III. CONCLUSION

A common practice in ICP analysis, and one of its major advantages, is to analyze major, minor, trace, and ultratrace concentration levels simultaneously. Variable matrix samples having multiple major elements present in wide concentration ranges are, for exploration pur- poses, better suited to selective extraction techniques.

The preceding results were obtained without any at- tempt to correct for background or spectral interferences. Collectively, the plasma values are within 15% of the atomic absorption values, exempting sample GXR-3. A majority of optical emission methods developed for min- eral evaluations in exploration geochemistry are of a semiquantitative nature. The accuracy (based on AA results) and precision presented here exceed that which is necessary to discriminate geochemical patterns and trends. For more quantitative analyses, investigation of neutral atom lines, more precise profiling, improvement in nebulization, power regulation, and aerosol flow rates are areas to be optimized to provide even lower detectidn limits, in addition to more precise and accurate analytical results.

Presently, the investigation has been limited to seven elements. However, a new method has been developed utilizing the same solvent which will extract As, Sb, Se, Te, Sn, Hg, T1, Ga, In, Pt, and Pd. 25 When two or more elements are sought, the limited amount of extractant (5 ml) available for analyses places the ICP, with its simul- taneous multielement capabilities, in an indispensable position. In addition, the expectation that hydrogeo- chemistry is becoming an increasingly important area, the same extraction procedure can be adapted as a field method for concentrating trace metals in water samples.

The Branch of Exploration Research of the U.S. Geo- logical Survey is concerned with the development of chemical methods sensitive enough to discriminate not

only the primary but also the secondary dispersion pat- terns of hidden or buried orebodies. The ability of argon plasma to accept organic solvents provides a partial but important answer to some of the analyst's problems. Preconcentration and purging before determination al- low for a greater degree of confidence by minimizing correction factors and elements that tend to introduce adverse effects.

ACKNOWLEDGMENT

The authors wish to thank Dave Cart and Jane Borst of Applied Research Laboratories for contributing their help and knowledge of organic solutions.

1. S. Greenfield, H. McD. McGeachin, and P. B. Smith, Tantala, 22, 1 (1975). 2. G. W. Dickinson and V. A. Fassel, Anal. Chem. 41, 1021 (1969). 3. P. W. J. M. Boumans and F. J. de Boer, Spectrochim. Acta 27B, 391 (1972). 4. V. A. Fassel and R. N. Kniseley, Anal. Chem. 46 lll0A, 1155A, (1974). 5. P. W. J. M. Boumans, ICP Inform. Newslett. 3, 71 (1977). 6. R. H. Scott and M. L. Kokot, Anal. Chim. Acta 75, 257 (1975). 7. H. R. Sobel, R. N. Kniseley, W. L. Sutherland, and V. A. Fassel, ICP Inform.

Newslett. 1, 14 (1975). 8. C. C. Butler, R. N. Kniseley, and V. A. Fassel, Anal. Chem. 47, 825 (1975). 9. W. J. Haas, V. A. Fassel, and R. N. Kniseley, ICP Inform. Newslett. 1, 67

(1975). 10. R. H. Scott and A. Strasheim, Anal. Chim. Acta 76, 71 (1975). 11. A. W. Varnes, M. S. Vigler, and A. Eskamani, ICP Inform. Newslett. 1, 56

(1975). 12. F. N. Abercrombie, ICP Inform. Newslett. 2, 309 (1977). 13. A. F. Ward, ICP Inform. Newslett. 2, 271, {1977). 14. R. L. Dahlquist and J. W. Knoll, ICP Inform. Newslett. 1, 97 (1975). 15. S. S. Karachi and F. L. Corcoran, Appl. Spectrosc. 27, 41 (1973). 16. J. G. Viers, Anal. Chem. 50, 1097 (1978). 17. P. W. J. M. Boumans and F. J. de Boer, Spectrochim. Acta 30B, 309 (1975). 18. S. S. Berman and J. W. McLaren, Appl. Spectrosc. 32, 372 (1978). 19. J. Warren, ICP Inform. Newslett. 2, 262 (1977). 20. A. F. Ward, ICP Inform. Newslett. 1, 266 (1976). 21. G. Horlick, Ind. Res./Dev. 20, 70 (1978). 22. G. H. Allcott and H. W. Lakin, U.S. Geological Survey Open-File Report 78-

163 (1978). 23. J. G. Viets, private communication (1978). 24. F. N. Ward, H. M. Nakagawa, T. F. Harms, and G. H. Van Sickle, U.S.

