Aus Wissenschaft und Forschung 28 Tribologie + Schmierungstechnik · 65. Jahrgang · 5/ 2018 Tribofilm formation of a-C: H coatings under influence of temperature in boundary lubricated oscillating sliding against alumina and silicon nitride R. Wäsche, A. R. Jayachandran, G. Brandt, T. Schmid, T. Tamura, T. Nakase* Der Einfluss von Temperatur und Gegenkörperwerkstoff auf die Bildung von Tribofilmen auf a-C: H Beschichtungen wurde mit einer Kugel-Ebene Anordnung im geschmierten Kontakt untersucht. Die Ergebnisse zeigen, dass die Verschleißvolumina exponentiell mit der Temperatur ansteigen. In Abhängigkeit von der Temperatur können zwei Bereiche mit unterschiedlichem Reibungsverhalten beobachtet werden. Die Reibungszahlen steigen von Raumtemperatur bis zu einer Übergangstemperatur von etwa 110 °C an und nehmen danach mit weitersteigender Temperatur bis 250 °C wieder ab. Ein Übergangsbereich bei Temperaturen um 110 °C wird auch im Verschleißverhalten beobachtet. Allerdings beschleunigt sich die Zunahme der Verschleißvolumina oberhalb der Übergangstemperatur, hauptsächlich durch Verschleiß der a-C: H Schicht. Obwohl ein direkter Transfer von a-C: H Material auf den Gegenkörper nicht festgestellt wurde, konnte teilweise in Abhängigkeit von der Temperatur die Bildung eines Tribofilmes auf dem Gegenkörper nachgewiesen werden. Die Bildung dieses Tribofilmes mit unterschiedlicher Zusammensetzung scheint die tribologischen Eigenschaften des Systems zu beeinflussen. Auf dem Gegenkörper aus Aluminiumoxid hat sich bei Raumtemperatur kein Tribofilm gebildet. Bei höheren Temperaturen enthält der Tribofilm auf dem Gegenkörper nur Phosphor, das aus dem Antioxidationsadditiv des Schmiermittels stammt. Auf der Kugel aus Siliziumnitrid bildet sich bei allen Temperaturen ein Tribofilm, dessen Hauptkomponente Sauerstoff ist, der aus der Tribo-oxidation des Siliziumnitrids stammt. Bei Temperaturen ab dem Übergangsbereich enthält der Tribofilm zusätzlich Phosphor und Kohlenstoff. Schlüsselwörter a-C: H, Temperatur, Ölschmierung, Aluminiumoxid, Siliziumnitrid, Tribofilm The influence of temperature and counter body material on the tribofilm formation of a-C: H coatings, has been investigated in a lubricated ball on disk contact situation. The results show that the wear volumes of the system increase exponentially with increasing temperature. Regarding friction, due to increasing temperature, two different areas have been identified. Friction increases from room temperature up to about the transition temperature range of 110 °C and decreases with further rising temperatures. The transition temperature range of 110 °C is observed in the wear behavior, too. However, contrary to friction, the wear volume is increasing with increasing temperature, mainly due to increased wear of the a-C: H coating. Although a distinct transfer layer of a-C: H has not been observed, a tribofilm has partly formed. The formation of these tribofilms with different composition seems to have a major influence on the tribological properties. On alumina ball at room temperature no tribofilm has formed. At higher temperatures, the tribofilm contains only a phosphorus compound which stems from the anti-oxidant additive of the lubricant. On silicon nitride ball, a tribofilm forms at all temperatures and oxygen is the main compound. This oxide tribofilm is formed due to the tribo-oxidation of silicon nitride. At higher temperatures above the threshold of 110 °C, the tribofilm also contains phosphorus and carbon. Keywords a-C: H, temperature, lubricated sliding, alumina, silicon nitride, tribofilm Kurzfassung Abstract * Dipl. Min., Dr.-Ing. Rolf Wäsche M. Tech., PhD Ashok Raj Jayachandran Dipl. Ing. (FH) Guido Brandt Dipl. Chem., Dr.rer.nat Thomas Schmid Bundesanstalt für Materialforschung und -prüfung (BAM), 12205 Berlin, Germany Tetsuya Tamura, M. Eng., Takuya Nakase, M. Eng., Dr. Eng. KYB corporation, 1-12-1 Asamizodai, Minami-ku, Sagamihara-shi, Kanagawa, 252-0328 JAPAN T+S_5_18.qxp_T+S_2018 28.08.18 13: 50 Seite 28 fore, to investigate and explore the tribological behavior of a-C: H coatings in the range of temperatures from room temperature up to 250 °C in the boundary lubrication regime, highlighting the influence of the counter body material. As counterbody materials, α-alumina and silicon nitride have been chosen. This decision was made for the following reasons. Besides the fact, that these materials are well known structural ceramics with many different applications in machine technology [14], they offer to study the effect of tribochemical influence on the friction and wear behavior. Both are hard materials with little influence of temperature on the mechanical properties in the intended temperature range. However, silicon nitride is prone to oxidation and the generation of silicon oxide due to tribological action may have significant influence on friction and wear, whereas α-aluminum oxide is practically inert. Ordinary steel as a counterbody material could not be used since annealing temperatures of different steels, which are available on the market in the form of a sphere, were just too low, so that these materials would lose their hardness and be prone to severe plastic deformation during the test with the predetermined parameters. 