JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 199 1 VOL. 6 353 Solid Sampling in Electrothermal Atomic Absorption Spectrometry Using Commercial Atomizers A Review Carlos Bendicho* and Margaretha T. C. de Loos-Vollebregt Laboratory of Analytical Chemistry Delft University of Technology De Vries van Heysfplantsoen 2 2628 RZ Delft The Netherlands Summary of Contents 1 introduction 2 Direct Solid Sampling 2.1 Atomization Systems 2.1.1 Graphite tube atomizer 2.1.2 Cup cuvette atomizer 2.1.3 L’vov platform techniques 2.1.3.1 Platform atomization 2.1 -3.2 Boat atomization 2.1.3.3 Microboat atomization 2.1.4 Cup atomizer 2.1.5 Cup-in-tube technique 2.1.6 Carbon rod atomizer 2.1.7 Graphite probe atomizer 2.1.8 Ring chamber atomizer 2.1.9 Second surface atomizer 2.2.1 Influence of particle size 2.2.2 Influence of sample homogeneity 2.2.3 Influence of sample mass 2.2.4 Influence of analyte location in the sample 2.2 Accuracy and Precision 2.3 Chemical Modification 2.4 Sample Introduction Systems 2.5 Sample Preparation 2.6 Calibration 2.7 Applications of Direct Solid Sampling 3 Slurry Sample Introduction 3.1 Preparation and Mixing of Slurries 3.1.1 Use of stabilizing agents 3.1.2 Magnetic stirring and vortex mixing 3.1.3 Ultrasonic agitation 3.1.4 Gas mixing of slurry 3.1.5 Pre-digestion of slurry 3.1.6 Electric dispersion in a condensed medium 3.2 Influence of Particle Size 3.3 Influence of Slurry Concentration 3.4 Chemical Modification 3.5 Calibration 3.6 Applications of Slurry Sample Introduction 4 Direct Solid Sampling Versus Slurry Sample Introduction 4.1 Sample Pre-treatment 4.2 Sample Introduction 4.3 Dilution of the Solid Sample 4.4 Sample Homogeneity 4.5 Peak Shape 4.6 Calibration 5 Conclusions 6 References Keywords Direct solid sampling; slurry sample introduction; electrothermal atomic absorption spectrometry; commercial atomizer; review * Present address Perkin-Elmer Hispania S..4. Calle La Mas6 2 28034 Madrid Spain354 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1991 VOL. 6 1 Introduction The success of the technique depends on the skill of the The direct analysis of solid samples without chemical pre- treatment is a goal widely pursued in order to simplify the analytical procedure. The application of electrothermal atomic absorption spectrometry (ETAAS) to the direct analysis of samples is as old as the technique itself.' Electrothermal AAS has received increased attention for the direct determination of metals at trace levels in a wide variety of samples owing to its hgh sensitivity the simplicity of the instrumentation and the relatively low cost.Several reviews have addressed the solid sampling tech- nique from a general point of vie^,^-^ making reference to either specific ~ a m p l e s ~ - ~ or comparing different atomic spectrometric techniques.' Specific features of solid sampl- ing have been claimed throughout the literature. The advantages of using solid sampling over acid digestion or fusion procedures are obvious and they can be summarized as follows (i) reduced sample pre-treatment and hence an increase in the speed of the whole analytical procedure; (ii) low contamination risk an essential requirement when trace levels of metals are to be determined; (iii) fewer possibilities of analyte losses during the sample pre- treatment or retention by insoluble residues; and (iv) the use of corrosive and hazardous chemicals is avoided.The number of papers published between 197 1 and 1990 shows the impact of the solid sampling technique (Fig. 1). In this review attention is mainly focused on the literature concerning the application of commercially available atom- izers for solid sampling. Among the different procedures for inserting solid samples into the electrothermal atomizer two approaches are widely used firstly direct introduction of the solid sample and secondly the introduction of a slurry prepared by suspending the finely divided solid sample in an appropriate liquid diluent. The period between 1971 the year when the first application of solid sampling with a commercial atomizer was reported,8 and 1980 reflects the early development of the solid sampling technique.The applications reported for solid sampling were characterized by the use of laboratory- built atomizers capable of accepting solid samples wall atomization slow heating rates for atomization (ramp times of 1 s and longer) peak absorbance measurements and matrix-matched solid standards for calibration in addition to the use of unreliable and manually-operated sample insertion devices. analyst in many instances. I t is clear that problems and drawbacks arise when chemical pre-treatment is avoided.These problems are listed below. (i) Insertion of the solid sample into the atomizer is not as easy as it is for liquid samples. Direct introduction of the sample requires extreme care and training. Loss of sample and contamination can occur during the weighing andor transfer from the balance to the atomizer. This problem is less critical when using slurry introduction. (ii) A different weighing is required for each replicate in contrast to liquid samples which use direct insertion. This can make the analytical procedure tedious and clearly unacceptable for routine determinations. With slurry intro- duction several determinations can be performed using only one sample preparation. Some improvement has been achieved by the automation of solid sample introduction for routine determinations (see section 2.1.3).(iii) Problems arise with standardization that do not occur when working with liquid samples. Matrix depen- dence of the peak shape and incomplete release of the analyte from the solid sample are commonly found. Calibration with aqueous standards is only successful sometimes and the use of the standard additions method fails when the added analyte is not equally affected by the matrix. Calibration with certified reference materials (CRMs) of very similar matrix to the samples constitutes the most reliable approach. (iv) Chemical modifiers are less effective for the direct atomization of solids as the chemical modifier cannot effectively interact with the analyte which is occluded within the sample particles.Several solid chemical modifi- ers have been tried in order to improve this interaction. (v) Solid samples offer fewer possibilities for dilution when the concentration of analyte is high for which a decrease in the instrumental sensitivity or prior dilution of the solid sample with an appropriate solid buffer are required. The slurry technique allows dilution but critical use of dilution is required in order to avoid errors e.g. when only a few particles remain in the slurry. (vi) Precision is worse than for liquid samples. Typically relative standard deviations (RSDs) are about 10%. Preci- sion can be improved by either decreasing the particle size or increasing the mass used for the determination when sample inhomogeneity is the main factor affecting preci- sion.For many CRMs sample inhomogeneity has to be taken into account for sample masses of below 200 mg; this is much higher than the 0.1-10 mg used in solid sampling. (vii) Parts of the matrix can remain in the furnace after the atomization step and thereby cause the deterioration of the atomizer. The build-up caused by the matrix can also block the light beam emitted from the radiation source and hence affect the analytical performance. A cleaning-out step is necessary after a series of measurements in order to avoid this problem. (viiz9 Spectral interferences are more pronounced when the solid sample is atomized directly. High and structured background signals can cause large errors. The use of a powerful background correction system is required if accurate results are to be obtained.Oxygen ashing can also be helpful to remove non-specific absorbance interferences. (ix) Specific problems are associated with slurry intro- duction such as the need to maintain a stable and homogeneous slurry and contamination by the liquid (x) Problems arise when determining refractory metals owing to the incomplete release during atomization and OCClUSiOn in the solid matrix. Higher atomization tempera- tures are required to release the analyte in a reasonable period of time in comparison with liquid samples. 1970 1975 1980 1985 1990 diluent. Year Fig. I Number of papers published per year in the period between 1971 and 1990 in direct solid sampling (solid line) and slurry sample introduction (broken line) based on the 249 references included in the reviewJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 199 1 VOL.6 355 (xi) Problems can arise in the preparation of blanks especially when using ground samples. Many of these problems have been partly solved during the last decade by the implementation of various new instrumental developments and by improving the analytical methodology. Progress in three different lines of develop- ment accounts for the present use of solid sampling as a reliable technique which is often preferred to the chemical pre-treatment of samples. These include the design of graphite atomizers specifically adapted to handle solid samples; the automation of solid sample introduction; and the use of stabilized temperature platform furnace (STPF) technology.The introduction in 1981 of improvements in furnace technology by Slavin and c o - w o r k e r ~ ~ ~ ~ ~ under the concept of STPF conditions can be considered as 'the break- through' allowing the solid sampling technique to become an alternative to the analysis of solutions in routine determinations. Although direct solid sampling has been widely docu- mented and reviewed no specific attention has been given to the introduction of slurries. The popularity of this technique as an alternative to direct solid sampling is reflected in the large number of papers published during the last decade (Fig. 1). In this review it is mainly commercially available atomizers that are discussed as laboratory-built atomizers have been extensively treated in previous review^.^-^ Direct solid sampling and slurry sample introduction are both reviewed independently so that the characteristic features of each approach can be easily overviewed; relevant applications are included.2 Direct Solid Sampling Early work reporting direct solid sampling showed that modifications to commercially available atomizers and the development of laboratory-built atomizers specifically de- signed for solid sampling were required. The main features of laboratory-built atomizers for direct solid sampling are (i) the use of larger dimensions than with the commercially available atomizers in order to facilitate sample introduc- tion thereby preventing the radiation beam from being blocked and avoiding a high background caused by the solid matrix; (ii) the separation of the volatilization and atomiza- tion of the analyte; (iii) the use of large sample masses in order to avoid errors caused by inhomogeneous distribution of the analyte; (iv) the lack of isothermality during atomiza- tion; and (v) the use of a low heating rate during atomiza- tion to avoid high background signals. The atomization of solid samples by using atomizers with larger dimensions than those used for liquid sampling was recommended by Langm~hr.~.~ Atomizers used for solid sampling can be heated inductively or resistively.Although resistively-heated furnaces are more common induction furnaces were designed to accommodate solid samples. Several manufacturers of electrothermal atomizers have implemented specifically designed atomizers for the direct analysis of solid samples.These designs include tube furnaces platforms cups and probes. Commercially avail- able atomizers are briefly described below. 2.1 Atomization Systems 2.1.1 Graphite tube atomizer Graphite tube atomizers are based on the Massmann design" (e.g. Perkin-Elmer and Varian). The solid sample can be introduced into the furnace either through a centrally located hole in the furnace or from the ends of the tube. Tube-type furnaces were widely used during the initial stage of solid sampling in ETAAS. In 1971 only two years after the introduction of commercial graphite furnace atomizers Kerbefl determined Au in polyester fibres by direct insertion of the sample. The problems associated with the use of this type of furnace include troublesome sample introduction; strong spectral interferences; and non- reproducible deposition of the solid sample on the wall.Accurate positioning of the solid sample is important in order to avoid non-reproducible vaporization as a result of the temperature gradient along the tube. 2.1.2 Cup cuvette atomizer In contrast to the Massman furnace the cup cuvette atomizer contains the sample in a cup-shaped furnace. This type of furnace is manufactured by Hitachi [Fig. 2(a)]. The main advantage of the cup cuvette atomizer is that the sample vaporization takes place slightly out of the absorp- tion volume so that blocking of the radiation beam as a result of the sample residue formed after atomization is avoided. In addition the cup cuvette atomizer gives reproducible sample deposition.Analytical applications using the cup cuvette atomizer have been reported by Takada and co-workers12-16 for the analysis of metals and alloys. Sensitivity can be improved by covering the cup cuvette with a graphite lid.'' The enhancement is due to an increased residence time of the analyte in the cup. 2.1.3 L'vov platform techniques L'vov et al.18 proposed the technique in which a platform made of pyrolytic graphite is inserted into the centre of a graphite tube. The sample is deposited on the platform. The platform is mainly heated by radiation from the hot tube wall. As a result a delay in sample vaporization takes place compared with the vaporization from the tube wall. Consequently the atomic vapour is formed when the gas temperature in the absorption volume has reached steady- state conditions. Isothermal atomization reduces both spectral and chemical interferences. Several electrothermal atomizers incorporate the platform technique for atomiza- tion.In all of them a graphite support of varying dimen- sions and relatively low mass is used. The temperature delay depends on the mass of the platform and the contact between the platform and the tube wall. Despite the extensive use of the platform technique for dissolved samples its application to solid sampling has been more restricted. Chakrabarti et d . 