摘要:

Rapid Indirect Determination of Very Low Levels of Cocaine by Tandem On-line Continuous Separation and Inductively Coupled Plasma Atomic Emission Spectrometric Detection Journal of Analytical Atomic Spectrometry I I ALBERTO MENENDEZ GARCIA ENRIQUE SANCHEZ URIA AND ALFREDO SANZ-MEDEL* Department of Physical and Analytical Chemistry Faculty of Chemistry University of Oviedo Oviedo Spain A very sensitive precise and automated method is proposed for the indirect determination of cocaine. The method is based on selective extraction into chloroform of the protonated alkaloid as an ion pair with Dragendorff reagent; BiH is generated directly from the organic phase and the volatile species is brought to an ICP-AES system for bismuth detection. Both separation steps extraction and gas-liquid separation are accomplished in a continuous manner with on-line final ICP- AES detection.The optimum chemical continuous flow and instrumental variables were determined. The proposed method allows the determination of cocaine over a wide range of concentration (20 ng ml-'-100 pg ml-') and the precision at a concentration of 0.5 pg ml-' was 3%. The method was applied to the determination of cocaine in confiscated samples. Keywords Cocaine; alkaloids; liquid-liquid extraction; continuous pow; hydride generation; inductively coupled plasma atomic emission spectrometry Cocaine a psychotropic drug is methylbenzoylecgonine a derivative of ecgonine which in turn is a carboxylic acid derivative of tropine. Cocaine has been used for medical and therapeutic purposes mainly as a local anaesthetic in ophthal- mology.Its euphorizing action on the nervous central system and the facility for its intake however have led to the spread of its consumption as a drug of abuse. This increasing consump- tion of cocaine necessitates the development of analytical methods in order to achieve adequate control of this drug in consignments of drugs or in biological samples e.g. urine to measure the consumption of cocaine by drug addicts and to control the therapy of drug addiction cure programmes as it is known that metabolites of cocaine such as benzoylecgonine and small amounts of the untransformed drug are excreted in the urine in 24 h; in some cases valuable information can be obtained from the determination of cocaine in hair. Immunoanalysis has been claimed to be the ideal technique for screening the water-soluble metabolite benzoylecgonine in urine samples.For this purpose there are commercial kits for the rapid quantification of cocaine in biological fluids. However positive results obtained by immunoassays must be confirmed by using non-immunological techniques. Gas chromatography is a useful technique for the determi- nation of cocaine (or its metabolites) but prior extraction of the drug'- is usually required for biological samples. HPLC usually in the reverse-phase mode has also been used for cocaine determination with non-polar Spherisorb C or pBondpack c18 column^.^-^ Capillary electroph~resis,~ supercritical fluid chromatographys and electrochemical methodsg have also been proposed for cocaine determination. * To whom correspondence should be addressed.However GC-MS is the technique most frequently used for the determination of cocaine and its metabolite^.'^-'^ Methods based on the formation of ion pairs extractable into organic solvents between the protonated alkaloid and a negative- ly charged metal complex e.g. CO(SCN),~- Ni(SCN),2- or BiI,- have also been applied for cocaine determination. Detection can be carried out by spectr~photometry'~-'~ of the organic phase. Such an extraction scheme has been used for indirect determination using AAS. In one method,17 protonated cocaine (Coc. H)' is continuously extracted into 1,2-dichloromethane with BiI,- providing the Bi in the organic solvent. Further the cocaine ion pair may be precipitated as the metal complex and re-dissolved in a flow injection system for final determination of Bi by AAS."*19 In this work the concept of tandem on-line continuous separation based on two continuous separation steps carried out on-line with ICP-AES detection previously proposed in our laboratory,20 was used for cocaine determination. Dragendorff reagent21 was selected for generating the ion pair (C0c-H)' Bi14- which was extracted in chloroform in a continuous mode; BiH was directly generated on-line from the organic phase and continuously introduced into the torch of the ICP for AES measurement of bismuth.EXPERIMENTAL Reagents All chemicals were of analytical-reagent grade unless stated otherwise. Alkaloids. A stock standard solution of cocaine (Sigma-Aldrich Quimica Madrid Spain) was prepared in ethanol at a concen- tration of 1000 pg ml-' and stored at 0-4 "C in PVC container.Ethanol solutions of the following alkaloids were used for the interference study heroin morphine codeine methadone atro- pine lidocaine quinine atropine and strichnine. Dragendor- reagent. A standard solution of Bi( NO) 5H20 (Merck Darmstadt Germany) was prepared by dissolving 5.0 g of the salt in 10 ml of HNO and diluting with distilled water to 100 ml. A 50 g amount of KI (Merck) was dissolved in 100ml of distilled water and Dragendorff reagent was obtained by mixing 5ml of the Bi"' solution 62.5ml of the iodide solution and 487.5ml of distilled water. The concen- tration of tetraiodobismuthate(n1) ( Bi14-) in this solution was 9.3 x lo- mol 1-' (pH 2.0).Solutions of this reagent of lower concentration were prepared by dilution of suitable volumes in doubly distilled water. Sodium tetrahydroborate(@ (NaBH,) solution 1.5% m/v DMF. The solution was prepared daily using NaBH obtained from Merck. Journal of Analytical Atomic Spectrometry August 1996 Vol. 1 1 (561 -565) 561Apparatus All glass material was cleaned with dilute nitric acid (1 +9) and rinsed with ultra-pure water (Milli-Q; Millipore Molsheim France) before use. A Perkin-Elmer (Norwalk CT USA) ICP/5000 spec- trometer equipped with a Perkin-Elmer Data System 10s laboratory computer was used for ICP measurements using the 223.06 nm bismuth emission line. A Minipuls-2 multi-channel pump from Gilson (Worthington OH USA) and an HP4 peristaltic pump from Scharlau Science (Villiers le Bel France) were used for pumping the reagent feeding the tandem on-line continuous separation device.A continuous phase separator made of PTFE with 3 pm pore diameter membranes (Millipore FSLW 02500 Type FS) was employed. Circular membranes (2.5 cm diameter) were used for continuous liquid-liquid extraction of the ion pair of protonated cocaine and tetraiodobismuthate(II1). The general operating mode of the tandem on-line separation device and its coupling to the ICP torch can be seen in Fig. 1. A description of its operation mode for this application is given below. General Procedure The device used in our tandem on-line system allowing for ion-pair formation and its simultaneous and continuous extrac- tion into CHC13 and the direct generation of BiH from this solvent is depicted in Fig.1. First a 2.9 ml min-' flow of cocaine (or blank) in ethanolic medium at pH 1.0 (adjusted with HN03) merges in a T-piece with a 0.5 ml min-' flow of lo- moll-' Dragendorff reagent. The resulting flow reaches the mixing coil and the formation of the (Coc-H)' BiI,- ion pair takes place. Then this aqueous combined flow (3.40 ml min-') merges with a 0.70 ml min-' flow of CHC& at the solvent segmenter where the continuous extraction of the ion pair starts to take place and continues along the extraction coil. The PTFE 'phase separator' leads the aqueous phase continuously to waste and the chloroform phase containing the analyte is merged with NaBH and glacial acetic acid with direct formation of BiH,.This volatile com- pound generated in the continuous mode when the organic phase mixes with a 0.80 ml min-' flow of 1.5% m/v NaBH in DMF solution and a 1.0ml min-' flow of glacial acetic acid is transferred to the ICP torch by a 0.3 1 min-' continuous stream of argon and the liquid and gaseous phases are separated in the gas-liquid separator (see Fig. 1). The BiH3 reaches the plasma where Bi is monitored at the 223.06nm emission line. The GC determination of cocaine in real samples was car- ried out in an external laboratory following a standardized procedure using a phenylsilicone Ultrados column (HP 10091B-012) of 12 m x 0.33 pm id.22 RESULTS AND DISCUSSION Choice of the Organic Solvent for Extraction and the Emission Line The solvents used in our previous work usually well tolerated by the ICP (without disturbance of the plasma stability) include xylene and IBMK.20323*24 Unfortunately the cocaine ion-pair extraction is accompanied by Bi14H complex extrac- tion and therefore it is unsuitable for indirect determination via Bi detection.For this extraction we verified that only low- polarity organic solvents such as chloroform or dichloro- methane,'7.21.25 were able to secure the selective extraction of the ion pair formed between cocaine and Dragendorff reagent. Chloroform provided an intensity-to-background ratio slightly higher than dichloromethane for Bi and superior plasma stability in the continuous system and was therefore selected for the extraction. In order to select the emission line for Bi measurement the four most sensitive lines recommended in the literature were tried. The highest signal-to-background ratio was observed in our system at the 223.06 nm emission line of bismuth.Optimization of the ICP-AES Parameters The three most critical instrumental parameters affecting the emission signals in ICP-AES (rf power carrier gas flow rate and viewing height) were optimized first in our instrument using a simplex optimization method utilizing the minimum 'background equivalent concentration' as the optimization criterion. The optimum experimental values found for these parameters are summarized in the top part of Table 1. Optimization of Chemical Parameters for the Extraction The concentration of BiI,- reagent and pH are the two main chemical parameters to be optimized in order to establish the best experimental conditions for the continuous extraction of the (Coc-H)' BiI,- ion pair into chloroform.For the optimiz- Fig. 1 and ICP Schematic diagram of the continuous liquid-liquid CocH + Bi14- extraction coupled on-line with a hydride generator gas-liquid separator 562 Journal of Analytical Atomic Spectrometry August 1996 Vol. 1 1ation of both parameters a 270 cm extraction coil and a 100 cm mixing coil (Fig. 1) along with a reference 10 pg ml-' cocaine solution in ethanol were used. The effect of BiI,- concentration on the continuous ion-pair extraction was tested in the concen- tration range 10-5-10-3 mol 1-' fixing the extraction pH at 2.0. As can be seen in Fig. 2 the concentration of BiI,- is critical for the efficient extraction of the ion pair for concen- trations of BiI,- higher than 2 x loF4 moll-' the extraction efficiency decreased slightly.A lo- moll-' Bi1,- solution was finally selected for subsequent experiments. The influence of pH on the extraction was then tested in the range 1-4. The results obtained are plotted in Fig. 3. As can be seen pH had a critical effect on the continuous extraction efficiency. Maximum ion-pair extraction was achieved at pH 1 decreasing dramatically at higher pH. These results agree well with the chemistry expected as an acidic medium will be necessary for the protonation of the NH groups of the cocaine molecule (pK= 8.5); in this medium cocaine forms (Coc*H)+ and an ion pair with BiI,-. For pH values higher than 1.0 the efficiency of extraction decreases because deprotonation takes place.On the other hand the effective concentration of BiI,- is probably lower (bismuth precipitated as a basic salt at pH 3.5 or higher). Consequently a pH value of 1.0 (adjusted with HNO,) was finally selected for extraction. Another chemical parameter to be optimized is the NaBH concentration in DMF (necessary for BiH formation directly in the chloroform phase). For its optimization a 10 pg ml-1 cocaine solution to feed the tandem on-line system was used and glacial acetic acid was selected to give an acidic medium for hydride generation from the organic phase. Concentrations of NaBH in DMF in the range 0.5-4% m/v were tested. The ICP-AES signal observed (Fig. 4) increased with increasing concentration of NaBH up to 1 YO. From 1 YO to 2% a plateau was reached and decreasing values of the net emission intensity of bismuth were observed at higher concentrarions probably Fig.2 Effect efficiency (I,) on-line system 0 2 4 6 8 1 0 [BiI;]/mol I-' x 10" of BiI,- concentration on the continuous extraction of 10pgml-' of cocaine in CHCI by the tandem 800 c v) -8 650 E 600 .- 550 0 1 2 3 4 5 PH Fig.3 efficiency (I,) into chloroform of 10 pg ml-' of cocaine Effect of pH on the tandem on-line continuous extraction 200 f 0 1 2 3 4 [NaBH,] in DMF (%) Fig. 4 Effect of NaBH concentration (in DMF) on the continuous generation of BiH by tandem on-line continuous separation with final ICP-AES Bi determination because larger amounts of excess hydrogen are produced that dilute the Bi and at the same time decrease the residence time of the analyte in the plasma leading to poor stability of the plasma.In consequence a 1.5% m/v NaBH concentration was selected for the general procedure. Optimization of the Continuous Flow Parameters Continuous flow parameters to be optimized for the automatic indirect determination of cocaine in our continuous tandem on-line device were the lengths of both the extraction and mixing coils the aqueous flow rate (qw) and the organic flow The lengths of the mixing and extraction coils (Fig. 1) were optimized using a 10 pg ml-' cocaine solution. First using a 270 cm length of the extraction coil the influence of the mixing coil length on the extraction was tested. This parameter seems not to be critical for the ion-pair extraction because the same analytical signal was obtained with coils between 30 and 100 cm long.The influence of the extraction coil length between 50 and 300 cm on the Bi ICP-AES signal was then examined. The results obtained (Fig. 5) show that the analytical signal increases with increasing length of the coil up to 200cm and for longer coils a plateau is reached. An extraction coil of 200cm is long enough for efficient ion-pair extraction so a safer 250cm long extraction coil was selected for the general procedure. The diameter of the coils was less critical as we have already verified in other similar experiments carried out with our tandem on-line system in previous s t ~ d i e s ; ~ ' ? ~ ~ a 0.7 mm id coil was selected. The influence of the total aqueous flow rate qw (sample + reagent solution flow rates) was optimized using a flow rate of chloroform of 0.70mlmin-' for the continuous extraction of a 10 pg ml-1 solution of cocaine at the previously optimized pH value and concentration of Dragendorff reagent.rate (40). 5 0 100 150 200 250 300 Coil lengthkm Fig. 5 of cocaine Effect of the length of the extraction coil on the determination Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 563The sample flow rate was varied between 0.3 and 3.0 ml min-'. With higher total aqueous flow rates a higher mass of cocaine is extracted and a higher analytical signal was observed. However when the sample flow rate was > 3.0 ml min-' the precision and the efficiency of extraction worsened. A 2.90 ml min- sample flow rate (RSD = 0.7%) was finally selected for subsequent experiments.Using this flow rate of 2.90mlmin-' the effect of the chloroform flow rate qo on the extraction was investigated. Values of qo ranging between 0.1 and 1.0 ml min-' were tested and the results obtained are shown in Fig.6. The Bi signal increased with increase in flow rate up to a maximum at 0.75 ml min-' probably as a consequence of the improvement in the efficiency of extraction due to a larger contact surface between the two phases. For qo values >0.75 ml min-' it seems that a lower preconcentration factor prevails as the analytical signal began to decrease. The precision worsened for both high and low values of qo. In consequence a compro- mise value of qo of 0.70 ml min-' (RSD =0.8%) was selected for the general procedure.This value provides a preconcen- tration factor in the organic phase of 4. The optimum values found for the instrumental flow and chemical parameters involved are all summarized in Table 1. Analytical Performance Characteristics Using the 3s IUPAC criterion the LOD attained using the experimental conditions given in Table 1 with the proposed tandem on-line ICP-AES system determination was found to 1200 ..& 1100 E 1000 -g 800 .- s C .- 900 C 0 5 700 600 0 0.2 0.4 0.6 0.8 1 FIOW rate/mt min-' Fig.6 Effect of organic phase flow-rate (qo) on the tandem on-line continuous extraction of 10 pg ml- ' cocaine solution with lo- moll-' BiI,- solution at pH 1.0 (sample flow rate 2.90 ml min-'). The precision of the intensity of the signals estimated by the standard deviation is given for each measurement in bar [numbers below each bar RSD (YO)] Table 1 Optimum parameters (instrumental flow and chemical) for the indirect determination of cocaine by tandem on-line continuous separation and ICP-AES final determination (A= 223.06 nm) Plasma ICP- Rf forward power Carrier gas flow Viewing height (above coil) Continuous pow- Mixing coil length Extraction coil length Internal diameter of coil Sample flow rate Chloroform flow rate Glacial acetic acid flow rate NaBH solution flow rate Bi14- concentration pH of sample Organic phase NaBH concentration Chemical- 1350 W 0.4 1 min-' 11 mm 50 cm 250 cm 0.7 mm 2.90 ml min- 0.70 ml min-' 1.00 ml min-' 0.80 ml min-' be 2 ng ml-'.The precision attained at a cocaine level of 0.5 pg ml-' was 3% and the linear analytical range extended from 10 x LOD to 100 pg ml-'.The sampling frequency was about 12 samples per hour. Interference Studies Only common alkaloids containing NR groups in their mol- ecule that could form ion pairs with BiI,- were investigated as potential interferents. The maximum permissible concen- trations of the alkaloids studied as interferences were estab- lished by determining 0.5 pg ml-' of cocaine in ethanolic solution in the presence of increasing amounts of each of the interferents following the general procedure. The tolerated limits of alkaloids potentially interfering with the proposed determination of cocaine are given in Table 2. As can be seen from the recoveries obtained in each case the method is sensitive and relatively selective because a 20-fold excess of quinine and heroin and a 65-fold excess of papaverine are tolerated.Application to Real Samples In order to validate the analytical capability of the proposed method the determination of cocaine in four confiscated samples was carried out. The samples were dissolved in ethanol and filtered through Sep-Pak cartridges into 50ml flasks. These ethanolic solutions were analysed for cocaine in our laboratory using the proposed method and by GC using a standardized procedure22 in an external laboratory. Comparative results are given in Table 3. Good agreement was obtained. CONCLUSIONS The tandem on-line continuous separation device using the Bi1,- as counter ion provides a clear separation of cocaine into chloroform as an ion pair and the indirect determination of this alkaloid by ICP-AES coupled on-line as a detector for bismuth was achieved.The proposed automatic method allows the indirect determination of this drug at pgl-' levels in the presence of other alkaloids and can be used as an alternative method to more conventional GC-ECD. The latter technique and HPLC determinations are laborious as they involve Table 2 Interferences in the determination of cocaine by the proposed method (analyte concentration 0.5 pg ml-' i= 223.06 nm) Interferent Heroin Morphine Codeine Quinine Papaverine Atropine Strychnine Methadone ~~ Maximum amount tolerated/ pg ml-' 10 15 18 10 34 15 20 20 Recovery 99 104 102 97 101 98 100 98 (%I Table 3 Cocaine determination in confiscated samples by routine GC-ECD and by the proposed method Cocaine ("/.) moll-' 1 .o Chloroform 1.5% m/v in DMF Sample no.1 2 3 4 This work 30.75 & 1.5 30.20 f 1 .O 26.13 f 0.7 27.24k0.5 GC-ECD 30.70 k0.7 30.18+ 1.0 26.19k0.9 27.31 k0.4 564 Journal of Analytical Atomic Spectrometry August 1996 Vol. 1 1several time-consuming extractions and purifications of the extracts prior to injection into the chromatograph. Other alternatives such as electrochemical or immunoassay tech- niques lack the desired selectivity. Moreover for more sensitive determinations by GC further preconcentration steps involving evaporation of the organic extracts obtained are needed leading to even longer analysis times. In contrast using the continuous on-line ICP-AES procedure recommended here comparatively low levels of cocaine (pg 1-I) can be monitored requiring only 5 min per sample.Indirect methodologies proposed using AAS d e t e ~ t i o n ' ~ - ~ ~ and the same extraction principle provided detection limits in the range 0.2-2.5 pg ml-' of cocaine. Financial support from the DGICyT (Project PB 94-1331) during this work is gratefully acknowledged. The provision of confiscated cocaine samples by the Laboratory of Conse- jeria de Sanidad del Principado de Asturias and Gas Chromatographic cocaine determination realized in the last laboratory by J. M. Cabeza is also deeply appreciated. REFERENCES 1 LeBelle M. J. Dawson B. Lauriault G. and Savard C. Analyst 1991 116 1063. 2 Okeke C. C. Wynn J. E. and Patrick D. S. Chromatographia 1994 38 52.3 Moore J. M. Casale J. F. Klein R. F. X. Copper D. A. and Lydon J. J. Chromatogr. A 1994 659 163. 4 Schwartz R. S. and David K. O. Anal. Chem. 1985 57 1362. 5 Jatlow P. and Nadim H. Clin. Chem. 1990 36 1436. 6 Sukbuntering J. Walters A. Chow H. H. and Mayersohn M. J. Pharm. Sci. 1995 84 799. 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Tagliaro F. Poiesi C. Aiello R. Dorizzi R. Ghielmi S. and Maringo M. J. Chromatogr. 1993 638 303. Crowther J. B. and Henion J. D. Anal. Chem. 1985 57 2711. Fernandez Abedul M. T. Barreira J. R. Costa A. and Tuiih P. Electroanalysis 1991 3 409. Okeke C. C. Wynn J. E. and Patrick K. S. Chromatographia 1994 38 52. Corburt M. R. and Koves E. M. J . Forensic Sci. Int. 1994 39 136. Kintz P. and Mangin P. Forensic Sci. Int. 1995 73 93. de laTorre R. Ortufio J. Gonzalez M. L. Farre M. Cami J. and Segura J. J. Pharm. Biomed. Anal. 1995 13 305. Chichuev Yu. A. Farmatsiya (Moscow) 1984 33 70. Hernandez A. Gutierrez P. and Thomas J. Farmaco Ed. Prat. 1986 41 300. Mirzaeva Kh. A. and Ivanova N. I. Zh. Anal. Khim. 1984 39 1691. Eisman M. Gallego M. and Valcarcel M. Anal. Chem. 1992 64 1509. Eisman M. Gallego M. and Valcarcel M. J. Anal. At. Spectrom. 1993 8 1117. Eisman M. Gallego M. and Valcarcel M. J. Pharm. Biomed. Anal. 1994 12 179. MenCndez Garcia A. Sanchez Uria J. E. and Sanz Medel A. Anal. Chim. Acta 1990 234 133. Nerin C. Garnica A. and Cacho J. Anal. Chem. 1986,58 2617. Leach H. and Ramsey J. D. in Clarke's Isolation and IdentiJcation ofDrugs ed. Moffat A. C. Pharmaceutical Press London 1986 Menendez Garcia A. Perez Rodriguez M. C. Sanchez Uria J. E. and Sanz Medel A. Fresenius' J . Anal. Chem. 1995,353,128. Menkndez Garcia A. Fernandez Sinchez M. L. Sanchez Uria J. E. and Sanz Medel A. Mikrochim. Acta 1996 122 157. Travnikoff B. Anal. Chem. 1983 55 795. pp. 178-201. Paper 6/02054B Received March 25 1996 Accepted May 28 1996 Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 565

