The Microstructure and Properties of Transition Metal Nitride Nanocrystalline Coatings

Home , Tantalum nitride

By Song Xu

A thesis in fulfilment of the requirements for the degree of

Doctor of Philosophy

School of Material Science and Engineering Faculty of Science

April 2016

PLEASE TYPE THE UNIVERSITY OF NEW SOUTH WALES Thesis/Dissertation Sheet

Surname or Family name: Xu

First name: Song Other name/s:

Abbreviation for degree as given in the University calendar: Ph.D.

School: Materials Science & Engineering Faculty: Science

Title: The microstructure and properties of transition metal nitride nanocrystalline coatings

Abstract 350 words maximum: (PLEASE TYPE)

The composition, microstructure, mechanical properties and biocompatibility of a number of tantalum nitride coatings, deposited on Ti-Al-V substrates by the double cathode glow discharge technique, were investigated. With increasing nitrogen partial pressure, the composition of the tantalum nitride coatings changed from hexagonal Ta2N to fcc TaN. Both coatings exhibited a nanoporous structure comprising fine (~ 10 nm) equiaxed grains together with a homogenous distribution of (~5-10 nm) nanopores. The relatively high hardness and low elastic modulus of these coatings led to improved damage resistance and wear resistance for these tantalum nitride coatings. Further, Ta2N-based coatings also showed good compatibility with hydroxyapatite. However, the introduction of oxygen during the deposition process led to significant degradation of the coating hardness, wear resistance and damage resistance of these coatings. This was due to the presence of Ta2O5 and an amorphous tantalum oxynitride phase arising from the higher oxygen pressure during deposition.

In addition to these tantalum nitride coatings, the composition, microstructure, and mechanical properties of zirconium nitride coatings again deposited by double cathode glow discharge technique were also investigated. Zirconium nitride coatings were deposited on both stainless steel and Ti-Al-V substrates. The zirconium nitride coatings on the stainless steel substrates exhibited a bimodal microstructure with both fine grains and more elongated coarser grains. In contrast, the coatings on the Ti-Al-V substrates showed a uniform microstructure comprising fine equiaxed grains and together with a number of nanopores. All these coatings showed comparable hardness values, but lower elastic modulus values, compared with zirconium nitride coatings deposited by other deposition techniques. Therefore, a relative high wear resistance and damage resistance may be expected for the zirconium nitride coatings studied in this thesis.

By exploring the influence of nanopores on the mechanical properties of nanocrystalline Ta2N coatings, it was found that the presence of nanoporosity may increase wear resistance and damage resistance of these coatings by significantly reducing elastic modulus without greatly decreasing hardness. This suggests a promising new strategy for improving the mechanical properties of nanocrystalline coatings.

Declaration relating to disposition of project thesis/dissertation

I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or in part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all property rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation.

I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstracts International (this is applicable to doctoral theses only).

……………………………………… ………………… ……………………………… ……..…………………….……………… Signature Witness Signature Date

The University recognises that there may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years must be made in writing. Requests for a longer period of restriction may be considered in exceptional circumstances and require the approval of the Dean of Graduate Research.

FOR OFFICE USE ONLY Date of completion of requirements for Award:

THIS SHEET IS TO BE GLUED TO THE INSIDE FRONT COVER OF THE THESIS

‘I hereby grant the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all proprietary rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation. I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstract International (this is applicable to doctoral theses only). I have either used no substantial portions of copyright material in my thesis or I have obtained permission to use copyright material; where permission has not been granted I have applied/will apply for a partial restriction of the digital copy of my thesis or dissertation.'

Signed ……………………………………………......

Date ……………………………………………......

AUTHENTICITY STATEMENT

‘I certify that the Library deposit digital copy is a direct equivalent of the final officially approved version of my thesis. No emendation of content has occurred and if there are any minor variations in formatting, they are the result of the conversion to digital format.’

Signed ……………………………………………......

Date ……………………………………………......

ORIGINALITY STATEMENT

I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged

Signed ……………………………………………......

Date ……………………………………………......

ACKNOWLEDGEMENT

Writing an “Acknowledgements” for a thesis is like writing down the last period at the end of a story. I thought about this time for a few times, however, when it really comes, I feel a little bit hesitating. This three years journey is a difficult but fantastic adventure for me. I can’t reach the destination without the kind help of a number of individuals. I owe them my deepest gratitude and best wishes. I would like to acknowledge their contributions here with my sincere thankfulness.

Firstly, I would like to give my deepest gratitude to my supervisor Prof. Paul Munroe for his guidance and support in these three years. I am so fortunate to have Paul as my supervisor. I have to say I came to this university with only curiosity and passion, Paul taught me how to be an independent researcher. He dedicates a lot of energy in guiding me, and also taught me how to write academically with great patience. He also encouraged me to start a project on my own interest. I really learned a lot from His profound knowledge and critical attitude. I appreciate his untiring effort in reviewing my paper and thesis. I hope I could have the opportunity to work with him in the future.

I would like to thank Prof. Jiang Xu for providing transition metal nitride specimens and for reviewing the manuscripts of my papers. We worked together in several projects. From him, I learned many strategies and experiences in paper writing and project designing. I also need to thank Dr. Zonghan Xie for his valuable comments during the paper writing stage. He did a lot of work in reviewing the manuscripts of my papers.

I am so grateful for the kind help come from the technical and administrative staff at UNSW. Dr. Bill Joe spent a lot of time with me in the Nanomechanics Testing Lab in these three years. His patience and endeavor deeply impressed me. I want to give my heartfelt respect to Dr. George Yang for his efficiency and professionalism. I would like to give my thanks to the staffs in EMU, especially Dr. Charlie Kong, Mr. Sean Lim, and Ms. Katie Levick, they were always there when I needed help. I need to thank Dr. Bill Gong, Dr. Yu Wang and Dr. Anne Rich for their kind assistance. They help me a lot in my sample

ii characterization. I also want to give my appreciation to Ms. Joanne Hallis, Dr. Anthony Zhang, Mr. Danny Kim, Ms. Laura McNally, Mr. Alan Chow, Ms. Lana Strizhevsky and Ms. Lucy Zhang from the School of Materials Science and Engineering. Only with their support, I can enjoy my three-year-life in UNSW.

I also want to say thank you to all my friends. I know they will never ask for my thanks. The courage and concern they gave to me supported me in those good and bad times. To those of my brothers and sisters in “Playing Mafia Day and Night” group, I don’t know how to express my gratitude to you guys. I am so lucky to have you. I will always remember the happiness we created together.

Youran, three years long-distance relationship, we made it. Thank you for your understanding and company.

At last, I will give my heartfelt thanks to my parents. They brought me into this world, and give me all their love. They encouraged me to start this adventure three years ago, and supported me from beginning to end. “Thank you” is not nearly enough to express my gratitude to them.

iii

ABSTRACT

By exploring the influence of nanopores on the mechanical properties of nanocrystalline

Ta2N coatings, it was found that the presence of nanoporosity may increase wear resistance and damage resistance of these coatings by significantly reducing elastic modulus without iv greatly decreasing hardness. This suggests a promising new strategy for improving the mechanical properties of nanocrystalline coatings.

TABLE OF CONTENTS

ORIGINALITY STATEMENT ...... i ACKNOWLEDGEMENT ...... ii ABSTRACT ...... iv TABLE OF CONTENTS ...... vi LIST OF PUBLICATIONS ...... xi Chapter 1 Introduction ...... 1 1.0 Introduction ...... 2 Chapter 2 Literature review ...... 5 2.0 Literature review ...... 6 2.1 Nanocrystalline transition metal nitride coatings ...... 6 2.1.1 Nanocrystalline hard coating ...... 6 2.1.2 Material selection for hard coatings...... 7 2.1.3 The advantages of transition metal nitrides ...... 9 2.1.4 Common transition metal nitride coatings ...... 10 2.1.4.1 Titanium nitride coatings ...... 10 2.1.4.2 Chromium nitride coatings ...... 11 2.1.4.3 Zirconium nitride coatings ...... 12 2.1.4.4 Tantalum nitride Coatings ...... 13 2.2 Mechanical and tribological properties of hard coatings ...... 15 2.2.1 Hardness and elastic modulus ...... 15 2.2.1.1 Hall-Petch effect and inverse Hall-Petch effect ...... 17 2.2.1.2 The “strongest size” ...... 19 2.2.2 Toughness and wear resistance ...... 20 2.2.2.1 The influence of H and E to the mechanical properties of hard coatings ..... 20 2.3 Deposition methods ...... 21 2.3.1 Chemical vapor deposition (CVD) ...... 22 2.3.1.1 Plasma-enhanced chemical vapor deposition...... 23

2.3.2 Physical vapor deposition (PVD) ...... 24 2.3.2.1 Sputtering deposition ...... 27 2.3.2.2 Ion plating ...... 28 2.3.3 Double cathode glow discharge plasma technique ...... 30 2.3.4 The influence of deposition conditions ...... 31 2.3.4.1 Substrate temperature ...... 31 2.3.4.2 Gas pressure ...... 32 2.3.4.3 Bias and ion bombardment ...... 32 2.3.4.4 Impurities ...... 33 2.4 Characterization techniques...... 34 2.4.1 Scanning electron microscopy (SEM) and energy dispersive spectrometery (EDS) ...... 34 2.4.2 Transmission electron microscope (TEM) and selected area electron diffraction (SAED) ...... 36 2.4.3 Focus ion beam microscopy ...... 38 2.4.3.1 Preparation of TEM samples with a FIB system ...... 40 2.4.4 X-ray photoelectron spectroscopy (XPS) ...... 41 2.4.5 X-ray diffraction (XRD) ...... 42 2.4.6 Raman spectroscopy ...... 44 2.5 Summary ...... 45 Chapter 3 Nanoindentation testing ...... 46 3.0 Nanoindentation testing ...... 47 3.1 Introduction of nanoindentation testing ...... 47 3.2 Data measurement analysis ...... 48 3.2.1 Indenters...... 48 3.2.2 Hardness...... 50 3.2.3 Elastic modulus ...... 52 3.2.4 Nanoindentation errors ...... 53 3.3 Deformation and fracture behavior ...... 53 3.4 Partial unloading testing ...... 55 3.5 Summary ...... 55

vii

Chapter 4 The microstructure and mechanical properties of tantalum nitride coatings ...... 56 4.1. Introduction ...... 57 4.2 Experimental methods ...... 57 4.3. Results ...... 60 4.3.1 Microstructure and phase analysis ...... 60 4.3.2 Nanoindentation testing ...... 63 4.3.3 Contact damage testing ...... 65 4.4 Discussion ...... 66 4.5 Conclusion ...... 69 Chapter 5 The influence of oxygen on the microstructure and mechanical properties of tantalum nitride coatings deposited by the double cathode glow discharge plasma technique ...... 71 5.1 Introduction ...... 72 5.2 Experiment methods ...... 72 5.3 Results ...... 74 5.3.1 XRD ...... 74 5.3.2 XPS ...... 75 5.3.3 SEM ...... 78 5.3.4 TEM ...... 80 5.3.4.1 Coating #1 ...... 80 5.3.4.2 Coating #2 ...... 81 5.3.5 Nanoindentation ...... 82 5.3.6 Contact damage testing ...... 84 5.4 Discussion ...... 85 5.5 Conclusion ...... 87

Chapter 6 The formation of bone-like hydroxyapatite on a Ta2N coating ...... 88 6.1 Introduction ...... 89 6.2 Experimental method ...... 90 6.3 Results ...... 91

6.3.1 XRD analysis of the as-deposited Ta2N coating ...... 91 viii

6.3.2 Raman spectroscopy ...... 92 6.3.3 SEM ...... 93 6.3.4 TEM ...... 95 6.4 Discussion ...... 100 6.5 Conclusion ...... 102 Chapter 7 The microstructure and mechanical properties of zirconium nitride coatings deposited by a double cathode glow discharge plasma technique ...... 103 7.1 Introduction ...... 104 7.2 Experimental methods ...... 105 7.3 Results ...... 107 7.3.1 XRD ...... 107 7.3.2 SEM ...... 109 7.3.2.1 Coating Q1 ...... 109 7.3.2.2 Coating Q2 ...... 110 7.3.2.3 Coating Q3 ...... 111 7.3.2.4 Coating Q4 ...... 112 7.3.3 TEM ...... 113 7.3.3.1 Coating Q1 ...... 113 7.3.3.2 Coating Q2 ...... 115 7.3.3.3 Coating Q3 ...... 117 7.3.3.4 Coating Q4 ...... 119 7.3.4 Nanoindentation testing ...... 121 7.3.5 Contact damage test ...... 124 7.4 Discussion ...... 126 7.5 Conclusion ...... 130 Chapter 8 The porosity-dependence of mechanical properties and damage mechanisms in nanoporous tantalum nitride coating ...... 131 8.1 Introduction ...... 132 8.2 Experimental method ...... 133 8.3 Results ...... 136 8.4 Discussion ...... 145

8.5 Conclusion ...... 147 Chapter 9 Conclusion ...... 148 9.0 Conclusion ...... 149 Reference...... 152

LIST OF PUBLICATIONS

1. Song Xu, Jiang Xu, Paul Munroe, Zong-Han Xie, “A critical role of nanoporosity in improved mechanical properties of tantalum nitride nanoceramic coating”, will be submitted to Scripta Materialia.

2. Song Xu, Paul Munroe, Jiang Xu, Zong-Han Xie, “The microstructure and mechanical properties of tantalum nitride coatings deposited by a double cathode glow discharge plasma technique”, submitted to Surface and Coatings Technology.

3. Jiang Xu, ZhengYang Li, Song Xu, Paul Munroe, Zong-Han Xie, “A nanocrystalline zirconium carbide coating as a functional corrosion-resistant barrier for polymer electrolyte membrane fuel cell application”, Journal of Power Sources, 2015 297 359-369.

4. Jiang Xu, Wei Hu, Song Xu, Paul Munroe, Zong-Han Xie, “Electrochemical Properties of a Novel β-Ta2O5 Nanoceramic Coating Exposed to Simulated Body Solutions”, ACS Biomaterials Science & Engineering, 2016 2 73-89.

5. Jiang Xu, Hao Jie Huang, ZhengYang Li, Song Xu, Hongliang Tao, Paul Munroe, Zong- Han Xie, “Corrosion behavior of a ZrCN coated Ti alloy with potential application as a bipolar plate for proton exchange membrane fuel cell”, Journal of Alloys and Compounds 2016 633 718-730.

6. Jiang Xu, Song Xu, Paul Munroe, Zong-Han Xie, “A ZrN nanocrystalline coating for polymer electrolyte membrane fuel cell metallic bipolar plates prepared by reactive sputter deposition”, RSC Advances 2015 5 67348–67356.

7. Xiaolei Liu, Yidan Huang, Jae Yun, Xiaoming Wen, Zhong Lu, Tian Zhang, Hongtao Cui Wei Li, Chang-Yeh Lee, Song Xu, Xiaojing Hao, Gavin Conibeer, “Characterization of a Cu2ZnSnS4 solar cell fabricated by sulfurization of metallic precursor Mo/Zn/Cu/Sn”, Physica Status Solidi (A) 2015 1-6. xi

Chapter 1

Introduction

1.0 Introduction

Nanocrystalline materials are generally defined as single or multi-phase polycrystalline materials with a nanoscale grain size, typically in a range from 1 nm to 250 nm [1]. As the grain size decreases, nanocrystalline materials exhibit improvements in their mechanical and physical properties, such as enhanced mechanical strength, increased tribological properties, as well as higher thermal and electrical resistivity etc., compared to conventional coarse-grained materials [2]. Therefore, in order to achieve better surface properties and satisfy property requirements under harsh working conditions, further studies in coating science have been performed on thin film materials exhibiting nanoscale structures. Many factors must be taken into consideration in designing nanocrystalline coatings, such as the volume fraction of grain boundary, grain size, coating thickness, surface and interfacial energy, texture, epitaxial strains etc. In turn, those factors are mainly determined by choices of the materials deposited, the deposition methods used and process parameters employed [3].

There are many different types of nanocrystalline coatings, including nanocomposite coatings, nano-scale multilayer coatings, nano-graded coatings etc. Among these coatings, the properties of transition metal nitride nanocrystalline coatings, such as titanium nitride, zirconium and chromium nitride, have been widely studied [4]. Furthermore, zirconium nitride and tantalum nitride nanocrystalline coatings have recently attract a lot of attention as corrosion resistant and biocompatible coatings [5, 6].

Many deposition methods have been developed to produce nanocrystalline coatings. The most frequently used deposition technologies include plasma enhanced chemical vapor deposition, magnetron sputtering deposition, ion-assisted deposition, and cathodic arc evaporation. In order to generate coatings with further improved properties at low cost, new, innovative deposition techniques are under development. The double cathode glow discharge plasma technique is a relatively newly developed deposition method. It was first used by Xu et al. to produce nanocrystalline coatings [7]. Compared to other

2 methods, this approach uses relatively simple apparatus and can be applied to a wide range of materials that make it an attractive deposition approach for many potential uses.

Techniques such as X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), Raman spectroscopy (RS), scanning electron microscope (SEM), scanning probe microscopy (SPM), focused ion beam (FIB) microscopy and transmission electron microscopy (TEM) are often employed to analyze the composition and determine the structure of coatings. These can be used collectively to generate a detailed understanding of the structure, chemistry and crystallography of advanced coatings.

Furthermore, to estimate the mechanical properties of as-deposited coatings nanoindentation is commonly used. By introducing a predesigned load to the coating surface, the hardness and elastic modulus can be readily calculated. Meanwhile, by combining nanoindentation with several other techniques such as FIB and TEM, the deformation mechanisms of coatings under external load can be investigated.

In this thesis, the composition, microstructure, mechanical properties and biocompatibility of transition metal nitride nanocrystalline coatings, including tantalum nitride, tantalum oxy-nitride and zirconium nitride coatings, deposited by the double cathode glow discharge deposition technique with diverse deposition parameters were studied. Further, the influence of nanoporosity on the mechanical behaviour of nanocrystalline Ta2N coating is also discussed. The aim of this thesis is to explore the effects of deposition parameters on the composition, microstructure, and mechanical properties of a range of as-deposited transition metal nitride coatings, and achieve a better understanding of the structure-property relationships in nanocrystalline materials.

The thesis is structured as follows: Chapter 2 presents the research background and key theoretical concepts relevant to this study. This chapter can be divided into 4 parts. The first part reviews the literature on nanocrystalline coatings, especially transition metal coatings. The second part discusses the mechanical properties of nanocrystalline coatings. The third part presents the most frequently used deposition techniques, including both

3 chemical vapor deposition (CVD) and physical vapor deposition (PVD) methods. The fourth part details the characterization techniques used in this thesis. In Chapter 3, the procedures for nanoindentation testing are detailed. Since each group of coatings was prepared under specific deposition parameters, the other experimental procedures are detailed, as required, in each chapter.

The description of the results and corresponding discussion commence in Chapter 4. In Chapter 4 the characterization of the microstructure, composition and mechanical properties of tantalum nitride coatings deposited by the double cathode glow discharge deposition technique are described and discussed. Chapter 5 reveals the influences of oxygen pressure during deposition on the microstructure and mechanical properties of tantalum nitride as-deposited coatings. The following chapter, Chapter 6, discusses the bioactivity of tantalum nitride coatings by testing the compatibility between an as- deposited tantalum nitride coating and hydroxyapatite. In Chapter 7, zirconium nitride coatings were produced on both stainless steel and Ti-Al-V substrates, and their microstructure, composition and mechanical properties are discussed in this chapter. Chapter 8 discusses the influence of nanoporosity on the mechanical properties of a nanocrystalline coating. It was found that for nanocrystalline coatings, nano-sized pores can increase their damage resistance and wear resistance. Finally, the results are summarized in Chapter 9, together with suggestions for further work.

Chapter 2

Literature review

2.0 Literature review

This chapter is divided into four parts. The first part provides an overview of the development, properties and application of nanocrystalline coatings, especially those based on transition metal nitrides. In the following sections, key concepts about the mechanical properties of nanocrystalline coatings are described. The third part introduces the more frequently used vapor deposition techniques and the double cathode glow discharge plasma technique. Finally, the characterization techniques used in this thesis are described.

2.1 Nanocrystalline transition metal nitride coatings 2.1.1 Nanocrystalline hard coating

Nanocrystalline materials have been widely studied in the last 30 years [1]. The definition of nanocrystalline materials is usually given by the limitation of microstructural length or the grain size of a material. Normally, nanocrystalline materials are composed of structural elements which exist within the nanoscale in at least one direction [8, 9]. When the grain size reduces to the nanometer (typically 1-250nm) range, a large fraction of atoms are located in the interphase region, such as grain boundaries and higher-order junctions. For a nanocrystalline material with an average grain size of 100 nm, the volume fraction of atoms located in these interphase regions is about 3%, but this value may increase to 30% and 50%, when the grain size decreases to 10 nm and 5 nm, respectively [1]. As a consequence, the properties of nanocrystalline materials are not only directly dependent on the polycrystals themselves, but are also greatly influenced by the interphase structure of the grains in the coatings. In comparison with conventional materials, a decrease in the size of the structural element (i.e. grain size) and increased importance of the interphase can bring attractive properties to nanocrystalline materials, such as increased mechanical strength, higher hardness, improved wear resistance, enhanced thermal stability and electrical resistivity etc. [2].

In order to improve the performance of bulk materials and to generate desirable properties, yet retain the bulk properties of the original substrate material, nanocrystalline materials are often applied to the surface of bulk substrates to form a functional coating [8, 10]. Protective hard coatings are one of the most important applications for nanocrystalline materials. Combining high hardness and good wear resistance with other properties, such as good corrosion resistance and low coefficient of friction between tools and machined surface, both machining speed and the lifetime of tools can be greatly improved by the deposition of a nanocrystalline hard coating. This may then contribute to increased industrial productivity and lower overall production cost by increasing the machining speed, yet reducing the costs of maintaining and replacing tools [11].

Today, hard coatings have become one of the most important and versatile methods of improving the performance of bulk materials [12]. The increasing demand for higher productivity and performance of tools in manufacturing industries requires continuous development in both the preparation methods used and the improvement of the mechanical properties of hard coatings.

2.1.2 Material selection for hard coatings

The mechanical properties of coating materials are determined by both the intrinsic properties of the grains and the structure of the coating [13, 14]. Therefore, it is important for the coating materials themselves to have excellent mechanical properties. However, good mechanical properties are not the only requirement for a coating material, other properties, such as the chemical activity and interaction between the coatings material to the substrate should also be considered during coating design to achieve better protection for tools against harsh working conditions. Therefore the coating material should be carefully chosen. Based on different regions, the requirements for a nanocrystalline coating can be summarized as below (Fig. 2.1):

1) At the coating/substrate interface, the degree of adherence between the coating and the substrate should be considered. The level of adhesion of the coating to the substrate

should be sufficient to withstand the stresses and heat generated during machining processes. Meanwhile the interaction between coating and substrate should be limited to reduce the extent of any influence deriving from the substrate. In order to avoid delamination, the extent of any thermal expansion misfit between the coating and substrate should also be taken into account to avoid the generation of large thermal stress generated during the cooling of coatings from the deposition temperature. Therefore the thermal expansion coefficient of coating material should be similar to that of substrate. Sometimes by inducing a graded interface the disparity in thermal expansion and thermal conductivity between substrate and coating material can be reduced, hence excellent adherence could be achieve during deposition process [15].

2) To provide good protection to the substrate, the coating material should provide high hardness, good fracture toughness and reasonable wear resistance. In most circumstances, hardness is the primary property and the hardness of coating material should be higher than the surface it will be used to machine.

3) At the coating/environment interface, the influence of the working environment on the coating material should be considered. Normally good chemical stability is required for a coating material to resist oxidation and corrosion.

Fig. 2.1 Important criteria for the material selection of coatings. [12]

In addition to the properties mentioned above, sometimes, other properties, such as optical behavior, electrical resistivity, should also be considered during the design of a coating [4, 12].

2.1.3 The advantages of transition metal nitrides

In order to meet increasingly stringent technological demands, a variety of coating materials have been designed and tested. Generally, based on the chemical bonding state, the most frequently used coating materials can be classified into three different groups: 1) metallic hard materials, e.g. TiN, ZrC, TiB2; 2) covalent hard materials, e.g. B4C, BN; and 3) ionic hard materials, e.g TiO2. Although each type of material exhibits relative advantages and disadvantages, it appears that metallic materials, including transition metal nitrides, carbides and borides, are very suitable for hard coating applications in a wide range of situations. [12]

Transition metal nitrides have attracted much attention as hard coating materials in many fields due to their attractive properties. The benefits from simultaneous contributions of metallic, covalent and ionic bonding between both metal-metal pairs of atoms and metal- nonmetal pairs of atoms, provide transition metal nitrides with attractive combinations of such as high toughness, attractive aesthetic appearance and excellent conductivity (with typical resistivities range from 10 to 30 μΩcm) [16]. Meanwhile, transition metal nitrides exhibit similar thermal expansion coefficients with many frequently-used substrate materials such as stainless steel and titanium alloys. This provides good adherence between the coating and the substrate, reducing the likelihood of delamination, cracking, or other adverse effects caused by residual thermal stresses due to thermal expansion mismatches [17].

However, transition metal nitrides also have a number of limitations, for example those arising from the low shear modulus of their non-directional metallic bonds. As such, it is difficult for transition metal nitrides to achieve ultra-high hardness values (hardness ≥ 70 GPa) [14]. Recent attempts to design ultra-hard coatings based on transition metal nitrides have focused on building nanocomposite ternary systems with nanocrystalline

9 metal nitride grains embedded in an amorphous matrix by adding elements such as Si into transition metal nitrides that undergo spinodal phase segregation. The strong covalent bonds at the interfacial region between titanium nitride grains and Si3N4 amorphous matrix can lead to remarkable increases in the hardness of nanocrystalline coating [18, 19]. In several studies, extremely high hardness values of up to 105 GPa were achieved by TiN/Si3N4 coatings fabricated using this approach [20].

2.1.4 Common transition metal nitride coatings 2.1.4.1 Titanium nitride coatings

Amongst all the transition metal nitride coatings, titanium nitride is the most widely studied and applied. TiN has a NaCl like fcc structure. It also exhibits a golden yellow color, which makes it attractive for applications in areas such as bathroom fittings [21]. Since the 1960s, TiN has been deposited as a protective coating via both CVD and PVD techniques. TiN coatings exhibit many excellent properties, high hardness, good wear resistance, a low friction coefficient, low maintenance and cost etc. The main disadvantage of TiN, however, is its low oxidation resistance at high temperature [22]. Research shows TiN coatings can be easily oxidized and softened when the temperature is greater than ~500 ℃. Moreover, the penetration of oxygen may lead to a porous surface structure for TiN coatings [23]. The generation of oxides into the nitride can greatly decrease the coating’s wear resistance and increases the coefficient of friction that leads to low productivity and structural failure of tools. To improve the high temperature performance of TiN coatings, aluminum is often added into TiN to form a TiAlN ternary system [24]. Due to the introduction of aluminium, TiAlN only starts to severally react with oxygen at 800℃ [25]. Moreover, TiAlN coatings also exhibit higher hardness and better wear resistance then TiN [26].

Normally, the thickness of TiN coatings deposited on tools is between 2 μm to 10 μm [21]. Based on the first principle calculations, the theoretical hardness for single crystal TiN is about 30 GPa [27]. However, the properties of TiN coatings can be greatly influenced by their microstructure and stoichiometry. The hardness of TiN coatings

10 decreases with increasing concentration of vacancies and impurities. Furthermore, TiN coatings often exhibit a columnar grain structure with grains elongated parallel to the deposition direction. The hardness of coatings can exhibit a great variation between the coating surface and their cross-section because of the anisotropy of the grain texture [21]. Meanwhile, both the hardness and the wear resistance of TiN coatings increase with decreasing atomic ratio of N/Ti [4, 21]. Normally, the reported hardness values for dense stoichiometric TiN coatings are in a range from 20 to 23 GPa [28].

Other than use as a protective coating, TiN coatings have also been applied as diffusion barriers in microelectronic devices and solar cells, selective transparent films and used for applications in high temperature photothermal conversion [28].

2.1.4.2 Chromium nitride coatings

CrN also attracts much of attention as a protective coating. Normally CrN coatings are deposited by PVD [29]. The deposition process for CrN coatings can be performed at temperatures as low as 250 ℃ [29]. The higher sputtering yield of CrN during deposition makes it suitable for mass production applications [30]. However, because of the low reactivity of Cr towards nitrogen, the Cr-N system presents two common phases CrN

(which exhibits a fcc structure) and Cr2N (which exhibits a hexagonal structure)[4]. CrN typically exhibits a high hardness of about 25 GPa [30-32]. Since the hexagonal Cr2N is harder than fcc CrN, the CrN based coatings shows increased hardness with increasing content of Cr2N [33].

