Effect of the type, size and concentration of solid lubricants on the tribological properties of the polymer PEEK

Wear 364-365 (2016) 31–39

Contents lists available at ScienceDirect

Wear journal homepage: www.elsevier.com/locate/wear

Effect of the type, size and concentration of solid lubricants on the tribological properties of the polymer PEEK M. Zalaznik a,b, M. Kalin a,n, S. Novak c,d, G. Jakša e Laboratory for Tribology and Interface Nanotechnology, Faculty of Mechanical Engineering, University of Ljubljana, Bogišičeva 8, 1000 Ljubljana, Slovenia Pladent d.o.o., Lokarje 19, 1217 Vodice, Slovenia c Department for Nanostructured Materials, Jožef Stefan Institute, Jamova cesta 39, 1000 Ljubljana, Slovenia d Jožef Stefan International Postgraduate School, Jamova cesta 39, 1000 Ljubljana, Slovenia e Department of Surface Engineering and Optoelectronics, Jožef Stefan Institute, Jamova cesta 39, 1000 Ljubljana, Slovenia a

b

art ic l e i nf o

a b s t r a c t

Article history: Received 4 April 2016 Received in revised form 1 June 2016 Accepted 11 June 2016 Available online 17 June 2016

Poly-ether-ether-ketone (PEEK) is a high-performance, temperature-resistance polymer that is finding an increasing range of applications. In order to even enhance PEEK's mechanical and tribological properties particles of different compositions, shapes and sizes are added into its matrix. PEEK has already been combined with many different particles; however, very rarely MoS2 and WS2 – two state-of-the-art solidlubricants – were added into the PEEK matrix. Furthermore, a comprehensive tribological study combining MoS2 and WS2 particles of different sizes and concentrations has not yet been reported. In this investigation we looked at the effect of micro- and nanosized MoS2 and WS2 particles in PEEK on the drysliding tribological behaviour against stainless steel (100Cr6). A non-conventional technique, i.e., the sintering of dry-pressed compacts, was used to prepare the PEEK composites. The results show that all the particles, irrespective of their composition and size, reduce the friction (up to 30%); however, the nanoscale particles require a higher concentration to form an effective low-friction tribofilm. The formation of a tribofilm is necessary to reduce the wear of all the composites (up to 51%); this is strongly promoted by the addition of nano- or microparticles of both the MoS2 and WS2 materials. In addition, the hardness, which is greatly increased by the addition of all the particles, significantly improves the wear behaviour. The results of XPS analyses showed that the oxidation occurs during tribological sliding, which reduces particles beneficial wear behaviour effects. & 2016 Elsevier B.V. All rights reserved.

Keywords: Polymer-matrix composite Solid lubricants Surface analysis Sliding friction Sliding wear

1. Introduction Polymers are taking an increasingly important position in almost every branch of industry; they are being used as the main material for a final product or simply as a replacement for other materials (e.g., carbon steel, stainless steel, titanium, aluminium, magnesium, brass, bronze, etc.). High-performance engineering plastics are used for the most demanding applications, such as gears, sealing rings, bushings, bearings, valves, etc. These polymers are required to withstand harsh operating conditions with high loads, high temperatures and long operating hours. Often, these polymers are considered in non-lubricated applications, where either the operating conditions and environmental demands or the nature of the application itself (e.g., sterility) restricts the use of lubrications. For such applications, the polymer itself, albeit with some difficulty, fulfils the required demands, especially when n

Corresponding author. Tel.: þ 386 1 4771 460; fax: þ 386 1 4771 469. E-mail address: [email protected] (M. Kalin).

http://dx.doi.org/10.1016/j.wear.2016.06.013 0043-1648/& 2016 Elsevier B.V. All rights reserved.

the application in question includes dry-sliding movements. Polyether-ether-ketone (PEEK) is a high-performance polymer with good mechanical and promising tribological properties. However, PEEK's coefficients of friction in dry-sliding contacts can reach values that are not desired for such conditions and can promote higher wear and premature material failure. Therefore different particles are added into PEEK's matrix in order to improve its tribological and mechanical properties [1–7]. In non-lubricated environments, the performance of the materials mostly depends on the properties of the interfaces and surfaces, and their ability to form anti-wear and low-shear boundary films that in lubricated conditions are usually provided by lubricants and additives. This can also be achieved by using appropriate self-lubricating particles, such as molybdenum disulphide (MoS2) and tungsten disulphide (WS2). These two solid lubricants are well known for their low-friction properties and wear resistance [8–11], but they have not yet been extensively used in polymer composites, especially in combination with PEEK. A literature review of PEEK composites indicated the importance of the particle size, the concentration and the transfer-film

32

M. Zalaznik et al. / Wear 364-365 (2016) 31–39

formation of the surfaces of the counterparts [6,12–14]. Also, the formation of a transfer film on the counterpart surfaces is believed to be the key to a low coefficient of friction and wear rate [15–17]. However, this information is still missing for the suggested MoS2 and WS2 particles in PEEK. Therefore, in this study we have used two different sizes – nano and micro – of MoS2 and WS2 particles, and four different concentrations. With different tribological, mechanical, chemical and optical analyses we were able to thoroughly assess the effect of different solid lubricants (MoS2/WS2), their size and concentration on the properties of the PEEK. Special attention was given to the transfer-film formation with respect to the particle type, size and concentration.

