Coulometric Determination of Ferrocyanide with Electrolytically Generated Ceric Ion ROBERT V. DILTS' and N. HOWELL FURMAN f r i c k Chemical Laboratory, Princeton University, Princeton,
5amples of potassium ferroc?anide from 1.7 to 62 microequivalents habe been titrated with electrol) tically generated ceric ion with an accuracg of +0.40% or hetter. The sensitite amperometric end point was eniplog ed. The chemical reaction near the equibalence point wds slow. but h! generating in small increments in this region, and waiting a few minutes for equilibr i u m to become established, accurate results were obLainetl. J erg dilute potassium ferrocg anide solutions were found to be unstable so that i t was necessar) to use, as samples, weighed aliquote of 0.01 \- solutions, which deterniiiiecl the lower limit of samples that coitltl I)e measured accuratelb.
T
FIE determination of ferrocyanide both for its own inherent v:ilue, arid :E :i retluction product in the indirect determinatioii of other suhstances-e.g., hydrazine, reduring sugars, etc.h:is heronie iiicrwsinply important. Ceric sulfate has been used I Fwlunietric*oxidizing agent in this determination by numer011s iiivc.stigator5. Tlie encl point of this determination has been t I c ~ t o c . t c dhy visixil indicators such as I . 10-phenanthroline ferrous aii1J':itcL ( 1 0 ) . diphenylamine sodium sulfate ( 2 ) . or the greenish cmloi, 01' fct,i,ic ferrocyanide ( 6 ) .and potentiometric methods ( 2 , 6, 't). ( ) f the \\-or!iers who hiivtl stiidied this reaction. B e r r , ~( 2 , Rivw I P W data a n d 11-illartl and \-oung( 1 0 )merely tabulate general caontlitions for tlie titration. Fnrm:tn and Evans (6) have made :I tlvt:iil(d study of t h e conditions that are optimiini for the direct of ferror>-tinirle,and also for the titration of ceric tic~tei~ii~iiiatior~ ion \\-ith fc~rrocy:iiiitlr. )lost of tlie c-onditions for the titrations pertornietl in tiiis inwstigation n-ere based on the material presentcd t)j. t h c w 1:ist Ruthore. 1lel:itively few inorganic siibstances have been determined coulomc~trically using electrolytically generated ceric ion. T o d:itrJ those include ferrous iron ( 5 ) . uranium (4),and iodide ion ( 7 ) h i i t not ferroc*y:inicle. Since a sensitive method for determining mic~rocjn:intitiesof ferrocyanide might be developed by nieaiis of a coiilonietric procedure. the present study was directed t o w rd this aini . APPARATUS
Figiir(> 1 rhon-s a schematic representation of the apparatus iiaed ioi, the roulonietric titrations. T h e constant current source, designed ti!. Reilley, Coolie, and Furman ( 8 ) , is capable of delivering cnrrents from 1 t o 5 ma. and from 25 t o 150 ma., in multiples lues intermediate t o these, several in the external circuit. When 25 Ma. and 1 ma., t h e apparatus source ivith high resistors, which t h e voltage supply t o provide t h e d e k e d current intensity (see Figure 1). Two Ohmite, T y p e ilB, v:ii,iable c:trt)on resistors were included in the circuit in order t o rn:ike fine adjustments of the current value. When large currents of 10 ma. or lnrger were used, these potentiometers were shortc,iwiiited to prevent them from being burned out. -4 10-ohm n-irv-wound resistor was included in the circuit as a dummy for t h e titration cell. Thus, verj- little change occurred in the resist: i n w in t h c circuit when the current was permitted to flow tlrroligh t h e titlntion cell. The, generating current as measured hy determining t h e ZR ilrop :icross :i standard 10,000-ohm resistance box with a Leeds ~-