Geological Survey Bulletin 1289, p. 33 (1969). 25. J. G. Viets and J. R. Clark, "Selective extraction of trace metals from halide

acid solutions using Aliquat 336 in Methyl Isobutyl Ketone," paper presented at annual ACS Meeting in Miami (1978).

A Stable Pulsed Hollow Cathode Power Supply

FREDERIC G. DEWALT, J O H N R. AMEND, and RAY W O O D R I F F Department of Chemistry, Montana State University, Bozeman, Montana 59717

A circuit for a stable pulsed hollow cathode power supply is described. The long term current stability of this power supply is shown to h a v e a r e l a t i ve standard deviation of ±0.0082%. T h i s p o w e r s u p p l y has been used to replace the optical chopper in conventional atomic absorption systems using lock-in amplifier t y p e data acquisition. Index Headings: H o l l o w cathode excitation.

INTRODUCTION

Optical choppers and rotating mirrors have tradition-

Received 20 February 1979; revision received 23 April 1979.

460 Volume 33, Number 5, 1979

ally been used to interrupt hollow cathode light sources in atomic absorption spectroscopy, and to switch between sample and reference light sources in background cor- rected systems. Commercial pulsed hollow cathode power supplies have been utilized for the same purpose. The power supply described in this paper, however, is much less expensive and at least as stable as commer- cially available units. (For example, the 0.0082% stability of this unit compares favorably with the 0.05% stability figure published for the Spectrogram Corporation model LPS-ID hollow cathode power supply.) This power sup- ply consists of an unregulated 400 V dc power supply and a programmable constant current regulator which drives

APPLIED SPECTROSCOPY

a hollow cathode or hydrogen lamp. The lamp is switched from a very low "idle" level to a high emission level by digital logic signals. Both "idle" and "on" levels for each channel are adjustable by front panel potentiometers. The power supply provides a high level of current regu- lation and emission stability (±0.0082%) with large changes (±40 V) in lamp supply voltage. The parts cost of the system is minimal--less than $20.00 plus an un- regulated 400 V dc power supply. It may be used to replace the optical chopper or commercial pulsed hollow cathode power supplies in conventional atomic absorp- tion systems using lock-in amplifier data acquisition.

I. CIRCUIT DESCRIPTION

The basic regulator circuit, sketched in Fig. 1, may be understood by following its operation from the time the hollow cathode lamp high voltage supply is turned on until the lamp current is stabilized. Let us begin by assuming that switch $1, the "lamp on" switch is open. The input of summing amplifier IC~ will then be a posi- tive voltage set by the "idle" potentiometer, producing a negative input to inverting amplifier IC2, and a positive voltage at the non-inverting (+) input of control amplifier ICa (point A). At this time the lamp has not yet fired, no current is flowing through the lamp circuit, and the voltage across the current sensing resistor R ~ e (point B) is zero. With zero input voltage into the inverting (-) input of control amplifier IC~ and a positive voltage applied to its non-inverting (+) input, the output of IC3 will be at its positive limit, about +12 V. Transistor TR~ will thus be gated full "on," and will provide very little resistance to current flow in the hollow cathode lamp circuit.

