PART I FUNDAMENTALS AND INSTRUMENTATIONPart I: Fundamentals and Instrumentation 3 1 Light Sources There continues to be steady progress in the construction, operation and understanding of light sources used in analytical atomic spectroscopy. Most papers, however, cover ground that has been examined previously : nevertheless, this repetition serves to consolidate knowledge and, by examining the subject from a different viewpoint, in some cases promotes its application to analytical situations.An example of this is the increasing use of the pulsed- mode operation of light sources. Hollow-cathode lamps continue to occupy a pre-eminent position in practical AA analysis, whilst EDLs and lasers attract research interest for their value in AF analysis. I t is unlikely that this relative emphasis will change in the near future, except in special cases, such as the use of EDLs for AA determination of Se and As and the use of tunable lasers if their cost were reduced and the ‘tuning range increased. 1.1 HOLLOW-CATHODE LAMPS 1.1.1 Spectral Line Profiles and Discharge Mechanisms Tilch and Wollbrandt (605), using a Fabry-Perot interferometer with pressure scanning, examined the current dependence of the emission from a Ca HCL and found that the line profile could be represented by Voigt functions, within experimental error, and that the Ca atoms were at a higher temperature than the Ne filler gas.The temperature depend- ence of the collisional width of the Ne line corresponded to a Van der Waals interaction. An approximation which simplifies the calculation of the Voigt function has been described (545).Wagenaar and de Galan (1641) conclude from an investigation of 18 atomic spectral lines of analytical interest that, except for a few elements (e.g., Ca and Si), hyperfine structure is a major factor in determining line profile, The width of the resonance lines of B emitted by an HCL depend to a large extent on the filler gas (532).Using Ar, Kr and Xe, the line half-widths were found to be 0.014-0.027 nm and the line profile was Gaussian with a small blue shift indicating Doppler broadening, an effect which does not occur for Ne or He (1540), caused by the high ejection velocities of sputtered B atoms. I t was concluded that the sputtered atoms were excited by collision with rare-gas atoms and that the cross-sections for collisional excitation differ for the various gases.Agreement between Monte Carlo calculations and experimental observation suggests that collisional excitation of the sputtered B atoms in Ar, Kr or Xe occurred within a very few collisions of their ejection from the cathode surface. From an investigation of the mechanism of the hot hollow-cathode discharge at various operating currents (744), it was concluded that sputtering was the dominant process up to 0-3 A, and that evaporation accompanied by fractionation, as the cathode temperature increased above 1400°K, occurred between 0.3 and 1.0 A.The authors also inferred that interatomic collisions played a major part in the hot hollow-cathode discharge. Kagan et al.(1354) have proposed a mechanism of formation of the distribution function of fast electrons in a hollow cathode. 1.1.2 Construction Lowe (523) has described a demountable high-intensity lamp in which the cathode holder was water cooled and a continuous flow of inert gas passed through the lamp. For low-melting- point cathodes, considerable gains in intensity were achieved by cooling the cathode, while high-melting-point materials give comparable performance to sealed high-intensity lamps.A method has been described (154) for the refilling and alteration of used HCLs, although the emission intensities of these lamps were less than those of the originals. In situations where replacement HCLs are not readily available, attempts at regeneration may be justified, but in view of the difficulties in manufacturing bright stable lamps with an acceptable life, it should not be lightly undertaken.4 Part I : Fundamentals and Instrumentation 1.1.3 Operation Interest continues in the use of pulsed HCLs for AFS.Cordos and Malmstadt (176, 177) have described a power supply which produces current pulses (200 mA peak current, 10 ms prrlse width) with a reproducibility of 0.002-0.003 (RSD) giving a light output stability of 0.0008 over a 10 minute interval, with a long-term drift (several hours) of 0.002.Human (894) investigated the use of pulsed HCLs in AF using current pulses up to 900 mA and 200 ms with a Cu lamp. For currents greater than 400 mA, however, there was no further increase in signal-to-noise ratio, despite the increase in the emission intensity of the lamp, due to severe self-reversal of the emission line.Additional information on pulsed and modulated HCLs will be found in references 31, 202, 528, 918, 1215, and in ARAAS, 1971, 1, 2, and ARAAS, 1972, 2, 2. Two papers have dealt with different aspects of the application of a magnetic field to a hdlow cathode. An A1 HCL placed in a magnetic field of up to 900 G and cooled with liquid nitrogen (835) produced double the intensity for the 396-15 nm line with a slight increase in line width.A development of potential value in AA analysis, where background absorption is a problem, is the modified HCL described by Stephens (1471), which is stable in a magnetic field. Zeeman splitting of the emitted atomic lines is achieved by using a low power electromagnet so that, by modulating the current to the electromagnet, the emission from the lamp is also modulated in perturbed and non-perturbed Zeeman components, which can then serve as reference and sample beams, respectively.This technique has previously only been applied to Hg analysis (ARAAS, 1972, 2, 5). 