Table 1 Compositions and denotations of the samples. Particle type

Concentration, wt%

Denotation

Pure PEEK MicroMoS2 NanoMoS2 MicroWS2 NanoWS2

– 0.5; 0.5; 0.5; 0.5;

PEEK M0.5; M1; M2; M5 NanoM0.5; nanoM1; nanoM2; nanoM5 W0.5; W1; W2; W5 NanoW0.5; nanoW1; nanoW2; nanoW5

1; 1; 1; 1;

2; 2; 2; 2;

5 5 5 5

to the surface roughness Ra¼0.030 70.005 μm (T8000, Hommelwerke GmbH, Schwenninger, Germany). 2.2. Tribological tests

2. Experimental 2.1. Materials and preparation All the samples used in this investigation were based on a polyether-ether-ketone (PEEK) matrix obtained from the company s Victrex (VICOTE 704, d50¼8.5 μm, Victrex plc., Thornton Cleveleys, UK). For the composite samples four different particles were used: microMoS2, nanoMoS2, microWS2 and nanoWS2. The microsized molybdenum disulphide powder was obtained from Sigma Aldrich (Sigma-Aldrich, St. Louis, USA). The average particle size was around 2 μm, with a purity of 99%. The nanosized molybdenum disulphide powder was obtained from Graphene Supermarket (Graphene Supermarket Co., USA). The average particle size was  90 nm, with a purity of 99%. The microsized tungsten disulphide powder was also obtained from Graphene Supermarket (Graphene Supermarket Co., USA). The particle size was in the range 0.4–1.0 μm, with a purity of 99%. The nanosized tungsten disulphide particles were obtained from the company NanoMaterials (NanoMaterials, Ltd., Yavne, Israel) in a fullerenelike shape. The fullerene-like WS2 had a uniform, symmetrical, spherical structure of 20–100 layers with a diameter of 30–70 nm. The particles were used in four different concentrations, i.e., 0.5 wt%, 1 wt%, 2 wt% and 5 wt%, and were compared to the pure PEEK sample. The samples were prepared by dry pressing the powder, followed by sintering. This procedure proved to be both time and cost efficient, and also showed a high potential for producing homogenous composite materials with a high flexibility of both the material and the processing parameters [18,19]. First, the pure PEEK powder and the different concentrations of particles were ultrasonically and magnetically mixed in ethanol for 5 min to achieve a homogenous suspension. The suspensions were dried (80–83 °C) to remove the ethanol and the powder was pressed in a disc-like shape with a diameter of 15 mm at a pressure of 100 MPa (PW 10, Paul-Otto Weber GmbH, Remshalden, Germany). The samples were inserted into a tube furnace (IJS, Slovenia) and heated at a temperature of 300 °C for 60 min, with a heating and cooling rate of 5 °C/min. Disc-like composite samples were embedded in universal embedding resin (Technovit 4071, Heraeus Kulzer GmbH, Germany) and, prior to the testing, were polished to a surface roughness of Ra¼0.030 70.005 μm (T8000, Hommelwerke GmbH, Schwenninger, Germany). The compositions of the samples and their denotations are presented in Table 1. For the counterpart material used in our the tribological tests a DIN 100Cr6 stainless steel was chosen, since it is a common material used for various mechanical components (bearings, gears, etc.) and it also commonly used in other tribological investigations. The counterpart sample was a flat-end cylinder with a diameter of 3 mm (Tinex, trgovska družba d.o.o., Šenčur, Slovenia). Before the tribological testing the cylinder surfaces were polished

