Plasma emission sources in analytical spectroscopy—II

Talantn, Vol 22, pp 553-562

Pergamon

Press, 1975 Prmted m Great Br~tam

TALANTA

REVIEW*

PLASMA EMISSION SOURCES IN ANALYTICAL SPECTROSCOPY-II S. GREENFIELD,

H. McD. MCGEACHINand P. B. SMITH

Albright & Wilson Limited, Industrial Chemicals Divlsion, P.O. Box 80, Oldbury, Warley, England (Recetued 13 December 1974. Accepted 10 January 1975) Summary-The use of microwave plasmas and of capacitively coupled high-frequency sources for emission spectrochemical analysis is reviewed.

The first part of this review gave a theoretical background and discussed plasma jets operated with d.c. This part deals with microwave plasmas (by which we mean plasmas operated at frequencies greater than 300MHz), and with high-frequency (hf) plasmas (at frequencies in the range l-3OOMHz) which are capacitively coupled to the power generator. In the first section we give a survey of the development of these sources as applied to spectrochemical analysis of solutions and solids; the great majority are concerned with solutions. In the second section some aspects of these systems are compared and discussed. The third section lists some specialized applications, mostly with gaseous samples, in gas analysis and isotope analysis, and as detectors in gas chromatography. We have excluded applications of electrodeless discharge lamps to atomic-absorption and atomic-fluorescence spectrometry, where the emission from the plasma does not characterize the sample but rather constitutes the excitation source for it. Happily the confusing literature on the preparation and operation of electrodeless discharge lamps has recently been critically reviewed.’ DEVELOPMENTOF PLASMASOURCES Capacitively coupled high-frequency plasmas

In 1941, Cristescu and Grigorovici3 reported their experiments on a discharge produced by applying the output of a high-frequency oscillator to two circular plates which were separated vertically by up to 15 cm. The lower plate had a copper cone with a platinum tip attached to it. The plates formed a condenser which was part of the circuit which determined the frequency of the oscillator, typically in the range 6090MHz. A discharge could be formed at the tip by touching it with an isolated conductor which could emit electrons easily when heated by the strong electric field. These electrons gain kinetic energy from * For reprints of this Review see Publisher’s announcement near end of this Issue. t Part I-Tulunta 1975.22. 1. TAl. 2217-A

plasmas as

the electric field and this energy can be transferred by collision to heat the gas and to produce by ionization the number of electrons necessary for a sustained discharge. A temperature of 4000 K was found when a power of 650W was supplied. The authors later gave a theoretical treatment4 of the discharge. This discharge was utilized in 1956 as a source for spectroscopic gas analysis by Stolov5 and in 1957 for spectrochemical analysis by Badarau, Giurgea, Giurgea and Trufia.6 They used a hollow cylinder for the upper electrode and noted that this electrode was not absolutely necessary, but had a useful function in concentrating the lines of force of the electric field. The system is therefore very different from an arc, where a substantial current flows through both electrodes, which are therefore both necessary. The essential electrode is a tip where a very strong electric field is developed. Discussions and applications of one- and two-electrode plasmas of this kind were given by Kapicka,7,8 Dunken, Mikkeleit and Kniesche,g Tappe and van Calker,” Dunken, Pforr and Mikkeleit,” Trunecek,l’ Dunken and Pforr,13 and Pforr and Langner.14 The glow discharge lamp of Vurek and Bowman’ 5-’ 7 can formally be included in this class, since it has a pointed electrode at which the discharge is formed and an annular electrode surrounding the discharge tube, but the very low power (10 W) applied gives rise to a discharge of much less luminosity. A new kind of discharge in which a d.c. voltage was superposed on Cristescu and Grigorovici’s highfrequency plasma was described by Cristescu” in 1960. While the hf voltage was applied across two electrodes as before, these electrodes were held at the same d.c. potential and an adjustable d.c. voltage was applied between them and a third electrode, movable between them. Altering the d.c. voltage so as to vary the d.c. in the range 0.1-l A changed the apparent excitation temperature and also the appearance of the plasma. Applications to spectrochemical analysis were made,‘g-20 and later studies were reported by Zakharov.21 In contrast to these systems where the upper electrode, if there is one, acts only as a guide for the

553

5.54

S. GREENFIELD,H.

McD.

MCGEACHIN and P. B. SMITH

lines of force and does not conduct a significant current, is the type which is the high-frequency analogue of d.c. arcs or plasma jets, where the electrodes carry a substantial current. Zheenbaev22~24 developed sources where the discharge space was subjected to hydrodynamic compression, either by sucking air and the discharge products through the central hole of the upper electrode, or by blowing a stream of air along the axis of the discharge. The increased pressure resulted in a higher temperature because of more collisions, and an improvement in sensitivity of one or two orders of magnitude was claimed. The samples were usually aqueous solutions and were continuously vaporized from a Perspex container above the lower electrode by an auxiliary electrode dipping into the solution. Gostkowska and Ekiert” used a high-frequency (20 MHz) arc at atmospheric pressure to determine trace impurittes m semiconductor materials by fashioning rods out of these solid samples and fixing them to copper electrodes, while Eramets and Kukkasjlrviz6 used pelleted samples and graphite electrodes in a helium atmosphere at reduced pressure, to determine bromine, chlorine and tellurium. Scholz” and Roddy and Green’s designed torches in which a single conductor was used to form both a central electrode and a coaxial coil surrounding it. There is thus both capacitive and inductive coupling to the plasma, which is formed at the tip of the central conductor. Torches of this kind were utilized for spectrochemical analysis by Mavrodineanu and Hughes,29 who passed the sample aerosol through the hollow conductor. to emerge through small holes at the tip; they found that other methods of introducing a sample aerosol led to partial short-circuiting through liquid films. Other users10.30.31 did not seem to observe this effect. Egorova32-34 used purely capacitive coupling with two external annular electrodes, one at each end of a water-cooled discharge tube. This type of coupling had been used earlier for isotope analysis in plasmas at reduced pressure.35p37 Egorova’s plasma was at atmospheric pressure in argon, which was introduced tangentially, and was viewed along its axis. She found that two kinds of discharge could be excited. The first was a constricted plasma like an arc channel while the second was a continuous diffuse luminescence which filled the tube. Lines of high excitation energy were excited in the first type whereas those of low energy predominated in the second. Both types were useful for analysis and the change from one to the other depended on the concentration and ionization potential of the sample atoms. A plasma excited by a single, water-cooled, external, annular electrode was described by Grigoriev, Frolov and Sanodze.38 Again the discharge consisted of a thin filament when pure argon was used but changed to a diffuse form when an easily ionized element (an alkali or alkaline earth metal) was mtroduced at a suitable concentration. The apparatus included two atomizers, one for the sample and the other for the easily ionized element.