ISSN:0267-9477    

DOI:10.1039/JA9961100561

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Determination of Dopants and Impurities in Optical Crystals of p=Barium Borate by Inductively Coupled Plasma Atomic Emission Spectrometry and Flame Atomic Absorption Spectrometry Journal of Analytical I Atomic I Spectrometry 1 ELISAVETA IVANOVA NONKA DASKALOVA SERAFIM VELICHKOV PETRANKA SLAVOVA AND GALJA GENTSCHEVA Institute of General and Inorganic Chemistry Bulgarian Academy of Sciences BG-SoJia 11 13 Bulgaria Inductively coupled plasma atomic emission spectrometry and flame atomic absorption spectrometry were applied to the determination of dopants (Ce Nd Eu and Er) and impurities (Fe Na Mg and Al) in the optical crystals of 8-barium borate dissolved in hydrochloric acid. The matrix interferences occurring in analyses carried out by means of ICP-AES and FAAS were studied. The detection limits of the analytes in a 0.8% solution of pbarium borate in 2 mol I-' hydrochloric acid were determined.The relative standard deviation of the obtained results lies within 2-6%. Keywords Dopants; impurities; P-barium borate; single crystals; inductively coupled plasma atomic emission spectrometry; flame atomic absorption spectrometry P-Barium borate is a relatively new and important material for non-linear optical applications in the visible and ultraviolet regions.'.2 It has a number of physical properties which make it particularly attractive. It exhibits large effective coefficients of second harmonic generation a wide transparent waveband high damage thresholds and high optical homogeneity. These properties suggest that P-barium borate will have important applications especially in the ultraviolet region.' The concept of improving or modifying the characteristics of non-linear optical crystals is not new.For many crystals of the solid-solution type it is possible to effect changes in certain material properties such as the birefringence and transparency by selecting an appropriate composition of the various elements in the c r y ~ t a l . ~ Doping is a way of varying this composition. On the other hand the incorporation of impurities in the crystal also affects its proper tie^.^ The correct interpretation of the properties of non-linear optical crystals and the control of their synthesis requires a knowledge of the concentration of dopants and impurities. This knowledge is acquired by analysis.The purpose of the present work was to determine the concentration of dopants (Ce Nd Eu and Er) and impurities (Na Fe A1 and Mg) in the optical crystals of P-barium borate with a view to the analytical control of their synthesis. Inductively coupled plasma atomic emission spectrometry is a technique particularly appropriate to this purpose.' Comparative data for the concentration of the trace impurities were obtained by flame atomic absorption spectrometry. EXPERIMENTAL Instrumentation The Jobin-Yvon (Longjumeau France) equipment and the operating conditions used in the ICP-AES experiments are specified in Table 1. T, was measured by the Boltzmann plot method.6 The FAAS measurements were performed with a Pye Unicam SP 192 spectrometer using the most sensitive lines of the elements under the operating conditions recommended in the manuals including deuterium background correction.Reagents and reference solutions Reagents of the highest available purity and triply distilled water from a quartz apparatus were used. The stock solutions of the elements (1 mg ml-') were prepared from Merck Titrisols when available (E. Merck Darmstadt Germany). The stock solutions of Ce Nd Eu and Er were prepared by dissolving the corresponding oxide (Specpure Johnson and Matthey Royston Hertfordshire UK) in hydrochloric acid as described el~ewhere.~ The reference solutions were prepared in 2 mol 1-' hydrochloric acid. The solvent blank was 2 mol 1-' hydrochloric acid while the matrix blank was a 0.8% m/v solution of undoped P-barium borate.Plastic laboratory ware was used throughout. Dissolution procedure A 200 mg amount of the finely ground crystalline material was dissolved in 2 mol 1-' hydrochloric acid in a polypropylene cup at room temperature. The solution was transferred into a Table 1 conditions Speciation of the spectrometer ICP source and operating Monochromator Mounting Grating Wavelength range Dispersion Entrance slit Exit slit Resultant spectral slit Practical spectral bandwidth Photomultiplier rf Generator Frequency Oscillator Power output Nebulizer Pump Operating conditions Incident power Reflected power Outer argon flow rate Carrier flow rate Liquid uptake rate Excitation temperature Transport efficiency of ICP system JY 38 (Jobin-Yvon) Czerny-Turner focal length 1 m Holographic 2400 grooves mm - ' 170-700 nm (1st order) 0.38 nm mm-' 0.02 mm 0.04 mm 15.2 pm 15.6 pm Hamamatsu TV R 446 HA Plasma Them Model HFP 1500 D Crystal controlled at 13.56 MHz Meinhard concentric glass Peristaltic ten-roller Gilson Minipuls I1 (Gilson Medical Electronics France) 27.12 MHz (+0.050/,) 0.5- 1.5 k W 1.0 kW 10 w 15 1 min-' 0.5 1 min -' 1.3 1 min-' 7200 K 3 yo Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 (567-570) 56725 ml polypropylene calibrated flask and was made up to the volume with the same acid.The concentration of P-barium borate in the solution obtained was 0.8% m/v. RESULTS AND DISCUSSION Dissolution of the Single Crystals of PBarium Borate The solubility of the crystalline material was tested in water nitric acid and hydrochloric acid.It was found that a-barium borate is not soluble in water and is hardly soluble in nitric acid. The most effective dissolution was achieved with hydro- chloric acid in the concentration range 2-3 moll-'. No heating was necessary. When hydrochloric acid of lower or higher concentration was used dissolution slowed down and even stopped. An optimum dissolution medium of 2 mol 1-1 HCl was chosen. Solutions of /?-barium borate with concentrations up to 2% m/v were prepared. The optimum concentration of the sample matrix in the solution was found to be 0.8% m/v which permitted a reliable determination of trace element concentrations at a relatively low level of matrix interference in both ICP-AES and FAAS. The solubility of the /?-barium borate samples doped with Ce Nd and Eu was the same as that of the undoped sample.Only the Er-doped sample exhibited lower solubility the 0.8% m/v solution of the Er-doped #?-barium borate was close to its saturated solution. Studies of the Determination of Ce Nd Eu Er Fe Mg Na and A1 in the Solution of #?-Barium Borate by Means of ICP-AES Spectral interferences and line selection Boumans discussed various aspects of spectral interferences in atomic emission spectros~opy.~~~ In the present study the spectral interferences around the prominent analysis lines of Ce Nd Eu Er Na Fe Mg and A1 were investigated in the presence of 5 mg ml-' of barium and 1 mg ml-' of boron as interferents in the solution these being their levels in the 0.8% m/v sample solution. Information on the interfering matrix lines in a A k300 pm spectral window for each of the interferents was derived from wavelength scans centred around the analysis lines Aa of the analytes.Fig. 1 shows an example. El~ewhere,'.~ we have accurately specified how the results were obtained and quantified. The interpretation of the results revealed that firstly all investigated analysis lines are free of line and wing interferences in the presence of 1 mg ml-' boron (the latter has a poor emission spectrum 94 identified emission lines are listed in the MIT Tables"); secondly barium has a richer emission spec- trum than boron (472 identified emission lines are listed in the MIT Tables"). With 5 mg ml-' of barium as interferent the analysis lines under investigation were influenced by wing interferences.Line interferences were established around two analysis lines Ce 11 413 380 pm was overlapped by Ba 413243pm and Eu I1 412970 pm by Ba I1 413066pm (see Fig. 1). The identified interfering barium lines in the spec- tral windows of cerium and europium analysis lines are not included in the Boumans Line Coincidence Tables" but are listed in the MIT Tables." These data provided the possibility of correct evaluation of the type of spectral interferences and optimum line selection.' To the latter purpose the true detection limit criterion of Boumans and Vrakking was applied.12 According to this approach Q-values for line interference [QI(Aa)] and Q-values for wing (background) interference [Q,(AA,)] for each of the above mentioned interferents were distinguished.The term QI(Aa) is expressed as the ratio SI(Aa)/SA where S,(A,) is the partial sensitivity of the interfering line defined as the signal -h Ba 41 3 243 412 970 WavelengWpm Fig. 1 Example of a spectral scan over a spectral region of & 300 pm around the analysis line. Central wavelength 412 970 pm (Eu line). Interferent barium. Analytes are 4 pg ml-' of europium in pure solvent or 5 mg ml-' of barium per unit interferent concentration produced by the interfering line at the peak wavelength of the analysis line A and SA is the sensitivity of the analysis line. The term Qw(Aha) is expressed as the ratio Sw(AA,)/SA where Sw(AAa) is the wing sensitivity of the interfering line in the spectral window AA and SA is as stated above. Table 2 lists Qw (Ah,) and Q,(h,) values in the presence of 5 mg ml-l of barium in solution (as was mentioned above boron does not contribute to the Q- values).Evidently the fi-barium borate matrix causes rather Table 2 Q,(AA,) and QI(Aa) values of the investigated analysis lines of Ce Nd Eu Er Na Fe A1 and Mg. Interferent 5 mg ml-' barium Analysis linefpm Ce I1 413 765 Ce I1 413 380* Ce I1 395254 Nd I1 401 225 Nd I1 430358 Nd I1 406 109 Eu I1 381 967 Eu I1 4129707 Eu I1 420505 Er I1 337271 Er I1 349910 Er I1 323058 Na I 588995 Fe I1 238204 A1 I 396 151 Mg I1 279 553 Q w ( W 6.0 x lo-' 7.4 x lo-' 8.2 x 10-5 1.3 x lo-' 1.1 x 9.6 x 1.9 x 3.3 x 1.6 x lop6 2.1 x lo-' 7.6 x 8.8 x 6.6 x lop6 3.5 x 10-5 5.9 x 1.2 x 10-6 QI (La ) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8.9 x 10-5* 2.8 x 10-5+ * Overlapping of Ce I1 413 380 pm with Ba 413 243 prn.'' Overlapping of Eu I1 412970 pm with Ba I1 413066 pm." 568 Journal of Analytical Atomic Spectrometry August 1996 VoL 1 1low interferences.This permitted us to use the most sensitive lines of the analytes as analysis lines. Efect of the P-barium borate matrix on the net line intensities of the analytes The net line intensities decreased by about 5% in the presence of 0.8% m/v P-barium borate as a matrix in the solution compared with the net line signals in the pure solvent (2 mol 1-1 HCl). The effect is negligible as regards the detection limits but reflects on the accuracy of the analysis. The correct calibration requires precise matching of the matrix content. Detection limits By use of the Qr(Aa) and Qw(AAa) values the conventional detection limits (CL,conv) and the true detection limits (C,,,,,) were determined in the 0.8% m/v solution of P-barium borate (Table 3).For the sake of comparison the detection limits in the pure solvent (2 mol 1-1 HC1) are also listed (C,). As can be seen the true detection limits of the ICP-AES determination in the presence of the P-barium borate matrix do not signifi- cantly differ from those in a pure solution (column 1) only in the case of wing interference. With additional line interference substantially higher true detection limits are registered. The detection limits with respect to the dissolved solid are of interest rather than the detection limits in solution. Table 4 shows the true detection limits of the analytes with respect to the dissolved solid (0.8% m/v P-barium borate) using the 'best' analysis lines (column 1).The fictitious detection limits i.e. the detection limits in the presence of the same matrix assuming no contribution of an interfering signal at the peak wavelength of the analysis line are also shown (column 2). Study of the Matrix Effects on the Flame AAS Determination of Fe Na Mg and Al in a Solution of /?-Barium Borate The detection limits of the FAAS determination of Fe Na Mg and A1 in /I-barium borate (0.8% m/v solution of P-barium borate in 2 mol 1-' HCl) determined from the fluctuations of the blank signal (30 criterion) are presented in Table 5. Table 3 CL CL,conv and CL,true values of the prominent analysis lines of Ce Nd Eu Er Na Fe A1 and Mg in the presence of 0.8% m/v P-barium borate in solution Analysis line/pm Ce I1 413 765 Ce I1 413 380 Ce I1 395254 Nd I1 401 225 Nd I1 430358 Nd I1 406 109 Eu I1 381 967 Eu I1 412970 Eu I1 420505 Er I1 337271 Er I1 349 910 Er I1 323058 Na I 588995 Fe I1 238204 A1 I 396 151 Mg I1 279 553 CL/ ng ml-' 17.0 17.0 17.0 40.0 50.0 50.0 0.7 1.4 2.0 3.0 3.7 3.7 19.0 8.4 19.0 0.1 CL,conv( CL true /' ng ml- ng ml- 25.0 40.0 215.0 28.0 - - 42.0 - 52.0 51.0 - - 1 .o 6.0 55.0 2.2 - - 6.0 5.0 4.7 - - - 20.0 - 13.0 - 20.0 - 0.2 - Table4 True detection limits with respect to the dissolved solid (0.8% m/v P-barium borate) for the 'best' analysis lines of the elements (column 1).Fictitious detection limits in the presence of the same matrix supposing no matrix interference (column 2) Detection limit/pg g-' 'Best' analysis line/pm Ce I1 413 765 Nd I1 401 225 Eu I1 381 967 Er I 337271 Na I 588995 Fe I1 238204 A1 I 396 151 Mg I1 279 553 1 3.14 5.25 0.13 0.75 2.38 1.63 2.38 0.25 2 2.13 5.00 0.09 0.38 2.38 1.05 2.38 0.02 Table 5 Detection limits (30 criterion) of the FAAS determination of the impurity elements with respect to P-barium borate Element Na Fe A1 Mg Detection limit/pg g-' 1.7 10.0 150 0.35 The effect of the P-barium borate matrix on the analytical signal of the elements was studied.To this purpose the ratio between the slope of the curve of standard additions in the 0.8% m/v solution of P-barium borate and that of an aqueous calibration curve was determined (see Table 6). Since the content of A1 in P-barium borate was found to be below its detection limit in flame AAS it was spiked to the test solutions at a level of l o x the detection limit.As can be seen the matrix P-barium borate significantly suppresses the analytical signal the slope ratio is less than unity for all analytes examined. This imposes the use of standard additions for adequate calibration. Analysis of Single Crystals of /?-Barium Borate Single crystals of undoped P-barium borate and P-barium borate doped with Ce Nd Eu and Er respectively were analysed for their dopant and impurity content after dissolution according to the procedure. The concentration of dopants was determined by ICP-AES. The results are presented in Table 7. The mean of three separate dissolutions is shown. The relative standard deviation for Er is 2.6%. It follows from the results Table6 Slope ratio between the curve of standard additions in a 0.8% m/v solution of P-barium borate in 2 mol 1-' HCl and the calibration curve in 2 mol 1-' HCl Element Na Fe Mi3 A1 Slope ratio 0.85 0.70 0.77 0.86 Table 7 Content of dopants in the optical crystals of P -barium borate Added to the initial Found in the RSD Dopant solid solution (% m/m) crystal (% m/m) (%) Ce 0.62 - Nd 0.64 Eu 0.68 Er 0.75 0.39 2.6 - * - - - - ~~ * Below the corresponding detection limit of the ICP-AES determi- nation (see Table 4).Journal of Analytical Atomic Spectrometry August 1996 VoI. 1 1 569Table8 Content of impurities in the optical crystals of /&barium borate ICP-AES FAAS Content/ RSD Content/ RSD Impurity P8 g-' W) P8 g-' (%) Na 460.0 2.0 489.0 2.5 Fe 36.0 2.5 35.0 2.9 A1 59.0 4.2 * - - Mg 7.0 5.9 5.9 5.5 * Below the detection limit of the FAAS determination (see Table 5).obtained that the only successful doping of @-barium borate achieved was that with Er. No appreciable amounts of Ce Nd or Eu have been incorporated in the crystals. The content of impurities in the single crystals of /?-barium borate was determined by ICP-AES and FAAS. The results obtained for the undoped sample are shown in Table 8. Each value is a mean of three separate dissolutions. The relative standard deviation of all results obtained by the two methods lies within 2-6%. The content of impurities in the doped crystals of @-barium borate was found to be at the same level as that shown for the undoped sample (Table8). Hence the impurity content primarily depends on the purity of the starting substances and of the environment during the synthesis (e.g.furnace utensils ambient air). The financial support from the National Fund for Scientific Research of the Ministry of Science Education and Technology of Bulgaria under registration no. x-335 is gratefully acknowledged. REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 Feigelson R. S. Raymakers R. J. and Route R.K. Prog. Cryst. Growth Charact. 1990 20 115. Miyazaki K. Sakie H. and Sato T. Opt. Lett. 1986 11 797. Laudise R. A. in Crystal Growth and Characterization ed. Ueda R. and Mullin J. B. North-Holland Amsterdam 1975. Morris P. A. J. Crystal Growth 1990 106 76. Daskalova N. Velichkov S. Slavova P. Ivanova E. and Aleksieva L. Spectrochim. Acta Part B submitted for publication. Mermet J. M. in Inductively Coupled Plasma Emission Spectroscopy Part 2 Applications and Fundamentals ed. Boumans P. W. J. M. Wiley New York 1987 ch. 10 p. 353. Daskalova N. Velichkov S. Krasnobaeva N. Slavova P. Spectrochim. Acta Electronica included in Spectrochim. Acta Part B 1992 47 E1595. Boumans P. W . J . M. Fresenius' Z . Anal. Chem. 1986 324 397. Boumans P. W. J . M. in Inductively Coupled Plasma Emission Spectroscopy Part I Methodology Instrumentation and Performance ed. Boumans P. W. J. M. Wiley New York 1987 p. 358. Harrison G. R. MIT Wavelength Tables The MIT Press Cambridge MA USA 1969 p. 1939. Boumans P. W . J . M. Line Coincidence Table for Inductively Coupled Plasma Emission Spectrometry Pergamon Press Oxford 1980. Boumans P. W. J. M. and Vrakking J. J . A. M. Spectrochim. Acta Part B 1987 42 819. Paper 6/01 980C Received March 21 1996 Accepted May 22 1996 570 Journal of Analytical Atomic Spectrometry August 1996 Vol. 11

ISSN:0267-9477    

DOI:10.1039/JA9961100567

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Ultratrace Determination of Cadmium by Atomic Absorption Spectrometry Using Hydride Generation With in situ Preconcent rat ion in a Pallad ium-coated Graphite Atomizer Journal of Analytical Atomic Spectrometry HEIDI GOENAGA INFANTE MARIA L. FERNANDEZ SANCHEZ AND ALFREDO SANZ-MEDEL* Department of Physical and Analytical Chemistry University of Oviedo Julian Claveria 8 33006-Oviedo Spain A very sensitive ETAAS method with in situ trapping of the metal is described for ultratrace determinations of Cd in liquid solutions. It is based on volatile Cd HG from a vesicular solution with subsequent trapping of the metal in a graphite platform coated with Pd. This Pd coating on a pyrolytic graphite L'vov platform acts as the atomizer. Sodium tetrahydroborate(Ir1) in a vesicular medium of didodecyldimethylammonium bromide is used to generate Cd hydride from the sample solution.This volatile species is trapped by the Pd coating pre-heated at 150"C the preconcentrated Cd then atomized at 1600 "C and the transient AAS signal measured. Detection limits (DL) observed are in the ng I-' range (e.g. using a 1.4 ml Cd sample volume for preconcentration the DL was 60 ng I-'). Higher preconcentrations up to a maximum of 1.4 ng of total Cd deposited are possible with the proposed methodology. The RSD for ten replicate analyses of 1.4 ml of 500 ng I-' Cd solution (0.7 ng of total Cd) was 1.7%. The combination of vesicles-mediated HG with the ease of decomposition of Cd hydride on the Pd coating at 150°C provides a convenient Cd preconcentration method for enhancing the sensitivity of direct determinations of Cd by conventional ETAAS or by HGAAS.Keywords Cadmium hydride generation; in situ preconcentration; palladium coated graphite tube; electrothermal atomic absorption spectrometry Trace and ultratrace determinations of cadmium in environ- mental and biological samples have become of increasing interest owing to the high toxicity of this metal.' Although several highly sensitive and selective analytical techniques have been applied for the determination of such low levels of cadmium ETAAS is probably the method of choice for this determination. Hydride generation techniques result in more efficient trans- port of the analyte to the atomizer and can be used to significantly improve flame AAS detection limits for a number of elemenk2 Several approaches have been proposed for enhancing atomic absorption based either on direct transfer of the generated hydride into the a t o m i ~ e r ~ - ~ or its collection (trapping) for concentration prior to its introduction into the atomization cell.7-" Recently 'trapping' procedures have been which utilize the graphite furnace as both a trapping device and an atomization cell.These procedures currently provide the most sensitive atomic spectrometric methods available for the detection of volatile hydride forming elements. '* Zhang et al." used a palladium-coated graphite tube for sorption of arsenic antimony and selenium. Their experiments * To whom correspondence should be addressed. proved that palladium on the surface of the graphite tube acted as an efficient adsorbent for the metallic hydrides.Sturgeon et aL2' further studied the trapping of volatile element hydrides by platinum group elements for in situ preconcen- tration-ETAAS using a graphite furnace. The analytical per- formance and mechanism of sequestration of the hydrides using different platinum metals as collection substrates were thoroughly evaluated. On the other hand our group has reported that organized media can be used advantageously to improve the generation of volatile species for AAS determination^.^^-^^ Particularly important was the development of a continuous-flow method for the generation of volatile cadmium species from an aqueous solution containing different surfactants using NaBH as the reduction agent.The most promising analytical results were obtained with didodecyldimethylammonium bromide (DDAB) vesicles as the organized medium for Cd2+ reduction.23 Experiments showed that such volatile species were able to transport the metal to the atomic measurement cell as the volatile species cadmium h~dride.~ However at room tem- perature CdH2 is very unstable and it would decompose generating Cdo and hydrogen. This peculiar mechanism was proposed recently2' to explain the first cold vapour-AAS determination of cadmium reported in the Moreover this mechanism would allow the transport of CdH and the preconcentration of CdO onto an adequate con- veniently heated L'vov platform surface. The decomposition reaction could be favoured by the well known adsorption properties of H2 in Pdo formed in the coating of the trapping graphite platform surface with a palladium nitrate s~lution.'~ This paper describes an ultrasensitive method for the precon- centration and determination of very low levels of cadmium based on the continuous generation of volatile cadmium from vesicles of DDAB with NaBH with subsequent trapping of the metal in a graphite platform coated with palladium.Atomization and AAS measurements are then carried out for the determination of very low levels (picograms) of the trapped cadmium metal. EXPERIMENTAL Apparatus Analytical AAS measurements were made using a Perkin- Elmer Model 3030 atomic absorption spectrometer fitted with an HGA-500 graphite furnace atomizer. A Perkin-Elmer cad- mium hollow cathode lamp was used. The analyte absorption peaks were printed out using a Perkin-Elmer PR-100 printer.The volatile cadmium hydride was introduced into the graphite tube uia a glass capillary tube (25 mm long x 1 mm id) with its final end in contact with the L'vov platform." Argon was used as the carrier gas. The other end of the glass Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 (571-575) 571NaBH'InDDAB C-n Peristaltic pump cd (in DDAB) + HCI i lkflon tubing Graphite atomizer rwT:-) Gar - liquid Gar ' Ar Fig. 1 the platform Instrumental set-up for continuous HG and Cd collection in tube was connected to the outlet of the gas-liquid separator2 by a PTFE tube (1 1 cm long x 3 mm id). A schematic diagram of the instrumentation used for HG generation and collection in the graphite tube (atomizer) is presented in Fig.1. A high-intensity ultrasonic processor (Sonics and Materials Danbury CT USA) Model 500 W was used for vesicles generation. Reagents A 1000mg 1-' cadmium (as cadmium nitrate) stock standard solution (Merck Darmstadt Germany) was used. Cadmium working solutions were freshly prepared daily by diluting appropriate aliquots of the stock solution in ultrapure water (Milli-Q Millipore). Sodium tetrahydroborate(II1) solutions were prepared by dissolving NaBH (PROBUS Barcelona Spain) in ultrapure water stabilized in 0.1 YO (m/v) sodium hydroxide solution. These solutions were prepared weekly and always filtered before use. moll-l) were prepared by dissolv- ing the surfactant powder (Fluka Buchs Switzerland) in ultrapure water and sonicating at room temperature at a power of 60 W for about 12 min with the tip of the high intensity ultrasonic processor (this procedure is referred to here as 'sonication').A 1000mg 1-' palladium (as palladium nitrate) standard solution (Merck) was used. All mineral acids and metal salts used were of analytical- reagent grade and ultrapure water was used throughout. Vesicles of DDAB ( Procedures Continuous volatile cadmium species generation from DDAB In the flow system shown schematically in Fig. 1 the cadmium sample (dissolved in 0.4 mol 1-1 HCI and lop2 moll-' DDAB vesicles) is continuously pumped through one of the channels of the peristaltic pump at 1.4 ml rnin-l and merged with a 4% (m/v) solution of NaBH (flow rate 1.4 ml min-l) dissolved in mol I-' DDAB vesicles.The resultant solution feeds the entrance of a grid nebulizer which is used here totally detuned in order to allow for separation of the gaseous species from the liquid sample which is sent to waste. In this way the grid nebulizer operates as a convenient gas- liquid separator23 allowing the volatile species to reach the graphite tube while the liquid phase goes to drain. Cd collection and atomization The sequence of operations for collection and atomization of the species of Cd was as follows before starting a typical HG cycle 10 pl of 500 mg 1-1 palladium solution were injected 572 Journal of Analvtical Atomic SDectrometrv Aunust 1996 onto the L'vov platform in the centre of the graphite tube using an Eppendorf microlitre pipette fitted with a disposable polypropylene tip and dried at 120 "C for 25 s. After drying the furnace was heated at 800 "C for 15 s to complete the thermal treatment for adequate Pd coating of the platform; the graphite tube was then cooled to the sorption temperature (150 "C).At this point the peristaltic pump was started for continuous volatile Cd generation followed by insertion of the glass capillary tip into the graphite tube to start the Cd collection period (60 s). The generated volatile species were swept with argon from the gas-liquid separator into the L'vov platform in the furnace (see Fig. 1). Once the collection period was completed the glass tube was removed from the graphite furnace to allow for atomization. The analyte was atomized at 1600 "C in a period of 4 s using maximum power and the internal argon interrupted flow mode.After measurement of the AAS signal the tube was fired at 2700 "C for 2 s for the final cleaning step. RESULTS AND DISCUSSION Preliminary Studies Optimization of Cadmium Determination for ETAAS Preliminary AAS conditions for conventional ETAAS determi- nation of Cd were optimized the optimum instrumental set- tings used are summarized in Table 1. The pyrolysis and atomization behaviour of 10 pl of a 10 pg I-' Cd aqueous solution were investigated in detail using palladium as a chemical modifier26 and a L'vov platform. A 5 pg portion of total Pd were added to the platform using the palladium nitrate solution. The Cd pyrolysis curve was obtained at an atomization temperature of 2000 "C while the metal atomization curve was carried out at a fixed pyrolysis temperature of 800 "C.Fig. 2 shows the observed results for pyrolysis and atomiz- ation of Cd in solution with Pd modifier. As can be seen optimum figures for the temperature programme were 800 "C for pyrolysis and 1600 "C for atomization. Once the pyrolysis and atomization temperatures were selec- ted the total amount of chemical modifier added per 10 p1 of sample was investigated; 10 pl of a 10 pg 1-1 Cd solution were injected into the graphite tube and the ETAAS signals observed for increasing palladium nitrate concentrations added to the platform were recorded. The results showed that addition of 3-5 pg of Pd before the Cd sample on the L'vov platform was enough to secured a constant AAS signal.Cadmium Hydride Generation-in Situ Preconcentration and Atomization in a Palladium-coated Graphite Tube Sorption-atomization temperature studies Following on from data obtained in previous we optimized the continuous flow CdH2 generation in this experi- Table 1 Optimal instrumental settings used Wavelength Lamp current Slit-width Gain Background correction Reactor length Final flow of sample Drain flow Tube type Signal processing Integration time Ashing temperature Atomization temperature 228.8 nm 4.0 mA 0.7 nm 75.0 V Deuterium 7.0 cm 2.8 ml min-' 2.8 ml min-' Pyrolytic graphite with L'vov platform Peak height 4.0 s 800 "C 1600 "C VOl. 11Fig. 2 Optimization of conventional Cd determination by ETAAS. (a) Pyrolysis curve at atomization temperature of 2000 "C; and (b) atomization curve at pyrolysis temperature of 800°C.Ten pl of 10 pg 1-' Cd aqueous solutions were used for both experiments. ment. Optimal flows observed and other instrumental param- eters selected for subsequent work are given in Table 1. By using preconcentration of 1.4 ml of a solution containing 500 ng 1-l of Cd and the conditions given in Table 1 the effect of temperature of the graphite tube (sorption temperature) on the absorbance (peak height) of analyte preconcentrated on the platform was studied. The results observed for different sorption and atomization temperatures are presented in Fig. 3. As can be seen in Fig. 3(a) a plateau of the analytical AAS signal was obtained for a sorption temperature range of 50-300°C. It is interesting to 0.1 2 I 20 420 820 1220 1620 Temperature / 'C Fig.3 Effect of sorption and atomization temperatures on absorbance. (a) Sorption temperature at an atomization temperature of 1600°C (for 0.7 ng of total Cd in the sample solution); and (b) atomization temperature at a sorption temperature of 150 "C (for 0.7 ng of Cd as above). note that even at room temperature the volatile species of Cd formed seems to deposit onto the L'vov platform with high efficiency. This is probably due to the instability of CdH which at room temperature appears to decompose producing a 'cold vapour' of cadmium according to:" CdH %Cdo + H This decomposition reaction (and therefore the analytically sought deposit/preconcentration of the metal onto the Pd coating) should be favoured by (a) the well known adsorption properties of Pd metal for molecular H which would displace the above reaction to the right; and (b) the rise in temperature of the platform trapping the hydride.Negligible AAS signals of cadmium were obtained at a temperature range of 100-500 "C using pyrolytic graphite L'vov platforms. Thus the efficiency of collection is enhanced by coating the graphite tube with palladium (probably uia adsorption of H,). High values of background absorption were observed at sorption temperatures under 100 "C (due to water vapour condensation in the graphite tube wall in the sorption step). Therefore a temperature of 150°C was finally selected for Cd trapping. An atomization curve was constructed by fixing the CdHz trapping temperature at 150 "C and varying the atomization temperatures.The results obtained have been plotted in Fig. 3( b) and indicate that atomization of cadmium is complete between 1400-1800 "C. Therefore 1600 "C was again selected for atomization of the Cdo deposit onto the Pd coating (as in the conventional ETAAS methodology see Table 1). Table 2 summarizes the different settings selected in the furnace for in situ preconcentration-atomization of Cd. Influence of carrier Arflow rate The influence of carrier Ar flow rate (see Fig. 1) on the absorbance of Cd (a 500 ng 1-' solution preconcentrated for 1 min under continuous flow at 1.4ml min-' and final measurement of the ETAAS signal under the conditions given in Table 2) was investigated. An increase of the Cd absorbance was observed up to 145ml min-' at which point a plateau was reached.An argon flow rate of over 220ml min-' orig- inated DDAB carry over into the graphite tube and thus high values of background absorption. Therefore a carrier gas flow rate of 150 ml min-' was finally selected. Table 2 Furnace programme for in situ pre-concentration and atomization Palladium coating of the graphite platform The effect of the total amount of palladium deposited onto the L'vov platform as a thermal coating on the cadmium ETAAS signal obtained from the preconcentration of 1.4 ml of 500 ng 1-' solution of Cd was investigated in more detail using integrated absorbance measurements. Fig. 4(a) shows the results obtained in these experiments. As can be seen the integrated absorbance (peak area) obtained increased rapidly over the range 1-4 pg of deposited palladium ( 10 pl of solutions Step Temperature/ Ramp Hold time/ "C time/s S Inner argon flow rate/ml min-' 1 Drying* 120 10 25 300 2 Pyroly sis/pre- trea tmen t * 800 10 15 300 3 Collection 150 10 80 10 4 Atomization 1600 0 4 0 5 Cleaning 2700 1 2 300 * Thermal treatment of the platform.Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 573Table 3 Comparison between conventional and in situ preconcen- tration-ETAAS analytical performance characteristics for determi- nation of cadmium 0 5 10 15 28 Palladium amount / ccg Fig.4 Effect of amount of palladium on integrated absorbance. (a) Pd for coating (1.4 ml of 500 ng 1-' Cd sample solution); and (b) Pd as chemical modifier (10 pl of 10 pg 1-1 Cd aqueous sample solution).of 100-400 mg 1-' of Pd added before thermal treatment and CdH sorption). In other words it seems that with increasing mass of Pd' the platform surface area available for analyte trapping should also increase,,' so its trapping efficiency of CdH increases accordingly until a final plateau is reached above about 5 pg of total Pd on the platform. We also investigated the influence of the volume of the coating palladium solution added to the platform on the Cd trapping efficiency for a constant total mass of Pd deposited of 5 pg. Over the entire range of volumes (2.5-50 p1 of Pd nitrate solutions added at different concentrations) studied a constant AAS signal was observed. It appears that the whole platform surface is Pd' coated even for low volumes of coating solution (good wettability) and this would explain the excellent efficiency of trapping observed.Therefore a treatment of the L'vov platform with 10 p1 of the palladium solution of 500 mg 1-' was selected as the optimum Pd coating of the graphite. The results obtained for the effect of Pd chemical modifier added on cadmium determination by conventional ETAAS are presented in Fig. 4( b). Interestingly the curves presented in Figs. 4(a) and 4(b) are the same shape (i.e. the optimum amount of total Pd on the graphite platform is about 5pg). However the integrated absorbance values obtained for con- ventional ETAAS determination [ Fig. 4( b)] were 2.5 times higher than those observed for HG-in situ preconcentration- ETAAS [Fig. 4(a)]. This observation could be ascribed to serious losses in the gas generation-transport-deposition pro- cesses responsible for preconcentration.According to previous work2' the efficiency of generation of the volatile cadmium species at room temperature is relatively low (49.8% & 3.4). Of course some losses of cadmium can be expected in the transport (metal deposition in tubing). Moreover the trapping of CdH may not be quantitative in the platform. All these processes account for the overall loss of the AAS signal observed (see Fig. 4). However by generating the CdH2 species at lower temperatures the generation efficiency increases [e.g. up to 75% by generation at 0°C (ref. 25)] and therefore the AAS signal should also increase thereby improving the precon- centration process. Analytical Performance Characteristic data defining the analytical performance of the proposed Cd in situ preconcentration-ETAAS system are Analytical Conventional In situ preconcentration characteristics ETAAS ETAAS Detection limit/ng 1-' 200 60* 13t Precision (%) 1.7 at 0.5 pg 1-' 2.5 at 4 pg 1-' Linear range/pg 1-' up to 10 up to 1* * For 1.4 ml of preconcentrated Cd solution.t For 7 ml of preconcentrated Cd solution. summarized in Table 3 in a comparison of the same analytical performance characteristics observed for our instrument using conventional ETAAS of Cd. This Table illustrates more clearly the benefits of in situ preconcentration of the metal. As shown the detection limit (30) was calculated to be 60 ng 1-l for 1.4 ml of sample solution preconcentrated.Of course lower detection limits could be obtained if higher sample volumes were preconcentrated. Experiments showed that the AAS signal obtained did not depend on the volume of sample solution (e.g. 5 ml of 100 ng 1-' Cd solution and 1 ml of 500 ng 1-' of Cd solution after preconcentration following the rec- ommended procedures brought about the same ETAAS observed signals). In fact preconcentration (7 ml of Cd solu- tion) over 5min resulted in a detection limit of 13 ng 1-' (Table 3). The reproducibility of the recommended method for ultra- traces of Cd was evaluated by assessing the precision of the observed AAS signals 1.4 ml of a 500 ng 1-' Cd solution. Analysis of 10 replicates by this procedure (i.e. HG-in situ trapping-ETAAS) resulted in an RSD of The HG-in situ trapping-ETAAS procedure for Cd pro- posed here was calibrated using standards in the range 100-2000 ng 1 -' with continuous generation of volatile cad- mium species from DDAB and preconcentration and atomiz- ation conditions as detailed in Tables 1 and 2.Calibration points were based on the average of triplicate measurements of each Cd standard corrected for its blank. The calibration graph fitting was carried out by linear regression. The graph was linear up to 1000 ng 1-' of Cd and a correlation coefficient of 0.998 was obtained between the ETAAS signal and Cd concentration. Potentially interfering elements on Cd hydride generation were tested (only interferences in the CdH generation step should be expected in our procedure). Hydride-forming elements and high levels of alkali alkaline earth metals or common anions were found not to affect cadmium HG in the DDAB medium.Only interferences from some elements present at very high concentrations in the sample such as Zn and Ni,23 were found to seriously effect the preconcentration- determination procedures. 1.7%. CONCLUSIONS The Cd preconcentration method described here has been developed using straightforward in situ ETAAS detection. The sensitivity precision and reliability of the proposed method however warrants its future application in connection with other high sensitivity detectors (e.g. ET-ICP-MS) for extremely low level determinations of this toxic metal. The authors wish to thank the International Cooperation Institute Madrid Spain for awarding a doctoral grant to H.Goenaga. Financial support from DGICYT (Spain) through project number PB 94-1331 is gratefully acknowledged. The authors also thank J. M. Marchante Gayon for helpful dis- cussions on ETAAS. 574 Journal of Analytical Atomic Spectrometry August 195'6 Vol. 1 1REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Robards K. and Worsfold P. Analyst 1991 116 549. Godden R. G. and Thomerson D. R. Analyst 1980 105 1137. Nakahara T. Prog. Anal. At. Spectrosc. 1983 6 163. Welz B. and Schubert-Jacobs M. At. Spectrosc. 1991 12 91. Dedina J. and Tsalev D. L. in Hydride Generation Atomic Absorption Spectrometry J. Wiley and Sons New York 1995. Welz B. and Stauss P. Spectrochim. Acta Part B 1993 48 951. Holak W. Anal. Chem. 1969 41 1712. Maher W. A. Anal.Lett. 1983 16 801. Lee D. S. Anal. Chem. 1982 54 1682. Sturgeon R. E. Willie S. N. and Berman S. S. J. Anal. At. Spectrom. 1986 1 115. Sturgeon R. E. Willie S. N. and Berman S. S. Anal. Chem. 1987 49 2441. Doidge P. S. Sturman B. T. and Rettberg T. M. J. Anal. At. Spectrom. 1989 4 251. Chaudhry M. M. Ure A. M. Cooksey B. G. Littlejohn D. and Halls D. J. Anal. Proc. 1991 28 45. Yan X.-P. and Ni Z.-M. J. Anal. At. Spectrom. 1991 6 483. Ni Z.-M. He B. and Han H.-B. J. Anal. At. Spectrom. 1993 8 995. Liao Y.-p. and Li A.-m. J. Anal. At. Spectrom 1993 8 633. 17 18 19 20 21 22 23 24 25 26 Tao G.-L. and Fang Z.-L. J. Anal. At. Spectrom. 1993 8 577. Vien S. H. and Fry R. C. Anal. Chem. 1988 60 465. Zhang L. Ni Z.-m. and Shan X.-q. Spectrochim. Acta Part B 1989 44 751. Sturgeon R. E. Willie S. N. Sproule G. I. Robinson P. T. and Berman S. S. Spectrochim. Acta Part B 1989 44 667. Fernandez M. R. Segovia E. Valdes-Hevia M. C. Aizpun B. Marchante J. M. and Sanz-Medel A. Spectrochim. Acta Part B 1995 50 377. Sanz-Medel A. Fernandez M. R. ValdCs-Hevia M. C. Aizpun B. and Liu Y. M. Talanta 1993 40 1759. ValdCs-Hevia y Temprano M. C. Fernandez de la Campa M. R. and Sanz-Medel A. J. Anal. At. Spectrom. 1993 8 847. Sanz-Medel A. ValdCs-Hevia y Temprano M. C. Bordel Garcia N. and Fernandez de la Campa M. R. Anal. Proc. 1995,32,49. Sanz-Medel A. Valdes-Hevia M. C. Bordel N. and Fernandez M. R. Anal. Chem. 1995,67 2216. Smeyers-Verbeke J. Yang Q. Penninckx W. and Vandervoort F. J. Anal. Atom. Spectrom. 1990 5 393. Paper 6100394 J Received January 18 1996 Accepted May 17 1996 Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 575