These coatings also exhibit several advantages compared to TiN coatings. Firstly, CrN has better corrosion resistance than TiN, especially thermal oxidation [34, 35]. The CrN can exhibit a stable phase composition and good adhesion to substrate even to a high temperature of 700 ℃. Secondly, the tribological properties of CrN are also better than TiN. For example, the friction coefficient of CrN is much lower than that of TiN [36, 37]. This results in a decrease in the likelihood in cracking under abrasive wear [38]. The wear rate of CrN is also about 40% slower than TiN [37]. This means that CrN can

11 greatly increase the productivity and lifetime of tools coated compared to TiN. Furthermore, CrN also shows better fracture toughness than TiN, that means more energy is needed for the propagation of cracks in CrN coatings [37]. Finally, CrN coatings typically exhibit a very fine grain size on deposition so a larger thickness can be achieved for CrN coatings compared with conventional TiN coatings [29]. The maximum thickness for CrN coatings can be as much as 50 μm [37]. That means CrN presents as a flexible choice of coating thickness for different working conditions. This, in turn, can prolong component lifetime.

Due to its excellent properties, CrN coatings are thought to be good substitution of TiN coatings in many applications, especially for those that require large coating thickness and high thermal oxidation resistance. Since the 1990s, CrN coatings have been applied on cutting, die casting and cold forming tools to increase their durability, and also deposited on machine parts to provide wear and corrosion protection. [29]

2.1.4.3 Zirconium nitride coatings

Similar to TiN, ZrN is also a group -IV transition metal nitride, therefore ZrN coatings share a similar structure and properties to TiN coatings [4]. Along with TiN coatings, ZrN coatings are also commercially used as protective coatings for cutting tools and machine tools [39]. Normally, ZrN coatings are deposited by PVD processes, such as reactive sputtering, laser deposition, ion-beam assisted deposition [39, 40]. The low electrical resistivity of 13.6 μΩ cm in bulk and the excellent thermal stability (ΔH=-87.3 kcal/mol) make ZrN a promising candidate for diffusion barriers in intergrated circuit technology [6]. ZrN coatings also have good biocompatibility and can be used in biomedical applications [41].

Recent research shows that ZrN coatings exhibit good resistance to the corrosion of proton exchange membrane fuel cells, and satisfy corrosion current density requirements at both anode and cathode potentials. Preliminary studies have demonstrated that ZrN could be used as a coating material to extend the service life and performance of metallic

12 bipolar plates. In an earlier study, a ZrN nanocrystalline coating was deposited onto a 316L SS substrate using a double glow discharge plasma technique. The electrochemical characteristics and surface wettability of the as-deposited coating were investigated and compared with uncoated 316 SS to evaluate ZrN's potential as a protective coating for fuel cell bipolar plates. According to the data acquired, the corrosion resistance of the ZrN coated 316L SS was significantly improved in comparison with that of uncoated 316L SS. Moreover, the ZrN-coated 316L SS is more hydrophobic than the uncoated steel, which prevents accumulated water from flooding the electrode system and lowers the extent of any corrosive attack.

2.1.4.4 Tantalum nitride coatings

In contrast to titanium, chromium and zirconium-based nitrides, tantalum nitride has attracted less attention as a coating material. This is presumably because of the high cost and more stringent deposition conditions required for this material [42]. Even so, tantalum nitride has recently attracted interest as a potential diffusion barrier coating material [43]. Further, some studies have demonstrated that tantalum nitride-based coatings have better histocompatibility and blood compatibility than some widely used biomedical titanium alloys, making them promising candidates for medical applications [5]. Such coatings have been prepared by a range of deposition methods including DC/RF magnetron sputtering, ion-beam-assisted deposition, reactive-electron-beam evaporation etc [44]. Their mechanical properties have been investigated in a number of studies [45- 49]. The results showed that mechanical properties of these coatings were dependent strongly upon the deposition method and deposition parameters employed, the hardness values for tantalum nitride coatings normally range from 20 GPa to 45 GPa, with an elastic modulus between 350 GPa and 450 GPa.

Commonly, tantalum nitride coatings are deposited using magnetron sputtering deposition methods [45, 48, 50]. In this case, with variations in nitrogen partial pressure, the resultant coatings can frequently exhibit complex microstructures containing a number of equilibrium and metastable phases. Phases such as bcc-Ta, hexagonal Ta2N,

13 hexagonal ε-TaN, and fcc TaN are often observed, in addition to a variety of other phases, including tetragonal-structured Ta, hexagonal δ-TaN, having a WC-type structure, hexagonal-Ta5N6, tetragonal-Ta4N5 and orthorhombic-Ta3N5 [42, 47]. Since the mechanical properties of tantalum nitride-based coatings are related to its phase composition, it is important to understand the coating microstructure development under different deposition method and conditions. The common phases of tantalum nitride are shown in Table 2.1.

Table 2.1 The Crystal structures for Ta-N phases. TaN phase Structure Lattice parameter Space group

TaN0.04 Cubic 푎 = 10.09Å -

TaN0.1 Cubic 푎 = 3.37Å Im3m

Ta2N Hexagonal 푎 = 3.04Å , 푐 = 4.91Å P63/mmc

TaN0.8 Hexagonal 푎 = 2. 93Å, c = 2. 86Å P-6m2

TaN Cubic 푎 = 4.33Å Fm3m TaN Hexagonal 푎 = 5.19Å, 푐 = 2.90Å P6/mmc

Ta3N5 Orthorhombic a = 3.89Å, b = 10.26Å, 푐 = 10.26Å Cmcm

Ta4N5 Tetragonal 푎 = 6.83Å, 푐 = 4.27Å I4/m

Ta5N6 Hexagonal 푎 = 5.17Å, 푐 = 10.30Å P63/mmc

Ta6N2.57 Hexagonal 푎 = 5.29Å, 푐 = 4.92Å P-31m

Ta4N Orthorhombic a = 5.16Å, b = 3.11Å, c = 9.94Å -

Ta6N2.5 Orthorhombic - -

In studies of Valleti et al. [51] TaN-based coatings were prepared by a reactive DC magnetron sputtering technique with varying nitrogen to argon ratios at 300℃. The TaN coatings are composed of mainly cubic-TaN0.1, orthorhombic-Ta4N, orthorhombic-

Ta6N2.5, hexagonal-TaN0.8, and cubic-TaN phases in different volume fractions. Based on the rule of mixtures:

퐻푚𝑖푥 = (푉훼 × 퐻훼) + (푉훽 × 퐻훽) (2.1)

where Hmix, Hα and Hβ are the hardness values of mixed and individual phases and. Vα,

Vβ are the volumes fractions of the two phases in the film, the approximate hardness of each phase was calculated. A superhigh hardness of 61.87 GPa was achieved for an orthorhombic-Ta4N phase. Meanwhile, hardness values close to 49.9 GPa, 15.5 GPa and

30.8 GPa have been achieved in single phase cubic-TaN, hexagonal-TaN0.8, and orthorhombic-Ta6N2.5, respectively. Notably, the hardness of nanocystalline coating can also be influenced by both the grain size and interface structure. In order to acquire more accurate values of hardness, multi-component models, in which the grain size and grain boundaries etc. were also considered as components during calculation, should be applied [52, 53]. Here, based on the studies of Valleti, the emphasis was only placed on the hardness of TaN-based phases.

2.2 Mechanical and tribological properties of hard coatings 2.2.1 Hardness and elastic modulus

The hardness of nanocrystalline coatings, which are commonly applied on cutting tools for drilling, milling, turning operations, plays a very significant role in preventing tools from chipping and, hence, premature failure [54].

Hardness, H, is a complex property, which can be correlated to various factors, such as shear strength, cohesive energy, grain size etc. From a mechanical perspective, the hardness of a material is usually defined as its resistance to plastic deformation under isostatic pressure. Microscopically, hardness is a measure of how readily dislocations are generated and are able to move through the material in response to shear stresses produced by an external load [13]. As to the hardness, materials are usually divided into three groups: (1) hard materials, having a hardness ˂ 40 GPa; (2) superhard materials, having a hardness ≅ 40 GPa~70 GPa; and (3) ultrahard materials, having a hardness ≥ 70 GPa [14, 55].

The hardness of a material is influenced by both the strength of interatomic forces and the dominant deformation mechanism. That means both intrinsic hardness and extrinsic properties play important roles in determining the hardness of a nanocrytalline coating. Having a high cohesive energy and short bond length, covalent bonds are much stronger than ionic and metallic bonds. Normally, the intrinsic hardness of a material reduces with increasing the concentration of ionic or metallic bonds present. The mechanisms of dislocation nucleation and propagation of a material is controlled by the grain size, microstructure, phase structure, crystallographic texture. In many cases, coatings that exhibit a columnar structure or high defect content show a lower hardness. Moreover, changes in preferred orientation can also result in a variation of hardness for a coating system due to a variation in measured hardness for different orientations. This is especially common for those materials with fewer slip systems. The relationship between hardness and grain size will be discussed later in this section. [4]

The elastic response of a material to hydrostatic pressure is described by the elastic modulus, E, which reflects the resistance to deformation under reversible deformation [14]. This determines the stress needed to form and move dislocations, as well as to form and propagate cracks in the material [56]. Generally, the strength of a material usually increases with increasing elastic modulus. It is also generally accepted that the hardness of a bulk materials scales with the value of shear modulus, G, because the plastic deformation of crystalline materials is due to dislocation activity, while the energy of a dislocation is proportional to shear modulus [57]. Since, shear modulus is also proportional to elastic modulus by the relation:

퐸 퐺 = (2.2) 2(1+휈) where ν is the Posisson’s ratio, elastic modulus can be used as an indicator for hardness. For the above reasons, generally, hard material design elastic modulus has been commonly used as a guide for theoretical predictions of the properties for new hard materials [19, 58]. Therefore, much attention has been paid to those materials with intrinsically high elastic moduli [14].

However, recent research has demonstrated that the relationship between hardness and elastic modulus or shear modulus is not always as simple. For example, calculations show that OsB2 has a bulk modulus of over 300 GPa, relatively close to that of diamond (bulk modulus ≈ 440 GPa; hardness ≈100 GPa) and cubic BN (bulk modulus ≈370 GPa; hardness ≈60 GPa) [59]. However, the hardness of OsB2 is less than 20 GPa [14]. This is due to the low shear strength of 9.1 GPa of the non-directional metallic Os-Os bonds in the (001)[010] slip system [59]. Furthermore, C3N4 has high bulk, as well as shear, modulus, but relatively low hardness of 26-28 GPa [14], because the electronic stable 3 pseudocubic C3N4 phase, with sp lone-pair states, will transform into an unstable graphitic like phase under small shear strains, leading to a much lower energy barrier, a strong tendency toward shear breaking under strain, much shorter bond deformation range and, consequently, a surprisingly low shear strength despite its high equilibrium shear modulus [60]. Therefore, materials with high values of elastic modulus may exhibit relatively low hardness due to low shear strength in some slip systems and the structural transformation to a softer phase in the shear process [14]. Thus, the high values of elastic modulus and shear modulus guarantee neither structural stability nor electronic stability. Conversely, it is also possible, even for the same system, to create a coating with high hardness, but relatively low elastic modulus, which may contribute to better mechanical properties for hard coatings [61]. The influence of high hardness and relatively low elastic modulus is discussed in the following sections.

2.2.1.1 Hall-Petch effect and inverse Hall-Petch effect

For conventional materials, the hardness of material, 퐻, the hardness of a single grain, 퐻표, and the grain size, 푑, satisfy the Hall-Petch effect:

1 − 퐻 = 퐻표 + 푘푑 2 (2.3) where 푘 is an experimental constant. That means by decreasing the grain size of a material, higher strength can be achieved [62]. The Hall-Petch effect is also applicable to

17 nanocrystalline materials. With decreasing grain size from the scale of microns down to few tens of nanometers, the dislocations generated during plastic deformation are expected to pile up at the grain boundaries, and become less mobile, thus the strength and hardness of a material increases [63]. One example of the hardness enhancement for a nanocrystalline, ZrN coating, due to the Hall-Petch effect can be observed from region 1 of Fig. 2.2 [62].

Fig. 2.2 Hall-Petch plot of hardness of nanocrystalline ZrN coatings against the inverse square-root of grain size. [62]

However, with further decreases in the crystallite size, an inverse Hall-Petch effect dominates. It can be seen from region 2 that the hardness decreases markedly with a negative k (the slope of the Hall-Petch plot) (Fig. 2.2). This is because accompanying the substantial decrease in grain size, atoms located in grain boundaries now occupy a significant fraction of the material volume. The pile-up of dislocations at grain boundaries ceases, since the stress concentration caused by the pile-up is sufficient to eliminate dislocations by moving them into the grain boundaries, and thus fewer dislocations can be observed in the grain. This results in the breakdown of the Hall-Petch

18 effect [14, 62]. With further decreases in grain size, the dislocation mechanism becomes invalid and plastic deformation is dominated by grain boundary shear instead of dislocation motion [64]. The deformation process becomes intergranular rather than intragranular. This leads to a decrease of hardness and strength of the nanocomposite, since grain boundary is a softer phase and the considerable decrease in grain sizes creates continuous shear zones across the nanocomposite [65, 66].

2.2.1.2 The “strongest size”

The Hall-Petch effect and the inverse Hall-Petch effect suggest a crossover in deformation mechanisms, and one would therefore expect to find a grain size of maximum hardness, the so-called ‘strongest size’, between the dislocation-based deformation region (regions 1 Fig. 2.2) and the grain-boundary-mediated deformation region (region 2 Fig. 2.2) [63].

In 1998, Yip et al. reported, by using computer simulations, the ‘strongest size’ for copper and palladium are grain sizes of 19.3 nm and 11.2 nm, respectively. Futher, similar simulations show that for most nanocrystalline materials the ‘strongest size’ranges from 10 to 20 nm [14]. Qi et al. used a modified Eshelby's equation to calculate the strongest size, D:

2nGb 퐷 = (2.4) Kπτ

퐸 where G is shear modulus, for isotropic materials G = , E is elastic modulus, b is 2(1+휐)

2 magnitude of the Burgers vector (b=√ 푎, a is lattice parameter), K is a constant factor 2 equal to (1-ν) for an edge dislocation, ν is Poisson's ratio, n is the number of dislocations in a pile-up, τ is critical shear stress. Invoking the Tresca shear stress criterion, plastic deformation occurs at τ=0.5Y, where Y is a yield stress equal to one third of the hardness H under plastic constraint [62]. From the standpoint of material design, the ‘strongest size’

19 effect offers a possible strategy for hard coatings to reach maximum hardness by depositing coatings with a controlled grain size.

2.2.2 Toughness and wear resistance

Wear resistance and toughness are two of the most important mechanical properties of a hard coating. Wear resistance refers to the ability of material to resist wear damage. It is macroscopically represented by mass loss or dimensional changes of the coating material under wear working conditions [67]. The term toughness, usually measured in terms of fracture toughness, is a reference to the stress resistance of a material. This reflects the highest stress intensity that a material can withstand without presence of a flaw or fracture, and is used to describe the ability of a material to absorb energy during deformation up to the commencement of fracture [68].

2.2.2.1 The influence of H and E to the mechanical properties of hard coating

From a conventional viewpoint, high hardness is considered as the prime property in defining high wear resistance and damage resistance. However, it was not until studies showed that some polymeric materials with low hardness could provide excellent wear and fracture behavior in impact and erosion conditions, that the importance of elastic modulus in characterizing the wear resistance and damage resistance was recognized. [69]

Recent studies in tribology show that elastic modulus plays an equally important factor in describing the wear resistance and damage resistance of hard coatings [69]. It is now widely accepted that elastic strain to failure value (H/E) is a more reliable indicator of wear resistance for ceramic, metallic and polymeric materials. The ratio of H3 to E2 (H3/E2) can also be used to describe the ability of coating to resist plastic deformation, since the yield pressure (the load required to introduce plastic deformation), Py, is proportional to the hardness and elastic modulus by [69, 70]:

3 2 Py ∝ H /E (2.5)

Meanwhile, the threshold cracking load Pc (the minimum load required to cause cracking) is correlated with hardness and elastic modulus by the relation [71]:

1 푃푐 ∝ ⁄퐸2퐻 (2.6) thus, the values of 1/E2H is frequently used as ranking parameter in describing the ability of a material to resist the formation of cracks.

According to Eq. 2.5 and 2.6, it is evident that the wear resistance and crack resistance of a coating increases with decreasing elastic modulus (higher H/E, H3/E2, and 1/E2H values are achieved). Musil et al.’s study [61] shows that the elastic modulus of a nanocrystalline material is not directly correlated to its hardness. That means, for a given composition, variable sets of values of hardness and elastic modulus can be created by carefully choosing the best deposition method and optimal conditions. Thus, it is possible to improve the wear resistance and elastic modulus by creating a coating with a lower elastic modulus of coating, but high hardness relatively.

2.3 Deposition methods

Surface engineering encompasses a wide range of surface modification techniques, such as coating deposition, surface chemical treatment, ion bombardment etc. [72]. The aims of surface engineering involves changing the properties of the surface and near-surface region in a way to improve the performance, and preventing property degradation of the bulk material caused by interactions between material and the surrounding environment [73].

Since the 1960s, coating technologies have been applied to the surface of metal cutting tools to enhance their durability in operation [11]. The techniques used to produce hard

21 coating can be broadly divided into two categories: deposition and surface modification [4]. In this thesis, only deposition technology will be discussed.

In 1998, it was reported that ~65% of metal-cutting tools in North America and West Europe are coated by chemical vapor deposition (CVD) or physical vapor deposition (PVD) techniques [74]. In this section, the general principles of both CVD and PVD method are described. Further, some other frequently used deposition techniques are also be introduced here.

2.3.1 Chemical vapor deposition (CVD)

Chemical vapor deposition (CVD) is used to describe a group of deposition methods whereby a chemical reaction of a vapor occurs during the deposition process. It is widely used to prepare functional coatings which can be used as hard coatings, passivation layers, diffusion barriers, corrosion-resistant coatings, heat-resistant coatings, and epitaxial layers for microelectronics. [75]

CVD is a well-established technique for coating deposition, and was first industrialized in the 1960s [11]. Generally, CVD contains the creation of vapor reactants, with the constituents of the coating material. The formation of the solid material occurs by the chemical reaction of gaseous reactants at a heated substrate surface in a sealed chamber [2]. Thermal, laser or plasma energies are often used to assist the decomposition of reactants. The basic principle of CVD is shown by Fig. 2.3 and Eq. 2.7.

Fig. 2.3 The schematic of CVD. [75]

Gaseous reactants (g) → Solid material (s) + Gaseous products (g) (2.7)

For example, during the deposition process of a TiN coating, gaseous TiCl4, NH3 and H2 are pumped into a deposition chamber. The gaseous reactants react following Eq. 2.8, and form a TiN coating on the substrate material:

TiCl4(g)+NH3(g)+1/2 H2(g)→TiN(s)+4 HCl(g) (2.8)

The characteristics of the coating can be controlled by varying the experimental conditions, including the substrate material, substrate temperature or composition of the reaction gas mixture, total pressure gas flows. The CVD process provides coatings with high purity and a wide range of chemical composition, and can be used to cover any geometry. The coatings deposited by CVD technique are usually uniform in thickness and properties and exhibit low porosity [75]. Another advantage of CVD, compared with other methods, is that it is a relatively economic process with simple apparatus requirements and can be easily used in large-scale production [2].

2.3.1.1 Plasma-enhanced chemical vapor deposition

Conventional thermally driven CVD requires a relatively high temperature to provide the kinetic energy required for the reaction of gaseous reactants and the growth of a coating. This requirement limits its application in depositing coatings on metal substrates with low

23 melting temperatures. Therefore, the plasma-enhanced chemical vapor deposition (PECVD) technique was developed to overcome this problem. [76]

PECVD is a low temperature deposition technique. By substituting thermal energy with additional electrical energy, PECVD allows deposition to occur at much lower temperatures, thus eliminating the disadvantages due to the high deposition temperatures used in conventional CVD. By using direct current or radio frequency glow discharge, gases can be incorporated into the plasma with energetic electrons. These energetic electrons activate the gaseous components in the chamber and provide the kinetic energy for reaction and deposition. Therefore, the thermal energy required to sustain the deposition can be partly replaced, which lowers the temperature of the deposition process. [77]

Except for the low deposition temperature, PECVD also has some other advantages. Since the energetic interaction between the plasma and the substrate surface is controllable by using bias-controlled or pulsed plasma techniques, a dense and environmental stable coating can be achieved with a variation of chemical compositions. Meanwhile, compared with thermally driven CVD, PECVD provides a higher deposition rate. The benefits of the low deposition temperature and high deposition rate are that PECVD is a low cost, but high productivity, coating approach. [78]

2.3.2 Physical vapor deposition (PVD)

Physical vapor deposition (PVD) encompass a wide range of deposition processes using different physical vapor-phase technologies, such as evaporation, sputtering, laser ablation or ion plating to eject solid materials as atoms or molecules [72]. The atoms or molecules transport through a vacuum or low pressure environment and form thin solid coatings onto various substrates by the condensation of vaporized solid materials. The atoms or molecules will chemically react with gases introduced into the deposition chamber, so called reactive deposition, to form new compounds [79]. The basic principle of PVD is shown in Fig. 2.4.

Fig. 2.4 A schematic diagram showing the basic principles of PVD.

Basically, a PVD process can be divided into three steps [72]:

1) The creation of vapor-phase by evaporation and sputtering.

2) The transport of the vapor species from the source target to the substrate. In an evaporation process, since the transport of vapor species usually occurs under vacuum, few collisions will take place during the transport and deposition is more of a line-of- sight process. However, when plasma or reactive gas is involved, a large number of collisions in the vapor phase will take place during transport to the substrate and the transport of vapor species is no longer a line-of-sight process. This may lead to a larger coverage and a more uniform coating thickness.

3) The nucleation and growth of a coating on the substrate. Compared with CVD processes, the main advantage of PVD is that all the three steps mentioned above occur independently. In contrast, in CVD process, all the three steps take place simultaneously at the substrate. During the PVD processes, the parameters of each step can be independently controlled, that means that it is possible to have a

much greater degree of flexibility in controlling the structure, properties and deposition rate of coating.[79]

For instance, the deposition process for CrN by thermal evaporation PVD can be described as follows:

1) The source metal is heated to the evaporation temperature of Cr.

2) The Cr atoms travel to the substrate and form a Cr layer;

3) Nitrogen gas is subsequently introduced into the deposition chamber and reacts with the Cr layer to form a CrN coating.

By adjusting the deposition time, the thickness of the CrN can be controlled. [37]

2.3.2.1 Sputtering deposition

Sputtering is a non-thermal vaporization technology [80], in which the atoms of either solid or liquid target are ejected by the bombardment of energetic particles, normally ions [72]. The energetic ions are usually produced by two approaches, namely either an ion beam gun or a plasma. Since ion beam sputtering is not suitable for large-scale applications, in most sputtering processes, the energetic ions are produced and accelerated by plasma from a gaseous phase.

Fig. 2.5 shows the basic principle of a plasma-based sputter deposition process. In a sputtering deposition, a cathode (source target) and an anode (substrate) are positioned opposite to each other in a deposition chamber. The deposition chamber ie pre-evacuated and vented with a noble gas, usually argon, and a direct current or a radio frequency bias is used to create a high density plasma from the noble gas which in turn gives rise to a greater degree of ionization. During the deposition process, gas ions produced and accelerated by the plasma bombard the source target from providing atoms. The source atoms pass into the vapor phase and then deposit onto the substrate to form a coating. [81]

Fig. 2.5 A schematic diagram showing the principles of a basic sputtering process. [79]

Sputtering is a relatively inefficient way to induce a solid to vapor transition, the phase change energy of sputtering could be three to ten times larger than a thermal evaporation deposition [79].

In order to decrease the energy cost of sputtering, magnetron sputtering deposition was developed. By trapping energetic electrons with orthotropic magnetic and electric fields, the plasma becomes more intense. The dense plasma greatly increases the efficiency of ionization of the gaseous reactant, and allows deposition processes to occur at relatively low pressures and target voltages with, overall, a higher deposition rate. The efficiency of ionization can be further increased by applying an unbalanced magnetron field to trap more electrons in the plasma, that is so called unbalanced magnetron sputtering. Meanwhile a column of plasma will contact with the substrate, which is conductive for the source ions. As such, larger energetic ion fluxes can be accelerated and bombarded onto the substrate concurrently. This increases the deposition rate as well as the adherence between the substrate and coating. [81, 82]

2.3.2.2 Ion plating

Fig. 2.6 A schematic diagram showing the principles of Ion plating deposition. [79]

Ion plating encompasses a wide range of coating growth techniques in which the growing coating is bombarded with low energy ions prior to, and during, the deposition process [81]. Ion plating was first described in 1960s and it is also known as ion assisted deposition (IAD), ion vapor deposition (IVD) or ion beam assisted deposition (IBAD) [83]. In a ion plating process the source material is normally vaporized by a thermal evaporation technique, normally with a high-voltage electron beam [72]. An ion gun or a plasma created from a noble gas, usually argon, by biasing the substrate to a high negative potential, is used to supply low energy ions [79]. Prior to deposition, the substrate surface is cleaned by energetic ions to provide a good adhesion between the coating and substrate. During deposition, the continuous bombardment of low energetic ions increases the kinetic energy of the deposited atoms, and supplies activation energy for chemical reactions to occur. Consequently, this results in a larger grain size and a more condensed structure of the coatings. Meanwhile the bombardment can also help modify any residual stresses [72]. However, the main disadvantages of ion plating are its high energy cost and a slower deposition rate since some deposited atoms on coating surface will be sputtered away [81].

2.3.3 Double cathode glow discharge plasma technique

Fig 2.7 Schematic diagram showing the double cathode glow discharge deposition technique.

The double cathode glow discharge plasma technique was developed by Prof. Xu (Nanjing University of Aeronautics and Astronautics, China) in the 2000s [7]. This deposition method is derived from Xu Tec, developed by Prof. Xu (Taiyuan University of Technology, China) in the 1980s in response to the requirement for an economic and effective approach to the production of high quality surface alloying layers [84]. As a plasma surface treatment technique, the double cathode glow discharge plasma technique combines the advantages of both plasma and sputtering deposition techniques [85]. With the application of two plasmas at both the target source and the substrate, respectively, dense and readily controllable coatings can be produced with low energy input.

A schematic diagram showing the double cathode glow discharge plasma deposition method is shown in Fig. 2.7. As indicated in this schematic, there are three electrodes in the deposition chamber: one anode and two cathodes. The two cathodes represent the substrate and the target source material. Two independent power supplies are used to generate two low temperature plasmas via glow discharge to heat both the substrate and the source material. The glow discharge plasma strikes the source material supplying the desired elements. Accelerated by the electric potential difference, the energetic ions

30 deposit onto the substrate and form a coating layer. The four most important parameters relating to glow discharge are the target electrode voltage, cathode voltage, working pressure and the distance between the source electrode and cathode. Notably, due to lack of any plasma intensification method, the level of ionization for double cathode glow discharge plasma technique is relatively low. Therefore, relatively high bias voltages were normally used during deposition. [86]

2.3.4 The influence of deposition conditions 2.3.4.1 Substrate Temperature

Substrate temperature plays an important role in deposition processes [87]. Normally, the deposition rate is related to the substrate temperature. A higher substrate temperature leads to a higher deposition rate and, usually, a denser microstructure. The influence of temperature on the microstructure of a coating can be explained by several structure zone models [88]. Firstly, a relatively high deposition temperature leads to greater interdiffusion between the coating material and substrate. This may form a transition zone at the surface of the substrate, which can, potentially, enhance the adhesion between coating and substrate [89]. Further, a higher deposition temperature increases the surface mobility of deposition ions, contributing to fewer defects in the coating [42].

A high substrate temperature may lead to potential tool failure by adversely affecting the tribological properties of the coating and the microstructure of the substrate. Coatings prepared at higher substrate temperatures typically exhibit a residual tensile stress at room temperature, arising from differences in the thermal expansion coefficients between the coating material and the substrate[4]. This may lead to coating cracking and tool fracture [90]. As such, this limits the choice of possible substrate materials [89, 91]. Furthermore, a higher substrate temperature may result in the formation of some brittle phases with low fracture strength, this limits the application of coatings on sharp edges [92]. A higher temperature also leads to problems related to the controlled coating growth [93].

2.3.4.2 Gas pressure

Both CVD and PVD involve gases to eventually form a thin film on a substrate. The partial pressure of the reactive gases can influence the composition of the thin film and determine its microstructure [94]. The gas pressure also influences the transport of vapor species from the source to the substrate. If the pressure of the gas is high, there may be a large number of collisions in the vapor phase during transport to the substrate [94]. This may result in a larger coverage and a more uniform distribution of the vapor that contributes to a more uniform coating thickness [79].

2.3.4.3 Bias and ion bombardment

In any plasma involved deposition processes, the substrate is exposed to the bombardment of energetic ions. The bombardment energy of the energetic coating ions, which is determined by the bias between voltages of the substrate or target and the ground potential, can greatly influence the crystallite size, formation of defects, and compressive stress of the coating [14].

A important impact of bias and ion bombardment is that the bombardment induces a high biaxial compressive residual stress, which enhances the hardness of a coating [14]. Generally, hardness of coatings decreases with increasing tensile residual stress and increases with increasing compressive stress (Fig. 2.8) [95]. By counteracting shear stress and hindering dislocation activity, a compressive stress of 5-7 GPa can enhance the hardness of the coating to values as high as 60 to 100 GPa [96]. For example, a hardness of 72 GPa for a HfB2 coating with a high biaxial compressive stress of approximately 7 GPa was reported by Herr et al. [97], and a 100 GPa hardness was also achieved for a (TiAlV)N coating in a study for similar reasons [98]. However, this high hardness is thermally unstable and deceases after annealing at relatively high temperatures [14, 57]. As a result, it is important to avoid the influence of biaxial compression during deposition process.

