depending upon the concentration of the anions present in the base solutions capable of depolarizing the mercury anode. This is a unique case in which the electrode constitutes the reduced form, and hence the apparent halfwave potential, obtained from the reduction wave alone, is not identical with the true half-wave potential of the composite wave. The apparent E l / , plotted us. log [Hg++], in a 2M sodium chloride supporting electrolyte, produced a straight line with a slope of 31 mv., in excellent agreement with a %electron reduction. The apparent E l / , became more negative by decreasing the concentration, an important, factor to be considered in polarographic determinations. Thc galvanometer deflection measurements must be carried out a t a suitable constant voltage corresponding to the upper plateau. The voltage chosen was approximately 0.2 volt more negative than the apparent El/, obtained for low concentration of mercuric ion (5 x 10-2mM) in the various supporting electrolytes (Table 111). Mercuric ion was best determined with a manual polarograph (polarometer), with the galvanometer deflection measured a t the limiting plateau. This deflection was corrected for the blank to obtain the diffusion current and it served for quantitative determinations. The advantages of this method were described previously
(0. Both the accuracy and precision
were tested in a 2M sodium chloride supporting electrolyte (Tables I and 11). The results were also satisfactory when other supporting electrolytes were investigated (Table 111). Both the Dead Sea end brine alone and the Dead Sea end brine saturated with isoamyl alcohol were used to demonstrate the possibility of using the medium in which mercuric ion is being tested as the supporting electrolyte itself. A supporting electrolyte composed of manganese sulfate, ammonium chloride, and ammonium sulfate w a s also tested. This medium proved suitable for the determination of low concentrations of mercuric ion in manganese sulfate solutions. Also, the use of O.BM sodium sulfite alone as a supporting electrolyte was found suitable; it appeared that sodium sulfite not only removed dissolved oxygen, but also formed a stable comples nith mercuric ion, possibly NazHg(SO&, generally increasing the solubility of mercuric ion in certain supporting electrolytcs, such as the ammoniaammonium chloride, and causing a negative shift in the apparent halfwave potential in this medium. The lowest detectable concentration with the polarographic method was 5 X 10-3 mmole of mercuric ion per liter; a t lower concentrations the variations from blanks were insignificant, but the plot of the voltage a t which a constant current is attained-e.g., 0.05 Ma.-vs. concentration (Figure 2) yielded significant differences in the range of
concentrations of 10-3 to 10-2mM. The determination of mercuric ion in this concentration range was thus possible. The mercury pool anode wa5 also suitable for the determination of mercuric ion in the various supporting electrolytes that were used for the saturated calomel electrode. The voltage variance due to concentration changes of mercuric ion was negligible, and the apparent half-wave potential of reduction was more positive in each case. The use of a constant voltage of -0.25 volt was suitable for all the supporting electrolytes involved in this work. ACKNOWLEDGMENT
The author thanks Avrahani Baniel, Israel Mining Industries Laboratories, for permission t o conduct and publish this research, Bruno Paschkes and Karol Jusskiewicz for assistance in some of the experiments, and Jaacov hIashall for valuable criticism. LITERATURE CITED
( I ) Israel,
Yecheekel, Vromen, Avrsham, ANAL.CHEM.31, 1470 (1959). (2) Kolthoff, I. M., Miller, C. S., J. A n .
Chem. SOC.63, 2732 (1941). (3) Meites, Louis, “Polarographic Techniques,” pp. 69-70, Intersciencc, New York, 1955. ( 4 ) Wattera, J. I., Mason, J. G., J. A m . Chem. SOC.78, 285 (1956).
RECEIVED for review December 15, 1958. Accepted June 10, 3959.
Controlled-Pote ntia I and Derivative PoI a rogra ph M. T. KELLEY, H. C. JONES, and D. J. FISHER Analytical Chemistry Division, Oak Ridge National Laboratory, Oak Ridge, Tenn.
b
Controlled-potential
polarography
has been exploited in the versatile controlled-potential and derivative polarograph. An amplifier continuously forces the potential of the polarized electrode with respect to the reference electrode to equal that of the linearly increasing control potential independently of circuit and of cell resistances. Controlled-potential polarography is especially suitable for high resistance media or for high current polarography in aqueous media where polarograms on ordinary instruments suffer distortion, because of iR drop in the electrolyte. The polarograph can be used for the analysis of irreversible and reversible species at yery low concentrations. It is possible to record the instantaneous
currents, the successive peak currents, the successive average currents, or the time derivative of the polarographic wave as a function of the potential of the polarized electrode with respect to the solution. Regular and derivative polarograms are presented that illustrate the advantage of controlledpotential over conventional polarography.
C
polarography, beset by many experimental problems, has led to more refined methods of analysis such as those of controlledpotential and desivative polarography. ONVEKTIONAL
PROBLEMS DUE TO iR LOSSES AND KINETICS
Regular Pdarography. The pres-
ence in a conventional polarographic circuit of the resistances of the solution, dropping mercury electrode capillary, measuring resistor, and salt bridge results in the potential of the polarized electrode with respect t o the solution being different from the potential applied from the polarograph. With aqueous solutions, solution, capillary, and salt bridge resisb ances totaling as high as 1000 ohms can usually be tolerated in regular polarography, because a t a cell current of 10 pa, this would represent an error of only 10 mv. in the potential of the polarized electrode. Square wave polarographs, however, require that t,he total cell, capillary, and salt bridge resistance be maintained a t about 50 ohms or less (6). The required high VOL. 31, NO. 9 , SEPTEMBER 1959
1475