1 9 determined several metals in oyster tissue using wall and platform atomization and selective volatilization of the matrix. Although similar precision was found by using both techniques accuracy was better when the sample was atomized from a platform. Likewise platform atomization offered better sensitivity for volatile metals. Chakrabarti et made a comparison between wall platform and probe atomization for the direct determination of lead in bovine liver.Accurate results were only found with platform and probe atomization and by measuring the integrated absorbance. A special platform design has been reported by Brown et ~zl.,~' which facilitates the positioning of solid samples on the platform [Fig. 2(b)]. Kufurstz2 compared the different commercial atomiza- tion systems that are available for direct solid sampling. It was concluded that large graphite tubes are required in order to facilitate insertion of the sample and that L'vov platform conditions should be used for atomization.The manufacturers of graphite furnace atomizers supply a variety of graphite supports such as platforms boats and microboats which cause some delay in sample vaporization. However theoretical calculations indicate that the vapour phase temperatures reached with these types of atomizers are lower than for the L'vov platform atomizer and356 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 199 1 VOL. 6 (a) Tube / Carbon Fig. 2 (a) Cup cuvette atomizer; (b) platform-tube atomizer designed by Brown el aZ.;2* (c) L'vov platform atomizer; ( d ) boat atomizer; (e) microboat atomizer; U miniature cup atomizer; (g) cup-in-tube atomizer; ( h ) carbon rod atomizer; (i) probe atomizer; 0) ring chamber atomizer designed by Schmidt and Falk;64 and (k) second surface atomizer designed by Rettberg and H~lcornbe~~ therefore increased chemical interferences can be ex- p e ~ t e d .~ ~ J ~ A brief description of commercially available platform atomizers is presented below. 2. I .3.1 Platform atomization. The L'vov platform sys- tem is supplied with the atomizers commercially available from Perkin-Elmer [Fig. 2(c)]. The mass of the platform is about 75 mg. The insertion of the platform into the tube is axial. Grooved and ungrooved tubes can be used to accommodate the platform. However this type of support is unsuitable for weighing the solid sample because part of the sample can be lost during the insertion operation although this system can be successfully used when the solid sample is introduced through the sampling hole of the tube by means of a solid injector.24 2.1.3.2 Boat atomization.The 'boat technique' is used in the atomizers manufactured by Grin Analysengerate. The behaviour of this system approaches that of the L'vov platform [Fig. 2(d)]. The deposition area (7 x 4 mm) is open and bordered. The boat has a mass of about 130 mg and is transferred from the microbalance used to weigh the solid sample to the atomizer by means of a special claw. Insertion is axial and is attained using a jerk-free mechani- cal guide. Sample masses of from 0.2 to 30 mg can be introduced with this system.25 An automated solid sample analysis system based on the boat technique has been developed.26 The AAS instrument includes a powder sampler an integrated microbalance and a transport and handling system for the sample boat.The applications reported for this technique include the analysis of metals and alloys,27 fly ash,28 sewage sludge,29 biological matter,3o p~lyethylene,~~ filter matter,32 marine foodstuff^^^ and salivary calculi.33 2.1.3.3 Microboat atomization. This technique is avail- able in the atomizers manufactured by Thermo Jarrell Ash. The microboat is a 6 x 4 mm and has an approximate mass of 120 mg. The deposition area is open and bordered [Fig. 2(e)]. The microboat containing the solid sample is trans- ferred into the atomizer in a similar way to the boat technique by using a special claw. The microboat is inserted into the graphite tube through a narrow slit. The insertion is radial and there is a low risk of contamination. A disadvantage of the microboat is that the delay in atomization is considerably reduced as the whole microboat surface is in contact with the tube wall.The microboat technique has been compared with atomi- zation using induction furnaces for the analysis of metallic ~ a m p l e s ~ ~ - ~ ~ and glass materials.37 Induction furnaces can only be operated at a maximum temperature of about 2600 "C which is 300 "C lower than that with resistively-heated furnaces. Resistively-heated furnaces permit lower charac- teristic masses to be obtained than do induction furnaces. In contrast to resistively-heated furnaces induction fur- naces allow the use of larger sample masses and hence lower background absorbances are produced.36 2.1.4 Cup atomizer The first application of a cup for solid sampling was reported by Price et a1.38 The use of graphite cups inserted into graphite tubes permits not only the safe transfer of sample from the microbalance to the atomizer but also facilitates its weighing.The sample introduction system ensures a reproducible sample-to-sample deposition and in contrast to the cup-in-tube technique (see section 2.1 S) the deposition can be viewed. The use of a cup for atomization also avoids deterioration of the graphite tube because there is no build-up of deposit on the wall.38 The cost is also significantly lower than that of the graphite tube. The cup atomizer is used in the electrothermal atomizers manufactured by Hitachi [Fig. 2m]. These atomizers use a miniature cup (2.5 mm diameter; 45 mg mass). The cups are inserted into the graphite tube through an enlarged hole in the middle of the tube; again a special claw is used for insertion.The risk of contamination is low but the area of contact between the cup and the tube wall is large and therefore the 'L'vov platform effect' is reduced. TheJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 199 1 VOL. 6 357 reproducibility from cup-to-cup is sufficient to permit the use of one cup for each determination. The mass of sample that can be introduced ranges from 0.1 to 2 mg.39 Miniature cups used for solid sampling in the Hitachi atomizers are U- shaped in contrast to those used for dissolved samples. Atomization with graphite cups has been used for the determination of metals in particulate matter suspended in air,40 steel Cu and Cu alloys,41 stainless steel magnesite Mg-Zn alloys silica and Si-A1 alloys.38 Several applications for the miniature cup system have been reported such as the analysis of bovine plant matter,44 oyster and biological matter39 and marine organisms.46 The miniature cup has also been used to determine several metals preconcentrated from solution by coprecipitation with S-hydro~yquinoline,~~ dimethylglyoxime (DMG)-Ni- 1 -(2-pyridylazo)-2-naphthol (PAN)48 and magnesium 8- q ~ i n o l i n a t e .~ ~ The concentration factors achieved with preconcentration can be as high as 18 OoO-f~ld.~~ 2.1.5 Cup-in-tube technique The cup-in-tube atomizer for solid sampling is manufac- tured by Perkin-Elmer. In consists of a capsule that has an aperture of 3.5 x 4.2 mm and has a mass of approximately 180 mg [Fig.2(g)]. The capsule is radially inserted through a hole located in the middle of the graphite tube. There is some risk of contamination during the insertion or removal of the capsule from the tube. The delay produced with this system is significantly larger than that in wall or platform a t o m i z a t i ~ n . ~ ~ The reason for the pronounced delay is the larger mass of the capsule and the protection of the sample from radiational heating. The sample deposition is difficult to observe as the deposition area is closed. The gas above the sample is also shielded from radiational heating by the capsule walls. Gas and sample remain cold until the capsule is heated so that the heating mechanism of the sample is by conduction from the capsule wall.The characteristic masses obtained with the cup-in-tube technique are slightly worse than with the L’vov platf~rm.~’ Typically sample masses of 0.3-1.5 mg are i n t r o d ~ c e d . ~ ~ The applications of the cup-in-tube technique include the analyses of plants and plastic film,53 poly(viny1 chloride) (PVC) and bovine liver,51 coal coal fly ash and particulate matter,s2 biological matter54 and Ni-based alloys.5s 2.1.6 Carbon rod atomizer Carbon rod atomizers (CRAs) are open atomizers in which the sample is placed in a cup or tube which in turn is transversely heated by two graphite electrodes [Fig. 2(h)]. These atomizers manufactured by Varian were widely used for solid sampling e.g. for the determination of metals in particulate matter suspended in air.The air samples were filtered through cups or tubes made of porous graphite which subsequently were fitted in the atom- izer.56-58 A CRA can be operated under isothermal or non- isothermal atomization. Lundberg and F r e ~ h ~ ~ reported that isothermal atomization using a CRA with a cup offered better precision and accuracy for the determination of lead in steel and Ni-based . alloys. Lundberg60 indicated that integrated absorbance is preferable for measuring the atomic absorption signal even when the temperature of the carbon cup rises continuously during the atomization step and when using non-isothermal atomization. Precision was reported to be better when using integrated absorbance in comparison with peak height. 2.1.7 Graphite probe atomizer Probe techniques proved to be useful for isothermal atomization.A probe of graphite is inserted into a graphite tube that is pre-heated to atomization temperature [Fig. 2(i)]. This atomizer is commercially available from Philips. As the probe is primarily heated by radiation transfer from the hot tube atomization of the analyte takes place at a constant temperature of the gas phase. The probe technique has some advantages over the platform technique.20 In solid sampling the high matrix concentration can promote the removal of the atomic vapour from the atomizer as a result of the rapid increase in temperature. This can change the residence time of the analyte in the absorption volume. In probe techniques the temperature of the tube is constant when the sample is inserted so that this effect is minimized.More importantly probe atomization facilitates the dissoci- ation of molecular species. Probe atomization was used by Chakrabarti et aL61 for the determination of Pb Cd Ni Cu and Mn in airborne particulate matter by prior collection of the solid particles on the probe. Chakrabarti et uI.~O made a comparison between wall platform and probe atomiza- tion for the determination of lead in bovine liver. Reliable results were obtained with both probe and platform atomization. Probe atomization offers the advantage that drying and pyrolysis of the sample can be achieved outside the tube thereby eliminating some of the interferences caused by the matrix. Background signals caused by smoke condensation on cooler parts of the tube is also avoided.Schron et a1.62 used a graphite rod to insert solid geological samples into a graphite tube atomizer. Khammas et aI.(j3 reported the use of wall and probe atomization to deter- mine copper in milk powder. Wall atomization required the use of the standard additions method in order to obtain accurate results. In contrast however calibration using aqueous standards was feasible when probe atomization was used. 2.1.8 Ring chamber atomizer Schmidt and Falk64 designed a special graphite tube for direct solid sampling. The tube has rotational symmetry and a sample chamber that is isolated from the absorption volume [Fig. 20)]. The sample is accurately positioned owing to the tube configuration. Relatively large amounts of the solid sample (up to 10 mg) can be introduced.The ‘ring chamber tube’ does not have temperature gradients along the area where the sample is contained and it avoids blocking of the radiation beam. The metal vapour enters the absorption volume when the gas phase temperature is stabilized because of the small temperature gradient that occurs along the sample area and the short ramp time. This design was used to determine Au in gold wire for microelec- tronics and Cu and Ni in plant matter.64 2.1.9 Second surface atomizer Rettberg and H o l ~ o m b e ~ ~ achieved an effective separation between vaporization and excitation processes by condens- ing the atomic vapour on a cooled Ta surface (second surface atomizer). The system combines the use of a Ta plug onto which the analyte species condense and a tube-type furnace [Fig.2(k)]. Although the atomizer was originally used for liquid sampling,66 it is particularly useful for solid sampling because of the separation of the vaporization and excitation processes. The solid sample (0.7-4 mg) is weighed into a graphite cup using a microspatula and then the cup is placed in the tube and the Ta plug inserted. The tube is heated so that analyte vaporization and deposition on the Ta plug take place (transfer step). In contrast to the cup atomizer (section 2.1.4) the contact between the358 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1991 VOL. 6 sample cup and the tube is good and consequently the temperature delay between cup and wall is small. During the transfer step the Ta plug is cooled by a gas stream in order to promote analyte condensation.Once the atomic vapour is deposited on the Ta plug the cup containing the sample matrix is removed from the tube (cool-down step). The Ta plug is inserted the cooling gas stopped and the tube heated to a high temperature so that the analyte is re- vaporized from the Ta surface. The atomic absorption signal is independent of the matrix and the background signal is low as a result of the separation of the analyte and the matrix during the transfer step. Recoveries are in the range 8 1 - 127% for citrus leaves fly ash river sediment and filter paper.65 A second surface atomizer was used by Rettberg and H ~ l c o m b e ~ ~ to measure the time-dependent vaporization rate of an analyte from an individual sample over several heating cycles.The data were related to the total amout of analyte in the sample using aqueous standards. The method was used to determine Pb in tin copper and steel?’ 