Moreover, increasing ion bombardment could also lead to increases in the nucleation density, surface mobility of adatoms, and reaction and diffusion rates, together with a decrease in the formation of interfacial voids. The growth of columnar microstructure could be disrupted by these effects and result in the formation of dense and fine grained structures with near equiaxed grains [99, 100].

Fig. 2.8 Hardness and compressive stress of TiNx coatings deposited by magnetron sputtering vs. the ratio of the N2/Ar flow rate [96]

Meanwhile, studies also show with increasing substrate negative bias, the grain size of the coating decreases, and then increases, after achieving a minimum value with increasing bombardment energy [101]. Also, a moderate bias voltage raises the energy of the deposition ions, increasing their surface mobility. This results in fewer vacancies and a higher coating density. However, a high bias may cause resputtering and decrease the deposition rate [102].

2.3.4.4 Impurities

The properties of coatings also have high sensitivity to impurities. Sometimes impurities may lead to catastrophic failure of components. It has been shown that superhard coatings can only be achieved in nanocomposites that are free of impurities [103].

Impurities can be introduced by a range of causes, for example leakage of the deposition apparatus, insufficient cleaning of the reactor walls prior to the deposition [104], diffusion between the substrate and coating [105] or incomplete vacuum outgassing. Meanwhile, coatings with low impurity concentrations can be achieved by periodic mechanical cleaning, extensive outgassing at high temperature [104], and appropriate setting of deposition parameters, such as plasma density, gas flux and deposition temperature [96].

By changing the structure of nanocomposites, impurities can greatly influence the properties of thin coatings. For example, oxygen impurities in the host lattice can modify the electronic structure and promote shifts in luminescence peaks of Al-N thin coatings [106]. A 1 at.% chlorine addition to TiN coatings degrades the long-term stability of these coatings, and coatings with a chlorine content of 1 to 2 at.% will lose hardness and delaminate from the substrate in humid air [104]. Moreover, research shows hydrogen impurities of several at.% in binary nc-TiN/a-Si3N4 nanocomposites limits their hardness to about 30-35 GPa[104]. The most dramatic degradation of properties is often caused by oxygen impurities. Due to its high electronegativity, oxygen impurities can weaken neighbouring bonds and reduce hardness. In the nc-TiN/a-Si3N4 system, a concentration of only 0.1 at.% of oxygen can cause an apparent decrease of hardness by at least 10 GPa [104]. It was shown by Veprek et al. that no superhard nanocomposites can be obtained when the oxygen impurity content reaches ＞0.3 at.%.

2.4 Characterization Techniques 2.4.1 Scanning electron microscopy (SEM) and energy dispersive spectrometery (EDS)

The scanning electron microscope is one of the most frequently used analytical tools in the material sciences [107, 108]. It permits the observation and characterization of materials at the nanoscale with high depth of field images [109]. Fig. 2.9 shows a schematic of the layout of a SEM. In a SEM, a beam of high-energy electrons, produced by an electron gun, is used to scan over the specimen surface in a raster pattern with the

34 control of electron lenses. The interaction of the electron beam with the sample produces a variety of signals (Fig. 2.10), including secondary electrons, backscattered electrons, and characteristic x-rays. The secondary electrons and backscattered electrons are collected by detectors and used to form images of the sample. Secondary electrons provide strong topographic contrast, whilst backscattered electrons can be used to generate compositional contrast. In addition to this, characteristic x-rays can be used to provide chemical analysis through energy-dispersive x-ray spectrometery (EDS)

Fig. 2.9 A schematic representation of the ray path and components for a SEM. [110]

Fig. 2.10 Electron beam specimen interactions in the scanning electron microscope. [110]

2.4.2 Transmission electron microscope (TEM) and selected area electron diffraction (SAED)

The transmission electron microscope is a high-resolution analytical technique [111]. It is commonly used in the microstructural characterization of coatings [112]. In the TEM, an electron beam of uniform current density is emitted by an electron gun. The electron energy is typically in the range of 60-300 keV (usually around 200 keV). The electron beam irradiates a thin, electron transparent specimen with the control of a two-stage condenser-lens system (Fig. 2.11), and the electrons interact with atoms in the sample through both elastic and inelastic scattering (Fig. 2.12). The transmitted electrons are imaged with a three- or four-stage lens system, onto a fluorescent screen. The image can then be recorded by CCD cameras. The most important lens in the TEM is the objective lens. It forms the images and diffraction patterns that are magnified by the other lenses. High resolution can only be achieved when the focal length of the objective lens is short, and this means the objective lens must be very strong [113].

Fig. 2.11 A schematic representation of the ray path for TEM. [114]

Fig. 2.12 A schematic representation of the signals in TEM. [114]

Bright field (BF) imaging is the most commonly used operational mode in the TEM. The BF images reflect the mass thickness contrast of the specimen. Diffraction contrast in BF images can be used to image crystalline defects and other similar structural features. The thicker areas of specimen or regions with higher atomic number will appear darker in the image. In dark field (DF), the diffracted beam is used for image formation. The DF image shows diffraction contrast and contrast due to crystalline defects in the sample. [114, 115]

According to Bragg’s law (Eq. 2.10), when electrons interact with a periodic atomic structure, they are scattered by the same angle and form a diffraction pattern with distinct reflections. In contrast, if the atomic structure is random, no clear diffraction pattern can be observed. Therefore, the diffraction pattern carries information on the crystallography of the TEM sample such as crystal structure, crystal orientation, lattice parameters, etc. This characterization technique, where information on crystal structure can be studied by electron diffraction based on the electron diffraction pattern of TEM sample, is called selected area electron diffraction (SAED).

Information related to the chemical composition of coatings can be obtained by techniques such as energy dispersive X-ray spectroscopy (EDS) and electron energy loss spectroscopy (EELS) [116]. In EDS, an x-ray spectrum is acquired from a small region of the specimen using a solid-state detector. X-rays which exhibit characteristic energies

37 based on the elements present in the specimen can be used to determine the concentrations of the different elements in the specimen. Energy losses of the transmitted electrons cause by inelastic scattering are measured by an energy loss spectrometer. Information about chemical composition, bonding structure and electronic structure can be obtained from features in the EELS spectrum.

The specimen for TEM must be very thin, typically of the order of 5nm-0.2µm for incident 200keV electrons, depending on the density and elemental composition of the specimen and the resolution desired. Dedicated preparation techniques, such as the use of focused ion beam microscopy (FIB) or electropolishing, are needed to prepare the specimens.

2.4.3 Focus ion beam microscopy

Focus ion beam (FIB) technology was first developed in late 1970s. It is frequently used to study the microstructure, especially cross-sectional microstructure, of thin film coatings and to prepare samples for transmission electron microscopy (TEM) [117].

The components of a FIB system can be broadly divided into three parts: the source (normally a liquid gallium ion source), the ion column, and the stage and beam control [118]. The basic principle of FIB is similar to that of SEM, although sometimes an ion column is combined with a SEM to form a dual-beam FIB system (Fig. 2.13). In a FIB, a group of electrostatic lenses and apertures are employed to focus and control an ion beam that is scanned over the sample surface held under vacuum. The main difference between a FIB and a SEM is that instead of energetic electrons, a focused beam of gallium ions is used, which are much heavier, and thus have much higher momentum than electrons [118]. The gallium ions will lose energy on impact with the surface of a sample, generating electron signals (secondary electrons), and they will sputter neutral and ionized atoms. By collecting the secondary electrons or ionized atoms with a bias voltage, the morphology and the composition of sample can be analyzed. Meanwhile, the sputtering effect, which is called milling, enables sections to be precisely cut into the

38 specimen surface. This can be used in a variety of applications such as cross-sectional observations, TEM sample preparation as well as micromachining operations or the preparation of 3D reconstructions of sample microstructure [117]

Fig. 2.13 A schematic diagram of a dual beam FIB.

The FIB can be used to analyze the cross-sectional morphology of thin film coatings [119]. A schematic diagram of the imaging process is shown in figure 2.14. Firstly, after the region of interest has been located, a protective layer (normally Pt) is deposited over the area of interest to minimize surface damage during milling. Then, a trench will be milled into the surface with relatively high beam current (typically in the nm range) to expose the cross-section of the region of interest. After the initial mill, the cross-section will be ‘clean-up’ milled with a lower beam current (typically ranging from a few hundred pA’s to 1 nA) to remove the redeposited material and clean the cross-section. Since, in a single beam FIB, the ion optical column is fixed, the sample will be tilted (generally by 45°) to enable imaging of the milled cross-section. The cross-section is then imaged with a low beam current, typically ~10 pA, to limit further ion beam damage on

39 the milled surface. The interaction of the incident ion beam generates secondary electrons which are detected and used to form secondary electron images.

+ Protective layer Ga+ beam Ga beam

Fig. 2.14 Preparing a cross-section using a FIB. The specimen is (a) milled using the ion beam and (b) tilted to an angle (θ) so that the cross-section can be imaged.

2.4.3.1 Preparation of TEM samples with a FIB system

Further, FIB provides a very convenient, and rapid, method to prepare TEM samples of thin film coatings [119, 120]. Generally, the sample preparation process includes the deposition of a protective coating on the region of interest. This is usually a platinum coating. Next, a strip defining the TEM specimen about 1 micron in width is made by milling two trenches at both sides of the deposited region. Then, further mills are performed to narrow down the thickness of the specimen and to make it electron transparent (normally less than about 100 nm). This is performed with a lower beam current typically around 100 pA. Finally, the sample is tilted to cut off the TEM specimen from the rest of the sample. The milled TEM sample is then lifted out with a micro- manipulator and placed on a carbon-coated copper grid for TEM observation.

2.4.4 X-ray photoelectron spectroscopy (XPS)

X-ray photoelectron spectroscopy, which is concerned with the measurement of the binding energy of electrons, is widely used to study the surface chemical composition of a wide range of specimens including thin film coatings [121, 122]. An energetic X-ray beam is projected onto the specimen. Electrons whose binding energies are less than the energy of the incident X-ray are ejected as photoelectrons (Fig. 2.15). By collecting the emitted electrons and measuring the kinetic energies of these photoelectrons with an electron spectrometer, the electron binding energy of the photoelectrons emitted can be calculated via the relationship:

Ek = hυ − Eb − Φ (2.9) where h is Planck’s constant, υ is the wavenumber of incident radiation (where hv represents the incident X-ray energy), Ek is the kinetic energy of the photoelectrons, Eb is the binding energy of electron in a specific orbital of an atom and Φ is the work function of a photoelectron analyzer.

Fig. 2.15 Principle of x-ray photoelectron-spectroscopy. [123]

The XPS spectrum is presented as a graph of the intensity of photoelectrons detected versus the binding energy. Those electrons which are excited and escape without further energy loss contribute to the characteristic peaks in the spectrum; those which undergo inelastic scattering and suffer energy loss contribute to the background of the spectrum. The binding energy of photoelectrons is an intrinsic property of an element, and can be slightly influenced by its chemical environment [124]. The categories of elements, the corresponding chemical state of each element, and the quantity of each element simultaneously determine the position and intensity of the peaks in the XPS spectrum of a specimen. By fitting the XPS curve with the characteristic binding energy of each element in different chemical states as recorded in a database, and the quantity and bonding state of each element presented in the specimen can be analyzed [123].

The depth of analysis in XPS is determined by its attenuation length (λ). 85% of the photoelectrons are emitted from a depth of less than 2λ. Since attenuation length is generally a few nanometers. The depth from which the information of chemical composition can be derived is very shallow. That means XPS is a surface analysis technique, and the obtained chemical information can be strongly affected by surface oxidation and contamination of the samples due to their exposure to air. Normally, specimen cleaning is needed to be performed before XPS analysis using argon ion bombardment to remove surface contamination [108, 122].

2.4.5 X-ray diffraction (XRD)

X-ray diffraction (XRD) is a powerful technique for the structural study of thin films [125]. X-ray diffraction is caused by the elastic scattering of x-ray photons by atoms in a periodic lattice (Fig. 2.16). Since the wavelength of X-ray is in the same order of the interatomic distance of a crystal (in the angstrom range), by using crystals as the diffraction grating of X-ray, the crystal structure of a material can be studied from Bragg’s law [126]:

푛휆 = 2푑 sin 휃 (2.10)

42 where n is an integer called the order of reflection, 휆 is the wavelength of the incident X- ray, d is the interplanar spacing and 휃 is the Bragg angle.

Fig. 2.16 The schematic representation of elastic scattering of x-ray photons. [126]

The main application of X-ray diffraction for thin films is the qualitative determination of the phase composition by comparing the XRD pattern with the standard line patterns in a Powder Diffraction File (PDF) database. However, X-ray diffraction can also be used to measure the crystalline size or the relative intensities of the orientations of different crystal planes.

The diffraction peaks from perfect crystals are usually precise and narrow. Sometimes peak broadening may occur due to incomplete destructive interference in scattering directions. The width of the peaks is related to the crystal size. The crystal size can be estimated by the Scherrer equation:

퐾휆 푎 = (2.11) 훽 푐표푠 휃 where a is a measure of the dimension of a grain in the direction perpendicular to the reflecting plane, λ is the x-ray wavelength, β is the peak width, θ is Bragg angle, and K is a constant.

The directions of the reflected beams are determined entirely by the geometry of the lattice, which in turn is governed by the orientation and spacing of the crystal planes[127]. If the lattice parameter, 푎, is known, the indices of crystal planes (ℎ 푘 푙) at which the beam is diffracted can be calculated from the interplanar spacing, d. For a cubic crystal, the equation is:

푎 푑( ) = (2.12) ℎ 푘 푙 √(ℎ2+푘2+푙2)

From the intensity data, the preferred orientation of coating can be evaluated using a texture coefficient (TC(hkl)) calculated by the following equation [128]:

퐼푚(ℎ 푘 푙)⁄퐼0(ℎ 푘 푙) 푇퐶(ℎ 푘 푙) = 1 (2.13) ∑푛퐼 (ℎ 푘 푙)⁄퐼 (ℎ 푘 푙) 푛 1 푚 0

where Im(hkl) is the measured intensity of the (hkl) plane, while I0(hkl) represent the intensity of (hkl) plane in the powder diffraction file, and n is the number of reflection planes. A TC(hkl)) value is greater than 1 means the coating material shows a preference in a specifc (hkl) plane, however, when the TC(hkl)) value is smaller than 1, a lack of grains in (hkl) orientation can be concluded.

2.4.6 Raman spectroscopy

Raman spectroscopy is a spectroscopic technique based on the Raman effect discovered by Krishna and Raman in 1928. When a sample is illuminated by a monochromatic beam of light photons, both elastic scattering (also called Rayleigh scattering), resulting in a beam of photons with the same energy (same frequency) as the incident beam, and inelastic scattering, resulting in a series of photons with lower (called the Stokes line) or or higher energies (called the anti-Stokes line) than the incident beam (hence lower or higher frequencies), will take place. These re-emited photons with frequencies shifted up or down in comparison with original monochromatic frequency, exhibit the energy difference between the incident and scattered photon, and are called Raman effect. The

Raman effect provides information about vibrational, rotational and other low frequency transitions in molecules. Different materials have different vibrational modes, and therefore Raman spectra can be used to characterize the composition of materials. For example, the Raman spectrum of Hydroxyapatite (HAP, Ca10(PO4)6(OH)2) mainly depends from the vibration modes of the phosphate groups [129]. Therefore, the peaks −1 related to the symmetric stretching mode (ν1, at 961 cm ), doubly degenerate bending −1 −1 mode (ν2, at 447 cm and 433 cm ), triply degenerate asymmetric stretching mode (ν3, at 1030 cm−1, 1046 cm−1, 1054 cm−1, and 1076 cm−1), and triply degenerate bending −1 −1 −1 −1 mode (ν4, at 582 cm , 594 cm , 610 cm and 620 cm ) of PO4 group can be observed from the characteristic Raman spectrum for HAP.

Normally, during a Raman analysis the sample is illuminated with a laser beam in the ultraviolet (UV), visible (Vis) or near infrared (NIR) range. The scattered light will be collected with a lens and sent through interference filter or spectrophotometer to obtain Raman spectrum of a sample. Raman spectroscopy can be used to study chemical characterization and micron-scale mapping of solid, liquid and gaseous samples. [130]

2.5 Summary

This thesis aims to study the composition, microstructure and properties of transition nitride metal coatings. The mechanical properties, such as hardness, elastic modulus of transition metal nitride coatings will be measured. The result will be compared with data from other studies. By using a combination of SEM, TEM, XRD and XPS techniques, the composition and microstructure of transition metal nitride coatings can be studied. The microstructure development under different deposition method and conditions, and the relationship between microstructure and the mechanical properties of transition metal nitride coatings will be explored. Raman spectroscopy will be used to investigate the biocompatibility of a tantalum nitirde coating.

Chapter 3

Nanoindentation testing

3.0 Nanoindentation testing

This chapter describes the experimental procedure of nanoindentation, as well as the measurement of hardness and elastic modulus properties. The nanoindentation errors and deformation behavior of coatings during the nanoindentation testing are discussed. Meanwhile, the procedure of a newly developed indentation technique-partial unloading testing is also described here.

3.1 Introduction of nanoindentation testing

Hardness and elastic modulus are very important properties for materials. These parameters are relatively easy to measure. The capacity of hardness and elastic modulus to indicate other mechanical properties also make them suitable for more broadly describing the mechanical properties of materials. Hardness measurements usually include three categories of hardness: scratch hardness, rebound or dynamic hardness and static indentation hardness. Static indentation measurements are the most widely used method for measuring the hardness of coating materials [131]. In a static indentation test, both elastic and plastic deformation data is generated under a diamond indenter of different geometrical shapes. The hardness of coating is calculated from the ratio of the applied load, P, and the contact area, A,[14] and the elastic modulus can be determined from the load to displacement curve using the Oliver-Pharr method [132, 133].

However, conventional indentation tests, which are based on directly observing the indent to determine it size and hence calculate hardness, are not suitable for analyzing the mechanical properties of sub-micron thin coatings, since the indentation impressions are normally not visible under an optical microscope. A mechanical property measurement, which is applicable even when the indentation cannot be imaged, is therefore required. Load- and depth-sensing indentation testing, also commonly called nanoindentation testing, has proven to be useful in probing the properties of submicron-scale coatings, since indentations as shallow as a few nanometers can be measured [95]. In a commonly used method, the load and displacement of the indentation are continuously measured and

47 described by a displacement-load curve. By analyzing the experimental data related to the contact area at peak load, hardness, H, and elastic modulus, E, can be deduced. Since some indenters are also sub-micron in size, the mechanical properties of some materials can be measured with sub-micron resolution [132] .

Although nanoindentation is mainly used for measuring hardness and elastic modulus, it can also be used to measure residual stress [95], and study the fracture behavior of thin coatings when accompanied by observation by transmission electron microscopy (TEM) or focused ion-beam (FIB) microscopy [134].

3.2 Data measurement analysis 3.2.1 Indenters

Indenters are usually made of diamond, because of its high strength and hardness. Indenters used in nanoindentation test can generally be classified into two categories, namely pointed or spherical indenters. [135].

A variety of pointed indenters can be used for determining hardness, such as pyramid indenters and conical indenters. With pointed indenters, the contact strain distribution in a homogeneous material is similar for any depth of displacement and the tested material is usually deformed permanently from the commencement of loading. Thus, pointed indenters clearly reflect the change of properties with depth [136]. The advantage of the conical indenter is that the conical shape can eliminate stress concentrations at sharp edges [137]. However, conical indenters are less widely used, because they are difficult to manufacture. Compared with other pyramid indenters, such as Vickers or Knoop indenters, a Berkovich indenter, with face angles of 65.3°, and only three equilateral sides, can readily achieve a sharp tip [135]. Since a sharper tip is desirable for extremely thin coatings requiring shallow indentation, the Berkovich (Fig. 3.10 (a)) indenter is commonly used to measure the hardness and elastic modulus of coatings [132]. However, the stresses at the pointed tip are very high, so cracks appear and propagate at relatively

48 low loads and depth it is unsuitable to study the failure mechanisms of coated materials [136].

A spherical indenter (Fig. 3.1 (b)) is a truncated cone with an inclined angle of 60° or 90° and a spherical shaped tip [138]. It is more effective to investigate the failure mechanisms in coatings. The contact pressure and stresses in the contact area of a spherical tip increase gradually with increasing load and depth of displacement. The deformation of coatings tested this way commonly starts at the edges or near the contact area after the tensile stress reaches a sufficient value at a relatively high load. As such, the generation and development of cracks can be observed more easily. The minimum critical load and depth of failure can be measured by steps in the load-displacement curve [136]. However, limited by the manufacturing technology, precise spherical geometries cannot be obtained by a diamond indenter with a diameter of only a few micrometers. This leads to inaccurate values for mechanical properties, such as hardness and elastic modulus, since the calculations are based on the assumption of a perfect spherical geometry of the indenter [139].

(a) (b) Fig. 3.1 SEM images of the tips of a pointed indenter (Berkovich indenter) (a) and a Spherical indenter (b). [138]

3.2.2 Hardness

The hardness of a material, H, is a measure of its ability to resist deformation upon loading [131]. It is defined as the ratio of the applied load, 푃, to the contact area, 퐴 , 푃 between the indenter and the specimen, 퐻 = , in the indentation test [14]. The hardness 퐴 of a coating is commonly calculated by the Oliver-Pharr method [132].

In the Oliver-Pharr method, hardness is given by the peak load, 푃푚푎푥, and the projected area of contact at peak load, 퐴푐:

푃 퐻 = 푚푎푥 (3.1) 퐴푐

푃푚푎푥 can be directly acquired from displacement-load diagram (Fig. 3.3). The contact area at peak load is calculated based on the geometry of the indenter and the vertical distance between the tip and the perimeter of the contact area (contact depth), ℎ푐, Fig. 3.12.

Fig. 3.2 A schematic representation of load versus indenter displacement showing quantities used in the analysis as well as a graphical interpretation of the contact depth. [132]

Because the indenter does not deform significantly in nanoindentation testing, it can be assumed that the projected area of contact at peak load is related to the geometry of the

50 indenter, and can be described by an area function (Eq. 3.2) which is related to contact depth, ℎ푐:

퐴 = 퐹(ℎ푐) (3.2)

Fig. 3.3 A schematic representation of a section through an indentation showing various quantities used in the analysis. [132]

The functional form, 퐹, can be determined experimentally, for which the hardness is known. For a Berkovich indenter:

2 퐴 = 24.5ℎ푐 (3.3)

The contact depth can be calculated by Eq. 3.4:

ℎ푐 = ℎ푚푎푥 − ℎ푠 (3.4)

where ℎ푠 is the displacement of the surface at the perimeter of the contact. Since the

ℎ푚푎푥 can be measured from the displacement-load diagram, ℎ푐 can be obtained if ℎ푠 is known. Oliver et al. show that the hs of a conical indenter satisfies Eq. 3.5:

푃 ℎ = ɛ 푚푎푥 (3.5) 푠 푆

51 where ɛ is the geometric constant for the indenter, and S is the unloading stiffness at peak load, which can be established by the slope of unloading curve:

푑푃 푆 = (3.6) 푑ℎ

Combining Eqs. (3.1) to (3.6), the hardness of a coating material can be calculated.

Although the Oliver-Pharr method was firstly developed with a conical indenter, it is also suitable for other indenters (e.g. Berkovich indenter, ɛ = 0.75) [139].

3.2.3 Elastic modulus

The elastic modulus can be calculated from a relationship between contact stiffness, contact area and elastic modulus, which is independent of geometry [133]:

푑푃 2 푆 = = 훽퐸 √퐴 (3.7) 푑ℎ √휋 푟

Where, β is the shape factor (1.034 for a Berkovich indenter and 1 for a spherical indenter), Er is called the reduced modulus, which is the combined modulus of the specimen and the indenter:

1 (1−휐2) (1−휐2) = + 푖 (3.8) 퐸푟 퐸 퐸푖

where 퐸 and 휐 are the elastic modulus and Poisson’s ratio for the specimen, and 퐸𝑖 and 휐𝑖 are the same parameters for the indenter [132].

3.2.4 Nanoindentation errors

The measurement of hardness of hard coatings by means of indentation can be affected by a number of factors. Since the deformation behavior of substrate is usually different from the coating material, the measured hardness and elastic modulus can be influenced by the substrate during nanoindentation testing. This substrate effect will increase with increasing depths of penetration. In order to eliminate the influence from the substrate, a limitation of load and indentation depth is required. It is widely accepted that the indentation depth should be less than 10% of the total thickness of coating in order to avoid any influence from the substrate.[112]

Furthermore, the indentation size effect is particularly common in nanoindentation testing. A small indentation of a few microns in size results in a large strain gradient, leading to geometrically necessary dislocations which will cause an apparent increase in the measured hardness [140]. Errors can also arise from surface contamination layers. Some coating materials, such as tantalum nitride [141], achieve chemical inertness by forming a oxides or carbides layers at the surface. Since the dislocations may pile up in the surface contamination layer during nanoindenation, a different mechanical property between the contamination layer and coating material will contribute to unrealistic hardness and elastic modulus values, especially when the penetration depth is very shallow. Meanwhile, the surface roughness of coatings may also impact the measured data resulting in an inaccurately measured contact area, especially if the roughness value is comparable to the indentation depth of few hundreds nanometers. Errors can also be caused by insufficient contact time or imperfect geometry of the indenter tip. [131]

3.3 Deformation and fracture behavior

The deformation behavior of coatings are complex and can be influenced by several factors, such as the coating material itself, substrate, coating/substrate interface and deposition method used [142]. In addition to the plastic and elastic deformation that occur during loading, coatings may also fracture or delaminate under a high load. Swain and

Menčík [143] studied the cross-sections of different coatings and reported the non-elastic displacement response of various combinations of coating and substrate materials during nanoindentation testing. According to their study, hard coatings, which are usually harder and more brittle than substrates, usually suffer cracking rather than delamination during nanoindentation testing.

By studying the deformation behavior of TiN, NbN, TiN/ZrN and TiN/NbN coatings deposited by cathodic arc deposition (the substrate for the TiN and TiN/NbN coatings was a powder metallurgical hard tool steel (ASP23), and the substrate for the NbN and TiN/ZrN coatings was a single phase austenitic stainless steel (304L, annealed condition)), the sequence of the deformation events of a hard brittle coating were summarized by Hainsworth et al. [142] as:

1) The initial loading leads to elastic and plastic deformation of the coating. At this stage the shear stress applied to the substrate is small, and not sufficient to deform the substrate.

2) The shear stress experienced by the substrate exceeds a critical value. Plastic yielding occurs in the substrate. The coating is forced to flex and bend to conform to the deformation of the substrate, causing significant tensile stresses around the outer periphery of the contact zone. The shear stresses in the coating cause plastic flow in areas around the axis of the contact and along the indenter edges.

3) Through-thickness cracks occur due to increasing tensile stress around the contact zone. The deformation is dominated by the substrate rather than the coating. Increasing load causes radial cracks as the coating is bent into the substrate.

4) Cracks emerge at the coating-substrate interface as the coating is continually loaded.

5) As the load is released, the system initially recovers in a completely elastic manner. Then, the propagation of interfacial cracks allows the coating to detach itself from the

54 substrate in order to reduce its own elastic strain energy due to bending. The interfacial cracks continue to propagate until the system comes to equilibrium.

3.4 Partial unloading testing

Partial unloading is an in-situ indentation technique, which requires several unloading and reloading cycles at different loads before achieving the maximum load. This technology is based on the fact that the unloading process is totally elastic, and would not impact the mechanical properties of a hard material [144]. During the partial unloading testing, the loading process is divided into a number of steps. After each loading step the indenter is unloaded to a certain percentage (this percentage should not be 0% in order to keep a constant contact) of the current load, then the indenter will be loaded to the next unloading load. The load to displacement curve at each unloading point is continually recorded. Based on the load to displacement curve, the hardness and elastic modulus can be obtained from each unloading point. The major advantage of this technique is it can be used to analyze the variation of hardness and elastic modulus with increasing indentation depth [145].

3.5 Summary

In this thesis, three indentation systems, namely Hysitron Triboindenter workstation (Hysitron, Minneapolis, USA), UMIS (Ultra-Micro Indentation System 2000, CSIRO, Sydney, Australia) workstation, and Durascan microindentation system (Struers, Denmark), were used to perform indentation testing on the as-deposited coatings. By measuring the hardness and elastic modulus of the as-deposited coatings, and introducing contact damage to the coating surfaces, the mechanical properties and fracture mechanism of the as-deposited coatings were studied.

Chapter 4

The microstructure and mechanical properties of tantalum nitride coatings

4.1. Introduction

In addition to conventional applications of transition metal nitrides, such as diffusion barrier layers, corrosion resistant coatings and hard wear-resistant coatings, tantalum nitride coatings, most commonly those based on Ta2N and TaN, show significant potential for use in medical applications. Studies by Leng et al. showed that tantalum nitride exhibits better blood compatibility than either low-temperature isotropic pyrolytic carbon (LTIC) or TiN, and can be used as a substitute for LTIC in the manufacture of artificial heart valves [5]. Meanwhile, tantalum-based implants also exhibit excellent biocompatibility in orthopedic and dental applications [146]. A new research direction for the application of tantalum nitride is using tantalum nitride coatings to increase the bioactivity of Ti alloy bio-implants in order to help the reconstruction of bone in the human body, since titanium alloys have relatively poor compatibility with bone-like tissues [147]. This requires tantalum nitride to have both excellent bioactivity and good mechanical properties to survive the long-term effects of load-bearing applications.

