ANALYST, JANUARY 1987, VOL. 112 17 Di- and Tributvltin SDecies in Marine and Estuarine Waters. Inter4aborato;y Comparison of Two Ultratrace Analytical Methods Employing Hydride Generation and Atomic Absorption or Flame Photometric Detection Aldis 0. Valkirs and Peter F. Seligman Marine Environment Branch, Naval Ocean Systems Center, San Diego, CA 92152, USA Gregory J. Olson and Frederick E. Brinckman Surface Chemistry and Bioprocesses Group, National Bureau of Standards, Gaithersburg, MD 20899, USA and Cheryl L. Matthias and Jon M. Bellama Department of Chemistry and Biochemistry, University of Maryland, College Park, MD 20742, USA Di- and tributyltin compounds present in marine and estuarine waters at sub-parts per billion (<pg 1-1) levels were determined using two different chemical speciation procedures.Generally, good analytical agree- ment was obtained from split samples independently analysed by a simultaneous hydride generation - dichloromethane extraction procedure followed by gas chromatographic separation and flame photometric detection (GC - FPD, performed at the National Bureau of Standards) and by a hydride generation procedure followed by purge and trap collection with boiling-point separation and atomic absorption detection (HG-AA, performed at the Naval Ocean Systems Center). Sea water samples containing tributyltin at sub-p.p.b. levels can be stored frozen (-20 "C) in polycarbonate containers for up to 2-3 months without any serious loss of analyte. Keywords: Antifouling marine biocides; atomic absorption and flame photometric detection; h ydride generation; inter-laboratory comparisons; organotin speciation Organotin compounds are increasingly used industrially as catalysts, plastics stabilisers and biocides.1 Tributyltin species are among the most effective organotin biocides and their worldwide use as active agents in antifouling coatings, especially for ships,2 is rapidly expanding. The increasing use of tributyltin-based antifouling coatings has raised concern about their environmental fate and the effects on non-target organisms of the toxic tributyltin species released from the coatings. Bioassays with algae,3 oyster ,4 crabs and mussel larvae,6 mysid shrimp,' copepods8 and fish9 have shown sub-lethal and lethal effects of tributyltin at parts per billion (pg 1-1) and lower levels in water.Consequently, some nations have issued regulations (France) or proposed regulations (UK) to control the use of tributyltin-based antifouling coatings on small craft in an attempt to protect marine life (generally shellfish) near harbours and marinas. The US Environmental Protection Agency recently initiated a special review on the environmental use of organotin com- pounds. 10 In order to evaluate effectively the environmental risks associated with tributyltin biocide usage and to develop monitoring strategies, analytical methods capable of the detection and speciation of key diagnostic butyltin species in environmental waters at concentrations as low as parts per trillion (ng 1-1) levels must be developed and employed. The determination of the toxic tributyltin species and its less toXic3J1J2 primary degradation product, dibutyltin, is of paramount importance.The relative abundances of these species offer clues to their persistence and toxic impacts at ambient levels in environmental systems. However, the chemical determinations of these species in natural waters are difficult and cannot be achieved by conventional total tin analysis. Only recently have methods been described for the requisite sub-pg 1-1 speciation of butyltins in aquatic environ- ments. These include chromatographiclsl6 or boiling- point17-20 separation of butyltin species, followed by atomic absorption,16-18320 mass spectrometric,l4 or flame photo- metricl3715>19 detection. Sample derivatisation