Corrosion Science 53 (2011) 36033611

Contents lists available at ScienceDirect

Corrosion Science
journal homepage: www.elsevier.com/locate/corsci

Three dimensional microstructure study of oxide scale formed on a high-speed steel by means of SEM, FIB and TEM
Q. Zhu a, H.T. Zhu a,, A.K. Tieu a, C. Kong b
a b

Faculty of Engineering, University of Wollongong, Northelds Avenue, Wollongong, NSW 2522, Australia Electron Microscope Unit, The University of New South Wales, Sydney, NSW 2052, Australia

a r t i c l e

i n f o

a b s t r a c t
Characteristics of the oxide scale formed on a high-speed steel (HSS) material in the temperature range 550650 C were examined. The surface morphologies of oxidised sample indicate that the temperature has a signicant inuence on the oxidation behaviour of the HSS samples. Differential oxidation is expected to occur due to high-alloyed components in the material. The carbide-free matrix has a good oxidation resistance due to the dissolved chromium; while vanadium rich carbide (MC) regions were oxidised heavily because of high free energy at the carbide/matrix interface and low thermal stability of the MC carbides. FIB/TEM cross-section observation shows that the oxide scale formed on the surface has a duplex-layer structure with vanadium oxides covering the top of the scale. 2011 Elsevier Ltd. All rights reserved.

Article history: Received 11 March 2011 Accepted 1 July 2011 Available online 8 July 2011 Keywords: A. Steel B. SEM B. TEM C. Oxidation

1. Introduction High-speed steels (HSS) has been successfully used to manufacture hot work rolls because it retains a higher hardness at high temperature compared to conventional high-chromium (Hi-Cr) and indenite chill iron (IC) rolls. However, oxidation is an ever present problem during hot rolling processing [1,2]. The working temperature of hot work rolls can exceeds 600 C although the ash temperature could in fact rise to 800 C due to the frictiongenerated heat [3,4]. The oxidation of the roll surface is one of four wear mechanisms (i.e. abrasion, adhesion, thermal fatigue and oxidation) of the work rolls which directly affects the roll performance and the surface quality of rolled material [3,57]. A number of studies have been carried out to investigate the oxidation behaviour of high-speed steels [3,812]. The formation of a thin oxide scale on the roll surface plays an important role during hot rolling. Vergne et al. [2] suggested that a compact thin oxide layer on the roll surface can prevent the sticking problem by establishing an oxideoxide contact instead of a metal-oxide contact with the hot rolled material. However, the tribological behaviour of the oxides is still not clear. Pellizzari et al. [13] found that the hard oxide layer on the roll surface can have a positive effect on the wear rate, while other researchers argued that the oxide layer has a protective effect as a solid lubricant against the roll wear [14]. Vergne et al. [2,15] pointed out that the tribological behaviour of the

oxides is very complex because it is inuenced by the nature and physical properties of the oxide scales. Detailed characterisation of the microstructure of the oxide scale is essential for a deeper understanding of the oxidation behaviour in the contact between the roll and strip. However, only a limited number of papers have been published on this topic. The conventional way to study oxide scale by metallographic crosssection preparation methods often causes damage to the brittle oxide scale at room temperature. In addition, the oxide layer formed on HSS is expected to be very thin due to its good oxidation resistance [9,10,12]. Thus, it is difcult to study microstructure of such thin oxide layers in details by traditional methods. Recently focused ion beam (FIB) microscope has been increasingly used to study the oxidation of metals because it can easily produce without damage a site-specic oxide cross-section and thin foil for transmission electron microscope (TEM) observation [1619]. In the present work, a roll grade HSS material was isothermally oxidised in dry air in the temperature range 550650 C which is generally considered as the working temperature of hot work roll surfaces [14,14,20]. A three dimensional study of the oxide scales from surface topography to site-specic cross-sections was investigated in details by means of SEM, FIB and TEM. 2. Experimental 2.1. Material and oxidation procedure

Corresponding author. Tel.: +61 2 4221 4549; fax: +61 2 4221 3101.
E-mail address: hongtao@uow.edu.au (H.T. Zhu). 0010-938X/$ - see front matter 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.corsci.2011.07.004

The chemical composition of the studied high-speed steel is listed in Table 1. Oxidation tests were carried out in a high

