Determination of Mercury by Electrothermal Atomic Absorption Spectrometry Using Different Chemical Modifiers or a Slurry Technique Journal of Analytical Atomic Spectrometry I. KARADJOVA P. MANDJUKOV S. TSAKOVSKY AND V . SIMEONOV Faculty of Chemistry University of Sofa 1 J. Bourchier Sofa 11 26 Bulgaria J. A. STRATIS AND G. A. ZACHARIADIS School of Chemistry Aristotle University Thessaloniki 540.06 Greece The applicability of different chemical modifiers for thermal stabilization and ETAAS determination of mercury is studied. The modifier effect is strongly influenced by the type of acid and acid content in the sample solution. A method is described for mercury determination in tuna fish and sediment after high pressure digestion with nitric acid using thioacetamide (TAC) as chemical modifier.The method permits determination of 0.5 pg g- ' mercury. A simpler and faster procedure using a slurry technique for the determination of mercury in reference materials (tuna fish spinach cabbage and sediments) was also evaluated. Optimal instrumental parameters for mercury determination in slurries are presented. On the basis of the results obtained a new procedure which allows determination of 0.1 pg g- ' mercury was developed. Keywords Mercury; electrothermal atomic absorption spectrometry; modiJer; slurry technique Cold vapour atomic absorption spectrometry (CVAAS) with or without amalgamation is one of the most widely used methods for mercury determination because of its high sensi- tivity and reliability. However this technique requires additional equipment to the conventional instrumentation and a large number of reagents are used which can give rise to a relatively high blank value.Solid-phase microsampling is not possible and also some matrix interferences on vapour gener- ation have been observed. Electrothermal atomic absorption spectrometry (ETAAS) is another common and relatively sensitive method for mercury determination although the high volatility of mercury and some of its compounds even at room temperature offers a lot of difficulties in its analytical application. A number of investigations have been presented for mercury stabilization in the graphite furnace. Thermal stability of mercury has been achieved by using different modifiers intro- duced into the furnace as a solution (NH4)2S,1 H202 and HCl,2-5 Te di~hromate,~ etc.Grobensky et aL8 proposed another scheme for mercury stabilization; injection of the modifier (PdCl,) into the furnace followed by appropriate heating and after cooling injection of the sample. However some of the conclusions of the papers cited above are ambigu- ous and depend strongly on the particular analysis. For optimal ETAAS determination of mercury efficient thermal stabilization is necessary in both the drying and pretreatment steps. In our opinion only chemical reactions can stabilize mercury in the drying step; the role of the modifier is in the formation of a thermally stable compound of mercury preventing reduction of Hg2+ to Hg'. In this study the use of different organic and inorganic modifiers (Pd as PdCI2 Ce as Ce(NH4),(N0,) TAC (thio- acetamide CH,CSNH,) and mixture of H202 and HCl) were investigated and compared for thermal stabilization and ETAAS determination of mercury in different reference mate- rials.Optimal instrumental parameters for determination of Hg after high pressure digestion with nitric acid using TAC as modifier are presented. A simple procedure for analysis of Hg after slurry preparation is also proposed. The results obtained by the above techniques are compared with those obtained by the conventional cold vapour technique. EXPERIMENTAL Apparatus The AAS measurements were carried out on a Perkin-Elmer Zeeman 3030 atomic absorption spectrometer coupled with an HGA 600 atomizer. The light source was an electrodeless discharge lamp for mercury.The spectral bandpass and the wavelength used were as recommended by Perkin-Elmer. Uncoated and pyrolytic graphite coated graphite tubes were used as atomizers. Solutions (10-20 pl) were introduced into the graphite furnace using an AS-60 autosampler (Perkin- Elmer). Atomic absorption signals were recorded on an Anadex printer. The apparatus for the cold vapour determinations included a 100 ml Dreschler bottle a humidity trap with Mg(C104)2 as desiccant and a 15 cm cylindrical cell mounted on the AA spectrometer. Reagents Nitric acid hydrochloric acid and hydrogen peroxide used for sample digestion and slurry preparation were of Suprapur grade (Merck). All other reagents were of analytical-reagent grade. Mercury stock solution 1000 mg 1-' was Spectrosol grade (Merck).Modifier