Self-lubricating triboactive (Cr,Al)N+Mo:S coatings for fluid-free applications

Within this study, self-lubricating and triboactive (Cr,Al)N+Mo:S coatings were developed and investigated for the deposition on components in a low-temperature physical vapor deposition (PVD) hybrid process. Therefore, direct current magnetron sputtering (dcMS) and high power pulse magnetron sputtering (HPPMS) PVD were combined by using an industrial coating machine. Hereby, it was possible to deposit dense and smooth triboactive, self-lubricating nitride coatings with different chemical compositions and architectures on 16MnCr5E samples. Two coating architectures, a matrix monolayer and a graded coating structure, were developed to evaluate the effect on the tribological behavior. The morphology and coating thickness were analyzed by means of scanning electron microscopy (SEM). Furthermore, the indentation hardness and modulus of indentation as well as the compound adhesion between substrate materials and coating were analyzed. Tribological analyses of (Cr,Al)N+Mo:S-coated and uncoated samples were conducted under fluid-free friction regime at room temperature T = (20 ± 3) °C, a velocity v = 0.1 m/s and a distance s = 1000 m by varying the Hertzian contact pressure from 400 MPa ≤ pH ≤ 1300 MPa against steel counterparts, 100Cr6, in a pin-on-disk (PoD) tribometer. The graded coating architecture of (Cr,Al)N+Mo:S enabled a significant wear and friction reduction. Furthermore, Raman analyses prove the formation of solid lubrication tribofilm containing MoS2, MoO3 MoO2 and MoxOy at the toplayer of a graded (Cr,Al)N+Mo:S coating, which are responsible for the improved tribological behavior.


Introduction
Today, gearboxes are conventionally lubricated with greases or oils. However, mineral oil or synthetic oilbased lubricants are often not applicable in medical technology, food technology, production under vacuum processes technology and aerospace applications due to extreme ambient conditions. Moreover, lubrication requires a high level of machine component design due to the use of complex sealing systems, e.g., in gearboxes. In the absence of conventional lubricants, friction increases, which leads to a high thermal load and thus to rising of mechanical stresses and accelerated failure of components [1,2]. Spur gears that are exposed to high mechanical loads without conventional grease or oil lubrication show considerable damage after a short operating time due to the enormous heat input caused by friction [3,4]. Typical damage patterns on gears are scuffing and hot running. A friction reduction can be achieved by using solid lubricants like graphite or molybdenum disulfide (MoS 2 ). However, these solid lubricants are rarely developed for high mechanical loads.
A possible solution could be the fluid-free lubrication with solid lubricants such as transition metal dichalcogenides (TMD) like molybdenum disulfide (MoS 2 ) which are widely used in the industry. The crystallographic hexagonal structure of MoS 2 is bonded by weak van der Waals forces between the planes. Here, a sliding on planes against each other can be initiated through low shear stresses and leading to a friction reduction in tribological systems. A possible application of MoS 2 coatings on components can be conducted by means of MoS x physical vapor deposition. Despite the friction reduction in tribological contacts, high wear rates can lead to fast coating failures depending on the structure and chemical composition under high loads [5]. For this reason, different coating architecture concepts to improve the tribological behavior of Mo-and S-containing coatings are pursued in the literature. One coating concept follows the approach of adding different metals Al, Cr, Ti to a MoS 2 monolayer [6][7][8]. Furthermore, hard soft combinations with a hard nitride-based coating and a soft MoS 2 toplayer are also a possible solution to improve the tribological behavior of MoS 2 containing PVD coatings [8]. However, the disadvantage of these coating architecture concepts is that the low wear resistant solid lubricant MoS 2 toplayer also forms the lubricant depot at the same time, while the hard support layer below only serves as wear protection.
An alternative approach pursues the incorporation of disulfide solid lubricants into a hard coating matrix, which serves to increase the wear resistance and as a depot for solid lubricant storage. The nitride hard coatings CrN, TiN and (Cr,Al)N are well known as wear protection coatings on cutting tools in industrial applications [9][10][11][12]. In general, the use of nitride-based coatings in dry-running tribological contacts leads to a wear reduction for the coated contact partner, whereby wear rises for uncoated counterparts. Hereby also the friction in the tribological contacts with uncoated components increases [13]. Investigations of TiN, CrN, and (Cr,Al)N as a matrix material for the incorporation of Mo and S were carried out by [14][15][16]. Here, the negative effects regarding wear could be reduced and a friction reduction could be achieved. The chemical analyses on the modification elements Mo and S in the (Cr,Al) N and TiN layer systems have shown that MoS 2 is not necessarily stoichiometrically present in the matrix [14,16].
