ANALYTICAL PROCEEDINGS. JULY 1989, VO1 26 255 Fourth Biennial National Atomic Spectroscopy Symposium The following are summaries of three of the papers presented at a Joint Meeting of the Atomic Spectroscopy Group and the Spectroscopy Group of the Institute of Physics held from June 29th to July lst, 1988, in the University of York. Twenty-four other papers appeared in full in the March issue of the Journal of Analytical Atomic Spectrometry. Use of the Hydride Generation Technique in the Determination of Bismuth with a d.c. Plasma Atomic Emission Spectrometer Paavo Peramski, Veli Matti Korvala and Lauri H. J. Lajunen Department of Chemistry, University of Oulu, SF-90570 Oulu, Finland The use of the hydride generation technique, together with various atomic spectrometric methods (atomic absorption, atomic emission and atomic fluorescence), has established its position in the determination of trace amounts of germanium, tin, lead, arsenic, antimony, bismuth, selenium and tellurium.1 Applications of the hydride generation method in connec- tion with a direct current plasma atomic emission spectrometer (DCP-AES) are not so numerous, however. The original work was done by Miyazaki et a l . , who determined arsenic and antimony in water samples with a batch-type hydride genera- tor. In their system the evolved hydrides were first collected in a liquid nitrogen trap.' Other uses of batch-type systems are the determination of selenium and tellurium3 and arsenic4 in aqueous solutions and germanium in ferrous dust.5 The determination of antimony in geological samples has also been reported.6 A batch type hydride generator is a simple piece of instrumentation, but a continuous-flow hydride generation system offers some advantages in contrast to a batch-type system, e .g . , the possibility of emission signal integration. Panaro and Krull used a continuous flow hydride generation method for the determination of arsenic in water and fish samples.' Krull and Panaro also interfaced a high-performance liquid chromatograph with a continuous flow hydride genera- tion DCP-AES for the determination of total tin and organotin compounds.8 Further, Ek and Hulden investigated the deter- mination of arsenic and selenium9 and Ebdon and Sparkes determined selenium in environmental samples with a continu- ous flow hydride generation DCP-AES.l o In this work bismuth was determined with a batch type hydride generator coupled to a DCP-AES. The method was applied to the determination of bismuth in geological reference samples. Experimental Instrumentation Bismuth emission signals were measured by use of a Spectra- Span IIIB single channel d:c. plasma atomic emission spec- trometer. A Hewlett-Packard 85 processor, interfaced with the spectrometer, was used for data processing when samples were introduced by pneumatic nebulisation. In the hydride genera- tion method emission signals were recorded by a Goertz 120E chart recorder. The peak heights were also read from the digital display of the spectrometer. The instrumental settings used in the measurements are given in Table 1.~~~~ ~ ~ Table 1. Instrumcntal parameters for determination of bismuth Sample introduction Equipment Hydride generation Nebulisation Spectromerer- Wavclengt h/nm 223.031 223.031,289.798 298.903.306.772 Entrance slit/pm 100.300 100. 300 Exit slit/pm loo. 300 100, 300 Ncbuliser pressurefib in-' 14 20 Integration time/s Active diagnostic 5 , lot Repeats 3.9t Sample volume/ml 10 Sample acidity. YO 5 NaBHl volume/ml 0.5 NaBH., concentration. YO m/V 2 .o Argon flow/l min- 1.5 Reaction time/s None PMT voltageN 800 * Gain 12 * Sleevc pressure/lb in-' 50 50 mode Hydride generutor- Recorder- Sensitivit y/mV 20-100 Chart speedcm min- * 0.5 * Variable settings were used. t Determination of background equivalent concentration and detec- tion limit. Reagents Only analytical-grade reagents were used. The reagents were obtained from Merck, except sodium tetrahydroborate, which was obtained from Fluka AG.The water used was