Adhesive Scale Formation on Low Silicon Steel Surface; Characterization and Mechanism

Abstract

A systematic study has been conducted to characterize adhesive/non-adhesive scale on low silicon containing hot rolled steel surface and mechanism for formation of such scale. Different characterisation tools like XRD, SEM-EDS, GDOES, Scanning Kelvin probe and XPS were used to identify the chemical composition and characteristics of adhesive and non-adhesive scales. The characterization studies confirmed that the adhesive scale consists of CaFeSi₂O₆ compound along with magnetite and haematite with red in colour whereas non-adhesive scale consists of grey colour hematite-magnetite phase. Humidity in environment promotes oxidation of steel and depending on relative humidity different oxides are formed on the steel surface. Low humidity persists in India during the winter season which lead to lower oxidation potential and steel surface is oxidised accordingly. The wüstite phase is formed under lower oxygen potential at high temperature when slab is released from the mould. Wüstite is reactive to mould powder (tricalcium silicate) at high temperature and forms CaFeSi₂O₆ compound on the steel surface which is very adhesive in nature. Lots of cracks are inheritance attributes of non-adhesive scale whereas eventually no crack exists in adhesive scale. The compound present in the adhesive scale remains on the steel surface even after subsequent hot rolling and acid pickling operation as chloride ions of acid may not able to dissolve or penetrate through such stable compound. The quality of the finished product is deteriorated due to the presence of such adhesive scale as it leads to poor product performance during subsequent processing like cathodic paint deposition.

1. Introduction

The steel substrate is oxidized during hot rolling operation. Steel slab of 210 mm thickness is hot rolled at 1100–1200°C through roughing and finishing mills of a hot strip mill to a thickness of 1.6 m–6 mm. Oxidation of iron and other alloying elements in steel occurs during hot rolling operation. For low-alloyed steels or carbon steels, the scale formed during rolling is thick enough to be applicable to Wagner’s model¹⁾ and to assume parabolic oxidation kinetics. Moreover, partial interfacial control of oxidation kinetics may also be expected.²⁾ Zheludkevich et al.³⁾ have studied the influence of oxygen dissociation on the iron oxidation rate: It was found oxygen dissociation is negligible at 300–800°C. However, in ambient air, the water vapor present in the atmosphere could affect interfacial reactions.⁴⁾ The activation energy of the oxidation process in wet air (2 vol% H₂O) was found to be 151±30 kJ/mol. This value can be compared with data of Sakai et al. (149±17 kJ/mol) in air⁵⁾ and Graham et al. (134 kJ/mol) at low oxygen pressure (1.79 × 10-5 bar).⁶⁾ Many fundamental theories of iron and steel oxidation^7,8,9) reported and are well understood but very few studies^10,11,12) have been directed to systematically examine the scale structures and their formation mechanisms on hot-rolled steel strip surface. Different scientists^{13,14,15,16,17,18,19,20)} have also been reported the evidence of different types of scales developed around the original wustite layer. A. Dias et al. have reported different oxides of manganese, iron and aluminum under different oxidation conditions.²¹⁾ They could be subdivided in three classes: first, the oxides observed in temperatures up to 800°C [Mn₂O₃ and (Mn,Fe)₂O₃]; second, those found between 800 and 1000°C [γ-Fe₂O₃, γ-Mn₂O₃, Fe₂O₃, FeO and Mn₃O₄]; and finally, the aluminum spinels [MnAl₂O₄ and FeAl₂O₄] observed in temperatures above 1100°C. The Si content in the steel influences the oxide morphology. When the Si content in the steel is more than 0.35 wt%, then fayalite^22,23) phase formation dominates and influences the pickling kinetics.^23,24,25) Good quality cold-rolled steel sheet is mandatory for electro deposition painting thereby Si content in steel should be kept as low as possible to avoid any issue related to sticky scale. This scale also poses a severe surface quality issue as it is not removed easily by pickling operation. There is a natural tendency for the formation of such type of scale at the steel surface irrespective of steel chemistry (i.e. low or high Si) and occurrence/severity of such scale is not observed to be the same throughout the year. Hence, some external conditions might be responsible for formation of such type of scale. As the adhesive scale leads to poor pickling behaviour, the overall surface quality of the finished product is deteriorated and it leads to poor performance of the steel during subsequent coating like cathodic paint deposition.

