was developed for the determination of thallium at the ng/ml

Transcription

1 Enhancement of Thallium Response by Flow Injection Hydride Generation AAS Using Palladium and Rhodamine B Zhu Daan and Xu Shukun* Research Center for Analytical Sciences, Northeastern University Box 332, Shenyang, P.R. China INTRODUCTION Flow injection hydride generation atomic absorption spectrometry (FI-HGAAS) offers significant advantages, including a reduction of more than 90% in sample and reagent consumption, 2 3-fold higher sampling frequencies, better precision, enhanced selectivity, and easy automated operation. Because of these benefits, the FI-HGAAS technique has been extended to the hydride-forming elements such as Se, Te, As, Sb, Bi, Pb, Ge, Sn, In, and Tl using a quartz tube atomizer or graphite furnace atomizer with in situ trapping preconcentration (1). But reports on the hydride generation of thallium are few. In 1984, Yan et al. (2) developed a batch hydride generation AAS method for the determination of thallium, where a peak height characteristic sensitivity of 0.12 mg/l was achieved. They also pointed out that tellurium (VI) has a positive effect on sensitivity. Recently, Liao et al. (3) developed a hydride generation and in situ trapping (in the graphite tube) system for the ETAAS determination of thallium using tellurium as the enhancement reagent at 60ºC. The peak height characteristic mass was 0.92 ng and the detection limit at 1 ng. Ebdon et al. (4) developed a continuousflow method for the HGAAS determination of thallium and reported that excess noise may be eliminated at 0ºC. They reported a peak height characteristic concentration of 4 ng/ml for thallium. Unfortunately, none of the studies *Corresponding author. xusk@mail.neu.edu.cn ABSTRACT A flow injection hydride generation atomic absorption spectrometric (FI-HGAAS) method was developed for the determination of thallium at the ng ml -1 levels. Palladium, an effective enhancement reagent, was added to the acidified sample solution. The presence of micro amounts of Rhodamine B increased the hydride generation efficiency further. The acidified sample solution containing palladium and Rhodamine B merged with the aqueous sodium tetrahydroborate solution at room temperature. The generated gaseous phase was separated in a gasexpanded gas-liquid separator and was led directly into a quartz tube atomizer at 1000ºC via the Ar carrier gas for detection by AAS. Chemical and flow injection (FI) parameters for the hydride generation of thallium were optimized, including the concentration of the enhancement reagents, acidity of sample solution and carrier solution, NaBH 4 concentration, reaction coil length, sample volume, carrier gas flow rate, and flow rate of the carrier as well as the reagent stream. With the optimized conditions, the precision was 0.9% RSD (200 ng/ml, n=11), with a sampling frequency of 120/h. A detection limit of 3.4 ng/ml Tl (3 ) was obtained with a sample volume of 500 µl. The interferences from coexisting elements were studied. The recoveries of thallium in water samples by spiking the samples with 100 ng/ml of thallium were in the % range. reported a successfull HGAAS determination of thallium at the ng/ml levels in real samples. There was no discussion about interferences from co-existing heavy metals, except about the enhancement effects from Te, Cu, Co, Se, and Ni as well as the suppressive effects of Ge and Cd (1, 3). To the best of our knowledge, the present paper is first to report a FI-HGAAS method for the determination of trace amounts of thallium at the ng/ml levels using palladium as the enhancement reagent. EXPERIMENTAL Instrumentation A PerkinElmer Model 2100 atomic absorption spectrometer with a deuterium lamp background corrector was used in this study, combined with a Model FIAS -200 flow injection system. A thallium hollow cathode lamp was operated at 9 ma. A wavelength of nm was used with a bandpass of 0.7 nm. Peak height absorbance was used for evaluating the results throughout this work. The time-resolved absorbance signals for thallium were recorded using high-resolution graphics on a high-resolution screen and printed using an Epson EX-800 printer (Epson Japan). The rotations of the two multichannel peristaltic pumps and the actuation of the valve were programmed and automatically controlled by the computer software. The gas-liquid separator (GLS, Zhao-Fa, Shenyang), similar to the transparent plastic separator, was filled to one-third with glass beads of 3 mm diameter as described by Fang (1). Highpurity argon was used as the carrier gas. Atomic Spectroscopy 136