2 Experimental 2.1 Materials The DLC layer is an amorphous carbon coating containing hydrogen, classified as a-C: H. The coating with denotation KYB Type A is an industrial a-C: H coating from KYB Corporation, Sagamihara, Japan. The intermediate metal layer (adhesion layer) is chromium with a thickness of about 50 nm, and it is graded from the substrate with 100 % chromium up to the a-C: H layer. The a-C: H layer has a coating thickness of about 700 nm as determined by measuring the cross section in the Scanning Electron Microscope (SEM). The interface between substrate and coating is free of voids or pores. The surface roughness of the coating was determined as-deposited without machining or finishing using of stylus profilometer (Hommel T8000, stylus tip TKL100 with 60° and 2 µm). The values found were R a = 5 nm, R pk = 6 nm and R vk = 12 nm, respectively. The surface roughness was also determined using an Atomic Force Microscope (AFM) from Asylum corporation. The root-mean-square roughness of the coating (as deposited) was 2.53 nm and R a = 1.98 nm. The hardness of the a-C: H coating is 13 GPa, and the indentation modulus is 139 GPa, as measured with Nano indenter ELIONIX ENT-1100a with a load of 9.8 mN. The ratio of sp 2 / (sp 2 + sp 3 ) as determined with near edge X-ray absorption fine structure (NEXAFS) is 0.49, corresponding to 51 % of sp3 hybridized carbon. The hydrogen content, as measured by Elastic Recoil Detection Analysis (ERDA), is 16 at%. The a-C: H coatings were deposited on the flat surface of a cylindrical disk with a 1 Introduction Diamond-like carbon coatings (DLC) combine a variety of very attractive properties like high hardness, chemical inertness and low adhesion to most counterpart materials [1]. They show therefore a good to excellent tribological behavior depending on the loading conditions [2]. The properties of the coatings are also largely dependent on composition as well as the deposition method and conditions [3,4]. DLC coatings are therefore applied in many manufacturing processes, with great success with respect to enhancing tool life times and decreasing maintenance cycles by supplying low friction and preventing wear. The properties of the coatings are widely dependent on composition as well as the deposition method and conditions [4,5]. Liu et al. stated that the DLC films against alumina showed low values of friction coefficient (< 0.1). The low friction coefficient in the intermediate friction stage is mainly attributed to transfer layer formation, whereas the occurrence of steady-state friction is related to the formation of a graphitized tribolayer under thermal and strain effects from repeated friction [6]. Vanhulsel et al. found that the tribological behavior of the DLC films started to change around 100 °C. Also, at an elevated temperature, the friction coefficient decreased, and the wear scars became larger and deeper due to a more severe wear process. It was postulated that a degradation of the bulk properties of a-C: H coatings occurs at a temperature around 400 °C owing to the dehydrogenation of coating and the transformation of sp 3 hybridized carbon atoms into sp 2 hybridized atoms [7]. Several papers have been published, dealing with the influence of ambient temperature on the tribological performance of a-C: H coatings. They show that the ambient temperature is crucial for the stability of the coating, if tribologically loaded [5, 8, 9, 10, 11]. Results on oil lubricated sliding experiments on a-C: H coatings are more scarcely published than on those tests performed under dry conditions. Currently, existing results of a-C: H in oil lubricated sliding mostly cover the temperature range up to about 80 - 120 °C, since this is the temperature range of steady state conditions for combustion engines and their related components. However, it can be estimated that under heavy load asperity contact situations can efficiently produce 200 °C due to flash temperatures. In previous studies [12,13], it was shown that the tribological performance of hard amorphous carbon coatings under unlubricated conditions is severely decreasing with increasing temperature. Since oil temperatures in the range of 200 - 250 °C bear the danger of oil decomposition due to thermal instability of carbon bonds, synthetic ester oil with an upper application temperature of 250 °C has been applied. The aim of this paper is, there- Aus Wissenschaft und Forschung 29 Tribologie + Schmierungstechnik · 65. Jahrgang · 5/ 2018 T+S_5_18.qxp_T+S_2018 28.08.18 13: 50 Seite 29 application temperature of 250 °C, low vapor pressure and burns without residues. The oil has excellent oxidation stability due to a sulfur-phosphorous compound as oxidation protection additive. The chemical analysis of the oil revealed the following elements: Ti is 10 mg/ kg, phosphorus is 924 mg/ kg, and sulfur is 2300 mg/ kg. No further additives are present. 