If the high voltage supplied to the hollow cathode lamp is sufficient to fire the lamp (around 400 V), current will begin to flow in the RL-TRI-tLe~ circuit, and a voltage drop Ese~e equal to

E . . . . = (hollow cathode CUrrent)(P~e~e) (1)

REGULATED ÷12vd¢ +HV

I '°°' ~'IDLE" OLLOW

IOOK 1(~4uf CATHODE 22K IK

~LAMP ON" 100 lOOK 1MEG "1 +1

OOK lOOK IK I 1K HEP 55020 ~ R I ) _L *~1 ~r I I / ~ ' t ~ P , , A LL/:,~ . . . . I "-1 '"

~TL 191 C ¢-PI, B

LAMP

Fro. 1. The circuit diagram for the pulsed hoUow cathode power supply is illustrated in this figure. The regulated 12 V reference supply provides a stable reference voltage for the "idle" and "lamp on" controls. These voltages are summed by IC~ when the "lamp on" switch is closed. The output of IC~ is returned to positive polarity by inverting amplifier IC2, and presented as a reference signal to the non-inverting input of control amplifier ICa. Lamp current is sensed by observing the voltage drop across l:t~.~; this voltage is compared to the programmable reference voltage by control amplifier ICa. The output of this amplifier controls the electron flow through regulator transistor TRy. Lamp current will rise until the IR drop across asense almost exactly equals the reference voltage at point A.

will develop at point B. This positive voltage is applied to the inverting input of control amplifier IC~, thus re- ducing the positive output of this operational amplifier and reducing the drive to regulator transistor TR1. The current through the hollow cathode lamp will increase until the voltage developed at point B (Ese~e) almost exactly equals the control voltage (ERef) at point A. As the voltage developed at point B approaches the refer- ence voltage presented at point A, the output of control amplifier IC3 falls, thus limiting the current through the hollow cathode lamp circuit. If for any reason the current in the lamp circuit should decrease (a decrease in the dc high voltage supply, for example), the difference between Eaef and E . . . . will increase, the output of ICa will become more positive, and the effective resistance of the regula- tor transistor TR1 will decrease causing it to pass more current.

A mathematical description of this circuit is relatively straightforward:

a) The voltage at point B is given by the equation

E . . . . . • (Ilamp) ( a . . . . ) (2)

b) The control amplifier output (IC3) is given by the equation

Econtrol -- (ERef -- E . . . . . )(control amplifier gain) (3)

Control amplifier IC3 is a differential operational ampli- fier operated with a gain on the order of 10 a at low frequencies.

c) The lamp current through the transistor regulator is given by the equation

II,mp (4) = (transistor base current)(transistor current gain)

where

transistor base current (5) Ucontrol

= transistor base resistor Rbase

Combining these equations, we find that

(Econtrol) Ilamp -- - - (transistor current gain); (6)

Rbase

o r

Ilamp (ERef -- E ..... ) (amplifier gain)

Rbase

• (transistor current gain);

(7)

o r

(ERef -- [(II,mp)(R ..... )](amplifier gain) Ilamp = Rbase (8)

• (transistor current gain).

Rearranging this equation, we find that

[ERef - - ( I I . m p ) ( R . . . . . ) ] {9)

(Ilamp) (Rb,~e) (amplifier gain)(transistor current gain)

Substituting values for a lamp current of 10 mA, a sensing

A P P L I E D S P E C T R O S C O P Y 4 6 1

resistor of 100 ~2, a transistor base resistor of 1 K~2 (Rbnse), an amplifier gain of 108, and a typical transistor current gain of 40, we find that

(10 -2 A)(10 a ~2) [Enef - (10 -2 A)(102 ~2)] ffi (103)(40)

ERef -- 1.000 V -- 2.5 X 10 -4 V (10)

ERef = 1.00025 V

Assuming the typical gain figures presented above, the reference voltage must exceed the voltage developed across the current sensing resistor by about 250 #V to maintain a 10 mA current through the hollow cathode lamp. A change in lamp current of 1/zA (1 part in 10,000 or 0.01% for a 10 mA lamp current) will produce a change in control amplifier input of (10 -s A)(102 ~2) = 10 -4 V, resulting in a change in control amplifier output of (10 -4 V)(103 gain) ffi 0.1 V. This ability to make large corrections in the transistor control voltage with minute changes in lamp current causes the circuit to maintain the hollow cathode lamp current at an extremely stable value. Other factors, principally the regulation of the control voltage to the non-inverting input of IC8 and the stability of the current sensing resistor R . . . . . have lim- ited the measured long term stability of this circuit to a relative standard deviation of about _+0.0082%.