solutions were prepared by dissolving appropriate amounts of PdCl Ce(NH,),(NO,) or TAC in doubly distilled water. Sample Preparation Procedure 1 wet digestion An accurately weighed amount (about 500mg dry mass) of spinach cabbage sediment or tuna fish was placed in a Teflon crucible and 7.0ml of concentrated HNO were added. The crucible was covered placed in a pressure bomb and left for 3 h at a temperature of 90°C. The resulting solution was transferred into a 10 ml (25 ml for CVAAS) calibrated flask and diluted to the mark with doubly distilled water. The procedure for digestion of the sediments was the same except that a mixture of 4 ml of concentrated HNO + 2.5 ml Journal of Analytical Atomic Spectrometry December 1995 Vol. 10 1065of concentrated HC104+ 1.5 ml of concentrated HF was used for the acid digestion and the temperature was raised to 145 "C.Table 1 Maximum pretreatment temperatures for the loss-free deter- mination of mercury (0.1 pg ml-') obtained with different Pd or Ce concentrations Procedure 2 slurry preparation The dried samples were ground in an agate ball-mill until the particle size was < 50 pm. An accurately weighed amount of sample (about 50-100 mg) was placed into a polyethylene vessel and mixed with 1 ml of an aqueous solution containing 6% v/v HC1+2% v/v HN03+4% v/v HzOz. Finally 0.1 ml of an aqueous solution of 0.5% v/v Triton X-100 was added and the slurry was prepared after manual shaking. Atomic Absorption Measurements E TAAS An aliquot of 10-20 p1 of the digested sample solution or the slurry solution was injected into the graphite furnace.The chemical modifiers were injected either separately or pre-mixed with the tested solution. CVAAS An aliquot of 10 ml of the digested sample solution (procedure 1) was transferred into the Dreschel bottle with 15 ml of doubly distilled water and a suitable portion of a 5% m/v solution of SnCl was added to reduce mercury. The physico- chemical and instrumental conditions were as described in ref. 9. RESULTS AND DISCUSSION ETAAS Determination of Mercury Pd as PdCl and Ce as Ce(NH4)2(N03)6 as modijiers Palladium and cerium were chosen as representatives of two types of modifiers. It is known that Pd exists as Pdo and Ce as CeO during the pretreatment step. Previous investigations showed that with these modifiers the losses in the drying step were strongly dependent on the type of acid and acid content in the sample solution.For this reason for model investi- gations aqueous standard solutions of mercury were used. Various Pd and Ce concentrations from 0.1-500 pg ml-' (stoichiometric ratio Pd:Hg) were studied for thermal stabiliz- ation of 0.1 pg ml-' Hg. As can be seen from the pretreatment and atomization curves (Fig. l ) even very low concentrations 0.15 -$ 0.13 0 $ 0.11 a 3 0.09 2 0 c) = 0.07 0.05 A o 200 400 600 aoo 1000 1200 Ternperat ure/"C Fig. 1 A Pretreatment and B atomization curves for mercury stabil- ization (20 pl of 0.1 pg ml-' Hg) in presence of different concentrations of Pd (10 pl) 0 0.2 pg ml-'; 0 1 pg ml-'; 0 50 pg ml-'; and .500 pg ml-' Pd/pg ml - T/"C Ce/pg ml - ' T/"C 0.2 200 0.2 200 0.5 300 0.5 250 1 350 1 300 5 3 50 5 3 50 50 400 250 3 50 500 400 500 400 of Pd or Ce stabilize mercury up to 200°C. Enhancement in the concentrations of Pd or Ce leads to an increase in the maximum pretreatment temperatures (Table 1). Also for modi- fier concentrations in the range 0.1-50 pg ml-' no difference in the sensitivity for mercury determination (integrated absorbance was used in measurements) is observed when maximum power or 1 s ramp time in the atomization step was used. An atomization temperature of 850°C gives the best sensitivity. Higher atomization temperatures decrease sensi- tivity owing to convection losses between the pretreatment and atomization step in connection with the high volatility of mercury species.The presence of Pd or Ce at 500 pg ml-' level allows higher pretreatment temperatures to be used about 400 "C but the sensitivity is lower. It might be assumed that a high modifier content promotes the trapping of mercury in a matrix crystal lattice thereby decreasing its equilibrium vapour pressure. This requires higher atomization temperatures and even with maximum power convection losses prior to the atomization step are remarkable. The absorbance signal profiles and appearance time for mer- cury as can be expected depend on the concentration of Pd or Ce. Using aqueous standard solutions and Pd as modifier losses during the drying step are around 25% (Table 2) and do not depend on Pd concentration. Probably the reduction of Pd in the graphite furnace prevents some losses of