Atomic Absorption Spectroscopy P. PLATT Colgate-Palmolive Limited Ordsall Law Salfovd Contents Introduction Nomenclature Sensitivity Limit of detection Noise level Resonance radiation Atomisation Atomiser Nebulisation Nebulisation efficiency Direct-injection burner Pre-mix system Separated flame Observation height Burner angle Resonance radiation detector Interference Matrix effect Ionisation buffer Releasing agent Nebuliser atomiser systems Monochromators Detector and read-out systems Automation Reagents for atomic absorption spectroscopy Interferences-Application of method Instrumentation Sources Experimental considerations 17 Applications The determination of macro components Elements determined in the nitrous oxide - acetylene flame Alkaline earth elements Aluminium Beryllium Boron Chromium Germanium Molybdenum Niobium Noble metals Rare earths Scandium Rhenium Silicon Tin Titanium Vanadium Zirconium Elements with analytical resonance lines between 190 and 230 nm Determinations by an indirect method Determination of mercury Determination of isotopic concentrations 17 ATOMIC ABSORPTION SPECTROSCOPY 179 I n trod uction This review is concerned with the developments that have taken place within the field of atomic absorption spectroscopy during the last 3 or 4 years.It is not aimed at the specialist in the field but rather is intended to provide the type of information required by the busy analyst who uses or has access to an atomic absorption spectrophotometer as just one of a number of instruments routinely used in the laboratory.He may perhaps use the technique regularly or only occasionally. It may also assist those who are debating whether or not an atomic absorption instrument can be justified in their own laboratory. Some idea of the rapid growth of the technique may be gained by noting that in 1966 there were between 1000 and 2000 atomic absorption spectrophotometers in use throughout the world1 whereas by 1969 this number had risen to more than 10000.2 West3 compares its impact on inorganic trace analysis to that of gas -liquid chromatography on organic analysis. Nevertheless despite it being now over 10 years since the first commercial unit was announced according to Reynolds* it is apparent that the technique is still essentially new to many laboratories.However atomic absorption spectroscopy must be accepted as being a very popular analytical tool in every field of activity because of the many advantages it can offer over other techniques. These advantages have already been well documented by various a ~ t h o r s . ~ ~ ~ ~ Whether or not all of the oft-claimed advantages are justified has been questioned by at least two prominent worker^.^-^ During the period covered by this review a number of books on the subject have been published and these are recommended for obtaining more detailed i n f o r m a t i ~ n . ~ ~ ~ - ~ ~ The House Journals of the various instrument manufacturers also provide much useful information and colour films on the theory and practice of atomic absorption spectroscopy have become available for free loan to qualified 0rganisations.l6,~~ In addition an excellent abstracting service dealing with all aspects of atomic absorption spectroscopy is now available.l* Nomenclature The definition of universally acceptable terms for use in atomic absorption spectroscopy is a subject that has recently received much active and necessary consideration by the Atomic Spectroscopy Group of the Society for Analytical Chemistry and by an I.U.P.A.C.Commission.18 The technique is still compara-tively new and there are obvious advantages to be gained by establishing an approved nomenclature as soon as possible. The ambiguity that can arise in the absence of such a nomenclature may be illustrated by the term ‘atomisation,’ which was until recently widely interpreted as the process that converts a liquid to a mist and not as a function producing atom^.^^^^^ The subject was given prominence by Powel121 when he considered terms selected with the analytical chemist in mind.A list of tentative proposals for the definition of terms used was proposed by the Atomic Spectroscopy Group and this was submitted for comment to the I.U.P.A.C. Commission to the Standard 180 PLATT Association of Australia and to several independent individuals. The following terms have been taken from the selected list and are suggested as the ones with which every analyst using the technique should be familiar. Under each term heading will be found any available comments and information taken from the literature concluding with the definition proposed by the Atomic Spectroscopy Group of the Society for Analytical Chemistry.It is hoped that in this way the importance of using a common set of defined terms may be encouraged, and hence facilitate the interchange of information and results. Considering their usefulness in making analytical comparisons it is not surprising that the terms ‘sensitivity’ and ‘detection limit’ have received the most publicity. Sensitivity Ramirez-Mufioz and Ulrich22 stated in 1966 that any change in the widely used ‘concentration for 1 per cent. absorption’ seemed unlikely. Many prominent workers including Elwell and GidleyJ6 Stupar and Dawson,lS Marshall and Schrenk23 and Slavinll have used the same definition. Ramirez-Muiioz Shifrin and Hell% defined sensitivity as the ratio between response obtained from an instrument and the concentration of the analyte in solution.Grant25 stated that the terns ‘sensitivity’ and ‘limit of detection’ had been used interchangeably and there was still no universal acceptance of definitions for these terms although there was a growing tendency to distinguish between them. He defined sensitivity as the ability to discern a small change in concen-tration of analyte at some specified concentration. Rendina26 complained of ‘a lack of definition’ and that conditions under which sensitivity measurements are made do not relate adequately to real life analytical situations. The latter complaint is a reasonable one in so much as it is true that the element in question is generally assumed to be the only solute present usually in water as the solvent.Also the quoted sensitivity figure may well have been determined by analysing a considerably more concentrated solution and then making a simple calculation to obtain the concentration corresponding to 1 per cent. absorption. With this in mind it will be readily appreciated that ‘limit of detection’ is a far more useful term for the analyst. It should be noted that ‘sensitivity’ is defined below as a direct measure of the slope of the calibration graph near zero concentration and therefore in common with the use of this definition in other branches of analytical chemistry it gives an indication of the rate of change of the observed signal with concentration.Proposed S.A.C. definition Sensitivity is the concentration in solution of the element to be determined that will produce a change compared to pure solvent, of 0.0044 absorbance units i.e. 1 per cent. absorption in the optical transmission of the atomic vapour at the wavelength of the radiation used. Limit of detection Ramirez-Muiioz and Ulrich22 suggested that this term may be defined as the concentration in parts per million of analyte giving a signal corresponding to th ATOMIC ABSORPTION SPECTROSCOPY 181 peak-to-peak noise obtained at 0 per cent. absorption level under some given operating conditions. These authors stressed that information concerning the operating conditions e.g. scale expansion and damping is necessary to properly evaluate and compare various systems.Kirkbright Semb and West2’ agreed with these latter observations and stated that it is necessary to agree upon a suitable definition of detection limit. While there is no official agreement for this limit an increasingly accepted definition is the concentration in water solution that gives a signal equal to twice the size of the background variability.28 In statistical terms, the concentration of an element at the detection limit can be determined with a coefficient of variation of 50 per cent. Marshall and S ~ h r e n k ~ ~ defined the limit of detection as that concentration of metal that produces an absorption equivalent to twice the magnitude of the fluctuation in the background at zero absorption. Slavinll used a similar definition and stated that for comparing analytical methods the concept of detection limit is more useful than sensitivity because it includes not only the factors that govern analytical sensitivity but also the factors that control the level of back-ground fluctuation in the analysis.Grant26 noted that there must be careful definition of the conditions under which the standard deviation was measured if the limit of detection value is to be meaningful. He defined the limit of detection as the smallest amount or concentra-tion of analyte that can be detected with certainty. Rains25 said the detection limit is not defined as readily as sensitivity and he defined it as the concentration in microgrammes per litre that produces an absorption signal equal to twice the magnitude of the fluctuations of the blank or background noise.RubeSka and Moldanl4 defined the detection limit as the concentration causing a deflection equal to three times the standard deviation of the fluctuations of the unabsorbed signal or twice the maximum noise level of the fluctuations. Assuming a normal distribution for the fluctuations both definitions are almost identical. The I.U.P.A.C. Commission recommended that three sigma should be used to give a minimum certainty of 95 per cent. at any form of distribution that can reasonably be anticipated. Cooke12 pointed out that detection limits of elements in atomic absorption, although less frequently quoted than sensitivities are nevertheless a better criterion of instrumental performance. He preferred to have detection limit in terms of standard deviation of the background noise and quoted what is essentially the proposed S.A.C.definition. Reynolds29 considered this definition to be more stringent and meaningful than the ‘twice the noise level’ one and made it clear that it is now more than ever necessary to ascertain exactly under what conditions and how published performance figures have been obtained and calculated. Proposed S.A.C. definition. The limit of detection is the minimum amount of an element that can be detected with a 95 per cent. certainty which is that amount of the element that gives a reading equal to twice the standard deviation of a series of at least ten determinations at or near blank level 182 PLAT" Noise level Proposed S.A.C. definition. Noise level is that concentration of the element to be determined that would give a signal equal to one fiftieth of the sum of twenty measurements taken as follows-The output of an atomic absorption spectrometer operating on a blank solution is recorded for ten time periods each of ten times the time constant of the instrument.The maximum displacements that occur to both sides of the median line in each of the ten periods are measured. These are the twenty measurements referred to above.3o Resonance radiation Proposed S.A.C. definition. Resonance radiation is the characteristic absorbed radiation that corresponds to the transfer of an electron from the ground state level to a higher energy level in the atom. Atomisation Stupar and Dawson19 defined this term as 'the production of atomic vapours.' Proposed S.A.C.definition. Atomisation is the process that converts the element to be determined or its compounds to an atomic vapour. Atomiser Proposed S.A.C. definition. An atomiser is the device usually a flame, used to produce and stabilise or maintain a population of free atoms. The I.U.P.A.C. Commission would accept this term only with great reluctance in view of its long common and somewhat unfortunate usage as a synonym for nebulizer. The S.A.C. committee however was of the opinion that it is the only term that follows naturally from atomisation. Nebulisation Stupar and Dawson19 defined this term simply as the production of aerosols. Proposed S.A.C. definition. Nebulisation is the process that converts a liquid to a mist.Nebulisation efficiency Proposed S.A.C. definition. Nebulisation efficiency is the ratio of the amount of sample reaching the atomiser to the total amount of sample entering the nebuliser. Djrect-injection burner Proposed S.A.C. definition. A direct-injection burner is one in which the liquid is nebulised directly into the flame. The flame obtained with such a burner is normally turbulent ATOMIC ABSORPTION SPECTROSCOPY 183 Pre-mix system Proposed S.A.C. definition. A pre-mix system is a sampling unit wherein the fuel oxidant gas and sample mist are mixed in a spray chamber before entering the flame. Flames obtained by using this system are normally laminar. Separated flame Proposed S.A.C. definition. A separated flame is one in which the diffusion combustion zone is so separated from the primary combustion zone as to enable the two zones to be observed independently.Observation height between the optical axis of the monochromator and the top of the burner. Proposed S.A.C. definition. The observation height is the vertical distance This term is sometimes referred to as burner height. Burner angle Proposed S.A.C. definition. The burner angle is the acute angle between the plane of the flame produced by a long path burner and the optical axis of the monochromat or. Resonance radiation detector Proposed S.A.C. definition. A resonance radiation detector is a selective detector in which atoms in an atomic vapour are excited by radiation from an external source and the intensity of the resulting fluorescence radiation is measured.Interference Stupar and Dawson19 defined this term as the discrepancy between the true and apparent concentrations of an element when estimated by comparing its absorption with that of a pure solution. Proposed S.A.C. definition. Interference is a general term for an effect that modifies the instrumental response to a particular concentration of the element to be determined. Matrix effect Proposed S.A.C. definition. A matrix effect is an interference caused by differences between the sample and a standard containing only the element to be determined and where appropriate a solvent. Ionisation buffer Proposed S.A.C. definition. An ionisation buffer is a spectroscopic buffer used to minimise or stabilise the ionisation of free atoms of the element to be determined 184 PLATT Releasing agent Proposed S.A.C.definition. A releasing agent is a spectroscopic buffer used to reduce interferences attributable to the formation of involatile compounds in the atomiser. I nst c u men t at i o n It is outside the scope of this paper to discuss the essential requirements of an atomic absorption spectrophotometer. These have already been more than adequately described in the original paper by Walsh5 and subsequently by a num-ber of well known authors in the field.334J0J1J3-15J9 Before discussing the particular developments of recent origin it may be of useful interest to indicate generally the points to be considered when choosing an atomic absorption instrument. According to Reynolds3l there are several ways of choosing such a unit.One method used more frequently than is desirable is that of selecting the most expensive unit within the laboratory budget and trusting that it will more than meet the essential requirements. He advises the chemist to give serious thought to the features he requires the instrument to possess. In the first place the characteristics and concentration levels of the elements to be determined should be considered together with the accuracy required. For example depending upon the level of the element present and the matrix in which it occurs the simplest equipment may allow the determination of elements such as magnesium and copper with the same level of sensitivity as the more expensive units. On the other hand, the determination of elements such as iron chromium and nickel or arsenic and selenium may require a higher performance monochromator and hence a more expensive instrument (see Sources).The possibility of using electrodeless discharge lamps for the determination of arsenic and selenium warrants serious consideration when choosing instrumentation with these two elements in Speed of operation is always important and must be given due consideration as must versatility. Most commercial instruments can be operated in the emission mode and offer a built-in scale expansion facility. Prices made it clear that to be of real value to the analyst an instrument must allow very ready interchange between all operating conditions for different elements e.g. lamps wavelengths gases and absorption - emission modes.It should also allow complete control over such instrumental parameters as burners, burner height and angle lamps scale expansion and read-out facilities. It should maintain high sensitivity for trace element determinations and high stability of zero and infinite absorbance levels for the determination of more major constituents and for long runs of routine analysis particularly when automatic sampling is being used. For the analysis of major constituents it is a big advantage according to Reynolds,31 to have an instrument capable of zero suppression as illustrated later in the determination of calcium although a number of workers have reported successful results without this facility (see Applications). Most commercial instruments use the single-beam optical system and for many applications this system is perfectly satisfactory.However the advantage ATOMIC ABSORPTION SPECTROSCOPY 185 of a double-beam system should be considered and weighed against the extra cost involved. Most are not genuine double-beam systems in contrast to those com-monly found in molecular spectroscopy because the reference beam does not pass through a reference flame. Hence a double-beam system does nothing to com-pensate for flame noise which with the advent of modern stable sources is the greatest producer of noise in atomic absorption measurement^.^^^^ It will however, compensate for source noise when present and this is an advantage in improving the detection limit and in compensating for zero drift during prolonged sample runs.This noise compensation will result in a higher signal-to-noise ratio which has a particular advantage when using scale expansion. However a double-beam system results in some loss of light energy and if the same slit-width is used more gain is necessary. More gain means more noise and hence a decrease in the signal-to-noise rati0.~~,~5 Therefore two opposite effects are operating which will in part at least cancel each other out. A single-beam system does not monitor source variations but does offer some advantages in that it allows the use of lower gain with modern stable sources and consequently allows a lower noise operation. This can mean better precision, detectability and signal-to-noise ratio.% Some instrumentation can have the output signal integrated over a fixed time peri0d.~6 This does not compensate noise but automatically ‘sums’ it and then feeds out the average as a noise-free reading.It is certainly a time saving con-venience but requires a drift free base-line otherwise results can be in error. In 1968 a most versatile instrument designed with a double monochromator system made its appearan~e.