摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL. 8 74 5 Fructose-6-phosphate Kinase Immobilized on Controlled-pore Glass as a Substrate for Selective Separation of Antimony(1ii) Maria Beatriz de la Calle-Guntihas Yolanda Madrid and Carmen Camara* Departamento de Quimica Analitica Facultad de Ciencias Quimicas Universidad Complutense de Madrid 28040 Madrid Spain Fructose-&phosphate kinase immobilized on controlled-pore glass and packed in a glass microcolumn was used for the selective separation and preconcentration of Sblll. Results obtained for mixtures of trivalent and pentavalent antimony showed that SblIl can be selectively separated from SbV after retention by the enzyme and elution with a 3% v/v lactic acid solution. After the elution step SblIl was determined by electrothermal atomic absorption spectrometry.An enrichment factor of 5 was achieved. Recovery studies of SblIl in de-ionized tap and sea-water gave values of around 75-t5% without any need to control the pH or temperature. The immobilized enzyme retains its activity for at least 200 elution cycles. Keywords Antimony(i1i); speciation; fructose-6-phosphate kinase; inmobilization; electrothermal atomic absorp- tion spectrometry Enzyme assays are highly selective and sensitive. These properties together with simplicity speed and ease of automation make enzyme methods highly promising for use in pharmacology agriculture. the food industry and environmental analysis. There are three major reasons for immobilizing enzymes they offer considerable operational advantages over freely mobile enzymes their chemical or physical properties can be selectively altered and they can serve as model systems for natural in vivo membrane-bound enzymes.Immobilization techniques have been extensively however little attention has been paid to enzyme-based methods for determining the metals that act as enzyme activators or inhibitors. Dolmanova et proposed a method for the enzymic determination of mercury and lead. Antimony at the trace level is a relatively common toxic element. Its derivatives are utilized in a number of industrial processes and in therapeutic agents against several major tropical parasitic diseases. and recently an antimonial compound has been found to be a potentially effective reverse transcriptase inhibitor of possible utility against the AIDS virus.4 Antimony is thought to act by bonding to thiol-containing enzymes and it is well known that antimony metabolism depends on its oxidation state.5 Certain antimony species act by covalently bonding to thiol groups of fructose-6-phosphate kinase the enzyme synthe- sized by the parasites being 80 times more sensitive than the human enzyme nevertheless these antimony species only act in the trivalent oxidation state.6 This paper reports an enzymic method for the speciation of Sb1I1 and SbV based on the enzyme fructose-6-phosphate kinase followed by electrothermal atomic absorption spec- trometry (ETAAS).Attention was initially concentrated on the optimization of the experimental conditions showing the potential capabilities of the method.No great effort has yet been made to understand the mechanism by which SblI* acts on the enzyme; accordingly these preliminary results will be extended with further investigations. Experimental Apparatus A four-channel peristaltic pump (Gilson HP4) was used to pass the samples through the column. A Perkin-Elmer Model I 1 OOB atomic absorption spectro- *To whom correspondence should be addressed. meter equipped with an HGA-400 graphite furnace and a deuterium lamp background corrector was used. A Perkin- Elmer antimony hollow cathode lamp was used as the light source. The spectral bandpass used to isolate the 2 17.6 nm antimony line was 0.2 nm. Argon (flow rate 300 ml min-') was used to purge air from the cuvette. Standards and Solutions All reagents were of analytical-reagent grade or higher purity and purified de-ionized water was obtained using a Milli-Q system (Millipore).A stock standard SbI1' solution (I000 pg ml-l) was prepared by dissolving 0.2740 g of potassium antimony1 tartrate (Carlo Erba) in de-ionized water and diluting to 100 ml. Working standard solutions were prepared daily. A stock standard SbV solution (1000 pg ml-I) was prepared by dissolving 0.2 160 g of potassium pyroantimo- nate (Carlo Erba) in de-ionized water and diluting to 100 ml. Working standard solutions were prepared daily. Calcium Mg Na K and Ni standard solutions for AAS (1000 pg ml-I) were acquired from BDH. Copper Pb Zn and Bi solutions (1000 pg m-l) were prepared by weighing suitable amounts of CU(NO,)~ Pb(N03)2 ZII(NO~)~ and Bi(NO,) (Carlo Erba) respectively; Bi( NO3) was dissolved in concentrated HCl and diluted to a final concentration of 1.2 mol 1 - I HCI.Arsenic(ir1) solution (1000 pg ml-I) was prepared by dissolving an appropriate amount of As,03 (Carlo Erba) in de-ionized water adding 10 ml of 1 mol 1 - I NaOH and diluting to a final volume of 250 ml with 2 mol I-' HCI. Arsenic(v) solution (1000 pg ml I - ] ) was prepared by dissolving an appropriate amount of As205-2H20 (Carlo Erba) in concentrated HCI and diluting to a final concentra- tion of 2 moll-' HCl. Selenium(1v) and SeV1 solutions ( 1000 pg ml I-') were prepared by weighing appropriate amounts of Na2Se03 (Aldrich) and Na,SeO (Aldrich) respectively and dissolving in de-ionized water. Fructose-6-phosphate kinase (Type VII from Bacillus stearothermophilus) and controlled-pore glass (CPG-240 80- 120 mesh mean pore diameter 22.6 nm) were obtained from Sigma.Enzyme Immobilization The CPG-240 (0.2 g) was boiled in 10 ml of 5% nitric acid for 30 min filtered on a glass filter washed with de-ionized water and dried in an oven at 95 "C. An aqueous aminoalkylating agent was prepared by adding 1 ml of746 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL. 8 3-aminopropyltriethoxysilane to 9 ml of water and adjust- ing the pH to 3.45 with 5 moll-' HCI. The dried glass was added the pH was re-adjusted to 3.45 and the mixture was kept at 75 "C on a water-bath for 150 min. The alkylami- nated glass was filtered through a sintered-glass filter (porosity G3) and washed and dried as before.The alkylamination process was repeated to ensure complete activation of the glass. The dried alkylamino-glass was stored in an air-tight bottle and kept until needed. The cross-linking agent glutaraldehyde (2.5% solution) was prepared by adding 2.5 ml of 50% glutaraldehyde solution (Fluka) to phosphate buffer (0.1 mol l-' pH 7.0) and diluting to 50 ml with the buffer. Alkylamino-glass (0.2 g) was added to 1 ml of the glutaraldehyde buffer in a well stoppered vessel through which argon had been bubbled to remove oxygen. The reaction was allowed to proceed for 1 h at room temperature with brief deoxygenation with argon every 10 min for the first 30 min. The activated glass was washed well with distilled water. Fructose-6-phosphate kinase [ 1 mg 250 U ( 1 U = 16.67 nkat)] was dissolved in 3.0 ml of cold (4 "C) phosphate buffer (0.1 mol l-l pH 8) and added to the activated glass. Argon was bubbled through the solution as before.The solution was kept at 4 "C for 2.5 h. The immobilized enzyme derivative was washed first with cold phosphate buffer and then with cold water to ensure the removal of any unlinked enzyme'. The resulting immobilized enzyme was packed by pumping into a glass column (10 x 2.5 mm i.d.) so that the length of the immobilized enzyme zone was 10 mm. The beds were stored at 4 "C in phosphate buffer (pH 8.0). The activity of the enzyme remained constant during 6 months of storage and 200 cycles could be run through the column without any decrease in the enzyme activity. Procedure A 5 ml volume of antimony solution was passed through the column at a flow rate of 0.2 ml min-' and de-ionized water was passed through the column for 1 min to clean the system.Then 1 ml of 3% v/v lactic acid was passed through the column to elute the retained antimony and the column was cleaned again with de-ionized water before passing a new aliquot of sample. Antimony was determined by ETAAS applying the temperature programme given in Table 1. Results and Discussion Effect of the Experimental Parameters on the Retention and Elution of SblI1 Optimization of sample flow rate Different sample flow rates (0.20 0.40 0.82 1.22 and 1.63 ml min-I) were tested to determine the efficiency of antimony retention; 30 pg 1-l SblI1 solutions were passed through the column and the eluate was measured and compared with a 30 pg 1-l SblI1 standard solution. The results showed that the efficiency of retention slowly decreases with increasing flow rate signifying a rapid uptake of Sb1*I by the fructose-6-phosphate kinase.A flow Table 1 Temperature programme for determination of antimony by ETAAS Parameter Drying Ashing Atomization Conditioning Temperature/"C 100 700 2300 2500 Ramp time/s 20 30 0 1 Hold time/s 5 5 2 2 rate of 0.2 ml min-' providing a retention of 75% was chosen for further experiments. Influence of lactic acid concentration Antimony reacts with a-hydroxy acids to form com- p l e x e ~ . ~ ~ ~ In previous ~ o r k ~ * ~ lactic citric and malic acids were used to form antimony complexes. Of these citric acid seems to form the strongest complex with antimony but it is well known that citric acid is a powerful inhibitor of the enzyme fructose-6-phosphate kinase and for this reason and to avoid possible destruction of the enzyme it was rejected for use as the eluent. Of lactic and malic acid the former provides faster kinetics of reaction and forms a stronger antimony complex than the latter.Based on these facts an attempt was made to elute SblI1 with lactic acid and promising results were obtained. Inorganic acids were not used as eluents as they would probably damage the enzyme and so the lifetime of the column would be reduced. Six different lactic acid concentrations were used to elute the SblI1 retained by the enzyme 0.5 1 3,6,8 and 10% v/v. A constant recovery of 75% was achieved in all instances and a concentration of 3% v/v was adopted in further experiments.Study of lactic acid flow rates used to elute SH1' Lactic acid flow rates of 0.20 0.40 0.82 and 1.22 ml min-l were tested to determine their influence on the efficiency of elution. The eluates obtained were measured and the absorbances were compared with those of a 30 pg 1 - I SblI1 standard solution in 3% v/v lactic acid. The elution efficiency slowly decreased with increasing flow rate. A lactic acid flow rate of 0.2 ml min-I providing a 75% recovery of SbIll was chosen for further experiments. The fact that no significant