3604

Q. Zhu et al. / Corrosion Science 53 (2011) 36033611

Table 1 Chemical composition of the studied high-speed steel (in weight percent, wt.%) analysed by X-ray uorescence spectroscopy. Fe Balance C 1.96 Cr 4.85 V 4.00 Mo 4.47 W 3.40 Mn 1.26 Si 0.99 Ni 0.78

(XRD) using a GBC MMA diffractometer with monochromated Cu-Ka radiation. 3. Results 3.1. SEM study for surface morphologies before and after oxidation

temperature microscope (HTM) with dry air at temperatures 550 650 C with an increment of 50 C for 30 min up to 120 min. Fig. 1 shows a schematic illustration of the transverse cross-section of the HTM chamber and the position of the specimens. The HSS samples were cut into cuboid specimens with dimensions 2.7 2.7 2 mm3. One broad face on the samples was polished and nished with 1 lm diamond paste. The sample surfaces were cleaned by means of ultrasonic agitation in ethanol and preserved in a desiccator before the oxidation test. The specimens were heated up and cooled down in a high purity argon atmosphere (Ar, 99.999%) before and after oxidation. The detailed experimental procedure can be found in [8]. 2.2. Analysis methodology After oxidation, a FEI XT Nova Nanolab 200 workstation which combines a dual beam of focused ion beam and a eld emission scanning electron microscope (FIB/SEM) was used to characterise the surface morphology of specimen and prepare thin-foil specimens for further TEM observation. The system is equipped with a platinum deposition system. The FIB column uses a ne, energetic beam of gallium ions that scan over the surface of a specimen for imaging. At high beam currents the ion beam rapidly sputters away the specimen surface which exposes the subsurface crosssection. Two steps were taken to prepare the TEM samples used in this work. Firstly, the electron beam was used to choose a typical feature, and a protective Pt lm was deposited on the selected surface to protect the oxide scale prior to the cutting; secondly, the ion beam was used to cut off the cross-sections with the dimension 15 lm 5 lm 0.50.6 lm, and mill these thin-foils down to less than 100 nm in thickness for a TEM examination. During the FIB cross-sectioning, an accelerating voltage of 30 kV and high ion currents were used and, for the milling lower currents of 10.1 nA were used. A JEOL JSM 6490 scanning electron microscope (SEM) equipped with energy dispersive X-ray analysis (EDAX) was also employed to characterise the surface morphology of the specimen before and after oxidation. A Philips CM200 eld emission gun transmission electron microscope (FEGTEM) equipped with a Brucker energy dispersive X-ray spectroscopy (EDAX) system was used to investigate the detailed microstructure of the FIB cross-sections of the oxide scales. It was operated with an accelerating voltage of 200 kV. Selected oxidised samples were examined by X-ray diffraction

Fig. 2a shows the backscattered electron (BSE) image of the polished high-speed steel sample prior to the oxidation tests. The material consists of three main carbides: the slender petal-like V-rich MC carbides; brous-like (Mo, W)-rich M2C carbides and network-like Cr-rich M7C3 carbides (M2C and M7C3 carbides are found precipitate together). High magnication observation in TEM indicates that there are many ne Mo, W-rich spherical carbides (M2C) with sizes $200 nm precipitate in the HSS matrix (Fig. 2b). The EDX analysis indicates that although more chromium precipitates with carbides, chromium is evenly distributed in the matrix. Fig. 3a shows the secondary electron (SE) image of the HSS sample surface after oxidation at 550 C for 30 min in dry air. It can be seen that very ne iron oxides cover the sample surface with some large rod-like crystals after oxidation. A higher magnication image indicates that the oxide scale does not cover the substrate evenly, several cavities are clearly visible and some locations are protruded by rod-like crystals (Fig. 3b). SEM/X-ray mapping analysis of oxidised surface shown in Fig. 4a conrms that the cavities correspond to the chromium and molybdenum rich carbides area, and rod-like crystals are rich of vanadium indicating that V-rich MC carbides have been oxidised after the test (detailed feature of vanadium oxide as seen in Fig. 4b). It is clearly seen that ne iron

Fig. 1. Schematic illustration for the transverse cross-section of the high temperature microscope chamber and the position of specimen placed in the chamber.

