University of Groningen Self-adaptive and self-healing nanocomposite tribocoatings Cao, Huatang

Self-adaptive and self-healing nanocomposite tribocoatings

Cao, Huatang

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.

Document Version

Publisher's PDF, also known as Version of record

Publication date: 2019

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

Cao, H. (2019). Self-adaptive and self-healing nanocomposite tribocoatings. University of Groningen.

Copyright

Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

2

2

C

HAPTER

2

In this chapter, we summarize the various techniques utilized for the production of WS2/a-C(H) coatings, and the subsequent experimental methods for the characterization and data analysis. Closed-field unbalanced magnetron sputtering was employed to prepare the expected tribocoatings. Characterization tools such as scanning electron microscope (SEM), transmission electron microscope (TEM), focused ion beam (FIB), atomic force microscope (AFM), energy dispersive spectroscope (EDS), glazing incidence X-ray diffraction (GI-XRD), X-ray photoelectron spectroscopy (XPS), Raman spectroscopy, nanoindentation and tribometry were used to determine the coating morphology, chemical composition, phases, microstructure and their mechanical and tribological properties. The basic principles of physical sputtering deposition and characterization techniques are explained.

2.1 P

HYSICAL

S

PUTTERING

The physical sputtering process involves the physical (not thermal) vaporization of atoms from a surface by momentum transfer from impinging energetic atomic-sized particles [1]. The energetic particles are commonly ions of a gaseous material accelerated in an electric/magnetic field from a plasma. In particular, sputtering is the deposition of particles that originate from the target surface being sputtered and transferred to the substrates, while resputtering is the simultaneous re-emission of the deposited material away from the substrates by ion or atom bombardment. The sputtering gas is often an inert gas such as argon (Ar). Free electrons flow from the negatively charged target source material (acting as cathode) in the plasma environment, colliding with the outer electronic shell of the Ar gas atoms. This results in an ionization of Ar gas whichare later accelerated by the negatively charged targets due to the potential difference at a very high velocity. The impingement onto the target materials “sputters off” atomic-size particles from the target source material. These particles cross the vacuum deposition chamber of the sputter coater and are deposited as a thin film of material coated on the surface of the substrate. Besides, secondary electrons are released accompanying to the sputtering process and are repelled away aiding the ionization process. The magnetron positioned underneath the target can increase the efficiency of the ionization as the helical patch that the free electrons follow in the magnetic field extend the effective length and consequently enhance the probability of ionizing collisions with neutral working gas atoms. The magnetron also confines the electron such that sputtering can be achieved at high rates with denser plasma. Consequently, enhanced ion currents can be achieved to the targets considerably increasing the deposition rate. Sputtering only takes place under high vacuum condition when the kinetic energy of the bombarding particles is high enough to overcome the threshold energy, and thus allowing more controlled thin film deposition on the atomic level than that could be achieved by melting a source material with conventional thermal means.

The ratio of atoms ejected or “sputtered off” from the target or source material to the number of high energy incident particles is called the sputter yield. The sputter yield varies and can be controlled by the energy and incident of angle of the bombarding ions, the deposition pressure, the relative mass of the ions and target atoms, as well as the surface binding energy of the target atoms. The bombarding particles produce a collision cascade and some of the momentum is transferred to surface atoms which can be ejected (or sputtered). The mass of the bombarding species is significant to the energy and momentum transferred to the film atom during the collision. From the laws of the conservation of energy and the conservation of momentum, the energy, Et,

2

2

transferred by the physical collision between hard spheres is given by the equation [1,2]:

𝐸𝑡

𝐸𝑖

=

4 𝑀𝑡𝑀𝑖𝑐𝑜𝑠2𝜃

(𝑀𝑖+𝑀𝑡)2 (2.1)

where E is the energy, M is the mass, i is the incident particle, t is the target particle and 𝜃 is the angle of incidence as measured from a line joining their centers of mass as shown in Figure 2.1. Obviously, when 𝜃 equals to zero (cos𝜃 = 1), it is a linear collision. Further if the masses are identical, a maximum energy could be transferred. Accordingly, the larger mass difference, the lower energy transfer ratio will be. Therefore, matching the atomic mass of the bombarding ion to the target atom is crucial to the sputtering yield.

Figure 2.1 Collision of particles.

For instance, from the energy transfer point of view, this makes xenon (131 amu) and mercury (201 amu) more attractive for sputtering heavy elements, and instead light elements such as nitrogen (only 14 amu) less attractive (this rule also applies to the resputtering from the deposited material). However, recently argon (40 amu) gains popularity as the sputtering gas due to its low cost and chemical inertness.

