GB2057695A - Method and Apparatus for Measuring the Oxygen Potential of an Ionic Conducting Melt - Google Patents

Abstract

A method and apparatus for measuring the oxygen potential of ionic melts such as sodium silicate and soda-lime glasses are described wherein a reference electrode 10 and an indicator electrode 11 are coupled by a potentiometer and inserted into the melt with the reference and indicator electrodes 10, 11 separated by a distance of at least 3 mm and the indicator electrode 11 insulated 13, 14 at its point of entry into the melt to prevent the formation of a three phase boundary between the melt, the electrode and the gas atmosphere above the melt. The method and apparatus avoid the problem of oxygen transfer through the solid electrolyte of the reference electrode which prevented previously known oxygen probes from measuring the oxygen potential of ionic melts. The reference electrode may be a conventional solid electrolyte reference electrode and the indicator electrode 11 comprises a platinum or molybdenum wire 12 which extends through alumina isolator 13 cemented into alumina tube 14. <IMAGE>

Description

SPECIFICATION Method and Apparatus for Measuring the Oxygen Potential of an ionic Conducting Melt The present invention relates to a method of measuring the oxygen potential of an ionic conducting melt and to an oxygen probe for use in the method.

In industries concerned with handling ionic conducting melts, such as glass, it is desirable to know the oxygen concentration in the melt. While it is not possible to directly measure the oxygen concentration it is possible to measure the oxygen potential directly. The oxygen potential is equal to the oxygen concentration multiplied by an activity coefficient for the particular melt. As the oxygen concentration of an ionic conducting melt, such as glass, will affect such things as the melting rate of the glass, gas solubility within the melt, the colour of the glass and the appearance of unwanted inclusions, a knowledge of the oxygen potential bf the glass, and thus an indication of the oxygen concentration, is of great help in controlling processes using or affecting the melt.

Oxygen probes are known for the measurement of the oxygen potential of non-conductive materials such as gasses or for electron conductive liquid materials such as steel, however, such probes are not suitable for the measurement of the oxygen potential of ionic conductive melts. These conventional oxygen probes include a solid electrolyte reference electrode around which is wrapped an indicator electrode. Such a probe is unsuitable for the measurement of the oxygen potential of ionic conducting melts due to oxygen transfer through the solid electrolyte of the reference electrode and due to the potential developed of a three phase boundary between the indicator electrode, the melt and the atmosphere above the melt.The present invention is designed to substantially overcome these deficiencies and to enable the simple and accurate measurement of the oxygen potential of an ionic conducting melt.

The present invention consists in a method for measuring the oxygen potential of an ionic conducting melt comprising inserting into the melt a reference electrode and an indicator electrode and measuring the electrical potential therebetween, characterized in that the reference electrode and the indicator electrode are separated within the melt by a distance of at least 3 mm. and in that the indicator electrode is insulated at its point of entry into the melt to prevent the formation of a three phase boundary between the electrode, the melt and a gas atmosphere above the melt.

In another aspect the present invention consists in means for measuring the oxygen potential of an ionic conducting melt comprising a reference electrode electrically connected through a potentiometer to an indicator electrode spaced more than 3 mm. from the reference electrode, the indicator electrode being disposed in an insulating support such that the indicator electrode may be inserted into an ionic melt without a three phase boundary existing between the electrode, the melt and a gas atmosphere above the melt.

The reference electrode is of conventional design and is preferably a solid electrolyte electrode such as a stabilised zirconia electrolyte electrode. The preferred stabilising material for the zirconia is yttria, however, magnesia or calcia may be used. Other known oxygen reference electrodes could, however, be used.

The indicator electrode is preferably made from platinum or molybdenum. The insulation applied to the indicator electrode may be any suitable heat resistant insulating material preferably a ceramic material such as alumina. In order to totally preclude a three phase boundary it is advantageous to fuse the insulating material about the electrode contact wire. This may be done by slip casting an alumina rod, embedding the contact wire, or the contact itself, in the alumina rod and firing it at 1 6O00C.

The electrical potential between the reference electrode and the indicator electrode are best measured using a high impedance potentiometer or multimeter.

The indicator electrode should be placed in the melt at least 3 mm. from the reference electrode to avoid false readings due to oxygen being transferred through the solid electrolyte of the reference electrode. It has been found that the oxygen gradient set up by the migration of oxygen passing through the electrolyte falls off rapidly as the distance between the indicator and reference electrodes increases. There is, however, in practice no limit to the maximum distance between the reference electrode and the indicator electrode. The spacing of the indicator electrode from the reference electrode also allows a plurality of indicator electrodes to be placed in the melt at spaced locations and compared with a single reference electrode.

A surprising discovery made by the present inventors is that when the reference electrode and the indicator electrode are spaced apart in an ionic conducting melt the oxygen potential measured is that of the indicator electrode/melt interphase and not the reference electrode/melt interphase. If this were not so the oxygen transfer problem would not have been eliminated.

Hereinafter given by way of example only is a preferred embodiment of this invention described with reference to the accompanying drawings in which: Fig. 1 is a vertical sectional view of a reference electrode which may form part of the means according to this invention.

