Comparative tribological behaviors of TiN, CrN and MoNCu nanocomposite coatings

Comparative tribological behaviors of TiN, CrN and MoNCu nanocomposite coatings

ARTICLE IN PRESS Tribology International 41 (2008) 49–59 www.elsevier.com/locate/triboint Comparative tribological behaviors of TiN–, CrN– and MoN–C...

2MB Sizes 0 Downloads 13 Views

ARTICLE IN PRESS

Tribology International 41 (2008) 49–59 www.elsevier.com/locate/triboint

Comparative tribological behaviors of TiN–, CrN– and MoN–Cu nanocomposite coatings A. O¨ztu¨rka, K.V. Ezirmika, K. Kazmanlıa, M. U¨rgena,, O.L. Eryılmazb, A. Erdemirb a

Department of Metallurgical and Materials Engineering, Istanbul Technical University, Maslak, Istanbul, 34469, Turkey b Argonne National Laboratory, Energy Technology Division, Tribology Section, IL, USA Received 24 March 2006; accepted 20 April 2007 Available online 21 June 2007

Abstract The purpose of this study is to investigate comparative tribological behaviors of Cu-doped TiN, CrN, and MoN coatings under a wide range of dry sliding conditions. TiN and CrN coatings have been developed and used by industry in numerous tribological applications including, machining, manufacturing and transportation. In contrast, MoN has attracted very little attention as a tribological coating in the past, despite being much harder than both TiN and CrN. In this paper, we will mainly concentrate on the Cu-doped versions of these coatings whose tribological properties have not yet been fully explored. The results of this study have confirmed that the addition of Cu into TiN, CrN and MoN coatings has indeed modified the grain size and morphology, but had a beneficial effect only on the friction and wear behavior of MoN. The tribological behavior of CrN did not change much with the addition of Cu but that of TiN became worse after Cu additions. Raman spectroscopy technique was used to elucidate the structural and chemical natures of the oxide films forming on sliding surfaces of Cu-doped TiN, CrN and MoN films. The differences in the friction and wear behavior of Cu-doped TiN, CrN, and MoN is fully considered and a mechanistic explanation has been provided using the principles of a crystal chemical model that can relate the lubricity of complex oxides to their ionic potentials. r 2007 Published by Elsevier Ltd. Keywords: Friction and wear; Crystal chemistry; Nanocomposite coatings; Hydrid PVD techniques; Raman spectroscopy

1. Introduction Thin hard and soft coatings are used extensively by industry to combat friction and wear in a wide range of machine tools and moving mechanical assemblies ranging in size from a few micrometers to several millimeters. Most coatings are now produced by vacuum-based deposition techniques. Soft coatings (such as Sn, Ag, MoS2, etc.) are mostly used to control friction in sliding bearing applications where use of liquid lubricants is neither desirable nor practical, while hard coatings are well-suited for controlling wear in a variety of metal-cutting and -forming operations as well as rolling, rotating and sliding bearing applications [1]. Among the many hard coatings available today, nitrides and carbides of certain transition metals (such as Ti, Cr, Zr, etc.) are the popular [2–5]. In most Corresponding author. Tel.: +90 212 2856999; fax: +90 212 2853427.

E-mail address: [email protected] (M. U¨rgen). 0301-679X/$ - see front matter r 2007 Published by Elsevier Ltd. doi:10.1016/j.triboint.2007.04.008

metal-cutting and -forming applications, it was found that the extent of improvement in performance and durability is dependent on the mechanical and tribological properties of hard coatings applied to the surface. Accordingly, in recent years, many researchers have focused their attention to further increase the hardness of these coatings by means of alloying at nanoscales or building a nanocomposite and nanolayered or superlattice coating architectures [6]. In the case of nanoscale alloying, addition of both the miscible and immiscible elements into nitride coatings has been tried and in some of the cases, remarkable improvements in hardness and hence wear resistance have been achieved. In the case of nanoscale layering, a series of novel films with extreme toughness have been produced in recent years and offered for large scale applications. Alloying of bulk materials and thin coatings is not a new practice and has certainly been employed at micro- and macro-scales for decades [7]. However, alloying or modification of the structure and chemistry of such materials

