Wear 302 (2013) 1310–1318

Contents lists available at SciVerse ScienceDirect

Wear

0043-16

http://d

* Corr

E-m

journal homepage: www.elsevier.com/locate/wear

A simulation of wear behaviour of high-speed steel hot rolls by meansof high temperature pin-on-disc tests

Hongtao Zhu a,*, Qiang Zhu a, Anh Kiet Tieu a, Buyung Kosasih a, Charlie Kong b

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

a r t i c l e i n f o

Article history:

Received 1 September 2012

Received in revised form

13 November 2012

Accepted 17 November 2012Available online 29 November 2012

Keywords:

Oxide scale

Tribology

High-speed steel

Pin-on-disc

48/$ - see front matter & 2012 Elsevier B.V. A

x.doi.org/10.1016/j.wear.2012.11.032

esponding author. Tel.: þ61 2 4221 4549; fax

ail address: hongtao@uow.edu.au (H. Zhu).

a b s t r a c t

In this paper, a high temperature pin-on-disc configuration was used to simulate the contact

established between a high-speed steel (HSS) work roll and a hot strip material in hot rolling, in

which the pin represented the HSS roll and the disc represented a strip steel. The pin surfaces were

oxidised due to the heat transfer from the disc while they were in contact. This work focused on the

contact behaviour of the oxide scale in the roll bite during hot rolling while the testing temperature was

close to the rolling temperature, the Hertzian pressure was similar to the contact pressure and the

sliding speed was close to those in the roll bite. The coefficient of friction during the tests was

monitored and recorded in-situ. It was found that the evolution of the coefficient of friction could be

divided into three stages. Associated with the evolution of the coefficient of friction, the morphologies

and micro-structures on the surface of pin were characterised by means of SEM, FIB and TEM

techniques to study the tribological behaviour of oxide scale in contacts. The results indicated that

the wear mechanism of pin surface varies in different stages. At the stages I and II, the oxide scale on

the pin surface is significantly deformed. At the stage III, which the coefficient of friction is stable, the

wear mechanism is a mixture of adhesion, abrasion and oxidation. The oxide transfer from the mild

carbon steel disc to HSS pin significantly contributed to the scale formed on the HSS pin surface.

& 2012 Elsevier B.V. All rights reserved.

1. Introduction

High-speed steels (HSS) have become the most favourable mate-rial choice of manufacturing hot rolls for their excellent mechanicalperformance, i.e. hardness and wear resistance, at elevated tempera-tures. During hot rolling process, a superficial oxide scale forms onthe work rolls due to thermal cyclic working conditions. This oxidescale plays a major role in hot rolling because it allows the startingand driving of metal sheet between the rolls [1,2]. Under thermo-mechanical working conditions this scale grows to a critical thicknessand then deteriorates and causes damage by inlaying and breakingdown on the surfaces of rolled product and tools. The wear mechan-isms of the work rolls in hot rolling are well known, i.e. thermalfatigue, abrasion, adhesion and oxidation. However, little attentionhas been given to the oxidation of rolling-mill roll surface and itseffects on the tribological behaviour [3]. The tribological behaviour ofoxides is very complex. It could be abrasive and therefore detrimental,or lubricated and form a protection against wear, either way it isinfluenced by the nature and physical properties of the oxide scalesinto which it makes contact [4–7].

ll rights reserved.

: þ61 2 4221 3101.

Although it is still a challenge for researchers to simulate hotrolling conditions in a laboratory, several test configurations havebeen developed. Many researchers used a disc-on-disc configura-tion to study the wear mechanism of the hot rolls [8–18], onesimulating the hot strip, the other one simulating the roll. Someresearchers referred to a pin-on-disc configuration [4,20,21]. Thetribological properties are system-dependent and have beenfound sensitive to the condition changes of the contact surfaces,such as the formation, accumulation, agglomeration and sinteringof wear particles on the surface [18,19]. Hokkirigawa et al. [20]developed an in-situ CCD microscope pin-on-disc tribosystem toanalyse the microscopic sliding wear mechanism of oxide films onhot roll surfaces. Using this system the dynamic wear process ofoxide films on hot roll surface were observed successfully. Vergneet al. [4] carried out an extensive investigation of the interactionbetween the oxides formed on the surface of a work roll and thecoefficient of friction established when hot rolling steel by pin-on-disc tests. Joos et al. [21] tried to identify and assess the wearmechanism involved in the hot tribological behaviour of highspeed steel roll grades by carrying out a series of pin-on-disc testsat elevated temperatures. They argued that the tribologicalbehaviour of roll grades at high temperature was strongly affectedby the nature, thickness, and thermo-mechanical properties oftheir oxide scales involved in the contact. In the current work, a

