1. Introduction

Tellurium can be determined with high sensitivity and selectivity by atomic spectrometry after generation of its covalenthydride through reaction with sodium borohydride. This strategy provides an excellent way for matrix isolation and increases the element intake inside the atomizer (1). Using hydride generation, tellurium has been determined by atomic absorption spectrometry (2, 3), graphite furnace atomic absorption spectrometry (4, 5), inductively coupled plasma atomic emission (6, 7) and inductively coupled plasma mass spectrometry (8-10). In recent years, the development of simple atomic fluorescence spectrometers has provided a suitable tool for tellurium determination (11-20). Tellurium has been determined by HG-AFS in metals and alloys (11-15), water (16, 17), urine (18), milk (19) and geological samples (20). From studies reported in the literature for Te determination in waters by HGAFS it can be noticed that a high limit of detection (80 g/L) was found in spite of the use of a preconcentration through cation exchange (16). On the other hand, total Te can be determined after reduction of Te(VI) with citric/tartaric acids, obtaining a limit of detection of 0.1 g/L in the diluted solution (17).
The aim of this paper is to develop a simple method for the direct determination of trace amounts of tellurium, by continuous hydride generation and atomic fluorescence spectrometry, in environmental samples. The sensitivity and limit of detection of HGAFS were improved by selecting the appropriated operational conditions. On the other hand, for total tellurium determination, Te(VI) was reduced to Te(IV) by using solid sodium bromide as reductant in order to avoid sample dilution. The developed method was evaluated by the analysis of a reference material (SGR-1).

2. Experimental

2.1. Apparatus An Unicam VP-90 continuous-flow vapour system equipped with a B-type gasliquid separator (Cambridge, UK) and a Permapure drier tube from PS Analytical (Sevenoaks, Kent, UK) was used for hydride generation. An Excalibur atomic fluorescence detector PSA 10033 from PS Analytical, equipped with a tellurium boosted discharge hollow cathode lamp (BDHCL) from superlamp Photron (Victoria, Australia), an hydrogen diffusion flame, a series of lenses and a specific filter and a solar blind photomultiplier, was employed for tellurium fluorescence measurements at the 214.3 nm resonance wavelength. Flexible polyvinyl-chloride peristaltic pump tubes of 1.85 mm and 1.14 mm id

Determination of tellurium by continuous hydride
were employed to transport HCl and NaBH4 respectively. All tubes employed to construct the manifold were from 0.8 mm id. polytetrafluoroethylene (PTFE). For sample digestion, a domestic LG microwave oven, model MS 1905-C (Manchester, U.K.), with a magnetron frequency of 2450 MHz and a maximum exit power of 750 W was used. 2.2. Reagents and standards All the reagents used were of the highest purity available, and reagent-grade water, obtained from a Milli-Q water purification system Millipore (Bedford, MA) with a resistivity of 18 M/cm, was used for preparation of standards and samples. A 0.1221 g/L Te(IV) standard stock solution was prepared by dissolving Na2TeO3 Aldrich Chem. Co, (Milwaky, WI) in deionized water. A 0.1228 g/L Te(VI) stock standard solution was prepared in deionized water by dissolving H2TeO6 Fluka (Buchs, Switzerland). The calibration solutions were prepared daily by dilution of stock standard ones with 5.0 mol/L HCl which was prepared from 37% d= 1.18 g/mL HCl Fluka (Buchs, Switzerland). For total tellurium determination 1.0 g of sodium bromide Panreac (Barcelona, Spain) was dissolved, in acidified samples and standards, and employed as reductant. Sodium tetrahydroborate was prepared from Fluka (Buchs, Switzerland) and dissolved in 0.5% (m/v) NaOH Probus (Barcelona, Spain). This solution was prepared daily and filtered before use with a nylon 0.45 m pore-size mesh from Lida (Kenosha, WI), and was used as the reducing reagent and as H2 source for the diffusion flame. Argon C-45 (purity 99.995%) and synthetic air C-45, obtained from Carburos Metalicos (Barcelona, Spain), were used to transport the hydride generated to the atomiser flame and to dry the vapour phase in the Perma Pure drier tube respectively.
2.3. Sample conditioning Mineral and drinking water samples were obtained from the local market in Morocco (Sidi Ali, Sidi Harazem) and Spain (Bejis) and tap waters taken from several towns near Valencia (Burjassot, Els Pous). Samples, collected in polyethylene bottles, were acidified with concentrated HCl to obtain a pH near 1.2 to minimize adsorption of trace elements on the bottle walls and microorganisms proliferation and were stored in the dark. Sediment reference material SGR-1 was obtained from U.S. Geological Survey. Table 1 Operating conditions for HG-AFS determination of tellurium Parameter Spectrometer operating conditions Resonance wavelength/nm Bandpass/nm Primary current/mA Boost current/mA Gain Potentiometer Hydride generation conditions Reagent concentration HCl / mol/L NaBH4 /% (m/v) NaBr /% (m/v) Reagent flow rate Sample/HCl / mL/min NaBH4 / mL/min Argon / mL/min Air / L/min Reaction coil length /cm 9.8 3.3.5.0 1.214.3 0.5 15.0 17.5 6.0 6.0