reduction was detected at high flow rates suggests fast kinetics for the elution process. Lower flow rates were not used because they entailed prolonged elution times.Influence of ionic strength The effect of salinity (ionic strength) was tested by adding increasing amounts of Mg(N03)* to the sample up to final concentrations of 0.001 0.01 and 0.1 mol 1-l and the recovery was compared with that of SblI1 solutions in de- ionized water. There was no salinity effect as all the solutions provided the same efficiency of retention. It can be deduced from the results obtained that the experimental parameters studied have little influence on the efficiency of either the retention or elution processes. Study of SbV Interference Once the optimum conditions for the retention of SbrI1 on the enzyme had been established solutions of 30 pg I-' of SbV were passed through the column under the optimized conditions and measurements before and after passage through the column showed a 25% retention of SbV.As the elution of the retained SbV with 3% v/v lactic acid was quantitative it was concluded that this 25% retained SbV would be eluted together with SblI1 when the solutions were prepared in de-ionized water. When spiked samples of SbV in sea- and tap water were analysed the retention of the SbV in the column was only about 10%. These results suggest that the recovery of SbV (and also that of Sblll which decreased to 65% for spiked sea-water samples) depends on the complexity of the matrix which involves a prior knowledge of antimony retention in order to obtain an accurate SbIil recovery correction factor or the application of the standard additions method.747 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL.8 loo 75 A $ 50 00 a 25 0 5 10 15 20 25 Sample volume/ml Fig. 1 Effect of sample volume on antimony retention efficiency total amount of SblI1= 150 ng; sample flow rate 0.2 ml min-I; eluent flow rate 0.2 ml min-I; and eluent 3% v/v lactic acid solution So far the most widely applied methods for Sblll and SbV speciation are those based on liquid-liquid extractionsJO and further determination by ETAAS or by hydride generation (HG) AAS under pH contro1,11J2 which allows the direct determination of Sblll by stibine generation and indirect determination of SbV as the difference between total antimony and Sb1I1. Normally liquid-liquid extraction is slow and suffers from poor repeatability and HG-AAS usually requires preconcentration of antimony species (e.g.by cold trapping' I ) . Both methods allow Sb1I1 and SbV speciation when they are present at an SbI1I:Sbv ratio of 1:9. The proposed method is very simple and permits the preconcentration and separation of Sb1I1 from the remainder of the matrix without any sample pre-treatment. Efficiency of Preconcentration To determine whether the described method achieves a satisfactory preconcentration of SbllI different sample volumes ( I 2 3 4 5 10 and 25 ml) containing the same total amout of Sblll (150 rig) were passed through the column and the eluate was measured. Fig. 1 shows that a constant recovery of 75% was achieved for up to 5 ml of sample; the efficiency of retention decreased drastically for higher sample volumes. No explanation can be given for this behaviour at present.The next step in determining the efficiency of Sb1I1 preconcentration was to determine the minimum volume of 3% v/v lactic acid required to elute the antimony. Aliquots of 0.5 ml were passed through the column after passing 5 ml of a 50 pg 1-l SblI1 solution. Quantitative recovery of antimony was obtained in the first two aliquots and negligible absorbance was found in the last three. According to these results Sb1I1 can be preconcentrated by a factor of 5. Nevertheless when more complex matrices were analysed the volume that could be run without a decrease in Sblll retention was less than 5 ml; when sea-water was analysed the retention of SblI1 decreased to 55% compared with 65% obtained for a 1 ml sample. The decrease in the retention of Sb1I1 with an increase in sample volume was even more pronounced for tap water when 1 ml was passed through the column the retention was 75% (the same as for de- ionized water) but with 5 ml the eluate gave a negligible signal.This drastic decrease in retention for tap water can be due to interference by tin (see Table 3) which shows similar behaviour; no interference was observed even for an Sb:Sn of 1 100 when a 1 ml sample was passed through the column whereas negligible absorbance was obtained when the procedure was applied to 5 ml of sample. Considering all these facts the standard additions method was recommended for analysis by the proposed procedure. This means that the retention efficiency for Sb1I1 is highly influenced by the complexity of the matrix which could be explained by the competition for the reaction sites (mainly SH groups) on the enzyme by other concomitant elements in the sample and presumably by competitive complexation of the Sblll preventing its reaction with the enzyme.Speciation of SblI1 and SbV in Different Matrices Speciation was carried out in three different matrices de- ionized tap and sea-water. Spiked samples were prepared in these media containing two Sb?Sbv ratios 20:lOO and 40 100. The results (calculated by applying a correction factor for SbV retention of 25% in de-ionized water and of < 10% in tap and sea-water) are given in Table 2. The study was carried out by passing 1 ml of spiked sample and eluting with 1 ml of 3% v/v lactic acid to avoid the effect of sample volume The results given in Table 2 suggest that the recoveries of both SblI1 and SbV depend on the complexity of the matrix being 75 75 and 65% for de-ionized tap and sea-water respectively for Sb1I1 and 25 <10 and <looh for de- ionized tap and sea-water respectively for SbV.Study of Interferences The effect of different ions on the retention of Sb1I1 and SbV was tested and the results are given in Table 3. Most of the ions studied did not affect the retention. Only Ca and Pb had a positive effect on the analytical signal of both antimony species and Nil1 had a positive effect on SbV. The effect of Ca can be attributed to its ability to act as a chemical modifier (this analyte might be retained in the column and eluted together with Sblll; the same explanation could explain the interference of Nil1 in the determination of SbV.The effect of Pb might be a spectral interference due to the proximity of the Sb line at 217.6 nm and the Pb line at 217.0 nm. This explanation was checked by adding known amounts of Pb to different concentrations of SbllI which resulted in a non-specific absorbance in all instances. Among the negative interferents SnIv had a characteristic effect when 1 ml of sample was passed through the column no interference was observed but when 5 ml were passed through a negligible absorbance was provided by the eluate; when an Sb-Sn mixture was measured at an Sb:Sn ratio of 1:lO no interferent effect was detected with 1 or 5 ml of sample. Table 2 Speciation of SbilI and SbV in de-ionized tap and sea-water Antimony added (ppb) Antimony recovered (ppb) 20( SblI') 40( Sb"') 20(Sb111)+ 100(SbV) 40( Sb"') + 1 00( SbV) De-ionized water Tap water Sea-water 15.2+ 1.4 15.0+ 1.2 11.2+0.6 30.3+4.2 29.3+0.7 22.0+ 1.5 12.0+ 3.1 16.0+2.4 11.0+0.6 26.6 + 3.4 28.0+ 1.0 20.7+3.2748 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL.8 ~ Table 3 Interfering effect of different ions in SblI1 and SbV retention Interference (O/O) In t e rferen t Sb:interferent SeIV SeV1 As111 AsV PblI SnIV Bill1 1 ml sample volume 5 ml sample volume HgI1 Na+ K+ Ca2+ Mg2 + A13+ Fe3 + CU" Znll Nil1 C1- *n.i.=No interference was observed. YNegligible absorbance was provided by the eluate. $SbV:Fe = I 100. 1 10 1:lO 1:lO 1:lO 1:lO 1:lOO 1:lOO 1:lOO 1:lO 1:1000 1:lOOO 1:lOOO 1:lOOO 1:lOO 1:lOOO 1:lOO 1:lOO 1:lOO 1:lOOO Sb1I1 (40 ppb) n.i.* n.i.n.i. n.i. n.i. + 35 n.i. -t - 43 n.i. n.i. + 60 n.i. n.i. - 47 n.i. n.i. n.i. n.i. SbV ( 1 PPm) - 49 - 62 - 23 - 28 n.1. + 39 n.i -t - 60 n.i. n.i. + 170 n.i. n.i. n.1. n.1. + 89 n.1. - 7.04 Table 4 Analytical characteristics of the method (n=6) Linear range (ppb) 20- 180 Sensitivity (ppb-') 9.09 x 10-4 Detection limit (30) (ppb) 4.1 Limit of determination (ppb) 9.92 Precision (O/O) 11.55 (for 20 ppb) 2.00 (for 180 ppb) Conclusions Although the results obtained in this study are only preliminary the potential of the new method for the selective determination of Sb1I1 is apparent. This selective determination is based on the retention of SblI1 on fructose- 6-phosphate kinase immobilized on CPG. The proposed method allows the preconcentration of SbIII which is especially important in environmental analysis where the concentration of SblI1 is much lower than that of SbV. However the need for prior knowledge of the recovery of antimony in the different samples is a restriction on the applicability of the method.Further improvements can be expected using more selective substrates for SbIIl which will be discussed in a subsequent paper. The authors thank M.E. de Leon by her collaboration in enzyme treatment and the Spanish Education Ministry and Direccion General de Investigacion Cientifica y TCcnica (Project 88/0094) for providing financial support. Max Gormann is thanked for revising the manuscript. 1 2 3 4 5 6 7 8 9 10 1 1 12 References Zaborsky 0. R. Immobilized Enzymes CRC Press Cleve- land OH 1973. Carr P. W. and Bowers D. Immobilized Enzymes in Analytical and Clinical Chemistry Wiley-Interscience New York 1980. Dolmanova I. F. Schekhovtsova T. N. and Kutcheryaeva U. V. Talanta 1987 34 201. Rosenbaum W. Dormont D. Spire B. Vilmer E. Gentilini M. Griscelli C. Montagnier L. Barre-Sinoussi F. and Chermann J. C. Lancet 1985 i 450. Fowler B. A. and Goering L. in Metals and Their Com- pounds in the Environment. Occurrence Analysis and Biologi- cal Relevance ed. Merian E. VCH Weinheim 1991 pp. Korolkovas A. and Burkhalter J. H. Compendio Esencial de Quimica Farmaceutica Revertk Barcelona 1978. Ledn-Gonzilez M. E. and Townshend A. Anal. Chim. Acta 1990,236 267. de la Calle-Guntiiias M. B. Madrid Y. and Camara C. Anal. Chim. Ada 1991 247 7. de la Calle-Guntiiias M. B. Torralba R. Madrid Y. Palacios M. A. Bonilla M. and CAmara C. Spectrochim. Acta Part B 1992 47 1165. Chung C. H. Iwamoto E. Yamamoto M. and Yamamoto Y. Spectrochim. Acta Part B 1984 39 459. Andreae M. O. Asmode J. F. Foster P. and Van't Dack L. Anal. Chem. 1981 53 1766. de la Calle-Guntiiias M. B. Madrid Y. and Camara C. Anal. Chim. Acta 1991 252 161. 743-750. Paper 2/05642I Received October 22 1992 Accepted February 10 1992

