Solutions are identified as follows: a solution containing 10^{-3} m chloride and 10^{-4} m iodide is identified as (pCl 3, pI 4), where pCl = -log [Cl^{-}]. Unless otherwise noted, pCl and pI refer to molal concentrations of electrolytes. The notation E(Cl, 1) refers to electrode potentials measured in solutions containing both Cl^{-} and I^{-}. The units of the electrical potential E(Cl, I) are always expressed in mV(sce), where sce is the saturated KCl calomel half-cell. The precision of a measurement or calculation is indicated by the number of significant figures in the numerical result.
Electrodes
The Ag/AgCl electrodes were manufactured by electrodeposition of AgCl onto chemically polished silver wire. The nominal purity of the commercially acquired silver was 0-999+. The wire diameter was 0-0508 cm. The silver wires were sealed into glass capillaries with epoxy resin, with varying lengths of wire protruding from the insulated region. The silver was anodized in 1-0 m KCl solution stirred continuously at 25^{o}C in conditions of room light for 60 min at 93 uA. Assuming 100% coulombic efficiency, each electrode was calculated to have 0.50 mg AgCl (3-5 micromol) in the deposited film. The surface area for each electrode was calculated from the length of the cylindrical wire exposed to anodizing solutions. Measurements of length were accurate to 0.02 cm. Electrode areas ranged from 0.06 to 0.13 cm^{2} and computed film thicknesses ranged from 6-8 to 13,um. Computed cds ranged from 0.7 to 1.4 mA/cm^{2}. The colour of electrodes anodized in this way was dark brownish-grey. Electrodes were soaked, unshorted, in solutions of 0.01 m KCl for 3-7 d prior to their use in the experiments. Potentials were determined for each electrode to 0.1 mV precision, and the standard error of the mean among the electrodes used was never more than 0.2 mV when measured in solutions of nominally pure KCl.
Solutions
Solutions were made from reagent grade chemicals and redistilled water. The KCl used was "Baker Analyzed" (J. T. Baker Chemical Co., Phillipsburg, N.J., U.S.A.) containing 0.001% iodide by weight. Contamination by Br- was 0.005%. Where required to maintain an ionic strength of 0.10 m, KNO_{3} was added to the solutions. The pH of all solutions was adjusted to 4.2 by the addition of HN0_{3}. The solutions were unbuffered. No attempt was made to exclude oxygen from the solutions.
Electrical recording
Potentials were measured against a commercially acquired saturated KCl calomel half-cell with fibre liquid junction. Since KCl was found to diffuse from the tip of the reference electrode at a rate of the order of 1 micromol/min, the reference electrode was left in solution only as long as required for the periodic potential measurements. In some experiments for which continuous records were required, the reference cell was connected to the solution through a KNO3 salt bridge.
Potentials were measured directly by analogue and digital recorders. The input of the recording apparatus was a field-effect-transistor probe compensated for leakage current. The input resistance of the probe was 10^{10} ohm and the leakage current was < 10^{-13} A. Two such channels were used for the recordings. Potentials from as many as 8 electrodes and the ground reference level were led in pairs to the input probes by means of relays with gold-plated contacts which could be operated either manually or automatically. The total series junction potential in the system was not known. Drift in the direct-coupled recording apparatus was corrected by referring to ground for each set of electrode measurements.
Potentials were measured by integrating digital voltmeters with gate durations of 100 ms. Usually, potential samples were measured for each pair of electrodes once per second for 5 s, and the average was computed. Electrodes were electrically floated between measurement intervals.
Mass spectrometric analysis
The relative I/Cl ratio in the electrode film was studied by a mass spectrographic technique. Ions were generated by a spark source in a spectrometer, Type 21-110 (Consolidated Electrodynamics Corp., Pasadena, California, U.S.A.). The I/Cl atomic ratio on the spectrographic plates was determined by a computer-controlled microdensitometer. The efficiency of transmission and detection of chlorine and iodine are unknown, but it is assumed that within the reported range, the relative ratio on the spectrographic plates was proportional to the ratio in the electrode film.^{5}
X-ray diffraction
The electrodes were well suited geometrically for analysis in a Debye-Scherrer powder camera. Samples were irradiated by X-rays of wavelength 1-54 Angstroms from a copper source. Measurement of the diffraction lines was facilitated by use of a microdensitometer. Systematic errors in the method were corrected and lattice constants for the fcc structure of AgCl were computed by linear extrapolation to 180^{o} (2Ø) against the function^{8}
Experiments were performed with solutions in 150-ml open beakers in a water bath maintained at 25.0^{o}C. Both the solutions and the water bath were stirred by Teflon covered magnetic bars driven from below by a common magnet rotated by an air motor.
The concentrations of I^{-} and Cl^{-} varied over wide ranges in the experiments. Electrode potentials were measured in solutions, initially 100.0 ml in volume, having pCl values of 1, 2, or 3. For each solution, the pI value was varied by the addition of known quantities of a solution containing KI, KCl and KNO, such that pCl and ionic strength remained constant and pI could be computed. Initial pI was calculated allowing for I^{-} contamination of KCl. The time of each introduction of I^{-} was recorded, and the electrode potential was measured at regular intervals.
