.d.)
were used for diagnostic and quantitative purposes, while a 30-m SE-54 column of the same specifications was used for gas-chromatographic/mass-spectrometricanalysis. Total hydrocarbon, FA, and KET fractions were analyzed by splitless injection (6), temperature programming the column 40-260 is0 "C a t 4 "C/min. The PLR fraction was silylated by using BSA (N,O-bis(trimethylsily1)acetamide) before injection (7). Analyses of the PLR fraction were performed in two ways: (1)direct injection of the sample in excess BSA a t 150 "C with temperature programming to 260,,, "C a t 4 "C/min and (2) transfer to 9:l hexane/ethyl acetate under dry N2 followed by injection a t 90 "C with temperature programming as before. The former method was used for examination of the sterols, and the latter for intermediate molecular weight species. Tentative compound identifications are based on gaschromatographic retention times and/or by analyses on a Finnigan Model 4000 quadrupole mass spectrometer interfaced with a Finnigan Model 9610 gas chromatograph. Spectrqmetric data were processed with a Finnigan INCOS Model 2300 data system. Quantitative results for THC and FA fractions were achieved by comparing samples with external standards. These standards contained compounds distributed over the molecular-weight range of interest and were run the same day as the samples. Repetitive extractions showed that the isola-
Table 1. General Characteristics of Total Hydrocarbon Fractions in Storm Runoff Samples sampling time
THC, a mg/L
En-alkanes, pg/L
UCM,
maxima
max n-alkane
mean OEP
(%-alkanes
> n-C24)/THC
Prln-Cl7
PrlPh
Unfiltered Samples 1000 1100 1200 1300 1400 1450
2.0 6.2 5.1 4.0 10 19 18
68 110 200 110 220 820
1600 1700
20 10 9.2
360 360 220 210
1800
16
240
1500 1550
1.o 1.7 1.4 1.2 1.4 1.3 1.2
0.1 0.7 0.3 0.5 0.9 0.5 0.5
c29
1.2 1.2 1.4
c29
1.4
0.5 0.6 0.7 0.7
1750/2800
c17
2750/1700 1750/2800 2800/ 1750
c29
2800/1900 2650/1550 2600 2800 2550 2550
c29
2700
c17
cis CIS Cl8 c17 c17
1.6 1.3 1.4 1.4 1.3 1.2 0.8 0.7 0.7 0.8 0.6
2.1 2.6 2.2 2.2 1.7
1.7 1.2 1.o 1.7 1.3 1.o 0.8
2.1 1.8 1.6 2.0 1.7 2.7 1.6
0.8 0.6 0.7 0.6
1.4 1.5
1.2 1.2 1.8 1.3 1.3
3.9 1.5 1.5 1.4 1.6 1.4
Filtered Samples
1400 1450 1500
0.42 0.38 0.36 0.41 0.50 0.37 0.39
18 12 20 10 9.2 23 12
1550 1600 1700 1800
0.40 0.36 0.43 0.30
6.4 5.3 4.0 2.3
1000 1100 1200 1300
1800/2850 2750/1750 1750/2750 2800/ 1750 2800
cis
2300/ 1650 2600 2800 2550 2500 2600
CIS
1.o
c17
1.o 1.o 0.9 1.o 1.2 1.o 1.o 1.1
c17
1.1
c23
1.o
1.5 1.3 0.8 0.5 0.7 1.o 0.8 0.3 0.5 0.4 0.4
1.4 1.0 1.2 1.3 1.4
0.1
1.8
(0.1 1.6 0.4 1.2
0.8
c23
cis CIS c23
CIS CIS
1.5
1.2
Particulate e
a
1000 1200 1450 1500
0.08 0.06 0.26 0.28
1700
0.68
0.29 0.03 9.9 2.7 8.4
Total hydrocarbons, measured gravimetrically.
2250 f 2300 2350 2600
c23 c23 c23 c23 c29
0.5 0.4
0.9
Maxima of the unresolved complex maxima given in Kovats indexes; if bimodal, numerator refers to index
of the dominant mode. Mean OEP is average of all running OEP values (8). Values are given as percentages. * "Bound" hydrocarbonsin particulates (BTHC). 'No data obtained because of insufficient samples.