ISSN:0267-9477    

DOI:10.1039/JA9930800745

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

749 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL. 8 Determination of Lead and Copper in Kerosene by Electrothermal Atomic Absorption Spectrometry Stabilization of Metals in Organic Media by a Three-component Solution* lvana A. Silva Reinaldo C. Campos and Adilson J. Curtiust Departamento de Quimica da Pontificia Universidade Catolica do Rio de Janeiro 22453 Rio de Janeiro RJ Brazil Silvia M. Sella Departmento de Quimica da Universidade Federal Fluminense 24220 Niteroi RJ Brazil The instability of copper and lead in kerosene and in analytical organic solutions is the main problem with their direct determination. When propan-1-01 is added to a mixture of kerosene and water a homogenous three- component solution is obtained. It was verified that the metals in this solution placed in an autosampler cup are stable for at least 3 h in a closed calibrated flask and remain stable for at least 24 h.The three-component solutions containing purified kerosene and enriched with metallic ions in aqueous solution or with copper or lead cyclohexanebutyrate or tetraethyllead are also stable. In spite of some differences in the absorption pulses the maximum pyrolysis temperatures characteristic masses and precision are similar for the three-component solutions in comparison to aqueous solutions. Keywords Atomic absorption spectrometry; organic medium; kerosene; copper and lead determination The determination of metals in aviation kerosene is of considerable importance to the petroleum industry because of its effects on the utilization and performance character- istics of the product.Some metals are known to catalyse oxidative reactions degrading the thermal stability of the kerosene and other petroleum fractions. Only low concen- trations of soluble metals can be tolerated by the fuel before its stability decreases to an undesirable extent. The changes in concentration of the metal in organic solutions produced by evaporation and/or deposition of the metal on the walls of the storage container lead to inaccurate results. Karwowska et a1.l and Hulanicki and Bulska2 have reported the instability of copper and lead cyclohexanebutyrates dissolved in chloroform and in p- xylene respectively. de La Guardia and Vida13 have found that the iron complex of ammonium pyrrolidin- 1 -yldithio- formate extracted into isobutyl methyl ketone is only stable for 2 h.Other workers have also reported the instability of many carbamates in organic ~olutions.~-~ Atomic absorption spectrometry is a very common technique for the determination of metals in petroleum products. A review of the analysis of petroleum and petroleum products by this and other related techniques has been published by Sychra et al.7 The direct determination of metals is advantageous in relation to the methods which require sample pre-treatment in that a more rapid analysis can be performed and the risks of contamination and analyte losses are reduced. Owing to its high sensitivity electrothermal atomic absorption spectrometry (ETAAS) is particularly suitable for direct analysis. For samples of petroleum fractions the metal to be determined is very often present in the sample as different organometallic species7 whose identity is usually not known.Several papers have reported the effects not only of the chemical form of the analyte but also of the nature of the organic solvent on the absorbance signal of different metals obtained by ETAAS. Two reviews of these effects have been published including flamesa and electrothermal atomi~ers.~.~ Recently Tserovsky and Arpadjan,'O using modern instrumentation *Presented at the Second Rio Symposium on Atomic Absorption tTo whom correspondence should be addressed. Spectrometry Rio de Janeiro Brazil June 21-22 1992. found no significant difference in the peak shape pyrolysis temperature curves and sensitivities by using ligands with different donor atoms for a number of metals.But the sensitivities were usually different for different solvents and tube surfaces being lower in comparison to the sensitivities in aqueous solution. The differences were not high and were explained by greater spreading of the organic sample in the furnace and deeper penetration into the graphite. Emulsions have been used in the analysis of gaso1ine,''J2 lubricating oils,l3 ointments14 and organic e ~ t r a c t s ~ J ~ ~ * ~ by flame AAS. Generally an oil-in-water emulsion using a suitable emulsifier is formed and aqueous inorganic stan- dards are used for calibration. The addition of aqueous copper solution to a mixture of propionic acid and gasoline produced good sensitivity and stability according to Kar- wowska et a!.' The first attempts to determine metals in kerosene by ETAAS injecting the sample directly into the graphite furnace were not successful because of the instability of the metal in the sample and in the analytical organic solutions.It was decided to try homogeneous solutions of water and kerosene obtained by the addition of propan-1-01 as a common solvent for the determination of copper and lead in kerosene. While an emulsion is strictly a heterogeneous system the three-component solution is a transparent homogeneous system. Experimental Instrumentation The measurements were carried out with a Perkin-Elmer Zeeman 3030 atomic absorption spectrometer equipped with an HGA-600 graphite furnace and an AS-60 auto- sampler. The light sources were hollow cathode lamps (HCLs).The atomic absorption signals were recorded on a PR-100 printer. The conditions used were those rec- ommended by the manufacturer. Pyrolytic graphite coated graphite tubes (Perkin-Elmer B0070699) with L'vov plat- forms made of totally pyrolytic graphite (Perkin-Elmer B0109324) were used. Argon 99.96% was used as purge gas. Polyethylene autosampler cups (Perkin-Elmer B0087056) were used. The temperature programmes are shown in750 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL. 8 Table 1 Temperature programme for copper with use of pyrolytic graphite coated eletrographite tubes with a pyrolytic graphite platform Step Temperature/"C Ramp/s Hold/s Argon flow rate/ml min-' 1 80 1 5 300 2 120 5 10 300 3 1000 10 20 300 4 2300 0 5 0 (Read) 5 2650 1 5 300 6 20 I 5 300 Table 2 Temperature programme for lead with use of pyrolytic graphite coated electrographite tubes with a pyrolytic graphite platform.Modifier 1 pg of tetrabutylammonium dihydrogenphosphate Step Temperat ure/"C Ramp/s Hold/s Argon flow rate/ml min-' 1 80 1 5 300 2 120 5 10 300 3 600 10 20 300 4 1800 0 5 0 (Read) 5 2650 1 5 300 6 20 1 5 300 Tables 1 and 2 for copper and lead respectively. The stabilized temperature platform furnace (STPF) conditions were applied except that no modifier was used for copper. For lead 1 pg of tetrabutylammonium dihydrogenphos- phate in ethanol was used as a modifier. The volumes pipetted onto the platform were 10 pl of the sample solution and 10 pl of the modifier unless otherwise specified.Reagents and Solutions Copper aqueous stock solution 1000 pg ml-I. Prepared from a Titrisol ampoule (Merck No. 9987) by dilution with 0.2% vlv nitric acid. Lead aqueous stock solution 1000 pg ml-I. Prepared from a Fixanal ampoule (Riedel-de Haen No. 38555) by dilution with 0.2% v/v nitric acid. Copper as cyclohexanebutyrate (CuCHB) in oil 1000 pg g-l. Merck No. 15055. Lead as cyclohexanebutyrate (PbCHB) in oil 1000 pg g- I . Merck No. 1505 1. Lead as tetraethyllead (PbEtJ 1 1.68% mlm in an organic mixture. Dupont. Dithizone (Dz diphenylthiocarbazone BDH No. 1 3063) 1 x mol 1-1 in chloroform. Pro analysi Merck No. 0 174. Propan-1-01 (Merck No. 997). Ethanol (Grupo Quimica No. 0 1004). Tetra but ylam mon ium dih ydrogen phosp hate (Sigma No. T1531) diluted with ethanol 100 pg ml-I.Palladium chemical modifier 10 g 1-l (Merck No. 7289). Commercial kerosenes used as samples. Produced by the Brazilian petroleum company Petrobras. PuriJed kerosene (Petrobrcis). Obtained by percolation through bauxite and containing copper and lead below their detection limits. The water was purified by the Milli-Q (Millipore) process. The organic analytical solutions were obtained by dilu- tion of the organic stock solutions with purified kerosene. The aqueous analytical solutions were obtained by dilution of the aqueous stock solution with 0.2% v/v nitric acid in water. The copper dithizonate was obtained by extracting a 2 pg ml-' aqueous solution in a pH 3 (acetic acid-ammon- ium acetate) buffer with a 1 x mol 1-1 dithizone solution in chloroform.The volume ratio was 1:l. The copper dithizonate complex was diluted with purified kerosene. The three-component solution was obtained by adding enough propan- 1-01 (at least 1 1 ml) to homogenize a water-in-oil emulsion (2 ml of 0.2% nitric acid in water and Oe300 t A PPLhd-sa 0 500 1000 1500 Te m per at u r ePC Fig. 1 Pyrolysis temperature curves for copper (atomization temperature 2300 "C) A as copper dithizonate in a three- component solution 0.8 ng of Cu; B as Cu2+ in aqueous 0.2% HN03 0.8 ng of Cu; and C kerosene sample in a three-component solution 5 ml of kerosene). The analyte was introduced into the three-component solution as an organic compound or as the metallic ion in aqueous solution. The resulting pH of the three-component solution was 1.5 and in some experiments was increased to 6.5 with an acetic acid-acetate buffer.Results and Discussion Pyrolysis Temperature Curves Pyrolysis temperature curves for copper are shown in Fig. 1. The curves are very similar for copper as dithizonate in a three-component solution (curve A) and as Cu2+ in aqueous solution (curve B) with a maximum pyrolysis temperature of around 1000 "C. For the kerosene sample in a three- component solution (curve C) the maximum pyrolysis temperature is about 200 "C higher indicating that some component or some copper ligand of the kerosene provides an additional thermal stabilization of the copper in the graphite tube. For lead using 1 pg of tetrabutylammonium dihydrogen- phosphate as modifier the curves are shown in Fig. 2. The solutions studied have about the same maximum pyrolysisJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY.AUGUST 1993 VOL. 8 75 1 BG 0.3 0 400 800 1200 1600 TemperaturePC ( a ) Fig. 2 Pyrolysis temperature curves for lead (atomization temper- ature 1800 "C) A as lead cyclohexanebutyrate in a three- component solution 0.8 ng of Pb; B as Pb2+ in aqueous 0.2% HN03 0.5 ng of Pb; C kerosene sample in a three-component solution; D as Pb2+ in a three-component solution 0.5 ng of Pb; E as PbCHB in a three-component solution 0.43 ng of Pb; and F diluted kerosene sample in a three-component solution. Modifiers 1 pg of tetrabutylammonium dihydrogenphosphate (curves A-C); and 10 pg of palladium (curves D-F) BG 0.1 0 I 1 ( b ) $ BG 0.1 al C a o (c) L BG 0.5 ( d ) 0 9 AA 0.5 d I 0 2.5 5.0 Time/s Fig.3 Absorption pulses for copper (pyrolysis temperature 1000 "C atomization temperature 2300 "C) (a) aqueous 0.2% v/v HN03 0.4 ng of Cu; (b) kerosene sample; (c) kerosene sample in a three-component solution; and (d) as copper dithizonate in a three- component solution 0.4 ng of Cu. Continuous line corrected absorbance; broken line background signal temperature of around 600 "C. The pyrolysis temperature curve for tetraethyllead in a three-component solution not shown in Fig. 2 is very similar to the curve for lead cyclohexanebutyrate under the same conditions. In the same figure curves using 10 pg of palladium as the modifier are also shown. Maximum pyrolysis temperatures were much higher at around 1200 "C showing the superiority of palladium over phosphate in stabilizing lead.For economic reasons in the study of the stability of lead in organic solutions phosphate was used as modifier. Absorption Pulses Absorption pulses for copper in different solutions are shown in Fig. 3. The pulses are very similar. Slightly higher appearance times shown in Table 3 were obtained for the three-component solutions with different copper species in comparison to copper in aqueous solution or in kerosene. Absorption pulses for lead in analytical solutions Fig. 4 show the effect of the modifier delaying the appearance time (Table 3). The appearance times for tetraethyllead in a three-component solution with and without modifier are much lower than for lead cyclohexanebutyrate showing the strong effect of the lead compound.The pulses for lead in the kerosene sample and in a three-component solution containing the same sample are shown in Fig. 5. The modifier effect delaying the appearance times is not so strong as in the analytical solutions (Table 3). The appear- ance times without modifier are higher in the solutions containing the kerosene sample in comparison to the analytical solutions which contain purified kerosene indi- cating an effect of lead ligands in the kerosene sample or of another component not present in the purified kerosene. al C - m o . ..