Pre-transformation
In solutions containing Cl^{-} and I^{-}, electrode
potentials E(Cl, I) were always negative to those in pure Cl^{-}
solutions, and decreased with increasing [I^{-}]. Figure 1 is a
plot of E(Cl, 1) versus pI. Each value for potential is the average for
10-16 electrodes, each determined 10 min after the addition of I^{-}.
Since the ionic strength of all solutions was maintained at a constant
value, the activity coefficients for I^{-}, Cl^{-}, N03-
and K+ should be nearly the same according to the Debye-Huckel equation.
The molal concentrations given in this report may be converted to activities
by adding 0.12 to the pI or pCl values. Each of the 3 sets of points plotted
in Fig. 1 represents measurements made in solutions of the specified pCl.
Included in each set of points are both time-dependent and time-invariant
data. In the range pI 6—pI 4, E(Cl, I) reached a time-invariant value within
2 min after the time of introduction of I^{-}. Potentials were
stable to within 0.1 mV from about 2 to at least 20 min. The lines shown
in Fig. 1 are fitted to the time-invariant data by the method of least
squares.
In the range pI 4—pI 3, the potentials departed from the linear
approximation of the time-invariant data; moreover, E(Cl, I) varied linearly
with time, and the slope of the relation (deldt) increased with diminishing
pI (Fig. 2).
The points in Fig. 2, reading from left to right and top to bottom, represent potentials sampled consecutively at 5-min intervals. The pI value was changed every 20 min during the experiment. There are too few points to characterize completely the relation between dE/dt and pI; however, the slope of the linear approximation is observed to increase with increasing pCl. In other words, high concentrations of Cl^{-} tend to diminish the dependence of dE/dt on pI.
Transformation
As the I^{-} concentration was increased the potentials
developed increasingly prominent, slow, irregular oscillations. In the
range of pI 4-pI 3 the peak to peak amplitude of these variations was of
the order of 1-2 mV, and the period was of the order of 3-5 s. In Fig.
3, the analogue record of a single electrode potential plotted against
time, the irregular oscillations can be seen superimposed upon the potential,
which is linearly related to time for the first part of the record. The
latter part of the record illustrates an increasingly negative dE/dt.
Figure 4 is a plot of digital potential measurements for the same
electrode extended over a longer time. The coefficients of the equation
plotted on the graph were empirically derived from the data. The computed
values of the exponential term of the equation with the coefficients given
were less than 1 mV for times from 0 to 6.6 min. The S-shaped character
of the plot is typical of the electrical event hereafter referred to as
transformation. The step seen near the end of the transformation
curve is a typical part of the overall potential/time relationship. The
electrode of Figs. 3 and 4 had been exposed to solutions (pCl 2, pI <
4) for 100 min prior to the interval shown on the graph.
Transformation marks the transition from one to another set of functional characteristics. Before the onset of transformation the characteristics of the electrodes approach those of the ideal Ag/AgCl electrode when measured in pure KCl solutions. By contrast, fully transformed electrodes exhibit the functional characteristics of Ag/Agl electrodes; their potentials are reproducible to within 0-5 mV and vary linearly with pI with a slope of -59 mV per 10-fold increase in [I^{-}].
One way to observe the effect of transformation on electrodes
is illustrated in Fig. 5. Potentials of pre-transformed and transformed
electrodes were compared with those of a nominally pure Ag/AgCl electrode
in KCl solutions containing no I^{-}. It can be seen that pretransformed
electrodes are identical to the Ag/AgCl electrode despite their prior exposure
to solutions containing I^{-}. By comparison, the potentials of
the transformed electrodes deviate substantially from those of the Ag/AgCl
electrode.
In solutions of pI 4-pI 3, each electrode potential varied linearly
with time for a distinct interval prior to the onset of the transformation
curve. The period of electrode exposure to these solutions terminating
with the inflexion of the transformation curve is defined as the latent
period of transformation, or latency. Electrodes exposed to the same experimental
conditions exhibiting small variances in E(Cl, I) during the pre-transformation
studies nonetheless showed wide variability in their latencies. One physical
parameter found to correlate fairly well with the latent period for transformation,
however, was the surface area. In Fig. 6, latency is plotted against the
reciprocal of area for each of the three pCl values. As the area increased,
the latent period of transformation decreased. Since the initial film volume
was the same for all electrodes, the latency may also be interpreted as
being directly related to film thickness.
Electrode composition
As the exposure of electrodes to I^{-} proceeded, the original brownish grey colour of the silver chloride electrode was replaced by an increasingly vivid yellow. Electrodes began to turn yellow within 5 min when exposed to solutions containing pI < 4. On electrodes that had been completely transformed, the yellow coating was very thick, of the order of 0-5 mm, and was sufficiently friable to be brushed off easily. Microscopic study of the yellow powder brushed from electrodes revealed polycrystalline aggregates of hexagonal prisms. The maximum crystalline dimensions ranged up to 50 micrometer.