316
Environmental Science & Technology
tion scheme used here was effective in removing 95+% of the extractable organic material. Because we were interested primarily in qualitative features, no attempt was made to correct values for either extraction or recovery efficiencies, the latter of which were found to be 96% for NDA and 68%for TIB.
Results Hydrocarbons. Table I summarizes general characteristics of the hydrocarbons in unfiltered and filtered samples and for bound hydrocarbons in particulates. All samples exhibited a broad envelope of unresolved components ranging approximately from n-Cl3 to n-C36+. Resolved components surmounting this envelope constituted, on the average, less than 17% of the total integrated area. The hydrocarbon distributions are characteristically of two types (Figure 1).Unfiltered and filtered samples collected up to, and including, 1450 hours (Figure 1A) had bimodally distributed, unresolved complex mixtures (UCM) with maxima occurring a t n-Ci7-19 and n-C26-zs. After 1450 hours, a unimodal UCM with a
maximum in the range of n-C23-28 was observed (Figure 1B). Bound hydrocarbons, examined a t several sampling times (e.g., Figure 1C) had highly symmetrical, unimodal UCM's with maxima a t n-Czz-26. When normal alkane distributions of unfiltered samples are plotted against time during progress of the storm, several features emerge (Figure 2). First, the levels of the total normal hydrocarbons ( Z n-alkanes) are highly variable and correspond roughly to fluctuations in the total hydrocarbon burden (cf. Table I). These variations do not show strong correlations with gross storm parameters such as flow or suspended solids ( 4 ) . Second, the normal alkanes in the range of n-C13-24 maximize a t n-C17,1s and demonstrate a smooth distribution as evidenced by values of odd-even predominance, OEP (8), near unity. Finally, excluding samples taken at 1000and 1200 hours, n-alkanes greater than n-C24 show strong odd-even predominance (OEP > 1)and maximize a t n-C29. The distribution of normal hydrocarbons in filtered samples is similar to that in unfiltered samples; however, the mean OEP values are consistently lower (Table I). Also, with the exception of
c.
A 1450 hours ETHC
._._ -- ...-
7
i
B 1550 hours ET%
C 1450 hours BTHC
do"
dd
160"
110"
140"
1BO0
1BO0
2td
2'm0
2kf
2h*
Isothermal
-
Figure 1. Gas chromatograms of total hydrocarbon fractions in storm runoff: (A) 1450 hours (unfilteredsample): (B) 1550 hours (unfiltered sample);
(C)1450 hours (particulates). Peak identities are listed in Table II. Volume 15, Number 3, March 1981
317
2SI
I' 1.0
05
.. . . . . * 16
"
18
'
I
'
' "
"
20 22 24 CARBON NO
"
26
' '
28
30
Figure 2. Distribution for normal hydrocarbons in unfiltered samples of storm runoff during the course of the storm along with running oddeven predominance values for the sample taken at 1450 hours.
the sample taken at 1000 hours, filtered samples contain reduced amounts of high molecular weight n-alkanes (relative to total n-alkanes; cf. Figure 3). Bound hydrocarbons represent only a small fraction (n-C24 hydrocarbons showed an enrichment. Besides the normal alkanes, an assortment of branched, unsaturated, and cyclic compounds were identified (Figure IA--C, Table 11).Among the branched species, a homologous series of isoprenoids, C14-21 (excluding C17), were prominent. Also, with the exception of the C15 homologue, c13-19 isoalkanes were tentatively identified. Olefins were found in minor abundance; however, a number of polynuclear aromatic hydrocarbons were identified in both ETHC and BTHC fractions, including several homologous series. Among these were naphthalene plus C1-5 homologues, biphenyl plus C1-4, and phenanthrene/anthracene plus Cl-2 homologues. Fluoranthene, pyrene, chrysene, xanthene, and benzopyrene were also tentatively identified. No attempt was made to refine the separation of aromatic compounds; therefore, the present list is undoubtedly incomplete. As noted previously ( 4 ) aromatic compounds were enriched in filtered samples relative to unfiltered samples. Benzothiophene and dibenzothiophene (plus alkylated homologues), found abundantly in Philadelphia storm waters ( 2 ) ,were detected only a t trace levels here. A series of alkyl cyclohexanes with side chains ranging in length from c8 to C15 (ETHC) and from C11 to Cis (BTHC) were identified as major components. Other cyclic compounds included multiple homologous series of steranes, diterpanes, and triterpanes. The steranes found in all samples were complex; however, no sterenes were detected. Among the 318