Current regulation was measured using the circuit illustrated in Fig. 2. The voltage drop across the sensing resistor R . . . . was monitored by a voltage-to-frequency converter (Analog Devices AD 452J 100 KHz unit, sta- bility equal to a relative standard deviation of about _+0.0042%) and the digital output from this unit accu- mulated in a scaler. The lamp was pulsed by a 100 Hz TTL signal derived from a 5 MHz crystal oscillator and divider string (Fig. 3), and a preset counter gated the scaler "on" for 200 lamp pulses. Results of this experi- ment are presented in Table I.

II. APPLICATIONS

This system has been applied to drive a single hollow cathode lamp in a flameless atomic absorption system

÷HV

( R t

• HOLLOW CATHODE

REGULATOR TOic3 ~ TRANSISTOR

Rsense GATE LAMP ~e~se ~ T ~ i A M M E E ~ S L E GATE CONTROL PROGRAMABLE CLOCK LINE

FIG. 2. Current stability was monitored by the circuit shown in this figure. A voltage-to-frequency converter digitized the voltage drop across the sensing resistor; this digitized signal was accumulated in a scaler. The lamp was pulsed by a 100 Hz crystal-controlled clock; data were collected for 200 lamp cycles.

SMHICRYS~AL -Sq

.o00a2u~ R~SEt

÷I0 ~I

~,747 out SOUARE

FIG. 3. Crystal clock (1 MHz to 1 Hz output capability).

TABLE I. Pe r fo rmance of the pulsed hollow cathode light source, Exper imenta l condit ions:

Hol low-cathode lamp: Cu Operating current : 10 mA Opera t ing frequency: 100 Hz

Sample size: 200 pulses

dc power supply voltage 320 340 360 380 400 No. of measu remen t s 15 15 15 15 15

S igna l (average counts) 46835.7 46838.8 46841.0 46843.9 46846.1 S t a n d a r d deviat ion 2.0702 0,0000 1.1952 1.8898 2.0702

% S t a n d a r d deviat ion 0.0044 0.0000 0.0026 0.0040 0.0044

To ta l count var ia t ion wi th 320-400 V power variat ion:

n = 75 Av = 46841.1

S t a n d a r d deviat ion = 3.8555 % s t anda rd deviat ion = 0.0082% var ia t ion

:,::::T::p,:: , .0. s,O.A, LOC'-'"AMP

% ATOM..,. ..... ].O.O...OMATOR HOLLOW -CATHODE | i

FIG. 4. Elimination of interference from furnace background light. The conventional system had been bothered by light from the furnace passing through the chopper blades, reflecting from the lamp window, and returning to the spectrometer as chopped broad-baud light. The pulsed hollow-cathode system eliminated this problem and resulted in increased stability and more than an order of magnitude reduction in the cost of the light source (power supply, hollow-cathode lamp, and optical chopper system).

using a dc powered hollow cathode lamp, an optical chopper, and a lock-in amplifier. Some trouble had been encountered with this system (Fig. 4), since light from the hot sample atomizer was passing through the chopper blades, reflecting from the window of the hollow cathode lamp, and returning as chopped broad-band light to the spectrometer and data acquisition system. The optical chopper was removed from the system and the regulated dc hollow cathode power supply replaced by the current regulated pulsed hollow cathode power supply. Although light from the atomizer still struck the lamp window and returned to the spectrometer slit, it was not chopped and the lock-in amplifier did not respond to it. This improve-

462 Volume 33, Number 5, 1979

ment was made more attractive by the fact that an expensive regulated hollow cathode power supply was replaced by an inexpensive unregulated dc supply, and the $600 optical chopper was replaced by an inexpensive pulsed current regulator.

ACKNOWLEDGMENTS

The authors would like to acknowledge the support of this work by the Environmental Protection Agency (EPA Grant R805935) and partial support by National Science Foundation Grant CHE-7415060 and DOE Grant ET-78-C-01- 3087.