Microwave Two

plasmas

types of apparatus have been used to excite microwave plasmas. A discharge tube with a central electrode can be used; the electrode IS coupled C%I a waveguide to the mrcrowave oscillator and the plasma is formed at the tip, like some of the highfrequency plasmas already discussed. The other method of excitation is to place a discharge tube (wrth no electrodes) in a resonant cavity. Most of the sources which use a central electrode are derived from the designs of Cobine and Wrlbur3” and Schmidt,40 which operate at 2450 MHz. The first application to the analysis of solutions was reported by Mavrodineanu and Hughesz9 in 1963. who used 2 kW power. They found a gas temperature in the range 2900-3300K by observing that molybdenum could be melted, but not tantalum or tungsten The excitation temperature was much higher. Various working gases were used; hydrogen or helium gave the least background and hydrogen was more effective in volatilizing refractory substances. An mcandescent tip. emitting electrons, was requned to sustain the discharge m hydrogen. Kessler and his assoctates studied a similar torch4’*42 and applied it to materials used in the glass industry43*44 and to general analysis of solutions.45 Tappe and van Calker’” reported briefly on sources at frequencies of 461 and 24OOMHz. Yamamoto and Murayama46 studied a source with a frequency of 520MHz before using4’ the almost universal microwave frequency of 2400 MHz. Yamamoto and MurayamaJh studied a source with a frequency of 520 MHz before using” tions to the analysis of steel have been reported “’ ” The first successful applicatron to spectrochemlcal analysis of a microwave dtscharge formed without electrodes and in a resonant cavity seems to have been the analysrs of nitrogen isotopes by Broida and Chapman s3 in 1958. An application to the analysis of soluttons was described by Yamamoto.55 McCormack, Tong and Cooke,5” in developing a detector for gas chromatography. found that at low pressure it was necessary for the diameter of the discharge tube to be at least 1cm, for smaller tubes resulted m an unstable plasma; at atmospheric pressure stable and more intense discharges were obtained m tubes as narrow as 1 mm. Runnels and Gibsons6 investigated the use of this type of plasma for the excitation of metals in solution. The low power used (about 25 W). made the introduction of the sample as an aerosol impracticable, since the plasma was easily extmguished by the solvent. The sample was therefore placed on a platinum filament which was heated to give complete sample evaporation outsrde the drscharge. The design of the vaporization chamber had large effects on the sensttlvity and reproducibihty of the emission; particular attention to the prevention of plating-out was necessary. A number of apphcations from the same school have appeared” 5v and the most recentreport, by Lichte and Skogerboe.“” mdicated that modifications to the resonant cavity allowed the introduction of desolvated aqueous solutions. This

Plasma emission sources in analytical spectroscopy-11

555

Table 1. Systems classified by power Power,

Type

Microwave

looa

High-frequency

10 250 500 1500 2ooo

hadpreviouslybeenachievedbyotherworkerswithlowpower systems.6’-64 Short residence time is a limitation to sensitivity and the possibility of sealing samples in a tube to overcome this has been investigated. 65, 66 All the work already mentioned on the use of discharges excited in a cavity was carried out with low-power sources (200 W or less). An exception is the generator used by Tsemko and his associate@‘* 68 which was capable of supplying more than 2 kW to an argon plasma with tangential flow at atmospheric pressure, the source being viewed either along its axis or at right angles to it.

REVIEW OF SYSTEMS USED FOR ELEMENTAL ANALYSIS OF SOLUTIONS

Plasma generators A survey of the different types of generator used must be incomplete, since authors frequently do not give details. Even when the power output of the generator is specified we do not usually know how much reaches the plasma. Kessler and Gebhardt’s paper 43 is an admirable exception: we are told that of the 2.5 kW available from the magnetron. the maximum they can utilize, because of the gas flows adopted, is 1 kW and that in fact they choose 860 W. Of this, 16OW is reflected, so 700W are transferred to the plasma. Thus a simple statement that they use a generator of maximum output 2.5 kW would not be very relevant. It is likely that the figures quoted in many instances are for the maximum output (since this figure is supplied by the manufacturer and requires no measurement). These figures are therefore likely to represent only an upper limit for the power actually supplied to the plasma. For microwave sources where a resonant cavity is used, its design is of great importance in the transfer Table

2. Methods

References

W

5660, 64, 69. 70 46, 61-63, 65, 66 45, 47, 52 29, 41-43, 48, 67, 68 15-17 29 30 32-34 2s

of power; some designs have been discussed.60*7’,72 Table 1 shows that there are many applications to the analysis of solutions or solids for which a lowpower microwave source is at least sufficiently promising to be thought worth reporting. This is in spite of the restriction to very small samples, necessary for the stability of the discharge. If applications to the analysis of gaseous samples were included here there would be a greater preponderance of these devices. The situation with hf sources is very different; with the exception of Varek and Bowman’s glow discharge,“-” there are no such systems reported with power of less than 250 W and it is possible to infer that this was considered too sma11z9 The impracticability of forming a discharge with a low-power hf oscillator is probably due to the relatively large skindepth at these frequencies, which means that the power is dissipated in a relatively large volume, giving a power density insufficient to maintain the discharge. The frequencies reported for the hf sources (with one early exception36 where it is ISOMHz) lie in the range 6-60 MHz. Apart from commercial availability, reasons are not advanced for the values chosen; it is our opinion that, within fairly wide limits, the choice is not of great importance. Sample introduction Table 2 gives a survey of methods used to introduce liquid and solid samples; we exclude cases where a solid sample is placed on, or used as, an electrode. In the cases noted, desolvation was used to prevent plasmas of low power from being extinguished by the solvent. Chemical generation included the production of hydrogen sulphide from cast iron5’ the reduction of mercury compounds3’.” and the generation of arsine.59 Kessler4’ gives a well-justified warning on the necessity of reproducible particle-size and rate of of Introducing

samples

Method

References

Pneumatic nebulization

6. 7, 9. 10, 19, 29, 32-34, 38. 43-48, 52*. 60*, 61, 62*, 67, 68

Ultrasonic nebulization Evaporation from heated filament Chemical generation of vapour In sealed tube As powder

11, 13, 14, 30, 32, 33, 64* 15-17. 56, 57, 63, 69, 70 31, 57-59 65. 66 41. 42

* With desolvation.

