Investigation of wear mechanisms through in situ observation during microscratching inside the scanning electron microscope

Wear 259 (2005) 18–26

Investigation of wear mechanisms through in situ observation during microscratching inside the scanning electron microscope J. Michler a,∗ , R. Rabe a,b , J.-L. Bucaille a , B. Moser a , P. Schwaller a , J.-M. Breguet b a

Swiss Federal Institute for Materials Testing and Research (EMPA), Feuerwerkerstr. 39, CH-3602 Thun, Switzerland b Swiss Federal Institute of Technology Lausanne (EPFL), CH-1015 Lausanne, Switzerland Received 31 July 2004; received in revised form 28 January 2005; accepted 14 February 2005 Available online 10 May 2005

Abstract Scratch experiments have been used in the past to investigate various contact phenomena such as abrasion mechanisms or asperity contacts. Depending on tip geometries, loading conditions and materials investigated, different crack propagation modes (radial or lateral cracks, etc.) or deformation modes (ploughing, chipping, etc.) may dominate the scratching process. The residual scratch path can yield some information about dominant deformation and fracture modes. It is, however, often not possible to uniquely correlate cracks and other phenomena with events on the recorded load–displacement curves. We have built a miniaturized microscratch device for use inside a scanning electron microscope (SEM) that allows the observation of the surface around the tip with sub-micrometer resolution during scratching. Using a conical indenter with spherical tip we demonstrate on different materials that the device is a powerful tool to observe initiation and propagation of cracks, to observe the flow of the material near the indenter (piling-up and sinking-in) and to study chip and particle formation mechanisms during microscratching. In GaAs, particles were observed to form in front and on the rear side of the tip via interaction of chevron cracks. In the case of a Fe-based bulk metallic glass, shear bands were observed to form in front of the tip leading to serrated chip formation. Discontinuities in the tip penetration during scratching of a polymer thin film were related to the onset of crack formation behind the tip and to the propagation of semi-circular cracks in front of the tip. The observed large elastic recovery of the polymer film at the rear side of the tip has to be taken into account for accurate contact area calculations. © 2005 Elsevier B.V. All rights reserved. Keywords: Microscratching; In situ observation; Bulk metallic glass; Polymer thin films; GaAs

1. Introduction Scratch resistance of materials is of importance in many technological applications such as automotive clear coats or glasses. In a wider sense, scratch testing serves as (a) model experiment to understand deformation and crack formation mechanisms in precision machining and (b) is fundamental to the understanding of deformation and removal mechanisms in grinding, polishing and abrasive wear, because all these processes are functions of the cumulative actions of repeated microscopic contacts. Scratching is also used as a test proce∗

Corresponding author. Tel.: +41 33 2284605; fax: +41 33 2884490. E-mail address: [email protected] (J. Michler).

0043-1648/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.wear.2005.02.111

dure to determine qualitatively the mechanical properties of coatings, in particular the coating-substrate adhesion [1]. In ductile bulk materials, the behaviour of the material during a scratch test depends mainly on the shape of the indenter that controls the imposed deformation level and on the mechanical properties of the material. Bucaille et al. showed that during scratching the average deformation is related to the ratio of the penetration h of the indenter to the contact radius a [2]: εr ∝

h . a

(1)

For a conical indenter having a spherical tip, the deformation level h/a increases as the penetration of the indenter

