(2) Aston, J. G., Fink, H. L., Tooke, J. W., Cinnes, M. R., IND. ENG. CHEM., ANAL.ED. 19, 218 (1947). (3) Furukawa, G. T., McCoskey, R. E., King, G. J., J . Research Natl. Bur. Standutds 47, 256 (1951). (4) Glasgow, A. R., Jr., Streiff, A. J., Rossini, F. D.. Zbid., 35, 355 (1945)
Johnston, H. L., Giauque, W. F., J . Am. Chem. SOC.51, 3194 (1929). Kester, E. B., Ind.Eng. Chem. 24,770 (1932). Mastrangelo, S. V. R., Dornte, R. W., J . Am. Chem. SOC.77, 6200 (1955). Tunnicliff, D. D., Stone, H., ANAL. CHEW27, 73 (1955).
(9) Weissberger, A., (‘Physical Methods of Organic Chemistry,” Vol. 1 , Chap. 10, Interscience, New York, 1945. (10) Werner, A. C., Mastrangelo, S. V. R., J . Am. Chem. SOC.75,5244 (1953). RECEIVEDfor review August 27, 1956. Accepted January 29, 1957.
Improved Instrument for High-Frequency Conductometric Titration FRANK KUPKA and W. H. SLABAUGH Oregon Sfate College, Corvallis, Ore.
.The purpose of this work was to develop an instrumental method for the detection of end points in titrations involving colloidal electrolytes. Operating at a fixed frequency, an amplitude-stabilized oscillator in conjunction with a specially shaped conductance cell produces a practically linear response over a range of an order of magnitude of conductance. Use of three oscillator frequencies with two cells allows titration of solutions with conductances between 4 X loy4 and 6.5 X 1 O P N potassium chloride in the “linear” response range of the instrument. A nonlinear but useful response i s produced down to conductances as low as that of 3 X 10-jN potassium chloride and as high as that of 1N. The device permits rapid titration in colloidal systems, even though the conductance change so produced i s only a few per cent of the total conductance of the solution.
A
the simple response of a conventional 1000-cycle bridge to changes in conductance is especially desirable in titrating colloidal electrolytes, the necessity of having electrodes in contact with the solution makes i t very inconvenient to use this type of instrument. External electrodes, which eliminate many of these inconveniences, require a high-frequency device, and the possibility of constructing an inexpensive high-frequency unit with a linear response to conductance was explored. As indicated by Hall and coworkers (3), the response of a high-frequency cell to changes in conductance can at best only approach t h a t of a conventional conductance cell. To make the high-frequency cell response as nearly as possible like t h a t of the conventional cell, i t is necessary t o minimize capaciLTHOUGH
G
R FC r
Figure 1 .
A. Two-pronged socket B. 221/2-volt battery, Burgess C1,Cs,Cb,CS. 0.001 mfd C3, C4. 0.05 mfd. Ce, C7. 0.01 mfd.
High-frequency source
U-15
Cs. 0.002-mfd. mica Dz. Germanium diode lN58 A
NE. Neon bulb NE-2. R1. 51 kilo-ohms Rz. 10 kilo-ohms. R3. 39 kilo-ohms, 2 watts Ra. 4.7 kilo-ohms RE,Rs. 100 kilo-ohms R7. 2-kilo-ohm potentiometer Rs. 470 kilo-ohms RFC. Radio-frequency choke, 15 mh. VI, V2. 6.4&5 beam power amplifier tube tance effects within the cell by proper cell design. An instrument by Blake (1) measures the admittance of a conductance tube, b u t i t has a nonlinear response over the conductance range of the solutions studied. To shift the response range, the conductance cell was modified by ihserting capillary sections between the electrodes. The amplitude stability of the crystal oscillator was improved by a clamp tube circuit operated by the signal from the cathode follower stage. The latter serves t o isolate the oscillator from the load and provides a low impedance signal source for the conductance cell.
V3. lBAT7 dual triode
X. Crystal socket for crystal holders
with I/2-inch spaced pins, 0.093-inch diameter (for Type FT-243 and FT-241-.4 holders) Mounted crystals, approximately 0.1, 1.0, and 3.0 mc. One 2 x 4 X 10 inch chassis One 2 x 5 X 4 inch chassis Cabinet to hold meter, switches, potentiometer, and shunt resistors. Can also contain power supply Power supply, regulated at 250 and 150 volts, 40 ma. Connecting cable, shielded 5-conductor with appropriate terminals and sockets Battery, 6 volts A standard constant input-impedance multiplier circuit is used with the meter. The conventional bucking current arrangement permits taking advantage of the maximum meter deflection in the course of a titration by canceling out current carried by ions not involved in the titration reaction.