In addition, sputtering is a non-equilibrium process. When sputtering is performed in a low pressure or vacuum environment, high energy reflected neutrals of the bombarding gas and high energy sputtered atoms from the target bombard the growing film and thus affect the film formation process. High energy bombardment can trigger resputtering of one or several particular elements from the depositing material giving an obvious decrease in the sputtering yield from the target and driving the deposited material far from its expected stoichiometry [3,4]. Ions are accelerated to the plasma during ion plating with a self-bias. This can be further aggravated when a negative bias is purposely applied on the collecting substrates. For instance, in real practices of sputtering, one normally resorts to a very high negative bias (up to several hundred voltages) to produce substantial high energy bombardment particles to remove the contaminants on the substrates prior to the coating deposition. Still, a high

pressure in the aggregation chamber can reduce the sputtering rate. This is due to that a higher pressure leads to a short mean free path of the species, which enhances the possibility of collisions with inert gas atoms and drains the energy of the ions such that fewer of them keep sufficient threshold energy for the subsequent sputtering [5]. On the other hand, high energy particle bombardment onto the growing surface induces “atomic peening” where surface atoms are struck and recoil into voids and interstitial sites in the lattice of the surface region [6]. This results in densification of the material and introduces also compressive stresses into the coating.

It is known that in WS2 based sputtered coatings, there is a heavy S preferential resputtering problem (corresponding to W-rich) in the coatings [7–9]. This can be due to the larger mass difference of Ar (40 amu) to W (180 amu) and the close mass difference of Ar to S (32 amu), whereby Ar can transfer higher energy ratio to S as compared with W causing resputtering. Another reason to be taken into account arise from the weak bonding of S to the substrate because of its higher vapor pressure (thus volatile) [10]. Therefore, Chapter 3 and Chapter 4 of this dissertation have investigated different methods to tune the chemical composition of the sputtered WS2 based coatings in various conditions (e.g. target substrate distance, carbon content, bias voltage, deposition pressure, etc.) and the corresponding mechanical properties were investigated. A key issue in this dissertation is to shed some new insights into the effects of S/W ratio and the total content of sulfur on the tribological properties.

2.2 C

OATING DEPOSITION

Nanocomposite WS2/a-C(H) coatings, unhydrogenated or hydrogenated, were deposited on silicon wafers (100) and M2 high speed steel substrates in this study, which were ultrasonically cleaned with acetone and rinsed with ethanol for three times and further Ar plasma etched before deposition. The coatings were deposited by magnetron sputtering in a TEER UDP400/4 closed-field unbalanced magnetron sputtering (CFUMS) system, as shown in Figure 2.2. Figure 2.3a illustrates the schematic view of the closed-field unbalanced magnets, in an argon atmosphere for nonreactive co-sputtering from one graphite targets and two face-to-face WS2 targets as indicated in Figure 2.3b. For reactive sputtering, a mixture of argon/acetylene was used, among them the acetylene gas provides the carbon source. Note that the C and Cr targets are powered by a direct current (DC) power source, whereas that for WS2 targets are powered by pulsed DC power source where the negative charge built-up over the target surface is removed via a short duration cycle reversal of an applied positive voltage. This exhibits advantages of dramatically reducing or eliminating the arcing in targets. Pulsed-DC power supplies that operate up to 350 kHz are

2

2

commercially available, allowing stable processing of reactive materials that form insulating compounds when sputtered in a reactive atmosphere.

Figure 2.2 (a) Teer UDP-400/4 deposition system, (b) a view of the inside of the deposition chamber with the front door open, the magnetrons and the sample holder are visible (courtesy by Damiano Galvan [11]) .

Figure 2.3 (a) Schematic view of the closed-field unbalanced magnetron sputtering (CFUMS) configuration, the polarity of the magnets facing the vacuum chamber is alternated among neighboring magnetrons. Pure Ar gas or a mixture of Ar and C2H2 is introduced into the chamber for nonreactive or reactive sputtering deposition process, respectively; (b) schematic illustration of the four positioned targets as indicated, with a rotating carrousel holder around the central axis of the chamber while with adjustable target-substrate distances (TSD) ranging from 70-290 mm.

Further, with this closed-field unbalanced technology, a magnetic trap can be formed to prevent electrons to escape the plasma region towards the chamber walls, which increases the overall ionization of the discharge, see Figure 2.3a. Furthermore, this unbalanced approach increases the magnetic field intensity and thus the plasma

ionization in proximity to the substrates. The extent of ion bombardment on the targets and substrate and the reactivity of the reactive gases are enhanced, which contributes to a denser coating and a higher deposition rate. Considering that ions are charged particles, magnetic fields can be used to control their velocity and behavior. A rotary pump acts as backing pump and a dry turbo-molecular pump plays the key role in enabling the system to reach the base pressure in the lower 10-6 _{mbar range.}

2.3 C

HARACTERIZATION TECHNIQUES

A very diverse set of experimental tools (microscopic, chemical, structural) have been employed to characterize various materials aspects, such as microstructure, phases, composition and mechanical properties. A concise summary of most of the microscopy and spectroscopy and mechanical testing machine is listed in Table 2.1. The abbreviations and a description of the main working principle of the lists are presented as follows.

Table 2.1 List of analysis equipment applied in this dissertation.

