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.

Self-adaptive and Self-healing

Nanocomposite Tribocoatings

ISBN: 978-94-034-1547-5 (printed version)

ISBN: 978-94-034-1546-8 (electronic version)

ENTEG Institute PhD thesis series 2019

ISBN: 978-94-034-1547-5

Self-adaptiv

e and S

elf-healing N

anocomposite

Tribocoatings

H

uatang C

ao

replenish

transfer film

tribofilm

W

S

O

_coating

5 nm

Self-adaptive and Self-healing

Nanocomposite Tribocoatings

The research presented in this thesis was performed in the Advanced Production Engineering (APE) group of Engineering and Technology institute Groningen (ENTEG) and Materials Science Group of Zernike Institute for Advanced Materials at the University of Groningen, The Netherlands.

This research was carried out under financial support by Chinese Scholarship Council (CSC, No. 201406160102).

Print: Zalsman Groningen B.V. Cover design: Huatang Cao

ISBN: 978-94-034-1547-5 (printed version) ISBN: 978-94-034-1546-8 (electronic version)

Self-adaptive and Self-healing

Nanocomposite Tribocoatings

PhD thesis

to obtain the degree of PhD at the

University of Groningen

on the authority of the

Rector Magnificus prof. E. Sterken

and in accordance with

the decision by the College of Deans.

This thesis will be defended in public on

Monday 25 March 2019 at 11.00 hours

by

Huatang Cao

born on 20 June 1987

in Hunan, China

Supervisors

Prof. Y.T. Pei

Prof. J.T.M. De Hosson

Assessment Committee

Prof. A. Cavaleiro

Prof. J. Post

Prof. A. Vakis

Chapter 1 Introduction ... 1

1.1 Tribology ... 1

1.2 Solid lubrication ... 3

1.3 Self-adaption and self-healing funcationality ... 7

1.4 Outline of the thesis ... 10

Reference ... 12

Chapter 2 Synthesis, characterization and testing of WS2/a-C(H) tribocoatings ... 17

2.1 Physical vapor sputtering ... 18

2. 2 Coating deposition ... 20 2.3 Characterization techniques ... 22 2.3.1 Microstructure characterization ... 22 2.3.2 Chemical composition ... 29 2.3.3 Mechanical testing ... 33 Reference ... 36

Chapter 3 Effect of carbon concentation and argon flow rate on the microstructrue and tribological properties of magnetron sputtered WS2/a-C coatings ... 39

3.1 Introduction ... 40

3. 2. Experimental procedures ... 41

3.2.1 Preparation of the WS2/a-C coatings ... 41

3.2.2 Characterization of the WS2/a-C coatings ... 42

3. 3. Results and discussions ... 46

3.3.1 Basical charcterisitic of WS2 ... 47

3.3.2 Chemical and structural characterization ... 48

3.3.3 Mechanical and tribological properties ... 51

3.4 Conclusion ... 58

Reference ... 59

Chapter 4 On the S/W stoichiometry and triboperformance of WSxC(H) coatings deposited by magnetron sputtering ... 63

4.1 Introduction ... 64

4.3.1 Chemical composition and structural characterization ... 66

4.3.2 Mechanical properties ... 74

4.3.3 Tribological properties ... 76

4.4 Conclusion ... 81

Reference ... 81

Chapter 5 Instant WS2 platelets reorientation of self-adaptive WS2/a-C tribocoating ... 87

5.1 Introduction ... 88

5. 2. Experimental procedures ... 88

5. 3. Results and discussions ... 89

5.3.1 Microstructure characterization ... 89

5.3.2 Tribological properties ... 90

5.3.3 Self-adaptive mechanism ... 91

5.4 Conclusion ... 94

Reference ... 94

Chapter 6 On the self-healing peroformance of WS2/a-C tribocoatings ... 97

6.1 Introduction ... 98 6. 2 Experimental procedures ... 99 6. 3 Results ... 100 6.3.1 Microstructural characterization ... 100 6.3.2 Self-healing of damages ... 101 6.4 Discussions ... 112 6.5 Conclusion ... 120 Reference ... 120

Chapter 7 Summary and outlook ... ……….……125

7.1 Summary ... 125

7.1 Outlook ... 127

Samenvatting ... ……….……133

Appendix ...137

1

C

HAPTER

1

This chapter highlights the motivation and objectives of the thesis work on the self-adaptive and self-healing tribocoatings and provides an overview of the structure of the dissertation. It briefly reviews how transition metal dichalcogenides (TMD), WS2

in particular, based coatings can be used in triboapplications and the reasons why they can be beneficial when embedded in an amorphous carbon matrix. It is shown that the nanocomposite characteristics of TMD in an amorphous carbon matrix provides capabilities of self-adaptation and of self-healing in different tribo-environments.