2. Experimental Procedure

The adhesive scale formation is prominent during low oxidation potential conditions (i.e. mainly during the winter season with lower relative humidity ~ 40% to 50%), variable external conditions during winter may be responsible for formation of such type of adhesive scale on hot rolled (HR) steel surface. A detailed characterisation of scale formed on steel surface during the low oxidation potential condition was performed. Experimental plan consisted of:

(i) Collection of samples for scale characterization (low oxidation potential vis-à-vis comparatively higher oxidation potential conditions),

(ii) Detailed characterisation of scale

(iii) Thermodynamic calculation for stability of different oxides under plant conditions

(iv) Thermal treatment of mould powder on the surface of HR steel plate and subsequent characterisation of reaction products.

The basic composition of steel is shown in Table 1, as analysed using optical emission spectrometer according to ASTM E 415-99A standard.²⁶⁾

Table 1. Basic steel composition (wt%).

C	Si	S	Mn	P	Al	Ti	B	Cr	N
0.05	0.018	0.004	0.29	0.031	0.045	0.005	0.001	0.023	0.004

The oxide scale on the steel surface was characterized by SEM-EDS, GDOES, XRD and XPS techniques. The top surface of the scale was examined by scanning electron microscopy (SEM. JEOL JXA 6400). LECO GDS-850A glow discharge optical emission spectroscopy (GDOES) was used for determination of elemental depth profile of both types of scales (adhesive vis-à-vis non-adhesive). The structure of the scale at the adhesive portion was determined using grazing X-ray diffraction (XRD, Philips Analytical X-ray B.V. Machine). X-ray diffraction analysis of the scale was conducted using CuKα target material with step size of 0.01 degree. The X-ray photo electron spectroscopic (XPS) technique was used to evaluate the state of all the elements presence in the adhesive scale region. A laboratory scale adherence test was conducted at 1000°C with mould powder on HR steel surface. 0.5 gm mould powder was dispersed at the central region of the steel plate by spatula. Post characterization of sample was performed to understand the reactivity of mould powder with iron oxide scale. The experimentation was conducted at Jamshedpur in the eastern state Jharkhand of India.

Tests were performed to see the effect of atmospheric humidity on scale characteristics. High humidity indicates more oxygen potential as high humidity means presence of more water vapour in atmosphere and water vapour acts as an oxidising agent for steel and like oxygen. Test were performed to check the effect of different oxygen potential on scale characteristics. Electrically resistance heated tube furnace (shown in Fig. 1) was used with provision to purse CO₂ gas under variable pressure which ensures change in oxygen potential inside the tube furnace.

Fig. 1.

The experimental set up used to perform the heating of steels at elevated temperature under different oxygen potential conditions by changing pursing pressure of CO₂ and subsequent spraying of mould powder on red-hot steel surfaces. (Online version in color.)

The experimental set up used to perform the heating of steels at elevated temperature under different oxygen potential conditions by changing pursing pressure of CO₂ and subsequent spraying of mould powder on red-hot steel surfaces. The main functions of mould powder are to provide strand lubrication and to control mould heat transfer in the horizontal direction between the developing steel shell and the water-cooled copper mould. During casting, the powder melts on the steel surface, forming a layer of liquid mould slag. Crystallinity plays a critical role in steel shell formation. Glassy fluxes have high heat transfer and result in a thicker steel shell. Steels with large volumetric shrinkage on cooling must have a crystalline flux to reduce the radiative heat transfer and avoid the formation of cracks in the shell.

Oxide free low Si containing steel samples were heated in tube furnace under different oxygen potentials up to temperature of 1000°C. The heated steel sample was removed from tube furnace after heating it in a fixed oxygen potential condition and immediately mould powder (tri calcium silicate) was sprayed by spray gun on the red-hot steel surface. The different oxide scales formed on different steel surfaces under different oxygen potential conditions were characterised by SEM-EDS analysis.

3. Results and Discussion

3.1. Characterisation of Scale

Figure 2 shows the HR steel samples processed under low oxidation potential – winter season (adhesive scale) and comparatively higher oxidation potential conditions – non-winter season (non-adhesive).

Fig. 2.