H. Zhu et al. / Wear 302 (2013) 1310–1318 1311

pin-on-disc test rig was used to identify the role of oxide scale onthe tribological behaviour, in which the pin represented the HSSroll and the disc represented a steel strip. This work is an attemptto study the wear of the high-speed steel, especially concerningthe oxide evolution during contact in the roll bite of hot striprolling. Although the continuously sliding contact in pin-on-disctests is different from cyclic sliding/rolling contact of work roll onhot strip, it can help us to gain a fundamental understanding ofthe role of the oxide scale in high speed work roll wear in hotrolling. The testing temperature was chosen close to the rollingtemperature, the Hertzian pressure was similar to the contactpressure and the sliding speeds were close to those in the rollingbite of hot rolled strip. The coefficient of friction was monitoredand recorded in-situ during the tribological tests. Scanningelectron microscope (SEM), focused ion beam (FIB) and transmis-sion electron microscope (TEM) techniques were used to char-acterise the morphologies and micro structures of the contactzones on the surface of the pin. The evolution of oxide on HSS pinduring the test, especially scale spallation and oxide transfer fromhot disc, and its influence on wear mechanism at the differentstages have been investigated.

2. Experimental procedure

The pin-on-disc tests were carried out on the CETR tribometer.This device is not a hot rolling simulator, but this laboratoryapparatus can reveal the oxidational wear mechanism involved inhot rolling [21]. In a pin-on-disc configuration, the pin repre-sented the HSS grade material and the disc represented a stripsteel (Fig. 1). The chemical compositions of these two materialsare listed in Tables 1 and 2, respectively. The pin is manufactured

Fig. 1. Schematic illustration of high temperature pin-on-disc test configuration.

Table 1Chemical composition of the HSS pin material (wt%), analysed by X-ray fluores-

cence spectroscopy.

Fe C Ni Mn Cr Mo V W Si P S

Balanced 1.96 0.78 1.26 4.85 4.47 4.00 3.40 0.99 0.028 0.034

Table 2Chemical composition of the disc material (wt%), analysed by atomic emission spectro

Fe C P Mn Si S

Balanced 0.215 0.015 0.79 0.34 0.029

into a mushroom shape with a 3 mm radius hemispherical end asshown in Fig. 1.

The tribological tests were carried out in the following twosteps: (i) in the first step, the disc was heated up to 900 1C in aheating chamber while the pin was kept at room temperatureoutside the heating chamber. A k-type thermocouple was placedunderneath the disc to monitor its temperature during the tests;(ii) in the second step, when the temperature of the disc reached900 1C, the pin was brought into contact with the disc and held for20 min for pre-oxidation before stating to rotate the disc. Theduration of the following wear test was up to 1 h. A K-typethermocouple was embedded into the pin, approximately 2 mmvertical distant away from the contact (as demonstrated in Fig. 1)to monitor its temperature during the test. The heating of the discwas not interrupted in order to maintain the high temperature(900 1C) during the sliding test. After each test, the pin was pulledup out of the heating chamber and cooled in the air.

Table 3 shows the operating conditions of the tribologicaltests. Test parameters such as the Hertzian pressure were chosenmild and close to the practical hot rolling conditions in order toobserve the behaviour of the antagonistic oxide scales in thecontact zone without destroying them too fast. The coefficient offriction was monitored and recorded in-situ during the test. Theinterrupted tests were performed to understand the interaction ofthe oxides corresponding to the evolution of friction coefficientcurve during the tests. After the tribological tests, SEM, FIB andTEM were used to characterise the morphologies, cross-sectionsand microstructures of the pin surface at both the contact andnon-contact zones.