Within the scope of this study, the goal was to investigate the effect of two different coating architectures on friction reduction by the solid lubricant MoS 2 and increased wear resistance through the use of (Cr,Al)N. For this purpose, the two coating architecture concepts (i) matrix and (ii) graded (Cr,Al)N coating with a nitride-based toplayer and embedded Mo and S were developed and analyzed. The intended application of the developed coatings is highly loaded spur gears running under dry conditions. Therefore, the case-hardened steel 16MnCr5E, which is conventionally used for gears, was chosen as substrate material, which requires a low-temperature coating PVD process due to the low annealing temperature T≤300°C. This can be achieved by means of a hybrid process of direct current magnetron sputtering (dcMS) and high power pulse magnetron sputtering (HPPMS) coating process [17]. Chemical analysis, analysis of the microstructure, indentation hardness and modulus of indentation enable a basic understanding of the coatings and there architecture. In addition, tribological analyses at different Hertzian contact pressure from 400 MPa≤p H ≤1300 MPa under fluid-free friction regime at room temperature T=(20 ±3)°C were conducted. After tribological testing, Raman analyses were conducted to prove the supply mechanism of the graded coating and transfer mechanism to the 100Cr6 counterpart. Particularly in the field of highly loaded tribological contacts there is substantial need for research regarding the design of dry-running tribological systems and their supply mechanisms of solid lubricants.

Experimental details Coating deposition
The molybdenum Mo-and sulfur S-containing (Cr, Al)N?Mo:S coatings were deposited in an industrial scale coating machine CC800/9 HPPMS, CemeCon AG, Wü rselen, Germany, by developing a low-temperature coating process T≤300°C. The coating unit was equipped with four dcMS cathodes and two HPPMS cathodes. Both technologies were combined in a dcMS/HPPMS process, due to the beneficial advantages of high ionization regarding improved mechanical properties, low roughness, dense microstructure, coating of complex geometries and enhanced compound adhesion between substrate and coating [18][19][20][21]. The arrangement of the cathodes and the layout of the deposition chamber are depicted in Fig. 1a.
For the deposition of the (Cr,Al)N?Mo:S coatings, Al plugged Cr targets (purity Cr 99.95%, purity Al 99.5%), a Cr plugged Al target (purity Al 99.5%, purity Cr 99.5%), and a MoS 2 target (purity 99.5%) were used. The notations CrAl20 and AlCr30 refer to a Cr target with 20 Al plugs or an Al target with 30 Cr plugs respectively. The dimensions of the targets were 500 mm988 mm, and the diameter of the Al or Cr plugs was approximately Ø ≈ 15 mm, Fig. 1b. The process parameters for the deposition of the reference coating (Cr,Al)N are presented in [22]. In this study, two different coating architectures, a monolayer matrix (M) and a graded (G) (Cr,Al)N?Mo:S structure, Fig. 2, were developed and compared with the reference coating (Cr,Al)N. The idea of the monolayer matrix coating was to produce a coating with a constant depot of embedded Mo and S depots from the intermediate layer to the toplayer. This should enable the in situ formation of the solid lubricant MoS 2 during the tribological contact, Fig. 2a.
An alternative coating concept is described by the graded coating architecture, whereby besides the changes of the chemistry at the monolayer matrix coating, a differing interlayer design and chemical composition of the toplayer are developed, Fig. 2b. Hereby, the ratio of Mo and S is increased toward the toplayer but still embedded in a nitride matrix. Therefore, three different coatings of each architecture were deposited and analyzed.