Fig. 2 is a vertical sectional view of one form of indicator electrode which may form part of the means according to this invention.

Fig. 3 is a vertical sectional view of another form of indicator electrode which may form part of the means according to the invention.

The reference electrode 10 is of generally conventional configuration and includes a Zirconia tube 21 formed from Zirconia 8% by weight of Yttria as a stabiliser. The tube 21 is closed at its lower end and the bottom inside surface of the tube 21 is provided with a porous platinum layer 22 formed by coating the surface with "Hanovia" (Registered Trade Mark) black platinum paste and heating the tube to develop the porous platinum layer.

The tube 21 is connected at its open end to an alumina tube 23 by a phosphate bonded cement joint 24. The upper end of the alumina tube 23 is joined to a brass tube 25 by an epoxy resin joint 26.

The upper end of the brass tube 25 is threaded and carries an annular collar 27 which encloses an 0ring 28 and causes it to sealingly engage about a further brass tube 29 which is disposed around a 4bore alumina tube 30 to which it is sealingly connected by epoxy resin joints 31 and 32.

The alumina tube 30 extends along the full length of the alumina tube 23 and terminates just short of the porous platinum layer 22. A platinum electrode 33 extends down one bore of the tube 30 and is connected to the porous platinum layer 2 by a platinum gauze plug 34 a platinum/platinum plus 13% rhodium thermocouple 35 is also disposed in the alumina tube 30. The thermocouple junction 36 occurs just short of the lower end of the alumina tube 30. The point of entry of the electrode 33 and the thermocouple 35 into the alumina tube 30 is sealed with an epoxy resin seal 36.

Gas inlet 37 joins tube 29 radially of its axis while gas outlet 38 similarly joins tube 25. The gas inlet 37 and gas outlet 38 allow the reference gas in the reference electrode 10 to be maintained at constant compositions and pressure during use of the electrode.

The indicator electrode 11 shown in Figure 2 comprises a platinum wire 12 which extends through a fixed alumina isolator 13 and is wound around the free end of the isolator 13. The other end of the isolator 13 is cemented into an alumina tube 14 by a bead of phosphate bonded alumina 15. The indicator electrode 11 of Figure 3 is similar in design, and similar parts have identical numbers to those appearing in Figure 2, with the exception that the platinum wire 12, instead of being wound about the alumina isolator 13, is formed into a self supporting helical coil coaxial with the isolator.

The indicator electrode 11 of Figure 2 has the advantage that the platinum wire 12 is well supported by the isolator 13 and is unlikely to break off if the indicator electrode is used to stir the melt.

This arrangement also has the advantage that it can be easily cleaned to remove glass from a previous melt. It has been found however that when the indicator electrode 11 is used in ionic melts having a low iron content there is a drift in the measured oxygen potential with time in an unstirred melt. This drift is caused by a reaction between the alumina of the isolator and the glass which changes the oxygen potential of the melt. The indicator electrode 11 of Figure 2 is more susceptible to this problem than is the indicator electrode 11 of Figure 3.

The measurement of the oxygen potential of ionic melts using the process of the present invention will now be described.

Example 1 Sodium disilicate glasses (1 80 g) containing 1.5-2 wt.% total Fe were made from acid washed sand and analytical grade anhydrous sodium carbonate. Iron was added as either ferric oxide or ferrous oxalate. In some melts high purity carbon was added for further reduction of the glass. The raw materials were mixed, then melted in either alumina or platinum crucibles. Homogeneity of the melt was achieved by stirring and/or bubbling through the melt. The glass melt was kept at the desired temperatures for at least 1 5 hours before measurements were made. The Y2O3-ZrO2 reference electrodes have been described previously in this specification with reference to Figure 1. The indicator electrode was described with reference to Figure 2.The isolator has a 10 mm outside diameter and the platinum wire was 80 mm long and wound into a coil of 10 mm length. Before use the platinum wire was cleaned in a reducing flame then washed with distilled water. Care had to be taken when this electrode was immersed in the melt to ensure that a three-phase boundary (atmosphere-melt-Pt) was not created.

EMF measurements were carried out with a high impedance multimeter(l O'OQ) and data logger unit (5 x 109Q). Total iron was analysed by atomic adsorption spectrophotometers. About 0--2% alumina contamination from the crucible was also detected by the same analysis. Ferrous iron analyses (in triplicate) were carried out on powdered samples of 1.0 g dissolved in HF-H2SO4 under high purity nitrogen. Reproducibility was +3% for this analysis.

With the experimental arrangement employed, up to 3 separate indicator electrodes and 2 bridging probes could be tested at the same time. In any measurement at least 2 reference electrodes were used to check the gas reference potential (air or CO-CO2). This procedure was applied to ensure that any change of potential of the probes was due to the indicator, and not the reference electrode.

Before any measurement was made with the oxygen probe, the atmosphere above the melt was changed and EMF variation observed. If the EMF was found to be atmosphere dependent, i.e. a mixed potential has been measured, a new electrode was used. This test was repeated at intervals throughout each experimental run.

