Determination of Organic Mercury and Inorganic Sediment, Soil and Aquatic Organisms by Atomic Absorption Spectrometry

Transcription

1 477 Differential Mercury Cold-Vapor in Determination of Organic Mercury and Inorganic Sediment, Soil and Aquatic Organisms by Atomic Absorption Spectrometry TakaShi TOMIYASU, Ayako NAGANO, Hayao SAKAMOTO and Norinobu YONEHARA Department of Chemistry; Faculty of Science, Kagoshima Korimoto, Kagoshima 890, Japan University, A method has been proposed for the differential determination of inorganic and organic mercury in environmental and biological materials, based on their successive extraction followed by a cold vapor atomic absorption spectrometric (CVAAS) determination. Mercury was first extracted by shaking a sample with 1 M hydrochloric acid containing 3% sodium chloride in the presence of copper(i) chloride. In order to separate organic mercury from inorganic mercury, the extract was shaken with chloroform to extract only organic mercury. The chloroform phase was transferred into a 50 ml volumetric flask and 0.4 ml of 0.2% dithizone-chloroform solution was added. The contents of the flask was evaporated to dryness, followed by addition of 2 ml of 1:1 nitric acid-perchloric acid solution and 5 ml of concentrated sulfuric acid and heating at ca. 230 C for 30 min. The hydrochloric acid phase containing inorganic mercury was transferred into a 50 ml volumetric flask, followed by addition of 1 ml of 60% perchloric acid and 5 ml of concentrated sulfuric acid and heating at ca. 230 C for 30 min. After cooling, the digested sample solution was diluted with water to 50 ml. The mercury in each solution was determined by CVAAS. When 0.5 g amounts of artificial standard soil samples spiked with a given amount of organic mercury and inorganic mercury {20 ng (as Hg) of mercury(ii) chloride and methylmercury chloride} were treated by the above-mentioned procedures, the recoveries were 102% and 97.0% with the relative standard deviations of 1.0% and 2.8%, respectively. This method was successfully applied to a differential determination of mercury in sediment, soil and fish meat samples. Keywords Differential determination, inorganic mercury, organic mercury, cold vapor atomic absorption spectrometry Mercury is widely distributed at various concentration levels throughout the environment and its toxicity depends on its chemical forms. The real danger to living organisms comes from the presence of organic mercury compounds, principally methylmercury, in the environment. Fish, plants or animals cannot, themselves, convert other mercury compounds into methylmercury1, but microorganisms in bottom mulls can do 50.2,3 This methylmercury is then taken up by aquatic organisms and concentrated in the food chains. That is to say, sediments are one of the most interesting types of environmental samples, to say nothing of biological materials. We must determine the total amount of mercury, and identify and quantify its chemical forms in such samples as well. A few methods have been proposed for the speciation of mercury in sediments. Most of them, however, were for determining the concentration of either inorganic mercury4,5 or organic mercury alone.6_g There are few methods for determination of both inorganic mercury and organic mercury in a single experimental run. We have reported a differential determination of organic mercury, mercury(ii) oxide and mercury(ii) sulfide based on a successive extraction of these mercury compounds with chloroform, a 0.05 M sulfuric acid solution and an 1 M hydrochloric acid solution containing 3% sodium chloride, respectively, followed by a cold vapor atomic absorption spectrometric (CVAAS) determination.9 The method, which involved backextraction of organic mercury into aqueous phase from chloroform, was applied to marine sediments with satisfactory results. But its sensitivity was not so good; the detection limit for the procedure with a 1 g sample were mg kg 1 Hg, mg kg-' Hg and 0.26 mg kg' Hg for organic mercury, mercury(ii) oxide and mercury(ii) sulfide, respectively. Furthermore, the method could not be applied to the differential determination of mercury in biological materials. The present paper describes the differential determination of organic mercury and inorganic mercury in environmental and biological materials. These mercury compounds were extracted from the materials with 1 M hydrochloric acid containing 3% sodium chloride in the presence of copper(i) chloride, and organic mercury was then extracted with chloroform from the hydrochloric acid solution. The separated inorganic and organic mercury were determined by CVAAS. The proposed method is simple and sensitive without requiring back-