The coatings were deposited on case-hardened 16MnCr5E steel samples (Ø=25 mm, H=8 mm) polished to a roughness Ra=0.02 µm.  parameters of low-temperature coating process of (Cr,Al)N published in [22] and the target configuration of (Cr,Al)N?Mo:S coating for tools given in [14]. Moreover, the in situ temperature measurement system developed at IOT presented in [23] for determining the substrate temperature depending on cathode powers during the coating process was used to avoid an excessive overshooting of the annealing temperature T≤200°C of 16MnCr5E. The target configurations for the deposition processes of the coatings are presented in Table 1. In order to increase the Mo:S content at the toplayer for M3, two MoS 2 targets were chosen. This concept was adopted for the graded coatings. Exchanging a CrAl20 with an AlCr30 target could lead to changes of the coating structure, which could be beneficial for the tribological behavior. The process parameters and bold changes of the developed matrix M1-M3 and graded G1-G3 (Cr,Al)N?Mo:S coatings for the deposition of interlayer and toplayer are shown in Tables 2 and 3.

Coating and compound analysis
The coating thicknesses and morphology were investigated by scanning electron microscopy (SEM) analyses, Oxford Link ISIS, Oxford Instruments plc, UK, at the Central Facility for Electron Microscopy (GfE), RWTH Aachen University, Aachen, Germany. The chemical composition over the coating thickness was analyzed by means of X-ray microanalysis (EPMA) on the basis of a calotte. The calotte was generated by rotating a hardened steel ball combined with diamond suspension for t=60 s on the coating. Phase analyses of the developed (Cr,Al)N?Mo:S coatings were conducted by X-ray diffraction (XRD) measurements using grazing incidence diffraction (GID), XRD 3003, GE Energy Germany GmbH,

Tribological analysis
The samples were tested under pure sliding conditions in a PoD tribometer, CSM Instruments, Peseux, Switzerland. Tribological tests were conducted at fluid-free friction regime in a steel/steel and coating/ steel contact at different initial Hertzian pressure p H = 400 MPa, p H =600 MPa, p H =800 MPa and p H = 1300 MPa calculated based on basic part and counterpart properties. Therefore, the normal force was adjusted to F N =0.5 N for p H =400 MPa and F N =2 N for p H =600 MPa by using an uncoated Ø=10 mm diameter 100Cr6 ball. With regard to the higher initial Hertzian loads, uncoated 100Cr6 balls with a diameter of Ø=6 mm were chosen at a normal force of F N =2 N for p H =800 MPa and F N =8 N for p H =1300 MPa. It has to be stated that all considered Hertzian pressures can be regarded as an initial value, which will change during the tribological contact due to wear of basic part and counterpart. The chosen initial Hertzian pressure enables a mapping of different applications like in equipment in food or pharmaceutical industries and optical devices. Gears are conventionally running under sliding and rolling conditions during the tribological contact. However, the PoD testing under pure sliding conditions causes high load inputs to the compound of coating and 16MnCr5E, so that an assessment can be obtained with regard to the application of the developed coatings on gears. The parameters relative velocity v rel =0.1 m/s, radius r=2.5 mm, distance s=1000 m room temperature T=RT ≙ T=(20±3)°C and relative humidity φ ≈ (45±5) % were kept constant. In addition, a repetition of each tribological test was carried out. The average coefficients of friction (CoF) were evaluated for the final s=500 m running distance. Topography and wear analysis of the wear tracks on the basic parts and wear areas on the counterparts were conducted on one of the two samples after tribological testing by means of CLSM.

Chemical analysis of tribologically tested samples by Raman spectroscopy
For chemical analysis after tribological testing, the Raman spectrometer, Renishaw InVia Reflex, Renishaw GmbH, Pliezhausen, Germany, with a λ= 532 nm laser with a spot size of d ≈ 1 µm and a diffraction grating g(λ=532 nm)=1800 l/mm, was used. Here, a graded (Cr,Al)N?Mo:S-coated basic part and a 100Cr6 counterpart were exemplarily analyzed regarding the tribofilm formation. The measuring parameters were adjusted to the different samples. The laser was calibrated before the measurement by using a silicon reference sample.