ARTICLE IN PRESS 50

A. O¨ztu¨rk et al. / Tribology International 41 (2008) 49–59

and coatings is relatively new and has markedly improved their performance and durability. At present, there exist several kinds of alloyed coatings that can reliably meet the increasingly more stringent application requirements of advanced manufacturing and transportation systems. For example, a series of exotic multicomponent (such as (TiAl)N, TiBN, BCN, Ti(CN)) and multilayered (such as WC/C, TiC/TiB2, etc.) coatings [8–10] have recently been developed and used to combat friction and wear in various metal-cutting and -forming applications. The latest trend in the field is to produce super-hard and -tough coatings that can further enhance property and hence performance when used in increasingly more demanding machining and manufacturing operations [6]. For the production of superhard coatings, researchers have proposed a number of new approaches. At present, the most popular approach is to reduce the grain size of these coatings by creating a nanostructured and/or nanocomposite coating architecture [11,12]. Recent studies have shown that by employing such an approach, it is possible to increase coating hardness to values greater than 40 GPa. Currently, there exist two main practices for the production of nanocomposite coatings. In one of the practices, the nanocomposite structure is obtained by the combination of two hard and immiscible phases in one coating (e.g., nc–TiN/a–Si3N4; a–TiSi2–nc–TiSi2 ) while in the other practice, the mixture of one hard (MeN) and one soft (X) immiscible phase (e.g., Zr–Cu–N, Cr–N– Cu, Ti–Cu–N and Mo–Cu–N) are used. Nanocomposite coatings based on hard/hard phases have attracted an overwhelming interest in recent years and there now exist numerous studies that deal with the production, characterization, and application of these coatings [13–23]. Less attention has been paid to those composite coatings that are based on hard/soft phases. Most of the nanocomposite coatings have been synthesized using PACVD, ion beam sputtering, vacuum arc deposition, magnetron sputtering and recently, by hybrid arc and magnetron sputtering methods [24–28]. A hybrid system based on cathodic arc physical vapor deposition (CAPVD) and magnetron sputtering is most suitable for the production of hard–soft or MeN–X type nanocomposite coatings by taking advantage of the very unique qualities of both techniques. In such hybrid techniques, the soft metal is sputtered from a target and the hard nitride phase is synthesized by the cathodic arc source. There are limited studies in the literature that use this technique for the production of MeN–X type coatings [27,29,30]. From a tribological point of view, hardness is an important property, but hardness alone does not insure long life or low friction in most sliding or machining systems. During sliding and/or machining, very high temperatures and shear stresses are generated at the real contact or machining zones and as a result, significant changes occur in structure, chemistry and hence properties of hard coatings. In particular, strong tribochemical interactions among coating, work piece, and the operating

environment (such as air or liquid lubricant) can significantly impact the friction and wear behavior of coatings. Therefore, any type of further improvements in tribological performance of hard coatings must take into account the type of tribochemical interactions that may take place under the specific conditions of an application. Nanocomposite coatings based on the incorporation of a metallic and/or compound phases, present a unique opportunity to achieve much better performance and durability in machining and manufacturing applications. In particular, with proper selection and effective control of secondary phases, these coatings can be made not only harder but also self-lubricating. However, there are limited numbers of studies in the literature that deal with the optimization of both the friction and wear properties of hard coatings [31–34]. In this study metallic copper is introduced into a series of hard nitride coatings and the changes in their tribological behavior is attempted to be explained by the character and the chemistry of the tribofilms created during sliding process. For the explanation of the behavior, the theoretical principles given in ‘crystal chemical approach’ [35,36] has also been used. Among the many hard nitride coatings, TiN, CrN are widely used by industry. In the other hand, MoN coating is also hard but has rarely been used in the past as a protective coating. Because of its ability to produce soft and lubricious oxides (such as MoO3), it may offer unique possibilities for applications not only in machining but also in sliding bearing applications as demonstrated in recent publications by our research group (especially when a second phase, such as Cu is present) [23,37,38]. 2. Experimental 2.1. Coating procedure The Me–N–Cu (Me: Ti, Mo, Cr) films were produced by a hybrid cathodic arc+magnetron sputtering physical vapor deposition technique. The schematic diagram of the specific system used during deposition is shown in Fig. 1. Me component of the hard coatings were evaporated by a cathodic arc source while Cu is sputtered from a d.c. magnetron source. The reactive gas (i.e., pure nitrogen (99.999%)) was introduced into the system using a flowmeter at flow rates sufficient to result in nearly stoichiometric TiN, MoN, and CrN phases. Hardened high-speed steel (HSS) was used as a substrate material and polished to a surface roughness of Rap0.09 mm. After the polishing process, samples were ultrasonically cleaned in a series of hot alkaline cleaning baths for 5 min, then soaked in trichloroethylene and dried. The specific deposition parameters are given in Table 1. Coating deposition time was between 60 and 65 min. The thickness of the coatings was determined using a ball cratering technique (Calotest). A steel ball of 10 mm diameter was used with a cratering speed and time were 30 rev min1 and 20 s, respectively. As slurry 1 mm diamond

ARTICLE IN PRESS A. O¨ztu¨rk et al. / Tribology International 41 (2008) 49–59

51

Ne laser beam with a wavelength of 632.81 nm and focused to a spot size of 1–2 mm. 2.4. Micro-hardness measurements Hardness measurements were carried out with an ultra micro-hardness tester (Fischercope) 20 mN load was applied in steps of 0.2 mN so that the Vickers pyramid indenter penetrates to the one tenth of coating thickness. Each measurement was repeated at least 12 times for obtaining an average hardness value. 2.5. Wear tests

Fig. 1. Schematics of the Hybrid CAPVD and UBMS system.