- . . +? - AA 0.3 I I AA 0.3 r 1 0 5.0 Time/s Fig. 4 Absorption pulses for lead (pyrolysis temperature 600 "C; atomization temperature 1800 "C) (a) aqueous 0.2% v/v HN03 0.5 ng of Pb no modifier; (6) as (a) with modifier; (c) as lead cyclohexanebutyrate in a three-component solution 0.43 ng of Pb no modifier; and (d) as (c) with modifier.Modifier 1 pg tetra- butylammonium dihydrogenphosphate. Continuous line corrected absorbance; broken line background signal752 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL. 8 Table 3 Appearance time (faPp) for copper and lead as different species in different solutions with and without modifier. Modifier 1 pg of tetrabutylammonium dihydrogenphosphate Modifier twdS c u 0.40 Cu2+ Aqueous 02% HN03 No 0.73 Kerosene saimple No 0.73 Kerosene in a three-component No 0.76 0.40 CuHDz* Three-component No 0.79 0.34 CuCHB.1. Three-component No 0.82 Analyte Masshg Species Solution Pb 0.50 Pb2+ Aqueous (0.2% HN03) 0.50 Pb2+ Aqueous (0.2% HN03) 0.43 PbCHB Three-component 0.43 PbCHB Three-component 0.35 PbEt,$ Three-component 0.35 PbEt Three-component Kerosene sample Kerosene sample Kerosene in a three-component Kerosene in a three-component No Yes No Yes No Yes No Yes No Yes 0.7 1 0.88 0.59 0.76 0.30 0.50 0.85 0.9 1 0.85 1 .oo * CuHDz=Copper dithizonate.7 CHB = cyclohexanebutyrate. $ PbEt = Tetraethyllead. AA 0.2 BG 0.2 (a) I I n AA0.2 1 L 2 AA 0.2 1 ~e .. .. . 1 AA 0.2 BG 0.2 ( d ) 1 I . _ 1 0 5.0 Ti me/s Fig. 5 Absorption pulses for lead in the kerosene sample (pyro- lysis temperature 600 "C; atomization temperature 1800 "C) (a) no modifier; (b) with modifier; (c) in a three-component solution no modifier; and (d) in a three-component solution with modifier. Modifier 1 ,ug tetrabutylammonium dihydrogenphosphate.Con- tinuous line corrected absorbance; broken line background signal Pulses of about the same area were chosen to avoid the influence of the analyte mass on the appearance time. Absorption pulses for lead using palladium as a modifier were not investigated. Stability of the Metals in Organic Solutions The change in the integrated absorbance of copper with time after placing the solution in an open polystyrene 0.150 I 1 I I 1 0 50 100 150 200 250 Time/min Fig. 6 Stability of copper solutions in the autosampler cup. A copper dithizonate in purified kerosene 0.4 ng of Cu; B CuHDz in a three-component solution 0.4 ng of Cu; C CuCHB in a three- component solution 0.4 ng of Cu; D kerosene sample in a three- component solution; E copper cyclohexanebutyrate in purified kerosene 0.4 ng of Cu; F kerosene sample; and G copper cyclohexanebutyrate in chloroform 0.4 ng of Cu autosampler cup for different solutions are shown in Fig.6. The signal for copper dithizonate (curve A) and for copper cyclohexanebutyrate (curve E) in purified kerosene de- creases linearly in the studied time period. The signal loss is higher for copper cyclohexanebutyrate 44% in 150 min. For the kerosene sample (curve F) the loss is also high 28% in 150 min. Copper cyclohexanebutyrate in chloroform (curve G) shows a loss of 55% in 150 min. These data imply that copper in the kerosene sample and the analytical solutions is not stable. However for the three-component solutions the signal is almost constant with time indepen- dently of the copper species that are present.In a closed 10 ml calibrated flask made of glass 62% of copper as cyclohexanebutyrate in kerosene and 42% in chloroform is lost after 24 h as shown in Table 4. The signal for copper in a kerosene sample decreases 77% in 24 h. Copper dithizonate is stable in kerosene or in chloroform. The different copper compounds in three-component solu- tions are also stable. The stability curves for lead solutions in open auto- sampler cups are shown in Fig. 7. Lead cyclohexanebutyrate in kerosene or in chloroform is not stable Tetraethyllead diluted with kerosene but not in a three-compoundJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL. 8 753 Table 4 Signal loss (integrated absorbance) for copper and lead in solutions kept in a closed 10 ml calibrated flask made of glass for 24 h.Modifier for lead 1 ,ug of tetrabutylammonium dihydrogenphosphate Analyte Species Solution Signal loss (%) 0.3 ng of Cu CuCHB* 0.4 ng of Cu CuHDzt CuCHB CuCHB CuHDz CuHDz cu - 0.43 ng of Pb PbCHB PbCHB PbCHB 0.35 ng of Pb PbEt4$ - PbEt4 Pb2+ Pb2+ Pb - * CHB =Cyclohexanebutyrate. t CuHDz=Copper dithizonate. $ PbEt,= Tetraethyllead. Kerosene CHC13 Three-component Kerosene Three-component Kerosene sample Kerosene in three-component Kerosene T hree-component Kerosene Three-component T h ree-componen t Aqueous 0.2% HN03 Kerosene in three-component CHCI3 CHCI3 62 42 0 0 3 9 77 4 81 49 - 4 99 9 -3 -3 3 u) n qnn -0 Y 0.050 D ?! C - L 1 I I I I 1 120 160 200 240 0 40 80 Time/m in Fig. 7 Stability of lead solutions in the autosampler cup.A kerosene sample in a three-component solution; B lead cyclohexa- nebutyrate in a three-component solution 0.3 ng of Pb; C kerosene sample; D lead cyclohexanebutyrate in purified kerosene 0.43 ng of Pb; and E lead cyclohexanebutyrate in chloroform 0.43 ng of Pb solution is very unstable since the lead absorbance signal is almost zero at the time of the first measurement. The signal for lead in kerosene samples decreases 50% in 150 min. The three-component solutions containing lead cyclohexane- butyrate tetraethyllead (not shown in Fig. 7) and the kerosene sample are stable. In a closed 10 ml calibrated flask as shown in Table 4 there are high losses of lead as cyclohexanebutyrate or as tetraethyllead in kerosene. Lead as Pb2+ and as lead cyclohexanebutyrate or tetraethyllead in three-component solutions is reasonably stable. Whatever the metal compound the three-component solutions are stable.Also Cu2+ or Pb2+ in three-component solutions not shown in Figs. 6 and 7 are stable. The three-component solutions are likely to be water-in- oil microemulsions where submicroscopic droplets be- lieved to have water at their centre are stabilized by a layer of alcohol (co-surfactant) and surfactant.” Formation of detergent-free microemulsions has also been reported.’* Different surfactants (decylammonium chloride cetyltri- methylammonium bromide and Triton X- 100) were also added to the three-component solutions but no further benefits were obtained and it was decided to work with detergent-free solutions. To verify whether the metal extraction into the acidic aqueous droplet could be responsible for stabilization in the three-component solution the pH was increased to 6.5 using an acetic acid-acetate buffer.The stability was maintained in the high pH solution indicating that stability does not depend on the hydrogen ion concentration. Characteristic Mass and Precision The characteristic masses for copper and lead in different solutions are shown in Table 5 (characteristic mass is defined as the analyte mass which gives an integrated absorbance of 0.0044 s). The stock organic solutions were weighed to avoid pipetting errors that could result from their high viscosity. Also the copper and lead concentra- tions in the organic stock solutions were checked by flame AAS after dilution with propan- 1-01 and using analytical solutions containing Pb2+ and Cu2+ also diluted with propan- 1-01.The characteristic masses for each element are similar indicating that about the same sensitivity is achieved in the three-component solution as in aqueous solution independently of the chemical forms of the analytes. The relative standard deviations (RSDs) for the different solutions are shown in Table 6. The three-component analytical solutions and the kerosene sample in the three- component solutions have adequate precision somewhat poorer than that for the metals in aqueous solutions. The very poor precisions for copper and lead in the kerosene sample are mainly due to the instability of these metals in the sample. Copper dithizonate diluted with purified kerosene is also stable and consequently the precision is reasonable.Table 5 Characteristic mass (mo) for copper and lead in different solutions containing copper dithizonate (CuHDz) metal cyclohex- anebutyrate (MCHB) tetraethyllead (PbEt,) or metallic ions. Modifier for lead 1 pg tetrabutylammonium dihydrogenphosphate Analyte Species Solution mdpg Cu CuHDz Three-component 12.8 CuCHB Three-component 15.0 cu2+ Aqueous 0.2Oh HN03 14.0 cuz+ Three-component 15.6 Pb PbCHB Three-component 21.0 PbEt Three-component 20.3 PbZ + Aqueous 0.2% HN03 20.0 PbZ+ Three-component 22.0754 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL. 8 Table 6 Relative standard deviation (RSD) for copper and lead in different solutions containing copper dithizonate (CuHDz) metal cyclohexanebutyrate (MCHB) tetraethyllead (PbEt,) or metallic ions.Modifier for lead 1 pg of tetrabutylammonium dihydrogenphosphate (10 pl); n=20 Analyte 0.8 ng of Cu 0.8 ng of Cu 0.8 ng of Cu c u c u c u 0.5 ng of Pb 0.43 ng of Pb 0.35 ng of Pb Pb Pb Species cuz+ CuHDz CuCHB CuHDz - - Pb2+ PbCHB PbEt - - Solution Aqueous 0.2% HN03 Three-component Three-component Kerosene Kerosene sample Kerosene sample in three-component Aqueous 0.2% HN03 Three-compone n t Three-compone nt Kerosene sample in three-component Kerosene sample Volume/pl 20 20 20 20 20 20 10 10 10 10 10 RSD (O/O) 1.5 1.8 2.0 2.8 15.9 1.7 2.0 3.0 2.8 4.4 24.0 Conclusions Copper and lead are not stable in kerosene samples. Also analytical solutions obtained by diluting metal cyclohex- anebutyrate in standard Merck oil with purified kerosene are not stable.Forming a transparent solution by adding propan- 1-01 to a two-phase mixture of kerosene and water is a convenient way to keep the metals in solution in the autosampler cup and also in a closed glass flask. The sample should be mixed with the propan- 1 -01-water solution in adequate proportions soon after collecting the kerosene from the storage tanks. The use of three-component solutions besides providing adequate stability makes the analyte determination very flexible. For calibration either aqueous analytical solutions or three-component solutions enriched with the metallic ion or with the metal compound can be used. Adequate precision is achieved. Also modifiers in aqueous solutions (more easily available) instead of organic modifiers can be used with the three-component solution as shown for lead.The transparent three-compo- nent solutions can possibly stabilize metals in other organic samples and other metal complexes in different organic solvents by the proper choice of the common solvent and the use of appropriate volume ratios. The authors are grateful to the Brazilian Petroleum Com- pany Petrobras for providing the kerosene samples and also financial support. I. A. S. and S. M. S. had scholarships from the Brazilian Research Agency Co-ordenaciio de AperfeiCoamento de Pessoal de Nivel Superior (CAPES). References 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 Hulanicki A. and Bulska E. XX VI Colloquium Spectroscopi- cum Internationale ed. Tsalev D. L. University of Sofia Bulgaria 1989 vol. VII 57. de La Guardia M. and Vidal M. T. Talanta 1984 31 799. Tao H. Miyazaki A. Bansho K. and Umezaki Y. Anal. Chim. Acta 1984 156 159. Danielsson L.-G. Magnusson B. and Westerlund S. Anal. Chim. Acta 1978 98 47. Hulanicki A. Talanta 1967 14 1371. Sychra V. Lang I. and Sebor G. Prog. Anal. At. Spectrosc. 1981 4 341. Komarek J. and Sommer L. Talanta 1982 29 159. Volynsky A. B. Spivakov Ya. and Zolotov Yu. A. Talanta 1984 31 449. Tserovsky E. and Arpadjan S. J. Anal. At. Spectrom. 1991 6 487. Polo-Diez L. Hernandez-Mendez J. and Petraz-Penalva F. Analyst 1980 105 37. de La Guardia M. and Sanchez M. J. At. Spectrosc. 1982,3 36. Hernandez-Mendez J. Polo-Diez L. and Bernal-Melchor P. Anal Chim. Acta 1979 108 39. Polo-Diez L. Hernandez-Mtndez J. and Rodrigues Gon- zales J. A. Analyst 1981 106 737. Vidal Gandia M. T. and de La Guardia Cirugeda M. Analusis 1985 13 233. Vidal M. T. and de La Guardia M. 7alanta 1987 34 892. Bunton C. A. and de Buzzaccarini F. J. Phys. Chem. 198 1 85 3159. Barden R. E. and Holt S. L. in Solution Chemistry of Surfactants ed. Mittal K. L. Plenum Press New York 1979 vol. 2 p. 707. Paper 310 1082A Received February 23 1993 Accepted March 15 1993 I Karwowska R. Bulska E. Barakat K. A. and Hulanicki A. Chem. Anal. (Warsaw) 1980 25 1043.