by Grignard reaction14-16 or hydride generation13J7-20 is usually employed.Unfortunately, there have been no standards on which to base assessments of the relative accuracy of these analytical methods. Consequently, an inaugural inter-laboratory com- parison of methods for the detection of tributyltin in de- ionised water was organised and recently performed by the National Bureau of Standards (NBS), employing a stable, chromatographically purified aqueous tributyltin research material.21 In general, the 35 participant laboratories per- formed well in determining total tin (as tributyltin) at the parts per million (mg 1-1) level, but speciation of the tin at sub-pg 1-1 levels is an entirely different problem. A new intercom- parison is planned using a more dilute mixed butyltin species research material.It is also necessary to analyse butyltin species in natural water samples to determine if it is possible to compare data from different laboratories at low, ambient levels (ng 1-1) in marine and estuarine waters. Similarly, sample preservation procedures such as freezing must also be evaluated in order to determine their applicability for the storage and exchange of environmental samples. Such procedures were effective in the storage of sea water samples containing organoarsenic species, with respect to distribution and stability, for a period of three months at -20 OC.22 For these reasons, and for reporting preliminary results on concentrations of butyltin species in a variety of coastal marine and estuarine waters, the Naval Ocean Systems Center (NOSC) and the NBS undertook a joint intercomparison to measure tributyltin and its primary degradation product, dibutyltin, in shared marine and estuarine water samples collected on the east and west coasts of the USA and in England.18 ANALYST, JANUARY 1987, VOL.112 Experimental Glassware or plasticware (polycarbonate) was used directly if new, or else it was leached with 10% nitric acid solutions for 8-12 h and rinsed repeatedly with de-ionised water. Samples were collected from ships or piers in new polycarbonate (8-20 1) containers. Sampling sites included the west (San Diego Bay, CA, nine sites) and east (Chesapeake Bay, MD, three sites) coasts of the USA and the east coast of England (two sites). The sampling containers were submerged to 0.5-1.0 m depth, the caps were removed by hand, the bottles filled and the caps replaced underwater prior to removal from the water.In this manner, we avoided collecting surface microlayer films, which can contain butyltin compounds in relatively high concentrations. 1 3 ~ 5 On return to the laboratory, samples were either imme- diately frozen (-20 "C) and shipped, or aliquots were transferred into 1-1 polycarbonate containers, frozen and shipped. Prior to analysis, the samples were thawed at room temperature or by gentle warming (40 "C), but were not allowed to exceed room temperature and were analysed while still cool. Frozen Sample Storage Evaluation An effective sample storage procedure was considered essen- tial to allow successful comparisons of the determination of butyltin concentrations measured in split samples, and for the development of monitoring procedures.Frozen sea water samples containing tributyltin at sub-pg 1-1 concentrations were evaluated in order to document their long-term stability. A large volume of sea water, circulated over panels painted with antifouling paint containing tributyltin, was collected in a 20-1 polycarbonate bottle. This unfiltered source solution was poured into individual 500-ml polycarbonate bottles and frozen at -20 "C. Individual bottles were then removed at various times and analysed by hydride generation - atomic absorption (HG - AA) for tributyltin. Very little, if any, dibutyltin or monobutyltin was initially present. Subsequent analysis