Experimental

The experiments were performed under a flow of 99.99 volume per cent oxygen in a high vacuum apparatus, shown schematically in Figure 1. The pressures were adjusted by a needle valve and were measured by an ionisation gauge, and the temperature of the inductively heated specimen was measured by an optical micropyrometer.

The weight change of the iridium discs was determined by a suspension balance having an accuracy of 1μN. The oxidation anneals were interrupted for weighing because the electromagnetic forces did not allow continuous weighing during the induction heating. The oxidation periods were chosen in a manner which gave reliable readings of the weight changes.

Polycrystalline iridium discs 10 mm diameter and 1 mm thick were employed as specimens, and before the experiments they were cleaned and then recrystallised and degassed by a vacumm heat treatment at 1800°C for 15 minutes. When the same sample was used for different runs it was vacuum annealed, at 1800°C for 10 minutes, before each series, to give defined starting conditions. An analysis of the impurities found in the samples is given in the Table.

Results

The weight change of the iridium specimens as a function of oxidation time is shown in Figure 2. It can be seen that the linear weight loss strongly increases with increaseing pressure, at a constant temperature of 2000°C; and with increasing temperature, at a pressure of 1 pascal. From these curves the volatilisation rates and recession rates, respectively, of iridium have been calculated. In Figure 3 these rates are plotted versus 1/temperature. Each line in Figure 2 is represented by one point in Figure 3. The curves show a continuous increase with rising temperature and at lower pressures approach the straight line plot of the evaporation of the pure metal in vacuum. The temperature dependence of the oxidation process lies in the range of 125 kJ/mol. Figure 3 shows that a marked increase of the iridium losses exists when it is annealed in oxygen atmospheres, which extends upwards to several orders of magnitudes, especially in the lower temperature range. In the upper temperature range the increase of the iridium losses by the oxygen atmosphere is less pronounced.

Figure 4 shows the iridium weight loss plotted against the oxygen pressure. The pressure dependence follows the relation v ∝ p^0.3 between 1 and 100 pascals and at temperatures between 1500 and 2000°C. At lower pressures and at temperatures greater than 2000°C the recession rate becomes pressure independent, because here the pure iridium metal evaporation predominates. Between 1500 and 1900°C and at lower pressures the rate is proportional to the pressure. In Figure 4 the pure metal evaporation rate at 1900 and 2000°C is plotted additionally at the ordinate.

Discussion

From the gravimetrical determination of the iridium losses in the present work no direct conclusion can be drawn on the mechanism of the reaction between oxygen and iridium. The weight change gives the iridium losses due to the pure iridium metal evaporation together with the iridium component of the volatilising iridium oxides. The oxygen component is not determined because in the steady state between the oxygen sorption from the gas phase and the evaporation of the volatile oxides this part is continuously compensated.

By mass spectrometric investigations (9) the oxides IrO₂(g) and IrO₃(g) have been detected by the Knudsen cell technique and partial pressures in the range of 10 pascals within the cell. The oxide evaporation from a free iridium surface in the non-equilibrium kinetically controlled process has not been investigated up to now as it has been done in similarly reacting systems, for example molybdenum (10), tungsten (11) and rhenium (12). In analogy to the rhenium system the reactions between oxygen and iridium at high temperatures and low pressures resulting in the steady states between oxygen sorption and oxide evaporation can be described by a kinetic model based on the following assumptions (12). The chemisorbed oxygen atoms can desorb as O atoms, as O₂ molecules, and as oxide molecules (predominantly IrO₂ and IrO₃).All reactions take place in the adsorption layer and the oxygen atoms are mobile on the surface at the experimental temperatures. The total iridium loss v_Ir can be described by the relation:

Where and k_Ir are temperature dependent rate constants of the volatilisation of IrO₂, IrO₃ and Ir, respectively, and θ_O is the oxygen coverage (5, 12, 13). The oxygen is balanced by the absorption and desorption reactions, equalising their rates:

In the low-pressure range under consideration the kinetics of all processes are surface-controlled and neither true equilibrium states between gas phase and adsorption layer are established nor mass transfer in the gas phase or through agaseous boundary layer near the surface have to be regarded, as in the case of high-pressure oxidation (7, 8).

The results of the present gravimetric investigation are compared in Figure 5 with those of Wimber et al (7), obtained by photographically measuring the reduction of the wire diameter during the oxidation anneal. Both studies cover pressure range from about 0.1 pascal up to atmospheric pressure, and temperatures from 1500 to 2260°C. In the overlapping pressure and temperature range the agreement is good. Again it is evident that at low pressures the weight losses become pressure independent at high temperatures, when the weight losses are predominantly caused by the pure iridium metal evaporation, for example at 2260, 2110 and 2000°C with oxygen pressures less than 2, less then 0.2 and less than 0.01 pascals, respectively.

Further oxidation studies at very low pressures, in the range 1.3×10⁻⁵ to 1.3×10⁻³ pascals, are reported by Olivei (5), where the fraction of incident molecules pumped by an iridium surface is determined as a function of pressure and temperature. Assuming an average composition of the volatile oxides of IrO_2·5 the weight loss can be estimated from these values but the results obtained seem to be too high by about two orders of magnitude.

The theoretical treatment of the oxidation process at higher pressures is rather difficult (6 to 8) since in addition to the chemical surface reactions, eventually occurring equilibrium reactions, the mass transfer in the gas phase and natural or forced convection have to be taken into consideration. Specifically, the influence of the convection fluxes cannot be definitely determined due to geometrical parameters like shape and size of the sample and of the receiver. A theoretical treatment of iridium oxidation based on the diffusion of the evaporating species through the gaseous boundary layer is given by Wimber et al (6, 7). The strong influence of the sample dimensions is shown, for example, by Fryburg (8) studying the oxidation of platinum wires 0.29 to 2.2 mm diameter. Further difficulties may arise in an inert gas/oxygen mixture like air, where the inert gas pressure reduces the evaporation rate (14, 15).

Finally, in Figures 6 and 7 the results of oxidation studies in air are compiled. Figure 6 shows the weight loss of iridium during oxidation at reduced pressures, 10 to 6.7×10⁴ pascals, reported by Wimber et al (7). Compared with oxygen the iridium losses lie half an order to one order of magnitude lower. Figure 7 gives the iridium losses during the oxidation in the range of atmospheric pressure. A relatively fair agreement is obtained over the large temperature range (800 to 2260°C) taking into consideration the difficulties of measuring the very low weight losses at low temperatures, and the factors influencing the oxidation rate, which were mentioned above. In particular at temperatures below 2000°C the iridium losses are enlarged by several orders of magnitude when compared with pure evaporation in vacuum.

Muhamed I. Ismail has been on leave of absence from the University of Alexandria, Faculty of Engineering, Alexandria, Egypt.