Table 1 Hybrid CAPVD+unbalanced magnetron sputtering (UBMS) coating steps and procedures Steps

Procedure

Glow discharge

600 V, 3 Pa, 15 min

Ion etching and heating with Me ions

1000 V bias 60 A for Ti and Cr 100 A for Mo

Coating

The same cathode current used in ion etching 100 V bias, 7.5–15 mTorr nitrogen pressure, Coating temperature 400–450 1C

The tribological properties of hard coatings were evaluated in a ball on disk tribotester and a reciprocating wear test machine (Plint&Partners, Model TE70). The unidirectional tests (ball on disc tests) were performed under a normal load of 5 N and a sliding velocity 0.2 m/s in open air of 4573% relative humidity. The counterface material was 10 mm diameter alumina balls. The bidirectional (reciprocating) wear tests were carried out at 30 Hz reciprocating frequency, under 5 N normal load, and 1 mm stroke length in open air of 4573% humidity. In these tests 10 mm alumina balls were also used as counterface material The total sliding distance was 54 m. Worn ball and flat surfaces were investigated by using optical profilometer (Veeco Wyco NT1100). 3. Results 3.1. Properties of the coatings

suspension is used. The coating thicknesses varied between 1.8 and 3.8 mm, depending on coating time and type. 2.2. Morphology and phase analysis of the coatings The structural details of the coatings were determined by a glancing angle X-ray diffractometer with a thin film attachment (Philips Model PW3710) using Cu Ka radiation over the 2y range of 20–1201. The y scan method with a fixed incidence angle of 21 was used. The phases in each film were identified by matching the diffraction peaks with those of The Joint Committee for Powder Diffraction (JCPDS) database. The cross-sectional film morphology and elemental analyses were conducted using a Field emission scanning electron microscope (FE-SEM) (Jeol, Model: JSM-7000F) equipped with energy dispersive spectroscopy (EDS) unit. 2.3. Micro RAMAN investigations Wear debris and the coatings were analyzed by micro Raman system (Jobin-Yvonne HR 800) which uses a He–

EDS analyses indicated that the hard coatings produced on HSS substrates contained 5.5–14.0 at% copper depending on coating process parameters. The thickness of the coatings (as determined by a calotest method and further verified by cross section SEM investigations) varied between 1.8 and 3.8 mm. The hardness of the TiN and CrN coatings decreased slightly with the addition of the copper, but such a decrease was not observed for the MoN coatings perhaps due to somewhat lower copper content of these coatings compared to CrN and TiN. A gradual decrease in hardness is generally observed in the MeN–X type nanocomposite coatings with the increase of copper contents over 2–3 at%. However, it is reported in the literature [39,40] that hardness decrease was not observed when a high bias voltage was applied during the coating process. For the coatings used in this study, the application of a high bias voltage (100 to 150 V) during the production might be the explanation for retaining of the hardness of the coatings with copper contents well above 2–3 at% (Table 2). Hardness, surface roughness, thickness, and copper content of the coatings are summarized in Table 2.

ARTICLE IN PRESS A. O¨ztu¨rk et al. / Tribology International 41 (2008) 49–59

52

3.2. Structure of the coatings A slight negative peak shift was observed in XRD spectra of all coatings that contained no copper. These shifts in peak positions are attributed to the compressive stresses in the coatings. By the introduction of copper into the coating structure the shifts in the XRD diffraction spectra became less discernible which indicated a decrease of compressive stresses in the films. Peak broadening which is typical for these nanocomposite structures was also observed. [1–15,25,39] indicating a grain size refinement. The XRD spectra of the coatings used in this study are given in Fig. 2(a)–(c). The presence of copper in the structure did not create a change in the crystal structure of the coatings. No diffraction peaks are found for copper in the coatings; suggesting that it is perhaps distributed interstitially within the nanocrystalline nitride phases or in an amorphous state. More structurally sensitive techniques such as TEM, EELS are needed for detecting the copper sites in the structure. Fig. 3 shows the cross-sectional SEM images of standard and Cu-doped MoN films. The cross-sectional images of other films were also similar to the ones shown in this figure. Specifically, with the addition of Cu, the film morphology changed from a columnar one to a nearly featureless structure. Such a dramatic change in structure is primarily attributed to the fact that Cu (which is immiscible in these TiN, CrN, and MoN coatings) is highly mobile and thus able to disrupt the columnar grain growth during deposition and hence control the grain size. These results are consistent with the findings of other researchers who have explored the effects of soft and/or noble metals on film morphology of nitride-based films [14,23,27,41–43]. 3.3. Tribological properties of the coatings 3.3.1. Ball on disc wear tests The results of tribological tests have further confirmed that the friction and wear coefficients of un-doped coatings were rather high for TiN but somewhat lower for both CrN and MoN (see Fig. 4). In other studies too, researchers have often reported very high friction for TiN and relatively low friction for CrN and MoN coatings [4,38]. The exact mechanisms that control friction and wear behavior of these coatings are not well-understood, but

