摘要:

Continuous Hydride Generation Low- pressure Microwave-induced Plasma Atomic Emission Spectrometry for the Determination of Arsenic Antimony and Selenium Journal of Analytical Atomic Spectrometry FLORIAN LUNZER AND ROSARIO PEREIRO-GARCIA Department of Physical and Analytical Chemistry Faculty of Chemistry C / Julian Claveria 8 University of Oviedo 33006 Oviedo Spain NEREA BORDEL-GARCIA Department of Physics Faculty of Physics University of Oviedo 33006 Oviedo Spain A LFREDO S ANZ- M ED EL* Department of Physical and Analytical Chemistry Faculty of Chemistry C/ Julian Claveria 8 University of Oviedo 33006 Oviedo Spain The direct coupling of on-line continuous hydride genera tion methods to low-pressure microwave-induced Ar and He plasmas sustained in a surfatron was evaluated for the determination of trace amounts of arsenic antimony and selenium by atomic emission spectrometry The effect of the hydrogen produced during the hydride generation step the optimization of the operating conditions for He and Ar plasmas and the figures of merit of the analytical systems are given.Detection limits (3s) of 0.7 0.9 and 4.1 ng ml-' were obtained for As Sb and Se respectively using an Ar plasma at 50 Torr (1 Torr = 133.322 Pa) and a power of 115 W. The relative standard deviations calculated at the 50 ng ml-I level were in the range 3-4% for the three elements. Poorer detection limits were obtained for He than for Ar discharges (by about two to five times depending on the emission line). Using an argon discharge the method was applied successfully to the determination of As in sea-waters.Keywords Hydride generation; low-pressure; microwaue- induced plasma; atomic emission spectrometry; arsenic antimony and selenium determination The production of volatile covalent hydrides from solution has proved to be a most effective means of sample introduction of suitable analytes into atom sources such as flames quartz tubes heated electrically or with flames,' and plasmas particu- larly the inductively coupled plasma ( ICP).2*3 Few papers however have been published concerning the use of hydride generation as a sample introduction system for low-power (< 300 W) microwave-induced plasmas (MIPs) because con- ventional atmospheric-pressure MIPs are highly perturbed by the hydrogen produced during the hydride generation step.Therefore most attempts to interface hydride generation tech- niques to MIPs included steps for prior separation of hydrogen from the hydrides e.g. a condensation tube immersed in liquid or a graphite furnace with metallic palladium6 to trap the hydrides temporarily; these trapped hydrides are subsequently released by volatilization during a higher tem- perature cycle. Tao and Miyazaki7 described the continuous on-line introduction of volatile hydrides and Hg into a helium MIP using a hydrogen separation membrane. Some hydride generation-MIP methods have been also described in which no hydrogen removal is performed. For * To whom correspondence should be addressed. example Barnett et ~ 1 . ~ 3 ~ coupled to the MIP a miniaturized hydride manifold aimed at producing a low hydrogen evol- ution.This system was based on the addition of a small portion of the sample volume to an NaBH pellet.*v9 A continuous sample hydride generation system coupled on-line to a TMolo Beenakker cavity was reported for selenium" and lead;" however poor detection limits by atomic emission spec- trometry (AES) were achieved. More recently the direct coup- ling of a continuous hydride generation system to a low-power MIP source known as the 'microwave plasma torch' (MPT) was described for the determination of arsenic antimony and tin,I2 and detection limits in the low-ngml-' range were reported. In 1974 Moisan and c o - w ~ r k e r s ~ ~ ~ ~ proposed a simple and compact surface-wave launcher known as the 'surfatron' for the generation of plasma columns using low-power microwave energy.In recent years the atmospheric pressure MIP created in a surfatron has been proposed for different applications (e.g. refs. 15-19). The results showed that the surfatron seems to provide a better tuning performance and higher resistance to detuning than resonance cavities. Moreover surface-wave discharges can be operated from a fewTorr to atmospheric pressure without requiring any change in matching conditions. Experiments have shown that the best operating pressure in a surfatron is not necessarily atmospheric p r e s ~ u r e ~ & ~ ~ and a change in plasma behaviour versus pressure was reported at around 100 T ~ r r . ~ ~ In this work the effect of the continuous addition of molecular hydrogen (mixed with the carrier gas) on Ar and He low-pressure plasmas sustained in a surfatron was in- vestigated.Optimization and analytical performance of a continuous hydride generation system coupled directly to the low-pressure surfatron were evaluated for the determination of arsenic antimony and selenium by MIP-AES. The appli- cation of the low-pressure Ar plasma to the determination of As"' in sea-water is also discussed. EXPERIMENTAL Instrumentation A diagram of the instrumentation is shown in Fig. 1 which also includes the scheme and dimensions of the continuous hydride generation system. Volatile hydrides are prepared by simultaneously pumping into a mixing T the sample solution (prepared in 2 mol 1-1 HC1) and a 1% solution of NaBH,. Journal of Analytical Atomic Spectrometry April 1995 Vol.10 31 1Computer n I 1 I I I - -1 Spectrometer [ !/ I I - I 1 Spectralink Fig. 1 Experimental set-up PP peristaltic pump; GLS gas-liquid separator (15 cm x 2 cm id.); C condenser; W waste; CL clamp; MWG microwave generator; S surfatron; and PMT photomultiplier tube The generated gases and accompanying solvent vapour pass through a glass column filled with 3 mm 0.d. glass beads which allow a smooth separation of gases from l i q ~ i d s . ' ~ ~ ~ - ~ ~ The hydrides and hydrogen are swept continuously through the column by the plasma carrier gas. Residual moisture is removed in a refrigerated c o n d e n ~ e r . ' ~ * ~ ~ A restriction was made with a throttling clamp on the plastic sampling tube prior to the plasma region to match the rate of aspiration of the vacuum pump with the total flow rate of carrier and evolved gases from the reaction system.A side- arm directed into a 50ml water-filled Erlenmeyer flask was incorporated into the sampling line as a means of controlling the equilibration of these flow rates. A possible mismatch reveals itself by producing bubbles within or by lifting the water level in the glass tube entering'the Erlenmeyer flask. By maintaining this water level stable a sampling efficiency of 100% could be obtained. Details of the MIP components the vacuum system and the spectrometric equipment are given in Table 1. The surfatron body was water-cooled by means of copper tubing ( 3 mm 0.d.) brazed to its exterior surface. Usually the coolant used for the desolvation condenser was also used to cool the surfatron.Different plasma pressures in the plasma were achieved by decreasing the displacement volume of the vacuum pump with the aid of a butterfly valve. Helium (99.999%) or alternatively argon (99.998%) were examined as plasma gases. In contrast to conventional MIPs held in a surfatron at atmospheric p r e s s ~ r e ~ ~ . ~ ' MIPs at low pressure in Ar or He were very easy to ignite at low pressure. In fact plasmas at pressures lower than 10 Torr ignited spon- taneously without the aid of any external source of electrons such as a Tesla coil or a tungsten wire. Reflected powers were always kept below 6 W. Reagents Stock standard solutions of arsenic(w) antimony(m) and selenium(rv) of lo00 mg 1-' were prepared from Merck atomic absorption standards.Working standard solutions were pre- pared daily by appropriate dilution of the respective stock standard solutions with 2 mol 1-' HCl which was used as the blank. Sodium tetrahydroborate(w) solution ( 1 YO m/v) was pre- pared by dissolving sodium tetrahydroborate(in) powder (Probus) in water stabilized with 0.5% m/v NaOH sodium hydroxide (final concentration). The solution was prepared freshly every day and filtered before use. 312 Journal of Analvtical Atomic Svectrometrv. ADrd 1995. Table 1 Instrumental equipment Components Description and specifications Microwave Generator AF GMW Microwave equipment- Microwave generator with reflected power meter Surfatron Microwave launcher Capillary Vacuum system Rotary ,vane pump Pressure measurement Spectrometer- Optical arrangement Monochromator Photomultiplier Read-out system 24-301 DR; frequency 2450 MHz forward power 30-300 W coaxial cable KMW 243 N 50 i2 .Brass structure described e l s e ~ h e r e . ~ ~ A brass chamber was connected to the open end of the quartz tube (see Fig. 1); this structure has two lateral symmetric outlets to a vacuum pump and a third outlet on its top for a pressure gauge. The plasma is viewed through a circular silica window of 38 mm diameter Fused-silica capillary 20 cm x 3 mm i.d. x 7 mm o.d. inserted concentrically to the surfatron Vacuubrand Model RZ8 MKS Baratron capacitance pressure transducer Model 122B; response 0.1- 100 Torr Plasma viewed axially and imaged 1 1 on the entrance slit using a fused-silica lens 1 m Jobin-Yvon HR-1000 M Czerny-Turner mount with a grating of 2400 grooves mm-'; slit width 0.1 mm Hamamatsu R-212 Jobin-Yvon Spectralink system controlled by a computer All reagents were of analytical-reagent grade and ultrapure water obtained with a Milli-Q system (Millipore) was used for the preparation of all solutions.Sample Preparation Sea-water samples were collected in a Spanish coastal region (Cantabric sea Salinas Asturias) and immediately acidified to a final pH of 2 for storage. The samples were then filtered through a 0.45 pm Millipore membrane. Immediately before analysis the samples were again acidified using concentrated HCl to a final HCl concentration of 2 mol 1-'. VOl. 10RESULTS AND DISCUSSION 3.0 2.5 2.0 -5 -' 1.5 1 .o 0.5 decrease in the background intensity (He/H,) was observed on increasing the pressure in the range assayed (10-90 Torr).Similar behaviour was observed when Ar was used in these experiments. However it is known that larger pressures lead to an increase in particle densities and therefore to a larger number of collisions which would give rise to an inverse trend. One explanation for the observed behaviour supported by some previous work,28 is that radiative de-excitation processes increase at the expense of non-radiative processes as the pressure is decreased below 100 Torr. As can be seen from Fig. 2 an optimum pressure range of 30-60 Torr for analytical purposes (ZN/ZB) was obtained for both Ar and He as plasma gases. It is interesting that this finding is in accordance with previous experiments with a low- pressure He sutfatron2' for a different application (determi- nation of pesticides by gas chromatography).It is also note- worthy that at low pressures Ar seems to be more efficient than He for atomization-excitation of arsenic (see Fig. 2). ' Plasma Features The size and appearance of the Ar and He discharges in the surfatron are strongly dependent on the pressure microwave forward power and flow rate of the support gases lowering the pressure increasing the microwave forward power or lowering the gas flow rates give rise to plasmas of considerably larger sizes; e.g. a light-pink Ar discharge that extended to about 10cm at 20Torr (50 W 300ml min-') shortened to about 4cm at 100Torr. Pink argon discharges were longer than white helium discharges under similar experimental conditions.Both discharges shortened with the addition of hydrogen (chemically generated by acidification of 1% NaBH in a continuous system) to the plasma carrier gas and the plasma became intensely pink for Ar and bright white-yellowish for He. As an illustration using the experimental conditions (hydride generation and plasma parameters) given in Table 2 a 115 W argon plasma extended about 7.5 cm from the surfa- tron structure at 20 Torr and 4 cm at 40 Torr. Discharges that filled the whole diameter of the discharge tube were obtained in all instances. It has been pointed out that low-pressure plasmas could withstand higher concentrations of molecular gases than atmospheric-pressure discharges.21 Using our experimental set- up it was observed that at 115 W Ar plasmas (130 ml min-') sustained at atmospheric pressure were extinguished by using flow rates of 1% NaBH as low as 0.12ml min-'.However for a low working pressure of 50Torr maintaining all the other conditions identical the plasma was not extinguished at 1% NaBH flow rates of even up to 1.3 ml min-' (the maximum flow rate assayed) in spite of the large amounts of molecular hydrogen produced continuously on acidification (see Fig. 1). Optimization of Plasma Parameters Arsenic(rrr) was the model analyte chosen for optimization studies. The line used for emission intensity measurements was As 1228.81 nrn,l2 which was found to provide the best signal- to-background ratio. The optimization studies were carried out for Ar and He plasmas independently following a univari- ant search.The ratio IN/ZB (where I is background-corrected net analyte emission intensity and ZB the background emission intensity) was the analytical parameter chosen as the criterion to be maximized. Selected cornpromize operating parameters for the hydride generation and atomization-excitation in the plasma used in these studies are given in Table 2. influence of Pressure The effect of pressure on the background signal and on the ZN/IB ratio at 228.81 nm was evaluated. An exponential Table 2 arsenic antimony and selenium Selected Operating Parameters for the Determination of Hydride generation system- Sample acidity NaBH concentration Sample flow rate NaBH flow rate Low-pressure plasma- Pressure Gas flow rate Microwave forward power 2 mol I-' HCI 1% (in 0.5% NaOH) 1.2 ml min-' 0.4 ml min-' Argon Helium 50 Torr 50 Torr 180 ml min-' 15 W 115 W 235 ml min-' Eflect of plasma gasJEow rate It was observed that especially at comparatively low flow rates (below 100 ml min-') an increase in the plasma gas flow-rate stabilizes the plasma either in pure gas (Ar or He) or in Ar-H or He-H mixtures.The effect of plasma gas flow rate in our experiments is shown in Fig. 3. Not surprisingly the greatest analytical signal occurs at modest carrier flow rates below 300ml min-'. Because the analyte is in a gaseous form (hydride) it can be transported into the plasma with nearly 100% efficiency at 3.5 I L I I 1 I I 0 2'0 40 60 80 100 Pressurenorr Fig. 2 Effect of pressure (1 Torr= 133.322 Pa) on the As I 228.81 nm analytical signal (net As signal over background Z,,/Zb) for 100 ng ml-' of As"' A 130 ml min-' of Ar as plasma gas and 110 W; and B 170 ml min-' of He as plasma gas and 120 W 3.5 f 3.0 2.5 2 .o 1.5 1 .o 0.5 P -= 1 I I I 1 I I 0 100 200 300 400 500 600 700 Flow rate/ml min-' Fig.3 Effect of the plasma gas flow rate on the As I 228.81 nm analytical signal observed A Ar plasma gas ( 1 10 W and 35 Torr); and B He plasma gas ( 1 10 W and 30 Torr) Journal of Analytical Atomic Spectrometry April 1995 Vol.10 313very low gas flow-rates; higher flow rates serve mainly to decrease the gas temperature and shorten the residence time of the analyte species in the discharge. Influence of microwave forward power For the selected hydride generation conditions given in Table 2 low-pressure Ar and He plasmas can be sustained at powers as low as 40 W with continuous introduction of the gaseous hydride-H mixture.The microwave forward power influences the plasma length and therefore the sample residence time in the discharge.23 As can be seen in Fig. 4 the arsenic analytical signal ratio (IN/IB) increased steadily in Ar and He plasmas with microwave forward power over the range 50-150 W and the relative increases were higher for Ar than for He discharges. One fact accounting for these results is related to the larger increase in background radiance as a function of microwave power observed for He plasmas than for the Ar discharges.28 Taking into account that the use of high powers shortens the tube lifetime and that attenuation of microwaves occurs in the coaxial cable,20 a power of 115 W was chosen for further experiments for both Ar and He plasmas. Analytical Performance The experimental conditions chosen for arsenic (Table 2) were also applied in the determination of antimony and selenium.Table 3 shows the detection limits (DLs) found in this study for AS"' Sb"' and Se" using the Ar discharge. The DLs were calculated as three times the standard deviation of the back- 10.0 r (1.25 9.0 1 B / I 3.0 1 I -= 0 40 ; 60 80 100 120 140 160 0.25 0 Power/W Fig. 4 Influence of microwave forward power on the As I 228.81 nm analytical signal selected (In/Ib) A 275 ml min-' Ar as plasma gas and 40 Torr; and B 170 ml min-' He as plasma gas and 30 Torr ground noise. Linear ranges up to 500 ng ml-' were observed for As and Sb and up to 100 ng ml-' for Sb.Relative standard deviations (n = 5) calculated for 50 ng ml-' of the correspond- ing element were 3.7% for As 3.0% for Sb and 3.0% for Se. Curiously and in disagreement with values reported for atmospheric-pressure MIPs,12 the DLs obtained for low- pressure He plasmas were poorer than those obtained with Ar as plasma gas. For example DLs of 1.4 ng ml-' for As I 228.81 nm 3.3 ng ml-' for As I 193.76 nm and 4.5 ng ml-' for Sb I 231.15 nm were obtained in the He discharge. This type of behaviour was reported a long time ago by Busch and Vi~kers,~ for low-pressure pure argon and helium MIPs in the measurement of the emission intensity produced by Hgo vapor- ized ele~trothermally.~~ It is known that electron temperatures are higher for He than for Ar MIPs and that on the other hand electron concentrations are greater for Ar plasmas.3o The observed analytical results emphasize the importance of the radiative ion recombination mode123.30 in the excitation processes occurring in the low-pressure MIP.Table3 also shows a compilation of DLs obtained with other previously published continuous hydride generation methods coupled to atmospheric-pressure MIPs and AES detection. The D.L.s obtained in this study especially for As and Sb compared well with those found for an atmospheric- pressure Ar-ICP,3 in spite of the much greater input power used for the ICP; in addition in the experiments using the ICP the sample flow rate was as high as 6.2ml min-' (five times higher than in this work).One can also note the superiority in terms of D.L.s of the low-pressure surfatron versus other atmospheric-pressure MIP source^.^*'^ The better DL reported for As at atmospheric pressure in ref. 7 can be explained by the removal of molecular hydrogen prior to entering the MIP discharge by using a hollow-fibre membrane. The ratio of sample and 1% NaBH flow rates selected for the evaluation of the analytical performance characteristics (Table 3) was about 3. This ratio was similar to those in the other methods that are included in Table 3.337,'2 However we observed that the DLs obtained in the MIP-AES low-pressure surfatron are strongly related to the amount of hydrogen entering the plasma. As can be seen in Fig. 5 the lower the flow rate of 1% NaBH (and the smaller the amount of H2 generated) the better was the As"' detection limit; this effect resulted from the combination of an increase in arsenic net emission intensity and a decrease in the background noise with decreasing hydrogen concentrations in the plasma.Therefore the results in Fig. 5 indicate that whenever the particular application makes it possible it could be very advantageous to decrease the flow rate of NaBH as consider- able improvements in the D.L.s can be achieved on lowering the hydrogen evolution. Table 3 Comparison of detection limits (ng ml-') obtained by published* continuous hydride generation methods MIP-AES and the present method Element As Sb Se Wavelength/ nm I 193.76 1200.33 I 228.81 I 234.98 1206.83 1231.15 1252.81 1259.81 I 196.09 1203.98 Low-pressure Ar MIP 115 W (this work) 1.2 2.8 0.7 1.5 1.4 0.9 1.2 11.1 4.1 - He MPT 200 Wl2 4.8 3.2 - - - - 5.9 - - - - 6.1 - * According to ref.3 the corresponding detection limits using ICP-AES for detection were As 1 Sb 0.4 and Se 0.5 ng m1-l. Hydrogen was removed on-line using a hollow-fibre membrane in this method. 31 4 Journal of Analytical Atomic Spectrometry April 1995 Vol. 100.8 1 0.7 - 1 0.6 - E 0.5 - 9 0.4 - 0 n 0.3 0.2 - . c - 0 - 0 0.1 0.2 0.3 0.4 0.5 Flow rate of 1% NaBHJml min-' Fig. 5 Effect of 1% NaBH flow rate on the LOD for As I 228.8 in an Ar discharge. Other conditions as in Table 2 Table 4 Recovery study of As"' added to sea-water using the con ous arsine generation system and the Ar low-pressure surfatron nm inu- As (111) added to As found* sea-water/ng ml-l ng ml-' s,* (Yo) 25 25.4 3.3 50 50.8 3.1 75 73.8 2.4 100 100.8 3.1 * Five replicates were carried out for each sample; S = relative standard deviation of the five replicates.Influence of Foreign Ions and Determination of Asm in Sea-water The proposed method was investigated for the determination of low levels of arsenic in sea-water. The effect on the As MIP- AES signal of adding increasing amounts of NaCl to the sample solution was investigated. Concentrations up to 30000 yg ml-I of sodium using the above hydride generation system and the Ar low-pressure surfatron were studied. As expected it was observed by monitoring the Na I 589.0nm emission that sodium does not reach the plasma. Hence no ionization interferences of this easily ionizable element in the MIP can be expected in the determination of arsenic.The evaluation of the effect of total inorganic carbon (TCO,) was also considered of interest as the resultant CO could give rise to plasma instabilities.26 The average concentration of TCO defined as the sum of the concentrations of C 0 3 2 - HC03- and dissolved COz in the ocean is low3 mol l-1.31 Experiments performed with the Ar low-pressure surfatron showed that no interferences in the determination of As are produced by concentrations of HC03- up to mol 1-I. Therefore the main interferences expected by foreign ions in the proposed method would be those of a chemical nature produced during the hydride generation step which are consequently common to all detection techniques using hydride generation. Sea-water samples in which arsenic could not be detected with our method were spiked with known amounts of As"' and analysed under the conditions given in Table 2.Table 4 shows the results of these recovery tests which demonstrate good agreement between expected and found As"' concentrations. CONCLUSIONS In contrast to conventional atmospheric-pressure MIPS the low-pressure discharges can be easily maintained in the pres- ence of molecular gases. This feature has been exploited for the study of the continuous on-line determination of some hydride-forming elements using low-pressure MIP-AES sus- tained in a surfatron. The good detection limits obtained and the adequate pre- cision observed with this technique prompt its direct coupling to high-performance liquid chromatographic (HPLC) methods via hydride generation at the exit of the column. Experiments with HPLC-hydride generation-MIP techniques based on the use of surfactants as mobile for the speciation of toxicologically important As and Hg species are in progress.The authors thank Dr. B. Fairman for his help with some steps in this work and P. Gonzalez Gonzalez of the machine shop of EITI of the University of Oviedo for the construction of the plasma vacuum brass chamber. Financial support from DGICYT (Spain) through project PB91-0669 is gratefully acknowledged. REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Nakahara T. Prog. Anal. At. Spectrosc. 1983 6 163. Nakahara T. Spectrochim.Acta Rev. 1991 14 95. De Oliveira E. McLaren J. W. and Berman S . S. Anal. Chem. 1993 5 2047. Robbins W. B. Caruso J. A. and Fricke F. L. Analyst 1979 104 35. Bulska E. Broekaert J. A. C. Tschopel P. and Tolg G. Anal. Chim. Acta 1993 271 171. Matusiewicz H. Sturgeon R. E. and Berman S. S. Spectrochim. Acta Part B 1990 45 209. Tao H. and Miyazaki A. Anal. Sci. 1991 7 55. Barnett N. W. Chen L. S. and Kirkbright G. F. Spectrochim. Acta Part B 1984 39 1141. Barnett N. W. Spectrochim. Acta Part B 1987 42 859. Ng K. C. Xu X. and Brechmann M. J. Spectrosc. Lett. 1989 22 1251. Ng K. C. and Shen W. L. Spectroscopy 1987 2 50. Pereiro R. Wu M. Broekaert J. A. C. and Hieftje G. M. Spectrochim. Acta Part B 1994 49 59. Moisan M. Beaudry C. and Leprince P. Phys. Lett.A 1974 50 125. Moisan M. and Zakrzewski Z. J. Phys.D. 1991 24 1025. Galante L. J. Selby M. and Hieftje G. M. Appl. Spectrosc. 1988 42 559. Luffer D .R. Galante L. J. David P. A. Novotny M. and Hieftje G. M. Anal. Chem. 1988 60 1365. Coulombe S. Tran K. C. and Hubert J. in Element-Specific Chromatographic Detection by Atomic Emission Spectroscopy ed. Uden P. C. American Chemical Society Washington DC Quintero Ortega M. C. Cotrino Bautista J. Saez M. Menendez Garcia A Sanchez Uria J. E. and Sanz Medel A Spectrochim. Acta Part B 1992 47 79. Calzada M. D. Quintero M. C. Gamero A. Cotrino J. Sanchez Uria J. E. and Sanz-Medel A. Talanta 1992 39 341. Rivibre B. Mermet J.-M. and Deruaz D. J . Anal. At. Spectrorn. 1987 2 705. Rivikre B. Mermet J.-M. and Deruaz D. J. Anal. At. Spectrom. 1988 3 551. Rivikre B. Mermet J.-M. and Deruaz D. J. Anal. At. Spectrorn. 1989 4 519. Granier A. Bloyet E. Leprince P. and Marec J. Spectrochirn. Acta Part B 1988 43 963. Pyen G. S. Long S. and Browner R. F. Appl. Spectrosc. 1986 40 246. Broekaert J. A. C. Pereiro R. Starn T. K. and Hieftje G. M. Spectrochim Acta Part B 1993 48 1207. Camuiia-Aguilar Pereiro-Garcia R. Sanchez-Uria J. E. and Sanz-Medel A. Spectrochim Acta Part B 1994 49 475. Duan Y. Du X. and Jin Q. J. Anal. At. Spectrom. 1994,9 629. Goode S. R. Buddin N. P. Chambers B. Baughman K. W. and Deavor J. P. Spectrochim. Acta Part B 1985 40 317. Busch K. W. and Vickers T. J. Spectrochim. Acta Part B 1973 28 85. Zander A. T. and Hieftje G. M. Appl. Spectrosc. 1981 35 357. Johnson J. S. Coale K. H. and Jannasch H. W. Anal. Chem. 1992 64 1065A. Liu Y. M. Fernandez M. L. Blanco E. and Sanz-Medel A. J. Anal. At. Spectrorn. 1993 8 815. 1992 189-204. Paper 41056430 Received September 16 1994 Accepted October 25 1994 Journal of Analytical Atomic Spectrometry April 1995 Vol. 10 315