1.1 T

RIBOLOGY

The surface of a component is usually the most important engineering component. In practice it is often the surface of a work-piece that is subjected to wear and corrosion, which causes losses or failures in functionality. The complexity of the tribological properties of materials and the economic aspects of friction and wear justify an increasing research effort. In industrialized countries around 30 % of all energy generated is ultimately wasted through wear and friction. In highly industrialized countries losses due to friction and wear are estimated between 1% and 2% of Gross Domestic Product (GDP) [1].

Wear from sliding surfaces is typically the result of one or more of the following main mechanisms [2,3]: (i) abrasive wear, whereby a hard counterface slides across a softer surface, which includes two-body abrasive wear (e.g. hard particles are generated and adhered to the surface) and third-body abrasive wear (e.g. hard particles are free to roll and slide); (ii) adhesive wear, whereby material is removed (material transfer) due to the adhesive forces; (iii) fatigue wear, whereby wear particles are detached due to crack nucleation and crack propagation in the subsurface area; and (iv) tribo-chemical wear, whereby wear particles are generated in a corrosive environment by a chemical reaction (e.g. corrosive wear) or a chemical reaction owing to friction or mechanical energy (e.g. tribo-chemistry). One should realize that wear is a rather complex process with many influential factors such as kinematics, mechanical stress, the formation of a lubricating tribofilm, chemical environment, temperature, humidity, and materials properties. In the evaluation of the tribological properties of a materials, the extrinsic conditions should be emphasized.

To an increasing degree continuous search is on the surface modification techniques and advanced coatings or films that can largely decrease friction and increase the wear resistance of the protected materials. There have been significant progress in surface modification and engineering fields providing very thick, hard, slippery and polished surfaces for severe tribological applications. Advancing wear resistance is typically accomplished by the introduction of a shear-accommodating layer between contacting surfaces. For instance, the field of plasma vacuum surface engineering utilizes a family of plasma based surface modification technologies, such as plasma-assisted chemical vapor deposition, ion beam implantation & deposition, magnetron sputtering, cathodic vacuum arc and plasma-enhanced electron beam evaporation. The layers produced with these techniques are called “thin films”, limiting to several µm thick, in comparison to “thick coatings” that can be built-up in a large scale up to several mm thick such as by laser surface modification (e.g. cladding, hardening, alloying) or

1

protective layer is sufficient to do the job, it is not necessary to produce a thick layer on the top of a substrate. This thesis highlights the novel thin functional protective thin coating (< ~ 2 µm) prepared by magnetron sputtering for advances in sliding wear and low friction.

1.2 S

OLID LUBRICANTS

:

DLC

AND

TMD

Liquid lubricants (e.g. oils) are used to suppress contact pressure, to reduce friction and to facilitate the sliding between solid bodies. However, in many industrial applications where the operating conditions are beyond the liquid realm: the presence of a liquid is inappropriate or even fully impossible to be introduced (e.g. the vacuum environments and high-temperature applications, or situations with incompatibility problems). Thus, for these reasons liquid lubricants should be avoided. Consequently, the design and development of solid self-lubricating coatings becomes an attractive subject of fundamental and applied research. The solid lubricating behavior of materials such as graphite and mica are well developed. Recent modern coating technologies (e.g. PVD) has greatly expanded the application of solid lubricant materials including diamond-like carbons (DLC) and transition metal dichalcogenides (TMD).

