September, 1980 POLLUTANTS I N THE TROPOSPHERE AND NATURAL WATERS 375 Different Forms of Phosphorus in Freshwater R. J. Stevens Department of Agriculture, Freshwater Biological Investigation Unit, Greenmount Road, Muckamore, Antrim, BT41 4PX, Northern Ireland Phosphorus Chemistry in Freshwater Phosphorus has been identified as a key factor governing the growth of algae in Lough Neagh. Like carbon, phosphorus can exist in freshwater in a vast number and variety of molecular structures. The dominant phosphorus compounds will contain phosphorus with a co- ordination number of 4. Chief among these are the phosphates, in which phosphorus exists in oxidation state +5. Compounds containing the phosphorus-carbon bond have been identified in biological systems. Such compounds could persist in freshwater because of the stability of the bond, but they are thought to be insignificant quantitatively compared with phosphate-containing compounds.Some phosphate properties governing the different forms of phosphorus likely to exist in freshwater are (i) condensation reactions resulting in esters and polyphosphates, (ii) precipitation reactions forming low-solubility compounds, (iii) adsorption on to surfaces of clay minerals and hydrous metal oxides and (iv) soluble complex formation with transition and alkaline earth metals. Phosphorus Analyses in Freshwater The analytical procedures are based on the reaction of orthophosphate with molybdate in acidic solution to give molybdophosphoric acid, which, when reduced, forms molybdenum blue. Murphy and Riley’s method1 using antimony as catalyst and ascorbic acid as reducing agent is widely accepted.The concentration of all forms of phosphorus (total phosphorus) is determined by an acid-persulphate digestion of the unfiltered sample prior to the analysis for orthophosphate. Total phosphorus can be divided into two main components, soluble and particulate, by filtration of the sample through a membrane filter, conventionally with 0.45pm pore size. The concentration of all forms of soluble phosphorus (total soluble phosphorus) is determined by an acid-persulphate digestion of the filtered sample prior to the analysis for orthophosphate. The concentration of phosphorus determined by direct analysis of the filtered sample for orthophosphate is called soluble reactive phosphorus. Soluble reactive phosphorus concentrations may be greater than true orthophosphate con- centrations because the acidic conditions of the test may hydrolyse labile esters and poly- phosphates, disintegrate metal - orthophosphate complexes or dissolve colloidal hydrous metal oxides containing adsorbed orthophosphate. Experimental evidence that soluble reactive phosphorus was greater than orthophosphate by about 20% was obtained by comparison with an enzymatic orthophosphate method.2 Soluble unreactive phosphorus is the difference between total soluble and soluble reactive phosphorus analyses.I t represents phosphorus compounds capable of passing through the membrane filter and requiring digestion in acid-persulphate to convert them into orthophosphate.The particulate phosphorus concentration can be obtained by the difference between the total and total soluble deter- minations, or by direct analysis of the material retained on the membrane filter. The relative amounts of soluble reactive phosphorus, soluble unreactive phosphorus and particulate phosphorus in the major river inputs to Lough Neagh are about SO%, 15% and 35%, respectively. It is known that algae can use orthophosphate for growth, and algal available phosphorus in freshwater is often regarded as being equivalent to soluble reactive phosphorus. The chemical nature and availability to algae of the soluble unreactive and particulate phosphxus fractions are largely unknown. Evidence has been obtained for the existence of inositol phosphates in the soluble unreactive f r a ~ t i o n .