Abbreviation Technique Model description OM Optical Microscopy Olympus VANOX-T Confocal Confocal Microscopy Nanofocus uSurf SEM Scanning electron

microscopy Philips XL-30 ESEM Philips XL-30 SEM Tescan Lyra SEM EDS Energy-dispersive X-ray

spectroscopy EDAX in SEM Bruker in TEM (HR-/EF-)TEM

High-resolution/energy-filtering transmission electron microscopy

JEOL 2010 Field-emission TEM

AFM Atomic force microscopy Veeco Dimension 3100 FIB Focus ion beam Tescan Lyra

FEI Helios G4 CX XPS X-ray Photoelectron

Spectroscopy Surface Science SSX-100 ESCA Raman Raman Spectroscopy Thorlabs HNL

XRD X-Ray diffraction D8 Advance, Bruker PANalytical XPert Pro MRD

2

2

SEM

SEM, as the most widely used type of electron microscope, examines the microscopic structure and surface morphology by the scanning the surface of materials, offering much higher resolution and greater field of view at larger depth as compared with optical microscope. In SEM the electron beam is focused to a spot (controlled by the size of the electron probe) and scans sequentially across the specimen. At each location, signals are emitted from the specimen and collected by detectors. The detector signal is synchronized with known location of the beam on the specimen, and the signal intensity is used to modulate the corresponding image pixel. The signals collected in series are combined to form an image whose dimensions/pixel distribution depends on the scanning pattern chosen. The acceleration voltage to generate an electron beam is in the range of 1-40 kV [12].

When high energy electrons strike a specimen, they produce both elastic and inelastic scattering, yielding two types of signal electrons forming SEM images. Among them, elastic scattering produces the backscattered electron (BSEs), which are incident electron scattered by atoms in the specimen. In contrast, inelastic scanning produces secondary electrons (SEs), which are electrons ejected from atoms in the specimens surface at a depth of 5–50 nm. It must be pointed out that SEs are deflected at small angles and show significantly low energy (3-5 eV) [12] compared with incident electrons; thus SEs are mostly used for achieving topographic contrast. Besides, SEs electrons emitted from the specimen surfaces facing the detector will be collected abundantly, and the corresponding sites in the image will show more brightly. Similarly, the electrons emitted from the surfaces not facing the detector reach the detector with more difficulty, thus rendering corresponding areas in the image appearing darker. On the other hand, BSEs are deflected from the specimen at larger angles and they have an energy level close to that of incident electrons (60-80%). Their high energy enables them to escape from a deeper level in the interaction zone, from depths of around 50–300 nm. Note in elastic scattering, larger atoms (corresponding with a greater atomic number, Z) have a higher probability of producing an elastic collision due to their greater cross-sectional area, consequently, the number of BSEs reaching a BSE detector is proportional to the mean atomic number of the sample. Thus, a brighter BSE intensity correlates with greater average Z (e.g. heavy metals) in the specimen and dark areas corresponds to a lower average Z, whereby BSEs provide information for elemental composition contrast. In this study, we mostly used an accelerating voltage of 5 kV in the Philips XL-30 ESEM for surface morphology observations of the deposited coatings and worn surface after wear testing.

Transmission electron microcopy (TEM) is a powerful versatile technique that makes use of a much higher energy electron beam with high energy up to several hundred of kiloelectronvolts (e.g. normally 200 keV) to transmit through an ultra-thin specimen, interacting with the specimen as it pass through so as to achieve high spatial resolution characterization of morphology (size, shape, arrangement of particles on scale of atomic diameters), crystallographic information (diffracted electrons, arrangement and order of atoms, or defects) and chemical identification (EDS, or electron energy loss spectroscopy) of various materials.

According to the Abe and Rayleigh criterion, the smallest distance that can be resolved in visible-light microscope (VLM), δ, is given approximately by the equation [13]:

δ=_𝜇sinβ0.61𝜆

,

λ = _𝐸1.22_0.5 (2.3)

where λ is the wavelength of radiation (nm), µ is the refractive index of the viewing medium, and β is the semi-angle of collection of the magnifying lens, E is the electron voltage (eV). Take the green light in the middle visible spectrum, λ is about 550 nm [13] corresponding to the resolution of a good VLM around 300 nm. However, this resolution limit is far from distinguishing materials down to a nano scale from an atomic view. In TEM, the high energy electron beam creates extremely short electron wave; for instance, when applying a voltage of 100 keV, the λ is only about 0.004 nm, which is much smaller than the diameter of an atom (0.1~0.5 nm). The higher voltage, the lower wavelength and better resolution will be. Thus, atomic resolution is expected readily obtained in TEM. However, it is in fact impossible to develop an ideal TEM that approaches the theoretic wavelength-limited limit of resolution due to substantial spherical aberrations in lens and also the chromatic aberrations due to the specimen itself. Currently, the spherical aberration corrected TEM normally can provide resolution down to sub angstrom.

