the faradaic responses at a given potential are additive, it is only necessary to measure Ai, in the mixture, a t the predetermined El,%of each of the individual components. Simultaneous equa. tions can then be set up, utilizing the information derived from Ai,,, nieasurements in solutions of the pure constituents.
INCREASING NEGATIVE POTENTIAL SCANNED IN 0.01 I INCREMENTS, STARTING AT ZERO
In
‘1
Pb
ACKNOWLEDGMENT
The authors thank W. A . Higinbotham and Seymour Rankowitz, whose ideas and support have constituted a major contribution to the design and construction of the incremental polarograph.
SCANNED IN 0 0 0 5 1 INCREMENTS. STARTlhG AT - 0 3 0 v
In
1?1
Pb 1
LITERATURE CITED
(1) Barker, G. C., Jenkins, I. L., Analyst
77,685 (1952).
(2) Barker. G. C.. Anal. Chim. Acta 18.
118 (1958).
(3) Barker, G. C., “Advances in Polarography,” p. 144, Pergamon Press, London,
Figure mixture
1960. (4) Breyer, B., Gutman, F., Hacobian, S., Australian J. Sci. Research A3, 558
7. Incremental polarograms of multicomponent 1 X 10-5MCuf2 ‘, 1 X 1 O - W Pb2i 1 X lO-6M TI+ ) (0.1M HCI) 0.9 X 10-6M In+$ 1 X 10-6MCd+? j
iimi.
j - . _ _ , .
(5) Glickstein, J., Rankonitz,lcS., Auerbach, C., Finston, H. L., ildvances in Polarography,’’ p. 183, Pergamon Press, London, 1960. (6) Hamm, R. E., -4NAL. CHE!d. 30, 350 (1958). (7) Kelley, M. T., Fisher, I). J., Cooke, W. D., Jones, H. C., “Advances in Polarography,” p. 158, Pergamon Press, London, 1960. (8) Kolthoff, I. hI., Lingane, J. J., “Polarography,” p. 70, 2nd ed., Interscience, Xew York, 1952. ( 9 ) Ibid., p. 87. (10) Kronenberger, K., Strehlow, H., Elbel, 4. IT.,Pohrographische H e r . 5 , 62 (1957). i l l ) Leveque, M. P., Roth, F.. J . chzm. phys. 46, 180 (1949).
1
Upper curve shows simultaneous recording of current-voltage curve, below base line.
12) Lingane, J. J., FVilliains, It., J . Am. Chem. SOC.74, 790 (1952). : 13) Meit?, L., Polarographic Techniques, Appendix B, Interscience, New York, 1955. (14) Milner, G. W. C., “The Principle,: and Applications of Polarography, p. 68, Longmans Green, London, 1957. I 15) Rankowitz, S., Higinbotham, W.A, Ulickstein, J., Paper presented at the ’
Iriterriational IRE Convention, Sew York, Lu. Y., March 1961. T o be published in the Convention Record. RECEIVED for review April 14, 1961. .iccepted July 10, 1961. Division of
.inalytical Chemistry, 139th Yketing, .ICs, St. Louis, M o . , March 1961. Work performed under the auspices of the c-. J. Atomic Energy Commission.
Stationary Electrode Polarography with a Staircase Voltage Sweep CHARLES K. M A N N Department of Chemistry, Florida Stafe University, Tallahassee, Flu.
