《复合材料 Composites》课程教学资源（学习资料）第五章 陶瓷基复合材料_Tribological behaviour of Si3N4 and Si3N4

Availableonlineatwww.sciencedirect.com WEAR ELSEVIER wear260(2006)634-641 www.elsevicr.com/locatewear Tribological behaviour of SinA and Si3N4-ooTiN based composites and multi-layer laminates M. Hadad a,*G. Blugan.J Kubler D.E. Rosset. L Rohr a J Michler a erkerstr: 39. CH-3600 Thun Swit b EMPA, Laboratory for High Performance Ceramics, 8600 Duebendorf, Switzerland Received 12 January 2004; received in revised form I March 2005: accepted 16 March 2005 Available online 24 June 2005 Abstract Si3N4-Tin based multi-layer laminates exhibit differences in residual stress between individual layers due to a variation of the thermal expansion coefficient between the layers. The residual stress distribution in these multi-layer laminates is known to improve the apparent macroscopic fracture toughness. In this work, the tribological behaviour of bulk, composites and multi-layers laminates are investigated. Si3N4 bulk, Si,Na based composites with 10, 20 and 30 wt% Tin and different multi-layer laminates have been tested under dry conditions with reciprocal movement using a ball-on-block configuration In particular, the infuence of sliding directions with respect to the layer orientations has been investigated The experimental results show that wear resistance increased with increasing tiN content in Si3N4-TIN composites. However, multi-layer laminates exhibit an up to three times higher apparent fracture toughness, but do not show an improvement of wear resistance compared to C 2005 Elsevier B V. All rights reserved Keywords: Silicon nitride; Composites and laminate ceramics; Ball-on-block wear testing; Anisotropy; Wear mechanism Introduction toughness up to 100% and to enhance fracture strength up to 1000 MPa [ 8]. This has been attributed mainly to resid Silicon nitride Si3 N4 based ceramics have attractive phys- ual stress at hetero-phase boundaries(Si3 N4-TIN) caused by ical properties such as high hardness, chemical inertness and mismatch of thermal expansion coefficient between TIN and low density. One of the most important applications are cut- Si3 N4. Similarly, multi-layer ceramic laminates have been ting tools and machining of hard materials, like cast iron and previously reported to have considerable improvement in nickel based alloys [1, 2]. The main limitations of ceramic macroscopic fracture toughness [11-13]. This was attributed cutting tools are their low tensile strength and fracture tough- to crack deflection due to the presence of residual stresses ness. Different approaches to improve mechanical properties between the layers. Si3N4 based laminates have therefore like fracture toughness and wear performance include:(a) the potential for considerable improvements in term of prod enforcement by a textured second phase or multi-layer uct lifetime and energy consumption reduction for several laminates [3-5] and composites and(b)second phase addi- applications, e.g. cutting tools [14]. Similar to composites, tions such as Tin or TiB2 being dispersed homogenously additional residual stresses are formed within laminates due into the matrix of Si3N4 [6-10]. Silicon nitride compos- to the difference in thermal expansion coefficient between ites with 230-40 wt% TiN have shown to increase fracture two layers(Si3 N4 and Si3N4-TIN composite)upon cooling from the sintering temperature Corresponding author. Tel: +41 33 228 2963; fax: +41 33 228 44 90. The abrasive wear resistance of materials has been E-mail address: mousab. hadad @empa.ch(M. Hadad) reported to increase generally with both increasing hardness 0043-1648/- see front matter O 2005 Elsevier B V. All rights reserved doi:10.1016/wear2005.03.027

Wear 260 (2006) 634–641 Tribological behaviour of Si3N4 and Si3N4–%TiN based composites and multi-layer laminates M. Hadad a,∗, G. Blugan b, J. Kubler ¨ b, E. Rosset c, L. Rohr a, J. Michler a a EMPA, Swiss Institute for Material Science and Technology, Feuerwerkerstr. 39, CH-3600 Thun, Switzerland b EMPA, Laboratory for High Performance Ceramics, 8600 Duebendorf, Switzerland c University of Applied Sciences of Geneva, Switzerland Received 12 January 2004; received in revised form 1 March 2005; accepted 16 March 2005 Available online 24 June 2005 Abstract Si3N4–TiN based multi-layer laminates exhibit differences in residual stress between individual layers due to a variation of the thermal expansion coefficient between the layers. The residual stress distribution in these multi-layer laminates is known to improve the apparent macroscopic fracture toughness. In this work, the tribological behaviour of bulk, composites and multi-layers laminates are investigated. Si3N4 bulk, Si3N4 based composites with 10, 20 and 30 wt% TiN and different multi-layer laminates have been tested under dry conditions with reciprocal movement using a ball-on-block configuration. In particular, the influence of sliding directions with respect to the layer orientations has been investigated. The experimental results show that wear resistance increased with increasing TiN content in Si3N4–TiN composites. However, multi-layer laminates exhibit an up to three times higher apparent fracture toughness, but do not show an improvement of wear resistance compared to composites. © 2005 Elsevier B.V. All rights reserved. Keywords: Silicon nitride; Composites and laminate ceramics; Ball-on-block wear testing; Anisotropy; Wear mechanism 1. Introduction Silicon nitride Si3N4 based ceramics have attractive phys￾ical properties such as high hardness, chemical inertness and low density. One of the most important applications are cut￾ting tools and machining of hard materials, like cast iron and nickel based alloys [1,2]. The main limitations of ceramic cutting tools are their low tensile strength and fracture tough￾ness. Different approaches to improve mechanical properties like fracture toughness and wear performance include: (a) reinforcement by a textured second phase or multi-layer laminates [3–5] and composites and (b) second phase addi￾tions such as TiN or TiB2 being dispersed homogenously into the matrix of Si3N4 [6–10]. Silicon nitride compos￾ites with 230–40 wt% TiN have shown to increase fracture ∗ Corresponding author. Tel.: +41 33 228 29 63; fax: +41 33 228 44 90. E-mail address: mousab.hadad@empa.ch (M. Hadad). toughness up to 100% and to enhance fracture strength up to 1000 MPa [8]. This has been attributed mainly to resid￾ual stress at hetero-phase boundaries (Si3N4–TiN) caused by mismatch of thermal expansion coefficient between TiN and Si3N4. Similarly, multi-layer ceramic laminates have been previously reported to have considerable improvement in macroscopic fracture toughness [11–13]. This was attributed to crack deflection due to the presence of residual stresses between the layers. Si3N4 based laminates have therefore the potential for considerable improvements in term of prod￾uct lifetime and energy consumption reduction for several applications, e.g. cutting tools [14]. Similar to composites, additional residual stresses are formed within laminates due to the difference in thermal expansion coefficient between two layers (Si3N4 and Si3N4–TiN composite) upon cooling from the sintering temperature. The abrasive wear resistance of materials has been reported to increase generally with both increasing hardness 0043-1648/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.wear.2005.03.027