certain investigators have attributed high friction of TiN to the formation of TiOx while the relatively low friction of CrN and MoN was also attributed to the oxygen-rich tribofilms forming on their sliding surfaces. Apparently, compared to TiOx, the oxides of Mo and Cr have lower shear strength and hence capable of providing lower friction, especially at elevated temperatures Certain TiOx films (especially the ones with Magnelli type phases) are also known to possess easy shear character and hence low friction [31,44]. However, based on our limited surface analytical and structural work in this study, we could not verify the presence and/or absence of such phases within the tribofilms. The addition of copper into CrN caused a slight increase in its average friction coefficient and the frictional trace has become very unsteady. However, the presence of Cu in MoN film slightly reduced its friction coefficient. In the case of TiN, a substantial increase was observed in friction after Cu addition as shown in Fig. 5. Examination of the wear tracks by a 3D surface profilometer has revealed a polishing type wear on the surfaces of CrN and MoN coatings with or without Cu-doping. Specifically, the surface asperities of these coatings were completely removed during sliding tests and the sliding wear tracks had become much smoother than the original surface. However, in the case of TiN, the addition of Cu appeared to have an adverse effect. Specifically, the wear depths of TiN coatings increased from 0.8 to 4.0 mm with the addition of copper and these tracks became very rough and rugged as shown in Fig. 6. The addition of Cu into hard coatings had a significant positive effect on the wear behavior of counter face alumina balls (Fig. 7). The least amount of wear was measured on balls that were slid against the doped and undoped MoN films, while the highest amount of wear was found on alumina balls that were slid against the TiN films. The greatest positive effect due to Cu-doping was seen on MoN films (more than 50% reduction in wear rate as shown in Fig. 7). 3.3.2. Reciprocating wear tests In an attempt to further confirm the results of unidirectional ball-on-disk tests and to determine the effect of copper doping on friction and wear more precisely, we performed a number of complementary reciprocating tests in another machine. After the tests, we also examined the

Table 2 Copper content, thickness, hardness and surface roughness of the coatings Coating

Cu (at%)

Cu (wt%)

Thickness (mm)

Hardness (Gpa)

Surface roughness Ra (mm)

MoN MoN–Cu TiN TiN–Cu CrN CrN–Cu

— 5.5 — 12.8 — 11.2

— 3.7 — 16.3 — 13.3

2.5 2.4 2.5 1.8 3 3.8

3770.6 4270.7 3270.6 3070.6 2870.5 2770.5

0.1970.05 0.1870.09 0.2370.02 0.1870.09 0.1470.08 0.1670.09

ARTICLE IN PRESS A. O¨ztu¨rk et al. / Tribology International 41 (2008) 49–59

53

c -TiN-JCPDS File no:38-1420 Cu -JCPDS File no:04-0836

Intensity (a.u.)

TiN

TiN-Cu

20

30

40

50

60

70 2θ

80

90

100

110

120

MoN - JCPDS File no:25-1367 Cu - JCPDS File no:04-0836

Intensity (a.u.)

MoN

MoN-Cu

20

30

40

50

60

70 2θ

80

90

100

110

120

c-CrN- JCPDS File no:11-0065 Cu - JCPDS File no:04-0836

Fig. 3. Cross-sectional SEM images of (a) MoN and (b) Mo–Cu–N coatings.

Intensity (a.u.)

CrN

CrN-Cu

20

30

40

50

60

70 2θ

80

90

100

110

120

Fig. 2. XRD spectra of the: (a) TiN and TiN–Cu, (b) MoN and MoN–Cu, (c) CrN and CrN–Cu coatings.

structural chemistry of the sliding surfaces and tribofilms that were formed during these reciprocating tests. It is well known that higher speeds and shorter stroke lengths in a

Fig. 4. The CoF–distance relation of TiN, CrN and MoN coatings under 5 N load (ball on disc tests).

ARTICLE IN PRESS 54

A. O¨ztu¨rk et al. / Tribology International 41 (2008) 49–59

Fig. 5. The CoF–distance relation of TiN–Cu, CrN–Cu and MoN–Cu coatings under 5 N load (ball on disc tests).

Fig. 6. 3D optical wear profiles of the: (a) TiN and (b)TiN–Cu coatings after ball on disc wear tests.

reciprocating test give rise to higher temperatures and shear stresses at the contact zone. Both of these effects are expected to create a situation that favors tribo-oxidation and/or tribochemical interactions between sliding pairs and surrounding atmosphere.

Fig. 7. Wear rates of alumina balls tested against the coatings.

As can be seen from Fig. 8, the frictional behaviors of both the un-doped and doped TiN, CrN, and MoN coatings during reciprocating tests were similar to the ones that were observed during unidirectional sliding tests. Again, the presence of copper in the coatings increased the friction coefficients of TiN and CrN coatings, while causing a decrease for the MoN coatings. In these tests, the TiN coatings exhibited the worst tribological performance. In addition to exhibiting the highest friction among all the coatings tested, they were also totally worn out even after sliding for only a few thousand cycles (Figs. 8a, 9a and b) The maximum wear depths measured on the wear tracks were 10.4 mm for TiN (see Fig. 10a). With the addition of copper into TiN, a slight decrease in wear depth from 10.4 to 6.1 mm was achieved as shown in Fig. 10(a). However, in all of the TiN coatings tested, the coating layers were totally worn out and the steel substrate was exposed (see Fig. 9a,b)). Therefore, the friction and wear results reported are representative of not only the TiN coating but also the substrate materials. In the case of CrN, the depth of wear track increased from 0.42 to 0.63 mm by the addition of copper into CrN coatings. However the wear depths were incomparably lower than those of the TiN coatings (Fig. 10a). MoN coatings showed a far better performance by affording not only a lower friction coefficient (i.e., E0.3) but also the shallowest wear depth among all the undoped nitride films tested in our study. The addition of copper into MoN resulted in further lowering of COF to less than 0.3 and almost no measurable wear as shown in Figs. 8(b), 9(c), (d) and 10(a). The most positive effect of Cu is observed on the wear volumes of the alumina balls as observed in ball on disc tests. The wear rates of balls showed a decrease with the addition of copper into the coating structure (Fig. 10b)