demineral- ised and distilled. A stock solution containing 1000 mg 1-1 of bismuth was prepared by dissolving 1.oooO g of bismuth metal in 50 ml of concentrated nitric acid and diluting the solution to 1000 ml with distilled water. Other stock solutions, containing lo00 mg 1-1 of each element, were prepared from the following reagents: iron, germanium, tin, antimony, selenium and tellurium (metallic elements), nickel(I1) chloride hexahydrate, copper(1) nitrate trihydrate and lead(I1) nitrate. A reductant solution was prepared by dissolving sodium tetrahydroborate in distilled water. The solution was stabilised by adding one sodium hydroxide pellet (approximate mass about 200 mg) per 50 ml of solution.A fresh solution was prepared daily.256 ANALYTICAL PROCEEDINGS. JULY 1989. VOL 26 Procedures Solution nebulisation When characteristics for different wavelengths were measured, solutions were made up in 1% nitric acid and the sample uptake rate during nebulisation was 1.6 ml min-1. The different plasma positions (although they could not be precisely calib- rated) were defined as follows. The position where the entrance slit was just placed inside the excitation region (crook of the “V”) was defined as “-1”. Then the vertical control knob of the jet assembly was turned counter-clockwise, one turn giving the position “0” and two turns “+l”. 110 1 22 2% NaBH, 4% NaBH, Y g 90 1% NaBH, \+ a %loo- -..- r“ Y I-\<\% a 70 \ 3% NaBH, 1 2 4 6 8 10 12 14 Reaction time/s 6o 0 Fig. 1. (reductant volume 0.5 ml) The effect of reaction time on the determination of bismuth Hydride generation In the measurements made with the hydride generation procedure samples were dissolved in 5% hydrochloric acid, which is a suitable medium for hydride generation with bismuth.11,12 During the optimisation procedure, and in interference studies, the bismuth concentration in test solu- tions was 30 pg 1-1. Dilute solutions were always prepared just before use. The hydride generation system has been described else- where.3.h However, a slightly modified system was used in the determination of bismuth. The drying and delaying tubes of the hydride generator were removed and only a cotton wool plug was placed in the outlet tube of the generator.The measure- ment procedure was essentially the same as described in reference 6, except that a 2% (mlv) reductant solu” lion was used instead of a 4% solution. Also, the bismuth hydride was swept into the plasma immediately after the addition of reductant, without any collection time. Digestion of geological samples Reference samples were accurately weighed (0.2 g) into 100-ml beakers. Six millilitres of aqua regia were added and the beakers covered with watch glasses. The beakers were next put in a sand-bath for 1.5 h ( t = 120°C). Then the samples were allowed to cool and they were diluted with water and centrifuged. The solutions were transferred to 100-ml calib- rated flasks and diluted to volume with 5% hydrochloric acid.Table 2. Signal to noise ratios for different bismuth wavelengths* Wavelength/nm Plasma position S:N ratio 223.031 + 1/2 23.0 289.798 + 1/2 14.5 298.903 0 16.4 306.772 - 1/2 6.8 * Analytical solution contained 100 mg 1-l of bismuth. Finally, the solutions were further diluted, and the analysis were made by a standard addition method. Results and Discussion Pneumatic Nebulisation Four different wavelengthsl3 for bismuth were studied in detail (Tables 2 and 3). Although the emission line at 306.772 nm was very sensitive, it had a poor signal to noise ratio. Therefore, the emission wavelength 223.031 nm, which gave the best detec- tion limit, was selected for further use with the hydride generation met hod. 4 - 3.5 - 3.0 - 2.5 8 a - 2.0 $ - 1.5 - 1.0 - 0.5 L 1 I ’0 0 2 4 Red u ct a n t concentration , O/O Fig.2. The effect of reductant concentration in the bismuth determination (no reaction time, VNaBHJ = 0.5 ml) Hydride Generation Op tim isation The hydride generation for bismuth was first tried with the original SpectraMetrics hydride