Recent studies have demonstrated that the double cathode glow discharge plasma technique is a simple, yet effective, coating method to enhance the surface properties of Ti alloys. Metal silicide or metal nitride coatings prepared using this technique exhibited a nanocrystalline structure, with good adhesion to the substrate [148, 149]. In the present study, tantalum nitride coatings were deposited using a double cathode glow discharge technique. Following deposition, the microstructures of the as-deposited coatings were characterized in detail and the mechanical properties of these coatings were assessed by indentation testing. The relationship between mechanical properties, deposition conditions and microstructure is discussed and compared with other studies focused on TaN-based coatings.

4.2 Experimental methods

The substrate material was Ti-6Al-4V in the form of disks with a diameter of 40 mm and a thickness of 3 mm. The nominal composition of this alloy in wt.% is given as: Al, 6.42;

V, 4.19; Fe, 0.198; O, 0.101; C, 0.011; N, 0.006; and the balance, Ti. Before sputter deposition, the substrates were ground and then polished consecutively with silicon carbide papers down to 1200 grit, followed by cleaning with pure acetone and distilled water in an ultrasonic bath. A 99.99% purity Ta disk, which was 100 mm in diameter and 5 mm thick, was used as the target in the deposition process. Tantalum nitride coatings were deposited onto the polished Ti-6Al-4V substrates using double cathode glow discharge apparatus. During the process of sputter-deposition, one cathode was the target composed of the desired sputtering material, and the other was the substrate. When voltages are applied to the two cathodes, glow discharge occurred, as described elsewhere [150]. The base pressure in the chamber was 5×10−4 Pa, and the working pressure was 35 Pa at a constant Ar gas flow rate of 100 sccm during deposition. Synthesis of the two coatings was conducted in a flowing Ar + N2 gas mixture, with the Ar:N2 flux ratios of 20:1 and 20:3, respectively. Other deposition parameters are shown in Table 4.1.

Table 4.1 The deposition conditions of tantalum nitride coatings #1 #2

Atmosphere Ar:N2=20:1 Ar:N2=20:3

Working pressure 35 Pa 35 Pa

Target bias voltage -750 V -750 V

Substrate bias voltage -300 V -300 V

Substrate temperature 700℃ 700℃

Target & substrate distance 10 mm 10 mm

Treatment time 2h 2h

The phase composition of the as-deposited Ta-N coatings was studied using an X-ray diffraction (XRD) system (PANalytical Empyrean) equipped with a Cu anode. Cu Kα (λ= 1.5406 nm) radiation was used during the analysis, and the instrument was operated at a current of 40 mA and an energy of 45 kV. The X-ray signal was collected over of 2 theta

58 values from 30° to 90°. The cross-sectional morphology of the two as-deposited coatings was examined by scanning electron microscopy (SEM, Hitachi S3400). The microstructure of two coatings was also investigated by a transmission electron microscope (TEM, Philips CM200) operating at 200 kV. TEM samples were prepared using a dual-beam focused ion beam system (FIB, XT Nova Nanolab 200), using methods described elsewhere [119]. A UMIS (Ultra-Micro Indentation System 2000, CSIRO, Sydney, Australia) workstation equipped with a Berkovich diamond indenter was used to measure the mechanical properties of as-deposited coatings. During nanoindentation, a maximum load of 400 mN was applied to the surface of each coating with a holding time at maximum load for 10 s. The maximum contact depths were smaller than 10% of the coating thickness. To ensure reproducible data, twelve indentations were performed on three different areas of each coating. The hardness (H) and Elastic modulus (Er) of the coatings were calculated according to the Oliver-Pharr method [132].

To investigate contact damage resistance, 200 g and 1000 g loads were applied to each as-deposited coating using a microindentation system (Durascan, Struers, Denmark) equipped with a Vickers diamond indenter. The cross-sectional structures of the coatings under and adjacent to these indents were characterized by a single beam focused ion beam microscope (FEI xP200, FEI instrument, Hillsboro, USA) [119]. All samples were sputter coated with a thin layer of gold prior to the FIB analysis to protect the near surface features and also minimize charging effects.

4.3. Results 4.3.1 Microstructure and phase analysis

Fig. 4.1 X-ray diffraction patterns of the Ta-N coatings prepared using two different Ar:N2 flux ratios.

The X-ray diffraction spectra collected from as-deposited coatings deposited on Ti-6Al-

4V substrates are presented in Fig. 4.1. The results show that hexagonal Ta2N (JCPDS

NO. 26-0985) is the dominant phase in the coating deposited with the Ar:N2 flux ratio of 20:1, while fcc-structured TaN (JCPDS NO. 49-1283) is the primary phase for the coating deposited with the Ar:N2 flux ratio of 20:3. In addition for this coating, the diffraction peak intensity of the (200) plane for the TaN phase is signiﬁcantly higher than that of the corresponding standard powder diffraction pattern, indicating that this coating exhibits a strong (200) preferred orientation. Moreover, the broad peaks appeared in these spectra suggest a fine grain size was present in both coatings.

Often, the presence of residual stresses and/or the substitution of impurity atoms with different atomic radii canl slightly change the lattice parameter and, thus, lead to a shift in the position of diffraction peaks. However, the XRD peak positions of the tantalum nitride phases present in each coating fit well with the corresponding powder diffraction file data. This observation suggests that there is no appreciable residual stress present in these coatings.

a b

31.5μm 20.5μm

30 μm 30 μm

Fig. 4.2 Typical Cross-sectional SEM micrographs of the two as-deposited coatings deposited on Ti-6Al-

4V substrates; (a) Ta2N (corresponding to a Ar:N2 flux ratio of 20:1); (b) TaN (20:3).

As shown by the SEM cross-sectional images (Fig. 4.2), both coatings display a uniform and featureless structure without any visible pores or cavities. The interfaces between the coatings and their respective substrates appear sharp and straight. Moreover, the coatings appear tightly adhered to the substrates without any evidence of delamination. It is also apparent that the Ta2N coating is ~32 μm thick, whilst the thickness of the TaN coating is

~20 μm thick. Clearly, this observation shows that with increasing Ar:N2 flux ratio, the

Ta2N coating exhibited a higher deposition rate than the TaN coating. A similar phenomenon was also found for tantalum nitride coatings prepared by reactive magnetron sputtering [151]. For the tantalum nitride coatings deposited on silicon substrate by reactive magnetron sputtering with a flow rate of 100 sccm, a RF input power of 100W, and a negative substrate bias of 100 V, the deposition rate decreased from about 11 nm/min to only 5 nm/min when the nitrogen partial flow rate increased from 0% to 15%.

a b

Deposition direction

100 nm

c d

Deposition direction

100 nm

Fig. 4.3 (a) Cross-sectional TEM bright-field micrograph of the Ta2N coating. (b) SAED pattern of the coating the Ta2N coating. (c) Cross-sectional TEM bright-field micrograph of the TaN coating. (d) SAED pattern of the TaN coating.

According to the bright field TEM micrograph (Fig. 4.3 (a)), the Ta2N coating is composed of equiaxed grains with a diameter of about 10 nm. However, the arrangement of grains shows a linear structure aligned parallel to the direction of growth. Nanopores, typically less than 10 nm in diameter, can also be widely observed in this coating. The

TaN coating (Fig. 4.3 (c)) shows a generally similar microstructure compared to the Ta2N coating. That is, nanopores ranging from 5~10 nm in diameter are homogenously distributed amongst an agglomeration of ~ 10 nm diameter equiaxed grains. Based on the TEM micrographs, the volume fracture of nanoporosity in the two coatings can be

62 calculated as ~10%. The corresponding selected area electron diffraction (SAED) patterns acquired from the center of the two coatings (Fig. 4.3 (b) and (d)) show the equiaxed grains in Ta2N coating deposited with Ar:N2 flux ratio of 20:1 were identified as Ta2N (Fig. 4.3 (b)), which is consistent with the XRD analysis. Whilst for the TaN coating, diffraction rings corresponding to the (111), (200) and (220) planes of fcc TaN were observed that confirmed the presence of TaN as the only phase in this coating. Notably, the diffraction rings for both coatings are continuous, suggesting that the coatings have a very fine grain size.

4.3.2 Nanoindentation testing

a b

Fig. 4.4 (a) Load-displacement curves for the as-deposited tantalum nitride coatings under nanoindentation; (b) Hardness and elastic modulus of as-deposited tantalum nitride coatings; (c) The H/E and H3/E2 ratios derived from the hardness and elastic modulus of as-deposited tantalum nitride coatings. (The error bars represent one standard deviation of the data)

It is generally accepted that if the ratio of the indentation depth to coating thickness is less 10%, the effect of the substrate on the hardness and elastic modulus values obtained can be ignored [112]. As shown by the load-displacement (p-h) curves (Fig. 4.4 (a)) the maximum contact depths for both coatings are much smaller than 10% of the coating thickness, indicating that the contribution to the deformation behavior from the underlying substrate is negligible. The smooth curves, without discontinuity, for both the loading and unloading stages, suggest that both coatings are free from cracking during nanoindentation testing. The hardness and elastic modulus values of both coatings are shown in Fig. 4.4 (b). The H and E for fcc TaN-based coating are 25±1GPa and 205±16

GPa, respectively. Compared to the fcc TaN coating, the Ta2N coating has both a higher hardness (H, 34±2 GPa) and elastic modulus (E, 266±15 GPa).

The influences of both hardness and elastic modulus on the mechanical properties of solid materials have been widely studied [69, 70]. It is widely accepted that a higher hardness to elastic modulus (H/E) value, and thus a higher plastic resistance ratio H3/E2, indicates a better resistance to wear damage and plastic deformation [152]. As shown in 3 2 Fig. 4.4 (c), with increasing N2 partial pressure, the changes in both H/E and H /E in tantalum nitride coatings show similar trends as the hardness and elastic modulus values 3 2 (Fig. 4.4 (c)). The H/E and H /E values of the Ta2N coating are 5% and 50% greater, respectively, than that of the TaN coating. This means that Ta2N coating is likely to exhibit better wear resistance and a higher threshold load for the initiation of plastic deformation than the TaN coating.

4.3.3 Contact damage testing

a b

2.3 μm 2.7 μm

c d

5 μm 5 μm

Fig. 4.5 FIB cross-sectional micrographs of the 200g Vickers indents for (a) Ta2N coating; (b) TaN coating;

FIB cross-sectional micrographs of the 1000g Vickers indents for (c) Ta2N coating; (d) TaN coating.

The result of contact damage testing was evaluated by FIB. As shown in Fig. 4.5 (a) and (b), no structural damage can be observed from the cross-sectional micrographs of the

200 g indents on both the Ta2N and TaN coatings. Furthermore, the indentation on the

Ta2N coating is about 0.4 μm shallower than that on the TaN coating, indicating a high 3 2 resistance to plastic deformation for the Ta2N coating, consistent with the H /E calculations. When the indentation load was increased to 1000g, edge cracks, emanating from the impression corner, were generated in the region close to the indentation impression edge in the Ta2N coating (Fig. 4.5 (c)). These edge cracks show only very shallow penetration and extend downwards by less than 1 μm from the free surface, and can barely be observed from the cross-sectional micrograph. It is also evident that no

65 cracks are visible in the sample sub-surface. In contrast, significantly larger edge cracks with greater penetration into the coating surface were visible for the 1000g indent on TaN coating. Beneath the edge cracks, inclined cracks, which are located below the indentation surface and propagate to a greater depth at an acute angle to the loading direction, can be observed in the sub-surface (Fig. 4.5 (d)).

Although at an applied load of 1000 g, structural damage was detected in both coatings, it is evident that more energy is required for the generation and propagation of cracks in the

Ta2N coating and therefore it exhibits a greater resistance to the formation and growth of cracks caused by mechanical loading.

4.4 Discussion

It is broadly accepted that the mechanical properties of nanocrystalline coatings are affected by several factors such as composition, interatomic bonding, grain size and preferred orientation, grain boundary structure and the presence of any internal residual stresses etc. [153, 154]. In this study, under different nitrogen partial pressure conditions, the two tantalum nitride coatings show similar microstructural features, comprising fine, and nanoscale equiaxed grains together with evenly distributed nanopores, and were free from microcracks and delamination.

Both the XRD and SAED analyses indicated the phase composition of tantalum nitride coatings changes with increases in nitrogen partial pressure. The coating deposited with the Ar:N2 flux ratio of 20:1 exhibits a hexagonal Ta2N phase, which is commonly observed at lower deposition nitrogen partial pressures [46]. When a higher nitrogen flow was introduced (Ar:N2 flux ratio of 20:3 here), the coating structure changed to fcc TaN. A similar transition was observed in tantalum nitride coatings deposited by either magnetic sputtering or ion beam assist deposition [47, 155]. It is worth noting that in these samples that some transient phases, such as hexagonal-TaN, were observed when these coatings were prepared by magnetic sputtering deposition, but these were not observed in this study.

Moreover, unlike the TaN coatings prepared by PVD processing, which commonly exhibit a (111) texture [47, 151], the fcc TaN coating fabricated in this study exhibited a strong (200) preferred orientation. Based on Pelleg’s theory [156], the driving force for the texture preference during grain growth is the tendency to form a coating with the lowest overall energy, which is determined by the competition between the strain and surface energies. For fcc TaN, which is isostructural to NaCl, the lowest surface energy is associated with the (200) plane and the lowest strain energy is exhibited by the (111) plane [157]. Since the strain energy is proportional to the coating thickness [158], a (200) preferred orientation is normally expected for fcc TaN at small coating thickness values as the overall energy is dominated by the surface energy. In contrast, a (111) preferred orientation will normally be developed at great coating thicknesses, since strain energy then becomes dominant. However, although the thickness of the fcc TaN coating is about 20 μm, the high thermal energy provided by the high substrate temperature increases the surface mobility of the deposited atom and consequently lead to the dominant factor being surface energy [158, 159]. Therefore, to lower the overall energy of the coating, a (200) preferred orientation was developed by the fcc TaN coating. A similar preference for the low surface energy (200) plane has also been reported for other fcc transition metal nitrides, such as CrN and TiN [160, 161].

As the consequence of the overall similar microstructure (i.e. very fine grain size), but differences in phase composition in these two coatings, it can be concluded that the difference in elastic modulus and hardness of the two coatings is mainly influenced by the intrinsic properties of the coatings’ phase constituents [13]. Since Ta2N exhibits a hexagonal structure and TaN exhibits an fcc-structure, the relatively fewer slip systems for hexagonal Ta2N would better inhibit dislocation motion. As such, this presumably contributed to the considerably higher hardness of the Ta2N coating compared with the TaN coating.

Based on recent experimental work [14, 162], a nanocrystalline material may achieve its highest hardness with a grain size of around 10 nm to 20nm. Compared with the tantalum nitride coatings deposited by other methods, which typically exhibit a denser structure

67 with larger grain sizes [163], a typical feature of group of tantalum nitride coatings in this study is the high density of nanopores distributed homogenously throughout the nanocrystalline coatings as well as the fine grain size close to 10 nm. In spite of the possible decrease in hardness due to the presence of nanoporosity [164], the finer grain size of the coatings in the present study leads to a relatively high level of hardness.

According to the literature, the hardness of hexagonal Ta2N coatings prepared by magnetron sputtering varies between 30 to 35 GPa [46, 48]. For fcc- TaN coatings hardness values of approximately 30 GPa were noted by a number of research groups who used magnetron sputtering [165, 166]. However, for a fcc TaN coating with (200) preferred orientation, the reported hardness was much lower. For example, Bernoulli et al. reported a hardness of 20 GPa for a fcc TaN coating with (200) a preferred orientation prepared by DC magnetron sputtering [50]. Moreover, a hardness of 26 GPa was measured from a fcc TaN coating deposited using RF magnetron sputtering with a strong (200) preferred orientation and a nonporous structure with a grain size of 50 nm [167]. In view of this, the tantalum nitride coatings prepared in our study, show comparable hardness (34±2 GPa for hexagonal Ta2N coating, and 25±1 GPa for fcc TaN coating) to those tantalum nitride based coatings prepared by magnetron sputtering deposition methods, even though the coatings studied here include the presence of nanopores [46, 48, 50].

It is commonly believed that porosity could act as stress concentrator for the nucleation of cracks. However, Reddy et al. [168] suggested that nanopores, with a diameter of few nanometers, may not act as a source of failure, since they are much smaller than the critical size that initiates a crack, and, further, an appropriate concentration of small nanopores may enhance the damage tolerance of nanocrystalline material by absorbing the applied mechanical energy. On the other hand, the existence of nanopores resulted in a lower E value for the nanocrysalline coating, since the elastic modulus of coating shows a decreasing trend with increasing porosity [169]. Moreover, as a result of the locally lower elastic modulus at a grain boundary, the decrease in elastic modulus is also related to the volume fraction of grain boundaries in the coating, which increases significantly with decreasing grain size [1]. Note, the elastic modulus of the Ta2N coating measured

68 here is 266±15 GPa, much lower than values of 350 GPa to 400 GPa determined for those full-dense hexagonal Ta2N coatings, with similar hardness values, prepared by dc magnetron sputter deposition [48, 163]. As a result, the hexagonal Ta2N prepared in the present study has higher H/E and H3/E2 values, potentially leading to improved damage resistance and wear resistance. For the fcc TaN coating, a significant increase in the H/E and H3/E2 values, respectively, was also obtained compared with some fcc TaN coatings prepared by dc magnetic sputtering [49]. Reddy et al.’s work suggested that damage resistance was increased because nanopores provide a space to accommodate plastic deformation, rather than the formation of an initial crack. This present study indicates that the decrease of elastic modulus of caused by the existence of nanopores also played an important role in promoting the damage resistance of tantalum nitride coatings. This suggests that the control of nanopore size and distribution during deposition may be an effective way to tune the microstructure and, hence, the mechanical properties of hard coatings. However, future studies need to be done to examine the effect of variation in porosity volume fraction on mechanical behaviors. Rather than creating a dense microstructure, by making a small compromise in hardness, the damage resistance and wear resistance of hard coatings could be greatly improved by introducing a finite amount of nanopores through selected deposition conditions.

4.5 Conclusion

The effect of increasing nitrogen partial pressure upon the phase composition, microstructure and mechanical properties of tantalum nitride coatings deposited by the double cathode glow discharge technique were investigated. The following conclusions can be drawn:

1. With increasing nitrogen partial pressure the primary phase of the tantalum nitride- based coatings transformed from hexagonal Ta2N to fcc TaN, the latter with a strong (200) preferred orientation.

2. Both coatings formed a homogenous microstructure composed of ~10nm diameter equiaxed grains together with 5~10nm diameter nanopores. The overall microstructure did not vary significantly with increasing nitrogen partial pressure. Compared with the tantalum nitride coatings deposited by other methods, this microstructure yielded a reasonably high hardness, but relatively low elastic moduli.

3. The hexagonal Ta2N coating exhibits higher hardness, elastic modulus, damage resistance and wear resistance compared to the fcc TaN coating.

4. The presence of nanopores formed in the tantalum nitride coatings prepared by double cathode glow discharge leads to improved damage resistance and wear resistance.

Chapter 5

The influence of oxygen on the microstructure and mechanical properties of tantalum nitride coatings deposited by the double cathode glow discharge plasma technique

5.1 Introduction

The oxidation resistance of tantalum nitride coatings mainly arises from the presence of a passive tantalum oxide layer that forms on the specimen surface. However, this tantalum oxide passive layer can be easily destroyed by aggressive thermal or chemical conditions [141, 170]. Without that protection, oxygen can penetrate into the tantalum nitride coatings and adversely affect its structure and properties. As was described in Chapter 2, these oxygen-related impurities are normally harmful to the mechanical properties of these coatings. In this chapter, the influence of oxygen on the microstructure and mechanical properties of tantalum nitride coatings, deposited by the double cathode glow discharge plasma technique, was studied.

Two tantalum nitride-based coatings were deposited on Ti-6Al-4V substrates under an O2 gas flow by the double cathode glow discharge plasma technique. By using XRD, XPS SEM and TEM, the composition and microstructure of the as-deposited coatings were then characterized in detail. The mechanical properties of the as-deposited coatings were assessed by nanoindentation testing. By comparing the results in this study with the data for tantalum nitride coatings presented in Chapter 4, the effects of oxygen on the microstructure and the mechanical properties of tantalum nitride coatings are discussed.

5.2 Experiment method

Ti-6Al-4V substrates and a Ta target were used in this study. The size and composition of both substrates and target were detailed in Chapter 4. Before deposition, the substrates and target were polished and cleaned following the procedures described in Chapter 4. Two tantalum oxy-nitride coatings were deposited onto the Ti-6Al-4V substrates using double cathode glow discharge apparatus with Ar + N2 +O2 mixed gas. The Ar:N2:O2 flux ratios were 20:1:1 and 20:3:1, respectively. The deposition conditions are shown in Table 5.1. Notably, higher negative biases were applied on both target and substrate in order to acquire a dense microstructure with equiaxed grain by increasing the nucleation

72 density and increases the surface mobility [99]. The two as-deposited coatings were designated #1 and #2.

Table 5.1 The deposition conditions of Ta-based coatings #1 #2

Atmosphere Ar:N2:O2=20:1:0.5 Ar:N2:O2=20:3:0.5

Working pressure 35 Pa 35 Pa

Target bias voltage -800 V -800 V

Substrate bias voltage -350 V -350 V

Substrate temperature 700℃ 700℃

Target & substrate distance 10 mm 10 mm

Treatment time 2h 2h

The composition of as-deposited coatings was studied by XRD (PANalytical Empyrean) equipped with a Cu anode. Cu Kα (λ= 1.5406 nm) radiation was used in the analysis. The X-ray signals were collected over 2 theta values from 15° to 90°. The cross-sectional morphology and microstructure of as-deposited coatings was observed by both SEM (Hitachi S3400) and TEM (Philips CM200, 200 kV). Details of TEM specimen preparation process were given in Chapter 4.

A Hysitron Triboindenter workstation equipped with a Berkovich diamond indenter was used to measure the hardness and elastic modulus of the as-deposited coatings. During nanoindentation testing, a maximum load of 8000 μN was applied to the surface of each coating at a loading rate of 800 μN/s and holding time of 5s at maximum load. 10 indents were performed on each coating. The maximum contact depth for each indent was carefully controlled to be less than 10% of the coating thickness. The hardness and elastic modulus of the coatings were calculated according to the Oliver-Pharr method [127].

The contact damage testing was performed using a UMIS workstation equipped with a spherical diamond indenter with a radius of 5 μm. The contact load was 500 mN. The cross-sectional structures of the indents were characterized by a FIB microscope. All the samples were sputter coated with a thin layer of gold prior to the FIB analysis to protect the near surface features and also minimize charging effects.

5.3 Results 5.3.1 XRD

Fig. 5.1 X-ray diffraction spectra for the two as-deposited coatings.

The X-ray diffraction spectra for the two as-deposited coatings collected over a 2 theta range from 15° to 90° are shown in Fig. 5.1. It can be seen that for the diffraction spectrum for Coating #1, deposited with a gas mixture Ar:N2:O2=20:1:0.5, contains reflections from both hexagonal Ta2N (JCPDS No. 26-0985) and orthorhombic Ta2O5 (JCPDS No. 25-0922). The peaks observed at 2 theta values of 33.8°, 36.3°, 38.5°, 50.1°, 60.3°, 66.7° and 72.8° were indexed to the (100), (002), (101), (102), (110), (103) and

(112) planes of hexagonal Ta2N. Meanwhile, peaks corresponding to the (1 11 0) and (3

11 1) planes of orthorhombic Ta2O5 were also present.

For Coating #2, with a gas ratio of Ar:N2:O2=20:3:0.5, no peaks consistent with the presence of hexagonal Ta2N were observed. All the tantalum nitride was present in the form of the fcc-structured TaN phase. The (111), (200), (220) and (311) peaks for fcc cubic TaN were observed at 2 theta values of 36.1°,41.7°,60.5° and 72.3°. Similarly, orthorhombic Ta2O5 was again present in this coating. It can be seen that with the introduction of oxygen during the deposition process, that Ta2O5 was formed in both coatings.

5.3.2 XPS

Fig.5.2 XPS spectra for the two as-deposited coatings argon ion beam cleaned for 300s.

The XPS spectra obtained from the two coatings are shown in Fig.5.2. Before the XPS survey scan was recorded, the surfaces of both coatings were cleaned by argon ion beam etching for 300s. Ta, N and O were identified from both coatings. However, prior research shows the intensity of Ta peaks arising from Ta-N-based coatings will decrease

75 and the peak positions will shift to lower binding energy after argon ion beam cleaning due a sputter-induced reduction of the Ta, since nitrogen can be sputtered away from the nitride by the Ar ions during the cleaning [50, 171]. To avoid the influence of both surface oxidation and peak shifts during ion beam cleaning, detailed analysis of the tantalum peaks was performed both on the original, uncleaned, surface, to investigate specifically the bonding state of Ta with N, and on the surface argon ion beam cleaned for 300s, to study specifically the bonding state of Ta with O, for each coating.

Coating 2

Coating 1

Fig. 5.3 High resolution Ta4f XPS spectra obtained from the original surface of as-deposited coatings, that is without ion beam cleaning.

According to the curve-fitted high resolution XPS spectra of the Ta4f peak acquired from the original, uncleaned surface of the two coatings. The Ta4f spectra of the two coatings can be deconvoluted to three sets of 4f doublets. For Coating #1, the binding energy of the Ta4f peaks can be indexed to the chemical state of Ta in Ta2N (Ta4f7A=22.7 eV, Ta4f5A= 24.6 eV) [151], Ta in Ta-O-N (Ta4f7B=25.4 eV, Ta4f5B= 27.3 eV) [42], and

Ta in Ta2O5 (Ta4f7C=26.2 eV, Ta4f5C= 28.1 eV) [172]. Meanwhile, the Ta4f doublets from Coating #2 are consistent with the binding energy of Ta in TaN (Ta4f7A=23.9 eV,

Ta4f5A= 25.9 eV) [42], Ta in Ta-O-N (Ta4f7B=25.2 eV, Ta4f5B= 27.1 eV) and Ta in

Ta2O5 (Ta4f7C=26.1 eV, Ta4f5C= 28.0 eV). The relatively strong signals for the two Ta- O binding states in the spectra are presumably due to surface oxidation of the tantalum nitride coatings [170, 171]. Consistent with the XRD pattern, the XPS spectra acquired from the original surface of two coatings confirm that the tantalum nitride phase in

Coating #1 presents as Ta2N, while Coating #2 contains mainly TaN.

Coating 2

Coating 1

Fig. 5.4 High resolution Ta4f XPS spectra obtained from the ion beam cleaned surface of as-deposited coatings.

The binding state for Ta-O for the two coatings can be studied from the Ta4f XPS spectra obtained from the argon ion beam cleaned surface (Fig. 5.4). It can be seen that the signals for the Ta4f peak representing the tantalum nitride phase in the two coating become stronger after ion beam cleaning. For Coating #1, the doublets for Ta in Ta-O-N

(Ta4f7B=24.3 eV, Ta4f5B=26.2 eV) and Ta in Ta2O5 (Ta4f7A=24.3 eV, Ta4f5A=26.2 eV) can be observed. Meanwhile, similar doublets are also found for Coating #2. It can be concluded from the high resolution XPS spectra that the oxygen in these tantalum nitride coatings exists in two distinct chemical states. The majority of oxygen was found in the form of Ta2O5, but a small amount of oxygen reacted with both tantalum and

77 nitrogen to generate a metastable Ta-O-N phase, which is an intermediate state between tantalum nitride and tantalum oxide [42].

5.3.3 SEM

a b

50μm 50μm SE BSE c d

50μm 50μm SE BSE

Fig. 5.5 Cross-sectional secondary electron and backscattered electron SEM images for the Ta-based coatings on Ti-6Al-4V substrates: (a) and (b) Coating #1; (c) and (d) Coating #2

The secondary electron and backscattered electron images for Coating #1 (Fig. 5.5 (a) and (b)) show that the as-deposited coating exhibits a dense and uniform cross-sectional structure without any visible structural flaws. The interface between coating and substrate is sharp and flat, it appears the coating is well bonded to the substrate. The thickness of Coating #1 is about 42 μm. The dark area at the coating-substrate interface is an artifact caused by metallographic polishing. Moreover, the cross-section for Coating #2 is also compact and uniform, with a uniform coating/substrate interface. The thickness of Coating #2 is about 36 μm.

a b

20 μm 20 μm

c d

20 μm 20 μm

Fig. 5.6 EDS elemental maps for Ti and Ta of the as-deposited coatings (a),(b) Coating #1;(c),(d) Coating #2.

The elemental distribution of Ta and the coating/substrate interfaces for the two coatings can be clearly observed from the EDS elemental maps for the Ta Lα and Ti Kα signals. It can be seen that the distribution of Ta in both Coating #1 and Coating #2 are uniform. Clear, sharp interfaces between the coatings and substrates can be observed for both Coating #1 and Coating #2.