Mixing valve

Figure 1. Manifold employed for the on-line HG-AFS determination of tellurium. 2.4. Analytical procedure Table 1 summarises the instrumental and experimental conditions for tellurium determination, and Figure1 shows the manifold employed in this study. For Te(IV) determination, a water sample volume of 12.5 mL was acidified with concentrated HCl and diluted to 25 mL with deionized water obtaining a 5.0 mol/L HCl solution. Standard solutions, reagent blank and samples, transported at a flow rate of 9.8 mL/min, were merged with a 1.2% (m/v) NaBH 4 solution, in 0.5% (m/v) NaOH, with a flow rate of 3.1 mL/min, to produce the tellurium hydride. The hydride was generated in a reaction coil of 100 cm, and the reagents mixture passed to the gas-liquid separator, being transported the Te hydride to the detector by an argon flow rate of 400 mL/min. For the total tellurium determination in water, the aforementioned procedure was used after a previous reduction of the samples. 1.0 g of NaBr was added to the acidified samples and then the solution heated at 60C during 10 minutes inside a water bath. All solutions were further cooled to room temperature before reaction with NaBH4 for hydride generation. 2.5. Microwave-assisted leaching procedure For quantitative leaching of tellurium from sediments, 200 mg of sediment were weighed and introduced into a 115 mL PTFE reactor. A volume of 3 mL of different concentrations of HCl (from 1 to 6 mol/L) was added to the sample and the reactor was closed hermetically and irradiated in a microwave oven, operated at 750 W, during three steps of 1 minute. After that, the reactor was let to cool and the partially dissolved sample was dispersed to 25 mL with water. 5 mL of the supernatant solution were diluted to 25 mL with HCl to obtain a final concentration of 5 mol/L and the Te determination was carried out as indicated before. Acid solutions were employed as standards for the analysis of leached samples and results were compared with the standard addition procedure to prevent possible matrix interference.
3. Results and discussion
3.1. Effect of hydrochloric acid and sodium tetrahydroborate concentrations on Te fluorescence The study of the HCl concentration effect on the atomic fluorescence signal for Te(IV), showed in Figure 2, indicated that signals obtained are strongly influenced by HCl concentration in the range between 0.5

Atomic fluorescence intensity

Te (IV) 10 ng/mL

and 7 mol/L. However, HCl did not influence blank signals. The increase of HCl increases Te fluorescence signal up to reach a maximum value for 6.0 mol/L HCl, and it was observed a strong reduction of Te(IV) fluorescence for a 6.5 mol/L HCl. So a HCl concentration of 5.0 mol/L was selected. Moreover, a 10 ng mL Te(VI) standard solutions gives the same fluorescence signal as the blank in the HCl range studied. Previous experiments demonstrated that a NaBH4 concentration higher than 1% (m/v) was required to maintain the hydrogen flame. Figure 3 summarizes studies carried out about the influence of NaBH4 concentration, on the Te(IV) hydride fluorescence using a 3.0 mL/min NaBH4 flow. As can be seen, it was observed a dramatic influence of the NaBH4 concentration on the Te(IV) and blank fluorescence signals, being found at 1.8% NaBH4 the maximum fluorescence. However, to obtain a good signal to noise relationship, and in order to obtain the best analytical performance, a NaBH4 concentration of 1.2% (m/v) was selected for Te(IV) determination. 3.2. Effect of carrier and sodium tetrahydroborate flow rates on Te hydride generation The effect of the sample flow rate on the tellurium hydride generation was studied between 4.0 to 12.5 mL/min. Results shown in Figure 4 indicate that tellurium fluorescence signal increases on increasing the flow rate up to reach a maximum value for 10.5 mL/min whereas the blank signal remains constant in all the range studied. Furthermore, the use of higher flow rates causes an increasing unstability of signals and was not recommended. So, to obtain the best signal to noise relationship, a 9.8 mL/min flow rate was considered as the most appropriate.
HCl concentration (mol/L)
Figure 2. Effect of HCl concentration on the fluorescence of 10 ng/mL Te(IV).