did not, therefore, address these species, as concen- trations were very near or below the detection limits (5 ng 1-1).Hydride Generation - Atomic Absorption Method At NOSC, the HG - AA methodl7J8 for producing volatile tin species with detection by modified hydrogen flame atomic absorption spectrometry was an adaptation of methods described by Braman and Tompkinslg and Hodge et a1.20 Inorganic tin and organotin compounds are derivatised to stannane and the respective alkyltin hydrides by sodium borohydride before detection. Briefly, a sample was placed into a 500-ml modified gas washing bottle and acidified to pH 5.0-5.5 with 2 M acetic acid. Hydride derivatives were formed by the addition of 4% mlV sodium tetrahydroborate(II1) prepared in a 1% mlV sodium hydroxide solution in distilled water.A ratio of 1 ml of sodium tetrahydroborate(II1) solution to 100 ml of sample was used to generate hydride species, which were purged from solution with helium carrier gas and trapped in a glass U-tube (2 mm i d . ) packed with 0.01-0.015 g of 3% OV-1 on Chromosorb W HP (80-100 mesh) and immersed in liquid nitrogen. Inlet and outlet lines to the U-trap and detector were made of FEP Teflon. The solution was purged for a total of 5 min after the addition of sodium tetrahydroborate(II1) to ensure the maximum removal of tin hydrides from solution. The trap was then removed from the liquid nitrogen bath and the tin species were separated and detected sequentially according to their boiling-points as they distilled from the trap. The tin species were carried into a quartz tube, atomised in a hydrogen - air flame and detected by an atomic absorption spectrometer at 286.3 nm.Gas flow-rates with respect to hydrogen, air and helium were 220, 140 and 40 ml min-1, respectively. The volatilisation of tributyltin hydride (Bu3SnH) required heating the trap in an oil-bath (180 "C). Standardisation was accomplished by the addition of an appropriate alkyltin standard (in ethanol) to the unknown, or by calibration graphs with the values calculated by peak integration. Concentrations of butyltin species are reported as ng 1-1 of tributyltin chloride or dibutyltin dichloride. The detection limit for dibutyltin and tributyltin hydrides was 5 ng 1-1. We demonstrated a relative standard deviation of 6.3% (standard deviation of a single measurement divided by the mean) for the analysis of five replicate determinations prepared in sea water at a concentration of 10 ng 1-1 of tributyltin chloride.The redistribution of butyltin groups resulting from the sodium tetrahydroborate(II1) derivatisa- tion process has not been detected in butyltin standards analysed at NOSC. The recent analysis of a tributyltin reference material prepared in distilled water at the NBS, containing tin only as the tributyl species,21 has shown no evidence of re-distribution. Gas Chromatographic Separation - Flame Photometric Detec- tion Method At the NBS, 100-200-ml aliquots of sea water samples were analysed by a simultaneous extraction (dichloromethane) and hydride generation method followed by gas chromatography - flame photometric detection (GC - FPD).13 Briefly, for a typical analysis of saline water with a butyltin concentration in the sub-pg 1-1 range, the following procedure was used. To 100 ml of sample in a 125-ml glass separating funnel, equipped with a Teflon stopcock and Teflon-lined screw-top, were added 2.8 ml of dichloromethane and 2.0 ml of 4% mlV aqueous NaBH4.In addition, a 10-p1 spike of a 0.5 p.p.m. aqueous solution of dipropyltin dichloride was added as an internal standard. The funnel was capped and shaken by hand for 1 min, vented and then shaken (240 strokes min-1) on a wrist-action shaker for 10 min. After a 5-min settling period, the lower organic layer was removed. An additional 1.4 ml of dichloromethane