Results and discussion
Temperature measurement of coating process parameters The in situ temperature measurement system proves that high HPPMS cathode power up to P=5000 W has a neglectable effect on the substrate temperature of the 16MnCr5E samples. In contrast thereto, the addition of four parallel running dc cathodes to the HPPMS cathodes limits the power of the dc cathodes to P=800-1000 W. These results enabled the development of a low-temperature coating process for the deposition of the different coating architectures (i) matrix monolayer and (ii) graded. In Fig. 3, the measured substrate temperature during the coating process of G3 divided in the four main sections is shown. It can be observed that during the coating deposition the annealing temperature of 16MnCr5E 150°C≤ T≤200°C is exceeded to T=264°C. However, the effect on the loss of hardness is negligible. The data sheet in [26] shows that the hardness loss at T=300°C is around 1 HRC≤H≤3HRC. These values are within the range of deviation for the case hardening process of H=(60±2) HRC. Also Rockwell indentation tests of steel samples after the deposition process did not show any changes of the hardness. Nevertheless, the process was developed close to the limits of the acceptable temperature load of 16MnCr5E during the coating deposition.

Matrix coating development
In a first step, self-lubricating (Cr,Al)N?Mo:S monolayer matrix coatings were developed. Figure 4 shows the cross-sectional micrographs of the three developed (Cr,Al)N?Mo:S monolayer matrix coatings, as well as the reference coating (Cr,Al)N.
It can be observed that the matrix coatings and the (Cr,Al)N coating possess a dense and fine columnar structure. Due to changes of bias voltage U B , HPPMS cathode power P HPPMS and nitrogen pressure p N2 in the coating processes M1-M3 highlighted in Table 2, the coating thicknesses are ranging between 1.3 µm≤ t≤2.1 µm. In addition, a smooth surface structure with only little defects at the toplayer can be found for all coatings, which can be explained by the use of the HPPMS technology. The depth-resolved chemical compositions of the coatings by means of EPMA are shown in Fig. 5.
For further discussion, only the toplayer compositions of the coatings will be considered. The addition of a second MoS 2 target in favor of a CrAl20 target in the coating process M3 led to a small increase in the Mo and S content in comparison with M2 and M1. The highest Al content of all developed coatings can be proven for the reference (Cr,Al)N. The quantitative values of the chemical composition of the toplayers of the matrix monolayer coatings and the reference coating are shown in Table 4. The crystallographic phase analyses of the developed (Cr,Al)N?Mo:S matrix coatings and (Cr,Al)N are presented in the XRD patterns in Fig. 6.
All patterns reveal a cubic CrN phase with corresponding peaks at 2θ ≈ 37.5°, 2θ ≈ 43.6°, 2θ ≈ 63.4°a nd 2θ ≈ 76.0°and a cubic AlN phase, 2θ ≈ 37.8°, 2θ ≈ 43.9°, 2θ ≈ 63.8°and 2θ ≈ 76.7°. Moreover, a cubic structure was also measured for molybdenum nitride (Mo 2 N) phase, 2θ ≈ 37.4°, 2θ ≈ 43.5°and 2θ ≈ 63.1°l eading to the conclusion that an additional distortion of the (Cr,Al)N lattice is generated. The peaks of M1 coating appear to be significantly lower in comparison with the M2 und M3 coating, which can be explained through an X-ray amorphous or nanocrystalline structure. The XRD spectra cannot prove the formation of MoS 2 . The reason are the low contents of Mo and S for the M1, M2 and M3 coatings. Nevertheless, it can be assumed that Mo and S are incorporated separately in the coating matrix. The crystalline structure of (Cr,Al)N can be proven regarding the cubic CrN and AlN peaks presented in Fig. 6. Furthermore, the indentation hardness H IT and the modulus of indentation E IT of the (Cr,Al)N?Mo:S matrix coatings and the reference coating (Cr,Al)N were determined by nanoindentation, Fig. 7.
The   step, the compound adhesion between the (Cr,Al)N? Mo:S matrix coatings, (Cr,Al)N and 16MnCr5E was measured by means of scratch test. The critical loads Lc 1-3 of the compound M1/16MnCr5E were improved by changing process parameters during the interlayer deposition, Table 2. In [27,28] it is shown that the parameters HPPMS cathode power, process pressure and bias voltage have an impact on the compound adhesion between coating and substrate material. Hereby, the compound adhesion between matrix coatings and 16MnCr5E could be improved from Lc 3 =45 N for M1 to Lc 3 =60 N for M2 or M3, Fig. 8. The comparison of the compound systems M2/16MnCr5E and M3/16MnCr5E shows equivalent critical loads but a different plastic deformation behavior. One possible explanation could be the differences in the chemical composition of M2 and M3. The cracks at Lc 3 of M3/16MnCr5E appear to be thinner and narrower and to be less critical than on M2/16MnCr5E. A differing plastic deformation behavior can also be found for the binary system (Cr, Al)N, which shows the highest critical load at Lc 3 = 70 N without huge cracks at the edge region of the scratch.
