JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY FEBRUARY 1993 VOL. 8 127 Accurate Determination of Selenium in the Presence of Iron by Deuterium Arc Electrothermal Atomic Absorption Spectrometry Sunil Jai Kumar and S. Gangadharan Analytical Chemistry Division Bhabha Atomic Research Centre Bombay-400085 India A method is described which eliminates the over-correction due to spectral interference of iron on the total absorbance of selenium. A pyrolytic graphite platform was pre-treated with 1 mg ml-l of platinum and 5% m/v ascorbic acid solution. A 2 pl volume of a 1 mg ml-' palladium solution in 4% nitric acid and 0.5% m/v ascorbic acid was used to stabilize and enhance selenium sensitivity. A novel way of simultaneous introduction of palladium and ascorbic acid solutions without the precipitation of palladium is discussed.Palladium is found to delay atomization of selenium to temperatures of up to 1500 "C. Studies were carried out on 0.6 ng of selenium with varying amounts of iron in the furnace. For an absolute iron content of 40 pg in the furnace a recovery of 101 YO was obtained. The method was tested on a solution containing 50 ng ml-l of selenium in the presence of 2.5 mg ml-' of iron. The standard additions method was used to analyse the solution. A characteristic mass of 25 pg was obtained. Keywords Selenium determination; spectral interference due to iron; platinum-ascorbic acid pre-trea tment; electrothermal atomic absorption spectrometry; deuterium-arc background correction Iron is a major source of interference in the determination of selenium by electrothermal atomic absorption spectro- metry (ETAAS) with deuterium-arc background correc- t i ~ n .' - ~ The problem is associated with an over-correction due to spectral interference of iron at the 196.0 nm selenium wavelength. In general platinum has been used to remove this spectral interference; Peile et al.4 used carbon monoxide-platinum modification. Iron in the presence of oxygen forms FeO which according to these workers4 is responsible for the spectral interference. Carbon monoxide is added as an oxygen scavenger and platinum is believed to catalyse the reaction (2CO + 02-2C02). Bauslaugh et al.5 are of the view that platinum is sufficient to remove the interference by forming a platinum-iron alloy retarding the atomization of iron.Voth-Beach and Shradefl have used ascorbic acid and other reducing agents to reduce palladium to its elemental state for effective modification. A palladium-ascorbic acid mixed modifier has been used by Knowles and Brodie' using Zeeman-effect background correction in the determination of selenium in blood. Sampson* has determined selenium in serum with deuterium-arc background correction where the background absorption and atomic absorption are separated in time using a copper-magnesium modifier. In this work pyrolytic graphite platforms were pre-t reated with a platinum-ascorbic acid mixture to remove the spectral interference due to iron. The palladium-ascorbic acid modifier was used to stabilize and increase the selenium sensitivity. The method was developed to analyse selenium in water samples in the presence of 2.5 mg ml-' of iron.Table 1 instrumental conditions for the determination of sele- nium in the presence of iron Spectrometer- Background correction On Beam mode Double beam Wavelengthhm 196.0 Slit-width/nm 1 .o Slit-height Normal integration time/s 5.0 Lamp current/mA 10 Step No. Temperature/*C Time/s Argon/l min-' Graphite furnace- 110 110 1100 1100 110 110 2200 2200 2400 10 20-60 10 10 4.8 2.0 1.1 3.0 2.0 3.0 3.0 3.0 3.0 3.0 0 0 0 Read 3.0 Sampling- Sampling mode Automatic standard additions Volume of 200 ng ml-l Se/,ul 1-3 Volume of blanWp1 0-8 Volume of sample/,ul 5 Total volume delivered/,ul 10 Volume of palladium solution/pl 2 Experimental Instrumental Parameters A Varian Model AA-875 series spectrometer equipped with a GTA-95 graphite tube atomizer and an autosampler was used and deuterium-arc background correction was em- ployed.Eppendorff micropipettes were used for dilution. The details of furnace and spectrometer operating para- meters for the determination of selenium are shown in Table 1. A cooling step has been included in the ashing step to increase the size of the peak a~ea.