918 Analyst August 1983 Vol. 108 $9. 918-924 Determination of the Chemical Forms of Cadmium and Silver in Sediments by Zeeman Effect Flame Atomic-absorption Spectrometry Ken R. Lum Environmental Contaminants Division National Water Research Institute CCI W P.O. Box 5050 Burlington, Ontario L7R 4A6 Canada and Duart G. Edgar Nissei Sangyo Canada Inc. 89 Galaxy Blvd. Suite 14 Rexdale Ontario M9W 6A4 Canada A polarised Zeeman flame atomic-absorption spectrometer has been used for the determination of cadmium and silver in chemical extracts of sediments using a procedure designed to provide information on the potential availability of trace elements. The limit of determination for complex solution matrices was found to be 1.1 pg 1-1 for both cadmium and silver. Minimum base-line noise for the instrument afforded very good stability of the calibration con-ditions at the 10 pg 1-1 level for the two elements.Accuracy tests performed with standard reference materials showed good agreement for the determi-nations in the extracts. The analysis of suspended and bottom sediments is used to demonstrate the value of this instrumental technique. Keywords ; Zeeman flame atomic-absorption spectrometry ; chemical forms ; cadmizcm determination ; silver determination ; sediments The toxicity of cadmium and its dissemination into the environment are of considerable current intere~t.l-~ In freshwaters silver is regarded as one of the most toxic metals to aquatic Because metal ion toxicity in aquatic systems can be influenced by sorption and binding to suspended and bottom sediments much work has been done on determining the chemical form and potential availability of elements associated with these substrates.2J-9 The determinations rely on a variety of extraction scheme^^^^-^^ that require the use of high ionic strength solutions such as 1 M magnesium chloride 1 M sodium acetate 0.2 M barium chloride or 0.22 M sodium citrate - 0.11 M sodium hydrogen carbonate - 1 .O g sodium dithionite.The large number of extracts that result from these procedures precludes the use of electro-thermal atomisation atomic-absorption spectrometry (ETA-AAS) techniques because of the additional time that is required to separate the elements of interest from the complex matrix, e.g. by chelation and solvent extraction. The advantages of greater sensitivity in ETA-AAS can in addition be minimised by contamination during the separation step and the longer analysis time needed for these analyses.The direct aspiration flame AAS approach thus offers a simple and rapid analytical tool which further calls for less skill on the part of the operat or. In the procedure12 adopted by us for studies of trace metal availability the flame detection limit for cadmium was reported to be 0.012 pg ml-1 which represents a detection limit of 0.3 pg g1 for a 1.0-g sample subjected to this sequential extraction scheme and a final volume of 25 ml. These detection limits are generally inadequate because background total cadmium concentrations for soils and sediments are in the range 0.4-0.6 pg g-1.13J4 Apart from grossly polluted sediments total cadmium concentrations are usually less than 10 pg g1.14 More-over for chemical extraction procedures the concentration of cadmium in the fractions that are deemed readily or potentially available does not normally exceed 50% of the t o t ~ d .