November 1979 Vol. 104 No. 1244 The Analyst Direct Analysis of Solids by Atomic-absorption Spectrophotometry A Review F. J. Langmyhr Department of Chemistry University of Oslo Oslo 3 Norway Summary of Contents Introduction Previous literature Instruments Atomisation methods Atomisation in flame cells Atomisation of undispersed powders Atomisation of powders suspended in solid dispersing agents Atomisation of powders suspended in liquid dispersing agents Atomisation of powders sampled by arcing or sparking and suspended Atomisation in tube or crucible cells and in cells in the form of a T inverted Tor + Atomisation from rod strip or braid cells Atomisation from an electrically heated rod or a closed tube placed in a flame Atomisation in d.c. or a.c.arcs Atomisation by cathodic sputtering Atomisation by laser or discharge lamps Other atomisation methods Conclusion in gaseous dispersing agents Atomisation of air- and water-suspended solids Reactions in atomisation cells Thermochemistry Kinetics Interferences Samples sampling sampling errors and sample preparation Standards standardisation and techniques of measurement Accuracy and precision Time of analysis Contamination control Applications Appendix Conclusion References Keywords Review ; solids analysis ; atomic-absorption spectrophotometry Introduction Since its introduction in 1955 by the Australian physicist Walsh,l atomic-absorption spectro-photometry (AAS) has developed into one of the most versatile techniques in analytical chemistry.The method permits the simple rapid and inexpensive determination of nearly 70 elements in materials ranging from geology mining and metallurgy to biology pharmacy and medicine. Today AAS is indispensible in those fields where large numbers of samples have to be analysed on a routine basis such as clinical chemistry and geochemical assay. 99 994 LANGMYHR DIRECT ANALYSIS OF SOLIDS Analyst Vol. 104 The high sensitivities and the low limits of determination make AAS particularly useful in trace analysis and pollution control. The accuracy and precision of AAS with flame atomisation are in general sufficiently high to allow the determination not only of the minor and trace constituents but also of the main components of many materials. AAS based on atomisation in flames is now a well established and universally employed technique.However during the past decade much work has been devoted to the problem of replacing the conventional method of flame atomisation by more sensitive and convenient methods. Among the new techniques atomisation in non-flame cells has been found to be particularly promising. Compared with flame atomisation non-flame cells have the advan-tages of requiring smaller amounts of sample of having lower limits of determination and of being better suited for the direct analysis of solid materials. The principle of analysing solids by atomising the components directly from the solid state offers some definite advantages the time-consuming decomposition step can be omitted and the analysis can be carried out without addition of reagents and without any separation and/or concentration steps; the risks of introducing contaminants and of losing the element to be determined are thus considerably reduced.The disadvantages of the direct AAS analysis of solids are that the method is destructive, that normally only one element can be determined at a time that the use of small samples may introduce sampling errors and that interferences may give systematic errors during the measurement of absorption. In this paper a survey is given of the equipment and the procedures that have been employed for the direct AAS analysis of solids; a bibliography on the applications of the technique is also included. The basic theory of AAS is not included; for comprehensive treatises on this subject reference is made to the many advanced textbooks in the field.Previous Literature Some previous survey the whole or parts of the present field. General informa-tion on electrothermal atomisation in AAS can be found in a recent monograph.6 Instruments A selection of instruments and atomisation cells for AAS is commercially available and annual surveys of the equipment are published.' The flame technique is still the most widely used method for atomisation and consequently all instruments have this basic equipment. Most manufacturers also offer other atomisation cells some of which are suitable for the direct analysis of solids. The discussion of the AAS instruments will be confined to pointing out the demands that the analyses of interest here make on the instrumentation.In many such analyses solid or liquid matrix particles and matrix molecules cause considerable background absorption which may introduce serious systematic positive errors. It is therefore imperative that the instrument is equipped with a device that compensates for this interference. Today most AAS instruments are or may be equipped with background correctors the correction for non-specific absorption of radiation being based on consecutive measurement of atomic absorption plus background and of the background alone. It should be noted that the arc-source deuterium lamps normally employed for background correction only cover the wavelength range 180-350 nm. Some manufacturers offer instru-ments equipped with two background correctors the second being an inexpensive tungsten lamp.This lamp compensates for non-specific absorption in the wavelength range 300-800 nm and it is recommended that this lamp be used down to about 300 nm; in this way the lifetime of the much more expensive deuterium lamp may be prolonged. The background correctors are only capable of compensating for a certain amount of non-specific absorption. Specifications from instrument manufacturers indicate maximum correction capacities varying from 0.4 to 1.7 absorbance units. The capacity of background correctors may be measured by inserting sieves of varying mesh size in the light path of the instrument November 1979 BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 995 In the direct AAS analysis of solids the amplitude and profile of the signal depend on the rates of free atom formation and loss.The rate of free atom formation depends on the temperature and rate of heating on the reactions occurring in the cell and on the composition of the matrix. The factors governing the loss of free atoms are the expulsion and diffusion of atoms out of the absorbing zone and if a purging gas is led through the cell of the flow-rate of the gas stream. The amplitude of the signal is attained when the rates of atom production and atom loss are identical. As the rate of free atom loss is difficult to control in atomisation cells the observed signal will depend mainly on the rate of formation of free atoms. If a number of free atoms are produced rapidly they give a large signal of short duration whereas if the same number is produced slowly the signal will be lower and of longer duration.An important problem is to define whether the analysis should be based upon measuring peak heights or peak areas. Traditionally the peak-height signal has been used and instru-ments of earlier design were not equipped with devices for measuring the areas under the peaks. Today most instruments measure both amplitudes and peak areas ; the amplitudes are either displayed digitally or may be read on an absorbance scale. The measurement of the peak heights of the fast transient signals obtained during the direct analysis of solids requires a fast electronic handling system. This system should be able to record signals of a duration of about 0.1 s. The chart recorders normally employed have a response time of 0.3-0.5 s for full-scale deflection.This may be sufficient for recording small transient signals but is not adequate for registering large signals of short duration. Instead of a laboratory recorder a transient recorder a storage oscilloscope or as in recently introduced instruments a video screen may be used for signal registration. The main objection to the use of peak-height measurements is that even for the more volatile elements and for high rates of heating the rate of free atom formation and conse-quently the profile and the amplitude of the peak may vary considerably with the compo-sition of the matrix. It is therefore strongly recommended to base the direct analysis of solids upon the measurement of peak areas. A measuring system that records both the absorption peak and the area under the peak is very useful.In addition to avoiding the errors of varying rates of atomisation the measurement of peak areas has the advantages of often giving a better precision of giving a calibration graph that is linear over a larger range and of permitting the use of lower atomisation temperatures; the last factor increases the life of the atomisation cell and reduces the risks of sputtering and of producing non-atomic species in such amounts that the compensating capacity of the background corrector is exceeded. The atomisation cell should preferably be connected to a power supply that permits both step and ramp heating (ramp heating allows gradual heating of the cell). A low rate of heating between the drying ashing and atomisation steps reduces the risks of violent reactions and sputtering.It is hoped that future developments in AAS may lead to the introduction of spectro-photometers and atomisation cells designed to meet the special requirements of the solid-sampling technique such as cells of a suitable size faster signal handling systems and improved background correction facilities. The use of the Zeeman effect may solve some of the problems adhering to the equipment employed today. The Zeeman instruments have a single radiation source so that the beam alignment is no longer a problem they have a simpler optical system the background measurements are made closer to the analyte line and they can compensate for a larger non-specific absorption of radiation.To the author’s knowledge only two instruments of this type are commeicially available [the Hitachi (Japan) and the Erdmann & Griin (West Germany) atomic-absorption spectrophotometers] . In addition to the Zeeman effect the Faraday effect i.e. a system utilising beam polarisa-tion has potentiality as a method for correcting the background. Atomisation in Flame Cells Atomisation of undispersed powders powdered solids directly into flames have not been described in the literature. Atomisation Methods To the author’s knowledge AAS methods based on the introduction of undispersed However 996 LANGMYHR DIRECT ANALYSIS OF SOLIDS Analyst Vol. 104 powdered samples have been placed into small platinum or graphite crucibles which were then heated in a flame.8 The low temperatures obtained in these atomisation cells allow the analysis of only the most volatile elements.Atomisation of powders suspended in solid dispersing agents An attempt at the AAS analysis of a powdered sample mixed with a dispersing agent was made9 with a miniature Archimedian screw for introducing the sample into a gas stream feeding a pre-mixed flame. The samples were diluted 100-fold with calcium carbonate of constant particle-size distribution. The equipment was used for determining palladium in cat a1 ysts . Another p r o c e d ~ r e l ~ - ~ ~ is based on mixing the sample with sodium chloride or sodium carbonate depositing the mixture between the threads of a steel screw and atomising the analyte by moving the screw into an air - acetylene flame.The technique was applied to the determination of lithium rubidium caesium and lead in silicate minerals and rocks. The principle and application of a flame cell based on the use of a solid fuel has been described.15-18 The powdered solid sample is mixed with a solid propellant the mixture is pressed into a tablet and one or more tablets are burned just below the radiation beam of the instrument. The method was used for the determination of various metals in their ores. Atomisation of powders suspended in liquid dispersing agents At an early stage of the development of AAS the method found widespread application in the analysis of wear metals in used lubricating and hydraulic oils and greases. The only pre-treatment necessary was to dilute the sample with an organic solvent in order to over-come the viscosity problems