5.3.4 TEM 5.3.4.1 Coating #1

a b

500 nm 500 nm

Fig. 5.7 (a) Bright-field TEM image taken from the center of Coating #1; (b) Bright-field TEM image taken from the coating-substrate interface of Coating #1; (c) SAED pattern taken from the center of Coating #1.

Fig. 5.7 shows bright-field TEM images of Coating #1 with corresponding selected area electron diffraction (SAED) taken from the center of the coating. According to the bright field image (Fig. 5.7 (a)), Coating #1 shows a compact structure containing slightly elongated coarse grains, which are typically greater than 200 nm in length. The micrograph for the coating/substrate interface shows the presence of a small number of pores with average size of about 100 nm at the coating/substrate interface. These pores were not observed in the SEM images, possibly due to either the limited resolution of the

SEM or that they exist in only low volume fractions. Both hexagonal Ta2N and orthorhombic Ta2O5 were identified from the SAED pattern acquired from the center of

80 the coating (Fig. 5.7 (c)), which is consistent with the XRD analysis. The discontinuous diffraction rings in the SAED pattern confirms that this coating is made up of relatively coarse grains.

5.3.4.2 Coating #2

a b

200 nm 200 nm

Fig. 5.8 (a) Bright-field TEM image taken from the center of Coating #2; (b) Bright-field TEM image taken from the coating-substrate interface of Coating #2; (c) SAED pattern taken from the center of Coating #2.

Bright field images of Coating #2 are shown in Fig. 5.8. Similar to Coating #1, Coating #2 is also comprised of elongated coarse grains with length of more than ~200 nm. The arrangement of grains is relatively compact without the presence of any nanopores present throughout the coating. However, a relatively rough and porous interface can be observed between the coating and the substrate (Fig. 5.8 (b)). The corresponding SAED from the coating center pattern indicates that Coating #2 exhibits a mixed phase structure

81 of fcc TaN and orthorhombic Ta2O5. The rings in the SAED pattern are not continuous consistent with the relatively coarse grain size of this coating.

5.3.5 Nanoindentation

Fig. 5.9 (a) Load-displacement curves of the as-deposited Coatings #1 and #2; (b) Hardness and elastic modulus of the as-deposited coatings; (c) The H/E and H3/E2 ratios derived from the hardness and elastic modulus of the as-deposited coatings.

Typical load-displacement (L-D) curves for Coatings #1 and #2 are shown in Fig. 5.9 (a). Indentation was performed with a maximum load of 8000 μm. Smooth curves were obtained from both the loading and unloading process for both coatings, suggesting that no cracking or other catastrophic events occurred during loading. Based on the L-D curves, the hardness (H) and elastic modulus (E) of the two coatings were calculated. As shown in Fig. 5.9 (b), Coating #1 exhibits a higher hardness and modulus. That is, the hardness for Coating #1 was measured to be 23.6 GPa ± 0.3 GPa and the elastic modulus of this coating was measured to be 307.2 GPa ± 2.08 GPa. For Coating #2 the hardness and elastic modulus decreased to 10.9 GPa ± 0.64 GPa and 211.1 GPa ± 6.6 GPa, respectively.

The H3/E2 and 1/E2H values for the two coatings are shown by the column graph in Fig. 5.9 (c). A higher H3/E2 (thus a higher H/E) value reveals the better ability of a material to resist the generation of deformation and abrasive damage, whilst a higher 1/E2H value reflects an increased difficulty to form cracks. As shown by this figure the H3/E2 values decrease from about 0.14 GPa to 0.03 GPa from Coating #1 to Coating #2 due to the

83 transition from the harder, less ductile hexagonal Ta2N phase to the softer fcc TaN phase. The 1/E2H values showed the reverse trend, the 1/E2H value of Coating #1 is 0.5E-6 GPa-3, lower than that of Coating #2, which is 2.0E-6 GPa-3.

5.3.6 Contact damage testing

a b # 1 # 0.4 μm 5.1 μm

c d # # 1 1 1.1 μm # # 6.8 μm

Fig. 5.10 Plan view and cross-sectional secondary electron FIB images of the residual indent impressions following a 500 mN spherical indentation on (a),(b) Coating #1; and (c),(d) Coating #2.

The contact surfaces are almost completely smooth for both coatings; no cracks were detected from the plan view images. Similarly, in the cross-sectional images no sub- surface cracks are visible. These featureless structures indicate that both coatings studied here exhibit a good damage tolerance. It can also be seen that the diameter and residual

84 indentation depth of the residual indent of Coating #1 (5.1 μm and 0.4 μm) are smaller than the equivalent features in Coating #2. That is the residual diameter and residual indentation depth for Coating #2 are 6.8 μm and 1.1 μm, respectively. This suggests that Coating #1 is harder than Coating #2 consistent with the mechanical data determined for hardness values.

5.4 Discussion

The research of Sundgren et al. [4] suggests oxygen can influence the structure and mechanical properties of hard coatings by substitution into the lattice. That is, occupying either interstitial positions or being incorporated into grain boundaries. Normally, the existence of tensile stresses and the substitution of atoms with smaller atomic radii will slightly decrease the lattice parameter, and thus lead to a shift in the diffraction peaks to higher 2 theta values. Conversely, a smaller 2 theta angle will be observed if there are residual compressive stresses or atoms with larger atomic radii substituting into the coating. By comparing the XRD patterns of the two coatings in this study with the XRD data for the tantalum nitride coatings deposited without the presence of oxygen describe in chapter 4, which exhibit same peak positions compared with the corresponding powder diffraction file data, it can be seen that the peak positions for hexagonal Ta2N in Coating

#1 and fcc TaN in Coating #2 are consistent with those of hexagonal Ta2N and fcc TaN deposited without oxygen. This suggests that oxygen was not present as either a substitutional or interstitial addition in Coatings #1 and #2. According to the XRD and

XPS results the oxygen was present in the form of both Ta2O5 and Ta-O-N phases in both coatings. A similar phenomenon was also found for tantalum oxynitride coatings produced by plasma enhanced magnetron sputtering [42] and reactive magnetron sputtering [173]. Dréo [173] ascribe this phenomenon to an amorphous structure or an extremely small grain size for the Ta-O-N phase, which would result in weak and broad diffraction peaks, therefore the peaks could then themselves be obscured by other strong signals from the other phases present or be obscured by background noise as an clear signal was not evident in the XRD spectra obtained. Ishihara[174] found that tantalum oxynitride prepared by RF magnetron sputtering exhibited an amorphous structure as the

85 substrate temperature increased to 500 ℃. Since discrete and clear reflections for Ta-O-N were not detected by either XRD or SAED, it is presumed that the Ta-O-N in the two tantalum oxynitride coatings studied here are present in the amorphous state.

In comparison to the two tantalum nitride coatings deposited without oxygen (described in Chapter 4), which exhibit a nanoporous structure with fine nanoscale equiaxed grains, the two tantalum-based coatings in this study exhibited a more compact structure with much coarser grains. The absence of nanopores and the larger grains is presumably due to the introduction of oxygen during the deposition process.

Prior research showed that Ta-O-N coatings deposited by reactive magnetron sputtering on to a Si substrate exhibited hardness values ranging from 7 to 14 GPa with elastic moduli ranging from 140 to 190 GPa [173]. Moreover, the hardness of a bulk Ta-O-N ceramic produced by hot pressing was approximately 16 GPa [175, 176]. Meanwhile, the hardness and elastic modulus values for Ta2O5 are typically about 10 GPa and 200 GPa, respectively [176-178]. Presumably, the presence of the softer Ta2O5 and Ta-O-N phases can significantly decrease the hardness of these tantalum nitride coatings [42]. Compared with the mechanical property values of the Ta2N coating described in Chapter 4, the hardness of Coating #1 (23.6 GPa ± 0.3 GPa) is about 30% less. Meanwhile for Coating #2, a ~ 54 % decrease in hardness can be obtained compared with the equivalent TaN coating studied in Chapter 4. However, similar decreases were not found for the elastic modulus values. The elastic modulus for Coating #1 was 307.2 GPa ± 2.1 GPa, that is much higher than that of the Ta2N coating (266±15 GPa) without the introduction of oxygen, and for the TaN coatings, the elastic modulus of Coating #2 is slightly higher than the value acquired from the equivalent TaN coating studied in Chapter 4. The higher elastic modulus for these coatings is possibly due to either the more compact structure or effects associated with the larger grain size. Since pores can be recognized, in effect, as a phase with an elastic modulus value of 0 [179], a higher elastic modulus would be expected with a denser structure [169]. Meanwhile, the elastic modulus of nanocrystalline materials is also directly related to the volume faction of the material bound up in grain boundaries due to the expected lower density and elastic modulus of the grain boundary

86 regions [180]. The larger grain sizes for both coatings results in a lower overall fraction of grain boundaries and hence a further increase in elastic modulus [1]. As a result of the low hardness and relatively high elastic modulus, the H3/E2 values of the tantalum oxynitrdie coatings are much lower (0.14 GPa and 0.03 GPa for Coating #1 and Coating

#2, respectively, compared with 0.69 GPa and 0.38 GPa for the Ta2N and TaN coatings studied in Chapter 4). This suggests a lower resistance for these tantalum oxynitride coatings to abrasive damage.

5.5 Conclusion

In summary, the presence of oxygen can influence the structure and mechanical properties of tantalum nitride coatings deposited by double cathode glow discharge technique by forming both orthorhombic Ta2O5 and a disordered Ta-O-N phase, in addition to Ta2N and/or TaN. Unlike the nanoporous structure for the tantalum nitride coatings described in Chapter 4, these tantalum oxynitride coatings present a dense structure, comprising coarse grains, without the presence of nanopores. The softer Ta2O5 and disordered Ta-O-N phases also lead to a decrease in the hardness of coating, however relatively higher elastic modulus values were obtained due to the much coarser grains and the absence of nanopores. This would be expected to lead to relatively lower wear resistance for these coatings. Therefore, in order to produce high performance tantalum nitride coatings, the presence of oxygen should be avoided during the deposition process.

Chapter 6

The formation of bone-like hydroxyapatite on a Ta2N coating

6.1 Introduction

Artificial implant materials are required to bond with living bone and tissues when they are implanted into the human body. Therefore, it is essential for the implant material to achieve effective fixation to bone and an exhibit a high affinity with surrounding tissues [181]. As noted in Chapter 4, the disadvantage of titanium alloys as bio-implant materials is their lack of bioactivity [182]. Therefore, tantalum nitride-based materials, which exhibit higher biocompatibility than Ti alloys [5, 182], are considered to have potential as coating materials to improve the bioactivity of titanium alloy artificial implants.

Hydroxyapatite (HAP, Ca10(PO4)6(OH)2), is a clinically successfully tested material, which can be used as fillers and spacers for bones [183]. Hydroxyapatite normally exhibits a dense and porous structure comprising a granular microstructure. The composition and properties of HAP are very similar to that of the apatite naturally present in bones. The essential requirement to generate effective adhesion between bone and bio- implants is that the apatite and collagen produced by osteoblasts can directly grow on the surface of bio-implants [184].

As HAP can be reproduced on the surface of bioactive materials in a simulated body fluid (SBF), it is often used as substitution for apatite in bones to assess the bioactivity of materials.

In this chapter, the vitro bioactivity of a hexagonal Ta2N coating was evaluated by observing the growth of HAP following exposure to simulated body fluid. A Ta2N coating was first deposited on a Ti-6Al-4V substrate by the double cathode glow discharge plasma technique. The as-deposited coating was then sectioned into two equal sections. The two samples were then immersed in SBF that contained the ions needed for the formation of HAP for two different times, 14 and 25 days, respectively. After immersion, the composition and microstructure of the HAP-like layers were studied in detail by a combination of Raman spectroscopy, XRD SEM and TEM.

6.2 Experimental method

The Ta2N coatings were deposited by the double cathode glow discharge technique onto a Ti-Al-V substrate following the procedure described in detail in Chapter 4. The deposition parameters are as follows: base pressure in the chamber, 5×10−4 Pa; working pressure 35 Pa with a constant Ar gas flow rate of 100 sccm; Ar:N2 flux ratios, 20:1; target bias voltage with direct current, -750 V; substrate bias voltage with impulse current, -300 V; substrate temperature, 700 ℃; parallel distance between the target and the substrate, 10 mm; and treatment time, 2 h.

The composition of the as-deposited Ta2N coating was analyzed by an XRD system

(PANalytical Empyrean) equipped with a Cu anode (λKα= 1.5406 nm). The X-ray signal was collected over 2 theta values ranging from 20° to 90°.

The as-deposited coating was then cut into two halves with a diamond saw. The two samples were soaked in SBF at 36.5℃ for 14 and 25 days, and were designated 14D and 25D, respectively. HAP layers are expected to be formed on the coating surface. The SBF solution, mimicking the composition of human blood, was prepared following the procedure described elsewhere by Kokubo [185]. The SBF solution is prepared by dissolving reagent NaCl, NaHCO3, KCl, K2HPO4·3H2O, MgCl2·6H2O, CaCl2 and

Na2SO4 in distilled water and buffered at a pH of 7.4 with tris-hydroxymethyl aminomethane ((CH2OH)3CNH2) and hydrochloric acid (HCl) at 36.5℃. To keep a stable chemical environment, the SBF solution was refreshed every two days. After soaking in SBF, the two coatings were gently cleaned by deionized water, and then dried in a vacuum desiccator.

To get a better understanding about the chemical composition of the hydroxyapatite growth on these Ta2N coatings, the Raman spectra for the HAP-like layers on the two coatings were acquired by a Raman Microscope (Renishaw inVia Raman Microscope) using 514 nm Argon ion laser (green) with 1800 l/mm grating. The Raman signals were collected over of Raman shift values from 300 cm-1 to 1100 cm-1. The thickness,

90 morphology and microstructure of the hydroxyapatite layers were characterized by SEM (Hitachi S3400) and TEM (Philips CM200, 200 kV) following the procedure described in Chapter 4.

6.3 Results

6.3.1 XRD analysis of the as-deposited Ta2N coating

Fig. 6.1 XRD pattern for as-deposited Ta2N coating.

It can be seen from Fig. 6.1 that the XRD pattern for the as-deposited Ta2N coating, prior to immersion in SBF, is consistent with the XRD data for hexagonal Ta2N (reference code: 00-026-0985). The diffraction peaks obtained at 33.8°, 38.2°, 50.6°, 60.5°, 67.8°, 72.4°, 74.0° and 82.6° can be indexed to the reflections from the (100), (101), (102),

(110), (103), (112), (201), (202) planes, respectively, of hexagonal Ta2N.

6.3.2 Raman spectroscopy

Fig. 6.2 The Raman spectra of the two HAP-like layers on samples 14D and 25D.

Raman spectroscopy was performed on samples 14D and 25D. The characteristic Raman signals for HAP arise mainly from the internal vibrations of the phosphate groups [129]. −1 The strongest characteristic peak for HAP, is the PO4ν1 peak at 961 cm . Apart from this −1 −1 −1 peak, two PO4ν2 peaks (at 447 cm and 433 cm ) four PO4ν3 peaks (at 1030 cm , 1046 −1 −1 −1 −1 −1 −1 cm , 1054 cm , and 1076 cm ) and four peaks PO4ν4 (at 582 cm , 594 cm , 610 cm and 620 cm−1) can also be frequently observed from the characteristic Raman spectrum for HAP.

Fig. 6.2 shows the Raman spectra obtained from the HAP-like layers on the surfaces of

14D and 25D. For the Ta2N coating soaked in SBF solution for 14 days, no clear Raman peaks were obtained from the coating surface suggesting HAP did not form on this sample, presumably due to insufficient soaking time. The HAP-like layer present on this

92 surface was possibly amorphous calcium phosphate (ACP), which has no clear Raman signature and often forms as an initial phase in the growth of HAP, [186, 187]. However, the Raman spectrum for the HAP-like layer on Coating 25D fits well with the signature −1 Raman spectrum for HAP. A very strong PO4ν1 peak was obtained at 962 cm . −1 −1 Furthermore, the two peaks for PO4ν2 can be observed at 430 cm and 442 cm . The −1 broad peak centered at 590 cm contains the three peaks from the vibrations of the PO4ν4 −1 −1 band. Finally, at 1042 cm and 1072 cm two peaks for PO4ν3 were also obtained. It can therefore be concluded that the Raman spectrum for the surface of Ta2N coating soaked in SBF solution for 25 days indicates the formation of a HAP layer. The small wave number shift as compared to the standard characteristic Raman signals for HAP were due to the heat induced by laser excitation [129].

6.3.3 SEM

a b

c d

Fig. 6.3 Plan view and cross-sectional SEM images for 14D ((a) & (b)) and 25D ((c) & (d)). (The dotted lines represent the thickness of the HAP-like layers grown on the two coatings)

Secondary electron images of the plan-view surface and cross-sectional morphology of the layers formed on the two Ta2N coatings are shown in Fig. 6.3. As can be seen from

Fig. 6.3 (b) and (d) the as-deposited Ta2N coating is about 20 μm in thickness. After 14 days exposure the sample 14D appears completely covered by an HAP-like layer with a thickness ranging from 1.1 μm to 2.7 μm (shown by the dotted line). The plan view image of the HAP-like layer on sample 14D shows a relatively uniform and smooth surface with only a few granular precipitates on the coating surface. Meanwhile, a relatively compact cross-section was obtained from the HAP-like layer (Fig. 6.3 (b)). For the Ta2N coating immersed in SBF solution for 25 days, the thickness of the HAP layer is between 4.6 μm and 8.4 μm (shown by the dotted line, Fig. 6.3 (d)). The HAP layer comprises a large number of HAP pillars. It can be seen that most of the HAP pillars have a broadly spherical cross-sectional morphology with a diameter of about 5 μm (Fig. 6.3 (c)). In some areas adjacent HAP growths have aggregated to form a clustered structure. Between the clusters, deep valleys can be observed. According to the cross-sectional image (Fig. 6.3 (d)), the HAP layer exhibits a rough cross-sectional morphology. Cracks and defects can be clearly observed from this image.

6.3.4 TEM

a b

c d

dark contrast

Fig. 6.4 (a) TEM BF images for 14D; (b) TEM image taken from the interface between HAP-like layer and substrate of 14D; (c) TEM image taken from the HAP-like layer of 14D; (d) higher resolution TEM image for the HAP-like layer of 14D (e) SAED pattern from the ACP-like layer; (f) EDS spectrum from the ACP- like layer.

Fig. 6.4 shows bright field TEM images which show the microstucture of the HAP-like layer on the Ta2N coating soaked for 14 days, together with the correspongding SAED pattern and EDS spectrum. The layer exhibits an average thickness of about 0.9 μm (Fig. 6.4 (a)). Expect for some growth defects, the layer exhibits a generally uniform, featureless structure (Fig. 6.4 (c)). No lattice fringes were observed even at higher magnification (Fig. 6.4 (d)). Moreover, some darker contrast can be observed from the layer in local regions. A sharp, flat and well bonded interface can be found between the

Ta2N coating and the growth layer, indicating good compatibility between these two layers. The SAED diffraction pattern aquired from the center of HAP-like layer (Fig. 6.4 (e)) shows only a diffuse halo that confirms the amorphous structure of this layer. According to the EDS spectrum from this layer (Fig. 6.4 (f)), this amorphous layer mainly contains O, Ca and P [188]. It can be concluded that the HAP-like layer is ACP. Elements such as Ta, Ga, Cu etc. were also obtained by EDS. These elements were introduced by either the TEM sample preparation process, such as gallium ions beam, redeposition, or the copper grid used to support the TEM sample [119, 120].

96 a b

c d

e f

Fig. 6.5 (a) TEM BF images for 25D; (b) TEM image taken from the ACP layer of 25D; (c) TEM image taken from the interface between the ACP layer and the HAP layer of 25D; (d) TEM image taken from the HAP layer of 25D (e) SAED pattern for ACP layer; (f) SAED pattern for HAP layer; (g) EDS spectrum for ACP layer; (h) EDS spectrum for HAP layer

A low magnification TEM image (Fig. 6.5 (a)) from the sample immersed in SBF for 25 days shows a two layered structure deposited on the Ta2N coating. The outer layer exhibits an amorphous structure, similar to that observed in the sample soaked for 14 days. The thickness of this layer is roughly 1.8 μm. With increasing depth from the outer surface, a darker contrast level can be observed, indicaing the emergance of crystalline

HAP. Therefore, it can be presumed the local regions of darker contrast in 14D are possibly evidence of nucleation of HAP crystals. A further increase in depth shows a transition in the microstructure from a featureless amorphous structure into an aggragtation of needle-shaped fine HAP grains with length of 30.5 ± 4.1 nm and width of 4.6 ± 0.5 nm. The thickness of this layer is about 7.1 μm. Growth defects, such as holes and cracks, can be found in both ACP layer and HAP layer. Meanwhile a flat and non-porous interface between HAP and Ta2N coating was evident in this sample.

Moreover, the HAP grains are tightly adhered to the Ta2N coating. According to the EDS spectra obtained from both the ACP and HAP layers (Fig. 6.5 (g) & (h)), the ratio of the intensity of Ca to the intensity of P increased between the ACP and HAP layers. This observation is consistent with the results of prior studies, where HAP exhibits a higher Ca/P ratio than ACP [183, 189]. The SAED pattern aquired from the ACP layer shows a dim ring consistent with the (300) plane of HAP, indicating the initial nicleation of HAP crystals within the ACP layer. Meanwhile, for the HAP layer (Fig. 6.5 (e)) the SAED pattern exhibits four clear diffraction rings for the (002), (211), (300) and (130) planes for the HAP phase. This confirms that the needle-shaped grains in this layer are HAP.

6.4 Discussion

The deposition parameters and mechanical properties of the Ta2N coating were previously described in Chapter 4. After soaking in SBF solution, a uniform ACP layer was initially formed on the Ta2N coating. Then, with increasing immersion time, the amorphous compound that formed initially gradually crystallized into a nanoscale needle- shape HAP grains. Growth defects, such as holes and cracks, can be found in both the

ACP layer present on the Ta2N coating after immersion for 14 days and the HAP layer on the Ta2N coating after immersion for 25 days. The ACP and HAP layers are well bonded to the coating material, indicating potential good bioactivity of the Ta2N coating.

The mechanism controlling the formation of HAP on tantalum nitride coatings has been previously studied by Xu et al. [189]. A schematic diagram showing this process is shown in Fig. 6.6. Based on Xu et al.’s theory, the formation procedure of HAP on Ta2N coating can be described as follows:

Firstly, the tantalum niride is hydrolyzed and forms a TaOxNy compound when initially immersed in SBF, the reaction is:

+ - Ta2N+H2O→TaOxNy+NH4 +OH (1)

The bridging oxygens of the TaOxNy provide negative charges to the coating surface. Because of Coulomb forces, the negatively charged surface will selectively combine with the calcium ions in the SBF. As more and more calcium ions accumulate on the surface of the Ta2N coating, the surface gradually becomes positively charged. As a result, the positively charged surface starts to combine with negatively charged phosphate ions in SBF solution. Since SBF, as well as human blood, is supersaturated with both calcium and phosphate, the absorption for calcium and phosphate will continuously alternate, once the coating surface is negatively charged, to form an ACP layer [182, 183]. As ACP is a metastable phase, over time, the ACP will then solutionize and re-nucleate as the stable bone-like crystalline HAP phase [190]. From the sample soaked for 25 days, the

100 transition layer from ACP to HAP can be clearly observed, demonstrating the ongoing crystallization process of ACP with increasing soaking period.

Fig. 6.6 Schematic of HAP nucleation and growth on the Ta-N coating soaked in SBF.[189]

Since the formation of ACP and the nucleation and growth of HAP is spontaneous in SBF. The efficiency of the apatite nucleation is governed by the concentration and the formation speed of TaOxNy [182, 183]. It can be seen from both the SEM and TEM images that the width for the HAP and ACP layers formed on the Ta2N coating soaked for 25 days is much thicker than the ACP layer on the sample immersed for 14 days. The maximum thickness of the HAP and ACP layers formed after 25 days is about 9.7 μm, ~7.0 μm thicker than that of the ACP layer formed in 14 days, indicating an accelerating speed of formation for ACP and HAP with increasing soaking period. This is possibly due to surface oxidation of Ta2N as tantalum nitride can be readily oxidized in air [141]. The slow hydration reaction of the passive tantalum nitride layer may inhibit the formation of TaOxNy, greatly slowing the speed of growth of the ACP layer in its early stages of formation [182].

101

6.5 Conclusion

In this chapter, the biocompatibility of a Ta2N coating deposited by double cathode glow discharge deposition was studied. After immersion in SBF, an ACP layer was initially formed on the Ta2N coating. With increasing immersion time HAP nucleated from the

ACP. The Ta2N coating showed good bioactivity. The Ta2N coating was fully covered by either ACP or HAP layer after immersion. The interface between Ta2N coating and HAP was uniform without any visible defects, suggesting the HAP layer is tightly bonded to the Ta2N coating. In Chapter 4, the excellent mechanical properties of hexagonal Ta2N were described. It seems that the application of Ta2N coatings to Ti alloys has high potential for use as bone implant materials.

102

Chapter 7

The microstructure and mechanical properties of zirconium nitride coatings deposited by a double cathode glow discharge plasma

technique

103

7.1 Introduction

Zirconium nitride coatings have already been widely used as a replacement for titanium nitride coatings to protect machining parts on substrates such as high speed steels [191]. Furthermore, ZrN also has potential for application as a biomaterial [41]. Recently, zirconium nitride coatings have attracted considerable interest as protective coatings that exhibit both low electrical resistivity and high corrosion resistance [192]. Prior research shows ZrN coatings can be deposited on metallic bipolar plates as a coating material to extend the service life and performance of these components [193]. As a material which is expected to be applied for use in various harsh working conditions, it is important to understand the coating microstructure and mechanical properties of zirconium nitride coatings prepared under a range of deposition conditions, such that structure-property relationships can be optimized.

In this chapter, two materials that are frequently used as metallic bipolar plates for fuel cells, stainless steel (SS), and Ti-6Al-4V (TC4) were used as the substrate materials. Zirconium nitride coatings were deposited on both substrates using a double cathode glow discharge technique. The phase, cross-sectional morphology, microstructure and mechanical properties of these coatings were assessed in this study.

104

7.2 Experimental methods

316L stainless steel and Ti-6Al-4V disks, with a diameter of 40 mm and a thickness of 3 mm, were used as substrates. A 99.99% purity Zr disk, which was 100 mm in diameter and 5 mm thick, was used as the target in the deposition processes. Before deposition, the substrates and target were polished consecutively with silicon carbide papers, and then cleaned with pure alcohol and distilled water in an ultrasonic bath. The zirconium coatings were deposited onto the polished substrates using double cathode glow discharge apparatus at two different substrate temperatures, 650 ℃ and 700 ℃. The other deposition parameters are shown in Table 7.1. The four as-deposited coatings were designed Q1, Q2 (both on stainless steel) Q3 and Q4 (both on Ti-6Al-4V).

Table 7.1 The deposition conditions of zirconium nitride coatings Q1 Q2 Q3 Q4

Atmosphere Ar:N2 =10:1 Ar:N2 =10:1 Ar:N2 =10:1 Ar:N2 =10:1

Working pressure 35 Pa 35 Pa 35 Pa 35 Pa

Target bias voltage -800 V -800 V -800 V -800 V

Target bias voltage -350 V -350 V -350 V -350 V

Substrate temperature 700 ℃ 650 ℃ 700 ℃ 650 ℃

Target/substrate distance 10 mm 10 mm 10 mm 10 mm

Treatment time 2h 2h 2h 2h

Substrate SS SS Ti-Al-V Ti-Al-V

The XRD patterns of the as deposited coatings were collected over of 2 theta values ranging from 15° to 80° using an X-ray diffraction (XRD) system (PANalytical Empyrean) equipped with a Cu anode using Cu Kα (λ= 1.5406 nm) radiation.

105

The cross-sectional morphology of as-deposited coatings was examined by scanning electron microscopy (SEM, Hitachi S3400). A transmission electron microscope (TEM, Philips CM200) operating at 200 kV was used to observe the microstructure of the four as-deposited coatings. TEM specimens were prepared using FIB-prepared samples using the lift-out method.

The mechanical properties of the zirconium nitride coatings were evaluated by a Hysitron Triboindenter workstation equipped with a Berkovich diamond indenter. The nanoindentation testing was operated at a maximum load of 8000 μN with a loading rate of 800 μN/s and holding time of 5s at maximum load. 10 indents were performed on each coating. The maximum contact depth for each indent was carefully controlled to be less than 10% of the coating thickness.

The contact damage testing was performed by a UMIS workstation equipped with a spherical diamond indenter with a radius of 5 μm. The maximum load was 500 mN. Using a FIB microscope, the cross-sectional structures of the indents were characterized. Prior to the FIB analysis, all the samples were sputter coated with a thin layer of gold to protect the near surface features and also to minimize charging effects.

106

7.3 Results 7.3.1 XRD

Fig. 7.1 X-ray diffraction patterns of the as-deposited zirconium nitride coatings.