F The use of a discontinuous voltage sweep for voltammetric determinations is described. Imposition of a sweep in the form of a voltage-time staircase permits the use of very high rates of polarization, with the result that the sensitivity of the method is increased and the time for a determination is made small. Repetitive measurements with very short intervals of intervening time can be made. Design of the apparatus and operating conditions are described in detail. 1484
D
ANALYTICAL CHEMISTRY
TI
recent research in electroanalytical chemistry has been concerned nith attempt,s to improve 3ensitivity and selectivity. These have generally involved either a n electrolytic concentration procedure follovr-ed by n voltametric determination, or various types of alternating current polarography in which the electrode double layer (EDL) charging current is separated from the faradaic current. One technique, square wave polarography, inr-olvrs nicasurernent’ of current a t a fised timcl lifter t h c change of polarity UCH
1
of tlw square wave ( 1 ) . Because the EDL charging current decays more rapidly than faradaic current, this eliminates charging current from the measurement. The present paper is intended to suggest a somewhat different approach to the same problem of dimination of the effects of charging current. The technique of voltammetry, involving imposition of a linear voltage sweep on a microelectrode, permits an increase in sensitivity as compared nith riolarography When a voltage
ic I
i
SINGLESWEEP
I
I
I
-
I
! Figure 1. tus
I
! Block diagram of appara-
sweepj rather than a constant voltage, is used, the current-voltage curvc that results exhibits a peak, the height of wtiirh is proportiond to concentration of elwtroreactive solute. This peak is not similar to a polarographic masimuni, but is caused by the large rurrcnt which f l o w as tho voltage sweep passes through the value required to initiate the vlectrcide reaction. K i t h :I moving w c c p there is a short period of time during which the applied volt'age is sufficient to sustain the elecatrode reartion, and a n apprcciable i:oncwitration of electroreactivc' solute, is prrscnt in the 1:iyer of solution in intimate contact with the electrodc. K h e n this solutr i p exhausted, thta iwrri,nt must (', .;in(.(, t h r r:it,t, of tht, c.lrr,-
trode reaction is then limited by thG rate a t which more solute can be brought in from the bulk of the solution. The relationship betn-een sensitivity and polarization rate has been derived bsRandles ( I I ) , Sevcik ( I S ) , and Delahay (4). Peak current is proportional to the square root of polarization rate. The enhanced sensitivity obtained by increasing thc polarization rate is limited in a t least tL7-o ways. First,. if tlic electrode reaction is slow, increasing the applied voltage rapidly niny not give a n appreciablr. current. Secondly, the EDL charging current inc>reases with inrreasc in polarization rate. Diff culties stemming from KDL 1:harging curwnt can be minimized b j - using a discaontinuous voltage i;n-c~p in the form of a voltage-time stairimse. The use of a staircase sweep was suggested earlier ( 2 ) . By taking adrantage of the difference in rate of do"ay, one can separate the charging (arrent from the faradaic currrnt. 'This rcasults in considerably increased (wrrent-concentration ratios, as coni1:ared nith polarography or linear s n - c ~ l i voltanimct~ry.
nections in detail, is given in Figure 1. The output of a staircase generator is fed to voltage and power amplifiers. The amplified signal is applied through a measuring resistor t o the electrolysis cell. The measuring resistor I R drop forms the input t o the vertical aniplifirr of a n oscilloscope. The cell voltage sweep is used to drive the horizontal amplifier. The resulting display has the conventional currentvoltage coordinates. Staircase Generator. The stail,case generator was adapted from a circuit given in the General Electric Transistor Manual (6). A detailed schematic diagram appears in the upper left corner of Figurp 2. Transistor &I ociated circuitry comprise an osci1l:itor having an output of rectangulai, voltage pulses. Pulse frerjuon~~y. hcnce t,he step n-idtli of the final \ \ - a x form, is controlled by potc&ometer R9. These pulses apiiear in the output of Q2 as current pulses n-hich are imposed on condenser CI. Transistor Q3 limits the height of the staircase by discharging C1 a t a definite voltage which is regulated by potentiometrr Rlu. The wive form produced has a n amplitude of 12 volts, stixp Ividth variable. from 0.28 to 6.8 insec. Choice of cap:iritor C, affuats thc n u m h r r of
EXPERIMENTAL
Reagents. .I11 reagent? \\-cw of :tnnlytical reagent grade a n d n-erc iisrd u.ithout further treatment. The cadniium(I1) stork solution was standardized by electrodeposition. manyanese(I1) by sodium hismuthate oxidation folloived by titration with iron(II), arid zinc l ~ ythe nicrcuryt,hi o cyan a t e ni e t h o d . Apparatus. X block diagrani of the :ippai,at,iis, slio\vinq t h e external con-
Table
1.
Components for Staircase Sweep Generator
(;,I:;. 254'31 Vnijunction tr:irisiator G.1.:. 2 S 5 2 5 trnnsistor 2-nifd. 1000-volt capacitor. '
wlirre i, and i :ire farziiaic and charging c.urrcnt,s. i se t ~ f the stnirr.i~ to - 1
4 1
Table
Cation Cd +2
Concn., mM 2.00
1.00 0,800 0.400
II.
Current-Concentration Data
Background Electrolyte 1M KCl
0,100
Cd +*
Zn + 2
0,050 0.020 2.00 1.00 0.400 0.050 0,020 2.00 1.OO
1000
1M KC1
0,050
Mn +2
>In
2.00 1.00 0.500 0.200 2.00 1.00 0.500 0.100 2.00 1.0 0.500 0.100
Standard Deviation,
% '
f1.3 1.2 1.0 0.9 4.2 2.4 12.5
1M XaC104
415 453 358 483 330 297 400
537 390
526 565
Polarization Rate, Volts/Sec. 34.0 34.1 34.8 34.1 34.1 34.0
Peak Voltage us.