ISSN:0267-9477    

DOI:10.1039/JA9951000311

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

Improved Thallium Hydride Generation Using Continuous Flow Methodologies LES EBDON PHILLIP GOODALL AND STEVE J. HILL Journal of Analytical Atomic Spectrometry Analytical Chemistry Research Unit Department of Environmental Sciences University of Plymouth Drake Circus Plymouth UK PL4 8AA PETER STOCKWELL P S Analytical Arthur House Sevenoaks UK TN1.5 6QY K. CLIVE THOMPSON LabServices Yorkshire Water Charlotte Road Shefield UK S2 4EQ The use of continous flow vapour generation to improve the determination of thallium by atomic spectrometry has been investigated. Attempts to generate volatile organothallium compounds were unsucessful but using sodium tetrahydroborate a volatile thallium species was produced. Evidence was obtained to suggest that the vapour phase species was thallium(1) hydride TIH.At room temperature noisy signals were obtained but when the reaction manifold was chilled to 0°C the signals were far more stable. Cooling did not significantly change the sensitivity. A characteristic concentration of 4 ng cm -3 was obtained for continous flow vapour generation determination of thallium by atomic absorption spectrometry (AAS). This is a 1500-fold improvement on the previous sensitivity for batch vapour generation AAS. Continous vapour generation also appears to eliminate any serious interferences from other hydride-forming elements. Flow injection of nitric acid between samples was used to remove memory effects. Keywords Thallium determination; vapour generation; thallium hydride; atomic absorption spectrometry Although the toxicity of thallium was recognized soon after its discovery in 1861 there is still very little in the literature concerning chronic exposure to low levels of the element.It is now well known that thallium is a cumulative poision and has mutagenic and teratogenic properties. Thallium has an abun- dance of in the earth's crust occurring in igneous rocks lead and zinc ores and some rare minerals. However one of the major environmental threats from thallium is to popu- lations living in the vicinity of coal-burning power stations and cement works. Although there are several analytical tech- niques available for the determination of thallium many lack the sensitivity required for environmental samples. For example direct flame atomic absorption spectrometry (FAAS) and inductively coupled plasma atomic emission spectrometry (ICP-AES) offer detection limits of 0.2 and 0.1 mg l-l res- pectively.Thus vapour generation techniques coupled with atomic spectrometry offer an attractive potential solution to low-level analysis. Thallium vapour generation presumably as the hydride was reported by Yan et al.' These workers used a batch method coupled with AAS and a characteristic mass of 1.2 pg TI for a 2 0 0 ~ 1 injection was obtained this being equivalent to a characteristic concentration of 6 pg ~ m - ~ . This method was therefore less sensitive than conventional FAAS whilst being subject to large positive interferences from the presence of other covalent hydride-forming elements. For example the addition of 2 pg of Te resulted in an 11-fold increase in the absorbance of thallium.For these reasons this report remained little more than a curiosity. In this paper we report improvements to thallium hydride generation AAS (HG-AAS) when compared with this previous report,' by the use of continuous flow methodologies. The volatile species was identified as probably thallium(1) hydride. Derivatization with sodium tetraethylborate (NaTEB) to yield metal alkyls has been shown to be a viable vapour generation technique for both ~ a d m i u m ~ . ~ and The possibility of utilizing organometallic derivatives of thallium was investi- gated in our studies as an alternative to thallium HG but no detectable quantity of thallium was evolved. EXPERIMENTAL Instrumentation Thallium HG was performed under continuous flow method- ology using a commercial hydride generator (PS Analytical Sevenoaks UK).A 2-1 mixing of the sample-blank stream (8 cm3 min-') with the reagent stream (4 cm3 min-l) was employed for the HG manifold. Atomic absorption measure- ments (SP9 Pye Unicam Cambridge) were accomplished using a flame heated quartz tube atom cell. Measurements were made at the 276 nm thallium resonance line. Control of the reactor collection and processing of the analytical signals (AAS only) were accomplished via a PC using commercial software (Touchstone Spinoff Technical Systems Benfleet UK). For refrigerated HG a pneumatic six- port switching valve (PS Analytical,) was used instead of the electrically operated valve of the hydride generator. Gas-liquid separation was accomplished using glass U-tube separators of either the Type A configuration (PSA H003-G101 Hydride Type) or Type B (PSA H003-G102 Mercury Type).The ICP-AES measurements were made using a Liberty 200 spectrometer (Varian Warrington Cheshire UK) with the commercial vapour generator operating independently of external computer control. Reagents Sodium tetrahydroborate and 27.5% v/v hydrogen peroxide solution (Aldrich Chemical Company Gillingham UK). Sodium tetrahydroborate solutions stabilized by the addition of alkali (NaOH 0.1 mol dm-3) were prepared freshly each day and filtered prior to use. Sodium hydroxide nitric mid and hydrochloric acid (Analytical reagent grade Merck Poole UK). Sodium tetraethylborate. Prepared in these laboratories but further purified by crystallization and recrystallization from sodium-dried diethyl ether at - 78 "C.Solutions were prepared Journal of Analytical Atomic Spectrometry April 1995 Vol. 10 31 7freshly each day and purged vigorously with argon prior to use to remove ether of crystallization. Standard thallium solution (T1 1000 pg ~ m - ~ Aldrich Chemical Company). RESULTS AND DISCUSSION Thallium Hydride Generation Thallium HG was investigated in some detail with either AAS (quartz tube atom cell) or ICP-AES detection of any evolved thallium species. Gas-liquid separation was accomplished via a conventional (Type A) U-tube gas-liquid separator. Preliminary experiments were performed with a reagent solution of sodium borohydride (NaBH 10 g dm-3 0.265 mol dm-3) dissolved in aqueous sodium hydroxide (NaOH 0.1 mol dm-3 4 g dm-3).Preliminary experiments indicated that an analytical signal for thallium was only observed when sample acidity (hydrochloric acid) was such that there was insufficient acid for complete neutralization and hydrolysis of the reagent solution ([H+]<0.18 mol dm-3). These initial experiments are summarized in Table 1. The sensitivity was obviously much better than that reported by Yan et a!.' e.g. characteristic concentration z 10 ng ~ m - ~ but the signal was extremely noisy and suffered from slow rise times severe memory effects and variable peak shapes (Fig. 1). The absence of an analytical signal at a non-absorbing line (T1 291.8 nm) confirmed that the analytical signal was specific to thallium. This was supported by the presence of three emission lines characteristic of thallium in the ICP emission spectrum of the gas purged from the gas-liquid separator.The application of deuterium background correction in the AAS measurements did not improve the noise characteristics of the Table 1 Thallium hydride vapour generation [H+]/mmol dm-3 Observation 25 50 75 150 180 No detectable signal Small absorbance Significant absorbance but noisy signal Large absorbance but very noisy signal No detectable signal Time/s - Fig- 1 Thallium hydride generation at room temperature. Zone 1 blank (0.17 mol dm-3 HCl) and reagent; Zone 2 sample and reagent; Zone 3 blank and reagent; Zone 4 sample and reagent; Zone 5 blank and reagent; Zone 6 acid purge (2 cm' 4 mol dm-' HCl); and Zone 7 blank and reagent analytical signals.This indicated that systematic instrumental effects could not explain the poor performance of thallium vapour generation. The possibility that some of the detrimental properties of thallium HG were attributable to atomization effects was discounted when ICP-AES detection was applied to the vapour generation system and no significant differences between detection with AAS and ICP-AES were observed. Thallium was identified unambiguously by the presence of three characteristic lines in the ICP emission spectrum. tnterferences [n the paper by Yan et al.,' a severe interference from other hydride forming elements (i.e. As Se and Te) was observed. This effect was attributed to the formation of a complex hydride. Using a continuous flow method no increase in sensitivity was observed upon the addition of equal ( gravimetric) concentrations of other hydride-forming species.The peak shapes were dramatically improved by the use of it nitric acid matrix for the sample/blank solutions. Typical rise times in hydrochloric acid (0.150 mol dm - 3 ) of 180 s (half maximum signal) were reduced to z 75 s in a nitric acid medium (0.150 mol dm-3). The memory effect and excessive noise demonstrated previously for the generation of thallium from a hydrochloric acid medium were retained in the nitric acid system. Identification of Vapour Phase Species The TI' and T1"' hydrides are both known6 and it was thought that the latter was the target species as covalent hydride formation normally stabilizes the higher oxidation states of main group metals (e.g.GeH SnH PbH,). Experimental evidence for the relative stability of the thallium hydrides is sparse but a number of theoretical treatments have suggested a sequence of stability TlH,- > TlH > TlH > T1H2+ >TlH+ .7 Experimentally only the four most stable hydride species have been observed.' Therefore by analogy with plumbane it was thought that the use of an auxiliary oxidant might be necessary for efficient and smooth generation of thallium hydride. In acid solution hydrogen peroxide (E" H,0,:H20 = 1.77 V) should oxidize T1' to TI"' (E" T13+:Tlt = 1.25 V). The latter couple is extremely sensitive to pH owing to the extreme insolubility of Tl(OH) [E" Tl(OH),:Tl OH =0.05 V] and therefore oxidation by hydrogen peroxide under alkaline con- ditions is also favourable.The addition of hydrogen peroxide under all conditions summarized in Table 1 resulted in the complete suppression of the thallium AAS signal. This indicates that the volatile species was probably not thallium(1rr) hydride. According to Wiberg et a1.,' thallium(1) hydride has been found experimentally to be more stable than the thallium(I1I) hydride with the T1' species being stable to 150 OC." Given this evidence it is likely that the vaporous species is therefore thallium(1) hydride. An interesting phenomenom was observed in that the residual boron signal caused by incomplete gas-liquid separ- ation was dependant upon thallium concentration (Fig. 2). This at first suggested that perhaps some thallium-borane compound was the vapour phase species.This was quickly discarded as the peak shapes of the boron emission did not follow a similar profile to that of the T1 emission signal i.e. the rate of appearance of the boron emission signals was more rapid than that of the thallium signal. The rise and decay times of the boron signal were at least twice as fast as the thallium signal. Similar behaviour of the residual sodium emission signal with respect to [Tl] was also observed (Fig. 2). This also excludes the possibility that the analytical signal was due to carry-over of colloidal thallium from the gas-liquid 318 Journal of Analytical Atomic Spectrometry April 1995 Vol. 100 2 4 6 8 10 12 14 16 18 20 [Tll/ng cmP3 2 0.40 0.30 0 0.20 2 0.10 Fig. 2 Dependance of thallium concentration on residual boron and sodium signals where 0 B; and 0 Na - - - - separator as again coincidence of the B/Na and TI signals would be indicative of this situation. This phenomenum of increased sodium and boron carry- over only occurs with incomplete hydrolysis of sodium tetra- hydroborate.Similar behaviour was observed for lead HG which also occurs under acid deficient conditions.' Under typical conditions for arsine generation (2 mol dmP3 HCl 1% NaBH in 0.1 mol dmP3 aqueous NaOH) i.e. complete hydrolysis no additional sodium and boron are carried over into the atom cell. Memory Effect and Variable Peak Shapes The variable peak shape appears to be a function of residual thallium from the previous sample. If the reagent stream was replaced temporarily by an acid blank the original baseline was recovered upon re-introduction of reagent to the system. A subsequent run exhibited the relatively slow rise times of the original experiment.After a number of runs a black deposit upon the wall of the gas-liquid separator directly opposite to the inlet for the reaction mixture could be seen. Test tube scale reactions in an inert atmosphere were revealing. The addition of an aliquot of thallium solution (20 cm3 1000 pg cm-3 T1 0.16 mol dm-3 HNO,) to sodium tetrahydroborate solution (10 cm3 10 g dm- NaBH 0.1 mol dm- NaOH) resulted in the formation of a finely divided black precipitate which redissolved slowly on standing. This precipitate was assumed to be thallium metal. Despite this rather complex behaviour of the reaction of thallium with aqueous sodium tetrahydroborate it is possible to make a number of deductions (1) thallium is probably transferred from solution to the carrier gas and thus to the atom cell initially as a vapour; (2) the analytical signal cannot be due to carry-over of colloidal thallium as the precipitation of thallium metal in the system is immediate and therefore the peak shapes for sodium boron and thallium would be co-incident; (3) residual thallium metal in the separator could not increase the kinetics of the transfer of colloidal thallium metal as an aerosol; and (4) a mechanism for transfer of thallium to the atom cell depending only upon carry-over of I Thalliim in v1 - Thallium oul Loop 1 6 s cooling coil 7 t To GLS Mixing T-piece I 1 I 6scoolingcoil I I I NaBH4in I Ice bath I Fig.3 Schematic diagram of continuous flow system manifold V injection value for cleaning cycle; V2 sample/blank switching valve; and hoop 1 injection loop for cleansing acid 0.50 I I streams and the introduction of cooling coils prior to merging of the sample/blank with the reagent stream. These changes were introduced to ensure that the reactants were cooled to the ice point prior to reaction. The immediate benefit was the elimination of the excess noise from the system (Fig. 4) with no loss of the enhanced sensitivity of the continuous flow method compared to the batch method of Yan et al..' This chilling made it possible to assess realistically the performance of the system and yielded a characteristic concen- tration of 4 n g ~ m - ~ .This is a factor of 1500 better than reported previously and additionally does not exhibit the extreme degree of positive interference from other hydride- forming elements. The elevated baselines due to the presence of residual thallium were retained between the refrigerated and ambient temperature vapour generation systems. This was eliminated by introducing a second switching valve (Fig. 3). This enabled the flow injection of relatively concentrated acid (1.0 cm3 4 mol dmP3 nitric acid) into the sample/blank stream to act as a cleaning stage in the analytical cycle. Using this system good repeatability was obtained for both peak shape and analytical signal on successive measurements e.g. 2% relative standard deviation on both peak areas and peak maxima measurements.Mechanism of Thallium Hydride Generation colloidal metal does not appear to be consistent with the reported increased sensitivity of thallium vapour generation at 273 K' nor the elimination of excessive noise from the system (this work). Refrigerated Thallium Hydride Generation Tl+(aq)+BH,-(aq)=TlBH Low temperature generation ( = 273 K) was used by Yan et al.' to improve the sensitivity of their system. This was also applied to the continuous flow method used here via the manifold shown in Fig. 3. The important features are the replacement of the electrical switching valve of the vapour generator with a pneumatic valve to alternate between sample and blank The mechanism of formation of T1H probably proceeds via the thallium tetrahydroborate species which has been prepared from aqueous sol~tions.~ Thallium tetrahydroborate is a white crystalline material which is insoluble in water and relatively inert with respect to hydrolysis.The products of hydrolysis are thallium metal and boric acid. The former being amphoteric will readily redissolve in the reaction mixture leading to the observed memory effect. The interesting reaction of thallium tetrahydroborate with Journal of Analytical Atomic Spectrometry April 1995 Vol. 10 319respect to thallium HG is that TlH can be prepared by heating the TlBH at 40"C9 2TlBH4 = 2TlH + B2H6 This suggests that thallium hydride is initially formed by this route in competition with the normal hydrolysis reactions of tetrahydroborate species. It is not clear how the formation of thallium hydride is catalysed by thallium metal nor why cooling the reaction mixture results in smoother generation of thallium hydride. The latter may be as a result of slowing the rate of hydrolysis of thallium tetrahydroborate thallium hydride or indeed sodium tetrahydroborate.Further investigation is required to identify unambiguously the exact mechanisms which are responsible for the formation of thallium hydride. Organothallium Compounds for Vapour Generation Thallium has an extensive organometallic chemistry based upon thallium(m) alkyls and aryls. Triethyl thallium is a volatile liquid [bp = 43 "C ( 1 mm Hg) decomposes at 125 "C] which seems at first sight to be an ideal candidate as the basis of a vapour generation system. Reaction of aqueous thallium(1) standards under alkaline or neutral pH with aqueous sodium tetraethylborate solutions did not yield detectable quantities of triethyl thallium.The use of an auxiliary oxidizing agent (hydrogen peroxide) did not promote the formation of triethyl thallium. This was assumed to be due to either the extreme reactivity of thallium(II1) alkyls e.g. they react violently with water are pyrophoric and explosively decompose on exposure to light or the trapping of the analyte in the inert and stable dialkyl cation [ T1( Et)J -+ . Thallium(1) cyclopentadienyl is stable and inert and the preparation is trivial. On the test tube scale treatment of an alkaline solution of thallium(1) with aqueous alkali saturated with cyclopentadiene immediately precipitates the organomet- allic compound.In practice the cyclopentadienyl derivative was found to be too involatile for use in a vapour generation system i.e. mp = 215 "C decomposes at = 300 "C. In both cases the type B configuration gas-liquid separator was employed as an interface due to previous experience with organometallic vapour generati~n.~?~ This work was supported financially under a Science and Engineering Research Council (SERC) Department of Trade and Industry (DTI) Link scheme with P S Analytical and Labservices Yorkshire Water. REFERENCES 1 2 3 4 5 6 7 8 9 10 Yan D. Yan Z. Cheng G.-s. and Li A.-m. Talanta 1984,31,133. DUlivo A and Chen Y.-w. J. Anal. At. Spectrom. 1989 4 319. Ebdon L. Goodall P. Hill S. J. Stockwell P. B. and Thompson K. C. J. Anal. At. Spectrom. 1993 8 723. Sturgeon R. E. Willie S. N. and Berman S. S. Anal Chem. 1989 61 1867. Ebdon L. Goodall P. Hill S. J. Stockwell P. B. and Thompson K. C. J. Anal. At. Spectrom. 1994 9 1417. Lee A. G. in The Chemistry of Thallium Elsevier Amsterdam 1971. Schwerdtfeger P. Boyd P. D. W. Bowmaker G. A. Mack H. G. and Oberhammer H. J. Am. Chem. Soc. 1989 111 15. Waddington T. C. J. Chem. SOC. 1958,4783. Wiberg E. Dittmann O. Noth H. and Schmidt M. Z. Naturforsch. B 1957 12 61. Wiberg E. Dittmann O. Noth H. and Schmidt M. 2. Naturforsch. B 1957 12 62. Paper 4/03988B Received July I 1994 Accepted December 1 1994 320 Journal of Analytical Atomic Spectrometry April 1995 Vol. 10

