摘要:

ANALYST, JULY 1992, VOL. 117 1129 Comparison of Solvent Extraction and Solid-phase Extraction for the Determination of Organochlorine Pesticide Residues in Water Guan H. Tan Department of Chemistry, University of Malaya, 59100 Kuala Lumpur, Malaysia Solid-phase extraction (SPE) of organochlorine pesticide residues from environmental water samples was evaluated using octadecyl (CI8)-bonded porous silica. The efficiency of SPE of these pesticide residues from reagent water samples at 1-5 pg dm-3 levels was compared with those obtained by solvent extraction with hexane and Freon TF (trichlorotrifluoroethane). Average recoveries exceeding 80% for these organochlorine pesticides were obtained via the SPE method using small cartridges containing 100 mg of 40 pm CI8-bonded porous silica.The average recovery by solvent extraction with hexane and Freon TF exceeded 90% in both instances. It was concluded that the recoveries and precision for the SPE of organochlorine pesticides were poorer than those for the solvent extraction method. Organochlorine pesticide residue levels in environmen- tal water samples from two major rivers flowing through predominantly rice-growing areas were monitored by gas chromatography using the solvent extraction method with hexane. Exceptionally high levels of organochlorine pesticide residues such as BHC, DDT, heptachlor, endosulfan and dieldrin were found in these water samples. Keywords: Solid-phase extraction; organochlorine pesticide; solvent extraction; water; gas Chromatography Solid-phase extraction (SPE) using pre-packed reversed- phase octadecyl (CI8)-bonded silica materials contained in cartridges has been widely used for sample preparation for quantitative analyses.l-10 Organic trace determinations in clinical, environmental and food samples using the SPE sample preparation technique have been reported."-14 Appli- cation of the SPE technique to the determination of pesticides in environmental waters has been widely investigated.1519 However, the use of commercially available SPE cartridges for sample preparation in pesticide analysis has been shown to give rise to interferences, especially when gas chromatography (GC) with electron-capture detection (ECD) is employed.20 Extraneous peaks which appear in the gas chromatograms have been attributed to phthalate plasticizers in the poly- propylene housing material of these cartridges.20 A cornpara- tive study of the SPE technique using commercially available cartridges and the conventional solvent extraction method with hexane and Freon TF (trichlorotrifluoroethane) was carried out for 12 organochlorine pesticides.These solvents were chosen as an alternative to the US Environmental Protection Agency (EPA) Method 608 for organochlorine compounds.21 A monitoring study of organochlorine pesticide residue levels in environmental waters was also carried out using the solvent extraction method followed by Florisil column chro- matographic clean-up and GC-ECD. Water samples from two rivers, Sungai (Sg.) Bernam and Sg. Selangor, which flow through predominantly rice-growing areas along the west coast of Peninsular Malaysia, were chosen for this study.Water samples were collected from different monitoring stations along the two rivers over a 12 month period. Experimental Apparatus An Eyela rotary vacuum evaporator was used for concentrat- ing the sample extracts. A Shimadzu GC 9A gas chromatograph fitted with an electron-capture detector suitable for on-column injection and connected to a Shimadzu CR 3A integrator was employed throughout. The column was 2 m X 3 mm i.d. glass, coated with 1.5% SP-2250-1.95% SP-2401 on Supelcoport (100-120 mesh), obtained from Supelco. Supelclean LC-18 SPE cartridges from Supelco, pre-packed in a polypropylene housing cartridge containing 100 mg of C18-bonded porous silica and with a 0.45 pm filter attached, were used.Reagents The use of high-purity reagents and solvents helped to minimize interference problems. The impurity levels of all solvents and reagents used did not exceed an acceptable blank when subjected to the complete procedure without the sample. The solvents used were obtained from various sources and were of different purities. Doubly distilled water from an all-glass distiller and stored for not longer than 7 d in a glass storage tank was used as reagent water. Baker GC-grade hexane (b.p. 68.5-69.0 "C) of 99.8% purity was used for the preparation of GC samples and Baker pesticide-grade hexane (b.p. 68.0-68.5 "C) of 99.8% purity was used for extraction purposes. Light petroleum (b.p. 60-80 "C) from BDH (analytical- reagent grade, 90% purity) was redistilled in glass twice prior to use.Diethyl ether (b.p. 34-40 "C) from Riedel-de Haen (GP grade, 99% purity) was passed through a chromato- graphic column packed with alumina to eliminate all peroxides and was glass distilled twice prior to use. Ethyl acetate (b.p. 77-77.5 "C) and methanol (b.p. 65 "C) were obtained from May and Baker (Rhone Poulenc) (HPLC grade, 99.8% purity). Freon TF (trichlorotrifluoroethane) (b.p. 42 "C), obtained from Du Pont, was glass distilled twice and collected at 42 "C prior to use as an extraction solvent. All organochlorine pesticide standards used were obtained from the US EPA. All standards came in a stock solution of 100 pg cm-3 of 99.8% purity. For calibration purposes, a portion of these stock solutions was diluted and a series of standard mixtures were prepared. All standard solutions and mixtures were stored in polytetrafluoroethylene (PTFE)-lined screw-capped amber-glass bottles at 4 "C.These solutions and mixtures were checked monthly for signs of degradation or evaporation. Standard mixtures were replaced after 3 months. Sigma pesticide-grade magnesium silicate (Florisil) of 100- 120 mesh size was stored in amber-glass bottles. Merck aluminiumoxid 90 (alumina) of 70-230 mesh ASTM was pre-heated at 400 "C prior to use and stored in amber-glass bottles. Granular anhydrous sodium sulfate (Fluka) of 99%1130 ANALYST, JULY 1992, VOL. 117 purity was purified by heating at 400 "C overnight and stored in PTFE-lined screw-cap glass bottles. Boiling chips were heated at 400 "C prior to use.Procedure The grab sample technique was used to take water samples from the river. Water samples were collected in duplicate in amber-glass bottles and were acidified to pH 2 with sulfuric acid to eliminate biological activity in the water. Liquid-liquid extraction (LLE) All glassware in this procedure was washed with the extraction solvents, hexane and Freon TF. A 250 cm3 volume of river water sample was extracted with 3 x 8 cm3 of extraction solvent. The extracts were combined and collected in a 250 cm3 round-bottomed flask and concentrated using a rotary vacuum evaporator to about 10 cm3 for Florisil column chromatographic clean-up. Florisil column chromatography Florisil column chromatographic clean-up was used for fractionation of organochlorine pesticides from the water samples by elution with solvents of increasing polarity.The chromatographic columns (50 cm x 15 mm i.d.) were slurry-packed with 7 g of Florisil (activated at 450 "C overnight, made into a 1.5% v/m water-Florisil slurry with distilled water and stirred for 3 h before use). Approximately 0.5 cm-3 of anhydrous sodium sulfate was placed at the top of the column to absorb any water in the sample or the solvent. The column was pre-eluted with 60 cm3 of light petroleum. Just prior to exposure of the sodium sulfate layer to the air, the washings from the flask containing the distilled-off solvent were placed in the column and allowed to sink below the sodium sulfate layer. A total of 200 cm3 of light petroleum was poured into the column and the eluate was collected in a round-bottomed flask.Elution was effected with five different solvent mixtures, each a total of 200 cm3 (100% light petroleum, 6% light petroleum-diethyl ether, 15% light petroleum-diethyl ether, 50% light petroleum4iethyl ether and 100% diethyl ether). The eluates were then evaporated to dryness using a rotary vacuum evaporator and the residues were dissolved in 2 cm3 of GC-grade hexane for GC analysis. Solid-phase extraction (SPE) The Supelclean LC-18 cartridges were pre-activated by passing through them two column volumes (2 x 2 cm3) of HPLC-grade methanol followed by three column volumes (3 X 2 cm3) of reagent water. Prior to the exposure of the C18 material to air, 100 cm3 of the river water sample were added via a syringe fitted with a 0.45 ym filter.The cartridge was then air-dried by drawing air through it. These SPE cartridges were eluted with 2 cm3 of HPLC-grade ethyl acetate for GC analysis. GC analysis The separation and determination of the pesticide residues were performed with the Shimadzu GC 9A gas chromato- graph fitted with an electron-capture detector using a mixed- phase glass column (1.5% SP-2250-1.95% SP-2401 on 1W 120 mesh Supelcoport, 2 m x 3 mm i.d.). The injector and detector temperatures were 250 "C. The oven temperature was maintained at 180 "C for 13 min, then programmed at 5 "C min-1 to 190 "C and held there for 3 min before being raised to 200 "C at 10 "C min-1 and held at 200 "C for 11 min. Nitrogen was used as the carrier gas at a flow rate of 50 cm3 min-1. The sample size was 2 mm3.Concentration measurement The direct comparison technique using external standards was chosen for this study because the pesticides identified in the samples clearly matched known organochlorine pesticide standards based on their retention times. The concentration of pesticide residues in the water was calculated using the equation where W = ng of pesticide standard, A, sample, V1 = extract volume (cm3), A2 standard, V2 = volume of extract injected volume of sample extracted (cm3). = peak area of = peak area of (mm3) and V3 = Efficiency of Extraction Methods A 250 cm3 volume of reagent water was spiked with 50 mm3 of 5.0 pg cm-3 of each of the 12 organochlorine pesticide standards. A duplicate of this aliquot was prepared. One aliquot was subjected to solvent extraction with hexane and the other with Freon TF and both were subjected to the entire analytical procedure as described above.This was repeated nine times and a set of data for the efficiency of the extraction method was obtained. Similarly, the extraction efficiency of the SPE method was determined by spiking 100 cm3 of reagent water with 50 mm3 of 5.0 yg cm-3 of each of the 12 pesticide standards, followed by the experimental procedure as described above. This was repeated nine times and a set of data for the efficiency of the SPE method was obtained. The efficiencies of extraction of the pesticides by solvent extraction with hexane and Freon TF and by the SPE method were also determined for spiked river water samples using the same procedure as above.Results and Discussion The gas chromatogram of a mixture of 12 organochlorine pesticide standards is shown in Fig. 1. All 12 pesticides can be resolved and eluted in a reasonable time by employing the GC conditions stipulated above. The efficiencies of extraction of the 12 organochlorine pesticides by solvent extraction with hexane and Freon TF and by the SPE technique are shown in Table 1. Average recoveries exceeding 90% were obtained by solvent extraction with both hexane and Freon TF. These results are comparable to those reported using the US EPA Test Method 608 for organochlorine pesticides via solvent extraction with dichlo- romethane .21 The standard deviations for the recovery of these organochlorine pesticides using solvent extraction with hexane and Freon TF ranged from 4 to 6%.These values are within acceptable limits for reproducible results in quanti- tative analysis. The recovery data using the SPE technique followed by GC of a 2 mm3 aliquot of the ethyl acetate eluate were much lower than those obtained by solvent extraction with hexane and Freon TF (Table 1). Average recoveries ranging from 80 to 87% were obtained for the 12 organochlorine pesticides. However, these results do agree fairly well with those reported by Junk and Richard14 for some of the organochlorine pesticides. The standard deviations for the recovery of these organochlorine pesticides using the SPE technique, ranging from 8 to 11%, are also greater than those obtained with the solvent extraction method. The efficiencies of extraction of the 12 organochlorine pesticides by solvent extraction with hexane and Freon TF and by the SPE technique from river water samples are shown in Table 2.The recoveries of these 12 pesticides spiked in river water ranged from 84 to 103% for the solvent extraction method with both hexane and Freon TF, but were much lowerANALYST. JULY 1992, VOL. 117 1131 n L 9 0 10 20 Retention ti m e/m i n Fig. 1 Gas chromatogram of organochlorine pesticide standards. Analytical column, GP 1.5% SP-2250-1.95Y0 SP-2401 on Supelcoport (100-120 mesh). Carrier gas N2 at 50 cm3 min-l. Column tempera- ture: 180 "C for 12 min, increased at 5 "C min-', then held at 190 "C for 3 min, and finally increased at 10 "C min-l, and held at 200 "C for 12 min. Electron capture detector.Injection volume 2 mm3. Pesticide standards, 100 pg cm-3, each. 1, a-BHC; 2, P-BHC; 3, y-BHC (lindane); 4, aldrin; 5 , heptachlor; 6, endosulfan I; 7, heptachlor epoxide; 8, o,p'-DDE, 9, endosulfan 11; 10, endrin; 11, dieldrin; and 12, p,p'-DDT Table 1 Recovery of organochlorine pesticide standards in reagent water using solvent extraction with hexane and Freon TF and SPE Mean recovery +_ SD (YO)* Organochlorine pesticide Hexane Freon TF SPE wBHC b-BHC y-BHC o,p'-DDE p,p'-DDT Heptachlor Heptachlor epoxide Endosulfan I Endosulfan I1 Endrin Dieldrin Aldrin * It = 10. 96.0 f 4.8 98.0 k 4.7 96.0 f 5.0 101.0 f 4.9 96.0 f 5.4 92.0 f 5.8 96.0 f 4.8 93.0 f 6.1 93.0 f 6.7 105.0 f 4.8 98.0 f 5.2 96.0 f 5.6 94.5 f 5.0 103.0 f 5.1 92.0 f 5.4 96.0 f 5.3 92.0 f 6.1 90.0 f 6.2 93.0 f 5.1 91.0 f 6.5 90.0 f 6.2 98.0 f 5.3 91.0 f 5.6 92.0 f 6.1 87.0-t 8.2 87.0-t 8.4 85.0f 8.0 86.0k 9.8 85.0k 9.2 80.03I 9.9 85.04 9.9 81.0 3I 10.1 80.0 k 10.4 87.0f 9.9 83.0 k 10.6 86.0 3I 10.1 Table 2 Recovery of organochlorine pesticide standards added to river water using solvent extraction with hexane and Freon TF and SPE Mean recovery 4 SD (%)* Organochlorine pesticide Hexane Freon TF SPE a-BHC P-BHC y-BHC o,p'-DDE p,p'-DDT Heptachlor Heptachlor epoxide Endosulfan I Endosulfan I1 Endrin Dieldrin Aldrin * n = 10.93.0 f 4.2 97.0 f 4.3 95.0 f 4.0 103.0 5 3.9 85.0 f 4.0 88.0 f 3.8 95.0 f 4.0 91.0 f 4.2 93.0 f 3.8 101 .O k 4.0 98.0 f 3.8 96.0 f 5.2 91.5f4.6 83.0-t 8.6 100.0f4.2 83.0-t 9.2 90.054.6 78.0f 8.6 95.0 +_ 4.2 82.0 -t 10.2 84.0 f 4.3 79.0 1- 10.6 88.0 f 4.4 73.0 k 11.8 91.0k4.4 78.031 8.8 90.0 f 4.5 76.0 f 10.6 90.0 -t 4.2 73.0 k 11.2 98.0 f 3.8 83.0 f 11.6 93.0 f 4.2 78.0 k 10.0 90.0 f 3.1 79.0 3I 10.6 with hexane or Freon TF.The SPE technique using CIS- bonded porous silica in cartridges has also been reported to give rise to extraneous peaks in the gas chromatograms of ethyl acetate eluates.20 In view of the above, the solvent extraction method was chosen for subsequent monitoring studies of these organo- chlorine pesticide residues in river water samples. As the recovery data and standard deviations for solvent extraction with Freon TF and hexane were similar (Table l), hexane was chosen because less evaporative losses occurred under the local laboratory conditions. The results of the monitoring studies for organochlorine pesticides in water samples from two rivers, Sg.Bernam and Sg. Selangor, along the west coast of Peninsular Malaysia, from June 1989 to July 1990 are shown in Table 3. The organochlorine pesticides that were identified in the water samples from stations along the two rivers were DDT, BHC, endosulfan, heptachlor and dieldrin. The presence of these organochlorine pesticides can be attributed to their wide usage during the 1950s and 1960s especially in the rice fields. These organochlorine pesticides degrade very slowly and accumulate in the soil of the rice fields and are subsequently leached out into the aquatic system of the surrounding areas. This situation still exists even though most of these organochlorine pesticides have been banned and discontinued from usage in the rice fields for many years.The residue levels of organochlorine pesticides in these river waters are high as regards their influence on aquatic life. For example, the total DDT level found in the water from Sg. Bernam ranges from 180 to 206 ng dm-3 and from Sg. Selangor it ranges from 60 to 190 ng dm-3. Both of these levels exceed the critical level tabulated in the Malaysian Interim Standards for aquatic life, viz., 4 ng dm-3.22 The levels of endosulfan in Sg. Bernam water ranges from 32 to 92 ng dm-3 and in Sg. Selangor water from 180 to 500 ng dm-3. Both of these levels also exceed the critical level of 10 ng dm-3 for aquatic life.22 Similarly, the levels of BHC, dieldrin and heptachlor in both river waters exceeded the critical levels for aquatic life.(73.0-83.0%) when the SPE method was employed. The standard deviations for the recovery using the SPE technique are also greater than those obtained with the solvent extrac- tion method. The values are comparable to those obtained for reagent water. The results show that the SPE method for sample enrich- ment of organochlorine pesticides in environmental water samples would lead to less reliable data, based on the recovery and standard deviation, than the solvent extraction method Conclusion These studies on organochlorine pesticides show that despite the numerous claims by both manufacturers and research groups on the usefulness of SPE cartridges for sample preparation and trace enrichment in environmental samples,14,16~19~20 solvent extraction with hexane or Freon TF can be a more accurate alternative method for organochlorine pesticide residue determinations in environmental waters .1132 ANALYST, JULY 1992, VOL.117 Table 3 Organochlorine pesticide levels in Sg. Bernam and Sg. Selangor river water (from June 1989 to July 1990; average of four data sets) Pesticide concentratiodng dm-3* Distance from River estuary/km t-BHC t-DDT t-Hept t-Endosulfan Dieldrin Total Sungai Bernam 17.6 270 206 97 32 57 662 162 370 180 140 92 22 804 Sungai Selangor 5.0 260 90 90 260 15 715 30.0 450 190 130 500 72 1342 75.4 130 60 90 180 50 510 * t-BHC = a-BHC + 6-BHC + y-BHC; t-DDT = o,p’-DDE + p,p’-DDT; t-Hept = heptachlor + heptachlor epoxide; t-Endosulfan = endosulfan I + endosulfan 11. The author thanks the Malaysian government for financial support of this work provided under R&D IRPA Grant No.4-07-04-045. References 1 Junk, G. A., Adv. Chem. Ser., 1987, 214, 201. 2 Bogus, E. R., Gallagher, P. A., Cameron, E. A., and Mumma, R. O., J. Agric. Food Chem., 1985,33,1018. 3 Puyear, R. L., Fleckenstein, K. J., Montz, W. E., and Brammer, J. D., Bull. Environ. Contam. Toxicol., 1981, 27, 790. 4 Saner, W. A., Jadamec, J. R., Sager, R. W., and Killen, T. J., Anal. Chem., 1979,51, 2180. 5 Chladek, E., and Marano, R. S., J. Chromatogr. Sci., 1984,22, 313. 6 Steinheimer, T. R., and Ondrus, M. G., Anal. Chem., 1986,58, 1839. 7 Renberg, L., and Lindstrom, K., J. Chromatogr., 1981, 214, 327. 8 Thome, J. P., and Vandaele, Y., Int. J. Environ. Anal. Chem., 1987, 29,95. 9 Lawrence, J. F., Organic Trace Analysis by Liquid Chromato- graphy, Academic Press, New York, 1981. 10 Frei, R. W., and Brinkman, U. A. Th., TrAC, Trends Anal. Chem. (Pers. Ed.), 1981, 1, 45. 11 Bushway, R. J., J. Chromatogr., 1981,211, 135. 12 13 14 15 16 17 18 19 20 21 22 _- Gay, D. D., and Lahti, R. A., Znt. J. Pept. Protein Res., 1981, 18, 107. Kraak, J. C., Smedes, F., and Meijer, J. W. A., Chromato- graphia, 1980, 13,673. Junk, G. A., and Richard, J. J., Anal. Chem., 1988,60,451. West, S . D., Dorulla, G. K., and Poole, G. M., J. Assoc. Ofj Anal. Chem., 1983, 66, 111. Drevenkar, V. Z., Frobe, F., Stengl, G., and Tkalcevic, B., Mikrochim. Acta, 1985, 143. Bardalaye, P. C., and Wheeler, W. B., Znt. J. Environ. Anal. Chem., 1986,25, 105. Hoke, S. H., Brueggemann, E. E., Baxter, L. J., andTrybus, T. J., J. Chromatogr., 1986,357,499. Richard, J. J., and Junk, G. A., Mikrochim. Acta, Part I , 1986, 387. Junk, G. A., Avery, M. J., and Richard, J. J., Anal. Chem., 1988, 60, 1347. US EPA Test Method 608, Organochlorine Pesticides and PCBs, EMSL-US EPA, Cincinnati, OH, 1984. Goh, S. H., Lim, R. P., and Yap, S. Y., Water Quality Criteria and Standards for Malaysia, DOE and Institute of Advanced Studies, University of Malaya, Kuala Lumpur, 1986, vol. 4, p. 59. Paper 1 I063366 Received December 18, 1991 Accepted February 3, 1992

