A Preliminary Evaluation of a New Electronic Polarographic Instrument SIR: We have recently developed and constructed an electronic, potentiostatic, 3-electrode polarographic instrument capable of operating as a normal polarograph, a potential sweep (chronoamperometric) polarograph, and as a normal and derivative pulse polarograph. While the 2-electrode pulse polarograph concept, as developed by Barker ( I ) , appears very useful for analytical purposes ( 4 ) , little work has been published utilizing the Barker instrument (which is commercially available from Southern Analytical Instruments Co., Camberly, Surrey, England). We indicate in the present communication some features of interest which we have found in preliminary work with our instrument. Figure 1 shows a derivative pulse polarogram of a 2 X 10-7Alf Cd(I1) solution in 0.1M KSOs. The curve was recorded in less than 10 minutes. The peak current is linear with conto 5 X 10-3.\f centration from 2 X Cd(I1) solutions. With our present in-trument, determination of a species undergoing a reversible 2-electron reduction is possible a t the lo-’ molar level, with the instrument operating in the derivative mode. The resolution shown by the instrument may be seen from Figure 2, where
a derivative pulse polarogram for a solution 5 X lO-‘M in Tl(1) and 1 x 10-4M in Pb(I1) in 0.1M K S 0 3 is shown. The two peaks, which are approximately 60 mv. apart, are resolved. In normal polarographic work, determination of In(II1) in the presence of large amounts of Sn(1V) is not possible without prior chemical separation. Figure 3 shows a derivative pulse polarograph of a solution 1.0 x 10-3.u in Sn(1V) and 5 x ~ o - ~ inM In(II1) in 1.O.V HCI; a calibration curve for In(II1) in a 20-fold excess of Sn(IV) is linear in In(II1) concentration in the ,1f tin. investigated range of ca. The two waves are separated by only 140 mv. The determination of both Sn(1V) and In(II1) under these conditions ran be carried out with no more difficulty than would be found in the usual polarographic technique. One interesting use of the instrument is in obtaining voltammetric curves at solid electrodes. Figures 4 and 5 are photographs of the current-potential curves for l O - 3 M Ag(1) in 0.1M KN03 obtained a t a 0.2-cn1.~ Beckman
1
-300
-400
-500
-800
-100
VOLTS VS SCE
Figure 3. Derivative pulse polarogram for a solution l X l 0-3 M in Sn(lV) and 5 X 1O-jM in In(lll) in 1.OM HCI
platinum button electrode; Figure 4 shows the curve as obtained in the normal pulse mode of the instrument. In the normal pulse mode, pulses of successively increasing amplitude are applied from a potential initially
-O M 3
I
Iu ~ O O I V F -
VOLTS V S S C E
Figure 1. Derivative pulse polarogram of a 2 X lO-’M Cd (11) solution in 0.1 M KNOa
1366
ANALYTICAL CHEMISTRY
F-0lV-l VOLTS VS
SCE
Figure 2. Derivative pulse polarogram for a solution 5 X 1 OP4Min TI(1) and 1 X 10-4M in Pb(ll) in 0.1M KN03 First peak is the lead reduction
VOLTS V S RCE
Figure 4. Normal pulse polarogram for a 1.0 X 10-aM Ag(l) solution in 0.1M KN03 at a platinum button e Iect rod e
anodic to that for silver deposition, returning to this initirtl potential (with subsequent silver stripping) after each pulse. Figure 5 is a derivative pulse polarograph current-potential curve for the same solution. The curves are quite reproducible and not affected by electrode rotation. The shape of the normal and derivative curves is in fair accord with what is theoretically expected, but rarely obcserved, for metal deposition a t solid e1el:trodes ( 3 ) . The limiting current for the normal mode current-voltage curve, in Figure 4, is in reasonable agreement with the current calculated from the Cotrell equation ( 2 ) . The current is measured 40 mseconds after the pulse is applied; this corresponds to the electrolysis time used in the calculations. Greater sensitivity may be obtained by making the current measurement at shortw times. I n the voltage swtlep mode it has proven possible to determine Pb(I1) in the presence of 500 to 1000 fold excess of Cu(I1) by initiating the cathodic sweep on a dropping mercury electrode from a potenial in the limiting current region for the Cu(I1) reduction, the Cu(I1) limiting current being unaffected by application of the sweep. By initiating a n anodic sweep from a
used to determine the Pb, with some slight increase in sensitivity. A detailed description of the instrument, capable of being used in both cathodic and anodic voltage scans in all modes of operation, will be published shortly ( 5 ) and intensive work on the application, performance, and limitations of the instrument is in progress.