~ , ~ The quantitative identification of components in the soluble unreactive and particulate phosphorus fractions is being investi- gated in this laboratory. Soluble Unreactive Phosphorus The concentrations of soluble unreactive phosphorus in the river inputs to Lough Neagh As these concentrations are too low for any direct identification are typically 10-20 pg 1-l.Anal. PYOC. of specific phosphorus compounds, a concentration step was developed. Evaporation, freeze-drying, ultrafiltration and adsorption - precipitation as methods of concentration were evaluated. A simple adsorption - precipitation technique was chosen using lanthanum as the coagulant, resulting in a concentration factor of at least 100 for soluble unreactive phosphorus.The soluble unreactive phosphorus concentrates were then subjected to gel filtration chromatography after selection of a suitable gel (Sephadex G-50), column (90 x 3.2 cm) and buffer (sodium tetraborate). Soluble unreactive phosphorus eluted at the void volume, indicating compounds with relative molecular mass > 10000, and together with soluble reactive phosphorus, indicating compounds with a relative molecular mass of about 1000. Similar observations have been made by gel filtration and anion-exchange chromato- graphy of freshwater without pre-concentration of soluble unreactive pho~phorus.~.~ Starting the concentration procedure with 10 1 of freshwater results in about 20pg of soluble un- reactive phosphorus eluting from the gel filtration column as the high relative molecular mass component.Attempts are now being made to process 100-1 batches of freshwater, perform gel filtration on a preparative scale and use a volatile buffer, so that soluble unreactive phosphorus fractions may be isolated in sufficient amounts to be able to perform more separation and identification techniques. Techniques such as mass spectrometry, gas chromatography of phosphate derivatives combined with mass spectrometry and 31P nuclear magnetic resonance spectroscopy could be useful. 376 POLLUTANTS I N THE TROPOSPHERE AND NATURAL WATERS Particulate Phosphorus Possible components of the particulate phosphorus fraction are sand, silt, coarse clay, medium and fine clay aggregates, algae, bacteria, zooplankton, organic debris and hydrous metal oxides with adsorbed orthophosphate, organic phosphates or polyphosphates.The hydrous iron( 111) oxide - orthophosphate component of particulate phosphorus is being investigated at present. Iron(II1) exists in freshwaters at pH 7-8 as hydrous iron(II1) oxide, on to which orthophosphate can be complexed and/or adsorbed. This hydrous oxide should be readily soluble in acid. Particulate phosphorus retained on a membrane filter was treated with sulphuric acid of similar concentration to that used in the soluble reactive phosphorus test. The orthophosphate and iron( 111) solubilised were then determined. There was a positive relationship between the amount of orthophosphate and the amount of iron(II1) solubilised for a range of river water samples.In the 76 samples analysed, acid- soluble orthophosphate accounted for 49% of the particulate phosphorus fraction. Acid- and alkali-soluble concentrates from particulate phosphorus are also being subjected to gel filtration studies in the same way as for soluble unreactive phosphorus concentrates. Organic phosphorus compounds that exist in the particulate phosphorus fraction because of association with particulate material may be identifiable. References 1. 2. 3. 4. 5. 6. Murphy, J., and Riley, J. P., Anal. Chim. Acta. 1956, 27, 31. Stevens, R. J., Water Res., 1979, 13, 763. Eisenreich, S. J., and Armstrong, D. E., Environ. Sci. Technol., 1977, 11, 497. Herbes, S. E., Allen, H. E., and Mancy, K. H., Science, 1975, 187, 432. Peters, R. H., J . Fish. Res. Board Can., 1978, 35, 315.Francko, D. A., and Heath, R. T., Limnol. Oceanogr., 1979, 24, 463. Reduction of Nutrient Input to Lough Neagh by Phosphorus Removal at Sewage Treatment Works A. V. Gray Water Quality Branch, Industrial Science Division, Department of Commerce, 5-1 1 Verner Street, Belfast, Northern Ireland The addition of extra nutrient to a lake can result in excessive algal blooms. A gradual enrichment of Lough Neagh has been taking place since 1900, due mainly to the advent of the water closet and changing agricultural practices involving the widespread use of artificialSeptember, 1980 POLLUTANTS IN THE TROPOSPHERE AND NATURAL WATERS 377 fertilisers. Fish life suffered, domestic and industrial water supplies were affected and the amenity value of the shoreline deteriorated.Published methods of controlling eutrophication1 include harvesting weed and/or algae, using fish as grazers, addition of copper sulphate or herbicides, flow diversion or dilution and