The ray diagram of imaging and diffraction mode is shown in Figure 2.4. In imaging mode, an objective aperture is inserted in the back-focal plane of the objective lens (thus the object plane of the intermediate lens is the image plane of the objective lens), as shown in Figure 2.4 (the right). For bright field imaging, a small objective aperture is used with the center beam selected and the rest signal blocked. The transmitted beam is subsequently magnified and further projected by the intermediate and projector lenses to obtain the sample image on CCD (e.g. a fluorescent screen). In bright field, contrast is formed directly by the absorption of electrons in the sample, namely, thicker regions of the sample, or regions with a higher atomic number will appear dark, whereas regions with no sample in the beam path will appear bright. For

2

2

dark field imaging, on the other hand, the small objective only allows a diffracted beam to pass and form the image. Image contrast is obtained by the interaction of the electron beam with the sample where denser areas or areas containing heavier elements appear darker because of elastic and inelastic scattering of the electrons in the sample. In addition, scattering from crystalline materials also introduces diffraction contrast. This contrast depends on the orientation of a crystalline area in the sample with respect to the electron beam.

Figure 2.4 Schematics of diffraction mode and imaging mode in TEM [13]. In diffraction mode, as shown in Figure 2.4 (the left), a selected area aperture is inserted into the beam path and placed in the back-focal plane (thus the back focal plane of the objective lens is the object plane for the intermediate lens) below the sample holder and allowing selection of the interested area where the diffraction patterns is projected onto the viewing screen. Diffraction contrast is formed by elastically scattered electrons. Samples can exhibit diffraction contrast, whereby the electron beam undergoes Bragg’s diffraction in a crystalline sample, resulting in

disperses electrons into discrete locations (e.g. diffraction spots) corresponding to a satisfied diffraction condition of the specific crystal structure. In particular, individual array spots come from a single crystal, whereas ring patterns come from polycrystalline materials and some weak halo may rise from amorphous materials. Energy-filtered transmission electron microscopy (EF-TEM) is a family of imaging techniques that utilize properties of the energy loss spectrum, whose contrast is formed by inelastically scattered electrons where inner shell ionization determines the presence of certain element, thereby increasing contrast and creating unique contrast effects in the image [13,14]. The selection of zero-loss electrons in images and diffraction patterns not only allows a better comparison with computer simulations but also eliminates the inelastic background in diffraction patterns from thicker samples and avoids the blurring of images due to the chromatic aberration [14]. Normally, some energy selecting slits are employed to transform the selected part of the spectrum into energy-filtered images. This technique is particularly useful in creating elemental/chemical maps at nanometer resolution. In this thesis, EF-TEM was used to compare the elemental contrast at the cross-section of a TEM lamella sliced at the wear interface in Chapter 5.

FIB

Focused ion beam (FIB) instrument uses a beam of ions (e.g. Ga+_{, He}+_{) rather than} electrons directly to modify/mill or image a surface at nanometer precision provided carefully controlling the energy and intensity of the ion beam [15].

Besides, ion beam enhanced chemical vapor deposition is also used to deposit material with comparable precision to FIB milling. In such case, a small quantity of a specific precursor gas such as tungsten hexacabonyl (W(CO)6) is introduced into the vacuum chamber and is allowed to chemisorb onto the sample. When injected under the scanning beam, the precursor gas will be decomposed into volatile and non-volatile components, with the nonvolatile remained as a deposition product on the specimen surface while the volatile products being extracted by the vacuum system. Therefore, the deposited metal (e.g. W) can be used as a sacrificial layer to protect the underlying sample from the destructive sputtering of the ion beam.

A FIB is mostly combined with an SEM, see Figure 2.5. In a FIB-SEM dual beam, the electron and ion beams intersect at a particular angle (e.g. 52°) at a coincident point near the sample surface offering high resolution SEM imaging of the FIB-milled surface. Such systems combine the advantages of both the SEM and FIB and provide complementary imaging and beam chemistry capabilities.

2

2

Figure 2.5 The 52° tilted stage positon for in-situ lifting out TEM lamella in the FEI Helios G4 CX dural beam system.

Figure 2.6 (a) depiction of protective of Pt layer on the wear track of WS2/a-C coating after high-temperature tribotest for 5000 laps at 457 °C; (b) bulking milling on sides of sample; (c) U shape cut; (d) sample lift-out and weld on the TEM grid; (e) sample thinning at a lowering voltage; (f) close-view of electron transparent sample (tribofilm in-situ formed in the wear track of interest) for later TEM observations.

In this thesis, we predominately make use of FIB to prepare site-specific (e.g. on the wear track or the worn ball scar) TEM lamella sample, which is otherwise unviable in conventional methods such as mechanical grinding, followed by dimpling and ion polishing by the precision ion polishing system (PIPS) or other chemically etching methods. In practice, we take the following procedures (see Figure 2.6) to in-situ lift out samples in Tescan Lyra or FEI Helios G4 FIB-SEM dual beam systems: (a) get started to use SEM to find the areas of interest, set eucentric height and then focus and link; (b) insert gas injection system (GIS) and use electron beam and ion beam to sequentially deposit two protective layers (e.g. Pt in this study); (c) perform the bulk milling, intermediate milling and U shape-cut; (d) in-situ lift out the lamella from the bulk sample and weld it to a Cu FIB lift-out grid; (e) gradually thin the lamella; (f) finally polish the lamella using low voltages until sample becoming electron transparent.