address, Department of Chemistry, Williams College, Williams-
N. J.
and S o r t h r u p Ftudent-type potentionic~tt~r..ill time measurements were made with a Alodel S-6, clutch-operated clock from t h e Standard Electric Time Co. The sensitive end-point procedure of Cooke, Itrille>.>and Furman (3) \vas used for detection of the end of t h e titration. The circuit for this consisted of a Leeds and S o r t h r u p Type P galvanometer, a Shallcross Aryton shunt, and two I.5-volt dry cells, 1vhic.h served as a voltage source. The generator electrodes were of platinuni-ii,idirim foil. The anode Y:P 2 X 4 cni. and the cathode w:iq 1 X 2 cni. T h e cathode was i3ol:ited from the hulk of the solution i i i the titration vessel by :t separate compartment, which had a sintered-glass disk at its hottom and ~ v : vfilled above t h e kvel of the, solution in ,the titration (le11 n i t h 15y0 ammonium sulfate soliition. Potassium sulfate and sodium sulfate could not be uped i n this compartment n-hen working with the cerous sulfate generating solution for a precipitate of cerous alum \vas formed from t h e small amount, of salt t h a t leaked from t h e cathode compartment into the titration v e s ~ c l . Tlie presence of this precipitate hindrreti the proper functioning of hoih the generator anode and the indicator electl.Od('R. The operating-indicator electrode co~isistedof a 2-cm. long platinum wire. B and S gage 28, for the I:rrgcr samples, hut with microsanip1e.y n-here a greater eensitivitj- w:is requirrd, a 1-cm. squnrc~p1:itiiiiim-iridiuni foil was suhstitutrd for the n-ire. T h e refrrc~nw(-ellconsisted of a saturated re11 of lead am:ilgttni-lead .sulfate in 1.V siilfiiric acid, as reconimr~ntiedby Cooke, Reilley, a n d Furmnn ( 3 ) . This cell hac a const:tnt potcmtial of -0.27 volt V ~ I Y I P the hydrogen half-cell and has t h e advantagc that it introcliicw no interfering ions into t h e titration c ~ I 1 . .\ S o . 7664 Leeds a n d S o r t h r u p pH meter was u r t d t o present the potential of t h e indicator cirruit for thr, sensitive end-point procedure. Tlie titration cell consisted of a Tveighing tiottlr of a 30-ml. capacity, covered n-ith a rubber stopper provided with openings for t l i r four electrodes, a gas inlet, and the addition of t h e sample. All samples n-ere added to the titration cell from a 5.000-ml. niicrohuret graduated in 0.01 ml., except those of 1.8-microequivalent size, which were added from a BD hypodermic syringe of 1.0-ml. capacity, that served as a weight buret. il piere of glass tuhing was drawn out to provide a fine tip for deliver?. of tht, sample. SOLUTIONS
The generating solution was prepared hy dissolving reagent fate tetrahydrate (G. F.Smith Chemical Co.) in id until the solution was saturated (ca. 12y0). T h e ferrocyanide solutions were prepared by dissolving neighed amounts of reagent grade potassium ferrocyanide trihydrate (Baker and .;\damson) in 2.0S sulfuric acid which had heen deaerated with tank nitrogen, and b y making u p t o volume with deaerated water. These solutions were standardized hy titration Tyith 0.05205.Y ceric sulfate, standardized against arsenious oxide, using I , 10-phenanthroline ferrous sulfate as a n indicator. T h e ferroc?.anide solutions were found t o be somexT-hat unstable, making it necessary t o prepare and standardize a fresh solution each day. When not in use throughout the course of the day, they were kept in the dark in order to minimize decomposition. PROCEDURE
