Prediction of Oxide Phases Formed upon Internal Oxidation of Advanced High-Strength Steels

The effect of Cr on the oxidation of Fe–Mn-based steels during isothermal annealing at different dew points was investigated. The Fe–Mn–Cr–(Si) phase diagrams for oxidizing environments were computed to predict the oxide phases. Various Fe–Mn steels with different concentrations of Cr and Si were annealed at 950 °C in a gas mixture of Ar or N2 with 5 vol% H2 and dew points ranging from − 45 to 10 °C. The identified oxide species after annealing match with those predicted based on the phase diagrams. (Mn,Fe)O is the only oxide phase formed during annealing of Fe–Mn binary steel alloys. Adding Cr leads to the formation of (Mn,Cr,Fe)3O4 spinel. The dissociation oxygen partial pressure of (Mn,Cr,Fe)3O4 in the Fe–Mn–Cr steels is lower than that of (Mn,Fe)O. The Si in the steels results in the formation (Mn,Fe)2SiO4, and increasing the Si concentration suppresses the formation of (Mn,Cr,Fe)3O4 and (Mn,Fe)O during annealing.


Introduction
Advanced high-strength steels (AHSS) are widely used for automotive applications to reduce the weight of car bodies and thereby reducing fuel consumption and CO 2 emissions; see, for example, [1]. To protect AHSS against corrosion, a zinc coating is applied onto the steel surface usually by hot-dip galvanizing [2]. Before entering the zinc bath the steel strip passes through continuous annealing furnaces. The annealing atmosphere is usually a gas mixture of N 2 and H 2 with dew points ranging from -60 to 10°C, whose oxygen partial pressure level is below the dissociation oxygen partial pressure of Wüstite (FeO). However, the alloying elements (e.g. Mn, Cr, Si, Al) in AHSS have a higher affinity to oxygen than iron, leading to the formation of oxides of these elements during the annealing process prior to galvanizing. The presence of these alloying element oxides at the steel surface reduces the zinc wettability [3] and thus impairs the galvanizing quality [4]. Therefore, it is imperative to understand the oxidation behaviour of AHSS in N 2 plus H 2 gas mixtures with various dew points and to find conditions to mitigate the formation of external oxides after annealing.
The oxidation behaviour of Fe-Mn binary and Fe-Mn-Si ternary steel alloys below the dissociation oxygen partial pressure of FeO has been studied in detail (see, for example, Refs. [5][6][7][8][9]). However, only a few studies have been reported on the oxidation of Cr-alloyed Fe-Mn-based steels [10,11]. The study on the effect of Cr on the type of oxides formed in AHSS during annealing is not complete, and the type of oxides formed during annealing of Fe-Mn-Cr ternary and Fe-Mn-Cr-Si quaternary steel alloys has not been confirmed by X-ray diffraction analysis. Since Cr is often added to AHSS to enhance the hardenability of austenite [1], the primary aim of the present investigation is to understand the effect of Cr on the type of oxides formed in Fe-Mn-based steels during annealing. Also, the effect of Si on the type of oxide species formed in steels alloyed with both Mn and Cr is studied. The type and composition of the oxides formed in Fe-Mn-Cr-(Si) steel alloys as a function of annealing temperature, alloy composition and the annealing dew points are first predicted with thermodynamic computation. Then, the experimentally identified oxide phases formed after annealing different Fe-Mn-Cr-(Si) steel alloys in gas mixtures of Ar or N 2 with 5 vol% H 2 and different dew points are compared with the thermodynamic predictions. Finally, the effect of Cr and Si on the internal and external oxidations of Fe-Mn-based steels is discussed.