concentration of supporting electrolyte demands the use of very pure supporting electrolyte materials, especially if one wishes to analyze concentrations as low as or loner than 10+M. I n common with sine wave alternating current polarography, the sensitivity of square wave polarography is limited severely, if the electrochemical reaction a t the polarized electrode is not kinetically rapid and chemically reversible. However, in some cases this effect can be utiliied advantageously to eliminate mutual interference between species that reduce reversibly and irreversibly a t or near the same potential. Traces of organic substances can strongly affect squam wave polarograms because of their effects on reaction kinetics and their changing adsorption a t the interface as a function of applied voltage (d, S). In organic systems, cell resistances of the order of 30,000 ohms may be encountered, so that correction must be made for very appreciable iR losses (5, 20). Errors in the potential of the polarized electrode caused by a current measuring resistor in series with the cell are easily eliminated by using a current amplifier so that this voltage drop is cancelled by the feedback system. Thia has also been done in the Leeds & Sorthrup Electro-Chemograph. With aqueous media, the resistance of the salt bridge and barrier between the saturated calomel electrode and the cell often is greater than that of the remainder of the cell. Potential drops arising from the flow of cell current through the high resistance of a commercial pH type saturated calomel electrode together with the possibility of polarizing it,. prevent the use of this very convenient type of electrode in conventional polarographic circuitry. The use of these electrodes would also he of possible advantage in very dilute solutions, because of the relatively small exposed surface area of an asbestos fiber barrier. Several approaches to the problems caused by the iR cell and the salt bridge potential errors in regular polarography have been reported (1, 11, 17, 21). Of these methods, the most interesting is the use, first reported by Piul Arthur, of an X - Y recorder to plot the polarographic current us. the corresponding value of the effective potential 'of the dropping mercury electrode. The method has the serious disadvantage that it observes the effect without automatically eliminating it, as does the controlled-potential and derivative polarograph. Polarograms of rather unusual appearance, which are difficult to interpret, are obtained by the X-Y recorder method. Derivative Polarography. Derivative polarography, in which one measures the derivative of the cell current with respect to the polarized 1476
ANALYTICAL CHEMISTRY
electrode potential as a function of that potential, offers several advantages over regular polarography. For example, this technique should enable the polarographic determination of a species in the presence of another having nearly the same half-wave potential more readily and with greater accuracy than with regular polarographic techniques. The magnitude of the necessary difference in half-wave potentials is influenced by the relative values of diffusion coefficients, n, and concentrations. With conventional polarographs, one can obtain a good derivative polarogram only if the circuit, cell, capillary, and salt bridge resistances arc low (6, 8, 16). The controlled-potential and derivative polarograph can produce excellent derivative polarograms in the presence of very high resistances. In conventional circuits, although the applied voltage be scanned a t a constant rate, the value of dE,,.,/dt, as seen by the polarized electrode, is not constant because of the variable iR losses in the solution, bridge, and measuring resistor. The derivative of an ideal wave made under these conditions is unsymmetrical-Le., the latter portion is "dragged out." The peak height is reduced, because it is dependent upon the value of dE,,,,,/dt as well as upon id. These deleterious effects are eliminated by the potential-control and current amplifiers. If the derivative be taken with an inserted RC network, the output voltage is equal to RC(dE/ dt) (I - e-,'""), where E is directly . the scan rate proportional to ~ D . M . E and is constant. While one needs a large value of RC to obtain a usable output signal magnitude, the other effect of a large RC time constant is to cause the output voltage to lag appreciably behind the desired signal, which is dE/dt. The expected discrimination of derivative polarography between waves having close values of half-wave potential can not be fully realized, because of this lag. The peak broadening effect is minimized by using an analog computer circuit to obtain the derivative polarogram. OTHER POLAROGRAPHIC PROBLEMS
Condenser Current. The linear compensator is very useful for regular polarography with current ranges of 1 pa. and less. In work a t very high sensitivity. where one needs to employ full scale recorder ranges well below 1 pa. to obtain measurable waves, the supporting electrolyte residual current is very far from approximating a reproducible linear function of the applied voltage (14). T.-nder these conditions, it is easier to measure a derivative than a regular polarogram. Square wave polarography, pulse polarography,
differential polarography, and preaccumulative (anodic or cathodic s t r i p ping suspended drop) techniques are other useful approaches to this problem. D.M.E. Current Fluctuations. The use of a quadruple parallel-T RC filter eliminates very wide fluctuations in the values of the derivative due to growth and fall of the D.M.E. drops (8, 9, 12, IS). Heavy damping of other types would severely broaden the derivative peaks so that those from waves of close half-wave potentials would overlap. For regular polarography, the parallel-T RC filter removes these fluctuations and an average currents polarogram may be recorded. The peak follower also eliminates these fluctuations, and with it one may record a maximum currents polarogram having an insignificant amount of damping. Reproducibility of Dfbpping Mercury Electrode. Imperfections in the drop to drop reproducibility of the dropping mercury electrode may be the major variable that defines the least amount of material that may be determined by this polarograph. Further work on this problem is in progress. CONTROLLED-POTENTIAL POLAROGRAPHY
This polarograph uses controlledpotential electrolysis to minimize the iR loss problems. Various workers have used Luggin capillaries for the measurement of interfacial impedance. Booman used controlled-potential electrolysis for coulometric titrations (4). Pecsok and Farmer controlled only the applied voltage and so their polarograph does not correct for salt bridge and solution iRlosses (18). In this polarograph, an operational amplifier is used as a potential control amplifier, so that it continuously controls electronically the potential of the polarized electrode with respect to the reference electrode and forces it to equal the applied control voltage, independently of the values of these resistances. In organic solvents of high resistance, the cell resistance is dependent upon the size of the mercury drop a t each instant during its life and so the actual potential of the dropping mercury electrode fluctuates widely unless it is electronically controlled. It is this variable iR term that renders difficult the interpretation of X-Y recorded polarograms. It also limits the utility of iR compensator schemes that assume a fixed value for the cell resistance. A three-electrode cell is used. The polarographic current does not flow through the reference electrode which, therefore, may conveniently be a commercial asbestos fiber-type saturated calomel electrode, A conventional dropping mercury electrode has been
n
_f--1
OJT SIGNAL BUS
I ’ I
I
1
I ! _
’ I
DIFFERENTIATOR
1 inverting 1
SCALECHANGER
OEAK FOLLOWER DERIVATIVE ANALOG COMPUTER
CELL
0 t o i400”.
G ;VIRTUAL GROUND POINT FLNCTION S W I T C H IN OPERATE PO5 TIO’.
1 . 7 -
1.1..
010-25“.
b o v E INDICATOR R L o A D
-
-
IlOOvJ
I 1-2-4-10
20011
-
-
b 1-2 C 4-10
P A R A L L E L - I , R C FILTER
Figure 1 .