556

S. GREENFIELD, H.

McD. MCGEACHIN and P. B. SMITH

feed in the quantitative analysis of powders. Ultrasonic nebulization often results in an improvement in sensitivities and detection limits: a comparison is given by Egorova. 32*33The most convenient method, well suited for automatic analysis of a series of samples on a rotating turntable, is pneumatic nebulization; none of the others is easy to use in this fashion. Limits of detection

Many of the limits of detection given in the papers reviewed are (following Kaiser73) that concentration which is predicted to give a signal of magnitude which is a certain multiple of the standard deviation of the background. The multiplier used is often two, sometimes three and occasionally one. Skogerboe, Heybey and Morrison74 give what seems to be a technically more correct procedure to take account of the fact that the background and its standard deviation are usually estimated from a small number of readings. The point had not been overlooked by Kaiser, who not only recommended that at least twenty blank analyses be used and that the value of the multiplier should be three but warned against accepting literally the corresponding probability of being in error. In spite of the fact that the published values are therefore not all comparable, we have made no attempt to adjust them to the same basis, for it seems to us that the differences due to the different methods of estimation are of little consequence to the spectroscopist engaged in practical analysis, because of other factors which affect the limit and may be beyond his control. It seems more important to us that the published limits should have been checked experimentally by analysing known samples having concentration which lie quite near these limits. Extrapolation from concentrations much higher can be misleading; in one case,66 the discrepancy between extrapolation and experiment pointed the way to an improvement in experimental procedure. Table 3 lists some of the best values of published limits for most of the elements which have been studied. They are often determined with the use of pure solutions although they are liable to be influenced by the matrix of a real sample. Many other factors,

such as the nature and acidity of the solution, gas pressures and flow-rates, available power, type of nebulizer used, and the part of the plasma viewed, also have their effects. If only one element is to be determined these parameters can be optimized if necessary but if the source is used for the simultaneous determination of several elements, either by photography or a multichannel photoelectric system. the conditions will not be optimum for all elements. It seems likely, therefore, that different operating conditions explain the large number of different references cited in Table 3 for the best detection limits; this was brought home to us when compiling the table, when we noted that different systems which might be expected to behave in a similar way frequently gave detection limits differing by factors of perhaps 100, one system favouring some elements, and another others. With the possible exception of systems which are designed and optimized for one element, we would therefore not recommend the blind acceptance of a low detection limit as an indicator of merit of the emission source. The use of ultrasonic nebulization in particular can often make a dramatic improvement. Matrix

eficts

One type of matrix effect 1s due to the presence of refractory radicals; atoms which are not free do not contribute to the atomic spectrum. The intensity of the emission of an element will depend therefore on the concentration in the matrix of other elements with which the element may combine. This effect does not occur if the gas temperature is sufficiently high, because then the radicals are dissociated; freedom from this type of interference is one of the great advantages of working with a high-power source. It is our opinion that the dissipation of several kilowatts in the plasma is necessary to obtain this advantage and that most of the sources reviewed here will therefore be subject to this effect, whether mentioned by the experimenter or not. The presence of radicals also leads to the emission of band spectra, which are likely to cause spectral interference.

Table 3. Detection limits obtained with microwave and capacitively coupled high-frequency plasma sources Element

Detection limit,

Wavelength,

PPm

nm

Reference

Aluminium

0.02

396.2

30

Antimony Arsenic Barium Beryllium Bismuth Boron

0.1 {0.1 0.03 0.05 0.05 1.0 0.01

231.2 259.8 193.7 455.4 234.9 472.3 249.8

64 312,61> 60 38 32 64 60

Bromine Cadmium

500

2 X 10-7

470.5 228.8

26 66

Other references to this element 6, 8, IO, 13, 14. 22, 25, 29, 32, 33r 38. 42-44. 4649. 64 25. 29. 33, 47. 63 25, 29, 47, 48, 59, 61, 63 6, 7, 9. 10, 14, 22, 30, 32, 33, 45, 47, 69, 76 6, IO, 22, 33. 63 25, 30, 47. 48 6, 10, 22. 24, 29, 32, 42, 45. 47, 48, 51, 63, 67. 68 10, 16, 21, 30. 32, 33, 45, 47, 48, 6&63, 65, 67-69

Plasma

emission

sources Table

Detection Element

iw

limit,

in analytical 3. Cmtrrwd

Wavelength, nm

Other references this element

Reference

Caesmm Calcium

2 0005

455.5 422.7

10 30

Carbon Carbon as CN Cerium Chlorine Chromium Cobalt Copper

100 4 1.5 500 OQOl 0001 00001

241.9 388.3 413.4 351.9 345.4 324.8

42 42 42 26 56 56 56

Dysprosium Erbium Europlum Gadolimum Gallium Germanium Gold Hafmum Holmium Indium Iodine Iron

0.04

396.8

15

0.01 2 0.04 1.5 @02

459.4 432.6 417.2 265.1 261.6

38 15 62 47 30

10-e 10 O+)ol

451.1 206.2 373.5

66 63 56

Lanthanum Lead

0.005

405.8

60

Lithium Magnesium

0.01 ow5

670.7 285.2

38 30

Manganese Mercury Molybdenum Neodynium Nickel Niobium Palladium Phosphorus Platinum Potassium Praseodymium Rhenium Rhodium Rubidium Samarium Scandium Selenium Silicon Silver Sodium Strontium Sulphur Tantalum Tellurium Terbium Thallium Thorium Thulium Tin