ISSN:0267-9477    

DOI:10.1039/JA9930800749

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL. 8 755 Determination of Tin in Indium Phosphide by Electrothermal Atomic Absorption Spectrometry and Inductively Coupled Plasma Mass Spectrometry* Marco Taddia and Michele Bosi University of Bologna G. Ciamician Department of Chemistry Via Selmi 2 1-40126 Bologna Italy Vanes Poluzzi University of Bologna School of Analytical Chemistry Via Selmi 2 1-40126 Bologna Italy Two independent methods for the determination of tin in tin-doped indium phosphide were developed and compared. The electrothermal atomic absorption spectrometry (ETAAS) method utilized both platform atomization and a chemical modifier composed of orthophosphoric acid and magnesium nitrate. The detection limit (6sJ is 5.0 pg g-I for a 250 mg sample. The inductively coupled plasma mass spectrometry (ICP-MS) method monitored the I2%n isotope and a typical detection limit (6sb) of 0.5 pg g-I for a 20 mg sample was obtained.The observed range is 0.2-0.75 pg g-I depending on the instrumental stability and response. The matrix effect was investigated in both methods and the results indicated that matrix-matched calibration standards should be used. Indium phosphide wafers obtained from a single crystal grown from a tin-doped melt were analysed by using both methods to determine the tin content and the change in concentration along the crystal. Results having means that did not differ significantly (P=0.05) were achieved. Both the ETAAS and ICP-MS techniques were useful in determining tin in the tin-doped InP however ICP-MS without internal standardization was less precise.When real samples containing 35-1 49 pg g-l of Sn were analysed the relative standard deviation was in the range 12.9-4.3% for ICP-MS and 3.4-2.0% for ETAAS. Keywords Inductively coupled plasma mass spectrometry; electrothermal atomic absorption spectrometry; tin; indium phosphide 111-V Semiconductor compounds e.g. gallium arseriide (GaAs) and indium phosphide (InP) are very important materials in the areas of microelectronics and optoelectron- ics. Indium phosphide has become the material of choice in the fabrication of sources (light-emitting diodes laser diodes) and detectors (photodiodes avalanche photodi- odes) operating at wavelengths between 1.3 and 1.6 pm. This is at present the ideal range for optical communi- cations based on 50,-GeO fibres.Electrical devices fabricated on InP substrates include high-speed microwave devices and optoelectronic circuits. Indium phosphide is also attractive for high conversion efficiency solar cells having a greater resistance to radiation degradation than silicon and gallium arsenide cells.' The n-type InP substrate for laser diode fabrication is routinely prepared by incorpo- rating tin in the lattice. Tin-doped InP single crystals are grown by the liquid-encapsulated Czochralski (LEC) tech- nique by adding elemental tin to the undoped polycrystal- line InP in the starting charge. The dopant concentration along the ingot is a function of the distribution coefficient of tin in InP. The electrical characterization of the wafers is usually performed by the Van der Pauw technique based on the Hall-effect measurement which allows for the concen- tration of the carriers without regard to their identity.In general the chemical characterization is performed by mass spectrometric techniques requiring a preliminary calibra- tion by using implanted or uniformly-doped samples as standards. The aim of this research was to develop an accurate and precise method for the determination of tin in tin-doped InP. Electrothermal atomic absorption spectrometry (ETAAS) was chosen for its various advantages and because of previous experience in the analysis of 111-V semiconduc- tor material^.^^^ Although the determination of tin by ETAAS is complicated owing to volatilization losses and *Presented at SAC '92 the 10th International Conference on Analytical Chemistry Reading UK September 20-26 1992.vapour-phase interference^,^-' the literature reports its use for environmental,8-Lo inorganic" and biological12 samples. The determination of tin in InAs has also been investigated. l 3 The second objective was to test the analytical capabili- ties of inductively coupled plasma mass spectrometry (ICP- MS) for the determination of tin in InP and to compare the analytical performances of ICP-MS and ETAAS. The determination of tin by ICP-MS in a variety of environmen- tal and standard reference samples has been r e p ~ r t e d ~ and the separation between indium and tin isotopes in a mass spectrum of an InP sample has been presented.14 Experimental Apparatus ETAAS All the atomic absorption measurements were made on a Perkin-Elmer 372 spectrometer equipped with an HGA- 500 graphite furnace atomizer and a deuterium-arc back- ground corrector.The operating conditions are given in Table 1. A Perkin-Elmer hollow cathode lamp was em- ployed as the spectral light source. In order to maintain an acceptable signal-to-noise ratio it was found necessary to increase the operating current gradually from 30 to 33 mA. Preliminary experiments were carried out at a wavelength of 224.6 nm and the background was evaluated using a wavelength of 226.5 nm from a cadmium hollow cathode lamp. Although the 224.6 nm wavelength showed the highest tin sensitivity a wavelength of 235.5 nm gave a better signal-to-noise ratio and in this case the background was evaluated by using a hydrogen hollow cathode contin- uum source.Pyrolytic graphite coated graphite tubes and pyrolytic graphite L'vov platforms were used. New tubes and platforms were conditioned according to the manufac- turer's recommendations. Although the aim was to develop a method for heavily doped samples the necessary precau- tions were taken to minimize the blank by keeping the756 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL. 8 Table 1 HGA-SO0 graphite furnace conditions Wavelengthhm Lamp current/mA Spectral bandpasdnm Drying step Temperature/"C Ramp time/s Hold time/s Temperature/"C Ramp time/s Hold time/s Atomization step Temperat ure/"C Ramp time/s Hold time/s Cleaning step Temperat ure/"C Ramp time/s Hold time/s Temperature/"C Ramp time/s Hold time/s Ashing step Cooling step Argon flow rate (step 3)/ml min-' Mode (peak height)/s Deuterium-arc background correction Sample Volumelpl 235.5 33.0 0.7 250 5 30 1300 1 30 2600 0 6 2700 1 3 20 1 20 50 Balanced 10 6.0 Table 2 Instrumental operating conditions for ICP-MS Forward power/kW Outer gas flow rate/l min-' Intermediate gas flow ratell min-' Nebulizer gas flow rate/l min-I Sample uptake rate/ml min-I Sampler orifice nickellmm Skimmer orifice nickel/mm Nebulizer Acquisition mode Mass range/u No.of scan sweeps Dwell time/ps No. of channels 1.3 0.7 0.8 1 .o 1 .o 0.75 13 V-Groove type Peak jumping 115.90-125.23 1600 80 512 contamination sources under control. To prevent memory effects after running each series of samples poly(tetraflu0- roethylene) (PTFE) test-tubes were boiled first with 0.9 mol 1 - I nitric acid and 0.8 mol 1-l hydrochloric acid and then with de-ionized distilled water.ICP-MS The ICP-MS measurements were carried out using a VG Elemental PlasmaQuad PQ2 + spectrometer. A Gilson autosampler was used. The ICP-MS operating conditions are given in Table 2. The instrumental stability mass resolution and calibration were checked daily by means of standards. Similarly the plasma torch position ion lenses tuning and gas flow rates were optimized daily. An accurate cleaning of both skimmer and sampler cones as well as the spray chamber and plasma torch were crucial in order to maintain a stable instrumental response. Reagents A stock standard solution ( 1000 mg 1 - I of Sn) was prepared by diluting the Prolabo standard tin tetrachloride solution with water and 4 mol 1-l hydrochloric acid to a final hydrochloric acid concentration of about 3.7 mol 1 - I .Working standards were prepared by dilution of the stock solution with de-ionized distilled water. The orthophos- phoric acid was of Suprapur quality other acids were electronic-reagent grade and the remaining reagents were of analytical-reagent grade. The undoped InP was received from the Istituto Materiali Speciali per Elettronica e Magnetism0 (MASPEC Institute of the National Research Council) Italy. Procedure ETAS The InP slice was preliminarily washed with acetone to remove traces of grease and briefly dipped in concentrated hydrochloric acid until the chemical attack began. At that point water was added to stop the reaction the slice was washed again with acetone and dried in an oven at 100 "C.After cooling the sample was pulverized and an appropri- ate weighed amount (50- 150 mg) was dissolved in 2 ml of 37% m/m hydrochloric acid in a PTFE test tube. After leaving for 15 min at room temperature the solution was transferred into a 25 ml calibrated flask and the volume adjusted with de-ionized distilled water. A 10 pl aliquot of the sample was injected into the furnace followed by 10 p1 of a chemical modifier solution containing 0.5 g 1-I of Mg as the nitrate and 0.125 mol 1-' orthophosphoric acid. A thermal cycle was carried out and the peak height absor- bance reading was taken. At least five replicate measure- ments were obtained for each sample.The unknown concentration was determined from a matrix-matched calibration graph. This was obtained by standard additions to a blank containing the same amount of the undoped InP. ICP-MS For the ICP-MS the preparation procedure used for ETAAS was slightly modified to keep the dissolved solid content in the sample solution within the tolerable limits ( ~ 0 . 2 % m/v). A 20 mg InP sample as dissolved in 2 ml of 37% m/m hydrochloric acid. The solution was transferred into a 50 ml calibrated flask and the volume was adjusted with de- ionized distilled water. A series of matrix-matched stan- dards in the range 1-200 pg 1-I of Sn were prepared by adding known amounts of tin to a solution of the undoped InP. First the standards then the unknowns were nebu- lized into the plasma and the lzoSn isotope was measured.Between the standards and samples hydrochloric acid (4% v/v) was aspirated into the nebulizer for a few minutes to restore the original blank signal of the acid. The unknown concentrations were determined from the calibration graph. Results and Discussion ETAAS The starting point for the systematic investigations was to compare wall versus platform atomization in terms of sensitivity and detection limit. At the 224.6 nm wavelength with standards prepared in 2% v/v nitric acid best results were achieved by using the platform. The characteristic mass of tin [peak height absorbance (A,)=0.0044] was 48 pg with the platform in the pyrolytic graphite coated graphite tube 393 pg with the coated tube alone and 124 pg with the uncoated tube alone.The detection limits ( 6 ~ ~ ) were 0.35 5.9 and 0.74 pg respectively. The poor detection limit resulting from the atomization in the pyrolytic graphite coated graphite tube without the platform is a consequence of the lack of signal repeatability. For example when atomizing 16 ng of tin the mean absorbance was 0.166 with an RSD of 19% (n=6). This should be worth investigating as it has been demonstratedIO that tungsten carbide coated tubes provide unsatisfactory results. In fact tungsten coating reducedJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL. 8 757 peak height sensitivity for tin by 30% relative to that obtained with the pyrolitic graphite coated graphite tube and platform. When tubes and platforms were both coated with tungsten the peak area sensitivity was similar to that obtained with pyrolytic graphite coated graphite tubes and platforms although the tungsten coating produced a highly distorted tin signal.Therefore tungsten-coated atomizers are not advantageous with respect to the pyrolytic graphite coated graphite tube and pyrolytic platform. The sample solution obtained by dissolving the InP in hydrochloric acid contains indium trichloride and phos- phorus is lost as phosphine. Such a matrix following vaporization in the atomizer can