ARTICLE IN PRESS A. O¨ztu¨rk et al. / Tribology International 41 (2008) 49–59

55

°

°

°

Fig. 8. CoF vs cycle diagram of: (a ) TiN and TiN–Cu, (b) Mo–N and MoN–Cu, (c) CrN and CrN–Cu coated HSS discs against alumina balls. Fig. 9. 3D optical wear profiles of the coatings after reciprocating wear tests. (a) TiN coating, (b) TiN–Cu coating, (c) MoN coating, (d) MoN–Cu coating.

ARTICLE IN PRESS 56

A. O¨ztu¨rk et al. / Tribology International 41 (2008) 49–59

spectrum of TiN–Cu coatings, the bands were somewhat broader and less intense suggesting much smaller grain size for the debris particles produced during reciprocating test. Moreover a slight shift was observed at bands around 250 and 400 cm1 compared to the debris particles of un-doped TiN. In this region, there are bands of copper oxide. Deconvolution of the peaks in this region has indeed confirmed the presence of CuO (Fig. 11b). These Raman results also suggested that CuO exists as a separate phase in the debris particles.

b

a

Fig. 11. (a) Raman spectrum of the wear debris formed on TiN and TiN– Cu coatings during reciprocating wear tests and reference spectra of anatase, rutile and CuO. (b) Deconvoluted part of the spectrum. Fig. 10. (a) Reciprocating wear scar maximum wear depth of MeN–Cu and MeN coatings, (b) ball wear rate of the alumina balls working against Me–N and MeN–Cu coatings.

3.4. Raman investigation of wear debris The debris formed in the wear tracks after reciprocating tests were investigated with a micro-Raman system and the Jobyn–Yvon library was used to determine the various types of oxides of Cr, Mo, Cu and Ti (TiO2 (rutile), TiO2 (anatase), Cr2O3, CuO, Cu2O and MoO3 that may have formed during sliding tests. For reference purposes, the complex oxide of Mo and Cu (i.e., CuMoO4) was prepared according to the recipe described by Ehrenberg et al. [45]. For the Raman investigation of the debris of TiN and TiN–Cu coatings without the contribution of steel substrate (hence iron oxides), the shorter-duration reciprocating wear tests were conducted and the tests were stopped just before reaching the substrate. The debris of TiN and TiN–Cu coatings was similar in their Raman character. The debris of TiN is primarily composed of the oxides of Ti, namely rutile and anatase (Fig. 11a). In the Raman

Fig. 12. Raman spectrum of the wear debris formed on CrN and CrN–Cu coatings during reciprocating wear tests and reference spectra of Cr2O3, and CuO.

ARTICLE IN PRESS A. O¨ztu¨rk et al. / Tribology International 41 (2008) 49–59

In the Raman spectrum of the debris particles of CrN coatings, Cr2O3 bands were clearly observed at 305, 546 and 606 cm1. One strong broad band was also observable at 686 cm1. (Fig. 12). A similar spectrum was also obtained for the CrN debris in previous studies [46]. However, no explanation for the bands observed at 686 and 868 cm1 had so far been given. These two broad bands gave very similar characteristics to CrO2 and Cr8O21 (a mixture of CrO2 and Cr2O3) compounds as reported by Stanoi et al. [47]. According to the Raman findings, we have concluded that the debris of CrN is mainly composed of a mixture of +3 and +4 valent oxides of chromium (i.e., Cr2O3 and CrO2). There are some differences between the Raman spectra of CrN and CrN–Cu coatings. The hump between 1050 and 1200 cm1 of standard CuO is also observed in the Raman spectrum of CrN–Cu. Furthermore, the bands of CuO between 250 and 350 cm1 overlaps with the bands of Cr2O3. These differences in the spectra can be interpreted as an indication of the presence of separate phases of chromium oxides and CuO in the debris of CrN–Cu coatings. The Raman spectra of the debris of MoN and MoN–Cu coatings (Fig. 13) are substantially different from those of the CrN, CrN–Cu, TiN, and TiN–Cu. The Raman spectrum of the wear debris particles of MoN coatings were consistent with the ones that were reported in a previous study [31]. Namely, in these spectra, we could clearly identify a number of Raman shifts that are related to MoO3 and molybdenum sub oxides. In the spectrum of MoN–Cu coatings, the Raman shifts due to the CuO, as observed in other coatings, was not present however broad peaks in the regions (800–1000 cm1) corresponding to copper molybdates became more pronounced. The formation of such copper molybdates has already been reported by others for Mo–Cu films upon heating in open air [48].