S.C.E. -0.69
34.0
0.6
35.1
-0.72
1.1 1 .o
36.0
-1.20
35.0
-1.68
1.3 0.8 2.2 3.1
36.5
-1.58
0.4
35.0
-1.94
2.5 1.6 9.4 11.4
0.500 0.100 Zn +*
Peak Current, ga./ Mmole 584 668 696 865 923 927
3.4 4.8 30 3.9 3.3 5.2 7.0
1.0
3.1
3.1
Since the polarization rate can be adjust,ed, it should be possible to increase the selectivity of voltammetry by choosing a rate such that little response is obtained from a substance having a slow t~lectrocie reaction while a large response :s obtained frotn wbstnnces hnving fast reactions. T?:: results report,ed here imlicate ihd: polarization rates fatter than 35 :.o!t.> per second tire necessary to obtain any degree of d e c t i v i t y ivhen the sinipie
p ~ i'his . identity -.i rrsuit pertairis Imth XI rur1.I' shape
3 v . l i to iprecision of !ntsasurPinent of pc.3.k current. hvrordingiy, most of the resuits reportpi1 ilertain to sviutions n-hifnh had not been deaerate(I Repetitive XlecArolyses. K h e n t h r a w r r l r generator i. not opPratir:g Figure 3 s h o w t h a t t,he 2ositivr ,~lectrode Is ahorteiieci t o gro\i:lil t,hroiigh R12 and R,. Since tile nq+t,ive rlwtrode is grounded, it will iw :inudxaily stripped uf any reversi hi) tieposited reactant. it should oiil!. be ncwssary. therefore, to w i t lonc ~ ~ n o i i gfor h the stripping process to 1~ mnpleted to make repeated esposiir.1.: of the same electrode with reprodu\ i h t results. The extremely shorf duration of the period of electrolysis make: possible closely spaced repetitive cs-
'40L.
33, NO. 1 I, OCTOBER 1961
* 1489
Table 111.
Experiments with Large Resistors Polarization Rate, Volts/Sec. Stair‘PJ Linear case Rm, Ma./ i&, sweep sweep Ohms Mmole Mv. 43.7 14.3 7.2
1000 900 800
88= 203 159
56 161 142
Value adjusted to account for difference in electrode areas. 4
posures. Using reduction of CdC12, a series of 40 sweeps was imposed on a single electrode; the signals were observed on a n oscilloscope having a persistent phosphor and spaced so that a new signal appeared before the trace of the preceding one was completely gone. N o change in shape of the response was observed. The precision for six or seven exposures of a single electrode is ordinarily slightly better than that for the same number of exposures of fresh electrodes. Presumably this reflects small deviations in the size of the mercury drops collected for the electrodes. This was checked in detail only for the reduction of cadmium but was observed more casually for reduction of zinc and manganese. This behavior did not occur when surface active materials were present in the solution. Surface Active Compounds. As expected, films of surface active compounds will interfere. The extent of interference will be determined by the concentration of surface active substance, t h e nature of the ion undergoing reduction, the type of supporting electrolyte, and the polarization rate. Very rapid electrode reactions, such as reduction of Cd(II), are less affected than the slower reactions and the presence of a polarizable anion such as chloride minimizes interference by adsorbed films ( 8 ) . To check the effect of surface active compounds, the detergent, Triton X-100 ( R o b & Haas Co.), was used. This compound, like gelatin, is nonionic and therefore adsorbed on both sides of the electrocapillary zero. I n the reduction of CdClt in 1M KC1 with a 7-volts per second sweep, a concentration of approximately 5 X Triton X-100 had no apparent effect. With about 1 X lo+% under the same conditions, a freshly formed electrode shows a normal response. When the electrode stands in contact with the solution for as much as a minute, the response then obtained is very much reduced. If repetitive voltage sweeps are imposed on the electrode, the response increases and becomes normal in size and shape. An electrode which 1490
ANALYTICAL CHEMISTRY
has been through this cycle will show exactly the same behavior if allowed again t o stand in contact with the solution. When a faster voltage sweep is used, the adsorbed film more effectively blocks the electrode reaction. Thus with 35 volts per second, a 2 X concentration prevented the reduction of 1.00mM CdClz in 1M KCl. Imposition of repetitive sweeps had no effect. The presence of an agar salt bridge can cause difficulty. The freshly prepared bridge apparently caused no trouble. After it had been in use a few weeks, however, the behavior described above became noticeable. Replacement with a fiber-type reference electrode eliminated the difficulty. The data in Table I1 pertain to electrolyses carried out in systems to which no organic matter was deliberately added, but from which no unusual effort was made to prevent contamination. Comparison with Linear Sweep Voltammetry. Comparison of different v o l t a m e t r i c and polarographic methods is complicated by the variety of current-measuring devices which are used. To make a valid