-
RC“ DAM PING
Block diagram of controlled-potential and derivative polarograph \
used as the polarized electrode, but other microelectrodes could be used. The third electrode is a working electrode consisting of a platinum wire dipping directly into the cell solution. It is the anode, if the polarized electrode is the cathode. It is not isolated by a salt bridge and barrier, because the products of electrolysis a t the working electrode have not been observed to reach the diffusion layer about the polarized electrode. The relative position of the three electrodes does nott seem to be critical. The saturated calomel electrode is placed close to the polarized electrode and on the opposite side from the working electrode in order to minimize any effect from potential gradients. There may be a very small, uncorrected iR drop existing mostly a t the interface of the polarized electrode. The resistance of the mercury thread in a standard dropping mercury electrode capillary is about 50 ohms; hence there will be an uncompensated iR loss. By using a specially designed dropping mercury electrode this resistance could be reduced, but the present work has not indicated a need for it. However, the sum of these errors is generally very low and not significant. FEATURES OF CONTROLLED-POTENTIAL AND DERIVATIVE POLAROGRAPH
The controlled-potential and derivative polarograph has the following features: potential-control and current amplifiers which eliminate all sppreciable iR losses, analog computer derivative-taking circuit, parallel-?‘ filter, peak follower, linear compensator, “zero” damping option, very low through very high current ranges, lineoperated potential supplies, operational tests function. magnetic clutch, reversible span drive, commercial asbestos
fiber S.C.E. option, moderate fabrication and maintenancecosts becauseof the use of commercial operational amplifiers, and use of any standard strip chart recorder. Block Diagram. The block diagram of the controlled-potential and derivative polarograph is shown in Figure 1. It is drawn with the “function” switch in the position, which is the one used while polarograms are being made. Electrolysis Cell. The single compartment electrolysis cell, shown schematically in the block diagram, is of conventional shape and size, but has a third working electrode, SO that no current will be passed through the reference electrode, which is utilized only in the potential control system. A sample volume of 20 ml. is used, but can be greatly reduced for special applications. Potential Sources. Five-volt constant potential sources are utilized instead of batteries for the initial and span potential, the linear compensator, and the zero set circuit. I n each of these, a silicon diode is used in a half-m-ave rectifier configuration that supplies an isolated (offground) input voltage to a Texas Instruments Type 652CO silicon voltage reference (Zener) diode which delivers a constant output voltage having excellent regulation and stability (23). Initial Potential. An adjustable initial potential of from 0 to + 3 volts is provided. The approximate value in use is indicated on a panel meter. The initial potential may be precisely adjusted to the desired value by use of the “test 1 pa. per volt” position of the function switch. Polarizing Potential. The span potential is also shown on a panel meter and may be similarly adjusted t o any precise value between 0 and
-1.5 volts. The algebraic sum of the initial potential and the effective portion of the span potential (indicated on a 10-turn precision dial) is the control potential. The span potential is applied t o a Helipot that is driven through a magnetic clutch by a reversible synchronous scan drive motor. A 1 r.p.m. motor is used, although other scanning rates are available with stock motors. The clutch is energized only when the scan motor sa itch is thrown to the “forward” or to the “reverse” position. Khen this switch is in the “off” position, the clutch is not energized so that the operator may very easily return the scan Helipot to its 0% scan position. This is much more convenient than overriding the drag of a mechanical friction clutch. The mechanical breadboard mounting technique that is used for the ganged Helipots, motor and clutch is discussed Table I (Xote 7 ) . The construction notes on the circuit diagram, Figure 2, are listed in Table I. Linear Compensator. The linear compensator utilizes the countercurrent method of compensation proposed by Ilkovic and Semerano for the residual current (10). A t concentrations corresponding to full scale sensitivities of the order of 1 pa., the residual current increases almost linearly over considerable portions of the control potential range. Thus, a current of opposite sign, increasing a t an equal rate, applied to the input of the current amplifier will eliminate most of the effect of this residual current. By adjustment of the potential applied to a Helipot driven by the scan drive motor, this rate may be vaned from 0 to 1.25 ba. per 10 minutes. Other maximum values of linear compensation can be obtained (Note 15) by changing the value of a 4megohm series resistor. A s a result of this compensation. most VOL. 31, NO. 9, SEPTEMBER 1959
1477
Table 1. Fabrication Note 1. All resistors are 0.5 watt, 5% unless specified otherwise. 2. VR1 through VR4 are bvolt Zener (silicon voltage reference) diodes, Texas Instruments 652 CO, or equivalent. 3. A GAP/R universal stabilized amplifier Model USA-3 (USA3-M3 modular packaged) plugs into each Amphenol 26420016s connector. GAP/R components (K2-X, USA-3, Rl O O A ) (George A. Philbrick Researches, Inc., Boston 16, Mass.) are used. 4. Current range and range multiplier switches, measuring and feedback resistors, and linear compensator resistor are to be fully shielded by an iron box. The Amphenol connector for the current amplifier must be mounted on this shielding box. All leads to the current range switch and to the input terminal of the USA-3 amplifiers must be mechanically stable. 6. Sarkes Tarzian type IN1084 silicon rectifiers also designated as M-500 (Sarkes Tarzian, Inc., Rectifier Div., Bloomington, Ind.) are used. 6. Potentiometer, two-section, precision, 10-turn, 10,000-ohm each section, resistance tolerance 3~575, linearity tolerance f 0.1 yo, temperature coefficient of resistance material f130 p.p.m./’ C. Electrical and me4” chanical rotation 3600” - 0” 0.5 watt, -55” to $80” C. Maximum starting torque 5 ounce inch, maximum running tor ue 4 ounce inch. Size 1 l S l e X 113/16 X 4.206 inches hetween mounting surfaces. ’Front bushing 3/8-32 NEF-2A X “/I6 inch long from mounting mrface. Rear bushing -
Notes for Controlled-Potential and Derivative Polarograph
Note
that a test current increasing at the rate of -0.1 &a. per minute produces a derivative output on the 1-pa. current range of Soy0 of full scale. If more derivative sensitivity be desired, it may be obtained by increasing the ratio of the feedback to the input resistor. 14. Resistor values shown in the zero set network result in a shift of +60 to -15 mv. They may be changed if other shift values are required. If more compensation is needed for prior reduction currents, it may be obtained by supplying an adjustable constant current to the input terminal of the current amplifier (’7). 15. Other maximum values of linear compensation may be obtained by changing the value of this resistor. With the 4-megohm value shown, up to 1.25 pa. of linear compensation current is available. 16. Adjustment of humbucker potentiometer. Throw the peak follower, parallel-T filter, and derivative network switches to out. Turn the current range switches to 0.05 pa. and the function switch to cell open. Connect a direct-current oscilloscope to the test jack. Adjust the humbucker potentiometer to minimize zero offset in the current amplifier (position producing a minimum directcurrent output). The offset should be less than 0.1 volt on the 0.05 ua. range and should decrease on the higher current ranges. 17. Stemag resistors (Arnhold CeramicsInc.,iYew York 22, N. Y.) are used.
the resulting position of the recorder pen will correspond to zero polarographic cell current. This position may be shifted, as desired, by use of the zero set control. The use of the cell open function also readily identifies the sign of the cell current being recorded. Test 1 pa./Volt Function. The test 1 pa./volt function has two purposes: precise adjustment of the initial and span potentials, and operational testing of the polarograph apart from the cell. For example, if the “current range” and “range multiplier” switches be positioned a t a full scale current range of 1 pa., a n initial potential increment of 100 mv. will cause a recorder pen movement of 10% of the full scale. If the span potential is exactly 1 volt, the scan motor will advance the pen downscale at the constant rate of 10% of full scale per minute. If the “derivative network” switch be thrown to the rrin”position, a constant pen deflection of 50% of full scale from the derivative zero position, corresponding to this
rate of change of current, will be observed. This derivative value will be downscale for “forward” and upscale for “reverse” scanning directions. In logical fashion, making use of Ohm’s law, all current ranges may be checked for accuracy and operational tests may be made of other components, including the peak follower, linear compensator, and the parallel-?’ RC filter. A procedure for the Oak Ridge National Laboratory Model Q-1988 controlledpotential and derivative polarograph. describes in dctail, operational tests and circuit performance specifications and is available upon request from the authors. Curve Follower. A curve follower may be added t o this polarograph for use a t very high sensitivities ( 1 4 ) . Its output would be inserted in series with the zero set potential. Power Supplies. A Model R-100A GAP/R compound, regulated dual power supply (George A. Philbrick Researches. Inc.) is used t o supply
7.