J. Michler et al. / Wear 259 (2005) 18–26

increases. For small deformations, the behaviour of the material is mainly elastic and no permanent groove can be observed after scratching. If the deformation imposed by the indenter increases, ductile ploughing is observed [3,4]. The transition between elastic and plastic domains also depends on the rheology of the scratched material. For an elasticperfectly plastic behaviour, the flow of the material near the indenter is strongly dependent on the ratio between Young’s modulus, E, and yield stress, σ y [5]. For high values of E/σ y pile-ups are created near the indenter, while for low values of E/σ y the material sinks-in in front of the indenter and elastic recovery at the rear of the indenter is large [2]. This means that for polymers, i.e. materials having a large elastic part (E/σ y ∼ 10–50), the deformation mode can vary from the elastic behaviour with almost no residual groove to ductile ploughing. For metals (E/σ y > 200), even for a small deformation level, ductile ploughing is the main observed mode, followed by chip formation as the deformation imposed by the indenter increases. In brittle solids, a large variety of cracks and deformation phenomena can be observed [6,7]. Soda-lime glass is an extensively studied brittle model material: Swain has carried out a detailed study using Vickers tips already in 1979 [8]. His observations of the morphology of the scratch path and of the extent of cracks below the surface after scribing can be summarised as follows: at loads below 0.05 N, there is no visible surface or sub-surface cracking around or within the residual groove. Well-developed median and lateral cracks are visible at intermediate loads (0.1–5 N), with lateral cracks not intersecting the surface at loads up to 1 N. At high loads, median cracks and poorly developed lateral cracks are observed and the region around the scratch track is seen to have a “crushed” appearance. Similar observations have also been reported by Ahn [9]. Bulsara investigated in situ the contact region between a Vickers tip and soda-lime glass using an optical microscope looking from the back side [10]. He observed that depending on the applied load median and lateral cracks are generated either during unloading behind the indenter or along with the sliding tip. Radial/chevron cracks were found to arise due to median-type cracks forming ahead of the sliding indenter and deviating to either side of the scratch track. Due to the geometrical self-similarity, the deformation imposed by the indenter and defined by Eq. (1) is constant for Vickers or conical indenters as the penetration of the indenter increases. The transition between these deformation modes can therefore not be related to a variation of the deformation imposed by the indenter. However, the volume solicited by the scratching process increases as the penetration depth increases which allows the activation of more defects which initiates brittle fracture. In this work, the deformation and fracture phenomena occurring during microscratching with a rounded conical indenter are studied using an in situ observation technique inside a scanning electron microscope (SEM). The electron microscope allows imaging a major part of the contact region with sub-micron resolution during microscratching. There have

19

been already excellent studies on deformation and fracture modes using scratch techniques inside an SEM [11,12]. For instance Hedenqvist and Hogmark revealed fracture modes of TiN coatings on steel during scratching [11,13]. Our experimental set-up differs significantly from approaches in the past, as we use piezo-based actuators that allow scratches with nanometer displacement resolution. In order to explore the potential of the technique, scratch tests were performed on a semiconductor, on a metallic glass and on a polymer thin film sample. Furthermore, the results of the in situ SEM measurements are compared to scratch results from a commercial nanoscratch instrument.

2. Experimental 2.1. Samples Materials belonging to several different categories in terms of mechanical properties were investigated to explore the potential of the SEM-based microscratch device: a Fe-based amorphous metal (also known as bulk metallic glasses) sample, single crystalline GaAs and a polymer thin film. The Fe-based bulk metallic glass sample has a nominal composition of Fe61 Zr8 Y2 Co5 Cr2 Mo7 B15 and was produced by arc melting and drop cast into a copper mold. The specimen was mechanically polished to a mirror finish. More details on the processing and full characterization of the microstructure can be found in [14]. The GaAs wafers were of (0 0 1)-type and not doped. The polymer thin film was a 5 ␮m thick polysiloxane coating containing 20 vol.% colloidal SiO2 particles on a thermosetting-based polycarbonate substrate. Such types of coatings are applied for eyeglass lenses. The material behaviour during a scratch is therefore of direct interest in the application of this type of coatings. For more details see Bucaille et al. [2]. 2.2. Experimental set-up 2.2.1. Conventional microscratching A Nanoindenter-XP (MTS/Nanoinstruments, Oak Ridge, TN) equipped with a lateral force measurement option for scratch experiments was used to determine the indentation hardness and the Young’s modulus of the investigated samples and to perform scratch experiments with a linearly increasing load from 0 to 300 mN. A conical diamond indenter (45◦ half-angle) with a spherical tip having a tip radius of around 1 ␮m was used for the scratches. The scratch length was 500 ␮m and the scratch velocity was 10 ␮m/s. Lateral force and penetration depth were continuously recorded as a function of the applied normal load. For each sample several scratches have been made to guarantee the unambiguousness of the obtained results. For each scratch the specimen’s surface profile has been measured prior and after the scratch using the same conical indentation tip with a constant load of 50 ␮N. The GaAs(0 0 1) surface was scratched along the

