JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY NOVEMBER 1994 VOL. 9 1289 Determination of Trace Amounts of Thallium in Nickel-based Alloys by Electrothermal Atomic Absorption Spectrometry Rajananda Saraswati* N. R. Desikan and T. H. Rao Defense Metallurgical Research Laboratory Kanchanbagh P. O. Hyderabad-500 258 lndia A method is described for determining trace amounts of thallium in nickel-based alloys by electrothermal atomic absorption spectrometry (ETAAS). The alloy was dissolved by two methods one a microwave digestion procedure and the other a hot-plate dissolution method. Different acids were used for each of the two different methods. For microwave digestion a combination of H2S04 HNO and HF acid was used whereas for hot-plate dissolution a combination of H,SO HF and H202 was used.Ascorbic acid (C6H&@6) was used as chemical modifier. Comparable results for TI are achieved by the two methods of dissolution. The installation of a stabilized temperature platform in the graphite furnace tube improves the absorbance signal and the precision of the measurement. The detection limit was 0.1 ng ml-' of TI ( 3 ~ ~ ) . The accuracy of the method was elucidated through the analysis of certified nickel-based alloys. Keywords Thallium; nickel-based alloy; atomic absorption spectrometry Nickel-based heat-resisting alloys are used extensively in the aerospace industry for the construction of turbine blades. The presence of certain trace elements in nickel-based alloys can have serious effects on the mechanical and physical properties.Thallium is one of the deleterious trace elements that reduces the workability elongation and creep-rupture life of nickel- based alloys when present even at trace levels.'-3 The determi- nation of trace elements including T1 in these alloys has been a challenge because of the complexity of the matrices. Many methods are available for the determination of trace amounts of T1; electrothermal atomic absorption spectrometry (ETAAS) is among the most widely Sample preparation is one of the important steps in the determination of T1 by ETAAS. Serious signal depression can be caused by the presence of HCI' or HC104.' Matrix interference is another difficulty that has to be overcome in the accurate determination of T1. The use of H2S04' has been suggested for the elimination of HCl interference; use of pyrolytic graphite coated graphite tubes could reduce the interference due to HC104.10 Chemical modi- fication with organic or inorganic compounds can be used for the suppression of matrix The objective of this work was to develop a microwave dissolution method for nickel-based alloys for the determi- nation of T1 and to study the effect of chemical modifiers on the accurate determination of T1 in certified reference materials.This work also proves the merits of using a stabilized tempera- ture platform furnace to determine the trace amounts of T1. Experimental Instrumentation A GBC Model 902 atomic absorption spectrometer equipped with a GBC Model GF 2000 graphite furnace was used. A Visimax I1 hollow cathode lamp (HCL) was used as a T1 light source and a deuterium lamp was used for the background correction.Pyrolytic graphite coated graphite tubes and pyro- lytic graphite platforms supplied by GBC Instruments Australia were used. High purity argon gas was used as the purge gas with a flow rate of 50mlmin-'. Solutions were injected into the graphite furnace by a GBC PAL 2000 auto- sampling system. Integrated absorbance (peak area) values were used in measurements which were recorded with an Epson Lx-800 printer. Microwave-digestion was carried out in a micro-wave oven * Present Address Inorganic Analytical Research Division National Institute of Standards and Technology Gaithersburg Maryland 20899 USA. Toshiba Model E855 BTC that operated between 72 and 720 W with 81 W increments. The Teflon perfluoroalkoxy (PFA) vessels used for the dissolution had a volume of 200 ml and tight-fitting screw-cap lids.Reagents High-purity Suprapure grade chemicals from Merck were used. The lab-ware was used only after soaking in 10% HNO for several hours and then equilibriating in Milli Q water (18 MR resistance; Millipore). A stock standard solution containing 1000 pg ml-' of T1 was reported by dissolving specpure thallium(1) nitrate or sulfate in nitric or sulfuric acid. Working T1 standard solutions were prepared every day by serial dilution of the stock stan- dard solution. Digestion Procedure Wet digest ion This procedure is similar to that described e1~ewhere.l~ Two samples from each of standard reference material (SRM) 897 SRM 898 and SRM 899 [Trace Alloys; National Institute of Standards and Technology (NIST) Gaithersburg MD USA] and certified reference material (CRM) 345 and CRM 346 [IN 100 Alloy; Bureau of Analysed Standards (BAS)] together with two blanks were prepared.An accurately weighed (1 g) sample of nickel-based alloy was placed into a platinum crucible and moistened with distilled water and 7.0 ml of concentrated HF and 7.0 ml of concentrated H2S04 were added. The crucible was covered with a platinum lid and heated gently for 3 h on a hot-plate till 90% of the sample was dissolved. Then 2 ml of 30% H202 were added and heated again gently for 1 h to dissolve the sample completely. The crucible was allowed to cool to room temperature and 5 ml of 5% m/v ascorbic acid were added. The contents of the crucible were carefully transferred into a 100 ml polypropylene Cali- brated flask and diluted up to the mark with de-ionized water.Microwave-oven digestion procedure Two samples from each of SRM 897 898 899 and CRM 345 346 together with two blanks (one blank for each set) were prepared. Samples of nickel alloy (1 g) were accurately weighed into the PFA microwave digestion vessels. To each vessel were added 15 ml of 10% v/v HNO 5 ml of concentrated H2S04 and 1 ml of concentrated HF. The vessels were capped tightly and were placed (six at a time) in the microwave-oven. The 1 h digestion programme was set with 40% power for 10 min,1290 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY NOVEMBER 1994 VOL. 9 60% power for 30 min and 40% power for 20 min. Following the digestion programme the samples were left for 15 min in the microwave oven and then removed from the microwave oven and placed in a tray of ice water for about 30 min.Then the vessels were carefully uncapped their contents transferred into 100 ml polypropylene calibrated flasks 5 ml of 5% m/v ascorbic acid were added and the solution was diluted to volume with distilled water. Instrumental Analysis Procedure The above alloy solutions were analysed by injecting 20-50 pl portions into a pyrolytic graphite coated graphite tube using the optimized conditions for drying ashing and atomizing steps as shown in Table 1. The corresponding blank solutions were also analysed. The calibration graphs were obtained by injecting different amounts of T1 standard solution (Fig.1). The solutions were also analysed by injecting onto the platform located in the graphite tube using the instrumental conditions detailed in Table 1. Blanks were run regularly and their values were subtracted to obtain the net absorbance values. Results and Discussion Atomization of T1 suffers seriously from the severe interference of matrix elements and acids in ETAAS. Chemical modifiers Table 1 in Ni-based alloys Optimum analytical conditions for thallium determination Instrument mode Beam mode Wavelength/nm Slit-width/nm Lamp current/mA Integration time/s Sampling mode Furnace Furnace step temperature/ No. "C 1 80 2 110 3 400 4 500 (550)* 5 1800 (2200)* 6 2500 Ramp time/ 5 10 15 15 2 5 S Absorbance BC on Double beam 276.8 0.5 (0.7)* 10.0 (8.0)* 3.0 Auto sampling Hold time/ Gas flow/ s mlmin-' 5 50 10 50 15 50 15 0 2 0 3 50 Read on No No No No Y No * Platform.0.30 0.25 0.20 a C $ 0.15 0 Q .i? 0.10 0 5 10 15 20 25 30 Concentrationhg mi-' Fig. 1 platform; B without platform Calibration graph obtained for T1 standard solutions; A with enhance the volatility of unwanted elements of the matrix and stabilize the analyte during the pyrolysis stage. Sulfuric acid and palladium are frequently used as effective chemical modi- fiers for the determination of T1. In the present study the solutions were prepared in an acid combingtion that included H2S04 to avoid the interference which results from HC1 or HC104 as decomposition acids. The interference of the alloy matrix could be removed without any separational technique by using ascorbic acid as a modifier.Studies were conducted to optimize the ascorbic acid concentration. It was found that 0.25 g of ascorbic acid is sufficient to suppress the matrix effect of 1 g of nickel alloy sample (in 100ml). In the presence of H2S04 T1 forms the stable oxide instead of the volatile halide. It was observed that 6-8% v/v H2S04 is optimum for the best absorbance signal. The effect of vigorous heating at high temperatures during the digestion step was monitored. It was observed that excess heating at higher temperature during the dissolution process resulted in evaporation of T1; therefore all precautions were taken to avoid this. Ashing temperature atomizing temperature and time all play a very important role in the absorbance of T1 in nickel- based alloys.The effect of varying ashing temperature on the absorbance of T1 was studied keeping the atomizing tempera- ture constant. Fig. 2 shows the influence of ashing temperature on the integrated absorbance of T1 in nickel-based alloys. The maximum absorbance signal of TI was observed at an ashing temperature of 500°C. The absorbance of T1 decreased at lower and