3. Results

Fig. 2 shows the evolution of temperatures of disc and pinsurface during the test. It can be seen that the temperature of thedisc remained stable at 900 1C throughout the whole test whilethe pin quickly reached approximately 660 1C (less than 5 min)after making contact with the heated disc. The pin was keptcontacting with the heated disc for 20 min for pre-oxidationbefore the test started. It should be noted that this monitoredpin temperature refers to the location 2 mm vertically away fromthe sliding contact zone.

Fig. 3a shows a friction coefficient curve of a typical tribolo-gical test under a normal load of 5 N and a sliding speed 0.05 m/s.It can be seen that the friction stabilised in a very short time (lessthan 300 s) from the start of the test and the friction coefficientremained at approximately 0.26–0.28 for the most of the testperiod. According to the friction coefficient curve, the pin-on-disctest can be divided into three stages, as shown in Fig. 3b; (i) thefirst stage, noted Stage I, the friction coefficient decreased

scopy.

Ni Cr Mo Cu Al

0.006 0.020 o0.002 0.014 o0.003

Table 3Operating conditions of the pin-on-disc test.

Normal

load

Hertzian

pressure

Sliding

speed

Disc

temperature

Testing duration

5 N 650 MPa 0.05 m/s 900 1C 90 s, 150 s, and

1 h

Fig. 2. Temperatures of disc and pin surface during a typical tribological test.

Fig. 3. (a) Evolution of friction coefficient during the pin-on-disc test with

conditions of normal load 5 N, and sliding speed 0.05 m/s; (b) three tribological

stages and two interrupted points considered in a typical pin-on-disc test.

Fig. 4. (a) Secondary electron (SE) image of the pin wear track after the

tribological test at normal load 5 N and sliding speed 0.05 m/s for 90 s (in Stage

I), (b) TEM bright field image of the cross section of the pin wear track, FIB was

used to prepare the cross section.

H. Zhu et al. / Wear 302 (2013) 1310–13181312

immediately after the start of the wear test; (ii) the second stage,noted Stage II, corresponds to an increase in the friction coeffi-cient after the minimum value; (iii) the last stage, noted Stage III,is the stabilisation step of the friction. To better understand thetribological behaviour, two interrupted pin-on-disc tests, whichhave been labelled in Fig. 3b, were performed to investigate theevolution of tribological contact mechanisms during the test. Thefirst interrupted test corresponds to the end of Stage I (around90 s from the start of the test), the second interrupted test refersto the middle of Stage II (around 150 s from the start of the test).

Fig. 4a shows the wear track of the pin surface after aninterrupted pin-on-disc test which terminated at 90 s since the

start of the test in Stage I. According to the friction coefficientcurve shown in Fig. 3b, it is near the end of stage I. It can be seen acompact and smooth ‘‘glaze’’ oxide scale was visible on thesurface of the worn pin after the test (Fig. 4a). Fig. 4b shows theTEM bright field image of the cross-section of the pin wear trackwhere this ‘‘glaze’’ oxide scale was approximately 850 nm thick.The oxide scale was highly compressed because the microstruc-ture of the scale is dense and free of pores. Although the interfacebetween the oxide scale and the HSS substrate can be clearlyobserved, the bond between the oxide scale and HSS substrate isstrong because there are no visible cracks and micro-pores at theinterface. Underneath the oxide the HSS substrate has beendeformed into the banding grains in an array and directionconsistent with a sliding direction.

Fig. 5a shows the wear track of the pin surface after aninterrupted pin-on-disc test which stopped at 150 s since thestart of the test (in Stage II). The morphology of the surface of thepin is quite different from that at the Stage I. A discontinuous‘‘glaze’’ oxide scale covers the worn surface of the pin. A highermagnification SEM observation on the wear track indicates thatthe discontinuity of the ‘‘glaze’’ oxide scale was due to the oxide

Fig. 5. (a) Secondary electron (SE) image of the pin wear track after the pin-on-disc test at normal load 5 N and sliding speed 0.05 m/s for 150 s (in Stage II), (b) higher

magnified SE image of the pin wear track, the sample is titled 531, and (c) TEM bright field image of the cross section of the pin wear track, FIB was used to prepare the

cross section.

