Liquid Sample Introduction Devices in Flow Injection Atomic Spectroscopy* Invited Lecture I Journal of I Analytical J. L. BURGUERA AND M. BURGUERA IVAIQUIM (Andean Institute for Chemical Research) Faculty of Sciences University of Los Andes P.O. Box 542 Mdrida 51 01-A Venezuela A general view concerning some of the advantages of using flow injection analysis manifolds as the interface between sample and/or reagents with different atomic spectroscopy techniques is presented. The development and current status of sample introduction devices in flow injection atomic spectroscopy is reviewed. The different time- and volume-based devices are presented with emphasis on their performance for reliability sample and reagent consumption degree of flexibility robustness and automation capabilities.Keywords Flow injection; automated analysis; sample introduction; atomic spectroscopy The determination of the concentration of a particular element in a sample irrespective of its chemical form is part of the broad analytical field of ‘elemental analysis’. For many years the procedures used in analytical chemistry were overwhelm- ingly dominated by reliance on the chemical reaction of the elements and the fields of gravimetric volumetric and electro- chemistry were widely used in this respect with very limited and primitive equipment. Usually the methods were time- consuming tedious and subject to numerous errors many of which could only be eliminated or corrected for by a high degree of skill and tender loving care on the part of the analyst to ensure precise and accurate results.’ More than any other technique in spectroscopy atomic spectroscopy (AS) has caught the imagination of analytical chemists during.recent years. The field of AS is actually not one technique but four AAS AES AFS,2 and XRF.3*4 These techniques provide some of the most sophisticated and elegant methods for the detection and determination of minor trace and ultra trace metals in clinical environmental agronomic and industrial samples. Among the spectroscopic techniques AAS ETAAS ICP-AES ICP-AFS and ICP-MS have found wide-spread acceptance as high-performance tools for elemen- tal analysis3 However ICP-MS is in some cases more valuable owing to its excellent sensitivity selectivity ability for isotope dilution and extended lineal dynamic range.’ Instruments for AS techniques have been developed with sophisticated computer programs which are able to perform highly reproducible procedures with a minimum of attention by the analyst.Atomic spectroscopy has the advantages of being specific sensitive widely applicable rapid and of wide- spread availability. Approximately 70 elements can be deter- mined directly and many more elements and compounds can be assayed by indirect procedures. Unfortunately sample composition often varies from one sample to the next which determines sample preparation and introduction. Ideally a sample introduction system for AS would transfer the sample reproducibly and efficiently to the atomization/excitation stage. The development of procedures for analytical AS dealing with solution solid and gaseous samples has been a long-term research theme;6 the major * Presented at the Third Rio Symposium on Atomic Spectrometry Caracas Venezuela November 6-12 1994.limitation being sample preparation which frequently requires the use of digestion chambers and reasonably large sample masses and reagent amounts. The batch methodology usually applied for AS evaluations is the most time-consuming step in the analysis procedure. An analysis of a sample can only be as good and less time consuming if the sample preparation is automated. Flow injection (FI)7 is a technique that has intro- duced total modernization of the sample handling steps that are automated to render a sample ready for introduction into the AS instrument without or with minimal intervention by the analyst.FI may be defined as the sequential insertion of discrete sample solutions into an unsegmented continuously flowing stream with subsequent detection of the analyte.8 The simplest flow analyser consists of a propelling unit which is used to impel forward the carrier stream; an injector or sample introduction device by means of which a well-defined volume of sample solution is injected or inter- calated into the carrier stream in a reproducible manner; a micro-reactor which consists of a piece of tubing or a mixing chamber where the sample zone disperses and reacts with the components of the carrier stream; and a dete~tor.