Experimental Procedures Samples
The chemical composition in atom per cent of the steel alloys investigated is listed in Table 1. The Fe-1.7Mn, Fe-1.8Mn-0.6Cr-0.5Si, Fe-1.8Mn-1.1Cr-0.5Si and Fe-2.8Mn-0.6Cr-0.5Si steel samples were cut with a plate cutter from a cold-rolled steel sheet. The Fe-1.8Mn-0.5Cr and Fe-1.7Mn-1.5Cr alloys were cut from ingots, and the Fe-1.9Mn-1.0Cr-0.1Si and Fe-1.9Mn-1.6Cr-0.1Si steels were cut from hot-rolled steel plates by electric discharge machining (EDM). Before annealing, the surface of specimens was ground with SiC emery paper and then polished with 1-micron diamond grains. Finally, the samples were cleaned ultrasonically in isopropanol and dried by blowing with pure nitrogen (purity better than 5 N). The samples were stored in air-tight membrane boxes (Agar Scientific G3319, the UK).

Annealing
The oxidation experiments below the dissociation oxygen partial pressure of Wüstite were carried out in a horizontal quartz tube furnace (Carbolite MTF 12/38/ 850, the UK) with an inner tube diameter of 30 mm. The temperature in the furnace tube was measured with a NiCr/NiAl (type K) thermocouple at the sample location. A gas mixture of N 2 or Ar with 5 vol% H 2 was passed through the furnace tube at atmospheric pressure. De-aerated and deionized water (18.2 MX cm at 25°C) was evaporated with a controlled evaporator mixer (CEM, Bronkhorst, the Netherlands) to create specific dew points of -45, -37, -30, -10 and 10°C (corresponding oxygen partial pressures of 8.1 9 10 -22 , 5.0 9 10 -21 , 2.2 9 10 -20 , 1.0 9 10 -18 and 2.3 9 10 -17 atm. at 950°C; see ''Appendix'') in the furnace. The pure water was de-aerated with nitrogen gas in a closed pressurized stainless steel vessel, and the dissolved oxygen gas in the pure water was below 100 ppb, as measured with an O 2 sensor (InPro 6850i, Mettler Toledo, the USA). The dew points of -10 and 10°C were monitored with a cooled mirror analyser (Optidew, Michell Instruments, the UK). The dew points of -45, -37 and -30°C were monitored with another cooled mirror analyser (S4000 TRS, Michell Instruments, the UK). The fluctuation of dew points during annealing was within ± 2°C. The partial pressure of evaporated H 2 O in the gas mixture is related to dew point according to the formula given in ''Appendix''. The gas mixture with dew points of -10 and 10°C consists of N 2 ? 5 vol% H 2 , while the gas mixture with dew points of -45, -37 and -30°C was created with Ar ? 5 vol% H 2 . However, the oxidation behaviour of the steel alloys used in this study is not affected by the type of carrier gas (i.e. Ar or N 2 ), because formation of nitride is not thermodynamically favourable for our samples and our annealing conditions. The flow rate of gas mixture with dew points of -30 to 10°C was 1500 sccm, while the flow rate of gas mixture with dew points of -45 and -37°C was 3000 sccm. Before annealing, the sample was placed onto a quartz boat located at the cold zone of the furnace tube. Then, the furnace was closed and flushed with the reaction gas for more than 30 min. Next, the sample was moved to the hot zone of the furnace with a quartz rod to start an annealing cycle. At the end of the annealing cycle, the sample was moved in the reverse direction, i.e. to the The oxidation experiments above the dissociation oxygen partial pressure of Wüstite were carried out with the Fe-1.9Mn-1.6Cr-0.1Si steel alloy in the same horizontal quartz tube, but with an infrared furnace (Model 4069-12R-05L, Research Inc.) at 950°C for 8 h in a gas mixture of Ar ? 20 vol% CO 2 ? 20 vol% CO (oxygen partial pressure of 8.8 9 10 -16 atm.; see ''Appendix'') at atmospheric pressure with a total gas flow of 500 sccm. The temperature of the sample during annealing was determined by measuring the temperature of a dummy sample which was spot-welded to a thermocouple. The dummy sample has the same dimensions and chemical composition as the target sample.