Since the test conditions can also have a major effect on the tribological properties [20,21] only one set of test conditions was used in order to focus only on the concentration, size and particles type. The test parameters were determined with preliminary tests and were kept the same throughout entire investigation, including our other work [18,19,22]. The tribological tests were performed on a CETR tribological tester (UMT-2, now Bruker, CA, USA) in a pin-on-disc configuration. The upper specimen (i.e., the pin) was a steel cylinder and the lower specimen (i.e., the disc) was a polymer sample. Tribological tests were conducted in a reciprocating sliding regime, with a stroke of 5 mm. The upper cylinder was loaded with a pressure of 1 MPa and the sliding velocity was 0.05 m/s. The test duration was 7 h with an overall sliding distance of 1260 m. The test duration was long enough to ensure a steady-state coefficient of friction for all the tested samples. All the tribological sliding tests were conducted in a dry-sliding contact (without additional lubrication) at 20 °C, RH¼ 307 5%. The tests on each sample were performed at least four times to ensure a relevant statistical evaluation. The presented values for the coefficient of friction show an average value of the steady-state coefficients of friction for each sample, with the corresponding standard deviations. 2.3. Characterisation and analyses Prior to the tribological testing, all the samples were characterised in terms of the particle distribution in the PEEK matrix and the hardness. The distribution of particles was evaluated with mapping-mode energy-dispersive X-ray spectroscopy (EDS). A scanning electron microscope (SEM; JEOL Ltd., Tokyo, Japan) combined with a light-element Si energy-dispersive X-ray detector (beryllium-window type EDS; Oxford Inst., Abingdon, UK), operated at an accelerating voltage of 20 kV was used for the evaluation. The Vickers hardnesses of all the samples were measured on a Miniload 2 hardness tester (Leitz Miniload, Wild Leitz GmbH, Wetzlar, Germany). The hardness was measured with a standard Vickers indenter, with a load of 491 mN for an indentation loading time of 15 s. The average value of 20 measurements and their standard deviation are presented. The wear rates and the corresponding 2D cross-sections were measured with white-light optical interferometry (Contour GT-K0, Bruker, Billerica, Massachusetts). The graphs show the average values of the normalised wear rates with the corresponding standard deviations. The worn surfaces were inspected with an optical microscope (Eclipse LV-150, Nikon, Tokyo, Japan) using 25  magnification and with a scanning electron microscope (SEM; JEOL JSM-T330A, JEOL Ltd., Tokyo, Japan) operated with an accelerating voltage of 10 kV and equipped with an Inca Energy data-processing unit (Oxford Inst., Analytical Ltd., Abingdon, UK). Representative worn surfaces are presented.

M. Zalaznik et al. / Wear 364-365 (2016) 31–39

Transfer films were evaluated in terms of their formation on the steel counterparts and thickness. Optical microscopy was used for the investigation of the counterpart surfaces after the tribological tests. Based on these images, the surface coverages of the transfer films were calculated with the open-source NIH software ImageJ. Firstly, the entire counterpart surface was calculated as a reference value. Secondly, with the manual settings of the threshold, only the transfer film was selected and its surface was calculated. Based on the ratio between the reference counterpart value and the transfer-film value the surface coverage was calculated. The surface coverage is presented as the percentage of covered surface. For the evaluation of the transfer-film thickness on the steel counterparts, white-light optical interferometry was used. The methodology for the transfer-film thickness evaluation was based on the surface-roughness parameter Sz (peak-to-valley) measured on the 3D topographies of the steel counterparts, recorded with 25  and 50  magnifications. The results show the average value of 15 measurements of the Sz parameters with the corresponding standard deviations. For the estimation of the oxidation, X-ray photoelectron spectroscopy (XPS) analyses were performed on the surfaces of selected polymer composites before the tribological testing and their steel counterparts after the tribological tests. Only samples with the highest concentration (5 wt% of particles) were analysed – M5, nanoM5, W5 and nanoW5. The XPS analyses were made using a PHI-TFA XPS spectrometer (Physical Electronics Inc., USA) with a monochromatic Al X-ray radiation source. The analysed area was 0.4 mm in diameter and the analysed depth was approximately 3– 5 nm. The analyses took place in an ultra-high vacuum, which was approximately 10  7 Pa during the analysis. The peaks C 1s, O 1s, Mo 3d, S 2p and W 4f were identified in the acquired spectra. From the XPS spectra the surface composition was calculated by dividing the peak intensities by the relative sensitivity factors provided by the manufacturer of the XPS spectrometer [23]. Prior to the spectra processing, the spectra were referenced to the C–C/C–H peak in the C 1s spectrum at a binding energy of 284.8 eV.

33

3. Results 3.1. Material characterisation Before the tribological tests the distribution of the particles in the PEEK matrix was evaluated with EDS mapping. The results of the analyses for the lowest (0.5 wt%) and the highest (5 wt%) concentrations for micro/nanoMoS2 are presented in Fig. 1. Similar results were obtained for micro/nanoWS2 particles. The MoS2 and WS2 particles were relatively well distributed in the PEEK matrix; however, the presence of agglomerates is obvious, in particular for the composites with the highest concentrations of microsized particles. Comparing the hardness behaviour shown in Fig. 2, all the composites, except the microWS2/PEEK combination, exhibited a 12–27% hardness increase at low concentrations (0.5 wt% and 1 wt%). From the 2 wt% of particles onwards, the hardness values decreased for both particle types (MoS2 and WS2) and sizes (micro and nano). In all cases the lowest hardness was achieved at the highest concentration of particles (5 wt%), with the samples nanoM5, W5 and nanoW5 having values lower than the PEEK matrix. The microMoS2/PEEK composites were the least affected by the particle concentration and were also the only composites with hardness values higher than that of the PEEK, irrespective of the concentration of particles. On the other hand, the hardnesses of the microWS2/PEEK composites were, for all concentrations, lower than that of the PEEK and were gradually decreasing with a higher concentration. 3.2. Tribological tests The friction behaviour was found to be comparable for all the composites; with an increasing concentration the coefficients of friction were reduced and the lowest coefficient of friction was, in all cases, achieved at the highest concentration (Fig. 3). Based on the particle size, the most notable differences were observed at concentrations below 2 wt%. The composites filled with microsized particles exhibited a significant decrease in the coefficient of friction already for concentrations of 0.5–1 wt%, while a higher

50 μm

50 μm

50 μm

50 μm

Fig. 1. EDS mapping of the molybdenum for the PEEK composites: (a) M0.5, (b) M5, (c) nanoM0.5 and (d) nanoM5.