Pictorial view of HR samples (a) non-adhesive and (b) adhesive scale. (Online version in color.)

It is evident from the Fig. 2 that adhesive scale formed under low oxidational potential condition consists of alternative brownish and grey layer whereas non-adhesive scale consists of alternative black and grey layers. The steel manufacturer does not face any difficulty for further processing of hot rolled coil with non-adhesive scale as it is removed easily by pickling operation. The brownish and grey alternative layers look like a tiger skin. So, sometimes this scale is also called as a ‘tiger scale’ and this scale is very adhesive in nature and difficult to remove by pickling operation. Hence further processing of this type of hot rolled steel product with adhesive scale is a real challenge for steel manufacturer.

Figure 3 shows SEM-EDS analysis of the brownish layer (in picture used as RED) and grey layer (in picture used as BLACK) of the adhesive scale.

Fig. 3.

SEM-EDS analysis of the adhesive oxide scale. (Online version in color.)

It confirmed the presence of high level of silicon along with calcium in the brownish layer of HR scale while the grey layer consists of mostly iron and oxygen with little amount of silicon which might be due to the silicon segregation from the bulk at the interface of steel – scale in the re-heating furnace. Figure 4 shows the SEM-EDS analysis in the brownish layer under higher magnification.

Fig. 4.

SEM-EDS analysis of the adhesive oxide scale under higher magnification. (Online version in color.)

It is evident that scale morphology of the brownish layer is not homogenous. It is also evident that silicon and calcium content within the brownish layer varied at different locations. The area with higher silicon and calcium content shows lower crack density in the scale compared to other areas within the brownish scale where the presence of these elements is less in concentration or completely absent. Low crack density observed in scale consisting of higher amount of silicon and calcium will be difficult to pickle since the pickling acid will not be able to penetrate through the scale and reach the underlying steel as compared to areas having higher crack density. Figure 5 shows the cross-sectional pictorial images of oxide scales formed during winter and summer seasons.

Fig. 5.

Cross sectional images of scale formed during winter and summer seasons after careful mounting and etching. (Online version in color.)

The EDS analysis indicates that winter scale consists predominantly four elements of iron, silicon, calcium and oxygen whereas summer scale consists only two elements of iron and oxygen. It is evident from Fig. 6(a) that the total thickness of scale is around 7 microns at the region of brownish layer. Presence of ~ 1.4 at% Si is detected consistently within the scale as shown in Fig. 6(b) by GDOES technique.

Fig. 6.

Depth profiling (a) Fe & O and (b) Si & Mn at brownish region of the HR scale using GDOES technique. (Online version in color.)

Thereby it can be said that adhesive scale contains a silicon bearing compound. It is evident from Fig. 7(a) that the total thickness of HR scale thickness is around 7 microns at the region of grey layer. Presence of marginal amount (0.3 at%) of silicon (see Fig. 7(b)) is detected at grey layer by GDOES technique.

Fig. 7.

Depth profiling (a) Fe & O and (b) Si & Mn at grey region of the HR scale using GDOES technique. (Online version in color.)

Silicon is known for formation of fayalite type of scale at the steel-scale interface but such low amount of silicon of 0.3 at% observed in the grey scale area is not sufficient for continuous layer formation of fayalite and such scale can be pickled easily. Figure 8 shows the XRD peaks of the adhesive scale.

Fig. 8.

Different compounds present in the scale identified by XRD technique.

It is evident that scale consists of mixture of different iron oxides and calcium silicates. Magnetite and wüstite were detected as iron oxides. Presence of hematite was not detected in XRD as its total content might be lesser than 5% in the scale. Two types of calcium silicate compounds were detected, one was plain calcium silicate and the other was calcium iron silicate. The details of phases detected in XRD are mentioned in Table 2.

Table 2. Different compounds present in the scale identified by XRD technique.

Visible	Ref. Code	Score	Compound Name	Displacement [°2Th.]	Scale Factor	Chemical Formula
*	98-001-7149	67	Magnetite low	0.054	0.980	Fe3O4
*	98-001-2493	0	Wuestite	−0.295	0.223	Fe0.932 O1
*	98-003-8290	33	Hedenbergite	−0.084	0.759	Ca1 Fe1 O6Si2
*	98-006-7062	23	Hatrurite	0.714	0.165	Ca3 O5 Si1

The work function (WF) as surface potential (mV) of this scale sample was also measured by Scanning Kelvin probe. Variation of WF values is shown in Fig. 9 for brownish and grey regions.