Prior to admitting the gas mixtures to any of the furnaces, each gas, i.e. Ar, N 2 , H 2 , CO 2 and CO (all with a purity better than 5 N vol%), was filtered to remove any residual hydrocarbons, moisture and oxygen, with Accosorb (\ 10 ppb hydrocarbons), Hydrosorb (\ 10 ppb H 2 O) and Oxysorb (\ 5 ppb O 2 ) filters (Messer Griesheim, Germany), respectively. The flow of each gas was regulated and monitored using mass flow controllers (Bronkhorst, the Netherlands).

Characterization
Grazing angle X-ray diffractometry (GA-XRD) was used to identify the oxide phases present at the surface of the oxidized samples. The XRD patterns were recorded, with a Bruker D8 Discover diffractometer in the grazing incidence geometry using Co Ka radiation, in the 2h region between 20°and 60°with a step size of 0.03°2h and a dwell time of 10 s. The incidence angle of the X-ray beam was fixed at 3°with respect to the sample surface. In this grazing incidence geometry, the depth below the surface corresponding to 70% of the diffracted intensity of pure iron is 1.27-1.34 lm for 2h ranging from 30°to 60° [12]. When the incidence angle of the X-ray beam was at 2°with respect to the sample surface, the depth below the surface corresponding to 70% of the diffracted intensity of FeO is 1.4-1.5 lm for 2h ranging from 30°to 60°. Also, XRD measurements in the Bragg-Brentano geometry were taken with the same diffractometer in the 2h region between 10°and 110°with a step size of 0.03°2h and a dwell time of 2 s. Then, the depth below the surface corresponding to 70% of the diffracted intensity of FeO for Bragg-Brentano geometry is about 5.7-10.9 lm for 2h from 30°to 60° [12].
The surface and cross section of the annealed samples were observed with scanning electron microscopy (SEM) using a JSM-6500F (JEOL, Japan) operated with an accelerating voltage of 5 or 15 kV. X-ray microanalysis (XMA) using energy-dispersive spectroscopy (EDS) was performed with the same SEM instrument, but equipped with an UltraDry 30 mm 2 detector (Thermo Fisher Scientific, the USA) to determine the chemical composition of oxides qualitatively.
X-ray photoelectron spectroscopy (XPS) was used to analyse the chemical composition of the oxides at the steel surfaces. The photoelectron spectra were recorded with a PHI 5400 ESCA equipped with an X-ray source operated at 200 W and 13 kV using an Al anode. The pass energy of the spherical capacitor analyser was set at 35.75 eV. Prior to the recording of the spectra, the steel sample surface was sputtered mildly with a 3 keV Ar ion beam with ion flux of 2 9 10 -13 mol/ mm 2 s, rastering over an area of 5 9 5 mm for 10 min.

Results and Discussion
Phase Diagrams and Oxide Phases The phase diagrams of Fe-Mn, Fe-Mn-Cr and Fe-Mn-Cr-Si alloys in an oxidizing environment at 950°C are shown in Figs. 1, 2, 3, 4, 5 and 6. These phase diagrams were constructed with FactSage [13]. The thermodynamic data of the stoichiometric compounds Cr 2 O 3 and SiO 2 as well as the solid solution oxides, namely (Mn,Cr,Fe) 3 O 4 spinel (Mn,Fe)O, (Mn,Fe) 2 SiO 4 and (Mn,Fe)SiO 3 , were taken from the FToxid database [14]. A solid solution of Fe-Mn binary, Fe-Mn-Cr ternary and Fe-Mn-Cr-Si quaternary alloy with fcc or bcc crystal lattice was created with the thermodynamic data in the FSstel database [14]. A gas mixture of Ar and O 2 with increasing oxygen partial pressure (in atm.) was created using the thermodynamic data in the FactPS database [15] to be in equilibrium with the alloy phase and the oxides.