ISSN:0267-9477    

DOI:10.1039/JA9961100571

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Determination of Trace Amounts of Antimony Germanium and Tin in High-purity Iron by Electrothermal Atomic Absorption Spectrometry After Reductive Coprecipitation With Palladium TETSUYA ASHINO AND KUNIO TAKADA Institute for Materials Research Tohoku University Katahira 2-2-1 Aoba Sendai Miyagi 980-77 Japan Trace amounts of antimony germanium and tin in high-purity iron were quantitatively separated by a reductive coprecipitation technique with palladium and determined by electrothermal atomic absorption spectrometry. When sodium phosphinate (NaPH,O,) was used as a reductant antimony and germanium could be separated simultaneously from large amounts of iron. Similarly when sodium tetrahydroborate (NaBH,) was used germanium and tin could also be separated simultaneously. The atomic absorbances of antimony germanium and tin were increased by about 1.5 3.7 and 4.5 times respectively in the presence of palladium.The limits of detection (corresponding to three times the standard deviation of the blank) of antimony germanium and tin were 0.Ol9 O.Ol0 and O.o& pg g-' respectively. Keywords Atomic absorption spectrometry; reductive coprecipitation; high-purity iron; antimony; germanium; tin; palladium The characteristics of high-purity metals have been widely investigated. For example it has become apparent that 99.999% (5N) iron differs from 99.95% (3N5) iron in physical and chemical properties.' In order to assess the purity trace elements have to be determined. The reported methods for the determination of antimony and tin2-5 cannot really be applied to 5N iron.Methods for the determination of germanium in iron have not been reported. Therefore more sensitive methods for the determination of trace amounts of antimony germanium and tin in high-purity iron are needed. When the elements in metal or alloy samples are to be determined the sample is usually decomposed and dissolved and the sample solution is used for the determination. For the deter- mination of trace amounts of antimony germanium and tin spectr~photometry,~.~ voltammetry,8-'0 HG-ICP-AES," HG-MIP-AES,I2 HG-1CP-MS,l3 FAAS,I4 HGAAS"-I8 and ETAAS'9-23 have been reported. In the HG methods the elements for determination can be separated from the metal matrix and the sensitivity may be higher than in ETAAS. However the conditions for generation of the hydride must be chosen according to the sample composition and it is not possible to determine many elements from a single sample.Therefore it is difficult to determine trace amounts of anti- mony germanium and tin in a large number of iron samples. In NAA,24 decomposition of the sample is not required but special apparatus is needed. ETAAS is suitable for the deter- mination of trace amounts of antimony germanium and tin in high-purity iron because these elements can be determined from a single sample by the same procedure and no special techniques are needed. However it may be necessary to separate and concentrate trace amounts of these elements from iron not only because the sensitivity of the deter- mination of these elements by ETAAS is not sufficient but Journal of Analytical Atomic Spectrometry also because the sensitivity may be decreased by the presence of iron.Therefore preconcentration methods have also been r e p ~ r t e d . ' > ~ * ~ ~ ~ ~ We have reported on the determination of trace amounts of selenium and tellurium in several metals and alloys by ETAAS after reductive coprecipitation with p a l l a d i ~ m . ~ ~ . ~ ~ This method is useful because trace amounts of selenium and tellurium can be separated simultaneously from large amounts of matrix metals and these elements can be sensitively determined by ETAAS when palladium is used as a chemical modifier. In order to develop a procedure for the highly sensitive determination of trace amounts of antimony germanium and tin in high-purity iron we investi- gated and employed ETAAS after reductive coprecipitation with palladium.EXPERIMENTAL Apparatus For the determination of antimony germanium and tin a 2-9000 simultaneous multi-element atomic absorption spec- trometer (Hitachi Tokyo Japan) was used with Zeeman-effect background correction. A hollow cathode lamp was used as the light source. The operating conditions are listed in Table 1. For dissolution and coprecipitation of the sample a PTFE beaker was used. For sealing the PTFE beaker a sealon film (Fuji Film Tokyo Japan) was employed. For filtration a membrane filter (Nuclepore polycarbonate pore size 0.2 pm Coster Cambridge MA USA) was used. Reagents Antimony standard solution (1.00 mg 1-I). A 0.100 g amount of metallic antimony (99.999%) was dissolved in 20ml of 7 moll-' HN03 together with 1 g of tartaric acid on a hot- plate and diluted to 100ml with water.The solution was diluted with tartaric acid and water before use. Germanium standard solution A (1.00 mg 1- '). A 0.100 g amount of metallic germanium (99.999%) was dissolved in 20 ml of 7 moll-' HNO together with 1 g of tartaric acid on a hot-plate and diluted to 100ml with water. The solution diluted with tartaric acid and water before use. Germanium standard solution B (1.00 mg 1-I). A 0.144 g amount of germanium(rv) oxide (GeO 99.999%) was dissolved in 2.5 moll-' NaOH solution on a hot-plate after which 20 ml of 14 moll-' HN03 were added and the solution was trans- ferred into a calibrated flask. The solution was then diluted to 100 ml with water.It was diluted with water before use. Tin standard solution (1.00 mg 1-l). A 0.100 g amount of metallic tin was dissolved in 20 ml of 6 mol I-' HCl on a hot- plate after which the solution was transferred into a calibrated flask and 40 ml of 12 mol I-' HCl and 1 g of tartaric acid were Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 (577-583) 577Table 1 Instrumental and operating conditions for ETAAS Instrument Hitachi 2-9000 :simultaneous multi-element atomic absorption spectrometer Sample injection Light source Background correction Carrier gas flow rate Interrupted gas flow rate Furnace Injection volume Element Wavelength Lamp current Temperature programme- Drying Ashing Atomization Cleaning Autosampler Hollow cathode lamp Polarized Zeeman effect 300 ml min-l 0 ml min-' Pyrolytic graphite coated graphite tube 10 pl Sb Ge 231.2 nm 8.0 mA 80-150°C 30 s 150-1200 "C 30 s 1200-1200 "C 30 s 25OO0C 5 s 3000"C 10 s 265.2 nm 10.0 mA 80-150 "C 30 s 150-1400 "C 30 s 1400-1400 "C 30 s 2600"C 5 s 3000"C 10 s Sn 224.6 nm 8.0 mA 80-150°C 30 s 150-1600 "C 30 s 1600-1600 "C 30 s 3000 "C 10 s 2700"C 5 s added.The solution was then diluted to 100ml with water. It was diluted with tartaric acid solution (5% m/v) before use. Palladium standard solution ( 3 g I-' solution was used for calibration 0.6 g 1-' solution was used for coprecipitation). A 0.300 g amount of metallic palladium was dissolved in 5 ml of 14moll-' HN03 and diluted to 100ml with water. The solution was used for calibration. A 20ml aliquot of the solution was diluted to 100ml with water and used for coprecipitation.Sodium ethylenediaminetetraacetate (EDTA-Nu) solution (0.4 moll-' = 15% m/v). A 75 g amount of EDTA-Na was dissolved in 20 ml of 15 moll-' aqueous NH3 on a hot-plate and diluted to 500ml with water. Sodium phosphinate (NaPH,O,) solution (1.14 mol I-' = 10% m/u). A 12g amount of NaPH2O2.H2O was dissolved in water and diluted to 10 ml with water. Sodium tetrahydroborate (NaBH,) solution (1.59 moll-' = 6% m/v). A 6 g amount of NaBH was dissolved in water and about 0.1 g of NaOH was added. The solution was then diluted to 100ml with water. EDTA-Na NaPH,O and NaBH solutions were kept in a polyethylene bottle and NaPH202 and NaBH solutions were kept at a temperature of less than 10°C. Distilled de-ionized water was used for the preparation of all standard and sample solutions.All the reagents used were of analytical-reagent grade. Samples The high-purity iron used was Grade 1 iron powder (Johnson Matthey Materials Technology Royston Hertfordshire UK). JSS 001-4 and 002-4 Pure Iron (Japanese Iron and Steel Certified Reference Materials The Japan Iron and Steel Federation Tokyo Japan) were used as reference samples. Procedure Preparation of sample A 1 g amount of the sample was weighed into a PTFE beaker then 10 ml of HN0,-HCl (1 + 1) and 10 ml of water were added and the sample was dissolved on a hot-plate. Subsequently 20 ml of H2S04-H,P04 (1 + 1) were added and the solution was heated to fumes. After cooling to room temperature the solution was diluted with 30 ml of water and 50 mi of EDTA-Na solution were added. Then 6.25 moll-' NaOH solution was added until the pH reached 4.0 whereupon the solution was boiled in order to produce the EDTA-Fe complex.After cooling to room temperature 1 g of ascorbic acid 578 Journal of Analytical Atomic Spectrometry August 1996 was added. For the determination of antimony 6.25 moll-' NaOH solution was added until the pH reached 7.0 after which 5 ml of 0.6 g 1-' palladium standard solution and 5 ml of NaPH solution were added. The solution was sealed with sealon film and left at room temperature for at least 3 h. For the determination of tin 6.25 mol I-' NaOH solution was added until the pH reached 10.0 after which 5 ml of 0.6 g 1-' palladium standard solution and 5 ml of NaBH solution were added.The solution was sealed with sealon film and left at room temperature for at least 4 h. For the determination of germanium both methods could be adopted. After standing the precipitate was collected on a membrane filter. Immediately thereafter the membrane filter with the collected precipitate was transferred into a glass beaker and the precipitate was dissolved in 1 ml of 0.66mol1-' (=lo% m/v) tartaric acid solution 1.5 ml of HNO and one drop (about 0.05 ml) of HC1 at room temperature. The solution was transferred into a calibrated flask and diluted to exactly 10 ml with water. The blank was prepared by the same procedure described above but without the sample. Determination by E TAAS A lop1 aliquot of the sample or blank solution was injected into the graphite furnace.The atomic absorbance was measured under the conditions shown in Table 1. Preparation of solutions for the calibration graph Volumes (0-15 ml) of the antimony germanium and/or tin standard solution were transferred into calibrated flasks. Then 5 ml of 3 g 1-' palladium solution 7.5 ml of HNO 5 ml of 0.66 moll-' tartaric acid solution and five drops (about 0.25 ml) of HC1 were added to each solution. Finally the solutions were diluted to exactly 50 ml with water. RESULTS AND DISCUSSION Effects of Iron and Chemical Modifier on the Determination of Antimony Germanium and Tin by ETAAS The effects of iron and chemical modifiers on the determination of antimony germanium and tin by ETAAS under the con- ditions mentioned in Table 1 were examined and the results are shown in Fig.1. The method employed was as follows solutions containing 0.1 mg I-' of antimony germanium and/or tin and lOOmgl-' (1000 times the concentration of these elements) of various chemical modifiers or 10 g 1-' (100000 times the concentration of these elements) of iron VOl. 115 I t n I I I I I 4 f !I3 !I U 0 Fig. 1 Effect of chemical modifier on the atomic absorbance of Sb Ge and Sn; Sb Ge and Sn 0.1 mg 1-l; Pd Mg and Cu 100 mg 1-'; Fe logl-'. Atomic absorbances were normalized to 1 when no modifier was present were prepared. Then the atomic absorbances of antimony germanium and/or tin were measured. Palladium palladium- magnesium and palladium-copper were used as chemical modifiers. For antimony germanium and tin the increase in atomic absorbance was highest in the presence of palladium only.Therefore of the chemical modifiers studied palladium appeared to be the most useful for the ETAAS determination of these elements. Also it was found that the absorbances of antimony germanium and tin were decreased in the presence of a 100 000-fold concentration of iron. In spite of the addition of palladium to the iron-containing solution the same results were obtained. Therefore for the determination of trace amounts of antimony germanium and tin in high-purity iron by ETAAS it is necessary to separate these elements from iron; in the method adopted palladium was used as a carrier for coprecipi t a t ion. Concentration of Palladium as Chemical Modifier The effect of the amount of palladium on the atomic absorbances of antimony germanium and tin was examined.The results obtained are shown in Fig. 2. The method employed was as follows solutions containing 0.1 mg I-' of antimony ger- manium and/or tin and palladium concentrations ranging from 0.06 to 0.66 g 1-' were prepared. Then the atomic absorbances of antimony germanium and/or tin were measured. For all three elements the highest atomic absorbance was obtained when the concentration of palladium was 0.3 g 1-'. Therefore the concentration of palladium adopted as a chemical modifier for the determination by ETAAS was 0.3 g 1-' i.e. the amount of palladium used as a carrier for coprecipitation was 3 mg. Ashing Temperature for ETAAS The effect of ashing temperature on the atomic absorbances of antimony germanium and tin was examined.The results o*20 I 0.15 8 0.10 $ < 0.05 0.00 I 0 400 800 1200 1600 2000 Ashing temperaturePC Fig. 3 Relationship between the ashing temperature and the atomic absorbance of Sb; Sb 0.1 mg I-'; Pd 300 mg 1-'; + Sb and Pd; -+- Sb only 0.16 0.12 1 0.08 51 .n 4 0.04 0.00 1 . I . I . I . I 400 800 1200 1600 2000 Ashing temperaturePC Fig. 4 Relationship between the ashing temperature and the atomic absorbance of Ge; Ge 0.1 mg 1-I; Pd 300 mg 1-'; + Ge and Pd; -+- Ge only [Pd] I g r' 0 400 800 1200 1600 2000 Ashing temperaturePC Fig. 2 Relationship between the palladium concentration and the absorbance of Sb Ge and Sn; Sb Sn and Ge 0.1 mg 1-'; + Sb; + Ge; and + Sn Fig. 5 Relationship between the ashing temperature and the atomic absorbance of Sn; Sn 0.1 mg 1-I; Pd 300 m g 1-I; - Sn and Pd; + Sn only Journal of Analytical Atomic Spectrometry August 1996 Vol.11 579obtained are shown in Figs. 3-5. The method employed was as follows solutions containing 0.1 mg I-' of antimony ger- manium and/or tin with and without 300 mg 1-' of palladium were prepared. The atomic absorbances of antimony ger- manium and/or tin were then measured in the ashing tempera- ture range from 200 to 2000°C. For the determination of antimony the highest absorbance was obtained at 1200 "C; the absorbance decreased above 1300 "C regardless of whether or not the solution contained palladium. Therefore the ashing temperature adopted for antimony was 1200 "C. For the deter- mination of germanium the absorbance was almost constant below 1600 "C; the highest absorbance was obtained at 1400 "C when palladium was not present.Therefore the ashing tem- perature adopted for germanium was 1400 "C. For the determi- nation of tin the highest absorbance was obtained at 1600°C; the absorbance decreased above 1700 "C when palladium was present. However the absorbance decreased above 1300 "C when palladium was not present. Therefore the ashing tem- perature adopted for tin was 1600°C. Atomization Temperature for ETAAS The effect of atomization temperature on the atomic absorbances of antimony germanium and tin was examined. The results obtained are shown in Fig. 6. The method employed was as follows solutions containing 0.1 mg I-' of antimony ger- manium and/or tin and 300 mg 1-' of palladium were prepared.The atomic absorbances of antimony germanium and/or tin were then measured in the atomization temperature range from 1700 to 3000°C. For the determination of antimony the absorbance was almost constant between 2000 and 3000"C and the highest absorbance was obtained at 2500 "C. Therefore the atomization temperature adopted for antimony was 2500 "C. For the determination of germanium the absorbance was almost constant between 2600 and 3000 "C. Therefore the atomization temperature adopted for germanium was 2600 "C. For the determination of tin the highest absorbance was obtained at 2700 "C. Therefore the atomization temperature adopted for tin was 2700 "C. Selection of Reductant for Coprecipitation In order to select a suitable reductant for the coprecipitation and separation of trace amounts of antimony germanium and tin from high-purity iron the effect of reductants was examined.The method employed was as follows high-purity iron (1 g) was weighed into a PTFE beaker and 1 ml of antimony germanium and/or tin standard solution was added. The Oa30 * mixture was dissolved in HNO and HCI then H2S04 and H,PO were added and the solution was heated to fumes. The analyte elements were separated and concentrated by coprecipitation with palladium using various reductants. The precipitate was dissolved and the solution diluted to 10m1 then the elements were determined by ETAAS. Ascorbic acid sodium hydrogensulfite-hydrazine sulfate (NaHS0,-hydrazine) NaPH202 NaBH and metallic zinc were examined as reductants and the results obtained are shown in Table 2.Ascorbic acid and NaHS0,-hydrazine have been used for coprecipitation of selenium and t e l l u r i ~ m . ~ ~ . ~ ~ NaPH 02 NaBH and metallic zinc are stronger reductants than ascorbic acid and NaHS0,-hydrazine. Antimony germanium and tin are not reduced as easily as selenium or tellurium because the redox potentials of Sb'" Ge" and Sn" are lower than those of SeIV and Te". Therefore it was decided to investigate stronger reductants than ascorbic acid or NaHS0,-hydrazine. When ascorbic acid or NaHS0,-hydrazine was used as the reductant antimony germanium and tin were not recovered. It was assumed that these elements could not be reduced to the metal by these reductants. When NaBH was used the same result was obtained.However it was assumed that antimony germanium and tin were vaporized as their hydrides because the reducing power of NaBH was too strong. On the other hand when NaPH,O was used hydrogen was generated on the surface of palladium formed by reduction. In solutions containing large amounts of iron these elements and palladium were not precipitated. It is believed that since the iron(1IF phosphinate complex is formed NaPH,O cannot act as a reductant. In order to overcome these problems a procedure was adopted in which the matrix iron was masked by EDTA-Na28.29 and antimony germanium and tin were coprecipitated with NaPH,O or NaBH at pH > 4.0. When NaPH,O was used both antimony and germanium could be quantitatively recovered at pH 7.0. When NaBH was used both germanium and tin could be quantitatively recovered at pH 10.0.When metallic zinc was used only tin was quantitatively recovered. Therefore for the determination of antimony NaPH,O2 is suitable as a reductant. For the deter- mination of tin NaBH is suitable. For the determination of germanium both NaPH20 and NaBH can be used. Effects of pH on the Coprecipitation The effect of pH on the recoveries of antimony germanium and tin by coprecipitation with palladium using NaPH20 or NaBH was examined. The results obtained are shown in Figs. 7 and 8. The method employed was as follows high- purity iron (1 g) was weighed into a PTFE beaker and 1 ml of antimony germanium and/or tin standard solution was added. The mixture was dissolved and heated to fumes using the same procedure as described above after which the analyte 1W ZOO0 2400 2800 3200 Atomizing temperaturePC Fig.6 Relationship between the atomization temperature and the atomic absorbance of Sb Ge and Sn; Sb Ge and Sn 0.1 mg 1-I; Pd 300mg1-'; + Sb; + Ge; and + Sn 580 Journal of Analvtical Atomic Svectrometrv. August 1986. Table2 Recoveries of Sb Ge and Sn by coprecipitation using several reductants Recovery (YO) Reductant Ascorbic acid* NaHS0,-hydrazine sulfate* NaPH2O2* NaPH202 (pH 4.0) NaPH,02 (pH 7.0) NaBH,* NaBH (pH 7.0) NaBH (pH 10.0) Zn metal* Sb NDt ND ND 3.45 101.0 ND 15.6 69.4 2.35 Ge ND ND ND 19.7 96.9 ND 51.0 97.5 90.2 Sn ND ND ND 19.2 41.1 ND 85.1 99.8 103.4 * No pre-reduction with ascorbic acid and no addition of EDTA-Na. t ND = Not detected. VOl. 11Fig.7 Relationship between the pH for coprecipitation using NaPH,O and the recoveries of Sb Ge and Sn; Fe 1 g; Sb Ge and Sn 1 pg; + Sb; It Ge; and + Sn p 6 4 4 0 20 0 Fig. 8 Relationship between the pH for coprecipitation using NaBH and the recoveries of Sb Ge and Sn; Fe 1 g; Sb Ge and Sn 1 pg; + Sb; + Ge; and -0- Sn elements were separated by coprecipitation with palladium. When NaPH,02 was used as the reductant the pH was adjusted to 4.0-8.0. On the other hand when NaBH was used the pH was adjusted to 7.0-12.0. The precipitate was dis- solved and the solution diluted to 10ml; finally the three elements were determined by ETAAS. When NaPH,O was used as the reductant antimony was quantitatively recovered at pH 6.5-7.0; germanium was quantitatively recovered at pH 7.0.However tin was not completely recovered in the pH range examined. Therefore for the determination of antimony or germanium NaPH,O was used as the reductant and the pH was adjusted to 7.0. On the other hand when NaBH was used germanium and tin were quantitatively recovered at pH 10.0. However antimony was not completely recovered in the pH range examined. Therefore for the determination of germanium or tin NaBH was used as the reductant and the pH was adjusted to 10.0. Amount of Reductant The effect of the amount of NaPH,O or NaBH used as reductant for coprecipitation on the recoveries of antimony germanium and tin was examined. The results obtained are shown in Fig. 9. The method employed was as follows high- purity iron (1 g) was weighed into a PTFE beaker and 1 ml of antimony germanium and/or tin standard solution was added.The mixture was dissolved and heated to fumes using the same procedure as described above after which 0.5-2.5 g of NaPH202 or 0.15-0.75 g of NaBH was added. The analyte 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Reductant mass / g Fig. 9 Relationship between the mass of reductant and the recoveries of Sb Ge and Sn; Fe 1 g; Pd 3 mg; Sb Ge and Sn 1 pg. With NaPH,O, + Sb; It Ge. With NaBH, + Ge; + Sn elements were separated by coprecipitation with palladium. The precipitate was dissolved and the solution diluted to 10 ml; finally the elements were determined by ETAAS. It was found that when 0.5-1.Og of NaPH202 was used both antimony and germanium were quantitatively recovered. On the other hand when 0.15-0.45 g of NaBH was used germanium and antimony were quantitatively recovered.It was assumed that when more than 1.0 g of NaPH202 and more than 0.45 g of NaBH was used all three elements were vaporized as their hydrides. Therefore it was decided to use 0.5 g of NaPH202 and 0.3 g of NaBH,. Standing Time for Coprecipitation The standing time for coprecipitation for the separation of antimony germanium and tin was examined. The results obtained are shown in Fig. 10. The method employed was as follows high-purity iron (1 g) was weighed into a PTFE beaker and 1 ml of antimony germanium and/or tin standard solution was added. The mixture was dissolved and heated to fumes using the same procedure as described above after which the analyte elements were separated by coprecipitation with palladium using standing times of 1-4 h.The precipitate was dissolved and the solution diluted to 10ml; finally the elements were determined by ETAAS. It was found that when the standing time was more than 3 h quantitative recoveries 1 2 3 4 5 Standing time / h Fig. 10 Relationship between the standing time for coprecipitation and recoveries of Sb Ge and Sn; Fe 1 g Sb Ge and Sn 1 pg. With NaPH,O, * Sb; -t Ge. With NaBH, + Ge; -+- Sn Journal of Analvtical Atomic Soectrometrv. AuPust 1996. Vil. 1 1 581of antimony and germanium were obtained whereas tin was quantitatively recovered after more than 4 h. Therefore when NaPH202 was used as the reductant the standing time employed was 3 h; when NaBH was used the standing time employed was 4 h. Acid Dissolution of Sample It has been reported that germanium is vaporized as its chloride during dissolution of samples in hydrochloric acid.,' The effect of hydrochloric acid on the recoveries of antimony germanium and tin was examined.The results obtained are shown in Fig. 11. The method employed was as follows high- purity iron (1 g) was weighed into a PTFE beaker and 1 ml of antimony germanium A (made from the metal) germanium B (made from the oxide) and/or tin standard solution was added. The mixture was dissolved in HCl-HNO solutions containing different proportions of HCl and heated to fumes by the same procedure as described above after which the analyte elements were separated by coprecipitation with p 8s cr! 8 0 - 7s - 0.5 1 .o 1.5 2.0 HCl HN03 Fig. 11 Relationship between the volume ratio of acid (HCI HNO,) used for dissolution and the recoveries of Sb Ge and Sn; Fe 1 g; Sb Ge and Sn 1 pg; .Sb; A Ge A (from metal); V Ge B (from oxide); and 0 Sn Table 3 Analytical results for mixed solution sample palladium. The precipitate was dissolved and the solution diluted to 10ml; finally the elements were determined by ETAAS. It was found that when the HCI:HNO ratio was greater than 1.5 the recovery of tin was also lower. It is assumed that not only germanium but also antimony and tin are vaporized as their chlorides during dissolution and heating to fumes. The recovery of germanium from solution B was less than from solution A when the HC1 HNO ratio was greater than 1. This was because the tartaric acid present in the germanium A standard solution reduced the vaporization of germanium markedly.Therefore a 1 + 1 HC1-HNO3 mixture was used for dissolution of samples. Calibration Graph For the preparation of calibration graphs an aqueous standard solution was used and the relationship between the atomic absorbance and the concentration of antimony germanium and/or tin was examined. A straight line passing through the origin was obtained for a concentration of <0.36 mg 1-1 of antimony c 0.40 mg 1-l of germanium and < 0.30 mg 1 - of tin. The limit of detection (three times the standard deviation of the blank n = 10) of the proposed method was 0.Ol9 pg g-' of antimony O.Ol0 pg g-' of germanium and 0.031 pg 8-l of tin in 1 g of sample. Analysis of Mixture Solution Sample Antimony germanium and tin in a mixture solution sample which consisted of 1 pg of each of these elements and 1 g of high-purity iron were determined according to the proposed method.The results obtained are shown in Table 3. All three elements were quantitatively determined. Analysis of Reference Samples The proposed method was applied to the determination of antimony germanium and tin in reference samples. The results obtained are shown in Table 4. In the determination of antimony and tin the analytical values agreed with the Added/ Reductant Pg Fe ( 1 g) Fe ( 1 g) NaBH Fe (1 g)+Sn 1.02 Fe (1 g) Fe (1 g)+Ge 1.00 Fe (1 €9 NaPH,O Fet (1 g)+Sb 1-04 - Fe (1 g)+Ge 1 .OO - - - Found/ 1.18 1.04 f 0.01 0.98 & 0.00 2.32 & 0.00 < 0.02 < 0.01 1.328 f 0.04 1.01 * 0.00 < 0.01 Recovery ("/.I n 100.3 6 4 98.1 6 4 98.0 5 4 100.9 5 4 - - - - RSD* 1.33 0.660 W O ) - - 0.392 3.29 2.18 - * Relative standard deviation.t High-purity iron; Johnson Matthey Grade 1 powder. Table 4 Analytical results for reference samples Sample Reduct ant Element Pure Iron NaPH,02 Sb (JSS 001-4) Ge NaBH Sn Ge Pure Iron NaPH 0 Sb (JSS 002-4) Ge NaBH Sn Ge Reference value*/ Analytical value/ P8 g-' pg g-' - 0.18 f 0.011 - 0.05~ 0.00 < 0.4 0.29 0.00 < 1 0.55 & 0.03 0.566 & 0.01 < 1 0.64 +_ 0.03 - 0.056 f 0.006 - - 0.59 0.02 RSDt 5.86 2.02 5.94 3.21 5.23 4.09 (%) 11.2 12.4 * Non-certified value. t Relative standard deviation. 582 Journal of Analytical Atomic Spectrometry August 1996; Vol. 1 1reference values. In the determination of germanium good agreement was found between the value obtained using NaPH,O and that obtained using NaBH,.CONCLUSION Trace amounts of antimony germanium and tin in high-purity iron were quantitatively separated and concentrated by reductive coprecipitation with palladium. The advantages of the proposed method are that when palladium is used as a chemical modifier the ETAAS determination of all three elements is highly sensitive. When using NaPH202 both antimony and germanium were separated simultaneously and when using NaBH both germanium and tin were separated simultaneously and quantitatively by a similar procedure from matrix iron. The limit of detection for these elements using the proposed method was about 10-100 times lower than with existing method^.^.^ Therefore the proposed method is useful for the determination of trace amounts of antimony germanium and tin in high-purity iron.The authors thank Professor Kichinosuke Hirokawa for his continuing guidance. REFERENCES Abiko K. Sci. Am. Jpn. Version 1993 23 20. Takada K. Inamoto I. and Okano T. Microchem. J. 1994 49 291. Takada K. Muter. Jpn. 1994 33 84. JIS G1235-1981 Methods for Determination of Antimony in Iron and Steel 1982. Hosoya M. Konno H. and Takeyama S. Bunseki Kagaku 1983 32 444. Danzaki Y. Bunseki Kagaku 1988 37 153. Sun Q. Wang H. T. and Mou S. F. J. Chromatgr. 1995,708,99. 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Adeloju S. B. O. and Pablo F. Anal. Chim. Acta 1992 270 143. Sun C. Q. Gao Q. A. Xi J. B. and Xu H. D. Anal. Chim.Acta 1995 309 89. Waller P. A. and Pickering W. F. Talanta 1995 42 197. Uggerud H. and Lund W. J. Anal. At. Spectrom. 1995,10,405. Lunzer F. Pereiro-Garcia R. Bodelgarcia N. and Sanz- Medel A. J. Anal. At. Spectrom. 1995 10 311. Feldrnann J. Grumping R. and Hirner A. V. Fresenius’ J. Anal. Chem. 1994 350,228. Venkaji K. Naidu P. P. and Rao T. J. P. Talanta 1994,41 1281. Liu X. Z. Xu S. K. and Fang Z. L. At. Spectrosc. 1994 15,229. Schulze G. and Lehmann C. Anal. Chim. Acta 1994 288 215. Burguera M. Burguera J. L. Rivasa C. Carrero P. Brunetto R. and Gallignani M. Anal. Chim. Acta 1995 308 143. Burns D. T. Chimpalee N. and Harriott M. Fresenius’ J. Anal. Chem. 1994 349 530. Rettberg T. M. and Beach L. M. J. Anal. At. Spectrom. 1989 4 427. Xuan W. Spectrochim. Acta Part B 1992 47 545. Sahayam A. C. and Gandadharan S. Can. J. Appl. Spectrosc. 1994 39 61. Ivanova E. Stoimenova M. and Gentcheva G. Fresenius’ J. Anal. Chem. 1994 348 317. Morishige Y. Hirokawa K. and Yasuda K. Fresenius’ J. Anal. Chem. 1994 350 410. Ira P. and Macfarlane A. M. J. Radioanal. Nucl. Chem. 1994 182 427. Sayama Y. Fukata T. and Kuno Y. Bunseki Kagaku 1995 44 569. Ashino T. Takada K. and Hirokawa K. Anal. Chim. Acta 1994 297 443. Ashino T. and Takada K. Anal. Chim. Acta 1995 312 157. Ishikuro M. Hosoya M. and Takada K. Bunseki Kagaku 1991 40 71. Itagaki T. Ishikuro M. and Takada K. Bunseki Kagaku 1994 43 569. Gridchina G. I. Zauod. Lab. 1968 34 1194. Paper 6/01 042C Received February 13 1996 Accepted May 17 1996 Journal of Analytical Atomic Spectrometry August 1996 Vol. 1 1 583