2.2 Tribological testing The tribological tests on the coatings were performed with an SRV III-tribometer (Optimol, Munich, Germany) using spheres of 10 mm diameter and polished surfaces as a counter body under oscillating conditions in gross slip fretting mode [15]. The test parameters are shown in Table 1. The coefficient of friction and linear wear were recorded online as primary tribological quantities. The friction force (F F ) is determined according to equation (1). The friction energy (F E ) is the energy dissipated during one cycle with stroke 2Δx and corresponds to the area within the friction loop hysteresis [15]. (1) Dividing the friction force (F F ) by the normal force (F N ) results in the coefficient of friction (f) or COF. Since the COF may scatter especially during running-in, the average of COF is formed by using the friction values recorded during the second half of the test, i. e. for a test lasting Aus Wissenschaft und Forschung 30 Tribologie + Schmierungstechnik · 65. Jahrgang · 5/ 2018 diameter of 24 mm and a height of 7.9 mm made from Cronidur 30 steel by magnetron sputtering, using a graphite target at a temperature of 380 °C in methane (CH4) atmosphere. Cronidur 30 is a nitrogen alloyed steel (1.4108) with an annealing temperature above 450 °C and a Vickers hardness HV10 of 6.8 GPa. This steel was chosen because its annealing temperature is higher than the deposition temperature of the coating so that the mechanical properties of the substrate are unaffected by the deposition process. Counterbody materials were α-alumina and silicon nitride. The form was a sphere with a polished surface. Silicon nitride is NBD 200, a bearing grade material with MgO sinter additive from Cerbec Europe (Saint Gobain), High Tech Ceram, Blankenheim Germany. This material is nearly 100 % dense due to encapsulated hot pressing. Density is 3.2 g/ cm 3 and flexural strength greater than 900 MPa. Young’s modulus is 320 GPa, and Vickers hardness HV10 is 15.5 GPa. The material of the alumina ball is sintered α-alumina Frialit F99.7 with a relative density of > 99 % from Friatec Mannheim, Germany. Impurity content is less than 0.5 %. Young’s modulus is 380 GPa, and Vickers hardness HV1 is 17.6 GPa. All surfaces of the balls are polished with the following surface roughness values Ra: silicon nitride 0.004 µm and α-alumina 0.22 µm. The lubricant used is high temperature ester oil OKS 352 (OKS, Spezialschmierstoffe GmbH, 82216 Maisach, Germany) for industrial use. This oil has an upper Table 1: Tribological system and test parameters Tribosystem Test parameters Ball 10 mm, polished Stroke ∆x 0.2 mm α-Al 2 O 3 and Si 3 N 4 Frequency ν 20 Hz Disk a-C: H on steel Cronidur30 Normal force F n 20 N Lubricating medium High temperature ester oil OKS352 Number of cycles n 10 5 Ambient Laboratory air Oil Temperature T 25 °C to 250 °C ! " ! # $%&' 100,000 cycles, this would mean to record the COF during the last 50,000 cycles and calculating the average. Figure 1 shows a schematic of the wear volumes as they appear on both ball and disk. The radius R’ corresponds to the radius of the ball after the wear test and is identical to the curvature of the wear scar of the disk. The wear scars are generally very minimal due to the high wear resistance of the coating. Wear volumes on the Figure 1: Schematic of wear volumes at ball and disk T+S_5_18.qxp_T+S_2018 28.08.18 13: 50 Seite 30 disk have therefore been evaluated by white light interferometry (WLI). The raw data as measured with WLI (New View 5022, Zygo LOT; Darmstadt, Germany) has been evaluated with the software package mountains map 6.2 (Digital Surf, Besancon, France). Figure 2 shows the schematic of wear profile as it appears on the disk specimen after wear. The radius R’ corresponds to the radius of the ball after the wear test and is identical to the curvature of the wear scar of the disc. The denoted radius R corresponds to the radius of the ball. The area difference between the curves of R and R’ corresponds to the worn part on the ball. The planimetric wear W q is the area between disk surface (coating) and the wear scar profile. The linear wear W l is simply the linear distance between disc surface and bottom of wear scar. The wear behavior is described quantitatively by the coefficient of wear, k, which is defined