2 Reagents and Standard Solutions All reagents were of analytical grade and deionized water was used throughout. To keep the reductant stable, a 0.3% (m/v) solution of sodium tetrahydroborate was prepared daily in deionized water containing sodium hydroxide (0.3%, m/v). 1,10-phenanthroline (0.4%, m/v, Shenyang Chemical Co. Shenyang, China) was prepared by dissolving 0.2 g of 1,10-phenanthroline in 50 ml deionized water. Rhodamine B (0.5%, m/v, Shenyang Chemical Co. Shenyang, China) was prepared by dissolving 0.25 g of Rhodamine B in 50 ml deionized water. Nitric acid (63%, w/v, Jinxi Chemical Co., Jinxi, China) was of pure reagent grade. A standard solution of 10 mg/l thallium was prepared by dilution of the stock standard solutions (1 g/l) with 2 mol/l HNO 3. A series of working standard solutions containing 50, 100, 150, 200, and 250 µg/l thallium was prepared by stepwise dilution of the 10-mg/L standard solution and made to contain 0.2 mol/l HNO 3 and 400 µg/l Pd. For the system using Rhodamine B, the solution was made to contain 0.1 mol/l HNO 3, 400 µg/l PD, and % Rhodamine B. Palladium stock solution (1g/L) was used throughout this work. Procedure The FI operating program and the operating parameters are listed in Table I, and the manifold used for HGAAS is shown in Figure 1. Thirty seconds were needed for the analysis of one sample, which included data processing and printout time. One cycle of an analysis includes three steps: First, four seconds to wash the sample loops using the standard or sample solution and to load the sample in the TABLE I Flow Injection Operation Program for the FI-CVGAAS System Step Time Valve Pum1 Pum2 Function (S) Position (ml/min) (ml/min) 1 4 Fill R:2 S:4.4 Prefill, change C:6 W:25 sample 2 12 Fill R:1.2 S:4.4 Sample filling C:3.6 W: Injection R:3 W:25 Sample injection C:9.2 R: NaBH 4 ; C: Carrier; S: Sample; W: Waste Fig.1. Flow injection manifold for the HGAAS determination of Thallium. (a) Sample filling; (b) Sample injection. P1, P2, peristaltic pumps; AAS, atomic absorption spectrometer; GLS, gas-liquid separator; S, sample; R, NaBH4; C, carrier; W1, W2, water; SL, sampling loop; Ar, argon carrier gas. pre-fill step. Second, in the loading step (Figure 1a), the standard or sample solution was filled into the sample loop. Third, in the injection step (Figure 1b), ending the sampling, the valve rotated automatically to injection position. Pump 2 was speeded up (as in Figure 1). The sample or standard solution stored in the sampling loop was carried out by the water carrier and merged with the reductant. The chemical reaction took place immediately to form hydride vapor. The vapor was separated from the waste in the gas-liquid separator (GLS) and the generated thallium hydride vapor was carried into the quartz tube (AAS) at 1000ºC by the argon gas for measurement. Sample Pretreatment The water sample was filtered and acidified to 0.1 mol/l HNO 3 and made to contain 400 µg/l Pd and % Rhodamine B. Method Development The main aim of this work was to select an appropriate enhancement reagent for the hydride generation of thallium and to establish a method for FI-HGAAS analysis. The relationship between absorbance of the signals and the 137