ISSN:0267-9477    

DOI:10.1039/JA9961100577

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Microwave-assisted Dilute Acid Extraction of Trace Metals From Biological Samples for Atomic Absorption Spectrometric Determination CHAO YAN ZHOU MING KEONG WONG* AND L I P LIN KOH Department of Chemistry National University of Singapore Lower Kent Ridge Road Singapore 11 9260 YEOW CHIN WEE Department of Botany National University of Singapore Lower Kent Ridge Road Singapore 119260 A microwave extraction method with dilute nitric acid for biological samples was developed using a pressure feedback microwave extraction system. The concentration of nitric acid pressure setting and extraction time were optimized by a mixed level orthogonal array design. The interactions among the selected parameters for the extraction were investigated. The effect of sample mass was also studied. Six replicate analyses of NIST SRM 1515 Apple Leaves were performed under the optimized extraction conditions (5 ml of 14% v/v nitric acid; pressure setting 1104 kPa; extraction time 30 min; sample mass 0.3 g).Eight elements oiz. Ca Cu Fe K Mg Mn Ni and Zn were determined by FAAS or ETA AS. Recoveries of 96-103% were achieved for all the elements. The RSDs of the test elements were less than 3.8% except for Ni. The extraction method was also employed to extract four other biological SRMs and CRMs. With the exception of Ni the recoveries of the eight test elements were 91-107% in all the reference materials. Keywords Microwave-assisted acid extraction; trace metal; biological sample ; atomic absorption spectrometry The successful dissolution of a biological sample is a crucial first stage in many spectrochemical and electrochemical analy- ses for trace metal constituents.There are two types of decomposition methods commonly employed wet decompo- sition and dry ashing.' Contamination and the loss of volatile elements are serious problems for the dry-ashing method. In wet decomposition various acid mixtures have been proposed for different sample matrices and different analytes. In most instances it is essential to employ an oxidizing agent to obtain complete decomposition. Both methods are tedious and time consuming but their careful execution is the backbone of a successful laboratory analysis.' Two recent developments in the sample preparation pro- cedure are the use of closed vessels to accelerate sample extraction or digestion and minimize contamination and losses of volatile elements and the use of microwave radiation to assist in extraction or dige~tion.~ With closed-vessel sample preparation systems various acid mixtures have been employed to destroy rapidly the organic matrix of the samples at elevated temperature or pre~sure.~-'~ The oxidizing power of acids is significantly increased with elevated temperature. The contami- nation loss of volatile elements and dissolution time are red~ced.~." However most of the established extraction or digestion methods employ concentrated acids.Extraction with dilute nitric acid is a simple and safe technique that has been applied with some success in the * To whom correspondence should be addressed. I Journal of I Analytical 1 Atomic 1 Spectrometry 1 determination of a range of elements in animal tissue CRMs at room temperat~re'~.'~ and of botanical CRMs with heat- ing.16 However generally it takes several hours or a day to obtain the digest.The problem of the dilute nitric acid extrac- tion can be overcome by the use of closed-vessel microwave extraction. The use of dilute nitric acid removes the danger of rapid pressure build-up as well as the possibility of systematic errors of using strong oxidizing acids. In this work orthogonal array design was used to evaluate the parameters of the microwave oven and acid concentration. The effect of sample mass was studied after other parameters had been optimized. Orthogonal array design is a chemometric approach combining the advantages of both the simplex method and factorial design.It provides an efficient and effective testing strategy.17-19 From the work a biological sample preparation method using a closed-vessel microwave extraction technique and dilute nitric acid was developed. The NIST SRM 1515 Apple Leaves was used to evaluate the selected parameters. EXPERIMENTAL Instrumentation All biological samples were extracted in an MDS-2000 Microwave Sample Preparation System (CEM Matthews NC USA). This system features a 630 W magnetron programmable in 1% power increments. An internal pressure control system allows for the control of pressure from 0 to 1380 kPa (200 psi) in five separate stages. A 12 vessel 360" revolving turntable operates via an optical sensor at 6 rev min-'. The extraction vessels have rupture membranes for safe operation under 1725 kPa (250 psi).The high concentration elements in the biological samples were determined with a Shimadzu (Tokyo Japan) flame atomic absorption spectrometer (AA-670) equipped with a gas control- ler and PR-4 graphic printer (FAAS). The low concentration elements were determined using a Perkin-Elmer (Norwalk CT USA) Model 4100ZL atomic absorption spectrometer equipped with an HGA-600 graphite furnace and Zeeman background correction (ETAAS). Specific instrumental para- meters for the various elements are given in Table 1. In all instances the determinations were carried out using a calibration graph. Chemicals The concentrated nitric acid used as well as all the other chemicals were of analytical-reagent grade (Merck Darmstadt Germany).All glass and poly(propy1ene) apparatus was washed with 5% v/v nitric acid and de-ionized distilled water. Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 (585-590) 585Table 1 Instrumental conditions for the determination of the elements FAAS- Gas flow rate/l min-' Element Acetylene Mn Fe Mg 1.9 2.0 1.6 ETAAS- Lamp Element current Wavelength c u 15 mA 324.8 nm Furnace programme for Cu- Furnace settings Dry 1 Temperature/"C 120 Ramp time/s 1 Hold time/s 20 Ar flow-rate/ml min-' 250 Air 8.0 8.0 8.0 Slit-width 0.7 nm Dry 2 140 5 40 250 Waveleng t h/nm 279.5 248.3 285.2 Aliquot Absorbance Tube volume Peak area THGA 20 pl Ash Atomize 1100 2050 10 0 20 5 250 0 Slit-width/nm 0.4 0.2 0.5 Modifier 0.03 mg Mg(NO3 )2 Clean 2400 1 2 250 The samples and acid were all diluted with de-ionized distilled water.Each stock standard solution of the test element containing 1000 mg 1-1 of the element in the nitrate form for AAS was obtained from BDH (now Merck Poole Dorset UK). All working standard solutions were prepared by immediate serial dilution with 0.5% v/v nitric acid solution. Microwave Extraction About 0.3g of NIST SRM 1515 Apple Leaves was weighed into a PTFE extraction vessel. After adding 5ml of dilute nitric acid the vessel was shaken for 3 min to allow for thorough mixing. The vessel was then capped and placed on the turntable. The microwave oven was operated according to the settings listed in Tables 2 and 3. Each trial had its own blank with the same amount of dilute nitric acid and microwave programme.After extraction the vessel was vented when the inner pressure was under 138 kPa (20 psi). The extract was filtered and then diluted to 15 ml with de-ionized distilled water. Assignment of Experiments For the extraction of biological samples using the closed-vessel microwave extraction procedure the most important variables are the extraction programme (power and time pressure and time) the power output the number of samples extracted the type and the concentration of acids used and the sample mass.'l In the pressure feedback microwave extraction pro- cedure the sample dissolution procedure was controlled by the pressure and power setting pr~gramme.~*~'.~' When a biological sample is extracted with strong oxidizing acids the power setting programme can be used to reduce the volatile reaction. Dilute nitric acid is not a strong oxidizing acid hence the maximum power output (100%) and rapid extraction programme were used.The number of samples extracted was four. Hence the following four variables were chosen in order to obtain the optimum trace element recoveries (i) the concen- tration of nitric acid (variable A); (ii) the pressure setting (variable B); (iii) the extraction time (variable C); and (iv) the mass of sample. The effect of sample mass was studied separ- ately as its interaction with the other variables. The other variables were evaluated by a mixed level orthogonal array design analysis. Orthogonal arrays are a classical design of experirnenk2' In forms orthogonal arrays are different from the way these arrays are usually displayed in the statistical literature.Orthogonal array design is in fact a saturated fractional factorial design. The precursor of orthogonal array design was Latin and Graceo-Latin squares,17 the theory of which was described in the 1940s by Ra022923 and B o ~ e . ~ ~ It was successfully applied in the engineering area for quality control by T a g ~ c h i . ~ ~ The theory and methodology of ortho- gonal array design as a chemometric method for optimization of analytical procedures have been discussed in previous p a p e r ~ . ' ~ - ~ ~ In orthogonal array designs orthogonal arrays are used to assign factors to a series of experiment combinations the results of which can then be analysed by using a common mathematical procedure.The main effects of the factors and preselected interactions are independently extracted. In an orthogonal array different combinations of numerals of any two columns have an equal appearance frequency. Here ortho- gonal means balanced. By arranging experiments orthogonally different effects can be separated. This method combines the advantages of both the simplex method and factorial design. Much more information can be obtained from a limited number of experirnent~.l~-'~ The method has been applied in liquid chromatography,26 gas ~hromatography~~ and micro- wave digestion of sediment samp1es.l' In this work the concen- tration of nitric acid (variable A ) was the most important variable considered. It was evaluated in a four-level design. The pressure setting (variable B) and extraction time Table 2 Assignment of factors and level settings for evaluating dilute nitric acid extraction of biological samples in the OA,,(4I x 212) matrix* Column No.Level 1 2 3 4 5 15 7 8 9 10 11 12 13 A B ( A x B) (A+B)z ( A X B)3 4 ( A X C) ( A X C) ( A x C)3 B x C 1 2% 1104 30 2 14% 552 15 3 10% 4 6 Yo * A Concentration of dilute nitric acid (v/v); B pressure setting (kPa); C extraction time (min). 586 Journal of Analytical Atomic Spectrometry August 19!36 Vol. 11Table 3 OA1,(4' x 212) matrix and experimental results ~- Column No. Trial 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 K1 K2 K3 K4 1 1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4 242.8 382.2 364.8 291.6 2 1 1 2 2 1 1 2 2 1 1 2 2 1 1 2 2 649.1 632.3 3 1 1 2 2 1 1 2 2 2 2 1 1 2 2 1 1 646.5 634.9 4 1 1 2 2 2 2 1 1 1 1 2 2 2 2 1 1 649.9 631.5 5 1 1 2 2 2 2 1 1 2 2 1 1 1 1 2 2 651.7 629.7 6 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 661.9 619.5 7 1 2 1 2 1 2 1 2 2 1 2 1 2 1 2 1 643.3 638.1 8 1 2 1 2 2 1 2 1 1 2 1 2 2 1 2 1 648.7 632.7 9 1 2 1 2 2 1 2 1 2 1 2 1 1 2 1 2 654.5 626.9 10 1 2 2 1 1 2 2 1 1 2 2 1 1 2 2 1 646.4 635.0 11 1 2 2 1 1 2 2 1 2 1 1 2 2 1 1 2 638.0 643.4 12 1 2 2 1 2 1 1 2 1 2 2 1 2 1 1 2 641.0 640.4 13 1 2 2 1 2 1 1 2 2 1 1 2 1 2 2 1 643.8 637.6 CO-R 71.5 58.5 61.3 51.5 95.0 94.6 96.6 96.0 93.6 89.0 92.0 90.2 77.9 69.0 74.0 70.7 (variable C) were evaluated in a two-level design.Because one four-level variable and two two-level variables were to be optimized and their interactions to be considered an 0A16(4' x 212) matrix was employed to assign variables and their interactions.Following a randomization process for vari- ables the level settings of the variables studied are listed in Table 2. The randomization process for level setting was important in reducing the errors from personal inclination on level assignment. The interactions were presented as A x B A x C and B x C . The positions of the variables chosen and level settings used in the mixed level orthogonal array design are displayed in Table 3. The column assignment was from Table 2. Column trials show the 16 experiments to be carried out. The variable settings are represented by 1 2 3 and 4. The composition response (Co-R) is listed in the last column. The Krn (rn = 1-4) values in the last four rows are the sum of Co- R with m level of variable assigned in the column.Krn is used to evaluate the effect of the variable. The analysis of variance (ANOVA) technique was employed to estimate the variable effects. For evaluating the effects of the variables at different levels from 16 experimental trials the recoveries of Mn Fe Mg and Cu in extracts were used as response functions. Their concen- trations in the extracts were determined by FAAS or ETAAS. For each test element the output response of one test element (Relement) is defined as Relement = accuracy of test element (%) element concentration from experiment x 100% certified concentration given by SRM 1515 The Co-R of the four elements was also used to judge the quality of the microwave extraction. It was calculated as - RESULTS AND DISCUSSION Considering the interactions of the selected variables three of them were evaluated by a mixed level orthogonal array design analysis OAl6(4l x 212).The results of Relement of Mn Fe Mg and Cu in the 16 experimental trials are listed in Table 4. The results of Co-R are listed in Table 3. The significance of different variables was quantitatively evaluated by the ANOVA tech- nique for Co-R. The ANOVA analysis results and F-test are Table4 Recoveries of the four elements used for evaluating the extraction efficiency in the 16 OA16(4' x 2") trials Trial 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Mn 88.0 85.6 83.7 82.0 95.0 96.3 93.9 88.7 91.9 90.7 92.6 91.1 85.6 83.3 89.4 90.2 Fe 87.4 45.2 51.9 26.9 92.2 88.6 93.6 90.2 80.4 76.1 83.1 71.2 64.7 73.0 67.0 100 Mg 93.3 90.0 87.1 86.6 96.5 93.3 99.0 96.1 92.4 100 100 95.3 97.7 91.2 95.8 88.9 c u 17.3 13.2 22.3 10.4 96.1 100 100 100 99.2 84.8 99.1 91.1 62.1 36.7 37.7 36.8 shown in Table 5.The sum of the squares are the statistical results from the Co-R for different variables or interactions. The equations of the sum of the squares and other items in the ANOVA table for testing the magnitude of the different variables in orthogonal array design were given in a previous paper.17 From the analysis it is seen that the concentration of nitric acid (A) and the extraction time (C) are statistically significant at p < 0.01. The pressure setting (B) and the inter- actions of A x B and A x C are significant at p < 0.05. The interaction of B x C is significant at p < 0.25. The effect of sample mass was studied later after the other variables had been optimized.Effect of the Concentration of Nitric Acid Most methods for the microwave extraction or digestion of biological samples have employed concentrated nitric acid or acid mixtures with strong oxidizing On the other hand to avoid the violent reaction of acids with the organic samples in closed vessels it is necessary to use the time and power or time and pressure pr~gramming.',~~ The violent reaction would be reduced if dilute nitric acid was employed because of its low oxidizing power. One of the purposes of this work was to identify the concentration of dilute nitric acid that could decompose the organic matrix within a short time Journal of Analytical Atomic Spectrometry August 1996 Vol.1 1 587Table 5 Results of ANOVA analysis of Co-R in the OA16(4' x 212) matrix 100 - 2 sf. 0 5 0 5 0 - E! E Source A B C A x B A x C B x C Error h M n - Fe - cu - Mg Sum of squares 3 160.4475 17.6400 112.3600 58.8200 65.3000 8.1225 4.247 5 - CO-R Degrees of freedom 3 1 1 3 3 1 3 Mean square F 1057.4825 774.09* 17.6400 12.467 112.3600 79.365 19.9400 14.089 21.7667 15.379 8.1225 5.747 1.4158 Best level 2 1 1 A B and A B1 A,C and A,C B1 c * p = 0.01 (99% confidence level) F(3,3) = 29.46. t p = 0.05 (95% confidence level) F(3.1) = 9.28. 5 p = 0.01 (99?'0 confidence level) F(3.1) = 34.12. 9 p = 0.05 (95% confidence level) F(3,3) = 10.17. 7 p = 0.25 (75% confidence level) F(3.1) = 2.02. and keep the violent reaction to a minimum. Four different concentrations of nitric acid were evaluated in the OA16(4' x 212) matrix. The ANOVA results of Co-R indicated that the concentration of nitric acid strongly influenced the sample decomposition.Fig. 1 shows the relationship between the output response of four test elements (Relement) and Co-R and the concentrations of nitric acid. Dilute nitric acid at 2 and 6% v/v did not completely destroy the organic sample (some solid substances were still present in the extract) and gave relatively low recoveries of the test elements. Relatively good Relement values were obtained with concentrations of nitric acid of 10 and 14% v/v. Considering the complicated matrix of the organic samples 14% v/v nitric acid was selected. Effect of Pressure Pressure is a very important parameter in pressure feedback microwave extraction.It is used to control the decomposition reaction in a closed ve~sel.'~ Fig 2 shows the pressure curves of trials 5 and 9. There was no overshooting of the pressure control setting point. The decomposition reaction was not violent and the pressure control was efficient in controlling the reaction. The effect of pressure setting was found to be statistically significant at p < 0.05 from ANOVA results of Co-R. The best element recoveries were obtained at a pressure setting of 1104 kPa (160 psi). On the other hand the effect of the interaction between the concentration of nitric acid (A) and the pressure setting (B) was also important. Table 6 shows the interaction of variables A and B which was obtained from the sum of Co-R values containing the different level settings of variables A and B in 16 trials.The high pressure setting was useful in improving the recoveries of test elements when samples were extracted with nitric acid at very low concen- tration. When the concentrations of nitric acid were at 10 and 14% v/v the effect of high pressure settings was insignificant. Fig. 1 Effect of the concentration of nitric acid on sample extraction 1200 s *O0 2 v) 400 * Trial9 0 = I I 1 0 10 20 30 Time/min Fig. 2 nitric acid Relation between pressure and extraction time using 14% (v/v) Table 6 Four-by-two table of the interaction between the nitric acid concentration (variable A) and the pressure setting (variable B) Concentration of nitric acid level Pressure level 1 2 3 4 1 130.0 189.6 182.6 146.9 2 112.8 192.6 182.2 144.7 Considering the fact that some organic samples would be more difficult to decompose than SRM 15 15 Apple Leaves a pressure setting of 1104 kPa (160 psi) was also used in subsequent sample extractions. Effect of Extraction Time For most element determinations the sample dissolution step is time consuming.In the normal sample extraction by the dilute nitric acid method it takes several hours to a day to obtain the homogenate ~ o l u t i o n . ~ ~ * ~ ~ One of the advantages of closed-vessel microwave extraction is the reduction of the extraction time. In this work the extraction time was evaluated at two levels of 30 and 15 min respectively. From the results of ANOVA analysis of Co-R the effect of the extraction time was statistically significant at p < 0.01.The best recoveries of the test elements were obtained with an extraction time of 30 min. The interactions of A x C and B x C were also signifi- cant at p < 0.05 and 0.25 respectively. Table 7 shows that the use of a 30min extraction improved the element recoveries when very low concentrations of nitric acid were employed. When 14% v/v nitric acid was used there was no difference in the element recoveries between the two levels. The best combi- 588 Journal of Analytical Atomic Spectrometry August 1996 Vol. 11Table 7 Four-by-two table of the interaction between the concen- tration of nitric acid (variable A ) and the digestion time (variable C ) Concentration of nitric acid level Time level 1 2 3 4 1 132.8 191.6 185.6 151.9 2 110.0 190.6 179.2 139.7 nation of variables B and C was B C1 (Table 8).Tables 7 and 8 were obtained in a similar way to Table 6. An extraction time of 30 min was used in subsequent experiments. Effect of Sample Mass According to the mixed level orthogonal array design analysis OA,,(4' x 212) the concentration of dilute nitric acid the pressure setting and the extraction time were optimized. Under the optimized conditions different sample masses were extracted in several groups. The recoveries of the four test elements are listed in Table 9. Using a mass of 0.4 g or less resulted in complete extraction by 5 ml of 14% v/v nitric acid as shown by the good recoveries obtained. When the sample mass was over 0.5 g the recoveries were reduced. Also some solid substances were present in the extract solutions.A sample mass of 0.3 g is thus recommended. Table 8 Two-by-two table of the interaction between the pressure setting (variable B) and the extraction time (variable C) Extraction time level Pressure level 1 2 1 332.6 324.0 2 306.2 308.4 Table 9 Recoveries (YO) of the four test elements using different masses of NIST SRM 1515 Apple Leaves (n = 3) Mass of sample/g Mn Fe Mg c u 0.2 9 7 f 2 101f3 102f3 99 * 2 0.3 98 f 5 9 6 f 4 9 9 f 4 102f3 0.4 9 6 f 5 100f4 101f6 96 f 5 0.5 98 f 5 8 6 f 9 9 2 f 4 8 1 f 7 0.6 8 9 f 6 83+9 9 2 k 6 63 f 9 Application Six replicate analyses of NIST SRM 1515 Apple Leaves were carried out under the optimized conditions. Ca Cu K Fe Mg Mn Ni and Zn were determined by FAAS or ETAAS. The results are listed in Table 10.Recoveries of 96-103% were obtained. With the exception of Ni the RSDs of the test elements were less than 3.8%. The optimized parameters were applied to the extraction of different biological samples namely NIST SRM 1557b Bovine Liver and CRMs No. 1 Pepperbush No. 3 Chlorella and No. 6 Mussel from the National Institute for Environmental Study (NIES) of Japan. The results obtained are summarized in Table 10. With the exception of Ni recover- ies of 91-107% were obtained for the eight test elements. CONCLUSION The microwave extraction technique can be used to improve the dilute nitric acid extraction of biological samples. By using a rapid extraction programme viz. 100% power output 1104 kPa (160 psi) pressure and a 30 min extraction time biological samples of 0.3 g were completely extracted with 5 ml of 14% v/v nitric acid.This extraction method is also suitable for other biological samples. C.Y.Z. thanks the National University of Singapore for a research scholarship. REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 Van Loon J. C. Selected Methods of Trace Metal Analysis Biological and Environmental Samples Wiley-Interscience New York 1985 pp. 77-141. Hoke F. Scientist 1992 6( 14) 19. Matusiewicz H. Sturgeon R. E. and Berman S . S. J. Anal. At. Spectrom. 1991 6 283. Bettinelli M. Baroni U. and Pastorelli N. Anal. Chim. Acta 1989 225 159. de la Fuente M. A. and Juhrez M. Analyst 1995 120 107. Yang Q. Penninckx W. and Smeyers-Verbeke J. J. Agric. Food Chem. 1994 42 1948. Krushevska A. Barnes R. M. and Amarasiriwaradena C.Analyst 1993 118 1175. Lajunen L. H. J. Piispanen J. and Saari E. At. Spectrosc. 1992 13 127. Baldwin S. Deaker M. and Maher W. Analyst 1994 119 1701. Tahan J. E. Granadillo V. A. Sanchez J. M. Cubillan H. S. and Romero R. A. J. Anal. At. Spectrom. 1993 8 1005. Kingston H. M. and Jassie L. B. Anal. Chem. 1986 58 2534. Mohd A. A. Dean J. R. and Tomlinson W. R. Analyst 1992 117 1743. Temminghoff E. J. M. and Novozamsky I. Analyst 1992,117,23. Table 10 Analytical results (pg g-') for NIST SRM and NIES CRM biological samples (n = 4) Element Ca Found Cu Found Fe Found K Found Mg Found Mn Found Ni Found Zn Found Certified Certified Certified Certified Certified Certified Certified Certified NIST SRM* 1515 15 100 & 200 15 260 f 150 5.5 f 0.1 5.6 f 0.2 8 0 k 3 8 3 f 5 16 600 f 300 16 100 f 200 2680 f 60 2710 & 80 5 2 k 2 54rfr 3 0.9 k 0.2 0.91 f 0.12 12.7 -t 0.3 12.5 f 0.3 NIST SRM 1577b 107 f 3 116f4 170 f 4 160 f 8 192 f 5 184 f 15 10 200 f 200 9940 f 20 600 f 40 601 f 28 11.0 f 0.8 10.5 f 1.7 - 118f4 127 f 16 NIES CRM No.1 14 200 f 500 13 800 f 700 12.0 k 0.7 1 2 k 1 219 f 4 205 f 17 14 800 f 100 15 100 & 600 4030 & 100 4080 f 200 2000 f 60 2030 f 170 8.5 f 0.5 8.7 k 0.6 343 f 11 340 f 20 NEIS CRM No. 3 5010 f 170 4900 f 300 3.6 0.2 3.5 f 0.3 1680 50 1860 f 100 11 800 f 100 12 400 _+ 60 3380 f 90 3300 k 200 6 3 f 2 69 & 5 0.6 f 0.2 20.4 f 0.3 20.5 f 1.0 - NIES CRM No. 6 1190 f 70 1300 f 100 4.7 & 0.2 4.9 f 0.3 158 f 7 158 f 8 5160 f 80 5400 f 200 2020 f 30 2100 f 100 16.0 k 0.8 16.3 rfr 1.2 1.1 & 0.1 0.93 f 0.06 9 9 k 2 106 f 6 * Six replicate analyses. Journal of Analytical Atomic Spectrometry August 1996 Vol. 1 1 58914 15 16 17 18 19 20 21 22 Asp T. N. and Lund W. Talanta 1992 39 563. Niazi S. B. Littlejohn D. and Halls D. J. Analyst 1993 118,821. Chow P. Y. T. Chua T. H. Tang K. F. and Ow B. Y. Analyst 1995 120 1221. Lan W. G. Wong M. K. Chen N. and Sin Y. M. Analyst 1994 119 1659. Lan W. G. Wong M. K. Chen N. and Sin Y. M. Analyst 1994 119 1669. Zhou C. Y. Wong M. K. Koh L. L. and Wee Y. C. Anal. Chim. Acta 1995 314 121. Hasty E. T. Littau S. E. and Revesz R. paper presented at the Pittsburgh Conference and Exhibition 1991. Kacker R. N. Lagergren E. S. and Filliben J. J. J. Res. Natl. Inst. Stand. Technol. 1991 96 577. Rao C. R. Bull. Calcutta Math. SOC. 1946 38 67. 23 Rao C. R. Proc. Edinburgh Math. SOC. 1947 8 119. 24 Bose R. C. Sankhya 1947 8 107. 25 Taguchi G. System of Experimental Design Kraus International Publishers New York 1987 vols. 1 and 2. 26 Wan H. B. Lan W. G. Wong M. K. and Mok C. Y. Anal. Chim. Acta 1994 289 371. 27 Oles P. J. J . Assoc. Off Anal. Chem. 1990 73 724. 28 Lamble K. and Hill S . J. Analyst 1995 120 413. 29 de Bore J. L. M. and Maessen F. J. M. J. Spectrochim. Acta Part B 1983 38 739. Paper 61001 49A Received January 8 1996 Accepted May 15 1996 590 Journal of Analytical Atomic Spectrometry August 1996 Vol. 11