20

J. Michler et al. / Wear 259 (2005) 18–26

tip was observed from the side with respect to the moving direction. A low scratch speed (around 0.2 ␮m/s, compared to 10 ␮m/s in the Nanoindeter-XP) was used to allow the acquisition of high quality SEM video sequences. Typical frame rates were one frame per second. This frame rate appeared to be a good compromise between time resolution and image quality at higher magnifications (>5000×). Y-scratches were realized starting at maximum load and keeping it constant during the whole scratch length (20 ␮m).

3. Results and discussion Fig. 1. Photograph of the SEM nanoscratch device: (1) load cell, (2) XYpositioning table, (3) sample holder, (4) tip holder and adapter, (5) piezo actuator, displacement sensor and (6) dovetail mounting.

1 1 0 direction. Postmortem SEM micrographs were taken from the scratched areas. Indentation load–displacement curves for the determination of the indentation hardness and the Young’s modulus for GaAs and the iron-based bulk metallic glass have been measured using a three-sided Berkovich-type diamond pyramid at a maximum applied load of 100 mN [15]. 2.2.2. The SEM microscratch device A detailed description of the construction and the complete technical specifications can be found in [16]. In the following, we only give a short description of the device. The instrument is shown in Fig. 1. A 20 ␮m range closed-loop stack piezo (5) with a built in displacement sensor applies the normal load by pushing a flexure parallelogram where the tip holder (4) is fixed. These components are assembled on a fine positioning stage (6) (also a flexure mechanism) having 1 mm range. It is driven by a fine pitch screw, connected by a cable to a knob installed on the SEM chamber’s door. The whole assembly is fixed on the main body by a dovetail allowing testing samples with a wide range of heights using different load cells. A XYstage (1) holds the load cell (2) and the sample (3). Only the normal force and the penetration depth are measured in the current set-up. A stick-slip actuator drives the X-axis and can be used for long scratches up to 10 mm. Due to the limited blocking force of the stick-slip actuator the maximal normal forces are – depending on the materials tested – on the order of 300 mN. The Y-axis can be positioned only when the SEM chamber is opened but contains a flexure-stack piezo construction (not visible on the picture) used for scratching the sample in the Y-direction over a range of 20 ␮m with higher forces. The identical conical diamond tip that was used in the MTS Nanoindenter-XP is also used for the in situ scratch experiments. In the present experimental set-up, the instrument has been installed inside a Zeiss DSM962 SEM (Carl Zeiss SMT AG, Oberkochen, Germany). Scratches were performed with a linearly increasing load from 0 up to 200 mN (GaAs, Fe-based bulk metallic glass) or up to 300 mN (polysiloxane). The typical scratch length (using the X-axis) was around 300 ␮m. In this experiment, the

3.1. GaAs(0 0 1) 3.1.1. Conventional microscratching Fig. 2 shows the penetration depth (thin line) and the lateral force (thick line) as a function of the applied normal load. Until about 35 mN of normal load the penetration curve is very smooth, without irregular variations of the penetration depth. From approximately 70 mN several pop-in and popout events are noticed with amplitudes larger than 100 nm. In the transition regime between 35 and 70 mN, the variations of the depth during loading are on the order of 20 nm. The lateral force increases with increasing load and shows a similar evolution of irregularities as a function of the applied load. Fig. 3 presents SEM micrographs of these three regions. A remaining depression indicating plastic deformation has been found by SEM for loads larger than approximately 5 mN. At the beginning of the intermediate load regime, significant ridge formation at the contact edges along the sides of the scratch becomes visible (see Fig. 3a). Also some small transverse cracks within the scratch path and small debris just outside the scratch can be discerned. Above 50 mN, but still in the intermediate load regime, chevron cracks appear, propagating from the side of the scratch path in a curved shape ending roughly perpendicular to the scratch path. In the transition to the high load regime chip-outs start to form

Fig. 2. Penetration depth and friction force during microscratching of GaAs as a function of applied normal load for 1 1 0 scratch direction.