was added and the extraction procedure repeated. The organic layers were combined (approximately 2 ml) in polypropylene centrifuge tubes and evaporated to 100-200 p1 or less under a gentle stream of air.Appropriate reagent blanks were carried through the entire procedure. A gas chromatograph equipped with a flame photometric detector was used for the determinations. Chromatographic separations were carried out on a 1.83 m (6 ft) x 2 mm i.d. glass column packed with 1.5% OV-101 (liquid methylsili- cone) on Chromosorb G HP (100-120 mesh). A hydrogen-rich flame was employed, supported by H2 flowing at the measured rate of 110 ml min-1, air at 70 ml min-1 and N2 (zero grade) carrier gas at 20 ml min-1. The FPD was equipped with a 600-nm cut-on interference filter (band pass 600-2000 nm) to monitor the SnH molecular emission.The output signal from the FPD was recorded simultaneously on a strip-chart recorder and an integrator - plotter. For all runs reported here, the column temperature was programmed at 23 "C for 2 min and then heated to 170 "C at 32 "C min-1. The detector temperature was maintained at 200 "C and the injection port at 150 "C. The determination of butyltin species was performed either by the method of standard additions or from calibration graphs using sea water or estuarine samples containing little or no measurable levels of butyltin species. Butyltin and di- propyltin chlorides were used as received for the preparation of standard solutions for determination. Concentrations of butyltin species are reported as nanograms of tributyltin chloride or dibutyltin dichloride per litre.The GC - FPD method gives detection limits for di- and tributyltin species ofANALYST, JANUARY 1987, VOL. 112 L SnH4 HG-AA T 19 I Pr2SnH2 GC - FPD approximately 5 ng 1-1, with a relative standard deviation of 1&15% at 10 ng l-V3 I Results Typical chromatograms for the HG - AA and GC - FPD methods are shown in Fig. 1. Di- and tributyltin species were detected in the 14 coastal marine and estuarine surface water samples in concentrations ranging from trace to several hundred ng 1-1. Tables 1 and 2 show the means ( X ) for 2-5 replicate analyses (depending on the sites), standard devia- tions of a single measurement (S,) and the percentage relative standard deviations (RSD) of single measurements. With a few exceptions, the concentrations of the dibutyltin species were lower than those of the tributyltin species.In general, agreement was good between the two analytical methods employed. In all but five samples the NBS and NOSC I SnH4 Blank L Blank Y I ' I I 1 Time/m i n 0 4 8 Fig. 1. Chromatograms from analysis of environmental samples by HG - AA and GC - FPD methods. For HG - AA, Bu2SnH2 = 16 ng 1-1 and Bu3SnH = 9 ng 1-l. For GC - FPD, Bu,SnH2 = 28 ng 1-l and Bu3SnH = 90 ng 1-1 values were within 20% of their mean concentration for tributyltin. With dibutyltin, only four of the values varied by more than 20% of the means. The RSDs for replicate analyses were also similar for the two methods, being in the range 11-15% for the two tin species. There was no consistent trend for one method yielding a higher concentration of butyltins than the other.In 5 of 14 instances (36%), the GC - FPD values were greater than the HG - AA values for tributyltin. Also, 5 of 14 analyses showed greater values of dibutyltin with GC - FPD. Only two samples showed serious discrepancies (more than a two-fold difference) for tributyltin concentrations: the San Diego Bay-5 (four-fold) and the Bradwell Ck. marina (which barely exceeded two-fold) sites. Three sites showed such discrepancies with dibutyltin: the San Diego sites 8 and 9 (two- and three-fold, respectively) and the MAFF Laboratory (five-fold) site. Plots of HG - AA values compared with GC - FPD values for dibutyltin and for tributyltin showed good