AlSiTiN and AlSiCrN multilayer coatings: Effects of structure and surface composition on tribological behavior under dry and lubricated conditions

Accepted Manuscript Title: AlSiTiN and AlSiCrN multilayer coatings: effects of structure and surface composition on tribological behavior under dry and lubricated conditions Author: Maria Giulia Faga Giovanna Gautier Federico Cartasegna Paolo C. Priarone Luca Settineri PII: DOI: Reference:

S0169-4332(16)00002-7 http://dx.doi.org/doi:10.1016/j.apsusc.2015.12.237 APSUSC 32220

To appear in:

APSUSC

Received date: Revised date: Accepted date:

30-6-2015 22-12-2015 31-12-2015

Please cite this article as: M.G. Faga, G. Gautier, F. Cartasegna, P.C. Priarone, L. Settineri, AlSiTiN and AlSiCrN multilayer coatings: effects of structure and surface composition on tribological behavior under dry and lubricated conditions, Applied Surface Science (2016), http://dx.doi.org/10.1016/j.apsusc.2015.12.237 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

AlSiCrN

AlSiTiN

WC-Co Uncoated

1.2 Top coating

Multilayer, n.5 alternate A+B layers, fixed stoichiometry, thickness = 0.5 T

Multilayer step B

0.8 0.6 0.4 0.2

Gradient layer

0

Adhesion layer

0

500

1000

1500

2000

cr

Metal substrate

ip t

Multilayer step A

1.0 Coefficient of friction

Total coating thickness = T

Top coating, different stoichiometry, thickness = 0.5 T

us

Laps

d

M

an

The demand for high performance nanostructured coatings has been increasing. AlSiTiN and AlSiCrN nanocomposite coatings were deposited by PVD technique. Coatings were analyzed in terms of structure, hardness and adhesion. Tribological properties under dry and lubricated conditions were studied. The effects of surface and bulk properties on friction evolution were assessed.

te

AlSiTiN and AlSiCrN multilayer coatings: effects of structure and surface composition on tribological behavior under dry and lubricated conditions

a

Ac ce p

Maria Giulia Fagaa, Giovanna Gautiera, Federico Cartasegnab, Paolo C. Priaronec, Luca Settineric*[email protected] National Research Council (CNR), Imamoter, Strada delle Cacce 73, 10135 Torino, Italy Clean NT Lab, Environment Park S.p.A., Via Livorno 60, 10144 Torino, Italy c Politecnico di Torino, Department of Management and Production Engineering, Corso Duca degli Abruzzi 24, 10129 Torino, Italy b

Phone: (+39) 011 0907230, fax: +39 011 0907299

Abstract Nanocomposite coatings have been widely studied over the last years because of their high potential in several applications. The increased interest for these coatings prompted the authors to study the tribological properties of two nanocomposites under dry and lubricated conditions (applying typical MQL media), in order to assess the influence of the surface and bulk properties on friction evolution. To this purpose, multilayer and nanocomposite AlSiTiN and AlCrSiN coatings were deposited onto tungsten carbide-cobalt (WC-Co) samples. Uncoated WC-Co materials were used as reference. Coatings were analyzed in terms of hardness and adhesion. The structure of the samples was assessed by X-ray diffraction (XRD), while the surface composition was studied by XPS analysis. Friction tests were carried out under both dry and lubricated conditions using an inox ball as counterpart. Both coatings showed high hardness and good adhesion to the substrate. As far as the friction properties are concerned, in dry conditions the surface properties affect the sliding contact at the early beginning, while bulk structure and tribolayer formation determine the main Page 1 of 19

behavior. Only AlSiTiN coating shows a low and stable coefficient of friction (COF) under dry condition, while the use of MQL media results in a rapid stabilization of the COF for all the materials. Keywords Nanocomposite coating; AlSiTiN; AlSiCrN; Mechanical performance; Microstructure; Friction.

us

cr

ip t

1. Introduction Veprek has reviewed the development of superhard materials, pointing out that nanocrystalline composites, such as nanocrystalline nitrides embedded into a matrix of amorphous silicon nitride, hold the best promise of meeting all the complex demands of technically applicable superhard materials [1]. The review describes the basis for strength and hardness, explaining how theoretical values are considerably different from the real ones, because of the strong influence of the microstructure, which affects the growth and propagation of micro-cracks (in ceramic materials). According to the Hall-Petch law, the material strength could be reasonably increased by decreasing the crystals particles [1].

Ac ce p

te

d

M

an

With respect to the potential described in the aforementioned review, innovative coatings with outstanding properties, as quaternary chromium aluminum silicon nitride (CrAlSiN) and aluminum silicon titanium nitride (AlSiTiN), have been developed over the last few years [2-4]. Several research efforts have been carried out aiming to correlate the coating structure/architecture with the mechanical and thermal properties. For instance, Lin et al. [5] focused on the microstructure evolution of CrAlN/SiNx coatings with different bilayer thickness, discussing the relationship between grain size and properties, while the microstructures and properties of Al70Cr30-xSixN (x = 0, 1, 2, 5 at.%) films deposited by reactive arc-evaporation have been studied by Soldán et al. [6]. As far as CrAlSiN nanocomposite coatings (deposited by lateral rotating cathod arc technique) are concerned, Ding et al. [7] have assessed the influence of different (Al+Si)/Cr atomic ratios on composition, structural, and mechanical properties. In their paper, the authors identified the threshold (Al+Si)/Cr ratio maximizing the coating hardness, and the abrasion resistance was found to be proportional to the hardness. Milling tests executed with CrAlSiN coated tools provided better results than those performed with TiAlN-coated tools. For CrAlSiN films, Ho et al. [8] have studied the effect of a heat treatment (at temperatures ranging from 400 to 800°C) in a nitrogen atmosphere, observing that CrAlSiN film heat-treated at 600°C showed the highest hardness, the lowest friction coefficient, and the best wear resistance among all the others test conditions. Such performance was confirmed also by means of cutting tests. Chen et al. [9] have investigated how silicon content affects oxidation behavior in Si-doped CrAlSiN coatings. According to Tritremmel et al. [10], who evaluated the effect of aluminum and silicon content when producing arc evaporated Al-Cr-Si-N thin films, the addition of Si results in an increase of hardness, accompanied by a reduced crystallinity and grain size. However, Al-Cr-Si-N films with high Si content show a reduced hardness due to grain boundary sliding promoted by the higher amorphous phase fraction. Overall, the improved thermal stability and oxidation resistance (Zhang et al. [11]; Polcar and Cavaleiro [12]), together with the excellent high temperature tribological properties (Polcar and Cavaleiro [13, 14]), make quaternary coatings suitable for several applications, like high performance cutting tools. In a previous research work, the authors evaluated the performance of AlSiTiN and AlSiCrN coatings produced via cathodic arc PVD, with multilayer and gradient microstructure, versus commercial AlTiN and AlCrN films [15]. Both gradient and multilayer AlSiTiN coatings guaranteed the lower tool wear values when contour milling an AISI M2 steel, due to the high hardness and the excellent wear resistance at high temperatures [16]. Moreover, when milling difficult-to-cut materials, as titanium aluminides, under severe conditions (e.g., in dry cutting), tests revealed that tool life can be increased by using nanocomposite coatings. In addition, AlSiTiN coating showed higher wear resistance than CrAlSiN coating [17]. Page 2 of 19