Fig. 9.

Work function (WF) as surface potential (mV) values of adhesive scale by Kelvin probe. (Online version in color.)

It is evident that there is a change in potential between two regions. Thermodynamic values were calculated using WF values from point 1 to 21 indicate presence of hematite whereas thermodynamic values calculated using WF values from point 21 to 45 indicate presence of iron-calcium-silicate compound. The surface potential values measured using kelvin probe equipment has a direct co-relation with Gibbs free energy change. Figure 10 shows the XPS results of the brownish layer which formed on the steel surface during winter season. Presence of hematite, silica and calcium oxides were confirmed by XPS.

Fig. 10.

XPS analysis of the brownish layer of the adhesive scale (a) Fe₂O₃, (b) SiO₂ and (c) CaO.

It was also confirmed that Fe is in Fe³⁺ state, Si is in Si⁴⁺ state and Ca is in Ca²⁺ state. Also, their orbitals were identified.

3.2. Thermodynamic Calculation for Stability of Different Oxides for HSM Conditions

Gibbs free energy was calculated with equilibrium oxygen concentration for three iron oxides in the temperature range of 0 to 1230°C using Thermocalc software. Figure 11 shows the stability diagram of different iron oxides under different equilibrium oxygen concentration and temperature.

Fig. 11.

Iron oxides stability diagram under variable equilibrium oxygen and temperature. (Online version in color.)

Free energy change is more negative for wüstite under lower oxygen potential whereas the free energy change is more negative for hematite under higher oxygen potential. It can be said that wüstite forms preferably under lower equilibrium oxygen concentration and tendency for formation of magnetite and hematite increases when equilibrium oxygen concentration increases. It is evident that wüstite phase is stable at high temperature (above 800°C) under lower oxygen potential. Hematite and magnetite are stable in-between temperature of 800 and 50°C. Hence it can be said that tendency for wüstite formation above 800°C is more whereas tendency for hematite and magnetite formation is more below 800°C.

3.3. Thermal Treatment of Mould Powder on the Surface of HR Steel Plate

The mould powder used in plant was collected and XRD analysis was done. Figure 12 shows the XRD peaks of mould powder indicating main component of mould powder is tricalcium silicate oxide.

Fig. 12.

XRD peaks of the mould powder used for thermal treatment.

Hot rolled steel plate was cut into small piece and thereafter oxide scale was removed by mechanical brushing technique from surface of the steel plate (see Figs. 13(a)–13(b)). 0.5 gm mould powder was placed near the central region of the steel plate by spatula and heated at 1000°C for 5 minutes. Figure 13(c) shows the pictorial view of steel surface after treatment with mould powder at 1000°C.

Fig. 13.

Pictorial view of steel surface (a) with oxide scale (b) after removal of oxide scale and (c) after treatment with mould powder at 1000°C. (Online version in color.)

It is evident that mould powder reacts with the steel surfaces and forms different reaction products. Figure 14 shows the SEM-EDS results of the steel surface after heat treatment with mould powder.

Fig. 14.

SEM-EDS analysis of the steel surface after thermal treatment with mould powder on the steel surface. (Online version in color.)

EDS analysis shows presence of Si, Ca, Na, K, Al along with Fe and O at the place where mould powder was placed before thermal treatment and the pattern of presence of these elements is like that observed for brownish scale layer in Figs. 4 and 6. The EDS results shows in Fig. 14 ensured absence of Si, Ca, Na, K, Al along with Fe and O in other locations where mould powder was not placed. So, it can be concluded that mould powder tends to react with steel surface and the reaction product formed on the steel surface difficult to remove by pickling operation.

