1060 Analyst September 1983 VoE. 108 $9. 1060-1066 Evaluation of Selenium Determination in Biological Material by Atomic-a bsorption Spectroscopy Rodney J. Mailer NS W Department of Agriculture Agricultural Research Institute Wagga Wagga New South Wales 2650, Australia and James E. Pratley School of Agriculture Riverina College of Advanced Education Wagga Wagga New South Wales 2650, A ustmlia Existing methods of selenium determination by atomic-absorption spectro-scopy using the hydride vapour generation technique were examined. Of the factors investigated it was found that the method of digestion could be a source of considerable error. Stirring time in the vapour generator, acid concentration and final solution volume were also critical factors affecting the consistency of results.Sodium tetrahydroborate(II1) pellets were found to have considerable advantages over a solution of the same chemical owing to their stability and the uniform rate of addition and reaction. Keywords Selenium determination ; biological material ; atomic-absorption spectroscopy ; hydride vapour generation The determination of nanogram levels of selenium in biological material has generally been based on a fluorimetric technique with 2,3-diamin~naphthalene,~ a method that has been improved by the use of semi-automated techniques.2 Other methods available include atomic-absorption spectro~copy,~s~ radiochemical methods15 gas - liquid chromatography6 and, more recently inductively coupled plasma emission ~pectroscopy.~ Analysis by atomic-absorption spectroscopy has been shown to be a quick and simple method of routine analysis using the hydride vapour generation technique3s8p9 and is currently used on a routine basis in many laboratories.Although the literature contains several references to this method there appear to be a number of factors generally overlooked that have considerable bearing on accuracy and precision. Through the necessity to mineralise all forms of selenium and at the same time prevent loss through volatilisation at high temperatures or in reducing conditions the method of sample dissolution is a major factor in obtaining consistently accurate results. Several workers have studied the types of oxidation mixtures used in dissolution including perchloric acid - sulphuric acidlg nitric acid - urea,6 phosphoric acid - hydrochloric acidlo and most frequently nitric acid - perchloric a ~ i d .~ * ~ - ~ v ~ Conditions within the vapour-generation chamber including volume of solution acid concentration and acid composition must be consistent. The solution contains hydrochloric acid used to reduce selenium(V1) to selenium(IV) the other components will depend on the digestion method employed. Volumes of 5-50 ml have been used and hydrochloric acid concentrations within the chamber of between 5 and 50% .499910 Increasing the concentration of the acid solution or the introduction of other acids e.g. nitric acid will reduce the peak absorbance. l1 The presence of oxygen within the silica cell causes significant absorbance a t 196.0nm. A concise routine of purging and stirring is therefore necessary prior to the addition of sodium tetrahydroborate(II1) .lo Sodium tetrahydroborate(II1) has been used both in sodium hydroxide solution at con-centrations of 2.5-6yO9 and in the dry state as pellets.The solution is relatively unstable and is added with a syringe via a septum and therefore at a variable rate. Pellets have been shown to be more reproducible and more convenient owing to their stability. This paper reports on some aspects of previous methods using the atomic-absorption technique outlining sources of variability and means of minimising them MAILER AND PRATLEY Experimental Apparatus and Conditions 1061 A thermostatically controlled aluminium block with a two-step heat control and over Holes were drilled in the block to A Varian Model AA875 double-beam atomic-absorption spectrophotometer was used in temperature cut-out was used for sample digestion.accommodate 25 Pyrex digestion tubes 32 mm diameter x 215 mm tall. conjunction with a Varian Model 65 hydride generator. Atomic-absorption conditions The conditions used were as follows light source Varian hollow-cathode lamp with a deuterium background correction lamp; lamp current 10 mA ; wavelength 196.0 nm ; spectral band pass 1.0 nm; fuel - support acetylene - air; and measurement mode peak height 15 s integration time. Recorder The recorder was a Varian 9176. A chart speed of 5 cm min-l and a pen response of 10 mV full-scale deflection were used. Hydride vapour generator The following conditions were used internal dead volume 150 ml [+30 cm of 6 mm poly(viny1 chloride) tubing] ; absorption cell quartz 30-cm path length ; nitrogen supply pressure 140 kPa; and flow-rate 2 1 min-l.Reagents The digestion acid used for blood samples was