as the total wear volume (W v ), divided by the product of normal force F N and total sliding distance s. Equations (2) - (5) are applied for calculation of the respective wear volumes on ball and disc and the wear coefficient k. (2) (3) (4) (5) W VBall = wear volume at ball W Vdisk = wear volume at disk d k = wear scar diameter perpendicular to sliding direction d || = wear scar diameter parallel to sliding direction R = radius of ball R’ = radius of ball after testing W q = planimetric wear, cross section of wear scar from profile Δx = stroke n = number of cycles F N = normal force W l = linear wear In data series one for each counter material, three independent test series were carried out over the temperature range of room temperature up to 250 °C. Scattering of data for determined wear volumes below 100 °C was relatively low with good reproducibility. Above 100 °C scattering of wear data increased between the data series one on the one hand, and data series two and three on the other hand. Reproducibility for friction coefficients in dependence on temperature was excellent over the whole temperature range. However, wear data showed increased scattering at temperatures above 150 °C, when all three data series are considered. Reproducibility within data series 1 (three independent tests at each temperature) was excellent. 2.3 Microanalytical investigations The wear scars have been characterized by scanning electron microscopy (SEM) and energy dispersive x-ray diffraction (EDX). 3 Results and Discussion 3.1 Friction Figure 3 shows the coefficient of friction of all coated samples under lubricated conditions as a function of temperature for counterbodies alumina and silicon nitride. The dotted lines are simple polynomial fit functions. The graphs show significant and well reproducible characteristics. First, the COF of sliding pair with alumina counterbody is significantly (about 20 %) smaller than the COF for sliding pair with silicon nitride as the counter material for all temperatures. Second, the COF of each sliding pair is increasing with temperature up to a temperature range of about 100 °C for silicon nitride and about 125 °C for alumina. With further temperature increase, the COF for both counter materials are decreasing. Aus Wissenschaft und Forschung 31 Tribologie + Schmierungstechnik · 65. Jahrgang · 5/ 2018 = Vball W ÷øö çèæ - × × × ^ ' 1 1 64 2 || 2 R R d d p ( ( q Vdisk W x R d d W × D + × × × = ^ ' 1 64 2 || 2 p ( ( ( Vdisk Vball V W W W + = ( n V n V F n x W F sW k × × D × = × = 2 ( Figure 2: Schematic of wear scar profile and the relations between measuring quantities T+S_5_18.qxp_T+S_2018 28.08.18 13: 50 Seite 31 the pairings with the alumina counter material is lower than those with silicon nitride as a counter material. This finding is turned into the opposite at higher temperatures above about 150 °C. In this temperature range wear volumes strongly increase much faster with temperature for alumina counter material than for silicon nitride. Figure 6 shows wear volumes after 100,000 cycles sliding distance, separately for disc (DLC layer) and ball (ceramic). These diagrams reveal that at low temperatures the wear is mainly taking place at the ceramic counter body, whereas at high temperatures the wear is significantly higher for the a-C: H coating. At a first glance these results may be understood by the different mechanical properties of the materials. The DLC coating has excellent chemical stability and a high hardness in the lower temperature range, comparable to the two ceramic materials. Consequently, the DLC coating wears less than the two ceramic materials as is observed in Figure 6. Most probably, the wear behavior in this temperature range, below the temperature threshold of 100-120 °C, is characterized by the difference in hardness. However, above this temperature threshold, this wear characteristic is changed dramatically towards wear of the DLC coating. This is most probably due to thermal instability and modification of the carbon network of the coating. DLC Aus Wissenschaft und Forschung 32 Tribologie + Schmierungstechnik · 65. Jahrgang · 5/ 2018 Figure 4 shows the COF for alumina counter material for three different temperatures as evolution with cycles, which also translates to time and sliding distance. The graphs reveal that the friction coefficient is relatively stable. It may also be noted that for 25 and 100 °C the level of COF is constant with increasing number of cycles, whereas the COF measured at 200 °C decreases with increasing number of cycles from initially 0.16 to 0.12 at 200 000 cycles. Similar observations have been made for silicon nitride counter body, but the frictional level is about 20 % higher when compared with alumina counter material. 