2.2 Accuracy and Precision Apart from the errors that can arise during the weighing of the solid sample the transfer operation and the sample introduction there are also errors associated with direct solid sampling which depend on particle size homogeneity of the sample sample mass and analyte concentration. 2.2.1 Influence of particle size Most of the materials require grinding in order to reduce the particle size of the sample to an appropriate size (e.g. 5 pm). Grinding of the material is also recommended for solid samples that are originally in powder form when the analyte is non-homogeneously distributed. Grinding im- proves the contact between the surface of the atomizer and the solid sample.The grinding operation is not always easy to perform (e.g. for metallic samples) and requires great care in order to avoid contamination. For refractory materials such as rocks and minerals vaporization of the analyte is often influenced by particle size. Improved accuracy and precision have been reported when the particle size is reduced. Eames and Matousek68 observed a decrease of the sampling error in the determina- tion of silver in quartzite rocks when the particle size was decreased by grinding to less than 10 pm. In addition accuracy can be improved by grinding as the analyte is more readily released. Stoeppler et al.69 concluded that small sample masses yield low concentration values when the sample is insuffici- ently ground.There is some controversy in the literature about the influence of grinding of the solid sample. In several types of samples (e.g. biological matter) particle size has no influence on accuracy but only on precision whereas for refractory materials (rocks minerals etc.) vaporization of the analyte is often influenced by particle size. Takada and HirokawaI2 concluded that the surface area of the sample particles influenced the vaporization of the analyte. Integrated absorbance was measured for re- peated atomizations of the same sample until complete vaporization of the analyte was achieved. Frech and Baxter70 found that increasing the mass of the sample did not always cause a proportional increase of the atomic absorption signal.Calibration with aqueous standards for the determination of A1 in biological material was feasible provided that the mass of the sample was lower than 2 mg. Precision is poor when the sample is introduced into the atomizer as a single particle. Takada and Koide” indicated this problem in the determination of Cu preconcentrated using an ion-exchange resin and using wall atomization. 2.2.2 Influence of sample homogeneity The influence of inhomogeneity is more critical when the sample mass decreases (typically sample masses of 0.1 - 10 mg are used). The lowest sample mass that can be handled reflects the difficulty of weighing and introducing the sample into the atomizer. Lundberg and F r e ~ h ~ ~ studied the trace element distributions for metallic samples and con- cluded that the influence of inhomogeneity on precision was negligible even working with sample masses as low as 2 mg.Siemer and Wei73 also found that the homogeneity did not affect the precision for sample masses of about 1 mg when determining Pb in rocks and glasses. On the other hand the relative homogeneity of solid samples can be estimated from precision measurements and sample mass.69v70 For materials that display non-homogeneous distribution of the analyte precision can be improved by grinding. For a large number of solid materials the influence of homogene- ity is not critical for sample masses of less than 200 mg. 2.2.3 Influence of sample mass The sample mass introduced into the atomizer is related to the analyte concentration in the original solid sample.The range of sample that is suitable is determined firstly by the availability of a microbalance to handle small sample masses and secondly from the capability of the atomizer to accept larger sample masses. Several workers suggested using larger graphite tubes than those commercially avail- able in order to introduce sample masses of up to 200 mg. The use of larger sample masses mainly improves the precision especially for non-homogeneous samples. The range of the sample mass that can be introduced is limited. Although sample masses of approximately 1 pg have been sample masses of lower than 0.1 mg are difficult to handle i.e. to weigh and transfer. When the concentration of the analyte of interest is too high for the analytically useful working range of the calibration graph either less sensitive l i n e ~ ~ ~ J ~ dilution of the sample with graphite4ss68*7s-78 or increase of the internal flow of ArS3 can be used.However these procedures can be a source of error. Vapour phase temperatures are lower when an internal gas flow is used and therefore matrix interferences are enhanced. 2.2.4 Influence of analyte location in the sample Different types of bonding between the analyte and the solid matrix in addition to different locations of the analyte in the solid particles (i.e. on the surface or occluded within the solid particles) can influence the absorbance measure- ment because of different vaporization kinetics. One example was the determination of several metals in steel.Signals with double peaks were observed as a consequence of the occurrence of the two locations of the analyte in the steel grain.13 The formation of several peaks in the atomic absorption signals was also observed for the determination of Pb in plasticss1 using the cup-in-tube technique and was attributed to the way the plastic decomposed when the temperature was increased. Several peaks were observed when lead was atomized at low pressure from copper alloys which were ascribed to different chemical forms of Pb or changes in its location.79 2.3 Chemical Modification Chemical modifiers are used in ETAAS for several pur- poses.*O The addition of a suitable reagent to the sample can change the relative volatility of the matrix or analyte so that no interference due to the matrix occurs during the atomization step.Stabilization of the analyte by a chemicalJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 199 1 VOL. 6 359 modifier avoids early losses that can occur during the pyrolysis and ramping steps of the atomization. Chemical modification for direct solid sampling is more complicated than for dissolved samples as the contact between the solid particles and the modifier is less effective. The addition of chemical modifiers has been used to diminish the volatility of the analyte to improve its release from the solid particles or to avoid the formation of double peaks. Chemical modifiers for direct solid sampling have been used in liquid and solid form. Chemical modifiers in liquid form are mostly used for the determination of highly volatile elements such as Pb Cd Se and As. The chemical modifiers most widely used are H3P0481 or H3P04 + (NH4)2HP0482 and HNO or HN03 + NH4H2P0421 for Pb; H3P0483 or NH4H2P04+HN0321~52 and H3P04 or Mg(N03)284 for Cd; HN03+Ni or HN03+Pd54 for Se; and H2S04+ HN03+Ni46 for As. Baxter and F r e ~ h ~ ~ reported that the use of Pd(NO& +Mg(N0,)2 rather than NH4H2P04 + Mg(N03)2 as a modifier for Pb and Cd determinations in solid samples gives superior results.In many instances a chemical modifier added in liquid form is ineffective for stabilization. de Kersabiec and BenedettiE4 carried out a study on the effect of chemical modifiers in liquid form including H3P04 Mg(NO& and Ni(N03)2 on the atomization of geological samples using the miniature cup and cup cuvette atomizers.Different samples were also atomized with the addition of chemical modifiers in the solid form. Double peaks were still obtained when liquid or solid modifiers were used. Graphite powder added to the solid sample is used to dilute the sample so that the mass of analyte atomized falls within the linear range of the calibration graph. In addition the graphite powder when mixed with the sample prevents the residue formed after atomization from fusing or sintering decreases chemical interferences and promotes the release of the analyte as the atomization occurs in a reducing environment.45~68.75-78,85,86 Moreover the lifetime of the atomizer is increased. Rapid deterioration of the surface of the atomizer was observed upon atomization of samples that contained large amounts of silicon.The silicon forms globules at high temperature which attack the pyrolytic graphite surface.68 The use of chemical modifiers in a solid form requires the modifier and sample to be ground and mixed so that the contact between them is good. Durnberger and co-work- e r ~ ~ ~ ~ ~ ~ found that Ni used as a finely divided powder was effective in stabilizing Se. It was essential that Ni was added in elemental form and in a 15-fold higher concentration compared with the amount used for liquid samples. 2.4 Sample Introduction Systems Four basic procedures have been used for sample introduc- tion in direct solid sampling. Firstly the introduction of a support on which the sample is weighed; once the sample is introduced into the atomizer the support is re-weighed.This procedure is used in most tube-type atomizers. The weighing and transfer operation requires care in order to minimize any contamination and possible sample losses. The Ta solid spoon commercially available from Perkin- Elmer belongs to this group of devices. The sample can be introduced through the dosing hole of the graphite tube provided the hole is enlarged.20 For solid samples consisting of ion-exchange resins used for preconcentration of metals from a solution a set of plastic tweezers and a small funnel can be used to introduce the resin particles into the tube.71 The second procedure a more advantageous approach is to use a graphite support for both weighing and atomiza- tion. Such systems are based on platforms boats micro- boats cups and probes as mentioned before.The use of inert supports made of ashless paper and cellulose acetate has also been attern~ted.,~ Inert supports tend to cause high blank values and moreover the residue formed after atomization can degrade the accuracy and reduce the lifetime of the atomizer. The third procedure involves the introduction of the sample by the use of special injectors. In order to facilitate the sample introduction through the sampling hole of the graphite tube solid sample injectors operated in a similar way to conventional autosamplers have been described. Grobenski et al.24 developed an injector to introduce fiiiely powdered samples into a tube-type atomizer. The injector works like a microsyringe and sample masses ranging from 0.5 to 5 mg were introduced.A different injector used for direct solid sampling was described by Kurfiurst et The reproducibility of the results is comparable to those obtained for manual solid sample introduction. The final procedure is the use of special sampling devices to collect the solid sample by filtration through porous graphite or electrostatic accumulation. These special sampl- ing devices are useful for solid materials suspended in gases.56-58,81-83 Collection on filters made of different ma- terials (ie. cellulose acetate or nitrate and PVC) has also been 2.5 Sample Preparation Sample treatment for solid sampling should be kept to a minimum in order to avoid possible contamination or loss of sample; although grinding sieving homogenization drying etc.are often required. Because of the influence of particle size on accuracy and precision grinding of the sample is carried out in order to convert the sample into powder form or to reduce the particle size. The grinding can be performed using agate mortars or vibrational pulveriz- ers. Grinding is suitable for brittle materials such as rocks minerals glass soils etc. Biological and vegetal materials can be converted into a solid that can be ground i.e. by drying dry ashing plasma ashing or Iyophilization. Grind- ing metallic samples is more troublesome and can cause contamination of the sample. 2.6 Calibration The solid matrix can strongly influence the shape of the signal obtained during atomization. This influence makes it troublesome to use peak height to quantify the atomic absorption signal.The fast transient signals require fast instrumental response and preferably integrated absor- bance meas~rements.~~ Calibration can be performed using several approaches (i) with CRMs; (ii) with synthetic solid standards; (iii) using the standard additions method; or ( i v ) with aqueous standards. The first approach requires appropriate CRMs to be available at the concentration level of the analyte of interest. This is not a general calibration method owing to the lack of reference materials with certified analyte concentrations covering the range of interest (e.g. for environmental samples). However calibration with CRMs has been used for solid sampling of metallic samples,92 sewage sludge,26 biological samples93 and plant It should be noted that a significant disadvantage of reference materials is that the certified concentration generally has an uncertainty which is large when compared with the measurement uncertainty of the method.Conse- quently the use of CRMs as calibrants will introduce a large and often ignored uncertainty in the final value. In some instances when CRMs are not available a synthetic solid sample that matches the composition of the original sample can be used.360 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 199 1 VOL. 6 The standard additions method has been used to suppress the influence produced by the bulk composition of the solid sample. This method requires that the analyte contained in the solid sample and the added analyte are equally affected by the matrix.Three different approaches are possible for the standard additions method in combination with solid ~ampling.