Amorphous carbon consist of a disordered network of carbon atoms with a mixture of both sp3_{- and sp}2_{-coordinated bonds. The family of amorphous carbon is called}

diamond-like carbon (DLC) [4,5]. In DLC, the carbon sp3 _{bond hybridization results in}

a diamond-like structure, which provides properties such as extremely high hardness, chemical inertness and low wear rate. The carbon sp2 _{bond hybridization represents a}

typical component of graphite which is internally coupled through weak Van der Waals force fields. DLC films can be in divided into two categories based on their hydrogen content, hydrogenated (a-C:H, 10-60 at.% H) and hydrogen-free (a-C, < 1 at.% H). Amorphous carbon films with a high fraction (> 50-90%) of tetrahedral sp3 _bond

called ta-C [5]. Comparted with conventional DLC, ta-C coating is favorably used in applications requiring ultrahigh strength and hardness such as tooling due to its super-high hardness. Also, it should be pointed out that the test environment has a profound influence on the friction and tribological performance of hydrogenated and hydrogen-free DLC films, respectively, e.g., in inert dry gases the coefficient of friction (CoF) of highly hydrogenated DLC is very low (< 0.02) in comparison to an increased value to 0.1-0.2 in humid air. On the contrary, such ultralow CoF values are typically measured for the hydrogen-free DLCs in humid environment, whereas they perform unsatisfactorily in dry conditions (CoF 0.4-0.7) [6]. The lubricating mechanisms for DLC have attributed to the hydrogen-terminated bonds with Van der Waals forces

acting in the contact between surfaces and the graphitization at micro-contract regions leading to transfer films [6]. In all, DLC has many significant features that contribute to excellent tribological characteristics, such as high hardness, anti-wear property with low friction coefficient and low wear rate as well as chemical inertness. DLC is probably the only material that can provide both high hardness and low friction under dry sliding conditions [7]. DLCs, however, face several drawbacks: for instance, it is only possible as a thin film material, not as a bulk glassy solid; besides its thin films tend to have a large residual stresses. In addition, it is not mechanically tough and it is usually alloyed with metals or other elements (a-C: Me) to solve these problems. Still, the thermal stability of DLC is relatively low (< 400 ℃), potentially limiting its application in high temperature.

Alternatively, appealing lubrication properties are also found in other compounds with a layered crystal structure, like transition-metal dichalcogenides (LTMD), i.e., compounds of transition-metal atoms (group Xe: IVb, Vb, VIb) and chalcogens (group X: S, Se, Te). The chemical bonds between the atoms within the sandwich are relatively powerfully covalent, whereas the bonding between adjacent sandwiches is weaker Van der Waals interactions [8], as illustrated by the structure of WS2 in Figure 1.1.

Consequently, glide occurs preferentially between the lamellae leading to the ultralow friction coefficient. Given the weak interactions between the sandwiched layers, they can be stacked into several ways (e.g. 1T, 2H, 3R) of which the most common form has a stacked sequence of two sandwiched layers (the stacking sequence is

X-Me-X-X-Me-X and is denoted 2H-MeX-Me-X-X-Me-X2), giving a trigonal symmetry and exhibiting a marked

two-dimensional behavior despite its three-two-dimensional atomic structure. In this anisotropic structure, a close-packed layer of transition-metal atoms is sandwiched in between two layers of chalcogen atoms. Each chalcogenide atom is equidistant from three metal atoms, and each metal atom is surrounded by six equidistant dichalcogenide atoms at the corners of a small trigonal prism.

Crystalline TMD coatings are often categorized into two categories depending on the orientation of the (002) basal planes. As illustrated in Figure 1.2, for typeⅠorientation the basal planes are perpendicular to the surface (namely with the c axis is parallel to the surface), while the typeⅡorientation has basal planes parallel to the surface (and the c axis vertical to it). Unless there are defects, the basal plane surface is unreactive while dangling bonds are found on the plane edges such as (110), which are active sites for environmental attacks such as oxidations. Therefore, the typeⅡorientation with horizontal basal planes is favorable for tribological properties and efforts have been made to grow such coatings [9–12], otherwise the resulting products are often typeⅠ orientation with standing platelets, making it highly porous with poor adhesion, as

1

appearance due to less reflection of visible light.

Figure 1.1 Crystal structure of 2H-WS2: (a) side view; (b) top view (Red: sulfur atoms;

Blue: tungsten atoms).