J. Michler et al. / Wear 259 (2005) 18–26

21

3.1.2. In situ SEM observation of microscratching In situ SEM observations during scratching showed that in the transition regime the pronounced plastic deformation might eventually cause the fracture strain of the GaAs to be exceeded. This was observed to result in cohesive GaAs fracture along the sides and in front of the tip leading to the removal of minute debris similar to observations during in situ SEM scratching of TiN thin films [13]. At higher loads, chevron-type cracks, in part together with lateral-type cracks, were identified to be responsible for chip-outs. Chip-out formation was observed all around the moving tip, i.e. in front, on the sides and sometimes behind the tip, where unloading occurs. Fig. 4 shows an image sequence that illustrates the chip formation in front of the tip. In the figure, the tip is moving towards the observer. In Fig. 4a, four cracks are visible, which was a typical configuration for loads around 100 mN. At loads around 50 mN, not displayed in Fig. 4, the two inner types of cracks were often observed to deviate outside

Fig. 3. Postmortem SEM micrographs from different regimes identified from Fig. 2. (a) Ridge formation at the contact edge at 38 mN. (b) Chevron cracks and debris formation at 50 mN. (c) Chevron cracks and onset of chipout and particle generation at 70 mN. (d) Chip-out and particle generation at 100 mN. The arrow indicates the direction of the scratch, i.e. the direction of increasing load.

originating from chevron cracks, see Fig. 3b. In addition, some micrometer-sized particles are found around the scratch path and at higher loads several chip-outs along the scratch path are visible (see e.g., Fig. 3c for 100 mN). Median cracks may be present, but they were not observed to reach the surface.

Fig. 4. SEM microscratching video sequence of GaAs at 100 mN load. The tip moves towards the observer. The chip formation takes place in front of the tip. Please note the low signal to noise ratio inherent to images extracted from SEM video sequences.

22

J. Michler et al. / Wear 259 (2005) 18–26

Fig. 6. Penetration depth and friction force during microscratching of Febased bulk metallic glass as a function of applied load. Fig. 5. SEM microscratching video sequence of GaAs at 50 mN. The tip moves away from the observer. Occasional chip formation in GaAs behind the tip involving a lateral crack.

the scratch path and to become two chevron cracks like the two outer cracks visible in Fig. 4a. At 100 mN, two chevron cracks in front of the tip interact and form a chip via brittle fracture: in Fig. 4b, a chip out in front of the moving tip is generated and in Fig. 4c and d, a few moments later on the side as well as again in front of the scratch. Fig. 5 shows a situation occasionally observed where the chip is generated in the wake of the scratching tip (tip is moving away from the observer). Two chevron cracks cross-link via a lateral crack that forms at the back side of the moving tip, i.e. during the unloading of the surface. This is very similar to the formation of chips in other brittle materials like soda-lime-silica glass [17]. Generally, on the micrometer scale the cracks were not observed to follow strictly cleavage planes. The in situ observations show that cracks extend at speeds similar to that of the indenter, i.e. stable crack propagation takes place related to the scratching speed. 3.2. Fe-based bulk metallic glass 3.2.1. Conventional microscratching Fig. 6 shows the penetration depth (thin line) and the lateral force (thick line) as a function of the applied normal load. Up to a normal load of about 120 mN the curve is again very smooth, without irregular variations in the penetration depth. Then a sudden increase in penetration depth as well as in lateral force is observed. At higher normal loads irregular depth variations with amplitudes above 100 nm appear. The lateral force increases with increasing load, doubles almost as the penetration depth suddenly increases (at around 120 mN) and then shows similar variations as the penetration depth. As indicated by Eq. (1), the deformation imposed by a conical indenter with a spherical tip increases as the penetration depth h increases. In the case of a half-angle of 45◦ and a tip radius

