Decarburization of 60Si2MnA in Atmospheres Containing Different Levels of Oxygen, Water Vapour and Carbon Dioxide at 700–1000 °C

Abstract

The decarburization behaviour of 60Si2MnA in atmospheres containing 0–21% O₂, < 20 ppm–17%H₂O, and with or without 8%CO₂, at 700–1000 °C, was investigated. The new findings of the current study were: (a) severe decarburization was associated with the formation of wüstite (FeO) scale on the steel surface, (b) the carbon activity at the steel–FeO interface was most likely determined by the reaction equilibrium between FeO and dissolved carbon in steel, (c) when a ferrite layer was able to form, the decarburization tendency was determined by the relative carbon permeability (defined as the product of carbon concentration difference at the two interfaces of the ferrite layer and carbon diffusivity) through the ferrite layer, and therefore, (d) the decarburization tendency at 800 °C was greater than those at 700 and 900 °C as the relative carbon permeability at 800 °C was the greatest. If FeO was absent when heating in dry O₂-containing gases, however, possibly as a result of the formation of a SiO₂ layer at the steel surface, decarburization was very much alleviated or avoided. At 1000 °C, the decarburization tendency was alleviated even when FeO was able to form because formation of a ferrite layer was not possible and carbon diffusivity in austenite was much lower than that in ferrite. A preformed oxide scale was effective in providing decarburization protection only when the steel was exposed to dry O₂-containing atmospheres.

Appendix 1: Determination of the Relative Carbon Permeability Through the Ferrite Layer [25]

In determining the equilibrium carbon composition at the FeO–steel interface, it was assumed that the dissolved carbon in steel, [C], could react with FeO via the following reactions:

$$ \left[ {\text{C}} \right] + 2{\text{FeO}} = 2{\text{Fe}} + {\text{CO}}_{2} $$

(15)

$$ \left[ {\text{C}} \right] + {\text{FeO}} = {\text{Fe}} + {\text{CO}} $$

(16)

and the overall reaction of these reactions was:

$$ {\text{CO}} + {\text{FeO}} = {\text{Fe}} + {\text{CO}}_{2} $$

(17)

Using the published data [37, 43], the standard Gibbs free energy of formation of Reaction (17) was given by

$$ \Delta G_{17}^{o} = - 22800 + 24.267T\left( {{\text{J/mole}}\;{\text{of}}\;{\text{CO}}} \right) $$

(18)

When Reaction (17) reached equilibrium,

$$ \frac{{P_{\text{CO}} }}{{P_{{{\text{CO}}_{2} }} }} = { \exp }\left( {\frac{ - 22800 + 24.267\;T }{RT}} \right) $$

(19)

where R is the gas constant (R = 8.314 J mol⁻¹ K⁻¹) and T temperature in K.

The $\frac{{P_{\text{CO}} }}{{P_{{{\text{CO}}_{2} }} }}$ thus obtained was used to determine the equilibrium carbon activity at the interface assuming $P_{\text{CO}}$ + $P_{{{\text{CO}}_{2} }}$ = 1 atm from the following reaction:

$$ \left[ {\text{C}} \right] + {\text{CO}}_{2} \left( g \right) = 2{\text{CO}}\left( g \right) $$

(20)

The standard Gibbs free energy of formation of Reaction (20) was given by [37, 43],

$$ \Delta G_{20}^{o} = - RT\ln \left[ {\frac{{P_{\text{CO}}^{2} }}{{a_{\text{c}} P_{{{\text{CO}}_{2} }} }}} \right] = 170700 - 174.5\;T\;\left( {{\text{J/mole}}\;{\text{of}}\;C} \right) $$

(21)

where $a_{c}$ was the equilibrium activity of carbon at the scale–steel interface with graphite being the standard state. From Eq. (21), we obtained,

$$ a_{\text{c}} = \frac{{P_{\text{CO}}^{2} }}{{P_{{{\text{CO}}_{2} }} }}\exp \left[ {\frac{170700 - 174.5\;T}{RT}} \right] $$

(22)

The calculated $a_{\text{c}}$ as a function of temperature is plotted in Fig. 14.