34

M. Zalaznik et al. / Wear 364-365 (2016) 31–39

Hardness, HV₀.₀₅

25 20 15 10 5 0 0

1

2

3

4

5

Filler, wt.% Micro MoS₂

Nano MoS₂

Micro WS₂

Nano WS₂

Fig. 2. Average Vickers hardness values of the PEEK composite samples.

Fig. 4. Average wear rates of the PEEK composite samples.

microMoS2 having the smallest wear rate at the highest concentration. The microMoS2/PEEK composites also seemed to be the least affected by the particle concentration, while the largest concentration effect was found for the nanoWS2/PEEK composites. The wear rates of the microWS2/PEEK composites were gradually increasing with a higher particle content and were, at all concentrations, higher than that of the PEEK. Regardless of the particle type or size, the wear rates were the highest at the highest concentration and were also higher compared to the pure PEEK. 3.3. Worn surfaces

Fig. 3. Average steady-state coefficients of friction for the PEEK composite samples.

concentration of nanosized particles was needed to achieve the same coefficients of friction. At 2 wt%, both the nanoMoS2 and the nanoWS2 displayed a coefficient of friction with values similar to those of the micro-sized filled composites, i.e., around 0.45. Therefore, when using low concentrations of particles, the microsized particles proved to be more effective at friction reduction than the nanosized particles, while at higher concentrations the differences among the various particles were much less evident, with no particular changes regarding the material or the size of particles. The greatest differences between the composites were observed in their wear behaviour, shown in Fig. 4. At low concentrations (0.5 wt% and 1 wt%) the composites filled with microMoS2, nanoMoS2 and nanoWS2 exhibited improved wear behaviour compared to the pure PEEK by up to 51%. All three composites exhibited their lowest wear rates at a concentration of 1 wt%. At 2 wt% the wear rates of the microMoS2 and the nanoWS2 were slightly increased, but were much lower than the nanoMoS2 or microWS2. A further increase in the concentration (5 wt%) resulted in poorer wear behaviour for all the samples, which was the most obvious for the nanoWS2, where the wear increased tremendously ( 2.5  ), matching the value of the microWS2. Both MoS2 composites exhibited a slower wear increase at higher concentrations compared to the WS2 composites, with the

The optical observations revealed the similar surface appearance of all the samples, regardless of the particle composition or size, with a combination of abrasion, areas of adhered material and areas of removed material. Because of the adhesion and the abrasion of the hard counterpart, the material was removed from the polymer surface and formed a transfer film with the counterpart. The removed material can be adhered back to the surface, which is revealed as positive or close-to-zero z-axes values on the 2D cross-sections (e.g., Fig. 5i and ii). The removed material adhered to the counterpart can also act as an abrasive particle, causing wide scratches on the surface (Fig. 5a and i). The amount of wide scratches (their width and depth) and the amount of adhered material is closely related to the wear rates measured on the samples. The removed material can adhere back to the originating polymer surface and cover the already-existing wide scratches, which results in a lower measured wear rate. The adhered material can also prevent the removal of additional material from the polymer surface, which also results in lower wear rates (Fig. 5b). There were no major differences observed in the wear mechanisms for the different composites, only in the intensity of the individual wear components. Some typical wear phenomena observed on the worn polymer surfaces are shown in Fig. 6. Adhesion, abrasion, and material removal in the form of spalling were observed on the worn surfaces; however, their intensity was mostly related to the wear rates and the hardness. The worn surfaces with a high hardness and low wear rates had a lot of adhered material (Fig. 6a), covering the already-existing wide scratches due to abrasion (Fig. 6b), and so preventing the formation of the new scratches. On the other hand, high wear rates were usually accompanied by large areas of removed material and by numerous wide scratches, but very little adhered material (Fig. 6c).

M. Zalaznik et al. / Wear 364-365 (2016) 31–39

i)

5

z axis, μm

Wide scratches

z axis, μm

ii)

iii) z axis, μm 500 μm

-5

0

1

2

3

-10 Wide scratches

-15

x axis, mm

5

Adhered material

0 -5

0

1

2

3

-10 -15 -20

500 μm

Removed material

0

-20

500 μm

Adhered material

35

x axis, mm

5 0 -5

0

1

2

3

-10 -15 -20

Removed material

x axis, mm

Fig. 5. Optical observations of selected worn polymer surfaces with the corresponding 2D cross-sections; (a)–(i) PEEK, (b)–(ii) nanoW1 and (c)–(iii) nanoM5.