ISSN:0267-9477    

DOI:10.1039/JA9951000317

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

Glucose as a Chemical Modifier for the Determination of Antimony and Selenium by Electrothermal Atomic Absorption Spectrometry Journal of Analytical Atomic Spectrometry M. TERESA PEREZ-CORONA M. BEATRIZ DE LA CALLE-GUNTINAS YOLANDA. MADRID AND CARMEN CAMARA" Departmento de Quimica Analitica Facultad de Ciencias Quimicas Universidad Complutense de Madrid 28040 Madrid Spain The effectiveness of glucose as a chemical modifier for the determination of Sb and Se by electrothermal atomic absorption spectrometry has been evaluated. The effect of increasing glucose concentration on the thermal behaviour of both elements and on the graphite tube surface is described. Glucose allows the quantitative stabilization of Sb and Se up to 1300°C compared with 700°C without glucose. Furthermore this chemical modifier significantly reduces interference by phosphorous in the determination of Se and by chloride in the determination of Sb.Keywords Antimony; selenium; glucose chemical modiJier; electrothermal atomic absorption spectrometry EXPERIMENTAL Apparatus A Perkin-Elmer Model 1 l00B atomic absorption spectrometer equipped with an HGA-400 graphite furnace and a deuterium lamp background corrector was used. Perkin-Elmer Sb and Se hollow cathode lamps operating at 20mA were used as the light sources. The spectral bandpass used to isolate the 217.6 nm Sb and 196.0 nm Se lines was 0.2 nm. An Ar flow- rate of 300ml min-l was used to purge air from the cuvette except in the atomization step in which stop flow was selected. Reagents The very low detection limits that can be achieved with electrothermal atomic absorption spectrometry (ETAAS) have made this technique one of the most widely applied of those based on AAS.However certain drawbacks especially in the determination of volatile elements are systematically associ- ated with analysis by this technique the most notable ones being loss of analyte by volatilization in the high-temperature pyrolysis step and interference by sample constituents which can give rise to a background that in some cases cannot be corrected even when Zeeman effect-based equipment is used. These drawbacks lead to a deterioration of the analytical All reagents were of analytical-reagent grade or higher purity and purified de-ionized water was obtained using a Milli-Q system (Millipore).Standard Sb (lo00 mg 1-I) was prepared by dissolving 0.2740 g of potassium antimony1 tartrate (Carlo Erba) in de-ionized water and diluting to 100 ml. Standard Se solution (loo0 mg 1-I) was prepared by dissolv- ing 0.5531 g of sodium selenite (Sigma) in water and diluting to 250 ml. Working solutions were prepared daily in glucose media by adding the appropriate amounts of glucose (Aldrich). Argon of 99.99% purity was used (Carburos Metalicos Spain). characteristics. Therefore the determination of volatile elements by ETAAS requires the use of chemical modifiers. Several chemical modifiers have been reported including inor- ganic reagents such as Mg(N03)2 Pd(N03)2,1-3 Pt cu Ni Pd Cu and Mo4,5 and organic reagents such as ascorbic acid oxalic acid and glucose6-* normally in combination with Procedure Graphite tube pretreatmentfor scanning electron microscope observation TO study how glucose affected the tube surface a set of tubes inorganic reagent^.^ Selenium and Sb (especially at high chloride levels) are troublesome to determine by this technique because they tend to form volatile compounds that can be lost during the drying and pyrolysis steps.An additional difficulty in determining Se arises when phosphate forms the molecules P2 and PO which cause major spectral interference when deuterium background correction is used." To avoid these problems the inorganic reagents mentioned above have been extensively used for the determination of these elements."-13 Organic compounds have not been used as often although for some analytical tasks they are more effective and do not give rise to equipment comprising two non-pyrolyitic graphite coated graphite and two pyrolytic graphite coated graphite was immersed in 2% glucose solution and a second identical set of tubes was immersed in 20% m/v glucose solutions.Low pressure was applied to force the glucose solutions into the graphite pores. Vacuum and atmospheric pressure were alternately applied until no air bubbles formed under vacuum which indicated that the glucose sQlution had completely filled the graphite pores. Then the tubes were placed in an oven and kept there at a temperature of 100°C for 4 h. The tubes were further dried by subjecting them to slow electrothermal heating to at least 900°C. Once the tubes had been pretreated as described contamination. The main aim of this work was to study the applicability of glucose as a chemical modifier in the thermal stabilization of Se and Sb and its effectiveness in the further reduction of chloride and phosphate interferences.above their surface was observed using a scanning electron microscope. Sampleanalysis A 20 p1 portion of standard Se (80 pg 1-') and Sb (40 pg 1-I) solutions in 0.001 % and 2% m/v glucose medium respectively were introduced into the graphite tube. The sequential dry- pyrolysis-atomization programme (Table 1) of the HGA-400 * To whom correspondence should be addressed. Journal of Analytical Atomic Spectrometry April 1995 Vol. 10 321Table 1 Temperature programme for the pyrolysis and atomization curves for antimony and selenium Whereas glucose concentrations in the 2-10% m/v range enabled thermal stabilization of Sb without altering sensitivity Temperature/ Hold time/ Ramp time/ Gas flow/ Step "C S S ml min-' Drying 110 5 20 300 Pyrolysis* 400- 1600 5 30 300 Atomization? 1300-240 2 0 stop Cleaning 2500 2 1 300 * Charing curves were recorded at atomization of 2300 "C.? Atomization curves were recorded at pyrolysis of 700 "C. was followed and the peak absorbance was registered as peak height. Glucose and Milli-Q water blanks were run regularly. RESULTS AND DISCUSSION Efficiency of Glucose as a Matrix Modifier in Antimony and Selenium Determination Experiments were carried out to determine the optimum temperatures and times for drying pyrolysis and atomization. The pyrolysis and atomization temperatures were optimized by preparing pyrolysis and atomization curves using 40 pg 1-1 of Sb and 80 pg I-' of Se in the presence and absence of glucose at different concentrations. This study was carried out with non-pyrolytic and pyrolytic tubes Previous studies showed that pyrolysis-atomization behaviour was similar for Sb"' and Sbv and for SeIV and SeV1.Thus for further studies only Sb as Sb"' and Se as SeIV were monitored. The results are shown in Fig. 1. In the absence of chemical modifier the maximum pyrolysis temperature that can be applied is 900°C and decreasing absorbance values are obtained at higher temperatures [Fig. l ( a ) and ( b ) ] . When glucose was added to the solution the pyrolysis temperature could be increased up to 1300°C without risk of Sb and Se losses.The optimum glucose concentration was more critical for Se than for Sb. 0.150 (a) 0.100 0.500 a 9 1 b 500 1000 1500 '. 2000 2500 4 0 0 x 0.200 a (b) t3=21OO c 0 0 Tern perature/"C Fig. 1 Pyrolysis (left) and atomization (right) curves for (a) antimony (glucose concentrations 0 / 0 0 % ; ./El 2%; A/A 6%; V/V lo%) and (6) selenium (=/El Se'"; A/A Se'"+2% glucose; V/V SeIV +0.001% glucose). Sb concentration 40 pg 1-l; Se concentration 80 pg I-'; sample volume 20 pl; pyrolysis temperature 700 "C; atomiz- ation temperature 2300 "C for Sb and 2100 "C for Se; non-pyrolytic tube [Fig. l ( a ) ] the thermal stabilization of Se by 2./0 glucose solution was accompanied by a sharp decrease in the analytical signal [Fig. l ( b ) ] .Glucose has also been used by other workers and it has been found that the sensitivity for Se was reduced in the presence of chloride and Ni(NO,) owing to decompo- sition products from glu~ose.'~ The better thermal stabilization of Sb and Se in the presence of glucose could be explained according to the mechanisms involving organic reagents proposed by Volinsky6 and Welz.* The addition of glucose results in the formation of carbon- aceous,residues and the Se and Sb oxides formed in the pyrolysis step are reduced by these carbonaceous residues to less volatile metal forms which are stable up to 1300°C. Thermal behaviour of Sb and Se was the same regardless of whether non-pyrolytic or pyrolytic graphite coated graphite tubes were used. However in the case of Sb there was a significant difference in sensitivity between the two types of tubes.The sensitivity with non-pyrolytic graphite tubes was the same for aqueous Sb solutions as for Sb solutions in glucose medium even when the glucose concentration was 10% m/v [Fig. l ( a ) ] . When atomization was in the pyrolytic graphite coated graphite tubes the sensitivity was drastically decreased by increasing the glucose concentration in the solu- tion. The Sb absorbance signal in 10% m/v glucose medium was approximately 10% of that in aqueous Sb solution. This lower sensitivity when antimony was atomized in pyrolytic tubes might be due to an inner spreading surface formed because of the addition of high amounts of glucose. This effect is especially marked in pyrolytic graphite coated graphite tubes because their surface is more polished than that 'of non-pyrolytic graphite tubes.For Se no significant differences were observed using either pyrolytic or non-pyrolytic tubes. Although the permissible pyrolytic temperature is higher in the presence of glucose the glucose concentration required for good sensitivity and analyte stabilization is significantly lower for Se (0.001 Yo) than for Sb (2-10%). The amount of glucose needed for Se might not be high enough to form an inner spreading surface and thus no differences were noted between the two types of graphite tube. To evaluate how glucose affects the tube walls tubes were treated as previously described and observed with a scanning electron microscope. In a first study tubes were immersed in 2% m/v glucose solutions.The amorphous structure of the non-pyrolytic graph- ite made it difficult to appreciate any significant differences between non-pretreated and pretreated non-pyrolytic graphite tubes. In pyrolytic graphite coated graphite tubes which have a much more highly organized structure some deposits were observed on the tube surface. Under a magnification of 1500 x (Fig. 2) the deposits seemed to be laminar structures adhering to the tube surface and not part of the tube itself. To rule out the possibility that the deposits were impurities Fig. 2 Appearance of a pyrolytic tube after pretreatment with 2% m/v glucose solution. Magnification 1500 x 322 Journal of Analytical Atomic Spectrometry April 1995 Vol. 10present in the tube a pyrolytic graphite coated graphite tube was pretreated as previously described but immersed in 20% m/v glucose solution.An image of the resulting surface showed that a more significant alteration of the surface took place. The laminar structures detected might be the carbonaceous residues that reduce the Sb and Se oxides and that are also responsible for the formation of a spreading surface. Use of Glucose to Remove Chloride and Phosphate Interferences in the Determination of Sb and Se It is well-known that one of the main problems in determining Se by ETAAS arises from phosphate-rich matrices. Thermal decomposition of phosphates in the graphite tube leads to the formation of polyatomic species mainly P PO and POz which result in serious spectral interferences." The present work showed that glucose significantly reduces interference by phosphate.Relatively high amounts of phosphate (about 1000 mg 1-I) seriously interfere in the determination of 80 pg 1-1 Se and the interference cannot be corrected by a deuterium lamp [Fig. 3(a)]. The presence of 0.001 YO glucose completely prevented phosphate interference [Fig. 3(b)] at 1000 and IOOOO mg -'. The effectiveness of glucose in removing chlorine interference in Sb determination was also tested. As mentioned in the introduction high chlorine concentrations increase Sb vola- tility and seriously interfere in Sb determination. As shown in Fig. 4(a) 20000 mg 1-l of chloride (the normal content of sea- water) results in a high background that cannot be corrected in the determination of 40 pg 1-l of Sb.However in the presence of 2% glucose the background signal is drastically reduced [Fig. 4(b)]. Another important consequence of the use of glucose is that there is no overlap between the maximum of the antimony absorbance signal and that of the background so overcorrection does not take place. Use of Glucose as a Matrix Modifier in Antimony Determination in Spiked Sea-water Glucose solution (2% m/v) was chosen as the optimum medium for Sb determination in sea-water spiked with 1Opg 1-l and 40 pg 1-' of Sb. Temperature was critical for precise and accurate recovery. The drying step was observed to be critical. If the temperature was increased quickly the sample spattered out of the tube increasing irreproducibility. Also once the sample was placed in the tube it spread over the walls and did not remain at the point where it was placed.These 0 3.0 Tim els Fig.3 Absorbance uersus time curves for 80pg I-' of Se plus 10 000 mg 1-' phosphate (solid line) and background (broken line) in (a) the absence of glucose and (b) the presence of 0.001% m/v glucose (1.000) ; ~,. . 1 .ooo L (IJ 2 2 1.000 2 (1.000) 0 I I . . ',. . . . . i . r.k"c- . -_. . ..... -.-=- 2.0 Timels Fig.4 Absorbance versus time curves for 4Opg 1-' of Sb plus 20 000 mg I-' chloride (solid line) and background (broken line) in (a) the absence of glucose and (b) the presence of 2% m/v glucose Table 2 in sea-water using a glucose chemical modifier Temperature programme for the determination of antimony Temperature/ Step "C Drying 120 Pyrolysis 700 900 1300 Atomization 2300 Cleaning 2500 Hold time/ 20 40 20 10 2 2 S Ramp time/ 120 60 60 20 0 1 S Gas flow/ ml min-' 300 300 300 300 stop 300 drawbacks were eliminated by increasing both the ramp and hold time in the drying step (see Table 2).The ashing step was subdivided into different parts. The temperature was increased very slowly up to 900°C (Sb is stable up to this temperature even in the absence of chemical modifiers) and then up to 1300°C over 20 s. If this latter temperature was reached quickly in a single step the concom- mitants in the sample were not properly eliminated and were atomized together with the Sb resulting in a background that was too high and could not be corrected by the deuterium lamp. The spiked sea-water samples were analysed by interpolation on a calibration curve constructed in 2% m/v glucose medium.The limit of detection (defined as 3s according to IUPAC) was 3 pg 1-' and the recovery was about 98%. The use of glucose allows the determination of Sb in samples with high chloride levels with good recovery values. This fact is presented as an advantage over the use of ascorbic acid which provides 25% lower Sb recoveries in matrices with high chloride content especially as potassium ~hloride.~ The proposed method was validated by using hydride generation (HG) AAS and the results obtained were 9.7 40.6 and 10.3 40.8 pg 1-1 using the proposed method and HGAAS respectively. CONCLUSIONS In view of the results obtained we conclude that glucose is a feasible chemical modifier for the determination of Sb and Se by ETAAS.Its use permits pyrolysis temperatures as high as those used with the most common inorganic matrix modifiers. The advantages of glucose over these matrix modifiers are it does not contaminate equipment with possible future analytes; it is completely harmless; it is easy to obtain Sb- and Se-free glucose so the glucose blanks are the same as those run on the tube alone; it does not diminish tube lifetime; and it is cheap. Another advantage of a glucose chemical modifier is that Journal of Analytical Atomic Spectrometry April 1995 V01.10 323it efficiently removes major interference by phosphate and chloride in the determination of Se and Sb respectively. The authors thank the Direccibn General de Investigacibn Cientifica y Tecnica (Project PB 92/0218) and M.B.de la Calle thanks the Spanish Ministry of Education and Science for providing financial support. M. Gormann is thanked for revising the manuscript. REFERENCES Schlemmer G. and Web B. Spectrochim. Acta Part B 1987 41 1157. Bermejo-Barrera P. Soto-Ferreiro R. M. and Bermejo- Barrera A. Fresenius’ J. Anal. Chem. 1993 345 60. Styris D. L. Prell L. J. Redfield D. A. Holcombe. J. A. Bass D. A. and Majidi V. Anal. Chem. 1991 62 508. Johannessen K. Gammelgaard B. Jons O. and Hansen S. H. J . Anal. At. Spectrom. 1993 8 199. 5 6 7 8 9 10 11 12 13 14 Ni Z . M. and Shan,X. Q. Spectrochim. Acta Part B 1988,42,937. Volinsky A. Tikhomirov S. V. Senin V. G. and Kaskin A. N. Anal. Chim. Acta 1993 284 367. Gilchrist G. F. R. Chakrabarti C . L. Byrne J. P. and Lamoureux M. J. Anal. At. Spectrom. 1990 5 175. Welz B. Akman S. and Schlemmer G. Analyst 1985 110 459. Peramaki P. and Lajunen L. H. J. J. Anal. At. Spectrom. 1992 7 735. Saeed K. and Thomassen Y. Anal. Chim. Acta 1981 130 281. Dahl K. Thomassen Y. Martinsen I. Radziuk B. and Salbu B. J. Anal. At. Spectrom. 1994 9 1. Radziuk B. and Thomassen Y. J. Anal. At. Spectrom. 1992 7 397. Laborda F. Viiiuales J. Mir J. M. and Castillo J. R. J. Anal. At. Spectrom. 1993 8 737. Dedina J. J. Anal. At. Spectrom. 1987 2 435. Paper 4/06091 A Received October 6 1994 Accepted December 5 1994 324 Journal of Analytical Atomic Spectrometry April 1995 VoE. 10