Fig. 13. Raman spectrum of the wear debris formed on MoN and MoN– Cu coatings during reciprocating wear tests and reference spectra of MoO3, CuMoO4 and CuO.

57

In summary, Raman spectroscopy has revealed that the wear debris particles of all the films tested contained large amounts of oxides. In the wear debris of copper containing TiN, and CrN coatings, copper oxides were present as separate phases, besides the oxides of Ti and Cr. However, in the wear debris of MoN–Cu coatings, there was a evidence for the formation of a complex oxides based on molybdenum and copper. 4. Discussion The results of this study further confirmed that it is possible to produce Cu-containing TiN, CrN, and MoN based nanocomposite (nc) coatings by a hybrid PVD system consisting of arc-PVD and magnetron sputtering. XRD and fracture cross section investigations of the copper-doped coatings revealed grain size refinement in the form of a featureless cross section morphology that is typical of most MeN-X type nc-coatings. Although the copper content of the films was above 2–3 at% a substantial decrease in the hardness of the nc coatings is not observed. However, from a tribological point of view, all of the coatings exhibited markedly different friction and wear behaviors. It is a well known fact that the character of the wear debris (oxides), produced during sliding contact, plays a very crucial role in determining the tribological behavior of sliding system [49]. In particular, the chemical nature of these debris particles formed during sliding test (single, complex, stoichiometric, sub-stoichiometric, etc.) may have a significant effect on friction and wear. One of the theoretical approaches for the explanation of the relationship between debris chemistry and lubricity is the ‘‘crystal chemical’’ approach [50]. According to this approach, those oxides or oxide mixtures with higher ionic potentials are more likely to provide low friction when present at a sliding interface. Conversely, the presence of those oxides with lower ionic potentials will lead to high friction and hence high third-body wear. The ionic potential is described as j ¼ Z/r. Z refers to formal cationic charge and r refers to the radius of the cation [36]. Those cations with very small radius are more effectively screened by the surrounding oxygen anions and hence their ionic potentials are much larger. Those oxides with high ionic potentials (such as Re2O7 with an ionic potential of E13) are soft and hence easy to deform and/or shear during sliding. Such oxides are often referred to as lubricious oxides in tribology literature. For binary oxides, the crystal chemical approach considers the difference in the ionic potentials of the two oxides that are present on sliding surfaces. In this case, the larger the difference, the lower the friction coefficients [50]. Hence oxides with a high ionic potentials are expected to function as lubricious compounds. TiN was the coating giving the worst tribological behavior in both of the wear tests utilized in this study (see Figs. 4, 5 and 8). The Raman analysis of the wear debris particles of these coatings showed that it is

ARTICLE IN PRESS A. O¨ztu¨rk et al. / Tribology International 41 (2008) 49–59

58

composed of a mixture of titanium oxides, namely rutile and anatase (Fig. 11). Considering the crystal chemical approach; the oxides of titanium with their low ionic potentials are not expected to possess easy shear property, but rather classified as one of the abrasive oxides. Therefore, the oxide-based debris particles of TiN would cause high friction and wear. For the mixed oxides, the crystal chemical approach predicts that an improvement in friction and wear performance should result if there is a high difference between the ionic potentials of the oxide constituents. The difference between the ionic potentials of titanium and copper oxides is rather small, i.e., 2.6, hence the presence of copper oxides is not expected to bring an improvement in the tribological behavior of TiN–Cu coating (Table 3). In the Raman analysis of TiN–Cu coating debris, CuO is detected to be present as a separate phase (Fig. 11). In the case of CrN, Cr2O3 which was the major wear debris detected at sliding interface and it is a well-known oxide for its abrasive character and has an ionic potential of 4. However CrN coatings exhibited a behavior that is not very typical of abrasive oxide forming coatings. As can be seen in Figs. 4, 5 and 8, regardless of Cu, friction coefficient of CrN was lower than that of TiN. In the Raman spectra of sliding surfaces of CrN coatings, we could detect bands related to CrO2 in addition to Cr2O3 (see Fig. 12). Formation of these oxides during sliding can provide an explanation for the much better tribological behavior of CrN coatings. The ionic potential of CrO2 is 7.3. This value is much higher than the ionic potential of Cr203 which is 4. As explained above, according to the crystal chemical approach, the lubricious character of oxides increases with the increase of ionic potentials. The formation of an oxide during wear process with higher ionic potential (CrO2) was perhaps the reason for good tribological behavior of CrN coatings in this study. The introduction of copper into the structure of CrN did not result in an improvement of the tribological behavior

Table 3 Ionic potentials of various oxides Compound

TiO2: titanium (IV) oxide

Coordination number Ionic potential, j, Z/r

6-coordinate, octahedral MoO3: molybdenum (VI) 6-coordinate, oxide tetrahedral CrO2: chromium (IV) oxide 4-coordinate, tetrahedral Cr2O3: chromium (III) 6-coordinate, oxide octahedral CuO copper (II) oxide 4-coordinate, tetrahedral

5.4 8.2 7.3 4 2.8

Ionic potentials shown in the table are easily calculated by considering the cationic charge and the radius for each cation that can be found at /http://www.webelements.comS and other sources.