ip t

The motivations for using nanocomposite multilayers in the field of hard films are predominantly (1) to facilitate a strong adhesion between film and substrate, (2) to obtain wear-protective films with a low chemical reactivity and low friction, and (3) to increase the hardness and toughness of the coating. The layered structure also hinders propagation of cracks through the coatings because of the ballistic effect (multilayer acts as a crack inhibitor) increasing the coating fracture resistance. Furthermore, it was recognized that a significant reduction of thickness of individual layers and of the crystallite sizes, down to a few nanometers, can result in a significant enhancement of the film features. [18].

Ac ce p

te

d

M

an

us

cr

Summing up, the quaternary coatings are suitable and commercially exploitable for the application under severe thermomechanical conditions (e.g., cutting tools). This aspect makes interesting to further explore their properties: in particular, the friction behavior under green lubricated conditions have not been widely explored yet, to the authors’ knowledge. Such lack of information prompted the authors to focus this paper on the effect of bulk and, especially, surface properties on the friction behavior of PVD AlSiTiN and AlSiCrN nanocomposite multilayer coatings using a Minimum Quantity Lubrication system. Such lubrication method is expected to reduce environmental impact and to be safe for operators [19].

Page 3 of 19

cr

ip t

2. Experimental approach 2.1. Substrate material and coatings The substrate material was a WC-Co hard metal (HM) supplied by Wolframcarb Rock Tools S.r.l. (Italy). The composition as well as the physical and mechanical properties are listed in Table 1 and Table 2, respectively. TiC and TaC(NbC) are carbides usually added to the basic HM composition, mainly made of WC and a metallic binder, as Co in most of the cases (also Ni can be used as binder). Depending on the grade of cemented carbides, a different percentage of such carbides is added. Basically, the constituents can be divided into three phases: alpha (WC), beta (the binder), gamma (single carbide or combination of carbides different from WC, e.g. TiC, TaC/NbC). Based on this classification, Table 1 shows three columns referring to the percentage volume of each phase. Prior to coating deposition, the surface of the substrate was polished to a mirror-like finish, corresponding to a surface roughness Ra of approximately 0.02 µm.

M

an

us

The substrate was coated with AlSiTiN and AlSiCrN nanocomposites, consisting in nanoparticles of aluminum and titanium nitrides (or mixed AlTi nitrides) or aluminum and chromium nitrides (or mixed AlCr nitrides), embedded into a matrix of amorphous Si3N4 (a-Si3N4 in the following). The coatings were prepared through a Physical Vapour Deposition (PVD) process carried out by means of a PL-55 prototype unit available at the Clean NT Lab of Environment Park S.p.A. (Italy). The unit was equipped with the Lateral Arc Rotating Cathodes (LARC) system. The machine was fitted out with two cathodes: one made of titanium (or chromium), and the second one made of an Al-Si eutectic alloy.

Ac ce p

te

d

The coating architecture (as shown in Fig. 1) can be divided into four different zones, starting from the substrate: (1) an adhesion layer, (2) a gradient layer, (3) a multilayer, and (4) a top layer. The adhesion layer (1) is composed of Ti(or Cr)N. This layer, characterized by a thickness of few nanometers, enhances the adhesion between the substrate and the other layers. Then, a thin gradient layer of Ti(or Cr)AlSiN is deposited (2). This layer allows the transition between the single (adhesion) layer and the multi-layer. There is a continuous variation in composition of the Ti(or Cr)AlSiN layer, that reaches on its top the same composition of the first layer of the multilayer. The multilayer (3) supports mechanical loads and thermal strains. The multilayer is composed of Ti(or Cr)AlSiN layers differing in chemical composition. This portion of the coating is therefore divided in alternating A+B layers, where the A layer presents a composition richer in Ti(or Cr), while the B layer has a composition richer in AlSi. The surface layer (4), labeled as top coating in Fig. 1, richer in aluminum, has the highest hardness in order to guarantee the best wear resistance, and shows the best performance in terms of thermal barrier and corrosion resistance. The coating distribution was designed in order to achieve a constant thickness of 3 µm.