Environmental Science & Technology
prominent constituents were the c27-29 5a-steranes with lesser amounts of methyl steranes (c28-30). Extended tricyclic diterpanes (C20-29) were also verified, where homologues from C20 to C24 have the inferred structure I (Figure 4), and those from Czs to C2g exist as diasteromeric pairs due to a methyl branch in the side chain. Triterpanes were dominated by the 17a(H)-hopane series; neither triterpenes nor the 17p(H) compounds were found. The 17ct(H) series consists of 17a(H)-trisnorhopane (11, R = H ) , 17a(H)-norhopane (11, R = CgH5), 17a(H)-hopane(11, R = C3H7),and an extended hopane series present as resolvable diastereomeric pairs in near 1:l abundance, (22R, 22s) ranging from C31 to C35 (111).A C28 triterpane identified as 17a(H),18a(H),21p(H)-28,30-bisnorhopane was also found (IV). Fatty Acids. In general, the extractable normal fatty acids were bimodally distributed with maxima a t 18:O and 24:O or 26:O (Table 111). Strong even-odd carbon-number predominance was observed in all cases. Palmitic acid (16:O) was the dominant species; however, the order of importance among the other major components (18:1,18:0, and 16:l) was highly variable. Branched fatty acids including iso-14:0-17:0, anteiso-15:0, -17:0, -19:0, and one isoprenoid, 4,8,12-trimethyltridecanoic acid, were detected in EFA fractions. Dehydroabietic acid (V) was found only in minor amounts. Various phthalates and adipates observed in all samples were present only in small quantities in procedural blanks. No triterpenoidal acids were detected. On the basis of gravimetric data, the bound fatty acids constituted from 23 to 50%of the total acids. This amount is similar to that found by Farrington et al. (9) and Cranwell ( I O ) , who studied both marine and freshwater sediments, but is lower than that observed for Mississippi River waters (11). The general distribution features of bound and extractable fatty acids are similar to each other. No systematic differences in 16:1/16:0 or 18:1/18:0 ratios are apparent, but measurable differences include the 16:0/18:0 ratio (Table 111), which is invariably higher for bound fatty acids, the ratio of even to odd carbon n-alkanoic acids, which is lower for the BFA fraction, and the percent of singly branched (Le., is0 and anteiso) acids which is 4-8 times higher in the bound fraction than for the corresponding extractable acids (cf. Table IV). The only isoprenoid identified in BFA fractions was 5,9,13-trimethyltetradecanoic acid. Diterpenoidal and triterpenoidal acids were not detected. Ketones. The ketone fractions were composed of relatively simple mixtures of compounds. In EKET fractions, a bimodal series of n-alkan-2-ones ranging from Cl2 to C27, maximizing a t C17 and C25 and having a mean OEP of 1.3, was found only in trace amounts. The predominant compound was the isotwo isomers of prenoid 6,10,14-trimethylpentadecan-2-one; phytenic acid y-lactoyes were also found in abundance (structure VI). A number of ketoaromatics including diphenylmethanone, fluorenone, anthracenone, phenanthrenone, anthracenedione, phenanthrenedione (plus a C1 homologue), and xanthenone were tentatively identified. Two ketosteroids were also found and a C2H5 hoin minor amounts: cholesta-3,5-diene-7-one mologue (structure VII, strong peak a t m/z 174). The BKET fractions also contained a bimodal n-alkan2-one series extending from Cl0 to C29 with a mean OEP value of 4.9, and maxima at C17 and CZ5.Again, the dominant ketone was the CIS isoprenoid; however, only trace amounts of as yet unidentified ketoaromatics were detected, and these did not elute in the same regions as aromatics found in EKET fractions. In addition, the y-lactone isomers of phytenic acid, found abundantly in the EKET fractions, were completely absent. Ketosteroids in the BKET fraction included a series of C27-29 homologues of cholesta-3,5-dien-7-one structure and C28,29 triene homologue compounds (VII). Finally, a series of
SOLVENT-EXTRACTABLE Hydrocarbons Unfiltered b
5
Filtered
TI C
0.8
-
a
a
d
a,
>
. I
-a,0
t
13 1
”
d
(.