lake modification. The Wood and Gibson report2 indicated that phosphorus was the limiting nutrient in Lough Neagh and estimated that 80% of the total phosphorus load entering it had originated from the major treatment works. Subsequent measurement by the Water Quality Branch (WQB) of the phosphorus load discharged from the major sewage treatment works3 revealed that the figure was only about 40% of the total. The remainder originates from agricultural run-off, land drainage, the rural population and rainfall.4 In 1967, massive blooms of Anabaena pas-aqua occurred.Chemical precipitation at source has also been employed. Phosphorus in Sewage The water-borne phosphorus entering a sewage treatment works comes from three main sources. The domestic contribution is split between the phosphorus originating from detergents and that of human origin from excreta. The third source is the industrial contri- bution. In Northern Ireland the detergents contribute about 45-50y0 of the total amount of phosphorus in sewage and industry is responsible for 15-20~0. The remainder is of human origin. The results of a survey5 of a Northern Ireland housing estate suggest that the domestic contribution may be lower than in the remainder of the UK. Untreated sewage contains a mixture of phosphorus compounds, including dissolved ortho- phosphate, dissolved and particulate organic phosphate, inorganic condensed phosphates and particulate inorganic phosphates, either precipitated or absorbed on to other solids.During the normal treatment stages at a sewage works, some of the particulate phosphorus will be biologically degraded and inorganic condensed phosphates will be hydrolysed to dissolved orthosphosphate. Detergent formulations include polyphosphates as builders and water conditioners, typically sodium tripolyphosphate (STPP) , Na,P,O,,. Human excreta contains various organic phosphates in addition to orthophosphate.6 Phosphorus Reduction at Sewage Treatment Works Some of the phosphorus present in sewage is removed during the normal treatment pro- cesses. Removals of 5-15y0 have been recorded at the primary sedimentation stage.’ During biological purification at the secondary treatment stage, phosphorus is required for normal cell growth, the amount being normally limited by the carbon requirements.Up to 20% reductions have been recorded at a normal activated sludge stage,s while 15% reductions at most were reported across a biological filter.g Enhanced phosphorus reductions have been obtained at some activated sludge works by modifiying the normal mode of operation, This “luxury uptake” process has received considerable attention in South Africa,lo-12 where consistent phosphorus removals of 85% have been claimed. Other physical methods for phosphorus removal include ion exchange, reverse osmosis, electrodialysis and adsorption using soil spreading1 As the control of biological processes for phosphorus removal is not straightforward and physical methods are generally economically prohibitive, most phosphorus reduction plants normally employ chemical addition.Chemical Methods The addition of a single chemical coagulant, or in some instances a mixture, normally results in the precipitation of soluble phosphates and coagulation of other species present, followed by particle agglomeration and subsequent settlement in the existing sedimentation tanks. The chemical solids so produced may have settlement characteristics inferior to the normal sewage solids and the existing sedimentation tanks may be somewhat ineffective in solid - liquid separation. Mechanical flocculation is often employed as a means of over- coming this problem by increasing the frequency of particle collision and causing larger378 POLLUTANTS I N THE TROPOSPHERE AND NATURAL WATERS Anal.Proc. aggregates to form, usually of increased density. Sometimes polyelectrolytes are added to improve floc aggregation further. A subsequent filtration stage may be employed to produce very small phosphorus residuals. Coagulants may be added at one of three different points in the normal sewage treatment process, viz., pre-primary, post-primary and post-secondary (or tertiary). Normally the chemical employed is a salt of calcium (most commonly lime), iron [as the iron(I1) or iron(II1) chloride or sulphate] or aluminium (as the sulphate, chlorohydrate