First the oxide phase that can be formed in a Fe-Mn binary alloy is considered; see Fig. 1 and Table 2. The dissociation oxygen partial pressure of MnO is lower than that of FeO. However, since both FeO and MnO have the same rock-salt crystal structure, FeO and MnO can form a continuous solid solution [16] denoted as (Mn,Fe)O. The Fe concentration in the (Mn,Fe)O increases with oxygen partial pressure, which agrees with the results reported in Refs. [17,18]. Above an oxygen partial pressure of about 1.6 9 10 -16 atm. at 950°C, all the Fe and Mn in the alloy are oxidized to (Mn,Fe)O. The prediction with the computed phase diagram is in agreement with our experimental results (see Table 2) and the results reported in Ref. [5], and (Mn,Fe)O is the only type of oxide phase that can be formed in a Fe-Mn binary alloy below an oxygen partial pressure of 1.6 9 10 -16 atm. However, when annealing the Fe-1.7Mn steel at the dew point of -45°C (oxygen partial    Table 2. Considering the Mn concentration in the alloy fixed, the dissociation oxygen partial pressure of (Mn,Cr,Fe) 3 O 4 spinel decreases, while the dissociation oxygen partial pressure of (Mn,Fe)O slightly increases with the Cr concentration in the alloy; see Fig. 2. The dissociation partial pressure of (Mn,Cr,Fe) 3 O 4 spinel is (It is noted that when the alloy matrix is in austenite and the pO 2 is below 3.2 9 10 -26 atm., the formation of (Mn,Fe)SiO 3 is also predicted, but is not shown here.) lower than the dissociation oxygen partial pressure of (Mn,Fe)O, when the Cr concentration in the bulk alloy is higher than about 0.2 at.%. This is confirmed by the observation that (Mn,Cr,Fe) 3 O 4 spinel is formed during annealing at low dew points, while (Mn,Fe)O appears at high dew points in the Fe-1.7Mn-1.5Cr, Fe- Table 2 Identified oxide species from XRD measurements (all XRD results obtained from measurements using Co Ka radiation with grazing incidence geometry) after annealing the Fe-Mn and Fe-Mn-Cr-(Si) steel alloys at 950°C for 1 h in a gas mixture of Ar or N 2 with 5 vol% H 2 and dew points of -45, -37, -30, -10 and 10°C (corresponding to oxygen partial pressures (pO 2 ) of 8.1 9 10 -22 , 5.0 9 10 -21 , 2.2 9 10 -20 , 1.0 9 10 -18 and 2.3 9 10 -17 atm., respectively)  Predicted oxide phases are denoted in bold and indicated with an asterisk a Data not available because the amount of oxides formed is below the detection limit of the XRD measurements * Predicted oxide phases Oxid Met 1.9Mn-1.0Cr-0.1Si and Fe-1.9Mn-1.6Cr-0.1Si steel alloys; see Table 2. The increase in the dissociation oxygen partial pressure of (Mn,Fe)O with the Cr concentration is evidenced by the identification of (Mn,Fe)O in the Fe-1.8Mn-0.5Cr, but is not detected in the Fe-1.7Mn-1.5Cr alloy annealed at the dew point of -45°C; see Table 2. This is due to the fact that the dissociation oxygen partial pressure of (Mn,Fe)O decreases with the concentration of Mn dissolved in the alloy. At certain