Results In work carried out in earlier studies by Johnston (J. Am. Ceram. Soc. 47(4), 198-201(1964)) glass oxygen potential was determined by equilibrating the melt with an atmosphere of known oxygen partial pressure and then relating this to the measured Fe2+/Fe3+ weight ratio. In this present study the oxygen potential in the glass was directly determined and related to the Fe2+/Fe3+ ratio without the need of equilibrating the melt and atmosphere. Reference electrodes of the same oxygen potential were found to come to O mV (i.e. equilibrated) in less than 30 min. Even faster response times were observed with the indicator electrodes where stable EMF's were achieved in less than 5 min. The EMF of the probes remained constant for at least 24 h.EMF results (with respect to air reference) at different temperatures and analysed Fe2+/Fe3+ ratios are tabulated in Table 1.

Table 1 Results Measured by Oxygen Probes Log Po2 Log E T corrected Run No. Fe2+/Fe3+ (+3m V) { K) to 1 1100 CC 1 -1.84 8 1373 -0.80 2 -1.66 55 1273 -1.32 3 -1.53 75 1339 -1.66 4 -1.58 85 1400 -2.01 5 -1.44 90 1362 -1.97 6 -1.41 148 1368 -2.84 7 -1.13 150 1354 -2.83 8 -0.423 318 1378 -5.35 9 -0.298 425 1387 -6.91 10 -0.0377 492 1301 -7.99 The data obtained during the present study and those from Johnston's have been analysed by Multiple Regression using the following expression:

a, b, c, are listed in Table 2, with other statistical coefficients.

Also listed in Table 2 are the regression coefficients obtained by combining the data of present study (1000-1100 C) with those of Johnston (1100-1450 ). Table 2 Results from Regression Analysis Slope of log Po2 vs Multiple Standard Uncertainty Fe2+ correlation error of Range for log Coefficients coefficient estimate E(mV) Fe3+ Work a b c R 1100 C 1300 C //c Johnston 1.927 -5332.5 0.232 0.998 0.05 #14 #16 4.31 This study -0.5514 -1974.2 -0.254 0.985 0.12 #31 #36 3.94 Combined data 1.791 -5131.3 -0.236 0.995 0.087 #25 #28 4.24 Multiple Regression Analysis of Johnston's data yielded lines of slot 4.31. Data from the present oxygen probe yielded lines of slope 3.94 (compared with theoretical value of 4).

The excellent agreement between the results obtained by direct measurement using the oxygen probe and Johnston's equilibrium data indicates the accuracy of Johnston's work and more importantly the value of the oxygen probe as a device for the rapid determination of oxygen potential in molten glasses.

The oxygen probe according to this invention was also found to be useful for the measurement of the oxygen potential of commercial soda-lime glasses. It was found that the present invention performed as well with the soda-lime glasses as it did with sodium silicate glasses although measurement temperatures had to be increased slightly to allow more rapid equilibration of the system.

Oxygen potential measurements were obtained for commercial amber, green and white (colourless) glasses and these measured values were found to correspond with published data. In carrying out these measurements it was found that the measured oxygen potential in commercial glasses drifted slowly with time due to a reaction of the alumina isolator of the indicator electrode with the glass. Such changes were not observed with sodium silicate glasses because their high iron content provided an oxygen "buffer" which resisted changes in oxygen potential.

The observed drift of oxygen potential is considered to be of little significance in commercial glass tanks where the continual flow of new glass would ensure that correct oxygen potentials are always being measured.

Claims (10)

Claims

1. A method for measuring the oxygen potential of an ionic conducting melt comprising inserting into the melt a reference electrode and an indicator electrode and measuring the electrical potential therebetween, characterised in that the reference electrode and the indicator electrode are separated within the melt by a distance of at least 3 mm. and in that the indicator electrode is insulated at its point of entry into the melt to prevent the formation of a three phase boundary between the electrode, the melt and a gas atmosphere above the melt.

2. A method as claimed in claim 1 in which the indicator electrode is made from platinum or molybdenum embedded in a suitable heat resistant insulating support.

3. A method as claimed in claim 2 in which the platinum or molybdenum electrode is fused into an insulating support formed of a ceramic material.

4. A method as claimed in claim 3 in which the ceramic material is alumina.

5. Means for measuring the oxygen potential of an ionic conducting melt comprising a reference electrode electrically connected through a potentiometer to an indicator electrode spaced more than 3 mm. from the reference electrode, the indicator electrode being disposed in an insulating support such that the indicator electrode may be inserted into an ionic melt without a three phase boundary existing between the electrode, the melt and a gas atmosphere above the melt.

6. Means as claimed in claim 5 in which the indicator electrode is made from platinum or molybdenum embedded in a suitable heat resistant insulating support.

7. Means as claimed in claim 6 in which the platinum or molybdenum electrode is fused into an insulating support formed of a ceramic material.

8. Means as claimed in claim 7 in which the ceramic material is alumina.

9. A method for measuring the oxygen potential of an ionic conducting melt substantially as hereinbefore described with reference to, and as illustrated in, the accompanying drawings.

10. Means for measuring the oxygen potential of an ionic conducting melt substantially as hereinbefore described with reference to, and as illustrated in, the accompanying drawings.