M. Hadad et al. /Wear 260(2006)634-641 and fracture toughness [15-18]. It has, however, been found that the improvement in mechanical properties of laminates does not lead to a higher wear resistance [5]. Furthermore, it has been shown that laminates with compressive stresses has higher wear resistance than in those of tensile stresses [14. For instance, in case of Si3N4, mechanical properties and abrasive wear resistance were reported to vary inversely to grain size of precipitated second phase like Bn or TiN [19]. Si3N4 was found to be unsuitable for machining of steel because of the chemical affinity between this pair of materials 2]. Studies on friction coefficient and wear of self-mated Si3N4 have been widely reported Under un-lubricated condi- (0.7-0.9)and high wear valu (5 x 10- to 10-mm/Nm) have been observed [20-22] Si3N4-nin composites under dry conditions show that wear resistance ofSi3 N4 was increased by addition of Tr ver stacking sequence of Si3N4/Si3N4+%TIN laminate system. However, under dry conditions, self-mated TiN has a slightly higher friction coefficient than self-mated SiN4. This has into dough. Tapes were then formed by twin roll compaction een attributed to the absence of titanium oxide formation, and were stacked in 4/1 system(a tensile layer Si3N4-X% which could reduce friction and wear rate [26]. In order TiN composites with 620 um of thickness and compres- to reveal if there is a tribological impact of the improved sive Si3N4 layer with 150 um of thickness as shown in mechanical properties of the laminate structure, the tribo- Fig. 1). The stacks were hot pressed in vacuum. Different logical behaviour of three materials: bulk Si3 N4, Si3N4-TIN residual stresses exist in the layers due to the difference in composites and Si3N4/Si3N4-TIN laminates have been inves- coefficient of thermal expansion(CTE)between two different tigated here. In particular, different sliding directions with components In the previous work, the mechanical proper- respect to the layers in laminates have been tested under oscil- ties were analysed [27-29]. The results are summarised in lating, un-lubricated sliding Table 2.2. Wear testing and analysis Friction and wear experiments were conducted with a 2.1. Specimens preparation reciprocating movement in a ball-on-block sliding wear est setup Is The specimens used in this investigation were hot pressed 95 un-lubricated wear testing. The top oscillating speci Si3N4, Si3N4-TiN composites and Si3N4 based laminates. men( ball) acts on the bottom block(tested specimen)at For composite materials, TIN was introduced during the pre-programmed frequency of oscillation, stroke and load milling of starting powders in order to insure homogeneous settings. The friction force is continually measured by a sen- dispersion of TIN into the Si3N4 resulting in composite mate- sor. The upper oscillating specimen is a 9.525 mm diameter rials with 10, 20 and 30 wt%TiN. Multi-layer laminate mate- bearing ball of silicon nitride having Ra=0.007 um surface rials consist of Si3 N4 bulk and Si3 N4-X%TIN composite roughness, -1500 HV hardness and a density of 3.2 g/cm layers. The layers of laminate were manufactured by milling Test conditions were as follows: 1.3 GPa of Hertzian pressure the powders, addition of a binder and drying of the mixture stroke length of 2 mm, reciprocating frequency of 10 Hz, rel- Table I Mechanical property of bulk, composites and laminates materials according to 27-29 Material Toughness KiC(MPam) Strength(MPa Young,s modulus(GE Hardness(Hv 0.5) 790.2 Si3Na-10% TN 4.4 1384 Si3Na-20% TnN Si3Na-30% TiN 4.7 784.7 Laminates Si3 Na/Si3 N4-10% TIN 13N4Si3 N4-20% TIN 9.24 Si3N4/Si3 Na-30% TIN