-
DESCRIPTION
OF APPARATUS
T h e high-frequency source (Figure 1) consists of a Pierce oscillator with a neon bulb connected as a limiter in the crystal circuit to prevent possible damage to the crystal. Frequency is changed by plugging in the appropriate crystal. Interchangeable mounted VOL. 29, NO. 5, M A Y 1957
845
M M.
MM
r Figure 3.
Figure 2.
Cell and equivalent circuit
Cell and meter circuits
AI. Two-pronged plug CIO. 0.001 mfd. D1. Germanium diode 1x35 H. Conductance cell holder pA. 25-pa. meter Rg. ' / 8 R M Rio. RM Rii. '/z RM Riz. R M Ria. ' / ~ R M Ria. '/7 RM
There RM is meter resistance (va1ues:best determined ex erimentally) Rls. 22 RI6. 10 kilo-ohm wire-wound potentiometer, 4-watt RI7. 600-ohni n-ire-n-ound potentiometer SI. Single-pole single-throw switch S2. Double-pole &position, nonshorting switch Sa. Double-pole single-throw switch (second pole used for power supply)
kilo-aims
CZ. Capacitance across solution
C3. External electrode capacitance Individual electrodes to solution Cd, Ck. capacitances d. Capillary diameter P I . Output electrode P2. Input electrode R1. Resistance of cell due to solution
I
6040 -
I
I
4
80
EFFECT OF DOUBLING FREQUENCY OR CAPACITANCE
cI=loo c2
20-
quartz crystals of a type formerly used in military transmitters are used for the 1000- and 3000-kc. frequencies and a secondary frequency standard type is used for 100 kc. The signal at the cathode of the cathode follower stage is rectified and applied to the grids of the dual triode, which has both sections connected in parallel. This acts as a clamp tube to maintain a constant signal output by reducing the screen grid voltage of the oscillator tube when the amplitude of the oscillation exceeds a value determined by battery B. The high-frequency source is built on a 5 X 2 X 4 inch chassis mounted on the conduction cell housing. The conductance cell is made of three sections of thin-walled borosilicate glass tubing 7 mm. in diameter, which are sealed to two sections of capillary tubing about 12 mm. long. The capillary in cell A has a n inside bore of 1.61 mm. and that in cell B a diameter of 0.66 mm. Each electrode is 0.005-inch copper foil wrapped around the 7-mm. tubing, soldered together, and cemented in place. The end sections extend about 4 inches beyond the electrodes. A diagram of a typical cell is shown in Figure 2, A . To facilitate changing conductance cells, modified l/r-inch fuse mounting clips are used to make electrical contact with the electrode and to provide mechanical support for the cell. The cell and associated germanium rectifiers and chokes are mounted in a 10 X 2 X 4 inch chassis. The indicating meter and power supply are housed in a 8 X 8 X 8 inch cabinet. A conventional voltageregulated power supply capable of delivering u p to 40 ma. at 250 and 150 volts is used. Connections from the 846
ANALYTICAL CHEMISTRY
C E L L CONDUCTANCE pMH0.S
Figure 4.
Theoretical response of cell
i
120-
Figure
5. Response of cell A a t 1 mc.
power supply to the oscillator and from the meter to the conductance cell are made of a multiwire shielded cable terminating in an octal plug and socket. The meter is a 41/2-inch rectangular model with a 25-pa. movement accurate to h 2 % . Experimentally determined shunts are used to produce meter ranges of 50, 100, and 200 pa. with a constant input impedance maintained by appropriate series resistors. A bucking current may be supplied to the meter circuit by means of a 6-volt battery and the associated resistor-potentiometer network. This feature is especially useful in performing titrations in solutions which contain high concentrations of ions not involved in the titration reaction. PRINCIPLE
AND
METHOD
OF OPERATION
The output stage of the oscillator applies a signal of constant frequency and amplitude to the middle electrode of the cell. The path through the cell to ground includes capacitance between the electrode and the solution, electrolytic conductance within the solution, and capacitive coupling from the solution to the other electrodes where rectification to direct current is accomplished by germanium diodes. Change in the electrolytic conductance of the solution produces a corresponding change in the current through the diodes and is observed on a microammeter. A frequency of 1 me. in conjunction with a cell A is suitable for most titrations. The solution to be titrated is drawn into the lower end of the conductance cell by a Sigmamotor pump and returned to the titration vessel, thermostated a t room temperature. The oscillator output and bucking current are selected to produce a nearly full scale meter deflection in the 50or 25-pa. range during a titration. Meter deflection is plotted against volume of titrant as is done with data obtained by an ordinary conductance bridge. The time required for an ordinary titration. including cleaning the cell for the nest titration, is less than 10 minutes. THEORETICAL CONSIDERATIONS