ISSN:0267-9477    

DOI:10.1039/JA9961100585

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Simultaneous Determination of Cobalt and Manganese in Urine by Electrothermal Atomic Absorption Spectrometry. Method Development Using a Simplex Optimization Approach I Journal of I Analytical 1 Atomic 1 Spectrometry 1 BENT SCHACK IVERSEN ANTONIA PANAYI JUAN PABLO CAMBLOR AND ENRICO SABBIONI Commission of the European Communities Joint Research Centre Via E. Fermi 21 020 Ispra (VA) Italy A rapid method for the simultaneous determination of Co and Mn in urine was developed using a multielemental Perkin- Elmer SIMAA 6000 graphite furnace atomic absorption spectrophotometer. A simplex optimization was used to establish the pyrolysis and atomization temperature and the ramp and hold time for the pyrolysis step. The sample preparation was a simple 1 + 1 dilution using a diluent of 2% v/v HN03 and 0.05% m/v Triton X-100 in MilliQ-water.The LODs were 0.18 pg 1-' for Co and 0.09 pg 1-l for Mn. The recovery studies to check for bias have shown acceptable accuracy for the procedure (104% for Co and 100% for Mn with 15 and 13% RSD respectively; 2 pg 1-' Co added to 10 different samples at 0.46-1.91 pg l-l 2 pg 1-' Mn added to 10 different samples at 0.11-1.09 pg 1-l). Keywords Simplex optimization; simultaneous multielement analysis; cobalt; manganese; urine; electrothermal atomic absorption spectrometry Cobalt is an essential element for mammals being an integral component of vitamin BIZ but it is also recognized as a toxic metal in workplace exposure. In particular occupational expo- sure to cobalt in the metals industry diamond polishing and the porcelain chemical and pharmaceutical industries is a potential health risk to humans causing or exacerbating diseases such as lung asthma and fibrosis which are probably all immunotoxic reactions.',2 Cobalt dust has also been impli- cated as an etiological agent in contact dermatitis in metal workers.Consideration of these sensitization reactions leads to the question of exposure to low doses of Co as a possible environmental problem. Manganese is also recognized as an essential and neurotoxic element with known exposures from Fe alloys dry cells oxidizers and organomanganese com- p o u n d ~ . ~ Abnormally high amounts of Mn in the brain of Parkinsonian subjects suggests a possible connection between the element and this di~ease.~ Biomonitoring of Co and Mn is therefore strongly recommended to minimize health risks arising from environmental and occupational exposure to these elements.As part of the EURO-TERVIHT project (Trace Element Reference Values in Human Tissues) which aims to establish and compare trace metal reference values from inhabi- tants of the European Union,' background levels of trace elements including Co and Mn in biological fluids and tissues from non-exposed individuals are being determined to establish baseline values for biomonitoring strategies. ETAAS has for many years been the method of choice for biomonitoring of trace elements. It is a reliable and easy to handle method with low costs per sample. A significant draw- back is the single-elemental character which can be time consuming if more than one element has to be measured.With the introduction of new multi-elemental instrumentation the analytical potential of ETAAS has increased. The aim of this paper is to improve the simultaneous determination of Co and Mn in urine. In order to optimize the ETAAS analytical system the modified simplex method as described by Aberg and Gustavsson6 was used. Simplex methods have previously been used to optimize analytical parameters for chromatographic procedures. Optimization is fast and reliable if there is an existing knowledge of the significant factors of the chemical measurement process. The simplex optimization is characterized by changing one or more factors for a new experiment as a result of the response of the previous experiments. This is different from the classical and more time consuming experimental approach where one factor is changed at a time while holding the other factors constant.MATERIALS AND METHODS Equipment An AA spectrometer SIMAA 6000 from Perkin-Elmer Norwalk USA was used for simultaneous multielemental analysis. Zeeman-effect background correction employed a built-in 0.8 T magnetic field oriented longitudinally to the optical path. Pyrolytically coated graphite furnaces with plat- form were used. The samples were injected from an AS70 autosampler with an 80-position tray and with a micro- dispenser selectable in increments of 0.1 pl. The wavelengths were Co 242.5 nm at 30 mA; Mn 279.5 nm at 20 mA. The injected volume was 20 p l ( l 0 p1 of urine and 10 pl of diluent). The procedure was controlled by the AA Winlab software version 1.1 (Perkin-Elmer Norwalk CT USA).The lamps used were hollow cathode lamps (HCL) from Perkin-Elmer. Reagents The Co and Mn 1.000 g I-' reference solutions were certified AA standards from Fisher Scientific Company New Jersey USA. Throughout the procedure Suprapure double sub-boiling distilled HNO (Romil Loughborough UK) was used. The Mg(N0,)2 -6H,O chemical modifier and Triton X-100 were from Merck Darmstadt Germany. The Pd chemical modifier was an Atomic Absorption Standard solution from Aldrich Milwaukee WI USA. All dilutions were made using Milli-Q ultrapure water (Milli-Pore Molsheim France). Pipette tips (Eppendorf Hamburg Germany) and sampler cups ( Perkin-Elmer) were made of polypropylene. Procedure Samples stored at - 20 "C in screw-capped polypropylene tubes then being allowed to defrost and mixed well were diluted 1+1 with a diluent of 2% HNO and 0.05% Triton X-100 in MilliQ-water; 20 pl of the solution were injected into the furnace.From the reference solutions containing 1 g 1-' of Co and 1 g 1-1 of Mn 100 pl volumes were pipetted into a calibrated flask and Milli-Q water was added to 50ml. From this stock solution (2 mg 1-' Co; 2 mg 1-l Mn) 100 pl were pipetted into a calibrated flask and Milli-Q water was added to 50ml. The resulting solution was used to spike a urine sample to give matrix matched calibration standards with the concentrations of Co and Mn shown in Table 1. Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 (591 -594) 591Table 1 Calibration scheme for Co and Mn in urine t B Calibration point Addition of Co/pg I-’ Addition of Mn/pg 1-1 1 0 0 2 1.20 1.20 3 2.40 2.40 4 3.60 3.60 Simplex Optimization System A simplex is a geometric model with n+ 1 vertices in a space with n dimensions.The n dimensions correspond to the n factors that need to be optimized. The first step is to construct the initial simplex choosing n + 1 combinations of the n factors. For the ETAAS system used four recognized important factors were optimized by the procedure pyrolysis temperature; ramp time to reach this temperature; hold time at this temperature; and atomization temperature. All other steps in the furnace programme were chosen in agreement with existing procedures employed in this laboratory for ETAAS determinations of urine samples.The constraints or the boundaries for each factor and for the response of the system were decided a priori. To check if an acceptable optimum for the system had been reached two criteria were used firstly as a convergence criterion the standard deviation of the responses of all vertices of the simplex should be less than a pre-set value; secondly characteristic masses for Co and Mn should be acceptably low indicating a sufficiently sensitive optimum for the system. To establish the initial simplex five vertices were calculated according to Table2;7 vertex no. 1 has coordinates of zero and indicates the starting levels chosen by the analyst. The vertices 2-5 are then constructed by multiplying the chosen step size for each factor by the values in the table and the results are added to the value of vertex 1.The vertices are illustrated for a two-dimensional system in Fig. 1 with the initial simplex (W-NW-B) in solid lines and the possible new simplexes in dashed lines. The new coordinates are calculated by using the following equations n+l C K - V w (1) T/ - i = l C - n where i is the number of vertices n is the number of factors and V 6 Vw and V are the coordinates for the vertex of the centroid vertex number worst vertex and new vertex respect- ively. V is calculated and then substituted into eqn. (2). New vertices V = AVc + Vw (2) The value of A is 2 for the normal reflection 3 for the expanded reflection 1.5 for positive contracted and 0.5 for negative contracted vertices. The responses from the ETAAS system are examined and the best vertex (B) the worst (W) and the next-to-worst (NW) are established.The worst vertex W is reflected through the centroid to find R (reflected vertex) and further experiments N W b Factor 1 Fig. 1 Possible new vertices for a two factor system using the modified simplex method. Original simplex in solid lines. C = Centroid B =best vertex W = worst vertex NW = next to worst vertex R = reflected vertex E =expanded vertex PC =positive contracted vertex and NC =negative contracted vertex are undertaken to evaluate the response from this new vertex illustrated in Fig. 2. RESULTS AND DISCUSSION The initial simplex was constructed using the a priori con- sidered best levels and combination of the factors as vertex 1 and vertices 2-5 were calculated according to Table 2.The step sizes were chosen in order to cause a new response sufficiently different from the previous also taking into account experimental error. The ramp and the hold time were approxi- mated to the nearest integer and the temperature settings were approximated to the nearest 10 degrees the smallest increment allowed by the instrument. Table 3 shows the constraints and step sizes and Table 4 the vertices of the initial simplex. In order to obtain a large signal for the optimization procedure the urine was spiked with 4 pg I-’ Co and 4 pg 1-l Mn and measured in duplicate for each combination of factors (vertex). In order to take into account the different sensitivity of the ETAAS system for Co and Mn the over-all response evaluated was a modified summation of the absorbances from each element Corrected sum = 2.7 x integrated absorbance for Co +integrated absorbance for Mn The factor 2.7 is the ratio of recommended characteristic mass for Co to the recommended characteristic mass for Mn under single element conditions with the characteristic mass defined as the mass of analyte (normally expressed in pg) which causes an integrated absorbance of 0.0044.When evaluating the differences between the responses from the five different vertices of the simplex decisions to accept or Table 3 Constraints and step sizes for the system to be optimized Factor Constraints Step sizes Pyrolysis temperature/”C 1100-1600 200 Ramp time/s 1-30 20 Hold time/s 1-40 20 Atomization temperature/”C 1700-2600 200 Table 2 Values used to establish the initial simplex Table 4 Vertices for the initial simplex Factors Factors Vertex Pyrolysis Atomization no.temperature Ramp time Hold time temperature 1 0 0 0 0 2 1 .000 0 0 0 3 0.500 0.866 0 0 4 0.500 0.289 0.817 0 5 0.500 0.289 0.204 0.791 Vertex Pyrolysis Atomization no. temperature Ramp time Hold time temperature 1 1300 5 10 2400 2 1500 5 10 2400 3 1 400 22 10 2400 4 1400 11 26 2400 5 1400 11 14 2560 592 Journal of Analvtical Atomic Svectrometrv. AuPust 1996 voz. 11B,NW,Win I Fig. 2 Flow chart of the decision process of the simplex optimization method ‘>’ indicates better than Y is the direction to take if ‘Yes’ is the answer to the comparison and N if the answer is ‘No’. XZY indicates that X substitutes Y (1) the control is not passed and a new optimization experiment is carried out.(2) the control procedures are fulfilled and the optimization is finished. Abbreviations as in Fig. 1 reject the responses were based on the differences of the responses without testing whether they were statistically sig- nificant. As a final criterion for the evaluation of the responses the shapes of the absorbance curves had to be acceptable from experimental experience and the RSD less than 10% for each element. If not the vertex was considered to be the worst. A problem in optimizing Co and Mn together is a difference of 500°C in atomization temperatures according to the rec- ommended settings for the instrument by the manufacturer. Because of this problem the criteria for accepting the optimiz- ation must be chosen carefully.The two criteria for the control procedure for an eventual acceptable optimum were the RSD of the responses from the five vertices of the simplex should be lower than 5% and the calculated characteristic masses for system settings according to the best vertex should not exceed the recommended charac- teristic masses for single element conditions by more than 20%. The characteristic masses were established by relating the differences in integrated absorbance with the differences in mass of the elements from the same urine sample spiked with 2 ppb Co and 2 ppb Mn and with 4 ppb of each element. An optimum which was acceptable according to the two control criteria was reached as vertex number 17 see Table 5.In total 13 experiments were carried out and four vertices were ignored because one or more factors were outside their constraints. The two criteria for an acceptable optimum were fulfilled because the RSD for the corrected sums of the final five vertices was 3.0% and the characteristic masses calculated for the optimum were for Co 20.9 pg; and for Mn 4.1 pg. For Mn the result is better than the value published by the manufacturer and for Co the value published by the manu- facturer is less than 20% better than the value determined for this system. It is important to stress that the responses should be independent of time. In an ETAAS optimization procedure the graphite tube will change during use therefore the firings should be limited in number so that the tube can be considered unchanged throughout the optimization.The optimization was not improved by adding 5pg of Pd-15 pg of Mg(NO,) as chemical modifier to each sample as addition the modifier increased the risk of sample contami- nation. A control of 20 calibration curves for Co and Mn was made by the multivariate procedure suggested by Mestek et a1.* The Mahalanobis distances are calculated according to the Journal procedure and drawn in the control charts as shown in Fig. 3. The control limits in the charts correspond for ICL (Inner Control Limit) to the distance from the base line to T0.95452 from the Hoteling’s T2 distribution and for OCL (Outer Control Limit) to the distance G,99732. The limits are compar- able to the well known limits +2s and +3s from Shewart control charts.Fig. 3 demonstrates that all 20 calibration curves for Co and Mn show good stability without disturbances. Recovery Study Ten different urine samples with and without addition of analyte were measured with 2 repetitions. As shown in Table 6 the results were satisfactory. The RSDs were quite high but taking the low concentrations into consideration this is acceptable. Certified Reference Materials The analytical usefulness of the proposed method could not be assessed by analysis of standard reference materials because they do not exist for Co and Mn in urine. Instead the Seronorm Trace Elements in Urine with a recommended content of 10.2 pg 1-’ (173 nmol 1-’) of Co was used. It was 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Calibration curves Fig.3 in urine Control charts of 20 calibration curves for (a) Co and (b) M n of Analytical Atomic Spectrometry August 1996 Vol. 1 1 593Table 5 The vertices and their responses (corrmm) and the action to be taken for the optimization procedure Vertex 1 2 3 4 5 6 77 8 9 107 11 1271 13 14 157 16 177 I' Pyrolysis 1300 1500 1400 1400 1400 1550 1360 1430 1420 1290 1190 1230 1090 1240 1360 1220 1270 /"C Atomiza- Ramp/s Hold/s tion/"C 5 10 2400 5 10 2400 22 10 2400 11 26 2400 11 14 2560 20 20 2480 9 13 2420 12 5 2470 22 11 2530 31 12 2590 27 14 2560 38 12 2630 29 13 2400 16 13 2520 15 16 2620 17 14 2560 13 -3 2490 Corr.* 0.05 18 0.0576 0.0379 0.0544 0.0127 0.0623 3 0.0470 0.0637 9 0.0644 6 5 6 $ 0.0661 0.0673 Action? R NC R PC R E R E R NC R PC * Corrected sum (see text).t Actions taken for the new vertex R = reflected E = expanded PC = positive contracted NC = negative con- tracted. $ Absorbance peaks not accepted. $ Factors outside the con- straints. 1 Last five vertices. 11 Optimum. Table 6 Recovery* of 2.0 pg 1-' Co and 2.0 pg 1-' Mn added to urine samples ~ ~~~ ~ Range/g 1-' Recovery RSD (%) (%) Wth addition Wthout addition co 0.46-1.91 2.60-3.9 5 104 15 Mn 0.11-1.09 1.76-2.28 100 13 * n= 10. diluted 4 times (2.55 pg 1-l) to be within the calibration range and the value found was 2.72 pg 1-1 (10.88 pg 1-l for the undiluted sample) with an s of 0.16 pg 1-l (n= lo) so the accuracy is acceptable. Comparison With Other Analytical Techniques Radiochemical neutron activation analysis RNAA is recog- nized as an outstanding analytical technique for many elements at trace levels because it has multielement capability good selectivity and accuracy with low LODs. However due primar- ily to the difficulty of obtaining access to irradiation and special radiochemical facilities RNAA as a routine method has lost ground to other powerful technique^.^.'^ Its role as a reference technique is still very important in the analysis of biological samples particularly in developing spectrochemical techniques such as ETAAS and ICP-MS.The method for Mn was compared with an independent analytical technique. Twelve spot urine samples were collected from subjects living in Ispra (Varese Italy) and submitted to RNAA which included the isolation of induced 56Mn and its counting by computer-based high resolution y-ray spec- trometry.A linear regression procedure with no special assumptions regarding the distribution of the samples and measurement errors was used for the comparison of the two methods." The statistical procedure of Passing and Bablok'l gave the following regression line for the relationship between RNAA ( X ) and ETAAS (Y) Y = -0.24+1.40X with 95% confidence interval [ - 0.80 0.221 for the intercept and [0.91:2.00] for the slope. According to the theory of the procedure the hypothesis of an intercept of 0 is accepted because the confidence interval contains 0 and the hypothesis of a slope of 1 is also accepted because the confidence interval contains 1. Table 7 The furnace programme for the Co-Mn method Temperature/ Vertex "C 1 110 2 130 3 800 4 1270 5 2560 6 2600 Gas flow/ Ramp/s Hold/s ml min-' Read 1 20 250 5 30 250 5 25 250 17 14 250 1 5 0 X 1 5 250 Analytical Performance The calibration curves showed linearity up to 6 pg 1-l for both elements.No attempt was made to check if the analytical ranges could be extended. The LOD (3s n=20) was 0.18 pg 1-l for Co and 0.09 pg l-' for Mn. Owing to its significant influence on the very low levels of Co and Mn great care must be taken in establishing the blank values for autozero. Conditions Selected for the Final Optimized Method The optimized temperature programme is shown in Table 7. CONCLUSION The SIMAA 6000 instrument can improve the determination of Co and Mn in urine at the ultra-trace level. The advantages of this instrument such as simultaneous multielement determi- nation and fast and reliable automatic handling of the samples are fully used in the Co-Mn method.The practical LOD are sufficiently low (Co 0.18 pg 1-I; Mn 0.09 pg 1-') for the determination of the metals from non-occupationally exposed subjects. The procedure is fast allowing more than 100 determi- nations per day and requires only a minimum of sample preparation thereby reducing the risk of contamination. In any case the use of ultra-pure acids and ultra-clean room conditions is still essential for accurate determination of Co and Mn at these levels. The optimization described employed four important factors pyrolysis temperature ramp time hold time and atomization temperature. However it cannot be excluded that further optimization including e.g.the sample volume injected or the concentration of HN03 or Triton X-100 could lead to further improvement. The study also confirms that the simplex method could be used successfully in atomic absorption optimization. REFERENCES 1 2 3 4 5 6 7 8 9 10 11 Lauwerys R. and Lison D. Sci. Tot. Environ. 1994 150 1. Christensen J. M. Sci. Tot. Environ. 1995 166 89. Sekt P. K. and Chandra S . V. in Metal Neurotoxicity eds. Bondy S. C. and Prassad K. N. CRC Press Inc. Boca Raton Florida 1988 pp. 19-33. Mena I. in Disorders of Mineral Metabolism eds. Bronner F. and Coburn J. W. Academic Press New York 1981 vol. I Sabbioni E. Minoia C. Pietra R. Fortaner S. Gallorini M. 4nd Saltelli A. Sci. Tot. Environ. 1992 120 39. Aberg E. R. and Gustavsson A. G. T. Anal. Chim. Acta 1982 144 39. Long D. E. Anal. Chim. Acta 1969 46 193. Mestek O. Pavlik J. and Suchanek M. Fresenius J. Anal. Chem. 1994 350 344. Pietra R. Sabbioni E. Gallorini M. and Orvini E. J. Radioanal. Nucl. Chem. Articles 1986 102 69. De Goeij J. J. M. and Woittiez J. R. W. J. Radioanal. Nucl. Chem. Articles 1993 168 429. Passing H. and Bablok W. J. Clin. Chem. Clin. Biochern. 1983 21 709. Paper 610095 7C Received February 9 1996 Accepted June 4 1996 pp. 233-270. 594 Journal of Analytical Atomic Spectrometry August 1996 Vol. 11