oxygen partial pressure and concentration of Mn in the alloy, the amount of Mn that reacts with Cr to form (Mn,Cr,Fe) 3 O 4 spinel increases and, hence, the concentration of Mn that remains in the alloy decreases with Cr concentration. The addition of Si to the Fe-Mn-Cr alloys leads to the formation of (Mn,Fe) 2 SiO 4 . The dissociation oxygen partial pressure of (Mn,Fe) 2 SiO 4 is lower than that of (Mn,Cr,Fe) 3 O 4 for the alloy compositions considered here; see Figs. 3 and 4. This is consistent with the oxide phase identified in the annealed Fe-1.8Mn-0.6Cr-0.5Si and Fe-1.8Mn-1.1Cr-0.5Si steels, i.e. a single (Mn,Fe) 2 SiO 4 oxide phase is formed at the dew point of -45°C; see Table 2. However, according to the phase diagram (Fig. 4), the formation of (Mn,Cr,Fe) 3 O 4 spinel is also predicted for the Fe-1.8Mn-0.6Cr-0.5Si and Fe-1.8Mn-1.1Cr-0.5Si alloys after annealing at the dew point of -45°C. This shows that adding Si to the Fe-Mn-Cr steel alloys suppresses the formation of (Mn,Cr,Fe) 3 O 4 spinel. Apparently, Si in the steel lowers the oxygen partial pressure at steel surface. But, after annealing of the Fe-1.8Mn-0.6Cr-0.5Si and Fe-1.8Mn-1.1Cr-0.5Si alloys at higher dew points than -45°C up to 10°C, also (Mn,Cr,Fe) 3 O 4 spinel is observed; see Table 2. This is in agreement with the phase diagram; see  Table 2. The effect of Si on the dissociation oxygen partial pressure of (Mn,Fe)O can be explained as follows. Since (Mn,Fe) 2 SiO 4 is much more stable than both (Mn,Cr,Fe) 3  Between the dissociation oxygen partial pressure of (Mn,Fe)O and 2.3 9 10 -17 atm. (dew point of 10°C) at 950°C, the effect of oxygen partial pressure has no effect on the type of oxides formed during annealing of Fe-Mn-Cr-(Si) steel alloys. For example, both (Mn,Fe)O and (Mn,Cr,Fe) 3 O 4 spinel were identified on Fe-1.8Mn-0.6Cr-0.5Si, Fe-1.8Mn-1.1Cr-0.5Si and Fe-2.8Mn-0.6Cr-0.5Si steels after annealing at -10 and 10°C; see Table 2. It is expected that annealing the Fe-1.8Mn-0.5Cr, Fe-1.7Mn-1.5Cr, Fe-1.9Mn-1.0Cr-0.1Si, Fe-1.9Mn-1.6Cr-0.1Si and Fe-2.8Mn-0.6Cr-0.5Si steel alloys above the dissociation oxygen partial pressure of (Mn The effect of the crystal lattice of the steel matrix on the equilibrium oxide phases formed in advanced high-strength steels is small. Figure 5 shows the computed phase diagrams of the Fe-Mn-Cr-Si quaternary alloys in an oxidizing gas atmosphere at 850°C with the concentration of Mn and Si fixed at 1.8 and 0.5 at.%, respectively. The constitution of the steel alloy matrix was fixed in either bcc or fcc phase in the computation, and thus the effect of alloy composition on the austenite-ferrite phase transformation was not considered. The dissociation oxygen partial pressure of (Mn,Cr,Fe) 3 O 4 spinel and (Mn,Fe)O at 850°C in austenite is only slightly higher than in ferrite. This is due to the fact that the chemical potential of the alloying elements Mn and Cr in ferrite is higher than that in austenite.