*~*~~ In the first approach increasing amounts of analyte are added to constant masses of solid sample. This is difficult to achieve as it requires weighing identical amounts of solid sample for each individual addition. The second approach involves the use of increasing masses of solid sample with the addition of constant amounts of analyte from a standard solution. In this approach only one standard solution is used and one addition is made and therefore the final result can be inaccurate. The third approach means varying both the solid sample mass and the standard additions amount. The absorbance signal is a function of two independent variables and can be extracted by multiple linear regression.The use of aqueous ~ t a n d a r d s ~ ~ - ~ ~ ~ ~ ~ requires that the absorption signal is independent of the bulk matrix and only dependent on the concentration of analyte. This requirement is usually not fulfilled because of both the spectral and chemical interferences caused by the matrix. For calibration with aqueous standards to be successful a set of instrumental requirements has to be met i.e. rapid signal processing integrated absorbance measurements Zeeman background correction and atomization under isothermal conditions. For solid samples the absolute calibration of the autosampler or pipette used to dispense the standard solutions is also important because the accuracy of the determinations is directly dependent on the accuracy of the pipette. 2.7 Applications of Direct Solid Sampling The applications of direct solid sampling in ETAAS are shown in Table 1.Applications are classified alphabetically with respect to the matrix. Information on the elements that are determined the type of atomizer used and the analytical performance are also given. Attention has been mainly focused on commercially available atomizers. 3 Slurry Sample Introduction The introduction of a suspension of the finely powdered sample (slurry) combines both the advantages of liquid and solid sampling.166 This technique was first used by Brady Table 1 Applications of direct solid sampling in ETAAS Sample Aerosols Airborne particulate matter Algal cells Alloys Alumina and synthetic Aluminium corundum Animal serum Biological matter Biological matter Element Pb Pb Mn Pb Fe Ni and Cr c u Ga Cr Cr Cr Co and Ni Pb and Ni Pb Pb and Cd Pb and Cd Hg Se c u As Cd Zn Pb Ba La and Mg Mn Cu and Cr A1 Mn TI Cd Pb Co Ni and Mo Atomizer Boat - Graphite tubet CRA (cup) Ta boat - Cup atomizer Graphite tube Graphite tube Graphite tube Graphite tube and PTS PT Boat Boat Boat Cup-in-tube Microboat and graphite tube Miniature cup CRA IC cuvetteg CRA Cup cuvette Cup atomizer Comments Atmospheric Pb is collected by membrane filters.RSD= 11-37% DL*=0.4 pg of Pb per filter-paper Freeze-dried sample RSD= 7% RSD = 20% RSD=2.6-16'41 RSD= 5.1-10.7% DL=0.18 ng DL=22.5 pg mg-l Lyophilysed sample. RSD=9-22Oh RSD= 5- 16% Introduction of 1 pg of tissue HN03 and NH4H2P04 as chemical modifier Study of homogeneity Air-dried and freeze- dried samples DL=O.15 pg g-l RSD= 9% for Ba; 1 7% for La; and 8% for Mg RSD= 10% for bovine liver and animal muscle Freeze-dried samples. - Coprecipitation in the cup with ammonium pyrrolidinedi thiocarbamatef Reference 95 90 96 59 97 85 98 99 100 101 74 21 102 103 30 54 104 39 105 70 106 107 108JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 199 1 VOL. 6 36 1 Table 1-continued Sample Element Atomizer c u Miniature cup Comments Three solid standards (NET(/) are used for Calibration coprecipitates with magnesium oxinate and DMG-Ni-PAN Pb and Cd profiles are determined in birds' feathers - Solid sampling and wet digestion are compared Deuterium and Smith- Hieftje background correction are compared Reference 93 109 110 1 1 1 112 113 24 114 44 49 115 94 116 117 Pb and Cd Boat A1 Cup-in-tube Cr Pb Cd Boat Cd Boat Ni and Hg Biological and vegetal matter P Graphite tube Pb Rb Cd Cu PT Ni and Mn Nine elements Boat Pb Cd Mn Cu Zn and Cr Al Cd Co Cu Mn Ni Pb and Zn Miniature cup Miniature cup Peak area.Special injector is described Coprecipitation with magnesium 8-quinolinate. Multi-element analysis - Generalized standard additions method for calibration Automated introduction - CRA Al Ag Ni and Se - Pb Cd Zn Mn and Se Cup atomizer Probe technique c u Ni and Ag chemical modifiers are attempted. For Se calibration with solid standards is required also freeze-dried samples Direct introduction and 118 Pb and Sb - Biological matter from a river ecosystem Cu Pb Cd and Zn Boat Biological matter and marine organisms As Miniature CUD 119 RSD= 3- 10% Standard additions method RSD = 5.5- 13%.Aqueous for calibration standards for calibration RSD = 9.6% RSD = 6.3-7.3% DL=5.1 pg graphite. RSD = 4.9% additions method for calibration Ni Calibration with aqueous standards is well suited for Pb and Zn but not for Mn. RSD= 514%. Pipette method for sampling Sample mixed with RSD = 10%. Standard RSD = 10% for Cd; 10-20% for Dilution with graphite. 46 120 Bovine liver Cu Zn Pb Co Fe and Cd Pb Graphite tube Graphite tube. and PT PT and probe 20 Cd Pb Miniature cup Miniature cup 42 43 Calcium carbonate Coal cu Miniature cup 121 Be Graphite tube 122 Cd and Ni Pb Zn and Mn Cup-in-tube 52 123 Cup cuvette Coal and particulate matter Coal and petroleum coke Copper Se Boat DL=O.l pg g-I RSD = 7-2 1 '/o Standards are prepared by quantitative doping of high-purity Cu 77 Cu Ni and V Ag Boat Graphite tube 124 92 Fibre and plastic paper Filter material Fly ash Freshwater mussels Cu Fe Mn and Si Cd Boat TI Boat A] Cr Cu Pb and Graphite tube Graphite tube Zn 125 126 28 I27 - - - Freeze-dried samples362 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1991 VOL.6 Table 1-continued Sample Gallium arsenide Geological samples Glass Graphite Hair Hair Inorganic samples Ion-exchange beads Malt and yeast Margarine Marine foodstuffs Metal and alloys Metals (Ta Fe and Metals and oxides Nickel-based alloys Fe-W) Oyster tissue Paper Particulate matter Pancreatic tissue Pancreatic tissue Pancreatic tissue Petroleum Phosphorized Cu alloy Pig liver Plankton Plant and biological matter Plant matter and oyster mussels Element Cr Cr As Cd Pb Hg Sb and Se Ag Pb and Bi Pb Pb 13 elements Cu Ag Coy Ni and Pb As Cd Cr Hg and Pb Pb and Cd T1 Ni Cd Pb and Hg Bi Pb Ag and Te Ca Several Pb Bi Se and Te As Sb Se and Te Cd In and Zn Pb T1 Bi Te Se and Pb Cd Pb Zn Cu Co and Fe Cd Hg Cu and Cr 10 elements Pb Be Cd Se Ag Hg Pb Hg Pb Be and Cd Cd Hg and Cu Ca Cd Cd Be Pb Pb Cd and Zn Cu and Cr Pb and Cd Cr Atomizer Comments Graphite tube - miniature cup and PT study - - Cup cuvette Chemical modifier Microboat and Cup cuvette induction furnace Graphite tube Graphite tube Boat - Cup-in-tube Boat Boat Microboat and induction furnace - Boat Graphite tube and Microboat and CRA induction furnace Microboat and induction furnace Cup-in-tube Graphite tube Boat CRA CRA (cup) CRA (CUP) CRA (CUP) CRA (tube) CRA (tube) Cup atomizer Boat Graphite tube CRA Graphite tube and CRA Graphite tube Boat Graphite tube Cup-in-tube - Miniature cup Dilution with graphite. DL=0.03 ng of Pb DL=O.Ol pg g-' Dry ashing followed by atomization Silk and animal hair are used as solid standards - Signal summation method is used to improve the signal-to-noise ratio.Calibration with aqueous standards is feasible - Sample is a chip of the metal DL= 50 pg of Ca - RSD=7-25'h DL=O.l 0.1 0.06 and 0.003 pg g-' for As Sb Se and Te respectively RSD= 12% for Zn 8% for Cd and 7Oh for In. DL=10 2 and 10 ng g-l for Zn Cd and In respectively Calibration with standard additions method.DL=0.5 pg g-' of Pb RSD = 4.9Oh Calibration with aqueous standards and use of chemical modifiers. RSD=6-14% Peak height. Standard additions method for calibration - - RSD=4.2Oh RSD = 1 5% - RSD = 7- 1 4Oh RSD = 6Oh Collection on filter Collection on filter. RSD = 1 Ooh RSD = 14.8Oh - - DL= 1 ng g-I of Be (HGA) and Atomizer is modified 10 ng g-I of Be (CRA) for vacuum atomization - Pd +magnesium nitrate is more suitable than NH4H2P04+ magnesium nitrate DL=7 pg RSD=6-9% Reference 128 129 84 37 130 131 132 133 76 134 135 136 25 34 137 27 138 35 36 139 55 19 140 141 81 56 83 57 58 40 32 142 143 144 145 79 146 147 23 45JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 199 1 VOL. 6 363 Table 1 -continued Sample Element Atomizer Comments Reference 50 148 53 51 8 31 149 150 151 152 71 153 154 82 73 38 155 33 47 86 28 26 68 75 156 157 13 12 60 158 159 160 72 41 15 16 17 14 161 77 162 163 Plant soil As Cd Cr Cu Pb and Mn Pb Cu Cd Mn and Rb Cup-in-tube CRA Cup-in-tube Cup-in-tube Graphite tube Boat Graphite tube CRA Boat RSD = 1 0% Plastic RSD= 15% RSD = 3.1 - 1 2% Plastic film PVC Polyester fibre Polyethylene Polymers PVC air filter Rat placenta and bovine liver Cr Pb and Cu Au Pb Cu Cr and Ni Fe Cu and Cr Be Cd RSD = 4-6% DL=0.02 pg g-' - Freeze-dried sample.DL=0.005 ng Comparison of microwave digestion and direct solid sampling. Samples were diluted with urea RSD=4.1% RSD= 10-20% RSD= 10-20% O2 ashing Red mud Cd Cr Pb Ag and PT Au Resin (cation-exchange) Cu Graphite tube Rocks and sediments Rock samples Cd Graphite tube Au and Ag CRA Pb PT Rocks glasses and Rocks and alloys Sediment soil fly ash leaves and water Pb CRA (cup) Pb Na Fe and Ti Cup atomizer RSD = 8% Pb Cd and Cu Boat Digestion is recommended for soil and solid sampling sediments and leaves Salivary calculi Sea-water Cd Pb and Zn Boat As Cd Zn Pb and Miniature cup Zn I Preconcentration with 8-hydroxyquinoline.RSD=6-14% Several NIST and BCR standards Pb and Cd Cup-in-tube Pb Cd and Hg Boat Cd Boat Ag CRA (cup) and boat Study of homogeneity Sewage sludge - Automated system Dilution of sample with graphite. RSD=8.l% Dilution with graphite Aqueous standards for calibration O2 ashing Study on double peaks DL=0.1 0.04 and 0.008 ng of Cu Mn and Ag respectively - RSD = 3.840% Silicate rocks Siliceous materials Ca Mg and Fe Ta boat Silicon Pb Sb and Mn Microboat Spinach Steel Rb Graphite tube Pb Bi Ag and Zn Cu Mn Ag and Pb Cup cuvette Cup cuvette Pb CRA Ag Bi Cd and Zn Pb and Bi CRA CRA (cup) Steel and Ni-based alloy Steel Fe-Mo and Ni-based alloy Steel Cu and Cu alloy Tin Pb Bi Zn and Ag Microboat RSD = 6% Pb and Sb Graphite tube Pb Cup atomizer Zn and Bi Cup cuvette Pb and Cd Cup cuvette DL=34 ppb RSD = 3.6-9% for Pb and Sensitivity is RSD = 9-43% 5- 12% for Cd enhanced by covering the cup with a lid RSD= 6-28Oh RSD = 7-9% RSD= 7- 12Oh Pb Cup cuvette Tin ingot Uranium oxide (U,O,) Cu and Ag Cup cuvette Cd and Li CRA Co Cr Cu Mn and CRA Ag Ca K Li Mg CRA c o Miniature cup Ni Na Pb Sn and Zn Vitamin B12 Calibration with a synthetic reference material.DL= 0.15 ng ml-1 of Co or 4 ng mg-I of vitamin B12364 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1 99 1 VOL.6 Table 1-continued Sample Element 'Atomizer Comments Reference Water co Miniature cup Sample was mixed with 164 20 mg of Mg and 2% 8-hydroxyquinoline (2-5 ml) for preconcentration Water borne Water samples Cd Cu Mn Pb and Miniature cup Coprecipitation Wheat particulate matter Cd Pb and Cu Boat - Zn with DMG-Ni-PAN Cd Boat - * DL detection limit. f Graphite tube heated graphite atomizer (HGA) and graphite tube atomizer (GTA). $ PT atomization with L'vov platform. 9 IC cuvette integrated contact cuvette. fi IUPAC name ammonium pyrrolidin- 1 -yldithioformate. )I NIST National Institute of Standards and Technology [formerly known as NBS (National Bureau of Standards)]. 89 48 165 and c o - ~ o r k e r s ~ ~ ~ J ~ * to determine Zn and Pb in leaves and marine sediments.The slurry technique overcomes some of the problems associated with direct solid sampling as it permits sample introduction using micropipettes and auto- samplers which are routinely used in liquid sampling. Another remarkable feature of slurry introduction is that the same atomizers that are used for liquid sampling can be successfully used for the atomization of slurries. Most of the work done in slurry sampling has been carried out using wall atomization and platform atomization with tube-type atomizers. The most critical factor in the slurry technique is probably the need for maintaining a stable slurry during the time required for sample introduction. So far homogeniza- tion procedures include the use of stabilizing agents magnetic and ultrasonic agitation vortex mixing gas mixing and pre-digestion of the slurry.The particle size of the solid sample can have a strong influence on accuracy and precision. Successful slurry analysis requires the char- acterization of (z] the sample homogeneity; (ii) the grinding and sieving; (iiz] the preparation of the slurry (i.e. the liquid medium by the addition of chemical modifiers and wetting agents and antifoam agents); (iv) the mixing and homogeni- zation of the slurry (i.e. the use of stabilizing agents or agitation systems); (v) the extraction of analyte into the slurry medium (i.e. influence of the acidic diluent and efficiency of the agitation system in achieving extraction of the analyte); (vi) the influence of particle size on deposition of the slurry sub-sample (i.e. the concentration range of the slurry required to obtain reproducible and representative sampling; and (vii) the influence of particle size on atomization efficiency (i.e.the determination of the rela- tive atomization efficiency for the slurry sample and the possibility of using aqueous standards for calibration). 