~ ~ ~ J ~ Although there are few data on the chemical forms of silver in soils and sediments total con-centrations of silver in estuarine sediments are ca. 1 pg g1.9 Recent work on the application of the Zeeman effect to atomic-absorption analysis has shown that significant reductions in base-line noise can be attained.l7J8 The superior capability of such instruments for the correction of non-atomic absorption compared with systems relying on deuterium lamp correction is now well documented.l+l LUM AND EDGAR 919 In this paper we report the results of an evaluation of a polarised Zeeman atomic-absorption spectrometer for the determination of cadmium and silver in sediment extracts.Experimental Instrumentation Atomic-absorption measurements were made on a Hitachi Model 180-80 polarised Zeeman instrument using a water-cooled pre-mix burner with air - acetylene. Hamamatsu hollow-cathode lamps were used. The configuration and operation of the Zeeman flame system has been described previou~ly.~~ Instrumental parameters were those recommended by the manufacturer and are given in Table I. Analytical results were obtained using three 5-s integrations of each sample. The data processing unit stores the absorbances of a blank and three standards and gives a mean concentration standard deviation and coefficient of variation (in yo) for each analysis.TABLE I INSTRUMENTAL PARAMETERS FOR THE DETERMINATION OF CADMIUM AND SILVER Element Wavelength/nm Slit width/nm Burner height/cm Flame composition Cadmium 228.8 1.3 7.5 Air 9.4 1 min-I Silver . . 328.1 2.6 7.5 Air 9.4 1 min-1 C,H 2.3 1 min-l C,H 2.3 1 min-l Reagents AnalaR-grade or equivalent quality reagents were used except for the following lithium chloride and caesium chloride Aldrich gold label (99.999%) ; sodium acetate E. Merck Suprapur ; nitric and hydrochloric acids prepared by sub-boiling distillation; and acetic acid, Baker Instra-analysed grade. Standards were made up in 0.16 M nitric acid by serial dilution of 1000 mg 1-1 stock solutions (Fisher Scientific Co.) .Sequential Chemical Extraction Procedure A five-part sequential scheme12 was followed except for the lake sediments for which a reagent composed of 0.75 M lithium chloride and 0.25 M caesium chloride in 60% methanol was used instead of 1 M magnesium chloride. The volumes of extractants used were identical with those recommended by Tessier et ~ 1 . 1 ~ and the procedure was applied to ca. 1-g (dry mass) samples of sediment (Fig. 1). The sequential extractions simulate to a certain extent various environmental conditions to which sediments and similar materials may be subjected Although these extraction schemes are not perfectly selective they can provide valuable information on the mobility and avail-ability of elements in soils sediments and Dissolution of Standard Reference Materials A modified room-temperature dissolution procedure was used for solubilising several stand-ard reference materials to assess the accuracy of the instrumental determination and the extrac-tion process.23 The samples were weighed into 125-ml acid-cleaned polyethylene bottles, wetted with 2 ml of doubly distilled water then 3 ml of aqua regia and 20 ml of hydrofluoric acid were added.The bottles were agitated on a wrist-arm shaker for 16 h then 75 ml of saturated boric acid solution were added and the bottles were replaced on the shaker for a further 4 h. The supernatants were analysed after allowing the small amounts of residue to settle. Five replicate 1-g samples of NBS SRM 1577 (bovine liver) were digested in PTFE beakers using 10 ml of 1 + 1 sulphuric acid and heating to white fumes.After allowing the beakers to cool sufficient 30% hydrogen peroxide was added in 0.2-ml aliquots to obtain a yellow extract. The solution was heated for a further 30 min cooled and made up to 25ml with doubly distilled water 920 LUM AND EDGAR DETERMINATION OF CHEMICAL FORMS Analyst VoZ. 108 Triplicate 0.2-g samples of the National Institute for Environmental Studies (NIES Japan) Certified Reference pond sediment were muffle ashed at 500 "C for 3 h. A separate 0.2-g sub-sample was dried in a forced air oven at 110 "C for 24 h to determine moisture content. The ashed samples were carefully wetted