~ The signal obtained in FI is transient in nature and analytical measure- ments on this signal can be based on different parameters such as peak height or area as in HPLC.Neither physical equilib- rium (flow homogenization) nor chemical equilibrium is neces- sarily attained at the instant of detection. FI offers several advantages in terms of considerable decrease in sample volume per measurement (normally using 10-50 pl) and reagent con- sumption high sample throughput (50-300 measurements per hour); reduced residence times (readout time about 3-40 s); shorter reaction times (3-60 s) easy switching from one analysis to another (manifolds are easily assembled and/or exchanged); reproducibility (usually < 2% RSD); reliability; low carryover; and a high degree of flexibility (it can be coupled with almost any detection system). Concerning the on-line coupling of an FI assembly to an atomic spectroscopic instrument FI systems can be regarded as an interface between samples and standards with the instru- ments providing useful analytical information about the quali- tative and quantitative composition of samples.The usefulness of integrating FI with AS was recognized very early in the development of FI Initial reports dealt with the use of the FI system as a mere alternative to manual introduc- tion of samples. In such systems the noise problems associated with the air bubbles in segmented sample insertion systems are eliminated the analyst can use sample volumes in the microlitre range per measurement and the sample flow rate is optimized to allow higher atomization rates in the nebulization process with a net increase in sample analysis throughput.However the main advantages of FI in resolving key steps when interfaced to atomic spectrometers are still under active research. These are the development of automated systems for the on-line chemical modification and manipulation of the analyte prior to its introduction into the spectrometer. This Journal of Analytical Atomic Spectrometry July 1995 VoZ. 10 473sample treatment may include minerali~ation,'~-~~ precipi- tation,19 preconcentration or analyte separation using ion- removal of interfering matrix c o m p o ~ n d s ~ ~ . ~ ~ liquid-liquid extraction,26-28 v ~ l a t i l i z a t i o n ~ ~ - ~ ~ etc. A necessary requirement when performing an analysis based on FI is by definition to insert reproducibly a well-defined zone of either sample or reagent into a carrier stream where the zone disperses in a controlled manner on its way toward and through the sensing system.Therefore a crucial and indispensable part of an FI analyser is the injector or sample introduction device. The present paper reviews selectively the different introduction devices used in FI-AS systems with emphasis on their performances in terms of reliability sample and reagent consumption degree of flexibility robustness and automation capabilities. FUNDAMENTAL ASPECTS OF A LIQUID SAMPLE INTRODUCTION SYSTEM Conceptually a sample introduction system (SIS) consists of a device interfaced with a detector by means of a liquid delivery unit which intercalates or injects sample and/or reagent plugs into the flow The entire SIS besides delivering discrete volumes of samples and/or reagents may also perform sampling cali- bration procedures and a series of sample/reagents treatment on-line.The different kinds of sample introduction devices can be divided into three ~ategories:~ (1) volume-based devices; (2) time-based devices; and (3) a combination thereof. In category ( 1 ) the insertion is based on the physical entrapment of a sample and/or reagent solution into a perfectly constant volu- metric cavity and its subsequent transfer in a reproducible manner into the non-segmented carrier stream. The inserted volume must be changeable at least in the range 20-200~1. The volume needed to fill the unit is inevitably larger than the inserted volume but it should not be significantly larger.These kind of devices have been utilized since the early stages of development of FI systems. In category (2) a certain volume of sample and/or reagent solution is drawn aspirated or displaced at a constant flow rate during a fixed period of time into a well-defined section of tubing from where the metered sections are introduced and mixed downstream into an FI system by a combination of hydrostatic and hydrodynamic forces. The sample/reagent volume needed to fill the system could be smaller than the volume of the respective tubing section and in theory the insertion of almost any volume is possible. However this volume may be