Aromatic Hydrocarbons in Waters of Port Phillip Bay ...

Aust. J. Mar. Freshw. Res., 1984, 35, 119-28

Aromatic Hydrocarbons in Waters of Port Phillip Bay and the Yarra River Estuary

J. David Smith and William A. MaherA Marine Chemistry Laboratory, School of Chemistry, University of Melbourne, Parkville, Vic. 3052. APresent address: Department of Inorganic and Physical Chemistry, University of Adelaide, Box 498 G.P.O., Adelaide, S.A. 5001.

Abstract Determination of aromatic hydrocarbons in coastal waters, using solvent extraction and fluorescence emission analysis, shows wide distribution of fuel oils with only a small contribution from crude oil. Oil concentrations are reported as equivalent amounts of m-terphenyl (T) or pyrene (P). Observed concentrations in Port Phillip Bay are generally 0.2-0.3 pg 1-' (T) and 0.1-0.2 pg 1-' (P), with higher values in Corio Bay and the Yarra River estuary.

Introduction Pollution of the marine environment by petroleum hydrocarbons is a widespread problem (Levy 1971; Gordon et al. 1974; Connell and Miller 1980). It is recognized that chronic, low level, hydrocarbon inputs may have long-term ecological effects, even where no evidence may exist of acute immediate detrimental effects (Levy and Walton 1973; Connell and Miller 1980). Information on the amounts and types of petroleum hydrocarbons present in the marine environment is an aid to understanding mechanisms of transport and the fates of the hydrocarbons, and in evaluation of the ecological effects of hydrocarbon inputs. Water-solubilized aromatic hydrocarbons, although comprising a small fraction of the total hydrocarbon input, can be toxic to marine organisms (Gordon and Prouse 1973; Anderson et al. 1974; Winters and Van Baalen 1975) or disrupt the behaviour of organisms (Connell and Miller 1980). The incorporation of aromatic hydrocarbons, including the polycyclic aromatic hydrocarbons (PAH), some of which are highly carcinogenic (Hoffmann and Wynder 1971), by organisms is also a potential hazard due to transfer through food chains and possible ingestion by man (Suess 1976). Polycyclic aromatic hydrocarbons are produced in a number of natural processes. Polycyclic quinone pigments are synthesized by bacteria, fungi, higher plants and some animals (Thomson 197 I), and these may be reduced in anoxic sediments to yield perylene (Wakeham 1977). The perylene would remain in the sediment and not significantly contribute to the water column. Evidence of direct biosynthesis of PAH in bacteria and plants is contradictory, but reported rates are low and would at most make a very minor contribution (Neff 1979). Some polycyclic aromatic hydrocarbons are formed naturally by high-temperature pyrolysis of organic materials, as would occur in bush fires. Low- to moderate-temperature diagenesis of sedimentary organic matter results in the formation of fossil fuels. Crude oils contain 7-35% (w/w) aromatic hydrocarbons (Gilchrist et al. 1972), and the proportion of polycyclic aromatic hydrocarbons increases with hightemperature refining. Large amounts of PAH are produced by incomplete combustion of fossil fuels (Suess 1976). 0067-1940/84/020119$02.00

J. David Smith and William A. Maher

Globally, the input to the ocean of PAH from petroleum hydrocarbons is small compared with the input from the land-based high-temperature processes (Suess 1976). Port Phillip Bay is a shipping route, with major ports on Corio Bay (Geelong) and Hobsons Bay (Melbourne), and there are oil refineries on the shore. In this environment, input of aromatic hydrocarbons from petroleum is likely to be dominant over atmospheric inputs of combustion products, with negligible contribution from biochemical processes. With the application of relatively simple analytical techniques, aromatic hydrocarbons provide a useful indicator for assessing the anthropogenic input of hydrocarbons into rivers and the sea. In this study, estuarine and seawater samples were analysed for particulate and dissolved aromatic hydrocarbons to establish the types of petroleum hydrocarbons present in Port Phillip Bay, the Yarra River estuary and Corio Bay. Two Bass Strait coastal waters were examined to provide a baseline for comparison of contamination of the other areas. Ultraviolet fluorescence spectroscopy was used to estimate the concentrations of aromatic hydrocarbons and to identify the possible sources of localized inputs.