Fig. 2. (a) Backscattered electron (BSE) microstructure image of the studied highspeed steel; (b) TEM bright eld microstructure image of the studied high-speed steel.

Q. Zhu et al. / Corrosion Science 53 (2011) 36033611

3605

Fig. 3. Secondary electron (SE) images of surface morphology of specimens after 30 min oxidation at 550 C (a and b), 600 C (c and d) and 650 C (e and f) in dry air. Higher magnication SE images are shown on the right hand side. Samples are tilted 52 in (b) and (f).

oxide whiskers and blades growing on the oxide scale in Fig. 4b and c. However, the iron oxides do not grow directly from the metal substrate. There is a thin, compact chromium rich oxide layer formed and covering the substrate. The ne chromium oxides are more visible at the Cr-rich M7C3 carbides/matrix interface due to higher chromium contents precipitating in these areas than in the matrix as seen in Fig. 4c. Fig. 3c shows the surface morphology of the HSS sample oxidised at 600 C in dry air for 30 min. After oxidation, a compact oxide layer covers the HSS matrix and the areas of V-rich MC carbides protrude above the surface due to the oxidation of carbides. At a higher magnication it can be seen that the vanadium oxides are parallelepiped grains and surrounded by porous iron oxides (Fig. 3d). The surface morphology of the samples oxidised at

650 C in dry air for 30 min is shown in Fig. 3e where it can be seen that the surfaces were oxidised more severely compared to the samples at 600 C. However, the whole surface was still oxidised unevenly. At a higher magnication, it can be observed that the carbide areas on the surface are oxidised severely, while the matrix area has a good oxidation resistance and are covered with a thin and compact oxide scale (Fig. 3f). Fig. 5a and b show the surface morphologies of the HSS samples oxidised in dry air for 120 min at 600 C and 650 C, respectively. The surface morphologies show no signicant difference to those for 30 min, particular for the case of 600 C (Fig. 5a). For the case of 650 C, after 120 min oxidation, the V-rich MC carbides areas were oxidised heavily and protrude prominently out of the surface. Although the matrix has a relatively strong oxidation resistance,

3606

Q. Zhu et al. / Corrosion Science 53 (2011) 36033611

Fig. 4. (a) FEG-SEM/X-ray mapping of oxidised surface after oxidation at 550 C for 30 min in dry air. (b) High magnication SE image of morphology of vanadium oxide, the sample is tilted 52. (c) High magnication SE image of surface morphology of Cr-rich carbides area after oxidation, the sample is tilted 52.

when the oxidation time increases, the area of the matrix is getting smaller due to the expansion of the oxides from the V-rich MC carbides areas (Fig. 5b). 3.2. FIB and TEM study of oxide scale cross-sections Fig. 6 shows the TEM bright eld image of the cross-section of the oxide layer formed at 600 C for 30 min in dry air. It can be seen that the oxide layer can be distinguished clearly into two sub-layers. However the thickness of the oxide scale uctuates signicantly from less than 0.5 lm (carbides-free matrix) to 1.28 lm (V-rich MC carbides region). Fig. 7a shows the cross-section which was cut off from the V-rich MC carbides area (Fig. 6). The outer layer consists of large oxide crystals, while the inner layer is more porous and the oxides are ne. EDS results indicate that the oxides in the outer layer (Fig. 7b, A in Fig. 7a) are rich of vanadium, manganese and iron, while in the inner layer the oxides consist of chromium and iron oxides (Fig. 7c, B in Fig. 7a). It is noted that the V-rich MC carbides were oxidised and decomposed into small

grains (Fig. 7d, C in Fig. 7a) and ne (Mo, W)-rich spherical M2C carbide was not oxidised after oxidation (Fig. 7e, D in Fig. 7a) in the inner oxide layer. There are no chromium elements appearing in the outer layer, which indicates that chromium oxides only form on the metal surface and remain in the same location during the growth of the oxide scale. With a longer oxidation time, the thin chromium oxide layer becomes the interface between the outer and inner oxide layers. Fig. 8a illustrates the cross-section of oxide scale after oxidation at 650 C for 120 min. The FIB cross-section was cut off from the V-rich carbides area. Fig. 8b and c show the higher magnication TEM bright eld image of the cross-section of oxide scale. It can be seen that in the carbide-free matrix, a thin and compact chromium-iron layer can be found on the matrix surface, and the total thickness of the oxide scale formed in this area is around 1.59 lm. However in the V-rich MC carbides area, a thick duplex-layer oxide scale formed (Fig. 8b and c) whose maximum thickness is 6.16 lm. The inner layer the oxide scale consists of ne oxides and high concentration of porosity and the outer layer consists of big columnar