~-ll Reagents Nitric acid. Suprapure Merck. Hydrochloric acid. Suprapure Merck. Selenium stock standard 2.5 mg ml-l. Prepared by dissolving high-purity selenium metal powder in nitric acid. Solution maintained in 5% v/v nitric acid. Palladium stock standard 1 mg ml-I. Prepared by dissolving palladium wire (99.97%) in nitric acid. Solution maintained in 4% v/v nitric acid.Platinum stock standard 1 mg ml-l. Prepared by dissolv- ing platinum wire (99.95%) in aqua regia [hydrochloric acid+nitric acid (4+ l)]. Final solution maintained in 5% v/v hydrochloric acid. Iron stock standard 10 mg ml-l. Prepared by dissolving high-purity iron in nitric acid. The final solution was maintained in 5% v/v nitric acid. Ascorbic acid general reagent grade. Ascorbic acid solu- tion (5% m/v) prepared by dissolving 250 mg of ascorbic128 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY FEBRUARY 1993 VOL. 8 Table 2 Furnace programme for platinum pre-treatment of the graphite platform Furnace temperature programme- Step No. Temperature/"C Time/s Argon11 min-' 1 2 3 4 5 6 7 8 9 10 110 110 250 1000 1000 110 110 2200 2200 2400 10 90 20 10 10 4.8 2.0 1.1 3.0 2.0 3.0 3.0 3.0 3.0 3.0 3.0 0 0 0 Read 3.0 Sampling- Sampling mode Automatic Volume of 1 mg m1-I of platinum solution/pl 10 10 Volume of 5Oh ascorbic acid solution/pl Multiple injection 20 No.of replicates 3 Last dry phase step 4 acid in 5 ml of freshly prepared de-ionized water. This was diluted ten times to prepare a 0.5% m/v ascorbic acid solution. Sample Preparation Selenium solution (200 ng ml-l) was prepared daily in 0.5% v/v nitric acid from dilution of the stock standard. Aliquots of 1 2 3 and 4 pl of 10 mg ml-l of iron solution corresponding to 10,20,30 and 40 pg absolute iron content respectively along with 3 pl of 200 ng ml-l of selenium solution (0.6 ng absolute selenium content) were introduced into the furnace for recovery tests. Selenium (50 ng ml-l) in 2.5 mg ml-l of iron was prepared from the stock solutions to check the validity of the method.A standard additions calibration graph was obtained by taking 1 2 and 3 pl of 200 ng ml-I of selenium standard and 5 pl of the sample solution in 0.5% ascorbic acid. A 2 p1 aliquot of 1 mg ml-l of palladium was used as modifier. An autosampler was used for the preparation of standards. Platform Pre-treatment The pyrolytic graphite was pre-treated by depositing 10 pl of 1 mg ml-l of Pt and 5% ascorbic acid each 20 times with the autosampler. The platform was heated in the graphite furnace under the conditions shown in Table 2. The entire procedure was repeated three times (overall 60 depositions) to obtain a lasting coating. A total volume of 20 pl was added for complete coverage of the platform surface.Care was taken to dry the solutions completely in the drying step. Results and Discussion Interference Due to Iron The baseline obtained in the presence of 20 pg of iron using a pyrolytic graphite platform in the absence of selenium is shown in Fig. 1 A. The negative dip in the baseline illustrates that the absorbance of the deuterium arc is greater than the total absorbance of the selenium hollow cathode lamp. This indicates the presence of a spectral interference as a result of which an over-corrected absor- bance of selenium is obtained (ie. total absorbance of selenium is lower than the actual value). The use of a platform pre-treated with platinum removes the spectral interference as a perfectly flat baseline is obtained (Fig.1 0.3 I 3000 0 3.0 Tirnels 6.1" Fig. 1 Traces of the selenium absorbance signal on the tempera- ture-time profile A 20 pg of iron on a pyrolytic graphite