The X-ray diffraction spectra collected from 2θ values ranging from 15° to 80° for the four zirconium nitride coatings, Q1, Q2, Q3, and Q4 are shown in Fig. 7.1. It can be seen that the spectra for all four coatings comprise reflections from mainly fcc ZrN (JCPDS

NO. 35-0753). In addition, relatively weak diffraction peaks for monoclinic ZrO2 (JCPDS NO. 01-0750), can also be observed from all these zirconium nitride coatings. For coatings Q1 and Q3, the most intense peak for fcc ZrN is from the (200) plane. However, for coatings Q2 and Q4, deposited at a lower substrate temperature, the reflections from the (111) plane are relatively stronger than those of the (200) plane. To evaluate the

107 orientation preference of the as-deposited zirconium nitride coatings, the texture coefficient (TC) of the fcc ZrN were calculated though the following equation [194]:

퐼푚(ℎ푘푙)⁄퐼0(ℎ푘푙) 푇퐶ℎ푘푙 = 1 (3) ∑푛 퐼 (ℎ푘푙)⁄퐼 (ℎ푘푙) 푛 1 푚 0

where Im(hkl) is the measured intensity for (hkl) plane, I0(hkl) is the relative intensity for (hkl) plane according to the powder diffraction data, n is the number of reflection planes.

A TC(hkl) value greater than 1 means a preference to grow along the (hkl) orientation, meanwhile a TC(hkl) value close to 1 signifies a more random orientation, and a TC(hkl) value below 1 shows a lack of orientation for such a plane. The TC value for fcc ZrN phase in all the four coatings are list in Table 7.2.

Table 7.2 The texture coefficients of different planes for the as-deposited zirconium nitride coatings.

Plane (111) (200) (220) (311) (222)

Sample

Q1 1.62 0.93 0.81 0.70 0.96

Q2 1.65 0.16 1.76 0.73 0.71

Q3 1.64 0.91 0.96 0.74 0.81

Q4 1.61 0.14 1.38 0.90 0.88

From Table 7.2, it is clear that the fcc ZrN phase in all the three coatings show a strong preference for the (111) orientation. However, for Coating Q2 and Coating Q4, a strong preference for fcc ZrN in (220) can also be observed. For zirconium nitride coatings on stainless steel, the TC value of fcc ZrN (220) increases from 0.81 for Coating Q1, to 1.76 for Coating Q2. Similarly, for zirconium nitride coatings on Ti-Al-V, the TC value of fcc ZrN (220) is 1.38 for Coating Q4, about 0.4 higher than that of Coating Q3. Meanwhile, it is also noteworthy that a reduced preference for (200) planes can also be found for

108

Coatings Q2 and Q4, (0.16 and 0.14 respectively), compared with Q1 and Q3, 0.93 and 0.91, respectively.

7.3.2 SEM 7.3.2.1 Coating Q1

20 μm 20 μm SE BSE

Fig. 7.2 Cross-sectional SEM micrographs of Coating Q1 with corresponding EDS elemental maps

The cross-sectional secondary and backscattered SEM images for Coating Q1 are shown in Fig. 7.2. Coating Q1 is ~19.3 μm thick. The cross-section of coating Q1 is compact and uniform, no cracks or structural flaws can be found throughout the coating. The coating/substrate interface for Coating Q1 is reasonably sharp, but a few pinholes can be observed at the interface. It appears that the coating is well bonded to the SS substrate. From the EDS elemental maps, it can be seen that the distribution of Zr and Fe in the coating and substrate, respectively, appears uniform. Further, the transition from Zr to Fe

109 is relatively sharp, according to these elemental maps, suggesting a sharply defined coating-substrate interface.

7.3.2.2 Coating Q2

10 μm 10 μm SE BSE

Fig. 7.3 Cross-sectional SEM micrographs of Coating Q2 with corresponding EDS elemental maps.

The secondary and backscattered electron images for Coating Q2 show a coating that is ~11.3 μm thick. This coating exhibits a dense and continuous structure without any obvious defects present. The coating and substrate are tightly adhered, no visible pores were found at the coating/substrate interface. According to the EDS elemental maps (Fig. 7.3 (c) and (b)), the distribution of Zr is uniform in the coating. The EDS elemental maps for Zr and Fe depict a reasonably sharp interface between the coating and the substrate.

110

7.3.2.3 Coating Q3

20 μm 20 μm SE BSE

Fig. 7.4 Cross-sectional SEM micrographs of Coating Q3 with corresponding EDS elemental maps.

It can be seen from Fig. 7.4 (a) and (b) that the thickness of Coating Q3 is ~16.7 μm. A uniform and compact cross-section without cavities or defects can be observed from the SEM images. The coating appears to be well bonded to the substrate with a sharp and continuous interface. According to the EDS elemental maps (Fig. 7.4 (c) and (d)) it appears that the coating-substrate interface between is relatively sharp, with negligible interdiffusion evident.

111

7.3.2.4 Coating Q4

10 μm 10 μm SE BSE

Fig. 7.5 Cross-sectional SEM micrographs of Coating Q4 with corresponding EDS elemental maps.

According to the secondary and backscattered images (Fig. 7.5 (a) and (b)) from coating Q4, a featureless coating with a thickness of ~7.3 μm is present. The substrate/coating interface of Coating Q4 appears continuous and sharp. No structural defects can be found associated with either the coating or at the interface. Corresponding EDS elemental maps are shown in Fig. 7.5 (c) and (d). The relatively distinct transition from Zr to Ti in the elemental maps suggests a sharp coating/substrate interface.

112

7.3.3 TEM 7.3.3.1 Coating Q1

a b

100 nm 100 nm

Fig. 7.6 (a) Cross-sectional TEM bright-field micrograph acquired from the centre of Coating Q1; (b) Cross-sectional TEM bright-field micrograph acquired from the coating/substrate interface of Coating Q1; (c) SAED pattern of Coating Q1 acquired from the centre of the coating.

Fig. 7.6 (a) and (b) show bright field TEM images of the microstructure of Coating Q1 and the coating/substrate interface. According to these images, Coating Q1 exhibits a dense microstructure. Fine equiaxed grains, ~ 30 nm in diameter, are embedded in a

113 matrix of elongated coarser grains with a width of less than 100 nm. Nanopores were not commonly observed from this coating. The coating/substrate interface is straight and the coating is generally well-bonded to the substrate, with only a small number of pinholes observed at the interface. The SAED pattern contains the reflections for both fcc ZrN. and monoclinic ZrO2 It can be seen that the diffraction pattern of fcc ZrN is made up of discrete segments, rather than continuous rings, consistent with the presence of coarser ZrN grains in the coating. The EDS elemental maps (Fig. 7.7) from this coating confirm that Coating Q1 exhibits a uniform distribution of Zr and a sharp transition from Zr to Fe can be found at the interface.

Zr Fe

Fig. 7.7 STEM micrograph and EDS elemental maps for Coating Q1.

114

7.3.3.2 Coating Q2

a b

100 nm 500 nm

Fig. 7.8 (a) Cross-sectional TEM bright-field micrograph acquired from the centre of Coating Q2; (b) Cross-sectional TEM bright-field micrograph acquired from the coating/substrate interface of Coating Q2; (c) SAED pattern acquired from the centre of Coating Q2.

Similar to Coating Q1, Coating Q2 also shows a microstructure comprised of both fine equiaxed fine grains, with an average grain size of 30 nm, together with coarser elongated grains with a width of approximately 100 nm. Nanopores can also be observed located at triple-junctions. Large, elongated cracks with a length of about 500 nm can be found near

115 the coating-substrate interface, which suggests relatively poor adhesion between the coating and the substrate. The SAED pattern from Coating Q2 contains the discontinuous diffraction segments of fcc ZrN. However, weak reflections from monoclinic ZrO2 can also be observed in this pattern. The corresponding EDS maps for the interface show a rough interface between coating and substrate. Notably, STEM images and elemental EDS maps shows that the interface between the coating and the substrate is not flat and perturbations up to several hundred nm in width can be observed. However, the elemental EDS maps suggest no evidence of inter-diffusion. The lack of diffusion suggests these perturbations arise from the preparation of the substrate prior to deposition.

Zr Fe

Fig. 7.9 STEM micrograph and EDS elemental EDS maps for Coating Q2.

116

7.3.3.3 Coating Q3

a b

200 nm 200 nm

Fig. 7.10 (a) Cross-sectional TEM bright-field micrograph acquired from the centre of Coating Q3; (b) Cross-sectional TEM bright-field micrograph acquired from the coating/substrate interface of Coating Q3; (c) SAED pattern acquired from the centre of Coating Q3.

Compared with the zirconium nitride coatings deposited on stainless steel substrates, coating Q3, which was deposited on a Ti-Al-V substrate, exhibits a structure consisting of equiaxed grains with a more uniform size distribution. The average grain size is ~35 nm. The grain boundaries can be clearly observed from the BF images. Nanopores, a few nanometers in diameter, can also be observed homogeneously distributed. As shown by

117 the Fig. 7.10 (b), the bonding between substrate and coating is relatively strong, but a small number of pinholes can be seen at the interface. Different from Coating Q1 and Q2, the SAED pattern from Coating Q3 shows more continuous diffraction rings for fcc ZrN, suggesting that Coating Q3 has a finer, more uniform grain size. Meanwhile, diffraction rings of monoclinic ZrO2 were also present in the SAED pattern. From the EDS elemental maps, a uniform distribution of Zr and a relatively clear interface between coating and substrate were obtained.

Ti Zr

Fig. 7.11 STEM micrograph and EDS elemental maps for Coating Q3.

118

7.3.3.4 Coating Q4

a b

100 nm 100 nm

Fig. 7.12 (a) Cross-sectional TEM bright-field micrograph acquired from the center of Coating Q4; (b) Cross-sectional TEM bright-field micrograph acquired from the coating/substrate interface of Coating Q4; (c) SAED pattern acquired from the center of Coating Q4.

The structure of Coating Q4 is generally similar to that of Coating Q3. Equiaxed grains, ~20 nm in diameter, appear to have agglomerated into a porous structure. Small nanopores, with an average size of 5 nm, can be observed throughout the coating. A larger fraction of nanoporosity can be observed for Coating Q4 compared with Coating

119

Q3. However, closer to the coating-substrate interface, the zirconium nitride coating shows a compact structure, and almost no nanopores can be seen from this region. Meanwhile, a slight decrease in grain size can also be found close to this interface. The coating/substrate interface is straight and the coating appears well bonded to the substrate. The sharp interface can also be observed from the relevant EDS elemental maps. The distribution of Zr is uniform throughout the coating. Moreover, the SAED of Coating Q4 is similar to that of Coating Q3. That is, indicates the presence of ZrN and monoclinic

ZrO2.

Ti Zr

Fig. 7.13 STEM micrograph and EDS elemental maps for Coating Q4.

120

7.3.4 Nanoindentation testing

Fig. 7.14 Load-displacement curves of the as-deposited zirconium nitride coatings under nanoindentation.

The mechanical properties of the zirconium nitride coatings coatings, including hardness and elastic modulus, were obtained using a Hysitron Triboindenter workstation equipped with a Berkovich diamond indenter. During nanoindentation, a maximum load of 8000μN was applied to the surface of each coating at a loading rate of 800 μN/s. Fig. 7.14 shows the representative load-displacement (P-H) curves for all the four coatings. The P-H curve for each coating is smooth for both the loading and unloading cycles. Based on these P-H curves the hardness (H) and elastic modulus (E) values for each coating were calculated using the Oliver-Pharr method.

121

Fig. 7.15 Hardness and elastic modulus of as-deposited zirconium nitride coatings

As shown in Fig.7.15, all the four coatings show hardness values above 10 GPa. The hardness of Coating Q1 was determined to be 17.6 ± 3.3 GPa, which is higher than the other three samples. Slightly lower than Coating Q1, the hardness of Coatings Q2 was measured to be 14.1 ± 1.7 GPa. For Coating Q3 and Coating Q4, the measured hardness values were 14.4 ±2.0 GPa and 12.1 ± 1.1 GPa, respectively. The corresponding elastic modulus for Coating Q1 was also the highest, which was 254 ± 36 GPa. For Coating Q2, this value decreased slightly to 246 ± 23 GPa. The elastic moduli for Q3 and Q4 on the titanium substrates were 217 ± 17 GPa and 212 ± 25 GPa, respectively, much lower than the two coatings on stainless steel substrates.

122

Fig. 7.16 The H/E and H3/E2 ratios derived from the hardness and elastic modulus of as-deposited zirconium nitride coatings.

The H3/E2 and 1/E2H values for the four zirconium nitride coatings were derived from the H and E values (Fig. 7.16). As shown in this figure, the H3/E2 value for Coating Q1 is 0.085 ± 0.021 GPa, higher than that of Coating Q2, 0.046 ± 0.021 GPa, indicating a better resistance for Coating Q1 to wear damage. However, Coating Q2 exhibited a better tolerance to the formation of cracks, since the 1/E2H value for Coating 2 is 1.2 E-6 ± 0.6 E-6 GPa-3, higher than that of Coating Q1, 0.9E-6 ± 0.6E-7. The H3/E2 and 1/E2H values for the two zirconium nitride coatings coatings deposited on Ti-Al-V coatings show similar trends. A higher H3/E2 value was obtained from Coating Q3, (0.064 ± 0.024 GPa) than Coating Q4 (0.039 ± 0.006 GPa), but Coating Q4 exhibits a greater 1/E2H, (1.8E-6 ± 1.0E-6 GPa-3) than Coating Q3 (1.5E-6 ± 0.5E-6 GPa-3).

123

7.3.5 Contact damage test

a b

0.51 μm

5.3 μm

c d

0.70 μm

6.1 μm

Fig. 7.17 Plan view and cross-sectional FIB micrographs 500 mN spherical indents for (a) (b) Coating Q1; (c) (d) Coating Q2; (e) (f) Coating Q3; (g) (h) Coating Q4.

124

e f

0.73 μm 5.7 μm

g h

6.4 μm

Fig. 7.17 Plan view and cross-sectional FIB micrographs 500 mN spherical indents for (a) (b) Coating Q1; (c) (d) Coating Q2; (e) (f) Coating Q3; (g) (h) Coating Q4.

A 500 mN load was applied to the surface of each coating using a spherical indenter using a UMIS nanoindentor to evaluate contact damage resistance. Fig. 7.17 shows the plan view and cross-sectional FIB images of the residual indent impressions for each coating. All four coatings exhibit good damage tolerance, the contact surface is smooth for each coating. Featureless and crack-free sections were obtained from all cross- sectional images. By measuring the radius and the depth of residual indent, it can be seen that these observations are consistent with the hardness values, Coating Q1 exhibits a smaller indent radius and shallower indentation depth than Coating Q2 (5.3 μm in radius and 0.51 μm in depth for Coating Q1, and 6.1 μm and 0.70 μm for Coating Q2). For the

125 zirconium nitride coatings coatings on Ti-Al-V substrates, similar results were obtained (5.7 μm and 0.73 μm for Coating Q3, 6.4 μm and 0.79 μm for Coating Q4). All the four coatings exhibited good damage resistance without any cracks or other structural failures observed.

7.4 Discussion

These as-deposited coatings mainly consisted of a fcc ZrN phase, together with a low concentration of monoclinic ZrO2. Since the substrate deposition temperature is much lower than 1200 ℃, it is not surprising that the ZrO2 phase presents with a monoclinic structure (the phase diagram of zirconium oxide is shown in Fig. 7.18) [195]. The presence of ZrO2 is presumably due to either a small fraction of residual air in the deposition chamber or an air leakage during operation [196].

Fig. 7.18 The phase diagram of zirconium oxide. [195]

For both the zirconium nitride coatings deposited on stainless steel and Ti-Al-V substrates, the fcc ZrN phase showed a strong crystallographic preference for the (111)

126 plane. This is due to a tendency of coatings to evolve a texture with the lowest overall energy, including both strain and surface energy [156]. The strain energy is mainly generated from the compressive residual stresses caused by the elastic deformation during the deposition of the coating, therefore the strain energy produced is proportional to the coating thickness [197]. However, the surface energy does not vary with the thickness of the coating. Because of the relatively large thickness for all four coatings, the overall energy is dominated by strain energy. To lower the overall energy, a texture with a lower strain energy was evolved, favoring the preference for the (111) planes, which have the lowest strain energy [198]. A similar preference for (111) preferred orientation for a ZrN coating was also reported by Liu et al. [40]. In their research a group of ZrN coatings were deposited by reactive DC magnetron deposition with a DC power of 150 W on Si (100) substrate at a substrate temperature of 50 ℃. An increasing preference in (111) plane was observed since the coating thickness increased from 35 nm to 925 nm.

Meanwhile, it is noteworthy that the intensity of fcc ZrN (200) peaks increase significantly with increasing substrate temperature. Further, more intense (220) peaks were found from the ZrN-based coatings deposited at lower temperatures. This is because with increasing deposition temperature, a higher mobility is acquired by the atoms due to their increased thermal energy, leading to a further increase in total surface energy [40]. Since the lowest surface energy is correlated to the (200) plane of ZrN [199], to lower the surface energy, increased preference for the (200) plane occurs in coatings Q1 and Q3, compared with the two coatings deposited at a lower substrate temperature. The increasing intensity in (200) for fcc ZrN with increasing deposited temperature have also been observed in several other studies [40, 197, 200, 201]. In summary, the texture of the coatings is determined by both the thickness of the coating and the deposition temperature.

Coatings deposited with the double glow discharge plasma technique normally exhibit a microstructure comprising fine grains with a grain size typically of tens nanometers in diameter (as shown in Chapters 4 and 5) [193]. However, the TEM images recorded from the ZrN coatings on stainless steel substrates exhibit a bimodal microstructure, which

127 exhibit both fine and elongated coarse grains. This distinct difference in the microstructure of the ZrN coatings on stainless steel substrates is possibly correlated to the structure of the substrate. The stainless steel is based on an austenitic (fcc) alloy, the same structure as the fcc ZrN, and this may promote local nucleation of fcc ZrN, that leads to the formation of coarser grains [202]. It can be seen from the TEM micrographs that at the coating-substrate interface of coatings Q1 and Q2 that coarse grains, with a grain size above 100 nm, can be found. However, for the ZrN coatings deposited on Ti- Al-V, which primarily consist of hcp Ti grains, this bimodal microstructure was not observed. Clearly the grain size at the coating-substrate interface for Coating Q3 and coating Q4 is fine and consistent. This effect from the substrate material suggests why coatings Q3 and Q4 exhibit a finer and more uniform grain size than coatings Q1 and Q2. Similar phenomenon was also reported by Hibbs et al. for fcc TiN coatings formed on two kind of steel substrates, one containing fcc carbide particles and the other without such particles [203]. The fcc TiN coating on the substrate with fcc carbide particles formed a bimodal structure with both small and large grains. While for the other coating a more uniform grain size was observed.

Nanoidentation results show that the hardness of these zirconium nitride-based coatings varied from approximately 12 GPa to 18 GPa. An increase in hardness with increasing substrate temperature can be observed by comparing the hardness of coatings Q2 and Q4 with coatings Q1 and Q3. There are two possible explanations. First, the higher substrate temperature increases the kinetic energy of the deposited atoms; this leads to a greater ability of atoms to move on the surface that further promotes the nucleation and growth of the deposited grains [197]. Consequently, this results in a denser structure [40]. It can be seen from the TEM bright field images that coatings Q1 and Q3 exhibit a more compact structure than coating Q2 and with a lower density of nanopores. This dense structure partly increases the hardness of these two coatings, since the hardness of nanocrystalline coatings normally shows a decreasing relationship with increasing fraction of nanopores [164, 204]. Meanwhile, the stronger preference for the (200) preferred orientation of coatings Q1 and Q3 may also lead to an increase in hardness due to texture strengthening effects [205, 206]. Similar increases in hardness have been

128 correlated to the stronger (200) texture in other studies of ZrN coatings [207-209]. For example, in Auger et al.’s research [208], the hardness of ZrN nitride coating with (200) preferred orientation is about 4 GPa harder than the coating with (111) preferred orientation. It is also noteworthy that the ZrN coatings on stainless steel substrates, Q1 and Q2, exhibit a slightly higher hardness than that of the ZrN coatings on Ti-Al-V substrates, Q3 and Q4. This is possibly due to the combination of both the strengthening from the reduced grain size with the strain hardening provided by dislocation activity in the larger grains due to the bimodal structure [1].

The mechanical properties of ZrN coating have been widely studied. The reported hardness values for ZrN produced by magnetron sputtering, ion-beam assisted deposition and other deposition methods typically range from about 10 GPa to 27 GPa [198, 208, 210]. However, the ZrN coating in this study does not show a very high hardness compared with some of the ZrN coatings deposited by other techniques. In addition to the nanoporous structure of these coatings, it is possible that the existence a small amount of monoclinic ZrO2 phase may slightly soften these coatings. This is because monoclinic

ZrO2 is a softer phase compared with fcc ZrN [42, 211]. However, similar to the tantalum nitride coatings detailed in Chapter 4. The ZrN coatings coating deposited by double cathode glow discharge technique exhibit a lower elastic modulus compared with ZrN coatings with comparable hardness values reported in the literature, i.e. 260-460 GPa [201]. For example, Auger et al. studied a group of ZrN coatings deposited on silicon substrates by RF magnetron sputtering. In their study, for a ZrN coating with a hardness of 11.9 GPa, similar to that of coating Q4 in this research (12.1 GPa), the measured elastic modulus was 257 GPa, about ~20% higher than the elastic modulus of coating Q4. Further, an elastic modulus of 290 GPa was achieved for a ZrN coating with a hardness of 18.7 GPa by Auger et al., also ~20% higher than that measured for coating Q1 with an comparable hardness of 17.6 GPa. The relative decrease in elastic modulus is especially notable for the two ZrN coatings deposited on Ti-Al-V substrates, which have a larger volume fraction of nanopores compared with the ZrN coatings on stainless steel substrates. The lower elastic modulus of ZrN coatings in this study is presumably related to the presence of nanopores and the high volume fraction of grain boundaries associated

129 with the very small grain size. Since the elastic modulus of grain boundary is much lower than that of the grain interior due to lower atomic packing [180], a significant decrease in elastic modulus of coating is expected with increasing volume fraction of grain boundaries. The benefits brought by the existence of nanoporosity in nanocyrstalline are coatings concluded in detail in Chapter 8, and are not repeated here. This decrease in elastic modulus correlates with the deposition method associated with higher H3/E2 and 1/E2H values for the zirconium nitride coatings in this study, thus greatly improved the damage resistance and wear resistance for these ZrN coatings would be expected. It can be seen that none of these four coatings exhibited cracking or other structural failures during the contact damage testing. Therefore, with the higher wear resistance and damage resistance, ZrN coatings prepared by double cathode glow discharge technique can offer a good protection to tools under abrasive wear conditions.

7.5 Conclusion

The microstructure and mechanical properties of zirconium nitride coatings deposited on both stainless steel and Ti-Al-V substrates at different substrate temperatures by double glow discharge plasma technique were studied. All the zirconium nitride coatings exhibited an fcc ZrN phase with strong (111) preferred orientation. An increase in the intensity of the (200) texture was found with increasing substrate temperature. An increase in the substrate temperature led to an increased hardness of the coatings. The zirconium nitride coatings on the stainless steel substrates exhibited a bimodal microstructure which both fine grains and more elongated coarse grains. This is thought to be influenced by the fcc structure of the stainless steel substrate. In contrast, the two coatings on the Ti-Al-V substrates show a uniform microstructure comprising fine equiaxed grains and together with a number of nanopores. The four coatings in this study show comparable hardness and much lower elastic modulus compared with the zirconium nitride coating deposited by other deposition techniques. The decrease in measured elastic modulus, relative to coating hardness, increases the H3/E2 and 1/E2H values of these coatings, which suggests further improvements in the wear resistance and fracture damage resistance of these coatings.

130

Chapter 8

The porosity-dependence of mechanical properties and damage mechanisms in nanoporous

tantalum nitride coating

131

8.1 Introduction

Recent theories about contact and wear mechanisms show that hardness (H) and elastic modulus (E) play equally important roles in defining mechanical properties such as wear resistance and fracture toughness [69]. The values of H/E and H3/E2 have been widely accepted as reliable indicators to evaluate the durability of a material when exposed to abrasive damage and plastic deformation [152]. Meanwhile, the threshold cracking load,

Pc (the minimum load required to cause cracking) is correlated with hardness and elastic modulus by the relation [71]:

1 푃푐 ∝ ⁄퐸2퐻 (2) thus, the value of 1/E2H is frequently used as a ranking parameter in describing the ability of a material to resist the formation of cracks. These key parameters suggest that a material with the high hardness, but lower elastic modulus, possesses a higher resistance to plastic deformation and more damage tolerance.

It is widely accepted for bulk materials that the hardness of a material normally increases with increasing elastic modulus [56, 212]. However, Musil et al.’s recent study [61] showed that the elastic modulus of a nanocrystalline material is not directly correlated to its hardness. Instead, it is strongly influenced by non-equilibrium grain-boundary structures and the presence of defects, such as porosity [213]. That means, for a given composition, it is possible to generate coatings with different sets of values of hardness and elastic modulus. Thus, the mechanical behavior can be tuned by adjusting the hardness and elastic modulus values of the coating material. Since the volume fracture of nanoporosity has a significant influence on the elastic modulus of nanocrystalline phases [169], we can hypothesize that by decreasing the elastic modulus, through the presence of nanopores, without significantly compromising its hardness, higher values of H/E, H3/E3 and 1/E2H can be achieved, therefore the tolerance of the coating to wear damage, plastic deformation and cracking can be significantly improved.

132

Conventionally, the existence of pores, which can behave as stress concentrators and local regions of weakness, is thought to be detrimental to the mechanical behavior of a material [214, 215]. Several models have been developed to describe the decease of fracture toughness with increasing porosity [215]. Thus, conventionally, effort is made to produce full-dense structures in order to avoid the degradation of mechanical properties caused by high volume fractions of porosity [168]. However, a recent study by Reddy et al. [168] suggested that pores with a diameter of only a few nanometers will not lead to catastrophic structural failure of materials under external load, since they are much smaller than the critical size needed to trigger the initiation of nanocracks. Instead, the existence of nanoporosity will provide a volume of material that will relieve the stresses associated with grain boundary sliding caused by deformation. In their study, the measured toughness of nanoporous B4C, with a grain size between 40 nm to 150 nm, showed a 75% increase compared with full-dense microcrystalline B4C. However, the effect of nanopores on the hardness and elastic modulus of nanocrystalline materials has not been well understanded. Moreover, in order to explore the influence of nanopores on the mechanical properties of nanocrystalline material, more research is required on investigating full-dense and nanoporous materials with a similar grain size.

In this study, two Ta2N coatings, one with a nanoporous structure and the other with a full-dense structure, but with similar grain sizes, were deposited using double cathode glow discharge. Nanoindentation and contact damage testing were performed on both coatings to analyze the influence of nanoporosity on the mechanical properties of these nanocrystalline coatings. Furthermore, by a combination of microindentation with transmission electron microscopy (TEM) and focused ion beam (FIB) microscopy, the mechanical behavior of these nanoporous Ta2N coatings under external load was investigated.

8.2 Experimental method

Two Ta2N coatings were deposited onto polished Ti-6Al-4V substrates by using double cathode glow discharge apparatus equipped with a 99.99% purity Ta target. Before the

133 coating process, all substrates were ultrasonically cleaned for 10 min, each successively in baths of acetone and alcohol, and blown dried. Prior to the deposition, the surfaces of the samples were further cleaned by Ar ion bombardment at −650 V for 10 min to remove surface contamination. The deposition was carried out in an Ar and N2 atmosphere of 99.99% purity. The base pressure was 5×10−4 Pa and deposition gas pressure was kept constant at 35 Pa with an Ar:N2 flux ratio of 20:1. The target/substrate distance was 10 mm; and treatment time was 2 h.

For the full-dense Ta2N coating the following parameters were used: target electrode bias voltage with direct current, -900 V; substrate bias voltage with impulse current, -350 V; substrate temperature, 800°C; while for the nanoporous Ta2N coating, different deposition parameters were used, specifically target electrode bias voltage with direct current, -750 V; substrate bias voltage with impulse current, -300 V; substrate temperature, 700°C. The two samples were designated as FD-Ta2N, for the full-dense

Ta2N coating, and NP-Ta2N, for the nanoporous Ta2N coating.

The phase composition of the as-deposited coatings was analyzed by a PANalytial Empyrean X-ray diffraction (XRD) system equipped with a Cu anode. The analysis was carried out at 40 mA and 45 kV using Cu Kα (λ=1.5406 nm) radiation. The X-ray reflections were collected over of 2 theta values ranging from 15° to 90°. The cross- sectional morphology of the two as-deposited Ta2N coatings was characterized in detail by scanning electron microscopy (SEM, Hitachi S3400). Meanwhile, the microstructure of coating were studied by using a combination of dual-beam focused ion beam microscopy (FIB, XT Nova Nanolab 200) and transmission electron microscopy (TEM, Philips, CM200), the detailed procedures have been described elsewhere [119].