H. Zhu et al. / Wear 302 (2013) 1310–1318 1313

scale spalling in some areas of the sliding contact zone on the pinsurface (Fig. 5b). In this spallation area the sub-surface of the pinwas re-oxidised because it had no protection from original oxidescale. Fig. 5c shows the TEM bright field image of the cross-section of the contact zone (non-spallation area) of the pinshowing a compact oxide scale of approximately 800 nm coveringthe pin surface. The bonding of the oxide scale and HSS substratewas still very good because there are no visible cracks and micro-pores at the interface.

Fig. 6a shows the wear track of the pin after 1 h’s pin-on-disctest (Stage III). It can be seen that the morphology of the weartrack at Stage III is different to Stage II. The contact zone of the pinsurface is covered by a discontinuous ‘‘glaze’’ oxide scale; how-ever, the adherent non-spallation ‘‘glaze’’ oxide scale is muchthicker compared with that at Stage II. As shown in Fig. 6d, theoxide spallation is prevalent and serious in the sliding contactzone. The fine oxide debris can be clearly observed left inside thecontact region; and the crack propagation occurs in the contactedoxide. As shown in Fig. 6b and c, there are clear interfacesbetween the sliding contact zone and non-contact area either atthe front and tail of wear track.

Fig. 7a shows the SE and TEM bright field images of the crosssection of wear track after 1 h’s pin-on-disc test (end of Stage III).It can be seen that the pin wear track is covered by an oxide scaleof non-uniform thickness and at the centre of the wear track, asmall piece of material has been rubbed off and cracks under-neath propagated into the substrate. It is possible that thismaterial rubbed off is vanadium MC carbide. Fig. 7b shows thedetailed microstructure of the cross-section of the pin wear trackwhere ‘‘glaze’’ oxide scale is adhering to the substrate and afterthe pin-on-disc test, a thick oxide has developed in the contact

zone. The total thickness of the adherent non-spallation oxidescale was approximately 9.5 mm, which is significantly higherthan that in Stage I (850 nm) and Stage II (800 nm). From a topdown view the ‘‘glaze’’ oxide scale seems to be compact andsmooth, but there are two large cracks inside the oxide scaleparallel to the surface, which actually divides the oxide scale intothree sub-layers (Fig. 7b). These two large cracks weaken theoxide integrity and promote the oxide spallation. Fig. 7c showsvery fine-grained oxides (less than 200 nm) formed at the contactfrontier; while the inner part of scale consists of a little largeroxide grain at about 400 nm (Fig. 7d). Fig. 8 shows the TEM/X-raymapping of the inner part of the oxide scale and its interfacebetween the oxide scale and HSS substrate. The results indicatedthat the adherent oxide scale mainly consists of iron oxides(Fe2O3 and Fe3O4). Chromium oxides and vanadium oxides onlyappeared in the most inner part of the oxide scale and at theinterface between the oxide scale and HSS substrate.

Fig. 9a shows the morphology of the pin surface outside thecontact zone (wear track) but very close, after the test for 150 s atthe end of Stage II. The pin surface is covered by iron oxides andlarge parallelepiped vanadium oxides. Fig. 9b shows the TEM brightfield image of the cross-section of this non-contact surface of theoxidised pin where the oxide scale consists of a large crystallineouter layer of iron oxide (Fe2O3) and a fine-grained inner oxide layerof Fe3O4, Cr2O3 and (Fe, Cr)-rich spinel oxides. There is a porous andclear interface between the oxide scale and HSS substrate. The oxidescale is approximately 1 mm thick. The oxidised pin surface becamerough due to protrusions of iron oxides and vanadium oxides.

Fig. 10a shows the morphology of the non-contact area of thepin, very close to the sliding contact zone after 1 h test at the endof Stage III. It can be seen that the surface is covered by a

Fig. 6. (a) Secondary electron (SE) image of the pin wear track after the pin-on-disc test at normal load 5 N and sliding speed 0.05 m/s for 1 h (end of Stage III),

(b) morphology of the wear track front, (c) morphology of the wear track tail, and (d) spallation and wear debris at the wear track.

Fig. 7. (a) Secondary electron (SE) image of the cross-section of the pin wear track after the pin-on-disc test at normal load 5 N and sliding speed 0.05 m/s for 1 h (end of

Stage III), (b) TEM bright field image of cross-section of the wear track, (c) higher magnified TEM bright field image of the oxide scale at the contact frontier, and (d) higher

magnified TEM bright field image of the oxide scale near the interface of oxide scale and the HSS matrix.