In the next step, tribological analyses were conducted for the matrix coatings, which lead to minimal differences in the CoF, whereby M3 was chosen as representative coating shown in Fig. 9. The decision was based on the comparable coating properties to (Cr,Al)N regarding thickness t, increased Mo content compared to M1 and M2, indentation hardness H IT , modulus of indentation E IT and compound adhesion Lc.
The tribological behavior in a dry-running coating/ steel and steel/steel contact was tested at a representative initial Hertzian pressure of p H = 800 MPa, Fig. 9. In order to analyze the effect of the added Mo and S in the toplayer of the matrix coatings on the CoF, the tribological systems (Cr,Al)N/steel and steel/steel are presented for comparison. The results show a slight friction reduction through the use of (Cr,Al)N?Mo:S regarding the higher CoF of the (Cr,Al)N/steel system. Moreover, the curve progression of the CoF is smooth in comparison with the CoF curve of the steel/steel contact and the (Cr,Al) N/steel contact. However, the CoF of the steel/steel contact still leads to the lowest CoF compared to tribological systems with (Cr,Al)N and (Cr,Al)N? Mo:S. It is believed that the Mo and S ratio in the matrix coating is not high enough for building the required quantity of the solid lubricant MoS 2 in the tribological contact, which would lead to a lowering of the CoF beneath the values of a steel/steel contact. It has to be stated that at this point wear was not analyzed, due to the primary goal of friction reduction in dry-running highly loaded tribological contacts.

Development of a graded (Cr,Al)N1Mo:S coating architecture
Based on the presented results, the coating architecture was varied from a monolayer matrix coating to a graded (Cr,Al)N?Mo:S coating structure. The coating processes were developed based on the promising coating and compound properties of M3/steel. The changes at the toplayer can be found in Table 3. The SEM images of the three graded coatings developed by means of dcMS/HPPMS process are presented in Fig. 10. The coatings show a columnar morphology and a thickness in a range of 4.0 µm≤t≤5.2 µm. Furthermore, a t=0.8 µm thick amorphous or nanocrystalline toplayer can be seen for the G3 coating. This was achieved by changing the process parameters during the toplayer deposition in comparison with the coating processes G1 und G2, Table 3. In addition, defects on the toplayer topography at G1 and G2 could also be reduced with regard to the coating G3.  The chemical compositions of the graded coatings are shown in Fig. 11. Due to changed process parameters compared to the matrix coatings, the Mo and S content is increased within the graded structure. For G2 and G1 an almost stoichiometric ratio of Mo and S can be observed at the toplayer. G3 shows a slight reduction of the Mo and S content in favor of a higher nitrogen, chrome and aluminum content. This is due to an increase in the nitrogen flow and the additional use of the two HPPMS cathodes with CrAl20 targets while the deposition of the toplayer compared to the deposition processes of G1 and G2. The approach was to strengthen the CrAlN nitride matrix in the toplayer, which contains Mo and S. The quantitative values of the chemical composition of the toplayers of the graded (Cr,Al)N?Mo:S coatings are shown in Table 5.