ISSN:0003-2654    

DOI:10.1039/AN9921701129

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, JULY 1992, VOL. 117 1133 Synthesis of Tin(iv) Vanadopyrophosphate as a Novel Stationary Phase for High-performance Liquid Chromatography and Its Application to Amino Acid Analysis Yao Xing-Dong,* Liu Jin-Chun,t Xia Jin-Lan, Cheng Jie-Ke and Zeng Yun'e Department of Chemistry, Wuhan University, Wuhan 430072, China The preparation and ion-exchange properties of a novel microcrystalline ion exchanger, tin(tv) vanadopyro- phosphate (SVPP), were studied. The exchanger has a high ion-exchange capacity (3.17 mequiv g-I), and its chemical and thermal stabilities are excellent. It was used as a stationary phase for the high-performance ion-exchange chromatographic separation o f amino acids. Optimum conditions for high-performance liquid chromatography were investigated. Five amino acids were separated rapidly on a short SVPP column.Keywords: Tin( IV) vanadopyrophosphate; high-performance liquid chromatography; amino acid During the last few decades, many important advances have been made in studies of inorganic ion exchangers. These materials have, on the other hand, rarely been utilized as stationary phases for high-performance liquid chromato- graphy (HPLC) owing to their poor mechanical stability and pressure-resistance properties. Therefore, an important aim in studies of inorganic ion exchangers is to improve the mechanical properties of these exchangers so that they can be used for HPLC.1 A series of pyrophosphate inorganic ion exchangers, such as tin(1v) pyrophosphate2 and tin(1v) selenopyrophosphate,3 have been synthesized. Both of these exchangers have excellent thermal and chemical stability with large ion- exchange capacities (IECs).It was concluded from many experiments that higher valencies, lower relative atomic masses and higher w values (where 4' = Z/R; and Z = the formal charge of the central cation and R = the covalent radius of the central cation plus the covalent radius of oxygen) of the central atoms in the exchangers would be beneficial in increasing the IEC values of these material^.^ Tin(1v) vanadophosphate, a heteropolyacidic exchanger with a maximum IEC of 1.70 mequiv 8-1 and good thermal and chemical stability, has been synthesized by Qureshi and Kaushik.5 If the phosphate group in this exchanger were to be substituted by a pyrophosphate group, the exchanger thus produced would beexpected to have better ion-exchange properties, as its w value would be raised considerably.In this paper a novel ion exchanger, tin(1v) vanadopyro- phosphate (SVPP), is reported. The optimum conditions for its synthesis and its ion-exchange behaviour were investigated in detail. The results show that SVPP is a polyfunctional microcrystalline exchanger with a large IEC and good thermal and chemical stability. It also has excellent pressure-resistant properties. The exchanger has previously been applied to the determination of trace amounts of lead in human hair by HPLC.6 In the present work, five amino acids were separated rapidly on a short SVPP column. Experimental Apparatus and Reagents The liquid chromatograph used consisted of a Shimadzu LC-6A equipped with a Shimadzu RF-530 fluorescence monitor.Post-column derivatization with o-phthalaldehyde (OPA) was employed for detection of the amino acids. * Present address: Department of Chemistry, Santou University, t To whom correspondence should be addressed. Santou, Guangdong 515063, China. Spectrophotometric measurements were made on a Shi- madzu UV-3000 spectrophotometer. The structure of the exchanger was investigated with a JEM-100 CX11 electron microscope. Infrared (IR) and laser Raman spectra of SVPP were measured using a Nicolet 170-sx Fourier transform infrared (FTIR) spectrometer and a Jobin-Yvon U-1000 laser Raman spectrometer. A Shimadzu Model DT-30B thermal analyser was used for both differential thermal analysis (DTA) and thermogravimetric analysis (TGA).The reagents SnC14.5H20, K2P207-3H20, NH4V03, NaOAc and methanol were all of analytical-reagent grade (Shanghai Second Reagent Company), whereas OPA and 2-mercaptoethanol were of chemical-reagent grade (Beijing Chemical Reagent Company). The concentration of the stock solutions of the amino acids was 1.0 mg cm-3; the stock solutions were diluted to 200 pg cm-3 for analysis. Preparation of SVPP A 1.0 rnol dm-3 K2P207 solution was mixed with a 0.05 mol dm-3 NH4V03 solution. Then, a 1.0 mol dm-3 SnC14 solution was slowly added to the mixture with continuous stirring. The mixing ratios are listed in Table 1. The pH of the mixture was adjusted to 1.5 with 1 rnol dm-3 KOH (or 1 mol dm-3 HN03). After stirring for 1 h, the mixture was allowed to stand for 24 h.All the above procedures were carried out at a temperature of less than 15 "C. The precipitate was then filtered under suction and washed with distilled water until the pH of the washings was 5 . The product was dried at 40 "C and then 'cracked' into small granules by the addition of cold water and converted into the H+ form with 1 .O rnol dm-3 HN03 (about 20 cm3 of HN03 per gram of exchanger). The precipitate was washed with water until the washings were of neutral pH. It was then dried at 40 "C, ground and sieved (300 mesh). Column Packing Procedure The high-pressure balanced-density technique was employed for column packing. Glyceryl alcohol-methanol (1 + 1) was chosen as the slurry-packing solvent and methanol as the solvent. A 1.5 g amount of SVPP was packed into a steel column (100 x 4.6 mm i d .) at a pressure of 3.04 X 107 Pa, and the column was pre-balanced with sodium acetate buffer solution prior to the analysis. Results and Discussion The effect of the conditions used for the synthesis of SVPP on both the chemical composition and ion-exchange properties of1134 ANALYST, JULY 1992, VOL. 117 Table 1 Effect of the conditions of the synthesis on the chemical composition and IEC values of SVPP Chemical Sn : V : P2 time/h Sn : V: P2 mequivg-1 Exchanger Mixing ratio, Ageing composition IEC/ No. 1 2 3 4* 5 6 7 8 9 10 11 * No precipitate. 2:l:l 2:1:2 3:1:2 1:l:l 10 : 1 : 10 10:1:20 15 : 1 : 20 20:1:20 10: 1 : 25 10: 1 :20 10:1:20 24 24 24 24 24 24 24 24 24 36 48 7.00 : 1.00 : 3.20 7.10: 1.00:3.65 7.50 : 1.00 : 3.66 19.00 : 1.00 : 10.00 30.00 : 1.00 : 20.00 24.40 : 1.00 : 14.30 24.40 : 1.00 : 14.00 30.00 : 1.00 : 20.20 30.00 : 1.00 : 19.80 - - 1.957 1.971 1.964 1.978 2.000 1.985 1.982 2.000 2.000 - - 1.83 2.15 2.05 2.38 3.17 2.68 2.64 3.15 3.10 - - 4000 3200 2400 1600 800 Q v/cm-' Fig.1 Infrared spectrum of SVPP the products is shown in Table 1. In order to determine the chemical composition of the exchanger, iodimetry,7 titration with ammonium molybd~phosphate~ and a spectrophoto- metric method with phosphatovanadotungstate7 were employed for determining Sn, P and V, respectively. It was obvious that the lower the Sn:P ratio in the exchanger, the higher would be its value, and the higher its IEC. Samples 6,lO and 11 (Table 1) were found to have the highest IEC values; the effect of ageing time on the IEC values of the exchangers is negligible.Therefore, sample 6 was chosen for further studies. The X-ray diffraction pattern8 and the electron diffraction pattern showed that SVPP has a microcrystalline structure. In the IR spectrum of SVPP (Fig. l), the absorption bands at 509, 750 and 1057 cm-1 are due to Sn-0, V-0 and P-0-P stretching vibrations, respectively. The two strong absorption bands at 1600 and 3200-3600 cm-1 (wide band) are obviously due to the OH stretching vibrations of the structural water (water of crystallization), the hydroxy groups in the exchanger and the water absorbed on the surface of SVPP. The existence of the three bonds was also verified by laser Raman spectrometry (bands at 500 cm-1 for Sn-O; 832,624 and 729 cm-1 for V-0; and 1044,1072 and 758 cm-1 for P-0-P).The TGA-DTA curves of SVPP showed that the percent- age mass loss of SVPP was lower than that of tin(1v) vanadophosphate in the range from room temperature to 500 "C. The effect of drying temperature on the IEC of SVPP is summarized in Table 2. A higher IEC was found for SVPP compared with tin(1v) vanadophosphate at temperatures up to 250 "C. The data showed that the thermal stability of SVPP was considerably better than that of tin(1v) vanadophosphate, as expected. The chemical stability of SVPP was determined by the method described in ref. 3. The contents of P, V and Sn in solution were determined spectrophotometrically with Phos- phatomolybdic Blue,9 phosphatovanadotungstate7 and ben- Table 2 Effect of drying temperature on the IEC (K+) of SVPP TemperaturePC 40 100 150 200 250 300 400 IEC(K+)/rnequivg-l 3.17 2.99 2.58 1.50 0.55 0.13 0.10 zofluorone,lO respectively.It was found from the solubilities of the three elements (Table 3) that SVPP is very stable in HN03, H2SO4 or citric acid medium. The order of the chemical stability of SVPP in different media is: citric acid > HN03 > H2SO4 > HCI > H20 > NaOH. The pH titration curve of SVPP shown in Fig. 2 was obtained by the method described by Topp and Pepper.11 It shows that SVPP is a polyfunctional group exchanger, and that H+ in the exchanger dissociates over a wider pH range than with other tin-based inorganic ion exchangers. The ion- exchange reactions take place mainly in the first dissociation step.On the basis of chemical composition, pH titration curve, thermogravimetric and IR studies, and the law of electro- neutrality, the following chemical formula is proposed for SVPP: [3Sn0 1.90H2H2P207 0.1V03];6.5H20. The SVPP column showed excellent pressure-resistant properties. No breakdown of the structure of SVPP was found at pressures up to 3.04 x 107 Pa, and the column is sufficiently stable for high-performance ion-exchange chromatography (HPIEC) under the conditions used here for HPLC; this is due to the microcrystalline structure of SVPP. In contrast, the mechanical strength of amorphous inorganic materials is usually poor. They tend to break into smaller pieces under high pressure. Therefore, they cannot be utilized as an HPIEC stationary phase.Effects of Chromatographic Conditions on Retention of Amino Acids As SVPP is an amphiprotic ion exchanger, the -OH groups present can dissociate in two ways: P-OH P-0- + H+ (cation exchange) P-OH S P+ + OH- (anion exchange) where P is the structural network of the exchanger. Amino acids are also amphiprotic compounds. The form in which they exist strongly depends on the pH of the solution: at high pH the amino acid will be in the anionic form, whereas at low pH it will be in the cationic form. At a pH near the isoelectric point (pl), amino acids exist in a neutral form in solution. Our experiments showed that each amino acid was retained on the SVPP column from both weakly acidic and weakly basic solutions. As the chemical stability of SVPP in acidic solution is better than that in basic solution, the weakly acidic solution was chosen for subsequent studies.ANALYST, JULY 1992, VOL.117 1135 Table 3 Solubilities of SVPP in different media (mg per 50 cm3) Solvent/mol dm-3 ~~~ ~ H2O HN03 HCl H2S04 Citric acid NaOH Element - 0.10 1.0 4.0 0.10 1.0 4.0 1.0 2.0 1.0 2.0 0.10 Sn 0.003 0.018 0.023 0.024 0.014 0.10 0.34 0.091 0.14 0.077 0.010 0.35 V 0.075 0.030 0.18 0.33 0.015 0.15 0.45 0.15 0.23 0 0 0.12 P 5.93 3.44 0.050 0.020 3.16 0.080 6.72 0.18 0.060 0 0 11.86 ~~~~~~~~~~~~~ ~~ ~ ~ ~ ~ ~ ~ Table 4 Effect of pH of mobile phase on k' of amino acids k' pH Gly Ala Leu Ile Val Ser Met Asp Phe Tyr Trp Glu 4.1 39 10.10 2.57 2.79 2.81 45 3.11 2.57 2.07 2.04 2.14 4.57 4.5 24.4 5.43 1.49 1.57 1.71 - 1.86 2.00 1.43 1.50 1.50 3.00 5.3 1.66 0.57 0.29 0.31 0.33 1.14 2.36 2.29 0.30 0.34 0.33 0.30 6.0 6.19 1.14 0.43 0.43 0.49 1.69 0.49 0.37 0.40 1.00 1.00 0.36 12 1 10 8 I a.6 4 2 0 4.0 8.0 12.0 [OH - ]/mequiv g - I Fig. 2 Titration curve of SVPP: A, NaCI-NaOH; and B, KCl-KOH J Various buffer solutions of different pH, viz., acetate, citrate, tartrate and their mixtures with some organic solvents, such as methanol, ethanol and acetone, were used as eluent. The retention of the amino acids is mainly affected by the pH of the eluent, as shown in Table 4. For most amino acids, the lowest capacity factors (k') were found at pH 5.3, which coincided with their p l values. These results show that the retention of amino acids on the SVPP column is based on ion exchange. The k' values would increase slightly with an increase in the pH of the eluent.In contrast, a marked increase would be found as the pH decreases from the pl. The lower the pH, the stronger the adsorption. These phenomena are all due to the formation of +NH3RCOOH or NH2RCOO-; the former is adsorbed more strongly on SVPP. The adsorption sequences for different types of amino acid are: Ser > Gly > acidic amino acids > amino acids containing sulfur > fatty amino acids > aromatic amino acids, this sequence is obviously due to steric effects from the size of the amino acids. The greater the size of the amino acid, the weaker the adsorption on SVPP. However, coordination may occur between Sn4+ in SVPP and the -SH or -OH groups in the amino acids, and the k' values of these amino acids, such as Ser and Met, would be higher than those of other amino acids.The maximum separation factor between two neighbouring amino acids was obtained at pH 4.5. As the retention of amino acids on SVPP is mainly controlled by the pH of the mobile phase, as described above, and as the affinity of microcrystalline and crystalline inorganic ion exchangers is higher for H+ than for alkali metal ions, it 0 3 6 9 12 15 18 21 Fig. 3 Chromatogram of five amino acids: 1, Leu; 2, Met; 3, Glu; 4, Ala; and 5, Gly. Eluent, 10 mmol dm-3 NaOAc buffer solution (pH 4.5); flow rate, 1 cm3 min-1; temperature, 30 "C; column, SVPP (in Na+ form) 100 X 4.6 mm i.d. Amount of each amino acid injected, 4 1.18 was to be expected that the concentration of the buffer solutions had little effect on the retention of the amino acids.A 10 mmol dm-3 NaOAc buffer solution (pH 4.5) was chosen as the eluent for the separation of amino acids. The effect of flow rate on k' was also studied. It was found that the k' values and the analytical period (the retention time required for elution of all the amino acids) for the amino acids decreased with increasing flow rate. However, both the column efficiency and the separation factor between two neighbouring amino acids decreased simultaneously. It was found that with a flow rate equal to or greater than 0.8 cm3 min-1, the separation factors decreased sharply, and most amino acids could not be separated. On the other hand, with a flow rate of less than 0.5 cm3 min-1, the chromatographic peaks were very broad.Therefore, the optimum flow rate of the eluent was taken to be 0.5 cm3 min-1. The effect of the temperature in the range 30-60 "C on the separation of the amino acids was studied. A higher column temperature was beneficial for the separation of the amino acids. However, as higher temperatures affected the stability of the exchanger in the mobile phase, the separations were performed at 30-40 "C. Typical results are shown in Fig. 3.1136 ANALYST, JULY 1992, VOL. 117 References 1 Yao, X. D., Liu, L. B., Liu, J. C., and Cheng,. J. K., Zon Exch. Adsorp., 1988,4459. 2 Xu, M., Liu, X., Liu, J. C., and Cheng, J. K., Zhongguo Xitu Xuebao, 1985,3 (3), 82. 3 Liu, X., Liu, J. C., and Cheng, J. K., Sep. Sci. Technol., 1989, 24, 63. 4 Yao, X. D., Liu, L. B., Liu, J. C., and Cheng, J. K., Wuhan Daxue Xuebao, Ziran Kexueban, 1989, (2), 77. 5 Qureshi, M., and Kaushik, R. C., Anal. Chem., 1977,49, 165. 6 Xia, J. L., Liu, J. C., Liu, L. B., and Cheng, J. K., Anal. Lab., 1990, 9, 5 . 7 Furman, N. F., Standard Methods of Chemical Analysis,Van Nostrand, Princeton, NJ, 6th edn., 1963, vol. 1. 8 Yao, X. D., Liu, J. C., Cheng, J. K., and Zeng, Y., Zon Exch. Adsorp., 1990, 6 , 107. 9 Snell, F. D., Photometric and Fluorometric Methods of Analy- sis. Nonmetals, Wiley, New York, 1981. 10 Charlot, G., Colorimetric Determination of Elements, Elsevier, Amsterdam, 1964. 11 Topp, N. F., and Pepper, K. N., J. Chem. SOC., 1949,3299. Paper 0102902E Received June 27, 1990 Accepted March 12, 1992