of the coatings. The Raman studies of the wear debris of copper containing coatings showed the presence of CuO as a separate phase besides the chromium oxides (Fig. 12). However, for these coatings, the detrimental effect of copper is not as severe as for TiN coatings. Again according to crystal chemistry approach, this can be attributed to the relatively higher difference of ionic potentials between CrO2 and CuO (7.32.8 ¼ 4.5). Molybdenum nitride coatings and their copper-doped versions gave the best overall tribological behavior during both the pin-on-disk and reciprocating tests (see Figs. 4, 5 and 8). The favorable tribological behavior of undoped and Cu-doped MoN is attributed to the formation of lubricious molybdenum oxides [38] and the presence of this oxide is clearly confirmed by the Raman Spectroscopy (Fig. 13). Also, according to the crystal chemical approach, molybdenum oxides are expected to give low friction coefficients due to their high ionic potential of 8.2. Introduction of copper into molybdenum nitride, contrary to TiN and CrN, showed a beneficial effect on tribological behavior by reducing both friction and wear. In the Raman spectrum of the wear debris of these coatings, the formation of a complex copper and molybdenum oxide is discernible (Fig. 13). Copper molybdenum mix oxides (copper molybdates) are well known for their lubricious character [48,49]. These results are in accordance with the crystal chemical approach; specifically the higher difference of ionic potentials between the copper and molybdenum oxides promotes the formation of lowfriction compounds. 5. Conclusions In this study, we investigated the friction and wear behaviors of TiN, CrN, and MoN with and without doping with Cu. Based on the results of structural, chemical, mechanical, and tribological studies, we can conclude the followings: 1. Doping TiN, CrN, and MoN with Cu significantly altered the structural morphology of these films. 2. Mechanically, the microhardness values of the coatings other than MoN were slightly reduced by Cu doping. 3. Friction and wear of base TiN, CrN, and MoN were affected significantly by the presence of Cu in the microstructure. Specifically, Cu-doping resulted in significant increases in the friction and wear of TiN. It had little or no significant effect on the friction and wear behavior of CrN; but had a significant positive effect on the behavior of MoN. 4. Raman spectroscopy has confirmed the formation of the oxides of Ti, Cr, Mo, and Cu on sliding surfaces during tribological tests. 5. In the case of TiN and CrN, the oxides of Ti, Cr, and Cu did not further react with each others to form more complex oxides during sliding contact, rather they remained as separate entities.

ARTICLE IN PRESS A. O¨ztu¨rk et al. / Tribology International 41 (2008) 49–59

6. In the case of MoN, however, the oxides of Mo and Cu further reacted during sliding and resulted in the formation of complex oxides based on molybdenum and copper, such as copper molybdate (CuMoO4) which is a wellknown lubricious oxide due to its high ionic potential. 7. Using the general principles of a crystal chemical approach, a mechanistic explanation was provided for the beneficial effect that Cu had on the friction and wear behavior of MoN films. 8. Further, the low-friction behavior of CrN (regardless of Cu-doping or not) was related to the formation of CrO2 on its sliding surface. The ionic potential of this oxide is also high and hence qualifies as a lubricious oxide.

Acknowledgments This work was supported by State Planning Organization of Turkey through the ‘‘Advanced Technologies in Engineering’’ and Production and Development of Low Wear and Low Friction Coefficient Providing Nanocomposite Coatings for High Temperature and Power Systems’’ projects. This work was also partially supported by the US Department of Energy, Office of Freedom Car and Vehicle Technologies Program, under contract W-31109-Eng-38. References [1] Donnet C, Erdemir A. Surf Coat Technol 2004;180–181:76–84. [2] Schwaller P, Haug FJ, Michler J, Patscheider J. Adv Eng Mater 2005;7(5):318–22. [3] Navinsek B, Panjan P, Milosev I. Surf Coat Technol 1997;97:182–91. [4] Su YL, Yao SH, Leu ZL, Wei CS, Wu CT. Wear 1997;213:165–74. [5] Kelesoglu E, Mitterer C, Kazmanli MK, Urgen M. Surf Coat Technol 1999;116–119:133–40. [6] Erdemir A, Eryilmaz OL, Urgen M, Kazmanli K, Mehta N, Prorok B. Tribology of nanostructured materials and coatings. In: Gogotsi Y, editor. Hanbook of nanomaterials. CRC Publication; 2005. p. 685–711. [7] Hauert R, Patscheider J. Adv Eng Mater 2000;2(5):247–59. [8] PalDey S, Deevi SC. Mater Sci Eng A 2003;342:58–79. [9] Wei C, Lin JF, Jiang T-H, Ai CF. Thin Solid Films 2001;381:94–103. [10] Wolfe DE, Singh J, Narasimhan K. Surf Coat Technol 2003; 165:8–25. [11] Musil J. Surf Coat Technol 2000;125:322–30. [12] Veprek S. J Vac Sci Technol A 1999;17(5):2401–20. [13] Musil J, Zeman P. Vacuum 1999;52:269–75. [14] Zeman P, Cerstvy R, Mayrhofer PH, Mitterer C, Musil J. Mater Sci Eng A 2000;289:189–97. [15] Musil J, Zeman P, Hruby H, Mayrhofer PH. Surf Coat Technol 1999;120–121:179–83. [16] Baker MA, Kench PJ, Tsotsos C, Gibson PN, Leyland A, Matthews A. J Vac Sci Technol A 2005;23(3):423–33. [17] Musil J, Leipner I, Kolega M. Surf Coat Technol 1999;115:32–7. [18] Andreasen KP, Jensen T, Petersen JH, Chevallier J, Bøttiger J, et al. Surf Coat Technol 2004;182:268–75.