with doubly distilled water transferred into PTFE beakers and 20 ml of freshly prepared aqua regia were added.The digests were reduced nearly to dryness 15 ml of hydrofluoric acid were added and heating was continued until the samples were dry care being taken to avoid baking. Hydrochloric acid (15 ml) and 15 ml of doubly distilled water were then added and the solutions were heated for 1 h to reduce the volume to ca. 15 ml. After cooling the volumes were made up to 25 ml in a calibrated flask. Results and Discussion Cadmium Calibration graphs were obtained for cadmium concentrations of 0.0 10.0 40.0 and 100.0 pg 1-l. The absorbances were stored in the instrument's data processing unit and for three separate days (and for standards of 0.0,20.0,40.0 and 60.0 pg 1-l) the computed correla-tion coefficient for the best linear fit line was 0.99. Using the former calibration range stand-ards run as samples gave the concentrations shown in Table 11.The data processing unit is capable of computing (within 4%) concentrations that were ten times that of the highest stored standard. Calibration stability and analytical reproducibility were assessed by aspirating the 10.0 pg 1-1 Sample a 0.75 M LiCl - 0.25 M CsCl - 60% CHBOH 10 min Room tem peratu re Residue Q 1 M CH,COONa,pH 5.0 Room temperature Extract I Residue 1 M NH2OH.HCI -25% CH3COOH 2h Room temperature Extract rn Residue H202 PH 2 90 "C 5 h Extract 1.2 M CH3COONHd - 20% HNO3 Residue 9 I Aqua regia - HF - HCI - H202 I Extract Readily exchangeable ions (A) Carbonate-bound su rface-oxide bound ions (B) Ions bound to Fe -Mn oxides (C) Organically and sulphide-bound ions (0) Ions bound to the residual phase Fig.1. Outline of the sequential chemical extraction procedure August 1983 OF CD AND AG IN SEDIMENTS BY ZEEMAN EFFECT AAS TABLE If LINEARITY OF CADMIUM CALIBRATION CONDITIONS 921 Concentration of standard/pg 1-1 10.0 25.0 40.0 60.0 100 250 500 1000 Concentration determined by instrument*/pg 1-1 11.6 f 1.1 24.7 f 0.8 42.6 f 1.7 61.3 f 0.9 101 f 0.4 246 f 0.8 480 f 1.6 961 f 0.6 * The data processing unit prints out the mean concentration of three 5-s integra-tions and the standard deviation. standard periodically during the analysis of sediment extracts. The standard was determined ten times over a 2-h period and the concentration measured varied from 10.0 to 11.6 pg 1-1 with a mean value of 10.8 5 0.6 pg 1-l.These data were obtained without any update of the calibration graph. In spectrochemical analysis the concentration of an element that will absorb 1% of the incident resonance energy of that element is defined as its reciprocal sen~itivity.~~ The absorbance of the 10.0 pg 1-1 cadmium standard was 0.001 5 which is below that required by this definition and which therefore is not applicable to this analysis in a calculation of the detection limit. A limit of determination was obtained from the analysis of 35 sediment extracts (seven samples of five fractions each) for which the majority (77%) gave concentra-tions less than 10.0 pg 1-1 (the range was 0.5-17.7 pg 1-l). The average of the standard deviations for the 35 analyses was 1.1 pg 1-1 and is used here as the instrumental limit of determination for cadmium in the complex matrices analysed in this study.A particularly valuable feature of this instrument was the low base-line noise observed even after aspirating samples containing 0.32 M magnesium chloride and 0.32 M sodium acetate. At no time during the three days of operation was it ever necessary to dismantle and clean the burner chamber and nebuliser. Occasional clogging of the nebuliser by particles in some of the samples was promptly rectified by the insertion of a cleaning wire. The accuracy of the determination was evaluated by analysing several standard reference materials (Table 