generator,3 but no bismuth emission signal was obtained. This was presumably a result of (thermal) instability* of the bismuth hydride formed’s (the same phenomenon has already been observed with tel- lurium3.16). Bismuthine is easily decomposed when it is swept through the drying and delaying tubes of the generator. Therefore, these tubes were removed and the bismuth hydride was purged directly into the plasma from the reaction vessel. When the hydrogen delaying tube is removed, extra care must be paid to the problems associated with the hydrogen that * The drying tube becomes hot when it absorbs water from t h e gas mixture.Table 3. Analytical characteristics for different bismuth emission lines measured by DCP-AES* Background Relative equivalent Detection Linear dynamic Wavelengt h/nm sensitivity concentratiodmg I- 1 limitlmg I- 1 range/mg I- I 223.031 0.11 2.644 0.056 0-100 289.798 0.04 5.294 0.116 O-lOOO 298.903 0.04 3.690 0.477 0-1o00 306.772 1 11.90 0.359 0-100 * The determinations were made according to procedures described in reference 14.ANALYTICAL PROCEEDINGS. JULY IYXY. VOL 26 5000 v) +- 5 4000 0 KJI 3000 -.. 2 .- Q) r 257 - - - Table 4. Results of the determination of bismuth in the reference samples (results in mg kg-1) This work Sample Material Individual result* Mean A B C GXR- 1 Jasperoid 14-46 1870 k 258 1700 > 100 1725 1876 1879 2037 2111 GXR-4 Copper mill 16.12 22.2 t 5.2 9 22 21.2 heads 1Y.Y7 22.12 22.58 30.37 D 1640 20 * Each value represents different weighting and determination. A.HCI digestion. MIBK extraction. FAAS.l‘’.’” B. Sodium hydrogen sulphate fusion, MIBK extraction, GFAAS.‘O C. KC103 - HCI digestion. MIBK extraction. FAAS.” D. HCI - Hz02 digestion, MIBK extraction, FAAS.” is a by-product in the reaction. A cotton wool plug was inserted in the outlet tube of the generator in order to damp the pressure wave generated. I t was also decided to purge the hydride immediately into the plasma after addition of the reductant so as to keep the hydrogen production to a minimum (Fig.1). A 4% sodium tetrahydroborate reductant solution used earlier3.6 was not suitable on this occasion, because the hydrogen was disturbing the plasma. Therefore, 0.5 ml of 2% reductant solution, giving satisfactory reproducibility for bis- muth hydride generation, was selected for use (Figs. 2 and 3). 120 I 1 8 0 0.2 0.4 0.6 0.8 1 .o Reductant volurne/rnl Fig. 3. bismuth ( CNaBH4 = 2%. no reaction time) The effect of reductant volume on the determination of The effect of argon flow through the reaction cell was also investigated. The best sensitivity was obtained when the flow was about 1.5 1 min-1 (this value was also used in the optimisation procedure). If the argon flow was higher, the plasma became unstable. When the nebuliser pressure was varied and argon flow through the cell kept constant (1.5 1 min-1).the best net signal to noise ratio was obtained at value 14 Ib in-’. The calibration graph for bismuth was linear up to 70 pg 1-1 (Fig. 4). A detection limit of 0.99 ng in a 10-mI sample was obtained (calculations were based on the mean of the blank plus three times its standard deviation, IZ = 10). The RSDs of the method were 2.2 and 1.5% for solutions containing 3 and 30 pg 1-1 of bismuth, respectively (n = 10). The matrix ion effects As a continuation of earlier studies,3.6 the interference effects of the other hydride forming elements and also iron, nickel and copper on the bismuth determinations were investigated. The bismuth concentration in test solutions was 30 pg 1 - 1 and increasing amounts of the other elements (one at a time) were added. As can be seen in Fig.5 , selenium, tellurium and copper interfaced at higher concentrations in the range studied. The interference effects caused by selenium and tellurium were similar to the effects of these elements on the determination of antimony.h Including the interferences of bismuth on the determination of selenium, tellurium and antimony,3.6 the interference effects seem to associate with the