of 1 ␮m, Eq. (1) becomes (h in nm): εr ∝

h 1 = a 1 + 410/ h

(2)

if the conical part of the tip is in contact with the surface. For large values of h, the deformation level h/a = 1, which corresponds to the deformation imposed by the conical indenter neglecting tip rounding. Although in our case, the material is in contact with the conical part of the indenter, i.e. for h > 290 nm, the deformation level imposed by the indenter is significantly less than 1 and continues to increase (h/a is 0.55 at 500 nm and 0.71 at 1 ␮m). The pop-in may therefore be related to a critical deformation level. Wang et al. studied Febased bulk metallic glass by nanoscratching with Berkovich indenters [18]. They also observed sudden pop-ins of the indenter under load and a corresponding increase in the friction force, which they associated to microcrack formation. Fig. 7 shows SEM micrographs of the scratch trace at 75, 105 and 150 mN. At 75 mN, the scratch appears uniform with no visible features at the borders and devoid of any debris. In contrast, at 105 mN shear bands at the border of the scratch can be resolved by SEM. Shear bands have also been observed in the case of instrumented indentation and compression tests [19,20] and are the result of a plastic instability that concentrates all plastic deformation inside those shear bands [21]. Shear bands have also been observed at the sides of wear tracks after tribological testing [22]. After the transition load of about 140 mN the edges of the scratch exhibit a smeared appearance and the shear bands are less pronounced at the borders, see example at 150 mN in Fig. 7c. Cracks were not detected by ex situ SEM in the tested load range. 3.2.2. In situ SEM observation of microscratching At low loads up to 16 mN, no chip is observed during in situ observation and simply a residual impression is formed, see Fig. 8a. The onset of chip formation is, however, difficult to determine as at low loads the contact edge is hidden by the tip and cannot directly be observed in the microscope.

J. Michler et al. / Wear 259 (2005) 18–26

23

indenter geometry [23,24]. In situ observations revealed the formation of small shear bands in front of the moving tip. The shear bands generate surface steps with a lateral extension smaller than the contact radius, which gives the formed chip a more debris-like appearance. At 105 mN, the situation

Fig. 7. Postmortem SEM micrographs from different loads: (a) at 75 mN uniform appearance with no visible features at the borders and without any debris, (b) at 105 mN shear bands at the border and (c) smearing out of the borders of the scratch path at 150 mN.

Ductile ploughing behaviour occurs at relatively small deformation levels, i.e. for loads smaller than around 16 mN, corresponding to a deformation level smaller than 0.27 (Eq. (2)). At loads above 16 mN, chip formation was observed to start. The chip formation seen in Fig. 8b at a load of 75 mN is directly related to the increase of the deformation level to 0.49 at a penetration of roughly 400 nm. This transition has been investigated during scratching of metals by varying the

Fig. 8. SEM micrographs extracted from a microscratching video sequence of bulk metallic glass: (a) no chip formation at 16 mN, (b) chip formation at 75 mN, (c) ongoing chip formation at 105 mN, (d) smearing-out of scratch borders at 150 mN and (e) discontinuous chip formation at 200 mN via popouts of single shear bands.

24

J. Michler et al. / Wear 259 (2005) 18–26

looks similar to 75 mN. Contrary to Fig. 8, the appearance of shear bands along the side of the residual scratch could not be resolved during the in situ experiments. The thickness of the chip increases significantly beyond the transition load (compare 150 mN) and shows shear bands. This may indicate a transition from wave removal to chip removal [25]. A transition from continuous to discontinuous plastic flow is also observed in metals with increasing deformation level [26]. At higher loads (200 mN) larger but fewer shear bands were observed in this study compared to lower loads leading to a more serrated chip. Once a critical stress was achieved one single shear band over the entire width of the contact was observed to pop-out in front of the tip and to form the corresponding part of the chip (see arrow in Fig. 8e). The chips show sharp edges and are similar in appearance to chips formed during high-speed single diamond turning [27]. The discontinuous chip formation process may be responsible for the discontinuities of the penetration under load in Fig. 6. At lower load, many small shear bands form in parallel in front of the tip and the overall curve appears therefore smooth. Cracks like the ones observed in GaAs were not observed in the investigated load range. In general, one could say that although bulk metallic glasses are known to have limited ductility in tension, the chip formation and the residual scratch much more resemble the one of a ductile metal than the one of a brittle material. This is in agreement with the findings of several other researchers that studied the wear behaviour of bulk metallic glass [22]. 3.3. Polysiloxane coating on a thermosetting-based polycarbonate substrate