100 75

NaBH 4 concentration % (m/v)
Figure 3. Effect of NaBH4 concentration on the fluorescence of 10 ng/mL Te(IV). Hydride generation conditions: Standard/sample carrier flow 7.5 mL/min, HCl 5.0 mol/L, NaBH4 flow 3.0 mL/min, Ar flow 300 mL/min, reaction coil length 100 cm. Further studies on the efficiency of the hydride generation for different flow rates of NaBH4, between 2.0 and 5.5 mL/min, were carried out. The flame was extinguished when flow rates lower than 2.0 mL/min were employed. The maximum signal was obtained for a 3.1 mL/min flow rate, being observed that the use of higher than 3.1 mL/min NaBH4 values only increases the unstability of blank signals.

3.3. Effect of reaction coil length on fluorescence signals The reaction coil length controls the hydride generation time, for a fixed flow rate, and it can affect the fluorescence signal depending on the speed of tellurium hydride formation. For a fixed total carrier flow of 11.9 mL/min it was evaluated the effect of the reaction coil length, from 50 to 300 cm using a PTFE tube of 0.8 mm of internal diameter. As can be seen in Figure 5, the increase of coil length from 50 to 100 cm increases slightly the tellurium hydride signal. On the other hand, coils longer than 100 cm cause signal decreasing, possibly due to the increase of dispersion and thus, a 100 cm coil was selected. Moreover, blank signal do not varies as a function of the coil length changes. 3.4. Effect of Argon flow rate on fluorescence signals The increase of the Ar flow rate increases the tellurium fluorescence signal up to reach a maximum value for a value of 400 mL/min. The fluorescence of the blank remains practically independent on the Ar flow. On the other hand, the use of an Ar flow higher than 500 mL/min extinguished the flame, and when Ar flow rates employed were lower than 200 mL/min the flame can not be ignited. According to that an Ar flow of 400 mL/min was chosen as the most adequate for tellurium determination. 3.5. Reduction of Te(VI) It is important to note that, in the conditions previously selected for Te(IV) hydride generation, Te(VI) solutions do not give any fluorescence signal. Because of that, the effect of a previous reduction step on the fluorescence intensity of tellurium was investigated in order to obtain as higher as possible signals from Te(VI). Some papers have reported the use of concentrated HCl (21), 4 M HCl (18), 3 M HCl (22) 20% HCl (23) or 0.5 M HCl (20) for the reduction of Te(VI) to the Te(IV). Other authors employed KI and ascorbic acid (9,
Standard/sample flow rate (mL/min)
Figura 4. Effect of the standard/sample carrier flow on the fluorescence of 10 ng/mL Te(IV). Hydride generation conditions: HCl 5.5 mol/L, NaBH4 3% (m/v) on NaOH 0.5% (m/v), NaBH4 flow 1.4 mL/min, Ar flow 400 mL/min, reaction coil length 150 cm.
Reaction coil length (cm)
Figura 5. Effect of reaction coil length on the fluorescence of 10 ng/mL Te(IV). Hydride generation conditions: NaBH4 flow 1.4 mL/min, HCl 5.5 mol/L, NaBH% (m/v) on NaOH 0.5% (m/v), Standard/sample carrier flow 10.5 mL/m, Ar flow 400 mL/min. 24) as reductant reagents before hydride generation. However, experiments carried out with 5 M HCl heating at 50C and 70C during 1 hour, do not provide evidences on reduction of Te(VI) to Te(IV).