Q. Zhu et al. / Corrosion Science 53 (2011) 36033611

3607

inner/outer layer interfaces. Si, Mo and W elements are mainly distributed in the inner layer similar to Cr element. Table 2 compares the thickness of oxide scale formed at V-rich MC carbides area and carbides-free matrix after oxidation at 600 and 650 C in dry air for 30 up to 120 min. The results indicate that the carbides-free matrix has a strong oxidation resistance resulting in a thin oxide scale formed after oxidation. At 600 C, the thickness of oxide scale formed on carbides-free matrix has less signicant increase when the oxidation time was increased from 30 to 120 min. However the V-rich MC carbides area is easy to be oxidised, and both of the temperature and oxidation time show signicant inuences on the oxidation rate. The oxide scale formed at V-rich MC carbides area after oxidation at 650 C for 30 min is twice the thickness of that formed at 600 C, while the oxide scales formed on the carbides-free matrix only increases from 0.42 to 0.56 lm. Fig. 10a and b show the XRD patterns of oxidised samples after oxidation at 650 C in dry air for 30 min and 120 min, respectively. The reections of iron substrate present in both spectra indicate that the thickness of oxide scales is less than the depth of X-ray diffraction. It is noted that the main phases of the oxide scale are Fe3O4 (magnetite) and Fe2O3 (hematite). The fact that chromium oxides and vanadium oxides were not detected may be due to a small percentage of these products compared to the whole oxide phases. 4. Discussion The surface morphology of the HSS sample changes greatly with temperature from 550 to 650 C, which indicates that temperature has a signicant inuence on the oxidation behaviour of the chosen HSS material (Fig. 3). The HSS samples have very good oxidation resistance at low temperatures, i.e. 550 and 600 C. At 550 C, there is a thin, compact chromium-iron oxide layer covering the top of the HSS matrix. The iron oxides are growing among or on this layer as a result of Fe ions diffusing through the layer from the substrate. The sample surfaces were oxidised unevenly at 650 C. The heavy oxidation occurred at the V-rich MC carbides area is due to a low thermal stability of V-rich MC carbides and high free energy at the carbides/matrix interfaces (Fig. 8b and d) [2123]. At the same time, the carbide-free matrix shows a good oxidation resistance with a thin compact chromium-iron rich oxide layer. Differential oxidation is expected to occur due to a high level of alloying elements in the HSS samples. These alloy elements possess different levels of oxidation resistance and precipitate preferentially as different types of carbides or dissolve in the matrix, leading to a different oxidation resistance between the matrix and the carbides. The results conrm the previous research work [3,8,12], where the MC carbides are oxidised heavily even though MC carbides dissolve chromium, and M7C3 carbides have a very good oxidation resistance. Vanadium carbides have a great afnity for oxygen at high oxygen partial pressure [24]. Hence, during the oxidation, vanadium ions (driven by the high partial pressure of oxygen) diffuse from the matrix and carbides across the thin oxide scale to feed the oxide layer formed on the outer layer. The development of vanadium rich oxide rod-like and parallelepiped crystals as shown in Fig. 4b and Fig. 7b can be considered as evidence of the diffusion of vanadium ions. The possible oxidation reaction of vanadium carbides are listed as follows:

Fig. 5. SE images of surface morphology of specimens after oxidation in dry air for 120 min, (a) at 600 C, and (b) at 650 C.