AFM

Atomic force microscopy utilizes a cantilever with an ultra-sharp tip to provide the means for sensing the forces exerted on the tip by the sample [16]. The tip runs over the ridges and valleys in the material revealing the surface morphology, see Figure 2.7. As the tip oscillates up and down due to the surface, the cantilever deflects. At the same time, a focused laser beam shines on the backside of the cantilever at an oblique angle and is subsequently reflected and collected by a position sensitive detector consisting of segmented split photodiodes whereby an image based on the displacement signal of the deflected cantilever may be created and magnified, revealing the configuration of the features being imaged from the machine. In comparison to SEM, AFM is superior in performing in ambient conditions on various materials and could produce higher atomic resolution (i.e. in the z-direction; the

Figure 2.7 (a) schematic of AFM operation in both tapping mode and contact mode (http://web.physics.ucsb.edu/~hhansma/biomolecules.htm), and typical AFM images of a WS2/a-C coating under the tapping mode (b) 2D profile and (c) 3D profile.

lateral resolution is still rather low). In this work, AFM was conducted for surface roughness measurement by the Veeco Dimension 3100 machine using a tapping mode

2

2

which follows the surface with the tip just oscillating over the surface, thus reducing the likelihood of scratching the trench in the surface and minimize the damage induced on a soft surface (e.g. pure WS2 coating). This mode also provides extra information about the sample surface in the phase image that corresponds to the height image. XRD

X-ray diffraction (XRD) is used to characterize the phases present in the crystalline material and regularity of the crystal. XRD contains crystallographic information over a large surface, thus obtaining volume averaged statistical information. In this work, the accelerating voltage and electron probe current in PANalytical-X'Pert MRD are 40 kV and 40 mA, respectively. The K-shell electrons are knocked out with the incident high-energy electron, followed by the vacancies being filled from the electrons from L-shell or M-L-shell, radiating the X-ray with a wavelength of 1.54 Å for the characteristic Ka spectrum; therefore the diffraction spectrum is generated by the elastic scattering of X-ray fulfilling the Bragg condition:

nλ = 2d sinθ (2.4)

where λ is the wavelength, θ is the scattering angle between the incident light and the direction normal to sample plane and n is the order of the diffraction peak. The lattice distance d can be calculated accordingly.

Figure 2.8 (a) the PANalytical-X'Pert MRD set-up for GI-XRD measurement; (b) a sketch of GI-XRD to detect WS2 nanoplatelets in a thin WS2/a-C coating.

For characterizing thin films using conventional θ-2θ scanning (e.g. powder XRD) method generally produces a weak signal from the film and an intense signal from the substrate. The way to avoid intense signal from the substrate and get stronger signal from the top film itself is to perform a 2θ scan with a fixed grazing angle of incidence (ω), namely known as GI-XRD. A small grazing angle (0.2°<ω<5°) is generally chosen to be slightly above the critical angle for total reflection of the film material. With GI-XRD in the PANalytical-X'Pert MRD equipment (Figure 2.8a), only the detector is moved and the spectral peaks correspond to the crystal planes satisfying the Bragg condition only coming from the surface structure in the order of nanometers (surface sensitive).

However, in GI-XRD crystal planes are inclined with respect to the sample surface whose normal is the bisector of the angle formed by the incident and diffracted beam (the diffraction vector s is changing its direction during the scanning, see Figure 2.8b), and the instrumental contribution to the width of the diffraction peak is higher as compared with the conventional powder XRD, making GI-XRD less appropriate to study orientation and crystallites size of the materials or samples with preferred orientation.

2.3.2 Chemical composition

Information about the surface chemical composition or phases can be obtained in various ways. For instance, the most common techniques depends on the required information, e.g. elemental distribution, valence state or a fine depth resolution. The techniques such as Energy-dispersive X-ray spectroscopy (EDS), X-ray photoelectron spectroscopy (XPS), X-ray diffraction (XRD) rely on various processes which take place because of irradiation, as shown in Figure 2.9. Some brief descriptions are given below.

Figure 2.9 The interaction of incoming primary electrons with a sample: (a) signals generated by electron-matter interaction in a thin sample; (b) absorption of SE, BSE, and X-rays in thick specimen, by inelastic scattering within the interaction volume, limits the sample depth where they can escape [17].

EDS

EDS relies on an incident electron beam on a zone of interests, either performed in SEM or TEM. The incoming electron knocks out a core shell (e.g. K, L, M, N…) electron of the

2

2

irradiated atoms, see Figure 2.10. The energy release by an electron form an outer shell to fill the vacancy of the inner core shell in the form of a characteristic X-ray.

Figure 2.10 Schematic of X-ray emission during the electron-atom interaction (https://nptel.ac.in/courses/115103030/module3/lec11/4.html).