a 2 + 1 V/V mixture of analytical-reagent grade nitric acid - perchloric acid. Hydrochloric acid 6 M. Distilled from borosilicate glass apparatus. Hydrochloric acid 20% V/V. Diluted from 6 M hydrochloric acid. Sodium tetrahydroborate(ll1) powder. Sodium tetrahydroborate(II1) pellets. Selenium standard solution 1000 pg ml-l. Nitric acid sfi. gr. 1.42. Perchloric acid 72% V/V. AnalaR grade from BDH Chemicals Ltd. For plant material a 5 + 2 mixture was used. Supplied by Alfa products Danvers MA USA. Supplied by Alf a products.Supplied by BDH Chemicals Ltd. AnalaR grade from BDH Chemicals Ltd. Preparation of Samples Blood samples A 1.00-ml sample of whole blood was measured into a clean test-tube using a positive displacement micropipette to ensure total dispensing of the viscous medium. Carefully, 3.0 ml of nitric acid - perchloric acid were added; the tube was covered and allowed to stand overnight. (Standing overnight was not found to be necessary with blood samples if a slow rate of heating was employed i.e. ambient to 150 "C over a period of 2 h.) The following morning the tubes were placed in a cold digestion block and heated to 150 "C. This tempera-ture was maintained until brown fumes of nitric acid had disappeared and white wispy fumes of perchloric acid had become visible.Charred samples were discarded although this was infrequent with the blood samples. The temperature of the block was raised to 200 "C and held for approximately 30 min at which time dense white fumes could be seen. This was to ensure the complete removal of nitric acid and the total conversion of any organic forms of selenium into selenous acid. The tubes were then removed and allowed to cool to ambient temperature prior to the addition of 0.2 ml of 6 M hydrochloric acid after which they were heated at 150 "C for 15 min to reduce selenium(V1) to selenium(1V). The tubes were removed and allowed to cool then 20 ml of 20% V/V hydrochloric acid were added to each. This process took about 90min. Plant samples Extra nitric acid was required to break down the fibrous material and it was necessary to add the perchloric acid separately, after the initial digest to avoid the loss of selenium due to charring.The digestion process was modified for plant material 1062 MAILER AND PRATLEY EVALUATION OF SELENIUM Analyst Vol. 108 A 1.0-g sample of dried plant material was weighed into each sample tube and 5 ml of nitric acid were added. A glass funnel was inserted into the top of the tube and was left to stand overnight. The next morning the tubes were placed in a cold aluminium digestion block and the temperature was increased slowly over a period of 30 min to 130 “C. Dense brown fumes were formed in the tubes. This temperature was maintained until the fumes became a light honey colour (approxi-mately 1 h) then 2 ml of perchloric acid were added through the funnel.After 2 min the funnels were removed and the temperature was increased to 150-170 “C and maintained until the nitric acid was removed and white wispy fumes of perchloric acid were evident. The temperature was further increased at this stage to 190-200 “C and held for a further 30 min to remove the final traces of nitric acid. The tubes were cooled to ambient temperature 0.2 ml of 6 M hydrochloric acid was added and then they were re-heated to 150 “C for 15 min. The tubes were again cooled to ambient temperature and 20 ml of 20% hydrochloric acid were added to each. Atomic-absorption Determination The method of calibrating the instrument was critical for achieving accurate and repro-ducible results. The Varian AA875 spectrophotometer has three modes of calculation the integration-repeat mode in which the average reading over a given integration time is displayed and continually updated ; the peak-height mode which indicates the highest reading over a time interval; and the peak-area mode which indicates the area under the graph during the hydride generation reaction.Although peak-area calculation has been reported to produce consistent res~lts,l~~J’ our work and that of others10 has shown peak-height calculation to be more reliable. The time required for peak area is in an excess of 25 s whereas peak height required only 10-15 s. The method of calibration was to allow the chamber to purge with nitrogen for about 2 min and then to calibrate zero in the integrate repeat mode. An integration time of 15 s was then called and the peak-height mode engaged.The instrument was zeroed using a blank solution as follows. The vapour generator lid was slightly raised above the chamber and 21 ml of blank solution (1 ml of perchloric acid 20 ml of 20y0 hydrochloric acid) were dispensed into the chamber The lid was closed quickly to avoid an excess of air from entering the chamber and the stop-watch was started. After 15 s the magnetic stirrer was turned on for 5 