4.1. Film growth and properties in high temperature corrosive environments
The slope of a linear curve, 0.5 in logarithmic scale means that the behaviour of film growth follows a parabolic law of oxidation [8] and that a protective film forms on the surface of mild steel. The slope of the linear curves in Fig. 3 was about 0.58 in the wet air and water steam environments. This indicated that the oxide films were weakly protective and probably weakened by multi-lamination (see Fig. 7). The fact that the slopes were approximately irrespective of the temperature is very important to obtain the activation energy of oxidation described below. The difference in weight gain could be scarcely found between the environments of wet air and water steam, as shown in Fig. 3. This is related to the lower equilibrium dissociation partial pressure of the oxide than the partial pressure of oxygen in the environments [8]. The corrosiveness of the environment seemed to increase both the slope of the linear curve and the absolute weight gain a little, as compared between Fig. 3 and Fig. 4, and then to weaken protectiveness of their films. The similar slope of weight loss, weight gain and film thickness curves on each environment as shown from Fig. 3, Fig. 4 and Fig. 5 indicates the same physical properties of the film through the testing time. The ratio of the amount of the weight loss to that of the weight gain was about 3, and implied the possibility of not iron chloride (Fe:Cl2 = 0.79:1) but iron oxide (Fe3:O4 = 2.6:1 or Fe2:O3 = 2.3:1).
The analyses of the corrosion products collected on films by XRD (Fig. 6) suggested that the films formed in any corrosive environments were only iron oxide and that iron chloride was not accumulated in the films. It is well known that volatile iron chloride generates and turns to iron oxide by the oxychlorination reaction [1]. It can be therefore deduced that the corrosiveness of the environments accelerates the oxidation of mild steel and sustains the increase in the slope in Fig. 4. The multi-layers of the films observed in Fig. 7 were considered to be randomly formed by temperature decreases in taking the specimens one by one from the furnace. The irregular growth of the multi-layers of the oxide films seemed to be one of the causes of the scattering data from Fig. 3, Fig. 4 and Fig. 5. Apparent densities of the oxide films in the corrosive environments of 773 and 873 K, which are obtained by the sum of the weight gain and loss (mg mm−2) divided by the film thickness, are listed in Table 1. The apparent density of the films in the 10 vol.% HCl environment took a lower value as compared with those in other environments. The oxide films in the 10 vol.% HCl environment look porous as shown in Fig. 7, and support the low density value.