Figure 1.2 (a) edge oreintaiton and (b) basal orientation.

Due to the weaker interlayer coupling, disordered stacking is also possible, for instance in the so-called turbostratic stacking, the sandwich layers are parallel but more or less randomly rotated around the c-axis [13,14]. Besides, other intercalants can also be inserted in between the layers [15]. In terms of lubrication, various mechanisms such as basal plane orientation, intra-granular shear and inter-crystallite glide were proposed for explaining the extraordinary lubrication [16]. For instance, Sokoloff [17] and Shinjo [18] pointed out that the intra-granular shear of MoS2 is associated with the

the counterpart. In all, this anisotropic property lies the physical foundation for explorations of 2H-TMDs as superior solid lubricants.

Figure 1.3 porous pure WS2 film: (a) by pulsed-DC magnetron sputtering in this study;

(b) by radio frequency magnetron sputtering in Ref. [23].

The LTMD family covers disulfides, diselenides and ditellurides of molybdenum, tungsten or niobium. The physical and tribological properties of LTMDs, particularly MoS2 and WS2, have been intensively studied for the last decades in lubrications [8]

Nowadays, they are used mainly as oil additives playing a low-frictional role. They are also being prepared as a thick coatings or a thin film deposited mainly by physical vapor deposition (PVD) method. Other TMDs such as diselenides and ditellurides of molybdenum and tungsten, although with similar structure to WS2 and MoS2, have not

demonstrated as lubricious as the first two and are of less technological interests

generally [19,20]. Compared with MoS2, WS2 is even more advantageous due to the

following reasons: (a) higher thermal stability (WS2 was reported to be about ~ 400 °C

while MoS2 starts to oxidize above ~ 300 °C [21,22]; (b) potential lower friction (WO3

is even superior to MoO3 at high temperature) and (c) less volatile. Therefore, WS2 is

attracting more scientific and industrial attentions.

TMD lubricants are mostly finding their wide applications in aerospace industry or dry environments. The large-scale industrial applicability of TMD is still limited and the operational impact is constrained by environmental issues particularly of humidity and elevated temperature. The main problems to face are as follows [24,25]:

 LTMD films are porous with potential cracks mostly due to the columnar morphology, see Figure 1.3. In reality, it turned out to be very difficult to produce pure LTMD films without cracks and porosity defects using conventional magnetron sputtering.

 Only (002) basal plane orientation is functional for a superlubricity. However, deposition of LTMD by magnetron sputtering mostly leads inevitably to a

1

which may cause a lack of replenishment.

 LTMD films are extremely sensitive to environmental attacks. When sliding in humid air, very likely a reaction occurs between the unsaturated dangling bonds leading to metal oxides MO3, see Figure 1.4. Only well orientated LTMD

with basal planes parallel to the surface may help resist oxidation.

 Extremely low hardness (0.3-2 GPa) with low load bearing capacity (corresponding to a large contact area). It is easily to peel off under high contact pressure and thus yield high wear rate particularly in high temperature applications.

As can be seen, all these above drawbacks make LTMD based coatings less attractive for operations in harsh environments and under high contact load. To achieve low friction and a low wear rate, other approaches for advancing TMDs are needed.

Figure 1.4 HR-TEM images indicating the formation of WO3 formed in WS2/a-C coating,

which may impede the sliding induced WS2 (002) basal plane reordering process as a

result of (a) blocking and (b) diverging [26].

1.3 S

ELF

-

ADAPTION AND SELF

-

HEALING FUNCTIONALITY

In the past decades, tribological coating and solid lubricant films developed towards from the (a) single component; (b) multicomponent; (c) multilayer; (d) nanostructured to the (e) smart adaptive coating. There is an increasing demand for environmentally robust solid lubricant coatings that can adapt themselves to different environments [27–29]. For instance, even if the targeted application is the friction and wear mitigation in space, usually the satellites and satellites launch vehicles requires some extended periods of time in humid coastal environments before launching [20]. This potentially expose the moving mechanical assemblies to humid environment,