111). The solution concentrations calculated from the certified values (pg g-') are compared with the measured concentrations.The recovery of cadmium from the coal fly ash (NBS SRM 1633) is in excellent agreement with the certified value and that reported by G l a d n e ~ . ~ ~ Similarly the determination of cadmium in digests of bovine liver (NBS SRM 1577) showed good agreement with the certified value. The lower than expected recovery for the river sediment (NBS SRM 1645) and the urban particulate matter (SRM 1648) is in part the result of their organic contents. For a similar room-temperature dissolution procedure the mass of the undissolved residue was found to be correlated with the carbon content of the ~ample.~3 The river sediment contains 1.71 yo of Freon-extractable oil and grease whereas the urban particulate matter contains 1.19%.An estimate of the organic content can be obtained from the loss on ignition (800 "C) value which is 10.72y0 for the river sediment; no value is provided for the urban particulate matter. In contrast the coal fly ash would be expected to contain very little organic matter. TABLE I11 ACCURACY OF THE DETERMINATION OF CADMIUM BY ZEEMAN EFFECT FLAME AAS Reference material NBS SRM 1633 coal fly ash 1 . . 2 3 . . NBS SRM 1645 river sediment . . NBS SRM 1648 urban particulate matter NBS SRM 1577 bovine liver (n = 5) . . Expected Measured Recovery, concentrationlpg 1-1 concentration/pg 1-1 yo 5.4 5.8 f 0.8 107 7.0 7.4 & 0.6 106 6.2 6.3 f 0.8 102 185 148 f 0.2 80 329 291 f 0.7 88 10.8 10.6 f 0.8 9 922 LUM AND EDGAR DETERMINATION OF CHEMICAL FORMS Analyst Vol.108 The possible effect of the high dissolved solids content (e.g. 0.7 M for boric acid) on nebulisa-tion and atomisation efficiency was checked by serial dilution and analysis. The concentra-tions obtained for 0 2 4 10 and 15x dilutions were 152 151 149 154 and 159 pg 1-1, respectively which indicate that such effects do not markedly suppress the cadmium atomic absorption. The accuracy of the determination was further assessed by analysing the NIES pond sediment26 and as shown in Table IV reasonable agreement was obtained for concentra-tions ca. five times the detection limit. It should be noted that the expected concentration has been corrected for a moisture content of 9.0%. TABLE IV ACCURACY OF THE DETERMINATION OF CADMIUM AND SILVER IN THE NIES CERTIFIED REFERENCE SEDIMENT Cadmiumlpg 1-1 Silverlpg 1-1 - Expected Measured Expected Measured Pond 1 .. 6.1 5.8 6.4 5.6 Pond2 . . 6.1 5.7 6.4 5.7 Pond 3 . . 6.1 5.7 6.4 5.6 Chemical forms of cadmium in suspended sediments Seven samples each of suspended sediments were collected by continuous-flow centrifugation from Hamilton Harbour and western Lake Ontario during the summer of 1981. The material was freeze-dried and 0. l-g sub-samples were subjected to the sequential extraction procedure. Total cadmium contents were obtained from the sum of the concentrations found in each extract. The distribution of the chemical forms in each fraction were then expressed as a percentage of the total. The results (Table V) show that the distribution of the chemical forms of cadmium is dominated by the iron and manganese oxide phase in Hamilton Harbour, which is to be expected as the harbour front is the location of the largest steelworks in Canada.For both areas the first two fractions are very important accounting for 28% and 48% for the harbour and lake respectively. The fact that for these fractions the solution concentrations (20 out of 28 extracts) were less than 10 pg 1-1 demonstrates the value of the Zeeman effect AAS determination at these low levels. TABLE V DISTRIBUTION OF THE CHEMICAL FORMS OF CADMIUM IN SUSPENDED SEDIMENTS FROM HAMILTON HARBOUR AND WESTERN LAKE ONTARIO Hamilton Harbour Lake Ontario Exchangeable forms % . . 