I

300

I

I

I

350

400

450

Wavelength (nrn)

Fig. 1. Fluorescence emission spectra. (a) Lubricating oil. (b) Outboard engine fuel and diesel fuel. (c) Gippsland crude oil. ( d ) Arabian crude oil. Excitation wavelength, 300 nm.

Materials and Methods Method Development Petroleum present in the marine environment is usually of mixed origins, continually altered by weathering processes, and its composition varies widely. Fluorescence emission spectroscopy has been used previously in oil petroleum studies with the fluorescence response at a single wavelength being calibrated using solutions of crude oil in a suitable solvent (Gordon and Keizer 1974; Anon. 1976). The complex composition of petroleum means that any method of determination can yield only an approximate concentration, and where different standards are used results obtained in different laboratories are not comparable. The fluorescence emission spectra of several crude oils and refined products, including lubricating oil and diesel oil, were examined to determine whether they had any common features (Fig. 1). The results showed that there were two broad groups: (i) oils with fluorescence arising predominantly from two- and three-ringed aromatic hydrocarbons with an emission maximum near 330 nm; and (ii) oils containing aromatic hydrocarbons with three or more rings with a fluorescence emission maximum near 380 nm. It was considered appropriate to report results based on the fluorescence at 330 and 380 nm. To ensure that the reference materials used were widely available and well characterized,

Aromatic Hydrocarbons in Victorian Coastal Waters

two pure compounds were selected: m-terphenyl with a fluorescence maximum at 330 nm, and pyrene with a fluorescence maximum at 380 nm. A wide slit width on the emission monochromator was used to produce single emission peaks from the reference materials without resolving detail. Calibrations using standard solutions showed that fluorescence emission intensity was directly proportional to concentration from less than 0 . 0 2 pg ml-I to 5 pg ml-I for both m-terphenyl and pyrene in cyclohexane. Comparison of the fluorescence intensity of solutions of the standards and crude oils showed that 5 pg ml-I of No. 2 fuel oil had the same fluorescence intensity as 1 pg ml-I of m-terphenyl, and that 2 pg ml-I of Qatar crude oil had the same fluorescence intensity as 1 ~g ml-I of pyrene. These ratios reflect the proportions of fluorescent material in the oils. The use of two standards and measurement of fluorescence intensity of the sample extracts at two wavelengths is novel, and it allows simultaneous estimation of oil products based on distillate (petrol and diesel fuels), and the heavier crude oils. Chrysene has been suggested as a single reference material (Anon. 1976), but m-terphenyl is more suitable where the fluorescent material in samples is largely distillate. Dissolved and particulate fractions of the aromatic hydrocarbons are differentiated on the basis of filtration through glass-fibre filters, and the aromatic hydrocarbons are separated from the water sample by solvent extraction. Recoveries of dissolved aromatic hydrocarbons through the analytical procedure, estimated from standard additions of 1 pg and 5 pg of Tapis/Pulai crude oil or Kuwait crude oil to filtered 1-litre aliquots of a bulk seawater sample were 87 i 5% and 85 i 7%, respectively. The recovery of particulate aromatic hydrocarbons was more difficult to assess. The complete recovery of fluorescent compounds was checked indirectly by extracting several filters (Y2, Y3 and Y4) with cyclohexane in a Soxhlet apparatus after the analytical extraction with dichloromethane and m&thanol. The cyclohexane extracts were concentrated by rotary evaporation, and spectrofluorimetq revealed no additional release of fluorescent material. To determine whether aromatic hydrocarbons from the sample were lost by absorption onto the walls of the sampling bottle, the bottles were emptied and rinsed with 100 ml of dichloromethane. Concentration by evaporation and spectrofluorimetric analysis of the rinse solvent showed that the amount of material adhering to the inside walls of the sample bottles was less than 2% of the total in the sample. Compounds containing a single aromatic ring in the structure are volatile, and as a large proportion is lost in the processing of water samples, they are excluded from quantitative treatment in the present study.