ISSN:0267-9477    

DOI:10.1039/JA9951000321

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

Electrothermal Atomic Absorption Spectrometric Determination of Ultratrace Amounts of Tellurium Using a Palladium-coated L'vov Platform After Separation and Concentration by Hydride Generation and Liquid Anion Exchange MARC0 GROTTI AND AMBROGIO MAZZUCOTELLI Zstituto di Chimica Generale Universita di Genova Viale Benedetto X V 3-1 61 32-Genova Italy An ultrasensitive method is described for the measurement of nanogram amounts of tellurium in the presence of interferents. Tellurium is separated from the matrix by hydride generation concentrated on a liquid anion exchanger (liquid Amberlite LA-2) and atomized from a platform previously treated with palladium. Analytical performances are reported and the possibility of application of the method to the analysis of complex samples (sea water sediments) is proved.Keywords Tellurium; electrothermal atomic absorption spectrometry; hydride generation; liquid anion exchanger; L'vov platform; coating; preconcentration; hydrogen telluride trapping Tellurium is widely employed in the electronic and metallurgi- cal industries. In geochemistry tellurium is used as a 'path- finder' or indicator of a certain type of deposit. Tellurium compounds are toxic and affect various organs as does sel- enium. Although tellurium has not caused any social problem because of its low abundance in the environment it is generally accepted as moderately toxic to plants and highly toxic to mammals even at trace levels.',2 The Italian Law No. 915/82 has included tellurium in the number of the elements that must be determined to classify a waste as 'toxic and harmful'.The determination of tellurium in real samples is complicated by its very low concentration and by matrix interference A widely employed technique for the determination of this element is hydride generation atomic absorption spec- trometry (HG-AAS) because of its relatively high sensitivity precision and reasonable selectivity. However this method is not sufficiently sensitive to determine very low concentrations of tellurium accurately and usually preconcentration processes are needed. The enrichment of hydrogen telluride in a cryogenic trap was attempted by Kobayashi et al.' but without success probably because of the instability of this hydride. In situ concentration of the volatile hydrides within the pre-heated graphite furnace and subsequent atomization was first pro- posed by Lee6 for the collection of bismuthine and sub- sequently applied by Andreae7 and by Yoon et a1.* for the determination of tellurium in environmental samples.An improvement in sensitivity and precision has been reported by Doidge et aL9 and Zhang et al.," by employing a palladium- coated graphite furnace. Ni et al." also studied the trapping capability of a silver-coated graphite tube. Hydrogen telluride has also been collected by bubbling it through a potassium iodide-iodine solution," but this method is limited by inter- ferences from copper mercury silver and selenium that must be removed prior to HG. Journal of Analytical Atomic Spectrometry I I Finally good results have been reported by Tsalev and Mandj~kov,'~ by trapping the hydrogen telluride in a cerium( 1V)-iodide absorbing solution.In the present work a new method based on the generation of hydrogen telluride its trapping on a liquid anion exchanger (Amberlite LA-2) and analysis by electrothermal AAS (ETA AS) is presented. Analytical performances are reported and the possibility of application of the method to the analysis of complex matrices (sea-water sediments) is proved. EXPERIMENTAL Apparatus Hydride generation was accomplished in a Varian VGA-76; the black fluoroelastomer tubing emerging from the gas-liquid separator was dipped through a glass pipette into the trapping solution and held in a graduated test-tube. Helium was used as the inert gas. A Varian SpectrAA 300 atomic absorption spectrometer equipped with a Zeeman graphite tube atomizer was used.Pyrolitic graphite coated graphite tubes with forked pyrolytic platforms were used. The operating parameters were lamp current 10 mA; wavelength 214.3 nm; bandpass 0.2 nm; and measurement mode integrated absorbance (QA). Reagents All the reagents were of reagent grade quality Tellurium standard solution 1000 ppm. A Spectrosol solution from Merck. Working standards were prepared daily by serial dilutions with Milli-Q water. Copper selenium mercury silver titanium and silicon 1000 ppm and iron calcium magnesium sodium potassium and aluminium standard solutions 10 OOO ppm. Spectrosol solutions from Merck. The required concentration of each element was obtained by diluting the respective standard solution with Milli-Q water. Calcium carbonate 99.995%.From Aldrich. NaBH solution 5% m/v. Prepared by dissolving 5.0 g of NaBH (analytical-reagent grade Aldrich) and 2.0 g of NaOH pellets (Carlo Erba pure reagent) in 100 mi of Milli-Q water. Liquid anion-exchange solution. Prepared by adding 10 ml of Amberlite LA-2 (Merck) to 5 ml of 6 ml 1-' HC1 stirring continuously and diluting with 10 ml of isobutyl methyl ketone (IBMK). Finally the organic phase was treated with 2 moll-' NH,C1 and 2 moll-' NH to pH 7. Journal of Analytical Atomic Spectrometry April 1995 Vol. 10 325Palladium solution 100 ppm. Obtained from a palladium standard solution 1000 ppm (Aldrich). Procedure The sample in 100 ml of 6 mol I-' HCl was boiled for 45 min and then delivered to the system for HG.The hydrogen telluride was collected into 1.8ml of the trapping solution (0.5 ml of aqueous phase and 1.3 ml of organic phase) held in a graduate test-tube. During the generation time part of the trapping solution evaporated and the final volume of the organic phase was z 1.0 ml. The concentrate was centrifuged at 2000 rev min-l for 10 min. Finally 10 pl of the organic phase were injected onto the palladium-treated platform and atomized according to the temperature programme reported in Table 1. Sample Preparation A 0.2g amount of a fine powdered Reference Sediment Material [MESS- 1 National Research Council Canada (NRCC)] were transferred into the Teflon vessel of the pressure 'bomb' (Perkin Elmer Italy) and 1 ml of a 100ngml-' tellurium standard solution was added.Then 4ml of aqua regia (hydrochloric acid-nitric acid 3 + 1) were added and the 'bomb' was heated at 150°C for about 20 h. After cooling and centrifugation the solution was transferred into a 100 ml graduated flask and diluted with 6mol1-1 HC1. 100ml of a Nearshore Seawater Reference Material (CASS-2 NRCC) were transferred into a 200ml graduated flask and 0.5 or 1 ml of a 100 ng ml -' tellurium standard solution was added. RESULTS AND DISCUSSION Reduction of Tew to TeN Since only Te" forms the hydrogen telluride it is necessary to reduce Tev' to Te" prior to HG. In agreement with several author^,'^*'^,'^ this was carried out by boiling aqueous solu- tions of tellurium in 6mol1-' HC1 for 45min. Also the selectivity of HG for TeIV allows the speciation of TeIV and ~ ~ v J .7 ~ 8 Table 1 Furnace temperature programme Temperature/ "C 50 150 350 1200 1200 2600 2600 Ramp time/ 5.0 50.0 30.0 10.0 0.0 0.7 0.0 S Hold time/ 0.0 0.0 0.0 2.0 2.0 2.0 2.0 S Gas flow/ 1 min-' 3.0 3.0 3.0 3.0 0.0 0.0 3.0 Hydrogen Telluride generation The effect of the concentration and the flow rate of HC1 and NaBH on the evolution of the hydrogen telluride had been already in~estigated.~,~,'~ Good results were obtained using the following conditions 6 mol I-' HCl; 5% m/v NaBH,; and flow rate for HCl NaBH and sample of 1 l and 4-5 ml min-' respectively. Trapping on Amberlite The collection of the hydrogen telluride was obtained by trapping it on the liquid anion exchanger Amberlite (LA-2). Since the formation of the species Se2- during the stripping of gaseous selenium hydride has been propo~ed,'~ an analogous species can be considered to explain the fixing of tellurium (which belongs to the same Group 16) on the liquid exchanger TeH- +Te2- + H+ (LA-2)Cl- +Te2- -+(LA-2)Te2- +C1- As can be deduced from the first two equilibria an acid environment does not facilitate the formation of Te2-.'Therefore the increase of the media pH by treating the anion exchanger with a base was expected to improve the trapping efficiency. By changing the pH value from 1 to 7 the efficiency of the process was increased from 40 to 90%. At higher pH values the exchange capability of LA-2 decreases probably owing to competition from OH- ions. ,Qtomization 'Tellurium was atomized from a platform previously treated with palladium by pre-injecting a solution of the modifier by carrying out the furnace temperature programme until the pyrolysis temperature was reached.The use of a palladium- coated platform has the following advantages the application of the L'vov platf~rm;'~.'~ the covering of the graphite surface; and the employment of a chemical m~difier.'*~'~ The result is an improvement of the atomization of tellurium as shown in 'Table 2 where the analytical performances obtained by employing different atomization systems (wall platform pal- ladium-treated platform) are reported for the analysis of a 25 ppb Te standard solution. Finally the higher pyrolysis temperature that can be used in the presence of palladium permits the organic matrix (IBMK LA-2) to be vaporized before the analyte is atomized.Interferences Hydride generation offers besides the possibility of a precon- centration of the analyte the achievement of a satisfying separation of the analyte from the matrix. Several elements were considered as possible interferents in the ETAAS determi- nation of tellurium in environmental samples (Table 3). No Table 2 Integrated absorbance repeatability and maximum pyrolysis temperature without loss of analyte T,,,(max) for the atomization of a 25 ppb Te standard solution from different atomization systems (n = 2) Atomization system Wall* Pyrolytic graphite coated graphite platform? Pd-treated platform$ Integrated Absorbance RSD (%) Tp&W/"C 0.079 & 0.004 5.5 0.1 63 +_0.008 4.8 0.1 65 & 0.007 4.3 700 700 1300 326 Journal of Analytical Atomic Spectrometry April 1995 Vol.10Table 3 List of interferents typically occuring in three environmental matrices Interferent NaCl NaF Na,SO NaHCO KBr KCl CaCI MgC12 SrCl H3BO3 Sea-water/ 27.65 g 1-1 0.019 2.660 0.133 0.020 0.466 0.533 1.560 0.008 0.020 Interferent Si A1 Fe K Mg Ca Na Ti Mn Sediment/ Calcium carbonate/ mg I-' Interferen t g 1-' 3 50 CaCO 24.97 500 250 150 100 100 50 30 10 interferences were found in the determination of 2 ng ml-1 of tellurium. It is well known that HG is susceptible to interferences from transition metals and hydride forming elements.20 In the deter- mination of tellurium by HG and trapping Maher" found severe interferent effects from copper mercury silver and selenium. The interferences of the same elements at the same analyte concomitant ratio were studied.No significant (< 10%) interferences were found. Analytical Figures of Merit The analytical figures of merit of the proposed method are reported in Table 4. The sensitivity of the method depends on the atomization efficiency and on the enrichment due to the preconcentration step. From the data in Table 2 we can deduce a characteristic mass of 16.7 pg for an integrated signal of 0.0044 s in agreement with the instrument's performance. Considering the whole process the characteristic concen- tration and the detection limits are 25 and 100 pg ml-1 respect- ively. The relative standard deviation (RSD) of the method evaluated by analysing nine samples on different days is 4.5%. The efficiency of the process has been evaluated by compar- ing the absorbance of a 1 ng ml-' tellurium standard solution after concentration with the absorbance of a lOOngml-' tellurium standard solution directly introduced into the furnace.The calibration graph shows a good linearity (r=0.9991) in the range 0-2ngml-' and can be adequately conveyed by the following equation QA = [Te]0.182+0.014 Higher analyte concentrations are assessable by working with lower samples volumes. Finally in order to evaluate the analytical accuracy of the method and to prove the possibility of application of the method to the analysis of real samples the method was applied in the analysis of seawater and sediment reference materials CASS-2 and MESS-1. Since these materials are not certified for tellurium it was added before sample preparation.The analytical results are summarized Table 4 Analytical figures of merit Integrated absorbancels Blank 1 ng ml-' tellurium standard solution after concentration 100 ng ml-' tellurium standard solution directly introduced into the furnace Characteristic concentration/ng m1-I Detection limit/ng ml-' Trapping efficiency (YO) RSD (%) Correlation coefficient 0.016 0.006 0.189 t0.009 0.209 k 0.004 0.025 0.100 4.5 0.999 1 90 Table 5 Analytical results Te added/ Sample Pg I-' CASS-2 1 .o CASS-2 1 .o CASS-2 0.5 MESS-1 l.O* MESS-1 1.0* MESS-1 1.0* Te found/ 1 .oo 1.18 0.44 0.68 0.79 0.70 I % 1-l Recovery 100 118 88 68 79 70 (%I * Concentration in solution (after solubilization). in Table 5. The recovery is very high for the sea-water samples and lower but satisfactory considering their very low concentration for the sediment samples.REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Cooper W. C. Tellurium Van Nostrand Reinhold New York 1971 ch. 7. Gerhardsson L. Glover J. R. Nordberg G. F. and Vouk V. in Handbook on the Toxicology of Metals eds. L. Friberg Nordberg G. F. and Vouk V. Elsevier Science Publishers B.V. Amsterdam 2nd edn. 1986. Sedykh E. M. Belyaev Yu. I. and Sorokina E. V. J. Anal. Chem. (USSR) 1980 35 2162. Ni Z.-m. and Shan X.-q. Spectrochim. Acta Part B €987 42,937. Kobayashi R. Imai M. and Hashimoto Y. Bunseki Kagaku 1982 31 467. Lee D. S. Anal. Chem. 1982 54 1682. Andreae M. O. Anal. Chem. 1984,56,2064. Yoon B. M. Shim S. C. Pyun H. C. and Lee D. S. Anal. Sci. 1990 6 561. Doidge P. S. Sturman B. T. and Rettberg T. M. J. Anal. At. Spectrom. 1989 4 251. Zhang L. Ni Z.-m. and Shan X.-q. Spectrochim. Acta. Part B 1989 44 751. Ni Z.-m. He B. and Han H.-b. J. Anal. At. Spectrom. 1993 8 995. Maher W. A. Analyst 1983 108 305. Tsalev D. L. and Mandjukov P. B. Microchem. J. 1987 35 83. Yammarnoto M. Yasuda M. and Yamamoto Y. Anal Chem. 1985,57 1382. Agterdenbos J. Bussink R. W. and Bax D. Anal. Chim. Acta 1990 232 405. L'vov B. V. Spectrochim. Acta Part B 1978 33 153. Slavin W. Manning D. C. and Carnrick G. R. At. Spectrosc. 1981 2 73. Tsalev D. L. Slaveykova V. I. and Mandjukov P. B. Spectrochim. Acta Rev. 1990 13 225. Weibust G. Langmyhr F. J. and Thomassen Y. Anal. Chim. Acta 1981 128 23. Welz B. and Melcher M. Analyst 1984 109 569. Paper 4/03359K Received June 6 1994 Accepted December 1 1994 Journal of Analytical Atomic Spectrometry April 1995 Vol. 10 327