3.2 Wear Figure 5 shows the wear volumes and the wear coefficients for the entire system (coated sample, lubricant and counter-body) for both counter materials as a function of temperature. The wear volumes and wear coefficients are increasing with increasing temperatures. The graphs reveal that in the low temperature range, up to about 100 °C, the wear volume for Figure 3: Coefficient of friction as function of temperature for aluminum oxide and silicon nitride ball on a-C: H coated Cronidur 30 discs Figure 4: Coefficient of friction as a function of number of cycles (sliding distance) for alumina counter material for three different temperatures Figure 5: Wear volumes and wear coefficients of system for both counter materials as a function of temperature. Please note the different scales for the two counterbody materials. T+S_5_18.qxp_T+S_2018 28.08.18 13: 50 Seite 32 coatings of the type a-C: H, as used in this investigation, contain a certain amount of hydrogen and are prone to transformation by enhanced oscillation of carbon bonds owing to temperature. The carbon network will eventually transform into the stable form of graphite when the temperature is high enough. Generation of graphite on DLC coatings due to sliding action at higher temperatures is a long-known phenomenon [6]. Figure 6 reveals a distinct difference between the two counter materials in the temperature range above 150 °C. It is unlikely that only the higher hardness of alumina is causing dramatically higher wear on the a-C: H coating. Another reason might be seen in the fact that silicon nitride is prone to tribochemical wear due to oxidation and that silicon oxide particles are released into the tribo-interface, subsequently modifying friction and wear behavior [16]. To clarify this, Figure 5 is considered again. It suggests that the temperature dependence of the wear process might be exponential. A more exact analysis is possible with graphical presentation of the logarithm of wear rates k versus the reciprocal of the absolute temperature, according to the following equations (6) and (7) [17](6) (7) with A being a pre-exponential factor, hence lnA is a constant. E a is the activation energy of the wear process representing the height of the energy barrier for the underlying reactions leading to wear, R is the universal gas constant 8.314 J · mol -1 · K -1 and T the absolute temperature in Kelvin. In this empirical evaluation of the experimental data, the wear mechanisms are considered to be governed mainly by chemical or physical reactions with specific rate constants. In this respect, reactions like desorption or adsorption of species such as water or hydrogen from the sliding surfaces as well as the transformation of the a-C: H network into graphite and/ or the oxidation of silicon nitride into SiO 2 according to equation (8) may be considered main factors to explain the experimental results observed. The destabilization of a-C: H coatings by temperature and the formation of graphite as the stable phase of carbon is a wellknown phenomenon, occurring around 500 - 600 °C under static conditions [18]. Under sliding conditions, this temperature is lowered to around 100 - 200 °C under dry sliding [6,19,] and under boundary lubricated conditions graphitization may occur around 100 °C [20]. (8) Equation (8) is only one example for multiple other possible oxidation reactions of silicon nitride [21,22,23]. Figure 7 shows the system wear volumes for both counter bodies on a logarithmic scale as a function of the reciprocal absolute temperature. Interestingly, these Arrhenius type graphs show for both systems two distinct temperature areas with different slopes. According to equation (7), the slopes m contains the activation energy E a of the respective wear processes and may be determined according to equations (9) and (10). (9) (10) This analysis reveals, therefore, the existence of a lower temperature area 1, extending from room temperature up to about 110 °C and a higher temperature area 2 which extends from about 110 to 250 °C for each sliding couple. In these areas, the slopes of the logarithm of the wear rate against the reciprocal temperature are both linear but different. This means that a. the evaluation according to equation (6) seems to be justified and b. the activation energies of the underlying chemical reactions and the subsequent wear mechanisms are different. In fact, Figure 7 points at two distinct different wear mechanisms in dependence on temperature for each counter material. This finding is in good agreement with the observed temperature dependence of the coefficients of friction as shown in Figure 3. Figure 3 reveals an increased friction coefficient up to a small temperature range of around 110 °C with decreasing friction coefficients with further increasing temperature. This temperature interval of about 10 - 15 ° around 110 °C corres- Aus