It is believed that the exchange of a CrAl20 target to an AlCr30 target as well as the adjustments of process parameters led to the formation of a hexagonal AlN phase, which could favor the formation of hexagonal Mo a S b . It is assumed that MoS 2 grows partially epitaxially on the AlN and therefore the hexagonal structure is preferred for G1 and G2. The formation of a cubic phase for CrN and AlN as it was proven for M1-M3 coatings cannot be matched to the coatings G1-G3. Thereby, it has to be taken into account that the XRD spectra are dominated by the structure of the toplayer. The formation of a cubic phase with a transition into a hexagonal phase could be possible at the intermediate layer. This can be concluded because the interlayer coating process part is based on the matrix coating process of M3. In case of G3, further analyses by means of transmission electron microscopy (TEM) have to be conducted, in order to prove, if the MoS 2 toplayer possesses an amorphous or nanocrystalline structure. The evaluation of the indentation hardness and the modulus of indentation of the graded (Cr,Al)N?Mo:S coatings, Fig. 13, shows an increase in the indentation hardness of G1 H IT = (1.4±0.2) GPa to G2 H IT =(2.9±0.2) GPa, which can be related to the exchange of a plugged CrAl20 target to a plugged AlCr30 target.  An explanation could be the slight increase in the alumina content in the coating leading to a higher AlN content. Nevertheless, the measurements are influenced by the soft toplayers of G1-G3 leading to low indentation hardnesses und modulus of indentation. A more significant increase can be measured for the modulus of indentation for G1 at E IT =(62.2± 3.6) GPa and G2 at E IT =(83.2±2.5) GPa, which could be related to changes of the target configuration. Further process parameter changes result in an indentation hardness of G3 of H IT =(7.4±1.0) GPa and a modulus of indentation of E IT =(101.4±11.3) GPa. A reason is the changed nitrogen gas flow Q(N 2 ) and the connected argon pressure p Ar adjustment, Table 3. Furthermore, the Mo:S-containing runningin layer of the coating G3 was stabilized by additional use of the two HPPMS cathodes at the toplayer deposition. Analysis of the compound adhesion between the graded (Cr,Al)N?Mo:S coatings and 16MnCr5E is shown in Fig. 14. The comparison of the critical loads between the coating compounds G1/ 16MnCr5E, G2/16MnCr5E at Lc 1 ≈ 20 N−30 N and G3/16MnCr5E at Lc 1 ≈ 15 N shows a reduction of the compound adhesion of the latter compound. An explanation cannot be given yet.
In sum, the coating delamination caused by the scratch tip appears to be lowered for the compound G3/16MnCr5E compared to the remaining two developed graded coating/steel compounds, due to less delamination at the edge region of the scratch. Even though the compound G3/16MnCr5E shows low values for Lc 1, the maximum critical load Lc 3 is in the same order as for the compounds G1/16MnCr5E and G2/16MnCr5E. Due to the promising results of M3 and G3 and the reference (Cr,Al)N, these coatings were chosen for tribological analysis at p H =400 MPa, p H =600 MPa, p H =800 MPa and p H =1300 MPa under dry-running in a coating/steel contact compared with a steel/steel contact. Tribological analysis Figure 15 shows the mean values of the CoF as a function of the initial Hertzian pressures for four different contacts under dry-running conditions. It can be seen that the CoF is reduced by increasing the Hertzian pressure from p H =400 to p H =1300 MPa. An explanation regarding the steel/steel contacts can be given through the possible increased formation of friction reducing iron oxides like Fe 3 O 4 as proposed in [29] which is also discussed in the model of Quinn [30]. At an initial Hertzian pressure of p H = 400 MPa and p H =600 MPa, the CoF of the systems (Cr,Al)N/steel and M3/steel is situated in the same area. With regard to the higher initial Hertzian pressures at p H =800 MPa and p H =1300 MPa, the CoF of the tribological system M3/steel does not show a significant difference to the CoF of (Cr,Al)N/steel. However, in case of M3/steel a formation of the solid lubricant MoS 2 during the tribological contact due to higher flash temperatures and wear is possible. Thereby, an opening of the Mo and S reservoirs at the toplayer of the coating is enabled. The lowest CoF was measured by the use of G3 for all considered tribological loads. In addition, the CoF further reduced from 0.52≤ CoF≤0.18 by increasing load conditions, leading to the possible conclusion that more MoS 2 is built in the tribological G3/steel contact. Moreover, the additional formation of Mo x O y in the tribofilm is possible. The differences between the four considered contacts steel/steel, (Cr,Al)N/steel, M3/steel and G3/steel can also be observed in the curve progressions as shown in Fig. 16. Herein, the results are exemplarily shown for an initial Hertzian pressure p H =1300 MPa. With regard to the curves of the contacts between (Cr, Al)N/steel, M3/steel it can be observed that the small shares of Mo ≈ 7 at.% and S ≈ 3 at.% at the toplayer of M3 has no friction reducing effect. The measured CoF for a steel/steel contact is significantly lower compared to the use of (Cr,Al)N and M3, which can be a result of formed oxides as explained above. A stable curve progression at a low CoF value was measured for the G3/steel contact.