ISSN:0267-9477    

DOI:10.1039/JA9951000325

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

Determination of Barium by Electrothermal Atomic Absorption Spectrometry I Journal of I Analytical 1 Atomic Spectrometry MARIA IN& c. MONTEIRO Departamento de Quimica da Pontificia Universidade Catblica do Rio de Janeiro 22453-900 Rio de Janeiro R.J. Brazil ADILSON JOSE CURTIUS" Departamento de Quimica da Universidade Federal de Santa Catarina 88040-900 Florianbpolis S.C. Brazil Barium atomization in electrographite pyrolytic graphite- coated (pyrocoated) and tungsten-coated graphite tubes using argon or nitrogen as sheath gas was studied. Sensitivity appearance time pyrolysis and atomization temperature curves tube lifetime and scanning electron micrographs were obtained for the different surfaces. The best characteristic mass and longest lifetime were obtained in a pyrocoated tube without a platform using argon as sheath gas.Nitrogen increases the appearance time and decreases the sensitivity owing to the formation of gaseous barium monocyanide. High- purity argon or a lower pyrolysis temperature does not increase the tube lifetime. The appearance temperature around 2200"C suggests that barium is atomized through the reduction of barium by the carbon. Complete atomization was not obtained with any of the conditions as a plateau was not verified on the atomization temperature curve up to 2650°C. Nitrogen also decreases the maximum pyrolysis temperature in a pyrocoated tube. The micrographs of the tungsten pre- treated tube explained its bad performance. The tungsten coverage is discontinuous and melting of the tungsten phase decreases the coverage and leads to the destruction of the pyrocoating.Non-spectral interferences were observed specially at high salt concentrations. The analysis of real samples is possible if the standard additions method is used. The concentration of barium found in Citrus Leaves agreed with the certified value. Keywords Barium determination; graphite furnaces; electrothermal atomic absorption spectrometry; scanning electron micrographs; surface and salt eflects The determination of barium is important for biochemical geochemical environmental and legal medicine studies. Soluble barium salts may produce cardiac dysfunction and other diseases.' Interest in the possible leaching of barium into drugs intended for parenteral administration originates from the presence of barium in glass (usually encasing the drug) the use of barium stearate as a plasticizer softener and pigment dispersant and the use of barium sulfate as a filler in the production of rubber.2 Barium sources in the environment are diesel fuel containing a barium base additive and barium polysulfide used as a fungicide.' The distribution of barium concentrations in sea-water could be applied to circulation rate s t ~ d i e s .~ - ~ Some examples for the determination of barium by electro- thermal atomic absorption spectrometry (ETAAS) using a graphite furnace have been reported e.g. barium was deter- mined in calcium carbonate,6 gunshot residues,738 drugs intended for parenteral administration,2 oilfield injection water and its suspended solids,' brines and caustic soda solution," paper mill stock and water systems'' and sea-water.12-" * To whom correspondence should be addressed.The high atomization temperatures used (between 2500 and 2900°C)'3~15 and in some situations the high salinity of the matrices1* decreased strongly the lifetime of pyrocoated tubes. A lifetime as low as 25 firings was found when barium was determined in sea-water diluted 1 + 1 with water using ammonium nitrate as chemical modifier and an atomization temperature of 2700 "C. The thin pyrolytic graphite layer does not protect the graphite tube efficiently and other coatings and impregnation materials have been studied in order to improve the atomizer surface properties." A method of impreg- nation of the graphite tube with tantalum was proposed for the determination of several element~.'~ However the effect of this form of pre-treatment on barium was not investigated.The scope of this work was the study of the effects of different graphite surfaces and sheath gases in the determi- nation of barium by ETAAS. Morphological studies by scan- ning electron microscopy (SEM) were used to explain the performance of the different surfaces. The possibility of the application of the optimum conditions will be evaluated through an interference study and analysis of certified reference materials. In particular tube pre-treatment with tungsten not previously used for barium will be investigated. EXPERIMENTAL Instrumentation A Perkin-Elmer Model 1100 atomic absorption spectrometer equipped with a Perkin-Elmer HGA-300 graphite furnace an AS-40 autosampler and an Epson FX-86E graphics plotter were used.A Perkin-Elmer Z 3030 instrument equipped with a Zeeman-effect background corrector a Perkin-Elmer HGA-600 graphite furnace an AS-60 autosampler and a PR-100 graphics plotter were also used. The measurements were made at the 553.6 nm resonance line with bandwidths of 0.2 nm (L) and 0.7 nm (L) when the Model 1100 and 2-3030 spectrometers were used respectively. Signal evaluation was based exclusively on integrated absorbance values. When only analytical solutions of barium in 0.1 moll-' nitric acid were measured the Model 1100 instrument was used without back- ground correction. In the studies of acid and salts effects and in the analysis of real samples the absorbance measurements were performed with the Model Z 3030 instrument without or with Zeeman-effect background correction.Scanning electron micrographs were obtained with a Hitachi Model S 450 scanning electron microscope. A JEOL Model JXA-840 A energy-dispersive X-ray fluorescence spectrometer was also used. Materials and Reagents Argon of 99.9 and 99.998% purity and nitrogen of 99.9% purity were used. The following graphite tubes and platforms Journal of Analytical Atomic Spectrometry April 1995 Vol. 10 329were used uncoated polycrystalline electrographite (EG) tubes Perkin-Elmer Part No. B0070699; pyrolytic graphite-coated graphite tubes (PG) Perkin-Elmer Part No. B0109322; and L'vov platform pyrolytic graphite Perkin-Elmer part No. B0109324. Before the tubes and platforms were used analytically a conditioning programme according to the manufacturer's recommendations was applied.The acids used in the experiments were of high-purity grade (Ultrex) from J. T. Baker except phosphoric acid which was of analytical-reagent grade from Merck. A barium stock standard solution containing 1000 mg 1-1 of Ba was prepared from a Titrisol ampoule (Merck). Chloride salts of analytical-reagent grade from Merck were used for most solutions. Stock standard solutions were prepared by dissolving the appropriate amounts of salts in de-ionized water. Calcium and magnesium stock standard solutions were prepared by dissolving calcium carbonate and magnesium oxide (high-purity grade from Johnson Matthey Chemicals) in hydrochloric acid. Iron stock standard solution was prepared by dissolving iron wire in hydrochloric acid.Tungsten stock standard solution was prepared by dissolving 15.6 g of analytical-reagent grade sodium tungstate hydrate (Na2W04.2H,0) in de-ionized water. A total volume of lop1 was used in all experiments using the furnace programme given in Table 1. No chemical modifier was added. Tube Treatment Tungsten-pre-treated graphite tubes were prepared according to the procedure described by Zatkalg with some modifications. The graphite atomizer tubes were vertically immersed in 8.8% m/v tungsten soaking solution contained in a plastic beaker. The beaker was transferred into a desiccator which was evacu- ated with a water pump and maintained under reduced pressure for 30 min. The air bubbles that formed on the tube walls were released by tapping the exterior of the desiccator.Atmospheric pressure was restored in the desiccator and the tubes were removed from the bath and dried first at the room temperature for 30 min and then at 105 "C for 1 h. Each tube was mounted in the atomizer unit and the barium atomization programme (Table 1) was used. This procedure was repeated once again but the tubes were soaked under reduced pressure for 18 h. Samples and Sample Decomposition The following certified reference materials were used Standard Reference Material (SRM) 1572 Citrus Leaves from the National Institute of Standards and Technology (NIST) and NASS-2 Open Ocean Seawater from the National Research Council Canada. Solid sample decomposition was carried out according to the procedure described by Curtius et a1.;20 the citrus leaves material was dried at 85 "C for at least 4 h before weighing.Four portions of 0.2000 g of material were weighed and transferred into the Table 1 Temperature programme used for the determination of barium PTFE beaker of a Perkin-Elmer Autoclave-3. A 5 mi volume of de-ionized water and 5 ml of concentrated nitric acid were added. The autoclave was closed heated to 155 5 "C on a hot- plate and kept at this temperature for 90 min. After cooling to the room temperature the autoclave was opened and the clear solution transferred quantitatively into a 25 ml calibrated flask and diluted to volume with de-ionized water. The sea-water sample was injected in the pyrocoated tube after a 1 + 1 dilution with de-ionized water.RESULTS AND DISCUSSION Pyrolysis and Atomization Temperature Curves Pyrolysis and atomization temperature curves for barium in 0.1 mol 1-1 nitric acid using different tube materials and argon or nitrogen as sheath gas are depicted in Fig. 1. No plateau was reached in any of the atomization temperature curves indicating that better sensitivities could probably be obtained 0 1000 2000 TemperaturePC ~ 3000 Fig. 1 Pyrolysis and atomization temperature curves for Ba in 0.1 moll- HNO (a) 0 PG Ar wall atomization (0.2 ng Ba); x PG N wall atomization (1.0 ng Ba); 0 PG Ar wall atomization W-pre- treated (0.5 ng Ba). (b) A EG Ar wall atomization (0.1 ng Ba); [I EG N wall atomization (0.5 ng Ba). (c) 0 PG Ar platform a.tomization (0.4 ng Ba); V PG N platform atomization (2.0 ng Ba) Step No.Furnace temperaturerc Time/s Internal gas flow rate/ ml min-l 1 90 1 10 300 2 150 15 10 300 3 1600 ( Ar); 1200 (N,) 1 30 300 4* 2650 0 10 0 5 2650 1 4 300 6 20 10 10 300 * Read in this step. 330 Journal of Analytical Atomic Spectrometry April 1995 Vol. I0using temperatures higher than 2650°C. It is interesting that other workers" found for atomization temperatures higher than 2550"C measured by a pyrometer that the signal decreased using Th-pre-treated pyrocoated tubes whereas for W-pre-treated tubes as used in this work a maximum was not found. In the pyrolysis temperature curves a plateau was obtained up to the maximum pyrolysis temperature which ranged from about 1600 to 2000°C for the different systems.The appearance temperature obtained from the atomization temperature curves using argon as sheath gas was 2200 "C for atomization from the wall which agrees with the appearance temperature found by Campbell and Ottaway,22 indicating that reduction of barium oxide by the carbon may occur. The appearance temperature obtained for the W-pre-treated pyro- coated tube was lower around 18OO"C which could imply a different atomization mechanism. Unexpectedly the slower rate of heating for platform atomization gave a lower appear- ance temperature around 1800 "C. Also nitrogen in compari- son with argon produced lower appearance temperatures around 2000°C for uncoated and pyrocoated tubes. Appearance Time and Characteristic Mass Appearance times and characteristic masses for barium in 0.1 mol 1-1 nitric acid are presented in Table 2.In order to avoid taking into account the decrease in sensitivity produced by the Zeeman-effect correction the measurements used to obtain the characteristic masses were performed without back- ground correction. The temperature programme shown in Table 1 was used except that the pyrolysis temperature was 1200°C in all instances. The sensitivity was determined after about ten initial firings for each condition. As expected platform atomization produced longer appearance times in comparison with wall atomization for pyrocoated tubes. The W-pre-treatment led to a slightly lower appearance time for the pyrocoated tube but a higher appearance time for the uncoated tube. The times were always higher when nitrogen was used instead of argon.The best sensitivity was obtained using pyrocoated tubes wall atomization and argon as sheath gas. The W pre-treatment did not improve the sensitivity for the coated tube but decreased the characteristic mass (the mass of the analyte that gives an integrated absorbance of 0.0044 s) for the uncoated tube. The use of nitrogen instead of argon had a strong effect decreasing the sensitivity and increasing the appearance times possibly owing to the formation of stable gaseous monocyanide molecules.23 The best characteristic mass obtained 6.6 .