agreements with the theoretical line of slope 1.0 (Figs. 2 and 3). The analytical results (using the HG - AA method) for three frozen sample storage sets measured at different times and containing tributyltin at sub-pg 1-1 concentrations are presented in Table 3.Individual analytical values represent single determinations performed by one of three analysts on one of two instruments. Data from set I1 show that a slow loss of tributyltin occurs during the frozen storage of sea water samples containing tributyltin at sub-pg 1-1 levels. The regression line for these data has a slope of -0.00149 pg 1-1 per week, with a standard deviation of 0.00027. This corre- sponds to a tributyltin loss of about 1.4% per week of the fitted time zero value (0.112 pg 1-1). After 41 weeks of storage, the tributyltin level was less than 50% of the initial concentration (Table 3).In this study, replicate analyses of sea water samples containing di- and tributyltin gave average RSDs of 11-15%. Given this level of uncertainty, the frozen storage of samples for up to 2-3 months should not cause serious analytical problems. A significant negative slope was not obtained from tributyltin values regressed with storage time using data from sets I and 111. This is probably due to the relatively short storage times of sets I and 111, which were only one-third and one-sixth of the time of set I1 storage. Discussion This study evaluated the determination of dibutyl- and tributyltin species at ng 1-1 concentrations in natural waters Table 1. Concentrations of tributyltin detected in marine and estuarine waters GC - FPD (NBS) HG - AA (NOSC) X S, R.S.D., '/o X S, R.S.D.,% SanDiegoBay-6 .. . . 5* -t - 9 2.0 22 BayBridge,MD . . . . 6 1.2 20 NMS Baltimore Harbor, MD . . 11 0.8 7 12 --$ SanDiegoBay-7 . . . . 22 1.1 5 21 1.5 7 SanDiegoBay-4 . . . . 23 2.5 11 18 2.1 12 MAFF Lab., England . . 37 11 30 68 0 0 SanDiegoBay-8 . . . . 79 8.2 10 95 10 11 SanDiegoBay-5 . . . . 90 33 37 19 7.1 37 Annapolis,MD . . . . 97 5.1 5 103 5.2 5 SanDiegoBay-1 . . . . 184 30 16 270 26 10 SanDiegoBay-2 . . . . 338 60 18 369 38 10 - - - SanDiegoBay-9 . . . . 50 8.1 16 93 24 26 SanDiegoBay-3 . . . . 162 2.7 2 209 13 6 Bradwell Ck., England . . 732 --t - 332 43 13 Average . . . . . . . . 15 13 * Tributyltin chloride, ng 1-1. t Single analysis. 3 Not measurable.20 ANALYST, JANUARY 1987, VOL. 112 Table 2. Concentrations of dibutyltin detected in marine and estuarine waters GC - FPD (NBS) HG - AA (NOSC) MAFF Lab., England Bay Bridge, MD .. Baltimore Harbor, MD San Diego Bay-6 . . San Diego Bay-7 . . San Diego Bay4 . . Annapolis,MD . . San Diego Bay-4 . . SanDiegoBay-9 . . Bradwell Ck., England San Diego Bay-5 . . San Diego Bay-3 . . SanDiegoBay-2 . . SanDiegoBay-1 . . Average . . . . . . * Dibutyltin dichloride, ng 1-1. t Single analysis. $ Not measurable. X . . 2* . . 5 . . 11 . . 11 . . 13 . . 17 . . 19 . . 20 . . 20 . . 23 . . 28 . . 60 . . 263 . . 270 . . S , R.S.D., "/o --t 0 0 1.6 15 1.5 14 0.3 2 2.8 16 1 .o 5 5.1 26 2.7 14 7.9 28 7.7 13 18 7 48 18 13 - - - X 11 NM$ 11 16 12 52 29 23 49 19 25 75 21 1 270 S, R.S.D., Oh 1.1 10 - - - - 1.2 8 0.6 5 9.5 18 4.1 14 4.6 20 4.2 9 4.0 21 2.1 8 0 0 14 7 21 8 11 Table 3.Frozen storage of sea water samples containing tributyltin at sub-pg I-' concentrations Set I: Set 11: Set 111: April-August 1984 April-September 1985 September-November 1985 Weeks 0 0.3 0.9 1.1 2.1 3.3 4.1 4.9 6.0 6.9 7.0 7.9 8.9 9.9 11.7 13.7 13.9 14.0 Concentration* 0.13 0.13 0.14 0.11 0.13 0.09 0.13 0.10 0.11 0.10 0.10 0.10 0.11 0.09 0.13 0.12 0.15 0.10 * Tributyltin chloride, pg I-'. Weeks 0 0.1 1 .o 2.1 2.1 2.3 2.3 2.4 2.9 3.0 3.0 3.1 3.3 4.1 4.1 4.9 4.9 5.0 5.0 5.1 Concentration * 0.112 0.070 0.093 0.086 0.130 0.120 0.128 0.102 0.102 0.103 0.133 0.110 0.121 0.080 0.090 0.111 0.090 0.090 0.087 0.083 Weeks 5.3 5.4 5.6 5.6 5.7 