Sample Collection The locations of sampling stations are shown in Fig. 2. Port Phillip Bay is a tidal marine bay with an area of about 2000 km2 and an average depth of about 20 m. The catchment land area is about 9700 km2 with a population of more than 2 . 5 x 106. The Yarra River estuary and Corio Bay have been described previously (Bagg et al. 1981). Two coastal water samples from Bass Strait were taken at locations with no evident sources of pollution. Water samples were taken at a depth of 0 . 5 m and 2 5 litres were collected in 4 min. The sampler, specially constructed to collect water at the depth required and to exclude contact with any other water, consists of a Teflon inlet valve mounted on a 2.5-litre glass bottle held in a stainless steel cradle. Analysis One-litre subsamples of water were filtered through precombusted Whatman Type F glass-fibre filters in an all-glass Millipore filtration apparatus. The hydrocarbons in the filtrate are described as 'dissolved', and the hydrocarbons retained on the filter as 'particulate'. Dissolved aromatic hydrocarbons were extracted into two 40-ml aliquots of dichloromethane in a separating funnel after adjusting the pH to 4-5 (Keizer and Gordon 1973), and the extracts were dried over anhydrous sodium sulfate. The dichloromethane was removed in a rotary evaporator and the residue dissolved in 5 ml of cyclohexane. The cyclohexane was purified by fractional distillation before use. Filters were extracted with 20 ml of dichloromethane with ultrasonic treatment for 20 min, followed by extraction in a Soxhlet apparatus with 30 ml of methanol (Burns and Smith 1980). The extracts were dried over anhydrous sodium sulfate, and the solvents removed by rotary evaporation. Both residues were dissolved in 2-ml aliquots of cyclohexane, then combined and made up to 5 ml with


cyclohexane. All extracts were stored frozen until analysed. The ~recombustedfilters gave no measurable contribution to the blank.

Port Phillip Ba Y

j

Corio Bay

Fig. 2. Locations of sample stations. (P) Port Phillip Bay. ( Y ) Yarra River estuary. (C) Corio Bay.

The aromatic hydrocarbon contents of the extracts were estimated by fluorescence emission spectroscopy using a Perkin-Elmer MPF 2A fluorescence spectrophotometer with 1-cm quartz cells.


Solutions were excited at 300 nm and the emission spectrum scanned from 310 to 390 nm. The fluorescence emission intensities at 330 and 380 nm were measured. The results are reported as concentrations equivalent to rn-terphenyl (T) and pyrene (P). Standard solutions of these compounds in cyclohexane were run with each batch of samples. The fluorescence emission intensities of impurities in the reagents were measured as the blank.

Wavelength (nm)

Fig. 3. Synchronous fluorescence emission spectra of aromatic hydrocarbon extracts of water samples from stations Y1, Y3, C1 and C4. s, 'Soluble' fraction; p, 'particulate' fraction.

To identify the probable sources of aromatic hydrocarbon inputs to these waters, synchronous scanning of excitation and emission wavelengths was used with the direct excitation mode (Lloyd 1971; Wakeham 1977). The excitation and emission wavelengths were offset by 25 nm, both slit widths set at 4 nm, and the emission spectrum from 265 to 500 nm was recorded. The synchronous fluorescence emission spectra obtained (Fig. 3) were compared with those of the reference materials


(Fig. 4) and the most likely source assigned to the material showing the closest similarity. Three major sources were taken: (i) fuel-distillates including gasoline and diesel fuel, with a synchronous fluorescence peak at about 300 nm; (ii) lube-refined products including lubricating oil, with a fluorescence peak at about 340 nm; (iii) crude-crude oils with significant fluorescence at wavelengths of 380 nm and greater.

Fig. 4. Synchronous fluorescence emission spectra of reference materials. (a) Arabian crude oil. (b) Gippsland crude oil. (c) Outboard engine fuel and diesel fuel. (d ) Lubricating oil.

-

300

400

Wavelength (nm)

Results Concentrations of aromatic hydrocarbons in the samples, expressed as equivalent concentrations of m-terphenyl and pyrene, are summarized in Table 1, and the probable sources of the fluorescent materials are listed in Table 2. Synchronous fluorescence emission spectra of the extracts of some samples are shown in Fig. 3.