~’ The design allowed the simultaneous use of a resonant line and a nearby non-resonant line to compensate background or matrix interference. It also allows as one alternative the simultaneous use of two resonant absorbing lines of a particular element and when the strengths of two signals at their selected wavelength are compared the flame noise which is common to both is eliminated. The instrument also incorporates double-beam compensation of source variations.Elwell and Gidley6 have placed at the end of their book a tabulated insert on commercially available equipment. This gives the essential characteristics of six-teen instruments and although published in 1966 it is a most useful introduction to available instrumentation. Eardley and Mountford38 offer an excellent evaluation of seven commercial atomic absorption spectrometers with manufacturers comments regarding the evaluation results and Dawson and Bro~ghton~~ have published a guide to the selection of instruments for use in clinical biochemistry. Hence a considerable amount of information and advice is available that the analytical chemist can use to guide him when choosing an instrument. However the person responsible for doing the choosing is well advised to evaluate in practice his short list of instru-ments together with their various attachments.The evaluation should be carried out with samples peculiar to the particular laboratory and taking into account the experience and ability of the operators who will be using the equipment 186 PLAT" Reynoldss1 believes that the surest way of assessing the suitability of all equipment is to see it demonstrated by a competent operator under fair conditions that reasonably simulate those in which it is required for use. This should be in the actual laboratory where the instrument is to be used if this can be arranged but all manufacturers will arrange for the prospective purchaser to visit their specialist laboratories and analyse his own samples under fair conditions.Finally the appli-cations service and servicing facilities that are available from the manufacturers should be taken into account as this can be most valuable. The basic layout of all atomic absorption instruments is shown in block diagram form in Fig. 1. Recent developments relating to the various components will be considered in their instrumental sequence. t r Mono- Detector amplifier Source --sc- Atomiser = - - and I e Read-out chroma tor -Nebuliser cloud chamber Fig. 1. Basic layout of atomic absorption instruments Sources It must be said at the outset that hollow-cathode lamps are by far the most commonly used spectral sources and it is unlikely that this situation will change, at least for some time to come.It is in the field of lamp technology that the most dramatic improvements in instrumental performance have been made and there has been a steady increase in the number of elements for which reliable and efficient hollow-cathode lamps are available.29 Slavin and Slavin2 pointed out that in the early 1960's the chief problem with practical atomic absorption methods was associated with the difficulty of making useful light sources. By 1966 all of the elements except the alkalis and possibly mercury were able to be determined most effectively and commonly with hollow-cathode sources.28 Some lamps notably those for arsenic were regarded as inadequate for general use.40v41 Kahn28 also pointed out that although hollow-cathode lamps were available discharge lamps were best for the alkali metals.Slavin and Ringhardtz*O agreed with this conclusion. The advantage of discharge lamps lay in the fact that they were many times brighter than the hollow-cathode types and hence in th ATOMIC ABSORPTION SPECTROSCOPY 187 near infrared where the monochromator and detector are weakest and the alkali metals have their most sensitive resonance lines the discharge sources were considerably more efficient. Despite these difficulties the sources that were commonly available showed, in all but a few particularly difficult cases such long life times that it was almost impossible to gather statistical data on failure rates.42 These hollow-cathode lamps although satisfactory produced a rather low output intensity that necessi-tated the use of fairly wide monochromator slit settings.This resulted in a lack of sensitivity and selectivity and also in non-linear response curves. During operation the current in conventional hollow-cathode lamps is usually adjusted to the maximum consistent with the line width requirements. It is not in general increased beyond the point where the line width shows a signifi-cant increase because this is accompanied by self-absorption or self-reversal and this too may result in loss of sensitivity and selectivity and produce non-linear response cuves. For many elements lower currents were acceptable and provided quite satisfactory sharp line sources. However in certain circumstances an improve-ment was required and this was provided by the high intensity hollow-cathode lamp.43 As the name implies these lamps produced a big improvement in intensity over the conventional lamps and it was anticipated that they would permit much more linear sensitive and stable atomic absorption measurements to be made.The design of high intensity lamps provides for the vaporisation and excitation functions of the lamp to be separated by using the hollow cathode merely to produce the metal vapour which is then excited by an auxilliary discharge of about 500 mA between two electrodes in front of the cathode. This system allowed significant enhancement of the resonance line emission without appreciable increase in line width.& WilliP gives three circumstances in which distinct advantages were gained with high intensity lamps.The first is where a conventional lamp necessarily runs at a low current and where the presence of non-absorbing lines close to the resonance line would require the use of narrow slits to satisfactorily isolate this line. Under these conditions a better signal-to-noise ratio without loss of sensitivity may be obtained. The second is where there is an ionised metal or filler gas line so close to the resonance line that it cannot be separated by the monochromator and a tremendous improvement ensued by using a high intensity lamp to enhance preferentially the resonance line. The final use is in analyses that require the bright and somewhat noisy nitrous oxide - acetylene flame. This flame requires an intense source i.e. less amplifier gain to provide a noise-free analysis.C0balt,4~*~~ nickel4' and iron44 are examples of elements that were better determined by high intensity lamps although it was not expected that they would supersede the standard lamps for such elements as magnesium zinc lead calcium and copper.4s However high intensity lamps have a number of disadvantages31 in that an extra power supply is required to provide the high auxillary current they are costly have restricted life times and can only be produced for certain elements. 188 PLATT Considering the rather serious disadvantages involved it is not surprising that improvements to conventional lamps were sought. The re-designed conven-tional lamps that soon became available embodied the necessary improvements and were called high spectral output lamps.The rapidly replaced the high intensity lamps which no longer provided significant advantages in most Indeed it seems most unlikely that the four-electrode high intensity lamps will be used in the future at least in their present form and it is reported that of the two companies that originally produced them one has virtually abandoned their manufact~re.~~ The superior performance of the new two-electrode high spectral output lamps is because of the incorporation of shielded cathodes re-positioned and re-shaped electrode assemblies and better choice of filler gases. No extra power supply is required of course they are comparatively cheap and have a long life time. Several companies now give a lamp warranty for 5 ampere hours and under normal conditions of operation this means that 200 to 500 h are guaranteed.In actual fact most lamps far exceed this W a l ~ h ~ ~ notes that for most analyses there is little to be gained by using lamps of high intensity but if an increased intensity is required it can be obtained by (a) the pulsing technique of Dawson and EllisJS0 (b) high intensity hollow-cathode lamps or (c) electrodeless discharge lamps. Mannine9 was able to say in 1968 that greatly increased output from sodium and potassium hollow-cathode lamps has now enabled them to be recommended for atomic absorption work. This means that the vapour discharge lamps with their special power supply are no longer required for these two elements. RubidiumS1 and caesium do not have reliable hollow-cathode lamps and the excellent vapour dis-charge lamps that are available should be used for these two elements.A suitable hollow-cathode lamp is now available for mercury as well as an excellent vapour-discharge lamp. Reliable hollow-cathode lamps are also available for arsenic and selenium although the use of electrodeless discharge lamps3e should be considered by anyone interested in determining these elements and also tellurium. Reynolds31 in 1969 said that improvements in lamp technology have raised the performance of simple atomic absorption equipment to that of more expensive units. This can be seen in their capability of selecting the 248.33 nm from the 248.82 nm iron lines and by the separation with a simple monochromator of the 357.87 359.35 and 36063 nm lines of chromium. Perhaps the most valuable success according to Reynolds was the production of lead lamps that permit the reliable use of the very sensitive 217.0 nm line on relatively simple equipment.In fact by 1969 several commercial suppliers were offering a near complete range of hollow-cathode lamps including those for the lanthanides. Although some of these are rather expensive all are as bright for each element as an atomic absorp-tion spectrophotometer can ~ t i l i s e . ~ ~ ~ ~ Two modifications or extensions of the high intensity type of lamp have been reported. Bowman et al. described the isolation of atomic resonance lines by selective modulation.% In this technique the resonance lines of an elemen ATOMIC ABSORPTION SPECTROSCOPY 189 emitted by a d.c. operated high intensity lamp are selectively modulated by passing them through a pulsating cloud of atoms of the same element produced by an ax.operated hollow cathode. The resonance lines of the element to be determined are selectively modulated and measured by the ax. detector. None of the other lines is modulated and hence is not measured. This technique permits the use of much larger slit-widths and less exacting monochromators even simple filters may suffice. KoirtyohannM pointed out that a severe limitation of selectively modulated lamps was the extra power supply required for their operation. LoweS5 described an improved lamp of this type that did not require any extra electrode the cloud of atoms producing the selective modulation being obtained by a series of short, high current pulses superimposed on the steady direct current.Van Rensburg and Zeeman56 described the determination of gold platinum, palladium and rhodium by using a multi-element high intensity lamp with selective modulation. Butler and Brink57 prepared selectively modulated lamps for several different elements and found that the analytical curves were steeper and more linear than with conventional lamps. Also neither the band-width nor slit-width was critical even for line-rich elements such as iron and nickel. Dawson and Ellisso described a method to increase the intensity of hollow-cathode lamps. They superimposed a large pulsed current on to the low steady direct current and obtained up to several hundred fold gain in intensity with no deterioration in absorption signal.In this fashion the working life of the lamp was extended considerably and the shelf-life also appeared to be increased. Again an extra power supply is required. Neither of these modified high intensity sources has become commercially available although Koirtyohann5* considered that the latter technique50 might find use in atomic fluorescence measurements where the added intensity is so necessary. An interesting use of pulsed hollow-cathode lamps in a multi-channel atomic fluorescence device has in fact recently been described by Mitchell and Johanss0n.~8 A relatively large number of papers have appeared dealing with the use of multi-element hollow-cathode lamps. In common with the single-element lamps, changes in design have led to improved lamps with brighter emission improved stability and a longer life.59 One still obtains less light per element than with the single lamps but this should now be at an acceptable level.A good quality monochromator may be required to avoid spectral interference with some combi-nations of elements e.g. iron cobalt and nickel. A further disadvantage is the possibility of one element sputtering off before another. For example a sodium -potassium lamp may become predominantly a sodium lamp after a time. There is an obvious advantage when the final cost is less per element and the time saving by eliminating lamp changes from element to element should be considered, although it will not now be as important with modern instrumentation offering lamp ‘warming-up’ devices and hence instant use.In 1966 Slavinl said lamps were provided that combined several elements without loss of performance and that a large variety of such combinations woul 190 PLAT" eventually become available. Yet in 1968 Manningg was not so optimistic. His experience had not shown the multi-element lamps to have the previously expected economic and convenience advantages. He considered that although the con-venience might outweigh the disadvantages for some users the single-element lamp was recommended for most analyses. Jaworowski and Weberling6O found an occurence of apparent spectral interference in the use of multi-element lamps and suggested that a high dispersion monochromator should be used to eliminate the effect. Galassi and Hell*l used a multi-element lamp for barium calcium magne-sium and strontium and found interference with the calcium line at 422.7 nm by a secondary strontium line at 421.6 nm although the calcium line could be isolated by the use of a narrow slit-width.Analytical results compared favourably with those for a single-element calcium lamp. Frank et aLB1 studied the use of an iron hollow-cathode lamp as a multi-element source for magnesium manganese nickel, copper barium and silver. They concluded that an effective intermediate sensitivity was obtained. Heneagea2 found that when using a five-element lamp for chromium, cobalt copper manganese and nickel the difference between the results obtained using this or single-element lamps was usually very small. - I By 1969 Fernandez et ~ 1 . ~ ~ were able to report that multi-element lamps offered a reduction in lamp cost per element and although the emission intensity from the individual elements was not as great as that from single-element lamps, the lower intensity had little or no effect on performance.The sensitivities and detection limits obtained with various multi-element combinations were about equal to those obtained with single-element lamps. In 1969 Barnett and Kahnl7 found that a multi-element lamp for aluminium silicon calcium magnesium iron, copper and zinc gave a level of analytical performance with an ease and speed that indicated that this lamp suffered less by comparison with single-element lamps than had been supposed. Cooke12 suggested that difficulty may be experienced in selecting the required line when using a six-element lamp.A number of manufacturers offer multi-element lamps the calcium - magne-sium one in particular and as they axe now capable of a good performance they should certainly be given serious consideration when new or replacement lamps are required. A number of other hollow-cathode lamp modifications such as molten ~athodes,3~*~ magnetic controls4 and demountable typesap66 have been used by various workers but there is as yet no evidence of commercial development. Fassel et d.67 evaluated spectral continua as primary sources and the results obtained showed that these sources could be successfully used in atomic absorption spectroscopy. For thirty-two elements the sensitivities observed were either comparable to or exceeded those then obtained (1965 to 1966) with hollow-cathode sources.These workers suggested further refinements such as greater resolution in the spectrometer better continuum primary sources lower noise flames and e'lec-tronic integration of the absorption signal. DeGalan et ~ 1 . 6 8 concluded that with a good medium dispersion monochromator a continuous source offers several distinct advantages e g . the possibility of qualitative analysis simplicity of backgroun ATOMIC ABSORPTION SPECTROSCOPY 191 correction and low cost compensating for the higher priced high resolution mono-chromator. Also it yields detection limits that are approximately the same as those obtained with a hollow-cathode lamp. McGee and Winefordner@ used a continuum source in conjunction with an extended flame cell an argon - hydrogen -air flame and a medium dispersion monochromator.They found this system to be competitive with a hollow-cathode source and a typical air - acetylene flame system. Despite these apparent advantages the continuous source has not developed commercially. Kahn42 attributed this to the severe instrumental difficulties associated with it while Slavinll and Willis& pointed out that the calibration curves are so bent as to be useless at high absorbance and not much better even at fairly low absorbance values. Dagnall and West70 considered that the continuous source had only been used to demonstrate an academic point. The use of microwave-excited electrodeless discharge lamps (E.D.L.’s) as an alternative source to hollow-cathode lamps in atomic absorption spectroscopy has recently been propo~ed.3~~~~-73 This type of source has been known for many years and used mainly for studies of spectral structure.The reason for this present development is to be found in the search for more intense sources to produce better linearity in calibration curves and a higher signal-to-noise ratio. E.D.L.’s should be much less expensive to produce than high intensity or high spectral output hollow-cathode lamps which were developed for similar reasons.7o They do, however require a micro-wave generator and cavity which will add considerably to the cost of equipment. Since these initial publications conflicting views have been expressed about electrodeless lamps. Some workers have indicated that they can be used satisfac-torily while others have said they are much less stable than hollow-cathode lamps and cannot be recommended for atomic absorption spe~troscopy.7~ Dagnall and West70 in 1968 concluded that the application of electrodeless discharge lamps in atomic absorption spectroscopy is at present only in its early stages and many of the experimental parameters involved in their preparation and operation have not been fully examined.However from the results that have recently appeared in the literature these authors suggest that there is little doubt that E.D.L.’s will occupy an increasingly important position in the future. The lamps prepared in their laboratory were primarily used for atomic fluorescence spectroscopy. also in 1968 reported that a comparison of the intensity of electrode-less discharge lamps and hollow-cathode lamps showed the former to be more intense in nearly every case but owing to the greater line-width produced the absorption was never greater.The lower sensitivity obtained for the electrodeless lamps was attributed to self-reversal. In 1969 W~odward~~ reported that electrode-less discharge lamps did not have the same reliability as hollow-cathode lamps but, nevertheless they proved extremely useful for analysis by atomic absorption spectroscopy. In several instances low analytical sensitivity was obtained but electrodeless lamps had been prepared and applied regularly during the previous 9 months for the atomic absorption determination of over twenty elements in a wide variety of samples 192 PLATT Slavin and Slavin2 considered that the published literature tended to over-simplify the problems associated with electrodeless lamps and observed several practical problems although they confidently expected an eventual improvement.They found an important difficulty in that many of the lamps produced were very unstable with time and concluded that the utility of atomic fluorescence probably depends upon successful improvement of electrodeless discharge lamps. Double-beam instrumentation will be an advantage in accounting for base-line drift. Reynolds31 also points out that the very high output intensities make E.D.L’s ideal for atomic fluorescence measurements and they could even replace some of the less efficient hollow-cathode lamps for atomic absorption analysis.Headridge and Richardson7* compared electrodeless discharge lamps with hollow-cathode lamps for atomic absorption spectroscopy and concluded that for many determinations where the highest precision is not required electrodeless discharge lamps are satisfactory as light sources and are much cheaper to make than hollow-cathode lamps. As expected the detection limits are poorer for the electrodeless lamps. Browner Dagnall and West76 found that the performance of electrodeless lamps for lead mercury silver thallium and tin both modulated and unmodulated, was comparable or better than the respective hollow-cathode lamp run under optimised conditions. These workers concluded that previous work had shown the advantages of electrodeless lamps in the short wavelength regions for such elements as arsenic and selenium and there were other instances where they gave rise to considerable increases in sensitivity and a wider working range compared with hollow-cathode lamps.For the mercury electrodeless discharge lamp the better spectral profile of the emitted resonance line gave these advantages. Fisher and Hayward in a recent pape*2 said that probably the most frequent criticism of electrodeless discharge lamps was on the grounds of instability. They found that lamps run in a high intensity mode gave poor analytical sensitivity and when this was improved by reducing power instability increased and it was not possible to match the performance of hollow-cathode lamps. However by using a tunable cavity adjusted to minimise reflected power from the cavity these workers were able to improve lamp performance for arsenic selenium bismuth, gallium germanium thallium tellurium and aluminium lamps such that the sensitivity was better than that obtained with hollow-cathode lamps.This was achieved at the expense of lower intensity. They studied arsenic and selenium lamps in detail and concluded that the performance of electrodeless discharge lamps as atomic absorption sources compares favourably with that of hollow-cathode lamps for these elements in particular. Comparable stability better analytical sensitivity and better detection limits were obtained. Electrodeless discharge lamps retail for about half the price of hollow-cathode lamps but when the time and facilities are available they can be ‘home-made’ easier than can hollow-cathode lamps and for much less than the retail price.The commercially available lamps are now said to have a virtually unlimited shelf-life and shorter warm-up periods than the earlier lamps ATOMIC ABSORPTION SPECTROSCOPY 193 It appears that these sources will develop more for atomic fluorescence measurements but certainly anyone interested in determining elements with resonance lines in the ultraviolet region e.g. arsenic and selenium should consider their use for atomic absorption spectroscopy and compare them as far as possible with the improved hollow-cathode lamps that are now available for these elements. Nebuliser - atomiser systems To ensure efficient evaporation