Wissenschaft und Forschung 33 Tribologie + Schmierungstechnik · 65. Jahrgang · 5/ 2018 ! " # $ %& ' / () ( ( ( ( *+ ! *+" , & ' (#) ( ( ( ( Si 3 N 4 + 5 O 2 à 3 SiO 2 + 4 NO ( m= - ! " ( ( ( # $ % &' ( ) ( Figure 6: Wear volumes after 100,000 cycles sliding distance, separately shown for disk (a-C: H layer) and ball (silicon nitride, a, and alumina, b) T+S_5_18.qxp_T+S_2018 28.08.18 13: 50 Seite 33 Unfortunately, the published results cover only the temperature range up to 110 °C but show clearly the linear relationship of the logarithm of wear volume with reciprocal absolute temperature. The observed differences in activation energy might be due to the different counter material and the various coatings with much higher hardness. Aus Wissenschaft und Forschung 34 Tribologie + Schmierungstechnik · 65. Jahrgang · 5/ 2018 ponds well with the transition temperature range of wear rates at around 110 °C for both alumina and silicon nitride counter materials, see shaded area in Figure 9. Figure 7 also reveals, that there is a distinct difference in the wear rates for the two counter body materials. In the lower temperature regime, wear rates of the system with alumina counter-body are smaller compared to the one with silicon nitride counter-body. This difference, which is also indicated in Figure 5, may be addressed to the tribo-oxidational wear on silicon nitride ball, as indicated by equation (8). Figure 8 shows the logarithm of the wear rates of both counter bodies. Over the whole temperature range, only one slope is observed for alumina as well as for silicon nitride, signifying that on the counter body only one single wear mechanism is active. The activation energies for the identified temperature ranges with constant slopes are compiled in Table 2. In the lower temperature range the table reveals slightly higher activation energies for alumina counter body than for silicon nitride. This holds also true for the wear of the a-C: H coating (disk). Activation energy for the wear process is slightly higher for alumina counter body than for silicon nitride counter material. It may be also noticed that the disks show higher activation energies than the counter bodies, although the absolute values are lower. The results are in reasonably good agreement with data published by Lafon-Lafayette et al. [24], on a-C: H and ta-C coatings sliding against steel as a counter material. Figure 7: Logarithm of system wear rates as a function of reciprocal absolute temperature Figure 8: Logarithm of wear volumes of ball as a function of reciprocal absolute temperature Table 2: Activation energies of system, ball and disk in kJ · mol -1 for temperature ranges 25 - 110 and 110 - 250 °C Alumina Silicon nitride Alumina Silicon nitride 25 -110 °C 25 -110 °C 110 -250 °C 110 -250 °C System 5.1 4.3 31.8 10.8 Ball 4.9 3.5 4.9 3.5 a-C: H coated Disk 9.2 7.4 44.3 12.4 T+S_5_18.qxp_T+S_2018 28.08.18 13: 50 Seite 34 Figure 9 reveals that the different wear regimes during increasing temperature are stemming from the wear behavior of the a-C: H coating of the disk. The raising COF in Figure 3 at temperatures up to 110 °C, area 1, cannot be explained experimentally. It might be speculated, that in this lower temperature range, desorption of water or of hydrogen from the a-C: H surface plays a vital role for an increasing COF. The desorption of humidity from the sliding surfaces with increasing interactive, attractive forces between the two sliding surfaces (and consequently increasing COF) are well known in lubricated contacts [25]. To get further clarity about the basic wear mechanisms and possible formation of tribofilms, wear scars on the ball have been investigated by scanning electron microscopy (SEM) and energy dispersive x-ray spectroscopy (EDX) Figure 10 shows the SEM image of the alumina and silicon nitride balls after a test run for 100,000 cycles at 25 °C, 100 °C and 200 °C, resp. The pictures show that a tribofilm has formed at all temperatures on silicon nitride ball, but on alumina only at temperatures of 100 and 200 °C. The chemical composition of the tribofilms is shown in Figure 11 and Figure 12 by recording mappings of carbon, aluminum, oxygen and phosphorus as well as sulfur (sulfur is only shown for alumina ball due to lacking space). On the Aus Wissenschaft und Forschung 35 Tribologie + Schmierungstechnik · 65. Jahrgang · 5/ 2018 Figure 9: Logarithm of wear volumes of disk as a function of reciprocal absolute temperature Figure 10: SEM of wear scar on ball; alumina in the upper line, a, b and c, silicon nitride in the lower line, d, e and f; increasing temperatures from left to right, identical scale 50 µm Figure 11: EDX element mapping of tribofilm on alumina ball at 25 °C (left), 100 °C (middle) and 200 °C (right), identical scale 50 µm T+S_5_18.qxp_T+S_2018 28.08.18 13: 50 Seite 35 coefficient of friction is higher for silicon nitride counter body. To explain the friction behavior as observed in Figure 3 and Figure 4 in more detail, more experimental data are needed. However, the wear volume above the transition temperature of 110 °C is lower for silicon nitride, see Figure 6. It may be therefore concluded, that tribo-oxidation of the silicon nitride ball causes the main difference in friction and wear. The fact, that on alumina ball no carbon transfer film has formed, and that obvious formation of graphite is not taking place, the decrease of friction above the 110 °C threshold, see Figure 3, is most likely not attributed to graphite as a solid lubricant. 