limited by the electronic system and length of each suction tube section used. Sample introduction device (3) is more versatile and reduces dead volumes per analysis.Its robustness is actually a topic of intense research. VOLUME-BASED INTRODUCTION DEVICES In the early stages of development of FI systems an aqueous sample was introduced manually using a hypodermic syringe34 and a flap35 valve into a continuously moving unsegmented carrier stream of water or reagent solution; an injection principle which was also adapted to FI-AS36-38 and FI-ICP39 and FI-ICP-MS40,41 systems. The use of syringes7p36-38,42 h as gained some popularity because of the simplicity economy and rapid availability of the experimental setup; and the injection of the required volume of sample per measurement. There were however some severe limitations in their use (1) the volume of material injected depends on the position of the plunger which makes its use tedious; (2) the form of the injected plug depends on the speed with which the plunger is being depressed especially with liquids of different viscosity; (3) the rubber septum which is part of this kind of injector may not be compatible with all samples and carrier stream solutions resulting in a chemical attack of the septum's elasto- mer with possible chemical leaching; (4) after repeated injec- tions pieces of the septum eventually become detached and block the flow line; and ( 5 ) the introduction of the syringe needle causes a temporal flow change which leads to physical mixing of the adhered sample in the needle with the carrier stream interface.Rocks et al.43 described an inexpensive double three-way stopcock valve for sample introduction into the flowing stream of an FI-AAS system.It consisted of a borosilicate glass barrel fitted with a tapered PTFE key. The key has two right-angled bores which can be aligned each with three limbs as required. By choosing different lengths of sample loops the injected volume may be varied. As well the valve contains a bypass tube of higher hydrodynamic resistance so that the valve allows the carrier flow to continue when the valve loop is filled. This valve is difficult to adapt to different FI systems must be rotated manually and is also prone to leak or stick when exposed to a heavy work load. Because of their limitations as the scaling down of FI systems gained momentum these kind of injectors were quickly abandoned and today are of merely historical interest. The use of syringe based injectors were rapidly replaced by HPLC-type rotary valves.44 These valves have one or two external loops which can be replaced for changing the injected volume.Although it reproducibly intercalates a well defined zone of sample or reagent fluid into the carrier stream by switching the valve from a functional point of view the use of a syringe to fill the loop is tedious and too slow for practical assays and the risk of contamination of the sample from the metallic needle of most syringes is troublesome when assaying ultra trace elements. This led to inexpensive and ingenious designs of low-pressure rotary valve introduction devices of approximately the same volumetric scale as the commercially available HPLC v a 1 v e ~ .~ ~ ~ * ~ ~ ~ ~ Although the different rotary valve devices have been designed and mechanically manufac- tured differently the function of this type of valve is the intercalation of a fluid into the carrier stream. Their functional principle is based on the loading and intercalation positions. In the load position while the carrier stream flows through the valve in a bypass channel the sample fills the loop by aspiration with a syringe or a pump. When the valve is turned to the alternate position the sample plug is inserted into the carrier stream and therefore swept out of the volumetric cavity. Each rotary valve consists of three layers where a central moveable rotor is sandwiched between two stators. The rotor has two pairs of holes accommodating the sample and the internal bypass loop drilled to match two pairs of parts in the stator.Plexiglas clear PVC poly (vinylidene difluoride) (PVDF) and PTFE are the most used materials to construct these rotary valves; however in commercial valves the parts in contact with the solutions are made from inert fluorocarbons in order to avoid problems from corrosion. Numerous designs of home made and commercially available rotary valves have been tried often characterized by increasing levels of com- plexity. Two four six and eight-port valves have been con- s t r ~ c t e d ~ ~ ~ ~ ~ and used in different FI configurations which allows simultaneous sample/reagent introduction and split and nested sample loops to provide (1) a wide range of functions such as merging zones and zone sampling procedures (e.g.simple alternate sample introduction metering of separate volumes of sample and reagent triple zone introduction. and introduction of two adjacent zones); and (2) physical treatment or chemical modification of samples by ion-exchange precon- centration dialysis gas-diffusion digestion dilution filtration immobilized enzymes and chemical Up to now the majority of workers have favoured the use of the above described rotary valves probably because they are commercially available with variable degrees of automated control and can be readily used in combination with automatic 474 Journal of Analytical Atomic Spectrometry July 1995 Vol. 10samplers.52 However all systems employing rotary valves suffer from a common fault. Not only does the sample fill the volume of the sample loop but the entry and exit tubes must be filled with sample.Additional sample is used to flush out the residual of the previous specimen. Therefore the injection process is wasteful of sample requiring a sample consumption of at least twice the volume actually delivered to the AS instrument. This waste of sample volume becomes critical when most clinical samples are analysed. Loops may be bent or otherwise dam- aged giving erroneous results which necessitate immediate corrective action. Rotary valves are also prone to leakage when exposed to the heavy work load of a routine laboratory (usually between 5000 and 10000 measurements). Nevertheless the cost of most rotary valves is relatively low and replacement of worn-out parts of the injector can be easily made.Bergamin and C O - W O ~ ~ ~ ~ S ~ ~ - ~ ~ have initiated the double proportional sliding valve insertion commutator which is capable of performing merging-zone (or the insertion of several zones simultaneously which penetrate into each other down- stream7) and zone sampling procedures. This can be achieved in two different ways by intermittent pumpings6 or by the use of multiple inje~tion.'~ The simplest version of this kind of introduction device consists of two external plates with a moveable centre bar of metal polyethylene or Perplex held together by two spring-loaded screws. Holes are drilled through the commutator pieces in accordance with the flow diagram used. Silicone rubber sheets with holes are placed between the commutator plates to avoid leakage.For most applications the performance of these commutators seems to be comparable to those of rotary valves including the possibility of an easy automation of the sample introduction and on-line manipulation operations. With the hydrodynamic injection technique described by Ruzicka and Han~en,'~-'~ the introduced volume is well defined without the need of moving parts. The valve is replaced by two T junctions connected by a channel of known volume. This volume defines the sample volume to be introduced. The carrier stream is stopped and the sampling stream is simul- taneously turned on to fill the volume to be introduced. The sampling stream is then turned off and the carrier stream is turned on to complete the insertion cycle.Two peristaltic pumps are used to control the movement of sample and carrier solutions. Zagatto et ~ 1 . ~ ~ have described a hydrodynamic sample introduction procedure based on commutation which requires only one peristaltic pump operating continuously. Although the precision is good (usually RSD <1%) the inclusion of a commutator valve defeats one of the main purpose for which hydrodynamic sample introduction was developed i.e. to eliminate moving parts and concomitant wear. Despite its inherently favourable characteristics hydro- dynamic sample introduction has not been widely used in AS routine analyses probably owing to the difficulty in balancing hydrodynamic forces which causes a much worse precision (z 2.5%) than rotary valve introduction. TIME-BASED INTRODUCTION DEVICES Riley and co-workers60*61 have introduced the valveless time- based introduction method controlled dispersion analysis which consumes minimum amounts of sample and reagent.The probe transfer mechanism was a simple cam timer- operated device that raised and lowered an arm and rotated it through 90". The arm normally carried a sample probe and a reagent probe with provision for fitting an additional reagent probe if needed. In the simplest form of this system the peristaltic pump is