8. 9.
10.
11.
12.
+
/i *I
of the current range is available for recording the actual wave, because the base line is brought parallel t o the potential axis. Such a parallel wave can be measured with greater precision than a skewed wave. Zero Set. The zero set circuit inserts a voltage adjustable from - 15 t o +60 mv. in series with the recorder and t h e output of the current amplifier (Note 14). Prior t o starting the scan motor, the recorder pen is normally set t o read nearly full scale by means of the zero set control. When the polarogram is made, the negative output from the current amplifier caused by the polarographic reduction current drives the recorder downscale, from right t o left, so that a wave of conventional appearance is obtained. T o record oxidation currents, one would initially set the recorder pen near the left-hand limit so that it could move upscale in recording the positive output of the current amplifier. If the function switch be rotated to “cell open,” 1478
ANALYTICAL CHEMISTRY
Note 32 X inch long from mounting surface. Front shaft extension 0.25 inch diameter X 13/ls inches long from mounting surface. Rear shaft 0.25 inch diameter 7/8 inch long from mounting surface. Helipot Model A. Motor, magnetic clutch, and ganged Helipots are connected with solid couplings and mounted in line to a base plate by hangers, basic type A-103 (Servomechanisms, Inc., Westbury, N. Y.). Magnetic clutch is Model C-125-1, 28V, (Autronics, Inc., Florissant, Mo.). The 1 r.p.m. synchronous motor, 115 volts, 60 cycle per second, is art 33A 613-3600 (Globe Indktries, Inc., Dayton, Ohio). After fabrication, all GAP/R K2-X amplifiers must be checked for zero offset. If more than 0.75 volt of offset exists, the input bias must be adjusted according to GAP/R instructions (19). Stabelex “D” capacitors (Industrial Condenser Corp., Chicago 18, Ill.) are used. Minneapolis Honeywell Brown Electronik Model Y153X12 single point, strip chart recorder or equivalent with standard (low impedance) circuit, Manual standardizing, 1 inch per minute chart speed, 4.5 seconds full scale pen speed, 10 mv. range, chart paper No. 58SdN. The ratio of these resistors is chosen so that the derivative peak height as seen with a particular current range is equal to about 4n times the height of the regular polarogram for reversible waves. The value of the S7K resistor is trimmed so
13.
+
and -300 volts to the operational amplifiers (19). A filament transformer having eight 6.3-volt secondaries is used for the tube heaters and for the power supplies for the silicon voltage reference diodes. An isolation transformer serves the bridge rectifier power supply which energizes the m a g n e h clutch. Potential-Control Amplifier. The potential-control amplifier is ~b GAP/R model USA-3 universal, stabilized amplifier (19). This plug-in, chopper stabilized, printed circuit ampkifier is inexpensive and reliable, has negligible distortion, noise, and drift, and is ideally suited for this instrumental application. The amplifier, t h a u g h negative feedback, mainhius its input terminal a t ground potential- (virtual ground, G). The input signal consists of the algebraic sum o&thecontrol potential and the potential of the polarized electrode with respect to that of the solution, as seen through the reference electrode. The amplifier can deliver an output voltage of up to 115 volts, or -, a t about 10 ma. The gain of the amplifier is a t least lo7 (directcurrent open loop) and drops to unity a t 1 megacycle, so that it can respond to a rapidly changing signal. Its output is connected to the platinum wire working electrode. The feedback loop is closed through the solution, so that the amplifier continuously delivers a current that forces the potential of the polarized electrode with respect to the reference electrode to equal the control potential. A “potential-control overload” limit indicator, located on the control panel, lights up if the output of the potential-control amplifier reaches 10, -0 volts. This happens 100 rarely, because a current of 5 pa. could be forced by a potential difference of 100 volts through a cell resistance of 20,000,000 ohms. With this polarograph, one has the combined advantages of conventional polarography, selective controlled-potential electrolysis, and “zero cell and circuit resistance” regular and derimtive polarography. Current Amplifier. The current amplifier is also a GAP/R Model USA-3 amplifier. There is no error in the polarized electrode potential due to a current measuring resistor iR drop, because the amplifier feedback loop prevents i t from being seen by the cell. The polarized electrode potential is held a t virtual ground G. The feedback configuration consists of a measuring, a feedback, and an output resistor. Various feedback ratios and values of the measuring resistor are used to give 22 full scale current ranges from 0 to: 0.005, 0.05, 0.1, 0.2, 0.3, 0 5 , 0.75, 1, 1.5, 2 , 3, 5 , 7.5, 10, 15, 20, 30, 50, 100, 200, 300, and 500 p a . I n contrast, the lowest full scale current range of most commercial polarographs
+
+
Table
II.
Polarographic Data
10d4M Cadmium, Regular
E,,*, Volts
Polarograms, Filtering “Undamped’’4 Peak follower Quadruple parallel-?’
us.
S.C.E.
-0.645 -0. G47 -0.1367
Derivative Polarograms, Filtering Peak follower4 Quadruple parallel-T 1O-aM Cadmium, Regular Polarograms, Filtering “Undamped”” Peak follower Quadruple parallel-T
Peak Voltage Volts, us. S.C.E. -0.655 -0.673
Full Peak Width at Half-Peak Height, mv. 47 58
-0.645 -0.647 -0,667
Derivative Polarograms, Peak Voltage Volts, Filtering vs. S.C.E. Peak follower“ -0.655 Quadruple parallel-T -0.674 Measurements made at mid-points of recorded oscillations.