3.4. Transfer films With the addition of the particles, the transfer-film formation on the steel counterpart was clearly promoted by the higher surface coverage and the thickness of the films, Fig. 7. The polymer transferfilm surface coverage on the steel counterparts, presented in Fig. 8a, appears to be more related to the concentration of the added particles rather that the particle composition or size. The transfer-film surface coverage was increasing in proportion to the concentration of particles and reached an almost identical final value (at 5 wt%), regardless of the particle type. A slightly lower surface coverage, of about 50%, was observed for the nanoWS2, while the highest surface coverage measured was 63% (microMoS2). Some more differences were observed in the transfer-film thicknesses, Fig. 8b. The transfer-film growth at low concentrations could be differentiated, depending on whether the composites were filled with micro- or nanoparticles. At 0.5–1.0 wt%, the transfer-film thicknesses were, in the case of micro-sized particles, much higher than those of the nano-sized particles. However, at 2 wt% all the samples had similar thickness values, but were again differentiated at the highest concentration. The MoS2-filled composites had slightly higher transfer-film thicknesses than the WS2-filled composite. 3.5. XPS analyses Figs. 9 and 10 show high-energy-resolution XPS spectra of Mo 3d and W 4f recorded before and after the tribological tests. The XPS spectrum of the Mo 3d before the tribological tests (Fig. 9a)

shows one doublet peak of Mo, while the small peak at a binding energy of 226.7 eV belongs to S 2s [23]. The Mo doublet (Fig. 9a) at binding energies of 229.4 eV and 232.5 eV can be identified as the Mo 3d5/2 and Mo 3d3/2 of Mo4 þ (MoS2). After the tribological tests the high-resolution XPS spectrum of Mo 3d can be fitted with two doublet peaks of Mo 3d. The first Mo doublet (solid line, Fig. 9b) at lower binding energies (229.4 eV and 232.5 eV) can be identified as the Mo 3d5/2 and Mo 3d3/2 peaks of Mo4 þ (MoS2), similar to the results before the tribological tests. The second Mo doublet (dotted line, Fig. 9b) with Mo 3d5/2 at 232.9 eV is very close to the literature value of Mo6 þ (232.4 eV), representing the MoO3 [23]. Very similar results were also found for the nanoMoS2/PEEK composites. Comparable observations were made for the WS2/PEEK composites. The high-resolution XPS spectrum of W 4f shows one doublet peak. The W doublet (Fig. 10a) at binding energies of 33.34 eV and 35.5 eV can be identified as the W 4f7/2 and W 4f5/2 peaks of W4 þ (WS2). In the XPS spectrum of W there is also a small peak at a binding energy of 39.0 eV, originating from the W 5p3/2 level, but it is not relevant for the interpretation of the W spectra. After the tribological tests the high-resolution XPS spectrum of W 4f shows multiple peaks that can be fitted with two doublet peaks. The first W doublet (solid line, Fig. 10b) at the lower binding energy (32.5 eV and 34.7 eV) can be identified as the W 4f7/2 and W 4f5/2 peaks of W4 þ (WS2). The second W doublet (dotted line, Fig. 10b) with W 4f7/2 at 35.8 eV represents W6 þ (WO3) [23]. Very similar results were found for the nanoWS2/PEEK composites. Based on the results of the XPS spectra of Mo 3d and

36

M. Zalaznik et al. / Wear 364-365 (2016) 31–39

Adhered material

50 μm

Wide scratches

50 μm

Removed material

50 μm Fig. 6. SEM observations of different wear mechanisms found on worn polymer surfaces; (a) adhesion (PEEK), (b) abrasion with wide scratches (M5) and (c) material removal, spalling (W2).

W 4f the entire oxidation of the MoS2 and WS2 particles occurred during the tribological sliding, since no oxidation products were observed on the unworn surfaces prior to the tribological tests. From the high-resolution XPS spectra the ratio of the individual components could be calculated, enabling an estimation of the oxidation for the MoS2 and WS2. Fig. 11 presents the ratio between the non-oxidised and oxidised particles (MoS2/MoO3 and WS2/ WO3) based on the ratios for Mo4 þ (MoS2): Mo6 þ (MoO3) and W4 þ (WS2): W6 þ (WO3). The results show that both MoS2 particles (micro and nano) had a very similar percentage of MoO3 present on the surface of the transfer film, i.e., 63–65%. On the other hand, the oxidation of the WS2 particles was higher compared to the MoS2, reaching values of 74–90%. Additionally, the nano-sized WS2 was more oxidised than the micro-sized WS2.

4. Discussion In general, the coefficient of friction for the composite PEEK samples was decreased with a higher concentration of particles, regardless of the particles' composition (MoS2 and WS2) or size (micro and nano), Fig. 3. Similar results were reported by McCook et al. [11] for nano- and microWS2 particles in a PEEK matrix and by Hu [24], investigating nano- and microMoS2 particles in HDPE. However, so far no tribological results were reported for MoS2/ PEEK composites.