E i g h t c w milliliters of the saturated cerous sulfate solution tvere placed in t h e titration cell and deaerated with tank nitrogen for 15 minutes. For samples of 1.8 microequivalents, 3.0 ml. of sirupy (€%Yo) phosphoric acid also were added t o eliminate a drift in t h e indicator system. .;\fter deaeration, t h e potential of this solution was adjusted t o 1.210 volts as indicated by zero current flon- on the galvanometer. Since the potential of the generating solution was usually below this value, this presetting of the solntion ,potential was easily accomplished b y merely generating ceric ion. However, occasionally it was found necessary t o add a drop of ferrocyanide t o reduce the potential t o a value helow t h e reference one, and then generate ceric ion. Then
1275
1276
ANALYTICAL CHEMISTRY
t h e sample of ferrocyanide was added t o the cell and generation of ceric ion a t constant current begun and continued until 781-0 current reading on the galvanometer was obtained again. I n t h e virinity of t h e end point. where t h e concentration of the reacting ions a a s very low, generation was carried out in small increments with a wait of 5 minutes (10 minutes in the case of a second or third sample titrated in the same background mixture) t o allow equilibrium to be established. T h e number of microequivalent. of ferrocyanide nere calculated, from t h e time of generation and t h e value of the current. T h e results of a series of such determinations are presented in Table I. DISCUSSION
T h e potential impressed across the indicator electrodes (1.210 volts) x m s determined b y potent,iometrictitrat,ion of ferrocyanide with electrolytically generated ceric ion. The potential of the midpoint of the break in thjs titration curve vias selected as the most sntisfactory one t o use as the preset end-point potential.
Table I. Coulometric Determination of Ferrocyanide with Electrolytically Generated Ceric Ion Current, Na.
Time, Nin,
Added
11.91 11.92 11.90 12.05
8.419 8.393 8 290 8.223
62.29 61.93 61.68 61.62 30.94 30.R9 30.69 30.63 5.492 5.449 3 433 5.179 1.949 1.867 1.794 1.734
4,090 4,090 4,130 4.100 1.087 1,090 1.090 1,090 0.49G2 0,4929 0.4956 0,4942
12.139 12.024 11.919 12 003 8.135 8.039 8.027 7.653 6.304 6.071 5.837 5.634.
Microequivalents Found Difference 62.34 62.13 61.34 61.61
fO.03 +0.20 -0.34 -0.01
30.87 30.58 30.61 30.60 5.498 5.448 5.440 5.187 1.945 1.8fil 1.799 1.731
-0.07 -0.11 -0 08 -0.03 +0.006 -0 001 4-0.005 i-0.008 -0.004 -0.006 f0.005 -0,003
Error,
background medium, a wait of 5 minutes m s found to be generally an adequate period. I n general, the ferrocyanide solutions Tyere sufficiently stable throughout an 8- or 9-hour period to permit their use Tvithin t h a t length of time without restandardization. Hon-ever, a decrease, observed in the titer in the case of the 0.003S solution, was so slight t h a t no significant error was introduced through its continued use. I n the case of 0.001S ferrocyanide. the decay in titer was so rapid t h a t successive titrations, when standardizing the solutions, gave significantl?. loTver values (approximately 0.20 ml. of the ceric sulfate) each time. Also, xithin 2 hours the solution had decreased in concentration from 0.001103 to 0.0010325. Thus, it m-as .impossible to use solutions t h a t were this dilute. Therefore for the 1.8-microequivalent samples, 0.01.V ferrocyanide was used, and the hypodermic syringe was employed for the addition of the sample instead of the microburet. K i t h ferrocyanide solutions of this concentration, samples much smaller than 1.8 microequivalents could not be determined conveniently Kith great precision since aliquots would have had weights only slightly greater than 100 mg. Perhaps the use of 0.005S ferrocyanide solutions would permit the determination of 1 microequivalent (422 of ferrocyanide), hut with samples smaller than that, the accuracy of the measurement of sample size would he less than t h a t of the determination of the concentration itself.