3.4. Mechanism of Adhesive Scale Formation on HR Steel Surface during Winter

Characterisation findings confirmed that non-adhesive scale mainly consists of magnetite and hematite whereas adhesive scale consists of calcium-iron-silicate compound along with magnetite and traces of hematite. We propose the following mechanism of brownish adhesive scale formation. In winter time when the adhesive scale formation occurs with highest frequency, there is marked difference in the humidity level inside and outside a steel plant. In winter, due to low ambient temperature the moisture carrying tendency of air is low i.e. low relative humidity values. In this case, the mould powder will have higher chances to absorb moisture compared to non-winter conditions when the ambient temperature is at least 15–20°C higher with higher relative humidity in air. In the steel plant, there is no separate mechanism to dry damp mould powder prior to charging in the caster. We suspect that the damp mould powder being fed in the caster at a high rate might not get enough time to get completely molten and form the glassy film but instead a little amount of the mould powder remains in fused condition and gets entrapped at certain locations along the surface of solidified steel shell. In the lower part of the mould, the fused mould powder forms hedenbergite (CaFeSi₂O₆) along with wüstite, which is stable at those temperatures. Henenbergite must be forming stable and crack free film after casting and the phase grows stronger and stable during reheating furnace exposure. Since the melting point of CaFeSi₂O₆ of 1227°C which is very close to the reheating furnace temperature of 1200°C it is quite possible that this compound might become semi molten or attain mushy state and form more smooth and uniform film visible as ‘tiger’ strips. The brownish colour observed for this scale can also be attributed to the hedenbergite phase. Since the morphology of this phase, as observed in section 3.1 shows low crack density areas this phase remains as un-pickleable and adhesive scale on the final hot rolled product. The hedenbergite (CaFeSi₆O₆) is a rock forming mineral and very compact in nature due to cumulative volume of different elements is matched with the volume of the final oxide compound as a consequence no crack is generated on the compound steel interface whereas other oxides of high temperature oxide scale on steel surface mainly consist of wustite, magnetite and hematite which are defective in nature also volume of iron does not match with its oxides as a consequence stress is developed and finally decohesion happens in between steel and scale interface. It is important here to discuss one more mechanism reported by Okada et al.²⁷⁾ and Fukugawa et al.²⁸⁾ where they have also red coloured scale formation on hot rolled steel sheet, irrespective of silicon content of steel. The red scale observed by these researchers was forming on generally a thicker scale (thickness > 20 μm) and they observed higher crack density in this phase. The researchers claimed that in low silicon steel, wüstite phase partially cut into the steel gets fractured during rolling and it immediately oxidizes to higher oxygen containing iron oxides phases of magnetite and hematite due to ample availability of oxygen in atmosphere and high temperature of rolling. This scale appears red and powdery. While a little different mechanism was proposed for higher silicon steel where the fayalite formation at steel-scale interface eventually binds wüstite which later fractures during hot rolling leading to hematite formation giving red and powdery scale. Hematite being the most stable iron oxide in terms of pickling ability, it does not get pickled and remains as a reddish patch after pickling.

The brownish scale observed in our case is different since the oxide scale thickness reported is lower than 10 μm and the presence of carryover mould powder material is observed in terms of hedenbergite phase in the oxide scale. The visual appearance is little similar to the reddish iron oxide layer reported in^24,25) from macroscopic point of view however on a microscopic scale, the brownish iron oxide scale reported here is more smooth and homogeneous with very less crack density compared to the high crack density and thicker scale compared to grey scale reported in other literature.^27,28) The two similar looking scales from macroscopic point of view have a distinct root cause. Figure 15 shows the pictorial view of scales formed on steel surfaces after heat treatment in tube furnace under different oxygen potential conditions at 1000°C and subsequent spraying of mould powder on red-hot steel surfaces.

Fig. 15.

Pictorial view after heating of two steel samples under different oxygen potentials (a) normal (b) lower by pursing CO₂ at elevated temperature and subsequent spraying of mould powder on steel surface at red-hot condition. (Online version in color.)

It is evident that the scale is grey in colour when heat treatment done on steel surface under normal oxygen potential whereas the scale is red in colour when heat treatment done on steel surface under lower oxygen potential by pursing CO₂. The different oxide scales formed on different steel surfaces under different oxygen potential conditions were characterised by SEM-EDS analysis and findings have been given in Table 3 and Fig. 16.

Table 3. The different scales formed on steel surfaces after heating in tube furnace under different CO₂ pursing pressure and subsequent spraying of mould powder.