during aspiration.In these samples the metals are largely present as suspended particles of metals and/or alloys the efficiency of aspiration and atomisation in flames depend on the particle size. Most of the workers who have made contributions in this field agree that large particles are not completely vaporised and quantitative results may therefore not be obtained. Refer-ences 19-34 pertain to the use of this technique. The analysis of suspensions of wear metals in oils and greases by flame atomisation is now replaced by atomisation in furnaces; by the latter method complete vaporisation of the suspended particles is secured. A device was con~tructed~~ for keeping an aluminium oxide catalyst suspended in water and in various organic solvents; among the dispersing agents tested methanol gave the highest readings for cobalt and molybdenum.The samples were run against closely matched solid standards. Dispersions of metals and alloys can be prepared by operating a spark immersed in ~ a t e r ~ 6 the dispersion being aspirated into a flame. The method has been applied successfully to the determination of minor and trace elements in steel and aluminium alloys. Other workers3' determined tin in suspensions of tin(1V) oxide tin(1V) sulphide and tin(I1) oxide in water without or with agents added to prevent settling. Good results were obtained when the samples and standards were closely matched. Fundamental studies on the efficiency of atomisation of suspended particles have been made.32938 In these investigations metal pa,rticles (of copper iron and silver) and geo-logical samples (sulphide ores zinc concentrate stream sediments and silicate rocks) were suspended in water or in mixtures of organic solvents and the suspensions were aspirated into flames.The results clearly demonstrated the influence of particle size on the atomisation efficiency, which was found to be well below loo% even for particles of diameter 1 ,urn. The determination of trace elements in titanium(1V) oxide pigments has been described.39 It has been shown that the particle size of this pigment is easily reduced by hand-grinding to below 1 pm and the atomisation efficiency can therefore be expected to be relatively high. The only advantages of the methods surveyed in this section are that the sample treat-ment in some methods such as those based upon dispersing the powdered sample in a liquid, is simple and rapid and that many determinations can be made on the same suspension.However a number of disadvantages some of which are serious are inherent in techniques based on atomising undispersed or dispersed solids in flames. The main objection is the incomplete atomisation of the suspended particles. As is apparent from the papers surveye November 1979 BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 997 above and also other theoretica140 and experimental s t ~ d i e s ~ l - ~ ~ complete vaporisation of particles in a flame is possible only when the particle size is reduced to below 1 pm. As most analysts know the reduction of the particle size of many samples to below 1 pm is virtually impossible.Other disadvantages are the difficulties connected with the preparation of homogeneous mixtures of the sample and a solid dispersing agent the risk that the solid mixture may sputter during heating and that particles settle in liquid dispersing agents. The techniques discussed in this section are not to be recommended in those instances where a high degree of accuracy and precision is required. The methods may however, be useful in such fields as geochemical prospecting. Atomisation of 9owders sampled by arcing OY sparking and suspended in gaseous dispersing agents Solid materials may be transformed into an aerosol of fine particles by sampling with a d.c. arc or a high-voltage spark. A gas (air or argon) fed through the arc or spark chamber, transports the particles into a flame for atomisation and analysis.The method allows convenient sampling of metals and alloys ; non-conducting solid powders may be sampled either from a conventional graphite supporting electrode44 or be mixed with copper powder and pressed into discs of a suitable form.45 By the present technique the mass nebulised is of the order of a few milligrams; these amounts will normally be sufficient to ensure representative sampling. From the data reported in the literature it seems that the accuracy and precision are satisfactory. The method is claimed by some workers46 to be rapid and simple while other workers45 state that the time of analysis is lengthy mainly because of the long pre-spark time necessary in order to reach a steady state.The mechanism of the formation of sample particles is not sufficiently well known. For all techniques based on the introduction of solid particles into flames it is of para-mount importance that the size of the particles is so small that complete vaporisation is secured. The results of some workers46 indicate that this may represent a problem. Unfortunately the effect of varying the parameters upon the particle size distribution has not been studied. Until the results of further work are available no definite conclusions can be d6awn as to the general applicability of the technique to the direct analysis of solids. The matrix and mutual interference effects should also be investigated. Atomisation in Tube or Crucible Cells and in Cells in the Form of a T Inverted T or + The atomisation cells preferably employed today for general non-flame work and for the direct analysis of solids are those in the form of a simple tube or crucible.To some extent more complicated constructions in the form of a T inverted T or + are also used. All of these cells are heated electrically by one or more resistance or induction circuits. Carbon tube furnaces for the study of absorption spectra are d e s ~ r i b e d ~ ~ ~ * ~ in the literature from the early part of this century. Pioneering work on the construction and application of atomisation cells of the present type was started by L’vov49 in 1959; in a series of papers he and his co-workers set a founda-tion for later developments in this field.General surveys6JO on the construction of electrothermal atomisers and special surveys2v3 on those employed in direct analysis of solids have been published. The cells are made of graphite or metal (mostly tantalum or tungsten); the preferred material today is graphite. Graphite can be heated rapidly to above 3300 K its reducing property is of advantage in many atomisations the material is pure or can easily be purified by heating and it is inexpensive and easy to form. The disadvantages inherent in cells made of graphite are that they are combustible that some require a considerable amount of electric power (up to 15 kW) and that they may exhibit memory effects when the element to be determined is incompletely atomised and accumulates in the cell; the latter problem is likely to occur with elements forming stable carbides in these instances higher atomisation temperatures and/or longer atomisation times are required.About 30 papers describe cells belonging to the present group 998 LANGMYHR DIRECT ANALYSIS OF SOLIDS Analyst Vol. 104 The graphite atomisation cells are normally made of the high-purity material used for emission spectrography. To reduce the porosity of the graphite (and consequently the loss of free atoms by diffusion) to minimise carbide formation for certain elements and to prolong the life of the cell the whole or interior surface may be coated with a layer of pyrolytic graphite. As the coating is progressively removed during use the cell has either to be re-coated or better be continuously coated by adding a hydrocarbon e.g.methane or propane to the purging gas.51 Similar beneficial effects are obtained by soaking the graphite cell in certain metal salt solutions; by heating protective coatings of carbides of boron, hafnium molybdenum niobium tantalum titanium tungsten silicon vanadium or zirconium are formed.52 It is also possible to make cells of glassy carbon which is denser and mechanically stronger than the normal or pyrolytically coated graphite. The cells offered by the manufacturers are of the tube or crucible type; to the author's knowledge cells of the T inverted T or + form are not commercially available. The tube cells should preferably have an inner diameter greater than 6 mm; in tubes of smaller diameter it is difficult to introduce solid samples and the sample is also likely to interfere with the beams from the radiation sources.In comparison with the tube cells cells of the crucible type have a shorter absorbing zone and are therefore likely to exhibit a poorer sensitivity. In some types of crucible cells the evaporated species also condense more rapidly than in tube cells. The crucible cells are normally heated by resistance heating the cell being clamped between two graphite rods. In order to obtain reproducible temperatures it is important that the parts forming the contacts are closely matched and that they are in good contact. It is not recommended to remove the crucible for weighing and then re-instal it for atomisation. The crucible cells have the advantage of requiring less power than most tube cells; as the sample is placed in the bottom of the crucible it does not interfere with the light.The cells of the T,53 inverted T54 and + 5 5 9 5 6 forms are of a more complicated construction; they are bigger than the tube or crucible cells and require more power and separate heating circuits may be necessary for the various parts. The maximum temperature obtained in most of these cells is 2800 K. Some of the cells of the + form55956 allow rapid and successive analyses of various materials. The long absorption tube of these cells gives a high sensitivity, and the well below the absorption tube keeps the samples away from the beams of the radiation sources. Among the various atomisation cells described in the literature the dual chamber cell should be mentioned5'; this cell has a closed cylindrical graphite chamber around a con-ventional graphite tube; its construction is similar to a type described in a previous paper.58 The sample is placed in the outer chamber and on heating the vapours diffuse at different rates into the inner tube where the absorption is measured.The cell can accommodate larger samples than most tube or crucible cells and may be useful for separating analytes from matrices producing large amounts of smoke. However the low temperature of the cell allows only elements of high volatility to be determined. Other devices for separating analytes from non-absorbing species have been suggested. One of the recent constructions is a modification of a T-type of furnace,59 in which the sample crucible is closed with a porous graphite cover that filters out and/or decomposes the smoke particles.However analytes forming stable carbides may be trapped in the cover. The cells constructed for the separation of analytes from non-absorbing species have a complicated design and can only be used for the more volatile elements. The versatility of the atomisation cells discussed in this section is demonstrated by their applications; inorganic and organic materials in the form of solid powders drillings cuttings of sheet or fibres samples of soft and hard tissue etc. have been analysed successfully. As is apparent from the above discussion of the analysis of wear metals in used lubricating and hydraulic oils and greases, the atomisation in flames of metals in suspension cannot be considered as quantitative; however by atomising these samples in the present type of cells accurate results can be expected for a number of metals normally encountered in used lubricating and hydraulic oils and grea~es.60-6~ The samples may also be added to the present types of cells as slurries in water to which either sodium hexametapho~phate~~ or a thixotropic thickening agent is added.66 The samples may also be added as suspensions in liquids November 1979 BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 999 Atomisation from Rod Strip or Braid Cells Horizontal graphite rods with a hole in or a slit along the upper surface for the sample, strips made of metal (tantalum tungsten molybdenum or platinum) and graphite braid have been described and some are also commercially available.In comparison