Based on the particle size used in our research, two decreasing trends for the coefficient of friction were observed. In the case of the composites filled with microparticles, the coefficient of friction was decreased at low concentrations. On the other hand, if the composites were filled with nanoparticles, a higher concentration of particles was needed to reach the same coefficient of friction as with the microparticles, Fig. 3. The higher concentration of nanoparticles needed for the same friction reduction compared to the microparticles suggests that more nanoparticles are needed to deposit a self-lubricating tribofilm at the interface. This was demonstrated by the lower transfer-film thicknesses of the nanoparticles measured at low concentrations compared to microparticles, Fig. 8b. However, regardless of the particle type or size, the coefficients of friction for all the composites were almost the same for the highest particle concentration. Despite the large friction reduction obtained in our research for all the tested composite materials (up to 30%, Fig. 3), even lower coefficients of friction were expected, based on the reported low friction values for MoS2 and WS2 in the literature [8]. One of the possible reason for the less-than-expected benefit is the fact that under certain conditions (exposure to oxygen, high temperature, humidity, etc.), MoS2 and WS2 can oxidise into MoO3 and WO3, causing poorer intra-crystalline slip, reduced lubrication properties, a shorter lifetime, a higher coefficient of friction and less wear performance [25–27]. Our XPS analyses conducted on the transfer films showed that more than 65% of the MoS2 and more than 75% of the WS2 was oxidised during the tribological sliding (Fig. 11), which confirmed this hypothesis. This led to the relatively high coefficients of friction measured in our study. Although the lower oxidation of the nano- and microMoS2 compared to nano- and microWS2 particles implies a difference between these particles, this was not represented in the coefficients of friction. For low concentrations of particles, all the composites, except those filled with microWS2, showed better wear behaviour than the pure PEEK. Nevertheless, the wear improvement was greater when using nanoparticles. The better wear behaviour of the polymer composites filled with MoS2/WS2 nanoparticles compared to the polymer composites filled with MoS2/WS2 microparticles were to some extent also reported by other research groups [9,11,24]. These three particles (micro/nanoMoS2 and nano WS2) all led to the best wear resistance at the same optimum concentration (i.e., 1 wt%). After the optimum concentration was exceeded a severe deterioration in the wear properties was observed, which was also found in the literature [9,11,24]. The wear behaviour observed in our research was closely related to the hardness of the composites. Composites with a higher hardness had a much better wear performance than the composites with a lower hardness. A lower hardness indicates that the material is more prone to plastic deformation, thereby enhancing the wear. In general, all the worn surfaces of the PEEK composites resulted from a combination of the same wear mechanisms, i.e., abrasion, adhesion and spalling, just that the intensities of the individual components were different. Composite materials with higher hardness and, therefore, lower wear rates had a lot of adhered material deposited on their own wear tracks (Fig. 5b). This adhered material covered the already-existing wide scratches (abrasion) and/or prevented the formation of new ones. Composite materials with lower hardness and high wear rates were usually accompanied by large areas of the removed material and by numerous wide scratches, but very little adhered material (Fig. 5c). To achieve an improved tribological behaviour, the formation of a transfer film on the counterparts is of great importance [15–17], since the transfer film is supposed to protect the polymer from the usually harder counterpart, can act as a lubricant and is responsible for the transition from the increasing to the steady-state wear behaviour [28]. In our tribological study, the PEEK without

M. Zalaznik et al. / Wear 364-365 (2016) 31–39

37

500 μm

500 μm

500 μm

100

Transfer film thickness, μm

Transfer film surface coverage, %

Fig. 7. Optical observations of the transfer films on the steel counterparts of; (a) PEEK, (b) M0.5 and (c) M5.

80 60 40 20 0

3

2

1

0

0

1

2

3

4

5

Filler, wt.% Micro MoS2

Nano MoS2

0

1

2

3

4

5

Filler, wt.% Micro WS2

Nano WS2

Fig. 8. (a) Transfer-film surface coverage and (b) transfer-film thickness measured on the steel counterparts of the PEEK composite samples.

any particles showed only a slight tendency to form a transfer film with a counterpart surface (Fig. 7a), which can also be observed in other literature [29]. The weak bonding of PEEK with a steel counterpart and the low surface coverage of the transfer film are also revealed as a higher coefficient of friction compared to the composite samples. With the addition of the particles, the transfer-film formation on the steel counterpart was obviously promoted via a larger surface coverage and the thickness of the films (Fig. 8), which resulted in lower coefficients of friction, probably due to the weaker interlayer forces inside the thick transfer film, enabling easier interfacial sliding, typical for MoS2 and WS2. However, an increased surface coverage and a greater transfer-film thickness did not also mean improved wear behaviour. On the contrary, less wear was recorded when thinner

transfer films were measured; while thicker transfer films were usually accompanied by high wear. This observation is consistent with other literature stressing the importance of a thin and uniform transfer film for improved wear behaviour [15,29–31]. It is believed that the MoO3 helps with the transfer of material to the counterpart surface and promotes thicker transfer films [32], but the adhesive strength of the transfer films is weak and the transfer film can be easily removed from the surface, leading to high wear rates. Based on the XPS results of our transfer films the oxidation rate of the MoS2 was lower than that of the WS2, which was also observed in other literature [33]. Composites filled with MoS2 particles indeed had lower wear rates compared to the composites filled with WS2 particles, which could be a direct effect of the oxidation.