0%04 lo- 5 O+l45 0.3 0.01 0.5 @4 10 @I 0.3 0.5

403.1 253.6 379.8 430.4 341.5 405.9 34@6 265.9 769.9 422.5

30 58 45 75 30 52 41 68 30 16 75

0.05 8 02 20 004 0.2 0.001 0~001 0.0005 0.2

369.2 4201 4424 424.1 196.0 251.6 328.1 589.2 460.7 216.9

47 10 75 32 60 47 56 14 38 51

0.2 1 10-e 8

214.3 432.6 377.6 401.9

61 75 66 48

Tungsten Uranium Vanadium Ytterbium Yttrium Zinc

0.2 0.1 0.1 0.4

303.4 365.4 334.9 429.5

30 52 64 > 48

0.015 0.5 0.03 0.0006

437.9 398.8 431.5 213.9

30 32 75 60

Zirconium

1

339.2

42, 52

Titanium

557

spectroscopy-H

to

6, 22, 29 9, 10, 13, 14, 16, 17, 22, 25, 29, 34, 38, 42, 44-48, 61, 64, 67, 68 8 6, 22, 48 9, 10, 14, 22, 32, 33, 38, 42, 45 10, 29, 30, 32, 38, 42, 45, 60, 64, 79 8, 10, 16, 25, 29, 30, 32, 38, 45, 48, 61, 64, 61-69 29 29 29 29 47 68 10, 25, 29, 32, 33, 47 29 29 22, 29, 32, 33, 47, 62, 64, 65 29 6. 7, 10, 13, 14, 16, 25, 29, 30, 32, 34, 4245. 48. 61. 63. 64. 67, 68 22, 19 6, 8, 10, 13, 14, 16, 19, 25, 29, 30, 32, 38, 48, 63-65, 69 10. 14. 30. 45 10, 13, 14, 16, 22, 25, 32-34, 42-45, 48, 64 9, 10, 13, 14, 25, 32-34, 38, 45-48 10, 16, 2931, 47, 60-63, 65, 70 8, 25, 29, 32, 33, 38, 42, 47, 50 22, 29 10, 25, 29, 32, 33, 38, 42, 45-47. 67, 68 29. 68 29 29, 32, 33, 67 10, 47 6, 7, 10. 14, 15 22, 29 29

63,

38,

33,

61,

22 22. 29 29 61, 63 25, 29, 67, 68 10, 25, 29, 32, 33, 45, 64, 69 10. 13, 15, 16, 25. 29, 42, 44, 45, 61 9, 10, 22, 24, 30, 67, 68 29 25. 68 26, 47 10, 25, 29, 30, 32, 38, 65 29 6, 22, 29. 48, 61, 64 10, 48, 68 8, 10, 47, 50, 68 29 6, 19, 29, 32, 42, 46, 47, 60 29 6, 7, 10, 13, 14, 16, 29, 30, 32, 33, 38, 4548, 6 l-64, 67-69 41, 68

558

S. GREENFIELD,

H. McD. MCGEACHINand P. B.

SMITH

Table 4. Practical applications of microwave and capacitively coupled high-frequency plasma sources Field of application Clinical chemistry Impurities in semiconductors Limestone, dolomite Glass 011

Steel

Elements determined Ca. Cd, Cu, Fe, Hg, K, Mg, Na, Pb, Zn Si, Mg, Cu, Mn, MO, Fe, Al, As, Bi, Tl, Sb, Pb, Ni, Na, Ca Mg. Fe, Al, Ca, Mn Na, Ca, Mg. Al, Fe Fe, Cr. Ni. Ag Al, W, MO, B, Nb. TI, Zr

Another effect which is liable to occur wtth apparatus of any power or frequency is a change in emission intensity if an element of low ionization potential is introduced. The immediate result of this is likely to be an increase in electron density, which will move the ionization equilibrium of the other atoms towards the neutral side. This simple consideration is sometimes adequate to explain at least qualitatively the effects observed: these effects are frequently that elements of high ionization potential are not affected, smce there are few ions of such species m any case, while those of low ionization potential, which are initially appreciably ionized, will show a noticeable change if the equilibrium is altered. Changes in the spatial distribution of the plasma32.38 have also been invoked by Murayama75,76 and other studies have been made by Cristescu and Giurgea,” Kitagawa and Takeuchi,77x78 and Pupyshev and Muzgin.79 Very pronounced enhancements are sometimes observed; the detection limits for rare earths were improved by factors of up to lOOtI by the addition of sodium,75 but reductions in sensitivity have also long been known.’

Many of the papers mentioned are exploratory and aim at establishing the potential of the technique for the analysis of solutions. Papers referring to specific elements can be traced from Table 3. More substantial practical applications are listed m Table 4. Plasma parameters

It is reasonable to assume that all the devices mentioned have high excitation temperatures, for otherwise they would be useless as excitation sources. We have already mentioned our opinion that most of the sources reviewed do not have powers sufficiently high to produce a high gas temperature. Most, therefore, will not be in local temperature equilibrium (LTE). which requires the different kinds of temperature to have the same value. Low-pressure plasmas are likely to be still further from LTE, because the lower number of collisions impairs the transfer of energy. The literature ofplasma parameters such as temperatures of various kinds and particle densities of different species is very large. A few papers with relevance to the small plasmas used for laboratory spectrochemical sources are those by M01lwo;~~ van Calker ;” Lochteand TravHoltgreven; a2 Lanz, Lochte-Holtgreven ing ;83 Jecht and Kessler; 84 Pforr and Kapicka;85 Kapicka;86 Egorova; 87 Baltin, Batenin, Goldberg and

References 15-17 25 34, 43, 44 44 45 49-52

Tsemko;88s9 Britske and Sukach;” Kapounova;91 Kapoun;92 and Busch and Vickers,93 as well as referenCeS4,h.20,23,29,41,48,62,64,75,77,78 already mentioned in the text. SPECIALIZED APPLICATIONS Gas mixtures Stolov’ used a high-frequency brush discharge formed at a pointed electrode to analyse binary mixtures of nitrogen and carbon dioxide. This mixture was also studied by Botschkowa, Frisch and Schreider37 who also undertook trace analysis with their ~-MHZ, 350-W source. White, Watkins and Fletcher94 analysed respiratory gases for oxygen and nitrogen. Vashman, Lipis and Teterina” used a microwave (3000-MHz) discharge for the analysis of argon-helium mixtures. Given, Magee and Wilson,96 and Chakrabarti, Magee and Wilsony7 investigated the possible application of Tesla-type discharges to gas analysis, and also used a 230-MHz source to study argon-carbon dioxide mixtures.98 An &MHz source was evaluated by Boos and WinefordnerQ9 for the detection of air, CO, C02, SO*, NH,, NO, N02, N2 and CH4. The effect of parameters such as gas pressure and power on the spectral intensity of nitrogenhelium mixtures was Investigated by Snopov.“’ Trace impurities in rare gases Servigne, de Montgareuil and Domine”’ used a microwave discharge at low pressure to determine nitrogen in the rare gases. Ishida”’ used an hf source for hydrogen and nitrogen in argon. Botschkowa, Frisch and Schreider 37 determined nitrogen, hydrogen, oxygen and carbon monoxide in the rare gases, and traces of rare gases in others, with an hf discharge. Fay, Mohr and Cook103 studied microwave discharges, Tesla discharges and Geissler tubes before opting for a silent, high-voltage (lo-kV) mainsfrequency discharge for detecting nitrogen in argon. Gutkina and Maslennikova,‘04 however, pointed out the advantages of high frequencies in studying traces of nitrogen in helium. Taylor, Gibson and Skogerboelo5 adapted the system of Runnels and Gibsons6 to the determination of carbon-, oxygen-, nitrogen-, and hydrogen-containing compounds in argon. Penchev, Belchev, Piperov and Belinov”” analysed helium for traces of neon with a 60-MHz plasma. lsotope

analysis

The spectroscopic analysis of isotopes depends on the shifting, often by as much as 1 A, of atomic lines