CO2 pressure	Appearance	Compactness	Scale composition
Without CO2	Greyish	Not compact	Fe and O
0.20 bar	Brownish	Less compact	Fe, Si and O
0.40 bar	Reddish	compact	Fe, Si, Ca and O
0.60 bar	Reddish	compact	Fe, Si, Ca and O
0.80 bar	Brownish	Less compact	Fe, Si and O
1 bar	Greyish	Not compact	Fe and O

Fig. 16.

SEM-EDS analysis of scales formed after heating of two steel samples under different oxygen potentials (a) normal (b) lower by pursing CO₂ at elevated temperature and subsequent spraying of mould powder on steel surface at red-hot condition. (Online version in color.)

Lot of cracks are evident on scale which obtained under normal oxygen potential whereas no such crack is formed on the scale which formed under lower oxygen potential. The scale contains mainly four elements of iron, silicon, calcium and oxygen when steel is heated to about 1000°C under lower oxygen potential and mould powder was sprayed on the red-hot steel surface whereas the scale contains only two elements of iron and oxygen when steel is heated to below 800°C under normal oxygen potential and mould powder was sprayed on the red-hot steel surface. More oxidizing conditions exists during summer season as humidity level is high which is favourable condition for formation of hematite whereas less oxidizing conditions exists during winter season which promotes for formation of wüstite. Tendency for wüstite phases formation is high during slab released from mould as wüstite is stable at high temperature (above 1000°C) at lower equilibrium oxygen concentration. Tricalcium silicate compound is used as mould powder and wüstite tend to react with silica and calcium oxide to form compound (CaFeSiO₄) at high temperature.   

$$\text{FeO}+{\text{CaO, 2SiO}}_{\text{2}}\to {\text{CaFeSi}}_{\text{2}}{\text{O}}_{\text{6}}$$

The stability and adhesive ness of both types of scales were tested in chloride medium by Tafel test which gives the dissolution of scale in chloride medium. Tafel test results (see Table 4) indicate extremely slower dissolution of scale which were obtained under lower oxygen potential condition whereas the same test results indicate faster dissolution of scale which was obtained under normal oxygen potential condition.

Table 4. The dissolution rate in chloride medium of different scales formed on steel surfaces after heating in tube furnace under different CO₂ pursing pressure and subsequent spraying of mould powder.

CO2 pressure	Dissolution rate
Without CO2	1.563 mpy
0.20 bar	1.203 mpy
0.40 bar	0.051 mpy
0.60 bar	0.034 mpy
0.80 bar	1.020 mpy
1 bar	1.962 mpy

Then, it can be concluded that the scale is formed on steel surface during winter season consists of iron-silicate compound which is adhesive/sticky in nature and may not be removed easily by pickling operation whereas the scale is formed on steel surface during summer season consists of iron oxide only which is non-adhesive/non-sticky in nature and can be removed easily by pickling operation. The layer does not get removed during de-scaling, hot rolling and pickling operations. The adhesive scale leads to poor surface quality of the finish product and it exhibits surface problems during subsequent processing of cathodic paint deposition. So, special arrangement is required during winter season.

4. Conclusions

Characterisation findings have been confirmed that the adhesive scale mainly consists of CaFeSi₂O₆ compound along with magnetite and hematite. During winter season, the dry atmosphere leads to low oxidation potential to steel surface. Wüstite phase is the most stable phase out of the three probable iron oxide phases as per thermodynamic stability is concerned. This iron oxide phase is reactive to ingredients (calcium oxide and silicon oxide) of mould powder at high temperature. Tricalcium silicate compound is used as a mould powder and wüstite tends to react with silica and calcium oxide to form the compound of CaFeSi₂O₆. This compound is red in colour and extremely stable and adhesive in nature to steel surface. Lots of cracks are formed of non-adhesive scale whereas eventually no crack exists in adhesive scale. The stable and adhesive compound remains on the steel surface even after subsequent hot rolling and acid pickling operation as chloride ions of acid may not able to dissolve or penetrate through such stable and adhesive compound. The quality of the finished product is deteriorated due to the presence of such stable and adhesive compound as it leads to poor product performances during subsequent processing like cathodic paint deposition.