Nanoindentation testing were performed on both the FD-Ta2N and NP-Ta2N coatings with a UMIS (Ultra-Micro Indentation System 2000, CSIRO, Sydney, Australia) workstation equipped with Berkovich diamond indenter with a face angle of 65.3°. Load cycle indentation was used to study the hardness and elastic modulus of both coatings with a maximum load of 400 mN. In order to get better understanding about the influence

134 of nanoporosity on the mechanical properties on the NP-Ta2N coating, partial unloading indentation [216] was also performed during nanoindentation testing. During the partial unloading indentation the loading process was divided into ten separate loading steps. After each loading step to an unloading point (the load where the unloading stage started), the indenter was unloaded to 20% percentage of the unloading point. Then, the indenter was loaded to the next unloading point at a higher load. The maximum load of 400 mN was reached at the 10th loading and unloading cycle. The load-displacement curve at each unloading point was continually recorded. Since the unloading process is totally elastic, the mechanical properties would not be greatly impacted by the repetitive unloading process [144, 145]. For both partial unloading and load cycle indentation, the maximum penetration depth was carefully controlled to be less than 10% of the coating thickness to avoid any influence from the substrate [112].

Moreover, 1000g indents were applied to the two Ta2N coatings using a microindentation system (Durascan, Struers, Denmark) equipped with a Vickers indenter to qualitatively analyze the damage tolerance of the two Ta2N coatings. The indent area and the cross- sectional morphology of the indents were observed by a single beam focused ion beam microscope (FIB, FEI xP200, FEI instrument, Hillsboro, USA)[119]. Then, a TEM sample was sliced from the center of the 1000g indent on the nanoporous Ta2N coating using a dual-beam FIB, to reveal the deformation behavior of this specimen.

135

8.3 Results

Fig. 8.1 X-ray diffraction patterns of the as-deposited Ta2N coatings

According to the X-ray diffraction patterns (Fig. 8.1), only hexagonal Ta2N (reference code: 00-026-0985) can be identified from the two as-deposited Ta2N coatings. For the

NP-Ta2N coating, the peaks match with the reflections from the (100), (101), (102), (110),

(103), (200), (112), (201), (202) planes of hexagonal Ta2N that are observed at 2θ values of 33.8°, 38.7°, 50.6°, 60.4°, 67.0°, 71.3°, 72.8°, 74.2°, and 82.8°, respectively. Since the (002) peak at a 2θ value of 36.5° is overlapped by a strong (101) peak, a broadening of the (101) peak can be observed from the spectrum. Generally, the FD-Ta2N coating

136 exhibited a similar diffraction pattern to NP-Ta2N coating. The only difference is that the

FD-Ta2N coating shows a stronger orientation preference for the (101) plane relative to the NP-Ta2N coating. The positions of diffraction peaks for both NP-Ta2N coating and

FD-Ta2N coating fit well with reference data. Any possible peak shifts caused by residual stresses present in these coatings were not observed.

a b

22 μm 21 μm

20 μm 20 μm

Fig.8.2 Cross-sectional SEM images of (a) NP-Ta2N coating and (b) FD-Ta2N coating.

As shown by the cross-sectional SEM images (Fig. 8.2 (a) and (b)), the thickness values of both the NP-Ta2N coating and the FD-Ta2N coating are about 22 μm and 21 μm, respectively. Both coatings exhibit a compact and continuous cross-sectional structure, no cracks or structural flaws can be observed from these secondary electron micrographs. The interfaces between coatings and substrates are sharp and straight, it appears the two

Ta2N coatings are well bonded to the substrates.

137

(a) (b)

50 nm 50 nm

Fig. 8.3 Cross-sectional TEM bright-field images for (a) the NP-Ta2N coating; (b) the FD-Ta2N coating.

The bright field TEM images (Fig. 8.3) show the microstructure of the two as-deposited

Ta2N coatings close to their outer surface. An intense Ta2N (101) ring in the inset selected area diffraction patterns shown in Fig. 8.3 (a) and (b) provides further evidence that both Ta2N coatings have a strong (101) texture, in agreement with the XRD data. It can be seen from the BF images that both coatings contain similar equiaxed grains with a grain size ranging from 15~30 nm. On closer inspection, it can be found that a large number of nanopores with an average diameter of 2 nm are uniformly distributed throughout on the NP-Ta2N coating, whereas no nanopores can be observed in the FD-

Ta2N coating. The volume fraction of porosity in nanoporous Ta2N coating can be calculated to be ~ 11%. This structural transition from the full-dense to the porous on a micro-scale for the two as-deposited Ta2N coatings is intimately associated with the technical parameters used for deposition. Compared with the NP-Ta2N coating, the FD-

Ta2N coating was prepared at a higher deposition temperature, combined with increased ion bombardment induced by raised negative bias of the substrate, which are responsible for the formation of a coating with a denser microstructure. The higher deposition temperature increases the surface mobility of adatoms, allowing the adatoms to diﬀuse to more equilibrium positions and to overcome self-shadowing eﬀects exerted by the previously deposited atoms. Meanwhile, the increased ion bombardment, through raising the negative bias of the substrate, enhances the nucleation density by forming nucleation sites and also decreases the formation of interfacial voids [100]. Remarkably, the

138 difference in deposition temperature has little influence on the grain size of the two Ta2N coatings.

139

Fig. 8.4 (a) Load-displacement curves for load cycle indentation for NP-Ta2N and FD-Ta2N; (b) Load- displacement curves for both load cycle indentation and partial unloading indentation for the NP-Ta2N coating; (c) elastic modulus calculated from all the unloading points above 50 mN in partial unloading testing indentation for the NP-Ta2N coating.

The corresponding hardness and elastic modulus for the two Ta2N coatings were measured by load cycle indentation testing with a Berkovich diamond indenter with a face angle of 65.3°. The maximum load was 400 mN. 5 indents were applied to each coating. The typical load-displacement curves for NP-Ta2N and FD-Ta2N are shown in Fig. 8.4 (a). The smooth curves, without detectable pop-in or pop-out events, for both loading and unloading stages suggest that cracking and delamination did not occur during the indentation testing for both coatings. The hardness and elastic modulus were calculated from the load-displacement curves based on the Oliver-Pharr method [ref]. Interesingly, the difference between the values of hardness for the two coatings obtained from the load cycle indentation method is negligible. That is, the values obtained were 34

GPa for the NP-Ta2N coating and 35 GPa for the FD-Ta2N coating, but the value of elastic modulus for the nanoporous Ta2N coating was about 10% lower than that for the full-dense Ta2N coating, due to the presence of nanopores in the former coating.

140

3 2 2 Based on the hardness and elastic modulus, the values of H/E, H /E and 1/E H for NP-

Ta2N and FD-Ta2N were calculated. Since NP-Ta2N exhibits a similar hardness, but 3 2 2 lower elastic modulus compared with FD-Ta2N, higher values of H/E, H /E and 1/E H were achieved for NP-Ta2N, indicating potential improvements in wear resistance, deformation resistance and damage tolerance for this nanoporous coating.

In order to achieve a better understanding of the influence of nanoporosity on the mechanical properties of these Ta2N coatings, 5 partial unloading indents were performed on NP-Ta2N coating with a maximum load of 400 mN close to the position where load cycle indents were carried out. Fig. 8.4 (b) shows the average load-displacement curves for partial unloading testing and the corresponding average load cycle indentation curve for the NP-Ta2N coating for comparison. Clearly, the partial unloading curve is smooth at each loading and unloading step without any pop-ins or pop-outs. At the lower loads, the loading components for the partial unloading curves are almost overlapped with the load cycle indentation curve. Then, with increasing contact load, the penetration depth for partial unloading shows a decreasing trend compared with the load cycle indentation at the same load. This suggests an increased resistance to local plastic deformation with a gradually densified structure of this coating during the partial unloading process. Fig. 8.4 (c) shows the scatter graph of elastic modulus calculated from all the unloading points above 50 mN for the five partial unloading indents (the data calculated from lower unloading load were not included here, since there is larger instrument error at lower loads). The solid line represents the elastic modulus for the FD-Ta2N coating acquired in load cycle indentation testing. It can be seen that the values of elastic modulus show a large variation at low loads. With increasing load, the elastic modulus gradually stabilizes at the solid line. The hardness and elastic modulus values were calculated from load- displacement curve at 400mN (Fig. 4 (b)). Both the hardness and elastic modulus increased during the partial unloading testing for NP-Ta2N coating compared with results for load cycle indentation. The average elastic modulus for NP-Ta2N at 400 mN in partial unloading testing was 291 GPa, which is broadly similar to the elastic modulus measured from the FD-Ta2N coating in load cycle indentation, suggesting the decrease in elastic modulus in the NP-Ta2N coating was due to the presence of nanoporosity.

141

a b

3.4 μm

c d

3.9 μm

Fig. 8.5 FIB images for the 1000 g indent on the NP-Ta2N coating (a) plan view image; (b) cross-section;

FIB images for the 1000 g indent on the FD-Ta2N coating (c) plan view image (d) cross-section.

To further evaluate the damage tolerance of the two coatings 1000g indents were introduced into both coatings using a Vickers indenter. It can be seen that edge cracks running parallel to the impression edge can be observed from the plan-view image of the

NP-Ta2N coating (Fig. 8.5 (a)). As shown by the cross-sectional FIB image (Fig. 8.5 (b)) the penetration of these edge cracks below the surface is very shallow. Generally, the NP-

Ta2N coating exhibits a featureless cross-section following indentation; no cracks can be observed beneath the contact surface. However, for the FD-Ta2N coating, except for the edge cracks, a concentric array of cracks, running parallel to each other can also be observed from the indented area (Fig. 8.5 (c)). The cross-sectional FIB image (Fig. 8.5

142

(d)) of the indent shows the existence of inclined cracks beneath the contact surface. The presence of surface cracks and inclined cracks suggest it is easier to introduce cracks and structural failures into the FD-Ta2N coating compared with the NP-Ta2N coating. This observation indicates a higher damage resistance of NP-Ta2N, which is consistent with the calculations based on the E and H values.

The significant surface and subsurface cracking for the FD-Ta2N coating during indentation not only suppresses elastic recovery due to release of strain energy, but also contributes to additional permanent (residual) deformation, manifested by the larger indentation depth at the macro level. The contact depth of the 1000 g indents on the FD-

Ta2N coating was 3.9 μm, about 0.5 μm deeper than that of the NP-Ta2N coating. Ritcher et al. [217] also found that the intergranular cracking within Vickers indents might generate additional displacements, resulting in larger deformation, and hence result in a lower measured hardness values.

143

a b

10 nm

Fig. 8.6 (a) Bright field TEM images for the deformed region of the NP-Ta2N coating (b) enlarged image for the contact surface (c) enlarged image for the area away from the contact surface

As shown by the bright field TEM images from a cross-sectional TEM specimen prepared from the center of 1000g indent for the NP-Ta2N coating (Fig. 8.6 (a)-(c)), it can be seen that the density of nanopores for the coating varies with the depth below the contact surface. At the outermost layer close to the contact surface, the deformed coating shows a denser structure without the presence of any nanocracks after indentation. The density of nanopores at the contact surface was reduced since the coating has evidently been compressed. Almost no nanopores can be observed near the contact surface (Fig. 8.6 (b)). However, as shown in Fig. 8.6 (c), away from the contact surface, again the structure of the NP-Ta2N coating remains nanoporous, and is similar to that of the undeformed region. It appears that rather than acting as an initiation point for the

144 formation of cracks during deformation, the nanopores provide a space to absorb deformation energy and relieve the influence of the external force. When plastic deformation occurred, the nanopores closer to the outer surface of NP-Ta2N coating were compressed and eliminated, and a denser, more compact layer was formed below the contact surface. The thickness of this compact layer would be presumably determined by the magnitude of the contact force. The above observations provide an explanation for why the values of elastic modulus for the NP-Ta2N coating increase with increases in the indent loads.

8.4 Discussion

Clearly, the nanoporous Ta2N coating shows better mechanical properties than the fully- dense coating. Compared with the Ta2N coatings deposited by other deposition methods, this improvement in mechanical properties through the presence of nanoporosity is more significant. Typically, the reported hardness and elastic modulus values for Ta2N coatings are 30~35 GPa and 350~400 GPa, respectively [46, 48]. That means, the calculated 3 2 2 maximum H/E H /E and 1/E H values of those Ta2N coatings are below 0.1, 0.35 GPa, -7 -3 and 2.3^10 GPa , respectively, obviously smaller than those of NP-Ta2N coating in this study, which are 0.13, 0.58 GPa, and 4.4^10-7 GPa-3.

Conventionally, structural failures, such as cracks, normally initiate from pores [218]. The presence of stress concentrations arising from pores that have formed during fabrication will evidently decrease the ductility of nanocrystalline materials, since the crack nucleation and propagation can be dramatically enhanced, causing catastrophic failure at low stress levels [215, 219]. However, the effect of stress concentration can be sharply decreased by reducing the pore size to the nanoscale or controlling pore morphology (e.g. to form near-spherical pores) [220]. Those nearly globular nanopores, with an average diameter of 2 nm, are much smaller than the critical crack size, and thus might not serve as the crack sources. As shown by the TEM analysis, the existence of nanopores did not cause failure in the nanoporous Ta2N coating even under a high load of 1000 g.

145

As is well known, the deformation mechanism will gradually be dominated by grain- boundary shear with decreasing grain size [64], which is similar to the shear banding observed in amorphous or metallic materials [221-223]. Normally, the formation of nanocracks occurs in order to release the accumulation of strain energy at triple junctions caused by sessile dislocations result from grain boundary sliding [219, 224]. Instead of acting as the initial point of a nanocrack [219], the nanopores provide a free space to accommodate the grain boundary plastic strains by the densification and elimination of nanopores and can have a positive impact on the enhanced toughness through blunting the tip of cracks [168]. Moreover, the dispersed nanopores may promote the proliferation of shear bands at low stresses. This is favorable for improved plasticity without sacrificing other mechanical properties [225]. This presumably is the reason why cracks are more readily formed in the fully-dense Ta2N coating during indentation.

Meanwhile, prior studies regarding the effects of porosity on the hardness and elastic modulus of porous solid materials suggest that hardness and elastic modulus have similar dependence on the volume fracture of porosity. In Jang et al.’s study concerning the influence of porosity on the hardness and elastic modulus of ZrO2- Y2O3 coatings [226], the hardness and the elastic modulus values of the coating show a similar decreasing tendency with increasing volume fracture of porosity. Another study investigating the porosity dependence of the mechanical properties of yttria-stabilized tetragonal zirconia polycrystals with a grain size on the micron scale, and a pores typically also a few microns in size, also shows with increasing porosity, the decrease in elastic modulus is linearly correlated to a decrease in hardness [164]. However, the experimental results obtained in this study suggest that for a nanocrystalline coating, with a nanoporous structure, the elastic modulus of coating is more sensitive to nanoporsity than hardness in a microstructure where the pore size is much smaller than grain size. With a similar grain size, ranging 15 to 30 nm, the hardness of the fully dense Ta2N coating is only 3% higher than that of the nanoporous Ta2N coating with an average pore size of 2 nm. In contrast, the elastic modulus of the fully dense Ta2N coating is 15% higher compared with the nanoporous Ta2N coating. As a result of the different sensitivity of hardness and elastic modulus to nanoporosity, higher H/E H3/E2 and 1/E2H values were achieved by the

146 nanoporous Ta2N, indicating that the presence of nanoporosity can greatly improve the wear resistance, deformation tolerance and damage resistance of the nanocrystalline coating.

8.5 Conclusion

In summary, this study suggests that a homogeneous distribution of nanopores will not lead to detrimental structural failure in a nanocrystalline coating under a large external force. On the contrary, the existence of nanoporosity may increase the mechanical properties of a nanocrystalline coating by significantly reducing its elastic modulus, without greatly decreasing the hardness of the coating. That means by exploiting the difference in sensitivity of hardness and elastic modulus to nanoporosity, the mechanical properties of nanocrystalline coating can be tuned and improved by introducing a controlled volume of nanopores.

147

Chapter 9

Conclusion

148

9.0 Conclusion

A number of tantalum nitride and zirconium nitride based coatings were prepared using a double cathode glow discharge deposition technique. A range of studies were then performed on these coatings to investigate a) the microstructure and mechanical properties of tantalum nitride coatings, b) the influence of oxygen on the mechanical properties of these tantalum nitride coatings, c) the bioactivity of tantalum nitride coatings, d) the microstructure and mechanical properties of zirconium nitride coatings, and e) the influence of nanoporosity on the mechanical properties of the nanocrystalline coatings studied in this thesis. The principal conclusions for this thesis are summarized as follows：

1. With increasing nitrogen partial pressure, the tantalum nitride coatings deposited by double cathode glow discharge technique transformed from hexagonal Ta2N, at a low nitrogen partial pressure, to fcc TaN with a higher nitrogen partial pressure. Both tantalum nitride coatings exhibit a homogenous microstructure composed of equiaxed fine grains together with distribution of nanopores.

2. The hardness measured for the hexagonal Ta2N coating was 34 GPa and 25 GPa for the fcc TaN coating; whilst the elastic modulus for these two coatings were 266 GPa and 205 GPa, respectively. A consequence of the relatively high hardness and low elastic modulus of these tantalum nitride coatings is the expectation that they may exhibit high damage resistance and wear resistance compared with similar coatings prepared by other deposition methods.

3. Oxygen impurities deliberately introduced during the deposition process result in the formation of orthorhombic Ta2O5 and amorphous Ta-O-N as well as tantalum nitrides. These phases may act to greatly decrease the hardness and wear resistance of the tantalum nitride coatings. Therefore, residual air and air leakage should be avoided during the deposition of tantalum nitride coatings.

149

4. Tantalum nitride (Ta2N) coatings exhibit very good compatibility with bone-like HAP. After immersion in a simulated body fluid solution for 25 days, the tantalum nitride coating was fully covered by a thick and relatively uniform HAP layer. Good adherence between the tantalum nitride coating and HAP was also obtained. As such, Ta2N coatings are a promising candidate for biomedical applications.

5. Zirconium nitride coatings deposited by the double cathode glow discharge technique exhibited an fcc ZrN phase with a strong (111) orientation. The ZrN coatings on stainless steels exhibited a bimodal microstructure including both fine grains and more elongated coarse grains. In contrast, the ZrN coatings on Ti-Al-V substrates exhibit a more uniform microstructure comprising fine equiaxed grains and together with a distribution of nanopores. With increasing substrate temperature, the hardness of the ZrN coatings was found to slightly increase. All the ZrN coatings exhibit broadly comparable hardness but much lower elastic moduli compared with zirconium nitride coating deposited by other techniques, suggesting a higher wear resistance and damage resistance of the coatings for the coatings prepared by this method.

6. By studying the influence of nanopores on the mechanical properties of nanocrystalline materials. It was found that elastic modulus is more sensitive to the presence of nanoporosity than hardness, especially when the size of nanopores is smaller than the critical size needed for the nucleation of cracks. Instead of causing structure failure under external load, the presence of nanoporosity can significant increase the damage resistance of nanocrystalline coatings by absorbing the deformation and the movement of dislocation along the grain boundaries under external load, and decreasing the elastic modulus of nanocrystalline material without significant influence on its hardness.

Suggestions for future work include performing wear testing and corrosion testing on these transition nitride coatings, in order to get better insight into the wear resistance, and the corrosion resistance of these coatings. The relationship between elastic modulus and the volume fraction of nanopores could also be investigated by modeling approaches. Variations in deposition parameters using the double cathode glow discharge deposition

150 technique should be explored for these coatings in order to get better control on the volume fraction and the size of the nanoporosity present.

151

Reference

1. M. A. Meyer et al., Mechanical properties of nanocrystalline materials. Progress in Materials Science, 2006. 51(4): pp. 427-556. 2. S. C. Tjong et al., Nanocrystalline materials and coatings. Materials Science and Engineering: R: Reports, 2004. 45(1-2): pp. 1-88. 3. S. Zhang et al., Recent advances of superhard nanocomposite coatings: a review. Surface & Coatings Technology, 2003. 167(2-3): pp. 113-119. 4. J. E. Sundgren et al., A review of the present state of art in hard coatings grown from the vapor phase. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films, 1986. 4(5): pp. 2259-2279. 5. Y. X. Leng et al., Biomedical properties of tantalum nitride films synthesized by reactive magnetron sputtering. Thin Solid Films, 2001. 398: pp. 471-475. 6. C. P. Liu et al., Systematic study of the evolution of texture and electrical properties of ZrNx thin films by reactive DC magnetron sputtering. Thin Solid Films, 2003. 444(1-2): pp. 111-119. 7. J. Xu et al., Double glow surface alloying of low carbon steel with electric brush plating Ni interlayer for improvement in corrosion resistance. Surface & Coatings Technology, 2003. 168(2-3): pp. 156-160. 8. K. Lukaszkowicz, Review of Nanocomposite Thin Films and Coatings Deposited by PVD and CVD Technology, in Nanomaterials. 2011, InTech, Croatia. pp. 145-162. 9. M. Cherkaoui et al. Structure, Mechanical Properties, and Applications of Nanocrystalline Materials. in Atomistic and Continuum Modeling of Nanocrystalline Materials. 2009, Springer, USA. pp. 29-52. 10. C. C. Koch, Structural nanocrystalline materials: an overview. Journal of Materials Science, 2007. 42(5): pp. 1403-1414. 11. S. Veprek et al., Industrial applications of superhard nanocomposite coatings. Surface & Coatings Technology, 2008. 202: pp. 5063-5073. 12. H. Holleck, Material selection for hard coatings. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films, 1986. 4(6): pp. 2661-2669. 13. M. Hebbache, Shear modulus and hardness of crystals: density functional calculations. Solid State Communications, 2000. 113: pp. 427-432. 14. S. Veprek, Recent attempts to design new super- and ultrahard solids leads to nano-sized and nano-structured materials and coatings. Journal of Nanoscience & Nanotechnology, 2011. 11: pp. 14-35. 15. T. Spalvins, A review of recent advances in solid film lubrication. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films, 1987. 5: pp. 212-219. 16. P. P. Eklund et al., Transition-metal-nitride-based thin films as novel energy harvesting materials. Journal of Materials Chemistry C, 2016. 17. K. Holmberg et al., Surface stresses and response to loading, in Coatings Tribology: Properties, Mechanisms, Techniques and Applications in Surface Engineering. 2009, Elservier, Oxford, UK. pp. 75-116. 18. S. Veprek et al., Design of ultrahard materials: Go nano! Philosophical Magazine, 2010. 90(31-32): pp. 4101-4115. 19. S. Veprek et al., Mechanical properties of superhard nanocomposites. Surface & Coatings Technology, 2001. 146-147: pp. 175-182.

152

20. S. Veprek et al., Origin of the hardness enhancement in superhard nc-TiN/a-Si3N4 and ultrahard nc-TiN/a-Si3N4/TiSi2 nanocomposites. Philosophical Magazine Letters, 2007. 87(12): pp. 955-966. 21. S. Y. Zhang, TiN coating of tool steels: a review. Journal of Materials Processing Technology, 1993. 39: pp. 165-177. 22. H. C. Barshilia et al., Transition Metal Nitride-Based Nanolayered Multilayer Coatings and Nanocomposited Coatings as Novel Superhard Materials, in Nanostructured Thin Films and Coatings: Mechanical Properties. 2010, CRC Press, USA. pp. 427-480. 23. W. H. Zhang et al., Tribological behavior of TiN and CrN coatings sliding against an epoxy molding compound. Surface & Coatings Technology, 2000. 130: pp. 240-247. 24. K. Y. Chen et al., Alloying effects on elastic properties of TiN-based nitrides. Journal of physics D: Applied physics, 2003. 36: pp. 2725–2729 25. H. C. Barshilia et al., Raman spectroscopy studies on the thermal stability of TiN, CrN, TiAlN coatings and nanolayered TiN/CrN, TiAlN/CrN multilayer coatings. Journal of Materials Research, 2011. 19(11): pp. 3196-3205. 26. S. Y. Yoon et al., Comparison for mechanical properties between TiN and TiAlN coating layers by AIP technique. Journal of Materials Processing Technology 2002. 130-131: pp. 260-265. 27. Y. Yang et al., First-principles calculations of mechanical properties of TiC and TiN. Journal of Alloys & Compounds, 2009. 485(1-2): pp. 542-547. 28. J. E. Sundgren, Structure and properties of TiN coatings. Thin Salid Films, 1985. 128: pp. 21-44. 29. B. Navinsek et al., Industrial applications of CrN (PVD) coatings, deposited at high and low temperatures. Surface & Coatings Technology, 1997. 97: pp. 182-191. 30. Z. H. Han et al., Effect of N2 partial pressure on the microstructure and mechanical properties of magnetron sputtered CrNx films. Surface & Coatings Technology 2003. 162: pp. 189-193. 31. H. Ichimura et al., Mechanical properties of arc-evaporated CrN coatings: Part I - nanoindentation hardness and elastic modulus. Surface & Coatings Technology, 2001. 145: pp. 88-93. 32. W. Y. Ni et al., Effects of the ratio of hardness to Young’s modulus on the friction and wear behavior of bilayer coatings. Applied Physics Letters, 2004. 85(18): pp. 4028-4030. 33. P. Hones et al., Structural and mechanical properties of chromium nitride, molybdenum nitride, and tungsten nitride thin films. Journal of Physics D: Applied Physics, 2003. 36: pp. 1023-1029. 34. I. Milosva et al., Comparison of TiN, ZrN and CrN hard nitride coatings: Electrochemical and thermal oxidation. Thin Solid Films 1997. 303: pp. 246-254. 35. C. Liu et al., EIS comparison on corrosion performance of PVD TiN and CrN coated mild steel in 0.5 N NaCl aqueous solution. Corrosion Science, 2001. 43: pp. 1953-1961. 36. Z.P. Huang et al., Friction behaviour of TiN, CrN and (TiAl)N coatings. Wear, 1994. 173: p. 13-20. 37. Y. L. Su et al., Comparison of tribological behavior of three films-TiN, TiCN and CrN- grown by physical vapor deposition. Wear, 1997. 213: pp. 165-174. 38. Y. Chiba et al., Wear resistance of arc ion-plated chromium nitride coatings. Journal of materials research, 1993. 8: pp. 1109-1115. 39. V. N. Zhitomirskya et al., Vacuum arc deposition and microstructure of ZrN-based coatings. Surface & Coatings Technology, 1997. 94-95: pp. 207-212.

153

40. C. P. Liu et al., Deposition temperature and thickness effects on the characteristics of dc- sputtered ZrNx films. Materials Chemistry & Physics, 2004. 86(2-3): pp. 370-374. 41. Y. Cheng et al., A Study of ZrN/Zr Coatings Deposited on NiTi Alloy by PIIID Technique. Transations on plasma science, 2006. 34: pp. 1105-1108. 42. X. Liu et al., Effect of deposition and annealing temperature on mechanical properties of TaN film. Applied Surface Science, 2011. 258: pp. 1033-1037. 43. M. Takeyama et al., Properties of TaNx films as diffusion barriers in the thermally stable Cu/Si contact systems. Journal of Vacuum Science & Technology B: Microelectronics and Nanometer Structures, 1996. 14: pp. 674-678. 44. M. Grosser et al., The impact of substrate properties and thermal annealing on tantalum nitride thin films. Applied Surface Science, 2012. 258: pp. 2894-2900. 45. S. K. Kim et al., Deposition of tantalum nitride thin films by D.C. magnetron sputtering. Thin Solid Films, 2005. 475: pp. 202-207. 46. R. Westergard et al., Mechanical and tribological characterization of DC magnetron sputtered tantalum nitride thin films. Surface & Coatings Technology, 1997. 97: pp. 779- 784. 47. G. R. Lee et al., Superhard tantalum-nitride films formed by inductively coupled plasma- assisted sputtering. Surface & Coatings Technology, 2007. 201: pp. 5207-5210. 48. N. R. Moody et al., Adhension and fracture of tantalum nitride films. Acta Materialia, 1998. 46: pp. 585-597. 49. C. S. Shin et al., Growth and physical properties of epitaxial metastable cubic TaN (001). Applied physics letters, 1999. 75: pp. 3808-3810. 50. D. Bernoulli et al., Magnetron sputter deposited tantalum and tantalum nitride thin films: An analysis of phase, hardness and composition. Thin Solid Films, 2013. 548: pp. 157-161. 51. K. Valleti et al., Studies on phase dependent mechanical properties of dc magnetron sputtered TaN thin films: evaluation of super hardness in orthorhombic Ta4N phase. Journal of Physics D: Applied Physics, 2008. 41(4): pp. 0454091-0454096. 52. C. C. Koch et al., The inverse Hall-Petch Effect-Fact or Artifact? Mat. Res. Soc. Symp. , 2001. 634: pp. B5.1.1-B5.1.11. 53. M. F. Tambwe et al., Haasen plot analysis of the Hall-Petch effect in Cu/Nb nanolayer composite. J. Mater. Res., 1999. 14: pp. 407-417. 54. J. X. Deng et al., Wear mechanisms of cemented carbide tools in dry cutting of precipitation hardening semi-austenitic stainless steels. Wear, 2011. 270(7-8): pp. 520- 527. 55. J. Musil, Hard and superhard nanocomposite coatings. Surface & Coatings Technology, 2000. 125: pp. 322–330 56. S. Veprek et al., A concept for the design of novel superhard coatings. Thin Solid Films, 1995. 268: pp. 64-71 57. S. Veprek et al., Mechanical properties of superhard nanocomposites. Surface & Coatings Technology, 2001. 146-147: pp. 175-182. 58. S. Veprek et al., Novel thermodynamically stable and oxidation resistant superhard coating materials. Surface and Coatings Technology 1996. 86-87: pp. 394--401 59. J. Yang et al., Is Osmium Diboride An Ultra-Hard Material? Journal of the American Chemical Society, 2008. 130: pp. 7200–7201. 60. Y. Zhang et al., Strain dependent bonding in solid C3N4: High elastic moduli but low strength. Physical Review B, 2006. 73(6) : pp. 0641091-0641094. 61. J. Musil et al., Relationships Between Hardness, Young's Modulus and Elastic Recovery in Hard Nanocomposite Coatings. Surface & Coatings Technology, 2002. 154: pp. 304-313.