3 enhancement reagent concentration, sample acidity, reductant concentration, and reaction coil length was investigated using the FI technique. Owing to the within-day variations in the hydride generation system, a universal-type approach was used for optimization. The flow injection parameters and chemical reaction conditions were optimized with the vapor generation efficiency as the main figure of merit and with simultaneous consideration on precision and interferences from co-existing elements. The median values of the parameter ranges used in previous FI vapor generation systems were taken into account and the parameters gradually adjusted to close-tooptimum values for final evaluation. RESULTS AND DISCUSSION Temperature of the Atomizer Preliminary experimental results showed that the temperature of the quartz tube atomizer had very good effects on the detection signal. High temperature can provide enough energy to atomize the thallium hydride, which leads to a linear increase in the temperature range from 700ºC to 1000ºC. Because of software limitations (maximum temperatures above 1000ºC are not allowed), 999ºC was used for the atomization step for further experiments. Selection of Enhancement Reagent The selection of an appropriate enhancement reagent to enhance the hydride generation efficiency of thallium is very important, because of the reportedly low sensitivity of HGAAS for thallium determination (2). In Reference 3, tellurium was used as the enhancement reagent for thallium hydride generation and its in situ trapping in a graphite furnace. In our work, an enhancement effect of tellurium on the signal of thallium hydride was observed, but only for thallium concentrations at Fig. 2. Effects of palladium, gold, and platinum on peak height absorbance of 200 mg/l thallium, other conditions are the same as in Figure 1. the mg L -1 level. It is still difficult to obtain low detection limits for ultratrace thallium in environmental samples. Even using ETAAS detection for a 200-µL sample (3), the peak height characteristic mass was 0.92 ng (4.6 ng/ml) and the detection limit at 1 ng (5 ng/ml). The sensitivity was not enough for the determination of trace or ultra-trace amounts of thallium. This also means that when using a quartz tube as the atomizer, the selection of a more efficient enhancement reagent is required. In this work, it was found that Au, Pt, and Pd in the sample solution at a certain concentration could enhance the signal of thallium at the µg/l level. The enhancement effects of these three metal ions at room temperature are shown in Figure 2. Of the three metal ions, palladium provided the best performance for the hydride generation of thallium and since it is the most economical, it was chosen as the enhancement reagent. As shown in Figure 2, there is a plateau for the concentration of palladium from 300 ng/ml to 450 ng/ml; therefore, 400 ng/ml of Pd in sample solution was used for further experiments. Compared to the data from Reference 2 without and with the presence of tellurium, the sensitivity was increased by a factor of 220 and 28, respectively. Compared with the signal of 10 mg/l thallium obtained using this FI system without the presence of palladium and with the presence of 4 mg/l tellurium, the sensitivity was increased by a factor of 40. Co-enhancement Effect of Rhodamine B The alkaline dye Rhodamine B is extensively used in the spectrophotometric determination of thallium (5). The experiments also showed that Rhodamine B enhances the peak height of thallium in the presence of palladium. The absorbance of peak height for thallium was enhanced by 40% with % of Rhodamine B in the sample solution using the same FI parameters and reagent concentration, except that the concentration of HNO 3 was from 0.2 mol /L down to 0.1 mol/l. The optimized concentration of HNO 3 in the reaction medium was 138

4 lower for best enhancement effect, which may be due to the fact that Rhodamin B is unstable at higher acidity. The relationships between absorbance, concentration of HNO 3 and Rhodamine B are shown in Figure 3. Optimization of FI and Chemical Parameters Acidity of Sample and Carrier Solution The hydride generation efficiency of thallium depends strongly on the acidity of the reaction medium and the acid species. The effect of various acidity conditions on the FI-HGAAS determination of thallium has been studied in this work. An attempt was made to investigate the effects of HCl, HNO 3 and H 2 SO 4 acidity from 0.05 mol/l to 2 mol/l. The curve shown in Figure 4 was plotted from the absorbance obtained under different acidity conditions without Rhodamine B but with water as the carrier. It can be seen that low acidity conditions resulted in higher sensitivity for thallium hydride generation. But when hydrochloric acid was used, the