M. Hadad et al. / Wear 260 (2006) 634–641 635 and fracture toughness [15–18]. It has, however, been found that the improvement in mechanical properties of laminates does not lead to a higher wear resistance [5]. Furthermore, it has been shown that laminates with compressive stresses has higher wear resistance than in those of tensile stresses [14]. For instance, in case of Si3N4, mechanical properties and abrasive wear resistance were reported to vary inversely to grain size of precipitated second phase like BN or TiN [19]. Si3N4 was found to be unsuitable for machining of steel, because of the chemical affinity between this pair of materials [1,2]. Studies on friction coefficient and wear of self-mated Si3N4 have been widely reported. Under un-lubricated condi￾tions, high friction coefficients (0.7–0.9) and high wear values (5 × 10−5 to 10−4 mm3/N m) have been observed [20–22]. Si3N4–TiN composites under dry conditions show that wear resistance of Si3N4 was increased by addition of TiN [23–25]. However, under dry conditions, self-mated TiN has a slightly higher friction coefficient than self-mated Si3N4. This has been attributed to the absence of titanium oxide formation, which could reduce friction and wear rate [26]. In order to reveal if there is a tribological impact of the improved mechanical properties of the laminate structure, the tribo￾logical behaviour of three materials: bulk Si3N4, Si3N4–TiN composites and Si3N4/Si3N4–TiN laminates have been inves￾tigated here. In particular, different sliding directions with respect to the layers in laminates have been tested under oscil￾lating, un-lubricated sliding. 2. Experiments 2.1. Specimens preparation The specimens used in this investigation were hot pressed Si3N4, Si3N4–TiN composites and Si3N4 based laminates. For composite materials, TiN was introduced during the milling of starting powders in order to insure homogeneous dispersion of TiN into the Si3N4 resulting in composite mate￾rials with 10, 20 and 30 wt% TiN. Multi-layer laminate mate￾rials consist of Si3N4 bulk and Si3N4–X% TiN composite layers. The layers of laminate were manufactured by milling the powders, addition of a binder and drying of the mixture Fig. 1. Macrostructure of layer stacking sequence of the 4/1 Si3N4/Si3N4 + %TiN laminate system. into dough. Tapes were then formed by twin roll compaction and were stacked in 4/1 system (a tensile layer Si3N4–X% TiN composites with ∼620
m of thickness and compres￾sive Si3N4 layer with ∼150
m of thickness as shown in Fig. 1). The stacks were hot pressed in vacuum. Different residual stresses exist in the layers due to the difference in coefficient of thermal expansion (CTE) between two different components. In the previous work, the mechanical proper￾ties were analysed [27–29]. The results are summarised in Table 1. 2.2. Wear testing and analysis Friction and wear experiments were conducted with a reciprocating movement in a ball-on-block sliding wear configuration. The test setup is similar to ASTM G133- 95 un-lubricated wear testing. The top oscillating speci￾men (ball) acts on the bottom block (tested specimen) at pre-programmed frequency of oscillation, stroke and load settings. The friction force is continually measured by a sen￾sor. The upper oscillating specimen is a 9.525 mm diameter bearing ball of silicon nitride having Ra = 0.007
m surface roughness, ∼1500 HV hardness and a density of 3.2 g/cm3. Test conditions were as follows: 1.3 GPa of Hertzian pressure, stroke length of 2 mm, reciprocating frequency of 10 Hz, rel￾Table 1 Mechanical property of bulk, composites and laminates materials according to [27–29] Material Toughness KIC (MPa m1/2) Strength (MPa) Young’s modulus (GPa) Hardness (Hv 0.5) Bulk Si3N4 4.26 790.2 303 1556 Composites Si3N4–10% TiN 4.47 685 311 1384 Si3N4–20% TiN 4.61 883.9 317 1336 Si3N4–30% TiN 4.71 784.7 330 1396 Laminates Si3N4/Si3N4–10% TiN 9.75 590 308 – Si3N4/Si3N4–20% TiN 9.24 699 305 – Si3N4/Si3N4–30% TiN 14.3 707 313 –

636 M. Hadad et al./ Wear 260(2006)634-641 ative humidity during testing of 20-40%RH, sliding distance an increase of Tin content in the matrix, e.g. Si3N4-30% 36 m and ambient temperature Tin revealed three times higher wear resistance than bulk The same tribological tests were performed at 300C Si3N4 only on Si3N4 bulk and Si3N4+30% TIN composite. Before and after each experiment, the specimens were cleaned and 3.1.1. The influence of the sliding directions on wear weighed. The cleaning process consisted of: (a) soaking the resistance specimens in acetone for 5 min, (b)in subsequent treatment Transversal sliding shows the highest wear resistance. In in an ultrasonic bath of acetone and alcohol for 5 min,(c) some cases the wear resistance is a factor three times higher Immersion the spe ecimens in alcohol for 5 min and d)drying compared to sliding on the outer layer as shown in Fig 2b the specimens under cold air for 5 min This difference in wear resistance is due to the tin containing The worn surfaces and wear debris were investigated in the inner layer with an optical profilometer and a scanning electron micro- scope(SEM). The SEM(LEO DSM 962)is equipped with 3. 1.2. The influence of the difference in laminates on the an energy dispersive X-ray spectrometer(EDS) for elemen- wear resistance tal analysis (Voyager 4, Noran Instrument, Liquid nitrogen With respect to the sliding directions(Fig 2b) and com- cooled Si(Li) detector). The EDS working conditions were a paring among all laminates, laminates 30 shows the highest take-off angle 35, working distance to the specimen 25 mm wear resistance, which could confirm the improvement of and accelerating voltage of 15 kV. We used EDS to analyse wear resistance by adding TiN. On the other hand, comparing the worn surface and wear particles. In this case, our EDs between laminates 20 and 10. the difference in wear resis- analysis delivers only semi-quantitative results as we anal tance could be within the experimental scatter. yse non-planar surfaces and light elements like oxygen. Also EDS reveals only information on compositions from the outer 3. 1.3. Comparison between laminates and composites surface as the beam penetrates only a certain depth into the materials sample surface, which is in our experimental conditions in Despite the high scatter, a comparison between laminates nge of 1-2 um de and composites materials is possible, in the case of laminate 30 compared to Si3N4-30% TiN composite and laminate 10 compared to Si3N4-10%TIN composite, laminates subjected 3. Results and discussion to transversal or longitudinal sliding solicitation show that the wear resistance of those laminates and composites was 3. Wear rate similar. On the other hand. laminate 20 shows lower wear resistance compared to Si3N4-20% TiN composite Slid- Each determined wear rate is an average value from ng on the outer layer shows also a lower wear resistance three wear tests. Fig. 2a shows the measured wear rates of than bulk materials. This may be attributed to the presence bulk and composites. The wear resistance increased with of porosity within the layer. Wear is found to be produced C Laminates 10 Laminates 20 5a Laminates 30 20E-04 20E04 口S3N4+10%TN 冒15E04 圈siN4+20%TN 15E04 g10E04 50E05 5.0E-05 Fig. 2. Wear rate results: (a)composites and bulk and (b) laminates with a schematic presentation of sliding direction

