Longer Term Oxidation Kinetics of Low Carbon, Low Silicon Steel in 17H₂O–N₂ at 900°C

Abstract

The longer term oxidation behaviour of a low carbon, low silicon steel in 17H₂O–N₂ at 900°C was examined. The oxidation kinetics exhibited a multi-stage pattern with the first stage being initially linear and then transitioned to parabolic with a rate constant two orders of magnitude smaller than prediction. This was followed by an accelerated stage, and then a second parabolic stage with a rate constant much closer to that of iron and steel oxidation in air or oxygen. A single-phase wüstite layer was observed throughout the oxidation period. Using Wagner’s scaling equation, the effective oxygen potential (oxygen potential achieved at the outer surface of the wüstite layer) was calculated. During the first parabolic stage, this potential was extremely low, leading to a very low oxidation rate. It was then increased continuously through the accelerated oxidation stage to a much higher level at the second parabolic stage. The possible mechanism responsible for the observed trend of effective oxygen potential change and the corresponding trend of oxidation kinetics was discussed.

Appendices

Appendix I: Diffusion Coefficient of Radioactive Iron in Wüstite Fe_1−δ O

From the data of Chen and Peterson [38], the following relationships are obtained,

$$D_{{{\text{Fe in Fe}}_{ 0. 9 4} {\text{O}}}}^{\text{s}} = 8.67 \times 10^{ - 3} { \exp }\left( {\frac{ - 122900\left( J \right)}{RT}} \right)$$

(7)

$$D_{{{\text{Fe in Fe}}_{ 0. 9 2} {\text{O}}}}^{\text{s}} = 1.64 \times 10^{ - 2} { \exp }\left( {\frac{ - 130070\left( J \right)}{RT}} \right)$$

(8)

$$D_{{{\text{Fe in Fe}}_{ 0. 9 0} {\text{O}}}}^{\text{s}} = 4.68 \times 10^{ - 2} { \exp }\left( {\frac{ - 141940\left( J \right)}{RT}} \right)$$

(9)

$$D_{{{\text{Fe in Fe}}_{ 0. 8 8} {\text{O}}}}^{\text{s}} = 5.24 \times 10^{ - 2} { \exp }\left( {\frac{ - 143300\left( J \right)}{RT}} \right)$$

(10)

Regression of the data of Himmel et al. [15] for iron diffusion in Fe_0.907O yields,

$$D_{{{\text{Fe in Fe}}_{ 0. 9 0 7} {\text{O}}}}^{\text{s}} = 1.18 \times 10^{ - 2} { \exp }\left( {\frac{ - 124350\left( J \right)}{RT}} \right)$$

(11)

Note that the original equation given by Himmel et al. could be incorrect, which is, as pointed out previously [1]

$$D_{{{\text{Fe in Fe}}_{ 0. 9 0 7} {\text{O}}}}^{\text{s}} = 0.118\;{ \exp }\left( {\frac{ - 124350\left( J \right)}{RT}} \right)$$

(12)

Appendix II: Determination of Parameters for Eq. (5)

Extensive studies had been carried out to determine the degree of deviation from stoichiometry for wusttie ${\text{Fe}}_{1 - \delta } {\text{O}}$ [28–35]. At 900 °C, the range of deviation is 1 − δ = 0.89–0.94 [Ref. 36, p. 222].

Himmel et al. [15] conducted experimental studies to determine the relationship between the radioactive trace diffusion coefficient of iron in wüstite and the composition of wüstite at 700 to 983 °C and found that at 900 °C, the value appeared to increase with temperature, although the actual data were rather scattered. Within the range of deviation from stoichiometry (1−δ = 0.89–0.94), the values obtained by Himmel et al. were within the range of 2.35 × 10⁻⁸ to 4.45 × 10⁻⁸ cm²/s at 900 °C, as shown in Appendix Fig. 8.

Fig. 8

Relationship between radioactive trace diffusion coefficient of iron in wustite, Fe_1−δ O, and deviation from stoichiometry δ at 900°C [15, 38]. Data of Chen and Peterson [38] were extrapolated from experimental data obtained at other temperatures (Appendix I)

Chen and Peterson [38] measured the diffusion coefficients of radioactive iron in wüstite at 800, 1000 and 1200 °C and found that the diffusivity of iron in wüstite increased with δ at 1200 °C, decreased with δ at 800 °C, and insensitive to the variation of δ at 1000 °C. Extrapolating the experimental data to 900 °C using the equations in Appendix I, it was found that the iron diffusivity increased slightly with increased δ, as seen in Appendix Fig. 8, a trend opposite to that claimed by Himmel et al. [15].

Considering the data obtained by Chen and Peterson [38] and those by Himmel et al. [15], it is reasonable to conclude that the diffusivity of iron in wüstite is insensitive to the variation of δ at 900 °C, having a value within the range of 2.5 × 10⁻⁸ to 3.0 × 10⁻⁸ cm²/s. This conclusion is further supported by the study of Engell [37] who measured the concentration of iron-ion vacancy δ across the wüstite layer electrochemically in the scales formed in oxygen between 750°C and 1000°C and found that the concentration gradient of iron-ion vacancy was constant over the entire wüstite layer. A condition for this to occur, as pointed out by Engell [37], was that the diffusion coefficient of the defects was independent of the oxide concentration. A value of $D_{\text{Fe}}^{ *} = 2.9 \times 10^{ - 8}$ cm²/s will be used in this study.

There has been no consensus in the determination of the correlation factor for ion diffusion in wüstite [38, 39], although it was proposed [31] that its value was most likely to be within the range of 0.55–0.82, depending on the actual diffusion mechanism (via iron vacancies or interstitial iron ions or both [35, 40]). In conducting theoretical calculation of oxide growth on iron at 1100 °C, Garnaud and Rapp [39] used a correlation factor of 0.5 and obtained good agreement between theoretical and experimental results, whereas in some other studies, $D_{\text{Fe}}^{ *}$ was treated as the same as $D_{\text{Fe}}^{\text{s}}$ [1, 3–5, 15]. A value of 0.7 will be used in the current study.

The absence of magnetite precipitates in all the oxidized samples suggests that the oxygen partial pressure was significantly lower than the wüstite/magnetite equilibrium value of 1.83 × 10⁻¹⁵ atm throughout the oxidation period, and therefore the corresponding δ value at the outer surface of the scale would be much smaller than 0.11. Consequently, in the wüstite scales obtained in this study, the value of $\frac{1}{1 - \delta }$ in Eq. (5) should satisfy the following relationship,

$$\frac{1}{0.95} \cong 1.05 < \frac{1}{1 - \delta } \ll 1.12 = \frac{1}{0.89}$$

(13)

With this small variation range, $\frac{1}{1 - \delta }$ in Eq. (5) will be approximated as a constant at $\frac{1}{1 - \delta } \approx 1.07$ or $1 - \delta \approx 0.935$.