Table 1 Indentation hardness (H) and Young’s modulus (E) values for Fe-based bulk metallic glass and GaAs(0 0 1) Sample

Poisson number (ν)

H (GPa)

E (GPa)

Fe-based BMG GaAs(0 0 1) Polysiloxane film Thermosetting PC

0.36 0.31 0.4 0.4

13.91 (0.32) 7.04 (0.09) 0.24 0.18

214 (6) 108 (3) 3.19 2.1

The values in brackets indicate the standard deviation from several tests. The Poisson coefficients used within the Oliver-Pharr model are indicated for information. The values for the polysiloxane film and its substrate are from Bucaille et al. [2].

very smooth. At this load some small serrations in the penetration depth as well as in the lateral force appear. For normal loads above about 200 mN discontinuities in the penetration curve become larger. Again, the behaviour of the lateral force as a function of the applied normal load shows a similar behaviour. A penetration depth of 16 ␮m was achieved at a maximum normal load of 300 mN. This value is 8 and 16 times higher than for the GaAs and the bulk metallic glass samples, respectively. This large difference is not surprising considering the difference in the hardness values and the Young’s moduli for these materials (Table 1). The penetration depth of the indenter into the surface of the sample is more than three times the film thickness but no delamination has been observed. 3.3.2. In situ SEM observation of microscratching Fig. 10 shows images from two intermediate loads, i.e. 120 and 240 mN. As previously observed by Gauthier and coworkers [28,29] and Bucaille et al. [2] such material with

3.3.1. Conventional microscratching Fig. 9 shows the penetration depth (thin line) and the lateral force (thick line) depending on the applied normal load. Until about 110 mN normal load the penetration curve is again

Fig. 9. Penetration depth and friction force during microscratching of a polysiloxane coating on a thermosetting-based polycarbonate substrate.

Fig. 10. SEM micrographs extracted from a microscratching video sequence of a polysiloxane coating on a thermosetting-based polycarbonate substrate: (a) crack formation behind the moving tip at 120 mN and (b) crack formation in front of the moving tip at 240 mN.

J. Michler et al. / Wear 259 (2005) 18–26

large elastic deformation capabilities tends to sink-in around an indenter (Fig. 10). A very large elastic recovery is observed in the wake of the moving indenter from small to large normal loads. A post-profiling of the scratch track using the same tip used for the scratch with an applied load of 50 ␮N revealed a residual depth of 1 ␮m at maximum, i.e. only about 6% of the maximum penetration depth of the indenter. Classically, the elastic recovery is decomposed into two parts: an instantaneous recovery and a time-dependent elastic recovery, the ratio between these two parts depending on the material itself and on the scratch speed of the indenter [29]. Fig. 10 shows that the instantaneous elastic recovery is large enough for the material to be in contact with the rear face of the indenter, which is of great importance to determine the contact pressure between the material and the indenter under such conditions. Fig. 10 allows to link the features observed on the lateral force and penetration curve at 110 mN to damages formed during scratching. One can see that at the rear side of the tip coating failure appears due to the large elastic recovery of the deformed substrate (Fig. 10a). For larger normal loads, cracks begin to propagate on the sides and to the front of the tip, which causes larger variations in the penetration depth of the indenter (Figs. 9 and 10b). This finding fairly agrees with the observation of smaller and larger discontinuities in the penetration curve of Fig. 9. While the cracks forming in the wake of the moving tip are probably caused by a tearing action by the tip pulling on the film by friction forces, the progression of cracks via the side to the front of the tip is probably caused by tensile stresses resulting from the bending stresses within the film due to the larger indentation depths [30]. From ex situ observations it is not possible to tell when and where (with respect to the moving tip) the cracks appear. This information is important in order to understand the formation of these cracks. Depending on where these cracks appear different stresses are responsible for their formation, which in turn allows conclusions on the failure behaviour of the thin film.