comparison between staircase and linear sweep voltammetry, electrolyses were performed by both methods, using the apparatus already described, under as nearly identical conditions as possible. The linear sweep experiments can then be compared with results from the literature and with those of the staircase sweep experiments. The data for linear sweep experiments, shown as curve 2 of Figure 4, were obtained using electrodes of a slightly different area from those used to get the data of curve 1. This difference was accounted for in the calculations. In addition, larger measuring resistors were necessary because of the lower sensitivity of the linear sweep method. A sensitivity limit of 5 x ~ o - ~isN given in the literature for linear sweep voltammetry (14). The experiments reported here are comparable. For example, take the fourth point of curve 2, Figure 4, which represents a sensitivity of about 80 pa. per mmole. It would require measurement of lbout 0.05 pa., which is reasonable, to obtain a sensitivity of 5 X IO%f. In view of the fact that the data of curve 2, Figure 4, agree reasonably well with literature descriptions of linear sweep voltammetry, comparison of curves 1 and 2 should allow valid comparison between the two techniques under the conditions of these experiments. On this basis, staircase voltammetry offers a n increase in sensitivity by a factor of 7.5. The actual sensitivity of staircase voltammetry, and that with linear sweep too, though probably to a lesser extent, can be in-
creased considerably by suitable current ainplification and potential control. The limited current response from the linear sweep experiments made it necessary to use larger measuring resistors than need be used with the staircase sweep. Resistance drop, zpRm, due to peak current through the measuring resistor varied from 23 to 54 mv. for curve 1 and from 56 to 260 mv. for curve 2 , Figure 4. Since the larger values of ipR, would affect the linear sweep experiments adversely, data are presented in Table I11 describing experiments in R-hich larger resistors were used with t h t staircase sweep. A staircase sweep of 7.2 volts per second with ipR, of 142 mv. affords a larger sensitivity than a 43.7-volt per second linear s\$eep nith a 55-mv, drop. To estimate the sensitivity of staircase voltammetry with the current amplifier mentioned, the value of 1000 pa. per mmole from Table I1 for 0.020m M CdCl? Rill be taken. -4s an estimate of the value of current which can be accurately measured, one fourth of the highest full scale sensitivity, which amounts to 0.0012 pa. (9), will be used. On this basis, the limiting sensitivity of the method would be 1 X IOp9 X . In evaluating this estimate, it should be realized that a value for the reduction of CdCI2, a very favorable electrode reaction, R as used. On the other hand, a polarization rate of only 35 volts per second was used to obtain it and the increase in sensitivity stemming from the use of a circuit of effectively zero-ohm resistance was not taken into consideration. I t seems probable, therefore, that the actual sensitivity would be limited by the linearity of the background signal.
ACKNOWLEDGMENT
The writer acknowledges the financial assistance given by the Petroleum Research Fund of the American Chemican Society.’ He is indebted to Verne1 C. Champeaux and John Simpson for advice and assistance in the design and construction of the apparatus described.
LITERATURE CITED
(1) Barker, G. C., Anal. Chim. Acta 18, 118 (1958). (2) Barker, G. C., Proc. 2nd International
Congress of Polarography, Cambridge, 1959.
(3) Breyer, B., Revs. Pure and Appl. Chem. 6 , 249 (1956).
(4) ~, Delahav. P.. J. Am. Chem. SOC.75.
1190 (1963). ’ (5) Delahay, P., “New Instrumental Methods in Electrochemistry,” Interscience, New York, 1954. (6) General Electric Transistor Manual,
4th cd., p. 143, Semiconductor Products Department, General Electric Co., Liverpool, N.Y., 1959. (7) Grahame, D. G., Chem. Reus. 41, 441 (1947). (8) Heyrovsk$, J., Dzscussions Faraday SOC.KO. 1, 212 (1947). (9) Kelley, h l . T., Jones, H. C., Fisher, D. J., ANAL.CHEM.31, 1475 (1959). (10) IZolthoff, I. hI., Lingaiie, J. J.,
“Polarography,” 2nd ed., p. 52, Interscience, h’ew York, 1952. i l l ) Randles. J. E. B.. Trans. Faradav Soc. 44.327 f 1948). ’ (12) Ross, J. iV., DeMars, R. D., Shain, I., ANAL.CHEM.28, 1768 (1956). (13) Sevcik, A,, Collection Czechoslw. Chem. Cmms. 13, 349 (1948). (14) Streiili, C. A., Cooke, W. D., ANAL. CHEM.25, 1691 (1943); J . Phys. Chem. 57, 824 (1953). ~
(15) Vielstich, W., Delahay, P., J . A m . Ckem. Soc. 79, 1874 (1957). (16) Willard, H. H., Furman, N. H., Bricker, C. E., “Elements of Quantitative Analysis,” 4th ed., p. 69, Van Nostrand, New York, 1956. RECEIVED for review December 7, 1960. Accepted June 15, 1961. Division of Analytical Chemistry, 140th Meeting, ACS, Chicago, Ill., September 1961.