Table 2 Effect of temperature, heating time and NaBr concentration on the hydride fluorescence signal ratio obtained from Te(VI) and Te(IV). Fluorescence ratio (in%) between signals found from Te(VI) and Te(IV) Time (min) 60 50C 2% NaBr 50C 4% NaBr 60C 2% NaBr 60C 3% NaBr 60C 4% NaBr
Sodium bromide is an efficient reductant for Te(VI) as can be seen in Table 2. This Table shows the ratio between Te(VI) and Te(IV) fluorescence signals from solutions of the same concentration that undergoes the same reduction procedure, that is, same temperature, heating time and NaBr concentration. Nevertheless, the reduction of Te(VI) to Te(IV) depends on both, NaBr concentration and temperature, being observed that using a 4% NaBr concentration and 50C, 20 minutes are enough to obtain comparable results between fluorescence of tellurium hydride obtained from Te(IV) and Te(VI) solutions. On the other hand, the increase of temperature till 60C provides a quantitative reduction of Te(VI) to Te(IV) in 10 minutes, thus offering the same fluorescence for solutions of 5 g/L tellurium obtained from Te(IV) and Te(VI). So, the aforementioned conditions were selected in order to obtain accurate results in total tellurium determination in aqueous or dissolved samples.
3.6. Analytical figures of merit Typical calibration lines obtained for the two species considered, when standards were prepared without any previous reduction for Te(IV) and with the previous reduction step for Te(VI), correspond to equations: IF = 6.79 (Te(IV)) - 0.719; r = 0.994 IF = 7.10 (Te(VI)) + 0.453; r = 0.998 As it has been indicated before, without a previous reduction step the fluorescence signal obtained from Te(VI) was practically undetectable. However, as can be seen from equations, the slope of calibration lines obtained from both, Te(IV) and Te(VI) after reaction with NaBr, were statistically comparable. Using the operating conditions indicated in Table 1, the calibration graphs were linear up to a concentration of 20 g /L, but could be enlarged if the sensitivity of the instrument was reduced. The detection limit,
established as 3sb/slope, where sb is the standard deviation of 10 blank measurements, were in the range between 0.4 and 0.7 g/L for Te(IV) and from 0.4 to 0.9 g/L for Te(VI) determination. The coefficient of variation, evaluated from the relative standard deviation of 3 independent determinations of a same sample in the whole range of calibration varied between 1 and 16% for Te(IV) and from 0.2 to 13% for Te(VI). The accuracy of the method was evaluated by recovery experiments carried out using water samples spiked with known concentration of Te(IV) and Te(VI) at different concentration levels from 0.4 to 2.0 g/L and from 1.0 to 3.0 g/L respectively. As can be seen in Table 3, the analytical recovery values achieved were closed to 100% in all the experiments. 3.7. Determination of total tellurium in sediments Additional experiments were carried out to confirm the accuracy of the developed procedure, and it was determined the total tellurium concentration in a certified sample (SGR-1 USGS).

To get the quantitative tellurium leaching from sediment, a microwave assisted treatment at 750 W was employed. From studies carried out, a three minutes digestion with HCl 6 mol/L, in 1 minute steps, was considered adequate. Results obtained were 0.250 0.003 ng/g in front of the 0.248 ng/g proposed value in the certified sample. On the other hand, to reduce the pressure problems produced from the high HCl concentration employed, it was studied the leaching effectiveness of HCl concentration, maintaining constant the heating steps. Figure 6 shows the obtained results and, as can be see, a 3 mol/l concentration of HCl is enough to achieve the quantitative leaching of tellurium from sediment. Additional experiments, in which slopes of external calibration and standard addition calibration were compared, for sediment analysis, evidenced that the HCl concentration and sample matrix do not affect the fluorescence measurements, obtaining slope ratios between 94 and 102%.

4. Conclusin

Hydride generation atomic fluorescence provides a sensitive methodology for direct determination of Te(IV) and total tellurium at g/L levels. The LOD obtained,
Table 3 a b Analytical recovery of Te (IV) and Te (VI) in mineral and drinking waters at different levels of concentration Water I II III IV V VI
Te(IV) Added (ng/mL) 0.4 0.8 0.8 0.4 2.0 1.6 R (%) 103 6
Te(VI) Added (ng/mL) 2.0 2.0 1.0 1.0 3.0 3.0 (R%) 100 1
Mineral waters: I, Bejis; II, Sidi Ali; III, Sidi Harazem. Tap waters: IV, Pous; V, Burjassot 1; VI, Burjassot 2. R (%): Recovery values obtained from three replicates standard deviation.