4. Conclusions We present an instrumented microindentation/scratch system that can be used inside a commercial SEM chamber. Microscratches on three different types of materials – GaAs, bulk metallic glass and polymer thin film – demonstrate that the SEM-based device is a powerful tool to observe with submicron resolution the initiation and propagation of cracks and the flow of the material near the indenter (piling-up and sinking-in) and can yield information not available from conventional microscratch experiments. In particular, crack formation in GaAs(0 0 1) was shown to take place both on the side and in front of the tip. Particles in GaAs were observed to form in front and on the rear side of the tip via interaction of chevron cracks. Discontinuous shear band formation in bulk

25

metallic glass in front of the moving tip was shown to lead to serrated chip formation. Crack formation in a polymer thin film was first observed behind and to the sides of the moving tip. At higher loads cracks were seen to emanate in front of the tip leading to the classical ring crack pattern often observed after scratching. Furthermore a large elastic recovery at the rear side of the tip was observed which has to be taken into account for contact pressure calculations.

Acknowledgements The authors would like to thank S. Fahlbusch and G. B¨urki for SEM assistance and V. Le Houerou, University of Rennes, France, for helpful discussions. B.M. would like to thank C.T. Liu for providing the bulk metallic glass sample. This work has been partly financially supported by the Swiss Office on Education and Science (OFES) in the framework of the European project: ROBOSEM–Growth: G1RD-CT-2002-00675 and by the Swiss Federal Office for Professional Education and Technology under Contract number CTI 6025.2.

References [1] R. Jacobs, J. Meneve, G. Dyson, et al., A certified reference material for the scratch test, Surf. Coat. Technol. 174 (2003) 1008–1013. [2] J.L. Bucaille, E. Felder, G. Hochstetter, Experimental and threedimensional finite element study of scratch test of polymers at large deformations, J. Tribol. ASME 126 (2004) 372–379. [3] B.J. Briscoe, E. Pelillo, S.K. Sinha, Scratch hardness and deformation maps for polycarbonate and polyethylene, Polym. Eng. Sci. 36 (1996) 2996–3005. [4] B.J. Briscoe, P.D. Evans, E. Pelillo, S.K. Sinha, Scratching maps for polymers, Wear 200 (1996) 137–147. [5] J.L. Bucaille, E. Felder, G. Hochstetter, Mechanical analysis of the scratch test on elastic and perfectly plastic materials with the threedimensional finite element modeling, Wear 249 (2001) 422–432. [6] B.R. Lawn, Fracture and deformation in brittle solids: a perspective on the issue of scale, J. Mater. Res. 19 (2004) 22–29. [7] A. Broese van Groenou, N. Maan, J.D.B. Veldkamp, Scratching experiments on various ceramic materials, Philips Res. Rep. 30 (1975) 320–359. [8] M.V. Swain, Proc. R. Soc. Lond. A 366 (1979) 575–597. [9] Y. Ahn, Deformation about sliding indentation in ceramics and its application to lapping, PhD Thesis, School of Industrial Engineering, PurDue University, West Lafayette, IN 47907–1287, August 1992. [10] V.H. Bulsara, S. Chandrasekar, Direct observation of contact damage around scratches in brittle solids, Proc. SPIE 3060 (1997) 76–88. [11] P. Nledengvist, S. Hogmark, Experiences from scratch testing of tribological PVD coatings, Tribol. Int. 30 (1997) 507–516. [12] M. Yoshino, T. Aoki, T. Shirakashi, R. Komanduri, Some experiments on the scratching of silicon: in situ scratching inside an SEM and scratching under high external hydrostatic pressures, Int. J. Mech. Sci. 43 (2001) 335–347. [13] P. Hedenqvist, M. Olsson, S. Jacobson, S. Soderberg, Failure mode analysis of TiN-coated high speed steel: in situ scratch adhesion testing in the scanning electron microscope, Surf. Coat. Technol. 41 (1990) 31–49. [14] Z.P. Lu, C.T. Liu, W.D. Porter, Role of yttrium in glass formation of Fe-based bulk metallic glasses, Appl. Phys. Lett. 83 (2003) 2581–2583.