Plasma emission sources in analytical spectroscopy-11 or molecular bands, caused by the different masses of the isotopes. Excitation of the spectra seems to be accomplished very conveniently in high-frequency and microwave discharges. Nirrogen. The most widely investigated element is nitrogen, perhaps because of the importance of biological experiments with labelled nitrogen, ’ 'N. Hoch and Weisser35 excited the discharge by applying the output of a diathermy unit of unspecified frequency to aluminium foil wrapped round the ends of a discharge tube. Broida and Chapmans experimented with both sealed and unsealed discharges and with frequencies of I50 and 2450MHz. A sealed tube had the advantage of requiring only a small sample and, although special care was required in filling it, was adopted. The higher frequency was chosen because no tuning was required and it was more sensitive. The pressure in the discharges was chosen to be 15 mmHg, and the power supplied was 125 W. Faust”’ used a frequency of about 7 MHz to excite spectra in a capillary tube and measured the intensity ratios with a step-filter rather than a rotating sector. Zaidei, Lazeeva and Petrov”’ applied excitation in a high-frequency electrodeless discharge to the determination of the isotopes of oxygen and nitrogen in metals. Nemets and Petrov”’ reported the simultaneous determination of the isotopes of nitrogen and hydrogen, and of carbon and oxygen isotopes. Sommer and Kick' lo used a 2450-MHz discharge and gave details of their sample handling. Leicknam, Figdor. Keroe and Muehli” designed an apparatus of high stability working at 100 MHz, and described suitable detection equipment; operating conditions were discussed by Leicknam. Middelboe and Proksch.’ ” With very small (less than 5 pg) samples of nitrogen, the discharge is easily extinguished and Goleb and Mlddelboe~13 added a mixture of helium and xenon to maintain it. Ferraris and Proksch’ i4 evaluated commercially available components for constructing a system to excite the spectra and discussed different methods of calculating the isotopic concentrations from the spectra. Kumazawa’ ’ 5 discussed some practical problems. Keeney and Tedesco’ I6 developed a method for sample preparation and evaluated a commercially available analyser (Straton. type NOI-4). Middelboe”’ discussed a high-resolution method for the calibration of commercialiy available analysers. Lloyd-Jones, Hudd and fill-Cottinghaml 1* reported ins~umen~l and procedural modifications for the use of the Straton NOIspectrometer and made recommendations on the calculations and sample preparation. Hydrogen. The analysis of hydrogen-deuterium mixtures was investigated by Broida and Moyer36 and by Broida and Morgan.“’ Sources at frequencies of 150 and 2450 MHz were tried: the ISO-MHz source was preferred since the microwave source showed strong self-absorption. A flow system was used, since with a closed system the calculated concentration decreased with time. (It is interesting to note that the opposite options were later chosen by Broida and

559

Chapman” for the analysis of nitrogen, for different but equally pertinent reasons.) A method for the analysis of mixtures with l&90% hydrogen was reported by Veinberg, Zaidel and Petrov,“’ with smaI1 amounts of deuterium by Zaidel and Ostrovskaya”’ and with small amounts of hydrogen by Nemets, Petrov and Shabdukarimov.‘22 Nemets and Petrovlo9 reported on the simultaneous detection of the isotopes of hydrogen and nitrogen. Carbon. Isotopic analysis of carbon was described by Zaidel and Ostrovskaya,‘23 who used the intensities of carbon monoxide bands. A similar technique was used by Nemets and Petrov.“’ Uranium. The application of electrodeless discharges to the isotopic analysis of uranium was reported by Capitani and co-workers.1”L’26 l?efecfors for gas c~o~fogrup~y

Although an hf discharge had been suggested by Karman and Bowman’*’ in 1959, and discharge excitation by a Tesla coil by Sternberg and Poulsonrz8 in 1960, practical applications to chromatographic detection were not reported until 1965. McCormack, Tong and Cooke55 ex~rimented with high-vol~ge a.c. and d.c. glow discharges, with a hollow-cathode d.c. discharge, and with electrodeless discharges in argon and helium at 8 and 2450 MHz. They found the intensities of the spectra emitted from the microwave plasma to be significantly higher than those from the others. The spectra emitted from this discharge showed that a variety of organic molecules had been fragmented to atoms or diatomic species; these spectra could be utilized as sensitive detectors of compounds containing carbon, iodine, sulphur, phosphorus, chlorine and fluorine, and applications to other compounds were suggested. An application of this detector by Bathe and Lisklz9 to the determination of organophosphorus insecticide residues was published simultaneously. This was the first of a series, subsequent members of which were concerned with the determination of iodinated herbicide residues and metabolites,130 with the effects of reducing the discharge pressure (which were to enhance by an order of magnitude the emission from phosphorus, but not that from iodine),‘31 with the substitution of helium for argon as the discharge gas (which enabled atomic lines of sulphur, chlorine and bromine to be observed and used in place of the band spectra previously utilized),132 and its application to pesticide analysis,133-135 and with improved matching of the microwave generator to the discharge cavity72 and applications to determination of methylmercury salts in fish.‘36 Bellet, Westlake and Guntheri3’ also used a similar apparatus for the detection of phosphorus. Moye,‘38 having obtained unsatisfactory results in phosphorus detection with a pure argon discharge, studied the effects of mixing nitrogen, carbon dioxide, oxygen and helium with the argon. Only helium proved satisfactory, the non-monatomic gases giving rise to very strong continuum radiation even in trace quantities, and producing very unstable discharges.