with the graphite tube or crucible cells the present group of atomisation devices have certain attractive advantages; they are small and of a simple construction, thus requiring less power they are heated and cooled more rapidly and those made of metal do not form carbides. The main disadvantage of using rod strip or braid cells is that no additional energy is available in the space above the cell to prevent relatively rapid condensation of the evapora-ted species and consequently the effective lifetime of the absorbing atoms is considerably shorter than that when tube cells are used. The braid cells do not seem to have been applied to the direct analysis of solids; the construction makes this type less adaptable to the analysis of powders.Metal cells become brittle as a result of reactions with the samples and with air the metals employed are rather expensive and some such as tungsten are often not sufficiently ductile to permit the production of more complicated forms. The tendency during recent years has been to replace the present types of cells with tube cells made of graphite. Some of the present types of cells have a small sample capacity. Atomisation from an Electrically Heated Rod or a Closed Tube Placed in a Flame Various w0rkers~~-70 have described atomisation cells that consist of an electrically heated rod or a closed tube placed horizontally in a flame; the pulverised solid samples are placed in either an open slit or a crater in the rod or in an axial hole in the rod the hole being closed with graphite powder or a graphite washer.The flame transports the evaporated species through the radiation beam and prevents rapid condensation. From the data published on the use of the closed graphite tube about 30 elements can be determined including such low-volatility elements as silicon titanium molybdenum and vanadium. Inorganic materials such as minerals rocks ores ceramics slags oxides and salts have been analysed successfully. A special type of atomisation cell consists of two graphite tubes one fitted coaxially into the other58; the cell is heated electrically and during heating it is enveloped in a reducing flame. The solid samples are placed in the cavity between the tubes and on heating the vapours diffuse through the porous walls of the inner tube; to prevent diffusion through the outer tube and the end parts these are either made of glassy carbon or covered with pyrolytic graphite.The different rates of diffusion make it possible to make measurements that are otherwise influenced by a large non-specific absorption of radiation. The cell can be heated only to 2900 K. Unfortunately there are few data to demonstrate the feasibility of this cell. Some of the constructions belonging to the present group have the advantage of having a large sample capacity (from tens to hundreds of milligrams) and the risk of introducing sampling errors is thus reduced. t o offer any definite advantages. The disadvantages are that facilities for both electric and gas heating are required.In comparison with the simple tube cells the present types of equipment cannot be seen The equipment surveyed in this section is not commercially available. Atomisation in D.c. or A.c. Arcs It is not unexpected that the well established excitation methods of emission spectro-graphy also have been employed for atomisation purposes. Various ~orkers~l-73 have reported the direct analysis of solids by atomisation in d.c. or a.c. arcs the atomic absorption being measured in the gap between the electrodes. However as the temperatures in the arcs are considerably higher than in the flames and furnaces normally employed in AAS the ratio of neutral atoms to excited and ionised atoms is less favourable, in particular for the alkaline and alkaline earth elements. The d.c.and a.c. arcs are powerful devices for vaporising solid materials 1000 LANGMYHR DIRECT ANALYSIS OF SOLIDS Analyst Vol. 1U4 The higher sensitivity obtained for a number of heavy metals when measured by the absorption instead of the conventional emission method may be utilised in laboratories equipped with these instruments. The above instruments for the direct analysis of solids by atomisation in arcs are not equipped with lamps for background correction and non-specific absorption has to be compensated for by repeating the measurements at the wavelength of a non-absorbing line. In comparison with atomisation of solids in graphite tube cells the present technique cannot be seen to offer any advantages. From a survey of the literature on the subject it appears that there is little interest in the method.Atomisation by Cathodic Sputtering In various paper^,^^-^^ in particular by Australian workers cathodic sputtering has been shown to be an efficient technique for producing atomic vapour from metals and alloys. The method is also applicable to the analysis of non-conductive materials such as rocks and ores but it is then necessary to mix the sample with metal powder e.g. copper in order to make it conductive. The particles released from a previously cleaned cathode are usually ground-state neutral atoms that leave the cathode surface with high energy; thermal equilibrium of the sputtered atoms with the filler gas is rapidly established and in the vicinity of the cathode the sputtering and the diffusion rates are such that the concentration of atoms reaches a steady-state value.The equipment and procedures for carrying out analyses by cathodic sputtering are relatively simple ; when the surface has been properly processed the amounts sputtered are sufficiently high to prevent any serious sampling errors. The time required for processing the surface of the material to be sputtered may vary considerably, the cleanliness of the cathode surface and the purity of the filler gas are critical and the complex mechanism by which particles are released from the cathode is not known. These problems and the conflicting results reported in the literature have made analysts and manufacturers sceptical about the future of cathodic sputtering. The small number of papers that have been published on this subject in recent years reflects the present lack of interest in the technique.Further work is obviously required to ascertain its potentiality as a method for analysis by absorption. However it should be mentioned that the method is being increasingly used for solid sample analysis by emission. Atomisation by Laser or Discharge Lamps When a pulse of laser light is focused on a solid sample a cloud of vaporised material is formed that contains neutral and excited atoms ions and other species. The laser-produced plasma has a bright emission; however the free atoms outlast the emission period thus allowing AAS measurements to be made. Direct analysis of solids by laser atomisation has been described by a number of worker^.^^-^^ In the method in one of these papers,82 the metal sputtered from the sample is collected on the inside of a small graphite cylinder which is subsequently heated for atomisation and analysis.The samples analysed have mainly been solid metals or alloys and powders of inorganic materials pressed into pellets; however one papers7 also describes the analysis of various biological materials such as plants liver blood and muscle. The workers in this field have employed widely different equipment instrumental arrangements and measuring techniques; while some workers have obtained satisfactory analytical results, others have obtained only semi-quantitative data. The advantages of the technique are that it is rapid it is applicable to solid materials in almost any form it allows the determination of a large number of elements and it requires a minimum of sample pre-treatment.The disadvantages are that the technique cannot match the detection limits achieved by conventional atomisation in tube or crucible cells it requires careful standardisation for each matrix system the small amounts of sample vapourised may cause sampling errors and the high energy of the laser pulse causes eruption of liquid or solid particles from the samples. In addition to the use of lasers discharge lamps have also been employed for atomisation There are however a number of problems inherent in the present technique. Conclusions on the potentiality of the present technique are difficult to draw November 1979 BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 1001 of solids. However as the sensitivity of analysis with sources of continuous spectra is normally one or two orders of magnitude worse than that obtained with the line sources normally used the discharge lamps are at present of less importance as the primary light source.A special applications1 of a discharge lamp as the atomisation cell for solid samples should be mentioned. The external surface of a lamp was coated with a thin film of an oxide a flash of the lamp evaporated the film and the atomic lines of the spectrum were used to determine the content of impurities. Other Atomisation Methods The plasma formed by an electric discharge in a dielectric channel has a high temperature and pressure and may be used both as an atomisation cell for solid samples and as a source of a continuous spectrum.It has been demonstrateds2 that elements that are difficult to atomise such as silicon aluminium and titanium are easily evaporated from quartz glass. As the work carried out on the technique and its applications is limited it is difficult to evaluate its potentiality. It seems that the rapid evaporation of the sample and the turbulent ejection of the vapours may cause sputtering of solid or melted particles. Conclusion Today electrothermal atomisation is the preferred method for solid sample analysis and the discussion in the following sections will therefore be limited to the use of this type of atomisation cell. Atomisation of Air- and Water-suspended Solids Within the field of environmental analytical chemistry the analysis of suspended particles in the atmosphere and in water constitutes an important branch.The direct analysis of leadv3 and cadmiums4 in air particulates is possible by introducing air directly into a specially designed atomisation cell. However most analyses of air particulates are made by separating the solid particles by filtration either through a filter made of cellulose glass or a plastic material,v5-104 or through a porous graphite cup or ~ y l i n d e r l ~ ~ - ~ ~ * ; in the former method the determinations are made by transferring the whole or part of the filter into a suitable atomisation cell whereas in the latter technique the graphite filter also serves as the atomisation cell. A patentlm describes equipment that collects electrostatically solid particles in gases in a graphite tube and then transfers the tube into a heating circuit for atomisation.Equipment has also been developed for sampling air particulates in the field. Particulate matter in water may be analysed directly after f i l t r a t i ~ n . ~ ~ l l ~ The direct analysis of solids suspended in air and water can be made rapidly and requires only small amounts of sample. Reactions in Atomisation Cells A knowledge of the reactions that occur in atomisation cells is highly desirable in order to choose the most favourable conditions and to control and minimise the effects of interferences. Theoretical and experimental studies are complicated by the non-equilibrium conditions that often prevail during the reactions the high temperatures and the correspondingly high rates of reaction and the lack of reliable thermochemical data.The ensuing discussion will be limited to a short survey of the reactions that are likely to occur in atomisation cells made of carbon or metal. More comprehensive treatments have been published.6*111J12 Atomisation cells are normally equipped with power supplies that are controlled auto-matically e.g. by a microprocessor and allow the sample to be heated in at least four stages, viz. the drying ashing (or pyrolysis) atomisation and cleaning steps. In addition the analyst can select the time of heating the interval and rate of heating between the steps and the flow-rate of the purging gas. During the drying stage at about 370 K the main reaction will be the removal of non-essential water. However some highly volatile elements and metal-containing compounds, such as mercury and