Three gold sensing electrodes were made from 3 mm gold rod. The rod was cut with a diamond wafering saw into 3 mm lengths, soldered to insulated copper wire, placed inside 9 cm lengths of 1 cm (outside diameter) Plexiglas tubing, and cast inside the tubing with epoxy. After the epoxy was set, the gold was polished with 600 grit emery paper. The exposed surface area of gold in the finished electrodes was 0.0707 cm².
A gel-filled, epoxy-body Ag/AgCl electrode was used as a reference electrode. The accuracy of this electrode was checked in Zobell standard redox solution using a freshly polished gold electrode. Unless otherwise indicated, all potentials are reported with respect to the Ag/AgCl reference potential.
ORP Probe Cleaning Procedures
Stability of the readings provided by the gold electrodes was measured in a molybdenite flotation slurry. Molybdenite flotation concentrate was slurried in deionized water at 10% by weight MoS2 pulp density in a Wemco two-liter glass flotation cell. The slurry was agitated by the flotation machine for approximately 8 h each day for the duration of the test. The slurry was open to the air during testing, but air was not introduced through the flotation machine. Calcium hydroxide was added to the slurry to maintain pH 9.2; NaHS·xH2O was added to maintain approximately -0.1 V measured against the AgCl reference electrode.
Three gold electrodes were prepared and tested simultaneously. The slurry potential was measured with each gold electrode against the AgCl reference electrode using a digital multimeter. One gold electrode was connected to a strip chart recorder and immersed continuously in the slurry for continuous reading of the potential. The second gold electrode was immersed continuously in the slurry; however, potential measurements were made only occasionally using this electrode. This electrode was tested for electrolytic cleaning during the test. The third electrode was used as a standard for comparing the response of the other two electrodes and was stored in deionized water except when used for reading redox potentials of the slurry.
Anodic cleaning of one of the electrodes was accomplished with a potentiostat. The gold electrode functioned as the working electrode of the potentiostat. A graphite rod immersed in the slurry functioned as the counter electrode. The Ag/AgCl electrode functioned as the reference electrode for the potentiostat. Currents and voltages from the potentiostat were recorded during and after the cleaning cycles. Following the test, the tip of each of the three gold electrodes was cut off with a wafering saw and examined under a scanning electron microscope.
Laboratory research was undertaken to better understand fouling of redox-sensing electrodes in mineral slurries, and develop methods to avoid fouling to allow for reliable redox potential control in flotation circuits.
In the laboratory tests, two gold electrodes were immersed in a molybdenite slurry. The potential as sensed by the first gold electrode (constantly immersed in slurry and connected to strip chart recorder) is plotted in figure 1. For comparison, the potential as measured by the clean gold electrode is also plotted. An observable coating accumulated on the first electrode within 14 d. The potentials sensed by the two gold electrodes, however, did not differ greatly during the test period. This agrees with previous experience in flotation plants by Bureau researchers and others (Labonte, 1988), where gold electrodes performed more reliably than expected.
After 20 d immersed in the slurry, the tip of the electrode was cut off and examined under the SEM. A photomicrograph of the electrode tip is shown in figure 2. The gold cannot easily be distinguished in the micrographs because of the coating. The gold occupies the lower portion of figure 2A; the box in the photograph straddles the edge of the gold. In the enlargement in figure 2B, a crack running nearly horizontally through the middle of the photo marks the interface between the gold (lower portion of the photograph) and the epoxy (upper portion) in the electrode. The electrode tip was uniformly coated with foreign material after 20 d in the flotation slurry. The coating was analyzed in the electron microscope by energy dispersive x-ray (EDX) analysis and found to contain a large amount of calcium, a small amount of sulfur, and trace amounts of copper and iron. The coating also fizzed in hydrochloric acid. Based on this information, the coating appears to be primarily CaCO3 with a minor amount of CaSO4.
The potential sensed by the second gold electrode (constantly immersed in slurry but not connected to strip chart recorder) is plotted in figure 3. A visible coating accumulated on this electrode within 14 d, as happened with the first electrode. At 14 d, electrolytic cleaning of the second electrode was attempted. Zhou and Ghander (1991) recommend anodic polarization to condition fresh electrodes. Following this recommendation, the gold electrode was anodically polarized to a potential of 1.0 V with respect to the standard hydrogen electrode (+800 mV vs. Ag/AgCl) for 10 min. No visible change in the coating at the tip of the electrode was observable. The potential sensed by the electrode following this cleaning treatment was -22.8 mV compared with -93.4 mV for the clean gold electrode. The fouled electrode was also immersed in Zobell solution, where it produced a reading of +91 mV compared with +214 mV for the clean gold electrode. Thus, 800 mV vs. Ag/AgCl was deemed insufficient for cleaning fouled gold electrodes.