higher ashing temperatures. This may be explained on the basis that at low pyrolysis temperatures T1 is co-volatilized with the matrix resulting in a reduction in T1 residence time and therefore atomic absorption of thallium decreases. At 500"C the reduction or decomposition of these compounds will become maximum and formation of T1 atoms will be higher; accordingly the absorbance of T1 is increased. However at higher ashing temperatures T1 compounds will evaporate as molecules and the result is a decrease in atomic ab~orpti0n.l~ The effect of varying the atomizing temperature while keep- ing the ashing temperature constant was also investigated. Fig.3 shows the influence of atomizing temperature on the absorbance of T1 in nickel-based alloys. An atomizing tempera- ture of 1800 "C (without platform) and 2200 "C (with platform) produced maximum absorbance for T1. The effect of installation of a L'vov platform which was designed to fit in the graphite furnace tube was investigated. It was observed that by installing the platform the absorbance of thallium was enhanced by almost a factor of three (Fig. 3 ) . 0.20 0.16 a EJ 0.12 e d a 0.08 0.04 0 A I I I 200 400 600 800 1000 Te m pe ra t u rePC Fig.2 Absorbance of TI as a function of ashing temperature at an atomizing temperature of 1800 "C; A with platform; B without platformJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY NOVEMBER 1994 VOL. 9 1291 8 g 0.12 e s a 0.08 a 0.04 0 1400 1600 1800 2000 2200 2400 TemperaturePC Fig.3 Absorbance of T1 as a function of atomizing temperature at an ashing temperature of 500 "C; A with platform; B without platform Table 2 Absorbance and RSD values of the calibration standards; n= 10 Concentration of Mean absorbance standard/ng ml- (platform value) 5 0.014 (0.051) 10 0.030 (0.106) 20 0.06 1 (0.196) 30 0.089 (0.288) Table 3 Results of T1 determinations in Ni-based alloys Concentration Concentration Sample found/pg 8-l certified/pg g-' SRM 897 0.50 & 0.02 0.51 & 0.03 SRM 898 2.70 k 0.05 2.75 & 0.02 SRM 899 0.25 & 0.01 0.252 & 0.003 CRM 345 1.90 & 0.05 < 0.2 CRM 346 1.98 k 0.04 (2)* * Information value.This is attributed to the platform which delays volatilization until the gas phase is at a higher unchanging temperature. This enhancement in the absorbance of T1 would help to enable the determination of T1 in complex matrices such as nickel-based alloys at low levels rather than with atomization from the tube wall. The precision was also improved by the installation of the platform in the furnace tube. The relative standard deviation (RSD) was about 2-10% without platform and 0.2-3% after installation of the platform. The mean absorbance values and RSD values were obtained from 10 repetitive measurements of the absorbance of T1 in the same sample solution.Measurement Uncertainty The calibration curve prepared in the range 0-30ngml-I of T1 was linear (Table 2). The detection limit defined as the T1 concentration corresponding to three times the standard devi- ation of the blank is 1.0 ng m1-l. The accuracy of this method was tested by determining TI in SRM 897 898 and 898 and CRM 345 and 346. The comparison of results obtained by this method (Table 3) shows good agreement with the certified values and no evidence of bias. Microwave digestion of nickel-based alloys with H,SO HN03 and HF in a pressurized PFA vessel provides an alternative to the conventional hot-plate wet digestion pro- cedure for the determination of T1. In the presence of H2S04 and ascorbic acid matrix interference can be suppressed to the required levels.It is evident from the absorbance values obtained that the installation of a stabilized temperature platform in the graphite furnace would help in extending the analysis to further lower levels of T1. The authors thank the Director Defense Metallurgical Research Laboratory for giving permission to communicate this work. References 1 2 3 4 5 6 7 8 9 10 11 12 13 Holt T. D. and Wallace W. Int. Met. Rev. 1976 21 1. Heardridge J. B. and Nicholson R. A. Spectrochim. Acta Part B 1984 39 551. Wood D. R. and Cook R. M. Metallurgica 1963 67 109. Griepink B. Sager M. and Tolg G. Pure Appl. Chem. 1988 60 1425. Marks Y. J. Welcher G. G. and Spellman R. J. Appl. Spectrosc. 1977 31 9. Kujirai O. Kobayashi T. Ide K. and Sudo E. Talanta 1983 30 9. Welcher G. G. Kriege 0. H. and Marks J. Y. Anal. Chem. 1974 46 1227. Koirytohann S. R. Glass E. D. and Lichte F. E. Appl. Spectrosc. 1981 35 22. Fuller C. W. Anal. Chim. Acta 1976 81 199. Slavin W. Carnrick G. R. and Manning D. C. Anal. Chim. Acta 1982 138 103. Shan X. Ni Z. and Zhang L. Talanta 1984 31 150. Manning D. C. and Slavin W. Spectrochim. Acta Part B 1988 43 1157. Kujirai O. Kobayashi T. and Sudo E. Fresenius' 2. Anal. Chem. 1979 297 398. Paper 4/02 794 I Received May I 1 I994 Accepted July 18 1994