2.2 Coating characterization The coating thickness was measured by means of the ball crater test (Calotest), while the adhesion was assessed by Rockwell indentations and scratch tests. For the Calotest, a ball made of hardened steel is turned in order to abrade the coating (also by using a standard metallographic diamond lapping paste). Once the film has been abraded off, the projection surface can be evaluated. For the Rockwell indentation, the resistance to crack propagation along the interface can be assumed as a measurement of the adhesion. The result is evaluated by optical microscope imaging. The adhesion level is classified by using reference images (coded as HF1 to HF8), according to the level of cracking or the coating delamination around the indentation. As far as the scratch test is concerned, a commercial scratch tester (CSM Instruments Revetest) was employed. The tests were performed by using a 200 µm radius Rockwell diamond stylus, continuously increasing the normal load within the range 1-200 N and applying a load rate of 10 N/mm. During the scratching event, the friction Page 4 of 19

us

cr

ip t

force and the acoustic emission (A.E.) were detected. Two different critical loads, namely LC1 and LC2, were determined. LC1 is defined as the load for which the first cracks close to the scratch are detected. LC2 is defined as the load for which coating delamination begins. The micro-hardness of the coatings was measured by a Durimet micro-hardness tester, applying a load of 2.9 N by means of a pyramidal indenter. Five measurements were taken and the average value was computed, per each test. Crystallographic structure of the samples was investigated by X-ray diffraction (XRD). All the measurements were carried out by using an Italstructures APD2000 diffractometer using copper Kα radiation and a scintillation detector. All the measurements were analyzed using the WINACQ32 software in a 2-θ range between 20° and 80° (with an angle step of 0.02° and a time step of 15 seconds). The search of the phases (Search Match) was performed with the WinSearch software (based on the database of the ICDD), while data processing was performed using the WINDUST32 and Fityk software. Both instrument and software can not separate the contributions of Kα1 and Kα2 (while Kβ is removed using a nickel filter), so the peak parameters were calculated using both Kα1 and Kα2. It has to be underlined that Kα2 contribution is almost negligible at low angles, but it increases when increasing 2-θ.

Ac ce p

te

d

M

an

Samples were investigated by means of XPS analysis (X-ray Photoelectron Spectroscopy), using a Scienta ESCA 200 instrument equipped with a hemispherical analyzer and a monochromatized Al Kα X-ray source (hν = 1486.6 eV). The chamber was kept constantly under ultra-high vacuum at a level of 5·10-9 mbar. Analyses were performed in standard mode (tilt 90°), and samples were fixed with a conductor graphite tape. Each XPS analysis consists of a survey, which is a scan on a wide range of energies (0-1200 eV). It allows identifying all the elements that may be present at the surface of the sample, and it provides detailed spectra of the core levels of the main identified elements. The core levels were acquired with a pass-energy of 150eV, which allows an energy resolution of about 0.4 eV. After spectra acquisition, the detection of different elements of the same core level and their identification in terms of chemical bonds was done by mathematical deconvolution (Shirley type background removal and Lorentzian-Gaussian type components). As far as quantification is regarded, the applied atomic sensitivity factors were provided by Physical Electronics as reported in the Handbook of X-Ray Photoelectron Spectroscopy [20]. The obtainable sensitivity in the quantification of the different components in relative atomic concentration is of the order of 0.5-1%. Each detected component corresponds to a different chemical bond on the surface sample, i.e. the convolution of signals resulting from multiple overlapped chemical bonds. Despite the high energy resolution of the instrument, the peak deconvolution has a certain indetermination, resulting from partial overlapping. Coatings cross sections were analyzed by scanning electron microscopy (SEM) using a ZEISS EVO 50 XVP with LaB6 source, equipped with detectors for secondary and backscattered electrons collection and Energy Dispersion Spectroscopy (EDS) probe for elemental analyses.

2.3. Tribological test. The friction behavior of AlSiTiN and AlSiCrN nanostructured coatings was investigated in sliding tests by means of a CSM high-temperature tribometer, by using a ball-on-disc configuration. The authors have focused their attention on the tribological properties against inox steel, using a ball of 6 mm in diameter. The tests were carried out using a 1 N constant load at a speed of 1 cm/s. Such parameters were chosen with the purpose to prevent wear phenomena. The friction test duration was 3000 laps (150.84 m), and it was done with and without lubricant supply. Lubrication was provided by an oil bath at room temperature. Table 3 shows the composition and the characteristic of the three environmentally-friendly and biologically decomposable lubricants, which were manufactured from non-toxic, renewable, vegetable resources.

Page 5 of 19

us

cr

ip t

3. Results and discussion 3.1. Coating characterization results Fig. 2 shows the tridimensional analysis (obtained by profilometer) of the sample surfaces, while Fig. 3 illustrates the SEM images of the coatings, highlighting the multilayer structure. The thickness of AlSiTiN and AlSiCrN coatings was 3.2 and 4.1 µm, respectively (Table 4). As far as the sum of the layers is concerned (excluding the adhesion and external layers), it ranged between 600 nm and 1 µm for both the coatings, but it was closer to 1 µm for AlSiCrN in most of the regions analyzed by SEM. The thickness assessment by SEM is affected by some uncertainty at this scale level, depending on the planarity of the surface, the level of polishing, and the coating homogeneity. Both coatings significantly increase the surface hardness, as reported in Table 4. The differences between the two multilayers are negligible. As for the Rockwell C test, both coatings had a quite good adhesion, whose classification is HF4. Fig. 4 shows the typical features of super-hard coating structures. Similar cracks are usually found when a high hardness monolayer supports the high normal load required for the test (150 kg). This phenomenon typically occurs in systems where a thin very hard film is deposited on a substrate with a significantly lower hardness [21].

M

an

A further evaluation of the coating adhesion is shown in Fig. 5, where the pictures of the scratch tests and the loads measured at the beginning of the coating removal (LC1) or at the total coating delamination (LC2) are reported. The highest the load values are, the highest the coating adhesion to the substrate is expected to be. The images of the scratch tracks show the same type of failure for both coatings, according to Rockwell C indentation data. At the early beginning, parallel cracks along the rim of the scratch scar and semicircular cracks at the bottom of the scratch channel can be seen. A good adhesion was found for both the coatings. AlSiTiN nanocomposite, showing slightly higher critical loads, led to achieve better outcomes (Fig. 5).