+O 08 +0.32 -0,55 -0.02 -0 23 -0.36 -0.26 -0.10 4-0.11 -0.02 fO.09 +O.l5 -0.21 -0.32 +0.28 -0.23
Deaeration of the cerous sulfate generating solution was necessary t o obtain precise results. for without it the solution became pale blue upon addition of the sample and the results were from 4 t o 13yo too low. It was found necessary to disconnect the indicator electrodes throughout the major part of the titration and only have the detection circuit in operation during the last minute of the titration. Othern-ise a slight blue precipitate formed on the platinum indicator electrode, which caused the results t o be about 1% too low. T h e indicator electrode was cleaned readily b y placing i t in boiling 1 to 1 nitric acid for 10 minutes and then holding i t in a soft flame for a fevi minutes. If the generation of ceric ions was carried out continuously until the galvanometer reading m-as zero, the results were again low. When the generation was halted there was no noticeable drift,in the downward direction of potential readings a t this point; thus t h e reaction appeared t o have gone to completion. However, if the generation n-as stopped before the galvanometer reached a reading of zero current, a slight downward drift was noted, indicating a sluggishness in the reaction near the equivalence point. This is not entirely unexpected from a consideration of the sizes and concentrations of the ions involved a t this portion of the titration curve. Furman and Evans (6) also report this in their macrotitrations and state that from 1 t o 3 minutes are required in the immediate vicinity of the end point t o obtain stable potential readings. I n this work, 5 t o 10 minutes were found t o be necessary for the attainment of stable galvnnomet,er readings-Le., two readings on the galvanometer scale werc identical for 1 minute. When a second sample had been added to the szme solution in the cell after completion of one titration the longer period of time was necessiry, while for t,he first ssmple in a
POTENTI-
(IT ANOAR0 RI?(II(ITANCC BOX
..'WM 00 K
CELL
+
0
11.
TIMER
t.t
I Y
'On 50
@
Figure 1.
ttOK
Y
K
5.0
M
6 VOLTS
Schematic generation circuit
The data in Table I show t h a t samples of ferrocyanide from 1.7 t o 62 microequivalents (or 0.72 to 26.2 mg.) can be determined with an accuracy of 1 0 . 4 0 %$by a coulometric procedure. B y using higher generation currents, larger amounts certainly could be determined and perhaps smaller samples could be titrated, although the accuracy of the knowledge of the amount of szmple added becomes less as the sample size decreases. It should also be possible to apply this determination of ferrocyanide t o such procedures as those involving reducing sugars and other substances in which the ferrocyanide produced by the reduction of ferricyanide is a measure of the amount of material present. LITERATURE CITED
(1) .4tanasiu, I. A , , a n d Stefanscu, V., Ber.. 61, 1343 (1928). (2) Berry, 4.J., Analyat, 54, 461 (1929).
V O L U M E 27, NO. 8, A U G U S T 1 9 5 5 ( 3 ) Cooke, W. D., Reilley, C . S . , and Furman, N. H., ANAL. CHEM.,23, ltiG2 (1951). (4) Furman, S . H., Bricker, C. E., and Dilts, R. V., ANAL.CHEM., 25, 482 (1953). (5) Furnian, S . H.. Cooke, W.D., and Reilley, C. K.,Ibid..23, 945
(1961).
( 6 ) Furman, X . H., and Evans, o., J . Am. C h e m . SOC.,51, 1128 (1929). ( 7 ) Lingane. J. J., prirate communication. 1964.
1277 Reilley, C. N., Cooke, W. D., and Furman, S . H., .Is.AI.. CHEM., 23, 1030 (1951). (9) Someya, K., Z . anorg. allgem. Chem., 181, 183 11929). (10) Willard, H. H., and Young, P., J . Am. Chwm. Soe., 55, 3260 (1953). (8)
RECEIVED for review March 18, 1955. Accepted April 22, 1955. Based o n dissertation submitted b y Robert V. Dilts in partial fulfillment of the requirements for the degree of doctor of philosophy a t Princeton University, 1954.
Thermogravimetric Pyrolysis of Cupferron Complexes of Scandium, Yttrium, and Rare Earth Elements WESLEY W. WENDLANDT Department o f Chemistry and Chemical Engineering, Texas Technological College, Lubbock, Tex.