ISSN:0267-9477    

DOI:10.1039/JA9961100591

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Determination of Lead and Cadmium in Environmental Samples Optimized by Simplex Optimized Atomic Absorption Methods I Journal of 1 Analytical Atomic Spectrometry CONSTANTINE D. STALIKAS AND GEORGE A. PILIDIS European Environmental Research Institute Dodonis 42 45221 Ioannina Greece MILTIADES I. KARAYANNIS Laboratory of Analytical Chemistry University of loannina 451 10 loannina Greece Simplex optimized methods for the determination of cadmium and lead in environmental samples by use of electrothermal atomic absorption spectroscopy (ETAAS) are presented. The optimum experimental conditions were attained by applying the composite modified simplex optimization method after considering the following variables ramp to ashing temperature ashing temperature atomization ramping time atomization temperature and modifier concentration.The way of introducing the matrix modifier in the graphite furnace is one of the most important among the numerous factors that must be considered for the development of an ETAAS method. The matrix modifier was injected into a graphite tube provided with a L'vov platform prior to the injection of the sample leaving the modifier wet. The analytical investigations showed that the optimized systems are ideally suited to the analysis of cadmium and lead in samples of environmental origin offering good accuracy and precision. Keywords Lead; cadmium; environmental samples; simplex optimization; atomic absorption spectrometry Cadmium is a toxic element present at low concentrations in nature.' Cadmium occurrence stems from anthropogenic sources such as mining operations waste incineration and combustion of coal and oil while it occurs naturally in the environment as a result of volcanic emissions.2 The contami- nation from Cd has increased rapidly in recent years; it is commonly found in aquatic and terrestrial environments and is characterized by a long environmental per~istence.~?~ The monitoring of Cd in marine species (seafood) can serve as an indicator of variations in marine pollution.' Lead in the atmosphere comes mainly from the combustion of gasoline that contains tetraethyllead as an anti-knock agent.The accurate determination of lead in foodstuffs is important since the prolonged intake of even low concentrations of lead can cause serious toxic effect^.^.^ The low concentration levels of cadmium and lead during their assay in most of the samples unavoidably require a preconcentration step or a preliminary separation of cadmium in the bulk matrix.Most analytical procedures require a solution of the analyte and in principle solid samples must be dissolved prior to quantification. A number of methods of varying complexity have been used to determine trace amounts of Cd and Pb spectrophotometry atomic absorption spectrometry (AAS) inductively coupled plasma atomic emission spectrometry or mass spectrometry and electroanalytical methods. Among these the fastest and most sensitive that have conquered the field of analytical chemistry are electrothermal atomic absorption spectrometry (ETAAS) and inductively coupled plasma atomic emission spectrometry or mass spectrometry.The two techniques differ from each other in relative simplicity cost of analysis and interferences en~ountered.~.~ ETAAS combines high sensitivity simplicity low detection limits and a low sample volume requirement. All of these features have been responsible for its widespread use for a diverse range of samples. Since the performance of ETAAS is strongly affected by several instrumental parameters it is advisable to carry out optimization prior to the adoption of any procedure. The most convenient method of sample introduction into the graphite furnace is the injection of liquid samples. Thus solid samples require mineralization and destruction of the organic matrix before the implementation of any analytical technique.Volatilization losses can occur during wet oxidation if the temperature is allowed to exceed 250°C.10 The open- beaker wet ashing technique appears to be gaining popularity since it does not require complex and expensive instrumen- tation. Microwave digestion has received considerable atten- tion as an alternative to conventional wet Bomb digestion has been found to be preferable for the multi-element analysis of samples because problems of losses of volatile analytes have been observed using both microwave or hot plate dige~tion.'~ The direct introduction of solid samples or slurries is increasing in popularity because it combines the simultaneous destruction of the sample matrix and the atomiz- ation or excitation of the analyte but suffers from poor precision although the accuracy is good e n ~ u g h .' ~ - ' ~ Chemical modifiers of different types have been used but none of them seems to be suitable for general application. The combination of chemical modifiers and the L'vov platform in the interior of graphite tubes has largely eliminated inter- ferences on Cd and Pb signal in various matrices showing improved precision and ac~uracy.'~ In this paper the composite modified simplex (CMS) method is employed for the determination of Cd and Pb in seafood. ,Unlike the classical optimization procedures the CMS takes into account possible interactions among the parameters involved while minimizing the number of the required experi- ments accessing the optimum parameter value^.'^.'^ To the best of our knowledge no attempt has been made to date to optimize by use of the CMS method the analysis of these two important naturally occurring metals by ETAAS.However the method has been applied for the optimization of several analytical systems.20-2z EXPERIMENTAL Apparatus An Atomic Absorption Spectrometer Varian (Mulgrave Victoria Australia) SpectrAA-300 was used for this work linked with the Varian GTA-96 graphite furnace atomizer and Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 (595-599) 595the automatic sample dispenser. The results were recorded with an Epson (Nagano Japan) LX-400 printer. The hollow cathode lamps (SpectrAA) in conjunction with the deuterium background corrector were used for the lead and cadmium determinations. A lamp current of 4 mA at 228.8 nm was employed for the analysis of cadmium while the 283.3 nm lead resonance line was selected and used with a lamp current of 5 mA for the analysis of lead.The slit width in both cases was 0.5 nm. The atomization was performed with a pyrolytic graphite tube with a L'vov platform. The peak height was used for quantitative analysis. Argon was used as the inert gas for purging the graphite tubes. In obtaining the data a 15 pl volume of sample solution plus a 5 pl volume of modifier at appropriate concentration were used throughout. Reagents Used and Sample Treatment All solutions were prepared with particle-free deionized doubly distilled water. Merck standard solutions for AA were used to prepare the working solutions in 5% HNO,. Ammonium dihydrogenphosphate was selected as the appro- priate matrix modifier based on previous investigation^.^^ Analytical reagent grade crystals of the modifier were dissolved in water to give several solutions which were subsequently used as the chemical modifier.A mussel tissue RM 278 with certified Cd and Pb contents of 0.34k0.02 pg g-' and 1.91 f 0.04 pg g- ' respectively was used as the reference mate- rial (RM) for the validation of the methods. Marine Sediment (NRC-PACS l) Estuarine Sediment (CEC-CRM 277) and Tea (NRC-CRM CS5-05) were selected for the implementation of the developed methods. The certified reference materials were treated as follows OSOg dry mass was slowly digested by heating in a Teflon digestion vessel with 6ml of concentrated HNO,. The sedi- ments were then further subjected to HF-HC104 attack until the entire sample had been dissolved.The digestates were quantitatively transferred and diluted to a final volume of 20ml with 10% HNO,. Appropriate dilutions with the blank (10% HNO,) were made by the programmable sample dis- penser during the analysis. Optimization Procedure The simplex process was carried out using software which can run on any IBM compatible computer. The optimization was carried out on a sample of mussel tissue in order to establish the 'real' experimental conditions and to eliminate problems arising during the analytical procedure. This reference material has a matrix similar to most of the edible environmental samples and is thus suitable for the validation of the method. The effect of five experimental parameters on the analyses of the metals was investigated ramp to ashing temperature ashing tempera- ture atomization ramping time atomization temperature and modifier concentration were optimized by the CMS optimiz- ation method.Some preliminary univariate experiments were carried out prior to the CMS optimization in order to establish the boundaries of the values of each parameter. The lower and the higher values as well as the precision required for each variable were set accordingly as is shown in Table 1. The initial set of conditions was given and the program generated the rest of the experiments. For the five variable system six sets of experiments constitute the initial simplex. After the performance of each experiment the actual values of the variables and the peak height of the signal were entered.The first six experiments represent the first cycle of the simplex. The software continued the process of accessing the optimum by performing reflections (R) expansions (E) con- tractions (C) and lagrange interpolative fits (L). Each set of experiments was performed in three replicates. Between each experiment a blank corrective experiment was run to ensure stable and repeatable results. RESULTS AND DISCUSSION The optimum values of the parameters were deemed to have been reached after almost 30 experiments for Cd and Pb. Tables 2 and 3 summarize the optimization experiments carried out and the corresponding responses. Because of the volatility of many Cd and Pb compounds lower ashing temperatures are needed in comparison with those applied for the determination of most of the other metals. Higher temperatures result in losses of Cd and Pb during the charring step while lower temperatures retain matrix interferences Under the optimum ashing temperatures the ashing ramping times when restricted to low values do not appear to affect significantly the response on the analysis of Cd and Pb.High ashing ramping times in connection with high ashing tempera- tures decrease the absorbances due to the considerable loss of the analytes. The optimum atomization temperatures allow complete atomization of the analytes in the RM. By stopping the gas flow at this point the sensitivity increases while the lifetime of the graphite tube is not shortened. The atomization ramping time is a very critical parameter affecting the sensitivity of the system.It is advisable to heat the furnace rapidly to a preselected optimum temperature in order to achieve a sufficient temperature difference between tube and platform. Longer atomization ramping times during the atomization step led to a decrease in the peak height absorbance due to losses of the volatile analytes and to vapour phase interferences (physical interferences) when analysing samples with complex matrices. In the case of Pb the shorter ramping time was used whereas up to 2 s this variable does not seem to affect the analysis of Cd strongly. Optimum atomization temperatures have been established to ensure complete atomization and prolonged lifetime of the graphite furnace. The modifier concentration is critical when determining Cd and Pb in environmental samples although in these experi- ments it is not very pronounced.In both cases the concen- tration of matrix modifier was kept at 3.5 mg ml-I since the ammonium dihydrogenphosphate is a very cheap reagent thus keeping the cost of the analyses low. Moreover the analytes Table 1 of Pb and Cd Range and precision of experimental variables for the composite modified simplex (CMS) optimization of ETAAS for the determination Reverse boundary - - Variable Pb Cd Ashing ramping time/s 5 5 Ashing temperature/oC 300 300 Atomization ramping time/s 0.5 0.5 Atomization temperaturePC 1900 1600 Modifier concentration/mg ml- ' 0.5 0.5 Forward boundary Precision Pb Cd Pb or Cd 1000 800 50 2600 2400 100 50 50 5 8 8 0.5 7 7 0.5 596 Journal of Analytical Atomic Spectrometry August 1996 Vol.11Table 2 Simplex optimization of variables affecting the ETAAS response on the determination of Cd 0.16. Experiment no.* 1 2 3 4 5 6 7R 8E 9R 1 OE 11R 12R 13R 14C 15R 16R 17R 18C 19L 20R 21c 22R 23R 24C 25R 26C 27L 28R 29C 30R (a) Ashing ramping time/s 25 50 40 40 40 40 25 20 20 15 15 15 30 20 5 30 40 15 20 15 20 10 10 20 30 15 15 25 15 15 Ashing temperature/ "C 600 600 800 700 700 700 600 550 5 50 500 3 50 400 750 450 400 600 650 450 500 400 550 400 350 550 600 450 450 550 450 400 Atomization ramping time/s 5.0 5.0 5.0 8.0 6.0 6.0 2.5 0.91 1.5 0.61 1 .o 0.6$ 3.5 1.5 0.61 3.5 4.0 1.5 2.0 0.62 3.5 1.5 1 .o 2.0 2.0 1.5 2.0 2.5 1.5 2.5 Atomization temperature/ "C 2000 2000 2000 2000 2400 2100 2100 2200 1700 1600 1700 1600 1900 1800 1600 2000 2100 1700 1800 1600 1900 1600 1600 1900 2200 1700 1700 1900 1700 1800 Modifier concentration/ mg ml-' 1.5 1.5 1.5 1.5 1.5 4.5 2.0 2.5 2.5 3.5 3.5 0.5 0.5 2.5 2.5 4.0 2.5 2.5 2.5 3.5 2.0 3.5 3.5 2.5 1.5 3.0 3.0 2.5 3.0 3.5 Peak height? (arbitrary units) 0.160 0.143 0.122 0.11 1 0.1 14 0.137 0.175 0.169 0.180 0.172 0.141 0.154 0.127 0.184 0.157 0.180 0.138 0.185 0.171 0.160 0.165 0.177 0.171 0.173 0.1 72 0.190 0.193 0.170 0.193 0.195 * R Reflection; E Expansion; C Contraction.7 Average of three runs. $ The actual variable values used. The lowest possible value of ramping time under the instrumental specifications and the defined atomization temperature. are stabilized and the potential interfering constituents of the samples are presumably converted into more volatile com- pounds which are removed during the ashing process (see steps 3-5 in Table 4).The optimum furnace operating conditions for the analysis of both metals are given in Table4. The cooling down step was included in the furnace heating programme to ensure that the graphite had reached an adequately low temperature before the next injection. The recorded responses during the progress of the simplex as a function of each set of experiments are presented in Fig. 1. The method of introduction of the matrix modifier into the graphite furnace has proved to also affect the efficiency of the analysis of Cd and Pb in the environmental samples.24 There are two ways of introducing the matrix modifier into the graphite furnace when using a programmable sample dispenser in conjunction with ETAAS firstly simultaneous injection of the modifier and the sample (normal injection); and secondly injection of the modifier and optional drying prior to the injection of the sample (pre-injection). Pre-injection of the modifier can be performed by either proper drying of the modifier (ashing temperature) prior to the injection of the sample or wet deposition of the modifier at low temperature followed by injection of the sample. After the construction of the calibration curves by applying the specified parameters the accuracy the precision and the characteristic mass (defined as the mass of the analyte which yields a signal equal to 0.0044 absorbance) corresponding to the sensitivity of each injection method were evaluated and are shown in Table 5.All injection methods exhibited more or less similar and fairly good analyt- ical characteristics. More specifically the results indicate that pre-injection of the modifier by the wet deposition method proved to be slightly superior to the other two methods in AA 0*14t A A A ; I 8 0.12 e A Pb 0 0 5 10 15 20 25 30 35 I v a g 0.20 $ 0.18 0*16 t A A A A A A A O*12t A A A Cd 0.10; 5 10 15 20 25 30 Experiment no. Fig. 1 the determination of Pb and Cd Progress of the simplex during the optimization procedure for terms of the criteria which have been set for the evaluation. This could be attributed to the better mixing between sample and modifier and therefore to the higher efficiency of the reaction for stabilizing the analyte.Journal of Analytical Atomic Spectrometry August 1996 Vol. 1 1 597Table 3 Simplex optimization of variables affecting the ETAAS responst on the determination of Pb Experiment no.* 1 2 3 4 5 6 7R 8E 9R 1 OE 11R 12R 13R 14C 15R 16R 17C 18L 19R 20c 21L 22R 23R 24R 25C 26R 27C 28R 29R 30C 31R Ashing ramping time/s 10 35 25 25 25 25 25 25 25 25 40 30 20 30 10 20 25 25 40 20 20 20 15 15 20 15 20 10 5 20 10 Ashing temperature/ "C 600 600 950 700 700 700 400 300 3 50 300 3 50 300 300 500 450 5 50 3 50 400 400 450 450 300 300 400 500 500 450 450 550 400 450 Atomizatiomn ramping time/s 3.0 3.0 3.0 6.5 4.0 4.0 5.0 6.0 2.0 0.91 3.0 1.5 2.5 3.0 2.5 4.0 2.0 2.5 3.0 2.5 2.5 1.5 2.0 0.91 0.91 1 .o 1 .o 0.91 0.9$ 1.5 1 .O$ Atomization temperature/ "C 2000 2000 2000 2000 2300 2100 2200 2300 2300 2500 2500 2200 2600 2100 2000 2200 2200 2200 2500 2100 2200 2300 2300 2300 2300 2300 2300 2400 2400 2300 2500 Modifier concentration/ mg ml-' 1 .o 1 .o 1 .o 1 .o 1 .o 3.5 3.5 5.0 4.5 6.0 6.0 7.0 7.0 3.0 2.5 0.5 5.5 5.0 6.0 3.5 3.5 4.0 4.5 4.5 4.5 5.0 4.5 4.0 3.5 4.5 5.0 Peak height? (arbitrary units) 0.130 0.141 0.102 0.103 0.130 0.153 0.156 0.152 0.164 0.164 0.149 0.151 0.145 0.155 0.152 0.136 0.156 0.162 0.148 0.159 0.165 0.160 0.157 0.171 0.168 0.161 0.168 0.172 0.163 0.167 0.171 * R Reflection; E Expansion; C Contraction.? Average of three runs. $ The actual variable values used. The lowest possible value of rampirig time under the instrumental specifications and the defined atomization temperature. Table 4 Optimized instrumental parameters for the determination of Pb and Cd* Step 1 Drying 2 Drying 3 Ashing 4 Ashing 5 Ashing 6 Atomization 7 Atomization 8 Clean-out 9 Cooling down TemperaturerC Pb Cd - 150 200 500 500 500 2300 2300 2600 45 95 120 400 400 400 1800 1800 2400 45 'Time/s Pb Cd 25 35 10 10 10 15 1 .o 1 .o 2.0 2.0 0.9 2.0 2.0 2.0 2.0 2.0 12.8 11.8 Gas flow rate/] min-' Pb or Cd 3 3 3 3 0 0 0 3 3 Read command Pb or Cd No No No No No Yes Yes No No ~~ ~ ~~ * Injected volume 20 p1 (sample +matrix modifier).Table 5 The statistical results after the implementation of several injection techniques for the determination of Pb and Cd in the reference material Determined value/pg g- Accuracy 11%) Precision (RSD YO) Characteristic mass/pg* Injection method Pb Cd Pb Cd Pb Cd Pb Cd Normal injection 1.84 0.35 96 103 3.9 3.4 4.9 0.3 Pre-injection wet deposition 1.88 0.34 98 100 3.2 2.4 4.3 0.3 Pre-injection dry deposition 1.84 0.32 96 94 4.9 3.8 4.9 0.3 * Characteristic mass the mass of the analyte which yields a signal equal to 0.0044 absorbance.The conventional analytical procedures using working matrix effect was also investigated by performing recovery experiments on the RM. The results showed a satisfactory recovery for both metals ranging from 95 to 101%. The proposed techniques were also used to determine Cd and Pb content in certain certified environmental reference materials. The results are presented in Table 6 and indicate that the curves with standard solutions prepared in 5% HN03 were applied t o determine the concentrations of the analytes in the mussel tissue RM.Under the optimum experimental con- ditions these concentrations were found to be 0.34 pg g-' and 1.88 pg g-' respectively. The accuracy of dependence on the 598 Journal of Analytical Atomic Spectrometry August 199.6 Vol. 11Table 6 Results for the recovery of Pb and Cd from several environmental Certified Reference Materials applying the optimized techniques Type Certified value Found & s* Certified value Found & s* Marine Sediment (NRC-PACS 1) 404 & 20 392+ 17 2.38 k0.20 2.30 & 0.10 Estuarine Sediment (CEC-CRM 277) 146+ 3 152+5 11.9k0.4 1 1.4 & 0.5 Tea (NRC-CRM CS5-05) 1.06 _+ 0.10 1.01 f0.09 0.032 0.005 0.03 3 k 0.004 * Average of three runs &standard deviation. proposed optimized techniques are successfully applicable to a variety of environmental samples.CONCLUSION Taking into account the great importance and interest of Cd and Pb in nature as environmental pollutants we have pre- sented in this paper an optimized ETAA system for the analysis of these metals in environmental samples. The CMS optimiz- ation procedure has successfully been applied to improve the determination of Cd and Pb in samples with complex matrices. Even if the settings of the optimum experimental conditions cannot be exactly employed in other types of AAS the results are very promising regarding sensitivity accuracy and pre- cision. Since most of the ETAA systems can be automated the proposed techniques offer increased flexibility in developing procedures for the assay of various metals in real samples even if their matrix is unknown.REFERENCES 1 Aylett B. J. ‘Cadmium’ in Comprehensive Inorganic Chemistry ed. Trotman-Dickenson A. F. Pergamon Press Oxford 1973 p 254. 2 Fassett D. W. in Metals in the Environment ed. Waldrom H. A. Academic Press New York 1980 pp 61-1 10. 3 Chaney L. R. Beyer W. N. Gifford C. H. and Sileo L. Trace Subst. Environ. Health 1989 22 263. 4 Robards K. and Worsfold P. Analyst 1991 116 549. 5 Alikhan M. A. Bagatto G. and Zia S. Water Res. 1990,24,1069. 6 Nriagu J. O. Environ. Pollut. 1988 50 139. 7 tobinski R. and Adams F. C. ‘Lead and Organolead Compounds’ in Analysis of Contaminants in Edible Aquatic 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Resources ed. Kiceniuk J. W. and Ray S. VCH New York Fernandez F. J. and Giddings R. At. Spectrosc. 1982 3 61. Olesik J. W. Anal. Chem. 1991 63 1. Salisbury C. D. and Chan W. J. Assoc. Of. Anal. Chem. 1985 68 218. McCarthy H. T. and Ellis P. C. J. Assoc. Off. Anal. Chem. 1991 74 566. Cabrera C. Lorenzo M. L. Gallego C. Lopez M. C. and Lillo E. J . Agric. Food Chem. 1992 40 1631. Okamoto K. and Fuwa K. Anal. Chem. 1984,56 1758. Brown A. A. Lee M. Kullemer G. and Rosopulo A. Fresenius’ 2. Anal. Chem. 1987 328 354. Stephen S. C. Ottaway J. M. and Littlejohn D. Fresenius’ 2. Anal. Chem. 1987 328 346. L‘vov B. V. Specrochim. Acta Part B 1978 33B 153. May T. W. and Brumbaugh W. G. Anal. Chem. 1982,54 1032. Betteridge D. Wade A. P. and Howard A. G. Talanta 1985 32 709. Betteridge D. Wade A. P. and Howard A. G. Talantu 1985 32 723. Betteridge D. Taylor A. F. and Wade A. P. Anal. Proc. 1984 21 373. Stalikas C. D. Karayannis M. I. and Tzouwara- Karayanni S. M. Talanta 1994 41 1561. Pergantis S. A. Cullen W. R. and Wade A. P. Talanta 1994 41 205. Analytical Methods for Graphite Tube Atomizers ed. Rothery E. Varian Techtron Publication 1988. Johnson D. Vurian Analytical Instruments at Work AA-101 November 1990. 1994 pp. 115-157. Paper 6/01 4798 Received March 1 1996 Accepted May 8 1996 Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 599