fact that at high concentrations bismuth, selenium and tellurium are reduced to free metals or perhaps metal borides” in the reaction vessel during the reduction step. However, the formation of a dark precipitate in the reaction vessel was not seen at this time. The interference effect caused by copper is obviously due to the formation of an inter-element compound with bismuth.6000 r :IlL-- 0 20 40 Bisrnuthipg 60 80 I - 100 120 140 Fig. 4. Calibration graph for bismuth Determination of Bismuth in Geological Samples Bismuth was determined in US Geological Survey reference samples GXR-1 and GXR-4.t The analyses were made by a standard additions method, because interferences were expec- ted in a complex sample matrix (the effects of the other elements can be diminished by using various masking agents, as reviewed by Nakaharal.18). The results of the determinations are shown in Table 4. The variations between individual results ~ ~~ t. The analysis of samples GXR-2 and GXR-5 (soil samples) was also tried. Because thcse materials contained very low levels of bismuth. one gram of each material was weighed.However. the solutions could not be diluted sufficiently and. when analysis from the concentrated solutions was attempted. large amounts of gas were produced. extinguishing the plasma. This problem could probably have becn overcome by evaporating the samples to dryness and re-dissolving the residue.258 a a .- G 40 a - a 20 ANALYTICAL PROCEEDINGS. JULY 1989. VOL 26 - - O 1 2 4 Molar ratio of interferent analyte zot bcu I /I I 1 10 20 40 100 200 0 1 . r ’ 1 2 4 Molar ratio of interferent analyte Fig. 5. bismuth The effect of the other elements on the determination ot are quite large, but the mean values are satisfactory in comparison with other published values. Conclusions The instability of bismuth hydride creates specific require- ments for hydride generator design, as is observed in this work. When the hydrogen delaying tube is removed, hydrogen causes problems with a d.c. plasma.These difficulties may, however, be overcome by careful adjustment of instrumental paramcters and solution concentrations. When bismuthine is swept directly into the plasma, the sensitivity is improved, as long as the quantification is based on measurement of the peak height. However, the linear dynamic range of the method is dimin- ished. In conclusion, the proposed method proved to be a suitable way of monitoring bismuth at low levels, as indicated by the results obtained from the GXR samples. 1. 2. 3 . 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.20. 21. 22. References Nakahara. T., Prog. Anal. Atom. Specrrosc.. 1983, 6. 163. Miyazaki. A.. Kimura. A., and Umezaki, Y., Anal. Chim Acra. 1977. 90. 119. Hayrynen, H.. Lajunen. L. H. J.. and Peramaki. P.. Arorn. Specrrosc.. 1985.6. 88. Boampong, C.. Brindle, I . D.. and Ponzoni, C. M. C., J. Anal. Afom. Specfrom., 1987. 2. 197. Brindle. I. D.. and Ponzoni. C. M. C.. Analysr. 1987. 112. 1547. Peramaki. P.. and Lajunen. L. H. J.. Anul.yst. 1988. 113. 1567. Panaro. K. W.. and Krull. 1. S.. Anal. Lerr.. 1984. 17. 157. Krull. I. S., and Panaro. K. W.. Appl. Specrrosc.. 1985, 39. 960. Ek, P.. and Hulden. S.-G.. Talanra. 1987. 34. 495. Ebdon. L., and Sparkes. S. T.. Microchem. J . . 1987. 36. 198. Fernandez. F. J.. Atom. Ahsorpr. Newsl.. 1973, 12. 93. Thompson. M..Pahlavanpour. B.. Walton. S. J.. and Kirk- bright. G. F., Analyst. 1978. 103. 568. Meggers. W. F.. Corliss. C. H.. and Scribner. B. F.. “Tables of Spectral Line Intensities. Part 1-Arranged by Elements.” Second Edition. U.S. Government Printing Office. Washing- ton. 1975 (reprinted by SpectraMetrics). “Handbook of Spectral Line Characteristics for thc DC PlasmdEchelle Systems.” SpectraMetrics. Andovcr. MA, USA. Fujita. K.. and Takada. T.. Tulanfa. 1986. 33, 203. Chapman, J. F., and Dale, I. S., Anal. Chim. Acfa, 1979. 111, 137. Bye. R.. Talanra, 1986. 33. 705. Nakahara. T.. Nakanishi. K.. and Wasa. T.. Spectrochim. Acra, 1987.42B. 119. Headridge. J . B.. and Richardson. J.. Analysr. 