The electrical equivalent circuit of
0.000289 N A g N 4
the conductance cell is shown in Figure 2, B. Electrodes P1 (Figure 2, A ) are considered as a single electrode. Because the interelectrode capacity, Car is constant and both the frequency and amplitude are fixed, the portion of signal passed by C3is constant and need not be considered in a titration apparatus. C4 and Cs can be treated as a single capacitance, C1, located a t C4. This simplified circuit can be further reduced to a series equivalent circuit consisting of a series resistance, R,, and series capacitance, C,, the values of which can be calculated from equations referred to by Reilley and McCurdy ( 5 ) in their excellent discussion of high-frequency measurements. The current, I , passing through the series circuit may be calculated from the equation ( 2 ) I = d E m where E is the applied signal voltage and X , is the reactance of C, (2). X , becomes small for large values of C1 and small values of C2; thus the current becomes almost entirely a function of R. over a wide range of values of the latter. Small values of Cz make R, nearly equal to R1, except a t very high values of R1. As Cz decreases and R1 increases as the diameter of the section of the cell betreen the electrodes decreases, these parameters may be adjusted to some extent. By suitable selection of C1 and C2 the current becomes very nearly proportional to the solution conductance over a considerable range of the latter. The theoretical curve for a cell in which C1 = 100 C B ,E = 1, and Ci, Cp and the frequency have fixed values is shown in Figure 4. The relatively linear portion of the curve in the region of the inflection point becomes longer as the ratio of C1 to C p increases. As the curve illustrates, increasing C1 extends the upper portion of the curve, thus improving the high conductance response. Decreasing CB extends the low conductance response. Changing the frequency has the same effect as a proportional change in both C1 and Cz.
Table I.
Linear Ranges of Experimental Cells Capillary Fre- Linear Range Diameter, quency, (Nolarity Cell blm. hlc. of KC1)
B
0.664
1.0
3.0
3 . 5 x 10-3 to 3 . 5 x 10-2 6 X to 6.5 X
EXPERIMENTAL RESULTS
The experimental curve for cell A a t a frequency of 1 me. is shown in Figure 5. The conductance values approximate 1 x 10-4M potassium chloride concentrations. The current scale is in arbitrary units. Cell B was constructed with thinner walls and capillary sections of smaller diameter, thus shifting the response to higher specific conductances. Its response shows the same general features as cell A. The practically linear response ranges of these cells a t various frequencies are shown in Table I. In addition, a t 0.1 me., cell A gives a nonlinear but unambiguous response to conductance as low as t h a t of 3 X low5 iV potassium chloride. For cell B a t 3 mc., a useful response is obtained up to the conductance of 1M potassium chloride. Figure 6 shows the effect of frequency on the shape of a typical titration curve. Curve I illustrates the titration of a solution whose conductance is within the linear operating range of cell A a t 0.1 mc. The same solution a t 1.0 mc. has a conductance which moves the titration curve below the linear range shown in Figure 5. A solution whose conductance places it above the linear section of the response curve produces a titration curve that is convex upward. Figure 6 (right) is the titration curve for a suspension of calcium bentonite buffered a t p H 8.5 with potassium hydroxide and boric acid. The titrant is 0.107N dipotassium dihydrogen (ethylenedinitrilo)tetraacetate (EDTA). The curves show excellent reproducibility and provide data from which the base exchange capacity of calcium bentonite may be calculated. The values thus obtained agree with those determined by conventional ammonium acetate methods (4).
la-31GRAM1 A CALClU M
BENTONIT E
ACKNOWLEDGMENT
I
0
I
0.5
I
1.0 ML. KCl
I
1.5
2.0 0
2D
1.0 ML
3.0
4.0
EDTA.
Figure 6. Frequency effect (left) and titration of colloidal system (right)
The authors gratefully acknowledge support of the project by the Baroid Division, National Lead Co., and the assistance of H. R. T’inyard, Department of Physics, in the design and VOL. 29, NO. 5, MAY 1957
847
construction of the electronic equipment. LITERATURE CITED
(1) , , Blake. G. G.. “Conductimetric Anal-
ysis at Radio-Frequency,” pp. 34-
42, Chemical Publishing Go., New York, 1952. (2) Crufts Electronics Staff, “Electronic Circuits and Tubes,” McGraw-Hill, New York, 1947. (3) Hall, J. L., Gibson, J. A., Phillips, H. 0.. Critchfield. F. E.. ANAL. CHEM.’26, 1539 (1954).
’
(4) Peech, AI., IND.ENG.CHEX, ANAL. ED.,13, 436 (1941). (5) Reilley, C. N., McCurdy, W. H., ANAL.CHEM.25, 86 (1953).
RECEIVEDfor review Mav . . 7.. 1956. Accepted January 28, 1957.