2 478 ANALYTICAL SCIENCES JUNE 1996, VOL. 12 extraction of organic mercury and has been successfully applied to the differential determination of mercury in soil, sediment and fish meat samples. The detection limit of the method is estimated for an 1 g sample to be mg kg 1 Hg and mg kg-' Hg for organic mercury and inorganic mercury, respectively, and offers, therefore, an 8-fold improvement in sensitivity over the previous method9 for organic mercury. Experimental Apparatus The Sansou Automatic mercury analyzer Model Hg used, which was connected with a Rikadenki recorder, comprised an air circulation pump, a reaction vessel, an acidic gas trap, a four-way stop-cock and an atomic absorption spectrometer. Air was used as the carrier gas, fed through a potassium permanganate solution. A sample solution was placed in the reaction vessel and made up to 20 ml with water. The vessel was stoppered tightly and then attached to the closed circulation system. After 1 ml of 10% tin(ii) chloride solution was injected with a syringe, the air was circulated for 30 s through the four-way stop-cock to allow the mercury vapor to liberate from the sample solution completely. During this circulation, acidic gases leaving from the sample solution were removed by sodium hydroxide solution. After the circulation, the liberated mercury vapor was introduced into a quartz cell, in which the atomic absorption of mercury was measured at nm and recorded on the recorder. The maximum peak height obtained was used for the mercury determination. Reagents Pure water was prepared by purifying distilled water with a Millipore Milli-Q SP system just before use. Reagent-grade chemicals were used throughout. Sodium chloride was heated at ca. 700 C for 2 h in an electric furnace in order to remove any trace of mercury. A commercially available mercury(ii) standard (1000 mghg 1-1 in 0.02 M hydrochloric acid solution) was obtained from Wako. A working solution (0.1 mg Hg 1-1) were prepared by diluting this solution with 0.05 M sulfuric acid containing 3% sodium chloride. This working standard solution was stored in a brown glass bottle. Dithizone solution (0.2%) was prepared as follows: 1.2 g of dithizone (assay; 85%) was dissolved in 100 ml of chloroform and this solution was washed by shaking with 10 ml of 6 M hydrochloric acid. The washing was repeated two more times with an additional 10 ml aliquot of 6 M hydrochloric acid. After excluding the hydrochloric acid, the organic phase was washed by shaking with 10 ml of water. The water was excluded and the organic phase was diluted to 500 ml with chloroform. This solution was stored in a brown glass bottle and the surface of the solution was covered with 1% sodium sulfite solution in order to prevent the dithizone from air oxidation. A tin(ii) chloride (10%) solution was prepared by dissolving 25 g of tin(ii) chloride dehydrate in 50 ml of ca. l l M hydrochloric acid and diluting to 250 ml with water. Possible mercury contamination can be removed by bubbling nitrogen gas vigorously through the solution for 20 min. Artificial standard soil samples were prepared as follows; a known amount of each mercury compound was added to finely powdered and homogenized soil, which was made mercury-free 2 h previously. by heating at ca. 700 C for Procedure Procedure A (for determination of total mercury):10 a known amount of sample was placed in a 50 ml volumetric flask, to which 2 ml of 1:1 nitric acid-perchloric acid solution and 5 ml of concentrated sulfuric acid were added, and heated on a hot plate at 230 C for 30 min. After cooling, the digested sample was made up to 50 ml with water and the mercury in a suitable aliquot of the resulting solution (<20 ml) was analyzed by CVAAS. Since the heating at such high temperature might cause serious changes in the volume of the flask, the volumes of the volumetric flasks after being used many times in this procedure were measured; the weight of the pure water filled in the flasks was measured. There was no significant variation in the volume of the flasks t49.79 g g (ca.10 C), mean g, relative standard deviation 0.18%, n=15}. Procedure B (for differential determination of mercury): to a known amount of sediment (or soil) sample in a 50 ml glass centrifuge tube, 0.4 g of copper(i) chloride and 10 ml of 1 M hydrochloric acid containing 3% sodium chloride were added. The centrifuge tube was then stoppered, shaken with a shaker for 10 min and finally centrifuged at 3000 rpm for 2 min. After the supernatant was transferred into another centrifuge tube, the extraction was repeated three times with 10 ml aliquots of 1 M hydrochloric acid containing 3% sodium chloride. The residue in the centrifuge tube was transferred into a 50 ml volumetric flask, and the mercury in the residue was determined by Procedure A. To the combined hydrochloric acid extract in the centrifuge tube, 7 ml of chloroform was added, shaken for 2 min and centrifuged. The chloroform phase was transferred into a 50 ml volumetric flask. The extraction was repeated three times. To the combined chloroform extract in the 50 ml volumetric flask, 0.4 ml of 0.2% dithizone solution was added. The contents of the flask were evaporated to dryness. Organic mercury in the contents was then determined by Procedure A. A suitable aliquot of the hydrochloric acid phase (<5 ml), which contained only inorganic mercury, was transferred into a 50 ml volumetric flask and the mercury content was determined by Procedure A, except for the use of 1 ml of 60% perchloric acid instead of 2 ml of 1:1 nitric acid-perchloric acid solution.