driven by a stepping motor controlled by a microcomputer. The sample probe normally rests in a reagent container so reagent is pumped through the system between samples. When a sample is presented to the machine the pump is stopped and the probe is transferred to a sample container.The pump then rotates through a precise predetermined angle and again stops. The probe is returned to the reagent container and the pump restarted. The sample slug travels up the probe through the pump and onwards into the reactor tubing and finally to the detector no part of the sample being wasted. The introduction of air bubbles in this kind of sample introduc- tion device is not possible because the pump (driven by a stepping motor) is stopped while the probe is transferred either to the sample or reagent container. Notwithstanding the instrumental likeness to auto sample arms there are a number of essential differences between the two the most marked of which are the aspiration systems used by auto sampler arms withdraw the samples from the sampler vials and dispense them to the analyser cups or cuvettes; the time is not such a decisive factor; and high-precision syringes are the commonest option for measuring and transferring the liquid samples.Whereas in a valveless time-based device the sample and reagents are drawn by controlled movement of a peristaltic pump. The major drawback of this mode of sample introduc- tion is the momentaneous negative over-pressure created when the pump restarts which greatly deteriorates the reproducibility when used for AAS measurements. Jorgensen et a1.62 described a method to achieve a time- based introduction of sample using a two-position valve. The valve is opened to the sampling position for a precise time then is switched to the carrier position.The advantage cited for using a two-position valve is that the pump never stops thus the introduced volumes are more precise. The performance of this kind of system appears to be comparable to those of hydrodynamic sample introduction techniques and a valve. A different approach to sample introduction into a carrier stream with a time-based device and a simple timer circuit which can switch the current needed to activate a solenoid valve on and off at desired fixed intervals has been described by Burguera et ~ 1 . ~ ~ In this way the sample or carrier streams are either closed or opened,64 allowing the intercalation of variable and precise microlitre volumes of sample into the carrier stream with an average deviation of about 0.6%. While the carrier stream was propelled with a peristaltic pump the sample plug was introduced through a gas-pressurized reser- voir. With this sample introduction device calibration graphs can be constructed from a single solution65 and on-line dilution of samples is also possible.66 Further improvements such as the incorporation of a doubly stopping shutter (to close or to open any selected set of tubing at a fixed time) and an electronically controlled timer allow the performance of on-line programmed functions which may include sequential sample and reagent introduction merging zones zone trapping sample preconcentration and in uiuo sample uptake for measurements of metal species by AAS and ETAAS.16-'8,23,67-70 Gas diffusion units filters and ion-exchange micro-columns can be easily included to perform different on-line sample pretreatment procedures.Besides sample and reagents are not consumed during the stoppage of the sample zone the dead volume is reduced to a minimum value almost any desirable volume of sample and reagents can be readily inserted and the set-up is economical and easily adapted to a particular analytical application. VOLUME AND TIME-BASED DEVICES The hydrodynamic principle can be used to combine volu- metric sample metering with time-based insertion using two and three pumps which operate at different volumetric flow rates.7 All of these exploit the use of confluence points at which a well-defined sample zone is formed by means of the alternate motion of sample and carrier stream. This approach has not yet been adapted to systems where AS fulfils the role of the Journal of Analytical Atomic Spectrometry July 1995 Vol.10 475detector probably because the movement of the sample con- tainer has to be accurately sequenced with the operation of the pumps in order not to impair the performance of the system. This drawback could be eliminated in the future by electronically controlling the whole operation. Tyson et al. have described a variable tube dimension manifold.71 By switching the same injected volume by means of a rotary injection valve along tubes of different dimensions (length and internal) dispersion coefficients ranging from 6 to 40 in six discrete steps were obtained. The manifold has been applied to the determination of calcium chromium and nickel by FAAS. These kind of manifolds can be readily used to produce various calibrations of varying sensitivity by splitting the zone sample and confluence points at different dimen- s i o n ~ .