(zl - zZ)/(zz - zs), regardless of whether Reaction 2 takes place or not. This conclusion contrasts with the interpre- tation of the experimental value for the ratio of the height of the oxidation wave from I- to Iz to the height of the suc- cessive oxidation wave from I2 t o 1 0 3 - in 1M HCIOa, as given by Beran and Bruckenstein (5).

We have seen that, if the standard rate constant for the electrode process Equation 1 b is sufficiently high, then the boundary value problem relative to the partial cathodic wave A P +. A S involves exclusively the linear combinations cp and I) and therefore is formally identical both when the solution contains either A1 or Az . Hence, under the present conditions, the second cathodic wave Az +. As obtained from a solution of A 1 of concentration C1* is identical both in shape and in height to the unique cathodic wave given by a solution of Az of concentration C:* = I1C1*/h. If the substance AP is not stable in solution or if its preparation is difficult, we can therefore advantageously study the properties of the A z + A l step starting from a solution of A1 without worrying about kinetic complications due to Reaction 2. The preceding considerations hold even if Equilibrium 2 is

perfectly mobile and may be easily extended to the case in which A1 is electroreduced in more than two steps. An eventual deviation from the preceding theoretical expectations may be attributed to the existence of a notable difference among the diffusion coefficients of the various species. Thus, for example, the two successive reversible steps I- + Iz and IZ --c IC1 of the overall oxidation wave of I- to IC1 in hydro- chloric acid medium on smooth platinum, exhibit different heights (9 , 10) on account of the noticeable difference among the diffusion coefficients of I-, Iz, IzCl-, ICl, and ICIz-.

Institute of Analytical Chemistry ROLANDO GUIDELLI University of Florence Florence, Italy

RECEIVED for review October 19, 1970. 1971.

Accepted June 8,

(9) A. L. Beilby and A. L. Crittenden, J . Phys. Chem., 64, 177

(10) G. Piccardi and R. Guidelli, ibid., 72, 2782 (1968). (1960).

A Digital Potentiostat

SIR: Reported, for the first time, is an instrument for potentiostatic control of an electrochemical cell employing a new technique-digital current feedback. Digital current pulses, which can be scaled to read-out directly on a scaler the weight of sample oxidized or reduced, pass through the auxil- iary electrode of the cell and are used to maintain a control- potential. The potential-control rise-time of this device is limited primarily by the R C characteristics of the cell and, to a minor extent, by the size of the digital current driver and the dead time between pulses. This development entirely obvi- ates the need for elaborate phase-correcting networks pres- ently required to prevent potentiostatic overshoot or oscilla- tion. In addition, current-time information appears at the output of the device in the form of a pulse train and is thus directly computer compatible. The use of a current-to-volt- age converter and a voltage-to-frequency converter required for computer compatibility with conventional analog systems is eliminated. Since the instrument operates simultaneously as a potentiostat, a current-to-frequency converter, and sub- sequently as a digital integrator, a new term-digipotentio- grator, abbreviated DPI-has been coined t o identify it. The title of this article may, however, find more general acceptance. The further advantages of this new design are improved sensitivity and stability, design simplicity, and a large reduction in costs for components and assembly. Ex- tensive use of integrated circuits is made. Also the instrument size and weight are small and the power requirements are quite low.

The arrangement shown in Figure 1 is used to digitally maintain a control-potential and t o digitally measure the charge transfer which occurs in an electrochemical reaction. Two charge sources are connected to the auxiliary electrode. One source injects charge digitally, while the other source, which is a constant current driver, extracts charge on a con- tinuous basis. Both drivers are designed to deliver up to 1 mA of average current and to perform independently of the potential of the auxiliary electrode.

The cell reference electrode is connected to the positive terminal of a differential comparitor through a unity gain impedance coupling amplifier and the desired control-poten- tial is applied to the negative terminal. Any positive output from the differential comparitor-balance detector initiates operation of the digital charge pump for fixed time periods and at rates determined by the charge demands of the cell. A local 500-kHz crystal clock furnishes digital injector drive intervals up to the clock frequency and therefore provides very good digital resolution. Since the charge demand sam- pling is made only during the times between possible injector drive intervals, the occurrence of a partial current drive inter- val is prevented.