38

M. Zalaznik et al. / Wear 364-365 (2016) 31–39

1200

Mo 3d Intensity, c/s

1000

Mo4+

800 600 400 200 0

240

235

230

225

220

Binding energy, eV

Intensity, c/s

Fig. 11. Oxidation of micro/nanoMoS2 and micro/nano WS2 particles.

Mo 3d

1800

Mo6+

1400

Mo4+

1000 600 200

240

235

230

225

220

Binding energy, eV

Our XPS analyses also revealed that the oxidation rate of both sizes of MoS2 particles were very similar, which is different to the observations made in [9], where the oxidation rate of nanoMoS2 particles in POM was higher than that of microMoS2 particles in POM. On the other hand, the oxidation rate of the nanoWS2 particles used in our investigation was much higher compared to the microWS2 particles. This findings in PEEK-based composite tribology suggest that the oxidation is very important, or even key, for the wear (and friction) of PEEK composites and must be studied further to reveal the optimum particle type, size and concentration for any polymer application.

Fig. 9. High-resolution XPS spectra of Mo 3d obtained on (a) unworn surface of M5 and (b) transfer film of M5 after the tribological tests.

5. Conclusions 450

W 4f

Intensity, c/s

400 350

 The coefficients of friction for composite materials were, com-

W4+

300 250



200 150 100 50

44

40

36

32

28

Binding energy, eV



Intensity, c/s

2500

W 4f

2000

W6+

1500

W4+

1000



500



0 46

42

38

34

30

26

Binding energy, eV Fig. 10. High-resolution XPS spectra of W 4f obtained on (a) unworn surface of W5 and (b) transfer film of W5 after the tribological tests.

Furthermore, the XPS analyses performed before the tribological tests (on the unworn polymer surface) and after the tribological tests on the selected transfer films formed on the steel counterparts unambiguously confirmed that the entire oxidation process for MoS2 and WS2 occurs during the tribological sliding.

pared to pure PEEK, decreased by up to 30%, regardless of the particle type or size. Lower coefficients of friction are attributed to the transfer-film formation on the steel counterpart. All the composites, except those filled with microWS2, showed lower wear rates at low concentrations of particles, compared to the pure PEEK. The wear behaviour was greatly dependent on the hardness of the composites; higher hardness led to a much better wear performance compared to composites with lower hardness. Although a higher concentration of nanoparticles was needed to reduce the friction compared to microparticles, the wear rates were lower when nanoparticles were used. Addition of low concentration of nanoparticles into PEEK matrix immensely enhanced the hardness of the material which led to the better wear performance but also to the higher coefficients of friction due to a lesser self-lubricating tribofilm at the interface. Based on the XPS results of our transfer films the oxidation rate of MoS2 was lower than that of WS2. The composites filled with MoS2 particles indeed had lower wear rates than the composites filled with WS2 particles. XPS analyses before and after the tribological tests unambiguously confirmed that the entire oxidation process for MoS2 and WS2 occurs during the tribological sliding. XPS analyses also revealed that the oxidation rate of both sizes of MoS2 particles were very similar, while the oxidation rate of the nanoWS2 particles was much higher than for the microWS2 particles.

Acknowledgements The authors greatly acknowledge Marko Lukek for the preparation of the samples, dr. Janez Kovač and Tatjana Filipič (Department

M. Zalaznik et al. / Wear 364-365 (2016) 31–39

of surface engineering and optoelectronics, Jožef Stefan Institute) for their help with the XPS measurements and the analyses. This work was in major part funded by the Slovenian Research Agency (ARRS), Contract number J2-4191.