is about 1 pa., although the Radiometer Model PO4 polarograph has a maximum sensitivity of 0.02 p a . The output of the current amplifier corresponding to a full scale recorder pen deflection is -25 volts. A “current-amplifier overload” indicator, which is mounted on the control panel, lights up whenever 10, the amplifier output reaches 100 -0 volts. The zero set circuit provides the necessary recorder scale shift to compensate for cell currents as great as three times the full scale value of the current range in use, such as those due to residual currents or diffusion currents of ions having lesser half-wave potentials. It is possible (Note 14) to provide a greater amount of suppression. The range of the recorder is 10 mv. Types of Polarograms Recorded. Three units, the peak follower, parallel-T RC filter, and derivative analog computer, may be inserted into the signal bus a t the throw of the corresponding toggle s\vitches. By means of these constituent circuits, one may record, instead of a regular damped or undamped polarogram, a maximum currents polarogram with essentially zero damping, an average currents polarogram, or a derivative polarogram, all made with selective controlled-potential electrolysis. Peak Follower for Maximum Current Polarograms. A peak follower, functional operator circuit has been designed that uses a nonlinear circuit element and a voltage follower operational amplifier (19). This is an improved version of the diode filter built for a polarograph designated model Q-1673 ( 7 , I S ) . A diode is connected so that a capacitor may be readily charged, but not discharged by the negative going output of the current amplifier. The output of the peak follower consists of the peak values of the oscillations of the input signal. The recorded polarographic wave is not distorted by
+
Full Peak Width at Half-Peak Height, mv. 47 60
damping; therefore selectivity is maintained and peak currents reached during the lives of successive drops of the dropping mercury electrode are recorded as a function of the true controlledpotential of the dropping mercury electrode. The current oscillations due to the growth and the fall of drops of the dropping mercury electrode are thus eliminated without introducing distortion due to damping. These maximum values are of theoretical significance, and a slow speed recorder is able to keep up with the actual peak values envelope. A GAP/R Model K2-X operational amplifier is wired as a voltage follower (19). It has an input impedance of over 100 megohms and an output impedance of less than 1 ohm, so that the capacitor is not loaded by the recorder and the signal is returned to the bus from a low impedance source. Because the signal level in the bus is high, bus wiring is not critical and it is not necessary to stabilize the KZ-X amplifier with a K2-P chopper amplifier. The capacitor memory can be erased by momentarily depressing a “diode discharge” switch located on the control panel. Within the precision of the experimental measurement, the measured half-wave potential of a maximum currents polarographic wave does not differ from that of an undamped and unfiltered wave (Table 11). The peak follower is of value in recording derivative waves, although its output contains discontinuities while the current flowing through a drop builds up to the maximum value reached with the preceding drop. Application of the peak follower is limited to situations where the maximum polarographic currents are increasing or are constant as a function of the potential of the polarized electrode. The diode is connected in such manner that it will conduct and charge the capacitor only for a negative VOL. 31, NO. 9, SEPTEMBER 1959
1479
1480
0
ANALYTICAL CHEMISTRY
going signal from the current amplifieri.e., only for reduction current in the cell. A diode reversing switch is not provided. At high sensitivity such conditions n-ill hold only over limited potential spans. The linear compensator is sometimes of value to create a suitable input signal for the peak follower. -2 preliminary polarogram, made without the peak follou-er, will establish vhether it may be used legitimately. Parallel-T RC Filter for Average Current Polarograms. The low-pass parallel-T RC filter, described bv Kelley and Fisher (12), may be regarded as a means of electronically averaging the polarographic current during the life of each drop. The oscillations due to the growth and fall of the drops are removed from the record with but small distortion of the polarographic wave form. Because the fundamental null frequency is cycle per second, drop times must be shorter than 5 seconds, or another filter section tuned to reject 1/2.5 cJ-cle per second may be inserted ahead of the four filter sections noF provided. This filter is particularly helpful in sifting out the polarographic information contained in the cell currents in a form suitable for derivative taking purposes. I n this polarograph, by means of a “filter sections” switch, one may choose any of three combinations of parallel-T RC networks that correspond to 1and 2, 4,and 10, or 1,2, 4,and 10 times the null frequency of the network of lowest null frequency. Thus, in recording average currents polarograms, one may use the least possible amount of filtering to eliminate the dropping mercury electrode current fluctuations. In this polarograph, the parallel-T RC filter is given its optimum impedance environment. The input terminal of the filter is driven by a low impedance source. The output is terminated by a 2-second RC filter and is connected to a GAP/R Model K2-X operational amplifier that is used as a voltage follower so t,hat the filter looks into a very high impedance and the filtered signal is returned to the bus from a very low impedance source. This filter has been applied by Hamm in his square wave polarograph (8). RC Damping. A relatively small amount of conventional RC damping is available. 4 10-turn 25,000-ohm potentiometer and a 10-microfarad capacitor are used so that the maximum damping time constant when the parallel-T filter is not in use is only 0.25 second. The output of the quadruple parallel-T RC filter contains the highest frequency components of the D.M.E. current fluctuations, which are removed by the terminating RC damping network. There is also an RC filter inserted in the peak- follower. The time con-
stants are low. While transients are filtered out, the polarographic wave is not appreciably damped by these networks. Analog Computer for Derivative By means of the Polarograms. analog derivative computer, one may record the time derivative of the polarographic wave with respect t o the polarized electrode potential as a function of that potential. The maximum value of the derivative peak, for .a kinetically rapid and electrochemically reversible electrochemical reaction conforming t o the HeyrovskyIlkoviE equation, is given by:
This value occurs when E = In general the peak value will be a linear function of concentration. For reactions conforming to the simple model, corresponding to the Heyrovsky-IlkoviE equation, the width of the derivative polarographic peak a t the half-peak position is 90.6/n mv. From this relationship, one can determine experimental values of n, test the wave for “reversibility,” and evaluate the derivative taking performance of the polarograph with established redox reactions. Derivative polarography is of interest, because it should be more useful than regular polarography a t very low concentrations and when one needs t o resolve waves separated by less than 200 mv. The derivative computer consists of a GAP/R Model K2-X operational amplifier, wired as an inverting differentiator, followed by another K2-X amplifier wired as an inverting, proportioning element (19). The differentiator time constants are so chosen that the differentiator output is the time derivative of the polarographic wave but the differentiator has a limited high frequency response (19). The proportioning element, or scale changer, has two purposes: The derivative signal is reinverted so that it has the same polarity a t the recorder as the regular signal; and for the convenience of the operator, the ratio of the feedback to the input resistor is chosen so that the derivative peak height of the reversible waves is equal to about 4% times the regular height, as seen with the same current range. If more derivative sensitivity be desired, it may be obtained by faster scanning rates, by use of a lower current range, or by increasing the rzltio of these resistors (Note 13). Should the current amplifier limit, the value of the derivative will return to zero. This eve1-t will be signaled by the lighting of the (‘current amplifier overload” indicator and the corrective step is to use a higher current range. Potentiometric Chart Recorder. Any 10-mv. recorder may be used, including those having an input impedance as low as 300 ohms. The
recommended iecorder ( S o t e 12) is a Minneapolis Honeywell Brown Electronik single point strip chart recorder (Type 153) having manual standardization, a 1-inch per minute chart speed, and a 4.5-second full scale (162 r.p.m.) pen speed, The recorder is not modified, which keeps the cost low and simplifies maintenance. The signal on the signal bus is divided by a resistive network so that -25 volts on the signal bus delivers - 10 mv. to the recorder. Circuit Diagram. The circuit diagram is shown in Figure 2. The circuits of the GAP/R components are supplied with them and are available from the manufacturer (19). The use of these excellent, reliable, inexpensive components greatly reduces the fabrication cost of the polarograph and simplifies maintenance. The procedure referred to is intended also for use in the routine maintenance of this instrument. Signal Bus. Figure 3 shows the layout of the polarograph. The functional arrangement of the toggle switches that insert various operational components such as the peak follower is based on the signal bus concept. The operator may see a t a glance exactly which functions are in use. ELECTRICAL TESTS OF PERFORMANCE
After the prototype polarograph had been fabricated, tests were made to check the electrical operation of its circuits. Satisfactory performance mas observed. The potentials of the two virtual ground points, G, were measured and found to be and to remain a t ground potential, including a t the instant of detachment of each D.M.E. drop. With a direct current oscilloscope attached to the output of the polarograph, no tendency could be detected for the current amplifier, potential control amplifier, derivative taker, parallel-?’ filter, or peak follower to be unstable a t the instant of drop detachment. The effects seen on the oscillographic display of the wave forms of the individual drops caused by each parallel-T RC filter option, RC damping, the peak follower, and the derivative computer are fasrinating. Thus, it was shown that the peak follower output does indeed lie at the maximum values of the currents. TYPES O F FILTERING AVAILABLE
The effects of the various kinds of filters on the chart recorded wave form of the individual drop currents are illustrated in Figure 4. The relative vertical positions and magnitudes of the zero cell current and each of the cell currents epochs are maintained in the figure. The available amount of conventional RC damping (0.25 second) is seen to have a very slight effect by comparing the appearance of this epoch with that of the VOL. 31, NO. 9, SEPTEMBER 1959
1481
unfiltered epoch. It is sufficient, however, to remove the vertical “pips” seen on the output of the quadruple parallel-T RC filter when RC is zero. (This epoch was recorded without the Z-second terminating filter.) Comparison of the two parallel-T RC epochs shows that the 1 and 2 sections eliminate the lower fpequency components of the D.M.E. fluctuations and that all four sections remove all but the highest harmonics, Even with RC = 0, the apparent maximum values, as seen on a 4.5-seconds re-
corder, are not the true maximum currents, which are recorded when the peak follower is used. WhemRC = 0, the response of the recorder is the limiting response and one records “zero” damped polarograms. For regular polarography the output of the peak follower does not require RC damping.