The small value capacitor C1 is attached from the reference electrode to the grounded working electrode to maintain potential-control when the cell is driven through its electro- capillary maximum potential.

Figure 2 illustrates the current and voltage response in time to a -500 mV to ground square-wave applied t o the control input of the DPI when employing a capacitive-resistive simu- lated cell load. The digital current feedback pulses, which are initially applied at the clock frequency and thus appear as a blur on the time base selected, drive the control potential from -500 mV to ground. Once ground is reached, only occasional pulses are required to maintain the control-poten- tial. These pulses have a n average frequency of about 4 kHz and a displacement in the control-potential of approximately 10 mV. Note that, in this system, the descent time is limited only by the current capability of the digital charge source and by the dead time between pulses. The use of a IO-mA charge source would, for example, reduce the descent time by a factor of 10.

Figure 3 shows a repeat of the above experiment but now using a modified Matson et ai. cell (1) and employing 0.01M

(1) W. R. Matson, D. K. Roe, and D. E. Carritt, ANAL. CHEM., 37, 1594 (1965).

1718 ANALYTICAL CHEMISTRY, VOL. 43, NO. 12, OCTOBER 1971

d Reference electrode I control potenliol 1 , :g!i+ , pump

i, =Constant current flow

A + Figure 1. Digipotentiogmtor @PI) block diagram

Figure 2. DPI transient response for a dummy cell

Upper trace: Current pulses Lower trace: Control-potential response Time base: 250 Mecldivisian

Figure 3. DPI transient response for an electrochemical cell (1). Parameters the same as for Figure 2

Figure 4. Anodic-stripping voltammogram

SoluUon: 0.01M KCI, 61 ppb. Pb2+ Conditions: -1.000 V os. SCE starting potential. Ine ramp step - 5 mV, ungated analyzer input employed. Dwel 1 sec/channel

KCI as the supporting electrolyte. The potential er is less than that shown in Figure 2 because the squa is now applied against a SCE. Although the descent somewhat longer owing to the solution resistance , dynamically changing cell capacitance, it is adequat for our present purposes. A pulsed current extractioi roughly 8 kHz is required to satisfy the background demand at a control-potential of 0.0 V vs. SCE. ‘I quency is greater than the electron-transfer rate of active substances and is therefore of no consequence.

The current drift of this device amounts to only 0.01 a period of 1.4 hours. The study was made by imp control-potential of - 1.000 V against the simulated cel average count rate was 24,000 for a IO-sec current G

interval. The internal potential control of the DPI i than 1 mV.

remental II time =

rcursion re-wave !time is and the ely fast I rate of current

h i s fre- electro-

I% over losing a 1. The impling s better

ANALYTICAL CHEMISTRY, VOL. 43, NO. 12, OCTOBER 1971 4 b 1719

Figure 4 is an anodic-stripping voltammogram taken with the DPI employing the required peripherial gear described in a previous paper ( 2 ) and using the cell mentioned earlier (1).

The presented instrument is a unipolar device because it was developed solely for anodic-stripping analysis in conjunction with an ongoing environmental study. A more sophisticated bipolar instrument is presently under development and will be reported in the near future. One should note that direct grounding of the working electrode of the cell eliminates bothersome induced currents present in the floating work- ing electrode designs. Location of the drop separation time when applying the future instrument t o polarography will be simplified. A drop separation signal can easily be derived by sensing the abrupt current-pulse spacing changes which occur at the end of the life of a drop.

In conclusion, this approach to direct conversion of charge to a digital number appears to be quite generally applicable. The same principles which resulted in the development of the digital integrator (2) and the digital nuclear spectrometer (3) have now been applied to electrochemistry. If the pulsed

(2) R. G. Clem and W. W. Goldsworthy, ANAL. CHEM., 43, 918

(3) W. W. Goldsworthy, Nucl. Instrum. Methods, 94, 221 (1971). (1971).

current source in the present application were replaced with a broad spectrum pulsed light source and the electrochemical cell was replaced with a phototube, direct digital conversion of spectrometric data for an absorber in the pulsed light path would result.