or aluminate). Iron(I1) sulphate and lime have been dosed together on occasions.Development Work in Northern Ireland Conventional Dosing Taking account of local circumstances, the former Water Pollution Research Laboratory (WRRL), now the Water Research Centre (WRC), suggested that lime should be added as the coagulant at the tertiary treatment stage to effect a reduction in the phosphorus content of the sewage. Laboratory experiments were then undertaken by WQB to establish dose rates, floc settlement characteristics and size of flocculation tank required in order that a pilot-scale plant could be constructed. The results of the laboratory experiments indicated that the higher the alkalinity of the sewage, the greater is the amount of lime required to attain the same degree of phosphorus removal. Further, the process appeared to be pH dependent.By employing filtration lower lime doses could be used, with less sludge pro- duced (Fig. 1). 0 10 20 30 a? 2 6 40 E 50 3 60 z t? 70 L 80 30 1 m 6 20 $ --. - 0 a, 0, U 10 3j 90 100 8 9 10 1 1 12 PH Fig. 1. Percentage of phosphorus removed and increase in sludge volume with increase in pH. The initial pilot-plant investigations were carried out using pH-controlled dosing followed by flocculation, sedimentation and filtration to achieve phosphorus reduction. The results were encouraging, about a 75% phosphorus removal being achieved after settlement alone, and greater than 85% by subsequent pass'age through an Immedium filter (Fig. 2). Split-effluent Dosing This method involves splitting the sewage into two unequal portions, adding the total lime dose to the smaller volume and mixing for a certain period, then blending with the remainder of the sewage and finally mixing for a further period.The benefits of the split-effluent process are that phosphorus removals similar to those normally produced by the conventional process can be achieved using smaller amounts of coagulant, and the concomitant reduction in the pH of the final effluent reduces neutralisation costs (Fig. 3). At roughly the same time a coagulant was produced by the Industrial Science Division from local bauxite deposits. It was found to have a potential for removing phosphorus from sewage effluents. Consequently, laboratory investigations were carried out on severalSeptember, 1980 POLLUTANTS IN THE TROPOSPHERE AND NATURAL WATERS c ! - B /-- \ ......D . . , , . ,.: , , , , ,. , ,, , , , , ,., , , ,, ,, , . , . ... ....... . . . . .... . . .. . .._. . . . .. .. .. .. . ... . ..... 1 I 379 12 24 12 24 12 Wed. Thur. Fri. Fig. 2. Results for Rallymena pilot plant. A, Un- treated; B, after settlement; C, Immedium filtered; and D, 0.45-pm filter. coagulants. The final costings which the Department of the Environment produced showed clearly that the split-effluent lime process should be implemented at biological filtration plants. At activated sludge works dosing with iron(I1) sulphate to the mixed liquor stage was the most beneficial and economic process. 4 .o t I m -. - E 3.0 0 r Q L a - B 2.0 + m 3 - 0 $ a 1 .o 0 100 200 300 400 Lime dose/mg I-' Ca(OHh Fig. 3. Addition of lime to sewage: comparison of methods.Sample PF 112. Alkalinity, 99 mg 1-l: Split, 200 : 800. Phosphorus concentrations : settled, conventional (B), split effluent (C) ; filtered ( < 1 pm), conventional (F), split effluent (E). Sludge volumes: conventional (A) ; split effluent (D). WQB then carried out laboratory work to establish the necessary design parameters for The optimum ratio for flow Minimum mixing times of 2 min were required at large-scale implementation of the split-effluent lime process. separation was found to be about 1 : 4.380 POLLUTANTS I N THE TROPOSPHERE AND NATURAL WATERS Anal. Proc. both the first and second stages of the process if goDd phosphorus removals were to be achieved with sedimentation alone. Thz initial results were encouraging, and 75% phosphorus removals were achieved using a 1:4 effluent split and a lime dose of 220 mg 1-1 of Ca(OH), (Fig.4). Floc shear was critical on the second stage mixing. Modifications were then made to the existing pilot plant. Q m 2.0 c 0 I - Average = 3.56 mg % removal = 84.55 I- After sedimentation L--J Average = 0.25 mg I-' P I I I I I I-' P '. '. t Average = 0.55 mg I-' P U nf i I tered 24 24 24 24 Sat. 15.5.76 Sun. 16.5.76 Mon. 17.5.76 Tue. 18.5.76 Fig. 4. Pilot-plant phosphorus concentrations. Run PPE. Dose, 240 mg 1-l of Ca(OH),. Split, 19:81. Sludge production at the pilot plant was found to be about 0.6-0.7% of the total treated volume, compared with about 1.5% found in the laboratory work. The concentrations of solids in the sludge produced varied between 48 and 61 kg m-3 (4.8-6.1y0 solids).The pilot plant was then further modified to enable it to treat the total flow at the works. During periods of smooth operation excellent phosphorus removals of greater than 90% could be achieved (Fig. 5). 