Determination of Uranium in the Presence of Molybdenum by Controlled-Potential Coulometric Titration H. E. ZITTEL, LOUISE 6. DUNLAP, and P. F. THOMASON Analytical Chemistry Division, Oak Ridge National Laboratory, Oak Ridge, Tenn.
b A coulometric method for the determination of uranium(V1) in the presence of molybdenum has been developed. Uranium(V1) is titrated coulometrically in a solution of sodium tripolyphosphate and sodium sulfate (pH 7.5 to 9.5)a t a potential of - 1.40 volts vs. the S.C.E. Molybdenum(V1) does not interfere when the molybdenum-to-uranium weight ratio does not exceed 1:7. In the range from 1 to 5 mg. of uranium titrated, the error of the method is about * l % . Various possible interferences were studied; most of the interferences can b e eliminated easily. The method is relatively simple and rapid.
T
production of reactor fuels composed of uranium and various amounts of molybdenum has created the need for a method of determining uranium in such fuels. The method must be applicable to highly radioactive materials and must be more accurate than the conventional polarographic and fluorometric methods for uranium in molybdenum-bearing uranium fuels. Following the work of Harris and Kolthoff (5) and of Susic, Gal, and Cuker (I%’), Propst (8) developed a polayographic method for uranium in which the molybdenum intcrference encountered in acid media was prevented by use of a supporting electrolyte of p H 6 that was 0.5M in potassium phthalate and 0.0221 in potassium ascorbate. Propst ( 8 ) reported the precision of the method to be 2.5y0,. Pfibil and Blazek (7) w r e able to estimate uranium in an alkaline carbonate medium. Although the controlled-potential coulometric procedures for uranium in acid media that are discussed by Shults and Thomason (IO) and by Farrar, Thomason, and Kelley (1) are not applicable t o the HE
determination of uranium in the presence of molybdenum, the general method is useful for the analysis of highly radioactive materials and is much more precise than the polarographic method. The coulometric method described herein is precise and eliminates the molybdenum interference by the use of a complexing alkaline supporting electrolyte. The method is based on the facts that sodium tripolyphosphate in an alkaline medium complexes uranium (VI) sufficiently t o prevent its precipitation and that, in such a medium, molybdenum(V1) is not electroactive. Uranium(V1) is coulometrically reduced in this medium at a potential of -1.40 volts us. the S.C.E.The method is accurate to within =t0.5yoat the 5mg. level of uranium. Molybdenum (VI) does not interfere when i t is present in a weight ratio to uranium of not more than 1:7. The relative standard deviation is about 0.5% in the range from 1 t o 5 mg. of uranium. Trace amounts of iron(III), chromium(III), and nickel (11) do not interfere, whereas chromium(VI), copper (11))ruthenium(IV),
and nitrate do interfere. However, these interferences can be eliminated very simply prior to the coulometric titration. EXPERIMENTAL
Instrumentation
and
Apparatus.
ORNL Model Q-2005 electronic controlled-potential coulometric titrator was used throughout this study (6). Potentiometer; Rubicon, 0- to 1.6volt range. pH Meter; Beckman Model G. Polarograph; ORNL Model Q-1673
($1
*
The titration cell assembly is shown in Figure 1. Reagents. Stock solution of suuportini electrolyte was prepared 6 y adding 40 ml. of 0.16M Na5P3010 solution t o 60 ml. of 0.1M NaZS04 solution and adjusting the resulting solution t o p H 10 by the addition of 1M KaOH. Commercial grade NasPsOla was obtained from hlonsanto Chemical CO. Standard stock solution of uranium (VI), approximately 25 mg./ml. was prepared by diqsolving iiational Bureau of Standards U308 in “Os. The nitrate mas destroyed by fuming with STIRRER
HELIUM---
BECKMAN CALOMEL ELECTQODE
D M E
TEFLON C A P 5 0 - m i BEAK W C O R TUBE
K z S 0 4 -AGAR PLUG
Hg 2
\P+ SEALED IN CONTACT TUBE
Figure 1.
Titration cell assembly VOL 33, NO. 1 1 , OCTOBER 1961
1491