Ac ce p

te

d

Fig. 6 shows the diffraction pattern detected on the AlSiTiN coating. A series of narrow peaks at 31.7°, 35.8°, 48.4°, 64.1°, 73.3°, 75.6° and 77.3° due to the substrate and allocated to orientations 001, 100, 101, 101, 111, 200 and 102 of crystallographic planes of the WC-Co was observed. Apart from the reflections corresponding to the substrate, the pattern shows an intense peak at approximately 42.7°, assigned to 200 reflection of cubic titanium nitride (preferred orientation), and a less intense signal at 62.2°, assigned to 220 reflection of cubic titanium nitride. After profile fitting (omitted for sake of brevity), some other peaks of very low intensity, assigned to reflections of a cubic phase of titanium nitride (c-TiN) and aluminum nitride (c-AlN), reported in the literature as Ti(Al)N, could be observed. A partial segregation of silicon in the matrix of Ti(Al)N with possible formation of amorphous Si3N4 (a-Si3N4) and amorphous TiSi2 (a-TiSi2) can explain the presence of the amorphous component, responsible for the signals broadening. The possible formation of amorphous AlN cannot be completely ruled out.

It is important to remark that lattice parameters may change if the relative concentration change. The introduction of atoms with different atomic radius into the lattice can lead to the expansion or contraction of the lattice itself (peak position shift). Moreover, the incorporation of silicon into the matrix of nanocomposite (Ti, Al) N plays an important role in the crystal structure of the coating. An increase of Si content can lead to a widening of peak shape (caused by the decrease of crystallite size or the residual stress) and a shift of the peak position. High concentration of silicon (~19%) can cause a partial disappearance of the reflections and the pattern tends to be similar to that of an amorphous material. The amorphous or partially amorphous components (broadened peaks and structured background) in the two patterns may be due to the formation of silicon nitrides and/or amorphous aluminum.

Page 6 of 19

us

cr

ip t

Fig. 7 shows the diffraction pattern detected on the AlSiCrN coating deposited on the HM substrate. The four main peaks at 37.7°, 43.6°, 63.3° and 76.4° correspond to reflections 111, 200, 220 and 311 of cubic chromium nitride. Less intense signals at 44.8° and 65.0° were evidenced by fitting (omitted for sake of brevity). The two components correspond to the cubic phase of aluminum nitride. For the case of AlSiCrN coating, studies in literature [22 and references therein] tend to identify such structures as belonging to a more general mix of nitrides of chromium and aluminum. As reported in ICCD cards [11-0065 and 25-1495] the positions of the peaks of cubic AlN and cubic CrN are actually very close, and this may increase the difficulties in data interpretation. The overlapping of the patterns detected on the sample and on the HM substrate indicate that there is no interference between the structures of the two spectra. Only the reflections at higher angle of the substrate have values that are in the 2-θ region of 220 and 311 reflections of the coating. However, they do not belong to the substrate because the most intense reflections of WC-Co are missing, suggesting that such signals can be attributed to the coating.

d

M

an

In Fig. 8 the XPS survey detected on AlSiTiN coating is shown. All the elements on the surface are identified. Peaks attributed to coating (Ti, N, Si, Al) and to surface contamination (C, O and Ca) are observed. The presence of carbon is due to hydrocarburic contamination and oxidized components, since the sample was analyzed as received. In fact, etching process was not performed in order to analyze the same surface used for tribological tests, having contaminations and oxide layers that might deposit on the surface because of the presence of oxidizable metals. The core levels of all the elements is reported in Table 5. For sake of brevity, only Al2p and Si2p profiles resulting from the deconvolution and interpretation of the various components are shown (Fig. 9).

Ac ce p

te

In Table 6, components concerning the PVD coating have been highlighted in bold, thus the others result from the oxide layer and/or contaminations on the surface. The oxide layer is composed of an organic component linked to carbon-based compounds and an inorganic component (MeO) resulting from the oxidation of metals in the upper layer of the nanocomposite. In particular, titanium and aluminum are easier oxidizable while the contribution of silicon oxide is not notable so much. The formation of Ti-O-N bonds (oxide nitrides) is possible. The nanocomposite is composed of a mixture of nitrides: on the surface the presence of aluminum nitride is relevant, less that of titanium and silicon nitride. The correctness of the deconvolution and their assignments for the various components of the coating is based on a cross-check among the different core levels and not only nitrogen core level. Normally, N1s core level detected on a nitride with known chemical composition appears as a simple Gaussian with FWHM of about 1.6-1.9eV. Any enlargement means the presence of added chemical states. In this study, the line shape of N1s is the result of the overlapping of at least three bond components, characterized by a very small chemical shift (often less than 1 eV). So the deconvolution contains some uncertainty.

In Fig. 10, the survey detected on AlSiCrN deposited on HM substrate is shown, while in Table 7 the quantification of the different elements in relative atomic percentage is reported. All the elements on the surface are identified. Peaks attributed to coating (Cr, N, Si, Al) and to surface contamination (C, O and Ca) are observed. The presence of carbon is due to hydrocarburic contamination and oxidized components. In fact, also in this case the sample was analyzed as received for the same reasons briefly above described. Table 8 shows all the components found for the AlSiCrN sample, while Fig. 11 shows the deconvolution of Cr2p, N1s, Al2p and Si2p.

Page 7 of 19

an

us

cr

ip t

The surface oxide layer is composed of (besides an organic component linked to carbon-based compound) an inorganic component (Me-O) resulting from the oxidation of metals present in the upper layer of the nanocomposite, as for AlSiTiN coating. As for this AlSiCrN quaternary coating (even more than for AlSiTiN) the signal of core levels resulting from different components overlaps in terms of binding energy. In literature, many authors do not make any distinctions between the different contributions, especially regarding the study of nitrogen line, which is often fitted by a single peak (because of its high symmetry) attributed to overlap signals deriving from different metallic nitrides. Despite the difficulties, in this analysis a detailed fit of all core levels was made, considering all possible contributions. The nanocomposite is made of a mixture of nitrides, on the surface the presence of aluminum nitride is relevant, while that of chromium and silicon nitride is less relevant. The core level of nitrogen is very complex: oxidized components overlap in separation zone of two pair components. As for aluminum the main peak, symmetrical, resulting from nitride component, has a tailless at high binding energy due to its oxidized component. In addition, it presents a kind of "satellite peak" (more than 4 eV from the main peak), whose attribution is doubtful. However, a study on AlCrN coating attributes such peak to Cr3s. This evidence could be consistent also with the actual case [23].