ISSN:0267-9477    

DOI:10.1039/JA9961100595

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Effect of Ascorbic Acid and Sucrose on Electrothermal Atomic Absorption Signals of Indium Journal of Analytical Atomic Spectrometry SHOJI IMAI Department of Chemistry Faculty of Integrated Arts and Sciences University of Tokushima Tojushima 770 Japan NORIYUKI HASEGAWA YASUKO NISHIYAMA YASUHISA HAYASHI Department of Chemistry Joetsu University of Education Joetsu Niigata 943 Japan KENGO SAITO Nissei Sangyo SI Center Sakamachi Shinjuku Tokyo 160 Japan A significant enhancement of sensitivity with the formation of an unresolved double peak signal was observed for indium when deposited in a pyrolytic graphite (PG) furnace pyrolysed with an organic matrix solution and when deposited in a bare PG furnace as a matrix-added solution such as ascorbic acid and sucrose. As the pyrolysis temperature increased the integrated absorbance increased owing to an increase in absorbance at the second peak.The temperature at the second peak remained nearly constant. Integrated across the maximum the integrated absorbance decreased at the decrease in absorbance at the second peak which was shifted to low temperature. The treated PG furnace gave better thermal stability for the effect of pyrolysis temperature for In atomization whereas the organic matrix additive did not. The double peaks were analysed using Arrhenius plots in the treated PG furnace with a low pyrolysis temperature. The same atomizing species smaller sized droplets of In(l) on either active sites or in the thermally stable amorphous carbon were attributed to the species in the rate-determining step for the first and second signals respectively in the treated PG furnace with a low pyrolysis temperature.With the matrix additive the two atomizing species In,O(g) and smaller sized droplets of In(1) on active sites were attributed to the first and second signals respectively. When the pyrolysis temperature increased the atomizing species in the matrix additive did not change to the larger sized droplets of In(1) except in the treated PG furnace. It was concluded that smaller sized droplets dispersed on the active sites vaporize easily before the larger sized droplets were formed because of a decreased probability of movement due to interaction with the active sites. Those dispersed in the amorphous carbon with porous morphology form larger sized droplets by collision and coalescence of the smaller sized droplets.Keywords Ascorbic acid; sucrose; chemical modifier; electrothermal atomic absorption spectrometry; indium; kinetic analysis Ascorbic acid has been widely used as an organic chemical modifier in the analysis of complex samples in ETAAS. It has been reported that the effectiveness of ascorbic acid is due to the formation of active carbon species and reductive gases.'-' Reductive gases such as H CH CO and CO affect the ETAAS signal for lead in the gas phase.' Schcherbakov et al. proposed that the addition of ascorbic acid results in a more uniform dispersion of the analyte on the furnace surface prior to atomization and the formation of reductive centres which are activated in the temperature range 950-1070 K.Vaporization of the active carbon species in the temperature range 970-1070 K was observed as a background absorption pulse in the analytical line for gold.6 Recently the character- istics of the pyrolysis of ascorbic acid were investigated by electrothermal vaporization inductively coupled argon plasma mass spectrometry (ETV-ICP-MS) and Raman spectrometry (i) formation of gaseous products such as hydrocarbons CO and COz below 580 K; (ii) formation and release of the active carbon species in the range 600-1100 K; (iii) formation of thermally stable amorphous carbon in the range 1000-1200 K and release of the amorphous carbon uia decomposition into active carbon species in the range 1200-2400 K; and (iv) for- mation of a lower oriented pyrolytic graphite (PG) phase with porous surface morphology after 100 times reported pyrolysis with 5% m/v ascorbic acid at a temperature above 2500 K.6-8 The effect of the two types of carbon produced by the pyrolysis of ascorbic acid on atomization of gold has been reported.6 Gold for which the metal-graphite interaction is as weak as that for indium is adsorbed on both the active carbon species and the thermally stable amorphous carbon.Smaller sized droplets of gold are formed on the active carbon species resulting in a low-temperature shift of the gold signal and larger sized droplets are formed on the thermally stable amorphous carbon resulting in a high-temperature shift of the gold signal. In the presence of 0.02% m/v ascorbic acid since a significant low-temperature shift of the gold signal was observed the less pronounced high-temperature shift the effectiveness of the formation of thermally stable amorphous carbon was less than that with an organic solvent such as methanol or ethan01.~ In the presence of 1% m/v ascorbic acid most of the gold is adsorbed in the thermally stable amorphous carbon resulting in the formation of larger sized droplets.Indium is one of the elements that exhibit a large loss of analyte in ETAAS particularly when using a PG furnace. A non-pyrolytic graphite (NPG) furnace with a porous and active surface gives better sensitivity." McAllister," using graphite furnace mass spectrometry and thermochemical pre- dictions proposed that the thermal dissociation of In,03 (s) to form In,O(g) during atomization cycles above 1000 K causes substantial losses in sensitivity This sensitivity loss reaction is catalysed by carbon atoms at active sites on the graphite wall In,O,(s) + 2C(s)+In,O(g) + 2CO where C(s) is the active carbon atom." A double peak signal with better sensitivity was observed in the NPG furnace and in the PG furnace a single peak signal with two unresolved Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 (601 -606) 601pulses was reported with identical mechanisms for both fur- naces."*" McNally and Holcombe" noted in kinetic investi- gations that two activation energies may arise due to the droplet size such as bulk metal (larger sized droplets) and smaller sized droplets.Gilmutdinov and c o - ~ o r k e r s ~ ~ * ' ~ how- ever described a gas-phase distribution of In atoms and oxides (pg levels of In) in graphite furnaces by the shadow spectral filming technique which is an imaging technique in atomic and molecular absorption spectrometry. They presented the following direct heterogeneous reduction scheme as an atomiz- ation mechanism In,O(g) + C(s) + 2In(g) + CO (3) where C(s) represents the carbon atoms of the hot graphite wall.Recently Imai et a1." proposed the following atomization process to be identical for the PG and NPG furnaces + ZC(S) + C(S) - 2co - co In,O(g) - 2In(g) (first peak) (4) InzO,(s) + 3C(S) - 2In(l) -- In2(g) -+ 2In(g) - 3CO (second peak) The temperature at the second peak in the NPG furnace was higher than that in the PG furnace. With gallium which is one of the elements exhibiting a large loss of analyte in the same way as indium the use of the NPG furnace or 3% m/v ascorbic acid additive gave a better sensitivity in a similar way to indium and it was concluded that this is due to the presence of active sites promoting the formation of a more easily atomized form of gallium rather than the volatile oxide form.15 In this work we found unresolved double peak signals with substantial sensitivity enhancement and thermal stability for the pyrolysis temperature effect when an indium sample solu- tion was deposited in a PG furnace after pyrolysis of the organic matrix solution above 1100 K.Although organic matrix additives gave a sensitivity enhancement the develop- ment of thermal stability was not observed. Further at low concentrations of organic matrix added a decrease in sensi- tivity was found.This work was undertaken to elucidate the effects of an organic matrix pyrolysed and added on the ETAAS of indium. EXPERIMENTAL Apparatus A Hitachi (Marunouti Tokyo Japan) Model 2-8000 flame and graphite furnace atomic absorption spectrometer equipped with a Zeeman-effect background corrector an optical temperature controller system (Hitachi Model 180-0341) and an automatic data processor was used. A 20 pl volume of sample solution was injected by an automatic sampler. The analytical wavelength and spectral bandwidth were 325.6 nm and 1.3 nm for In respectively. An Oki (Toranomon Tokyo Japan) if-800 Model 50 personal com- puter was used to record the absorbance signal profiles at 20 ms intervals.The output data from the optical temperature controller were acquired at 4ms intervals by the computer and stored on a diskette. Temperature data were calibrated using a Chino (Shinjuku Tokyo Japan) Model IR-AH1S radiation pyrometer. The pyrometer was calibrated with a Pt-Rh thermocouple. For this pyrometer the wavelength was 960 nrn and the uncertainties were 0.5% from 870 to 1500 K 1.0% from 1500 to 2300K and 2.0% from 2300 to 3300K. The standard atomizer conditions are given in Table 1. Each measurement was carried out five times. Table 1 Standard atomization conditions Inner gas Ramp Hold flowrate/ Stage Temperature*/"C time/s time/s ml m-' Drying 120 30 0 200 Pyrolysis 800 30 10 200 Cleaning 2900 0 3 200 Atomization? 2500 0 3 30 * Programmed for the atomizer unit.t The optical temperature controller was used. Raman spectra were measured at room temperature by means of a Jobin-Yvon (Atogo Bussan Shinbashi Tokyo Japan) Ramanor T64000 spectrometer based on a triple 0.64 m focal length monochromator equipped with three 1800 grooves mm-' gratings and a 1024 x 256 element CCD detector. A triple subtractive configuration was used with a spectral bandwidth of about 3.5 cm-' with 20 mW of laser power at the sample using argon radiation at 514.5 nm and 180" scat- tering. The wavenumbers of the observed Raman spectra were calibrated using argon plasma lines. Reagents An aliquot of a commercially available standard solution (Wako Chuoku Osaka Japan) was appropriately diluted with 0.1 mol dmP3 HNO before use. Ascorbic acid and sucrose were of laboratory-reagent grade (Wako).Distilled de-ionized water was purified with a Milli-QII system (Waters Milford MA USA). High-purity argon prepared by Takachiho Chemical Industry (Shibuya Tokyo Japan) was used. RESULTS AND DISCUSSION When investigating the pyrolysis products of ascorbic acid and sucrose it is useful to discuss the difference in the effectiveness of the two compounds. It has been reported that two types of carbon species are formed as pyrolysis products of an organic matrix one of which is an active carbon species vaporising in the temperature range 970-1070 K and the other a thermally stable amorphous carbon existing above 1100 K. The vaporis- ation of the active carbon species was observed as a back- ground absorption pulse the response of which corresponded to the organic matrix concentration in the range 970-1070 K.The value of the background absorption for 1% m/v sucrose solution (0.004) was less than that for 1% m/v ascorbic acid solution (0.050) indicating smaller production of active carbon species for the 1 YO m/v sucrose than the 1 YO m/v ascorbic acid solution. To investigate the degree of coating of the PG furnace wall by the thermally stable amorphous carbon Raman spectra were measured after pyrolysis of 5% m/v ascorbic acid and 5% m/v sucrose solutions at 1575 and 1394 K repsectively. The spectra observed with the bare PG furnace wall are shown in Fig. 1. For the PG furnace wall two Raman bands a sharp band corresponding to the E mode near 1584cm-' and a broad band with weak intensity at about 1361 cm-' (disorder mode) appeared.'.'' The broad bands (1603 and 1356 cm-') for the thermally stable amorphous carbon were observed as the laser beam was focused at the centre of the sample compartment of the PG furnace after pyrolysis of 5% m/v ascorbic acid and 5% m/v sucrose solutions at 1575 and 1394 K respectively.The E band for the PG surface was not observed for the ascorbic acid solution but only for the sucrose solution. The intensity of the disorder mode for sucrose was less than that for ascorbic acid. Hence the proportion of surface area coated by the thermally stable amorphous carbon relating to the area of laser beam spot for sucrose is less than that for ascorbic acid. It can also be found from Fig. 3 in ref. 8 602 Journal of Analytical Atomic Spectrometry August 1996 Vol.11Fig. 1 Raman spectra of a surface of the sample compartment of PG furnace. 1 Bare furnace wall; 2 after pyrolysis of 20 pl of 5% m/v ascorbic acid solution at 1575 K; 3 after pyrolysis of 20 pl of 5% m/v sucrose solution at 1394 K that as the bandwidth for the two Raman bands corresponding to the thermally stable amorphous carbon residues decreased with increase in the treatment temperature the amorphous carbon becomes organized with increasing temperature in the range 1200-1700 K. The amorphous carbon is graphitized above 2500 K.* Fig. 2 shows typical transient signals for the atomic absorp- tion observed under the standard atomizer conditions for only indium deposited in PG and NPG furnaces and a PG furnace treated with 1% m/v ascorbic acid solution at 1281 K.The time axis is labelled in such a way that zero time has elapsed before the atomization stage of the heating cycle. The atomiz- ation mechanism for the double peak signal in the NPG furnace with a porous surface was investigated. The single peak signal observed in the PG furnace also consists of two unresolved pulses whose atomization process is identical with that in the NPG furnace. In the PG furnace treated with ascorbic acid an unresolved double peak signal was also observed similar to that in the NPG furnace. When organic matrix solution was pyrolysed in the PG furnace above 1100 K a thin layer of thermally stable amorphous carbon was formed on the PG wall after the active carbon species had been released.The thin layer of amorphous carbon also has a surface with macropores which play an important role in the admission diffusion or transport into the inside of the thin layer producing a decrease in the tendency of smaller sized droplets to diffuse on the graphite A kinetic approach is a useful method although the inter- Time/s Fig.2 Atomic absorption signals for 6ng of indium deposited in bare NPG (1) and PG (2) furnaces and for 2 ng of indium deposited in a PG furnace treated with 20 p1 of 1% m/v ascorbic acid solution at 1281 K (3) with temperature profile. Temperature profile 4 the NPG furnace; and 5 the PG furnace mediate species involved are not observed but speculated upon. A kinetic approach including the rate controlling mechanism for the reaction has been proposed for the single peak signal in the absorbance-time Since unresolved double peak signals were observed for indium deposited in the PG furnace treated with the 1% m/v ascorbic acid at 1281 K the absorbance for the first peak involved in the rising side of the signal must be corrected for the contribution from the second peak to utilize a kinetic appr0a~h.l~ A kinetic approach based on the following equation was adopted to obtain information about the Arrhenius activation energy (E,) in the rate con- trolling mechanism in ETAAS 1n[g(a)Tp2] = -E,R-l T-' + ln[ARP(dT/dt)-'E,-l] (6) where g(a) is a function that depends on the process controlling the reaction rate a is the fractional conversion (a = AJA,, where A, and A are the absorbance at maximum absorption and time t respectively) dT/dt is the heating rate of the furnace wall A is the frequency factor P = 1-2X+6X2-24X3+120X4+*..,X=RT/E,and TandR have their usual meanings.E zpp T, and g(a) for the first and second peaks were defined as E, KPpl Tmaxl and g(a)l and Ea2 zpp2 Tmax2 and g ( ~ ) ~ respectively. A typical Arrhenius plot for 2ng of indium deposited in the PG furnace treated with 1% m/v ascorbic acid at 1281 K is shown in Fig. 3. A straight-line segment was observed in the Arrhenius plots above 1680 K. A first-order reaction (Fl) for the rate-con- trolling mechanism was chosen in order to give the best possible fits; the E value was obtained as 100 f 15 kJ mol-I. To correct the absorbance of the first peak for that of the second peak the absorbance corresponding to the second peak in the period of the first peak was evaluated by extrapolating the straight line for the second peak to each atomization time.For the corrected absorbances for the first peak the F mechanism was chosen and the E value was obtained as 110 & 10 kJ mol-'. When the absorbance correction was not carried out the E value was obtained as 130 f 20 kJ mol-' with the F mechanism. The KPp2 and Tmaxl values can also be estimated with the absorbance correction. In the treatment with 5% m/v ascorbic acid solution the absorbance corrections were carried out in the same way. The E g(a) zPp and Tma values are summarized in Table2 with those obtained at a pyrolysis temperature of 1620K in the PG furnace treated with 1% m/v ascorbic acid at 1360 K.The rate-controlling mechanism was chosen for each case. The process-controlling mechanism can be established from the specific dependence of the absorbance-time profiles on the analyte mass.,* When the -13.0 n 1 b I = t Y .!? -19.0 -2 I .o \ ' I -23.0 5.00 6.00 7.00 8-00 9.00 lo4 T-' I K" Fig. 3 Arrhenius plot with the first-order process-controlling mechan- ism for 2 ng of indium deposited in the PG furnace treated with 20 pl of 1% m/v ascorbic acid solution at 1281 K and calculation of the second signal absorbances in the period of the first signal using eqn. (5). 0 observed value; and 0 calculated value Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 603Table 2 Arrhenius activation energy (E,) appearance temperature ( zpp) temperature at maximum absorbance (T,,,) and process-con- trolling mechanism [g(a)] of indium atomic absorption signal in the PG furnace treated with ascorbic acid solution Concentration of ascorbic acid in the treatment solution E,,/kJ mol-' 1% m/v (1281 K)t 800 1190 k 13 1456 k 20t (1520 f 20) (130 k 20) F1 t 1340 f 151 1890 f 15 100 f 15 F1 110 f 101 5% m/v 1% m/v (1281 K) (1360 K) 800 1620 1190 f 20 1590 k 15t 67 f lot Fl 1 1450 Tt_ 151 1990 f 15 130 f 15 (75 f 10) 1740 & 15 1950 f 15 262 f 20 F1 Fl * Tpyr = pyrolysis temperature for indium atomization; qppl Tmaxl E and g(a)l =parameters for the first peak; ~ p p z TrnaxZ Ea2 and g ( ~ 1 ) ~ = parameters for the second peak.t Treatment temperature. 1 Estimated by using the absorption correction. atomic absorption signals were measured over various indium concentrations the time at the maximum absorbance and at the shoulder maintained a constant value.The dependence of the absorbance-time profiles on the analyte mass indicates that the atomization for both peaks occurs with the F1 mechanism. It was found that there is a decrease in E and an increase in Ea2 with an increase in xPp2 and Tmax2 with increase in the concentration of ascorbic acid in the treatment solution. The amount of active sites and thermally stable amorphous carbon increases with increase in the matrix concentration. When gold sample solution was deposited in the PG furnace treated with organic matrix solution low- and high-temperature shifts of the signal with a decrease and increase in E values respect- ively take place at the same The low-temperature shift was caused by the formation of smaller sized droplets of gold on active sites and the high temperature shift was caused by the formation of larger sized droplets of gold in thermally stable amorphous carbon.McNally and Holcombe12 obtained two sequential atomization energies 142 & 15 and 222 & 8 kJ mo1-l for E and Ea2 for indium deposited by an aerosol in a bare PG furnace without any absorbance correc- tion. The Eal value for the aerosol deposition is close to the Eal and Ea2 values for the treated PG furnace. The first and second atomizations for the aerosol deposition take place via vaporization from smaller sized droplets and bulk vaporization from the larger sized droplets respectively. Thus the observed decrease and increase in Eal and Ea2 values are attributed to the decrease and increase in indium droplet size in different situations such as on the active site and in the amorphous carbon respectively.The integrated absorbance for 2 ng of indium deposited in a PG furnace treated with 20 p1 of an organic matrix solution (0 0.01 1 and 5% m/v ascorbic acid and 1 2 and 5% m/v sucrose) at various treatment temperatures and atomized according to the standard atomization conditions is shown in Fig. 4. Fig. 5 shows typical transient signals for atomic absorp- tion with the PG furnace treated with 1% m/v ascorbic acid solution. At treatment temperatures above 1100 K at which the active carbon species has been released the absorbance imreased with increase in the ascorbic acid concentration indicating the sensitivity enhancement to be in response to the proportion of the area coated by the thermally stable amorph- ous carbon.The absorbance increased with increase in the 0.12 0.10 3 I 0.08 1 0.06 0.04 a CI 0.02 0.00 700 )# 1100 1300 ls00 1700 1900 2100 2300 ZSOI Pyrolysis temperature I K Fig. 4 Integrated absorbance for 2 ng of indium deposited in a PG furnace treated with 20p1 of organic matrix solution at various treatment temperatures. Ascorbic acid concentration 0 0; A 0.01; 0 1; and A 5% m/v. Sucrose concentration 0 1; 4 2; and x 5% m/v 0.20 a 0 a 1 .oo 2 .oo 3.00 s o L 0 2 0 a - 3273 273 0 1 .oo 2.00 3.00 Time/s Fig. 5 Atomic absorption signals for 2 ng of indium deposited in a PG furnace treated with 20 pl of 1% m/v ascorbic acid solution at various treatment temperatures.Treatment temperature 1,840; 2 1128; 3 1281; 4 1384; 5 1683; 6 1779; and 7 1882 K. 8 temperature profile treatment temperature especially for the second peak and reached a maximum at 1480 and 1740 K with 1 and 5% m/v ascorbic acid respectively. This increase in absorbance is due to the suppression of analyte loss according to eqn. (2). When the treatment temperature increased from 1281 to 1384 K the bandwidth of the first peak decreased and that of the second peak increased with increase in the integrated absorbance (lines 3 and 4 in Fig. 5). In the temperature range 1200-1700K with bandwidth for the two Raman bands for the pyrolysed ascorbic acid decreased with increase in the treatment temperature due to organization of the amorphous carbon.The decreases of the first and second peaks are ascribed to a decrease in the disordered sites and to ordering of the amorphous carbon respectively. Across the maximum the second peak was shifted slightly to low temperature with 604 Journal of Analytical Atomic Spectrometry August 1996 Vol. 11increase in the treatment temperature owing to thermal destruction of the amorphous carbon and the formation of active sites. At 1 and 5% m/v ascorbic acid the enhancement was still observed at 2000 and 2400 K respectively. At 2390 K the contribution of amorphous carbon to the gold signal was observed as a broadening of the signal.6 The sensitivity enhancement was less pronounced for sucrose than ascorbic acid at the same concentration of the matrix which is also in response to the Raman spectrometric data.Fig. 6 shows the effect of the pyrolysis temperature on the integrated absorbance for indium deposited in the PG furnace treated with the 1% m/v ascorbic acid at 1360 K deposited in the bare PG and NPG furnaces and for indium sample solution containing the organic matrix in the PG furnace. The absorbance was lost at temperatures below 1400 K in the bare PG and NPG furnaces in which indium atomizes via inter- mediate species In,O(g) and In(1) for the first and second peaks respectively. However in the treated PG furnace a higher absorbance was observed at 1360K above which the first peak disappeared and the signal pulse could be observed at 1800 K. No enhancement of thermal stability was observed for indium sample solution containing the organic matrix.As a single peak signal corresponding to the second peak was only observed at a pyrolysis temperature of 1620 K in the treated PG furnace the kinetic analysis was carried out without the absorbance correction (Table 2). As the Ea2 value is close to the enthalpy of bulk vaporization of indium 240 kJ mol-1,f9 and the E value for the aerosol deposition,12 the indium atomization mechanism is attributed to bulk vaporization from larger sized droplets. It was proposed for the aerosol deposition that the larger sized droplets are formed by the collisions and coalescence of the small particles of indium on the graphite surface. With gold a growth of droplet size in the macropores in a PG furnace treated with 5% m/v ascorbic acid solution at 1270 K was reported.6 Hence an acceptable explanation for the redistribution of indium for the second peak is that the smaller sized droplets dispersed in the thermally stable amorph- ous carbon probably move along the macropores with increas- ing temperature ultimately to form larger sized droplets in the amorphous carbon by collisions.Fig. 7 shows atomic absorption signals for 2 ng of indium in a 1% m/v organic matrix and those with variation in the matrix concentration. The kinetic approach was applied to the second peak for 1% m/v ascorbic acid and sucrose. The results are given in Table 3 with those for the first peak in 0.1% m/v ascorbic acid and 1% m/v sucrose. The Fl mechanism chosen - 0.m I 1 3273 0.01 0 (w( 1m 1m 1400 1600 1880 2mo 22m Pyrolysis temperature I K Fig.6 Effect of pyrolysis temperature on integrated absorbance for indium in the PG and NPG furnaces. PG furnace 0 absence of organic matrix 6 ng In (1.3 scale); A 1% m/v ascorbic acid additive 2 ng In; x 1 YO m/v sucrose additive 2 ng In; and A furnace treated with 20 pl of 1% m/v ascorbic acid solution at 1360 K 2 ng In. NPG furnace 0 absence of organic matrix 6 ng In (1.3 scale) 0.10 R 2.00 3.00 LI 1 .oo s o 42 9 O*'O [(b) 0 Timeis Fig.7 Atomic absorption signals for 2ng of indium dissolved in various concentrations of ascorbic acid (a) and sucrose (b). Concentration 1 0 2,O.Ol; 3,O.l; and 4,1% m/v. 5 temperature profile was supported by the results of the dependence of the profiles on the analyte mass.The Eal values correspond to those reported for CO additive rather than those for pure argon." The Ea2 values obtained correspond to the E and Ea2 values for indium deposited in the PG furnace treated with ascorbic acid solution which indicates the formation of smaller sized droplets of indium metal. The study of the effect of pyrolysis temperature indicated a lower thermal stability for the matrix additive than that for the treated PG furnace (Fig. 6). The absorbance was almost completely lost at a charring tempera- ture of 1400K similar to the absorbance at the first peak observed in the treated PG furnace and the integrated absorbance in the bare NPG furnaces. Hence indium deposited as a solution containing the organic matrix vaporizes before larger sized droplets are formed during the pyrolysis stage.This is due to a decreased probability of movement of the droplets due to interactions with the active sites. It has been reported that 3% m/v ascorbic acid additive gives a better sensitivity for gallium which is also one of the elements that exhibits a substantial sensitivity loss in ETAAS like indium due to volatile monoxide formation at temperatures above 1000K.15 This is due to the reduction of oxides to the metal at a lower temperature before volatile species such as the monoxide are formed.I5 If the reduction mechanism is the dominant mechanism for sensitivity enhancement with indium no decrease in absorbance due to an organic matrix additive should be observed. In this case significant sensitivity losses were observed with a low-concentration organic matrix addi- tive as shown in Fig.8. The sensitivity loss with sucrose is less pronounced than that with ascorbic acid owing to the strength of background absorption for the vaporization of the active carbon species. As it has also been reported that the reduction of In203(s) to In20(g) due to active sites on the graphite takes place the sensitivity loss occurs as a result of the mechanism in eqn. (2) due to active sites in the pyrolysed organic matrix. Above 0.1% m/v ascorbic acid and at 1% m/v sucrose the sensitivity was enhanced at both peaks. It has been reported Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 605Table 3 Arrhenius activation energy (Eal and Ea2) and rate-controlling mtechanism [g(cx)l and g(cx)J for the first and second atomizations of indium First atomization Furnace Conditions Gas Ea,/kJ mol-' g(41 PG* Only indium Ar 226 f 10 Fl NPG* Only indium Ar 201 f 15 F1 PG* Only indium 1.01% co 175 f 10 Fl NPG* Only indium 1.01% co 150 f 15 F PG Ascorbic acid additive Ar 140 f 20t Fl PG Sucrose additive5 Ar 176 f 20 Fl Second atomization E,,/kJ mol-' g ( 4 2 103 f 10 75 f 10 103 f 10 l l O f 15$.F1 120 f 20 Fl * Ref. 10. t 0.1% m/v ascorbic acid. $ 1% m/v ascorbic acid. 0 1% m/v sucrose. that 1.01 YO CO additive gave a sensitivity enhancement due to the increase in the first peak by promotion of the gas-phase dissociation of In,O(g). The E values in the presence of 0.1% m/v ascorbic acid and 1% m/v sucrose correspond to those for the CO additive. The increase in absorbance at the first peak takes place as a result of the promotion of the dissociation of In,O(g).The increase in absorbance at the second peak is due to the increase in the amount of smaller sized droplets dispersed on the active sites of the amorphous carbon. The sensitivity enhancement is attributed predominantly to the reduction of In,O,(s) to In(1) due to active sites and trapping of the In(1) as smaller sized droplets on the active sites. CONCLUSIONS Pyrolysis of an organic matrix provides two types of carbon species. One is the active carbon species formed above about 600 K and released in the temperature range 950-1 100 K. The other is the thermally stable amorphous carbon formed during release of the active carbon species and existing above 1100 K.The active carbon species plays the role of a condensed-phase reductant to reduce In,O,(s) to In,O(g). The gas-phase reduc- tants formed from the organic matrix give a sensitivity enhance- ment owing to promotion of the gas-phase dissociation of In,O(g). The thermally stable amorphous carbon which forms a thin layer directly reduces In203(s) to In(1). Smaller sized droplets are formed on the active sites of the amorphous carbon and in the amorphous carbon. The smaller sized droplets dispersed on the active sites vaporize easily before the larger sized droplets are formed because of a decreased prob- ability of movement due to interaction with the active sites. The smaller sized droplets dispersed in the amorphous carbon layer form larger sized droplets by collision and coalescence of the droplets.Although larger sized droplets are not formed when indium sample solution containing the organic matrix is deposited in the bare PG furnace they are formed when only indium is deposited in the PG furnace treated with organic matrix solution at temperatures above 1100 K. The amorphous carbon layer plays an important role as an area for movement collision and coalescence of indium droplets and produces a decrease in the tendency of larger sized droplets to diffuse on the graphite surface. REFERENCES 1 2 Regan J. G. T. and Warren J. At. Absorpt. Newsl. 1978 17 89. Hydes J. D. Anal. Chem. 1980 52 959. 0.05 I -3 -2 1 0 1 log {[organic matrix](%)} Fig. 8 Integrated absorbance for 2 ng of indium dissolved in various concentrations of organic matrices.Organic matrix broken line none; 0 ascorbic acid; and A sucrose. Errors are represented by bars 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Tominaga M. and Umezaki Y. Anal. Chim. Acta 1982,139,279. Schcherbakov V. I. Belyaev Yu. I. Myasodov Y. B. F. Marov I. N. and Kalinichenko N. B. Zh. Anal. Khim. 1982 37 1717. Gilchrist G. F. R. Chakrabarti C. L. and Byrne J. P. J. Anal. At. Spectrom. 1989 4 533. Imai S. and Hayashi Y. Bull. Chem. SOC. Jpn. 1992 65 871. Imai S. and Hayashi Y. Anal. Chem. 1991 63 772. Imai S. Nishiyama Y. Tanaka T. and Hayashi Y. J. Anal. At. Spectrom. 1995 10 439. Imai S. Okuhara K. Tanaka T. Hayashi Y. and Saito K. J . Anal. At. Spectrom. 1995 10 37. Imai S. Hasegawa N. Hayashi Y. and Saito K. J. Anal. At. Spectrom. 1996 11 515. McAllister T. J. Anal. At. Spectrom. 1990 5 171. McNally J. and Holcombe J. H. Anal. Chem. 1991 63 1918. Gilmutdinov A. Kh. Zakharov Yu. A. Ivanov V. P. and Voloshin A. V. J. Anal. At. Spectrom. 1991 6 505. Gilmutdinov A. Kh. Zakharov Yu. A. Ivanov V. P. Voloshin A. V. and Dittrich K. J. Anal. At. Spectrom. 1992 7 675. Botha P. V. and Fazakas J. Anal. Chim. Acta 1984 162 413. Yoshikawa M. Muter. Sci. Forum 1989 52-53 365. Iwamoto E. Miyazaki N. Ohkubo S. and Kumamaru T. J. Anal. At. Spectrom. 1989 4 433. McNally J. and Holcombe J. H. Anal. Chem. 1987 59 1105. Handbook of Chemistry and Physics ed. Weast R. C. CRC Press Boca Raton FL 75th edn. 1994. Paper 61036626 Received May 28 1996 606 Journal of Analytical Atomic Spectrometry August 1996 Vol. 1 1