Consequently, these X-rays have a characteristic energy whereby the parent atom can be traced. The best way to quantify the analysis of targeted element is to compare the peak intensities to the intensity of the pure elements collected under identical condition, which usually needs a ZAF correction:

K-ratio = Intensity Element Sample

Intensity Pure Element , (2.5)

Weight % = K−ratio_{Z ∙A∙ F} × 100% (2.6) where Z describes how the electron beam penetrates in the sample (Z-dependent and density dependent) and loose energy, A is the absorption correction and F is the fluorescence correction.

However, compared with wavelength dispersive spectroscopy (WDS), EDS is normally used as a quantitative method to determine the concentration of atom distribution and is particularly less accurate for lighter elements. Note the accurate electron beam is only able to irradiate a rather small area, whereas a large interaction volume can be activated where the X-rays are originating, which is largely dependent on the electron accelerating voltage resulting in the order of a few micrometers deep and wide. EDS can be performed in small area (spot), linear line scan and area elemental mapping to distinguish elemental variations for providing additional information combined with the microstructure analysis. Considering the elements of interests are C, S, O and W, we employed 20 kV as the accelerating voltage in SEM. The atomic percentage was averaged from three results from random areas in spot mode.

XPS

In contrast with EDS capable of digging out elemental information of several micron deep, X-ray photoelectron spectroscopy (XPS) is a very surface sensitive technique with an information depth about less than 10 nm for perpendicular to the surface incident X-rays.

In principle, the photoemission process involves three steps: (a) photoelectrons are generated through the interaction of the X-ray with atomic core level electrons; (b) the photoelectrons move though the sample to the surface with some inelastically scattered along the way; (c) the electron escaping the surface are emitted in the vacuum and absorbed into the analyzer. The photoelectron without any inelastic scattering will appear as narrow lines in the spectrum, while those losing energy will be part of the background. The X-ray energy hυ is absorbed by core level electrons with binding energy Eb, leading to emitted photoelectrons with kinetic energy Ek, which can be measured by the electron energy analyzer. According to the photoelectric effect demonstrated by Einstein[18],

EB = hυ - EK - Φ (2.7)

where Φ is the work function. The binding energy Eb is characteristic of the energy transition at the parent atom. The intensity of the XPS spectrum is a function of the binding energy of the photoelectron. The detailed energy spectrum could reveal chemical shifts in the atom bonding, where the valence of the atoms in a certain structure or molecule as well as the stoichiometry can then be determined. In this project, XPS measurements were performed with a Surface Science SSX-100 ESCA instrument equipped with a monochromatic Al K X-ray source (hυ = 1486.6 eV) at vacuum pressures lower than 2 × 10-9 _{mbar during data acquisition. The electron} take-off angle with respect to the surface normal was 37°. Due to the spot size of the X-ray beam, the lateral resolution is generally limited to a spot with a diameter of 1000 µm. The energy resolution was 1.16 eV and 0.1 eV for full and local elemental spectra, respectively. Binding energies were plotted with respected to the carbon 1s photoemission peak at about 285.6 eV. At least three different spots were measured on each sample for reproducibility. Note that due to the element of sulfur in the investigated coatings is very sensitive to Ar+ _{ion bombardment, potentially leading to} sulfur loss, therefore no prior surface cleaning process by bombardment was executed. Instead, measurements were directly conducted on freshly prepared samples. After that, XPS spectra were analyzed using the least-squares curve-fitting program Winspec software developed at the LISE Laboratory, University of Namur, Belgium. Raman

2

2

Raman spectroscopy is highly sensitive to study the structural properties of amorphous carbon materials. For instance, diamond like carbon (DLC) is a metastable form of amorphous carbon containing a mixture of Sp2 _{and Sp}3 _{type C bonds. The} Raman spectra of disordered graphite show two quite sharp modes, the G peak around 1580–1600 cm-1 _{and the D peak around 1350-1380 cm}-1_{, usually assigned to zone} center phonons of E2g symmetry and K-point phonons of A1g symmetry, respectively [19]. The G band is the first-order Raman band of all Sp2 _{hybridized carbon materials.} The D band is a defect activated band in Sp2 _{hybridized carbon materials. Based on the} ratio of D and G bands, we can get an estimation about defect densities and as well as Sp2 _{and Sp}3 _{ratio. WS2 peaks are also Raman active, whereby Raman spectrum could} compare the relative intensity of DLC/WS2 at various situations.

In this work, Raman spectra were also recorded using an Olympus BX51M upright microscope with excitation at 632.8 nm (Thorlabs HNL 120-1 HeNe laser) and around 1-2 mW at a sample, with appropriate laser line clean up filters from Semrock. Excitation was delivered using a dichroic mirror (Semrock) and light collected via a round to line multicore fiber (which acted as slit) and delivered to a Shamrock 163 spectrograph and dispersed with a SRT-SHT-9003 grating onto a iDus-418 CCD detector (Andor Technology). Calibration was performed using the spectrum of polystyrene. In general, Raman measurement needs very little sample preparation process, therefore the Raman measurements were directly conducted on the bulk raw coating, worn surfaces on the wear track and counterpart surface. By adjusting suitable laser power intensity without irradiating damages onto the samples, rapid nondestructive spectrum can be easily achieved.