References [1] J. Sandler, P. Werner, M.S.P. Shaffer, V. Demchuk, V. Altstadt, A.H. Windle, Carbon-nanofibre-reinforced poly (ether ether ketone) composites, Compos. Part A: Appl. Sci. Manuf. 33 (2002) 1033–1039. [2] J. Sandler, A. Windle, P. Werner, V. Altstädt, M. Es, M. Shaffer, Carbonnanofibre-reinforced poly (ether ether ketone) fibres, J. Mater. Sci. 38 (2003) 2135–2141. [3] J.P. Davim, R. Cardoso, Effect of the reinforcement (carbon or glass fibres) on friction and wear behaviour of the PEEK against steel surface at long dry sliding, Wear 266 (2009) 795–799. [4] J.P. Davim, R. Cardoso, Tribological behaviour of the composite PEEK-CF30 at dry sliding against steel using statistical techniques, Mater. Des. 27 (2006) 338–342. [5] Z. Lu, K. Friedrich, On sliding friction and wear of PEEK and its composites, Wear 181 (1995) 624–631. [6] D.L. Burris, W.G. Sawyer, A low friction and ultra low wear rate PEEK/PTFE composite, Wear 261 (2006) 410–418. [7] B. Stuart, Tribological studies of poly (ether ether ketone) blends, Tribol. Int. 31 (1998) 647–651. [8] B. Bhushan, Modern Tribology Handbook, CRC Press LLC, USA, 2001. [9] K. Hu, J. Wang, S. Schraube, Y. Xu, X. Hu, R. Stengler, Tribological properties of MoS2 nano-balls as filler in polyoxymethylene-based composite layer of threelayer self-lubrication bearing materials, Wear 266 (2009) 1198–1207. [10] J. Wang, K. Hu, Y. Xu, X. Hu, Structural, thermal, and tribological properties of intercalated polyoxymethylene/molybdenum disulfide nanocomposites, J. Appl. Polym. Sci. 110 (2008) 91–96. [11] N. McCook, M. Hamilton, D. Burris, W. Sawyer, Tribological results of PEEK nanocomposites in dry sliding against 440C in various gas environments, Wear 262 (2007) 1511–1515. [12] Q.J. Xue, Q.H. Wang, Wear mechanisms of polyetheretherketone composites filled with various kinds of SiC, Wear 213 (1997) 54–58. [13] M. Kuo, C. Tsai, J. Huang, M. Chen, PEEK composites reinforced by nano-sized SiO2 and Al2O3 particulates, Mater. Chem. Phys. 90 (2005) 185–195. [14] N. Ma, G. Lin, G. Xie, G. Sui, R. Yang, Tribological behavior of polyetheretherketone composites containing short carbon fibers and potassium titanate whiskers in dry sliding against steel, J. Appl. Polym. Sci. 123 (2012) 740–748. [15] S. Bahadur, C.J. Schwartz, The influence of nanoparticle fillers in polymer matrices on the formation and stability of transfer film during wear, Tribol. Interface Eng. Ser. 55 (2008) 17–34. [16] N. Myshkin, M. Petrokovets, A. Kovalev, Tribology of polymers: adhesion, friction, wear, and mass-transfer, Tribol. Int. 38 (2006) 910–921.

39

[17] K. Laux, C. Schwartz, Effects of contact pressure, molecular weight, and supplier on the wear behavior and transfer film of polyetheretherketone (PEEK), Wear 297 (2013) 919–925. [18] M. Zalaznik, M. Kalin, S. Novak, Influence of the processing temperature on the tribological and mechanical properties of poly-ether-ether-ketone (PEEK) polymer, Tribol. Int. 94 (2016) 92–97. [19] M. Kalin, M. Zalaznik, S. Novak, Wear and friction behaviour of poly-etherether-ketone (PEEK) filled with graphene, WS2 and CNT nanoparticles, Wear 332–333 (2015) 855–862. [20] A. Greco, R. Erck, O. Ajayi, G. Fenske, Effect of reinforcement morphology on high-speed sliding friction and wear of PEEK polymers, Wear 271 (2011) 2222–2229. [21] B.J. Briscoe, S.K. Sinha, Tribological applications of polymers and their composites: past, present and future prospects, Tribol. Interface Eng. Ser. 55 (2008) 1–14. [22] M. Zalaznik, S. Novak, M. Huskić, M. Kalin, Tribological behaviour of a PEEK polymer containing solid MoS2 lubricants, Lubrication, Science 28 (2016) 27–42. [23] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, in: J. Chastain, R.C. King (Eds.), Handbook of X-ray Photoelectron Spectroscopy, Physical Electronics, Inc., Minnesota, USA, 1995. [24] K.H. Hu, X.G. Hu, J. Wang, Y.F. Xu, C.L. Han, Tribological properties of MoS2 with different morphologies in high-density polyethylene, Tribol. Lett. 47 (2012) 79–90. [25] K. Wong, X. Lu, J. Cotter, D. Eadie, P. Wong, K. Mitchell, Surface and friction characterization of MoS2 and WS2 third body thin films under simulated wheel/rail rolling–sliding contact, Wear 264 (2008) 526–534. [26] B.C. Windom, W. Sawyer, D.W. Hahn, A Raman spectroscopic study of MoS2 and MoO3: applications to tribological systems, Tribol. Lett. 42 (2011) 301–310. [27] R.R. Sahoo, S.K. Biswas, Microtribology and friction-induced material transfer in layered MoS2 nanoparticles sprayed on a steel surface, Tribol. Lett. 37 (2010) 313–326. [28] S. Bahadur, The development of transfer layers and their role in polymer tribology, Wear 245 (2000) 92–99. [29] J. Vande Voort, S. Bahadur, The growth and bonding of transfer film and the role of CuS and PTFE in the tribological behavior of PEEK, Wear 181 (1995) 212–221. [30] G. Jintang, Tribochemical effects in formation of polymer transfer film, Wear 245 (2000) 100–106. [31] L. Yu, S. Yang, W. Liu, Q. Xue, S. Bahadur, Action of transfer film in improving friction and wear behaviors of iron- and copper-filled poly (ether ether ketone) composites, J. Appl. Polym. Sci. 76 (2000) 179–184. [32] J. Wang, M. Gu, B. Songhao, S. Ge, Investigation of the influence of MoS2 filler on the tribological properties of carbon fiber reinforced nylon 1010 composites, Wear 255 (2003) 774–779. [33] J. Kogovšek, M. Kalin, Various MoS2  , WS2  and C-based micro- and nanoparticles in boundary lubrication, Tribol. Lett. 53 (2014) 585–597.