154

62. Z.B. Qi et al., The inverse Hall–Petch effect in nanocrystalline ZrN coatings. Surface & Coatings Technology, 2011. 205(12): pp. 3692-3697. 63. S. Yip, Nanocrystalline metals Mapping plasticity. nature materials, 2004. 3: pp. 11-12. 64. A. S. Argon et al., The strongest size. Philosophical Magazine Letters, 2006. 86(11): pp. 713-720. 65. P. Barai et al., Mechanics of very fine-grained nanocrystalline materials with contributions from grain interior, GB zone, and grain-boundary sliding. International Journal of Plasticity, 2009. 25(12): pp. 2410-2434. 66. S. Yip, The strongest size. Nature, 1998. 391: pp. 532-533. 67. K. L. Rutherford et al., A micro-abrasive wear test, with particular application to coated systems. Surface & Coatings Technology, 1996. 79: pp. 231-239. 68. S. Zhang et al., Toughness evaluation of hard coatings and thin films. Thin Solid Films, 2012. 520(7): pp. 2375-2389. 69. A. Leyland et al., On the significance of the H/E ratio in wear control: a nanocomposite coating approach to optimised tribological behaviour. Wear, 2000. 246: pp. 1–11 70. M. E. Roy et al., Diamond-like carbon coatings enhance the hardness and resilience of bearing surfaces for use in joint arthroplasty. Acta Biomater, 2010. 6: pp. 1619-1624. 71. A. Miserez, J.C.W., P. J. Thurner, J. Aizenberg, Y. Dauphin, P. Fratzl, D. E. Morse, F. W. Zok, Effects of Laminate Architecture on Fracture Resistance of Sponge Biosilica: Lessons from Nature. Advanced Functional Materials, 2008. 18(8): p. 1241-1248. 72. P. M. Martin, Handbook of Deposition Technologies for Films and Coatings. 2010, USA, Elsevier Inc. 73. K. Komvopoulos, Surface engineering and microtribology for microelectromechanical systems. Wear, 1996. 200: pp. 305-327. 74. B. North, Six issues for the hard coatings community. Surface & Coatings Technology 1998. 106: pp. 129–134. 75. J. O. Carlsson et al., Chemical Vapor Deposition, in Handbook of Deposition Technologies for Films and Coatings (Third Edition), P. M. Martin, Editor. 2009, Elsevier Inc: Burlington, USA. pp. 314-363. 76. W. G. Lee et al., Preparation and properties of amorphous TiO2 thin films by plasma enhanced chemical vapor deposition. Thin Solid Films, 1994. 237: pp. 105-111. 77. S. Veprek, Conventional and new approaches towards the design of novel superhard materials. Surface & Coatings Technology 1997. 97: pp. 15-22. 78. L. Martinu et al., Plasma-Enhanced Chemical Vapor Deposition of Functional Coatings. In Handbook of Deposition Technologies for Films and Coatings, 2009, Burlington, USA, Elsevier Inc. 79. S. I. Shah et al., Evaporation: Processes, Bulk Microstructures, and Mechanical Properties, in Handbook of Deposition Technologies for Films and Coatings (Third Edition), P. M. Martin, Editor. 2009, Elsevier Inc.: Burlington, USA. pp. 135-252. 80. D. M. Mattox, Handbook of Physical Vapor Deposition (PVD) Processing. 2010, Oxford, UK, Elsevier Inc. 81. N. Savvides, Ion-assisted deposition and metastable structures. Thin Solid Film, 1988, 163: pp. 3-32. 82. J. J. Cuomo et al., Sputter deposition of dense diamond-like carbon films at low temperature. Applied Physics Letters, 1991, 58(5): pp. 466. 83. D. M. Mattox, Ion plating: past, present and future. Surface & Coatings Technology 2000, 133-134: pp. 517-521.

155

84. Z. Xu, A Novel Plasma Surface Metallurgy: Xu-Tec Process Surface & Coatings Technology, 1990. 43/44: pp. 1065-1073. 85. J. Xu et al., Ni-based superalloy surface alloying by double-glow plasma surface alloying technique. Vacuum, 2004. 72(4): pp. 489-500. 86. C. H. Zhou et al., Reactive sputter deposition of alumina films on magnesium alloy by double cathode glow-discharge plasma technique. Materials Characterization, 2010. 61(2): pp. 249-256. 87. S. O. Chuan et al., Analysis of Ti base hard coating performance in machining process: a review. Journal of Applied Sciences, 2012. 12: pp. 1882-1890. 88. J. A. Thornton, High rate thick film growth. Annu. Rev. Mater. Sci., 1977. 7: p. 239-260. 89. H. E. Hintermann, Adhesion, Fabriction and Wear of Thin Hard Coatings. Wear, 1984. 100 pp. 381 - 397. 90. J. Destefani, Cutting Tool 101: Coatings. Manufacturing Engineering, 2002. 129(4): pp. 47-56. 91. Sproul, W.D., Multilayer, multicomponent, and multiphase physical vapor deposition coatings for enhanced performance. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films, 1994. 12(4): p. 1595. 92. A. T. Santhanam et al., Comparison of the steel-milling performance of carbide inserts with MTCVD and PVD TiCN coatings. International Journal of Refractory Metals & Hard Materials, 1996. 14(1-3): pp. 31-40. 93. R. Yigit, E.C., F. Findik, S. Koksal, Tool life performance of multilayer hard coatings produced by HTCVD for machining of nodular cast iron. International Journal of Refractory Metals & Hard Materials, 2008. 26(6): p. 514-524. 94. P. M. Martin, Deposition Technologies An Overview, in Handbook of Deposition Technologies for Films and Coatings, P.M. Martin, Editor. 2009, Elsevier Inc: Burlington, USA. pp. 1-31. 95. J. G. Swadener et al., Measurement of residual stress by load and depth sensing indentation with spherical indenters Journal of Materials Research, 2001. 16. 96. S. Veprek et al., Different approaches to superhard coatings and nanocomposites. Thin Solid Films, 2005. 476(1): pp. 1-29. 97. W. Herr et al., The influence of a heat treatment on the microstructure and mechanical properties of sputtered coatings. Surface & Coatings Technology, 1997. 97(1-3): pp. 335- 340. 98. J. Musil et al., New Results in Reactive Magnetron Deposittion of TiN Films. Thin Solid Films, 1988. 167: pp. 107-I 19. 99. D. M. Mattox, Particle bombardment effects on thin-film deposition: A review. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films, 1989. 7(3): pp. 1105- 1114. 100. P. H. Mayrhofer et al., Structure–property relationships in single- and dual-phase nanocrystalline hard coatings. Surface & Coatings Technology, 2003. 174-175: p. 725- 731. 101. F. A. Sarott et al., Effect of substrate bias on the properties of microcrystalline silicon films. Solid State Communications, 1982. 42(6): pp. 465-468. 102. Q. M. Wang et al., Microstructural control of Cr–Si–N films by a hybrid arc ion plating and magnetron sputtering process. Acta Materialia, 2009. 57(17): pp. 4974-4987. 103. S. Veprek et al., Limits to the preparation of superhard nanocomposites: Impurities, deposition and annealing temperature. Thin Solid Films, 2012. 522: pp. 274-282.

156

104. S. Veprek et al., A. Niederhofer, D. Ma, S. Mukherjee, Degradation of superhard nanocomposites by built-in impurities. Journal of Vacuum Science & Technology B: Microelectronics & Nanometer Structures, 2004. 22(2): pp. L5-L9.

105. T. Fu, Y.Z., Y. G. Shen Interface structure of sputter deposited CNx film on silicon substrate. Materials Letters, 2008. 62(17-18): p. 2685-2687. 106. M. G. Méndez et al., Characterization of AlN Thin Films Fabricated by Reactive DC Sputtering: Experimantal Meansurement and Huckel Calculation. International Journal of Modern Physics B, 2009. 23(9): pp. 2233-2251. 107. H. Liu et al., Characterization of (Al, Si)N films deposited by balanced magnetron sputtering. Thin Solid Films, 2009. 517(21): pp. 5988-5993. 108. J. W. Lee et al., Microstructure, mechanical and electrochemical properties evaluation of pulsed DC reactive magnetron sputtered nanostructured Cr–Zr–N and Cr–Zr–Si–N thin films. Surface & Coatings Technology, 2010. 205(5): pp. 1331-1338. 109. J. Goldstein et al., Scanning Electron Microscopy. 2002, New York, USA, Springer. 110. H. Schatten, Scanning Electron Microscopy for the life sciences. 2013, New York, USA, Cambridge University Press. 111. L. Reimer, Transmission Electron Microscopy. Springer Series in Optical Sciences. 1989, New York, USA, Springer. 112. P. C. Wo et al., Factors governing the mechanical behaviour of CrSiN coatings: Combined nanoindentation testing and transmission electron microscopy. Materials Science & Engineering A, 2012. 534: pp. 297-308. 113. D. B. Williams et al., Transmission Electron Microscopy A Textbook for Materials Science. 2009, New York, USA: Springer. 114. Z. L. Wang et al., Electron Microscopy Techniques for Imaging and Analysis of Nanoparticles, in Developments in Surface Contamination and Cleaning Fundamentals and Applied Aspects. 2008, William Andrew Publishing: New York, USA. pp. 531-584. 115. M. Rühle et al., Transmission electron microscopy, in Physical Metallurgy, 1996, Elsevier Ltd, Burlington, USA. 116. B. Fultz et al., Transmission Electron Microscopy and Diffractometry of Materials Forth Editor. 2001, New York, USA, springer. 117. S. Reyntjens et al., A review of focused ion beam applications in microsystem technology. Journal of Micromechanics & Microengineering, 2001. 11: pp. 287-300. 118. I. Utke et al., Gas-assisted focused electron beam and ion beam processing and fabrication. Journal of Vacuum Science & Technology B: Microelectronics and Nanometer Structures, 2008. 26(4): pp. 1197-1276. 119. P. R. Munroe, The application of focused ion beam microscopy in the material sciences. Materials Characterization, 2009. 60: pp. 2-13. 120. P. C. Wo et al., A novel technique for microstructure characterization of garnet films. Optical Materials, 2009. 32(2): pp. 315-322. 121. J. Swartz, The availability of ESCA instrumentation and the qualitative and quantitative nature of the data produced promise much for future applications in chemical research and analytical problem solving. Analytical Chemistry 1973. 45(9): pp. 788-800. 122. A. A. Mel et al., XPS study of the surface composition modification of nc-TiC/C nanocomposite films under in situ argon ion bombardment. Thin Solid Films, 2011. 519(12): pp. 3982-3985. 123. J. F. Watts, An introduction to surface analysis by XPS and AES. 2003, UK: WILEY. 124. J. F. Watts, X-ray photoelectron spectroscopy Vacuum, 1994. 45: pp. 653-671. 125. R. Guinebretière, X-ray Diffraction by Polycrystalline Materials. 2007, CA, USA: ISTE Ltd.

157

126. A. K. Chatterjee, X-Ray Diffraction, in Handbook of Analytical Techniques in Concrete Science and Technology, J.J.B. V.SRamachandran, Editor. 2001, William Andrew Publishing: New York, USA. pp. 275-332. 127. R. E. Smallman et al., Physical Metallurgy and Advanced Materials Engineering (Seventh Edition). 2007, Burlington, USA: Elsevier Ltd. 128. M. I. Jones et al., Effect of substrate preparation and deposition conditions on the preferred orientation of TiN coatings deposited by RF reactive sputtering. Surface & Coatings Technology, 2000. 132: pp. 143-151. 129. S. Koutsopoulos, Synthesis and characterization of hydroxyapatite crystals: A review study on the analytical methods. Journal of biomedical materials research, 2002. 62: pp. 600-612. 130. R. L. McCreery, Raman Spectroscopy for Chemical Analysis, ed. J.D. Winefordner. 2000, USA: Wiley interscience. 131. S. Veprek, The search for novel, superhard materials. Journal of Vacuum Science & Technology A, 1999. 17(5): pp. 2401-2421. 132. W. C. Oliver et al., An improve technique for determining hardness and elastic modulus using load and displacement sensing indentation experiments. Journal of Materials Research, 1992. 7: pp. 1564-1583. 133. G. M. Pharr et al., On the generality of the relationship among contact stiffness, contact area, and elastic modulus during indentation. Journal of Materials Research, 1991. 7(3): pp. 613-617. 134. Z. H. Xie et al., Deformation mechanisms of TiN multilayer coatings alternated by ductile or stiff interlayers. Acta Materialia, 2008. 56(4): pp. 852-861. 135. T. J. Bell et al., The Determination of Surface Plastic and Elastic Properties by Ultra Micro-indentation. Metrologia, 1991. 28: pp. 463-469. 136. J. Menčík et al., Design of Contact Loaded Coated Components Using Data From Depth Sensing Micronindentation Tests. Surface Engineering, 1997. 13(6): pp. 498-504. 137. R. A. Mirshams et al., Correlation of nanoindentation measurements of nickel made using geometrically different indenter tips. Acta Materialia, 2006. 54(4): pp. 1123-1134. 138. A. J. Haq, Deformation Behaviour of Diamond-like Carbon Coatings on Silicon Substrates A Thesis by Ayesha Jabeen, in School of Material Science & Engineering, 2008, UNSW. 139. G. M. Pharr, Measurement of mechanical properties by ultra-low load indentation. Materials Science and Engineering A, 1998. 253: pp. 151–159. 140. W. D. Nix et al., Indentation Size Effects In Crystalline Materials: a Law for Strain Gradient Plasticity. Journal of the Mechanics & Physics of Solids, 1998. 46(3): pp. 411- 425. 141. A. O. Ibidunni, The Oxidation and Resistance of Tantalum-Nitride Thin Films. Oxidation of Metals, 1993. 40: pp. 5-20. 142. S. V. Hainsworth et al., The effect of coating cracking on the indentation response of thin hard-coated systems. Surface & Coatings Technology, 1998. 102(1-2): pp. 97-107. 143. M. V. Swain et al., Mechanical property characterization of thin films using spherical tipped indenters. Thin Solid Films, 1994. 253: pp. 204-211. 144. A. J. Bushby, Nano-Indentation Using Spherical Indenters. Nondestructive Testing & Evaluation, 2001. 17(4-5): pp. 213-234. 145. T. J. Bell et al., The Determination of Surface Plastic and Elastic Properties by Ultra Micro-indentation. Metrologia, 1992. 28: pp. 463-469. 146. B. R. Levine et al., Experimental and clinical performance of porous tantalum in orthopedic surgery. Biomaterials, 2006. 27(27): pp. 4671-81.

158

147. A. Fukuda et al., Bone bonding bioactivity of Ti metal and Ti-Zr-Nb-Ta alloys with Ca ions incorporated on their surfaces by simple chemical and heat treatments. Acta Biomater, 2011. 7(3): pp. 1379-86. 148. J. Xu et al., The nature and role of passive films in controlling the corrosion resistance of MoSi2-based nanocomposite coatings. Journal of Materials Chemistry A, 2013. 1: pp. 10281-10291. 149. L. L. Liu et al., The roles of passive layers in regulating the electrochemical behavior of Ti5Si3-based nanocomposite films. Journal of Materials Chemistry A, 2013. 1: pp. 2064- 2078. 150. J. Xu et al., Novel high damage-tolerant, wear resistant MoSi2-based nanocomposite coatings. Applied Surface Science, 2013. 270: pp. 418-427. 151. C. C. Chang et al., Microstructural and electrical characteristics of reactively sputtered Ta-N thin films. Thin Solid Films, 2002. 413: pp. 46-51. 152. A. Leyland et al., Design criteria for wear-resistant nanostructured and glassy-metal coatings. Surface & Coatings Technology, 2004. 177-178: pp. 317-324. 153. K. Valleti et al., Studies on phase dependent mechanical properties of dc magnetron sputtered TaN thin films: evaluation of super hardness in orthorhombic Ta4N phase. Journal of Physics D: Applied Physics, 2008. 41(4): pp. 45409-45415. 154. J. Xu et al., Connecting structural, mechanical and tribological characteristics of Al alloyed nanocrystalline molybdenum silicide coatings. Journal of Physics D: Applied Physics, 2013. 46(6): pp. 65304-65320. 155. K. Valleti et al., Studies on phase dependent mechanical properties of dc magnetron

sputtered TaN thin films: evaluation of super hardness in orthorhombic Ta4N phase. Journal of Physics D: Applied Physics, 2008. 41: p. 1-6. 156. J. Pelleg et al., Reactive-Sputter-Depositied TiN Films on Glass Substrates. Thin Solid Films, 1991. 197: pp. 117-128. 157. U. C. Oh et al., Effects of strain energy on the preferred orientation of TiN thin films. Journal of Applied Physics, 1993. 74: pp. 1692-1696. 158. D. H. Lai et al., Structure-property-performance of nanocomposite Cr–Si–N films synthesized by a duplex surface treatment technique. Journal of Materials Research, 2011. 26: pp. 3020-3031. 159. J. Xu et al., Mechanical and corrosion-resistant properties of Ti–Nb–Si–N nanocomposite films prepared by a double glow discharge plasma technique. Ceramics International, 2014. 40: pp. 8621-8630. 160. J. W. Lee et al., The effects of pulse frequency and substrate bias to the mechanical properties of CrN coatings deposited by pulsed DC magnetron sputtering. Thin Solid Films, 2006. 494: pp. 161-167. 161. C. H. Ma et al., Texture evolution of transition-metal nitride thin films by ion beam assisted deposition. Thin Solid Films, 2004. 446(2): pp. 184-193. 162. Z. B. Qi et al., The inverse Hall–Petch effect in nanocrystalline ZrN coatings. Surface & Coatings Technology, 2011. 205(12): pp. 3692-3697. 163. K. H. Min, Comparative study of tantalum and tantalum nitrides (Ta2N and TaN) as a diffusion barrier for Cu metallization. Journal of Vacuum Science & Technology B: Microelectronics and Nanometer Structures, 1996. 14(5): pp. 3263-3269. 164. J. Luo et al., Porosity-dependence of elastic moduli and hardness of 3Y-TZP ceramics. Ceramics International, 1999. 25: pp. 281-286. 165. C. S. Shin et al., Growth and physical properties of epitaxial metastable cubic TaN (001). Applied physics letters, 1999. 75: pp. 3808-3810.

159

166. C. S. Shin et al., Epitaxial NaCl structure δ-TaNX (001): Electronic transport properties, elastic modulus, and hardness versus N/Ta ratio. Journal of Applied Physics, 2001. 90: pp. 2879-2885. 167. H. B. Nie et al. Liew Structural and electrical properties of tantalum nitride thin films fabricated by using reactive radio-frequency magnetron sputtering. Applied Physics A Materials Science & Processing, 2001. 73: pp. 229-236. 168. K. M. Reddy et al., Enhanced mechanical properties of nanocrystalline boron carbide by nanoporosity and interface phases. Nature Communication, 2012. 3: pp. 1052-1058. 169. T. Nakamura et al., Effects of Pores on Mechanical Properties of Plasma-Sprayed Ceramic Coatings. Journal of the American Ceramic Society, 2000. 83: pp. 578-584. 170. A. O. Ibidunni, The Oxidation and Resistance of Tantalum-Nitride Thin Films. Oxidation of Metals, 1993. 40: pp. 5-20. 171. A. O. Ibidunni et al., Characterization of the Oxidation of Tantalum Nitride. Surface & Interface Analysis, 1993. 20: pp. 559-564. 172. I. V. Sieber et al., Porous Tantalum Oxide Prepared by Electrochemical Anodic Oxidation. Journal of the Electrochemical Society, 2005. 152(9): pp. C639-C644. 173. H. Demiryont et al., Effects of oxygen content on the optical properties of tantalum oxide films deposited by ion-beam sputtering. Applied optics, 1985. 24(4): pp. 490-495. 174. A. Ishihara et al., Tantalum (oxy)nitrides prepared using reactive sputtering for new nonplatinum cathodes of polymer electrolyte fuel cell. Electrochimica Acta, 2008. 53(16): pp. 5442-5450. 175. W. R. Matizamhuka et al., Preparation, Hardness, and Fracture Toughness of Tantalum Oxynitride Ceramics. Journal of the American Ceramic Society, 2007. 90(7): pp. 2280- 2282. 176. O. B. Shcherbina et al., Mechanical properties of Nb2O5 and Ta2O5 prepared by different procedures. Inorganic Materials, 2012. 48(4): pp. 433-438. 177. Y. N. Wu et al., First-principles studies of Ta2O5 polymorphs. Physical Review B, 2011. 83(14) : pp. 1441051-1441059. 178. H. Szymanowski et al., Optical properties and microstructure of plasma deposited Ta2O5 and Nb2O5 films. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films, 2005. 23(2): pp. 241-247. 179. A. R. Day et al., The elastic moduli of a sheet containing circular holes. Journal of the Mechanics & Physics of Solids, 1992. 40: pp. 1031-1051. 180. O. Yeheskel et al., Elastic moduli of grain boundaries in nanocrystalline MgO ceramics. Journal of Materials Research, 2011. 20(03): pp. 719-725. 181. M. Uchida et al., Apatite formation on zirconium metal treated with aqueous NaOH. Biomaterials 2002. 23: pp. 313-317. 182. T. Miyazaki et al., Mechanism of bonelike apatite formation on bioactive tantalum metal in a simulated body fluid. Biomaterials, 2002. 23: pp. 827-832. 183. T. Kokubo et al., Novel bioactive materials with different mechanical properties. Biomaterials, 2003. 24(13): pp. 2161-2175. 184. T. Kokubo, Apatite formation on surfaces of ceramics, metals, and polymers in body environment. Acta materialia, 1998. 46: pp. 2519-2527. 185. T. Kokubo et al., How useful is SBF in predicting in vivo bone bioactivity? Biomaterials, 2006. 27(15): pp. 2907-2915. 186. J. Mahamid et al., Amorphous calcium phosphate is a major component of the forming fin bones of zebrafish: Indications for an amorphous precursor phase. Proceedings of the National Academy of Sciences, 2008. 105(35): pp. 12748-53

160

187. L. Cleries et al., Dissolution behaviour of calcium phosphate coatings obtained by laser ablation. Biomaterials, 1998. 19: pp. 1483-1487. 188. C. Combes et al., Amorphous calcium phosphates: synthesis, properties and uses in biomaterials. Acta Biomater, 2010. 6(9): pp. 3362-3378. 189. J. Xu et al., Promoting bone-like apatite formation on titanium alloys through nanocrystalline tantalum nitride coatings. Journal of Materials Chemistry B, 2015. 3(19): pp. 4082-4094. 190. M. Weinlaender et al., Raman microprobe investigation of the calcium phosphate phases of three commercially available plasma-flame-sprayed hydroxyapatitecoated dental implants. Journal of Materials Science: Materials in Medicine, 1992. 3: pp. 397-401. 191. E. J. Harju et al., Optimization of Wear and Corrosion Resistance of Triode-Ion-Plated Nitride Coatings. Surface & Coatings Technology, 1990. 41: p. 157-166. 192. C. Turan et al., Contact resistance characteristics of coated metallic bipolar plates for PEM fuel cells – investigations on the effect of manufacturing. International Journal of Hydrogen Energy, 2012. 37(23): pp. 18187-18204. 193. J. Xu et al., A ZrN nanocrystalline coating for polymer electrolyte membrane fuel cell metallic bipolar plates prepared by reactive sputter deposition. RSC Advances, 2015. 5(82): pp. 67348-67356. 194. J. Xu et al., A nanocrystalline zirconium carbide coating as a functional corrosion- resistant barrier for polymer electrolyte membrane fuel cell application. Journal of Power Sources, 2015. 297: pp. 359-369. 195. Temperature-Dependent Phase Transitions of ZrO2. Materials Design Application Note, 2009. 196. A. Rizzo et al., Synthesis and characterization of titanium and zirconium oxynitride coatings. Thin Solid Films, 2009. 517(21): pp. 5956-5964. 197. H. Jiménez et al., Effect of the substrate temperature in ZrN coatings grown by the pulsed arc technique studied by XRD. Surface & Coatings Technology, 2006. 201(3-4): pp. 1594-1601. 198. J. H. Huang et al., Evaluation of fracture toughness of ZrN hard coatings by internal energy induced cracking method. Surface & Coatings Technology, 2014. 258: pp. 211- 218. 199. M. M. Larijani et al., Structural and mechanical properties of ZrN films prepared by ion beam sputtering with varying N2/Ar ratio and substrate temperature. Vacuum, 2006. 81(4): pp. 550-555. 200. D. Roman et al., Effect of deposition temperature on microstructure and corrosion resistance of ZrN thin films deposited by DC reactive magnetron sputtering. Materials Chemistry & Physics, 2011. 130(1-2): pp. 147-153. 201. Y. C. Chieh et al., Microstructure evolution of ZrN films annealed in vacuum. Surface & Coatings Technology, 2006. 200(10): pp. 3336-3340. 202. C. V. Thompson, Grain growth in thin films. Annual Review of Materials Research , 1990. 20: pp. 245-268. 203. M. K. Hibbs et al., Effects of substrate temperature and substrate material on the structure of reactively sputtered TiN film. Thin Solid Films, 1984. 122: pp. 115-129. 204. M. A. Lopes et al., Glass-reinforced hydroxyapatite composites: fracture toughness and hardness dependence on microstructural characteristics. Biomaterials, 1999. 20: pp. 2085-2090.

161

205. C. S. Chen et al., Influence of growth temperature on microstructure and mechanical properties of nanocrystalline zirconium carbide films. Thin Solid Films, 2005. 479(1-2): pp. 130-136. 206. S. Hsu et al., Mechanical behaviour of cadmium-boron and cadmium- tungsten particulate composites. Journal of Materials science, 1977. 12: pp. 131-140. 207. L. C. Chang et al., Mechanical properties and oxidation behavior of ZrNx thin films fabricated through high-power impulse magnetron sputtering deposition. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films, 2016. 34(2): pp. 02D1071-02D1077. 208. M. A. Auger et al., Hardness and tribology measurements on ZrN coatings deposited by reactive sputtering technique. Vacuum, 2007. 81(11-12): pp. 1462-1465. 209. A. Mitsuo et al., Mechanical properties of zirconium films prepared by ion-beam assisted deposition. Nuclear Instruments & Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 2003. 206: pp. 366-370. 210. G. López et al., High-temperature tribological characterization of zirconium nitride coatings. Surface & Coatings Technology, 2005. 200(7): pp. 2092-2099. 211. A. Kobayashi, Formation of high hardness zirconia coatings by gas tunnel type plasma spraying. Surface & Coatings Technology, 1997. 90: pp. 197-202. 212. Y. W. Bao et al., Investigation of the relationship between elastic modulus and hardness based on depth-sensing indentation measurements. Acta Materialia, 2004. 52(18): pp. 5397-5404. 213. M. Dao et al., Toward a quantitative understanding of mechanical behavior of nanocrystalline metals. Acta Materialia, 2007. 55(12): pp. 4041-4065. 214. V. D. Krstic, Porosity dependence of strength in brittle solids. Theoretical & Applied Fracture Mechanics, 1988. 10: pp. 241-247. 215. A. S. Wagh et al., Dependence of ceramic fracture properties on porosity. Journal of Materials Science, 1993. 28: pp. 3589-3593. 216. J. S. Field et al., A simple predictive model for spherical indentation. Journal of Material Research, 1993. 8: pp. 297-306. 217. V. Richter et al., On hardness and toughness of ultrafine and nanocrystalline hard materials. International Journal of Refractory Metals & Hard Materials, 1999. 17: pp. 141-152. 218. R. W. Rice, Grain size and porosity dependence of ceramic fracture energy and toughness at 22 ℃. Journal of Material Science, 1996. 31: pp. 1969-1983. 219. I. A. Ovid'ko, Review on the Fracture Process in Nanocrystalline Materials. Journal of Material Science, 2007. 42: pp. 1694-1708. 220. S. Zhang et al., Toughening of hard nanostructural thin films: a critical review. Surface & Coatings Technology, 2005. 198(1-3): pp. 2-8. 221. H. Conrad et al. Mechanisms for grain size hardening and softening in Zn. Acta Materialia, 2002. 50: pp. 5067–5078 222. G. He et al., Novel Ti-base nanostructure-dendrite composite with enhanced plasticity. Nature materials, 2003. 2(1): pp. 33-7. 223. M. Hatherly, A.S.M., Shear Bands in deformed metals. Scripta Materialia, 1984. 18: p. 449-454. 224. I. Ovid'ko, Triple junction nanocracks in deformed nanocrystalline materials. Acta Materialia, 2004. 52(5): pp. 1201-1209. 225. T. Wada et al., Enhancement of room-temperature plasticity in a bulk metallic glass by finely dispersed porosity. Applied Physics Letters, 2005. 86(25): pp. 2519071-2159073.

162

226. B. K. Jang et al., Influence of porosity on hardness and Young's modulus of nanoporous EB-PVD TBCs by nanoindentation. Materials Letters, 2005. 59(27): pp. 3462-3466.

163