Ac ce p

te

d

M

3.2. Friction results. Figs. 12 and 13 show the evolution of the coefficient of friction (COF) of coatings and HM substrate sliding against the inox ball in dry and lubricated conditions, respectively. As far as the coatings are concerned, the COF show the same values at the early beginning of the sliding contact. After few laps, AlSiTiN COF slightly decreases and rapidly stabilizes, while the one for AlSiCrN increases and becomes almost stable after about 900 laps. Friction coefficient depends on many aspects, like particles size, as reported by Senda et al. [24], who observed that a slight COF decrease occurs when reducing ceramic crystallite dimensions from 11.6 to 3.4 µm. The crystalline phases have also an important influence on friction values, as it can be observed in the paper of Martin et al. [25]. Furthermore, temperature and tribofilm formed as a consequence of sliding contact between two surfaces is of paramount importance [14]. Despite the phenomenon complexity and the lack of literature dealing with the correlation between microstructure and friction properties of nanocomposites-inox systems, some hypothesis devoted to explain the friction behavior can be drawn. The similarity of the surface composition of the coatings, in contact with the inox ball at the beginning of the test, can account for the first instants of friction test behaviour. The external layer of both coatings contains a similar percentage of oxidized silicon, which was found to affect significantly the friction properties of nanocomposites. COF is reduced when increasing the silicon percentage [26], because of the formation of SiOx self-lubricating films. In the actual case, silicon percentage is similar for both the coatings (2.5 % and 2.8 % for AlSiTiN and AlSiCrN, respectively), so that the self-lubricating effect is expected to be the same. As for the other elements on the external layer, a mix of various nitrides and oxides are present, together with hydrocarbons. The main differences between the two coatings composition (consisting of TiN and CrN in AlSiTiN and AlSiCrN, respectively) are reduced on the surface because of the formation of oxidized phases, consisting of Al2O3, TiOx, CrOx, and the wide amount of organic phases. For that reason, no significant differences are expected in terms of friction, according to the actual results. After few laps, the external layer of few nanometers observed by XPS, is likely destroyed as a consequence of the sliding contact of inox ball, resulting in a diversification in friction because of the bulk differences between the coatings, as shown by diffractometry. Interestingly, AlSiTiN shows a significantly lower and more stable COF with respect to AlSiCrN and uncoated material.

Page 8 of 19

cr

ip t

The instability of friction of CrN films is reported in [25]: the authors attributed such a kind of behavior to a plowing action of the hard coating debris generated by the coating failure. In the present study, a very similar COF trend is clearly visible for the coating containing CrN. Due to the higher friction observed for AlSiCrN, a flash temperature significantly higher than R.T. can be expected and, then, a limited phase transformation can not be excluded. More likely, FeOx, can be expected to form as tribofilm, which usually leads to a significant increase of COF [27]. As for the effect of TiN nanoparticles, some studies present friction results of nitrides-based coatings [28], showing how an increase of aluminum content results in a friction increase. Furthermore, TiN phase reduces COF if compared to ternary nitrides coatings, also including AlCrN. As for the uncoated material, its friction starts from lower values with respect to coated ones and reaches values similar to AlSiCrN at the end of the test, suggesting again that the flash temperature is significantly high in this case, leading to higher wear and phase transformations, including soldering of material coming from the steel, leading to the formation of oxidized iron.

M

an

us

Tests under lubricated conditions (applying MQL media) show that WC material presents the lowest friction coefficient and that the type of lubricant does not affect the results. Furthermore, the lubricant allows a rapid friction stabilization for all the materials. These results are in agreement with the above mentioned hypothesis, made when describing friction evolution in dry condition. In fact, lubricants act also as a cooling medium, preventing both a temperature increase and the consequent formation of tribofilm, so allowing to keep almost the same friction coefficient for the whole test duration. This hypothesis is supported by the fact that friction values in lubricated conditions are very similar to those observed at the beginning of the test in dry condition for all the materials (almost 0.2 for uncoated and 0.4 for coated materials).

Ac ce p

te

d

4. Conclusions AlSiTiN and AlSiCrN nanocomposite coatings were deposited by Physical Vapour Deposition technique. Coatings were analysed in terms of hardness and adhesion. Their structure was assessed by XRD, and the surface composition was studied by XPS. Friction tests were carried out. The results can be summarized as follows: both coatings show quite good adhesion to the substrate, and comparable high hardness; XRD analysis for AlSiTiN coating reveals a preferred orientation of cubic titanium nitride (200), and a partial segregation of silicon in the matrix of Ti,(Al)N with possible formation of a-Si3N4 and a-TiSi2; XRD analysis on AlSiCrN coating reveals a cubic phase of CrN (200), very close to AlN, as principal structure; a mixture of nitrides, oxides, and organic contaminations are present on the external layer of both coatings, so that the differences arising from bulk composition (TiN and CrN) are reduced on the surface; such poor differences are responsible for the similar behaviour observed at the beginning of the sliding test using inox steel as counterface. Friction coefficient evolution significantly changes after few minutes of tribo test. The diversification is attributable to bulk materials properties, particularly to the presence of CrN in AlSiCrN, making the COF higher and significantly more instable with respect to that observed for AlSiTiN; the presence of lubricants allows keeping the COF low and stable for all the materials, including WC-Co, as a consequence of the cooling effect preventing both phase transformation and soldering.