24 12 24 12 24 12 24 12 24 12.4.78 13.4.78 14.4.78 15.4.78 Fig. 5 . Full-scale plant phosphorus concentrations. Dose, Solid lines, unfiltered; broken line, filtered. 220 mg 1-1 of &(OH),. Iron(I1) Sulphate Dosing A full-scale phosphorus reduction plant capable of dosing iron( 11) sulphate was constructed at the Ballynacor sewage treatment works at about the same time as the full-scale split- effluent lime plant was built at the Ballymena works. Owing mainly to shortages of staff, WQB were not able to operate this plant concurrently with work at the Ballymena works.September, 1980 POLLUTANTS IN THE TROPOSPHERE AND NATURAL WATERS 381 The plant at the Ballynacor works consists of a debagging area and a feed hopper with a screw dry feeder leading into solution make-up storage tanks.These lead to a break- pressure tank, which is piped directly to the metering pumps. Control instrumentation has been installed to provide for automatic flow related dosing. Typical dose rates envisaged were 6-16 mg 1-1 of Fe. It was expected that effluent total residual phosphorus concentration of 1-1.5 mg 1-1 would be achieved. Recent Developments During the period of running the split-effluent lime process, bulk supply of lime from local producers ceased. The very significant increased costs of imported lime plus the potential availability of the locally produced alumino-ferric liquor have recently led to a re-appraisal of the choice of coagulant for phosphorus reduction. In the long term the capital and operational costs of using lime in the split-effluent process could be considerably greater than for dosing the alumino-ferric liquor.The Department of the Environment therefore requested WQB to carry out a feasibility study on the use of this coagulant added to crude sewage as a means of phosphorus reduction. The initial results indicate that dosing with the alumino-ferric liquor at three different rates depending on the time of day could be a viable alternative to the processes previously envisaged for phosphorus reduction at sewage treatment works in the Lough Neagh catchment. Full-scale trials are about to commence, the results of which will be used in deciding future management policy.References 1. 2. 3. 4. 5. 6. 7. Porcella, D. B., and Bishop, A. B., “Comprehensive Management of Phosphorus Water Pollution,” Ann Arbor Science Publishers, Ann Arbor, Mich., 1975. Wood, R. B., and Gibson, C. E., Water Res., 1973, 7, 173. Wilson, J. A., “Contribution of Point Sources Especially Domestic Sewage,” Lake Pollution Eutro- Smith, R. V., “Lough Neagh Investigations,” Eutrophication of Lakes and Reservoirs, Symposium Alexander, G. A., and Stevens, R. J., Water Res., 1976, 10, 757. Thorpe, W. V., Bray, H. G., and Jamej, S. P., “Biochemistry for Medical Students,” Churchill, Edinburgh, 1964, p. 515. Jenkins, D., and Menar, A. B., “The Fate of Phosphorus in Sewage Treatment Process, Part I, Primary Sedimentation and Activated Sludge,” SERL Report No.67-6, University of California, Berkeley, 1967. Jenkins, D., and Menar, A. B., “The Fate of Phosphorus in Sewage Treatment Processes, Part 11, Mechanism of Enhanced Phosphate Removal by Activated Sludge,” SERL Report No. 68-6, University of California, Berkeley, 1968. Vollenweider, R. A,, 1968, “Scientific Fundamentals of the Eutrophication of Lakes and Flowing Waters, with Particular Reference to Nitrogen and Phosphorus as Factors in Eutrophication,” Report to OECD, Paris, DAS/CSI/SS, 27, 1. phication Control, Seminar of the National Science Council, Killarney, 1977. of the Institution of Public Health Engineers, 1976. 8. 9. 10. 11. 12. Barnard, J.L., Water Res., 1975, 9, 485. Barnard, J. L., “Nutrient Removal in Biological Systems,” J . I m t . Water Pollut. Control, 1975, 74, Barnard, J. L., Water S.Afr., 19’76, 2, 136. 