of the sample solution it must be introduced into the flame in a finely dispersed state.To carry out this operation pneumatic nebulisers are invariably used.13~14 In this system the supporting gas which is usually air or nitrous oxide aspirates the liquid sample and converts it to a mist. The mist may be either admitted directly into the flame as with total consumption burners or it may reach the flame via a cloud chamber where the larger droplets of liquid are deposited and rejected. The latter system is known as a pre-mix one and it allows the fine mist to be led through a burner slot of convenient length to produce a long path flame cell for absorption.ls The cloud chamber system is relatively inefficient in that only about 10 per cent. of the original sample volume sprayed reaches the flame but it has the advantage of sustaining a highly stable laminar flame the characteristics of which are influenced very little by the sample mist except when non-aqueous solutions are sprayed.12 Willis20 studied the mode of operation of the nebulising system and concluded that the sensitivity of the instrument and the effect of chemical interferences in the flame are critically dependent on the construction of the nebuliser and particu-larly on the rate of liquid uptake.Reynolds31 considered that at the present time (1969) the capabilities of all commercial nebulisers are very similar and it is in this sector that most workers recognise that improvements in the performance of atomic absorption equipment need to be made. RubeSka14 also expects further develop-ments in atomic absorption to depend on the improvement of methods for produc-ing atomic vapours.Rains25 suggests that the nebuliser - burner system is probably the most important component in any absorption measurement. Some workers have tried to avoid the inefficient waste of 90 per cent. of the solution when using pre-mix burnem2 Perhaps the most commonly tried method is the use of an ultrasonic nebuliser which will convert a larger fraction of the solution to the fine mist utilised in the pre-mix burner than will the common pneumatic type of nebuliser. Most workers have found that a smaller amount of sample will produce a given signal i.e. improved efficiency but that it is difficult to convert this into a real improvement in detection limits. Slavin and Slavin2 confirmed the general findings of Hoare et aL7' in that about one fifth of the sample up-take rate produced about the same absorption signal for some fifteen elements, but it was not possible to increase the uptake rate of the ultrasonic nebuliser to take advantage of this improvement in efficiency.West78 stated that in spite of the apparent advantages associated with ultra-sonic nebulisation the technique has not become widespread perhaps because o 194 PLATT the inconvenience of sample changing and clean-up of previously used systems. Several worker^^^-^^ improved the design to make sample changing more con-venient. Stupar and Dawsonl9 studied the theoretical and experiment+ aspects of the production of aerosols by both pneumatic and ultrasonic means. They concluded that unless operated at the high frequency (>ti00 kHz) and high power necessary to generate fine mist at sample flow-rates of 1 to 5 ml min-l the sensi-tivity and interference is worse than that obtained with a pneumatic nebuliser.Nevertheless at the time (1968) the ultrasonic type offered a practical means of increasing the efficiency with the possibility of reduced interference. However the usefulness of the ultrasonic nebulisers has remained remarkably disappointing2 and there appears to be no advantage in commercial development on the part of the instrument manufacturers although of course an ultrasonic generator may be purchased separately if so desired. Other methods used to improve the efficiency of the pre-mix system include heating the support gas before mixing and heating of the cloud chamber.Rawson81 and Riley and Taylors2 used the former technique and obtained increases in effi-ciency of up to 16 and 3 fold respectively. Heating of the cloud chamber has been used by a number of and at least one instrument manufacturer incorporates this system into commercially available equipment .86 Gray and Gallwas87 compared the use of a laminar flow high solids burner with a heated chamber laminar flow burner for the determination of serum calcium and con-cluded that the heated chamber type was less convenient for use with high solids but was nevertheless much more sensitive. Heating the support gas or the cloud chamber causes the sample mist to be dried and hence concentrated before entering the flame thus obtaining increased efficiency.Unfortunately according to WillisJM the increase is usually accompanied by some instability or drift in the readings and of carry-over or memory from one sample to the next. Three methods of introducing the sample into the flame without nebulisation have been reported. Venghiattisss described a method for the atomisation of solid samples. He did not claim that solid sampling would replace the conventional nebuliser type of sample introduction but showed that for some elements at least, the sensitivity attained was greater than with the conventional method. A solid sampling system is commercially available from one instrument manufacturers9 and has been successfully used for the determination of copper nickel gold silver, mercury lead and bismuth. The refractory elements e.g.aluminium vanadium, titanium silicon and the rare earths cannot at present be determined by this method. The sampling boat technique was described by Kahn et aL90 and was designed to get the maximum detection limit with small amounts of sample. The boat is constructed of tantalum and holds about 1 ml of sample which is dried by holding it near the flame and is then passed into the flame for atomisation. Quite dramatic improvements in detection limits were obtained for easily atomised elements such as arsenic zinc lead and mercury. Curry et aL9l applied the sample boat technique to the determination of thallium in biological material and found an increase i ATOMIC ABSORPTION SPECTROSCOPY 195 sensitivity of at least twenty five times over conventional methods ; sensitivities for lead and cadmium were also increased.The same workers reported that in contrast to conventional atomic absorption analytical methods for blood and urine, the results obtained with the tantalum boat are susceptible to inter-element interferences and calibration by the method of standard additions (see section dealing with ‘Experimental Considerations’) is essential for accurate quantitative results. However they also noted the advantages of rapid quantitative analysis and minimum sample pre-treatment which made the technique ideal for screening large numbers of samples for such toxic elements as thallium lead mercury and cadmium while using very small samples of blood and urine of 50 to 200 pl. The third technique adopted by a number of workers employed a wire to trans-port the sample into the flame.White92 supported the sample on a loop of platinum wire and mechanically introduced it into the flame; the atomic vapour thus pro-duced is directed into a nickel tube adsorption cell to increase the sensitivity and duration of the measurement. The principle application of this technique was the determination of lead in blood when an ear or a finger prick of 0.1 ml can be used rather than the more usual 5 ml extracted intravenously by syringe. Delvesg3 combined the techniques of Kahn and White and provided a rela-tively simple but accurate method in which the tantalum boat was replaced by a nickel micro-boat in association with an absorption tube as used by White. The concentration of lead in lo-$ samples of whole blood was accurately and rapidly determined.Along with the continued reign of the hollow-cathode lamp as the main primary source the flarne remains as the most commonly used atomiser. Price13 states that the combustion flame is one of the most convenient and efficient atom producing plasmas. Rubeska14 considered that the use of chemical flames has enjoyed the greatest popularity partly because of a carry-over from the emission technique and partly because of its simplicity and the ease with which it can be applied and controlled. The problem of atomisation according to W i l l i ~ ~ ~ is central to the whole atomic absorption technique and although several other techniques have been tried (see later) the use of a suitable flame still remains as the method favoured by the overwhelming majority of workers in the field.While the flame has certain disadvantages the equipment required is simple inexpensive , easy to use and well adapted to the rapid measurement of a series of different solutions so that it is unlikely that any other method of atomisation will replace it for the great bulk of analytical work. Slavinll also considered that the advantages of the flame are so compelling that it is unlikely any other technique will com-pletely replace it although he consideredzit to be the residual problems and limita-tions to a flame that have initiated a large number of research programmes of considerable interest. WestS made it clear that virtually the only atom reservoirs in use are flames although he said there are many factors indicating that other media may be more advantageous and these will undoubtedly come into use in the future.Apart from the different gas mixtures used there are two distinct types o 196 PLATT flame. In the first which is called the pre-mixed laminar flame and has a well defined inner cone or reaction the support gas converts the sample solution to a mist that passes into the cloud chamber where the larger droplets settle out and go to waste. The fuel gas may be added either before or after the chamber and the mixture of fuel gas support gas and sample mist proceed to the burner where it is burnt.94 Most workers agree that for atomic absorption work the laminar flow burner is ~ u p e r i o r ~ ~ J ~ * ~ ~ and it is significant that almost all commercial instruments currently produced use this type.The second type results from the fuel gas and the supporting gas not being mixed until the point at which they enter the flame and the sample solution is also introduced at this point. This one is called a total consumption system and is a combination nebuliser - burner. It is necessarily circular in design and cannot conveniently be used to provide a long absorption path for atomic a b s o r p t i ~ n . ~ ~ ~ ~ ~ At first sight one might expect total consumption burners to be more efficient and might question why pre-mix burners are used at all. However numerous examples from the literature have shown interferences to be present in the total consumption flame that are absent from pre-mixed flames2?31 With a total con-sumption burner a large proportion of the solution goes straight through un-evaporated cooling the flame and leading to poor sensitivity.Cowley et aLg5 indicated that a large part of the problem associated with turbulent flames results from disruption of the suitable chemical environment found in the well defined zones of the pre-mixed flame. Mossotti and Duggang6 found improvement in over-all performance when using a total consumption burner for emission measurements in which the gases were pre-mixed and produced a more laminar flame with well defined zones. Hence there appears to be no need for total consumption burners to be limited to turbulent flames. The total consumption burner is unlikely to flash-back as there is no large chamber containing a potentially explosive gas mixture but modem pre-mix burners are designed so that the chance of flash-back is minimal.At present the long path pre-mix burner is superior for atomic absorption purposes because it gives good sensitivity acceptable signal-to-noise ratios has remarkably little memory and excellent quantitative stability. The relatively low temperature of the flame(s) was one of its major disadvan-tages because it strictly limited the number of elements that could be determined.14 The air - propane flame at a temperature of about 1925 "C gives the best sensitivity for the alkali metals and for certain other elements forming compounds that are not thermally stable. The most commonly used and generally useful gas mixture is air - acetylene because the temperature and reducing conditions within this flame can be varied within fairly wide limits.For most routine determinations it is operated in a non-luminous manner at about 2300 "C while the luminous mode, providing reducing conditions is suitable for some of the elements forming refrac-tory oxides including tin barium chromium and molybden~m.~?~~J~?~7 However, a large number of elements forming refractory oxides require a much higher tem-perature than can be provided by acetylene burning in air. Higher temperature flames are provided by acetylene or hydrogen burning i ATOMIC ABSORPTION SPECTROSCOPY 197 oxygen and by acetylene burning in a mixture of oxygen and nitrogen. However, owing to the high burning velocity of these mixtures they are prone to flash-back and require special burners with narrow slots which easily become clogged by the crystallisation of salts or the formation of ~ 0 0 t .l ~ ~ ~ ~ Hence they are not as easy and safe to use as are air - acetylene flames. The problems involved in progressing to a high temperature flame were over-come most satisfactorily by the nitrous oxide - acetylene flame described by W i l l i ~ . ~ ~ It is predicted that this system will probably retain the privileged position of chemical flames in atomic absorption spectroscopy for at least a few years to c ~ m e . ~ ~ l ~ The development of the nitrous oxide - acetylene flame and suitable burners for its use resulted in the extension of the number of elements readily amenable to atomic absorption methods from thirty-five to about sixty-five which is approaching the theoretical maximum.loO It can thus be readily appreciated that the introduction of this flame was a major step forward in atomic absorption analysis.Nitrous oxide - acetylene has a low burning velocity similar to that of air -acetylene combined with a high temperature (of about 2900 "C). Measurement of the temperature was discussed by deGalan and Samaey,lo2 who obtained values agreeing well with literature data.lol WillisM and Rubeska14 state that the high temperature obtained is caused in part by the energy liberated by the decomposi-tion of the nitrous oxide. The burning velocity is limited by this decomposition reaction so that it is quite safe to work with the mixture when using standard pre-mix systems.Kirkbright et aZ.lo3 investigated the reactions that occur in the nitrous oxide - acetylene flame and concluded that these are more complex than can be explained by the reducing action of incandescent carbon particles. It appears that nitrous oxide decomposes to provide oxygen at the primary reaction zone and hence raises the temperature and burning velocity of the flame. Some of the nitrous oxide possibly via nitric oxide appears to react directly with fuel molecules in the primary zone to produce more -CN and -NH radicals than are produced in conventional hydrocarbon - air flames. A reducing atmosphere is thus provided to protect the metal atoms. Amos and Willis98 in describing the use of high temperature pre-mixed flames considered that the new flames apart from allowing the addition of some twenty-five metals to those already determinable possessed the further advantage of permitting the determination of the alkaline earth metals which are only partially atomised in cooler flames with higher sensitivity and greater freedom from chemi-cal interference.The higher temperature removes the need for the presence of lanthanum which was commonly added as a releasing agent when determining alkaline earth metals in the air - acetylene flame although it may still be used as an effective ionisation buffer in the latter flame. Potassium is more often used in this c a p a ~ i t y . ~ * J ~ It may still be true that air - acetylene is the most generally useful gas mix-ture but nitrous oxide - acetylene which has normally been considered most use-ful for elements forming refractory compounds is often found to have advantage 198 PLATT for a number of other elements normally determined in an air - acetylene flame and it is interesting to note that chemical interference is largely eradicated52 and better linearity of calibration data is often obtained.In general the hotter the flame the less chemical interferen~e.~ However because ionisation will be greater in the hotter flame the sensitivity will be less and whenever this is important a31 ionisation buffer should be added. In actual fact the 50-mm slot solid stainless-steel burner normally used for nitrous oxide - acetylene often gives adequate sensitivity when burning air - acetylene and for many of the analyses involving both gas mixtures it may not be necessary to change burner heads98 One of the worst faults of the nitrous oxide - acetylene burner head was its tendency to carbon up particularly when operated fuel rich.This fault has now been removed by re-designing the burner head. On the rare occasions when flash-backs do occur it is usually when the flame is being ignited or extinguished. Pampello5 described a method of avoiding flash-backs by using air as the oxidising gas during ignition and turning off. There is a difference of opinion among the instrument manufacturers about the necessity of this safety procedure some including it and others favouring direct ignition of nitrous oxide - acetylene. Should a flash-back occur modern instrumentation is designed so that no serious damage results.In 1966 Slavin et aZ.lo6 reporting recent experiences with the nitrous oxide -acetylene flame said that the technique had proved to be completely practical and many laboratories were utilising it on a routine basis for a number of elements. Nonetheless the method was still very new and much development work remained to be done. These workers studied chemical and ionisation effects and the effects of varying slot dimensions in relation to the determination of a number of elements including calcium barium silicon and titanium. Later in the same year NlanninglO7 listed the operating conditions together with sensitivities and detection limits for twenty-eight elements determined in the nitrous oxide - acetylene flame. The same author discussed effects caused by chemical interferences ionisation and flame noise.Walsh1°8 pointed out that it must not be expected that this new flame will completely eliminate chemical interferences in all analyses. West3 considered that the hotter flame had made the study and determination of elements such as aluminium molybdenum and niobium relatively easy and that further progress will no doubt improve quite considerably the sensitivities obtain-able. In 1968 Willis9* reviewed the features successes and remaining limitations of the nitrous oxide - acetylene flame. He traced the development from low tem-perature to high temperature flames and indicated that the nitrous oxide - acety-lene system had been adopted by the manufacturers of commercial equipment as the standard method of atomising metals that are not satisfactorily atomised in the air - acetylene flame.Willis discussed ionisation effects chemical interferences and inter-element enhancement effects in the use of high temperature flames in chemical analysis. He concluded that the development of high temperature flames for use in atomic absorption has not only provided a powerful technique fo ATOMIC ABSORPTION SPECTROSCOPY 199 the solution of analytical problems but has also led to a more critical consideration of the factors determining the performance of flames as atomising systems. Bowman and Willislo9 presented results showing the usefulness of the nitrous oxide - acetylene flame in the analysis of rocks oils minerals and steels. They studied several effects that need to be considered and concluded that any metal with an ionisation potential of below about 6.5 V for example aluminium and several of the rare earths will be affected to a significant extent by ionisation if this is not suppressed.Also the atomisation of some metals such as titanium is incomplete to an extent that is strongly dependent on the other substances present. In 1969, Shifrin et aZ.l1° reported recent results with a nitrous oxide burner showing detec-tion limits better than other values currently reported in the literature for several elements. These workers used hot and cold cloud chamber operation with aqueous and organic solvents. Also in 1969 Slavin and Slavin2 reported that the nitrous oxide - acetylene flame had made it possible to determine almost all metals with detection limits that are usually better than 1 pg ml-l.Finally a useful practical point is suggested by Julietti and Wilkinson.lll At the normal high flow-rate of nitrous oxide the expansion of gas in the regulator causes considerable cooling which may have an appreciable influence on the per-formance of the regulator. This was overcome by placing a 100-W light bulb a few centimetres above the regulator. Most atomic absorption work has been carried out with acetylene flames supported by air or nitrous oxide. Butler and Fulton112 and Fleming115 investi-gated the use of an acetylene flame supported by a mixture of air and nitrous oxide. This system allowed a continuously variable flame temperature between the lower air - acetylene and the higher nitrous oxide - acetylene.It was flexible and safe but the results seemed to indicate that for most applications there were no advantages significant enough to justify the additional complexity of gas mixings Acetylene burning