the organomercurials may be lost during this operation 1002 LANGMYHR DIRECT ANALYSIS OF SOLIDS Analyst Vol.104 The ashing stage involves one or more heating steps in the range from 620 K to above 1800 K. The main purposes of this operation are to destroy organic matter to remove essential water and interfering elements and compounds by evaporation or sublimation to transform carbonates sulphates nitrates hydroxides etc. into the corresponding oxides, or to transform the analytes into compounds that are less volatile and more convenient for the atomisation processes. In the analysis of certain inorganic materials such as metals and alloys the ashing step may be omitted. The volatility of the analyte or its compounds sets the upper limit of the ashing step.When volatile metals such as cadmium zinc and bismuth are to be determined the ashing temperature should not exceed 620 K. During low-temperature ashing arsenic selenium and tellurium may be completely or partly volatilised either in elemental form or as volatile compounds. Losses of these elements during ashing can be avoided by adding a metal solution e.g. of copper nickel, silver or molybdenum to the sample the metals forming thermally stable compounds with the ana1~tes.l~~ Certain metals such as molybdenum and vanadium allow ashing temperatures above 1800 K. The ashing temperatures recommended in the literature vary considerably with the type of cell employed and the material being analysed and it may therefore be necessary to establish the upper limit of the ashing temperature.During the atomisation step the cell is heated to temperatures in the range 1100-3300 K, the analyte is vaporised and the absorption of the free atoms is measured. The amplitude and shape of the peak are determined by the rate of the atomisation process. If the analysis is based on the measurement of peak heights the rate of free atom formation should be high for both sample and standard; this is normally obtained by maintaining a high rate of heating. However when peak areas are measured instead of peak heights the kinetics of the reactions lose their significance and assuming that the analyte is fully atomised and that all free atoms spend the same time in the light beam the shape of the signal may vary without influencing the analytical results.As recommended above direct AAS analysis of solids should preferably be based on the measurement of peak areas. The purpose of the fourth and final heating step is to clean the cell and is carried out by heating the cell for a few seconds at the maximum temperature. Any residue remaining in the cell should be removed by blowing or brushing; as the latter operation usually affects the interior surface of graphite cells it should be followed by a cleaning step. Thermochemistry Despite the possibility that equilibrium may not be attained in atomisation cells thermo-chemical calculations may still be of value for explaining the reactions in graphite and metal cells. These calculations are complicated by the large number of processes taking place in the solid and gaseous state and the complexity of the transport mechanisms.On the other hand the high temperatures limit the number of compounds in the gaseous phase to the simple thermally stable diatomic molecules. The tendency of some elements such as selenium tellurium bismuth and potassium to form polyatomic species has also to be taken into account. Various ~ ~ r k e r ~ ~ ~ ~ ~ ~ ~ ~ have related the experimentally established appearance tempera-tures of elements atomised in graphite cells with the temperatures at which the free energies of the reduction of the oxides with carbon to free gaseous metals were zero. The agreement between the temperatures was satisfactory for some elements but it was poor for other systems.In those instances where the appearance temperatures were higher than the temperatures at which the free energies were zero the discrepancies can be explained by assuming that at the latter temperatures the rates of reaction or the vapour pressure of the analytes were too low. L’vovl12 discussed the formation of thermally stable carbides pointing out that these reactions may play a more important role than had previously been assumed. On the basis of the heat of carbide formation and its relationship with the appearance temperature the experimentally determined and estimated appearance temperatures for 35 elements were compared. For many analytes the agreement was satisfactory; however for a number of systems in particular the alkali and alkaline earth elements there were considerabl November 1979 BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 1003 discrepancies.A possible reason for these differences is the formation of intercalation compounds i.e. the tendency of some elements in particular those which are easily ionised, to enter the interlamellar space in graphite and to form compounds with carbon. The complete decomposition of intercalation compounds requires prolonged heating at temperatures above 2300 K ; these compounds are less likely to be formed in cells made of glassy carbon or in graphite cells covered with a metal carbide or with pyrolytic graphite. Calculations of the free-energy changes occurring in the formation of stable carbide com-pounds show that metal oxides frequently react to form stable carbides a t temperatures below that at which the reduction of the metal oxide by carbon to gaseous metal atom occurs.Some metal oxides e.g. antimony(II1) oxide have relatively high vapour pressures at the appearance temperatures of the free gaseous metal atoms and some oxides may there-fore be lost by evaporation. In calculations of the thermal dissociation of oxides in a graphite environment the equilibria between carbon and oxygen also have to be taken into account. In atomisation cells made of metal of which the tantalum strip cell is a well known type, the most likely path to the formation of gaseous atoms is the thermal dissociation of metal oxides or other compounds. The reduction of a number of metal oxides with tantalum is thermodynamically feasible ; however in comparison with the dissociations the rates of reaction for the reduction of oxides in contact with the dense metal are probably low.When nitrogen is used as a purging gas for graphite atomisation cells CN absorption bands may be observed; at temperatures above 2300K the absorption of these bands increases exponentially with increasing temperature. In graphite cells purged with nitrogen some elements react with CN molecules to form cyanides,l12 and the formation of these compounds reduces the sensitivity of the determina-tions. By replacing nitrogen with argon the difficulties in the use of the former gas are avoided. Today argon is the preferred purging gas. From the above survey it appears that thennochemical calculations may explain many of the reactions that occur in carbon or metal atomisation cells; however for a number of systems the processes are not yet fully understood and further work is required.Kinetics In a studyll6 of the behaviour of copper in the range 1720-2220 K the atomisation processes were explained by assuming a slow first-order reaction involving reduction of copper oxide by carbon followed by rapid volatilisation of copper. In a series of papers117-11s other workers described the supply of the analyte by an Arrhenius-type rate constant; they relate the removal of the analyte to convection and diffusion. The studies were made with the use of a home-made open graphite rod cell and the data are therefore not directly applicable to the commercial graphite tube cells. The removal of analyte vapour from a graphite tube furnace has been found120 to occur by a loss (about 20%) through the aperture for the sample injection by diffusion through the tube walls (about 20%) and by diffusion to the cooler parts of the cell where condensation takes place.In a recent paper,121 previous contributions in the field are critically surveyed and a more advanced theory is introduced. The paper describes a study of the supply and removal of analyte vapours in graphite crucible and tube atomisation cells. In both cells the release of the atoms from the graphite surface was found to be determined by the temperature of the surface and was explained by an Arrhenius-type rate constant. The removal of the atoms from the cells varied with the type of cell employed. In the crucible cell diffusion was found to be the dominating process while in the larger tube cell operated under gas flow conditions convection is the dominating process.When the latter cell was operated under gas stop diffusion and to some extent expansion are the main factors in the removal process. From theoretical considerations and experiments it was demonstrated that under gas stop conditions only 25% of the sample is contained in the tube cell; under flow con-ditions this amount is reduced to less than 10%. It should therefore be possible to increase considerably the sensitivity of these determinations. The cited paper121 gives in a con-densed form an excellent review of the kinetics of the reactions in graphite atomisation cells. Papers on the kinetics of atomisation processes are limited.The atomisations were made in a graphite tube furnace 1004 LANGMYHR DIRECT ANALYSIS OF SOLIDS Analyst Vol. 104 Interferences The present technique is subject to various interferences that may introduce serious negative or positive systematic errors; a survey of commonly occurring interferences is given below. The main source of interferences is the matrix. During atomisation the matrix may produce atom molecule and radical interfering species and solid or liquid particles formed by sputtering from the sample by destruction products of inorganic or organic matter or by condensation of vaporised atoms or compounds in the cooler part of the atomisation cell. If the matrix produces atoms that absorb close to the absorption line of the analyte direct spectral interference occurs.A small number of such interferences have been reported122; however in most instances this interference can be avoided by making the measurements at another wavelength. Another type of spectral interference is encountered when the matrix gives off atoms with absorption lines that are not close enough to cause direct spectral interference but sufficiently close to be included in the spectral range of the slit (approximately 0.5 nm); these lines will absorb radiation from the background corrector lamp but not from the analyte light source, and consequently a negative correction (error) will be applied to the absorption signal of the analyte. The latter situation was recently documented123 in the determination of selenium in the presence of large amounts of iron.Iron has a number of resonance wave-lengths close to the most sensitive selenium line at 196.0 nm and these iron lines absorb emission from the deuterium background corrector but not from the selenium emission source. Again the interference can be avoided by selecting another selenium line. It was recently demonstrated12* that the interference of iron on the 196.0-nm selenium line can be avoided by carrying out the measurement with an instrument based on the Zeeman principle for background correction. The matrix may also produce molecules or radicals that absorb radiation; absorption spectra of molecules of commonly appearing salts oxides and hydroxides have been recorded by various workers.126 These molecules exhibit one or more broad maxima in the spectral range used for AAS measurements; in addition to the spectral continua they may also contain line-rich electronic spectra.While the background corrector compensates for the former interference the corrector does not correct for the fine structure of the molecular background. The broad-band spectra of the molecules of salts etc. will normally be compensated for by the background corrector system. However when large amounts of absorbing species are produced rapidly the consecutive measurement of the atomic absorption and of the background is not operating sufficiently rapidly and a negative signal is observed. This disturbance can be avoided by introducing a smaller sample and employing a lower rate of he at ing . The matrix will of course also affect the rate of the atomisation process and thus the