.i1.,,,
a, 0
3
C
I
d
J
1s
Bound Hydrocarbons
K
la
Ib 1
1
Carbon Number
d
le
.
e
b
Figure 3. Normal alkane distribution plots for selected unfiltered, filtered, and particulate (bound) samples of storm runoff: (a) 1000, (b) 1200, (c) 1450, (d) 1500, and (e) 1700 hours. Concentrations are normalized to that of most abundant n-alkane.
methoxy steroids (c27-29, structure VIII) coeluted with the ketone band. These were identified by comparison with mass spectra of Idler et al. (12). The parent compound was 3methoxycholesta-5-ene; lesser amounts of C1- and C2-substituted 3-methoxycholesta-5,22-diene homologues were also found. In the case of both keto- and methoxysteroids, the order of predominance was C2g > C27 > C28 (steroid skeleton only). Polar Compounds. The polar fraction represents a rather complex assemblage of molecular types. Gas chromatograms of the derivatized EPLR and BPLR fractions (1450 hours) show that the majority of the polar compounds are incorporated in a broad UCM extending from CZO.In contrast, branched species, tentatively identified as is0 and anteiso C15 and C17 compounds, are principally in the form of the 0isomer; the remaining members of the series have various isomeric compositions. The major constituents are Cl2, c14, “iso-C15”, c16, CIS, and c24. The w-hydroxy acids detected in BPLR samples are c20-26 even-carbon,homologues. Except for C20, which was present in small amounts, their relative abundances are roughly equivalent (Table V). The a,w-dicarboxylicacids were found in the range of c9-26; however, in some samples the (29-19 compounds were comVolume 15, Number 3, March 1981
319
Table 11. Selected Compounds Identified in Los Angeles River Stormwaters a peak no.
campd
peak no.
compd
Hydrocarbons (Figure 1) branched 2 3 8 9 12 17
2-methyldodecane ( iso-CI3)
1
2,6,1O-trimethylundecane (isopr-Ci4)
4
2-methyltridecane ( iso-Ci4) 2,6,1O-trimethyldodecane (isopr-CI5) 2,6, IO-trimethyltridecane (isopr-C16)
5 6 7
24
2,6,10-trimethylpentadecane (isopr-Cis) 2,6,10,14-tetramethylheptadecane(isopr-CZi)
10 13
11 18 20 23 28
cyclics n-octylcyclohexane (C14) n-decylcyclohexane (C16) n-undecylcyclohexane (C17) n-dodecylcyclohexane (Cis) n-tetradecylcyclohexane (Cz0)
14 15 16 19 21 22
30
n-pentadecylcyclohexane (C2,)
25
31 32
C23H42 extended diterpane (structure I, R = C6HI3) C24H44 extended diterpane (structure I, R = C7H15)
26 27
C25H46 extended diterpane (structure I, R = CaH17)
29
33 34 35
aromatics naphthalene 1-methylnaphthalene 2-methylnaphthalene triisopropylbenzene (internal standard) biphenyl Cz naphthalenes C1 biphenyls CB naphthalenes Cp biphenyls C4 naphthalenes Cs naphthalenes C3 biphenyls phenanthrene or anthracene C i phenanthrenelanthracenes Cp phenanthrenelanthracenes pyrene fluoranthene
extended diterpane (structure I, R = C9HZ0) C28H52 extended diterpane (structure I , R =
84C & ip
c 1iH23) 36 37
17a(H),21p(~)-30-norhopane(structure 11, R = C2H5) 170!(H),21p(H)-hOpane (structure 11, R = C3H7) Polar Compounds (Figure 4)
1 3 4 5 7 8 9 10 11 12 13 14 15 16 19 20 22 24 26 27 29 30 32 33 35 36 38 40 45 a
hydroxy acids alp-hydroxydecanoic acid (n-Clo)c alp-hydroxyundecanoic acid (n-Ci1)
1 6
p-hydroxydodecenoic acid (Clp.,) alp-hydroxydodecanoic acid (n-Clp)
17 21
p-hydroxytridecanoic acid (n-Cls) alp-hydroxytetradecanoic acid (%C14)
23 25
P-hydroxypentadecanoic acid ("iso-Ci5") alp-hydroxypentadecanoic acid ("aCCI5") alp-hydroxypentadecanoic acid (n-Cl5) p-hydroxyhexadecanoic acid (" !s&l!j") alp-hydroxyhexadecanoic acid (n-CI6)
31 34 37 42 44
6-hydroxyheptadecanoic acid ("iso-Ci7") p-hydroxyheptadecanoic acid ("aCC1,") alp-hydroxyheptadecanoic acid (n-CI7) alp-hydroxyoctadecanoic acid (n-Cls) a/p-hydroxynonadecanoic acid (n-Cj9)
2 18 39
alp-hydroxyeicosanoic acid (n-Czo) a-hydroxyheneicosanoic acid ( ~ - C Z ~ ) ~