3 479 From biological samples, mercury was extracted by shaking for 20 min with 20 ml of 1 M hydrochloric acid containing 3% sodium chloride solution in the presence of 0.4 g of copper(i) chloride. After a centrifugation, the supernatant was transferred into another centrifuge tube and 7 ml of chloroform was added. Other procedures were the same as those for the sediment. Results and Discussion Effect of the amount of copper(i) chloride on the extraction of mercury from real samples By shaking artificial standard soil samples with 1 M hydrochloric acid containing 3% sodium chloride solution in the presence of copper(i) chloride, spiked methylmercury chloride, mercury(ii) oxide and mercury(ii) sulfide were extracted quantitatively.9 However, the mercury compounds originally present in real sediments or soils may be more highly associated with the silicate matrix of the sediments or soils than the spiked mercury compounds in artificial standard soil samples and may resist the extraction into hydrochloric acid. Thus the extraction efficiency of mercury with the hydrochloric acid was investigated for real samples. As the addition of copper(i) chloride was quite effective for the elucidation of mercury(ii) sulfide from the artificial standard soil samples9, the effect of the added amount of copper(i) chloride was examined in the first place. The soil samples taken in Kagoshima University were shaken with 1 M hydrochloric acid containing 3% sodium chloride in the presence of various amounts of copper(i) chloride; the extraction was repeated three times with 10 ml aliquot of 1 M hydrochloric acid containing 3% sodium chloride. The mercury in the residue was determined by Procedure A. The ratio of the residual mercury decreased with an increase in the amounts of copper(i) chloride and did not change significantly at larger than 0.3 g of copper(i) chloride (Fig. 1). An amount of copper(i) chloride of 0.4 g was chosen. The soil samples were shaken for 10 min with successive 10 ml portions of 1 M hydrochloric acid containing 3% sodium chloride in the presence of 0.4 g of copper(i) chloride; mercury in the residue was determined by Procedure A. About 90% of mercury in the samples could be extracted by two successive extractions and no significant change of the extraction efficiency was observed by the increase in the number of successive extraction (2-4 times). As we expected the extraction efficiency to be improved by grinding samples, these examinations were done for finely ground soil samples. No improvement of the extraction efficiency, however, was observed. Hydrochloric acid extraction was thus repeated three times for sediment and soil samples in the recommended procedure. The effect of concentrations of hydrochloric acid and sodium chloride on the extraction efficiency of mercury were also examined; no significant change was observed in the examined concentration ranges of M and of 0-5%, respectively. The shaking Fig. 1 Effect of amount of copper(i) chloride on the extraction of mercury from soil. Total mercury concentration in soil samples (mg kg-1): ~, 0.097; Q, 0.12; 0, 0.073; p, 0.086; p, 0.037; +, The point marked X is the mean value of these 6 samples. time had no effect for the efficiency of the first extraction in the investigated range of 5-30 min. At selected extraction conditions, about 90% of mercury could be extracted into the hydrochloric acid solution from soil samples (86-93%, mean 91%, n=6). Inorganic mercury concentration of biological materials is usually considerably lower than that of sediments; thus, a smaller volume of 1 M hydrochloric acid containing 3% sodium chloride solution is favorable. Because the smaller the volume of the solution was, the higher the mercury concentration in extract became. The effect of the volume of hydrochloric acid solution was examined in the range of ml by treating a fish meat sample. No significant change of the extraction efficiency was observed at larger than 20 ml with 93-98% of the efficiency of first extraction at shaking time 20 min in the presence of 0.4 g of copper(i) chloride (mean 96%, n=3), whereas in the absence of copper(i) chloride about 15% of mercury remained in a residue after the extraction. In the case of biological materials homogenization was required for obtaining a good extraction efficiency. Thus mercury in biological materials was extracted by shaking a homogenized sample with a 20 ml aliquot of 1 M hydrochloric acid 3% sodium chloride for 20 min in the presence of 0.4 g of copper(i) chloride. Extraction of organic mercury from hydrochloric acid into chloroform The effect of the amount of chloroform on the extraction of organic mercury from hydrochloric acid containing 3% sodium chloride was examined. A 30 ml of