in nitric oxide produces a flame that has a slower burning velocity and a slightly higher temperature than has nitrous oxide - acety-lene. However nitric oxide is relatively expensive and not readily available. Amos and WiUisg8 considered further investigation of this flame might be worth-while but Slavin et aZ.1°6 indicated no significant advantage in its use. Hence its price availability and its corrosive and toxic nature have ruled it out for atomic absorption spectroscopy. Temperatures higher even than the nitrous oxide -acetylene flame would probably limit their usefulness by the loss through ionisation of neutral atoms.s* Other pre-mixed flame types may however offer advantages for some ele-ments.The argon - entrained air - hydrogen flame has much the same characteris-tics as the air - hydrogen one and both are of use in determining elements having their resonance lines below 220 nm and when lower temperatures are required to avoid interference effects. Air - acetylene and air - propane flames absorb about 60 per cent. of the resonance radiation from elements like arsenic and selenium whereas argon and air supported hydrogen flames are considerably more trans-parent at the lower wavelengths required for these elements and only absor 200 PLATT about 15 per cent. of the radiation. A number of workers have recommended these flames particularly for arsenic and s e l e n i ~ m ~ ~ ~ - - ~ ~ ~ and for some other ele-m e n t ~ ~ ~ ' - ~ ~ ~ The sensitivity and detection limits for these elements are usefully increased but unfortunately owing to the lower flame temperature interferences are also liable to increase at the same time.A multi-slot burner head supporting three parallel flames provides the most satisfactory performance with these cooler flames presumably because the central flame is protected from disturbances to a greater degree than with a single-slot burner.116s118 Reynolds4 considered that tin is best determined with the air - hydrogen flame. Dagnall et ~ 2 . l ~ ~ showed the strongly reducing high temperature nitrous oxide - hydrogen flame to have con-siderable promise as an atom reservoir for a wide range of metals whereas Willis et found data that did not support this evidence and suggested only a limited usefulness for this flame compared with the nitrous oxide - acetylene system.The multi-slot burner was first described by Boling122 and has since been widely adopted as a burner capable of coping with solutions of high solids content. As already mentioned the protected inner flame is used for absorption and this leads to the marked reduction of flame noise. A more recent development that has found commercial application is the use of inert gas shielded or separated flames. Kirkbright et ~ 2 . l ~ ~ described an air -acetylene burner that was shielded from the atmosphere by a flow of nitrogen and the same workers also described a nitrous oxide - acetylene burner that was shielded by argon or nitrogen.12* The inert gas lifts off and separates the secondary diffusion zone leaving the primary and hottest zone for viewing with considerably less interference.Separation of the air - acetylene flame is particularly beneficial for elements with resonance lines below 200 nm where atmospheric oxygen entrained in the secondary zone of conventional flames absorbs strongly. In the determina-tion of arsenic and selenium improved detection limits result from the lower background ab~orpti0n.l~~ Separation of the nitrous oxide - acetylene flame pro-vides protection of the reducing atmosphere in the interconal zone and results in a larger population of ground state atoms of the refractory oxide forming elements.This is most useful in the determination of aluminium boron and silicon when a threefold increase in detection limit re~ults.1~~ During the period covered by this review Rubegka and Moldan have reported further investigations with long path absorption Considerable increases in sensitivity result and the best sensitivities are achieved under fuel rich condi-tions when the tube shields the flame gases from the oxidising atmosphere and separates the flame into primary and secondary reaction zones along the horizontal axis. Despite their almost exclusive use flames possess several disadvantage~,~~8~~~ and a number of workers have used various techniques for the flameless atomisa-tion of samples. Slavinll briefly discusses the non-flame methods none of which has as yet found any commercial application although it is possible that one or two may be offered as accessories in the future.L ' v o Y ~ ~ ~ used a lined graphite furnace electrically heated in an argon or nitroge ATOMIC ABSORPTION SPECTROSCOPY 201 atmosphere to produce atomic vapour from samples that are in either solution or powder form. Massrnannl3* also investigated the atomisation of samples in a graphite crucible heated in an argon atmosphere. West and Williams128 and Anderson et a1.l3l described an electrically heated carbon filament atom reservoir. This is a simple and efficient device which can readily be fitted or adapted for use with most commercial atomic absorption equipment and is capable of completing an analysis within 5 s and repeating it every 2 min with the filament having no memory effects and also being self-purging.These methods can all analyse very small samples with spectacular detection limits. Further work with high temperature furnaces has been described by a number of a ~ t h o r s . l ~ ~ - l ~ ~ Other flameless techniques used include 1asers,137$138 plasma t o r ~ h e s l ~ ~ - l ~ ~ and hot-wire amalgams or c o a t i n g ~ . l ~ J ~ ~ Monochromators The monochromator in an atomic absorption instrument does not fulfil quite the same function as it does in other types of spectrophotometer. As the source produces a series of monochromatic wavelengths the monochromator serves only to select the resonance line from unwanted radiation.52 Either a prism or a diffraction grating may be used as the monochromator, although commercial instrumentation now appears to favour the use of gratings.Reynolds31 discusses the general features and factors affecting the performance of monochromators and concludes that there is more to recommend a grating instru-ment than one utilising a prism. The same author also states that for most common metals a very moderate monochromator is quite suitable but for metals producing complex spectra or for which the sources possess poor output a better mono-chromator is necessary. Owing to the vast improvements in lamp technology the latter group of metals is few and diminishing in number. A new type of monochromator was described by Sullivan and W a l ~ h l ~ ~ and depends for its action on the absorption and re-emission of resonance radiation by the atomic vapour of the element being determined.The atomic vapour may be produced by the electrical heating of a small block of the particular metal or for metals of higher melting point and lower vapour pressure by cathodic sputtering as in a hollow-cathode lamp. Sullivan and W a l ~ h l ~ ~ discuss the applications of resonance monochromators and point out the favourable case of the Group I1 elements which have a resonance spectrum consisting of one line only. They mention the possible advantages in the design of instruments for the simultaneous determination of several elements. The same authors and also W i l l i ~ ~ ~ list the major advantages of these systems com-pared with dispersion types. The main advantages are that the effective resolution at about 0*001 nm enables the flame noise compared to the signal to be far less than that obtained with a dispersion monochromator where the slit-width is nearer 0.1 nm.A resonance monochromator is permanently tuned to the wavelength of the resonance lines it is designed to isolate so it can therefore withstand far more rigorous conditions and should be ideally suited to continuous routine use o 202 PLATT in-line systems. A conventional monochromator on the other hand is subject to thermal or mechanical drift. Constant calibration data may be facilitated by the over-all control that the monochromator has over the effective resolution this being independent of the resonance lines emitted by the source. Disadvantages include a limited life loss in sensitivity with increasing com-plexity of the resonance radiation spectrum and the possibility of having to use high apertures to achieve an adequate signal.A number of workers in the field have used resonance monochromators,~4~-1~0 but only one manufacturer has included them in commercially available equip-ment.151 Whether or not they become more generally accepted depends very largely on manufacturing Detector and read-out systems Almost all atomic absorption instruments are now fitted with photomultiplier detectors which enable the widest possible range of metals to be determined.44 Apart from the light of a particular wavelength originating from the source there will also be light of the same wavelength arising from the flame falling on the detector.It is necessary to distinguish between these two radiations as it is imperative that only that from the source is measured. This requirement is met by modulating the source radiation either mechanically or electronically and tuning the detector to the particular frequency of m~dulation.~~ Conventionally after detection the signal is amplified and displayed on a meter. Single-beam instruments have microammeters calibrated for transmission and absorbance measurements whereas double-beam instruments generally use null-point systems. Alternatively a chart recorder may be used to provide a permanent record and to display any signal fluctuations. In addition some type of recording device is essential for use with automated multi-sample handling.For atomic absorption spectroscopy it is convenient to use a logarithmic recorder that will automatically provide a read-out linear in absorbance units and hence providing the Beer - Lambert law is obeyed a linear concentration scale. It is even more convenient although no more accurate to use one of the digital concentration read-out or print-out devices offered by some manufacturers. If the Beer-Lambert law is not obeyed it is desirable to correct for the curvature a relatively difficult requirement. Kah1115~ described an instrument that provided a correction of the deviations from Beer’s law and a read-out directly in concentration. Integration techniques may be applied to the amplified signal either to pro-duce a steady noise-free r e a d i r ~ g l ~ ~ ~ or to accumulate low absorbance signals and recover them from the noise signal thus improving the signal-to-noise ratio.l= The majority of instruments have a scale expansion facility to increase small read-ings and improve the precision of measurement.For the determination of macro constituents an instrument with the ability to expand selected portions of the scale rather than the whole range from zero upwards is most useful. This is known as ‘zero suppression’ and when used in conjunction with an integrating system, gives most impressive results.36,1s ATOMIC ABSORPTION SPECTROSCOPY 203 W i l l i ~ ~ ~ considered there was no doubt that the growing demand for large numbers of routine analyses will lead to a widespread application of automated sampling techniques in conjunction with digital read-out and print-out facilities.To cope with the large amount of information resulting from automatic instrumen-tation computer techniques were described by Ramirez-Mufioz et ~ 1 . l ~ ~ More recently Wendt15s has described the application of a computer programme to handle all the calculations involved in converting percentage absorption readings into sample concentrations. Malakoff and c ~ - w o r k e r s ~ ~ ~ J ~ ~ have also described the application of computer techniques to atomic absorption methods. SlavinlG0 discussed the acquisition of data on a typewriter read-out which automatically punches paper tape for computer processing and this type of equipment has recently been announced comrner~ially.~~~ Automation According to Slavin ,11 the time required to introduce the prepared sample and record the signal is a small part of the analytical time and automation would not be a great practical advantage.Far more important is the time required to convert the absorption signal to actual concentration data and the time required to dilute the sample to the optimum analytical range. Slavin describes an apparatus for a fully automatic analysis that includes introducing diluting and mixing the sample , and the result in concentration terms is presented on a paper strip with sample identification numbers. Similar equipment to that described by Slavin is available commercially from at least one ma,nufacturer.lsl Dawson et aZ.162 described an automatic high speed scanning multi-channel instrument for the determination of sodium potassium calcium and magnesium in clinical samples by simultaneous emission and absorption measurements.Partial automation is readily available in the form of automatic sampling devices and automatic sampling plus dilution has been accomplished by connecting a Technicon AutoAnalyzer system to the nebuliser of an atomic absorption i n ~ t r u m e n t . l ~ - ~ ~ ~ A number of workers have recommended the use of automatic d i l u t e r ~ ~ ~ ~ J ~ ~ and syringesle8 to speed up solution preparation prior to atomic absorption measurement. When a manually operated instrument becomes overloaded with work the type of work involved will decide whether automation or a second manually operated instrument is the answer. Automation is worth considering only when the work consists of similar samples for a small number of elements.lel Reagents for atomic absorption spectroscopy With the rapid development and increasing popularity of the atomic absorption technique there was a need to consider the availability of reagents with a special degree of purity for use in the preparation of samples for analysis by atomic absorption spectroscopy.With this aim in mind a reagent sub-com-mittee of the Atomic Spectroscopy Group was set up in 1967 to approach the major reagent manufacturers. 204 PLATT The specifications of a list of reagents widely used as releasing agents and ionisation buffers and also inorganic and organic solvents was circulated to the suppliers. These reagents are used in large amounts relative to the concentration of elements actually being determined and need to be in a high state of purity with respect to those elements.Two further types of reagents were also considered. One was the supply of stock solutions of elements of accurately known strength and as such solutions are much diluted in use the accent was on accuracy rather than purity. The second consideration was for the provision of suitable chelating agents for the extraction of metals into organic solvents. Amongst the latter, ammonium pyrrolidine dithiocarbamate has been widely recommended for the extraction and concentration of heavy metals. Hopkin and Williams Ltd. have published a monograph on the use of ammonium pyrrolidine dithiocarbamate,leg Excellent mutual progress was made and it is now possible to purchase from B.D.H.Chemicals Ltd. and from Hopkin and Williams Ltd. a substantial range of the reagents mentioned above. Additionally Kodak Ltd. market in the United Kingdom a range of atomic absorption reagents from the Fischer Scientific Com-pany of America and Johnson Matthey offer their Spec. Pure Chemicals. For the analyst who prefers to make his own standard solutions or for those not commercially available Roth17* provides details for the preparation of 1000 p.p.m. solutions for sixty-five elements. Standard stock solutions (not lead) should be stored in polythene bottles and should be freshly diluted into standard working solutions each day. Experimental Considerations According to Ramirez-Muiio~,~~~ atomic absorption like most instrumental methods requires several conditions to be fulfilled to guarantee acceptable analy-tical results.These are (i) sufficient knowledge of the fundamentals by the analyst who has to plan new applications and realistically interpret the results; (ii) the availability of suitable equipment and (iii) a preliminary study of operating conditions for each analytical system including results of preliminary calibration data (linearity and sensitivity as well as a study of interferences both physical and chemical). As RubeSka and Moldanf4 point out detailed instruction manuals come with the equipment and it is only necessary to mention one or two general hints in this connection. The instrument should be placed on a solid ‘island type’ bench or trolley with an exhaust hood placed over the burner to efficiently remove the various toxic fumes.Sensitivity of the detection system may be altered by changing either the slit-width or the electrical gain setting. In general the broadest slit compatible with an admissible spectral band-pass is used or according to WalshF5 the slit-width used should be the maximum that will both permit isolation from all other lines to be achieved and give a maximum signal-to-noise ratio at the detector. Ramirez-MuiiozS points out that an operator facing the problem of increasing the instru-mental sensitivity to a maximum has to expect an increase of noise. Nois ATOMIC ABSORPTION SPECTROSCOPY 205 suppression devices or damping can then be used which corresponds to an increase in the time constant although this is not advisable when the sample size is limited.It is important when using wide slits to check that the signal caused by background does not result in a loss of absorption sensitivity. Price13 advises that although any continuous background signal is not recorded it is nevertheless good practice to keep it minimal otherwise the detector may become saturated. The use of highly luminous flames should therefore be avoided. Ramirez-Muiioz17* has dis-cussed the calculation of signal size and signal-to-noise ratio and the same author and co-workers have detailed the relationship between precision and sensitivity24 and between accuracy and More recently Ramirez-Muiioz has dis-cussed the application of sensitivity diagrams to atomic absorption spectroscopy.174 These diagrams have been applied in plane and space forms to facilitate the repre-sentation and comparison of the analytical behaviour of different elements in terms of sensitivity.They also aid in the study of single elements under diverse experimental conditions the performance of different instruments and in the variation of sensitivity in the presence of a releasing agent(s) under the same or differing experimental conditions. Limiting interference ratios can also be easily calculated with the help of these diagrams. It is necessary to clearly distinguish between sensitivity and limit of detection, and reference to the diagrammatic explanations given by Elwell and GidleyJ6 Slavinll or the Eel Bulletin36 is recommended. Detection limit is a valuable guide to instrument performance for a particular deterrnination.Grunder and B ~ e t t n e r l ~ ~ studied various instrumental parameters and com-pared different types of burners and hollow-cathode lamps from an analytical point of view the information obtained being used to study precision and sensiti-vity. They concluded that the analytical curve appears to be the single most important piece of information available to the analyst. The curve and its asso-ciated slope together with the signal-to-noise ratio can give information on how well the burner and the hollow-cathode lamp are functioning. Interferences Lewis176 indicated that despite their apparent simplicity atomic absorption methods require careful laboratory technique and a thorough knowledge of potential sources of error and their control.According to GrantF5 the most serious aspect of interferences is a lack of information about them. Simple methods for their correction can usually be devised once they are well understood. Spectral interference caused by the presentation of unwanted light to the detector may be caused by the spectrum lines of unwanted elements molecular bands and background from the flame or from major elements. With thehigh source resolution and modulation techniques used in modem equipment this type of interference is hardly ever encountered in atomic absorption.177 Some spectral line interferences have recently been observed but these are uncommon and can usually be avoided if the analyst is aware of their possible existence.178 Koirtyohann and Pickett17* have discussed the background interference resultin 206 PLATT from molecular absorption which is obtained when determining a number of elements in the presence of the alkali halides and alkaline earth elements.Earlier observations reported by other workers indicating light losses because of scattering by particles in the flame were not observed by Koirtyohann and Pickett and they considered that previous workers were actually measuring molecular absorption. They concluded that the number of specific spectral interferences found is small, and freedom from such interferences will continue to be one of the major advantages of the atomic absorption method. This is particularly true when using the hotter nitrous oxide - acetylene flame.14 RubeSka and Moldan14 and Kahnl80 consider that the question still exists as to what proportion of the light loss is caused by scatter rather than absorption.Kahn makes no effort to distinguish between them as the analytical effect of both phenomena is the same. Cooke and Price177 mention the mutual spectral interference encountered when sodium and potassium occur together caused by the intense background continua. This effect is nullified either by adding the correct amount of the second element