Yarra River Estuary The concentrations of dissolved and particulate aromatic hydrocarbons in the Yarra River waters were an order of magnitude higher than the concentrations found in the Bass Strait samples. The highest concentrations of dissolved and particulate aromatic hydrocarbons were found in a water sample taken at the confluence of Moonee Ponds Creek and the Yarra River (stn Y3). Although this sample may represent an isolated instance of input of a pollutant, previous examination of sediment from the same area showed the presence of oil and abnormally high levels of polycyclic aromatic hydrocarbons (Bagg et al. 1981). The synchronous fluorescence emission spectrum (Fig. 3) also indicated the presence of crude oil in this sample. The synchronous fluorescence emission spectra of all other extracts of Yarra River water showed that the one- and two-ring aromatic hydrocarbons are dominant, indicating that they originate mainly from refined oil products rather than from crude oil. Corio Bay Concentrations of dissolved aromatic hydrocarbons from most of the Corio Bay samples were intermediate between the high levels found in the Yarra River estuary and the low levels in Bass Strait. The water sample (C4) taken near the oil refinery on Corio Bay contained high levels of aromatic hydrocarbons in the dissolved and particulate fractions


(Table 1). The synchronous fluorescence emission spectra indicated a dominance of refined oil products in the water. In Corio Bay, refined products form the major source of aromatic hydrocarbons, and spills of crude oil apparently contribute only a small fraction of the total fluorescent hydrocarbons.

Table 1. Aromatic hydrocarbon concentrations in the Yarra River estuary, Corio Bay, Port Phillip Bay and Bass Strait coastal waters Waters were sampled at a depth of 0 5 m. Results are expressed in equivalent concentrations of the pure calibration compounds rn-terphenyl (T) and pyrene (P) Station

Temp. ("C)

SalinityA

Aromatic hydrocarbon concn (pg I-') Dissolved Particulate (PI (TI (P) (T) Yarra River estuary 3.5 2.7 18 3.5 3.0 1.2 0 9 Corio Bay 0 95 0.55 0 4 0.5 0.85 Port Phillip Bay 0 30 0 48 0.27 0.22 0.25 0.25 0.21 0.22 Bass Strait 0.15 0.26

A

Practical salinity scale 1978 (Anon. 198 1).

Port Phillip Bay Concentrations of low molecular weight dissolved fluorescent material in the two samples from Bass Strait, and in most of the samples from Port Phillip Bay, ranged from 0.15 to 0 . 4 8 ~g 1-I m-terphenyl equivalent. Concentrations of material fluorescent at the pyrene wavelength were below the limit of detection in both dissolved and particulate phases. The concentrations of dissolved aromatic hydrocarbons were higher near the mouth of the Yarra River (Bl) and Mordiallic Creek (B2). The aromatic hydrocarbon concentrations in the Port Phillip Bay waters were too low to identify the sources from the synchronous fluorescence emission spectra. Fluorescence emission was elevated only


in the 330-nm wavelength region, indicating an increase in typical fuel-oil components but not crude oil. The major portion of the aromatic hydrocarbons in Port Phillip Bay is attributed to pollution by refined products, most probably petrol and diesel fuel oil. Table 2. Probable sources of aromatic hydrocarbons in the dissolved and particulate phases from the Yarra River estuary, Corio Bay and Port Phillip Bay All samples not listed in the table contained oils at concentrations too low to be identified. Crude, crude oil; lube, refined petroleum products including lubricating oil and residual fuel oil; fuel, distillates including gasoline and diesel fuel Station

Probable source of hydrocarbons Dissolved Particulate phase phase Yarra River estuary Fuel, lube Fuel, lube Fuel, lube, crude Fuel, lube Fuel Fuel Fuel