produce background signals ascribable to scattering and molecular absorption of stable gaseous indium monochloride16 and probably of the oxide (In,O). The presence of In20 in pyrolytic graphite coated graphite tubes at 1000- 1500 "C has recently been confirmed by mass spectrometry.17 Background studies showed that the non-specific absorb- ance signal produced by 52 pg of the undoped InP (HC1) at 226.5 nm (Cd 11) was lost only by ashing at 1400 "C which is too high a temperature for tin. Even though the background absorbance at 235.5 nm as measured with the hydrogen source was approximately half of that at 226.5 nm a similar temperature was necessary.This required the use of an appropriate chemical modifier in order to increase the maximum permissible ashing temperature. The literature'* reports several chemical modifiers for the determination of tin. A comparative study of the influence of various inorganic compounds on the ETAAS signal for tin has also been performed by Dittrich et a1.,l3 in a paper devoted to the analysis of InAs.They concluded that nickel nitrate was the best modifier since the tin signal was enhanced by a factor of four. The present evaluation was restricted to Nil1 Co" Pdll and to the mixed modifier magnesium nitrate plus ortho- phosphoric acid. All of them allowed an increase in the maximum ashing temperature in the absence of indium Ni and Co to 1200 "C Pd to around 900 "C and magnesium nitrate plus orthophosphoric acid to 1300 "C. However in the presence of the indium matrix nickel strongly depressed the tin signal thus negating the above benefit. Unlike nickel the mixed modifier magnesium nitrate plus orthophosphoric acid did not alter the results with or without the presence of indium. It should be noted that orthophosphoric acid was used instead of the more common ammonium mono- or dihydrogenphosphate as it is available to a high degree of p ~ r i t y .~ Experiments showed that the modifier itself pro- duced a background signal at the analytical wavelength of 235.5 nm. However ashing at 1300 "C reduced the total background absorbance to a level that could be handled by the deuterium-arc correction system. (The y-intercept of the tin calibration graph did not differ significantly from zero over the whole matrix concentration range.) The influence of the matrix mass on the tin peak absorbance was investigated by comparing the standard additions calibration graphs obtained using sample solutions prepared from different masses of InP. The results indicated that the variation in slope caused by different matrix masses (17-47 pg of In injected) was lower than the residual standard deviation.However as the slope varied from day to day calibration was repeated for each sample run by performing standard additions on a matrix-matched blank. In the presence of 47 pg of In the linear regression parameters (95% confidence) for a typical five points graph in the range 0-7.5 ng of Sn (upper linearity limit) were a=0.16x 1 . 5 5 ~ lo- b=6.20x 10-2+0.42~ and correlation coefficient (r)=0.9993. The average characteristic mass of tin [peak height absorbance (A,) = 0.00441 in the presence of different matrix masses (17-47 pg of In) was 72k3 pg (95% confidence n = 7). The detection limit (64,) is 5.0 pg g-' for a 250 mg sample. This was calculated as the abscissa value from the calibration function y=a + 6syl where a is the y-intercept and sylx is the standard deviation of the y residuals. This is the concentration which can be detected with 99.87% certainty.l9 The use of the y-intercept in place of the mean blank signal and sylx in place of the blank standard deviation is justified by the fundamental assumptions of the unweighted least squares method.20 Recovery tests were performed by adding known aliquots of standard tin solution to the undoped InP before dissolving the samples.The tin content of the undoped InP was below the detection limit of the method and negligible compared with the content in the doped material. The mean A given by the independent blanks prepared from the undoped InP was 0.002 k 0.002 (n= 8).Results of recovery experiments are shown in Table 3. The tin recovery varied between 98.2 and 99.1% in the range 50-100 pg g-' of Sn. ICP-MS Indium and tin are elements of neighbouring masses two of their isotopes lIsIn and lI6Sn representing a typical resolution problem. l 4 This resolution is complicated in the presence of a large excess of indium and trace amounts of tin. Fortunately the most abundant tin isotope lZ0Sn does not suffer from either the above interference or interference from certain cadmium tellurium and xenon isotopes. For these reasons the lzoSn was monitored in the present work. When compared with calibration graphs of samples in 4% v/v hydrochloric acid calibration graphs in the presence of InP decomposition products revealed a moderate matrix effect.The parameter chosen to measure the matrix effect was the variation in tin sensitivity (calibration function slope) in the presence of the matrix and in 4% v/v hydrochloric acid. The slope ratio (InP matrix/4% v/v HCl) varied from day to day in the range 0.82-0.95 depending on the instrumental conditions. For this reason matrix- matched standards were used. Another problem was sensi- tivity drift which occurred after prolonged nebulization of InP solutions into the plasma. A decrease of about 10- 1 5% was observed after 1 h. This could be due to the presence of a yellowish deposit probably In203 on the sample and Table 3 Results of recovery experiments for tin by using the ETAAS procedure Added*/ Found?/ Recovery g-' Pg g-' (O/O) 50.0 49.5 st 0.2 99.1 ( n = 5 ) 100.0 98.2 k 0.6 98.2 ( n = 6 ) *Added to 100.0 mg of the undoped InP.?Mean value & standard deviation of n determinations. Table 4 Results of recovery experiments for tin by using the ICP- MS procedure Added*/ 15.0 25.0 50.0 75.0 100.0 125.0 150.0 200.0 Pg g-' Found+/ Pg g-' 15.9 2 0.2 22.7 2 0.4 48.9 2 1.7 79.524.1 96.0 1- 3.9 126.4 f 0.9 153.8 1- 1.8 21 3.6 st4.1 Recovery 106.0 90.8 97.8 106.0 96.0 101.1 102.5 106.8 *Added to 20.0 mg of the undoped InP. TMean Value +- standard deviation of four determinations.758 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL. 8 ~ ~~ ~ Table 5 Results of comparative determinations of tin in tin-doped InP samples by ETAAS ICP-MS and Van der Pauw electrical measurements ETAAS ICP-MS Van der Pauw Wafer carrier concentration§/ No.y/L* pgg-'? atom ~ m - ~ Pug g-'4 atom cm-3 ~ m - ~ 5 0.13 35.52 1.2 ( 0 . 8 6 f 0 . 0 3 ) ~ 10I8 38.7k5.0 (0.94k0.12)~ 1OI8 1 . 1 5 ~ 10l8 17 0.26 52.0 k 1.8 ( I .26 f 0.04) x loi8 53.8 k 3.5 ( I .3 I f 0.09) x 10I8 1.35 x 10l8 27 0.43 74.6+ 1.9 (1.81 k 0 . 0 5 ) ~ 10I8 76.7k3.3 ( 1 . 8 6 f 0 . 0 8 ) ~ 10l8 1 . 8 0 ~ loL8 0.66 147-t3 (3.57-tO.07)~ 10I8 149f 1 1 (3.62f0.27)~ 10I8 3 . 0 ~ 1018 37 *Zone length solidified (y= seed end distance L= total crystal length). ?Mean value f standard deviation of six determinations. $Mean value f standard deviation of three determinations. 4Approximately 10% uncertainty. skimmer cones. Operating parameters that shortened the time of analysis were adopted to prevent such a drawback. The detection limit (6sb) of the ICP-MS procedure was typically 0.5 pg g-I for a 20 mg InP sample.It was calculated from the analytical calibration function as the abscissa value giving y=yb+ 6sb where yb is the mean value of 20 blank signal measurements with standard deviation sb. A typical blank signal was around 200 counts s-l with standard deviation of about 20 counts s-I. Detection limits for the determination of tin in biological material and sediments previously r e p ~ r t e d ~ were in the range 0.025-0.5 pg g-l of Sn. In the present work the observed range extended from 0.2 to 0.75 pg g-l the lowest limit being achieved only after thorough cleaning of the whole apparatus and the highest one after a few days of operation. The detection limit of 0.5 pg g-l quoted above is representative of typical instrumental stability conditions achievable in routine work.Recovery tests were performed in the range 15-200 pg g-l of Sn by using a procedure similar to that employed in the ETAAS method. As shown in Table 4 recoveries varied from 90.8-1 06.8%. Analysis of Real Samples and Comparison of Methods Both methods were applied to the determination of tin in tin-doped InP samples. These were wafers cut at four different heights of the ingot in order to compare the dopant concentrations corresponding to different solidified regions. The LEC-grown doped InP was obtained from the MASPEC Institute. Results are reported in Table 5 and compared with the carrier concentrations given by the Hall-effect measure- ments. The ETAAS and the ICP-MS gave results having means which do not differ significantly (P=0.05).The relative standard deviation (RSD) for the ETAAS results varied from 2.0 to 3.4% whereas for ICP-MS it varied from 4.3 to 12.9%. The results confirm how during the growth of the InP crystal from a tin-doped melt tin is rejected by the solid and is accumulated by the liquid (K<1 where K= distribution coefficient). Although both ETAAS and ICP-MS are valuable tech- niques for the determination of tin in heavily doped samples the analytical performances were slightly different. A comparison of both techniques is summarized in Table 6. In terms of the detection limit ICP-MS is better than ~ ~~ ~ Table 6 Detection limit (DL) and precision obtained in the determination of tin in InP using ETAAS and ICP-MS Method DL/pg g-l RSD* (O/O) ETAAS 5.0 3.4-2.0 ICP-MS 0.5 12.9-4.3 *Obtained in the concentration range 35- 149 pg g-' of Sn in InP.ETAAS although the precision of ICP-MS without internal standardization is lower. By adding an internal reference element of a concentration equal to both samples and standards and then measuring the ratio of the tin and the internal standard signals the precision of the ICP-MS results should be improved. This is expected to be particu- larly evident when samples with a high content of dissolved solids are analysed. In fact internal standardization allows correction for long-term signal instability arising from nebulizer clogging and other sources mentioned above. The authors thank Dr. R. Fornari (Consiglio Nazionale delle Ricerche MASPEC Parma Italy) for providing InP samples and the relevant electrical data.1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 19 20 References Yamamoto A. Yamaguchi M. and Uemura C. Appl. Phys. Lett. 1984 44 6 I 1. Gauneau M. Rupert A. Minier M. Regreny O. and Coquille R. Anal. Chim. Acta 1982 135 193. Taddia M. and Filippini O. Fresenius'Z. Anal. Chem. 1988 330 506. Taddia M. and Clauser G. Fresenius' 2. Anal. Chem. 1989 334 148. Wendl W. and Muller-Vogt G. Spectrochim. Acta Part B 1984 39 237. Volynsky A. B. Sedykh E. M. Spivakov B. Ya. and Havezov I. Anal. Chim. Acta 1985 174 173. Frech W. Lundberg E. and Cedergren A. Prog. Anal. At. Spectrosc. 1985 8 257. Apte S. C. and Gardner M. J. Talanta 1988 35 539. Brzezinska-Paudyn A. and Van Loon J. C. Fresenius' Z. Anal. Chern. 1988 331 707. Subramanian K. S. Talanta 1989 36 1075. Goyal N. Purohit P. J. Dhobale A. R. Patel B. M. Page A. G. and Sastry M. D. J. Anal. At. Spectrom. 1987 2 459. Iwamoto E. Shimazu H. Yokota K. and Kumamaru T. J. Anal. At. Spectrom. 1992 7 421. Dittrich K. Mothes W. Yudelevich I. G. and Papina T. S. Talanta 1985 32 195. Shushan B. Quan E. S. K. Boorn A. Douglas D. J. and Rosenblatt G. in Microelectronics Processing Inorganic Ma- terials Characterization ed. Casper L. A. American Chemical Society Washington DC 1986 ch. 17. Vickrey T. M Harrison G. V. Ramelow G. J. and Carver J. C. Anal. Lett. 1980 13 781. Dittrich K. Talanta 1977 24 725. McAllister T. J. Anal. At. Spectrom. 1990 5 171. Tsalev D. L. Slaveykova V. I. and Mandjukov P. B. Spectrochim. Acta Rev. 1990 13 225. Massart D. L. Vandeginste B. G. M. Deming S. N. Michotte Y. and Kaufman L. Chemometrics A Textbook Elsevier Amsterdam 1988 pp. I 12-1 13. Miller J. C. and Miller J. N. Statistics for Analytical Chemistry Ellis Horwood Chichester 1988 p. 1 17. Paper 3/00151B Received January I I 1993 Accepted April 26 I993