+ 0.3 pg (0.0044 s)-' is about ten times higher than the value calculated by L'vov et They assumed several conditions in their calculations including total vaporization and atomization of the analyte no ionization of the free atoms and isothermality of the furnace over its lengths.They attributed the discrepancy between the calculated and the measured characteristic masses primarily to the formation of gaseous molecules of carbides and monocyanides. In fact Styri~,~' vaporizing barium chloride in vitreous carbon and in pyrolytic graphite furnaces observed gas-phase barium dicarbide appearing concurrently with free barium using atomic absorption spectrometry and mass spec- trometry. Later L'vov~~ obtained a degree of ionization of 0.35 for barium at 2430"C showing that ionization may also affect the determination of the characteristic mass. Tube Lifetime Tube lifetimes were studied for untreated and W-pre-treated pyrocoated tubes using argon as sheath gas.As shown in Fig. 2(a) using argon of 99.9% purity the barium signal increased in the first determinations needing around 30 firings to reach a reasonable stability. After 170 firings the signal decreased almost continuously. The relative standard deviation presented Fig. 2(a) shows a good precision for more than 300 firings. The tube lifetime decreased when argon of higher purity (99.998%) was used [Fig. 2(b)] or when a lower pyrolysis temperature (1200 instead of 1600 "C) was used [Fig. 2(c)]. This unexpected behaviour could not be explained. The worse performance concerning the lifetime was shown by the W-pre- treated tube for which the sensitivity decreased almost continu- ously with the number of determinations reaching about half of the initial value after 100 firings as shown in Fig.2(d). Mass Loss of Pyrocoated Graphite Tube Using argon of 99.9% purity as sheath gas a pyrocoated graphite tube lost 0.35% of its original mass after 11 firings without barium injection and 0.44% when a total of 330 ng of barium were injected into the tube. Using nitrogen of 99.9% purity the losses were higher 0.61 and 0.67% without and with 330 ng of barium respectively. The formation of gaseous carbon may occur in all situations. In the presence of barium the loss is enhanced probably owing to the formation of gaseous barium carbide and monocyanide. In a nitrogen atmosphere apparently the formation of monocyanide is favoured. Scanning Electron Micrographs of Pyrocoated Tubes without and with Tungsten Pretreatment The morphology of the sampling area of a pyrocoated tube [the same tube as shown in Fig.2(a)] after 375 firings using argon (99.9% purity) with a total load of 187.5 ng of barium in 0.1 moll-' nitric acid is shown in Fig. 3(a). The character- istic smooth nodular surface of the pyrocoated tube has disappeared and the surface is similar to that of an electrolytic graphite showing that the pyrolytic layer was destroyed. An energy-dispersive X-ray fluorimetric (EDXRF) analysis of this surface gave a barium signal suggesting the formation of Table 2 Appearance time and characteristic mass for barium in 0.1 mol 1-1 nitric acid Tube material* PG PG EG EG PG PG PG-W EG-W Atomization Wall Wall Wall Wall Platform Platform Wall Wall Sheath gas Ar N2 Ar N2 Ar N2 Ar Ar Appearance time/s (0.5 ng of Ba; n=2) 0.61 0.73 0.64 1 .o 0.92 1 .o 0.59 0.7 Characteristic mass/pg 6.6 0.3 40.0 f 0.2 20.0 5 0.1 160+7 11.4k0.4 90+2 7.0+ 1 16+1 (0.0044 s)-1 (n=4) * PG = pyrolytic graphite-coated graphite; EG = polycrystalline electrographite; PG-W = W-pre-treated pyrocoated graphite; EG-W = W-pre- treated uncoated tube.Journal of Analytical Atomic Spectrometry April 1995 Vol. 10 331i 40 0.4 0.2 0 0. *. . - 20 0 ,"" 000 00 ooo O a o o O Y O o O ~ O o ~ %OO ooooooooooooooIm 00" oo 0 " " 00 I I I n Jo 100 200 300 400 40 (I) 0 m . +? 0 A2 m 1 2o 0.2 . 00 O 0 000 On L 0-J a C 4- - 0 0 0 100 200 300 *(d) 0 .. . 0. C ) 0 0.4 0.2 40 0.4 20 0.2 - 0 . 0 0 0. . 0 0 0. O.0 O .. .O 0. 0 0 0 0 0 0 O O - 0 0 0 O0 0 O 0 0 0 0 0 0 00000000000 oo 0 0 0 - 0 0 oooooooooo oo 0 ooo 0 0 0 00 0 0 0 8 0 100 200 0 Firing No.100 200 Fig. 2 Integrated absorbance and RSD(%) for 0.5 ng of Ba in 0.1 moll-" HNO over the lifetime of a pyrocoated tube. Each datum point is the average of five determinations (a) Ar of 99.9% purity; (b) Ar of 99.99896 purity; (c) Ar of 99.9% purity and pyrolysis temperature of 1200 "C; . and ( d ) Ar of 99.9% purity and W-pre-treated pyrocoated tube barium carbide. 'Also a barium signal was detected in the rim of the inner wall of the tube indicating vapour-phase reactions. The morphology of the nodular deposits exfoliation and delamination of the pyrocoated layer in the sample injection hole of the same tube as shown in Fig.3(a) can be seen in Fig. 3(b). The morphology of the sampling area of a pyrocoated tube [the same tube of as in Fig.2(b)] after 200 firings with a total load of 100 ng of barium in 0.1 mol 1-1 nitric acid using argon of higher purity (99.998%) can be seen in Fig. 4(a). It exhibits corrosion of the nodular surfaces but delamination and destruction of the pyrocoated layer were not observed as in the tube discussed previously [Fig. 4(b)]. This could be due to the different purity of the argon or to the different number of firings and total load of barium. An EDXRF analysis of this sampling area and of the rim of the inner wall of the tube also gave a barium signal. Many nodular deposits and exfoliation of the pyrocoated layer in the sample injection hole area can be seen in Fig. 4(b). The increased nodular deposits may be due to a lower oxygen concentration in the sheath gas.The morphology of a W-pre-treated pyrocoated tube using argon of 99.9% purity shows a discontinuous covering of the tungsten layer in an unused tube especially in the top of the carbon nodules [Fig. 5(a)]. Fig. 5(b) shows the morphology of the sampling area of a W-pretreated pyrocoated tube after 155 firings with a total load of 77.5 ng of barium in 0.1 mol 1-1 nitric acid using argon of 99.9% purity [the same tube as in Fig. 2(d)]. The tungsten phase became more discontinuous probably resulting from the fusion of the phase. As the tungsten carbide has a melting point of 2870+50°C,27 it is more probable that a eutectic mixture of tungsten with carbon which has a lower melting-point 2425-2732 0C,28,29 was 332 Journal of Analytical Atomic Spectrometry April 1995 Vol.10Fig. 3 Morphology of the surface of a pyrocoated tube after 375 firings with a total load of 187.5 ng of Ba in 0.1 mol I-' HNO and Ar of 99.9% Purity (a) sampling area with destroyed pyrocoated layer; and (b) injection hole from the inside with carbon nodules and Fig. 5 Morphology of a W-pre-treated pyrolytic graphite tube (a) dis- continuous covering of the tungsten layer in an unused tube; and (b) sampling area with discontinuous tungsten phase after 155 firings with a total load of 77.5 ng of Ba in 0.1 HNO using Ar of delamination 99.9% purity Fig.4 Morphology of the surface of a pyrocoated tube after 200 firings with a total load of 100 ng of Ba in 0.1 mol 1-1 HN03 and Ar of 99.998% purity (a) sampling area with corrosion of the nodular surface; and (b) injection hole from the inside with many nodular deposits and exfoliation of the pyrocoated layer obtained.The incomplete tungsten covering and the destruc- tion of the pyrocoated layer are the reasons for the non- improvement of the barium sensitivity in this tube. Pre- treatment of the tube with yttrium produced a sensitivity enhancement of about 30% which remained approximately unchanged only during the first 5-10 firings.24 Effects of Acids and Salts The superiority of the pyrocoated tube in an argon atmosphere with regard to tube lifetime and sensitivity is clear as already discussed. Using these conditions the possible use of the ETAAS in the determination of barium in real samples was further explored. First the effects of some acids and salts on the barium signal was investigated.Argon of 99.9% purity was used and blanks were run in parallel with all determinations. The effects of nitric hydrochloric phosphoric sulfuric and perchloric acid in the concentration range 0.1-4.0 mol 1-1 on the signal of 0.2 ng of barium are shown in Table 3. The effect was less than 13% except for phosphoric acid which in the concentration range 2.0-4.0 mol 1-1 increased the barium signal by about 20%. The influence of salts as chlorides on 0.2ng of barium in 0.1 mol 1-1 nitric acid with and without Zeeman-effect back- ground correction is shown in Table 4. Sodium above 1.0 pg and strontium above 0.1 pg caused spectral interferences which were corrected using the Zeeman-effect corrector. The increased barium signal in the presence of calcium in the mass range 0.1-10 pg even with correction of the background may poss- ibly be explained by the favoured formation of calcium mono- cyanide and carbide and by the reduction of the oxygen in the gas phase due to its reaction with calcium.A higher mass of calcium 100 yg caused complete elimination of the barium signal probably because of the occlusion of barium in the calcium matrix. The increased barium absorbance in the presence of more than 0-1 yg of potassium was relatively high indicating that the barium ionization may be higher than e ~ p e c t e d . ~ ~ . ~ ~ Correction of the background 'does not reduce the positive interference of potassium. Magnesium iron and nickel can be tolerated in the studied mass range even without background correction.In conclusion of the studied salts high masses of potassium and calcium caused non-spectral interferences. Table 3 Relative integrated absorbance of 0.2 ng of barium in the presence of acids (n=2) Acid concentration moll- 0.0 0.1 0.2 0.5 1 .o 2.0 3.0 4.0 Acid Nitric Hydrochloric Phosphoric Sulfuric Perchloric 100 100 100 100 100 105 102 - - - - - - - - 92 103 99 113 95 95 102 108 108 96 101 96 116 112 109 107 99 128 85 111 109 98 121 91 90 Journal of Analytical Atomic Spectrometry April 1995 Vol. 10 333Table 4 Relative absorbance of 0.2 ng of barium in 0.1 moll-' nitric acid using a pyrocoated tube and argon of 99.9% purity without and with Zeeman-effect background correction in the presence of salts as chlorides (n = 2) Na K Ca Fe Ni Sr without 100 104 96 114 138 20 1 - * Cation mass/Clg 0 0.001 0.01 0.1 1 10 100 Mass ratio cation/barium 0 5 50 500 5 x 103 5 x 104 5 x 105 without 100 99 107 115 116 126 0 with 100 95 100 124 130 208 0 without 100 101 98 116 119 123 124 without 100 101 102 100 115 99 - * without 100 98 101 95 97 114 - * with 100 105 102 109 108 109 96 without 100 110 116 119 165 186 168 with 100 94 114 106 162 188 * - without with 100 100 99 99 104 90 115 104 101 103 120 116 80 -* with 100 100 94 96 104 103 - * with 100 106 106 96 90 106 - * with 100 99 102 98 115 108 - * * Not determined. 7 8 9 10 11 12 13 Sherfinski J.H. At. Absorpt. Newsl. 1975 14 26. Newbury M. L. Can. SOC. Forensic Sci. J. 1980 13 19. Lieu V. T. and Woo D. H. At. Spectrosc. 1980 1 149. Jasim F.and Barbooti M. M. Talanta 1981 28 353. Parigi J. S. At. Spectrosc. 1982 3 126. Beaty R. D. and Cooksey M. M. At. Absorpt. Newsl. 1978,17,53. Rollemberg M. C. E. and Curtius A. J. Mikrochim. Acta Part ZZ 1982 441. Sugiyama M. Fujino O. and Matsui M. Bunseki Kagaku 1984 33 E123. Hoenig M. Dehairs F. and Kersabiec A. J. Anal. At. Spectrom. 1986 1 449. Roe K. K. and Froelich P. N. Anal. Chem. 1984 56 2724. Epstein M. S. and Zander A. T. Anal. Chem. 1979 51 915. Bishop J. K. B. Anal. Chem. 1990 62 553. Zatka V. Anal. Chem. 1978 50 538. Curtius A. J. Schlemmer G. and Welz B. J. Anal. At. Spectrom. 1987 2 311. Carnrick G. R. and Slavin W. J. Anal. At. Spectrom. 1988 3 1023. Campbell W. C. and Ottaway J. M. Talanta 1974 21 837. L'vov B. V. in XXI Colloquium Spectroscopicum Znternationale and 8th Conference on Atomic Spectroscopy Cambridge Heyden London 1979 ch.12 p. 152. L'vov B. V. Nikolaev V. G. Norman E. A. Polzik L. K. and Mojica M. Spectrochim. Acta Part B 1986 41 1043. Styris D. L. Anal. Chem. 1984 56 1070. L'vov B. V. Spectrochim. Acta Part B 1990 45 633. Handbook of Chemistry and Physics ed. Weast R. C. 49th Edn. Chemical Rubber Co. Cleveland OH 1968-69 p B-259. Nadler M. R. and Kempter C. P. Rev. Sci. Znstrum. 1961,32,43. Hall B. F. and Spooner N. F. ISA Trans. 1965 4 355. Analysis of Real Samples As matrix effects were verified for ions usually present in real samples even using the Zeeman-effect corrector in the analysis of the two certified materials calibration was performed using the standard additions method with three replicates. The concentration obtained for barium in NIST SRM 1572 Citrus Leaves was 21.1k0.5 ppm which agrees with the certified value of 21 & 3 ppm. For the sea-water sample (NASS-2) the determined concentration was 20 & 3 fig 1- '. Roe and Froelich,16 using a continuous background corrector chemical modifier and the standard additions method found barium concentrations in the range 5.4kO.9-21.3 k2.8 pg 1-' for waters from the Central Northwest Pacific Ocean of different depths. The detection limit (signal-to-noise ratio = 3) for barium in 0.1 moll-' nitric acid was 25 pg using 10 p1 of the solution. 14 15 16 17 18 19 20 21 22 23 The authors are grateful to the Brazilian Science and Technology Ministry and to the Research Center of Petrobras for financial support. 24 REFERENCES 1 Edelbeck L. and West P. W. Anal. Chim. Acta 1970 52 447. 2 Szydlowski F. J. and Vianzon F. R. Anal. Lett. 1978 B1 161. 3 Wolgemuth K. and Broecker W. S. Earth Planet. Sci. Lett. 1970 8 372. 4 Bacon M. P. and Edmond J. M. Earth Planet. Sci. Lett. 1972 16 66. 5 Bender M. Sned T. Chan L. H. Bacon M. P. and Edmond J. M. Earth Planet. Sci. Lett. 1972 16 81. 6 Thompson K. C. and Godden R. G. Analyst 1975,100 198. 25 26 27 28 29 Paper 4/00034 J Received January 5 1994 Accepted August 17 1994 334 Journal of Analytical Atomic Spectrometry April 1995 Vol. 10