3.1 Preparation and Mixing of Slurries Preparation of the slurry consists of adding a liquid diluent to the solid material that has been previously ground and sieved (when necessary) weighed and placed into a con- tainer in which the slurry is stable during the time required for sampling. The amount of solid material that is weighed depends on the concentration of the analyte and the dilution in the final volume of the slurry.For samples of low homogeneity the precision is improved when larger amounts of sample are used to make the slurry. The introduction of the slurry sub-sample into the atomizer can be carried out manually or by using an autosampler. When autosampler cups are used in the weighing of the solid material the maximum volume of slurry is limited to the cup volume (about 2 ml) which in turn limits the maximum amount of material that can be suspended. Once the slurry is prepared the solid sample must be equally distributed in the volume of the liquid. The different approaches available to maintain a stable and homogeneous slurry are discussed. 3.1.1 Use of stabilizing agents Slurry preparation in aqueous solution is rarely suitable because most powdered materials undergo rapid sedimen- tation.Sedimentation of the suspended material usually occurs after mixing the slurry and can be quantified by using Stoke's law. The sedimentation rate depends on the densities of the diluent and solid material the viscosity of the diluent medium and the radius of the sample particles. The slurry can be stabilized using a highly viscous liquid medium. So far gly~erol,l~*J~~ non-ionic surfactants' 74 and organic solvents of high have been used as slurry stabilizing agents. The stabilization capability of these reagents largely depends on sample characteristics and particle size. As pointed out by Majidi and H o l ~ o m b e ~ ~ ~ the time interval between complete mixing of the slurry and the removal of an aliquot for analysis can be increased in a highly viscous medium with a density similar to that of the particles.Fuller and Thompson169 used a thixotropic thickening agent to stabilize slurries prepared from rock samples. The procedure consisted of the addition of sodium hexameta- phosphate as wetting agent and ammonia to neutralize the medium. The slurry was stable for several days. Stabilizing agents are useful when an autosampler is used to introduce the sample as the slurry can be left in the sampling cups without further homogenization. 17* Little- john et all7' observed that the stabilization of the slurry was dependent on both the particle size and the concentra- tion of the stabilizing agent. Sample deposition was difficult owing to the high viscosity of the medium when the Viscalex concentration was greater than 2-3Oh.On the other hand the maximum slurry concentration to be used depends on the concentration of the stabilizing agent. Stephen et ~ ~ 1 . l ~ ~ found that the maximum slurry concentration was 10% when a stabilizing agent concentra- tion of the order of 3Oh was used for slurries prepared from a spinach sample (50 pm particle size). Similar behaviour was observed when glycerol was used as the stabilizing agent. Problems are reported with respect to the use of viscous media in slurry preparation. The sample aliquot is ineffici- ently pipetted into the atomizer when the concentration ofJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY. AUGUST 1991 VOL. 6 365 the stabilizing agent is high.The sample can remain around the dosing hole which can degrade the precision. This problem has been observed using mi~ropipettesl~~ and autosamplers.166 The use of an autosampler also requires a cleaning procedure for the capillary between each of the successive determination~.'~~J~~ It is also necessary to add an additional pyrolysis step in order to remove the excess of stabilizing agent. 169~172 Stabilizing agents are usually ineffici- ent when the slurry contains particles of high density.177 It has been reported that glycerol causes interference for Cd and Pb even under isothermal atomization condi- t i o n ~ . ' ~ ~ Owing to the carbon build-up formed on pyrolys- ing the glycerol both Cd and Pb absorption signals shifted to lower temperatures which caused non-isothermal atomi- zation.These problems can be minimized by cooling the slurry to 5 "C and using the smallest amount of glycerol possible in order to reduce the interference. Apart from stabilizing agents the addition of wetting and antifoaming compounds can improve the dispersion of the slurry. The application of wetting and dispersive additives was discussed by Tsalev et a1.80 in an extensive review on chemical modification in ETAAS. In slurry analysis Triton X-100 is useful to disperse solid particles that might otherwise tend to float on the top of the l i q ~ i d . ' ~ ~ J ~ ~ Viscous samples containing solid particles (for example petroleum products) can be sampled as a slurry. Dilution with an organic solvent is required to reduce the viscosity so that sample introduction is facilitated.l g 0 3. I .2 Magnetic agitation and vortex mixing Magnetic agitation and vortex mixing have been widely used mainly during the early work carried out using the slurry technique. Slurries are prepared in a beaker and the slurry is stirred in order to achieve a homogeneous distribution of the solid material (usually for 3-5 min). The stirring action is stopped and a sample aliquot is withdrawn and introduced into the atornizer.lg1 The sample aliquot can also be withdrawn under continuous stirring.lg2 The effectiveness of magnetic stirring and vortex mixing basically depends on the sedimentation rate of the sus- pended material. Hinds et allg3 reported that differential sedimentation of sample particles produced during work with soil samples contributed to an incomplete recovery of Pb.The error associated with sedimentation is a function of both particle size and the presence of particles having different composition. As a result large errors can be expected when the analyte is predominantly distributed in particles of high density which undergo rapid sedimentation. Miller-Ihli 73 compared vortex mixing with ultrasonic agitation systems for the homogenization of slurries for ETAAS. The precision achieved with these systems was not significantly different. However vortex mixing caused lower recoveries of Fe. After magnetic agitation or vortex mixing slurries are usually introduced with a micropipette because of the difficulty of incorporating these systems in an autosampler. Lynch and LittlejohnIs4 reported a miniature magnetic stirring device that could be used in combination with an autosampler.Slurries were homogenized in the autosampler cups using small magnetic bars coated with polytetrafluoro- ethylene (PTFE). The system showed good performance for the determination of Pb in freeze-dried samples. 3.1.3 Ultrasonic agitation Ultrasonic agitation has been used as an effective system to homogenize slurries for ETAAS. Ultrasonic agitation can be used in combination with both manual and automated introduction of the slurry. An advantage of this system in comparison with magnetic agitation and vortex mixing is that the analyte of interest is partly extracted into the liquid phase owing to the ultrasonic action when the slurries are prepared in an acidic medium.Ultrasonic agitation is more effective than other agitation systems. This is important when the analyte is predominantly located in the larger sample particles that undergo rapid sedimentation. Miller- Ihli173 reported obtaining improved recoveries for the determination of Fe in wheat flour and bovine liver when the sample was sonicated. Agitation with a small ultrasonic probe also permits the slurry to be prepared directly in the sample cups of the autosampler. van Loenen and reported the modifi- cation of an autosampler in order to use ultrasonic agitation of a slurry prepared from coal fly ash samples. They reported that ultrasonic stirring could not maintain a homogeneous slurry during the time required for the autosampler to withdraw the sample aliquot unless com- plete analyte extraction was achieved.Preparation of the slurry outside the autosampler cup permits work with larger sample masses than those used when the slurry is prepared directly in the sample cup (200-300 mg versus 1-2 mg) which facilitates the work with less homogeneous samples. The ultrasonic treatment of slurries requires special care in order to avoid contamination when the titanium probe touches the wall of the cup. The leaching of metallic impurities from the cup wall has been observed with polyethylene cups but not with PTFE cups.173 Miller-Ihli18s automated the system combining ultra- sonic agitation and sample introduction with the autosam- pler. A similar design has been reported by Carnrick et and is commercially available from Perkin-Elmer.Auto- matic slurry sample introduction in combination with ultrasonic agitation has been used by Epstein et to study the sources of variability in the measurement. It was concluded that the precision depends on both the extraction yield of the analyte and the homogeneity of the sample. Precision of 2-1Oo/o was obtained by Bradshaw and Sla- vinIB8 when determining Se Pb and T1 in coal and coal fly ash. The time required for each determination was less than 1 min by using ultrasonic agitation of the slurry and elimination of the pyrolysis step. This method was also used to determine Mn.IB9 3.1.4 Gas mixing of slurry Bendicho and de Lo~s-Vollebregt~~~ reported a procedure in which effective homogenization of the slurry was achieved by passing an Ar stream through a narrow capillary tube introduced into the slurry medium.This system allows sample preparation directly in the cups. The homogeniza- tion requires a 'bubbling time' with the Ar stream of only 30 s. The precision obtained for the determination of Cu Co Cr Mn Fe and Ni in different samples of glass was about 6%. This system is easy to handle and it does not require the use of stabilizing agents or special devices for agitation. However the effectiveness of mixing largely depends on both the particle size and the characteristics of the solid sample. 3.1.5 Pre-digestion of slurry Pre-treatment of the slurry can be helpful to extract the analyte of interest into the liquid phase. In contrast to the extensive sample pre-treatment inherent in fusion and digestion procedures that are conventionally used pre- digestion of a slurry only requires a partial decomposition of sample hence it is not time-consuming.Fagioli and co- w o r k e r ~ ' ~ ~ - ' ~ ~ were the first to use partial wet oxidation of several biological and vegetal materials by concentrated sulphuric acid and subsequent analysis of the carbonaceous366 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 199 1 VOL. 6 slurry thus formed. This method is much faster and more reliable than those based on dry ashing or extensive wet oxidation. Slurry pre-digestion can also be accomplished in the sample cups of an autosampler. Hoenig et ~ 1 . l ~ ~ reported the difficulties of stabilizing slurries containing large par- ticles in the analysis of sediments and particulate matter. In order to solve these problems pre-digestion was performed by adding a small amount of concentrated nitric acid to the sampling cups containing the slurry. Partial extraction of the analytes into the liquid part of the slurry was obtained.The pre-digestion step significantly improved both the precision and accuracy. A rapid pre-digestion procedure has been proposed by Bendicho and de L~os-Vollebregt~~~ for the determination of several metals in glass. Pre-digestion of the slurry was based on the addition of hydrofluoric acid at a low concentration (3%) and homogenization by gas mixing.177 The pre-digestion step is accomplished in the sample cups of an autosampler 3 min being sufficient to achieve an effective extraction of the analytes from the powdered glass (15 pm particle diameter).This approach gives 2-fold improvement of precision for the determina- tion of Cu Co Cr Mn Fe and Ni in comparison with that obtained without pre-digestion. Cold extraction using di- lute hydrofluoric acid is found to be useful for those materials in which the analyte is predominantly contained within the siliceous fraction. 3.1.6 Electric dispersion in a condensed medium The electrical dispersion of metallic samples in a condensed medium gives rise to a colloidal suspension which is suitable for the slurry technique. Pchelkin and co-work- ers19591% studied in detail the characteristics of the sols formed upon spark ablation. The colloidal suspension contains metallic particles with particle sizes of less than 1 pm and displays high stability.Using this method slurries of metallic materials are obtained without the need for machining in order to convert the sample into a powder and therefore there is a reduced risk of contamination. The spark ablation can be performed in aqueous and organic solvents. This method is attractive for metallic samples that are difficult to dissolve by conventional acid digestion. 