2.3.3 Mechanical testing

Nanoindentation is a useful method to probe the local material properties of a solid. Its effective use lies in interpreting the data collected from a nanoindentation experiment with an associated analytical solution. Nanoindentation builds on a relationship between force, displacement and the contact area of the indenter (punch) for a linear elastic material. Based on the measured indentation load-displacement curves (see Figure 2.11), the influences on the non-rigid indenters on the load-displacement behavior can be effectively accumulated via defining a reduced modulus, 𝐸𝑟 through the equation [20]:

1 𝐸𝑟

=

+

;

S =

=

𝐸

√𝐴

(

Figure 2.11 A schematic representation of load versus indenter displacement data for an indentation experiment. The quantities illustrated as Pmax: the peak indentation load; hmax: the indenter displacement at peak load; hf: the final depth of the contact impression after unloading; and S: the initial unloading stiffness [20].

where E and

𝑣

𝑣i are the same parameters for the indenter, H is the hardness, Pmax is the peak indentation load, A is the projected area of the hardness impression. S = dP/dh is the experimentally measured stiffness of the upper portion of the unloading data. By measuring the initial unloading stiffness and proposing that the contact area is equal to the optically measured area of the hardness, the modulus can be thus be derived. For plotting the mechanical properties of thin films by nanoindentation, normally the depth-controlled method is used, with the indented depth controlled less than the 1/10 of the thickness of thin films or coatings to avoid the influence from the bulk substrates. Also note that modern theories of wear point out that while the hardness of the coating is an important factor determining the wear property, other parameters should also be taken into account. They can be H/E (the elasticity index, related to the elastic strain to failure, a higher H/E ratio points to a higher wear resistance) and H3_/E2 _{(related to the resistance to plastic deformation). Sometimes a lower Young’s} modules of a coating contribute to the distribution of the applied load over a larger area and reduce the peak contact stress. The resistance of a coated surface to plastic deformation, i.e., its yield strength, is proportional to H3_/E2 _{as the expression of the} yield pressure in a rigid ball contact on an elastic/plastic surface is according to the equation:

P = 0.78 r2_{× (}𝐻3

2

2

where the r is the radius of a rigid ball. As can be seen, a higher H3_/E2 _{supposes to result} in an enhanced elastic recovery of the coating under contact.

2.3.4 Tribometry

The coefficient of friction (CoF, µ), is the ratio between the friction force (Ft) and the normal load (FN) and could be given by the following equation:

µ = 𝐹𝑡

𝐹𝑁 (2.12)

Friction between two surfaces in relative motion is a complex phenomenon that involves phonon dissipation, deformation and ploughing of asperities or debris, bond breaking and formation, strain-induced structural transformation and local surface reconstruction and adhesions between the contacted surfaces [21]. Earlier studies indicated that friction is influenced by the real contact area and the type and the strength of the bond formed at the interface of the surfaces and also whether the material around the contacted area is sheared and ruptured during sliding [22]. Coulomb, however, found that the friction force to resist sliding is independent of the contact area and independent of the sliding velocity, however, this friction law was challenged by the developed self-lubricating nanocomposite coatings where a higher velocity may facilitate a reduced friction due to the formation of a thin transfer film resulting in that interfacial sliding actually takes places between the transfer film on the ball and the surface of the coating [21].

Similarly, wear is defined as the volume of materials removed from the surface during the sliding. Accordingly, the wear rate Wr of the coatings is defined as the volume of wear per unit track length, per Newton of normal load and per lap (mm3_{/Nm) and is} thus calculated according to the following equation:

WR = V/(L × s) (2.13)

where V is the wear volume, s the total running distance of the ball over the disk, and L the normal load. The lower WR, the higher resistance of the material to wear removal. Normally there is a trade-off between CoF and WR and to fabricate coating with both a low CoF and higher wear resistance is among the tribologists’ long-lasting pursuit. The main research task in this thesis is to investigate the friction and wear behavior of the WS2/a-C(H) nanocomposite coatings by using a ball-on-disk high-temperature enabled tribometer (CSM Instruments, Switzerland) under both dry and humid sliding conditions, as displayed in Figure 2.12. It should be pointed out that during tribological tests the counterpart steel ball was stationary, while the sample disks were rotating anti-clockwise underneath the stationary ball. The normal load was transferred from a horizontal arm (where the ball holder is fixed) to the ball/sample contact interface.

The normal load used in study was normally fixed at 5 N yielding a Hertz pressure of about 0.75 GPa, which could avoid plastic deformation of the substrate underneath the coating and thus is suitable to investigate the potential self-lubricating mechanisms provided by the coating itself. The linear sliding velocity is set at 10 cm/s.

Figure 2.12 (a) Image of the high-temperature CSM ball-on disk tribometer; (b) the schematic illustration of the rotational sliding configuration of the ball-on-disk tribometer with relevant parts as indicated.