amplitude and the shape of the peak.The rate of reaction may seriously affect analyses based on the measurement of peak heights whereas it does not affect data obtained from peak-area measurements. The scattering of radiant energy by very small particles (with a diameter less than one tenth of the wavelength of the radiation measured) can be expected to occur according to the Rayleigh theory Le. the scattering (s) is proportional to the fourth power of the inverse of the wavelength (A) (s = l/X4) and the effect of scattering is therefore much greater in the ultraviolet than in the visible range. Scattering by large particles can be expected to occur according to the Mie theory and in these instances only a slight wavelength dependency would be expected.Within the range of slit widths of most AAS instruments the absorption by scattering can be considered to be constant. Non-specific losses of radiation by scattering are normally compensated for by employing a background corrector. However the rapid formation of large amounts of scattering particles may exceed the compensating capacity of the corrector. In the analysis of biological and other materials the analyst is often faced with the problems associated with the presence of large amounts of halides in particular sodium chloride. The halides affect the measurements in various ways; in addition to the effect on the rate of atomisation and the non-specific absorption of molecules or particles the formation o November 1979 BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 1005 volatile gaseous monohalides of the analytes represents a serious source of interference.The problems associated with the formation of gaseous monohalides were discussed in a recent review.l12 The interference from chloride (and presumably other halides) can be removed or be compensated for by the following means (a) addition of sulphuric orthophosphoric nitric or ascorbic acid and removal of chloride by evaporation; (b) addition of ammonium nitrate solution and removal of chloride as the ammonium saltll3; (c) removal of sodium chloride by evaporation at about 1500 K ; ( d ) binding chloride in a compound that is more stable than the monochloride of the analyte112; (e) carrying out the measurements according to the method of standard additions; and (f) adding hydrogen to the purging gas and removing chloride as hydrogen chloride gas.The last method has been used successfully in the analysis of lead in steels,126 the separation being effected at 900-1OOO K. Samples Sampling Sampling Errors and Sample Preparation As is apparent from the bibliography on the use of the present technique (see below) a variety of solid materials has been analysed by the techniques surveyed under Atomisation Methods. In addition to the samples originally present as solids materials in the liquid state e.g. biological fluids can be analysed by transforming them into solids by drying, dry ashing plasma ashing or lyophilisation. These operations serve the purpose of con-centrating analytes present in amounts near or below the lower limit of determination.After being transformed into the solid state the material should be homogenised by grinding. It should be noted that certain elements and compounds such as mercury may be lost during these operations. Losses of volatile elements during the above operations can be avoided by adding a stabilising agent.l13 In most methods for the direct AAS analysis of solids the amounts of sample range from less than 1 mg and up to about 20 mg; some atomisation cells allow amounts of sample up to several hundred milligrams to be atomised. Whereas the determination of minor or trace elements in a sample weighing say 500 mg does not normally introduce any serious sampling errors the determination of the trace elements in samples with a mass of the order of 1 mg raises problems relating to the sampling errors and the risks of introducing contaminants.All sampling operations should of course, be made with the utmost care in order to obtain a representative and uncontaminated sample and to reduce the sampling errors to a minimum. Brittle inorganic substances such as minerals rocks ores salts and certain elements and alloys should not be crushed or pulverised in conventional equipment made of alloyed steel, but should be crushed or ground in agate corundum or carbide mortars and pestles. The particles of these samples normally can be readily ground manually to pass a 270-mesh (63-pm) sieve; further grinding to pass a 400-mesh (37-pm) sieve may take a considerable time.However with the use of automatically operated grinding equipment the particle size of inorganic brittle materials can be reduced to a considerably finer state of subdivision. In the author's laboratory,127 about 500 mg of the granite GH (a reference sample issued by Centre de Recherches Pdtrographiques et Geochimiques France) was pulverised for 3 h in an automatic agate mortar and pestle. The particle-size distribution (established with a Coulter Counter) is shown in Table I. From Table I it appears that about 86% of the particles were smaller than 8 pm. About 1% of the particles had a size in the range 3240 pm and this fraction may consist of biotite, the particle size of which is difficult to reduce. However the whole sample passed a 400-mesh sieve i.e.the maximum particle size was 37 pm. It should be noted that prolonged grinding may result in losses of volatile elements e.g., mercury from a silicate or sulphide matrix. Samples that have been in contact with metal sieves should not be used in the subsequent analysis and the portions of the sample taken for the sieve tests should be discarded. If the sample has to be sieved before analysis nylon-meshed plastic sieves should be employed. Brittle inorganic materials are usually analysed in the form of powders. The errors introduced by the sampling of powders for trace analysis vary considerably with the distri-bution pattern of the analyte. The trace element may either be evenly distributed in one or more of the main components of the matrix or be present as discrete particles disseminate 1006 LANGMYHR DIRECT ANALYSIS OF SOLIDS Analyst Vol.104 TABLE I PARTICLE SIZE DISTRIBUTION OF THE GRANITE GH Particle size/pm Content % <1.6 2.6 1.6-2.0 4.7 2.0-2.6 9.0 2.6-3.0 12.6 3.0-4.0 16.4 4.0-6.0 16.2 6.0-6.0 14.8 6.0-8.0 10.9 8.0-9.0 7.6 9.0-12.7 3.6 12.7-1 6.0 1.3 16.0-20.0 1.8 20.0-24.0 0.3 24.0-32.0 0.0 32.0-40.3 1.1 40.3-60.8 0.0 Sum 101.6 throughout the matrix; as would be expected the former instance gives a smaller sampling error. To demonstrate the order of magnitude of the sampling error under the conditions of the less favourable situation a method described in the literature1% was used to estimate the error for a distribution of mercury(I1) sulphide in a matrix of iron(I1) sulphide.l29 In Table I1 the relative standard deviation of the content of mercury is given for 5-mg samples con-taining 1 3 10 and 30 p.p.m.of mercury and for various particle sizes. As is apparent from Table 11 the sampling error may be so large under combinations of unfavourable conditions that reliable analytical data are unobtainable. TABLE I1 SAMPLING ERROR FOR A DISTRIBUTION OF MERCURY(II) SULPHIDE IN A MATRIX OF IRON(II) SULPHIDE Particle ASTM diameter/ mesh CCm number 108 140 63 270 37 400 20 10 --Approximate number of particles per 6 mg of sample 1.6 x lo9 2.4 x lo6 1.9 x 106 1.3 x 104 3.8 x 104 Relative standard deviation of the Hg content % for samples containing : 1 p.p.m. 3 p.p,m. 10p.p.m. 30p.p.m. I A I 232 134 73 43 81 47 26 16 48 27 16 9 19 11 6 3.6 6.7 3.9 2.1 1.2 However when the samples are ground to a very fine state of subdivision the sampling error under the conditions of an unfavourable distribution pattern may be reduced to such low levels that the sample size can be reduced to below 1 mg.This is exemplified by the direct AAS determination of copper in granite GH.12' The particle size distribution of this sample after grinding is shown in Table I. It was assumed that copper was present in the granite as discrete particles of chalcopyrite (CuFeS,) that all particles were present as spheres having an average diameter of 5 p m and that the specific gravity of the sample was 2.7. The number of particles per milligram was calculated to be 5.7 x lo6 and a 0.2-mg sample therefore contains about 106 particles.Using the above method for estimating the sampling error and a sample size of 0.2 mg the relative standard deviation of the content of copper was calculated to be 20%. This is a large error; however as the number of particles is probably considerably higher than lo6 the error obtained in actual analysis may prove to be smaller. The content of copper in granite GH was then determined by direct AAS; from 1 + 1 mixtures of granite and graphite powder six portions corresponding to amounts of granite ranging from 0.16 to 0.34 mg were transferred into a graphite tube furnace and copper wa November 1979 BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 1007 determined by measurements against the National Bureau of Standards (NBS) Standard Reference Material No.614 (Trace Elements in Glass). The average content of copper was found to be 13 p.p.m. with a relative standard deviation of 16% ; the latter is substantially lower than the above estimated value. The recommended value for copper in granite GH is 14 p.p.m. It can be concluded that the error associated with the sampling of finely powdered inorganic materials is acceptable for amounts of sample down to fractions of 1 mg. Ductile metals and alloys are normally sampled by drilling. The sampling error may again be serious when the element to be determined is present as or in irregularly distributed particles. However as is apparent from the data from applications (see below) direct AAS analysis has been used successfully in the determination of a number of trace constituents in irons and steels; in these analyses the amounts of sample ranged from 1 to 12 mg.Recent studieslS0 of the distribution of lead in mild steel stainless steel a nickel-base alloy and ferromolybdenum and of antimony in mild steel demonstrated that the two trace metals are evenly distributed and that amounts of sample of 2 mg can be taken without introducing any serious sampling errors. In the sampling and treatment of solid materials of human animal and plant origin the diversity of the samples and the widely differing properties of the analytes make it difficult to give any general procedure. The amount of material available is often small and the elements to be determined may be inhomogeneously distributed and consequently sampling errors may occur.Many samples of biological origin can be homogenised by grinding in the wet state; the grinding equipment should be made of materials that do not contaminate the sample. Samples of suspensions and tissues can be homogenised by solubilisation in quaternary ammonium bases; it should be noted that these reagents may contain appreciable amounts of contaminants. Standards Standardisation and Techniques of Measurement The direct AAS analysis of solids is normally based on measurements against suitable standards; in principle it is possible to carry out “absolute” analyses but this possibility does not seem to have been applied to the AAS analysis of solids. The following standardisation methods have been employed measurements against solid standard samples of natural materials or industrial products; measurements against synthetic solid standards; measurements according to the standard additions technique ; measurements against standard solutions.When reliable solid standards are available with certified values for the elements to be determined and with matrices corresponding to those of the samples measurements against such materials are to be preferred. The selection of such standard samples in particular for the analysis of trace components is limited with respect to both materials and elements; however the selection of such materials is steadily increasing. Lists of suppliers of reference and standard materials have been publi~hed.