10 f 3 17 f 6 Carbonate and surface-oxide bound yo .. 18 f 12 31 f 10 Bound to Fe and Mn oxides yo . . f . 53 f 8 34 f 7 Bound to organic matter % . . 7 f 3 12 f 5 Residual forms % . . ,. 12 f 9 6 f 3 Total cadmium concentration rangelpg g-l . . 5.0-8.0 3.5-8.0 Silver Calibration graphs were obtained for silver concentrations of 0.0 20.0 40.0 and 60.0 pg 1-l. Correlation coefficients ranged from 0.95 to 0.99. Although the absorbance of the 20 pg 1-1 standard is not displayable (<0.00099) the microprocessor has been designed to record and calculate concentrations based on detecting absorbances to six significant figures. Calibration stability was ascertained by aspirating a 10.0 pg 1-1 standard periodically during the analysis of sediment extracts. The average value for ten measurements for this standard was 9.3 & 0.8pg.l-l.The limit of determination was calculated from the average of the standard demations of the analysis of 16 sediment extracts that contained less than 10.0 pg 1-1 of silver (nine of which were between 0.5 and 2.0 pg 1-l) and found to be 1.1 pg 1-l. None of the reference materials examined in this study have been certified for silver content Augwst 1983 OF CD AND AG IN SEDIMENTS BY ZEEMAN EFFECT AAS 923 However the NIES pond sediment has been analysed for this element by isotope dilution mass spectrometry. As for cadmium the measured concentrations are in reasonable agreement with those calculated from the value provided by the NIES (Table IV). Although the con-centrations of silver in these samples yield “undetectable” absorbances <O.OOO 99 the high spectral intensity of the hollow-cathode lamp permits a very low gain to be used.The resulting low noise enhances the ability of the microprocessor to discriminate atomic-absorption signals at silver concentrations below 10 pg 1-l. For the urban particulate matter (SRM 1648) a value of 6 pg g1 is provided by the NBS as information only. We obtained 6.4 pg g-l for a single sample subjected to the room-tempera-ture dissolution procedure which is in good agreement with the value reported by Greenb~rg.~’ Data on the chemical extraction of a sediment core from Moira Lake Ontario and coal fly ash (SRM 1633) are presented in Table VI. There is no value given by the NBS for the silver content of the coal fly ash. G l a d n e ~ ~ ~ reported a range of 0.25-1.3 pg g1 for coal fly ash with a 1975 certificate date.No value was listed for the SRM having a 1979 certificate date that was analysed in this study. Nevertheless the concentrations obtained here demonstrate the value of the polarised Zeeman effect AAS system for the determination of silver in extracts of environmental materials. In a recent study of weak acid-soluble silver in marine sediments,% the determination by direct aspiration AAS was considered unreliable for samples containing less than 2.5 pg g1 of silver. It was therefore necessary for these workers to resort to an ammonium tetramethylene dithiocarbamate - isobutyl methyl ketone extraction in order to determine silver in the range 0.1-2.5 pg g-l. As with the cadmium determinations the high concentrations of alkali and alkaline earth elements in our extracts did not cause burner problems as was noted in a determination of silver in corrosion test solutions containing 0.9% and 3% of sodium ~hloride.~g TABLE VI DISTRIBUTION OF THE CHEMICAL FORMS OF SILVER IN A SEDIMENT CORE AND IN COAL FLY ASH (SRM 1633) Total concentration/ Sediment depth/cm rg 8-l 0-1 .. 5.63 9-10 . . 8.05 14-15 ,. . . 6.60 18-19 . . 3.66 21-22 . . 4.30 32-33 . . 1.00 Coal fly ash 1 . . . 0.15 Coal fly ash 2 . . . . 