Acknowledgments Page 9 of 19

This publication was made possible by NPRP grant nr. NPRP 5-423-2-167 from the Qatar National Research Fund (a member of Qatar Foundation). The statements made herein are solely the responsibility of the authors.

cr

ip t

References [1] S. Veprek, The search for novel, superhard materials, Journal of Vacuum Science & Technology A 17 (1999) 2401-2420. [2] S. Zhang, W. Wu, W. Chen, S. Yang, Structural optimisation and synthesis of multilayers and nanocomposite AlCrTiSiN coatings for excellent machinability, Surface & Coatings Technology 277 (2015) 23-29.

us

[3] D. Yu, C. Wang, X. Cheng, F. Zhang, Microstructure and properties of TiAlSiN coatings prepared by hybrid PVD technology, Thin Solid Films 517 (2009) 4950-4955.

an

[4] D. Philippon, V. Godinho, P.M. Nagy, M.P. Delplancke-Ogletree, A. Fernández, Endurance of TiAlSiN coatings: Effect of Si and bias on wear and adhesion, Wear 270 (2011) 541-549.

Ac ce p

te

d

M

[5] C.-H. Lin, Y.-Z. Tsai, J.-G. Duh, Effect of grain size on mechanical properties in CrAlN/SiNx multilayer coatings, Thin Solid Films 518 (2010) 7312-7315. [6] J. Soldán, J. Neidhardt, B. Sartory, R. Kaindl, R. Čerstvý, P.H. Mayrhofer, R. Tessadri, P. Polcik, M. Lechthaler, C. Mitterer, Structure-property relations on arc-evaporated Al-Cr-Si-N coatings, Surface & Coating Technology 202 (2008) 3555-3562. [7] X.-Z. Ding, X.T. Zeng, Y.C. Liu, Structure and properties of CrAlSiN Nanocomposite coatings deposited by lateral rotating cathod arc, Thin Solid Films 519 (2011) 1894-1900. [8] W.-Y. Ho, C.-H. Hsu, C.-W. Chen, D.-Y. Wang, Characteristics of PVD-CrAlSiN films after post-coat heat treatments in nitrogen atmosphere, Applied Surface Science 257 (2011) 3770-3775. [9] H.-W. Chen, Y.-C. Chan, J.-W. Lee, J.-G. Duh, Oxidation behavior of Si-doped nanocomposite CrAlSiN coatings, Surface & Coatings Technology 205 (2010) 1189-1194. [10] C. Tritremmel, R. Daniel, M. Lechthaler, P. Polcik, C. Mitterer, Influence of Al and Si content on structure and mechanical properties of arc evaporated Al-Cr-Si-N thin film, Thin Solid Films 534 (2013) 403-409. [11] S. Zhang, L. Wang, Q. Wang, M. Li, A superhard CrAlSiN superlattice coating deposited by a multi-arc ion plating: II. Thermal stability and oxidation resistance, Surface & Coatings Technology 214 (2013) 153-159. [12] T. Polcar, A. Cavaleiro, High temperature properties of CrAlN, CrAlSiN and AlCrSiN coatings - Structure and oxidation, Materials Chemistry and Physics 129 (2011) 195-201. [13] T. Polcar, A. Cavaleiro, Structure and tribological properties of AlCrTiN coatings at elevated temperature, Surface & Coatings Technology 205 (2011) S107-S110. [14] T. Polcar, A. Cavaleiro, High-temperature tribological properties of CrAlN, CrAlSiN and AlCrSiN coatings, Surface & Coatings Technology 206 (2011) 1244-1251. [15] L. Settineri, M.G. Faga, G. Gautier, M. Perucca, Evaluation of wear resistance of AlSiTiN and AlSiCrN nanocomposite coatings for cutting tools, CIRP Annals - Manufacturing Technology 57 (2008) 575-578. [16] M.G. Faga, G. Gautier, R. Calzavarini, M. Perucca, E. Aimo Boot, F. Cartasegna, L. Settineri, AlSiTiN nanocomposite coatings developed via Arc Cathodic PVD: Evaluation of wear resistance via tribological analysis and high speed machining operations, Wear 263 (2007) 1306-1314. [17] P.C. Priarone, S. Rizzuti, L. Settineri, G. Vergnano, Effects of cutting angle, edge preparation, and nano-structured coating on milling performance of a gamma titanium aluminide, Journal of Materials Processing Technology 212 (2012) 2619-2628.

Page 10 of 19

us

cr

ip t

[18] M. Stueber, H. Holleck, H. Leiste, K. Seemann, S. Ulrich, C. Ziebert, Concepts for the design of advanced nanoscale PVD multilayer protective thin films, Journal of Alloys and Compounds 483 (2009) 321-333. [19] K. Weinert, I. Inasaki, J.W. Sutherland, T. Wakabayashi, Dry Machining and Minimum Quantity Lubrication, CIRP Annals - Manufacturing Technology 53/2 (2004) 511-537. [20] Handbook of X-Ray Photoelectron Spectroscopy (1995), Physical Electronics Inc., USA. [21] R.M. Souza, A. Sinatora, G.G.W. Mustoe, J.J. Moore, Numerical and experimental study of the circular cracks observed at the contact edges of the indentations of coated systems with soft substrates, Wear 251 (2001) 1337-1346. [22] T. Polcar, T. Vitu, J. Sondor, A. Cavaleiro, Tribological Performance of CrAlSiN Coatings at High Temperatures, Plasma Processes and Polymers 6/S1 (2009) S935-S940. [23] J.L. Endrino, G.S. Fox-Rabinovich, A. Reiter, S.V. Veldhuis, R. Escobar Galindo, J.M. Albella, J.F. Marco, Oxidation tuning in AlCrN coatings, Surface & Coatings Technology 201 (2007) 4505-4511.