ISSN:0267-9477    

DOI:10.1039/JA9951000329

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

Studies of Jet Configurations for Jet- enhanced Sputtering Devices HYO J. KIM YANG S. PARK AND JUNG H. CHO Department of Pharmacy Dong Duck Women' University Seoul 136-71 4 Korea GAE H . LEE Department of Chemistry Chung Nam National University Daejon 305-764 Korea KYU H . CHO KEE B. LEE AND HA S . KIM Department of Chemistry Seoul National University Seoul 152-742 Korea A modified gas jet-enhanced glow discharge device for atomic spectrometry has been developed. Three types of jet configurations were constructed and investigated six-jet cone- jet and cylinder-jet. The discharge current pressure and gas flow rate were studied for optimum discharge conditions. The sample mass loss for a Cu sample increases three-fold using a six-jet and cone-jet configuration with a gas flow rate increase of 0-200 sccm.However the sample mass loss increases from 0.8 to 1.5 mg as the gas flow rate increases from 0 to 200 sccm with a cylinder-jet configuration. The maximum absorbance was obtained with a cone-jet configuration at 1.7 mbar 150 sccm and 30 mA discharge current. Keywords Jet-enhanced glow discharge device; atomic absorption; atomic emission; sputter rate The use of gas jets in sputtering makes this direct sampling method practical for atomic absorption spectrometry (AAS) measurements. An atomizer called the Atomsource (Analyte Medford OR USA) uses six high-velocity gas jets that help not only to reduce the re-deposition but also appear to increase the rate of removal of sample material before re-deposition is considered as evidenced by the craters in the sample surface where gas jets are directed.These systems have been applied primarily to AAS applications but to some extent to atomic emission spectrometry (AES) and mass spectrometry (MS). Ohlsl used the Atomsource for the analysis of metals and alloys by AAS. Winchester and Marcus2 analysed non- conducting powder samples by mixing the non-conductor with a conducting powder. Kim and Piepmeier3 investigated the effect of gas flow rate on sputtering characteristics by measuring atomic absorption emission and sample loss rate. They reported an increase in the sample loss rate and atomic absorption signal with increase in gas flow rate. Banks and Blades4-* studied the characteristics of a jet-assisted glow discharge source for AES. At constant pressure and power the gas flow increased the sample loss rate by a factor of 3 times compared with no flow.However some emission lines were decreased in intensity (large absorption coefficients) while others were. enhanced (small absorption coefficients). They concluded that this behaviour is suggestive of a self-absorption problem. These suggestions were supported by Chakbrabarti and c o - w o r k e r ~ ~ * ~ ~ who investigated the effects of changes in applied power discharge gas pressure and flow rate on atomic absorption and emission signals obtained from a jet-assisted cathodic sputtering atomizer. A six-jet configuration makes a typical pattern of sputtering craters on the sample surface that may not be appropriate when more uniform surface information is needed such as in-depth analysis.Therefore a six-jet cone-jet and cylinder-jet Journal Journal of Analytical Atomic Spectrometry configuration were constructed and tested for their perform- ances. The effect of pressure current and gas flow rate on atomic absorption emission and sample loss rate were studied with these three types of jet configurations. EXPERIMENTAL Glow Discharge-AES System Fig. 1 is a schematic diagram of the glow discharge-AES (GD-AES) system which consists of three components a GD cell spectrometer and control unit. The power supply was an Emace 2000 (Seoul Korea) which is capable of supplying up to 2000 V and 250 mA. The power supply is controlled either manually or by a computer. The sputtering gas of 99.995% pure argon was used with a moisture trap (RGF-125-400 Labclear Oakland CA USA) to remove water vapour from the sputtering gas.The sputtering gas was passed through the mass flow controller when the power of a 3-way solenoid valve was on. However the gas was directed into the discharge cell when the power of the 3-way solenoid valve was off to flush the sample surface after exchanging the sample. The solenoid valve (CKD Tokyo Japan AG43-02-5) was automatically operated by a solid state relay & relay driver board (PCLD-786 Advantech Taiwan). The operating flow rate of the argon gas between 0 and 2000 sccm was controlled by a mass flow controller (MKS Andover MA USA). The pressure inside the sputtering cell was monitored on the outlet side using a Pirani gauge (Balzers TPG300 Furstentum Liechtenstein).To measure the pressure difference between the inlet side of the jet configuration and the inside of the discharge chamber an additional pressure gauge (gauge 1 ) was used. The pressure was independently controlled with the gas flow rate by a Whitey regulating valve installed between the cell and vacuum pump. A two stage vacuum pump of 450 1 min-' (Varian SD450 Lexington MA USA) with an oil mist filter provided a sufficient vacuum inside the discharge cell. Modified Gas-jet Sputtering Cell The solid sample atomizer shown schematically in Fig. 2 consists of a Teflon chamber with a de-mountable quartz window at each end and an emission window at the front of the sample surface. A water cooled copper cathode plate on the side of the chamber helps to stabilize the sample tempera- ture during analysis.The flat sample is sealed over a 0.9 cm diameter hole in the cathode plate with an O-ring. Any piece of metal or alloy with a 2.5 cm minimum flat surface may be analysed directly. To measure the emission from the sputtering chamber an emission window brass extension tube 60 cm long and 2cm in internal diameter with a de-mountable quartz of Analytical Atomic Spectrometry April 1995 Vol. 10 335- ressure gauge 1 Dc power supply Gas filter Solenoid valve Fig. 1 Schematic diagram of glow discharge atomic emission spectrometry system Emission H E 0 CD II Absorption x wl D 4.. 114 mm Fig. 2 Schematic diagram of glow discharge atomizer where A cathode; B sample; C anode; D arrestor (Macor); E gas inlet block; F cell body; G absorption window; H emission window; and I vacuum port window at the end was used to prevent the deposition of sputtered material on the window.Fig. 3 is a schematic diagram of a six-jet cone-jet and cylinder-jet configuration. The six-jet configuration has six jet channels of 0.5 x 0.5 mm at the anode and is formed with the insulator so that the high velocity gases are directed toward the sample surface at an angle of 60". The cone-jet and cylinder- jet configurations have a 0.5mm radial gap between the jet nozzle and the insulator so that high velocity gases are directed to the sample surface at an angle of 60" for the cone-jet and 90" for the cylinder-jet configuration. Measurement System For the atomic emission measurements a 1.0 m Czerny-Turner grating monochromator (Spex Model 1000M Edison NJ USA) with a Hamamatsu R928 PMT was used.Thef/8 1.0 m monochromator employs a 110 mm square plane grating (blazed for 500 nm) with 1200 lines mm-l giving it a reciprocal linear dispersion of 0.8 nm mm-' in the first order. Emission from the discharge cell is transferred to the monochromator with a 25 mm diameter symmetric convex spherical silica lens. The nominal focal length of this f/6.0 lens is 150 mm. Data acquisition and monochromator control were performed with a DSlOOO (Spex Edison NJ USA) and homemade software written in Quick BASIC language. For the atomic absorption measurements the modified glow discharge atomizer was installed in a commercial AAS instru- ment (Varian Model Spectra AA-30 Techtron Pty. Victoria Australia) in place of a flame burner.The absorption line chosen for the analysis was Cu I (324.7 nm) for a resonance line and Cu I(521.1 nm) for a non-resonance line. The hollow cathode lamp (Varian Victoria Australia) current was 4 mA and the slit-width of the monochromator was 0.2 nm. Sample Loss Measurement Thin oxygen-free copper sheets 30x 30mm in area and 0.60mm thick were sputtered and the amount of loss was determined by weighing on an analytical balance. The net exposed area was circular and 0.8 cm in diameter. Various combinations of current pressure gas flow rate and sputter time with the three types of jet configurations were used. RESULTS AND DISCUSSION Voltage-Current Characteristic Curves The effect of gas flow rate and pressure on the voltage-current relationship was studied with the six-jet configuration as shown in Fig.4. The discharge voltage decreases from 600 to 300 V as pressure increases from 2 to 6mbar at 40mA discharge current and 50 sccm gas flow rate. The discharge voltage also decreases from 600 to 450 V as the gas flow rate increases from 50 to 400 sccm at 2 mbar and smaller voltage decreases are obtained with increases in gas flow rate from 50 to 400 sccm at 6mbar pressure. Kim and Piepmeier3 mentioned as the reason for this decrease in discharge voltage as the gas flow rate increases that a gas jet provides a locally high pressure region between the cathode and anode; higher pressures require lower cell voltage. This assumption was supported by ~ t h e r s ~ . ~ . ~ and recently Lazik and Marcus" measured the reduction of the dc bias voltage in an rf glow discharge system when the gas flow rate increased at constant pressure.They conclude that the effect of gas flow rate on the dc bias would seem to be pressure related where increasing gas flow provides i l higher localized pressure at the cathode surface. 336 Journal of Analytical Atomic Spectrometry April 1995 Vol. 10Gas inlet I---= J Anode MacorR I G a q T [ k I I 1 Cathode (sample) I I Anode Six-jet configuration ( b) 33.0 mm C) v 23.0 mm u I-. / I 9.0 mm dTb 3 Gap 0.5 mm Cone-jet configuration C y I i n d e r-jet co nf i g u r a t i on Fig. 3 Schematic diagram of various shapes of jet configuration (a) 6-jet configuration; (b) cone-jet configuration; and (c) cylinder-jet configuration 500 ;5 10 15 20 25 30 35 40 CurrenttmA Fig.4 Current-voltage characteristic curves for the six-jet configur- ation at A 2 mbar 50 sccm; B 2 mbar 400 sccm; C 6 mbar 50 sccm; and D 6 mbar 400 sccm Fig. 5 shows the voltage-current relationship with various jet configurations. There is little difference among the jet configurations with a discharge current of c25mA. At > 25 mA the cylinder-jet configuration shows the lower dis- charge voltage at constant current than the cone and six-jet 1000 1 I A l 5 10 15 20 25 30 35 40 C u rren ttm A Fig. 5 Current-voltage characteristic curves for A cone-jet; B 6-jet; and C cylinder-jet configurations at 1.5 mbar 150 sccm gas flow rate and copper sample configurations. These voltage-current differences among jet configurations might be due to the change in the flow pattern at the discharge region.Mass Loss Experiment Fig. 6 shows the effect of sputtering time on sample mass loss with various jet configurations at 3 mbar 30 mA and 200 sccm gas flow rate. The sample mass loss increases as the sputtering time increases up to 30 min with all the jet configurations. The six-jet configuration shows the highest sample mass loss followed by the cone-jet and cylinder-jet configurations which shows the lowest sample mass loss. The relative standard deviation of the mass loss measurements for 30 min sputtering time for 10 replicates is 5%. Fig. 7 shows the effect of discharge current and pressure on the sample mass loss at a constant gas flow rate of 150 sccm. Sample mass loss increases as the current increases at a constant pressure with all the jet configurations.At a constant pressure and gas flow rate the sample mass loss increases as the current increases owing to the net power increase. At a constant current and gas flow rate the sample loss rate increases as the pressure decreases. In order to maintain a constant current while the pressure is decreased at a constant I A 0 10 20 Ti me/m i n 30 Fig.6 Effect of sputtering time on sample mass loss for A 6-jet; B cone-jet; and C cylinder-jet configurations at 3 mbar 30 mA and 200 sccm gas flow rate with copper sample Journal of Analytical Atomic Spectrometry April 1995 VoZ. 10 33730 2mbar ,+ A '2 Cur rentlm A Fig.7 Effect of discharge current on sample weight loss for two different pressures at 200sccm gas flow rate 10min sputtering time with copper sample where A cone-jet; B 6-jet; and C cylinder-jet configurations gas flow rate the discharge voltage must increase. The sample mass loss varies little with the various jet configurations at 4mbar pressure however the sample loss rate is enhanced with the cone-jet configuration at the pressure of 2 mbar.Even though the voltage decreases at a constant current (the same as the power decreases) with an increase in gas flow rate the mass loss increases up to a certain level. Therefore sample loss rate has been measured to investigate the effect of gas flow rate by several author^?