560

S. GREENFELD,H. McD. MCGZACHIN and P. 3. .%ITH

Maximum sensitivity was found with a com~sition of 85% He and 15% Ar. Appli~t~ons to pesticide residues were reported. A series of investigations by Dagnall and West with collaborators began with a study of suiphur compounds i39 in which it was pointed out that the response, and wavelength for maximum emission, were dependent o‘n the identity of the sulphur-containing compound, because of incomplete fragmentation; improvements were obtainedi4* by restricting the volume of the discharge, by using a platinum catalyst anda different~av~ty to increase fragmen~tion, by using all the eluted sample and by stabilizing the discharge pressure. A return to au argon discharge at atmospheric pressure was advocatedi4’ on the grounds of simplicity, wide application and encouraging performance after optimization for routine analysis. The Feasibility of determining i~terelement ratios was established.14” Highly selective and sensitive response in the analysis of volatile metal chelates was reported.‘43 An emlssive helium discharge was used144 for the determination of carbon monoxide, carbon dioxide, nitrous oxide and sulphur dioxide in air, following separation by gas ~~o~~~ography. Luippold and Beauchamp established the feasibility of usmg a low-pressure microwave discharge in helium to find the ratio of deuterium to hydrogen in order to determine the presence and extent of labelling in hydrocarbons separated by gas chromatography. Houpt and Compaan’4” studied the determmation of organic compounds containing phosphorus, sulphur, the halogens or mercury; they developed a method particularly suitable for organomercury compounds in foods, as did Grossman, Eng and Tong.14’ Detection limits for carbon, hydrogen, deuterium. oxygen, nitrogen, Buorme, chlorine, bromine, iodine. suiphur and phosphorus in organic compounds were presented by Lowings:‘QS these were less than 0-i ngjsec. with the exception of oxygen and nitrogen, which were about 3 ng/sec, Almost identical Iimits were reported by McLean, Stanton and Penketh,‘49 who were able to determine atomic ratios and hence empIrical formulae of organic compounds. They stress the importance of mixing small quantities of oxygen or nitrogen with the helium working gas, to prevent the depositlon of carbon on the wall of the plasma tube. Helium was chosen because the higher energy of a helium plasma gave freedom from band spectra. An account of the dete~m~natiou of metal chelates (which slightly antedates reference 143) was given by Kawaguchi, Sakamoto and M~uike.15~ Ail the applications listed above use microwave excitation at a frequency of 2450 MHz. West 151 has compared systems operating at 30 and 2450 MHz. He suggests, rather tentatively, that the 30-MHz discharge has the Following advantages: (I) it may be somewhat more sensitive; (2) a plasma can be formed in any gas at atmospheric pressure; (3) It is probably somewhat cheaper; (4) it is simpler to adjust and operate; (5) it is Iess likely to foul the plasma tube.

He suggests that the 2450-MHz system has these advan~ges: (1) it is ~ommerciaIly available; (2) it is perhaps more readily adapted to gas chromatographic systems; (3) it is more easily thermostated; (4) it does not require extensive electrical shielding. It might be thought that West’s comparison favours a radiofrequency system; nevertheless the literature shows the overwhelming popularity of microwave excitation for gas chromatographic detectors. This is possibly due to the ready availability of low-power microwave generators intended for therapeutic purposes. The working gases used are argon or helium or a mixture. There is a consensus that with the low power used a stable plasma can be formed at atmospheric pressures only in argon, and that, for stability, helium must be used at reduced pressure.

REFERENCES

1. S. Greenfield, H. McD. McGeachin and P. B. Smith, Tulanta, 1975,22.1.

2. J. P. S. Haarsma, G. L. deJong and J. Agterdenbos, Spectrochim. A&z, 1974. 298. 1. t G. Qistescu and R. &igo;ovici, Bufl. Sot. &urn. _. P&V., 194t, 42, 37. 4. Idem, Opt. i S~ectrosko~jya, 1959, 6, 129. 5. A. L. Stolov. Uch. Zap. Ka~a?~sk. Gas Unrv. f956, 116, No. 1, 118. 6‘ E. Badarau, M. Giurga, Gh. Giurgea and A. T. H. Trutia, Spectruchim. Acta. 1957, 11, 44I. 7. V. Kapicka, Spisv ~r~~odo~dec~ Fak. Umv, Brne, 1961, 269. 8. fdem, Cesk.

Casopis

Fys., 1962 12, 505.

9. H. Dunken, W. ‘Mikkeleit and W. Kniesche. Acta Chum. Acad. Sci. Huna.. 1962. 33. 67. 10. W. Tappe and J. v& balke;, Z’ Anal. Chem., 1963, 198, 13.

11. H. Dunken, G. Pforr and W. Mlkkeleit, Z. Chew., 1964,4, 237. 12. V. Trunecek, ibid., 1964.4.3.58. 13. H. Dunken and G. Pforr, Z. Phvsik. Chem. iLe!PztuI, 1965, 230, 48. 14. G. Pforr and K. Langner, Z. Chem., 1965, 5, 115. 15. G. G. Vurek and R. L. Bowman, Scrence, 1965, 149, 448. 16. G. G. Vurek, Anal. Chem., 1967, 39, 1599 17. D. C. Marcusand R. L. Jam&n, rbtd., 1972.44, 1523, 18. 6. D. Cristescu, Arm. Physik, 1960, 6, 153. 19. G. D. Cristescu and M. Giurgea, Opt. i Spektroskoplya, 1961, 11, 424. 20. ldem, Z. Chem., 1967, 7, 360. 21. V. K. Zakharov, Issled. Obl. Khim. Fiz. Metodov Anal. Miner. Syr’ya., 2971, 116. Zheenbaev, lzv. ~~sshIkh Uehebn. 2avedenii. FE.. 1960, No. 3, 103. 23. F. A. KoroIev and Zh. Zheenbaev, Rz. Probl. Spektroskopii, Akud. Nauk SSSR, materials I&A ~oveshch Leni~~ad 1960. 1962. 1, 107. 24. Idem, fzv. Akad.~~auk Kirg. SSR, her. Estestv. I Tekhf~ Nauk, 1964, 6, 33. 25. El. Gostkowska and H. Ekiert, Przeglad. Elektron., 1966, 7, 229. 22. Zh.

26 0. Eriimetti and K. Kukkasjirvl, Aetu Chem. Fenn., 1969, 42, 295. 27. 0. Scholz, Umschau Wss. Techn., 1959, 59, 716. 28. C. Roddy and B. Green, Elecrronics War/d, 1961 29.

(Feb.), 29. R. Mavrodineanu

and R. C. Hughes, S~ectrochrm.