In a modification of this treatment, the fouled electrode was anodically cleaned for 10 min at 1.2 V vs. Ag/AgCl. The overall cell voltage for the cleaning was approximately 8 V. The current density averaged 9.1 mA for the electrode, or approximately 1,300 A/m². The potential of the electrode was monitored following the cleaning treatment until the potential stabilized. The potential readings are plotted in figure 4.
Approximately 30 min was required to reach a stable potential after cleaning at 1.2V. The stabilized potential reading (-93 mV) agreed well with the potential of the standard gold electrode of -88 mV. Visual inspection of the electrode tip indicated that the coating fouling the electrode tip had been removed. Several mechanisms are possible that would result in removal of the scale. The strong oxidation would cause anodic dissolution of metal sulfides in the scale. Anodic dissociation of water would produce gaseous oxygen, which would force scale from the electrode surface.
This polarization treatment removed the coating on the electrode tip and restored proper functioning of the electrode; however, the time required to attain stable readings (=30 min) is more than would be desirable in plant operation, where the electrode may be needed for continuous, on-line process control. In an attempt to reduce the stabilization time following potentiostatic cleaning, a potentiodynamic conditioning procedure described by Zhou and Chander (1991) was tested. The gold electrode in the molybdenite slurry was polarized for 10 min at 1.2 V as before. The anodic treatment was followed by 45 s of potentiodynamic cycling, during which the potential was oscillated between -600 mV and +800 mV at 400 mV/s scan rate; six potentiodynamic cycles were completed.
Following this combined potentiostatic-potentiodynamic treatment, the potential of the gold electrode was monitored. The plots in figure 4 show that the potential of the treated electrode stabilized within 1 min following the cycling treatment compared to 30 min stabilization time required without the potentiodynamic treatment. The potential sensed by the gold electrode in both the molybdenite slurry and the Zobell solution after cycling was found to be acceptable.
Following cleaning, the tip of the cleaned electrode was cut from the electrode and examined under the SEM. The photomicrographs in figure 5 show the gold to be free of calcium deposits. The area free of coating extended approximately 10 µm beyond the edge of the gold. The bulk of the electrically insulating epoxy material, however, remained coated with calcarious material. From EDX measurements, the composition of the coating was similar to that of the material observed in figure 2. For comparison, the tip of a freshly polished electrode was cut from the electrode and examined under the SEM; a photomicrograph of this electrode is shown in figure 6. The only observable difference is more debris from the polishing on the freshly polished electrode. Thus, electrolytic cleaning of the gold electrode followed by a brief period of rapidly cycled potential appears beneficial in maintaining gold redox sensing electrodes in flotation slurries.
Assuring a stable potential from the AgCl reference electrode in operation was critical to this test. The potential of the reference electrode was checked using a clean gold electrode in Zobell solution. Readings were stable over the course of several hours, but during 12 days of testing, the AgCl reference electrode drifted from +202 mV in Zobell solution to +220 mV. In general, however, the gel-filled, epoxy-body reference electrode appears to be robust in the flotation system.
Fouling of redox-sensing electrodes is a problem in monitoring and controlling the electrochemistry of sulfide flotation slurries. Recent laboratory work at the USBM has shown that gold electrodes may be cleaned by anodically polarizing the electrode to 1.2 V vs. Ag/AgCl to remove the accumulated scale. To reduce the time required to attain stable readings following cleaning, a 45 s period of potentiodynamic cycling was included in the cleaning procedure. Examination under the electron microscope showed that this method is effective in removing accumulated scale from gold redox-sensing electrodes. Electrochemical measurements following cleaning and cycling showed rapid return of the sensing electrode to stable, reliable potential readings. Work is continuing to assess the effectiveness of this procedure over an extended period of time and to test the procedure for cleaning of gold redox-sensing probes in control circuits at commercial flotation plants.
The optimum cleaning method will likely depend on pH, Eh, and chemical composition of the specific mineral systems.