which calls on development of multi-phase, nanocomposite materials known as adaptive lubricants since no single phase, either DLC or TMD, currently could provide solid lubrication by themselves under varying degrees of humidity. In an ideal case, chameleon wear protective coatings should be able to reversibly self-adapt their surface chemistry and structure to maintain low friction and wear while cycling over a broad range of ambient environments and temperatures. Self-adaptive coating selection is guided by several factors including operational environment, temperature, geometry, load, and speed considering that different materials perform optimally under a limited range of environmental and loading conditions. Chameleon tribological coatings are being designed to tackle the existing problem: single-constituent tribological materials cannot operate reliably over extended ranges of ambient humidity or temperatures. The key feature of chameleon tribocoating concept is to provide reversible self-adjustment of surface chemistry, structure, and mechanical behavior in the contact zone to obtain low friction and prevent wear at all anticipated operational conditions.

Figure 1.5 Schematic representation of a solid lubrication with a chameleon adaptive

coating with environmental cycles [2,30].

The compliant amorphous characteristic of DLC in a nanocomposite coating render it to be dense and homogenous, generating a high density of interphase interfaces that assist in crack deflection and termination of crack growth and also increase in the load-bearing capacity as well as protection of TMD from oxidation [31,32]. All these features make DLC a promising matrix candidate for supporting the soft LTMD for self-adaptive functionality. In 1999, Voevodin and co-authors were the first to present a

1

combined WS2/graphite target. The motivation was to combine the low friction and

high wear resistance of amorphous DLC with the wear resistance of WC and the lubricating property of WS2. The assumption is that TMD phase is attributed to be

responsible for the lubrication and low-friction properties in dry environment, whereas DLC plays a role in lubrication in high humidity. Also it offers high strength and resistance to wear, and the coating becomes versatile in switching environments and present “chameleon behavior”, as illustrated in Figure 1.5. These smart coatings, at least consisting of one TMD component (WS2 or MoS2), were largely developed later

in combination with a-C(N)[35,36], yttrium-stabilized zirconia (YSZ)/Au [37], Mo2N/Ag [38], Ti [15], Cr [39], Sb2O3/Au [27], Pb [40]. The basic lubricating

mechanism underlying the self-adaptation was proposed that in vacuum or dry gas conditions, friction induces TMDs to be crystallized and reorientated with their basal planes parallel with the sliding direction although sputtering nanocomposite of DLC and TMD may produce type I orientation TMD or even an amorphous state, as illustrated in Figure 1.6. While in humid air, graphite-like transfer film is formed by friction induced sp3 _{→ sp}2 _{phase transition leading to low friction; however, the role of}

carbon is still controversial [37,39,41,42]. Besides, the YSZ, Au, Ag, MO3 were reported

as beneficial in high temperature triboapplications.

Figure 1.6 Schematic representations of (a) two crystallographic growth textures with basal planes perpendicular or parallel (preferred) to the substrate, and (b) amorphous structure. The process of sliding induced (a → c → d) orientation of perpendicular (or randomly oriented) basal planes parallel to the sliding direction to allow ultralow friction, or (b → c → e) amorphous to crystalline transformation to achieve low friction. Corresponding

cross-section TEM images from the wear track of (d) crystalline MoS2/Au coating and (e)

amorphous MoS2/Sb2O3/Au coating [20].

Self-healing materials possess the inherent capability to repair damage by themselves, or alternatively with some external stimulation such as temperature, light, chemical change, pressure change or some mechanical actions to adaptively reproduce or restore their original functionality [43]. It is known that coating prepared by PVD are mostly columnar-like, such columns are potential weak links for crack propagations (see Figure 1.7), which may consequently lead to catastrophic failures such as branch or delamination. In real severe wear conditions, accumulating fatigue or some sudden impact or damage further aggravate the degradation of coatings. It took numerous efforts for surface scientists and engineers to counteract the growth of columnar structures [31,44]. Therefore, the damage self-healing ability is a particularly intriguing property for tribocoatings to protect the material of a construction against wear.

Figure 1.7 BF-TEM images of the sputtered TiC/a-C:H coating with columnar structure as indicated by an arrow in (a, b) and (c) shows a crack may propagate through a column boundary in the coating leading to potential failure. Note the circle in (c) is due to the presence of a sample –supporting holey carbon grid [45].