ISSN:0003-2654    

DOI:10.1039/AN9921701133

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, JULY 1992, VOL. 117 1137 Comparison of Ion Chromatographic Methods for the Determination of Organic and Inorganic Acids in Precipitation Samples Venghuot Cheam National Water Research Institute, Research and Applications Branch, Box 5050,867 Lakeshore Road, Burlington, Ontario, Canada L 7R 4A6 A comparison of three ion chromatographic methods for the determination of organic and inorganic acids in precipitation samples has been made, in order to identify which method or combination of methods is most suitable for the determination of these acids. If only one method is to be used, the gradient method is the most suitable as it allows the determination of nearly all the acids of interest. However, to cover adequately all the common acids, the best solution is to use a combination of gradient and ion-exclusion methods, as the gradient method provides the most reliable results for formic, hydrofluoric and methanesulfonic acids, and the ion-exclusion method provides the best results for acetic and lactic acids in addition to giving corroborating analyses of some of the other organic acids of interest.The inorganic acids are adequately determined by the gradient method. This combination is preferred over the isocratic-ion exclusion combination as the latter would not allow satisfactory determination of the complete range of acids of interest. The detection limits for the gradient (0.014.08 ppm) and isocratic (0.01-0.25 ppm) methods are comparable for most acids but are lower than those obtained by the ion-exclusion method (0.1-0.4 ppm).Overall, the precision of the results generated by the isocratic method is better than that for the gradient method, which is in turn better than the precision of the ion-exclusion method. The recoveries obtained by the three methods are within 100 & 10% except for some organic acids determined by the ion-exclusion method. Keywords : Method comparison; ion chroma tog rap h y; gradient; isocra tic; ion -exclusion The importance of organic acids in precipitation samples has been well documented.1-10 For example, it has been reported that over 60% of the free acidity in remote sites in Venezuela and Australia is due to formic and acetic acids and that in Northern Territory, Australia, formic and acetic acids make up about 80% of total acidity in both cloud droplets and rainwater.1-3 There are numerous organic and inorganic acids present in precipitation samples and both acid types need to be determined accurately for proper accounting of atmospheric chemistry processes and precipitation ionic balances. The major and minor acids commonly found and reported are shown in Table 1. The natural tendency is to develop a single method that can determine all of the acids. A gradient anion-chromatographic method" has been developed (referred to as the gradient method) that can virtually achieve that aim. However, there were some weaknesses: co-elution of acetic and lactic acids and occasional poor resolution of some early peaks of organic monoprotic acids. Notwithstand- ing, it is a valuable method. However, it is questionable whether it would be the method of choice for a thorough analysis of the acids of interest. Given the weaknesses of the gradient method and the myth among analytical chemists that a gradient method is routinely not as reliable as an isocratic method, an alternative may be preferred: for example, using two different methods, one mainly for organic acids (for example, an ion-exclusion chromatographic method, referred to as the ion-exclusion method) and the other mainly for inorganic acids (for example, an isocratic anion-chromatographic method, referred to as the isocratic method).This paper will address these alternatives and compare the three methods. By use of real, relevant water samples and by assessing the analytical data generated, pertinent to method characteristics, it will be shown that the myth is unfounded and that the gradient method combined with the ion-exclusion method is superior to the combination of isocratic and ion-exclusion methods for determining the complete range of acids of interest. Experimental Reagents High-purity reagents were used throughout together with Milli-Q water (18 MQ; Millipore, Bedford, MA, USA).For the ion-exclusion (IE) method, the following reagents were used: HC1 (Baker Instra-Analyzed, Phillipsburg, NJ, USA), octanesulfonic acid (OSA) (0.1 ma1 dm-3 solution) (Dionex, Table 1 Acids commonly reported in precipitation-related samples Acid Formic Acetic Oxalic Glycolic Propionic Lactic Butyric Citric MSt HOMSt Sulfuric Nitric Hydrochloric Hydrofluoric Phosphoric Nitrous Concen- tration range* (Yo ) Formula 0.7-60 HC02H 0.3-30 CH3COzH 0-1.6 HOzCC02H 0-3 HOCH2C02H 0-6.4 CH3CH2COZH 0-3 CH3CHOHC02H 0-0.8 CH~(CH~)ZCO~H 0-0.9 HOC(CH~CO~H)~COZH 0-4.7 CH3S03H 0-10 HOCHzS03H 3-53 HN03 8-45 H2SO4 0-59 HCl 0.5-7 HF 0-15 H3P04 0-0.9 HNOz Refs. 4,6-9,11,14-18 4,6-9,11,15-18 4,9,11 11,16 16,17 16 16,18 8 , l l 15 16 -$. -$.-$. --$ --$ -$. * Approximate range (YO of total organic and inorganic acids) compiled from cited and some uncited references for many different types of precipitation samples around the world. t MS = Methanesulfonic acid; HOMS = hydroxymethylsulfonic acid. $. References are given for organic acids only, as inorganic acids are already well documented in the literature.1138 Load/i n ject valve ANALYST, JULY 1992, VOL.117 - Analytical Micromem brane - 1 , - column suppressor \ Sunnyvale, CA, USA), tetrabutylammonium hydroxide (TBAOH) (0.1 mol dm-3 solution) (Dionex), butyric acid (BDH, Poole, Dorset, UK) , sodium hydroxymethylsulfonate (Dr. H. Puxbaum, Technical University, Vienna, Austria), and the following sodium or potassium salts (Fluka, Buchs, Switzerland; Fisher Scientific, Fair Lawn, NJ, USA; and BDH AnalaR): citrate, fluoride, glycolate, lactate, formate, acet- ate, propionate, fluoride, methanesulfonate, chloride, nitrite, nitrate, sulfate and phosphate. For the isocratic method, the reagents used were: sulfuric acid (BDH Aristar), butyric acid (BDH), sodium hydroxymethylsulfonate (Dr. H. Puxbaum) , and the following sodium and potassium salts (Fluka, Fisher Scientific and BDH AnalaR): carbonate, hydrogen carbon- ate, chloride, nitrite, nitrate, methanesulfonate, sulfate, oxalate, phosphate, citrate, fluoride, glycolate, lactate, form- ate, acetate and propionate. The chemicals used in the gradient method are given elsewhere.11 A stock solution of lo00 ppm (mg dm-3) was prepared for each acid and the organic acids were preserved with 0.2% high-performance liquid chromatographic grade CHCI3.All standards and natural unspiked or spiked samples were preserved likewise. Equipment and Operating Conditions Ion-exclusion method The equipment consists of a Dionex ion-chromatographic system with an APM analytical pump, separator and suppres- sor columns, CDM-1 conductivity detector and a data processing system using the autoion 400 software (Dionex).The APM continuously pumps the eluent through the whole system, which is shown schematically in Fig. 1. The suppressor column chemically converts the highly conductive species of the eluent (HC1) into the less conductive ion pair species of TBA+/Cl- (effective background conductivity =99 pS) , thus enhancing the detector sensitivity for the analytes. After the system reaches equilibrium, a test sample is loaded and injected via the loadhject valve into the path of the eluent, which carries the sample into the separator column; here some analytes (charged species) are ‘excluded’ based on the Donnan exclusion phenomenon, whereas others are ‘allowed’ to permeate a semi-permeable membrane into the pores of the resin, where they are separated according to factors such as pK,, eluent pH and resin type.At the suppressor the resolved analytes displace C1- and become conductive species of TBA+/analyte- ion pairs, which are then detected by the CDM-1 conductivity detector; the detection signals are converted to digital signals by the interface module (CIM, Dionex) and stored in an AT computer (IBM, Boca Raton, FL, USA). The A1400 software controls the pump, the valve and the detector via CIM, and permits the treatment of stored analytical data or sends results of interest to a printer (Epson FX-86e; Epson America, Torrance, CA, USA) for hard copies. The operating conditions are given in Table 2. k 1 - t 1 Waste Printer Eluent Fig. 1 Schematic diagram of the system Isocratic method The same equipment as described above is used for this method, but the column types, eluents, regenerant and operating conditions are different and are described in Table 3. The equipment set-up and operating procedures are similar to those described above.Here, the separation mechanism is based on the exchange of anions between the mobile eluent and the stationary cationic sites and on the different affinities of anions towards the fixed exchange sites. The suppressor chemically converts the eluent Na2C03-NaHC03 into car- bonic acid species (effective background conductivity = 18 ps) and the analytes into free acid species. The equipment and operating conditions and the optimiza- tion process for the gradient method are given in ref. 11, where a gradient pump and NaOH eluents were used. Here, the anion-exchange mechanism is dealt with in a similar manner as in the isocratic method, but the NaOH eluent is converted into water (effective background conductivity --3 pS); the detected analytes are also free acid species. Optimization and Eluents Ion-exclusion method Either 1 mmol dm-3 HCI or OSA can be used as the eluent.Various flow rates ranging from 0.5 to 1.2 cm3 min-1 were tested. Any one of the flow rates could be used to resolve the acids, but in this work 0.8 cm3 min-1 is the optimum choice, giving a reasonable analysis time and acceptable resolution for the four closely eluted peaks (HF, lactic, glycolic and formic acids). However, these four peaks are sometimes not well resolved regardless of the flow rate, giving rise to imprecise results.Eight acids can be resolved: citric, HF, lactic, glycolic, formic, acetic, propionic and butyric acids, whereas the inorganic acids, methanesulfonic, hydroxymethylsulfonic, and oxalic acids are co-elutants in peak 1 or elute in the water dip region (Fig. 2). The OSA eluent produces a better baseline (less noise) and a better detection limit than does the HCI eluent. However, with OSA there is a large unwanted peak co-eluting with the propionic acid peak, which is unacceptable as propionic is one of the acids of interest. A run of a pure C03*--HC03- solution shows that the carbonic acid peak has, for all practical purposes, an identical retention time with the propionic acid Table 2 Operating conditions for IE chromatography Eluent 1 mmol dm-3 HCl Suppressor AMMS-ICE (Dionex) Flow rate 0.8 cm3 min-1 Regenerant/flow rate 5 mmol dm-3 TBAOH*/2 cm3 min-1 Sample loop 50 mm3 Detector Conductivity, CDM-1 Separator Analytes HPICE-ASl(250 x 9 mm i.d.) (Dionex) Citric, hydrofluoric, lactic, glycolic, formic, acetic, propionic and butyric acids * Tetrabutylammonium hydroxide.Table 3 Operating conditions for isocratic anion chromatography Eluent Flow rate Sample loop Guard column Separator Suppressor Regenerantlflow rate Detector Analytes Na2C03-NaHC03 (either one of the three mixtures in mmol dm-3): 2.6013.27; 2.3612.98 ; 1.7712.23 2.0 cm3 min-1 50 mm3 HPIC-AG4A (50 x 4 mm i.d.) (Dionex) HPIC-AS4A (250 X 4 mm i.d.) (Dionex) Anion MicroMembrane Suppressor (AMMS) 0.0125 mol dm-3 H2S04/2 cm3 min-l Conductivity, CDM-1 Methanesulfonic, hydrochloric, nitrous, phosphoric, nitric, h ydroxymet h ylsulfonic , sulfuric and oxalic acidsANALYST, JULY 1992, VOL.117 7.70 -0.69 1139 ( d) I I I I - I NO3- 7 8 2 3 I1 I I I I I peak for both eluents, but the peak height with OSA is higher than that with HCl. As shown in Table 4, at the propionic acid elution peak at 14.1 min, the peak height with OSA for Milli-Q water is 1.2 x 106, which is significantly higher than zero, the peak height with HCl for Milli-Q water. The peak height should be zero. Likewise, the peak heights using OSA, 2.0 2 . >. c .- .g 1.0 u U 0 0 2 7 0 4 8 12 16 Ti me/mi n Fig. 2 Chromatogram of a standard generated by the IE method: 1, inorganic acids (HCl, H2SO4, HN03. HN02, H3PO4), CH3S03H, HOCH2S03H and oxalic acid; 2, citric acid (5 ppm); 3, HF (1 pprn); 4, lactic acid (2.5 ppm); 5, glycolic acid (2.5 pprn); 6, formic acid (2.5 ppm); 7, acetic acid (5 pprn); 8, propionic acid (5 ppm); and 9, butyric acid (5 ppm) 0 7.70 -0.69 7.70 -0.69 for the standard and rain samples, are larger than those with an HCl eluent by about 1.2 x 106.Thus the peak heights with the OSA eluent are apparently biased high. Furthermore, the analysis of the three samples by the gradient method gives a value of zero for Milli-Q water and small peaks for both the standard and rain samples, which agrees with the results from the HC1 runs. Hence, it seems that the unwanted peak observed with the OSA eluent is the carbonic acid peak, as observed earlier by Franklin.12 Therefore, an HC1 eluent was chosen in spite of its noisier baseline and higher background conductivity of 99 yS (versus 43 yS for OSA).Table 4 Peak height comparison between OSA and HCl eluents at the propionic acid elution peak in the IE method Peak height (arbitrary units) Sample tRlmin OSA HCl 0 0 Eluent - Milli-Q water 14.12 1.20 x 106 0 Standard 14.11 1.56 x 106 0.58 x 106 Rain sample 14.14 1.42 x 106 0.20 x 106 1 I I I 7.70 - 8 -0.69 I I I Time/min Fig. 3 Elution patterns at four different eluent strengths: (a) 1.1 X standard eluent; (b) standard eluent (1 g of Na2C03 + 1 g of NaHCO,); (c) 0.75 x standard eluent; and (d) 0.65 x standard eluent. 1, Methanesulfonate (2.5 ppm); 2, C1- (0.5 pprn); 3, NO2- (0.5 ppm); N03- (1 ppm); HP042- (1 pprn); 6, HOMS (5 ppm); 7, S042- (2.5 ppm); and 8, oxalate (2.5 ppm)1140 ANALYST, JULY 1992, VOL. 117 Weaker acids such as benzoic or tridecafluoroheptanoic acid may also be used as eluents but were not tested.A silver suppressor column can be used with an HCl eluent to allow the acids to be detected as free acids (as in the anion-chromato- graphic methods), and this in theory should enhance sensitiv- ity to a level comparable to that of the gradient method; however, the column may be clogged up with silver chloride, resulting in a futile suppression process. Isocratic method The C032--HC03- eluent system used previously13 for anion analysis of unpreserved samples is applied here to preserved samples (0.2% CHC13). The working standard eluent used was prepared with 1 g of Na2C03 and 1 g of NaHC03 in 4 dm3 of water. Three other eluents were prepared likewise using 1.1, 0.75 and 0.65 g of each salt.Four chromatograms of a standard using the four eluents are shown in Fig. 3. As can be seen, each eluent is suitable for use and has its own merit. For example, the eluent with 1.1 g of salts has the shortest analysis time and well defined, sharp peaks but with samples having markedly different concentrations, the closely eluted peaks are prone to showing peak overlap or shouldering. On the other hand the eluent with 0.65 g of salts has very well separated peaks but the later peaks are very broad and elute very late. It can also be observed from Fig. 3 that the elution patterns of phosphate and nitrate depend on the eluent strength, i.e., phosphate elutes before nitrate at the 1.1 and 1 g strengths but elutes after nitrate at the 0.75 and 0.65 g strengths.In all four instances, the methanesulfonic acid peaks are at the edge of the water dip region, which can cause its analytical results to be unreliable. Also, HF co-elutes with other analytes (glycolic, formic, acetic, lactic, propionic and butyric acids) in the water dip region and cannot, therefore, be determined satisfactorily. Finally, sulfuric and citric acids co-elute but sulfuric acid is about 100 times more sensitive than citric acid. Changing the carbonate-hydrogen carbonate ratio (chang- ing pH) will resolve the methanesulfonic acid peak and the water dip better, but it may also adversely affect the resolution of others. For example, Saigne et a1.15 were able to separate the methanesulfonic acid peak from the water dip but saw their C1- peak elute at a time of 10 min, an extremely long retention time.Of course the retention times for nitrate and sulfate will be much longer. As our C032--HC03- ratio is a good working one, which has been shown to work ruggedly for a wide range of samples,13 there was no real need to test different ratios. Results and Discussion Commonly Reported Acids The acids commonly found in precipitation-related samples as reported in the literature are sulfuric, nitric, hydrochloric, hydrofluoric, formic and acetic acids, in addition to propionic, lactic, glycolic, butyric, methanesulfonic, oxalic, hydroxy- methylsulfonic, nitrous, phosphoric and citric acids, as shown in Table 1.4&-9,11,14-18 Analytes and Analysis Time The resolution of analytes and analysis time using each of the three methods for a standard containing acids commonly found in precipitation samples is shown in Fig.4. The resolution quality, however good it may be for ea'ch method, can be marred by some drawbacks. For example, the 0.30 0.30 0 4 8 12 16 Ti me/m i n Fig. 4 Analytes and analysis time for the three methods. (a) Ion-exclusion; (b) isocratic; and (c) gradient method1141 ANALYST, JULY 1992, VOL. 117 3.30 - Id Ic Ib -0.30 1 1 0 4 8 12 Ti me/min Fig. 5 l c , acetate; ld, fluoride; 2, chloride; 3, nitrate; 4, sulfate; and MS, methanesulfonate Chromatograms of a Burlington rain-dry fall sample. ( a ) Isocratic method; (b) and ( c ) gradient method. la, Formate; lb, resolution of acetic, propionic, glycolic and butyric acids may sometimes be poor with the gradient method.In the isocratic method, methanesulfonic acid elutes at the edge of the void volume; and the resolution of phosphoric-nitric acids might be problematic although it can be easily remedied. In the ion-exclusion (IE) method, the citric acid peak is very close to the void volume, and hydrofluoric, lactic, glycolic and formic acids are not always well resolved. The resolution and efficiency of the gradient method (16 analytes in 16 min) is outstanding when compared with the isocratic method (8 analytes in 10 min) and the IE method (8 analytes in 19 min). However, it has some limitations: acetic and lactic acids co-elute and, as mentioned above, the resolution of some organic monoprotic acids can be poor, especially after a change of eluent, as they are closely eluted. As acetic acid is a major organic acid, the acetic-lactic acid co-elution could cause inaccurate determination even though lactic acid is a minor acid.Ion-exclusion on the other hand allows effective resolution of the two acids, and the acetic acid peak is consistently reliable in the chromatograms. The resolution of formic acid, however, is more reliable with the gradient than with the IE method. Thus, the combination of the gradient power (HF, formic, methanesulfonic, inorganic acids) and the IE ability to resolve key analytes (lactic, acetic, propionic acids) would be capable of satisfactorily determin- ing all the acids of interest. This combination also gives some duplicate analyses for double-checking purposes.The isocratic method has the following advantages over the gradient method: shorter analysis time (10 versus 16 min) and a more stable and more easily handled eluent system. In terms of the number of analytes that can be analysed, it appears that the isocratic-IE combination is superior to the gradient glycolate; Table 5 Comparison of detection limits, based on a 50 mm3 sample loop Detection limit (ppm) Anion- Anion- exchange exchange gradient isocratic IE Formic 0.04 - 0.4 Acid mode mode mode Acetic 0.06 - 0.2 Glycolic 0.04 - 0.2 Propionic 0.06 - 0.3 Lactic - - 0.3 Butyric 0.07 - 0.4 Oxalic 0.07 0.20 Citric 0.07 - 0.4 MS 0.05 0.05 HOMS 0.08* 0.25 - HF 0.01 - 0.1 0.02 0.015 - HC1 HNO;? 0.01 0.01 HN03 0.01 0.01 - - - - 0.04 0.04 - - H2S04 H3P04 0.03 0.03 determined for HOMS as it co-eluted with succinic acid.* 0.08 is the detection limit for succinic acid. The limit was not method. However, this combination will not allow satisfactory determination of the complete range of analytes because of the drawbacks mentioned above. For example, a major acid such as formic acid cannot be determined by the isocratic method and, even though IE will permit its determination, the1142 ANALYST, JULY 1992, VOL. 117 poor precision would give uncertain results (see below). The same reasoning applies to HF. Methanesulfonic acid (MS) cannot be determined by IE and the results obtained from the isocratic method can be unreliable as MS occurs at the edge of the water dip. For example, it is not possible to identify peak 1 in a chromatogram of a rain-dry fall sample [Fig.5, chromatogram(a)]; based on retention time, it could be MS, formic, HF or acetic acid or a combination of these acids. An analysis of the same sample by the gradient method [chromat- ogram(b)] reveals that peak 1 of chromatogram (a) definitely contains formic acid and hardly any MS. When chromatogram (6) is expanded ten times [chromatogram (c)] it can be seen that peak 1 of chromatogram (a) consists mainly of four different analytes; namely HF, acetic, glycolic and formic acids. Thus the isocratic-IE combination, if used, will require another method to determine adequately all the acids of interest and this would be a definite disadvantage. Further- more, although this combination can resolve and determine acetic and lactic acids, which the gradient method cannot (Table 5), the uncertain results given by this combination for HF, MS and formic acids hardly make it superior to the latter method.Detection Limit and Sensitivity For the IE and isocratic methods, the detection limit was defined as three times the standard deviation of 8-10 replicate analyses of a standard containing five times the calculated detection limit, the latter being the concentration giving a signal-to-noise ratio of 2-3. For the gradient method, the detection limit was defined as twice the standard deviation, as it was determined from data generated on five different days in a 3 week period involving a weekly change of eluents, which led to less precise results and thus larger detection limits; in this way, it was similar to dealing with interlaboratory detection limits, which should be larger than a single-labora- tory detection limit.19 The detection limits obtained for the three methods, based on a 50 mm3 sample loop, are compared in Table 5.The limits by IE are higher than those by the gradient and isocratic methods, the last two being compar- able. Method sensitivity has several definitions in the literature both in general terms, and in terms specific to the gradient method.11 It is depicted here as the response as a function of concentrations at the low end of the analytical curves for the IE and isocratic methods (Figs. 6 and 7). Each point represents an average value of three analyses. For each acid at four concentration levels, a line was drawn manually and the regression parameters were calculated and are shown in the legends.Although the detection limit is the most commonly refer- enced performance indicator, sensitivity (which depends on the difference between the equivalent conductances of sample and eluent species and also on the over-all change in conductivity) can offer some intriguing information. For example, if sensitivity is taken as the response per unit concentration, it can be equated to the slope of each line (Figs. 6 and 7, Fig. 5 in ref. 11) and can be easily used in arithmetic computations. The ratio of sensitivities and detection limits of the three methods are compared in Table 6 and it can be seen that the sensitivity is much greater for the gradient method than for the IE method, and as expected the detection limit for the gradient method is smaller than that for the IE method.However, even though the sensitivity is greater for the gradient method than the isocratic method, the detection limits are comparable (Table 6). This may be explained by the e 1600 CitricA v) 3 E 800 8 % g 400 Q a a 0 0.5 1 2 3 4 5 Fig. 6 Method (IE) sensitivity for the acids. Regression equations ( r = correlation coefficient): HF, 0.013 + 0.00148R ( r = 1.OOO); formic acid, 0.12 + 0.0022R ( r = 1.0oO); citric acid, 0.43 + 3.82 x lO-3R - 6.12 x lO-’R*; acetic acid, 0.054 + 0.00398R (r = 1.0oO); glycolic acid, 0.026 + 0.00463R ( r = 1.0oO); lactic acid, 0.005 + 0.0056R ( r = 1.OOO); propionic acid, 0.067 + 0.0069R (I = 1.000); and butyric acid, 0.144 + 0.0135R ( r = l.0oO) Concentration (c) (ppm) Table 6 Sensitivity ratio and detection limit ratio for gradient (G) versu ion-exclusion (IE) method, and gradient (G) versus isocratic (ISO) method based on a 50 mm3 sample loop I Gradient versus Gradient versus ion-exclusion isocratic I - f / Sensitivity ratio (G : IE) Detection limit ratio (IE : G) Sensitivity ratio (G : ISO) 5.3 5.0 2.8 5.9 5.3 12.5 - - - 9.1* Detection limit ratio (IS0 : G ) 1.0 1.0 0.75 1 .o 1.0 2.9 - - - 2.1* r‘ 0) 0, 3400 .- Acid Nitrous Nitric Hydrochloric H y drofluoric Sulfuric Phosphoric Oxalic Formic Glycolic HOMS Acetic Lactic Citric Methanesulfonic Propionic Butyric L Y tu Q - 2600 C 3 - 15.4 - 10.0 * 1800 s c 0 Q K 8 1000) - 7.0 11.8 8.0 6.6 11.0 16.2 - - - - 10.0 5.0 3.3 5.7 5.0 5.7 - - - 200 0.025 0.5 1 2 3 4 5 Concentration (c) (ppm) Fig. 7 Method (isocratic) sensitivity for the acids.Regression equations ( r = correlation coefficient): C1-, -0.016 + 0.0002373R ( r 0.0002622R ( r = 0.9999); MS, 0.106 + 0.0004976R ( r = 0.9998); HP04*-, 0.02134 + 0.0007822R (r = 0.9998); S042-, 0.015 + 0.000833R ( r = 0.9997); oxalate, 0.08577 + 0.002346R ( r = 1.oooO); and HOMS, 0.1463 + 0.003684R ( r = 0.9999) = 0.9990); N02-, 0.0087 + 0.0002567R ( r = 0.9990); N03-, 0.0081 + - 1.0 - 1 .o * Values from the gradient method are for succinic acid. The sensitivity and detection limit for HOMS were not determined as HOMS co-elutes with succinic acid in the gradient method.ANALYST, JULY 1992, VOL. 117 1143 Table 7 Comparison of precision, for Sibley water and quality control (EU-ANI-1) (in parentheses) samples Relative standard deviation (YO)* Anion-exchange Anion-exchange Acid gradient mode isocratic mode Formic 7.3 (8.3) - Acetic 5.3 (10.8) - Glycolic 2.5 (9.7) - Propionic 1.7 (10.0) - Butyric 2.1 (6.7) - Citric 3.2 (2.0) - HF 4.0 (7.3) - HC1 6.1 (6.2) 4.9 (5.4) Lactic - - Oxalic 0.9 (2.4) 6.9 (2.1) MS 3.9 (3.6) 8.4 (1.6) HOMS 4.3 (2.l)t 1.3 (1.3) HN02 12.9 (2.0) 4.2 (2.3) HN03 1.9 (1.7) 2.7 (5.9) H2S04 2.2 (1.9) 2.1 (2.6) H3P04 3.8 (3.0) 1.8 (4.2) * n 2 5 .t Values for succinic acid. IE mode 16.4 (13.1) 11.5 (6.6) 20.0 (13.8) 14.6 (8.4) 12.9 (23.4) 31.0 (12.9) 5.9 (8.2) - - - 17.3 (8.8) - - - - - Table 8 Comparison of accuracy for Sibley water and quality control (EU-ANI-1) (in parentheses) samples Recovery (YO) Anion-exchange Anion-exchange Acid gradient mode isocratic mode Formic 94 (94) - Acetic 103 (96) - Glycolic 98 (103) - Propionic 106 (106) - Butyric 103 (102) - Oxalic 108 (100) 97 (99) Citric 106 (100) - HF 106 (110) - - - Lactic MS 108 (101) 105 (102) HOMS 104 (99)* 98 (95) HCl 105 (105) 98 (97) HNO2 106 (104) 96 (100) HN03 loo (95) 97 (103) H2S04 99 (99) 100 (103) H3P04 104 (99) 98 (102) * Values for succinic acid.IE mode 97 (86) 103 (102) 103 (105) 105 (100) 106 (106) 99 (108) 101 (122) - - - 100 (96) - - - - - methods, the instruments are different, the main difference being the pumps, the gradient pump (up to 27 580 kPa or 4000 psi) was used for the gradient method, and the analytical pump (up to 13 790 kPa or 2000 psi) was used for the isocratic method. The gradient method also had four more columns between the pump and the loadinject valve than the isocratic method.These differences must have contributed to the higher noise for the gradient method, which subsequently ‘neutralizes’ the higher signal, thus leading to the similarity in detection limits. Furthermore, at very low sample concentra- tions, very little carbonic acid (eluent of the isocratic method) is being replaced by sample species so the apparent drop in response due to loss of conducting carbonic acid is less marked and the over-all increase in response due to the change in conductivity is comparable to that of the gradient method where the eluent with lower conductance (water) is being replaced by the same species. Precision and Accuracy The relative standard deviations obtained for two natural waters, one water collected from Sibley, Ontario, and the other a quality control sample (EU-ANI-1) prepared using several rainfalls for the Eulerian quality control programme, are compared in Table 7.It can be seen that the IE method is less precise over-all than the gradient and isocratic methods. Specifically, IE has inferior reproducibility for HF, lactic, glycolic and formic acids. The gradient method also has poor reproducibility for acetic, propionic, glycolic and butyric acids. The isocratic method produced precise results through- out, and this could perhaps further explain its small detection limits as discussed above (as detection limit is a precision statement). The percentage recoveries were likewise obtained and are compared in Table 8.The recoveries are in general well within 100 k 10% except for some organic acids determined by the IE method. The invaluable reviews by the referees are acknowledged. fact that the gradient detection limits were considered as interlaboratory limits as explained above. Another explanation for the similarity in detection limits between the gradient and isocratic methods is the following. Given the fact that the background conductivity for the gradient method (=3 pS) is smaller than that for the isocratic method ( ~ 1 8 pS), and that the peaks for the gradient method are more needle shaped than those for the isocratic method (15 peaks in about 5 min versus 8 peaks in about 8 min resolution time, see Fig. 4), it is not surprising that in general the apparent sensitivity ratio (gradient : isocratic, Table 6) is significant. However, on a peak area basis and at lower concentrations where very little carbonic acid is replaced by sample species, this ratio would definitely be less marked.Sensitivity is a function of the difference between the equivalent conductances, whereas detection limit is a function of noise and is often defined in terms of the signal-to-noise ratio. Noise varies from instrument to instrument and depends on factors such as signal magnitude, temperature variations and electronics. Although the same detector was used for both 1 2 3 4 5 6 7 8 9 10 11 12 13 References Sanhueza, E., Elbert, W., Rondon, A., Arias, M. C., and Hermoso, M., Tellus, Ser. B, 1989,41, 170. Likens, G. E., Keene, W. C., Miller, J. M., and Galloway, J. N., J. Geophys. Res., 1987, 92, 13299. CSIRO, Division of Atmospheric Research, Research Report 1985-1988, Aspendale, 1989, p. 27. Grosjean, D., van Neste, A., and Williams, E. L., 11, Prepr. Pap. Natl. Meet.-Am. Chem. SOC., Div. Environ. Chem., 1988, 28, 59. Grosjean, D:, Atmos. Environ., 1988, 22, 1637. Keene, W. C., and Galloway, J. N., Atmos. Environ., 1984,18, 2491. Galloway, J. N., and Gaudry, A., Atmos. Environ., 1984, 18, 2649. Backman, S. R., and Peden, M. E., Water, Air Soil Pollut., 1987, 33, 191. Baltensperger, U., and Kern, S., J. Chromatogr., 1988, 439, 121. Ferek, R. J., Eynon, B. P., and Endlich, R. M., J. Appl. Meteorol., 1988,27, 1344. Cheam, V., J. Chromatogr., 1989, 482, 381. Franklin, G. O., Tappi, 1982,65, 107. Cheam, V., and Chau, A. S. Y., Analyst, 1987, 112,993.1144 ANALYST, JULY 1992, VOL. 117 14 Solomon, P. A., Fall, T., Salmon, L. G., and Cass, G. R., Prepr. Pap. Natl. Meet.-Am. Chem. SOC., Div. Environ. Chem., 1988,28,72. 15 Saigne, C., Kirchner, S., and Legrand, M., Anal. Chim. Acta, 1987, 203, 11. 16 Tsitouridou, R., and Puxbaum, H., Int. J. Environ. Anal. Chem., 1987,31,11. 17 Murray, F., Arch. Environ. Contam. Toxicol., 1989, 18, 131. 18 Haddad, P. R., and Jackson, P. E., J. Chromatogr., 1988,447, 155. 19 Cheam, V., and Chau, A. S. Y., Specification Studies and Statements. I. Trace Metals, Major Ions, Nutrients, Physical Parameters and Miscellaneous Inorganic Parameters in Waters at Three Concentration Levels, National Water Research Institute, Ontario, 1981, manuscript No. 16-AMD-T-6-81-VC. Paper 1 I021 87G Received May 9, 1991 Accepted January 24, 1992