signal was unstable, and a significant decrease in peak height was observed with a concentration of hydrochloric acid up to or beyond 0.3 mol/l. The reduction in sensitivity for high chloride ions was at least partially due to the formation of the Tl - Cl complex, which eliminates any reaction of the thallium ion with sodium tetrahydroborate to form thallium hydride. With the H 2 SO 4 medium, SO 2-4 can form a yellowish precipitation with the thallium ion and thus may not be the best choice either. Using nitric acid as the sample medium resulted in a plateau from 0.1 mol/l to 0.3 mol/l. The high concentration of nitric acid has negative effects on the determination due to the dilution effect of large amounts of hydrogen generated in the reaction process. Fig. 3. Relationships of Rhodamine B, HNO 3 and absorbance. All other conditions are the same as in Figure 1. Fig. 4. Effects of concentration of HNO 3, H 2 SO 4 and HCl in sample solution on the absorbance of 200 mg/l thallium. All other conditions are the same as in Figure

5 In the selection of a carrier solution, water was compared to nitric acid having the same acidity as the sample medium. The experimental results showed that the sensitivity of thallium for the two kinds of carrier solutions was almost the same. For deionized water as the carrier solution, the peak recordings were much smoother than for 0.2 mol/l HNO 3 as the carrier solution. To obtain best precision and to use the least amount of nitric acid, deionized water was chosen as the carrier solution throughout this work. Effects of Concentration of Sodium Tetrahydroborate The effects of tetrahydroborate concentration (in 0.3% NaOH and without the use of Rhodamine B) on peak height absorbance for 100 µg/l thallium were investigated. Based on the experimental results for peak height absorbance and NaBH 4 concentration, 0.3% NaBH 4 in 0.3% NaOH solution was chosen to obtain best sensitivity. Effects of Reaction Coil Length and Sample Volume The effects of a reaction coil length from 5 25 cm on peak height absorbance were investigated. The shortest length of 5 cm corresponded to the length connecting the reagent merging point and the gas-liquid separator. For the 0.2 mol/l HNO 3 sample medium with deionized water as the carrier solution and the experimental conditions used as shown in Figure 1, the relationship between absorbance and reaction coil length are shown in Figure 5. Because hydride generation is an instantaneous reaction, shorter reaction times are beneficial to suppress interfering reactions that are slower than the main reaction. A longer reaction coil would have a negative effect on the determination. A 10-cm reaction coil length is the optimum length for highest sensitivity. On the other hand, thallium hydride is unstable and often decomposes, resulting in Fig. 5. Effect of reaction coil length on absorbance of 200 mg/l thallium, other conditions are the same as in Figure 1. the formed thallium being adsorbed onto the reaction coil. In addition, after a longer reaction time, the color of the reaction coil turns blackish. For these reasons, a shorter reaction coil of 10 cm was used to avoid loss of thallium. Since there was no evident difference in sensitivity for a sample volume beyond 500 µl, a 500-µL sample volume was chosen as the optimum volume. Flow Rate of Sample, Reagent, and Carrier Gas Preliminary tests showed that best sensitivity was obtained with a 3:1 ratio of carrier and reductant flow rate. Based on that ratio, the effects of flow rate of the carrier and reagent on peak height absorbance were investigated. It was found that a carrier flow rate of 9.2 ml/min was optimum, which was used for further experiments. In this work, argon was used as the carrier gas to transport the thallium hydrides into the atomizer. The flow rate and flow stability of the carrier gas usually has a significant effect on the sensitivity and repeatability of the method. Experimental results showed that lower gas flow rates (100 ml/min) gave higher sensitivity, but the peak of the signal tailed since the thallium hydride vapor flowing into the quartz tube cannot be expelled in time. Higher carrier gas flow rates (>140 ml/min) decreased the sensitivity sharply, which may be due to the dilution effect. Considering the effects on sensitivity and peak shape, a carrier gas flow rate of 120 ml/min was chosen. Interferences The possible co-existing metal ions in water were tested for the interference study. Of the potential interfering ions, Cd, Hg, Cu, and Pb interfere in the process of generating thallium hydride. Because the optimum acidity for the generation of thallium hydride is about 0.2 mol/l, which is near the vapor generation of cadmium (6), cadmium was the most seriously interfering element among the hydride-forming species studied. Other potential 140