Acta. 1963, 19, 1309

Plasma emission sources in analytical spectroscopy-II 30. C. D. West and D. N. Hume, Anal. Chem., 1964, 36, 412. 31. R. W. April and D. N. Hume, Science, 1970, 170, 849. 32. K. A. Egorova, Zh. Prikl. Spektrosk., 1967, 6, 22. 33. K. A. Egorova and A. 1. Perevertun, lzu. Sib. Otd. Akad. Nauk SSSR, Ser. Khim. Nauk, 1967, (41, 132. 34. K. A. Egorova and V. D. Kiselev, in Spektrosk., Tr. Sib. Soveshch., 4th 1965 ed. N. A. Prilezhaeva, p. 345. Izd, “Nauka”, Moscow, 1969. 35. M. Hoch and H. R. Weisser, Helv. Chim. Acta, 1950, 33, 2128. 36. H. P. Broida and J. W. Moyer. J. Opt. Sot. Am., 1952, 42, 37. 31. 0. Botschkowa, S. Frisch and E. Schreider, Snecrrochim. Acta, 1958, 13, 50. 38. I. G. Grieoriev. N. M. Frolov and M. L. Sanodze. Soobshch.Akad: Nauk Gruz., 1970, 59, 329. 39. J. D. Cobine and D. A. Wilbur, J. Appl. Phys., 1951, 22, 835. 40. W.*Schmidt, E~ektron. Rundsc~u, 1959, 13, 404. 41. U. Jecht and W. Kessler, Z. Anal. Chem., 1963, 198, 27. 42. W. Kessler, Glastech. Ber., 1971, 44, 479. 43. W. Kessler and F. Gebhardt, ibid., 1967, 40, 194. 44. F. Gebhardt and H. Horn. ibid.. 1971. 44. 483. 45. M. Sermin. Analusis, 1973,.2, 186. ’ 46. M. Yamamoto and S. Murayama, Spectrochim. A&z, 1967, UA, 773. 47. S. Murayama, H. Matsuno and M. Yamamoto, ibtd., 1968, 23B, 513. 48. H. Got& K. Hirokawa and M. Suzuki, Z. Anal. Chem., 1967, 225, 130. 49. M. Suzuki, Bunseki Kagaku, 1968, 17, 1529. 50. Idem, ibid.. 1970, 19, 207. 51. Idem, ibtd., 1970, 19, 204. 52. R. Nakashima, S. Sasaki and S. Shibata, Anal. Chim. Acta, 1974, 70, 265. 53. H. P. Brolda and M. W. Chapman. Anal. Chem.. 1958, 30, 2049. 54. M. Yamamoto, German Patent 1,208,100. 5.5. A. J. McCormack, S. C. Tong and W. D. Cooke, Anal. Chem., 1965, 37, 1470. 56. J. H. Runnels and J. H. Gibson, ibid., 1967, 39, 1398. 51. H. E. Taylor. J. H. Gibson and R. K. Skogerboe, g&d., 1970, 42, 1569.

58. F. E. Lichte and R. K.. Skogerboe. ibid., 1972, 44, 1321. 59. Idem, ibrd., 1972, 44, 1480. 60. idem ibid., 1973, 45, 399 61. D N. Hingle, G. F. KirkbrIght and R. M. Bailey, T~~u~tu,1969, 16, 1223. 62. K. Fallgatter, V. Svoboda and J. D. Winefordner, Appl. Spectry., 1971, 25, 347. 63. K. M. Aldous, R. M. Dagnall, B. L. Sharp and T S. West, Anal. Chum. Acta, 1971, 54, 233. 64. H. Kawaguchi, M. Hasegawa and A. Mlzuike, Spectrochim. Acta, 1972, 27B, 205. 65. A. van Sandwgk, P. F. E. van Montfort and J. Agterdenbos, ~~an~~, 1973, 20, 495. 66. A. van SandwiJk and J. Agterdenbos, ibid., 1974, 21, 360. 67. L. G. Bachurina, I. I. Devyatkin, V. M. Perminova and N. 1. Tsemko, Zh. Prikl. Spektrosk., 1971, 15, 401. 48. L. G. Bachurina, I. I. Devyatkin, V. M. Perminova, S. A. Savostin, N. I. Tsemko and L. K. Chuprina, Zuvodsk. Lab., 1973, 39, 225.

69. H. Kawaguchi, M. Hasegawa and A. Mizuike, Bunko Kenkyu, 1972, 21, 36. 70. R. M. Dagnall, J. M. Manfield, M. D. Silvester and T. S. West, Spectrosc. Letters, 1973, 6, 183. 71. F. C. Fehsenfeld, K. M. Evenson and H. P. Broida, Rev. Sci. Inst., 1965, 36, 294.

561

72. C. A. Bathe and D. J. Lisk, Anal. Chem., 1968, 40, 2224.

73. H. Kaiser, Z. Anal. Chem., 1965, 209, 1. 74. R. K. Skogerboe, A. T. Heybey and G. H. MorrIson, Anal. Chem., 1966, 38, 1821. 75. S. Murayama, S~ec~ocbim. Acta., 1970, 25B, 191. 76. fdem, Bunko Kenkyu, 1970, 19, 237. 77. K. Kitagawa and T. Takeuchi, Anal. Chim. Acta, 1972, 60, 309.

78. Idem, ibid., 1973, 67, 453. 79. A. A. Pupyshev and V. N. Muzgin, Zh Analit. Khim., 1973, Zs, 890. 80. L. Mollwo, Ann. Physik, 1958, 2, 97 81. J. van Calker, Colloq. Spec~~sc. Intern. Lyons. lY61, 1962, 2, 48. 82. W. Lochte-Holtgreven, Z. Anal. Chem., 1963, 198, 1. 83. S. Lanz, W. Lochte-Holtgreven and G. Traving, Z. Physik, 1963, 176, 1. 84. U. Jecht and W. Kessler, ibid., 1964, 178, 133. 8.5. G. Pforr and V. Kapicka, Collection Czech. Chem. Commun., 1966, 31, 4710.

86. V. Kapcka, Cesk. Casopis Fys., 1967, 17, 5. 87. K. A. Egorova, Zh. Prikl. Spectrosk., 1967, 6, 288. 88. L. M. Baltin. V. M. Batenin, V. R. Gol’dberg and N. I. Tsemko, Generatory Nlzkotemp. P&my, 7’r. Vses. Nauch.--Tekh. Konf 3rd, 1968, ed. A. V. Lykov, “Energiya”, Moscow, 1969. 89. L. M. Baitin. V. M. Batenin, V. R. Gol’dberg, N. I. Tsemko and L. K. Spuskanyuk, Przmen, P~s~~ronu Spektrosk., Mater, Vses. Simp. 1968, ed. Zh. Zheenbaev, p. 88. “Ilim”, Frunze, USSR, 1970. 90. M. E. Britske and Yu. S. Sukach, ihld., p. 96 91. J. Kapounova. Sb. Ved. Pracl Vysoka Skola Banske Ostrave, Rada Horn.-Geol, 1973, 18, 87. 92. K. Kapoun, Ibtd., 1973, 18, 119. 93. K. W. Busch and T. J. Vickers, Spec~oc~~im. Acta., 1973, 28B, 85. 94. C. S. White, L. C. Watkms and E. E. Fletcher. J Am&on Med., 1956, 27, 414. 95. A. A. Vashman, L. V. Lipis and N. A. Teterina, Opt. i Spektroskopiya. 1959, 6, 260. 96. T. Given, R. J. Magee and C. L. Wilson, Talanta,