41 43
w-hydroxyeicosanoic acid (n-Cno)
46
a,w-dicarboxylic acids a,o-nonanedioic acid (n-Cg)e a,o-undecanedioic acid (rrCll)e a,w-hexadecanedioic acid (*Cia) a,w-octadecanedioic acid (n-CIs) a,@-nonadecanedioic acid (n-CI9) a,o-eicosanedioic acid (rrCz0) a,@-docosanedioic acid (n-Cpp) a,w-tricosanedioic acid (n-Cp3) a,w-tetracosanedioic acid (n-Cn4) a,w-pentacosanedioic acid (n-Cp5) a,&-hexacosanedioic acid (n&) sterols and others diethyl phthalate 1,2,3-propanetrioic acid, 2-(acetyloxy)-, tributylester 5p-cholestan-3p-ol (coprostanol, structure IX) cholesta-5-ene-3p-ol (cholesterol) 24-methylcholesta-5-ene-3~-ol(campesterol) 24-ethylcholesta-5-ene-3p-ol @-sitosterol)
a-hydroxydocosanoic acid (n-Cpp) w-hydroxyheneicosanoic acid (n-Cp1) a-hydroxytricosanoic acid (n-C23) w-hydroxydocosanoic acid (n-Cpp) a-hydroxytetracosanoic acid (n-Cp4) w-hydroxytricosanoic acid (fbC23) a-hydroxypentacosanoic acid (n-Cp.5) w-hydroxytetracosanoic acid (fiCp4) a-hydroxyhexacosanoic acid (n-C26) w-hydroxyhexacosanoic acid (n-cpe)
Major components and significant constituents are listed: identifications based on mass-spectral interpretation. Normal and branched hydroxy acids listed; Small amount of a-hydroxy acid isomer present. Small amount
"iso-" and "ai-" identifications based upon relative retention times and mass-spectral data. of @-hydroxyacid isomer present. e Trace amount coeluting with another peak.
320 Environmental Science & Technology
II
I Extended diterpanes
Hopanes
m
Extended hopanes
R = H,CH3 ,Cz H,
R=H,CH,,C,H,
pm
Cholesta-3,5-dien=l-one series
3-methoxy-cnaleota-5-ene series
no
E Coprostanol
X Eplcoprostanol
no
li m.Cholestanol
Figure 4. Structures of organic compounds identified in storm runoff.
pletely absent, whereas in others they constituted 99+% of the total a p d i a c i d s (Table V). Compounds greater than CIS showed strong even-odd predominance and maximized at Czz. Unsaturated species were not observed. Bound sterols included cholesterol, campesterol, and p-sitosterol in the order of predominance C29 > C27 > CZS.Cholestanol was the only stanol observed. As in the case of EPLR fractions, phytol, dihydrophytol, and triterpenoidal alcohols were not detected. Discussion
Hydrocarbons. Hydrocarbons constitute roughly 60% of the total solvent-extractable organics in Los Angeles River stormwaters ( 4 ) . The vast majority of these hydrocarbons (-94%) are associated with particulate matter and are primarily derived from petroleum residues. Molecular evidence presented here in support of this latter assertion includes the following: (1) a broad envelope of unresolved species extending from I n - C l 3 to n-C36+ and comprising >80% of the total hydrocarbons (14),(2) a homologous series of normal alkanes (n-CI3-24) with OEP 1.0 ( 1 4 ) , (3) abundant branched homologues including isoprenoids, iso- and anteisoalkanes, (15,16),(4) multiple homologous series of alicyclic and polycyclic compounds such as the alkyl cyclohexanes, steranes, diterpanes, and triterpanes, and ( 5 ) a variety of parent polynuclear aromatic compounds in association with alkyl-substituted homologue assemblages. In addition, the absence of 17@(H)-hopaneisomers and the distribution of the l7a(H)-hopane series >C30 (i.e., 22R and 22s diastereomeric pairs occur in near 1:1 abundance) point