ISSN:0003-2654    

DOI:10.1039/AN9921701137

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, JULY 1992, VOL. 117 1145 Anion-exchange Chromatography of Mixed Cyano Complexes : Separation and Determination of Dicyanoaurate(1) Emmanuel 0. Otu, Campbell W. Robinson and John J. Byerley Department of Chemical Engineering, University of Waterloo, Waterloo, Ontario, Canada N2L 3G I Two eluent systems are described for the separation and determination of gold and other metal cyano complexes. The elution order of the cyano species was dependent on the sodium perchlorate concentration. Greater ionic column selectivity for gold was achieved with an eluent containing ammonia. The complexes were detected spectrophotometrically at 215 nm. The detection limit for gold is 5 pg dm-3 using direct injection. Keywords: Gold determination; metal cyano complexes; ion chromatography The hydrometallurgy of gold involves the leaching of gold ore in oxygenated cyanide solution.This treatment results in the formation and stabilization of Au' as the dicyanoaurate(1) anion Au(CN)2-. As the cyanide leach system is not particu- larly selective, a wide range of metal cyano complexes also appear in the hydrometallurgical circuit. Many of these co-leached metal cyano complexes can build up in the leach circuit to concentrations far in excess of the gold concentration and can interfere in the determination of gold in the various process liquors. Further problems may arise from the high background level of various oxysulfur species and thiocyanate when sulfidic gold ores are processed. There are numerous established methods for gold determination.1-4 These include the early gravimetric and spectrophotometric procedures3.4 and modern instrumental methods such as atomic absorption and emission spectrometry,l-3 X-ray fluorescence spec- trometry2.3 and certain electrochemical techniques.5 All these methods have found successful application in many industrial situations requiring gold determination. The chromatographic separation and detection of both derivatized and underivatized metal species now form the basis of a number of reliable techniques for determining metal ions in aqueous solution. This methodology has been particu- larly useful for determining a specific metal species in the presence of a broad range of related metal species, in addition to providing multi-metal ion analyses of liquors associated, for example, with hydrometallurgical process streams.It is noteworthy that the identification of oxidation states is possible with this technique. The term ion-exchange or simply ion chromatography is often used to identify a number of related methods. One method for separating anionic metal species employs an eluent containing a lipophilic ion such as tetrabutylammonium hydroxide together with suitable organic and inorganic modifiers, e.g., acetonitrile and car- bonate.611 The method which employs a common reversed- phase silica column is referred to as reversed-phase ion-pair chromatography (RPIPC) and has been used successfully by Haddad and co-workers68 to separate metal cyano complexes with ultraviolet (UV) detection. A variation of this method, termed mobile phase ion chromatography (MPIC), also employs a neutral non-polar column but detection is via suppressed conductivity.9-11 An alternative method involves the use of anion-exchange chromatography employing a resin with -NR3 functionality.Rocklin12 reported the separation of precious metal chloro complexes using a perchlorate- hydrochloric acid eluent. Ultraviolet and amperometric detec- tion were used. In this paper the separation of metal cyano complexes in alkaline cyanide solution using anion-exchange chromato- graphy is described. Attention is focused on the determination of low levels of gold as Au(CN)~- in the presence of a number + of related metal cyano species and other anions. Preconcen- tration procedures were not considered in this work. Experimental All chromatography was carried out using a Dionex System 2010i, which included an ion chromatograph module, a gradient pump and a UVhisible detector.A Hewlett-Packard (NP3394A) millivolt recorder was used. The separator was a Dionex HPIC-AS5 column in which the anion-exchange resin was composed of a surface-sulfonated poly(styrene)-divinylbenzene core with a particle diameter of 15 pm coated with fully aminated latex particles with a diameter of about 0.1 pm. The ion exchange functionality is -NR3. The separation column was protected by a Dionex HPIC-AG4 guard column. Eluents were prepared from analytical-reagent grade NaC104-H20 (Fisher), NaCN (Baker), 50% NaOH (Mal- linckrodt) and ammonia solution [BDH (now Merck)]. Potassium salts of the cyanides of gold [KAu(CN)2], silver [KAg(CN)2] and cobalt [K&o(CN)~] and copper(1) cyanide (CuCN) and zinc cyanide [ZII(CN)~] were obtained from Johnson Matthey Chemicals (Aesar grade).Potassium cyan- ide salts of iron(rr) and iron(m) (Fisher) were analytical- reagent grade. Potassium cyanonickelate(I1) was obtained from Strem Chemicals. Stock standard solutions containing 100 mg dm-3 as metal of Au(CN)2-, Ag(CN)2-, Ni(CN)42-, Co(CN)b3-, Fe(CN)64- and Fe(CN)$ were prepared by dissolving the appropriate amount of salt in de-ionized water. Stoichiometric amounts of NaCN to form CU(CN)~~- and ZII(CN)~~- were added to dissolve CuCN and Zn(CN)2 to make 100 mg dm-3 stock solutions with respect to the metal. In preparing a set of synthetic solutions from the metal cyanide stock solutions, a number of other components were added to simulate indus- trial conditions.These included S042- as Na2S04, S2O32- as Na2S203, SCN- as NaSCN, As02- as NaAs02, Ca2+ as CaC12, Hg2+ as HgC12 and humic acid. Technical-grade humic acid (sodium salt) was supplied by Aldrich. Synthetic solutions of the following compositions were prepared from the stock standard solutions: synthetic solution A, Au' 2, Ag' 2, Cul 2, Ni" 2, Col" 1, Fell, 0.25 and Fe"l0.25 mg dm-3; and synthetic solution B, Au' 2.5, Cul 50, Fe1I 2.5, Zn" 5, Co'" 0.5, NilI 2.5, Ag' 1, As"' 25, Hg" 5, NaCN 250, humic acid 10, SCN- 125, S042- 250, S2O32- 50 and Ca 125 mg dm-3. All chromatographic experiments were carried out at ambient temperature (approximately 298 K) and a flow rate of 1.0 cm3 min-1.The sample loop volume for direct injection was either 50 or 150 mm3. The detection wavelength was 215 nm. The eluents contained a constant amount of NaCN (15 +1146 ANALYST, JULY 1992, VOL. 117 mmol dm-3) to maintain the integrity of the cyano complexes. When NaOH was used its concentration was 20 mmol dm-3. When ammonia was substituted for NaOH the concentration was 22 mmol dm-3. The concentration of the displacing perchlorate varied from 20 to 120 mmol dm-3. All eluents were vacuum de-gassed and maintained under nitrogen. The eluents were pumped through the system until a stable baseline was obtained. The detector zero was offset and the sample was then injected through a pneumatically operated injection valve. Peak heights were used for quantification and retention times were used for ion identification with reference to standards.Results and Discussion A typical chromatogram obtained by direct injection of a 1.0 mg dm-3 standard Au' solution using an eluent composition of 50 mmol dm-3 NaC104, 15 mmol dm-3 NaCN and 20 mmol dm-3 NaOH is shown in Fig. 1. The injection loop volume was 150 mm3. The chromatogram shows the Au(CN)2- peak together with two other peaks. These peaks were identified as CU(CN)~~- and Fe(CN)64--Fe(CN)63- and resulted from impurities in the standard solution. Calibration showed the concentrations of these impurities to be less than 4 pg dm-3. Under the chromatographic conditions the retention times for Cu(CN)32-, Au(CN)~- and Fe(CN)64--Fe(CN)63- were 3.9, 4.8 and 7.4 min, respectively.t - m C P) iij 0 Au I T 0.002 A 1 cu 4 Time/mi n 8 Fig. 1 Chromato ram for standard Au(CN)*- solution: loop volume = 150 mm3; [Auf= 1.0 mg dm-3; h = 215 nm; 0.02 a.u.f.s.; and eluent, 50 mmol dm-3 NaC104 + 20 mmol dm-3 NaOH + 15 mmol dm-3 NaCN s- Ir AU Ni -e 10.003 A co 0 4 8 12 16 20 Time/m i n Fig. 2 Isocratic separation of Agl, Cull, A d , Nil1, Fell-Felll and ColI1 in synthetic solution A: loop volume = 50 mm3; h = 215 nm; 0.02 a.u.f.s.; and eluent, 50 mmol dm-3 NaC104 + 20 mmol dm-3 NaOH + 15 mmol dm-3 NaCN The chromatography used for the standard gold test solutions was applied to a solution containing low levels of metal cyano complexes, including gold. This solution is identified as synthetic solution A. The chromatogram shown in Fig. 2 indicates early elution of Ag(CN)2- (2.4 min), followed by CU(CN)~~- (3.7 min) and Au(CN)~- (4.7 min) on the tail of the Cu(CN)32- peak and Ni(CN)42- (6.5 min).The Fe(CN)64- and Fc?(CN)~~- complexes co-elute (7.3 min) with CO(CN)~~- eluting at 17.6 min. The co-elution of Fe" and Fe"' cyano complexes may be the result of the reduction of Fe(CN)$ under the chromatographic conditions. While this chromatography does afford some separation of the various cyano complexes, determination of Aul is not particularly convenient owing to the interference from CU(CN)~~-. It is apparent that when the concentrations of the various metal cyano complexes are many times greater than that of gold, e . g . , in gold processing solutions, the gold peak will be swamped by that of the Cu' complex. These problems can be overcome by gradient pumping of the eluent whereby the perchlorate concentration is varied from 20 to 95 mmol dm-3 using a linear-curvilinear programme.A chromatogram obtained from a 50 mm3 injection of the same sample as used to obtain Fig. 2 is shown in Fig. 3. The gradient conditions are given in Table 1. In this instance improved separation between the various metal cyano complexes is obtained. The gold complex Au(CN)2- is satisfactorily separated, with the peak representing Cu(CN)32- now appearing after rather than before the Au(CN)2- peak. This gradient pumping scheme allows the convenient determination of Au' and could form the basis for determining the other metal cyano complexes. In order to evaluate fully the gradient pumping modification for gold determination in process solutions, samples contain- ing metal cyano species with concentrations many times that of 10.003 A Ag sys cu SYS 4 Ni 1 I I I I I 0 4 8 12 16 20 24 Time/min Fig.3 Gradient separation of Ag', Au', Cu', Nil', Fell-Felll and CoI1' in synthetic solution A: loop volume = 50 mm3; h = 215 nm; 0.02 a.u.f.s.; Sys = peaks due to gradient programme. The gradient conditions are given in Table 1 Table 1 Gradient conditions for separation using an eluent containing NaOH Eluent composition (%)* Time/min El 0.7 10 15 20 25 100 95 75 25 0 E2 0 5 25 75 100 * El, eluent 1 = 20 mmol dm-3 NaC104 + 20 mmol dm-3 NaOH + 15 mmol dm-3 NaCN; E2, eluent 2 = 120 mmol dm-3 NaC104 + 20 mmol dm-3 NaOH + 15 mmol dm-3 NaCN.ANALYST, JULY 1992, VOL. 117 1147 I I I I I I 0 4 8 12 16 20 24 Time/min Fig.4 Gradient separation of metal cyano complexes in synthetic solution B with conditions as in Fig. 3. Sys = peaks due to gradient programme r 1 Au I 10.002 A '1- 0 4 8 12 16 Timehi n Fig. 5 Chromatogram for standard Au CNh- solution: loop volume = 150 mm3; [Ad] = 1.0 mg dm-3 + [Cub = 0.1 mg dm-3; h = 215 nm, 0.02 a.u.f.s.; and eluent, 50 mmol dm-3 NaC104 + 22 mmol dm-3 ammonia solution + 15 mmol dm-3 NaCN 1 I 1 I 1 0 4 8 12 16 20 Time/min Fig. 6 Interference of copper on the Au(CN)y signal: [Ad] = 0.1 mg dm-3; [Cu'] = 142 mg dm-3; and loop volume = 50 mm3. Other conditions as in Fig. 5 gold, in addition to other anionic and cationic species, must be considered. This condition is represented by synthetic solution B.The chromatogram obtained for a 50 mm3 injection of synthetic solution B is shown in Fig. 4. In this instance, the 50 mg dm-3 Cu' concentration in the sample produces a very broad and off-scale peak very soon after the elution of AU(CN)~-, whose peak represents 2.5 mg dm-3 Au'. It is apparent that if Au' is to be determined at levels characteristic 0.0002 A I 0 3 6 Timehin Fig. 7 Ultraviolet detection of Au(CN)*- without preconcentration at concentrations of (a) 5, (b) 10 and (c) 25 pg dm-3: 0.002 a.u.f.s. Other conditions as in Fig. 5 of mill effluents (1Cb100 pg dm-3), some means of eliminating this interference of Cu' must be implemented. Note that Zn" and Hg" cyano complexes, which are part of synthetic solution B, do not appear on the chromatogram shown in Fig.4. The ZJI(CN)~~- may be eluted in the void volume owing to its low stability (log p = 16.7) in comparison with the other complexes. For Hg" a high-stability complex is formed [Hg(CN)42-, log p = 41.51 but its absorbance at 215 nm is reported to be very weak.* In situ complexations in ion chromatography have proved useful in the development of new chromatographic separation methods.13J4 It was decided to replace sodium hydroxide in the eluent with ammonia and at the same time maintain the pH possibly to provide for competitive equilibria with cyanide for copper. The chromatogram obtained by injecting 150 mm3 of the standard 1 mg dm-3 Au' solution as shown in Fig. 1, but using an eluent composition of 50 mmol dm-3 NaC104, 15 mmol dm-3 NaCN and 22 mmol dm-3 NH3 solution did not yield the defined CU(CN)~~- impurity peak.However, as1148 E E .- m50 9) 0) r Y a 2 ANALYST, JULY 1992, VOL. 117 - 100 I 1 I 1 I I 1 I I 0 4 8 12 16 20 24 Time/min Fig. 8 Gradient separation of Ag', Au', Nil', Cu', Fell-Fel'' and Co"' in synthetic solution A: loop volume = 50 mm3; h = 215 nm; 0.02 a.u.f.s. Eluent switched after 13.5 min: eluent 1, 50 mmol dm-3 NaC104 + 22 mmol dm-3 ammonia solution + 15 mmol dm-3 NaCN; and eluent 2, 120 mmol dm-3 NaC104 + 22 mmol dm-3 ammonia solution + 15 rnmol dm-3 NaCN cu 0 4 8 12 Ti me/mi n Fig. 9 Separation of Cu' (0.2 mg dm-3), Nil' (1 .O mg dm-3), Fel'-Fel'' (0.25 mg dm-3) and ColI1 (0.5 mg dm-3): loop volume = 50 mm3; h = 215 nm; 0.02 a.u.f.s.; and eluent, 100 mmol dm-3 NaC104 + 22 mmol dm-3 ammonia solution + 15 mmol dm-3 NaCN shown in Fig.5, on injecting a 150 mm3 sample of the standard 1 mg dm-3 Au' solution to which had been added 0.1 mg dm-3 Cu', the CU(CN)~~- peak appears much later (15.6 min) and is well separated from the Au(CN)~- peak. The Au(CN)~- peak height was enhanced by 30% over that observed when using NaOH in the eluent and the retention time was slightly longer (5.1 min). The Fe(CN)64--Fe(CN)63- species were not eluted in a reasonable chromatographic time with this eluent. A chromatogram obtained by injecting 50 mm3 of a 0.1 mg dm-3 Au' solution containing 142 mg dm-3 Cu' is shown in Fig. 6. It is observed that the gold peak is well separated from the broad Cu' signal. Using test solutions prepared from the standard Au' solution, the limit of accurate determination of gold was calculated to be 10 pg dm-3 using the criterion of a signal-to-noise ratio of 3 : 1.The limit of detection was calculated to be 5 pg dm-3. Ten replicate injections of a 10 pg dm-3 standard gold solution gave a relative standard deviation of 1.25%. The above performance was obtained without preconcentration. Chromatograms obtained by injecting 150 mm3 of standard solutions containing 5, 10 or 25 pg dm-3 Aul are shown in Fig. 7. For simplicity, the copper impurity peaks (elution time 13.7 min), which are amplified ten times, are not shown in Fig. 7. A chromatogram obtained by injecting 50 mm3 of synthetic solution A is shown in Fig. 8 using a delayed step change in NaC104 concentration from 50 to 120 mmol dm-3. This 0 20 40 60 [NaC1041/mmol dm-3 Fig.10 Effect of NaC104 concentration on the peak height for Au(CN)~-: loop volume = 50 mm3, h = 215 nm, 0.02 a.u.f.s. A. Eluent as in Fig. 5 and B, eluent as in Fig. 1 Table 2 Comparison of determination of gold (mg dm-3) using ion chromatography (IC), AAS and DCP Method Test solution as made up IC AAS DCP 0.1* 0.104 0.102 0.095 0.17 0.100 0.179 0.097 0.1$ 0.106 0.252 0.084 0.5* 0.509 0.498 0.502 0.5t 0.504 0.604 0.526 * Sample matrix, distilled water. 7 Sample matrix, 50 mmol dm-3 NaC104 + 15 mmol dm-3 NaCN + $ Sample matrix, synthetic solution B, 25-fold dilution, 0.15 22 mmol dm-3 NH3 solution. mol dm-3 NaCN. change in eluent composition allowed the early elution of the strongly retained cobalt and iron species which otherwise would interfere with further chromatographic separations.The concentration of nickel in industrial gold process solutions is not as high as copper and its position on the chromatogram will not seriously affect the determination of Au' Determination of all the cyano complexes using gradient pumping did not allow the convenient determination of the metal cyano complexes in a single injection, as shown in Fig. 8. For the purposes of separating and determining CU(CN)~~-, Fe(CN)64--Fe(CN)63-, Ni(CN)42- and Co(CN)$, an eluent composition of 100 mmol dm-3 NaC104, 15 mmol dm-3 NaCN and 22 mmol dm-3 NH3 solution was used isocratically. The resulting chromatogram for a solution containing 0.25 mg dm-3 FeI1I-Fe", 0.20 mg dm-3 Cu', 1.0 mg dm-3 Nil' and 0.5 mg dm-3 C O ~ ~ ' as cyano complexes is shown in Fig.9. It was found that the amount of sodium perchlorate in the eluent was critical to the sensitive detection of Au(CN)Z- complexes. The retention times decrease with increasing sodium perchlorate concentration. The effect of sodium perchlorate concentration on the peak height of the AU(CN)~- complex is shown in Fig. 10. Although the objective of this study was to examine the anion-exchange chromatography of Au(CN)~-, it is appro- priate to comment briefly on how the technique compares with established instrumental methods such as atomic absorption spectrometry (AAS) and d.c. plasma spectrometry (DCP). A number of experiments were carried out whereby the instru- ments were first calibrated using a standard gold sample. Subsequent samples with the same matrix compared favour- ably, but a change of matrix had a significant effect on the AAS data, a lesser effect on the DCP data and virtually no effect on the anion-exchange chromatographic data.The results are reported in Table 2. Given the complexity of the matrix of industrial samples, anion-exchange chromatography would appear to be an attractive analytical method.ANALYST, JULY 1992, VOL. 117 1149 Conclusion This study has shown that Aul as Au(CN)~- can be effectively separated and determined from a sample matrix containing related metal cyano complexes and other species. By using a perchlorate-cyanide-ammonia eluent, competitive equilibria succeed in selectively delaying the elution of the metal cyano complexes, especially Cu', which is commonly present at high concentrations in gold process solutions.The detection limit of Au' as Au(CN)2- is 5 pg dm-3. Preconcentration will further lower the detection limit. Copper(1) and FelI-FelI1 are best separated and detected using NaOH in the eluent in view of the higher sensitivity obtained. The authors thank the Natural Sciences and Engineering Research Council of Canada (NSERC), the Province of Ontario University Research Incentive Fund (URIF) and the Centre for Chemical Process Metallurgy (CCPM) , University of Toronto, Canada, for financial assistance. References 1 Van Loon, J. C., and Barefoot, R. R., Determination of Precious Metals: Selected Instrumental Methods, Wiley, New York, 1991. 2 3 4 5 6 7 8 9 10 11 12 13 14 Van Loon, J. C., and Barefoot, R. R., Analytical Methods for Geochemical Exploration, Academic Press, New York, 1989. Beamish, F. E., and Van Loon, J. C., Recent Advances in the Analytical Chemistry of Noble Metals, Pergamon Press, Oxford, 1972. Beamish, F. E., The Analytical Chemistry of Noble Metals, Pergamon Press, Oxford, 1966. Schmid, G. M., and Curley-Florino, M. E., in Encyclopedia of Electrochemistry of the Elements, ed. Bard, A. J., Marcel Dekker, New York, 1975, vol. IV, p. 101. Haddad, P. R., and Rochester, N. E., Anal. Chem., 1988,60, 536. Haddad, P. R., and Rochester, N. E., J . Chromatogr., 1988, 439, 23. Hilton, D. F., and Haddad, P. R., J. Chromatogr., 1986, 361, 141. Weiss, J., Handbook of Ion Chromatography, Dionex, Sunny- vale, CA, 1986. Haak, K. K., and Franklin, G. O., in The 70th AES Annual Technical Conference Proceedings, Indianapolis, American Electroplaters' Society, Winter Park, FL, 1983, pp. 1-28. Haak, K. K., Plating Surf. Finish., 1983, 70, 34. Rocklin, R. D., Anal. Chem., 1984, 56, 1959. Sevenich, G. J., and Fritz, J. S., Anal. Chem., 1983, 55, 12. Heneghan, G., and Wallace, G. G., Anal. Proc., 1986,23,29. Paper 1 I0581 7G Received November 18, 1991 Accepted February 11, 1992