Acknowledgements

The authors are grateful to Tata Steel for given approval to publish this research work in technical journal. The authors are also grateful to Ms. Nitu Rani for conducting the experimentation, Mr. V. Sharma for surface characterization and Dr Sudipto Sarkar for giving such as an interesting research subject.

References

• 1)   C.  Wagner: Corros. Sci., 9 (1969), 91.
• 2)   C.  Wagner: Corros. Sci., 10 (1970), 641.
• 3)   M. L.  Zheludkevich,  A. G.  Gusakov,  A. G.  Voropaev,  A. A.  Vecher,  K. A.  Yasakau and  M. G. S.  Ferreira: Oxid. Met., 62 (2004), 223.
• 4)   A.  Galerie,  Y.  Wouters and  M.  Caillet: Mater. Sci. Forum, 369–372 (2001), 231.
• 5)   H.  Sakai,  T.  Tsuji and  K.  Naito: J. Nucl. Sci. Technol., 22 (1985), 158.
• 6)   M. J.  Graham,  S. I.  Ali and  M.  Cohen: J. Electrochem. Soc., 117 (1970), 513.
• 7)   T.  Maruyama,  N.  Fukagai,  M.  Ueda and  K.  Kawamura: Mater. Sci. Forum, 461–464 (2004), 807.
• 8)   N.  Birks and  G. H.  Meier: Introduction to High Temperature Oxidation of Metals, Edward Arnold, London, (1983), 72.
• 9)   O.  Kubaschewski and  B. E.  Hopkins: Oxidation of Metals and Alloys, 2nd ed., Butterworths, London, (1962), 63.
• 10)   L.  Hachtel: Prakt. Metallogr., 32 (1995), 332.
• 11)   S.  Garber: Metall. Soc. Conf., 6 (1960), 41.
• 12)   S.  Garber: J. Iron Steel Inst., 192 (1959), 155.
• 13)   K.  Sachs and  G. T. F.  Jay: J. Iron Steel Inst., 195 (1960), 180.
• 14)   J. W.  Pickens: (Philadelphia), Vol. 21, Iron and Steel Society, Philadelphia, (1983), 39.
• 15)   J.  Baud,  A.  Ferrier and  J.  Manenc: Oxid. Met., 12 (1978), 331.
• 16)   R. Y.  Chen and  W. Y. D.  Yuen: Oxid. Met., 53 (2000), 539.
• 17)   R. Y.  Chen and  W. Y. D.  Yuen: Oxid. Met., 56 (2001), 89.
• 18)   P.  Promdirek,  G.  Lothongkum,  S.  Chandra-Ambhorn,  Y.  Wouters and  A.  Galerie: Oxid. Met., 81 (2014), 315.
• 19)   N. N.  Kalasin,  S.  Yenchum and  T.  Nilsonthi: Mater. Today: Proc., 5 (2018), 9359.
• 20)   H.  Han,  T.  Zhou,  W.  Wang,  Z.  Chen and  C.  Yang: Mat. Res., 22 (2019), e20190083. https://doi.org/10.1590/1980-5373-mr-2019-0083
• 21)   A.  Dias and  V.  de Freitas Cunha Lins: Corros. Sci., 40 (1998), 271.
• 22)   E. J.  Song,  D. W.  Suh and  H. K. D. H.  Bhadeshia: Ironmaking Steelmaking, 39 (2012), No. 8, 599.
• 23)   A.  Chattopadhyay and  T.  Chanda: Scr. Mater., 58 (2008), 882.
• 24)   R.  Bhattacharya,  G.  Jha,  S.  Kundu,  R.  Shankar and  N.  Gope: Surf. Coat. Technol., 201 (2006), 526.
• 25)   A.  Chattopadhyay,  P.  Kumar and  D.  Roy: Surf. Coat. Technol., 203 (2009), 2912.
• 26)  ASTM E415-99a: 2005, Standard Test Method for Optical Emission Vacuum Spectrometric Analysis of Carbon and Low-Alloy Steel.
• 27)   H.  Okada,  T.  Fukagawa,  H.  Ishihara,  A.  Okamoto,  M.  Azuma and  Y.  Matsuda: ISIJ Int., 35 (1995), No. 7, 886.
• 28)   T.  Fukagawa,  H.  Okada and  Y.  Maehara: ISIJ Int., 34 (1994), No. 11, 906.