References

% Leaching
HCl concentration (mol/L )
HORLICK G. Anal Chem 56: 278R-292R, 1984. KOREZ A., EROGLU A.E., VOLKAN M., ATAMAN O.Y. J Anal At Spectrom 15: 1599-1605, 2000. HALL G.E.M., MACLAURIN A.I., PELCHAT J.C., GAUTHIER G. Chem Geol 137: 79-89, 1997. YOON B.M., SHIM S.C., PYUN H.C., LEE D.S. Anal Sci 6: 561-566, 1990. LIAO Y.P., HAUG H.O. Microchem J 56: 247-258, 1997. RIGBY C., BRINDLE I.D. J Anal At Spectrom 14: 253-258, 1999. WICKSTROM T., LUND W., BYE R. Analyst 120: 2695-2698, 1995. MORROW A., WILTSHIRE G., HURSTHOUSE A. At Spectrosc 18: 23-28, 1997. HALL G.E.M., PELCHAT J.C. J Anal A Spectrom 12: 97-102, 1997. HALL G.E.M., PELCHAT J.C. J Anal At Spectrom 12: 103-106, 1997. DULIVO A., MARCUCCI K., BRAMANTI E., LAMPUGNANI L., ZAMBONI R. Spectrochim Acta 55-B: 1325-1336, 2000. MARCUCCI K., ZAMBONI R., DULIVO A. Spectrochim Acta 56-B: 393-407, 2001. FENG X.J., FU B. Anal Chim Acta 371: 109-113, 1998. JIA J.D. Lihua Jianyan Huaxue Fence 37: 104-106, 2001. GUO D.J., LI L.S., WANG G.M. Guangpuxue Yu Guangpu Fenxi 18: 719-723, 1998. TANG Z.Y., ZHANG S.Y., QIO H.O., WANG G.Q., JIN Z.X. Fenxi Shiyanshi 16: 21-24, 1997. GUO X.W., GUO X.M. Guangpuxue Yu Guangpu Fenxi 16: 88-92, 1996. LU Y.K., SUN H.W. Fenxi Ceshi Xuebao 19: 24-27, 2000.

4. 5. 6. 7. 8. 9. 10. 11.
Figura 6. Effect of HCl concentration on leaching percentage of total tellurium from sediment sample during microwave-assisted sample treatment (n=3). 0.4 to 0.7 ng/mL, are of the same order or better than those found in the literature, 0,1 ng/mL (17) and 80 ng/mL (16). However, the concentration of Te(IV) and Te(VI) present in natural waters make necessary the use of a preconcentration step before the determination of these species in waters. The use of a prereduction step, based on the use of solid NaBr, provides a way to do the determination of Te(IV) and total tellurium, being confirmed by the recovery studies carried out on spiked water samples. On the other hand, the developed procedure for the determination of tellurium by HG-AFS has been validated analyzing a sediment that has a Te concentration level higher than that found in natural and mineral waters.

12. 13. 14. 15.

Acknowledgments
Authors acknowledge the financial support of the Spanish DGES project PB96-0779, and the financial support of the Conselleria de Cultura, Educaci y Cincia de la Generalitat Valenciana project GV99-115-1-02.

17. 18.

F. El Hadri / Ciencia Vol. 13, N 2 (2005) 218 - 227 19. CAVA-MONTESINOS P., CERVERA M.L., PASTOR A., DE LA GUARDIA M. Anal Chim Acta 481: 291-300, 2003. HOU M. Yejin Fenxi 10: 20-22, 1990. KOBAYASHI R., IMAIZUMI K. Anal Sci 6: 83-85, 1990. 24. 22. 23.
227 ANDREAE M.O. Anal Chem 56: 20642066, 1984. TAMARI Y. Bunseki Kagaku 48: 499-504, 1999. HALL G.E.M., PELCHAT J.C. Geostand Newsl 21: 85-91, 1997.

20. 21.