H. Zhu et al. / Wear 302 (2013) 1310–13181314

Fig. 8. TEM/X-ray mapping of the cross-section pin wear track after tribological test at normal load 5 N and sliding speed 0.05 m/s for 1 h (inner part of the oxide scale and

the HSS matrix).

H. Zhu et al. / Wear 302 (2013) 1310–1318 1315

continuous porous oxide scale with iron oxide whiskers andparallelepiped vanadium oxides protruding from the scale.Fig. 10b clearly shows the morphology of the scale in a cross-sectional view. TEM/X-ray mapping of the cross-section of theoxide scale revealed that the scale consists of two sub-layers, withthe outer layer being mainly large columnar iron oxide (Fe2O3)and the inner layer a mixture of fine iron oxides and chromiumoxides (Fe3O4, Cr2O3 and (Fe, Cr)-spinel oxides). The parallele-piped vanadium oxides are protruding from the oxide scale (asseen in Fig. 10c). The oxide scale in non-contact area was around2.8 mm thick after 1 h pin-on-disc test (the protruding vanadiumoxides were not included in measuring the scale thickness).

Fig. 11 shows the SE image of the cross-section of oxide scaleformed on the mild carbon steel disc after 1 h pin-on-disc test at

900 1C. The oxide scale was around 70 mm thick after the test. Theoxide scale consists of three sub-layers; the outer layer ishaematite (Fe2O3) approximately 4.7 mm, the columnar crystal-line middle layer is magnetite (Fe3O4) approximately 19 mm thick,and the large-grained inner layer is wustite (FeO) approximate46 mm thick.

4. Discussion

The temperature of the pin and disc were well controlledduring the pin-on-disc tests (Fig. 2). The surface temperature ofthe disc was kept at 900 1C with fluctuations less than 2 1C. Thesurface temperature of the pin reached 660 1C in a very short time

Fig. 9. (a) Secondary electron (SE) image of non-contact pin surface after the pin-

on-disc test at normal load 5 N and sliding speed 0.05 m/s for 150 s in Stage II,

(b) TEM bright field image of the cross section of the non-contact zone, FIB was

used to prepare the cross section.

H. Zhu et al. / Wear 302 (2013) 1310–13181316

(less than 300 s) after being into contact with the disc due tothermal conduction and radiation. Although the temperature ofthe pin was monitored from a position located 2 mm above thecontact zone, by comparing the morphologies of pin close to thecontact zone (Figs. 9 and 10) with our previous study [22–24], thetemperature of the pin in the contact zone was probably around700 1C. During hot rolling, the temperature of strip ranges from800 to 1200 1C, and the flash temperature of the surface of thework rolls can reach up to 650 1C, or even up to 700 1C due to heatgenerated by plastic deformation and friction [25–31]. Therefore,the pin-on-disc tests reproduced the temperatures of hot rollingvery well. The temperature of disc was well kept at 900 1C duringthe tests which led a thick oxide scale forming on the surface(around 70 mm thick after 1 h). The pin was pre-oxidised for

nearly 20 min after being in contact with the disc before the pin-on-disc test began, the contact mechanism between the pin anddisc is oxide-to-oxide, which reproduces the real contact condi-tion in the hot strip rolling.

The testing condition of the pin-on-disc test with a normalload 5 N and sliding speed 0.05 m/s is close to the tribologicalconditions in the roll bite during hot rolling. Three stages wereidentified after analysing friction coefficient curve: Stage I wherethe friction coefficient decreased dramatically from around 0.4(static contact) to 0.23 (dynamic contact) in a very short time(less than 100 s). Stage II where the friction coefficient increasedfrom the minimum value (0.23) to approximately 0.29 in a shorttime (about 150 s); and Stage III where the friction coefficientremained at about 0.27 for the rest of the test. Vernge et al. [4]reported the similar phenomenon. But, Vernge defined the periodbefore the friction stabilised as a running-in period. In the currentwork the running-in period consists of Stages I and II, and Stage IIIcorresponds to the friction stabilisation period.