to the ancient character of these hydrocarbons (17).An indication of the influence of locally producedlconsumed petroleum was found in the molecule l?a(H),18a(H),21p(H)28,30-bisnorhopane, identified as a major terpenoid constituent of the Monterey shale off Santa Barbara, CA, and in California crude oils (18).More recently, it has been found in marine sediments and sediment-trap particulates collected in San Pedro Basin, located offshore from the mouth of the Los Angeles River (19, 20). Biogenic hydrocarbons, in minor amounts, were evidenced by the high molecular weight n-alkanes (>n-C24) with OEP values >LO. These are presumably derived from higher plant epicuticular waxes (21)and in no case exceed 1.6%of the total hydrocarbons (Table I). However, because bacteria have been known to display little or no carbon preference in metabolically synthesized n-alkanes (22, 23), we cannot exclude the possibility that minor amounts of bacterial hydrocarbons might also be present in stormwater runoff. On the basis of the aforementioned assemblage of characteristics, we conclude that petroleum, not recent biogenic, hydrocarbons predominate. A storm event not only removes but effectively homogenizes a diverse set of organic source materials ( 4 ) . Distributions observed in the laboratory, then, must represent a composite of multiple inputs. Among the probable sources of hydrocarbons in the Los Angeles River basin are (1)vehicular exhaust particles, (2) lubricating oils, (3) atmospheric fallout (rain and dry, e.g., forest fires, combustion of fossil fuels, and eolian transport of bioorganics), (4) fuel oils, ( 5 ) spillage of crude and refined petroleum products during production, processing, or transportation, (6) leached/eroded pavement, (7) natural biogenic sources on land, (8) erosion of organic-bearing sedimentary rocks, and (9) others. The difficulty in defining the composite arises from both the complexity of this input array and the possibility of postdepositional alterations (24, 25). On the basis of GC, IR, and lead data, Zurcher et al. ( 1 ) suggested that automobile exhaust particulates were the primary contributors to hydrocarbon burdens in Swiss motorway runoff. These particulates contain hydrocarbons distributed essentially as a UCM from n-Czn to n-C34+ and maximizing a t n-Cz9 (26).We found a similar pattern in our samples, particularly during the later stages of the storm; however, n-alkanes, not generated to any great extent in combustion experiments ( 2 6 ) ,are quite abundant in stormwater samples. Dewaxed lubricating oils and transmission fluids consist essentially of high molecular weight UCMs with no detectable normal alkanes (27). Thus, the high molecular weight mode of the UCM which dominated the hydrocarbons in samples collected during later stages of the storm may have its origin in any combination of these materials. The abundance of normal and isoalkanes and the relatively low UCM suggest that extensive biodegradation has not occurred, as it is generally agreed that microbial utilization of petroleum proceeds approximately in the following sequence: normals, branched, cyclics, and aromatics (14, 24). Normal and perhaps branched alkanes should have been greatly reduced or eliminated if intense microbial breakdown had taken place, leaving a strongly pronounced UCM. This type of distribution is common in petroleum-contaminated river and marine sediments (28-30) and biodegraded crude oils (31). The data also suggest that a minimum of physical weathering has occurred. Evaporative losses may be prevented by association of liquid hydrocarbons with particulate matter (32) or microencapsulation onhn pavement surfaces (e.g., solution of crankcase drippings into asphalt), Attention should be drawn to a consistent change in the hydrocarbons that occurred during the storm. Samples collected up to and including 1450 hours exhibited bimodal Volume 15, Number 3, March 1981
321
Table 111. Total Extractable Fatty Acids in Unfiltered Samples and Bound Fatty Acids in Particulates of Storm Runoff a 1000 compd
EFA
lo:o 1 l:o 12:o 13:O 14:O 150 16:O 17:O 18:O 19:o 20:o 21:o 22:o 23:O 24:O 250 26:O 27:O 28:O 29:O 30:O 16:l 18:l 17:l 19:lb i-13:0 ai-13:O i-14:0 C15:O ai-15:O i-16:0 i-17:O ai-17:O 16:0/18:0
1200 BFA
0.6