regular and derivative polarograms of 9 y of cadmium per ml. (10-4M) in 0.1N hydrochloric acid. Bath the “undamped” polarogram and the maximum currents polarogram are recorded with zero damping. Consequently, the observed values of E,,%are essentially the same (see Table 11). Furthermore, the latter polarogram is a record of actual successive peak currents, unaffected by the mechanical damping conventionally associated with chart recorded polarograms. Not only is the wave height of a maximum-currents polarogram easily measured and theoretically interpreted, but the wave form is not distorted. The selectivity would be expected to be better than that obtainable by conventions1 .~~ ~. .~~... polarography, but ‘this advani,age has not yet been quantitatively ev aluated. The average currents polarograirn represents the results obtained by the electronic averaging action of the quadruple parallel-T RC filter. There is some time lag in the parallel-T RC filter resulting in an error in the observed E,,, value (Table 11). This error may be eliminated for reversible waves, by taking the average of two values obtained by measuring polarograms made in the forward and reverse directions of scanning(l.9). T h i s t i e l a g i s of no practical disadvantage in regular or derivative concentration measurements because, for a fixed amount of parallel-T RC filtering, the observed id and peak values are linearily related to concentration (1.9). The corresponding derivative polarogram is also shown. This derivative polarogram was made Kith the ratio of the resistors in the inverting derivative amplifier equal to 0.25 of the ratio shown in the circuit diagram (Figure 2, Table I, Note 13). The increased seusitivity is attained with no increase in recorded noise and with equally good Pleak form. It is seen that the derivative peak is sharp and has good form.
TYPICAL POLAROGRAMS
The four kinds of polarograms that may be recorded with this polarograph are shown in Figure 5, which consists of
Figure 3. Frontview
ofORNLcontrolledpotential and depolarorivative graph
I LLLL
RCiO
RENT--
Rc:+iec
:Em APING’’
1-2 11-T.RC FILTER; RC=O
Ouod.il-T,RC FILTER; RCiO
PEAK FOLLOWER; RCiO
CONDITIONS : ’3pg.Cd“lml.in 0.4 N HCI D.M.E. et -0.3v.us.S.C.E.
CONTROLLED-POTENTIAL AND DERIVATIVE POLaIROGRAIPH
+.
Electronic recording of average and maximum values of current during the life of each successive drop of I
nyw=
D.M.E.
alp0 I
I ZERO
-0.3 -0.5 -0.7
-0.9
ED,M,E,.voils, v%S.C.E. UNDAMPED POLAROGRAM RC;O
Figure
u -0.3 -0.5 -0.7 -0.3
1
ED,M,E,. Volts. v%.S.C.E. AVERAGE CURRENTS WLAROGRAM RC=O Quod. 11-T. RC Filter
8
1
1
-0.3 -0.5 Yolie
-
1
,
1
u -0.9
-0.3 -0.5
-0.7
E POLAROGRAM I
MAXIMUM CURRENl RC=O Peak Follower
4 see -T. RC Filter 11ower I F = 0.1 “‘/mi”.
5. Regular and derivative polarograms made with controlled-potential and derivative polarogroph 9 Y Cd++/ml. in 0.1N hydrochloric acid
1482
.’I DIODE START DISCHAF OF %A,.
CELL CURRENT--
ANALYTICAL CHEMISTRY
CCNTROLLEC-POTEhTlAL AND DERiVAT Vi POLAROGRAPH iE'ect,on,co!ly
TO describe the performance of the polarograph in greater detail, the data shown in Table I1 are reported. X Sargent 2-to &second capillary was used. No maximum suppressor was added, The temperature was 24" C. The supporting electrolyte was 1N hydrochloric acid. The value of Eliz for cadmium in 1M hydrochloric acid has been reported to be -0.64 volt vs. S.C.E. a t 25" C. (16). The theoretical value for the peak width a t half peak height is 45.3 mv. The value of the derivative peak height with the parallel-T filter for millimolar cadmium is about 50% of full scale on the 100-pa. current range. This sensitivity is obtained with the derivative circuit wired as shown in Figure 2. The range of replicate measured halfwave potentials is less than 5 mv.
ORNL MODEL 0-16'3 POLAROGRAPH (Not Corrected tor ,R D r o p )
Corrected for ,R D r c p )
R: 0
I
J
OF SCAN
'START
-
TEST OFiCONTROL AMPLIFIER BY INSERTING SERIES RESISTORS
- S T A R T SF S C A N L - S T A R T OF SCAN
L ! , I , -07
-05
-07
'
-09
1 -
The gratifying success of the controlled-potential polarographic technique is clearly shown in Figures 6 t o 9. In Figure 6, resistors of 0, 510,000, and 20,000,0(30 ohms are inserted in series ivith the @arographic cell. The predictable effect af %e corresponding dR loss is seen with theModelQ-1673 polarograph which does not have controlledpotential circuitry. On the other hand, the three polarograms made with the controlled-potential and derivative polarograph are identical and may be superimposed throughout the waves.
t.3
I.!
E,ME.~d:s'~~.S.CE.