ISSN:0267-9477    

DOI:10.1039/JA9930800755

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL. 8 759 Elimination of the Interfering Effect of Transition Metals in the Determination of Tin by Hydride Generation Atomic Absorption Spectrometry Amin M. Abdallah Mohamed M. El-Defrawy* Nagwa Nawar and Manal M. El-Shamy Department of Chemistry Faculty of Science University of Mansoura 3551 6 P. 0. Box 30 Mansoura Egypt The determination of tin in the presence of a variety of metal ions has been studied by hydride generation atomic absorption spectrometry. The results show serious depressing interference from transition metals. Such interference can be completely eliminated using potassium cyanide. A mechanism is put forward to interpret the effect of cyanide via the formation of a stable cyano-metal complex with the interfering metal before the harmful depressing effect of the latter takes place in the quartz cell.Keywords Tin determination; hydride generation atomic absorption spectrometry; interference; cyanide Since about 1970 the hydride generation technique has found wide application in the determination of hydride forming elements.Iv2 Several transition metal ions can cause interference unless appropriate precautions are taken.3-5 In the hydride generation technique the interference reported is found to be independent of the ana1yte:interferent ratio but dependent upon the concentration of the interferent in the solution.6 Few procedures with limited applications such as solvent extraction' are met with in the literature. Agterdenbos and Bax8 reported that the interference by transition metal ions is caused by their enhancing the catalytic decomposition of NaBH,.The literature survey carried out for this research revealed the possibility that the harmful catalytic action of the transition elements resulting in the decomposition of tetrahydroborate could be removed by use of potassium cyanide. Experimental Instrumentation A Perkin-Elmer 2380 atomic absorption spectrometer was used with a Unicam tin hollow cathode lamp operated at 286.3 nm with a 0.7 nm slit-width. Absorbances (peak height) were recorded with a PM 8251 single-pen recorder (Unicam) at a chart speed of 10 cm min-l. A Perkin-Elmer MHS- 10 hydride generation system equipped with a 10 cm single slot burner was used. An acetylene-air flame was employed with flow rates of 20 1 min-l ( 10 psig) and 45 1 min-I(60- 100 psig) respectively.Nitrogen (1 1 min-l) was used to purge the generation system. The catalytic study was carried out using the Schlenk-line t e c h n i q ~ e . ~ A 20 pg 1-* solution of SnIV was prepared from a standard SnIV solution from BDH ( 1000 pg 1-I) by appropri- ate dilution. A 2% m/v sodium tetrahydroborate (Fluka) in 1 O/o m/v sodium hydroxide solution was prepared and filtered immediately before use. All chemicals used were of the highest purity available. Recommended Procedure An appropriate volume of the standard or sample containing up to 50 pg 1-I of SnIV with or without interferent is placed with a micropipette (< 1 pl) in the reaction vessel containing 10 ml of 1% hydrochloric acid. The reaction vessel is attached to the hydride generation unit.After 30 s the * To whom correspondence should be addressed. solution in the reductant vessel containing the tetrahydro- borate with or without cyanide (CN) ethylenediamine- tetraacetic acid (EDTA) Alizarine Red Sulfonate (ARS) Eriochrome Cyanine RC (ECRC). 8-hydroxyquinoline (8- HQ) catechol (CAT) or pyrogallol (PYRO) is then trans- ferred into the reaction vessel. The run is carried out as usual. Results and Discussion Preliminary investigation on the effect of different metal ions including transition metal ions on the absorbance of Sn is shown in Fig. 1. Trials to eliminate interference of transition metal ions using different chelating agents such as EDTA ARS ECRC 8-HQ CAT PYRO or CN added directly to the solution in the reaction flask by taking the absorbance signals and relating them to the signal of the standard Sn solution without addition of the interferents were made.The experiments carried out under such conditions were not successful. Changing the sequence of addition of the reagents results in the addition of the chelating agent to the reductant reservoir in the presence of the alkaline tetrahydroborate solution. Under this condi- tion the controlled automatic transfer of the mixture from the reductant reservoir to the reaction flask makes the reaction of the chelating agent take place more conveni- ently. The experimental results tabulated in Table 1 indicate that the ineffectiveness of most of the ligands investigated could possibly be due to the weak complex formation between the interferents and the ligand under the prevailing conditions.In view of the results obtained it is convenient to add cyanide to solutions of Sn plus interfer- ents. Fig. 1 shows clearly that when potassium cyanide is added in increasing concentration to solutions containing Sn in the presence of different interferents the depressive effect of the latter gradually decreases and disappears as the concentration of the cyanide reaches 2 x mol I-'. The absorbance of Sn returns to the same value as the standard after which the absorbance becomes almost level irrespec- tive of addition of cyanide. A cyanide concentration of 2.0 x l 0-3 mol 1-1 is found to be capable of removing the interfering effect of different transition metal ions.Dupli- cate determinations were performed throughout the experi- ments. Results are related to the signal obtained from an SnIV solution plus the recommended cyanide concentration which has no significant effect on the absorbance signal of the analyte. A detection limit of 0.4 pg 1-l was recorded with a precision of 1.6% for ten determinations. The hydride technique is based on the conversion of the analyte to its volatile hydride transport of the hydride to a760 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL. 8 ~- ~ Table 1 Effect of some chelating agents on the absorbance signal of 20 pg I-' of SnIV in the presence of interferents. The concentration of the ligands was 2 x mol I-' in each case while the concentration of interfering metal ion was 20 pg ml-' Sn recovery (Yo) Interferent None Ni c u Au Cu + Co + Ni Mn+Au+Pd Pd + Fe + Cr Rd+ Ru + Fe EDTA ARS ECRC 100 37 52 10 32 12 10 1 1 32 19 33 5 15 21 64 23 65 19 35 5 22 23 67 32 8-HQ 100 21 63 1 1 PPt 25 CAT PYRO CN 100 100 100 24 21 100 50 52 98 10 - 100 64 68 98 70 72 101 15 66 102 99 - - KCN/103 mol I-' 1 2 3 4 5 6 100 80 20 C B LL I 1 I - ,. I 0 20 40 60 80 I nte rferenvpg m I-' Fig.1 Variation in absorbance signal of 20 pg I-' of SnIV in presence of interferent A Ni B Cu C Au D Ni with interferent plus KCN E Cu with interferent plus KCN F Au with interferent plus KCN solution and G Mg or Ca with or without addition of KCN. The absorbance signal of SnIV only or in the presence of CN- is taken as 100% for comparison heated quartz cell by a carrier gas and the dissociation of the volatile hydride to its constituent atoms.Many pro- cesses occur before reaching this final step. The main equilibria that exist in these processes are BH4- + 3H20 + 4H+ +H3B03 + 1 1 H (1) 1 1H+3BH4-+4H2Sn03-4SnH4+3H20+3H3B03 (2) SnH4+Sn+ 2H2 (3) Many researchersLOJ have studied the atomization of gaseous hydrides during the determination of hydride forming elements. They have pointed out that the atomiza- tion is not caused by the thermal decomposition but is due to collisions during the sweeping process to the heated quartz cell. However the presence of metal ions might catalyse the recombination of the constituents. Therefore a depressing effect on the analyte signal might be observed in the presence of the interfering ions.Most probably the depressing effect is caused by increasing the disorder of the system which prevents the completion of reaction (2) thus decreasing the production of the analyte. Moreover an investigation into the evolution rate of hydrogen as a criterion for the decomposition rate of tetrahydroborate in the absence and presence of some interfering elements vit Cu Ni Co Mg Ca and Au was carried out. To follow the hydrogen evolved as an indica- tion of the decomposition of the tetrahydroborate [reaction (I)] a soap bubble flowmeter connected to a three-necked flask with Schlenk tubes and vacuum lines was used.'* A constant flow rate of nitrogen as the purge gas of the 0 20 40 60 80 100 120 140 I nte rfe renvpg m I-' Fig. 2 Effect of different metal ions A Ni; B Cu; C Au; and D Ca or Mg on the rate of evolution of hydrogen as an indicator of the decomposition rate of tetrahydroborate; the evolution rate in a solution of 1% m/v HCI and 20 ,ug I-' SnIV is taken as 100% and all other results are related to this value generating system was maintained the number of hydrogen bubbles evolved counted and the evolution rate of hydrogen calculated. The conditions under which these experiments were carried out were principally similar to those taking place in the analysis cell used for hydride generation. Fig.2 shows the effect of some metals on the rate of evolution of hydrogen. It is clear that the general trend of transition metals is to increase the rate of evolution of hydrogen i.e. to increase the decomposition rate of tetrahydroborate. In the same context when cyanide plus tetrahydroborate are added to a hydrochloric acid solution containing Sn and interferent (Cu Ni Co Mg Ca or Au) no significant change in the evolution rate of hydrogen was recorded. The implication of the foregoing results and observations is that it is possible to inhibit the action of transition metal ions of increasing the decomposition rate of tetrahydro- borate by complexing the former with cyanide in solution before the component is swept away into the quartz cell.Formation of a stable cyano-complex with the interferent and not with the analyte Sn13 makes it possible to determine the latter safely in the presence of a variety of transition metal ions. References 1 Hatch W. R. and Ott W. L. Anal. Chem. 1968 40 2085. 2 Bailey B. W. and Lo F. C. Anal. Chem. 1971 43 1521. 3 Camail M. Loiseau B. Margaillan A. and Vernet J. Analusis 1983 11 358. 4 Brindle I. D. and Le X-c. Analyst 1988 113 1377. 5 Castillo J. R. Mir J. M. and Perez I. Microchem. J. 1989 39 119.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL. 8 76 1 6 Bertha A. and Ikrenyi K. Anal. Chim. Acta. 1982 139 329. 7 Narasaki H. and Ikeda M. Fresenius’ J. Anal. Chem. 1990 336 5 . 8 Agterdenbos J. and Bax D. Fresenius’ 2. Anal. Chem. 1986 323 783. 9 Shriver D. F. The Manipulation of Air Sensitive Compounds McGraw-Hill New York 1969. 10 Welz B. and Melcher M. Analyst 1984 109 569. 1 1 Welz B. and Melcher M. Analyst 1984 109 577. 12 Miles M. L. Harris T. M. and Hauser C. R. J. Org. Chem. 1965,30 1007. 13 Sharpe A. G. The Chemistry of Cyano Complexes of Transi- tion Metals Academic Press London 1976. Paper 2/06 748J Received December 21 1992 Accepted March 17 I993