The shares of Mo ≈ 28 at.% and S ≈ 31 at.% at the toplayer of G3 appears to be the decisive factor in case of friction reduction. In addition, a stable curve progression can be observed as well. After tribological testing, the 100Cr6 counterparts as well as the wear tracks were analyzed by means of CLSM, Fig. 17. All counterparts show abrasive wear scars regarding the grooves on the topography, Fig. 17b. A difference can be observed in the diameter of the wear areas of the counterparts, which is increased in the steel/steel contact as well as for the (Cr,Al)N/ steel and M3/steel contact compared to the wear area of the counterpart in the G3/steel contact, Fig. 17a. Herein, the deposition of a thick layer at the edge region of the circle area for the tribological system G3/steel can be observed. It is assumed that a Mo a-S b ?Mo x O y layer is formed during the tribological contact, which is then transferred from the G3 coating to the counterpart, Fig. 17a.
In comparison of the width of the wear tracks, Fig. 17b, a reduction for the Mo-and S-containing coatings can be seen. All tracks show abrasive wear marks, which correspond to the grooves on the counterparts. With regard to the wear track of the steel/steel contact, adhesive wear scars cannot be excluded. Moreover, the topography of the wear track on G3 shows black clustered areas, which are not identified to be a result of adhesive wear. Indeed due to the low friction results, the images of the counterparts and the presence of MoS 2 in the toplayer of the as-deposited coatings, it is believed that MoS 2 is transferred in the grooves of the wear track. Furthermore, the formation of Mo x O y is possible due to the presence of oxygen in the fluid-free tribological contact. Figure 18 presents the wear volume of the counterparts W VCP in dependence of different loading conditions. In contrast to the CoF, wear is increased by rising Hertzian pressure from p H =400 to p H =1300 MPa, which can be explained by an enlargement of abrasive wear in the steel/steel contact. Due to the increasing initial Hertzian load, a local micro-welding might occur more frequently leading to adhesive wear at the 100Cr6 ball in steel/ steel contacts. A significant wear reduction was measured for the tribological contact with the (Cr,Al) N or matrix coating in comparison with the steel/ steel contact, Fig. 18.
As a result of the addition of Mo and S at the toplayer of M3 in contrast to (Cr,Al)N, a further reduction of the wear of the counterpart is measured. With regard to the small differences between (Cr,Al) N/steel and M3/steel regarding the CoF, the wear reduction cannot be explained by the formation of MoS 2 but probably due to formation of molybdenum oxide Mo x O y as in [31]. In case of the graded (Cr,Al) N?Mo:S coating, the wear of the counterparts appears to be the lowest, which can be attributed to the possible transfer of MoS 2 as well as the formation of Mo x O y . The wear volume of the wear tracks after tribological testing is shown in Fig. 19. The highest wear volume can be observed for steel/steel contacts for all considered Hertzian pressures p H . A significant reduction of wear volume from p H =600 MPa to p H =800 MPa can be observed.
Here, it is possible that at p H =800 MPa, the tribological steel/steel system enables the formation of a Fe c O d /Cr e O f tribofilm, which works in an ideal state. It is assumed that in this ideal state the tribological parameters lead to a continuous formation and wear of the oxidic tribofilm in a balanced ratio. At higher or lower Hertzian pressures, this equilibrium shifts to the disadvantage between the formation of a protective oxidic tribofilm and wear. Moreover, the wear of the (Cr,Al)N and M3 coated samples was not measurable on the samples due to neglectable changes irrespective of the mechanical loads. In contrast thereto, the wear volume of the G3 coated basic parts slightly increases from p H =400 to p H =1300 MPa. Nevertheless, the wear volume of G3 increases from p H =400 MPa till p H =800 MPa but stays at the same level even at a higher load of p H =1300 MPa. Furthermore, the wear volume is significantly reduced in comparison with the basic parts in the steel/steel contacts. The results prove again that hardness and wear resistance are not in any linear relationship to each other regarding the presented mechanical properties of M3 and G3 coating in correlation with the wear reduction. Therefore, also the chemical composition and the formation of tribo-induced layers have to be taken into account.