Lube Lube Lube, crude Unidentified Fuel, lube Fuel, lube Lube

Corio Bay Fuel, lube Fuel, lube Fuel, lube Lube Lube

Fuel Fuel Unidentified Lube Unidentified

Port Phillip Bay Fuel

Fuel

Discussion Differences in sampling techniques, analytical methods and reference standards, added to the natural variability in the composition of petroleum pollutants, make comparisons with other studies difficult. Total hydrocarbons in the C14 to C34 paraffin boiling range have been reported by Burns and Smith (1980). They used 60-80-litre water samples, each taking 3-4 h to pump through an absorption column in situ. They eluted the hydrocarbons and determined their concentrations by gas chromatography. For Corio Bay and Hobsons Bay they found hydrocarbon concentrations of 0.2-22.6 pg I-], and in general a dominance of fuel and lubricating oil residues rather than crude oils. Our use of fluorescence spectroscopy gave similar results using 1-litre water samples, and the use of pure compounds as reference standards will facilitate the comparison of the present results with those obtained in future investigations. The reference compounds m-terphenyl and pyrene are readily available and well characterized, unlike the various oils previously used as standards (Levy 1971; Keizer and Gordon 1973; Gordon et al. 1974). Although it cannot be assumed that the fluorescence emission measured is due only to aromatic hydrocarbons of petroleum origin, it is unlikely that other contributions to fluorescence emission will be significant. Other fluorescent materials, including chlorophylls and humic acids, either do not extract into dichloromethane or methanol or fluoresce at wavelengths other than those used in this study (Smart et al. 1976).


Aromatic hydrocarbon concentrations in the Yarra River and Corio Bay were elevated compared to those of Port Phillip Bay and Bass Strait. This is expected as the Yarra River and Corio Bay receive effluent from storm-water drains carrying runoff from roads, and domestic and industrial effluents. These types of effluent have been shown to be major sources of aromatic hydrocarbons in the environment (Wakeham 1977). It has been suggested that disposal of used lubricating oils is a significant source of aromatic hydrocarbons in urban waterways (Brummage 1975). This is consistent with the present results: all areas with elevated levels of aromatic hydrocarbons contained major contributions from refined oils (including lubricating oils), and the areas with aromatic hydrocarbon concentrations approaching background levels appeared to have fuel oil as the major source of contamination. The concentrations of aromatic hydrocarbons decreased with increasing distance from shore-based inputs. This trend has been observed before and removal of the aromatic hydrocarbons by association with particulate matter and sedimentation has been suggested as a possible mechanism (Bums and Smith 1980). Flocculation and deposition of suspended material in the Yarra River estuary have been reported previously (Smith and Milne 1979). Concentrations of dissolved, compared with particulate, aromatic hydrocarbons were substantially higher at most sample stations. Unsaturated hydrocarbons are relatively soluble in seawater and are expected to occur in the dissolved fraction. Saturated hydrocarbons are hydrophobic and are transported mainly in association with particulate matter (Burns and Smith 1980). Suspensions of oil in water are non-homogeneous (Levy 197I), and physical fractionation and coagulation may occur during sampling or filtration. For these reasons, the division of aromatic hydrocarbons into dissolved and particulate concentrations must be treated with caution. This study has shown that elevated concentrations of aromatic hydrocarbons are present in the Yarra River estuary, in Corio Bay and to a lesser degree in Port Phillip Bay. Ubiquitous pollution by refined oil products has occurred in these areas. The most widespread products are the light fractions, gasoline and diesel fuel, with additional inputs of the heavier lubricating or residual fuel oils near the urban and industrial areas. The contribution of crude oil appears to be small when compared with that of refined oil products. Acknowledgments We thank Melbourne University Programme in Antarctic Studies for financial support, and J. Powell, Esso-BHP for supplying reference oil samples.