197 3.2 Influence of Particle Size The particle size of the solid material used to make a slurry can influence the stabilization deposition and atomization efficiency of the slurry which in turn can influence both accuracy and precision. Fuller et reported that the precision achieved for slurry sample introduction was largely influenced by particle size although this was not a critical factor in achieving effective atomization.Relative atomization efficiencies of about 100% were obtained for the determination of several metals in titanium ores and silicate rocks provided that the diameter of the solid particles was less than 25 pm. Under those conditions calibration could be carried out using aqueous standards. The decreased influence of particle size on atomization is an advantage of ETAAS over other atomic spectrometric techniques used in combination with slurry sampling such as flame AAS and inductively coupled plasma atomic emission spectrometry. Several workers have reported quantitative recovery of the analyte for particles of signifi- cantly larger diameter (25 pm). Hinds et ~ 1 . ~ ~ ~ reported incomplete recovery for lead when determined in soil; owing to the slower vaporization of the larger particles Pb is inefficiently atomized within the observation time.Pipetting efficiency was only 80% when the particle size was larger than 50 pm. Inefficient atomization was observed unless the particle size was between 50 and 2 pm. Karwowska and Jackson199 used solid samples with particle sizes of less than 20 pm during their atomization studies of Pb from a synthetic soil matrix to obtain better precision. In contrast Miller-Ihli173 pointed out that the precision achieved was better for particles in the range 250-600 pm in comparison with the precision achieved with small particles. Stephen et aL2O0 found improved precision in the analysis of liver tissue when the finest fraction of the sample (particle diameter less than 21 pm) was used.Chen and Xuzo1 established that particle size should be lower than 76 pm in order to obtain quantitative recoveries when deter- mining trace metals in aluminium oxide. It is obvious that smaller particles facilitate sample preparation and improve recovery. Errors associated with large particles (diameter > 100 pm) arise from the difficulty of maintaining a homogeneous distribution of the large particles in suspension and the lower pipetting efficiency for large particles. Grinding of the original sample is usually required in order to minimize the errors. The optimum range for particle size depends on sample composition. 3.3 Influence of Slurry Concentration Another important factor in the slurry technique is the slurry concentration.Samples of high analyte content can be analysed more readily by the slurry technique than by direct solid sampling as the slurry can be easily diluted. However dilution of the slurry can only be carried out within a limited range; precision is degraded when working with highly diluted slurries because only a small number of particles remains in the slurry. On the other hand when the analyte content of the original sample is very low the concentration of the slurry can be increased accordingly although pipetting efficiency can deteriorate if slurries are more concentrated. Another factor to be taken into account is the increase in the matrix effects that can arise when the slurry concentration is increased. Lynch and L i t t l e j ~ h n ~ ~ ~ established an optimum slurry concentration range for the analysis of food.Precision was less than 4% when working with particles of less than 50 pm in diameter provided that the concentration of slurry was below 10% m/v. The accuracy deteriorated when the slurry concentration was above 5% as a result of the excess of matrix. Slurry concentrations higher than 5% caused inefficient deposition of the slurry aliquot. Holcombe and MajidiZoZ characterized the errors associ- ated with slurry sampling. Errors can arise from uncertain- ties in the sample volume the number of particles in the sample volume and the variation in the mass of the individual particles. It was concluded that errors can be minimized by working with small particles concentrated slurries and narrow particle size distribution.3.4 Chemical Modification Slurry sampling in ETAAS allows the use of chemical modifiers. The interaction between the chemical modifier and the particles of the solid sample is closer than for direct solid sampling. Most of the work on chemical modification for the slurry technique has been carried out in order to stabilize highly volatile elements such as Pb and Cd. 172,199.203-208 Karwowska and J ~ c ~ s o ~ ~ ~ ~ * ~ ~ ~ studied the release mecha- nism of Pb from alumina particles as a function of the location of analyte in the particles i.e. adsorbed or occluded. A mixture of (NH4)2HP04 and Mg(N03) was recommended as a chemical modifier and successful calibration with aqueous standards was reported.The assumption was made that in soil samples Pb is adsorbed onto the particles through covalent bonding with 0 and OH. However soil is a more complex matrix than aluminaJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 199 1 VOL. 6 367 as it also contains organic material. MareEek and Synekfo9 studied the effect of the Ga occluded or adsorbed in alumina samples. The increase in peak height of the signal was attributed to the delay in the atomization caused by the alumina particles. Hinds and Jacksonzo5 used a mixture of montmorillonite and humic acid to simulate the behaviour of a soil matrix. A chemical modification study was carried out for Pb. Among the modifiers studied the best perform- ance was obtained with a mixture of magnesium and palladium nitrates.This chemical modifier leads to similar appearance times for the atomic absorption signal of slumes and aqueous standards. Palladium is efficiently reduced by the organic matter present in the Other procedures attempted for chemical modification in the determination of Pb in soil samples such as oxygen ashing pre-slurry ashing in a muflle furnace and the addition of nitric acid are less suitable as the peak characteristics differ significantly from those obtained with aqueous stan- d a r d ~ . ~ ~ ~ Hoenig and van Hoey~eghenl~~ used a mixture of (NH4)zHP04 Mg(N03)z and HN03 as chemical modifier for the determination of Pb and Cd in plant tissue sediment and lyophilized animal material; (NH4)zHP04 has also been used for the determination of Cd in olive leaves and mussels.zo3 For As Ni(N03)2178 and a mixture of Ni(N03)2 Mg(N03)z and HN03210 have been used; Ni(N03)z has also been used as a modifier for Se in selenized and in Ebdon and Parryz1 found that Ni(N03)z gave better performance than CU(NO~)~ or Pd(N03)2 as chemical modifiers for the determination of Se in coal.Graphite can be used for chemical modification in the slurry technique by adding fine graphite powder to the slurry. Carbon black has been used as a chemical modifier for the determination of Be and Mo in beryllium oxidezlz and molybdenum d i s ~ l p h i d e ~ ~ ~ respectively. The addition of carbon black prevented the interference of silica and caused a signal enhancement of about 3Ooh. Likewise this modifier enhanced A1 and Fe absorption signals by a factor of 1.5 when determined in aluminium oxide and iron oxide respectively.175 Graphite powder has been added to slurries consisting of non-radioactive simulated waste for the determination of Mo Ru Rh and Pd.214 Barnett et d Z l 5 used a mixture of Ca and Mg or Ni and Mg and La as chemical modifiers to determine boron a difficult element using the ETAAS technique in plant samples. Totally pyrolytic graphite tubes were used in order to minimize carbide formation. As for direct solid sampling the oxygen or air ashing technique facilitates pyrolysis of samples with high organic material contents and avoids any build-up of carbonaceous residues in the atomizer. 166*208~z1 lv2l6 The possibility of eliminating both the chemical modifier and the pyrolysis step in slurry sampling has also been studied.Those approaches require the use of isothermal conditions for atomization and Zeeman background correc- tion in order to avoid matrix interferences. Bradshaw and Slavinlg8 reported the rapid determination of As Pb T1 and Se in coal and coal fly ash matrices by automated ultrasonic agitation and elimination of the pyrolysis step. The preci- sion for replicates of slurries varied from 2 to looh depending upon the amount of solid sample delivered to the furnace. For the determination of As in fly ashls8 the background was lower when Pd was used as a modifier. However the background peak was separated from the As peak so that good precision was obtained. Bendicho and de L~os-Vollebregt'~~ have studied the influence of chemical modifiers such as magnesium nitrate palladium nitrate and a mixture of both in the atomization of Cu Co Cr Mn Ni and Fe in several glass materials. Studies by scanning electron microscopy demonstrated that the pyrolysis step caused an increase in particle size of the slurry sub-sample on the platform and therefore could be omitted.Chemical modifiers did not have any effect on recovery. Further research is required in order to establish the conditions in which rapid analysis can be carried out by avoiding the pyrolysis step and chemical modification without any deterioration in the analytical performance. 3.5 Calibration Calibration in slurry analysis can be accomplished by using the same approaches as mentioned for direct solids analy- sis. In contrast to direct solid sampling the slurry technique allows aqueous standards to be used for calibration in many instances.Quantification using aqueous standards has been carried out successfully in combination with different atomization conditions. Platform atomization and inte- grated absorbance quantificati~n,l~J~~ platform atomiza- tion and peak absorbance quantification wall atomiza- tion and integrated absorbance quantification181 and wall atomization and peak absorbance quantification2l7 have all been reported. Integrated absorbance measurements pro- vide accurate results even for wall atomization.181 The standard additions method or calibration using standard slurries were required for calibration when wall atomiza- tion was ~ s e d . ~ ~ ~ J ~ ~ Platform atomization does not always guarantee good accuracy when aqueous standards are used for calibration.The use of a chemical modifier is sometimes necessary in order to match the absorption signals produced by the analyte when in the aqueous standard and when in the slurry.205 The implementation of Zeeman background correc- tion,188 Smith-Hieftje background correctionz1 or wave- length facilitates slurry analysis because they allow effective correction for high background absorbances that arise from the atomization of the matrix. The use of STPF conditions as proposed by Slavin and co-w~rkers.~J~ has been recommended in order to eliminate spectral and chemical interferences. Recent work carried out with slurry sampling proved that the STPF conditions can also be successfully used for slurry atomization.z19 3.6 Applications of Slurry Sample Introduction The applications of slurry sample introduction in ETAAS are shown in Table 2.Applications are presented in a similar form as for direct solid sampling (Table l) the elements that are determined type of sample homogeniza- tion technique and relevant analytical results are also included. 4 Direct Solid Sampling Versus Slurry Sample Introduction The choice of one of the two approaches available for introduction of solid samples into electrothermal atomizers requires a critical look at the factors affecting the whole analytical procedure. The different factors to be considered are discussed below. 4.1 Sample Pretreatment Although it is evident that both direct solid introduction and slurry introduction can be successfully used for a variety of samples several criteria can be followed in order to indicate which procedure is preferable.For instance most applications reported for the analysis of pure metals alloys geological materials and plastics are based on direct introduction of the sample whereas for the materials that are originally in powder form such as sediments soils or particulate matter slurry introduction is preferred. From368 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1 99 1 VOL. 6 Table 2 Applications of slurry sample introduction in ETAAS Sample Element Na and Pb Alumina Ca Cr Cu Fe Pb Homogenization Magnetic Comments RSD=2-6% Reference 181 207 199 202 220 209 22 1 175 203 222 212 191 104 223 219 216 224 173 189 186 225 210 170 21 1 188 188 226 178 Magnetic Study of Pb atomization Study of Pb release from alumina RSD=2-5.4% Pb Magnetic K Na Mn Mg and Cr Ca Cr Cu Fe Na and Pb Ga Manual RSD=2-6% Alumina Ultrasonic and magnetic Suspension in ethanol- water (9+ 1).Wall and PT atomization. Peak height increases owing to the 'mini-platform' effect caused by the alumina Aluminium fluoride and Aluminium oxide and iron aluminium phosphate A1 oxide A1 and Fe Stabilization with high viscosity solvents RSD=4.3% for A1 and 2.4% for Fe Aquatic plants mussels and olive leaves Cd Pb NH4H2P04 as chemical Ascorbic acid as chemical modifier modifier Grinding and Ultrasonic agitation with zirconium spheres Beryllium oxide Be Carbon black as chemical modifier. RSD = 2.9% Biological and vegetal matter Pb and Cd Vortex mixing Carbonaceous slurry (digestion with sulphuric acid) Comparison of direct solid and slurry Comparison between digestion and slurry Single and multi-element determinations Use of an air-ashing step to reduce interferences PT atomization c u Pb Several Ultrasonic Biological matter - Ultrasonic probe Cr Co Pb and Mn Au and Ag Vortex mixing and Triton X- 100 Biological and organic matter Biological materials Biological and environmental samp.:s Chelating sorbents Coal 12 elements Mn Ultrasonic Multi-element determinations Rapid analysis.Automated slurry introduction Ultrasonic Ultrasonic Description of an ultrasonic slurry sampler Preconcentration of noble metals on sorbents Ni+Mg as chemical modifier. DL= 0.1 Au Pt Pd Ir and Rh Magnetic As Pug g-' 10% m/v slurrv DL=0.03 pg g-* for a As Se Magnetic + stabilizing agent Magnetic RSD= 5Oh; DL=*0.05 pg g-l for 15Oh m/v slurry.Air ashing in situ RSD=2-5%. Rapid analysis (1 min per sample) PT atomization and chemical modifiers. RSD = 3.3Oh As Pb T1 and Se Ultrasonic Be Automated mixing with magnetic stirring Ultrasonic +glycerol Coal fly ash As. Cd and PbJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 199 1 VOL. 6 Table 2-continued 369 Sample Foods Foods tuffs Element Homogenization Comments Reference 208 184 200 227 177 194 202 228 229 230 23 1 180 168 63 232 213 214 Cd Pb O2 ashing in situ Pd as modifier. Freeze-dried sample. RSD = 4% O2 ashing and NH4H2PO4 as chemical modifier Batch sorption of elements from solution Rapid analysis. Study of chemical modifiers RSDs are