Chemical analysis of the wear tack by Raman spectroscopy
Finally, G3 and 100Cr6 were chosen for chemical analysis by Raman spectroscopy due to the low CoF and wear of the counterpart during tribological testing. The wear track of a G3 basic part and a 100Cr6 counterpart after tribological testing at p H =400 MPa, T=20°C, s=1000 m and v=0.1 m/s were exemplarily analyzed, Figs. 20a and 21a. In addition, a reference spectrum was recorded in each case, Figs. 20b and 21b. Due to curve fittings of the Raman spectra, the positions of the individual peaks could be identified as shown in Figs. 20, 21, Tables 6, 7 and 8.
The chemical analysis of the as-deposited G3 basic part shows the formation of MoO 2 and MoO 3 , Table 6, which can be a result of the laser power applied on the surface. This leads to the assumption that molybdenum is available in an unbounded metallic form at the toplayer of G3. The peaks at ῦ= 361 cm −1 and ῦ=407 cm −1 can be assigned to MoS 2 .
The first peak of MoS 2 appears to be hidden beneath the MoO 2 peak at ῦ=335 cm −1 . Moreover, amorphous carbon was proven on the as-deposited surface, which can be a result of an alcohol cleaning or a contamination during the handling of the samples. The analysis inside the wear track proves the formation of iron oxide like hematite α-Fe 2 O 3 , magnetite Fe 3 O 4 , maghemite γ-Fe 2 O 3 as well as molybdenum oxides MoO 2 , MoO 3 , Mo x O y and MoS 2 . The appearance of amorphous carbon in the wear track was predominantly excluded due to the high shares of iron in contrast to carbon of the uncoated 100Cr6 counterpart. However, the assignment for the peak was listed in the area of 1100 cm −1 ≤ῦ≤1600 cm −1 . Based on the intensities in the Raman spectra, the shares of molybdenum oxides is significantly higher in comparison with MoS 2 , leading to the assumption that the friction and wear reduction is mainly determined by the formation of the molybdenum oxides. The iron oxides are a result of the abrasion of the 100Cr6 counterpart. Due to the low shares of iron oxides within the Raman spectra, it is assumed that they have a subordinated effect on friction and wear.
Based on these results, a 100Cr6 counterpart was analyzed to get an impression of the transfer layer during the tribological contact, Fig. 21  and Mo x O y and transfer of the tribofilm from the basic part to the 100Cr6 counterpart during the tribological contact, Table 8. Due to the tribofilm transfer from the basic part to the counterpart, a friction reduction can be observed in the tribological analysis. The analysis of the untested 100Cr6 counterpart proves that the formation of iron oxides is a result of the tribological contact.
In order to prove the exact chemical composition, further chemical analysis by Raman spectroscopy and X-ray photoelectron spectroscopy (XPS) will be needed. Based on these results and further optical spectrometry, an analysis of the supply mechanism of the graded (Cr,Al)N?Mo:S coating and the transfer mechanism of the tribofilm is possible. Here, a deeper understanding of the tribological contact can be achieved.

Summary and conclusion
In this study the process and material development of self-lubricating triboactive (Cr,Al)N?Mo:S coatings for fluid-free applications were shown. Matrix   and graded coating architectures were analyzed regarding morphology, chemical composition, crystallographic structure and indentation hardness and modulus of indentation as well as the compound adhesion between coating and steel. In the next step, tribological tests at different loads ranging from an initial Hertzian pressure p H =400 MPa to p H = 1300 MPa were conducted. The key results can be summarized as follows: In the next step, chemical analyses by means of Raman spectroscopy on the wear tracks and on the counterparts will be proceeded to get a deeper view on the supply mechanism of self-lubricating triboactive (Cr,Al)N?Mo:S coatings. The results can be supported by SEM analyses of the wear tracks and counterparts of coated and uncoated samples after tribological testing. Furthermore, photoemission spectroscopy (XPS) can help to prove the results by Raman spectroscopy and reveal the binding character of the tribo-induced layers.