SEM and TEM investigations (Figs. 4 and 5) showed that a thin,compact, and smooth ‘‘glaze’’ oxide scale around 800–850 nmformed on contact zone of the pin during the running-in period ofStages I and II. Compared to the thickness of the oxide scaleformed at the non-contact zone (approximate 1 mm) in the period(Fig. 9), the thickness of the oxide scale remained almostunchanged. However, the TEM cross-section observation(Fig. 5c) shows that the ‘‘glaze’’ oxide scale in the contact zonehas been heavily compressed and deformed, resulting a verydense microstructure. Therefore, the phenomenon of oxidestransferring from the disc to the pin surface is supposed to occurin this stage. The ‘‘glaze’’ oxide scale consists of iron oxides (Fe2O3

and Fe3O4), (Fe, Cr)-rich oxides and some vanadium oxides (V2O5).It has a good adhesion to the HSS matrix, as seen in Figs. 4b and5b, and the interface between the oxide scale and HSS matrix wasvery compact and free from pores and cracks. The formation ofthe ‘‘glaze’’ oxide scale protects the pin from surface wear, whichcan be reflected from the coefficient of friction whose value in therunning-in period was lower than the rest of the pin-on-disc test.From the beginning of Stage I, the pin began to slide from static todynamic, with an oxide to oxide contact configuration. The oxideon an HSS pin was deformed and a continuous and complete‘‘glaze’’ oxide scale (Fig. 4a) covered the whole contact area to actas a solid lubricant and leading to the minimum friction. How-ever, this protection did not last long, at the Stage II running inperiod the oxide scale was spalled off due to shear stressgenerated by friction, oxide scale thickening, thermal impact,and stress growth inside the oxide scale (seen in Fig. 5c). Thecontact surface became rougher due to the oxide spallation, hencethe coefficient of friction quickly increased from its minimumvalue. However, there was no debris or cracks on the contactsurface of the pin at this stage.

As the test proceeded, friction between the pin and the discstabilised (Stage III), while the wear mechanisms became com-plicated. Local spallation of the oxide scale kept occurring duringthe rest of the test and oxide debris generated on the wear track.On one hand, spallation of the oxide scale may cause superficialdamage and increase the rate of oxidation on the HSS pin. On theother hand, oxide debris on the wear track may act as free thirdbodies abrading the antagonistic surfaces [32]. Vanadium MCcarbides have been rubbed off the surface due to the relativeweak binding energy between them (Fig. 7a). The adhesive non-spallation oxide scale in the contact region was around 9.5 mmthick after 1 h pin-on-disc test, whereas the oxide scale at thenon-contact zone was only 2.8 mm thick. A much thicker non-spallation oxide in the contact region is not possible only fromoxidation by the HSS pin itself. It can be deduced that oxidestransferring from the mild carbon steel disc to the pin

Fig. 10. (a) Secondary electron (SE) image of pin surface morphology at the non-contact area after 1 h pin-on-disc test in Stage III, (b) TEM bright field image of the cross

section of the non-contact area, (c) TEM/X-ray mapping of the cross section of the non-contact area.

Fig. 11. Secondary electron (SE) image of the cross-section of oxide scale formed

on the disc surface after 1 h pin-on-disc test (outside the contact zone).

H. Zhu et al. / Wear 302 (2013) 1310–1318 1317

significantly contributes to the scale thickness formed on the HSSpin surface. Because the external oxide scale on the surface of theHSS pin consisted mainly of iron oxides (Fe2O3 and Fe3O4), asshown in Fig. 8, which was the same phase composition of oxide

on the disc surface, it provided further evidence that oxidetransferred from the disc to the pin. The dense, fine-grained andequiaxed oxides (mainly Fe2O3) found in the contact zone is theresult of the combination of compression, sliding, oxidation, oxidetransfer from disc, agglomeration and sintering of wear particlesduring the high temperature pin-on-disc testing. Although theunbroken adhesive oxide seemed compact and smooth from a topview, there were large cracks inside the oxide scale. Bondingbetween the oxide scale and the HSS matrix also weakenedbecause cracks and pores appeared at the interface. It could bepredicted that catastrophic spalling might occur under certaincircumstances, which may be the possible mechanism for whythe ‘‘banding’’ phenomenon [20,33] occurred during hot rolling.