1970. 95, 930. Ficklin. W. H.. and Ward, F. N., U.S. Geol. S u n ) . J . Res.. 1976.4, 217. Viets. J. G.. Anal. Cliem.. 1978, 50. 1097. O’Leary. R. M., and V i m , J . G.. Arorn. Spectrosc.. 1986.7.4. Improvement in the Determination of Bismuth by Flame Atomic Absorption Spectrometry Using a New Design of Slotted Tube Atom Trap Narong Chimpalee, Michael Harriott and D. Thorburn Burns Department of Pure and Applied Chemistry, The Queen‘s University of Belfast, Belfast BT9 5AG, Northern Ireland The use of a slotted quartz tube as an atom trap in flame spectroscopy was first described by Watling. 1-3 The system has since been studied in a commercial format (STAT, slotted tube atom trap) by Brown et d.4-8 The elements that showed sensitivity improvements were volatile and formed compounds which decomposed thermally in the primary reaction zone of an air - acetylene flame.The improved sensitivity using a STAT was attributed to reduced flame speed, a longer optical path and the partial exclusion of the entrained air. This produces a stable chemical environment and enhanced concen- tration of neutral atoms. This system has been studied by using different tube designs to obtain the best sensitivities for bismuth. Experimental Apparatus A Perkin-Elmer Model 403 atomic-absorption spectropho- tometer was used with a Perkin-Elmer bismuth hollow cathode lamp. Signals were recorded using a Perkin-Elmer Model 56 chart recorder set at the 10-mV range. The spectrometer conditions for bismuth were: wavelength, 223.1 nm; lamp current, 10 mA; band-pass, 0.7 nm; flame, lean air - acetylene; aspiration rate, 5.6 ml min-1.A cradle similar to that of Brown and Taylor5 was built to align the slotted tube directly above the burner head slot along the instrument’s optical path (Fig. 1). The cradle is attached toANALYTICAL PROCEEDINGS, JULY 1989. VOL 26 259 Table 2. Sensitivity for bismuth with the various STATs Fig. 1. STAT and cradle the burner head by using the two burner safety wire retention screws; the slotted tube is held in place on two V-shaped adjustable plates on the loading arm by springs. The design allows a change over from analysis by STAT to that by conventional flame in seconds. The various STAT designs examined are described in Table 1. Table 1. Slotted tube atom trap designs STAT Tube material, dimensions and entrance slot A STAT tubes (A-E) were of B translucent quartz, 140 mm C long, 15 mm i.d.Entrance D slot, 115 mrn X 3 mm wide E I in each instance Exit at 180" to entrance slot, dimensions 1lOmm x 3 mm 90 rnm x 3 mm 50 mrn x 3 mm 30mrn x 3mm STAT E contained a row of 8 holes, 6 mm diameter, 9 mm apart 100mm x 3mrn 90mrnx3mm 50mrn x 3mm 30mmx3mm STAT J contained a row of 6 holes. 6 rnm diameter, 9 rnm apart Reagents All acids and salts used in the interference studies were of AnalaR grade (BDH Chemicals). Trace element stock solution was supplied by BDH Chemicals as 1000 mg 1-1 solution (SpectrosoL). All water used was doubly purified by distilla- tion followed by de-ionisation (Egla C114). Quartz tubes. Satin surface translucent and transparent grades from Thermal Syndicate. Results and Discussion Optimisation and Interference Study The position of the STAT relative to flame was varied by means of the V-ended adjustable plates.The STAT was adjusted in relation to the spectrometer's optical axis by means of the normal lateral and vertical burner position adjustment controls. The highest signals were obtained for STAT A-E with 10-mm gap between the burner head and the bottom of the STAT and for STAT F-J with an 8-mm gap. An enhancement of the bismuth absorbance signal com- pared with that given by the flame was obtained with all STAT tubes (Table 2 ) . The greatest enhancement was with STAT design I; however, because of the length of the tube and the STAT A B C D E F G H I J Flame alone Sensitivity*/pg ml- 1 0.34 0.27 0.26 0.25 0.27 0.30 0.26 0.26 0.24 0.25 0.50 * Concentration to give absorbance of 0.0044 small exit slot the flame deposited