Fig. 6. TEM bright eld cross-section image of oxide scale formed at 600 C in dry air after 30 min oxidation.

oxide crystals. EDS analysis (Fig. 8d) indicates that the columnar oxide crystals are pure iron oxide and vanadium oxides formed on the top of the oxide scale. There is a signicant difference in thickness of oxide scale formed in the V-rich MC carbides area and the carbide-free matrix. Fig. 9 shows the TEM/X-ray maps of the cross-section of the oxide scale. It conrms that vanadium carbides were oxidised and the vanadium oxides diffuse to the top of the oxide scale. Manganese was also found on the top of the scale. Chromium accumulates at the substrate/inner oxide layer and

VCs O2 g ! VOs COg 2VCs 3:5O2 g ! V2 O5 s 2COg

1 2

Although the morphology of VO crystal is similar to that of V2O5 whose crystal structure is orthorhombic, thermodynamic calculations of the Gibbs free energy [25] that involved the reaction of

3608

Q. Zhu et al. / Corrosion Science 53 (2011) 36033611

Fig. 7. (a) High magnication TEM bright eld cross-section image of oxide scale on a vanadium carbides (MC) area formed at 600 C in dry air after 30 min oxidation. (b) EDS spectra of area A; (c) EDS spectra of area B; (d) EDS spectra of area C; (e) EDS spectra of area D.

vanadium carbide with oxygen to form vanadium oxide VO or V2O5 show that the formation of V2O5 (DG = 1357 kJ/mol, 650 C) is more favoured than the formation of VO (DG = 457 kJ/mol, 650 C). Therefore, the oxidation product should be V2O5. The steel matrix has a good oxidation resistance because chromium elements are distributed homogenously. It is well known that chromium is commonly added to steel to increase the oxidation resistance [26]. Chromium has a greater afnity to oxygen than iron, and a Cr2O3 lm is quickly formed and thickens slowly by the outward movement of cations at the base of the scale. This chromium rich oxide layer owns a characteristic of a low ionic conductivity leading to a low oxidation rate. In the present work, a complex layer consisting of iron and chromium oxides at the matrix/oxides interface was formed. The formation of this layer may display a protective behaviour, and reducing the oxidation rate, especially when the (Cr, Fe)3O4 spinel are formed in the layer [27,28]. Cross-section investigation indicates that differential oxidation produces a different growth mechanism of oxide scale. At 600 C, V-rich MC carbides were oxidised and decomposed, while (Mo, W)-rich ne M2C carbides were not oxidised and retained the original shape (Fig. 7a). Because V-rich MC carbides are easier to be oxidised than the carbide-free matrix as discussed above, the oxide scale formed on these areas is much thicker than that in the carbides-free matrix. The thickness of oxide formed on the V-rich MC carbides area is more than twice the thickness on the

carbide-free matrix in Fig. 6. FIB cross-section observation indicates that the thickness of the oxide scale after oxidation for 120 min on the HSS matrix changes little compared to that for 30 min, but the oxide layer in the V-rich MC carbides areas is clearly getting thicker. After 120 min oxidation at 650 C, the carbide-free matrix area was covered with a thin oxide layer with 1.59 lm thickness, in which a compact chromium-iron sub-layer is approximately 250 nm thick (Fig. 8b); while heavy oxidation took place in V-rich MC carbide area where the total thickness of the oxide scale is around 6.16 lm (Table 2 and Fig. 8b). Both Figs. 7 and 8 show that the oxide scale formed in V-rich MC carbides area has a duplex-layer structure. It is believed that the structure of oxide scale shown in Fig. 7a indicates the initial stage of oxide scale formation due to a lower oxidation temperature and shorter oxidation period. Results of TEM/X-ray mapping of cross-section of the oxide scale formed at 650 C for 120 min indicate that a ne grained structure decient in iron and rich in chromium oxide layer was found next to the substrate; the outer layer consists of columnar iron oxide crystals with vanadium oxides covering the top surface (Fig. 8c). These results are similar to Molinari et al.s research [10] which indicates that three sub-layers in the oxide scale after oxidation at 700 C for 24 h are: (i.) a thin a-Fe2O3 outer layer; (ii.) a thick and porous c-Fe2O3 intermediate layer; (iii.) the inner M3O4 spinel (M = Fe, Cr, V) produced by the internal oxidation. X-ray diffraction analysis in the present work shows that the oxide layer mainly consists of Fe3O4 (magnetite) and Fe2O3 (hematite).