636 M. Hadad et al. / Wear 260 (2006) 634–641 ative humidity during testing of 20–40% RH, sliding distance 36 m and ambient temperature. The same tribological tests were performed at 300 ◦C only on Si3N4 bulk and Si3N4 + 30% TiN composite. Before and after each experiment, the specimens were cleaned and weighed. The cleaning process consisted of: (a) soaking the specimens in acetone for 5 min, (b) in subsequent treatment in an ultrasonic bath of acetone and alcohol for 5 min, (c) immersion the specimens in alcohol for 5 min and (d) drying the specimens under cold air for 5 min. The worn surfaces and wear debris were investigated with an optical profilometer and a scanning electron micro￾scope (SEM). The SEM (LEO DSM 962) is equipped with an energy dispersive X-ray spectrometer (EDS) for elemen￾tal analysis (Voyager 4, Noran Instrument, Liquid nitrogen cooled Si(Li) detector). The EDS working conditions were a take-off angle 35◦, working distance to the specimen 25 mm and accelerating voltage of 15 kV. We used EDS to analyse the worn surface and wear particles. In this case, our EDS analysis delivers only semi-quantitative results as we anal￾yse non-planar surfaces and light elements like oxygen. Also EDS reveals only information on compositions from the outer surface as the beam penetrates only a certain depth into the sample surface, which is in our experimental conditions in the range of 1–2
m depth. 3. Results and discussion 3.1. Wear rate Each determined wear rate is an average value from three wear tests. Fig. 2a shows the measured wear rates of bulk and composites. The wear resistance increased with an increase of TiN content in the matrix, e.g. Si3N4–30% TiN revealed three times higher wear resistance than bulk Si3N4. 3.1.1. The influence of the sliding directions on wear resistance Transversal sliding shows the highest wear resistance. In some cases, the wear resistance is a factor three times higher compared to sliding on the outer layer as shown in Fig. 2b. This difference in wear resistance is due to the TiN containing in the inner layer. 3.1.2. The influence of the difference in laminates on the wear resistance With respect to the sliding directions (Fig. 2b) and com￾paring among all laminates, laminates 30 shows the highest wear resistance, which could confirm the improvement of wear resistance by adding TiN. On the other hand, comparing between laminates 20 and 10, the difference in wear resis￾tance could be within the experimental scatter. 3.1.3. Comparison between laminates and composites materials Despite the high scatter, a comparison between laminates and composites materials is possible, in the case of laminate 30 compared to Si3N4–30% TiN composite and laminate 10 compared to Si3N4–10% TiN composite, laminates subjected to transversal or longitudinal sliding solicitation show that the wear resistance of those laminates and composites was similar. On the other hand, laminate 20 shows lower wear resistance compared to Si3N4–20% TiN composite. Slid￾ing on the outer layer shows also a lower wear resistance than bulk materials. This may be attributed to the presence of porosity within the layer. Wear is found to be produced Fig. 2. Wear rate results: (a) composites and bulk and (b) laminates with a schematic presentation of sliding direction

M. Hadad et al. /Wear 260(2006)634-641 637 by mechanical contact and subsequent particle detachment ature and high relative humidity, tribochemical reactions are that takes place at the micrometer and submicrometer scale known to be the dominate wear mechanisms [22]. Titanium The residual stresses in laminates did not contribute to any oxide was shown to form under mild conditions(up to 0.9 GPa improvement of wear resistance compared to the composites. of Herzian pressure and 0. 1 m/s of velocity)[26, 30]. The The residual stresses in layers of laminate were calculated won surface of Si3 N4 bulk tested under the ambient temper and reported in the previous work [29]. A possible explana- ature(Fig 3a) looks like a polished area containing porosity tion is that local contact pressures in particle abrasion are However, the worn surface of composites(Fig. 3b) shows much higher than residual stresses. Similar results of lami- cavities in the order of the grain size of Tin(bright areas) nates structured materials have been published, where resid- This may result from intergranular fracture at the interfaces ual stresses have not contributed to improve wear resistance Si3 N4/TIN, and subsequent pull-out of Tin particles. Bulk and composites generate very smooth wear debris, agglom erated and attached to each other Fig 4). The wear debris has 3.1.4. Wear rate of materials at 300C both spherical and irregular shape between 0. 1 and 0.5 um of An increase in temperature to 300C led to a decrease of diameter. The loose wear debris and the worn surfaces were 10 times in wear resistance of Si3 N4-30% TiN composites. analysed by EDs. We could not detect the presence of oxy- Similarly, the wear resistance of Si3 N4 bulk decreased by a gen. Therefore, we conclude that the Tin O2n-I film removal factor five times. However, wear resistance of Si3N4-30% is not the main wear mechanism These results are in agree- TIN composites still was two times higher than bulk Si3N4 ment with the literature [25] wear resistance under the same conditions The worn surface of laminate in Fig. 3c shows the two inner layers, Si3N4 and Si3N4-30% TIN composite(longitu- 3. 2. Wear mechanism and friction dinal sliding). The composite inner layer within the laminate shows a worn surface similar to worn composites. Concerning Mechanical and/or tribochemical were the dominant wear the Si3N4 inner layers, the polished area with porosity could mechanisms of Si3 N4 based ceramics[30]. Mechanical wear not be evidenced The coefficient of friction has been reported modes include intergranular fracture, delamination and grain in many reviews to vary between 0.85 and 1 [22, 30].Our boundary fatigue at the interface between B-Si3N4 and glass experimental results of all materials at the ambient tempera phase on and below the surface. Furthermore, the elongation ture revealed high friction values from I to 1.2 (Figs. 5 and 6) of these cracks below the surface may lead to the delamina- This may be due to the absence or removal of the tribo-layer ion of larger surface region. The tribochemical reactions of like titanium oxide because of the high contact pressure. This Si3 na based ceramic mainly leads to the formation of phases is coherent with the absence of oxygen in the contact area. such as SiO2, SiOH and SiO,N, [20]. Si3 N4-TiN composites The observed wear debris was repulsed out of contact area by are able to form a stable oxide film(Tin O2n-i), which can be mechanical push. Otherwise the wear debris could have had beneficial to the tribological behaviour. Under high temper- a bearing role as a third body, which could reduce friction Bulk inner layer Composite 30%iN inner tayer aces at ambient temperature: (a)Si3N4 bulk, (b)Si3N4-30% nin composite worn surface inclined some degree and(c) on(not inclined)