^^^",'^ Fig. 8 represents the effect of gas flow rate on mass loss variation with three different jet configurations at a constant pressure and constant current.Sample mass loss increases almost three times when gas flow rate increases from 20 to 200 sccm with the six-jet and cone- jet configurations. Then sample mass loss decreases about 30% as the gas flow rate further increases up to 900sccm. In the case of the cylinder-jet configuration sample mass loss increases about two times when the gas flow rate increases from 20 to 200 sccm followed by a decline of about 20% with further increases in gas flow rate. Initial increase in sample mass loss with gas flow rate up to 200 sccm is consistent with other authors r e s ~ l t s . ~ ~ ~ ~ " * ' ~ Br oekaert et al.l2 measured the ablation rates with a 0.2 and 0.5mm jet diameter with gas flow rate. Ablation rates increased up to 3 times with increases of gas flow rate from 70 to 210 sccrn and much more rapidly when the 0.5mm jets were employed.The reason is that an increase in the gas flow rate reduces the re-deposition; however further decreases in sample loss rate were attributed to the localized high pressure where the gas jets'are directed onto the sample surface which reduced the discharge voltage. This was verified by Lazik and Marcus'' by measuring the dc bias voltage with different gas flow rates in an rf glow discharge system. From the previous results the differences in sample mass loss due to the shapes of jet configuration may be A ... . .. -f . . . . . . . . . . . . . I I I I 0 200 400 600 800 1000 Ar flow rate/sccm Fig.8 Effect of gas flow rate on sample weight loss for A 6-jet; B cone-jet; and C cylinder-jet configurations at 3 mbar 30 mA and 10 min.sputtering time with copper sample 338 Journal of Analvtical Atomic Svectrometrv. Auril 1995. attributed to the difference in flow pattern for 0 to 200 sccm and localized pressure changes. Under conditions of constant current and voltage assuming a constant sputtering ratio the main reason for increase in sample loss rate with increase in gas flow rate is the reduction in re-deposition. Absorbance Measurements The main difference between the Grimm type and Atomsource was that the Atomsource used the gas flow and enhanced both sampling efficiency and transport efficiency. Kim and Piepmeier3 showed that for the Atomsource the absorption signal for Cu increased with argon flow rate and reached a maximum of 200 sccm at constant pressure and current. Hutton et ~ 2 .' ' also investigated the effect of gas flow rate on the change in the normalized absorbance signals of Cu Cr and Ni at a constant power. The absorbance signal showed an initial increase with increasing gas flow rate up to 500sccm after which all curves levelled off. At flow rates >lo00 sccm all absorbance signals decreased because of a decreased residence time for atoms within the analysis volume. They concluded that the transport efficiency of atoms is at a maximum at a flow rate of 500 sccm for a pressure of 10 Torr. The reason for the signal maximum occurring at a different flow rate between the Atomsource and the Hutton's cell is most likely due to the fact that they used a different cell design and cell volume.To investigate the optimum gas flow rate and pressure for the maximum sensitivity of our developed cell the absorption was measured for various jet configurations. Fig. 9 shows the absorbance of Cu 1324.7 from the stainless- steel sample at various currents for four different pressures at a constant gas flow rate with various jet configurations. Absorbance increases as discharge current increases for all jet A 0.08 1 0.06 0.02 L 0 10 20 30 40 50 60 Cu rrent/mA Fig. 9 Effect of discharge current on Cu 324.8 nm absorbance for A 1.7 mbar; B 2.5 mbar; C 3.0 mbar; and D 4.0 mbar at 150 sccm gas flow rate with steel sample 0.06 0.05 $ 0.04 $ 0.03 4 0.02 0.01 0 C $ 0 =' B 1.7 m b m 4.0 mbar / 0 100 200 300 400 Ar flow rate/sccm Fig. 10 Effect of gas flow rate on Cu 324.8 nm absorbance for A 6-jet; 13 cylinder-jet; and C cone-jet configurations at 1.7 mbar and 4.0 mbar pressure 40 mA discharge current with steel sample VOl.10configurations with the exception of the pressure at 1.7 mbar with a cylinder-jet configuration in the range 40-50mA. An increase in the applied current at a constant pressure and gas flow rate causes an increase in the absorbance most likely because of an increase in the amount of sputtered analyte atoms. The absorbance decreases as the pressure increases for the cylinder-jet and cone-jet configurations presumably because of increased re-deposition of sputtered atoms. However in the case of the six-jet configuration shown in Fig. 8 curve A absorbance does not change with the variation of the pressure.Fig. 10 shows the absorbance as a function of gas flow rate at a constant current with 1.7 and 4.0mbar pressures with various jet-configurations. Absorbance increases as gas flow rate increases up to 150sccm at 1.7mbar followed by a decrease or flat region with further increases in gas flow rate. Compared with other jet configurations absorbance with a cone-jet configuration was enhanced about two times at 150 sccm gas flow rate and 1.7mbar pressure. The maximum absorbance for each pressure and gas flow rate with a constant current was obtained when the gas flow rate was between 100 and 200sccm and 1.7mbar pressure with the cone-jet con- 150000 v) C 3 c .- 2 E 3 Y looooo h u) a C 0 v) c .- E 50000 .- .- 'E u1 0 10 20 30 40 50 60 C urrent/m A Fig.11 Effect of discharge current on Cu 324.8 nm emission intensity for two pressures at 150 sccm gas flow rate with steel sample where A 6-jet; B cone-jet; and C cylinder-jet configurations 4000 - .r = 2000 L. e -2 0 c .- s v figuration. Thus to obtain the maximum sensitivity it is necessary to control the argon gas flow rate independently of the argon gas pressure." Emission Measurements To investigate the effect of experimental parameters on self absorption with several jet configuration shapes the emission intensity of Cu I 324.7nm as a resonance line and Cu I 521.82 nm as a non-resonance line were chosen. Gas flow rate pressure and current were chosen as experimental parameters. Fig. 11 shows the emission intensity of the Cu I 324.7 nm from a brass sample at various currents for two different pressures at a constant gas flow rate of 200 sccm with the various jet configurations.With all jet configuration emission intensities increase as current increases with all pressures. Fig. 12 shows the effect of gas flow rate on Cu emission intensities of Cu I 324.7 nm and Cu I 521.82 nm at 3 mbar and 30mA with (a) six-jet (b) cone-jet and (c) cylinder-jet configuration. Fig. 12 (a) shows an initial significant increase followed by a decline in the emission intensities of Cu I 324.7 nm and Cu I 521.82 nm with increasing gas flow rate. The maximum emission intensity of Cu I 521.82 nm is obtained at 100 sccm however the maximum emission intensity of Cu I 324.7 nm is obtained at 50 sccm. Fig. 12 (b) and (c) shows similar shapes with the six-jet configuration for Cu I 521.82 nm.However the emission intensity of Cu 324.7 nm decreases as gas flow rate increases and the maximum emission intensity is obtained at 50 sccm. Compared with the absorption graphs this is an interesting result because the maximum emission intensity is expected at the gas flow rate of maximum sample loss rate. Both emission lines show a maximum for the six-jet configuration but at a lower flow rate than the maximum in the sample loss rate. This difference may be caused by a difference in excitation conditions for different flow rates. For example high flow rates may cool the excitation in the region where the atom concentration is high. However the resonance line of 324.7 nm for the cone-jet and cylinder-jet configurations shows a continuous decrease in emission intensity as the gas flow rate increases up to 100sccm while the non-resonance It \ \ 1000 800 600 400 \ .I I I 0 200 400 600 800 1000 ~ ~~~~ Ar flow rate/sccm Fig. 12 Effect of gas flow rate on A Cu 324.754 nm and B Cu 521.82 nm emission intensity at 30 mA discharge current and 3 mbar pressure with steel sample for (a) 6-jet; (b) cone-jet; and (c) cylinder-jet configurations Journal of Analytical Atomic Spectrometry April 1995 Vol. 10 339line of 521.8 nm still shows a maximum. This may reveal a possible self-absorption problem for the cone and cylinder-jet configurations. The cone-jet configuration shows similar sputtering charac- teristics to the six-jet configuration type. However the cone- jet configuration produces a more uniform sputtering pattern and can be adapted more easily to in depth analysis. However the cone-jet shows more serious self absorption than the six- jet configuration Tor resonance lines.Further studies on the stability calibration depth resolution and spectral line emis- sion under various conditions will be reported in another paper. Acknowledgment is made to the Korean Science and Engineering Foundation under grant number 9 13-0304-006-2. REFERENCES 1 2 Ohls K. Z . Anal. Chem. 1987 327 118. Winchester M. R. and Marcus R. K. Appl. Spectrosc. 1988 42 941. 3 4 5 6 7 8 9 10 11 12 Kim H. J. and Piepmeier E. H. Anal. Chem. 1988,60,2040. Banks P. R. and Blades M. W. Spectrochim. Acta Part B 1989 44 1 1 17. Banks P. R. and Blades M. W. Spectrochim. Acta Part B 1991 46 501. Banks P. R. and Blades M. W. Spectrochim. Acta Part B 1992 47 1203. Banks P. R. and Blades M. W. Spectrochim. Acta Part B 1992 47 1287. Banks P. R. and Blades M. W. Spectrochim. Acta Part B 1992 47 1435. Chakrabarti C. L. Headrick K. L. Hutton J. C. Bertels P. C. and Back M. H. Spectrochim. Acta Part B 1991 46 183. Hutton J. C. Chakrabarti C. L. Bertels P. C. and Back M. H. Spectrochim. Acta Part B 1991 46 193. Lazik C. and Marcus R. K. Spectrochim. Acta Part B 1993 48 863. Broekaert J. A. C. Bricker T. Brushwyler K. R. and Hieftje G. M. Spectrochim. Acta Part B 1992 47 131. Paper 4/01 874E Received March 29 1994 Accepted October 28 1994 340 Journal of Analytical Atomic Spectrometry April 1995 Vol. 10

ISSN:0267-9477    

DOI:10.1039/JA9951000335

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

CUMULATIVE AUTHOR INDEX JANUARY-APRIL 1995 Aboal-somoza Manuel 227 Adams Freddy C. 111 Allen Lloyd A. 267 Arruda Marco A. Z. 55 Barciela-alonso M. C. 247 Barnes Barbara S. 177 Becker-Ross Helmut 61 127 Belazi Abd Ulhafid 233 Bermejo-barrera Adela 227 247 Bermejo-barrera Pilar 227 247 Boonen Sylyie 81 Bordel-Garcia Nerea 3 1 1 Brueggemeyer Thomas W. 177 Bryant M. F. 295 Bulska Ewa 49 Burden Trevor J. 259 Camara Carmen 321 Caruso Joseph A. 7 Gastle Laurence 303 Cernohorsky TomaS 155 Chan W. T. 295 Cho Jung H. 335 Cho Kyu H. 335 Cimadevilla Enrique Alvarez- Cloud Jacques 287 Corns Warren T 287 Cossa Daniel 287 Crews Helen M. 303 Crowe John B. 177 Curtius Adilson Jose 329 Dams Richard 81 Davidson Christine M 233 241 de la Calle Guntinas M. Beatriz 111 321 Ebdon Les 317 Ebihara Mitsuru 25 Efstathiou Constantinos E.221 145 Cabal 149 Ek Paul 121 Fairman Ben 281 Fell Gordon S. 215 Ferron-novais M. 247 Florek Stefan 61 127 145 Fordham Peter J. 303 Furuta Naoki 25 Gallego Mercedes 55 Golloch Alfred 161 Gomez Gomez M. M. 89 Goodall Phillip 317 Gramshaw John W. 303 Grazhulene Svetlana S. 161 Greenfield Stanley 183 Grotti Marco 325 Halls David J. 169 Harnly James M. 187 197 Harrison Iain 215 Hayashi Yasuhisa 37 Heitkemper Douglas T. 177 Hill Steve J. 317 Houk R. S. 267 Huang Meng-Fen 31 Hulden Stig-Goran 121 Hwang Chorng-Jev 31 Imai Shoji 37 Imbert Jean-Louis 93 Ivaska Ari 121 Jackson Kenneth W. 43 Jantzen Eckard 105 Jedral Wojciech 49 Jiang Shiuh-Jen 31 Jones Phil 281 Keating Gillian E 233 Kerrich Rob 99 Khvostikov Vladimir A. 161 Kim Ha S.335 Kim Hyo J. 335 Kinard W. F. 295 Kirschner Stefan 161 Kotrly Stanislav 155 Lajunen Lauri H. J. 117 Lee Gae H. 335 Lee Kee B. 335 Lerat Yannick 137 Littlejohn David 215 233 241 Lobinski Ryszard 11 1 Lunzer Florian 3 11 Lyon Ian C. 273 Madrid Yolanda 321 Mahmood Tariq M. 43 Mao X. L. 295 Marawi Isam 7 Masera Eric 137 Massart D. Luc 207 Mauchien Patrick 137 Mazzucotelli Ambrogio 325 McCartney Martin 233 McCurdy Ed 303 McLeod C. W. 89 Miyazaki Akira 1 Moens Luc 81 Monteiro Maria In& C. 329 Moreda-piiieiro Antonio 227 Okuhara Kyoichi 37 Olson Lisa K. 7 Paama Lilli 117 Pang Ho-ming 267 Park Yang S. 335 Parry Susan J. 303 Pasullean Benyamin 241 Penninckx Wim 207 Peramaki Paavo 117 Pereiro-Garcia Rosario 3 1 1 Perera Indral K. 273 Perez-Corona M. Theresa 321 Perkins Charles V.253 253 Piiri Lindy 117 Pin Christian 93 Piperaki Efrosini A. 221 Powell J. J. 259 Prange Andreas 105 Qiao Huancheng 43 Radziuk Bernard 127 197 Rodel Gunther 127 Russo R. E. 295 Sadler Daran A. 253 Saito Kengo 37 Sanjuan Jane 287 Sanz-Medel Alfredo 149 281 Schmecher Gisela R. 61 Smeyers-Verbeke Johanna 207 Smith Clare M. M. 187 Stenz Herbert 127 Stockwell Peter B 287 317 Tanaka Toshiyuki 37 Tao Hiroaki 1 Taylor P. D. 259 Telgheder Ursula 161 Telouk Philippe 93 Thomaidis Nikolaos S. 221 Thompson Diana 303 Thompson K. Clive 317 Thompson R. P. H. 259 Tischendorf Reinhard 61 Turner Grenville 273 Uchino Tomonori 25 Valcarcel Miguel 55 Vanhaecke Frank 8 1 Vankeerberghen Peter 207 Wang Jiansheng 7 Warren Arnold R. 267 Wolnik Karen A. 177 Wrobel Katarzyna 149 Xie Qianli 99 311 Journal of Analytical Atomic Spectrometry April 1995 Vol. 10 341

ISSN:0267-9477    

DOI:10.1039/JA9951000341

出版商:RSC    

年代:1995

数据来源: RSC