material on the spectrometer windows.Designs D and J gave the next highest responses, but the use of D also caused damage to the spectrometer windows. STAT design J was the best compromise between enhance- ment of signal and avoidance of deposition of material on to the spectrometer windows. After coating with lanthanum oxide the STAT tubes were able to withstand prolonged use. Devitrification, spreading outwards from the slots or holes, limited the tube lifetimes to about 300 discrete samples sprayed and measured. Linear calibration graphs were obtained for bismuth stan- dards (in 2% V/V nitric acid) over the concentration ranges 0-10 pg ml-1 (Fig. 2). The enhancement factor for STAT versus flame at the optimum observation height was ~ 2 .2 . The sensitivity, detection limits and precision data are as follows: sensitivity, 0.25 pg ml-* (= 0.0044 absorbance); detection limit, 0.143 pg ml-1 (= 2~ base-line noise); precision, 0.8% (rsd for 10 replicates at 4.0 pg ml-1). 60 1 Concentration of bismuth, p.p.m. Fig. 2. Calibration graph for bismuth. Enhancement factor = 2.2 An interference study was carried out on 4 pg ml-1 of bismuth for the elements expected (calcium, magnesium, sodium and zinc ions) over concentration ranges for calcium and magnesium of up to 500 pg ml-1 and for sodium and zinc of up to 1000 pg ml-1, exceeding the concentrations found in pharmaceutical formulations. The results showed no signifi- cant interferences. However, devitrification of the tube was observed, as noted by Brown and T a y l ~ r , ~ and this was caused by the presence of a high concentration of sodium in the formulations.In order to prolong the STAT tube lifetime, the surface was coated with lanthanum oxide by continuous aspiration of 1% m/V lanthanum chloride solution for 15 min. Analysis of Pharmaceutical Formulations Samples were placed in 100-ml Kjeldahl flasks and 72% perchloric acid and 30% hydrogen peroxide were added. TheANALYTICAL PROCEEDINGS. JULY 1989. VOL 26 260 Table 3. Analysis of selected bismuth containing pharmaceutical formulations Bismuth found/mg Declared bismuth Spectro- Sample con t en t/mg STAT* Flame* photomct rict Bismuth carbonate Bismuth carbonate Bismuth subgallate ointment ( 1 .(K) g ) 19 .x3 19.46 2 0.13 lY.65 f 0.39 19.75 2 0.3') tablet ( 1 tablet) 10.65 11 .O() f 0.07 11.21 * 0.20 10.91 * 0.10 suspension ( 1 ml) 49.17 50.45 f 0.32 52.24 k 0.72 5 1 .ox * 0.06 * Each value is the mcan and standard deviation for 1 0 analyses.t Each value is the mcan and standard deviation of 5 analyses. mixtures were heated gently to complete oxidation and placed (filtered if necessary) into 1 0 ml calibrated flasks and made up to volume with distilled water. Sample solutions were then diluted with 2% ( V / V ) nitric acid to bring them within range of the calibration graph. Results and Conclusions The results for the pharmaceutical formulations by STAT and conventional flame atomic absorption spectrophotometry arc given in Table 3. The formulations were also analysed independently by an ultraviolet - visible spectrophotometric method," the results being in good agreement with those obtained by STAT.With the use of STAT it is possible to determine bismuth with aqueous standard solutions. Results obtained by using a conventional flame system show a slight systematic positive bias compared with those by STAT. The precision attained using STAT was superior to that attained from the conventional flame system. The method described herein is thus considered suitable for the routine determination of bismuth in pharmaceutical formulations. 1. 3. 3. 4. 3. 6 . 7. 8 . 9. References Watling. R. J . . Atzul. Chitn. Actu, 1977. 94. 181. Watling. R. J .. Atiul. C'hitn. Actu, 1978. 97. 395. Watling. R. J . . and Watling. H. R.. Spctrodzitn.Actu, 1980. 35B. 451. Taylor. A . , and Brown. A. A.. Anulyst, 1983. 108. 1159. Brown. A. A.. and Taylor. A.. AtzuIysta 1984. 109. 1455. Brown. A. A.. Milner. B. A.. and Taylor, A.. Atiulyst, 1985, 110. 510. Brown. A. A.. and Taylor. A.. Atznly.