Q. Zhu et al. / Corrosion Science 53 (2011) 36033611

3609

Fig. 8. (a) Overview TEM bright eld cross-section image of oxide scale formed at 650 C in dry air after 120 min oxidation, (b) and (c) higher magnied TEM bright eld cross-section images of oxide scale, (d) EDS spectra of areas A and B labelled in (c).

FeO (wstite) was not detected from X-ray diffraction analysis of the samples. The reason might be that although FeO (wstite) is stable at 650 C, its formation was suppressed by the existence of chromium which increases the number of defects in the crystal structure of FeO (wstite) and favours its transformation to magnetite or chromium rich spinel structure [27,29]. The absence of FeO (wstite) in the oxide layer is in agreement with the previous observations found in the oxidation studies of work rolls [3,8,30]. There is a clear interface between the inner and outer layers, which is also the initial metal surface (Fig. 8b and c). TEM/X-ray mapping indicates that chromium and silica accumulate at the interface. The accumulation of silica at the scale/substrate interface is frequently reported in the literature [21,31]. It has been suggested that the formation of a silica sub-scale can contribute signicantly to the oxidation resistance just like chromium does [32,33]. However, in the present case, the silica at the interface can hardly act as a diffusion barrier since it does not form a continuous lm. We believe that this chromium and silicon rich band is a marker showing the original metal/gas interface. This implies that the inner layer grows mainly by inward oxygen diffusion while outer layer grows by outward cation diffusion. High porosity is found in the inner layer which means that inner layer does not provide much resistance against oxidation. The formation of pores in the inner layer may be due to two reasons. First, Fe and V ions diffuse outwards leaves behind a vacancy; second, the escape of CO produced by the

reduction of carbides leave the voids in the inner layer [34]. The columnar iron oxide outer layer is believed to be formed due to Fe ions diffusing outward from the inner layer. The different oxidation behaviour between the carbide-free matrix and V-rich MC carbide areas leads to an inhomogeneous surface after oxidation. This has an important inuence on the contact friction and wear of hot work rolls. From our experimental results there are two mechanisms of oxidation behaviour happening at V-rich MC carbides area and carbides-free matrix area, respectively. At carbide-free matrix area, the oxidation behaviour is similar to the behaviour of FeCr alloy. At the beginning of the oxidation process both chromium and iron elements are oxidised, forming a thin oxide layer on the matrix surface. Then, a competitive process starts. As the chromium oxide is more stable than iron oxide, chromium diffuses to the metal/ oxide interface and is oxidised in this region where the partial pressure of available oxygen is lower than that in the external oxidising gas. Furthermore iron diffuses across the internal chromium oxide layer to the external oxide/gas interface, reacting with oxygen and producing the outer growth of the scale [27,35,36]. In this way, a chromium oxide layer grows beneath the iron oxide layer. This compact chromium rich oxide inner layer effectively prohibits the iron elements diffusing to the external interface of oxide scale, and reduces the oxidation rate signicantly. At the V-rich MC carbides area, however, a continuous and compact chromium oxide

3610

Q. Zhu et al. / Corrosion Science 53 (2011) 36033611

Fig. 9. TEM/X-ray mapping of the oxide scale formed at 650 C after oxidation in dry air for 120 min.

sub-layer does not form on the original metal surface due to high free energy on carbides/matrix interfaces and a great afnity for oxygen of V-rich MC carbides. The aggressive oxidation in this region leads to a porous chromium oxide inner sub-layer which provides sufcient channels (mircocracks, pores) for iron and

vanadium elements diffusing outwards and oxygen diffusing inwards as well. As the oxidation processes, an outer iron oxide layer consisting of columnar crystals forms due to a sufcient supply of iron elements from the substrate and fast oxidation rate. Vanadium element has a great oxygen afnity at high oxygen

Q. Zhu et al. / Corrosion Science 53 (2011) 36033611 Table 2 Comparison of oxide scale thickness formed at vanadium carbides area and carbidesfree area after oxidation at 600 and 650 C in dry air for 30 up to 120 min. The thickness values were obtained by measuring from TEM images. Temperature/C Oxidation time/min 30 120 30 120 Oxide scale thickness/lm MC rich area 1.28 3.18 2.31 6.16 Carbides-free matrix 0.42 0.65 0.56 1.59

3611

600 650

Fig. 10. XRD patterns of the oxidised sample at 650 C in dry air. (a) for 30 min and (b) for 120 min.