1959, 3, 191. 97. C. L Chakrabartl, R J. Magee and C. L. Wilson, ibid., 1962, 9, 43. 98. Idem, Ibid., 1962, 9, 639. 99. L. E. Boos. Jr. and J. D. Winefordner, Anal. Chem., 1972. 44, 1020. 100. N. G. Snopov, Zh. Prlkl. Spektrosk, 1973, 18, 371. 101. M. Serviene. P. G. de Monteareuil and D. Domine. Compt. iend., 19.56,242, 2827. 102. R. Ishida, Repr. Govt. Chem. Ind. Res. inst., Tokyo, 1956, 51, 342. 103. H. Fay. P. H. Mohr and G. A. Cook, Anal. Chem., 1962. 34, 1254. 104. R. I. Gutkma and E. I. Maslennikova, Materuxly Ural’sk. Soveshch, po Spektroskopii. 4th, Sverdlosk, 1963, 98. 105. H E. Taylor, J. H. Gibson and R. K. Skogerbge, Anal. Chem.. 1970, 42, 876. 106. N Penchev, St. Belchev, N. Piperov and G. Belinov, lzu. Ftz. Inst. ANEB, Bufg. Akad. Nauk, 1971, 21, 119.

107. H. Faust, Z. Anal. Chem., 1960, 175, 9. 108. 4. N. Zaidel, G. S. Lazeeva and A. A. Petror, Zh. Prlkl. Spektroskoprr, Akad. Nnuk Balorussk. SSR, 1964, 1, 236. 109. V. M. Nemets and A. A. Petrov. Spekrrosk., 7?. Sib. Soveshch., 4th, 1945, ed. N. A. Prilezhaeva, p. 329. Izd. “Nauka”, Moscow, 1969. 110. K. Sommer and H. Kick, Z. Anal. Chem., 1966,220,21. 111. J. P. Leicknam, H. C. Figdor, E. A. Keroe and A. Muehl, Intern. J. Appf. Raduztlon Isotopes, 1968. 19, 235.

562

S. GREENFIELD, H. McD.

MCGEACHIN and P. B. SMITH

112. .I. P. Leicknam, V. Middelboe and G. Proksch, Anal. Chrm. Acta. 1968, 40, 487. 113. J. A. Goleb and V. Middelboe, ibid., 1968, 43, 229. 114. M. M. Ferrarls and G. Proksch. rbld., 1972, 59, 177. 115. K. Kumazawa, Radioisotooes. 1972. 21. 623. 116. D. R. Keeney and M. J. Tedesco, A&. Chrm. Acta, 1973, 65, 19. 117. V. Middelboe, Appi. Sprccry., 1974, 28, 274. 118 C. P. Lloyd-Jones, G. A. Hudd and D. G. Hill-Cottingham, Analyst, 1974, 99, 580. ii9. H. P. Broida and G. H. Mornan. Anal. Chem., 1952. 24, 799. 120. G. V. Veinberg, A. N. Zaldel and A. A. Petrov, Opt. i Spektroskopiva. 1956. 1, 972. 121. A. N. Zaidei and G. V. Ostrovskaya, Vestn. Lenmgr. Univ. 14, No. 16, Ser. Fz. I K/urn., 1959, 3, 39. 122 V. M. Nemets, A. A. Petrov and B. A. Shubdukarimov, Zk. Prikl. Spektrosk., 1971. 15, 790. 123. A. N. Zaidel and G. V. Ostrovskaya, Opt. I Spektroskoprya, 1960, 9, 142. 124. J P Lelcknam and R. Capitmi, Commis. Energle At., Rappt. CEA-R-2428, 1964, 1. 12.5. R. Capltini, M. Ceccaldi and J. P. Leicknam, Commzs. Energie At., Rappt. CEA-R-3457, 1968. 126. R. Capitmi, M. Ceccaldi. J. P. Leicknam and J. Rabec, Commrs. Energle. At., Rappt CEA-R-3457(2), 1970. 127. A. Karmen and R. L Bowman, Ann. N.Y Acad. Ser., 1959, 72, 714. 128. J. C. Sternberg and R. E. Paulson, J. Chromatog.. 1960, 3, 406. 129. C. A. Bathe and D. J. Lisk, Anal. Chem.. 1965. 37, 1477.

130. 131. 132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142. 143. 144. 145. 146. 147. 148. 149. 150. 151.

Idem, ibid., 1966, 38, 783. Idem, ibId., 1966, 38, 1757. Idem. ibid., 1967, 39, 786. Idem, J. Assoc. Ofic. Anal. Chem.. 1967, SO, 1246. Idem, Biomed. Appl. Gas Chromatog., 1968, 2, 165. ldem, J. Gas Chromatog., 1968, 6, 301. ldem, Anal. Chem., 1971, 43, 950. E. M. Belief W. E Westlake and F. A. Gunther, Bull. Envir. Contam. Tosicol.. 1967, 2. 255. H. A. Moye, Anal. Chem., 1967, 39, 1441. R. M. Dagnall, S. J. Pratt, T. S. West and D R. Deans, Talanta, 1969, 16, 797. Idem, Ibid., 1970, 17, 1009. R. M. Dagnall, T. S. West and P Whitehead, Allal. Chim. Acta, 1972, 60, 25. Idem, Anal. Chem., 1972. 44, 2074. Idem, Analyst, 1973, 98, 647. R. M. Dagnall, D. J. Johnson and T. S. West, Sprcwosc. Lett., 1973, 6, 87. D. A. Luippold and J. L Beauchamp, Anal. Chem.. 1970, 42, 1374. P. M. Houpt and H. Compaan, Analusis, 1972, 1, 37. W. E. L. Grossman, J. Eng and Y. C. Tong, Anal. Chun. Acta, 1972, 60, 447 B. J. Lowings, Analusis, 1972, 1, 510. W. R. McLean, D. L. Stanton and G. E. Penketh. Analyst, 1973, 98, 432. H. Kawaguchl, T. Sakamoto and A. Mlzuike. Talanta. 1973, 20. 32 I. C. D. West, Anal. Chem., 1970, 42, 811.