\t, 1985. 110. 579. Lee, M.. Brohn. A. A.. and Lcfort. C.. Bioj ( N u n c y ) , 1985. 16. 31. Burns, D. T.. and Tungkananuruk. N.. And. C'hitn. Acrtr. 1987. 285. 197. Use of Surface Coatings in the Determination of Bis(tributy1tin) Oxide in Freshwater by Using Graphite Furnace Atomic Absorption Spectrometry C. Donaghy, M. Harriott and D. Thorburn Burns Department of Pure and Applied Chemistry, The Queens University of Belfast, Belfast BT9 5AG, Northern ire Ian d The rate of generation of the atomic vapour of tin from a graphite surface is a complex function of: surface temperature; the rate of temperature increase; and the physical and chemical nature of the surface.1.' These variations, coupled with the possibility of tin being in some molecular form rather than an atomic cloud, have led to the experimental development of graphite surface treatments.Instrumentation All impregnation studies were carried out with a Perkin-Elmer double-beam atomic absorption spectrophotometer (Model 403) in conjunction with the flameless atomisation device HGA76. The measurements were made with a hollow cathode lamp at the 286.3 nm resonance line with a band width of 0.7 nm. The signals were recorded on a Philips PM8251 chart recorder at 10-mV range. The optimised furnace programme Table 1.Furnace programme conditions Dry Ash Atomisc Cleanout Temperature 150°C 800 "C 3500 "C 200 "C Gas-stop 0 f f Off On Off Time 30 s 30 s -.. ' 5 s 5 s conditions for the atomisation of tin and organotin compounds are given in Table 1. Graphite Surface Treatments A range of metal salts were impregnated on to the graphite surface in order to improve the atomisation efficiency of tin.3 The impregnation method was simply the application of a 10% suspension of a metal salt on to the graphite surface using a BCL 1OO-pl micropipette. This procedure, plus the heating cycle in Table 2, was applied six times in order to give a durable surface coating. Table 2. The heating cycle Step Temperature Time 1 Drying 1 50 "C 60 s 2 Ashing 1 (MU) "C 30 s 3 Atomise 4 Cleanout 2(WX! "C 5 s 26" "C 1 0 s (GS) Standard calibration graphs, Fig.1. for tributyltin oxide, using different impregnated surfaces, were obtained by inject- ing SO yl of each standard solution on to each surface. The most effective metal salts were NH4V03,Ta205 and.4NALYTICAl. PKOCEEDINGS. JULY IYSY. VOL 26 26 1 0.75 7 1 a 0.5 e n <0.25 u C m m U 0.1 0.2 0.3 0.4 0.5 0.6 TBTO, p.p.m. Fig. 1. Calibration graphs ( NHl),Mo7021.3H20. In contrast, thc surface NHJVOl showed a much better long-tcrm stability with enhanced sensitivity, without any significant degradation in prccision. A statistical analysis of the results obtaincd using thesc surface matcrials is given in Tablc 3. Atomisation Profile Comparison Initial studies on the atomisation bchaviour of tin in the graphite furnace as a function of timc, as shown in Fig.2. rcvealed very little scnsitivity and poor precision. v h-.. - Ta I 1 Uncoated 1 t Time (300 mm min-1) Fig. 2. Atomisation profilcs This was attributed to the fact that tin forms a carbidc stable enough to slow down considerably the mass transfer of thc clement to the gaseous phase, even at thc highcst operating temperature of the furnace. Thus, a competitive rcaction for the available free carbon sites on the surface of the graphite tube might improvc the atomisation. A selection of transition metal salts as competitors in the formation of stable carbides were therefore screened. The solutions containing molyb- denum. vanadium and tantalum gave positive results in thcse scrccning cxperimcnts. The best over-all sensitivity and precision were obtained with tubes impregnated with vana- dium. Irish Sea Belfast River Land over 120 m Built-up area Strangford Lough - Key:- organotin Sn4+ 200 “c 100 v) 0 l- m I- 50 L 0 Seawater Estuarine Water Site No. 2 Site No.1 Fig. 3. Biological survey Solvent Extraction and Recoveries using a Vanadium Coated Tube The extraction, determination and speciation of tributyltin in seawater is