4 480 ANALYTICAL SCIENCES JUNE 1996, VOL. 12 Fig. 2 Effect of amount of chloroform (a) and dithizone solution (b) on the recovery of organic mercury. Methyl mercury chloride of 10 ng (as mercury) was added into (a) 30 ml of 1 M hydrochloric acid containing 3% sodium chloride and (b) 21 ml of chloroform. aliquot of 1 M hydrochloric acid containing 3% sodium chloride spiked with a given amount (10 ng as Hg) of methylmercury chloride was taken into a centrifuge tube. After the mercury compound was extracted with three x ml (x=3, 5, 7 and 10) portions of the chloroform, each of the first, second and third extracts were treated as in Procedure B. As shown in Fig. 2(a), recovery of added methylmercury chloride increased with an increase in the amount of chloroform; by extraction with three portions of 7 ml of chloroform, the methylmercury chloride added was recovered completely. Thus the extraction with three aliquots of 7 ml of chloroform was used in the recommended procedure. The effect of the shaking time in the 1-8 min range was examined with 7 ml of chloroform. The efficiency of the extraction did not change significantly in these variations in the shaking time. Thus, 2 min of shaking time was chosen as the recommended procedure. Determination of organic mercury in chloroform Mercury in an organic phase can not be determined directly by CVAAS. Although the mercury could be back-extracted from organic phase into aqueous phase by the use of thiosulfate solution9, the thiosulfate ion interfered with reduction of mercury by tin(ii) chloride on the determination of mercury by CVAAS. Thus, in this work, chloroform was excluded by evaporation. Ohkoshi et al. discussed the concentration of benzene sample solution by evaporation, but the recovery of added methylmercury chloride from their concentration Table 1 Separation and recovery of mercury compoundsd procedure was only 22%.11 Our investigation found that the addition of dithizone into chloroform prevents methylmercury chloride losses upon evaporation of the solvent. The effect of the added amount of 0.2%dithizone-chloroform solution on the recovery of mercury was thus examined with 21 ml of chloroform spiked with a given amount of methylmercury chloride (10 ng as Hg); the experimental conditions except for the amounts of dithizone solution were according to Procedure B. The result is shown in Fig. 2(b). The quantitative recovery (%) of spiked methylmercury chloride was obtained in the range of more than 0.3 ml. Thus, an amount of 0.2% dithizonechloroform solution of 0.4 ml was chosen. Separation and recovery of inorganic and organic mercury In order to check the efficiency of the procedures for

5 481 Table 2 Differential determination of mercury in real samples the separation of inorganic and organic mercury, a 0.5 g amount of artificial standard soil sample spiked with a given amount (20 ng as Hg) of methylmercury chloride and/or mercury(ii) chloride was treated by Procedure B. Table 1 shows that a nearly 100% recovery was achieved for each mercury compound. The relative standard deviations for 6 replicate determinations were 2.8% for organic mercury and 1.0% for inorganic mercury, respectively. Application to real samples The proposed method was applied to the differential determination of mercury in soil, marine sediment and fish meat samples. Table 2 shows the results, together with those obtained by Procedure A. The sum of the three values (i.e., organic mercury, inorganic mercury and residual mercury concentrations) agreed well with that for the total mercury obtained by Procedure A. The method was further checked by adding known amounts of methylmercury chloride and mercury(ii) chloride to the soil and fish meat samples. Good recoveries of % (mean 96%) and % (mean 99%) were obtained for the former and latter mercury compounds, respectively. The present method which can be applied to the differential determination of mercury in both sediment and biological materials is useful for elucidation of the behavior of mercury in environment. The authors are particularly indebted to Dr. Akagi of The National Institute For Minamata Disea and advice with CVAAS determination. References Hirokatsu se for help 1. K. Borg, H. Wanntorp, K. Erne and E. Hando, J. Appl. Ecol. Suppl., 3, 171 (1966). 2. F. A. J. Armstrong and J. F. Uthe, Atomic Abs. Newsl.,10, 101 (1971). 3. G. Oshima and K. Nagasawa, Eisei Kagaku, 16, 78 (1970). 4. N. W. Revis, T. R. Osborne, D. Sedgley and A. King, Analyst [London], 114, 823 (1989). 5. L. W. Jacobs and D. R. Keeney, Environ. Sci. TechnoL, 8, 267 (1974). 6. H. Hintelmann and R. D. Wilken, Appl. Organomet. Chem., 7, 173 (1993). 7. K. Matsunaga and S. Takahashi, Anal. Chim. Acta, 87, 487 (1976). 8. K. Tanaka, K. Fukaya, S. Fujiki and S. Kanno, Eisei Kagaku, 20, 349 (1974). 9. H. Sakamoto, T. Tomiyasu and N. Yonehara, Anal. Sci., 8, 35 (1992). 10. H. Akagi and H. Nishimura, "Advances in Mercury Toxicology", ed. T. Suzuki, N. Imura and T. W. Clarkson., p. 53, Plenum Press, New York, S. Ohkoshi, T. Takahashi and T. Sato, Bunseki Kagaku, 22, 593 (1973). (Received January 24, 1996) (Accepted March 18, 1996)