mercury.The same is almost true for standard solutions of mercury in the presence of hydrochloric acid with Pd or Ce as modifiers. Unfortunately in the presence of nitric acid (nitric acid content around 10% v/v in the sample solution) losses of mercury during drying step are around 80% even with 500 pg ml-' Pd or Ce as modifiers (Table2). A probable explanation is the formation of Hg(N03) in the presence of nitric acid which decomposes at a much lower temperature (79 "C) in compari- son with HgC1 (mp 276°C and bp 302"C).10 Therefore it is impossible to employ the modifiers Pd or Ce for mercury determination in nitric acid digested samples. Mixture of HC1+ H202 + Pd and TAC as modijiers A mixture of HCI and H,Oz as modifier especially for the drying step has been pr~posed.~ This modifier permits a maximum pretreatment temperature of 240 "C.Taking into account our previous investigations it is evident that the mixed modifier HCl + H202 + Pd will ensure a maximum pretreat- Table 2 Mercury losses during the drying step (130 "C) using Pd as modifier mercury concentration 0.05 pg ml-' 20 pl sample aliquot Peak area (arbitrary units) t/s Pd 5 pg ml-'* Pd 50 pg ml-'* Pd 500 pg ml-'t 15 0.071 0.073 30 0.066 0.068 60 0.053 0.055 0.065 0.03 5 0.014 * Dissolved in 10% v/v HC1. t Dissolved in 10% v/v HNO,. 1066 Journal of Analytical Atomic Spectrometry December 1995 Vol. 10ment temperature of around 350 "C without losses during the drying step. This modifier was used for the determination of mercury in various samples (tuna fish and sediments) dissolved after high-pressure digestion with nitric acid.Unfortunately in the presence of a high nitric acid content (>2% in the sample solution) this modifier does not work. Marked losses of mercury were observed when these samples were introduced into the graphite furnace. Further investigations showed that in the presence of nitric acid the modifier which gives suitable pretreatment stabiliz- ation as well as temperature stabilization during the drying step is TAC with an optimal concentration of 10 g 1-'. The TAC hydrolyses in acidic solution according to the reaction CH3CSNH2 + 2H20 + H + +CH3COOH + NH4+ + H2S and thus permits the formation of HgS during the drying step. The role of this modifier is also to reduce the excess of nitric acid and thus to prevent the formation of thermally unstable Hg(N03)2.According to the pretreatment and atomization curves (Fig. 2) obtained for mercury standards with high nitric acid content and TAC as modifier the maximum pretreatment temperature is 230°C. This modifier was used in the determi- nation of mercury in reference samples tuna fish and sediments prepared according to procedure 1. Modifier solution 10 pl of l o g 1-' TAC was injected separately and mixed with the sample solution in the furnace using the AS-60 autosampler. For both tuna fish and sediment using instrumental conditions summarized in Table 3 and TAC as modifier the results for the determination of mercury were in good agreement with v) 0.05 \ 0 m 0.04 8 Ll 'p 2 0.03 2 CI) L = 0.02 0.06 r- - - - - A f I \" b I L 0.01 1 I 1 I 0 200 400 600 800 1000 1200 Temperature/"C Fig.2 A Pretreatment and B atomization curves for mercury deter- mination in presence of different matrices 0 tuna fish acid digestion sample with TAC as modifier; 0 sediment IAEA 158 slurry technique; and 0 cabbage IAEA 359 slurry technique Table 3 Instrumental parameters for mercury determination the certified values. The detection limit obtained was 0.5 pg g- ' (3s criteria six parallel determinations). The standard additions method is recommended for calibration. Results obtained for tuna fish and sediments are compared with those obtained with the cold vapour technique and good agreement was achieved (Table 4). ETAAS Determination of Mercury by Slurry Atomization During any digestion procedure used for sample preparation in mercury determination mercury losses may occur.An alternative approach for determining such a volatile element is direct solid microsampling or slurry atomization. Slurries were prepared according to procedure 2. On the basis of some preliminary results a mixture of 6% vjv HC1+2% v/v HN03 + 4% H202 ensures thermal stabilization of leached mercury in the drying step and also is very suitable for slurry homogenization. The addition of Triton X- 100 prevents slurry precipitation for a few minutes. Slurries are manually shaken before every manual injection into the graphite furnace. The optimal instrumental parameters were obtained using pretreat- ment and atomization curves for every sample treated accord- ing to procedure 2 (typical examples for