~$l~~ In the certificates of analysis such as those issued by the NBS the certified values are based on a sample size of at least 260 mg.When these samples are employed as standards in the direct AAS analysis of solids the amount taken for atomisation will often be of the order of 1 mg and the sampling error has to be taken into consideration. However as is apparent from the precision obtained from the use of various NBS standard reference materials the sampling error for a number of analytes does not seem to be serious for amounts of sample of the order of 1 mg. When no suitable solid standard is available it is possible to use synthetic solid standards. Such standards are produced commercially with a graphite or silica matrix; they can of course also be prepared from pure substances and with a matrix resembling that of the sample.However the preparation of these samples is tedious and time consuming. It is also possible to apply the standard-additions technique by adding standard solutions to a series of accurately weighed solid samples of approximately the same mass. In these analyses it is recommended to add constant volumes of the standard solutions to let the standard solution soak into the sample and to establish the position of the standard-additions graph by the method of least squares 1008 LANGMYHR DIRECT ANALYSIS OF SOLIDS Analyst VoE. 104 In a few instances it has been found possible to analyse solids by measurements against calibration graphs established by atomising aqueous standard solutions. The standard solutions should be prepared from high-purity metals or compounds (preferably dissolved in nitric or sulphuric acid).In these instances the sample and the standards have a different matrix composition the atomisation mechanisms will be different and the absorption signals are also likely to differ. The measurements should again be based on the addition of equal volumes of the standard solutions and on the measurement of peak areas instead of peak heights. The use of some of the above standardisation techniques is illustrated by the data listed in Table 111 obtained from electrothermal atomisation and AAS determination of cadmium, manganese and lead in hydroxyapatite and two samples of animal bone.182 TABLE I11 RESULTS FOR CADMIUM MANGANESE AND LEAD IN HYDROXYAPATITE AND TWO SAMPLES OF ANIMAL BONE Sample Method* Hydroxyapatitet .A B Animal bone IAEA . . Values recommended by IAEA C A D Animal bone Weider C A D %t p.p.m. 0.33 0.34 Cd 1 Srr 18 18 % Mn - 2 S r l P.P*m* % 0.94 3 0.84 6 Pb - 0 Srt P.P.m. % 2.6 12 2.5 12 Not given 0.072 - ---0.036 -12 --22 32 16§ 33 12 32 11 6.4 20 0.0 22 - -- -6.8 16 6.9 15 0.8 18 9.8 14 9.7 10 - -- -* Method A (standard addition) standard solution is added to the sample solution with atomisation Method B (standard addition) standard solution is added to the solid Method C (standard addition) standard solution Method D solid sample is atomised in the graphite furnace. sample with atomisation in the graphite furnace. is added to the sample solution with atomisation in the flame.in the graphite furnace with solid hydroxyapatite as standard. t Average result with relative standard deviation (sc). 1 The sample employed as the solid standard. tj Relative standard deviation of the mean value. The content of an element in a reference material that has a suitable matrix composition can of course also be established by analysis. Trace element contents can often be deter-mined by decomposing the sample and carrying out the analysis by a conventional flame AAS method by atomising the sample solution in a furnace or by any other suitable method. These important determinations should be made by at least two reliable methods. The techniques of measurement may vary from the plotting of a full calibration graph before atomising the sample to the alternating measurement of standards and samples.As the sensitivity may vary during the measurements the latter approach is to be preferred. By carrying out at least five determinations of the analyte statistical methods e.g. for the rejection of data can be applied. Accuracy and Precision A number of systematic errors may affect the accuracy of the present technique and no estimate of the errors can be given. The most serious sources of systematic errors are the losses of the analyte by evaporation or sputtering during the drying and/or ashing steps incomplete atomisation chemical and spectral interferences the inability of the background corrector to compensate for large non-specific absorptions different rates of atomisation for the sample and standard (when peak heights are measured) varying times of residence of the atoms in the cell incorrec November 1979 BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 1009 adjustment of the sources of radiation swelling of the sample during drying ashing or atomisation the use of slow signal measuring systems for rapid signals and incorrect values assigned to the standard samples.The accuracy should of course always be established by analysing a reliable standard with the same or a similar matrix composition; the number of analyses should be sufficient to apply statistical methods such as the t-test to decide whether or not systematic errors are present. The precision is established by performing a series of analyses and assuming that the errors are random and have a normal distribution calculating the standard deviation.These calculations are normally made without establishing the actual distribution ; it should be noted that analytical results for trace elements may exhibit other distributions such as a log-normal distribution. In the direct analysis of solids the precision is mainly affected by the errors associated with the sampling and the absorption measurements. Relative standard deviations of 5-10% are frequently obtained for elements present at the 1 p.p.m. level; similarly at the 1 p.p.b. (parts per log) level values in the range l0-30% have to be considered as normal. Time of Analysis Assuming that the instrument is ready for measurement and that the sample has been prepared for analysis approximately ten solid samples can be analysed per hour.Contamination Control The direct AAS analysis of solids has the important advantage of not requiring any decomposition separation and/or concentration steps; compared with many other analytical methods the risks of introducing contaminants are considerably reduced. However the sample may take up impurities during the sampling and sample preparation steps such as crushing grinding sieving and homogenisation from the ambient atmosphere and from contact with glass plastic materials metal knives and spatulas. To avoid or reduce the risks of introducing metallic contaminants it is recommended to employ whenever possible, thoroughly cleaned non-metal equipment such as agate mortars and pestles nylon-meshed sieves plastic sample bottles and spatulas and silica knives.The impurities from the ambient air may be reduced by working in a glove-box fitted with filters or in a laboratory kept under a constant positive pressure of filtered air. For more comprehensive treatments of the contamination problems in trace analysis, reference is made to a monograph133 and a recent review paper.lM Applications In Table IV a bibliography is given of the applications of the direct AAS analysis of solids; in this list of applications all atomisation methods have been considered. The applications are classified according to the composition of the matrix the first group consisting of materials with an inorganic matrix and the second listing substances with an organic matrix. TABLE IV APPLICATIONS Refer-Materials analysed Elements determined ences Materials analySea Elements determined A .Mat& wifh a- inorganic matrix Acids (see Salts) Airsuspendedmatter Pb Pb Be Pb - Air suspended matter (cont.) Cd Pb 93 Ph _. 10s 106 9s cd Cd Cu Mn Pb Pb Pb 97 Alumina (see Aluminium Pb 96 oxide) -Ag Be Cd Hg Pb Se 107 Aluminium Cr Pb 108 Ca Cd 98 Zn Pb 100 Cu Mg Zn Cd Cu Pb Zn 101 T1 Be 104 cu Refer-ences 09 102 94 103 136 138 137 138 139 140 4 1010 LANGMYHR DIRECT ANALYSIS OF SOLIDS Analyst Vol. 104 TABLE IV-continued Refer-ences Refer-ences 138 77 Materials analysed Aluminium(cont.) . . . . Aluminium-base alloys . . Elements determined Materials analysed Iron cont). . . . . . . Elements determined Zn Cr Cu Fe Mn Ni Si Ti cu 141 77 85 Bi c cu Mn Mo Ni Si V BI 162 163 164 Zn Zn Cu Mg Mn Cu Mn Cu Ni c u 138 141 36 9 0 7R 80 67 69 35 137 9 142 i% Fe cu In Iron ores .. . . . . Iron oxide . . . . . . Laterite . . . . . . Lead . . . . . . Lead-base aliok . . . . Leadores . . . . . . Lithium compounds (see Magnesite brick : :: Magnesiumoxide . . . . Manganeseore . . . . Mercury ores Metal oxide films (see Tin ’ ‘ oxide films) Molecular sieve . . . Molybdenum . . . . Molybdenum oxide . . . . Nickel-base alloys . . . . Salts) 16 67 137 136 161 138 16 Aluminium oxide . . . . Ag Cd Cu Sb Zn Bi Bi ~~ c u Cu Fe Mg Co. Mo Pb G i Pd Zn -Ti cu cu -2 67 80 69 2 16 Antimony Apatite (see Hydroxyapatitk and Phosphate rock amcentrates) .. . . Ash. * . ~ m u c a r G l ; a t e . . . . Bases (seesalts) . . . . Bauxite . . . . Born nitride’ . . . . Brass . . . . . . co Ti Hg -Pb c u Ga In Cd Cu Ag Be Cd Fe Mn Ni, Fe Ni Cu Ni Cu Ni A1 Pb Zn T1 c u -Pb Sb --143 67 137 144 77 90 78 78 --Pb -9 166 69 88 130 166 166 82a 78 166 167 69 168 83 68 69 9% 169 170 80 60 171 --71 172 173 174 176 8 176 Al Zn Ca Cd Cu Eu Mg Zn Fe Pb Bronze . . . . . . Bi Pb SC Sn Te T1 Bi Pb Zn Ni 82 82a 146 80 Cadmium . . . . . . Calcium oxide Carbon (see Graphite) : Cast iron (see Iron) .. catalysts . . . . . . Chromiumoxide ._ Clay . . . . . . . . Coal . . . . . . . . Nickel compounds . . . . Niobium . . . . . . Niobium oxide . . Phosphate rock concentrai& Quartz . . . . . . Al Zn Si Al Cu Fe Ag Bi Cd In Pb T1 Zn cu --9 36 84 67 137 146 145 148 69 149 76 160 44 151 77 152 16 44 88 --a 69 67 153 153 69 84 -Pd Co Mo c u c u Ga In Cd Be Ca Fe Na Ni Pb Cd Cu Pb Zn Cu NI V Mn Zn P Fe Ni Si Zn Cd Pb Zn c u Fe K Na Na Ti co 2 --Au Au Quartz amorphous Quartz glass . . Rare earth oxides . . Rocks (see Silicates) Salts acids bases . . Eu Ti Ca Cu Fe K Mg Mn, Na ‘Li ‘Li c u Al Co Cu Fe Na Ni, Pb Zn -2; 2 sn TI copper .. . . . . Sediments (see also Silicates) . . silica (see Silicon &.ihe) Silicates (see also Sediments) Copper ores . . . . Copper slags (see Slags) . . corundum . . . . . . Crayons (see Paints) . . Dolomite . . . . . . Perrites . . . . . . Ferromanganese . . . . Ferrosilicon . . . . . . Fly ash (see Ash) . . . . Gadoliniumoxide . . . . Glass (see also Quartz) . . 3 Cr Cu Ni F% Cd cd CU Cd Pb Cd Pb 10 11 177 84 12 178 83 179 13 69 14 180 66 181 127 143 69 69 80 66 66 69 2 127 69 78 182 183 184 186 -Ca cd Cu Eu Mg Zn Ag Cu Mn Zn Cu Mn Zn Cu Ni Pb Au Au Cs Rb c u Ag Cd W T1 Zn Pb 164 127 143 Goldores . . . . . . Gold ore tailings .. . . Graphite . . . . . . 68 17 68 166 79 71 co Cs Li Rb cd Cr Cu Cs Li Rb c u Au Al. Zn Ag Ca Cu Ag’ Bi’cd Cr c u Mu &o. bb. )Sb.’Sn.kl. i n Pb Fe Fe c u Pb c u Feihin Pb ’ ’ . 144 9 Silicon carbide . . Silicon nitride . . Silicon oxide. . . . Pd As Cd Cu Fe Hg Pb, Se. Zn 156 86 167 168 68 69 169 132 160 69 161 Mn Ag Au Bi Cd H g Mu, hi $b gb i n .ri 27 elements Ag Bi Cd Cr Cu Fe, Mn Pb Cd Mn Pb Al Pb Bi As: ~~ cd c u P. v Slag . . . . . . Sludge . . . . CU Cd Cu Pb cd Cu Pb Cr Ni Zn Cd Cr Cu Ni, Al Bi Ca Cd, - Fe Mg fi, Pb Zn Pb Zn a cu Hpdroxyapatite . . . . Iodineoxide . . . . Inm . . . . . . . . Soil . . . . . . Steel . . . . .. 70 186 81 Cr Gr. Mn Ni S November 1979 BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 101 1 TABLE IV-continued Refer- Refer-Materials analysed Elements determined ences Materials analysed Elements determined e n c e Steel (cont.) . . . . . . Cr Mn Ni Bi 46 161 89 141 138 78 162 90 187 188 163 189 164 130 190 Leaves (see also Plants) 215 204 01s 219 146 59 164 220 415 146 204 221 Al Al Sb Sn Al Sb Sn Cr Cu Fe Ni Bi Al Cr Mo V Ag Bi Pb Zn co Pb ~~ Cd Cd Cu Pb Zn Be Al Cu Fe Mn Pb Sr, H g Cd H g Ag Cu Mn Pb Se Pb Zn Be Cd Cu Pb Zn cd Zn Liver . Al3 Cr Cu Mn Ni Pb Pb Ag Bi Cd Zn Pb Zn Ag Bi Cd I n Bi Cd Cu In Pb Sb, Te TI In Fe c u Al Zn Cr Cu Fe Mn Ni Si Fe Sn Sb Al Pb Cu Fe Mn Pb Cu Ni Al Zn Cd Li Co Cr Cu Mn Ni Cd Mg Mn Cd 1 B b I J 1%3U rrqj arru 113 81 154 59 82a 191 137 Sulphide ores .. . . 