ACKNOWLEDGMENT

The authors thank Dr. E. H. Huffman and Dr. E. K. Hyde

WILLIAM W. GOLDSWORTHY RAY G. CLEM‘

for their support in this undertaking.

Nuclear Chemistry Division and

University of California Berkeley, Calif. 94720

Lawrence Radiation Laboratory

RECEIVED for review June 1, 1971. Accepted July 13, 1971. Work performed under the auspices of the U. S. Atomic Energy Commission.

1 To whom correspondence should be sent.

Alternate or Simultaneous Electron Impact-Chemical Ionization Mass Spectrometry of Gas Chromatographic

SIR : Gas chromatography-electron impact ionization mass spectrometry (GC-EIMS) is by now a well established analyt- ical technique ( I ) . Gas chromatography-chemical ioniza- tion mass spectrometry (GC-CIMS) is, on the other hand, a recent development (2, 3). In GC-CIMS the carrier gas for gas chromatography on entering the ion source becomes the reactant gas for chemical ionization and the carrier-reactant gas may be introduced directly (2) into the ion source without an interfacing splitter or separator. Methane has thus far been used mostly as the carrier-reactant gas although GC- CIMS is far from being limited to this gas. In general, chemical ionization (CI) ( 4 ) mass spectra have fewer ions and more abundant high mass ions than electron impact ionization (EI) mass spectra although this is not always the case. How- ever, CI mass spectra alone are of limited value in such ap- plication as structure elucidation of organic compounds if they show too few characteristic fragment ions. E1 and CI are in fact complementary and should be used as intimately as possible; furthermore these two modes of ionization are highly compatible (5) and can be used intimately. A mass spectrometer with two ion sources, one E1 and one CI, is de- scribed herein. The two-source instrument allows for the

(1) J. T. Watson, in “Ancillary Techniques of Gas Chrorna- tography,” L. S . Ettre and W. H. McFadden, Ed., Wiley- Interscience, New York, N. Y., 1969, pp 145-225.

(2) G. P. Arsenault, J. J. Dolhun, and K . Biernann, Chem. Commun., 1970, 1542.

( 3 ) D. M. Schoengold and B. Munson, ANAL. CHEM., 42, 1811 (1970).

(4) M. S. B. Munson and F. H. Field, J . Amer. Chem. Soc., 88, 2621 (1966).

(5) G . P. Arsenault, in “Biochemical Applications of Mass Spec- trometry,’’ G. R. Waller, Ed., Wiley-Interscience, New York, N. Y., in press, 1971.

Effluent

first time alternate or simultaneous EI-CI mass Spectrometry of G C effluent.

Figure 1 shows a line drawing of the modified QUAD 300 quadrupole mass spectrometer. The CI and E1 sources are in series and have a ground lens (G) in common. The GC effluent enters the CI chamber (I) which is gas tight and is a t a pressure of ca. 0.5 torr. The G C effluent then exits through the electron entrance and ion exit apertures of the CI chamber (I) into the vacuum envelope surrounding it which is a t ca. 4 X The pressure in the quadrupole mass filter is maintained at 3 X

Alternate E1 and CI mass spectra of G C effluent are obtained by turning the power off

Torr and in which the E1 source is located.

Torr under these conditions.

CI E 1 Source Source Quadrupole -Mass Fil ter

-*r \

;C

Pumping I System

# I

I Pumping I System # 2

Figure 1. Line drawing of m a s spectrometer fitted with two ion sources-not drawn to scale F: filament; R: repeller; D: drawout lens; I: ionization chamber; G: ground lens as well as differentially pumped : awture

1720 ANALYTICAL CHEMISTRY, VOL. 43, NO. 12, OCTOBER 1971