M. Hadad et al. / Wear 260 (2006) 634–641 637 by mechanical contact and subsequent particle detachment that takes place at the micrometer and submicrometer scale. The residual stresses in laminates did not contribute to any improvement of wear resistance compared to the composites. The residual stresses in layers of laminate were calculated and reported in the previous work [29]. A possible explana￾tion is that local contact pressures in particle abrasion are much higher than residual stresses. Similar results of lami￾nates structured materials have been published, where resid￾ual stresses have not contributed to improve wear resistance [5]. 3.1.4. Wear rate of materials at 300 ◦C An increase in temperature to 300 ◦C led to a decrease of 10 times in wear resistance of Si3N4–30% TiN composites. Similarly, the wear resistance of Si3N4 bulk decreased by a factor five times. However, wear resistance of Si3N4–30% TiN composites still was two times higher than bulk Si3N4 wear resistance under the same conditions. 3.2. Wear mechanism and friction Mechanical and/or tribochemical were the dominant wear mechanisms of Si3N4 based ceramics [30]. Mechanical wear modes include intergranular fracture, delamination and grain boundary fatigue at the interface between -Si3N4 and glass phase on and below the surface. Furthermore, the elongation of these cracks below the surface may lead to the delamina￾tion of larger surface region. The tribochemical reactions of Si3N4 based ceramic mainly leads to the formation of phases such as SiO2, SiOH and SiOxNy [20]. Si3N4–TiN composites are able to form a stable oxide film (TinO2n−1), which can be beneficial to the tribological behaviour. Under high temper￾ature and high relative humidity, tribochemical reactions are known to be the dominate wear mechanisms [22]. Titanium oxide was shown to form under mild conditions (up to 0.9 GPa of Herzian pressure and 0.1 m/s of velocity) [26,30]. The worn surface of Si3N4 bulk tested under the ambient temper￾ature (Fig. 3a) looks like a polished area containing porosity. However, the worn surface of composites (Fig. 3b) shows cavities in the order of the grain size of TiN (bright areas). This may result from intergranular fracture at the interfaces Si3N4/TiN, and subsequent pull-out of TiN particles. Bulk and composites generate very smooth wear debris, agglom￾erated and attached to each other (Fig. 4). The wear debris has both spherical and irregular shape between 0.1 and 0.5
m of diameter. The loose wear debris and the worn surfaces were analysed by EDS. We could not detect the presence of oxy￾gen. Therefore, we conclude that the TinO2n−1 film removal is not the main wear mechanism. These results are in agree￾ment with the literature [25]. The worn surface of laminate in Fig. 3c shows the two inner layers, Si3N4 and Si3N4–30% TiN composite (longitu￾dinal sliding). The composite inner layer within the laminate shows a worn surface similar to worn composites. Concerning the Si3N4 inner layers, the polished area with porosity could not be evidenced. The coefficient of friction has been reported in many reviews to vary between 0.85 and 1 [22,30]. Our experimental results of all materials at the ambient tempera￾ture revealed high friction values from 1 to 1.2 (Figs. 5 and 6). This may be due to the absence or removal of the tribo-layer like titanium oxide because of the high contact pressure. This is coherent with the absence of oxygen in the contact area. The observed wear debris was repulsed out of contact area by mechanical push. Otherwise the wear debris could have had a bearing role as a third body, which could reduce friction. Fig. 3. SEM micrographs of worn surfaces at ambient temperature: (a) Si3N4 bulk, (b) Si3N4–30% TiN composite worn surface inclined some degree and (c) laminates at longitudinal sliding direction (not inclined).