216 57 169 205 220 217 216 222 217 223 87 224 225 226 227 228 229 2 3.0 231 232 233 54 234 235 236 Pb Pb Cd Cu Mn Pb Cu Fe Mn Pb Zn Pb Cd Pb Ph Co Cr Cu Pb Zn Ag Bi Co Cu Fe Mn, Al Cr Cu Pb Zn c u c u Cu Zn Pb Pb Cd Pb Pb Cu Fe Mn Si Cd Cu Mn Ph Cu Pb Cu Fe Ag cd Co Cr Cu Fe, Co Cr Cu Fe Mn Ni Ni Pb Zn Mg Mn Pb Sb 192 137 69 80 155 193 69 37 91 82 194 65 127 155 195 196 197 197a 136 Sulphide ore concentrates . . Sulphur . . . . . . Synthetic mixtures.. . . Tantalum . . . . . . Lung . . . . . . Tin ores Tinoxide fi& 1 :: Titanium . . . . Titaniumoxide . . . . Molluscs . . . . Muscle . . . . Mussels . . . . Nails . . . . . . Tungsten . . . . . . Uranium . . . . . . Uranium oxide . . . . Zinc . . . . . . . Paint . . . . Paper . . . . 198 161 138 _ _ cd Cd Cd Cr Cu Fe Mn Ni Pb. Si 77 198 138 67 199 144 69 Zinc-base alloys . . . . Zinc oxide . . . . . . Zirconium . . . . . . Zirconium oxide . . . . Cd Cd c u La Y Bi Cd Pb Sb Sn Sb Plankton . . Plants (see also Grass, Leaves Tomatoes and Wheat) . . . . 87 70 70a 237 232 68 54 238 239 240 241 242 232 233 239 243 244 213 220 220 245 246 160 203 207 140 72 110 207 ---B.Materials with an organic m Algal cells . . . . . . Amoeba Blood-lyophiiised ashed ' . (seealsoSerum) . . . . Bone . . . . . . . . Dental material (see Teeth) Fibres (see Silk) . . . . Fish . . . . . . . . Fish meal . . cattix Mn Mn Polymers . . . 200 201 202 160 132 --203 204 205 204 206 207 208 209 210 211 -Au Al Cu Fe Au Cu Pb Cr Cu Fe Ti Si V Ag Zn Cd Mn P b Pb co P Co. Mn. Zn Pulp . . . . . . Resin (see Polymers) Rubbex . . . . Seaweed Serum ashed . : Silk . . . . . . Silkworm . . . . Silkworm egg . . Skin . . . . . . Teeth . . . . Textiles (see Silk) . . Tissue . . . . Tomatoes (see also Plant Urea . . . . Water suspended matter Wax crayons (see Paint) Wheat (see also Plants) Cu Fe Mn Si Cd Cu Mn Pb -Pb Si Sn Pb Grain (see Wheat) .. . . Grass (see also Plants) . . Hair . . . . . . . . -c u Pb co Ag Cu Fe Pb Cu Fe Mn Zn Cu Fe Mn Zn Cu Pb Ag AI Co Cr Cu Fe, Mn Xi Si Pb Ag '41 Bi Co Cr Cu, Fe Mg Mn Ni Pb, Si Zn Ag Co Cu Fe Ni Pb, Co Cu Fe Mn Pb Cr Cd Pb Ag Mn Zn 212 Si 213 214 132 215 216 21 7 Co Mn Zn c u TI Ag Cd Mn Pb Pb Ivory . . . . . . Kidney . . . . . . Hg Cd Pb -c u The appendix gives references to some studies that were found to give Iess reliable or only semi-quantitative results or which proved to be unsuccessful. Appendix The literature contains descriptions of atomisation cells which according to the a ~ t h o r s ~ ~ ~ ~ ~ are applicable to the direct analysis of solids; however in these papers no data relating to applications are given 1012 LANGMYHR DIRECT ANALYSIS OF SOLIDS Analyst Vol.104 AAS was usedlM to examine various grades of spectroscopic carbon in the form of rods, and arsenic cadmium copper iron mercury lead selenium and zinc were detected. The impurities were found to be distributed throughout the matrix as well as on the surface. By heating the rods at 1570 K for varying periods of time the content of impurities could be reduced; however cleaned rods were rapidly re-contaminated by absorption of metal impurities from the ambient atmosphere. For the atomisation of volatile elements microcrucibles made of platinum or graphite have been employed.8 The equipment was tested by atomising cadmium from a sample of syenite ; however analytical results were not given.Calibration graphs were reported for silver and gold in mixtures with graphite but no data were given for applications to inorganic and/or organic materials. Atternpt~2~~ at the direct AAS determination of cobalt in feed grains and forages were unsuccessful the reasons being stated to be a large non-specific absorption of radiation and other interference problems. Solid samples of polymers and filter-paper discs were atomised in an inverted T-type of furnace.24e The elements detected or determined in these samples were not reported. The direct determination of aluminium and vanadium in lyophilised and ground powders of bovine liver and of cat and rabbit brain has been attempted2S0; however the accuracy and precision obtained were unsatisfactory .A carbon dioxide laser has been employedse to atomise silver from copper-base alloys but the method was found to give only semi-quantitative results. The content of chromium in brewer’s yeast has been determined by the solid sampling technique261; however the results were much lower than those from samples wet ashed in a closed vessel. The former analysis included ashing at temperatures up to 1620 K and it seems highly probable that chromium was lost at this high temperature. An a.c. arc has been used73 as an atomiser for AAS. Conclusion From the applications listed in Table IV it appears that 40 elements have been determined by the present technique the concentrations of the analytes ranging from more than 100 p.p.m.to about 0.1 p.p.b. (lo9). All types of solid materials can be analysed the only limitation being that the sampling error should be acceptable for amounts of sample of the order to 1-10 mg. Direct analyses of solids have been made with a wide variety of equipment; a number of atomisation cells were home-made and of special designs and the instrumental arrangements have also varied considerably. The most useful and extensively employed atomisation cell for solids is the simple cylindrical graphite tube. The direct AAS analysis of solids has considerable potentiality as a general method for trace analysis and will be of particular value in those fields such as medicine biology and environmental chemistry where only small amounts of sample are available.1. 2. 3. 4. 6. 6. 7. 8. 9. 10. 11. 12. 13. 14. References Walsh A. Spectrochim. Ada 1955 7 108. L‘vov B. V. Talanta 1976 23 109. Razumov V. A. Zh. Prikl. Spektrosk. 1976 24 1117. Wanninen E. Editor. “Analytical Chemistry. Essays in Memory of Anders R.ingbom,” Pergamon Langmyhr F. J. Talanta 1977 24 277. Fuller C. W. “Electrothermal Atomization for Atomic Absorption Spectrometry,” Analytical “Annual Reports on Analytical Atomic Spectroscopy,” The Chemical Society London. Prudnikov E. D. Zh. Prikl. Spektrosk. 1972 17 352. Coudert M. A. and Vergnaud J. M C.R. Hebd. Skanc. Acad. Sci. Paris 1969 268 1225; Analyt. Govindaraju K Mevelle. G and Chouard C. Chem. Geol. 1971 8 131.Govindaraju K. Mevelle G. and Chouard C. Bull. SOC. Fr. Ckram. 1972 96 47. Govindaraju K. Herman R. Mevelle G. and Chouard C. Atom. Absorption Newst. 1973 12 73. Govindaraju K. Mevelle G. and Chouard C. Analyt. Chem. 1974. 46 1672. Govindaraju K. Analusis 1976 3 164. Press Oxford 1977 p. 461. Sciences Monographs No. 4 The Chemical Society London 1977. Chem. 1970 42 1303 November 1979 BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 1013 16. 16. 17. 18. 19. 20. 21. 22. 23. 24. 26. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 61. 52. 53. 54. 66. 66. 67. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 70a.71. 72. 73. 74. 76. 76. 77. 78. Venghiattis A. A. Atom. Absorption Newsl. 1967 6 19. Venghiattis A. A. and Whitlock L. Atom. Absorption Newsl. 1967 6 136. Venghiattis A. A. Spectrochim. Acta 1967. 23B 67. Miaud P. and Robin J. Bull. Soc. Chim. Fr. 1968 No. 2 864. Sprague S. and Slavin W. Atom. Absorption Newsl. 1963 2 20. Sprague S. and Slavin W. Atom. Absorption Newsl. 1966 4 367. Burrows J. A. Heerdt J . C. and Willis J . B. Analyt. Chem. 1966 37 579. Means E. A. and Ratcliff D. Atom. Absorption Newsl. 1966 4 174. Slavin S. and Slavin W. Atom. Absorption Newsl. 1966 5 106. Moore E. J. Milner 0. I. and Glass J. R. Microchem. J. 1966 10 148. Mostyn R. A. and Cunningham A. F. J. Inst. Petrol. 1967 53 101. Schallis J. E. and Kahn H. L. Atom.Absorption Newsl. 1968 7 84. Bartels T. T. and Slater M. P. Atom. Absorption Newsl. 1970 9 76. Kriss R. H. and Bartels T. T. Atom. Absorption Newsl. 1970 9 78. Kahn H. L. Peterson G. E. and Manning D. C. Atom. Absorption Newsl. 1970 9 79. Jackson D. R. Salama C. and Dunn R. Can. Spectrosc. 1970 15 17. Taylor J. H. Bartels T. T. and Crump. N. L. Analyt. Chem. 1971 43 1780. Lacour A. Teinturier C. and Isabelle D. B. Me’thodes Phys. Anal. 1971 7 49. Barnett W. B. Kahn H. L. and Peterson G. E. Atom. Absorption Newsl. 1971 10 106. Jackson K. W. Aldous K. M. and Mitchell D. G. Appl. Spectrosc. 1974 28 669. Kashiki M. and Oshima S. Analytica Chim. Ada 1970 51 387. Ghiglione. M. Eljuri E. and Cuevas C. APPZ. Spectrosc. 1976 30 320. Harrison W. W. and Juliano P. O. Analyt.Chem. 1971 43 248. Willis J. B. Analyt. Chem. 1976 47 1752. Fuller C. W. Analyst 1976 101 961. L’vov B. V. “Atomic Absorption Spectral Analysis,” Second Edition Adam Hilger Bristol 1970. Lebedev V. I. Zh. Analit. Khim. 1969 24 337. Bastiaans G. J. and Hieftje G. M. Analyt. Chem. 1974 46 901. Panichev N. A. Prudnikov E. D. Tatti A. V. and Turkin Y. I. Zh. Prikl. Spektrosk. 1975 22, Kantor T. and Pungor E. Paper presented at the 17th Colloquium Spectroscopicum Internationale, Human H. G. C. Scott R. H. Oakes A. R. and West C. D. Analyst 1976 101 265. Winge R. K. Fassel V. A. and Kniseley R. N. Appl. Spectrosc. 1971 25 636. King A. S. Astrophys. J. 1908 27 363. Maller L. A. Annln Phys. 1927 82 39. L‘vov B. V. J. Eng. Phys. 1969 2 44. Kirkbright G. F. Analyst 1971 96 609.Szydlowski F. J. Peck E. and Bax B. Appl. Spectrosc 1978 32 402. Fritzsche H. Wegscheider W. Knapp G. and Ortner H. M. Talanta 1979 26 219. Woodriff R. Stone R. W. and Held A. M. Appl. Spectrosc. 1968 22 408. Robinson J . W. Wolcott D. K. and Rhodes L. Analytica Chim. Acta 1976 78 286. Talmi Y. and Morrison G. H. Analyt. Chem. 1972 44 1466. Headridge J. B. and Smith D. R. Talanta 1972 19 833. Hadeishi T. and McLaughlin R. D. Analyt. Chem. 1976 48 1009. Katskov D. A. Kruglikova L. P. L’vov B. V. and Polzik L. K . Zh. Prikl. Spektrosk. 1974 20, Nichols J. A. Jones R. D. and Woodriff R. Andyt. Chem. 1979 50 2071. Brodie K. G. and Matougek J. P. Analyt. Chem. 1971 43 1667. Omang S. Analytica Cham. Ada 1971 56 470. Alder J. F. and West T. S Analyfica Chim.Acta 1972 58 331. Reeves R. D. Molnar C. J. and Winefordner J. D. Analyt. Chem. 1972 44 1913. Reeves R. D. Molnar C. J. Glenn M. T. Ahlstrom J. R. and Winefordner J . D. Analyt. Chem., See ref. 39. Fuller C. W. and Thompson I. Analyst 1977 102 141. Borzov V. P. L’vov B. V. and Plyushch G. V. Zh. Prikl. Spektrosk. 1969 11 217. Men’shikov V. I. Malykh V. D. and Shestakova T. D. Zh. Andit. Khim. 1974 29 2132. Katskov D. A. Kruglikova L. P. and L’vov B. V. Zh. Analit. Khim. 1976 30 238. Razumov V. A. Zh. Analit. Khim. 1977 32 383. Jackson K. W Ebdon L. and Webb D. C. Paper presented a t the 21st Colloquium Spectro-Belyaev Y. I. Ivantsov L. M. Karyakin. A. V. Pham Hung Phi and Shemet V. V. Zh. Analit. Belyaev Y. I. and Gordeev V. V. Oceanology 1972 12 766.Malykh V. D. Men’shikov V. I. Morozov V. N. and Shipitsyn S. A. Zh. Prikl. Spekfrosh. 1972, Russel B. J. and Walsh A. Specfrochim. Acfa 1969 15. 883. Gatehouse B. M. and Walsh A. Spectrochim. Acta 1960 16 602. Walsh A. Appl. Spectrosc. 1973 27 336. Gough D. S Analyf. Chem. 1976 48 1926. McDonald D. C. A d y f . Cbm. 1977 49 1336. 391. Florence 1973. 739. 1972 44 2206. scopicum and 8th International Conference on Atomic Spectroscopy Cambridge 1979. Khim. 1968 23 508. 16 12 1014 LANGMYHR DIRECT ANALYSIS OF SOLIDS Analyst Vol. 104 79. 80. 81. 82. 82a. 83. 84. 86. 86. 87. 88. 89. 90. 91. 92. 93. 94. 96. 96. 97. 98. 99. 100. 101. 102. 103. 104. 106. 106. 107. 108. 109. 110. 111. 112. 113.114. 116. 116. 117. 118. 119. 120. 121. 122. 123. 124. 126. 126. 127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139. Mossotti V. G. Laqua K. and Hagenah W.-D. Spectrochim. Acta 1967 23B 197. Karyakin A. V. and Kaigorodov V. A. Zh. Analit. Khim. 1968 23 930. Krivchikova 8. P. and Demin V. S. Zh. Prikl. Spektrosk. 1971 14 592. Nikolaev G. I. and Podgornaya V. I. Zh. Prikl. Spektrosk. 1972 16 911. Thomerson D. R. and Godden R. G. Baird-Atomic Ltd. Braintree Essex personal communication. Karyakin A. V. Pchelintsev A. M. Shidlovskii A. I. Vul’fson E. K. and Tsingarelli M. N. Zh. Vul’fson E. K. Karyakin A. V. and Shidlovskii A. I. Zh. Analit. Khim. 1973 28 1253. Osten D. E. and Piepmeier E. H. Appl. Spectrosc.1973 27 165. Vul’fson E. K. Karyakin A. V. and Shidlovskii A. I. Zav. Lab. 1974 40 945. Vul’fson E. K. Gribovskaya I. F. Karyakin A. V. and Yanushkevich A. F. Zh. Analit. Khim., KQntor T. Pdos L. Fodor P. and Pungor E. Talanta 1976 23 585. MatonSek J. P. and Orr B. J. Spectrochim. Acta 1976 31B 475. Ishizuka T. Uwamino Y. and Sunahara H. Analyt. Chem. 1977 49 1339. Kaigorodov V. A. Zh. Prikl. Spektrosk. 