based on the solvent extraction of organotin species directly into toluene and of inorganic tin as its tin(1V) - quinolin-8-01 chelate into chloroform. Thc detection limit was 10-12 ng with 00-100% recoveries.J.5 However, this concentra- tion - separation procedure, when applied to freshwater estua- rine samples. is hindered by the emulsification of the toluene in the aqueous layer.Recoveries are poor and strong interfer- ences occur in the graphite furnace AAS analysis. However, this emulsified layer can be greatly reduced by the addition of 20 ml of 2.5% analytical-reagent grade sodium chloride and 2 ml of methanol to the extract and centrifugation for 15-20 min at 3000 rev min-1. This procedure produces a clear toluene layer, removes graphite furnace interferences and gives 97- 98% organotin recoveries. Environmental Application There is currently considerablc interest in the monitoring andANALYTICAL PROCEEDINGS. JULY 1989. VOL 26 262 Table 3. Statistical analysis Increase in Detection R.S.D.. peak height. limitl Coating n X S.D. YO O/O I% None 5 72.6 2.32 3.19 Nil 6.0 NHJVO~ Ta205 5 108.1 1.64 1.52 50 5.2 ( N H J ) , , M O ~ O ~ J .~ H ~ O 5 121.4 3.14 2.59 70 4.5 5 150.0 1.82 1.21 100 3.0 impact assessment of tin and organotin compounds on aquatic systems. However, there are problems in the determination of these compounds: uncomplexed inorganic salts hydrolyse very rapidly in water; the solubility of inorganic tin is less than 1-2 p.p.m. over the pH range of 2-8; organotin compounds have a very low solubility in water, of the order of 5-100 p.p.m. Therefore, both inorganic and organotin compounds are likely to be found on sludges and solids in contact with water containing these species. The over-all solvent extraction and detection procedures have lowered the amount of tin and organotin that can be determined in aquatic systems. These methods have been applied to a biologically significant area of Strangford Lough in Northern Ireland (Fig.3). I t is apparent that a consistent and sensitive method of monitoring the possible pathways of organotin compounds into aquatic systems can be achieved. References 1. 2. 3. 3. 5. Vickrey. T. M.. and Harrison. G . V . . Anal. Chem., 1981. 53. 1573. Fritzsche, H., Wedscheider, W., and Knapp, G . , Talanta, 1979, 26, 219. Burns. D . T . . Dadgar, D . , and Harriott. M.. Analyst. 1984. 109. 1099. Burns. D. T.. Glockling. F . . and Harriott. M.. Analyst, 1981, 106, 921. Burns, D. T.. Harriott. M., and Glockling, F . . Fresenius Z . And. Chem.. 1987. 327, 7019. Analytical Applications of Spectroscopy Edited by C.S.Creaser, Uniuersity of East Anglia and A.M.C. Davies, Institute of Food Research, Norwich This book provides a ’ State-of-the- Art’ review of the applications of the major spectroscopic techniques and will prove invaluable to researchers involved in this form of analysis, The book provides wide-ranging coverage of recent developments in analytical spectroscopy, and in particular the common themes of chromatography - spectroscopy combinations, Fourier transform methods and data handling techniques. Each section includes a review of key areas of current research, written by spectroscopists who have made major contributions in their respective disciplines, as well as short reports of new developments in these fields. These common themes have played an increasingly important part in recent advances in spectroscopic techniques and emphasise the multidisciplinary approach of present research, 502 pages ISBN 0 85186 383 3 Price $47.50 ($99.00) ~~~~ ~ ~~ To order or for further information, please write to: Royal Society of Chemistry, Distribution Centre, Blackhorse Road, Letchworth, Herts SG6 IHN, UK or telephone (0462) 672555 quoting your credit card details. We now accept Access/Visa/MasterCard /EuroCard. RSC Members are entitled to a discount on most RSC publications and should write to: The Membership Manager, Royal Society of Chemistry, Thomas Graham House, Science Park, Milton Road, Cambridge CB4 4WF, U.K.