ISSN:0003-2654    

DOI:10.1039/AN9921701145

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, JULY 1992, VOL. 117 1151 Determination of Osmium Abundance in Molybdenite Mineral by Isotope Dilution Mass Spectrometry With Microwave Digestion Using Potassium Dichromate as Oxidizing Agent Katsuhiko Suzuki, Qi-Lu, Hiroshi Shimizu and Akimasa Masuda Department of Chemistry, Faculty of Science, University of Tokyo, Hongo, Tokyo 113, Japan An efficient and reliable method was developed for the accurate determination of osmium abundance in molybdenite (MoS2) by isotope dilution mass spectrometry. The sample solution of molybdenite was prepared by acid decomposition using the microwave digestion technique with addition of potassium dichromate (K2Cr207) as oxidizing agent and osmium in the sample solution was purified by distillation. There was no change in the ratios of 1870s derived from molybdenite relative to 1920s of an osmium standard with progress of distillation, indicating that isotopes of osmium derived from the two sources were completely homogenized.The precision of the osmium abundance obtained in the developed method was 1% (la). The technique was applied to a natural molybdenite sample for Re-Os age determination and the result was in good agreement with that obtained by the U-Pb method for zircon. Keywords: Osmium abundance; molybdenite; microwave digestion; isotope homogenization; isotope dilution mass spectrometry The nuclide 187Re is radioactive and decays to the stable 1870s by emission of a @-particle with a half-life of 4.23 X 1010years.l The Re-0s system has been thought to be highly suitable for the dating of sulfide minerals, whose formation ages cannot easily be determined by other conventional geochronometers.In particular, the application of the Re-0s system to the dating of molybdenite (MoS2) is of interest, because the rhenium concentration in molybdenite is relatively high (from a few ppm to a few per cent) compared with those in other sulfide minerals .273 Nevertheless, its application has been hindered by analytical difficulties with osmium isotope measurement and osmium abundance determination in molybdenite. Recent developments in various mass spectrometric tech- niques have allowed easier access to osmium isotope measure- ments.e Therefore, the precise determination of osmium in molybdenite has remained as an unsolved problem to be tackled from the analytical chemistry aspect.It is well known that osmium can easily escape as osmium tetroxide even in weakly oxidative conditions, which is one of the major reasons why osmium abundance determination has been difficult. In contrast, the determination of rhenium in molybdenite experiences little difficulty. In this study, efforts were focused on the development of a chemical procedure for the precise determination of osmium in molybdenite by isotope dilution mass spectrometry. Luck and Allkgre9 reported Re-0s ages of molybdenites, but the ages given for some of the molybdenites analysed seemed dubious. It is considered that problems remain with their analytical method, because little osmium could be recovered when we carried out repeated experiments10 following their method. Osmium might escape during the analytical procedure, and this method of Luck and Allkgre is judged to be impractical.Recently, it was observed that, in the determination of osmium in molybdenite by the isotope dilution method, complete isotopic exchange between the sample and spike isotopes was not guaranteed by the conventional acid diges- tion of molybdenite. 11 This isotope disequilibrium effect results in erroneous results for the determination of osmium in molybdenite and hence in the Re-0s age of molybdenite. Suzuki and MasudalO found that microwave digestion was effective in promoting osmium isotope exchange between the sample and spike or standard. Here, a further improved method for the precise determina- tion of osmium in molybdenite is presented, with improved conditions of decomposition and distillation.Complete iso- topic homogenization can be achieved by the procedure described here, which allows the accurate determination of osmium in molybdenite. This method for the determination of osmium has been applied to the dating of a natural molyb- denite. Experimental Apparatus A PlasmaQuad inductively coupled plasma mass spectrometer from VG Elemental was used for the measurement of osmium and rhenium isotope ratios. The operating conditions are presented in Table 1. For the decomposition of samples, an MDS-81D microwave digestion system (CEM) with a magnetron frequency of 2450 MHz was used. Mass discrimination in osmium isotope measurements was corrected by a method modified from that of Hirata et a1.12 This correction method is based on the assumption of a linear relationship between mass discrimination factor and mass number.Table 1 Instrumental operating conditions Plasma conditions- Frequency Incident power Reflected power Argon gas flow rate: Cooling Auxiliary Nebulizer Sampling conditions- Sampling aperture diameter Skimmer diameter Load coil aperture distance Dwell time No. of scans No. of channels Measurement time Scanning mass Data acquisition- 27.12 MHz 1.35 kW <5 w 13 dm3 min-l 1.5 dm3 min-1 0.75 dm3 min-' 1.0 mm 1 .O mm 10 mm 160 CLS 300 1024 33 s 182-195 u1152 ANALYST, JULY 1992, VOL. 117 Reagents In the following decomposition and distillation experiments, molybdenite (MoS2) powder from Wako Pure Chemical Industries (LAH2013) was used.Osmium contained in the molybdenite was confirmed to be pure radiogenic 1870s only. In the determination of osmium in molybdenite by means of the isotope dilution method, 'common' osmium can be used as a type of isotopic internal standard because of the predomi- nant existence of radiogenic 1870s in the molybdenite, free from 'common' osmium. A standard solution of common osmium was prepared by dissolving (N&)20~C16 powder (Johnson Matthey Material Technology) in 6 mol dm-3 HCl. The standard (N&)20SC16 was used because its powder can be easily dissolved and the osmium species is stable in the prepared standard solution. This solution was stored in a sealed glass vessel and diluted to the appropriate concentration when it was used. In order to determine precisely the concentration of osmium, perfect isotope exchange is required between radio- genic 1870s accumulated in molybdenite and common osmium added as an isotopic internal standard.Procedure Osmium can easily escape as osmium tetroxide even under weakly oxidative conditions because osmium tetroxide is highly volatile. Therefore, great care is required during chemical treatment to avoid loss of osmium. The osmium determination procedures tried in this study are as follows and are summarized in Table 2. A. Simple heating decomposition in a glass ampoule and distillation with addition of oxidizing agent ( K2Cr207) The sample powder, together with the osmium standard solution having a common isotope abundance, was placed in a glass ampoule with a mixture of 12 mol dm-3 HN03 and 4.5 mol dm-3 H2SO4 (1 + 4). After closing the glass ampoule, the sample was decomposed by heating the ampoule in a hot water-bath at about 80 "C for 2 d.After complete decomposi- tion, the ampoule was cooled to 0 "C to prevent loss of osmium as osmium tetroxide when opening the ampoule. The decom- posed solution was rapidly transferred into a distillation vessel and K2Cr207 was added to oxidize osmium to osmium tetroxide completely. Subsequently, osmium was separated by distillation and distilled osmium was trapped in about 0.1% thiourea solution. B. Microwave decomposition with addition of oxidizing agent (K2Cr207) in the course of decomposition B-1. Microwave decomposition with addition of K2C1-207 at the middle stage of decomposition.After addition of the osmium standard solution, the sample powder was decom- posed in acid solution (HN03 and H2SO4) in a poly- (tetrafluoroethylene) (PTFE) vessel by microwave digestion for 45 min at a power of about 200 W. After decomposition, the decomposition vessel was cooled and K2Cr207 was added to the decomposed solution. The amount of K2Cr207 was twice that of the sample powder. The vessel was again heated in the microwave system under the same conditions as in the first decomposition in order to oxidize osmium completely and obtain a high recovery of osmium. After cooling, the decomposed solution containing acid solution and oxidizing agent ( K2Cr207) was transferred into a distillation vessel. Subsequently, osmium was separated as osmium tetroxide by distillation and trapped in about 0.1% thiourea solution.B-2. Microwave decomposition with addition of K2Cr207 from the beginning of decomposition. After addition of the osmium standard solution, the sample powder was decomposed in acid solution (HN03 and H2SO4) with addition of oxidizing agent (K2Cr207) from the beginning of decomposition, under the same conditions as in procedure B-1. After decomposition, osmium was separated by distillation in the same way as in procedure B-1. C. Microwave digestion and distillation with addition of oxidizing agent The sample was decomposed in acid solution (HN03 and H2SO4) using the microwave digestion technique under the same conditions as in procedure B, except that no oxidizing agent was added. The decomposed solution was transferred into a distillation vessel and osmium was distilled in the following three different ways: (C-1) with addition of K2Cr207, (C-2) with addition of 9 mol dm-3 HC104 solution and (C-3) without addition of oxidizing agent. Results and Discussion Change of 187Od1920s Ratio with Progress of Distillation It has been found previously11 that simple decomposition by heating resulted in osmium isotopic disequilibrium between 1870s derived from molybdenite and common 0 s isotopes added as an internal standard.The change in the mass spectrum of osmium with progress of distillation for procedure A is shown in Fig. 1. It can be clearly seen that radiogenic 1870s derived from molybdenite increased gradually with time relative to common 0 s isotopes added as an internal standard, indicating that the isotope homogenization of osmium could not be achieved by procedure A, even after complete decomposition.Molybdenite is considered to be completely decomposed because silver-coloured molybdenite powder changed to a white precipitate of molybdic acid and this white precipitate was easily dissolved by ammonia solution. If Table 2 Analytical procedures employed Acid reagent* and vessel Addition of K2Cr207 Oxidizing agent Procedure Decomposition used in decomposition in decomposition added in distillation A Simple heating HN03 and H2SO4, B-1 Microwave HN03 and H2SO4, B-2 Microwave HN03 and H2S04, c-1 Microwave HN03 and H2S04, c-2 Microwave HN03 and H2S04, c-3 Microwave HN03 and H2S04, glass ampoule PTFE vessel PTFE vessel PTFE vessel PTFE vessel PTFE vessel No K2Cr207 At the middle stage of From the beginning of (K2Cr207)t (K2cr207) t decomposition decomposition No K2Cr207 No HClO, No No * Acid solution used for decomposition was a 1 + 4 mixture of 12 rnol dm-3 HN03 and 4.5 mol dm-3 H2SO4.t The decomposed solution contains K2Cr207 used during decomposition.ANALYST, JULY 1992, VOL. 117 1153 molybdenite was not decomposed and converted into mol- ybdic acid, the precipitate could not be dissolved by ammonia solution. In Table 3 are shown the variations in the 187Os/192Os ratio with distillation time for the procedures examined, together with osmium abundance variations for the molybdenite. Here the osmium abundance is calculated from the time-integrated osmium isotope ratio, which is based on isotope ratios and ion intensities for individual distillations.The change in the calculated osmium abundance with progress of distillation is also shown in Fig. 2. The 187Os/192Os value depends on the ratio of the amount of sample to that of osmium standard solution added, whereas osmium abundances calculated from integrated 187Os/192Os ratios should be definite once isotope homogenization between the sample and standard has been achieved, irrespective of the analytical procedure. As shown in Table 3, in procedures B-1 and C-1-(a), -(b) and -(c), changes in the 187Os/1920s ratio with distillation time cannot be observed, within the limits of analytical errors. In addition, it is worth noting that all the 0 s contents calculated for procedures B-1 and C-1-(a), -(b) and -(c) agree with each other, suggesting reliability of the calculated 0 s abundance. These results indicate that complete exchange of osmium isotopes between the sample and standard could be attained by these procedures.For practical use, procedure B-1 is thought to be preferable to procedures C-1-(a), -(b) and -(c), a 1 I % I 1 I I I 184 186 188 190 192 mlz Fig. 1 Change of osmium mass spectrum with time of distillation (a) 0-1 h; (b) 1-3 h (spectrum enlarged X3.5), obtained by procedure A: sim le heating decomposition and distillation with addition of K28r207. Amount of sample about 60 mg and volume of acidic solution for decomposition (HN03 and H2S04), 15 cm3 because errors in the 187Os/192Os ratios in procedure B-1 are smaller than those in the others.On the other hand, in procedures A, B-2, C-1-(d), -(e), -(f) and -(g), C-2 and C-3, the 1870s/1920s ratios increase with progress of distillation. These results indicate that, in these procedures, osmium isotopic equilibrium could not be achieved between 1870s derived from molybdenite and common osmium added as a standard. In these procedures, the resultant osmium abundances calculated from the inte- grated 187Os/192Os ratios also change with distillation time. However, the changes in osmium abundance with distillation time are minor for procedures B-2 and C-1-(d), -(e), -(f) and -(g), in which the peak intensities in the first distillation are much higher than those in subsequent distillations and these changes in intensity yield minor changes in the resultant osmium abundances with distillation time.It should be noted that, among the series of procedures C-1, changes in 1870s/1920s ratios with distillation time are observed for procedures C-1-(d), -(e), -(f) and -(g) and not for C-1-(a), -(b) and -(c); the sample amounts used are more than 74 mg in the first four experiments and less than 24 mg in the last three. These results imply that a low molybdenum concentration in the decomposed solution promotes osmium isotope exchange between the sample and standard added, owing to less polymerization of molybdic acid at low molyb- denum concentrations in the solution. 300 A a 0 P 1 t l 250 +- ou 6 200 6i gz; I I I 0 1 2 3 Di st i I I at io n ti m e/h Fig. 2 Change of osmium concentrations calculated from the time-integrated 1870s/1920s ratios given in Table 3.Procedure: 1, C-3; 2, C-2; 3, A-(c); 4, B-2; 5 , C-1-(a); 6, B-1; and 7, C-lc(g) Table 3 Change of 1870s/1920s ratios and osmium concentrations with distillation time. Errors are 2omean 1870s/1920s ratio Count ratelcounts s-1 Amount of Amount of 0 s Procedure sample/g standardhg 0-lh 1-3 h 0 - l h 1-3 h 0.015 0.030 0.060 0.161 0.007 0.012 0.024 0.074 0.098 0.149 0.399 0.109 0.094 5.8 5.4 15.3 34.4 2.3 3.8 8.7 13.4 15.2 26 1 174 14.8 17.1 1.73 f 0.20 3.49 k 0.11 2.01 f 0.08 2.90 f 0.08 2.15 f 0.14 2.31 f 0.07 1.97 f 0.06 4.04 k 0.08 4.61 f 0.18 4.22 f 0.18 1.68 f 0.04 3.52 f 0.12 2.34 f 0.12 6.51 f 2.42 10.1 f 2.1 22.0 f 2.2 8.79 f 1.30 2.25 f 0.72 2.01 f 0.46 2.00 f 0.44 8.28 f 0.48 17.7 k 4.2 31.5 f 3.6 21.2 f 3.0 45.7 f 3.2 18.9 f 4.4 16 OOO 36 OOO 19 600 22 200 9 100 5 520 12 300 6 900 6oOOO 51 100 96 200 49 900 21 700 125 1014 4 047 510 170 80 210 100 360 510 710 570 1600 Distillation 187Od1920s Count rate/ time/min ratio count s-l B-1 0.105 17.6 0-10 4.39 f 0.44 12 400 10-20 4.46 f 0.12 64Ooo 20-30 4.36 f 0.26 15 400 30-100 4.59 f 1.04 270 * Osmium concentrations calculated from time-integrated 1870s/1920s ratios.0 s content (ppb)* 0-lh 275 f 32 261 f 18 201 f 6 245 k 6 264 f 21 293 f 6 281 f 10 291 f 6 289 f 6 295 f 14 286 f 6 19058 171 f 3 0 s content 294 f 30 297 f 12 296 k 12 296 f 12 (PPb)* 0-3 h 279 _+ 36 276 ?I 25 246 f 12 257 f 10 268 f 20 290 f 10 282 f 10 298 f 6 290 f 12 298 f 16 299 f 8 217 k 8 197 4 161154 ANALYST, JULY 1992, VOL. 117 Effects of Microwave Digestion and Oxidation With K2C1-20, The results shown in Fig.2 and Table 3 indicate that, in addition to the low molybdenum concentration in the decom- posed solution, the experimental treatments involving micro- wave digestion and oxidation using K2Cr207 in the digestion and distillation procedures promote osmium isotope hom- ogenization between the sample and added standard. First, the microwave digestion effect will be examined. For this purpose, the data obtained by procedure C-1 are compared with those from procedure A. These two proce- dures, with the same 0 s distillation procedures using K2Cr207 as an oxidizing agent, differ in sample decomposition method, viz., sample decomposition by microwave digestion in proce- dure C-1 and simple heating in a glass ampoule in procedure A.As shown in Table 3, a change in the 187Os/192Os ratio with distillation time is clearly observed in all experiments with procedure A even with small sample sizes, whereas no such changes are observed with procedure C-1 at sample sizes smaller than 30 mg. Further, the calculated osmium abun- dances in procedure A are always lower than those obtained with procedures B-1 and C-1-(a), -(b) and -(c), in which isotopic exchange between the sample and standard is considered to be substantially completed. This result shows that sample decomposition by microwave digestion is useful for complete isotopic exchange between the sample and standard. It is likely that the vibration effect caused by the microwaves prevents polymerization of molybdic acid and also absorption of osmium on molybdic acid.Next, the effect of oxidation with K2Cr207 in the decompo- sition procedures will be examined. The positive effect of this oxidation on the osmium isotope equilibrium is shown in the results with procedure B-1, where K2Cr207 is added at the intermediate stage of microwave digestion and is also con- tained in the solution during distillation. As shown previously, in this procedure, the 1870~/19~Os ratios remained constant with progress of distillation and a reliable osmium abundance is obtained. In this procedure, formation of molybdic acid polymer is further lowered by microwave digestion combined with oxidation using K2Cr207. As a result of less poly- merization, osmium absorption on molybdic acid is suppressed. Finally, the effect of oxidation using K2Cr207 in the distillation procedures will be examined.Comparison of data obtained by procedures C-1, C-2 and C-3 will give an insight into this effect. The difference in these three procedures lies in the distillation conditions: distillation with addition of K2Cr207 in C-1, with addition of HC104 in C-2 and without addition of oxidizing agent in C-3 (Table 2). As shown in Table 3, marked changes in the 187Os/192Os ratios with progress of distillation are observed for procedures C-2 and C-3, whereas the ratios are constant with procedure C-1. Moreover, the osmium abundances obtained using procedures C-2 and C-3 are much lower than the reliable abundances obtained with procedures B-1 and C-1-(a), -(b) and -(c). These results suggest that a considerable amount of radiogenic 1870s derived from MoS2 powder remains trapped on the poly- merized molybdic acid during distillation without the use of K2Cr207 as an oxidizing agent.Osmium in Molybdic Acid Polymer Here the reasons why the complete isotopic equilibrium of osmium is difficult in conventional molybdenite decomposi- tion are considered. Note that, in the decomposed solution of molybdenite, molybdic acid polymer is considered to be easily formed and osmium atoms in the decomposed solution are trapped on or in the molybdic acid polymer. It is also noted that, during distillation, osmium is vaporized as osmium tetroxide. Perfect isotopic exchange between osmium from molybdenite and that from the standard would be prevented by differential trapping of the two types of osmium on or in polymerized molybdic acid and by differential oxidation to Os04 between the two types of osmium.Increases in the 1870s/192 0 s ratio with distillation time are observed in some procedures. These results imply that osmium in molybdenite is more easily trapped in or on molybdic acid polymer than that in the standard prepared from (NH4)20~C16 during decomposition and that the preferential trapping of osmium from molybdenite results in differential vaporization of the two types of osmium during distillation; osmium atoms, which originate from molybdenite and are trapped on or in molybdic acid polymer, are considered to be vaporized at a later stage of the distillation than those of the standard, under weakly oxidative conditions. The difference in the chemical species of osmium between molybdenite and the standard prepared from (NH4)20~C16 is concluded to result in the isotope disequilibrium between them.In order to understand further the effect of the difference in chemical species of osmium between the sample and standard on isotopic exchange, another osmium standard was prepared from osmium metal powder (Koch-Light Laboratories) by dissolving it in a 3 + 1 mixture of 6 mol dm-3 HCl and 12 mol dm-3 HN03, in addition to the osmium standard prepared from (NH4)20sC16 powder. After addition of the standard solution prepared from osmium metal to molybdenite powder, the molybdenite was decomposed in a mixture of HN03 and H2SO4 by the microwave decompostion method. Subsequently, osmium was separated by distillation without addition of an oxidizing agent.This procedure corresponds to procedure C-3 in Table 2. The change in the osmium mass spectrum with progress of distillation for both procedures using the different osmium standards, one prepared from osmium metal and the other from (NH4)20sC16, is shown in Fig. 3. It is observed that the 187Os/192Os ratio decreases gradually with distillation time in the spectra for the mixture of molybdenite and the standard prepared from osmium metal, whereas the ratio increases with time in the spectra for the mixture of molybdenite and the standard prepared from (NH4)20sC16. This opposite tendency of a change in the 187Os/192Os ratio with time might be explained in terms of the difference in the extent of absorption of the different osmium species on the molybdic acid polymer.Osmium in the standard prepared from the metal is absorbed most strongly in the polymer; osmium in molybdenite is more weakly absorbed; osmium in the standard prepared from (N&)20~C16 is the most weakly absorbed. This order might be determined by the atoms surrounding the osmium atoms and their numbers per osmium atom. The species of surround- ing atoms are considered to be oxygen in the standard prepared from osmium metal, sulfur in molybdenite and chlorine in the standard prepared from (NH4)20sC16 (Fig. 4). The number of atoms surrounding one osmium atom is the highest for osmium in the standard prepared from (NH4)20SCl6, where the surrounding chlorine atoms interfere with the formation of 0s-0 bonding and decrease the absorption of osmium onto the molybdic acid polymer.On the other hand, it is natural that the molecule of osmium tetroxide formed in the standard prepared from osmium metal is most easily absorbed on or in the molybdic acid polymer consisting of Mo-0 bonds. Distillation of Decomposed Solution With Less Polymerized Molybdic Acid Under Intense Oxidizing Conditions Using The opposite changes observed in 187Os/192Os ratios demon- strate that the difference in osmium chemical species in the standard solutions results in differences in the mode of isotopic exchange between the sample and standard under weakly oxidizing conditions. On the other hand, the results obtained with procedures B-1 and C-1-(a), -(b) and -(c) show that intense oxidation with K2Cr207 during distillation com- K2Cr207ANALYST, JULY 1992, VOL.117 1155 1 I 1 1 I G ‘5 184 186 188 190 192 I I I I 1 184 186 188 190 192 mlr Fig. 3 Osmium mass spectra of the distillates from the solutions prepared from mixtures of molybdenite and osmium standards: (a mixture of molybdenite and osmium standard prepard from metal; (b{ that of molybdenite and osmium standard prepared from 20sC16. A, &1 h; B, 1-2 h (spectrum enlarged x8); C, 2-3 h X60); and F, 1-3 h (spectrum enlarged x25) enlarged X10); D, 0-0.5 h; E, 0.5-1 h (spectrum enlarged pletes the isotopic exchange between the sample and standard if the formation of molybdic acid polymer is limited. In these procedures, formation of molybdic acid polymer is lowered by a combination of microwave digestion with (1) use of K2Cr207 in the decomposition using procedure B-1 or (2) lowering the molybdenum concentration in the decomposed solution using procedures C-1-(a), -(b) and -(c).In these instances, the Mo-0 bond in the molybdic acid polymer may be broken by Cr2O72,- and subsequently osmium trapped on the polymer becomes free osmium tetroxide. During distillation using K2Cr207, it is observed that the colour of the solution changes from orange to green. This change corresponds to a valence change of chromium from Cr2O72- to Cr3+, endorsing the oxidation of the molybdic acid polymer and osmium. It is also (a) Standard prepared from metal + oso, Partly captured So,ution %@ (b) Standard prepared from (NH4)20sCle OSClS~ - acid dncaptured o Radiogenic 1870s 0 Standard 0s 0 Oxygen Molybdenum 0 Sulfur @ Chlorine Fig.4 Two-dimensional diagram illustratin the difference in atomic arrangement in the decomposed solutions: (a! mixture of molybdenite and osmium standards prepared from the metals; (b) molybdenite and osmium prepared from (NH&OsCk considered that Mo atoms in polymerized molybdic acid are replaced with Cr atoms, and it is likely that the polymerization of molybdic acid is lowered. Hence the addition of K2Cr207 is considered to be effective in two ways. During distillation using K2Cr207 in the solution with less polymerized molybdic acid, osmium released from molybdic acid polymer and osmium not captured in the polymer are distilled together as OsO4 and the resulting 1870s/1920s ratios remain constant, from which reliable osmium abundances can be obtained.However, the osmium abundance obtained using procedure B-2 shows that K2Cr207 oxidation is not effective when the oxidant is added at the beginning of the decomposition of molybdenite. In this instance, most of the K2Cr207 is probably consumed in the oxidation of MoS2 before attacking the Mo-O and 0s-O bonds. Moreover, if the molybdic acid polymer is fully developed, distillation with K2Cr207 is not effective in releasing osmium trapped on the polymer. In these instances, osmium atoms from the standard and molybdenite are differentially distilled and the 1870s/1920s ratios change with distillation time, as observed for procedures A-1 and C-1-(d), -(e), -(f) and -(g). The microwave decomposition technique combined with oxidation by K2Cr207 during decomposition and distillation is concluded to complete the osmium isotope exchange between the sample and osmium standard.The precise osmium abundance in molybdenite can be obtained by this method. Review of the Analytical Procedure Presented by Luck and In the method of Luck and All&gre,g molybdenite was decomposed by HF and HN03, molybdenum oxide or molybdenum fluoride being thought to be formed in the decomposition solution. Like molybdenum oxide, molybde- num fluoride is known to form a polymer easily, e.g., tetramer. Therefore, also in this instance, it is likely that complete isotopic exchange between molybdenite and the standard would be prevented by differential trapping and differential vaporization between osmium atoms from molyb- denite and the standard.For example, the reported Re-Os ages of 6370-6800 million years for Lacorne Preissac molybde- nite might be due to escape of osmium during the chemical procedure before completion of osmium isotope exchange between molybdenite and the standard, in particular, owing to ~11egre91156 Samde ANALYST, JULY 1992, VOL. 117 Table 4 Re-0s age of molybdenite from Matasvaara (Finland). Errors are 2amean l- c- Addition of "35Re spike and 0 s standard I Microwave decomposition (HN03+H2S04) Cooling - Addition of K2Cr207 t Microwave decomposition Distillation 1 0 s solution Addition of ethanol I I Drying up Dilution Anion-exchange resin separation (Dowex 50W I-X8,200-400 mesh) Mo, major elements $..., 1 mol dm-3 HCI 12 mol dm-3 H N 0 3 Drying up Dissolution with 1% H N 0 3 1 1 Re-dissolution C L ICP-MS Fig.5 Analytical procedure for a natural molybdenite preferential escape of osmium derived from the standard during addition of hydrazine and the following drying stage. This preferential escape of osmium might result in an apparent age of 6370-6800 million years. In addition, our careful experiments using their procedure showed that the recovery of osmium is often low in their procedure. In conclusion, it is very difficult to obtain reliable osmium contents and reliable Re-Qs ages using the procedure of Luck and Alkgre.9 R e O s Ages of a Natural Molybdenite From Finland The Re-0s age of a molybdenite from Matasvaara (Finland) was obtained using procedure B-1, i.e..microwave decompo- sition with addition of K2Cr207 in the middle stage of microwave decomposition. The sample was confirmed to contain only radiogenic 1870s as osmium. Therefore, the Re-Os age for this molybdenite can be obtained by determin- ing the rhenium and osmium abundances. The age (7) is given T = {ln[(1870sPRe) + l]}/h by where h is the decay constant of 187Re. The analytical procedure is summarized in Fig. 5 . The osmium abundance was determined by procedure B-1. The Re-0s age by Luck and Re-0s Reference All&gre9/106 Re (ppm) 0 s (ppb) age/l06 years ageA06 years years 25.1 f 0.2 701 f 5 2550 k 90 2700* 1840 k 60t 27.0 f 0.3 728 f 6 2590 2 90 26.6 f 0.1 728 k 10 2650 -C 90 * U-Pb age of zircon.14 t Re-0s age of the m~lybdenite.~ rhenium abundance was determined by the isotope dilution technique using a 185Re-enriched spike with inductively coupled plasma mass spectrometry.The 185Re spike solution was added at the decomposition stage and rhenium was separated from matrix elements on an anion-exchange resin column with 1 mol dm-3 HCl solution and subsequently 12 mol dm-3 HN03. This rhenium separation method is based on that of Huffman et al. 13 Rhenium and osmium abundances and calculated Re-Os ages are listed in Table 4, together with the U-Pb age reported for an associated mineral and the Re-Os obtained by Luck and Allkgre.9 Our Re-0s ages of Matasvaara molybdenite are 2550 f 90, 2590 5 90 and 2650 +_ 90 million years, in good agreement with the reference14 age of 2700 million years. Luck and Allkgreg reported an Re-0s age of 1840 f 60 million years for this mineral. These results clearly demonstrate that procedure B-1 is sufficiently reliable to obtain osmium abundances in molyb- denite. It is concluded that reliable Re-0s ages can be obtained by the method described here. We express our gratitude to Dr. 0. Kouvo, Geological Survey of Finland, for the donation of the mineral sample and for valuable correspondence. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 References Lindner, M., Leich, D. A., Russ, G. P., Bazan, J. M., and Borg, R. J., Geochim. Cosmochim. Acta, 1989,53, 1597. Hintenberger, W., Herr, W., and Voshage, H., Phys. Rev., 1954, 95, 1690. Fleischer, M., Econ. Geol., 1959, 54, 1406. Allegre, C. J., and Luck, J. M., Earth Planet. Sci. Lett., 1980, 48, 148. Walker, R. J., and Fassett, J. D., Anal. Chem., 1986,58,2923. Fehn, U., Teng, R., Elmore, D., and Kubik, P. W., Nature (London), 1986,323,707. RUSS, G . P., Bazan, J. M., and Date, A. R., Anal. Chem., 1987, 59,984. Creaser, R. A., Papanastassiou, D. A., and Wasserburg, G. J., Geochim. Cosmochim. Acta, 1991,55,397. Luck, J. M., and Allkgre, C. J., Earth Planet. Sci. Lett., 1982, 61,291. Suzuki, K., and Masuda, A., Proc. Jpn. Acad., Ser. B, 1990,66, 173. Suzuki, K., Shimizu, H., and Masuda, A., Abstracts Issue, ICOG7, Canberra, Australia, 1990, p. 98. Hirata, T., Shimizu, H., Akagi, T., and Masuda, A., ICP Inf. Newsl., 1988, 13, 731. Huffman, E. H., Oswalt, R. L., and Williams, L. A., J. Inorg. Nucl. Chem., 1956,3,49. Wetherill, G . , Kouvo, O., Tilton, G. R., and Gast, P. W., J. Geol., 1962, 70,74. Paper 11052234E Received October 17, 1991 Accepted February 12, 1992