1975 23 389. Kaigorodov V. A. and Grishko G. A. Zh. Analit. Khim. 1977 32 205. Loftin H. P. Christian C. M. and Robinson J . W. Spectrosc. Lett. 1970 3 161. Christian C. M. 11 and Robinson J . W. Analytica Chim. Acta 1971 56 466. Matousek J. P. and Brodie K. G. Analyt. Chem. 1973 45 1606. Amos M. D. Brodie K. G. and Matoulek J. P. “Proceedings of the Pittsburgh Conference on Roques Y.and Mathieu J. Analusis 1973 2 481. Brodie K. G. and Matougek J. P. Analytica Chim. Acta 1974 69 200. Brodie K. G. and Matoulek J . P. Boll. Lab. Chim. Prov. (Bologna) 1975 1 279. Noller B. N. and Bloom H. Atmos. Environ. 1975 9 505. Noller B. N. Bloom H. and Parker C. R. “Proceedings of the 5th International Conferenceon Atomic Spectroscopy Melbourne Australia 1976,” Australian Academy of Science Dominion Press Victoria 1976. Noller B. N. and Bloom H. Analyt. Chem. 1977 49 346. Thomassen Y. Solberg R. and Hanssen J. E. Analytica Chim. Acta 1977 90 279. Havezov I. and Tamnev B. 2. Analyt. Chem. 1978 291 127. Woodriff R. and Lech J. F. Analyt. Chem. 1972 44 1323. Siemer D. Lech J. F. and Woodriff R. Spectrochim. Ada 1973 28B 469. Siemer D.D. and Woodriff R. Spectrochim. Ada 1974 29B. 269. Lech J. F. Siemer D. and Woodriff R. Environ. Sci. Technol. 1974 8 840. Kriese M. Kuhl J. Neumann S. and Torge R. German (BRD) Patent 2435091 1976. Favretto Gabrielli L. Pertoldi Marletta G. and Favretto L. Atom. Absorption Newsl. 1977 16 4. Sturgeon R. E. and Chakrabarti C. L. Prog. Analyt. Atom. Spectrosc. 1978 1 5. L‘vov B. V. Sfiectrochim. Acta 1978 33B. 153. Ediger R. D. Atom. Absorption Newsl. 1975 14 127. Aggett J. and Sprott A. J. Analytica Chim. Acta 1974 72 49. Campbell W. C. and Ottaway J. M. Talanta 1974 21 837. Fuller C. W. Analyst 1974 99 739. Paveri-Fontana S. L. Tessari G. and Torsi G. Analyt. Chem. 1974 46 1032. Torsi G. and Tessari G. Analyt. Chem. 1975 47 839. Tessari G. and Torsi G.Analyt. Chem. 1975 47 842. Sturgeon R. E. and Chakrabarti C. L. Analyt. Chem. 1977 49 1100. van den Broek W. M. G. T. and de Galan L. Analyt. Chem. 1977 49 2176. West T. S. Pure APPL. Chem. 1978 50 837. Manning D. C. Atom. Absorption Newsl 1978 17 107. Fernandez F. J. Myers S. M. and Slavin W. Perkin-Elmer Corp. Norwalk Conn. personal Massmann H. El Gohary Z. and Giicer S. Spectrochim. Acta 1976 31B 399. Frech W. and Cedergren A. Analytica Chim. Acta 1976 82 93. Wibetoe G. Thesis Department of Chemistry University of Oslo Norway 1978. Wilson A. D. Analyst 1964 89 18. Solberg. R. Thesis Department of Chemistry University of Oslo Norway 1973. Lundberg E. and Frech W. Analytica Chim. Acta 1979 104 75. Koch 0. G. Pure Appl. Chem. 1978 50 1551. Langmyhr F. J.and Kjuus I. Analytica Chim. Ada 1978 100 139. Zief M. and Mitchell J. W. “Contamination Control in Trace Element Analysis Chemical Analysis, Volume 47,” John Wiley New York 1976. Mizuike A. and Pinta M. Pure Appl. Chem. 1978 50 1519. Siemer D. D. Koteel P. Wei H.-Y. Emmel R. and Sotera J. in Ewing G. W. Editor “Environ-mental Analysis,” Academic Press New York 1977 p. 175. Massmann H. Spectrochim. Acta 1970 25B 393. Langmyhr F. J. and Rasmussen S. Analytica Chim. Acta 1974 72 79. Headridge J . B. Lab. Pract. 1974 23 5 . Quentmeier A. Hagenah W.-D. and Laqua K. “Proceedings of the 5th International Conference on Atomic Spectroscopy Melbourne Australia 1975,” Australian Academy of Science, Dominion Press Victoria 1976. Prikl. Spektrosk. 1973 18 610. 1975 30 1913.Analytical Chemistry and Applied Spectroscopy Cleveland Ohio 1973.” communication 140. 141. 142. 143. 144. 145. 146. 147. 148. 149. 150. 151. 152. 153. 154. 166. 156. 157. 158. 159. 160. 161. 162. 163. 164. 165. 166. 167. 108. 169. 170. 171. 172. 173. 174. 175. 176. 177. 178. 179. 180. 181. 182. 183. 184. 185. 186. 187. 188. 189. 190. 191. 192. 193. 194. 195. 196. 197. 197a. 198. November 1979 BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY Headridge J. B. and Smith D. R. Talanta 1971 18 247. 1015 Gries W. H. and Norval E. Analytica Chim. Acta 1975 75 289. Ashy M. A. Headridge J . B. and Sowerbutts A. Talanta 1974 21 649. Nikolaev G. I. Zh. Analit. Khim.1964 19 6s. Siemer D. D. and Wei H.-Y. Analyt. Chem. 1978 50 147. Katskov D. A. and L‘vov B. V. Zh. Prikl. Spektrosk. 1969 10 382. Norval E. and Gries W. H. Analytica Chim. Acta 1976 83 393. Talmi Y. Analyt. Chem. 1974 46 1005. Gladney E. S. Atom. Absorption Newsl. 1977 16 42. O’Reilly J. E. and Hale M. A. Analyt. Lett. 1977 10 1095. Langmyhr F. J. and Aadalen U. Department of Chemistry University of Oslo Norway, Pemsler J . F. and Rapperport E. J. Analytica Chim. Acta 1972 58 15. Sullivan J. V. personal communication cited in Walsh A. APPZ. Spectrosc. 1973 27 335. KAntor T. Fodor P. and Pungor E. Analytica Chim. Acta 1978 102 15. Langmyhr F. J. Thomassen Y. and Massoumi A. Analytica Chim. Acta 1973 67 460. Owens J. W. and Gladney E. S. Atom. Absorption Newsl.1975 14 76. Nikolaev G. I. Zh. Analit. Khim. 1965 20 445. Robinson J. W. and Hindman G. D. Spectrosc. Lett. 1972 5 169. Belyaev Y. I. Koveshnikova T. A. and Kostin B. I. Zh. Analit. Khim. 1973 28 2111. Belyaev Y. I. and Koveshnikova T. A. Zh. Analit. Khim. 1974 29. 659. Grushko L. F. Krasil’shchik V. Z. Lifshits M. G. and Chupakhin M. S. Zh. Analit. Khim. 1977, Langmyhr F. J. Lind T. and Jonsen J Analytica Chim. Acta 1975 80 297. Headridge J. B. and Smith D. R. Talanta 1972 19 833. Andrews D. G. and Headridge J . B. Analyst 1977 102 436. Aziz-Alrahman A. M. and Headridge J . B. Talanta 1978 25 413. Andrews D. G. Aziz-Alrahman A. M. and Headridge J. B. Analyst 1978 103 909. Marks J . Y. Welcher G. G. and Spellman R. J. Appl. Spectrosc. 1977 31 9. Headridge J .B. and Thompson R. Analytica Chim. Acta 1978 102 33. Timofeev €3. F. Ivanova M. I. Elyutin A. V. Maiorov I. A. Nedler V. V. and Filimonov L. N., Langmyhr F. J. Solberg R. and Thomassen Y. Analytica Chim. Acta 1977 92 105. Koizumi H. and Yasuda K. Analyt. Chem. 1976 48 1178. Goleb J . A. and Yokoyama Y. Analytica Chim. Acta 1964 30 213. Brady D. V. Montalvo J . G. Jr. Glowacki G. and Pisciotta A. Analytica Chim. Ada 1974, 70 448. Katskov D. A. and L’vov B. V. “The Analytical Chemistry of Rare Metals and Semiconductor Materials.” (in Russian) Moskovski Dom Nauchno-tekhnicheskoi Propagandi Imeni F. E. Dzerhinsky Moscow 1970. Belyaev Y. I. Pchelintsev A. M. and Zvereva N. F. Zh. Analit. Khim. 1971 26 1295. Luecke W. Eschermann F. Lennartz U. and Papastamataki A.J. Neues Jb. Miner. Abh. 1974, Belyaev Y. I. Pchelintsev A. M. Zvereva N. F. and Kostin B. J. Zh. Analit. Khim. 1971 26, Goleb J. A. Appl. Spectrosc. 1971 25 522. Data from University of Gottingen Department of Mineralogy Federal Republic of Germany, Langmyhr F. J. and Thomassen Y. 2. Analyt. Chem. 1973 264 122. Langmyhr F. J. Stubergh J. R. Thomassen Y. Hanssen J . E. and DoleZal J. Analytzca Chim. Gong H. and Suhr N. H. Analytica Chim. Acta 1976 81 297. Bouvier J.-L. and Abbey S. Analusis 1977 5 332. Mitchell D. G. Mills W. N. Ward A. F. and Aldous K. M. Analytica Chim. Acta 1977 90 275. Lester J. N. Harrison R. M. and Perry R. Sci. Total Environ. 1977 8 163. Stoveland S. Astruc M. Perry R. and Lester J . N. Sci. Total Environ. 1978 9 263. Carrondo M. J.T. Perry R. and Lester J. N. Analytica Chim. Acta. 1979 106 309. Ghiglione M. and Eljuri E. Acta Cient. Venez. 1970 21 123. Backman S. Analytical Chemistry Section Sandvik AB 81 1 01 Sandviken Sweden personal Tsujii K. and Kuga K. Bunseki Kagaku 1978 27 415. Eljuri E. Lunar J. and Ghiglione M. Appl. Spectrosc. 1979 33 170. Lundberg E. and Frech W. Analytica Chim. Acta 1979 108 75. Langmyhr F. J. Solberg R. and Wold L. T. Analytica Chim. Acta 1974 69 267. Karmanova N. G. and Pogrebnyak Y. F. Zh. Analit. Khim. 1979 34 213. Ishibashi W. Sat6 M. and Hashimoto M. Bunseki Kagaku 1974 23 597. Kanda M. Hori Y. and Matsumoto I. Bunseki Kagaku 1975 24 299. Goleb J . A. Analytica Chim. Acta 1966 34 135. Goleb J. A. Analyt. Chem. 1963 35 1978. Page A. G. Godbole S. V. Deshkar.S. B and Joshi B. D. Analyt. Lett. 1978 All 619. Page A. G. Godbole S. V. Kulkarni M. J. Shelar S. S. and Joshi B. D. Z. Analyt. Chem. 1979, unpublished data. 32 218. Zh. Analit. Khim. 1977 32 507. 120 178. 492. reported by Welz B. CZ-Chem.-Techn. 1972 1 460. Acta 1974 71 35. communication. 296. 40 1016 LANGMYHR 199. Krivchikova E. P. and Vasileva N. M. Zh. Analit. Khim. 1973 28 928. 200. Lorch D. W. Atom. Absorption Newsl. 1978 17 95. 201. Data from Carlsberg Laboratory Copenhagen Denmark reported by Welz B. CZ-Chem.-Techn., 202. Langmyhr F. J. Busch S. G. and Knutsen T. Department of Chemistry University of Oslo, 203. Lundgren G. and Johansson G. Talanta 1974 21 257. 204. Hadeishi T. Appl. Phys. Lett. 1972 21 438. 205. Langmyhr F. J and Aamodt J.Analytica Chim. Ada 1976 87 483. 206. Langmyhr F. J. and Orre S. Department of Chemistry University of Oslo to be published. 207. Grobenski Z. Melcher M. and Welz B. 2. Analyt. Chem. 1978 290 144. 208. Renshaw G. D. Pounds C. A. and Pearson E. F. Nature Lond. 1972 238 162. 209. Renshaw G. D. Pounds C. A. and Pearson E. F. J . Forens. Sci. 1973 18 143. 210. Alder J . F. Alger D. Samuel A. J. and West T. S. Analytica Chim. Acta 1976. 87 301. 211. Graef V. J. Clin. Chem. Clan. Biochem. 1976 14 181. 212. Alder J F. Samuel A. J. and West T. S. Analytica Chim. Acta 1976 87 313. 213. Alder J . F. Pankhurst C. A. Samuel A. J. and West T. S. Analytica Chim. Acta 1977 91 407. 214. Alder J . F. Samuel A. J. and West T. S. Analytica Chim. Acta 1977 94 187. 216. Church D.A. Hadeishi T. Leong L. McLaugNin R. D. and Zak B. D. Analyt. Chem. 1974, 216. Jackson K. W. and Mitchell D. G. Analytica Chim. Acta 1975 80 39. 217. Jackson K. W. Marczak E. and Mitchell D. G. Analytica Chim. Acta 1978. 97 37. 218. Kahn H. L. Fernandez F. J. and Slavin S. Atom. Absorption Newsl. 1972 11 42. 219. Brady D. V. Montalvo J. G. Jr. Jung J. and Curran R. A. Atom. Absorption Newsl. 1974 13, 220. Fujiwara K. Umezawa Y. Numata Y. Fuwa K. and Fujiwara S. Bunseki Kagaku 1977 26, 221. Pickford C. J. and Rossi G. Atom. Absorption Newsl. 1976 14 78. 222. Stegavik K. Mikalsen G. Ophus E. M. and Mylius E. A. Bull. Environ. Contam. Toxicol. 1976, 223 Lord D. A. Breck W. G. and Wheeler R. C. ASTM Spec. Techn. Pubn 1976 573 95. 224. Lord D. A. McLaren J. W. and Wheeler R.C. Analyt. Chem. 1977 49 257. 226. Barnett W. B. and Kahn H. L. Clin. Chem. 1972 18 923. 226. Van Stekelenburg G. J. Van de Laar A. J. B. and Van der Laag Clinica Chim. Ada 1976 59, 227. Sohler A. Wolcott P. and Pfeiffer C. C. Clinica Chim. Acta 1976 70 391. 228. Henn E. L. Atom. Absorption Newsl. 1973 12 109. 229. Mitchell D. G. Aldous K. M. and Ward A. F. Atom. Absorption Newsl. 1974 13 121. 230. Lau 0. W. and Li K. L. Analyst 1975 100 430. 231. Castellani F. Riccioni R. Gusteri M. Bartocci V. and Cescon P. Atom. Absorption Newsl., 232. Kerber J . D. Koch A. and Peterson G. E. Atom. Absorption Newsl. 1973 12 104. 233. Langmyhr F. J. Thomassen Y. and Massoumi A. Analytica Chim. Acta 1974 68 305. 234. Sychra V. Kolihova D. and Dudova N. Chemicki Listy 1976 70 737.236. Simon P. J Giessen B. C. and Copeland T. R. Analyt. Chem. 1977 49 2285. 236. Machiroux R. and Dupont J . C. Analytica Chim. Acta 1976 85 231. 237. Kerber J. D. Atom. Absorptiort Newsl. 1971 10 104. 238. Henn E. L. Analytica Chim. Acta 1974 73 273. 239. Hwang J. Y . Instrumentation Laboratory Inc. Wilmington Massachusetts USA personal 240. Girgis-Takla P. and Chroneos I. Analyst 1978 103 122. 241. Hein H. and Schrader W. “Applied Atomic Absorption Spectroscopy,” Pamphlet No. 3 1976, 242. Ediger R. D. Knott A. R. and Peterson G. E. Perkin-Elmer Corp. Norwalk Conn. USA. 243. Favretto L. Atom. Absorption Newsl. 1976 15 98. 244. Muzzarelli R. A. A. and Rocchetti R. Talanta 1976 22 683. 245. LidCn S. and Lundberg E. J. Invest. Derm. 1979 72 42. 246. Langmyhr F. J. Sundli A. and Jonsen J. AnaZytica Chim. Actu 1974 73 81. 247. Morrison G. H. and Talmi Y. Analyt. Chem. 1970 42 809. 248. Hageman L. Torma L. and Ginther €3. E. J. Ass. Ofi. Analyt. Chem. 1976 58 990. 249. Robinson J. W. and Wolcott D. K. Analytica Chim. Acta 1975 74 43. 260. Krishnan S. S. Quittkat S. and Crapper D. R. Can. J. Spectrosc. 1976 21 25. 261. Behne D. Bratter P. Gessner H. Hube. G. Mertz W. and Rosick. U. 2. Analyt. Chem. 1976, Received Awil 30th 1979 Accepted June 26th 1979 1972 1 460. Norway unpublished data. 46 1352. 118. 736. 15 734. 233. 1977 16 57. communication. Bodenseewerk Perkin-Elmer p. 48. 278 269