Composition of Oxide Phases
The Fe concentration in the (Mn,Cr,Fe) 3 O 4 spinel formed during the oxidation of Fe-Mn-Cr steel alloys increases with ambient oxygen partial pressure. A spinel oxide can be written in the general form of AB 2 O 4 [19]. For MnCr 2 O 4 spinel, Fe cations can substitute both Mn cations at A-site and Cr cations at B-site [20]. The lattice constant of (Mn 1-x Fe x )(Cr 2-y Fe y )O 4 spinel increases with the value of y, but decreases with the value of x [20]. The amount of Fe dissolved in the (Mn,Cr,Fe) 3 O 4 spinel (i.e. the value of (x ? y)/3 in (Mn 1-x Fe x )(Cr 2-y Fe y )O 4 ) can be predicted [13,15]. For example, the amount of Fe in (Mn,Cr,Fe) 3 O 4 increases from 0 to 0.12 when annealing a Fe-1.8 at.% Mn-1.1 at.% Cr alloy at 950°C in Ar ? 5 vol% H 2 gas mixture while increasing the dew point from -45 to 10°C; see Fig. 7. Moreover, the calculations show that non-stoichiometry of the spinel is negligible and that the Fe dissolved in the (Mn 1-x Fe x )(Cr 2-y Fe y )O 4 spinel mainly resides at the A-site after annealing at 950°C with an oxygen partial pressure of 2.3 9 10 -17 atm. (i.e. corresponding to a dew point of 10°C in Ar ? 5 vol% H 2 gas mixture). The measured stress-free lattice constant of (Mn,Cr,Fe) 3  An oxide scale is formed at the surface of the Fe-1.9Mn-1.6Cr-0.1Si steel after annealing at 950°C for 8 h in the Ar ? 20 vol% CO 2 ? 20 vol% CO gas mixture, i.e. at an oxygen partial pressure of 8.8 9 10 -16 atm. This oxide scale fully covers the steel surface, but the scale thickness is not uniform; see Fig. 9. An internal oxidation zone is formed underneath the oxide scale. (Mn,Cr,Fe) 3 O 4 spinel, MnO and Wüstite were identified from the diffraction pattern obtained by XRD with Bragg-Brentano geometry. However, in the diffraction pattern recorded by XRD using the grazing incidence geometry (having a smaller analysis depth), (Mn,Cr,Fe) 3 O 4 spinel and MnO can hardly be observed and the relative intensity of the diffraction peak of iron is significantly lower; see Fig. 10. This shows that the diffraction pattern recorded by XRD using the grazing incidence geometry mainly contains the information of the oxide scale at steel surface. These observations suggest that the external oxide scale is composed of Wüstite only, while the (Mn,Cr,Fe) 3 O 4 spinel and (Mn,Fe)O are formed as internal oxide precipitates. The establishment of local thermodynamic equilibrium between oxide precipitates and dissolved oxygen in alloy matrix upon internal oxidation of Fe-Mn binary steel alloys has been reported [18]. For Fe-Mn-Cr-(Si) steel alloys, the agreement between the computed phase diagrams and the experimentally identified oxide species formed during annealing at different temperatures and oxygen partial pressure indicates that (local) thermodynamic equilibrium between the gas ambient and the steel surface was established. Finally, the thermodynamic data used allowed the prediction of the oxide formed in advanced high-strength steels having a complex composition and microstructure.

Internal and External Oxides
The oxidation mode of Fe-Mn-Cr steel alloys annealed at 950°C in a gas mixture of Ar or N 2 ? 5 vol% H 2 changes from external to internal oxidation with increasing dew point from -45 to 10°C, which is similar as for Fe-Mn steels [5]. For example, when annealing the Fe-1.9Mn-1.6Cr-0.1Si steel in an Ar ? 5 vol% H 2 gas mixture with the dew point of -45°C the oxides are formed mainly at the sample surface; see Fig. 11. Increasing the dew point of the annealing gas mixture to 10°C, an internal oxidation zone is observed below the Fe-1.9Mn-1.6Cr-0.1Si steel surface. According to XMA and thermodynamic computations, the internal oxide precipitates should be composed of (Mn,Fe)O and (Mn,Cr,Fe) 3 O 4 spinel. Adding Si to the Fe-Mn-Cr steel promotes the formation of (Mn,Fe) 2 SiO 4 along original austenite grain boundaries; see Fig. 12. However, (Mn,Fe) 2 SiO 4 also forms inside grains and an individual precipitate can be composed of more than one type of oxide species; see Fig. 13.
Adding Cr to a Fe-Mn binary steel alloy increases the amount of oxides formed at the steel surface; see, for example, Fig. 14. Oxides at the surface of the Fe-