partial pressure; hence, the vanadium oxides appear on top of the oxide scale after oxidation. 5. Conclusions Surface morphologies and microstructure of oxide scale after oxidation in dry air were investigated and the related oxidation mechanism is discussed. The following conclusions can be drawn. Differential oxidation occurred during the oxidation due to highly alloyed component in the high-speed steel. V-rich MC carbide is easily oxidised because it has a great afnity for oxygen. V-rich MC carbides areas are oxidised severely due to the low thermal stability of MC and high free energy at the carbides/matrix interfaces. Duplex-layer oxide scales form in these areas after oxidation consisting of porous, ne crystalline (Cr, Fe)-rich oxides inner layer and large columnar iron oxides outer layer with vanadium oxides on the top. The carbides-free matrix shows a good oxidation resistance which may be due to a compact and thin (Cr, Fe)3O4 spinel reducing the oxidation rate. Due to the different oxidation behaviour between the carbide-free matrix and V-rich MC carbide areas, the surface morphology becomes inhomogeneous. References
[1] L. Surez, P. Rodrguez-Calvillo, Y. Houbaert, R. Cols, Oxidation of ultra low carbon and silicon bearing steels, Corros. Sci. 52 (2010) 20442049. [2] C. Vergne, C. Boher, R. Gras, C. Levaillant, Inuence of oxides on friction in hot rolling: experimental investigations and tribological modelling, Wear 260 (2006) 957975. [3] H.H. Kim, J.W. Lim, J.J. Lee, Oxidation behavior of high-speed steels in dry and wet atmospheres, ISIJ Int. 43 (2003) 19831988.