59

[19] Abadias G, Tse YY, Michel A, Jaouen C, Jaouen M. Thin Solid Films 2003;43:166–73. [20] He JL, Setsuhara Y, Shimizu I, Miyake S. Surf Coat Technol 2001;137:8–42. [21] Li ZG, Miyake S, Kumagai M, Saito H, Muramatsu Y. Surf Coat Technol 2004;183:62–8. [22] Joseph MC, Tsotsos C, Baker MA, Kench PJ, Rebholz C, Matthews A, et al. Surf Coat Technol 2005;190:345–56. [23] O¨ztu¨rk A. Manyetik sıc- ratma teknig˘i ile u¨retilmis- Mo2N ve nanokompozit Mo–N–Cu kaplamaların kazımalı as-ınma davranıs-larının kars-ılas-tırılması, M.Sc. thesis, I.T.U¨. Institute of Science and Technology, Istanbul, 2003. [24] Ma S, Prochazka J, Karvankova P, Ma Q, Niu X, Wang X, et al. Surf Coat Technol 2005;194:143–8. [25] Veprek S, Niederhofer A, Moto K, Bolom T, Mannling H-D, Nesladek P, et al. Surf Coat Technol 2000;133–134:152–9. [26] Veprek S, Haussmann M, Reiprich S. J Vac Sci Technol A 1996; 14(1):46–51. [27] Turutoglu T, Urgen M, Cakir AF, Ozturk A. Key Eng Mater Euro Ceramics VIII 2004;264–268(I):489–92. [28] Eryılmaz OL, Ark Fiziksel Buhar Biriktirme Yo¨ntemiyle Nanokompozit Mo–N–Cu Kaplamaların U¨retimi ve Karakterizasyonu, PhD thesis, I.T.U¨. Institute of Science and Technology, I˙stanbul, 2001. [29] Myung HS, Han JG, Boo JH. Surf Coat Technol 2004;177–178: 404–8. [30] Kim JS, Kim GJ, Kang MC, Kim JW, Kim KH. Surf Coat Technol 2005;193:249–54. [31] Suszko T, Gulbin´ski W, Jagielski J. Surf Coat Technol 2005; 194:319–24. [32] Endrino Jose L, Nainaparampil Jose J, Krzanowski James E. Scripta Materialia 2002;47:613–8. [33] Endrino Jose L, Nainaparampil Jose J, Krzanowski James E. Surf Coat Technol 2002;157:95–101. [34] Suszko W, Gulbinski J, Jagielski T. Sur Coat Technol 2006;200: 6288–92. [35] Erdemir A. Tribol Lett 2000;8:97–102. [36] Erdemir A, Shenghua L, J Yunasheng J. Int J Mol Sci 2005;6:203–18. [37] Kazmanlı MK, U¨rgen M, C - akır AF. Surf Coat Technol 2003;167: 77–82. [38] U¨rgen M, Eryılmaz OL, C - akır AF, Kayalı ES, Nilu¨fer B, Is-ık Y. Surf Coat Technol 1997;94–95:501–6. [39] Andreasen KP, Jensen T, Petersen JH, Chevallier J, Bottiger J, Schell N. Surf Coat Technol 2004;182:268–75. [40] Li ZG, Miyake S, Kumagai M, Saito H, Muramatsu Y. Surf Coat Technol 2004;183:62–8. [41] Myung HS, Lee HM, Shaginyanb LR, Han JG. Surf Coat Technol 2003;163–164:591–6. [42] Barna PB, Adamik M. Thin Solid Films 1998;317:27–33. [43] Baker MA, Kench PJ, Joseph MC, Tsotsos C, Leyland A, Matthews A. Surf Coat Technol 2003;162:222–7. [44] Gardos MN. Tribol Lett 2000;8:65–78. [45] Ehrenberg H, Weitzel H, Paulus H, Wiesmann M, Wltschek G, Geselle M, et al. J Phys Chem Solids 1997;58(I):153–60. [46] Constable CP, Yarwood J, Hovsepian P, Donohue LA, Lewis DB, Munz W-D. J Vac Sci Technol A 2000;18(4):1681–9. [47] Stanoi D, Socol G, Grigorescu C, Guinneton F, Monnereau O, Tortet L, et al. Mater Sci Eng B 2005;118:74–8. [48] Wahl KJ, Seitzman LE, Bolster RN, Singer IL, Peterson MB. Surf Coat Technol 1997;89:245–51. [49] Gulbinski W, Suszko T, Sienicki W, Warcholinski B. Wear 2003; 254:129–35. [50] Erdemir A. Surf Coat Technol 2005;200:1792–6.