Selective Precipitation of Thorium Iodate from a Tartaric Acid-Hydrogen Peroxide Medium Application to Rapid Spectrophotometric Determination of Thorium in Silicate Rocks and in Ores F. S. GRIMALDI, LILLIE B. JENKINS, and MARY Geological Survey, Washington 25, D. C.
H.
FLETCHER
U. S.
)This paper presents a selective iodate separation of thorium from nitric acid medium containing d-tartaric acid and hydrogen peroxide. The catalytic decomposition of hydrogen peroxide is prevented b y the use of 8quinolinol. A few micrograms of thorium are separated sufficiently clean from 30 rng. of such oxides as cerium, zirconium, titanium, niobium, tantalum, scandium, or iron with one iodate precipitation to allow an accurate determination of thorium with the thoronmesotartaric acid spectrophotometric method. The method is successful for the determination of 0.001% or more of thorium dioxide in silicate rocks and for 0.01% or more in black sand, monazite, thorite, thorianite, eschynite, euxenite, and zircon.
T
HE precipitation of thorium iodate from nitric acid medium (4) is a generally reliable and widely used method for the separation of thorium. Lead, mercury, tin, niobium, tantalum, tungsten, cerium(IV), uranium(IV), zirconium, titanium, silver, and to a smaller extent scandium, bismuth, and iron(II1) also precipitate from this medium. A clean separation of thorium iodate is obtained from the rare earth elements by reprecipitation. A dense, less contaminated, and easily filterable precipitate is obtained from homogeneous solution (6). Tillu and Athavale (7) used oxalic acid to prevent the precipitation of 20 mg. each of titanium and bismuth and 40 mg. of zirconium. The procedure was not applied to the determination of small
848
ANALYTICAL CHEMISTRY
amounts of thorium. Kronstadt and Eberle (3) used mercury as a carrier for the precipitation of 20 y or more of thorium. The present investigation concerns the separation of thorium iodate from nitric acid medium containing hydrogen peroxide, d-tartaric acid, and S-quinolinol. Tartaric acid minimizes the coprecipitation of zirconium, tungsten, scandium, and bismuth. Hydrogen peroxide minimizes the precipitation of titanium, niobium, and tantalum; 8-quinolinol prevents the catalytic decomposition of hydrogen peroxide, which is especially serious in the presence of cerium. Although less than 10% of the iron added is precipitated, the mixed carrier of mercury and iron used is more effective than mercury alone for the precipitation of microgram amounts of thorium. This separation procedure combined with the recently developed (9) spectrophotometric determination of thorium with the thoron-mesotartaric acid system is applied successfully to the determination of 0.001% or more of thorium dioxide in silicate rocks and 0.01% or more in black sand, monazite, thorite, thorianite, eschynite, euxenite, and zircon. REAGENTS AND APPARATUS
All chemicals used are reagent grade. Ferric nitrate (carrier solution), 1 ml. equivalent to 2 mg. of Fe20B. Dissolve 0.875 gram of ferric nitrate hexahydrate in 100 ml. of (1 plus 99) nitric acid. Potassium hydroxide (precipitating solution), 50% by weight aqueous. Potassium hydroxide (wash solution). Dilute 2 ml. of 50’% potassium hydroxide solution to 500 ml. with water.
Ammonium nitrate (wash solution), 1% aqueous. 8-Quinolinol. Dissolve 0 . 5 gram of reagent in 100 ml. of (1 plus 99) nitric acid. Hydrogen peroxide solution, 3%. Dilute 10 ml. of 30% hydrogen peroxide to 100 ml. with mater. &Tartaric acid solution. Dissolve 600 grams of tartaric acid in sufficient water to make 1 liter of solution. Filter through a dry paper and do not mash. Potassium iodate solution, 6% aqueous. Filter through a dry paper and do not wash. Mercuric nitrate (carrier solution), 1 ml. equivalent to 1 mg. of HgO. Dissolve 1.58 grams of mercuric nitrate monohydrate in 10 ml. of (1 plus 1) nitric acid and dilute with water to a liter. Iodate mash solution. hlix 60 ml. of nitric acid, 6 ml. of 30% hydrogen peroxide, and 200 ml. of 6% potassium iodate solution with enough water to make a liter of solution. PROCEDURE
The preparation of the solution for analysis should present no problems, except occasionally for a niobium and tantalum ore. illthough the medium for the precipitation of thorium will keep niobium and tantalum in solution, there may be a problem in preparing the solution of the sample in this medium without prior hydrolysis of niobium and tantalum. Two alternative procedures are given. The first procedure is the simplest but may fail on high-grade tantalates containing very little titanium. The second procedure is of general applicability. Procedure 1 (Ores and Silicate Rocks). hlix 0.0500 gram of a finely