[4] S. Spuzic, K.N. Strafford, C. Subramanian, G. Savage, Wear of hot rolling mill rolls: an overview, Wear 176 (1994) 261271. [5] M. Boccalini, A. Sinatora, Microstructure and wear resistance of high-speed steels for rolling mill rolls, in: Proceedings of the 6th International Tooling Conference, Karlstadt University, September, 2002, pp. 425438. [6] J.C.G. Milan, M.A. Carvalho, R.R. Xavier, S.D. Franco, J.D.B. De Mello, Effect of temperature, normal load and pre-oxidation on the sliding wear of multicomponent ferrous alloys, Wear 259 (2005) 412423. [7] R. Cols, J. Ramrez, I. Sandoval, J.C. Morales, L.A. Leduc, Damage in hot rolling work rolls, Wear 230 (1999) 5660. [8] Q. Zhu, H.T. Zhu, A.K. Tieu, M. Reid, L.C. Zhang, In-situ investigation of oxidation behaviour in high-speed steel roll material under dry and humid atmospheres, Corros. Sci. 52 (2010) 27072715. [9] F.J. Prez, L. Martnez, M.P. Hierro, C. Gmez, A.L. Portela, G.N. Pucci, D. Duday, J. Lecomte-Beckers, Y. Greday, Corrosion behaviour of different hot rolled steels, Corros. Sci. 48 (2006) 472480. [10] A. Molinari, G. Straffelin, A. Tomasi, A. Biggi, G. Corbo, Oxidation behaviour of ledeburitic steel for hot rolls, Mater. Sci. Eng., A 280 (2000) 255262. [11] L. Zhou, F. Liu, C.S. Liu, D.L. Sun, High temperature oxidation behaviour of highspeed steel for roll in water vapour, Trans. Mater. Heat Treat. 25 (2004) 134 138. [12] A. Molinari, M. Pellizzari, A. Biggi, G. Corbo, A. Tremea, Primary carbides in spincast HSS for hot rolls and their effect on the oxidation behaviour, in: Proceedings of the 6th International Tooling Conference, Karlstadt University, September, 2002, pp. 365377. [13] M. Pellizzari, A. Molinari, G. Straffelini, Tribological behaviour of hot rolling rolls, Wear 259 (2005) 12811289. [14] P.A. Munther, J.G. Lenard, A study of friction during hot rolling of steels, Scand. J. Metall. 26 (1997) 231240. [15] C. Vergne, C. Boher, C. Levaillant, R. Gras, Analysis of the friction and wear behaviour of hot work tool scale: application to the hot rolling process, Wear 250 (2001) 322333. [16] F. Liu, J.E. Tang, H. Asteman, J.-E. Svensson, L.-G. Johansson, M. Halvarsson, Investigation of the evolution of the oxide scale formed on 310 stainless steel oxidised at 600 C in oxygen with 40% water vapour using FIB and TEM, Oxid. Met. 71 (2009) 77105. [17] M. Sun, X. Wu, E.H. Han, J. Rao, Microstructural characteristics of oxide scales grown on stainless steel exposed to supercritical water, Scripta Mater. 61 (2009) 996999. [18] B. Pujilaksono, T. Jonsson, M. Halvarsson, J.-E. Svensson, L.-G. Johansson, Oxidation of iron at 400-600 C in dry and wet O2, Corros. Sci. 52 (2010) 1560 1569. [19] M.A.A. Motin, J. Zhang, P.R. Munroe, D.J. Young, Internal oxidation and metal dusting of FeSi alloys, Corros. Sci. 52 (2010) 32803286. [20] N.F. Garza-Montes-de-Oca, W.M. Rainforth, Wear mechanisms experienced by a work roll grade high-speed steel under different environmental conditions, Wear 267 (2009) 441448. [21] D.N. Hanlon, W.M. Rainforth, The rolling sliding wear response of conventionally processed and spray formed high-speed steel at ambient and elevated temperature, Wear 255 (2003) 956966. [22] C. Rodenburg, W.M. Rainforth, A quantitative analysis of the inuence of carbides size distributions on wear behaviour of high-speed steel in dry rolling/sliding contact, Acta Mater. 55 (2007) 24432454. [23] A. Molinari, G. Straffelini, A. Tomasi, A. Bigi, G. Corbo, Inuence of microstructure and chromium content on oxidation behaviour of spin cast high-speed steels, Mater. Sci. Technol. 17 (2001) 425430. [24] C. Santaf, C. Borgianni, Study of the oxidation kinetics of vanadium carbide, Oxid. Met. 9 (1975) 415425. [25] JANAF, Thermochemical tables, Ofce of standard reference data, Natural Bureau of Standards, Washington D.C., USA, 1971. [26] Ulick R. Evans, The corrosion and oxidation of metals, second supplementary volume, Edward Arnold, 1976. [27] A.S. Khanna, Introduction to high temperature oxidation and corrosion, ASM International, 2002. [28] I. Kvernes, M. Oliveira, P. Kofstad, High temperature oxidation of Fe-13Cr-xAl alloys in air/H2 O vapour mixtures, Corros. Sci. 17 (1977) 237252. [29] Y.N. Chang, F.I. Wei, High temperature oxidation of low alloy steels, J. Mater. Sci. 24 (1989) 1422. [30] V. Gonzlze, R. Rodrguez, Z. Haduck, R. Cols, Modelling oxidation of hot rolling work rolls, Ironmaking Steelmaking 28 (2001) 470473. [31] L. Mikkelsen, S. Linderoth, J.B. Bilde-Srensen, The effect of silicon addition on the high temperature oxidation of a FeCr alloy, Mater. Sci. Forum 461-464 (2004) 117122. [32] Y. Wouters, G. Bamba, A. Galerie, M. Mermous, J.-P. Petit, Oxygen and water vapour oxidation of 15Cr ferritic stainless steels with different silicon, Mater. Sci. Forum 461-464 (2004) 839846. [33] C. Pettersson, T. Jonsson, C. Proff, M. Halvarsson, J.-E. Svensson, L.-G. Johansson, High temperature oxidation of the austenitic (35Fe27Cr31Ni) alloy sanicro 28 in O2 + H2 O environment, Oxid. Met. 74 (2010) 93111. [34] R.Y. Chen, W.Y.D. Yuen, Review of the high-temperature oxidation of iron and carbon steels in air or oxygen, Oxid. Met. 59 (2003) 433468. [35] M.J. Monteiro, F.C. Rizzo, Effect of chromium content on the oxidation behaviour of high-speed steels under dry and moist air environments, Mater. Sci. Forum 522-523 (2006) 171180. [36] P. Kofstad, High temperature corrosion, Elsevier Applied Science Publishers Ltd, London and New York, 1988.