to the standards when determining the first or by adding a large excess of the second element to both standards and sarnples if the sample content of this element is variable. A number of authors have discussed the use of non-absorbing lines to correct for background ab~orption,llJ~J8~ although Koirtyohann and Pickett17@ considered this to be unsatisfactory in many cases owing to the possible presence of structure in the background absorption and the fact that frequently no suitable non-resonant line is available.Kerberls2 has described a method to select non-absorbing wave-lengths by using the scanning accessory normally provided for flame emission work. Koirtyohann and Pi~kettl'~ used a continuous source to measure and correct for background absorption in the ultraviolet region. More recently Kahnlso has described a background compensation system in which the continuous light from a deuterium arc and that from a hollow-cathode lamp are alternately passed through the flame and a ratio taken to cancel out any background absorption.Kahn discusses other compensation methods and criticises the non-absorbing line method because some elements such as zinc cadmium and mercury have no suitable lines. The same author points out that the method of standard additions does not compensate for background and that the easiest method currently used is to have a blank solution adjusted to the same composition as the sample but without the element of interest. Unfortunately the preparation of such a blank is not always feasible. Kahn28 has listed the three major types of interference as chemical ionisation and bulk or matrix. The most common interference is chemical in nature arising from a failure to break the chemical bonds formed between the element being determined and other materials in the matrix solution during evaporation of the sample mist in the flame.Rains2%eported the necessity of optimising various para-meters to eliminate or control chemical interferences. These include oxidant and fuel flow-rates flame temperature the flame region used for the measurement and the sample medium. Parsons and VVinef0rdnerl8~ have provided an excellent dis-cussion about the optimisation of critical instrumental parameters. The ATOMIC ABSORPTION SPECTROSCOPY 207 considered the importance of the various parameters and indicated those that can be systematically varied to obtain optimum performance including fuel and oxidant flow-rate sample solution flow-rate temperature of cloud chamber height of measurement in the flame slit-width and flame path length. RohledeP described a method of providing constant and adjustable sample solution feed rates by using peristaltic or piston pumps.Cellier and S t a ~ e l ~ ~ discussed a statistical approach to the determination of optimum operating conditions that has several advantages over the traditional method of varying one factor at a time. As already pointed out the use of a higher temperature flame such as that provided by the nitrous oxide - acetylene mixture is a simple and efficient means of removing the majority of chemical interferences although it is still very neces-sary to optimise the other parameters mentioned by Rains. The use of hotter flames increases ionisation interference and this must be suppressed by the addition of relatively large amounts of an easily ionisable element such as sodium or potas-siurn,2,6$%J77J86 Conversely the presence in a particular sample of an easily ionisable metal will enhance the absorption of other elements and produce errone-ously high results for the element being measured52 unless the standards are properly compensated.The majority of workers stress the importance of adjusting the burner height or the region of measurement in the flame to obtain maximum absorption,~~20,34~1s3~1~~1~7~18~ but Reynolds,3l working with very low gas flow-rates, found it unnecessary to have a separate adjustment for burner height. He simply kept the support gas flow constant and adjusted the fuel flow-rate for maximum absorption. Elwell and Gidleys justify a careful investigation of cationic inter-ference with different fuel-to-oxidant ratios and heights of measurement because by so doing interferences can be minimised and sometimes eliminated.It is useful to remember that altering the burner angle relative to the optical path effectively alters the flame cell size and offers one means of coping with various concentrations of the required element.3~2e Whenever it becomes more expedient to use cooler flames the analyst must make a serious appraisal of all the possible interferences and effectively counteract them. In such cases chemical interference can be minirnised by the addition of releasing or chelating agents. When added in a sufficient amount these agents should restore the absorption of the required element to the value obtained in the absence of the interferent (s).Price13 advises that interference effects caused by other major constituents in the sample should be investigated one element at a time. Elwell and Gidley6 suggest the addition of various amounts of anions to a constant dilute solution of the required element. Whenever interference is observed, a corresponding amount of the offending anion should be added to all of the standard solutions. It may also be advisable to examine any matrix cations in 10,000 fold excess over the element to be determined. The possible mechanisms of chemical interference have been discussed by a number of workers,6~13~14~189-192 and RainsWi provides a useful table linking the required element with typical interferents and the type of releasing - chelatin 208 PLATT agent required together with the appropriate references to any work carried out in this respect.It is interesting to note that although lanthanum is more effective than glycerol or EDTA as a releasing agent for the determination of magnesium or calcium in the presence of aluminium if glycerol is used in conjunction with lanthanum the releasing action is considerably more efficient than with lanthanum itself. RubeSka and hLZanlg0 observed a similar improvement when using mixtures of lanthanum with EDTA or 8-hydroxyquinoline. Willis20 provides evidence to support the view that the principal factor in the atomisation of elements subject to chemical interference is the rate of evaporation of the particles formed from the droplets of solution in the flame. The atomisation efficiency is dependent both on droplet size and on the position in the flame at which the measurement is made and chemical interference is not always least in the part of the flame where the concentration of metal atoms is a maximum.Many of the conflicting data in the literature on chemical interferences arise from a lack of appreciation of these facts. Koirtyohann and Pickettlg3 reported a new type of interference in the nitrous oxide - acetylene flame from the presence of mineral acids in the sample solution. Enhancement of calcium zinc and aluminium absorptions in the presence of 0.3 to 1.0 M sulphuric phosphoric and perchloric acids or 1 to 2 per cent. of ammonium phosphate and sodium chloride were ob-served near the centre of the flame. Nitric acid produced a smaller enhancement, while hydrochloric acid and organic materials such as sucrose and glycerol produced no effect.The enhancements are not changed by variations in fuel or oxidant flow-rate but are reduced when the optical path is higher in the flame and dis-appear at about 15 mm above the primary reaction zone. In every case the en-hancement disppeared or was drastically reduced when the burner was turned at a right angle to the optical axis. The expected decrease in absorption because of reduced nebuliser efficiency was noted at higher concentrations of the matrix materials. Thus these workers concluded that a new type of interference was present involving the spatial distribution of the sample within the flame and although the cause is not yet known they postulated that it involved the rate of diffusion of the salt particles outwards from the centre of the flame.Sastri et aZ.lg4 investigated the effect of the metal - oxygen bond on the sensi-tivity of absorption measurements and concluded that for metals that form refrac-tory oxides in flames low sensitivities are obtained if the metals are used as simple or oxy-salts or even as complexes with ligands having an oxygen atom as the donor. The sensitivities of these metals can be enhanced if the metal - oxygen bond in solution can be avoided or reduced by using metallocenes or fluoro-complexes in aqueous or organic solvents. These workers extended their studies195s196 to deter-mine the r6le of mixed metal polynuclear oxygen-bonded species in solution reducing the sensitivity of the metal to be determined.Hartlage197 reported a depressive interference exhibited by various amines on the absorption signal of several metal ions in the air - acetylene flame. This inter-ference which is well illustrated by the presence of 6 per cent. of triethylamine in ethanol causing a 40 per cent. error in the determination of cobalt is apparentl ATOMIC ABSORPTION SPECTROSCOPY 209 explained by the metal - amine complexes not dissociating at the temperature of the air - acetylene flame. It may be avoided by using the nitrous oxide - acetylene flame. Pricelo4 in discussing the analysis of biological materials by atomic absorption spectroscopy mentions the possibility of using dry or wet ashing procedures when the interference effects are not understood or when a method free from such effects is required.Ashing may also be used to separate metals from an organic matrix or to effect a concentration step. The ashing procedure must be carefully chosen to avoid volatilisation of the required element and it should result in a purely inorganic solution which may then be compared directly with properly constituted inorganic standard solutions. This procedure can be used as a check against more rapid procedures involving only dilution or de-proteinisation. Bulk or matrix interferences are generally of a physical nature caused by the viscosity surface tension and vapour pressure of the solvent used in the sample solutions being different from that of the standards. According to Slavin,ll most workers have found it necessary to match the standards and samples with respect to the materials that are present in excess of 1 per cent.of the total solution. The match is usually made by adding salts to the standards to equalise the concentra-tion of major constituents. Ramirez-Muiiozlg8 has discussed the analysis of systems with high sodium contents and concluded that in some cases involving the use of laminar flow burner and heated cloud chamber the low interference ratios obtained in the presence of high sodium chloride concentrations enabled determinations to be made with uncompensated standards. At levels of interference higher than the limiting interference ratio standards should be compensated with equivalent concentrations of sodium ions. The presence of an organic solvent in the sample solution will affect the physical nature of the solution and cause enhanced ab~orption,~~s~~ thus making it necessary to compensate the standards ac~ording1y.l~~ The enhancing effect of a miscible organic solvent mixed with an aqueous sample solution may be utilised to increase the response from a required element,4Js9 although Rains25 points out that such an addition is of limited value because the increase in sensitivity is less than 5 fold and the element required is inevitably diluted.MitchelI2O0 has discussed the r6le of the solvent in flame photometry and indicates that the combustion of the solvent mixture will alter flame size flame temperature and flame background continuum. Hence it becomes necessary to re-adjust completely the instrument operating conditions.The direct nebulisation of an organic solvent solution is useful in the analysis of water insoluble materials such as lubricating oils,201 vegetable oils and and has been used for the determination of metals in phospholipids.206 Orren206 concludes his paper on the basic principles of atomic absorption spectroscopy by commenting that the technique provides a rapid sensitive and precise analytical method which while not interference free is much less subject to interference than are most other techniques of comparable sensitivity and scope. RubeBka and Moldanf4 indicate three ways of eliminating interference when it does occur the most suitable method from the andytical point of view being th 210 PLATT addition of buffer solutions to both the standards and samples.Alternatively it may be possible to prepare standards that imitate the sample composition or the method of standard additions may be used. There are two further ways of reducing interference. For most purposes the choice of wavelength is limited to a selection of the most absorbing line but according to Reynolds4 it is sometimes necessary to choose a less absorbing line to avoid the high background noise that can occur over certain wavelength ranges. RubeSka and Moldan14 also point out a number of cases when a line other than the commonly recommended one may be used to advantage. Also to avoid excessive dilutions it is often convenient to use a less sensitive line than the recommended 0ne.4969207-210 Solvent extraction frequently removes chemical as well as physical interferences and allows concentration of the required element whenever this is necessary.25 Extraction into an organic solvent after complexation with a suitable reagent is an effective method as the organic solvent is itself responsible for a further increase in the sensitivity of many elements over and above the concentration factor.lo4 Highly selective extraction is not necessary and is in fact not desirable.Separations should be between groups of elements rather than between single e l e m e n t ~ . l ~ ~ ~ ~ ~ l The most popular complexing reagent for this purpose is ammonium pyrrolidine dithiocarbonate (APDC) and dithizone has also been used extensively. Methyl iso-butyl ketone is the most commonly used solvent. Carbon tetrachloride and chloroform can be used but give considerably less enhancement and also introduce the possibility of forming poisonous gases on c ~ m b ~ ~ t i o n .~ ~ ~ ~ ~ J ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ Although the latter disadvantage can be overcome it adds a somewhat more time consuming step.lsg The use of APDC makes possible the simultaneous extraction of approximately fourteen elements over a consider-able pH range.14 A number of authors have discussed the use of APDC for the determination of various elements including arsenic and se1enium169~212~213 as well as alternative complexing agents and solvent ~ y ~ t e m ~ . ~ 1 ~ 9 ~ ~ * ~ ~ ~ Munro200 has dis-cussed the effects of acidity and extraction ratio with reference to the APDC-copper complex and emphasises the need to equalise these factors for variable samples and standards.Campbel121e points out the need to saturate the organic solvent with water when adjusting the zero base-line and also the fact that organic solvent becomes a fuel when nebulised and a preliminary adjustment of the fuel -oxidant mixture should be made. Chau and Lum-Shue-Chan217 mention the serious errors involved if the equilibrium state is altered while using volumetric flasks for the extraction. Before leaving the subject of interferences it is convenient to mention two further points that have been reported to cause erratic and inconsistent results. Shepherd and Johnson218 noted the presence of acetone in acetylene cylinders causing an unstable and noisy flame particularly if the cylinder had been stored in a horizontal position.This was later ~ e r i f i e d ~ l ~ ~ ~ ~ ~ and the acetone removed by passing the acetylene stream through an activated charcoal column. The second undesirable effect was reported by Scheub and Stromsky221 and concerned the presence of dissolved air in distilled water. Air bubbles in the nebuliser syste ATOMIC ABSORPTION SPECTROSCOPY 21 1 apparently caused a steady decrease in sample uptake and hence a steady down-ward drift in absorption. The problem was overcome by de-aerating the distilled water under reduced pressure. Application of method Once the sample is in solution it is often only necessary to dilute the sample to the optimum analytical range which is usually 20 to 200 times the limit of detection although concentrations nearer the detection limit may be determined with less p r e c i s i ~ n .~ ~ J ~ ~ ~ ~ s ~ ~ If the dilution is such that the total sample content is less than 0.1 per cent. of the solution the standards will usually be satisfactory when they contain simply known amounts of the element of interest in the same solvent as the sample. Whenever the solids content is greater than 0.1 per cent. it is usually necessary to match the standards with respect to the principal sample constituents. The match need not be better than &25 per cent. in most cases. Standard solutions may of course be prepared containing more than one elementll and must also contain any releasing chelating or suppressing agents that may be necessary in the sample analysis. With the instrument conditions optimised the standard solutions are used for testing sensitivity limit of detection reproducibility and to prepare a calibra-tion curve.Ramirez-Muiio~~~~ lists five principal methods of calibration. To obtain a calibration curve absorbance obtained either directly or by calculation is plotted against concentration. Theoretically the resulting line should be straight but in actual practice it will for various reasons nearly always curve slightly towards the concentration axis0,11J3s25 This method is suitable for the majority of routine work provided that the number of standard solution readings is increased over the non-linear portion of the curve and standard solutions are measured regularly between samples to observe any changes that may occur for any reason during the total analysis time.Scale expansion should be used to measure very low concentrations near the detection limit. Under these conditions the relationship between percentage absorption and absorbance is reasonably linear and it is usually unnecessary to convert the scale expanded reading to absorbance before p10tting.l~~~~ When measuring very small absorbance values a recorder trace offers better precision. If a measurement with high precision is required as in macro determinations, two standard solutions with slightly higher and lower concentrations of the re-quired element than the sample solution are prepared. Scale expansion is used so that the lower standard assumes zero absorbance and the scale is concentrated on the particular range of interest.Certain instrumentation permits this so-called zero s~ppression.~~J~ It is not always feasible to prepare standards that match the sample solution and it may be advantageous to use a standard addition method as described by a number of authors.g~11~14~25~171 However this procedure should be used only as a last resort and then only for determining elements at trace level where a lower leve 212 PLAT" of accuracy may be acceptable. As pointed out in the section Experimental Con-siderations the method of additions does not differentiate between residual con-centrations and background absorbance. Having developed the basic method including calibration data samples of known composition e g . National and International Standards or chemically analysed samples should be checked against the calibration data and recovery experiments of known amounts of the required element will determine whether or not any interference effects have been completely over~ome.