ISSN:0267-9477    

DOI:10.1039/JA9930800759

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL. 8 763 INTER-LABORATORY NOTE Simple Stirring Device for a Slurry Sampling Technique in Electrothermal Atomic Absorption Spectrometry Bohumil DoCekal Institute of Analytical Chemistry Czech Academy of Sciences Veveri 97 CS-61142 Bmo Czechoslovakia A simple magnetic stirring device for continuous homogenization of slurries during sampling with conventional autosamplers has been designed as a small turbine with rotating magnetic pieces. The device can be easily incorporated into an autosampler tray and can be driven by compressed air or cooling water. Keywords A u toma tic slurry sampling ; magnetic stirring device; elec tro the rma I atomic absorption spectrometry The introduction of a suspension of a finely powdered sample combines the advantages of both liquid and solid sampling.' A remarkable feature of the slurry-suspension technique is the successful application of conventional atomizers and especially autosamplers in the automated analysis of suspended solids.The slurry technique over- comes some of the problems associated with true direct solid sampling for example tedious labour-intensive sam- ple weighing and manual transfer routines. In order to achieve accurate and repeatable results production of a stable homogeneous suspension should be maintained throughout the pipetting of sample aliquots. As reviewed recently,' re-digestion of sample particles application of stabilizing agents (thixotropic thickening agents surfac- tants and wetting agents) magnetic stirring ultrasonic agitation vortex mixing and gas swirling and bubbling are the procedures applied for this purpose.Magnetic stirring vortex mixing and ultrasonic agitation are the most attrac- tive techniques as they can be automated and they avoid the necessity of using stabilizing agents. The application of stabilizing agents might be problematic owing to increasing viscosity or adherence to the sampling capillary or high reagent blanks and contamination risk especially when analysing high-purity materials2 For the on-line homogenization of slurries a magnetic stirrer and an ultrasonic device have been used in combina- tion with a conventional auto~ampler.~-~ Homogenization has been performed by stirring with a small poly(tetraflu0- roethylene) (PTFE)-coated magnetic bar3-6 or by ultrasoni- cation7.* with a tapered titanium probe immersed in the slurry. A miniature magnetic device driven by electrical motors has been described by Lynch and Littlejohd for mixing food slurries in autosampler cups placed on the modified autosampler tray.During the analysis of powders used in the production of high-performance ceramics i.e. aluminium oxide3v4 and silicium ~ a r b i d e ~ stirring has been carried out by a remote-controlled rotating magnet situated over the covered sample beaker placed in the position usually used for the vessel with a modifier s o l u t i ~ n . ~ - ~ A new device has been designed for the latter technique. The aim of this paper is to describe this new simple magnetic stirring device which can be easily incorporated in conventional AS-40 -60 and -70 Models of the Perkin- Elmer autosampler in order to perform continuous homo- genization during sampling of slurries.Description of the Device The stirring device presented has been designed as a small turbine which is schematically shown in Fig. 1. Apart from magnetic pieces all parts are made of polymethacrylate glass. In the main body [Fig. 1(4] two channels are drilled for the tangential introduction of a driving medium into a rotor chamber. For easy connection of input and output tubings with a driving medium two fittings are positioned on the cover body [Fig. l(a)]. The medium is injected through a nozzle onto blades of a rotor [Fig. l(c)] in which two pieces of small magnets [Fig. l(b)] are fastened in an opposite magnetic direction.During an operation moving magnetic pieces interact with a small magnetic bar in a separate vessel placed on the cover body and containing a sample slurry (see Fig. 2). In the proposed design the device is readily adaptable to the existing Perkin-Elmer instrumentation. It can be easily placed (and also removed) in a few seconds in the autosampler tray in the position usually designated for the beaker with a modifier solution (see Fig. 2). In contrast to other devices this one requires no additional electrical supplies. The turbine can be driven by compressed air which is usually available in a laboratory equipped with atomic absorption spectrometric instrumentation or eco- nomically by water whereby cooling water is connected in ( a ) Fig. 1 Schematic diagram of stirring device (a) cover body with input and output fittings for the connection of tubings with driving medium; (b) two magnetic pieces mounted in an opposite magnetic direction; (c) rotor with holes for fixing of magnetic pieces; and (4 stator main body with tangentially drilled channels764 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL.8 Fig. 2 View of the mounted stirring device in a Model AS-70 Perkin-Elmer autosampler tray series with a furnace atomizer. The turning speed is adjusted by the flow rate of the medium. The design of the device and of the autosampler tray enables the agitation by a small PTFE-coated magnetic bar of the whole volume of slurry prepared approximately 20-25 ml so that sample aliquoting errors can be mini- m i ~ e d .~ A sample mass of 0.2-0.5 g can usually be sl~rried.*~~ The standard additions method can be per- formed by simple direct spiking of the slurry with negligible volumes (1 0-1 00 pl) of aqueous standard solutions. Chemi- cal modification can be accomplished by adding an appro- priate reagent directly into the slurry or by immediately dispensing a modifier solution from any position of the autosampler tray usually used for sample cups prior to expelling the slurry. Conclusion In spite of the reported adherence of magnetic particles of a few sample types on the stirrer bar e.g. iron-rich soils,'O the device appears to be considered adequate for routine application in the analysis of many samples. For example it has been successfully used in the analysis of ultra-high purity molybdenum oxide and silicon carbide.For homo- geneous samples the reproducibility of dispensing slurries was comparable to the typical reproducibility of dispensing of aqueous solutions.*?* A simple device could be imple- mented in any autosampler by changing the design of the tray and control programme. 1 2 3 4 5 6 7 8 9 10 References Bendicho C. and de Loos-Vollebregt M. T. C. J. Anal. At. Spectrom. 1991 6 353. DoEekal B. and Krivan V. J. Anal. At. Spectrom. 1993 8 000. Slovak Z. and DoEekal B. paper presented at the 5th Czechoslovak Conference on Atomic Spectrometry Nitra Czechoslovakia September 1-5 1980. Slovak Z. and DoEekal B. Anal. Chim. Acta 1981 129,263. DoEekal B. and Krivan V. J. Anal. At. Spectrom. 1992 7 521. Lynch S. and Littlejohn D. J. Anal. At. Spectrom. 1989 4 157. Miller-Ihli N. J. J. Anal. At. Spectrom. 1988 3 73. Miller-Ihli N. J. J. Anal. At. Spectrom. 1989 4 295. Epstein M. S. Carnrick G. R. Slavin W. and Miller-Ihli N. J. Anal. Chem. 1989 61 1414. Hinds M. W. and Jackson K. W. At. Spectrosc. 1991 12 109. Paper 3/00912B Received February 15 1993 Accepted April 14 I993

ISSN:0267-9477    

DOI:10.1039/JA9930800763

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL. 8 765 Abdallah Amin M. 759 Akatsuka Kunihiko 279 Alvarado Jose 253 Andreae Meinrat O. 119 Aoki Hiroyuki 41 5 Arpadjan Sonja 85 Axner Ove 375 Azerado M. A. 279 Bailey Elizabeth H. 551 Barciela-Alonso C. 649 Barciela-Garcia J. 649 Barnes Ramon M. 467 Bastos M. Lourdes 655 Baxter Malcolm 69 1 Beceiro-Gonzblez E. 649 Bermejo-Barrera A. 649 Bermejo-Barrera P. 649 Berman Shier S. 279 Berndt Harald 243 Blades Michael W. 26 1 Blais Jean-Simon 659 Bloxham Martin J. 499 Bloom Nicolas S. 591 Boer Peter 61 1 Boer Walther H. 61 1 Boonen Sylvie 71 1 Borer Matthew W. 333 Bosi Michele 755 Botto Robert I. 51 Brenner Isaac B. 475 Brindle Ian D. 287 Brockman Andreas 397 Brown Garrett N. 21 1 Burguera J. L. 229 235 Burguera M. 229 235 Byrne John P.599 Cai Yong 119 Chmara Carmen 745 Campos Reinaldo C. 247 Cafiada Rudner P. 705 Canals Antonio 109 Can0 Pavon J. M. 705 Cao Jieshan 379 Caruso Joseph A. 427 545 Carrion Nereida 493 Castillo Juan R. 643 665 Chakrabarti Chuni L. 599 Chakraborti Dipankar 643 Chang Mou-sen 379 Chen Yalei 379 Cheng Jianguo 623 Chenery Simon 299 Chikuma Masahiko 4 15 Chirinos Jose 493 Ciocan Adeline 273 Cook Jennifer M. 299 Corns Warren T. 71 Cortez Jesus Arroyo 103 Cresser Malcolm S. 269 Crews Helen 691 Cristiano Ana Rita 253 Curtius Adilson J. 247 749 Dams Richard F. J. 433 71 1 Darke Susan A. 145 Dawson J. B. 517 339 749 57 1 737 623 CUMULATIVE AUTHOR INDEX FEBRUARY-AUGUST 1993 de la Calle Guntiiias Maria de Oliveira Elisabeth 367 DoEekal Bohumil 637 763 Doidge Peter S. 403 Eastgate Alan R.305 Ebdon Les 71 691 723 El-Defrawy Mohamed M. Eloi Corinne 217 El-Shamy Manal M. 759 Evans E. Hywel 1 427 Fang Zhaolun 577 Ferreira Margarida A. 655 Fischer Johann L. 487 Fisher Andrew S. 691 Fransen Rent 61 1 Freedman Philip A. 19 Foner Henry 467 Fry Robert C. 305 Galley Paul J. 65 7 15 Gangadharan S. 127 Garcia Alonso Jose Ignacio Garcia de Torres A. 705 Giglio Jeffrey J. 1 Gilchrist Glen F. R. 623 Gill C. G. 261 Gilmutdinov Albert Kh. 387 GinC Maria F. 243 Giovanonne Bruno 673 Gluodenis Thomas J. Jr. Golloch Alfred 397 Gomez M. M. 461 Goodall Phillip 723 Gower Grant H. 305 Granadillo Victor A. 6 15 GrCgoire D. Conrad 599 Hahn Lothar 223 Haigh P. E. 585 Halicz L. 475 Hanna C. P. 585 Hansen Steen Honore 557 Harnly James M. 3 17 Hasegawa Ryosuke 48 1 Herndndez Cordoba Manuel Hernandis Vincente 109 Hiddemann Lars 273 Hieftje Gary M.65 333 Hill Steve J. 71 499 723 Hillamo Risto E. 79 Holclajtner-Antunovic Ivanka D. 349 359 Honda Masatake 453 Huang Chung-Wen 68 I Hughes Dianne M. 623 Inui Syn-ya 595 Israel Yecheskel 467 Jarvis Kym E. 25 Jiang Shiuh-Jen 68 1 Jimenez Maria S. 665 JurasoviC Jasna 4 19 Karadjova Irina 85 Kemp Anthony J. 551 Koch Lothar 673 Kojima Isao 115 Kondo Shinji 115 Beatriz 745 759 673 697 103 339 715 Koomans Hein A. 61 1 Krivan Viliam 637 Krug Francisco J. 243 Krushevska Antoaneta P. Kujirai Osamu 481 Kumar Sunil Jai 127 Laborda Francisco 643 737 Lam Joseph W. H. 279 Lamoureux Marc M. 599 Larsen Erik H. 557 659 Li Anmo 633 Liang Lian 591 Liao Yiping 633 Littlejohn D. 325 Longerich Henry P.371 439 Lopez Garcia Ignacio 103 Luong Van T. 41 Ly Tam 599 Madrid Yolanda 745 Maenhaut Willy 79 Mahalingam T. R. 565 Majidi Vahid 2 17 Marchante Gaybn Juan M. Martines Laura J. 467 Masuda Kimihiko 687 Mathews C. K. 565 Mazzetto G. 89 McAllister Trevor 403 McIntosh S. 585 McLaren James W. 279 McLeod C. W. 461 Milella E. 89 Mir Jose M. 643 737 Miyazaki Akira 449 Mizuno Takayuki 595 Moens Luc J. 71 1 Mohammad Bashir 325 Morgan C. A. 539 Muller-Vogt German 223 Murillo Miguel 493 Nagai Hisao 453 Navarro Janeth A. 6 15 Nawar Nagwa 759 Niemax Kay 273 Nobrega Joaquim A. 243 Nonn Christine 397 Norberg M. 375 Ohlsson K. E. Anders 41 Ohta Kiyohisa 595 Ozaki Elisa Akemi 367 Pakkanen Tuomo A. 79 Platzner I. 19 Poluzzi Vanes 755 Potts Philip J. 293 Prabhu R. Krishna 565 Pretty Jack R.545 Price W. J. 517 Pritzl Gunnar 557 Qi Wenqi 379 Rademeyer Cor J. 487 Radziuk Bernard 409 Ragnarsdottir K. Vala 55 1 Rapsomanikis Spyridon 1 19 Reimer Paul A. 449 Reis Boaventura F. 243 Ren J. M. 59 Rezende Mario do Carmo 467 73 1 247 Richner Peter 45 Robertson J. David 2 17 Romero Romer A. 6 15 Ruiz Ana I. 109 Salin Eric D. 59 SAnchez Uria J. Enrique 731 Sanz-Medel Alfredo 73 1 Sass Vinia A 243 Sella Silvia M. 749 Sen Gupta Joy G. 93 Sentimenti E. 89 Sesi Norman N. 65 Shan Xiao-quan 409 Shimamura Tadashi 453 Silva Ivana A. 749 Sjostrom S. 375 Snook R. D. 517 Smith B. W. 539 Soares Elisa M. 655 Steers Edward B. M. 309 Stockwell Peter B. 7 1 723 Story W. Charles 571 Stroh Andreas 35 Stuhne-Sekalec Lidija 445 Sturgeon Ralph E. 41 Styris David L.21 1 Sugiyama Takehiko 595 Suzuki Tohru 595 Taddia Marco 755 Takahashi Takako 453 687 Takaku Yuichi 687 Tao Guanhong 577 TeliSman Spomenka 4 19 Templeton Douglas M. 445 Thoby-Schultzendorff Dominique 673 Thompson K. Clive 723 Thorne Anne P. 309 TripkoviC Mirjana R. 349 Tserovsky Emil 85 Tyson Julian F. 145 585 Uebbing Jurgen 273 Ure Allan M. 325 Vandecasteele Carlo 433 Vanhaecke Frank 433 Verrept Peter 71 1 Vifiuales Jorge 737 Viswanathan K. S. 565 Vijayalakshmi S. 565 Vollkopf Uwe 35 Voloshin A. V. 387 VyskoEilova Olga 409 Walder Andrew J. 19 Watson John S. 293 Webb Peter. C. 293 Welz Bernhard 409 Wendl Wolfgang 223 Williams John G. 25 Willie Scott N. 41 Winefordner J. D. 539 Worsfold Paul J. 499 691 Wunderli Samuel 45 Xu Sonny X. 445 Yamada Kei 481 Yoffe O. 475 Zakharov Yu. A. 387 Zheng Shaoguang 287 687 359 697

ISSN:0267-9477    

DOI:10.1039/JA9930800765

出版商:RSC    

年代:1993

数据来源: RSC