ISSN:0003-2654    

DOI:10.1039/AN9921701151

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, JULY 1992, VOL. 117 1157 Determination of Nickel in Biological Materials After Mi.crowave Dissolution Using Inductively Coupled Plasma Atomic Emission Spectrometry With Prior Extraction into Butan-1-01 Elisa Vereda Alonso, Amparo Garcia de Torres and Jose M. Can0 Pavon* Department of Analytical Chemistry, Faculty of Sciences, University of Malaga, 29071 Malaga, Spain A sensitive procedure has been developed for the determination of ultratrace amounts of nickel in biological materials by inductively coupled plasma atomic emission spectrometry after extraction of the nickel ion into butan-1 -01 by using 1,5-bis(di-2-pyridylmethylene)thiocarbonohydrazide as the extracting reagent. Fast, efficient and complete sample digestion is achieved by an HN03-HCI poly(tetrafluoroethy1ene) bomb dissolution technique using microwave heating.Results obtained for eleven certified reference materials agreed with the certified values. Keywords: Nickel determination; inductively coupled plasma atomic emission spectrometry; 7,5-bis(di-2-pyridylmethylene)thiocarbonohydrazide; extraction; biological materials Liquid-liquid extraction is one of the most frequently used sample pre-treatment techniques in the determination of trace metals by inductively coupled plasma atomic emission spec- trometry (ICP- AES) .I,* 1,5-Bis( di-2-pyridylmethylene) thio- carbonohydrazine (DPTH) is a suitable complexing reagent for a number of metal ions. Most of the complexes formed by this reagent are coloured and can be extracted into a variety of organic solvents.Its synthesis and properties have been described in detail,3 as has its use in the extraction of cobalt into chloroform, followed by the determination of this metal by flame atomic absorption spectrometry.4 It has also been used for the spectrophotometric determination of cobalt- nickel mixtures.5 In this paper, an ICP-AES method for the determination of ultratrace amounts of nickel after extraction of the nickel ion into butan-1-01 (a solvent commonly used in ICP-AES) containing DPTH as the extracting reagent is described. The complex formed is soluble in the organic phase, which allows the use of aqueous-to-organic phase volume ratios of up to 8. In addition, the extraction step enhances the selectivity. The method thus developed was applied to the determination of nickel in various biological materials.Nine certified reference materials prepared by microwave digestion in sealed poly- (tetrafluoroethylene) containers and two certified human urines without mineralization were analysed. The use of closed vessels for the decomposition of biological samples, proposed by Kingston and Jassie6 and Aysola et al. ,' makes it possible to eliminate uncontrolled trace element losses of volatile molecular species that are present in a sample and leads to a decrease in blank values compared with open- beaker work, because contamination from the laboratory environment is much lower and also because closed vessels make it possible to use smaller amounts of reagents.8 Experimental Apparatus The ICP-AES measurements were made on a Perkin-Elmer 40 sequential emission spectrometer controlled by an IBM XT-286 computer.A Meinhard nebulizer was used each time butan-1-01 was employed. The r.f. generator of the nebulizer is internally mounted with a 'free-running' oscillator-type with a nominal central frequency of 40 MHz and a nominal operating power level of 900 W. The pH measurements were made with the aid of a Crison Digit-501 pH meter furnished * To whom correspondence should be addressed. with a combined glass+alomel electrode. Separating funnels were shaken on a Gallenkamp flask agitator. For sample digestion, a domestic microwave oven, Pana- sonic Model NN-8507/8557 with a 700 W magnetron, was used. The oven power (0-100% power in 20% increments) and the time range (0 s to 99.0 min) can be pre-set digitally; up to four sequential steps of different power and duration can be programmed.The oven used has a rotating antenna, so that the power is distributed homogeneously throughout the oven volume. The oven was placed in a laboratory fume-hood. Reagents All reagents used were of analytical-reagent grade. Doubly distilled, de-ionized water was used throughout. The ligand for the DPTH solution was synthesized as described elsewhere.3 A stock 0.1% solution in butan-1-01 was prepared by dissolving 0.1 g of DPTH in 9 cm3 of N,N-dimethylformamide and diluting to 100 cm3 with butan- 1-01. A stock solution of Nil* was prepared from the nitrate and standardized gravimetrically with dimethylglyoxime. Stan- dards of working strength were prepared by appropriate dilution as required.Standard solutions were prepared daily. A glycine-HC1 buffer of pH 3.6 was prepared by mixing 25 cm3 of 0.2 mol dm-3 glycine and 2.5 cm3 of 0.2 mol dm-3 HC1 in a 100 cm3 calibrated flask and diluting to the mark with water. A 1 mol dm-3 NaC104 solution was also used. Procedures Liquid-liquid extraction of nickel In 100 cm3 separating funnels were placed between 0.3 and 20 pg of Ni" plus 1 cm3 of 1 mol dm-3 NaC104, 5 cm3 of the glycine-HC1 buffer of pH 3.6 and distilled water up to a final volume of 20 cm3. Then, 10 cm3 of the 0.1% butan-1-01 solution of DPTH were added and the funnel was shaken vigorously on the mechanical agitator for 10 min, after which the phases were allowed to separate and the organic phase was collected. The organic phase was pumped into the plasma bulk by means of a peristaltic pump for the determination of nickel.Emission readings for nickel were taken on the 341.476 nm line (this wavelength was chosen as it resulted in the largest differences between the samples and the noise signals). A calibration graph was prepared from nickel standards treated in the same way.1158 ANALYST, JULY 1992, VOL. 117 Table 1 Working conditions of the microwave oven used with the different samples studied Sample* Power and time No. of treatments BCR CRM 062 Olive Europoea (Olive Leaves) BCR CRM 060 Lagarosiphon Yajor (Aquatic Plant) BCR CRM 061 Platihypnidium Ripariodides (Aquatic Plant) NRCC CRM DORM-1 Dogfish Muscle NRCC CRIPISOLT-1 Dogfish Liver NRCC CRM TORT-1 Lobster Hepatopancreas NIST SRM 1572 Citrus Leaves BCR CRM 186 Pig Kidney BCR CRM 184 Bovine Muscle 2minat480 W + 4min at 180 W 2.5 rnin at 480 W + 4.5 rnin at 360 W 2min at 700 W + 4minat 360 W 4minat360W+ 10minat180W 4minat360W + 10minat 180W 4minat360W+ 10minat180W 4minat360W+ 10minat180W 4 min at 360 W + 10 min at 180 W 4minat360W+ 10minat180W * BCR = Community Bureau of Reference; NRCC = National Research Council Canada; NIST = National Institute of Standards and Technology; CRM = certified reference material; and SRM = standard reference material.100 g 75 5 50 - C .- c c w 25 2 4 6 8 1 0 1 2 PH Fig. 1 Extraction of nickel into butan-1-01 versus pH using DPTH as extracting reagent. Concentration of nickel in the aqueous phase: 2 pg cm-3; volume of aqueous phase: 20 cm3; volume of organic phase: 10cm3 Determination of nickel in biological samples The samples were placed in Parr 478 microwave acid digestion bombs. The bombs were cleaned before use with 2 mol dm-3 nitric acid for 1 d followed by repeated rinsing with water.All glassware used was treated in the same way. Samples (about 200 mg) were weighed directly into the digestion vessels on a digital electronic balance and 4 cm3 of concentrated nitric acid were added. The mixture was allowed to stand for 30 min after which 2 cm3 of concentrated hydrochloric acid (12 mol dm-3) were added. The bomb was sealed and placed in the microwave oven with a beaker filled with 10 cm3 of water. This procedure minimizes the risk of damage to the magnetron in that small sample loads can cause magnetron failure, because most of the microwave power is reflected back to the waveguide and magnetron.The working conditions of the microwave oven are listed in Table 1 . After digestion, the solutions were evaporated by heating to a small volume (about 0.5 cm3) to eliminate the nitric acid. The pH was adjusted to 3.6 with 1 mol dm-3 NaOH solution and 10 cm3 of buffer plus 2 cm3 of 1 mol dm-3 NaC104 solution were added. The solutions were diluted to 250 cm3 with de-ionized water in a calibrated flask. Then, in a 250 cm3 separating funnel were placed 125 cm3 of sample solution and 15 cm3 of a 0.05% butan-1-01 solution of DPTH. Thefunnel was shaken vigorously on the mechanical agitator for 5 min, after which the phases were allowed to separate and the organic phase was collected in order to measure the Ni concentration by ICP-AES at 341.476 nm, as described above.Human urine can be analysed without mineralization. The urine was acidified with concentrated nitric acid until 1% v/v and stored frozen until required for the determination of nickel. In 250 cm3 separating funnels were placed 100-150 cm3 of acidified urine, which was analysed as described above. In all instances it is convenient to analyse the reagent and buffer solutions previously in order to test for the presence of nickel. 100 s. ; 75 0 .- w 5 50 25 c W I I I I I I I I I I I I I I I 2 4 6 8 10 2 4 6 8 0.5 1.0 1.5 2.0 2.5 3.0 [NaC1041/10-2 Shaking time/min [DPTH]/10-3 rnol dm-3 mot dm-3 Fig. 2 Influence of experimental variables on the extraction of the NP-DPTH com lex into butan-1-01.( a ) Influence of perchlorate concentration; (8, influence of shaking time; and (c) influence of reagent concentration (organic phase). Nickel concentration, 2 pg ~ m - ~ (aqueous phase); initial volume of aqueous phase, 20 cm3; initial volume of organic phase, 10 cm3 (total in each instance) Results and Discussion Effect of pH on the Extraction The influence of pH on the extraction of the NP-DPTH complex into butan-1-01 was studied over the range 1.0-12.5. The results (Fig. 1) showed that the extraction is quantitative between pH 3.0 and 7.0, and that it takes place to a much smaller extent outside this range. Selection of the Buffer Solution Three buffer solutions were tested: one of pH 3.6 consisting of glycine-HC1 and two of pH 3.8 consisting of potassium hydrogen phthalate-HC1 and acetic acid-soldium acetate.All yielded very similar results. Nevertheless, the glycine-HC1 buffer was chosen on account of its greater complexing ability, which could be of use in overcoming interferences. Effect of Perchlorate Concentration, Shaking Time and Reagent Concentration The extraction of the nickel-DPTH complex into an acidic medium was investigated by varying the NaC104 concentra- tion up to 6.5 x 10-2 mol dm-3. The extracted fraction increased on increasing the perchlorate concentration to 3 X 10-2 mol dm-3 [Fig. 2(a)]; however, further increases in the concentration, which were obtained by adding 1 cm3 of 1 mol dm-3 NaC104 to an aqueous volume of 20 cm3, had no appreciable effect for subsequent experiments.A study of the influence of the shaking time [Fig. 2(b)] revealed that the extent of extraction levelled off after 3 min. A shaking time of 5 rnin was selected for subsequent experiments. Next, the reagent concentration in the organic phase was varied while keeping the final volume at 10 cm3. The-results obtained [Fig. 2(c)] showed that the extracted fraction remained constant for DPTH concentrations equal to orANALYST, JULY 1992, VOL. 117 1159 greater than 6.83 x 10-4 mol dm-3 (0.03%). A concentration of 0.05% was used in practice, in order to prevent depletion of the reagent by other extractable ions that might be present in the aqueous medium. Effect of Phase-volume Ratio The volume of the aqueous phase was varied from 10 to 150 cm3 (0.05% m/v DPTH); phase volume ratios between 1 and 15 were tested.Nickel could be quantitatively extracted up to a phase ratio of 8, above which phase separation was unsatisfactory and the procedure was, therefore, inapplicable. On the other hand, when the volume of the aqueous phase was increased, the volume of the organic phase decreased, because butan-1-01 is not totally miscible with water; however, the quantitative extraction of nickel could be verified from the calibration graphs. When a phase ratio of 8 was used (80 cm3 of aqueous phase and 10 cm3 of organic phase), only approximately 2.5 ml of organic phase were obtained after extraction; hence the phase ratio is actually 32. This phase ratio allows the sensitivity of the direct non-extractive method to be increased by a factor of approximately 32.Determination of Nickel by ICP-AES A wavelength of 341.476 nm was selected as it resulted in the largest differences between the sample and the noise signals (minimum background equivalent concentration). The nebu- lizer was then optimized: the optimum signal-to-noise ratio was obtained at a value of 1.9 on an arbitrary scale from 0 to 6; the plasma gas flow rate was 12 dm3 min-1; and the auxiliary gas flow rate was 0.5 dm3 min-l. Other experimental variables were optimized, for which the following values were selected: torch height, first position on an arbitrary scale from 1 to 4 (the fourth position is the ignition position); photomultiplier voltage, 750 V; element time, 360 ms; integration time, 100 ms; read delay, 15 s; and spectral range, 1-00 nm.Table 2 Tolerated levels of foreign species in the determination of NiII* Species Tolerated ratio ( d m ) Alkali metal ions, alkaline earth metal ions, Cr1I1, AP, bromide, chloride, fluoride, nitrate, sulfate, phosphate, thiourea, thiosulfate >4OOo Mo"', Vv, Selv, ascorbic acid, oxalate 3000 As111 ? , pbll MnII 9 Bill1 2000 Hg",Agl,Sblll,Cull 1500 Fell, Fell1 1000 Zn" 200 * Nil1 = 10 ng ~ r n - ~ in the aqueous phase. The limits of detection and determination of the method were established according to American Chemical Society Committee of Environmental Improvement definitions.9 The detection limit thus achieved for an aqueous-to-volume phase ratio of 8 was 0.2 ng cm-3 (aqueous phase) and the determination limit was 1.7 ng cm-3 (aqueous phase).Relative standard deviations were 3.4% for 2.5 ng cm-3 and 1.8% for 25 ng cm-3 (all concentrations are referred to the aqueous phase). Study of Interferences The effect of various ions on the determination of nickel by the proposed method was examined under the optimum working conditions. For this study, different amounts of the ionic species tested were added to a 10 ng cm-3 solution of Ni in the aqueous phase, the volume of which was 160 cm3, whereas that of the organic phase was 20 cm3. The starting point was an m/m interferent-to-Ni ratio of 4000; if any interference occurred, the ratio was gradually lowered until the interfer- ence disappeared. The tolerance limits found (Table 2) show that nickel can be determined in the presence of a variety of ions including most of those that commonly occur with nickel in natural and synthetic samples.Cobalt interferes with the determination of nickel up to at least a ratio of 40; hence, another wavelength, 352.454 nm, was chosen to eliminate this interference. With this wavelength, the tolerated ratio of cobalt can be increased to 1000. Determination of Nickel at 352.454 nm The different instrumental variables for this wavelength were optimized. The following values were selected: nebulizer flow rate, 2.0 on an arbitrary scale from 0 to 6; plasma gas flow rate, 12 dm3 min-1; auxiliary gas flow rate, 0.5 dm3 min-1; torch height, first position on an arbitrary scale from 1 to 4; photomultiplier voltage, 900 V; element time, 360 ms; integration time, 40 ms; read delay, 15 s; and spectral range, 1.00 nm.The detection limit thus achieved for an aqueous-to-volume phase ratio of 8 at this wavelength was 3.2 ng cm-3 (aqueous phase) and the determination limit was 8.1 ng cm-3 (aqueous phase). Relative standard deviations were 3.9% for 10 ng cm-3 and 3.2% for 25 ng cm-3 (all concentrations are referred to the aqueous phase). Analysis of Biological Materials In order to ascertain the accuracy of the method, the optimized procedure was applied to the following certified biological materials: Community Bureau of Reference Table 3 Application of the proposed method to the determination of nickel in biological samples Sample BCR CRM 062 Olive Europoea (Olive Leaves) BCR CRM 060 Lagarosiphon Major (Aquatic Plant) BCR CRM 061 Platihypnidium Ripariodides (Aquatic Plant) NRCC CRM DORM-1 Dogfish Muscle NRCC CRM DOLT-1 Dogfish Liver NRCC CRM TORT-1 Lobster Hepatopancreas NIST SRM 1572 Citrus Leaves BCR CRM 186 Pig Kidney BCR CRM 184 Bovine Muscle NIST SRM 2670 Freeze-Dried Urine (Toxic Metals at Normal Levels) NIST SRM 2670 Freeze-Dried Urine (Toxic Metals at Elevated Levels) * Mean of three separate determinations. t Values in pg crn-3.Ni certifiedpg g-1 8.0 40.0 420.0 1.20 0.26 2.3 0.6 0.42 0.27 0.07t 0.30t Ni found*/pg g-1 7.9 zk 0.4 38.8 f 1.7 430.2 k 30.0 1.40 k 0.15 0.31 f 0.11 2.6 f 0.1 0.6 _+ 0.1 0.39 * 0.03 0.30 f 0.01 0.09 * 0.02-l 0.28 f 0.Olt1160 ANALYST, JULY 1992, VOL. 117 (BCR), Certified Reference Materials (CRMs) 062 Olive Europoea (Olive Leaves); 060 Lagarosiphon Major (Aquatic Plant); 061 Platihypnidium Ripariodides (Aquatic Plant); 186 Pig Kidney; 184 Bovine Muscle; National Research Council Canada (NRCC), CRMs DORM-1 Dogfish Muscle; DOLT-1 Dogfish Liver; TORT-1 Lobster Hepatopancreas; National Institute of Standards and Technology (NIST), Standard Reference Materials (SRMs) 1572 Citrus Leaves; and 2670 Freeze-Dried Urine (Toxic Metals at Normal and Elevated Levels).The results obtained are given in Table 3, and agree well with the certified values. The limits of detection found for the different biological samples dissolved in the aqueous phase are (in ng (3111-3): CRM 062 Olive Leaves, 0.4; CRM 060 Aquatic Plant, 0.2; CRM 061 Aquatic Plant, 0.2; DORM-1 Dogfish Muscle, 1.2; DOLT-1 Dogfish Liver, 1.7; TORT-1 Lobster Hepatopancreas, 1.1; SRM 1572 Citrus Leaves, 0.5; CRM 186 Pig Kidney, 1.1; CRM 184 Bovine Muscle, 0.9; and SRM 2670 Freeze-Dried Urine, 0.5. We thank the Direccion General de Investigacidn Cientifica y Tkcnica (DGICYT) for supporting this study (Projects PB87- 0711 and PB90-0809), and also the Junta de Andalucia. 5 6 7 8 9 References Miyazaki, A., Kimura, A., Banzho, K., and Umezaki, Y., Anal. Chim. Acta, 1982, 114,213. Whiteley, R. V., and Merril, R. M., Fresenius’ Z . Anal. Chem., 1982, 311, 7. Bonilla Abascal, J. R., Garcia de Torres, A., and Can0 Pavon, J. M., Microchem. J., 1981,26,55. Bustos, A., Sanchez-Rojas, F., Bosch Ojeda, C., Garcia de Torres, A., and Can0 Pavon, J. M., J. Anal. At. Specrrom., 1987, 2,253. Can0 Pavon, J. M., Garcia de Torres, A., and Bosch Ojeda, C., Analyst, 1985, 110, 1137. Kingston, H. M., and Jassie, L. B., Anal. Chem., 1986,58,261. Aysola, P., Anderson, P., and Langford, C. H., Anal. Chem., 1987,59, 1582. Introduction to Microwave Sample Preparation, eds. Kingston, H. M., and Jassie, L. B., American Chemical Society, Washing- ton, DC, 1988. ACS Committee of Environmental Improvement, Anal. Chem., 1980,52,2242. Paper I f06438J Received December 23, I991 Accepted February 26, 1992