LITERATURE CITED
( 1 ) Barker, G.
i VOLTS VS S C E
Figure 5. Derivative pulse polarogram for a 1.0 X 10-3M Ag(l) solution in 0.1M KNO, a t a platinum button electrode
potential cathodic to the Pb(I1) reduction, also on a dropping electrode, the peak current due to oxidation of the P b ( 0 ) from the electrode may also be
C., Gardner, A. W., At. Energy Res. Est. C/R 2297 August (1958); 2. Anal. Chem. 175. 79 (1960). ( 2 ) Delahay, P., “New Instrumental Methods in Electrochemistry,” p. 51, Interscience, New York, 1954. (3) Kolthoff, I. M., Lingane, J. J., “Polarography,” Tol. I, p. 203, Interscience, Kew York, 1952. (4) Schmidt, H., von Stackelburg, M., “Modern Polarographic Methods,” p. 70, Academic Press, S e w York, 1963. (5) Schlein, H., Parry, E., Osteryoung, R., in press.
E. P.PARRY R. A. OSTERYOUNG North American Aviation Science Center Canoga Park, Calif. RECEIVED for review March 16, 1964. Accepted April 9, 1964.
Study of IExperimental Parameters in Atomic Fluorescence FIame Spectrometry SIR: The basis of atomic fluorescence spectrometry was der,cribed by Winefordner and Vickers ( 6 ) , and the application of atomic fluorescence flame spectrometry to the analysis of zinc, cadmium, and mercury was discussed by Winefordner and Stas b ( 6 ) . This new method of spectral analysis was compared with other spectrometric methods in the two previous papers. I n this communication, greatly increased sensitivities of analysis are reported for zinc and mercury anc the atomic fluorescence of thallium i:, presented. The influence of flame type, excitation source type, solvent, sheath gas surrounding the flame gases, and inert gas added to the fuel gas on the atoinic fluorescenceparticularly the limit of detection-of zinc, cadmium, mercury, and thallium have been investigated. INSTRUMENTAL
The same instrumental setup previously described (6, 6) was used for all studies. However, 2,dditional equipment was necessary for part of the studies performed and will be described below. Most studies were performed using
the total consumption atomizer-burner with medium bore capillary ( 5 ) . For several of the studies the influence of a sheath of gas surrounding the flame was studied. For these investigations, the sheathed atomizer-burner previously described by Gilbert ( 2 ) was used. With a three-way valve foreign gasese.g., argon or nitrogen--could be added to the fuel gas (H2 or C2H2). For several studies, a chamber-type atomizer-Bunsen burner assembly was used (3). Several sources of excitation were used in these studies. For zinc and cadmium, Osram lamps with I-inch diameter holes cut in their outer soft glass envelopes to allow unobstructed pasiage of resonance radiation from the qu‘artz inner bulb were used. For mercury, several lamps were uqed in addition to the Hanovia lamp previously described ( 5 ) . A mercury Osram spectral lamp, type Hg-S, operated a t 1.1 amperes, a Philips mercury spectral lamp (The Ealing Corp., Cambridge 38, Mass.) operated a t 0.9 ampere, and a mercury electrodeless discharge tube were used. The mercury electrodeless discharge tube (Ophthos Instrument Co., Rockville, Xld.) was held at one end by a thermometer clamp in
the center of a 2450-mc. resonant cavity (Ophthos Instrument Co.) with a n observation port optically aligned with the flame cell. Power was supplied to the cavity by a coaxial cable from a 100-wat t microwave power generator (Model PGM-10 X 1, The Raytheon Co., Waltham 54, Mass.). All other lamps were placed in appropriate sockets and mounted as previously described ( 5 ) . For thallium, both the Osram spectral lamp operated at 1.0 ampere and the electrodeless discharge tube were used. Electrodeless lamps for gallium, indium, and selenium were also available (Ophthos Instrument Co.). The electrodeless discharge tubes were operated at only a percentage of full power (Hg and T1, 40%; Gal 60%; In, 50%; and Se, 60%). T o avoid overheating of the mercury and gallium lamps and to minimize self-reversal of source lines, these lamps were cooled to some extent by placing a small centrifugal blower below the resonant cavity and allowing a stream of air to circulate within the cavity. When the electrodeless tube or the Hanovia mercury vapor lamp was used, radiation was focused on the flame cell by a quartz lens with 5.0-mm. focal length. For the Osram and Philips VOL. 36, NO. 7, JUNE 1964
1367