5.7 7.6 7.6 11.4 16.6 16.7 16.9 17.9 18.4 18.4 18.7 18.9 20.9 41.4 41.6 Concentration* 0.105 0.085 0.090 0.080 0.071 0.082 0.097 0.097 0.073 0.057 0.060 0.075 0.089 0.073 0.103 0.054 0.071 0.099 0.043 0.046 Weeks 0 0.3 4.1 4.3 4.3 4.4 4.7 5.3 5.4 5.6 5.7 6.4 Concentration* 0.063 0.071 0.058 0.061 0.068 0.068 0.088 0.069 0.110 0.076 0.092 0.086 300 r I m - 8 2oo (3 I 0 100 200 300 400 GC - FPD, ng I-' Fig.2. Graph of HG - AA values versus GC - FPD values for tributyltin species from environmental sea water samples. The line is the theoretical slope, 1 .O 200 7 I 0 c - 8 100 GC - FPD, ng I-' Fig. 3. Graph of HG - AA values versus GC - FPD values for dibutyltin species from environmental sea water samples. The line is the theoretical slope, 1.0ANALYST, JANUARY 1987, VOL. 112 using two different measurement methods. The HG - AA method employs the direct hydridisation of the bulk sample, purging by an inert gas, trapping on a chromatographic substrate at liquid nitrogen temperatures and butyltin detec- tion by highly element-specific atomic absorption spec- trometry following boiling-point elution from the trap.The GC - FPD method employs hydride generation coupled with simultaneous extraction into an organic solvent (CH2C12). It is clear from this study that both methods give similar analytical results for di- and tributyltin species in marine and estuarine waters. It is possible that the few instances of disagreement (showing a 2-fold or greater difference in results) resulted from factors related to the chemical or physical composition of samples that affected one method more than the other. However, no obvious anomalies (oil pollution, excessive particulates, algal blooms, etc.) were noted in these samples.Frozen Sample Storage The data presented in Table 3 indicate that frozen storage of sea water samples containing tributyltin is a reasonably effective method of sample preservation for a period of approximately 2-3 months. Data from set 11, which was obtained over a 10-month period, showed that a slow loss of tributyltin occurred during frozen storage. The mechanism of this loss may not have involved de-alkylation as increases in degradation products (mono- and dibutyltin) were not noted. About half of the initial tributyltin concentration was present after a 10-month storage period. Data from sets I and 111, representing 6.4-14.0 weeks of frozen storage, did not show statistically significant negative slopes, and also confirm that frozen storage for several weeks does not result in significant losses of tributyltin.Frozen storage in polycarbonate bottles has proved to be effective in preserving sea water samples. Previous work has shown that storage of sea water samples containing tributyltin in polyethylene plastic containers resulted in substantial (62%) adsorptive losses from initial values after a one-week period at 4 OC.23 Samples stored in polycarbonate plastic, Pyrex glass and Teflon containers exhibited adsorptive losses of 3, 4 and 7%, respectively.23 Efforts are continuing to evaluate the stability of frozen sea water samples containing tributyltin for a period of approximately 1 year. Chemical speciation of butyltins in marine and estuarine waters at ambient (ng 1-1) levels is difficult and as a result there is little data available for toxicologists and environmen- tal agencies to evaluate. As legislation restricting the use of organotin antifouling paints is growing worldwide, there is a rapidly increasing need for monitoring data for the toxic tributyltin species and its less toxic degradation product, dibutyltin. This study is the first effort to make inter-labora- tory comparisons of methods for butyltin speciation in natural waters.Efforts of this type are important to establish intercomparability of methods and data as environmental monitoring programmes