platform without selenium; B 20 pg of iron on a platform pre-treated with platinum and ascorbic acid without selenium; C 0.6 ng of selenium 2 pg of palladium and 10 pg of ascorbic acid on a platform pre-treated with platinum and ascorbic acid; and D 0.6 ng of selenium 2 pg of palladium 10 pg of ascorbic acid and 20 pg of iron on a platform pre-treated with platinum and ascorbic acid Table 3 Selenium recoveries with varying iron content in the furnace Recovery (O/o) Iron content/pg Selenium added/ng (n=4) 10 20 30 40 0.6 0.6 0.6 0.6 103 96.3 96.9 101 B) in the presence of 20 pg of iron without selenium.Peak shape due to 0.6 ng of selenium with a palladium-ascorbic acid mixed modifier in the absence of iron (Fig. 1 C) is similar to that in the presence of 20 pg of iron (Fig. 1 D) and has the same integrated absorbance. The recoveries of 0.6 ng of selenium with varying iron concentrations are compared in Table 3. Recoveries in the range of 96-103% were obtained for 10-40 pg of absolute iron content in the furnace. Ascorbic acid upon pyrolysis in the furnace is known to produce carbon monoxide and hydrogen. 12~13 The carbon monoxide produced along with platinum may be instru- mental in removing the overcorrection by acting as an oxygen scavenger which prevents the formation of Fe0.4 However a definite explanation regarding the actual mechanism of removal of this overcorrection requires further investigation.The relative standard deviation (RSD) obtained varied from 3 to 10%. The large RSD obtained even in the absence of iron is because the formation of selenium atoms depends on the surface properties of platinum deposited on the platform. These may not remain constant from one firing to the next. Hence there is a slight irregularity in the peak shapes obtained which affects the precision. Palladium-Ascorbic Acid Modification As platinum combines4 with selenium affecting its sensitiv- ity palladium-ascorbic acid modification has been used to stabilize and improve the selenium sensitivity. Palladium solution in the presence of ascorbic acid is readily reduced to elemental palladium even at room temperat~re.~.' Hence the sequence in which the palladium and ascorbic acid solutions are added is very important. The selenium absorbance is shown in Table 4 when ascorbic acid and palladium solutions are taken in by the autosampler capillary in the following sequences A ascorbic acid standard palladium solution; and B blank palladium solution standard ascorbic acid.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY FEBRUARY 1993 VOL.8 129 0.15 \ 8 e 2 % 0.1 (D 0 + z ' 0.05 m C Life of Pre-treated Platforms Pre-treatment of the platform with platinum and ascorbic acid by this method is found to have a lasting effect. For up to 200 firings there was no sign of any over-correction due to iron. Deterioration started only after 230 firings. The use of a small amount of platinum in each aliquot (10 pg) ensures a uniform and lasting coating.However the lifetime of the tube is affected by highly concentrated acidic solutions; the coating is removed by dissolution of platinum in acidic solutions. 1 I I I 1 I 1100 1300 1500 1700 1900 Ashing temperature (wall temperature)/"C Fig. 2 Effect of varying the ashing temperature on A 0.6 ng of selenium 2 pg of palladium 10 pg of ascorbic acid and 20 pg of iron on a pyrolytic graphite platform pre-treated with platinum and ascorbic acid; and B 0.6 ng of selenium and 2 pg of palladium on a pyrolytic graphite platform Results The method was tested on sample solutions containing 50 and 100 ng ml-I of selenium respectively and 2.5 mg ml-I of iron. Values of 49 and 103 ng m1-I were obtained with RSDs of 7.5 and 7.8% respectively for n= 3.The character- istic mass (amount of selenium giving an integrated absorbance of 0.0044 s) was found to be 25 pg. Table 4 Comparison of integrated absorbance for 0.6 ng of Se with sampling sequences A and B; (see Fig. 2 ) Absorbance