1

micocracks are healed and the tribo-system becomes repairing and even self-curing. We intend to prepare WS2 coating alloyed with amorphous carbon by using the

sophisticated magnetron sputtering method. In particular, we have abandoned the standard idea to deposit “ideal” coating; instead, we focus on exploitation of the self-adaptive and self-healing capabilities of tribocoatings, as well as the high flexibility of the coating operating in different environments. The coating itself will consist of two different compounds: WS2 in the form of nano-grains or nano-platelets as lubricants

and amorphous carbon acting as the supporting matrix. The research follows the idea to allow the reorientation of the crystallized WS2 phases to the “frictionless” direction

inside the carbon matrix.

In Chapter 2, the techniques employed for the production of WS2/a-C(H) coatings, and

the subsequent experimental methods applied for the sample characterization and data analysis are described in details. Closed-field unbalanced magnetron sputtering was employed to prepare the expected tribocoatings. Characterization tools including scanning electron microscope (SEM), high resolution/energy-filtering transmission electron microscope (HR-/EF-TEM), 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 technologies were used to determine the coating morphology, chemical composition, phases, microstructure and their mechanical and tribological properties. Focused ion beam (FIB) was used to prepare site-specific TEM lamella at the wear track or the ball scar for HR-TEM to evaluate the self-adaptive mechanism in the coatings.

In Chapter 3, WS2/a-C coatings with various carbon contents (0-65 at.%) were

deposited on single crystal silicon wafers by magnetron co-sputtering one graphite and two WS2 targets under different Ar flow rates. Thus, a broad spectrum of coatings were

prepared under varied sputtering parameters provided sufficient information for the determination of the microstructure, the phase, and the mechanical properties. The effects of the argon flow rate (deposition pressure) on the chemical stoichiometric S/W ratio and coating density were investigated. At the same time, the effect of carbon on the S/W ratio was also investigated on the nano-hardness and elastic modulus and tribological properties. The mechanical and tribological properties of pure WS2 coating

and nanocomposite WS2/a-C are also compared. The coating carbon concentration,

deposition pressure and the testing atmospheres on the friction and wear were studied. Chapter 4 further follows the Chapter 3 to explore some other parameters influencing the S/W ratio in the coating: target-substrate distance and negative bias voltage applied to the substrates. The underlying physical understanding of sulfur preferential

resputtering was presented and the solutions to overcome the sulfur substoichiometry were provided. The real effect of S/W ratio on the tribological properties was discussed. Apart from the S/W ratio, the influence of total content of sulfur in the nanocomposite coating was discussed. In addition, the nonreactive sputtered WS2/a-C and reactive

sputtered WS2/a-C:H coatings with similar composition and mechanical performance

are scrutinized for the tribological tests in both humid and dry sliding conditions. This study also revealed the real role of carbon playing in the self-lubricating in the humid environment. Based on the triboperformance, a specific sputtering route to deposit WS2 based coating is proposed for the favorable development of self-adaptive coatings

versatile in harsh environments.

Chapter 5 reports on the self-adaptive behavior of WS2/a-C coating by switching the

sliding conditions, i.e., alternating the testing atmospheres from dry air (5% relative humidity) to humid air (55% relative humidity) several times. Reversible tribological behavior was exploited with the testing atmosphere cycling between dry air and humid air. Focused ion beam technique was then employed to slice TEM lamellae at both the wear track and the worn ball scar, and HR-TEM observations reveal the self-adaptive mechanism.

Based on the findings in Chapter 5, Chapter 6 further explores the potential self-healing capabilities in tribocoatings from nano- to micro-scale, as illustrated by the nanocomposite WS2/a-C coating. We first use a scratcher to induce two types of

notched cracks to mimic the potential damages. After that, a sliding stimulus was exerted aiming to self-repair the damages autonomously. The FIB-TEM observations reveal the self-healing mechanism. The study sheds new light on the possibility to release the critical requirements of producing flawless coatings for triboapplications in industry.

Chapter 7 summarizes the outcome of the thesis work, and gives an outlook of possible future perspective involving the further optimization of the coatings in extreme service conditions. Possible new developments of advanced versatile healing and self-adaptive coatings are proposed.