Fig. 14

Equilibrium carbon activity at FeO–steel interface as a function of temperature

After the equilibrium carbon activity at the interface was determined, the corresponding carbon concentration in the steel at the FeO–steel interface could be calculated using the known relationships between carbon activity and carbon composition. For dissolved carbon in α-Fe, the relationship given by Lobo and Gaiger [44] to express the activity coefficient of carbon in ferrite for carbon steel, $\varUpsilon_{\text{C}} \left( {\text{ferrite}} \right)$, was used,

$$ \log \varUpsilon_{C} = { \log }\left( {\frac{{a_{\text{C}} }}{{X_{\text{C}} }}} \right) = \frac{5846}{T\left( K \right)} - 2.687 $$

(23)

where $X_{\text{C}}$ was the equilibrium mole fraction of carbon in ferrite.

If the steel phase in equilibrium with the scale was γ-Fe, there were several equations available in the literature [45, 51,52,53] to relate the carbon activity to steel carbon composition. In this study, the equation given by Ellis et al. [45] was used:

$$ \log a_{\text{c}} = \log \left[ {\frac{{X_{\text{c}} }}{{1 - 5X_{\text{c}} }}} \right] + \frac{2080}{T} - 0.639 $$

(24)

Using Eqs. (22) to (24), the equilibrium carbon concentrations at the FeO–steel interface, depending on which steel phase was stable, was calculated.

When a ferrite layer formed on the steel surface, the difference in the carbon concentration between two interfaces of the ferrite layer, $\Delta C_{{C\;{\text{in}}\;\alpha }}$, provided a driving force for carbon diffusion through the ferrite layer. When the ferrite layer was thin, the carbon distribution in it could be approximated as having a linear composition gradient and the carbon diffusion flux through this layer could be described using the simplified Fick’s first law:

$$ J_{\text{C}}^{{\alpha - {\text{Fe}}}} = D_{\text{C}}^{{\alpha - {\text{Fe}}}} \cdot \frac{{C_{{C\;{\text{in}}\;\alpha }}^{\alpha /\gamma } - C_{{C\;{\text{in}}\;\alpha }}^{{\alpha /{\text{FeO}}}} }}{X} $$

(25)

where $J_{\text{C}}^{{\alpha - {\text{Fe}}}}$ = diffusion flux of carbon through the ferrite layer in mol cm⁻² s⁻¹ (moles per square centimetre per second); $D_{\text{C}}^{{\alpha - {\text{Fe}}}}$ = diffusion coefficient of carbon in ferrite, in cm² s⁻¹; $C_{{C\;{\text{in}}\;\alpha }}^{\alpha /\gamma }$ = carbon composition on the ferrite side at the BCC–FCC interface in mol cm⁻³; $C_{{C\;{\text{in}}\;\alpha }}^{{\alpha /{\text{FeO}}}}$ = carbon composition in ferrite at the BCC–FeO interface in mol cm⁻³; $X$ = thickness of the ferrite layer, in cm.

From Eq. (25), it was seen that for a certain thickness of the decarburization layer, $X$, the rate of carbon diffusion was determined by the product of carbon diffusivity $D_{\text{C}}^{{\alpha - {\text{Fe}}}}$ and the carbon composition difference between the two interfaces of the ferrite layer, $\Delta C_{{C\;{\text{in}}\;\alpha }} = C_{{C\;{\text{in}}\;\alpha }}^{\alpha /\gamma } - C_{{C\;{\text{in}}\;\alpha }}^{{\alpha /{\text{FeO}}}}$. Following the approach used by Smith [54], the following product was defined as the relative permeability ($P_{\text{C}}^{{\alpha - {\text{Fe}}}}$ in mol cm⁻¹ s⁻¹) of carbon through the ferrite layer,

$$ P_{\text{C}}^{{\alpha - {\text{Fe}}}} = D_{\text{C}}^{{\alpha - {\text{Fe}}}} \cdot (C_{{C\;{\text{in}}\;\alpha }}^{\alpha /\gamma } - C_{{C\;{\text{in}}\;\alpha }}^{\alpha /FeO} ) = D_{\text{C}}^{{\alpha - {\text{Fe}}}} \cdot \Delta C_{{C\;{\text{in}}\;\alpha }} $$

(26)

Substitution of Eq. (26) in Eq. (25) yielded

$$ J_{\text{C}}^{{\alpha - {\text{Fe}}}} = \frac{{P_{C}^{{\alpha - {\text{Fe}}}} }}{X} $$

(27)

From Eq. (27), it was seen that a greater relative permeability immediately led to a greater carbon flux for a given ferrite layer thickness.