Ac ce p

te

d

M

an

[24] T. Senda, E. Yasuda, M. Kaji, R.C. Bradt, Effect of Grain Size on the Sliding Wear and Friction of Alumina at Elevated Temperatures, Journal Of The American Ceramic Society 82/6 (2004) 1505-1511. [25] C. Lorenzo-Martin, O. Ajayi, A. Erdemir, G.R. Fenske, R. Wie, Effect of microstructure and thickness on the friction and wear behavior of CrN coatings, Wear 302 (2013) 963-971. [26] I.-W. Park, S.R. Choi, J.H. Suh, C.-G. Park, K.H. Kim, A superhard CrAlSiN superlattice coating deposited by multi-arc ion plating: I. Microstructure and mechanical properties, Surface & Coatings Technology 214 (2013) 160-167. [27] C.-W. Cho, Y.-Z. Lee, Effects of oxide layer on the friction characteristics between TiN coated ball and steel disk in dry sliding, Wear 254 (2003) 383-390. [28] L. Aihua, D. Jianxin, C. Haibing, C. Yangyang, Z. Jun, Friction and wear properties of TiN, TiAlN, AlTiN and CrAlN PVD nitride coatings, International Journal of Refractory Metals and Hard Materials 31 (2012) 82-88.

Fig. 1 Typical architecture of AlSiTiN and AlSiCrN coatings. Fig. 2 From the top to the bottom: tridimensional analysis of substrate, AlSiCrN coating, and AlSiTiN coating. Fig. 3 SEM images of AlSiTiN (left) and AlSiCrN (right) coatings. Fig. 4 Rockwell imprints on AlSiTiN (left) and AlSiCrN (right) coatings. Fig. 5 Scratch test main results. Fig. 6 XRD patterns detected on AlSiTiN. Fig. 7 XRD patterns detected on AlSiCrN deposited on HM substrate. Fig. 8 AlSiTiN/HM sample: XPS survey. Fig. 9 AlSiTiN: Al2p (a) and Si2p (b) core levels. Fig. 10 AlSiCrN/HM sample: XPS survey. Fig. 11 AlSiCrN: Cr2p (a), N1s (b), Al2p (c), Si2p (d) core levels. Fig. 12 Coefficient of friction (COF) under dry conditions. Fig. 13 Coefficient of friction (COF) under lubricated conditions. Table 1 Composition of substrate material. Element

WC

Binder

TiC+TaC(NbC)

vol. %

85 ± 3

9±3

6±2

Page 11 of 19

Table 2 Chemical and physical properties of substrate material (provided by the supplier). Apparent Hardness Density T.R.S. Metallographic average grain (HRA) (gr/cm3) (N/mm2) Structure size (µm) 91.2

13.6

1700

Uncombined carbon (%)

≤1

Fine

0.06

Density (g/cm³) at

Kinematic viscosity

15°C

(mm²/s) at 40°C

natural, highly refined triglycerides

0.92

35

LB5000

fatty alcohol

0.84

18

LB10000

natural, refined triglycerides

0.92

35

LB2000

> 300

> 160

> 250

an

Table 4 Coating thickness, Rockwell C and hardness results. Sample

Thickness (µm)

Flash point (°C)

cr

Composition

us

Lubricant

ip t

Table 3 Accu-Lube lubricant specifications.

Rockwell C

Vickers Hardness

-

-

1400 ± 23

AlSiTiN

3.2

HF4

3800 ± 12

AlSiCrN

4.1

HF4

3750 ± 18

M

WC-Co

Table 5 Quantification of the different elements in relative atomic percentage. C1s (at.%) O1s (at.%) Al2p (at.%) Si2p (at.%)

Ti2p (at.%)

N1s (at.%)

AlSiTiN/HM

7.74

21.32

Conc. (at.%) 21.14 4.63 2.21 3.03 6.90 10.39 2.44 0.51 3.07 1.79 0.50 1.46 0.46 15.94 1.43 3.69 0.31 15.15 2.33 2.40 0.20

Phases C-C / C-H COx COx COx MeO COx COx / SiOx (SiO2) TiN (TiNx) Ti-O-N/TiyOx TiOx

17.42

d

19.94

te

31.12

Ac ce p

Table 6 Chemical composition obtained for the sample AlSiTiN on HM. Core line Area BE (eV) Height FWHM 622.95 285.00 24.56 1.19 C1s001 136.37 285.89 5.84 1.10 C1s002 64.97 286.90 2.11 1.44 C1s003 89.34 288.62 2.15 1.98 C1s004 488.42 530.13 14.30 1.60 O1s001 735.08 531.93 18.21 1.90 O1s002 172.75 534.05 3.68 2.21 O1s003 90.69 455.17 3.39 1.20 Ti2p001 549.20 456.68 18.78 1.37 Ti2p002 321.12 458.24 8.72 1.73 Ti2p003 90.25 461.22 2.21 1.92 Ti2p004 260.78 462.80 6.42 1.87 Ti2p005 82.95 464.40 2.09 1.82 Ti2p006 756.69 396.24 27.99 1.27 N1s001 68.04 396.94 3.96 0.81 N1s002 175.40 397.40 6.33 1.30 N1s003 14.95 398.68 0.59 1.25 N1s004 291.00 73.50 10.14 1.35 Al2p001 44.81 74.43 1.59 0.99 Al2p002 67.70 101.33 1.95 1.64 Si2p001 5.66 103.83 0.25 1.06 Si2p002

2.46

RSF 0.296 0.296 0.296 0.296 0.711 0.711 0.711 1.798 1.798 1.798 1.798 1.798 1.798 0.477 0.477 0.477 0.477 0.193 0.193 0.283 0.283

AlN TiN (TiNx) SixNy Ti-N-O AlN Al2O3 SixNy SiOx (SiO2)

Table 7 Quantification of the different elements in relative atomic percentage. C1s (at.%) O1s (at.%) Al2p (at.%) Si2p (at.%)

Cr2p (at.%)

N1s (at.%)