0.14 Exchangeable forms/ r g g-l 0.52 0.63 0.90 0.12 0.10 0.03 0.02 0.02 Carbonate surf ace oxide bound/pg g-l 0.03 0.03 0.03 0.03 0.03 0.03 0.02 0.02 Fe Mn oxide bound/ 0.03 0.03 0.03 0.03 0.03 0.03 0.02 0.02 rg g-l Organic sulphide bound/ 2.31 2.02 1.30 0.24 0.36 0.16 0.03 0.02 r g l2-l Residual forms/ 2.70 5.40 4.40 3.30 3.84 0.12 0.12 Clg g-l 0.84 Hence using this instrumental technique it is now possible to determine directly the distri-bution of the chemical forms of silver and cadmium in environmental materials such as sus-pended and bottom sediments atmospheric particulates and soils.The data in Table VI show that significant amounts of silver can be found in sediments as readily extractable forms (mean value 7% of the total silver concentration; range 2-14y0). These forms are regarded as bioavailable because they are weakly bound and may equilibrate rapidly in water.2 The environmental significance of these results is at present being considered in a study of the factors affecting trace metal availability in aquatic systems.We are grateful to the referees for their helpful comments. The sample of NIES pond sediment was kindly supplied by Dr. Kensaku Okamoto. References 1. Associate Committee on Scientific Criteria for Environmental Quality “Effects of Cadmium in the Canadian Environment,” National Research Council of Canada Ottawa 1979 Issued as NRCC No. 16743. Forstner U. in Hutzinger O. Editor “The Handbook of Environmental Chemistry,” Volume 3, Part A Springer-Verlag New York 1980 p. 59. 2 924 LUM AND EDGAR 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28.29. Nriagu J. O. Editor “Cadmium in the Environment Part 1 Ecological Cycling,” Wiley-Inter-science New York 1980 682 pp. Davies P. H. Goettl J. P. and Sinley J . R. Water Res. 1978 12 113. Birge W. J. Hudson J . E. Black J. A. and Westerman A. G. in Samuel D. E. Hocutt C. H., and Mason W. T. Editms “Surface Mining and Fish/Wildlife Needs in the Eastern United States,” US Department of the Interior Fisheries Wildlife Service FWS/OBS-78/81 1978 p. 97. Lima A. R. Curtis C. Hammermeister D. E. Call D. J. and Felhaber T. A. Bull. Environ. Contam. Toxicol. 1982 29 184. Symeonides C. and McRae S. G. J. Environ. Qual 1977 6 120. Luoma S. N. and Jenne E. A. in “Biological Implications of Metals in the Environment,” Luoma S. N. and Bryan G. W. Sci. Total Environ. 1981 17 165.Gibbs R. J. Science 1973 180 71. Gupta S. K. and Chen K. Y. Environ. Lett. 1975 10 129. Tessier A. Campbell P. G. C. and Bisson M. Anal. Chem. 1979 51 844. Ure A. M. and Berrow M. L. in “Environmental Chemistry,” Volume 2 Specialist Periodical Katz A. and Kaplan I. R. Mar. Chem. 1981 10 261. Forstner U. in Nriagu J . O. Editor “Cadmium in the Environment Part 1 Ecological Cycling,” Brodie K. G. and Liddle P. R. Anal. Chem. 1980 52 1059. Fernandez F. J. and Giddings R. At. Spectrosc. 1982 3 61. Dawson J. B. Grassam E. Ellis D. J. and Keir M. J. Analyst 1976 101 315. Koizumi H. Yamada H. Yasuda K. Uchino K. and Oishi K. Spectrochim. Acta 1981 363 608. Tessier A. Campbell P. G. C. and Bisson M. Can. J . Earth Sci. 1980 17 90. Harrison R. M. Laxen D. P. H. and Wilson S. J. Environ. Sci. Technol. 1981 15 1378. Lum K. R. Betteridge J. S. and MacDonald R. R. Environ. Technol. Lett. 1982 3 57. Silberman D. and Fisher G. L. Anal. Chim. Acta 1979 106 299. Price W. J . “Spectrochemical Analysis by Atomic Absorption,” Heyden London 1979 p. 142. Gladney E. C. Anal. Chim. Acta 1980 118 385. Okamoto K. Editor “Preparation Analysis and Certification of ‘Pond Sediment’ Certified Research Report No. 38 National Institute for Environmental Studies, Greenburg R. R. Anal. Chem. 1979 51 2004. Dillon J. J. and Martin E. A, At. Spectrosc. 1982 3 66. Varma A. Talanta 1981 28 701. CONF-75029 US NTIS Springfield VA 1977 p. 213. Report Royal Society of Chemistry London 1982 p. 9. Wiley-Interscience New York 1980 p. 305. Reference Material, Japan 1982 112 pp. Received November 22nd 1982 Accepted March 22nd. 198