~J~ Ramirez-Muiioz222 has discussed the calculation of percentage recovery.Reproducibility and there-fore analytical accuracy should in the optimum concentration range be of the order of &1 per ~ e n t . ~ ? ~ ~ Applications Even the more recently published applications of atomic absorption spectro-scopy are far too numerous to be fully covered in this review. The ones discussed have been chosen either because they are somewhat novel or because they form part of the general developments that have taken place. The determination of macro components The major emphasis in the literature about atomic absorption has been its capability for trace level determinations and in fact as R ~ o n e y ~ ~ ~ points out the technique was specifically designed for this field.However many workers have reported the successful determination of metals present in high concentrations. Price13 indicates that the method can often be used for the determination of alloy-ing or comparatively major constituents because the reproducibility where the concentrations range is optimum should be about &-1 per cent. and this can be improved by replication. The analyst must of course be particularly careful when preparing the solution for analysis. When determining major constituents precision rather than sensitivity is the important factor; Kahn2% has indicated a procedure for obtaining optimum precision.It is necessary to observe the following three criteria (i) that the element required is present in the solution at a concentration within its optimum analytical range; (ii) the sample is sufficiently homogeneous so that a truly repre-sentative sample can be obtained; and (iii) the dilution of the sample does not contribute any error. Kahn discusses the effects of dilution and indicates that it can be avoided by reducing the optical path through the flame or by using a less sensitive resonance line although the latter must be carefully chosen. The same author also recommends the zero suppression technique when optimum precision at high concentration is required. Feldman et aZ.225 have reported the determination of major constituents by using an internal standard technique.These workers have analysed a wide range of materials and show the technique to be very accurate and to exhibit increased precision compared with the direct method ATOMIC ABSORPTION SPECTROSCOPY 213 The types of materials that have been analysed for major constituents by atomic absorption fall into two broad categories metallurgical and silicate types including cements. Sattur226 described the routine analysis of non-ferrous alloys and obtained good agreement with certificate or chemical analyses in many cases. Aqueous standards were used and no sample preparation other than acid dissolution was necessary. Sattur determined tin potassium copper bismuth lead antimony, zinc indium and cadmium all in concentrations of greater than 5 per cent.Copper was accurately determined in bronzes to concentrations as high as 83 per cent. Meddings and Kaiser227 determined nickel cobalt copper and iron and compared the precision with that obtained by wet chemical methods. By using equipment that included digital read-out and high brightness lamps these workers reported the coefficient of variation for routine analysis to be 0.3 to 0.8 per cent. for atomic absorption spectroscopy compared with 1 to 2 per cent. by wet chemical analysis. Ellis and Roger@ hawe determined lead and zinc in non-routine samples by using automatic dilution apparatus. The problem of sampling was overcome by success-ive grinding and rifBing. Welcher and Kriege228 developed specific procedures for the precise determination of major alloying elements in high temperature nickel-base alloys.These workers made a comprehensive study of techniques used for suppressing the numerous chemical interferences and found that for the most accurate analysis close matching of samples and standards with respect to both cation and anion concentrations was necessary. They preferred to extend the working range by rotating the burner rather than diluting the solutions. Working at the optimum concentration excellent agreement was obtained between the atomic absorption results and the National Bureau of Standards or wet chemical values for chromium aluminium molybdenum cobalt titanium iron tungsten, niobium tantalum and vanadium. Niobium and tantalum were determined after extraction into methyl iso-butyl ketone from an aqueous phase that was 10 M in hydrofluoric acid and 6~ in hydrochloric acid.The results quoted showed good precision for each element determined. Johns and Price229 described a comprehen-sive scheme for the analysis of brasses and bronzes in which the usual elements associated with these alloys were determined at levels from 0.1 to 10 per cent. The air - acetylene flame was suitable for the determination of iron lead manganese, nickel and zinc. Apart from viscosity differences introducing physical interference, no other interference was found from elements normally present in brasses and bronzes. When using the hotter nitrous oxide - acetylene flame ionisation inter-ferences were sometimes encountered that appeared as slight enhancements of aluminium and tin absorptions by copper.Both interferences were overcome by adding an equivalent amount of copper and solvent acid mixture to the standards. Whenever several elements present in a sample required the same dilution it was possible to use multi-element standards. Agreement with the certificate values was good and demonstrated the suitability of the recommended methods for routine analysis. discussed the application of atomic absorption spectro-scopy to the analysis of major constituents in alloys and steels. Zero suppressio 214 PLATT with scale expansion was used to achieve relative precisions of 0-1 to 0-3 per cent. The use of less sensitive lines and shorter flame path length was also explored. Cobb and Harrison231 determined aluminium in iron ores slags and refractory materials and achieved a reproducibility of k0.55 per cent.in the 20 to 70 per cent. range and the mean result agreed closely with the accepted chemical value. Langmyhr and Paus have published a series of papers concerned with the analysis of inorganic siliceous materials by atomic absorption spectroscopy and this work has been summarised by the same authors.232 A hydrofluoric acid decom-position technique was used which retained the silicon in solution. Primary standard solutions were prepared from highly purified reagents and details of potential interferences and the means of overcoming them are given. Good agreement with certificate values was obtained for silicon aluminium iron, magnesium calcium sodium and potassium and for titanium manganese chrom-ium and vanadium at somewhat lower levels.Langmyhr and PausSB have also provided some data on the intra-laboratory precision of atomic absorption analysis as applied to silicate rock. This precision can be classified as good to moderate and compares favourably with other methods. During the last 2 years a number of authors have described methods for the decomposition of silicates followed by a comprehensive analysis by using atomic absorption spectroscopy. B e r n a 9 found that a fluoboric - boric acid system provided a f avourable decomposition medium for the rapid and reliable determina-tion of silicon aluminium titanium and vanadium. The acid matrix used was found to eliminate the need for releasing compensating or complexing agents and it was not necessary to closely match the matrix composition to that of the un-known sample.Thus the use of single element standards for the analysis of complex materials is brought nearer. Hence whereas Langmyhr and Paus list interferences in several cases Bernas finds no inter-elemental effects. Van Loon and used lithium metaborate fusion and conducted extensive tests of possible inter-ferences. They found it necessary to correct for different types of interference in the determination of aluminium potassium sodium titanium magnesium calcium and manganese. No interference was found in the determination of iron and silicon. Similar fusion techniques using lithium borates followed by solution in nitric acid have been proposed by a number of other ~ o r k e r s .~ ~ ~ - - ~ ~ ~ In each case interferences were easily overcome and the precision was comparable with that obtainable by conventional methods of analysis. O r n a r ~ g ~ ~ ~ verified the results quoted by Van Loon but disagreed with Bernas over the use of cations which the latter worker considered to be a disadvantage. Karmie Galle242 developed a routine procedure for the determination of major constituents in geological samples by atomic absorption in which it was unnecessary to change from nitrous oxide-acetylene to air - acetylene to analyse non-refractory elements. The samples were obtained in solution by treatment with hydrofluoric acid followed by fusion with potassium pyrosulphate. CampbellM1 determined silicon (>5 per cent.) in alumi-nium alloys after treatment with sodium hydroxide and hydrogen peroxide.It wa ATOMIC ABSORPTION SPECTROSCOPY 215 necessary to prepare standards containing approximately the same amounts of aluminium and total solids as the sample solutions. The analysis of cement by atomic absorption spectroscopy has been discussed by a number of people; Crow et aLN3 found that a simple acid digestion step was sufficient sample preparation. The presence of silica affected the aluminium determination but this and other matrix interferences were compensated for by calibrating with National Bureau Standards similarly prepared. According to these workers cement chemists would probably accept errors of 1 per cent. for the determination of any oxide in cement except calcium.Satisfactory accuracy could be obtained for this element by using calibration curves but the determina-tion was limited by unacceptable precision and it was necessary to resort to the zero suppression technique to overcome this problem. Other workers have also recommended this technique for the determination of calcium in ~ e m e n t ~ ~ J ~ and very impressive results were obtained. Crow determined aluminium in cement with good precision and accuracy by using standard cement samples to prepare a standard curve. Capacho-Delgado and Manning2u obtained good precision for the determination of aluminium in cement by using aqueous standards containing about the same amount of calcium chloride and hydrochloric acid as was present in the sample solutions. Roos and PriceN6 determined calcium in cement without resorting to the zero suppression technique.They used an emission burner head with an absorption path of 1 cm to reduce the dilution factor but obtained a standard deviation of 0-7 per cent. which might not be acceptable when compared with standard gravimetric procedures giving nearer 0.2 per cent. The determina-tion of silicon also tended to lack precision. However one advantage of this procedure lay in its use of simple aqueous standards with the addition of lanthanum as a releasing agent. Elements determined in the nitrous oxide - acetylene flame Ramirez-MuiiozlsS recommended the use of the nitrous oxide - acetylene flame for thirty-three elements and in addition for the alkaline earth metals when high chemical interference is expected.Amos and Willisgs have determined the sensitivity at different wavelengths for twenty-seven elements many of which show none or only litttle absorption in the air - acetylene flame. Willisg4 lists sensi-tivities and detection limits for thirty-five elements and also indicates the likely applications for the determination of these elements. Shifrin et aZ.l10 reported recent results with a nitrous oxide - acetylene burner for twenty elements includ-ing seven that were normally determined in the air - acetylene flame but the analytical task was simplified by retaining the same flame throughout. Alkaline earth elements Amos and WilIisg8 called these the border-line elements. They can be deter-mined in an air - acetylene flame but generally show improved absorption in the hotter nitrous oxide - acetylene flame.Slavin et aZ.246 have illustrated the improve-ments shown when determining calcium or barium in the presence of phosphate 216 PLAT" and also point out the ease with which strontium in cement can be determined in the hotter flame compared with using the air - acetylene flame. In discussing the severe interference of phosphate on calcium Ulrich and Ramirez-Muiio~~~ indicated that the simple compensation of standards with phosphates leads to very small signals and it is better to use hot flames and to help the system with the action of some releasing agent (lanthanum) when massive amounts of phosphate are present. The lanthanum will also act as an ionisation buffer. The well known depressive effect of aluminium on calcium is overcome by using the nitrous oxide - acetylene flame together with lanthanum4 or by the addition of aluminium to the standards.25 Barium is remarkable for its poor sensitivity and is better determined in the nitrous oxide - acetylene flame.4J4 Kerber and BarnetF prefer to use an air -acetylene flame for the determination of barium in the absence of interferences, but when these are present or at lower concentrations the nitrous oxide flame is more convenient.At even lower concentrations these workers chose flame emis-sion with a nitrous oxide - acetylene flame as the best method. Ionisation must be suppressed by the addition of a suitable metal. Strontium is similar to calcium in its behaviour except that owing to the relatively low contents of strontium in natural materials the sample solutions cannot be diluted as much and the con-centration of interfering elements is thus likely to be higher than for calcium.By using a nitrous oxide - acetylene flame however chemical interference is efficiently overcome including that provided by iron(II1) .248 Magnesium is considerably more sensitive than calcium and its resonance line at 285.2 nm lies in the most favourable wavelength region. Hence high dilution can be used and this limits interferences. For example phosphates have at the most only a slight depressive effect. Interference caused by aluminium or silicate is not encountered when using the nitrous oxide - acetylene flame.14@9~250 The presence of alkali metals causes an enhancement interference when determining any alkaline earth metal in the hotter flame and the standards should be compensated accordingly or a large excess of the alkali metal added to both samples and standards.Klein and co-workers1@ adopted a more novel approach by coupling an auto-analyser to an atomic absorption instrument and simultaneously determining calcium and phosphate. An air - acetylene flame with lanthanum as the releasing agent was used but a nitrous oxide-acetylene flame could be used in such a system. Aluminium Until the advent of the nitrous oxide- acetylene flame aluminium was excluded from those elements determined by atomic absorption. At the present time aluminium is easily determined in this flame with few interferences. Potas-sium should be added to both samples and standards as an ionisation buffer.ll Ramakrishna et reported on the determination of aluminium and indicated that low concentrations could be determined accurately in the presence of various other ions and compounds.P a w l ~ k ~ ~ ~ determined aluminium in soil samples an ATOMIC ABSORPTION SPECTROSCOPY 217 obtained results that compared very well with those obtained by gravimetric analyses. Laflamme2= also reported the atomic absorption determination of alumi-nium in soils to be precise rapid and very useful. Van Loon254 obtained results agreeing well with those obtained by standard procedures in the analysis of high silica materials. The method used a hydrogen fluoride - sulphuric acid dissolution to remove the major component-silica-and is readily recommended.Dagnall et uZ.120 found the nitrous oxide - hydrogen flame showed considerable promise for the determination of aluminium and considered it worthy of further examination. Beryllium This element is barely detectable in the air - acetylene flame but according to Amos and Willis9* the use of a nitrous oxide - acetylene flame enables its deter-mination with a sensitivity approaching that for magnesium. Ramakrishna et ~ 1 . ~ ~ ~ reported a sensitivity of 0.025 p.p.m. and found a lack of interference from several ions and compounds. Manning255 confirmed the quoted sensitivity and found the addition of ionisation buffers to be unnecessary. B o k o ~ s k i ~ ~ ~ found the atomic absorption determination of beryllium in biological materials to be sufficiently rapid and sensitive for use as a monitoring procedure.provided a method for determining small amounts of beryllium in aluminium alloys. Boron Boron is not detected in an air - acetylene flame but is readily determined in aqueous solutions in the nitrous oxide - acetylene flame.ll Boron is not significantly ionised in the hotter flame and Manning255 found the sensitivity to be 35 pg ml-1 for 1 per cent. absorption with the detection limit at 6 pg m1-1 in fuel-rich flame with a 2 -nm slit-width. Bader and Brandenbe~-ger:~~ like Manning used ten times scale expansion to determine boron. They analysed biological materials and found a detection limit of 15 pg ml-l with a O.65-nm slit-width. Bader and Brandenberger determined sub-toxic levels of boron in serum and urine and solubilised tissue by a wet ashing procedure; dry ashing led to severe boron losses.H a r r i ~ ~ 5 ~ found it absolutely necessary to use the optimum conditions to obtain reproducible results when analysing low concentrations of boron flame conditions and burner height adjustment were both found to be critical. Owing to the low absorbance the method of standard additions was used by Harris to determine boron in potassium chloride. The particular isotope distribution in the sample made little difference to the final evaluation thus confirming earlier reports by other workers. Chromium It is well known that iron seriously interferes with the determination of chromium in an air - acetylene flame.260-262 Rarnirez-Mufioz and Roth262 con-sidered the suppression observed to be apparently caused by the formation o 218 PLATT iron - chromium compounds which axe hard to dissociate at the temperature of the air - acetylene flame.These workers suggested the use of a nitrous oxide - acetylene flame to help in overcoming the problem. Roos~~O found it necessary that the oxidation state in the sample solutions should be the same as that in the standards when using the air - acetylene system. Chromium(II1) gave considerably greater sensitivity than chromium(V1). Also in the cooler flame maximum absorbance and interference effects were found to be critically dependent on the flame gas mixture used and on the burner height. Investigations reported by Rooney and Pratt261 found that the nitrous oxide -acetylene flame overcame the inter-element effects and gas ratios were much less important and although there was still an effect from the presence of iron it was an enhancement rather than a suppression.It could be completely overcome by the addition of a small amount of iron to the final solutions. Wilson263 determined chromium in aluminium alloys by using a nitrous oxide - acetylene flame and found excellent agreement between the classical and the atomic absorption procedures. It was unnecessary to use standard solutions containing aluminium. Germanium Germanium shows some absorption in the fuel-rich air - acetylene flame but to attain really useful results a high temperature flame is required.ll Popham and Schrenk2G4 reported that by proper choice of experimental conditions germanium could be determined in a fuel rich nitrous oxide - acetylene flame with a detection limit of 0.5 p.p.m.and a sensitivity of 3 p.p.m. in a 50 per cent. aqueous acetone solution. This represented an approximately 2-fold enhancement over pure aqueous solutions. Interferences were more pronounced in the acetone medium. Kirkbright et aZ.,124 using separated flames obtained a significant improvement in the detection limit and extended the linear portion of the calibration graph. Molybdenum Molybdenum may be determined in either the fuel-rich air - acetylene or nitrous oxide - acetylene flames with about equal sensitivity. It is in fact one of the border-line Molybdenum is similar to chromium in that iron seriously interferes in an air - acetylene flame,14*260 and Kirkbright et aZ.266 recom-mend the use of a nitrous oxide - acetylene flame to remove interferences other than iron and then to compensate the standards and samples when necessary by the addition of iron.Absorbance with this flame is not as dependent on the flame composition as it is with an air - acetylene flame. Kirkbright et aZ.124 improved the detection limit and linearity of calibration by using separated flames. Ramakrishna et aZ.266 reported that most ions affect molybdenum absorption in the nitrous oxide supported flame some suppressing but most causing an enhancement in absorption. The method was made more selective by the addition of aluminium. Molybdenum has been determined in steel by using a nitrous oxide - acetylen ATOMIC ABSORPTION SPECTROSCOPY 219 flame and potassium sulphate as a buffer.267 It has also been determined by using an air - acetylene flame in fuels and lubricant^^^^^^^^ and in lake waters by using a nitrous oxide supported flame after complexation and extraction into an organic solvent .270 Butler and Mathew~~~l determined trace amounts of molybdenum in waters plant and silicate samples by using an air - acetylene flame after extraction into an organic solvent.Ure272 reported the determination of molybdenum in soil extracts by using conventional and separated nitrous oxide - acetylene flames. The absorption was enhanced by the presence of ammonia and therefore the molyb-denum was stripped from an organic complex by the addition of ammonia solution. With a 50-g soil sample a detection limit of better than 0.004 p.p.m.of molybdenum in the soil can be obtained. Kirkbright et recommend the use of a nitrous oxide - acetylene flame to achieve good sensitivity while avoiding the unpleasant, strongly luminous fuel-rich air - acetylene flame that can cause fatigue of the detection system. Niobium Manning255 reports a sensitivity of 20 p.p.m. and a detection limit of about 5 p.p.m. with potassium added at 1000 p.p.m. and a 10-fold scale expansion by using a nitrous oxide - acetylene flame. The niobium spectrum is very complex and Manning lists thirty-two ‘atom and ion lines,’ a number of which are un-resolved pairs. Noble metals Beamish et ~ 1 . ~ ’ ~ have reviewed the use of atomic absorption spectrochemical and X-ray fluorescence methods for the determination of the noble metals.The main difficulty in determining each of the noble metals is the extremely low concentration in which they are usually present in most materials.14 This necessitates the use of some concentration technique such as co-precipitation, cupellation ion-exchange or extraction into an organic solvent after complexation prior to the determination by flame methods. Gold palladium platinum rhodium ruthenium and iridium have all been determined in the air - acetylene flame by using a mufti-slot b ~ r n e r ~ 6 ~ ~ ~ 6 while reported determinations of osmium have nearly all used the nitrous oxide -acetylene Osolinski and Knight278 reported that the method for osmium was rapid precise and accurate. A sensitivity of 1 p.p.m.was attained and the method was applicable to both aqueous and non-aqueous solutions. The oxidation state of the osmium was not important. Fernandez2T7 found a 4-fold increase in sensitivity when using a nitrous oxide supported flame rather than air - acetylene for the determination