are undertaken. A related, parallel effort at NBS21 has chromatographically generated a stable aqueous tributyltin research material, used in conducting the first worldwide methods intercomparison for organotin measurements. The next phase of this effort is underway and will generate a mixed butyltin species research material for use in a inter-laboratory comparison involving speciation of the material diluted to sub-pg 1-1 levels.21 Development Center under program element 63724N. Ana- lytical support provided by Giti Vafa and Peter Stang is gratefully acknowledged. The NBS research was supported in part by the Office of Naval Research and the David Taylor Naval Ship Research and Devlopment Center. We gratefully acknowledge shiptime on the RN Ridgeley Warfield made available to us by the University of Maryland. We also thank Dr. Robert Paule of NBS for his helpful advice on data presentation and statistical considerations.Portions of this work will be included in the dissertation of C. L. M. to be submitted as a requirement for the PhD degree from the University of Maryland. The technical assistance provided by the advanced technol- ogy division of Computer Sciences Corporation, San Diego, CA, USA is gratefully acknowledged. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. References Blunden, S. J., Hobbs, L. A., and Smith, P. J., in Bowen, H . J. M., Editor, “Environmental Chemistry,” Royal Society of Chemistry, London, 1984, p. 51. Gitlitz, M. H., J. Coatings Technol., 1981, 53, 46. Walsh, G. E., McLaughlan, L. L., Lores, E. M., Louie, M. K., and Deans, C. H . , Chemosphere, 1985, 14, 383.Waldock, M. J., andThain, J. E., Mar. Pollut. Bull., 1983, 14, 411. Laughlin, R . B., French, W., Johannesen, R . B., Guard, H. E., and Brinckman, F. E., Chemosphere, 1984, 13, 575. Beaumont, A. R., and Budd, M. D . , Mar. Pollut. Bull., 1984, 15, 402. Salazar, M. H., and Salazar, S. M., in “Proceedings of the 11 th US - Japan Experts Meeting on Management of Sediments Containing Toxic Substances,” November 4-6, 1985, Seattle, WA. U’ren, S. C., Mar. Pollut. Bull., 1983, 14, 303. Seinen, W., Helder, T., Vernig, H., Penninks, A . , and Leeuwangh, P., Sci. Total Environ., 1981, 19, 155. Fed Regist., January 8th, 1986, 51, No. 5. Wong, P. T. S., Chau, Y . K., Kramer, O., andBengert, G. A . , Can. J. Fish Aquat. Sci., 1982, 39, 483. Laughlin, R. B., Jr., Johannesen, R . B., French, W., Guard, H. E., and Brinckman, F. E., Environ. Toxicol. Chem., 1985, 4, 343. Matthias, C. L., Olson, G. J., Brinckman, F. E., andBellama, J. M., Environ. Sci. Technol., 1986, 20, 609. Mueller, M. D., Fresenius Z . Anal. Chem., 1984, 317, 32. Maguire, R . J., Chau, Y . K., Bengert, G. A . , Hale, E . J., Wong, P. T. S., and Kramer, O., Environ. Sci. Technol., 1982, 16, 698. Maguire, R. J., Tkacz, R. J., J. Chromatogr., 1983, 268, 99. Valkirs, A. O., Seligman, P. F., Vafa, G., Stang, P. M., Homer, V., and Lieberman, S. H., Technical Report No. 1037, Naval Ocean Systems Center, San Diego, CA, 1985. Valkirs, A. O., Seligman, P. F., Stang, P. M., Homer, V., Lieberman, S. H., Vafa, G., and Dooley, C. A , , Mar. Pollut. Bull., 1986, 17, 319. Braman, R. S., and Tompkins, M. A., Anal. Chem., 1979,51, 12. Hodge, V. F., Seidel, S. L., and Goldberg, E . D., Anal. Chem., 1979, 51, 1256. Blair, W. R., Olson, G. J., and Brinckman, F. E., NBSIR 86-3321, National Bureau of Standards, Gaithersburg, MD, 1986. Yamamoto, M., Fujishige, K., Tsubota, H., and Yamamoto, Y . , Anal. Sci., 1985, 1, 47. Dooley, C. A., and Homer, V., Technical Report No. 918, Naval Ocean Center, San Diego, CA, 1983. Research performed at NOSC was sponsored by the Office of the Chief of Naval Research, Energy Research and Develop- ment Program and the David Taylor Naval Ship Research and Paper A61135 Received May 6th, 1986 Accepted July 7th, 1986