Identification of AgCl and Agl in the electrodes was made by X-ray
diffraction, and the relative amounts of I and Cl were determined by mass
spectrometry. The change in electrode composition during transformation
is summarized in Table 1. It can be seen from the table that the relative
I/Cl ratio increased by about 100-fold within the period of transformation.
The X-ray diffraction studies indicated that the I^{-}and Cl^{-}
in the electrode were present as their silver salts.
The lattice constant of the AgCl crystal was computed for each of 9 electrodes to discover whether there was evidence of Ag(Cl, 1) homogeneous solid solution formation. The electrodes studied included those with no AgI present as well as those with mixed AgCl and Agl crystalline structures. The AgCl lattice constant was 5.55 A for each electrode; thus, there was no evidence of lattice distortion despite the presence of AgI.
It was recognized early in the course of the experiments that the electrode potential is affected by several physical variables in addition to pCl and pI. Preliminary results suggested that these variables include ionic strength, temperature, pH, stirring rate, and electrode area. In addition, time was regarded as a parameter in all experiments. In this set of experiments, temperature, pH and ionic strength were held constant. The effect of stirring could not be determined quantitatively, but experiments with unstirred solutions demonstrated that the qualitative effects described here were not altered by stirring. Although it was not possible to maintain constant stirring conditions for all electrodes, stirred solutions were used because it was observed that the time-dependent effects of interest were enhanced. Electrode surface areas were allowed to vary by a factor of two. As a result, the inverse relation between the latent period of transformation and area was observed.
Silver/silver-chloride electrodes exposed to solutions containing Cl^{-} and I^{-} (pI < 4) become transformed to Ag/Agl electrodes after variable latent periods. The electrode potentials accompanying this process are sufficiently reproducible to substantiate the conclusion that they are functions of pCl, pI and time. The definition of time-invariant potentials depended upon the resolution of the experimental method. The minimum detectable [dE/dt] was 0.005 mV/min. Therefore, time-invariant potentials are defined as those having [dE/dtl less than this value. Potentials were time-invariant within the range of pI 6-pI 4, and decreased linearly with pI for each of three pCl values, the slopes increasing with increasing pCl.
Potentials having detectable time-dependence were observed in solutions of pI 4 - pI 3. Typically, in solutions of unchanging pCl and pI, [dE/dt] was constant initially; hence during this period, [dE/dtl was determined by pI and pCl. Later, [dE/dtl increased, reached a maximum, then decreased to zero. In this interval, the potential was S-shaped with time and the electrode was transformed from a Ag/AgCl electrode to a Ag/Agl electrode, as defined on a functional basis. The lattice constant for AgCl was determined by X-ray diffraction to discover whether there was any evidence of solid solution formation in the process of iodide-chloride exchange. Since the lattice constant was 5.55 A whether or not AgI was present, it is concluded that iodine does not appreciably enter a homogeneous solid solution phase with AgCl. The Agl structure identified by X-ray diffraction was hexagonal, and AgCl was fcc. Under certain conditions of precipitation AgI can assume a fcc structure having a lattice constant of 6.48 Angstrom, but there was no evidence of this structure in the X-ray diffraction patterns studied. Hexagonal prism crystals, presumably AgI, were observed with a compound microscope in the yellow powdery coating surrounding the transformed electrodes. Evidently AgI is formed in the solution phase from Ag+ made available by the solution of AgCl.
In solutions ranging from pI 4 - pI 3, prominent oscillations were observed superimposed upon the electrode potentials during the latent period. These oscillations disappeared after transformation. It is suggested that the fluctuations in potential reflect cyclic precipitation of Agl, resulting in variation of [Ag+] and [I^{-}] between the limits of the metastable saturation region for AgI. It is unlikely that the fluctuations are related to those reported under conditions of anodizing of Ag/AgCl electrodes in KCl solutions,^{7} since no current passed through the electrodes.
The experimental evidence reported in this paper has been interpreted
qualitatively. It is concluded that electrode potentials are related to
chloride and iodide abundances, in both solution and crystalline phases.
The quantitative description of the potential in terms of its physical
variables will depend upon experimental resolution of processes at the
solution/solid interface.
This work was performed under the auspices of the United States Atomic Energy Commission.
2. G. Pinching and R. G. Bates, J. Res. Natl. Bur. Stand. 37, 311 (1946).
3. W. Jaenicke, Z. Elektrochem. 57, 843 (1953).
4. D. M. Yee and D. E. Watson, Proc. Annual Conf. Engineering Med. Biol. 9,154 (1967).
5. R. E. Honig, in Mass Spectrometric Analysis of Solids, ed. A. J. Ahern, Ch. 2. Elsevier, New York (1966).
6. H. S. Peiser, H. P. Rooksby and A. J. C. Wilson (eds), X-Ray Diffraction by Polycrystalline Materials, p. 644. Reinhold, New York (1960).
7. H. Lal, H. R. Thirsk and W. F. K. Wynne-Jones, Trans. Faraday
Soc. 47, 70 (1951).