ISSN:0267-9477    

DOI:10.1039/JA9961100601

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Selective Extraction and One-drop Flame Atomic Absorption Spectrometric Determination of Trace Amounts of Silver in Highly-pure Copper and Lead Analytical Atomic Spectrometry ISAO KOJIMA AND ASAKO TAKAYANAGI Laboratory of Analytical Chemistry Department of Applied Chemistry Nagoya Institute of Technology Gokiso-cho Showa-ku Nagoya 466 Japan After the dissolution of copper and lead in sulfuric or nitric acid silver was selectively extracted as a ternary silver- 1,2-bis(ethylthio)ethane-picrate complex into a small volume of chloroform at pH 1.0-1.5 leaving copper and lead in an aqueous solution without a masking agent. Direct nebulization of 50 yl of the chloroform extract into a fuel-lean air- acetylene flame gave a sensitive signal height without background correction from chloroform.The present method was successfully applied in the determination of silver in copper and highly-pure copper and lead metals. Keywords Flame atomic absorption spectrometry; silver; ternary siluer-l,2-bis(ethylthio)ethane-picrate complex; highly-pure copper and lead Silver can be determined extremely well by AAS even with use of low temperature flames.' For the determination of trace silver the combination of solvent extraction and direct nebuliz- ation of the extract is very efficient.2 Even by use of 'one-drop' FAAS with direct nebulization of the chloroform extract of a silver complex the sensitivity for the determination of silver was improved about 4.4 fold.3 Various organic reagents have been proposed for the spectro- photometric determination of silver. Of these reagents some were used for the selective extraction of silver in the presence of a masking agent4 Recently it has been reported that thiacrown ethers obtained by replacing the oxygen atoms in crown ethers with sulfur atoms have a great affinity for soft metals such as silver copper(1) and mercury(11)~ and extract these metal ions as ternary complexes in the presence of a bulky anion such as p i ~ r a t e .~ - ~ In addition 1,2- bis(alky1thio)ethane was used as an extracting agent for these metals in the presence of a counter anion.*.'' The present paper describes the selective extraction of silver(1) with a commercially available thioether [ 1,2-bis(ethylthio)ethane] at pH 1.0-1.5 into a small volume of chloroform in the presence of picric acid leaving copper and lead ions in aqueous solution without use of a masking agent followed by 'one-drop' FAAS determination of silver.The method was applied in the determination of silver in highly-pure copper and lead. EXPERIMENTAL Apparatus An atomic absorption spectrometer equipped with a 100 mm burner head and a deuterium background corrector (Seiko Chiba Japan Model SAS-727) was used for metal measure- ments with a discrete nebulization technique using a fuel-lean air-acetylene flame under the same operating conditions as previously r e p ~ r t e d . ~ Sample solutions were injected with a micropipette (Gilson Villiers le Bel France P-200) into a PTFE funnel coupled directly to the nebulizer needle. Signal intensity was recorded using a strip-chart recorder (National Osaka Japan VP-6612a).The pH of the solutions was meas- ured with a Radiometer (Copenhagen Denmark) Type PHM-22 pH meter with a combined electrode. Reagents Stock standard silver solution 2 mg g-' in 0.5 mol dm-3 nitric acid was prepared by dissolving silver (99.99% purity; Nacalai Tesque Kyoto Japan) in nitric acid directly in a PTFE bottle (120cm3) followed by dilution by mass with water. Stock standard copper and lead solutions 2 mg g-' in 0.10 mol dm-3 nitric acid were prepared by dissolving spectrographic stan- dard grade copper and lead (99.999% purity; Mitsuwa Chemicals Osaka Japan) in nitric acid directly in a PTFE bottle (120 cm3 capacity) followed by dilution by mass with water. Working aqueous standard metal solutions were pre- pared by diluting the stock solution in 5-10 cm3 PTFE bottles to the appropriate concentrations by mass with 0.1 mol dmP3 nitric acid.Working standard silver solutions of chloroform extracts were prepared by extracting silver with 1,2- bis(ethy1thio)ethane (8-2s) at pH 1.0-1.5 in the presence of picric acid into an appropriate volume of chloroform the precise volume of which was calculated from the mass of the volume taken using a micropipette and the density. Chloroform of analytical-reagent grade ( Wako Osaka Japan) was used as received. Commercially available 8-2s (Wako) of analytical-reagent grade was dissolved in chloroform (0.05 mol drnp3). Aqueous ammonium pyrrolidin-l-yldithiof- ormate (APDC) solution (1%) was prepared by dissolving APDC (Nacalai Tesque) in water.Picric acid (0.01 mol dm-3) and sodium tetraphenylborate (0.01 mol drnp3) (Wako) of ana- lytical-reagent grade were also dissolved in water. Working standard chloroform solutions containing the copper- and lead-APDC complex were prepared by extracting the mixed solution containing 2.5 cm3 of copper (5 pg) or lead (10 pg) standard solution 0.5 cm3 of 1% APDC solution and 2 cm3 of water with 0.5cm3 of chloroform to test the percentage extraction of copper and lead with 8-2s at different pH values. Sodium acetate of analytical-reagent grade (Wako) was used to control the extraction pH. The other acids used were of analytical-reagent grade. De-ionized Milli-Q water was used throughout. Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 (607-61 0) 607Dissolution of Metals Into a PTFE bottle of 100 cm3 capacity were added 1 g of copper 4 cm3 of concentrated nitric acid 0.5 cm3 of concen- trated sulfuric acid and 3 cm3 of water.The bottle was then warmed on a hot plate to dissolve the metal and then heated to remove nitric acid at a temperature of about 150°C. The residue was diluted to 50 g with water. One g of lead taken in a 100cm3 PTFE bottle was dissolved with 9.7cm3 of warm 2mol dm-3 nitric acid and finally diluted to 50g with 0.1 mol dm-3 nitric acid. Copper solution was prepared as a sulfate solution to avoid the possibility of the extraction of the ternary Cu-8-2s-nitrate complex because divalent sulfate ion does give a ternary complex. For lead nitric acid was used to avoid the precipitation of lead sulfate.Extraction Procedure Into a 10 cm3 stoppered glass test-tube were placed 0.5-1.5 cm3 of sample solution (weighed) 0.5 cm3 of 0.2 mol dm-3 sodium acetate solution 0.5 cm3 of picric acid solution 3.5-2.5 cm3 of water and 0.5 cm3 of 8-2s chloroform solution (weighed). After shaking for 20 s (three times) by hand the test-tube was set aside until complete phase separation had occurred. A 50 p1 portion of the chloroform extract was then directly injected into the flame for AAS measurement of silver. Signal heights were measured for the determination of metal content. Calibration Graph To a 10cm3 stoppered glass test-tube were added aliquots (0-0.5 cm3 weighed) of a standard aqueous silver solution of 1.0 pg g-l 0.5 cm3 of picric acid and 4.5-4.0 cm3 of water.After shaking for 20 s (three times) by hand with 0.5 cm3 of 8-2s chloroform solution (weighed) the test-tube was set aside until complete phase separation had occurred. Portions (50 pl) of the chloroform extracts were directly nebulized into the flame for AAS measurement of the silver calibration standards. For copper and lead the calibration graph was also constructed by nebulizing 50 p1 of the chloroform extracts of the corresponding APDC complexes. 9 0 1 - I V 1 2 3 4 5 PH Fig. 1 Effect of pH on percentage extraction of silver (10 pg; circles) copper (25 pg; squares) and lead (30 pg; triangles) with 8-2s (0.05 mol dm-3) in chloroform in the presence of picric acid (open circles 2 x mol dm-3; filled circles squares and triangles 1 x lo-' mol dm-3 and open triangles 2 x mol dm-3) and tetra- phenylborate 2 x mol dm-' in aqueous phase; open squares).Vaq Vorg= 5 cm' 0.5 cm3 and the amount of organic substances extracted was reduced; hence the background signal was reduced. Effect of Injection Volume on the Signal Height The direct injection of the chloroform extract of a ternary silver-8-2s-picrate complex gave sensitive spike-like and/or very smooth signal profiles as in the case of zinc.3 The background corrected signal obtained by nebulizing a small volume of the chloroform extract increased with an increase in the injection volume up to a volume of about 70 pl at a constant silver concentration in the extract. But the increase in the signal intensity was not large for an injection volume of 50-70 pl. The background signal appeared only at injection volumes greater than 70 pl.This means that background correction is not necessary if an injection volume of less than 601.11 of chloroform extract is used. In order to enhance the extraction concentration of silver and to reduce the background signal an injection volume of 50 pl was used in the present study. Even an injection volume of 50 pl gave a satisfactory reproducibility. RESULTS AND DISCUSSION Effect of pH on Extraction Effect of Some Factors on the Signal Height The effect of pH on the extraction of silver copper and lead was studied using a chloroform solution of 0.05 moldm-3 8-2S in the presence of various amounts of picric acid (2 x 10-4-2 x mol dm-3) or tetraphenylborate anion (2 x lop4 mol dm-3). The results are shown in Fig.1. At a picrate concentration of 1.0 x low3 mol dmp3 98.5% of silver was extracted at pH 1.0-4.0. At a picrate concentration of 2.0 x mol dm-3 silver was quantitatively extracted at pH 1.0. This is expected as the picrate ion is the main species in the picric acid-picrate system at pH 1.0 because the logarith- mic formation constant of picric acid is 0.33." On the other hand only a few ppm of copper were extracted at pH 2.0 under the same conditions. Also extraction of copper increased with increasing extraction pH but lead was not extracted at all at pH 1-3. In addition the extraction of silver decreased with decreasing picrate concentration. When tetraphenylborate anion was used as a counter ion silver was completely extracted at pH 0.5-4.0 (this was omitted in Fig.1 for simplicity) and a relatively high concentration of copper was also extracted (see Fig. 1 ) but lead was not extracted at all. Thus a picric acid concentration of 1.0 x mol dmP3 was used to extract silver selectively from a solution containing large amounts of copper or lead. Copper or lead was left in the sample solution With a constant injection volume of 50 pl of a chloroform extract containing 1.0 p g ~ m - ~ of Ag the effects of the burner height (10.5-15.0 mm above burner head) sample flow rate (3.0-4.0 cm3 min-' of HzO) and acetylene flow rate (3.0-3.8 dm3 min-') on the signal height were studied with and without deuterium background correction. Judging from the baseline stability and the reproducibility of the signal height the same operating conditions as already reported in ref.3 were selected. Calibration Curve The calibration curve obtained by injecting 50 p1 of the chloro- form extracts was a straight line with a zero intercept. A typical example is presented in Fig. 2 with the calibration curve obtained by nebulizing aqueous solutions. The signals were obtained by repeated nebulization of standard extracts and were compared with the signals obtained by injecting sample extracts. The signal height obtained by injecting 50 pl of the chloroform extract containing 0.1 p g ~ r n - ~ of Ag was very constant the RSD obtained in this case was 1.2% (n= 10) (see A in Fig. 2) and the detection limit was 4 ng cm-3 of Ag in chloroform based on the 3 (T criterion. The calibration 608 Journal of Analytical Atomic Spectrometry August 1996 Vol.1 1Table 1 Analytical results for silver in copper metals Sample 5N-CU-A 5N-CU-B Np-Cu Cu taken/mg 10.25 10.32 10.31 10.66 10.52 31.78 31.01 31.01 31.45 31.60 10.50 10.52 10.54 31.61 31.53 31.46 4.23 4.23 4.25 Ag found/pg 0.092 0.096 0.090 0.097 0.095 0.282 0.282 0.286 0.278 0.280 Standard addition method 0.093 0.096 0.096 0.29 1 0.288 0.288 0.385 0.381 0.383 Ag content (ppm) 9.0 9.3 8.8 9.1 9.0 9.2 9.1 9.2 8.9 8.9 8.9 9.1 9.1 9.2 9.1 9.1 91.2 90.8 90.1 Mean s (%RSD) 9.1 k0.2 (1.7) 9.0 9.0 t- 0.1 (1.3) Table 2 Recovery of silver added to sample solution Cu taken/ 10.60 10.57 10.30 30.78 31.01 31.01 Sample mg 5 N - C u - A Ag added/ Ag found/ Recovery Pg I % (%) - 0.096 - 0.101 0.196 99.0 0.098 0.192 98.0 0.097 0.379 102.1 0.248 0.534 102.4 - 0.280 - curve for silver was not affected even in the presence of up to 50 pg cmP3 of Cu in the chloroform extract.Determination of Silver in Metals The present method was applied in the determination of silver in three copper samples C99.999 plus % purity (5N-Cu-A) 99.999% purity (SN-Cu-B) and normal purity (NP-Cu)] and one lead sample C99.999 plus % (5N-Pb)]. The final results were easily calculated from the sample mass used the final sample solution (weighed) an aliquot (weighed) taken from the final sample solution the volume (or mass) of chloroform and the silver content found in the chloroform extract. The results are given in Tables 1-3 and the signal height obtained for the lead sample is given in Fig. 2B. The analytical results obtained by the calibration method and the standard addition method compared well with one another and were reproduc- ible.The recovery of silver added to the sample solution was satisfactory (see Table 2). Table 3 Analytical results for silver in lead [Agl/pg ml-’ in CHCI [Agl/wg ml-’ in aq. soh. Fig. 2 Typical signals obtained by injecting 50 pl of chloroform extracts and 100 p1 of aqueous solutions where V Vorg = 5 cm3 0.5 cm3; 8-2s 0.05 mol dm-3 in chloroform; picric acid 1 x mol dmP3 in aqueous phase; A reproducibility (0.1 pg Ag in cm3 of chloroform); and B sample signal height in pure lead REFERENCES 1 Welz B. in Atomic Absorption Spectrometry 2nd edn. VCH Weinheim 1985 pp. 324-325. 2 West F. K. West P. W. and Ramakrishna T. V. Environ. Sci. Technol.1967 1 717. 3 Kojima I. Inagaki K. and Kondo S. J . Anal. At. Spectrom. 1994 9 1161. Sample Pb taken/mg Ag added/ng 5N-Pb 101.6 101.9 102.1 102.6 102.3 102.6 102.3 - 6.1 12.1 Standard addition method Ag founding 4.1 4.1 4.0 4.1 4.0 10.2 16.1 Ag content (ppb) Recovery (%) 40 40 40 40 39 40 39 41 - 102 100 Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 6094 El-Ghamry M. T. and Frei R. W. Anal. Chem. 1968 40 1986. 5 Oue M. Akarna K. Kimura K. Tanaka M. and Shono T. Anal. Sci. 1989 5 165. 6 Sevdec D. Fekete L. and Meider H. J. Inorg. Nucl. Chem. 1980 42 885 (and papers cited therein) 7 Saito K. Masuda Y. and Sekido E. Anal. Chim. Acta 1983 151 447. 8 Ohki A. Takagi M. and Ueno K. Anal. Chim. Acta 1984 159 245. 9 Oue M. Kimura K. and Shono T. Anal. Chim. Acta 1987 194 293. 10 Dietze F. Gloe K. Jacobi R. Muhl P. Beger J. Petrich M. Beyer L. and Hoyer E. Solvent Extr. Ion Exch. 1989 7 223. Martell A. E. and Smith R. H. in Critical Stability Constants vol. 3 Other Organic Ligands Plenum Press New York and London 1977. 11 Paper 6102372 J Received April 4 1996 Accepted June 11 1996 61 0 Journal of Analytical Atomic Spectrometry August 1996 Vol. 11

ISSN:0267-9477    

DOI:10.1039/JA9961100607

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

CUMULATIVE AUTHOR INDEX JANUARY-AUGUST 1996 Abou-Shakra Fadi R. 61 Adams Freddy C. 201 Akatsuka Kunihiko 69 Akiyama Masayuki 69 Allen Lori B. 526 Angeles Quijano M. 407 Apostoli Pietro 519 Arruda Marco A. Z. 169 Ashino Tetsuya 577 Barinaga Charles J. 317 Barnes Ramon M. 343 Barren James 279 Bavazzano Paolo 5 19 Beato Emilio Romero 37 Begerow Jutta 303 Belkin M. 491 Benzo Zully 445 Besteman Arthur D. 479 Bin He 165 Blades M. W. 43 Bloxham Martin J. 145 509 BrCbion Sophie 497 Bredendiek-Kamper Susanne Brenner I. B. 91 Brockhoff Carol A. 504 Byrne John P. 549 Caldwell Kathleen L. 339 Camara Carmen 407 Camblor Juan Pablo 591 Can0 Pavon J. M. 107 Caruso J. A. 491 Ceulemans Michiel 201 Chamberlain Isa 504 Chapple Graeme 549 Chenery Simon 53 177 Chiappini Remo 497 Chirinos Jose 253 Clevenger Wendy L.393 Cordero Bernard0 Moreno 37 Crain Jeffrey S. 523 Creed John T. 504 Dams Richard 543 Daskalova Nonka 567 Debrah Ebenezer 127 Denoyer Eric R. 127 DePalma Jr. Patrick A. 483 Ding W-W. 225,421 Dolan Scott P. 307 Dunemann Lothar 303 Ebdon Les 427 Efstathiou Constantinos E. 31 Eiden Gregory C. 317 El-Hagrasy Maha A. 379 El-Kourashy Abdel-Ghany 379 Ellis Lyndon A. 259 Elmahadi Hayat 99 Fang Zhaolun 1 Fell Gordon S. 297 Feng Xinbang 287 Fernandez Sanchez Maria L. Flint Colin D. 53 Foulkes Michael 427 Gachanja Anthony 145 537 57 1 Gallego Mercedes 169 Garcia De Torres A. 107 Garcia Sanchez Soledad 37 Garcia Alberto Menendez 561 Gentscheva Galja 567 Gercken Berthold 371 Goodall Phillip S. 57 469 Gregoire D. C. 359 Greibrokk Tyge 117 Gutierrez Ana Maria 407 Haraguchi Kensaku 69 Hartmann C.237 Hasanen Erkki K. 365 Hasegawa Noriyuki 513 601 Hayashi Yasuhisa 5 13 601 Helliwell T. R. 133 Hieftje G. M. 401 Hill Steve J. 145 509 Holclajtner-AntunoviC Ivanka Houk R. S. 247 Hutton Robert 187 Hwang Tarn-Jiun 139 353 Ilkov Atanas 313 Imai Shoji 513 601 Infante Heidi Goenaga 571 Ivanova Elisaveta 567 Iversen Bent Schack 591 Jalkanen Liisa M. 365 Jiang Shiuh-Jen 139 353 555 Jin Qinhan 331 Jin Qun 331 Johnson Stephen G. 57,469 Kabil Mohamed A. 379 Karayannis Miltiades I. 595 Katoh Takunori 69 Kelly S. A. 133 Kiely James T. 523 Kim S. 91 Kingston H. M. 187 Klenerman L. 133 Klockenkamper Reinhold 537 Klockow Dieter 537 Knight Kevyn 53 KO Fu-Hsiang 413 Koh Lip Lin 585 Kojima Isao 607 Koller Dagmar 187 Koppenaal David W.3 17 Kotrebai Mihaly 343 Krivan Viliam 159 371 Krushevska Antoaneta 343 Kumamaru Takahiro 11 1 Lasztity Alexandra 343 Lau Nancy 479 Lerat Yannick 213 Li Gangqiang 401 Liang Feng 331 Liaw Ming-Jyh 555 Littlejohn D. 207 463 Liu Don-Yuan 479 Liu Huiying 307 Lobinski Ryszard 193 Lonardo Robert F. 279 Luan Shen 247 Lyon Thomas D. B. 297 Magnuson Matthew L. 504 325 Mahoney Patrick P. 401 MaloviC Gordana 325 Maquieira Angel 99 Marcus R. Kenneth 483 Masera Eric 213 Massart D. L. 149 237 Matveev Oleg I. 393 Mauchien Patrick 213 McCrindle Robert I. 437 McGaw Brian 297 Michel Robert G. 279 Moens Luc 543 Moissette A. 177 Monod Jean-Louis 193 Montaser Akbar 307 Montero Thais 445 Montoro Rosa 271 Mordoh Leah S. 393 Mostafa M. A. 455 Murillo Miguel 253 Naka Hirohito 359 Nakamura Seiji 69 Nishiyama Yasuko 601 Nogay Donald J.187 O’Hanlon Karen 427 Ohtsuka Hideyuki 69 Olson L. K. 491 Panayi Antonia 591 Pang Ho-Ming 247 Parsons Patrick J. 25 Paschal Daniel C. 339 Patriarca Marina 297 Pavel Jiri 37 1 Pedersen-Bjergaard Stig 117 Penninckx W. 237 Perez-Conde M. Concepcion Perez Pavon Jose Luis 37 Perico Andrea 519 Perkins C. V. 207,463 Pilidis George A. 595 Pinto Carmelo Garcia 37 Piperaki Efrosini A. 31 Polydorou Christoforos K. 31 Puchades Rosa 99 Quentmeier Alfred 537 Quintal Manuelita 445 Rademeyer Cornelius J. 437 RaspopoviC Zoran 325 Rayman Margaret P. 61 Remy Bernard 213 Riter Ken L. 393 Roberts David J. 231 259 Roberts N. B. 133 Rosendahl Kerstin 519 Rubio Marcelo 123 Ruette Fernando 445 Sabbioni Enrico 591 Sadler D. A. 207 463 Saito Kengo 513 601 Sanchez Hector J.123 Sanz-Medel Alfredo 561 571 Schmitt Vincent O. 193 Schwartz Robert S. 307 Shepherd T. J. 177 407 Shkolnik J. 91 Sierraalta Anibal 445 Siitonen Paul H. 526 Siles Cordero M. T. 107 Sivaganesan Manohari 504 Skelly Frame E. 279 Slavin Walter 25 Slavova Petranka 567 Smeyers-Verbeke J. 149 237 Smith Benjamin W. 393 479 Stalikas Constantine D. 595 Sturgeon R. E. 225,421 Sun Han-Wen 265 Szpunar Joanna 193 Taillade Jean-Michel 497 Takada Kunio 577 Takayanagi Asako 607 Tao Guanhong 1 Tao Shiquan 11 1 Taylor Daniel B. 187 Thomaidis Nikolaos S. 31 Thompson Jr. Harold C. 526 Thompson Michael 53 Ting Bill G. 339 Tripkovic Mirjana 325 Turner Andrew D. 231 Tyson Julian F. 127 Uria Enrique Sanchez 561 Valcarcel Miguel 169 Vanhaecke Frank 543 van Holderbeke Mirja 543 Vankeerberghen P.149 Velez Dinoraz 271 Velichkov Serafim 567 Vereda Alonso E. I. 107 Volynsky Anatoly B. 159 von Buhlen Alex 537 Walsh H. P. J. 133 Ward Neil I. 61 Wee Yeow Chin 585 Weir D. G. 43 Wildhagen Dieter 371 Winefordner James D. 393,479 Wittmeier Adolph 287 Wong Ming Keong 585 Worsfold Paul J. 145 509 Wu Shaole 287 Xu Shukun 1 Yang Karl X. 279 Yang Kuei-Lin 139 Yang Li-Li 265 Yang Mo-Hsiung 413 Yang Wenjun 331 Ybaiiez Nieves 271 You Jianzhang 483 Yuzefovsky Alexander I. 279 Zander A. 91 Zhang De-Ciang 265 Zhang Hanqi 331 Zhao Yu-Hui 287 Zhe-Ming Ni 165 Zhou Chao Yan 585 Zochowski Stan W. 53 Zong Yan Y. 25 Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 61 IAnalytical Spectroscopy Monographs Series Editor Neil Barnett Deakin University Victoria Australia Ref No 11 57 Chemometrics in Analytical Spectroscopy By Mike J.Adams University of Wolverhampton Chemometrics in Analytical Spectroscopy provides students and practising analysts with a tutorial guide to the use and application of the more commonly encountered techniques used in processing and interpreting analytical spectroscopic data. In detail the book covers the basic elements of univariate and multivariate data analysis the acquisition of digital data and signal enhancement by filtering and smoothing feature selection and extraction pattern recognition exploratory data analysis by clustering and common algorithms in use for multivariate calibration techniques. An appendix is included which serves as an introduction or refresher in matrix alegra.The extensive use of worked examples throughout gives Chemometrics in Analytical Spectroscopy special relevance in teaching and introducing chemometrics to undergraduates and post-graduates undertaking analytical science courses. It assumes only a very moderate level of mathematics making the material far more accessible than other publications on chemometrics. The book is also ideal for analysts with little specialist background in statistics or mathematical methods who wish to appreciate the wealth of material published in chemometrics. "a worthwhile addition to the bookshelf of any chemist". Chromatographia Vol. 41 No.9/10 November 1995 Hardcover ISBN 0 85404 555 4 viii + 216 pages 1995 Price f 39.50 Ref No 1314 Inductively Coupled and Microwave Induced Plasma Sources for Mass Spectrometry By E.Hywel Evans University of Plymouth Jeffrey J. Giglio University of Cincinnati Ohio USA Theresa M. Casti I lano University of Cincinnati Ohio USA Joseph A. Caruso University of Cincinnati Ohio USA Plasma Sources for Mass Spectrometry provides an efficient means of ionizing most of the elements. This book looks at the most popular and widely used of these sources - inductively coupled plasma. It shows the problems associated with the method such as spectral overlap from polyatomic ions and the inefficiency of element ionization by plasmas formed with argon. Alternative gases to highly purified argon are discussed as well as microwave induced plasmas. Inductively Coupled and Microwave Induced Plasma Sources for Mass Spectrometry looks at how the primary mass spectrometric applications solve problems of this sort and includes the following discussions * interfacing atmospheric plasmas formed in a variety of gases with * atomization and ionization characteristics * polyatomic ion interferences * interfacing chromatographic techniques lhis book presents practical application of the theory in a "tutorial" style for the most important research in the field.It places particular emphasis on interfacing microwave and inductively coupled plasmas with mass spectrometry and interfacing chromatographic techniques and should be read by graduate students and researchers in instrumentation chromatography applied spectroscopy and other areas of analytical chemistry. Hardcover viii + 108 pages l!SBN 0 85404 560 0 1995 Price f32.50 Ref No 1161 mass spectrometry Flame Spectrometry in Environmental Chemical Analysis A Practical Guide By M.S.Cresser University of Aberdeen Flame Spectrometry in Environmental Chemical Analysis is the first title to be published in this new series and is a simple user-friendly guide to safe flame spectrometric methods for environmental samples. It explains key processes involved in achieving accurate and reliable results in atomic absorption spectrometry atomic fluorescence spectrometry and flame emission spectrometry showing the inter-relationship of the three techniques and their relative importance. Flame Spectrometry in Environmental Chemical Analysis presents the important information with thoroughness and clarity and in a style that makes it valuable to students and researchers using these techniques.It also offers straightforward reading for environmentalists with interests in such areas as pollution research agriculture ecology soil science geology and forestry; informing researchers of exactly what they can expect to be able to determine by flame spectrometric methods. Newcomers to flame spectrometry will gain increased confidence job skills and many handy tips and ideas from this book. It will impart a strong working knowledge that can be translated into sound data in the laboratory. '...he (the author) has provided a simple practical guide for environmental scientists who want accurate analytical results. Theory has been kept to a necessary minimum in this concise and clearly written monograph.A Laboratory Equipment Digest January 1995 Hardcover x + 108 pages I!%N 0 851 86 734 0 1994 Price f32.50 To order please contact The Royal Society of Chemistry Turpin Distribution Services Limited Blackhorse Road Letchworth Herts SG6 1 HN United Kingdom. Telephone +44 (0) 1462 672555. Fax +44 (0) 1462 480947. E-mail (Internet) turpin@rsc.org Please quote your credit card details. We can now accept Access/Visa/Mastercard/EuroCard/Amex. Turpin Distribution Services Limited i s wholly owned by The Royal Society of Chemistry. For information on other books and journals please contact Sales and Promotion Department The Royal Society of Chemistry Thomas Graham House Science Park Milton Road Cambridge CB4 4WF United Kingdom. Telephone +44 (0) 1223 420066. E-mail (Internet) sales@rsc.org. RSC members are entitled to a discount on most RSC publications. Details available from the Membership Administration Department at the Cambridge address above. Fax +44 (0) 1223 423429 world wide web http//chemisi:ry.rsc.org/rsc/

ISSN:0267-9477    

DOI:10.1039/JA9961100611

出版商:RSC    

年代:1996

数据来源: RSC