638 M. Hadad et al. /Wear 260(2006)634-641 Fig 4. SEM micrographs of wear debris collected after sliding test: (a)Si3 N4 bulk and (b)Si3N4-30%Tn composites Wear mechanism and friction at 300oC: literature data and velocities seems to lead a removal of the oxide layer suggests that the wear of Si3 N4/Si3 Na sliding pairs at 300C under mechanical action, which leads to an increase in wear is caused mainly by a thermomechanically dominated wear rate. The EDS analysis of the worn surface suggested oxy mechanism with grain boundary fatigue at the interface gen contents in the tenth of atomic percents. Despite the between B-Si3N4 and glass phase on and below the surface. semi-quantitative character of the EDs system in case of For the sliding pairs of Si3N4-TiN/Si3N4-TiN composites, non-planar surfaces and of the analysis of light elements, the tribochemically dominated wear mechanism was observed amount of oxygen indicated here is significant compared to [20]. The oxidation rate increases with increase in temper- ambient temperature results. This suggests on the one hand ature at the surface of contact. Increasing contact pressure that a tribochemical wear mechanism is active, which is in Bulk Si3N4 Si3N4-10%TiN composite g08 504 85 Time [min Time (min] Si3NA-20%TiN composite Si3N4-40%TiN composite 洲A L02 34578911121415 Time [min Time [min] Fig. 5. Friction coefficients of bulk and composites at ambient temperat

638 M. Hadad et al. / Wear 260 (2006) 634–641 Fig. 4. SEM micrographs of wear debris collected after sliding test: (a) Si3N4 bulk and (b) Si3N4–30% TiN composites. Wear mechanism and friction at 300 ◦C: literature data suggests that the wear of Si3N4/Si3N4 sliding pairs at 300 ◦C is caused mainly by a thermomechanically dominated wear mechanism with grain boundary fatigue at the interface between -Si3N4 and glass phase on and below the surface. For the sliding pairs of Si3N4–TiN/Si3N4–TiN composites, tribochemically dominated wear mechanism was observed [20]. The oxidation rate increases with increase in temper￾ature at the surface of contact. Increasing contact pressure and velocities seems to lead a removal of the oxide layer under mechanical action, which leads to an increase in wear rate. The EDS analysis of the worn surface suggested oxy￾gen contents in the tenth of atomic percents. Despite the semi-quantitative character of the EDS system in case of non-planar surfaces and of the analysis of light elements, the amount of oxygen indicated here is significant compared to ambient temperature results. This suggests on the one hand that a tribochemical wear mechanism is active, which is in Fig. 5. Friction coefficients of bulk and composites at ambient temperature.

M. Hadad et al. /Wear 260(2006)634-641 63 inate 20 三 三8 0.4 0134578911121415 0134578911121415 0134578911121415 Time [min] Time [min] HaLMMMWuhA+ 三985 884 2.a8 0.2 0134578911121415 0134578911121415 0134578911121415 Time [min] Time [min] 0.8 06 后06 e04 0134578911121415 0134578911121415 01345 911121415 Time [min Time [min Time [min] Fig. 6. Friction coefficients of laminates with different sliding orientations at the ambient temperature agreement with the literature [20, 23]. On the other hand, we at 300C. Contrary to the literature data [20], the friction can,however, not exclude that a simple oxidation process coefficient did not decrease compared to room tempera- within the contact area took place and that the increased ture experiments, see Fig. 7. However, the roughness of the due to a degradation of the mechanical material properties compare Figs. 3 and Compared to room temperature data, wear rate is related to increased thermomechanical wear worn surface increased Si3N4 under 300oC SiaNa-10%TiN composite under 2 memers~ww灬mNmy 八 0.8 0.2 345789 0134578911121415 Time [min] Time [min] Fig. 7. Friction coefficients Si3 N4 bulk and Si3 N4-30% TIN composite at 300C

M. Hadad et al. / Wear 260 (2006) 634–641 639 Fig. 6. Friction coefficients of laminates with different sliding orientations at the ambient temperature. agreement with the literature [20,23]. On the other hand, we can, however, not exclude that a simple oxidation process within the contact area took place and that the increased wear rate is related to increased thermomechanical wear due to a degradation of the mechanical material properties at 300 ◦C. Contrary to the literature data [20], the friction coefficient did not decrease compared to room tempera￾ture experiments, see Fig. 7. However, the roughness of the worn surface increased compared to room temperature data, compare Figs. 3 and 8. Fig. 7. Friction coefficients Si3N4 bulk and Si3N4–30% TiN composite at 300 ◦C