improved by 1.5-2-fold with pre-digestion Carbonaceous slurry Resins are used as solid standards for hair analysis.Resins are used as solid standards DL,=0.6 pg g-' RSD= 8- 1 1 W Miniature magnetic agitation device Pb and Cd Stabilizing agent (Viscalex) Gels As Sb and Bi Magnetic Glasses Cu Co Cry Mn Fe and Ni Gas mixing Glasses Cu Coy Cry Mn Gas mixing and Fe and Ni pre-digestion of slurry with HF Several - Cu and A1 - Hashish Ion-exchange resin Hg Co Ni and Mo Magnetic Iron oxide Pb As Magnetic. Use of Triton X-100 and ammonium phosphate Magnetic Iron pigments Suspension in 0.1W m/v Triton X-100 and 0.1% m/v Ni. DL=0.2 pg g-' of As - Lubricating oil 10 elements Dilution with Zn Vortex mixing c u Water-ammonia- kerosene 1 ,4-dioxan as dispersion medium Cd Se Pb Zn - Cs Cu Cr and Mo Marine sediments Milk powder - Comparison between dry ashing wet digestion and slurry Comparison of direct solid sample (cup-in- tube) and slurry.RSD= 5-20°h (slurry) and 9-26% (solid sampling) black. RSD = 3-9% Addition of carbon Milk Molybdenum disulphide Mo Mo Ru Rh and Pd Ultrasonic - Non-radioactive simulated waste Comparison of direct solid sampling and slurry introduction. Samples are mixed with graphite RSD=4.3% Nutritional supplements Se Use of non-ionic surfactants for stabilization 174 Orchard leaves and Pine needles Plant matter pine needles Cd Cu Cry Fe and Pb Pb Mn Fey Cu Cr A1 and Mg RSD = 4Oh RSD=2-9% - Sources of contamination are identified and minimized (sample cups grinding reagents etc.) 190 218 167 233 - Ultrasonic Vortex mixing Automated ultrasonic mixing Plant and biological matter Mn T1 and Pb Vortex mixing and ultrasonic agitation Comparison of ETAAS and electrothermal atomization laser excited atomic fluorescence spectrometry 234 172 Plant tissue and sediments Stabilization with glycerol RSD=0.45-13Oh.DL=0.4 pg of Cd and 5 pg of Pb Pb and Cd370 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1991 VOL. 6 Table 2-continued Element Homogenization Comments Reference Sample Plant tissue Cu Zn Fe and Mn B Magnetic Ultrasonic RSD=6% 217 Use of totally 215 pyrolytic tubes. Ni + Mg + La as modifier Pol ymer-supported River sediment catalysts Pd Stabilizing agent RSD= 0.07% 235 Pb Mn As and Fe Ultrasonic and Automated slurry 187 vortex mixing introduction.Study of error sources c u Use of a thixotropic - 169 thickening agent for stabilization Cd Cr Cu Ni Ultrasonic agitation - 236 Pb and Zn Rocks Sediments Sediments particulate matter Cu Cd Pb Co Stabilization with Cr and Ni glycerol. Pre-digestion of slurry Ultrasonic Al Ca Fe and Mg Cd Cr Cu Ni Pb and Zn Al Ca Fe and Mg - Pd modifier mixed 193 with matrix elements - 24 1 RSD = 8Oh 242 Sewage Sewage sludge RSD = 6- 1 1% for Al 243 1.7% for Ca 2.6% for Fe and 1.9Oh for Mg Wide range of sludge types 237 - 238 Comparison with 239 acid digestion and dry ashing acid digestion and dry ashing and colorimetric methods Comparison with 240 Comparison with FAAS 245 - Cd Cr Ni and Zn Ultrasonic Sb As Bi Te and Ultrasonic agitation T1 Ag Co Mn Mo and Ultrasonic agitation Sn Various - Sewage sludge and sewage efluent Pb Cu Cd Cr Ni and Zn Pb - 244 Stabilization with Stabilization with Magnetic Magnetic Viscalex (3Oh m/v) Viscalex (3% m/v) RSD = 3% 166 Spinach Pb DL=0.05 pg g-' 171 RSD=8% 182 RSD = 2-4Oh for Cd and 183 8-9Yo for Pb Study of chemical 204 components Study of chemical 205 modifiers 0.6 pg of Pd+ 1 pg of Mg 206 as chemical modifier Systematic study of 246 various Pd + Mg mixtures (Pd as nitrate and chloride). 0.6 pg Pd+ 1 pg Mg is optimum O2 ashing Pb Pb and Cd Soil Pb Magnetic Magnetic Pb Pb Magnetic Pb Steel alloys and pure metals Cu Fe Mg Mn Slurries prepared STPF conditions and 197 Pb and Zn with spark ablation calibration with aqueous standards Sulphate catalysts Fe Pt Stabilizing Tungsten strip agents heater for atomization Use of Triton X- 100 and stirring for 3-5 min - 247 179 Tissue samples Titanium ore and silicate rock Co Cr Cu and V Use of stabilizing Comparison between 198 agents flame plasma and electrothermal atomization Cu Fe Mn and Pb Sodium hexametaphosphate DL=O.l-0.5 pg g-I 248 is added to the slurry. Magnetic stirring and atomization Titanium oxideJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 199 1 VOL.6 371 Table 2-continued Sample Element Homogenization Comments Reference Water Cd Cu Pb Coy Ni and Cd Cd c u Freeze-dried alga 249 is used for Cd preconcentration time required 4 min per sample with chelating sorbents (20-30 pm). RSD < 16.5% Cd preconcentration and analysed as slurry with an alga. Extraction of Cu is insensitive to pH algal mass and alkali and alkaline earth metals Metals preconcentrated 2 50 An alga is used for 25 1 Cu is preconcentrated 252 the point of view of sample pre-treatment some samples are more favourable for direct solid sampling owing to the difficulty of grinding e.g.metallic samples which are commonly analysed in the form of chips drillings or turnings. In addition slurries prepared from metallic particles are difficult to stabilize (unless spark ablation is used) or mix in order to yield a homogeneous suspension. Contamination risks exist when grinding materials such as rocks and refractories. Slurry preparation might be less useful for samples that are difficult to disperse and tend to float on the top of the liquid medium such as very light powders. 4.2 Sample Introduction Sample introduction is straightforward for slurries as conventional sample introduction systems in widespread use for liquid samples can be used without further modification.Sample introduction can be automated when using the autosampler; an advantage for routine determina- tions. Errors that can arise during the weighing and transfer operation in direct solid sampling are minimized in slurry sampling. Comparisons of direct solid sampling and slurry sample introduction indicate that although the same accuracy is reached precision is better when using the slurry technique. 4.3 Dilution of the Solid Sample Slurries can be diluted if required so that the absorbance falls into the analytically useful range of the calibration graph. When the analyte concentration is too high direct solid sampling requires dilution of the sample with graphite powder which can cause contamination or a change in the analytical conditions so that the sensitivity is decreased.Changing the analytical conditions i.e. the gas flow rate during atomization and the use of less sensitive lines requires additional optimization. 4.4 Sample Homogeneity For the slurry technique the representative sub-sample is based on the fraction of the total slurry volume that is actually introduced into the furnace. Extraction of the analyte into the slurry medium takes place when the slurry is prepared in acidic media and provides improved preci- sion for non-homogeneous samples. For direct solid sampl- ing the representative sub-sample is based on the amount of sample that is weighed and transferred into the atomizer.The influence of sample non-homogeneity on precision is similar for both direct solid sampling and slurry introduction. 4.5 Peakshape The dependence of the shape of the peak of the absorption signal on the solid matrix is more pronounced when the sample is atomized directly. Vaporization kinetics for the release of the analyte are usually slower when the particle size increases as reported for metallic chips or geological samples that were insufficiently ground. Peak shape for the volatile metals is similar using both approaches but broader peaks are usually obtained for the determination of refractory metals when using direct solid sampling; broad peaks are difficult to integrate in reasonable periods of time.4.6 Calibration Reliable results are usually obtained for direct solid sampling when samples of CRMs are available. In contrast calibration with aqueous standards constitutes a more realistic possibility in the slurry technique provided that appropriate analytical and instrumental methodology is used. 5 Conclusions Solid sampling in ETAAS has received great attention in recent years mainly because of the availability of commer- cial atomizers that facilitate sample introduction and satisfy the requirements for isothermal atomization. The most advantageous designs are probably those that use a graphite support (platform boat microboat and cup) for both sample weighing and atomization. However these systems are manually operated and therefore sample introduction requires training of the analyst. Although sample homogeneity is a critical factor influencing preci- sion it is not a determining factor in achieving accurate results.In many instances sample homogeneity can be improved by using a suitable grinding procedure. In addition sample homogeneity is not critical for many samples and therefore sample masses of about 1 mg can be successfully analysed. The slurry technique gives better analytical performance than direct solid sampling. The slurry concentration can be easily changed so that the analyte concentration falls into the analytically useful range of the calibration graph. More importantly slurry sample introduction can be easily automated. The most trouble- some aspect of the slurry technique is the stabilization of the slurries during the time required to carry out the determination by ETAAS.Reliable procedures for homo- genization such as magnetic and ultrasonic agitation and vortex and gas mixing are available. Chemical modification is easier to perform in the slurry technique than in the direct solid sampling because of more effective contact between372 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 199 1 VOL. 6 the sample particles and the modifier. The use of modern furnace technology enables the analysis of solid samples with difficult matrices and also makes it possible to calibrate using aqueous standards. References 1 L'vov B. V. Znzh. Fiz. Zh. 1959 2 44. 2 L'vov B. V. Talanta 1976 23 109. 3 Langmyhr F. J. Analyst 1979 104 993. 4 Langmyhr F. J. and Wibetoe G.Prog. Anal. At. Spectrosc. 5 Headridge J. B. Spectrochim. Acta Part B 1980 35 785. 6 Sturgeon R. E. Spectrochim. Acta Part B 1989 44 1209. 7 Van Loon J. C. Anal. Chem. 1980,52,955A. 8 Kerber J. D. At. Absorpt. Newsl. 1971 10 104. 9 Slavin W. and Manning D. C. Spectrochim. Acta Part B 1982 37 955. 10 Slavin W. Camrick G. R. Manning D. C. and Pruszkow- ska E. At. Spectrosc. 1983 4 69. 11 Massmann H. Spectrochim. Acta Part B 1968 23 215. 12 Takada K. and Hirokawa K. Fresenius Z. Anal. Chem. 1982,312 109. 13 Takada K. and Hirokawa K. Talanta 1982 29 849. 14 Takada K. Bunseki Kagaku 1983,32 197. 15 Takada K. and Shoji T. Fresenius Z. Anal. Chem. 1983 16 Takada K. and Hirokawa K. Talanta 1983,30 329. 17 Takada K. Talanta 1985 32 921. 18 L'vov B. V. Pelieva L.A. and Sharnopolsky A. I. Zh. Prikl. Spectrosk. 1977 27 395. 19 Chakrabarti C. L. Wan C. C. and Li W. C. Spectrochim. Acta Part B 1980 35 547. 20 Chakrabarti C. L. Karwowska R. Hollebone B. R. and Johnson P. M. Spectrochim. Acta Part B 1987 42 1217. 21 Brown A. A. Lee M. Kullemer G. and Rosopulo A. Fresenius Z. Anal. Chem. 1987 328 354. 22 Kurfiirst U. Fresenius Z. Anal. Chem. 1985 322 660. 23 Baxter D. C. and Frech W. Fresenius J. Anal. Chem. 1990 24 Grobenski Z. Lehmann R. Tamm R. and Welz B. 25 Grobecker K. H. and Kluessendorf B. Fresenius Z. Anal. 26 Kurfiirst U. Kempeneer M. Stoeppler M. and Schuierer 27 Sommer D. and Ohls K. Fresenius Z. Anal. Chem. 1979 28 Kurfiurst U. and Grobecker K. H. LaborPraxis 1981,5,28. 29 Kurfiurst U. and Rues B. Fresenius Z. Anal.Chem. 1981 308 1. 30 Fleckenstein J. Fresenius Z. Anal. Chem. 1985 322 704. 31 Janssen A. Brueckner B. Grobecker K.-H. and Kurfiurst U. Fresenius Z. Anal. Chem. 1985 322 713. 32 Grobecker K.-H. and Muntau H. Fresenius Z. Anal. Chem. 1985,322 728. 33 Strubel G. Rzepka-Glinder V. and Grobecker K.-H. Fresenius Z. Anal. Chem. 1987 328 382. 34 Headridge J. B. and Riddington I. M. Mikrochim. Acta 1982 2 457. 35 Headridge J. B. and Nicholson R. A. Analyst 1982 107 1200. 36 Busheina I. S. and Headridge J. B. Anal. Chim. Acta 1982 142 197. 37 Headridge J B. and Riddington I. M. Analyst 1984 109 113. 38 Price W. J. Dymott T. C. and Whiteside P. J. Spectrochim. Acta Part B 1980 35 3. 39 Atsuya I. Itoh K. and Akatsuka K. Fresenius Z. Anal. Chem. 1987,328 338. 40 Noller B.N. and Bloom H. Anal. Chem. 1977 49 346. 41 Hirokawa K. and Takada K. Bunseki Kagaku 1980 29 675. 42 Atsuya I. and Itoh K. Bunseki Kagaku 1982 31 713. 43 Atsuya I. and Itoh K. Bunseki Kagaku 1982,31 708. 44 Atsuya I. and Itoh K. Spectrochim. Acta Part B 1983 38 1985 8 193. 315 34. 337 253. Mikrochim. Acta 1982 1 1 15. Chem. 1985,322 673. O. Fresenius J. Anal. Chem. 1990 337 248. 298 123. 1259. 45 Itoh K. Akatsuka K. and Atsuya I. Bunseki Kagaku 1984 46 Atsuya I. Itoh K. Akatsuka K. and Jin K. Fresenius Z. 47 Akatsuka K. and Atsuya I. Fresenius Z. Anal. Chem. 1987 48 Atsuya I. and Itoh K. Fresenius Z. Anal. Chem. 1988,329 49 Akatsuka K. and Atsuya I. Anal. Chem. 1989 61 216. 50 Vollkopf U. Grobenski Z. Tamm R. and Welz B. Analyst 1985 110 573. 5 1 Carnrick G. R.Lumas B. K. and Barnett W. B. J. Anal. At. Spectrom. 1986 1 443. 52 Schlemmer G. and Welz B. Fresenius Z. Anal. Chem. 1987 328 405. 53 Vollkopf U. Lehmann R. and Weber D. J. Anal. At. Spectrom. 1987 2 455. 54 Lindberg I. Lundberg E. Arkhammar P. and Berggren P.- O. J. Anal. At. Spectrom. 1988 3 491. 55 Irwin R. Mikkelsen A. Michel R. G. Dougherty J. P. and Preli F. R. Spectrochim. Acta Part B 1990 45 903. 56 Siemer D. Lech J. F. and Woodriff R. Spectrochim. Acta Part B 1973 28 469. 57 Seimer D. D. and Woodriff R. Spectrochim. Acta Part B 1974 29 269. 58 Seimer D. D. Lech J. and Woodriff R. Appl. Spectrosc. 1974 28 68. 59 Lundberg E. and Frech W. Anal. Chim. Acta 1979,104,75. 60 Lundberg E. Appl. Spectrosc. 1978 32 276. 61 Chakrabarti C. L. Xiuren H. Shaole W. and Shroeder W.H. Spectrochim. Acta Part B 1987 42 1227. 62 Schron W. Bombach G. and Beuge P. Spectrochim. Acta Part B 1983 38 1269. 