~ ~ Although versatile the main limitations are (1) the number of connecting lines which at the same time make these kind of manifolds cumbersome; and (2) a decrease in peak height and peak area in proportion to the tubing length which greatly reduces the sensitivity of the analyses. A multipurpose flow insertion system composed of a combi- nation of a sliding insertion commutator with up to eight solenoid valves has been recently described by Reis et ~ 1 . ~ ~ The sample and standard solutions are introduced by discrete commutation in three or more modules each composed of two solenoid valves (three-way type).By sequentially coupling the different modules and electronically actioning the different valves the merging zones zone sampling intermittent flow sequential injections zone trapping and stopped flow functions can be performed. On-line programmed dilution and standard additions for simultaneous determinations of metal species on plant digests by ICP-AES were chosen to demonstrate some features of the system. CONCLUSIONS It is clear that FI focuses its attention on the weakest link of modern instrumentation which is the sample introduction and pretreatment processes. Although some inadequacies still exist FI improves the analytical performance of AS offering many unique advantages for the handling and on-line pretreatment of sample^.^^-^^ The recent development of improved sample and/or reagents introduction devices determines the trends of its essential features.Electronically controlled volume-based devices will increasingly continue to offer solutions to some of the sample preparation steps with low sample consumption per measurement for a number of years. However further research with fully automated time-based introduction devices will greatly simplify the different and common preparation steps necessary for converting matrices into an acceptable form for AS determinations with a simul- taneous decrease in sample and reagent consumption per analysis. REFERENCES 1 Robinson J. W. Atomic Spectroscopy M. Dekker New York 1990. 2 Burguera M. Burguera J. L. and Pacey G. E. in Flow Injection Atomic Spectroscopy ed.Burguera J. L. Marcel Dekker New York 1989 pp. 293-321. 3 Greenfield S. J. Anal. At. Spectrom. 1994 9 565. 4 Janssens L. Vincze L. Rubio J. Adams F. and Bernasconi G. J. Anal. At. Spectrom. 1994 9 151. 5 Sheppard B. S. and Caruso J. A. J. Anal. At. Spectrom. 1994 9 145. 6 Sneddon J. Sample Introduction in Atomic Spectroscopy Elsevier Amsterdam 1990. 7 Ruzicka J. and Hansen E. H. Flow Injection Analysis John Wiley New York 2nd edn. 1988. 8 Stewart K. K. General Introduction in Flow Injection Atomic 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 Spectroscopy ed. Burguera J. L. Marcel Dekker New York Valcarcel M. and Luque de Castro M. D. Flow Injection Analysis Principles and Applications Ellis Horwood Chichester 1987.Stewart K. K. Talanta 1981 28 789. Zagatto E. A. G. Krug F. J. Bergamin Fo. F. H. Jorgensen S. S. and Reis B. F. Anal. Chim. Acta 1979 104 279. Broekaert J. A. C. and Leis F. Anal. Chim. Acta 1979 109 73. Burguera M. Burguera J. L. and Alarcbn 0. M. Anal. Chim. Acta 1986 179 351. De la Guardia M. Salvador A. Burguera J. L. and Burguera M. J. Flow Injection Anal. 1988 5 121. Haswell S. J. and Barclay D. Analyst 1992 117 117. Burguera J. L. and Burguera M. J. Anal. At. Spectrom. 1993 8 235. Burguera M. Burguera J. L. Rondon C. Rivas C. Carrero P. and Brunetto M. R. J. Anal. At. Spectrom. 1995 10 343. Burguera M. and Burguera J. L. Lab. Robotics Autom. (LRA) 1993 5 277. Valcarcel M. Gallego M. and Martinez-JimCnez P. Anal. Proc. 1986 23 233.Valcarcel M. in Sample Zntroduction in Atomic Spectroscopy ed. Sneddon J. Elsevier Amsterdam 1990 pp. 289-327. Fang Z. and Welz B. J. Anal. At. Spectrom. 1989 4 543. Welz B. Sperling M. and Sun X. Fresenius’ J. Anal. Chem. 1993 346 550. Burguera J. Burguera M. Carrero P. Marcano J. Rivas C. and Brunetto M. R. J. Autom. Chem. 1995 in the press. Kamson 0. F. and Townshend A. Anal. Chim. Acta 1983 155 253. Burguera J. L. Burguera M. Rivas C. Gallignani M. Carrero P. and Brunetto M. R. presented at the Third Rio Symposium Caracas 1994. Valcarcel M. and Gallego M. in Flow Injection Atomic Spectroscopy ed. Burguera J. L. Marcel Dekker 1989 Memon M. A. Zhuang Z. X. and Fang Z. L. At. Spectrosc. 