M. Hadad et al./ Wear 260(2006)634-641 Fig 8. SEM micrographs of wom surfaces morphology at 300.C:(a)Si3N4 bulk and (b)Si3N4-30% TiN composite 4. Conclusion 3]Y. Zhou, Mechanical properties and toughening mechanisms of The addition of up to 30% of titanium nitride to silicon nitride matrix led to an improvement of wear resistance of 4]J. Huang, Y.L. Chang, H.H. Lu, Fabrication of multi-laminated SigN4-Si3 N4/TiN composites and its anisotropic fracture behaviour, composites of up to three times, which has been attributed to ariations of residual stresses at the scale of the grain size due [5]A. Tarlazzi, et al., Tribological behaviour of Al2O3/ZrO2-ZrO2 lam- to the mismatch of thermal expansion coefficient between inated composites, Wear 244(2000)29-40 Si3N4 and tiN. Concerning laminate material, sliding on transversal orientation shows the highest wear resistance (7) B.-T. Lee, Y.J. Yoon, K-H. Lee, Microstructural characterization among all orientations. Comparing the different laminates of electroconductive Si3N4-TIN composites, Mater. Lett. 47(2001) laminates 30 shows the highest wear resistance On the other 71-76 hand,comparing laminates to composites, laminate struc- [8]A. Bellosi, A. Fiegna, A Giachello, Microstructure and Properties ture with additional stresses between layers that leads to of Electrically Conductive Si3N4-TIN Composites, Elsevier Science Publishers B V, 1991, pp 225-234 an increase macroscopic fracture toughness did not improve [9]HJ. Choi,K-S. Cho, J-G. Lee, R-curve behaviour of silicon wear resistance compared to Si3N4-%TIN composites. This tride- titanium nitride composites, J. Am. Ceram. Soc. 80(1997) is because wear is produced by mechanical contact, and sub- 2681-2684. sequent particle detachment that takes place at the micrometer [10] J.-L. Huang, S-Y. Chen, Investigation of silicon nitride composites and submicrometer scale. At 300oC. the wear resistance of toughened with prenitrided TiB2 21(1995)77-83 Si3N4-30% TIN composites still was two times greater than [11C -H. Yeh, M.-H. Hon, The SiN by slip casting, Cer SiaNa bulk [12]C. Wang, Control of composition ture in laminated sili- con nitride/boron nitride composites, J. Am. Ceram Soc. 85(2002) 2457-2461 [13]J.-L. Huang, F-C. Chou, H -H. Lu, Investigation of Acknowledgement Si3N4-TiNSi3N4-Si3N4 trilayer composites with residual surface compression, J. Mater. Res. 12(9)(1997)2357-2365 We would like to thank the gebert -Ruf Foundation [14 M.F. Amateau, Performance of laminated ceramic composite cutting Switzerland for the support under project VERBUNDLOTE [15]J.A. Hawk, DA(1995)317- Alman, J.J. Petrovic, Abrasive wear of (No. 1125.045)as well as the European Commission and SigNa-MoSi2 composites, Wear 203-204(1997)247-256 Swiss BBW under FP5 INCO Project LAMINATES(ICA- [16]MN. Gardos, R.G. Hardisty, Fracture ess-and hardness. CT-2000-10020; BBW contract 99.0785) ceramIcs. In 36(1993)652-660. [17E. Rabinowicz, Friction and Wear of Materials, second ed, A. wiley nescience Publication, 1995. [18]S.-T.Buljan, S.E. Wayne, Wear and design of ceramic cutting tool References [19S.F. Wayne, Microstructural aspects of Si3N4-TiC composites affect- [X. Zhao, Tribological characteristics of Si3N4 ceramic sliding on ing abrasion and erosion resistance, Tribol. Trans. 45(1990) ainless steel, Wear 206(1997)76-82 2]X.Z. Zhao, et al., Wear behaviour of Si N4 ceramic cutting tool [20]A. Skopp, M. Woydt, K. Habig, Tribological behavior of sill- Ceram.Int.25(4)(1999)309315 1000°C, Wear I8l-183(1995)571-580

640 M. Hadad et al. / Wear 260 (2006) 634–641 Fig. 8. SEM micrographs of worn surfaces morphology at 300 ◦C: (a) Si3N4 bulk and (b) Si3N4–30% TiN composite. 4. Conclusion The addition of up to 30% of titanium nitride to silicon nitride matrix led to an improvement of wear resistance of composites of up to three times, which has been attributed to variations of residual stresses at the scale of the grain size due to the mismatch of thermal expansion coefficient between Si3N4 and TiN. Concerning laminate material, sliding on transversal orientation shows the highest wear resistance among all orientations. Comparing the different laminates, laminates 30 shows the highest wear resistance. On the other hand, comparing laminates to composites, laminate struc￾ture with additional stresses between layers that leads to an increase macroscopic fracture toughness did not improve wear resistance compared to Si3N4–%TiN composites. This is because wear is produced by mechanical contact, and sub￾sequent particle detachment that takes place at the micrometer and submicrometer scale. At 300 ◦C, the wear resistance of Si3N4–30% TiN composites still was two times greater than Si3N4 bulk. Acknowledgements We would like to thank the Gebert-Ruf Foundation, ¨ Switzerland for the support under project VERBUNDLOTE (No. 1125.045) as well as the European Commission and Swiss BBW under FP5 INCO Project LAMINATES (ICA￾CT-2000-10020; BBW contract 99.0785). References [1] X. Zhao, Tribological characteristics of Si3N4 ceramic sliding on stainless steel, Wear 206 (1997) 76–82. [2] X.Z. Zhao, et al., Wear behaviour of Si3N4 ceramic cutting tool material against stainless steel in dry and water-lubricated conditions, Ceram. Int. 25 (4) (1999) 309–315. [3] Y. Zhou, Mechanical properties and toughening mechanisms of Si3N4 matrix laminated ceramic composite, Key Eng. Mater. 161–163 (1999) 353–356. [4] J. Huang, Y.-L. Chang, H.H. Lu, Fabrication of multi-laminated Si3N4–Si3N4/TiN composites and its anisotropic fracture behaviour, J. Mater. Res. Soc. 12 (1997) 2337–2344. [5] A. Tarlazzi, et al., Tribological behaviour of Al2O3/ZrO2–ZrO2 lam￾inated composites, Wear 244 (2000) 29–40. [6] M. Woydt, Wear engineering oxides/anti-wear oxides, Wear 218 (1998) 84–95. [7] B.-T. Lee, Y.-J. Yoon, K.-H. Lee, Microstructural characterization of electroconductive Si3N4–TiN composites, Mater. Lett. 47 (2001) 71–76. [8] A. Bellosi, A. Fiegna, A. Giachello, Microstructure and Properties of Electrically Conductive Si3N4–TiN Composites, Elsevier Science Publishers B.V., 1991, pp. 225–234. [9] H.J. Choi, K.-S. Cho, J.-G. Lee, R-curve behaviour of silicon nitride-titanium nitride composites, J. Am. Ceram. Soc. 80 (1997) 2681–2684. [10] J.-L. Huang, S.-Y. Chen, Investigation of silicon nitride composites toughened with prenitrided TiB2, Ceram. Int. 21 (1995) 77–83. [11] C.-H. Yeh, M.-H. Hon, The Si3N4 and Si3N4/TiC layered composites by slip casting, Ceram. Int. 23 (1996) 361–366. [12] C. Wang, Control of composition and structure in laminated sili￾con nitride/boron nitride composites, J. Am. Ceram. Soc. 85 (2002) 2457–2461. [13] J.-L. Huang, F.-C. Chou, H.-H. Lu, Investigation of Si3N4–TiN/Si3N4–Si3N4 trilayer composites with residual surface compression, J. Mater. Res. 12 (9) (1997) 2357–2365. [14] M.F. Amateau, Performance of laminated ceramic composite cutting tools, Ceram. Int. 21 (1995) 317–323. [15] J.A. Hawk, D.E. Alman, J.J. Petrovic, Abrasive wear of Si3N4–MoSi2 composites, Wear 203–204 (1997) 247–256. [16] M.N. Gardos, R.G. Hardisty, Fracture toughness-and hardness￾dependent polishing wear of silicon nitride ceramics, Tribol. Trans. 36 (1993) 652–660. [17] E. Rabinowicz, Friction and Wear of Materials, second ed., A. Wiley￾Interscience Publication, 1995. [18] S.-T. Buljan, S.F. Wayne, Wear and design of ceramic cutting tool materials, Wear 133 (1989) 309–321. [19] S.F. Wayne, Microstructural aspects of Si3N4–TiC composites affect￾ing abrasion and erosion resistance, Tribol. Trans. 45 (1990) 553–558. [20] A. Skopp, M. Woydt, K. Habig, Tribological behavior of sili￾con nitride materials under unlubricated sliding between 22 ◦C and 1000 ◦C, Wear 181–183 (1995) 571–580