Replicate Sequence 1 2 3 4 Mean RSD (Oh) A 0.080 0.076 0.056 0.056 0.067 19 B 0.118 0.129 0.124 0.132 0.125 5 In the case of sequence A the selenium absorbance continues to decrease with increasing number of replicate measurements (Table 4). For sequence B the absorbance remains constant with increasing number of replicate measurements. In sequence A ascorbic acid solution is taken into the capillary first and palladium last hence a large area of the capillary surface is contaminated by ascorbic acid and palladium solution remains in contact with this surface for the entire duration of its time in the capillary. This results in the deposition of palladium on the surface of the tip.While the solution is being delivered some of the selenium is withheld by palladium deposited on the tip of the capillary resulting in reduced absorbance. For sequence B the surface area contaminated by ascorbic acid is smaller and palladium solution comes in contact with this surface for a very short time (just while leaving the capillary). Therefore deposition of palladium does not occur resulting in a constant absorbance. No palladium precipitation was noticed on the tip for the entire duration of analysis for 2-5 p1 of 0.5% m/v ascorbic acid solution. For sequence A palladium deposited on the tip of the capillary can be removed by rinsing it with 20% v/v nitric acid for 3-4 min.A wash solution of 0.5% nitric acid is preferable for rinsing in between samplings by the autosam- pler. The effect of varying ashing temperature (wall tempera- ture) on the integrated absorbance due to 0.6 ng of selenium with platinum pre-treated platforms in the presence of 20 pg of iron 2 pg of palladium and 10 pg of ascorbic acid and with pyrolytic graphite platforms in the presence of 2 pg of palladium is shown in Fig. 2. Selenium loss in both cases starts at temperatures above 1500 "C which is close to the melting-point of palladium of 1554 OC.14 In the presence of ascorbic acid higher integrated absorbance is observed because of the effective reduction of palladium.6 Conclusion It has been shown that spectral interference due to iron is effectively removed by a novel method of pre-treating the platform with a platinum-ascorbic acid mixture.The method of pre-treatment is simple and does not require any additional reorganization to allow mixing of gases.4 The pre-treated platform was found to be effective for up to 200 firings in removing the over-correction due to iron. A novel method of simultaneous introduction of palladium and ascorbic acid solutions with an autosampler without palladium precipitation has been described. Use of a palladium-ascorbic acid mixed modifier permitted a higher ashing temperature. Interference-free determination of sele- nium in 2.5 mg ml-1 of iron solution has been demon- strated. 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 References Slavin W. and Carnrick G. R. At. Spectrosc. 1986 7 9 . Carnrick G. R. Manning D. C. and Slavin W. Analyst 1983 108 1297. Fernandez F. J. and Giddings R. At. Spectrosc. 1982 3 61. Peile R. Grey R. and Starek R. J. Anal. At. Spectrum. 1989 4 407. Bauslaugh J. Radziuk B. Saeed K. and Thomassen Y. Anal Chim. Acta 1984 165 149. Voth-Beach L. M. and Shrader D. E. J. Anal. At. Spectrum. 1987 2 45. Knowles M. B. and Brodie K. G. J. Anal. At. Spectrum. 1988 3 51 1. Sampson B. J. Anal. At. Spectrom. 1987 2 447. Falk H. Gilismann A. Bergann L. Minkwitz G. Schubert M. and Skole J. Spectrochim. Acta Part B 1985 40 533. Parsley D. H. J. Anal. At. Spectrum. 1991 6 289. Frech W. Li K. Berglund M. Baxter D. C. J. Anal. At. Spectrom. 1992 7 141. Gilchrist G. F. R. Chakrabarti C. L. and Byme J. P. J. Anal. At. Spectrom.,1989 4 533. Gilchrist G. F. R. Chakrabarti C. L. Byme J. P. and Lamoureux M. J. Anal. At. Spectrum. 1990 5 175. CRC Handbook of Chemistry and Physics ed. Weast R. C. CRC Press 70th edn. 1989-90 B-112. Paper 2/0 124 7B Received March 9 1992 Accepted June 15 1992