April, 1959
CELLSHAVINGA FUSEDZINC CHLORIDE ELECTROLYTE
2); if Di comes out -, x is set to equal -2. The procedure is repeated as many times as is necessary to get \Oil- F < O . In the present work, solutions of the individual equilibria were carried out using the simplifying assumption that y in equation (D) is negligible compared with 8. It is believed that avoidance of this assumption would have shortened the calculations only to a minor extent. It is recommended that F,the subiteration error, be set an order of magnitude smaller than E, else there may be danger of an infinite loop occurring. Furthermore, E should not be specified with too small a value such that it represents digits beyond those carried by floating point numbers. Extension to Additional Elements.-The exten-
525
sion of the present procedure to include additional (chemical) elements is quite simple. The first step, of course, is to decide which are the important reactions. Usually an estimate will be available from previous work as to the most reasonable species to choose as components. Then, after rewriting the chemical equations to correspond to reactions of formation of the other species from the components, one need merely provide for inclusion of the additional data and code in appropriate calling sequences for the new reactions. It is proposed to carry out such extensions in subsequent work. Availability of Program.-The author will be pleased to reproduce a copy of the program for any one wishing to see it.
EXPERIMENTS WITH THE CELL Bi, Bi203,ZnC12/ZnC12/ZnC12,Zn BETWEEN 450 AND 510' BY REUBENE. WOOD National Bureau of Standards, Washington
D. C.
Rsosived J u l y SI, 1968
As a p v t ,of study of cells having a fused zinc chloride electrolyte, experiments were carried out on such a cell in which the negative electrode was liquid zinc and the positive electrode was liquid bismuth on top of which had been put some BiZOa. The cell potentials are quite rapidly and accurately reproducible after temperature changes and after the passage of current. The cell Dotentials and temperature coefficient are reported. Phase equilibria in this system may be rather complicated and have not yet been studied.
Introduction The calomel electrode has long been of interest and utility sts a reference electrode in aqueous electrochemistry, but it is difficult to devise a close, fused-salt analog of this electrode because of the relatively high mutual solubility of the metal halides. As R possible basis for a satisfactory fused-salt reference electrode, similar if not exactly analogous to the aqueous calomel electrode, the use of bismuth and bismuth oxychloride in a suitable fused halide electrolyte suggested itself. The cell that was assembled and studied to test this idea may be represented as Bi, BizOsIZnCln/ZnClz/ZnCln, Zn
(1)
What solid phases are present at the bismuth electrode when this cell is operating reversibly is not yet known. It seems unlikely that they include Bi2O3, more likely that they include bismuth oxychloride and ainc oxide or zinc oxychloride. In fact, although he did not report studies of systems containing ainc, SillBn's work' on two-metal oxyhalide compounds suggests the possibility that reaction between zinc chloride and bismuth and zinc oxides could produce a series of double oxychlorides. However, the above representation is used for the cell to indicate that BizOaand the other substances shown were the only ones put into the cell. This cell was studied over the temperature range 450 to 510" and over a period of several weeks. (1)
L. G . SillBn, Nalurwissenscha/lsn, 80, 318 (1942).
Apparatus and Procedures The cell system used was similar t o the conventional H cell except that there were three vertical tubes and two cross connect,ions between each outside vertical tube and the middle vertical tube. The two lower cross connections provided the electrolyte path between the electrodes, the u per cross connections permitted the passage of helium wiich was maintained as an inert atmosphere above the cells. The pu ose of having three vertical tubes was to allow the use two Zn-ZnCls reference electrodes against which to measure the potential of the bismuth electrode. One of these zinc reference electrodes was in the center tube, the other wah in one of the outside tubes. At the bottom of each vertical tube were ap roximately 4 ml. of molten metal, zinc in two tubes, bismutlin the other. 2.4 g. of B&Oswas laced above the bismuth and then about 100 g. of liquid Zn& was introduced into the cell. A glassenclosed thermocouple junction dipped into the molten metal of each electrode. Electrical connection was made to the molten metals through graphite rods introduced into the cells through side arms so that these rods were not in contact with the electrolvte. Reagent-grade chemicals were used exclusively except that the zinc was an especially pure product asserted by the manufacturer to be 99.99% The zinc chloride was dried by bubbling anhydrous !%?through it at about 500' for 10 hours, then dried oxygen for about two hours. The cell was kept hot in a cylindrical electric furnace. This furnace was supplied with current from a constantvoltage transformer through a variable resistor system. No automatic temperature regulator was used but because of the constancy of power input and of the room temperature, the temperature conditions within the furnace remained quite static &R long as it was undisturbed and the resistance was not changed. Under such conditions, gradual temperature variations at the rate of 0.5" per hour would be typical. All potentials were measured with a T e K potentiometer using a standard cell calibrated in g e s e laboratories. The thermocouples were calibrated only at the melting
2
526
REUBEN E. W O O D
point of zinc. The deviation from standard tables2 found a t this temperature (about 1.5') was assumed to represent a satisfactory correction over the temperature range of these experiments to within the probable accuracy of the values of E and dE/dT reported here.