6 TABLE II Interference of Co-existing Ions in the Presence of 400 ng/ml Pd Co-existing ions and Nitric acid Interference Change in In Addition Concentration( mg/l) (mol/l) (%) Sensitivity (%) Cd Cd Cd Cd ,10-Phenanthroline Hg Hg ,10-Phenanthroline Hg ,10-Phenanthroline Cu Se(VI) Se(IV) Mn(II) Ni Zn Co Cr Sn Bi As Pb Pb Pb Ion exchange interfering elements such as Co, Ni, and Bi did not interfere in the hydride generation. In this study we attempted to eliminate the interference by increasing the acidity of the sample solution or by using a masking reagent. When the sample acidity was raised to 1 mol/l, interference of the Cd and Hg ions was almost eliminated, but the sensitivity was lost up to 39% and 20%, respectively. The higher tolerance of transition metals at higher acid concentrations is presumably due to a reduction in the local formation of the interference precipitate associated with these metals (1). When 0.002% 1,10-phenanthroline is added to the sample solution, the tolerable concentration of the interferents Cd, Hg, and Cu can be increased to 1.0, 0.5, and 0.5 mg/l, respectively, in 0.2 mol/l HNO 3. The interference from lead was not easily eliminated with the masking reagent due to the characteristic of the thallium ion, which is similar to the lead ion. By manually packing a column with sulfydryl resin and keeping the sample flow rate at 14.2 ml/min can eliminate the interference of lead. The interference resultsof co-existing ions are listed in Table II. 141

7 TABLE III Peak Height Characteristic Performance Data of the Method Medium DL (3 ) RSD Calibration curve Linear HNO 3 Pd ( µg/l) (n=11, %) range (mol/l) ( µg/l) (µg/l) A=3.49x10-4 C ~250 r= * A=4.98x0-4 C ~400 * With % (m/v ) Rhodamine B. Analytical Performance Using the optimum conditions and the FI-HGAAS system as described above, the peak height characteristic data for the typical analytical performances are listed in Table III. River water, tap water, and mineral water samples were analyzed using this method. The thallium in these water samples was not detectable because the levels were too low. The recoveries obtained by spiking the samples with 100 µg/l thallium were 109%, 106%, and 101%, respectively. r=0.999 CONCLUSION An efficient and automated FI- HGAAS method was developed for the determination of trace amounts of thallium. Micro amounts of palladium (or Pt or Au) in sample solution are an efficient enhancement reagent for thallium hydride generation. Rhodamine B showed an enhancing effect on the reaction. By using palladium and Rhodamine B as the enhancement reagent, the FI-HGAAS determination of thallium at mg/l levels, using a quartz tube atomizer at 1000 o C, will result in low detection limits and good precision. The study on the mechanism of the enhancement effects of palladium and Rhodamine B are ongoing. ACKNOWLEDGMENT The authors are grateful to Professor Zhaolun Fang for useful discussions, to the Natural Science Foundations of China for financial support, and to PerkinElmer Bodenseewerk, Germany, for providing the atomic absorption spectrometer and for partial financial support. REFERENCES 1. Z.-L. Fang, Flow Injection Separation and Preconcentration, VCH, Weiheim, Germany (1993). 2. D. Yan, Z. Yan, G. Cheng, and A.-M. Li, Talanta 31,133 (1984). 3. Y.-P. Liao, G.Chen, D. Yan, and A.-M. Li, Anal. Chim. Acta. 360, 209 (1998). 4. L. Ebdon, P.Goodall, S.J.Hill, P.Stockwell, and K.C.Thompson, J.Anal.At. Spectrom. 10, 317 (1995). 5. H. Luo, S.Liu, and Z.Liu, Fenxi- Huaxue 21, 1179 (1993). 6. M.-Y. Liu and S.-K. Xu, At. Spectrosc. 18, 195 (1997). Received March 30,