The thermogravinietric pyrolysis of the scandium, > ttrium, and the rare earth cupferrates was determined. It was found that it is not necessary to ignite the complexes to 900" C. but that the oxide level is reached at 500" to 600" C. The construction and operation of a simple therniobalance are described.
T
HE use of cupferron (ammonium salt of S-nitroso-pheiiylhydrox) lamine) as a precipitant for the rare earth elements has recently been investigated ( 5 ) . It n a s found t h a t the rare earth elements-lanthanum, cerium(III), praseodymium, neodymium, saniai ium, and gadolinium-were quantitatively precipitated from solution a t a p H of 3 to 4. As the precipitates were contaminated with an excess of cupferron, they were ignited and weighed as the corresponding oxides. This study was nndertaken to determine the optimum temperature limits for the ignition to the oxide and also t o investigate any intermediate products t h a t may be formed during the decomposition.
The heating rate of the furnace was controlled b>- a 6 revolutions-per-day 110-volt synchronous motor connected to the shaft of a Powerstat. The heating rate was linear from 35' to 950" C.. a t about 4.5' per minute. A slower heating rate could be obtained by decreasing the input voltage into the motor-driven Powerstat. The temperature of the furnace was measured by a n ironconstantan thermocouple, T , using a n ice bath as the reference junction. T h e potential of the thermocouple was detected by a Gray portable potentiometer, Model E-3042-6, made by the Gray Instrument Co., Philadelphia, Pa. T h e thermocouple was calibrated against the freezing point and boiling point of water, and the freezing points of cadmium metal and potassium chloride. The electromotive force of the thermocouple could be read to within j ~ 0 . 0 5mv., which corresponds t o about & l o .
w EXPERIMENTAL
Reagents. Cupferron was obtained from the G. F. Smith Chemical Co., Columbus, Ohio, and the Matheson, Coleman, Bell Co., East Rutherford, K. J. It was used without further purification. T h e rare earths were obtained as the oxides of better than 99% purity from Research Chemicals, Inc., Burbank, Calif., and the Lindsay Chemical Co., K e s t Chicago, Ill. .4s the contaminants were other rare earth elements, no further purification x a s neces. sary. Scandium oxide of 99.870 purit,y was obtained from A . D. l l a c k a y , Inc.. S e x - Tork. X . T. Yttrium chloride of 99% purity was obtained from Research Chemicals, Inc., Burbank, Calif. All other reagents were of analytical reagent grade. Thermobalance. I n a recent book, Duval ( 1 ) summarizes the development of the thermobalance. H e found that if a standard beam balance were used, the vibration soon dulled the knife edges and could also result in the beam's being displaced from the center position. These difficulties were overcome by the use of a multiple-range, precision torque balance. The balance was converted into a thermobalance as shown in Figure 1. The torque halance, 0 to 100 mg. in range, was made by the Vereenigde Draadfabrieken, Nijmegen. Holland. T h e smallest scale division was 0.2 mg.; thus, weighings could be made to i z O . 1 mg. The sample was placed in a platinum boat, 1 em. in diameter, SUBpended by a platinum wire. in a Vycor glass tube, 2.5 cm. in diameter and about 25 cm. in length. This tube was connected 29/52 joint at B. T h e furnace, -4,was to the balance by a constructed by first winding IA feet of Yo. 22 gage, Nichrome inch alloy V, resistance wire (1.01 ohms per foot), into a coil in diameter, then winding this coil onto the asbestos-covered tube a t about '/c-inch spacings. The completed windings were covered adequately with asbestos insulation.
Figure 1.
Schematic diagram of thermobalance
It was found t h a t during the pyrolysis of the cupferron complexes, the decomposition products adhered to the sides of the tube and the suspending wire. T o revent this difficulty, a suction was applied to the tube, S, whic1 allowed a slow stream of air to pass over the sample. No adverse effects could be detected