4 Conclusions In the present work, COF and wear behavior were investigated of alumina and silicon nitride balls sliding against a-C: H on Cronidur30 disk under lubricated conditions at temperatures up to 250 °C. From the results obtained, the following conclusions may be drawn: 1. The coefficient of friction increases up to a transition area of 110 °C, then decreases. 2. Arrhenius type analysis of wear coefficients reveals two distinct areas of wear with different wear regimes. The temperature ranges for rising and falling friction coefficients of friction agree well with the temperature ranges for the two different areas of wear. 3. Increasing wear above the transition temperature is mainly due to increased wear on the a-C: H coating. 4. The formation of a tribo-film is observed for both counter materials. However, in the case of alumina counter body, the tribo film consists mainly of a phosphorous compound which stems from the antioxidant additive of the oil. In case of silicon nitride counter body, the tribo-film contains oxygen and phosphorus at 25 °C, and carbon, phosphorous and oxygen at higher temperatures. Aus Wissenschaft und Forschung 36 Tribologie + Schmierungstechnik · 65. Jahrgang · 5/ 2018 alumina ball, only a tribo-film has formed at temperatures of 100 °C and 200 °C. These films on alumina ball are mainly composed of a phosphorus compound which stems from the phosphorus compound tri-cresylphoshate, which is contained in the oil as an anti-oxidant. It also seems, that the tribofilm is more concentrated on the rim of the wear scar and that its thickness is increasing with temperature. Sulfur, which is also part of the antioxidant mix in the oil, is not present in the film. It is interesting to note, that the element analysis for carbon does not show a significant carbon concentration in the wear scar on alumina ball, see Figure 11. From this experimental observation it may be concluded that a transfer of carbon from the a-C: H layer to the aluminum oxide ball has not taken place. The SEM pictures in Figure 10 show that a tribo-film has formed in the wear scar of the silicon nitride ball at all three temperatures. Figure 12 shows the composition of these films. The oxygen mapping reveals that oxidation has taken place and an oxide containing tribo-film is present, as was already expected. This supports the hypothesis of the tribo-oxidation process of silicon nitride due to tribological movement. But, as primary difference regarding the composition of the tribo-film on alumina ball, on silicon nitride ball a distinct carbon concentration is located in the wear scar, but only at temperatures of 100 and 200 °C, resp. It may be noted that temperature seems to have also a slight but distinct influence on the composition of the film. Going from 100 °C to 200 °C in oil temperature, the carbon intensity of the tribo-film is increased, whereas the phosphorous intensity is decreased at 200 °C. It also seems that the amount of oxygen is increased at the higher temperature. At 25 °C the films contain only oxygen (main) and little phosphorus, but no carbon. The difference in composition of the tribo-film formed may be the main reason for differences in friction and wear behavior. Although friction develops parallel for both counter bodies, it is obvious that in all conditions the Figure 12: EDX element mapping of tribofilm on silicon nitride ball at 25 °C (left), 100 °C (middle) and 200 °C (right), identical scale 50 µm T+S_5_18.qxp_T+S_2018 28.08.18 13: 50 Seite 36 5. Influence of counter material on friction and wear is seen to be mainly due to tribo-oxidation of silicon nitride and the differences within the formed tribo-film. Acknowledgement The authors wish to express their sincerest gratitude for the invaluable help of Mrs. C. Neumann and Mrs. C. Slachciak in creating the experimental results and Mrs. S. Benemann for carrying out SEM and EDX measurements. Conflicts of interests: The authors declare there has been no conflicts of interest. References [1] Robertson, J.: Diamond-like amorphous carbon. Mat. Sci. Eng. R 37: 129-281 (2002) [2] Donnet, C., Erdemir, A. (eds.): Tribology