5. Conclusions

High temperature pin-on-disc tests have been successfullycarried out to simulate the tribological behaviour of oxide scalein the role bite during hot rolling. The evolution of the frictioncoefficient during the pin-on-disc test can be divided into threestages. Stage I and II can be summarised as the running-in periodwhich lasts less than 300 s from the test commencing, and Stage

H. Zhu et al. / Wear 302 (2013) 1310–13181318

III is the stabilisation period of friction coefficient. The interruptedtests indicated that adhesive wear is the dominated mechanismsof the pin in running-in period. A thin, continuous, compact andsmooth ‘‘glaze’’ oxide scale formed on the pin at the Stage I, actedlike a solid lubricant and lead to a sharp decrease in friction. AtStage II of the running-period, oxide scale spallation occurred atthe contact zone due to shear stress generated by friction, oxidescale growing, thermal impact, stress growth inside the oxidescale, so the contact surface became rougher and the coefficient offriction quickly increased from its minimum value. At Stage III thewear mechanism on the pin becomes complicated. In addition tooxidation on the HSS pin, the oxides transfers from the disc to thepin which significantly thicken the oxide scale on the pin. Largecracks and pores can be found inside the oxide scale, indicatingthat the severe ‘‘banding’’ phenomena could happen when theoxide scale reaches a critical value. A large amount of wear debrisobserved on the pin wear track confirms that abrasive wearhappens at this stage. The wear mechanism at this stage is amixture of adhesive, abrasion and oxidation.

Acknowledgement

The authors acknowledge the financial support from TheAustralian Research Council (ARC), Australia.

References

[1] C. Vergne, C. Boher, C. Levaillant, R. Gras, Analysis of the friction and wearbehaviour of hot work tool scale: application to the hot rolling process, Wear250 (2001) 322–333.

[2] M. Schutze, Mechanical properties of oxide scales, Oxidation of Metals 44(1995) 29–61.

[3] J.C.G. Milan, M.A. Carvalho, R.R. Xavier, S.D. Franco, J.D.B. DeMello, Effect oftemperature, normal load and pre-oxidation on the sliding wear of multi-component ferrous alloys, Wear 259 (2005) 412–423.

[4] C. Vergne, C. Boher, R. Gras, C. Levaillant, Influence of oxides on friction in hotrolling: experimental investigations and tribological modelling, Wear 260(2006) 957–975.

[5] F.H. Stott, J. Glascott, G.C. Wood, Factors affecting the progressive develop-ment of wear-protective oxides on iron-base alloys during sliding at elevatedtemperatures, Wear 97 (1984) 93–106.

[6] J. Jiang, F.H. Stott, A mathematical model for sliding wear of metals atelevated temperatures, Wear 181–183 (1995) 20–31. (4, 166);J. Jiang, F.H. Stott, A mathematical model for sliding wear of metals atelevated temperatures, Wear 181–183 (1995) 20–31.

[7] M.B. Peterson, S.J. Calabrese, S. Li, X. Jiang, Friction of alloys at hightemperature, Journal of Materials Science & Technology 10 (1994) 313–320.

[8] Y. Sano, T. Hattori, M. Haga, Characteristics of high-carbon high speed steelrolls for hot strip mill, ISIJ International 32 (1992) 1194–1201.

[9] Y.J. Kang, J.C. Oh, H.C. Lee, S. Lee, Effects of carbon and chromium additions onthe wear resistance and surface roughness of cast high-speed steel rolls,Metallurgical and Materials Transactions A 32A (2001) 2515–2525.

[10] J.W. Park, H.C. Lee, S. Lee, Composition, microstructure, hardness, and wearproperties of high-speed steel rolls, Metallurgical and Materials TransactionsA 30A (1999) 399–409.

[11] O. Kato, H. Yamamoto, M. Ataka, K. Nakajima, Mechanisms of surfacedeterioration of roll for hot strip rolling, ISIJ International 32 (1992)1216–1220.

[12] M. Pellizzari, A. Molinari, G. Straffelini, Tribological behaviour of hot rollingrolls, Wear 259 (2005) 1281–1289.

[13] J.H. Lee, J.C. Oh, J.W. Park, H.C. Lee, S. Lee, Effects or tempering temperatureon wear resistance and surface roughness of a high speed steel roll, ISIJInternational 41 (2001) 859–865.