I
R 2: 300 x 3 1 i
cw3 -5-bfiT
OC
SCAN
\el's i c , S t E
Figure 6. E f h c t of iR drop due to resistor inserted in d e s with.polaragraph cetl
TYPICAL POLAROGRAMS IN H I G H RESISTANCE MEDIA
9 y Cd++,/ml. in 0.1N hydrochloric acid
ORNL Model Q - 1 6 7 3 polarogroph (not corrected for i R c e l l drop) Damping Regular RC = 112 set. Derivative RC = 1 / 2 sec. Quad. 11-T, RC; diode Controlled-potential and derivative polarograph (corrected electronically for iRe,li drop) Damping Regular RC = 0 Derivative RC = 1/4 sec. Quad. 11-T, RC; peak follower Conditions: 0.05M LiCIOd Roe11
-
Regular and derivative polarograms of
r
lO+M cadmium in 1-propanol with
LSTART OF SCAFI
O F SCAN
&TART T
8000 0
Scan rate = 0.1 voIt/rninute
LSTART OF SCAN
. 4
-02
-04
-06 -08 ED E v o l t s
,,
L -02
apparent, vs.Aq
d -04
sc
-
U
-06 E
-08
Figure 7. Effect of iRcel1 drop regular and derivative polarogram of 1 O-3M Cd++ in 1 -propanol
0.05M lithium perchlorate and of 4.4 X 10-~iM2-nitropropane in glycerol with 0.3X lithium chloride are shown, respectively, in Figures 7 and 8. The cell resistances were about 8000 and 35,000 ohms, respectively. The reduction of 2-nitropropane is highly irreversible. The improvement in the regular polarograms resulting from the controlledpotential technique is not immediately obvious until one measures Elizvalues or plots log i/(&-i) US. ED M E, The improvement in derivative polarography resulting from the controlled-potential electrolysis technique is, however, apparent and striking. By this technique, the improvement in derivative polarograms of, for example, 90 y cadmium per ml. in N hydrochloric acid is also noticeable. A typical comparison between the semiloq plots made from conventional and from controlled-potential regular polarograms is shown in Figure 9. The data in Figure 9 were taken from polaroM acetate in glacial grams of ~ o - ~ zinc acetic acid with 0.25M ammonium acetate as supporting electrolyte. The cell resistance was about 30 000 ohms miniVOL. 31, NO. 9 SEPTEMBER 1959
* 1483
niuin (maximum drop size), and was much greater when each dropping mercury electrode drop was small. By measuring the potential of the platinum wire working electrode with a vacuum tube voltmeter, it was seen that the potential-control amplifier typically pumped the voltage applied to the working electrode up and d o n during the lives of each of the successive drops, because the size of the drop has a large and direct effect upon the instantaneous value of the cell resistance. Yet, the control potential was a t all times applied to the dropping mercury electrode with respect to the saturated calomel electrode.
/I (from plot log
&
VI
ORNL Model Q - I 6 7 3 polaro-
ED E )
DIODE DISCHbR&E START OF SCAN
RATIO OF AVERAGE TO INSTANTANEOUS VALUES OF DIFFUSION CURRENT
The IlkoviE equation predicts that a t 25" C. with the dropping mercury electrode the ratio of the average to the maximum values of the diffusion controlled current is 0.859. Taylor et al. have reported oscillographic measurements at 25" C. on millimolar cadmium solutions (22). K i t h this polarograph, quadruple parallel-T and peak follower regular polarograms of millimolar cadmium in IN hydrochloric acid were made a t 24" C. The ratio, in the limiting current region, of the diffusion current as recorded with the quadruple parallel-?' filter to the diffusion current as recorded with the peak follower is 0.81, which is in good agreement with the oscillographic data of Taylor, Smith, and Cooter. hleasurements made with this polarograph on a lO-zJi cadmium chloride solution gave a ratio of 0.813.
I I I 1 _ ~ - l . i . l i J -0.5 -0.7 -0.9 - { . I -4.3 -1.5 E ~ , M E ~ , v o l t ~ , a ? p o r e n l , v s , bSqG E .
The planned evaluation program for this instrument includes four principal phases: determination of the practicality of iR errors elimination by means of the controlled-potential feature, establishment of sensitivity limits in typical determinations, observation of the discrimination obtained betxeen derivative waves having close half-wave potentials and various values of n, and oscillographic polarography, CONCLUSIONS AND SUMMARY O F PERFORMANCE OF POLAROGRAPH
The results obtained in phase one of this program are illustrated by Figures 4 to 9. Only enough supporting electrolyte is required t o override the migration currents, which is of advantage when working with organic solvents or a t high sensitivity. Excellent polarograms are obtained with high cell resistances. The use of a small commercial asbestos fiber saturated calomel electrode is very convenient and satisfactory. The use of selective controlled-potential electrolysis is of value in polarography. Striking 1484
ANALYTICAL CHEMISTRY
EC
l
~
!
~
-1.3 - 1 5
- 4 {
E , vo115,opporent. rs.Aq
S.C.E.
Figure 8. Effect of i R o e l l drop, regular and derivative polarograrns for 4.4 1 O-3M 2-nitropropane
x
i
7 -
j CONTROLLED- POTENTIAL R
-1 i
(Corrected Electronically
I
I
* /
i //-
-0RNL MODEL Q-1677 POLAROGRAPH ( N o t Corrected f o r i R DROP) CELL
I
CON DIT IONS :
0.1 APPLICATIONS O F POLAROGRAPH
l ~ l , -0.5 -0.7 - 0 9
graph (not corrected for iROeii drop) Damping Regular RC = 112 sec. Derivative RC = 112 sec; Quad. 11-T, RC; Diode Controlled-potential and derivative polarograph (corrected electronically for i R c e i i drop) Damping Regular RC = 0 Derivative RC = 0; Quad. Il.T, RC; Peak follower Conditions: S.C.E. to D.M.E. 35,OOOQ (min.1 R,,IIS.C.E.to Pt W.E.26,OOOQ Solvent: Glycerol Supporting electrolyte: 0.3M LiCl 0.1 voIt/min. l Scan , rate: l
S.C.E. to Pt W.E.: 23,000R 'S.C.E,to D.M.E.:30,0OOfi(min) 4Opa.Current Range 0.1 v o l i / s e c . S c a n Rate 30 min. Np Sparge
C.-P.
8; D.P. 0-4673 P.
improvements are obtained by the eliniination of iR losses in both regular and derivative polarography. General observation of instrument performance indicates that the peak follower enables the recording of maximum currents regular polarograms and drop oscillations are removed without introduction of a time lag. The linear compensator is useful a t current sensitivities of the order of 1 pa. Initial aiid span potentials may be adjusted precisely by means of the "test 1 pa./Volt" function without recourse t o external aids. The cell open function readily identifies the sign of the cell current being recorded. Replicate regular and
1 Damping I RC = 0
I
R C = 1 / 2 sec
I
derivative polarograms may be superimposed. Kork in other phases of this program is in progress. Preliminary results support the folloning conclusions. It is frequently easier at high sensitivity to measure derivative than regular waves. A linear region of the residual current is inherently compensated for 111 derivative polarography. The loner limit of concentrations and the precisionoi the measurement of diffusion currents is limited primarily bv irregularities in electrochemical mass transfer mechanisms and in dropping mercury electrode behavior, and not by instrumental considerations. Current research indicates that the prac-
tical detection limit both for reversible and for irreversible systems is at least l O + M . The results of further investigation will be reported in a subsequent paper. ACKNOWLEDGMENT
The assistance of W.D. Cooke of Corne11 University in evaluating this polarograph is acknowledged with thanks. The polarograph photographed in Figure 3 was fabricated by the ORKL Instrument Department under the direction of G. A. Holt, C. C. Courtney, and D. D. Walker. LITERATURE CITED
(1) Arthur, Paul, Lewis, P. A,, Lloyd, N. A , , ASAL.CHEM.26, 1853 (1954). ( 2 ) Barker, G. C., Cockbaine, D. R.,
Atomic Energy Research Establishment C/R 1404, pp. 3-4, Her Majesty’s Stationery Office, York House, Kingsway, London W.C. 2, 1957.