143. Determination of Polynuclear Aromatic Hydrocarbons in Water N. T. Crosby and D. C. Hunt Depavtment of Industry, Laboratory of the Government Chemist, Cornwall House, Stamford Street, London, SE1 9NQ Polynuclear aromatic hydrocarbons (PAHs) are fused-ring compounds formed during the pyrolysis or incomplete combustion of organic matter. Benzo [alpyrene (3,4-benzpyrene) is probably the best known and most extensively studied member of this family of compounds following its identification as a carcin0gen.l About 200 closely related PAH compounds have now been identified but not all are carcinogenic.These compounds are only very slightly soluble in water, their volatility is low and the absence of polar groups results in little reactivity apart from intense characteristic fluorescence emission spectra. Hence there is a real need to develop reliable analytical methods that can identify and discriminate382 POLLUTANTS I N THE TROPOSPHERE AND NATURAL WATERS Anal. PYOC. between a large number of individual compounds likely to be present in a wide variety of substrates only at extremely low levels. Other sources of PAHs include vehicle exhaust emissions and industrial effluents so that the compounds are found widely distributed through- out the environment. PAHs may, therefore, occur in water supplies as a result of pollution by industrial effluents, run-off from tarred or asphalted roads or by leaching from soil or from atmospheric fall-out.The potential danger to human health of these compounds was recognised in 1971 by the World Health Organization (WHO). In their International Standards for Drinking Water2 they recommended that amounts of six representative compounds (fluoranthene, 3,4- benzfluoranthene, 1 1 ,12-benzfluoranthene, 3,4-benzpyrene, 1,12-benzperylene and indeno- [1,2,3-cd]pyrene) should not, in general, exceed 0.2 pg 1-1 in potable waters. Higher con- centrations would be indicative of pollution and insufficient treatment. A similar standard has been proposed by the Council of the European Economic Community.3 Whilst the literature contains numerous papers on the determination of many different PAHs in a wide variety of substrates, few have dealt specifically with the separation and measurement of the six compounds recommended by WHO.Borneff and Kunte4 have described a two- dimensional thin-layer chromatographic method, following extraction of the sample with cyclohexane, which has a detection limit of about 2-3 ng 1-1 for each individual PAH. However, the potential of high-performance liquid chromatography coupled with fluorescence detection for this type of work has been investigated, initially for the determina- tion of residual amounts of PAHs in foods. The available column packing materials were unsatisfactory for the separation of these compounds and, hence, a new chemically bonded stationary phase, phthalimidopropyltrichlorosilane (PPS) , was prepared and evaluated.This particular material was selected bearing in mind both the structure of PAH compounds with their aromatic ring systems as well as the known selective extraction properties of dimethylformamide. The preparation of the new material has been described by Hunt eta1.536 Separations using the PPS column can be achieved that are not possible with other packing materials, e.g., octadecylsilane (ODs), and with both columns useful changes in retention times are observed that can be used to confirm the identity of unknown compounds. The PPS column can be used in both normal- and reversed-phase modes, the order of elution of the six compounds being different from that obtained on the ODS column. Of particular interest is the change in retention time of benzo [alpyrene and the fact that benzo[k]fluoranthene and perylene are separated on the PPS column but not on the ODS column.Many of the problems encountered in the analysis of different samples for PAHs arise from the difficulty of separating the compounds from the sample matrix. For example, we were asked to determine benzo[a]pyrene in samples of smoke produced during the burning of stubble. The smoke samples were collected on impactor filters and particles of soot were obviously present as well. In developing a method activated charcoal was used as a test substrate to check the recovery of PAH compounds from the