References

[1] K. Holmberg, A. Erdemir, Influence of tribology on global energy consumption, costs and emissions, Friction 5 (2017) 263–284.

[2] Q.J. Wang, Y.W. Chung, Encyclopedia of tribology, Springer 2013. [3] B. Bhushan, Introduction to tribology, 2nd_{, Wiley 2013.}

1

experiments to worldwide applications, Surf. Coat. Technol. 242 (2014) 214–225. [6] C. Donnet, A. Erdemir, Tribology of diamond-Like carbon films, Springer 2007. [7] C. Donnet, A. Erdemir, New horizon in the tribology of diamondlike carbon films,

Surf. Eng. 24 (2008) 399–401.

[8] N. Tak, R. October, Direct resolution of the layer lattice of molybdenum disulfide by transmission electron microscopy in relation to lubrication, Wear 124 (1988) 279–289.

[9] C. Muratore, A.A. Voevodin, Control of molybdenum disulfide basal plane orientation during coating growth in pulsed magnetron sputtering discharges,

Thin Solid Films 517 (2009) 5605–5610.

[10] K. Ellmer, C. Stock, K. Diesner, I. Sieber, Deposition of c(perpendicular to)-oriented tungsten disulfide (WS2) films by reactive DC magnetron sputtering

from a W-target in Ar/H2S, J. Cryst. Growth 182 (1997) 389–393.

[11] M. Regula, C. Ballif, M. Remškar, F. Lévy, F. Le, Crystallinity and texture promotion in crystallinity and texture promotion in WS2 thin films, J. Vac. Sci. Technol. A

15

(2014) 2323–2328.

[12] V. Weiß, S. Seeger, K. Ellmer, R. Mientus, Reactive magnetron sputtering of tungsten disulfide (WS2-x)films: Influence of deposition parameters on texture,

microstructure and stoichiometry, J. Appl. Phys. 103502 (2013) 1–9.

[13] D. Yang, R.F. Frindt, Powder X-ray diffraction of turbostratically stacked layer systems, J.Mater. Res. 11 (1996) 1733–1738.

[14] J. Moser, F. Lévy, Random stacking of MoS2-x sputtered thin films, Thin Solid Films

240 (1994) 56–59.

[15] D.G. Teer, New solid lubricant coatings, Wear 251 (2001) 1068–1074.

[16] J.M. Martin, H. Pascal, C. Donnet, T. Le Mogne, J.L. Loubet, T. Epicier, Superlubricity of MoS2: Crystal orientation mechanisms, Surf. Coat. Technol. 68 (1994) 427–432.

[17] J.B. Sokoloff, Theory of energy dissipation in sliding crystal surfaces, Phys. Rev. B

42 (1990) 760–765.

[18] K. Shinjo, M. Hirano, Dynamics of friction: superlubric state, Surf. Sci. 283 (1993) 473–478.

[19] E. Bergmann, G. Melet, C. Miiller, Friction properties of sputtered dichalcogenide layers, Tribol. Int. 14 (1981) 329–332.

[20] T.W. Scharf, S. V. Prasad, Solid lubricants: A review, J. Mater. Sci. 48 (2013) 511– 531.

[21] W.A. Brainard, The thermal stability and friction of the disulfides, diselenides, and ditellurides of molybdenum and tungsten in vaccum (10-9 _{to 10}-6 _{Torr), NASA}

[22] H.E. Sliney, Solid lubricant materials for high temperatures - A review, Tribol. Int.

15 (1982) 303–315.

[23] D.S. Wang, M. Hu, D. Jiang, X.M. Gao, Y.L. Fu, Q.Q. Wang, J. Wang, Preparation and characterization of the CrN nanocone array textured WS2 film, Mater. Lett. 188

(2017) 267–270.

[24] F. Gustavsson, S. Jacobson, Diverse mechanisms of friction induced self-organisation into a low-friction material - An overview of WS2 tribofilm formation,

Tribol. Int. 101 (2016) 340–347.

[25] T. Polcar, A. Cavaleiro, Review on self-lubricant transition metal dichalcogenide nanocomposite coatings alloyed with carbon, Surf. Coat. Technol. 206 (2011) 686–695.