(3) Barker, G. C., Jenkins, I. L., Analyst 77,685-96 (1952). (4) Booman. Glenn L.. ANAL.CHEM.29. \ - - -
,
(5) Bruss, D. B., DeVries, Thomas, J . Am. Chem. Soc. 78,733 (1956). (6) Ferrett. D. J.. blilner. G. W. C.. Analyst 80, 132-40 (1955).’ (7) Fisher, D. J., “Polarograph, ORXL Model Q-1673, High-Sensitivity, Diode
Filter, Derivative, Recording,” ORNL Master Analytical Manual, TID-7015 (section l ) , Method 1 003042 and 9 003042 (2-13-57), Office of Technical Services, DeDt. of Commerce, Washington 25, D. C: (~, 8 ) Hamm. R. E.. ANAL. CHEM.30. 350 I
(1958). ’ (9) Hume, D. N., Zbid., 30,675 (1958). (10) Ilkovic, D., Semerano, G., Collection Czechoslov. Chem. Communs. 4, 176 (1932). ( 1 1 ) Jackson, W., Jr., Elving, Philip J., ANAL.CHEM.28.378 f 1956). (12) Kelley, $1, T., Fisher, D. J., ZbLd., 28, 1130 (1956). . (13) Zbid., 30, 929 (1958). (14) Kelley, M. T., Miller, H. H., Ibid., 24, 1895 (1952).
(15) Kolthoff, I. M., Lingane, J. J., “Polarography,” 2nd ed., p. 504, Interscience. New York. 1952. (16) Lingane, ’J. J., Williams, R., J . .4m. C‘hem.SOC. 74, 790 (1952). (17) Nicholson, M. N . , A N A L . CHEY. 27, 1364 (1955). (18) Pecsok. R. L.. Farmer. R. W.. Zbid.. 28, 985 (1956). (19) George A. Philbrick Researches, Inc.,
Boston, Mass., “GAP/R Electronic -4nalog Computers,” “Applications Manual” and catalog data sheets. (20) Radin, N., DeVries, Thomas, ASAL. CHEX24,971 (1952). (21) Sawyer, D. T., Pecsok, R. L., Jensen, K. K., Zbid., 30, 481 (1958). ( 2 2 ) Taylor, J. K., Smith, R. E., Cooter, I. L., J . Research Natl. Bur. Standards 42, 387 (1949). (23) Texas Instruments, Inc., Dallas, Tex., *:?ata sheet, types 650C through 65329.
RECEIVEDfor review June 23, 1958. .%ccepted May 11, 1959. Division of .inalytical Chemistry, 134th Lleeting, .%CS,Chicago, Ill., September 1958.
Pola rogra phic Determination of Pentae rythrito I Trinitrate in the Presence of Nitroglycerin WILLIAM M. AYRES and GUY WILLIAM LEONARD Chemistry Division, Research Department, U . S. Naval Ordnance Test Station, China Lake, Calif.
A rapid polarographic method for the determination of pentaerythritol trinitrate in nitrocellulose propellants was developed for systems that may contain nitroglycerin, nitrocellulose, 2nitrodiphenyla mine, and dibutyl phtha Iate as major constituents. A sample of the propellant is dissolved in acetone. One aliquot i s polarographed without further treatment and the wave height obtained is ascribed to total nitrate-i.e., nitroglycerin, pentaerythritol trinitrate, and 2-nitrodiphenylamine. The second aliquot is reacted at room temperature with sodium hydroxide in ethyl alcohol, the nitroglycerin being decomposed by the sodium hydroxide, and the wave height ascribed to pentaerythritol trinitrate and 2-nitrodiphenylamine. The second wave height is corrected for the 2-nitrodiphenylamine present (determined spectrophotometrically) and the pentaerythritol trinitrate content is calculated or taken from a previously prepared standard curve. The first, or total, wave height is then corrected for both pentaerythritol trinitrate and 2-nitrodiphenylamine and the nitroglycerin content is calculated or taken from a prepared standard curve.
S
the development of a simple one-step synthesis ( 2 ), pentaerythritol trinitrate has been used in various nitrocellulose propellant formulations. The pentaerythritol trinitrate currently manufactured contains the di- and tetranitrated molecules as impurities. These impurities are added t o the propellant and determined by these procedures as pentaerythritol trinitrate. Typical constituents present in propellants which could be determined polarographically are dibutyl phthalate, 2-nitrodiphenylamine, nitroglycerin, and pentaerythritol trinitrate. As methods of analysis are available for dibutyl phthalate (4), nitroglycerin (S), and 2-nitrodiphenylamine ( I ) , a rapid method for determining pentaerythritol trinitrate in the presence of nitroglycerin was needed. The polarographic waves for all these constituents (except dibutyl phthalate) overlap and are indistinguishable with techniques now employed. A recent study in these laboratories indicated that nitroglycerin could be destroyed by alcoholic alkaline hydrolysis and that some other organic nitrates (such as the tri- and tetranitrates of pentaerythritol) were only slightly, if a t all, attacked by the base under the conINCE
ditions imposed. After several modifications of the Whitnack et al. procedure for nitroglycerin (S), a technique was developed for polarographically determining the total nitrate in the sample, decomposing the nitroglycerin by alkaline hydrolysis, and polarographically determining the remaining nitrate. iilthough no previous separation of nitrocellulose \Vas made, varying the amount of nitrocellulose present in the sample produced no apparent effect on the polarographic waves for pentaerythritol trinitrate or nitroglycerin ivith either the regular or reacted procedure. The interference due t o the presence of heavy nietal salts is negligible with the amounts normally found in propellant formulations. MATERIALS AND APPARATUS
-4Cary llodel XI recording spectro-
photometer with 1-em. quartz cells a as used for the determination of 2-nitrodiphenylamine. A Sargent hIodel XXI recording polarograph, employing a dropping mercury electrode and a mercury pool anode, was used to obtain and record the polarographic data. All samples were polarographed with no damping in 30-nil. beakers immersed in a constant temperature bath and VOL. 31, NO. 9, SEPTEMBER 1959
1485