M. Hadad et al. /Wear 260(2006)634-641 64 21]SW. Lee, M.C. Shen, S M. Hsu, ps: silicon nitride, 227]G. Blugan, M. Hadad, J. Janczak-Rusch, J. Kuebler, T. frac. ey Eng Mater. 89-91(1994)7 tography, mechanical properties, and microstructure of commercial [22]TE. Fischer, H. Tomizawa, silicon nitride-titanium nitride composites, Am. Ceram Soc. 88(4) microfracture in the friction an silicon nitride. Wear 105 2005)926-93 985)29-45 [28]G. Blugan, J. Kubler, N keramische Mehrschicht 23]Y. Imada, The tribological reaction accompanying on and m Komposit aus Silizium- of silicon nitride containing titanium nitride, Tr ASME 114 14 Symposium Verband. erkstoffe und Werkstoffverl 3-527-30762-1,pp.337 [24]Z. Da-Ming, A com ady on the microstructure and logical properties of Si3N4 and TIN films produced by the 229M. Lugovy, V. Slyunyayev, N. Orlovskaya, G. Blugan, J. Kuebler, method, Tribol. Int. 29(6)(1996)507-513 M. Lewis, Apparent fracture toughness of Si3 Na-based laminates [25]F. Brenscheidt, et al., Wear mechanisms in titanium implanted silicon or tensile stresses in surface lavers. Acta nitride ceramics, NIMB 129(1997)483-486 Mater.53(2005)289296 [26]C. Melandri, M.G. Gee, G de Portu, S Guicciardi, High temperature [30]K Kato, K. Adachi, Wear of advanced ceramics, Wear 9233(2002) friction and wear testing of silicon nitride ceramic, Tribol. Int. 28

M. Hadad et al. / Wear 260 (2006) 634–641 641 [21] S.W. Lee, M.C. Shen, S.M. Hsu, Ceramic wear maps: silicon nitride, Key Eng. Mater. 89–91 (1994) 751–756. [22] T.E. Fischer, H. Tomizawa, Interaction of tribochemistry and microfracture in the friction and wear of silicon nitride, Wear 105 (1985) 29–45. [23] Y. Imada, The tribological reaction accompanying friction and wear of silicon nitride containing titanium nitride, Trans. ASME 114 (1992) 230–235. [24] Z. Da-Ming, A comparative study on the microstructure and tribo￾logical properties of Si3N4 and TiN films produced by the IBED method, Tribol. Int. 29 (6) (1996) 507–513. [25] F. Brenscheidt, et al., Wear mechanisms in titanium implanted silicon nitride ceramics, NIMB 129 (1997) 483–486. [26] C. Melandri, M.G. Gee, G. de Portu, S. Guicciardi, High temperature friction and wear testing of silicon nitride ceramic, Tribol. Int. 28 (1995) 403–413. [27] G. Blugan, M. Hadad, J. Janczak-Rusch, J. Kuebler, T. Graule, Frac￾tography, mechanical properties, and microstructure of commercial silicon nitride–titanium nitride composites, Am. Ceram. Soc. 88 (4) (2005) 926–933. [28] G. Blugan, J. Kubler, N. Orlovskaya, keramische Mehrschicht- ¨ Laminates aus Siliziumnitrid und einem Komposit aus Silizium￾nitrid/Titannitrid, Verbundwerkstoffe, 14 Symposium Verbandw￾erkstoffe und Werkstoffverbunde, ISBN 3-527-30762-1, pp. 337– 342. [29] M. Lugovy, V. Slyunyayev, N. Orlovskaya, G. Blugan, J. Kuebler, M. Lewis, Apparent fracture toughness of Si3N4-based laminates with residual compressive or tensile stresses in surface layers, Acta Mater. 53 (2005) 289–296. [30] K. Kato, K. Adachi, Wear of advanced ceramics, Wear 9233 (2002) 1–8