ISSN:0003-2654    

DOI:10.1039/AN9921701157

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, JULY 1992, VOL. 117 1161 Determination of Butyltin Species in Sewage and Sludge by Gas Chromatography-Atomic Absorption Spectrometry Y. K. Chau, Shuzhen Zhang* and R. J. Maguire National Water Research Institute, Burlington, Ontario, Canada L7R 4A6 A method for the extraction and determination of butyltin compounds in sewage and sludge is reported. Sewage and sludge samples are acidified and shaken for 2 h. The various butyltin species are extracted quantitatively by tropolone (cycloheptatrienone) in toluene, followed by ethylation to their tetraalkyl- substituted forms, BuSnEt3, BupSnEt2, Bu3SnEt and Et4Sn, all of which can be separated and determined by a gas chromatographic-atomic absorption spectrometric technique. The non-pesticidal octyltin species and acid-leachable SnIV species can also be determined by this method.Detection limits expressed as Sn are 40 ng dm-3 and 2 ng g-1 dry mass for sewage and sludge, respectively. Analyses of some samples from Canadian treatment plants are given. Keywords: Monobutyltin; dibutyltin; tributyltin; sewage; sludge Organotin compounds are widely used in industry as poly- (vinyl chloride) stabilizers, industrial and agricultural bio- cides.l.2 Butyltin compounds have been found in various types of environmental samples as a result of their use as anti- fouling agents in paint formulation.3.4 Recently, a study has been carried out in this laboratory to examine the occurrence of butyltin residues in sewage treatment plants before and after treatment, in order to assess their treatability in such plants.Sewage samples consist of extremely complex and unknown matrices, and digestion is often required to release the analytes before analysis. Analy- tical procedures designed for speciation are not usually compatible with rigorous acid digestion. Limited information on the occurrence and analysis of butyltin species in sewage and sludge materials is available$ application of this work is described in a study of the occurrence of organotin compounds in a Swiss treatment plant.6 There is no information on the extraction of butyltin compounds from sewage and sludge. The present study reports the evaluation of several digestion/ extraction methods with the aim of developing an analytical procedure for the speciation of the different butyltin com- pounds in the influent, effluent, and sludge samples from municipal sewage treatment plants. These methods are being used in a survey of the occurrence of butyltin compounds in several major municipal treatment plants in Canada.Experimental Materials and Methods Monobutyltin chloride, dibutyltin dichloride and tributyltin chloride were obtained from Alfa Products, Ward Hill, MA, USA. Tropolone (cycloheptatrienone) and ethylmagnesium bromide (2.0 mol dm-3 in tetrahydrofuran) were supplied by Aldrich, Milwaukee, WI, USA. Standard butyltin solutions (lo00 pg cm-3 as Sn) were prepared by dissolving appropriate amounts of the butyltin compounds in water. The internal standard additions technique was used for the calculation of the concentrations of different butyltin species.The butyltin species in sewage were first extracted into tropolone, followed by ethyl-derivatization to the tetraalkyl- substituted forms to be determined by a gas chromatographic- atomic absorption spectrometric (GC-AAS) method.7 The system makes use of the separation power of the gas chromatograph and the element-specific atomic spectrometer as a detector. The GC-AAS system is element- and species- * On leave from the Center for Eco-Environmental Sciences, Academia Sinica, Beijing, China. specific, and is highly sensitive. The construction of the system and its application to the speciation of organometallic and organometalloidal compounds has been well documented.8 The gas chromatographic conditions for methylbutyltin species have been described in a previous pub1ication;g the following modifications were made for the separation of ethylated butyltin compounds.A 30 m fused silica wide bore capillary column, 0.53 mm i.d., with a 1.5 nm film thickness of dimethylpolysiloxane coating was used (Cat. No. 2-5303, Supelco, Bellefonte, PA, USA). The oven temperature was initially 90"C, and was programmed to increase at a rate of 18°C min-1 to a final temperature of 220°C at which it remained for 2 min. The nitrogen carrier gas flow rate was 30 cm3 min-1, the injection port temperature was llO"C, and the transfer line connecting the GC and AAS instruments was at a temperature of 165 "C. Parameters for the AAS instrument were as follows: the electrodeless discharge lamp (Sn) was operated at 8 mV and the 224.6 nm line was used; the furnace temperature was 850-900°C; furnace gas flow rates were, H2, 130 cm3 min-1 and air, 26 cm3 min-1; a deuterium background corrector was used.Peak areas were integrated with an HP 3392A integrator. Ionic organometallic compounds such as the monobutyltin BUS++, dibutyltin Bu2Sn2+, tributyltin Bu3Sn+, and the inorganic SnlV species generally have rather high boiling- points; a derivatization step is required to lower their boiling-points to values appropriate for GC separation. In connection with this, alkylation with a Grignard reagent (R'MgBr) to convert the ionic organometallic species, R,Sn@-l)+, and the inorganic SnIv species to their tetraalkyl- substituted forms, RnR'(4-n)Sn and R'4Sn has been commonly adopted. Ethylation with ethylmagnesium bromide has been found to give derivatives with appropriate boiling-points for GC operation.This derivatization does not change the identity of the original alkyl group, thus making it possible to identify the methyl- and butyltin species. Details of derivatiza- tion techniques for speciation of ionic organometallic com- pounds have been discussed in a recent review.8 The butyltin and SnlV species were extracted into tropolone followed by ethyl-derivatization to the tetraalkyl-substituted forms for GC-AAS determination. Dioctyltin species can also be determined by this technique. A chromatogram showing the various Sn species is shown in Fig. 1. Sewage samples used for the investigation were influent samples obtained from the Toronto Humber and Greater Vancouver Sewage Treatment Plants.Sewage samples were cidified to pH 1 immediately after collection and stored in a cold room at 5 "C in brown bottles until use. Sludge samples were not treated, but were stored in a cold room after1162 ANALYST, JULY 1992, VOL. 117 t i 2 4 6 8 1 0 1 2 Ti me/m i n Fig. 1 Chromatogram showing the ethyl derivatives of Snlv and butyltin species. 1, SnEt4; 2, BuSnEt3; 3, Bu2SnEt2; 4, Bu3SnEt; and 5 , Oct2SnEt2. Each peak represents approximately 2 ng as Sn collection. For assessment of extraction recoveries, the samples were spiked with 2.0 yg of each of the butyltin standards and equilibrated by shaking for 2 h before analysis. Digestion of sludge samples by acid treatment at room temperature Samples (150 cm3) of a well mixed sludge, after spiking with known amounts of each of the butyltin species, were first subjected to acid treatment by adding approximately 10 cm3 of concentrated hydrochloric acid in a 250 cm3 separating funnel and then allowed to stand inside a fume cupboard for 1-2 h until the gas evolution subsided. After adjustment of the pH of the solution to 1-2, the samples were shaken mechanically for 2 h and then extracted with 20 cm3 of 0.5% tropolone in toluene for 4 h with addition of 60 g of NaCl.An aliquot (5 cm3) of the extract was removed and evaporated almost to dryness on a heating block at 30°C with the aid of a nitrogen stream. The volume of the extract was adjusted to 1 cm3 with hexane and the mixture was allowed to react with 0.2 cm3 of ethylmagnesium bromide (2.0 mol dm-3 in tetrahydrofuran) for 10 min.The excess of ethylmagnesium bromide was destroyed by shaking with 2 cm3 of 0.5 mol dm-3 sulfuric acid. The hexane phase containing the derivatized butyltins and SnIV was subjected to a clean-up procedure by loading 0.5 cm3 of the hexane extract onto a glass column (15 x 1.5 cm i.d.) containing approximately 6 g of silica gel (containing 5% water), packed from bottom to top in the following order: glass wool, 1 cm of anhydrous Na2S04, silica gel bed, 1 cm of Na2S04. The butyltin compounds were eluted with 30 cm3 of hexane at a rate of 0.5 cm3 min-1. After reduction of the eluate volume to 0.5 cm3 in a rotary evaporator, a suitable aliquot was injected into the GC-AAS system for the determination of the butyltin and SnIV species.The results were expressed as yg of Sn per gram of sludge on a dry mass basis. The dry mass of sludge was obtained by drying several well-mixed aliquots of sludge samples at 30°C to obtain their averaged dry mass. Digestion of sludge samples by acid reflux After acid treatment as described above, samples (150 cm3) of sludge were placed in 500 em3 round-bottomed flasks and heated under reflux at approximately 60°C for 2 h. After cooling, the samples were extracted with 20 cm3 of 0.5% tropolone in toluene for 4 h with addition of 60 g of NaCl. The toluene extract was evaporated almost to dryness in a rotary Table 1 Recovery of butyltin species from sludge by different extraction methods Recovery (% ) * Toluene -t 46 82 Dichloromethane - - - 0.5% Tropolone- 76 hexane - - 0.5% Tropolone- dichloromethane - - 103 0.5% Tropolone- toluene 96 90 114 0.5% Tropolone- toluene * Recovery (%) is evaluated by spiking 2.0 pg of each of th (after acid reflux) 88 68 95 Extractant/solvent Sn4+ BuSn3+ Bu2Sn2+ Bu3Sn+ 94 - 49 113 107 95 butyltin and SnIV species into 150 cm3 (approximately 4.5 g dry mass) of sludge samples.After blank correction, the values are compared with extraction of the same amount of standards without the sludge. The recovery represents absolute recovery from sludge materials. Recovery is the average of two replicates. t (-) No recovery. evaporator and made up to 1 cm3 with hexane for subsequent derivatization and clean-up as described above. Extraction efficiency for butyltin species from sludge The room temperature acid-treated samples of sludge (150 cm3) were extracted once with 20 cm3 of the following extractants for 4 h with mechanical shaking after the addition of 60 g of NaCl.After extraction, the samples were processed in a similar manner to that described under Digestion of sludge samples by acid reflux to evaluate their extraction recoveries. The extractants tested were: 1, toluene; 2, dichloromethane; 3, 0.5% tropolone in hexane; 4, 0.5% tropolone in dichloromethane; 5,0.5% tropolone in toluene; and 6,0.5% tropolone in toluene after acid reflux for 2 h. Extraction of butyltin species from sewage samples Sewage samples (200 cm3) were acidified to pH 1 with HCl at the time of collection. The pH was checked and the dissolved gases were allowed to evolve before the extraction process.The sample was extracted with 20 cm3 of 0.5% tropolone in toluene for 4 h after adding 60 g of NaCl. The extract was processed and ethylated in the same manner as described above for the sludge samples. The clean-up was carried out using a Pasteur pipette as a column (15 cm X 5 mm i.d.) containing approximately 1 g of silica gel (5% water), packed in a similar manner as for the regular column described above, and eluted with 25 cm3 of 10% diethyl ether in hexane at a rate of 0.7 cm3 min-1. After reduction of the eluate volume to 0.5 cm3, the extract was analysed in the GC-AAS system. Results and Discussion Sewage and sludge materials consist of complex and unknown matrices, and must be subjected to proper digestion proce- dures in order to break down the organic materials to release the molecular and ionic butyltin compounds.In any direct speciation procedure, the digestion procedure must be care- fully chosen so as to destroy the matrices but not to affect the authentic forms of the analyte molecules. Of all the extractants investigated for the extraction of butyltin compounds from the acid-treated sludge (Table 1), toluene alone can be used to recover the BuZSn and Bu3Sn species but not the BuSn or Sn'" species. Tropolone must be used as extractant and toluene as solvent for quantitative recovery of BuSn and Sn'" species. It is imperative that the sludge samples be acidified prior to extraction for two reasons:ANALYST, JULY 1992, VOL. 117 1163 firstly to expel all gases from the sample, and secondly to release the analytes from the particulate matter.Similar phenomena were also observed in the extraction of butyltin compounds from sediment .lo Acid treatment at room temper- ature released the butyltin species as efficiently as the acid reflux technique. For the sake of convenience, acid treatment at room temperature is recommended as the procedure for sludge digestion. A second extraction with any of the extractants did not improve the recovery to any significant extent. In the acid-reflux sample treatment, more inorganic SnIV was released than in the 2 h acid treatment at room temperature. The Snrv species were still found in three consecutive extractions with tropolone in toluene, whereas no significant amounts of butyltin species were found in the second extraction.Therefore, the Snrv values from the acid sample treatment technique can only be regarded as acid- Table 2 Recovery of butyltin compounds from spiked sewage and sludge samples Recovery (% ) Spi ke/pg Sn4+ BuSn3+ Bu2Sn*+ Bu3Sn+ Sewage samples*- 1 110 101 99 99 2 91 101 107 90 5 106 109 109 104 Sludge samples-t- 1 84 103 95 95 2 105 100 89 113 5 91 96 94 91 * Sewage (effluent) samples were from the Toronto Humber Sewage Treatment Plant with the following butyltin content: Sn'", 5.2 pg dm-3; no detectable butyltin species. Spikes were added to 200 cm3 of sewage samples, equilibrated by shaking for 2 h before determina- tion. Results given are averages of independent duplicate analyses. t Sludge samples were from the Vancouver Sewage Treatment Plant with the following butyltin content: Sniv, 5.9 pg 8-l dry mass; no detectable butyltin species.Spikes were added to 100 cm3 (approxi- mately 3.0 g dry mass) of the well shaken sludge suspension, equilibrated by shaking for 2 h before analysis. Results given are averages of independent duplicate analyses. leachable SnIV. If accurate determination of SnIV is required, acid-reflux sample treatment must be used and repeated extractions should be performed up to three times, or alternatively, the SnIV may be determined as total Sn by atomic absorption methods after rigorous digestion. The effect of extraction time was investigated by shaking the sludge samples with a 0.5% tropolone in toluene solution for 2 and 4 h intervals. It was observed that shaking for 2 h allowed recovery of only 54% of the monobutyltin species, while all of the other butyltin species in sludge were recovered quantita- tively after 4 h shaking.For thorough extraction, a 4 h period was also adopted for sewage samples. Clean-up for sludge must be incorporated in the procedure. Silica gel (5% water) was found to be satisfactory for this purpose. A larger capacity column was necessary for sludge, whereas a smaller, more convenient Pasteur pipette could be used as the column for sewage samples which contained relatively small amounts of organic matter. As sludge contains a more complex matrix, the volume of the eluent used must be investigated for a particular type of sludge. Hexane could be used to release all of the butyltin and SnIV species without co-eluting the organic matter.Under the present conditions, the 15 x 1.5 cm i.d. column was sufficient to retain the organic matter in a brown band 2 cm wide at the top when 150 cm3 (approximately 4.5 g dry mass) of sludge suspension were used and eluted with 35 cm3 of hexane. In a smaller column operation for sewage samples, there was less coloured material and a light-brown band was observed. The extraction procedure for sludge was readily applicable to sewage samples. In both instances, acidification of the sample was necessary. The recovery of SnIV and butyltin species from sewage and sludge was investigated by spiking known amounts of the compounds into samples of relatively low butyltin content, and carrying out replicate analyses. It was observed that the recovery ranged from 84 to 114% at a spike level of 1-5 pg (Table 2).The precision of the method was assessed by running replicate (n = 6) analyses of spiked samples contain- ing 2 pg each of SnIV and butyltin species in 150 cm3 of sewage and in 100 cm3 of sludge, respectively. Relative standard deviations for sewage at the 13.33 pg dm-3 level were 4,6,11 and 8%, respectively, for the SnIV, BuSn3+, Bu2Sn2+ and Table 3 Analysis of butyltin and acid-leachable Sniv species in sewage and sludge in some Canadian treatment plants* Sn4+/pg dm-3 BuSn3+/pg dm-3 Bu2Sn2+/pg dm-3 Bu3Sn+/pg dm-3 Sewage Influent Effluent Influent Effluent Influent Effluent Influent Effluent Plant A- - - Sept. 90 --t 10.7 f 0.7 8.7 f 0.1 2.4 t- 0.1 2.0 f 0.3 Dee. 90 - - 10.3 f 0.6 5.2 f 0.4 - - - - - Plant B- Sept.90 - - 3.4 k 0.3 3.3 f 0.3 - - - - Sept .90 - - 13.7 f 0.9 6.9 f 0.1 - - - - Dee. 90 - - 5.5 f 0.4 1.3 f 0.3 - - - - Plant C- Dec. 90 27.5 f 2.1 9.2 f 0.2 5.7 f 1.1 9.5 f 0.2 - - - - Sludge Plant A- Sept. 90 Dec. 90 Plant B- Sept .90 Dec. 90 Plant C- Sept .90 Dec. 90 Sn4+/ BuSn3+/ Bu2Sn2+/ Bu3Sn+/ pgg-' kg-' pg kg- CLg kg-' - - - 9.3 k 0.5 24.0 f 0 440 f 30 210 f 35 277.5 f 23 5.8 f 0.3 - 20.5 f 0.5 130.5 f 2 3.2 f 0.4 16.4 f 0 35.0 f 0.5 45.5 f 3 192.6 f 7.4 - - - - - - 40.7 f 0 * Results given are averages of duplicate analyses; sewage samples, 200 cm3; sludge samples, 150 cm3 suspension; results expressed on dry mass -t (-) Not detected. Detection limits (as Sn) for sewage, 40 ng dm-3; for sludge, 2 ng g-1 dry mass. basis, using density of sludge for calculation.1164 Bu3Sn+ species.Relative standard deviations for sludge samples spiked at 20 pg cm-3 were 10,8,9 and 9% for SnIV, BuSn3+, Bu2Sn2+ and Bu3Sn+, respectively. Detection limits (as Sn) as measured using a signal-to-noise ratio = 3 and under the present operating conditions and sample sizes were 40 ng dm-3 and 2 ng g-1 dry mass for sewage and sludge, respectively. Analyses of some samples from Canadian treatment plants are given in Table 3 to illustrate the application of these methods. In all the samples taken from five sewage treatment plants in Canada, no octyltin species was detected. Results of the survey will be published elsewhere.11 For sample preservation, it is recommended that 2 cm3 of concentrated HC1 per dm3 of sewage be added to the sample immediately after collection, to reduce the absorption of the analytes on the container walls and to release the analytes from particulate matter in solution.Sample bottles for sewage and sludge should be placed in coolers with lids loosely capped during transportation especially in summer time. Plastic bottles and other plastic ware should not be used in sampling because of the use of dibutyltin and monobutyltin as poly- (vinyl chloride) stabilizers. Sludge samples should be analysed as soon as possible after collection to minimize biological action resulting in chemical species transformation. It has been observed that sewage samples, after acidification, can be stored for a period of up to 1 month in brown bottles in the dark at 5 "C without changes occurring in butyltin species distribution.As sludge materials are highly bioactive, storage is not advisable unless the sample has been chemically processed. The stability of butyltin species in sludge samples was not investigated. The sample treatment and digestion section of the proposed procedure is lengthy; however, the samples can be left unattended overnight. Twelve samples can be conveniently handled in a working day. In summary, the method is applicable to the determination of butyltin species in sewage and sludge materials. The acid treatment is efficient for the release of butyltin compounds quantitatively from complex ANALYST, JULY 1992, VOL. 117 sewage matrices. The GC-AAS technique is specific and sensitive for this purpose. This study was supported in part by PESTFUND and a Canadian Environmental Protection Act fund (CEPA) of the Department of Environment.1 2 3 4 5 6 7 8 9 10 11 References Thompson, J. A. J., Pierce, R. C., Shaffer, M. G., Chau, Y. K., Cooney, J. J., Cullen, W. R., and Maguire, R. J., Organotin Compounds in the Aquatic Environment: Scientific Criteria for Assessing Their Effects on Environmental Quality. National Research Council of Canada, Associate Committee on Scientific Criteria for Environmental Quality, NRCC Publ. No. 22494, National Research Council of Canada, Ottawa, 1985. Maguire, R. J., Appl. Organomet. Chem., 1987, 1,475. Maguire, R. J., Tkacz, R. J . , Chau, Y. K., Bengert, G. A., and Wong, P. T. S., Chemosphere, 1986,15, 264. Seligman, P. F., Grovhoug, J. G., Valkirs, A. O., Stag, P. M., Franshan, R., Stallard, M. O., Davidson, B., and Lee, R. F., Appl. Organomet. Chem., 1989,3, 31. Muller, M. D., Anal. Chem., 1987,59, 617. Fent, K., and Muller, M. D., Environ. Sci. Technol., 1991, 25, 489. Tian, S., Chau, Y. K . , and Liu, D., Appl. Organomet. Chem., 1989,3, 249. Chau, Y. K., and Wong, P. T. S., in Analysis of Trace Organics in the Aquatic Environment, eds. Afghan, B. K., and Chau, A. S. Y., CRC Press, Boca Raton, FL, 1989, ch. 8, pp. 283-312. Chau, Y. K., Wong, P. T. S., and Bengert, G. A., Anal. Chem., 1982,54,246. Zhang, S . , Chau, Y . K., Li, W. C., and Chau, A. S. Y., Appl. Organomet. Chem., 1991,5,431. Chau, Y. K., Zhang, S., and Maguire, R. J . , Sci. Total Environ., 1992, in the press. Paper 1 I04046 D Received August 5, 1991 Accepted January 20, 1992

ISSN:0003-2654    

DOI:10.1039/AN9921701161

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, JULY 1992, VOL. 117 1165 Determination of Mercury in Fluorescent Lamp Cullet by Atomic Absorption Spectrometry Ryszard Dobrowolski and Jerzy Mierzwa Central Laboratory, M. Curie-Sklodowska University, 20-03 I Lublin, Poland Two methods for the determination of mercury in fluorescent lamp cullet samples were developed. Cold vapour atomic absorption spectrometry (AAS) was applied to samples that were digested to dissolve the attached, mercury-containing phosphor. In the other method, solid phosphor material stripped from the glass cullet was used in a solid sampling technique employing electrothermal AAS with a specially designed ring chamber graphite tube. The results for the determination of mercury by the two methods were comparable. The relative standard deviations were 3.2-3.5% for the cold vapour AAS technique and 8.5-9.9% for direct solid sampling AAS at mercury levels of about 1.5 and 2.5 pg g-1, respectively. The proposed digestion procedure and mercury determination methods (especially the solid sampling AAS method) can be successfully applied to the rapid monitoring of the mercury level in fluorescent lamp cullet and facilitate its further use as recycled glass.Keywords: Mercury determination; glass; atomic absorption spectrometry; cold vapour and electrothermal atomization; solid sampling The high toxicity of mercury and most of its compounds requires its determination at very low concentrations in a variety of samples, e.g., water, air and food. One of the sources of pollution by mercury is the manufacture of fluorescent lamps.Moreover, the working life of such lamps is limited and their disposal is very hazardous. For this reason, fluorescent lamp cullet should be used to monitor the level of mercury. For mercury determination at nanogram levels, atomic absorption spectrometry (AAS) is frequently applied. 1~ Mercury determination by cold vapour AAS (CVAAS) is now a popular technique, but some problems may occur during the sample pre-treatment and digestion steps and in the prepara- tion of standards.3.4 Because of the high volatility of mercury and its com- pounds, the determination of mercury by AAS with elec- trothermal atomization (for both liquid or solid samples) in a graphite furnace (ETAAS) is very difficult. Different stabiliz- ing agents have been used to minimize mercury losses in the thermal pre-treatment step.On the basis of the literature, it can be concluded that there is no agreement on an effective method of stabilization. Edigers stated that addition of ammonium sulfide to the mercury solution provided better thermal stability and allowed higher temperatures of thermal pre-treatment (up to 300 "C) to be used. Kunert et aZ.6 stated that the presence of sodium sulfide or oxidizing agents (KMn04, K2Cr20, and H202) on the graphite surface increased the sensitivity for mercury(i1) because of decreased losses of mercury during the pre-atomization periods. Issaq and Zielinski7 used hydrogen peroxide (2% H202) for stabilization of mercury. Alder and Hickmans reported that addition of H202 had a beneficial effect only when H2O2-HC1 or H202-HCl-HN03 mixed solutions were used.Lender0 and Krivang concluded that stabilization with HCl-H202 was more convenient than with gaseous hydrogen sulfide and more effective and reliable than that with ammonium sulfide solution. Some examples of solid sampling mercury determination have been described. Idzikowski and Michalewska'o deter- mined mercury in copper ores (containing sulfides) by solid sampling ETAAS. Standards for calibration were prepared with 3% Na2S203. This method was rapid and did not require any pre-treatment of the analyte materials. de Kersabiec and Benedettill determined mercury in geological samples by solid sampling using the Hitachi cup-type cuvette and 'dilu- tion' of the samples with graphite powder. The purpose of this work was to develop a simple, rapid and reproducible method for the determination of mercury in cullet from fluorescent lamps using AAS.Two methods were investigated and compared. In one, mercury-containing phosphor was dissolved from the cullet by acid digestion and the digest was analysed for mercury by CVAAS. In the other, phosphor was manually stripped from pieces of glass cullet and analysed by solid sampling ETAAS. Experimental Apparatus The experiments were performed with an AAS-3 atomic absorption spectrometer (Carl Zeiss, Jena, Germany) equipped with a mercury cold vapour generation system, an EA-3 graphite furnace atomizer and a Servoscribe-1s recorder (Smith Industries). A deuterium lamp background corrector was used in all instances. A Juniper hollow cathode lamp (at 4 mA) was used as the light source.Non-pyrolytic graphite coated graphite tubes (ring chamber tubes) specially designed by Schmidt and Falk12.13 for solid sample analysis (for the AAS-3 spectrometer) were used. These tubes (see Fig. 1) allow the introduction of a relatively large amount (up to 10 mg) of solid into a separate chamber around the middle part of the tube. For precise regulation of the graphite tube temperature, the real temperatures were set up before the measurements using an Ni-Cr-Ni thermocouple. Samples were weighed with a Sartorius Model R-200 D analytical balance. Reagents and Standards Standard mercury solutions were prepared from a mercury stock standard solution (Merck) and doubly distilled water. .-.-.-.-.-. t Fig.1 Cross-section of the ring chamber graphite tube1166 ANALYST, JULY 1992, VOL. 117 Table 1 Operating parameters for the determination of mercury in cullet samples by solid sampling ETAAS Step Parameter 1 2 3 4 TemperaturePC 75 90 500 2400 RampPC s-l 20 2 50 500 Hold time/s 10 30 5 5 Argon flow rate/cm3 min-l 4 0 4 0 0 Max. Working standard solutions for solid sampling mercury determination were prepared in a 0.1 mol dm-3 solution of sodium sulfide and for cold vapour determination in 1 mol dm-3 nitric acid. Tin(u) chloride reducing solution (5% d v ) , sodium sulfide solution and dilute nitric acid were prepared from analytical-reagent grade reagents and doubly distilled water. Wet Digestion and CVAAS Determination Crushed cullet from fluorescent mercury lamps (the size of the pieces should not exceed 6-7 mm) was treated with concen- trated nitric acid in a closed digestion Bethe system14 made of quartz.The sample containing 10 zk 0.01 g of cullet was transferred into a round-bottomed flask and then 10 cm3 of concentrated nitric acid were introduced into the flask. The sample was digested at about 85 "C until complete dissolution of the phosphor. After cooling, the solution was separated from the glass cullet by decantation and then transferred quantitatively into a 100 cm3 calibrated flask and diluted to volume with doubly distilled water. A blank was prepared in the same way from 10 cm3 of concentrated nitric acid and doubly distilled water. Just before the measurements the samples and blank were suitably diluted.The samples were then analysed by CVAAS. For a sample volume of 10 cm3, a 1 cm3 volume of the 5% SnClz reductant was used. Air was bubbled through the solution at a rate of 2 dm3 min-1 into the measurement cell and then to exhaust. The measured spectral bandpass was 0.7 nm and the analytical spectral line of mercury was at 253.7 nm. The absorbance was recorded and the maximum peak height was read. The calibration graph method was used for determination. Determination by Solid Sampling ETAAS A thin layer of phosphor was scraped manually from pieces of fluorescent lamp cullet and the powder was transferred into a closed vessel. The sample was thoroughly mixed by rotating the vessel and in this way about 100 mg of phosphor contaminated by mercury were collected.This sample was used for further investigations. The mass of the sample (in the range 0.25-5 mg) was determined by differential weighing (with a precision of k0.02 mg) of one part of the ring chamber tube, then the tube was inserted in the EA-3 atomizer. The lamp current, mercury wavelength and spectral band- pass were the same as those used for CVAAS. Argon was used as the purge gas. The time-temperature programme is shown in Table 1. Integrated absorbance was measured. Mercury was determined by the standard additions method. Aqueous mercury standards in 0.1 mol dm-3 sodium sulfide solution were used as additions. Results and Discussion The proposed method for the preparation of cullet samples by nitric acid digestion followed by CVAAS determination is relatively simple and rapid.It has been found experimentally that the minimum mass of cullet representative (guaranteeing 0 100 200 300 400 500 600 700 TemperaturePC Fig. 2 Thermal pre-treatment graphs for mercury: A , solid sample of halophosphate phosphor; and B, water standard plus Na2S 0.2 , I 9) C -2 0.1 s U 0 3 6 Time/s Fig. 3 Signals obtained from the atomization of mercury in: A , solid sam le (1 mg of phos hor) with solution added; B, solution alone (10 the background attenuation. Integrated absorbance: A, 0.31; B, 0.22; and C, 0.11 mm r , 0 . 3 pg cm-3 H$; and C, solid sample alone. The broken line is Table 2 Comparison of results of mercury determinations in fluores- cent lamp cullet by CVAAS, solid sampling ETAAS and inverse voltammetry Amount of mercury/mg kg-1* Solid sampling Inverse Sample CVAAS ETAAS voltammetry Immediately after break 2.43 2.50 2.76 (RSD 3.5%) (RSD 8.5%) (RSD 12.0%) Cullet taken from dump 1.58 1.49 1.75 * Relative standard deviations (RSDs) were evaluated from five (RSD 3.2%) (RSD 9.9%) (RSD 10.6%) individual measurements.an acceptable reproducibility of the analytical signal) of the material investigated is about 10 g. The efficiency of extraction of mercury with nitric acid was checked by determining mercury after repeated extraction of the same sample of cullet. The recovery of mercury by a one-step extraction was 98%. Initial investigations on direct solid sampling showed that most of the mercury contained in the phosphor isolated from broken lamps was evolved at temperatures above 200 "C (see Fig.2). This suggested that the mercury was not physically adsorbed on the phosphor surface but was chemically bonded on this surface. For this reason the standard additions methodANALYST, JULY 1992, VOL. 117 1167 with the addition of aqueous mercury solutions stabilized by sodium sulfide can be employed with good results (see Fig. 3). The weighed portion suitable for the accurate determination of mercury ranged from 0.8 to 5.0 mg of phosphor. The calculated detection limit of the proposed method (30 criterion) was about 3 ng of mercury. The results of mercury determination by CVAAS and solid sampling ETAAS in two samples of fluorescent lamp cullet are presented in Table 2. These results are compared with those for the determination of mercury in the same samples (after wet digestion) by inverse voltammetry.15 The results obtained are comparable; the small differences in the values may be due to the poorer precision (relative standard deviations 10.6 and 12.0% at mercury levels of about 1.5 and 2.5 pg g-1, respectively) of the inverse voltammetric method.For AAS determinations the relative standard deviations (for five individual measurements) at mercury levels of about 1.5 and 2.5 pg g-1 were 3.2 and 3.5%, respectively, for CVAAS and 8.5 and 9.9%, respectively, for direct solid sampling ETAAS. The direct determination of mercury contained in solids (phosphor isolated from the cullet of fluorescence lamps) by ETAAS confirmed the applicability of this method to the rapid monitoring of mercury levels in glass cullet without the time- and reagent-consuming step of wet digestion.On the basis of these results, the parameters for thermal pre-treat- ment of fluorescent lamp cullet to remove mercury and allow the cullet to be recycled can be investigated and optimized. 5 6 7 8 9 10 11 12 13 14 15 References Chilov, S., Talanta, 1975, 22, 205. Ure, A. M., Anal. Chim. Acta, 1975, 76, 1. Koirtyohann, S. R., and Khalil, M., Anal. Chem., 1976,48,136. Christmann, D. R., and Ingle, J. D., Jr., Anal. Chim. Acta, 1976,86, 53. Ediger, R. E., At. Absorpf. Newsl., 1975, 14, 127. Kunert, I . , Komarek, J., and Sommer, L., Anal. Chim. Acta, 1979, 106, 285. Issaq, H. J., and Zielinski, W. L., Jr., Anal. Chem., 1974, 46, 1436. Alder, J. F., and Hickman, D. A., Anal. Chem., 1977,49,336. Lendero, L., and Krivan, V., Anal, Chem., 1982,54,579. Idzikowski, A., and Michalewska, M., Chern. Anal. (Warsaw), 1980, 25, 35. de Kersabiec, A. M., and Benedetti, M. F., Fresenius’ 2. Anal. Chem., 1987,328,342. Schmidt, K. P., and Falk, H., Spectrochim. Acta, Part B , 1987, 42, 431. Schmidt, K. P., and Falk, H., paper presented at the Xth CANAS, Torun, Poland, September 5-9, 1988. Minczewski, J., Chwastowska, J., and Dybczynski, R., Analiza Sladowa, Metody Rozdzielania i Zagpzczania, WNT, Warsaw, 1973, p. 99. Dobrowolski, R., and Mierzwa, J., paper presented at Euro- analysis VIII, Vienna, August 26-31, 1990. Paper 1 lO6083J Received December 2, 1991 Accepted February 10, 1992