Daysa
Experiments and Observations
10
The electrolyte over the bismuth electrode had the pale pink t o lavender color which hasbeen found to characterize all such electrodes. This is presumably the color of some bismuth-containing ion. The potential difference observed between the two zinc electrodes was about 0.15 millivolt a few hours after the cell was assembled, and it gradually increased to about 0.45 millivolt during a period of two weeks. It varied between 0.41 and 0.52 millivolt during the next two weeks, at the end of which time the cell was disassembled. Temperature differences of as much as three degrees were observed from time to time among the three electrodes. By changing the position of the cell in the furnace the relative temperatures of the electrodes with respect to each other could be changed. However, whether the two zinc electrodes were at nearly the same temperature or whether the middle electrode was two degrees hotter or two degrees colder than the other one, the middle electrode was always negative with res ect to the outer one, and any changes in the magnitude of t&s potential difference resulting from such temperature changes could not be distinguished from the random fluctuations of this potential difference within the ranges already stated. It was concluded from these facts that the temperature of the zinc electrodes had relatively little influence on the cell potentials. In view of this conclusion, although the temperatures of all three electrodes were measured re ularly only those of the bismuth electrode are recorded in &able I. The potentials recorded in Table I are those of the cell comprising the bismuth electrode and the outside zinc electrode. The zinc electrode is negative with respect to the bismuth electrode, of course. The outer zinc electrode was chosen as reference electrode for pu oses of this tabulation mainly because contamination by Tiffusion from the bismuth electrode should have been less at the outer electrode than a t the middle one. I t may be borne in mind, however, ;that the choice is not very important because the maximum difference ever observed between the two zinc electrodes was 0.52 mv. It was found that the potential of the cell could be represented by the equation E = 0.6727 - 0.00055(1 - 450) (2, where the potential is in volts and t is the temperature in degrees centigrade. The last column in Table I represents the deviation of the measured value from the value that would be calculated by this equation. Many more readings than those included in the table were recorded. The values are representative. They were chosen by the arbitrary rule of taking the last measurement on the indicated day or, on days when the furnare current was readjusted, the last reading before this readjustment. The potentials of the Bi, Bi208, ZnClZ/ZnClz/ZnClz, Zn cell were very steady and easily measurable to one hundredth of a millivolt. As can be seen from the data in Table I and from equation 2, the cell temperature coefficient is large but the tern erature effects are quite reproducible and reversible. d s t of the readings recorded in Table I were made at least one day after any previous major temperature change. However, the cell potentials followed the temperature changes quite rapidly even without stirring. In two instances the potentials were read every few minutes while the temperature was changing at the rate of about 10' per hour. In one of these instances the temperature was rising, in the other it was falling. In neither case did the potential lag by more than a millivolt behind the values corresponding to equation 2. Stirring seldom was resorted to. It has
11 12 13 14 15 16 17 19 21 22 22 23 24 26
(2) "Reference Tables for Tl~errnocouplea," Circular 661, National Bureau of Standards, 1955.
Vol. 63
TABLE I CELLPOTENTIALS AT VARIOUS TEMPHRATURES Bob&,
(I
1,
OC.
E o b d . . V.
Eoalod.,b V.
Eded.
-
mv.
0.6408 -0.5 0,64029 507.9 .64066 509.0 ,6402 .5 504.7 .6426 .8 ,64182 507.4 .64131 .6411 .2 507.6 .64158 ,6410 .6 482.0 ,65434 ,6551 - .8 480.3 ,65543 ,6560 - .6 ,6556 - .6 481.1 ,65504 480.6 .65613 ,6559 .2 .67167 452.1 ,6716 .1 ,6649 .o 464.2 .66494 473.2 ,66033 ,6600 .3 ,6548 - .5 482.6 ,65431 ,6501 - .3 ,64983 491.1 500.4 ,64590 ,6450 .9 Since cell was aasembled. b Calculated by equation 2.
++ +
+ + + +
the disadvantage of promoting diffusionof bismut'h ions and temporarily aggravating the effect of any vertical temperature gradients in the furnace. When necessary, the stirring was done without exposing the electrodes to air. The potential of the bismuth electrode always fell with stirring. This may have been due at least in part to temporary unsaturation of the solution a t the bismuth surface by dilution with unsaturated electrolyte from above. In any case, normal potential values (values consistent with equation 2) were restored within less than one-half hour after stirring. The cell appears to be relatively non-polarizable. In one experiment a load of 1000 ohms was connected between the bismuth electrode and the adjacent zinc electrode. The potential difference between the bismuth electrode and the remote zinc electrode was observed. Before the load was applied this potential difference was 0.645 volt. Immediately upon application of the load it fell to 0.439 volt. It fell only 0.002 volt further during the 40 minutes the load was left on. Within one minute after the load was removed the measured potential difference had gone back to 0.644 volt. The temperature a t the start of this efrperiment was 501 and it fell by about one-half degree durlng the expenment. The current drawn by the load was 0.42 ma. and the electrode current density about 0.25 ma. cm.-*.
Summary Until the pha.se equilibria in the system BiBi2O3-Zn0-ZnCl2 have been studied it would seem impossible to relate the cell potentials reported here to free energies and entropies of formation. The following two statements summarize the conclusions to which these experiments lead. 1. The measured potentials of the cell Bi, Bi203,ZiiCla/ZnC12/ZnCI2,Zn are reproducible .with respect to temperature changes and after the flow of current. The reversible E of this cell a t 450" is 0.6727 0.0010 volt and dE/dT = -0.00055 f 0.00001 volt deg.-'. 3. Within the stated temperature range a t lead, either electrode of this cell should serve as a reliable reference electrode in cells whose electrolyte is mainly fused zinc chloride. Acknowledgment.-This work was donc under a research conttract with the Office of Ordnancc Rcsearch, Ordnance Corp, U. S. Army.
*
t