filter media. Simple extrac- tion using a variety of solvents gave unacceptably low recoveries. Hence, the Soxhlet apparatus over a 7-h period was used.Recoveries of benzo[a]pyrene with four different solvents are shown in Table I. Clearly, benzene is far superior to other solvents for this particular substrate and yet other solvents are often used when particulate matter is known to be present. Superficially, water should be one of the easiest substrates for analysis as PAHs are only very slightly soluble in aqueous systems. However, at the low concentrations encountered TABLE I RECOVERY OF BENZO [a]PYRENE FROM ACTIVATED CHARCOAL USING A SOXHLET APPARATUS Recovery, yo v- 7 Solvent Range Mean Benzene . . . . . . 80-100 87 Cyclohexane .. .. 5-6 5.6 Dichloromethane . . .. 4720 8.6 Acetone . . .. .. 8-17 14September, 1980 POLLUTANTS I N THE TROPOSPHERE AND NATURAL WATERS 383 adsorption on to glass surfaces and sediment (if present) may cause difficulties for the analyst.In our work we have taken samples of River Thames water at Waterloo Bridge for analysis in the belief that other surface waters of drinking water quality would give fewer problems. Methods of analysis are generally checked by the addition of known amounts of the deter- minant to the sample, followed by analysis to check on recoveries. In this instance, the poor solubility of PAHs in water makes this operation difficult and the compounds are usually added in an organic solvent (often the same solvent used in the subsequent extraction stage). Clearly, this does not adequately reflect the real-life sampling situation. We attempted to “spike” tap water with PAHs dissolved in methanol. After shaking, the sample was filtered through a membrane filter and 60-80y0 of the PAHs were recovered.However, after standing overnight before extraction and analysis, only 45-60Y0 of the original amounts of PAHs were recovered, suggesting that adsorption on to the walls of the glass bottle was taking place. Direct extraction of the water sample and bottle using isooctane gave recoveries of SO-SO% even after the sample had been left to stand for 3 days before extraction. This is shown in Table 11. If, however, the solvent was added to the bottle before taking the sample, increased recoveries in the range 75-95y0 were obtained. Hence, we recommend that some of the extraction solvent be added to each sample bottle before sampling the water for analysis.TABLE 11 EFFECT OF PRIOR ADDITION OF SOLVENT TO SAMPLE BOTTLE Unspiked/ng 1-1 A r \ Compound No solvent With solvent Fluoranthene . . .. . . 10.2 15.3 Benzo[b]fluoranthene . . . . 0.7 1.3 Benzo[k]fluoranthene . . . I 0.2 0.3 Benzo[a]pyrene . . .. . * 1.8 4.2 Tndeno[1,2,3-cd]pyrene . . . . ND* ND* Benzo[ghi]perylene . . .. 2.4 4.4 Spiked e R G no solvent, yo solvent, yo 69 74 66 83 71 81 81 96 60 88 59 82 * Not detected. The effect of filtration is shown in Table 111. A sample of raw River Thames water was sampled both with and without filtration and both with and without addition of solvent prior to sampling the water. Enhanced recoveries of PAH compounds were obtained for both unfiltered and filtered samples through prior addition of solvent. Using the Perkin- Elmer 3000 detector, picogram amounts of PAHs can be measured, giving a detection limit in favourable circumstances as low as 0.01 ng 1-1 in water. TABLE I11 EFFECT OF FILTRATION AND SAMPLING ON PAHs IN RIVER THAMES WATER PAH compounds/ng 1-1 Filtered Unfiltered f A \ Compound e W - b o solvent hWith solvent ’ Fluoranthene . . .. . . 10.2 15.3 540 667 Benzo;blfluoranthene . . .. 0.7 1.3 85 99 Benzo[k]fluoranthene . . . . 0.2 0.3 17 20 Renzo:a]pvrene . . . . .. 1.8 4.2 294 430 IndenoC1,2,3-cd]pyrene . . .. ND* ND* ND* ND* Benzo[ghi]perylene . . .. 2.9 4.4 500 541 * Not detectxl References 1. Cook, J. W., Hieger, I., Kennaway, E. L., and Mayneord, W. V., Proc. R. SOC., London, Ser. B, 1932, 111, 455.384 EQUIPMENT NEWS Anal. Proc. 2. 3. 4. 5. 6. World Health Organization, “International Standards for Drinking Water,” WHO, Geneva, 1971. Council Directive 75/440/EEC, Off. J . Eur. Commun., L.194, 25 July 1975, p. 29. BornefT, J., and Kunte, H., Arch. Hyg. Bakterzol., 1969, 153, 220. Hunt, D. C., Wild, P. J., and Crosby, N. T., J . Chrornatogr., 1977, 130, 320. Hunt, D. C., Wild, P. J., and Crosby, N. T., Water Res., 1978, 12, 643.