The counterpart balls used in the tribo-tests were ø6 mm ball, including 100Cr6 bearing steel balls and Si3N4 balls, normally with the later used for the high temperature tribotesting. Before the tribo-tests, the counterpart balls were cleaned with acetone and dried with clean dry compressed air for three times. Since WS2-based coating is rather sensitive to humidity, a wide range of relative humidity (RH%) ranging from 5% to 55% was chosen. The humidity was maintained with a feedback controlled (home-made humid adjustor) flux of dry air or water vapor into the protection container, depending on the ambient humidity. For the tribotests in high-temperature explorations, the tribometer could continuously increase the temperature until 600 ℃ in a heating burner.

After the tribo-tests, the morphology of the worn surfaces of the composites and of the balls were observed using optical microscopy and scanning electron microscopy EDS was used to analyze the elemental composition on the worn surfaces on the wear track and the counterpart ball scar. Confocal microscopy (Nanofocus μSurf) was used to measure the surface profile of the worn surfaces of the composites and the balls at different positions, for the assessment of the wear volume by a Matlab code with an error of ± 5% for most cases. The Matlab code is attached in the Appendix 1 at the end of the dissertation. In the ball-on-disk configuration, an area of the ball is continuously in contact with the coating, whereas the corresponding areas on the coating are in contact only once during a lap. The sliding stopping condition can be dependent either

2

2

on the sliding distance (equal to the product of laps and perimeter of the wear track) or the sliding laps, and will be detailed in later chapters.

R

[1] D.M. Mattox, Handbook of physical vapor deposition (PVD) processing second edition, Elsevier, 2009.

[2] T.L. Alford, Atomic collisions and backscattering spectrometry, in: Fundam. nanoscale film anal., Springer US, Boston, MA, 2007: pp. 12–33.

[3] W. Bauer, Intrinsic resputtering during film deposition investigated by Monte Carlo simulation, J. Vac. Sci. Technol. 12 (1994) 3157-3164.

[4] D.W. Hoffman, Intrinsic resputtering-theory and experiment, J. Vac. Sci. Technol.

8 (1990) 3707–3712.

[5] M. Ohring, Materials science of thin films: deposition and structure, Second edition, Academic Press, 2011.

[6] K.H. Müller.

Ion beam induced epitaxial vapor-phase growth: A

molecular-dynamics study

[7] A.A. Voevodin, J.P. O’Neill, J.S. Zabinski, Nanocomposite tribological coatings for aerospace applications, Surf. Coat.Technol. 116–119 (1999) 36–45.

[8] H. Dimigen, H. Hubsch, P. Willich, K. Reichelt, Stoichiometry and friction properties of sputtered MoSx layers, Thin Solid Films 129 (1985) 79–91. [9] M. Evaristo, T. Polcar, A. Cavaleiro, Tribological behaviour of C-alloyed transition

metal dichalcogenides (TMD) coatings in different environments, Int. J. Mech. Mater. Des. 4 (2008) 137–143.

[10] R.S. Goeke, S. Prasad, T.W. Scharf, P.G. Kotula, Synthesis of MoS2-Au nanocomposite films by sputter deposition, Sandia Report (2012), Unlimited Release.

[11] D. Galvan, Nanocomposite Coatings: Processing, structure and tribological performance, Thesis , University of Groningen, 2007.

[12] Y. Leng, Materials characterization: Introduction to microscopic and spectroscopic methods. Wiley, 2008.

[13] D.B. Williams, C.B. Carter, Transmission electron microscopy: A textbook for materials science, Springer, Boston MA, 1996.

[14] L. Reimer, Energy-filtering transmission electron microscopy, Springer, 1994. [15] N. Yao, Focused ion beam systems: Basics and application, Cambridge University

Press, 2007.

[16] P. Eaton, P. West, Atomic force microscopy, Oxford University Press, 2010. [17] B.J. Inkson, 2-Scanning electron microscopy (SEM) and transmission electron

microscopy (TEM) for materials characterization, in: G. Hübschen, I. Altpeter, R. Tschuncky, H.-G.B.T.-M.C.U.N.E. (NDE) M. Herrmann (Eds.), Woodhead Publishing, 2016: pp. 17–43.

[18] A. Einstein, Über einen die Erzeugung und Verwandlung des Lichtes betreffenden heuristischen Gesichtspunkt, Ann. Phys. 322 (1905) 132–148.

[19] A.C. Ferrari, J. Robertson, Interpretation of Raman spectra of disordered and amorphous carbon, Phys. Rev. B 61 (2000) 14095–14107.

[20] W.C. Oliver; G.M. Phar, An improved technique for determining hardness and elastic modulus using load and displacement sensing indentation experiments, J. Mater. Res. 7 (1992), 1564-1583.

[21] J.T.M. De Hosson, Y.T. Pei, C.Q Cheng, Jerky-type phenomena at nanocomposite surfaces: The breakdown of the Coulomb friction law, JOM 59 (2007) 45–49. [22] D. Tabor, Friction—the present state of our understanding, J. Tribol. 103 (1981)