of osmium, Atwell and Herbert279 compared the air - acetylene flame with the nitrous oxide - acetylene flame for the determination of rhodium and concluded that the serious interferences encountered with the former were essentially removed by using the hotter flame. Johns and Price2go recommended the nitrou 220 PUTT oxide - acetylene flame for the determination of rhodium to improve linearity of calibration and to overcome all interferences except the ionisation effect for which a compensating buffer is required.Aldous et aZ.281 described the preparation of electrodeless discharge lamps for palladium silver platinum and gold. Rare earths Amos and WillisQ8 pointed out that the atomic absorption technique with a high temperature flame should prove very valuable in the analysis of rare earth mixtures as no mutual interference was expected. The sensitivities for the heavy rare earths are considerably better than for light ones and it is necessary to include an ionisation buffer to obtain the best sensitivity. Jaworowski et aZ.282 list sensitivi-ties for rare earth elements determined in aqueous and organic solutions. These range from 1.5 p.p.m. in 80 per cent. of alcohol up to 100 p.p.m. in a purely aqueous medium. Kinnunen and Lindsjo283 found that the most serious interferences were caused by fluoride silica and aluminium the degree of interference depending on the element to be determined.Van Loon et aZ.28Q found aluminium to be the most serious interference and it was concluded that no releasing agent could completely eliminate its effect. It was however accounted for by adding 150 p.p.m. of alumi-nium to the standards and 1 per cent. of lanthanum eliminated all other commonly encountered interferences. The relative freedom of atomic absorption methods from matrix and spectral interferences makes the technique potentially useful in rare earth analysis although the poor sensitivity obtained for some elements precludes its use for a complete analysis in natural materials. Hingle et aZ2= found the detection limits for eight rare earths by using an emission technique and a separated nitrous oxide - acetylene flame to be equal or , superior to those obtainable by atomic absorption spectroscopy.Scandium Chau286 discussed the determination of trace amounts of scandium by atomic absorption in the nitrous oxide - acetylene flame. The flame gas mixture was found to be critical for optimum absorption. The presence of EDTA enhanced the absorption and a sensitivity of 0.06 p.p.m. was observed by using extraction into an oxine - butanol mixture. Rhenium Biechler and Long287 determined rhenium in a fuel-rich nitrous oxide -acetylene flame by using solvent extraction and obtained a sensitivity of 3.5 p.p.m. Molybdenum did not interfere at 7000 times the rhenium concentration.The burner height was adjusted for maximum response by using the highest standard, and it was ensured that the volumes of the aqueous and organic phases were the same for standards and samples ATOMIC ABSORPTION SPECTROSCOPY 221 Silicon Silicon is one of the less sensitive elements and optimum sensitivity requires careful adjustment of the instrument conditions. The fuel-rich nitrous oxide -acetylene flame used causes rapid carbon build-up around the burner slot unless one of the later re-designed burners is available. Price and Roos28s determined silicon in steel cast iron aluminium alloys and cement. They investigated the optimum operating conditions and discussed the effect of foreign ions and the preparation of standards and samples. The sensitivity was found to be 8 to 10 p.p.m.and the detection limit was 3 p.p.m. in aqueous solu-tion. McAuliffeBg reported a method for the determination of silicon in cast iron and steel. The results obtained were compared with certificate values and with those obtained by colorimetric analysis and showed the atomic absorption method to have the accuracy and repeatability necessary for control or referee analysis. Morrow and Dean290 used atomic absorption as a specific silicon detector for the gas-chromatographic determination of silylated aliphatic alcohols. Paralus~~~l described the determination of trace silicon in an organic matrix and Mario and GernerB2 determined silicon in a commercial hand lotion. Kirkbright et aZ.124 using a separated flame found that the high fuel flow required resulted in an instability at the ends of the burner slot when nitrogen shielding was used.The effect was not experienced with argon shielding and hence the latter gave a much more pronounced improvement in detection limit. The sensitivity and detection limit obtained were 2.5 and 0.24 p.p.m. respectively, and a calibration graph was linear from 5 to 200 p.p.m. Dagnall et aZ.293 compared the detection limits obtained for silicon by atomic absorption with those obtained by flame emission and atomic fluorescence. Absorption with a high intensity lamp or an electrodeless discharge lamp was 4 p.p.m. whereas emission and atomic fluorescence gave values of 20 and 5 p.p.m. respectively when using an argon-separated flame. Tin The most sensitive conditions for tin are obtained with an air - hydrogen flame in spite of its lower temperature but the cooler flame allows interferences to occur.The nitrous oxide-acetylene flame gives a greater freedom from any possible interference^.^^^^ Amos and Willisgs quote sensitivities of 2-5 p.p.m. in the hot flame compared with 1 p.p.m. in the cooler one. Interferences present in the determination of tin when using an air - hydrogen flame have been discussed by Capacho-Delgado and Manning294 and by Juliano and Harri~on.2~~ used the nitrous oxide - acetylene flame to eliminate the interferences during the determination of tin in tin ores and concentrates. He obtained a detection limit of 0.02 per cent. in the sample and the results were similar to those obtained by standard chemical methods.Shannon297 reported that the nitrous oxide - acetylene flame was unsatisfac-tory for the determination of tin in the presence of glycerol 222 PLAT" No satisfactory explanation appears to have been given for the high sensitivity obtained for tin when using the cooler air - hydrogen flame.l14 Titanium Amos and WillisS8 reported the determination of titanium in the nitrous oxide - acetylene flame with a sensitivity of 3.5 p.p.m. These workers also indicated the enhancement of titanium in the presence of hydrofluoric acid and hydrofluoric acid plus iron. They attribute this behaviour to the formation of a stable complex, fluoroacid which inhibits the formation of titanium oxide but is more readily decomposed at high temperatures to yield a higher population of titanium atoms.Nevertheless it was considered that further investigation was required before this element could be determined in a routine fashion. Headridge and H ~ b b a r d ~ ~ * determined titanium in steels permanent magnet alloys and cast iron by using a nitrous oxide - acetylene flame. Hydrofluoric acid was used as the solvent and the use of aqueous ethanolic solutions more than doubled the sensitivity. When the correct conditions were used there wasno interference from the other elements commonly found in these materials. The results obtained agreed well with certificate values for titanium contents of between 0.1 and 1.2 per cent. Kirkbright et aZ.lN used separated flames to determine tita-nium and found linear calibration curves from 10 to 250 p.p.m.The addition of potassium chloride gave an increase in absorbance at all concentrations of titanium and had no effect on the linearity of calibration. With potassium chloride present, the detection limit was 0.035 p.p.m. Mostyn and C~nningham~~~ avoided the use of hydrofluoric acid and used aqua regia as the solvent with 2000 pg ml-l of potassium chloride to obtain reliable analytical results that compared well with those obtained by chemical and S-ray fluorescence analysis on various alloys. Vanadium By using a high intensity lamp in conjunction with a nitrous oxide - acetylene flame Capacho-Delgado and Manning30° determined vanadium in steel and gas oils. An apparent compensating effect of sulphuric and phosphoric acids enabled results in good agreement with certificate values to be obtained in the analysis of steels.Gas oils were analysed by the method of standard additions after dilution in xylene and a detection limit of 0.05 pg ml-l was obtained. Sachdev et aLW1 found an enhancing effect by many potential interferents which may have been caused by competition in the formation of oxides. The same workers302 studied the r61e of mixed organic solvents and found that the addition of diethylene glycol and similar compounds increased the absorption by about 50 per cent. Hall et aLm3 reported that suitable standards were prepared for the determination of vanadium in steels simply by adding 1 per cent. of trivalent iron. Good agreement was obtained with certificate results. Kirkbright et aZ.124 obtained a detection limit of 0.04 p.p.m.in the presence of potassium chloride when using separated flames. Calibration curves were linea ATOMIC ABSORPTION SPECTROSCOPY 223 between 5 and 100 p.p.m. but curved towards the concentration axis between 100 and 200 p.p.m. Zirconium Amos and WillisB8 found a similar effect with zirconium in the presence of hydrofluoric acid and iron as for titanium. A detection limit of 5 pg ml-l has been reported.ll Slavinll found an enhancement with hydrochloric as well as with hydrofluoric acid. TyleSM found that potassium added as the sulphate entirely suppressed the absorption of zirconium instead of producing the expected enhance-ment. Tyler attributed this to the formation of zirconium sulphate and Slavinll found the expected enhancement when potassium was added as the chloride.Kirkbright et aZ.12* reported a detection limit of 0.24 p.p.m. in the argon-shielded flame and in the presence of potassium chloride. Calibration curves were linear between 500 and 1000 p.p.m. Elements with analytical resonance lines between 190 and 230 nm The air - acetylene flame strongly absorbs any radiation in the wavelength region between 190 and 230 nm in which arsenic antimony selenium tellurium, cadmium zinc and lead have their resonance lines. Strong light absorption by the flame leads to background instability and a reduction in the available energy reaching the detector. For these reasons and also because of difficulties in producing satisfactory sources little detailed information was available on the determination of elements such as arsenic and selenium until recently.However a number of ways of tackling the problems involved have now been described and suitable instrumentation has been made commercially available. The primary source intensity can be increased by the use of high spectral output hollow-cathode lamps or by using electrodeless discharge lamps. This allows the use of an air - acetylene flame to keep chemical interferences minimal, while still leaving ample energy to reach the detector. Alternatively the flame type may be changed to make it more transparent and hence absorb considerably less energy. The latter may be accomplished by using a cooler flame such as air-hydrogen or argon - hydrogen in which chemical interferences will be greater or by retaining the air - acetylene flame and separating it by nitrogen shielding.Dagnall et uZ.= have described the use of electrodeless lamps as sources for the determination of antimony arsenic and selenium. The intensity of the resonance lines was sufficient to allow the use of narrow slits and the stability was such that a 10-fold scale expansion could be used. Detection limits with different flame systems varied between 0.1 and 0.5 pg ml-l. Air - acetylene and the cooler more transparent nitrogen - hydrogen flames were used on single and triple-slot burners. For arsenic the latter flame absorbed 50 per cent. of the radiation on a single-slot burner but only 5 per cent. on a triple-slot one. This was attributed to the wider flame allowing all of the light from the lamp to pass through the centre of the flame.The cooler flame may give rise to chemical interference in whic 2 24 PLATT case it will be necessary to resort to a suitable separation technique such as solvent extraction or ion exchange. Menis and Rains306 determined arsenic in cast iron and in high purity selenium metal by using an electrodeless discharge lamp after solvent extraction followed by re-extraction into an aqueous solvent. Fisher and H a y ~ a r d ~ ~ studied the determination of arsenic and selenium by using electrode-less discharge lamps and found better detection limits and stability compared with the use of hollow-cathode lamps. Kahn and Schallis found considerable improvements in both sensitivity and detection limits for arsenic selenium and cadmium when using an argon - hydrogen flame and moderate improvements for zinc and lead,lf4 Pre-mixed air - hydrogen flames were also shown to produce better detection limits than air - acetylene for elements tellurium cadmium zinc and lead.ll8 Like Dagnall et ~ 1 .~ ~ these workers used a three-slot burner but in contrast they found little success when using nitrogen - hydrogen flames. The detection limit for arsenic was less than 0.05 pg ml-1 and for selenium less than 0.1 pg ml-l compared with 0.25 and 0.5 pg ml-1 respectively in the air - acetylene flame. Because of the background absorption from the air - acetylene flame at 217 nm the less sensitive 283.3-nm line has been preferred for the determination of lead and this gives a detection limit of about 0.03 pg ml-l.However with the air - acetylene flame the most sensitive lead line at 217 nm can be successfully used. As Kahn and Schallis point out the cooler flames produce at best improve-ment factors of only 2 or 3 and operators who are using air - acetylene are well advised to add hydrogen only if the degree of improvement is really necessary. They found no advantage in the air - hydrogen flame for the determination of antimony at the 217.6-nm line. This element is in fact determined easily in the air - acetylene flame.ll Cookella reported that a good hollow-cathode lamp was now available for arsenic. The lamp is of the high spectral output type which ensures that the intensity is sufficiently high to give a good signal-to-noise ratio in the final measure-ment.Cooke investigated air - acetylene and argon - hydrogen flames on a multi-slot burner and recommended the air - acetylene system to overcome interferences. He considered it desirable to use an instrument incorporating a silica prism mono-chromator as transmission and stray light characteristics are superior to diffrac-tion grating monochromator characteristics. In addition the dispersion of a prism is high at these short wavelengths thus allowing the use of relatively wide slit-widths while retaining adequate resolution. Kirkbright et uZ.123 determined arsenic and selenium in a nitrogen separated air - acetylene flame. This system offers good sensitivity and precision combined with relative freedom from chemical interferences. Kahn et ~ 1 . ~ 0 have reported greatly improved detection limits for arsenic selenium cadmium lead and zinc by using the sampling boat technique.Hill25 determined arsenic in steels iron ores and spelters in an argon - hydrogen flame following a rapid distillation procedure. Holak306 converted arsenic to arsine which was collected in a U-tube immersed in liquid nitrogen. The arsine was swep ATOMIC ABSORPTION SPECTROSCOPY 225 into the cloud chamber with nitrogen gas and determined in an air - acetylene flame on a three-slot burner. Ando et aL307 greatly increased the sensitivity for arsenic by using a nitrogen - hydrogen flame in conjunction with a Vycar long tube. A sensitivity of 0.006 p.p.m. was obtained. Chakrabarti308 found the air - acetylene flame to be more sensitive than air -hydrogen for the determination of selenium.A sensitivity limit of 0.72 p.p.m. in aqueous solutions was improved to 0.30 p.p.m. when extracted into methyl iso-butyl ketone as the diethyl-dithiocarbamate complex. The same author also in-vestigated the determination of tell~rium.30~ Barnett and Kahn310 described the determination of tellurium in steel by using the ‘Deuterium Background Corrector’ to compensate for the broad-band absorbance caused by the matrix material. Air -acetylene air - hydrogen and argon - hydrogen flames were examined and the first one selected because it showed good sensitivity and was likely to exhibit the fewest interferences. Taylor119 used the air - hydrogen flame to determine lead on a three-slot burner head. The determination of lead in methyl iso-butyl ketone was three times more sensitive than when using water as the solvent and the increased sensitivity at the expense of signal-to-noise ratio is illustrated for the 217-nm line.Slavin and SattuPl have noted a spectral interference of lead on antimony when using the primary antimony resonance line at 217.6 nm and a spectral slit-width of 0.7 nm. It is suggested that the resonance lines at 206.8 or 231.1 nm are used to avoid the interference. Determinations by an indirect method A number of indirect methods have been proposed whereby the component to be determined reacts quantitively with a metal that can subsequently be deter-mined by atomic absorption spectroscopy. However as Koirtyohann points these indirect methods must be used very cautiously because the specificity of the determination depends on the chemistry of the reactions and not on the final determinations.The methods can be quite valuable in some cases. Cations anions and organic materials have been determined indirectly most often following quantitative precipitation or complexation - extraction but some-times by utilising a quantitative interference effect. A brief selected summary is given in Table I. Before leaving the indirect methods a few points are worthy of further mention. Thorium cannot be determined directly by atomic absorption and so the indirect procedure described by Kirkbright et U Z . ~ ~ although complicated by many potential interferences does at least provide a means of determining thorium by this technique.Dunk et aL319 determined sulphate in textiles and found the atomic absorption results to compare reasonably well with volumetric and gravi-metric methods and to require less operator time. The determination of sugar in plant materials reported by Potter et aZ.,= gave results in close agreement with the existing Association of Official Agricultural Chemists method and had the advantage of simplicity. Ezel1325 found the atomic absorption determination o 226 PLATT chloride in plant liquors to be capable of high precision and to provide more accurate determinations than titrimetric methods. TABLE I Component required Metal used References Phosphate Phosphorus arsenic silicon Sulphur dioxide Sulphur dioxide Sulphate Nitrate Thioc yanate Fluoride Chlorine and chloride Ammonia Ammonia Thallium Thorium Phthalic acid Pentachlorophenol Sugar Non-ionic surfactants Molybdenum 312 Lead 317 Mercury 318 Barium 319 320 Copper 321 Copper 322 Magnesium zirconium or titanium 323 Silver 324 325 326 327 Molybdenum 328 Zirconium 329 Molybdenum 328 Molybdenum 330 Copper 331 Iron 332 Molybdenum 3 13-31 6 Copper 333 Molybdenum 334 Determination of mercury K;oirtyohann@ indicates that the use of static atomic vapours rather than flames allows us to expect big improvements in sensitivity and the available information for mercury permits an estimation of the general improvement to be expected.According to Koirtyohann about 5 pg of mercury are required to pro-duce 1 per cent. of absorption in the air - acetylene flame.In a cell of cross-section 0.25 cm2 the same absorption is produced by 0.000 05 pg of mercury in the static vapour. Brandenberger and Bade? described a dynamic vapour determination of mercury by using a commercial atomic absorption instrument. The mercury was amalgamated from solution on to a copper wire which was then electrically heated to vaporise the mercury in an absorption cell placed in the light beam of the spectrophotometer. The mercury vapour absorbed radiation while being pumped through the cell and the resulting absorption was plotted as a function of time by a recorder. The same workers later reported a static vapour approachs6 that was less complicated than the dynamic method and produced equivalent accuracy, sensitivity and detection limit.Between 10 and 200 ng of mercury was determined with an accuracy of &3 per cent. and an accuracy of &lo per cent. was obtained at 0.2 ng of mercury. ThillieF reported a similar procedure for the determination of mercury pollution in atmospheric air ATOMIC ABSORPTION SPECTROSCOPY 227 Determination of isotopic concentrations Atomic absorption has been used to determine isotopic concentrations of lithium lead and uranium. An attempt to determine boron isotopes was not successful .= K i r ~ h h o f ~ ~ measured the concentrations of lead-206 and lead-208 by using the absorption of the hyperfine components of the 405.8 and 283.3 nm lines. Known isotopic mixtures were incorporated into hollow cathodes which were used as primary radiation sources.The concentration of the individual isotopes was determined by measuring the absorption of isotopically pure vapours. BrimhaPo measured the concentrations of lead-206 lead-207 and lead-208 in solutions by using the hyperfine components of the 283.3-nm line. Three hollow cathodes each enriched in one of the isotopes were used and standard solutions of known isotopic composition were prepared. By using a standard atomic absorption system to measure the absorptions isotopic concentrations of an unknown were calculated by solving a set of three simultaneous equations. GolebS1 was able to determine uranium isotope ratios by using a hollow-cathode lamp as the absorption tube. Butler and SchroedersP2 reported the deter-mination of lithium isotope ratios by using a special instrument to enable the rapid and precise measurements of absorbance ratios.These atomic absorption procedures produced results in good agreement with mass spectrometric data. I thank the representatives of the various instrument manufacturers for their willingness to discuss various aspects and in particular Mr. W. J. Price for his valuable comments. I also thank my colleagues at Colgate-Palmolive Ltd. for their valuable assistance without which this work could not have been carried out. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 References Slavin W. Appl. Spectrosc. 1966 20 281. Slavin W. and Slavin S. Ibid. 1969 23 421. West T. S. Endeavour 1967,26 44. Reynolds R. J. Lab. Equip. Dig. 1968 6 79. Walsh A.Spectrochim. Acta 1955 7,‘108. Elwell W. T. and Gidley J. A. F. Atomic Absorption Spectrophotometry.’ Second Edition Pergamon Press 1966. Fassel V. A. and Becker D. A. Analyt. Chem. 1969,41 1522. Fassel V. A. Plenary Lecture International Atomic Absorption Spectroscopy Con-ference Sheffield July 1969. Winefordner J. D. y d Vickers T. J. Analyt. Chem. j970,42 206R. Ramfiez-Muiioz J . Atomic Absorption Spectroscopy Elsevier Publishing Company, 1968. Slavin W. ‘Atomic Absorption Spectroscopy,’ Interscience Publishers 1968. Cooke P. A. ‘Atomic Absorption Spectrophotometry,’ Second Edition Pye-Unicam Ltd., Price W. J. in Browning D. R. Editoor ‘Spectroscopy,’ McGraw-Hill 1969. RubeSka I. and Moldan B. ‘Atomic Absorption Spectrophotometry,’ Iliffe 1969 p. 154. Reynolds R.J. Aldous K. and Thompson K. C. ‘Atomic Absorption Spectroscopy-Pye-Unicam Ltd. Cambridge England. 1969. a Practical Guide.’ Charles Griffin 1970 228 PLATT 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 Perlun-Elmer Corporation Instrum News 19f9.20 15 Masek P R and Sutherland I Edztors Atomic Absorption and Flame Emission Stupar J and Dawson J B Appl Optzcs 1968,7 1351 Willis J B Spectrochzm Acta 1967 23A 811 Powell R J W ‘International Atomic Absorption Spectroscopy Conference,’ Sheffield, 1969 Ramirez-Mufioz J and Ulrich W F Flame Notes 1966 I(2) 33 Marshall D and Schrenk W G Spectroscopy Lett 1968 1 87 Ramfrez-Muiioz J Shifrin N and Hell A Mzcrochem J 1966 11 204 ‘Atomic Absorption Spectroscopy,’ Speczal Technzcal Publzcatzon No 443 American Rendma J Analyt Advances 1968 1 (2) Kxkbright G F Semb A and West T S Spectroscopy Lett 1968 1 (2) 97.Kahn H L J Chem Educ 1966 43A 7 Reynolds R J Metallurgza 1970 81 197 St John P A McCarthy W J and Winefordner J D Analyt Chem 1967,39 1495 Reynolds R J Lab Equzp Dzg 1969 7 75 Fisher A A and Hayward G C 21st Pittsburgh Conference on Analytical Chemistry Dagnall R M Thompson K C and West T S Atomzc Absorptzon Newslett 1967, Price W J Spectrovzszon 1966 No 15 2 Ramfrez-Muiioz J Amer Lab 1969 (August) Eel Bulletin No 7 1969 Muller R H Analyt Chem 1968,40 85A Eardley R P and Mountford A H Speczal Technzcal Publzcatzon No 59 British Dawson J B and Broughton P M G I Technzcal Bulletzn No 14 1968 Association Slavin W and Ringhardtz I Kern Teollzsuus 1966 11 985 Galassi M and Hell A Flame Notes 1966 1 (2) 28 Kahn H L J Chem Educ 1966,43 (2) A103 Sullivan J V and Walsh A Spectrochzm Acta 1965 21 721 Willis J B Rev Pure Appl Chem 1967 17 111 Cartwnght J S and Mannmg D C Atomzc Absorptzon Newslett 1966 5 114 Ure A M and Wtchell R I Spectrochzm Acta 1967 23B 79 Cartwright J S Sebens C and Slavm W Atomzc Absorptzon Newslett 1966 5 22 Price W J Spectrovzszon 1966 No 15 2 Manning R J Flame Notes 1968 3 ( l ) 11 Dawson J B and Ellis D J Spectrochzm Acta 1967,23A 565 Hussein A M I and Nickless G I InternationaI Atomic Absorption Spectroscopy Con-Cooke P A Process Bzochem 1967 2 (6) 27 Bowman J A Sullivan J V and Walsh A Spectrochzm Acta 1966 22 205 Kontyohann S R Atomzc Absorptzon NewsZett 1967 6 77 Lowe R M Spectrochzm Acta 1969 24B 191 Van Rensburg H C and Zeeman P B Analytzca Chzm Acta 1968,43 173 Butler L R P and Brink J A International Atomic Absorption Spectroscopy Mitchell D C and Johansson A Spectrochzm Acta 1970 25B 175 Fernandez F J Mamng D C and Vollmer J Atomzc Absorptzon Newslett 1969, Jaworowsh R J and Weberling R P Ibzd 1966 5 125 Frank C W Schrenk W G and Meloan C E Analyt Chem 1966 38 1005 Heneage P Atomzc Absorptzon Newslett 1966 5 67 Vollmer J Ibzd 1966 5 35 Schrenk W G and Meloan C E Spectrosc Lett 1968 1 237 Popham R E and Schrenk W G AppE Spectrosc 1968,22 192 Woodriff R A Wheeler G V and Ryder W A Ibzd 1968 22 348 Spectroscopy Abstracts,’ Science and Technology Agency Society for Testmg Materials Philadelphia 1968 and Applied Spectroscopy March 1970 6 117.Ceramic Research Association 1968 of Clinical Biochemists Scientific and Techmcal Communications ference Sheffield July 1969 Conference Sheffield July 1969 8 11 ATOMIC ABSORPTION SPECTROSCOPY 229 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 Fassel V. A. Mossotti V. G. Grossman W. E. L. and Kniseley R. N. Spectrochim. DeGalan L.McGee W. W. and Winefordner J. D. Analytica Chim. Acta 1967,37,436. McGee W. W. and Winefordner J. D. Ibid. 1967 37 429. Dagnall R. M. and West T. S. Appl. Optics 1968 7 1287. Dagnall R. M. Pribil R. and West T. S. Analyst 1968 93 281. Mansfield J. M. Bratzel M. P. Norgardon H. O. Knapp D. O. Zacha K. E. and Zacha K. E. Bratzel M. P. Winefordner J. D. and Mansfield J. M. Analyt. Chem., Headridge J. B. and Richardson J. Lab. Pract. 1970 19 372. Woodward C. Atomic Absorption Newslett. 1969 8 121. Browner R. F. Dagnall R. M. and West T. S. Analytica Chim. Acta 1969 45 163. Hoare H. C. Mostyn R. A. and Newland B. T. N. Ibid. 1968,40 181. West C. D. Analyt. Chem. 1968 40 253. Stupar J. and Dawson J. B. Atomic Absorption Newslett. 1969 8 38. Kirsten W.J. and Bertilsson G. 0. B. Analyt. Chem. 1966,38 648. Rawson R. A. G. Analyst 1966 91 630. Riley J. P. and Taylor D. Analytica Chim. Acta 1968 42 548. Veillon C. and Margoshes M. Spectrochim. Acta 1968 23B 553. Venghiattis A. A. AppL Optics 1968 7 1313. Hell A. Ulrich W. F. Shifrin N. and Ramirez-Muiloz J. Ibid. 1968,7 1317. Beckman Atomic Absoption Systems Bulletin 7134 p. 13. Gray J. M. and Gallwas G. E. Flame Notes 1967 2 2. Venghiattis A. A. Spectrochim. Acta 1967 23B 67. Instrum. News 1967 17 1. Kahn H. L. Peterson G. E. and Schallis J. E. Atomic Absorption Newslett. 1968,7 35. Curry A. S. Read J. F. and Knott A. R. Analyst 1969 94 744. White R. A, International Atomic Absorption Spectroscopy Conference Sheffield, Delves H. T. Analyst 1970 95 431 Willis J.B. Appl. Optics 1968 7 1295. Cowley T. G. Fassel V. A. and Kniseley R. N. Spectrochim. Acta 1968 23B 771. Mossotti V. H. ,and Duggan M. Appl. Ofltics 1968 7 1325. Price W. J. Efl. Wat. Treat. J. 1967,7 218. Amos M. D. and Willis J. B. Spectrochim. Acta 1966 22 1325. Willis J. B. Nature 1965 207 715. Amos M. D. Atomic Absorption Spectroscopy Symposium Cardiff April 1967. Willis J. B. Rasmuson J. O. Kniseley R. N. and Fassel V. A. Spectrochim. Acta, deGalan L. and Samaey G. F. International Atomic Absorption Spectroscopy Con-Kirkbright G. F. Peters M. K. and West T. S. Talanta 1967 14 789. Price W. J. Process Biochem. 1970 5 35. Pampel C. Atomic Absorption Newslett. 1967 6 133. Slavin W. Venghiattis A. and Manning D. C. Ibid. 1966 5 84. Manning D. C. Ibid.1966 5 127. Walsh A. J. N.Z. Inst. Chem. 1966 30 7. Bowman J. A. and Willis J. B. Analyt. Chem. 1967 39 1210. Shifrin N. Hell A. and Ramirez-Muiioz J. Appl. Spectrosc. 1969 23 365. Julietti R. J. and Wilkinson J. A. E. Ibid. 1969 23 173. Butler L. R. P. and Fulton A. Appl. Optics 1968 7 (lo) 2131. Fleming H. D. Sflectrochim. Acta 1967 23B 207. Kahn H. L. and Schallis J. E. Atomic Absorption Newslett. 1968 7 5. Dagnall R. M. Thompson K. C. and West T. S. Talanta 1968 15 677. Cooke P. A, International Atomic Absorption Spectroscopy Conference Sheffield, Ramirez-Mufioz J. Brace R. O. Claus R. W. and Shifrin N. Flame Notes 1967,2 30. Schallis J. E. and Kahn H. L. Atomic Absorption Newslett. 1968 7 75. Taylor J. H. Ibid. 1969 8 95. Acta 1966’22 347. Winefordner J.D. Spectrochim. Acta 1968 23B 389. 1968,40 1733. July 1969. 1968,22B 725. ference Sheffield July 1969. July 1969 PLATT Dagall R. M. Thompson K. C. and West T. S. Analyst 1968 93 153. Willis J. B. Fassel V. A. and Fiorino J. A. Spectrochzm. Acta 1969 24B 157. Bolmg E. A. Ibzd. 1966 22 425. Krrkbnght G. F. Sargent M. and West T. S. Atomzc Absorptzon Newslett. 1969 8 34. Kirkbright G. F. International Atomic Absorption Spectroscopy Conference Sheffield, RubeSka I. and Moldan B. Analyst 1968 93 148. - and - Appl. Optzcs 1968 7 (7) 1341. West T. S and Wdliams X. K Analytzca Chzm. Acta 1969 45 27, L’vov B. V. Spectrochzm. Acta 1969 24B 53. Massmann H. Ibzd. 1968,23B 215. Anderson R. G. Maines I. S. and West T. S International Atomic Absorption Spectro-Tomkms R.S. and Ercoli B. Appl. Optzcs 1967 6 (8) 1299. Woodriff R. and Stone R. W. Ibzd. 1968 7 (7) 1337. Woodriff R. Stone R. W. and Held A. M. Appl. Spectroscopy 1968,22 408. Woodriff R. and Ramelow G. Spectrochzm. Acta 1968,23B 665. Mornson G. H. and Talmi Y. Analyt. Chem 1970,42 809. Mossotti V. G. Laqua K. and Hagenah W. D. Spectrochzm. Acta 1967,2333 197. Karyakin A. V. and Kagarodov V. A. Zh. Analzt. Khzm 1968 23 930. Wendt R. H. and Fassel V. A. Analyt. Chem. 1966 38 337. Hoare H. C. and Mostyn R. A. Ibzd. 1967 39 1153. Vedlon C. and Margoshes M. Spectrochzm. Acta 1968,23B 503. Aldous K. M Dagnall R. M. Thompson K. C. and West T. S. Analytzca Chzm. Ada, Greenfield S. Smith P. B Breeze A. E. and Chdton N. M D. Ibzd 1968,41 385. Sullivan J. V.and Walsh A. Spectrochzm. Acta 1966,22 1843. Brandenberger H. Chzmza 1968,22 449. Sullivan J. V. and Walsh A. Appl. Optzcs 1968 7(7) 1271. Rawling B. S and Sullivan J. V. Mzneral Process. Extract. Metall. 1967 76 (c) 238. Boar P. L. and Sullivan J. V. Fuel 1967 46 47. Bowman J. A. Analytzca Chzm. Acta 1967 37 465. Johnson J. D. Yamasaki G. K and Burger J. C zn Grove E L and Perhns A J., Edztors Developments in Applied Spectroscopy Symposium,’ Vol IIIA Plenum Press 1969. Muller R. H. Analyt. Chem. 1967 39 99A. Kahn H. L. International Atomic Absorption Spectroscopy Conference Sheffield, Harrison W. W. and Berry I;. E. Analytzca Chzm Acta 1969 47 415. Evans Electroselenium Ltd. ‘Model 240 Brochure.’ Ramirez-Mufioz J Malakoff J. L and Aime C P. Analytzca Chzm.Acta 1966’36 328. Wendt R. H. Atomzc Absorptzon Newslett. 1968 7 28. Malakoff J. L Ramfrez-Muiioz J and Aime C. P Analytzca Chzm. Acta 1968,43 37. Malakoff J. L. Ramfrez-Muiioz J and Scott W. Z. Ibzd. 1968,42 515. Instrum News 1969,20 15. Slavin W. International Atomic Absorption Spectroscopy Conference Sheffield, Instrum. News 1968 19 6. Dawson J. B. Ellis D. J and Mdner R. Spectrochzm. Acta 1968,23B 695. Lacy J. Analyst 1965 90 65. Klein B Kaufman J. H andMorganstern S. Clzn. Chem 1967.13 388. Crawford L. R jun and Greweling T Appl Spectroscopy 1968,22 793. Ellis D J. and Rogers J Speclrovzszon 1967 17 2. Lewis R. R Atomzc Absorptzon Newslett. 1969 8 94. Kokot M. L and Butler L. R. P. Ibzd. 1969 8 92. Watson C. A. Monograph 74 Hopkin & Williams Ltd.Roth M. E. Flame Notes 1969 4 4. Ramfrez-Mufioz J. J. Forens. Scz. Soc. 1967 7 151. - Flame Notes 1967 2 54. - and - Talanta 1969 16 1467. July 1969. scopy Conference Sheffield July 1969. 1968 41 380. July 1969. July 1969. 230 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 17 ATOMIC ABSORPTION SPECTROSCOPY 231 173 Ramirez-Muiioz Micvochem. J. 1967 12 196. 174 - J. Talanta 1969 16 1037. 175 Grunder F. I. and Boettner E. A. in Grove E. L. and Perkins A. J. op. cit. p. 201. 176 Lewis L. L.Analyt. Chem. 1968,40 28A. 177 Cooke P. A. and Price W. J. Spectrovision 1966 16 7. 178 Fassel V. A. Rasmuson J. O. and Cowley T. G. Spectrochim. Acta 1968,23B 579. 179 Koirtyohann S. R. and Pickett E. E. Analyt. Chem. 1966,38 585. 180 Kahn H. L. Atomic Absorption Newslett. 1968 7 40. 181 Perry C. J. and Keyworth D. A. Can. Spectroscopy. 1967 12 47. 182 Kerber J. D. Atomic Absorption Newslett. 1967 6 131. 183 Parsons M. L. and Winefordner J. D. Appl. Spectrosc. 1967 21 368. 184 Rohleder H. A. International Atomic Absorption Spectroscopy Conference Sheffield, 185 Cellier K. M. and Stace H. C. T. Appl. Spectrosc. 1966 20 26. 186 Manning D. C. and Capacho-Delgado L. Analytica Chim. Acta 1966 36 312. 187 Brace R. O. Flame Notes 1966 1 2. 188 Ramirez-Muiioz J. in Grove E.L. and Perkins A. J. op. cit. p. 169. 189 Reynolds R. J. Wld Med. Instrumn 1969 7 10. 190 RubeSka I. and Moldan B. Analytica Chim. Acta 1967 37 421. 191 Halls D. J. and Townshend A. Ibid. 1966 36 278. 192 Popham R. E. and Schrenk W. G. in Grove E. L. and Perkins A. J. op. cit. p. 189. 193 Koirtyohann S. R. and Pickett E. E. Analyt. Chem. 1968,40 2068. 194 Sastri V. S. Chakrabarti C. L. and Willis D. E. Can. J. Chem. 1969,47 587. 196 - - and - International Atomic Absorption Spectroscopy Conference, 197 Hartlage F. R. jun. Analytica Chim. Ada 1967 39 273. 198 Ramirez-Muiioz J. Analyt. Chem. 1970,42 517. 199 Munro D. C. Appl. Spectrosc. 1968 22 199. 200 Mitchell D. G. Lab. Pract. 1967 16 587. 201 Kohlenberger D. W. Atomic Absorption Newslett. 1969 8 108.202 Guillaumin R. Revue Fr. Cps Gas 1969 16 497. 203 PrbvBt A. Atomic Absorption Newslett. 1966 5 13. 204 Piccolo B. and O’Connor R. T. J Amer. Oil Chem. SOC. 1968 45 789. 205 Montford B. and Cribbs S. C. Tuluntu 1969 16 1079. 206 Orren M. J. S.A. Chem. Processing 1968,3 36. 207 Peterson G. Atomic Absorption Newslett. 1966 5 117. 208 Margoshes M. Anulyt. Chem. 1967,39 1093. 209 Ramirez-MuEioz J. and Roth M. E. Flame Notes 1969 4 28. 210 - and - Ibid. 1969 4 62. 211 Takeuchi T. Suzuki M. and Yanagisawa M. Anulytica Chim. Acta 1966,36 258. 212 Mulford C. E. Atomic Absorfition Newslett. 1966 5 88. 213 Brooks R. R. Presley B. J. and Kaplan R. Talantu. 1967 14 809. 214 Montford B. Can. Spectrosc. 1968 13 126. 216 Husler J. W. and Cruft E. F. Analyt. Chem 1969,41 1688.216 Campbell D. J. Atomic Absorfition Newslett. 1967 6 49. 217 Chau Y. K. and Lum-Shue-Chan K. Anulytica Chim. Acta 1969,48 434. 218 Shepherd G. A. and Johnson A. J. Atomic Absorption Newslett. 1966,s; 142. 219 Manning D. C. Ibid. 1968 7 44. 220 Manning D. C. and Chabot H. Ibid. 1968 7 94. 221 Scheub W. H. and Stromsky C. J. Ibid. 1967,6 95. 222 Ramirez-Mufioz J. Flame Notes 1969 4 12. 223 Rooney R. C. Electron. Equip. News May 1967. 224 Kahn H. L. Atomic Absorption Newslett. 1967 6 61. 226 Feldman F. J. Blasi J. A. and Smith S. B. jun. Analyt. Chem. 1969,41 1095. 226 Sattur T. W. Atomic Absorption Newslett. 1966 5 37. 227 Meddings B. and Kaiser H. Ibid. 1967,6 28. 228 Welcher G. G. and Kriege H. Ibid. 1969 8 97. 229 Johns P. and Price W. J. Metallurgia 1970 81 75.July 1969. 195 - - and - Talanta 1969 16 1093. Sheffield July 1969 232 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 25 1 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 PLATT Welz B. Joint Symposium on Accurate Methods of Analysis for Major Constituents; S.A.C. & Koninklijke Nederlandse Chemishe Vereniging (Analyt. Sectn) London, April 1970. Cobb W. D. and Harrison T. S. Details as Ref. 230. Langmyhr F. J. and Paus P. E. Atomic Absor$tion Newslett. 1968 7 103. - and - Ibid. 1969 8 131. Bernas B. Analyt. Chem. 1968 40 1682. Van Loon J.C. and Parissis C. M. Analyst 1969 94 1057. Omang S. H. Analytica Chim. Acta 1969 46 225. Muter R. B. and Cockrell C. F. Ap$L Spectrosc. 1969 23 493. Medlin J. H. Suhr N. H. and Bodkin J. B. Atomic Absorption Newslett. 1969 8 25. Yule J. W. and Swanson G. A. Ibid. 1969 8 30. Boar P. L. and Ingram L. K. AnaZyst 1970 95 124. Campbell D. E. Analytica Chim. Acta 1969 46 31. Karmie Galle O. Appl. S$ectrosc. 1968 22 404. Crow R. F. Hime W. G. and Connolly J. D. J . Portland Cem. Ass. Res. Dev. Labs., Capacho-Delgado L. and Manning D. C. Analyst 1967,92 553. Roos J. H. T. and Price W. J. Analyst 1969 94 89. Slavin W. Venghiattis A. and Manning D. C. Atomic Absorption Newslett. 1966 5 84. Kerber J. D. and Barnett W. B. Ibid. 1969 8 113. Fulton A. and Butler L.R. P. S9ectrosc. Lett. 1968 1 317. Ramakrishna T. V. West P. W. and Robinson J. W. Analytica Chim. Acta 1968, Newsbitt R. W. Jbid. 1966 35 413. Ramakrishna T. V. West P. W. and Robinson J. W. Ibid. 1967 39 81. Pawluk S. Atomic Absorption Newslett. 1967 6 53. Laflamme Y. Ibid. 1967 6 70. Van Loon J. C. Ibid. 1968 7 3. Manning D. C. fbid. 1967 6 35. Bokowski D. L. Ibid. 1967 6 97. Peterson E. A. Ibid. 1969 8 53. Bade1 H. and Brandenberger H. Ibid. 1968 7 1. Harris R. Ibid. 1969 8 42. Roos J. T. H. International Atomic Absorption Spectroscopy Conference Sheffield, Rooney R. C. and Pratt C. G. International Atomic Absorption Spectroscopy Con-Ramirez-Mufioz J. and Roth M. E. Flame Notes 1968 3 2. Wilson L. Analytica Chim. Acta 1968 40 503. Popham R.E. and Schrenk W. G. Spectrochim Acta 1968,23B 543. Kirkbright G. F. Smith A. M. and West T. S. Analyst 1966 91 700. Ramakrishna T. V. West P. W. and Robinson J. W. Analytica Chim. Act@ 1969,41, Endo Y. Hata T. and Nakahara Y. JaFan Analyst 1969 18 878. Juliettu R. J. and Wilkinson J. A. E. Analyst 1968 93 797. Mostyn R. A. and Cunningham A. F. J . Inst. Pet. 1967 53 101. Chan Y . K. and Lum-Shue-Chau K. Analytica Chim. Acta 1969 48 205. Butler L. R. P. and Mathews P. M. Ibid. 1966 36 319. Ure A. M. International Atomic Absorption Spectroscopy Conference Sheffield July Kirkbright G. I?. Peters M. K. and West T. S. Analyst 1966 91 705. Reamish F. E. Lewis C. L. and Van Loon J. C. TaZanta 1969 16 1. Van Loon J. C. 2. Analyt. Chem. 1969,246 122. Fuhrman D. L. Atomic Absorption Newslett.1969 8 105. Fernandez F. Ibid. 1969 8 90. Osolinski T. W. and Knight N. H. A$pZ. Spectrosc. 1968 22 532. Atwell M. G. and Herbert J. Y. Ibid. 1969 23 480. Johns P. and Price W. J. Pittsburgh Conference on Analytical and Applied Spectro-1967 (May) 60. 40 350. July 1969. ference Sheffield July 1969. 437. 1969. scopy March 1970 ATOMIC ABSORPTION SPECTROSCOPY 233 281 282 283 284 285 286 287 288 289 290 29 1 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 32 1 322 32 3 324 325 326 327 328 329 330 33 1 332 Aldous K. M. Dagnall R. M. and West T. S. Analyst 1969 94 347.Jaworowski R. J. Weberling R. P. and Bracco D. J. Analytica Chim. Acta 1967 37, Kinnunen J. and Lindsjo O. Chemist Analyst 1967 56 25. Van Loon J. C. Aarden D. and Galbraith J. International Atomic Absorption Hingle D. N. Kirkbright G. F. and West T. S. Analyst 1969,94 864. Chau Y. K. Talanta 1968 15 421. Biechler D. G. and Long C. H. Atomic Absorption Newslett. 1969 8 56. Price W. J. and Roos J. T. H. Analyst 1968 93 709. McAuliffe J. J. Atomic Absorption Newslett. 1967 6 69. Morrow R. W. and Dean J. A. J . Chromatog. Sci. 1969 7 572. Paralusz C. M. Appl. Spectrosc. 1968 22 520. Mario E. and Gerner R. E. J . Pharm. Sci. 1968 57 1243. Dagnall R. M. Kirkbright G. F. West T. S. and Wood R. Analytica Chim. Acta 1969, Capacho-Delgado L. and Manning D. C. Spectrockim.A d a 1966,22 1505. Juliano P. O. and Harrison W. W. Analyt. Chem. 1970 42 84. Bowman J. A. Analytica Chim. Acta 1968,42 285. Shannon I. L. Caries Res. 1969 3 339. Headridge J. B. and Hubbard D. P. Analytica Chim. Acta 1967 37 151. Mostyn R. A. and Cunningham A. F. Atomic Absorption Newslett. 1967 6 86. Capacho-Delgado L. and Manning D. C. Ibid. 1966 5 1. Sachdev S. L. Robinson J. W. and West T. W. Analytica Chim. Acta 1967 37 12. and - Ibid. 1967 37 156. Hall G. Cochrane I. G. and Dorman R. W. International Atomic Absorption Spectro-scopy Conference Sheffield July 1969. Tyler J. B. Atomic Absorfltion Newslett. 1967 6 14. Menis O. and Rains T. C. Analyt. Chem. 1969 41 952. Holak W. Ibid. 1969’41 1713. Ando A. Suzuki M. Fuwa K. and Vallee B. Ibid. 1969 41 1974.Chakrabarti C. L. Analytica Chim. Acta 1969,42 379. - Ibid. 1967 39 293. Barnett W. B. and Kahn H. L. Atomic Absorption Newslett. 1969 8 21. Slavin S. and Sattur T. W. Ibid. 1968 7 99. Zaugg W. S. and Knox R. J. Analyt. Biochem. 1967 20 282. Zaugg W. S. Atomic Absorption Newslett. 1967 6 63. Kirkbright G. F. Smith A. M. and West T. S. Analyst 1967 92 411. Hurford T. R. and Boltz D. F. Analyt. Chem. 1968,40 379. Ramakrishna T. V. Robinson J. W. and West P. W. Analytica China. Acta 1969,45, Rose S. A. and Boltz D. F. Analytica Chim. Acta 1969 44 239. Jungreis E. and Anavi Z. Ibid. 1969 45 192. Dunk R. Mostyn R. A. and Hoare H. C. Atomic Absorption Newslett. 1969,8 79. Kadow A. and Rabban E. International Atomic Absorption Spectroscopy Conference, Yamamoto Y. Kumamaru T. Hayashi Y. Otani Y. Japan Analyst 1969 18 359. Danchik R. S. and Boltz D. F. Analyt. Chem. 1968 40 2215. Bond A. M. and O’Donnell T. A, Ibid. 1968 40 560. Reichel W. and Acs L. Analyt. Chem. 1969 41 1886. Ezell J. B. jun. Atomic Absorption Newslett. 1967 6 84. Bartels H. Ibid. 1967 6 132. Westerland-Helmerson U. Ibid. 1966 5 97. Danchik R. S. and Boltz D. F. Analyt. Lett. 1968 1 891. Bond A. M. and Willis J. B. Analyt. Ckem. 1968 40 2087. Kirkbright G. F. Rao A. P. and West T. S. Sflectrosc. Lett. 1969,2 69. Kamamaru T. Hayashi Y. Okamoto N. Tao E. and Yamamoto Y. Analytica Chim. Yamamoto Y. Kamamaru T. and Hayashi Y. Talanta 1967 14 611. 284. Spectroscopy Conference Sheffield July 1969. 47 407. - , 43. Sheffield July 1969. Acta 1966 35 524 234 PLATT Potter A. L. Ducay E. D. and McCready R. M. J . Ass. 08. Agric. Chem. 1968 51, Sheridan J. C. Lawu E. P. K. and Senbowski B. Z. Analyt. Chem. 1969 41 247. Brandenberger H. and Bader H. Atomic Absorption Newslett. 1967 6 101. - and - Ibid. 1968 7 53. Thilliez G. Ckim. Analyt. 1968 50 226. Goleb J. A. Anal. Chim. Acta 1966 36 130. Kirchhof H. Spectrochim. Acta 1969 24B 235. Brimhall W. H. Analyt. Chem. 1969 41 1349. Goleb J. A. Analytica Chim. Acta 1966 34 135. Butler L. R. P. and Schroeder W. W. International Atomic Absorption Spectroscopy 748. Conference Sheffield July 1969. 333 334 335 336 337 338 339 340 341 34