63 Khammas Z. A. Marshall J. Littlejohn D. Ottaway J. M. and Stephen S. C. Mikrochim. Acta 1985 1 333. 64 Schmidt K. P. and Falk H. Spectrochim. Acta Part B 1987 42 431. 65 Rettberg T. M. and Holcombe J. A. Anal. Chem. 1986 58 1462. 66 Rettberg T. M. and Holcombe J. A. Spectrochim. Acta Part B 1984,39,249. 67 Rettberg T. M. and Holcombe J. A. Anal. Chem. 1988,60 600. 68 Eames J. C. and Matousek J. P. Anal. Chem. 1980 52 1248. 69 Stoeppler M. Kurfiirst U. and Grobecker K.-H. Fresenius Z. Anal. Chem. 1985 322 687. 70 Frech W. and Baxter D. C. Fresenius Z. Anal. Chem. 1987 328 400. 71 Takada T. and Koide T. Anal. Chim. Acta 1987 198,303.72 Lundberg E. and Frech W. Anal. Chim. Acta 1979,104,67. 73 Siemer D. D. and Wei H.-Y. Anal. Chem. 1978 50 147. 74 Alder J. F. and Batoreu M. C. C. Anal. Chim. Acta 1983 75 Munoz F. E. Calvo A. and Leon L. E. Anal. Lett. 1983,16 76 Esser P. Fresenius Z. Anal. Chem. 1985 322 677. 77 Durnberger R. Esser P. and Janssen A. Fresenius Z. Anal. 78 Page A. G. Godbole S. V. Kulkarni M. J. Shelar S. S. and 79 Hassell D. C. Rettberg T. M. Fort F. A. and Holcombe J. 80 Tsalev D. L. Slaveykova V. I. and Mandjukov P. B. 81 Matousek J. P. and Brodie K. G. Anal. Chem. 1973 45 82 Grobenski Z. and Lehmann R. At. Spectrosc. 1983,4 1 11. 83 Brodie K. G. and Matousek J. P. Anal. Chim. Acta 1974 84 de Kersabiec A. M. and Benedetti M. F. Fresenius 2. Anal. 85 Takada K. Anal. Sci. 1987 3 221.86 Mohl C. Grobecker K. H. and Stoeppler M. Fresenius Z. 87 Esser P. and Durnberger R. Fresenius Z. Anal. Chem. 88 Kurfiurst U. Rues B. and Wachter K. H. Fresenius Z. Anal. 33 301. Anal. Chem. 1987 326 53. 329 453. 750. 155 199. 835. Chem. 1987,327 343. Joshi B. D. Fresenius Z. Anal. Chem. 1979 296 40. A. Anal. Chem. 1988 60 2680. Spectrochim. Acta Rev. 1990 13 225. 1606. 69 200. Chem. 1987,328 342. Anal. Chem. 1987 328 413. 1987,328 359. Chem. 1983,314 1.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 199 1 VOL. 6 373 89 Van Soon M. and Muntau H. Fresenius Z. Anal. Chem. 1987,328 390. 90 Yamamuchi M. Okubo N. Yokoyama H. Ui T. and Asada E. Bunseki Kagaku 1982 31 56 1. 91 Siemer D. D. and Baldwin J. M. Anal. Chem. 1980 52 295. 92 Pauwels J. De Angelis F.Peetermans F. and Ingelbrecht C. Fresenius J. Anal. Chem. 1990 337 290. 93 Atsuya I. Akatsuka K. and Itoh K. Fresenius J. Anal. Chem. 1990,337,294. 94 Baxter D. C. J. Anal. At. Spectrom. 1989 4 415. 95 Low P. S. and Hsu G. J. Fresenius J. Anal. Chem. 1990 96 Lorch D. W. At. Absorpt. Newsl. 1978 17 95. 97 Zolotovitskaya E. S. and Fidel’man B. M. Zh. Anal. Khim. 98 Awad N. A. N. and Jasim F. Microchem. J. 1989,40 187. 99 Graf-Harsanyi E. and Langmyhr F. J. Anal. Chim. Acta 1980 116 105. 100 Behne D. Bratter P. Gebner H. Hube G. Mertz W. and Rosick U. Fresenius Z. Anal. Chem. 1976 278 269. 101 Langmyhr F. J. and Orre S. Anal. Chim. Acta 1980 118 307. 102 Luecker E. Rosopulo A. and Kreuzer W. Fresenius 2. Anal. Chem. 1987 328 370. 103 Luecker E. Rosopulo A. Koberstein S.and Kreuzer W. Fresenius Z. Anal. Chem. 1987 329 31. 104 Ebdon L. and Evans E. H. J. Anal. At. Spectrom. 1987 2 317. 105 Berggren P.-O. Anal. Chim. Acta 1980 119 161. 106 Rorsman P. and Berggren P.-O. Anal. Chim. Acta 1982 140 325. 107 Matsuno K. Iwao S. and Kodama Y. Bunseki Kagaku 1984 33 125. 108 Nichols J. A. and Woodriff R. J. Assoc. Of Anal. Chem. 1980 63 500. 109 Hahn E. Hahn K. Mohl C. and Stoeppler M. Fresenius J. Anal. Chem. 1990 337 306. 110 Nordahl K. Radziuk B. Thomassen Y. and Weberg R. Fresenius J. Anal. Chem. 1990 337 310. 111 Strubel G. Rzepka-Glinder V. Grobecker K. H. and Jarrar K. Fresenius J. Anal. Chem. 1990 337 316. 112 Ohls K. Spectrochim. Acta Part B 1984 39 1105. 1 13 Langmyhr F. J. and Dahl I. M. Anal. Chim. Acta 198 1,131 114 Rosopulo A.Grobecker K.-H. and Kurfiurst U. Fresenius 115 Fujiwara K. Umezawa Y. Numata Y. Fuwa K. and 11 6 Luecker E. Kreuzer W. and Busche C. Fresenius Z. Anal. 117 Aadland E. Aaseth J. Radziuk B. Saeed K. and Thomas- 118 Lichtenberg W. Fresenius 2. Anal. Chem. 1987 328 367. 119 Fleckenstein J. Fresenius Z. Anal. Chem. 1987 328 396. 120 Chakrabarti C. L Wan C. C. and Li W. C. Spectrochim. 121 Nakamura T. Okubo K. and Sato J. Anal. Chim. Acta 122 Gladney E. S. At. Absorpt. Newsl. 1977 16 42. 123 Ali A. H. Smith B. W. and Winefordner J. D. Talanta 1989 36 893. 124 Langmyhr F. J. and Aadalen V. Anal. Chim. Acta 1980 115 365. 125 Kerber J. D. Koch A. and Peterson G. E. At. Absorpt. Newsl. 1973 12 104. 126 Schothorst R. C. Geron H. M. A. Spitsbergen D. and Herber R.F. M. Fresenius 2. Anal. Chem. 1987 328 393. 127 Lord D. A. McLaren J. W. and Wheeler R. C. Anal. Chem. 1977,49,257. 128 Johnson D. Headridge J. B. McLeod C. W. Jackson K. W. and Roberts J. A. Anal. Proc. 1986 23 8. 129 Guenais B. Poudoulec A. and Minier M. Analusis 1982 10 78. 130 Takada K. Bunseki Kagaku 1989,38 356. 131 Kowalska A. Kedziora M. and Kedziora A. At. Spectrosc. 1980 1 33. 132 Alder J. F. Samuel A. J. and West T. S. Anal. Chim. Acta 1976 87 313. 337 299. 1981 36 1564. 303. Z. Anal. Chem. 1984 319 540. Fujiwara S. Bunseki Kagaku 1977 26 735. Chem. 1989,335 176. sen Y. Fresenius 2. Anal. Chem. 1987 328 362. Acta Part B 1980 35 93. 1988 209 287. 133 Alder J. F. Pankhurst C. A. Samuel A. J. and West T. S. Anal. Chim. Acta 1977 91 407. 134 Sansoni B.and Brunner W. Fortschr. Atomspektrom. Spurenanal. 1984 1 187. 135 Hennig W. Bemdt H. and Schaldach G. Fresenius J. Anal. Chem. 1990,337 275. 136 Krebs B. Fresenius Z. Anal. Chem. 1987 328 388. 137 Ishibashi W. Kikuchi R. and Yamamoto K. Bunseki Kagaku 1979 28 394. 138 Marks J. Y. Welcher G. G. and Spellman R. J. Appl. Spectrosc. 1977 31 9. 139 Zhao S. Liu G. and Liu G. Lihua Jianyan Huaxue Fence 1989 25 11. 140 Knecevic G. and Kurfiurst U. Fresenius Z. Anal. Chem. 1985,322 7 17. 141 Simon P. J. Glessen B. C. and Copeland T. R. Anal. Chem. 1977,49 2285. 142 Berggren P.-O. Berglund O. and Hellman B. Anal. Bio- chem. 1978 84 393 143 Grobecker K. H. and Kurfiurst U. Fortschr. Atomspektrom. Spurenanal. 1984 1 365. 144 Nilsson T. and Berggren P.-O. Anal. Chim. Acta 1984,159 381.145 Robbins W. K. Runnels J. H. and Merryfield R. Anal. Chem. 1975,47 2095. 146 Klubendorf B. Rosopulo A. and Kreuzer W. Fresenius Z. Anal. Chem. 1985 322 721. 147 Machiroux R. and Dupont J. C. Anal. Chim. Acta 1976,85 231. 148 Girgis-Takla P. and Chroneos I. Analyst 1978 103 122. 149 Henn E. L. Anal. Chim. Acta 1974 73 273. 150 Havezov I. and Tamnev B. Fresenius 2. Anal. Chem. 1978 291 127. 151 Herber R. F. M. Roelofsen A. M. Roelfzema W. H. and Copius Peereboom-Stegeman J. H. J. Fresenius Z. Anal. Chem. 1985,322 743. 152 Alvarado J. and Petrola A. J. Anal. At. Spectrom. 1989 4 41 1. 153 Gong H. and Suhr N. H. Anal. Chim. Acta 1976 81 297. 154 Prudnikov E. D. and Rantsev A. A. Zh. Anal. Khim. 1979 155 Lehmann C. Lorber K.-E. Ruthenberg K. and Schulze G.156 Headridge J. B. Johnson D. Jackson K. W. and Roberts J. 157 Grobenski Z. Weber D. Welz B. and Wolff J. Analyst 158 Lundberg E. and Frech W. Anal. Chim. Acta 1979,108,75. 159 Frech W. Lundberg E. and Barbotti M. M. Anal. Chim. Acta 1981 131 45. 160 Backman S. and Karlsson R. W. Analyst 1979 104 1017. 161 Page A. G. Godbole S. V. Deshkar S. B. and Joshi B. D. Anal. Lett. Part A 1978 11 619. 162 Page A. G. Godbole S. V. Kulkami M. J. Porwal N. K. Shelar S. S. and Joshi B. D. Talanta 1983 30 783. 163 Akatsuka K. and Atsuya I. Fresenius Z. Anal. Chem. 1989 164 Akatsuka K. Nobuyama N. and Atsuya I. Anal. Sci. 1989 165 Homer E. and Kurfurst U. Fresenius Z. Anal. Chem. 1987 166 Stephen S. C. Littlejohn D. and Ottaway J. M. Analyst 167 Brady D. V. Montalvo J.G. Jung J. and Curran R. A. At. 168 Brady D. V. Montalvo J. G. Glowacki G. and Pisciotta 169 Fuller C. W. and Thompson I. Analyst 1977 102 141. 170 Ebdon L. and Parry H. G. M. J. Anal. At. Spectrom. 1987 17 1 Littlejohn D. Stephen S. C. and Ottaway J. M. Anal. Proc. 172 Hoenig M. and Van Hoeyweghen P. Anal. Chem. 1986,58 173 Miller-Ihli N. J. Anal. At. Spectrom. 1988 3 73. 174 Thompson D. D. and Allen R. J. At. Spectrosc. 198 1,2,53. 175 Yoshimura C. and Shinya N. Nippon Kagaku Kaishi 1987 34 1725. Fresenius Z. Anal. Chem. 1989 333 707. A Anal. Chim. Acta. 1987 201 3 11. 1983 108 925. 335 200. 5 475. 328 386. 1985 110 1147. Absorpt. Newsl. 1974 13 118. A. Anal. Chim. Acta 1974 70 448. 2 131. 1985 22 376. 2614. 8 1545.374 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 20 1 202 203 204 205 206 207 208 209 210 21 1 212 213 214 215 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 199 1 VOL.6 Majidi V. and Holcombe J. A. Spectrochim. Acta Part B 1990 45 753. Bendicho C. and de Loos-Vollebregt M. T. C. Spectrochim. Acta Part B 1990 45 679. van Loenen D. C. and Weers C. A. Fortschr. Atomspektrom. Spurenanal. 1986 2 635. Priesner D. Sternson L. A. and Repta A. J. Anal. Lett. 1981 14 1255. Saba C. S. Rhine W. E. and Eisentraut K. J. Appl. Spectrosc. 1985 39 689. Slovak Z. and Docekal B. Anal. Chim. Acta 1981 129 263. Jackson K. W. and Newman A. P. Analyst 1983,108,261. Hinds M. W. Jackson K. W. and Newman A. P. Analyst 1985,110 947. Lynch S. and Littlejohn D. J. Anal. At.Spectrom. 1989 4 157. Miller-Ihli N. J. J. Anal. At. Spectrom. 1989 4 295. Carnrick G. R. Daley G. and Fotinopoulos A. At. Spectrosc. 1989 10 170. Epstein M. S. Carnrick G. R. Slavin W. andMiller-Ihli N. J. Anal. Chem. 1989 61 1414. Bradshaw D. and Slavin W. Spectrochim. Acta Part B 1989,44 1245. Jordan P. Ives J. M. Carnrick G. R. and Slavin W. At. Spectrosc. 1989 10 165. Fagioli F. Landi S. Anal. Lett. 1983 16 1435. Fagioli F. Landi S. Locatelli C. and Bighi C. At. Spectrosc. 1986 7 49. Fagioli F. Locatelli C. Scanavini L. Landi S. and Berti Donini G. Anal. Sci. 1986 2 239. Hoenig M. Regnier P. and Wollast R. J. Anal. At. Spectrom. 1989 4 631. Bendicho C. and de Loos-Vollebregt M. T. C. Spectrochim. Acta Part B 1990 45 695. Pchelkin A. I. Kharlamov I. P. Gusinskii M.N. and Shipova E. V. Zh. Anal. Khim. 1987 42 2138. Karyakin V. Yu. Kharlamov I. P. and Pchelkin A. I. Zavod. Lab. 1988 54 36. L'vov B. V. and Novichikhin A. V. At. Spectrosc. 1990 11 1. Fuller C. W. Hutton R. C,. and Preston B. Analyst 1981 106 913. Karwowska R. and Jackson K. W. J. Anal. At. Spectrom. 1987 2 125. Stephen S. C. Ottaway J. M. and Littlejohn D. Fresenius Z. Anal. Chem. 1987 328 346. Chen T. and Xu C. Guangpuxue Yu Guangpu Fenxi 1988 8 45. Holcombe J. A. and Majidi V. J. Anal. At. Spectrum. 1989 4 423. Ebdon L. and Lechotycki A. Microchem. J. 1987,36,207. Hinds M. W. and Jackson K. W. J. Anal. At. Spectrom. 1987 2 441. Hinds M. W. and Jackson K. W. J. Anal. At. Spectrom. 1988 3 997. Hinds M. W. Katyal M. and Jackson K. W. J. Anal. At. Spectrom.1988 3 83. Karwowska R. and Jackson K. W. Spectrochim. Acta Part B 1986 41 947. Olayinka K. O. Haswell S. J. and Grzeskowiak R. J. Anal. At. Spectrom. 1986 1 297. MareEek J. and Synek V. J. Anal. At. Spectrom. 1990 5 385. Ebdon L. and Pearce W. C. Analyst 1982 107 942. Ebdon L. and Parry H. G. M. J. Anal. At. Spectrom. 1988 3 131. Yoshimura C. and Huzino T. Nippon Kagaku Kaishi 1984 7 1132. Yoshimura C. and Huzino T. Nippon Kagaku Kaishi 1985 5 882. Schmiedel G. Mainka E. and Ache H. J. Fresenius 2. Anal. Chem. 1989,335 195. Barnett N. W. Ebdon L. Evans E. H. and Ollivier P. Anal. Proc. 1988 25 233. 216 217 218 219 220 22 1 222 223 224 225 226 227 228 229 230 23 1 232 233 234 235 236 237 238 239 240 24 1 242 243 244 245 246 247 248 249 250 25 1 252 Ebdon L. Fisher A.S. Parry H. G. M. and Brown A. A. J. Anal. At. Spectrom. 1990 5 321. de Benzo Z. A. Fernandez R. Carrion N. and Eljuri E. At. Spectrosc. 1988 9 87. Carridn N. de Benzo Z. A. Moreno B. Fernhndez A. Eljuri E. J. and Flores D. J. Anal. At. Spectrom. 1988 3 479. Miller-Ihli N. Spectrochim. Acta Part B 1989 44 1221. Slovak Z. and Docekal B. Anal. Chim. Acta 198 1,129,273. Yoshimura C. and Huzino T. Nippon Kagaku Kaishi 1985 7 1392. Ebdon L. and Lechotycki A. Microchem. J. 1986,34 340. Favretto L. Pertoldi Marletta G. and Favretto G. Mikro- chim. Acta 1981 1 387. Thunus L. and Dauphin J.-F. Anal. Chim. Acta 1990 235 393. Kubrakova I. V. Varshal G. M. and Kudinova T. F. Zh. Anal. Khim. 1987 42 126. Haraldsen L. and Bruno Pougnet M. A. Analyst 1989 114 1331. Slovak Z. and Docekal B. Anal. Chim. Acta 1980,117,293. Alder J. F. and Batoreu M. C. C. Anal. Chim. Acta 1982 135 229. Slovak Z. Anal. Chim. Acta 1979 110 301. Ldpez Garcia I. and Hernhndez Cdrdoba M. J. Anal. At. Spectrom. 1989 4 701. Mpez Garcia I. and Hernhndez Cdrdoba M. J. Anal. At. Spectrom. 1990 5 647. Wagley D. Schmiedel G. Mainka E. and Ache H. J. At. Spectrosc. 1989 10 106. Miller-Ihli N. J. Fresenius J. Anal. Chem. 1990 337 271. Butcher D. J. Irwin R. L. Takahashi J. Su G. Wei G.-T. and Michel R. G. Appl. Spectrosc. 1990 44 1521. Chiricosta S. Gum G. Gallo R. Spadaro A. and Vitarelli P. At. Spectrosc. 1982 3 185. Carrondo M. J. T. Reboredo F. Ganho R. M. B. and Oliveira J. F. S. Talanta 1984 31 561. Lester J. N. Harrison R. M. and Perry R. Sci. Total Environ. 1977 8 153. Stoveland S. Astruc M. Perry R. and Lester J. N. Sci. Total Environ. 1978 9 263. Kempton S. Stemtt R. M. and Lester J. N. Talanta 1982 29 675. Sterritt R. M. and Lester J. N. Analyst 1980 105 616. Carrondo M. J. T. Lester J. N. and Perry R. Analyst 1979 104 831. Carrondo M. J. T. Perry R. and Lester J. N. Anal. Chim. Acta 1979 106 309. Carrondo M. J. T. Lester J. N. and Perry R. Talanta 1979 26 929. Carrondo M. J. T. Perry R. and Lester J. N. Sci. Total Environ. 1979 12 1. Stoveland S. Astruc M. Perry R. and Lester J. N. Sci. Total Environ. 1979 13 33. Hinds M. W. and Jackson K. W. J. Anal. At. Spectrom. 1990 5 199. Nakamura S. Suzuki H. and Kubota M. Bunseki Kagaku 1989 38 399. Fuller C. W. Analyst 1976 101 961. Majidi V. and Holcombe J. A. J. Anal. At. Spectrom. 1989 4 439. Ishmiyarova G. R. Shcherbininina N. I. Sedykh E. M. Myasoedova G. V. and Vul'fson E. K. Zh. Anal. Khim 1988,43 1981. Majidi V. and Holcombe J. A. Spectrochim. Acta Part B 1988,43 1423. Mo S. J. and Holcombe J. A. Anal. Chem. 1990,62 1994. Paper I /004696 Received February Ist 1991 Accepted April I Oth 1991