s before “Cal Zero” and “READ” were pressed and a sodium tetrahydro-borate(II1) pellet was immediately dispensed into the chamber. After 15 s the indicator displayed 0.000 and the instrument was zeroed. The drain plug was pressed to remove the previous sample. The lid was raised and the sample dispensed into the chamber.Immediately the lid was closed and the stop-watch started. After 15 s the stirrer was started and stirring was continued for 5 s before “READ” was pressed. Immediately a pellet was dispensed into the chamber and the absorbance value was obtained from the digital indicator 15 s later. This process was repeated for all standards and samples. A chart recording was run sequentially to ensure that a characteristic response was obtained for each sample and the peak height was within the integration time allowed. Re-calibration of zero was necessary as the recorder pen failed to return to the base line following sample reading. Rinsing the chamber between samples was found to be unnecessary. Reading of the sample solutions was essentially the same as for calibrating zero.Results and Discussion Sample Preparation Cleanliness of glassware The utmost care needs to be taken with cleanliness of glassware to avoid contamination. The glassware was new and was kept separate from all other glassware. Following each digest tubes were immediately rinsed several times with distilled water and inverted to drain dry. No detergents or acids were used for washing up. The use of selenium as a catalyst in Kjeldahl digests is a potential source of significant contamination if both analyses are undertaken in the same laboratory September 1983 DETERMINATION IN BIOLOGICAL MATERIAL BY AAS 1063 Digestion pyocess Rapid heating in the initial stages of digestion caused vigorous boiling which gave low readings from samples of known concentrations of selenium possibly due to physical loss.Charring may also result and this leads to the volatilisation of selenium. All samples showing signs of charring were discarded. Fig. 1 shows the effect of charring on duplicate samples of blood. It was found that plant samples showing slight darkening when the temperature was increased from 150 to 190 "C could be recovered by adding 0.5-1.0 ml of nitric acid and re-digesting at 150 "C. It was more reliable however to discard the digest and start anew. Complete removal of nitric acid from the digested solutions is also stressed as its presence affects the absorbance and creates erroneous results (Fig. 2). The majority of nitric acid should be removed whilst the temperature is maintained at 150 "C to avoid vigorous boiling.Final traces of nitric acid are removed at 200 "C in the final digestion step. 0.10 Q) C m e s 3 0.05 0 C Fig. 1. (A) Effects of charring during digestion on the absorbance value of a blood sample compared with (B) the same sample without charring (without background correction). (C) Blank solution. 0.10 8 $ a m e 0.05 n 0 1 2 3 Volume of nitric acid addedhl Fig. 2. Effects of nitric acid on absorbance. Sample concentration 100 ng per 20 ml. Reduction of selenium The success of the reduction of selenium(V1) to selenium(1V) is critical to the reliable determination of selenium by atomic-absorption spectroscopy. The reduction step occurs according to the reaction Se0,2- + 2HC1+ Se0,2- + C1 + H,O 6 C D E F G H Fig.3. Selenium recovery in standard solutions from 100 ng of selenium with different methods of reduction. A 0.2 ml of 6 N HC1 at 150 "C for 5 min; B 0.2 ml of 6 N HC1 at room temperature; C 20 ml of 20% HC1 at 150 "C for 15 min; D 10 ml of 20% HCl at 150 "C for 5 min; E 5 ml of 20% HC1 a t 150" C for 15 min; F 5 ml of 20% HC1 a t 150 "C for 10 min; G 5 ml of 20% HCl at 150 "C for 5 min; and H 5 ml of 20% HC1 a t room temperature 1064 MAILER AND PRATLEY EVALUATION OF SELENIUM Analyst VoZ. 108 The need to obtain a complete reduction of selenium has been indicated.14 The conditions for our determinations were 150 "C for 10-15 min as outlined elsewhere,2 although other workers have used lower temperatures99l4 and in some instances no heating at all.4 Hydrochloric acid was provided by adding 0.2 ml of 6 N hydrochloric acid.A larger volume was found to be unnecessary and required longer to reach the temperature necessary to reduce selenium(V1) to selenium(1V). Standards did not require reduction. Nine samplings containing 100 ng of selenium and reduced by different methods resulted in readings with a mean of 0.081 ng and a range of 0.080-0.084 (Fig. 3). All samples were however, subjected to the same reduction step in order to standardise conditions. Y 0 50 150 250 Seleniuming mi-' Fig. 4. Effect of background correc-tion on absorbance. A With back-ground correction; and B with no background correction. 