[14] G. Savage, R. Boelen, A. Horti, H. Morikawa, Y. Tsujimoto, Hot wear testing ofroll alloys, in: Proceedings of the 37th MWSP Conference, ISS, vol. XXXIII,1996, pp. 333–337.

[15] S. Lundberg, T. Gustafsson, The influence of rolling temperature on roll wear,investigated in a new high temperature test rig, Journal of MaterialsProcessing Technology 42 (1994) 239–291.

[16] D.N. Hanlon, W.M. Rainforth, C.M. Sellars, The rolling-sliding wear responseof conventionally processed and spray formed high chromium content castiron at ambient and elevated temperature, Wear 225–229 (1999) 587–599.

[17] D.N. Hanlon, W.M. Rainforth, The rolling sliding wear response of conven-tionally processed and spray formed high speed steel at ambient andelevated temperature, Wear 255 (2003) 956–966.

[18] M. Pellizzari, D. Cescato, M.G. De Flora, Hot friction and wear behaviour ofhigh speed steel and high chromium iron for rolls, Wear 267 (2009) 467–475.

[19] F.H. Stott, M.P. Jordan, The effects of load and substrate hardness on thedevelopment and maintenance of wear protective layer during sliding atelevated temperatures, Wear 250 (2001) 391–400.

[20] K. Hokkirigawa, T. Kato, T. Fukuda, M. Shinooka, Analysis of sliding wearmechanism of oxide films on hot roll surfaces base on in-situ observation ofwear process by the CCD microscope tribosystem, Journal of Japanese Societyof Tribologists 42 (1997) 777–784.

[21] O. Joos, C. Boher, C. Vergne, C. Gaspard, T. Nylen, F. Rezai-Aria, Assessment ofoxide scales influence on wear damage mechanisms of HSM work rolls, Wear263 (2007) 198–206.

[22] Q. Zhu, H.T. Zhu, A.K. Tieu, M. Reid, L.C. Zhang, In-situ investigation ofoxidation behaviour in high-speed steel roll material under dry and humidatmospheres, Corrosion Science 52 (2010) 2707–2715.

[23] Q. Zhu, H.T. Zhu, A.K. Tieu, C. Kong, Three dimensional microstructure studyof oxide scale formed on a high-speed steel by means of SEM, FIB and TEM,Corrosion Science 53 (2011) 3603–3611.

[24] A.K. Tieu, Q. Zhu, H.T. Zhu, C. Lu, An investigation into the tribologicalbehaviour of a work roll material at high temperature, Wear 273 (2011)43–48.

[25] H.H. Kim, J.W. Lim, J.J. Lee, Oxidation behaviour of high-speed steels in dryand wet atmospheres, ISIJ International 43 (2003) 1983–1988.

[26] S. Spuzic, K.N. Strafford, C. Subramanian, G. Savage, Wear of hot rolling millrolls: an overview, Wear 176 (1994) 261–271.

[27] Y. Yu, J.G. Lenard, Estimating the resistance to deformation of the layer ofscale during hot rolling of carbon steel strips, Journal of Materials ProcessingTechnology 121 (2002) 60–68.

[28] V. Vitry, S. Nardone, J.-P. Breyer, M. Sinnaeve, F. Delaunois, Microstructure oftwo centrifugal cast high speed steels for hot strip mills applications,Materials & Design 34 (2011) 372–378.

[29] M.P. Guerrero, C.R. Flores, A. Perez, R. Colas, Modelling heat transfer in hotrolling work rolls, Journal of Materials Processing Technology 94 (1999)52–59.

[30] A.A. Tseng, F.H. Lin, A.S. Gunderia, D.S. Ni, Roll cooling and its relationship toroll life, Metallurgical Transactions A 20A (1988) 2305–2320.

[31] A. Tudball, S.G.R. Brown, Practical finite element heat transfer modelling forhot rolling of steels, Ironmaking and Steelmaking 33 (2006) 61–66.

[32] N. Chelliah, S.V. Kailas, Synergy between tribo-oxidation and strain rateresponse on governing the dry sliding wear behaviour of titanium, Wear 266(2009) 704–712.

[33] L.C. Erickson, S. Hogmark, Analysis of banded hot rolling rolls, Wear 165(1993) 231–235.