AlSiCrN/ HM

8.60

24.38

20.62

19.07

24.52

2.81

Page 12 of 19

Phases C-C / C-H COx COx COx CrO AlO COx COx / SiOx (SiO2) CrxNy CrN CrxOy

cr

ip t

Conc. (at.%) 15.78 1.88 1.02 1.89 3.48 6.00 7.79 1.68 1.49 2.94 1.46 0.68 1.23 0.82 4.90 18.01 1.78 18.26 3.61 2.55 2.25 0.50

us

RSF 0.296 0.296 0.296 0.296 0.711 0.711 0.711 0.711 2.201 2.201 2.201 2.201 2.201 2.201 0.477 0.477 0.477 0.193 0.193 0.193 0.283 0.283

CrN AlN CrxNy / SixNy AlN AlxOy (Al2O3) uncertain attribution SixNy SiOx (SiO2)

te

Ac ce p

Total coating thickness = T

Top coating, different stoichiometry, thickness = 0.5 T

d

M

an

Table 8 Chemical composition obtained for the sample AlSiCrN on HM. Core line Area BE (eV) Height FWHM 2537.65 285.00 112.69 1.06 C1s001 302.77 286.08 13.44 1.06 C1s002 164.34 287.01 7.30 1.06 C1s003 303.46 288.60 7.99 1.81 C1s004 1342.96 530.59 34.79 1.81 O1s001 2316.61 531.35 57.30 1.90 O1s002 3008.68 532.16 86.92 1.63 O1s003 649.42 533.51 15.27 2.00 O1s004 1779.22 574.61 68.69 0.82 Cr2p001 3515.51 575.70 92.58 1.65 Cr2p002 1746.12 577.02 29.97 2.49 Cr2p003 813.10 584.35 21.31 1.68 Cr2p004 1467.02 585.05 32.88 1.70 Cr2p005 981.33 586.20 14.49 3.17 Cr2p006 1270.95 396.20 31.58 1.89 N1s001 4665.60 396.53 158.66 1.38 N1s002 461.53 397.66 11.68 1.87 N1s003 1914.52 73.66 57.73 1.56 Al2p001 378.22 74.77 8.93 1.82 Al2p002 267.58 77.95 3.81 3.00 Al2p003 346.06 101.30 11.66 1.40 Si2p001 76.91 102.15 3.25 1.11 Si2p002

Multilayer, n.5 alternate A+B layers, fixed stoichiometry, thickness = 0.5 T

Top coating

Multilayer step A Multilayer step B

Gradient layer Adhesion layer Metal substrate

Page 13 of 19

µm 1.5 1.4 1.3 1.2 WC-Co

1.1

ip t

1.0 0.9 0.8

cr

0.7 AlSiCrN

0.6

0.4 0.3 0.2 AlSiTiN

an

0.1

us

0.5

Ac ce p

te

d

M

0.0

AlSiTiN

1 µm

AlSiTiN

AlSiCrN

1 µm

AlSiCrN

Page 14 of 19

20

25

30

35 50

40

AlSiTiN

d

te

AlSiCrN

cr

Lc1 Lc2

50 µm 50 µm

100 Load (N)

50 55 60 2-Theta (deg)

45

ip t

50 µm

us

an

M

AlSiTiN 50 µm

TiN 220 WC (101) / AlN (220)

WC (101)

AlSiCrN

Lc2

150

65

WC (111) WC (200) WC (102)

TiN 200

0

AlN 200

WC (100)

Ac ce p

AlN (111) / TiN (111)

WC (001)

Intensity (a.u.)

Lc1

70 Lc2

Lc1

200

75

80

Page 15 of 19

Cr(Al)N 111

Cr(Al)N 220

25

30

35

40

45 50 55 60 2-Theta (deg)

75

80

M

500 O

70

65

an

20

us

Cr(Al)N 311

cr

ip t

Intensity (a.u.)

Cr(Al)N 200

N

d

Ti2p Ti2s

te

300 O KLL

200

Na1s

Ac ce p

Intensity (cps)

400

Si2p

C

Si2s

Al2p ? Ca

100

1200

1000

800

600

Ti3p Al2s

400

200

0

Binding Energy (eV)

Page 16 of 19

20

11.0

(a)

Al2p

(b)

Si2p 10.5

Intensity (cps)

10

10.0 9.5 9.0

5 8.5

75

74

73

72

104

3500 O1s

3000

N1s

Cr2p

O KLL Cr2s

M

1500

Al2s

?Sc2s

d

1000

1200

Si2s

te

Ca2p

500

1000

800

101

100

C1s

Cr KLL

Ac ce p

Intensity (cps)

Na1s

2000

102

an

Zn2p

2500

103

Binding Energy (eV)

us

Binding Energy (eV)

cr

76

ip t

Intensity (cps)

15

600

400

Al2p

Cr3p

Si2p

200

0

Binding Energy (eV)

Page 17 of 19

200

250

Cr2p

N1s

(a)

(b)

200

100

50

150

100

50

585

580

575

399

398

Al2p

Si2p

(c)

20

78

76

74

AlSiCrN

103

AlSiTiN

102

101

100

Binding Energy (eV)

WC-Co Uncoated

Ac ce p

1.2

25

72

te

Binding Energy (eV)

30

d

80

Coefficient of friction

(d)

an

Intensity (cps)

M

Intensity (cps)

40

0.2

394

35

60

0.4

395

40

80

0.6

396

Binding Energy (eV)

us

Binding Energy (eV)

0.8

397

cr

590

1.0

ip t

Intensity (cps)

Intensity (cps)

150

0 0

500

1000

1500

2000

Laps

Page 18 of 19

AlSiTiN

AlSiCrN 1.2

1.2

1.2 Lubricant: Accu-Lube LB2000

Lubricant: Accu-Lube LB5000

0.4

0.8 0.6 0.4

1000

1500

2000

0

500

1500

2000

Laps

0

500

1000

1500

2000

Laps

d

M

an

us

Laps

1000

cr

500

te

0

0.4

0

0

0

0.6

0.2

0.2

0.2

0.8

ip t

0.6

Coefficient of friction

Coefficient of friction

0.8

Ac ce p

Coefficient of friction

Lubricant: Accu-Lube LB10000

1.0

1.0

1.0

WC-Co Uncoated

Page 19 of 19