1993 14 50. Kuban V. Crit. Rev. Anal. Chem. 1991 22 477. Astrom O.Anal. Chem. 1982 54 190. Liversage R. R. Van Loon J. C. and de Andrade J. C. Anal. Chim. Acta 1984 161 275. Morita H. Kimoto T. and Shimomura S. Anal. Lett. 1983 16 1187. Burguera M. Burguera J. L. Brunetto M. R. De la Guardia M. and Salvador A. Anal. Chim. Acta 1991 261 105. Burguera M. and Burguera J. L. J. Anal. At. Spectrom. 1993 8 224. Ruzicka J. and Hansen E. H. Anal. Chim. Acta 1977 89 241. Ruzicka J. Hansen E. H. and Mosbaek H. Anal. Chim. Acta 1977 92 235. Rocks B. F. Sherwood R. A. and Riley C. Clin. Chem. 1982 28 440. Fukamachi K. and Ishibashi N. Anal. Chim. Acta 1980,119,383. Burguera J. L. Burguera M. and Gallignani M. An. Acad. Bral. Cienc. 1983 55 209. Jacintho A. O. Zagatto E. A. G. Bergamin F. H. Krug F. J. Reis B. F. Bruns R. E. and Kowalski B.R. Anal. Chim. Acta 1981 130 243. Bloxham M. J. Hilland S. J. H. and Worsfold P. I. J. Anal. At. Spectrom. 1994 9 935. Ebdon L. Fisher A. S. and Wrosfold P. I. J. Anal. At. Spectrom. 1994 9 611. Van Staden J. F. in Flow Injection Atomic Spectrocopy ed. Burguera J. L. Marcel Dekker New York 1989 pp. 49-101. Rocks B. F. Sherwood R. A. Turner Z. J. and Riley C. Ann. Clin. Biochem. 1983 20 72. Ruzicka J. and Hansen E. H. Flow Injection Analysis John Willey New York 1981 pp. 104-109. Karlberg B. and Pacey G. E. Flow Injection Analysis. A Practical Guide Elsevier Amsterdam 1989. Valcarcel M. and Luque de Castro M. D. in Andisis por Injeccidn en Flujo University of Cordoba Cordoba Spain 1984. Ruzicka J. Fresenius’ J. Anal. Chem. 1986 324 745. Martinez-Jimenez P. Gallego M.and Valcarcel M. Anal. Chem. 1987,59 69. Kuban V. Crit. Rev. Anal. Chem. 1992 23 15. Welz B. Sperling M. and Sun X. Fresenius’ J. Anal. Chem. 1993 346 550. 1989 pp. 1-17. pp. 157-224. 476 Journal of Analytical Atomic Spectrometry July 2995 Vol. 1051 52 53 54 55 56 57 58 59 60 61 62 63 64 65 Santelli R. E. Gallego M. and Valcarcel M. Talanta 1994 41 817. Van Staden J. F. and Van Rensburg A. Analyst 1990 115 605. Bergamin H. F. Zagatto E. A. G. Krug F. J. and Reis B. F. Anal. Chim. Acta 1978 101 17. Zagatto E. A. G. Krug F. J. Bergamin H. F. Jorgensen S. S. Reis B. F. Anal. Chim. Acta 1979 104 279. Zagatto E. A. G. Krug F. J. Bergamin H. F. and Jorgensen S . S. in Flow Injection Atomic Spectroscopy Marcel Dekker New York 1989 pp. 225-257. Ruzicka J. and Hansen E.H. Anal. Chim. Acta 1980 114 19. Ruzicka J. and Hansen E. H. Anal. Chim. Acta 1983 145 1. Clark G. D. Whitman D. A. Christian G. D. and Ruzicka J. Crit. Rev. Anal. Chem. 1990 21 357. Zagatto E. A. G. Bahia F. O. Gine M. F. and Bergamin H. F. Anal. Chim. Acta 1986 181 265. Riley C. Aslett L. H. Rocks R. Sherwood R. A. MacK Watson J. D. and Morgan J. Clin. Chem. 1983 29 332. Riley C. Rocks B. and Sherwood R. A. Talanta 1984 31 879. Jorgensen S. S. Petersen K. M. and Hansen L. A. Anal. Chim. Acta 1985 169 51. Burguera J. L. Burguera M. Rivas C. De la Guardia M. and Salvador A Anal. Chim. Acta 1990 234 253. Carrero P. Burguera J. L. Burguera M. and Rivas C. Talanta 1993 40 1967. Burguera J. L. Burguera M. Rivas C. De la Guardia M. Salvador A. and Carbonell V.J. Flow Injection Anal. 1990,7 11. 66 67 68 69 70 71 72 73 74 75 76 De la Guardia M. Morales-Rubio A. Carbonell V. Salvador A. Burguera J. L. and Burguera M. Fresenius’ J. Anal. Chem. 1993 345 579. Burguera J. L. Burguera M. Matousek de Abel de la Cruz A. Aiiez N. and Alarcon 0. M. At. Spectrosc. 1992 13 67. Burguera J. L. and Burguera M. Lab. Robotic Autom. (LRA) 1991 3 119. Burguera J. L. Burguera M. Carrero P. Rivas C. Gallignani M. and Brunetto M. R. Anal. Chim. Acta 1995 in the press. Burguera M. Burguera J. L. Rivas C. Carrero P. Brunetto M. R. and Gallignani M. Anal. Chim. Acta 1995 in the press. Tyson J. F. Anal. Chim. Acta 1986 179 131. Tyson J. F. in Advances in Atomic Spectroscopy ed. Sneddon J. JAI Press Greenwich CT 1992 vol. 1 pp. 161-228. dos Reis B. F. GinC M. F. Krug F. J. Bergamin H. F. J. Anal. At. Spectrom. 1992 7 865. McLeod C. M. J. Anal. At. Spectrom. 1987 2 549. Burguera J. L. and Burguera M. J. Trace Elem. Electrolytes Health Dis. 1993 7 9. Christian G. and Ruzicka J. Spectrochim. Acta Part B 1987 42 157. Paper 4/07094A Received November 21 1994 Accepted March 8 1995 Journal of Analytical Atomic Spectrometry July 1995 Vol. 10 477