References Anderson, T. W., Neff, J. M., Cox, B. A., Tatem, H. E., and Hightower, G. M. (1974). Characteristics of dispersions and water soluble extracts of crude oils and their toxicity to estuarine crustaceans and fish. Mar. Biol. 27, 75-88. Anon. (1976). Guide to operational procedures for the IGOSS pilot project on marine pollution (petroleum) monitoring. UNESCO, Manuals and Guides No. 7. Anon. (1981). Tenth Report of the Joint Panel on Oceanographic Tables and Standards. UNESCO Technical Papers in Marine Science No. 36. Bagg, J., Smith, J. D., and Maher, W. A. (1981). Distribution of polycyclic hydrocarbons in sediments from estuaries of south-eastern Australia. Aust. J. Mar. Freshw. Res. 32, 65-83. Brummage, K. G. (1975). The sources of oil entering the sea. In 'Background Papers for a Workshop on Inputs, Fates and Effects of Petroleum in the Marine Environment'. pp. 1-6. (National Academy of Science: Washington D.C.) Burns, K. A., and Smith, J. L. (1980). Hydrocarbons in Victorian coastal waters. Aust. J. Mar. Freshw. Res. 31, 251-6.


Connell, D. W., and Miller, G. J. (1980). Petroleum hydrocarbons in aquatic ecosystems-behaviour and effects of sublethal concentrations, Part 1. C.R.C. Critical Reviews in Environmental Control, No. 11, pp. 37-104. Gilchrist, C. A,, Lynes, A., Steel, G., and Whitham, B. T. (1972). The determination of polycyclic aromatic hydrocarbons in mineral oils by thin layer chromatography and mass spectrometry. Analyst 97, 880-8. Gordon, D. C., and Keizer, P. D. (1974). Estimation of petroleum hydrocarbons in seawater by fluorescence spectroscopy: improved sampling and analytical methods. Fish. Res. Board Can., Tech. Rep. No. 481. Gordon, D. C., Keizer, P. D., and Dale, J. (1974). Estimates using fluorescence spectroscopy of the present state of petroleum hydrocarbon contamination in the water column of the northwest Atlantic Ocean. Mar. Chem. 2, 251-61. Gordon, D. C., and Prouse, N. J. (1973). The effects of three oils on marine phytoplankton photosynthesis. Mar. Biol. 22, 329-33. Hoffmann, D., and Wynder, E. L. (1971). Respiratory carcinogens: their nature and precursors. Proceeding of the International Symposium on Identification and Measurement of Environmental Pollutants, pp. 9-16. (Natl Res. Counc. Can.: Ottawa.) Keizer, P. D., and Gordon, D. C. (1973). Detection of trace amounts of oil in sea water by fluorescence spectroscopy. J. Fish. Res. Board Can. 30, 1039-46. Levy, E. M. (1971). The presence of petroleum residues off the east coast of Nova Scotia, in the Gulf of St Lawrence and the St Lawrence River. Water Res. 5, 723-33. Levy, E. M., and Walton, A. (1973). Dispersed and particulate petroleum residues in the Gulf of St Lawrence. J. Fish. Res. Board Can. 30, 261-7. Lloyd, J. B. F. (1971). Synchronised excitation of fluorescence emission spectra. Nature (Lond.) 231, 64-5. NeK J. M. (1979). 'Polycyclic Aromatic Hydrocarbons in the Aquatic Environment.' (Applied Science Publishers: London.) Smart, P. L., Finlayson, B. L., Rylands, W. D., and Ball, C. M. (1976). The relation of fluorescence to dissolved organic carbon in surface waters. Water Rex 10, 805-1 1. Smith, J. D., and Milne, P. J. (1979). Determination of iron in suspended matter and sediment of the Yarra River estuary, and the distribution of copper, lead, zinc and manganese in the sediments. Aust. J. Mar. Freshw. Res. 30, 731-9. Suess, M. J. (1976). The environmental load and cycle of polycyclic aromatic hydrocarbons. Sci. Total Environ. 6 , 25 1-7. Thomson, R. H. (1971). 'Naturally Occurring Quinones.' 2nd Edn. (Buttenvorths Scientific Publ.: London.) Wakeham, S. G. (1977). Synchronous fluorescence spectroscopy and its application to indigenous and petroleum-derived hydrocarbons in lacustrine sediments. Environ. Sci. Technol. 11, 272-6. Winters, K., and Van Baalen, C. (1975). The effects of a No. 2 fuel oil and two oils on the growth and photosynthesis of microalgae. Mar. Biol. 28, 87-94.

Manuscript received 26 April 1983, accepted 12 October 1983