ISSN:0003-2654    

DOI:10.1039/AN9921701165

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, JULY 1992, VOL. 117 1169 Matrix Effect Corrections for the Quantitative X-ray Fluorescence Determination of Iron Using Scattered Radiation Maria Teresa Garcia-Gonzalez Centro de Ciencias Medioambientales (CSIC), C/Serrano I 15,28006 Madrid, Spain Maria Dolores Haro-Ruiz lnstituto Tecnologico Geominero de Esparia, C/Rios Rosas 23, 28003 Madrid, Spain Alfonso Hernandez-Laguna lnstituto de Estructura de la Materia (CSIC), C/Serrano 123, 28006 Madrid, Spain Scattered radiation was used as the internal standard in order to correct matrix effectsfor the determination of Fe by X-ray fluorescence. The K a , the Compton K a of the Rh tube and the 14" 28 continuum background were tested. The first- and last-named scattered radiation were found to correct the matrix effects with better standard deviations and correlation coefficients than the fit without correction.The improved functions were fitted to non-linear polynomial equations. The results suggest that the characteristic scattered radiation of the X-ray tube must be taken into account when determining certain analytes. Keywords : X-ray flu0 rescen ce; 4 ua n tita tive &term in a tion; iron; matrix corrections; scattered radia ti0 n Scattered radiation had long been considered to be a perturbation in X-ray fluorescence analysis, because in many instances it may hide interesting lines of the analyte elements. However, since the work by Andermann and Kemp,l scat- tered radiation has been used as an internal standard for correcting matrix effects, instrumental errors, packing den- sity, texture and heterogeneity effects.The use of scattered radiation as an internal standard has the following advantages: (a) no foreign element is introduced into the samples, in contrast to the normal internal standard method; (b) no additional manipulations of the sample are necessary; and (c) there is no dilution of the analyte. This method is usually employed taking into account the marked similarities between samples and standards. Usually, a low average atomic number for the matrix elements and a narrow range of analyte concentration are suitable.2-3 In order to obtain good matrix effect corrections, it is preferable that the element present at the highest concentration should have no absorption edge between the fluorescence and the scattered wave1engths.M In addition, it is advantageous to apply this methodology to the elements of higher atomic number present in the sample.Usually, only one scattered radiation is employed for correcting matrix effects, and it is chosen empirically. However, in some instances, a special combination of two scattered radiations is employed, as, for example, with the coherent and incoherent scattered radiation relationships.7 Clark and Mitchell8 determined Fe in metallurgical com- pounds, and in Si and Cu matrices. No reliable corrections were obtained using continuum scattered radiation from a W tube. Martin-Rubi et aZ.9 studied the correction of matrix effects in the determination of Fe, by means of different scattered radiations in Fe compounds, using a Cr X-ray tube.They obtained a good correction with the 14" 28 LiF (200) continuum background. This correction was fitted to non- linear polynomial functions. The function had a negative deviation from the linear function (curvature below an ideal straight line which should fit the relationship between radiation and concentration). In contrast, the correction obtained by means of the Ka scattered radiation was not satisfactory. These results were fitted to non-linear poly- nomials, but the curves deviated in the opposite direction from the straight line, compared with the other correction method used. Therefore, a geometric average between these two curves yielded a linear function with good standard deviation and correlation coefficient. Based on the results of the work described in refs.8 and 9, it could be argued that the scattered radiation of the target element of the X-ray tube could be an important feature of a method for correcting matrix effects. This paper describes a method to correct matrix effects by means of X-ray scattered radiation using a Rh tube with the aim of determining the influence of the nature of the target element of the X-ray tube. Experimental A Siemens SRS 300 sequential X-ray fluorescence spec- trometer on-line with a PDP 11/23 digital computer was used. An end-window Rh X-ray tube and LiF (200) crystal was also used. British Chemical Standard (BCS) Certified Reference Materials (CRMs) containing between 2 and 66% Fe were used; the concentrations of the major elements are shown in Table 1. The methodology described by Romand2 was followed in order to obtain the optimum measurement conditions for X-ray fluorescence and scattered radiation lines.Optimum Table 1 Major elements in BCS-CRMs. All values in % Sample BCS-CRM 17512 BCS-CRM 378 BCS-CRM 377 BCS-CRM 303 BCS-CRM 301/1 BCS-CRM 382 BCS-CRM 315 Fe 66.1 61.8 52.5 36.0 23.8 11.6 2.11 Si02 2.58 5.2 8.6 16.50 7.4 19.5 51.2 A1203 1.08 0.80 3.40 6.81 4.25 7.51 42.4 CaO 0.08 6.58 10.8 19.6 22.5 33.0 0.43 MgO 0.03 0.2 1.08 1.84 1.72 0.57 10.7 Table 2 Instrumental X-ray fluorescence measurement parameters Measurement angles [LiF (200) 28"] Power Rh target Fe KP 51.76 60 kV; 20 mA LiF (200) Background 50.25 Coarse collimator Rh Ka 17.55 60 kV; 30 mA Vacuum Rh background 1 17.10 Flow detector Rh background 2 17.85 Compton 18.48 Compton background 1 17.85 Compton background 2 21.00 60 kV; 30 mA1170 ANALYST, JULY 1992, VOL.117 excitation conditions were those that gave the maximum analyte-to-background intensity ratio. The Fe KP line was used for measurements because the Ka line intensities were too large for samples with high concentrations of Fe. The Ka and Compton Ka scattered lines for the Rh tube were used to correct matrix effects. The 14" 28 continuum background radiation was studied with the same aim. The measurement conditions are presented in Table 2. The intensity measure- ments were repeated four times. All the concentration-intensity equations were fitted to a polynomial by using the least-squares method included in the BMDP packagelo such as: 3 i = O c = 2 U i X I ' (1) The standard deviations of the data, a, and coefficients of the fit, a,,, and the correlation coefficients, r, were calculated.The coefficient of determination (CD)11 and the model selection criterion (MSC)1* were also evaluated in order to include other statistical parameters to increase the significance of the different polynomial fits. The BCS-CRMs were ground until 200 mesh. Then, a solution of elvacite in acetone was mixed with the powdered sample, and the mixture was pressed at 40 tonnes cm-2 to form a disc 4 cm in diameter and 5 mm thick.13 Results and Discussion Firstly, the Fe KP line was measured as a function of the concentration (Fig. 1). A linear equation was obtained with a = 2.5, and r = 0.9958 (Table 3). Higher-order polynomials were used, but with no significant improvement.The standard 70 60 50 - 40 I - Q, LL 30 20 10 0 20 40 60 80 100 120 140 IF, ~~4103 counts s-1 Fig. 1 Fe concentration as a function of the Fe KP X-ray fluores- cence intensities. The value of of ranges from 10 to 40 deviation of the intercept of the linear fit is higher than the value of the coefficient. A second fit without any intercept is shown in Table 3. Most of the statistical parameters of the goodness of the fit are better than before. However, a,, is statistically less meaningful than before. In principle, it is possible to carry out a reliable measurement of Fe, within the margin of error characterized by a. However, this method is free from matrix effect corrections, particle size effects, instrumental error and packing density.Accordingly, it is possible to correct for scattered radiation and obtain improved results. The scattered Rh Ka radiation was measured under the optimum excitation conditions. Similar conditions applied for the Fe KP radiation. The relationship IFe Kfi IRh KCK Is = - as a function of the Fe concentration is shown in Fig. 2. A homogeneous distribution of points along the curve and a negative deviation from a straight line with respect to the x-axis (intensity) was obtained. The fitted equation is third- order without a squared coefficient. The standard deviation of the intercept is approximately three times the absolute value of the coefficient and a new fit without an intercept was used. In this instance, the two coefficients are more significant than before.The matrix effect correction obtained by introducing the scattered Rh Ka radiation is mainly described by the linear term, and the cubic term is only a correction and consequently is not so significant, although this term yields the curvature at the extreme end of the curve. In both fits an improved standard deviation and correlation coefficient were obtained (a = 1.5 and r = 0.9986 for the fit without an intercept). The CD is slightly lower than in the previous fit, but the MSC is higher than all the values shown. The standard deviation is lower by a value of 0.9 than that obtained from free matrix effect corrections. A similar correction using the Cr tube gives positive deviations from a straight 1ine;g however, the values of u and r obtained here are better than those found previously.9 The Cr Ka scattered line is near the Fe KP line, and lies on the long wavelength side of the Fe absorption edge. Unfortunately, this scattered radiation is not convenient, as it does not lie on the short wavelength side of the absorption edge of the element present at the highest concentration, at least for samples with high concentrations where the Fe KP line is used.This latter condition is preferable in order to obtain a linear fit to eqn. (l).14 The Rh Ka scattered radiation is on the short wavelength side of the absorption edge of the analyte, but is far away from the Fe KP line. Accordingly, the scattered radiation located on the long wavelength side of the analyte absorption edge would yield matrix corrections with a Table 3 Coefficients of the polynomial [eqn.(l)] with and without an intercept, standard deviations of the coefficient (aai) and the data of the fits (a), coefficient of determination (CD), model selection criterion (MSC), averaged non-signed error (151) and standard deviations of the errors (9) ZFe Is 1, Parameter -1.1 1.8 5.458 x 10-4 2.2 x 10-6 - - - - 2.50 0.9958 0.9916 4.21 2.22 2.41 - - 5.34 x 10-4 1.1 x 10-5 - - - 2.37 0.9958 0.9910 4.42 2.28 2.40 -0.5 1.4 0.314 0.018 - - -9.6 x 10-7 2.2 x 10-7 1.61 0.9986 0.9972 5.03 1.75 2.15 - - 0.3078 0.0093 - - -9.1 x 10-7 1.5 x 10-7 1.47 0.9986 0.9971 5.28 1.73 2.10 -0.60 0.95 0.01964 -1.43 x 10-6 7.8 x 10-4 1.2 x 10-7 - 1.14 0.9993 0,9986 5.72 1.21 1.33 - - 0.01924 4.4 x 10-4 ,1.382 X 10-6 8.3 x 10-8 - - 1.07 0.9993 0.9985 5.91 1.09 1.24ANALYST, JULY 1992, VOL.117 1171 70 r I 60 50 30 20 10 0 50 100 150 200 250 300 IFe KI3 : IRh Kn Fig. 2 Fe concentration as a function of Z F ~ K ~ Z R , , Ka (&). The value of of, ranges from 1 to 2 70 60 50 40 Lc 30 - - 0, 20 10 0 1000 2000 3000 4000 5000 6000 7000 /Fa KP : /con1 14 2H Fig. 3 Fe concentration as a function of Z F ~ ~dZ~4 2e (Zc). The value of qC ranges from 1 to 158 positive deviation from a straight line, whereas the scattered radiation located on the short side of the absorption edge of the analyte would yield a negative deviation from a straight line. The same corrections were also tested on the Compton scattered radiation, but the standard deviation was worse than that of the correction free test. The correlation coefficient was of the same order.This is in agreement with results reported previously.9 Further, this scattered radiation appears to be more suitable when applied to matrix corrections to elements present at low concentration.2 The 14" 28 continuum background radiation was used to form the quotient: IF€. I, = - I14 (3) Results are shown in Fig. 3. A negative deviation from a straight line with respect to the intensity axis is observed. The best fit to this curve was a second-order polynomial (Table 3). The fit shows the same features as before and a polynomial without an intercept is obtained. This fit is more significant than the non-linear equation given earlier. This correction is mainly described by the linear term, whereas the squared term is a modification which accounts for the curvature.Again the lack of an intercept leads to a matrix effect correction that does not introduce additional noise. Further, the fits of I, and I, with intercept coefficients show lower values than the fit without correction. Improved values for the parameters r (0.9993) and <J (1.1) are obtained. The CD for the fit without an intercept is slightly lower than that with an intercept. The MSC of the last fit is higher than those of the other fits shown in Table 3. All the results are better than those obtained with the previous corrections. However, the standard deviation of the intensity relationship, arc, has higher values than those given by the previous corrections. On the other hand, the I, correction provides a more linear fit than I, because the coefficient of the non-linear term and the quotient between this and the linear coefficient of I, are much lower.These results are in agreement with those reported previ- ously,9 which exhibit the same type of deviation from a straight line, and the statistical coefficients have similar values. In this instance, the continuum background radiation appears to be more reliable when corrections are performed with the scattered radiation (at least for this type of sample). The results would appear to be independent of the nature of the target. The coherent scattered radiations coming from the main lines of the element of the target are more specific, as a consequence of their different wavelengths. As can be seen in Figs. 2 and 3, the two corrections have the same negative deviation with respect to the intensity axis and, therefore, it is not possible to obtain corrections by mixing I, and I,, which could be combined to give a linear equation as in ref.9. Clark and Mitchell8 obtained no reliable corrections for the determination of Fe in metallurgical compounds using conti- nuum scattered radiation. Copper, however, has the shortest characteristic wavelengths in the samples used by these workers, and the Fe lines are between the continuum scattered radiation and the absorption edge of Cu. In order to test the reliability of the methods described above, one of the data points was removed from the fits. The removed standard was re-introduced as a sample, which allowed the precision of the method to be tested.All the standards were removed one at a time, with the exception of those at lower and higher concentrations in order to ensure that the extreme features -of the fits were retained. The average non-signed error, [El, and the standard deviation of the errors, aE, are shown in Table 3. Application of the 14 28 correction gives lower values for bcth quantities, and an improved difference of 1.17 for both IEI and a~ is obtained. Conclusions For the types of sample described above, it is possible to obtain meaningful and reliable matrix effect corrections using scattered radiation arising from a Rh X-ray tube, particularly if the continuum radiation arising from the maximum of the scattered spectrum is used. The results appear to be indepen- dent of the nature of the target element.These results, and those obtained previously,9 show that 14" 28 LiF is a reliable continuum scattered radiation. The scattered radiation arising from the main lines of the tube also provides a correction, but is worse than that arising from the continuum radiation. Both types of correction yield reliable non-linear equations where the curves pass through the origin. The position of the absorption edge of the element present at the highest concentration appears to be important for the correction of I,, and a suitable mix of corrected intensities would yield linear equations. The authors thank Drs. G. A. Arteca, J. L. M. Abboud and D. C. Moule for discussions. References 1 Andermann, G., and Kemp, J. W., Anal. Chem., 1958, 30, 1306. 2 Romand, M., 8eme Colloque sur lilnalyse de la Matiere, Florence, September, 1969.1172 ANALYST, JULY 1992, VOL. 117 3 4 5 6 7 8 9 10 Vie la Sage, R., Quisefit, J. P., Dejean de la Batriz, R., and Fancherre, J., X-ray Spectrom., 1979,8,121. Tertian, R., and Claisse, F., Principles of Quantitative X-Ray Fluorescence Analysis, Heyden, h n d o n , 1982. Berth, E. P., Principles and Practice of X-Ray Spectrometric Analysis, Plenum Press, New York, 1975. Muller, R. O., Spectrochemical Analysis by X-Ray Fluores- cence, Adam Hilger, London, 1972. Rayland, A. L., Natl. Meet. Am. Chem. SOC., Philadelphia, 1964, 147. Clark, N. H., and Mitchell, R. J., X-Ray Spectrom., 1973,2,47. Martin-Rubi, J. A., Haro-Ruiz, M. D., and HernBndez-Lag- una, A., An. Quim., 1987, 83B, 305. BMDP Statistical Software Manual, California University Press, Berkeley, 1987 release, 1R program. 11 Pefia, D., Estadistica. Modelos y Mktodos, Alianza Universi- dad, Madrid, 1991. 12 MINSQ program, MSC calculated in Taller de Precisidn y Centro ElectrotCcnico de Artilleria of Ministerio de Defensa, Madrid, Spain. 13 Barba, F., Valle, F., and Martin-Rubi, J. A., Silic. Znd., 1983, I, 13. 14 Reynolds, R. C., Am. Mineral., 1967,52, 1493. Paper 1104042A Received August 5, 1991 Accepted January 7, 1992

ISSN:0003-2654    

DOI:10.1039/AN9921701169

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, JULY 1992, VOL. 117 1173 Oxalate-catalysed Oxidimetric Assay of Thiourea With Permanganate Alice Kurian Central Electrochemical Research Institute, Karaikudi-623 006, India A method for the determination of thiourea in acidic solution by titration with permanganate in the presence of a known amount of oxalic acid is described. The reaction is a four-equivalent oxidation of thiourea by permanganate. The oxalic acid in the solution effects a change in the mechanism such that thiourea becomes a four-electron reductant in its reaction with permanganate. Keywords: Thiourea; four-electron reductant; oxidimetry; permanganate An oxidimetric method for the assay of thiourea with permanganate was reported in an earlier paper.' It is a simple, rapid and direct titrimetric method based on the oxidation of thiourea in acidic solution wherein the end-point is reached when 1 mol of thiourea is allowed to react with 2.5 equiv of permanganate.It is an excellent method for routine analysis. On further investigation it has been found that the stoi- chiometry of the reaction of thiourea with permanganate is increased when oxalic acid is included in the sample solution, and that the thiourea becomes a four-electron reductant when thiourea and oxalic acid are allowed to react simultaneously with permanganate in acidic solution. The minimum amount of oxalic acid necessary to achieve this increase in reaction stoichiometry corresponds to a molar ratio of 1:8, oxalic acid: thiourea in the titrating solution. Hence, in the present study, a simple and rapid method for the determination of thiourea in pure solutions has been developed on the basis of the four-electron reaction of thiourea with permanganate in acidic solution in the presence of a known amount of oxalic acid.The proposed method has considerable advantages over other titrimetric method$-5 owing to its simplicity and the easy accessibility of the reagents required. Experimental Reagents Thiourea solution, 0.025, 0.05 and 0.10 mol dm-3. Potassium permanganate solution, 0.01 and 0.02 mol dm-3. Oxalic acid solution, 0.025 and 0.05 mol dm-3. Sulfuric acid solution, 12 mol dm-3. All solutions were prepared with analytical-reagent grade reagents in doubly distilled water. Method Aliquots (10 and 20 cm3) of thiourea solution were mixed with aliquots (5, 10 and 20 cm3) of the standard solution of oxalic acid in a conical flask and acidified with the required amount of sulfuric acid solution to give an acidity greater than 1.50 mol dm-3.These solutions were titrated with the standard permanganate solutions at room temperature. A convenient number of titrations were carried out in each of eight different combinations chosen for this purpose with molar ratios of oxalic acid to thiourea ranging from 1 : 8 to 4 : 1 and typical 0.9 2 0.7 (D C al L .- 4- w 0.5 0.3 24.8 25.0 25.2 Volume of KMnOd solution/cm3 Fig. 1 Potentiometric titration of a thiourea-oxalic acid solution [composition: 0.025 mol dm-3 thiourea solution (20 cm3) and 0.05 mol dm-3 oxalic acid solution (10 cm3)] with KMn04 at an H2S04 concentration of: 0, 2; and A , 3 mol dm-3 Table 1 Results obtained €or the titration of thiourea-oxalic acid mixtures with standard permanganate solutions.As there is only one inflection point in the potentiometric titration curve, the thiourea content is calculated by subtracting the oxalic acid contribution from the titration value Concentration Volume of of thiourea Sample thiourea solution/ 1 10.0 5.0 2 10.0 10.0 3 20.0 2.5 4 20.0 5.0 No. solution/cm3 mol dm-3 * Titration value. Concentration Concentration Volume of of oxalic of KMn04 Volume of oxalic acid acid solution/ solution/ KMn04 Standard solution/cm3 10-2 mol dm3 10-2 mol dm-3 solution*/cm3 Error (YO) deviation (YO) 5.0 2.5 1.0 45.15 0.40 0.55 5.0 2.5 2.0 42.45 0.10 0.30 20.0 5.0 2.0 40.05 0.25 0.28 - - 2.0 25.05 0.20 0.251174 ANALYST, JULY 1992, VOL.117 results for those with low and high oxalic acid contents are presented in Table 1 (Entries 1-3) together with that obtained for the blank titration of thiourea solution (Entry 4) for comparison. In order to illustrate the sharpness and reliability of the end-point, potentiometric titrations were carried out at different concentrations of acid from 1.50 to 3.50 mol dm-3 (Fig. 1). Results and Discussion The proposed method is rapid and reliable for routine analysis. The reaction is spontaneous and very fast at room temperature. The end-point is sharp and precise with higher concentrations of sulfuric acid (2.5-3.5 mol dm-3). The end-point corresponds to the oxidation of 1 mol of oxalic acid by 2 equiv, and 1 mol of thiourea by 4 equiv of permanganate. The purpose of this work was to investigate the mechanism for the reaction of thiourea with 2.5 equiv of the permanga- nate oxidizing agent as stated earlier. The experiments involved in this investigation led to the discovery that the oxalic acid could effect one further step in the oxidation and that the reaction was exactly quantitative for use as an analytical method.The primary objective of this paper is to highlight the analytical implications of this result. However, it is not possible to interpret the exact mechanism at present and no explanation is offered in the existing literature .+9 References Kurian, A., and Suryanarayana, C. V., Analyst, 1972, 97, 576. Gupta, K. S., Swamp, R., and Sharma, D. N., Labdev, Part A, 1974, 12, 84. Srivastava, A., and Bose, S., J. Indian Chem. SOC., 1985, 62, 546. Jayasree, N., and Indrasenan, P., Talanta, 1985, 32, 1067. Srinivasan, K., and Indrasenan, P., J. Indian Chem. SOC., 1987, 64,451. Werner, E. A., J. Chem. SOC., 1919, 115, 1168. Grover, K. C., and Mehrotra, R. C., Freseniw’ 2. Anal. Chem., 1958, 160, 267. Sarwar, M., and Thibert, R. J., Anal. Lett., 1968, 1, 381. Dal Nogare, S., in Organic Analysis, eds., Mitchell, J., Jr., Kolthoff, I. M., Poskauer, E. S., and Weissberger, A., Interscience, London, 1953, vol. 1, p. 392. Paper 110.56480 Received November 6, 1991 Accepted January 28, 1992

ISSN:0003-2654    

DOI:10.1039/AN9921701173

出版商:RSC    

年代:1992

数据来源: RSC