1.2 MICROWAVE DISCHARGE TUBES 1.2.1 Temperature Control and Muld-element Lamps Elcctrodeless discharge lamps are an intense source of narrow-profile atomic spectra, but their application to AAS and AFS has been limited by the practical problems of operation (519, 811, 1303, 1449).In some cases, an increase in the light output of a lamp is accomp- anied by an increased spectral line width and hence a loss in absorption sensitivity. Attempts to improve the stability of EDLs by changing the manufacturing technique (975) and by using optically controlled feedback (1243) continue to be made, but temperature control of the lamp appears to be the most promising (26, 1244, 1304).In the case of multi-element lamps it was found (174) that the optimum temperature depended solely on the element, while the intensity of the emission could depend on the other elements present.Pate1 c't al. (26, 264) have presented the results of fluorescence measurements with a graphite rod atomizer using multi-element lamps under the following conditions : Cd-Hg-Zn at 275OC, Cd-FeHg-Tl-Zn at 285 OC and Ag-Co-Cr-Cu-Fe-Mn-Ni-Pb-Sn at 45OOC. An alterna- tive system of multi-element operation using AF employs two dual element lamps driven by a single power supply (181). 1.2.2 spectral Overlap In a few cases advantage can be taken of spectral overlap to determine one element by using the radiation from another. Norris and West (1040) have employed this approach for the determination of Cr by AA and AF using the overlap of the Ne 359.352 nm line with the Cr 359.349 nm line.* 12.3 Gas-flow Microwave Lamps Dagnall et al.(716) have described a new type of discharge lamp in which argon flows through a quartz tube at atmospheric pressure. A quartz cup containing the element of interest was suspended in the tube in a region where a micro-wave cavity excited the discharge. Although the lamp life was short, an intense emission was obtained for elements * See also: Manning, D.C., Atom. Absorp. Newsl., 1971, 10, 97.Part I : Fundamentals and Instrumentation 5 having compounds with boiling points in the range 1240-2200°C, and they could easily be recharged. Bazhov and Balshin ( 5 5 5 ) have described a lamp consisting of an opaque bulb divided into two chambers by a transparent window. One chamber was an EDL, and the other was filled with vapours of the sample.The latter was provided with a side window for the outlet of fluorescence emission. 1.2.4 Safety Stanley et al. (506) have sounded a warning note in the use of microwaves, as the health hazards are not yet predictable. They have found that radiation levels in excess of the prescribed safety standards have been detected in the vicinity of commonly used microwave cavities.Consequently, shielding precautions and on-site radiation surveys are to be recom- mended. It has been reported that an r.f. spectral lamp may be readily ignited by irradiation with light from a photographic flash-gun (1 194). 1.3 LASERS 1.3.1 Application to Atomic Fluorescence Analysis Winefordner (881) and Omenetto (1526) have demonstrated that there are advantages in using high-intensity pulsed sources to achieve near saturation of atomic energy levels in AF.These advantages are (a) the fluorescence signal is not greatly affected by source stability, ( b ) for a two-level system, the fluorescence signal is no longer dependent on the quantum efficiency, and so one can use more-reducing flames, and (c) the linearity of the analytical curve is greatly increased, particularly at high optical density. A limitation of the pulsed high-power laser is its low pulse-repetition rate.Using tunable laser excitation (200) only one source was required to excite the fluorescence spectra of 35 elements in air/H,, air/C2H1 and N20/C2Hz flames. Flame back- ground and analyte emission interferences were essentially absent, and non-resonance fluorescence was used to eliminate noise due to scatter of source radiation by the flame gases. 1.3.2 Applications to Fundamental Studies of Atomic Spectra From the point of view of the practising analyst the additional complexity and expense of laser-excited AF is not justified, as the technique has little to offer beyond what can be obtained using simpler AA techniques.The greatest value of laser-based studies is likely to be in the insight this approach provides to atomic and molecular energy levels and transitions, c.g., the fluorescence spectra of Iz excited with 528.7 nm radiation (232) and of CH in an O,/CzHz flame excited at 431.5 nm (1068); the location and pressure broadening of absorption lines (1001) and laser-saturated atomic resonance fluorescence (7 lo).Konjevic and Konjevic (997) have introduced an aidnatural gas flame into the cavity of a dye laser and used it for the determination of Na by AA down to 0.2 ng 1-'. This approach was proposed by Peterson et al. (ARAAS, 1971, 1, ref. 229) for the enhancement of absorption spectra and would appear to justify further investigation. The literature on the design and operation of lasers is voluminous, but it is not appropriate to record it here; however, typical material relating to dye lasers is given in references 1242, 1350, 1360.A comprehensive literature survey of chemical lasers has been prepared by Arnold and Rogjeska (240). 1.4 CONTINUUM SOURCES The suitability of a continuum light source for use in AA has again been investigated (604).A high-current (100 A) stabilised graphite arc which proved to be a better source than tungsten-filament, deuterium or xenon lamps has been investigated theoretically for the determination of Mg by AA (1129). The temperature of tungsten-ribbon lamps has been determined (949) by relative measurements of the spectral and integral radiant energies, in6 Part I : Fundamentals and Instrumentation the wavelength range 500-800 nm; the results in the temperature range 2000-3000°K agreed with the certificated values within 30°. I t was found that the electrical power consumed was approximately proportional to the integral radiation. With the increasing use of pulsed line sources in AF, parallel studies of the suitability of flash tubes would be of value. A source of high intensity for ultraviolet (184.9 and 194.2 nm) radiation based on a mercury vapour discharge lamp has been described (250), in which the efficiency is improved by using an electron-emitting hot-gas filament inside the source envelope and by employing high current densities (20 A cm-').