of Diamondlike Carbon films - fundamentals and applications. ISBN 978-0-387-30264-5, Springer Science and Business Media LLC, 2008 [3] Robertson, C.: Improving the properties of diamond-like carbon. Diamond and related materials 12, 79-84 (2003) [4] Hauert, R., Müller, U.: An overview on tailored tribological and biological behavior of diamond-like carbon. Diamond and related materials 12, 171-177 (2003) [5] Bremond, F., Foumier, P., Platon, F.: Test temperature effect on the tribological behavior of DLC coated 100 Cr 6 couples in dry friction. Wear, 254, 774-783 (2003) [6] Liu, H., Tanaka, A., Umeda, K.: The tribological characteristics of diamond-like carbon films at elevated temperatures. Thin Solid Films, 346 167-168 (1999) [7] Vanhulsel, A., Velasco, F., Jacobs, R., Eersels, L., Havermans, D., Roberts, E.W., Sherrington, I., Anderson, M.J., Gaillard, L.: DLC solid lubricant coatings on ball bearings for space applications. Tribology International, 40 (7), 1186-1194 (2007) [8] Krumpiegl, T., Meerkamm, H., Fruth, W., Schaufler, C., Erkens, G., Böhner, H.: Amorphous carbon coatings and their tribological behavior at high temperatures and in high vacuum. Surface and Coatings Technology 120- 121, 555-560 (1999) [9] Klaffke, D., Hartelt, M., Brand, J., Brand, C., Wittorf, R.: Effect of temperature on the tribological behavior of DLC coatings. Materialwissenschaften u. Werkstofftechnik 35, 936-941 (2004) [10] Ouyang, J.H., Sasaki, S.: Friction and wear characteristics of a Ti-containing diamond-like carbon coating with SRV tester at high contact load and elevated temperature. Surface and Coatings Technology 195, 234-244 (2005) [11] Ronkainen, H., Holmberg, K.: Environmental and thermal effects on the tribological performance of DLC coatings. In: Tribology of diamond-like carbon films, by C. Donnet and A. Erdemir (eds.) ISBN 978-0-387-30264- 5, Springer Science and Business Media LLC, 155-200, (2008) [12] Wäsche, R., Klaffke, D.: Tribology of DLC Films Under Fretting Conditions. In: C. Donnet and A. Erdemir (eds.) Tribology of Diamond-Like Carbon Films. ISBN 978-0- 387-30264-5, Springer Science and Business Media LLC, 362-382 (2008) [13] Waesche, R., Weihnacht, V.: Influence of counterbody material on wear of ta-C coatings under fretting conditions at elevated temperatures. Wear 267, 12, 2208-2215 (2009) [14] Gomes, J.R., Camargo Jr., S.S., Sim-o, R.A., Carrapichano, J.M., Achete, C.A., Silv, R.F.: Tribological properties of silicon nitride ceramics coated with DLC and DLC-Si against 316L stainless steel. Vacuum 81, 1448- 1452 (2007) [15] Klaffke, D.: Towards a tribological reference test - fretting test? Tribotest 12 133-147 (2006) [16] Peng, D.-X., Chen, C.-H., Kuang, Y., Chang, Y.-P., Chang, S.-Y.: Size effects of SiO2 nanoparticles as oil additives on tribology of lubricant. Industrial Lubrication and Tribology 62, 111-120(2010) [17] Laidler, K.J.: Chemical kinetics, 3rd edition 1987, Harper and Row, New Yorck, ISBN 0-06-043862-2, chapter 2.9, 39-53 [18] Fink, J., Muller-Heinzerling, T., Pfluger, J., Bubenzer, A., Koidl, P., Crecelius, G.: Structure and bonding of hydrocarbon plasma generated carbon films: an electron loss study. Solid State Communications 47, 687- 691(1983) [19] Liu, Y., Meletis, E.I.: Evidence of graphitization of diamond-like carbon films during sliding wear. Journal of Material Science 32, 3491-3495(1997) [20] Mobarak, H.M., Masjuki, H.H., Niza Mohamad, E., Kalam, M.A., Rashedul, H.K., Rashed, M.M., Habibullah, M.: Tribological properties of amorphous hydrogenated (a-C: H) and hydrogen free tetrahedral (ta-C) diamondlike carbon coatings under jatropha biodegradable lubricating oil at different temperatures. Applied Surface Science 317, 581-592(2014) [21] Butt, D.P., Albert, D., Taylor, T.N.: Kinetics of thermal oxidation of silicon nitride powders. J.Am.Ceram.Soc. 79, 2809-2814 (1996) [22] Stern, K.H.: Oxidation of silicon, silicon carbide (SiC) and silicon nitride (Si3N4). National Research Laboratory NRL Memorandum Report 5774, June 4, 1986. [23] Enomoto, T., Ando, R., Morita, H., Nakayama, H.: Thermal oxidation rate of a Si3N4 film and its masking effect against oxidation of silicon. Japanese Journal of Applied Physics 17[6] 1049-1058 (1978) [24] Lafon-Playette,S., Fontaine, J., De Barros Bouchet, M.- I., Heau, C.: Critical role of a metallic counterpart on the tribochemical wear of ta-C coatings in base oil. Wear 402-403, 91-99(2018) [25] Matveevskij, R.M., Schulze, D.K.-W., Bujanowski, I. A.: Zusammenwirkung zwischen der thermischen Stabilität ep-wirksamer Modelsubstanzen und ihren tribochemischen Eigenschaften bei Reibung auf der KT-2 Maschine. Schmierungstechnik 11, 78-81 (1980) (in german) Aus Wissenschaft und Forschung 37 Tribologie + Schmierungstechnik · 65. Jahrgang · 5/ 2018 T+S_5_18.qxp_T+S_2018 28.08.18 13: 50 Seite 37