[26] H.T. Cao, J.T.M. De Hosson, Y.T. Pei, Self-healing of a pre-notched WS2/a-C coating,

Mater. Res. Lett. 07 (2019) 103–109.

[27] T.W. Scharf, P.G. Kotula, S.V. Prasad, Friction and wear mechanisms in MoS2/Sb2O3/Au nanocomposite coatings, Acta Mater. 58 (2010) 4100–4109.

[28] J.J. Hu, R. Wheeler, J.S. Zabinski, P.A. Shade, A. Shiveley, A.A. Voevodin, Transmission electron microscopy analysis of Mo-W-S-Se film sliding contact obtained by using focused ion beam microscope and in situ microtribometer,

Tribol. Lett. 32 (2008) 49–57.

[29] C. Muratore, A.A. Voevodin, Chameleon coatings: adaptive surfaces to reduce friction and wear in extreme environments, Ann. Rev. Mater. Res. 39 (2009) 297– 324.

[30] A.A. Voevodin, C. Muratore, S.M. Aouadi, Hard coatings with high temperature adaptive lubrication and contact thermal management: Review, Surf. Coat.

Technol. 257 (2014) 247–265.

[31] Y.T. Pei, D. Galvan, J.T.M. De Hosson, Nanostructure and properties of TiC/a-C:H composite coatings, Acta Mater. 53 (2005) 4505–4521.

[32] L. Gu, P.L. Ke, Y.S. Zou, X.W. Li, A.Y. Wang, Amorphous self-lubricant MoS2-C

sputtered coating with high hardness, Appl. Surf. Sci. 331 (2015) 66–71.

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

[34] A.A. Voevodin, J.P. O’Neill, J.S. Zabinski, WC/DLC/WS2 nanocomposite coatings for

aerospace tribology, Tribol. Lett. 6 (1999) 75–78.

[35] A. Nossa, A. Cavaleiro, The influence of the addition of C and N on the wear behaviour of W-S-C/N coatings, Surf. Coat. Technol. 142–144 (2001) 984–991. [36] A. Nossa, A. Cavaleiro, Mechanical behaviour of W-S-N and W-S-C sputtered

coatings deposited with a Ti interlayer, Surf. Coat. Technol. 163–164 (2003) 552-560.

1

chameleon coatings for space and ambient environments, Thin Solid Films 515

(2007) 6737–6743.

[38] S.M. Aouadi, Y. Paudel, B. Luster, S. Stadler, P. Kohli, C. Muratore, C. Hager, A.A. Voevodin, Adaptive Mo2N/MoS2/Ag tribological nanocomposite coatings for

aerospace applications, Tribol. Lett. 29 (2008) 95–103.

[39] T. Polcar, F. Gustavsson, T. Thersleff, S. Jacobson, A. Cavaleiro, Complex frictional analysis of self-lubricant W-S-C/Cr coating, Faraday Discuss. 156 (2012) 383–401. [40] H. Li, G.G. Zhang, L.P. Wang, Low humidity sensitivity of MoS2/Pb nanocomposite

coatings, Wear 350–351 (2016) 1–9.

[41] T. Polcar, M. Evaristo, A. Cavaleiro, Self-lubricating W-S-C nanocomposite coatings,

Plasma Process. Polym. 6 (2009) 417–424.

[42] T. Polcar, M. Evaristo, R. Colaço, C. Silviu Sandu, A. Cavaleiro, Nanoscale triboactivity: The response of Mo-Se-C coatings to sliding, Acta Mater. 56 (2008) 5101–5111.

[43] B.J. Blaiszik, S.L.B. Kramer, S.C. Olugebefola, J.S. Moore, N.R. Sottos, S.R. White, Self-healing polymers and composites, Annu. Rev. Mater. Res. 40 (2010) 179–211. [44] Y.T. Pei, C.Q. Chen, K.P. Shaha, J.T.M. De Hosson, J.W. Bradley, S. A. Voronin, M.Čada,

Microstructural control of TiC/a-C nanocomposite coatings with pulsed magnetron sputtering, Acta Mater. 56 (2008) 696–709.

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

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, E,

2

2

[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

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, 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

2

2

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

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.

2

2

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

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.