sediment and cabbage are depicted in Fig.2) and are summarized in Table 3. Better sensitivity was achieved when uncoated graphite tubes were used as atomizers. In order to reach higher pretreatment temperatures without mercury losses Pd (50 pg ml- ') was mixed with the slurries before injection. The results obtained are almost the same as without Pd for tuna fish cabbage and spinach slurries. The pretreatment temperature of 240 "C achieved without Pd is good enough to minimize non-specific absorption. Also it is better to use a 1 s ramp time in the atomization step than maximum power. For slurries obtained from sediments the atomization temperature is critical. It is important to atomize mercury before the matrix element atomization so an atomization temperature of 900°C is used.In the presence of Pd as modifier a higher atomization temperature has to be used and thus very high values of non- specific absorption were observed (Fig. 3). As a conclusion for mercury determination by slurry atomization under the experimental conditions described there is no need to use Pd as modifier. Slurries have different physicochemical properties compared with aqueous solutions and it is clear that it is best to calibrate against certified reference material prepared in the same way as the sample. For slurries obtained from sediments this is the only possibility. Results achieved using slurry atomization with such calibration agreed very well with results achieved after acid digestion and ETAAS or CVAAS determination of Step Drying Pretreatment Atomization Cleaning Conditions T/"C Ramp/s Hold/s Gas flow/ml min-' T/"C Ramp/s Hold/s Gas flow/ml min-' T/"C Ramp/s Hold/s Gas flow/ml min-' T/"C Ramp/s Hold/s Gas flow/ml min-' Wet digestion (tuna sediment) 130 10 10 300 240 10 20 300 850 1 8 0 2200 1 3 300 Slurry (tuna spinach cabbage) 130 10 10 300 220 15 25 300 900 1 8 0 2300 2 4 300 Slurry (sediment) 130 10 10 300 220 20 35 300 900 1 8 0 2500 3 5 3 00 Journal of Analytical Atomic Spectrometry December 1995 Vol.10 1067Table 4 Comparison of results for mercury content (pg g-') in reference materials; n=6 BG 0.1 ( a ) ETAAS Reference material wet digested sample slurry technique Tuna fish (IAEA 350) 3.72 k0.06 3.97 k 0.08 Cabbage (IAEA 359) < DL 0.1 5 0.07 Spinach (IAEA 331) < DL 0.1910.07 Pine needles NIST SRM 1575 < DL 0.17+0.06 River sediment NIST SRM 1645 1.24 + 0.09 Sediment (Quasimeme 1993-94) 1.05f0.07 1.22f0.09 0.97 & 0.07 Certified CVAAS value 3.82 +0.08 3.9 0.1 1 f 0.08 NA* 0.17 k 0.08 NA - 0.15 1.1 1.03 k 0.08 NA - * NA = not available.0 8 Timds Fig. 3 ETAAS determination of Hg in sediment IAEA 1588 by slurry technique. Effect of atomization temperature ( xt) on absorbance signal (AA) profile for Hg (solid line) and background absorbance (BG) signals profile (broken line) at a pretreatment temperature of 300 "C (a) T = 900 "C; and (b) T = 1600 "C mercury (Table4). For slurries obtained from tuna fish cab- bage or spinach the standard additions method also might be used for calibration.The slurry analysis results for these samples compare well with concentrations obtained using a method based on acid digestion and analysis by ETAASl or CVAAS (Table 4). The detection limit for mercury by slurry technique is 0.1 pg 8-l ( 3 s criteria n= 10). CONCLUSION Some of the difficulties associated with mercury determinaf ion in some sample types e.g. soils vegetation and biota can be overcome by using chemical modifiers or a slurry technique. The application of well known modifiers such as Pd or Ce does not always provide a ready solution for any type of matrix. For nitric acid digested samples correct results for mercury content can be achieved by using TAC as modifier. The slurry sampling technique is proposed as a useful and effective method for rapid quantitative mercury determination even in a complex sediment matrix. REFERENCES 1 2 3 4 5 6 7 8 9 10 Ediger R. At. Absorpt. Newsl. 1975 14 127. Issaq H. and Zielinski W. Anal. Chem. 1974 46 1436. Alder J. and Hickman D. Anal. Chem. 1977 49 336. Owens J. and Gladney E. Anal. Chem. 1976,48 787. Lendero L. and Krivan V. Anal. Chem. 1982,54 579. Analytical Methods for Furnace Atomic Absorption Spectroscopy Perkin-Elmer Uberlingen Germany 1979. Rattonetti A. Instrumentation Laboratory Inc.; Report No. 12 Wilmington MA 1980. Grobenski Z. Erler W. and Vollkopf U. At. Spectrosc. 1985 6 91. Zahariadis G. and Stratis J. J. Anal. At. Spectrom. 1991 6 239. Handbook of Chemistry and Physics 56th edn. CRC Press Cleveland OH 1975-1976. Paper 51002788 Received January 17 1995 Accepted June 7 1995 1068 Journal of Analytical Atomic Spectrometry December 1995 Vol. 10