26

J. Michler et al. / Wear 259 (2005) 18–26

[15] W.C. Oliver, G.M. Pharr, An improved technique for determining hardness and elastic modulus using load and displacement sensing indentation experiments, J. Mater. Res. 7 (1992) 1564–1583. [16] R. Rabe, J.-M. Breguet, P. Schwaller, S. Stauss, J. Patscheider, J. Michler, Observation of fracture and plastic deformation during indentation and scratching inside the scanning electron microscope, Thin Solid Films 469/470 (2004) 206–213. [17] V. Le Houerou, J.C. Sangleboeuf, S. Deriano, T. Rouxel, G. Duisit, Surface damage of soda-lime-silica glasses: indentation scratch behavior, J. Non-Cryst. Solids 316 (2003) 54–63. [18] J.G. Wang, B.W. Choi, T.G. Nieh, C.T. Liu, Nano-scratch behavior of a bulk Zr–10Al–5Ti–17.9Cu–14.6Ni amorphous alloy, J. Mater. Res. 15 (2000) 913. [19] C.A. Schuh, T.G. Nieh, A survey of instrumented indentation studies on metallic glasses, J. Mater. Res. 19 (2004) 46–57. [20] W.J. Wright, R. Saha, W.D. Nix, Deformation mechanisms of the Zr40 Ti14 Ni10 Cu12 Be24 bulk metallic glass, Mater. Trans. 42 (2001) 642–649. [21] C.A. Schuh, A.S. Argon, T.G. Nieh, J. Wadsworth, The transition from localized to homogeneous plasticity during nanoindentation of an amorphous metal, Philos. Mag. 83 (2003) 2585–2597. [22] A.L. Greer, KL. Rutherford, M. Hutchings, Wear resistance of amorphous alloys and related materials, Int. Mater. Rev. 47 (2002) 87–112.

[23] K. Hokkirigawa, K. Kato, Z. Li, The effect of hardness on the transition of the abrasive wear mechanism of steels, Wear 123 (1988) 241–251. [24] A.J. Black, E.M. Kopalinsky, P.L.B. Oxley, An investigation of the different regimes of deformation which can occur when a hard wedge slides over a soft surface: the influence of wedge angle, lubrication and prior plastic working of the surface, Wear 123 (1988) 97–114. [25] J.A. Williams, Analytical models of scratch hardness, Tribol. Int. 29 (1996) 675–694. [26] T. Sakamoto, T. Tsukizoe, Friction and prow formation in a scratch process of copper by a diamond cone, Wear 44 (1977) 393–403. [27] M. Bakkal, A.J. Shih, R.O. Scattergood, Chip formation, cutting forces, and tool wear in turning of Zr-based bulk metallic glass, Int. J. Mach. Tool Manuf. 44 (2004) 915. [28] C. Gauthier, R. Schirrer, Time and temperature dependence of the scratch properties of poly(methylmethacrylate) surfaces, J. Mater. Sci. 35 (2000) 2121–2130. [29] C. Gauthier, S. Lafaye, R. Schirrer, Elastic recovery of a scratch in a polymeric surface: experiments and analysis, Tribol. Int. 34 (2001) 469–479. [30] J. Michler, E. Blank, Analysis of coating fracture and substrate plasticity induced by spherical indentors: diamond and diamond-like carbon layers on steel substrates, Thin Solid Films 381 (2001) 119– 134.