50 2q 10 5 Volume of HCl/ml Fig. 5. Absorbance spectra from 100ng of selenium added in various volumes of 20% m/V HCI.Determination by Atomic-absorption Spectroscdpy Background correction Although background correction was not found to be essential it resulted in more repro-ducible peaks less noise and greater accuracy. Two calibrations were obtained with and without background correction (Fig. 4). The mean values of the standards were obtained (130 ng) which would be expected to give absorbance values of 0.144 and 0.223 respectively. To test the value of the two methods, 0 5 10 20 50 Acid concentration YO V/V Fig. 6. Effect of hydrochloric acid concentration on absorbance readings. Selenium concentration, 100 ng per 20 ml September 1983 DETERMINATION IN BIOLOGICAL MATERIAL BY AAS 1065 95% tolerance limits for a selenium content determination were ca1c~lated.l~ The limits obtained were (125 135) and (98 170) respectively.The recorder trace in Fig. 1 without background correction indicates the extent of background absorbance. 0.10. 0 (II e a a 0 -Volume The volume of the sample used in the vapour generator chamber is a critical factor. An increase in volume produced a wider and flatter peak (Fig. 5). This is in agreement with the findings of Brodiell for arsenic determination but conflicts with those reported by Hobbins.lo However the findings in each instance indicate the importance of consistency of volume in obtaining reproducible results. Concentration of acid Increasing acid concentration in the solution resulted in decreased absorbance (Fig. 6). In our investigations 20% hydrochloric acid gave good reproducibility and was used for all analyses.Pellets versus sodium tetrahydro borate(III) solution (2.5%) With the use of solution it was necessary to use a fine gauge needle when penetrating the septum to prevent leakage. It has also been shown that concentrations of less than 4% can cause tailing of peaksg although a 2.5% solution is commonly used.16 Pellets were more consistent despite absorbance levels being lower in solutions of equivalent selenium concentration (Fig. 7). An additional disadvantage of the solution is its instability requiring fresh preparation after 2 h. This is an important consideration when a large number of samples are to be analysed . Consequently the injection time was slow and variable. I C D I Solution Fig. 7. Comparison of the absorbance from a sample con-taining 100 ng of selenium using sodium tetrahydroborate(II1) C and D solution and A and B pellets.Purging and stirring time Purging of the sample with nitrogen to remove all traces of oxygen and stirring prior to the addition of the pellets were necessary in order to obtain reproducible results (Figs. 8 and 9). We used a purging time of 20 s the last 5 s of which the sample was stirred as recommended by Hobbins.lo Conclusion Determining deficiencies of selenium in plant material or blood often involves samples containing 20ngml-l of selenium or less. The ability to measure these values require 1066 MAILER AND PRATLEY Stirring time/s Fig. 8. Effect of stirring time on sample absorbance with no initial purging. Selenium concentration, 100 ng per 20 ml.5 20 40 80 Purging time/s Fig. 9. Effect of purging time on sample absorbance. Selenium con-centration 100 ng per 20 ml. Stirring time 10 s. extreme accuracy cleanliness and in particular consistency in the method. The importance of acid volume and concentration and the large variations in absorbance with purging and stirring times indicate the sensitivity of the method employed. The aspects of the method outlined in this paper allows precise and accurate measurements of trace amounts of selenium. The method has the advantage of speed and simplicity provided that the precautions outlined are observed. 1. 2. 3. 4. 6. 6. 7. 8. 9. 10. 11. 12. 13. 14. 16. 16. References Watkinson J. H. Anal. Chem. 1966 38 92. Watkinson J. H. Anal. Chim. Ada 1979 105 319. Thompson K. C. and Thomerson D. R. Analyst 1974 99 595. Clinton 0. E. Analyst 1977 102 187. Gorsuch T. T. Analyst 1959 84 135. Kurahashi K. Inoue S. Yonekura S. Shimoishi Y. and Tbei K. Analyst 1980 105 690. Pahlavanpour B. Pullen J. H. and Thompson M. Analyst 1980 105 274. Hall R. J. and Gupta P. L. Analyst 1969 94 292. Lloyd B. Holt P. and Delves H. T. Analyst 1982 107 927. Hobbins. W. B. “Varian Instruments a t Work AA-11,” Varian AA Resource Centre Park Ridge, Brodie K. G. Am. Lab. 1979 11 58. Nhve J. Hanocq M. Molle L. and Lefebvre G. Analyst 1982 107 934. Brodie K. G. “Atomic Absorption Spectroscopy Asian Seminar Series,” Varian Techtron Pty. Ltd. Springvale Australia 1980. Brodie K. G. Am. Lab. 1977 9 73. Williams E. J . “Regression Analysis,” John Wiley New York 1979. “Model 65 Vapour Generation Accessory Operation Manual,” Publication No. 85-100338-00 Varian Illinois 1981. Techtron Pty. Ltd. Springvale Australia. Received March 141h 1983 Accepted March 31st 198