An Electronic, Adjustable-Waveform Current Generator for Use with a Quarter-wave Resonant Spark Source

TSUTOMU ARAKI and J O H N P. WALTERS* Department of Chemistry, University of Wisconsin, Madison, Wisconsin 53706

A n e w l y des igned current genera tor for use wi th a quarter- w a v e spark source for a f u n d a m e n t a l spec troscop ic s tudy of spark d i s c h a r g e c h e m i s t r y is descr ibed. The generator outputs a t rapezo ida l d i s charge current w a v e f o r m w h o s e durat ion and ampl i tude are adjustable i n d e p e n d e n t l y o f the source repet i t ion r a t e u p to 5 psec and 30 A, respec t ive ly . A multi-~r sec t ion LC l adder n e t w o r k and a f a s t SCR are combined for the n e w gen- e r a t o r to produce the adjustable d i s charge current w a v e f o r m , in p lace of a s ing le RC fi l ter for a convent iona l current g e n e r a - tor . Th e LC n e t w o r k funct ions as a current w i d t h s tre tcher and the SCR s h a p e s the pu l se into the des ired durat ion b y shunt c l ipp ing the tail . To eva luate the per formance o f the n e w c u r r e n t generator , t rans ient e m i s s i o n w a v e f o r m s o f Cu l ines are com- p a r e d for a t rapezo ida l current w a v e f o r m and a s imple r e l a xa - t ion discharge w a v e f o r m wi th that genera ted from the conven- t ional generator . Index Headings: Spark source; E m i s s i o n spec troscopy; Pulse gen- e ra to r .

INTRODUCTION

It is well-known that the use of spark discharge for combined sampling and excitation is highly effective in the spectrochemical analysis of metal alloys. Among the many kinds of spark sources available for such applica- tions, 1-5 the quarter-wave source in recent use here 6 could offer several superior characteristics. However, to realize full advantage in both basic and applied spectroscopy, we felt it necessary to increase the flexibility of the types of current waveforms used in previous experiments. 7' 8 This motivated the design and construction of the new current generator reported here.

Spectroscopic study of spark discharge with sharp- edged, variable-width trapezoidal current pulses can give important information on the behavior of cathodically sampled analyte vapor. The vapor can be observed not only during and after the current pulse, but also, more importantly, at the current pulse boundary. Since this is one time when analytically useful radiation is produced, 9 such study is of direct influence on new methods of

Received 9 March 1979. * Author to whom requests for reprints should be sent.

Volume 33, Number 5, 1979

spectrochemical analysis using positionally stable spark discharges. 1°

The new current generator outputs a trapezoidal cur- rent waveform whose width is adjustable up to 5 #sec independently of the repetition rate and peak pulse cur- rent. The generator is composed of an inductor-capacitor (LC) ladder network and a shunt circuit using a fast silicon-controlled rectifier (SCR). The ladder network functions as a current pulse-width stretcher and the shunt circuit shapes the pulse. These components replace the conventional resistor-capacitor (RC) filter used with- out active components in earlier work. ~8 We report here on the design, construction, and typical performance of the new current generator.

I. QUARTER-WAVE SPARK SOURCE

Fig. 1 shows a schematic diagram of the 323 MHz quarter-wave source used here. It is shown with a con- ventional current generator. This unit is a third genera- tion device ~ developed from an original 162 MHz driven quarter-wave source. The heart of the system is a coaxi- ally resonant cavity whose length corresponds to a quarter-wavelength of the 323 MHz driving signal. This signal is applied as a burst of rf through a sending point near the cavity node.

The applied voltage leads to a progressively building standing wave. Each applied cycle of this wave is re- flected as a mirror image at the open end of the cavity and returns to the sending point. The reflected waves add vectorially to the input waves. In this manner, the voltage at the gap is progressively and rapidly increased to the point where ionization begins at the sharply pointed electrode. The arithmetic details of the electrical process have been presented. 6-8

When ionization starts, the impedance of the receiving end of the cavity begins to fall. This decreases the amount of power reflected, and the rf voltage on the electrodes either stabilizes, if cumulative ionization does not occur, or falls to a low value, if the gap becomes conducting. During the time that full ionization is taking place (ca. 10 to 15 nsec in argon) the source is in a "fragile"

APPLIED SPECTROSCOPY 463